6-Azaspiro[2.5]octanes as small molecule agonists of the human glucagon-like peptide-1 receptor.

Activation of the glucagon-like peptide-1 (GLP-1) receptor stimulates insulin release, lowers plasma glucose levels, delays gastric emptying, increases satiety, suppresses food intake, and affords weight loss in humans. These beneficial attributes have made peptide-based agonists valuable tools for the treatment of type 2 diabetes mellitus and obesity. However, efficient, and consistent delivery of peptide agents generally requires subcutaneous injection, which can reduce patient utilization. Traditional orally absorbed small molecules for this target may offer improved patient compliance as well as the opportunity for co-formulation with other oral therapeutics. Herein, we describe an SAR investigation leading to small-molecule GLP-1 receptor agonists that represent a series that parallels the recently reported clinical candidate danuglipron. In the event, identification of a benzyloxypyrimidine lead, using a sensitized high-throughput GLP-1 agonist assay, was followed by optimization of the SAR using substituent modifications analogous to those discovered in the danuglipron series. A new series of 6-azaspiro[2.5]octane molecules was optimized into potent GLP-1 agonists. Information gleaned from cryogenic electron microscope structures was used to rationalize the SAR of the optimized compounds.

TorsionNet: A Deep Neural Network to Rapidly Predict Small-Molecule Torsional Energy Profiles with the Accuracy of Quantum Mechanics.

Fast and accurate assessment of small-molecule dihedral energetics is crucial for molecular design and optimization in medicinal chemistry. Yet, accurate prediction of torsion energy profiles remains challenging as the current molecular mechanics (MM) methods are limited by insufficient coverage of drug-like chemical space and accurate quantum mechanical (QM) methods are too expensive. To address this limitation, we introduce TorsionNet, a deep neural network (DNN) model specifically developed to predict small-molecule torsion energy profiles with QM-level accuracy. We applied active learning to identify nearly 50k fragments (with elements H, C, N, O, F, S, and Cl) that maximized the coverage of our corporate compound library and leveraged massively parallel cloud computing resources for density functional theory (DFT) torsion scans of these fragments, generating a training data set of 1.2 million DFT energies. After training TorsionNet on this data set, we obtain a model that can rapidly predict the torsion energy profile of typical drug-like fragments with DFT-level accuracy. Importantly, our method also provides an uncertainty estimate for the predicted profiles without any additional calculations. In this report, we show that TorsionNet can accurately identify the preferred dihedral geometries observed in crystal structures. Our TorsionNet-based analysis of a diverse set of protein-ligand complexes with measured binding affinity shows a strong association between high ligand strain and low potency. We also present practical applications of TorsionNet that demonstrate how consideration of DNN-based strain energy leads to substantial improvement in existing lead discovery and design workflows. TorsionNet500, a benchmark data set comprising 500 chemically diverse fragments with DFT torsion profiles (12k MM- and DFT-optimized geometries and energies), has been created and is made publicly available.

Are 2D fingerprints still valuable for drug discovery?

Recently, molecular fingerprints extracted from three-dimensional (3D) structures using advanced mathematics, such as algebraic topology, differential geometry, and graph theory have been paired with efficient machine learning, especially deep learning algorithms to outperform other methods in drug discovery applications and competitions. This raises the question of whether classical 2D fingerprints are still valuable in computer-aided drug discovery. This work considers 23 datasets associated with four typical problems, namely protein-ligand binding, toxicity, solubility and partition coefficient to assess the performance of eight 2D fingerprints. Advanced machine learning algorithms including random forest, gradient boosted decision tree, single-task deep neural network and multitask deep neural network are employed to construct efficient 2D-fingerprint based models. Additionally, appropriate consensus models are built to further enhance the performance of 2D-fingerprint-based methods. It is demonstrated that 2D-fingerprint-based models perform as well as the state-of-the-art 3D structure-based models for the predictions of toxicity, solubility, partition coefficient and protein-ligand binding affinity based on only ligand information. However, 3D structure-based models outperform 2D fingerprint-based methods in complex-based protein-ligand binding affinity predictions.

Comprehensive Assessment of Torsional Strain in Crystal Structures of Small Molecules and Protein-Ligand Complexes using ab Initio Calculations.

The energetics of rotation around single bonds (torsions) is a key determinant of the three-dimensional shape that druglike molecules adopt in solution, the solid state, and in different biological environments, which in turn defines their unique physical and pharmacological properties. Therefore, accurate characterization of torsion angle preference and energetics is essential for the success of computational drug discovery and design. Here, we analyze torsional strain in crystal structures of druglike molecules in Cambridge structure database (CSD) and bioactive ligand conformations in protein data bank (PDB), expressing the total strain energy as a sum of strain energy from constituent rotatable bonds. We utilized cloud computing to generate torsion scan profiles of a very large collection of chemically diverse neutral fragments at DFT(B3LYP)/6-31G*//6-31G** or DFT(B3LYP)/6-31+G*//6-31+G** (for sulfur-containing molecule). With the data generated from these ab initio calculations, we performed rigorous analysis of strain due to deviation of observed torsion angles relative to their ideal gas-phase geometries. Contrary to the previous studies based on molecular mechanics, we find that in the crystalline state, molecules generally adopt low-strain conformations, with median per-torsion strain energy in CSD and PDB under one-tenth and one-third of a kcal/mol, respectively. However, for a small fraction (<5%) of motifs, external effects such as steric hindrance and hydrogen bonds result in strain penalty exceeding 2.5 kcal/mol. We find that due to poor quality of PDB structures in general, bioactive structures tend to have higher torsional strain compared to small-molecule crystal conformations. However, in the absence of structural fitting artifacts in PDB structures, protein-induced strain in bioactive conformations is quantitatively similar to those due to the packing forces in small-molecule crystal structures. This analysis allows us to establish strain energy thresholds to help identify biologically relevant conformers in a given ensemble. The work presented here is the most comprehensive study to date that demonstrates the utility and feasibility of gas-phase quantum mechanics (QM) calculations to study conformational preference and energetics of drug-size molecules. Potential applications of this study in computational lead discovery and structure-based design are discussed.

Comparative pharmacokinetic profile of cyclosporine (CsA) with a decapeptide and a linear analogue.

The synthesis and in vivo pharmacokinetic profile of an analogue of cyclosporine is disclosed. An acyclic congener was also profiled in in vitro assays to compare cell permeability. The compounds possess similar calculated and measured molecular descriptors however have different behaviors in an RRCK assay to assess cell permeability.

A potentiator of orthosteric ligand activity at GLP-1R acts via covalent modification.

We report that 4-(3-(benzyloxy)phenyl)-2-ethylsulfinyl-6-(trifluoromethyl)pyrimidine (BETP), which behaves as a positive allosteric modulator at the glucagon-like peptide-1 receptor (GLP-1R), covalently modifies cysteines 347 and 438 in GLP-1R. C347, located in intracellular loop 3 of GLP-1R, is critical to the activity of BETP and a structurally distinct GLP-1R ago-allosteric modulator, N-(tert-butyl)-6,7-dichloro-3-(methylsulfonyl)quinoxalin-2-amine. We further show that substitution of cysteine for phenylalanine 345 in the glucagon receptor is sufficient to confer sensitivity to BETP.

Discovery of spirocyclic-diamine inhibitors of mammalian acetyl CoA-carboxylase.

A novel series of spirocyclic-diamine based, isoform non-selective inhibitors of acetyl-CoA carboxylase (ACC) is described. These spirodiamine derivatives were discovered by design of a library to mimic the structural rigidity and hydrogen-bonding pattern observed in the co-crystal structure of spirochromanone inhibitor I. The lead compound 3.5.1 inhibited de novo lipogenesis in rat hepatocytes, with an IC50 of 0.30 µM.

Development and Comprehensive Benchmark of a High-Quality AMBER-Consistent Small Molecule Force Field with Broad Chemical Space Coverage for Molecular Modeling and Free Energy Calculation.

Biomolecular simulations have become an essential tool in contemporary drug discovery, and molecular mechanics force fields (FFs) constitute its cornerstone. Developing a high quality and broad coverage general FF is a significant undertaking that requires substantial expert knowledge and computing resources, which is beyond the scope of general practitioners. Existing FFs originate from only a limited number of groups and organizations, and they either suffer from limited numbers of training sets, lower than desired quality because of oversimplified representations, or are costly for the molecular modeling community to access. To address these issues, in this work, we developed an AMBER-consistent small molecule FF with extensive chemical space coverage, and we provide Open Access parameters for the entire modeling community. To validate our FF, we carried out benchmarks of quantum mechanics (QM)/molecular mechanics conformer comparison and free energy perturbation calculations on several benchmark data sets. Our FF achieves a higher level of performance at reproducing QM energies and geometries than two popular open-source FFs, OpenFF2 and GAFF2. In relative binding free energy calculations for 31 protein-ligand data sets, comprising 1079 pairs of ligands, the new FF achieves an overall root-mean-square error of 1.19 kcal/mol for ΔΔG and 0.92 kcal/mol for ΔG on a subset of 463 ligands without bespoke fitting to the data sets. The results are on par with those of the leading commercial series of OPLS FFs.

On-resin N-methylation of cyclic peptides for discovery of orally bioavailable scaffolds.

Backbone N-methylation is common among peptide natural products and has a substantial impact on both the physical properties and the conformational states of cyclic peptides. However, the specific impact of N-methylation on passive membrane diffusion in cyclic peptides has not been investigated systematically. Here we report a method for the selective, on-resin N-methylation of cyclic peptides to generate compounds with drug-like membrane permeability and oral bioavailability. The selectivity and degree of N-methylation of the cyclic peptide was dependent on backbone stereochemistry, suggesting that conformation dictates the regiochemistry of the N-methylation reaction. The permeabilities of the N-methyl variants were corroborated by computational studies on a 1,024-member virtual library of N-methyl cyclic peptides. One of the most permeable compounds, a cyclic hexapeptide (molecular mass = 755 Da) with three N-methyl groups, showed an oral bioavailability of 28% in rat.

Broadening the Scope of Binding Free Energy Calculations Using a Separated Topologies Approach.

Binding free energy calculations predict the potency of compounds to protein binding sites in a physically rigorous manner and see broad application in prioritizing the synthesis of novel drug candidates. Relative binding free energy (RBFE) calculations have emerged as an industry-standard approach to achieve highly accurate rank-order predictions of the potency of related compounds; however, this approach requires that the ligands share a common scaffold and a common binding mode, restricting the methods' domain of applicability. This is a critical limitation since complex modifications to the ligands, especially core hopping, are very common in drug design. Absolute binding free energy (ABFE) calculations are an alternate method that can be used for ligands that are not congeneric. However, ABFE suffers from a known problem of long convergence times due to the need to sample additional degrees of freedom within each system, such as sampling rearrangements necessary to open and close the binding site. Here, we report on an alternative method for RBFE, called Separated Topologies (SepTop), which overcomes the issues in both of the aforementioned methods by enabling large scaffold changes between ligands with a convergence time comparable to traditional RBFE. Instead of only mutating atoms that vary between two ligands, this approach performs two absolute free energy calculations at the same time in opposite directions, one for each ligand. Defining the two ligands independently allows the comparison of the binding of diverse ligands without the artificial constraints of identical poses or a suitable atom-atom mapping. This approach also avoids the need to sample the unbound state of the protein, making it more efficient than absolute binding free energy calculations. Here, we introduce an implementation of SepTop. We developed a general and efficient protocol for running SepTop, and we demonstrated the method on four diverse, pharmaceutically relevant systems. We report the performance of the method, as well as our practical insights into the strengths, weaknesses, and challenges of applying this method in an industrial drug design setting. We find that the accuracy of the approach is sufficiently high to rank order ligands with an accuracy comparable to traditional RBFE calculations while maintaining the additional flexibility of SepTop.

Biaryl-bridged macrocyclic peptides: conformational constraint via carbogenic fusion of natural amino acid side chains.

A general method for constraining peptide conformations via linkage of aromatic sidechains has been developed. Macrocyclization of suitably functionalized tri-, tetra- and pentapeptides via Suzuki-Miyaura cross-coupling has been used to generate side chain to side chain, biaryl-bridged 14- to 21-membered macrocyclic peptides. Biaryl bridges possessing three different configurations, meta-meta, meta-ortho, and ortho-meta, were systematically explored through regiochemical variation of the aryl halide and aryl boronate coupling partners, allowing fine-tuning of the resultant macrocycle conformation. Suzuki-Miyaura macrocyclizations were successfully achieved both in solution and on solid phase for all three sizes of peptide. This approach constitutes a means of constraining peptide conformation via direct carbogenic fusion of side chains of naturally occurring amino acids such as phenylalanine and tyrosine, and so is complementary to strategies involving non-natural, for example, hydrocarbon, bridges.

Use of 3D properties to characterize beyond rule-of-5 property space for passive permeation.

The application of conformationally dependent measures of size and polarity to characterize beyond rule-of-5 (Ro5) space for passive permeation was investigated. Specifically, radius of gyration, an alternative to molecular weight, and three-dimensional polar surface area and the generalized Born/surface area dehydration free energy, alternatives to hydrogen-bond donor and acceptor counts, were computed on models of the permeating conformations of over 35 000 molecules. The resulting guidelines for size and polarity, described by the 3D properties, should aid the design of Ro5 violators with passive permeability.

Exploring aromatic chemical space with NEAT: novel and electronically equivalent aromatic template.

In this paper, we describe a lead transformation tool, NEAT (Novel and Electronically equivalent Aromatic Template), which can help identify novel aromatic rings that are estimated to have similar electrostatic potentials, dipoles, and hydrogen bonding capabilities to a query template; hence, they may offer similar bioactivity profiles. In this work, we built a comprehensive heteroaryl database, and precalculated high-level quantum mechanical (QM) properties, including electrostatic potential charges, hydrogen bonding ability, dipole moments, chemical reactivity, and othe properties. NEAT bioisosteric similarities are based on the electrostatic potential surface calculated by Brood, using the precalculated QM ESP charges and other QM properties. Compared with existing commercial lead transformation software, (1) NEAT is the only one that covers the comprehensive heteroaryl chemical space, and (2) NEAT offers a better characterization of novel aryl cores by using high-evel QM properties that are relevant to molecular interactions. NEAT provides unique value to medicinal chemists quickly exploring the largely uncharted aromatic chemical space, and one successful example of its application is discussed herein.

A Small-Molecule Oral Agonist of the Human Glucagon-like Peptide-1 Receptor.

Peptide agonists of the glucagon-like peptide-1 receptor (GLP-1R) have revolutionized diabetes therapy, but their use has been limited because they require injection. Herein, we describe the discovery of the orally bioavailable, small-molecule, GLP-1R agonist PF-06882961 (danuglipron). A sensitized high-throughput screen was used to identify 5-fluoropyrimidine-based GLP-1R agonists that were optimized to promote endogenous GLP-1R signaling with nanomolar potency. Incorporation of a carboxylic acid moiety provided considerable GLP-1R potency gains with improved off-target pharmacology and reduced metabolic clearance, ultimately resulting in the identification of danuglipron. Danuglipron increased insulin levels in primates but not rodents, which was explained by receptor mutagensis studies and a cryogenic electron microscope structure that revealed a binding pocket requiring a primate-specific tryptophan 33 residue. Oral administration of danuglipron to healthy humans produced dose-proportional increases in systemic exposure (NCT03309241). This opens an opportunity for oral small-molecule therapies that target the well-validated GLP-1R for metabolic health.

Addressing limitations with the MM-GB/SA scoring procedure using the WaterMap method and free energy perturbation calculations.

The MM-GB/SA scoring technique has become an important computational approach in drug design. We, and others, have demonstrated that for congeneric molecules the correlation with experimental data obtained with the physics-based scoring is usually superior to scoring functions from typical docking algorithms. Despite showing good accuracy when applied within a series, much work is necessary to improve the MM-GB/SA method in order to gain greater efficiency in drug design. Here, we investigate the poor estimation of protein desolvation provided by the GB/SA solvation model and the large dynamic range observed in the MM-GB/SA scoring compared to that of the experimental data. In the former, replacing the GB/SA protein desolvation in the MM-GB/SA method by the free energy associated with displacing binding site waters upon ligand binding estimated by WaterMap provides the best results when ranking congeneric series of factor Xa and cyclin-dependent kinase 2 (CDK2) inhibitors. However, the improvement is modest over results obtained with the MM-GB/SA and WaterMap methods individually, apparently due to the high correlation between the free energy liberation of the displaced solvent and the protein-ligand van der Waals interactions, which in turn may be interpretable as estimates of the hydrophobic effect and hydrophobic-like interactions, respectively. As for the large dynamic range, comparisons between MM-GB/SA and FEP calculations indicate that for the factor Xa test set this problem has its origin in the lack of shielding effects of protein--ligand electrostatic interactions; that overly favors ligands that engage in hydrogen bonds with the protein.

Design Principles for Intestinal Permeability of Cyclic Peptides.

One of the most exciting facets of cyclic peptides is that they have the potential to be orally bioavailable, despite having physical properties well beyond the traditional "Rule-of-5" chemistry space (Lipinski et al., Adv Drug Deliv Rev. 23(1): 3-25, 1997). An important component of meeting this challenge is to design cyclic peptides with good intestinal permeability. Here we discuss the design principles for intestinal permeability that have been developed in recent year. These principles can be subdivided into three regimes: physical property guidelines, design strategies for the macrocyclic ring, and design strategies for side chains. The most important overall aims are to minimize solvent-exposed polarity while keeping size, flexibility, and lipophilicity within favorable ranges, thereby allowing peptide chemists to achieve intestinal permeability in addition to other important properties for their compounds, such as solubility and binding affinity. Here we describe a variety of design strategies that have been developed to help peptide chemists in this endeavor.

Nonclassical Size Dependence of Permeation Defines Bounds for Passive Adsorption of Large Drug Molecules.

Macrocyclic peptides are considered large enough to inhibit "undruggable" targets, but the design of passively cell-permeable molecules in this space remains a challenge due to the poorly understood role of molecular size on passive membrane permeability. Using split-pool combinatorial synthesis, we constructed a library of cyclic, per-N-methlyated peptides spanning a wide range of calculated lipohilicities (0 < AlogP < 8) and molecular weights (∼800 Da < MW < ∼1200 Da). Analysis by the parallel artificial membrane permeability assay revealed a steep drop-off in apparent passive permeability with increasing size in stark disagreement with current permeation models. This observation, corroborated by a set of natural products, helps define criteria for achieving permeability in larger molecular size regimes and suggests an operational cutoff, beyond which passive permeability is constrained by a sharply increasing penalty on membrane permeation.

Discovery of Potent and Orally Bioavailable Macrocyclic Peptide-Peptoid Hybrid CXCR7 Modulators.

The chemokine receptor CXCR7 is an attractive target for a variety of diseases. While several small-molecule modulators of CXCR7 have been reported, peptidic macrocycles may provide advantages in terms of potency, selectivity, and reduced off-target activity. We produced a series of peptidic macrocycles that incorporate an N-linked peptoid functionality where the peptoid group enabled us to explore side-chain diversity well beyond that of natural amino acids. At the same time, theoretical calculations and experimental assays were used to track and reduce the polarity while closely monitoring the physicochemical properties. This strategy led to the discovery of macrocyclic peptide-peptoid hybrids with high CXCR7 binding affinities (Ki < 100 nM) and measurable passive permeability (Papp > 5 × 10-6 cm/s). Moreover, bioactive peptide 25 (Ki = 9 nM) achieved oral bioavailability of 18% in rats, which was commensurate with the observed plasma clearance values upon intravenous administration.

Probing the Physicochemical Boundaries of Cell Permeability and Oral Bioavailability in Lipophilic Macrocycles Inspired by Natural Products.

Cyclic peptide natural products contain a variety of conserved, nonproteinogenic structural elements such as d-amino acids and amide N-methylation. In addition, many cyclic peptides incorporate γ-amino acids and other elements derived from polyketide synthases. We hypothesized that the position and orientation of these extended backbone elements impact the ADME properties of these hybrid molecules, especially their ability to cross cell membranes and avoid metabolic degradation. Here we report the synthesis of cyclic hexapeptide diastereomers containing γ-amino acids (e.g., statines) and systematically investigate their structure-permeability relationships. These compounds were much more water-soluble and, in many cases, were both more membrane permeable and more stable to liver microsomes than a similar non-statine-containing derivative. Permeability correlated well with the extent of intramolecular hydrogen bonding observed in the solution structures determined in the low-dielectric solvent CDCl3, and one compound showed an oral bioavailability of 21% in rat. Thus, the incorporation of γ-amino acids offers a route to increase backbone diversity and improve ADME properties in cyclic peptide scaffolds.

Discovery of an in Vivo Tool to Establish Proof-of-Concept for MAP4K4-Based Antidiabetic Treatment.

Recent studies in adipose tissue, pancreas, muscle, and macrophages suggest that MAP4K4, a serine/threonine protein kinase may be a viable target for antidiabetic drugs. As part of the evaluation of MAP4K4 as a novel antidiabetic target, a tool compound, 16 (PF-6260933) and a lead 17 possessing excellent kinome selectivity and suitable properties were delivered to establish proof of concept in vivo. The medicinal chemistry effort that led to the discovery of these lead compounds is described herein together with in vivo pharmacokinetic properties and activity in a model of insulin resistance.