Sensitive dissection of a genomic regulatory landscape using bulk and targeted single-cell activation

Abstract

Transcriptional enhancers are non-coding DNA elements that regulate gene transcription in a temporal and tissue-specific manner. Despite advances in computational and experimental methods, identifying enhancers and their target genes essential for specific biological processes remains challenging. Determining target genes for enhancers is also complex and often relies on indirect, low-resolution, and/or assumptive methodologies. To identify and functionally perturb enhancers at their endogenous sites without altering their sequence, we performed a pooled tiling CRISPR activation (CRISPRa) screen surrounding PHOX2B, a master regulator of neuronal cell fate and a key player in neuroblastoma development. This screen allowed the de novo identification of CRISPRa responsive elements (CaREs) that alter cellular growth within the 2 Mb genomic region. To determine CaRE target genes, we developed TESLA-seq (TargEted SingLe cell Activation), which combines CRISPRa screening with targeted single-cell RNA-sequencing and enables the parallel readout of the effect of hundreds of enhancers on all genes in the locus. While most TESLA-revealed CaRE-gene relationships involved neuroblastoma-related regulatory elements already active in the system, we found many CaREs and target connections normally active only in other tissue types or with no previous evidence and induced out of context by CRISPRa. This highlights the power of TESLA-seq to reveal gene regulatory networks, including edges active outside of a given experimental system. HighlightsO_LISystematically perturbed regulatory landscape in a 2 Mb genomic region surrounding PHOX2B to identify hundreds of CRISPRa-responsive elements that affect cellular growth C_LIO_LIDeveloped TESLA-seq as a principled molecular approach to find gene targets of dozens of candidate regulatory elements C_LIO_LIValidated interactions between identified regulatory elements and target genes and characterized their genomic features C_LIO_LIIntegrated a compendium of epigenomic datasets to identify regulatory relationships induced out of context or with no previous evidence C_LI