Abstract

Controlling gene expression in human cells is essential for interrogating many biological questions and for developing new therapies. Fusions between programmable DNA binding domains (eg. CRISPR) and different transcription factors or chromatin regulators can be targeted to specific genomic locations to change gene expression and chromatin organization. However, expressing fusions to full-length transcriptional or chromatin regulators makes delivery of these tools inefficient in human cells, especially when multiple regulators are needed to achieve efficient, fast, or long-term control. Moreover, we need to systematically test regulators in different contexts to determine how gene control varies with locus and cell type. We have developed a high-throughput method for testing tens of thousands of compact effector domain candidates (80aa each) for their effects on gene activation and silencing in human cells either at reporter genes (HT-recruit, Tycko et al, Cell 2020) or at endogenous genes and enhancers (CRISPR HT-recruit, unpublished). In this assay, pooled libraries of protein domains are delivered to human cells such that a different domain is recruited to the target gene in each cell. Cells are then magnetically separated based on target gene expression, and the transcriptional effect of each candidate regulator is measured by counting with NGS how often the regulator is found in cells with high versus low target gene expression. We found that some domains can have opposing effects at different promoters, and used context-independent effectors to engineer better synthetic repressors and activators. We developed an improved CRISPR interference tool containing the KRAB domain from ZNF705F, and showed that it can silence promoters and enhancers more efficiently than the standard CRISPRi containing the ZNF10 KRAB. Using combinations of activator domains from NCOA3, FOXO3 and ZNF473, we developed a compact human activator combination (NFZ, 145aa) for better CRISPR activation. These tools are useful for individual gene perturbations, high-throughput CRISPRi/a functional screens, and therapeutic applications where high potency and small size are essential.