A role for DNA polymerase alpha in epigenetic control of transcriptional silencing in fission yeast

Abstract

In the fission yeast Schizosaccharomyces pombe, transcriptional silencing at the mating-type region, centromeres and telomeres is epigenetically controlled, and results from the assembly of higher order chromatin structures. Chromatin proteins associated with these silenced loci are believed to serve as molecular bookmarks that help promote inheritance of the silenced state during cell division. Specifically, a chromodomain protein Swi6 is believed to be an important determinant of the epigenetic imprint. Here, we show that a mutation in DNA polymerase alpha (pol(alpha)) affects Swi6 localization at the mating-type region and causes a 45-fold increase in spontaneous transition from the silenced epigenetic state to the expressed state. We also demonstrate that pol(alpha) mutant cells are defective in Swi6 localization at centromeres and telomeres. Genetic analysis suggests that Polalpha and Swi6 are part of the same silencing pathway. Interestingly, we found that Swi6 directly binds to Pol(alpha) in vitro. Moreover, silencing-defective mutant Pol(alpha) displays reduced binding to Swi6 protein. This work indicates involvement of a DNA replication protein, Pol(alpha), in heterochromatin assembly and inheritance of epigenetic chromatin structures.

Fig. 1. Mutations in swi1, swi3 and swi7/polα suppress PEV at the mat2/3 locus and neighboring sequences. (A) The line drawing shows the physical map of the mating-type region in KΔ::ura4⁺ cells (not drawn to scale). (B) Effect of swi mutations on KΔ::ura4⁺ expression. The ura4-off derivatives of non-switching mat1-Msmto KΔ::ura4⁺ strains with the wild-type (WT, SPG32), swi7-1 (SPG106), swi1-S28 (SPG112) or swi3-146 (SPG114) mutant background were allowed to grow on YEA-rich medium. Colonies formed on YEA plates were then replicated onto AA-URA medium and incubated at 33°C for 72 h, except for the swi7-1 mutant strain that was grown for only 24 h. (C) Serial dilution plating assay. Cells were suspended in water and then 10-fold serial dilutions were spotted onto non-selective (N/S), AA-URA or conterselective FOA medium and grown for 3 days before being photographed. The KΔ::ura4⁺ strains used in (B) and (C) carried a mutation at the endogenous ura4 locus. The Luria and Delbruck fluctuation test was employed to measure the transition rates quantitatively. (D) Mutation in polα suppresses PEV of ade6⁺ expression at the L region. Cells were plated on adenine-limiting YE medium and incubated at 33°C for 3 days before being photographed. Representative colonies of wild-type (WT) or swi7-1 (swi7) mutant strains with ade6⁺ at the indicated sites in the L region are shown. The red or white colonies on YE medium imply ade6-off or ade6-on phenotypes, respectively. Strains used were: L(BglII)::ade6⁺, SPG1218 (WT) and SPG1326 (swi7); L(SacI)::ade6⁺, SPG1217 (WT) and SPG1327 (swi7).

Fig. 2. Mutation in swi7/polα causes mat2-P derepression. (A) A schematic diagram of the mat2/3 region. The trans-acting loci, clr1-clr4 and swi6, affect silencing in the entire mat2/3 interval and define one silencing pathway, while the cis-acting sequences between BglII and BssHII sites, indicated by the black box, define a second pathway specific for mat2-P silencing. (B) Effect of swi mutations and Δ(Bg-Bs)-mat2-P deletion on mat2-P expression. RT–PCR analysis of mat2-P transcripts was performed as described in Materials and methods. Strains used were as follows: mat1-Msmto mat2-P derivatives, SP1125 (WT), SP814 (swi1) SP815 (swi3) and SPG151 (swi7-1); mat1-Msmto Δ(Bg-Bs)-mat2-P derivatives, SP1152 (WT), SPG153 (swi1), SPG154 (swi3) and SPG152 (swi7-1). (C) Analysis of mat2-P derepression by iodine staining. The colonies of mat1-Msmto or mat1-Msmto Δ(Bg-Bs)-mat2-P cells carrying the wild-type (swi7⁺) or mutant swi7 (swi7^–) allele were sporulated and exposed to iodine vapor before being photographed. Because mat1-Msmto strains fail to switch, their intensity of brown staining indicates the level of ‘haploid meiosis’, a phenotype caused by expression of both P and M information in haploid cells, thus reflecting the level of mat2-P derepression.

Fig. 3. Mutation in polα affects Swi6 localization at the mat locus. (A) Competitive PCR strategy for quantitating the amount of ura4⁺ integrated at the mating-type region versus the mini-ura4 (ura4DS/E) allele at the endogenous locus. (B) The loss of silencing correlates with the decrease in Swi6 levels near mat2-P. A map of the mat2/3 interval indicating the location of the mat2-P::ura4⁺ insertion is shown (not drawn to scale). Serial dilution analysis on non-selective (N/S), AA-URA or FOA medium was used to study the effect of swi7-1 mutation on mat2-P::ura4⁺ expression (middle panel). Quantitative estimation of Swi6 levels was carried out using CHIP methodology. DNA isolated from either crude extracts or anti-Swi6 IPed fractions was subjected to competitive PCR assay. PCR products of 426 and 694 bp from ura4DS/E and mat2-P::ura4⁺, respectively, were separated on a polyacrylamide gel and then quantified using a phosphoimager. The ratio of ura4⁺ and ura4DS/E signals present in total DNA prepared from crude extract was normalized to one. Then the same correction factor was applied to the ratio of ura4⁺ to the ura4DS/E signal in IPed sample lanes to calculate the relative enrichment. Strains used were SPG1228 (WT) and SPG1229 (swi7).

Fig. 4. swi6⁺-333-mediated ‘on’ to ‘off’ state transition and the corresponding increase in Swi6 levels at the mating-type region require swi7/polα. (A) Mutation in polα affects ade6-on to ade6-off transition. Three copies of swi6⁺ inserted at its normal endogenous location (swi6⁺-333 allele) were combined with ade6-on derivatives of KΔ::ade6⁺ swi7⁺ or KΔ::ade6⁺ swi7^– strains by using genetic crosses. The resulting strains were grown on adenine-limiting YE medium before being photographed. Formation of red colonies by swi7⁺ cells indicated efficient ade6-on to ade6-off transition. In contrast, no change in ade6⁺ expression state was observed in swi7^– cells as suggested by formation of white colonies by all cells. Strains used were SPG1304 (WT) and SPG1305 (swi7). (B) The KΔ::ura4⁺ ura4-on derivatives of the swi7⁺ and swi7^– strains were combined with swi6⁺-333. The resultant cultures were subjected to CHIP analysis and serial dilution plating assay to assess the levels of Swi6 at the mat locus and KΔ::ura4⁺ expression, respectively. Strains used were SPG1301 (swi7) and SPG1302 (WT).

Fig. 5. Mutation in swi7/polα affects Swi6 localization at centromeres and telomeres. (A) A map of the cen1 indicating marker gene insertion sites (not drawn to scale) and results from serial dilution plating analysis are shown (top and middle panel). CHIP analysis was used to quantitate Swi6 levels at cnt, imr and otr sites within cen1 (bottom panel). Strains used were: cnt1::ura4⁺, FY336 (WT) and SPG1060 (swi7); imr1R::ura4⁺, FY498 (WT) and SPG1062 (swi7); otr1R::ura4⁺, FY648 (WT) and SPG1067 (swi7). (B) Multiplex PCR was used to examine Swi6 levels at telomeres. DNA isolated from crude extracts or anti-Swi6 IPed fractions was analyzed for telomere-associated sequences (TAS) as well as mat2-P::ura4⁺ and ura4DS/E sequences, used as controls. A diagram of a telomere indicating TAS, short repeats (gray triangles) and PCR primer-binding sites (black triangles) is shown (top panel). The PCR product marked with an asterisk resulted from primers binding to an additional site within TAS, as confirmed by its cloning and sequencing. Strains used were SPG1228 (WT) and SPG1229 (swi7).

Fig. 6. Direct interaction between Polα and Swi6. (A) Binding of Polα to Swi6 protein. Yeast cell lysates from wild-type (WT; SP976) or polα mutant (swi7; SPJ5) cells were loaded onto a Swi6-coupled affinity column (+) or a mock-coupled blank column (–). The bound proteins were eluted and subjected to western blotting using affinity-purified anti-Polα antibodies. (B) The N- and C-terminal parts of Polα or luciferase protein were expressed in vitro and incubated with GST or GST–Swi6 beads. After extensive washes, the proteins that remained bound to beads were separated using SDS–PAGE and detected by fluorography. (C) The C-terminal domain of Polα contains the MIR consensus sequence found in mouse TIF1α, TIF1β and CAF-1, and in Drosophila Su(var)3–7. Corresponding regions of Polα in other species are also aligned (Hu, human; Mu, mouse; Dm, Drosophila). Conserved residues are indicated by boxes.

Fig. 7. Mutation in swi7/polα disrupts the Swi6 localization. Wild-type (A and B) or swi7/polα mutant cells (C and D) were stained with affinity-purified anti-Swi6 antibody. DAPI staining of each cell is also shown (E–H). Strains used were SP976 (WT) and SPJ5 (swi7).