Harnessing DNA-Encoded Libraries to Discover Bioactive Small Molecules

Abstract

Given the ever-increasing number of proteins, nucleic acids, and metabolites implicated in human disease, it is highly desirable to develop small molecules to probe therapeutically relevant biological pathways and to serve as leads for the development of new medicines. Recently, DNA-encoded chemical libraries – solution-phase collections of compounds each covalently linked to and specifically barcoded by a unique DNA sequence – have played an increasingly important role in the discovery of such bioactive compounds. DNA-encoded chemical libraries have several advantages, leveraging the extremely high sensitivity afforded by DNA amplification and the remarkable accessibility of modern high-throughput DNA sequencing. Taken together, these properties enable the efficient synthesis of large, diverse DNA-linked compound collections and the facile discovery of novel molecular interactions from these libraries. In vitro affinity selections on DNA-encoded chemical libraries have led to the discovery of new classes of synthetic small-molecule ligands against a variety of protein targets. Previous work in the Liu group validated our ability to identify potent probe molecules from our DNA-templated libraries, as well as explore the biology of their protein targets through chemical means. I have applied in vitro selections on our DNA-templated macrocycle libraries to a large number of proteins and protein complexes associated with human disease. In one case, our efforts led to new inhibitors of IDE from a library of 256,000 macrocycles, thus validating this library as a source of new bioactive compounds. In addition, I describe our use of the IDUP system to simultaneously evaluate all possible protein-ligand interactions out of combined libraries of DNA-tagged proteins and DNA-encoded small molecules. Not only were we successful in recapitulating known binding interactions in this assay format, I also discovered a previously unknown covalent inhibitor, ethacrynic acid, of the human protein kinase MAP2K6. I further probed the mechanistic basis of this binding interaction and showed that inhibition is due in part to ethacrynic acid’s ability to alkylate a nonconserved cysteine residue in MAP2K6. These results are illustrative of the potential of unbiased in vitro binding selections to uncover bioactive molecules with novel modes of protein target engagement.