Novel Role for Mediator Complex Subunit Srb5/Med18 in Termination of Transcription^,

Background: Mediator plays a role in the recruitment and post-recruitment steps of transcription by RNAP II.

Results: Mediator subunit Srb5 plays a role beyond the recruitment of preinitiation complex initiation and elongation steps of transcription.

Conclusion: Mediator has a novel role in termination of transcription in budding yeast.

Significance: Mediator-dependent termination may enhance transcription efficiency by coupling termination to reinitiation.

Abstract

Mediator complex functions at the recruitment as well as the post-recruitment steps of transcription. Here we provide evidence for a novel role of Mediator in termination of transcription. Mediator subunit Srb5/Med18 cross-links to the 5′ and 3′ ends of INO1 and CHA1. In srb5⁻ cells, recruitment of TATA-binding protein (TBP) and transcription factor IIB (TFIIB) onto the promoter of these genes remained unaffected, but cross-linking of the cleavage-polyadenylation factors Rna15 and Pta1 toward the 3′ end of genes was compromised. In these cells, RNA polymerase II accumulated near the 3′ end of genes and beyond. Transcription run-on analysis confirmed a transcription readthrough phenotype in the absence of Srb5/Med18. These results strongly suggest that Mediator subunit Srb5/Med18 is required for proper termination of transcription of a subset of genes in budding yeast.

Introduction

Mediator is a multisubunit megacomplex that plays a key role during transcription of protein-encoding genes in eukaryotes (1–4). It was discovered as a factor that helps activators communicate with the general transcription machinery in budding yeast (5–7). Since then, the role of Mediator in activator-dependent transcription has become well established among eukaryotes (3). Recent evidence suggests that Mediator is required for basal transcription as well (8–13). Although Mediator is structurally and functionally conserved throughout eukaryotes, the precise subunit structure and composition of the complex vary within a species and among species (4, 14). Mediator is recruited to the promoter region before binding of RNAP II³ and the general transcription factors. The promoter-bound Mediator then facilitates the recruitment of general transcription factors and RNAP II to form a preinitiation complex (PIC).

It is generally believed that Mediator helps in the assembly of the PIC during the first round of transcription and further stabilizes the assembled PIC for subsequent rounds of transcription (15, 16). Several lines of evidence suggest that the role of Mediator is not restricted to the recruitment of basal transcription machinery but that the complex has a “post-recruitment” role as well (3). First, genome-wide analysis revealed the association of Mediator with the coding region of transcriptionally active genes in budding yeast (17). Second, genetic analysis in yeast revealed the interaction of Mediator subunits with Set2, a histone methyl transferase, which functions during transcription elongation (18, 19). Third, Mediator subunit Med23 is required for stimulation of transcription of the serum-responsive gene, EGR1, at a step after the assembly of the PIC on the gene (20). Fourth, Mediator-dependent transcriptional activation in a reconstituted in vitro system required the elongation factor DRB sensitivity inducing factor (DSIF) (21). Lastly, Mediator subunit Med26 recruits the elongation complex containing eleven-nineteen lysine-rich leukemia (ELL)/ELL associated factor (EAF) and positive transcription elongation factor B (p-TEFB) to facilitate the release of paused polymerase into productive elongation (22). These studies reflect an extension of Mediator function into the initiation and early elongation steps of transcription. Despite the presence of Mediator near the 3′ ends of some genes in yeast (17), involvement of the complex in termination of transcription has not yet been demonstrated. Here we provide evidence that the function of Mediator extends well beyond its role in PIC assembly, initiation, and elongation into the last phase of the transcription cycle, that is, termination.

EXPERIMENTAL PROCEDURES

Yeast Strains

The yeast strains used in this study are listed in supplemental Table S1. Strain pairs FY23 (WT) and BPM2 (srb5⁻); SP10 and BPM12; BPM9 and BPM36; BPM34 and BPM35; and ABP1 and BPM37 are isogenic. BPM2 and BPM12 were constructed by replacing the entire ORF of SRB5/MED18 by KANMX6 and TRP genes, respectively, as described in Ref. 23. The C-terminal Myc-tagged Rna15 (SAM51, BPM33), HA-tagged Pta1 (BPM9, BPM36), Myc-tagged Srb4 (BPM34, BPM35), and tandem affinity purification-tagged Med15 (ABP1, BPM37) strains were derived from the FY23 and BPM2 strains by transforming with a PCR product amplified from pFA6-13Myc-TRP1, pFA6-3HA-TRP1, pFA6-13Myc-TRP1, or pBS1479 (TRP marker), respectively.

Cell Culture

Cultures were grown and processed for RT-PCR and ChIP as described in Ref. 24.

ChIP

ChIP experiments were performed as described previously (24). TBP and Rpb1 (8WG16) monoclonal antibodies were purchased from Santa Cruz Biotechnology and Covance, respectively.

Transcription Analysis

Transcription analysis of INO1 and CHA1 in wild type and srb5⁻ strains was performed by the RT-PCR approach as described previously (24).

Transcription Run-on (TRO) Assay

TRO assay was performed by the modification of protocols described in Refs. 25 and 26. All TRO signals were quantified and normalized with respect to 18 S rRNA controls.

Primers

Primers used in RT-PCR, ChIP, and TRO experiments are listed in supplemental Table S2.

Quantification

All quantifications and data analyses were performed as described in Ref. 24.

RESULTS

Genetic and biochemical approaches as well as electron microscopy studies have identified three evolutionarily conserved, structural modules termed “head,” “middle,” and “tail” in the core Mediator complex (27, 28). A fourth “kinase” module dynamically associates with the core complex to form a larger Mediator complex (29). The tail module interacts with the gene-specific transcription regulators, whereas the head module facilitates the direct recruitment of RNAP II and the general transcription factors to the promoter region (30, 31).

Genetic and structural studies have revealed that subunits of the head module make extensive contacts with the RNAP II subunits. X-ray diffraction analysis of the head module of yeast Mediator identified seven subunits organized into distinct domains (32). One of these subunits, Srb5/Med18, forms a heterodimer with Med20 and occupies a peripheral position in the overall architecture of the head module (32). SRB5/MED18 is not an essential gene, and srb5⁻ yeast cells are viable. In the absence of Srb5/Med18, however, cells grow about 2-fold slower than their wild type counterparts (supplemental Fig. S1). Srb5/Med18 is required for the transcription of a subset of genes in budding yeast (33). It exhibits a genetic as well as a physical interaction with several RNAP II subunits. Srb5/Med18 was discovered as one of the suppressors of the C-terminal domain (CTD) truncation mutation of Rpb1 (7). Recently, Srb5/Med18 was identified as a suppressor of an L111A mutation in the Rpb11 subunit of polymerase (34). Srb5/Med18 also exhibits a genetic interaction with Rpb4 (31). Structural studies showed the Srb5-Med20 heterodimer in close physical proximity of Rpb4/7 subunits of RNAP II (31, 32). Intriguingly, all the Srb5-interacting subunits of RNAP II (Rpb1, Rpb11, and Rpb4) are implicated in the termination of transcription. Rpb1 is not directly involved in termination, but phosphorylation of Ser-2 of its CTD is a prerequisite for the recruitment of cleavage factor 1 (CF1) 3′ end processing/termination complex toward the 3′ end of genes.

Mediator Subunit Srb5/Med18 Occupies the Terminator Region of a Subset of Genes

The interaction of Srb5/Med18 with RNAP II subunits involved in the termination of transcription prompted us to investigate its role in this process in budding yeast. Because Srb5/Med18 is not required for transcription of all genes, we selected two genes, INO1 and CHA1, whose induced transcription is dependent on Srb5/Med18 (supplemental Fig. S2, B and D). The transcription of INO1 is induced upon depletion of inositol from the medium (24), whereas CHA1 is induced in the presence of serine/threonine in the culture medium (35). We reasoned that if Srb5/Med18 plays a role in termination of transcription of INO1 and CHA1, it will cross-link to the 3′ end of these genes. ChIP analysis revealed that Srb5/Med18 indeed is localized to the terminator as well as the promoter regions of both genes upon induced transcription (Fig. 1, B, lanes 5 and 8, and E, lanes 13 and 16). Srb5/Med18 ChIP signal at the 3′ end, however, was marginally lower than that at the 5′ end (Fig. 1, C and F, gray bars, regions A and D). Furthermore, there was absolutely no signal for Srb5/Med18 in the body of the genes (Fig. 1, B, lanes 6 and 7, and E, lanes 14 and 15). These results are consistent with a role for Srb5/Med18 in the termination of transcription of these genes.

FIGURE 1.

Srb5/Med18 cross-links to the promoter and terminator regions of INO1 and CHA1. A and D, schematic depiction of INO1 and CHA1 genes indicating the positions of ChIP primer pairs used in B and E. B and E, ChIP analysis showing cross-linking of tandem affinity purification-tagged Srb5/Med18 to INO1 and CHA1. The input signal represents DNA prior to immunoprecipitation. C and F, quantification of the data shown in B and E, respectively. The error bars represent one full unit of S.D. based on a minimum of three trials.

Recruitment of 3′ End Processing/Termination Factors toward the 3′ End of Genes Is Affected in the Absence of Srb5/Med18

Mediator is a coactivator that facilitates the recruitment of general transcription factors and RNAP II onto the promoter of transcribing genes. Biochemical, genetic, and structural studies have demonstrated direct physical interaction of subunits of the head module of Mediator with TBP, TFIIB, RNAP II, and TFIIH (31, 32, 36, 37). However, there is no published study suggesting a role for Mediator in recruiting 3′ end processing/termination machinery toward the 3′ end of transcriptionally active genes. In budding yeast, there are two multisubunit complexes called the CF1 complex and the cleavage and polyadenylation factor (CPF) complex that participate in 3′ end processing of pre-mRNA as well as the termination of transcription (38). Because Srb5/Med18 was found occupying the extremities of transcriptionally induced INO1 and CHA1 genes, we examined the role of Srb5/Med18 in the recruitment of factors at both the 5′ and the 3′ end of these two genes. Our experimental strategy involved analyzing the binding of the general initiation factors TBP and TFIIB near the promoter and the 3′ end processing/termination factors Rna15 and Pta1 toward the terminator regions of INO1 and CHA1. The binding studies were performed in the cells deleted for srb5/med18 and isogenic wild type cells. Our results indicate that TBP and TFIIB occupancy of the promoter region of INO1 and CHA1 remained unaffected in the absence of Srb5/Med18 (supplemental Fig. S3B, lanes 1 and 5; supplemental Fig. S3C; supplemental Fig. S3E, lanes 9 and 13; and supplemental Fig. S3F), but cross-linking of TFIIB to the 3′ end was compromised (supplemental Fig. S3B, lanes 4 and 8; supplemental Fig. S3C; supplemental Fig. S3E, lanes 12 and 16; andsupplemental Fig. S3F). The recruitment of CF1 subunit Rna15 and CPF component Pta1 toward the 3′ end of both genes exhibited a dramatic decrease in srb5⁻ cells (Fig. 2B, lanes 4 and 8; Fig. 2C, region D; Fig. 2E, lanes 12 and 16; Fig. 2F, region D). Because Rna15 and Pta1 are integral components of the CF1 and CPF complexes, respectively, we infer that Srb5/Med18 may directly or indirectly facilitate recruitment of the CF1 and CPF complexes to the 3′ end of INO1 and CHA1.

FIGURE 2.

The recruitment of Rna15 and Pta1 to the 3′ end is adversely affected in srb5⁻ cells. A and D, schematic depiction of INO1 and CHA1 indicating the positions of ChIP primer pairs. B and E, ChIP analysis showing cross-linking of Myc-tagged Rna15 and HA-tagged Pta1 to different regions of INO1 and CHA1 upon induced transcription. The input signal represents DNA prior to immunoprecipitation. C and F, quantification of the data shown in B and E, respectively. The error bars represent one full unit of S.D. based on a minimum of three trials.

Srb5/Med18 Is Required for the Termination of INO1 and CHA1 Transcription

The failure of recruitment of CF1 and CPF 3′ end processing/termination complexes at the 3′ end of INO1 and CHA1 in srb5⁻ cells may be due to a termination defect. An initiation or an elongation defect, however, will lead to a similar consequence. To determine the step in the transcription cycle that affected the recruitment of 3′ end processing/termination factors in srb5⁻ cells, we measured RNAP II accumulation in different regions of INO1 and CHA1 during activated transcription of these genes. An RNAP II ChIP assay was performed in srb5⁻ cells and isogenic wild type cells. The results show that RNAP II was almost uniformly distributed throughout INO1 and CHA1 genes (Fig. 3, B and D, black bars) during induced transcription in wild type cells. In the absence of Srb5/Med18, however, RNAP II signal at the promoter of both INO1 and CHA1 decreased by 3–4-fold (Fig. 3, B and D, region A). These results are in agreement with the well established role of Mediator in the recruitment of RNAP II to the promoter region. Intriguingly, the polymerase concentration progressively increased from the promoter toward the terminator of both genes in srb5⁻ cells (Fig. 3, B and D). The greatest concentration of RNAP II was near the 3′ end and in the downstream intergenic region (Fig. 3, B and D, regions E, F, and G). On average, the polymerase ChIP signal near the 3′ end was about 3–4-fold higher in srb5⁻ cells as compared with the wild type cells (Fig. 3, B and D, region F). An interpretation of these observations is that RNAP II reads through the termination signal in the absence of Srb5/Med18. After passing the termination signal, RNAP II is not released from the template, but tends to accumulate beyond the 3′ end of genes. The readthrough of the termination signal and the accumulation of engaged RNAP II beyond the 3′ end of the gene are characteristics of a termination defect (39). These results, therefore, strongly argue in favor of a role for Srb5/Med18 in the termination of transcription of INO1 and CHA1.

FIGURE 3.

RNAP II accumulates in the 3′ end of genes in the absence of Srb5/Med18. A and C, schematic depiction of INO1 and CHA1 indicating the positions of ChIP primer pairs. B and D, quantification of ChIP analysis showing density of RNAP II in different regions of INO1 and CHA1 in srb5⁻ and isogenic wild type strains. The error bars represent one full unit of S.D. based on a minimum of three trials.

ChIP analysis indicates the position of template-bound RNAP II that may or may not be transcriptionally active. To corroborate the role of Srb5/Med18 in termination of transcription, it was important to show that the polymerase that was reading through the termination signal and accumulating near the 3′ end of genes was transcriptionally active. A nuclear run-on analysis was therefore carried out for CHA1, which has a long intergenic region flanking its 3′ end (Fig. 4A). The detection of any run-on transcripts now reveals active engagement of polymerase. The result shows that the polymerase reading through the terminator region of CHA1 into the intergenic region in srb5⁻ cells is transcriptionally engaged (Fig. 4, B, lanes E, F, and G, and C). No such transcription readthrough, however, was observed in wild type cells (Fig. 4, B, lanes E, F, and G, and C). These results strongly suggest that Mediator subunit Srb5/Med18 is involved in the termination of transcription of INO1 and CHA1 in budding yeast.

FIGURE 4.

RNAP II reads through the termination signal in the absence of Srb5/Med18. A, schematic depiction of CHA1 indicating the positions of TRO probes. B, TRO analysis of CHA1 in wild type and srb5⁻ cells. C, quantification of TRO data shown in B. The error bars represent one full unit of S.D. based on a minimum of three trials.

Recruitment of Mediator Complex to the Promoter of INO1 and CHA1 Is Not Affected in srb5⁻ Cells

Srb5/Med18 is not essential for cell viability. The Mediator complex retains its stability even in the absence of Srb5/Med18 (11). To determine whether the observed role of Srb5/Med18 on the termination of transcription was due to the lack of recruitment of the whole Mediator complex, rather than the specific role of Srb5/Med18, we performed ChIP analysis for Srb4/Med17 and Med15, which are the subunits of head and tail Mediator modules, respectively. Both Srb4/Med17 and Med15 were recruited to the promoter region of INO1 and CHA1 in srb5⁻ cells with wild type efficiencies (supplemental Fig. S4, B and D). Thus, the recruitment of Mediator to the 5′ end of genes is not affected in srb5⁻ cells. Taken together, our findings implicate the Srb5/Med18 subunit of Mediator in termination of transcription.

DISCUSSION

We provide four lines of evidence that Mediator subunit Srb5/Med18 is involved in the termination of transcription of a subset of genes in budding yeast. First, Srb5/Med18 cross-linked to the terminator region of INO1 and CHA1 during their activated transcriptional state. Second, CF1 and CPF 3′ end processing/termination complexes failed to get recruited near the terminator region of both genes in the absence of Srb5/Med18. Third, RNAP II ChIP demonstrated accumulation of polymerase near the 3′ end of genes and in the intergenic region. Fourth, TRO analysis confirmed a transcription readthrough phenotype in the absence of Srb5/Med18. There is a possibility that the observed defect in termination of transcription in the absence of Srb5/Med18 is an indirect effect. A termination factor may not be expressed or underexpressed in srb5⁻ cells, leading to the observed transcriptional termination defect. The cross-linking of Srb5/Med18 to the terminator region of CHA1 and INO1 upon induced transcription, however, favors the possibility of a direct role for the Mediator subunit in termination of transcription.

There are several possible explanations for the role of Srb5/Med18 in the termination of transcription. First, Srb5/Med18 may be directly facilitating the recruitment of CF1, CPF, or both complexes. Second, Srb5/Med18 may be involved in regulating phosphorylation of Ser-2 of CTD, which in turn facilitates the recruitment of CF1 and CPF complexes (40). Third, the role of Srb5/Med18 in termination may be through its interaction with the Rpb4 subunit of RNAP II. Structural studies have identified Srb5/Med18 in close physical proximity to Rpb4 in the RNAP II holoenzyme complex (31). Intriguingly, Rpb4 has been implicated in the termination of transcription in yeast (41).

Srb5/Med18 was found cross-linked to the promoter and the terminator regions of INO1 and CHA1 exclusively during their activated transcriptional states (Fig. 1, B, C, E, and F). This raises the possibility that the molecule of Srb5/Med18 cross-linking to the terminator could be different from the one occupying the promoter region. However, the promoter-terminator occupancy of Srb5/Med18 is only observed during activated transcription. It has been previously shown that yeast genes are in a looped configuration during such a transcriptionally activated state (24, 42, 43). It was also shown that TFIIB cross-links to both the 5′ and the 3′ ends only when the genes are in looped configuration (24, 42–44). The cross-linking of TFIIB to both the ends of INO1 and CHA1 upon induced transcription (supplemental Fig. S3, B, C, E, and F) suggests that the two genes are in a looped architecture with their promoter and terminator regions lying in close physical proximity. We therefore propose that it is the promoter-bound Srb5/Med18 that is contacting the juxtaposed terminator region of looped INO1 and CHA1 during induced transcription of genes.

In the absence of Srb5/Med18, a unique distribution pattern of RNAP II was observed on INO1 and CHA1. There were few RNAP II molecules in the promoter regions of these genes during induced transcription. RNAP II density progressively increased toward the 3′ end, with the highest concentration observed in the intergenic region beyond the terminator. Because genes are in a looped configuration during activated transcription in budding yeast (24, 43), our findings raise the intriguing possibility that Srb5/Med18 may be facilitating the transfer of RNAP II from the terminator to the juxtaposed promoter for reinitiation. In the absence of Srb5/Med18, such a transfer is not possible, and consequently, there is more polymerase near the 3′ end of genes as compared with the promoter. Thus, Srb5/Med18 may play a role even beyond termination in coupling termination to reinitiation of transcription.