AbstractEpigenetics maintains cell-identity specific gene-expression patterns. However, within a population of isogenic cells of the same identity, a substantial variability in gene expression and responsiveness is still observed. Transcription bursting is a substantial source of this gene-expression variability or ‘noise’, contributing to phenotypic heterogeneity and potentially driving both physiological and pathological processes such as differentiation or tumorigenesis and drug resistance. Identification of transcription-bursting dynamics at a genome-wide scale has been restricted to inferring bursts in mRNA production computationally from the heterogeneity of mRNA levels in single cell transcriptomic data. Systematic characterisation of the genomic and epigenetic chromatin context of genes with defined transcription bursting behaviour has been incomplete. Here, we measured the bursting of transcription itself by genome-wide nascent RNA sequencing of breast cancer MCF-7 cells upon synchronisation of transcription with a transcription elongation inhibitor and by calibration using live cell imaging of nascent PP7-tagged GREB1 transcription. Comparing across the entire genome, we find transcription bursting to be ubiquitous, with burst sizes of up to 160 transcripts. Transcription bursting attributes ~85% to a trend in the variation in steady state gene expression between genes, whereas both burst frequency and nascent transcript degradation attribute minimally. Individual genes deviate strongly from this trend and engage both in anomalous burst size and frequency. We find that the presence of the TATA box or Inr sequence within gene promoters significantly predicts a larger burst size, as do promoter-associated YY1 and E2F1 transcription-factor binding motifs. Enrichment of the transcription start site with epigenetic marks such as H3K79me2 and H3Kl4ac is also strongly associated with the transcription burst size. Finally, we show that in these MCF-7 breast-cancer cells, genes with a larger transcription burst size exhibit a larger immediate transcriptional response following endocrine drug treatment. Our genome-wide transcription-bursting analysis method paves the way to elucidate the dynamic role of epigenetic regulation on dynamic transcription in pathophysiology.
Genome-wide lentiviral shRNA screen identifies serine/arginine-rich splicing factor 2 as a determinant of oncolytic virus activity in breast cancer cells
