Explore the vast range of titles available.

Pharmacological responses of G protein-coupled receptors (GPCRs) can be fine-tuned by allosteric modulators. Structural studies of such effects have been limited due to the medium resolution of GPCR structures. We reengineered the human A 2A adenosine receptor by replacing its third intracellular loop with apocytochrome b⁵⁶² RIL and solved the structure at 1.8 angstrom resolution. The high-resolution structure allowed us to identify 57 ordered water molecules inside the receptor comprising three major clusters. The central cluster harbors a putative sodium ion bound to the highly conserved aspartate residue Asp 2.50 . Additionally, two cholesterols stabilize the conformation of helix VI, and one of 23 ordered lipids intercalates inside the ligand-binding pocket. These high-resolution details shed light on the potential role of structured water molecules, sodium ions, and lipids/cholesterol in GPCR stabilization and function.

The increase of publicly available bioactivity data in recent years has fueled and catalyzed research in chemogenomics, data mining, and modeling approaches. As a direct result, over the past few years a multitude of different methods have been reported and evaluated, such as target fishing, nearest neighbor similarity-based methods, and Quantitative Structure Activity Relationship (QSAR)-based protocols. However, such studies are typically conducted on different datasets, using different validation strategies, and different metrics. In this study, different methods were compared using one single standardized dataset obtained from ChEMBL, which is made available to the public, using standardized metrics (BEDROC and Matthews Correlation Coefficient). Specifically, the performance of Naïve Bayes, Random Forests, Support Vector Machines, Logistic Regression, and Deep Neural Networks was assessed using QSAR and proteochemometric (PCM) methods. All methods were validated using both a random split validation and a temporal validation, with the latter being a more realistic benchmark of expected prospective execution. Deep Neural Networks are the top performing classifiers, highlighting the added value of Deep Neural Networks over other more conventional methods. Moreover, the best method (‘DNN_PCM’) performed significantly better at almost one standard deviation higher than the mean performance. Furthermore, Multi-task and PCM implementations were shown to improve performance over single task Deep Neural Networks. Conversely, target prediction performed almost two standard deviations under the mean performance. Random Forests, Support Vector Machines, and Logistic Regression performed around mean performance. Finally, using an ensemble of DNNs, alongside additional tuning, enhanced the relative performance by another 27% (compared with unoptimized ‘DNN_PCM’). Here, a standardized set to test and evaluate different machine learning algorithms in the context of multi-task learning is offered by providing the data and the protocols. Graphical Abstract .

Rational drug design often starts from specific scaffolds to which side chains/substituents are added or modified due to the large drug-like chemical space available to search for novel drug-like molecules. With the rapid growth of deep learning in drug discovery, a variety of effective approaches have been developed for de novo drug design. In previous work we proposed a method named DrugEx , which can be applied in polypharmacology based on multi-objective deep reinforcement learning. However, the previous version is trained under fixed objectives and does not allow users to input any prior information ( i.e. a desired scaffold). In order to improve the general applicability, we updated DrugEx to design drug molecules based on scaffolds which consist of multiple fragments provided by users. Here, a  Transformer model was employed to generate molecular structures. The Transformer is a multi-head self-attention deep learning model containing an encoder to receive scaffolds as input and a decoder to generate molecules as output. In order to deal with the graph representation of molecules a novel positional encoding for each atom and bond based on an adjacency matrix was proposed, extending the architecture of the Transformer. The graph Transformer model contains growing and connecting procedures for molecule generation starting from  a given scaffold based on fragments. Moreover, the generator was trained under a reinforcement learning framework to increase the number of desired ligands. As a proof of concept, the method was applied to design ligands for the adenosine A 2A receptor (A 2A AR) and compared with SMILES-based methods. The results show that 100% of the generated molecules are valid and most of them had a high predicted affinity value towards A 2A AR with given scaffolds.

In polypharmacology drugs are required to bind to multiple specific targets, for example to enhance efficacy or to reduce resistance formation. Although deep learning has achieved a breakthrough in de novo design in drug discovery, most of its applications only focus on a single drug target to generate drug-like active molecules. However, in reality drug molecules often interact with more than one target which can have desired (polypharmacology) or undesired (toxicity) effects. In a previous study we proposed a new method named DrugEx that integrates an exploration strategy into RNN-based reinforcement learning to improve the diversity of the generated molecules. Here, we extended our DrugEx algorithm with multi-objective optimization to generate drug-like molecules towards multiple targets or one specific target while avoiding off-targets (the two adenosine receptors, A 1 AR and A 2A AR, and the potassium ion channel hERG in this study). In our model, we applied an RNN as the agent and machine learning predictors as the environment . Both the agent and the environment were pre-trained in advance and then interplayed under a reinforcement learning framework. The concept of evolutionary algorithms was merged into our method such that crossover and mutation operations were implemented by the same deep learning model as the agent . During the training loop, the agent generates a batch of SMILES-based molecules. Subsequently scores for all objectives provided by the environment are used to construct Pareto ranks of the generated molecules. For this ranking a non-dominated sorting algorithm and a Tanimoto-based crowding distance algorithm using chemical fingerprints are applied. Here, we adopted GPU acceleration to speed up the process of Pareto optimization. The final reward of each molecule is calculated based on the Pareto ranking with the ranking selection algorithm. The agent is trained under the guidance of the reward to make sure it can generate desired molecules after convergence of the training process. All in all we demonstrate generation of compounds with a diverse predicted selectivity profile towards multiple targets, offering the potential of high efficacy and low toxicity.

The crystal structure of CCR2 chemokine receptor in a complex with two different antagonists—one orthosteric the other allosteric—which functionally cooperate to inhibit CCR2. Small-molecule chemokine receptor antagonists Chemokine receptors are a family of G-protein-coupled receptors that regulate the migration of immune cells; their function has been implicated in a range of diseases. Two groups reporting in this issue of Nature describe crystal structures of two different chemokine receptors bound to small-molecule inhibitors. Tracy Handel and colleagues describe the structure of CCR2—a promising drug target for autoimmune, inflammatory and metabolic diseases as well as cancer—bound to orthosteric (BMS-681) and allosteric (CCR2-RA-[ R ]) antagonists. Fiona Marshall and colleagues describe the structure of CCR9—involved in immune cell recruitment to the gut and a promising drug target in inflammatory bowel disease—in complex with the selective CCR9 antagonist vercirnon. Both CCR2 and CCR9 structures reveal an allosteric pocket on the cytoplasmic face of the receptor. This allosteric pocket appears to be highly druggable, and homologous pockets may be present on other chemokine receptors. CC chemokine receptor 2 (CCR2) is one of 19 members of the chemokine receptor subfamily of human class A G-protein-coupled receptors. CCR2 is expressed on monocytes, immature dendritic cells, and T-cell subpopulations, and mediates their migration towards endogenous CC chemokine ligands such as CCL2 (ref. 1 ). CCR2 and its ligands are implicated in numerous inflammatory and neurodegenerative diseases 2 including atherosclerosis, multiple sclerosis, asthma, neuropathic pain, and diabetic nephropathy, as well as cancer 3 . These disease associations have motivated numerous preclinical studies and clinical trials 4 (see http://www.clinicaltrials.gov ) in search of therapies that target the CCR2–chemokine axis. To aid drug discovery efforts 5 , here we solve a structure of CCR2 in a ternary complex with an orthosteric (BMS-681 (ref. 6 )) and allosteric (CCR2-RA-[ R ] 7 ) antagonist. BMS-681 inhibits chemokine binding by occupying the orthosteric pocket of the receptor in a previously unseen binding mode. CCR2-RA-[ R ] binds in a novel, highly druggable pocket that is the most intracellular allosteric site observed in class A G-protein-coupled receptors so far; this site spatially overlaps the G-protein-binding site in homologous receptors. CCR2-RA-[ R ] inhibits CCR2 non-competitively by blocking activation-associated conformational changes and formation of the G-protein-binding interface. The conformational signature of the conserved microswitch residues observed in double-antagonist-bound CCR2 resembles the most inactive G-protein-coupled receptor structures solved so far. Like other protein–protein interactions, receptor–chemokine complexes are considered challenging therapeutic targets for small molecules, and the present structure suggests diverse pocket epitopes that can be exploited to overcome obstacles in drug design.

Transporters are important therapeutic but yet understudied targets due to lack of available assays. Here we describe a novel label-free, whole-cell method for the functional assessment of Solute Carrier (SLC) inhibitors. As many SLC substrates are also ligands for G protein-coupled receptors (GPCRs), transporter inhibition may affect GPCR signalling due to a change in extracellular concentration of the substrate/ligand, which can be monitored by an impedance-based label-free assay. For this study, a prototypical SLC/GPCR pair was selected, i.e. the equilibrative nucleoside transporter-1 (SLC29A1/ENT1) and an adenosine receptor (AR), for which adenosine is the substrate/ligand. ENT1 inhibition with three reference compounds was monitored sensitively via AR activation on human osteosarcoma cells. Firstly, the inhibitor addition resulted in an increased apparent potency of adenosine. Secondly, all inhibitors concentration-dependently increased the extracellular adenosine concentration, resulting in an indirect quantitative assessment of their potencies. Additionally, AR activation was abolished by AR antagonists, confirming that the monitored impedance was AR-mediated. In summary, we developed a novel assay as an in vitro model system that reliably assessed the potency of SLC29A1 inhibitors via AR signalling. As such, the method may be applied broadly as it has the potential to study a multitude of SLCs via concomitant GPCR signalling.

The adenosine class of heterotrimeric guanine nucleotide-binding protein (G protein)-coupled receptors (GPCRs) mediates the important role of extracellular adenosine in many physiological processes and is antagonized by caffeine. We have determined the crystal structure of the human A₂A adenosine receptor, in complex with a high-affinity subtype-selective antagonist, ZM241385, to 2.6 angstrom resolution. Four disulfide bridges in the extracellular domain, combined with a subtle repacking of the transmembrane helices relative to the adrenergic and rhodopsin receptor structures, define a pocket distinct from that of other structurally determined GPCRs. The arrangement allows for the binding of the antagonist in an extended conformation, perpendicular to the membrane plane. The binding site highlights an integral role for the extracellular loops, together with the helical core, in ligand recognition by this class of GPCRs and suggests a role for ZM241385 in restricting the movement of a tryptophan residue important in the activation mechanism of the class A receptors.

Transmembranal G Protein-Coupled Receptors (GPCRs) transduce extracellular chemical signals to the cell, via conformational change from a resting (inactive) to an active (canonically bound to a G-protein) conformation. Receptor activation is normally modulated by extracellular ligand binding, but mutations in the receptor can also shift this equilibrium by stabilizing different conformational states. In this work, we built structure-energetic relationships of receptor activation based on original thermodynamic cycles that represent the conformational equilibrium of the prototypical A 2A adenosine receptor (AR). These cycles were solved with efficient free energy perturbation (FEP) protocols, allowing to distinguish the pharmacological profile of different series of A 2A AR agonists with different efficacies. The modulatory effects of point mutations on the basal activity of the receptor or on ligand efficacies could also be detected. This methodology can guide GPCR ligand design with tailored pharmacological properties, or allow the identification of mutations that modulate receptor activation with potential clinical implications.

Over the last 5 years deep learning has progressed tremendously in both image recognition and natural language processing. Now it is increasingly applied to other data rich fields. In drug discovery, recurrent neural networks (RNNs) have been shown to be an effective method to generate novel chemical structures in the form of SMILES. However, ligands generated by current methods have so far provided relatively low diversity and do not fully cover the whole chemical space occupied by known ligands. Here, we propose a new method (DrugEx) to discover de novo drug-like molecules. DrugEx is an RNN model (generator) trained through reinforcement learning which was integrated with a special exploration strategy. As a case study we applied our method to design ligands against the adenosine A 2A receptor. From ChEMBL data, a machine learning model (predictor) was created to predict whether generated molecules are active or not. Based on this predictor as the reward function, the generator was trained by reinforcement learning without any further data. We then compared the performance of our method with two previously published methods, REINVENT and ORGANIC. We found that candidate molecules our model designed, and predicted to be active, had a larger chemical diversity and better covered the chemical space of known ligands compared to the state-of-the-art.

The sodium taurocholate co-transporting polypeptide (NTCP, SLC10A1 ) is the main hepatic transporter of conjugated bile acids, and the entry receptor for hepatitis B virus (HBV) and hepatitis delta virus (HDV). Myrcludex B, a synthetic peptide mimicking the NTCP-binding domain of HBV, effectively blocks HBV and HDV infection. In addition, Myrcludex B inhibits NTCP-mediated bile acid uptake, suggesting that also other NTCP inhibitors could potentially be a novel treatment of HBV/HDV infection. This study aims to identify clinically-applied compounds intervening with NTCP-mediated bile acid transport and HBV/HDV infection. 1280 FDA/EMA-approved drugs were screened to identify compounds that reduce uptake of taurocholic acid and lower Myrcludex B-binding in U2OS cells stably expressing human NTCP. HBV/HDV viral entry inhibition was studied in HepaRG cells. The four most potent inhibitors of human NTCP were rosiglitazone (IC 50 5.1 µM), zafirlukast (IC 50 6.5 µM), TRIAC (IC 50 6.9 µM), and sulfasalazine (IC 50 9.6 µM). Chicago sky blue 6B (IC 50 7.1 µM) inhibited both NTCP and ASBT, a distinct though related bile acid transporter. Rosiglitazone, zafirlukast, TRIAC, sulfasalazine, and chicago sky blue 6B reduced HBV/HDV infection in HepaRG cells in a dose-dependent manner. Five out of 1280 clinically approved drugs were identified that inhibit NTCP-mediated bile acid uptake and HBV/HDV infection in vitro .