Foundations of biomolecular modeling.

The 2013 Nobel Prize in Chemistry has been awarded to Martin Karplus, Michael Levitt, and Arieh Warshel for "development of multiscale models for complex chemical systems." The honored work from the 1970s has provided a foundation for the widespread activities today in modeling organic and biomolecular systems.

Proof-of-concept studies with a computationally designed Mpro inhibitor as a synergistic combination regimen alternative to Paxlovid.

As the SARS-CoV-2 virus continues to spread and mutate, it remains important to focus not only on preventing spread through vaccination but also on treating infection with direct-acting antivirals (DAA). The approval of Paxlovid, a SARS-CoV-2 main protease (Mpro) DAA, has been significant for treatment of patients. A limitation of this DAA, however, is that the antiviral component, nirmatrelvir, is rapidly metabolized and requires inclusion of a CYP450 3A4 metabolic inhibitor, ritonavir, to boost levels of the active drug. Serious drug-drug interactions can occur with Paxlovid for patients who are also taking other medications metabolized by CYP4503A4, particularly transplant or otherwise immunocompromised patients who are most at risk for SARS-CoV-2 infection and the development of severe symptoms. Developing an alternative antiviral with improved pharmacological properties is critical for treatment of these patients. By using a computational and structure-guided approach, we were able to optimize a 100 to 250 µM screening hit to a potent nanomolar inhibitor and lead compound, Mpro61. In this study, we further evaluate Mpro61 as a lead compound, starting with examination of its mode of binding to SARS-CoV-2 Mpro. In vitro pharmacological profiling established a lack of off-target effects, particularly CYP450 3A4 inhibition, as well as potential for synergy with the currently approved alternate antiviral, molnupiravir. Development and subsequent testing of a capsule formulation for oral dosing of Mpro61 in B6-K18-hACE2 mice demonstrated favorable pharmacological properties, efficacy, and synergy with molnupiravir, and complete recovery from subsequent challenge by SARS-CoV-2, establishing Mpro61 as a promising potential preclinical candidate.

Chemistry of the Au-Thiol Interface through the Lens of Single-Molecule Flicker Noise Measurements.

Chemistry of the Au-S interface at the nanoscale is one of the most complex systems to study, as the nature and strength of the Au-S bond change under different experimental conditions. In this study, using mechanically controlled break junction technique, we probed the conductance and analyzed Flicker noise for several aliphatic and aromatic thiol derivatives and thioethers. We demonstrate that Flicker noise can be used to unambiguously differentiate between stronger chemisorption (Au-SR) and weaker physisorption (Au-SRR') type interactions. The Flicker noise measurements indicate that the gold rearrangement in chemisorbed Au-SR junctions resembles that of the Au rearrangement in pure Au-Au metal contact breaking, which is independent of the molecular backbone structure and the resulting conductance. In contrast, thioethers showed the formation of a weaker physisorbed Au-SRR' type bond, and the Flicker noise measurement indicates the changes in the Au-anchoring group interface but not the Au-Au rearrangement like that in the Au-SR case. Additionally, by employing single-molecular conductance and Flicker noise analysis, we have probed the interfacial electric field-catalyzed ring-opening reaction of cyclic thioether under mild environmental conditions, which otherwise requires harsh chemical conditions for cleavage of the C-S bond. All of our conductance measurements are complemented by NEGF transport calculations. This study illustrates that the single-molecule conductance, together with the Flicker noise measurements can be used to tune and monitor chemical reactions at the single-molecule level.

Enthalpies and entropies of hydration from Monte Carlo simulations.

The changes in free energy, enthalpy, and entropy for transfer of a solute from the gas phase into solution are the fundamental thermodynamic quantities that characterize the solvation process. Owing to the development of methods based on free-energy perturbation theory, computation of free energies of solvation has become routine in conjunction with Monte Carlo (MC) statistical mechanics and molecular dynamics (MD) simulations. Computation of the enthalpy change and by inference the entropy change is more challenging. Two methods are considered in this work corresponding to direct averaging for the solvent and solution and to computing the temperature derivative of the free energy in the van't Hoff approach. The application is for neutral organic solutes in TIP4P water using long MC simulations to improve precision. Definitive results are also provided for pure TIP4P water. While the uncertainty in computed free energies of hydration is ca. 0.05 kcal mol-1, it is ca. 0.4 kcal mol-1 for the enthalpy changes from either van't Hoff plots or the direct method with sampling for 5 billion MC configurations. Partial molar volumes of hydration are also computed by the direct method; they agree well with experimental data with an average deviation of 3 cm3 mol-1. In addition, the results permit breakdown of the errors in the free energy changes from the OPLS-AA force field into their enthalpic and entropic components. The excess hydrophobicity of organic solutes is enthalpic in origin.

Monte Carlo simulations for free energies of hydration: Past to present.

A summary of the development of Monte Carlo statistical mechanics simulations for the computation of free energies of hydration of organic molecules is followed by presentation of results with the latest version of the optimized potentials for liquid simulations-all atom force field and the TIP4P water model. Scaling of the Lennard-Jones interactions between water, oxygen, and carbon atoms by a factor of 1.25 is found to improve the accuracy of free energies of hydration for 50 prototypical organic molecules from a mean unsigned error of 1.0-1.2 to 0.4 kcal/mol.

Design, synthesis, and biological testing of biphenylmethyloxazole inhibitors targeting HIV-1 reverse transcriptase.

We report non-nucleoside inhibitors of HIV-1 reverse transcriptase (NNRTIs) using a biphenylmethyloxazole pharmacophore. A crystal structure of benzyloxazole 1 was obtained and suggested the potential viability of biphenyl analogues. In particular, 6a, 6b, and 7 turned out to be potent NNRTIs with low-nanomolar activity in enzyme inhibition and infected T-cell assays, and with low cytotoxicity. Though modeling further suggested that analogues with fluorosulfate and epoxide warheads might provide covalent modification of Tyr188, synthesis and testing did not find evidence for this outcome.

Assessing Metadynamics and Docking for Absolute Binding Free Energy Calculations Using Severe Acute Respiratory Syndrome Coronavirus 2 Main Protease Inhibitors.

Absolute binding free energy (ABFE) calculations can be an important part of the drug discovery process by identifying molecules that have the potential to be strong binders for a biomolecular target. Recent work has used free energy perturbation (FEP) theory for these calculations, focusing on a set of 16 inhibitors of the severe acute respiratory syndrome coronavirus 2 main protease (Mpro). Herein, the same data set is evaluated by metadynamics (MetaD), four different docking programs, and molecular mechanics with generalized Born and surface area solvation. MetaD yields a Kendall τ distance of 0.28 and Pearson r2 of 0.49, which reflect somewhat less accuracy than that from the ABFE FEP results. Notably, it is demonstrated that an ensemble docking protocol by which each ligand is docked into the 13 crystal structures in this data set provides improved performance, particularly when docking is carried out with Glide XP (Kendall τ distance = 0.20, Pearson r2 = 0.71), Glide SP (Kendall τ distance = 0.19, Pearson r2 = 0.66), or AutoDock 4 (Kendall τ distance = 0.21, Pearson r2 = 0.55). The best results are obtained with "superconsensus" docking by averaging the 52 results for each compound using the 4 docking protocols and all 13 crystal structures (Kendall τ distance = 0.18, Pearson r2 = 0.73).

Ensemble Geometric Deep Learning of Aqueous Solubility.

Geometric deep learning is one of the main workhorses for harnessing the power of big data to predict molecular properties such as aqueous solubility, which is key to the pharmacokinetic improvement of drug candidates. Two ensembles of graph neural network architectures were built, one based on spectral convolution and the other on spatial convolution. The pretrained models, denoted respectively as SolNet-GCN and SolNet-GAT, significantly outperformed the existing neural networks benchmarked on a validation set of 207 molecules. The SolNet-GCN model demonstrated the best performance on both the training and validation sets, with RMSE values of 0.53 and 0.72 log molar unit and Pearson r2 values of 0.95 and 0.75, respectively. Further, the ranking power of the SolNet models agreed well with a QM-based thermodynamic cycle approach at the PBE-vdW level of theory on a series of benzophenylurea derivatives and a series of benzodiazepine derivatives. Nevertheless, testing the resultant models on a set of inhibitors of the macrophage migration inhibitory factor (MIF) illustrated that the inclusion of atomic attributes to discriminate atoms with a higher tendency to form intermolecular hydrogen bonds in the crystalline state and to identify planar or nonplanar substructures can be beneficial for the prediction of aqueous solubility.

Computation of Absolute Binding Free Energies for Noncovalent Inhibitors with SARS-CoV-2 Main Protease.

Accurate, routine calculation of absolute binding free energies (ABFEs) for protein-ligand complexes remains a key goal of computer-aided drug design since it can enable screening and optimization of drug candidates. For development and testing of related methods, it is important to have high-quality datasets. To this end, from our own experimental studies, we have selected a set of 16 inhibitors of the SARS-CoV-2 main protease (Mpro) with structural diversity and well-distributed BFEs covering a 5 kcal/mol range. There is also minimal structural uncertainty since X-ray crystal structures have been deposited for 12 of the compounds. For methods testing, we report ABFE results from 2 µs molecular dynamics (MD) simulations using free energy perturbation (FEP) theory. The correlation of experimental and computed results is encouraging, with a Pearson's r2 of 0.58 and a Kendall τ of 0.24. The results indicate that current FEP-based ABFE calculations can be used for identification of active compounds (hits). While their accuracy for lead optimization is not yet sufficient, this activity remains addressable in separate lead series by relative BFE calculations.

Understanding Current Density in Molecules Using Molecular Orbitals.

While the use of molecular orbitals (MOs) and their isosurfaces to explain physical phenomena in chemical systems is a time-honored tool, we show that the nodes are an equally important component for understanding the current density through single-molecule junctions. We investigate three different model systems consisting of an alkane, alkene, and even [n]cumulene and show that we can explain the form of the current density using the MOs of the molecule. Essentially, the MOs define the region in which current can flow and their gradients define the direction in which current flows within that region. We also show that it is possible to simplify the current density for improved understanding by either partitioning the current density into more chemically intuitive parts, such as σ- and π-systems, or by filtering out MOs with negligible contributions to the overall current density. Our work highlights that it is possible to infer a non-equilibrium property (current density) given only equilibrium properties (MOs and their gradients), and this, in turn, grants deeper insight into coherent electron transport.

Visualizing and comparing quantum interference in the π-system and σ-system of organic molecules.

Quantum interference effects in conjugated molecules have been well-explored, with benzene frequently invoked as a pedagogical example. These interference effects have been understood through a quantum interference map in which the electronic transmission is separated into interfering and non-interfering terms, with a focus on the π-orbitals for conjugated molecules. Recently, saturated molecules have also been reported to exhibit destructive quantum interference effects; however, the very different σ-orbital character in these molecules means that it is not clear how orbital contributions manifest. Herein, we demonstrate that the quantum interference effects in conjugated molecules are quite different from those observed in saturated molecules, as demonstrated by the quantum interference map. While destructive interference at the Fermi energy in the π-system of benzene arises from interference terms between paired occupied and virtual orbitals, this is not the case at the Fermi energy in saturated systems. Instead, destructive interference is evident when contributions from a larger number of non-paired orbitals cancel, leading to more subtle and varied manifestations of destructive interference in saturated systems.

Trusting our machines: validating machine learning models for single-molecule transport experiments.

In this tutorial review, we will describe crucial aspects related to the application of machine learning to help users avoid the most common pitfalls. The examples we present will be based on data from the field of molecular electronics, specifically single-molecule electron transport experiments, but the concepts and problems we explore will be sufficiently general for application in other fields with similar data. In the first part of the tutorial review, we will introduce the field of single-molecule transport, and provide an overview of the most common machine learning algorithms employed. In the second part of the tutorial review, we will show, through examples grounded in single-molecule transport, that the promises of machine learning can only be fulfilled by careful application. We will end the tutorial review with a discussion of where we, as a field, could go from here.

Understanding In Vitro Pathways to Drug Discovery for TDP-43 Proteinopathies.

The use of cellular models is a common means to investigate the potency of therapeutics in pre-clinical drug discovery. However, there is currently no consensus on which model most accurately replicates key aspects of amyotrophic lateral sclerosis (ALS) and frontotemporal dementia (FTD) pathology, such as accumulation of insoluble, cytoplasmic transactive response DNA-binding protein (TDP-43) and the formation of insoluble stress granules. Given this, we characterised two TDP-43 proteinopathy cellular models that were based on different aetiologies of disease. The first was a sodium arsenite-induced chronic oxidative stress model and the second expressed a disease-relevant TDP-43 mutation (TDP-43 M337V). The sodium arsenite model displayed most aspects of TDP-43, stress granule and ubiquitin pathology seen in human ALS/FTD donor tissue, whereas the mutant cell line only modelled some aspects. When these two cellular models were exposed to small molecule chemical probes, different effects were observed across the two models. For example, a previously disclosed sulfonamide compound decreased cytoplasmic TDP-43 and increased soluble levels of stress granule marker TIA-1 in the cellular stress model without impacting these levels in the mutant cell line. This study highlights the challenges of using cellular models in lead development during drug discovery for ALS and FTD and reinforces the need to perform assessments of novel therapeutics across a variety of cell lines and aetiological models.

Quantification of the Helicality of Helical Molecular Orbitals.

The frontier molecular orbital (MO) topology of linear carbon molecules, such as polyynes, can be visually identified as helices. However, there is no clear way to quantify the helical curvature of these π-MOs, and it is thus challenging to quantify correlations between the helical curvature and molecular properties. In this paper, we develop a method that enables us to compute the helical curvature of MOs based on their nodal planes. Using this method, we define a robust way of quantifying the helical nature of MOs (helicality) by their deviation from a perfect helix. We explore several limiting cases, including polyynes, metallacumulenes, cyclic allenes, and spiroconjugated systems, where the change in helical curvature is subtle yet clearly highlighted with this method. For example, we show that strain only has a minor effect on the helicality of the frontier orbitals of cycloallenes and that the MOs of spiroconjugated systems are close to perfect helices around the spiro-carbon. Our work provides a well-defined method for assessing orbital helicality beyond visual inspection of MO isosurfaces, thus paving the way for future studies of how the helicality of π-MOs affects molecular properties.

Indoloxytriazines as binding molecules for the JAK2 JH2 pseudokinase domain and its V617F variant.

Small molecules that selectively bind to the pseudokinase JH2 domain over the JH1 kinase domain of JAK2 kinase are sought. Virtual screening led to the purchase of 17 compounds among which 9 were found to bind to V617F JAK2 JH2 with affinities of 40 - 300 µM in a fluorogenic assay. Ten analogues were then purchased yielding 9 additional active compounds. Aminoanilinyltriazine 22 was particularly notable as it shows no detectable binding to JAK2 JH1, and it has a 65-µM dissociation constant K d with V617F JAK2 JH2. A crystal structure for 22 in complex with wild-type JAK2 JH2 was obtained to elucidate the binding mode. Additional de novo design led to the synthesis of 19 analogues of 22 with the most potent being 33n with K d values of 2-3 µM for WT and V617F JAK2 JH2, and with 16-fold selectivity relative to binding with WT JAK2 JH1.

From in silico hit to long-acting late-stage preclinical candidate to combat HIV-1 infection.

The HIV-1 pandemic affecting over 37 million people worldwide continues, with nearly one-half of the infected population on highly active antiretroviral therapy (HAART). Major therapeutic challenges remain because of the emergence of drug-resistant HIV-1 strains, limitations because of safety and toxicity with current HIV-1 drugs, and patient compliance for lifelong, daily treatment regimens. Nonnucleoside reverse transcriptase inhibitors (NNRTIs) that target the viral polymerase have been a key component of the current HIV-1 combination drug regimens; however, these issues hamper them. Thus, the development of novel more effective NNRTIs as anti-HIV-1 agents with fewer long-term liabilities, efficacy on new drug-resistant HIV-1 strains, and less frequent dosing is crucial. Using a computational and structure-based design strategy to guide lead optimization, a 5 µM virtual screening hit was transformed to a series of very potent nanomolar to picomolar catechol diethers. One representative, compound I, was shown to have nanomolar activity in HIV-1-infected T cells, potency on clinically relevant HIV-1 drug-resistant strains, lack of cytotoxicity and off-target effects, and excellent in vivo pharmacokinetic behavior. In this report, we show the feasibility of compound I as a late-stage preclinical candidate by establishing synergistic antiviral activity with existing HIV-1 drugs and clinical candidates and efficacy in HIV-1-infected humanized [human peripheral blood lymphocyte (Hu-PBL)] mice by completely suppressing viral loads and preventing human CD4+ T-cell loss. Moreover, a long-acting nanoformulation of compound I [compound I nanoparticle (compound I-NP)] in poly(lactide-coglycolide) (PLGA) was developed that shows sustained maintenance of plasma drug concentrations and drug efficacy for almost 3 weeks after a single dose.

Targeting the TS dimer interface in bifunctional Cryptosporidium hominis TS-DHFR from parasitic protozoa: Virtual screening identifies novel TS allosteric inhibitors.

Effective therapies are lacking to treat gastrointestinal infections caused by the genus Cryptosporidium, which can be fatal in the immunocompromised. One target of interest is Cryptosporidium hominis (C. hominis) thymidylate synthase-dihydrofolate reductase (ChTS-DHFR), a bifunctional enzyme necessary for DNA biosynthesis. Targeting the TS-TS dimer interface is a novel strategy previously used to identify inhibitors against the related bifunctional enzyme in Toxoplasma gondii. In the present study, we target the ChTS dimer interface through homology modeling and high-throughput virtual screening to identifying allosteric, ChTS-specific inhibitors. Our work led to the discovery of methylenedioxyphenyl-aminophenoxypropanol analogues which inhibit ChTS activity in a manner that is both dose-dependent and influenced by the conformation of the enzyme. Preliminary results presented here include an analysis of structure activity relationships and a ChTS-apo crystal structure of ChTS-DHFR supporting the continued development of inhibitors that stabilize a novel pocket formed in the open conformation of ChTS-TS.

Metadynamics as a Postprocessing Method for Virtual Screening with Application to the Pseudokinase Domain of JAK2.

With standard scoring methods, top-ranked compounds from virtual screening by docking often turn out to be inactive. For this reason, metadynamics, a method used to sample rare events, was studied to further evaluate docking poses with the aim of reducing false positives. Specifically, virtual screening was performed with Glide SP to seek potential molecules to bind to the ATP site in the pseudokinase domain of JAK2 kinase, and promising compounds were selected from the top-ranked 1000 based on visualization. Rescoring with Glide XP, GOLD, and MM/GBSA was unable to differentiate well between active and inactive compounds. Metadynamics was then used to gauge the relative binding affinity from the required time or the potential of mean force needed to dissociate the ligand from the bound complex. With consideration of previously known binders of varying affinities, metadynamics was able to differentiate between the most active compounds and inactive or weakly active ones, and it could identify correctly most of the selected virtual screening compounds as false positives. Thus, metadynamics has the potential to be a viable postprocessing method for virtual screening, minimizing the expense of buying or synthesizing inactive compounds.

Covalent inhibitors for eradication of drug-resistant HIV-1 reverse transcriptase: From design to protein crystallography.

Development of resistance remains a major challenge for drugs to treat HIV-1 infections, including those targeting the essential viral polymerase, HIV-1 reverse transcriptase (RT). Resistance associated with the Tyr181Cys mutation in HIV-1 RT has been a key roadblock in the discovery of nonnucleoside RT inhibitors (NNRTIs). It is the principal point mutation that arises from treatment of HIV-infected patients with nevirapine, the first-in-class drug still widely used, especially in developing countries. We report covalent inhibitors of Tyr181Cys RT (CRTIs) that can completely knock out activity of the resistant mutant and of the particularly challenging Lys103Asn/Tyr181Cys variant. Conclusive evidence for the covalent modification of Cys181 is provided from enzyme inhibition kinetics, mass spectrometry, protein crystallography, and antiviral activity in infected human T-cell assays. The CRTIs are also shown to be selective for Cys181 and have lower cytotoxicity than the approved NNRTI drugs efavirenz and rilpivirine.

Novel allosteric covalent inhibitors of bifunctional Cryptosporidium hominis TS-DHFR from parasitic protozoa identified by virtual screening.

Protozoans of the genus Cryptosporidium are the causative agent of the gastrointestinal disease, cryptosporidiosis, which can be fatal in immunocompromised individuals. Cryptosporidium hominis (C. hominis) bifunctional thymidylate synthase-dihydrofolate reductase (TS-DHFR) is an essential enzyme in the folate biosynthesis pathway and a molecular target for inhibitor design. Previous studies have demonstrated the importance of the ChTS-DHFR linker region "crossover helix" to the enzymatic activity and stability of the ChDHFR domain. We conducted a virtual screen of a novel non-active site pocket located at the interface of the ChDHFR domain and crossover helix. From this screen we have identified and characterized a noncompetitive inhibitor, compound 15, a substituted diphenyl thiourea. Through subsequent structure activity relationship studies, we have identified a time-dependent inhibitor lead, compound 15D17, a thiol-substituted 2-hydroxy-N-phenylbenzamide, which is selective for ChTS-DHFR, and whose effects appear to be mediated by covalent bond formation with a non-catalytic cysteine residue adjacent to the non-active site pocket.