A General Requirement for the Sin3-Rpd3 Histone Deacetylase Complex in Regulating Silencing in Saccharomyces cerevisiae

Corresponding author: Michael Hampsey, Department of Biochemistry, UMDNJ–Robert Wood Johnson Medical School, 675 Hoes Lane, Piscataway, NJ 08854. E-mail:hampsemi@umdnj.edu

Communicating editor: F. WINSTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The Sin3-Rpd3 histone deacetylase complex, conserved between human and yeast, represses transcription when targeted by promoter-specific transcription factors. SIN3 and RPD3 also affect transcriptional silencing at the HM mating loci and at telomeres in yeast. Interestingly, however, deletion of the SIN3 and RPD3 genes enhances silencing, implying that the Sin3-Rpd3 complex functions to counteract, rather than to establish or maintain, silencing. Here we demonstrate that Sin3, Rpd3, and Sap30, a novel component of the Sin3-Rpd3 complex, affect silencing not only at the HMR and telomeric loci, but also at the rDNA locus. The effects on silencing at all three loci are dependent upon the histone deacetylase activity of Rpd3. Enhanced silencing associated with sin3, rpd3, and sap30 is differentially dependent upon Sir2 and Sir4 at the telomeric and rDNA loci and is also dependent upon the ubiquitin-conjugating enzyme Rad6 (Ubc2). We also show that the Cac3 subunit of the CAF-I chromatin assembly factor and Sin3-Rpd3 exert antagonistic effects on silencing. Strikingly, deletion of GCN5, which encodes a histone acetyltransferase, enhances silencing in a manner similar to deletion of RPD3. A model that integrates the effects of rpd3, gcn5, and cac3 on silencing is proposed.

EPIGENETIC effects are heritable, but reversible, changes in gene expression due to alterations in chromatin structure or DNA methylation (reviewed in HENIKOFF and MATZKE 1997 ). Classic examples of epigenetic phenomena include X-chromosome inactivation in mammals (PANNING and JAENISCH 1998 ), position-effect variegation in Drosophila (WAKIMOTO 1998 ), imprinting of specific loci in mammals (JAENISCH 1997 ), and silencing of the cryptic mating-type loci (HM) in Saccharomyces cerevisiae (SHERMAN and PILLUS 1997 ). Epigenetic control usually involves gene silencing, defined as position-dependent, gene-independent transcriptional repression, involving formation of specialized chromatin structures that encompass large regions of the genome (SHERMAN and PILLUS 1997 ; GRUNSTEIN 1998 ). Despite the importance of silencing in regulating cell growth and development, the molecular mechanisms responsible for establishing and maintaining a particular chromatin structure are not well defined.

In addition to silencing at the cryptic HM mating loci, silencing in yeast has been described for reporter genes integrated proximal to telomeres (telomere position effect; GOTTSCHLING et al. 1990 ) and within the rDNA array (BRYK et al. 1997 ; FRITZE et al. 1997 ; SMITH and BOEKE 1997 ). In contrast to stable silencing at the HM and rDNA loci, telomeric silencing is variegated, resulting in stochastic patterns of repression for RNA pol II-transcribed genes integrated at telomeres. This effect is comparable to the spread of heterochromatin that accounts for position-effect variegation in flies.

The combination of yeast genetics and biochemistry has led to the discovery of many factors that affect silencing. These include the silent information regulatory (SIR) proteins, the repressor-activator protein Rap1, and the core histones H3 and H4 (IVY et al. 1986 ; RINE and HERSKOWITZ 1987 ; KAYNE et al. 1988 ; APARICIO et al. 1991 ; KYRION et al. 1993 ; THOMPSON et al. 1994 ). Models for silencing suggest that SIR proteins are recruited to DNA by Rap1 and then spread along the DNA by interaction with the N-terminal tails of H3 and H4 (reviewed in SHERMAN and PILLUS 1997 ; GRUNSTEIN 1998 ). Nonetheless, silencing occurs by distinct mechanisms at each of the silenced loci. For example, telomeric silencing requires Sir2-Sir4, but is independent of Sir1, which is required for the establishment of silencing at the HM loci. The only SIR protein required for rDNA silencing is Sir2. Indeed, deletion of the SIR4 gene enhances rDNA silencing (BRYK et al. 1997 ; FRITZE et al. 1997 ; SMITH and BOEKE 1997 ; SMITH et al. 1998 ), a consequence of redistribution of limiting Sir2 to the nucleolus in the absence of Sir4 (SMITH et al. 1998 ).

Other proteins also play important roles in silencing. Rad6 (Ubc2) is an E2 ubiquitin-conjugating enzyme involved in many cellular processes, including DNA repair, UV-induced mutagenesis, N-end rule protein degradation, sporulation (reviewed in PRAKASH et al. 1993 ), and Ty1 integration specificity (PICOLOGLOU et al. 1990 ; KANG et al. 1992 ; LIEBMAN and NEWNAM 1993 ). Recent evidence demonstrated that RAD6 is required for silencing at telomeres and the HM loci and that deletion of RAD6 derepresses Ty1 transcription and mitotic recombination at the rDNA locus. The ubiquitin-conjugating activity of Rad6, but not Rad6-mediated N-end rule protein degradation, is essential for these processes (BRYK et al. 1997 ; HUANG et al. 1997 ). In addition, Rad6 ubiquitinates histones H2A, H2B, and H3 in vitro (SUNG et al. 1988 ; HAAS et al. 1990 ). These results suggest a role for Rad6 as a modifier of localized chromatin structure (PICOLOGLOU et al. 1990 ; KANG et al. 1992 ; LIEBMAN and NEWNAM 1993 ). Consistent with this hypothesis, Ubp3, a ubiquitin hydrolase, physically interacts with Sir4, and deletion of UBP3 enhances telomeric and HML silencing in yeast (MOAZED and JOHNSON 1996 ). Another (putative) ubiquitin hydrolase, encoded by the DOT4 gene, also disrupts silencing when overexpressed (SINGER et al. 1998 ).

The chromatin assembly factor I (CAF-I) also affects silencing. Yeast CAF-I is composed of three subunits encoded by the CAC1, CAC2, and CAC3 genes (KAUFMAN et al. 1997 ). The cac1 and rlf2 alleles of CAC1 alter Rap1 localization, perturb telomeric chromatin, and reduce telomeric silencing (ENOMOTO et al. 1997 ; KAUFMAN et al. 1997 ; MONSON et al. 1997 ). Silencing at the HM loci is also reduced in cac1 and rlf2 mutants, and similar effects on telomeric and HM silencing are conferred by cac2 and cac3 mutations (KAUFMAN et al. 1997 ; ENOMOTO and BERMAN 1998 ). CAF-I affects the maintenance, but not the reestablishment, of silent chromatin (ENOMOTO and BERMAN 1998 ). Cac3 (Msi1) is structurally related to Hat2, the regulatory subunit of the yeast B-type histone acetyltransferase (PARTHUN et al. 1996 ), to the RbAp48 and RbAp46 subunits of mammalian histone deacetylase complexes (PARTHUN et al. 1996 ; TAUNTON et al. 1996 ; ZHANG et al. 1997 ), and to the NURF-55 subunit of a Drosophila chromatin remodeling complex (MARTINEZ-BALBAS et al. 1998 ). Thus, Cac3 provides a structural link among four distinct complexes that affect histone metabolism.

Silent DNA is packaged into hypoacetylated nucleosomes that exhibit a pattern of histone acetylation reminiscent of metazoan heterochromatin (TURNER et al. 1992 ; BRAUNSTEIN et al. 1993 , BRAUNSTEIN et al. 1996 ). These results implicate histone acetyltransferases and deacetylases in silencing. Indeed, the Rpd3 histone deacetylase, and its associated protein Sin3, affect silencing at telomeric and HM loci (DE RUBERTIS et al. 1996 ; RUNDLETT et al. 1996 ; VANNIER et al. 1996 ). However, in contrast to their role in gene-specific repression, Rpd3 and Sin3 disrupt, rather than establish or maintain, silencing. A similar effect on silencing was reported for Rpd3 in Drosophila (DE RUBERTIS et al. 1996 ).

Mammalian and yeast Sin3 and Rpd3 proteins exist in large multisubunit complexes, estimated to be >2 MD in the case of the yeast Sin3-Rpd3 complex (KASTEN et al. 1997 ). In addition to Sin3, which appears to function as a platform for complex assembly, the human Sin3-Rpd3 complex includes the histone deacetylases HDAC1 and HDAC2, the histone-binding proteins RbAp46 and RbAp48, and two novel proteins designated SAP30 and SAP18 (Sin3-associated protein; ZHANG et al. 1997 , ZHANG et al. 1998 ). A homolog of human SAP30 has been identified in yeast (ZHANG et al. 1998 ). Like SIN3 and RPD3, the yeast SAP30 gene is not essential for cell viability. However, deletion of SAP30 confers a set of phenotypes that are shared among sin3, rpd3, and sap30 mutants; furthermore, Sap30 coimmunoprecipitates with Rpd3 (ZHANG et al. 1998 ). Thus, Sap30 is a novel protein of undefined function, conserved between the yeast and human Sin3-Rpd3 complexes.

It is presently unknown how many proteins affect silencing in yeast. Furthermore, the mechanisms by which these factors mediate silencing are unknown. In this study we have examined the role of the Sin3-Rpd3 complex in silencing at the telomeric, HMR, and rDNA loci. Our results demonstrate that the Sin3-Rpd3 complex plays a general role in silencing. Surprisingly, loss of the Gcn5 histone acetyltransferase exerts the same effect on silencing as loss of the Rpd3 histone deacetylase, yet Rpd3 and Gcn5 exert opposite effects on promoter-dependent, position-independent transcription. We propose a model to account for these results.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Yeast strains and media:
The yeast strains used in this study are listed in Table 1. The YMH strains were derived from strain UCC506 (RENAULD et al. 1993 ), strains CFY559 or CFY559sir2 (FRITZE et al. 1997 ), and strain yLP91 (PEMBERTON and BLOBEL 1997 ) by one-step disruption (ROTHSTEIN 1991 ) of the indicated genes. All yeast media were prepared according to standard recipes (BOEKE et al. 1984 ; SHERMAN 1991 ).

Plasmids:
Plasmids used in this study are listed in Table 2. Vectors pRS303 and pRS306 (SIKORSKI and HIETER 1989 ) and YCplac33, YEplac112, YIplac128, and YIplac204 (GIETZ and SUGINO 1988 ) are described elsewhere. The rpd3::URA3-disruption construct, pM1061, was generated by transferring the SalI-SstI fragment of M1436 (rpd3::LEU2) to the same sites of pRS306. YEplac112-RPD3 includes the entire RPD3 open reading frame inserted between the ADH1 promoter and CYC8 terminator in YEplac112. YEplac112-rpd3 (H188A) is identical to YEplac112-RPD3, except that it encodes a form of Rpd3 lacking detectable histone deacetylase activity in vitro (KADOSH and STRUHL 1998 ). pM1288 and pM1289 are identical to YEplac112-RPD3 and YEplac112-rpd3 (H188A), respectively, except that the vector is YCplac33. pM1176 was constructed by PCR amplification of the XbaI-NotI N-terminal fragment of SAP30 (nucleotides 1–590) and ligation into SpeI-NotI sites of pRS303. The sap30::LEU2 -disruption construct, pM1177, was constructed by ligation of the PCR-amplified PstI-SalI C-terminal fragment of SAP30 (nucleotides 385–1370) and the BamHI-SstI fragment (nucleotides 1–590) from pM1176 into PstI-SalI and BamHI-SstI sites of YIplac128, respectively. The sap30::TRP1 construct, pM1183, was generated by transferring the SphI-SstI fragment of pM1177 (sap30::LEU2) to the same sites of YIplac204.

Assays for telomeric, HMR, and rDNA silencing:
Telomeric silencing was scored as described previously (APARICIO et al. 1991 ). Tenfold serial dilutions of overnight cultures of UCC506 derivatives, containing the URA3 gene integrated at the right-end telomere of chromosome V (URA3-TEL-V-R), were spotted onto 5-fluoroorotic acid (5-FOA) and synthetic complete media and incubated at 30° for 3 days. Silencing at the rDNA and HMR loci was scored in the same manner, except that strains containing an ADE2 reporter at the rDNA or HMR loci were spotted onto synthetic complete and -Ade media to monitor the expression of ADE2.

The rpd3 rad6 and rpd3 gcn5 double mutants display synthetic slow-growth phenotypes (data not shown). Therefore, silencing at the telomeric URA3 gene (URA3-TEL-V-R) in these strains was scored by measuring cell viability on 5-FOA medium as described previously (GOTTSCHLING et al. 1990 ). Cells from overnight cultures were serially diluted and plated onto synthetic complete and 5-FOA media. After 3–4 days of incubation at 30°, the numbers of colonies on each plate were counted. The fraction of 5-FOA-resistant cells in a population was determined from at least three independent experiments and is expressed as the average ratio of colonies formed on 5-FOA medium to those formed on synthetic complete medium. Quantification of rDNA silencing was performed in the same way, except that the fraction of Ade⁺ cells is expressed as the average ratio of colonies formed on -Ade medium to those formed on synthetic complete medium.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Deletion of SAP30 enhances silencing at the HMR locus:
To determine whether the Sap30 component of the Sin3-Rpd3 complex plays a general role in silencing, we asked if deletion of SAP30 affects silencing at the HMR locus. Strain yLP19, which contains the ADE2 gene integrated at the hmrA locus (hmrA::ADE2; PEMBERTON and BLOBEL 1997 ), and isogenic rpd3 (YMH348), sin3 (YMH345), and sap30 (YMH349) strains, were used in this analysis. Expression of ADE2 allows cell growth on medium lacking adenine (-Ade) and results in a white colony phenotype, whereas enhanced silencing impairs cell growth on -Ade medium and confers a pink or red colony phenotype due to accumulation of a red pigment. Tenfold serial dilutions of each strain were spotted onto -Ade and synthetic complete media (+Ade) and incubated at 30° for 3 days. The sap30 deletion clearly impaired cell growth on -Ade medium, albeit to a lesser extent than either the rpd3 or sin3 deletions (Figure 1A). The sap30 mutant, similar to the rpd3 and sin3 mutants, also formed pink colonies on YPD medium, compared to white colonies for the wild-type strain (data not shown). We conclude that SAP30 counteracts HMR silencing in a manner similar to RPD3 and SIN3.

View larger version (109K): In this window In a new window Download PPT slide   Figure 1. Deletion of components of the Sin3-Rpd3 complex enhance silencing at the HMR and rDNA loci. Tenfold serial dilutions of the indicated strains were spotted onto -Ade and synthetic complete (+Ade) medium, followed by incubation at 30° for 3 days. Impaired growth relative to the wild-type strains is indicative of enhanced silencing. (A) The wild-type strain (yLP19) carries the ADE2 gene integrated at the hmrA locus; the rpd3 (YMH348), sin3 (YMH345), and sap30 (YMH349) strains are isogenic derivatives of yLP19. (B) The wild-type strain (CFY559) carries the ADE2 (and CAN1) genes integrated at the rDNA (RDN1) locus; the rpd3 (YMH335), sin3 (YMH333), and sap30 (YMH337) strains are isogenic derivatives of CFY559. (C) The wild-type strain (YMH171) carries the ADE2 gene at its normal chromosomal locus; the rpd3 (YMH270), sin3 (YMH265), and sap30 (YMH277) strains are isogenic derivatives of YMH171.

The Sin3-Rpd3 complex affects rDNA silencing:
To determine whether the Sin3-Rpd3 complex plays a general role in silencing, we examined the effects of sin3, rpd3, and sap30 deletions on rDNA silencing. Strain CFY559 and isogenic sin3 (YMH333), rpd3 (YMH335), and sap30 (YMH337) deletion mutants were used in this study. CFY559 is an ade2 can1 mutant carrying an ADE2-CAN1 double marker integrated at the rDNA array (FRITZE et al. 1997 ). Expression of the ADE2 marker within the rDNA array was scored by plating efficiency on -Ade medium and by the colony color phenotype. Compared to the wild-type strain, an ~10⁴-fold decrease in colony formation on -Ade medium was observed for the sin3 and rpd3 mutants, and a 10²- to 10³-fold decrease for the sap30 mutant (Figure 1B). The diminished ADE2 expression associated with the rpd3, sin3, and sap30 mutations cannot be attributed to loss of the ADE2-CAN1 marker by recombination between the rDNA repeats because these mutants exhibited a uniform pink colony phenotype rather than the red phenotype associated with deletion of the ADE2-CAN1 marker by recombination. Furthermore, transformation of pink rpd3 mutants with plasmid-borne RPD3 rescued the white colony phenotype, an effect that would not occur if the ADE2-CAN1 were deleted (data not shown). Also, the rpd3, sin3, and sap30 mutations do not cause an Ade^- phenotype when ADE2 is expressed from its normal chromosomal locus (Figure 1C), demonstrating that the Ade^- phenotypes associated with these mutations are specific for ADE2 expression from the rDNA locus. Taken together, these results establish that the Sin3-Rpd3 complex also affects silencing at the rDNA locus.

Enhanced silencing associated with rpd3 is SIR dependent:
To determine if the enhanced rDNA silencing associated with loss of components of the Sin3-Rpd3 complex is SIR protein dependent, the SIR2 and SIR4 genes were individually deleted in wild-type (CFY559) and isogenic rpd3 strains containing ADE2-CAN1 integrated at the rDNA array. Silencing at the rDNA locus was again scored by the efficiency of colony formation on -Ade medium and by colony color. Results are shown in Figure 2A. Whereas the rpd3 mutation dramatically increased silencing, the rpd3 sir2 double mutation restored growth to ~83% of the wild-type strain (cf. rows 1–3). In addition, the rpd3 sir2 mutant exhibited a white colony phenotype, compared to the pink phenotype of the rpd3 single mutant, and this phenotype can be rescued by plasmid-borne RPD3 (data not shown). However, comparison of the rpd3 sir2 double mutant with the sir2 single mutant revealed increased silencing associated with rpd3 in the sir2 background (cf. rows 3 and 5). These results demonstrate that enhanced silencing associated with rpd3 at the rDNA array is SIR2 dependent, but that sir2 is not completely epistatic to rpd3.

View larger version (123K): In this window In a new window Download PPT slide   Figure 2. Enhanced silencing by rpd3 at the rDNA and telomeric loci is SIR protein dependent. (A) Isogenic wild-type (CFY559), rpd3 (YMH335), rpd3 sir2 (YMH377), rpd3 sir4 (YMH381), sir2 (CFY559sir2), and sir4 (YMH379) strains with ADE2-CAN1 integrated at the rDNA locus were spotted onto -Ade and synthetic complete (+Ade) medium, followed by incubation at 30° for 3 days. Impaired growth on -Ade medium indicates enhanced silencing of ADE2; enhanced growth relative to the rpd3 mutant indicates loss of silencing. (B) Isogenic wild-type (UCC506), rpd3 (YMH320), rpd3 sir2 (YMH360), rpd3 sir4 (YMH364), sir2 (YMH358), and sir4 (YMH362) strains carrying the URA3 gene integrated 2 kb from the right telomere of chromosome V (TEL-V-R) were spotted onto 5-FOA (+FOA) and synthetic complete (-FOA) medium, followed by incubation at 30° for 3 days. Enhanced growth on 5-FOA medium indicates enhanced silencing of URA3; impaired growth indicates loss of silencing. (C) The same set of strains from B were spotted onto -Ura and synthetic complete (+Ura) medium, followed by incubation at 30° for 3 days. Impaired growth on -Ura medium indicates enhanced silencing of URA3; enhanced growth indicates loss of silencing.

The sir4 deletion, however, did not counteract the increase in silencing associated with rpd3 (Figure 2A; cf. rows 2 and 4). Also, the rpd3 single mutant and rpd3 sir4 double mutants displayed comparable pink colony phenotypes (data not shown). Consistent with previous results (FRITZE et al. 1997 ; SMITH and BOEKE 1997 ), deletion of SIR4 in an RPD3 wild-type background enhanced silencing, resulting in an ~10-fold decrease in colony formation on -Ade medium (Figure 2A; cf. rows 1 and 6). Therefore, enhanced rDNA silencing associated with the rpd3 deletion is Sir2 dependent, but Sir4 independent.

In contrast to the requirements for silencing at the rDNA array, both SIR2 and SIR4 are essential for silencing at the telomeric and HM loci (reviewed in SHERMAN and PILLUS 1997 ). We therefore asked whether the enhanced telomeric silencing associated with deletion of RPD3 is Sir2 and Sir4 dependent. Strain UCC506, which contains the URA3 gene positioned 2.0 kb (2+) from the right-end telomere of chromosome V (URA3-TEL-V-R; RENAULD et al. 1993 ), and a set of isogenic rpd3, sir2, and sir4 derivatives, were used in these experiments. The levels of URA3 silencing were monitored by cell growth on medium containing 5-FOA, which is toxic to cells expressing URA3 (BOEKE et al. 1984 ). The wild-type URA3-TEL-V-R strain grew poorly on 5-FOA medium, whereas the isogenic rpd3 mutant grew well (Figure 2B; cf. rows 1 and 2). However, no 5-FOA-resistant colonies were observed for the rpd3 sir2 and rpd3 sir4 double mutants or for the sir2 and sir4 single mutants (Figure 2B, rows 3–6), indicating that silencing of the telomeric URA3 gene was disrupted in these strains. We also did the reciprocal assay, scoring growth of the same strains on -Ura medium to determine whether rpd3 might increase TEL silencing in the sir background to an extent that might not be apparent in the FOA assay. No significant growth difference between the rpd3 sir2 or rpd3 sir4 double mutants and the sir2 or sir4 single mutants was observed (Figure 2C). [Growth of the double mutants is slightly impaired relative to the single mutants, but this difference can be accounted for by the weak slow-growth phenotype associated with rpd3, which is reflected in the +Ura control (Figure 2C)]. Thus, in contrast to the SIR4 independence of rDNA silencing, both SIR2 and SIR4 are required for the enhanced telomeric silencing associated with rpd3.

The Rpd3 effect on silencing is dependent upon histone deacetylase activity:
A histone deacetylase motif, containing evolutionarily invariant histidine residues at positions 150, 151, and 188 (H150, H151, and H188), was recently identified in the Rpd3 protein (KADOSH and STRUHL 1998 ). Amino acid replacements of any of these conserved histidine residues abolished enzymatic activity in vitro and weakened transcriptional repression of targeted genes in vivo. These replacements did not affect either Rpd3 stability or Sin3-Rpd3 interaction.

To determine if the enzymatic activity of Rpd3 is required to counteract silencing, plasmid-borne RPD3 and rpd3 (H188A) alleles were introduced into the rpd3 deletion mutants YMH335 and YMH348, which carry the ADE2 marker at rDNA and HMR loci, respectively. Whereas RPD3 rescued the growth defect of strain YMH335 on -Ade medium (Figure 3A, row 3), the rpd3-H188A strain remained Ade^- (Figure 3A, row 4). This result demonstrates that RPD3, but not rpd3-H188A, restores the expression of the ADE2 gene integrated at the rDNA locus in the rpd3 mutant. Consistent with this result, strains containing RPD3 or rpd3-H188A formed white and pink colonies, respectively, on YPD medium (data not shown). Similar results were obtained for the strains carrying an ADE2 reporter inserted in the HMR locus (Figure 3B, cf. rows 3 and 4; and data not shown). We also note that plasmid-borne expression of RPD3 results in better growth on selective medium in the rpd3 background than does chromosomally expressed RPD3 for both the RDN1::ADE2 and hmr::ADE2 strains (Figure 3A and Figure B, rows 3 vs. 1), suggesting that overexpression of RPD3 might weaken silencing. However, overexpression of RPD3, SIN3, or SAP30 in wild-type backgrounds did not weaken silencing at either rDNA or HMR (data not shown).

View larger version (61K): In this window In a new window Download PPT slide   Figure 3. The RPD3-encoded histone deacetylase activity is required for disruption of silencing at the rDNA, HMR, and TEL-V-R loci. In A–C the RPD3 cellular genotype is indicated before the slash and the plasmid-borne genotype after the slash. In A and B, disruption of silencing is scored as enhanced growth on -Trp -Ade medium, whereas in C disruption of silencing is scored as diminished growth on -Trp +FOA medium. (A) Strain CFY559 (RDN1::ADE2 RPD3) or isogenic strain YMH335 (RDN1::ADE2 rpd3) was transformed with vector alone (YCplac33) or its derivatives carrying either RPD3 or the rpd3-H188A allele, which encodes catalytically inactive Rpd3. Tenfold serial dilutions of the resulting Ura+ transformants were spotted onto -Ura -Ade or -Ura medium and incubated for 3 days at 30°. (B) Strain yLP19 (hmrA::ADE2 RPD3) or isogenic strain YMH348 (hmrA::ADE2 rpd3) was transformed with vector alone (YEplac112) or its derivatives carrying either RPD3 or rpd3-H188A. The resulting Trp+ strains were spotted onto -Trp -Ade or -Trp medium and incubated for 3 days at 30°. (C) Strain UCC506 (URA3-TEL-V-R RPD3) or isogenic strain YMH272 (URA3-TEL-V-R rpd3) was transformed with the same plasmids defined in B. The resulting Trp+ strains were spotted onto -Trp +FOA or -Trp medium and incubated for 3 days at 30°.

We also examined the requirement for Rpd3 activity in regulating telomeric silencing. In this case strain YMH272 (URA3-TEL-V-R rpd3) was used as the host and silencing was scored as enhanced growth (diminished URA3 expression) on 5-FOA medium. The host rpd3 strain is 5-FOA-resistant due to enhanced silencing of the URA3 marker (Figure 3C, row 2). This phenotype is rescued by plasmid-borne RPD3, resulting in 5-FOA sensitivity (row 3), but not by the rpd3-H188A plasmid (row 4). Taken together, the results in Figure 3 clearly demonstrate that the enzymatic activity of Rpd3 is required to counteract silencing at telomeric, HMR, and rDNA loci.

Enhanced silencing associated with rpd3 is RAD6 dependent:
Several studies have implicated Rad6-mediated ubiquitination as a regulator of silencing in both S. cerevisiae and Schizosaccharomyces pombe (BRYK et al. 1997 ; HUANG et al. 1997 ; SINGH et al. 1998 ). To investigate the possible relationship between ubiquitination and deacetylation in regulating silencing, we tested whether the enhanced silencing associated with loss of Rpd3 activity can bypass the requirement for Rad6. Due to the synthetic slow growth defect associated with rpd3 rad6 double mutants (data not shown), we assayed telomeric silencing at URA3-TEL-V-R by measuring viability of isogenic wild-type (UCC506), rpd3 (YMH320), rad6 (YMH405), and rpd3 rad6 (YMH407) strains on 5-FOA medium, rather than by the spotting assays described above. Results are presented in Table 3. As expected, rpd3 enhanced silencing, resulting in a 9000-fold increase in cell viability on 5-FOA medium, whereas rad6 weakened silencing, causing a 17-fold decrease in cell viability. Strikingly, the rpd3 rad6 double deletion further weakened silencing, resulting in a 1.4 x 10⁶-fold decrease in cell viability relative to the rpd3 single mutant. Similar results were observed for sin3 rad6 and sap30 rad6 mutants (data not shown). Thus, the enhanced telomeric silencing associated with loss of the Sin3-Rpd3 complex is Rad6 dependent.

View this table: In this window In a new window   Table 3. Effects of rpd3, rad6, and gcn5 on telomeric silencing

Because transcriptional silencing at the rDNA locus is mediated by a novel mechanism that depends on only a single SIR gene, SIR2 (BRYK et al. 1997 ; FRITZE et al. 1997 ; SMITH and BOEKE 1997 ), and deletion of SIR4 increases rDNA silencing (FRITZE et al. 1997 ; SMITH and BOEKE 1997 ; Figure 2A, row 6), we asked if rad6 exerts an effect on rDNA silencing similar to its effect on telomeric silencing. Deletion mutants comparable to those described above were generated using strain CFY559 (RDN1-ADE2-CAN1) and cell viability was scored on -Ade medium. Results are presented in Table 4. Again, rpd3 (YMH335) enhanced silencing, in this case resulting in a 1500-fold decrease in cell viability. In contrast, rad6 conferred a negligible effect on its own (YMH413), yet fully suppressed the effect of rpd3 in the rpd3 rad6 double mutant (YMH415), resulting in an 1100-fold increase in cell viability. In addition, wild-type (CFY559), rad6 (YMH413), and rpd3 rad6 (YMH415) strains formed white colonies on YPD medium, indicating efficient ADE2 expression, whereas the rpd3 strain was pink (data not shown). Thus, enhanced silencing at the rDNA locus is also Rad6 dependent.

View this table: In this window In a new window   Table 4. Effects of rpd3 and rad6 on silencing at the rDNA (RDN1) locus

Enhanced silencing associated with loss of Sin3-Rpd3 occurs in the absence of CAC3:
Components of the CAF-I complex are required for silencing at the HM and telomeric loci (ENOMOTO et al. 1997 ; KAUFMAN et al. 1997 ; MONSON et al. 1997 ; ENOMOTO and BERMAN 1998 ). To determine if components of the Sin3-Rpd3 complex interact with Cac3 to regulate silencing, we tested the ability of sin3, rpd3, and sap30 deletions to restore telomeric silencing in a cac3 strain. Silencing at the URA3-TEL-V-R locus was assayed by scoring cell growth on 5-FOA medium, as described above, using an isogenic set of strains with different combinations of cac3, rpd3, sin3, and sap30 deletions. The results are shown in Figure 4. As expected, cac3 weakened silencing, scored as enhanced 5-FOA sensitivity (row 2), whereas sap30 enhanced silencing (row 4). However, the double mutants (row 3) exhibit intermediate phenotypes corresponding to an ~10-fold increase in silencing relative to the wild-type strain (row 1). These effects were the same for deletion of all three components of the Sin3-Rpd3 complex (A, B, and C). Thus, Cac3 and the Rpd3-Sin3 complex exert opposite effects on silencing in a partially offsetting manner. Interestingly, the human counterpart of Cac3, RbAp46, is found as a component of the Sin3-Rpd3 complex. However, there is no evidence that yeast Cac3 is a component of the yeast Sin3-Rpd3 complex.

View larger version (79K): In this window In a new window Download PPT slide   Figure 4. CAC3 (MSI1) is dispensable for maintenance of telomeric silencing in the absence of an intact Sin3-Rpd3 complex. The effects of sap30 (A), rpd3 (B), and sin3 (C) on silencing at the URA3-TEL-V-R locus are shown. All strains are isogenic derivatives of strain UCC506 and were spotted as 10-fold serial dilutions onto medium either containing (+FOA) or lacking (-FOA) 5-FOA. Impaired growth relative to the wild-type control indicates disruption of telomeric URA3 silencing, whereas enhanced growth indicates enhanced silencing.

Deletion of GCN5 and RPD3 exerts similar effects on silencing:
Whereas RPD3 encodes a histone deacetylase that is required for transcriptional repression of targeted genes (KADOSH and STRUHL 1998 ; RUNDLETT et al. 1998 ), GCN5 encodes a histone acetyltransferase required for activation of targeted genes (KUO et al. 1998 ; WANG et al. 1998 ). These activities suggest that Rpd3 and Gcn5 exert opposite effects on transcriptional control of genes targeted by both factors. Indeed, defective activation of the HO gene by deletion of GCN5 can be suppressed by deletion of RPD3 (PEREZ-MARTIN and JOHNSON 1998 ).

These results suggested that gcn5, in contrast to rpd3, might weaken silencing. We tested this possibility by deleting GCN5 in the URA3-TEL-V-R reporter strain UCC506. Surprisingly, gcn5 dramatically enhanced silencing, resulting in a 1500-fold increase in cell viability on FOA medium (Table 3; cf. UCC506 and YMH366). Furthermore, the rpd3 gcn5 double mutation (YMH370) did not increase cell viability beyond the effect of rpd3 alone (YMH320; 9000-fold). A similar effect on silencing was observed at the hmrA::ADE2 locus, where gcn5 resulted in formation of pink colonies, yet the isogenic wild-type strain remained white (data not shown). These effects are not a consequence of position-independent effects on URA3 and ADE2 expression, because neither rpd3 nor gcn5 mutations confer uracil auxotrophy, FOA resistance, or the pink colony phenotype associated with impaired URA3 or ADE2 expression in an otherwise normal strain. Thus, Gcn5 histone acetyltransferase, like Rpd3 histone deacetylase, counteracts silencing.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The Sin3-Rpd3 complex plays a general role in silencing:
A role for Sin3 and Rpd3 in silencing at telomeric and HM cryptic mating loci has been shown previously (DE RUBERTIS et al. 1996 ; RUNDLETT et al. 1996 ; VANNIER et al. 1996 ). The results presented here confirm and extend those results by establishing that disruption of silencing at the rDNA array also requires Sin3 and Rpd3, as well as Sap30, a recently defined subunit of the Sin3-Rpd3 complex. Recently, the RPD3, SIN3, and SAP30 genes were also identified as IRS genes in a genetic screen for mutations that increase rDNA silencing (SMITH et al. 1999 ). Furthermore, disruption of silencing at telomeric, HMR, and rDNA loci is dependent upon the enzymatic activity of Rpd3. These results establish a general requirement for the Rpd3 histone deacetylase in epigenetic control of gene expression.

Relationship of Rpd3 to Rad6:
Recent studies demonstrated that deletion of RAD6 counteracts silencing at telomeric, HM, and rDNA loci in S. cerevisiae, and at the silent mating loci in S. pombe (BRYK et al. 1997 ; HUANG et al. 1997 ; SINGH et al. 1998 ). These effects are dependent upon the ubiquitin-conjugating activity of Rad6, but not its N-end rule protein-degrading activity. Conversely, the Ubp3 and (putative) Dot4 ubiquitin hydrolases counteract silencing (MOAZED and JOHNSON 1996 ; SINGER et al. 1998 ). Several models have been proposed to account for these results (HUANG et al. 1997 ). One suggests that repression is dependent upon ubiquitination of silencing regulators. Consistent with this idea, histones H2A, H2B, and H3 are ubiquitinated by Rad6 in vitro (SUNG et al. 1988 ; HAAS et al. 1990 ) and Ubp3 physically interacts with Sir4 in vivo (MOAZED and JOHNSON 1996 ). Ubiquitination has also been linked to silencing in Drosophila (HENCHOZ et al. 1996 ), and inactivation of a ubiquitin-conjugating enzyme has been associated with chromatin defects in mice (ROEST et al. 1996 ).

A striking result presented here is that rad6 is epistatic to the effect of rpd3 on silencing at the telomeric and rDNA loci (Table 3 and Table 4). One possible explanation for this result is that the Rpd3 histone acetyltransferase regulates expression of RAD6, which in turn is required for silencing. However, Western blot analysis showed that Rad6 protein levels are essentially unchanged in isogenic wild-type, rpd3, and gcn5 strains (data not shown). Therefore, neither Rpd3 nor Gcn5 affects silencing indirectly through RAD6. The more direct effect of Rad6 on silencing is consistent with the possibility that Rad6 affects silencing by ubiquitination of direct effectors of silencing.

Effects of Rpd3 and Gcn5 on silencing:
The Sin3-Rpd3 complex facilitates transcriptional repression as a consequence of targeted recruitment by DNA-binding transcriptional repressors (KADOSH and STRUHL 1997 ; RUNDLETT et al. 1998 ). Yet the Sin3-Rpd3 complex exerts the opposite effect on silenced loci, enhancing silencing in the absence of Sin3, Rpd3, and Sap30. How does the Sin3-Rpd3 complex counteract silencing, yet repress transcription at promoter-specific targets?

A distinct possibility is that the effect of the Sin3-Rpd3 complex is indirect. For example, Sin3-Rpd3 might repress expression of genes generally required for silencing such that in the absence of Sin3-Rpd3 these genes are derepressed, leading to increased silencing. This scenario seems likely given the general role of Rpd3 in transcriptional repression. Nonetheless, the effect of rpd3 on silencing cannot be accounted for by increased expression of either RAD6 (above) or SIR genes. Overexpression of SIR4 does not enhance silencing, but instead weakens silencing at HMR, telomeric, and rDNA loci (SUSSEL and SHORE 1991 ; RENAULD et al. 1993 ; SMITH et al. 1998 ). Also, the negative effect of SIR4 on HM silencing can be compensated by co-overexpression of SIR3 (MARSHALL et al. 1987 ). Sir2 is the only SIR protein required for silencing at all three loci. However, increased SIR2 expression slightly weakens silencing at the HM loci (M. COCKELL and S. M. GASSER, personal communication), yet enhances rDNA silencing (FRITZE et al. 1997 ; SMITH et al. 1998 ; Z.-W. SUN, unpublished results). Thus, enhanced silencing associated with loss of Sin3-Rpd3 function cannot be accounted for by overexpression of SIR genes.

A notable and unexpected result presented here is that deletion of GCN5 enhances silencing (Table 3). Accordingly, loss of Gcn5 histone acetyltransferase activity has the same effect on silencing as loss of Rpd3 histone deacetylase activity. This result was surprising because GCN5 and RPD3 exert opposite effects on transcriptional control of genes targeted by both factors. Indeed, defective activation associated with deletion of GCN5 can be suppressed by deletion of RPD3 (PEREZ-MARTIN and JOHNSON 1998 ).

The similar effects of rpd3 and gcn5 on silencing might be an important clue toward understanding how Rpd3 and Gcn5 influence silencing. Even though rpd3 and gcn5 affect the expression of a broad range of genes, rpd3 generally enhances transcription, whereas gcn5 impairs transcription. A stimulatory effect of gcn5 on expression of silencing factors would be opposite to its effect on most genes. This possibility seems especially unlikely if the same factors are also affected by rpd3. An alternative possibility, described below, is that Rpd3 and Gcn5 affect silencing directly by generating the histone acetylation pattern specific to silent chromatin.

A model for the role of Sin3-Rpd3 in silencing:
A substantial body of evidence indicates that silencing is a consequence of modified chromatin structure (KAYNE et al. 1988 ; MEGEE et al. 1990 ; PARK and SZOSTAK 1990 ; GOTTSCHLING 1992 ; SINGH and KLAR 1992 ; CHEN-CLELAND et al. 1993 ). Analysis of the patterns of histone acetylation at the HM loci revealed that histones H3 and H4 are hypoacetylated relative to their counterparts in transcriptionally active regions of the genome (BRAUNSTEIN et al. 1993 ). Moreover, the acetylation pattern of H4—hypoacetylation of lysines at positions 5, 8, and 16 (K5, K8, and K16) and hyperacetylation of lysine at position 12 (K12)—is identical to the H4 acetylation pattern in Drosophila heterochromatin (TURNER et al. 1992 ; BRAUNSTEIN et al. 1996 ). These results underscore the importance of acetylation and deacetylation of specific histone residues in regulating silencing.

Perhaps Rpd3 and Gcn5 affect silencing by catalyzing formation of the histone acetylation pattern resident in silent chromatin. Newly synthesized histone H4 is acetylated at K5 and K12, which are conserved modifications among humans, Drosophila, and Tetrahymena, and this pattern is thought to be important for assembly of H4 onto replicating DNA (ALLIS et al. 1985 ; SOBEL et al. 1994 , SOBEL et al. 1995 ). The CAF-I chromatin assembly factor deposits newly synthesized histones H3 and H4, but not those from bulk chromatin, onto DNA (SMITH and STILLMAN 1991 ). Indeed, human CAF-I exhibits substrate specificity for H4 acetylated at K5, K8, and/or K12 (VERREAULT et al. 1996 ). A similar chromatin assembly complex has also been identified in yeast (KAUFMAN et al. 1997 ). Furthermore, acetylation of one or more K5, K8, and K12 residues of H4 provides the recognition signal for chromatin assembly (MA et al. 1998 ), suggesting that the substrate specificities of the human and yeast CAF-I complexes are similar. Mutations in the CAC1, CAC2, and CAC3 genes, which encode yeast CAF-I (KAUFMAN et al. 1997 ), decrease telomeric, HM, and rDNA silencing (ENOMOTO et al. 1997 ; KAUFMAN et al. 1997 ; MONSON et al. 1997 ; ENOMOTO and BERMAN 1998 ; SMITH et al. 1999 ). These observations led to the proposal that CAF-I provides the substrate specificity to ensure that nucleosomes are assembled from appropriately acetylated histones (MONSON et al. 1997 ; ENOMOTO and BERMAN 1998 ).

The following model is proposed to explain the role of Rpd3 in silencing (Figure 5). Accordingly, Rpd3 would play a direct role in silencing by affecting the relative levels of the "inactive" (heterochromatin) and "active" (euchromatin) forms of histone H4. This ratio would affect the efficiency of formation of silent chromatin in much the same way that components of silent chromatin and a transcriptional activator compete to establish either the silent or active state of gene expression at telomeres following the disassembly of silent chromatin during DNA replication. This effect was proposed to account for the random nature of phenotypic switching in variegated gene expression (APARICIO and GOTTSCHLING 1994 ). Our model is consistent with the recent identification of multiple genes associated with DNA replication and chromatin modification in a genetic screen for rDNA silencing defects (SMITH et al. 1999 ). In the sin3, rpd3, or sap30 strains, the relative levels of H4 acetylated at K5 and K12 would increase due to loss of histone deacetylase activity. H4 acetylated at K12 is the inactive form, thereby accounting for enhanced silencing associated with loss of Sin3-Rpd3 function. This scenario is dependent upon substrate specificity of Rpd3 for H4 K12. Indeed, rpd3 enhances acetylation of H4 residues K5 and K12 (RUNDLETT et al. 1996 ).

View larger version (13K): In this window In a new window Download PPT slide   Figure 5. A model to account for the general effects of rpd3, gcn5, and cac3 on silencing. The rpd3 (sin3 and sap30) mutation eliminates Sin3-Rpd3 histone deacetylase (HDA) activity. Because Rpd3 has specificity for K5 and K12 of histone H4 (RUNDLETT et al. 1996 ), rpd3 would increase the level of the "inactive" or "prosilencing" form of H4. The gcn5 mutation eliminates Gcn5 histone acetyltransferase (HAT) activity. Because Gcn5 has specificity for K8 and K16 of H4 (KUO et al. 1996 ), gcn5 would decrease the level of the "active" or "antisilencing" form of H4. Consequently, both rpd3 and gcn5 would enhance silencing by increasing the relative level of the inactive form of H4. The cac3 deletion would inactivate the CAF-I chromatin assembly complex, which has been proposed to ensure that only inactive histones are assembled into silent chromatin and that local SIR protein concentrations are elevated to form a "wall" of silent chromatin (MONSON et al. 1997 ; ENOMOTO and BERMAN 1998 ). Consequently, cac3 would weaken silencing in both the wild-type (++ vs. +) and rpd3 background (+++++ vs. ++++) due to loss of substrate specificity for the inactive form of H4, and to decreased local SIR protein concentrations (denoted by large and small SIR fonts). The asterisk preceding gcn5 denotes that the effect of cac3 has not been tested in a gcn5 background. The black circles denote acetylation at the indicated lysine residues (K) of histone H4. The vertical arrows denote increased levels of the indicated forms of H4. This model is also applicable for histone H3, because H3 can be a substrate for Rpd3 (RUNDLETT et al. 1996 ), Gcn5 (KUO et al. 1996 ), and CAF-I (VERREAULT et al. 1996 ), and is a structural component of silent chromatin (GRUNSTEIN 1998 ). However, the specific acetylation pattern of H3 in silent chromatin has yet to be defined.

This model would also account for the enhanced silencing associated with gcn5 (Table 3). Accordingly, the Gcn5 histone acetyltransferase would directly affect silencing by catalyzing acetylation of H4 residues K8 and K16. Consistent with this premise, an H4 K16Q replacement, which simulates acetylated K16, disrupts the interaction between H4 and Sir3 (HECHT et al. 1995 ). This result led to the proposal that K16 hypoacetylation might be important for H4 interaction with Sir3 in heterochromatin (GRUNSTEIN 1998 ). In the gcn5 strain, the levels of H4 acetylated at K8 and K16 would decrease, thereby increasing the relative levels of the inactive form of H4 acetylated at K5 and K12. Again, this proposal is consistent with the specificity of Gcn5 for H4 residues K8 and K16 (KUO et al. 1996 ). This model is also applicable for histone H3, because H3 can be a substrate for Rpd3 (RUNDLETT et al. 1996 ), Gcn5 (KUO et al. 1996 ), and CAF-I (VERREAULT et al. 1996 ), and is a structural component of silent chromatin (GRUNSTEIN 1998 ). However, the specific acetylation pattern of H3 in silent chromatin has yet to be defined.

To facilitate inheritance of silencing, CAF-I would ensure that only appropriately acetylated inactive histones (both newly synthesized and recycled from the previous cell cycle) are assembled into silent chromatin (MONSON et al. 1997 ; ENOMOTO and BERMAN 1998 ). CAF-I might also exclude histones with the active acetylation pattern from being recycled into silent chromatin. In the case of a derepressed silent locus from the previous cell cycle, this function would be especially relevant (ENOMOTO and BERMAN 1998 ). In the cac mutants, new nucleosomes must be assembled by an alternative pathway (MONSON et al. 1997 ; KAUFMAN et al. 1998 ; QIAN et al. 1998 ). If the alternative assembly complex lacks the substrate specificity of CAF-I, then the increased level of inactive histones associated with the absence of either Rpd3 or Gcn5 would facilitate silent chromatin assembly. This would account for the offsetting effects of cac3 and either rpd3, sin3, or sap30 (Figure 4).

A second function of CAF-I would be to ensure that local Sir2, Sir3, and Sir4 protein concentrations are sufficiently elevated to permit assembly of a strong silencer. This conclusion is based on improved silencing associated with elevated levels of Sir2, Sir3, or Sir4 in cac1 mutants, and on disruption of silencing associated with limiting amounts of Sir2 or Sir3 in an otherwise wild-type background (ENOMOTO and BERMAN 1998 ). Therefore, the decreased local SIR protein concentrations associated with cac3 would partially weaken the enhanced silencing caused by sin3, rpd3, and sap30. This is consistent with the observation that loss of the Sin3-Rpd3 complex does not bypass the SIR protein requirement for maintaining silencing (Figure 2 and VANNIER et al. 1996 ).

A key feature of this model is that the acetylation state of histones affects the efficiency of assembly of silent chromatin. The model does not propose that the acetylation pattern at silent loci would necessarily change upon deletion of RPD3 or GCN5. Indeed, chromatin immunoprecipitation experiments, demonstrating that rpd3 and sin3 alter the acetylation pattern of histone H4 at Ume6-regulated promoters, showed that the H4 acetylation pattern at a telomeric locus is unchanged by rpd3 and sin3 (RUNDLETT et al. 1998 ), despite the dramatic effects of these mutations on telomeric silencing.

	FOOTNOTES

¹ Present address: Department of Biochemistry and Molecular Genetics, University of Virginia, Charlottesville, VA 22908.

	ACKNOWLEDGMENTS

We are grateful to Jeff Smith, Jef Boeke, Susan Gasser, Ken Robzyk, and Mary Ann Osley for communicating results prior to publication. We also thank Mike Christman, Shelley Esposito, Dan Gottschling, and Lucy Pemberton for yeast strains; Kevin Struhl for rpd3 alleles; and Jim Broach, Leonard Guarente, Kiran Madura, David McNabb, Ines Pinto, Louise Prakash, David Stillman, and Fred Winston for plasmids. We also acknowledge David Gross, Yi Zhang, and Danny Reinberg for fruitful discussions and critical review of the manuscript. Research in M.H.'s laboratory is supported by National Institutes of Health grant GM-39484.

Manuscript received January 22, 1999; Accepted for publication April 16, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALLIS, C. D., L. G. CHICOINE, R. RICHMAN, and I. G. SCHULMAN, 1985  Deposition-related histone acetylation in micronuclei of conjugating Tetrahymena. Proc. Natl. Acad. Sci. USA 82:8048-8052[Abstract/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].

BOEKE, J. D., 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].

BRAUNSTEIN, M., A. B. ROSE, S. G. HOLMES, C. D. ALLIS, and J. R. BROACH, 1993  Transcriptional silencing in yeast is associated with reduced nucleosome acetylation. Genes Dev. 7:592-604[Abstract/Free Full Text].

BRAUNSTEIN, M., R. E. SOBEL, C. D. ALLIS, B. M. TURNER, and J. R. BROACH, 1996  Efficient transcriptional silencing in Saccharomyces cerevisiae requires a heterochromatin histone acetylation pattern. Mol. Cell. Biol. 16:4349-4356[Abstract].

BRYK, M., M. BANERJEE, M. MURPHY, K. E. KNUDSEN, and D. J. GARFINKEL et al., 1997  Transcriptional silencing of Ty1 elements in the RDN1 locus of yeast. Genes Dev. 11:255-269[Abstract/Free Full Text].

CHEN-CLELAND, T. A., M. M. SMITH, S. LE, R. STERNGLANZ, and V. G. ALLFREY, 1993  Nucleosome structural changes during derepression of silent mating-type loci in yeast. J. Biol. Chem. 268:1118-1124[Abstract/Free Full Text].

DE RUBERTIS, F., D. KADOSH, S. HENCHOZ, D. PAULI, and G. REUTER et al., 1996  The histone deacetylase RPD3 counteracts genomic silencing in Drosophila and yeast. Nature 384:589-591[Medline].

ENOMOTO, S. and J. BERMAN, 1998  Chromatin assembly factor I contributes to the maintenance, but not the re-establishment, of silencing at the yeast silent mating loci. Genes Dev. 12:219-232[Abstract/Free Full Text].

ENOMOTO, S., P. D. MCCUNE-ZIERATH, M. GERAMI-NEJAD, M. A. S