分子动力学模拟技术在生物分子研究中的进展

曹了然 张春煜 张鼎林 楚慧郢 张跃斌 李国辉,*

(1中国科学院大连化学物理研究所分子模拟与设计研究组，分子反应动力学国家重点实验室，辽宁 大连 116023；2辽河油田总医院，辽宁 盘锦 124010)

分子动力学模拟技术在生物分子研究中的进展

曹了然1,#张春煜2,#张鼎林1,#楚慧郢1张跃斌1李国辉1,*

(1中国科学院大连化学物理研究所分子模拟与设计研究组，分子反应动力学国家重点实验室，辽宁 大连 116023；2辽河油田总医院，辽宁 盘锦 124010)

在当今科学探索中，分子动力学模拟作为实验的辅助，替代甚至引导，有越来越大的重要性。在这篇综述里，我们概括介绍了分子模拟领域的发展历史，特别是专注于近期的进展，包括新的分子力场，增强采样技术以及在研究生物分子体系方面的最新成果。

分子动力学模拟；分子力场；增强采样；酶促反应

1 Introduction

Ever since the first modeling of elastic collision between rigid spheres1, the molecular simulation technique has been greatly developed to achieve atomic-level information and widely used in both material science and studying biological systems, including proteins, nucleic acids, lipid membranes.Due to the increasing requirement on a deep understanding of not only the static structure of the target system but also its functionality such as how protein and ligand bind, how enzyme reactions are performed, how does a protein generate its three dimensional (3D) structure, etc. In order to give a reasonable explanation of those problems, an accurate but not too time-consuming modeling of those systems is necessary.Besides, to study those complex biochemical processes that usually have time scales beyond the calculation capabilities of current computational resources, more effective strategies of sampling in the conformational space are developed to achieve a balance between simulation time and sampling integrity.Thanks to all those progress, molecular dynamic simulation can be used to explore the mechanisms of biological processes, and by that means it is possible to build the intrinsic relationship between the microscopic molecule structure and the corresponding macroscopic biological behaviors.

In this review, a brief introduction on the history of the development in molecular force field and sampling schemes will be proposed. Due to the fact that numerous works has been published in this field, this review will be focused on recent published work and new technologies. Other than those,this review will also discuss recent important works of using molecular dynamics simulation approaches to study biological processes.

2 Development of force field

Computational treatment of molecular dynamics is based on inter-atomic forces, which can be derived by solving the Schrӧdingerʹs equation and related approaches are categorized as quantum mechanical methods. However, high calculation costs limit the use of those approaches to relatively simple systems. Molecular dynamics simulation, on the other hand,builds an empirical function to model the potential energy of the system. The function can be constructed via estimating the intermole- cular interaction energies from isolated monomer wave functions (namely the perturbative method) or via the energy differences between isolated monomers and corresponding dimer (namely the super-molecular calculations)2. For different atoms or atoms in different environments, parameter sets are introduced as the variables in the potential energy function. Due to the diversity of interatomic interactions in biological systems, as well as the complex electrostatic environments, it is challenging to build a set of parameter which can model the motions of atoms in different situations.

ZHANG Chun-Yu, was a visiting scholar in DICP. Her research focuses on biomedicine.

ZHANG Ding-Lin, received his Master degree from Jilin University in 2007. He is currently an assistant researcher in Prof. LI Guo-Huiʹs group since 2009. His research focuses on computational biology.

LI Guo-Hui, received his Ph.D degree from Dalian Institute of Chemical Physics in 2000, and is now working as a professor there. His research focuses on systematic methodology developments and applications on the mechanical analysis and theoretical predictions of biomolecular structures and functions,including coarse-grained model with all atomistic accuracy,polarizable molecular models and simulation strategy with a combination of software and hardware acceleration methods.

2.1 Additive force field

The classical way to estimate the interatomic interactions includes treating atoms as rigid spheres with fixed charges located on the nucleus. Electron distribution among bonded atoms based on electronegativity can be empirically illustrated as a partial charge, either positive or negative, on each atom.Thus, each atom responds to the surrounding electrostatic environment in an average way (mean-field approximation).From this simple treatment, interatomic electrostatic energies can be simply estimated via the Coulomb’s law, and the total electrostatic energy of the system is the summation of those pairwise energies. Then the complete potential energy function can be estimated via a summation of bonded energy terms(including bond lengths, bond angles, dihedral angles, and may also include improper dihedrals and other empirical correction terms), van der Waals interaction term (usually described by the Lennard-Jones potential), and the electrostatic energy. Dependent on the targeting system, different force fields have been built. To accurately model the properties of small organic or inorganic molecules such as metals, crystals, polymers and nanoparticles in materials, force fields such as CFF3−6, MM37,MMFF948−10, UFF11, DREIDING12are implemented, while for the dynamics of macromolecules, a number of widely used force fields have been built, including AMBER13,14,CHARMM15−22, GROMOS23−34, OPLS35,36.

As a well-established technique, additive force field has its intrinsic limitations. For systems involving a frequent and large change of electrostatic environment, such as the passage of small molecules or ions through lipid membrane bilayers, or the binding of substrate to the hydrophobic interior of an enzyme in water solution, the electron distribution change of those molecules can hardly be reflected by the fixed charge model. Thus, extended descriptions of the electrostatic interactions have been proposed to include polarization effects into the force field.

2.2 Polarizable force field

Different treatments of the polarization effects have led to different polarizable force fields, which can be generously distributed to three major groups: charge equilibration model,Drude particle model and induced dipole model. Numerous publications have been made on the theoretical background and force field products from those models37−40. Some relatively new polarizable force fields include the polarizable multipole interaction with supermolecular pairs (PMISP)41, the molecular fractionation with conjugate caps (MFCC)42, the atom bond electronegativity equalization method (ABEEM)43,44, the sum of interactions between fragments ab initio computed(SIBFA)45, the residue-specific force field (RSFF1 and RSFF2)46−48, etc. In this review, a few recent progresses in the establishment and application of polarizable force fields will be discussed.

The AMOEBA force field49,50, a notable product of the induced dipole model, has achieved great scientific research interests since its initial development due to its far better accuracy in modeling biological macromolecules. A recent work focused on the ion permeation of gramicidin A51showed that the AMOEBA force field was able to estimate the ion conductivity of the membrane in an order of magnitude that was very close to experimental results, while research beforehand using traditional force fields all led to tremendous deviations unable to be eliminated even with complicated and arbitrary correction terms included, thus further validated the necessity of introducing polarizability effects into simulation of biological processes in which the change in electrostatic environment cannot be ignored. However, the high computational cost caused by the calculation of induced dipoles and multipoles in AMOEBA greatly limits its application, even though the simulation speed can be increased via implementing enhanced sampling approaches and graphics processing units.

To handle this issue, a self-consistent multiscale treatment is very attractive but also technologically challenging. As a substitution, it is much simpler to seek the high-level energy accuracy via reweighting from a trajectory generated in low-level force fields. This strategy has been widely used to overcome the gap between quantum mechanical (QM) and molecular mechanical (MM) energy estimation. Typical work includes the non-Boltzmann Bennett reweighting scheme52−54and the multiscale Bennett acceptance ratio scheme55. Recently this approach was also applied to build a bridge between OPLS-AA and AMOEBA40. The authors reported a dual-force-field approach which performed 95% of the sampling using OPLS-AA. The required simulation time was decreased by a factor of 40 while via reweighting the mean free energy deviation comparing to direct AMOEBA simulation was only 0.2 kcal·mol−1(1 kcal·mol−1= 4.187 kJ·mol−1), and the deviation comparing to experimental deposition thermodynamics results was far better than that from uncorrected OPLS-AA results. However, it shall be noticed that approach was only tested on small organic molecules in their report,while for biological macromolecules, the major target of AMOEBA force field, the sampling deviation caused by different force field may be unneglectable and will make the trajectory generated via classical force field to be not very trustable.

Polarizability effects also play important roles in the investigation of enzyme reactions. It was shown by Warshel and Levitt in 1976 in their study of the lysozyme reaction56that an energetically unfavorable intermediate was formed without the polarization effects being counted, while the intermediate could be stabilized by the induced dipoles formed along with the binding of enzyme and substrate. Thus, a number of different polarizable force models, including the Drude oscillator57,58, fluctuating charge59,60, and induced dipole39,61−69,have been incorporated into the quantum mechanical/molecular mechanical (QM/MM) approach. Especially in the year 2016,three independent works39,70,71were reported focusing on replacing the classical molecular mechanical part in QM/MM with the AMOEBA force field. However, currently all those trials were only tested by QM/MM calculations of small molecules, while probably due to the high computational cost as well as the sampling obstacles, not so much progresses have been made to incorporate polarizable force fields in practical QM/MM simulations, such as the mechanical study of enzyme reactions.

Other than the previously described polarizable force fields which have achieved reasonable establishment and validated by a number of productive applications, there are also some theories dealing with atom polarizabilities with distinct strategies but have not been well recognized. As an important example, the explicit polarization (X-Pol) potential72−76, mainly proposed by Dr. Truhlar and Dr. Gao. Different to other polarizable force fields such as AMOEBA or CHARMM Drude,in which quantum chemical calculation results are used as the target for the fitting of force field parameters, in the X-Pol potential, the electron distribution among atoms is estimated with intrinsic QM approaches. To apply this strategy to deal with macromolecules or other large and complex systems, the X-Pol model is fragment-based, in which the whole system is divided into molecular subunits such as individual molecules,ligand and substrates, amino acid residues, etc. Then the complete wave function of the system is assumed to be a Hartree product of the wave function of each fragment77−80. The X-pol model has been incorporated into NAMD81and CHARMM82simulation package, and based on the model a QM force field for water, namely the XP3P (explicit polarization with three-point-charge potential) water model was developed and validated on both gas-phase clusters and liquid water83. By comparing the results from XP3P model and that of experiments, additive force fields (TIP3P) and two other polarizable force fields (AMOEBA and SWM4-NDP84), the XP3P did provide results in reasonable accord with the experimental values in some dimensions (density, dipole moments, etc.), and the overall accuracy was comparable to other empirical force fields. However, although the intrinsic theory behind XP3P model is more physically solid because its electron distribution is explicitly described based on QM treatment without introducing empirical parameters, the model did not provide obvious advantage in either accuracy or speed comparing to other polarizable force fields used as references in this work, not to mention the parameters of AMOEBA is being constantly optimized. Thus, the X-pol model has not proved its complete practical abilities and further work remains to be done on the realization of this theory.

3 Enhanced sampling schemes

Although the development in high-performance computer has greatly extended the applicability of molecular simulation,it is still not routinely feasible to study complex biological processes with time scales of milliseconds or beyond. Thus, in reasonable simulation time it is challenging to observe the occurrence of “rare events”85that involve crossing large kinetic energy barriers between metastable states. This is a common issue when studying the functionalities of biological systems,such as enzyme reactions, protein-ligand binding, folding and reallocation of proteins, etc. Anton86−89, the leading cooperation of super computers and specific optimization for molecular dynamics, does have expanded the accessibility of theoretical simulations. On the other hand, as an important group of research implements, a great number of enhanced sampling schemes have been proposed to achieve efficient sampling in the conformational space, while still unbiased free energy landscapes can be rebuilt via reweighting algorithms such as weighted histogram analysis method (WHAM)90, multistate Bennett acceptance ratio (MBAR)91and transition-based reweighting analysis method (TRAM)92. Different sampling strategies can be roughly distributed into the “constraint sampling” group and the “unconstrained sampling” group,depending on whether explicit constraints are applied to selected collective variables (namely the reaction coordinate).The choice of collective variables ranges from simple bond distances to more complexed ones such as root mean square deviation or native contact of protein residues. Constraint sampling schemes, including the popular target molecular dynamics93−97, meta-dynamics98−102, umbrella sampling103−109,string method110−112,milestoning113−116, etc., are very efficient in sampling along the given reaction coordinate, and the theoretical backgrounds, milestones in the developing processes and benchmark applications have achieved deep discussion in a variety of review articles85. However, arbitrarily selected reaction coordinates may not be the ideal candidacy that can represent the target biological process monotonically,especially when the process is too complicated to be described by simple coordinates. Under such condition, a “hidden barrier” resulted from the fluctuation of the environment (all coordinates other than the selected ones) can lead to poor convergence of the simulation results117. To solve the “hidden barrier” problem, the orthogonal space random walk(OSRW)118scheme was developed by Dr. Yangʹs group to include the sampling in the space orthogonal to the selected reaction coordinate space. On the other hand, several enhanced sampling schemes requiring no reaction coordinate definition have also been developed. A good reference review of unconstrained enhanced sampling approaches (replica exchange119−123, self-guided Langevin dynamics124−126,Essential energy space random walk127−129and accelerated molecular dynamics130−134) was provided by Dr. Miao and Dr.McCammon135. There are some other unconstrained methods such as the integrated tempering sampling (ITS)136and selective integrated tempering sampling (SITS)137, while this review will be focused on techniques newly proposed or the latest developments and refinements of old approaches published in recent years.

3.1 Recent developments of umbrella sampling

As one of the most widely used sampling approach in solving biological problems, umbrella sampling has achieved great scientific interests due to its quick convergence, trustable estimation of the potential of mean force and the compatibility for parallel computer sets. The technique can also be used along with other sampling algorithms such as replica exchange138,139, metadynamics140and target molecular dynamics141to further increase the sampling efficiency.Certainly, the approach also has its intrinsic drawbacks. First,umbrella sampling is performed based on a group of “windows”evenly distributed along a selected reaction coordinate, so the sampling results are highly limited to the initial numerical values of collective variables corresponding to those windows.The second problem is related to the previously mentioned hidden barrier. In typical umbrella sampling procedure,independent simulation is performed around each window, thus although the continuity along the reaction coordinate can be achieved (though still the free energy zero of each window is manually modified when “connecting” free energy profile of each window), the continuity of environment is routinely ignored. For windows with unneglectable hidden barriers orthogonal to the reaction coordinate space (i.e. transition states in enzyme reactions), the simulation will suffer from the long relaxation timescale and lead to a slow speed of convergence.

To solve the first problem, a method so called as the“adaptive umbrella sampling”105-107,142scheme has long been proposed. In this algorithm, not only samples can move in a small range near the center of each window, but also the initial collective variable settings can fluctuate during the simulation.One recent progress in this direction was made in Dr.Nakamuraʹs group. They proposed the virtual-system-coupled adaptive umbrella sampling scheme143, in which a virtual degree of freedom is added to the target system and the virtual system and the real system interact in an arbitrary way without breaking detailed balance condition144. During the simulation,the real system is restrained to windows with different states of the virtual degree of freedom (like the conventional idea of umbrella sampling that restraints the system to windows with different states along the reaction coordinate), and with a given time interval the system has a certain probability to be transferred from one virtual state to its neighboring states, with the transition probability satisfying the detailed balance equation. This strategy of transition between states is very like that of the Window-Exchange Umbrella sampling approach145-147, in which the system is allowed to swap between neighboring windows based on a Metropolis Monte Carlo exchange probability. The technique was initially validated via calculating the docking process of two flexible peptides and then used to achieve a fast estimation of binding energy landscape148.

For the second issue, which is a common obstacle for most kinds of constraints simulations, researchers have been trying to find a method that enables adaptive refinement of simulation parameters to accelerate convergence. One of the major developers of the umbrella sampling scheme, Dr. Rouxʹs group recently proposed the divergence analysis to identify problematic windows149. In this approach, the trajectory generated in the umbrella sampling simulation is binned to discrete conformational states and further used to construct the transition count matrix using transition-based reweighting analysis methods150. The matrix provides the unbiased distribution of those discrete states, and then by applying corresponding biasing potentials to the unbiased distribution one can predict the consensus distribution of a converged enhanced sampling biased simulation of each window. Based on the discrepancy between the distribution in real simulation and this consensus distribution, an adaptive sampling protocol can be used in the following simulation to enhance the sampling in those windows exhibit large discrepancies, either via increasing simulation time length or coupling with other enhanced sampling methods.

Another strategy to handle this problem is to construct an optimal biasing potential along the reaction coordinate before the umbrella sampling. Based on this strategy, a variety of memory based sampling approaches with adaptive and time-dependent potentials have been proposed. The local elevation umbrella sampling (LEUS) approach151, a combination of metadynamics and umbrella sampling scheme,was an important example. In typical memory-based approaches, the biasing potential in each window is made to be proportional to the number of previous visits. Thus, in a converged simulation, the biasing potential shall be able to,ideally, flatten the free energy surface along the selected reaction coordinate. The LEUS method was later coupled with the λ-dynamics approach152as an attempt to achieve more efficient sampling and automatized procedure while keep the accuracy of free energy153, with λ being introduced to the system as an order parameter representing the progress of the target process (i.e. λ = 0 for the beginning microstate and λ = 1 for the ending state). During the simulation, λ dynamically propagates along with the dynamics of the real system, and the biased potential applied to λ allows the sampling of a large conformational space. Recently the method was further developed to handle the multistate situation in a single simulation154. This time the order parameter λ evolves in a cyclic fashion, with the corresponding multiple states evenly distributed on the circle. This novel approach was initially validated by calculating absolute binding free energies of alkali ions to crown ethers in different solvents.

3.2 Recent developments in metadynamics

The metadynamics approach is a conventional and well-established algorithm to bypass the barriers on the free energy surface and allow the system to visit microstates of low probabilities, such as transition states. Due to this feature, the method is widely implemented in studying enzyme reactions.Metadynamics is most attractive in its high sampling efficiency because the free energy barriers between different states can be quickly overcome by additive Gaussian biasing potentials,while it is also this biasing potential oscillating around the actual free energies that lead to the difficulty to achieve simulation convergence.

To solve this problem, the well-tempered metadynamics approach was proposed155, in which the height of additive Gaussian potential smoothly decreases to zero as the current biasing potential increases. The Gaussian height is decreased exponentially based on an arbitrarily determined bias factor.This technique provided much better convergence in the resulting free energies, while the simulation efficiency was decreased as a trade-off156. As a further improvement, recently a new method, namely the transition-tempered metadynamics was developed to achieve more adaptive decreasing of the Gaussian potential height157. In this approach, Gaussian height is decreased based on the number of transitions between basins,which are roughly selected along the reaction coordinate using a combination of experimental results and physical insights before performing the simulation. Applications on a number of model systems showed that this new approach was able to sample along the selected collective variables as efficient as non-tempered metadynamics in the early stages of simulation while also achieve a good convergence similar to the welltempered metadynamics, and further the robustness of this method was validated by estimating the free energy profiles of the permeation process of small molecules through lipid bilayers158.

On the other hand, Dr. Parrinelloʹs group, the original developer of the metadynamics approach, recently proposed a new sampling strategy for quantum systems with a combination of Feynmanʹs path integral representation159of quantum statistical physics and metadynamics sampling160. The path integral formulation in computational studies, in which each quantum particle is mapped into a ring polymer of classical beads harmonically linked to each other, is a practical tool to explore systems in which quantum effects are important, which means that the tunneling effects can help crossing free energy barriers and enhance sampling. However, merely this effect is often not enough for sufficient sampling within a reasonable simulation time. By using the metadynamics sampling approach it is possible to enhance the configuration fluctuation of the ring polymer, thus allows tunneling to take place161.This path integral metadynamics approach was validated on a few model systems of different complexities. This sampling scheme was further improved by combining with the idea of replica exchange, with either the number of classical beads162or the de Broglie wavelengths163being used as the control of quantum effects in each replica and has been tested in a variety of model systems. In order to reduce the computational costs,it was reported that the number of beads can be reduced to two in a newly published work164and the practicability of this method was also tested in some simple systems.

3.3 Some new sampling algorithms

3.3.1 Accelerate sampling using bayesian interference

Trying to use experimental results to guide theoretical learning processes has long been an attractive research area. To take advantage of those experimental results containing systematic and random errors, as well as predictions based on those intrinsically inaccurate experimental observations, the Bayesian interference165−169is a widely-used strategy. The Modeling Employing Limited Data (MELD)170,171approach was proposed by Dr. Ken A. Dill’s group in 2015 as a method to make use of coarse physical insights (CPI), typically those limited or uncertain information from experiments to guide the sampling in molecular dynamics simulation. Based on the theory of Bayesian interference, those ambiguous and general information were introduced as a biasing potential to the Hamiltonian of the simulation system as a factor of conditional probability that can be used to optimize the initial probability distribution of atom coordinates. The technique was incorporated into the replica exchange method to refine protein structures170. As an example, for the coarse physical insights,the researchers used four rules as their guidance in their initial method-validation paper: (i) that proteins have secondary structures, (ii) that proteins have hydrophobic cores, (iii) that β-strands pairup, and (iv) that proteins have compact structures.Then, very like the machine learning strategy, the probability of the simulation trajectory that satisfies the rules will be used as a grade to judge the sample, which can then be further optimized.The MELD method was reported to be able to produce near-native protein structures well with a much faster simulation speed than regular molecular dynamics simulation.Later, the MELD method was further validated in the optimization of structures of complex macro protein molecules based on the native structures as well as a set of semi-reliable data, which are the distances between residues171. It was shown that the MELD method, different from the constraint sampling schemes that requires the correctness of given info, MELD could provide reasonable structures from imperfect data.

3.3.2 Variationally enhanced sampling

One major technical problem of enhanced sampling methods with constraints on collective variables is to construct a decent biasing potential that can flatten the free energy surface.Typical strategies of constructing the potential include metadynamics, in which a Gaussian potential whose height can be adjusted is continuously added; the adaptive biasing force method172-174, in which the potential is built via potential of mean forces estimated by current samples. In 2014 Valsson and Parrinello proposed a new technique, the variation approach175, to robustly construct the biasing potential. In this approach, a convex function of the biasing potential is minimized. With the optimal biasing potential that reaches the global minimum of this convex function, sampling along the reaction coordinate with a predefined, arbitrary probability distribution can be achieved176. This approach allows the construction of an ideal biasing potential via a small number of parameters that can be optimized by minimizing the convex function. After validating the method in model systems, its developers tried to extend its usage to construct high-dimensional free energy landscape as the sum of many two-dimensional terms and an additional term for initial estimation. Their first attempt was made on a small model system, Chignolin177, and then the technique was further refined by adding additional terms representing a set of coarse-grained collective variables to build the free energy landscape more efficiently but still keep its accuracy178.

Thanks to the continuous progress made in force fields and sampling techniques, molecular dynamics simulation is being more and more accepted as a useful tool to understand the mechanism of biological processes. In the next part of this review article, some recent and important explorations of biophysical systems using molecular dynamics will be discussed.

4 Using molecular dynamics to study biophysical systems

In general, molecular dynamics simulation is mostly used to study two kinds of biological processes: (1) conformational change of macro molecules, such as protein folding and unfolding, protein-ligand binding, DNA torsions, etc. In this area, tremendous important work has been done by Shawʹs group179-181, thanks to their unique capability to simulate biological processes of long time scales. (2) enzyme reactions,for which the research can be further divided into two directions, including simulation focused on the reaction center to collect the detailed but local information of reaction mechanism, and the study of enzyme activity via exploring conformational properties before and after the reaction. This kind of work can be further extended to the design of molecular probes, inhibitors and drugs. Currently, a hot research topic is the functionality of membrane proteins. Molecular simulation has been used to explore various structural and functional properties of membrane proteins, such as the source of selectivity in membrane permeation, structural flexibilities, etc.In a recent study using AMOEBA to accurately model the ion permeation of gramicidin A51the authors discussed the differences of behavior of the charged residues located in the interior region of the ion channel during the permeation of sodium cation and potassium cation, though they did not provide a strong conclusion on this problem because the publication was mainly a validation of the AMOEBA force field. In a recent publication, the drug binding mechanism of the M2 proton channel of influenza A virus was discussed in detail182. Taking advantage of the well-tempered metadynamics approach (which was discussed previously), the authors proposed a “bind and flip” mechanism, which involves a shift between a transiently populated “up” state and a thermodynamically favored “down” state, and investigated the influence of mutation of key residues as well as the inhibitory activity. The conformational change of proteins embedded in membrane bilayers is also an important target of computational studies. One recent study on the destabilization effects of G-protein-coupled receptors183showed that the loss of α-helicity due to the conformation shift as well as mobile detergent molecules in the membrane environment have a significant influence on the structural stability.

Application of molecular dynamics simulation on enzyme reactions is still highly limited by the sampling efficiency because the necessity of using quantum mechanical/molecular mechanical (QM/MM) hybrid approach to model bond breaking and bond formation greatly increase the simulation cost. However, with a strongly biased sampling the conformational space that can be visited is restricted, thus making it difficult to achieve a complete understanding of the reaction mechanism. Nevertheless, scientists are trying to achieve trustable mechanical investigations within limited simulation timescales. Recently a very important reaction, the nucleic acid polymerization catalyzed by DNA/RNA polymerases was studied using ab initio QM/MM simulation184.Based on a preserved hydrogen bond formed by the nucleophilic 3ʹ-OH and the nonbridging oxygen of the β-phosphate in the nucleotide shown in all available X-ray structures of DNA/RNA polymerases, researchers proposed a self-activated mechanism characterized by a concerted closed-loop catalytic sequence, and the mechanism was validated via QM/MM simulation on the human DNA polymerase η. The polymerization mechanism has long been a challenging research topic and in this work, the authors provided a novel mechanism which allows very high polymerization efficiency and the use of high-level QM/MM simulation greatly support their hypothesis.

To access the simulation of huge proteins containing hundreds of residues or more, coarse-grained (CG) force fields have been developed, such as MARTINI185, PACE186, ELBA187,and can be used with multi-scale simulation methods like the adaptive resolution simulation approaches188,189to achieve a balance between speed and accuracy. Enhanced sampling was also cooperated with coarse-grained force fields. In a recent published work190, the replica exchange method was used along with CG model recently developed191to tune the backbone hydrogen-bond of transmembrane peptide and reduce the time for the protein to stay in misfolded states. Comparing to the regular simulation (performed as a reference) using the same force field, the enhanced sampling provided much higher simulation efficiency, while still the structural features of the peptide in membrane could be characterized.

5 Summary and outlook

Using theoretical tools to pursue knowledge not accessible via experimental approaches for complicated systems has become increasingly important in scientific research, especially in solving biophysical problems. Other than molecular dynamics simulation, a few other computational approaches also play important roles, such as using the machine learning strategy to predict secondary structures of proteins192,193. For molecular dynamic simulation, there are also lots of challenging problems left to be dealt with. As one major problem, although current experimental approaches such as high-resolution Cryo-electron microscopy has been developed to provide near atom-level details of systems with millions of atoms, it is still not routinely possible to find a theoretical method to study the microscopic dynamical mechanisms of the self-assembly processes and the corresponding functionalities.However, with the continuous development in both the computer hardware and the simulation algorithm, molecular dynamics simulation is becoming more useful and more trustable. Although currently this approach is usually implemented as a support or supplement for experimental observations, it shall not be very long for molecular dynamics simulation to be used as a guidance and judgment of experiments.

(1) Alder, B. J.; Wainwright, T. E. J. Chem. Phys. 1959, 31, 459.doi: 10.1063/1.1730376

(2) McDaniel, J. G.; Schmidt, J. R. Annu. Rev. Phys. Chem. 2016,67, 467. doi: 10.1146/annurev-physchem-040215-112047

(3) Karplus, S.; Lifson, S. Biopolymers 1971, 10, 1973.doi: 10.1002/bip.360101014

(4) Warshel, A. Israel J. Chem. 1973, 11, 709.

(5) Warshel, A.; Levitt, M.; Lifson, S. J. Mol. Spectrosc. 1970, 33,84. doi: 10.1016/0022-2852(70)90054-8

(6) Warshel, A.; Lifson, S. J. Chem. Phys. 1970, 53, 582.doi: 10.1063/1.1674031

(7) Allinger, N. L.; Yuh, Y. H.; Lii, J. H. J. Am. Chem. Soc. 1989,111, 8551. doi: 10.1021/ja00205a001

(8) Halgren, T. A. Abstr. Pap. Am. Chem. S. 1992, 204, 38.

(9) Halgren, T. A.; Bush, B. L. Abstr. Pap. Am. Chem. S. 1996, 212, 2.

(10) Halgren, T. A.; Nachbar, R. B. Abstr. Pap. Am. Chem. S.1996, 211, 70.

(11) Rappe, A. K.; Casewit, C. J.; Colwell, K. S.; Goddard, W. A.;Skiff, W. M. J. Am. Chem. Soc. 1992, 114, 10024.doi: 10.1021/ja00051a040

(12) Mayo, S. L.; Olafson, B. D.; Goddard, W. A. J. Phys.Chem-Us. 1990, 94, 8897. doi: 10.1021/j100389a010

(13) Case, D. A.; Cheatham, T. E., 3rd; Darden, T.; Gohlke, H.;Luo, R.; Merz, K. M., Jr.; Onufriev, A.; Simmerling, C.;Wang, B.; Woods, R. J. J. Comput. Chem. 2005, 26, 1668.doi: 10.1002/jcc.20290

(14) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz,K. M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.;Caldwell, J. W.; Kollman, P. A. J. Am. Chem. Soc. 1996, 118,2309. doi: 10.1021/ja955032e

(15) Best, R. B.; Mittal, J.; Feig, M.; MacKerell, A. D., Jr.Biophys. J. 2012, 103, 1045. doi: 10.1016/j.bpj.2012.07.042

(16) Guvench, O.; Hatcher, E. R.; Venable, R. M.; Pastor, R. W.;Mackerell, A. D. J. Chem. Theory Comput. 2009, 5, 2353.doi: 10.1021/ct900242e

(17) Hart, K.; Foloppe, N.; Baker, C. M.; Denning, E. J.; Nilsson,L.; Mackerell, A. D., Jr. J. Chem. Theory Comput. 2012, 8,348. doi: 10.1021/ct200723y

(18) MacKerell, A. D., Jr.; Banavali, N.; Foloppe, N. Biopolymers 2000, 56, 257. doi: 10.1002/1097-0282(2000)56:4＜257::AID-BIP10029＞3.0.CO;2-W

(19) Mallajosyula, S. S.; Guvench, O.; Hatcher, E.; Mackerell, A.D., Jr. J. Chem. Theory Comput. 2012, 8, 759.doi: 10.1021/ct200792v

(20) Raman, E. P.; Guvench, O.; MacKerell, A. D., Jr. J. Phys.Chem. B 2010, 114, 12981. doi: 10.1021/jp105758h

(21) Vanommeslaeghe, K.; Hatcher, E.; Acharya, C.; Kundu, S.;Zhong, S.; Shim, J.; Darian, E.; Guvench, O.; Lopes, P.;Vorobyov, I.; Mackerell, A. D., Jr. J. Comput. Chem. 2010,31, 671. doi: 10.1002/jcc.21367

(22) Yu, W.; He, X.; Vanommeslaeghe, K.; MacKerell, A. D., Jr. J.Comput. Chem. 2012, 33, 2451. doi: 10.1002/jcc.23067

(23) Daura, X.; Oliva, B.; Querol, E.; Aviles, F. X.; Tapia, O.Proteins 1996, 25, 89. doi: 10.1002/(Sici)1097-0134(199605)25:1＜89::Aid-Prot7＞3.0.Co;2-F

(24) Hansen, H. S.; Hunenberger, P. H. J. Comput. Chem. 2011,32, 998. doi: 10.1002/jcc.21675

(25) Horta, B. A. C.; Lin, Z. X.; Huang, W.; Riniker, S.; van Gunsteren, W. F.; Hunenberger, P. H. J. Comput. Chem. 2012,33, 1907. doi: 10.1002/jcc.23021

(26) Kouwijzer, M. L. C. E.; vanEijck, B. P.; Kooijman, H.;Kroon, J. Aip. Conf. Proc. 1995, 330, 393.

(27) Lins, R. D.; Hunenberger, P. H. J. Comput. Chem. 2005, 26,1400. doi: 10.1002/jcc.20275

(28) Oostenbrink, C.; Soares, T. A.; van der Vegt, N. F. A.; van Gunsteren, W. F. Eur. Biophys. J. Biophy. 2005, 34, 273.doi: 10.1007/s00249-004-0448-6

(29) Ott, K. H.; Meyer, B. J. Comput. Chem. 1996, 17, 1068.doi: 10.1002/(Sici)1096-987x(199606)17:8＜1068::Aid-Jcc14＞3.3.Co;2-T

(30) Pol-Fachin, L.; Rusu, V. H.; Verli, H.; Lins, R. D. J. Chem.Theory Comput. 2012, 8, 4681. doi: 10.1021/ct300479h

(31) Reif, M. M.; Hunenberger, P. H.; Oostenbrink, C. J. Chem.Theory Comput. 2012, 8, 3705. doi: 10.1021/ct300156h

(32) Smith, M. D.; Rao, J. S.; Segelken, E.; Cruz, L. J Chem Inf.Model 2015, 55, 2587. doi: 10.1021/acs.jcim.5b00308

(33) Soares, T. A.; Hunenberger, P. H.; Kastenholz, M. A.;Krautler, V.; Lenz, T.; Lins, R. D.; Oostenbrink, C.; Van Gunsteren, W. F. J. Comput. Chem. 2005, 26, 725.doi: 10.1002/jcc.20193

(34) Suardiaz, R.; Maestre, M.; Suarez, E.; Perez, C. J. Mol.Struc-Theochem. 2006, 778, 21.doi: 10.1016/j.theochem.2006.08.030

(35) Jorgensen, W. L.; Maxwell, D. S.; TiradoRives, J. J. Am.Chem. Soc. 1996, 118, 11225. doi: 10.1021/ja9621760

(36) Kaminski, G. A.; Friesner, R. A.; Tirado-Rives, J.; Jorgensen,W. L. J. Phys. Chem. B 2001, 105, 6474.doi: 10.1021/jp003919d

(37) Rick, S. W.; Stuart, S. J. Rev. Comp. Ch. 2002, 18, 89.doi: 10.1002/0471433519.ch3

(38) Lamoureux, G.; Roux, B. J. Chem. Phys. 2003, 119, 3025.doi: 10.1063/1.1589749

(39) Kratz, E. G.; Walker, A. R.; Lagardere, L.; Lipparini, F.;Piquemal, J. P.; Andres Cisneros, G. J. Comput. Chem. 2016,37, 1019. doi: 10.1002/jcc.24295

(40) Nessler, I. J.; Litman, J. M.; Schnieders, M. J. Phys. Chem.Chem. Phys. 2016. doi: 10.1039/c6cp02595a

(41) Soderhjelm, P.; Ryde, U. J. Phys. Chem. A 2009, 113, 617.doi: 10.1021/jp8073514

(42) Zhang, D. W.; Zhang, J. Z. H. J. Chem. Phys. 2003, 119,3599. doi: 10.1063/1.1591727

(43) Yang, Z. Z. Abstr. Pap. Am. Chem. S. 2006, 231.

(44) Wang, C. S.; Zhao, D. X.; Yang, Z. Z. Chem. Phys. Lett.2000, 330, 132. doi: 10.1016/S0009-2614(00)00938-6

(45) Piquemal, J. P.; Gresh, N.; Giessner-Prettre, C. J. Phys.Chem. A 2003, 107, 10353. doi: 10.1021/jp035748t

(46) Jiang, F.; Zhou, C. Y.; Wu, Y. D. J. Phys. Chem. B 2014, 118,6983. doi: 10.1021/jp5017449

(47) Xun, S. N.; Jiang, F.; Wu, Y. D. J. Chem. Theory Comput.2015, 11, 1949. doi: 10.1021/acs.jctc.5b00029

(48) Zhou, C. Y.; Jiang, F.; Wu, Y. D. J. Phys. Chem. B 2015, 119,1035. doi: 10.1021/jp5064676

(49) Ponder, J. W.; Wu, C. J.; Ren, P. Y.; Pande, V. S.; Chodera, J.D.; Schnieders, M. J.; Haque, I.; Mobley, D. L.; Lambrecht,D. S.; DiStasio, R. A.; Head-Gordon, M.; Clark, G. N. I.;Johnson, M. E.; Head-Gordon, T. J. Phys. Chem. B 2010,114, 2549. doi: 10.1021/jp910674d

(50) Shi, Y.; Xia, Z.; Zhang, J. J.; Best, R.; Wu, C. J.; Ponder, J.W.; Ren, P. Y. J. Chem. Theory Comput. 2013, 9, 4046.doi: 10.1021/ct4003702

(51) Peng, X. D.; Zhang, Y. B.; Chu, H. Y.; Li, Y.; Zhang, D. L.;Cao, L. R.; Li, G. H. J. Chem. Theory Comput. 2016, 12,2973. doi: 10.1021/acs.jctc.6b00128

(52) Konig, G.; Hudson, P. S.; Boresch, S.; Woodcock, H. L.J. Chem. Theory Comput. 2014, 10, 1406.doi: 10.1021/ct401118k

(53) Konig, G.; Pickard, F. C. t.; Mei, Y.; Brooks, B. R.J. Comput. Aided Mol. Des. 2014, 28, 245.doi: 10.1007/s10822-014-9708-4

(54) Konig, G.; Mei, Y.; Pickard, F. C.; Simmonett, A. C.; Miller,B. T.; Herbert, J. M.; Woodcock, H. L.; Brooks, B. R.; Shao,Y. H. J. Chem. Theory Comput. 2016, 12, 332.doi: 10.1021/acs.jctc.5b00874

(55) Dybeck, E. C.; Konig, G.; Brooks, B. R.; Shirts, M. R.J. Chem. Theory Comput. 2016, 12, 1466.doi: 10.1021/acs.jctc.5b01188

(56) Warshel, A.; Levitt, M. J. Mol. Biol. 1976, 103, 227.

(57) Boulanger, E.; Thiel, W. J. Chem. Theory Comput. 2014, 10,1795. doi: 10.1021/ct401095k

(58) Boulanger, E.; Thiel, W. J. Chem. Theory Comput. 2012, 8,4527. doi: 10.1021/ct300722e

(59) Lipparini, F.; Cappelli, C.; Barone, V. J. Chem. Phys. 2013,138, 234108. doi: 10.1063/1.4811113

(60) Lipparini, F.; Cappelli, C.; Barone, V. J. Chem. Theory Comput. 2012, 8, 4153. doi: 10.1021/ct3005062

(61) Caprasecca, S.; Jurinovich, S.; Lagardere, L.; Stamm, B.;Lipparini, F. J. Chem. Theory Comput. 2015, 11, 694. doi:10.1021/ct501087m

(62) Caprasecca, S.; Jurinovich, S.; Viani, L.; Curutchet, C.;Mennucci, B. J. Chem. Theory Comput. 2014, 10, 1588. doi:10.1021/ct500021d

(63) Thellamurege, N. M.; Si, D.; Cui, F.; Zhu, H.; Lai, R.; Li, H.J. Comput. Chem. 2013, 34, 2816. doi: 10.1002/jcc.23435

(64) Caprasecca, S.; Curutchet, C.; Mennucci, B. J. Chem.Theory Comput. 2012, 8, 4462. doi: 10.1021/ct300620w

(65) Sneskov, K.; Schwabe, T.; Christiansen, O.; Kongsted, J.Phys. Chem. Chem. Phys. 2011, 13, 18551.doi: 10.1039/c1cp22067e

(66) Schwabe, T.; Olsen, J. M.; Sneskov, K.; Kongsted, J.;Christiansen, O. J. Chem. Theory Comput. 2011, 7, 2209. doi:10.1021/ct200258g

(67) Olsen, J. M.; Aidas, K.; Mikkelsen, K. V.; Kongsted, J.J. Chem. Theory Comput. 2010, 6, 249.doi: 10.1021/ct900502s

(68) Curutchet, C.; Munoz-Losa, A.; Monti, S.; Kongsted, J.;Scholes, G. D.; Mennucci, B. J. Chem. Theory Comput.2009, 5, 1838. doi: 10.1021/ct9001366

(69) Nielsen, C. B.; Christiansen, O.; Mikkelsen, K. V.; Kongsted,J. J. Chem. Phys. 2007, 126, 154112.doi: 10.1063/1.2711182

(70) Loco, D.; Polack, E.; Caprasecca, S.; Lagardere, L.;Lipparini, F.; Piquemal, J. P.; Mennucci, B. J. Chem. Theory Comput. 2016, 12, 3654. doi: 10.1021/acs.jctc.6b00385

(71) Dziedzic, J.; Mao, Y.; Shao, Y.; Ponder, J.; Head-Gordon, T.;Head-Gordon, M.; Skylaris, C. K. J. Chem. Phys. 2016, 145,124106. doi: 10.1063/1.4962909

(72) Han, J.; Truhlar, D. G.; Gao, J. Theor. Chem. Acc. 2012, 131,1161. doi: 10.1007/s00214-012-1161-7

(73) Leverentz, H. R.; Gao, J.; Truhlar, D. G. Theor. Chem. Acc.2011, 129, 3. doi: 10.1007/s00214-011-0889-9

(74) Xie, W.; Orozco, M.; Truhlar, D. G.; Gao, J. J. Chem. Theory Comput. 2009, 5, 459. doi: 10.1021/ct800239q

(75) Song, L.; Han, J.; Lin, Y. L.; Xie, W.; Gao, J. J. Phys. Chem.A 2009, 113, 11656. doi: 10.1021/jp902710a

(76) Xie, W.; Gao, J. J. Chem. Theory Comput. 2007, 3, 1890.doi: 10.1021/ct700167b

(77) Xie, W. S.; Song, L. C.; Truhlar, D. G.; Gao, J. L. J. Chem.Phys. 2008, 128. doi: Artn 23410810.1063/1.2936122

(78) Xie, W. S.; Song, L. C.; Truhlar, D. G.; Gao, J. L. J. Phys.Chem. B 2008, 112, 14124. doi: 10.1021/jp804512f

(79) Gao, J. L. J. Chem. Phys. 1998, 109, 2346.doi: 10.1063/1.476802

(80) Gao, J. L. J. Phys. Chem. B 1997, 101, 657.doi: 10.1021/jp962833a

(81) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.;Tajkhorshid, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.;Schulten, K. J. Comput. Chem. 2005, 26, 1781.doi: 10.1002/jcc.20289

(82) Brooks, B. R.; Brooks, C. L.; Mackerell, A. D.; Nilsson, L.;Petrella, R. J.; Roux, B.; Won, Y.; Archontis, G.; Bartels, C.;Boresch, S.; Caflisch, A.; Caves, L.; Cui, Q.; Dinner, A. R.;Feig, M.; Fischer, S.; Gao, J.; Hodoscek, M.; Im, W.;Kuczera, K.; Lazaridis, T.; Ma, J.; Ovchinnikov, V.; Paci, E.;Pastor, R. W.; Post, C. B.; Pu, J. Z.; Schaefer, M.; Tidor, B.;Venable, R. M.; Woodcock, H. L.; Wu, X.; Yang, W.; York,D. M.; Karplus, M. J. Comput. Chem. 2009, 30, 1545. doi:10.1002/jcc.21287

(83) Han, J.; Mazack, M. J. M.; Zhang, P.; Truhlar, D. G.; Gao, J.L. J. Chem. Phys. 2013, 139. doi: Artn 05450310.1063/1.4816280

(84) Stukan, M. R.; Asmadi, A.; Abdallah, W. J. Mol. Liq. 2013,180, 65. doi: 10.1016/j.molliq.2012.12.023

(85) Rohrdanz, M. A.; Zheng, W. W.; Clementi, C. Annu. Rev.Phys. Chem. 2013, 64, 295.doi: 10.1146/annurev-physchem-040412-110006

(86) Shaw, D. E.; Deneroff, M. M.; Dror, R. O.; Kuskin, J. S.;Larson, R. H.; Salmon, J. K.; Young, C.; Batson, B.; Bowers,K. J.; Chao, J. C.; Eastwood, M. P.; Gagliardo, J.; Grossman,J. P.; Ho, C. R.; Ierardi, D. J.; Kolossvary, I.; Klepeis, J. L.;Layman, T.; McLeavey, C.; Moraes, M. A.; Mueller, R.;Priest, E. C.; Shan, Y. B.; Spengler, J.; Theobald, M.; Towles,B.; Wang, S. C. Conf. Proc. Int. Symp. C 2007, 1.

(87) Shaw, D. E.; Deneroff, M. M.; Dror, R. O.; Kuskin, J. S.;Larson, R. H.; Salmon, J. K.; Young, C.; Batson, B.; Bowers,K. J.; Chao, J. C.; Eastwood, M. P.; Gagliardo, J.; Grossman,J. P.; Ho, C. R.; Ierardi, D. J.; Kolossvary, I.; Klepeis, J. L.;Layman, T.; Mcleavey, C.; Moraes, M. A.; Mueller, R.;Priest, E. C.; Shan, Y. B.; Spengler, J.; Theobald, M.; Towles,B.; Wang, S. C. Commun. Acm. 2008, 51,91. doi: 10.1145/1364782.1364802

(88) Shaw, D. E.; Dror, R. O.; Salmon, J. K.; Grossman, J. P.;Mackenzie, K. M.; Bank, J. A.; Young, C.; Deneroff, M. M.;Batson, B.; Bowers, K. J.; Chow, E.; Eastwood, M. P.;Ierardi, D. J.; Klepeis, J. L.; Kuskin, J. S.; Larson, R. H.;Lindorff-Larsen, K.; Maragakis, P.; Moraes, M. A.; Piana, S.;Shan, Y. B.; Towles, B. Proceedings of the Conference on High Performance Computing Networking, Storage and Analysis 2009.

(89) Grossman, J. P.; Towles, B.; Greskamp, B.; Shaw, D. E. Int.Parall. Distrib. P 2015, 860. doi: 10.1109/Ipdps.2015.42

(90) Kumar, S.; Bouzida, D.; Swendsen, R. H.; Kollman, P. A.;Rosenberg, J. M. J. Comput. Chem. 1992, 13, 1011.doi: 10.1002/jcc.540130812

(91) Shirts, M. R.; Chodera, J. D. J. Chem. Phys. 2008, 129.doi: Artn 12410510.1063/1.2978177

(92) Mey, A. S. J. S.; Wu, H.; Noe, F. Phys. Rev. X 2014, 4.doi: ARTN 04101810.1103/PhysRevX.4.041018

(93) Compoint, M.; Picaud, F.; Ramseyer, C.; Girardet, C. J.Chem. Phys. 2005, 122, 134707. doi: 10.1063/1.1869413

(94) Aci, S.; Mazier, S.; Genest, D. J. Mol. Biol. 2005, 351, 520.doi: 10.1016/j.jmb.2005.06.009

(95) Kruger, P.; Verheyden, S.; Declerck, P. J.; Engelborghs, Y.Protein Sci. 2001, 10, 798. doi: 10.1110/ps.40401

(96) Ferrara, P.; Apostolakis, J.; Caflisch, A. Proteins 2000, 39, 252.

(97) Schlitter, J.; Engels, M.; Kruger, P. J. Mol. Graph 1994, 12, 84.

(98) Bussi, G.; Laio, A.; Parrinello, M. Phys. Rev. Lett. 2006, 96,090601. doi: 10.1103/PhysRevLett.96.090601

(99) Laio, A.; Rodriguez-Fortea, A.; Gervasio, F. L.; Ceccarelli,M.; Parrinello, M. J. Phys. Chem. B 2005, 109, 6714.doi: 10.1021/jp045424k

(100) Gervasio, F. L.; Laio, A.; Parrinello, M. J. Am. Chem. Soc.2005, 127, 2600. doi: 10.1021/ja0445950

(101) Asciutto, E.; Sagui, C. J. Phys. Chem. A 2005, 109, 7682.doi: 10.1021/jp053428z

(102) Micheletti, C.; Laio, A.; Parrinello, M. Phys. Rev. Lett. 2004,92, 170601. doi: 10.1103/PhysRevLett.92.170601

(103) Kurtovic, Z.; Marchi, M.; Chandler, D. Mol. Phys. 1993, 78,1155. doi: 10.1080/00268979300100751

(104) Ding, K. J.; Valleau, J. P. J. Chem. Phys. 1993, 98, 3306.doi: 10.1063/1.464102

(105) Hooft, R. W. W.; Vaneijck, B. P.; Kroon, J. J. Chem. Phys.1992, 97, 6690. doi: 10.1063/1.463947

(107) Mezei, M. J. Comput. Phys. 1987, 68, 237.doi: 10.1016/0021-9991(87)90054-4

(108) Harvey, S. C.; Prabhakaran, M. J. Phys. Chem-Us. 1987, 91,4799. doi: 10.1021/j100302a030

(109) Shing, K. S.; Gubbins, K. E. Mol. Phys. 1981, 43, 717.doi: 10.1080/00268978100101631

(110) Peters, B.; Heyden, A.; Bell, A. T.; Chakraborty, A. J. Chem.Phys. 2004, 120, 7877. doi: 10.1063/1.1691018

(111) Weinan, E.; Ren, W. Q.; Vanden-Eijnden, E. Phys. Rev. B 2002, 66. doi: ARTN 05230110.1103/PhysRevB.66.052301

(112) Tait, R. J.; Zhong, J. L. Int. J. Nonlinear. Mech. 1993, 28,713. doi: 10.1016/0020-7462(93)90031-F

(113) West, A. M. A.; Elber, R.; Shalloway, D. J. Chem. Phys.2007, 126. doi: Artn 14510410.1063/1.2716389

(114) Majek, P.; Elber, R. J. Chem. Theory Comput. 2010, 6, 1805.doi: 10.1021/ct100114j

(115) Aristoff, D.; Bello-Rivas, J. M.; Elber, R. Multiscale Model Sim. 2016, 14, 301. doi: 10.1137/15m102157x

(116) Cardenas, A. E.; Elber, R. J. Phys. Chem. B 2016, 120, 8208.doi: 10.1021/acs.jpcb.6b01890

(117) Abrams, C.; Bussi, G. Entropy-Switz 2014, 16, 163.doi: 10.3390/e16010163

(118) Min, D.; Zheng, L.; Harris, W.; Chen, M.; Lv, C.; Yang, W.J. Chem. Theory Comput. 2010, 6, 2253.doi: 10.1021/ct100033s

(119) Nymeyer, H.; Gnanakaran, S.; Garcia, A. E. Methods Enzymol. 2004, 383, 119.doi: 10.1016/S0076-6879(04)83006-4

(120) Rhee, Y. M.; Pande, V. S. Biophys. J. 2003, 84, 775.doi: 10.1016/S0006-3495(03)74897-8

(121) Jang, S.; Shin, S.; Pak, Y. Phys. Rev. Lett. 2003, 91, 058305.doi: 10.1103/PhysRevLett.91.058305

(122) Yamamoto, R.; Kob, W. Phys. Rev. E Stat. Phys. Plasmas Fluids Relat. Interdiscip Topics 2000, 61, 5473.

(123) Machta, J.; Newman, M. E.; Chayes, L. B. Phys. Rev. E Stat.Phys. Plasmas Fluids Relat. Interdiscip Topics 2000, 62, 8782.

(124) Wu, X.; Brooks, B. R.; Vanden-Eijnden, E. J. Comput. Chem.2016, 37, 595. doi: 10.1002/jcc.24015

(125) Wu, X.; Damjanovic, A.; Brooks, B. R. Adv. Chem. Phys.2012, 150, 255. doi: 10.1002/9781118197714.ch6

(126) Damjanovic, A.; Wu, X.; Garcia-Moreno, E. B.; Brooks, B.R. Biophys. J. 2008, 95, 4091.doi: 10.1529/biophysj.108.130906

(127) Lv, C.; Zheng, L.; Yang, W. J. Chem. Phys. 2012, 136,044103. doi: 10.1063/1.3678220

(128) Zheng, L.; Yang, W. J. Chem. Phys. 2008, 129, 014105.doi: 10.1063/1.2949815

(129) Li, H.; Min, D.; Liu, Y.; Yang, W. J. Chem. Phys. 2007, 127,094101. doi: 10.1063/1.2769356

(130) de Oliveira, C. A.; Hamelberg, D.; McCammon, J. A. J. Phys.Chem. B 2006, 110, 22695. doi: 10.1021/jp062845o

(131) Hamelberg, D.; Shen, T.; Andrew McCammon, J. J. Chem.Phys. 2005, 122, 241103. doi: 10.1063/1.1942487

(132) Choudhary, D.; Clancy, P. J. Chem. Phys. 2005, 122, 154509.doi: 10.1063/1.1878733

(133) Miron, R. A.; Fichthorn, K. A. Phys. Rev. Lett. 2004, 93,128301. doi: 10.1103/PhysRevLett.93.128301

(134) Hamelberg, D.; Mongan, J.; McCammon, J. A. J. Chem.Phys. 2004, 120, 11919. doi: 10.1063/1.1755656

(135) Miao, Y. L.; McCammon, J. A. Mol. Simulat. 2016, 42, 1046.doi: 10.1080/08927022.2015.1121541

(136) Gao, Y. Q. J. Chem. Phys. 2008, 128.doi: Artn 06410510.1063/1.2825614

(137) Yang, L.; Liu, C. W.; Shao, Q.; Zhang, J.; Gao, Y. Q. Acc.Chem. Res. 2015, 48, 947. doi: 10.1021/ar500267n

(138) Murata, K.; Sugita, Y.; Okamoto, Y. Slow Dynamics in Complex Systems 2004, 708, 332.

(139) Murata, K.; Sugita, Y.; Okamoto, Y. Chem. Phys. Lett. 2004,385, 1. doi: 10.1016/j.cplett.2003.10.159

(140) Awasthi, S.; Kapil, V.; Nair, N. N. J. Comput. Chem. 2016,37, 1413. doi: 10.1002/jcc.24349

(141) Wang, Q.; Xue, T.; Song, C. N.; Wang, Y.; Chen, G. J. Int. J.Mol. Sci. 2016, 17. doi: 10.3390/ijms17050692

(142) Bartels, C.; Karplus, M. J. Comput. Chem. 1997, 18, 1450.doi: 10.1002/(Sici)1096-987x(199709)18:12＜1450::Aid-Jcc3＞3.0.Co;2-I

(143) Higo, J.; Dasgupta, B.; Mashimo, T.; Kasahara, K.;Fukunishi, Y.; Nakamura, H. J. Comput. Chem. 2015, 36,1489. doi: 10.1002/jcc.23948

(144) Higo, J.; Umezawa, K.; Nakamura, H. J. Chem. Phys. 2013,138. doi: Artn 18410610.1063/1.4803468

(145) Park, S.; Beaven, A. H.; Klauda, J. B.; Im, W. J. Chem.Theory Comput. 2015, 11, 3466. doi:10.1021/acs.jctc.5b00232

(146) Park, S.; Im, W. J. Chem. Theory Comput. 2014, 10, 2719.doi: 10.1021/ct500504g

(147) Park, S.; Im, W. J. Chem. Theory Comput. 2013, 9, 13.doi: 10.1021/ct3008556

(148) Dasgupta, B.; Junichi, H.; Nakamura, H. Biophys. J. 2016, 110, 55a.

(149) Jo, S.; Suh, D.; He, Z. W.; Chipot, C.; Roux, B. J. Phys.Chem. B 2016, 120, 8733. doi: 10.1021/acs.jpcb.6b05125

(150) Wu, H.; Noe, F. Multiscale Model Sim. 2014, 12, 25.doi: 10.1137/120895883

(151) Hansen, H. S.; Hunenberger, P. H. J. Comput. Chem. 2010,31, 1. doi: 10.1002/jcc.21253

(152) Wu, P.; Hu, X.; Yang, W. J. Phys. Chem. Lett. 2011, 2, 2099.doi: 10.1021/jz200808x

(153) Bieler, N. S.; Häuselmann, R.; Hünenberger, P. H. J. Chem.Theory Comput. 2014, 10, 3006. doi: 10.1021/ct5002686

(154) Bieler, N. S.; Tschopp, J. P.; Hunenberger, P. H. J. Chem.Theory Comput. 2015, 11, 2575.doi: 10.1021/acs.jctc.5b00118

(155) Barducci, A.; Bussi, G.; Parrinello, M. Phys. Rev. Lett. 2008,100. doi: ARTN 02060310.1103/PhysRevLett.100.020603

(156) Dama, J. F.; Parrinello, M.; Voth, G. A. Phys. Rev. Lett. 2014,112. doi: ARTN 24060210.1103/PhysRevLett.112.240602

(157) Dama, J. F.; Rotskoff, G.; Parrinello, M.; Voth, G. A.J. Chem. Theory Comput. 2014, 10, 3626.doi: 10.1021/ct500441q

(158) Sun, R.; Dama, J. F.; Tan, J. S.; Rose, J. P.; Voth, G. A.J. Chem. Theory Comput. 2016, 12, 5157.doi: 10.1021/acs.jctc.6b00206

(159) Goodall, M. C. Nature 1962, 196, 370.doi: 10.1038/196370a0

(160) Quhe, R. G.; Nava, M.; Tiwary, P.; Parrinello, M. J. Chem.Theory Comput. 2015, 11, 1383. doi: 10.1021/ct501002a

(161) Chandler, D.; Wolynes, P. G. J. Chem. Phys. 1981, 74, 4078.doi: Doi 10.1063/1.441588

(162) Peng, Y. X.; Cao, Z.; Zhou, R. H.; Voth, G. A. J. Chem.Theory Comput. 2014, 10, 3634. doi: 10.1021/ct500447r

(163) Nava, M.; Quhe, R.; Palazzesi, F.; Tiwary, P.; Parrinello, M.J. Chem. Theory Comput. 2015, 11, 5114.doi: 10.1021/acs.jctc.5b00818

(164) Nava, M.; Palazzesi, F.; Perego, C.; Parrinello, M.arXiv:1607.04846 2016.

(165) Smith, C. A. B. Int. Stat. Rev. 1975, 43, 242.doi: 10.2307/1402913

(166) Kadane, J. B. J. Am. Stat. Assoc. 1975, 70, 248.doi: 10.2307/2285412

(167) Titekar, V. G. Curr. Sci. India. 1974, 43, 327.

(168) Scott, A. J. Aust. J. Stat. 1974, 16, 186.

(169) Hill, B. M. Technometrics 1974, 16, 478.

(170) Perez, A.; MacCallum, J. L.; Dill, K. A. Proc. Natl. Acad. Sci.U S A 2015, 112, 11846. doi: 10.1073/pnas.1515561112

(171) MacCallum, J. L.; Perez, A.; Dill, K. A. Proc. Natl. Acad. Sci.U S A 2015, 112, 6985. doi: 10.1073/pnas.1506788112

(172) Lelievre, T.; Rousset, M.; Stoltz, G. Nonlinearity 2008, 21,1155. doi: 10.1088/0951-7715/21/6/001

(173) Darve, E.; Rodriguez-Gomez, D.; Pohorille, A. J. Chem.Phys. 2008, 128. doi: Artn 14412010.1063/1.2829861

(174) Kim, J. G.; Fukunishi, Y.; Nakamura, H. Phys. Rev. E 2004,70. doi: ARTN 05710310.1103/PhysRevE.70.057103 Valsson, O.; Parrinello, M. Phys. Rev. Lett. 2014, 113.doi: Artn 09060110.1103/PhysRevLett.113.090601

(175) McCarty, J.; Valsson, O.; Parrinello, M. J. Chem. Theory Comput. 2016, 12, 2162. doi: 10.1021/acs.jctc.6b00125

(176) Shaffer, P.; Valsson, O.; Parrinello, M. P Natl. Acad. Sci.USA 2016, 113, 1150. doi: 10.1073/pnas.1519712113

(177) Shaffer, P.; Valsson, O.; Parrinello, M. J. Chem. Theory Comput. 2016, 12, 5751. doi: 10.1021/acs.jctc.6b00786

(178) Lindorff-Larsen, K.; Piana, S.; Dror, R. O.; Shaw, D. E.Science 2011, 334, 517. doi: 10.1126/science.1208351

(179) Jensen, M. O.; Jogini, V.; Borhani, D. W.; Leffler, A. E.;Dror, R. O.; Shaw, D. E. Science 2012, 336, 229. doi:10.1126/science.1216533

(180) Dror, R. O.; Green, H. F.; Valant, C.; Borhani, D. W.;Valcourt, J. R.; Pan, A. C.; Arlow, D. H.; Canals, M.; Lane, J.R.; Rahmani, R.; Baell, J. B.; Sexton, P. M.; Christopoulos,A.; Shaw, D. E. Nature 2013, 503, 295. doi:10.1038/nature12595

(181) Llabrés, S.; Juárez-Jiménez, J.; Masetti, M.; Leiva, R.;Vázquez, S.; Gazzarrini, S.; Moroni, A.; Cavalli, A.; Luque,F. J. J. Am. Chem. Soc. 2016, 138, 15345. doi:10.1021/jacs.6b07096

(182) Lee, S.; Mao, A.; Bhattacharya, S.; Robertson, N.;Grisshammer, R.; Tate, C. G.; Vaidehi, N. J. Am. Chem. Soc.2016, 138, 15425. doi: 10.1021/jacs.6b08742

(183) Genna, V.; Vidossich, P.; Ippoliti, E.; Carloni, P.; Vivo, M.J. Am. Chem. Soc. 2016, 138, 14592.doi: 10.1021/jacs.6b05475

(184) Marrink, S. J.; Risselada, H. J.; Yefimov, S.; Tieleman, D. P.;de Vries, A. H. J. Phys. Chem. B 2007, 111, 7812.doi: 10.1021/jp071097f

(185) Wan, C. K.; Han, W.; Wu, Y. D. J. Chem. Theory Comput.2012, 8, 300. doi: 10.1021/ct2004275

(186) Orsi, M.; Essex, J. W. Plos. One 2011, 6.doi: Artn e2863710.1371/journal.pone.0028637

(187) Zavadlav, J.; Melo, M. N.; Marrink, S. J.; Praprotnik, M. J.Chem. Phys. 2014, 140. doi: Artn 05411410.1063/1.4863329

(188) Zavadlav, J.; Podgornik, R.; Melo, M. N.; Marrink, S. J.;Praprotnik, M. Eur. Phys. J-Spec. Top 2016, 225, 1595.doi: 10.1140/epjst/e2016-60117-8

(189) Bereau, T.; Deserno, M. J. Membr. Biol. 2015, 248, 395.doi: 10.1007/s00232-014-9738-9

(190) Bereau, T.; Wang, Z. J.; Deserno, M. J. Chem. Phys. 2014,140, 115101. doi: Artn 11510110.1063/1.4867465

(191) Zamani, M.; Kremer, S. C. Ieee. Int. C. Bioinform. 2015, 1304.

(192) Saha, S.; Ekbal, A.; Sharma, S.; Bandyopadhyay, S.; Maulik,U. Adv. Intell. Syst. 2013, 182, 57.

Recent Developments in Using Molecular Dynamics Simulation Techniques to Study Biomolecules

CAO Liao-Ran1,#ZHANG Chun-Yu2,#ZHANG Ding-Lin1,#CHU Hui-Ying1ZHANG Yue-Bin1LI Guo-Hui1,*
(1Laboratory of Molecular Modeling and Design, State key Laboratory of Molecular Reaction Dynamics, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, Liaoning Province, P. R. China;2Liaohe Oil Field General Hospital, Panjin 124010, Liaoning Province, P. R. China)

Molecular dynamics simulation (MDS) has gained increasing importance in current-day scientific research, as the supplement, guidance, or even replacement of experiments. In this review, we briefly introduce the history of the development of molecular dynamics simulation, focusing on recent progress including new-generation force fields, modern enhanced sampling schemes, and application for the investigation of complex biomolecules.

Molecular dynamics simulation; Force field; Enhanced sampling; Enzyme reaction

his Ph.D degree from the key laboratory for molecular enzymology & engineering, Jilin University in 2010. From 2010-2013, he worked as a post-doc at the state key laboratory of superamolecular structure and Materials,Jilin University. From July 2013, He joined the state key laboratory of molecular reaction dynamics, the Dalianinstitute of chemical physics,Chinese Academy of Sciences. His research aims at the understanding of the physics and function of proteins, protein complexes, and other biomolecular structures at the atomic level using computer simulations.

December 1, 2016; Revised: March 21, 2017; Published online: April 14, 2017.

her PhD degree from Jinlin University in 2009. Since 2009, she has been working in Dalian Institute of Chemical Physics, Chinese Academy of Sciences as an associate professor, a lecturer and a postdoctoral researcher. Her research interests now mainly focus on the construction of polarizable force field.

O641

10.1080/08927028908031382

[Review]

10.3866/PKU.WHXB201704144 www.whxb.pku.edu.cn

#These authors contribute equally to this paper.

*Corresponding author. Email: ghli@dicp.ac.cn; Tel: +86-41-84379875.

The project was supported by the National Natural Science Foundation of China (21573217, 91430110, 31370714, 21625302).国家自然科学基金(21573217, 91430110, 31370714, 21625302)资助项目

© Editorial office of Acta Physico-Chimica Sinica

CAO Liao-Ran, has been working in Dr.Liʹs research group since 2015 and his research mainly focuses on using enhanced sampling techniques to study complex biochemical processes, including enzyme reactions and conformational change of macromolecules.

CHU Hui-Ying,