Absolute Binding Free Energies with OneOPES.

The calculation of absolute binding free energies (ABFEs) for protein-ligand systems has long been a challenge. Recently, refined force fields and algorithms have improved the quality of the ABFE calculations. However, achieving the level of accuracy required to inform drug discovery efforts remains difficult. Here, we present a transferable enhanced sampling strategy to accurately calculate absolute binding free energies using OneOPES with simple geometric collective variables. We tested the strategy on two protein targets, BRD4 and Hsp90, complexed with a total of 17 chemically diverse ligands, including both molecular fragments and drug-like molecules. Our results show that OneOPES accurately predicts protein-ligand binding affinities with a mean unsigned error within 1 kcal mol-1 of experimentally determined free energies, without the need to tailor the collective variables to each system. Furthermore, our strategy effectively samples different ligand binding modes and consistently matches the experimentally determined structures regardless of the initial protein-ligand configuration. Our results suggest that the proposed OneOPES strategy can be used to inform lead optimization campaigns in drug discovery and to study protein-ligand binding and unbinding mechanisms.

DeepLNE++ leveraging knowledge distillation for accelerated multi-state path-like collective variables.

Path-like collective variables (CVs) can be very effective for accurately modeling complex biomolecular processes in molecular dynamics simulations. Recently, we have introduced DeepLNE (deep-locally non-linear-embedding), a machine learning-based path-like CV that provides a progression variable s along the path as a non-linear combination of several descriptors. We have demonstrated the effectiveness of DeepLNE by showing that for simple models such as the Müller-Brown potential and alanine dipeptide, the progression along the path variable closely approximates the ideal reaction coordinate. However, DeepLNE is computationally expensive for realistic systems needing many descriptors and limited in its ability to handle multi-state reactions. Here, we present DeepLNE++, which uses a knowledge distillation approach to significantly accelerate the evaluation of DeepLNE, making it feasible to compute free energy landscapes for large and complex biomolecular systems. In addition, DeepLNE++ encodes system-specific knowledge within a supervised multitasking framework, enhancing its versatility and effectiveness.

Conformational plasticity and allosteric communication networks explain Shelterin protein TPP1 binding to human telomerase.

The Shelterin complex protein TPP1 interacts with human telomerase (TERT) by means of the TEL-patch region, controlling telomere homeostasis. Aberrations in the TPP1-TERT heterodimer formation might lead to short telomeres and severe diseases like dyskeratosis congenita and Hoyeraal-Hreidarsson syndrome. In the present study, we provide a thorough characterization of the structural properties of the TPP1's OB-domain by combining data coming from microsecond-long molecular dynamics calculations, time-series analyses, and graph-based networks. Our results show that the TEL-patch conformational freedom is influenced by a network of long-range amino acid communications that together determine the proper TPP1-TERT binding. Furthermore, we reveal that in TPP1 pathological variants Glu169Δ, Lys170Δ and Leu95Gln, the TEL-patch plasticity is reduced, affecting the correct binding to TERT and, in turn, telomere processivity, which eventually leads to accelerated aging of affected cells. Our study provides a structural basis for the design of TPP1-targeting ligands with therapeutic potential against cancer and telomeropathies.

OneOPES, a Combined Enhanced Sampling Method to Rule Them All.

Enhanced sampling techniques have revolutionized molecular dynamics (MD) simulations, enabling the study of rare events and the calculation of free energy differences in complex systems. One of the main families of enhanced sampling techniques uses physical degrees of freedom called collective variables (CVs) to accelerate a system's dynamics and recover the original system's statistics. However, encoding all the relevant degrees of freedom in a limited number of CVs is challenging, particularly in large biophysical systems. Another category of techniques, such as parallel tempering, simulates multiple replicas of the system in parallel, without requiring CVs. However, these methods may explore less relevant high-energy portions of the phase space and become computationally expensive for large systems. To overcome the limitations of both approaches, we propose a replica exchange method called OneOPES that combines the power of multireplica simulations and CV-based enhanced sampling. This method efficiently accelerates the phase space sampling without the need for ideal CVs, extensive parameters fine tuning nor the use of a large number of replicas, as demonstrated by its successful applications to protein-ligand binding and protein folding benchmark systems. Our approach shows promise as a new direction in the development of enhanced sampling techniques for molecular dynamics simulations, providing an efficient and robust framework for the study of complex and unexplored problems.

Subcellular location defines GPCR signal transduction.

Intracellular G protein-coupled receptors (GPCRs) can be activated by permeant ligands, which contributes to agonist selectivity. Opioid receptors (ORs) provide a notable example, where opioid drugs rapidly activate ORs in the Golgi apparatus. Our knowledge on intracellular GPCR function remains incomplete, and it is unknown whether OR signaling in plasma membrane (PM) and Golgi apparatus differs. Here, we assess the recruitment of signal transducers to mu- and delta-ORs in both compartments. We find that Golgi ORs couple to Gαi/o probes and are phosphorylated but, unlike PM receptors, do not recruit ß-arrestin or a specific Gα probe. Molecular dynamics simulations with OR-transducer complexes in bilayers mimicking PM or Golgi composition reveal that the lipid environment promotes the location-selective coupling. We then show that delta-ORs in PM and Golgi have distinct effects on transcription and protein phosphorylation. The study reveals that the subcellular location defines the signaling effects of opioid drugs.

Open Binding Pose Metadynamics: An Effective Approach for the Ranking of Protein-Ligand Binding Poses.

Predicting the correct pose of a ligand binding to a protein and its associated binding affinity is of great importance in computer-aided drug discovery. A number of approaches have been developed to these ends, ranging from the widely used fast molecular docking to the computationally expensive enhanced sampling molecular simulations. In this context, methods such as coarse-grained metadynamics and binding pose metadynamics (BPMD) use simulations with metadynamics biasing to probe the binding affinity without trying to fully converge the binding free energy landscape in order to decrease the computational cost. In BPMD, the metadynamics bias perturbs the ligand away from the initial pose. The resistance of the ligand to this bias is used to calculate a stability score. The method has been shown to be useful in reranking predicted binding poses from docking. Here, we present OpenBPMD, an open-source Python reimplementation and reinterpretation of BPMD. OpenBPMD is powered by the OpenMM simulation engine and uses a revised scoring function. The algorithm was validated by testing it on a wide range of targets and showing that it matches or exceeds the performance of the original BPMD. We also investigated the role of accurate water positioning on the performance of the algorithm and showed how the combination with a grand-canonical Monte Carlo algorithm improves the accuracy of the predictions.

Phosphorylation of FAM134C by CK2 controls starvation-induced ER-phagy.

Selective degradation of the endoplasmic reticulum (ER) via autophagy (ER-phagy) is initiated by ER-phagy receptors, which facilitate the incorporation of ER fragments into autophagosomes. FAM134 reticulon family proteins (FAM134A, FAM134B, and FAM134C) are ER-phagy receptors with structural similarities and nonredundant functions. Whether they respond differentially to the stimulation of ER-phagy is unknown. Here, we describe an activation mechanism unique to FAM134C during starvation. In fed conditions, FAM134C is phosphorylated by casein kinase 2 (CK2) at critical residues flanking the LIR domain. Phosphorylation of these residues negatively affects binding affinity to the autophagy proteins LC3. During starvation, mTORC1 inhibition limits FAM134C phosphorylation by CK2, hence promoting receptor activation and ER-phagy. Using a novel tool to study ER-phagy in vivo and FAM134C knockout mice, we demonstrated the physiological relevance of FAM134C phosphorylation during starvation-induced ER-phagy in liver lipid metabolism. These data provide a mechanistic insight into ER-phagy regulation and an example of autophagy selectivity during starvation.

Drug Repurposing on G Protein-Coupled Receptors Using a Computational Profiling Approach.

G protein-coupled receptors (GPCRs) are the largest human membrane receptor family regulating a wide range of cell signaling. For this reason, GPCRs are highly desirable drug targets, with approximately 40% of prescribed medicines targeting a member of this receptor family. The structural homology of GPCRs and the broad spectrum of applications of GPCR-acting drugs suggest an investigation of the cross-activity of a drug toward different GPCR receptors with the aim of rationalizing drug side effects, designing more selective and less toxic compounds, and possibly proposing off-label therapeutic applications. Herein, we present an original in silico approach named "Computational Profiling for GPCRs" (CPG), which is able to represent, in a one-dimensional (1D) string, the physico-chemical properties of a ligand-GPCR binding interaction and, through a tailored alignment algorithm, repurpose the ligand for a different GPCR. We show three case studies where docking calculations and pharmacological data confirm the drug repurposing findings obtained through CPG on 5-hydroxytryptamine receptor 2B, beta-2 adrenergic receptor, and M2 muscarinic acetylcholine receptor. The CPG code is released as a user-friendly graphical user interface with numerous options that make CPG a powerful tool to assist the drug design of GPCR ligands.

DDT - Drug Discovery Tool: a fast and intuitive graphics user interface for docking and molecular dynamics analysis.

MOTIVATION: The ligand/protein binding interaction is typically investigated by docking and molecular dynamics (MD) simulations. In particular, docking-based virtual screening (VS) is used to select the best ligands from database of thousands of compounds, while MD calculations assess the energy stability of the ligand/protein binding complexes. Considering the broad use of these techniques, it is of great demand to have one single software that allows a combined and fast analysis of VS and MD results. With this in mind, we have developed the Drug Discovery Tool (DDT) that is an intuitive graphics user interface able to provide structural data and physico-chemical information on the ligand/protein interaction. RESULTS: DDT is designed as a plugin for the Visual Molecular Dynamics (VMD) software and is able to manage a large number of ligand/protein complexes obtained from AutoDock4 (AD4) docking calculations and MD simulations. DDT delivers four main outcomes: i) ligands ranking based on an energy score; ii) ligand ranking based on a ligands' conformation cluster analysis; iii) identification of the aminoacids forming the most occurrent interactions with the ligands; iv) plot of the ligands' center-of-mass coordinates in the Cartesian space. The flexibility of the software allows saving the best ligand/protein complexes using a number of user-defined options. AVAILABILITY AND IMPLEMENTATION: DDT_site_1 (alternative DDT_site_2); the DDT tutorial movie is available here. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

How many antimicrobial peptide molecules kill a bacterium? The case of PMAP-23.

Antimicrobial peptides (AMPs) kill bacteria mainly through the perturbation of their membranes and are promising compounds to fight drug resistance. Models of the mechanism of AMPs-induced membrane perturbation were developed based on experiments in liposomes, but their relevance for bacterial killing is debated. We determined the association of an analogue of the AMP PMAP-23 to Escherichia coli cells, under the same experimental conditions used to measure bactericidal activity. Killing took place only when bound peptides completely saturated bacterial membranes (10(6)-10(7) bound peptides per cell), indicating that the "carpet" model for the perturbation of artificial bilayers is representative of what happens in real bacteria. This finding supports the view that, at least for this peptide, a microbicidal mechanism is possible in vivo only at micromolar total peptide concentrations. We also showed that, notwithstanding their simplicity, liposomes represent a reliable model to characterize AMPs partition in bacterial membranes.