CDK12 globally stimulates RNA polymerase II transcription elongation and carboxyl-terminal domain phosphorylation

Abstract

Cyclin-dependent kinase 12 (CDK12) phosphorylates the carboxyl-terminal domain (CTD) of RNA polymerase II (pol II) but its roles in transcription beyond the expression of DNA damage response genes remain unclear. Here, we have used TT-seq and mNET-seq to monitor the direct effects of rapid CDK12 inhibition on transcription activity and CTD phosphorylation in human cells. CDK12 inhibition causes a genome-wide defect in transcription elongation and a global reduction of CTD Ser2 and Ser5 phosphorylation. The elongation defect is explained by the loss of the elongation factors LEO1 and CDC73, part of PAF1 complex, and SPT6 from the newly-elongating pol II. Our results indicate that CDK12 is a general activator of pol II transcription elongation and indicate that it targets both Ser2 and Ser5 residues of the pol II CTD.

Figure 1.

CDK12^as inhibition globally decreases RNA synthesis. (A) Experimental design. (B) Metagene analysis of TT-seq signal for expressed genes after treatment of cells with DMSO (black) versus 7.5 μM NM treatment (blue) for 15 min (left) or 30 min (right). The TT-seq coverage was averaged and aligned at their transcription start sites (TSSs) and polyadenylation (pA)-sites. Shaded areas around the average signal (solid lines) indicate confidential intervals. (C) Box plots of different length classes show the ratio of transcribed bases after 15 min (left) and 30 min (right) inhibition of CDK12^as compared to control. (D) Metagene analysis for different length classes comparing the average TT-seq signal before (DMSO treatment, black) and after CDK12^as inhibition (7.5 μM NM treatment, blue) for 15 min (upper panel) and 30 min (lower panel). The TT-seq coverage was averaged and aligned at their transcription start sites (TSSs) and polyadenylation (pA)-sites. Shaded areas around the average signal (solid lines) indicate confidential intervals. (E) Schematic representation of TT-seq signal changes along the gene body upon elongation velocity change. Upper panel: steady state transcription. Lower panels show TT-seq signal recovery spreading from the TSS.

Figure 2.

Inhibition of CDK12^as affects transcription elongation. (A) Experimental design. mNET-seq using an antibody against total pol II was performed in CDK12^as HEK293 cells after treatment with 15 min of DMSO solvent control or 7.5 μM NM. (B) Metagene analysis comparing the average mNET-seq signal before and after CDK12^as inhibition for 15 min of expressed genes. (C) Violin plots of the different length classes noted show the ratio of mNET-seq signal after 15 min inhibition of CDK12^as compared to control. (D) Metagene analysis for the different length classes noted comparing the average mNET-seq signal before and after CDK12^as inhibition for 15 min. The mNET-seq reads were averaged and aligned at their TSSs and (pA)-sites.

Figure 3.

CDK12 phosphorylates transcribing pol II. (A) Meta-analyses of scaled expressed genes of mNET-seq for Ser2P with and without normalization to pol II after treatment of CDK12^as cells with DMSO (black) or 7.5μM NM (orange) for 15 min. (B) Meta-analyses of scaled expressed genes of mNET-seq for Ser5P with and without normalization to pol II after treatment of CDK12^as cells with DMSO (black) or 7.5 μM NM (red) for 15 min. (C) mNET-seq profiles across CCND2 using a total pol II antibody, two Ser2P antibodies and one Ser5P antibody.

Figure 4.

ChIP-seq analysis of CDK12 phosphorylation of transcribing pol II. (A) Meta-analyses of scaled expressed genes of ChIP-seq for Ser2P (ab5095) with and without normalization to pol II after treatment of CDK12^as cells with DMSO (black) or 7.5μM NM (orange) for 15 min. (B) Meta-analyses of scaled expressed genes of ChIP-seq for Ser5P (ab5131) with and without normalization to pol II after treatment of CDK12^as cells with DMSO (black) or 7.5 μM NM (red) for 15 min.

Figure 5.

CDK12 activity is required for stable association of elongation factors with chromatin. (A) Meta-analysis of LEO1 subunit of the PAF1 complex (PAF1C) ChIP-seq across scaled expressed genes. (B) Meta-analysis of SPT6 ChIP-seq across scaled expressed genes. (C) Meta-analyses of LEO1 ChIP-seq ratioed to the total pol II signal before and after CDK12^as inhibition for 15 min across scaled expressed genes for the different length classes noted. (D) Meta-analyses of SPT6 ChIP-seq ratioed to the total pol II signal before and after CDK12^as inhibition for 15 min across scaled expressed genes for the different length classes noted. (E) Western blots of CDK12^as cell chromatin extracts. Cells are either treated with 7.5 μM 1-NM-PP1 or DMSO for 15 and 30 min as noted. The antibodies used are indicated on the right. Histone H3 was used as a loading control. (F) Quantitation of chromatin protein levels of SPT6, LEO1, CDC73, and histone H3 in CDK12^as cells relative to DMSO controls. Error bars = s.e.m. (n = 3 biological replicates). Statistical test: two-tailed unpaired t test, ns = not significant, * P < 0.05, *** P < 0.001.

Figure 6.

Model of CDK12 function in transcription. During the transcription cycle, CDK12 phosphorylates Ser2 and Ser5 of the pol II CTD and possibly other factors involved in transcription such as SPT6. Inhibition of CDK12 affects the pol II elongation rate by affecting the recruitment and/or stability on the chromatin of LEO1 and CDC73 from the PAF1 complex (PAF1C) and SPT6. CDK12 inhibition does not affect pol II pause release, which is mediated by CDK9. Orange line with ‘cap’: capped mRNA; orange and red dot: Ser2P and Ser5P, respectively. Red and violet arrows: phosphorylation mediated by CDK9 and CDK12, respectively.