Implementing a high-throughput parallel CRISPRi screening platform to identify functional lncRNAs

Technological advances in RNA-sequencing revealed that the human genome is pervasively transcribed, resulting in the production of thousands of long non-coding RNAs (lncRNAs). Several lncRNAs are now recognized as key components of diverse physiological processes. However, molecular genetics lacks a more comprehensive view of lncRNome functionality and the mechanisms through which lncRNAs operate. Current high-throughput approaches to study lncRNA function (i.e. pooled CRISPR library screens) are typically limited to a single cellular phenotype or dedicated molecular reporter based on which functional candidates are selected. Here we present a scalable platform enabling serial cellular and molecular phenotyping to catalog lncRNA functions in a high-throughput and arrayed approach using CRISPR interference (CRISPRi). We applied PCR, in vitro transcription and bead-based purification for high-throughput production of single gRNAs in 96-well plate format. Using lipid-based transfection, sgRNAs were delivered to HEK293T cells stably expressing a deficient Cas9 protein fused to repressive complexes (dCas9-KRAB-MeCP2). Cells were monitored in real-time using the Incucyte platform to quantify growth, proliferation and apoptosis. After 48 hours, cells were lysed and RNA-sequencing libraries were generated directly from crude cell lysates, followed by shallow RNA-sequencing to infer the molecular phenotype associated to each condition. A proof-of-concept screen including 10 sgRNAs for 20 lncRNA targets demonstrated the feasibility of our approach, revealing differentially expressed genes and pathways upon lncRNA knockdown. Subsequently, we have initiated the systematic silencing of over 300 lncRNAs through this platform in order to characterize their associated cellular and molecular phenotypes. Additional dCas9-KRAB-MeCP2 models are being generated to probe the functionality of the long non-coding transcriptome in various disease-relevant model systems.