C. Koch, Department of Biology, Chair for Biochemistry, Friedrich-Alexander-University Erlangen-Nürnberg, Staudtstr. 5, 91058 Erlangen, Germany Fax: +49 9131 8528254 Tel: +49 9131 8528257 E-mail: firstname.lastname@example.org
Saccharomyces cerevisiae cells control their cell size at a point in late G1 called Start. Here, we describe a negative role for the Sin3/Rpd3 histone deacetylase complex in the regulation of cell size at Start. Initiation of G1/S-specific transcription of CLN1, CLN2 and PCL1 in a sin3Δ strain occurs at a reduced cell size compared with a wild-type strain. In addition, inactivation of the transcriptional regulator SIN3 partially suppressed a cln3Δ mutant, causing sin3Δcln3Δ double mutants to start the cell cycle at wild-type size. Chromatin immunoprecipitation results demonstrate that Sin3 and Rpd3 are recruited to promoters of SBF (Swi4/Swi6)-regulated genes, and reveal that binding of Sin3 to SBF-specific promoters is cell-cycle regulated. We observe that transcriptional repression of SBF-dependent genes in early G1 coincides with the recruitment of Sin3 to specific promoters, whereas binding of Sin3 is abolished from Swi4/Swi6-regulated promoters when transcription is activated at the G1 to S phase transition. We conclude that the Sin3/Rpd3 histone deacetylase complex helps to prevent premature activation of the S phase in daughter cells.
Most eucaryotic cells regulate their commitment to cell division at the G1 to S phase transition. In the budding yeast Saccharomyces cerevisiae, the events in late G1 leading to S-phase entry are collectively referred to as ‘Start’ [1–3]. During the G1 phase, yeast cells monitor their size and ensure that they have reached a sufficiently large size for entry into the mitotic cell cycle. One of the earliest events occurring as cells pass through Start is the transcriptional activation of a large set of G1/S-specific genes including the G1 cyclins CLN1 and CLN2 and S-phase regulators [4–6]. Cln1 and Cln2 with their associated cyclin-dependent kinase Cdc28 (CDK1) activate the subsequent steps, leading to the accumulation of Clb5/6–CDK1 activity, DNA synthesis, budding and spindle pole body duplication.
The periodic expression of G1/S-specific RNAs depends on the two transcription factor complexes SBF (Swi4/Swi6) and MBF (Mbp1/Swi6) which share the common subunit Swi6 but contain different DNA-binding proteins [7–9]. Swi4 recognizes short cis-acting sequences called Swi4/6 cell-cycle box (SCB) elements originally identified in the HO promoter, whereas Mbp1 binds to MluI cell-cycle box (MCB) elements found in many S-phase genes, including cyclins CLB5 and CLB6 [7,10–12]. Genes regulated by SBF include the G1 cyclins CLN1, CLN2 and PCL1 [13,14]. The timing of CLN1 and CLN2 transcription is of particular importance for the control of cell size because their ectopic expression leads to early entry into the S phase [3,15]. Inactivation of SWI4 causes a defect in Start-specific transcription resulting in abnormally large cells with problems in morphogenesis [13,14,16].
Different cyclins are responsible for regulating G1/S-specific transcription. Whereas repression in G2 is caused by Clb1–4/CDK1 activity and leads to the dissociation of Swi4/Swi6 (SBF) from the promoter, activation in late G1 requires Cln3/CDK1 activity [3,15,17–19].
In early G1, SBF is already bound to the promoter but does not activate transcription [18,19]. This inactivity is largely because of binding of the Whi5 repressor to SBF [20,21]. Whi5 is thought to be the key target for the Cln3/CDK. Phosphorylation of Whi5 leads to its dissociation from SBF and its subsequent export from the nucleus [20,21].
This mode of regulation is strikingly similar to the activation of metazoan E2F transcription factors by cyclin D/Cdk4, which phosphorylates and thereby inactivates the Rb repressor before S phase .
In yeast, the G1 cyclin Cln3 is the key regulator that integrates signals about cell size and growth rate to promote cell-cycle progression at Start [1,2,23]. Differences in Cln3 protein levels and stability have a profound influence on cell size at Start. Activated alleles of CLN3 lead to smaller cells, whereas a cln3Δ mutant, although viable, enters the S phase at a larger cell size . Consistent with a function as a repressor and important target for Cln3/CDK activity, inactivation of WHI5 advances cell-cycle entry and largely bypasses the requirement for CLN3 [20,21]. Studies at the HO promoter have shown that CDK activation in late G1 is important for polymerase recruitment, whereas recruitment of Srb/mediator complex by SBF occurs prior to CDK activation [24,25]. A number of additional regulators were shown to affect the amount and timing of G1/S-specific transcription. These include, in particular, BCK2, which becomes essential in the absence of CLN3 , CCR4 , XBP1 , MSA1 , NRM1  and STB1 . Despite their similar architecture, SBF and MBF are not identically regulated. For example, the corepressor Nrm1 specifically regulates MBF target genes . STB1 was reported to have different effects on MBF- and SBF-regulated genes although it binds to the common subunit Swi6 . Deletion of STB1 in a cln3Δ strain caused a delay in G1/S transcription and the accumulation of large unbudded G1 cells  suggesting that Stb1 may act as an activator. Further experiments showed that the interaction of Stb1 with Swi6 is abolished upon phosphorylation of Stb1 through Cln–Cdc28 kinase complexes [32,33]. Earlier studies suggested that Stb1 may specifically act on MCB elements , whereas recent chromatin immunoprecipitation (ChIP) assays provided evidence that Stb1 is recruited to both SCB and MCB elements in the G1 phase . Stb1 was originally found to interact with the transcriptional corepressor Sin3 in a two-hybrid assay . Recent analysis of G1-specific mRNA levels in stb1Δ and sin3Δ mutants suggested a role for Sin3 and Stb1 in regulating these genes . Sin3 and its associated histone deacetylase Rpd3 act together in large multiprotein complexes on transcriptional repression of many genes [35–39]. Through interaction with DNA-binding proteins, Sin3 recruits the deacetylase Rpd3 to specific promoters. In particular, the DNA-binding protein Ume6 was shown to recruit Sin3 and Rpd3 deacetylase activity to genes involved in phospholipid biosynthesis, meiosis and sporulation [37,40–43]. Genome-wide acetylation studies  and genome-wide binding studies for Rpd3  showed that genes involved in cell growth and cell-cycle control, including the G1-specific gene PCL1, are targeted by the Rpd3 deacetylase.
In this study, we uncover a role for Sin3 and its associated histone deacetylase Rpd3 in cell size homeostasis at the Start of the cell cycle. We find that SIN3 represses SBF-dependent transcription in early G1 and show that Sin3 is bound to promoters in G1 and released around the onset of Start transcription. We conclude that Sin3 is important for the correct timing of SBF-dependent transcription in G1.
Sin3 represses SBF-dependent transcription
Mutations that accelerate cell division relative to cell growth lead to a reduced cell size at Start . The timing of Start is mostly determined by the initiation of G1/S-specific cyclin transcription. Activated alleles of the regulator CLN3 lead to smaller cells, whereas loss of CLN3 delays CLN1,2 transcription causing cells to start the cell cycle at a larger size . We exploited this phenotype in a screen for novel dose-dependent regulators of G1/S-specific transcription. We transformed cln3Δ mutants with a multicopy genomic library derived from YEplac181 and used centrifugal elutriation to identify transformants with a reduced cell size in G1. Not surprisingly, we found plasmids encoding the known regulators CLN1, CLN2, CLN3 and SWI4 (data not shown). In addition, we identified a plasmid encoding a truncated version of SIN3, lacking the C-terminal part of the coding region (2μSIN3ΔC) (Fig. 1) that led to a reduction of cell size in cln3Δ cells (Fig. 1A). This was accompanied with increased levels of CLN2 RNA and an increased budding index (Fig. 1B). The change in cell size may therefore be the result of increased G1 cyclin expression. This was unexpected because Sin3 has been described as a repressor of transcription [46,47]. We therefore tested whether the phenotype could be explained by a dominant-negative effect of the truncated SIN3 allele on the function of wild-type SIN3 gene. Sin3Δ mutants were originally identified because they allow HO expression in the absence of SWI5 [46,47]. Using a Swi5-dependent HO-ADE2 reporter gene we found that swi5 mutants transformed with the SIN3ΔC plasmid expressed HO-ADE2, suggesting that the truncated allele has a dominant-negative effect (data not shown).
To test directly whether SIN3 has an effect on the regulation of G1/S-specific transcription, we compared synchronized wild-type and sin3Δ mutant cells. Because we were interested in the timing of G1 cyclin expression, we analysed small G1 cells isolated by centrifugal elutriation. The collected G1 cells were diluted in fresh media and followed as they progressed through the cell cycle (Fig. 2). Isolation of small unbudded sin3Δ cells turned out to be difficult and yielded populations with a minimum content of 8–9% budded cells. The elutriated sin3Δ cells initiated budding at a size of 24–26 fL. This was smaller than for the congenic wild-type cells, which initiated budding at 32–35 fL (Fig. 2A). It is unlikely that the observed difference is caused by a lack of synchrony in the sin3Δ culture, because cells from the elutriated sin3Δ population were, on average, slightly smaller than those from the wild-type population (20.9 fL for sin3Δ and 21.8 fL for wild-type), although more sin3Δ cells had already passed the S phase (Fig. 2C). To analyse SBF-dependent gene regulation, the mRNA level of G1/S-specific genes was determined (Fig. 2B). Transcripts of the SBF-regulated genes CLN2 and PCL1 started accumulating at ∼ 23 fL in sin3Δ cells compared with ∼ 30 fL in the wild-type population, around the time of bud emergence (Fig. 2A,B). FACS analysis showed that sin3Δ mutant cells also replicated their DNA at a smaller cell size (Fig. 2C). These observations suggest that Sin3 is involved in repression of Start-specific transcription in G1 and thereby negatively regulates cell-cycle initiation. In most instances, Sin3 acts together with the histone deacetylase Rpd3 . We therefore analysed gene expression in congenic rpd3Δ cells. Rpd3Δ cells synchronized by elutriation initiated budding at a size of 24–26 fL, comparable with the sin3Δ strain (Fig. 2A). Transcription of SBF-regulated genes in the elutriated cells also started at around the same cell size as observed for the sin3Δ mutant (Fig. 2B). It is therefore likely that Sin3 acts together with Rpd3 in the regulation of G1-specific transcripts. Because we observed precocious activation of G1-specific transcription in elutriated sin3 mutant cells, we expected that asynchronously growing mutant cells would be, on average, smaller than a corresponding wild-type population. Interestingly, analysis of mean cell size from asynchronous sin3Δ cultures showed no reduced average size compared with a wild-type population (Fig. 3A–C). The average cell size of a population, however, also depends on the time spent in G2. Indeed, we found an increased budding index of 73% in sin3Δ cultures compared with 48% for wild-type cells. We also observed that log phase sin3Δ cells had a significantly increased percentage of cells that have entered S phase and replicated their DNA. This may, therefore, explain why cells from the sin3 population are, on average, not smaller than the wild-type population.
Inactivation of SIN3 suppresses the CLN3 requirement for Start
When wild-type cells reach a critical cell size, activation of G1-specific transcription by Cln3/Cdk1 is rate limiting for the further events at Start [3,15]. Therefore, inactivation of CLN3 results in a large cell phenotype. To test whether Cln3 is involved in releasing cells from a Sin3-dependent repression of G1-specific transcription, we analysed the consequences of deleting sin3 in a cln3 mutant. The cell size of sin3Δcln3Δ double mutants from a logarithmically growing culture was compared with that of single mutants and wild-type cells (Fig. 3A–C). The average cell size of cln3Δ cells was reduced to approximately the size of sin3Δ in the double mutant, suggesting that sin3 is partly epistatic to cln3 (Fig. 3A–C). The critical cell size for the initiation of budding and the activation of G1/S-specific transcription was investigated in small G1 cells elutriated from an asynchronous sin3Δcln3Δ double-mutant culture (Fig. 3D,E). Similar to the sin3Δ population (see above), it proved difficult to isolate small unbudded sin3Δcln3Δ cells, suggesting partial deregulation of cell-cycle entry. As can be seen from the FACS profile and cell size measurements 15 min after putting cells into fresh medium (Fig. 3F), inactivation of SIN3 in the cln3Δ mutant led to a reduction in cell size at birth. Moreover, although cln3Δ mutants started budding at a nearly twice the size of wild-type cells, sin3Δcln3Δ double-mutant cells initiated budding at around the size of wild-type cells (Fig. 3D) and activated G1/S-specific transcription at a much smaller size than cln3Δ mutants (Fig. 3E). Inactivation of SIN3 in a cln3Δ deletion mutant also caused the cells to replicate their DNA at a smaller cell size (Fig. 3F). Hence, inactivation of SIN3 advances Start in a cln3Δ mutant. These data suggest that Cln3 is also involved in releasing cells from a Sin3 dependent repression.
To elucidate if this repression by Sin3 is dependent on SBF (Swi4/Swi6), the cell size of sin3swi4 double mutants was determined. Cell size analysis of log-phase cultures from double mutants revealed that sin3Δ does not reduce the cell size of swi4Δ mutants, and did not advance transcriptional activation of G1/S-specific genes, but instead increased the average size of swi4Δ mutants from 84 to 109 fL (data not shown).
Sin3 is recruited to SBF-specific promoters
The effect of Sin3 on cell size and S-phase entry suggests a role for Sin3 in the timing of CLN1 and CLN2 transcription by SBF. If Sin3 were directly involved in regulating SBF (Swi4/Swi6)-dependent transcription in late G1, it should be present at the relevant promoters. Sin3 does not directly bind to DNA, but is known to be recruited to specific promoter regions by other DNA-binding proteins [40,48]. We therefore asked whether Sin3 is targeted to promoters of G1/S-specific genes in a Swi4/Swi6-dependent manner. For this, SIN3 was replaced by an epitope-tagged version at the SIN3 locus. Binding of epitope-tagged Sin3–myc to G1/S-specific promoters was assayed by ChIP experiments. Coprecipitated promoter DNA fragments encompassing the SBF-binding sites from the promoter regions of CLN1 and CLN2 were amplified by multiplex PCR along with control fragments from their coding regions and from a nontranscribed region on chromosome V. As shown in Fig. 4, the promoter elements of CLN1 and CLN2 were significantly enriched compared with control fragments from the coding region and the nontranscribed region of chromosome V. Immunoprecipitations were performed in triplicate to control for variations in the efficiency of immunoprecipiation. As an additional control for the specificity of Sin3 binding, cells not expressing the epitope tag were analysed in parallel. Sin3–myc binding to the promoter sequences of CLN1 and CLN2 was strongly reduced in swi4 and swi6 null mutants (Fig. 4), which further demonstrated the specificity of the observed interaction. These results suggest that Sin3 effects G1/S transcription directly, and that Sin3 is recruited to G1 cyclin promoters by SBF or by factors associated with Swi4 or Swi6.
Sin3 recruitment is regulated in a cell-cycle-dependent manner
To detect whether the recruitment of Sin3 and its associated histone deacetylase Rpd3 to G1-specific promoters is regulated during the cell cycle, we tested promoter occupancy in cells that were arrested at different stages of the cell cycle. We analysed cdc28-13 cell-cycle mutants arrested in G1 at 37 °C, as well as cells that were arrested in G2 with the microtubule depolymerizing drug nocodazole. Cdc28-13 mutants arrest in late G1 prior to the activation of G1/S-specific transcription. Strong Sin3 binding was detected in such cells at the CLN1, CLN2 and PCL1 promoter regions (Fig. 5A,D). The stronger signal in the arrested cultures is most probably a simple reflection of cell-cycle-dependent binding. Indeed, Sin3 was not associated with the promoters during G2, as we could not significantly coprecipitate CLN2 or PCL1 promoter elements with Sin3–myc from cells arrested with nocodazole (Fig. 5C,D). To provide further evidence for the specific binding of Sin3, we analysed the recruitment of Rpd3, the catalytic component of the Sin3/Rpd3 histone deacetylase complex, to G1 cyclin promoters in G1 (Fig. 5B). Cdc28-13 mutants expressing an epitope-tagged RPD3–HA6 were synchronized in late G1 by shifting log-phase cultures to 37 °C for 3 h until all cells were arrested as large unbudded cells. In ChIP assays with extracts prepared from arrested Rpd3–HA6 cells we observed recruitment of Rpd3–HA6 to SBF-dependent promoters, although the signal was weaker than the signal for Sin3–myc (Fig. 5B). The binding of Sin3 and Rpd3 to promoters of SBF-regulated genes in G1-arrested cells correlates with the transcriptional repression of CLN1, CLN2 and PCL1 in G1.
To analyse if the release of Sin3 from SBF-regulated genes coincides with transcriptional activation of G1 cyclins, we performed an arrest–release experiment. Cdc28-13 mutants were shifted to 37 °C until they were arrested as unbudded cells in G1. The cells were subsequently released from cell-cycle arrest by shifting the culture to 25 °C. RNA levels and Sin3 binding were analysed from samples taken every 10 min. For the arrested culture, ChIP analysis demonstrated strong binding of Sin3–myc to the CLN1 and CLN2 promoter (Fig. 6A; 0 min). When cells were released from the cell-cycle block, they synchronously entered the cell cycle (Fig. 6C). A peak of G1/S-specific transcription was observed between 10 and 20 min after release. Shortly thereafter, cells entered the S phase (FACS profile in Fig. 6D) and started budding (Fig. 6C). The ChIP signal began to fade 10 min after the cells were released (Fig. 6A,B). The decrease in promoter occupancy by Sin3 correlated best with the timing of transcriptional activation (Fig. 6B). The timing of Sin3 binding is therefore consistent with a role for Sin3 in repression of CLN transcription in the G1 phase.
Because of their abnormally large size, G1-arrested cell-cycle mutants may not accurately reproduce the situation found in small wild-type daughter cells in the early G1 phase. We therefore analysed promoter occupancy of Sin3–myc in elutriated wild-type cells. Small G1 cells were isolated by centrifugal elutriation and allowed to progress through G1. The presence of Sin3–myc at the CLN2 and PCL1 promoter was compared with cell-cycle progression. Because many cells were needed for ChIP assays it was not possible to analyse more than three time points. At each time point, samples from the culture were analysed by ChIP assay and the RNA levels of G1/S-specific genes and the DNA content of cells were determined (Fig. 6E–G). The data confirmed that SBF-specific promoters are occupied by Sin3 in the G1 phase. At 160 min, most cells in the culture had left G1 and exhibited no Sin3 binding to the promoter (Fig. 6E). Analysis of G1 cyclin expression showed that binding of Sin3 to the SBF-dependent promoters correlated with repression in G1, whereas disappearance of Sin3 from the promoter elements coincided with induction of Start-specific transcription (Fig. 6). To elucidate whether Sin3 leaves the promoter together with Rpd3, we performed an arrest-release experiment with cdc28-13 cells expressing Rpd3–HA. As shown in Fig. 7, the binding of Rpd3 is very similar to the kinetics of Sin3 binding (Fig. 6) although the signal intensities for Rpd3 are generally weaker. We therefore conclude that Sin3 acts together with Rpd3 at SBF-dependent promoters (Fig. 7).
In this study, we provide evidence that SIN3 is involved in the correct timing of G1 cyclin expression in Saccharomyces cerevisiae at the G1–S phase transition. We have found that inactivation of SIN3 leads to an advanced induction of Start-specific transcription in G1 daughter cells and that budding is initiated at a smaller cell size. Consistent with a direct role for Sin3 in repressing gene expression prior to Start, we find that Sin3 is present at the promoters of SBF-regulated genes in G1, but leaves the promoter around the time cells enter the S phase. Furthermore, inactivation of Sin3 suppresses the phenotype of cln3 mutants, allowing them to activate G1/S-specific transcription at a smaller cell size (Fig. 3). Such a phenotype would be expected if Cln3 with its associated Cdc28 kinase were involved in the inactivation or repression of Sin3/Rpd3-dependent histone deacetylation at the CLN1,2 promoters. These data raise several questions concerning the regulation of G1 cyclin transcription. In particular, whether CLN3 acts directly on Sin3/Rpd3 and how Sin3 is recruited to SBF-regulated genes like CLN2 in the G1 phase.
How is Sin3 recruited to SBF-regulated promoters?
Sin3 does not bind DNA directly, but associates with transcriptional regulators to bring the Rpd3 histone deacetylase to specific sites in chromatin [40,49]. There are different DNA-binding proteins thought to bind to Sin3. Besides the well-characterized interaction with Ume6, these include Ash1, Mcm1 and Ssn6 [35,50,51]. At the HO promoter, Sin3 is thought to be recruited in part by Ash1 [35,51]. Veis et al. reported cell-cycle-dependent binding of Sin3 to the G2/M-specific CLB2 promoter . Their data further showed that the recruitment of Sin3 is dependent upon an interaction with Fkh2 and Mcm1. The removal of Sin3 and the deacetylase complex does not require B-type cyclins but Cdc28/Cln activity . Similar to the recruitment of Sin3 to G2/M-specific promoters by the regulatory factors Mcm1 and Fkh2, we propose that Sin3 is recruited to G1/S promoters by SBF. The DNA-binding protein Ume6 was shown to be responsible for Sin3/Rpd3 recruitment at many other sites, for example, at SPO13, INO1, IME2 [40,43,52]. We found no Ume6 consensus sites  in the promoter regions of CLN1 and CLN2. Any one of the proteins present at the CLN2 promoter in early G1 could, in principle, be responsible for recruiting Sin3 to the promoter. These include Swi4, Swi6, Whi5 and Stb1 [18,19,21,33,35]. Although recruitment of Sin3 to the CLN2 promoter strongly depends on Swi4 and Swi6 (Fig 4), we found no significant effects of whi5 or stb1 mutants on the binding of Sin3 to the CLN2 promoter (data not shown). In addition, deleting WHI5 in a sin3 mutant did not reduce the cell size to the level of whi5Δ single mutants (data not shown). This makes it unlikely that Sin3 is recruited to the promoter via Whi5. Because the absence of Stb1 was observed to increase cell size of a cln3Δ mutant , it is not likely to mediate Sin3-dependent repression, although it could be important for releasing from Sin3-dependent repression later in the cell cycle. However, the timing of SBF binding [18,19], which arrives at the promoter as cells exit mitosis, would be consistent with a direct role as a Sin3/Rpd3 recruiting factor. Earlier ChIP results showed that the histone deacetylase Rpd3 is associated with the promoters of cell-cycle genes regulated by SBF, MBF, Fkh1, Fkh2, Mcm1 and Ndd1, and showed that SBF affected Rpd3 binding to CDC20 and PCL1, suggesting that Rpd3 can be recruited by several different transcription factors . This is consistent with our observation that both Sin3 and Rpd3 are recruited to G1/S-specific promoters in a cell-cycle-dependent manner. A situation in which transcriptional activators also directly recruit corepressors is in fact quite common, for example, in the case of E2F transcription factors in metazoans .
How is Sin3 removed from the promoter in the S phase?
The observation that deleting SIN3 partly suppresses the size phenotype of cln3 mutants suggests a possible role for Cln3 in the inactivation or subsequent removal of Sin3/Rpd3 complexes from the promoter. The only well-characterized, and presumably critical substrate for Cln3/CDK1 is the repressor Whi5 [20,21]. Removal of Sin3 at the beginning of the S phase is probably not a consequence of Whi5 inactivation caused by phosphorylation by Cln3/CDK1 [20,21], because there is no evidence for a direct interaction between Whi5 and Sin3 or Rpd3.
Alternatively, Sin3 may be a direct target for Cln3 kinase. Sin3 is a phosphoprotein  and was found to coprecipitate with Cln2 in a proteomics study of yeast CDKs . The timing of Sin3s removal from the promoter (Fig. 6) would also be compatible with a scenario in which the downstream Cln1/2–CDKs rather than Cln3/CDK are responsible for inactivating Sin3. Such a mechanism could contribute to a positive feedback loop of CLN activation  assisting in making S-phase entry irreversible. A role for Clns in the removal of Sin3 from promoters was suggested by Veis et al. in the case of CLB2 .
Given that Sin3, together with the Rpd3 histone deacetylase, is involved in modifying chromatin at many sites not concerned with cell-cycle control, we consider it more likely that Sin3/Rpd3-dependent chromatin changes at the CLN2 promoter are regulated by reversible recruitment of Sin3 and or Rpd3, rather than by regulating Sin3/Rpd3 directly.
A good candidate for a factor regulating Sin3/Rpd3 binding is Stb1, because it binds to both Swi6 and Sin3 [20,21,32,34]. In fact, Stb1 was originally identified as an interacting protein of Sin3 in two-hybrid assays (Stb1 for Sin three binding) . STB1 transcription was shown to be cell-cycle regulated and peaks in late G1 phase . Earlier studies showed that Stb1 binds only to synthetic MBF promoters , but a recent study provided evidence for in vivo binding to SBF and MBF promoters in G1 via an interaction with Swi6 . ChIP assays provided evidence that phosphorylation of Stb1 coincides with its dissociation from promoters at G1/S transition . This study also found Stb1 to be associated with G1/S-specific promoters until CLN transcription is inactivated . Phosphorylation of Stb1 inhibits interaction between Swi6 and Stb1 . Our finding of cell-cycle-specific promoter binding by Sin3 is in agreement with the proposal of de Bruin et al.  for a combined role of Stb1 and Sin3 in regulating G1/S transcription. We were able to show that Sin3 binds to promoter sequences prior to transcriptional activation and leaves the promoter around the time of transcriptional activation. The observation that Stb1 and Sin3 bind to promoters in G1 and leave the promoters at the time of transcriptional activation suggests that Stb1 and Sin3 may exert effects on G1/S-specific transcription in a concerted manner.
A possible model for the regulation of SBF by Sin3/Rpd3 may be that Sin3 is directly recruited by SBF in early G1 and that changes in Stb1 remove Sin3 from the promoter. Data concerning the timing of Sin3 removal from the promoter are consistent with a function for both Cln3 and the downstream cyclins Cln1 and Cln2 in removing Sin3 from the promoter. The timing is, however, not compatible with a model in which removal of Sin3 is a simple consequence of SBF removal from the promoter by Clb-kinase activity, because Swi6 remains associated with the CLN2 promoter for much longer . Cell-cycle-dependent binding of Sin3 has also been observed at the CLB2 locus . At the G2-specific CLB2 gene we found that Sin3 is recruited by Fkh2 in G1 and lost from the promoter after activation of Cln1,2 associated kinases . Although the timing is clearly different from the situation at the SBF-regulated genes analysed here, the inactivation by Cln-kinases may occur by a similar mechanism.
How important is Sin3 for the regulation of G1/S-specific transcription?
Sin3 mutants are not obviously smaller than wild-type cells, when the average size in the population is analysed, although we found a larger proportion of post-S-phase cells. This, and the fact that Sin3 mutants are rather pleiotropic, may explain why the deregulation of G1/S-specific genes becomes evident only in isolated G1 cells or in cln3Δ mutants. In addition, the effect of ectopically expressing the dominant-negative allele of SIN3 on cell size was most obvious in cln3 mutants. The importance of Sin3 for regulating CLN2 expression is therefore not so obvious. In the absence of Cln3, the timing of Start execution becomes quite variable, whereas once activated, all Start-related events occur in a coherent fashion [55,56]. It has been argued that cln3 mutants are particularly dependent on a positive feedback mechanism for G1/S transcription, i.e. Cln2 and Cln1 have to accumulate to a certain level before firing the positive feedback loop [55,56]. As a consequence, every mutation that partly deregulates CLN2 expression will potentially lower the threshold at which such a positive loop will fire. This may be one of the reasons why in cln3 mutants cell size is particularly sensitive to the inactivation of SIN3 and may make SIN3 apparently more important for repressing transcription in daughter cells with little Cln3. Remarkably, Aparicio et al.  reported similar effects of Sin3 in the regulation of S-phase timing. They showed that the S phase is advanced in the absence of the Sin3/Rpd3 histone deacetylase complex. In summary, we have identified an additional level of control at the G1- to S-phase transition that contributes to the astonishing precision of transcriptional timing observed in cell cycle regulated transcription in late G1.
Materials and methods
Strains and DNA
Strains used in this study were derived from strains W303, BY4741 or BY4742 (Table 1). Gene deletions were created by integrational transformation of PCR cassettes, as described previously . Double mutants were created by mating. After incubation on sporulation media plates for 2–10 days at 25 °C, tetrads were dissected with a micromanipulator (Singer Instruments, Roadwater, UK) and distributed on YEPD plates. After 4 days at 25 °C the phenotypes were analysed by replica plating. Genotypes of meiotic segregants were confirmed by PCR. Epitope tagging of yeast genes at the C-terminus was performed using a PCR-based strategy to introduce epitope tags to the chromosomal loci . Deletion mutant strains used for cell size measurements were obtained from BIOCAT (BY4741, BY4742, CY5713 and CY5538) and double mutants (CY5715) were created by mating (Table 1). The genomic library was a gift from R. Jansen (LMU Munich, Germany) and was generated by inserting genomic Sau3A fragments into the multicopy YEplac181 vector. Plasmid pCK1509 contains a fragment that comprises 488 bp of the 5′-UTR and 2436 bp of the SIN3 coding region.
Table 1. Yeast strains.
a Strains were obtained from BIOCAT (Heidelberg, Germany). b Strain was created by crossing strain CY5538 with CY5713.
MATalpha, congenic to By4741, except for sin3::KANMX, cln3::KANMXb
Growth conditions and cell-cycle arrests
Yeast cells were cultivated in YEP-based media with 2% glucose (YEPD) or 2% galactose (YEPGal). Cell-cycle arrests of temperature-sensitive mutants (cdc28-13) were performed by growing the cells to a titre of 107 cells·mL−1 at 25 °C (YEPD) in a water bath and then shifting the culture to 37 °C for 165 min. For centrifugal elutriation, cells were grown to D600 = 2.0 in YEPGal. The elutriation chamber was loaded with a total of 8000 D600 cells. Fractions of small cells were collected, pooled and cultivated at 25 °C in fresh media. Samples from the culture were taken at specific time points and cell size and cell-cycle progression were monitored. Budding index was determined by counting ∼ 250 cells. Flow cytometry was used to observe cell-cycle distribution as described previously .
Cell size measurements
Cell number and average cell size were analysed by using a CASY®1 cell counter model TT from Schärfe Systems (Innovatis AG, Reutlingen, Germany). To determine cell number and size distribution of yeast cultures, the cell suspensions were diluted in CASY-ton® isotonic buffer and sonicated for 30 s before the measurement.
ChIPs were carried out with modifications as described previously [24,60]. Crude extracts were prepared from cultures (55 mL, 2.5 × 107 cells·mL−1) treated with 1% formaldehyde for 20 min at room temperature before harvesting. After addition of 135 mm glycine and incubation for 5 min at room temperature, cells were harvested and washed four times with 2 mL NaCl/Tris buffer (20 mm Tris/HCl pH 7.5, 150 mm NaCl) to remove residual formaldehyde. Cells were resuspended in 600 μL lysis buffer (50 mm Hepes KOH pH 7.5, 140 mm NaCl, 1 mm EDTA, 1% Triton X-100, 0.1% deoxycholic acid) and cell breakage was carried out by addition of glass beads, and usage of a IKA® Vibrax VXR basic (25 min 2500 rpm, 4 °C). Extracts were sonicated five times for 30 s using a Bandelin Sonoplus HD2070/SH70G and the debris was removed by centrifugation (16 000 g 5 min, 4 °C). The supernatant was applied to 50 μL Dynabeads® Pan Mouse IgG (Dynal® Invitrogen (Karlsruhe, Germany); 400 000 beads·μL−1) that were loaded with epitope tag-specific antibodies (0.05 μg of antibodies per μL of bead suspension) and incubated for 3 h at 7 °C on a rotator. Antibodies used for ChIP assays were anti-myc 9E11 (Dianova, Hamburg, Germany) and anti-HA 12CA5 (lab preparation). Thereafter, the beads were washed twice with 1 mL lysis buffer, high salt buffer (50 mm Hepes KOH pH 7.5, 500 mm NaCl, 1 mm EDTA, 1% Triton X-100, 0.1% deoxycholic acid), washing buffer (10 mm Tris/HCl pH 8, 250 mm LiCl, 1 mm EDTA, 0.5% NP40, 0.5% deoxycholic acid) and TE buffer. Precipitates were eluted in 50 μL elution buffer (50 mm Tris/HCl pH 8, 10 mm EDTA, 1% SDS) at 65 °C for 10 min. Eluted proteins were analysed by SDS/PAGE and western blotting. We added 120 μL 1% SDS/TE buffer to 30 μL of the supernatant and incubated for 16 h at 65 °C. The eluates were treated with 0.5 μg·μL−1 proteinase K for 2 h at 37 °C. Thereafter, coprecipitated DNA was purified by phenol extraction. DNA fragments of promoters and coding regions were amplified by multiplex PCR and analysed on 2% agarose gels. As a control, we coamplified an untranscribed region of chromosome V (10562–10699) together with either the promoter or the coding region of CLN1, CLN2 and PCL1 in the same PCR. The primers for amplification of the control region on chromosome V were CK2229 (CAGTTTAACCCGAAGTTCTG) and CK2230 (AACAACGCAGCTGCTTTAAC). Primers used for the amplification of fragments from promoter fragments were:
CLN2, CK2148 (ATCTTTTTCGTATCCTCCGC) and CK2149 (AAAGGGCCAACAGTTGTTTC); CLN1, CK2158 (TAGGGTAGCGTGCCACAAAA) and CK2159 (CGTCTCTTGCAGGCTGAACA); PCL1, CK2364 (GCTAACAACTGAGAATGCGA) and CK2366 (ACACAAGAGTTAAGGACAAG).
The primers used for the amplification of fragments from the coding regions were:
CLN2, CK1724 (ATAGTGATGCCACTGTAGAC) and CK1725 (CATGATGGGGTTGATATGGT); CLN1, CK2254 (TAGTTCACCGCAAAGTACTG) and CK2255 (TATTGTAGAGGCCAGTTGCA); PCL1, CK2348 (CCATCCATCGCATTTTCTTG) and CK2349 (CTGTGTTGTTCGCTATGTTG).
Yeast cells were harvested and chilled on ice before they were washed with cold TE buffer. Cell pellets were frozen in liquid nitrogen and stored at −80 °C. Precipitated cells were resuspended in RNA buffer 1 (10 mm Tris/HCl pH 7.5, 300 mm NaCl, 1 mm EDTA, 0.2% SDS) and vortexed vigorously with phenol/chloroform/isoamylalcohol and glass beads for 15 min. After centrifugation the aqueous phases were mixed with ethanol (1 : 3 v/v) and incubated for 20 min at −20 °C. Tubes were centrifuged for 15 min at 16 000 g. Pellets were resuspended in RNA buffer 2 (10 mm Tris/HCl pH 7.5, 1 mm EDTA, 0.2% SDS) and incubated for 5 min at 65 °C. RNA content was measured at 260 nm with a Smartspec™ 3000 from BioRad (Munich, Germany). RNA (20 μg per sample) was separated on 1.3% agarose gels containing 1.3% formaldehyde. RNA was transferred to Gene Screen membranes. Hybridizations with 32P-labelled DNA probes were performed at 65 °C for 16 h in Church buffer (0.5 m NaCl/Pi pH 7.2, 7% SDS, 10 mm EDTA, 1% BSA). Filters were washed twice for 5 min in 2× NaCl/Cit/0.1% SDS and twice for 10 min in 1× NaCl/Cit/0.1% SDS at 65 °C. CMD1 RNA levels were determined as internal loading control.
We gratefully thank Marlis Dahl, Rosi Söllner, Alexander Schwahn, Uwe Sonnewald and Martin Korn for their support and helpful discussions. We thank Gustav Ammerer and Kim Nasmyth for strains. We thank Michael Schwenkert and Christian Kellner for their help with the FACS analysis.