Enzyme inhibition remains a dominant strategy for therapeutic intervention given the diverse roles enzymes play in catalyzing critical biochemical reactions. Despite successes in disrupting enzyme activity using small molecules, this modality struggles when specificity is required to distinguish between enzymes with high degrees of homology within their active sites. While larger biologics such as antibodies can be engineered to have exquisite binding specificity, isolating binding proteins with inhibitory properties remains challenging. In this work, we sought to leverage the complementary properties of antibodies (binding specificity) and small molecules (potent inhibitory properties) to discover hybrid structures capable of effectively inhibiting enzymes selectively. By utilizing genetic code expansion and yeast surface display, we have been able to generate large libraries of protein-small molecule hybrids. With a clickable noncanonical amino acid and complementary small molecule warhead, we have constructed multiple “chemically expanded” antibody libraries. Two distinct antibody libraries in our lab enable small molecule presentation either within or outside of the CDR-H3 region, respectively (Complementarity determining region H-3 (CDR-H3) is the primary specificity determinant of many antibodies). In this work, we seek to compare the performance of these two libraries in the discovery of hybrids targeting the well-behaved model protein bovine carbonic anhydrase (bCA). Screening against bCA using magnetic bead sorting and fluorescence-activated cell sorting (FACS) yielded strongly enriched populations capable of binding to bCA as well as a number of single clones that bind to the target in hybrid form on the yeast surface. Removal of the small molecule warhead from each clone results in a subsequent loss of binding detection, highlighting the importance of the warhead in these discovery campaigns. Additionally, next generation sequencing (NGS) enabled a detailed analysis into both residue-specific enrichments in the CDR-H3 as well as sequence clusters that identify potential epistatic interactions between residues in enriched populations. While additional characterization of clones remains in progress, the isolation of diverse collections of hybrids from two distinct, chemically expanded antibody libraries suggests that there are multiple effective ways to integrate small molecule functionalities into libraries of antibodies. These results motivate further efforts to screen protein-small molecule hybrids in this yeast display library screening format against challenging, therapeutically relevant enzymes for which no selective inhibitors exist.
