A recent study published in the journal Science described a crowdsourced, open-science, and structure-enabled drug discovery program for the main protease (Mpro) of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2).
Study: Open science discovery of potent noncovalent SARS-CoV-2 main protease inhibitors. Image Credit: Corona Borealis Studio / Shutterstock
The Importance of Antiviral Therapeutics in COVID-19 Control
The failure to abate coronavirus disease 2019 (COVID-19) will cause the virus to become endemic unless an accessible treatment is available. Antiviral therapeutics are essential to control COVID-19, and several oral antiviral agents have been approved, including molnupiravir, nirmatrelvir, and ensitrelvir. SARS-CoV-2 Mpro has been an attractive target for drug development, given its role in replication, high degree of conservation across CoVs, and dissimilarity to human proteases.
Crowdsourcing Drug Discovery: The COVID Moonshot Initiative
In the present study, researchers reported open science discovery of potent antivirals for SARS-CoV-2. This program, "COVID Moonshot," was aimed at SARS-CoV-2 Mpro, building off rapid fragment screening that assessed unique fragment-soaked crystals and identified 71 hits populating the active site. The non-covalent fragment hits lacked inhibitory activity in an enzymatic assay but offered a high-resolution map of interactions.
The team initiated an online crowdsourcing platform in March 2020 and asked participants to submit compounds designed based on fragment hits.
Synthesis and Screening
Biochemical assays and X-ray crystallography were used to evaluate compounds selected for synthesis, and their results were also released on the same platform. Designs were contributed by the core group (of labs and medicinal chemists) and the community.
A contract research organization (CRO), Enamine, was tasked with synthesizing compounds. The team computed synthetic routes of all submissions using the CRO's building block inventories and estimated synthetic complexity. The predicted synthetic complexity correlated with the actual time required for target compound synthesis. Next, alchemical free-energy calculations were used to estimate the potency of designs and analogs from virtual synthetic libraries.
The researchers used a global distributed computing network (Folding@home) to estimate the free energy of binding of all submissions. Initially, a small study was undertaken using the data generated from the first week of crowdsourced designs. The results of these calculations correlated well with experimentally determined affinities. Next, alchemical free-energy calculations were another criterion to guide selection and iterative design.
Three chronologically distinct design campaigns were observed – benzopyran ring decoration, benzopyran system replacement, and isoquinoline system replacement. These calculations helped select potent analogs from virtual libraries and highlighted where significant synthetic effort would be needed. As such, the team prioritized small libraries suggested for synthesis.
Optimizing Antiviral Potency Through Structure-Activity Analysis
Rapid structure-activity relationship (SAR) evaluation was complemented by using nanomole-scale high-throughput chemistry (HTC), as exemplified by the optimizations of amide coupling to extend MAT-POS-4223bc15-21 and Chan-Lam reaction to extend ADA-UCB-6c2cb422-1. Seven and 20 compounds from the Chan-Lam and amide series, respectively, were selected for resynthesis.
The extended compounds had similar binding as the parent compounds. One of the compounds in the Chan-Lam series had a slightly higher half-maximal inhibitory concentration (IC50) than the parent compound. On the other hand, several compounds in the amide series showed up to 300-fold improvement in IC50 relative to the parent compound.
Crystal soaking and x-ray diffraction yielded 587 structures. A subset of structures was analyzed that revealed ligand engagement hotspots and binding pockets' plasticity. The P1 and P2 pockets were hotspots of interactions. Some salient interactions in the P1 pocket sampled by the ligands included N145 (hydrophobic), H163 (hydrogen-bond donor), and E166 (hydrogen-bond acceptor).
By contrast, hydrophobic interactions with M165 and π-stacking interactions with H41 were predominant in P2. Next, the team explored P1, which has a steep SAR due to its preference for directional hydrogen-bond interactions and rigidity. Potency was increased by replacing pyridine with isoquinoline, introducing additional interactions with N142.
SAR around P2 was substantially tolerant to change. A shift in potency was possible by rigidifying the scaffold. Specifically, a tetrahydropyran ring was introduced to transform the substituent into a chromane moiety. Next, the chromane was substituted with tetrahydroisoquinoline to maintain potency.
Finally, a library was constructed through sulphonamide Schotten-Baumann coupling that increased inhibitory activity and antiviral efficacy. Overall, this resulted in a series of potent antivirals with a low brain penetrance, enhanced oral bioavailability, and favorable safety profile but a moderate in vitro-in vivo correlation for clearance.
Promising Results: Lead Compounds Show High Efficacy
One of the final leads, MAT-POS-e194df51-1, was evaluated in antiviral assays in various cell lines and was not cytotoxic, displaying a median effective concentration of 126 nM in HeLa-ACE2 cells and 64 nM in A549-ACE2-TMPRSS2 cells. This compound was also cross-reactive against SARS-CoV-2 variants. Crystallographic analysis showed that the interactions of this compound with the Mpro binding site were distinct from that of approved Mpro inhibitors.
Implications and Future Prospects of COVID Moonshot Findings
Overall, the study demonstrated the success of an open-science, patent-free antiviral discovery program in developing a differentiated lead. Notably, ensitrelvir approved in Japan was identified, in part, based on crystallographic data shared by COVID Moonshot. This project and lead COVID-19 series have been adopted into the Drugs for Neglected Diseases Initiative for further optimization of leads and preclinical development.
Study shows no developmental delays in toddlers post-COVID-19 infection