Establishment of expression-state boundaries by Rif1 and Taz1 in fission yeast

Abstract

The Shelterin component Rif1 has emerged as a global regulator of the replication-timing program in all eukaryotes examined to date, possibly by modulating the 3D-organization of the genome. In fission yeast a second Shelterin component, Taz1, might share similar functions. Here, we identified unexpected properties for Rif1 and Taz1 by conducting high-throughput genetic screens designed to identify cis- and trans-acting factors capable of creating heterochromatin-euchromatin boundaries in fission yeast. The preponderance of cis-acting elements identified in the screens originated from genomic loci bound by Taz1 and associated with origins of replication whose firing is repressed by Taz1 and Rif1. Boundary formation and gene silencing by these elements required Taz1 and Rif1 and coincided with altered replication timing in the region. Thus, small chromosomal elements sensitive to Taz1 and Rif1 (STAR) could simultaneously regulate gene expression and DNA replication over a large domain, at the edge of which they established a heterochromatin-euchromatin boundary. Taz1, Rif1, and Rif1-associated protein phosphatases Sds21 and Dis2 were each sufficient to establish a boundary when tethered to DNA. Moreover, efficient boundary formation required the amino-terminal domain of the Mcm4 replicative helicase onto which the antagonistic activities of the replication-promoting Dbf4-dependent kinase and Rif1-recruited phosphatases are believed to converge to control replication origin firing. Altogether these observations provide an insight into a coordinated control of DNA replication and organization of the genome into expression domains.

Keywords: DNA replication program; chromatin boundaries; fission yeast; gene silencing; heterochromatin.

Fig. 1.

Boundary trap and trapped elements. (A) Edge of heterochromatic domain in the mating-type region showing the two reporters used for the screen: (EcoRV)::ade6⁺ and (BlpI)::LEU2. (B) Boundary effects of trapped elements. Ten-fold dilutions of strains with indicated IR-R replacements were spotted onto selective media to monitor expression of (EcoRV)::ade6⁺ and (BlpI)::LEU2. Red colony color on YE indicates repression of (EcoRV)::ade6⁺. STAR3 and STAR4 were each independently isolated from two genomic libraries. IR-R⁺, PG3950; IR-RΔ, PG3947. (C) Quantification of (EcoRV)::ade6⁺ and (BlpI)::LEU2 transcripts by RT–quantitative PCR (RT-QPCR) for strains displayed in B, expressed as percent of act1⁺ transcript. (D) Endogenous chromosomal locations of trapped elements, shown in red. Corresponding heterochromatic islands are numbered in red; other islands are in black.

Fig. S1.

Trapped genomic elements that weakly restored (EcoRV)::ade6⁺ repression. Ten-fold serial dilutions of the indicated strains were spotted onto selective or indicator media. IR-R⁺, PG3950; IR-RΔ, PG3947.

Fig. 2.

Boundary formation by STAR elements requires Taz1 and Rif1. (A) Sequence of smallest subclones of STAR elements capable of restoring (EcoRV)::ade6⁺ repression, with putative Taz1-binding sites in red. (B) Repression of (EcoRV)::ade6⁺ by STAR subclones, assayed as in Fig. 1C. (C–E) Derepression of (EcoRV)::ade6⁺ in taz1Δ and rif1Δ mutants with indicated boundary elements, and in a double taz1Δ rif1Δ mutant with the STAR2 boundary, assayed as in Fig. 1 B and C . The (EcoRV)::ade6⁺ transcript levels were normalized to act1⁺ measured in the same samples and to the (EcoRV)::ade6⁺ transcript in IR-RΔ cells propagated in parallel.

Fig. S2.

Deletion analysis of boundary elements (part 1). (A) Location of subclones relative to H3K9me profiles (dotted lines) and Taz1 binding sites (yellow crosses). The primers used to generate these subclones, and the corresponding plasmids and strains are listed in Table S4. (B) (EcoRV)::ade6⁺ and (BlpI)::LEU2 expression tested in strains with the indicated subclones in the place of IR-R. The IR-RΔ strain spotted was PG2897, used as recipient for boundary replacements; it does not contain (BlpI)::LEU2.

Fig. S3.

Deletion analysis of boundary elements (part 2). (A and B) Same as Fig. S2 for the three other elements.

Fig. S4.

Test of chromosomal elements for boundary activity. (A) Tenfold serial dilution series of indicated strains spotted on YE plates. In each strain, the IR-R boundary was replaced by a chromosomal element endogenously located near a late/inefficient origin that shows increased activity in taz1Δ and rif1Δ mutants. Each chromosomal element contains at least two Taz1 binding sites (GGTTA) and is labeled according to the number of its adjacent origin. Both large and small amplicons were tested in each orientation (A or B) and were inserted several times in some instances (indicated by x2 or x3). Notable repression is marked by a red asterisk. (B) Chromosomal locations of origins and elements tested.

Fig. 3.

Regional control of replication by STAR elements and requirement for Rif1- and Mcm4 replication regulatory domains for boundary positioning. (A) Replication was assayed by BrdU incorporation in the presence of HU in cell-cycle-synchronized cells with the indicated boundaries. (B) Proposed regulation of MCM complex by CDK/DDK and Rif1-associated phosphatases (PP1), depicting amino-terminal serine-rich domain (NSD) of Mcm4 that represses origin firing unless phosphorylated. (C) Rif1 mutations that reduce its interaction with PP1 (rif1Δ; rif1-7A-PP1; rif1-PP1; and rif1-12D) reduce boundary formation by STAR2 unlike mutations that prevent phosphorylation by CDK/DDK (rif1-7A). (D and E) Derepression of (EcoRV)::ade6⁺ in mcm4ΔNSD mutant assayed as in Fig. 2 C–E.

Fig. S5.

Regional control of replication by STAR elements. (A) Replication efficiency at test chromosomal locations. The BrdU-containing immunoprecipitated DNA used to establish the replication profiles shown in Fig. 3 was also used to monitor replication efficiency at three test locations: an early origin, a late origin, and a nonorigin, showing similar immunoprecipitation (IP) efficiencies for the four strains. (B) Replication was assayed by BrdU incorporation at 25 min and 30 min following release into M phase in the absence of HU. Note that in this experiment, the IP efficiency is decreased at later times (e.g., at matK, 0.8% at 25 min to 0.6% at 30 min) due to broader BrdU-incorporation genome-wide, which causes the antibody to become limiting. In both A and B, the IP efficiency was measured as immunoprecipitated DNA relative to input chromatin (percent).

Fig. S6.

Mutational analysis of Rif1 mutants. As shown in Fig. 4 for STAR2, Rif1 mutations that reduce its interaction with PP1 (rif1Δ; rif1-7A-PP1; rif1-PP1; and rif1-12D) reduce boundary formation by STAR3 and by the 90-bp STAR2-S element, unlike mutations that prevent Rif1 phosphorylation by CDK/DDK (rif1-7A).

Fig. 4.

Restoration of heterochromatic domain integrity by tethered Taz1, Rif1, or protein phosphatases. (A) Taz1, Rif1, Sds21, and Dis2 expressed as fusions with the Gal4-binding domain (GBD) mediate (EcoRV)::ade6⁺ silencing when targeted to five Gal4-binding sites (5xGBS). Repression by GBD-Rif1, GBD-Sds21, and GBD-Dis2 occurs in the absence of Taz1. (B) Repression requires tethering to 5xGBS. (C) RT-QPCR analysis of transcripts as in Fig. 2E.

Fig. S7.

(EcoRV)::ade6⁺ is intact in tethering experiments. HindIII digests produced bands at 2759 and 3414 for ade6-DN/N and 1879 and 5204 for (EcoRV)::ade6+ when probed with the SpeI-PvuII ade6-containing fragment for all strains examined (-, PD7; GBD, PD80; GBD-Sds21, PD15; GBD-Dis2, PD20; GBD-Rif1, PD14; and GBD-Taz1, PD155).