Does a Machine-Learned Potential Perform Better Than an Optimally Tuned Traditional Force Field? A Case Study on Fluorohydrins.

We present a comparative study that evaluates the performance of a machine learning potential (ANI-2x), a conventional force field (GAFF), and an optimally tuned GAFF-like force field in the modeling of a set of 10 γ-fluorohydrins that exhibit a complex interplay between intra- and intermolecular interactions in determining conformer stability. To benchmark the performance of each molecular model, we evaluated their energetic, geometric, and sampling accuracies relative to quantum-mechanical data. This benchmark involved conformational analysis both in the gas phase and chloroform solution. We also assessed the performance of the aforementioned molecular models in estimating nuclear spin-spin coupling constants by comparing their predictions to experimental data available in chloroform. The results and discussion presented in this study demonstrate that ANI-2x tends to predict stronger-than-expected hydrogen bonding and overstabilize global minima and shows problems related to inadequate description of dispersion interactions. Furthermore, while ANI-2x is a viable model for modeling in the gas phase, conventional force fields still play an important role, especially for condensed-phase simulations. Overall, this study highlights the strengths and weaknesses of each model, providing guidelines for the use and future development of force fields and machine learning potentials.

A knowledge-based, structural-aided discovery of a novel class of 2-phenylimidazo[1,2-a]pyridine-6-carboxamide H-PGDS inhibitors.

Through an internal virtual screen at GlaxoSmithKline a distinct class of 2-phenylimidazo[1,2-a]pyridine-6-carboxamide H-PGDS inhibitors were discovered. Careful evaluation of crystal structures and SAR led to a novel, potent, and orally active imidazopyridine inhibitor of H-PGDS, 20b. Herein, describes the identification of 2 classes of inhibitors, their syntheses, and their challenges.

ParaMol: A Package for Automatic Parameterization of Molecular Mechanics Force Fields.

The ensemble of structures generated by molecular mechanics (MM) simulations is determined by the functional form of the force field employed and its parameterization. For a given functional form, the quality of the parameterization is crucial and will determine how accurately we can compute observable properties from simulations. While accurate force field parameterizations are available for biomolecules, such as proteins or DNA, the parameterization of new molecules, such as drug candidates, is particularly challenging as these may involve functional groups and interactions for which accurate parameters may not be available. Here, in an effort to address this problem, we present ParaMol, a Python package that has a special focus on the parameterization of bonded and nonbonded terms of druglike molecules by fitting to ab initio data. We demonstrate the software by deriving bonded terms' parameters of three widely known drug molecules, viz. aspirin, caffeine, and a norfloxacin analogue, for which we show that, within the constraints of the functional form, the methodologies implemented in ParaMol are able to derive near-ideal parameters. Additionally, we illustrate the best practices to follow when employing specific parameterization routes. We also determine the sensitivity of different fitting data sets, such as relaxed dihedral scans and configurational ensembles, to the parameterization procedure, and discuss the features of the various weighting methods available to weight configurations. Owing to ParaMol's capabilities, we propose that this software can be introduced as a routine step in the protocol normally employed to parameterize druglike molecules for MM simulations.

The exploration of aza-quinolines as hematopoietic prostaglandin D synthase (H-PGDS) inhibitors with low brain exposure.

GlaxoSmithKline and Astex Pharmaceuticals recently disclosed the discovery of the potent H-PGDS inhibitor GSK2894631A 1a (IC50 = 9.9 nM) as part of a fragment-based drug discovery collaboration with Astex Pharmaceuticals. This molecule exhibited good murine pharmacokinetics, allowing it to be utilized to explore H-PGDS pharmacology in vivo. Yet, with prolonged dosing at higher concentrations, 1a induced CNS toxicity. Looking to attenuate brain penetration in this series, aza-quinolines, were prepared with the intent of increasing polar surface area. Nitrogen substitutions at the 6- and 8-positions of the quinoline were discovered to be tolerated by the enzyme. Subsequent structure activity studies in these aza-quinoline scaffolds led to the identification of 1,8-naphthyridine 1y (IC50 = 9.4 nM) as a potent peripherally restricted H-PGDS inhibitor. Compound 1y is efficacious in four in vivo inflammatory models and exhibits no CNS toxicity.

The discovery of quinoline-3-carboxamides as hematopoietic prostaglandin D synthase (H-PGDS) inhibitors.

With the goal of discovering more selective anti-inflammatory drugs, than COX inhibitors, to attenuate prostaglandin signaling, a fragment-based screen of hematopoietic prostaglandin D synthase was performed. The 76 crystallographic hits were sorted into similar groups, with the 3-cyano-quinoline 1a (FP IC50 = 220,000 nM, LE = 0.43) being a potent member of the 6,6-fused heterocyclic cluster. Employing SAR insights gained from structural comparisons of other H-PGDS fragment binding mode clusters, the initial hit 1a was converted into the 70-fold more potent quinoline 1d (IC50 = 3,100 nM, LE = 0.49). A systematic substitution of the amine moiety of 1d, utilizing structural information and array chemistry, with modifications to improve inhibitor stability, resulted in the identification of the 300-fold more active H-PGDS inhibitor tool compound 1bv (IC50 = 9.9 nM, LE = 0.42). This selective inhibitor exhibited good murine pharmacokinetics, dose-dependently attenuated PGD2 production in a mast cell degranulation assay and should be suitable to further explore H-PGDS biology.

Fragment-Based Discovery of Allosteric Inhibitors of SH2 Domain-Containing Protein Tyrosine Phosphatase-2 (SHP2).

The ubiquitously expressed protein tyrosine phosphatase SHP2 is required for signaling downstream of receptor tyrosine kinases (RTKs) and plays a role in regulating many cellular processes. Genetic knockdown and pharmacological inhibition of SHP2 suppresses RAS/MAPK signaling and inhibit the proliferation of RTK-driven cancer cell lines. Here, we describe the first reported fragment-to-lead campaign against SHP2, where X-ray crystallography and biophysical techniques were used to identify fragments binding to multiple sites on SHP2. Structure-guided optimization, including several computational methods, led to the discovery of two structurally distinct series of SHP2 inhibitors binding to the previously reported allosteric tunnel binding site (Tunnel Site). One of these series was advanced to a low-nanomolar lead that inhibited tumor growth when dosed orally to mice bearing HCC827 xenografts. Furthermore, a third series of SHP2 inhibitors was discovered binding to a previously unreported site, lying at the interface of the C-terminal SH2 and catalytic domains.

Assessing the lipophilicity of fragments and early hits.

A key challenge in many drug discovery programs is to accurately assess the potential value of screening hits. This is particularly true in fragment-based drug design (FBDD), where the hits often bind relatively weakly, but are correspondingly small. Ligand efficiency (LE) considers both the potency and the size of the molecule, and enables us to estimate whether or not an initial hit is likely to be optimisable to a potent, druglike lead. While size is a key property that needs to be controlled in a small molecule drug, there are a number of additional properties that should also be considered. Lipophilicity is amongst the most important of these additional properties, and here we present a new efficiency index (LLE(AT)) that combines lipophilicity, size and potency. The index is intuitively defined, and has been designed to have the same target value and dynamic range as LE, making it easily interpretable by medicinal chemists. Monitoring both LE and LLE(AT) should help both in the selection of more promising fragment hits, and controlling molecular weight and lipophilicity during optimisation.

Generation of Quantum Configurational Ensembles Using Approximate Potentials.

Conformational analysis is of paramount importance in drug design: it is crucial to determine pharmacological properties, understand molecular recognition processes, and characterize the conformations of ligands when unbound. Molecular Mechanics (MM) simulation methods, such as Monte Carlo (MC) and molecular dynamics (MD), are usually employed to generate ensembles of structures due to their ability to extensively sample the conformational space of molecules. The accuracy of these MM-based schemes strongly depends on the functional form of the force field (FF) and its parametrization, components that often hinder their performance. High-level methods, such as ab initio MD, provide reliable structural information but are still too computationally expensive to allow for extensive sampling. Therefore, to overcome these limitations, we present a multilevel MC method that is capable of generating quantum configurational ensembles while keeping the computational cost at a minimum. We show that FF reparametrization is an efficient route to generate FFs that reproduce QM results more closely, which, in turn, can be used as low-cost models to achieve the gold standard QM accuracy. We demonstrate that the MC acceptance rate is strongly correlated with various phase space overlap measurements and that it constitutes a robust metric to evaluate the similarity between the MM and QM levels of theory. As a more advanced application, we present a self-parametrizing version of the algorithm, which combines sampling and FF parametrization in one scheme, and apply the methodology to generate the QM/MM distribution of a ligand in aqueous solution.

C-H functionalisation tolerant to polar groups could transform fragment-based drug discovery (FBDD).

We have analysed 131 fragment-to-lead (F2L) examples targeting a wide variety of protein families published by academic and industrial laboratories between 2015-2019. Our assessment of X-ray structural data identifies the most common polar functional groups involved in fragment-protein binding are: N-H (hydrogen bond donors on aromatic and aliphatic N-H, amides and anilines; totalling 35%), aromatic nitrogen atoms (hydrogen bond acceptors; totalling 23%), and carbonyl oxygen group atoms (hydrogen bond acceptors on amides, ureas and ketones; totalling 22%). Furthermore, the elaboration of each fragment into its corresponding lead is analysed to identify the nominal synthetic growth vectors. In ∼80% of cases, growth originates from an aromatic or aliphatic carbon on the fragment and more than 50% of the total bonds formed are carbon-carbon bonds. This analysis reveals that growth from carbocentric vectors is key and therefore robust C-H functionalisation methods that tolerate the innate polar functionality on fragments could transform fragment-based drug discovery (FBDD). As a further resource to the community, we have provided the full data of our analysis as well as an online overlay page of the X-ray structures of the fragment hit and leads: https://astx.com/interactive/F2L-2021/.

Fragment-to-Lead Medicinal Chemistry Publications in 2019.

Fragment-based drug discovery (FBDD) has grown and matured to a point where it is valuable to keep track of its extent and details of application. This Perspective summarizes successful fragment-to-lead stories published in 2019. It is the fifth in a series that started with literature published in 2015. The analysis of screening methods, optimization strategies, and molecular properties of hits and leads are presented in the hope of informing best practices for FBDD. Moreover, FBDD is constantly evolving, and the latest technologies and emerging trends are summarized. These include covalent FBDD, FBDD for the stabilization of proteins or protein-protein interactions, FBDD for enzyme activators, new screening technologies, and advances in library design and chemical synthesis.

Fragment-to-Lead Medicinal Chemistry Publications in 2018.

This Perspective, the fourth in an annual series, summarizes fragment-to-lead (F2L) success stories published during 2018. Topics such as target class, screening methods, physicochemical properties, and ligand efficiency are discussed for the 2018 examples as well as for the combined 111 F2L examples covering 2015-2018. While the overall properties of fragments and leads have remained constant, a number of new trends are noted, for example, broadening of target class coverage and application of FBDD to covalent inhibitors. Moreover, several studies make use of fragment hits that were previously described in the literature, illustrating that fragments are versatile starting points that can be optimized to structurally diverse leads. By focusing on success stories, the hope is that this Perspective will identify and inform best practices in fragment-based drug discovery.

Discovery of novel inhibitors of Trypanosoma cruzi trans-sialidase from in silico screening.

trans-Sialidase from Trypanosoma cruzi (TcTS) has emerged as a potential drug target for treatment of Chagas disease. Here, we report the results of virtual screening for the discovery of novel TcTS inhibitors, which targeted both the sialic acid and sialic acid acceptor sites of this enzyme. A library prepared from the Evotec database of commercially available compounds was screened using the molecular docking program GOLD, following the application of drug-likeness filters. Twenty-three compounds selected from the top-scoring ligands were purchased and assayed using a fluorimetric assay. Novel inhibitor scaffolds, with IC(50) values in the submillimolar range were discovered. The 3-benzothiazol-2-yl-4-phenyl-but-3-enoic acid scaffold was studied in more detail, and TcTS inhibition was confirmed by an alternative sialic acid transfer assay. Attempts to obtain crystal structures of these compounds with TcTS proved unsuccessful but provided evidence of ligand binding at the active site.

Fragment-to-Lead Medicinal Chemistry Publications in 2017.

This Miniperspective is the third in a series reviewing fragment-to-lead publications from a given year. Following our reviews for 2015 and 2016, this Miniperspective provides tabulated summaries of relevant articles published in 2017 along with some general observations. In addition, we discuss insights obtained from analysis of the combined data set of 85 examples from all three years of publications.

Structure-Activity and Structure-Conformation Relationships of Aryl Propionic Acid Inhibitors of the Kelch-like ECH-Associated Protein 1/Nuclear Factor Erythroid 2-Related Factor 2 (KEAP1/NRF2) Protein-Protein Interaction.

The KEAP1-NRF2-mediated cytoprotective response plays a key role in cellular homoeostasis. Insufficient NRF2 signaling during chronic oxidative stress may be associated with the pathophysiology of several diseases with an inflammatory component, and pathway activation through direct modulation of the KEAP1-NRF2 protein-protein interaction is being increasingly explored as a potential therapeutic strategy. Nevertheless, the physicochemical nature of the KEAP1-NRF2 interface suggests that achieving high affinity for a cell-penetrant druglike inhibitor might be challenging. We recently reported the discovery of a highly potent tool compound which was used to probe the biology associated with directly disrupting the interaction of NRF2 with the KEAP1 Kelch domain. We now present a detailed account of the medicinal chemistry campaign leading to this molecule, which included exploration and optimization of protein-ligand interactions in three energetic "hot spots" identified by fragment screening. In particular, we also discuss how consideration of ligand conformational stabilization was important to its development and present evidence for preorganization of the lead compound which may contribute to its high affinity and cellular activity.

Fragment-to-Lead Medicinal Chemistry Publications in 2016.

The popularity of fragment-based drug discovery (FBDD) is demonstrated by the number of recent successful fragment-to-lead (F2L) publications. This Miniperspective provides a tabulated summary of the F2L literature published in the year 2016, along with discussion of general trends. It uses the same format as our summary of the 2015 literature and is intended to be a resource for both FBDD practitioners and medicinal chemists in general.

Diverse, high-quality test set for the validation of protein-ligand docking performance.

A procedure for analyzing and classifying publicly available crystal structures has been developed. It has been used to identify high-resolution protein-ligand complexes that can be assessed by reconstructing the electron density for the ligand using the deposited structure factors. The complexes have been clustered according to the protein sequences, and clusters have been discarded if they do not represent proteins thought to be of direct interest to the pharmaceutical or agrochemical industry. Rules have been used to exclude complexes containing non-drug-like ligands. One complex from each cluster has been selected where a structure of sufficient quality was available. The final Astex diverse set contains 85 diverse, relevant protein-ligand complexes, which have been prepared in a format suitable for docking and are to be made freely available to the entire research community (http://www.ccdc.cam.ac.uk). The performance of the docking program GOLD against the new set is assessed using a variety of protocols. Relatively unbiased protocols give success rates of approximately 80% for redocking into native structures, but it is possible to get success rates of over 90% with some protocols.

Predicting "Hot" and "Warm" Spots for Fragment Binding.

Computational fragment mapping methods aim to predict hotspots on protein surfaces where small fragments will bind. Such methods are popular for druggability assessment as well as structure-based design. However, to date researchers developing or using such tools have had no clear way of assessing the performance of these methods. Here, we introduce the first diverse, high quality validation set for computational fragment mapping. The set contains 52 diverse examples of fragment binding "hot" and "warm" spots from the Protein Data Bank (PDB). Additionally, we describe PLImap, a novel protocol for fragment mapping based on the Protein-Ligand Informatics force field (PLIff). We evaluate PLImap against the new fragment mapping test set, and compare its performance to that of simple shape-based algorithms and fragment docking using GOLD. PLImap is made publicly available from https://bitbucket.org/AstexUK/pli .

Fragment-Based Discovery of Potent and Selective DDR1/2 Inhibitors.

The DDR1 and DDR2 receptor tyrosine kinases are activated by extracellular collagen and have been implicated in a number of human diseases including cancer. We performed a fragment-based screen against DDR1 and identified fragments that bound either at the hinge or in the back pocket associated with the DFG-out conformation of the kinase. Modeling based on crystal structures of potent kinase inhibitors facilitated the "back-to-front" design of potent DDR1/2 inhibitors that incorporated one of the DFG-out fragments. Further optimization led to low nanomolar, orally bioavailable inhibitors that were selective for DDR1 and DDR2. The inhibitors were shown to potently inhibit DDR2 activity in cells but in contrast to unselective inhibitors such as dasatinib, they did not inhibit proliferation of mutant DDR2 lung SCC cell lines.

Fragment-based approaches to the discovery of kinase inhibitors.

Protein kinases are one of the most important families of drug targets, and aberrant kinase activity has been linked to a large number of disease areas. Although eminently targetable using small molecules, kinases present a number of challenges as drug targets, not least obtaining selectivity across such a large and relatively closely related target family. Fragment-based drug discovery involves screening simple, low-molecular weight compounds to generate initial hits against a target. These hits are then optimized to more potent compounds via medicinal chemistry, usually facilitated by structural biology. Here, we will present a number of recent examples of fragment-based approaches to the discovery of kinase inhibitors, detailing the construction of fragment-screening libraries, the identification and validation of fragment hits, and their optimization into potent and selective lead compounds. The advantages of fragment-based methodologies will be discussed, along with some of the challenges associated with using this route. Finally, we will present a number of key lessons derived both from our own experience running fragment screens against kinases and from a large number of published studies.

Efficient exploration of chemical space by fragment-based screening.

Screening methods seek to sample a vast chemical space in order to identify starting points for further chemical optimisation. Fragment based drug discovery exploits the superior sampling of chemical space that can be achieved when the molecular weight is restricted. Here we show that commercially available fragment space is still relatively poorly sampled and argue for highly sensitive screening methods to allow the detection of smaller fragments. We analyse the properties of our fragment library versus the properties of X-ray hits derived from the library. We particularly consider properties related to the degree of planarity of the fragments.