In proliferating S. cerevisiae, genes whose products function in DNA replication are regulated by the MBF transcription factor composed of Mbp1 and Swi6 that binds to consensus MCB sequences in target promoters. We find that during meiotic development a subset of DNA replication genes exemplified by TMP1 and RNR1 are regulated by Mbp1. Deletion of Mbp1 deregulated TMP1 and RNR1 but did not interfere with premeiotic S-phase, meiotic recombination, or spore formation. Surprisingly, deletion of MBP1 had no effect on the expression of CLB5, which is purportedly controlled by MBF. Extensive analysis of the CLB5 promoter revealed that the gene is largely regulated by elements within a 100-bp fragment containing a cluster of MCB sequences. Surprisingly, induction of the CLB5 promoter requires MCB sequences, but not Mbp1, implying that another MCB-binding factor may exist in cells undergoing meiosis. In addition, full activation of CLB5 during meiosis requires Clb5 activity, suggesting that CLB5 may be regulated by a positive feedback mechanism. We further demonstrate that during meiosis MCBs function as effective transcriptional activators independent of MBP1.
IN the budding yeast Saccharomyces cerevisiae a large family of genes is expressed specifically at the boundary between G1- and S-phase (Cho et al. 1998; Spellman et al. 1998). This family can be divided into genes that are regulated by SCB (Swi4 cell cycle boxes) sequences and those regulated by MCB (MluI cell cycle boxes) sequences in their promoters (Koch and Nasmyth 1994). SCB genes are largely regulated by the transcription factor SBF (SCB binding factor), composed of Swi4, a DNA-binding protein, and Swi6, a transcriptional activator (Breeden and Nasmyth 1987; Andrews and Moore 1992). MCB genes are regulated largely by MBF (MCB binding factor), composed of Mbp1 and Swi6 (Koch et al. 1993). SBF regulates two G1 cyclin genes, CLN1 and CLN2, and a series of genes required for cell wall synthesis and bud formation (Stuart and Wittenberg 1994; Igual et al. 1996). MBF regulates a large number of genes that encode factors required for DNA synthesis and repair (Lowndes et al. 1991; Johnston and Lowndes 1992). In addition, MBF has been proposed to control the expression of two B-type cyclins: CLB5 and CLB6 (Schwob and Nasmyth 1993). Inactivation of MBF results in the deregulation of MCB genes such that their periodic expression is replaced with a moderate uniform level of transcript abundance throughout the cell cycle (Dirick et al. 1992; Lowndes et al. 1992). The importance of SBF and MBF gene regulation in controlling cell cycle progression is underscored by the observation that deletion of both MBF and SBF results in a lethal arrest in G1-phase (Nasmyth and Dirick 1991; Koch et al. 1993).
CLB5 and CLB6 encode a pair of B-type cyclins that are expressed and accumulate at the G1/S-phase boundary (Epstein and Cross 1992; Kuhne and Linder 1993; Schwob and Nasmyth 1993). These two cyclins in conjunction with the Cdk Cdc28 have an important role in initiating DNA replication. In the absence of Clb5 and Clb6, cells are viable but experience a delay in initiating DNA synthesis (Epstein and Cross 1992; Schwob and Nasmyth 1993; Donaldson et al. 1998). During mitotic proliferation other members of the B-type cyclin family encoded by CLB1-CLB4 can induce DNA replication and thus have some degree of functional redundancy with CLB5 and CLB6 (Schwob et al. 1994). Despite this redundancy Clb5 and Clb6 are more effective at activating origins of DNA replication than are the other B-type cyclins (Cross et al. 1999; Donaldson 2000).
In addition to mitotic proliferation, budding yeast cells are capable of adopting different developmental fates on the basis of environmental influences. In response to starvation, MATa/MATα diploids can abandon proliferation and embark on a developmental program that proceeds through meiosis and spore formation (reviewed in Kupiec et al. 1997). For simplicity we refer to this process as meiosis or meiotic development. Orderly progression through meiosis is controlled by regulated waves of gene expression (Chu et al. 1998; Primig et al. 2000). There are, at minimum, four families of meiosis-specific genes, classified as early, middle, mid-late, and late (Chu et al. 1998; Primig et al. 2000). This temporal pattern of gene expression helps to ensure that proteins encoded by specific gene families are produced at the time that their functions are required. Induction of the early meiosis-specific gene family is dependent on the transcription factors Ime1 and Ume6 (Kassir et al. 1988; Strich et al. 1994). The early gene family encodes factors required for meiotic DNA replication, chromosome pairing, synaptonemal complex formation, and meiotic recombination (Mitchell 1994). Accumulation of the early gene products leads to the activation of the middle-sporulation genes (Chu and Herskowitz 1998). The members of this family are induced by the transcription factor Ndt80, which binds to a DNA motif referred to as the middle-sporulation element (MSE) in their promoters (Hepworth et al. 1995; Chu and Herskowitz 1998). The middle-sporulation gene family encodes proteins essential for chromosome division, activation of the late genes, and the formation of spores (Chu et al. 1998). Products encoded by the mid-late and late gene classes are required for spore wall synthesis and spore morphogenesis (Mitchell 1994). Temporally regulated meiosis-specific gene expression is not confined to S. cerevisiae. Regulated waves of gene expression are a feature of meiotic progression in Schizosaccharomyces pombe, Drosophila melanogaster, Caenorhabditis elegans, mice, rats, and humans (reviewed by Schlecht and Primig 2003). This cascading pattern of gene regulation helps to ensure orderly progression through meiotic development.
Meiotic development includes features that are distinct from the cell cycle such as synaptonemal complex formation and two consecutive chromosome divisions, meiosis I (MI) and meiosis II (MII), in the absence of an intervening S-phase. Despite these unique features, progress through meiotic development has S-phase in common with the mitotic division cycle. Many of the same gene products required for DNA synthesis in proliferating cells are also utilized during premeiotic S-phase (Simchen 1974; Zamb and Roth 1977). In addition, the same origins of replication utilized by proliferating cells are also used in premeiotic S-phase (Collins and Newlon 1994). We have previously established that entry into premeiotic S-phase is dependent on the S-phase cyclins Clb5 and Clb6 (Stuart and Wittenberg 1998). In addition, Cdc28, the kinase activated by these cyclins, is required for premeiotic S-phase (Stuart and Wittenberg 1998; Benjamin et al. 2003). Thus the preponderance of data suggests that meiotic and mitotic S-phases may be similarly regulated.
CLB5 and many of the genes whose products are required for mitotic S-phase are regulated by MBF, by SBF, or by both (Koch et al. 1993; Iver et al. 2001). However, the importance of MBF and SBF in promoting premeiotic S-phase has not been extensively investigated. Many of the genes regulated predominantly by SBF are not expressed in meiosis and enforced expression of CLN1 or CLN2 blocks entry into meiosis (Colomina et al. 1999; Purnapatre et al. 2002). In addition, it has been demonstrated that cells lacking Swi4 or Swi6 effectively complete meiosis and spore formation (Leem et al. 1998). In contrast, a global investigation of gene expression demonstrated that a majority of the MBF-regulated genes are expressed during meiotic development in S. cerevisiae (Iver et al. 2001).
Our investigation has revealed that Mbp1 influences a subset of MCB-regulated genes during meiotic development, but is not essential for meiosis or spore formation. Surprisingly, we found that Mbp1 does not control the expression of CLB5 during meiotic development. However, MCB sequences in the CLB5 promoter are required for effective expression of CLB5, implying that another MCB-binding factor may regulate CLB5 during meiosis. These data reveal a remarkable complexity in the regulation of CLB5 during meiosis and suggest that there may be subclasses of MCB-containing genes that are regulated by different MCB-binding factors.
MATERIALS AND METHODS
Strains and growth conditions:
All strains used in this study are in the SK1 genetic background (Kane and Roth 1974) and were derived from DSY1030 (MATa lys2 ho::LYS2 ura3 leu2::hisG trp1::hisG arg4Bgl his4X) and DSY1031 (MATα lys2 ho::LYS2 ura3 leu2::hisG trp1::hisG arg4Nsp his4B). The strains used in this study and their relevant genotypes are listed in Table 1. All of the strains were generated by standard genetic procedures (Rose et al. 1990). Yeast strains were routinely propagated in YEP (1% yeast extract, 2% peptone, 30 mg/liter adenine) supplemented with 2% dextrose (YEPD) or 2% potassium acetate (YEPKAc). The sporulation medium (SPM) used in this study was 1% potassium acetate. Synchronous meiotic development and sporulation were accomplished as previously described (Stuart and Wittenberg 1998). Meiotic recombination was assayed as previously described (Padmore et al. 1991). The clb5::KanR, clb6::TRP1, swi4::LEU2, mbp1::URA3, ndt80::KanR deletions and Ndt80-HA have been previously described (Stuart and Wittenberg 1994, 1995, 1998; Sopko et al. 2002). Mbp1-Myc was generated by amplification of the 3′-end of MBP1, including the C terminus of the open reading frame. This was ligated with YIplac204 and site-directed mutagenesis was used to insert a cassette encoding 15 copies of the c-Myc epitope. The tagging construct was targeted for integration at the C terminus of MBP1. SWI4 was tagged at the C terminus with three tandem copies of the HA epitope using the vector pURA-SWI4HA as described (Flick et al. 2003). The CLB5 open reading frame was amplified as a BamHI-SacI fragment and placed under the regulation of 900 bp of IME2 promoter sequence (IME2-CLB5) or 800 bp of MET3 promoter sequence (MET3-CLB5). In both cases these constructs were targeted for integration at the URA3 locus in the clb5 clb6 haploid strains DSY1064 and DSY1065; diploids were generated by mating. For analysis of the CLB5 promoter a 2.6-kb XhoI-SacI fragment containing the CLB5 gene was ligated with YIplac211. The CLB5 gene in this construct included six tandem copies of the HA epitope at its C terminus. This construct included 980 bp of sequence upstream of the CLB5 start codon and was sufficient to provide proper regulation of the CLB5 gene. A nonfunctional clb5 reporter gene was generated by deleting the 250-bp BspEI-Bsu36I fragment from the CLB5 open reading frame; additionally, the transcriptional stop sequence in this construct was altered to allow us to distinguish it from the endogenous CLB5 mRNA. All MCB mutations were generated by site-directed mutagenesis as described (Stuart and Wittenberg 1994) and the promoter constructs were sequenced to ensure the inclusion of the mutations and that no additional mutations had been generated. The following mutations were generated: MCB (−389) GACGCGCC changed to GACGAGCC, MCB (−354) GGCGCGTC changed to GGACGTTC, MCB (−337) CACGCGCT changed to CACGAGCT, MCB (−315) TAGCGCCC changed to TAGCATGC, and MCB (−248) ACCGCGAA changed to ACCTAGAA. The number in parentheses refers to the position of the putative MCB relative to the CLB5 start codon; the core of each putative MCB is underlined. The MSE at position −230 was changed from AACGCAAAT to GATCCGGCT by PCR-mediated mutagenesis (Horton et al. 1989). The 176-bp deletion in the Δ179 promoter and the 100-bp deletion created in the Δ178 promoter were generated by PCR-mediated mutagenesis (Horton et al. 1989). The plasmid that places CLB5 under the regulation of only two MCB sequences (MCB-CLB5) was constructed by annealing the oligonucleotides WTKXF 5′-CTCTCGAGTGAAGACGCGCCCTTGATGGC-3′ and WTKXR 5′-TCGAGCCATCAAGGGCGCGTCTTCACTCGAGAGGTAC-3′ and inserting this duplex into the Δ179 promoter so that the oligonucleotide replaced the 176 bp deleted from the CLB5 promoter. CLB5 under the regulation of one mutant and one wild-type MCB (mcbx-CLB5) was similarly constructed; however, the oligonucleotides used encoded a mutant MCB sequence in which the core CGCG had been changed to CGAG. The clb5 reporter constructs were integrated at the URA3 locus in wild-type CLB5 CLB6 strains DSY1030 and DSY1031. The CLB5 constructs were integrated into the clb5 clb6 mutant strains DSY1064 and DSY1065. The presence of the constructs in single copy at the URA3 locus was confirmed by Southern blot analysis. For the experiments shown in Figure 6, DSY1064 was transformed with constructs in which CLB5 was driven either by two wild-type MCB elements (MCB-CLB5) or by one wild-type and one mutant MCB (mcbx-CLB5), and each strain was crossed to DSY1475, an mbp1::KanR strain. The diploid was sporulated and colonies derived from spores that were clb5::KanR clb6::TRP1 mbp1::KanR and URA3::MCB-CLB5 or URA3::mcbx-CLB5 were identified.
RNA extraction and Northern blot analysis:
Total RNA was isolated and subjected to Northern blot analysis as previously described (Stuart and Wittenberg 1994). Northern blots were probed for CLB5 with a 1.5-kb AflII DNA fragment containing the CLB5 open reading frame. SPS1 and SPS2 were detected with a 3-kb ClaI fragment from plasmid p18 (Percival-Smith and Segall 1984). ACT1 was detected with a 1.6-kb BamHI-HindIII fragment containing the ACTI gene. MBP1, TMP1, and RNR1 were all detected with PCR fragments encoding the open reading frame of their respective genes. All Northern blot probes were labeled with [α-32P]dCTP by random priming. Quantitation of Northern blot signals was achieved by the use of a Molecular Dynamics (Sunnyvale, CA) STORM phosphorimager.
Protein extraction, kinase assays, and Western blotting:
Protein samples used for Western blot were prepared by trichloroacetic acid extraction as described (Foiani et al. 1995). Western blots were probed with monoclonal antibody 12CA5 (Babco) at a dilution of 1:10,000 to detect Swi4-HA and Ndt80-HA or with monoclonal antibody 9E10 (Babco) at a dilution of 1:10,000 to detect Mbp1-Myc. As a loading control blots were probed with monoclonal antibody anti-PSTAIR (Sigma, St. Louis) at a dilution of 1:10,000. All of the primary antibodies were detected with horseradish peroxidase conjugated anti-mouse antibodies (Jackson Labs). Immunoprecipitation and histone H1 kinase assay were preformed as described (Stuart and Wittenberg 1998). The kinase reactions were separated on 10% SDS-polyacrylamide gels that were subsequently dried and the phosphorylated substrates were detected by exposing the dried gels to phosphor-storage screens that were subsequently developed on a Molecular Dynamics STORM phosphorimager.
Sporulation was assayed by microscopic examination of cultures that had been incubated in SPM for 24 hr. Two hundred cells per culture were counted and the percentage of cells that had formed asci were scored. Progression through the meiotic chromosome divisions MI and MII was assayed by staining the nuclear DNA with 4′,6-diamidino-2-phenylindole (DAPI) as previously described (Stuart and Wittenberg 1998). The nuclei were visualized and counted with a Zeiss Axioskop II. Nuclear DNA content of fixed propidium-iodide-stained cells was determine by fluorescence-activated cell sorting (FACS) as previously described (Stuart and Wittenberg 1998). Cell volume data for cells growing asynchronously in YEPD were collected using a Coulter (Hialeah, FL) Z2 particle size analyzer.
The S-phase cyclin CLB5 is essential for effective sporulation:
Wild-type SK1 diploid cells sporulate at a frequency approaching or >90%. In contrast, clb5 clb6 cells cannot form complete asci 24 hr following the induction of meiotic development (Table 2). The meiotic defect in clb5 clb6 cells is most likely due to an absence of the Clb-dependent Cdk activity needed to initiate S-phase. However, another explanation could involve defects in chromosome metabolism occurring during mitotic proliferation that persist to prevent successful premeiotic DNA replication. We tested this possibility by generating a strain whose only source of S-Cdk was Clb5 regulated by a meiosis-specific IME2 promoter. During mitotic growth the IME2 promoter is repressed; however, in sporulation medium the IME2 promoter is induced and CLB5 is expressed. The clb5 clb6 IME2-CLB5 cells proliferating in YEPD medium were similar in size to clb5 clb6 mutants (110 fl), indicating that they suffer a similar delay in entering S-phase. However, clb5 clb6 IME2-CLB5 cells initiate meiotic chromosome divisions at the same time as CLB5 CLB6 cells and achieve a similar degree of tetrad formation and spore viability (Table 2). A similar outcome was observed when the only source of S-Cdk was CLB5 regulated by a methionine repressible MET3 promoter. A clb5 clb6 MET3-CLB5 diploid strain was grown in rich medium supplemented with methionine (MET3 promoter off) and was then induced to enter meiosis in SPM lacking methionine (MET3 promoter on). These cells effectively progressed through meiotic S-phase and completed both meiotic divisions (Table 2). The number of tetrads formed by these cells was somewhat reduced but spore viability was similar to that achieved by wild-type cells (Table 2). In contrast, if the same cells were grown in the absence of methionine (MET3 promoter on) and were then induced to enter meiosis in the presence of methionine (MET3 promoter off), the cells displayed a significant reduction in the number of cells forming tetrads and a reduction in spore viability (Table 2). The concentration of methionine used in this medium (1 mm) caused a minor delay in the progression of wild-type cells but within 10 hr >80% had completed MII (data not shown). These observations imply that Clb5 must be synthesized de novo in meiosis to promote effective progression through meiosis and spore formation.
Transcriptional profile of CLB5 during meiotic development:
In cells induced to enter meiotic development CLB5, mRNA began to accumulate after 2 hr, concurrent with the onset of premeiotic S-phase, followed by a peak of transcript accumulation between 5 and 8 hr, coincident with the MI and MII chromosome divisions (Figure 1A, wild type). The induction of CLB5 observed under these conditions is dependent on the initiation of the meiotic program since ime1 mutants that cannot induce meiosis-specific genes did not accumulate CLB5 mRNA (Figure 1A, Δime1). In contrast, CLB5 mRNA did accumulate in ime2 mutants that are competent to activate early genes but are defective in middle gene activation (Figure 1A, ime2Δ). As expected, CLB5 mRNA accumulated effectively in swi4 mutants; indeed, the peak of CLB5 transcript accumulation is slightly earlier than that in wild-type cells (Figure 1A, swi4Δ). Surprisingly, we did not observe any difference in the meiotic mRNA profile of CLB5 when comparing MBP1 wild-type diploids with mbp1 mutants (Figure 1A; compare wild type with Δmbp1). This surprising observation compelled us to investigate the activity of Mbp1 further. An analysis of transcript abundance verified that MBP1 mRNA displayed a relatively uniform level throughout the meiotic time course (Figure 1B). Additionally, Mbp1 protein abundance does not change throughout meiotic development (Figure 1C, Mbp1-Myc). Similarly, it has been shown that Swi6 is present throughout meiotic development (Clyne et al. 2003). The constitutive abundance of Mbp1 and Swi6 is in contrast to the apparent instability of Swi4. We were able to detect full-length Swi4-HA in proliferating cells but only lower molecular weight degradation products could be detected when cells initiated meiosis (Figure 1C, Swi4-HA). Ndt80 is a meiosis-specific regulator of CLB5 and the entire middle-sporulation gene family (Chu and Herskowitz 1998). Ndt80 accumulates just prior to the induction of the middle-sporulation genes, coincident with the burst of CLB5 expression at ∼6 hr into the meiotic time course (Figure 1C, Ndt80-HA). Deletion of NDT80 abolished the peak accumulation of CLB5 transcript normally seen between 5 and 8 hr, resulting in a relatively constitutive CLB5 expression (Figure 1A, Δndt80). We considered it possible that both Mbp1 and Ndt80 regulate CLB5 but that the activity of Mbp1 might be masked by the burst of CLB5 mRNA induced by Ndt80. We tested this possibility by deleting both NDT80 and MBP1. The timing of CLB5 accumulation and the abundance of the CLB5 transcript were very similar in either ndt80 or ndt80 mbp1 diploid cells (Figure 1A; compare Δndt80 and Δmbp1 Δndt80). These data suggest that Mbp1 does not play a critical role in the regulation of CLB5 during meiotic development (Figure 1A).
Mbp1 is not required for efficient, timely DNA replication and meiotic recombination:
Although MBP1 does not appear to have a profound regulatory effect on CLB5, we considered it possible that Mbp1 might influence the expression of other genes and thus influence progression through S-phase. Analysis of MBP1 NDT80, mbp1 NDT80, and MBP1 ndt80 diploid cells by FACS indicated that neither Mbp1 nor Ndt80 is required for efficient premeiotic DNA replication; wild-type, mbp1, and ndt80 cells all successfully completed S-phase within 4 hr after being induced to initiate meiotic development (Figure 2A).
Extensive homologous recombination is a hallmark of meiotic development in most eukaryotes. In S. cerevisiae this process is dependent on successful completion of DNA replication (Borde et al. 2000; Smith et al. 2001). Delays in the initiation or progression of premeiotic S-phase would be manifested as a delay in the accumulation of recombination events. We compared MBP1 and mbp1 diploids in a “return to growth assay” where intragenic meiotic recombination was analyzed at two separate loci, ARG4 and HIS4. The rate at which recombinants could be recovered was very similar between the mbp1 mutant and MBP1 wild-type cells, suggesting that deletion of MBP1 does not lead to any defects in either meiotic DNA replication or recombination (Figure 2B).
Mbp1 regulates a subset of MCB-controlled genes during meiotic development:
Although Mbp1 does not appear to regulate CLB5 in meiosis, we wished to determine if it might influence the expression of other MCB-containing promoters. Northern blot analysis showed that in wild-type cells two established MBF-regulated genes, RNR1 and TMP1, display defined temporal patterns of transcript accumulation with discrete peaks of transcript abundance occurring 2–6 hr following the induction of meiotic development (Figure 2C, MBP1/MBP1). These peaks of transcript abundance occur coincident with the time at which premeiotic S-phase would be initiated. In mbp1 mutants, these distinct peaks were abolished and replaced by a relatively uniform pattern of transcript abundance that gradually decreased as the meiotic program neared completion (Figure 2C, mbp1/mbp1). Although inactivation of MBP1 reduced the periodic expression of RNR1 and TMP1 during meiotic development, this mutation did not have a general effect on meiotic gene expression as CLB5, SPS1, and SPS2 expression was unaffected (Figure 2C).
Mutational analysis of the CLB5 promoter:
Sequence scanning of the CLB5 promoter revealed a cluster of potential MCB elements and a putative MSE element (Figure 3A). To investigate the importance of these regulatory elements, we generated a series of mutations in the CLB5 promoter and fused these mutant promoters to a clb5 reporter gene. This reporter gene could be monitored in wild-type cells as the mutant transcript displays a mobility distinct from that of the endogenous CLB5 mRNA on Northern blots (Figure 3A; the reporter transcript labeled clb5 migrates more slowly than the endogenous CLB5 transcript labeled CLB5).
Northern blot analysis of the reporter gene under the regulation of a wild-type CLB5 promoter revealed a pattern of mRNA accumulation identical to that displayed by the endogenous CLB5 (Figure 3A). To characterize the role of sequence elements within the CLB5 promoter, we examined the expression of two constructs that contained deletions of the CLB5 upstream region encompassing the major recognizable regulatory elements in the CLB5 promoter. The deletion in Δ179 removes sequences between −222 and −389 relative to the ATG codon of CLB5. This promoter mutation resulted in the accumulation of significantly reduced levels of clb5 transcript throughout the meiotic time course (Figure 3B, Δ179). Although a single putative MCB had been juxtaposed closer to the clb5 open reading frame, it appeared to have a very minor effect on inducing clb5 expression. In addition, transcripts driven by the Δ179 promoter did not appear to display any middle-sporulation-specific accumulation; this is consistent with the likelihood that we eliminated the potential Ndt80-binding site. This result suggests that all of the major meiosis-specific regulatory elements in the CLB5 promoter reside between −222 and −389; consequently, we have designated this region as UASCLB5. The Δ178 mutation eliminates four of the five consensus MCB elements in the CLB5 promoter but retains two potential MCBs and the putative MSE sequence (Figure 3C). The Δ178 promoter also induced the accumulation of a significantly reduced abundance of clb5 transcripts; however, in this case the clb5 transcript displayed a weakly regulated pattern of accumulation with a peak occurring coincident with middle sporulation as evinced by the accumulation of SPS1 and SPS2 transcripts (Figure 3C, SPS1). This pattern of mRNA accumulation is consistent with the presence of an MSE element in the Δ178 promoter. The clb5 mRNA driven by the Δ178 promoter accumulated to much lower levels than that driven by a wild-type CLB5 promoter even during the middle phase of sporulation when Ndt80 would be expected to drive transcription via the MSE sequence. It is unclear whether this effect can be attributed solely to the potential MSE being a relatively weak activator or whether this deletion has removed flanking sequences that enhance Ndt80-binding interaction with cofactors.
The MSE element at position −230 in the CLB5 promoter was anticipated to be responsible for the Ndt80-dependent transcription of CLB5. However, inactivation of this sequence by site-directed mutagenesis did not significantly alter expression of the clb5 reporter gene (Figure 3D). In particular, the clb5 mRNA continued to display a peak of accumulation between 4 and 8 hr, consistent with activation by Ndt80. This observation implies that there may be another Ndt80-binding site in the CLB5 promoter. Since Ndt80-dependent expression was lost when UASCLB5 was deleted, Ndt80 must interact with a sequence between −222 and −398. Indeed, Ndt80 will bind effectively to this MSE mutant promoter in vitro (S. A. Raithatha and D. T. Stuart, unpublished results).
Mbp1 does not appear to influence the transcriptional regulation of CLB5 during sporulation. However, since multiple MCB sites reside within the CLB5 promoter, we decided to scrutinize these potential MCBs to determine whether or not these sites are active regulatory elements during meiosis. Mutation of five of the six MCB sequences (Δmcb5) in the reporter gene promoter had a very modest effect on its transcription relative to a wild-type promoter (Figure 3E). This initial analysis suggested that all important regulatory elements other than MCBs must be located between −222 and −298 in the CLB5 promoter.
We next wanted to determine if alterations to the CLB5 promoter would be of any physiological consequence to the cells. To investigate this, we generated constructs in which a functional CLB5 open reading frame was placed under the regulation of a wild-type or one of the mutant CLB5 promoters. These constructs were introduced into clb5 clb6 cells so that their only source of CLB5 was that under the regulation of promoter constructs that we generated. Northern blot analysis of the CLB5 open reading frame in these mutant diploids gave results that largely mimicked the result seen with the clb5 reporter in CLB5 CLB6 diploids. The Δ178 and Δ179 promoters were profoundly defective in driving the accumulation of CLB5 transcripts (Figure 4A, Δ178 and Δ179). However, site-directed inactivation of five MCBs in the CLB5 promoter resulted in a much greater reduction in transcript abundance than was observed for the clb5 reporter gene (Figure 4A, Δmcb5). The difference between Δmcb5 clb5 and Δmcb5 CLB5 may be due to a positive-feedback effect (see discussion). We quantitated the CLB5 mRNA relative to ACT1 mRNA that accumulated at 1, 4, 6, and 12 hr following induction of meiosis in cells whose only source of CLB5 was driven by a wild-type promoter, a Δ178 promoter, a Δmcb4 promoter in which four of six MCBs had been inactivated, or a Δmcb5 promoter. It is clear that, compared to the wild-type promoter, a Δ178 promoter yielded much lower levels of transcript abundance at all points throughout the time course (Figure 4B). Inactivation of four of the six MCBs (at positions −345, −337, −315, and −248) reduced accumulation of CLB5 mRNA at the 4- and 6-hr time points, and there was a delay in the peak accumulation of CLB5 mRNA (Figure 4B). Deletion of five of the six MCBs in the CLB5 promoter (Δmcb5) reduced the level of CLB5 transcript accumulation particularly at the 4- and 6-hr time points and resulted in a significant delay before the CLB5 transcripts could accumulate above the initial levels (Figure 4B).
MCB elements are frequently found in multiple copies within the promoters of genes that they regulate (Johnston and Lowndes 1992). We observed that mutation of five of the six MCBs in the CLB5 promoter reduced transcription much more than deletion of four of the MCB sequences. This might suggest a cooperative or an additive nature to the transcriptional activation directed by the MCBs. However, an alternative possibility is that the fifth MCB at position −389 is particularly potent at activating transcription. To determine whether inactivation of the MCB at −389 alone was responsible for the decrease in mRNA driven by the Δmcb5 promoter, a single mutation within this MCB was introduced into an otherwise wild-type CLB5 promoter (Figure 4A, Δmcb1). Homozygous clb5 clb6 Δmcb1-CLB5 diploids displayed a wild-type pattern of CLB5 mRNA expression (compare CLB5 and mcb1 in Figure 4A). Hence, mutation of the MCB at −389 alone is not responsible for the large decrease in CLB5 transcript abundance driven by the Δmcb5 promoter during meiotic development.
Histone H1 kinase activity associated with Clb5 follows a profile directly related to the pattern of mRNA accumulation; this is true both during mitotic proliferation and during meiotic development. Inactivation of four MCBs in the CLB5 promoter resulted in a decrease in the amount of Clb5-associated kinase that could be isolated from an asynchronous population of mitotically proliferating cells (Figure 4C; compare CLB5, asynch lane, with Δmcb4, asynch lane). A further reduction in the abundance of kinase activity resulted from deletion of five MCB elements from the CLB5 promoter (Figure 4C, Δmcb5). During meiotic development, cells whose CLB5 is under the regulation of a wild-type promoter accumulated Clb5-associated kinase activity about 2 hr after the initiation of meiosis. This kinase activity reached peak abundance between 6 and 8 hr (Figure 4C, CLB5). Although Clb5-dependent kinase activity peaks during the middle-sporulation events, the kinase activity detected between 2 and 4 hr following the induction of meiotic development is likely the activity relevant to fulfilling its critical role in promoting premeiotic DNA replication. The loss of four MCB elements from the CLB5 promoter (Δmcb4) reduced kinase activity during early time periods, eventually allowing peak levels to be achieved after 10 hr (Figure 4C, Δmcb4). The CLB5-associated kinase activity in the Δmcb4 strain, while delayed, did eventually accumulate to levels higher than that achieved by wild-type cells at 4 hr. We predicted that this level of Clb5-associated kinase activity would be sufficient to promote premeiotic S-phase and indeed these cells are capable of completing DNA replication although with some delay relative to wild-type cells (Figure 4D). In contrast, inactivation of five MCB regulatory elements within the CLB5 promoter (Δmcb5) produced very low levels of Clb5-associated kinase activity and resulted in a defect in DNA replication (Figure 4C, Δmcb5; Figure 4D).
CLB5 promoter mutations result in altered cell morphology and reduced sporulation efficiency:
Due to a delay in the initiation of DNA replication, proliferating clb5 clb6 cells spend a greater proportion of time in G1 and achieve a larger cell volume (Figure 5: compare A and B). Introduction of the wild-type promoter-driven CLB5 construct into clb5 clb6-deficient cells completely rescued the elongated phenotype and reduced cell volume from 114.64 fl to 74.67 fl, similar to wild-type cells, indicating that functionally relevant CLB5 expression was achieved from this construct (Figure 5C). A promoter containing deletions of UASCLB5 (Δ178) maintained the elongated, large-cell clb5 clb6 phenotype (Figure 5D). Inactivation of four MCBs may still allow adequate expression of CLB5 necessary to promote DNA replication since these cells did not appear to display as severe a phenotype as the clb5 clb6 mutant (Figure 5E). In contrast, Δmcb5 cells strongly exhibited the mutant morphology and increased cell volume during mitotic proliferation, suggesting insufficient expression of CLB5 (Figure 5F).
Introduction of a wild-type promoter-driven CLB5 restored the ability of clb5 clb6 diploid cells to complete both meiotic divisions and achieve wild-type levels of tetrad formation (Figure 5G). In contrast, CLB5 regulated by the deletion promoter Δ178 did not rescue the clb5 clb6 sporulation defect (Figure 5G), consistent with the inability of this promoter to drive sufficient CLB5 to restore wild-type morphology to proliferating clb5 clb6 diploid cells (Figure 5D). Inactivation of a single MCB element at −398 in the CLB5 promoter had little or no effect on sporulation efficiency (Figure 5G, Δmcb1). In contrast, loss of four MCBs (Δmcb4) reduced the efficiency of completing meiosis to ∼30%, whereas mutating five MCBs (Δmcb5) reduced the efficiency of cells completing meiosis to <5% (Figure 5G).
MCBs are effective upstream-activating elements during meiosis:
It has been demonstrated that in cells lacking Mbp1 many MBF-regulated genes lose the periodicity in their expression, albeit maintaining a relatively high uniform level of transcript abundance (Koch et al. 1993). Although Mbp1 may not be necessary for the expression of all MCB-regulated genes, the MCBs themselves may be critical for transcriptional activation of these genes. We investigated the effectiveness of MCB elements as transcriptional-activating sequences during meiosis by generating constructs in which a CLB5 open reading frame was placed under the regulation of either a wild-type MCB sequence (MCB) or a mutant (mcbx). Either the wild-type or the mutant MCB was inserted into the CLB5 Δ179 promoter. Although this promoter has a single core MCB sequence, we have shown that it is insufficient to promote CLB5 expression (Figure 3B). In two independent experiments, the addition of a wild-type MCB to this promoter induced the accumulation of CLB5 mRNA to a significantly greater level than was achieved when a mutant MCB was added to it (Figure 6A, top). The physiological consequences of this elevated level of CLB5 mRNA are reflected in the higher degree of spore formation achieved by cells expressing CLB5 from a promoter containing the wild-type MCB elements. Greater than 50% of cells expressing CLB5 from a wild-type MCB promoter successfully completed two meiotic divisions; this contrasts with only 3% seen in cells expressing CLB5 from an MCB mutant promoter (Figure 6B). We wanted to determine whether Mbp1 was responsible for the meiosis-specific transcription driven by this MCB element, so yeast strains lacking Mbp1 carrying CLB5 driven by either a wild-type or a mutant MCB were generated. In these cells, CLB5 driven by the wild-type MCB was expressed at elevated levels throughout meiotic development when compared to expression from the mutant MCB. At early times in meiosis the relative levels of mRNA were similar to cells harboring a functional MBP1; however, at later times the reporter in MBP1 strains achieved higher levels. (Figure 6A, bottom). The reason for this difference is unclear but may be an indirect effect driven by other MBP1-regulated genes. Those mbp1 cells expressing CLB5 from the wild-type MCB also achieved a sporulation efficiency of ∼50%, similar to that seen in MBP1 cells (Figure 6B). MBP1 cells expressing CLB5 from a mutant MCB promoter achieved a sporulation efficiency of 3%, whereas mbp1 cells expressing CLB5 from a mutant MCB promoter achieved a slightly greater sporulation efficiency (∼12%; Figure 6B). These data indicate that the MCB element can act in cis to activate gene expression during meiotic development and that the ability of the MCB to function as an activator (at least in the context of the CLB5 promoter) is not dependent on Mbp1.
Clb5 expression is essential for efficient completion of meiotic development:
The simplest explanation for the meiotic defect in clb5 clb6 mutants is that these cyclins must be expressed during meiotic development to promote meiotic S-phase and perhaps other functions. While it may be that clb5 clb6 mutant cells do experience a defect in chromosome metabolism during mitotic proliferation that hinders premeiotic S-phase, it is clear that expression of CLB5 in meiosis is sufficient to correct that defect. The fact that cells expressing CLB5 during proliferation are defective in completing meiotic development when they are deprived of CLB5 synthesis in meiosis further suggests that Clb5 expressed during mitotic proliferation is insufficient to promote premeiotic S-phase. The need for CLB5 expression to induce S-phase during both mitotic proliferation and meiotic development has clearly prompted the development of a mechanism to promote expression of CLB5 during both of these alternative fates. In this regard we predicted that CLB5 would be under the same form of regulation as other genes required for DNA synthesis. However, our analysis has suggested that CLB5 is regulated in a fashion distinct from other genes encoding factors required for DNA synthesis.
Regulation of CLB5 expression during meiotic development:
Genes regulated by MBF predominantly encode factors involved in DNA replication. With this in mind we anticipated that MBF would be involved in promoting premeiotic S-phase and, potentially, CLB5 expression, while SBF might not. Mbp1 is stable during meiosis while Swi4 is degraded and this may be a key regulatory mechanism that allows the activation of MBF-regulated genes while preventing activation of SBF-regulated genes. This partitioning of SBF- and MBF-regulated genes is critical for meiosis since expression of the SBF-regulated CLN1 and CLN2 genes prevents entry into meiotic development.
Synchronized cultures of proliferating mbp1 mutant cells display a reduction in the periodicity of CLB5 expression during the cell cycle (Koch et al. 1993). Given this, it is surprising that the pattern of CLB5 transcript accumulation during meiotic development is unaffected by the deletion of MBP1. This implies that another factor must regulate CLB5 during meiosis. When proliferating cells are synchronized by deprivation and then reinduction of G1 cyclins, CLB5 mRNA displays an identical pattern of accumulation in MBP1 and mbp1 strains, whereas POL1 and TMP1 mRNA both display reduced periodicity in mbp1 strains (Koch et al. 1993). This suggests that high levels of Cln/Cdk activity can promote periodic expression of CLB5 independently of Mbp1 and provides evidence that CLB5 regulation is distinct from other genes controlled by MCB sequences in their promoters.
Although MBP1 does not appear to regulate CLB5 during meiosis, it does regulate other MCB-containing genes. Mbp1 regulates genes whose products play roles in DNA replication, yet mbp1 mutants display little or no defect in the timing or completion of meiotic S-phase, meiotic recombination, or spore viability. This is somewhat surprising since it has been reported that swi6 mutants display reduced frequencies of recombination and decreased spore viability (Leem et al. 1998). This has been attributed to a reduction in the expression of genes required for meiotic recombination (Leem et al. 1998). However, Swi6 also participates in the SBF complex that regulates genes whose products function in cell wall synthesis. Thus the reduced viability of swi6 spores may be more related to cell wall integrity than to defects that occur during meiosis.
In contrast to RNR1 and TMP1, CLB5 displays a peak of transcript accumulation that is coincident with the onset of MI and MII and is consistent with CLB5 being regulated by Ndt80. However, since neither Ndt80 nor Mbp1 appears to influence the expression of CLB5 early in meiosis, it is likely that another activator regulates CLB5. In support of this hypothesis the activation of CLB5 in meiosis is dependent on the meiosis-specific transcription factor Ime1. Despite the requirement for Ime1, CLB5 is not regulated like other early meiosis-specific genes. The CLB5 promoter lacks any consensus URS1 sequences and there are no T4C elements or consensus-binding sites for Abf1. Thus, it is likely that there are both regulators and regulatory elements in the CLB5 promoter that remain to be characterized.
On the basis of the ability of MBF to modulate CLB5 transcript abundance during mitotic proliferation and the ability of Mbp1 to bind the CLB5 promoter in vitro, the presumption has been that MCBs regulate CLB5 by binding Mbp1. Although MBF is not required for meiotic expression of CLB5, the MCB elements are regulators of its expression. In addition, this analysis has revealed that DNA sequences between −298 and −398 of the CLB5 promoter contain as-yet-unidentified elements that promote expression of CLB5 in meiosis. This DNA fragment contains four consensus MCB sequences but mutation of these elements has a very small effect on the expression of the reporter gene in a wild-type cell. Clearly, this indicates that the −298 to −398 fragment contains important regulatory elements that have not been previously identified. This is not surprising considering that many meiosis-specific genes lack any of the sequence elements known to act as binding sites for transcriptional regulators. Interestingly the −298 to −398 region of the CLB5 promoter contains two sequences (CGCGCTTT and CTCACTTT) that match a consensus element identified in at least 15 genes expressed during middle sporulation (Chu et al. 1998). In addition, this region of the CLB5 promoter also contains the sequence GGTACAAAA, which is similar to a binding site for Ndt80. Indeed, this sequence is a likely Ndt80-binding site since deletion of the consensus MSE at −230 does not eliminate Ndt80-dependent transcript accumulation. We do not have definitive evidence that these specific sequences promote CLB5 expression in meiosis but our deletion analysis suggests that they are likely candidates.
Deletion of five of the six MCB sequences in the CLB5 promoter leads to a very modest reduction in the accumulation of clb5 reporter gene mRNA in wild-type cells. This indicates that other sequence elements within UASCLB5 can promote gene expression. In contrast, when the only source of CLB5 is driven by a promoter that lacks MCB sequences, the expression of CLB5 mRNA and protein are profoundly reduced. An economical interpretation of these observations is that Clb5 can promote the activation of its own promoter in a positive-feedback loop. If CLB5 is activated by positive feedback in meiosis, this must be independent of the MCB elements and would presumably be important only for promoting CLB5 expression at early times since Ndt80 would activate the gene as cells approach the middle phase of sporulation. An additional complexity to this interpretation is that MBF is not required for CLB5 expression so its activation through MCBs and potentially other uncharacterized elements must make use of a regulator that we have not yet identified.
The requirement for MCB sequences in the CLB5 promoter likely indicates that an MCB-binding factor independent of Mbp1 is responsible for CLB5 induction in meiosis. This idea is supported by the observation that deletion of MBP1 leads to constitutive expression of TMP1 and POL1 while deletion of the MCBs in their promoters essentially eliminates their expression (Gordon and Campbell 1991; McIntosh et al. 1991). An MCB-binding activity distinct from Mbp1 has previously been reported in proliferating cells (Verma et al. 1991). Our preliminary gel-shift experiments also suggest that an MCB-binding activity can be detected in mbp1/mbp1 diploids undergoing meiosis. It is unclear whether the activity that we have detected is meiosis specific but it suggests that an alternative MCB-binding factor may exist. S. cerevisiae encodes several candidate transcription factors (SOK1, PHD1, and GAT1) that possess sequence similarity to the DNA-binding domain of Mbp1 and thus may bind to MCBs. Whether any of these function to replace Mbp1 at some or all MCB-regulated promoters during meiosis is as yet a matter of conjecture.
Despite the importance of the MCBs in its promoter, CLB5 is regulated differently from other genes that have been shown to be under the control of MCB sequences. A possible explanation for the difference in regulation between CLB5 and other MCB-regulated genes is that perhaps not all MCB sequences are equal. This has clearly been shown to be the case in the TMP1 promoter where mutational inactivation of one of the two consensus MCBs has a profound effect on periodic transcription whereas mutation of the other MCB has a very minor effect (McIntosh et al. 1991). It may be that sequences flanking the core MCB element confer the ability of accessory factors to bind and influence whether Mbp1 or an alternative factor binds to that MCB. This might explain why deletion of Mbp1 has a significant effect on the expression of RNR1 and TMP1 but not on CLB5.
CLB5 presents an example of how a cell-cycle-controlled gene has its regulatory elements reprogrammed to function during meiotic development. Our analysis of CLB5 expression during meiosis has revealed complex regulation of the CLB5 promoter in contrast to the relatively simple models previously put forth. Further analysis of the UASCLB5 region will likely reveal new transcriptional activating factors, some of which may be unique to meiosis. In addition, our data point to the existence of an MCB-binding factor that regulates a subset of MCB genes independently of Mbp1. Thus MCB sequences provide a common link between cell-cycle-regulated genes in mitotic proliferation and the expression of these genes during meiotic development.
We thank Catherine Hui for help in construction of the Δmse constructs used in this study. We also thank James DeCesare, Matt Rawluk, and Chantelle Sedgewick for helpful discussions. This research was support by an operating grant (MOP 62700) to D.S. from the Canadian Institutes for Health Research and by scholar awards to D.S. from the Canadian Institutes for Health Research and the Alberta Heritage Foundation for Medical Research.
Communicating editor: A. P. Mitchell
- Received September 8, 2004.
- Accepted December 8, 2004.
- Genetics Society of America