Chromatin structure and function, and consequently cellular phenotype, is regulated in part by a network of chromatin-modifying enzymes that place post-translational modifications (PTMs) on histone tails. These marks serve as recruitment sites for other chromatin regulatory complexes that ‘read’ these PTMs. High-quality chemical probes that can block reader functions of proteins involved in chromatin regulation are important tools to improve our understanding of pathways involved in chromatin dynamics. Insight into the intricate system of chromatin PTMs and their context within the epigenome is also therapeutically important as misregulation of this complex system is implicated in numerous human diseases. Using computational methods, along with structure-based knowledge, we have designed and constructed a focused DNA-Encoded Library (DEL) containing approximately 60,000 compounds targeting bi-valent methyl-lysine (Kme) reader domains. Additionally, we have constructed DNA-barcoded control compounds to allow optimization of selection conditions using a model Kme reader domain. We anticipate that this target-class focused approach will serve as a new method for rapid discovery of inhibitors for multivalent chromatin reader domains.

Gene silencing via short hairpin mediated RNAi (shRNA) is a valuable experimental tool and has promise as a therapeutic strategy. Several shRNA platforms make use of the loop and flanking sequences from the endogenous microRNA (miRNAs) miR-30a or other miRNAs to provide an RNA structure for efficient and accurate biogenesis of the RNA trigger. However, the stem regions of these shRNAs are typically designed as perfect duplex structures which is an uncommon feature for endogenous miRNA precursors. A limitation of these designs is that shRNAs with perfect duplex stems undergo extensive stem cleavage analogous to the Dicer independent miRNA miR-451, destroying the shRNA trigger sequence that is present in the 3P arm. We employed an unbiased screen of > 9000 shRNA structures to identify features that prevent stem cleavage and promote canonical biogenesis and loading into the effector complex RISC. We find that a central stem bulge or kink reduces central stem cleavage and improves accuracy of Dicer processing. Furthermore, 9 -10 GC nucleotides in the guide strand improves shRNA efficiency. These design rules enable more effective shRNA tools and are compatible with existing sets of optimized guide/target sequences.