CDK8 is a positive regulator of transcriptional elongation within the serum response network

Abstract

The Mediator complex allows communication between transcription factors and RNA polymerase II (RNAPII). Cyclin-dependent kinase 8 (CDK8), the kinase found in some variants of Mediator, has been characterized mostly as a transcriptional repressor. Recently, CDK8 was demonstrated to be a potent oncoprotein. Here we show, using a human tumor cell line, that CDK8 is a positive regulator of genes within the serum response network, including several members of the activator protein 1 and early growth response family of oncogenic transcription factors. Mechanistic studies show that CDK8 is not required for RNAPII recruitment or promoter escape. Instead, CDK8 depletion leads to the appearance of slower elongation complexes carrying hypophosphorylated RNAPII. CDK8-Mediator regulates precise steps in the assembly of the RNAPII elongation complex, including the recruitment of positive transcription elongation factor b and BRD4. Furthermore, CDK8-Mediator specifically interacts with positive transcription elongation factor b. Thus, we have uncovered a role for CDK8 in transcriptional regulation that may contribute to its oncogenic effects.

Figure 1. CDK8 is a positive coregulator of serum responsive immediate early genes

(a) Western blot and Q-RT-PCR analysis of CDK8 depletion using short-hairpin RNA (shRNA) in HCT116 cells. (b) Microarray. The left panel is a heat map showing the 29 genes induced more than two-fold in control cells. The right panel is a heatmap of the relative level of expression of the same set of genes in CDK8 depleted cells. CDK8-dependent activation of the genes highlighted in red is further studied. Asterisks denote the only three genes whose expression increases upon CDK8 depletion. (c) Validation of microarray data. Immediate early gene mRNA expression was measured using quantitative-RT-PCR (Q-RT-PCR). mRNA levels were normalized to 18S rRNA. Data from at least three independent experiments is represented as mean −/+ standard error of the mean.

Figure 2. CDK8 positively affects RNAPII CTD phosphorylation at Ser5 and Ser2 without an effect on total RNAPII occupancy at FOS, EGR1, EGR2 and EGR3

The upper panel depicts the basic gene structure (enhancer, exons and introns) of FOS, EGR1, EGR2 and EGR3 and positions of amplicons at each loci used in the ChIP assays. Chromatin immunoprecipitation assay results for cyclin-dependent kinase 8 (CDK8), total RNA polymerase (RNAPII), phospho-Ser5 CTD (S5P) and phospho-Ser2 CTD (S2P) at the FOS, EGR1, EGR2 and EGR3 loci are shown in the lower panel. The mean from at least three independent experiments is represented.

Figure 3. CDK8 promotes RNAPII elongation at IEGs

(a) Q-RT-PCR analysis of IEG primary transcripts. Serum starved cells were treated with serum for the indicated times and nascent message was analyzed by Q-RT-PCR using intronic primers for each gene. (b) Nuclear Run-On analysis. Serum starved cells were treated with serum for 10 minutes, transcription was stalled and then allowed to continue for either 5 or 60 minutes in the presence of biotin-labelled UTP. Labelled RNA was purified and analysed by Q-RT-PCR as in a. Data from at least three independent experiments is represented as mean −/+ standard error of the mean. Small schematics of each locus, as depicted in Figures 2a and 3a, and position of the intronic amplicon used are shown above the expression data for each gene.

Figure 4. CDK8 affects serum-dependent recruitment of Mediator but not the levels of ELK1 and SRF, ELK1 phosphorylation, tetra-acetyl histone H4 (AcH4), TBP, and TFIIB

ChIP assays with the indicated antibodies and FOS and/or EGR1 directed primers were performed as described in Figure 2. Average data from at least three independent experiments is represented.

Figure 5. CDK8 is required for recruitment of CDK7, CDK9 and BRD4 to IEGs

(a) ChIP for TFIIH (CDK7), P-TEFb (CDK9), NELF(A), DSIF (SPT5), FACT (SPT16), BRD4, histone H4 acetylated at lysines 8 and 12 (H4 K8Ac and K12 Ac), and histone H3 acetylated at lysine 9 (H3 K9Ac) at the FOS locus as described in Figure 2. (b) ChIP for TFIIH (CDK7), P-TEFb (CDK9) and BRD4 at the EGR1 locus as described in Figure 2. Average data from at least three independent experiments is represented.

Figure 6. CDK9 inhibition mimics the post-recruitment defects in RNAPII activity observed in shCDK8 cells

(a) mRNA levels of FOS, EGR1, EGR2 and EGR3 were measured by Q-RT-PCR as in Figure 1, in serum starved cells treated with serum for 30 minutes with and without 1 hour of flavopiridol (150 nM) pre-treatment. Data from at least three independent experiments is represented as mean −/+ standard error of the mean. (b) ChIP for RNAPII, S5P and S2P at the EGR1 and EGR2 loci following serum stimulation with or without flavopiridol (150 nM) pre-treatment. Average data from at least three independent experiments is represented.

Figure 7. P-TEFb associates with CDK8-Mediator

(a) Silver stain of CDK8 and MED1 immunoprecipitates (IPs). Mediator and associated cofactors were immunopreciptated from Hela nuclear lysates using antibodies specific for CDK8 (CDK8-Mediator and CDK8-submodule) and MED1 (core Mediator and CDK8-Mediator). (b) Quantitative immunoblotting of IP elutions show CDK9 and cyclin T1 enriched in the CDK8 IP. c, Immunoblotting purified fractions of CDK8-Mediator and CDK8-submodule shows that CDK9 is detected in both the CDK8-submodule enriched (P0.3M/QFT/ASP/SREBP FT) and CDK8-Mediator enriched (P0.5M/QFT) fractions.