(19f) Engineering Improved Crispr Repressors for Targeted Mammalian Gene Regulation

The ability to precisely silence endogenous gene expression is vital for performing robust whole-genome screens and discovering non-coding transcriptional regulatory motifs in mammalian cells. CRISPR interference (CRISPRi) technology has emerged as a powerful tool enabling site-specific transcriptional repression for these applications. The CRISPRi platform employs the RNA-guided, catalytically inactive endonuclease, dCas9, to deliver conjoined transcriptional repressor domains near to the transcription start site of target genes. Almost all CRISPRi systems implement direct fusions of dCas9 with the Krüppel-associated box (KRAB) domain from the human protein KOX1, often referred to only as KRAB, but here called KOX1-KRAB to prevent ambiguity. Previous studies have demonstrated improved CRISPRi transcriptional repression by combining KOX1-KRAB with additional repressor domains or by utilizing KRAB domains from other human proteins. Despite this progress, the CRISPRi platform still suffers from incomplete repression and high performance variability across cell lines, as well as a general overreliance on KOX1-KRAB.

In order to improve the CRISPRi platform in mammalian cells, we sought to engineer more efficient repressors by discovering novel dCas9-repressor fusions, characterizing their transcriptional mechanisms of action, and incorporating this insight to construct highly efficient repressors that incorporate multiple modalities to silence local gene expression. We first screened single putative transcriptional repressor domains using a reporter assay in HEK293T cells and identified eleven novel protein domains compatible with the CRISPRi system, including one domain that outperformed KOX1-KRAB (Figure 1).

Next, we computationally quantified gene expression patterns of regulatory co-factors that interact with each novel repressor domain (as well as KOX1-KRAB) and discovered that numerous interaction partners for repressor domains have highly variable expression levels across human cell lines. To further investigate the generalizability of our novel repressors, we compared performance of individual repressors across three diverse human cell lines, and we are further assembling and screening bipartite repressor domains to evaluate potential synergistic activity when two domains, with diverse interaction co-factors, are directly fused together in conjunction with dCas9. We envision these efforts will mark a significant step forward in applying the CRISPRi platform to regulate transcription in human cells from diverse lineages.