Stick chemistry: High-throughput method for finding new molecular glues

Molecular glues are enjoying the spotlight, but discovering new ones is often a matter of luck. A new method, developed by Scripps Research chemist Michael Erb and colleagues, aims to discover new glues more intentionally with a “target-based” approach (Nat. Chem. Biol. 2026, DOI: 10.1038/s41589-025-02137-2).

To start, researchers take a protein of interest and an already known ligand for that protein, and then make many variants of that ligand using high-throughput sulfur fluoride exchange (SuFEx) click chemistry. These modified ligands can then be screened for degradation activity in cells.

Erb says the inspiration for the new method came from how previously discovered glues, like those for the proteins cyclin-K and BCL6, have structures similar to those of already-discovered ligands for those proteins. “We thought maybe we could find those types of examples in an intentional and prospective way, where we could just start with the ligand and prospectively convert them into molecular glues by sampling very large areas of chemical space,” he says.

In the new work, the researchers didn’t discriminate based on how the discovered glues actually work; they were first just interested in whether they did work. That flexibility helped lead them toward finding pathways beyond the protein cereblon, which Erb says has typically taken center stage as a recruiter for the E3 ligase because of its versatility.

One identified glue, for example, acts through an unusual mechanism in which it recruits FBXO3, a protein that had never before been identified in protein degradation. Another does act through cereblon, but it recruits its target protein before it recruits cereblon, whereas Erb says cereblon is typically recruited first.