Crystal structures of the novel coronavirus protease guide drug development

With the novel coronavirus, SARS-CoV-2, spreading across the globe, labs are shutting down as people are told to stay at home. But some labs are still hard at work, looking for druggable targets to treat COVID-19, the illness caused by the virus. Two groups of researchers are using different approaches to find new inhibitors of a key part of the virus’ self-replication machinery.

There aren’t many targets for anticoronavirus drugs because the virus doesn’t produce many proteins, says Matthew Todd, an expert in drug discovery at University College London who isn’t involved with either of the two projects. But one focus for medicinal chemists has been the virus’s main protease known as Mpro or 3CLpro. This enzyme processes a polyprotein chain coded by the virus’s RNA, chopping up the chain into functional proteins that the virus then uses to assemble itself and multiply. Disrupting this key piece of the virus’s self-replication machinery could bring an infection screeching to a halt.

“The protease is essential, but has no human homologues,” Todd says. So inhibitors of the protease have less of a chance of hitting a human protease, he adds.

To try and develop inhibitors for Mpro, two different international collaborations have used synchrotron X-rays to get high-resolution structures of the protease with potential inhibitors bound. While the approaches the teams are taking are different, the researchers are keen to stress that their approaches are complementary, and that the lead compounds they are developing are a long way from being used on patients with COVID-19.

At the University of Lübeck, Rolf Hilgenfeld has been studying coronaviruses for years. During the Middle East respiratory syndrome outbreak in 2013, which was caused by a coronavirus called MERS-CoV, Hilgenfeld’s team started working on protease inhibitors, giving the researchers a head start on SARS-CoV-2.

Earlier this year, Hilgenfeld’s team used high-intensity synchrotron X-rays to get a crystal structure of the new virus’s Mpro at high resolution and used that information to optimize an existing α-ketoamide inhibitor developed for fighting other coronaviruses (Science 2020, DOI 10.1126/science.abb3405). The synthesis of the new inhibitors was delayed by shutdowns in China, a result of the outbreak, but recent lab tests have now shown that one of these optimized compounds (shown) can bind to and inhibit the protease. Tests on healthy mice also suggest that the molecule could be administered by inhalation.

Hilgenfeld hopes this lead compound will be developed by a consortium set up by the European Union to fight COVID-19, but he expects that the road from this molecule to a functional drug could take years, perhaps ready in time for the next coronavirus that pops up.

Meanwhile, another consortium of scientists is trying to speed up the process of developing a viral protease inhibitor with a huge crowdsourced initiative. Earlier in March, crystallographers at the Diamond Light Source also solved the structure of the SARS-CoV-2 main protease at high resolution. They then completed a large fragment-screen by soaking protein crystals with small molecules representing fragments of possible drugs to see which fragments bound to the enzyme.

Based on those data, the team started the crowdsourced initiative to combine the expertise of multiple labs and researchers around the world to process as many possible protease inhibitor structures as possible. Chemists are invited to design new compounds—or submit existing ones that might bind to the protease—on a website created by an artificial intelligence medicinal chemistry start-up called PostEra. Submitted structures get prioritized based on factors like ease of synthesis and possible toxicity before a custom-synthesis firm makes the molecules. Researchers then screen the molecules for binding activity.

“The goal of the project is to find an antiviral that can get to the clinical stage as soon as we can,” explains Alpha Lee, who is part of the team behind the initiative. The strategy, he says, is to open the funnel of candidate compounds as wide as possible. They hope to find several lead compounds and then use their AI algorithms to suggest changes to the molecules to help speed up drug development.

Todd describes the two approaches as “both interesting projects.” But he cautions that finding promising candidates is just one step of the process. The bottleneck is often optimizing those candidates into viable drugs and, he says, “that’s difficult to do under the right conditions.” With labs shutting down to reduce social contact, both teams hope they can keep their projects running.

“It’s always key to develop several things at once,” adds Nir London, whose lab is involved in screening crowdsourced fragments. “I think both approaches are required and necessary and we can learn from each other.”

“I think this is a wake up message for the world that infectious diseases have long been neglected and it’s time to not neglect them,” Lee says. Hilgenfeld agrees. The development of antiviral drugs should be decoupled from recurring outbreaks he says, so we can ensure the sustainable development of antiviral drugs.