Combining Simulations and Solution Experiments as a Paradigm for RNA Force Field Refinement

Comprehensive Assessment of Force-Field Performance in Molecular Dynamics Simulations of DNA/RNA Hybrid Duplexes

Mixed double helices formed by RNA and DNA strands, commonly referred to as hybrid duplexes or hybrids, are essential in biological processes like transcription and reverse transcription. They are also important for their applications in CRISPR gene editing and nanotechnology. Yet, despite their significance, the hybrid duplexes have been seldom modeled by atomistic molecular dynamics methodology, and there is no benchmark study systematically assessing the force-field performance. Here, we present an extensive benchmark study of polypurine tract (PPT) and Dickerson-Drew dodecamer hybrid duplexes using contemporary and commonly utilized pairwise additive and polarizable nucleic acid force fields. Our findings indicate that none of the available force-field choices accurately reproduces all the characteristic structural details of the hybrid duplexes. The AMBER force fields are unable to populate the C3'-endo (north) pucker of the DNA strand and underestimate inclination. The CHARMM force field accurately describes the C3'-endo pucker and inclination but shows base pair instability. The polarizable force fields struggle with accurately reproducing the helical parameters. Some force-field combinations even demonstrate a discernible conflict between the RNA and DNA parameters. In this work, we offer a candid assessment of the force-field performance for mixed DNA/RNA duplexes. We provide guidance on selecting utilizable force-field combinations and also highlight potential pitfalls and best practices for obtaining optimal performance.

Comprehensive Comparison and Critical Assessment of RNA-Specific Force Fields

Molecular dynamics simulations play a pivotal role in elucidating the dynamic behaviors of RNA structures, offering a valuable complement to traditional methods such as nuclear magnetic resonance or X-ray. Despite this, the current precision of RNA force fields lags behind that of protein force fields. In this work, we systematically compared the performance of four RNA force fields (ff99bsc0χOL3, AMBERDES, ff99OL3_CMAP1, AMBERMaxEnt) across diverse RNA structures. Our findings highlight significant challenges in maintaining stability, particularly with regard to cross-strand and cross-loop hydrogen bonds. Furthermore, we observed the limitations in accurately describing the conformations of nonhelical structural motif, terminal nucleotides, and also base pairing and base stacking interactions by the tested RNA force fields. The identified deficiencies in existing RNA force fields provide valuable insights for subsequent force field development. Concurrently, these findings offer recommendations for selecting appropriate force fields in RNA simulations.

Encoding prior knowledge in ensemble refinement

The proper balancing of information from experiment and theory is a long-standing problem in the analysis of noisy and incomplete data. Viewed as a Pareto optimization problem, improved agreement with the experimental data comes at the expense of growing inconsistencies with the theoretical reference model. Here, we propose how to set the exchange rate a priori to properly balance this trade-off. We focus on gentle ensemble refinement, where the difference between the potential energy surfaces of the reference and refined models is small on a thermal scale. By relating the variance of this energy difference to the Kullback-Leibler divergence between the respective Boltzmann distributions, one can encode prior knowledge about energy uncertainties, i.e., force-field errors, in the exchange rate. The energy uncertainty is defined in the space of observables and depends on their type and number and on the thermodynamic state. We highlight the relation of gentle refinement to free energy perturbation theory. A balanced encoding of prior knowledge increases the quality and transparency of ensemble refinement. Our findings extend to non-Boltzmann distributions, where the uncertainty in energy becomes an uncertainty in information.

Boosting Ensemble Refinement with Transferable Force-Field Corrections: Synergistic Optimization for Molecular Simulations

A novel method combining the force-field fitting approach and ensemble refinement by the maximum entropy principle is presented. Its formulation allows us to continuously interpolate between these two methods, which can thus be interpreted as two limiting cases. A cross-validation procedure enables us to correctly assess the relative weight of both of them, distinguishing scenarios in which the combined approach is meaningful from those in which either ensemble refinement or force-field fitting separately prevails. The efficacy of their combination is examined for a realistic case study of RNA oligomers. Within the new scheme, molecular dynamics simulations are integrated with experimental data provided by nuclear magnetic resonance measures. We show that force-field corrections are in general superior when applied to the appropriate force-field terms but are automatically discarded by the method when applied to inappropriate force-field terms.

Accurate and Efficient SAXS/SANS Implementation Including Solvation Layer Effects Suitable for Molecular Simulations

Small-angle X-ray and neutron scattering (SAXS/SANS) provide valuable insights into the structure and dynamics of biomolecules in solution, complementing a wide range of structural techniques, including molecular dynamics simulations. As contrast-based methods, they are sensitive not only to structural properties but also to solvent-solute interactions. Their use in molecular dynamics simulations requires a forward model that should be as fast and accurate as possible. In this work, we demonstrate the feasibility of calculating SAXS and SANS intensities using a coarse-grained representation consisting of one bead per amino acid and three beads per nucleic acid, with form factors that can be corrected on the fly to account for solvation effects at no additional computational cost. By coupling this forward model with molecular dynamics simulations restrained with SAS data, it is possible to determine conformational ensembles or refine the structure and dynamics of proteins and nucleic acids in agreement with the experimental results. To assess the robustness of this approach, we applied it to gelsolin, for which we acquired SAXS data on its closed state, and to a UP1-microRNA complex, for which we used previously collected measurements. Our hybrid-resolution small-angle scattering (hySAS) implementation, being distributed in PLUMED, can be used with atomistic and coarse-grained simulations using diverse restraining strategies.

Simple Adjustment of Intranucleotide Base-Phosphate Interaction in the OL3 AMBER Force Field Improves RNA Simulations

Molecular dynamics (MD) simulations represent an established tool to study RNA molecules. The outcome of MD studies depends, however, on the quality of the force field (ff). Here we suggest a correction for the widely used AMBER OL3 ff by adding a simple adjustment of the nonbonded parameters. The reparameterization of the Lennard-Jones potential for the -H8···O5'- and -H6···O5'- atom pairs addresses an intranucleotide steric clash occurring in the type 0 base-phosphate interaction (0BPh). The nonbonded fix (NBfix) modification of 0BPh interactions (NBfix0BPh modification) was tuned via a reweighting approach and subsequently tested using an extensive set of standard and enhanced sampling simulations of both unstructured and folded RNA motifs. The modification corrects minor but visible intranucleotide clash for the anti nucleobase conformation. We observed that structural ensembles of small RNA benchmark motifs simulated with the NBfix0BPh modification provide better agreement with experiments. No side effects of the modification were observed in standard simulations of larger structured RNA motifs. We suggest that the combination of OL3 RNA ff and NBfix0BPh modification is a viable option to improve RNA MD simulations.

Multibody Terms in Protein Coarse-Grained Models: A Top-Down Perspective

Coarse-grained models allow computational investigation of biomolecular processes occurring on long time and length scales, intractable with atomistic simulation. Traditionally, many coarse-grained models rely mostly on pairwise interaction potentials. However, the decimation of degrees of freedom should, in principle, lead to a complex many-body effective interaction potential. In this work, we use experimental data on mutant stability to parametrize coarse-grained models for two proteins with and without many-body terms. We demonstrate that many-body terms are necessary to reproduce quantitatively the effects of point mutations on protein stability, particularly to implicitly take into account the effect of the solvent.

Recognizing the power of machine learning and other computational methods to accelerate progress in small molecule targeting of RNA

RNA structures regulate a wide range of processes in biology and disease, yet small molecule chemical probes or drugs that can modulate these functions are rare. Machine learning and other computational methods are well poised to fill gaps in knowledge and overcome the inherent challenges in RNA targeting, such as the dynamic nature of RNA and the difficulty of obtaining RNA high-resolution structures. Successful tools to date include principal component analysis, linear discriminate analysis, k-nearest neighbor, artificial neural networks, multiple linear regression, and many others. Employment of these tools has revealed critical factors for selective recognition in RNA:small molecule complexes, predictable differences in RNA- and protein-binding ligands, and quantitative structure activity relationships that allow the rational design of small molecules for a given RNA target. Herein we present our perspective on the value of using machine learning and other computation methods to advance RNA:small molecule targeting, including select examples and their validation as well as necessary and promising future directions that will be key to accelerate discoveries in this important field.

The development of nucleic acids force fields: From an unchallenged past to a competitive future

Molecular dynamics simulations have strongly matured as a method to study biomolecular processes. Their validity, however, is determined by the accuracy of the underlying force fields that describe the forces between all atoms. In this article, we review the development of nucleic acids force fields. We describe the early attempts in the 1990s and emphasize their strong influence on recent force fields. State-of-the-art force fields still use the same Lennard-Jones parameters derived 25 years ago in spite of the fact that these parameters were in general not fitted for nucleic acids. In addition, electrostatic parameters also are deprecated, which may explain some of the current force field deficiencies. We compare different force fields for various systems and discuss new tests of the recently developed Tumuc1 force field. The OL-force fields and Tumuc1 are arguably the best force fields to describe the DNA double helix. However, no force field is flawless. In particular, the description of sugar-puckering remains a problem for nucleic acids force fields. Future refinements are required, so we review methods for force field refinement and give an outlook to the future of force fields.

Small molecules targeting the disordered transactivation domain of the androgen receptor induce the formation of collapsed helical states

Intrinsically disordered proteins, which do not adopt well-defined structures under physiological conditions, are implicated in many human diseases. Small molecules that target the disordered transactivation domain of the androgen receptor have entered human trials for the treatment of castration-resistant prostate cancer (CRPC), but no structural or mechanistic rationale exists to explain their inhibition mechanisms or relative potencies. Here, we utilize all-atom molecular dynamics computer simulations to elucidate atomically detailed binding mechanisms of the compounds EPI-002 and EPI-7170 to the androgen receptor. Our simulations reveal that both compounds bind at the interface of two transiently helical regions and induce the formation of partially folded collapsed helical states. We find that EPI-7170 binds androgen receptor more tightly than EPI-002 and we identify a network of intermolecular interactions that drives higher affinity binding. Our results suggest strategies for developing more potent androgen receptor inhibitors and general strategies for disordered protein drug design.

IDPConformerGenerator: A Flexible Software Suite for Sampling the Conformational Space of Disordered Protein States

The power of structural information for informing biological mechanisms is clear for stable folded macromolecules, but similar structure–function insight is more difficult to obtain for highly dynamic systems such as intrinsically disordered proteins (IDPs) which must be described as structural ensembles. Here, we present IDPConformerGenerator, a flexible, modular open-source software platform for generating large and diverse ensembles of disordered protein states that builds conformers that obey geometric, steric, and other physical restraints on the input sequence. IDPConformerGenerator samples backbone phi (φ), psi (ψ), and omega (ω) torsion angles of relevant sequence fragments from loops and secondary structure elements extracted from folded protein structures in the RCSB Protein Data Bank and builds side chains from robust Monte Carlo algorithms using expanded rotamer libraries. IDPConformerGenerator has many user-defined options enabling variable fractional sampling of secondary structures, supports Bayesian models for assessing the agreement of IDP ensembles for consistency with experimental data, and introduces a machine learning approach to transform between internal and Cartesian coordinates with reduced error. IDPConformerGenerator will facilitate the characterization of disordered proteins to ultimately provide structural insights into these states that have key biological functions.

Inferring spatial source of disease outbreaks using maximum entropy

Mathematical modeling of disease outbreaks can infer the future trajectory of an epidemic, allowing for making more informed policy decisions. Another task is inferring the origin of a disease, which is relatively difficult with current mathematical models. Such frameworks, across varying levels of complexity, are typically sensitive to input data on epidemic parameters, case counts, and mortality rates, which are generally noisy and incomplete. To alleviate these limitations, we propose a maximum entropy framework that fits epidemiological models, provides calibrated infection origin probabilities, and is robust to noise due to a prior belief model. Maximum entropy is agnostic to the parameters or model structure used and allows for flexible use when faced with sparse data conditions and incomplete knowledge in the dynamical phase of disease-spread, providing for more reliable modeling at early stages of outbreaks. We evaluate the performance of our model by predicting future disease trajectories based on simulated epidemiological data in synthetic graph networks and the real mobility network of New York State. In addition, unlike existing approaches, we demonstrate that the method can be used to infer the origin of the outbreak with accurate confidence. Indeed, despite the prevalent belief on the feasibility of contact-tracing being limited to the initial stages of an outbreak, we report the possibility of reconstructing early disease dynamics, including the epidemic seed, at advanced stages.

Automatic Learning of Hydrogen-Bond Fixes in the AMBER RNA Force Field

The capability of current force fields to reproduce RNA structural dynamics is limited. Several methods have been developed to take advantage of experimental data in order to enforce agreement with experiments. Here, we extend an existing framework which allows arbitrarily chosen force-field correction terms to be fitted by quantification of the discrepancy between observables back-calculated from simulation and corresponding experiments. We apply a robust regularization protocol to avoid overfitting and additionally introduce and compare a number of different regularization strategies, namely, L1, L2, Kish size, relative Kish size, and relative entropy penalties. The training set includes a GACC tetramer as well as more challenging systems, namely, gcGAGAgc and gcUUCGgc RNA tetraloops. Specific intramolecular hydrogen bonds in the AMBER RNA force field are corrected with automatically determined parameters that we call gHBfix_opt. A validation involving a separate simulation of a system present in the training set (gcUUCGgc) and new systems not seen during training (CAAU and UUUU tetramers) displays improvements regarding the native population of the tetraloop as well as good agreement with NMR experiments for tetramers when using the new parameters. Then, we simulate folded RNAs (a kink–turn and L1 stalk rRNA) including hydrogen bond types not sufficiently present in the training set. This allows a final modification of the parameter set which is named gHBfix21 and is suggested to be applicable to a wider range of RNA systems.

Using Unassigned NMR Chemical Shifts to Model RNA Secondary Structure

NMR-derived chemical shifts are sensitive probes of RNA structure. However, the need to assign NMR spectra hampers their utility as a direct source of structural information. In this report, we describe a simple method that uses unassigned 2D NMR spectra to model the secondary structure of RNAs. As in the case of assigned chemical shifts, we could use unassigned chemical shift data to reweight conformational libraries such that the highest weighted structure closely resembles their reference NMR structure. Furthermore, the application of our approach to the 3'- and 5'-UTR of the SARS-CoV-2 genome yields structures that are, for the most part, consistent with the secondary structure models derived from chemical probing data. Therefore, we expect the framework we describe here will be useful as a general strategy for rapidly generating preliminary structural RNA models directly from unassigned 2D NMR spectra. As we demonstrated for the 337-nt and 472-nt UTRs of SARS-CoV-2, our approach could be especially valuable for modeling the secondary structures of large RNA.

Simulation-based inference with approximately correct parameters via maximum entropy

Inferring the input parameters of simulators from observations is a crucial challenge with applications from epidemiology to molecular dynamics. Here we show a simple approach in the regime of sparse data and approximately correct models, which is common when trying to use an existing model to infer latent variables with observed data. This approach is based on the principle of maximum entropy (MaxEnt) and provably makes the smallest change in the latent joint distribution to fit new data. This method requires no likelihood or model derivatives and its fit is insensitive to prior strength, removing the need to balance observed data fit with prior belief. The method requires the ansatz that data is fit in expectation, which is true in some settings and may be reasonable in all settings with few data points. The method is based on sample reweighting, so its asymptotic run time is independent of prior distribution dimension. We demonstrate this MaxEnt approach and compare with other likelihood-free inference methods across three systems: a point particle moving in a gravitational field, a compartmental model of epidemic spread and molecular dynamics simulation of a protein.

Toward Convergence in Folding Simulations of RNA Tetraloops: Comparison of Enhanced Sampling Techniques and Effects of Force Field Modifications

Atomistic molecular dynamics simulations represent an established technique for investigation of RNA structural dynamics. Despite continuous development, contemporary RNA simulations still suffer from suboptimal accuracy of empirical potentials (force fields, ffs) and sampling limitations. Development of efficient enhanced sampling techniques is important for two reasons. First, they allow us to overcome the sampling limitations, and second, they can be used to quantify ff imbalances provided they reach a sufficient convergence. Here, we study two RNA tetraloops (TLs), namely the GAGA and UUCG motifs. We perform extensive folding simulations and calculate folding free energies (ΔG_fold^°) with the aim to compare different enhanced sampling techniques and to test several modifications of the nonbonded terms extending the AMBER OL3 RNA ff. We demonstrate that replica-exchange solute tempering (REST2) simulations with 12-16 replicas do not show any sign of convergence even when extended to a timescale of 120 μs per replica. However, the combination of REST2 with well-tempered metadynamics (ST-MetaD) achieves good convergence on a timescale of 5-10 μs per replica, improving the sampling efficiency by at least 2 orders of magnitude. Effects of ff modifications on ΔG_fold^° energies were initially explored by the reweighting approach and then validated by new simulations. We tested several manually prepared variants of the gHBfix potential which improve stability of the native state of both TLs by ∼2 kcal/mol. This is sufficient to conveniently stabilize the folded GAGA TL while the UUCG TL still remains under-stabilized. Appropriate adjustment of van der Waals parameters for C-H···O5' base-phosphate interaction may further stabilize the native states of both TLs by ∼0.6 kcal/mol.

The Ad-MD method to calculate NMR shift including effects due to conformational dynamics: The 31 P NMR shift in DNA

A method for averaging of NMR parameters by molecular dynamics (MD) has been derived from the method of statistical averaging in MD snapshots, benchmarked and applied to structurally dynamic interpretation of the ³¹ P NMR shift (δ_31P ) in DNA phosphates. The method employs adiabatic dependence of an NMR parameter on selected geometric parameter(s) that is weighted by MD-calculated probability distribution(s) for the geometric parameter(s) (Ad-MD method). The usage of Ad-MD for polymers is computationally convenient when one pre-calculated structural dependence of an NMR parameter is employed for all chemically equivalent units differing only in dynamic behavior. The Ad-MD method is benchmarked against the statistical averaging method for δ_31P in the model phosphates featuring distinctively different structures and dynamic behavior. The applicability of Ad-MD is illustrated by calculating ³¹ P NMR spectra in the Dickerson-Drew DNA dodecamer. δ_31P was calculated with the B3LYP/IGLO-III/PCM(water) and the probability distributions for the torsion angles adjacent to the phosphorus atoms in the DNA phosphates were calculated using the OL15 force field.

Probabilistic Modeling of RNA Ensembles Using NMR Chemical Shifts

NMR-derived chemical shifts are structural fingerprints that are sensitive to the underlying conformational distributions of molecules. Thus, chemical shift data are now routinely used to infer the dynamical or conformational ensembles of peptides and proteins. However, for RNAs, techniques for inferring their conformational ensembles from chemical shift data have received less attention. Here, we used chemical shift data and the Bayesian/maximum entropy (BME) approach to model the secondary structure ensembles of several single-stranded RNAs. Inspection of the resulting ensembles indicates that the secondary structure of the highest weighted (most probable) conformer in the ensemble typically resembled the known NMR structure. Furthermore, using apo chemical shifts measured for the HIV-1 TAR RNA, we found that our framework reproduces the expected structure yet predicts the existence of a previously unobserved base pair, which we speculate may be sampled transiently. We expect that the chemical shift-based BME (CS-BME) framework we describe here should find utility as a general strategy for modeling RNA ensembles using chemical shift data.

Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems

Machine learning models are poised to make a transformative impact on chemical sciences by dramatically accelerating computational algorithms and amplifying insights available from computational chemistry methods. However, achieving this requires a confluence and coaction of expertise in computer science and physical sciences. This Review is written for new and experienced researchers working at the intersection of both fields. We first provide concise tutorials of computational chemistry and machine learning methods, showing how insights involving both can be achieved. We follow with a critical review of noteworthy applications that demonstrate how computational chemistry and machine learning can be used together to provide insightful (and useful) predictions in molecular and materials modeling, retrosyntheses, catalysis, and drug design.

The structural plasticity of nucleic acid duplexes revealed by WAXS and MD

Double-stranded DNA (dsDNA) and RNA (dsRNA) helices display an unusual structural diversity. Some structural variations are linked to sequence and may serve as signaling units for protein-binding partners. Therefore, elucidating the mechanisms and factors that modulate these variations is of fundamental importance. While the structural diversity of dsDNA has been extensively studied, similar studies have not been performed for dsRNA. Because of the increasing awareness of RNA's diverse biological roles, such studies are timely and increasingly important. We integrate solution x-ray scattering at wide angles (WAXS) with all-atom molecular dynamics simulations to explore the conformational ensemble of duplex topologies for different sequences and salt conditions. These tightly coordinated studies identify robust correlations between features in the WAXS profiles and duplex geometry and enable atomic-level insights into the structural diversity of DNA and RNA duplexes. Notably, dsRNA displays a marked sensitivity to the valence and identity of its associated cations.

A method of incorporating rate constants as kinetic constraints in molecular dynamics simulations

From the point of view of statistical mechanics, a full characterization of a molecular system requires an accurate determination of its possible states, their populations, and the respective interconversion rates. Toward this goal, well-established methods increase the accuracy of molecular dynamics simulations by incorporating experimental information about states using structural restraints and about populations using thermodynamics restraints. However, it is still unclear how to include experimental knowledge about interconversion rates. Here, we introduce a method of imposing known rate constants as constraints in molecular dynamics simulations, which is based on a combination of the maximum-entropy and maximum-caliber principles. Starting from an existing ensemble of trajectories, obtained from either molecular dynamics or enhanced trajectory sampling, this method provides a minimally perturbed path distribution consistent with the kinetic constraints, as well as modified free energy and committor landscapes. We illustrate the application of the method to a series of model systems, including all-atom molecular simulations of protein folding. Our results show that by combining experimental rate constants and molecular dynamics simulations, this approach enables the determination of transition states, reaction mechanisms, and free energies. We anticipate that this method will extend the applicability of molecular simulations to kinetic studies in structural biology and that it will assist the development of force fields to reproduce kinetic and thermodynamic observables. Furthermore, this approach is generally applicable to a wide range of systems in biology, physics, chemistry, and material science.

Using Cross-Correlated Spin Relaxation to Characterize Backbone Dihedral Angle Distributions of Flexible Protein Segments

Crucial to the function of proteins is their existence as conformational ensembles sampling numerous and structurally diverse substates. Despite this widely accepted notion there is still a high demand for meaningful and reliable approaches to characterize protein ensembles in solution. As it is usually conducted in solution, NMR spectroscopy offers unique possibilities to address this challenge. Particularly, cross-correlated relaxation (CCR) effects have long been established to encode both protein structure and dynamics in a compelling manner. However, this wealth of information often limits their use in practice as structure and dynamics might prove difficult to disentangle. Using a modern Maximum Entropy (MaxEnt) reweighting approach to interpret CCR rates of Ubiquitin, we demonstrate that these uncertainties do not necessarily impair resolving CCR-encoded structural information. Instead, a suitable balance between complementary CCR experiments and prior information is found to be the most crucial factor in mapping backbone dihedral angle distributions. Experimental and systematic deviations such as oversimplified dynamics appear to be of minor importance. Using Ubiquitin as an example, we demonstrate that CCR rates are capable of characterizing rigid and flexible residues alike, indicating their unharnessed potential in studying disordered proteins.

Data-Driven Molecular Dynamics: A Multifaceted Challenge

The big data concept is currently revolutionizing several fields of science including drug discovery and development. While opening up new perspectives for better drug design and related strategies, big data analysis strongly challenges our current ability to manage and exploit an extraordinarily large and possibly diverse amount of information. The recent renewal of machine learning (ML)-based algorithms is key in providing the proper framework for addressing this issue. In this respect, the impact on the exploitation of molecular dynamics (MD) simulations, which have recently reached mainstream status in computational drug discovery, can be remarkable. Here, we review the recent progress in the use of ML methods coupled to biomolecular simulations with potentially relevant implications for drug design. Specifically, we show how different ML-based strategies can be applied to the outcome of MD simulations for gaining knowledge and enhancing sampling. Finally, we discuss how intrinsic limitations of MD in accurately modeling biomolecular systems can be alleviated by including information coming from experimental data.

Structure and dynamics of a nanodisc by integrating NMR, SAXS and SANS experiments with molecular dynamics simulations

Nanodiscs are membrane mimetics that consist of a protein belt surrounding a lipid bilayer, and are broadly used for characterization of membrane proteins. Here, we investigate the structure, dynamics and biophysical properties of two small nanodiscs, MSP1D1ΔH5 and ΔH4H5. We combine our SAXS and SANS experiments with molecular dynamics simulations and previously obtained NMR and EPR data to derive and validate a conformational ensemble that represents the structure and dynamics of the nanodisc. We find that it displays conformational heterogeneity with various elliptical shapes, and with substantial differences in lipid ordering in the centre and rim of the discs. Together, our results reconcile previous apparently conflicting observations about the shape of nanodiscs, and pave the way for future integrative studies of larger complex systems such as membrane proteins embedded in nanodiscs.

Toward empirical force fields that match experimental observables

Biomolecular force fields have been traditionally derived based on a mixture of reference quantum chemistry data and experimental information obtained on small fragments. However, the possibility to run extensive molecular dynamics simulations on larger systems achieving ergodic sampling is paving the way to directly using such simulations along with solution experiments obtained on macromolecular systems. Recently, a number of methods have been introduced to automatize this approach. Here, we review these methods, highlight their relationship with machine learning methods, and discuss the open challenges in the field.

Extended Experimental Inferential Structure Determination Method in Determining the Structural Ensembles of Disordered Protein States

Proteins with intrinsic or unfolded state disorder comprise a new frontier in structural biology, requiring the characterization of diverse and dynamic structural ensembles. We introduce a comprehensive Bayesian framework, the Extended Experimental Inferential Structure Determination (X-EISD) method, that calculates the maximum log-likelihood of a disordered protein ensemble. X-EISD accounts for the uncertainties of a range of experimental data and back-calculation models from structures, including NMR chemical shifts, J-couplings, Nuclear Overhauser Effects (NOEs), paramagnetic relaxation enhancements (PREs), residual dipolar couplings (RDCs), hydrodynamic radii (R _h ), single molecule fluorescence Förster resonance energy transfer (smFRET) and small angle X-ray scattering (SAXS). We apply X-EISD to the joint optimization against experimental data for the unfolded drkN SH3 domain and find that combining a local data type, such as chemical shifts or J-couplings, paired with long-ranged restraints such as NOEs, PREs or smFRET, yields structural ensembles in good agreement with all other data types if combined with representative IDP conformers.

Fine-Tuning of the AMBER RNA Force Field with a New Term Adjusting Interactions of Terminal Nucleotides

Determination of RNA structural-dynamic properties is challenging for experimental methods. Thus, atomistic molecular dynamics (MD) simulations represent a helpful technique complementary to experiments. However, contemporary MD methods still suffer from limitations of force fields (ffs), including imbalances in the nonbonded ff terms. We have recently demonstrated that some improvement of state-of-the-art AMBER RNA ff can be achieved by adding a new term for H-bonding called gHBfix, which increases tuning flexibility and reduces risk of side-effects. Still, the first gHBfix version did not fully correct simulations of short RNA tetranucleotides (TNs). TNs are key benchmark systems due to availability of unique NMR data, although giving too much weight on improving TN simulations can easily lead to overfitting to A-form RNA. Here we combine the gHBfix version with another term called tHBfix, which separately treats H-bond interactions formed by terminal nucleotides. This allows to refine simulations of RNA TNs without affecting simulations of other RNAs. The approach is in line with adopted strategy of current RNA ffs, where the terminal nucleotides possess different parameters for terminal atoms than the internal nucleotides. Combination of gHBfix with tHBfix significantly improves the behavior of RNA TNs during well-converged enhanced-sampling simulations using replica exchange with solute tempering. TNs mostly populate canonical A-form like states while spurious intercalated structures are largely suppressed. Still, simulations of r(AAAA) and r(UUUU) TNs show some residual discrepancies with primary NMR data which suggests that future tuning of some other ff terms might be useful. Nevertheless, the tHBfix has a clear potential to improve modeling of key biochemical processes, where interactions of RNA single stranded ends are involved.

Assessing the accuracy of direct-coupling analysis for RNA contact prediction

Many noncoding RNAs are known to play a role in the cell directly linked to their structure. Structure prediction based on the sole sequence is, however, a challenging task. On the other hand, thanks to the low cost of sequencing technologies, a very large number of homologous sequences are becoming available for many RNA families. In the protein community, the idea of exploiting the covariance of mutations within a family to predict the protein structure using the direct-coupling-analysis (DCA) method has emerged in the last decade. The application of DCA to RNA systems has been limited so far. We here perform an assessment of the DCA method on 17 riboswitch families, comparing it with the commonly used mutual information analysis and with state-of-the-art R-scape covariance method. We also compare different flavors of DCA, including mean-field, pseudolikelihood, and a proposed stochastic procedure (Boltzmann learning) for solving exactly the DCA inverse problem. Boltzmann learning outperforms the other methods in predicting contacts observed in high-resolution crystal structures.

A novel high-dimensional NMR experiment for resolving protein backbone dihedral angle ambiguities

Intrinsically disordered proteins (IDPs) are challenging established structural biology perception and urge a reassessment of the conventional understanding of the subtle interplay between protein structure and dynamics. Due to their importance in eukaryotic life and central role in protein interaction networks, IDP research is a fascinating and highly relevant research area in which NMR spectroscopy is destined to be a key player. The flexible nature of IDPs, as a result of the sampling of a vast conformational space, however, poses a tremendous scientific challenge, both technically and theoretically. Pronounced signal averaging results in narrow signal dispersion and requires higher dimensionality NMR techniques. Moreover, a fundamental problem in the structural characterization of IDPs is the definition of the conformational ensemble sampled by the polypeptide chain in solution, where often the interpretation relies on the concept of 'residual structure' or 'conformational preference'. An important source of structural information is information-rich NMR experiments that probe protein backbone dihedral angles in a unique manner. Cross-correlated relaxation experiments have proven to fulfil this task as they provide unique information about protein backbones, particularly in IDPs. Here we present a novel cross-correlation experiment that utilizes non-uniform sampling detection schemes to resolve protein backbone dihedral ambiguities in IDPs. The sensitivity of this novel technique is illustrated with an application to the prototypical IDP [Formula: see text]-Synculein for which unexpected deviations from random-coil-like behaviour could be observed.

Conformational ensembles of an RNA hairpin using molecular dynamics and sparse NMR data

Solution nuclear magnetic resonance (NMR) experiments allow RNA dynamics to be determined in an aqueous environment. However, when a limited number of peaks are assigned, it is difficult to obtain structural information. We here show a protocol based on the combination of experimental data (Nuclear Overhauser Effect, NOE) and molecular dynamics simulations with enhanced sampling methods. This protocol allows to (a) obtain a maximum entropy ensemble compatible with NMR restraints and (b) obtain a minimal set of metastable conformations compatible with the experimental data (maximum parsimony). The method is applied to a hairpin of 29 nt from an inverted SINEB2, which is part of the SINEUP family and has been shown to enhance protein translation. A clustering procedure is introduced where the annotation of base-base interactions and glycosidic bond angles is used as a metric. By reweighting the contributions of the clusters, minimal sets of four conformations could be found which are compatible with the experimental data. A motif search on the structural database showed that some identified low-population states are present in experimental structures of other RNA transcripts. The introduced method can be applied to characterize RNA dynamics in systems where a limited amount of NMR information is available.

Interpretation of HDX Data by Maximum-Entropy Reweighting of Simulated Structural Ensembles

Hydrogen-deuterium exchange combined with mass spectrometry (HDX-MS) is a widely applied biophysical technique that probes the structure and dynamics of biomolecules without the need for site-directed modifications or bio-orthogonal labels. The mechanistic interpretation of HDX data, however, is often qualitative and subjective, owing to a lack of quantitative methods to rigorously translate observed deuteration levels into atomistic structural information. To help address this problem, we have developed a methodology to generate structural ensembles that faithfully reproduce HDX-MS measurements. In this approach, an ensemble of protein conformations is first generated, typically using molecular dynamics simulations. A maximum-entropy bias is then applied post hoc to the resulting ensemble such that averaged peptide-deuteration levels, as predicted by an empirical model, agree with target values within a given level of uncertainty. We evaluate this approach, referred to as HDX ensemble reweighting (HDXer), for artificial target data reflecting the two major conformational states of a binding protein. We demonstrate that the information provided by HDX-MS experiments and by the model of exchange are sufficient to recover correctly weighted structural ensembles from simulations, even when the relevant conformations are rarely observed. Degrading the information content of the target data—e.g., by reducing sequence coverage, by averaging exchange levels over longer peptide segments, or by incorporating different sources of uncertainty—reduces the structural accuracy of the reweighted ensemble but still allows for useful insights into the distinctive structural features reflected by the target data. Finally, we describe a quantitative metric to rank candidate structural ensembles according to their correspondence with target data and illustrate the use of HDXer to describe changes in the conformational ensemble of the membrane protein LeuT. In summary, HDXer is designed to facilitate objective structural interpretations of HDX-MS data and to inform experimental approaches and further developments of theoretical exchange models.

How to learn from inconsistencies: Integrating molecular simulations with experimental data

Molecular simulations and biophysical experiments can be used to provide independent and complementary insights into the molecular origin of biological processes. A particularly useful strategy is to use molecular simulations as a modeling tool to interpret experimental measurements, and to use experimental data to refine our biophysical models. Thus, explicit integration and synergy between molecular simulations and experiments is fundamental for furthering our understanding of biological processes. This is especially true in the case where discrepancies between measured and simulated observables emerge. In this chapter, we provide an overview of some of the core ideas behind methods that were developed to improve the consistency between experimental information and numerical predictions. We distinguish between situations where experiments are used to refine our understanding and models of specific systems, and situations where experiments are used more generally to refine transferable models. We discuss different philosophies and attempt to unify them in a single framework. Until now, such integration between experiments and simulations have mostly been applied to equilibrium data, and we discuss more recent developments aimed to analyze time-dependent or time-resolved data.

Computing, Analyzing, and Comparing the Radius of Gyration and Hydrodynamic Radius in Conformational Ensembles of Intrinsically Disordered Proteins

The level of compaction of an intrinsically disordered protein may affect both its physical and biological properties, and can be probed via different types of biophysical experiments. Small-angle X-ray scattering (SAXS) probe the radius of gyration (R_g) whereas pulsed-field-gradient nuclear magnetic resonance (NMR) diffusion, fluorescence correlation spectroscopy, and dynamic light scattering experiments can be used to determine the hydrodynamic radius (R_h). Here we show how to calculate R_g and R_h from a computationally generated conformational ensemble of an intrinsically disordered protein. We further describe how to use a Bayesian/Maximum Entropy procedure to integrate data from SAXS and NMR diffusion experiments, so as to derive conformational ensembles in agreement with those experiments.

Integrating Molecular Simulation and Experimental Data: A Bayesian/Maximum Entropy Reweighting Approach

We describe a Bayesian/Maximum entropy (BME) procedure and software to construct a conformational ensemble of a biomolecular system by integrating molecular simulations and experimental data. First, an initial conformational ensemble is constructed using, for example, Molecular Dynamics or Monte Carlo simulations. Due to potential inaccuracies in the model and finite sampling effects, properties predicted from simulations may not agree with experimental data. In BME we use the experimental data to refine the simulation so that the new conformational ensemble has the following properties: (1) the calculated averages are close to the experimental values taking uncertainty into account and (2) it maximizes the relative Shannon entropy with respect to the original simulation ensemble. The output of this procedure is a set of optimized weights that can be used to calculate other properties and distributions of these. Here, we provide a practical guide on how to obtain and use such weights, how to choose adjustable parameters and discuss shortcomings of the method.

Maximum Entropy Optimized Force Field for Intrinsically Disordered Proteins

Intrinsically disordered proteins (IDPs) constitute a significant fraction of eukaryotic proteomes. High-resolution characterization of IDP conformational ensembles can help elucidate their roles in a wide range of biological processes but remains challenging both experimentally and computationally. Here, we present a generic algorithm to improve the accuracy of coarse-grained IDP models using a diverse set of experimental measurements. It combines maximum entropy optimization and least-squares regression to systematically adjust model parameters and improve the agreement between simulation and experiment. We successfully applied the algorithm to derive a transferable force field, which we term the maximum entropy optimized force field (MOFF), for de novo prediction of IDP structures. Statistical analysis of force field parameters reveals features of amino acid interactions not captured by potentials designed to work well for folded proteins. We anticipate its combination of efficiency and accuracy will make MOFF useful for studying the phase separation of IDPs, which drives the formation of various biological compartments.

Accuracy of MD solvent models in RNA structure refinement assessed via liquid-crystal NMR and spin relaxation data

Molecular dynamics (MD) simulations play an important role in characterizing Ribonucleic Acid (RNA) structure, augmenting information from experimental techniques such as Nuclear Magnetic Resonance (NMR). In this work, we examine the accuracy of structural representation resulting from application of a number of explicit and implicit solvent models and refinement protocols against experimental data ranging from high density of residual dipolar coupling (RDC) restraints to completely unrestrained simulations. For a prototype A-form RNA helix, our results indicate that AMBER RNA force field with either implicit or explicit solvent can produce a realistic dynamic representation of RNA helical structure, accurately cross-validating with respect to a diverse array of NMR observables. In refinement against NMR distance restraints, modern MD force fields are found to be equally adequate, with high fidelity cross-validation to the residual dipolar couplings (RDCs) and residual chemical shift anisotropies (RCSAs), while slightly over-estimating structural order as monitored via NMR relaxation data. With restraints trimmed to encode only for base pairing information, cross-validation quality significantly deteriorates, now exhibiting a pronounced dependence on the choice of the solvent model. This deterioration is found to be partially reversible by increasing planarity restraints on the nucleobase geometry. For completely unrestrained MD simulations, the choice of water model becomes very important, with the best-performing TIP4P-Ew accurately reproducing both the RDC and RCSA data, while closely matching the NMR-derived order parameters. The information provided here will serve as a foundation for MD-based refinement of solution state NMR structures of RNA.

Determination of Structural Ensembles of Flexible Molecules in Solution from NMR Data Undergoing Spin Diffusion

Spin diffusion is a formidable problem when interpreting NMR data of chemical compounds. We developed a method to reconstruct the conformational ensemble of flexible molecules displaying spin diffusion, which minimizes the subjective bias in the interpretation of experimental data and which can be used routinely to obtain sets of structures with the correct thermodynamic weights. We showed in the case of a flexible molecule that the correct conformational ensemble is quite different from that obtained with standard methods.

Efficient Ensemble Refinement by Reweighting

Ensemble refinement produces structural ensembles of flexible and dynamic biomolecules by integrating experimental data and molecular simulations. Here we present two efficient numerical methods to solve the computationally challenging maximum-entropy problem arising from a Bayesian formulation of ensemble refinement. Recasting the resulting constrained weight optimization problem into an unconstrained form enables the use of gradient-based algorithms. In two complementary formulations that differ in their dimensionality, we optimize either the log-weights directly or the generalized forces appearing in the explicit analytical form of the solution. We first demonstrate the robustness, accuracy, and efficiency of the two methods using synthetic data. We then use NMR J-couplings to reweight an all-atom molecular dynamics simulation ensemble of the disordered peptide Ala-5 simulated with the AMBER99SB*-ildn-q force field. After reweighting, we find a consistent increase in the population of the polyproline-II conformations and a decrease of α-helical-like conformations. Ensemble refinement makes it possible to infer detailed structural models for biomolecules exhibiting significant dynamics, such as intrinsically disordered proteins, by combining input from experiment and simulation in a balanced manner.

Improving the Performance of the Amber RNA Force Field by Tuning the Hydrogen-Bonding Interactions

Molecular dynamics (MD) simulations became a leading tool for investigation of structural dynamics of nucleic acids. Despite recent efforts to improve the empirical potentials (force fields, ffs), RNA ffs have persisting deficiencies, which hamper their utilization in quantitatively accurate simulations. Previous studies have shown that at least two salient problems contribute to difficulties in the description of free-energy landscapes of small RNA motifs: (i) excessive stabilization of the unfolded single-stranded RNA ensemble by intramolecular base-phosphate and sugar-phosphate interactions and (ii) destabilization of the native folded state by underestimation of stability of base pairing. Here, we introduce a general ff term (gHBfix) that can selectively fine-tune nonbonding interaction terms in RNA ffs, in particular, the H bonds. The gHBfix potential affects the pairwise interactions between all possible pairs of the specific atom types, while all other interactions remain intact; i.e., it is not a structure-based model. In order to probe the ability of the gHBfix potential to refine the ff nonbonded terms, we performed an extensive set of folding simulations of RNA tetranucleotides and tetraloops. On the basis of these data, we propose particular gHBfix parameters to modify the AMBER RNA ff. The suggested parametrization significantly improves the agreement between experimental data and the simulation conformational ensembles, although our current ff version still remains far from being flawless. While attempts to tune the RNA ffs by conventional reparametrizations of dihedral potentials or nonbonded terms can lead to major undesired side effects, as we demonstrate for some recently published ffs, gHBfix has a clear promising potential to improve the ff performance while avoiding introduction of major new imbalances.

Development and Testing of the OPLS-AA/M Force Field for RNA

Significant improvements have been made to the OPLS-AA force field for modeling RNA. New torsional potentials were optimized based on density functional theory (DFT) scans at the ωB97X-D/6-311++G(d,p) level for potential energy surfaces of the backbone α and γ dihedral angles. In combination with previously reported improvements for the sugar puckering and glycosidic torsion terms, the new force field was validated through diverse molecular dynamics simulations for RNAs in aqueous solution. Results for dinucleotides and tetranucleotides revealed both accurate reproduction of 3 J couplings from NMR and the avoidance of several unphysical states observed with other force fields. Simulations of larger systems with noncanonical motifs showed significant structural improvements over the previous OPLS-AA parameters. The new force field, OPLS-AA/M, is expected to perform competitively with other recent RNA force fields and to be compatible with OPLS-AA models for proteins and small molecules.

Structural Characterization of Biomolecules through Atomistic Simulations Guided by DEER Measurements

Double electron-electron resonance (DEER) is a popular technique that exploits attached spin labels to probe the collective dynamics of biomolecules in a native environment. Like most spectroscopic approaches, DEER detects an ensemble of states accounting for biomolecular dynamics as well as the labels' intrinsic flexibility. Hence, the DEER data alone do not provide high-resolution structural information. To disentangle this variability, we introduce a minimally biased simulation method to sample a structural ensemble that reproduces multiple experimental signals within the uncertainty. In contrast to previous approaches, our method targets the raw data themselves, and thereby it brings forth an unbiased molecular interpretation of the experiments. After validation on the T4 lysozyme, we apply this technique to interpret recent DEER experiments on a membrane transporter binding protein (VcSiaP). The results highlight the large-scale conformational movement that occurs upon substrate binding and reveal that the unbound VcSiaP is more open in solution than the X-ray structure.

Inferring Structural Ensembles of Flexible and Dynamic Macromolecules Using Bayesian, Maximum Entropy, and Minimal-Ensemble Refinement Methods

The flexible and dynamic nature of biomolecules and biomolecular complexes is essential for many cellular functions in living organisms but poses a challenge for experimental methods to determine high-resolution structural models. To meet this challenge, experiments are combined with molecular simulations. The latter propose models for structural ensembles, and the experimental data can be used to steer these simulations and to select ensembles that most likely underlie the experimental data. Here, we explain in detail how the "Bayesian Inference Of ENsembles" (BioEn) method can be used to refine such ensembles using a wide range of experimental data. The "Ensemble Refinement of SAXS" (EROS) method is a special case of BioEn, inspired by the Gull-Daniell formulation of maximum entropy image processing and focused originally on X-ray solution scattering experiments (SAXS) and then extended to integrative structural modeling. We also briefly sketch the "minimum ensemble method," a maximum-parsimony refinement method that seeks to represent an ensemble with a minimal number of representative structures.

Analyzing and Biasing Simulations with PLUMED

This chapter discusses how the PLUMED plugin for molecular dynamics can be used to analyze and bias molecular dynamics trajectories. The chapter begins by introducing the notion of a collective variable and by then explaining how the free energy can be computed as a function of one or more collective variables. A number of practical issues mostly around periodic boundary conditions that arise when these types of calculations are performed using PLUMED are then discussed. Later parts of the chapter discuss how PLUMED can be used to perform enhanced sampling simulations that introduce simulation biases or multiple replicas of the system and Monte Carlo exchanges between these replicas. This section is then followed by a discussion on how free-energy surfaces and associated error bars can be extracted from such simulations by using weighted histogram and block averaging techniques.