The bromodomain-containing proteins BRD2, BRD3, BRD4 and BRDT are epigenetic readers of acetyllysine that are currently of great interest as therapeutic targets for a wide range of disorders. Thousands of small molecules targeting the acetyllysine-binding pocket have been synthesized and a subset used in more than 70 clinical trials to date, mostly focused in the field of oncology. These molecules, however, are not selective for one family member over the others and most trials have been disappointing given the extremely promising preclinical data.
In an effort to find better bromodomain inhibitors, we have used an mRNA display approach to screen a library of approximately a trillion cyclic peptides. Our biochemical data reveal a subset of peptides that are capable of recognizing a single BET-family paralogue with specificities of up to 10,000-fold.
Strikingly, the peptides achieve this specificity not by recognizing unique sequence motifs in each paralogue but rather by exploiting subtle differences in conformational dynamics (wiggly-ness!) whose origins are remote from the peptide-protein interface.
Our data suggest that such peptides might open the door to a completely new class of bromodomain inhibitors that have unprecedented paralogue-level specificity and therefore potentially superior properties as therapeutic agents.