AI-Accelerated Design of Targeted Covalent Inhibitors for SARS-CoV-2.

Direct-acting antivirals for the treatment of the COVID-19 pandemic caused by the SARS-CoV-2 virus are needed to complement vaccination efforts. Given the ongoing emergence of new variants, automated experimentation, and active learning based fast workflows for antiviral lead discovery remain critical to our ability to address the pandemic's evolution in a timely manner. While several such pipelines have been introduced to discover candidates with noncovalent interactions with the main protease (Mpro), here we developed a closed-loop artificial intelligence pipeline to design electrophilic warhead-based covalent candidates. This work introduces a deep learning-assisted automated computational workflow to introduce linkers and an electrophilic "warhead" to design covalent candidates and incorporates cutting-edge experimental techniques for validation. Using this process, promising candidates in the library were screened, and several potential hits were identified and tested experimentally using native mass spectrometry and fluorescence resonance energy transfer (FRET)-based screening assays. We identified four chloroacetamide-based covalent inhibitors of Mpro with micromolar affinities (KI of 5.27 µM) using our pipeline. Experimentally resolved binding modes for each compound were determined using room-temperature X-ray crystallography, which is consistent with the predicted poses. The induced conformational changes based on molecular dynamics simulations further suggest that the dynamics may be an important factor to further improve selectivity, thereby effectively lowering KI and reducing toxicity. These results demonstrate the utility of our modular and data-driven approach for potent and selective covalent inhibitor discovery and provide a platform to apply it to other emerging targets.

High-Throughput Virtual Screening and Validation of a SARS-CoV-2 Main Protease Noncovalent Inhibitor.

Despite the recent availability of vaccines against the acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the search for inhibitory therapeutic agents has assumed importance especially in the context of emerging new viral variants. In this paper, we describe the discovery of a novel noncovalent small-molecule inhibitor, MCULE-5948770040, that binds to and inhibits the SARS-Cov-2 main protease (Mpro) by employing a scalable high-throughput virtual screening (HTVS) framework and a targeted compound library of over 6.5 million molecules that could be readily ordered and purchased. Our HTVS framework leverages the U.S. supercomputing infrastructure achieving nearly 91% resource utilization and nearly 126 million docking calculations per hour. Downstream biochemical assays validate this Mpro inhibitor with an inhibition constant (Ki) of 2.9 µM (95% CI 2.2, 4.0). Furthermore, using room-temperature X-ray crystallography, we show that MCULE-5948770040 binds to a cleft in the primary binding site of Mpro forming stable hydrogen bond and hydrophobic interactions. We then used multiple µs-time scale molecular dynamics (MD) simulations and machine learning (ML) techniques to elucidate how the bound ligand alters the conformational states accessed by Mpro, involving motions both proximal and distal to the binding site. Together, our results demonstrate how MCULE-5948770040 inhibits Mpro and offers a springboard for further therapeutic design.

Selective class IIa histone deacetylase inhibition via a nonchelating zinc-binding group.

In contrast to studies on class I histone deacetylase (HDAC) inhibitors, the elucidation of the molecular mechanisms and therapeutic potential of class IIa HDACs (HDAC4, HDAC5, HDAC7 and HDAC9) is impaired by the lack of potent and selective chemical probes. Here we report the discovery of inhibitors that fill this void with an unprecedented metal-binding group, trifluoromethyloxadiazole (TFMO), which circumvents the selectivity and pharmacologic liabilities of hydroxamates. We confirm direct metal binding of the TFMO through crystallographic approaches and use chemoproteomics to demonstrate the superior selectivity of the TFMO series relative to a hydroxamate-substituted analog. We further apply these tool compounds to reveal gene regulation dependent on the catalytic active site of class IIa HDACs. The discovery of these inhibitors challenges the design process for targeting metalloenzymes through a chelating metal-binding group and suggests therapeutic potential for class IIa HDAC enzyme blockers distinct in mechanism and application compared to current HDAC inhibitors.

Structure-guided optimization of small molecule c-Abl activators.

c-Abl kinase is maintained in its normal inactive state in the cell through an assembled, compact conformation. We describe two chemical series that bind to the myristoyl site of the c-Abl kinase domain and stimulate c-Abl activation. We hypothesize that these molecules activate c-Abl either by blocking the C-terminal helix from adopting a bent conformation that is critical for the formation of the autoinhibited conformation or by simply providing no stabilizing interactions to the bent conformation of this helix. Structure-based molecular modeling guided the optimization of binding and activation of c-Abl of these two chemical series and led to the discovery of c-Abl activators with nanomolar potency. The small molecule c-Abl activators reported herein could be used as molecular tools to investigate the biological functions of c-Abl and therapeutic implications of its activation.

Potent and selective covalent inhibition of the papain-like protease from SARS-CoV-2.

Direct-acting antivirals are needed to combat coronavirus disease 2019 (COVID-19), which is caused by severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2). The papain-like protease (PLpro) domain of Nsp3 from SARS-CoV-2 is essential for viral replication. In addition, PLpro dysregulates the host immune response by cleaving ubiquitin and interferon-stimulated gene 15 protein from host proteins. As a result, PLpro is a promising target for inhibition by small-molecule therapeutics. Here we design a series of covalent inhibitors by introducing a peptidomimetic linker and reactive electrophile onto analogs of the noncovalent PLpro inhibitor GRL0617. The most potent compound inhibits PLpro with kinact/KI = 9,600 M-1 s-1, achieves sub-µM EC50 values against three SARS-CoV-2 variants in mammalian cell lines, and does not inhibit a panel of human deubiquitinases (DUBs) at >30 µM concentrations of inhibitor. An X-ray co-crystal structure of the compound bound to PLpro validates our design strategy and establishes the molecular basis for covalent inhibition and selectivity against structurally similar human DUBs. These findings present an opportunity for further development of covalent PLpro inhibitors.

Hepatitis C virus NS3/4A inhibitors and other drug-like compounds as covalent binders of SARS-CoV-2 main protease.

Severe acute respiratory syndrome-coronavirus 2 (SARS-CoV-2), which causes coronavirus disease 2019 (COVID-19), threatens global public health. The world needs rapid development of new antivirals and vaccines to control the current pandemic and to control the spread of the variants. Among the proteins synthesized by the SARS-CoV-2 genome, main protease (Mpro also known as 3CLpro) is a primary drug target, due to its essential role in maturation of the viral polyproteins. In this study, we provide crystallographic evidence, along with some binding assay data, that three clinically approved anti hepatitis C virus drugs and two other drug-like compounds covalently bind to the Mpro Cys145 catalytic residue in the active site. Also, molecular docking studies can provide additional insight for the design of new antiviral inhibitors for SARS-CoV-2 using these drugs as lead compounds. One might consider derivatives of these lead compounds with higher affinity to the Mpro as potential COVID-19 therapeutics for further testing and possibly clinical trials.

Structural, Electronic, and Electrostatic Determinants for Inhibitor Binding to Subsites S1 and S2 in SARS-CoV-2 Main Protease.

Creating small-molecule antivirals specific for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) proteins is crucial to battle coronavirus disease 2019 (COVID-19). SARS-CoV-2 main protease (Mpro) is an established drug target for the design of protease inhibitors. We performed a structure-activity relationship (SAR) study of noncovalent compounds that bind in the enzyme's substrate-binding subsites S1 and S2, revealing structural, electronic, and electrostatic determinants of these sites. The study was guided by the X-ray/neutron structure of Mpro complexed with Mcule-5948770040 (compound 1), in which protonation states were directly visualized. Virtual reality-assisted structure analysis and small-molecule building were employed to generate analogues of 1. In vitro enzyme inhibition assays and room-temperature X-ray structures demonstrated the effect of chemical modifications on Mpro inhibition, showing that (1) maintaining correct geometry of an inhibitor's P1 group is essential to preserve the hydrogen bond with the protonated His163; (2) a positively charged linker is preferred; and (3) subsite S2 prefers nonbulky modestly electronegative groups.

Design and synthesis of orally bioavailable serum and glucocorticoid-regulated kinase 1 (SGK1) inhibitors.

The lead serum and glucocorticoid-related kinase 1 (SGK1) inhibitors 4-(5-phenyl-1H-pyrrolo[2,3-b]pyridin-3-yl)benzoic acid (1) and {4-[5-(2-naphthalenyl)-1H-pyrrolo[2,3-b]pyridin-3-yl]phenyl}acetic acid (2) suffer from low DNAUC values in rat, due in part to formation and excretion of glucuronic acid conjugates. These PK/glucuronidation issues were addressed either by incorporating a substituent on the 3-phenyl ring ortho to the key carboxylate functionality of 1 or by substituting on the group in between the carboxylate and phenyl ring of 2. Three of these analogs have been identified as having good SGK1 inhibition potency and have DNAUC values suitable for in vivo testing.

Crystal structure of the kinase domain of serum and glucocorticoid-regulated kinase 1 in complex with AMP PNP.

Serum and glucocorticoid-regulated kinase 1 (SGK1) is a serine/threonine protein kinase of the AGC family which participates in the control of epithelial ion transport and is implicated in proliferation and apoptosis. We report here the 1.9 A crystal structure of the catalytic domain of inactive human SGK1 in complex with AMP-PNP. SGK1 exists as a dimer formed by two intermolecular disulfide bonds between Cys258 in the activation loop and Cys193. Although most of the SGK1 structure closely resembles the common protein kinase fold, the structure around the active site is unique when compared to most protein kinases. The alphaC helix is not present in this inactive form of SGK1 crystal structure; instead, the segment corresponding to the C helix forms a beta-strand that is stabilized by the N-terminal segment of the activation loop through a short antiparallel beta-sheet. Since the differences from other kinases occur around the ATP binding site, this structure can provide valuable insight into the design of selective and highly potent ATP-competitive inhibitors of SGK1 kinase.

Discovery of aminofurazan-azabenzimidazoles as inhibitors of Rho-kinase with high kinase selectivity and antihypertensive activity.

The discovery, proposed binding mode, and optimization of a novel class of Rho-kinase inhibitors are presented. Appropriate substitution on the 6-position of the azabenzimidazole core provided subnanomolar enzyme potency in vitro while dramatically improving selectivity over a panel of other kinases. Pharmacokinetic data was obtained for the most potent and selective examples and one (6n) has been shown to lower blood pressure in a rat model of hypertension.

A critical assessment of docking programs and scoring functions.

Docking is a computational technique that samples conformations of small molecules in protein binding sites; scoring functions are used to assess which of these conformations best complements the protein binding site. An evaluation of 10 docking programs and 37 scoring functions was conducted against eight proteins of seven protein types for three tasks: binding mode prediction, virtual screening for lead identification, and rank-ordering by affinity for lead optimization. All of the docking programs were able to generate ligand conformations similar to crystallographically determined protein/ligand complex structures for at least one of the targets. However, scoring functions were less successful at distinguishing the crystallographic conformation from the set of docked poses. Docking programs identified active compounds from a pharmaceutically relevant pool of decoy compounds; however, no single program performed well for all of the targets. For prediction of compound affinity, none of the docking programs or scoring functions made a useful prediction of ligand binding affinity.

Structure activity relationships of 5-, 6-, and 7-methyl-substituted azepan-3-one cathepsin K inhibitors.

The syntheses, in vitro characterizations, and rat and monkey in vivo pharmacokinetic profiles of a series of 5-, 6-, and 7-methyl-substituted azepanone-based cathepsin K inhibitors are described. Depending on the particular regiochemical substitution and stereochemical configuration, methyl-substituted azepanones were identified that had widely varied cathepsin K inhibitory potency as well as pharmacokinetic properties compared to the 4S-parent azepanone analogue, 1 (human cathepsin K, K(i,app) = 0.16 nM, rat oral bioavailability = 42%, rat in vivo clearance = 49.2 mL/min/kg). Of particular note, the 4S-7-cis-methylazepanone analogue, 10, had a K(i,app) = 0.041 nM vs human cathepsin K and 89% oral bioavailability and an in vivo clearance rate of 19.5 mL/min/kg in the rat. Hypotheses that rationalize some of the observed characteristics of these closely related analogues have been made using X-ray crystallography and conformational analysis. These examples demonstrate the potential for modulation of pharmacological properties of cathepsin inhibitors by substituting the azepanone core. The high potency for inhibition of cathepsin K coupled with the favorable rat and monkey pharmacokinetic characteristics of compound 10, also known as SB-462795 or relacatib, has made it the subject of considerable in vivo evaluation for safety and efficacy as an inhibitor of excessive bone resorption in rat, monkey, and human studies, which will be reported elsewhere.

Azepanone-based inhibitors of human cathepsin L.

The extension of a previously reported cathepsin K azepanone-based inhibitor template to the design and synthesis of potent and selective inhibitors of the homologous cysteine protease cathepsin L is detailed. Structure-activity studies examining the effect of inhibitor selectivity as a function of the P3 and P2 binding elements of the potent cathepsin K inhibitor 1 revealed that incorporation of either a P3 quinoline-8-carboxamide or a naphthylene-1-carboxamide led to increased selectivity for cathepsin L over cathepsin K. Substitution of the P2 leucine of 1 with either a phenylalanine or a beta-naphthylalanine also resulted in an increased selectivity for cathepsin L over cathepsin K. Molecular modeling studies with the inhibitors docked within the active sites of both cathepsins L and K have rationalized the observed selectivities. Optimization of cathepsin L binding by the combination of the P3 naphthylene-1-carboxamide with the P2 beta-naphthylalanine provided 15, which is a potent, selective, and competitive inhibitor of human cathepsin L with a K(i) = 0.43 nM.

First X-ray cocrystal structure of a bacterial FabH condensing enzyme and a small molecule inhibitor achieved using rational design and homology modeling.

The first cocrystal structure of a bacterial FabH condensing enzyme and a small molecule inhibitor is reported. The inhibitor was obtained by rational modification of a high throughput screening lead with the aid of a S. pneumoniae FabH homology model. This homology model was used to design analogues that would have both high affinity for the enzyme and appropriate aqueous solubility to facilitate cocrystallization studies.

Indole naphthyridinones as inhibitors of bacterial enoyl-ACP reductases FabI and FabK.

Bacterial enoyl-ACP reductase (FabI) is responsible for catalyzing the final step of bacterial fatty acid biosynthesis and is an attractive target for the development of novel antibacterial agents. Previously we reported the development of FabI inhibitor 4 with narrow spectrum antimicrobial activity and in vivo efficacy against Staphylococcus aureus via intraperitoneal (ip) administration. Through iterative medicinal chemistry aided by X-ray crystal structure analysis, a new series of inhibitors has been developed with greatly increased potency against FabI-containing organisms. Several of these new inhibitors have potent antibacterial activity against multidrug resistant strains of S. aureus, and compound 30 demonstrates exceptional oral (po) in vivo efficacy in a S. aureus infection model in rats. While optimizing FabI inhibitory activity, compounds 29 and 30 were identified as having low micromolar FabK inhibitory activity, thereby increasing the antimicrobial spectrum of these compounds to include the FabK-containing pathogens Streptococcus pneumoniae and Enterococcus faecalis. The results described herein support the hypothesis that bacterial enoyl-ACP reductases are valid targets for antibacterial agents.

Discovery of aminopyridine-based inhibitors of bacterial enoyl-ACP reductase (FabI).

Bacterial enoyl-ACP reductase (FabI) catalyzes the final step in each cycle of bacterial fatty acid biosynthesis and is an attractive target for the development of new antibacterial agents. Our efforts to identify potent, selective FabI inhibitors began with screening of the GlaxoSmithKline proprietary compound collection, which identified several small-molecule inhibitors of Staphylococcus aureus FabI. Through a combination of iterative medicinal chemistry and X-ray crystal structure based design, one of these leads was developed into the novel aminopyridine derivative 9, a low micromolar inhibitor of FabI from S. aureus (IC(50) = 2.4 microM) and Haemophilus influenzae (IC(50) = 4.2 microM). Compound 9 has good in vitro antibacterial activity against several organisms, including S. aureus (MIC = 0.5 microg/mL), and is effective in vivo in a S. aureus groin abscess infection model in rats. Through FabI overexpressor and macromolecular synthesis studies, the mode of action of 9 has been confirmed to be inhibition of fatty acid biosynthesis via inhibition of FabI. Taken together, these results support FabI as a valid antibacterial target and demonstrate the potential of small-molecule FabI inhibitors for the treatment of bacterial infections.

Discovery and characterization of a cell-permeable, small-molecule c-Abl kinase activator that binds to the myristoyl binding site.

c-Abl kinase activity is regulated by a unique mechanism involving the formation of an autoinhibited conformation in which the N-terminal myristoyl group binds intramolecularly to the myristoyl binding site on the kinase domain and induces the bending of the αI helix that creates a docking surface for the SH2 domain. Here, we report a small-molecule c-Abl activator, DPH, that displays potent enzymatic and cellular activity in stimulating c-Abl activation. Structural analyses indicate that DPH binds to the myristoyl binding site and prevents the formation of the bent conformation of the αI helix through steric hindrance, a mode of action distinct from the previously identified allosteric c-Abl inhibitor, GNF-2, that also binds to the myristoyl binding site. DPH represents the first cell-permeable, small-molecule tool compound for c-Abl activation.

Blind docking of pharmaceutically relevant compounds using RosettaLigand.

It is difficult to properly validate algorithms that dock a small molecule ligand into its protein receptor using data from the public domain: the predictions are not blind because the correct binding mode is already known, and public test cases may not be representative of compounds of interest such as drug leads. Here, we use private data from a real drug discovery program to carry out a blind evaluation of the RosettaLigand docking methodology and find that its performance is on average comparable with that of the best commercially available current small molecule docking programs. The strength of RosettaLigand is the use of the Rosetta sampling methodology to simultaneously optimize protein sidechain, protein backbone and ligand degrees of freedom; the extensive benchmark test described here identifies shortcomings in other aspects of the protocol and suggests clear routes to improving the method.

Validation studies of the site-directed docking program LibDock.

The performance of the site-features docking algorithm LibDock has been evaluated across eight GlaxoSmithKline targets as a follow-up to a broad validation study of docking and scoring software (Warren, G. L.; Andrews, W. C.; Capelli, A.; Clarke, B.; Lalonde, J.; Lambert, M. H.; Lindvall, M.; Nevins, N.; Semus, S. F.; Senger, S.; Tedesco, G.; Walls, I. D.; Woolven, J. M.; Peishoff, C. E.; Head, M. S. J. Med. Chem. 2006, 49, 5912-5931). Docking experiments were performed to assess both the accuracy in reproducing the binding mode of the ligand and the retrieval of active compounds in a virtual screening protocol using both the DJD (Diller, D. J.; Merz, K. M., Jr. Proteins 2001, 43, 113-124) and LigScore2 (Krammer, A. K.; Kirchoff, P. D.; Jiang, X.; Venkatachalam, C. M.; Waldman, M. J. Mol. Graphics Modell. 2005, 23, 395-407) scoring functions. This study was conducted using DJD scoring, and poses were rescored using all available scoring functions in the Accelrys LigandFit module, including LigScore2. For six out of eight targets at least 30% of the ligands were docked within a root-mean-square difference (RMSD) of 2.0 A for the crystallographic poses when the LigScore2 scoring function was used. LibDock retrieved at least 20% of active compounds in the top 10% of screened ligands for four of the eight targets in the virtual screening protocol. In both studies the LigScore2 scoring function enhanced the retrieval of crystallographic poses or active compounds in comparison with the results obtained using the DJD scoring function. The results for LibDock accuracy and ligand retrieval in virtual screening are compared to 10 other docking and scoring programs. These studies demonstrate the utility of the LigScore2 scoring function and that LibDock as a feature directed docking method performs as well as docking programs that use genetic/growing and Monte Carlo driven algorithms.