An abundance of allostery found by combining biochemical assays and AI

Researchers at the Centre for Genomic Regulation have developed a new technique for identifying allosteric sites in proteins (Nature 2022, DOI: 10.1038/s41586-022-04586-4).

While many drugs target the active site of a protein, where the chemistry happens, others take a more sideways approach. Allosteric sites are additional pockets that drugmakers can use to regulate the activity of enzymes, dialing their activity up or down by shifting the shape of the protein. Allosteric inhibitors can also be used to target disease-causing mutant proteins, such as those made by cancer cells, without affecting the analogous healthy proteins.

The team, led by Ben Lehner, analyzed two well-characterized human protein–binding domains and found many more sites with allosteric potential than had previously been identified. The researchers say the new technique could be a boon for drug discovery.

Until recently, finding allosteric sites has been expensive and time consuming. In their new work, the researchers used common cell assays that measure protein binding or abundance. In these assays, if the enzyme is functional, the cells will gain resistance to an antibiotic, meaning cells’ growth rate in the presence of the antibiotic is a proxy that lets scientists quantify protein binding. The researchers made thousands of slightly different versions of a protein by systematically mutating it; they then ran assays on all versions of the protein. By feeding all their assay data into a neural net, the team could pinpoint potential allosteric sites, characterize them, and explore how they could be modified.