Mediator regulates non-coding RNA transcription at fission yeast centromeres
© Thorsen et al.; licensee BioMed Central Ltd. 2012
Received: 25 August 2012
Accepted: 1 November 2012
Published: 21 November 2012
Skip to main content
© Thorsen et al.; licensee BioMed Central Ltd. 2012
Received: 25 August 2012
Accepted: 1 November 2012
Published: 21 November 2012
In fission yeast, centromeric heterochromatin is necessary for the fidelity of chromosome segregation. Propagation of heterochromatin in dividing cells requires RNA interference (RNAi) and transcription of centromeric repeats by RNA polymerase II during the S phase of the cell cycle.
We found that the Med8-Med18-Med20 submodule of the Mediator complex is required for the transcriptional regulation of native centromeric dh and dg repeats and for the silencing of reporter genes inserted in centromeric heterochromatin. Mutations in the Med8-Med18-Med20 submodule did not alter Mediator occupancy at centromeres; however, they led to an increased recruitment of RNA polymerase II to centromeres and reduced levels of centromeric H3K9 methylation accounting for the centromeric desilencing. Further, we observed that Med18 and Med20 were required for efficient processing of dh transcripts into siRNA. Consistent with defects in centromeric heterochromatin, cells lacking Med18 or Med20 displayed elevated rates of mitotic chromosome loss.
Our data demonstrate a role for the Med8-Med18-Med20 Mediator submodule in the regulation of non-coding RNA transcription at Schizosaccharomyces pombe centromeres. In wild-type cells this submodule limits RNA polymerase II access to the heterochromatic DNA of the centromeres. Additionally, the submodule may act as an assembly platform for the RNAi machinery or regulate the activity of the RNAi pathway. Consequently, Med8-Med18-Med20 is required for silencing of centromeres and proper mitotic chromosome segregation.
Mediator is a large (approximately 1 MDa) protein complex that conveys regulatory signals to RNA polymerase II (Pol II). The Saccharomyces cerevisiae Mediator was the first to be characterized but Mediators have since then been described in many other species. A comparative genomics approach of approximately 70 eukaryotic genomes shows that although its exact subunit composition varies, Mediator is conserved across the eukaryotic kingdom . The Schizosaccharomyces pombe Mediator consists of at least 20 subunits, all of which appear to have orthologues in Drosophila melanogaster, Caenorhabditis elegans and Homo sapiens.
Three distinct domains (head, middle and tail) have been identified by electron microscopy on single Mediator particles from S. cerevisiae. Electron microscopy on the S. pombe Mediator also shows a head and a middle domain, but no tail domain consistent with the lack of S. pombe orthologues of the S. cerevisiae tail components . The head domain can structurally be further divided (for example, a head domain submodule consisting of Med8-Med18-Med20 is found in both S. pombe and S. cerevisiae) [5, 6]. In S. pombe, Med27 may also be part of this submodule . A specific role for the Med8-Med18-Med20 submodule has hitherto not been described, although it is known from work in S. cerevisiae that Med18-Med20 interacts directly with the RNA Pol II subunits Rpb4 and Rpb7 .
In spite of the central role played by non-coding RNAs at S. pombe centromeres, little is known regarding the regulation of transcription in pericentromeric repeats. Transcription of the dg and dh repeats peaks during the S-phase of the cell cycle in a window where histone modifications change as a consequence of other cell-cycle regulated events [22–24]. Presently, only one promoter controlling transcription of a centromeric repeat has been described . Consistent with transcription being performed by RNA Pol II, centromeric transcripts are poly-adenylated  and specific mutations in RNA Pol II subunits impair heterochromatin formation [25, 27, 28]. The involvement of RNA Pol II in heterochromatin assembly indicates that the Mediator complex may also play a role in heterochromatin biology. Indeed, deletion of med1+ or med6+ was shown to lead to a moderate loss of centromeric silencing in a high throughput study . Further, Med15 was shown to interact with the chromatin-remodelling factor Hrp1 thus associating chromatin state with the Mediator complex . Mediator has also been associated with regulation of chromatin in HeLa cells as Med12, Med19 and Med26 interact with the silencing factor REST and the methyltransferase G9a, which methylates H3K9 at target genes [31, 32]. Here, we present a systematic analysis of S. pombe Mediator deletion mutants in relation to heterochromatin, and we identify roles played by the Med8-Med18-Med20 submodule in the transcriptional regulation of centromeric repeats and thus in heterochromatin formation, centromere function and chromosome segregation.
Genes encoding non-essential subunits of Mediator were individually deleted in FY498, a strain with the S. pombe ura4 + gene ectopically inserted in the centromere of chromosome 1, at imr1R(NcoI). In addition, a med8 ts allele  was crossed into FY498. We found that silencing of ura4 + at imr1R(NcoI) depends on all three components of the Med8-Med18-Med20 Mediator submodule, whereas the other four Mediator subunits tested (Med1, Med12, Med27, and Med31) were dispensable for silencing ura4 + at this location (Figure 1B-D). A variegated phenotype was observed for both med18 Δ and med20 Δ as some clones showed a robust silencing of ura4 + whereas others showed only weak silencing. Likewise, deletion of med1+ did occasionally show derepression of centromeric ura4 + ; however, this was a modest phenotype compared to the phenotype of med18Δ and med20Δ. Quantification of ura4 + transcript by RT-qPCR confirmed derepression of imr1R(NcoI)::ura4 + in strains with a compromised Med8-Med18-Med20 submodule (Figure 1E).
More generally, we noticed that the genome-wide expression profiles of clr4 and Mediator mutants display striking similarities indicating the Med8-Med18-Med20 submodule and H3K9me act in concert at many locations other than centromeres. A total of 42/110 genes upregulated more than 1.5x in clr4-481 are upregulated more than 2x in the med8ts mutant (; 164 genes are upregulated more than 2x in the med8ts mutant). A total of 24/58 genes upregulated more than 1.5x in clr4Δ are upregulated more than 2x in the med8ts mutant. These genes are enriched in large subtelomeric regions extending approximately 100 kb into chromosomes 1 and 2; 39/164 genes upregulated more than 2x in the med8ts mutant are subtelomeric. These regions share properties with centromeric heterochromatin [26, 38, 39] The same subtelomeric gene clusters are controlled by Spt6  suggesting Spt6, Clr4, and the Med8-Med18-Med20 Mediator submodule cooperate in heterochromatic gene silencing both at centromeres and at other chromosomal locations.
Mini-chromosome loss rate is higher in strains deleted for med18 + or med20 +
Mechanistically, the interaction between the Med18-Med20 sub-complex and the Rpb4/Rpb7 sub-complex of Pol II has been proposed to alter the conformation of the Pol II clamp domain to facilitate opening of its active-site cleft and thereby the access of promoter DNA to the Pol II cleft . This interaction would facilitate pre-initiation-complex (PIC) formation. We suggest that in heterochromatin specific interactions of other components with Mediator and/or Pol II might prevent clamp movement and thereby the productive interaction of Pol II with DNA.
Since the above proposed function of Med8-Med18-Med20 might not account for the decrease in siRNA or H3K9me in the mutants, we suggest that the Med8-Med18-Med20 submodule also facilitates the processing of long non-coding RNAs into siRNA. This second function might be carried out together with the two largest S. pombe RNA Pol II subunits, Rpb1  and Rpb2 . A mutation in Rbp2, rpb2-m203, increases the steady-state levels of centromeric transcripts and reduces siRNA to undetectable levels . The rpb2-m203 phenotype has been taken to suggest that Rpb2 provides an interaction interface with RNAi complexes and/or a means of distinguishing non-coding centromeric transcripts from mRNA, triggering processing of the former into siRNA . This presumed function of RNA Pol II, which would be compromised by the rpb2-m203 mutation, may also be affected by mutation in the Med8-Med18-Med20 submodule. A non-mutually exclusive possibility is that Med8-Med18-Med20 facilitates processing of centromeric non-coding RNA into siRNA together with Rpb1 . The S. pombe C-terminal domain of Rpb1 contains 28 conserved YSPTSPS repeats acting as an assembly platform for various mRNA processing factors, thus coupling transcription to pre-mRNA processing and export. A mutant form of Rpb1 (rpb1-11) retaining 16 of the 28 hepta-repeats apparently does not affect transcription of the pericentromeric repeats, but nevertheless compromises downstream RNAi function . As for Rpb2, given the ubiquitous interactions between the Mediator complex and active RNA Pol II, it seems plausible that a mutation in Med8-Med18-Med20 might disturb the Rpb1-dependent RNAi machinery assembly function. Alternatively, the Med8-Med18-Med20 submodule might itself be a site where pre-siRNA processing is regulated.
Consistent with our conclusions, a very recent study by Zhu and colleagues , published during the writing of this article, reports an accumulation of centromeric non-coding RNA and reduced processing of the dh repeat transcript into siRNA in a med20Δ strain. In addition, an independent large-scale epistasis map revealed genetic interactions between subunits of the Mediator and RNAi and heterochromatin components . Neither med8 nor med18 mutants were included in this screen but probing the bioGRID  with Osprey  lists 101 genetic interactions for med20 including interactions with dcr1 + , ago1 + , hrr1 + , swi6 + , cid12 + , clr3 + , hda1 + , hst2 + , pob3 + , set3 + , swc2 + and epe1 + . These interactions with heterochromatin-associated factors are fully consistent with the notion that the Med8-Med18-Med20 submodule participates in S. pombe heterochromatin formation. The data presented here, which are corroborated by Carlsten et al. , clearly demonstrate a role for Mediator in regulating centromeric chromatin.
Schizosaccharomyces pombe s trains used in the study
h + ura4-DS/E ade6-210 imr1R(NcoI)::ura4 + ori1
h + ura4-DS/E ade6-210 imr1R(NcoI)::ura4 + ori1 med12Δ::KanMX
h + ura4-DS/E ade6-210 imr1R(NcoI)::ura4 + ori1 med27Δ::KanMX
h + ura4-DS/E ade6-210 imr1R(NcoI)::ura4+ori1 med1Δ::KanMX
h+ura4-DS/E ade6-210 imr1R(NcoI)::ura4+ori1 med31Δ::KanMX
h+ura4-DS/E ade6-210 imr1R(NcoI)::ura4+ori1 med20Δ::KanMX
h+ura4-DS/E ade6-210 imr1R(NcoI)::ura4+ori1 med18Δ::KanMX
h+ura4-DS/E ade6-210 imr1R(NcoI)::ura4+ori1 dcr1Δ::KanMX
h+ura4-DS/E ade6-210/216 Ch16m23::ura4+-Tel
h+ura4-DS/E ade6-210/216 Ch16m23::ura4+-Tel med20Δ::KanMX
h+ura4-DS/E ade6-210/216 Ch16m23::ura4+-Tel med18Δ::KanMX
h A Ch16m23::ura4+-Tel leu1-32 ura4-DS/E ade6-210/216 clr4Δ::LEU2
mat1-Msmt-0 mat2-P(XbaI)::ura4+leu1-32 ura4-DS/E ade6-210 clr4Δ::LEU2
Oligonucleotides used in the study
Act1 q-PCR FW
Act1 q-PCR RV
dhA q-PCR FW
dhA q-PCR RV
ura4 q-PCR FW
ura4 q-PCR RV
RNA extraction and RT-PCR were as in  except for the final step where quantification was performed by ethidium-bromide staining using a Bio-Rad Laboratories imaging station and the Quantity One image analysis software (Bio-Rad Laboratories, Hercules, CA, USA). Primer sequences are listed in Table 3. For RT-PCR, the oligonucleotides GTO-265 and GTO-266 were used to amplify ura4+ and ura4-DS/E; GTO-223 and GTO-226 were used to amplify RNA originating from centromeric repeats or mating-type region; OKR70 and OKR71 were used to amplify actin mRNA. Strand-specific RT-PCR was achieved by using GTO-226 to prime reverse transcription on centromeric forward transcripts or GTO-223 on centromeric reverse transcripts prior PCR amplification.
RNA used in RT-qPCR was isolated using an RNeasy™ mini kit (Qiagen, Hilden, Germany) and an RNase-Free DNase set (Qiagen, Hilden, Germany). Reverse transcription of the purified RNA was performed using the RevertAidTM First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA) and random hexamer primers. qPCR was performed on a CFX96 real time PCR system (Bio-Rad Laboratories, Hercules, CA, USA) using the QuantiTect SYBR Green PCR kit (Qiagen, Hilden, Germany) supplied with SYBR Green Reference Dye. Three technical replicates were performed for each of the biological triplicates. Technical replicates with standard deviations above 10% were repeated or excluded from the experiment. Primers used to amplify act1 + and the dh repeat are shown in Table 3.
Mitotic chromosome loss was assayed as previously described  using cells containing the ade6-M210 allele on chromosome 3 and the ade6-M216 allele on the nonessential minichromosome Ch16m23::ura4+-Tel . Cells with this genotype are phenotypically Ade+ due to the interallelic complementation between ade6-M210 and ade6-M216. They form white colonies on media containing low concentrations of adenine. Loss of Ch16m23::ura4+-Tel results in red colonies or sectors. White and sectored colonies were counted following plating of the strains of interest on yeast extract plates to which no adenine had been added. The rate of minichromosome loss was determined as the number of colonies with a red sector equal to or greater than half the colony size (that is, the number of cells having lost their minichromosome at the first division after plating) divided by the number of white or sectored colonies.
For siRNA Northern blots, total RNA was isolated with Tri Reagent (Sigma Chemical Co., St. Louis, MO, USA) and 20 μg RNA was run on a 17.5% polyacrylamide/7 M urea gel and blotted onto a positive nylon membrane (Roche Diagnostics, Mannheim, Germany). siRNA were detected using a random-primed probe radioactively labeled with [α-32P]-dCTP (3000 Ci/mmol, PerkinElmer, Waltham, MA, USA). The template for random priming was a dh repeat PCR product amplified from genomic DNA with the dhH-siRNA and Cen-dh-FOR2 primers. Northern blots detecting the dg and dh repeats were obtained following electrophoresis of 10 μg total RNA prepared by a hot phenol protocol from the strains of interest. The gels used were 1% agarose in MOPS buffer with 6.7% formaldehyde. RNA was blotted onto a Hybond-XL membrane (GE Healthcare, Little Chalfont, United Kingdom). The dg and dh repeats were detected by a random-primed [α-32P]-dCTP radioactively labeled probe made on PCR products amplified from genomic DNA using p30F and p30R (dh repeat) or p33F+p33R (dg repeat). Hybridizations were performed overnight at 42°C in PerfectHyb PLUS hybridization buffer (Sigma Chemical Co., St. Louis, MO, USA).
ChIP was performed according to standard procedures. Antibodies used to immunoprecipitate RNA Pol II and H3K9me2 were ChIPAb RNA Pol II (Merck Millipore, Billerica, MA, USA) and histone H3 (dimethyl K9) antibody ChIP Grade ab1220 (Abcam, Cambridge, MA, USA), respectively. Protein G Dynabeads were used to pull down the antibody captured proteins. Rabbit Anti-Mouse Immunoglobulins (Dako, Glostrup, Denmark) were covalently coupled to the surface of Dynabeads with the Dynabeads Antibody Coupling Kit (Invitrogen, Life Technologies, Carlsbad, CA, USA) and these beads were used to pull down the Mediator complex through a TAP-tagged Med7. Presence of RNA Pol II, Mediator or dimethyl H3K9 was detected by qPCR using the primers dhA q-PCR FW and dhA q-PCR RV for the dh repeat, oMiT142 and oMiT143 for the dg repeat, or oMiT127 and oMiT128 for the putative promoter.
RNA polymerase II
Reverse transcription PCR
Quantitative reverse transcription PCR.
We are very grateful to Janne Hansen for excellent technical assistance. SH was supported by grants from the Danish Research Council, The Novo Nordisk Foundation and Manufacturer Vilhelm Pedersen and Wife Memorial Legacy (this support was granted on recommendation from the Novo Nordisk Foundation) and GT by grants from the Danish Research Council and the University of Copenhagen Center of Excellence MolPhysX.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.