Mediator binding to UAS s is broadly uncoupled from transcription and cooperative with TFIID recruitment to promoters
2016; Springer Nature; Volume: 35; Issue: 22 Linguagem: Inglês
10.15252/embj.201695020
ISSN1460-2075
AutoresSebastian Grünberg, Steven Henikoff, Steven Hahn, Gabriel E. Zentner,
Tópico(s)CAR-T cell therapy research
ResumoArticle20 October 2016free access Transparent process Mediator binding to UASs is broadly uncoupled from transcription and cooperative with TFIID recruitment to promoters Sebastian Grünberg Corresponding Author Sebastian Grünberg [email protected] orcid.org/0000-0002-5694-0574 Basic Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA Search for more papers by this author Steven Henikoff Steven Henikoff Basic Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA Howard Hughes Medical Institute, Fred Hutchinson Cancer Research Center, Seattle, WA, USA Search for more papers by this author Steven Hahn Steven Hahn Basic Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA Search for more papers by this author Gabriel E Zentner Corresponding Author Gabriel E Zentner [email protected] orcid.org/0000-0002-0801-7646 Department of Biology, Indiana University, Bloomington, IN, USA Search for more papers by this author Sebastian Grünberg Corresponding Author Sebastian Grünberg [email protected] orcid.org/0000-0002-5694-0574 Basic Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA Search for more papers by this author Steven Henikoff Steven Henikoff Basic Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA Howard Hughes Medical Institute, Fred Hutchinson Cancer Research Center, Seattle, WA, USA Search for more papers by this author Steven Hahn Steven Hahn Basic Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA Search for more papers by this author Gabriel E Zentner Corresponding Author Gabriel E Zentner [email protected] orcid.org/0000-0002-0801-7646 Department of Biology, Indiana University, Bloomington, IN, USA Search for more papers by this author Author Information Sebastian Grünberg *,1, Steven Henikoff1,2, Steven Hahn1 and Gabriel E Zentner *,3 1Basic Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA, USA 2Howard Hughes Medical Institute, Fred Hutchinson Cancer Research Center, Seattle, WA, USA 3Department of Biology, Indiana University, Bloomington, IN, USA *Corresponding author. Tel: +1 206 667 5263; E-mail: [email protected] *Corresponding author. Tel: +1 812 856 7377; E-mail: [email protected] The EMBO Journal (2016)35:2435-2446https://doi.org/10.15252/embj.201695020 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Abstract Mediator is a conserved, essential transcriptional coactivator complex, but its in vivo functions have remained unclear due to conflicting data regarding its genome-wide binding pattern obtained by genome-wide ChIP. Here, we used ChEC-seq, a method orthogonal to ChIP, to generate a high-resolution map of Mediator binding to the yeast genome. We find that Mediator associates with upstream activating sequences (UASs) rather than the core promoter or gene body under all conditions tested. Mediator occupancy is surprisingly correlated with transcription levels at only a small fraction of genes. Using the same approach to map TFIID, we find that TFIID is associated with both TFIID- and SAGA-dependent genes and that TFIID and Mediator occupancy is cooperative. Our results clarify Mediator recruitment and binding to the genome, showing that Mediator binding to UASs is widespread, partially uncoupled from transcription, and mediated in part by TFIID. Synopsis A high-resolution map of Mediator distribution in the yeast genome shows preferential binding to upstream activating sequences, poor correlation with transcription levels, and a cooperative role for TFIID and Mediator on target genes. ChEC-seq provides a ChIP-independent means by which to map the genomic distribution of Mediator. Mediator binds to UASs rather than core promoters or gene bodies. Mediator binding to UASs is broadly uncoupled from transcriptional activity. TFIID associates with the promoters of both annotated SAGA- and TFIID-dependent genes. Mediator and TFIID chromatin association is cooperative. Introduction The Mediator complex is a conserved coactivator that is broadly required for eukaryotic transcription. Mediator integrates regulatory signals from DNA-bound transcriptional activators and cis-regulatory elements to modulate the basal RNA polymerase II (Pol II) transcription machinery. Mediator appears to exert its effects on transcription in part through interactions with other coactivator complexes such as SAGA and TFIID. Previous work has suggested that the Mediator tail module is preferentially required at SAGA-dependent promoters (Ansari et al, 2012). In vitro studies have also described cooperative DNA binding between TFIID and Mediator (Baek et al, 2002; Johnson et al, 2002; Johnson & Carey, 2003; Takahashi et al, 2011), but it is unclear whether such a cooperative relationship exists in vivo. In metazoans, Mediator associates with distal enhancer elements and is critical for looping of enhancers to promoters (Kagey et al, 2010). Despite recent advances in understanding the structure and functions of Mediator (Allen & Taatjes, 2015), the mechanisms by which Mediator co-regulates global Pol II transcription remain poorly understood. Key to understanding the in vivo function of Mediator is accurate determination of its genomic binding sites. However, genome-wide mapping of Mediator using various chromatin immunoprecipitation (ChIP)-based methods has yielded ambiguous results in budding yeast, complicating analysis of its global transcriptional role. Mediator has been reported variously to bind both upstream activating elements (UASs) (Jeronimo & Robert, 2014) and core promoters (Ansari et al, 2009, 2012). Recent findings also indicate that Mediator accumulates at yeast core promoters only upon inhibition of the TFIIH subunit Kin28 (Jeronimo & Robert, 2014; Wong et al, 2014). Multiple studies have also argued for (Andrau et al, 2006; Zhu et al, 2011; Wong et al, 2014; Paul et al, 2015) and against (Fan et al, 2006; Fan & Struhl, 2009; Jeronimo & Robert, 2014) gene body association of Mediator. Despite a decade of genome-wide Mediator mapping, ambiguity regarding its genome-wide binding persists: Two recently published Mediator ChIP-seq studies indicate predominant gene body binding of Mediator (Wong et al, 2014; Paul et al, 2015), while two other recent mapping studies show little binding of Mediator to gene bodies but robust association with upstream regions under normal growth conditions (Eyboulet et al, 2013; Jeronimo & Robert, 2014). Notably, ChIP-seq for Med17 using an antibody against Med17 (Paul et al, 2015) or HA-tagged Med17 (Eyboulet et al, 2013) gives substantially different results. Issues potentially leading to these conflicting ChIP results include low Mediator ChIP efficiency and corresponding low enrichment values, artifactual signals in sonicated input chromatin (Teytelman et al, 2009; Vega et al, 2009; Grokhovsky et al, 2011; Poptsova et al, 2014), and the reported hyper-ChIPability of highly expressed genes (Park et al, 2013; Teytelman et al, 2013). While genome-wide binding of Mediator has been studied mainly in budding yeast, the structure of Mediator and its mechanisms of action are conserved throughout the eukaryotic lineage (Cai et al, 2009; Allen & Taatjes, 2015). As such, the uncertainty surrounding the binding of Mediator to the budding yeast genome has implications for understanding Mediator function in all eukaryotes. Here, we apply chromatin endogenous cleavage and high-throughput sequencing (ChEC-seq) (Zentner et al, 2015) to map Mediator binding to the yeast genome and ascertain its relationship to TFIID. ChEC-seq employs fusion of micrococcal nuclease (MNase) to chromatin-associated proteins, directing calcium-dependent cleavage to specific sites on chromatin in vivo. ChEC-seq is thus immunoprecipitation-independent and as such does not require cross-linking, chromatin solubilization, or antibodies, and is quantitative. ChEC-seq therefore provides a ChIP-independent means by which to establish high-confidence profiles of Mediator binding. Profiling two Mediator head module subunits (Med8 and Med17), we observe that Mediator globally associates with UASs, rather than core promoters or gene bodies, under all conditions tested. Unique patterns of Mediator enrichment at SAGA- and TFIID-dependent genes suggest distinct promoter architectures of their respective transcription initiation complexes on chromatin. A striking finding is that Mediator binding to UASs is widespread and at most genes only weakly correlated with expression levels, suggesting that Mediator occupancy is partly uncoupled from gene expression. However, loss of the Mediator tail subunit Gal11/Med15 strongly reduced Mediator recruitment to a subset of genes upregulated upon Gcn4 activation. Lastly, we find that Mediator is generally necessary for full recruitment of TFIID to TFIID- and SAGA-dependent genes and that TFIID is also required for full Mediator recruitment to chromatin. Our results clarify the genome-wide binding locations of Mediator and reveal a functional relationship between coactivators Mediator and TFIID in transcription initiation. Results ChEC-seq profiling of Mediator binding to the budding yeast genome ChEC-seq uses strains containing a C-terminal fusion of the calcium-dependent endo/exonuclease MNase to a chromatin-binding protein. Addition of calcium to permeabilized cells activates MNase and cleaves DNA in proximity to the chromatin-bound factor. We previously showed that ChEC-seq provides high-resolution maps of binding of the general regulatory factors Abf1, Rap1, and Reb1 to the yeast genome with additional information regarding their orientation on DNA (Zentner et al, 2015). As MNase must be near DNA for cleavage to occur, structural consideration of the protein(s) under study is essential. While this is relatively straightforward for transcription factors with defined DNA binding domains, the yeast Mediator complex consists of 25 subunits, which are distributed between four modules (head, tail, middle, kinase), without any documented DNA binding ability. Available high-resolution structural information is limited to parts of the Mediator middle module (Larivière et al, 2013; Wang et al, 2014) and the Mediator head domain (Imasaki et al, 2011; Lariviere et al, 2012). Head subunits Med8, Med17, and Med20 were fused with 3×FLAG-MNase based on their exposed carboxyl-terminal ends (Appendix Fig S1A). When assessed by agarose gel electrophoresis following ChEC, DNA from the Med8-MNase and Med17-MNase strains displayed a moderate amount of cleavage, while DNA from the Med20-MNase strain displayed little cleavage (Appendix Fig S1B). This was not due to lack of expression of Med20-MNase, as all three MNase-tagged subunits were appropriately expressed (Appendix Fig S1C). We sequenced endogenously cleaved DNA fragments from the Med8-MNase and Med17-MNase strains and mapped fragment ends back to the yeast genome. For comparison, we also plotted published Med14 ChIP-seq data (Wong et al, 2014) and Gal11/Med15 ChIP-chip data which had been normalized to a control ChIP from a strain without tagged Gal11/Med15 (Jeronimo & Robert, 2014). At three exemplary highly transcribed SAGA-dependent genes (CDC19, ILV5, PDC1), ChEC-seq revealed robust enrichment of Med8 and Med17-MNase cleavages upstream of TSSs but not within gene bodies (Fig 1A). Enrichment of specific ChEC cleavages was specific to fusion of MNase to Mediator subunits, as it was not observed in a strain expressing untethered MNase under the control of the MED8 promoter (Fig 1A and B). Med14 ChIP-seq showed enrichment upstream and within the coding regions of all three genes, with gene body signal being particularly pronounced at CDC19 (Fig 1A). Gal11/Med15 ChIP-chip effectively captured upstream enrichment at all three genes as well as modest gene body signal. We next examined Med8 and Med17 cleavages at three exemplary strongly transcribed TFIID-dependent genes (EFB1, RPS5, YEF3). As observed for SAGA-dependent genes, Med8 and Med17 cleavages were enriched upstream of TSSs but not within coding regions (Fig 1B). Med14 ChIP-seq displayed upstream of EFB1, across the upstream region and coding region of RPS5, and within the coding region of YEF3 (Fig 1B). Gal11/Med15 ChIP-chip captured enrichment upstream of EFB1 and YEF3, with some coding region signal at YEF3, and across the upstream region and coding region of RPS5 (Fig 1B). In summary, ChEC shows Mediator interaction at many intergenic gene regulatory regions but not within coding sequences of mRNA genes. In addition, the contrasting patterns of genome-wide ChIP enrichment we observed for Med14 and Gal11/Med15 using recent datasets underscores the continuing lack of clarity regarding the genome-wide distribution of Mediator. Figure 1. ChEC-seq mapping of Mediator at SAGA- and TFIID-dependent genes A, B. Signal tracks showing cleavages generated by Med8-MNase, Med17-MNase, and PMED8-MNase in YPD at three highly expressed SAGA-dependent (A) and three highly expressed TFIID-dependent (B) loci. All time points for a given factor were scaled to the same data range, and PMED8-MNase tracks were scaled to the lower Mediator-MNase fusion range. Med14 ChIP-seq and Med15 ChIP-chip data are shown for comparison. TSSs are indicated by arrows. Download figure Download PowerPoint Distinct patterns of Mediator association with SAGA- and TFIID-dependent UASs We next investigated the position of Mediator binding relative to TSSs. To this end, we assessed the average distance of the cleavage peak summit for each Mediator subunit and promoter class profiled, pooling data from ChEC time points (Fig 2A). For Med8, the distance from the peak summit to TSS at SAGA-dependent genes was 267 bp (332 bp for Med17) and 165 bp at TFIID-dependent promoters (127 bp for Med17). The distances of the Mediator cleavage maxima to TSSs support preferential binding to UASs, which are generally located 250–400 bp upstream of TSSs in yeast (Chambers et al, 1988; de Bruin et al, 2001; He et al, 2012; Yan et al, 2015), rather than core promoters, which typically span 75 bp upstream and 50 bp downstream of TSSs (Lubliner et al, 2013). We observed similar results when single ChEC time points were analyzed (Appendix Fig S2), indicating that pooling of ChEC time points does not distort average profiles. On the single gene level, our average observations were confirmed by robust Med8 cleavage over the previously characterized UASs of the CLB2 (SAGA-dependent) (Van Slyke & Grayhack, 2003) and RPS5 (TFIID-dependent) (Li et al, 2002) genes (Fig 2B). We also explored the previously reported association of Mediator with gene bodies. Consistent with our single-locus results (Fig 1), we observed little Med8 or Med17-MNase cleavage as far as 1 kb into gene bodies, indicating that the gene body enrichment of Mediator detected in many ChIP studies is not representative of Mediator's location. As Med8 and Med17 displayed very similar cleavage profiles at the UASs of SAGA- and TFIID-dependent genes in the preceding analyses, only Med8 was profiled in subsequent ChEC-seq experiments. Figure 2. Mediator preferentially associates with the UASs of SAGA-dependent genes Average plots of Mediator cleavages around the TSSs of SAGA- and TFIID-dependent genes. The distance of cleavage maximum to TSS is indicated by the dotted lines and corresponding numbers. Signal tracks of Med8-MNase and free MNase cleavages at the previously characterized UASs of the CLB2 and RPS5 genes. TSSs are indicated by arrows. Download figure Download PowerPoint Inactivation of Kin28 does not cause Mediator displacement from UASs Two recent studies suggested that Ser5 phosphorylation of the Pol II C-terminal domain by the TFIIH-associated kinase Kin28 is important for Mediator release from the core promoter (Jeronimo & Robert, 2014; Wong et al, 2014). We thus wondered whether Mediator binding, as measured by ChEC-seq, would shift from UASs to core promoters in kin28-analog sensitive (Kin28AS) cells after treatment with the inhibitor NA-PP1. As shown earlier, 6 μM NA-PP1 was sufficient to fully inhibit cell growth of the Kin28AS strain (Appendix Fig S3). NA-PP1 treatment resulted in a moderate decrease in Mediator binding at the UASs of mainly SAGA-dependent, and, to a lesser extent, TFIID-dependent genes (Fig 3A, Appendix Fig S4). However, we did not detect Mediator cleavages at the core promoter (Fig 3A), indicating that a majority of Mediator remained bound to UASs. This observation was confirmed at the upstream regions of four genes (GAP1, BAT1, ECM33, RPL2B) previously shown to have increased Mediator association with core promoters following NA-PP1 treatment of a Kin28AS strain (Wong et al, 2014) (Fig 3B). These results contrast with those of two recent studies showing accumulation of Mediator at core promoters upon Kin28 inhibition (Jeronimo & Robert, 2014; Wong et al, 2014). We speculate that, when looped to core promoters from UASs, Mediator-tethered MNase may be too far from DNA for efficient cleavage or blocked from access to DNA by the PIC, and that Mediator may be detected at core promoters in ChIP experiments through cross-linking to the PIC or other promoter-associated factors. Figure 3. Lack of RNA Pol II CTD Ser5 phosphorylation does not lead to robust Mediator-UAS dissociation Average plots of Med8 cleavages around the TSSs of SAGA- and TFIID-dependent genes in Kin28AS cells treated with DMSO (-NA-PP1) or 6 μM NA-PP1 (+NA-PP1). Signal tracks of Med8-MNase cleavages at the upstream regions of several genes previously shown to have increased core promoter association of Mediator by ChIP-qPCR following NA-PP1 treatment of a Kin28AS strain (Wong et al, 2014). All time points for a given treatment were concatenated to generate a combined track. TSSs are indicated by arrows. Download figure Download PowerPoint Mediator binding is widespread and uncoupled from transcriptional activity at most genes To further clarify the relationship of Mediator binding to the expression of SAGA- and TFIID-dependent genes, we stratified average Med8 cleavage levels at SAGA- and TFIID-dependent genes by expression level. As a measure for active transcription, previously published native elongating transcript sequencing (NET-seq) (Churchman & Weissman, 2011) signal within 200 bp downstream of the TSS was used. We found that Med8 occupancy was highest at the most highly transcribed quintile in both gene classes (Fig 4A), with significantly higher cleavage observed at the UASs of SAGA-dependent genes. To ensure that this was not due to the smaller number of genes in the SAGA-dependent quintiles (87 SAGA-dependent to 810 TFIID-dependent), we repeated the analysis using the 87 most highly transcribed TFIID-dependent genes. Again, we found that highly transcribed SAGA-dependent genes showed higher levels of Med8 cleavage at their UASs compared to the smaller TFIID-dependent gene set (Fig 4B), despite the fact that, on average, the top 87 most highly transcribed TFIID-dependent genes were expressed at significantly higher levels (Fig 4C, P = 2.23 × 10−13 by unpaired t-test). To more systematically assess the relationship between Mediator occupancy and transcription, we correlated average Med8 ChEC-seq signal in a 1-kb window upstream of the TSS with average NET-seq counts in a 200-bp window downstream of the TSS. In agreement with our above findings (Fig 4A), this revealed only modest correlations (SAGA-dependent Spearman's ρ = 0.4126, P < 0.0001; TFIID-dependent Spearman's ρ = 0.3467, P < 0.0001) (Fig 4D). We wondered whether the weak correlation of Med8 cleavage with transcriptional activity was driven by distinct subsets of TFIID- or SAGA-dependent genes. Hence, we determined correlations between Med8 cleavage and NET-seq for the SAGA- and TFIID-dependent NET-seq quintiles described in Fig 4A. While correlations were relatively poor across all quintiles, the best correlations were observed in the most highly transcribed quintile for both SAGA- and TFIID-dependent genes (Appendix Fig S5). Taken together, these observations indicate that Mediator occupancy and gene expression are largely but not completely uncoupled, and that Mediator binding to UASs is widespread. Figure 4. Mediator binding is widespread and partially uncoupled from transcription Average plots of Med8 cleavages around SAGA- and TFIID-dependent gene TSSs stratified into quintiles by the level of NET-seq signal in a 200-bp window downstream of the TSS. Average plots of Med8 cleavage around the TSSs of the 87 most highly transcribed SAGA- and TFIID-dependent genes. Boxplots of the transcription levels of the 87 most highly transcribed SAGA- and TFIID-dependent genes as determined by NET-seq. Significance was assessed by t-test. Horizontal line = mean, box range = 10–90th percentile, error bars = min to max. Scatterplots of average NET-seq counts in a 200-bp window downstream of the TSS versus average Med8 cleavages in a 1-kb window upstream of the TSS for SAGA- and TFIID-dependent genes. Correspondence between the datasets was assessed by Spearman correlation. Download figure Download PowerPoint The Mediator head module is recruited to activated genes and remains bound to downregulated genes We next analyzed how a global perturbation of transcription affects Mediator association with the genome. We treated cells with sulfometuron methyl (SM), which mimics amino acid starvation (Jia et al, 2000). This treatment results in upregulation of the Gcn4 transcription factor, which in turn activates the transcription of amino acid biosynthetic genes (Hinnebusch, 2005) in part through interactions with the Med15/Gal11 subunit of the Mediator tail module (Herbig et al, 2010; Jedidi et al, 2010). We first assessed Mediator recruitment to Gcn4 binding sites previously determined by ChIP-chip (MacIsaac et al, 2006) and observed a robust increase in Med8 binding following SM treatment (Fig 5A). We also mapped Mediator binding via Med17 ChEC-seq in a strain lacking Med15 and observed a substantial reduction in signal at Gcn4 binding sites, confirming the tail module dependence of activator recruitment (Fig 5A). The majority of Gcn4 sites tested displayed an increase in Mediator cleavage upon SM induction, and these increases were strongly attenuated in med15Δ (Appendix Fig S6). We next analyzed Med8 occupancy around the TSSs of genes ≥ twofold up- or downregulated by SM treatment (Saint et al, 2014). We found that Mediator recruitment to UASs was increased at approximately 20% of upregulated genes (Fig 5B; Appendix Fig S6) and these increases were diminished in med15Δ (Fig 5B, Appendix Fig S6). Unexpectedly, we observed only a modest decrease in Med8-dependent cleavage near SM-downregulated genes (Fig 5B). This decrease was observed at only a small minority of genes (Appendix Fig S6). Our findings suggest that the majority of Mediator remains bound to UASs even upon transcriptional downregulation, further supporting the hypothesis that Mediator occupancy and transcription are uncoupled at many genes. Figure 5. The Mediator head module is recruited to Gcn4 sites and upregulated genes in SM and only moderately dissociates from SM-downregulated genes Average plot of Med8 cleavages around 193 Gcn4 ChIP-chip peak midpoints (MacIsaac et al, 2006) in WT and med15Δ. Average plots of Med8 cleavages around the TSSs of genes upregulated and downregulated ≥ twofold in SM (Saint et al, 2014) in WT and med15Δ. Control cleavages were subtracted from SM cleavages at each base position. qRT–PCR analysis of SAGA- and TFIID-dependent genes in WT and med15Δ. Bars represent mean + SEM for two biological replicates performed in triplicate. *P < 0.05; †P < 0.01; ‡P < 0.005 by unpaired Student's t-test. Download figure Download PowerPoint The Mediator tail has previously been implicated in SAGA-dependent activated transcription (Ansari et al, 2012). We therefore used qRT–PCR to examine the effects of deleting Med15 on the expression of a set of representative Gcn4-activated genes consisting of seven SAGA-dependent and seven TFIID-dependent genes following SM induction. Deletion of the Mediator tail significantly downregulated 5/7 of the SAGA-dependent genes tested (ARG3, HIS4, ARG1, STR3, and ARG5), and the remaining SAGA-dependent genes (PCL5 and ALD5) showed a trend toward decreased expression. Of the TFIID-dependent genes tested, TRP3 expression was significantly increased, and three of the other tested TFIID-dependent genes (LEU3, RTG3, and SNO1) showed a trend toward increased expression (Fig 5C). These results confirm our genome-wide results that recruitment of Mediator through its tail module is important for activated transcription of predominantly SAGA-dependent genes. TFIID depends on Mediator for maximal promoter interaction To characterize the different recruitment pathways for Mediator at SAGA and TFIID-dependent genes, we analyzed the effect of Mediator loss on TFIID binding to chromatin, as previous studies have suggested Mediator-TFIID DNA binding cooperativity (Baek et al, 2002; Johnson et al, 2002; Johnson & Carey, 2003; Takahashi et al, 2011). To map TFIID binding genome-wide, we tagged the Taf1 subunit with MNase and performed ChEC-seq. Strikingly, we observed notable enrichment of Taf1 cleavages at both SAGA- and TFIID-dependent TSSs, though cleavages were slightly higher at annotated TFIID-dependent promoters (Fig 6A). As expected, the major site of Taf1 binding was within the nucleosome-depleted region (NDR) at core promoters (peak summit to TSS distance: 51 bp for SAGA-dependent, 68 bp for TFIID-dependent). We surprisingly observed a periodic enrichment of Taf1 cleavages into gene bodies, both upstream and downstream of TFIID-dependent TSSs in a pattern that is reminiscent of the nucleosomal arrays present in gene bodies. Comparison of MNase-seq (Henikoff et al, 2011) to Taf1 ChEC-seq data revealed a striking inverse relationship between Taf1 cleavages and nucleosome occupancy (Fig 6A), suggesting moderate cleavage of linker DNA between nucleosomes, perhaps due to interaction of the TFIID-associated and bromodomain-containing subunit Bdf1 with acetylated nucleosomes in the promoter region (Matangkasombut et al, 2000; Durant & Pugh, 2007). This pattern was also observed downstream of SAGA-dependent TSSs, though to a lesser extent. Figure 6. Mediator is required for full TFIID recruitment Average plot of Taf1 cleavages around the TSSs of SAGA- and TFIID-dependent genes. A nucleosome occupancy profile as determined by MNase-seq is also shown for comparison. Dotted lines represent Taf1 cleavage maxima in TFIID-dependent gene bodies. The NDR, taken to be the −1 to +1 nucleosome midpoint distance, is indicated by a black rectangle. Average plots of Taf1 cleavages around the TSSs of SAGA- and TFIID-dependent genes ± rapamycin to deplete Med14-FRB. Download figure Download PowerPoint We next examined whether loss of Mediator would affect TFIID recruitment and subsequent PIC formation. We used anchor-away (Haruki et al, 2008) to rapidly deplete nuclear Med14, which functions as connector between all four Mediator modules (Tsai et al, 2014). Depletion of nuclear Med14 (Appendix Fig S7) resulted in a reduction of TFIID occupancy at the core promoter, but a slight increase in and downstream shift of gene body cleavages regardless of coactivator dependence (Fig 6B). Approximately 70% of promoter NDRs displayed a modest decrease in Taf1 binding (Appendix Fig S8), indicating a moderate but widespread role for Mediator in TFIID recruitment. Loss of TFIID impairs Mediator recruitment Having found that disruption of Mediator impairs TFIID recruitment to promoters, we sought to determine whether the converse is true. We performed Med8 ChEC-seq following anchor-away depletion of Taf1 (Appendix Fig S7), an essential subunit of TFIID involved in promoter DNA binding (Louder et al, 2016). Med8 cleavages were, on average, strongly reduced at both SAGA- and TFIID-dependent promoters upon Taf1 depletion (Fig 7), implying an important role for TFIID in Mediator binding. Unexpectedly, Mediator recruitment to UASs of SAGA-dependent genes appeared to be more strongly affected by Taf1 depletion. Indeed, ~40% of SAGA-dependent upstream regions showed a noticeable decrease in average Med8 cleavages, while ~20% of TFIID-dependent upstream regions showed decreased average Med8 cleavages (Appendix Fig S9). We also noted an increase in average Med8 cleavages upstream of ~20% of SAGA-dependent genes, but no such increase upstream of TFIID-dependent genes (Appendix Fig S9). Combined with our data indicating that TFIID recruitment to the majority of genes is partially dependent on Mediator (Fig 6B), these observations indicate mutual dependency between TFIID and Mediator for chromatin recruitment to many genes. Figure 7. Loss of Taf1 impairs Mediator recruitmentAverage plots of Med8 cleavages around the TSSs of SAGA- and TFIID-dependent genes ± rapamycin to deplete Taf1-FRB. Download figure Download PowerPoint Discussion The in vivo genomic distribution of the Mediator complex in budding yeast has presented a considerable obstacle to understanding its in vivo functions. We present a genome-wide map of Mediator binding in budding yeast generated using ChEC-seq, a method based on a completely different principle than ChIP, allowing us to identify Mediato
Referência(s)