Tail and Kinase Modules Differently Regulate Core Mediator Recruitment and Function In Vivo

Abstract

Mediator is a highly conserved transcriptional coactivator organized into four modules, namely Tail, Middle, Head, and Kinase (CKM). Previous work suggests regulatory roles for Tail and CKM, but an integrated model for these activities is lacking. Here, we analyzed the genome-wide distribution of Mediator subunits in wild-type and mutant yeast cells in which RNA polymerase II promoter escape is blocked, allowing detection of transient Mediator forms. We found that although all modules are recruited to upstream activated regions (UAS), assembly of Mediator within the pre-initiation complex is accompanied by the release of CKM. Interestingly, our data show that CKM regulates Mediator-UAS interaction rather than Mediator-promoter association. In addition, although Tail is required for Mediator recruitment to UAS, Tailless Mediator nevertheless interacts with core promoters. Collectively, our data suggest that the essential function of Mediator is mediated by Head and Middle at core promoters, while Tail and CKM play regulatory roles.

Keywords: Cdk8/Srb10; TFIIH/Kin28; kinase module; mediator; tail module.

Figure 1. Mediator CKM Subunits Do Not Associate with Core Promoters in Living Cells

(A) TFIIB (tan), RNAPII (Rpb3, light gray), Mediator scaffold subunit Med14 (dark gray), Mediator Tail subunits Med15 and Med16 (blue), Mediator Head subunits Med17, Med22, Med8, and Med6 (gold), Mediator Middle subunits Med19 and Med7 (green), and Mediator CKM subunits Cdk8, CycC, and Med13 (red) occupancy (normalized ChIP-chip log₂ ratio of tagged versus no-tagged samples) in KIN28 WT (top tracks) and kin28as (bottom tracks) cells, all treated with NAPP1, along a segment of chromosome VII. Vertical dashed lines indicate the position of the two UAS bound by Mediator in this region. Genes (gray boxes) are shown at the bottom. Data for TFIIB, RNAPII, and Med15 in this and other figures are from (Jeronimo and Robert, 2014). (B) Scatterplots showing the occupancy of Mediator (the average of all subunits profiled) versus that of the indicated CKM subunits over 498 UAS regions in WT cells. (C) Scatterplots showing the occupancy of various Mediator subunits at core promoters (n = 5,128) in kin28as versus WT cells, all treated with NAPP1. (D) Distribution of the differential promoter occupancy in WT and kin28as cells for various Mediator subunits, all treated with NAPP1. (E) Average Cdk8 occupancy around UAS regions (n = 498) in WT (solid trace) and in med13Δ cells (dashed trace). (F) Average Cdk8 occupancy around UAS regions (n = 498) in WT (solid trace) and in med3Δ/med15Δ (TailD, dashed trace) cells. The coordinates of the UAS used in this and other figures were determined using our Mediator ChIP-chip data (see Experimental Procedures). See also Figure S1.

Figure 2. Mediator Accumulates at UAS Regions in the Absence of CKM

(A) Med15 (blue) and Med7 (green) occupancy in WT and kin28as cells, all treated with NAPP1, along a segment of chromosome VII. Dark gray traces show the effect of deleting MED13 in these strains. The Med15 and Med7 peaks appearing in kin28as cells correspond to promoter peaks and are not enhanced when deleting MED13 (med13Δ, dark gray trace). Arrows are pointing toward a UAS Mediator peak, enhanced upon MED13 deletion. (B) Average Med15 (blue) and Med7 (green) occupancy around the transcription start site (TSS) of transcriptionally active genes (n = 165, see Experimental Procedures) in kin28as cells (solid traces) and in kin28as/med13Δ cells (dashed traces). See also Figure S2A. (C) Average Med15 (blue) and Med7 (green) enrichment around UAS regions (n = 498) in WT (solid traces) and med13Δ (dashed traces) cells. (D) Heatmap representation of the hierarchical clustering of Pearson correlations between expression profiles of various Mediator subunit mutants. Data for srb10-3, a strain with a point mutation (D290A) in the catalytic domain of the CDK8 gene (Liao et al., 1995), were from Holstege et al. (1998), while all other data sets are from Kemmeren et al. (2014). See also Figure S2B. (E) Average (left) and heatmap representation (right) of Med15 occupancy around UAS regions (n = 498) in WT (solid traces) and cdk8D290A (dashed traces) cells.

Figure 3. The Tail Domain Is Required for Mediator-UAS Interactions but Dispensable for Mediator to Associate with Core Promoters

(A) Med7 (green) occupancy at the SAGA-dominated gene PMA1 (left) and the TFIID-dominated gene SCW4 (right) in WT and med3Δ/med15Δ (TailΔ) cells, both in a KIN28 WT and kin28as backgrounds. All samples were treated with NAPP1. TFIIB (tan) is shown as a placeholder for core promoters. Vertical dashed lines indicate the position of the UAS and TSS. (B) Heatmap representation of Med7 occupancy in WT (left) and med3Δ/med15Δ (TailΔ, right) cells around UAS regions (n = 498, see Experimental Procedures) sorted by decreasing Mediator occupancy. (C) Heatmap representation of Med7 occupancy in kin28as (WT, left) and kin28as/med3Δ/med15Δ (TailΔ, right) cells around the TSS of genes associated to the UAS regions used in (B). Traces at the bottom of each heatmap from (B) and (C) show the average occupancy across these loci. (D) Cumulative plot of the change in Med7 occupancy in WT versus med3Δ/med15Δ (TailΔ) cells at UAS (solid trace) and at core promoters (CP, dashed traces). UAS and core promoter traces were calculated from experiments performed in KIN28 WT and kin28as cells treated with NAPP1 respectively, using the same loci as in (B) and (C). See also Figure S3.

Figure 4. The Tail Module Contributes to PIC Assembly/Stability to a Similar Extent in SAGA- and TFIID-Dominated Genes

(A) Cumulative plot of the change in TFIIB promoter occupancy in med18-AA cells after a 90 min treatment with rapamycin (solid trace) and in med3Δ/med15Δ cells (TailΔ, dashed trace), relative to their respective WT. All genes with a TFIIB log₂ ratio > 2 were used in this analysis. (B) Cumulative plot of the change in TFIIB promoter occupancy in med3Δ/med15Δ cells (TailΔ) cells, relative to WT, for the SAGA-dominated (red) and TFIID-dominated (blue) genes with a TFIIB log₂ ratio > 2. (C) Cumulative plot of the change in TFIIB promoter occupancy in med18-AA cells after a 90 min treatment with rapamycin, relative to WT, for the SAGA-dominated (red) and TFIID-dominated (blue) genes with a TFIIB log₂ ratio > 2 (high TFIIB, solid traces) or for genes having equivalent TFIIB levels between the SAGA- and TFIID-dominated gene groups (equivalent TFIIB, dashed traces). (D) Distribution of TFIIB occupancy at the promoter of SAGA-dominated genes (red) and TFIID-dominated genes (blue). See also Figure S4.

Figure 5. A Topological Model for Mediator Interactions at the UAS and Promoter

(A) TFIIB, RNAPII, and various Mediator subunits (as in Figure 1A) occupancy in WT and kin28as cells, all treated with NAPP1, at the PMA1 (left) and CDC19 (right) loci. Vertical dashed lines indicate the position of the UAS and TSS, respectively defined as the peak summit of Mediator and TFIIB in WT cells. (B) Distribution of the difference in occupancy at the UAS versus the core promoter in kin28as cells treated with NAPP1, for various core Mediator subunits. All gene-UAS pairs separated by more than 300 bp are used (n = 108). (C) Read density for TFIIB (tan), Med15 (blue), and Med7 (green) ChIP-exo experiments performed in kin28as cells, treated with NAPP1, at the PMA1 (top) and PGK1 (bottom) loci. Vertical dashed lines indicate the position of the UAS and TSS. (D) A model for the topology of core Mediator, transiently interacting with the PIC. The topological model was built using data from Plaschka et al. (2015), onto which the various Mediator subunits were positioned using data from previous Mediator models (Robinson et al., 2015; Tsai et al., 2014) (see Experimental Procedures). Transparent surfaces colored in shades of blue (Tail), gold (Head), and green (Middle) show the positions of the Mediator modules based on Robinson et al. (2015). Ribbons for GTFs TBP, TFIIB, and TFIIF are shown in tan, for RNAPII are shown in gray, and for Head mediator subunits are shown in gold based on Plaschka et al. (2015). A cavity in the structure that coincides with the position were the CTD emerges from Rpb1 is indicated by an arrow. The structure of the Gcn4 DNA-binding domain bound to the UAS (Keller et al., 1995) (PDB accession number 2DGC) is shown as an example of an activator interacting with the Tail subunit Med15. (E) The minimal distance between the center of mass of each core Mediator subunits and the predicted position of the UAS DNA (gray) and core promoter DNA (CP; white) based on the structural model shown in (D). See also Figure S5.