Deconstructing lincRNA regulation during ESC to NPC differentiation. Deconstructing lincRNA regulation during ESC to NPC differentiation

Using single-cell and bulk RNA-sequencing, we profiled mouse embryonic stem cells and neural precursor cells to identify lincRNAs with cell type- and subpopulation-specific expression. In contrast to mRNAs, we found that lincRNAs are prone to prematurely terminate transcription, are less abundant across single cells, and are degraded by the nuclear exosome. These properties indicate that lincRNA genes widely function without requiring their RNA product. To test this prediction experimentally, we developed a method that separates the role of a lincRNA gene from the role of its RNA product. This revealed RNA-independent lincRNA gene activities, underscoring the importance of distinguishing the roles of transcription, the genomic locus, and the RNA product when studying non-coding loci, for which we are providing suitable tools. Overall design: Two mESC lines (BL6xCAST and CASTxBL6) were differentiated into neural precursor cells, as well as being sampled in 2i+LIF and in serum+LIF medium. Bulk RNA-seq was performed for all timepoints (2i+LIF, serum+LIF, day3, day6 and day8 of differentiation). Most timepoints were also sampled by single-cell RNA-seq (96 cells per sample, 7 samples in total). Cell lines with lincRNA genomic deletions, or with the hammerhead ribozyme genomically integrated into lincRNA loci, were also analysed by bulk RNA-seq (140 cell lines in total). We also performed CRosslinking and Analysis of CDNAs (CRAC) in wild-type cells or cells where endogenous Mtr4 is tagged with a FLAG-Bio tag, to examine transcriptome-wide targets of the nuclear RNA surveillance factor Mtr4. For normalisation, we performed NET-seq. To measure global RNA half-lives, we treated wild-type mESCs with actinomycin D, and performed bulk RNA-seq at several timepoints, in triplicate (3 wells of cells were analysed in parallel per timepoint).