Genetics, Vol. 155, 509-522, June 2000, Copyright © 2000
Two Classes of sir3 Mutants Enhance the sir1 Mutant Mating Defect and Abolish Telomeric Silencing in Saccharomyces cerevisiae
Elisa M. Stone1,a,
Cheryl Reifsnyderb,
Mitch McVeyb,
Brandy Gazob, and
Lorraine Pillusa,b
a Department of Biology, University of California, San Diego, California 92093-0347
b Department of Molecular, Cellular and Developmental Biology, University of Colorado, Boulder, Colorado 80309-0347
Corresponding author:
Lorraine Pillus, Department of Biology, University of California, 9500 Gilman Dr., San Diego, CA 92093-0347., lpillus{at}ucsd.edu (E-mail)
Communicating editor: F. WINSTON
 | ABSTRACT |
|---|
Silent information regulators, or Sir proteins, play distinct roles in chromatin-mediated transcriptional control at the silent mating-type loci, telomeres, and within the rDNA repeats of Saccharomyces cerevisiae. An unusual collection of sir3 mutant alleles was identified in a genetic screen for enhancers of the sir1 mutant mating-defective phenotype. These sir3-eso mutants, like the sir1 mutant, exhibit little or no mating defects alone, but the sir1 sir3-eso double mutants are essentially nonmating. All of the sir3-eso mutants are defective in telomeric silencing. In some mutants, this phenotype is suppressed by tethering Sir1p to telomeres; other mutants are dominant for mating and telomeric silencing defects. Additionally, several sir3-eso mutants are nonmating in combination with the nat1 N-terminal acetyltransferase mutant. The temperature-sensitive allele sir3-8 has an eso phenotype at permissive temperature, yet acts as a null allele at restrictive temperature due to loss of sir3-8 protein. Sequence analysis showed that eight of the nine sir3-eso alleles have mutations within the N-terminal region that is highly similar to the DNA replication initiation protein Orc1p. Together, these data reveal modular domains for Sir3p and further define its function in silencing chromatin.
TRANSCRIPTIONAL silencing is one means by which cells regulate gene expression. Silencing occurs when chromatin structure is modified at certain regions of chromosomes, inactivating the genes in those regions. Examples of silencing include the inactive mammalian X chromosome, position effect variegation in Drosophila, and the silent mating-type loci in fission and budding yeasts (for reviews, see LAURENSON and RINE 1992 
; WEILER and WAKIMOTO 1995 ; ALLSHIRE 1996 ; LYON 1999 ). In Saccharomyces cerevisiae, at least three genetic loci are subject to transcriptional silencing: the silent mating-type loci, telomeres and the rDNA. Numerous genes play a role in silencing at these loci. SIR1, SIR2, SIR3, and SIR4 were originally identified in mutant strains that inappropriately expressed the silent mating-type genes, generally leading to a mating-defective phenotype (RINE and HERSKOWITZ 1987 ; LAURENSON and RINE 1992 ). SIR2, SIR3, and SIR4 also function in silencing genes positioned near telomeres (APARICIO et al. 1991 ; VEGA-PALAS et al. 1997 ; PRYDE and LOUIS 1999 ). Sir2p, Sir3p, and Sir4p exist in a multiprotein complex that interacts with site-specific DNA binding proteins and with nucleosomes to mediate silencing (MORETTI et al. 1994 ; HECHT et al. 1995 , HECHT et al. 1996 ; MOAZED et al. 1997 ; STRAHL-BOLSINGER et al. 1997 ). Sir proteins are likely to be targeted to silent chromatin by Rap1p, Abf1p, and/or origin recognition complex (ORC) proteins, which bind directly to DNA sites within silencer sequence elements (SHORE et al. 1987 ; BUCHMAN et al. 1988 ; SUSSEL and SHORE 1991 ; BELL and STILLMAN 1992 ; FOSS et al. 1993 ; LIU et al. 1994 ; LOO et al. 1995 ). Chromatin is thought to be silenced through the interaction of Sir proteins with the N-terminal tails of histones H3 and H4 (HECHT et al. 1995 ). Sir2p plays an additional role in chromatin of the nucleolar rDNA array (BRYK et al. 1997 ; SMITH and BOEKE 1997 ). However, the molecular details of how Sir protein complexes achieve silencing remain incompletely defined.
A unique role for Sir1p in the establishment of silencing was demonstrated with the discovery that sir1 mutants exhibit a heritable yet epigenetic mating-defective phenotype (PILLUS and RINE 1989 ). In a population of sir1 mutant cells, two subpopulations exist: one is mating competent and normally silenced as in wild type, the other is mating defective due to derepression of the silent mating-type loci. Although sir1 mutants do not have a defect in the maintenance of silencing, it appears that silencing is not established efficiently in the subpopulation that is transcriptionally derepressed. How Sir1p functions in silencing is unclear, but mechanistic clues come from experiments in which Sir1p is tethered to regions of DNA lacking silencer sequences. When Sir1p is fused to the DNA binding domain of Gal4p, the fusion protein can be targeted to Gal4p-binding sites engineered near reporter genes to result in transcriptional silencing (CHIEN et al. 1993 ; FOX et al. 1997 ). Moreover, Sir1p can be shown to interact physically with the DNA replication initiation subunit Orc1p (TRIOLO and STERNGLANZ 1996 ). A central 17-amino-acid domain of Sir1p appears to direct it to silencers and is required for the interaction with Orc1p (GARDNER et al. 1999 ). Sir1p is not known to participate directly with ORC in DNA replication, however, so the mechanistic significance of the Sir1p-Orc1p interaction is not yet understood.
Sir3p is a key component of silent chromatin (reviewed in STONE and PILLUS 1998 ). It is an integral subunit of the multiprotein complex that functions at the silent mating-type loci and at telomeres. The sir3 null mutant is nonmating and defective in telomeric silencing (RINE and HERSKOWITZ 1987 ; APARICIO et al. 1991 ). Indeed, Sir3p is a limiting factor in telomeric silent chromatin (RENAULD et al. 1993 ) and when tethered to DNA appears to recruit other proteins to achieve silencing (LUSTIG et al. 1996 ). Several sir3 mutants were previously identified that suppress silencing defects of mutants in the histone H4 N terminus or the Rap1p C terminus (JOHNSON et al. 1990 ; LIU and LUSTIG 1996 ), providing genetic evidence for Sir3p-histone and Sir3p-Rap1p interactions. A recent study revealed that an N-terminal fragment consisting of approximately half of the Sir3p protein (GOTTA et al. 1998 ) is sufficient for enhanced telomeric silencing previously seen with SIR3 overexpression (RENAULD et al. 1993 ). Additionally, three broad domains were identified to have different properties in nucleating telomeric silencing by assaying the ability of tethered Sir3p fusion proteins to silence in conjunction with a rap1 telomere-defective mutant (PARK et al. 1998 ). When an N-terminal region and a C-terminal region of Sir3p are expressed simultaneously, partial complementation of the sir3 null mutant mating defect is observed, suggesting that the two halves can function independently (LE et al. 1997 ; GOTTA et al. 1998 ). From these studies and the work described in this article, a picture of distinct functional domains for Sir3p emerges.
In a genetic screen for enhancers of the sir one mutant mating defect (REIFSNYDER et al. 1996 ), we uncovered a collection of mutants including those termed the sir3-eso mutants to emphasize their distinct phenotypes described here. A genetic interaction between SIR1 and SIR3 had previously been noted, in that overexpression of SIR1 can suppress the mating defect associated with certain sir3 alleles (STONE et al. 1991 ). The sir3-eso mutants provide additional evidence for specific SIR1-SIR3 genetic interactions. Analysis of sir3-eso mutations revealed that the N-terminal domain of Sir3p is critical for silencing the HM silent mating-type loci and telomeres in the absence of SIR1. From sequence and phenotypic classification, the N-terminal region of Sir3p that is highly similar to the DNA replication initiation protein Orc1p is highlighted, suggesting a functional link between Sir1p, Sir3p, and Orc1p.
| MATERIALS AND METHODS |
|---|
Yeast strains, growth conditions, and transformation:
Genotypes of yeast strains used in this study are listed in Table 1. Yeast extract/peptone/dextrose (YPD) rich medium, supplemented synthetic medium lacking the appropriate nutrient for plasmid selection, and minimal medium were prepared as described (SHERMAN 1991 ). 5-Fluoroorotic acid (5-FOA) plates were prepared by adding 5-FOA to a final concentration of 0.1% (SIKORSKI and BOEKE 1991 ) to supplemented synthetic medium. Transformations into various yeast strains were performed with lithium acetate as described (SCHIESTL and GIETZ 1989 ). pLP1202 was used to delete HML in LPY4441. In JRY4603 and JRY4623, the SIR3 or SIR1 open reading frames, respectively, were deleted by standard methods (BAUDIN et al. 1993 ). Crosses were performed with descendants of AMR7 and RS862 to make LPY1132; LPY4441, JRY4603, and EY957 to make LPY2709; AMR27 and YDS631 to make LPY3237; RS862 and YDS631 to make LPY3238 and LPY3620; JRY4623 and LPY3620 to make LPY3320 and LPY3321; and RS862 and YDS634 to make LPY4417.
eso mutant screen:
A total of 259,000 colonies from 35 independent cultures of AMR27 transformed with SIR1 plasmid pJR910 (also known as strain LPY94) and 80,000 colonies from 8 independent cultures of JRY3010 with pJR910 (also known as LPY122) were mutagenized and plated on supplemented synthetic medium (REIFSNYDER et al. 1996 ). Resulting colonies were then screened at 30° for mutants that mated when the SIR1 plasmid was present but that did not mate without the plasmid. Genetic linkage analysis and plasmid complementation tests were performed to determine if the eso mutants were in previously identified silencing genes SIR2, SIR3, SIR4, NAT1, and ARD1 (REIFSNYDER 1996 ). Twenty-nine mutants in six complementation groups were uncovered. These included 1 allele of sas2 (REIFSNYDER et al. 1996 ), 5 alleles of sir2 (S. GARCIA and L. PILLUS, unpublished results), 13 alleles of sir3, 3 alleles of sir4, 2 alleles of ard1, and 5 alleles of nat1. Eight of the sir3-eso alleles were rescued by gap repair as described below. As preliminary mating analysis did not distinguish novel phenotypes of the 5 alleles that were not rescued (data not shown), we continued detailed analysis for only those 8 that were. The sir3-eso allele designations in the original mutant strains prior to gap repair and sequencing are as follows: LPY221, 1.6.o; LPY222, 2.9.a; LPY225, 3.25.a; LPY238, 6.1.b; LPY275, 10.16.a; LPY521, 2o; LPY669, H9b; LPY683, 3.i.j; and JRY188, sir3-8.
Plasmids:
pJR910 (also known as pLP17) contains SIR1 on a CEN-URA3 plasmid. pLP27 (STONE and PILLUS 1996 ) was used to complement sir3 mutants in above crosses where appropriate. pJR273 contains SIR3 as a 4.5-kb SalI fragment in the CEN-URA3 plasmid pSEYC58 (EMR et al. 1986 ). pRS313 and pRS315 are CEN-based low-copy plasmids with HIS3 or LEU2 markers, respectively (SIKORSKI and HEITER 1989). YEp351 is a LEU2, 2µ plasmid (HILL et al. 1986 ). pLP143 was constructed by inserting the SalI HindIII 5' SIR3 fragment, containing an NdeI site engineered by site-directed mutagenesis at codons -1/+1, into pKS+ Bluescript (Stratagene, La Jolla, CA) and subsequently inserting the HindIII 3' SIR3 fragment from pJR273; thus pLP143 contains SIR3 as a SalI fragment flanked by a nonstandard polylinker. An ApaI-BamHI fragment that contains the SalI fragment was then inserted into pRS313 to make pLP187, or into pRS315 to make pLP190. pLP189 was made in parallel with pLP187, except that it additionally contains the A2T mutation at codon 2 made by site-directed mutagenesis. pLP465 and pLP468 were created by gap repair of pJR273, containing sir3-eso mutants R92K and T135I, respectively. The SalI fragments of pLP465 and pLP468 were subcloned into pRS313 to make pLP1048 and pLP946, respectively. pLP1131, pLP464, pLP675, pLP469, pLP473, and pLP472 were created by gap repair of pLP187, containing sir3-eso mutants R30K, L96F, sir3-8(E131K), E140K, L208S, and S813F, respectively. The wild-type SIR3 gene was cloned from strains W303-1a and 381G by gap repair of pLP187 to make pLP1130 and pL1133, respectively. pLP304 is the wild-type SIR3 gene in YEp351 (STONE and PILLUS 1996 ); pLP535, pLP1190, pLP828, pLP681, pLP791, pLP516, pLP526, pLP534, and pLP586 contain the sir3-eso mutations as SalI fragments in YEp351, made from the corresponding CEN-HIS3 plasmids, in the following order: A2T, R30K, R92K, L96F, sir3-8E131K, E140K, L208S, and S813F. A BamHI digest was performed to direct integration of pLP1202, an hml
::TRP1 construct. The following constructs were previously reported (BELL et al. 1995 ): pSIR3.12 (referred to here as ORC1N-SIR3C, containing the first 231 amino acids of Orc1p fused to the final 677 amino acids of Sir3p), pSIR3.15 (SIR3C, deleting the N-terminal 241 amino acids of Sir3p), pSPB1.34 (SIR3N-ORC1C, containing the first 235 amino acids of Sir3p fused to the final 679 amino acids of Orc1p), pSPB1.36 (ORC1-SIR3-ORC1, substituting amino acids 457–680 of Orc1p with amino acids 557–779 of Sir3p), pSIR3.13 (SIR3-ORC1-SIR3, substituting amino acids 505–834 of Sir3p with amino acids 405–738 of Orc1p), and pSPB1.43 (ORC1C, deleting the N-terminal 235 amino acids of Orc1p).
Gap repair and DNA sequencing:
The sir3-eso mutant alleles were rescued from their chromosomal locations by standard methods of gap repair (ROTHSTEIN 1991 ). The gap-repaired plasmids that rescued the sir3-eso mutants were made by introducing pJR273, digested with BamHI and ClaI, into strains LPY238 and LPY222 for pLP465 and pLP468, respectively; and from pLP187, digested with ClaI and StuI, into strains LPY275 and JRY188 for pLP464 and pLP675, or with StuI and NruI, into LPY669, LPY683, and LPY225 for pLP472, pLP473, and pLP676, or with StyI into LPY521 for pLP469, or with HpaI into W303-1a, LPY221, and 381G for pLP1130, pLP1131, and pLP1133. Because modestly increased gene dosage of the sir3-eso mutants appeared to be sufficient to restore mating in the strains from which gap repair was done, several plasmids were rescued from each strain and transformed into appropriate sir3 single and sir1 sir3 double null mutants to identify those with an eso phenotype. During the process of sequencing we discovered that the sir3-eso mutations often lay outside of the gapped regions, an observation that has been previously described (ROTHSTEIN 1991 ). Therefore the entire open reading frame was sequenced for each mutant allele. Because pLP676 and pLP469 were identical, we chose to use only one, pLP469, in further analysis. Additionally, sequencing pLP1130 and pLP1131 revealed that the wild-type SIR3 alleles from strains W303-1a and 381G (the parental strain used in the genetic screen that yielded the sir3-8 mutant; HARTWELL 1980 ) contain the identical sequence to that reported in the Saccharomyces Genome Database (open reading frame YLR442C; http://genome-www.stanford.edu/Saccharomyces/). Note that these sequences differ at several positions from the original SIR3 sequence reported (SHORE et al. 1984 ).
Sequencing was performed using an Applied Biosystems (Foster City, CA) automated facility. The oligonucleotides used for sequencing were
- SIR3-10: 5'GAGACTGCATGTGTACATAGGC3'
- SIR3-12: 5'GCAGCCCTTTCATCACCTTCC3'
- SIR3-131: 5'TAACTGCTGAGCTATCAGAGAT3'
- SIR3-14: 5'CAGAGGAAATACCAATAAACTC3'
- SIR3-15: 5'TTTAGACCGGTTTGCACCAG3'
- SIR3-16: 5'AGAAAATATGGTTCGCCATTTC3'.
Immunoblot analysis:
Preparation of protein extracts, SDS-PAGE electrophoresis, and immunoblotting for Sir3p detection were performed as described (STONE and PILLUS 1996 ). High-copy 2µ plasmids were used to facilitate detection of Sir3p from wild-type and mutant strains. For all parameters tested previously, results with high-copy plasmids and endogenous genes were identical (STONE and PILLUS 1996 ).
Quantitative mating and telomeric silencing assays:
For quantitative mating assays, cells were grown to midlogarithmic phase in a supplemented synthetic medium for plasmid selection. They were then diluted appropriately to obtain ~100–300 colonies per plate and plated on the same medium to quantitate total number of cells. At the same time, appropriate dilutions for testing mating were mixed with MATa or MAT
mating-type testers LPY142 or LPY78, respectively, and plated onto minimal medium for diploid selection to quantitate the number of mating-competent cells. The mating efficiency is defined as the number of cells that mated per total number of cells. At least two experiments were performed for each strain, for which mean values were determined and the range of each of those values was indicated.
For monitoring telomeric silencing from a URA3 reporter gene (APARICIO et al. 1991 ), cells were grown to saturation for 2–3 days at room temperature in a supplemented synthetic medium for plasmid selection. These cultures were then plated in serial fivefold dilutions onto the same medium either lacking or containing 5-FOA, and plates were incubated at room temperature until colony growth was visible. Colonies resistant to 5-FOA are silenced for URA3.
| RESULTS |
|---|
A genetic screen identifies enhancers of the sir1 mutant mating-defective phenotype:
The SIR1 gene functions in establishing stable and heritable patterns of gene expression at the silent mating-type HM loci (PILLUS and RINE 1989 ). The sir1 null mutant exhibits a partial mating-defective phenotype due to an epigenetic phenomenon in which the silent mating-type genes in some cells of the population are completely derepressed yet in other cells the HM genes are fully repressed (PILLUS and RINE 1989 ). To identify genes that contribute to silencing in the population of transcriptionally repressed sir1 mutant cells, we performed a genetic screen for enhancers of the sir one mutant mating phenotype (eso) mutants (REIFSNYDER et al. 1996 ). Mutants were identified that were completely mating defective in the absence of SIR1, but mating competent in its presence. This was done by replica-plating mutagenized sir1 mutant cells on two media, either selecting for or against a SIR1 plasmid, then testing for mating in the presence or absence of the plasmid (REIFSNYDER et al. 1996 ). By looking for mutants that were mating defective only in the absence of SIR1, we sought to avoid isolating previously characterized mutants that were completely silencing defective.
Six complementation groups were found to affect mating in sir1 mutant cells. One of the eso mutants was in SAS2, a gene that encodes a member of a conserved family of acetyltransferases (REIFSNYDER et al. 1996 ). Five other complementation groups contained mutant alleles of genes known to be involved in silencing: ARD1, NAT1, SIR2 (S. GARCIA and L. PILLUS, unpublished results), SIR3, and SIR4 (see MATERIALS AND METHODS for details). Alleles of ARD1 and NAT1 were predicted to arise from the eso screen, as null mutants were previously shown to be mating defective in combination with sir1 mutants (WHITEWAY et al. 1987 ; STONE et al. 1991 ). This report focuses on characterization of the sir3-eso mutants.
Eight sir3-eso mutant alleles were rescued on plasmids by gap repair (see MATERIALS AND METHODS). In addition to these eight alleles from the eso screen, two other independently isolated sir3 mutants were found to have an eso phenotype and thus were included in our analysis. One mutant, called A2T, was a site-directed mutant in which the alanine residue at codon 2 was changed to threonine for a separate study (E. M. STONE and L. PILLUS, unpublished data). The other was the sir3-8 allele, previously described to be temperature sensitive for mating (HARTWELL 1980 ; RINE and HERSKOWITZ 1987 ). The sir3-8 allele was recovered on a plasmid by gap repair and, similar to the behavior of sir3-8 at its chromosomal locus, sir3 null mutant strains bearing the sir3-8 plasmid were temperature sensitive for mating. Transformants containing the sir3-8 plasmid were completely mating defective at the restrictive temperature of 37° (data not shown) but fully mating competent at the permissive temperature of 23° in a SIR1 sir3 strain. The plasmid was tested for its ability to complement the mating-defective phenotype of a sir1 sir3 double mutant at the permissive temperature and was found to be unable to fully complement the mating defect (see below). Therefore sir3-8 was classified as an eso mutant. Other previously identified sir3 alleles, however, were not eso mutants. The SIR3R1 and SIR3R3 alleles, originally recovered as suppressors of the mating defect of histone H4 N-terminal mutants (JOHNSON et al. 1990 ), did not enhance the mating defect of the sir1 mutant (T. LEWIS and L. PILLUS, unpublished results). Moreover, SIR3L31 (and SIR3N205, identical to SIR3R3), isolated as suppressors of the telomeric silencing defect of rap1 C-terminal mutants (LIU and LUSTIG 1996 ), did not exhibit an eso phenotype (data not shown). Thus the sir3-eso mutants are distinguished by their SIR1-specific phenotype.
Altered residues in sir3-eso mutant alleles cluster at the N terminus:
The entire SIR3 open reading frame was sequenced for each gap-repaired allele to identify the changes that conferred the eso phenotype. Only 1 bp was found to be mutated for each allele. The alleles were thus renamed to reflect the nature of the mutations (Table 2). One mutation, leading to the E140K substitution, was independently isolated twice, but all others were distinct. Therefore, a total of nine different altered residues were identified in the collection of sir3-eso mutants. Interestingly, eight of these nine substitutions clustered at the N terminus of Sir3p. This is the 214-amino-acid region most similar to Orc1p (50% identical, 63% similar; BELL et al. 1995 ). Seven of the mutated residues that mapped to this region were identical in wild-type Sir3p and Orc1p, and one was conserved (L96 in Sir3p; V96 in Orc1p). The remaining allele, S813F, was found in a larger region encoding the C terminus that is also conserved in Sir3p and Orc1p, although not as extensively as is the N-terminal region. Five sir3-eso mutations led to changes in conserved residues in the recently defined BAH domain (bromo-adjacent homology, at Sir3p N-terminal amino acids 48–189; CALLEBAUT et al. 1999 ). The BAH module is widely conserved among diverse proteins, including Sir3p and Orc1p, DNA methyltransferases, and DNA replication proteins; it has been suggested to be important for protein-protein interactions in processes that link methylation, replication, and transcriptional regulation. Thus, mutations that lead to changes in conserved residues within an N-terminal domain of Sir3p resulted in the eso phenotype and may serve to define a functionally significant domain.
The protein encoded by the sir3-8 mutant is thermolabile:
Because changes in Sir3 protein levels might account for silencing defects seen in sir3-eso mutants, immunoblot analysis was performed to determine if steady-state levels of Sir3 mutant proteins were similar to those of wild-type Sir3p. Immunoblot analysis of whole-cell protein extracts, using a polyclonal Sir3p antiserum, demonstrated that Sir3p levels and electrophoretic mobility were comparable to wild type for all of the sir3-eso mutants (data not shown). The mobility of sir3-eso mutant proteins was also evaluated during the pheromone and starvation responses, previously shown to result in Sir3p hyperphosphorylation and associated mitogen-activated protein (MAP) kinase pathway modulation of silencing (STONE and PILLUS 1996 ). None of the alleles examined appeared defective in pheromone- or starvation-induced Sir3p hyperphosphorylation (data not shown). However, when protein was examined from a sir3-8 mutant culture grown at the restrictive temperature of 37°, little to no protein was detected (Fig 1). When grown at the permissive temperature of 23°, sir3-8 mutant protein migrates normally (compare first and third lanes, Fig 1), but it disappears when shifted for 3 hr or more to 37° (final lane on right, Fig 1). When grown for as many as 16 generations after shifting to the restrictive temperature, sir3-8p remains undetectable (data not shown).
View larger version (106K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1.
The sir3-8 protein is thermolabile at the restrictive temperature. Immunoblot of whole-cell lysates was probed with an anti-Sir3p antiserum. Transformants of sir3 null mutant strain LPY2709 contained wild-type SIR3 plasmid pLP304 or sir3-8 mutant plasmid pLP791. Cultures were grown at 23°, and one-half of each culture was shifted to 37° for 3 hr before harvesting. Note that the temperature shift causes a change in Sir3p mobility (compare lanes 1 and 2) due to hyperphosphorylation as previously described (STONE and PILLUS 1996 ).
|
|
SIR1 overexpression was previously shown to partially suppress the mating defect of sir3-8 mutant cells (STONE et al. 1991 ). This suppression appears not to result from stabilization of sir3-8p, however, as the mutant protein levels are not restored by SIR1 overexpression (data not shown). Our results identifying sir3-8p as a thermolabile protein, together with previous data demonstrating that the sir3-8 mutant is completely mating defective at 37° (HARTWELL 1980 ; RINE and HERSKOWITZ 1987 ), support the interpretation that sir3-8 behaves as a conditional null allele.
Two classes of sir3-eso alleles exhibit different mating-defective phenotypes:
To quantitate sir3-eso mutant mating phenotypes, each sir3-eso allele was introduced into appropriate strains on a centromeric plasmid expected to be present in approximately one or two copies per transformed cell. Plasmids were used in quantitative experiments to ensure that the strain backgrounds were isogenic. The mutants behaved similarly when present on plasmids or at their endogenous chromosomal locus (data not shown). The sir3-eso plasmids were tested for their ability to complement the chromosomal null sir3 mutant, and mating phenotypes were compared for wild-type SIR1 and sir1 mutant strains in MAT and MATa backgrounds. Because the sir3-8(E131K) mutant is temperature sensitive for mating, all assays were performed at the permissive temperature of 23°. Strains carrying the other alleles exhibited no defect at 37° and behaved similarly at 23° and 30°.
Quantitative mating assays revealed that strains carrying the sir3-eso plasmids were severely mating defective in a MAT sir1 sir3 background but had little or no mating defect in a MAT SIR1 sir3 background (Table 3). Thus, the sir3-eso mutants clearly did not represent null or complete loss-of-function alleles. For seven of the sir3-eso plasmids, transformants in the sir1 sir3 mutant background mated with 10-5–10-6 efficiency or less, similar to the vector control transformant lacking wild-type SIR3 altogether. The two remaining sir3-eso mutants, sir3-A2T and sir3-8(E131K), were partially mating impaired in sir1 sir3 strains. These results are in contrast to the sir1 mutant carrying a wild-type SIR3 plasmid, which exhibited only a mild decrease in mating compared to the SIR1 strain (Table 3). Thus, all of the sir3-eso mutants significantly enhance the sir1 mutant mating-defective phenotype in the MAT background.
Quantitative mating analysis revealed that all nine sir3-eso mutants also showed severely decreased mating efficiency in a MATa sir1 sir3 strain (Table 4). Interestingly, a subset of strains carrying the sir3-eso alleles exhibited a mating defect in the presence of SIR1 in the MATa sir3 background. Two strains bearing different alleles, sir3-T135I and sir3-E-140K, mated 100-fold less efficiently than wild-type SIR3 strains. A third mutant, sir3-L208S, was completely nonmating in a MATa strain. It should be noted that this allele would not have been recovered in the eso screen in the MATa background, although it clearly fits the definition of an eso mutant when present in a MAT strain. Because none of these mutants was mating defective in the MAT strain (Table 3) and because sir3-T135I, sir3-E140K, and sir3-L208S cluster near one another within the region encoding the N terminus of Sir3p, they define a MATa-specific class of alleles that may identify an N-terminal functional subdomain important for silencing HML but not HMRa in a wild-type SIR1 strain background (see DISCUSSION).
We determined whether the sir3-eso alleles were dominant or recessive by quantitative mating assays in a MAT sir1 strain that was wild type for SIR3. The MAT background was used to avoid the MATa-specific effects of some of the alleles noted above. This analysis revealed that three of the alleles, sir3-T135I, sir3-E140K, and sir3-L208S, had partially dominant phenotypes, exhibiting a decreased mating efficiency of 10-3 (Table 3). These three alleles also constitute the class with MATa-specific mating-defective phenotypes. We grouped these alleles, calling those that are primarily recessive class I and those three alleles that exhibit MATa-specific and dominant eso mating defects class II. None of the sir3-eso mutants had dominant mating defects in a wild-type SIR1 mutant background, as wild-type MATa and MAT strains carrying plasmids containing the sir3-eso alleles mated with normal efficiency (data not shown).
To determine if sir3-eso mutants, like sir1 mutants, inherit alternate states of HM locus gene expression, consistent with a role in establishing silencing, we performed pedigree analysis on several cell lineages (PILLUS and RINE 1989 ). Two sir3-eso mutants, R30K and S813F, exhibited lineages in which cells transcriptionally silenced at the HM locus gave rise to transcriptionally active cells in subsequent generations (data not shown). Thus, these two alleles appear to be defective in maintenance of silent chromatin, and an establishment defect is not a general property of sir3-eso mutants.
Sir3p has been hypothesized to dimerize (MORETTI et al. 1994 ), and partial trans-complementation between an N-terminal coding region and C-terminal coding region of SIR3 has been observed (LE et al. 1997 ; GOTTA et al. 1998 ). Therefore heterodimer formation between different sir3-eso mutant proteins might lead to interallelic complementation. When MATa sir1 or MAT sir1 strains bearing the alleles sir3-R30K, sir3-R92K, sir3-L208S, and sir3-S813F at their endogenous chromosomal locus were transformed with the entire panel of sir3-eso plasmids, mating ability was not restored (data not shown). The failure of different sir3-eso mutants to function in interallelic complementation may reflect a requirement for Sir1p in heterodimer formation, or may be due to the inability of sir3-eso heterodimers to achieve an appropriate tertiary structure.
Mating defects are observed in a subset of nat1 sir3-eso double mutants:
NAT1 and ARD1 encode subunits of an N-terminal acetyltransferase (MULLEN et al. 1989 ; PARK and SZOSTAK 1992 ). Null nat1 and ard1 single and double mutants have identical phenotypes, including a partial mating defect in a MATa strain background (MULLEN et al. 1989 ) and synergistic loss of mating in the absence of SIR1 (WHITEWAY et al. 1987 ; STONE et al. 1991 ). Thus nat1 mutants have an eso phenotype in both MATa and MAT strains since they enhance the sir1 mating defect and in fact were identified in the eso screen described here (see MATERIALS AND METHODS). Because both sir3-eso mutants and nat1 mutants enhance the mating defect of sir1 mutants, we asked if sir3-eso mutants enhance the nat1 mating defect in a wild-type SIR1 background. The sir3-eso plasmids were introduced into a MAT nat1 sir3 strain, as MAT nat1 mutants and MAT sir3-eso mutants are completely mating competent, in contrast to either of these mutants in a MATa background.
Quantitative mating data revealed that some but not all of the sir3-eso alleles have more severe phenotypes in combination with nat1 mutants (Table 5). Strains with the sir3-T135I and sir3-L208S alleles exhibited an ~100-fold decreased mating efficiency and are found among the class II mutants with MATa-specific and dominant mating defects. The third class II mutant, sir3-E140K, did not exhibit worsened mating in the nat1 mutant. Moreover, two additional sir3-eso mutants, sir3-8(E131K) and sir3-S813F, were completely mating defective in the absence of NAT1. The remaining mutants had little or no effect on mating efficiency in combination with nat1 mutants. In addition, mating defects were tested in sir3-eso sas2 double mutants. SAS2 is also a member of a gene family with acetyltransferase activity and, like sir3-eso mutants, the sas2 mutant is severely mating defective only in the absence of SIR1 (REIFSNYDER et al. 1996 ). Mating efficiency for all sir3-eso sas2 double mutants was comparable to that of the single mutants (data not shown), suggesting a specific interaction between the NAT1 and SIR3 genes.
sir3-eso mutants are defective in telomeric silencing:
In addition to its requirement for silencing at the HM loci, SIR3 is essential for telomeric silencing. The URA3 reporter gene placed proximal to telomeric sequences is transcriptionally repressed, and this telomeric silencing, observed as sensitivity to 5-FOA, is abolished in a sir3 null mutant (APARICIO et al. 1991 ). Resistance to 5-FOA is a measure for silencing of the telomere-positioned URA3 gene, because cells expressing URA3 are sensitive to 5-FOA and only silenced cells are resistant and able to form colonies. We tested whether sir3-eso mutants functioned in telomeric silencing by plating a dilution series of transformants on growth medium selecting for plasmid maintenance (-his) and on 5-FOA-containing medium to monitor silencing and maintain selection for the plasmid (5-FOA-his). Control transformants of a sir3 mutant strain showed no growth with vector only and good growth with a wild-type SIR3 plasmid (Fig 2A). Transformants containing the sir3-eso plasmids were completely defective for eight of the sir3-eso mutants, and the sir3-8(E131K) mutant exhibited a partial defect in telomeric silencing. The telomeric silencing phenotype was apparent in the presence of SIR1, and additional telomeric defects were not detected in a sir1 mutant background (data not shown). Therefore the sir3-eso mutants are unable to function in telomeric silencing.
View larger version (108K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2.
sir3-eso mutants exhibit telomeric silencing defects. Serial fivefold dilutions were plated on supplemented synthetic medium selecting for plasmid maintenance to monitor growth (-his, left) or the same medium containing 5-FOA to monitor silencing of the telomere-proximal URA3 reporter gene (right). Transformants of sir3 mutant strain LPY3238 for complementation test (A), or of SIR3 wild-type strain YDS631 for dominance test (B), contained the following CEN-based plasmids ordered from top to bottom: vector control pRS313; SIR3 wild-type control pLP187; and sir3-eso alleles pLP189, pLP1131, pLP1048, pLP464, pLP675, pLP946, pLP469, pLP473, and pLP472.
|
|
Because several sir3-eso mutants exhibited dominant effects in silencing at the HM loci, the panel of mutants was also examined for dominant effects on telomeric silencing. Plasmids bearing each mutant allele were transformed into a telomeric reporter strain that was wild type for SIR3 and tested for silencing as above. Four mutants showed a striking loss of telomeric silencing in the presence of wild-type SIR3 (Fig 2B). Three of these dominant mutants, sir3-T135I, sir3-E140K, and sir3-L208S, are class II mutants, which exhibit a dominant mating-defective phenotype in the absence of SIR1, whereas one, sir3-R92K, is a class I mutant, among those that are recessive for the eso mating defect. The mutants with dominant telomeric silencing defects presumably form nonfunctional complexes with other silencing proteins, thereby interfering with wild-type SIR3 silencing at telomeres.
The Orc1p N terminus functionally replaces that of Sir3p in mating-type silencing in the absence of Sir1p, but not in telomeric silencing:
Given the high degree of sequence similarity between the N termini of Sir3p and Orc1p, a chimeric protein between the first 231 amino acids of Orc1p and the Sir3p C-terminal 677 amino acids was created and tested for silencing function (BELL et al. 1995 ). This ORC1N-SIR3C construct is capable of substituting for Sir3p in mating-type silencing (BELL et al. 1995 ). Because the sir3-eso mutations affected Orc1p-conserved residues, we hypothesized that the N terminus of Orc1p might also functionally replace the Sir3p N terminus in the absence of Sir1p. Indeed, the Orc1N-Sir3Cp chimera was mating proficient in both MATa and MAT sir1 mutant backgrounds (Table 4 and data not shown). The Orc1N-Sir3Cp chimera was also tested for complementation of the sir3 mutant telomeric silencing defect. In contrast to its efficient mating, the Orc1N-Sir3Cp chimera was defective in telomeric silencing (Fig 3A), exhibiting only partial function reflected by its intermediate growth on 5-FOA-containing medium. The Orc1N-Sir3Cp telomeric silencing defect was comparable in both sir1 mutant and SIR1 wild-type strain backgrounds (data not shown). Moreover, the Orc1N-Sir3Cp chimera appeared to be partially dominant, inhibiting telomeric silencing in SIR3 strains (Fig 3B). Thus, the N terminus of Sir3p is distinguished from that of Orc1p by its function at the telomere, perhaps through interactions with telomere-specific silencing proteins.
View larger version (105K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3.
The Orc1p N terminus cannot substitute for that of Sir3p in telomeric silencing. Serial fivefold dilutions were plated on supplemented synthetic medium for plasmid selection to monitor growth (-leu, left) or the same medium containing 5-FOA to monitor silencing of the telomere-proximal URA3 reporter gene (right). Transformants of sir3 null mutant strain LPY3238 for complementation test (A), or of SIR3 wild-type strain YDS631 for dominance test (B), contained the following plasmids: vector control pRS315; SIR3 wild-type control pLP190; and ORC1N-SIR3C (pSIR3.12), SIR3C (pSIR3.15), SIR3N-ORC1C (pSPB1.34), ORC1-SIR3-ORC1 (pSPB1.36), SIR3-ORC1-SIR3 (pSIR3.13), or ORC1C (pSPB1.43) (BELL et al. 1995 ). Original plasmid designations are noted in parentheses.
|
|
To further dissect the effect of SIR3 and ORC1 sequences on telomeric silencing in sir3 mutant and SIR3 wild-type strain backgrounds, the remaining chimeric and deletion constructs in the series reported by BELL et al. 1995 were examined. The additional constructs tested were a SIR3N-ORC1C chimera, which is the reciprocal chimera to the ORC1N-SIR3C noted above; two sandwich chimeras in which internal ORC1 and SIR3 regions were swapped to make ORC1-SIR3-ORC1 and SIR3-ORC1-SIR3; and N-terminal deletions to result in SIR3C and ORC1C. None of the constructs supported telomeric silencing in the sir3 mutant strain (Fig 3A), consistent with their inability to promote mating-type silencing (BELL et al. 1995 ). However, partial dominance was observed with the SIR3C and SIR3-ORC1-SIR3 constructs (Fig 3B), in addition to the ORC1N-SIR3C chimera described above. None of the other constructs exhibited dominant effects on telomeric silencing. As the C terminus of Sir3p was unable to complement the sir3 mutant, the partial telomeric silencing seen for the Orc1N-Sir3Cp chimera must be due to a low level of function of the Orc1p N-terminal region. In contrast, the partial dominance of Orc1N-Sir3Cp, Sir3Cp, and Sir3-Orc1-Sir3p may be due to Sir3p sequences common to all three constructs, which might sequester other silencing proteins into nonfunctional complexes.
Tethered Sir1p suppresses the telomeric silencing defect of sir3-eso mutants:
SIR1 does not appear to function in silencing telomeric reporter genes in standard assays, although it may play a modest role at native telomeres (APARICIO et al. 1991 ; PRYDE and LOUIS 1999 ). However, strong silencing is achieved by tethering Sir1p directly to telomeric sequences via the Gal4p DNA binding domain (GBD; CHIEN et al. 1993 ). Because silencing at the HM loci occurs only in the presence of SIR1 in sir3-eso mutant strains, we asked if the telomeric silencing defect of the sir3-eso mutants could be suppressed by tethering Sir1p to the telomere. A sir3 null mutant strain with the GAL4-upstream activating sequence positioned near a telomeric reporter gene was transformed with either the GBD control vector or GBD-SIR1 plasmid and the sir3-eso genes or appropriate controls on high-copy 2µ plasmids.
The sir3-eso mutants exhibited telomeric silencing defects even when present at elevated dosage on 2µ plasmids, as observed in control transformants containing GBD alone (Fig 4A). However, partially restored telomeric silencing was observed for several of the mutants when they were overexpressed, particularly sir3-A2T and sir3-8 (compare Fig 2A and Fig 4A). Significantly, when Sir1p was directed to telomeres with the GBD-SIR1 plasmid, six of the sir3-eso mutants exhibited near or complete restoration of telomeric silencing (Fig 4B). In contrast, GBD-SIR1 did not suppress the telomeric silencing defect of any of the Class II sir3-eso mutants that were dominant for mating-type silencing defects, T135I, E140K, and L208S. A control experiment, in which GBD-SIR1 and the sir3-eso plasmids were coexpressed in an isogenic strain without tethering sites, showed that SIR1 overexpression itself did not suppress the sir3-eso telomeric silencing defect (data not shown). Together, these data suggest that sir3-eso mutant telomeric silencing phenotypes, like the mating-defective phenotypes, can also be made dependent on Sir1p function. SIR1-mediated suppression may occur in one of several ways. For example, suppression may occur via protein-protein interactions of tethered Sir1p and sir3-eso mutant proteins or by the ability of Sir1p to independently establish silencing at telomeres when tethered, thereby compensating for sir3-eso mutant defects.
View larger version (97K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4.
Tethered Sir1p suppresses sir3-eso mutant telomeric silencing defects. Serial fivefold dilutions were plated on supplemented synthetic medium to monitor growth, selecting simultaneously for maintenance of two plasmids (-his -leu, left), or the same medium containing 5-FOA to monitor silencing of a telomere-proximal URA3 reporter gene that contained GAL4 binding sites for tethering (right). Transformants of sir3 mutant strain LPY4417 containing control GBD plasmid pMA424 (A), or pKL5 Sir1p tethering plasmid GBD-SIR1 (B), as well as the following 2µ plasmids carrying sir3-eso mutants and controls, are ordered from top to bottom: pLP304, pLP535, pLP1190, pLP828, pLP681, pLP791, pLP516, pLP526, pLP534, pLP586, and YEp351.
|
|
| DISCUSSION |
|---|
In a genetic screen for enhancers of the sir1 mutant silencing defect, we identified a collection of sir3 mutant alleles. Unlike the sir3 null mutant, which is completely mating defective in the presence of SIR1, the mating defects of these sir3-eso mutants are seen primarily in the absence of SIR1. Some of the mutants exhibit dominant effects; all of the sir3-eso alleles are defective in telomeric silencing. Mutated residues cluster in an N-terminal region that exhibits a high degree of sequence similarity with the N terminus of the DNA replication initiator Orc1p. The clustering of the mutations in the sir3-eso mutants thus identifies a domain of Sir3p that contributes to silencing in the absence of SIR1. The sir3-eso mutants also provide clues for the role of this shared domain in Orc1p and Sir3p and further evidence for a functional relationship between Sir1p and Sir3p.
sir3-eso mutants define functional domains in the Sir3 protein:
The phenotypic profile of the sir3-eso mutants provides insight into the emerging picture of different functional domains within Sir3p (for review, see STONE and PILLUS 1998 ). Sir3p can be viewed as consisting of two large domains (Fig 5): an N-terminal region with a high degree of similarity to Orc1p and an extended C-terminal region of the protein that can associate with other silencing proteins, including Sir2p, Sir4p, Rap1p, and the histones. Eight of the nine sir3-eso mutants identified cluster within the N-terminal domain (Fig 5), disrupting Sir3p silencing function at both HM loci in the absence of SIR1. Two predominant classes of sir3-eso alleles were uncovered: (1) those that are recessive for the eso phenotype and (2) those that are dominant and also MATa specific (i.e., affecting HML silencing) in wild-type SIR1 strains. The class II mutants cluster in the N-terminal region between amino acid residues 135 and 208. In SIR1 strains, both class I and II mutants function normally at HMR, but the class II mutants are defective at HML. Thus, the class II domain may be required for interaction with a silencing factor only at HML. The function of the N-terminal subdomain may be supplied by a redundant mechanism at HMR, consistent with the redundancy observed within the HMRE silencer (see, for example, BRAND et al. 1985 ).
View larger version (21K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 5.
Sir3p domains confer specific silencing functions. Nine different sir3-eso alleles defining two distinct classes of mutants were identified in this study as indicated. Eight of these cluster in the N-terminal region previously noted to be highly similar to Orc1p (in dark green, amino acids 1–214; BELL et al. 1995 ). The sir3-8(E131K) allele is distinguished in that it is temperature sensitive for mating. The C-terminal two-thirds of the protein has been implicated in a number of physical interactions with other proteins, including Rap1p, Sir4p, and histones H3 and H4 (light green; for review, see STONE and PILLUS 1998 ). Sir2p associates indirectly with Sir3p through a mutual Sir4p interaction (MOAZED et al. 1997 ; STRAHL-BOLSINGER et al. 1997 ). Asterisks indicate three previously identified suppressors of mutations in histone H4 or Rap1p (JOHNSON et al. 1990 ; LIU and LUSTIG 1996 ): SIR3L31 (S31L), SIR3R1 (W86R), and SIR3R3/SIR3N205 (D205N), respectively, none of which exhibit an eso mutant phenotype. The sir3-eso mutants R92K, L96F, E131K, T135I, and E140K disrupt conserved residues within the BAH domain found in DNA methyltransferases and other proteins with replication and chromatin functions, including both Sir3p and Orc1p (CALLEBAUT et al. 1999 ).
|
|
All of the sir3-eso alleles are defective in telomeric silencing. This phenotype is seen even in SIR1 strains, consistent with previous suggestions that Sir1p does not play a major role in silencing telomeric reporter genes (APARICIO et al. 1991 ). However, Sir1p is known to promote silencing when tethered to telomeric sequences (CHIEN et al. 1993 ), and we showed that tethered Sir1p suppresses telomeric silencing defects of the class I sir3-eso mutants. Thus Sir1p function can compensate for the sir3-eso mutant silencing defect at the HM loci and when tethered at the telomeres. Therefore, Sir1p may directly recruit Sir3p and when tethered may substitute for a telomeric factor that can no longer interact normally with Sir3p in the sir3-eso mutants. The telomeric silencing defect exhibited by the Orc1p-Sir3p chimera is consistent with there being a specific role for the Sir3p N terminus at telomeres, and this function might involve interaction with a silencing factor that is either redundant (with Sir1p, for example) or unnecessary at the HM loci. Orc1p itself may play a role at telomeres since mutations in two other ORC subunits, ORC2 and ORC5, cause telomeric silencing defects (FOX et al. 1997 ).
The class II sir3-eso mutants are characterized by dominant effects on mating and telomeric silencing phenotypes, yet these mutant proteins must have some productive interactions with other silencing proteins, because they all function in silencing under some circumstances. However, these dominant mutants must interfere with the function of factors with which they interact in other situations. For example, dominant mating defects are observed in the absence of SIR1, but not in its presence, suggesting an inability to disrupt the silencing function that is performed by SIR1 at the HM loci. Moreover, as tethered Sir1p does not suppress the class II alleles, it must not be able to function with those sir3-eso proteins at the telomeres. Interestingly, one of the class I mutants (R92K) is dominant for telomeric silencing but behaves like the other members of its class in that it is suppressed by tethered Sir1p.
Several sir3-eso mutants enhance the nat1 mutant mating defect. Interestingly, Sir3p is a potential substrate of Nat1p/Ard1p N-terminal acetyltransferase activity, indicated by its alanine residue at codon 2 (SHERMAN et al. 1993 ). However, the nat1 mutant phenotype does not appear to result from the absence of N-terminal acetylation on Sir3p, and so presumably acetylation of some other silencing protein by Nat1p must be required for normal silencing (E. M. STONE and L. PILLUS, unpublished data). Because nat1 mutants, like sir3-eso mutants, act as enhancers of the sir1 mating-defective phenotype (this study and STONE et al. 1991 ), the observed nat1 sir3-eso interactions imply that Nat1p, Sir1p, and Sir3p provide interdependent means through which silencing may be achieved.
The temperature-sensitive mating phenotype of the sir3-8(E131K) mutant is unique among sir3 mutants (HARTWELL 1980 ; RINE and HERSKOWITZ 1987 ). We determined that the sir3-8 allele has the additional phenotype of enhancing the sir1 mutant phenotype at permissive temperatures. Additionally, we discovered that sir3-8 encodes a thermolabile protein, thereby explaining the nature of this well-known mutant in molecular terms and allowing more detailed interpretations of earlier experiments using this allele (MILLER and NASMYTH 1984 ; HOLMES and BROACH 1996 ). For example, to evaluate the contribution of cell cycle control, silencing was abolished by raising sir3-8 mutant cells to the restrictive temperature. Then, cells were arrested at G1 and released at permissive temperature. Silencing was restored only during S-phase, demonstrating that establishment of silencing required DNA replication or some other S-phase event. In contrast, shifting sir3-8 mutants from permissive to restrictive temperature revealed that maintenance of silencing can be destroyed throughout the cell cycle (MILLER and NASMYTH 1984 ). Previously, the loss of silencing in the sir3-8 mutant at the restrictive temperature could be explained in at least two ways: inappropriate protein folding of sir3-8p within an intact multiprotein complex might disrupt its function, vs. protein instability might result in loss of function of sir3-8p within the complex or disassembly of the complex itself. We now favor the latter hypothesis, since the sir3-8 mutant behaves as a conditional null allele.
We hypothesize that different interaction surfaces of the Sir3p molecule are disrupted by the different classes of sir3-eso mutations. These mutants will be valuable for extending analysis of different domains of the Sir3p molecule as detailed structural information becomes available.
How might Sir1p, Sir3p, and Orc1p functions be linked?
Silencing is a heritable regulatory state, resulting from separable establishment and maintenance functions (PILLUS and RINE 1989 ; reviewed in RIVIER and RINE 1992 ). Sir1p functions in establishing silencing but has no apparent role in its maintenance. In contrast, other Sir proteins, including Sir3p, have clear roles in maintenance. In addition to Sir1p function in establishment, there is presumably some additional mechanism for establishing silencing since a subpopulation of sir1 mutant cells initiates and propagates the silenced state (PILLUS and RINE 1989 ). Candidate proteins for this function include those that promote telomeric silencing when tethered to engineered sites for ectopic DNA binding proteins, such as Sir3p, Sir4p, and Rap1p (BUCK and SHORE 1995 ; LUSTIG et al. 1996 ; MARCAND et al. 1996 ). One goal of the screen for eso mutants was to find other genes that function in a manner similar to SIR1. Because sir3-eso mutant alleles were identified in the screen, they were tested for establishment defects. Pedigree analysis suggested that although sir3-eso mutants are weakly defective in the maintenance of silencing, they are unlikely to have an establishment defect in a SIR1 wild-type background. The gene(s) responsible for establishment in the silencing-competent subpopulation of sir1 mutant cells have not yet been unambiguously identified, but may include RAP1, or other genes uncovered in the eso mutant screen like SIR2 or SIR4, or other as yet uncharacterized genes.
Several possibilities exist to explain the eso phenotype of the sir3 alleles. The weak sir3-eso maintenance defect, in combination with the sir1 establishment defect, may lead to complete derepression of the silent mating-type loci. Alternatively, SIR1 may have an unsuspected role in maintenance of silencing that is normally redundant with that of SIR3, leading to the failure to maintain silent chromatin only in the sir1 sir3-eso double mutants. Another possibility is that SIR3 may indeed function in establishing silencing, not constitutively, but in the absence of SIR1. This would be analogous to the situation seen for MAP kinase pathway genes in which KSS1 substitutes for FUS3 only in a fus3 null mutant (MADHANI et al. 1997 ). Finally, there may not be any other protein that functions similarly to Sir1p, but instead SIR1-independent establishment may occur by default at a low frequency (for example, other Sir proteins may occasionally find their way to silencers in the absence of potential recruitment by Sir1p). The molecular definition of Sir1p function and the mechanism of establishing silencing thus remain to be resolved. It is possible that Sir1p's function is distinct from the nucleation role revealed by tethering and mutant studies in which Sir3p, Sir4p, or Rap1p have been implicated in recruiting silencing factors to silent loci (SUSSELL et al. 1993; LUSTIG et al. 1996 ; MARCAND et al. 1996 ).
Genetic interactions between SIR1 and SIR3 raise the possibility that the two proteins may physically associate with one another. Our studies demonstrate that sir3-eso mutants can enhance the sir1 mutant mating defect and that the telomeric silencing defect of the sir3-eso mutants can be suppressed by tethered Sir1p. Furthermore, SIR1 overexpression suppresses the mating defects of certain sir3 mutant alleles (STONE et al. 1991 ). Additionally, Sir1p physically interacts with the N terminus of Orc1 (TRIOLO and STERNGLANZ 1996 ), and Orc1p and Sir3p share significant sequence similarity at their N termini (BELL et al. 1995 ). It is plausible that Sir1p interacts with the N terminus of Sir3p. However, physical interactions between Sir1p and Sir3p have yet to be observed in either two-hybrid analysis (TRIOLO and STERNGLANZ 1996 ) or coimmunoprecipitation experiments (E. M. STONE and L. PILLUS, unpublished data). A transient interaction between Sir1p and Sir3p, perhaps occurring only at a specific point in the cell cycle, might be responsible for the establishment of silencing. If Sir1p does not interact with the Sir3p N terminus, perhaps another domain of the Sir3p molecule actively inhibits a potential Sir1p-Sir3p interaction.
The high degree of similarity between the Sir3p and Orc1p N termini suggests a shared function between these domains that may be imagined in several ways. For example, Sir3p may in some instances substitute for Orc1p in the traditional ORC, thereby creating an alternative complex with modified or inhibitory function in DNA replication. Conversely, Orc1p may substitute for Sir3p in a subset of the multiprotein complexes containing the other silencing proteins; such a model would then predict that Orc1p is capable of performing some function independent of the other Orc subunits. The proposal that ORC may play a role independent of DNA replication is supported by the observation that silencing and DNA replication functions of ORC are separable, although they are still SIR dependent (FOX et al. 1995 ; DILLIN and RINE 1997 ). An additional possibility is that Sir2p, Sir3p, Sir4p, Rap1p, and histones may coexist with Sir1p and Orc1p in the same multiprotein complex. It is noteworthy that the N-terminal regions of both Sir3p and Orc1p contain the recently identified BAH domain, which is also found in DNA methyltransferases and other proteins thought to act in transcriptional control. The BAH domain is proposed to direct these proteins to their sites of action within chromatin (CALLEBAUT et al. 1999 ). Because five of the sir3-eso mutations are within the BAH domain, their further analysis may lead to greater understanding of the role of this domain in chromatin structure and function.
| FOOTNOTES |
|---|
1 Present address: Science and Health Education Partnership, University of California, 100 Medical Center Way, Woods Bldg., San Francisco, CA 94143-0905. 
| ACKNOWLEDGMENTS |
|---|
We thank A. Clarke, S. Garcia, J. Heilig, S. Jacobson, J. Lowell, R. Sternglanz, and R. West for critically reading the manuscript; S. Loo for his contribution in an initial phase of the eso mutant screen; L. Hartwell, J. Rine, and R. Sternglanz for plasmids and strains; and Y. Han for sequencing assistance. This work was initiated with funding from the National Science Foundation (NSF) and continued with support from the National Institutes of Health. M.M. received support from a Summer Fellowship from the Cancer Center at the University of Colorado Health Sciences Center, and B.G. received support from an REU supplement from NSF.
Manuscript received September 2, 1999; Accepted for publication January 21, 2000.
| LITERATURE CITED |
|---|
ALLSHIRE, R. C., 1996 Transcriptional silencing in fission yeast: a manifestation of higher order chromosome structure and functions, pp. 443–466 in Epigenetic Mechanisms of Gene Regulation, edited by V. E. RUSSO, R. A. MARTIENSSEN and A. D. RIGGS. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
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].
BAUDIN, A., O. OZIER-KALOGEROPOULOS, A. DENOUEL, F. LACROUTE, and C. CULLEN, 1993 A simple and efficient method for direct gene deletion in Saccharomyces cerevisiae.. Nucleic Acids Res. 14:3329-3330.
BELL, S. P. and B. STILLMAN, 1992 ATP-dependent recognition of eukaryotic origins of DNA replication by a multiprotein complex. Nature 357:128-134[Medline].
BELL, S. P., J. MITCHELL, J. LEBER, R. KOBAYASHI, and B. STILLMAN, 1995 The multidomain structure of Orc1p reveals similarity to regulators of DNA replication and transcriptional silencing. Cell 83:563-568[Medline].
BRAND, A. H., L. BREEDEN, J. ABRAHAM, R. STERNGLANZ, and K. NASMYTH, 1985 Characterization of a "silencer" in yeast: a DNA sequence with properties opposite to those of a transcriptional enhancer. Cell 41:41-48[Medline].
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].
BUCHMAN, A. R., W. J. KIMMERLY, J. RINE, and R. D. KORNBERG, 1988 Two DNA-binding factors recognize specific sequences at silencers, upstream activating sequences, autonomously replicating sequences, and telomeres in Saccharomyces cerevisiae.. Mol. Cell. Biol. 8:210-225[Abstract/Free Full Text].
BUCK, S. W. and D. SHORE, 1995 Action of a RAP1 carboxy-terminal silencing domain reveals an underlying competition between HMR and telomeres in yeast. Genes Dev. 9:370-384[Abstract/Free Full Text].
CALLEBAUT, I., J.-C. COURVALIN, and J.-P. MORNON, 1999 The BAH (bromo-adjacent homology) domain: a link between DNA methylation, replication and transcriptional regulation. FEBS Lett. 446:189-193[Medline].
CHIEN, C.-T., S. BUCK, R. STERNGLANZ, and D. SHORE, 1993 Targeting of SIR1 protein establishes transcriptional silencing at HM loci and telomeres in yeast. Cell 75:531-541[Medline].
DILLIN, A. and J. RINE, 1997 Separable functions of ORC5 in replication initiation and silencing in Sacchar
TOP
ABSTRACT
MATERIALS AND METHODS
