Deep learning path-like collective variable for enhanced sampling molecular dynamics.

Several enhanced sampling techniques rely on the definition of collective variables to effectively explore free energy landscapes. The existing variables that describe the progression along a reactive pathway offer an elegant solution but face a number of limitations. In this paper, we address these challenges by introducing a new path-like collective variable called the "deep-locally non-linear-embedding," which is inspired by principles of the locally linear embedding technique and is trained on a reactive trajectory. The variable mimics the ideal reaction coordinate by automatically generating a non-linear combination of features through a differentiable generalized autoencoder that combines a neural network with a continuous k-nearest neighbor selection. Among the key advantages of this method is its capability to automatically choose the metric for searching neighbors and to learn the path from state A to state B without the need to handpick landmarks a priori. We demonstrate the effectiveness of DeepLNE by showing that the progression along the path variable closely approximates the ideal reaction coordinate in toy models, such as the Müller-Brown potential and alanine dipeptide. Then, we use it in the molecular dynamics simulations of an RNA tetraloop, where we highlight its capability to accelerate transitions and estimate the free energy of folding.

SWISH-X, an Expanded Approach to Detect Cryptic Pockets in Proteins and at Protein-Protein Interfaces.

Protein-protein interactions mediate most molecular processes in the cell, offering a significant opportunity to expand the set of known druggable targets. Unfortunately, targeting these interactions can be challenging due to their typically flat and featureless interaction surfaces, which often change as the complex forms. Such surface changes may reveal hidden (cryptic) druggable pockets. Here, we analyze a set of well-characterized protein-protein interactions harboring cryptic pockets and investigate the predictive power of current computational methods. Based on our observations, we developed a new computational strategy, SWISH-X (SWISH Expanded), which combines the established cryptic pocket identification capabilities of SWISH with the rapid temperature range exploration of OPES MultiThermal. SWISH-X is able to reliably identify cryptic pockets at protein-protein interfaces while retaining its predictive power for revealing cryptic pockets in isolated proteins, such as TEM-1 ß-lactamase.

Determinants of Neutral Antagonism and Inverse Agonism in the ß2-Adrenergic Receptor.

Free-energy profiles for the activation/deactivation of the ß2-adrenergic receptor (ADRB2) with neutral antagonist and inverse agonist ligands have been determined with well-tempered multiple-walker (MW) metadynamics simulations. The inverse agonists carazolol and ICI118551 clearly favor single inactive conformational minima in both the binary and ternary ligand-receptor-G-protein complexes, in accord with the inverse-agonist activity of the ligands. The behavior of neutral antagonists is more complex, as they seem also to affect the recruitment of the G-protein. The results are analyzed in terms of the conformational states of the well-known microswitches that have been proposed as indicators of receptor activity.

Addressing Suboptimal Poses in Nonequilibrium Alchemical Calculations.

Alchemical transformations can be used to quantitatively estimate absolute binding free energies at a reasonable computational cost. However, most of the approaches currently in use require knowledge of the correct (crystallographic) pose. In this paper, we present a combined Hamiltonian replica exchange nonequilibrium alchemical method that allows us to reliably calculate absolute binding free energies, even when starting from suboptimal initial binding poses. Performing a preliminary Hamiltonian replica exchange enhances the sampling of slow degrees of freedom of the ligand and the target, allowing the system to populate the correct binding pose when starting from an approximate docking pose. We apply the method on 6 ligands of the first bromodomain of the BRD4 bromodomain-containing protein. For each ligand, we start nonequilibrium alchemical transformations from both the crystallographic pose and the top-scoring docked pose that are often significantly different. We show that the method produces statistically equivalent binding free energies, making it a useful tool for computational drug discovery pipelines.

Activation/Deactivation Free-Energy Profiles for the ß2-Adrenergic Receptor: Ligand Modes of Action.

We use enhanced-sampling simulations with an effective collective variable to study the activation of the ß2-adrenergic receptor in the presence of ligands with different efficacy. The free-energy profiles are computed for the ligand-free (apo) receptor and binary (apo-receptor + G-protein α-subunit and receptor + ligand) and ternary complexes. The results are not only compatible with available experiments but also allow unprecedented structural insight into the nature of GPCR conformations along the activation pathway and their role in the activation mechanism. In particular, the simulations reveal an unexpected mode of action of partial agonists such as salmeterol and salbutamol that arises already in the binary complex without the G-protein. Specific differences in the polar interactions with residues in TM5, which are required to stabilize an optimal TM6 conformation that facilitates G-protein binding and receptor activation, play a major role in differentiating them from full agonists.

Architecture of the MKK6-p38α complex defines the basis of MAPK specificity and activation.

The mitogen-activated protein kinase (MAPK) p38α is a central component of signaling in inflammation and the immune response and is, therefore, an important drug target. Little is known about the molecular mechanism of its activation by double phosphorylation from MAPK kinases (MAP2Ks), because of the challenge of trapping a transient and dynamic heterokinase complex. We applied a multidisciplinary approach to generate a structural model of p38α in complex with its MAP2K, MKK6, and to understand the activation mechanism. Integrating cryo-electron microscopy with molecular dynamics simulations, hydrogen-deuterium exchange mass spectrometry, and experiments in cells, we demonstrate a dynamic, multistep phosphorylation mechanism, identify catalytically relevant interactions, and show that MAP2K-disordered amino termini determine pathway specificity. Our work captures a fundamental step of cell signaling: a kinase phosphorylating its downstream target kinase.

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.

Structure of the connexin-43 gap junction channel in a putative closed state.

Gap junction channels (GJCs) mediate intercellular communication by connecting two neighbouring cells and enabling direct exchange of ions and small molecules. Cell coupling via connexin-43 (Cx43) GJCs is important in a wide range of cellular processes in health and disease (Churko and Laird, 2013; Liang et al., 2020; Poelzing and Rosenbaum, 2004), yet the structural basis of Cx43 function and regulation has not been determined until now. Here, we describe the structure of a human Cx43 GJC solved by cryo-EM and single particle analysis at 2.26 Å resolution. The pore region of Cx43 GJC features several lipid-like densities per Cx43 monomer, located close to a putative lateral access site at the monomer boundary. We found a previously undescribed conformation on the cytosolic side of the pore, formed by the N-terminal domain and the transmembrane helix 2 of Cx43 and stabilized by a small molecule. Structures of the Cx43 GJC and hemichannels (HCs) in nanodiscs reveal a similar gate arrangement. The features of the Cx43 GJC and HC cryo-EM maps and the channel properties revealed by molecular dynamics simulations suggest that the captured states of Cx43 are consistent with a closed state.

The action of physiological and synthetic steroids on the calcium channel CatSper in human sperm.

The sperm-specific channel CatSper (cation channel of sperm) controls the intracellular Ca2+ concentration ([Ca2+]i) and plays an essential role in sperm function. It is mainly activated by the steroid progesterone (P4) but is also promiscuously activated by a wide range of synthetic and physiological compounds. These compounds include diverse steroids whose action on the channel is so far still controversial. To investigate the effect of these compounds on CatSper and sperm function, we developed a high-throughput screening (HTS) assay to measure changes in [Ca2+]i in human sperm and screened 1,280 approved and off-patent drugs including 90 steroids from the Prestwick chemical library. More than half of the steroids tested (53%) induced an increase in [Ca2+]i and reduced the P4-induced Ca2+ influx in human sperm in a dose-dependent manner. Ten of the most potent steroids (activating and P4-inhibiting) were selected for a detailed analysis of their action on CatSper and their ability to act on sperm acrosome reaction (AR) and penetration in viscous media. We found that these steroids show an inhibitory effect on P4 but not on prostaglandin E1-induced CatSper activation, suggesting that they compete for the same binding site as P4. Pregnenolone, dydrogesterone, epiandrosterone, nandrolone, and dehydroepiandrosterone acetate (DHEA) were found to activate CatSper at physiologically relevant concentrations within the nanomolar range. Like P4, most tested steroids did not significantly affect the AR while stanozolol and estropipate slightly increased sperm penetration into viscous medium. Furthermore, using a hybrid approach integrating pharmacophore analysis and statistical modelling, we were able to screen in silico for steroids that can activate the channel and define the physicochemical and structural properties required for a steroid to exhibit agonist activity against CatSper. Overall, our results indicate that not only physiological but also synthetic steroids can modulate the activity of CatSper with varying potency and if bound to CatSper prior to P4, could impair the timely CatSper activation necessary for proper fertilization to occur.

General Metadynamics Protocol To Simulate Activation/Deactivation of Class A GPCRs: Proof of Principle for the Serotonin Receptor.

We present a generally applicable metadynamics protocol for characterizing the activation free-energy profiles of class A G-protein coupled receptors and a proof-of-principle study for the 5HT1A-receptor. The almost universal A100 activation index, which depends on five inter-helix distances, is used as the single collective variable in well-tempered multiple-walker metadynamics simulations. Here, we show free-energy profiles for the serotonin receptor as binary (apo-receptor + G-protein-α-subunit and receptor + ligand) and ternary complexes with two prototypical orthosteric ligands: the full agonist serotonin and the partial agonist aripiprazole. Our results are not only compatible with previously reported experimental and computational data, but they also allow differences between active and inactive conformations to be determined in unprecedented atomic detail, and with respect to the so-called microswitches that have been suggested as determinants of activation, giving insight into their role in the activation mechanism.

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.

Revealing druggable cryptic pockets in the Nsp1 of SARS-CoV-2 and other ß-coronaviruses by simulations and crystallography.

Non-structural protein 1 (Nsp1) is a main pathogenicity factor of α- and ß-coronaviruses. Nsp1 of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) suppresses the host gene expression by sterically blocking 40S host ribosomal subunits and promoting host mRNA degradation. This mechanism leads to the downregulation of the translation-mediated innate immune response in host cells, ultimately mediating the observed immune evasion capabilities of SARS-CoV-2. Here, by combining extensive molecular dynamics simulations, fragment screening and crystallography, we reveal druggable pockets in Nsp1. Structural and computational solvent mapping analyses indicate the partial crypticity of these newly discovered and druggable binding sites. The results of fragment-based screening via X-ray crystallography confirm the druggability of the major pocket of Nsp1. Finally, we show how the targeting of this pocket could disrupt the Nsp1-mRNA complex and open a novel avenue to design new inhibitors for other Nsp1s present in homologous ß-coronaviruses.

Design, Synthesis, and Anticancer Activity of a Selenium-Containing Galectin-3 and Galectin-9N Inhibitor.

Galectins are soluble ß-D-galactoside-binding proteins whose implication in cancer progression and disease outcome makes them prominent targets for therapeutic intervention. In this frame, the development of small inhibitors that block selectively the activity of galectins represents an important strategy for cancer therapy which is, however, still relatively underdeveloped. To this end, we designed here a rationally and efficiently novel diglycosylated compound, characterized by a selenoglycoside bond and the presence of a lipophilic benzyl group at both saccharide residues. The relatively high binding affinity of the new compound to the carbohydrate recognition domain of two galectins, galectin 3 and galectin 9, its good antiproliferative and anti-migration activity towards melanoma cells, as well as its anti-angiogenesis properties, pave the way for its further development as an anticancer agent.

A different perspective for nonphotochemical quenching in plant antenna complexes.

Light-harvesting complexes of plants exert a dual function of light-harvesting (LH) and photoprotection through processes collectively called nonphotochemical quenching (NPQ). While LH processes are relatively well characterized, those involved in NPQ are less understood. Here, we characterize the quenching mechanisms of CP29, a minor LHC of plants, through the integration of two complementary enhanced-sampling techniques, dimensionality reduction schemes, electronic calculations and the analysis of cryo-EM data in the light of the predicted conformational ensemble. Our study reveals that the switch between LH and quenching state is more complex than previously thought. Several conformations of the lumenal side of the protein occur and differently affect the pigments' relative geometries and interactions. Moreover, we show that a quenching mechanism localized on a single chlorophyll-carotenoid pair is not sufficient but many chlorophylls are simultaneously involved. In such a diffuse mechanism, short-range interactions between each carotenoid and different chlorophylls combined with a protein-mediated tuning of the carotenoid excitation energies have to be considered in addition to the commonly suggested Coulomb interactions.

Amphiphilic Histidine-Based Oligopeptides Exhibit pH-Reversible Fibril Formation.

We report the design, simulation, synthesis, and reversible self-assembly of nanofibrils using polyhistidine-based oligopeptides. The inclusion of aromatic amino acids in the histidine block produces distinct antiparallel ß-strands that lead to the formation of amyloid-like fibrils. The structures undergo self-assembly in response to a change in pH. This creates the potential to produce well-defined fibrils for biotechnological and biomedical applications that are pH-responsive in a physiologically relevant range.

Concurrent mutations in RNA-dependent RNA polymerase and spike protein emerged as the epidemiologically most successful SARS-CoV-2 variant.

The D614G mutation in the Spike protein of the SARS-CoV-2 has effectively replaced the early pandemic-causing variant. Using pseudotyped lentivectors, we confirmed that the aspartate replacement by glycine in position 614 is markedly more infectious. Molecular modelling suggests that the G614 mutation facilitates transition towards an open state of the Spike protein. To explain the epidemiological success of D614G, we analysed the evolution of 27,086 high-quality SARS-CoV-2 genome sequences from GISAID. We observed striking coevolution of D614G with the P323L mutation in the viral polymerase. Importantly, the exclusive presence of G614 or L323 did not become epidemiologically relevant. In contrast, the combination of the two mutations gave rise to a viral G/L variant that has all but replaced the initial D/P variant. Our results suggest that the P323L mutation, located in the interface domain of the RNA-dependent RNA polymerase, is a necessary alteration that led to the epidemiological success of the present variant of SARS-CoV-2. However, we did not observe a significant correlation between reported COVID-19 mortality in different countries and the prevalence of the Wuhan versus G/L variant. Nevertheless, when comparing the speed of emergence and the ultimate predominance in individual countries, it is clear that the G/L variant displays major epidemiological supremacy over the original variant.

Structural basis of the effect of activating mutations on the EGF receptor.

Mutations within the kinase domain of the epidermal growth factor receptor (EGFR) are common oncogenic driver events in non-small cell lung cancer. Although the activation of EGFR in normal cells is primarily driven by growth-factor-binding-induced dimerization, mutations on different exons of the kinase domain of the receptor have been found to affect the equilibrium between its active and inactive conformations giving rise to growth-factor-independent kinase activation. Using molecular dynamics simulations combined with enhanced sampling techniques, we compare here the conformational landscape of the monomers and homodimers of the wild-type and mutated forms of EGFR ΔELREA and L858R, as well as of two exon 20 insertions, D770-N771insNPG, and A763-Y764insFQEA. The differences in the conformational energy landscapes are consistent with multiple mechanisms of action including the regulation of the hinge motion, the stabilization of the dimeric interface, and local unfolding transitions. Overall, a combination of different effects is caused by the mutations and leads to the observed aberrant signaling.

Allosteric communication in class A ß-lactamases occurs via cooperative coupling of loop dynamics.

Understanding allostery in enzymes and tools to identify it offer promising alternative strategies to inhibitor development. Through a combination of equilibrium and nonequilibrium molecular dynamics simulations, we identify allosteric effects and communication pathways in two prototypical class A ß-lactamases, TEM-1 and KPC-2, which are important determinants of antibiotic resistance. The nonequilibrium simulations reveal pathways of communication operating over distances of 30 Å or more. Propagation of the signal occurs through cooperative coupling of loop dynamics. Notably, 50% or more of clinically relevant amino acid substitutions map onto the identified signal transduction pathways. This suggests that clinically important variation may affect, or be driven by, differences in allosteric behavior, providing a mechanism by which amino acid substitutions may affect the relationship between spectrum of activity, catalytic turnover, and potential allosteric behavior in this clinically important enzyme family. Simulations of the type presented here will help in identifying and analyzing such differences.

Antibiotics are crucial drugs for treating and preventing bacterial infections, but some bacteria are evolving ways to resist their effects. This 'antibiotic resistance' threatens lives and livelihoods worldwide. ß-lactam antibiotics, like penicillin, are some of the most commonly used, but some bacteria can now make enzymes called ß-lactamases, which destroy these antibiotics. Dozens of different types of ß-lactamases now exist, each with different properties. Two of the most medically important are TEM-1 and KPC-2. One way to counteract ß-lactamases is with drugs called inhibitors that stop the activity of these enzymes. The approved ß-lactamase inhibitors work by blocking the part of the enzyme that binds and destroys antibiotics, known as the 'active site'. The ß-lactamases have evolved, some of which have the ability to resist the effects of known inhibitors. It is possible that targeting parts of ß-lactamases far from the active site, known as 'allosteric sites', might get around these new bacterial defences. A molecule that binds to an allosteric site might alter the enzyme's shape, or restrict its movement, making it unable to do its job. Galdadas, Qu et al. used simulations to understand how molecules binding at allosteric sites affect enzyme movement. The experiments examined the structures of both TEM-1 and KPC-2, looking at how their shapes changed as molecules were removed from the allosteric site. This revealed how the allosteric sites and the active site are linked together. When molecules were taken out of the allosteric sites, they triggered ripples of shape change that travelled via loop-like structures across the surface of the enzyme. These loops contain over half of the known differences between the different types of ß-lactamases, suggesting mutations here may be responsible for changing which antibiotics each enzyme can destroy. In other words, changes in the 'ripples' may be related to the ability of the enzymes to resist particular antibiotics. Understanding how changes in one part of a ß-lactamase enzyme reach the active site could help in the design of new inhibitors. It might also help to explain how ß-lactamases evolve new properties. Further work could show why different enzymes are more or less active against different antibiotics.

A combined activation mechanism for the glucagon receptor.

We report on a combined activation mechanism for a class B G-protein-coupled receptor (GPCR), the glucagon receptor. By computing the conformational free-energy landscape associated with the activation of the receptor-agonist complex and comparing it with that obtained with the ternary complex (receptor-agonist-G protein) we show that the agonist stabilizes the receptor in a preactivated complex, which is then fully activated upon binding of the G protein. The proposed mechanism contrasts with the generally assumed GPCR activation mechanism, which proceeds through an opening of the intracellular region allosterically elicited by the binding of the agonist. The mechanism found here is consistent with electron cryo-microscopy structural data and might be general for class B GPCRs. It also helps us to understand the mode of action of the numerous allosteric antagonists of this important drug target.