Bias-exchange Metadynamics Simulations Research Articles

Influence of force field choice on the conformational landscape of rat and human islet amyloid polypeptide.

Human islet amyloid polypeptide (hIAPP) is a naturally occurring, intrinsically disordered protein (IDP) whose abnormal aggregation into toxic soluble oligomers and insoluble amyloid fibrils is a pathological feature in type-2 diabetes. Rat IAPP (rIAPP) differs from hIAPP by only six amino acids yet has a reduced tendency to aggregate or form fibrils. The structures of the monomeric forms of IAPP are difficult to characterize due to their intrinsically disordered nature. Molecular dynamics simulations can provide a detailed characterization of the monomeric forms of rIAPP and hIAPP in near-physiological conditions. In this work, the conformational landscapes of rIAPP and hIAPP as a function of secondary structure content were predicted using well-tempered bias exchange metadynamics simulations. Several combinations of commonly used biomolecular force fields and water models were tested. The predicted conformational preferences of both rIAPP and hIAPP are typical of IDPs, exhibiting dominant random coil structures but showing a low propensity for transient α-helical conformations. Predicted nuclear magnetic resonance Cα chemical shifts reveal different preferences with each force field towards certain conformations, with AMBERff99SBnmr2/TIP4Pd showing the best agreement with the experiment. Comparisons of secondary structure content demonstrate residue-specific differences between hIAPP and rIAPP that may reflect their different aggregation propensities.

A thermodynamic investigation of amyloid precursor protein processing by human γ-secretase

Human γ-secretase cleaves the transmembrane domains (TMDs) of amyloid precursor protein (APP) into pathologically relevant amyloid-β peptides (Aβs). The detailed mechanisms of the unique endoproteolytic cleavage by the presenilin 1 domain (PS1) of γ-secretase are still poorly understood. Herein, we provide thermodynamic insights into how the α-helical APP TMD is processed by γ-secretase and elucidate the specificity of Aβ48/Aβ49 cleavage using unbiased molecular dynamics and bias-exchange metadynamics simulations. The thermodynamic data show that the unwinding of APP TMD is driven by water hydration in the intracellular pocket of PS1, and the scissile bond T32-L33 or L33-V34 of the APP TMD can slide down and up to interact with D257/D385 to achieve endoproteolysis. In the wild-type system, the L33-V34 scissile bond is more easily hijacked by D257/D385 than T32-L33, resulting in higher Aβ49 cleavage, while the T32N mutation on the APP TMD decreases the energy barrier of the sliding of the scissile bonds and increases the hydrogen bond occupancy for Aβ48 cleavage. In summary, the thermodynamic analysis elucidates possible mechanisms of APP TMD processing by PS1, which might facilitate rational drug design targeting γ-secretase.

In Silico Engineering of Hydrate Anti-agglomerant Molecules Using Bias-Exchange Metadynamics Simulations

We use bias-exchange metadynamics simulations to quantify the adsorption of Antiagglomerant molecules to the surface of sII methane-propane gas hydrate. Multiple steps of sampling binding configura...

Interplay of hydrophobic and hydrophilic interactions in sequence-dependent cell penetration of spontaneous membrane-translocating peptides revealed by bias-exchange metadynamics simulations

Spontaneous Membrane Translocating Peptides (SMTPs) can translocate silently across the bilayer and, thus, have the best potential to improve the delivery of therapeutic molecules to cells without toxicity. However, how their translocation mechanisms are affected by a specific peptide sequence remains poorly understood. Here, bias-exchange metadynamics simulations were employed to investigate the translocation mechanisms of five SMTPs with the same composition of amino acids (LLRLR, LRLLR, LLLRR, RLLLR, and LRLRL). Simulation results yield sequence-dependent free energy barrier using the FESs along the z-directional distance. An in-depth analysis of sequence-dependent interactions in different regions of the bilayers indicates that the free-energy barrier height of a specific sequence is resulted from the accessibility balance of isolated or clustered hydrophobic residues (L) and hydrophilic residues (R) that leads to different levels of resistance for moving of a peptide into the hydrophobic center of the membrane. At the maximal of the free-energy barrier, all peptides have a conformation parallel to the membrane surface with the barrier height determined by their affinity to the hydrophobic region. The appropriate bilayer perturbation and GDM+ pairing are beneficial for peptide translocation. These results provide an improved microscopic understanding of how peptide sequence influences the translocation efficiency and mechanism.

Different effects of cholesterol on membrane permeation of arginine and tryptophan revealed by bias-exchange metadynamics simulations.

Experiments have shown that cholesterol influences the membrane permeability of small molecules, amino acids, and cell-penetrating peptides. However, their exact translocation mechanisms under the influence of cholesterol remain poorly understood. Given the practical importance of cell-penetrating peptides and the existence of varied cholesterol contents in different cell types, it is necessary to examine the permeation of amino acids in cholesterol-containing membranes at atomic level of details. Here, bias-exchange metadynamics simulations were employed to investigate the molecular mechanism of the membrane permeation of two amino acids Arg and Trp important for cell-penetrating peptides in the presence of different concentrations of cholesterol. We found that the free energy barrier of Arg+ (the protonated form) permeation increased linearly as the cholesterol concentration increased, whereas the barrier of Trp permeation had a rapid increase from 0 mol. % to 20 mol. % cholesterol-containing membranes and nearly unchanged from 20 mol. % to 40 mol. % cholesterol-containing membranes. Arg0 becomes slightly more stable than Arg+ at the center of the dipalmitoylphosphatidylcholine (DPPC) membrane with 40 mol. % cholesterol concentrations. As a result, Arg+ has a similar permeability as Trp at 0 mol. % and 20 mol. % cholesterol, but a significantly lower permeability than Trp at 40 mol. % cholesterol. This difference is caused by the gradual reduction of water defects for Arg+ as the cholesterol concentration increases but lack of water defects for Trp in cholesterol-containing membranes. Strong but different orientation dependence between Arg+ and Trp permeations is observed. These results provide an improved microscopic understanding of amino-acid permeation through cholesterol-containing DPPC membrane systems.

Proline 411 biases the conformation of the intrinsically disordered plant UVR8 photoreceptor C27 domain altering the functional properties of the peptide

UVR8 (UV RESISTANCE LOCUS 8) is a UV-B photoreceptor responsible for initiating UV-B signalling in plants. UVR8 is a homodimer in its signalling inactive form. Upon absorption of UV radiation, the protein monomerizes into its photoactivated state. In the monomeric form, UVR8 binds the E3 ubiquitin ligase COP1 (CONSTITUTIVELY PHOTOMORPHOGENIC 1), triggering subsequent UV-B-dependent photomorphogenic development in plants. Recent in vivo experiments have shown that the UVR8 C-terminal region (aa 397–423; UVR8C27) alone is sufficient to regulate the activity of COP1. In this work, CD spectroscopy and NMR experiments showed that the UVR8C27 domain was non-structured but gained secondary structure at higher temperatures leading to increased order. Bias-exchange metadynamics simulations were also performed to evaluate the free energy landscape of UVR8C27. An inverted free energy landscape was revealed, with a disordered structure in the global energy minimum. Flanking the global energy minimum, more structured states were found at higher energies. Furthermore, stabilization of the low energy disordered state was attributed to a proline residue, P411, as evident from P411A mutant data. P411 is also a key residue in UVR8 binding to COP1. UVR8C27 is therefore structurally competent to function as a molecular switch for interaction of UVR8 with different binding partners since at higher free energies different structural conformations are being induced in this peptide. P411 has a key role for this function.

AMBER and CHARMM Force Fields Inconsistently Portray the Microscopic Details of Phosphorylation.

Phosphorylation of serine, threonine, and tyrosine is one of the most frequently occurring and crucial post-translational modifications of proteins often associated with important structural and functional changes. We investigated the direct effect of phosphorylation on the intrinsic conformational preferences of amino acids as a potential trigger of larger structural events. We conducted a comparative study of force fields on terminally capped amino acids (dipeptides) as the simplest model for phosphorylation. Our bias-exchange metadynamics simulations revealed that all model dipeptides sampled a great heterogeneity of ensembles affected by introduction of mono- and dianionic phosphate groups. However, the detected changes in populations of backbone conformers and side-chain rotamers did not reveal a strong discriminatory shift in preferences, as could be anticipated for the bulky, charged phosphate group. Furthermore, the AMBER and CHARMM force fields provided inconsistent populations of individual conformers as well as net structural trends upon phosphorylation. Detailed analysis of ensembles revealed competition between hydration and formation of internal hydrogen bonds involving amide hydrogens and the phosphate group. The observed difference in hydration free energy and potential for hydrogen bonding in individual force fields could be attributed to the different partial atomic charges used in each force field and, hence, the different parametrization strategies. Nevertheless, conformational propensities and net structural changes upon phosphorylation are difficult to extract from experimental measurements, and existing experimental data provide limited guidance for force field assessment and further development.

Bias-Exchange Metadynamics Simulation of Membrane Permeation of 20 Amino Acids.

Thermodynamics of the permeation of amino acids from water to lipid bilayers is an important first step for understanding the mechanism of cell-permeating peptides and the thermodynamics of membrane protein structure and stability. In this work, we employed bias-exchange metadynamics simulations to simulate the membrane permeation of all 20 amino acids from water to the center of a dipalmitoylphosphatidylcholine (DPPC) membrane (consists of 256 lipids) by using both directional and torsion angles for conformational sampling. The overall accuracy for the free energy profiles obtained is supported by significant correlation coefficients (correlation coefficient at 0.5–0.6) between our results and previous experimental or computational studies. The free energy profiles indicated that (1) polar amino acids have larger free energy barriers than nonpolar amino acids; (2) negatively charged amino acids are the most difficult to enter into the membrane; and (3) conformational transitions for many amino acids during membrane crossing is the key for reduced free energy barriers. These results represent the first set of simulated free energy profiles of membrane crossing for all 20 amino acids.

Equilibrium Ensembles for Insulin Folding from Bias-Exchange Metadynamics

Earliest events in the aggregation process, such as single molecule reconfiguration, are extremely important and the most difficult to characterize in experiments. To this end, we have used well-tempered bias exchange metadynamics simulations to determine the equilibrium ensembles of an insulin molecule under amyloidogenic conditions of low pH and high temperature. A bin-based clustering method that uses statistics accumulated in bias exchange metadynamics trajectories was employed to construct a detailed thermodynamic and kinetic model of insulin folding. The highest lifetime, lowest free-energy ensemble identified consisted of native conformations adopted by a folded insulin monomer in solution, namely, the R-, the Rf-, and the T-states of insulin. The lowest free-energy structure had a root mean square deviation of only 0.15 nm from native x-ray structure. The second longest-lived metastable state was an unfolded, compact monomer with little similarity to the native structure. We have identified three additional long-lived, metastable states from the bin-based model. We then carried out an exhaustive structural characterization of metastable states on the basis of tertiary contact maps and per-residue accessible surface areas. We have also determined the lowest free-energy path between two longest-lived metastable states and confirm earlier findings of non-two-state folding for insulin through a folding intermediate. The ensemble containing the monomeric intermediate retained 58% of native hydrophobic contacts, however, accompanied by a complete loss of native secondary structure. We have discussed the relative importance of nativelike versus nonnative tertiary contacts for the folding transition. We also provide a simple measure to determine the importance of an individual residue for folding transition. Finally, we have compared and contrasted this intermediate with experimental data obtained in spectroscopic, crystallographic, and calorimetric measurements during early stages of insulin aggregation. We have also determined stability of monomeric insulin by incubation at a very low concentration to isolate protein-protein interaction effects.

Mechanics of water pore formation in lipid membrane under electric field

Transmembrane water pores are crucial for substance transport through cell membranes via membrane fusion, such as in neural communication. However, the molecular mechanism of water pore formation is not clear. In this study, we apply all-atom molecular dynamics and bias-exchange metadynamics simulations to study the process of water pore formation under an electric field. We show that water molecules can enter a membrane under an electric field and form a water pore of a few nanometers in diameter. These water molecules disturb the interactions between lipid head groups and the ordered arrangement of lipids. Following the movement of water molecules, the lipid head groups are rotated and driven into the hydrophobic region of the membrane. The reorientated lipid head groups inside the membrane form a hydrophilic surface of the water pore. This study reveals the atomic details of how an electric field influences the movement of water molecules and lipid head groups, resulting in water pore formation.

Uptake Dynamics in the LacY Membrane Protein Transporter

Membrane protein transporters govern important cellular processes and are therefore central to human health. To accomplish transport, these proteins rearrange their structures to alternatively expose an internal binding site to either side of the membrane. Recent advances in protein structural determination methods have resulted in a steadily increasing number of high-resolution structures of membrane transporters trapped in different intermediate states. However, to understand the underlying transport mechanism, the molecular details of uptake and release need to be determined. We have used specialized simulation hardware to simulate uptake of galactoside sugar into the Lactose permease (LacY) of Escherichia coli. The extended brute-force simulation revealed large-scale structural rearrangements, lipid and amino acid interactions, and hydration associated with sugar uptake. The free energy landscape of sugar entry was determined by parallel bias-exchange metadynamics simulations and identified a global free energy minimum that coincided with the crystallographic binding site and also a local free energy minimum in the larger periplasmic cavity of LacY. Together, our observations show a putative molecular mechanism for sugar uptake in this prototype membrane transporter.

Allosteric-Activation Mechanism of Bovine Chymosin Revealed by Bias-Exchange Metadynamics and Molecular Dynamics Simulations.

The aspartic protease, bovine chymosin, catalyzes the proteolysis of κ-casein proteins in milk. The bovine chymosin-κ-casein complex is of industrial interest as the enzyme is used extensively in the manufacturing of processed dairy products. The apo form of the enzyme adopts a self-inhibited conformation in which the side chain of Tyr77 occludes the binding site. On the basis of kinetic, mutagenesis, and crystallographic data, it has been widely reported that a HPHPH sequence in the P8-P4 residues of the natural substrate κ-casein acts as the allosteric activator, but the mechanism by which this occurs has not previously been elucidated due to the challenges associated with studying this process by experimental methods. Here we have employed two computational techniques, molecular dynamics and bias-exchange metadynamics simulations, to study the mechanism of allosteric activation and to compute the free energy surface for the process. The simulations reveal that allosteric activation is initiated by interactions between the HPHPH sequence of κ-casein and a small α-helical region of chymosin (residues 112-116). A small conformational change in the α-helix causes the side chain of Phe114 to vacate a pocket that may then be occupied by the side chain of Tyr77. The free energy surface for the self-inhibited to open transition is significantly altered by the presence of the HPHPH sequence of κ-casein.

Structural Diversity of Ligand-Binding Androgen Receptors Revealed by Microsecond Long Molecular Dynamics Simulations and Enhanced Sampling.

Androgen receptor (AR) plays important roles in the development of prostate cancer (PCa). The antagonistic drugs, which suppress the activity of AR, are widely used in the treatment of PCa. However, the molecular mechanism of antagonism about how ligands affect the structures of AR remains elusive. To better understand the conformational variability of ARs bound with agonists or antagonists, we performed long time unbiased molecular dynamics (MD) simulations and enhanced sampling simulations for the ligand binding domain of AR (AR-LBD) in complex with various ligands. Based on the simulation results, we proposed an allosteric pathway linking ligands and helix 12 (H12) of AR-LBD, which involves the interactions among the ligands and the residues W741, H874, and I899. The interaction pathway provides an atomistic explanation of how ligands affect the structure of AR-LBD. A repositioning of H12 was observed, but it is facilitated by the C-terminal of H12, instead of by the loop between helix 11 (H11) and H12. The bias-exchange metadynamics simulations further demonstrated the above observations. More importantly, the free energy profiles constructed by the enhanced sampling simulations revealed the transition process between the antagonistic form and agonistic form of AR-LBD. Our results would be helpful for the design of more efficient antagonists of AR to combat PCa.

Fusing simulation and experiment: The effect of mutations on the structure and activity of the influenza fusion peptide.

During the infection process, the influenza fusion peptide (FP) inserts into the host membrane, playing a crucial role in the fusion process between the viral and host membranes. In this work we used a combination of simulation and experimental techniques to analyse the molecular details of this process, which are largely unknown. Although the FP structure has been obtained by NMR in detergent micelles, there is no atomic structure information in membranes. To answer this question, we performed bias-exchange metadynamics (BE-META) simulations, which showed that the lowest energy states of the membrane-inserted FP correspond to helical-hairpin conformations similar to that observed in micelles. BE-META simulations of the G1V, W14A, G12A/G13A and G4A/G8A/G16A/G20A mutants revealed that all the mutations affect the peptide’s free energy landscape. A FRET-based analysis showed that all the mutants had a reduced fusogenic activity relative to the WT, in particular the mutants G12A/G13A and G4A/G8A/G16A/G20A. According to our results, one of the major causes of the lower activity of these mutants is their lower membrane affinity, which results in a lower concentration of peptide in the bilayer. These findings contribute to a better understanding of the influenza fusion process and open new routes for future studies.

Mapping the Dynamics Landscape of Conformational Transitions in Enzyme: The Adenylate Kinase Case

Conformational transition describes the essential dynamics and mechanism of enzymes in pursuing their various functions. The fundamental and practical challenge to researchers is to quantitatively describe the roles of large-scale dynamic transitions for regulating the catalytic processes. In this study, we tackled this challenge by exploring the pathways and free energy landscape of conformational changes in adenylate kinase (AdK), a key ubiquitous enzyme for cellular energy homeostasis. Using explicit long-timescale (up to microseconds) molecular dynamics and bias-exchange metadynamics simulations, we determined at the atomistic level the intermediate conformational states and mapped the transition pathways of AdK in the presence and absence of ligands. There is clearly chronological operation of the functional domains of AdK. Specifically in the ligand-free AdK, there is no significant energy barrier in the free energy landscape separating the open and closed states. Instead there are multiple intermediate conformational states, which facilitate the rapid transitions of AdK. In the ligand-bound AdK, the closed conformation is energetically most favored with a large energy barrier to open it up, and the conformational population prefers to shift to the closed form coupled with transitions. The results suggest a perspective for a hybrid of conformational selection and induced fit operations of ligand binding to AdK. These observations, depicted in the most comprehensive and quantitative way to date, to our knowledge, emphasize the underlying intrinsic dynamics of AdK and reveal the sophisticated conformational transitions of AdK in fulfilling its enzymatic functions. The developed methodology can also apply to other proteins and biomolecular systems.

Atomic‐resolution dissection of the energetics and mechanism of isomerization of hydrated ATP‐Mg2+ through the SOMA string method

The atomic mechanisms of isomerization of ATP-Mg(2+) in solution are characterized using the recently developed String Method with Optimal Molecular Alignment (SOMA) and molecular-dynamics simulations. Bias-Exchange Metadynamics simulations are first performed to identify the primary conformers of the ATP-Mg(2+) complex and their connectivity. SOMA is then used to elucidate the minimum free-energy path (MFEP) for each transition, in a 48-dimensional space. Analysis of the per-atom contributions to the global free-energy profiles reveals that the mechanism of these transitions is controlled by the Mg(2+) ion and its coordinating oxygen atoms in the triphosphate moiety, as well as by the ion-hydration shell. Metadynamics simulations in path collective variables based on the MFEP demonstrate these isomerizations proceed across a narrow channel of configurational space, thus validating the premise underlying SOMA. This study provides a roadmap for the examination of conformational changes in biomolecules, based on complementary enhanced-sampling techniques with different strengths. © 2015 Wiley Periodicals, Inc.

Bias-Exchange Metadynamics Simulations: An Efficient Strategy for the Analysis of Conduction and Selectivity in Ion Channels.

Conduction through ion channels possesses two interesting features: (i) different ionic species are selected with high-selectivity and (ii) ions travel across the channel with rates approaching free-diffusion. Molecular dynamics simulations have the potential to reveal how these processes take place at the atomic level. However, analysis of conduction and selectivity at atomistic detail is still hampered by the short time scales accessible by computer simulations. Several algorithms have been developed to "accelerate" sampling along the slow degrees of freedom of the process under study and thus to probe longer time scales. In these algorithms, the slow degrees of freedom need to be defined in advance, which is a well-known shortcoming. In the particular case of ion conduction, preliminary assumptions about the number and type of ions participating in the permeation process need to be made. In this study, a novel approach for the analysis of conduction and selectivity based on bias-exchange metadynamics simulations was tested. This approach was compared with umbrella sampling simulations, using a model of a Na(+)-selective channel. Analogous conclusions resulted from both techniques, but the computational cost of bias-exchange simulations was lower. In addition, with bias-exchange metadynamics it was possible to calculate free energy profiles in the presence of a variable number and type of permeating ions. This approach might facilitate the definition of the set of collective variables required to analyze conduction and selectivity in ion channels.

Conduction and Selectivity in Na+ Channels Analyzed by Bias-Exchange Metadynamics Simulations

Bacterial Na+ channels have been the subjects of numerous computational studies since the first experimental structure of a Na+ selective channel was solved in 2011. Molecular Dynamics simulations revealed the presence of 2 binding sites for Na+ ions, respectively at the intracellular and extracellular entrance of the selectivity filter, separated by low energy barriers. While there is a general agreement about these features, there are also important differences among the various computational studies. In particular, ion conduction has been described both as a 2-ions or a 3-ions process, and this difference has been correlated to the direction of conduction, or to the state of the intracellular gate. A current limit of the computational strategies usually adopted to estimate the energy profiles for permeation events, is that the number of permeating ions has to be defined in advance. As consequence, it is difficult to compare energetically conduction mechanisms characterized by different number of ions, and this could explain the lack of congruence in the literature. In order to overcome this limit, we tested a novel approach for the analysis of ion conduction based on bias-exchange metadynamics simulations. In bias-exchange, several replicas of the system are simulated in parallel. A metadynamics simulation is performed for each replica, along one or a few collective variable, and at fixed time intervals swaps of configurations between replicas are attempted. Using this approach it was possible to analyze by a single set of simulations the free energy for permeation events with different number of ions. The analysis revealed that several conduction mechanisms are indeed possible for Na+ channels. This computational strategy could find wide applications for the study of ion channels, in particular to characterize conduction of ion-mixtures, or channels that exhibit heterogeneous conduction events.

The Computational Studies of Co-Translational Protein Folding

Protein synthesis in all kingdoms of life occurs on the ribosome, a macromolecular machine of >2.5 MDa. Most of the current knowledge about the crucial process of protein folding is based on in vitro investigation of isolated polypeptide chains, which typically consider the refolding of full length proteins previously denaturated by various chemical or thermal conditions. However, in vitro folding is likely to differ from in vivo, as in the latter proteins start to fold while they are still gradually emerging through the ribosomal exit tunnel. The available conformational space to the nascent polypeptide chain is therefore different from that of the full-length protein.Excellent system to study impact of protein vectorial synthesis on protein structure and dynamics are protein C-terminal truncations. Specifically, carried in our group NMR and computational investigations of C-terminal truncations (delta4 and delta6) of an immunoglobulin fold - ddFLN, the F-actin cross-linking gelation factor from Dictyostelium discoideum - are allowing us to present, the energy landscape that emerges from these co-translational folding mimetics In my work I provide atomistic details to this energy landscape by carrying out bias exchange metadynamics simulations with chemical shifts used as structural restrains. Using this approach I overcome, in large part, the two main limitations of molecular dynamics simulations in structural studies of proteins, firstly the inaccuracies in the use of force field alone and secondly the limitations inherent in sampling of conformational space.My study presents a structural and dynamical characterization of the free energy landscape of this C-terminal truncation as well as changes in the landscape, while a protein is folding in the vectorial manner. Hence, it is providing insights into a better understanding of co-translational folding, which still represents a major open problem in molecular biology.

Toward structure prediction of cyclic peptides.

Cyclic peptides are a promising class of molecules that can be used to target specific protein-protein interactions. A computational method to accurately predict their structures would substantially advance the development of cyclic peptides as modulators of protein-protein interactions. Here, we develop a computational method that integrates bias-exchange metadynamics simulations, a Boltzmann reweighting scheme, dihedral principal component analysis and a modified density peak-based cluster analysis to provide a converged structural description for cyclic peptides. Using this method, we evaluate the performance of a number of popular protein force fields on a model cyclic peptide. All the tested force fields seem to over-stabilize the α-helix and PPII/β regions in the Ramachandran plot, commonly populated by linear peptides and proteins. Our findings suggest that re-parameterization of a force field that well describes the full Ramachandran plot is necessary to accurately model cyclic peptides.