NPEX: Never give up protein exploration with deep reinforcement learning

Elucidating unknown structures of proteins, such as metastable states, is critical in designing therapeutic agents. Protein structure exploration has been performed using advanced computational methods, especially molecular dynamics and Markov chain Monte Carlo simulations, which require untenably long calculation times and prior structural knowledge. Here, we developed an innovative method for protein structure determination called never give up protein exploration (NPEX) with deep reinforcement learning. The NPEX method leverages the soft actor-critic algorithm and the intrinsic reward system, effectively adding a bias potential without the need for prior knowledge. To demonstrate the method's effectiveness, we applied it to four models: a double well, a triple well, the alanine dipeptide, and the tryptophan cage. Compared with Markov chain Monte Carlo simulations, NPEX had markedly greater sampling efficiency. The significantly enhanced computational efficiency and lack of prior domain knowledge requirements of the NPEX method will revolutionize protein structure exploration.

An interoperable implementation of collective-variable based enhanced sampling methods in extended phase space within the OpenMM package

Collective variable (CV)-based enhanced sampling techniques are widely used today for accelerating barrier-crossing events in molecular simulations. A class of these methods, which includes temperature accelerated molecular dynamics (TAMD)/driven-adiabatic free energy dynamics (d-AFED), unified free energy dynamics (UFED), and temperature accelerated sliced sampling (TASS), uses an extended variable formalism to achieve quick exploration of conformational space. These techniques are powerful, as they enhance the sampling of a large number of CVs simultaneously compared to other techniques. Extended variables are kept at a much higher temperature than the physical temperature by ensuring adiabatic separation between the extended and physical subsystems and employing rigorous thermostatting. In this work, we present a computational platform to perform extended phase space enhanced sampling simulations using the open-source molecular dynamics engine OpenMM. The implementation allows users to have interoperability of sampling techniques, as well as employ state-of-the-art thermostats and multiple time-stepping. This work also presents protocols for determining the critical parameters and procedures for reconstructing high-dimensional free energy surfaces. As a demonstration, we present simulation results on the high dimensional conformational landscapes of the alanine tripeptide in vacuo, tetra-N-methylglycine (tetra-sarcosine) peptoid in implicit solvent, and the Trp-cage mini protein in explicit water.

Energetics and J-coupling constants for Ala, Gly, and Val peptides demonstrated using ABEEM polarizable force field in vacuo and an aqueous solution

The development of an atom-bond electronegativity equalisation method at the σπ-level (ABEEM) polarisable force field (PFF) for peptides is presented. ABEEM PFF utilises a fluctuating charge model to explicitly describe...

Boosting the conformational sampling by combining replica exchange with solute tempering and well-sliced metadynamics

Methods that combine collective variable (CV) based enhanced sampling and global tempering approaches are used in speeding-up the conformational sampling and free energy calculation of large and soft systems with a plethora of energy minima. In this paper, a new method of this kind is proposed in which the well-sliced metadynamics approach (WSMTD) is united with replica exchange with solute tempering (REST2) method. WSMTD employs a divide-and-conquer strategy wherein high-dimensional slices of a free energy surface are independently sampled and combined. The method enables one to accomplish a controlled exploration of the CV-space with a restraining bias as in umbrella sampling, and enhance-sampling of one or more orthogonal CVs using a metadynamics like bias. The new hybrid method proposed here enables boosting the sampling of more slow degrees of freedom in WSMTD simulations, without the need to specify associated CVs, through a replica exchange scheme within the framework of REST2. The high-dimensional slices of the probability distributions of CVs computed from the united WSMTD and REST2 simulations are subsequently combined using the weighted histogram analysis method to obtain the free energy surface. We show that the new method proposed here is accurate, improves the conformational sampling, and achieves quick convergence in free energy estimates. We demonstrate this by computing the conformational free energy landscapes of solvated alanine tripeptide and Trp-cage mini protein in explicit water.

Sampling of the conformational landscape of small proteins with Monte Carlo methods

Computer simulation provides an increasingly realistic picture of large-scale conformational change of proteins, but investigations remain fundamentally constrained by the femtosecond timestep of molecular dynamics simulations. For this reason, many biologically interesting questions cannot be addressed using accessible state-of-the-art computational resources. Here, we report the development of an all-atom Monte Carlo approach that permits the modelling of the large-scale conformational change of proteins using standard off-the-shelf computational hardware and standard all-atom force fields. We demonstrate extensive thermodynamic characterization of the folding process of the α-helical Trp-cage, the Villin headpiece and the β-sheet WW-domain. We fully characterize the free energy landscape, transition states, energy barriers between different states, and the per-residue stability of individual amino acids over a wide temperature range. We demonstrate that a state-of-the-art intramolecular force field can be combined with an implicit solvent model to obtain a high quality of the folded structures and also discuss limitations that still remain.

Pressure dependence of vibrational optical activity of model biomolecules. A computational study

Change of molecular properties with pressure is an attracting means to regulate molecular reactivity or biological activity. However, the effect is usually small and so far explored rather scarcely. To obtain a deeper insight and estimate the sensitivity of vibrational optical activity spectra to pressure-induced conformational changes, we investigate small model molecules. The Ala-Ala dipeptide, isomaltose disaccharide and adenine-uracil dinucleotide were chosen to represent three different biomolecular classes. The pressure effects were modeled by molecular dynamics and density functional theory simulations. The dinucleotide was found to be the most sensitive to the pressure, whereas for the disaccharide the smallest changes are predicted. Pressure-induced relative intensity changes in vibrational circular dichroism and Raman optical activity spectra are predicted to be 2-3-times larger than for non-polarized IR and Raman techniques.

Hydrophobicity-driven unfolding of Trp-cage encapsulated between graphene sheets

Understanding the interaction between proteins and graphene not only helps elucidate the behaviors of proteins in confined geometries, but is also imperative to the development of a plethora of graphene-based biotechnologies, such as the graphene liquid cell transmission electron microscopy. To discuss the overall geometrical-thermal effects on proteins, we performed molecular dynamics simulations of hydrated Trp-cage miniprotein sandwiched between two graphene sheets and in the bulk environment at the temperatures below and above its unfolding temperature. The structural fluctuations of Trp-cage were characterized using the backbone root mean square displacement and the radius of gyration, from which the free energy landscape of Trp-cage was further constructed. We observed that at both temperatures the confined protein became adsorbed to the graphene surfaces and exhibited unfolded structures. Residue-specific analyses clearly showed the preference for the graphene to interact with the hydrophobic regions of Trp-cage. These results suggested that the conformation space accessible to the protein results from the competition between the thermodynamic driving forces and the geometrical restraints. While confinement usually tends to restrict the conformation of proteins by volume exclusion, it may also induce the unfolding of proteins by hydrophobic interactions.

Understanding the Nature of Amino Acid Interactions with Pd(111) or Pd-Au Bimetallic Catalysts in the Aqueous Phase

The interaction of methionine (Met) with different bimetallic-segregated surfaces comprising a uniform distribution of strips and islands of Au on the Pd(111) surface was examined using molecular dynamics (MD) simulations. Out of all the segregated and uniformly doped surfaces studied, the design of Pd-Au islands showed some reduction in the interaction energy (E_int = -43.7 kJ/mol) as compared to that of the pure Pd(111) surface (E_int = -50 kJ/mol) for a single Met molecule. However, at a higher coverage of 9 Met molecules/simulation cell, none of the Pd-Au alloy surfaces showed any improvement as compared to the Pd(111) surface. In order to develop a comprehensive understanding of the nature of the nonbonded interaction of aqueous biogenic impurities with the Pd catalyst surface, the MD study was extended to include a variety of aliphatic, S-containing, aromatic, and polar amino acids. The potential of mean force (PMF) profiles were observed to be distinct for each class of amino acids with substantial differences among amino acids with acidic and basic side chains. The side chains of all the polar and aromatic amino acids showed direct contact with the surface while aliphatic amino acids had their hydrophobic side chain aligned away from the surface. Interestingly, lysine (Lys) and tyrosine (Tyr) were the only two amino acids which interacted preferentially via the distant backbone nitrogen and backbone oxygen, respectively, despite their side chains being in direct contact with the metal surface. The strength of interaction was correlated with the size of the amino acid; the interaction energies were observed to be the maximum for large molecules such as arginine (Arg, E_int = -87.7 kJ/mol) and tryptophan (Trp, E_int = -73.4 kJ/mol), while it was a minimum for aliphatic amino acids such as alanine (Ala, E_int = -10.9 kJ/mol). The study is focused on examining the sensitivity of the choice of the preferential interaction site, conformational preferences, and interaction energies to the side-chain specificity.

Shielding effect in protein folding

One of the most important interactions responsible for protein folding and stability are hydrogen bonds between peptide groups. There is a constant competition between the water molecules and peptide groups in a hydrogen bond formation. Also side-chains take part in this process by reducing hydration of peptide group (shielding effect) that promotes the protein folding. In this paper, a new approach to take into account a shielding effect is presented. A modification of the energy function is derived and incorporated into the UNited RESidue (UNRES) force field. Canonical Molecular Dynamics and Replica Exchange Molecular Dynamics with UNRES force field is applied to study the influence of this effect on protein structure, folding kinetics and free energy landscapes. The results of test calculations suggest that even small contribution of this effect into energy function changes force field behavior as well as speeds up the folding process significantly.

Influence of Hyperglycemic Conditions on Self-Association of the Alzheimer's Amyloid β (Aβ1-42) Peptide

Clinical studies have identified a correlation between type-2 diabetes mellitus and cognitive decrements en route to the onset of Alzheimer's disease (AD). Recent studies have established that post-translational modifications of the amyloid β (Aβ) peptide occur under hyperglycemic conditions; particularly, the process of glycation exacerbates its neurotoxicity and accelerates AD progression. In view of the assertion that macromolecular crowding has an altering effect on protein self-assembly, it is crucial to characterize the effects of hyperglycemic conditions via crowding on Aβ self-assembly. Toward this purpose, fully atomistic molecular dynamics simulations were performed to study the effects of glucose crowding on Aβ dimerization, which is the smallest known neurotoxic species. The dimers formed in the glucose-crowded environment were found to have weaker associations as compared to that of those formed in water. Binding free energy calculations show that the reduced binding strength of the dimers can be mainly attributed to the overall weakening of the dispersion interactions correlated with substantial loss of interpeptide contacts in the hydrophobic patches of the Aβ units. Analysis to discern the differential solvation pattern in the glucose-crowded and pure water systems revealed that glucose molecules cluster around the protein, at a distance of 5-7 Å, which traps the water molecules in close association with the protein surface. This preferential exclusion of glucose molecules and resulting hydration of the Aβ peptides has a screening effect on the hydrophobic interactions, which in turn diminishes the binding strength of the resulting dimers. Our results imply that physical effects attributed to crowded hyperglycemic environments are incapable of solely promoting Aβ self-assembly, indicating that further mechanistic studies are required to provide insights into the self-assembly of post-translationally modified Aβ peptides, known to possess aggravated toxicity, under these conditions.

The sensitivity of folding free energy landscapes of trpzips to mutations in the hydrophobic core

The sensitivity of the stability of folded states and free energy landscapes to the differences in the hydrophobic content of the core residues has been studied for the set of 16-residue trpzips, namely, Trpzip4, Trpzip5 and Trpzip6.