The SBF- and MBF-associated Protein Msa1 Is Required for Proper Timing of G1-specific Transcription in Saccharomyces cerevisiae*

In the budding yeast Saccharomyces cerevisiae, cell cycle initiation is prompted during G1 phase by Cln3/cyclin-dependent protein kinase-mediated transcriptional activation of G1-specific genes. A recent screening performed to reveal novel interactors of SCB-binding factor (SBF) and MCB-binding factor (MBF) identified, in addition to the SBF-specific repressor Whi5 and the MBF-specific corepressor Nrm1, a pair of homologous proteins, Msa1 and Msa2 (encoded by YOR066w and YKR077w), as interactors of SBF and MBF, respectively. MSA1 is expressed periodically during the cell cycle with peak mRNA levels occurring at the late M/early G1 phase and peak protein levels occurring in early G1. Msa1 associates with SBF- and MBF-regulated target promoters consistent with a role in G1-specific transcriptional regulation. Msa1 affects cell cycle initiation by advancing the timing of transcription of G1-specific genes. Msa1 binds to SBF- and MBF-regulated promoters and binding is maximal during the G1 phase. Binding depends upon the cognate transcription factor. Msa1 overexpression advances the timing of SBF-dependent transcription and budding, whereas depletion delays both indicators of cell cycle initiation. Similar effects on MBF-regulated transcription are observed. Based upon these results, we conclude that Msa1 acts to advance the timing of G1-specific transcription and cell cycle initiation.

Regulation of cell division is the means by which multicellular organisms ensure proper growth and differentiation and unicellular organisms achieve fidelity in duplication of genetic and cellular components in future generations. The primary mechanisms utilized by cells to control cell cycle progression are periodic transcription and proteolysis. Although there are distinctions between the metazoan cell cycle and that of budding yeast, recent work in our laboratory strongly suggests that the molecular mechanisms utilized to regulate periodic transcription during the cell cycle, specifically during cell cycle ini-tiation, are highly conserved (1). The pivotal role of the retinoblastoma (Rb) 3 pocket protein family and, in particular, pRb, in the control of periodic transcription during the cell cycle and in cell cycle initiation (2)(3)(4), argued that its function would be conserved in yeast, as well as metazoan species. Like the E2F1/DP transcription factor complex in metazoans, two heterodimeric transcription factor complexes, known as SBF and MBF, are responsible for transcribing genes specific to the G 1 phase of the cell cycle (reviewed in Ref. 5). These complexes are comprised of a shared component, Swi6 and the DNA-binding component, Swi4 or Mbp1, respectively, that recognize specific DNA sequence elements, known as SCBs and MCBs, within the promoters of G 1 -specific genes. Although it had been established that Cln3/CDK was the critical determinant of the timing of cell cycle initiation in the budding yeast, until recently no CDK-regulated target that was functionally analogous to pRb in the regulation of G 1 transcription had been identified (6 -8). Recent work showing that Whi5 has an analogous function to pRb in the Cln3/CDK-mediated repression of G 1 -specific transcription in budding yeast has addressed this apparent inconsistency (1,9).
To date, over 110 pRb binding partners have been identified (10). Although pRb is associated with other functions besides cell cycle regulation, many of which have no obvious analogy in yeast, such as apoptosis and cellular differentiation (11), several pRb-related molecular pathways pertaining to cell cycle regulation are conserved in yeast as well. For example, in metazoans, commitment of cells to transition through the G 1 to S phase of the cell cycle is known as the "restriction point" and involves phosphorylation and subsequent inactivation of pRb by cyclin D-Cdk4/Cdk6 to allow transcriptional activation by E2F1/DP, the pRb target transcription factor during G 1 (reviewed in Refs. 4, 12, and 13). Similarly, in yeast, commitment to transition from G 1 to S phase, known as "Start," requires that Whi5 be inactivated by Cln3/CDK phosphorylation so that it no longer represses transcription via SBF, its target transcription factor complex, thus allowing transcription of G 1 -specific genes (1,9). Also, recent studies suggest that analogous chromatin remodeling complexes in mammals and yeast are linked to G 1 -specific transcriptional regulation. Specifically, both human BRG 1 and hBRM chromatin remodeling complexes and their yeast homologs Swi/Snf have been shown to be involved in the recruitment and activation of their respective G 1 -specific transcription factors (E2F1 and SBF, respectively) (14 -16). Furthermore, these studies in human showed a direct correlation between the pattern of histone acetylation and deacetylation and E2F1 transcriptional activation, suggesting that these histone modification complexes may regulate the accessibility of G 1 -specific transcription factors to their target promoters.
Taken together, these studies demonstrate the high level of conservation between species in the regulatory mechanisms that exist to control entry into the cell cycle. Understanding both the parallels and the distinctions that exist across species is critical to unraveling how abnormalities in these pathways deregulate the cell cycle and lead to human diseases including congenital malformations and cancer. Although the discovery of the parallels between Whi5 and Rb function has broadened our understanding of the molecular mechanisms that control G 1 transcription in yeast, much still remains to be learned about G 1 -specific transcriptional regulation.
In an effort to broaden our understanding of G 1 -mediated transcriptional regulation in budding yeast, we have used multidimensional protein identification technology (MudPIT) to screen interactors of known SBF and MBF components to identify novel regulators of these G 1 -specific transcription factor complexes (1). This screen led to the identification of Whi5, an inhibitor of SBF-mediated transcription that is antagonized by Cln3/CDK, and Nrm1, an MBF-specific transcriptional corepressor required for the timely repression of MBF-regulated transcription during exit from the G 1 phase. In addition, two homologous proteins, Msa1 and Msa2, were identified. Further analysis of Msa1 revealed that it is expressed periodically during the cell cycle, displaying peak mRNA levels at late M/early G 1 and peak protein levels at early G 1 , similar to Cln3. Msa1 interacts with both SBF and MBF at their target promoters. Furthermore, association of Msa1 with promoters is regulated during the cell cycle, with peak binding occurring during G 1 . Msa1 promotes G 1 -specific transcription and cell cycle initiation. Taken together, this study suggests that Msa1 is a novel regulator of G 1 -specific transcription.

MATERIALS AND METHODS
Strains and Methods-All of the Saccharomyces cerevisiae strains used in this study were derived from 15Dub (MATa ade1 leu2-3,112 his2 trp1-1 ura3⌬ ns bar1⌬). The relevant genotypes of the strains utilized in these studies are listed in Table  1. The PCR-based Longtine gene modification method (17) was utilized to create various MSA1, SWI4, MBP1, and SWI6 alleles, including gene deletion, integration of the GAL1 promoter, and epitope tagging with TAP, 3xHA and 13xmyc. All of the strains were grown and maintained in rich YEP medium containing either 2% glucose or 2% galactose at 30°C.
Tandem Affinity Purification (TAP) and MudPIT Analysis-TAP purification and analysis of Msa1 was performed as described previously (1,18). To identify novel interactors, we utilized cln1⌬ cln2⌬ strains carrying a GAL1-regulated allele of CLN3 as the only G 1 cyclin and one of the components of SBF or MBF, either Swi6, Swi4, or Mbp1, TAP-tagged at the endogenous locus. The cells were arrested in the G 1 phase by depletion of CLN3 expression in the absence of galactose or released from arrest by induction of CLN3 expression in the presence of galactose, at which time they were allowed to progress to maximal levels of transcription for analysis. All TAP-tagged proteins were checked by cell size analysis and Western blot to ensure proper function and expression of these proteins. Once purified, protein complexes were digested, and associated proteins were identified using MudPIT analysis as described by McDonald and Yates (19).
Yeast Cell Extract Preparation-The cells growing in log phase were collected by centrifugation and washed with 0.5 ml of cold H 2 O. The pellets were resuspended in lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% Nonidet P-40, 10% glycerol) containing phosphatase inhibitors (50 mM NaF, 0.1 mM sodium orthovanadate, 10 mM sodium pyrophosphate, 5 mM EDTA, 5 mM EGTA), 0.4 mM phenylmethylsulfonyl fluoride, and leupeptin/aprotonin (1 g/ml). To lyse cells, the samples were vortexed with glass beads for four cycles of 40 s with 3-min rest intervals at 4°C between cycles. To collect extracts, the samples were centrifuged twice at maximum speed for 15 min at 4°C.
Co-immunoprecipitation and Western Blotting-0.5-1 mg of yeast protein extract was precleared with 20 l of 50% protein A-Sepharose (Sigma):lysis buffer slurry for 20 min at 4°C. The samples were spun down for 15 min at maximum speed. The supernatant was collected and incubated with 2.5-5 g of appropriate primary antibody (␣-HA (12CA5), ␣-Myc) for 1-2 h at 4°C. The samples were spun at maximum speed for 15 min. The supernatant was again collected and incubated with 25 l of 50% protein A-Sepharose slurry for 2-4 h at 4°C. the samples were spun down briefly to remove supernatant. The remaining beads were washed three times with 1 ml of modified buffer solution (lysis buffer with 0.5% Triton X-100, 0.5 mM NaCl concentration, with phosphatase and protease inhibitors as described previously). 20 l of 4ϫ sample buffer was added to each sample in preparation for Western blot.
Cell Synchronization-For mating pheromone synchronization at G 1 , the cells were grown in YEPD to log phase. ␣ factor was added at a concentration of 100 ng/ml, and the cells were incubated at 30°C for 1.5 h to synchronize. In addition to mating pheromone synchronization, G 1 cells were also obtained for synchronization studies by isolating small G 1 cell populations by centrifugal elutriation as described (20).
Real Time RT-PCR-Total RNA was isolated from cells using the RNeasy kit (Qiagen). The QuantiTech SYBR Green RT-PCR Kit (Qiagen) was used to prepare samples for analysis by RT-PCR. The reactions were then run on the Chromo-4 qPCRI system (MJ Research) using standard RT-PCR conditions. The obtained data were subsequently analyzed using MJ Opticon Monitor Analysis Software 3.0 (MJ Research).
Chromatin Immunoprecipitation Analysis-Chromatin immunoprecipitation analysis was performed as previously described by Flick et al. (21).
Cell Size Analysis-Actively growing cell cultures were diluted and permitted to grow for three to four generations until they reached log phase. Once cells reached log phase, at an approximate A 260 of 0.4 -0.7, 100 l of culture were collected and subjected to sonication for 30 s at maximum power using a Sonicator 3000 Misonix Inc. sonicator. The samples were then diluted in 10 ml of Isoton II Diluent and immediately subjected to cell size analysis using a Coulter Z2 Particle Cell Analyzer (Beckman-Coulter). For each analysis, the cell size of ϳ5-30 ϫ 10 3 was resolved and their distribution determined using Z2 AccuComp software (Beckman-Coulter).

Identification and Characterization of Msa1 and Msa2-
In an effort to further our understanding of G 1 -specific transcriptional regulation, we sought to identify novel components of SBF and MBF. To screen for novel components of these G 1 -specific transcription factors, Swi4, Mbp1, and Swi6 were tagged with a carboxyl-terminal TAP epitope at their endogenous locus in a cln1⌬cln2⌬cln3⌬ background containing a GAL1regulated CLN3 as its only source of G 1 cyclin. TAP purifications of associated protein complexes were performed in cells that were arrested in early G 1 , prior to transcriptional initiation, and in cells in late G 1 phase at the peak of G 1 transcription, as described previously by de Bruin et al. (1).
Analyses of protein complexes obtained from these purifications were achieved using MudPIT to perform mass spectrophometry-based identification of individual components. Along with known interactors of SBF and MBF, several novel proteins were also identified, including Whi5 and Nrm1, which function as negative regulators of SBF and MBF, respectively (1,20). Among the novel interactors, two, encoded by the open reading frames YOR066w and YKR077w, were identified by MudPIT as interactors of MBF and SBF components, respectively. Because we later found that Msa1 interacts with SBF (shown below) and Msa2 interacts with MBF (data not shown), we have named the genes encoding these proteins MSA1 and MSA2, respectively, for "MBF and SBF associated." Msa1 was identified as a novel interactor of MBF components with a high degree of confidence based on sequence coverage and spectral quality (Fig. 1). Specifically, Msa1 was found to interact with Mbp1, demonstrating 4.3 and 14.1% sequence coverage in arrested and released cells, respectively, as well as Swi6, showing 10.5% coverage in released cells (Fig. 1A). Msa2 was identified during our MudPIT analysis as an interactor of Swi4 (Fig. 1A). It was also recently reported to associate with Swi6 based upon genome-wide protein-protein interaction analysis (22). Msa2, which is related to Msa1, sharing 28% amino acid identity and 43% similarity (Fig. 1B), awaits further investigation. Here we report the analysis of the role of Msa1 in the regulation of G 1 -specific transcription.
Expression of Msa1 Is Regulated during the Cell Cycle-Identifying Msa1 as a novel component of MBF implicated it in the regulation of cell cycle initiation events during G 1 . This led to the prediction that Msa1, similar to other regulators of G 1 -specific transcription, including Whi5 and Cln3 (1,8), would accumulate periodically during the cell cycle. In fact, the MSA1 transcript was shown to accumulate during the early G 1 phase (23) and has more recently been assigned to the Yox1/Mcm1 target cluster, which is expressed during the M/G 1 phase (24). To look at accumulation of the Msa1 protein during the cell cycle, we synchronized cells during the G 1 phase with mating pheromone and then released them into the cell cycle. The samples were taken every 15 min and subjected to both Western blot analysis to evaluate protein levels and analysis of budding index to establish the cell cycle position and degree of synchrony. Our results demonstrate that the Msa1 protein accumulates periodically during the cell cycle and achieves peak abundance between 0 and 15 min after release from pheromone arrest, followed by a rapid decrease after 30 min (Fig. 2). This pattern is reinitiated at 75 min after release as cells enter a new cell cycle. The highest level of accumulation of Msa1 occurs during G 1 , prior to budding, and decreases rapidly as cells enter into the S phase consistent with the pattern of gene expression. These studies reveal a cell cycle expression pattern for Msa1 that is similar to that observed for other regulators of G 1 -specific transcription and is consistent with a role for Msa1 in regulation of transcriptional activation at the time of cell cycle initiation.
Msa1 Interacts with both SBF and MBF Components-To confirm the interactions revealed by MudPIT analysis, we performed co-immunoprecipitations of Msa1 with individual components of the SBF and MBF transcription factors. Msa1 was epitope-tagged at its genomic locus along with either Swi4, Mbp1, or Swi6. Whole cell lysates were collected from these strains, and protein complexes were immunoprecipitated using antibody specific for the epitope on Msa1 or the G 1 transcription factor components, as indicated (Fig. 3). Analysis of the constituents of these protein complexes by Western blot revealed that Msa1 co-precipitated with all of the G 1 transcription factor components. Although it was not apparent from the MudPIT analysis of G 1 transcription factor complexes, which revealed an interaction only with MBF, this analysis demonstrates that Msa1 interacts with SBF and MBF. This in not an entirely surprising result because, considering its qualitative nature, the MudPIT technique is an insufficient basis upon which to exclude an interaction between proteins.
Msa1 Associates with SBF-and MBF-regulated Promoters-We have shown that Msa1 interacts with both SBF and MBF transcription factors. Furthermore, we observed that its periodic expression is similar to some of the established regulators of G 1 -specific transcriptional activation. We therefore postulated that Msa1 is involved in the regulation of SBFand/or MBF-regulated gene expression. If so, then we expect Msa1 to be present at SBF-and MBF-regulated promoters. To analyze whether Msa1 is recruited to SCB and MCB promoter elements, we performed chromatin immunoprecipitation assays of Myc-tagged Msa1 along with Swi4 and Mbp1, the DNA-binding components for SBF and MBF, respectively, for comparison. Our results demonstrate that Msa1 associates with both MBF-regulated (RNR1 and CDC21) and SBF-regulated G 1 -specific promoters (CLN2 and SVS1) (Fig. 4). Furthermore, inactivation of Swi4 strongly reduces Msa1 binding to SBF target promoters, whereas inactivation of Mbp1 reduces binding to MBF target promoters, without appreciable effects on binding of the FIGURE 1. Msa1 and Msa2 associate with components of SBF and MBF. A, MudPIT was used to identify novel proteins that associate with TAP-tagged SBF and MBF components Swi6, Swi4, and Mbp1 in cln1⌬ cln2⌬ GAL1-CLN3 cells that were either arrested in G 1 or released into G 1 by induction of CLN3 and allowed to reach maximal levels of expression for analysis. The number of spectra with predicted masses for peptides corresponding to each of the analyzed proteins, Swi6, Swi4, Mbp1, Msa1, and Msa2 were tabulated. % Coverage refers to the percentage of sequence of the full-length protein that was covered by the unique peptides that were identified byMudPIT.PeptidesUnique/No.referstothenumberofuniquepeptidesidentifiedoverthetotalnumberofpeptidesforeachprotein.B,alignmentofthepredictedaminoacid sequence of Msa1 and Msa2. The gray blocks represent aligned amino acids that are similar, whereas the black blocks represent amino acids that are identical.  other transcription factor to its target promoters. We conclude that Msa1 binds to both SBF and MBF target promoters.

Binding of Msa1 to Promoters Is Regulated during the Cell
Cycle-Our finding that Msa1 binds to G 1 -specific promoter elements suggests that this protein functions at promoters to regulate transcription during the G 1 phase of the cell cycle. To test this hypothesis, we analyzed whether changes in binding status occurred during the cell cycle and whether binding was affected by the absence of SBF. To monitor periodic changes in binding status of Msa1 at promoters, we performed chromatin immunoprecipitation analysis of Myc-tagged Msa1 on cells that had been synchronized at G 1 by pheromone arrest and released to allow progression through the cell cycle. Binding of Msa1 to both SBF-and MBF-dependent promoters was observed to fluctuate during the cell cycle (Fig. 5A). The level of binding was maximal during early G 1 prior to maximum budding, after which Msa1 rapidly dissociated from the promoter as cells entered the S phase.
We then determined whether association of Msa1 with promoter elements was dependent on G 1 -specific transcription factors. Binding was analyzed in swi4⌬ strains to determine whether Msa1 association with promoter elements required SBF. In the absence of SWI4, the binding of Msa1 to SBF-specific promoter elements is severely disrupted, whereas binding to MBF-specific promoter elements is not affected (Fig. 5B). This is consistent with the behavior of Msa1 observed in asynchronous populations of swi4⌬ cells (Fig. 4). Furthermore, it appears that, although the level of association of Msa1 to SBF and MBF target promoters is compromised, the periodicity of binding during the cell cycle is unaffected (Fig. 5B). These studies demonstrate that Msa1 associates periodically with SBF and MBF target promoter elements with the highest levels of association occurring at early G 1 . Furthermore, association Msa1 with SBF is dependent on Swi4. Taken together, these results are consistent with a role for Msa1 in G 1 transcriptional regulation.
Msa1 Promotes Cell Cycle Initiation-Because our findings were consistent with a role for Msa1 in G 1 transcriptional acti-  vation, we sought to determine the biological effects of altering MSA1 expression on cell cycle initiation. One biological effect of misregulation of G 1 -specific transcription is a change in cell size. In fact, Whi5, an inhibitor of SBF-dependent transcription, was identified by virtue of a null mutation that gave rise to premature G 1 -specific transcriptional activation and, thereby, a significant reduction cell size (1,25). We analyzed the effect of both overexpressing and depleting Msa1 expression by placing MSA1 under the control of a GAL1 promoter at the endogenous locus. We have confirmed that MSA1 is overexpressed during the cell cycle in these strains (data not shown). Overexpression of GAL1-MSA1 by growth on galactose-containing medium led to a decrease in cell size relative to galactose-grown wild type cells (Fig. 6A, top panel), whereas depletion of Msa1 by growing cells carrying GAL1-MSA1 in glucose led to an increase in cell size (Fig. 6A, bottom panel). A similar result was obtained comparing glucose-grown msa1⌬ mutants to wild type cells growing under the same conditions (data not shown). Taken together, these observations suggest that Msa1 promotes cell cycle initiation. Together with our finding that it interacts with SBF and MBF promoter elements, our results suggest that the effect of Msa1 on cell size may reflect a role in activating G 1 -specific transcription.
To determine whether the changes observed in the relative size of cells misexpressing sMSA1 are a reflection of changes occurring in the timing of cell cycle initiation or of other factors that can affect cell size outside of G 1 , we synchronized wild type cells and cells having GAL1-MSA1 as their sole source of Msa1 protein. Strains were grown under inducing conditions to cause a small cell size, and small daughter cells were isolated by centrifugal elutriation and inoculated into fresh medium under either repressing (glucose) or inducing (galactose) conditions. Small G 1 daughter cells progressing synchronously into the cell cycle were analyzed for cell size and budding. We found that overexpression of Msa1 advances the size at which purified G 1 daughter cells initiate a new cell cycle (Fig. 6B, top panel), and depletion delayed cell cycle initiation (Fig. 6B, bottom panel) as determined by budding, suggesting that the effect of Msa1 on cell size is a reflection of its effect on the timing of cell cycle initiation.
To determine the relationship between Msa1 and established regulators of G 1 -specific transcription, we evaluated the effect of inactivating Msa1 on the cell size in strains deficient in those regulators. Cln3 and Whi5 are the best understood regulators of cell cycle initiation and, thereby, cell size during the G 1 phase. Consequently, the cell size profiles of asynchronous populations of cln3⌬ msa1⌬ and whi5⌬ msa1⌬ cells was determined and compared with that of each of the single mutants (Fig. 7). The msa1⌬ cln3⌬ mutants were comparable in size to cln3⌬ mutants (Fig. 7, top panel), which are large because of the failure to phosphorylate and inactivate Whi5 (1,9), suggesting that Msa1 acts downstream of Cln3. This is consistent with the observation that Msa1 associates with SBF and MBF, which are dependent upon Cln3/CDK for activation (6,7). In contrast, the msa1⌬ whi5⌬ mutant was smaller than the msa1⌬ mutant (Fig.  7, bottom panel), consistent with a role for Msa1 that is independent of Whi5, perhaps a reflection of a Whi5-independent role of Cln3 in modulating the activity of SBF and/or MBF.
Msa1 Promotes Transcriptional Activation of G 1 -specific Genes-Because it has previously been shown that the critical cell size at which bud emergence occurs in yeast S. cerevisiae is FIGURE 5. Msa1 association with G 1 -specific promoter elements varies during cell cycle progression. A, cells containing Msa1-13xmyc were synchronized in G 1 (0Ј) by ␣-factor arrest as described under "Materials and Methods." The cells were then released from arrest and allowed to progress through the cell cycle. The samples were collected every 15 min and subjected to chromatin immunoprecipitation of SBF (SVS1/CLN2)-and MBF (CDC21/RNR1)-regulated promoters by Msa1-Myc. A strain containing no tagged protein (no tag) and whole cell lysates (WCE) from each of the strains analyzed with ACT1 primers are provided as negative and positive controls, respectively. B, synchronization of swi4⌬ cells containing Msa1-13xmyc was performed as described under "Materials and Methods." Briefly, the cells were synchronized in G 1 (0Ј) by ␣-factor arrest then released from arrest and allowed to progress through the cell cycle. The samples were collected every 15 min and subjected to chromatin immunoprecipitation of SBF (SVS1/CLN2)-specific and MBF (CDC21/RNR1)-specific promoters by Msa1-Myc. A strain containing no tagged protein (no tag) and whole cell lysates (WCE) from each of the strains analyzed with ACT1 primers are provided as negative and positive controls, respectively. tightly correlated the timing of G 1 -specific transcriptional activation (6, 7), we followed the expression of the G 1 -specific transcriptional targets CLN2, a target for SBF and primary indicator of Start (4,12,26) and CDC21, a target of MBF encoding thymidylate synthase (20,27). The same wild type and GAL-MSA1-containing populations of synchronized G 1 daughter cells that were analyzed in Fig. 6B were used to analyze expression of the SBF and MBF target genes. Cultures grown in galactose (Fig. 8, left panels) or glucose (Fig. 8, right panels) to induce or repress GAL1-MSA1 expression, respectively, were used to analyze expression of SBF and MBF target genes. As expected, CLN2 and CDC21 mRNA accumulation precedes bud emergence in both wild type and MSA1 overexpressing cells. However, in cells misexpressing MSA1 , the effect on budding reflects the effect on G 1 -specific gene expression (Fig. 8, solid  lines). In MSA1 overexpressing cells, peak CLN2 and CDC21 mRNA levels occur at a smaller mean cell size compared with wild type cells (Fig. 8, left panels). Similarly, peak CLN2 and CDC21 mRNA levels for MSA1 depleted cells occur at a larger cell size than wild type cells (Fig. 8, left panels). Thus, the  We show that the altered level of Msa1 expression can lead to advancement of the timing of G 1 -specific gene expression without affecting the level of expression. Because Cln2 drives cells through Start (4,28), the earlier accumulation of CLN2 is expected to drive cells through Start at a smaller size. A similar, but more pronounced, effect on cell size is observed in cells overexpressing CLN3, a known activator of G 1 -specific transcription (29, 30). Again, the timing of transcriptional induction of CLN2 and CDC21 is delayed in Msa1-depleted cells without a substantial effect on the maximal level of expression (Fig. 8, right panels). These results suggest that the MSA1 acts to regulate transcription by affecting activation of G 1 -specific transcription factors, SBF and MBF, to which it binds. Together these data suggest that the primary activity of Msa1 is as an activator of SBF and MBF that primarily affects the timing, rather than the extent, of transcriptional activation. However, Msa1 can apparently interfere with another as yet unknown pathway affecting the timing of transcriptional activation in the absence of SBF.

DISCUSSION
Coordination of events of the cell division cycle is governed in part by the periodic expression of genes required for the execution of the events that define cell cycle phases. Cell cycle initiation in many eukaryotic cells is defined by the coordinate transcription of family of G 1 -specific genes that leads to the accumulation of activities required for DNA replication and cell growth. Faithful regulation of those genes is paramount for the coordination of cell growth and proliferation.
Here we describe Msa1, a novel regulator of G 1 -specific transcription in the budding yeast S. cerevisiae that is required for the proper timing of cell cycle initiation. Although Msa1 was identified as a protein associated with Mbp1 and Swi6 by MudPIT analysis, we show by co-immunoprecipitation that it binds to both MBF and SBF. Chromatin immunoprecipitation experiments show that it binds both SBF and MBF target promoters. In line with these observations both MBF and SBF target genes are affected by overexpression and depletion of Msa1. Consistent with its role in G 1 -specific transcriptional regulation, the MSA1 gene is expressed early in the G 1 phase and accumulates during arrest by mating pheromone consistent with its importance for timely activation of G 1 -specific gene expression.
Altering Msa1 abundance affects the size at which cells activate transcription of G 1 -specific targets and, as a consequence, the size at which cells produce a new bud, the first overt indicator of the initiation of a new cell cycle (31). Overexpression of Msa1 promotes early transcription and budding. Conversely, depletion of Msa1 leads to a delay in those events. Consistent with a role of Msa1 in the pathway by which Cln3 activates G 1 -specific transcription, inactivating Cln3 abrogated the effect of Msa1 inactivation on cell cycle initiation. This suggests that Msa1 acts downstream in a Cln3-dependent pathway. However, inactivation of the Cln3/CDK target, Whi5, still prompts a cell cycle advance in cells lacking Msa1, suggesting that Whi5 and Msa1 lie, at least in part, on separate pathways. Samples from the same elutriation synchrony experiment shown in Fig. 6B, using wild type (circles) and GAL-MSA1 (squares) cells grown in either galactose-containing (left panels) or glucose-containing (right panels) medium, were analyzed by quantitative RT-PCR to determine the abundance of both CLN2 (top panels) and CDC21 (bottom panels) transcripts (solid lines) and visually examined to determine the extent of budding (dotted lines).
Together, these findings define Msa1 as a size-sensitive component of cell cycle initiation prior to transcriptional activation (32).
Cln3-CDK has multiple substrates in the G 1 -specific transcriptional apparatus including Whi5 and Swi6. In addition to its effect on the activity and localization of Whi5, Cln3/CDK also directly phosphorylates Swi6, a component of both SBF and MBF. Although it is possible that Msa1 affects the interaction of Cln3/CDK with either of those targets, Msa1 may also act as a substrate. Msa1 contains multiple consensus sites for CDK and has been reported to be substrate for CDK in vitro (33). Although no role in regulation of Msa1 has yet been attributed to CDK, its participation in a function downstream of Cln3 and parallel with Whi5 is consistent with such a role.
It is unclear how Msa1 affects the timing of transcriptional activation without affecting the amplitude of transcriptional activation. However, when and to what extent transcription becomes activated do not appear to be tightly coordinated. For example, overexpression of Cln3 appears to advance transcriptional activation without substantially affecting the maximal level of transcription during the G 1 phase. In fact, if the advance in transcriptional activation is associated with an advance in cell cycle entry, maximal expression is largely unaffected. Conversely, if there is a delay in cell cycle entry without a concomitant effect on transcriptional activation, maximal transcription is actually increased (6,7). This is presumably a consequence of the delay in the accumulation of B-type cyclin associated CDKs that limit SBF activation (34,35). Similarly, repression of MBF targets during exit from G 1 phase occurs via negative feedback, in this case mediated by Nrm1 (20). Consequently, it is difficult to predict precisely how an effect on transcriptional activation will affect maximal transcription during the G 1 phase.
Strikingly, Ykr077w, the closest ortholog of Msa1, which we have named Msa2, was also isolated in our MudPIT analysis based upon its interaction with Swi4. It has also been reported to interact with Swi6 in a genome-wide affinity capture study (22). Consistent with a role in transcriptional activation, Msa2 localizes to the nucleus (36) and, like Msa1, was identified in a screen for strong transcriptional activators in a yeast two-hybrid experiment (37). We have shown that both Msa1 and Msa2 associate with G 1 -specific promoters during the G 1 phase (Fig.  4 and data not shown). Although we considered the possibility that Msa1 and Msa2 might play an important redundant function with regard to G 1 -specific transcriptional activation, inactivation of both genes has no synergistic effect on cell size or other overt phenotypes (data not shown). The effect of Msa2 on G 1 -specific transcription remains poorly characterized.
Msa1 is just one of a number of proteins that affect the timing of activation or magnitude of expression of G 1 -specific genes. In addition to Cln3 and Whi5, which are the primary determinants of the timing of transcriptional activation of SBF and MBF target genes, other factors play more subtle roles. Notably, Bck2 can promote transcriptional activation of G 1 -specific genes via a mechanism that is independent of CLN3 (30). Although it clearly acts via a pathway that is independent of Msa1 (data not shown), its mechanism of action is unknown. In addition, Stb1 promotes transcriptional activation of SBF-and MBF-regulated genes (38,39). Like Msa1, Stb1 binds to both transcription factors at promoters and is a target for CDK-dependent phosphorylation. However, unlike Msa1, it primarily affects the magnitude, rather than the timing, of G 1 -specific transcription. 4 Together, these proteins contribute to fine tuning of activation of the G 1 -specific genes at the start of the cell cycle and, perhaps, participate in the integration of nutritional and stress signals with the cell cycle transcription program.
Similarly, G 1 -specific transcription in mammalian cells is subject to a broad array of environmental and internal influences. Consistent with the greater complexity of those cells and genomes, the E2F family of transcription factors is diverse and subject to many types of regulatory signals. We speculate that, like in yeast where G 1 transcription is subject to modulation by multiple factors, the many Rb-associated proteins provide for that regulatory complexity. We are hopeful that understanding the mechanisms by which those yeast modulators, including Msa1, influence G 1 -specific transcription will facilitate the analysis of proteins having similar roles in mammals.