TRAP230/ARC240 and TRAP240/ARC250 Mediator subunits are functionally conserved through evolution

Abstract

In Saccharomyces cerevisiae Mediator, a subgroup of proteins (Srb8, Srb9, Srb10, and Srb11) form a module, which is involved in negative regulation of transcription. Homologues of Srb10 and Srb11 are found in some mammalian Mediator preparations, whereas no clear homologues have been reported for Srb8 and Srb9. Here, we identify a TRAP240/ARC250 homologue in Schizosaccharomyces pombe and demonstrate that this protein, spTrap240, is stably associated with a larger form of Mediator, which also contains conserved homologues of Srb8, Srb10, and Srb11. We find that spTrap240 and Sch. pombe Srb8 (spSrb8) regulate the same distinct subset of genes and have indistinguishable phenotypic characteristics. Importantly, Mediator containing the spSrb8/spTrap240/spSrb10/spSrb11 subunits is isolated only in free form, devoid of RNA polymerase II. In contrast, Mediator lacking this module associates with the polymerase. Our findings provide experimental evidence for recent suggestions that TRAP230/ARC240 and TRAP240/ARC250 may indeed be the Srb8 and Srb9 homologues of mammalian Mediator. Apparently Srb8/TRAP230/ARC240, Srb9/TRAP240/ARC250, Srb10, and Srb11 constitute a conserved Mediator submodule, which is involved in negative regulation of transcription in all eukaryotes.

The Mediator complex is essential for basal and regulated expression of nearly all RNA polymerase II (pol II)-dependent genes in Saccharomyces cerevisiae (1), and depletion of human Mediator from nuclear extracts abolishes transcription by pol II (2). Mediator conveys regulatory information from enhancers and other control elements to the promoter. The functional activities identified for Mediator include stimulation of basal transcription, support of activated transcription, and enhancement of phosphorylation of the C-terminal domain (CTD) of pol II by the transcription factor IIH kinase (3, 4). Mediator-like complexes have been isolated from Caenorhabditis elegans, Drosophila, mouse, and human cells (5).

The SRB genes encode a prominent group of Mediator subunits. These genes were originally identified by Nonet and Young (6) in a genetic screen in S. cerevisiae for suppressors of the cold-sensitive phenotype associated with a truncation of the CTD. A subgroup of Srb proteins (Srb8, Srb9, Srb10, and Srb11) form a specific module, which is present in holoenzyme preparations from cells growing exponentially in rich glucose medium, but is absent in stationary-phase cells (7, 8). The SRB11 and SRB10 genes encode cyclin C and the cyclin C-dependent kinase, respectively (9). Genetic analysis indicates that the Srb8–11 module is involved in the negative regulation of a small subset of genes (10). Srb8 is required for stable association of Srb10 and Srb11 with the holoenzyme inasmuch as holoenzyme preparations from srb8 deletion strains lack Srb10 and Srb11 (11). Homologues of Srb10 and Srb11 are found in some human Mediator preparations, whereas no homologues of Srb8 and Srb9 have been reported. Recently, however, we and others have suggested that weak similarities on the primary sequence level may exist between Srb8 and TRAP230/ARC240 (12–14), as well as between Srb9 and TRAP240/ARC250 (13, 14).

Here we identify a Mediator complex in Schizosaccharomyces pombe containing conserved homologues not only of Srb8, Srb10, and Srb11, but also of mammalian TRAP240/ARC250 (spTrap240). Genetic characterization and gene expression analysis demonstrates a close functional relationship between spTrap240 and spSrb8. Our findings offer, to our knowledge, the first experimental evidence for the existence of a conserved Mediator submodule containing Srb8/TRAP230/ARC240, Srb9/TRAP240/ARC250, Srb10, and Srb11. Our data thus suggest that the mechanisms for Mediator-dependent transcriptional repression are the same in all eukaryotes from yeast to humans.

Materials and Methods

Immunoblot Analyses. We performed immunoblot analysis for spSrb4, Pmc3, and Pmc6 as described (15). The tandem affinity purification (TAP)-tag was identified with rabbit antiperoxidase IgG (16). Dot blot analysis was performed as described (17). Protein concentrations were determined in triplicate with the Bradford reagent, and the amounts of cell extracts applied to dot blots were adjusted accordingly.

Protein Purification and Identification. We purified Mediator/pol II holoenzyme from TAP-spTrap240 and TAP-spMed7 Sch. pombe strains by following the TAP-tag purification protocol (16), with the following modifications. We grew 20 liters of Sch. pombe to midlogarithmic phase in yeast extract supplement (YES) medium (18) supplemented with 0.2 g/liter adenine. Cells were collected by centrifugation (JA-10, Beckman Coulter, at 2,500 rpm, for 7 min at +4°C), washed once with ice-cold water, and suspended in 0.5 ml of buffer A [200 mM KOH–Hepes, pH 7.8/15 mM KCl/1.5 mM MgCl₂/0.5 mM EDTA/15% glycerol/0.5 mM DTT, and protease inhibitors] per g of cell pellet. Cells were lysed by bead beating (Bead-Beater, Stratech, London) at 25 cycles, with each cycle consisting of 30 s of beating and 90 s of rest. We cleared the supernatant by centrifugation (JA-10, at 9,000 rpm, for 15 min, at +4°C), added 1/9 vol of 2 M KCl, and stirred for 15 min. After ultracentrifugation (Ti45, Beckman Coulter, at 42,000 rpm, for 20 min, at +4°C), we added 500 μl of IgG beads (Amersham Biosciences) and incubated for 1 h at +4°C. IgG beads were collected by centrifugation (JA-17, Beckman Coulter, at 1,000 rpm, for 2 min, at +4°C), loaded into a column, and washed with 30 ml of IgG buffer without Nonidet P-40. After washing with 20 ml of tobacco etch virus (TEV) protease cleavage buffer, we eluted the Mediator by incubation for 1 h at +16°C with 200 units of TEV protease in 2 ml of TEV buffer. The identification of spSrb8, spTrap240, spSrb10, and spSrb11 by matrix-assisted laser desorption ionization/time-of-flight (MALDI-TOF) MS analysis of in-gel digested proteins was described (19).

SRB10/11 Kinase Assay. The Srb10/11 CTD-kinase assay was performed as described (15).

DNA Microarray Analysis. Sch. pombe strains were grown in YES medium at 30°C until harvested at 5 × 10⁶ to 1 × 10⁷ cells per ml. RNA was extracted with hot phenol and was further purified by using RNeasy (Qiagen, Valencia, CA). We used Sch. pombe DNA microarrays from Eurogentech (Brussels) spotted with 5,029 annotated ORFs. Methods for preparation of cDNA, labeling, and hybridization will be published elsewhere. For each mutant five array experiments were performed, including dye swap. Arrays were read on a GMS418 array scanner (Genetic Microsystems, Woburn, MA) and images were analyzed with IMAGENE 4.0 and GENESITE LIGHT (BioDiscovery, Marina del Rey, CA). For each ORF the presented number represents data from at least four arrays. All numbers have standard deviations <20% and are within 95% significance.

Genetic Manipulation. Strains used are as follows: L972 (h⁻), kindly provided by M. Sipiczki (University of Debrecen, Debrecen, Hungary); Mp12 (h⁻ ade6-M210) and Mp13 (h⁺ ade6-M216), kindly provided by I. Hagen (Paterson Institute for Cancer Research, Manchester, U.K.); TP28 (h⁻ Δspsrb8::G418^R), TP33 (h⁻ Δsptrap240::G418^R), TP43 (h⁻ ade6-M216 spmed7⁺::TAP), and TP68 (h⁺ ade6-M216 sptrap240⁺::TAP). A C-terminally TAP-tagged spMed7⁺ strain was made by using long flanking PCR-generated fragments. Primers used for strain construction are listed in Table 4, which is published as supporting information on the PNAS web site, www.pnas.org. The two PCR-generated fragments (with genomic WT Sch. pombe DNA as template) were ligated into plasmid pFA6a-kanMX6-CTAP2 (20) by using restriction sites SalI, XmaI, SacI, and EcoRV, respectively. Cloned fragments were then sequenced. For yeast transformation the generated plasmid was digested with AvrII and EcoRV. A C-terminally TAP-tagged spTRAP240⁺ strain was made by using short flanking PCR-generated fragments with plasmid pFA6a-kanMX6-CTAP2 as template (20). TAP-spMed7⁺ and TAP-spTRAP240⁺ were generated in the diploid strain Mp12/Mp13 by lithium acetate transformation (21). G418-resistant colonies having correctly inserted TAP-tags were subjected to random spore analysis (22). In all cases G418 resistance segregated 2:2. Haploid G418-resistant strains were genotyped and subjected to Western analysis with antiperoxidase IgG. The TAP-tagged strains were phenotypically WT. Null mutants of either sptrap240⁺ or spsrb8⁺ were prepared by using the primers in Table 4 as described (19) in the diploid strain Mp12/Mp13. The obtained G418-resistant diploids were sporulated, and asci (>15) were dissected on YES plates. All four spores from tetrads of both diploids germinated and segregated 2:2 for G418 resistance, demonstrating that sptrap240⁺ and spsrb8⁺ are nonessential genes. For microarray analysis both deletions were made in the haploid h⁻ strain L972. For verification G418-resistant colonies were subjected to PCR and Southern blot analysis.

Flocculation Assay. A 10-ml culture grown to late exponential phase in YES medium was sonicated, the tube was vigorously vortexed, and cells were allowed to settle. The top 100 μl of cell suspension was removed at the appropriate time, spun down, and resuspended in 50 mM EDTA, and the cell number was determined by reading OD₆₀₀.

Results

We have previously used conventional chromatography to purify Sch. pombe Mediator in complex with pol II (15, 19). This pol II holoenzyme did not contain homologues of Srb8, Srb9, Srb10, and Srb11, which in S. cerevisiae form a submodule present in certain holoenzyme preparations (the Srb8–11 module; ref. 11). To investigate the presence of alternative forms of Mediator in Sch. pombe we fused a TAP tag in-frame with the spMed7 core Mediator subunit and used the TAP-tag purification protocol (16). In S. cerevisiae, members of the Srb8–11 module are broken down under conditions of nutrient limitations (7, 8). We therefore harvested cells early and tried to minimize proteolysis by rapid whole-cell extraction. Holoenzyme purified from these extracts contained not only core Mediator subunits and pol II but also, as revealed by Coomassie brilliant blue staining, two additional weak protein bands in the high molecular weight range (Fig. 1A).

Fig. 1.

Substoichiometric amounts of spSrb8 and spTrap240 in Mediator prepared from Sch. pombe.(A left) High molecular weight region of Mediator purified over IgG, from TAP-spMed7 cells. Proteins were separated by SDS/10% PAGE and revealed by staining with Coomassie blue. (Right) Densitometry of the SDS gel showed that the Mediator preparation contained substoichiometric amounts of Rbp1, Rbp2, spSrb8, and spTrap240 compared with the core Mediator subunit spRgr1/Pmc1. (B) Dot blotting with rabbit anti-peroxidase IgG of dilution series of whole-cell extracts from Sch. pombe (TAP-spMed7 and TAP-spTrap240). Crude extracts were prepared and blotted as described. The protein quality in our extracts was verified by separation of proteins on an SDS/10% polyacrylamide gel and immunoblot analysis.

The larger of the two proteins migrated with an apparent molecular mass of 140 kDa and was identified with MALDI-TOF mass fingerprinting as the product of SPAC589.02C, an uncharacterized ORF (Table 1). We named this protein spTrap240, because a BLAST sequence similarity search demonstrated significant similarity to the mammalian Mediator component TRAP240/ARC250 (23, 24). TRAP240/ARC250 displayed 26% identity and 44% similarity (E = 8.5 × 10⁻⁴) in a region spanning amino acids 660–811, and 23% identity and 44% similarity (E = 8.5 × 10⁻⁴) in a region spanning amino acids 1025–1071 in spTrap240. The smaller protein of ≈130 kDa was the product of SPAC688.08 (Table 1). This gene product demonstrated significant similarity with S. cerevisiae Srb8 (25% identity, 44% similarity, E = 6.8 × 10⁻⁹, in a region spanning amino acids 31–269 in Srb8). We named this protein spSrb8.

Table 1. Proteins in the spTrap240/Mediator complex.

Protein	No. of peaks	Mass, kDa	Deletion phenotype	Gene name*
spTrap240	16	139	Viable	SPAC589.02c
spSrb8	22	131	Viable	SPAC688.08
spSrb10	12	41	Viable†	SPAC23H4.17c
spSrb11	3	26	ND	SPBC12D12.06

Verified with MALDI-MS postsource decay analysis of the peptide ion at m/z 1068.72 (VLATAIVLLR). ND, no data available.

^*From the Sanger Centre, www.sanger.ac.uk.

^†Results taken from ref. 31.

We noted that the presence of spSrb8 and spTrap240 in our pol II holoenzyme preparation was accompanied with lower levels of pol II. To verify this observation, we tried to quantify individual subunits with densitometry (Fig. 1 A Right). Rpb1 and Rpb2 were present at ≈30% of the level of the spRgr1/Pmc1 core Mediator component, whereas spSrb8 and spTrap240 were present at ≈15% (Fig. 1 A; data not shown).

The substoichiometric amounts of spSrb8 and spTrap240 suggested that these subunits might be present in a distinct form of the Mediator complex. To test this hypothesis, we constructed a Sch. pombe strain in which the spTRAP240 gene product was fused in-frame with the TAP-tag. We first quantified the relative levels of TAP-tagged spTrap240 and spMed7 in the whole-cell extracts by using dot blot analysis (Fig. 1B). The TAP-tag can easily be detected in whole-cell extracts by using anti-peroxidase IgG (16), and this technique was recently used to estimate amounts of general transcription factors and Mediator subunits in S. cerevisiae (17). In nice agreement with our purified pol II holoenzyme, we found that spTrap240 is 5- to 10-fold less abundant in our Sch. pombe extracts than in the spMed7 core Mediator component. A similar experiment in S. cerevisiae, comparing the levels of Srb11 with the Med8 core Mediator component gave comparable results (data not shown). We next used the TAP-tag on spTrap240 for purification. We observed copurification of spTrap240 with the known Mediator components, spSrb4, Pmc3, and Pmc6 over IgG Sepharose (Fig. 2A). After purification over calmodulin Sepharose, we separated the spTrap240 Mediator complex on SDS/PAGE and stained with silver (Fig. 2B). Immunoblot and MALDI-TOF mass fingerprinting analysis identified spSrb8 as well as several previously characterized Mediator components in the purified Mediator. Surprisingly, we could not detect any pol II subunits in our analysis. The absence of pol II was further verified by immunoblotting (Fig. 2C). Apparently the spSrb8/spTrap240-containing complex, in contrast with Mediator we previously isolated, does not associate with pol II. The conclusion could be that the TAP-Med7 holoenzyme preparations (Fig. 1A) were, in fact, a mixture of different Mediators, one containing spSrb8/spTrap240, and another lacking these subunits, but associating with pol II. In support of this notion we could separate TAP-Med7 Mediator over heparin Sepharose and identify two separate Mediator peaks, one containing the pol II holoenzyme, and a second containing spSrb8/spTrap240, but lacking pol II (data not shown).

Fig. 2.

A spTrap240 containing Mediator complex. (A) TAP-Trap240 copurifies with Mediator proteins over IgG Sepharose. We analyzed the input (whole-cell extract, 15 μl) and the eluted proteins from the IgG Sepharose (15 μl) by immunoblotting. (B) After TAP purification, the TAP-spTrap240-containing Mediator complex was analyzed in an SDS/10% polyacrylamide gel and revealed by silver staining. The two bands denoted with asterisks (*) are contaminants corresponding to ribosomal proteins. (C) Immunoblotting reveals that Rpb1 is present in TAP-spMed7-purified Mediator, but is absent in the TAP-spTrap240 Mediator preparation. (D) CTD kinase assay with TAP-spMed7 and TAP-spTrap240 Mediator preparations. Purified S. cerevisiae pol II (100 ng per reaction) was used as substrate.

The TAP-spTrap240 Mediator also contained two previously uncharacterized proteins in the low molecular range. We identified these proteins as the Sch. pombe homologues of Srb10 (spSrb10) and Srb11 (spSrb11; Table 1). S. cerevisiae Srb10 has a CTD-kinase activity, and in agreement with this result, we detected significant CTD phosphorylation with the spTrap240/Mediator, but no detectable levels of phosphorylation with the pol II holoenzyme (Fig. 2D).

The association of spTrap240 and spSrb8, spSrb10, and spSrb11 in one form of the Sch. pombe Mediator indicated that these four proteins formed a subcomplex, similar to what has been described for the Srb8–11 complex in S. cerevisiae. A BLAST sequence similarity search did not reveal any spTrap240 homologues in the S. cerevisiae protein database (data not shown). However, we could find a clear homology between spTrap240 and Candida albicans Srb9 (caSrb9) and generate a multiple alignment with spTrap240 and TRAP240/ARC250 from human and Drosophila melanogaster (Fig. 3). This similarity has also been reported in two recent publications (13, 14). We found the possibility attractive that TRAP240 in Sch. pombe and higher eukaryotic cells were homologous to Srb9. Human Mediator has been reported to exist in two major forms, one larger (ARC-L, SMCC/TRAP) containing TRAP230/ARC240, TRAP240/ARC250, hSRB10, and hSRB11, and one smaller form (PC2, CRSP) lacking these proteins (25, 26, 27). The smaller CRSP interacts with the pol II CTD, whereas the larger ARC-L does not. We reasoned that TRAP230/ARC240 and TRAP240/ARC250 indeed could be the mammalian homologues of Srb8 and Srb9. In support of this notion, we and others (12, 13, 14) have also previously reported that TRAP230/ARC240 displays weak sequence similarity to Srb8. Because Sch. pombe contained conserved homologues of Srb8 and TRAP240/ARC250, we could explore a possible functional relationship between these two proteins. To this end we first compared the genetic and phenotypic characteristics of spsrb8⁺ and sptrap240⁺. First, strains with null mutations of spsrb8⁺ and sptrap240⁺ strains were constructed in the diploid Mp12/Mp13 (h⁻/h⁺ ade6-M210/ade6-M216). After sporulation and tetrad analysis, viability segregated 4:0, demonstrating that the two gene products are nonessential. For transcriptome analysis the null mutants of spsrb8⁺ and sptrap240⁺ were prepared in the haploid h⁻ strain L972. The two haploid deletion strains showed reduced growth rate in YES medium compared with that of WT cells. Both deletion strains are highly flocculent, the flocculation is divalent cation dependent, and the flocculation characteristics of the two strains are indistinguishable (Fig. 4). Thus, genetically and phenotypically, our analysis suggests a functional relationship between spsrb8⁺ and sptrap240⁺.

Fig. 3.

Alignment of spTrap240 and homologues from higher eukaryotic cells. The database accession codes for the proteins are NP_594050 (Sch. pombe), NP_005112 (Homo sapiens), and AAF50591 (D. melanogaster). The Srb9 sequence from C. albicans was obtained by an analysis of unfinished sequence data from the C. albicans genome project (www-sequence.stanford.edu/group/candida/index.html).

Fig. 4.

Flocculation level of Δspsrb8, Δsptrap240, and WT Sch. pombe under rich growth conditions. The cell number in the supernatant was determined by an OD at 600 nm at the indicated times. Both Δspsrb8 and Δsptrap240 are highly flocculent.

We next used whole-genome DNA microarrays to investigate the generality of spTrap240 and spSrb8 requirement for gene expression. If spTrap240 and spSrb8 belong to the same subcomplex, they should also regulate the same set of genes in vivo. Differences in specific transcript levels between Δspsrb8 or Δsptrap240 and WT strains were determined under rich growth conditions (Tables 2 and 3). The global changes in gene expression were surprisingly mild for both the Δspsrb8 and Δsptrap240 mutant. We classified genes into spTrap240/spSrb8-dependent and spTrap240/spSrb8-repressed, based on ratios ≤0.58 and ≥1.7, respectively. A subset of only 10 genes had their expression levels increased when the sptrap240 deletion mutant was compared with WT (complete data are available on request). The effects of the Δspsrb8 mutation correlated perfectly with the effects of Δsptrap240: all 10 genes induced in the Δsptrap240 mutant were also increased in the Δspsrb8, demonstrating that the two Mediator subunits have similar effects on gene expression and that both play a role in negative regulation of transcription. In support of this notion we also analyzed transcript levels under poor growth conditions and, although we observed more affected genes, the expression patterns were still identical for Δspsrb8 and Δsptrap240 (C.O.S., unpublished observations). Some of the genes most affected were homologues to the S. cerevisiae FLO genes, possibly explaining the observed flocculation phenotype. Our transcriptome analysis also revealed that both sptrap240⁺ and spsrb8⁺ have positive regulatory functions (Table 3). Four genes had reduced expression levels when the sptrap240 deletion mutant was compared with the wild type. Again, the effects of the Δsptrap240 mutation correlated perfectly with the effects of Δspsrb8. All four genes reduced in the Δsptrap240 strain were also reduced in the Δspsrb8 strain. We conclude that the two Mediator subunits have similar effects on gene expression and that both are part of the same Mediator submodule.

Table 2. Genes up-regulated in Δsptrap240 and Δspsrb8.

			Fold up-regulation
ORF	Gene	Description	Δsptrap240	Δspsrb8
SPBPB2B2.01	—	Putative amino acid permease	3.627	3.708
SPCPB16A4.07c	—	Homologue to S. cerevisiae FLO1	3.039	4.165
SPAPB24D3.07c	—	No information	2.964	3.276
SPAC56F8.12	—	No information	2.421	2.190
SPBC947.04	—	Homologous to S. cerevisiae FLO1, FLO9, and FLO5	2.293	2.818
SPAC186.01	—	Putative cell-surface Ser/Thr-rich protein	1.964	3.246
SPBC1198.02	—	Putative adenosine deaminase	1.893	1.882
SPAC343.12	rds1	Stress-response protein	1.792	2.087
SPBC1861.02	abp2	Autonomously replicating sequence-binding protein	1.747	2.063
SPAC1F8.06	—	Putative cell-surface protein	1.706	1.874

Table 3. Genes down-regulated in Δsptrap240 and Δspsrb8.

			Fold down-regulation
ORF	Gene	Description	Δsptrap240	Δspsrb8
SPACUNK4.10	—	Putative 2-hydroxyacid dehydrogenase	0.585	0.576
SPBC31F10.03	—	No information	0.539	0.442
SPAC2E1P3.05c	—	Putative cellulose binding β-glucosidase	0.512	0.442
SPBPB7E8.01	—	No information	0.465	0.427

Discussion

Here, we present experimental evidence for the notion that TRAP230/ARC240 and TRAP240/ARC250 are the functional homologues of Srb8 and Srb9. We have isolated a form of Sch. pombe Mediator, which contains conserved homologues not only of Srb8, Srb10, and Srb11 but also of mammalian TRAP240/ARC250. We find identical phenotypes for Δsptrap240 and Δspsrb8 deletions, and DNA microarray analysis reveals that spSrb8 and spTrap240 are involved in the control of the expression of the same distinct subset of genes.

The few genes found to be affected in Δsptrap240 and Δspsrb8 deletion strains were unexpected, but they correlate well with the displayed phenotypes. Apart from flocculation and a slightly lower growth rate, the Δsptrap240 and Δspsrb8 deletion strains behave essentially as WT. We do not believe that our results reflect a lower general dependence of Mediator for transcriptional regulation in Sch. pombe because deletion of another nonessential subunit, Pmc6, generated a more severe phenotype and affected many more genes (>150) in DNA microarray experiments (C.O.S., unpublished observation). Our observations are also in agreement with experiments in S. cerevisiae, where mutations in the conserved Srb10 subunit have very limited effect on global gene expression compared with other Mediator subunits, such as Srb4, Srb5, and Med6 (10). However, direct comparison between our data and gene expression data for S. cerevisiae Δsrb10 is difficult, because certain promoters reportedly differ in their sensitivity to Δsrb9 and Δsrb10 deletions (28).

One of the most affected genes in Δsptrap240 and Δspsrb8 is the homologue of S. cerevisiae FLO1 (29). This gene, which is also up-regulated in the S. cerevisiae srb10 kinase dead mutant strain (10), encodes a lectin-like cell-wall protein, which adheres to cell-wall components and causes aggregation of cells. The up-regulation of FLO1 could thus explain the flocculation phenotype observed for the srb8–11 gene deletions in both S. cerevisiae and Sch. pombe. Increased flocculation has also been reported by deletion of the prk1⁺/srb10⁺ and lkh⁺ genes, respectively (30, 31). Flocculation of the Δsrb10 strain further strengthens the functional relationship among the Srb8–11 subunits. In the case of deleting lkh⁺, the cells also exhibited filamentous adhesion. This last phenotype is not seen in either the Δspsrb8 or the Δsptrap240 strain, indicating regulation through different signaling pathways.

Our data suggest that Mediator interacts with either pol II or Srb8–11. This observation contradicts a recent report (32) on the isolation of a complex containing pol II, Srb8–11, and core Mediator from S. cerevisiae. However, this form of the holoenzyme was isolated with a FLAG-tag on Srb5, Srb6, or Rgr1, which are all core Mediator components, and the authors also report that the levels of Srb8–11 and pol II were substoichiometric to core Mediator components. Based on our observations, we therefore suggest that the described complex actually may correspond to a mixture of Mediator containing the Srb8–11 and Mediator lacking Srb8–11, but associating with RNA polymerase II. In agreement with this notion, we can isolate the S. cerevisiae Mediator with a TAP-tag on Srb11 and we cannot detect any pol II in these preparations (data not shown).

Recently Tjian and coworkers (33) demonstrated a specific interaction between the human CRSP Mediator complex and CTD. Interestingly the larger form of human Mediator, ARC-L, could not bind to CTD under these conditions. ARC-L and CRSP differ only slightly in subunit composition, because CRSP lacks TRAP230/ARC240, TRAP240/ARC250, hSrb10, and hSrb11, which are all present in ARC-L. These findings correlate nicely with the absence of pol II from Sch. pombe Mediator containing spSrb8/spTrap240/spSrb10/spSrb11 and would indicate that the ability of the Srb8–11 module to negatively regulate Mediator-pol II interactions is evolutionary conserved.

One obvious reason for the absence of the Srb8–11 module in the original Mediator preparation is its negative effects on transcription, because Mediator was originally purified on the basis of its ability to support activated transcription (3). The Srb8–11 submodule is degraded when yeast are grown under conditions of nutrient limitations (7, 8) and are also sensitive to the extraction procedure (data not shown). It is tempting to speculate that this degradation is a specific process and that the switch between Srb8–11 containing Mediator, and the smaller, active form lacking Srb8–11, may be a regulated event. Our results suggest that a substantial fraction of the Mediator complexes present in yeast contains the Srb8–11 module. This finding is in agreement with observations in higher eukaryotes, where Srb8–11-containing Mediator seems to be the predominant form. It is thus possible that the Srb8–11 proteins play a general role in transcriptional regulation. Perhaps the submodule is present in all nonactivated Mediator complexes and represses promoter-dependent transcription before activation. A possible model for the function of Srb8–11 is depicted in Fig. 5. Mediator in free form is in complex with Srb8–11. Activators interact directly with this pol II-free Mediator and recruit the complex to specific promoters. On signaling, the Srb8–11 complex is degraded, allowing interactions between Mediator and pol II. This simple model could perhaps also explain the mild phenotypes that are identified for Δspsrb8 and Δsptrap240. The recruitment of Mediator to the yeast HO promoter and to Drosophila HSF-binding sites without pol II supports this view (34, 35, 36). Thus, lack of Srb8–11 may have more profound effects on the kinetics/timing of transcriptional activation at specific promoters, rather than on the overall levels of gene expression.

Fig. 5.

A transcription activation domain (TAD) interacts directly with Srb8–11-containing Mediator and recruits the complex to specific promoters. The Srb8–11 complex is degraded at the promoter, allowing interactions between Mediator and pol II.

Here, we have identified spTrap240 as the Srb9 homologue of Sch. pombe Mediator. This identification provides evidence that the TRAP230/TRAP240/Srb10/Srb11 complex in human Mediator is homologous of the Srb8–11 complex in S. cerevisiae. We believe these observations further strengthen the functional similarities between yeast and mammalian Mediator. It appears increasingly clear that the basic mechanisms of Mediator-dependent transcriptional activation are conserved through evolution. Lessons from yeast can evidently be applied to the understanding of Mediator complexes in higher cells. A molecular description of the mechanisms by which the Srb8–11 interaction with core Mediator is regulated may prove central to our understanding of Mediator-dependent transcriptional repression and perhaps also activation, in all eukaryotes, from yeast to humans.