The PNUTS phosphatase complex controls transcription pause release

Abstract

Gene expression is regulated by controlling distinct steps of the transcriptional cycle, including initiation, pausing, elongation, and termination. Kinases phosphorylate RNA polymerase II (RNA Pol II) and associated factors to control transitions between these steps and to act as central gene regulatory nodes. Similarly, phosphatases that dephosphorylate these components are emerging as important regulators of transcription, although their roles remain less well understood. Here, we discover that the mouse PNUTS-PP1 phosphatase complex plays an essential role in controlling transcription pause release in addition to its previously described function in transcription termination. Transcription pause release by the PNUTS complex is essential for almost all RNA Pol II-dependent gene transcription, relies on its PP1 phosphatase subunit, and controls the phosphorylation of factors required for pause release and elongation. Together, these observations reveal an essential new role for a phosphatase complex in transcription pause release and show that the PNUTS complex is essential for RNA Pol II-dependent transcription.

Keywords: PNUTS; PP1 phosphatase; RNA Pol II; pause release; termination; transcription.

Figure 1:. The PNUTS complex is essential for RNA Pol II-dependent transcription

(A) A schematic illustrating the PNUTS-dTAG depletion approach. (B) Western blot for PNUTS complex components following immunoprecipitation (IP) with PNUTS or control antibody, in WT or PNUTS-dTAG ESCs. HDAC1 is a loading and negative control for IPs. (C) Western blot for PNUTS complex components in wild-type (WT) ESCs and PNUTS-dTAG ESCs before (UNT) and after dTAG13 treatment. HDAC1 is a loading control. (D) A snapshot of cTT-seq signal in UNT and dTAG13-treated PNUTS-dTAG ESCs. Read-through transcription beyond the TES after PNUTS depletion is indicated. (E) A box plot illustrating the transcription read-through index in UNT and dTAG13-treated PNUTS-dTAG ESCs. (F) A snapshot of cTT-seq signal illustrating major reductions in transcription after PNUTS depletion (dTAG13). (G) A metaplot of cTT-seq signal across all active genes (n= 11790) in UNT and PNUTS-depleted cells (dTAG13). (H) Heatmaps illustrating cTT-seq signal at active genes (n= 11790) in UNT and dTAG13-treated PNUTS-dTAG cells and a differential heatmap showing the log2 fold change (Log2 FC) in cTT-seq signal upon PNUTS depletion (right). Genes sorted by decreasing transcription in UNT cells. (I) A MA-plot illustrating the log2 fold change in cTT-seq signal after PNUTS depletion. Significant reductions in transcription (fold-change <−1.5 and p-value <0.05) are coloured red (n=11350) and significant increases (fold-change >1.5 and p-value >0.05) are coloured blue (n=3138). (J) A snapshot of cTT-seq signal illustrating reductions in upstream antisense transcription after PNUTS depletion. (K) A metaplot of cTT-seq signal in the upstream antisense direction of all active genes (n=11790) in untreated cells (UNT) and after PNUTS depletion (dTAG13).

Figure 2:. Entry into productive elongation relies on PNUTS

(A) A schematic illustrating endogenously Halo-tagged RNA Pol II (HT-RPB1) labelling with Halo ligand for single particle tracking. (B) Western blot for RPB1 and PNUTS in WT ESCs and PNUTS-dTAG;HT-RPB1 ESCs before (UNT) and after dTAG13 treatment. HDAC1 is a loading control. (C) A representative image of HT-RPB1 in live cells demonstrating appropriate nuclear localisation. (D) Example single molecule tracks for HT-RPB1. (E) Apparent diffusion coefficient histograms for HT-RPB1 in either UNT or PNUTS-depleted (dTAG13) cells (n=3). (F) The bound fraction of HT-RPB1 in UNT or PNUTS-depleted (dTAG13) cells. Dots represent individual biological replicates (n=3), black horizontal lines show mean values, and error bars represent standard deviation. P-value corresponds to a one-sided t-test. (G) A snapshot of cChIP-seq signal for total RNA Pol II and PNUTS in UNT and PNUTS-depleted cells (dTAG13). (H) Metaplots of cChIP-seq signal for total RNA Pol II and PNUTS across all active genes (n=11790) in UNT and PNUTS-depleted cells (dTAG13). (I) Heatmaps of cChIP-seq signal for total RNA Pol II and PNUTS across all active genes (n=11790) in UNT and PNUTS-depleted cells (dTAG13), sorted by decreasing RNA Pol II signal in UNT cells. A differential heatmap of log2 fold change (Log2 FC) between RNA Pol II cChIP-seq signal in dTAG13 and UNT cells is shown.

Figure 3:. PNUTS is essential for transcription pause release

(A) A schematic illustrating factors examined by cChIP-seq corresponding to distinct phases of transcription. (B) Snapshots (left) and metaplots (right) across all active genes (n=11790) for Ser5P and Ser2P RNA Pol II cChIP-seq in untreated (UNT) and PNUTS-depleted cells (dTAG13). (C) As (B) but for SPT5 and TFIIS. (D) As (B) but for NELF-B and INTS11. (E) As (B) but for PAF1 and SPT6. (F) As (B) but for BRD4 and CDK9.

Figure 4:. PNUTS interaction with PP1 is essential for exit from promoter-proximal pausing

(A) A schematic illustrating PNUTS WT and PNUTS W401A rescue systems where cells are treated with doxycycline (DOX) at 0hr to induce expression of PNUTS WT or PNUTS W401A, and treated with dTAG13 at 2hr to rapidly deplete endogenous PNUTS-dTAG. Cells were harvested for analysis at 4hr. (B) PNUTS immunoprecipitation (T7-IP) was performed from untreated (UNT) or rescued (Rescue) WT or W401A cells. Input (left panel) and T7 IP material (right panel) was subjected to western blot for PNUTS complex components. HDAC1 is loading and negative control for IP. (C) A snapshot showing cTT-seq signal in the untreated (UNT), PNUTS-depleted (dTAG13), and PNUTS WT and PNUTS W401A rescue lines. (D) A metaplot of cTT-seq signal over all active genes (n=11790) in the UNT, PNUTS-depleted (dTAG13), and PNUTS WT rescue ESCs. (E) As in (D) but for PNUTS W401A rescue ESCs. (F) A snapshot illustrating cChIP-seq signal for RNA Pol II and PNUTS following PNUTS WT or W401A rescue. (G) A metaplot of RNA Pol II cChIP-seq signal over all active genes (n=11790) following PNUTS WT or W401A rescue. (H) As in (G) but for PNUTS cChIP-seq.

Figure 5:. PNUTS regulates the phosphorylation of pause release and elongation factors

(A) A schematic illustrating the workflow for quantitative proteomics. (B) Volcano plots illustrating the log2 fold change in nuclear protein abundance after 30 min and 2 hrs PNUTS depletion (dTAG13). Blue dots correspond to proteins significantly reduced (fold change <−1.5 and p-value <0.05) and red dots correspond to proteins significantly increased (fold change >1.5 and p-value <0.05) in abundance. Components of the PNUTS complex are labelled. (C) A schematic illustrating the workflow for quantitative phosphoproteomics. (D) Volcano plots illustrating the log2 fold change in nuclear protein-corrected phosphoproteomics after 30 min and 2 hrs PNUTS depletion. Blue dots correspond to phosphopeptides significantly reduced (fold change <−1.5 and p-value <0.05) and red dots correspond to phosphopeptides significantly increased (fold change >1.5 and p-value <0.05) in abundance. Significantly changing phosphopeptides of potential interest are labelled. (E) GO term analysis of proteins that show increases in phosphorylation after PNUTS depletion (n=76). Counts indicates the number of proteins with increased phosphorylation within each GO term group. Associated p-values are shown on the right. (F) A Coomassie-stained gel of the reconstituted and purified PNUTS complex. (G) An in vitro phosphatase activity assay using purified PNUTS complex. Peptides phosphorylated at the indicated residues were used as substrates and the amount of free phosphate released was measured. Error bars represent standard deviation (n=4).

Figure 6:. The PNUTS TND guides interactions with substrates

(A) Top: a schematic of the PNUTS protein with key domains labelled. Bottom: AlphaFold prediction of the structure of PNUTS HEAT-TND/TOX4 complex interacting with the PAF1 TIM (orange). Apparent TIM-interacting residues within the PNUTS TND (mutated in PNUTS TND-M) are indicated. (B) A schematic illustrating the PNUTS WT and PNUTS TND-M rescue systems. (C) PNUTS immunoprecipitation (IP-T7) was performed from untreated (UNT) or rescued (Rescue) PNUTS WT or TND-M ESCs. Input (left panel) and T7-IP material (right panel) was subjected to western blot using antibodies specific for components of the PNUTS complex. HDAC1 is a loading and negative control for the T7-IP. (D) A volcano plot illustrating the log2 fold enrichment in protein interactions identified by IP-MS for PNUTS WT rescue compared to uninduced control (UNT). Proteins that are significantly enriched (fold change >1.5 and p-value <0.05) are indicated by larger circles and coloured according to the protein category they belong to. (E) A volcano plot showing depletion of PNUTS-interacting proteins following PNUTS TND-M rescue (log2 fold change TND-M/WT). PNUTS-interacting proteins are coloured as in (D). (F) Heatmaps showing log2 fold change in PNUTS-interacting proteins (TND-M/WT) as defined in (D), separated into groups based on their TND dependency.

Figure 7:. The PNUTS TND is essential for pause release and transcription

(A) A snapshot illustrating cTT-seq signal in the UNT, PNUTS-depleted (dTAG13), and PNUTS rescue conditions (WT or TND-M). (B) A metaplot of cTT-seq signal across all active genes (n=11790) in the UNT, PNUTS-depleted (dTAG13), and PNUTS WT rescue ESCs. (C) As in (B) but for PNUTS TND-M rescue ESCs. (D) A snapshot illustrating cChIP-seq signal for total RNA Pol II and PNUTS in the PNUTS WT or TND-M rescue ESCs. (E) A metaplot of total RNA Pol II cChIP-seq signal over all active genes (n=11790) in the PNUTS WT or TND-M rescue ESCs. (F) As in (E) but for PNUTS cChIP-seq.