Sum1 and Hst1 repress middle sporulation‐specific gene expression during mitosis in Saccharomyces cerevisiae

Jianxin Xie, Michael Pierce, Valérie Gailus‐Durner, Marisa Wagner, Edward Winter, Andrew K. Vershon

Author Affiliations

  1. Jianxin Xie1,
  2. Michael Pierce1,
  3. Valérie Gailus‐Durner1,
  4. Marisa Wagner2,
  5. Edward Winter2 and
  6. Andrew K. Vershon*,1
  1. 1 Waksman Institute and Department of Molecular Biology and Biochemistry, 190 Frelinghuysen Road, Rutgers University, Piscataway, NJ, 08854‐8020, USA
  2. 2 Department of Biochemistry and Molecular Pharmacology, Thomas Jefferson University, Philadelphia, PA, 19107, USA
  1. *Corresponding author. vershon{at}


Meiotic development in yeast is characterized by the sequential induction of temporally distinct classes of genes. Genes that are induced at the middle stages of the pathway share a promoter element, termed the middle sporulation element (MSE), which interacts with the Ndt80 transcriptional activator. We have found that a subclass of MSEs are strong repressor sites during mitosis. SUM1 and HST1, genes previously associated with transcriptional silencing, are required for MSE‐mediated repression. Sum1 binds specifically in vitro to MSEs that function as strong repressor sites in vivo. Repression by Sum1 is gene specific and does not extend to neighboring genes. These results suggest that mechanisms used to silence large regions of chromatin may also be used to regulate the expression of specific genes during development. NDT80 is regulated during mitosis by both the Sum1 and Ume6 repressors. These results suggest that progression through sporulation may be controlled by the regulated competition between the Sum1 repressor and Ndt80 activator at key MSEs.


Meiosis and sporulation in Saccharomyces cerevisiae are characterized by the sequential expression of large numbers of genes (Kupiec et al., 1997; Chu et al., 1998). Sporulation‐specific genes are expressed exclusively during this developmental program and can be broadly divided into early, middle or late temporal categories. Genome‐wide analysis of the transcriptional program during sporulation revealed that >300 genes are induced during the middle period of sporulation around the time of meiotic chromosome segregation (Chu et al., 1998). Mid‐sporulation genes are activated by the Ndt80 transcription factor that binds to a conserved sequence (gNCRCAAAA/T) termed the middle sporulation element (MSE) found in most middle gene promoters (Hepworth et al., 1995, 1998; Ozsarac et al., 1997; Chu and Herskowitz, 1998). NDT80 is itself a mid‐sporulation‐specific gene that has been implicated as a target of checkpoint controls that operate during meiosis. MSEs are also present in several B‐type cyclin promoters (CLB1, and CLB3–6) and Ndt80 has a role in the expression of these genes during meiosis (Chu and Herskowitz, 1998; Hepworth et al., 1998). These observations have led to the idea that Ndt80 plays a central role in the transcriptional cascade and in coordinating checkpoint signals that regulate the progression through meiosis and sporulation.

One of the genes that is activated by Ndt80 is SMK1, which encodes a mid‐sporulation‐specific MAP kinase homolog that is required for spore wall morphogenesis (Krisak et al., 1994; Wagner et al., 1997, 1999). The SMK1 promoter contains an MSE that is required not only for transcriptional activation during mid‐sporulation but also for repression of the Abf1‐dependent activator site during vegetative growth and early meiosis (Pierce et al., 1998). Here we show that MSEs found in other mid‐sporulation genes can also function to repress transcription during vegetative growth, and have identified SUM1 and HST1, genes previously associated with transcriptional silencing, as being involved in this repression.


A subclass of MSEs function as mitotic repression sites

The MSE in the SMK1 promoter represses expression during vegetative growth and early sporulation (Pierce et al., 1998). We were interested in whether other MSEs also function as repressor sites. MSEs found in the promoters of several sporulation‐specific genes were used to replace the URS1 site in a HOP1‐LacZ reporter plasmid. These constructs were assayed for their ability to regulate transcription during vegetative growth and mid‐sporulation, and under conditions in which NDT80 is ectopically expressed (Figure 1). The MSEs from the SMK1 and BBP1 promoters function as strong vegetative repressor sites and reduce expression of the promoter by ≥20‐fold. The SPR3 MSEs also repress transcription but are weaker than the SMK1 site. In contrast, the SPS4, DIT1 and CLB6 MSEs do not repress expression detectably. The NDT80 promoter contains two MSEs; one (at −78) is a strong repressor site, while the other (at −221) is not a repressor site in the assay. Therefore, MSEs found in a variety of promoters that are expressed during mid‐sporulation are able to repress gene expression in vegetative cells.

Figure 1.

Repression and activation by MSEs in a heterologous promoter. Oligonucleotides containing MSEs from the indicated promoters were cloned into the XhoI site of the HOP1‐LacZ reporter vector, pAV124 (Pierce et al., 1998). The mitotic repression activity of each construct was measured in strain W303A. Meiotic activation was measured in strain RSX2‐7B 10 h after transfer to sporulation media. Ndt80‐dependent activation of the reporters was measured in transformants of strain MPY2 grown in galactose. Values are β‐galactosidase units and are the average of three independent transformants. The fold repression and activation were calculated by comparison of the level of β‐galactosidase activity with a vector lacking an MSE.

We also measured the ability of the sites to induce expression during the middle stages of sporulation (Figure 1). All of the MSEs tested showed an increase in the level of LacZ expression during meiosis relative to the level of expression during vegetative growth. This result suggests that during meiosis, not only are the MSEs derepressed, but they also serve as activator sites, presumably by an Ndt80‐dependent mechanism. To test more directly whether activation by these sites is dependent on Ndt80, we examined LacZ expression in a strain in which NDT80 is ectopically expressed from the GAL1‐10 promoter (Chu and Herskowitz, 1998). We found that under these conditions, all of the MSEs are induced and function as activator sites. However, there are significant differences in the magnitude of Ndt80‐dependent activation. Taken together, these results show that although the MSEs contain a common core element, there are distinct classes of MSEs which have different regulatory activities during mitosis and meiosis.

Sum1 and Hst1 are required for MSE‐mediated repression

We have shown previously that mutations in SSN6, TUP1, SIN3, RPD3 or UME6, genes known to be involved in repressing some sporulation‐specific genes during mitosis, have little effect on the ability of the SMK1 MSE to repress gene expression in vegetative cells (Pierce et al., 1998). To identify genes required for MSE‐mediated repression, we performed a genetic screen to isolate mutants that are unable to repress expression of the HOP1‐LacZ reporter containing the MSE from the SMK1 promoter. Haploid transformants were mutagenized with ethylmethane sulfonate (EMS) and screened for blue colonies by X‐gal filter assays. Twenty‐seven independent recessive mutants were isolated that fell into two complementation groups comprised of eight and 19 isolates. The corresponding genes were cloned from a genomic plasmid library by complementation of the LacZ expression phenotype using a representative mutant from each group. Subcloning and sequence analysis of the clones obtained in the screen of the library show that SUM1 is required for complementation of the first group and HST1 is required for complementation of the second group.

The SUM1 and HST1 genes were identified previously based on their involvement in transcriptional silencing at the silent mating type loci. SUM1 was isolated originally as a dominant allele, SUM1‐1, which suppresses the defects in silencing at HMR and HML in sir2 mutants (Klar et al., 1985; Lin et al., 1990). Interestingly, SUM1‐1 also suppresses defects in silencing at the silent mating type loci in sir1, sir3 and sir4 mutants, mutations in the Abf1‐, Rap1‐ or Orc1‐binding sites in the HMR silencer E‐box element, and point mutations in the N‐terminal tails of the histone H3 and H4 proteins (Livi et al., 1990; Laurenson and Rine, 1991; Chi and Shore, 1996). The HST1 gene was identified initially as a sequence homolog to SIR2, sharing 63% identity and 76% sequence similarity between the entire proteins (Brachmann et al., 1995; Derbyshire et al., 1996). Although these findings suggest that both proteins may have roles in silencing, sum1 and hst1 null mutants do not show any apparent defects in silencing at the silent mating type loci or telomeres and, therefore, the normal functions of these two proteins in the cell were unknown (Brachmann et al., 1995; Chi and Shore, 1996; Derbyshire et al., 1996).

To verify that Sum1 and Hst1 are required for MSE‐mediated repression, we constructed sum1 and hst1 null mutants and measured LacZ expression from promoter–reporter constructs containing the SMK1, NDT80(−78) and SPR3 MSE sites (Figure 2). Expression of LacZ from the reporter plasmid containing the SMK1 MSE in the sum1 and hst1 null strains is 170‐ and 34‐fold higher, respectively, than in the wild‐type strain. The LacZ reporters containing the NDT80(−78) and SPR3 MSEs also show significant derepression in the sum1 and hst1 null mutants. In contrast, the sum1 and hst1 mutations have no effect on transcriptional repression mediated by the α2‐Mcm1 or Ume6 repressor‐binding sites (data not shown). These observations indicate that both Sum1 and Hst1 are specifically required for MSE‐mediated transcriptional repression during vegetative growth.

Figure 2.

Mutations in SUM1 and HST1 derepress MSEs. LacZ expression of HOP1‐LacZ reporter plasmids containing the SMK1 (pJX43) and NDT80(−78) (pMP15), SPR3(−14) (pMP38) and SPR3(−289) (pMP42) MSEs was measured in wild‐type (W303‐1A), sum1 (JXY3) and hst1 (JXY5) strains during vegetative growth as in Figure 1. The level of expression of the HOP1‐LacZ reporter plasmid containing the SMK1 MSE (pJX43) in a SUM1‐1 strain (JRY2456) is also shown.

SUM1‐1 suppresses the silencing defects in a number of different mutants (Livi et al., 1990; Laurenson and Rine, 1991; Chi and Shore, 1996). It was possible that this mutation suppresses the defect in silencing by altering the regulation of genes containing MSEs. We have assayed repression of the LacZ reporter construct containing the SMK1 MSE in a SUM1‐1 mutant and found that it has wild‐type levels of repression (Figure 2). This result suggests that the effect of SUM1‐1 on silencing is not an indirect consequence of derepressing MSE‐containing promoters.

We next examined the expression of a series of sporulation‐specific genes in sum1‐Δ or hst1‐Δ mutants during vegetative growth by hybridization analysis (Figure 3A). The levels of SMK1 and SPR3 mRNAs are at least 50‐fold higher in the sum1 mutant than in the wild‐type strain. In contrast, SMK1 is not derepressed in the hst1 mutant strain. SPR3 is derepressed in the hst1 mutant 7‐ to 9‐fold, significantly less than the level of derepression seen in the sum1 mutant. NDT80 is not derepressed in either mutant despite the fact that it contains an MSE that is repressed in a Sum1‐ and Hst1‐dependent fashion in the LacZ reporter assay (Figure 2). The level of SMK1 and SPR3 expression in a sum1 mutant is indistinguishable from that seen in the sum1 ndt80 double mutant (data not shown). These data show that derepression of SMK1 or SPR3 is not an indirect effect of derepressing the Ndt80 transcriptional activator.

Figure 3.

Sum1 and Hst1 repress expression of middle‐sporulation‐specific genes. (A) Expression of mid‐sporulation‐specific genes under vegetative conditions. RNA was prepared from wild‐type (W303‐1A), sum1‐Δ (JXY3) and hst1‐Δ (JXY5) strains grown under vegetative conditions. The same blot was hybridized with radiolabeled DNA fragments specific for the coding regions of the SMK1, NDT80, SPR3 and NHP6A genes. An ethidium bromide‐stained gel before transfer is shown as a control for RNA loading. (B) Expression of mid‐sporulation genes in NDT80‐expressing strains. RNA was prepared from wild‐type (W303‐1A) or sum1 (JXY3) strains grown in galactose‐containing media. Strains contained (+) or lacked (−) the GAL1‐10 promoter fused to NDT80 as indicated (Chu and Herskowitz, 1998; Chu et al., 1998). (C) Expression of SMK1, NDT80 and the early meiotic HOP1 gene in wild‐type (W303‐1A), sum1 (JXY3), ume6 (RSY431) and sum1 ume6 (JXY15) strains grown under vegetative conditions.

The absence of derepression of NDT80 upon deletion of SUM1 or HST1 does not necessarily mean that its promoter is not subject to Sum1 or Hst1 regulation. In addition to the two MSEs, the NDT80 promoter contains elements at −158 and −296 that conform to the URS1 consensus sequence (Luche et al., 1990). The Ume6 protein binds to URS1 sites and functions as a repressor in vegetative cells by recruiting the Sin3–Rpd3 histone deacetylase complex to the promoter (Kadosh and Struhl, 1997, 1998a,b). To test the possibility that NDT80 is repressed in vegetative cells through both the Sum1‐ and Ume6‐dependent pathways, NDT80 mRNA levels were assayed in ume6 and sum1 ume6 mutant strains (Figure 3C). While NDT80 mRNA is not expressed in either sum1 or ume6 mutants, it is expressed in the sum1 ume6 double mutant. This result indicates that Sum1 can regulate the expression of NDT80 under conditions where the Ume6‐dependent repression is derepressed.

Ectopic expression of NDT80 has been shown to activate a relatively large number of mid‐sporulation genes, and Ndt80 has also been shown to interact with MSE DNA (Chu and Herskowitz, 1998). We compared the ability of NDT80 to activate several middle genes in wild‐type and sum1 mutant strains (Figure 3B). NDT80 driven by the galactose‐inducible GAL1 promoter activated SMK1 and SPR3 expression in vegetative wild‐type and in sum1 mutant backgrounds. These results show that Sum1 is not required for MSE‐dependent transcriptional activation by Ndt80 and, furthermore, demonstrate that high‐level expression of NDT80 is able to bypass SUM1‐dependent repression.

Sum1 represses SMK1 but not adjacent genes

Under some conditions, the Sum1 and Hst1 proteins are involved in the transcriptional silencing of large regions of the chromosome (Klar et al., 1985; Lin et al., 1990; Livi et al., 1990; Laurenson and Rine, 1991; Brachmann et al., 1995; Chi and Shore, 1996; Derbyshire et al., 1996). If Sum1 and Hst1 function in the same manner as the Sir proteins to repress transcription, then other genes in the vicinity of the MSEs may also be silenced. The NHP6A gene is transcribed divergently from SMK1, and their translation initiation sites are separated by only 306 bp. Although SMK1 is repressed by Sum1 during vegetative growth, NHP6A is expressed at comparable levels in wild‐type, sum1 and hst1 strains (Figure 3A). These results suggest that unlike silencing at the telomeres, rDNA and silent mating type loci, repression mediated by Sum1 through the MSEs is gene specific, highly localized and does not spread to neighboring genes.

Overexpression of SIR2 partially suppresses defects in repression of an hst1 mutant

Overexpression of HST1 partially suppresses the effects on silencing of a sir2 mutant, suggesting that the two proteins have related activities (Brachmann et al., 1995; Derbyshire et al., 1996). We have performed the converse experiment and found that overexpression of SIR2 partially suppresses the MSE‐repression defect of an hst1 mutant (Figure 4). The suppression by SIR2 is specific for hst1 because overexpression of SIR2 does not alter expression of a reporter lacking an MSE and it does not suppress the defect of a sum1 mutant (data not shown). In addition, overexpression of SIR4 has no effect on the repression defect in an hst1 mutant (data not shown). These data further support the idea that Sir2 and Hst1 have similar activities.

Figure 4.

Overexpression of SIR2 partially suppresses the hst1 mutant phenotype. The SMK1 MSE reporter vector, pJX43, was co‐transformed with a high copy pGAL10‐SIR2 plasmid, pAR14 (Holmes et al., 1997), into wild‐type (W303a), hst1‐Δ (JXY5) and hst1‐42A (JXY2) strains. The level of β‐galactosidase activity was measured during vegetative growth in 3% raffinose (derepressing conditions) or 2% glucose (repressing conditions). A LacZ reporter vector lacking the MSE expresses the same level of β‐galactosidase in wild‐type and mutant strains grown in glucose or raffinose.

Since overexpression of SIR2 can partially restore the repression to an hst1 mutant, it was possible that Sir2 normally is involved in MSE‐mediated repression. However, the level of the MSE‐mediated repression of the LacZ reporter promoter containing the SMK1 MSE in a sir2 mutant is approximately the same as in the wild‐type strain and the level of repression in a sir2 hst1 mutant is the same as in the hst1 strain. We also found that the level of SPR3 mRNA is comparable between hst1 and hst1 sir2 strains during vegetative growth (data not shown). These data suggest that while Hst1 and Sir2 are functionally related, they play non‐overlapping roles in repressing mid‐sporulation genes and in silencing, respectively, in wild‐type cells. This finding is supported further by the observation that mutations in sir3 and sir4, which affect silent mating type loci and telomere silencing, as well as mutations in SIR2 homologs, hst2, hst3 and hst4, do not affect MSE‐mediated repression of our reporter constructs (data not shown).

Sum1 binds specifically to the SMK1 MSE

The Sum1 and Hst1 proteins do not show sequence similarity to a known DNA‐binding motif. To determine whether either protein binds directly to the SMK1 MSE, an electrophoretic mobility shift assay (EMSA) was developed to monitor binding to the site. Two slowly migrating MSE‐binding activities are detectable in cell extracts from the wild‐type strain (Figure 5A). The slower migrating MSE‐binding complex (arrow) is competed specifically by unlabeled MSE DNA, but not by non‐specific DNA or an MSE containing a mutation in the consensus sequence (lanes 2 and 3; data not shown). The MSE‐specific shift is missing in extracts prepared from the sum1 null mutant but is present in hst1 null mutant extracts (lanes 4 and 7). This result indicates that Sum1 is required for the MSE‐specific shift.

Figure 5.

Binding activity of the SMK1 MSE. (A) An EMSA of a 23 bp double‐stranded labeled oligonucleotide containing the SMK1 MSE site with crude extracts from wild‐type (W303‐1A, lanes 1–3), sum1 (JXY3, lanes 4–6) and hst1 (JXY5, lanes 7–9) strains (Gailus‐Durner et al., 1997). Binding competition experiments were performed by pre‐mixing an excess of unlabeled MSE DNA (S, lanes 2, 5 and 8) or a non‐specific site (N, lanes 3, 6 and 9) before the addition of the extract. No competitor (−) was added in lanes 1, 4 and 7. The minor increase in complex formation seen in the hst1 samples is not reproducible. (B) An antibody supershift experiment to an HA‐tagged version of Sum1. Extracts were made from JXY3 (sum1‐Δ) transformed with either a negative control vector (pRS415, lanes 10–12), wild‐type SUM1 (pSUM1, lanes 13–15) or SUM1 tagged with three copies of the HA epitope (pSUM1‐HA, lanes 16–18) (Chi and Shore, 1996). A 1 μl aliquot of HA antibody was added to the reaction at a 1:5 dilution in lanes 11, 14 and 17, and a 1:25 dilution in lanes 12, 15 and 18. No antibody was added to lanes 10, 13 and 16. (C) DNA‐binding activity of an MBP–Sum1 fusion to the SMK1 MSE. Lanes 19–22 show shifts by partially purified bacterial extracts from strains containing a pMAL‐C2 blank expression vector (V, lane 19) or an MBP–Sum1 expression vector (lanes 20–22). Lane 21 contains an excess of unlabeled MSE DNA and lane 22 contains a similar quantity of a non‐specific site.

To determine whether Sum1 is in the shifted complex, an extract from a sum1 strain harboring a plasmid expressing an epitope‐tagged Sum1‐hemagglutinin (HA) fusion protein was incubated with HA antiserum and DNA‐binding activity was monitored in an EMSA. The presence of the HA antibody inhibited the formation of the MSE‐specific complex with HA‐tagged Sum1 (Figure 5B, lanes 17 and 18) but not with the untagged protein (lanes 14 and 15). This result indicates that Sum1 is a component of the MSE‐specific shifted complex.

To test whether Sum1 can bind to MSE DNA directly in vitro, we constructed an MBP‐SUM1 expression vector and purified the fusion protein from bacteria. The maltose binding protein (MBP)–Sum1 fusion protein produces a specific shift of the SMK1 MSE, while MBP purified from extracts lacking the fusion does not (Figure 5C, lanes 19–22). These results show that Sum1 can bind to MSE DNA directly and specifically.

We have shown that MSEs from the promoters of different sporulation‐specific genes vary in their ability to repress transcription (Figure 1). To determine whether the ability of these different sites to repress transcription correlates with their Sum1‐binding affinities, we conducted competition EMSAs with unlabeled MSE duplex oligonucleotides (Figure 6). MSEs that function as strong repressor sites in vivo, such as SMK1 and NDT80(−78), function as strong competitors for Sum1 binding. In contrast, the DIT1 sites, which do not function as repressor sites in vivo, are unable to compete for Sum1 binding to the SMK1 MSE. These data demonstrate that the ability of Sum1 to bind to an MSE in vitro correlates with the ability of the site to repress transcription in vivo.

Figure 6.

Competition for Sum1 binding to different MSEs. An EMSA is shown of MBP–Sum1 partially purified from bacterial extracts binding to a labeled SMK1 MSE in the presence of different amounts of cold competitor MSE DNA. MSEs from SMK1(−69) (lanes 1–3), NDT80(−78) (lanes 4–6), NDT80(−221) (lanes 7–9), SPR3(−14) (lanes 10–12), SPR3(−289) (lanes 13–15), DIT1(−342) (lanes 16–18) and DIT1(−555) (lanes 19–21) were used as cold competitors for Sum1 binding. Cold competitor was added at 333‐fold (lanes 1, 4, 7, 10, 13, 16 and 19), 33‐fold (lanes 2, 5, 8, 11, 14, 17 and 20) and 3‐fold (lanes 3, 6, 9, 12, 15, 18 and 21) excess to labeled wild‐type SMK1 MSE.


The MSE was first identified as an element required for the activation of the mid‐sporulation‐specific SPS4 and SPR3 genes (Hepworth et al., 1995; Ozsarac et al., 1997). Analysis of the genome‐wide transcriptional program during sporulation (Chu et al., 1998) showed that sequence elements conforming to the MSE consensus proposed by Ozsarac (gNCRCAAAA/T) are found in the promoters of most middle sporulation genes. In addition, Ndt80 has been shown to bind directly to MSE DNA, and ectopic expression of NDT80 in vegetative cells has been shown to transcriptionally activate many MSE‐containing promoters (Chu and Herskowitz, 1998; Chu et al., 1998).

In this report, we have shown that a subclass of MSEs can function as potent repressor sites (Figure 1). Not all MSEs are strong repressor sites and, even within a single promoter with multiple MSEs, some are strong repressor sites while others are not (e.g. the NDT80 −78 and −221 MSEs; Figure 1). Although repressing and non‐repressing MSEs conform to the consensus sequence, the different regulatory activities of the sites suggest that the sequence specificity of MSE recognition is more complex than previously proposed. Differences of non‐conserved positions in the core element and base pairs flanking this site must therefore be important for contributing to the specificity of Ndt80 and Sum1 binding and/or their regulatory activities.

Our results suggest that there can be two occupancy states at certain MSE sites: a Sum1‐bound transcriptionally inactive state, which is associated with mitotic growth, and an Ndt80‐bound transcriptionally active state, which is associated with mid‐sporulation. NDT80 is expressed slightly before most mid‐sporulation‐specific genes (Hepworth et al., 1998). Thus, it is likely that during the interval between induction (early sporulation) and meiotic chromosome segregation (mid‐sporulation), the ratio of Sum1:Ndt80 binding activity decreases, leading to the derepression, as well as activation, of MSE‐containing promoters. Comparison of the repression and activation properties of the different MSEs in the heterologous promoter assay (Figure 1) suggests that different MSEs have different relative affinities for Sum1 and Ndt80. We have shown that the differences in the repression in vivo correlate with differences in Sum1‐binding affinity in vitro. These differences, in combination with variations over time in the ratio of the two proteins with opposing transcriptional regulatory effects, could provide a simple binary mechanism to vary both the precise timing and the magnitude of expression of a large number of target genes. Indeed, there is considerable variation in the timing of expression of genes containing MSEs during the middle stages of sporulation (Chu et al., 1998).

The idea that DNA‐binding proteins with opposing transcriptional effects can determine developmental choices is firmly established in the life cycle of λ phage by the Cro or cI proteins, which bind to the same sites but have different regulatory activities (Ptashne, 1992). Our results raise the possibility that the regulated competition for occupancy at key MSEs by Sum1 and Ndt80 may control the irreversible commitment step of meiotic chromosome segregation and subsequent spore morphogenesis. Additional layers of complexity that could be applied to this model would include regulated changes in Sum1 or Ndt80 activities in response to nutritional or perhaps meiotic checkpoint signals that regulate progression of sporulation.

Chu and Herskowitz (1998) previously demonstrated that Ndt80 can activate its own expression, presumably through the MSEs found in its promoter. It has been proposed that Ndt80 may be expressed at a low level early during sporulation (e.g. by a URS1‐dependent mechanism), which in turn could lead to high levels of Ndt80 during mid‐sporulation through a positive (MSE‐dependent) feedback loop (Chu and Herskowitz, 1998; Hepworth et al., 1998). Our results show that NDT80 is derepressed in a double ume6 sum1 mutant but not in a ume6 or a sum1 single mutant background. These observations show that NDT80 is regulated through both the Ume6 and Sum1 pathways and are consistent with the transcriptional cascade model for NDT80 induction.

The sequence similarity of the Hst1 and Sir2 proteins, the genetic interactions between SUM1 and SIR genes, and our demonstration that sum1 and hst1 mutants are defective in MSE repression suggest that mechanisms used to repress mid‐sporulation‐specific genes in vegetative cells may be partially shared with those used for silencing. The deletion of SUM1 leads to more robust derepression than the deletion of HST1 for all MSEs tested in the LacZ reporter assay. In addition, SMK1 and SPR3 are expressed at high levels in vegetative sum1 mutants, while in hst1 mutant backgrounds SMK1 is not expressed and SPR3 is expressed at only moderate levels. Furthermore, the level of derepression of the SMK1 and SPR3 MSEs in the LacZ expression assay, as well as expression of these mRNAs in a sum1 hst1 mutant strain, is comparable with that seen in the single sum1 mutant (data not shown). These data show that Sum1 is able at least partially to repress transcription in the absence of Hst1. Therefore, while Sum1 may function to recruit Hst1 to MSE‐containing promoters, it also has an intrinsic ability to repress expression in the absence of Hst1.

SUM1 was first identified as a dominant mutant, SUM1‐1, that suppresses the effects of sir2, as well as other mutants, in silencing of the silent mating type loci. We have shown that the SUM1‐1 mutant retains wild‐type levels of MSE‐mediated repression. Therefore, it is likely that the SUM1‐1 mutation does not alter its normal DNA‐binding specificity. Our data suggest that Sum1 and Hst1 may interact, and raise the possibility that the SUM1‐1 mutation causes alterations in binding affinity and/or specificity towards other Sir proteins. If so, these gain‐of‐activity interactions might play a role in recruiting Hst1, which is known to suppress sir2 defects partially when overexpressed, to silencing complexes.

Materials and methods


Oligonucleotides containing MSEs from the indicated promoters were cloned into the XhoI site of the HOP1‐LacZ reporter vector, pAV124, as described previously (Pierce et al., 1998). The sequences of the MSEs are shown in Figure 1, and each of these sites is flanked with TCGA ends to facilitate its cloning into a XhoI site of the vector.

The MBP–Sum1 fusion was constructed by cloning a PCR‐generated fragment containing the entire SUM1 open reading frame (ORF) between the BamHI and PstI sites of pMAL‐C2 (New England Biolabs). The PCR fragment encodes amino acid residues 2–1062 of Sum1 and includes 28 bp 3′ to the termination codon. Clones were screened by restriction analysis and confirmed by their ability to express an MBP fusion protein as monitored by Western analysis.

Isolation and analysis of mutants defective in MSE repression

A list of the strains used in this study is shown in Table I. Strains W1011‐3B and W1346‐3C, derivatives of W303 harboring pJX43, a hop1‐LacZ reporter vector with the SMK1 MSE site (Pierce et al., 1998), were mutagenized with 3% EMS to 20% survival, and 2.4 × 104 colonies were screened for LacZ expression by X‐gal filter assays. The mutants were rescreened for repression and sorted into complementation groups by pair‐wise mating between the mutants and screening for the inability to repress the LacZ reporter promoter. Genes that complement the mutations were cloned by co‐transforming a CEN/LEU2 plasmid S.cerevisiae genomic library (ATCC 77163) with pJX43 into a mutant from each complementation group (JXY1 and JXY2) and screening the transformants for white colonies in X‐gal filter assays. The complementing library plasmids were purified, reassayed for repression and the end points of each insert were determined by sequencing. Four and seven independent clones were isolated which complement the sum1–50A and hst1‐42A mutants, respectively. All of the clones contained a full‐length copy of their respective genes. Subclones of the individual ORFs contained on the complementation plasmids were constructed in pRS415, transformed back into the mutant strains and assayed for complementation of repression by X‐gal filter assay.

View this table:
Table 1. Yeast strains

To construct sum1‐Δ and hst1‐Δ null mutants, PCR‐generated DNA fragments containing the KanMX4 gene (Wach, 1996) with short flanking regions containing homology to the target gene on both ends were transformed into a W303 diploid strain and transformants were selected by plating on YEPD plates containing G418 (200 mg/l) (Gibco‐BRL). Colonies were rescreened for growth on G418 plates and the integration was verified by PCR using primer pairs that hybridize within and outside of the transformed fragment. Haploid sum1‐Δ (JXY3) and hst1‐Δ (JXY5) null mutant strains were obtained by dissection of sporulated heterozygous diploid strains. JXY3 and JXY5 were mated with JXY1 and JXY2, respectively, and sporulated. The diploid strains and dissected spores all show derepression of the lacZ reporter containing the SMK1 MSE, indicating that strains JXY1 and JXY2 contain mutations in the SUM1 and HST1 genes, respectively.

Liquid β‐galactosidase assay

β‐galactosidase activities of the HOP1‐LacZ constructs containing different MSE sites were determined in W303A for repression activity in vegetative cells and RSX2‐7B for activation during mid‐sporulation. To measure the suppression of the hst1 mutant by overexpression of SIR2, strain JXY5 (hst1‐Δ) that was co‐transformed with pAR14, a 2μ pGAL10‐SIR2 vector (Holmes et al., 1997), and pJX43 were grown in SD −ura −leu medium to saturation, diluted 1:25 into SRaf −ura −leu after being washed once in water, grown overnight and then assayed for β‐galactosidase activity. All β‐galactosidase activity assays in liquid were performed as described previously (Gailus‐Durner et al., 1997).

Electrophoretic mobility shift assays

Yeast cell extracts were prepared from wild‐type (W303‐1A), sum1‐Δ (JXY3) and hst1‐Δ (JXY5) strains as described previously (Gailus‐Durner et al., 1997). Oligonucleotides containing the SMK1 MSE were end‐labeled with [γ‐32P]ATP using polynucleotide kinase and purified by Nensorb columns (NEN) according to the manufacturer. The oligonucleotides were made double‐stranded by mixing with a 3‐fold excess of the matching strand, incubating at 90°C for 20 min and slowly cooling to 25°C overnight in a water bath. Binding reactions for the various protein preparations were carried out in 10 mM Tris–HCl pH 7.5, 40 mM NaCl, 4 mM MgCl2, 6% (w/v) glycerol, 10 μg/ml of sonicated salmon sperm DNA and 32P‐labeled oligonucleotide pairs (10 000 c.p.m.) in a total volume of 20 μl at room temperature for 20 min. Competition experiments were performed by pre‐mixing an excess of unlabeled MSE DNA or a non‐specific site before the addition of the extract. Protein dilutions were made in 20 mM Tris–HCl pH 8, 50 mM NaCl, 1 mM EDTA, 1 mg/ml bovine serum albumin (BSA), 5 mM β‐mercaptoethanol and 1 mM phenylmethylsulfonyl fluoride (PMSF). In the antibody supershift experiment, extracts were made from strain JXY3 transformed with either a negative control vector (pRS415), wild‐type SUM1 (pSUM1) or SUM1 tagged with three copies of the HA epitope (pSUM1‐HA) (Chi and Shore, 1996). A 1 μl aliquot of antibody to the HA epitope (Boehringer Mannheim) was added to the binding reaction at a 1:5 or 1:25 dilution and incubated for 30 min. Samples were analyzed on a 6% polyacrylamide gel (run in 0.5× TBE buffer for 60 min at 200 V). Gels were dried after electrophoresis, exposed to a phosphor screen and scanned on a Model 425E Molecular Dynamics phosphorimager.


We are deeply indebted to Lenore Neigeborn, Randy Strich, Jasper Rine, David Shore, Marc Gartenberg and Jef Boeke for providing strains, plasmids and helpful advice for the genetic screens and analysis. The authors would like to thank Randy Strich and David Norris for comments on the manuscript. This work is supported by Busch Postdoctoral and Predoctoral Fellowships to V.G.D. and M.P., respectively, a grant from the National Science Foundation (E.W.) and grants from the American Cancer Society (E.W. and A.K.V.).