Similarities and Differences in Ligand Binding to Protein and RNA Targets: The Case of Riboflavin.

It is nowadays clear that RNA molecules can play active roles in several biological processes. As a result, an increasing number of RNAs are gradually being identified as potentially druggable targets. In particular, noncoding RNAs can adopt highly organized conformations that are suitable for drug binding. However, RNAs are still considered challenging targets due to their complex structural dynamics and high charge density. Thus, elucidating relevant features of drug-RNA binding is fundamental for advancing drug discovery. Here, by using Molecular Dynamics simulations, we compare key features of ligand binding to proteins with those observed in RNA. Specifically, we explore similarities and differences in terms of (i) conformational flexibility of the target, (ii) electrostatic contribution to binding free energy, and (iii) water and ligand dynamics. As a test case, we examine binding of the same ligand, namely riboflavin, to protein and RNA targets, specifically the riboflavin (RF) kinase and flavin mononucleotide (FMN) riboswitch. The FMN riboswitch exhibited enhanced fluctuations and explored a wider conformational space, compared to the protein target, underscoring the importance of RNA flexibility in ligand binding. Conversely, a similar electrostatic contribution to the binding free energy of riboflavin was found. Finally, greater stability of water molecules was observed in the FMN riboswitch compared to the RF kinase, possibly due to the different shape and polarity of the pockets.

Simultaneous refinement of molecular dynamics ensembles and forward models using experimental data.

A novel method combining the maximum entropy principle, the Bayesian-inference of ensembles approach, and the optimization of empirical forward models is presented. Here, we focus on the Karplus parameters for RNA systems, which relate the dihedral angles of γ, ß, and the dihedrals in the sugar ring to the corresponding 3J-coupling signal between coupling protons. Extensive molecular simulations are performed on a set of RNA tetramers and hexamers and combined with available nucleic-magnetic-resonance data. Within the new framework, the sampled structural dynamics can be reweighted to match experimental data while the error arising from inaccuracies in the forward models can be corrected simultaneously and consequently does not leak into the reweighted ensemble. Carefully crafted cross-validation procedure and regularization terms enable obtaining transferable Karplus parameters. Our approach identifies the optimal regularization strength and new sets of Karplus parameters balancing good agreement between simulations and experiments with minimal changes to the original ensemble.

On the allosteric puzzle and pocket crosstalk through computational means.

Allostery is a constitutive, albeit often elusive, feature of biomolecular systems, which heavily determines their functioning. Its mechanical, entropic, long-range, ligand, and environment-dependent nature creates far from trivial interplays between residues and, in general, the secondary structure of proteins. This intricate scenario is mirrored in computational terms as different notions of "correlation" among residues and pockets can lead to different conclusions and outcomes. In this article, we put on a common ground and challenge three computational approaches for the correlation estimation task and apply them to three diverse targets of pharmaceutical interest: the androgen A2A receptor, the androgen receptor, and the EGFR kinase domain. Results show that partial results consensus can be attained, yet different notions lead to pointing the attention to different pockets and communications.

Reweighting of molecular simulations with explicit-solvent SAXS restraints elucidates ion-dependent RNA ensembles.

Small-angle X-ray scattering (SAXS) experiments are increasingly used to probe RNA structure. A number of forward models that relate measured SAXS intensities and structural features, and that are suitable to model either explicit-solvent effects or solute dynamics, have been proposed in the past years. Here, we introduce an approach that integrates atomistic molecular dynamics simulations and SAXS experiments to reconstruct RNA structural ensembles while simultaneously accounting for both RNA conformational dynamics and explicit-solvent effects. Our protocol exploits SAXS pure-solute forward models and enhanced sampling methods to sample an heterogenous ensemble of structures, with no information towards the experiments provided on-the-fly. The generated structural ensemble is then reweighted through the maximum entropy principle so as to match reference SAXS experimental data at multiple ionic conditions. Importantly, accurate explicit-solvent forward models are used at this reweighting stage. We apply this framework to the GTPase-associated center, a relevant RNA molecule involved in protein translation, in order to elucidate its ion-dependent conformational ensembles. We show that (a) both solvent and dynamics are crucial to reproduce experimental SAXS data and (b) the resulting dynamical ensembles contain an ion-dependent fraction of extended structures.

Kinetics of Drug Binding and Residence Time.

The kinetics of drug binding and unbinding is assuming an increasingly crucial role in the long, costly process of bringing a new medicine to patients. For example, the time a drug spends in contact with its biological target is known as residence time (the inverse of the kinetic constant of the drug-target unbinding, 1/koff). Recent reports suggest that residence time could predict drug efficacy in vivo, perhaps even more effectively than conventional thermodynamic parameters (free energy, enthalpy, entropy). There are many experimental and computational methods for predicting drug-target residence time at an early stage of drug discovery programs. Here, we review and discuss the methodological approaches to estimating drug binding kinetics and residence time. We first introduce the theoretical background of drug binding kinetics from a physicochemical standpoint. We then analyze the recent literature in the field, starting from the experimental methodologies and applications thereof and moving to theoretical and computational approaches to the kinetics of drug binding and unbinding. We acknowledge the central role of molecular dynamics and related methods, which comprise a great number of the computational methods and applications reviewed here. However, we also consider kinetic Monte Carlo. We conclude with the outlook that drug (un)binding kinetics may soon become a go/no go step in the discovery and development of new medicines.

Pressure control using stochastic cell rescaling.

Molecular dynamics simulations require barostats to be performed at a constant pressure. The usual recipe is to employ the Berendsen barostat first, which displays a first-order volume relaxation efficient in equilibration but results in incorrect volume fluctuations, followed by a second-order or a Monte Carlo barostat for production runs. In this paper, we introduce stochastic cell rescaling, a first-order barostat that samples the correct volume fluctuations by including a suitable noise term. The algorithm is shown to report volume fluctuations compatible with the isobaric ensemble and its anisotropic variant is tested on a membrane simulation. Stochastic cell rescaling can be straightforwardly implemented in the existing codes and can be used effectively in both equilibration and production phases.

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.

Predicting Residence Time and Drug Unbinding Pathway through Scaled Molecular Dynamics.

Computational approaches currently assist medicinal chemistry through the entire drug discovery pipeline. However, while several computational tools and strategies are available to predict binding affinity, predicting the drug-target binding kinetics is still a matter of ongoing research. Here, we challenge scaled molecular dynamics simulations to assess the off-rates for a series of structurally diverse inhibitors of the heat shock protein 90 (Hsp90) covering 3 orders of magnitude in their experimental residence times. The derived computational predictions are in overall good agreement with experimental data. Aside from the estimation of exit times, unbinding pathways were assessed through dimensionality reduction techniques. The data analysis framework proposed in this work could lead to better understanding of the mechanistic aspects related to the observed kinetic behavior.

Fully Flexible Docking via Reaction-Coordinate-Independent Molecular Dynamics Simulations.

Predicting the geometry of protein-ligand binding complexes is of primary importance for structure-based drug discovery. Molecular dynamics (MD) is emerging as a reliable computational tool for use in conjunction with, or an alternative to, docking methods. However, simulating the protein-ligand binding process often requires very expensive simulations. This drastically limits the practical application of MD-based approaches. Here, we propose a general framework to accelerate the generation of putative protein-ligand binding modes using potential-scaled MD simulations. The proposed dynamical protocol has been applied to two pharmaceutically relevant systems (GSK-3ß and the N-terminal domain of HSP90α). Our approach is fully independent of any predefined reaction coordinate (or collective variable). It identified the correct binding mode of several ligands and can thus save valuable computational time in dynamic docking simulations.

Binding Residence Time through Scaled Molecular Dynamics: A Prospective Application to hDAAO Inhibitors.

Traditionally, a drug potency is expressed in terms of thermodynamic quantities, mostly Kd, and empirical IC50 values. Although binding affinity as an estimate of drug activity remains relevant, it is increasingly clear that it is also important to include (un)binding kinetic parameters in the characterization of potential drug-like molecules. Herein, we used standard in silico screening to identify a series of structurally related inhibitors of hDAAO, a flavoprotein involved in schizophrenia and neuropathic pain. We applied a novel methodology, based on scaled molecular dynamics, to rank them according to their residence times. Notably, we challenged the application in a prospective fashion for the first time. The good agreement between experimental residence times and the predicted residence times highlighted the procedure's reliability in both predictive and refinement scenarios. Additionally, through further inspection of the performed simulations, we substantiated a previous hypothesis on the involvement of a protein loop during ligand unbinding.

Development of a multisite model for Ni(II) ion in solution from thermodynamic and kinetic data.

Force-field parameters are developed for a multisite model of Ni(II) ions to be used in molecular dynamics simulations combined to enhanced sampling methods. The performances of two charge-partitioning schemes are validated by taking into account structural, thermodynamic, and kinetic observables. One of the two models, featuring partial charges on the dummy atoms only, matches both Ni(II) free energy of solvation and water exchange rates. Such model is particularly suited to study complexation events at a fully dynamic description. © 2017 Wiley Periodicals, Inc.

Probing allosteric communication with combined molecular dynamics simulations and network analysis.

Understanding the allosteric mechanisms within biomolecules involved in diseases is of paramount importance for drug discovery. Indeed, characterizing communication pathways and critical hotspots in signal transduction can guide a rational approach to leverage allosteric modulation for therapeutic purposes. While the atomistic signatures of allosteric processes are difficult to determine experimentally, computational methods can be a remarkable resource. Network analysis built on Molecular Dynamics simulation data is particularly suited in this respect and is gradually becoming of routine use. Herein, we collect the recent literature in the field, discussing different aspects and available options for network construction and analysis. We further highlight interesting refinements and extensions, eventually providing our perspective on this topic.

Integrating experimental data with molecular simulations to investigate RNA structural dynamics.

Conformational dynamics is crucial for ribonucleic acid (RNA) function. Techniques such as nuclear magnetic resonance, cryo-electron microscopy, small- and wide-angle X-ray scattering, chemical probing, single-molecule Förster resonance energy transfer, or even thermal or mechanical denaturation experiments probe RNA dynamics at different time and space resolutions. Their combination with accurate atomistic molecular dynamics (MD) simulations paves the way for quantitative and detailed studies of RNA dynamics. First, experiments provide a quantitative validation tool for MD simulations. Second, available data can be used to refine simulated structural ensembles to match experiments. Finally, comparison with experiments allows for improving MD force fields that are transferable to new systems for which data is not available. Here we review the recent literature and provide our perspective on this field.

Molecular Dynamics Simulations with Grand-Canonical Reweighting Suggest Cooperativity Effects in RNA Structure Probing Experiments.

Chemical probing experiments such as SHAPE are routinely used to probe RNA molecules. In this work, we use atomistic molecular dynamics simulations to test the hypothesis that binding of RNA with SHAPE reagents is affected by cooperative effects leading to an observed reactivity that is dependent on the reagent concentration. We develop a general technique that enables the calculation of the affinity for arbitrary molecules as a function of their concentration in the grand-canonical ensemble. Our simulations of an RNA structural motif suggest that, at the concentration typically used in SHAPE experiments, cooperative binding would lead to a measurable concentration-dependent reactivity. We also provide a qualitative validation of this statement by analyzing a new set of experiments collected at different reagent concentrations.

Molecular Simulations Matching Denaturation Experiments for N6-Methyladenosine.

Post-transcriptional modifications are crucial for RNA function and can affect its structure and dynamics. Force-field-based classical molecular dynamics simulations are a fundamental tool to characterize biomolecular dynamics, and their application to RNA is flourishing. Here, we show that the set of force-field parameters for N6-methyladenosine (m6A) developed for the commonly used AMBER force field does not reproduce duplex denaturation experiments and, specifically, cannot be used to describe both paired and unpaired states. Then, we use reweighting techniques to derive new parameters matching available experimental data. The resulting force field can be used to properly describe paired and unpaired m6A in both syn and anti conformation, which thus opens the way to the use of molecular simulations to investigate the effects of N6 methylations on RNA structural dynamics.

Computational drug discovery under RNA times.

RNA molecules play many functional and regulatory roles in cells, and hence, have gained considerable traction in recent times as therapeutic interventions. Within drug discovery, structure-based approaches have successfully identified potent and selective small-molecule modulators of pharmaceutically relevant protein targets. Here, we embrace the perspective of computational chemists who use these traditional approaches, and we discuss the challenges of extending these methods to target RNA molecules. In particular, we focus on recognition between RNA and small-molecule binders, on selectivity, and on the expected properties of RNA ligands.

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.

Enhanced Molecular Dynamics Simulations of Intrinsically Disordered Proteins.

Molecular dynamics simulations represent a powerful tool to gain insights into structural and dynamical features of biomolecular systems. Nevertheless, their recognized limitation in terms of achievable timescales becomes particularly severe when dealing with slow processes. In such cases, the employment of enhanced sampling methods, which allow accelerating the characterization of rare events in a timeframe consistent with conventional computational resources, results as crucial. In particular, such advanced techniques have proven highly valuable in the context of protein folding and, specifically, to explore the conformational ensemble spanned by intrinsically disordered proteins (IDPs). Here, we describe how to set up molecular dynamics simulations with one of these enhanced sampling approaches (namely, Parallel Tempering Metadynamics in the Well-Tempered Ensemble) using the NTAIL peptide as a test case.

An Integrated Markov State Model and Path Metadynamics Approach To Characterize Drug Binding Processes.

Unveiling the mechanistic features of drug-target binding is of central interest in biophysics and drug discovery. Herein, we address this challenge by combining two major computational approaches, namely, Molecular Dynamics (MD) simulations and Markov State Models (MSM), with a Path Collective Variables (PCVs) description coupled with metadynamics. We apply our methodology to reconstruct the binding process of the antagonist alprenolol to the ß2-adrenergic receptor, a well-established pharmaceutical target. The devised protocol allowed us to estimate the binding free energy and identify the minimum free energy path leading to the protein-ligand complex. In summary, we show that MSM and PCVs can be efficiently integrated to shed light upon mechanistic and energetic details underlying complex recognition processes in biological systems.

Structural and Kinetic Characterization of the Intrinsically Disordered Protein SeV NTAIL through Enhanced Sampling Simulations.

Intrinsically disordered proteins (IDPs) are emerging as an important class of the proteome. Being able to interact with different molecular targets, they participate in many physiological and pathological activities. However, due to their intrinsically heterogeneous nature, determining the equilibrium properties of IDPs is still a challenge for biophysics. Herein, we applied state-of-the-art enhanced sampling methods to Sev NTAIL, a test case of IDPs, and constructed a bin-based kinetic model to unveil the underlying kinetics. To validate our simulation strategy, we compared the predicted NMR properties against available experimental data. Our simulations reveal a rough free-energy surface comprising multiple local minima, which are separated by low energy barriers. Moreover, we identified interconversion rates between the main kinetic states, which lie in the sub-µs time scales, as suggested in previous works for Sev NTAIL. Therefore, the emerging picture is in agreement with the atomic-level properties possessed by the free IDP in solution. By providing both a thermodynamic and kinetic characterization of this IDP test case, our study demonstrates how computational methods can be effective tools for studying this challenging class of proteins.