MGA2 and SPT23 Are Modifiers of Transcriptional Silencing in Yeast

Corresponding author: Scott G. Holmes, Department of Molecular Biology and Biochemistry, Lawn Ave., Wesleyan University, Middletown, CT 06459., sholmes{at}wesleyan.edu (E-mail)

Communicating editor: M. JOHNSTON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Transcriptional silencing at the HM loci and telomeres in yeast depends on several trans-acting factors, including Rap1p and the Sir proteins. The SUM1-1 mutation was identified by its ability to restore silencing to strains deficient in one or more of these trans-acting factors. The mechanism by which SUM1-1 bypasses the requirement for silencing proteins is not known. We identified four loci that when reduced in dosage in diploid strains increase the ability of SUM1-1 strains to suppress silencing defects. Two of the genes responsible for this effect were found to be MGA2 and SPT23. Mga2p and Spt23p were previously identified as functionally related transcription factors that influence chromatin structure. We find that deletion of MGA2 or SPT23 also increases the efficiency of silencing in haploid SUM1-1 strains. These results suggest that Mga2p and Spt23p are antagonists of silencing. Consistent with this proposal we find that deletion of MGA2 or SPT23 also suppresses the silencing defects caused by deletion of the SIR1 gene or by mutations in the HMR silencer sequences. However, we find that Mga2p and Spt23p can positively affect silencing in other contexts; deletion of either MGA2 or SPT23 decreases mating in strains bearing mutations in the HML-E silencer. Mga2p and Spt23p appear to be a novel class of factors that influence disparate pathways of transcriptional control by chromatin.

YEAST uses a position-effect mechanism to control the expression of genes specifying yeast cell type (HERSKOWITZ et al. 1992 ; THOMPSON et al. 1993 ; HOLMES et al. 1996 ). Yeast cell mating type is determined by the genes present at the expressed MAT locus, near the middle of chromosome III. Similar or identical genes are present at the ends of the same chromosome, at loci known as HML and HMR, but these genes are expressed only after a transposition event that moves them to the MAT locus. The genes at HML and HMR must be kept repressed to avoid haploid yeast cells from expressing both a and information in the same cell, which causes a nonmating phenotype. Silencing of HML and HMR is mediated by a repressive chromatin structure that by a variety of criteria is the yeast equivalent of metazoan heterochromatin.

Each of the silent mating-type loci are flanked by cis-acting "silencer" sequences, known as E and I, that are composed of binding sites for three protein factors: Rap1p, Abf1p, and origin recognition complex (ORC), a six-subunit complex. Rap1p and Abf1p are essential proteins that act as transcriptional activators at other locations in the genome, while ORC is an essential factor that also binds yeast origins of DNA replication. One function of the silencer-binding factors is to recruit a second set of factors, the Sir (silent information regulator) proteins, to HML and HMR. None of the four SIR genes is essential for viability, but null mutations in SIR2, SIR3, or SIR4 completely abolish silencing, while loss of the SIR1 gene causes a partial loss of silencing. The Sir2, Sir3, and Sir4 proteins can be found as a complex in yeast cells (HOLMES et al. 1997 ; MOAZED et al. 1997 ) and have been shown to be constituents of silent chromatin (HECHT et al. 1996 ; STRAHL-BOLSINGER et al. 1997 ). The Sir proteins are likely drawn to HML and HMR by associating with silencer-binding factors. Sir3p and Sir4p physically interact with the Rap1 protein (MORETTI et al. 1994 ; COCKELL et al. 1995 ), while Sir1 interacts with a subunit of the ORC and Sir4p (TRIOLO and STERNGLANZ 1996 ). The Sir3 and Sir4 proteins can also bind the N-terminal tails of histone H3 and histone H4 in vitro (HECHT et al. 1995 ). These observations provide a straightforward model for the formation of a nucleoprotein complex at HML and HMR that includes the Sir proteins and nucleosomes.

To gain insights on the mechanism of silencing and to identify additional factors that interact with the Sir proteins, genetic screens have been carried out to identify suppressors of the silencing defects caused by mutations in the SIR genes. The SUM1-1 mutation was isolated as a suppressor of the mating defect caused by a mutation in the SIR2 gene, the only such suppressor identified to date (KLAR et al. 1985 ). SUM1-1 was subsequently found to suppress a variety of mutations that lead to silencing defects, including deletions of SIR2, SIR3, or SIR4; specific mutations in the RAP1 gene or histone H4 genes; and partial deletions of the cis-acting silencer sequences (KLAR et al. 1985 ; LAURENSON and RINE 1991 ; CHI and SHORE 1996 ). The SUM1 gene has been cloned, and it codes for a novel protein. Deletion of the wild-type SUM1 gene causes mild defects in silencing and does not suppress sir mutants (CHI and SHORE 1996 ). Sum1p has been shown to be a site-specific DNA-binding factor that is involved in the repression of sporulation genes in mitotically dividing cells (XIE et al. 1999 ). Thus far the mechanism by which Sum1-1p bypasses the requirement for several factors essential for mating-type silencing is not clear.

To understand the mechanism by which Sum1-1p exerts its influence we screened for genes that can modify the ability of Sum1-1p to mediate silencing. We have identified four loci that increase mating efficiency in SUM1-1 strains when reduced in dosage. We report here the identification of two of these loci as the MGA2 and SPT23 genes. The Mga2 and Spt23 proteins were previously shown to be functionally related transcription factors that influence gene expression by affecting chromatin structure (BURKETT and GARFINKEL 1994 ; ZHANG et al. 1997 , ZHANG et al. 1999 ). Mga2p and Spt23p are found to suppress a variety of defects in the silencing apparatus, placing these proteins in a novel class of factors that influence disparate pathways of transcription control by chromatin.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Chromosome loss assay:
To identify loci that influence the mating ability of SUM1/SUM1-1 diploids we used the induced chromosome loss method of WAKEM and SHERMAN 1990 . A MAT set of strains, each containing FRT-URA3 sequences adjacent to the centromere of a different chromosome, was obtained from the Yeast Genetic Stock Center. This set of strains was made sir3::LEU2 using pAR78 (BRAUNSTEIN et al. 1993 ). SIR3 was reintroduced using pSH104, which is pRS414 (SIKORSKI and HIETER 1989 ) containing the complete SIR3 gene. The strains were then crossed to YSH196, maintaining selection for the URA3 marker. Selection for the SIR3 episome was relaxed, and we used a colony PCR assay to screen for strains that had lost pSH104. The resulting diploid strains were MAT/mat sir3::LEU2/sir3::LEU2 SUM1-1/SUM1. Selection for the URA3 marker was then relaxed, and cultured isolates of each diploid were spread on 5-fluoroorotic acid (5-FOA) plates to select for those that had lost the URA3 marker. A minimum of 200 5-FOA resistant colonies from each diploid were patched onto YPD plates. When grown out, these plates were replica-plated onto a lawn of MATa haploid tester strain YSH102. After overnight growth these plates were replica-plated to SD minimal media, which supports only the growth of strains that have mated. Individual patches were judged qualitatively to determine whether they displayed improved mating compared to a similar nonaneuploid diploid strain. We did not assay chromosome III, as loss of the MAT locus was predicted to always lead to loss of -mating. In addition, we found that the chromosome V strain was not made sterile by the deletion of SIR3, and thus this strain was also not assayed.

Strains and plasmids:
Strains used in this study are listed in Table 1. Telomere reporter strain YSH313 was made using plasmid ADH4UCA-IV, which is used to integrate the URA3 gene at the ADH4 locus, adjacent to telomere repeat sequences (GOTTSCHLING et al. 1990 ). To delete the SIR1 gene pES17 was used (STONE et al. 1991 ). Yeast strains deleted for MGA2 were made using either pML2 or pML3; yeast strains deleted for SPT23 were made using pML5 or pML6. pML1 was constructed first by amplifying a 1956-bp MGA2 fragment by PCR using primers SP61 and SP63. These primers add HindIII sites to the ends of the MGA2 fragment. This MGA2 fragment was subcloned into pBR322 at the HindIII site. pML2 was made by digesting pML1 with SacI and XbaI and then inserting the LEU2 gene. pML3 was made by digesting pML1 with SacI and XbaI and inserting the HIS3 gene. To delete the MGA2 gene, pML2 or pML3 was digested with HindIII and transformed into yeast. The resulting deletions remove promoter sequences and sequences coding for the first 180 amino acids of the Mga2 protein. pML4 was constructed by PCR amplifying a 2777-bp SPT23 fragment using primers SP96 and SP98. Primer SP96 adds a NotI site while primer SP98 adds an ApaI site. We then cloned this SPT23 fragment into pRS403 (SIKORSKI and HIETER 1989 ) at the NotI and ApaI sites. pML5 was made by digesting pML4 with XbaI and SalI and then inserting the LEU2 gene. To delete SPT23 we digested this plasmid with ApaI and NotI. This results in a deletion of promoter sequences and sequences coding for the first 590 amino acids of Spt23p. pML6 was made by inserting the HIS3 gene into pML4 between the XbaI and EcoRI sites. To use this plasmid for SPT23 knockouts we digested it with ApaI and NotI and transformed yeast, resulting in the deletion of promoter sequences and sequences coding for the first 361 amino acids of Spt23p. Primers used were as follows: SP61, GACTAAGCTTGGAAAGCAACGGAA TAGG; SP63, GGTGAAGCTTGATGGAATATGGGAATGG; SP96, GTACGTACGCGGCCGCCCTGCACAGTTCATCCTG; and SP98, GTACGTACGGGCCCCTTGGTCGTTCAAATGCG.

Mating assays:
Quantitative mating assays were performed as described (SHEI and BROACH 1995 ). For qualitative plate-mating assays two methods were used. In one method an equal number of cells from the strains to be tested were spotted on YPD plates and allowed to grow for one or two days. These patches were then replica-plated onto lawns of tester strains (either YSH102 or YSH103). Between 4 and 16 hr later these mating plates were replica-plated to minimal media and assayed for diploid formation. In the second method equal numbers of tester and testee cells were spotted together on YPD plates. Mating was allowed to proceed for between 4 and 16 hr, when the mating plates were replica-plated to minimal media. The data shown are representative of a minimum of four independent experiments.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Dosage modifiers of the SUM1-1 phenotype:
The SUM1-1 mutation was first characterized as recessive (KLAR et al. 1985 ), but later found to behave as a dominant mutation in some strain backgrounds (LAURENSON and RINE 1991 ; CHI and SHORE 1996 ). The present study, initiated before the gene was cloned, grew out of an attempt to map the location of the SUM1 gene. We assayed the ability of the SUM1-1 mutation to suppress the silencing defects caused by a deletion of the SIR3 gene (Fig 1). Deletion of the SIR3 gene abolishes silencing, leading to an inability of haploid strains to mate. As previously established, the SUM1-1 mutation restores silencing and allows mating in strains deleted for the SIR3 gene. In diploid strains heterozygous for SUM1-1 we found that the ability of Sum1-1p to suppress the SIR3 deletion was reduced, as indicated by a reduction in mating efficiency compared to SUM1-1 sir3 haploid strains. We concluded that the SUM1-1 allele is not completely dominant to the wild-type allele. Our mapping strategy relied on the premise that removal of the wild-type copy of the SUM1 gene would increase the mating ability of diploids containing both SUM1 and SUM1-1. To remove the wild-type copy of SUM1 from the heterozygous diploid we used the induced chromosome loss method of WAKEM and SHERMAN 1990 (Fig 2). The SUM1-1 haploid strain was crossed to a set of strains that have the FRT sequence integrated near the centromere of a single chromosome. FRT derives from the endogenous yeast plasmid, the 2µ circle. FRT sequences are recognized by a 2µ circle-encoded recombinase, the Flp protein. Flp normally recognizes two FRT sites present on the 2µ circle and mediates recombination between them. Recognition of a single centromere-linked FRT site by Flp often results in loss of the chromosome containing the introduced FRT site (FALCO et al. 1982 ; WAKEM and SHERMAN 1990 ). The FRT-marked strains lack the 2µ circle, so the marked chromosome is stable. When crossed to a cir⁺ strain containing the 2µ circle, such as the SUM1-1 haploid in our experiments, Flp recombinase frequently induces loss of the chromosome, creating a diploid strain aneuploid for a single chromosome (Fig 2).

View larger version (27K): In this window In a new window Download PPT slide   Figure 1. Mating ability of SUM1-1 strains. A partial genotype of each strain is shown, as well as the results of qualitative and quantitative mating assays. Values are expressed relative to YSH399.

View larger version (11K): In this window In a new window Download PPT slide   Figure 2. Mapping modifiers of silencing by the induced chromosome loss method. A cir0 strain (lacking the 2µ circle) with FRT-URA3 linked to the centromere of a single chromosome is crossed to a cir+ strain. Flp protein, coded for by the 2µ circle, acts on the FRT site to destabilize the chromosome, resulting in either loss of the chromosome or loss of the FRT-URA3 linked chromosome arm. A strain with a marked chromosome IX, allowing the different loss events to be distinguished, is shown as a specific example.

We crossed the set of FRT-marked strains to the SUM1-1 haploid and then assayed mating of the aneuploid strains. We predicted that when the marked chromosome containing the wild-type SUM1 gene was lost, mating would improve. The URA3 gene is integrated adjacent to the FRT site in each marked strain. We selected for strains that had lost their marked chromosome by plating on media containing 5-FOA, which selects for ura3^- cells (BOEKE et al. 1987 ). Loss of the URA3 marker in these strains represents three possible events: loss of the entire marked chromosome, loss of the marked chromosome arm, or loss of the marker without chromosome loss (WAKEM and SHERMAN 1990 ). 5-FOA resistant isolates were screened by a qualitative patch mating assay for a change in mating behavior. The results of this experiment are presented in Table 2. At least 200 5-FOA resistant isolates of each diploid were examined to improve the probability of assaying the result of true chromosome loss events. We were able to screen 14 of the 16 yeast chromosomes by this approach (see MATERIALS AND METHODS). Unexpectedly, loss of 4 individual chromosomes significantly improved mating in this assay. This indicated that genes besides SUM1 could influence the SUM1-1 phenotype when reduced in dosage; as mating improved when the dosage of these genes was reduced, it appeared that we were identifying antagonists of silencing. We did not identify any locus that decreased mating efficiency by this method. In addition, we did not identify chromosome IV, the location of the SUM1 gene (CHI and SHORE 1996 ), by this approach.

The change in mating we observed varied among different diploids both in the frequency of isolates scored as increased maters (Table 2), as well as in the degree of improvement in mating efficiency. The difference in frequency is not a clear indication that we were observing a variable level of penetrance, as the true frequency of chromosome loss, as opposed to simple loss of the URA3 marker, was not determined. We elected to continue our characterization of the dosage modifier present on chromosome IX, as it qualitatively produced the greatest improvement in mating efficiency.

By mapping the SUM1-1 mutation to markers on chromosome IX using tetrad analysis we confirmed that SUM1 did not reside on chromosome IX in our strain background (not shown). Since the induced chromosome loss can at some frequency cause loss of only one chromosome arm (WAKEM and SHERMAN 1990 ), we used a strain with markers on both arms of chromosome IX to more narrowly define the location of the modifying gene (Fig 2). In this marked diploid strain loss of either arm of chromosome IX is detected by uncovering histidine or lysine auxotrophies. 5-FOA resistant strains from this diploid were scored for mating ability, as well as the ability to grow on media lacking histidine or lysine. We find that mating efficiency improved in only 13% of His^- Lys⁺ isolates, while 83% of His^- Lys^- auxotrophs exhibit improved mating. This indicates that the dosage modifier is on the same chromosome arm as LYS1, the short arm of chromosome IX.

MGA2 and SPT23 antagonize silencing in SUM1-1 strains:
The yeast genome sequencing project reveals 39 potential open reading frames of at least 100 amino acids on the short arm of chromosome IX (CHURCHER 1997 ). This group includes the MGA2 gene. MGA2 was initially isolated in a screen for high expression suppressors of a transcription defect caused by a null mutation in the SNF2 gene (ZHANG et al. 1997 ). Snf2p is a key component of the Swi/Snf nucleosome remodeling complex, which has a role in activating the transcription of many genes (KORNBERG and LORCH 1999 ; TRAVERS 1999 ). Mga2p was subsequently shown to positively influence the transcription of several genes (ZHANG et al. 1997 , ZHANG et al. 1999 ) and to activate transcription when tethered to DNA (ZHANG et al. 1997 ). Due to its involvement in regulation of transcription and its association with chromatin, we considered MGA2 to be a candidate for the gene responsible for the chromosome IX dosage effects. To test this we deleted the MGA2 gene in the SUM1/SUM1-1 heterozygous diploid strain. We found that deletion of MGA2 in this background increased mating efficiency (Fig 3A). We conclude that MGA2 is likely to be the chromosome IX locus modifying suppression by SUM1-1.

View larger version (42K): In this window In a new window Download PPT slide   Figure 3. Mga2p and Spt23p antagonize silencing in SUM1-1 strains. (A) The mating ability of congenic diploid strains heterozygous for the SUM1-1 mutation but differing in their MGA2 gene dosage is compared. (B) Deletion of MGA2 or SPT23 increases mating in haploid SUM1-1/sir2 strains. The mating ability of congenic haploid strains differing only at the indicated loci is shown.

The MGA2 gene is not essential for viability (ZHANG et al. 1997 ). Therefore we were able to determine if deletion of MGA2 would also improve mating efficiency in SUM1-1 haploid strains. When we compared the mating efficiency of a SUM1-1 haploid with a congenic SUM1-1 mga2 strain, we observed that absence of MGA2 increased silencing in SUM1-1 haploids (Fig 3B). Therefore, either reduction of MGA2 gene dosage or the complete absence of Mga2p increases mating efficiency in SUM1-1 backgrounds.

MGA2 has significant structural and functional similarity to another yeast gene, SPT23 (ZHANG et al. 1999 ). SPT23 was first identified as a high-copy suppressor of the transcription defects caused by insertion of a transposable element in promoters of the HIS4 and LYS2 genes (BURKETT and GARFINKEL 1994 ). Like MGA2, SPT23 overexpression also has the ability to suppress some of the transcriptional defects caused by deletion of SNF2 (ZHANG et al. 1997 ). SPT23 is also nonessential, but strains deleted for both MGA2 and SPT23 are inviable, indicating that the Spt23 and Mga2 proteins have a redundant, essential function (ZHANG et al. 1997 , ZHANG et al. 1999 ). SPT23 is located on chromosome XI, one of the four chromosomes identified by our screen for gene dosage modifiers. This raised the possibility that SPT23 was the gene responsible for the chromosome XI dosage effects we observed. We determined if deletion of SPT23 increased silencing in the haploid SUM1-1 background. These results are shown in Fig 3B. We find that deletion of SPT23 significantly improves mating in the SUM1-1 background. Thus, by this assay SPT23 also appears to be an antagonist of silencing and is likely to be the chromosome XI modifier.

Loss of MGA2 or SPT23 suppresses SIR1 deletions and HMR silencer mutations:
We next wished to determine if MGA2 and SPT23 were general antagonists of silencing or if their action was specific to the SUM1-1 background. Silencing is extremely efficient in wild-type cells, so we did not expect to be able to measure an increase in silencing as a result of deletion of MGA2 or SPT23 in a completely wild-type strain. Therefore, we examined the influence of MGA2 and SPT23 in strains compromised for silencing due to mutations in the cis-acting silencer sequences or due to deletions of the SIR1 gene. Fig 4 shows the results of an experiment in which we deleted SPT23 in a strain lacking SIR1. Deletion of SIR1 in a wild-type strain causes a marked decrease in silencing efficiency that leads to a decrease in mating. We find that deletion of SPT23 in the sir1 background suppresses the mating defect, restoring mating efficiency to wild-type levels. As an additional test of the specificity of this effect we conducted this experiment in MAT strains, which must silence HMRa to allow mating, and MATa strains, which must silence HML to mate. We find that deletion of SPT23 is able to suppress the deletion of SIR1 in either of these backgrounds.

View larger version (37K): In this window In a new window Download PPT slide   Figure 4. Loss of SPT23 suppresses deletion of the SIR1 gene. Mating assays were performed on a set of strains differing only at the MAT, SIR1, and SPT23 loci. A partial genotype is shown for each strain.

We next assayed the ability of Mga2p and Spt23p to influence silencing in strains bearing mutations in the HMR silencer sequences. We used strains containing the yeast ADE2 gene integrated at the HMR locus, placing its transcription under control of the HMR silencer sequences. Yeast strains that do not express the Ade2 protein accumulate a red pigment. Therefore, a change in silencing in this strain produces detectable changes in colony color. These strains also have mutations in the HMR-E silencer, making it easier to measure changes in silencing efficiency. Deletion of MGA2 or SPT23 leads to a clear increase in the frequency of red colonies in these strains (Fig 5A). These deletions have no effect on colony color when the ADE2 gene is at its normal location (not shown). As this determination of silencing efficiency is independent of mating proficiency, this assay distinguishes mutations that affect transcriptional silencing from those that influence the mating process. Therefore, this result strongly suggests that deletion of MGA2 or SPT23 causes an increase in transcriptional silencing at HMR. We observed that many colonies produced by strains lacking the MGA2 gene did not exhibit a uniform increase in red pigment, but instead showed red and white sectors (Fig 5A). This suggests that the absence of Mga2p or Spt23p does not increase silencing to an equal extent in all cells, but instead increases the probability of establishing a heritable expression state and/or reinforces the inheritance of that state.

View larger version (93K): In this window In a new window Download PPT slide   Figure 5. Spt23p and Mga2p can suppress or exacerbate silencer mutations. (A) Colony color assay for silencing at HMR. Transcription of the ADE2 gene is placed under control of the HMR silencer sequences. A decrease in the expression of ADE2 leads to an increase in red color. A partial genotype of each strain is shown. hmrA strains lack the HMR-E ACS element (the ORC binding site); hmrB strains lack the HMR-E Abf1p binding site. (B) Plate-mating assays were performed on a set of strains bearing mutations in the HML silencers. All strains lack functional HML-I silencer sequences; HML-E (79-113) strains lack the Rap1p binding site; HML-E (90-95) strains lack the ACS element.

Mga2p and Spt23p can act as positive silencing factors at HML:
We also examined the influence of Mga2p and Spt23p on HML silencing in SUM⁺ SIR⁺ strains. For these experiments we used two different strains compromised for silencing due to mutations in the HML-E silencer sequences. Each strain contains a deletion of the HML-I silencer; strain DMY26 also contains a mutation in the HML-E Rap1p binding site, while DMY116 contains a mutation in the HML-E ACS sequence. If Mga2p and Spt23p are general antagonists of silencing we expected to observe increases in silencing efficiency when their genes are deleted in these backgrounds. Surprisingly, deletion of either MGA2 or SPT23 in these strains led to a decrease in mating efficiency (Fig 5B). As absence of Mga2p or Spt23p leads to a decrease in mating efficiency, this indicates that these factors are acting as positive silencing factors at this locus.

An additional method to assay subtle effects on silencing is to monitor telomere position effect. While silencing at telomeres relies on most of the same factors as mating-type silencing, including Sir2p, Sir3p, and Sir4p, repression is weaker than at the silent mating-type loci and is more sensitive to minor perturbations in the silencing machinery. Therefore, we deleted either MGA2 or SPT23 in a strain bearing a reporter of telomere position effect. This strain contains the URA3 gene integrated adjacent to telomere repeat sequences. We performed serial dilution assays to determine the level of telomeric silencing in strains deleted for MGA2 and SPT23, as well as in wild-type control strains. Cells were plated on complete media, media lacking uracil, and media containing 5-FOA. The results of our experiment with SPT23 are shown in Fig 6. We find that deletion of SPT23 has little or no effect on telomere position effect by this assay; we observed similar results in a strain deleted for MGA2 (not shown).

View larger version (12K): In this window In a new window Download PPT slide   Figure 6. Spt23p does not influence telomere position effect. Expression of a telomere-proximal URA3 gene was assayed in a wild-type strain and a congenic strain lacking the SPT23 gene. Cells were grown to high density in nonselective media and then serially diluted on the indicated plates.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We identified the Mga2 and Spt23 proteins as modifiers of the SUM1-1 phenotype; deletion of either the MGA2 or SPT23 gene in the SUM1-1 background causes an increase in the efficiency of silencing. Loss of Mga2p or Spt23p was further shown to suppress the mating defects caused by mutations in the HMR-E silencer and deletions of the SIR1 gene. These results are generally consistent with the prior characterization of the Mga2 and Spt23 proteins (BURKETT and GARFINKEL 1994 ; ZHANG et al. 1997 , ZHANG et al. 1999 ). First, Mga2p and Spt23p are structurally similar and appear to have redundant but not completely overlapping functions in the yeast cell. In all of the contexts we tested, deletion of either MGA2 or SPT23 produced similar phenotypes, although the effects of SPT23 deletions were generally of greater magnitude. Second, most of our results are in accord with the prior assignment of Mga2p and Spt23p as transcriptional activators. An activating function regulated by Mga2p/Spt23p may compete with the silencing machinery to determine the expression state at the silent mating-type loci. In this model the effect of Mga2p and Spt23p does not have a noticeable effect in wild-type strains, but in strains with weakened silencing the Mga2p/Spt23p regulated activity is able to exert a significant antagonizing influence.

However, Mga2 and Spt23p have a complex effect on silencing at HML that indicates the assignment of Mga2p and Spt23p as transcriptional activators is not clear cut. Mga2p and Spt23p act as antagonists of silencing at HML in a strain deleted for the SIR1 gene but appear to act as positive silencing factors in strains with HML silencer mutations. These seemingly contradictory results could reflect quantitative or qualitative differences in the strains we examined. First, the strains bearing mutations in the HML silencer sequences retain significantly less transcriptional repression than the sir1 strains used in this study. Perhaps the interaction of the Mga2p/Spt23p-dependent activity with the silencing machinery differs depending on the efficiency of silencing in the strain background. Alternatively, the silencing defect caused by deleting the SIR1 gene could be qualitatively different from the defect caused by mutations in the HML-E silencer sequences. The possibility that Mga2p and Spt23p have activating or repressing functions depending on context has a precedent in yeast. For instance, the silencer-binding factors Abf1p and Rap1p have activating roles at the promoters of many genes (SHORE 1994 ), yet have a key role in establishing repression at the HML and HMR silencers. Interestingly, deletion of MGA2 and SPT23 did not affect the efficiency of silencing of a telomere-linked gene, placing Mga2p and Spt23p in a small group of proteins that include Sir1p (APARICIO et al. 1991 ) and Sum1p (CHI and SHORE 1996 ) that preferentially affect silencing at the HM loci.

Our results reinforce and extend the conclusion that Mga2 and Spt23 exert their influence on transcription by regulating chromatin structure (BURKETT and GARFINKEL 1994 ; ZHANG et al. 1997 , ZHANG et al. 1999 ). Overexpression of Mga2p or Spt23p was previously shown to suppress the loss of Snf2p, part of the Swi/Snf nucleosome remodeling complex that has a role in activating gene expression. Spt23p can also suppress the loss of Snf5p, another component of the Swi/Snf factor, and was proposed to function in a complex that partially substitutes for the absence of Swi/Snf activity (ZHANG et al. 1997 ). Mga2 and Spt23 are selective in their ability to suppress Snf2 mutations and appear to affect a relatively small number of genes (ZHANG et al. 1999 ), indicating that they are not likely to be basal transcription factors. The modes of altering chromatin structure mediated by Swi/Snf and the Sir proteins have thus far been independently defined; members of the Swi/Snf complex have not been shown to influence mating-type silencing, while the Sir protein's effects on transcription have been limited to the HM loci, telomeres, and the rDNA locus. The identification of Mga2p and Spt23p as proteins that influence both these pathways may indicate that they are novel factors that have a more global influence on chromatin-mediated effects on transcription.

The Swi/Snf complex has been shown to alter nucleosome mobility in vivo and in vitro and is thought to activate transcription by removing nucleosomes from promoter sequences to allow entry of the basal transcriptional machinery (KORNBERG and LORCH 1999 ; TRAVERS 1999 ). Given that high levels of Mga2p or Spt23p suppress defects caused by loss of Swi/Snf components, a straightforward prediction for Mga2p and Spt23p's function is that they affect transcription by influencing nucleosome mobility. Silencing may involve a localized modification of nucleosomes (BRAUNSTEIN et al. 1993 ; TANNY et al. 1999 ; IMAI et al. 2000 ; LANDRY et al. 2000 ), and the efficiency of these modifications could be influenced by alterations in nucleosome mobility. Alternatively, changes in nucleosome mobility may influence the ability of a specific chromatin structure to be preserved during DNA replication and mitosis, a feature that contributes to the high efficiency of silencing (PILLUS and RINE 1989 ). These proposals lead to a hypothesis that Swi/Snf or related factors may influence specific aspects of silencing. However, it is also possible that Mga2p and Spt23p influence HML and HMR transcription indirectly by altering the expression of an independent silencing factor.

The multicomponent complexes involved in chromatin modification may be particularly sensitive to gene dosage effects (LOCKE et al. 1988 ). For example, position-effect variegation, a heterochromatin-mediated gene-silencing phenomenon in Drosophila, is known to be affected by inceasing or decreasing the gene dosage of numerous enhancers and suppressors of variegation (HENIKOFF 1996 ). The two contexts to which the Mga2 and Spt23 proteins have been linked are also known to be sensitive to gene dosage effects. For instance, yeast mating-type silencing can be affected by subtle alterations in the copy number of the SIR1 gene (STONE et al. 1991 ) or the SIR4 gene (SUSSEL et al. 1993 ), while the SPT group of mutants includes strains in which the gene dosage of H2B is reduced from two to one (CLARK-ADAMS et al. 1988 ). As our screen did not identify all genes known to affect mating-type silencing when reduced in dosage, this may reflect some level of specificity of the modifiers for the SUM1-1/sir3 background. In addition, we failed to identify chromosome IV, the location of the wild-type copy of SUM1, as predicted by the premise of our mapping strategy. This suggests that the decrease in mating in the SUM1/SUM1-1 heterozygote is due to an increase in the dosage of wild-type modifying genes and not to incomplete dominance of the SUM1-1 allele. Here we discovered four loci that behave as modifiers of the SUM1-1 phenotype and defined two of the genes responsible. The two remaining modifiers await further characterization.

The induced chromosome loss method provided a rapid means to identify dosage modifiers of mating-type silencing in diploid cells. While similar screens have not been commonly performed in yeast, dosage modifiers have been routinely isolated in obligate diploid organisms such as Drosophila. Screening for dosage modifiers offers the advantages of being able to discover phenotypes caused by loss-of-function mutations in essential genes and, in some contexts, may allow identification of subtle gene interactions not detectable by other means. The use of marked transposons to create mutations would simplify the isolation of dosage mutants and allow rapid identification of the affected gene (CHUN and GOEBL 1996 ; ROSS-MACDONALD et al. 1999 ). Heterozygous diploid strains with mutations in essential genes are also being generated in yeast by systematic gene deletion projects (WINZELER 1999 ). Our results suggest that screening these strains for specific mutant phenotypes will yield interesting new genetic interactions. Our identification of Mga2p and Spt23p as silencing factors by using this approach should help direct new experiments toward understanding in detail the mechanism of their effects in the yeast cell.

	ACKNOWLEDGMENTS

S.G.H. gratefully acknowledges James Broach, in whose lab this project was initiated. We also thank Danielle Margalit for assistance in strain construction, other members of the lab for support and helpful discussions, and David Shore, James Broach, Daniel Gottschling, and Rolf Sternglanz for providing strains and plasmids. This work was supported by research project grant RPG-98-351-01-MGO from the American Cancer Society.

Manuscript received February 25, 2000; Accepted for publication June 29, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

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., J. TRUEHEART, G. NATSOULIS, and G. R. FINK, 1987  5-Fluoroorotic acid as a selective agent in yeast molecular genetics. Methods Enzymol. 154:164-175[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].

BURKETT, T. J. and D. J. GARFINKEL, 1994  Molecular characterization of the SPT23 gene: a dosage-dependent suppressor of Ty-induced promoter mutations from Saccharomyces cerevisiae.. Yeast 10:81-92[Medline].

CHI, M.-H. and D. SHORE, 1996  SUM1-1, a dominant suppressor of SIR mutations in Saccharomyces cerevisiae, increases transcriptional silencing at telomeres and HM mating-type loci and decreases chromosome stability. Mol. Cell. Biol. 16:4281-4294[Abstract].

CHUN, K. T. and M. G. GOEBL, 1996  The identification of transposon-tagged mutations in essential genes that affect cell morphology in Saccharomyces cerevisiae. Genetics 142:39-50[Abstract].

CHURCHER, C. E. A., 1997  The nucleotide sequence of Saccharomyces cerevisiae chromosome IX. Nature 387(Suppl.):84-87[Medline].

CLARK-ADAMS, C. D., D. NORRIS, M. A. OSLEY, J. S. FASSLER, and F. WINSTON, 1988  Changes in histone gene dosage alter transcription in yeast. Genes Dev. 2:150-159[Abstract/Free Full Text].

COCKELL, M., F. PALLADINO, T. LAROCHE, G. KYRION, and C. LIU et al., 1995  The carboxy termini of Sir4 and Rap1 affect Sir3 localization: evidence for a multicomponent complex required for yeast telomeric silencing. J. Cell Biol. 129:909-924[Abstract/Free Full Text].

FALCO, S. C., Y. LI, J. R. BROACH, and D. BOTSTEIN, 1982  Genetic properties of chromosomally integrated 2m DNA in yeast. Cell 29:573-584[Medline].

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

HECHT, A., T. LAROCHE, S. STRAHL-BOLSINGER, S. M. GASSER, and M. GRUNSTEIN, 1995  Histone H3 and H4 N-termini interact with SIR3 and SIR4 proteins: a molecular model for the formation of heterochromatin in yeast. Cell 80:583-592[Medline].

HECHT, A., S. STRAHL-BOLSINGER, and M. GRUNSTEIN, 1996  Spreading of transcriptional repressor SIR3 from telomeric heterochromatin. Nature 383:92-96[Medline].

HENIKOFF, S., 1996 Position effect variegation in Drosophila: recent progress, pp. 319–334 in Epigenetic Mechanisms of Gene Regulation, edited by V. E. A. RUSSO, R. MARTIENSSEN and A. RIGGS. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

HERSKOWITZ, I., J. RINE and J. STRATHERN, 1992 Mating-type determination and mating-type interconversion in Saccharomyces cerevisiae, pp. 583–656 in The Molecular and Cellular Biology of the Yeast Saccharomyces, edited by E. W. JONES, J. R. PRINGLE and J. R. BROACH. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

HOLMES, S. G., M. BRAUNSTEIN and J. R. BROACH, 1996 Transcriptional silencing of the yeast mating type genes, pp. 467–487 in Epigenetic Mechanisms of Gene Regulation, edited by V. E. A. RUSSO, R. MARTIENSSEN and A. RIGGS. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

HOLMES, S. G., A. B. ROSE, K. STEUERLE, E. SAEZ, and S. SAYEGH et al., 1997  Hyperactivation of the silencer proteins Sir2p and Sir3p causes chromosome loss. Genetics 145:605-614[Abstract].

IMAI, S., C. M. ARMSTRONG, M. KAEBERLEIN, and L. GUARENTE, 2000  Transcriptional silencing and longevity protein Sir2 is an NAD-dependent histone deacetylase. Nature 403:795-800[Medline].

KLAR, A. J., S. N. KAKAR, J. M. IVY, J. B. HICKS, and G. P. LIVI et al., 1985  SUM1, an apparent positive regulator of the cryptic mating-type loci in Saccharomyces cerevisiae.. Genetics 111:745-758[Abstract/Free Full Text].

KORNBERG, R. D. and Y. LORCH, 1999  Chromatin-modifying and -remodeling complexes. Curr. Opin. Genet. Dev. 9:148-151[Medline].

LANDRY, J., A. SUTTON, S. T. TAFROV, R. C. HELLER, and J. STEBBINS et al., 2000  The silencing protein SIR2 and its homologs are NAD-dependent protein deacetylases. Proc. Natl. Acad. Sci. USA 97:5807-5811[Abstract/Free Full Text].

LAURENSON, P. and J. RINE, 1991  SUM1-1: a suppressor of silencing defects in Saccharomyces cerevisiae. Genetics 129:685-696[Abstract].

LOCKE, J., M. A. KOTARSKI, and K. D. TARTOF, 1988  Dosage-dependent modifiers of position effect variegation in Drosophila and a mass action model that explains their effects. Genetics 120:181-198[Abstract/Free Full Text].

MOAZED, D., A. KISTLER, A. AXELROD, J. RINE, and A. JOHNSON, 1997  Silent information regulator protein complexes in Saccharomyces cerevisiae: a SIR2/SIR4 complex and evidence for a regulatory domain in SIR4 that inhibits its interaction with SIR3. Proc. Natl. Acad. Sci. USA 94:2186-2191[Abstract/Free Full Text].

MORETTI, P., K. FREEMAN, L. COODLY, and D. SHORE, 1994  Evidence that a complex of SIR proteins interacts with the silencer and telomere-binding protein RAP1. Genes Dev. 8:2257-2269[Abstract/Free Full Text].

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

ROSS-MACDONALD, P., A. SHEEHAN, C. FRIDDLE, G. S. ROEDER, and M. SNYDER, 1999  Transposon mutagenesis for the analysis of protein production, function, and localization. Methods Enzymol. 303:512-532[Medline].

SHEI, G. J. and J. R. BROACH, 1995  Yeast silencers can act as orientation-dependent gene inactivation centers that respond to environmental signals. Mol. Cell. Biol. 15:3496-3506[Abstract].

SHORE, D., 1994  RAP1: a protean regulator in yeast. Trends Genet. 10:408-412[Medline].

SIKORSKI, R. and P. HIETER, 1989  A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae.. Genetics 122:19-27[Abstract/Free Full Text].

STONE, E. M., M. J. SWANSON, A. M. ROMEO, J. B. HICKS, and R. STERNGLANZ, 1991  The SIR1 gene of Saccharomyces cerevisiae and its role as an extragenic suppressor of several mating-defective mutants. Mol. Cell. Biol. 11:2253-2262[Abstract/Free Full Text].

STRAHL-BOLSINGER, S., A. HECHT, K. LUO, and M. GRUNSTEIN, 1997  SIR2 and SIR4 interactions differ in core and extended telomeric heterochromatin in yeast. Genes Dev. 11:83-93[Abstract/Free Full Text].

SUSSEL, L., D. VANNIER, and D. SHORE, 1993  Epigenetic switching of transcriptional states: cis- and trans-acting factors affecting establishment of silencing at the HMR locus in Saccharomyces cerevisiae.. Mol. Cell. Biol. 13:3919-3928[Abstract/Free Full Text].

TANNY, J. C., G. J. DOWD, J. HUANG, H. HILZ, and D. MOAZED, 1999  An enzymatic activity in the yeast Sir2 protein that is essential for gene silencing. Cell 99:735-745[Medline].

THOMPSON, J. S., A. HECHT, and M. GRUNSTEIN, 1993  Histones and the regulation of heterochromatin in yeast. Cold Spring Harbor Symp. Quant. Biol. 58:247-256[Medline].

TRAVERS, A., 1999  An engine for nucleosome remodeling. Cell 96:311-314[Medline].

TRIOLO, T. and R. STERNGLANZ, 1996  Role of interactions between the origin recognition complex and SIR1 in transcriptional silencing. Nature 381:251-253[Medline].

WAKEM, L. P. and F. SHERMAN, 1990  Chromosomal assignment of mutations by specific chromosome loss in the yeast Saccharomyces cerevisiae.. Genetics 92:803-821.

WINZELER, E. A., 1999  Functional characterization of the Saccharomyces cerevisiae genome by gene deletion and parallel analysis. Science 285:901-906[Abstract/Free Full Text].

XIE, J., M. PIERCE, V. GAILUS-DURNER, M. WAGNER, and E. WINTER et al., 1999  Sum1 and Hst1 repress middle sporulation-specific gene expression during mitosis in Saccharomyces cerevisiae.. EMBO J. 18:6448-6454[Medline].

ZHANG, S., T. J. BURKETT, I. YAMASHITA, and D. J. GARFINKEL, 1997  Genetic redundancy between SPT23 and MGA2: regulators of Ty-induced mutations and Ty1 transcription in Saccharomyces cerevisiae. Mol. Cell. Biol. 17:4718-4729[Abstract].

ZHANG, S., Y. SKALSKY, and D. J. GARFINKEL, 1999  MGA2 or SPT23 is required for transcription of the delta9 fatty acid desaturase gene, OLE1, and nuclear membrane integrity in Saccharomyces cerevisiae. Genetics 151:473-483[Abstract/Free Full Text].

This article has been cited by other articles:

				K. L. Auld, A. L. Hitchcock, H. K. Doherty, S. Frietze, L. S. Huang, and P. A. Silver The Conserved ATPase Get3/Arr4 Modulates the Activity of Membrane-Associated Proteins in Saccharomyces cerevisiae Genetics, September 1, 2006; 174(1): 215 - 227. [Abstract] [Full Text] [PDF]

				M. Matecic, K. Martins-Taylor, M. Hickman, J. Tanny, D. Moazed, and S. G. Holmes New Alleles of SIR2 Define Cell-Cycle-Specific Silencing Functions Genetics, August 1, 2006; 173(4): 1939 - 1950. [Abstract] [Full Text] [PDF]

				L. A. Papacs, Y. Sun, E. L. Anderson, J. Sun, and S. G. Holmes REP3-Mediated Silencing in Saccharomyces cerevisiae Genetics, January 1, 2004; 166(1): 79 - 87. [Abstract] [Full Text] [PDF]

				Y. Huang Transcriptional silencing in Saccharomyces cerevisiae and Schizosaccharomyces pombe Nucleic Acids Res., April 1, 2002; 30(7): 1465 - 1482. [Abstract] [Full Text] [PDF]

				E. Lebrun, E. Revardel, C. Boscheron, R. Li, E. Gilson, and G. Fourel Protosilencers in Saccharomyces cerevisiae Subtelomeric Regions Genetics, May 1, 2001; 158(1): 167 - 176. [Abstract] [Full Text]