NELF Focuses Sites of Initiation and Maintains Promoter Architecture

Many factors control the elongation phase of transcription by RNA polymerase II (Pol II), a process that plays an essential role in regulating gene expression. We utilized cells expressing degradation tagged subunits of NELFB, PAF1 and RTF1 to probe the effects of depletion of the factors on nascent transcripts using PRO-Seq and on chromatin architecture using DFF-ChIP. Although NELF is involved in promoter proximal pausing, depletion of NELFB had only a minimal effect on the level of paused transcripts and almost no effect on control of productive elongation. Instead, NELF depletion increased the utilization of downstream transcription start sites and caused a dramatic, genome-wide loss of H3K4me3 marked nucleosomes. Depletion of PAF1 and RTF1 both had major effects on productive transcript elongation in gene bodies and also caused initiation site changes like those seen with NELFB depletion. Our study confirmed that the first nucleosome encountered during initiation and early elongation is highly positioned with respect to the major TSS. In contrast, the positions of H3K4me3 marked nucleosomes in promoter regions are heterogeneous and are influenced by transcription. We propose a model defining NELF function and a general role of the H3K4me3 modification in blocking transcription initiation. Overall design: To investigate the function of NELFB, PAF1, and RTF1, we created individual cell lines in which the endogenous locus for each of these factors were tagged with FKBP12^F36V. This allowed us to deplete rapidly and selectively each of these factors upon the addition of 400 nM dTAGV-1. NELFB was depleted for 2h and 48h in the presence or absence of 1µM Flavopiridol, RTF1 was depleted for 2h, and PAF1 for 48h. We then performed PRO-Seq to quantitatively determine genome-wide levels of nascent transcripts. We also performed Pol-II and H3K4me3 DFF-ChIP-Seq on NELFB control and dTAGV-1 treated cells. For the Pol-II experiment, we depleted cells for 2h, while for the H3K4me3 experiment, we depleted cells for 2h and 48h. Finally, we profiled H3K36me3 genome-wide in control NELFB cells with DFF-ChIP-Seq. The Sf21 moth and 4T1 mouse cell lines were used for spike-in controls for PRO-Seq and DFF-Seq, respectively. Two biological replicates per sample were sequenced twice (two different runs).