Perspectives on Ligand/Protein Binding Kinetics Simulations: Force Fields, Machine Learning, Sampling, and User-Friendliness

Investigating the unbinding mechanisms and kinetics of MmpL3 inhibitors: A computational study

Mycobacterial membrane protein Large 3 (MmpL3) is responsible for transporting trehalose monomycolates across the inner membrane for cell wall biosynthesis, a process driven by the proton motive force and essential for the survival of Mycobacterium tuberculosis. As a result, MmpL3 has become a promising target for anti-tuberculosis drugs. Although many inhibitors targeting MmpL3 have been discovered, their unbinding mechanisms and kinetics remain poorly understood. In this study, the τ-random acceleration molecular dynamics (τRAMD) and steered molecular dynamics (SMD) methods were employed to investigate the unbinding mechanisms and kinetics of four representative MmpL3 inhibitors: SQ109, AU1235, NITD349, and BM212. Analysis of 320 RAMD dissociation trajectories revealed considerable diversity in the dissociation pathways for these inhibitors, dissociating into intracellular, extracellular, or transmembrane regions. Notably, the H4H5H10 pathway, dissociating to the intracellular region, was the primary route. Also, τRAMD results demonstrated a strong correlation between the computed relative residence times and experimental data. Furthermore, SMD simulations along the H4H5H10 pathway indicated that SQ109, AU1235, and NITD349 disrupted hydrogen bonding with MmpL3 prior to dissociation. Meanwhile, inhibitor BM212 underwent conformational adjustments within the binding pocket. All these inhibitors must traverse the channel formed by Phe255 and Phe644 via the H4H5H10 pathway, necessitating the overcoming of significant energy barriers. Based on these findings, we suggest that enhancing inhibitor interactions with MmpL3, such as through hydrogen bonding or increasing inhibitor size to create larger physical barriers (e.g., interactions with Phe255 and Phe644), may prolong the inhibitors' residence times.

Exploring Membrane Cholesterol Binding to the CB1 Receptor: A Computational Perspective

Cholesterol (CHOL) is a potential allosteric modulator of the CB1 receptor. In this work, we use atomistic molecular dynamics simulations to study how CHOL interacts with CB1 and to identify its binding sites (BS) and residence times on specific receptor zones. Our results evince minimal changes in CB1 conformational dynamics and secondary structure due to CHOL. We report five BSs, three of which coincide with previously described interaction regions (BS1, BS2, and BS3), while BS4 and BS5 are proposed as new BSs. Quantum descriptors of bonding such as Natural Bond Orbitals (NBO), Quantum Theory of Atoms in Molecules (QTAIM), and Noncovalent Interactions (NCI) analyses are employed to characterize the CHOL-BS interactions. The results show an exponential correlation between the strength of the interactions (mainly hydrogen bonds and hydrophobic contacts) and the residence time at the BSs. Although other approaches exist to identify high-affinity protein sites, our methodology integrates classical and quantum descriptions to better characterize BSs and predict ligand residence times in CB1, distinguishing persistent from transitory contacts. Since CHOL has been suggested as a potential endogenous allosteric ligand, our flexible strategy allows studying interactions that stabilize CHOL in CB1, could be extended to cannabinoid binding, and contribute to designing improved receptor ligands.

Trust the force, but pull wisely: Structural insights into non-equilibrium response forces from pulling MD simulations

We analyze the coarse-grained equations of motion of molecular systems subject to external driving. As exemplary processes, we study by means of targeted and steered molecular dynamics simulations the dissociation of a sodium-chloride ion pair in water and ligand-protein unbinding of trypsin-benzamidine. We derive an exact generalization of Mori's Langevin equation that contains the memory kernel of the stationary process, an additive driving force, and a non-equilibrium response force describing the effects of the perturbed environment. We show that both the fluctuating force in the stationary case and the non-equilibrium response force in the driven cases exhibit spatial structure in their first and second moments. The latter depends sensitively on the employed driving protocols. For sodium chloride, we find that the first moment of the non-equilibrium response force matches the mean force for slow constrained pulling. In contrast, for all tested restrained pulling protocols, significant differences arise between the two properties in both systems. We conclude that the non-equilibrium response of the solvent needs to be taken into account carefully when analyzing data from pulling simulations.

Functional dynamics of G protein-coupled receptors reveal new routes for drug discovery

G protein-coupled receptors (GPCRs) are the largest human membrane protein family that transduce extracellular signals into cellular responses. They are major pharmacological targets, with approximately 26% of marketed drugs targeting GPCRs, primarily at their orthosteric binding site. Despite their prominence, predicting the pharmacological effects of novel GPCR-targeting drugs remains challenging due to the complex functional dynamics of these receptors. Recent advances in X-ray crystallography, cryo-electron microscopy, spectroscopic techniques and molecular simulations have enhanced our understanding of receptor conformational dynamics and ligand interactions with GPCRs. These developments have revealed novel ligand-binding modes, mechanisms of action and druggable pockets. In this Review, we highlight such aspects for recently discovered small-molecule drugs and drug candidates targeting GPCRs, focusing on three categories: allosteric modulators, biased ligands, and bivalent and bitopic compounds. Although studies so far have largely been retrospective, integrating structural data on ligand-induced receptor functional dynamics into the drug discovery pipeline has the potential to guide the identification of drug candidates with specific abilities to modulate GPCR interactions with intracellular effector proteins such as G proteins and β-arrestins, enabling more tailored selectivity and efficacy profiles.

Detection of Putative Ligand Dissociation Pathways in Proteins Using Site-Identification by Ligand Competitive Saturation

Drug efficacy often correlates better with dissociation kinetics than binding affinity alone. To study binding kinetics computationally, it is necessary to identify all of the possible ligand dissociation pathways. The site identification by ligand competitive saturation (SILCS) method involves the precomputation of a set of maps (FragMaps), which describe the free energy landscapes of typical chemical functionalities in and around a target protein or RNA. In the current work, we present and implement a method to use SILCS to identify ligand dissociation pathways, termed "SILCS-Pathway." The A* pathfinding algorithm is utilized to enumerate ligand dissociation pathways between the ligand binding site and the surrounding bulk solvent environment defined on evenly spaced points around the protein based on a Fibonacci lattice. The cost function for the A* algorithm is calculated using the SILCS exclusion maps and the SILCS grid free energy scores, thereby identifying paths that account for local protein flexibility and potential favorable interactions with the ligand. By traversing all evenly distributed bulk solvent points around the protein, we located all possible dissociation pathways and clustered them to identify general ligand unbinding pathways. The procedure is verified by using proteins studied previously with enhanced sampling molecular dynamics (MD) techniques and is shown to be capable of capturing important ligand dissociation routes in a highly computationally efficient manner. The identified pathways will act as the foundation for determining ligand dissociation kinetics using SILCS free energy profiles, which will be described in a subsequent article.

Pathway Specific Unbinding Free Energy Profiles of Ritonavir Dissociation from HIV-1 Protease

Investigation of protein-drug recognition is key to understanding drug selectivity and binding affinity. In combination, the binding/unbinding free energy landscape and intermolecular interactions can be used to understand drug binding/unbinding mechanisms. This information is vital for the development of drugs with improved efficacy and explanation of mutation effects. This study investigated the dissociation processes of ritonavir unbinding from HIV protease (HIVp). Analyzing unbinding trajectories modeled by accelerated molecular dynamics (MD) simulations, three distinct pathways, pathways A-C, were characterized. Using a reduced dimensionality strategy with the principal component analysis, we carried out short classical MD runs with explicit water to sample local fluctuation during ritonavir dissociation and applied the milestoning theory to construct an unbinding free energy landscape. We found that each pathway showed similar values of binding free energy, albeit pathway A accounts for over 50% of dissociation trajectories. Interestingly, residue-residue correlation network analysis showed that in pathway A, a broad correlation network outside the flap region governs protein motions during ritonavir unbinding, which includes residues with reported mutation effects. However, the other two pathways showed limited correlation networks where no reported mutated residues were involved, explaining the favorability of pathway A. Guided by the free energy profile, we investigated each energy barrier and minimum, demonstrating that hydrogen bonding governed movement of the flap regions, directly impacting the calculated energy. Our study provided a new strategy to estimate ligand binding free energy and demonstrated the importance of the transient interactions during ligand-protein dissociation pathways in understanding drug unbinding.

Host-Guest Binding Free Energies à la Carte: An Automated OneOPES Protocol

Estimating absolute binding free energies from molecular simulations is a key step in computer-aided drug design pipelines, but the agreement between computational results and experiments is still very inconsistent. Both the accuracy of the computational model and the quality of the statistical sampling contribute to this discrepancy, yet disentangling the two remains a challenge. In this study, we present an automated protocol based on OneOPES, an enhanced sampling method that exploits replica exchange and can accelerate several collective variables to address the sampling problem. We apply this protocol to 37 host-guest systems. The simplicity of setting up the simulations and producing well-converged binding free energy estimates without the need to optimize simulation parameters provides a reliable solution to the sampling problem. This, in turn, allows for a systematic force field comparison and ranking according to the correlation between simulations and experiments, which can inform the selection of an appropriate model. The protocol can be readily adapted to test more force field combinations and study more complex protein-ligand systems, where the choice of an appropriate physical model is often based on heuristic considerations rather than systematic optimization.

Current Status of Computational Approaches for Small Molecule Drug Discovery

2024 has been an exciting year for computational sciences, with the Nobel Prize in Physics awarded for "artificial neural network" and the Nobel Prize in Chemistry presented for "protein structure prediction and design". Given the rapid advancements in Computer-Aided Drug Design (CADD) and Artificial Intelligence in Drug Discovery (AIDD), a document summarizing their current standing and future directions would be timely and relevant to the readership of Journal of Medicinal Chemistry. This piece of commentary aims to highlight recent developments, key challenges, and potential synergies between these fields, contributing to ongoing discussions in the literature and scientific blogs.

Computational screening of the effects of mutations on protein-protein off-rates and dissociation mechanisms by τRAMD

The dissociation rate, or its reciprocal, the residence time (τ), is a crucial parameter for understanding the duration and biological impact of biomolecular interactions. Accurate prediction of τ is essential for understanding protein-protein interactions (PPIs) and identifying potential drug targets or modulators for tackling diseases. Conventional molecular dynamics simulation techniques are inherently constrained by their limited timescales, making it challenging to estimate residence times, which typically range from minutes to hours. Building upon its successful application in protein-small molecule systems, τ-Random Acceleration Molecular Dynamics (τRAMD) is here investigated for estimating dissociation rates of protein-protein complexes. τRAMD enables the observation of unbinding events on the nanosecond timescale, facilitating rapid and efficient computation of relative residence times. We tested this methodology for three protein-protein complexes and their extensive mutant datasets, achieving good agreement between computed and experimental data. By combining τRAMD with MD-IFP (Interaction Fingerprint) analysis, dissociation mechanisms were characterized and their sensitivity to mutations investigated, enabling the identification of molecular hotspots for selective modulation of dissociation kinetics. In conclusion, our findings underscore the versatility of τRAMD as a simple and computationally efficient approach for computing relative protein-protein dissociation rates and investigating dissociation mechanisms, thereby aiding the design of PPI modulators.

Increased throughput in methods for simulating protein ligand binding and unbinding

By incorporating full flexibility and enabling the quantification of crucial parameters such as binding free energies and residence times, methods for investigating protein-ligand binding and unbinding via molecular dynamics provide details on the involved mechanisms at the molecular level. While these advancements hold promise for impacting drug discovery, a notable drawback persists: their relatively time-consuming nature limits throughput. Herein, we survey recent implementations which, employing a blend of enhanced sampling techniques, a clever choice of collective variables, and often machine learning, strive to enhance the efficiency of new and previously reported methods without compromising accuracy. Particularly noteworthy is the validation of these methods that was often performed on systems mirroring real-world drug discovery scenarios.

The Open Force Field Initiative: Open Software and Open Science for Molecular Modeling

Force fields are a key component of physics-based molecular modeling, describing the energies and forces in a molecular system as a function of the positions of the atoms and molecules involved. Here, we provide a review and scientific status report on the work of the Open Force Field (OpenFF) Initiative, which focuses on the science, infrastructure and data required to build the next generation of biomolecular force fields. We introduce the OpenFF Initiative and the related OpenFF Consortium, describe its approach to force field development and software, and discuss accomplishments to date as well as future plans. OpenFF releases both software and data under open and permissive licensing agreements to enable rapid application, validation, extension, and modification of its force fields and software tools. We discuss lessons learned to date in this new approach to force field development. We also highlight ways that other force field researchers can get involved, as well as some recent successes of outside researchers taking advantage of OpenFF tools and data.

Mechanism of Ligand Binding to Theophylline RNA Aptamer

Studying RNA-ligand interactions and quantifying their binding thermodynamics and kinetics are of particular relevance in the field of drug discovery. Here, we combined biochemical binding assays and accelerated molecular simulations to investigate ligand binding and dissociation in RNA using the theophylline-binding RNA as a model system. All-atom simulations using a Ligand Gaussian accelerated Molecular Dynamics method (LiGaMD) have captured repetitive binding and dissociation of theophylline and caffeine to RNA. Theophylline's binding free energy and kinetic rate constants align with our experimental data, while caffeine's binding affinity is over 10,000 times weaker, and its kinetics could not be determined. LiGaMD simulations allowed us to identify distinct low-energy conformations and multiple ligand binding pathways to RNA. Simulations revealed a "conformational selection" mechanism for ligand binding to the flexible RNA aptamer, which provides important mechanistic insights into ligand binding to the theophylline-binding model. Our findings suggest that compound docking using a structural ensemble of representative RNA conformations would be necessary for structure-based drug design of flexible RNA.