The Ime2 Protein Kinase Enhances the Disassociation of the Sum1 Repressor from Middle Meiotic Promoters

Abstract

Meiotic development in Saccharomyces cerevisiae (sporulation) is controlled by the sequential transcription of temporally distinct sets of meiosis-specific genes. The induction of middle genes controls exit from meiotic prophase, the completion of the nuclear divisions, and spore formation. Middle promoters are controlled through DNA elements termed middle sporulation elements (MSEs) that are bound by the Sum1 repressor during vegetative growth and by the Ndt80 activator during meiosis. It has been proposed that the induction of middle promoters is controlled by competition between Ndt80 and Sum1 for MSE occupancy. Here, we show that the Sum1 repressor can be removed from middle promoters in meiotic cells independent of Ndt80 expression. This process requires the phosphorylation of Sum1 by the meiosis-specific cyclin-dependent kinase-like kinase Ime2. The deletion of HST1, which encodes a Sir2 paralog that interacts with Sum1, bypasses the requirement for this phosphorylation. These findings suggest that in the presence of Ndt80, Sum1 may be displaced from MSEs through a competition-based mechanism but that in the absence of Ndt80, Sum1 is removed from chromatin in a separate pathway requiring the phosphorylation of Sum1 by Ime2 and the inhibition of Hst1.

Meiotic development is the pathway that produces haploid gametes from diploid precursors. Following the induction of the pathway, cells duplicate the genome and then enter an elongated prophase when homolog pairing, synapsis, and genetic recombination take place. Exit from prophase is followed by two sequential rounds of chromosome segregation. In many organisms, exit from prophase is a key control point where meiotic progression is regulated.

Meiotic development in the yeast Saccharomyces cerevisiae (sporulation) is tightly regulated by a transcriptional program. The transcriptional program of sporulation is characterized by the sequential expression of temporally distinct sets of genes that are induced as different steps in the program take place (5, 30). Although there are at least 12 temporally distinct subclasses of sporulation-specific genes, they can be broadly grouped into early, middle, and late classes (reviewed in reference 39).

Early promoters are activated by the Ime1 transcription factor, which is expressed and activated in diploids in response to starvation signals (reviewed in references 13 and 39). Early genes are expressed throughout meiotic S phase and prophase (5, 30). Toward the end of prophase, at the pachytene stage, homologs are fully connected by synaptonemal complexes (SCs), joint molecules have been formed but crossover recombinants have not been resolved (1, 24), and the spindle pole body (centrosome) has duplicated in preparation for the first meiotic division (3, 43).

Exit from pachytene, entry into the meiotic divisions, and spore formation require the induction of middle genes. Middle promoters are activated by the Ndt80 transcription factor, which specifically binds to DNA elements termed middle sporulation elements (MSEs) (6, 8). ndt80Δ mutants block meiotic development in pachytene (43). ndt80Δ-blocked cells can be held in pachytene for extended periods of time and efficiently resume mitotic growth if nutrients are replenished (43). In contrast, cells that have induced middle promoters and exited pachytene are largely unable to resume mitotic growth (33). These observations suggest that Ndt80 plays a role in the phenomenon of commitment to meiosis. Ndt80 also appears to be regulated by the recombination checkpoint (also known as the pachytene checkpoint) (6, 8, 26, 38) that blocks meiotic development in response to recombination intermediates such as double-strand breaks and defects in SC formation (see references 9 and 31 for reviews of the pachytene checkpoint).

Ndt80 promotes the disassembly of SCs and crossover formation by activating the expression of the Cdc5 polo-like kinase (36). Ndt80 also promotes the nuclear divisions and spore morphogenesis by activating the expression of M-phase cyclins, cell cycle regulatory molecules, and gene products required for spore formation (5). NDT80 is itself a tightly regulated meiosis-specific gene that is expressed shortly before most other middle genes, and it has been classified as an early/middle (25) or subclass 3a (30) gene. Its expression is controlled through Ndt80-MSE interactions in its own promoter in a positive autoregulatory loop (6). Despite the critical role that NDT80 induction plays in meiotic regulation, the molecular events that trigger the NDT80 autoregulatory loop have not yet been elucidated.

Sum1 is a DNA-binding protein that represses a subset of middle promoters in vegetative cells (42). Sum1 interacts with the core MSE and with adjacent bases (28). Ndt80-inducible MSEs therefore vary in their affinities for Sum1. The Sum1 and Ndt80 DNA-binding domains have been shown previously to compete for occupancy of MSE DNA in vitro. About 75 of the 150 middle promoters are Sum1 repressible, and these promoters are enriched with MSEs that are predicted to bind Sum1 (28, 40). The NDT80 promoter is controlled by a Sum1-repressible MSE, suggesting that the removal of Sum1 repression can regulate meiotic progression (25, 42). Sum1 represses transcription through the Hst1 NAD⁺-dependent histone deacetylase (Sir2 paralog) that is bound to Sum1 via the Rfm1 bridging protein (20). However, only a subset of Sum1-repressible promoters are derepressed in hst1Δ and rfm1Δ mutants, suggesting that Sum1 can repress transcription by an Hst1-Rfm1-independent mechanism. Consistent with the hypothesis that Sum1 regulates prophase exit, sum1Δ mutants bypass the pachytene checkpoint and carry out the nuclear divisions in the presence of recombination intermediates (broken chromosomes) to produce nonviable haploid products (15, 26).

In addition to fulfilling a role in middle meiotic gene regulation, Sum1 interacts with a subset of origins of DNA replication and positively regulates S phase by an Hst1-dependent mechanism (12, 41). Sum1 is also required to repress α-specific promoters in a cells (44). Sum1 and Hst1 also regulate the expression of promoters controlling the salvage pathway for NAD⁺ (2). However, NAD⁺ salvage pathway genes and α-specific genes are not induced during meiosis (5, 30). It is unclear how Sum1's roles in S phase regulation, cell type specification, NAD⁺ biosynthesis, and meiosis are differentially regulated.

Ime2 is a meiosis-specific cyclin-dependent kinase (CDK)-like kinase that regulates multiple steps in meiotic development (11). Ime2 has been shown previously to activate Ndt80 (34, 35). In addition, ime2Δ mutants show defects in middle meiotic gene expression that are partially suppressed by sum1Δ (25). These observations suggest that Ime2 promotes the transcription of middle promoters by positively regulating Ndt80 and by negatively regulating Sum1.

Two models can explain how Sum1 repression is removed in meiotic cells. The first model posits that Sum1 is removed from chromatin by a mechanism that requires competition with Ndt80. The findings that the purified DNA-binding domain of Sum1 can be displaced from an MSE with purified Ndt80 in vitro (28) and that high-level ectopic Ndt80 expression in vegetative haploid cells can induce SMK1 and other Sum1-repressible middle genes in mitotic haploid cells (5) demonstrate that competition can occur. According to this competition model, Sum1 would function as a transcriptional damper to prevent adventitious induction of middle promoters by Ndt80 or other activators. The second model posits that Sum1 is actively removed from chromatin in response to meiotic signals in S-phase or prophase cells. According to this sequential model, the regulated removal of Sum1 repression would create a state that would permit the NDT80 autoregulatory loop to be induced. Distinguishing between the competition and sequential models for the removal of Sum1 repression has important implications for how meiotic progression is regulated.

The SMK1 middle meiotic promoter provides an excellent model for the study of Sum1 regulation since it is controlled by a single, well-characterized MSE that interacts with Sum1 and Ndt80 and by an upstream activation sequence that binds the constitutive Abf1 transcriptional activator (28, 29, 42). Thus, Smk1 is not expressed when Sum1 is present, and it is expressed to a moderate level in an Abf1-dependent fashion when Sum1 has been removed. The derepressed SMK1 promoter is further activated by Ndt80 during meiotic development to yield peak levels of SMK1 expression (15).

In this study, we show that Sum1 is removed from middle meiotic promoters in a pathway that does not require competition with Ndt80. We also show that the Ime2 CDK-like kinase phosphorylates Sum1 on T306 in vivo and that a form of Sum1 containing a nonphosphorylatable substitution at this position (Sum1-T306A) is not removed in ndt80Δ meiotic cells. Although the deletion of Hst1 does not derepress SMK1 in vegetative cells, it does bypass the Sum1 removal defect of a sum1(T306A) ndt80Δ strain. These findings suggest that Ime2 promotes a state that is permissive for middle-promoter expression by antagonizing Hst1. These data indicate that a regulated pathway promotes the removal of the Sum1 brake from chromatin prior to NDT80 induction and suggest that this pathway controls the expression of middle promoters and meiotic progression. Our findings also suggest that when this pathway is compromised, Sum1 can be removed in an Ndt80-dependent pathway.

MATERIALS AND METHODS

Yeast strains and plasmids.

The rapidly sporulating SK1 genetic background was used for all the experiments described in this study (Table 1). The amino acid sequences of the SK1 and S288C forms of Sum1 differ at 14 positions (16). Therefore, only the SK1 form of Sum1 was used in these studies. Point mutations in SUM1 were generated using plasmid pMES39, which contained the SK1 form of SUM1 (chromosome IV coordinates 1084440 to 1081110) in which the initiator ATG had been changed to CGG in the XhoI and NotI sites of the URA3-based plasmid pRS306. The sum1(T306A) mutant was generated in pMES39 by changing codon 306 from ACT to GCT by using the QuikChange site-directed mutagenesis system to yield pMES42, which was digested with HindIII prior to transformation. Transformants were counterselected using 5-fluoroorotic acid, and the resulting sum1(T306A) mutant was analyzed by PCR and DNA sequencing. Eight histidine codons and the hemagglutinin (HA) epitope coding sequence were fused to the end of the SUM1 coding region corresponding to the Sum1 carboxy terminus to generate a His₈-HA (HH)-tagged Sum1 construct by using an HH-URA3 cassette as described by Chen et al. (4). SUM1-MYC-kanMX4 was generated as described by Longtine et al. (18).

TABLE 1.

Yeast strains

Strain	Genotype	Source or reference
ALY60	MATa/MATα SMK1-HA::kan/SMK1-HA::kan ura3/ura3 leu2::hisG/leu2::hisG trp1::hisG/trp1::hisG lys2/lys2 his4/his4 ho::LYS2/ho::LYS2	21
NAY176	MATa/MATα NDT80-HA::URA3/NDT80-HA::URA3 SUM1-MYC::TRP1/SUM1-MYC::TRP1 SMK1-HA::kan/SMK1-HA::kan ura3/ura3 leu2::hisG/leu2::hisG trp1::hisG/trp1::hisG lys2/lys2 his4/his4 ho::LYS2/ho::LYS2	This study
NAY28	MATa/MATα ndt80::LEU2/ndt80::LEU2 SMK1-HA::kan/SMK1-HA::kan ura3/ura3 leu2::hisG/leu2::hisG trp1::hisG/trp1::hisG lys2/lys2 ho::LYS2/ho::LYS2	This study
NAY185	MATa/MATα ndt80::LEU2/ndt80::LEU2 SMK1-HA::kan/SMK1-HA::kan SUM1-MYC::TRP1/SUM1-MYC::TRP1 ura3/ura3 leu2::hisG/leu2::hisG trp1::hisG/trp1::hisG lys2/lys2 ho::LYS2/ho::LYS2	This study
NAY228	MATa/MATα sum1(T306A)/sum1(T306A) SMK1-HA::kan/SMK1-HA::kan ndt80::LEU2/ndt80::LEU2 ura3/ura3 leu2::hisG/leu2::hisG trp1::hisG/trp1::hisG lys2/lys2 ho::LYS2/ho::LYS2	This study
NAY226	MATa/MATα sum1(T306A)/sum1(T306A) SMK1-HA::kan/SMK1-HA::kan ura3/ura3 leu2::hisG/leu2::hisG trp1::hisG/trp1::hisG lys2/lys2 ho::LYS2/ho::LYS2	This study
NAY278	MATa/MATα SMK1-HA::kan/SMK1-HA::kan hst1::kan/hst1::kan ndt80::LEU2/ndt80::LEU2 ura3/ura3 leu2::hisG/leu2::hisG trp1::hisG/trp1::hisG lys2/lys2 ho::LYS2/ho::LYS2	This study
NAY282	MATa/MATα sum1(T306A)/sum1(T306A) SMK1-HA::kan/SMK1-HA::kan hst1::kan/hst1::kan ndt80::LEU2/ndt80::LEU2 ura3/ura3 leu2::hisG/leu2::hisG trp1::hisG/trp1::hisG lys2/lys2 ho::LYS2/ho::LYS2	This study

Growth and sporulation of cells.

Cells were propagated on YEPD (1% yeast extract, 2% peptone, 2% glucose), SD (0.67% yeast nitrogen base without amino acids, 2% glucose, plus nutrients essential for auxotrophic strains), or YEPA (1% yeast extract, 2% peptone, 2% potassium acetate). For sporulation, cells were grown in liquid YEPA to a density of 10⁷ cells/ml. Cells were harvested by centrifugation, washed once in SM (2% potassium acetate plus 10 μg of adenine, 5 μg of histidine, 30 μg of leucine, 7.5 μg of lysine, 10 μg of tryptophan, and 5 μg of uracil per ml), resuspended in SM at 4 × 10⁷ cells/ml, and incubated at 30°C.

Immunological analysis of proteins.

Cellular extracts were prepared using the NaOH lysis-trichloroacetic acid precipitation procedure to limit proteolytic degradation during sample preparation as described previously (32). Samples were electrophoretically resolved through an 8% polyacrylamide gel, transferred onto Immobilon-P membranes (Millipore), and blocked overnight at 4°C in a mixture of phosphate-buffered saline, 0.1% Tween 20, and I-Block (Tropix). Antibody staining was performed overnight at 4°C with goat polyclonal anti-Sum1 (1:100; Santa Cruz) or monoclonal anti-HA (Covance; 1:2,500) and visualized using horseradish peroxidase-conjugated anti-goat or alkaline phosphatase-conjugated anti-mouse antibodies, respectively.

T306 phospho-specific analyses.

A peptide containing residues 300 to 310 of Sum1 followed by a cysteine residue (GKERPSTANSSC), with phosphothreonine at position 306 (underlined), was coupled to keyhole limpet hemocyanin, and the complex was injected subcutaneously into two rabbits. Booster injections were performed at 2, 6, and 8 weeks after the primary immunization. Animals were sacrificed at 12 weeks, and antibody was affinity purified from pooled sera by using the phosphorylated peptide conjugated to Affigel and absorbed against the nonphosphorylated peptide conjugated to Affigel (OpenBiosystems). We were unable to detect a specific signal by using total cellular extracts from meiotic cells with the affinity-purified phospho-T306 antisera. We therefore analyzed the phospho-T306 immunoreactivities of Sum1-HH or Sum1-T306A-HH proteins that had been purified from 10⁹ cells by using nickel beads under denaturing conditions as described by Chen et al. (4). In these experiments, analytical aliquots (2%) of the purified material were first assayed by immunoblot analyses using an HA antibody to estimate the relative concentrations of Sum1-HH. Subsequently, equivalent amounts of Sum1-HH were analyzed using a 1:500 dilution of the phospho-T306 antisera in ultrapure bovine serum albumin in Tris-buffered saline-Tween 20. Antibody staining was detected with a horseradish peroxidase-conjugated secondary anti-rabbit antibody.

ChIP.

Chromatin immunoprecipitation (ChIP) assays were performed as described by Strahl-Bolsing et al. (37) with the following modifications. Cells (2.0 × 10⁷) were cross-linked with 1% formaldehyde for 30 min at room temperature, after which glycine (final concentration, 125 mM) was added to quench the reaction. Cells were harvested, washed with water, and resuspended in 1 ml of FA-lysis buffer containing 140 mM NaCl (FA is 0.1% Triton X-100, 50 mM HEPES [pH 7.5], 1 mM EDTA, 0.1% Na-deoxycholate, 1 mM phenylmethylsulfonyl fluoride, 1 μg/ml pepstatin A, 1 μg/ml leupeptin, and 10 μg/ml aprotinin). Breakage of cells was performed by three cycles of beating with acid-washed glass beads for 1.25 min at 4°C and cooling on ice for 3 min. The bottom of the tube was punctured, and the lysate was collected by brief centrifugation. Fixed chromatin was sonicated using a Bioruptor cell disrupter (Cosmo Bio) for 10 min at 4°C (30 s on, 1 min off) to yield a fragment size of less than 500 bp. Immunoprecipitations were performed using 1 to 2 mg of sonicated chromatin solution and either 5 μl of anti-Sum1 antisera, 5 μg of anti-HA (Covance), or 2 μg of anti-Myc overnight at 4°C with rotation. The anti-Sum1 antisera used in the ChIP experiments was generated in rabbits against a fusion of maltose-binding protein to residues 770 to 1062 of Sum1 that was expressed and purified as described by Xie et al. (42). Purification of the immunoprecipitated protein-DNA complexes was achieved by adding 40 μl of a 1:1 slurry of protein A/G-Sepharose beads (CL-4B Amersham), 500 μg/ml bovine serum albumin, and 200 μg/ml sheared salmon sperm DNA for 1 h. Following binding, beads were washed twice with 1 ml of FA-lysis buffer-140 mM NaCl, once with FA-lysis buffer-500 mM NaCl, twice with a mixture of 0.25 M LiCl, 1% NP-40, 1% deoxycholate, 10 mM Tris HCl, and 1 mM EDTA (pH 8.0), and once with a mixture of 20 mM Tris-HCl and 1 mM EDTA (pH 8.0) for 5 min each time. The elution of bound protein-DNA was achieved by a 30-min incubation with 260 μl of 1% sodium dodecyl sulfate-0.1 M NaHCO₃ prewarmed to 65°C. Cross-links to DNA were reversed by overnight incubation at 65°C followed by ethanol precipitation. All samples were treated with RNase A and proteinase K (Roche) prior to purification with QIAquick PCR purification columns (Qiagen). DNA was quantified using real-time PCR with the Rotor-Gene 3000 system (Corbett Robotics). PCR was performed using Sensimix (Quantance Sybr green). The primer sequences were chosen to surround Sum1-binding sites present in middle meiotic promoters that were identified using the data of Chu et al. (5) and Primig et al. (30) and the consensus elements described by Pierce et al. (28) and Wang et al. (40). The sequences of the forward/reverse primers, respectively (shown 5′ to 3′), for each of the promoters analyzed are as follows: SMK1, GTGATTCGAAAAGTATCGCGC/GCGCCGAATTCTACCCTCA; NDT80, CTGACAAAGCTCCAGAACGGT/AGGGACCTTGGCTTTTCGAA; YAL018C, GCACGTGACTCAATAATACGACG/TCCATGTGTCACAAAACCAGAAA; SPR3, GTAAACCAATCAATGGCCCG/CATGGGAAGGATGCAGCAA; SPO74, TCCCGGTTATGTTGTATTGGC/CCATCACTCAACCAATTCAGCTC; SPO75, TGAAATTGTCAAGCCTGTCTCG/GAGATGAAGGGTTGTCGTTGGT; YOR365C, ATGCTGCTATGGAGCGATCC/GACCGCCATGAACTAAAGCTGT; HXT14, ATTATACTTCCCTTGACAAATCGATACG/TTTCCTCTTCTTTTGAGGGCC; and STE3, CAGATTTATGAACTCTGGGTATGGG/ACCAGCGCTTCTAACGAAAATTA. The sequences of the negative control primers for the open reading frame of DIT1 are as follows: TGCCAGCGTTTCCATGTAAG/CGGAACCGTACACTTTTTGCTC.

Quantitation of DNA was performed on the basis of a standard curve generated from serially diluted sonicated yeast genomic DNA. The percent input DNA recovered in the immunoprecipitation was calculated from averages for three replicates of immunoprecipitated DNA, negative control immunoprecipitates (no antibody), and input DNA as follows: (immunoprecipitated DNA signals − negative control signals)/input signals. Increases were calculated by dividing the percent input DNA recovered from the promoter of interest by the percent input recovered from the DIT1 intragenic (negative control) region. Each data point shows the average of results from three independent experiments.

Miscellaneous assays.

To monitor meiotic divisions, sporulating cells were taken at various times, fixed in 90% ethanol, and stained with 1 μg/ml of 4′,6-diamidino-2-phenylindole (DAPI). The completion of meiosis was scored by counting cells containing more than one DAPI-staining body by fluorescence microscopy. One hundred cells of at least three independent isolates per strain were counted. The preparation of total RNA and Northern blot hybridization analyses were carried out as described previously (14) using the DNA probes described by Lindgren et al. (15).

RESULTS

NDT80-independent derepression of SMK1.

Smk1 is not produced in mitotic cells when Sum1 is present, but it is produced when Sum1 is absent (15, 42). We compared Smk1-HA levels in wild-type and ndt80Δ meiotic cells to determine whether competition with Ndt80 is required for the removal of Sum1 repression. As shown in Fig. 1A, Smk1-HA starts to be produced in wild-type cells shortly before 5 h postinduction, with peak levels of Smk1 observed around 8 to 9 h postinduction. Smk1-HA in ndt80Δ diploids was observed at the same time as that in wild-type cells. However, the level of Smk1-HA observed was lower, and it did not increase to the level seen in wild-type cells. These findings are consistent with Sum1 repression being lifted in an NDT80-independent pathway prior to or at the pachytene exit control point. The further increase in Smk1 levels in wild-type cells is consistent with NDT80-dependent transcriptional activation, which occurs as pachytene exit, the nuclear divisions, and spore formation are taking place.

FIG. 1.

Sum1 repression of the SMK1 promoter is removed in an NDT80-independent fashion. (A) SMK1-HA or SMK1-HA ndt80Δ cells were transferred into sporulation medium, and samples were withdrawn at the indicated times. Total cellular extracts were analyzed with an HA antibody to monitor Smk1 expression (Smk1−) and with a PSTAIRE antibody that detects a closely spaced doublet consisting of Cdc28 and Pho85 as a loading control (Con−). The exposure time for the ndt80Δ samples was approximately four times longer than that for the wild-type samples. (B) NDT80-HA SUM1-MYC vegetative (V) cells and meiotic (M) cells (6.5 h postinduction) were analyzed by ChIP with HA or Myc antibodies. Values shown are increases in a PCR product containing the MSE in the SMK1 promoter relative to the DIT1 open reading frame PCR product (nonspecific control). Each bar represents the average of results from three separate ChIP experiments, each analyzed in triplicate. (C) ChIP assays were carried out with Sum1 antisera, and the results were analyzed as described in the legend to panel B.

Sum1 is removed from the SMK1 promoter independently of Ndt80.

The SMK1 promoter is controlled by a single MSE (located 70 to 79 bp upstream of the initiator ATG) that is necessary and sufficient for Sum1-dependent repression of SMK1 (29). To measure changes in Sum1 and Ndt80 occupancy at the SMK1 MSE, ChIP assays were carried out using a SUM1-MYC NDT80-HA strain. The results of these experiments demonstrate that the SMK1 MSE is bound by Sum1-Myc in mitotic cells and by Ndt80-HA in meiotic cells (Fig. 1B). Therefore, a switch from a Sum1- to an Ndt80-occupied state occurs at the SMK1 MSE during meiotic development in vivo.

To test whether the removal of Sum1 protein requires Ndt80, patterns of Sum1 occupancy at the SMK1 MSE in wild-type and ndt80Δ cells were compared. The data from these experiments show that Sum1 was removed from the SMK1 MSE in the presence or absence of Ndt80 (Fig. 1C). While the results presented were obtained using an antibody against the DNA-binding domain of Sum1, similar results were obtained when Sum1-Myc was analyzed using a monoclonal Myc antibody (data not shown). These findings demonstrate that Sum1 is removed from chromatin in an Ndt80-independent pathway in meiotic cells.

Sum1 is present during middle meiosis in wild-type and ndt80Δ strains.

We previously reported that the level of Sum1-HA expressed from a multicopy vector decreases around the time that cells complete the meiotic divisions and that the level of Sum1 is restored following the completion of meiosis (15). One explanation for the reduction in Sum1 at the SMK1 promoter is that Sum1 protein is not present in meiotic cells. We measured endogenous levels of untagged Sum1 protein using a polyclonal Sum1 antiserum. Although the level of endogenous Sum1 decreases around the time that the meiotic divisions occur, the decrease is less substantial than that seen with the overproduced epitope-tagged form of Sum1 (15). Thus, significant levels of Sum1 appear to be present in both wild-type and ndt80Δ samples as meiosis is taking place (Fig. 2).

FIG. 2.

Sum1 is present during middle meiosis. (A) Cells of the indicated genotypes were transferred into sporulation medium, and samples were withdrawn at the indicated times. Total cellular extracts were analyzed by immunoblotting with the indicated antibodies as described in the legend to Fig. 1. The arrows indicate the points where 50% of the cells had completed the nuclear divisions. Con, control.

Sum1 is phosphorylated by Ime2.

It has been shown previously that Ime2 phosphorylates Sum1 on residue T306 in vitro (22). We purified HH-tagged Sum1 and Sum1-T306A proteins from cells at various times following transfer to sporulation medium and analyzed the proteins with a phospho-specific antibody generated against T306 of Sum1. The results of these assays show that Sum1 is phosphorylated on T306 in vivo specifically in meiotic cells in an Ime2-dependent fashion (Fig. 3A). The kinetics of T306 phosphorylation indicate that this modification starts around the time that cells enter meiotic S phase and that the fraction of Sum1 that is phosphorylated on this residue increases during prophase. Interestingly, T306 phosphorylation appears to be reversed around the time that the meiotic divisions are completed (Fig. 3B). However, the addition of the HH tag to Sum1 slows down the timing of the nuclear divisions by 1.5 to 2 h, and it is conceivable that the kinetics of phosphorylation relative to the divisions in this background differ from that which occurs in cells without the HH tag. Regardless, these observations indicate that the phosphorylation of T306 by Ime2 can precede the nuclear divisions and are consistent with this modification's playing a role in regulating the removal of Sum1 repressor activity.

FIG. 3.

The meiosis-specific Ime2 kinase phosphorylates Sum1 at residue T306 in vivo. (A) Strains expressing HH-tagged forms of Sum1 were incubated in sporulation medium for the indicated times, and Sum1-HH was purified and analyzed by immunoblotting using phosphospecific T306 antisera (T306P) or an HA antibody to control for total Sum1-HH levels. (B) SUM1-HH diploids were transferred into sporulation medium, and samples were withdrawn at the indicated times and analyzed as described in the legend to panel A. The arrow indicates the point at which 50% of the cells had completed meiosis. The amount of purified Sum1-HH that was recovered from meiotic cells was smaller than that recovered from mitotic cells in all of these experiments. In the blots in both panels A and B, a larger fraction of the material purified from the meiotic cells than of that from the mitotic cells was loaded in order to equalize the amounts of Sum1 analyzed.

SMK1 expression is delayed in a sum1(T306A) mutant.

It has been shown previously that sum1(T306A) mutants complete meiosis and form spores (22). We compared Smk1-HA expression patterns and meiotic kinetics in wild-type cells and isogenic cells that differed only at codon 306 of Sum1. These assays revealed that the sum1(T306A) diploids induced Smk1-HA roughly 1.5 h later than the wild type (Fig. 4A). The production of Ndt80 (Fig. 4A) and the completion of meiosis (Fig. 4B) were only slightly delayed in the sum1(T306A) strain. Thus, the delay in Smk1-HA induction is not a consequence of the meiotic program's being delayed. One explanation for the delayed removal of the mutant Sum1-T306A protein during middle meiosis is that the T306A substitution alters the degradation of Sum1. However, the levels of Sum1-T306A and Sum1 were indistinguishable during meiotic development (Fig. 2, bottom). Therefore, delayed induction of Smk1-HA is not due to differences in the stability of the wild-type and T306A forms of Sum1.

FIG. 4.

Smk1-HA expression is delayed in a sum1(T306A) mutant in the presence of Ndt80. (A) Logarithmic cells of the indicated genotypes were transferred into sporulation medium, and samples were withdrawn at the indicated times. Total cellular extracts were analyzed by immunoblotting with the indicated antibodies as described in the legend to Fig. 1. Wild-type cells reached 50% meiosis at 6.4 h and sum1(T306A) cells reached 50% meiosis at 6.7 h in this experiment. Con, control. (B) Three independent isolates of coisogenic wild-type and sum1(T306A) mutants were sporulated, and the fraction of cells that had completed MI was quantified.

The NDT80-independent removal of Sum1 repression does not occur in a sum1(T306A) mutant.

To test whether the phosphorylation of T306 affects the removal of Sum1 repression in the NDT80-independent pathway, Smk1-HA expression patterns in ndt80Δ and ndt80Δ sum1(T306A) cells were compared. Smk1-HA was expressed in the ndt80Δ strain but not in the ndt80Δ sum1(T306A) strain (Fig. 5A). This observation suggests that the phosphorylation of Sum1 on T306 by Ime2 is required for the removal of Sum1 repression in the Ndt80-independent pathway. ChIP assays of Sum1 and Sum1-T306A from chromatin in ndt80Δ cells demonstrated that the T306A substitution prevents the removal of Sum1 from SMK1 chromatin (Fig. 5B).

FIG. 5.

Smk1 is not expressed and Sum1-T306A is not removed from the SMK1 promoter in ndt80Δ cells. (A) Cells of the indicated genotypes were transferred into sporulation medium, and samples were withdrawn at the indicated times. Total cellular extracts were analyzed by immunoblotting with the indicated antibodies as described in the legend to Fig. 1. (B) Cells of the indicated genotypes were analyzed by ChIP during vegetative growth (V) and during meiosis (M) (6.5 h postinduction). Values shown are increases (% input) of a PCR product containing the MSE in the SMK1 promoter relative to an intragenic DIT1 (negative control) PCR product. Each bar represents the average of results from three separate ChIP experiments, each analyzed in triplicate.

The deletion of HST1 restores the expression of Smk1-HA in an ndt80Δ sum1(T306A) strain.

Sum1 recruits the Hst1 histone deacetylase through the Rfm1 tethering factor. Hst1 and Rfm1 repress only a subset of Sum1-repressible middle promoters, and the deletion of HST1 or RFM1 has no effect on SMK1 in vegetative cells (20). The effect of hst1Δ on Smk1-HA expression was tested in ndt80Δ and ndt80Δ sum1(T306A) mutants following meiotic induction. hst1Δ had no effect on the expression of Smk1-HA in the ndt80Δ background. However, hst1Δ bypassed the ndt80Δ sum1(T306A) defect in Smk1-HA expression (Fig. 6A). ChIP assays demonstrated that Sum1 was removed from the SMK1 promoter in the sum1(T306A) ndt80Δ hst1Δ strain (Fig. 6B). These findings suggest that the phosphorylation of Sum1 by Ime2 inhibits the Hst1 deacetylase activity at this promoter and that the downregulation of Hst1 promotes Sum1 removal.

FIG. 6.

Deletion of HST1 restores Smk1-HA expression and the removal of Sum1-T306A in ndt80Δ cells. (A) Cells of the indicated genotypes were transferred into sporulation medium, and samples were withdrawn at the indicated times. Total cellular extracts were analyzed by immunoblotting with the indicated antibodies as described in the legend to Fig. 1. Con, control. (B) Cells of the indicated genotypes were analyzed by ChIP during vegetative growth (V) and during meiosis (M) (6.5 h postinduction) as described in the legend to Fig. 5.

Middle-gene mRNAs are regulated by the Ndt80-independent removal of Sum1.

To further investigate the Ndt80-independent pathway for Sum1 removal, ndt80Δ, ndt80Δ sum1(T306A), hst1Δ ndt80Δ, and hst1Δ ndt80Δ sum1(T306A) cells were incubated in sporulation medium for various times, and total RNA was purified and analyzed by Northern blot hybridization (Fig. 7). IME2, which is induced as an early meiotic gene, was expressed similarly in all of the genetic backgrounds tested, indicating that the sum1(T306A) and hst1Δ mutations do not influence the induction of the program or the accumulation of early meiosis-specific transcripts. The accumulation of SMK1 mRNA in the ndt80Δ strain started between 5 and 6.5 h postinduction. This increase in SMK1 mRNA was not observed in the ndt80Δ sum1(T306A) strain, but it was observed in the hst1Δ ndt80Δ sum1(T306A) strain. These findings are consistent with the expression characteristics of the Smk1-HA protein (Fig. 1A, 5A, and 6A).

FIG. 7.

Multiple Sum1-repressible middle-gene mRNAs accumulate in an Ndt80-independent process that is regulated by Ime2 and Hst1. Strains of the indicated genotypes were collected at the indicated times after being transferred into sporulation medium, and total RNA was prepared and analyzed by Northern blot hybridization with the indicated probes. The rRNA sample from the ethidium bromide (EtBr)-stained electrophoretic gel prior to Northern transfer is shown as a loading control. The wild-type (WT) RNA samples were included in this analysis to demonstrate the contribution of Ndt80 to the accumulation of the indicated mRNAs.

YAL018C and SPR3 are MSE-regulated middle genes that have been shown previously to be repressed by Sum1 in vegetative cells (20, 42). YAL018C is derepressed in hst1Δ and rfm1Δ vegetative cells, while SPR3 is not (2, 20). Similar to SMK1 mRNA, YAL018C mRNA accumulates in an Ndt80-independent fashion when cells are transferred into sporulation medium. Also similar to SMK1 mRNA, YAL018C mRNA in the sum1(T306A) ndt80Δ strain is reduced compared to the sum1(T306A) ndt80Δ hst1Δ strain. Thus, the regulatory interactions between sum1(T306A) and hst1Δ are not specific to the SMK1 middle promoter. However, it is worth noting that the level of SMK1 mRNA in sum1(T306A) ndt80Δ cells remained low throughout the time course but that YAL018C mRNA increased modestly at the later time points tested. Thus, there may be differences in the extent to which YAL018C and SMK1 are responsive to the sum1 T306 substitution. It is also worth noting that although YAL018C is derepressed in hst1Δ vegetative cells (2, 20; also our unpublished observations), the extent of derepression is modest compared to the maximally induced level in meiosis (see Discussion).

The expression characteristics of the SPR3 middle gene in the different genetic backgrounds are also similar to those of SMK1. Thus, SPR3 mRNA accumulated in ndt80Δ cells upon transfer into sporulation medium, and the sum1(T306A) substitution caused a delay in this accumulation. Moreover, SPR3 expression in the hst1Δ ndt80Δ sum1(T306A) strain was increased compared to that in the ndt80Δ sum1(T306A) strain. However, SPR3 was expressed slightly sooner in the hst1Δ ndt80Δ strain than in the ndt80Δ strain, and this phenomenon was not seen with SMK1 or YAL018C. Previous studies showed that the SPR3 and SMK1 promoters are similar in their regulation by MSE and Abf1-binding sites but that additional DNA elements control SPR3 expression in response to nutritional signals (23). It is possible that the differences in SMK1 and SPR3 regulation are related to these additional elements (see Discussion).

Sum1 is removed from multiple middle promoters.

We next tested the removal of Sum1 from multiple different Sum1-regulated promoters in the ndt80Δ, ndt80Δ sum1(T306A), hst1, and hst1 sum1(T306A) strains by using ChIP assays. YAL018C, SPR3, SPO74, SPO75, YOR365C, and HXT14 are middle meiosis-specific genes (5, 30) that are repressed by Sum1 (29) and have Sum1-binding sites in their promoters (28, 40). The degrees of occupancy of all of these promoters by Sum1 decreased in ndt80Δ cells that had been transferred into sporulation medium. These decreases were attenuated in the ndt80Δ sum1(T306A) strain. Moreover, a pattern that resembled that seen in the ndt80Δ strain was observed in the ndt80Δ sum1(T306A) hst1Δ background. Thus, the Sum1 occupancy patterns for all the Sum1-repressible middle promoters tested were qualitatively similar to that for the SMK1 promoter. The ndt80Δ allele used in this study lacks the open reading frame but retains the full promoter. Similar patterns were also observed at this promoter.

Sum1 has recently been shown to interact with the α-specific STE3 promoter (44). Sum1 was removed from the STE3 promoter in ndt80Δ cells transferred into sporulation medium, despite the fact that STE3 is not transcriptionally induced in meiotic cells (5, 30). We also observed a pattern of Sum1 occupancy changes at the STE3 promoter similar to that observed at SMK1 in the ndt80Δ sum1(T306A) and ndt80Δ sum1(T306A) hst1Δ strains. These findings suggest that the Ndt80-indpendent removal of Sum1 is a widespread response to meiotic induction that is regulated by Ime2 and Hst1.

DISCUSSION

In S. cerevisiae, the transient induction of middle meiotic promoters controls exit from prophase, nuclear segregation, and spore formation. In this study, we have shown that the Sum1 repressor can be removed from multiple middle promoters during meiotic development in the absence of Ndt80 competition. We have also shown that Sum1 is phosphorylated on T306 in a meiosis-specific fashion by Ime2 in vivo and that this modification is required for the Ndt80-independent removal of Sum1 from DNA. hst1Δ bypasses the removal defect in the sum1(T306A) ndt80Δ strain. These findings demonstrate that an Ndt80-independent pathway for Sum1 removal operates in meiotic cells and that this pathway is regulated by Ime2 and Hst1. These findings are relevant to how middle meiotic gene expression is controlled in yeast and also to the larger question of how genes are transiently expressed in developmental programs.

Consequences of T306 phosphorylation.

How does the phosphorylation of Sum1 on T306 by Ime2 regulate the removal of Sum1 from chromatin? The findings that Sum1-T306A is not removed from promoters in ndt80Δ cells but is removed in ndt80Δ hst1Δ cells suggest that one role of T306 phosphorylation is to negatively regulate Hst1. SMK1 is derepressed in sum1Δ (28) but not in hst1Δ or rfm1Δ (2, 20) vegetative cells. In addition, Sum1 is present at SMK1 and other promoters in vegetative hst1Δ cells (Fig. 6B and 8). These observations demonstrate that, while the removal of Hst1 permits Sum1 removal from chromatin in meiotic cells, it is insufficient to cause removal during vegetative growth. Taken as a whole, our findings are most consistent with a model in which the phosphorylation of T306 by Ime2 controls one step in a multistep pathway that leads to Sum1 removal and ultimately middle-gene induction (Fig. 9). We propose that T306 phosphorylation by Ime2 downregulates the Hst1 deacetylase at Sum1-occupied promoters. One possible mechanism would involve the inhibition of Rfm1-Hst1 binding to Sum1 (Fig. 9). However, Hst1/Sum1 activity can be regulated at NAD⁺ biosynthetic genes by a mechanism that does not appear to involve regulated changes in Hst1-Rfm1 binding to Sum1 (2). Thus, further work is required to determine whether T306 phosphorylation directly causes the dissociation of Rfm1-Hst1 from Sum1. The Rfm1-Hst1 complex has been shown to preferentially deacetylate H4 on K5 to positively regulate the initiation of Sum1-bound origins of DNA replication (41). In contrast to Hst1, Sir2 negatively regulates origins of DNA replication (7, 27). We propose that the preferential increase in the acetylation of specific residues in histone or nonhistone chromatin components following the downregulation of the Hst1 deacetylase generates a chromatin state in meiotic cells that is permissive for Sum1 removal (Fig. 9). Subsequently, Sum1 removal would be promoted by additional mechanisms. Superimposed on this pathway is the ability of Ndt80 to directly compete with Sum1 for binding to MSE DNA. The removal of phosphate from T306, as illustrated in Fig. 3B, may be involved in reestablishing repression.

FIG. 8.

Sum1 is removed from multiple promoters in an NDT80-independent pathway that is regulated by Ime2 and Hst1. Sum1 ChIP samples were analyzed by PCR with primers specific for the indicated promoters as described in the legend to Fig. 5B. The same samples were used for the analyses presented in Fig. 5B and 6B. Each bar represents the average of results from three separate ChIP experiments, each analyzed in triplicate. V, vegetative cells (6.5 h postinduction).

FIG. 9.

Multistep model for the Ndt80-independent removal of the Sum1 repressor from chromatin in meiotic cells. Sum1 recruits the Hst1 deactylase via the Rfm1 bridging protein to middle-promoter chromatin in vegetative cells (state 1). The phosphorylation of Sum1 by the meiosis-specific Ime2 protein kinase is shown to negatively regulate the associated Rfm1-Hst1 complex to generate state 2. While this regulation may occur by directly promoting the removal of Rfm1-Hst1 as indicated, this possibility has not been tested. The influence of Hst1 implies that a histone acetyltransferase (HAT) acts after Hst1 downregulation and that the increase in acetylation of an unidentified chromatin component generates a state that is permissive for Sum1 removal (state 3). The deletion of HST1 or RFM1 is insufficient to derepress many middle genes in mitotic cells, and this finding implies that the removal of Sum1 requires an additional regulatory factor (indicated by X) that leads to Sum1 removal (state 4). The changes leading to state 4 regulate the NDT80 promoter. The removal of Sum1 repression at the NDT80 promoter increases the Ndt80 protein concentration in the cell, leading to Ndt80 occupancy and transcriptional induction (state 5). Ndt80 can also promote one or more of the preceding steps as indicated (for example, by direct competition) to enhance the feed-forward and switch-like attributes of the system.

Genome-wide microarray studies have shown that only about one-third of Sum1-repressible genes are also repressed by Hst1 (20). Notably, the extent of derepression in hst1Δ cells is substantially less than that in sum1Δ cells for all genes tested. YAL018C is among the genes most significantly repressed by Hst1, yet the derepressed level of this gene is only a small percentage of the fully induced level seen in meiotic cells (Fig. 7 and data not shown). These observations suggest that Hst1's major role in meiotic gene regulation may be to promote transient changes in chromatin structure that permit downstream inductive events (such as the removal of Sum1) rather than to directly regulate gene transcription.

Relevance to the middle meiotic transcriptional switch and prophase exit.

We previously reported that levels of an HA-tagged form of Sum1 expressed from a multicopy vector decrease during meiosis and that this decrease is not observed in cells arrested at the checkpoint (15). These observations raised the possibility that fluctuations in the level of Sum1 in cells could regulate pachytene exit. In the present study, we confirmed that the level of endogenous Sum1 decreases during meiosis. However, the decrease is not as substantial as that seen with overproduced/epitope-tagged Sum1 (15; also data not shown). It is likely that the overexpression of Sum1 increases the fraction of the protein that is not bound to MSEs. If unbound Sum1 is less stable than bound Sum1 in meiotic cells, this could explain the more rapid degradation of overexpressed Sum1. Importantly, Sum1-T306A levels decrease with kinetics similar to that of wild-type Sum1, yet Sum1-T306A remains associated with chromatin in ndt80Δ cells transferred into sporulation medium. These observations indicate that the decrease in Sum1 levels seen in meiotic cells is insufficient to cause Sum1 removal from chromatin.

This study has shown that Sum1 can be removed from MSE DNA through a meiosis-specific Ndt80-independent pathway at all middle promoters tested, as well as the STE3 promoter, which is not induced during meiosis. These observations indicate that Sum1 is removed at many and perhaps most sites in chromatin during meiotic development. The NDT80 promoter is controlled by at least two MSEs (one of which is Sum1 repressible and a second that is not), as well as two upstream repression sequence 1 (URS1) elements that promote early meiosis-specific gene expression (25). It has been proposed previously that the regulated transition at URS1 from a repression complex (occupied by Ume6/Rpd3/Sin3) to an activation complex (occupied by Ime1) (19) functions in a priming step that is required to initiate NDT80 transcription but that Sum1 functions as a brake that holds the promoter in an inactive state until the preconditions for pachytene exit have been met (25, 42). Our data suggest that the phosphorylation of Sum1 on T306 is involved in the establishment of NDT80 expression. However, T306 phosphorylation cannot be the sole trigger for NDT80 transcription, since sum1(T306A) cells complete meiotic development. Our data suggest that the SMK1, YAL018C, and SPR3 promoters differ in their responsiveness to the sum1(T306A) and hst1Δ mutations. These differences are likely due to additional DNA elements that act in collaboration with MSEs. One possibility is that URS1 or other elements in the NDT80 promoter can overcome Sum1 repression and produce sufficient Ndt80 in meiotic cells to remove Sum1 by direct competition when T306 is not phosphorylated by Ime2. A second possibility is that redundant mechanisms negatively regulate the Sum1 protein. Sum1 contains multiple CDK sites, and Ime2 and Cdc28 are known to coregulate meiotic substrates (10). In addition, Lo et al. have suggested that Cdc7 regulates Sum1 removal (17). These observations suggest that multiple signals converge on Sum1 to control prophase exit and the completion of meiotic development.