Position Effect Variegation at the Mating-Type Locus of Fission Yeast: A cis-Acting Element Inhibits Covariegated Expression of Genes in the Silent and Expressed Domains

Corresponding author: Amikam Cohen, Department of Molecular Biology, The Hebrew University-Hadassah Medical School, Jerusalem 91010, Israel., amikamc{at}cc.huji.ac.il (E-mail)

Communicating editor: G. R. SMITH

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Schizosaccharomyces pombe switches its mating type by transposing a copy of unexpressed genes from the respective mat2 or mat3 cassettes to mat1. The donor cassettes are located in a silent domain that is separated from the expressed mat1 cassette by the L region. We monitored the expression of ade6 from sites in the L region and examined the relationship between the expression state at these sites and at sites within the silent domain. Results indicate that: (1) the silent domain extends into the L region, but repression is gradually alleviated with increasing distance from mat2, and overexpression of swi6 enhances PEV in the L region; (2) a transcriptionally active chromatin state, associated with reporter gene expression in the L region, spreads toward the silent domain; (3) a cis-acting element, located at the junction between the L region and mat2-P, ensures repression in the silent domain, regardless of the expression state in the L region; and (4) repression in mat1-P cells is less stringently controlled than in mat1-M cells. We discuss the functional organization of the mat region and genetic elements that ensure separation between repressed and derepressed domains.

IF a euchromatic gene is juxtaposed with a heterochromatic domain, it may be subjected to variable, but clonally heritable, repression. This phenomenon, named position effect variegation (PEV), is functionally related to other phenomena of transcriptional silencing that are inherited by epigenetic mechanisms. Some examples are X chromosome inactivation, genomic imprinting, and silencing of the mating type donor loci in yeast (reviewed in MCBURNEY 1993 ; RASTAN 1994 ; BARLOW 1995 ; EISSENBERG et al. 1995 ; HENDRICH and WILLARD 1995 ; HENIKOFF 1995 ; LOO and RINE 1995 ; ALLSHIRE 1996 ; HOLMES et al. 1996 ; RUSSO et al. 1996 ). PEV was first observed in Drosophila, where X-ray-induced translocation of the white gene to a site adjacent to centromeric heterochromatin resulted in red-white variegation of eye pigmentation (MULLER 1930 ). Since then, it has been investigated in several experimental systems, including trans-inactivation, telomere position effect, and translocation of heterochromatin genes to euchromatic regions (reviewed in KARPEN 1994 ; WEILER and WAKIMOTO 1995 ; HENIKOFF 1997 ). In the budding yeast Saccharomyces cerevisiae and the fission yeast Schizosaccharomyces pombe, position effect repression of the exogenous reporter gene was observed at the donor mating type loci and in subtelomeric regions (KLAR et al. 1981 ; NASMYTH 1982 ; PILLUS and RINE 1989 ; GOTTSCHLING et al. 1990 ; APARICIO et al. 1991 ; LORENTZ et al. 1992 ; SUSSEL et al. 1993 ). In S. pombe, it was also observed at the centromeres (ALLSHIRE et al. 1994 ). A relationship between PEV in Drosophila and S. pombe is suggested by the homology shared by Swi6p and Clr4p, which confer silencing in S. pombe, and the heterochromatin-specific chromosomal protein (HP-1) that is associated with suppression of PEV in Drosophila melanogaster (LORENTZ et al. 1994 ; IVANOVA et al. 1998 ). The homology is in the chromodomain that was also identified in the Polycomb protein of D. melanogaster, the mouse M31, M32, and M33 proteins, and in the human HSM1 protein (PARO and HOGNESS 1991 ; PLATERO et al. 1995 ).

S. pombe switches its mating type by transposing a copy of the unexpressed genes from the respective silent mat2 and mat3 cassettes to the closely linked mat1 cassette. mat2 contains the Pc and Pm genes, and mat3 contains the Mc and Mm genes. The P and M genes confer the respective P or M mating types on the cell when expressed from mat1. The silent cassettes possess all the regulatory sequences present at the transcriptionally active mat1, but both are located in a region of ~17 kb that is repressed by heritable transcription silencing. The two silent cassettes are separated from each other by a region named K, and the mat2-K-mat3 domain is separated from the transcriptionally active mat1 by a region of ~17 kb named L (Figure 1; reviewed in KLAR 1992 ; ALLSHIRE 1996 ). Mutations that disrupt silencing at the mat2-K-mat3 region also affect expression of reporter genes at the centromeres (ALLSHIRE et al. 1994 ). Some of these mutations affect the proper functioning of the centromeres (ALLSHIRE et al. 1994 , ALLSHIRE et al. 1995 ) and the directionality of mating-type switch (THON and KLAR 1992 ; GREWAL and KLAR 1996 , GREWAL and KLAR 1997 ; THON and FRIIS 1997 ). These results suggest that silencing at the mat2-K-mat3 region shares a common mechanism with the assembly of the heterochromatin-like structure at the centromeres. A common mechanism is also suggested by the 96% identity shared by a 4.3-kb sequence, located within the K region, and the dgII and dhIIa repeats of centromere 2 (GREWAL and KLAR 1997 ) and by the observation that Swi6p localizes to the centromeres and the mating-type loci in a clr4- and rik1-dependent manner (EKWALL et al. 1995 , EKWALL et al. 1996 ).

View larger version (22K): In this window In a new window Download PPT slide   Figure 1. PEV of ade6+ expression from sites in the L region. Cells were plated on YE medium and incubated at 33°, as described in MATERIALS AND METHODS. Representative colonies of strains with ade6+ insertions at the indicated sites (a, AP135; b, AP136; c, AP144; d, AP330) of isogenic clr1 (a', AP133; c', AP141) or swi6 (b', AP139; d', AP340) derivatives and of control ade6+ (SP412) and ade6-210 (AP137) strains are shown. The percentages of Ade+ (white) colonies of the respective (a–c) strains are indicated. In each case, the indicated range of the proportion of white colonies is based on data of at least four independent experiments. Bar, 2 mm.

Heritable transcription silencing also depends on cis-acting elements called silencers (reviewed in DILLIN and RINE 1995 ; LOO and RINE 1995 ; KAMAKAKA 1997 ). Results of a plasmid-based study in S. pombe identified four putative silencers adjacent to the mat2 locus (EKWALL et al. 1991 ). Deletion of a BssHII-BglII fragment, containing two of these elements, from the centromere-proximal side of mat2 derepressed the mat2-P genes on a plasmid, but it had only a marginal effect in a chromosomal context. However, this deletion greatly enhanced chromosomal mat2-P expression in strains mutated in any of the swi6 and clr1-clr4 genes (THON et al. 1994 ). This observation suggests redundancy of two overlapping silencing mechanisms: one depends on swi6 and clr1-clr4 genes, and the other on one or more cis-acting elements at the centromere-proximal side of mat2. A group of trans-acting genes, defined by mutations that enhance mat2-P and mat3-M expression in swi6 and clrI mutants, has been described recently. This phenotype suggests that products of these genes participate in the swi6- and clr1-clr4-independent silencing pathway (THON and FRIIS 1997 ; GREWAL et al. 1998 ).

This study focuses on the functional organization of the mating-type donor region of S. pombe and the genetic elements that ensure repression at the silent domain. Our results indicate that (1) silencing extends into the L region, but is gradually alleviated as the distance from mat2 increases; (2) a transcriptionally active chromatin state can spread from the L region toward the silent domain, but the previously identified cis-acting REII element governs repression within the silent domain, regardless of the expression state at the L region; and (3) repression at the mating-type region is more stringently controlled in mat1-M than in mat1-P cells.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strain construction and plasmids:
All strains used in this study and their genotypes are listed in Table 1. Molecular manipulations of cloned HindIII or HindIII-BglII fragments, carrying mat2 and the parts of the flanking L and K regions (BEACH and KLAR 1984 ), were performed on derivatives of Bluescript (ALTING-MEESE and SHORT 1989 ) or pIRTH1 plasmids. pIRTH1 was derived from pIRT2 (HINDLEY et al. 1987 ) by mutational inactivation of the HindIII site between the pUC118 fragment and LEU2. To insert reporter genes into the designated sites in the cloned mat2 fragment, chimera plasmids were cleaved by the appropriate restriction endonucleases, and either a 1.8-kb HindIII fragment carrying the ura4⁺ gene (BACH 1987 ) or a 3.5-kb PvuII-SmaI fragment carrying the ade6⁺ gene (SZANKASI et al. 1988 ) was ligated to the linearized plasmids. Where necessary, single-stranded ends were converted to blunt ends by DNA polymerase I (Klenow fragment) or T4 DNA polymerase-mediated synthesis or digestion. A mat2-K-mat3 fragment with an ade6⁺ replacement of the 7.5-kb K deletion was constructed essentially as described by GREWAL and KLAR 1996 , except that single-stranded ends were converted to blunt ends before insertion of the ade6⁺ fragment. To insert ade6⁺ into the PvuII site in the L region, a 10.4-kb HindIII genomic fragment containing the mat1-M cassette (BEACH 1983 ) was cloned in pBR322, and the fragment carrying the ade6⁺ gene was inserted by ligation into the PvuII site (pNA216). To construct AP224, pGB2 (CHURCHWARD et al. 1984 ) was linearized by SalI digestion and inserted into the SacI site of a cloned mat2-P HindIII fragment with a ura4⁺ insertion at the XbaI site. To construct AP357, a 2.7-kb XhoI-PvuII fragment of Vibrio fischeri luxA luxB genes was excised from pRF918 (FRIEDMAN-OHANA et al. 1998 ) and substituted for the 230-nt BssHII-EcoRI L region fragment. Targeted integration into the mat region was accomplished by transformation with purified restriction fragments of the appropriate plasmids. To overcome the recombination block at the mat2-mat3 interval (EGEL 1984 ), targeted integration to this interval was performed in a clr1 mutant (PG377). Integration of molecular constructs at the desired sites was confirmed by Southern hybridization analysis (SOUTHERN 1975 ). In all experiments reported here, ade6 genes were inserted at the L region in the same orientation, with the 5' end of the gene at the centromere-proximal side of the inserted fragment. Standard genetic crosses and transformation procedures (MORENO et al. 1991 ) were used in strain construction. For overproduction of Swi6p, cells were transformed with pAL2 (LORENTZ et al. 1994 ) and selected on leucine-depleted minimal medium (AA-Leu).

Culture conditions:
Strains were grown on rich medium (YEA), adenine-deficient rich medium (YE), sporulation medium (PM-N), uracil-depleted sporulation medium (PM-N-Ura), and minimal medium supplemented with the appropriate growth requirements (AA, MORENO et al. 1991 ). AA-Ura is a uracil-depleted minimal medium. Cells were selected for the Ura^- phenotype on FOA (5-fluoroorotic acid) medium, because in this medium, ura4 expression leads to the synthesis of a toxic product from FOA (BOEKE et al. 1984 ). FOA medium was minimal medium supplemented with 1 mg/ml FOA and 0.1 mg/ml uracil. Liquid cultures were grown at 33°, and plates were incubated at the indicated temperature. For scoring Ade phenotypes on YE plates, standard incubation periods were 3 days at 33°, 4 days at 30°, and 5 days at 25°. For full development of colony color, YE plates, incubated at 33°, were incubated at 25° for an additional 24 hr before photography. To monitor ade6 expression by cells harboring pAL2 derivatives, cells were plated on AA-Leu medium with a low concentration of adenine (30 mg/liter).

Iodine staining:
Haploid meiosis phenotype in heterothalic strains was examined by staining colonies on sporulation medium with iodine vapors, because spores, but not vegetative cells, contain a starch component. Plates were incubated for 4 days at 30° before staining (BRESCH et al. 1968 ).

Photography:
Colonies were photographed with overhead illumination using Ektachrome slide film (Eastman Kodak, Rochester, NY). Slides were computer scanned using the Adobe Photoshop program.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Repression spreads with decreasing strength from mat2 into the L region:
The repressed mat2-K-mat3 domain is separated from the transcriptionally active mat1 by the L region (Figure 1). To ascertain whether repression is extended beyond mat2, we constructed strains with an ade6⁺ reporter gene inserted at sites along the L region, and we compared ade6⁺ expression from these sites to that from a site within mat2. Cells of the respective strains were plated on low-adenine medium (YE, MORENO et al. 1991 ), and colony colors were examined. The occurrence of white or red colonies on this medium implies Ade⁺ or Ade^- phenotypes, respectively. Sectored colonies indicate that in some cells, ade6⁺ switched from one state of expression to another, and pink colonies suggest intermediate levels of repression (ALLSHIRE et al. 1994 ).

An ade6⁺ reporter gene inserted at the L region was subjected to PEV. When inserted within mat2 (BamHI), ade6 was repressed in the majority of the colonies (Figure 1). However, a small proportion of the colonies (~0.5% at 33°) expressed ade6⁺, as was evident by their white or sectored appearance. The derepressed phenotype was unstable. About 60% of the cells from the Ade⁺ colonies yielded red or sectored colonies upon replating. Repression extended into the L region, but it was gradually alleviated as the distance from mat2 increased. This was apparent from the gradual reduction in the proportion of Ade⁺ colonies, the hue of the red color (Figure 1), and the stability of the repressed state (Figure 2). About half of the colonies of a strain with an ade6⁺ insertion at 1.5 kb from mat2 (BglII) were either pink or sectored. As the distance of the ade6 insertion site from mat2 increased to 2.5 kb (SacI), most colonies were white (85–95%) and the others were light pink. The "pink" phenotype was associated with decreased stability of the repressed state, as was apparent from the high proportion of colonies with multiple white and pink sectors. At 14 kb from mat2 (PvuII), ade6⁺ was derepressed in all colonies (Figure 1).

View larger version (25K): In this window In a new window Download PPT slide   Figure 2. The effect of distance from mat2 and temperature on the stability of the ade6+ repressed state at sites in the L region. Cell suspensions of Ade- (red) colonies picked from YE plates at 25° were diluted and replated on the same medium. Colony colors were scored after incubation at the indicated temperature (see MATERIALS AND METHODS). Five classes of colony color on YE plates were scored: white, red (see Figure 1D for an example), pink (example in Figure 1C), light pink (lighter than the colony in Figure 1C), and sectored colonies. The percentage of colonies in each class is presented. More than 500 colonies of five independent colonies from each strain were examined. Error bars represent standard deviation. Map locations of the ade6 insertions in the respective strains are indicated.

We tested the effect of mutations in the swi6 and clr1-4 genes on repression of ade6⁺ within mat2 and at the L region. Consistent with earlier reports on the effect of these mutations on silencing at the mat region and at other locations (EKWALL and RUUSALA 1994 ; THON et al. 1994 ; THON and FRIIS 1997 ; ALLSHIRE et al. 1995 ; GREWAL and KLAR 1997 ), repression of ade6⁺ was alleviated in the corresponding mutants (see Figure 1 for examples). Some of these mutations affect genes that are involved in the assembly of a heterochromatin-like structure (GREWAL et al. 1998 ; EKWALL and RUUSALA 1994 ; LORENTZ et al. 1994 ; THON et al. 1994 ; ALLSHIRE et al. 1995 ; IVANOVA et al. 1998 ). Therefore, these observations suggest that the heterochromatin structure extends from the repressed mat2-K-mat3 domain into the L region, and this is accompanied by a variegated phenotype.

We compared the stability of the ade6 repressed state in strains with ade6⁺ insertions at sites along the L region, within mat2, or as a substitution for a deleted 7.5-kb fragment of the K region (K::ade6). Cells of red (Ade^-) colonies on YE medium were replated on the same medium, and colony colors were examined (Figure 2). Consistent with earlier reports (GREWAL and KLAR 1996 ; THON and FRIIS 1997 ), ade6 expression by the K::ade6⁺ strain variegated, and the alternative expression states were mitotically stable. In contrast, the stability of the repressed state of ade6, inserted at sites in the L region, decreased as the distance of the insertion site from mat2 increased. However, the stability of the derepressed state increased with the distance from mat2 (data not shown).

PEV in Drosophila and yeast is frequently sensitive to temperature, with repression being suppressed at higher temperatures (GOWEN and GAY 1933 ; SPOFFORD 1976 ; ALLSHIRE et al. 1995 ). PEV of reporter genes at the L region was strongly affected by temperature (Figure 2). For example, the proportion of cells from an Ade^- colony of an L(SacI)::ade6⁺ strain that converted to Ade⁺ was 38% at 33°, but only 4.2% at 25°. Unlike PEV at the L region, variegated expression of ade6⁺ by the K::ade6 strain was insensitive to temperature, and the alternative Ade epitypes of this strain were steadily maintained at 25° and 33° (Figure 2 and data not shown).

Repression enhancement at the L region by a K region deletion or swi6⁺ overexpression:
The extent of the silent domain in the L region may be limited by the availability of components for the spreading of the heterochromatin-like structure. Predictions of this proposition are that deletion of an internal fragment of the silent domain or overproduction of a limiting heterochromatin component would enhance the PEV of reporter gene expression from the periphery of the region. To test the first prediction, we constructed a strain with an ade6⁺ insertion at the SacI site in the L region and a 1.8-kb ura4⁺ substitution for a 7.5-kb fragment of the K region. Deletion of this fragment induced variegated expression of markers in the mat2-mat3 interval, but each expression state was relatively mitotically stable and changed rarely (GREWAL and KLAR 1996 ; THON and FRIIS 1997 ). We selected for the Ura⁺ and Ura^- phenotypes on selective media and then monitored ade6⁺ expression from the SacI site in the K::ura4⁺ mutant and an isogenic strain with the K region intact (Figure 3A). After selection for ura4⁺ expression from the K region, ade6⁺ in the L region was derepressed in cells of both strains. However, when cultures were selected for ura4⁺ repression, the proportion of colonies expressing ade6⁺ from the L region was exceedingly higher in the control strain than in the strain with the K::ura4⁺ mutation. Whereas colonies of strain AP165 with the K region intact were either white or light pink, most colonies of the deletion mutant AP153 were red. We also determined the effect of Swi6p overproduction from a multicopy plasmid on PEV of ade6 expression from the SacI site at the L region (Figure 3B). More than 90% of the colonies from a control strain harboring the vector plasmid pIRTH1 exhibited an Ade⁺ (white) phenotype, and the rest were light pink. Overexpression of swi6⁺ in an isogenic strain enhanced the PEV of ade6 expression from the SacI site. Less than 10% of the colonies of a strain harboring a swi6-expressing plasmid (pAL2) were white or sectored, and the rest were red. These observations suggest that the availability of heterochromatin components controls repression in the L region.

View larger version (110K): In this window In a new window Download PPT slide   Figure 3. Enhancement of PEV at the L region by (A) a deletion at the K region or (B) overexpression of swi6+. (A) Cells of a mat1-M L(SacI)::ade6+ K::ura4+ strain (AP165) and of an isogenic strain with the K region intact (AP153) were selected for ura4+ expression or repression on AA-Ura and FOA media, respectively, and plated on YE medium to examine expression of ade6. Relevant genotypes and previous growth conditions are indicated. (B) Cells of a mat1-M L(SacI)::ade6+ strain (AP136) harboring pAL2 (ade6+) or a plasmid vector pIRTH1 were plated on low-adenine AA-Leu medium. Colonies were photographed after 4 days incubation at 30°. Bar, 2 mm.

Deletion of a repression element near mat2 suppresses PEV in the L region:
The cis-acting element(s) near mat2 are involved in a swi6-independent mechanism that represses mat2 (THON et al. 1994 ). Deletion of one of these elements, named Repressive Element II (REII), enhances mat2P expression in swi6 and clr1-4 mutants, but not in wild-type cells (I. GOLDSHMIDT, M. MARASH and A. COHEN, unpublished results). To examine whether this element plays a role in repression of exogenous genes at the L region, we inserted ade6⁺ into the junction between mat2-P and the L region (BssHII), or at a distance of 230 nucleotides (nt) from mat2-P (EcoRI). We then determined the dependence of ade6⁺ repression on a cis-acting element located between the two insertion sites. The severe repression affecting ade6⁺ within mat2 (BamHI, Figure 1) was extended to the BssHII site at the junction between mat2-P and the L region (Figure 4). Less than 0.5% of the L (BssHII)::ade6 cells formed white colonies on YE medium, and the rare Ade⁺ epitype was unstable (data not shown). Repression was partially alleviated when the ade6⁺ insertion site was moved to the centromere-proximal side of REII (EcoRI). About 10% of the L(EcoRI)::ade6⁺ colonies were white, and the rest were pink or sectored. Deletion of REII markedly enhanced ade6⁺ expression. More than 60% of the colonies of a strain with an ade6⁺ replacement of the BssHII-EcoRI fragment were white, and the rest were light pink or sectored. The suppressive effect of this deletion on the PEV of ade6⁺ expression from the flanking sites demonstrates that chromosomal REII is functional in wild-type cells. It also indicates that this element does not act specifically on mat2-P genes, but can also repress exogenous reporter genes. Because ade6 was less stringently repressed in the deletion mutant than in strains with ade6 insertions at one of the two flanking sites, we conclude that the effect spreads to both sides of the element. However, the marked difference between repression efficiency at the EcoRI and BssHII sites (Figure 4) suggests that this element acts preferentially to its centromere-distal side.

View larger version (41K): In this window In a new window Download PPT slide   Figure 4. REII participates in repression of adjacent ade6 genes in the L region. Cells with ade6+ insertions at the BssHII (AP312) or EcoRI (AP326) site or with an ade6+ replacement of the BssHII-EcoRI deletion (AP342) were plated from YEA liquid cultures (107 cells/ml) onto YE plates. The percentages of Ade+ (white) colonies on the respective plates are indicated. All incubations were at 30°.

A cis-acting element in the L region excludes mat2-P from a derepressed state affecting its flanking regions:
Selection for ura4⁺ expression from the K region or from the EcoRV site at the centromere-distal side of mat3-M led to derepression of ade6 at sites along the L region (Figure 3A and data not shown). This indicates that a derepressed state, selected for within the silent domain, is extended into the L region. Because mat2-P is located between the K and L regions, we explored the possibility that it was also derepressed in cells selected for ura4⁺ expression from the K region. Expression of mat2-P was monitored in mat1-Msmt-0 derivatives by plating cells on a uracil-depleted sporulation medium and assaying for haploid meiosis. In these strains, M genes are expressed from mat1. Therefore, derepression of mat2-P would lead to simultaneous expression of P and M genes in the same haploid cell and subsequently to haploid meiosis (KELLY et al. 1988 ). Haploid meiosis in colonies on sporulation medium was monitored by iodine staining (BRESCH et al. 1968 ) and microscopic screening for azygotic spores.

mat2-P was excluded from the derepressed state affecting its flanking regions in a mat1-Msmt-0 K(XbaI)::ura4 strain and in isogenic derivatives with ade6⁺ insertions at 1.5 kb (BglII) or 2.5 kb (SacI) from mat2 (Table 2). Colonies of these strains resisted staining with iodine on a uracil-depleted sporulation medium, and haploid meiosis was not detectable by microscopic examination. Interestingly, when ade6⁺ was inserted at the junction between mat2 and the L region (BssHII), selection for ura4⁺ expression from the K region caused derepression of mat2-P. All colonies of this strain stained black by iodine on a uracil-depleted sporulation medium, and microscopic examination revealed that >80% of the cells in these colonies underwent haploid meiosis. Low levels of mat2-P expression were observed in a strain with an ade6⁺ insertion at the EcoRI site, 230 nt from mat2-P. About 1% of the colonies of this strain (AP326) stained lightly with iodine, and the proportion of cells undergoing haploid meiosis in the stained colonies was 5–10%. These results suggest that the ade6⁺ insertion at the BssHII site (and to a lesser degree at the EcoRI site) interfered with a mechanism that excludes mat2-P from a derepressed state affecting its flanking regions. These observations are consistent with earlier results showing alleviation of mat2-P repression in a plasmid context after insertion of an exogenous fragment at the junction between mat2 and the L region (EKWALL et al. 1992 ).

A low frequency of iodine-stained colonies was observed when cells with an ade6⁺ insertion at the BssHII site were plated on a uracil-supplemented sporulation medium (Table 2). Microscopic examination revealed that ~30% of the cells in the iodine-stained colonies underwent haploid meiosis. Replica-plating experiments, such as the ones depicted in Figure 5, revealed that cells of the rare Ade⁺ colonies also expressed mat2-P (not shown). This suggests covariegated expression of mat2-P and the adjacent ade6 in this strain.

View larger version (95K): In this window In a new window Download PPT slide   Figure 5. A cis-acting element on the BssHII-BglII fragment inhibits covariegated expression of mat2-P and ura4+ in the K region with ade6+ in the L region. Cells from colonies of L(SacI)::ade6+ strains on a nonselective (YEA) medium were plated on the same medium. Colonies were replica plated from the YEA plates onto YE, sporulation (PM-N), and, where applicable, AA-ura media. All incubations were at 30°. Expression of ade6+, mat2-P, and ura4+ was monitored as described in MATERIALS AND METHODS. (A) Uniform covariegation of ade6+ and mat2-P expression was observed in colonies of the mat1-Msmt0 L(SacI)::ade6+ (BssHII-BglII) strain (AP151). All Ade+ colonies (closed arrowheads) expressed mat2-P, and colonies that exhibited intermediate levels of ade6 repression (open arrowheads) were partially stained with iodine on a sporulation medium. (B) Regardless of ade6 expression state, mat2-P remained repressed in an isogenic strain, but with the wild-type allele of the BssHII-BglII deletion (AP136). (C) Covariegated expression of ade6+, mat2-P, and ura4+ in a mat1-Msmt0 L(SacI)::ade6+ (BssHII-BglII) K(XbaI)::ura4+ strain (AP174). (D) Covariegation was not observed in an isogenic strain with the wild-type allele of the BssHII-BglII deletion (AP165). A rare Ura+ colony of this strain represents ~0.2% of the colonies. (E) mat3-M remained repressed in a mat1-P17 L(SacI)::ade6+ (BssHII-BglII) K(XbaI)::ura4+ strain (AP186) regardless of the expression state of ade6+ in the L region and ura4+ in the K region. Closed and open arrowheads in C and D indicate Ade+ and Ade- colonies, respectively. Bar, 1 mm.

On the basis of these observations and the evidence for the existence of a cis-acting repression element near the BssHII site (EKWALL et al. 1991 ; THON et al. 1994 ), we hypothesized that this element inhibits covariegated expression of mat2-P with genes at its flanking regions and that the ade6⁺ insertion at the BssHII site interfered with its activity. To examine the role of cis-acting elements near mat2 in the exclusion of mat2-P from a derepressed state affecting the regions, we carried out a genetic analysis of this region. Cells of a BssHII-BglII deletion mutant (SP1151; see Figure 1) and of an isogenic strain, with the wild-type allele of this deletion (SP1124), were tested for mat2-P expression on a uracil-depleted sporulation medium. Haploid meiosis was observed in all Ura⁺ colonies of the deletion mutant (Table 2), but not in Ura⁺ colonies of the control strain or in colonies of the deletion mutant that were selected for ura4⁺ repression on FOA plates. Unexpectedly, a few colonies of the deletion mutant expressed mat2-P on a uracil-supplemented sporulation medium. These colonies stained positive with iodine, and cells in these colonies underwent haploid meiosis. This phenotype escaped detection in an earlier study of this strain (THON et al. 1994 ). These observations suggest that a cis-acting element(s) on the BssHII-BglII fragment participates in a mechanism that excludes mat2-P from a derepressed state affecting its flanking regions.

An ade6⁺ gene in the L region acts in synergy with deletion of an L region fragment to alleviate repression in the silent domain:
The low proportion of SP1151 colonies expressing mat2-P genes under nonselective conditions was comparable to the proportion of cells producing colonies on a uracil-depleted medium. This correlation suggested that deletion of the BssHII-BglII fragment allowed covariegated expression of mat2-P with the adjacent ura4⁺ at the K region. We examined whether this deletion would allow covariegated expression of mat2-P with ade6⁺ at the SacI site in the L region. Because ade6⁺ expression from this site is only partially repressed (Figure 1 and Figure 5), covariegated expression of the two genes should markedly increase the proportion of colonies expressing mat2-P. Colonies of mat1-Msmt-0 L(SacI)::ade6 strains were replica plated from a nonselective (YEA) medium onto YE and sporulation media, and the effect of the BssHII-BglII deletion on mat2-P expression was examined. A BssHII-BglII deletion or an ade6⁺ insertion at the SacI site had a minor suppressive effect on mat2-P repression (Table 2). However, a combination of the two L region mutations had a synergistic effect. About 90% of the double mutant's colonies stained dark with iodine on sporulation medium after incubation at 30° (Table 3), and the frequency of cells undergoing haploid meiosis in these colonies exceeded 80%. Furthermore, expression of ade6⁺ and mat2-P covariegated uniformly (Figure 5A), and colonies that showed partial derepression of ade6⁺ (pink colonies on a YE medium) also showed partial derepression of mat2-P (partial staining with iodine on sporulation medium) (~1000 colonies were examined).

Next, we examined the effect of a combination of ade6⁺ at the L region and deletion of the BssHII-BglII fragment on the expression state of ura4⁺ at the XbaI site in the K region. Either of the two L region mutations increased the frequency of Ura⁺ cells, but a combination of both had a synergistic effect (Figure 5 and Figure 6). Expression of ade6⁺, mat2-P, and ura4⁺ covariegated uniformly in the strain with the two L region mutations (Figure 5C), but not in an isogenic strain with the wild-type allele of the deletion (Figure 5D). These results indicate that covariegated expression of markers in the mat2-K region with ade6⁺ at the L region was inhibited by an element(s) on the BssHII-BglII fragment.

View larger version (65K): In this window In a new window Download PPT slide   Figure 6. Repression is more stringently controlled in mat1-M than in mat1-P cells. The effect of an ade6+ gene inserted at the L region (SacI) and deletion of the L region BssHII-BglII fragment on ura4+ expression from sites near mat2 (XbaI) or mat3 (EcoRV) was determined in mat1-M and mat1-P strains. After growth on YEA medium at 30°, 10-fold serial dilutions of cell suspensions of strains with the indicated L and K regions mutations were spotted on complete (AA) and selective media. In mat1-M cells, insertion of ade6 at the SacI site, deletion of the BssHII-BglII fragment, or a combination of the two L region mutations enhanced ura4+ expression from a site near mat2 (XbaI), but not from the centromere-distal side of mat3 (EcoRV). Conversely, in mat1-P cells, these mutations enhanced ura4+ expression from both insertion sites. Relevant restriction sites are indicated on the map. Control ura4+ and ura4-D18 strains of both mating types are presented.

The ade6⁺ insertion at the SacI site may have alleviated mat2-P repression in the BssHII-BglII deletion mutant by disrupting an unknown cis-acting element or interfering with its activity. Alternatively, an activity associated with ade6 expression may have affected the chromatin state at the insertion site. As a result, the transcriptionally active chromatin would spread toward the silent domain and disrupt mat2-P repression. To test the first possibility, we compared the effect of an ade6 insertion at the SacI site to that of a bacterial plasmid DNA, inserted at the same site. Unlike the ade6 insertion, the inserted 4-kb plasmid DNA did not enhance mat2-P or ura4⁺ expression from the K region (Table 3). This result and the covariegated expression of mat2-P, and an adjacent ura4⁺ gene, with ade6⁺ in the L region are consistent with the second possibility.

We note that the effect of the BssHII-BglII deletion on the expression state of ade6⁺ at the SacI site in the L region was smaller than that on the expression state of mat2-P or ura4⁺ in the K region. PEV of ade6 expression from the SacI site was not suppressed appreciably by this deletion. This observation, like the data presented in Figure 4, suggests that a repression element on the deleted fragment acts preferentially on its centromere-distal side.

We asked whether the derepressed state, induced by the combination of the ade6 insertion and the BssHII-BglII deletion, affects the expression state of mat3-M. To monitor ade6⁺, ura4⁺, and mat3-M expression in a mat1-P17 L(SacI)::ade6⁺ (BssHII-BglII) K(XbaI)::ura4⁺ strain, colonies were replica plated from a nonselective medium (YEA) onto YE, uracil-depleted, and sporulation media. Because P genes are expressed in this strain from mat1, expression of mat3-M would lead to haploid meiosis that can be monitored by iodine staining of colonies on sporulation medium. Expression of ura4⁺ from the K region and ade6⁺ from the L region covariegated in the BssHII-BglII deletion mutant. However, mat3-M remained repressed, regardless of the expression state of the exogenous markers at the L and K regions (Figure 5E). Ade⁺ Ura⁺ colonies did not stain with iodine, and haploid meiosis was not detected by microscopic examination. The cumulative effect of the ade6 insertion at the L region and deletion of the BssHII-BglII fragment on ura4⁺ expression from centromere-distal side of mat3 is described below (Figure 6).

REII inhibits covariegated expression of mat2-P with ade6 at the L region:
A study of mat2-P silencing in a plasmid context identified two cis-acting repression elements on the deleted BssHII-BglII fragment, REI and REII (EKWALL et al. 1991 ). We examined whether deletion of either one of these elements is sufficient to enhance expression of chromosomal mat2-P (Table 3). Deletion of a 1.3-kb EcoRI-EcoRI fragment containing REI had no detectable effect on the repressed state of mat2-P. Conversely, REII was required to maintain silencing of mat2-P when ade6⁺ was expressed in the adjacent L region. Deletion of the 230-nt BssHII-EcoRI fragment containing REII alleviated mat2-P repression in colonies expressing ade6 from the SacI site. The frequency of iodine-stained colonies in the REII deletion mutant was similar to that in the strain with the longer BssHII-BglII deletion (Table 3). However, unlike cells in colonies of the strain with the BssHII-BglII deletion, only 20–30% of cells in most colonies of the (BssHII-EcoRI) strain underwent haploid meiosis, and the rate of haploid meiosis exceeded 80% in only ~5% of the colonies.

We also asked whether an ade6⁺ gene, inserted elsewhere at the L region, would suppress mat2-P repression. The BssHII-EcoRI fragment containing REII was replaced by ade6⁺ or by a bacterial DNA fragment, and the expression state of mat2-P was examined (Table 3). Deletion of REII or its replacement by a 2.7-kb bacterial DNA fragment had only a minor effect on the repressed state of mat2-P, yet replacement of REII by ade6⁺ alleviated mat2-P repression. Close to 30% of the colonies stained dark with iodine, and the frequency of cells undergoing haploid meiosis in most colonies was ~80%. Altogether, these results suggest that an activity associated with ade6 expression is involved in derepression of markers in the silent domain. Because derepression was observed only in REII deletion mutants, we propose that REII counteracts the effect of this activity on the repressed state at the silent domain.

Repression is less stringently controlled in mat1-P than in mat1-M cells:
A ura4⁺ reporter gene inserted within the silent domain is less stringently repressed in mat1-P than in mat1-M cells (Figure 6; THON et al. 1999 ). A difference in repression stringency is also suggested by the difference in the hue of the red colony color between mat1-P and mat1-M derivatives, with an ade6⁺ gene at the L region (compare Figure 5C and Figure E). The less stringent repression in mat1-P cells may allow the derepressed state, associated with ade6⁺ expression from the L region, to extend farther into the silent domain than in mat1-M cells. To examine this possibility, we constructed mat1-M and mat1-P derivatives with ura4⁺ reporter genes near mat2 (XbaI) or on the centromere-distal side of mat3 (EcoRV). We then examined the effect of the mating type, an inserted ade6⁺ gene at the L region (SacI), and deletion of a fragment containing REII on ura4⁺ expression from the respective locations (Figure 6). In mat1-M cells, an ade6⁺ insertion at the L region (SacI) had only a minor (~10-fold) suppressive effect on ura4⁺ repression at a site near mat2 (XbaI) and no detectable effect on ura4⁺ repression at a site near mat3 (EcoRV). However, the same ade6⁺ insertion increased the frequency of Ura⁺ cells in cultures of mat1-P strains with ura4⁺ insertions at either one of the two locations by 50- to 100-fold. Furthermore, in mat1-M cells, a combination of ade6⁺ expression from the L region with deletion of a fragment containing REII markedly enhances ura4⁺ expression from the XbaI site near mat2, but not from the EcoRV site near mat3. Conversely, in mat1-P cells, this combination of L region mutations alleviated ura4⁺ repression at both insertion sites. The difference in stringency of repression between mat1-M and mat1-P cells is also apparent from the difference in plating efficiency on FOA medium between the corresponding derivatives with an ade6 insertion at the L region and deletion of the BssHII-BglII fragment. Altogether, these results indicate that in mat1-P cells, the repression alleviation effect, associated with ade6 expression from the L region, extended farther than in mat1-M cells. Because silencing at the mat2-mat3 region depends on activities that help assemble heterochromatin (EKWALL and RUUSALA 1994 ; LORENTZ et al. 1994 ; THON et al. 1994 ; ALLSHIRE et al. 1995 ; GREWAL et al. 1998 ; IVANOVA et al. 1998 ), these results suggest differences in chromatin structure between cells of the two mating types.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

PEV at the L region:
An ade6⁺ reporter gene at the mat2-proximal end of the L region is subjected to position effect repression that is gradually alleviated as the distance from mat2 increases (Figure 1, Figure 2, and Figure 4). Repression dependence on proteins that help assemble the heterochromatin structure (Figure 1) suggests that the variegated phenotype reflects alternative chromatin states at the insertion site. Thus, two chromatin states exist along the L region: a transcriptionally active state at the mat1-proximal end is evident from the presence of the essential let1 gene at this region (MICHAEL et al. 1994 ) and the expression of ade6 from the PvuII site (Figure 1). A heterochromatic state at the mat2-proximal end is indicated by the swi6- and clr1-clr4-dependent repression of a reporter gene at this region. Separation of the two chromatin states is essential for cell viability, control of meiosis, and mating-type switching.

Reporter genes, inserted at internal sites of the silent domain, are constitutively repressed. But sporadic and transient derepression affects a small proportion (<0.5%) of the cell population (Figure 6). Deletion of a 7.5-kb K-region fragment, containing a 4.3-kb centromeric sequence, stabilizes the derepressed state, thereby leading to a variegated phenotype (GREWAL and KLAR 1996 , GREWAL and KLAR 1997 ; THON and FRIIS 1997 ). The mode of the variegated ade6⁺ expression by L::ade6⁺ strains is distinct from that by the K::ade6⁺ strain. Consistent with earlier reports (GREWAL and KLAR 1996 ; THON and FRIIS 1997 ), the alternative states of ade6⁺ expression in the deletion mutant were mitotically stable. On the other hand, the stability of the repressed state in the L::ade6⁺ strains decreased gradually as the distance between ade6⁺ and mat2 increased and at elevated temperatures (Figure 2). Another distinction between the K::ade6⁺ and L::ade6 strains is in the extent of the derepressed state. In K strains, genes in the L region were derepressed in all cells selected for expression from internal sites (Figure 3A). Thus, the variegated phenotype is likely to indicate alternating chromatin states along the entire domain. On the other hand, in the L::ade6⁺ strains, the proportion of cell lineages expressing ade6 increased with the distance from mat2. Assuming that repression extends continuously into the L region, this observation may reflect clonal variation in the lengths of the heterochromatin extension into this region. Deletion of an internal fragment or overproduction of Swi6p enhanced PEV at the L region (Figure 3). These data suggest that the extent of the heterochromatic state into the L region may be limited by the availability of heterochromatin components. This conclusion is supported by the following observation: normally, an ade6⁺ gene at 9 kb from mat2 (HpaI) (Figure 6) is fully expressed, but when swi6⁺ is overexpressed, ade6 expression is subjected to PEV (N. AYOUB and A. COHEN, unpublished results). Direct evidence for spreading of Sir3 into S. cerevisiae telomere-distal chromatin after Sir3 overproduction has been presented (HECHT et al. 1995 ).

Evidence for competition between alternative states of expression:
In mat1-M cells, REII deletion or ade6⁺ insertion at the L region SacI site had only a marginal effect on silencing in the mat2-mat3 region. However, the two mutations acted in synergy to overcome repression of mat2-P and an adjacent ura4⁺ gene. A possible explanation for this result is that the inserted fragment physically disrupted an unknown cis-acting repression element or interfered with its activity. Because REII also contributes to repression at the silent domain, inactivation of both elements would have a synergistic effect. An interesting alternative explanation is that a transcriptionally active chromatin state, associated with ade6 expression, spread toward the constitutively repressed domain. Deletion of REII allowed for the extension of the active chromatin state across mat2-P, to overcome repression at internal sites of the silent domain. The following evidence argues for the latter possibility:

1. The suppressive effect of an ade6⁺ insertion in the L region on silencing at the mat2-K region was not specific to the SacI insertion site. A similar effect was observed when ade6 replaced REII (Table 3).

2. Unlike the ade6⁺ insertions, insertion of a bacterial plasmid DNA fragment into the SacI site of an L(BssHII-BglII) strain or replacement of REII by a bacterial DNA fragment had no detectable effect on mat2-P expression (Table 3).

3. Covariegated expression of genes at internal sites of the silent domain and ade6⁺ at the L region (Figure 5) implies a relationship between ade6⁺ expression and derepression of the affected genes.

We therefore postulate that transcription factor(s) interacting with elements on the inserted ade6⁺ fragment establish a transcriptionally active chromatin state at the insertion site. The active chromatin state propagates along the L region toward mat2. Deletion of REII, which normally overrides the effect of ade6 expression, alleviates repression at internal sites of the domain.

The derepressed state, induced by the combination of ade6⁺ expression from the L region and REII deletion, did not affect mat3-M (Figure 5E). This may result from a recently described cis-acting element located near mat3 (THON et al. 1999 ). This proposition is supported by the observation that deletion of the repression element near mat3 acts in synergy with the REII deletion to enhance mat3-M expression, and this depends on ade6 insertion at the L region (A. AYOUB, G. THON and A. COHEN, unpublished results). Disruption of silencing by an active chromatin state, induced by interaction of trans-activators with a reporter gene, has been observed before at the subtelomeric region of S. cerevisiae (APARICIO and GOTTSCHLING 1994 ). Here, we present evidence that a derepressed state, established at the periphery of the silent domain, may propagate into internal parts of the domain to overcome silencing and that cis-acting elements participate in a mechanism that counteracts derepression by the expanding active chromatin state.

The colony patterns in Figure 5 imply clonal stability of the alternative epitypes in the L(SacI)::ade6⁺ (Bss-HII-BglII) mutant. Assuming that these epitypes were established after competition between mechanisms that assemble the alternative chromatin states, their stability indicates that once a chromatin state is established, it is inherited in a semistable manner. To explain the clonal stability of the alternative epitypes, we hypothesize that competition between the alternative chromatin states may occur at any replication cycle, but the established state is preferentially inherited (KAMAKAKA 1997 ). This principle may provide a simple explanation for the clonal inheritance of the alternative expression states in PEV.

REII-mediated silencing mechanism:
Deletion of the chromosomal REII, or of a longer fragment containing this element, had only a marginal effect on repression of mat2-P or an adjacent ura4⁺ gene. Nevertheless, these deletions act in synergy with mutations in trans-acting silencing genes to suppress repression (THON et al. 1994 ; I. GOLDSHMIDT, M. MARASH and A. COHEN, unpublished results). Here, we show that REII contributes to the repressed state of mat2-P and adjacent reporter genes in cells with functional silencing genes and an ade6⁺ insertion in the L region (Table 2 and Table 3; Figure 5 and Figure 6). In an attempt to understand the mode of action of REII, we consider two possibilities. One is that REII participates in a general repression mechanism that is partially redundant with the mechanism that is inactivated by ade6 expression from the L region. For example, if ade6⁺ expression from the L region interferes with the inheritance of the repressed state and REII is required for its establishment, the two L region mutations would act in synergy to alleviate repression. As a result, in strains with the double mutations, expression of ade6⁺ and markers along the affected domain would covariegate. The other possibility is that REII acts as a boundary element, thereby preventing the spread of the transcriptionally active chromatin state into the mat2-mat3 region. The occurrence of boundary elements at the junction between silent and expressed domains has been demonstrated in Drosophila and mammals (reviewed in GERASIMOVA and CORCES 1996 ; GEYER 1997 ), but not in yeast. The results presented here do not discriminate between the two possibilities. However, the observation that the REII deletion also acts in synergy with swi6 or clr1-4 mutations to alleviate repression of mat2-P is more easily explained by the "redundant mechanisms" than by the "boundary element" hypothesis.

We note that deletion of REI had no detectable effect on the repressed state of chromosomal mat2-P or of an adjacent ura4⁺ reporter gene in the K region (Table 3; unpublished results). This observation indicates that the autonomous replication sequence (ARS) in REI (OLSSON et al. 1993 ) is not essential for repression at the mat2 region. However, it does not rule out a role for this element in the repression mechanism, because another ARS or ARS-like element may substitute for REI in the deletion mutant. An accessory role for REI in mat2-P silencing is consistent with the observation that in most colonies of the REII deletion mutant (AP347), the frequency of cells undergoing haploid meiosis is lower than that in colonies of an isogenic strain with the BssHII-BglII deletion. However, this difference may also be explained by the difference in the distance of ade6⁺ from mat2-P between the two strains.

Repression stringency is affected by the mating type:
Repression along the silent domain is less stringently controlled in mat1-P than in mat1-M cells. This allows for the repression alleviation activity, associated with ade6 expression from the L region, to extend farther along the mat2-mat3 interval. Likewise, the cumulative effect of ade6 expression from the L region and deletion of the BssHII-BglII fragment on repression of a reporter gene near mat3 is much stronger in mat1-P than in mat1-M cells (Figure 6). To explain these observations, we postulate that a transcriptionally active chromatin state propagating from the ade6⁺ insertion site competes with the heterochromatin-like structure at the mat2-mat3 region, and the outcome of this competition is affected by the stringency of the general repression mechanism. Thus, partial relaxation of the general repression control mechanism observed in mat1-P cells helps the active chromatin to overcome repression along the entire length of the silent domain and allows its propagation despite the activity of REII and the cis-acting element near mat3.

The reason repression is less stringently controlled in mat1-P than in mat1-M cells is not yet understood. One intriguing possibility is that the difference in repression stringency reflects a difference in chromatin organization at the mating-type region between the M and P cells. Differences in chromatin organization between the two mating types were proposed to explain the directionality of mating-type switching (THON and KLAR 1993 ). Thus, the chromatin organization in homothallic mat1-P cells would favor mat3-M as the donor in a gene conversion event leading to the formation of a mat1-M cell. A different chromatin organization in mat1-M cells would make mat2-P the preferred donor in the gene conversion reaction.

	FOOTNOTES

¹ These authors contributed equally to this work.

	ACKNOWLEDGMENTS

We are grateful to Amar Klar and Genevieve Thon for helpful discussions, S. pombe strains, and communication of results before publication. We thank Henning Schmidt and Gerry Smith for plasmids; Nissim Benvenisty, Howard Cedar, Genevieve Thon, the editor, and anonymous reviewers for useful comments; and Eyal Rand for help in preparing figures. This work was supported by a grant from the U.S.-Israel Binational Science Foundation (93-291). N.A. was supported by a fellowship to minority students from the Israeli Ministry of Science.

Manuscript received September 8, 1998; Accepted for publication February 22, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALLSHIRE, R. C., 1996 Transcriptional silencing in the fission yeast: a manifestation of higher order chromosome structure and functions, pp. 443–466 in Epigenetic Mechanisms of Gene Regulation, edited by V. E. A. RUSSO, R. A. MARTIENSSEN and A. D. RIGGS, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

ALLSHIRE, R. C., J.-P. JAVERZAT, N. J. REDHEAD, and G. CRANSTON, 1994  Position effect variegation at fission yeast centromeres. Cell 76:157-169[Medline].

ALLSHIRE, R. C., E. R. NIMMO, E. EKWALL, J. P. JAVERZAT, and G. CRANSTON, 1995  Mutations derepressing silent centromeric domains in fission yeast disrupt chromosome segregation. Genes Dev. 9:218-233[Abstract/Free Full Text].

ALTING-MEESE, M. A. and J. M. SHORT, 1989  pBluescript II: gene mapping vectors. Nucleic Acids Res. 17:9494[Free Full Text].

APARICIO, O. M. and D. E. GOTTSCHLING, 1994  Overcoming telomeric silencing: a trans-activator competes to establish gene expression in a cell-cycle-dependent way. Genes Dev. 8:1133-1146[Abstract/Free Full Text].

APARICIO, O. M., B. L. BILLINGTON, and D. E. GOTTSCHLING, 1991  Modifiers of position effect are shared between telomeric and silent mating-type loci in S. cerevisiae.. Cell 66:1279-1287[Medline].

BACH, M.-L., 1987  Cloning and expression of the OMP decarboxylase gene URA4 from Schizosaccharomyces pombe.. Curr. Genet. 12:527-534[Medline].

BARLOW, D. P., 1995  Gametic imprinting in mammals. Science 270:1610-1613[Abstract/Free Full Text].

BEACH, D. H., 1983  Cell-type switching by DNA transposition in fission yeast. Nature 305:682-688.

BEACH, D. H. and A. J. S. KLAR, 1984  Rearrangements of the transposable mating-type cassettes of fission yeast. EMBO J. 3:603-610[Medline].

BOEKE, J., F. LACROUTE, and G. R. FINK, 1984  A positive selection for mutants lacking orotidine-5'-phosphate decarboxylase activity in yeast, 5-fluoro-orotic acid resistance. Mol. Gen. Genet. 197:345-346[Medline].

BRESCH, C., G. MULLER, and R. EGEL, 1968  Genes involved in meiosis and sporulation of a yeast. Mol. Gen. Genet. 102:301-306[Medline].

CHURCHWARD, G., D. BELIN, and Y. NAGAMINE, 1984  A pSC101-derived plasmid which shows no sequence homology to other commonly used cloning vectors. Gene 31:165-171[Medline].

DILLIN, A. and J. RINE, 1995  On the origin of a silencer. Trends Biochem. Sci. 20:231-235[Medline].

EGEL, R., 1984  Two tightly linked silent cassettes in the mating type of Schizosaccharomyces pombe.. Curr. Genet. 8:199-203.

EISSENBERG, J., S. C. R. ELGIN and R. PARO, 1995 Epigenetic regulation in Drosophila: A Conspiracy of Silence, pp. 141–171 in Chromatin Structure and Gene Expression, edited by S. C. R. ELGIN. IRL Press, Oxford.

EKWALL, K. and T. RUUSALA, 1994  Mutations in rik1, clr2, clr3 and clr4 genes asymmetrically derepress the silent mating-type loci in fission yeast. Genetics 136:53-64[Abstract].

EKWALL, K., O. NIELSEN, and T. RUUSALA, 1991  Repression of a mating type cassette in the fission yeast by four DNA elements. Yeast 7:745-755[Medline].

EKWALL, K., T. OLSSON, and T. RUUSALA, 1992  Trans-acting factors and properly positioned DNA elements repress mating-type genes in fission yeast. Curr. Genet. 21:331-338[Medline].

EKWALL, K., J.-P. JAVERZAT, K. LORENTZ, H. SCHMIDT, and G. CRANSTON et al., 1995  The chromodomain protein Swi6: a key component at fission yeast centromeres. Science 269:1429-1431[Abstract/Free Full Text].

EKWALL, K., E. R. NIMMO, J.-P. JAVERZAT, B. BORGSTROM, and R. EGEL et al., 1996  Mutations in the fission yeast silencing factors clr1⁺ and rik1⁺ disrupt the localisation of the chromo domain protein Swi6p and impair centromere function. J. Cell Sci. 109:2637-2648[Abstract].

FRIEDMAN-OHANA, R., I. KARUNKER, and A. COHEN, 1998  Chi-dependent intramolecular recombination in Escherichia coli.. Genetics 148:1-13[Free Full Text].

GERASIMOVA, T. I. and V. G. CORCES, 1996  Boundary and insulator elements in chromosomes. Curr. Opin. Genet. Dev. 6:185-192[Medline].

GEYER, P. K., 1997  The role of insulator elements in defining domains of gene expression. Curr. Opin. Gen. Dev. 7:242-248[Medline].

G