Determining Long Noncoding Rna Mechanisms In Stem Cells Using Automated Image Analysis And Single Molecule Rna-Fluorescent In Situ Hybridization

Long noncoding RNAs (lncRNAs) are increasingly recognized as involved in human physiology and disease. Antisense lncRNAs are ubiquitous in the human genome and have many mechanisms of the action proposed, but the single-cell and single-molecule information that is needed to support or refute these mechanisms are lacking. Although single molecule RNA-Fluorescent in Situ Hybridization (RNA-FISH) is a single molecule method sometimes used to investigate lncRNAs, single-molecule counts and the spatial localization of these RNAs are often ignored. One reason is a lack of easily implemented and broadly applicable experimental and computational approaches for utilizing this data. We propose an RNA-FISH-based quantitative approach to test different mechanisms of gene regulation by antisense transcription. We applied our approach to investigate the mammalian antisense lncRNA pair Tsix and Xist known to have inhibitory interactions. by utilizing two-color, single-molecule RNA-FISH and quantitative image acquisition and analysis pipelines, we were able to quantify Tsix and Xist at single-molecule resolution in more than 8,000 cells including several biological replicas. Our analysis revealed that there was no significant colocalization of Xist and Tsix molecules, providing evidence against mechanisms of inhibition involving the binding of the two transcripts. We also found that anticorrelation between Xist and Tsix only occurred at sites of transcription, which suggests mechanisms of inhibition in cis based on the act of transcription. Surprisingly, our analysis of RNA-FISH data utilizing intronic probes provided evidence against transcriptional interference playing a significant role. Colocalization of transcription sites with histone marks elucidated possible relationships between active transcription and chromatin state. This data demonstrates the importance of single-molecule analysis when investigating noncoding RNA interactions. Our generalizable pipeline can be applied to transcripts across diverse species, cell types, and imaging modalities.