A Fission Yeast Repression Element Cooperates With Centromere-like Sequences and Defines a mat Silent Domain Boundary

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

REII is a Schizosaccharomyces pombe repression element located at the centromere-proximal end of the mat silent domain. Here we show that inversion of REII enhances silencing on its centromere-proximal side while suppressing silencing on its centromere-distal side. Transplacement of REII to a position 2.5 kb from its native locus extends the region of stringent repression to the new REII site. These results suggest that REII defines a mat silent domain boundary by acting preferentially toward its centromere-distal side. To investigate cooperation between REII and a K-region sequence that shares homology with the centromeric dg dh repeats (cen2 homology), we targeted combinations of these elements to an ectopic site and monitored expression of an adjacent reporter gene. Centromeric dh-like sequences conferred low-level silencing on the adjacent reporter gene, and REII, which did not display silencing activity on its own, enhanced cen2 homology-mediated silencing. Cooperation was also apparent at the mat locus, where deletion of REII impaired repression stability. We propose that REII and the cen2 homology play different yet complementary roles in silencing establishment and inheritance at the mat locus.

CHROMOSOMES of eukaryotic cells are organized into discrete functional domains that delineate independent units of gene activity. The expression state within each unit is controlled autonomously by cis-acting elements that recruit transcription factors or chromatin remodeling proteins and by boundary elements that protect internal genes from the long-range effect of external enhancers or silencers (reviewed in GERASIMOVA and CORCES 1996 ; GEYER 1997 ; KAMAKAKA 1997 ; KIOUSSIS and FESTENSTEIN 1997 ; SHERMAN and PILLUS 1997 ; KELLUM and ELGIN 1998 ; SUN and ELGIN 1999 ; UDVARDY 1999 ). Boundaries between constitutively expressed and repressed domains are not always distinct, and genes in peripheral regions may be subjected to stochastic, but clonally inherited repression. This phenomenon, named position effect variegation (PEV), was first described in Drosophila (MULLER 1930 ). Since then, PEV was observed in other organisms, including budding and fission yeast (GOTTSCHLING et al. 1990 ; ALLSHIRE et al. 1994 ; KARPEN 1994 ; EISSENBERG et al. 1995 ; HENIKOFF and MATZKE 1997 ; AYOUB et al. 1999 ). A variegated phenotype may also be caused by alternative states of chromatin assembly, affecting the entire length of the silent domain (PILLUS and RINE 1989 ; GREWAL and KLAR 1996 ; THON and FRIIS 1997 ; CAVALLI and PARO 1998B ; KLAR et al. 1998 ).

The fission yeast Schizosaccharomyces pombe switches its mating type by transposing a copy of unexpressed genes from the respective mat2-P or mat3-M donor cassettes to the transcriptionally active mat1 (BEACH 1983 ; BEACH and KLAR 1984 ). The donor cassettes flank a 10.9-kb region, named K, which is also stringently repressed (THON and KLAR 1992 ; THON et al. 1994 ; GREWAL and KLAR 1997 ). The silent mat2-K-mat3 domain is separated from the transcriptionally active mat1 by the L region (Fig 1). Several lines of evidence indicate that repression at the mat locus is accomplished by chromatin remodeling: silencing is regional, rather than gene specific (THON et al. 1994 ; GREWAL and KLAR 1997 ; AYOUB et al. 1999 ); mutations in silencing genes increase DNA accessibility in the repressed region to in vivo methylation (SING et al. 1998 ); two silencing genes, swi6 and clr4, encode chromodomain proteins (LORENTZ et al. 1994 ; GREWAL et al. 1998 ) that, like their Drosophila and mammalian counterparts, are likely to play a role in maintaining epigenetically controlled chromatin states (PLATERO et al. 1995 ; CAVALLI and PARO 1998A ); two other silencing genes, clr3 and clr6, encode putative histone deacetylases (IVANOVA et al. 1998 ); and transient exposure to trichostatin A, a specific inhibitor of histone deacetylases (YOSHIDA et al. 1990 ), has a long-lasting derepression effect (GREWAL et al. 1998 ; OLSSON et al. 1999 ).

View larger version (14K): In this window In a new window Download PPT slide   Figure 1. The mating-type region of S. pombe. The mat2-P and mat3-M mating-type donor cassettes are located within a silent domain of 17 kb and are separated from each other by a repressed region named K. The silent mat2-K-mat3 domain is separated from the transcriptionally active mat1 by a region of 15 kb named L. S. pombe switches its mating type by transposing a copy of the unexpressed genes from the respective mat2-P and mat3-M cassettes to mat1. The locations of the silent donor cassettes (solid boxes), mat1 (open box), cen2 homology, REII, mat3 silencer, and relevant restriction sites are indicated.

Repression is gradually alleviated with increasing the distance from mat2-P, leading to variegated expression of reporter genes along a stretch of 3 kb in the L region (AYOUB et al. 1999 ). Separation between the stringently repressed domain and the region of variegated expression appears to be controlled by a cis-acting element, named REII, located at the junction between mat2-P and the L region. REII was first identified as one of four cis-acting elements that cooperatively repress plasmid-borne mat2-P genes (EKWALL et al. 1991 ). Its deletion from the chromosome has only a subtle effect on the repressed state of mat2-P. Yet this deletion markedly enhances mat2-P expression in swi6 or clr1-clr4 mutants (THON et al. 1994 , THON et al. 1999 ; AYOUB et al. 1999 ). Intriguingly, REII inhibits the propagation of an active state, associated with gene expression in the L region, into the silent domain. Thus, in REII deletion mutants, expression of mat2-P and an adjacent ura4⁺ gene in the K region covariegates uniformly with ade6⁺ expression from the L region. Yet in strains with REII intact, genes in the silent domain are stringently repressed, regardless of the expression state in the L region (AYOUB et al. 1999 ). The mechanism by which REII ensures separation between the regions of stringent and variegated repression is not yet understood. One possibility is that REII acts as a repression element with a preferred directionality that helps assemble a heterochromatin complex on its centromere-distal side. The other is that REII acts as an insulator that inhibits the spreading of a transcriptionally active chromatin state into the mat2-K-mat3 domain. These two possibilities are not necessarily mutually exclusive.

Repression at the chromosomal mat2-K-mat3 region is controlled by at least two additional cis-acting elements. One element, located near mat3-M, controls the repressed state of mat3 and of markers at the centromere-distal part of the silent domain (THON et al. 1999 ). The other element, located within the K region, affects the expression state along the entire silent domain. Deletion of a 7.5-kb fragment from the K region lowers repression frequency, but the alternative epigenetic states are stably inherited in mitosis and meiosis. The K region also plays an important role in controlling mating-type switching directionality, as is evident from the observation that in K mutants switching competence covariegates with silencing (GREWAL and KLAR 1996 ; THON and FRIIS 1997 ). Remarkably, about one-third of the K region is homologous to the dgIIa and dhIIa centromeric repeat (GREWAL and KLAR 1997 ). Because transgenes in the centromeres are subjected to position effect repression (ALLSHIRE et al. 1994 ), it has been postulated that the cen2 homology is an active repression element in the K region (GREWAL and KLAR 1997 ).

To investigate the role of REII in promoting silencing at the mat silent domain, we examined the effect of its deletion, transplacement, or inversion on the expression state of reporter genes within the silent domain and at its periphery. We also attempted to create a synthetic silent domain at an ectopic site by inserting a reporter gene with various combinations of the centromeric outer repeat homology and REII and by monitoring reporter gene expression. The results indicate that dh sequences within the cen2 homology were sufficient to establish repression at an ectopic site. REII has no detectable silencing activity on its own. Yet it acts with a preferred directionality and cooperates with the cen2 homology to enhance silencing stability and define the boundary of the silent domain at the mat locus.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strain construction:
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 parts of the flanking L and K regions (BEACH and KLAR 1984 ), were performed on derivatives of Bluescript (ALTING-MEESE and SHORT 1989 ). To insert an ade6⁺ gene or an REII cassette into the designated mat locations, chimera plasmids were linearized by the appropriate restriction endonuclease and ligated to a 3.5-kb PvuII-SmaI fragment carrying ade6⁺ (SZANKASI et al. 1988 ) or a 0.23-kb EcoRI-BssHII fragment containing an REII cassette (EKWALL et al. 1991 ). Where necessary, single-stranded ends were converted to blunt ends by T4 DNA polymerase-mediated synthesis or digestion. To insert ade6⁺ into the HpaI site in the L region, the appropriate 5.5-kb SacI fragment of the L region was isolated from the CM740 cosmid (MIZUKAMI et al. 1993 ) and cloned into the SacI site of Bluescript. The Bluescript derivative (pNA1) was linearized by HpaI digestion and ligated to a SmaI-PvuII fragment carrying the ade6⁺ gene (pNA2).

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).

Molecular manipulations of a cloned 1.8-kb HindIII fragment carrying the ura4⁺ gene (BACH 1987 ) were conducted on a Bluescript derivative (pNA3). The SpeI site in Bluescript was inactivated before ura4⁺ insertion by SpeI digestion, conversion of the single-stranded overhangs to blunt ends and religation. ade6⁺ or combinations of ade6⁺ with the indicated REII and/or K region fragments (Fig 4 and Fig 5) were inserted into the StuI site within ura4. The 5.8-kb cen2 homology fragment was generated by PCR using pSGK (GREWAL and KLAR 1997 ) as a template and ATGTCTACTTCAAAACTCGC and CCATGTTCCATTACATATCC as primers. Targeted integration of molecular constructs into the ura4 locus was accomplished by transformation with purified SacI-ApaI fragments of the appropriate plasmids. Integration of molecular constructs at the desired sites was confirmed by Southern hybridization analysis (SOUTHERN 1975 ). Standard genetic crosses (MORENO et al. 1991 ) were used in strain constructions.

View larger version (19K): In this window In a new window Download PPT slide   Figure 2. The effect of REII deletion or transplacement on the expression state of ade6+ at sites in the mat region. The diagrams show the locations of the various ade6+ insertions and the transplacement or inversions of REII. Cells from patches on YEA plates were plated on low adenine (YE) medium and the percentages of red (Ade-), white (Ade+), and pink (intermediate levels of ade6 repression) were determined. The locations of mat2-P, REII (arrowhead), and relevant restriction sites are indicated in the top diagram. A shaded box indicates an ade6+ insertion, an open arrowhead indicates deleted REII, and a reversed arrowhead designates REII inversion.

View larger version (35K): In this window In a new window Download PPT slide   Figure 3. Inversion of REII enhances silencing on its centromere-proximal side. ade6 and his1 probes were used in Northern blot hybridization analysis of RNA preparations from cells with an ade6+ insertion at the SacI site in the L region and the indicated REII genotypes. Cultures were of cells from colonies on YE medium, showing Ade+ (white) or partially repressed (pink) phenotypes. The first two control lanes are of RNA preparations from an ade6-DN/N mutant (FY370) and an ade6+ strain (AP128). The REII+ strain is AP293, the REII inversion mutant is AP294, and the REII deletion mutant is AP295. The indicated values under each lane are arbitrary values of the respective ade6 hybridization signal divided by the signal in the control ade6+ lane and corrected for the ratio of the his1 signals.

View larger version (68K): In this window In a new window Download PPT slide   Figure 4. Inversion of REII suppresses mat2-P silencing. Colonies of stable mat1-M strains with the indicated REII and clr1 genotypes were grown on sporulation medium (PM-N), exposed to iodine vapor, and photographed. The percentage of clr1-165 cells showing a haploid meiosis phenotype (H.M.) was determined by microscopic examination of at least five fields. Each field was of an independent colony and 80–100 cells were scored in each field. Photographed colonies are of the following strains: SP1172 (clr1+ REII+), AP125 (clr1+ REII), AP262 (clr1+ REIIinv), PG377 (clr1-165 REII+), AP377 (clr1-165 REII), and AP262 (clr1-165 REIIinv). See Table 1 for detailed genotypes.

View larger version (22K): In this window In a new window Download PPT slide   Figure 5. Cooperative effect of a K-region fragment and REII enhances repression at an ectopic site. The figure includes structures of the various molecular constructs inserted in the ura4 locus, as well as the name of the strain and its swi6 genotype. Cells of Ade- (red) or Ade+ (white) colonies were replated on YE medium and the percentage of colonies showing the indicated phenotypes was determined by the colony color assay. clr1-clr4 derivatives of AP383 (not shown) had a similar phenotype as swi6 derivative (AP384). An arrow marks the orientation of the inserted ade6 gene. The polarity of REII (arrowhead) with respect to the K region in AP383 is the same as at the native mat locus. The polarity of REII in AP389 is inverted. n.r., not relevant.

Northern analysis:
Total cellular RNA was isolated from 10-ml aliquots of growing cultures (10⁷ cells/ml) in YEA medium, using the TRIZol reagent (GIBCO-BRL, Gaithersburg, MD), according to the manufacturer's protocol. All strains used for Northern analysis had an ade6-DN/N mutation (EKWALL et al. 1997 ), and the DNA probe used to detect the ade6 transcript was a 150-nucleotide (nt) NcoI fragment homologous to the DN/N deletion. Thus, only transcripts of the reporter gene were detectable. Quantitation of relative transcript levels was by phosphorimaging analysis. Transcript levels were standardized relative to his1 RNA, detected by hybridization to a 503-nt his1 fragment, generated by PCR, using pBS-his1 plasmid (R. WEISMAN, personal communication) as a template and ACAAGGTCGAGAAGAAAGCG and CCATCCAGGTTCATCCAAAG as primers.

Culture conditions:
Strains were grown on rich medium (YEA), adenine-limiting rich medium (YE), or sporulation medium (PM-N) (MORENO et al. 1991 ). All incubations were at 30°. For scoring Ade phenotypes on YE plates, standard incubation periods were 4 days at 30°. More than 500 colonies from each of 5 independent colonies of each strain were examined. One standard deviation values (±) are presented.

Iodine staining:
Haploid meiosis phenotype in heterothallic strains was examined by staining colonies on PM-N sporulation medium (MORENO et al. 1991 ) with iodine vapors, since spores, but not vegetative cells, contain a starch component. Plates were incubated for 4 days at 30° before staining (BRESCH et al. 1968 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

To explore the possibility that REII acts as a repression element with a preferred directionality, we determined the effect of its deletion, transplacement, or inversion on the expression state within the silent domain and its periphery. This was achieved by targeting an ade6⁺ reporter gene to sites within mat2-P or the L region and monitoring ade6 expression by examining colony color on low adenine medium (YE). On this medium, red and white colonies imply Ade^- and Ade⁺ phenotypes, respectively (MORENO et al. 1991 ), and pink colonies indicate intermediate levels of ade6 repression (ALLSHIRE et al. 1994 ). Unless indicated otherwise, all strains used in these experiments carried the ade6-210 mutation at the endogenous ade6 locus.

REII enhances silencing stability:
An ade6⁺ reporter gene, located within the mat2-P cassette (BamHI), is stringently repressed in >97% of the cell lines (AYOUB et al. 1999 ). Deletion of REII alleviated repression (Fig 2, AP363). Most colonies of the REII deletion mutant displayed an Ade⁺ phenotype and the rest exhibited intermediate levels of repression. Deletion of REII had a lesser effect on ade6 expression from the SacI site in the L region. Normally, 20% of the colonies of the L(SacI)::ade6⁺ strain (AP165) exhibit partial ade6 repression (AYOUB et al. 1999 ). Deletion of REII (AP347) decreased the proportion of colonies exhibiting any degree of ade6 repression to 10%.

Deletion of a 7.5-kb K-region fragment alleviates repression at the mat silent domain, but the alternative expression states are stably maintained (GREWAL and KLAR 1996 ; THON and FRIIS 1997 ). To compare the effect of K and REII deletions on the stability of the alternative expression states within mat2-P, we constructed the appropriate derivatives of a mat2-P(BamHI)::ade6⁺ strain. Cell suspensions from white (Ade⁺) or red (Ade^-) colonies on YE medium were replated on the same medium and incubated at 30°, and the proportion of red, white, and sectored colonies was determined (Table 2). The expressed state in the REII⁺ K⁺ control strain (AP313) was unstable. Less than 5% of the cells retained the Ade⁺ phenotype upon replating and the rest yielded red or sectored colonies. Conversion from the expressed to the repressed state in the REII mutant (AP363) was lower than that in the strain with REII intact, but higher than in the K mutant (AP421). Consistent with earlier studies (GREWAL and KLAR 1996 ; THON and FRIIS 1997 ), deletion of the 7.5-kb K-region fragment had little or no effect on the stability of the repressed state within mat2-P. Less than 0.5% of the cells from Ade^- colonies yielded Ade⁺ colonies upon replating and the frequency of sectored colonies was 1%. Unlike the K mutation, the REII mutation impaired the stability of the repressed state. About 6% of the cells from Ade^- colonies yielded white (Ade⁺) colonies and the frequency of sectored colonies exceeded 40% (Table 2). About half of the sectored colonies had multiple white (Ade⁺) sectors. To estimate the rate of change per cell division, we determined the frequency of half-sectored colonies (ALLSHIRE et al. 1995 ). This frequency was 2.4% for the REII strain and <0.1% for the isogenic strains with REII intact or with a K mutation. These results indicate a role for REII in assuring silencing stability.

Transplacement of REII toward mat1 extends the silent domain:
The decline in repression stringency on the centromere-proximal side of REII (AYOUB et al. 1999 ) and the effect of REII deletion on reporter gene repression within mat2-P (Fig 2) suggest that REII is a repression element that defines the boundary of stringent repression. To test this proposition, we asked whether transplacement of REII toward mat1 would extend the region of stringent repression. Transplacement of REII to the centromere-proximal side of ade6⁺ at the SacI site in the L region (Fig 1) enhanced ade6⁺ repression. The frequency of colonies exhibiting any degree of ade6⁺ repression increased from 20% in the wild-type strain (Fig 2, AP165) or 10% in the REII mutant (AP347) to >99% in the transplacement mutant (AP394). Furthermore, unlike in the REII strain or in the strain with REII intact, where intermediate levels of repression were observed, repression at the SacI site of the transplacement mutant was stringent (Fig 2). Silencing at the extended region of repression was alleviated by any of the swi6 and clr1-clr4 mutations (data not shown), thus indicating a role for chromatin remodeling proteins in REII-mediated repression at the L region. Further translocation of REII to a distance of 6.0 kb from its native locus (HpaI) did not extend the silent domain. The Ade⁺ phenotype of a strain with an ade6⁺ insertion at the HpaI site was not affected by transplacement of REII to the centromere-proximal side of the reporter gene (Fig 2, AP407). These results suggest that REII has no repression activity on its own. However, if located close enough to the silent domain it cooperates with an internal cis-acting element(s) to enhance repression on its centromere-distal side. Attempts to confirm this hypothesis and identify elements that cooperate with REII are described below.

Orientation dependence of REII activity:
If REII acts as a repression element with a preferred directionality, its inversion should enhance silencing on its centromere-proximal side. To test this prediction, we constructed a strain with an inverted REII at its native locus and an ade6⁺ insertion at the SacI site in the L region (Fig 1). We then compared the expression states of ade6⁺ in this strain to that in an isogenic strain with REII in the original orientation. The differences in the stability of the alternative ade6 expression states between the strains with REII in the original and inverted orientation (Table 3) indicate that inversion of REII enhances repression on its centromere-proximal side.

To rule out alternative interpretations of the colony color assay, we conducted Northern hybridization experiments with an internal ade6⁺ sequence as a probe (Fig 3). All strains used in this experiment had an ade6-DN/N deletion (EKWALL et al. 1997 ), and the hybridization probe was homologous to the fragment that was deleted from ade6 at its native locus. Thus, only transcripts of the reporter gene at the L region generated a hybridizable RNA product. Results of the Northern hybridization experiments clearly indicate that inversion of REII enhances repression on its centromere-proximal side. The levels of ade6 expression from the SacI site were lower in cultures of the inversion mutant than in cultures of the wild-type strain. The corrected ratios of these values were 0.1 for cultures originated from Ade^- or Ade⁺ colonies.

If a transplaced REII at the SacI site acts with a preferred directionality, its inversion should alleviate ade6 repression on its centromere-distal side. We tested this prediction by determining the effect of REII orientation on the expression state of an adjacent ade6⁺ gene in REII transplacement mutants. REII enhanced ade6⁺ repression at the SacI site in the L region if placed on its centromere-proximal side in the same orientation as in its native location (Fig 2, AP394). However, when placed in an inverted orientation (AP396), its silencing enhancement activity was markedly reduced.

We next asked whether inversion of REII at its native locus would affect the expression state of mat2-P. Deletion of REII has only a subtle effect on mat2-P repression in wild-type strains, but a combination of a REII mutation with any one of the swi6 or clr1-clr4 mutations has a synergistic derepression effect (THON et al. 1994 ; AYOUB et al. 1999 ). Therefore, to enhance the sensitivity of the assay, we examined the effect of REII inversion on the expression state of mat2-P in clr1 mutants as well as in wild-type cells. mat2-P expression was monitored by assaying for haploid meiosis in heterothallic mat1-M strains. In these strains derepression of mat2-P results in simultaneous expression of P and M genes in the same haploid cell, and this leads to haploid meiosis (KELLY et al. 1988 ). Haploid meiosis was monitored by microscopic examination and iodine staining of colonies on sporulation medium (BRESCH et al. 1968 ). Deletion or inversion of REII had no detectable effect on the repressed state of mat2-P in clr1⁺ cells. Nevertheless, either one of these REII mutations markedly enhanced mat2-P expression in clr1 mutants (Fig 4).

Taken together, these observations are consistent with the hypothesis that REII is a repression element that defines the boundary of the silent domain by acting with a preferred directionality toward its centromere-distal side.

Creation of a silent domain at an ectopic site:
The results of the preceding experiments suggest that REII has no silencing activity on its own. Yet it seems to cooperate with an internal cis-acting element to enhance silencing at the centromere-proximal end of the silent domain. To explore this possibility, we attempted to assemble a synthetic silent domain at an ectopic site by combining the activities of REII and K-region DNA fragments. Molecular constructs consisting of an ade6⁺ reporter gene and various combinations of REII and a 6.3-kb K-region DNA fragment (BseRI-BseRI; Fig 1) were targeted to the ura4 locus on chromosome 3. The expression state of ade6⁺ in the respective strains was monitored by the colony color assay. Consistent with the data in Fig 2 (AP407), expression of ade6⁺ from the ura4 locus was not affected by an adjacent REII cassette (Fig 5, AP379). On the other hand, the K-region fragment conferred low repression frequency at the ectopic site (AP179). A small proportion of the ura4::ade6⁺-K colonies (0.1%) displayed a partial Ade^- phenotype (pink colonies on YE medium) and 15% of the cells in these colonies maintained this phenotype upon replating. Replating of cells from white (Ade⁺) colonies yielded a very low proportion of red or sectored colonies.

To test for cooperation between REII and the K-region fragment, we targeted a molecular construct containing an ade6 reporter gene flanked by the two elements to the ura4 locus (AP383). The expression state of ade6 in the ura4::REII-ade6⁺-K strain was compared to that in the isogenic strains with K-ade6⁺ or ade6⁺-REII insertions. REII markedly enhanced K-region-mediated repression. About 15% of the ura4::REII-ade6⁺-K cells from Ade⁺ colonies yielded red or sectored colonies, and the majority of the cells from Ade^- colonies retained a repressed or partially repressed state upon replating. As in its native location, REII activity had a preferred directionality at the ectopic site. Inversion of the REII cassette with respect to the reporter gene (AP389) lowered its silencing-enhancing capacity.

To confirm that ade6 repression at the ura4 locus occurred by a chromatin remodeling mechanism, we examined silencing dependence on swi6 and clr1-clr4 genes. ade6 repression in the ura4::REII-ade6⁺-K strain was totally alleviated by mutations in any of the tested silencing genes (AP384 is an example). These results indicate that cooperation of the K-region fragment with REII establishes and stably maintains a repressed chromatin state at an ectopic site.

Silencing activity of dh-like sequences within the cen2 homology:
The cen2 homology within the K region has been postulated to promote heterochromatin assembly at the mat locus (GREWAL and KLAR 1997 ). To test this hypothesis, we replaced the 6.3-kb K-region fragment in molecular constructs at the ura4 locus with a cen2 homology sequence and monitored ade6 expression in the respective strains. A 3.6-kb SnaBI-HaeIII fragment (Fig 6, AP804), containing 84% of the cen2 homology, acted as a weak silencer outside the mat region context. Cells from white colonies rarely grew to red or pink colonies and 50% of the cells from red colonies maintained a partially repressed phenotype upon replating. As expected, cen2-mediated silencing depended on the activity of swi6 and clr1-clr4 silencing genes (data not shown). These observations imply that sequences within the cen2 homology are sufficient to establish a heterochromatin structure at an ectopic site.

View larger version (43K): In this window In a new window Download PPT slide   Figure 6. Silencer activity of cen2 homology sequences. The indicated restriction fragments of the cen2 homology and PCR-generated dg and dh fragments (top diagram) were incorporated into the molecular constructs at the ura4 locus. The distribution of Ade phenotypes of colonies originated by cells from red (Ade-) or white (Ade+) colonies on YE plates was determined by the colony color assay as in Fig 5. n.r., not relevant.

We tested for cooperation between the 3.6-kb cen2 homology and REII. A molecular construct consisting of an ade6⁺ reporter gene flanked by REII and the cen2 homology was targeted to the ura4 locus, and the expression state of ade6 was determined by the colony color assay. REII enhanced cen2 homology-mediated repression. More than 95% of the cells from Ade^- colonies retained a fully or partially repressed state upon replating, and 6% of the cells from Ade⁺ colonies regained it (Fig 6, AP270). These results demonstrate that cen2 homology-mediated silencing is enhanced by cooperation with REII.

In all the molecular constructs of Fig 5 and in the first two constructs of Fig 6 (AP804 and AP270) ade6⁺ was inserted with its promoter distal to the cen2 homology. Inversion of ade6⁺ (AP259) had a minor effect on the stability of the alternative expression states. This inversion enhanced the stability of the repressed state and lowered the stability of the derepressed state.

To determine whether the cen2 homology acts with a preferred directionality, we inverted the 3.6-kb fragment in the molecular construct at the ectopic site and determined the stability of the alternative expression states. Inversion of the cen2 homology with respect to the reporter gene (AP263) did not abolish its activity. Yet silencing activity in one orientation was higher than in the other.

Studies of centromere activities indicate functional redundancy among sequences of the centromeric outer repeats (BAUM et al. 1994 ; NGAN and CLARKE 1997 ). To test for redundancy in silencing-promoting activity, we compared the ability of four cen2 homology fragments, ranging in length from 0.56 to 2.2 kb, to confer silencing at the ectopic site (Fig 6). The three fragments sharing homology with the dhII repeat (AP278, AP287, AP288) conferred chromatin-mediated repression on an adjacent ade6⁺ gene. Yet the 0.56-kb dgII-like sequence (AP803) had no detectable silencing activity. Assuming that the proportion of cells from red colonies maintaining the Ade^- phenotype upon replating reflects repression stability, the data imply that repression stability in strains with dhII homology at the ectopic site was similar to that in the strain with the longer cen2 homology insertion. However, significant differences in the proportions of cells from white colonies yielding pink colonies upon replating were observed. Fragments of 0.58 kb (AP288) and 1.4 kb (AP287) that share 99% homology with the centromeric dhII repeats were more effective than the entire 3.6-kb cen2 homology fragment, whereas a 2.2-kb fragment (AP278), consisting mainly of dgII sequences, was somewhat less effective.

We asked whether the configuration of cooperating elements with respect to the reporter gene affects cooperation efficiency. Specifically, is a reporter gene flanked by the two elements more stringently repressed than a reporter gene with both elements on its same side? To address this question, we placed REII and the cen2 homology on the same side of ade6⁺ and targeted the molecular construct to the ura4 locus. REII silencing-enhancing activity at the ectopic site was abolished when placed together with the cen2 homology on the same side of the ade6⁺ gene. Less than 0.1% of the cells from Ade⁺ colonies acquired the Ade^- phenotype and only 37% of the cells from Ade^- colonies maintained partial ade6 repression upon replating (Fig 6, AP419). This suggests that cooperation between the two elements outside their native chromosomal context affects only the region between the elements. However, the possibility that the orientation of the ade6⁺ reporter gene with respect to the cen2 homology or to that surrounding DNA sequences affects silencing cannot be ruled out.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Silencing along the chromosomal mat2-K-mat3 domain is mediated by several trans-acting proteins and by at least three cis-acting elements: the cen2 homology, REII, and the mat3 silencer (GREWAL and KLAR 1996 ; THON and FRIIS 1997 ; AYOUB et al. 1999 ; THON et al. 1999 ). Here we show that sequences within the cen2 homology promote silencing at an ectopic site through a mechanism that depends on chromatin-modifying proteins and that REII cooperates with the cen2 homology to enhance silencing. We also show that REII enhances the stability of the repressed state at the mat locus and acts with a preferred directionality to define the boundary of stringent repression at the junction between mat2-P and the L region.

By acting with a preferred directionality, REII defines the boundary of stringent repression:
A cis-acting element may define the boundary of a silent domain by acting as an insulator that blocks the propagation of enhancer or silencer activities (SUN and ELGIN 1999 ; UDVARDY 1999 ) or by organizing repressive chromatin in a unidirectional manner (GDULA et al. 1996 ; BI et al. 1999 ). The following evidence suggests that REII defines the boundary of stringent repression at the junction between mat2-P and the L region by promoting silencing to its centromere-distal side: (a) repression is stringently controlled on the centromere-distal side of REII and is gradually alleviated on its centromere-proximal side (AYOUB et al. 1999 ); (b) inversion of REII enhances silencing on its centromere-proximal side while suppressing silencing on its centromere-distal side; (c) transplacement of REII to a site within the L region extends the region of stringent repression, and this extension depends on a direct orientation of REII and the activity of chromatin modifying proteins. Thus, the location and orientation of REII define the boundary of stringent repression at the centromere-proximal end of the mat silent domain. This conclusion is consistent with the observation that REII ensures repression in the silent domain, regardless of the expression state in the L region (AYOUB et al. 1999 ).

The reason that silencing enhancement capacity of a translocated REII is diminished as its distance from mat2-P increases from 2.5 (SacI) to 6.0 kb (HpaI; Fig 2) is not yet understood. One possibility is that REII activity depends on cooperation with an internal repression element, like the cen2 homology, and this cooperation can take place only within a limited chromosomal distance. Another possibility is that as the distance between the two elements increases, the length of the silent domain becomes limited by the availability of heterochromatin components. A third possibility is that a cis-acting element, located between the HpaI and SacI sites, interferes with the cooperation between REII and the cen2 homology. Further analysis of the L region should help distinguish between these possibilities.

Cis-acting repression elements at the mat locus:
Deletion of any one of the three cis-acting elements that promote silencing at the mat locus alleviates repression of reporter genes within the silent domain and leads to a variegated phenotype. However, different deletion mutants display different modes of variegation. Whereas K mutants adopt one of two stable states of reporter gene expression (GREWAL and KLAR 1996 ; THON and FRIIS 1997 ; and this study), the repressed state in strains with deletions of REII or the mat3 silencer is highly unstable (AYOUB et al. 1999 ; THON et al. 1999 ; and this study). The stability of the alternative expression states in K mutants implies that an element within the K region is involved in repression establishment, but repression inheritance is independent of this element activity (GREWAL and KLAR 1996 ; THON and FRIIS 1997 ). Consistent with this hypothesis we show here that different sequences of the cen2 homology promote the establishment of chromatin-mediated repression at an ectopic site. However, the repressed state conferred by the cen2 homology alone is relatively unstable. Remarkably, REII, which does not display independent silencing activity outside the mat silent domain, enhances the stability of cen2-mediated repression at the ectopic site. The effect of the REII deletion on the stability of the repressed state within mat2-P suggests that it is likely to play a similar role at the mat locus.

REII activity is limited to the centromere-proximal end of the silent domain and a functionally similar element, the mat3 silencer, enhances repression at the centromere-distal end of the domain (THON et al. 1999 ). The proposition that REII and the mat3 silencer are functionally similar is consistent with our recent observations that the mat3 silencer has no detectable silencing activity on its own. However, like REII, it cooperates with the cen2 homology to enhance silencing stability at the ectopic site (I. GOLDSHMIDT and A. COHEN, unpublished results).

Altogether, these observations are consistent with the notion that the cen2 homology plays a key role in the assembly of a repressive chromatin structure at the mat locus. Furthermore, cooperation of this element with REII on its centromere-proximal side, and with the mat3 silencer on its centromere-distal side, enhances repression along the entire length of the mat2-K-mat3 silent domain. The phenotypes of the various deletion mutants and of strains with different combinations of the three repression elements at the ectopic site suggest that the cen2 homology mediates the establishment of the repressed state, while REII and the mat3 silencer contribute to silencing stability.

Silencing activity of centromeric DNA:
The intriguing discovery that one-third of the K region shares 96% homology with the centromeric dgII dhII repeats suggests that shared cis-acting elements may play similar roles in heterochromatin assembly at the mat and cen loci. The cen2 homology in the K region may also promote silencing at the mat locus through homologous DNA-DNA interactions with its centromeric counterparts (GREWAL and KLAR 1997 ). Such interactions have been postulated before to promote regional silencing by playing a functional role in nuclear organization (HENIKOFF 1997 ; HENIKOFF and MATZKE 1997 ). If mainly homologous interactions are involved, the length of the homology is likely to affect silencing proficiency. This is clearly not the case in the experiments described above. A 0.56-kb dgII homology has no detectable silencing-promoting activity. On the other hand, silencing-promoting activity of the 0.58-kb dhII homology is higher than that of the 3.6-kb or 2.2-kb cen2 homologies and similar to that of the 1.4-kb dhII homology. Thus, it is unlikely that merely homologous interactions are involved in cen2-homology-mediated silencing.The observation that each one of the nonoverlapping fragments of the cen2 homology is sufficient to promote silencing at the ectopic site is consistent with the notion that heterochromatization is promoted by redundant DNA elements distributed along centromeric sequences. The 580-bp fragment that promotes silencing at the ectopic site contains a 310-bp sequence that shares >99% homology with dh repeats in all three centromeres. Further analysis of the 580-bp fragment should define the minimal requirements for dh-like sequences to establish a repressed state and may reveal specific cis-acting sequences through which silencing proteins may exert their function at the cen and mat loci.

Cooperation at a distance between cis-acting elements:
REII has no detectable repression activity on its own. Yet it cooperates with the cen2 homology to enhance silencing at the ectopic site. Cooperation between cis-acting elements may involve either one or a combination of the following mechanisms: transient or stable interaction between proteins, associated with the cooperating elements to create a loop structure that defines a silent domain (PIRROTTA and RASTELLI 1994 ; HENIKOFF 1996 ); enhancement of local concentration of weak binding sites for chromatin remodeling proteins; or functional complementation by the proteins associated with the two elements. The indications that REII defines the boundary of stringent repression and that cooperation between REII and the cen2 homology at the ectopic site affect only the region between the elements are consistent with the "looped domain" hypothesis. However, the observation that inversion of REII at the mat locus enhances silencing on its centromere-proximal side, while suppressing silencing on its centromere-distal side, argues against this possibility. This observation is more easily explained by unidirectional propagation of a silencing-enhancing complex from REII toward the cen2 homology and a cumulative effect of the complexes propagating from the two elements toward each other. It is likely that a similar mode of cooperation between the cen2 homology and the mat3 silencer promotes silencing at the centromere-distal end of the silent domain. Because genetic experiments suggest that REII and the cen2 homology play different yet complementary roles in silencing, we favor the "functional complementation" hypothesis. We therefore speculate that the two elements recruit partially overlapping or different sets of proteins that propagate toward each other and that the cumulative effect of the two complexes promotes stringent repression at the centromere-proximal end of the silenced domain.

The speculative model in Fig 7 is based on results of this study—the observations that REII activity is limited to the centromere-proximal end of the silent domain (THON et al. 1999 ) and that repression is gradually alleviated as the distance from REII toward mat1 increases (AYOUB et al. 1999 ). We assume that DNA elements within cen2 homology, like their counterparts at the centromere outer repeats, serve as nucleation points for heterochromatin components that propagate in a bidirectional manner toward mat2-P and mat3-M. Normally, the gradual decrease in heterochromatin density as the distance from the cen2 homology increases is compensated by the complementary structures that propagate from REII (Fig 7A). This explains why repression is stringently controlled within the mat2-K-mat3 domain and is gradually alleviated with distance at the centromere-proximal side of REII (AYOUB et al. 1999 ). The gradual decline in cen2-mediated repression leads to variegated expression of reporter genes within mat2-P in REII deletion mutants (Fig 7B). Because REII activity propagates in a unidirectional manner, inversion of REII enhances repression on its centromere-proximal side while suppressing repression on its centromere-distal side (Fig 7C). Likewise, transplacement of REII to sites within the region of variegated repression (SacI) extends the region of stringent repression to the new REII locus (Fig 7D).

View larger version (21K): In this window In a new window Download PPT slide   Figure 7. Model for cooperation of the cen2 homology (cenH) with REII in promoting silencing and in defining the centromere-proximal limit of stringent repression at the mat silent domain. This model assumes that DNA elements within the cen2 homology serve as nucleation points for heterochromatin components that propagate toward mat2-P and mat3-M. (a) Normally, the gradual decrease in heterochromatin density as the distance from the cen2 homology increases is compensated by a complementary structure propagating in a unidirectional manner from REII. Thus, repression is more stringently controlled on the centromere-distal side of REII than on its centromere-proximal side (AYOUB et al. 1999 ). (b) Deletion of REII alleviates repression at the centromere-proximal end of the silent domain but not at its centromere-distal end, where cooperation of the cen2 homology and the mat3 silencer controls repression. (c) An inverted REII enhances repression on its centromere-proximal side while suppressing repression on its centromere-distal side. (d) Cooperation of a translocated REII with the cen2 homology extends the silent domain to the new REII locus. Presumably, cooperation of the cen2 homology with the mat3 silencer enhances repression at the centromere-distal end of the silent domain (not shown). REII is designated by a solid arrowhead, REII deletion by an open arrowhead, mat2 by a shaded box, silencing activity of REII by a dotted line, and silencing activity of cen2 homology by a broken line. A solid line illustrates the degree of silencing along the centromere-proximal end of the silent domain.

Possible roles for REII in epigenetically modulated silencing inheritance:
Stable chromosomal inheritance of the repressed state in K mutants is regulated by an epigenetic mechanism (GREWAL and KLAR 1996 ; THON and FRIIS 1997 ). Thus, heterochromatin at the mat locus not only promotes mating-type switching and silencing of the donor cassettes but also templates its own reformation every cell cycle. The nature of the epigenetic marking or the mechanism that initially marks the affected region is not yet understood. Our current study reveals that the repressed state conferred by the cen2 homology alone is relatively unstable, but stability is enhanced by cooperation of the cen2 homology with REII. Thus, REII is likely to be involved in the formation or maintenance of the putative epigenetic marking. REII may enhance the stability of an epigenetic state through either one of two nonmutually exclusive mechanisms: (a) cooperation between REII and cen2 homology may promote the assembly of a higher order chromatin structure that is physically different and more stably inherited than the structure assembled by the cen2 homology alone or (b) REII, like S. cerevisiae HML-E silencer (HOLMES and BROACH 1996 ), may help provide the genomic memory that assures persistence of the repressed state from one generation to the next. For example, a protein complex may remain associated with REII during replication and serve as a nucleus for the reassembly of a complete heterochromatin structure on each sister chromatid after replication. Resolution between these hypotheses must await a comparative analysis of the chromatin structures assembled at the mat locus in the wild-type cells and REII mutants and the monitoring of the chromatin state following in vivo deletion of REII.

	FOOTNOTES

¹ These authors contributed equally to this work.

	ACKNOWLEDGMENTS

We thank Genevieve Thon for critical review of the manuscript and Shiv Grewal, Amar Klar, Ronit Weisman, and Robin Allshire for plasmids and strains. This work was supported by the U.S.-Israel Binational Science Foundation and by the Peter Hilston Foundation. N.A. was supported by a fellowship to minority students from the Israeli Ministry of Science.

Manuscript received April 20, 2000; Accepted for publication July 17, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

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].

AYOUB, N., I. GOLDSHMIDT, and A. COHEN, 1999  Position effect variegation at the mating-type locus of fission yeast: a cis-acting element inhibits covariegated expression of genes at the silent and expressed domains. Genetics 152:495-508[Abstract/Free Full Text].

BACH, M.-L., 1987  Cloning and expression of the OMP decarboxylase gene ura4 from Schizosaccharomyces pombe.. Curr. Genet. 12:527-534[Medline].

BAUM, M., V. K. NGAN, and L. CLARKE, 1994  The centromeric K-type repeat and the central core are together sufficient to establish a functional Schizosaccharomyces pombe centromere. Mol. Biol. Cell 5:747-761[Abstract].

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].

BI, X., M. BRAUNSTEIN, G.-J. SHEI, and J. R. BROACH, 1999  The yeast HMLI silencer defines a heterochromatin domain boundary by directional establishment of silencing. Proc. Natl. Acad. Sci. USA 96:11934-11939[Abstract/Free Full Text].

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

CAVALLI, G. and R. PARO, 1998a  Chromo-domain proteins: linking chromatin structure to epigenetic regulation. Curr. Opin. Cell Biol. 10:354-360[Medline].

CAVALLI, G. and R. PARO, 1998b  The Drosophila Fab-7 chromosomal element conveys epigenetic inheritance during mitosis and meiosis. Cell 93:505-518[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. O., 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, B. M. TURNER, G. CRANSTON, and R. C. ALLSHIRE, 1997  Transient inhibition of histone deacetylation alters the structural and functional imprint at fission yeast centromeres. Cell 91:1021-1032[Medline].

GDULA, D. A., T. I. GERASIMOVA, and V. G. CORCES, 1996  Genetic and molecular analysis of the gypsy chromatin insulator of Drosophila.. Proc. Natl. Acad. Sci. USA 93:9378-9383[Abstract/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. Genet. Dev. 7:242-248[Medline].

GOTTSCHLING, D. E., O. M. APARICIO, B. L. BILLINGTON, and V. A. ZAKIAN, 1990  Position effect at Saccharomyces cerevisiae telomeres: reversible repression of Pol II transcription. Cell 63:751-762[Medline].

GREWAL, S. I. S. and A. J. S. KLAR, 1996  Chromosomal inheritance of epigenetic states in fission yeast during mitosis and meiosis. Cell 86:95-101[Medline].

GREWAL, S. I. S. and A. J. S. KLAR, 1997  A recombinationally repressed region between mat2 and mat3 loci shares homology to centromeric repeats and regulates directionality of mating-type switching in fission yeast. Genetics 146:1221-1238[Abstract].

GREWAL, S. I. S., M. J. BONADUCE, and A. J. S. KLAR, 1998  Histone deacetylase homologs regulate epigenetic inheritance of transcriptional silencing and chromosome segregation in fission yeast. Genetics 150:563-576[Abstract/Free Full Text].

HENIKOFF, S., 1996 Position-effect variegation in Drosophila: recent progress, pp. 319–334 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.

HENIKOFF, S., 1997  Nuclear organization and gene expression: homologous pairing and long-range interactions. Curr. Opin. Cell Biol. 9:388-395[Medline].

HENIKOFF, S. and M. A. MATZKE, 1997  Exploring and explaining epigenetic effects. Trends Genet. 13:293-295[Medline].

HOLMES, S. G. and J. R. BROACH, 1996  Silencers are required for inheritance of the repressed state in yeast. Genes Dev. 10:1021-1032[Abstract/Free Full Text].

IVANOVA, A. V., M. J. BONADUCE, S. V. IVANOV, and A. J. S. KLAR, 1998  The chromo and SET domains of the Clr4 protein are essential for silencing in fission yeast. Nat. Genet. 19:192-195[Medline].

KAMAKAKA, R. T., 1997  Silencers and locus control regions: opposite sides of the same coin. Trends Biochem. Sci. 22:124-126[Medline].

KARPEN, G. H., 1994  Position-effect variegation and the new biology of heterochromatin. Curr. Opin. Genet. Dev. 4:281-291[Medline].

KELLUM, R. and S. C. R. ELGIN, 1998  Chromatin boundaries: punctuating the genome. Curr. Biol. 8:521-524.

KELLY, M., J. BURKE, M. SMITH, A. KLAR, and D. BEACH, 1988  Four mating-type genes control sexual differentiation in the fission yeast. EMBO J. 7:1537-1547[Medline].

KIOUSSIS, D. and R. FESTENSTEIN, 1997  Locus control regions: overcoming heterochromatin-induced gene inactivation in mammals. Curr. Opin. Genet. Dev. 7:614-619[Medline].

KLAR, A. J. S., A. V. IVANOVA, J. Z. DALGAARD, M. J. BONADUCE and S. I. S. GREWAL, 1998 Multiple epigenetic events regulate mating-type switching of fission yeast, pp. 87–103 in Epigenetics. Novartis Foundation Symposium 214, Wiley, Chichester, UK.

LORENTZ, A., K. OSTERMANN, O. FLECK, and H. SCHMIDT, 1994  Switching gene swi6, involved in repression of silent mating-type loci in fission yeast, encodes a homologue of chromatin-associated proteins from Drosophila and mammals. Gene 143:139-143[Medline].

MIZUKAMI, T., W. I. CHANG, I. GARKAVTSEV, N. KAPLAN, and D. LOMBARDI et al., 1993  A 13 kb resolution cosmid map of the 14 Mb Fission yeast genome by nonrandom sequence-tagged site mapping. Cell 73:121-132[Medline].

MORENO, S., A. J. S. KLAR, and P. NURSE, 1991  Molecular genetic analysis of fission yeast Schizosaccharomyces pombe.. Methods Enzymol. 194:795-823[Medline].

MULLER, H. J., 1930  Types of visible variations induced by X-rays in Drosophila.. J. Genet. 22:299-334.

NGAN, V. K. and L. CLARKE, 1997  The centromere enhancer mediates centromere activation in Schizosaccharomyces pombe.. Mol. Cell. Biol. 17:3305-3314[Abstract].

OLSSON, T. G. F., R. A. SILVERSTEIN, K. EKWALL, and P. SUNNERHAGEN, 1999  Transient inhibition of histone deacetylase activity overcomes silencing in the mating-type region in fission yeast. Curr. Genet. 35:82-87[Medline].

PILLUS, L. and J. RINE, 1989  Epigenetic inheritance of transcriptional states in S. cerevisiae.. Cell 59:637-647[Medline].

PIRROTTA, V. and L. RASTELLI, 1994  white gene expression, repressive chromatin domains and homeotic gene regulation in Drosophila.. Bioessays 16:549-556[Medline].

PLATERO, J. S., T. HARTNESS, and J. C. EISSENBERG, 1995  Functional analysis of the chromo domain of HP1. EMBO J. 14:3977-3986[Medline].

SHERMAN, J. M. and L. PILLUS, 1997  An uncertain silence. Trends Genet. 13:308-313[Medline].

SING, J., V. GOEL, and A. J. KLAR, 1998  A novel function of the DNA repair gene rhp6 in mating-type silencing by chromatin remodeling in fission yeast. Mol. Cell. Biol. 18:5511-5522[Abstract/Free Full Text].

SOUTHERN, E., 1975  Detection of specific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol. 98:503-517[Medline].

SUN, F.-L. and S. C. R. ELGIN, 1999  Putting boundaries on silence. Cell 99:459-462[Medline].

SZANKASI, P., W. D. HE