Scalable platform for cellular and biochemical screening of combinatorial small-molecule libraries in droplets

Abstract

A new pattern of hit discovery is emerging with the rise of chemocentric modalities, such as chemical induced proximity (CIP). Synthesis and functional screening of dynamic, purpose-built libraries, such as E3 ligase biased libraries for targeted protein degradation (TPD), is redefining the druggable landscape. However, this new paradigm arguably lies beyond the remit of conventional screening methods which have been optimized for static, diversity-oriented libraries. To address this gap, we developed microfluidic compound screening in droplets (MicDrop), a scalable platform where purpose-built DNA-encoded one-bead one-compound library members are individually tested at high concentrations in pico-liter sized droplets for either biochemical or cellular function. A proof-of-concept CRBN library demonstrated the robustness of the workflow through an IKZF3 degradation screen which discovered an unexpected succinimide-based IKZF3 degrader, while reproducing the relative potency ranking of reference compounds. Screening of a prospective VHL library for CDO1 recruitment and degradation demonstrated that orthogonal assays gave an informative consensus hit list and an actionable machine-learning (ML) recruitment model. These findings establish a blueprint for intentional discovery of chemical inducers of proximity, while producing reliable data to accelerate ML-based design-make-test-analyze (DMTA) cycles in pursuit of new therapeutics. O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=106 SRC="FIGDIR/small/730915v1_ufig1.gif" ALT="Figure 1"> View larger version (37K): org.highwire.dtl.DTLVardef@12de3cborg.highwire.dtl.DTLVardef@1c6618dorg.highwire.dtl.DTLVardef@12ea42eorg.highwire.dtl.DTLVardef@11cea77_HPS_FORMAT_FIGEXP M_FIG C_FIG