SPT6 functions in transcriptional pause/release via PAF1C recruitment

Abstract

It is unclear how various factors functioning in the transcriptional elongation by RNA polymerase II (RNA Pol II) cooperatively regulate pause/release and productive elongation in living cells. Using an acute protein-depletion approach, we report that SPT6 depletion results in the release of paused RNA Pol II into gene bodies through an impaired recruitment of PAF1C. Short genes demonstrate a release with increased mature transcripts, whereas long genes are released but fail to yield mature transcripts, due to a reduced processivity resulting from both SPT6 and PAF1C loss. Unexpectedly, SPT6 depletion causes an association of NELF with the elongating RNA Pol II on gene bodies, without any observed functional significance on transcriptional elongation pattern, arguing against a role for NELF in keeping RNA Pol II in the paused state. Furthermore, SPT6 depletion impairs heat-shock-induced pausing, pointing to a role for SPT6 in regulating RNA Pol II pause/release through PAF1C recruitment.

Keywords: NELF; PAF1; RNA polymerase II; SPT6; chromatin; elongation; gene expression; pause/release; promoter-proximal pausing; transcription.

Figure 1.. Rapid SPT6 depletion leads to increased RNA Pol II occupancy downstream of pause sites.

(A) Western blots of whole-cell lysates show SPT6 protein depletion by auxin-inducible degradation (AID). Parental or SPT6-AID DLD-1 cells were treated with 500 μM auxin for the indicated time. RPB1N blot shows the total levels of RNA Pol II subunit RPB1. Actin serves as a loading control. (B) Representative track example showing ChIP-seq signal for total RNA Pol II and SPT6 in SPT6-AID cells treated with auxin (500 μM, 2h). (C) ChIP-seq signal for total RNA Pol II and SPT6 around promoters, centered on pause sites. SPT6-AID cells were treated as in (B). Rows are sorted by RNA Pol II occupancy in the auxin (–) sample. RRPM, reference-adjusted reads per million; Log2FC, log2 fold change (+Auxin/−Auxin). N = 6,481 genes. (D) Distribution of pause-release ratio in SPT6-AID cells treated as in (B). *p < 0.001: KS test. N = 6,481 genes.

Figure 2.. Loss of SPT6 causes both RNA Pol II release and loss of elongation processivity.

(A) Schematic diagram describing dissection of RNA Pol II release and loss of processivity by measuring poly(A)+ mRNA. (B) Scatter plot showing log2 fold change of poly(A)+ mRNA levels versus gene length in SPT6-AID cells treated with auxin (500 μM, 6h). Differentially expressed genes (orange dots, adjusted p-value < 0.1) were identified from spike-in normalized RNA-seq data. Regression line (blue) and Spearman’s correlation coefficient 𝜌 are shown. n = 2 independent replicates, N = 25,715 transcripts. (C) Clustering of differentially expressed genes identified in (B). Top 25% short or top 25% long genes are indicated at the right. Z, z scores. n = 2 independent replicates, N = 7,459 genes. (D) ChIP-seq track examples for short and long genes in SPT6-AID cells treated with auxin (500 μM, 2h). Data in Figure 1B is shown again for the comparison. (E) Log2 fold change (+Auxin/−Auxin) of RNA Pol II ChIP signal, centered on gene start sites. SPT6-AID cells were treated as in (D). Rows are sorted by gene length. Dashed line indicates 3’ ends of genes. N = 6,481 genes.

Figure 3.. SPT6 loss causes decreased PAF1C-RNA Pol II interaction and increased NELF-RNA Pol II interaction.

(A) Scatter plot showing log2 fold change versus average abundance of RNA Pol II-pull downs detected by mass spectrometry in SPT6-AID cells treated with auxin (500 μM, 2 h). Differentially associated proteins (orange dots, adjusted p-value < 0.1 with > 2-fold change) and RNA Pol II subunits (dark gray) are shown. Dashed lines indicate 2-fold change. n = 6 independent replicates. (B) Log2 fold change of RNA Pol II-interacting complexes. Each dot indicates a subunit of each complex detected by mass spectrometry in SPT6-AID cells. (C) Gene set enrichment analysis of reactome pathway.

Figure 4.. SPT6 promotes the recruitment of PAF1C to elongating RNA Pol II.

(A) Representative ChIP-seq track examples in SPT6-AID cells treated with auxin (500 μM, 2h). Tracks for PAF1 complex (PAF1, LEO1, WDR61, CDC73, RTF1), total RNA Pol II, and SPT6 are shown. (B) ChIP-seq signal coverage at gene bodies (pause sites +150 bp to + 3kb) in SPT6-AID cells treated as in (A). Signal is normalized to total RNA Pol II ChIP signal. *p < 0.001: U test. N = 6,481 genes. (C) ChIP-seq mean signal around 3’ gene ends for RTF1, PAF1, and SPT6 in untreated SPT6-AID cells. Signal is centered on and normalized to 3’ gene ends. N = 6,481 genes. (D) Western blots of whole-cell lysates show double depletion of SPT6-AID and PAF1-AID proteins. Parental DLD-1, SPT6-AID, and SPT6-AID PAF1-AID cells were treated with auxin (500 μM, 2h). (E) ChIP-seq signal for total RNA Pol II around promoters, centered on pause sites. SPT6-AID PAF1-AID cells were treated with auxin (500 μM, 2h). Rows are sorted by RNA Pol II occupancy in the auxin (–) sample. N = 6,481 genes.

Figure 5.. SPT6 loss causes aberrant association of NELF with elongating RNA Pol II at promoters and gene bodies.

(A) Representative NELF-C ChIP-seq track in SPT6-AID cells treated with auxin (500 μM, 2h). PAF1, SPT6, and total RNA Pol II tracks are shown again for the comparison. (B) ChIP-seq signal at pause sites (± 150 bp) and gene bodies (pause sites +150 bp to + 3kb) in SPT6-AID cells treated as in (A). Signal is normalized to total RNA Pol II signal. *p < 0.001: U test. N = 6,481 genes. (C) Western blots of whole-cell lysates show double depletion of SPT6-AID and NELF-C-AID proteins. Parental DLD-1, SPT6-AID, and SPT6-AID NELF-C-AID cells were treated with auxin (500 μM, 2h). (D) ChIP-seq signal for total RNA Pol II around promoters, centered on pause sites. SPT6-AID NELF-C-AID cells were treated with auxin (500 μM, 2h). Rows are sorted by RNA Pol II occupancy in the auxin (–) sample. N = 6,481 genes.

Figure 6.. SPT6 is required for RNA Pol II pausing induced by the heat shock stress.

(A) Scatter plot showing log2 fold change versus average signal of PRO-seq signal in DLD-1 cells treated with heat shock (42°C, 1h). Orange dots indicate differentially expressed genes (adjusted p-value < 0.1). NHS, non-heat shock; HS, heat shock. n = 4 independent replicates. N = 6,481 genes. (B) Representative PRO-seq track at a HS-downregulated gene in SPT6-AID cells treated with auxin (500 μM, 2h), followed by heat shock (42°C, 1h). (C) PRO-seq signal around promoters of HS-downregulated genes, centered on pause sites. SPT6-AID cells were treated as in (B). Solid line, bootstrapped mean; shade, 75% confidence interval. N = 4,429 genes. The inset shows an enlarged view around pause sites. (D) Distribution of pause-release ratio in SPT6-AID cells treated as in (B). *p < 0.001: KS test. N = 4,429 genes.

Figure 7.. Model: SPT6 regulates pause/release via PAF1C recruitment and NELF removal.

Top: SPT6 facilitates the recruitment of PAF1 complex and removal of NELF complex. This leads to the release of RNA Pol II into gene bodies with high processivity. Note that SPT6 may indirectly regulate this NELF-PAF1C exchange, mediated through SPT6-interacting factors. Other factors such as MYC and P-TEFb may also be involved in this step. Bottom: In the absence of SPT6, RNA Pol II fails to recruit PAF1C and remains associated with NELF. This NELF-bound RNA Pol II is released into gene bodies with an elongation defect. Protein structures are adapted from (Vos et al., 2018a, 2018b), PDB: 6GML, 6GMH.