Molecular glues are small molecules that draw two proteins together, bringing a disease-related protein into contact with an enzyme that tags it for degradation. By tapping into the cell’s own recycling machinery, the approach could eliminate proteins linked to diseases that have long resisted conventional drugs. But finding new molecular glues has remained largely a matter of chance.
A new multiplexed screening platform that simultaneously tests multiple small molecules against a panel of E3 ubiquitin ligases, enzymes that tag proteins for degradation, has allowed a team of researchers to search for molecular glues more efficiently. Using this approach, the team identified a previously unknown molecular glue that becomes active only after being metabolized inside cells, pointing to a previously overlooked route for molecular glue discovery (Nature 2026, DOI:10.1038/s41586-026-10873-1).
Protein degradation expert Craig Crews, at Yale University, who was not involved in the work, says in an email to C&EN that “overcoming the serendipitous approach to molecular glue discovery has been a challenge. This new approach addresses this problem in a clever way and has revealed a very interesting new mechanism by which molecular glues can induce novel protein:protein interfaces.”
Although human DNA encodes roughly 600 E3 ligases, nearly all clinically relevant molecular glues exploit only a handful of them. “Molecular glue phenomena are rare,” says Hojong Yoon, a chemical biologist at the University of Texas MD Anderson Cancer Center and first author of the study. “I wanted to change the paradigm.”
While still a postdoctoral researcher at Dana-Farber Cancer Institute, Yoon did not want to modify the handful of molecular glues already known to researchers. Instead, he started by asking, “Why don’t we work on finding molecular glues in the first place, and then identify their more detailed mechanism?”
The screening technique yielded a compound dubbed M12, which appeared to recruit the RNA helicase DDX18, a protein implicated in several cancers, to the E3 ligase DCAF11.
But when Yoon and his team tried to validate the hit using purified proteins, the interaction disappeared. The researchers initially suspected that the proteins were at fault and repeated the experiments with freshly purified and commercially sourced proteins, but M12 still failed to act as a molecular glue.
M12 is modified with glutathione to produce a compound that acts as a molecular glue.
When the team added a crude cell lysate to the mix, the interaction was restored. That result suggested that some component of the cell was converting M12 into its active form.
Cryo-electron microscopy and follow-up biochemical experiments confirmed that M12 undergoes glutathione conjugation. The resulting metabolite, rather than the parent compound, binds DCAF11 and recruits DDX18 for degradation.
“We did not at all anticipate to find a metabolically activated molecular glue,” Eric S. Fischer, a cancer biologist at Dana-Farber Cancer Institute who was also part of the study, writes in an email. “But in hindsight, that represents one of the advantages of our screening approach using whole cell lysate and thereby keeping most cellular processes at least to some degree active.”
Yoon’s research group is now scaling up the screening platform to search for additional molecular glues and working to determine whether other drug metabolites can induce similar interactions. “We need to find more mass of new molecular glues,” he says, “and there are so many bottlenecks because we need to reassess drug discovery rules.”
