Activation of human STING by a molecular glue-like compound.

Stimulator of interferon genes (STING) is a dimeric transmembrane adapter protein that plays a key role in the human innate immune response to infection and has been therapeutically exploited for its antitumor activity. The activation of STING requires its high-order oligomerization, which could be induced by binding of the endogenous ligand, cGAMP, to the cytosolic ligand-binding domain. Here we report the discovery through functional screens of a class of compounds, named NVS-STGs, that activate human STING. Our cryo-EM structures show that NVS-STG2 induces the high-order oligomerization of human STING by binding to a pocket between the transmembrane domains of the neighboring STING dimers, effectively acting as a molecular glue. Our functional assays showed that NVS-STG2 could elicit potent STING-mediated immune responses in cells and antitumor activities in animal models.

Reactivities of acrylamide warheads toward cysteine targets: a QM/ML approach to covalent inhibitor design.

Covalent inhibition offers many advantages over non-covalent inhibition, but covalent warhead reactivity must be carefully balanced to maintain potency while avoiding unwanted side effects. While warhead reactivities are commonly measured with assays, a computational model to predict warhead reactivities could be useful for several aspects of the covalent inhibitor design process. Studies have shown correlations between covalent warhead reactivities and quantum mechanic (QM) properties that describe important aspects of the covalent reaction mechanism. However, the models from these studies are often linear regression equations and can have limitations associated with their usage. Applications of machine learning (ML) models to predict covalent warhead reactivities with QM descriptors are not extensively seen in the literature. This study uses QM descriptors, calculated at different levels of theory, to train ML models to predict reactivities of covalent acrylamide warheads. The QM/ML models are compared with linear regression models built upon the same QM descriptors and with ML models trained on structure-based features like Morgan fingerprints and RDKit descriptors. Experiments show that the QM/ML models outperform the linear regression models and the structure-based ML models, and literature test sets demonstrate the power of the QM/ML models to predict reactivities of unseen acrylamide warhead scaffolds. Ultimately, these QM/ML models are effective, computationally feasible tools that can expedite the design of new covalent inhibitors.

Benchmarking In Silico Tools for Cysteine pKa Prediction.

Accurate estimation of the pKa's of cysteine residues in proteins could inform targeted approaches in hit discovery. The pKa of a targetable cysteine residue in a disease-related protein is an important physiochemical parameter in covalent drug discovery, as it influences the fraction of nucleophilic thiolate amenable to chemical protein modification. Traditional structure-based in silico tools are limited in their predictive accuracy of cysteine pKa's relative to other titratable residues. Additionally, there are limited comprehensive benchmark assessments for cysteine pKa predictive tools. This raises the need for extensive assessment and evaluation of methods for cysteine pKa prediction. Here, we report the performance of several computational pKa methods, including single-structure and ensemble-based approaches, on a diverse test set of experimental cysteine pKa's retrieved from the PKAD database. The dataset consisted of 16 wildtype and 10 mutant proteins with experimentally measured cysteine pKa values. Our results highlight that these methods are varied in their overall predictive accuracies. Among the test set of wildtype proteins evaluated, the best method (MOE) yielded a mean absolute error of 2.3 pK units, highlighting the need for improvement of existing pKa methods for accurate cysteine pKa estimation. Given the limited accuracy of these methods, further development is needed before these approaches can be routinely employed to drive design decisions in early drug discovery efforts.

Leveraging Advanced In Silico Techniques in Early Drug Discovery: A Study of Potent Small-Molecule YAP-TEAD PPI Disruptors.

Disruption of the YAP-TEAD protein-protein interaction is an attractive therapeutic strategy in oncology to suppress tumor progression and cancer metastasis. YAP binds to TEAD at a large flat binding interface (∼3500 Å2) devoid of a well-defined druggable pocket, so it has been difficult to design low-molecular-weight compounds to abrogate this protein-protein interaction directly. Recently, work by Furet and coworkers (ChemMedChem 2022, DOI: 10.1002/cmdc.202200303) reported the discovery of the first class of small molecules able to efficiently disrupt the transcriptional activity of TEAD by binding to a specific interaction site of the YAP-TEAD binding interface. Using high-throughput in silico docking, they identified a virtual screening hit from a hot spot derived from their previously rationally designed peptidic inhibitor. Structure-based drug design efforts led to the optimization of the hit compound into a potent lead candidate. Given advances in rapid high-throughput screening and rational approaches to peptidic ligand discovery for challenging targets, we analyzed the pharmacophore features involved in transferring from the peptidic to small-molecule inhibitor that could enable small-molecule discovery for such targets. Here, we show retrospectively that pharmacophore analysis augmented by solvation analysis of molecular dynamics trajectories can guide the designs, while binding free energy calculations provide greater insight into the binding conformation and energetics accompanying the association event. The computed binding free energy estimates agree well with experimental findings and offer useful insight into structural determinants that influence ligand binding to the TEAD interaction surface, even for such a shallow binding site. Taken together, our results demonstrates the utility of advanced in silico methods in structure-based design efforts for difficult-to-drug targets such as the YAP-TEAD transcription factor complex.

Author Correction: Structural basis of indisulam-mediated RBM39 recruitment to DCAF15 E3 ligase complex.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Structural basis of indisulam-mediated RBM39 recruitment to DCAF15 E3 ligase complex.

The anticancer agent indisulam inhibits cell proliferation by causing degradation of RBM39, an essential mRNA splicing factor. Indisulam promotes an interaction between RBM39 and the DCAF15 E3 ligase substrate receptor, leading to RBM39 ubiquitination and proteasome-mediated degradation. To delineate the precise mechanism by which indisulam mediates the DCAF15-RBM39 interaction, we solved the DCAF15-DDB1-DDA1-indisulam-RBM39(RRM2) complex structure to a resolution of 2.3 Å. DCAF15 has a distinct topology that embraces the RBM39(RRM2) domain largely via non-polar interactions, and indisulam binds between DCAF15 and RBM39(RRM2), coordinating additional interactions between the two proteins. Studies with RBM39 point mutants and indisulam analogs validated the structural model and defined the RBM39 α-helical degron motif. The degron is found only in RBM23 and RBM39, and only these proteins were detectably downregulated in indisulam-treated HCT116 cells. This work further explains how indisulam induces RBM39 degradation and defines the challenge of harnessing DCAF15 to degrade additional targets.

Relative Binding Free-Energy Calculations at Lipid-Exposed Sites: Deciphering Hot Spots.

Relative binding free-energy (RBFE) calculations are experiencing resurgence in the computer-aided drug design of novel small molecules due to performance gains allowed by cutting-edge molecular mechanic force fields and computer hardware. Application of RBFE to soluble proteins is becoming a routine, while recent studies outline necessary steps to successfully apply RBFE at the orthosteric site of membrane-embedded G-protein-coupled receptors (GPCRs). In this work, we apply RBFE to a congeneric series of antagonists that bind to a lipid-exposed, extra-helical site of the P2Y1 receptor. We find promising performance of RBFE, such that it may be applied in a predictive manner on drug discovery programs targeting lipid-exposed sites. Further, by the application of the microkinetic model, binding at a lipid-exposed site can be split into (1) membrane partitioning of the drug molecule followed by (2) binding at the extra-helical site. We find that RBFE can be applied to calculate the free energy of each step, allowing the uncoupling of observed binding free energy from the influence of membrane affinity. This protocol may be used to identify binding hot spots at extra-helical sites and guide drug discovery programs toward optimizing intrinsic activity at the target.

Using Membrane Partitioning Simulations To Predict Permeability of Forty-Nine Drug-Like Molecules.

A simple descriptor calculated from molecular dynamics simulations of the membrane partitioning event is found to correlate well with experimental measurements of passive membrane permeation from the high-throughput MDCK-LE assay using a data set of 49 drug-like molecules. This descriptor approximates the energy cost of translocation across the hydrophobic membrane core (flip-flop), which for many molecules limits permeability. Performance is found to be superior in comparison to calculated properties such as clogP, clogD, or polar surface area. Furthermore, the atomistic simulations provide a structural understanding of the partitioned drug-membrane complex, facilitating medicinal chemistry optimization of membrane permeability.

Structure-Kinetic Relationships of Passive Membrane Permeation from Multiscale Modeling.

Passive membrane permeation of small molecules is essential to achieve the required absorption, distribution, metabolism, and excretion (ADME) profiles of drug candidates, in particular intestinal absorption and transport across the blood-brain barrier. Computational investigations of this process typically involve either building QSAR models or performing free energy calculations of the permeation event. Although insightful, these methods rarely bridge the gap between computation and experiment in a quantitative manner, and identifying structural insights to apply toward the design of compounds with improved permeability can be difficult. In this work, we combine molecular dynamics simulations capturing the kinetic steps of permeation at the atomistic level with a dynamic mechanistic model describing permeation at the in vitro level, finding a high level of agreement with experimental permeation measurements. Calculation of the kinetic rate constants determining each step in the permeation event allows derivation of structure-kinetic relationships of permeation. We use these relationships to probe the structural determinants of membrane permeation, finding that the desolvation/loss of hydrogen bonding required to leave the membrane partitioned position controls the membrane flip-flop rate, whereas membrane partitioning determines the rate of leaving the membrane.

Contributions of the membrane dipole potential to the function of voltage-gated cation channels and modulation by small molecule potentiators.

The membrane dipole potential (Ψd) constitutes one of three electrical potentials generated by cell membranes. Ψd arises from the unfavorable parallel alignment of phospholipid and water dipoles, and varies in magnitude both longitudinally and laterally across the bilayer according to membrane composition and phospholipid packing density. In this work, we propose that dynamic counter-balancing between Ψd and the transmembrane potential (ΔΨm) governs the conformational state transitions of voltage-gated ion channels. Ψd consists of 1) static outer, and dynamic inner leaflet components (Ψd(extra) and Ψd(intra), respectively); and 2) a transmembrane component (ΔΨd(inner-outer)), ariing from differences in intra- and extracellular leaflet composition. Ψd(intra), which transitions between high and low energy states (Ψd(intra, high) and Ψd(intra, low)) as a function of channel conformation, is transduced by the pore domain. ΔΨd(inner-outer) is transduced by the voltage-sensing (VS) domain in summation with ΔΨm. Potentiation of voltage-gated ion channels is of interest for the treatment of cardiac, neuronal, and other disorders arising from inherited/acquired ion channel dysfunction. Potentiators are widely believed to alter the rates and voltage-dependencies of channel gating transitions by binding to pockets in the membrane-facing and other regions of ion channel targets. Here, we propose that potentiators alter Ψd(intra) and/or Ψd(extra), thereby increasing or decreasing the energy barriers governing channel gating transitions. We used quantum mechanical and molecular dynamics (MD) simulations to predict the overall Ψd-modulating effects of a series of published positive hERG potentiators partitioned into model DOPC bilayers. Our findings suggest a strong correlation between the magnitude of Ψd-lowering and positive hERG potentiation across the series.

Discovery of phenyl acetamides as potent and selective GPR119 agonists.

The paper describes the SAR/SPR studies that led to the discovery of phenoxy cyclopropyl phenyl acetamide derivatives as potent and selective GPR119 agonists. Based on a cis cyclopropane scaffold discovered previously, phenyl acetamides such as compound 17 were found to have excellent GPR119 potency and improved physicochemical properties. Pharmacokinetic data of compound 17 in rat, dog and rhesus will be described. Compound 17 was suitable for QD dosing based on its predicted human half-life, and its projected human dose was much lower than that of the recently reported structurally-related benzyloxy compound 2. Compound 17 was selected as a tool compound candidate for NHP (Non-Human Primate) efficacy studies.

Structure-activity relationship of cytochrome bc1 reductase inhibitor broad spectrum antifungal ilicicolin H.

Ilicicolin H is a broad spectrum antifungal agent showing sub micro g/mL MICs against Candida spp., Aspergillus fumigatus and Cryptococcus spp. It is a potent inhibitor (C50 2-3ng/mL) of the mitochondrial cytochrome bc1 reductase with over 1000-fold selectivity against rat liver cytochrome bc1 reductase. Structure-activity relationship of semisynthetic derivatives by chemical modification of ilicicolin H and its 19-hydroxy derivative produced by biotransformation have been described. Basic 4'-esters and moderately polar N- and O-alkyl derivatives retained antifungal and the cytochrome bc1 reductase activities. 4',19-Diacetate and 19-cyclopropyl acetate retained antifungal and enzyme activity and selectivity with over 20-fold improvement of plasma protein binding.

Comparison of random forest and Pipeline Pilot Naïve Bayes in prospective QSAR predictions.

Random forest is currently considered one of the best QSAR methods available in terms of accuracy of prediction. However, it is computationally intensive. Naïve Bayes is a simple, robust classification method. The Laplacian-modified Naïve Bayes implementation is the preferred QSAR method in the widely used commercial chemoinformatics platform Pipeline Pilot. We made a comparison of the ability of Pipeline Pilot Naïve Bayes (PLPNB) and random forest to make accurate predictions on 18 large, diverse in-house QSAR data sets. These include on-target and ADME-related activities. These data sets were set up as classification problems with either binary or multicategory activities. We used a time-split method of dividing training and test sets, as we feel this is a realistic way of simulating prospective prediction. PLPNB is computationally efficient. However, random forest predictions are at least as good and in many cases significantly better than those of PLPNB on our data sets. PLPNB performs better with ECFP4 and ECFP6 descriptors, which are native to Pipeline Pilot, and more poorly with other descriptors we tried.

The design and synthesis of potent, selective benzodiazepine sulfonamide bombesin receptor subtype 3 (BRS-3) agonists with an increased barrier of atropisomerization.

Bombesin receptor subtype 3 (BRS-3) is an orphan G-protein coupled receptor expressed primarily in the hypothalamus which plays a role in the onset of both diabetes and obesity. We report herein our progress made towards identifying a potent, selective bombesin receptor subtype-3 (BRS-3) agonist related to the previously described MK-7725(1) Chobanian et al. (2012) that would prevent atropisomerization through the increase of steric bulk at the C-2 position. This would thereby make clinical development of this class of compounds more cost effective by inhibiting racemization which can occur over long periods of time at room/elevated temperature.

Combined Free-Energy Calculation and Machine Learning Methods for Understanding Ligand Unbinding Kinetics.

The determination of drug residence times, which define the time an inhibitor is in complex with its target, is a fundamental part of the drug discovery process. Synthesis and experimental measurements of kinetic rate constants are, however, expensive and time consuming. In this work, we aimed to obtain drug residence times computationally. Furthermore, we propose a novel algorithm to identify molecular design objectives based on ligand unbinding kinetics. We designed an enhanced sampling technique to accurately predict the free-energy profiles of the ligand unbinding process, focusing on the free-energy barrier for unbinding. Our method first identifies unbinding paths determining a corresponding set of internal coordinates (ICs) that form contacts between the protein and the ligand; it then iteratively updates these interactions during a series of biased molecular dynamics (MD) simulations to reveal the ICs that are important for the whole of the unbinding process. Subsequently, we performed finite-temperature string simulations to obtain the free-energy barrier for unbinding using the set of ICs as a complex reaction coordinate. Importantly, we also aimed to enable the further design of drugs focusing on improved residence times. To this end, we developed a supervised machine learning (ML) approach with inputs from unbiased "downhill" trajectories initiated near the transition state (TS) ensemble of the string unbinding path. We demonstrate that our ML method can identify key ligand-protein interactions driving the system through the TS. Some of the most important drugs for cancer treatment are kinase inhibitors. One of these kinase targets is cyclin-dependent kinase 2 (CDK2), an appealing target for anticancer drug development. Here, we tested our method using two different CDK2 inhibitors for the potential further development of these compounds. We compared the free-energy barriers obtained from our calculations with those observed in available experimental data. We highlighted important interactions at the distal ends of the ligands that can be targeted for improved residence times. Our method provides a new tool to determine unbinding rates and to identify key structural features of the inhibitors that can be used as starting points for novel design strategies in drug discovery.

Structure-Based Design and Preclinical Characterization of Selective and Orally Bioavailable Factor XIa Inhibitors: Demonstrating the Power of an Integrated S1 Protease Family Approach.

The serine protease factor XI (FXI) is a prominent drug target as it holds promise to deliver efficacious anticoagulation without an enhanced risk of major bleeds. Several efforts have been described targeting the active form of the enzyme, FXIa. Herein, we disclose our efforts to identify potent, selective, and orally bioavailable inhibitors of FXIa. Compound 1, identified from a diverse library of internal serine protease inhibitors, was originally designed as a complement factor D inhibitor and exhibited submicromolar FXIa activity and an encouraging absorption, distribution, metabolism, and excretion (ADME) profile while being devoid of a peptidomimetic architecture. Optimization of interactions in the S1, S1ß, and S1' pockets of FXIa through a combination of structure-based drug design and traditional medicinal chemistry led to the discovery of compound 23 with subnanomolar potency on FXIa, enhanced selectivity over other coagulation proteases, and a preclinical pharmacokinetics (PK) profile consistent with bid dosing in patients.

Hidden bias in the DUD-E dataset leads to misleading performance of deep learning in structure-based virtual screening.

Recently much effort has been invested in using convolutional neural network (CNN) models trained on 3D structural images of protein-ligand complexes to distinguish binding from non-binding ligands for virtual screening. However, the dearth of reliable protein-ligand x-ray structures and binding affinity data has required the use of constructed datasets for the training and evaluation of CNN molecular recognition models. Here, we outline various sources of bias in one such widely-used dataset, the Directory of Useful Decoys: Enhanced (DUD-E). We have constructed and performed tests to investigate whether CNN models developed using DUD-E are properly learning the underlying physics of molecular recognition, as intended, or are instead learning biases inherent in the dataset itself. We find that superior enrichment efficiency in CNN models can be attributed to the analogue and decoy bias hidden in the DUD-E dataset rather than successful generalization of the pattern of protein-ligand interactions. Comparing additional deep learning models trained on PDBbind datasets, we found that their enrichment performances using DUD-E are not superior to the performance of the docking program AutoDock Vina. Together, these results suggest that biases that could be present in constructed datasets should be thoroughly evaluated before applying them to machine learning based methodology development.

Structure of a transient intermediate for GTP hydrolysis by ras.

The flexibility of the conserved 57DTAGQ61 motif is essential for Ras proper cycling in response to growth factors. Here, we increase the flexibility of the 57DTAGQ61 motif by mutating Gln61 to Gly. The crystal structure of the RasQ61G mutant reveals a new conformation of switch 2 that bears remarkable structural homology to an intermediate for GTP hydrolysis revealed by targeted molecular dynamics simulations. The mutation increased retention of GTP and inhibited Ras binding to the catalytic site, but not to the distal site of Sos. Most importantly, the thermodynamics of RafRBD binding to Ras are altered even though the structure of switch 1 is not affected by the mutation. Our results suggest that interplay and transmission of structural information between the switch regions are important factors for Ras function. They propose that initiation of GTP hydrolysis sets off the separation of the Ras/effector complex even before the GDP conformation is reached.

Calculating Kinetic Rates and Membrane Permeability from Biased Simulations.

We present a simple approach to calculate the kinetic properties of lipid membrane crossing processes from biased molecular dynamics simulations. We demonstrate that by using biased simulations, one can obtain highly accurate kinetic information with significantly reduced computational time with respect to unbiased simulations. We describe how to conveniently calculate the transition rates to enter, cross, and exit the membrane in terms of the mean first passage times. To obtain free energy barriers and relaxation times from biased simulations only, we constructed Markov models using the dynamic histogram analysis method (DHAM). The permeability coefficients that are calculated from the relaxation times are found to correlate highly with experimentally evaluated values. We show that more generally, certain calculated kinetic properties linked to the crossing of the membrane layer (e.g., barrier height and barrier crossing rates) are good indicators of ordering drugs by permeability. Extending the analysis to a 2D Markov model provides a physical description of the membrane crossing mechanism.

Targeting structural flexibility in HIV-1 protease inhibitor binding.

HIV-1 protease remains an important anti-AIDS drug target. Although it has been known that ligand binding induces large conformational changes in the protease, the dynamic aspects of binding have been largely ignored. Several computational models describing protease dynamics have been reported recently. These have reproduced experimental observations, and have also explained how ligands gain access to the binding site through dynamic behavior of the protease. Specifically, the transitions between three different conformations of the protein have been modeled in atomic detail. Two of these forms were determined by crystallography, and the third was implied by NMR experiments. Based on these computational models, it has been suggested that binding of inhibitors in allosteric sites might affect protease flexibility and disrupt its function.