INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

An orderly progression through the cell division cycle requires that the relevant cellular events occur in a strict temporal sequence. In the budding yeast Saccharomyces cerevisiae, as in many other organisms, this is achieved by a combination of transcriptional and posttranslational regulations of various effectors. One of the best-studied examples of such posttranslational controls is the way in which the activity of the key protein kinase Cdc28 is regulated. Phosphorylation of specific tyrosine and threonine residues (19, 34, 35, 49), the association with a variety of cyclins (14, 43, 48, 53), and their proteolytic destruction at various stages of the cell cycle (2, 3, 10, 18) all contribute to sharply define not only the timing of Cdc28 activation-inactivation but also, presumably, its substrate specificity (35). Similarly, an anaphase-promoting complex whose activity is crucial for chromosome segregation appears to be regulated at the posttranslational level (24, 54, 55).

Although posttranslational modifications of the mitotic regulators are perhaps the most effective way for the cell to quickly respond to the changing cellular context, these responses may be further sharpened by controls at the transcriptional level. Indeed, it has been estimated by DNA microarray hybridization that about 800 genes of the budding yeast S. cerevisiae are transcribed only at specific stages of the cell cycle (50). Many of these periodically transcribed genes play crucial roles in cell cycle progression. For instance, the transcripts for the G₁ cyclins CLN1 and CLN2 appear in late G₁ phase approximately when the cells traverse START. The transcription of these genes is regulated by a sequence motif called SCB (Swi4/Swi6 cell cycle box) within their promoter (22). This motif is also found in the promoters of the genes encoding HO endonuclease and the cyclin-like protein Hcs26 (6, 11), which are also expressed in late G₁. A specific heteromeric transcription factor complex (SCF) containing Swi4 and Swi6 proteins drives the cell cycle stage-specific expression by binding to the SCB element (37, 38). The S-phase cyclin genes CLB5 and CLB6 and DNA synthesis genes such as POL1, TMP1, and CDC9 are also expressed in late G₁ but their expression is driven by another promoter element called MCB (Mlu1 cell cycle box) which binds a transcription factor complex (MBF) containing Mbp1 and Swi6 (23, 28). At the onset of the S phase, histones and, somewhat later, CLB3 and CLB4 cyclin genes are transcribed (15).

The transcription of another set of genes is activated near the completion of the S phase. This includes FAR1, a protein required for pheromone-induced G₁ arrest, the transcription factors SWI5 and ACE2, mitotic cyclin genes CLB1 and CLB2, and CDC5, a gene encoding a Polo-like protein kinase. It is not clear whether the periodic expression of these genes is regulated by a common promoter element and transcription factors. An analysis of SWI5 promoter revealed a 55-bp element that is sufficient for the periodic expression of SWI5 (29). It has been proposed that Mcm1, a DNA-binding protein important for the determination of mating-type specificity, forms a complex with Sff (SWI5 transcription factor) and binds to this 55-bp promoter element (29). The gene encoding Sff has not yet been identified. In subsequent studies, it was shown that Mcm1 is required for the expression of CLB1, CLB2, SWI5, CDC5, and ACE2 (1, 30). This is consistent with the fact that the promoter regions of CLB1, CLB2, and SWI5 contain consensus Mcm1 binding sites. Interestingly, Mcm1 occupies these sites throughout the cell cycle, implying that the timing of expression is perhaps determined either by specific modifications of the bound factors or by other associated factors (1). It has been recently reported that a novel promoter element ECB (early cell cycle box), which renders the expression of SWI4, CDC6, CDC47, and CDC46 M/G₁ specific, also binds Mcm1 (32). Since Mcm1 is also involved in the regulation of other cellular processes that are not directly linked to the cell cycle progression (26, 31, 33, 51), it seems unlikely that Mcm1 is the critical factor in determining the periodicity of gene expression.

As Cdc28 is the key protein kinase required at various stages for cell cycle progression, most CDC28 mutants exhibit defects in progression through both START and G₂/M transition (39, 41). However, cdc28-1N is a unique mutant in that, although normal with respect to its START functions, it fails to undergo nuclear division at the restrictive temperature and, consequently, arrests with a large elongated bud, a short mitotic spindle, 2N DNA content, high histone H1 kinase activity, and an undivided nucleus (39, 53). Although the molecular nature of its mitotic defect is still unknown, the cdc28-1N mutant has served well for the identification of a number of novel genes. CLB1, CLB2, and CLB4 were first isolated as allele-specific, high-dosage suppressors of the cdc28-1N mutation (53). Similarly, CIM3 and CIM5 genes encoding the 26S proteosome subunits were identified as mutations that exhibit synthetic lethal behavior in combination with cdc28-1N (17). We report here the isolation of a novel allele-specific, high-dosage suppressor of cdc28-1N, termed NDD1, and examine its role in cell cycle progression. We describe its basic characterization and show that NDD1 is an essential component, as is MCM1, of the mechanism that activates the expression of a set of late-S-phase-specific genes.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Yeast strains and growth media. Yeast strains (Table 1) used in this study were derivatives of the standard wild-type strain W303. Strain BJ2168 (US1005) was used in the mobility shift DNA-binding assay. Yeast cells were routinely grown in yeast extract-peptone medium containing adenine (50 mg/liter) supplemented with either glucose or galactose. Raffinose was also added to the galactose medium. All media were prepared as described by Rose et al. (44).

Isolation of NDD1. cdc28-1N cells were transformed with a library of yeast genomic fragments cloned into the BamHI site of the YEp13 (2µm) vector (36), allowed to recover at 24°C for 24 h, and then shifted to 37°C for 3 days. The plasmids were retrieved from various transformants and were categorized into seven distinct groups based on the patterns of fragments generated by restriction enzyme digestion. One group of plasmids harbored a 2.7-kb PvuII genomic fragment containing the NDD1 gene.

DNA manipulations. All DNA manipulations were performed as described by Sambrook et al. (46). The 2.7-kb PvuII fragment containing the entire NDD1 gene was blunt-end cloned into the SmaI-BglII site of pIC19H vector to yield pUS529. To construct the NDD1 gene under the control of GAL1 promoter, an NcoI site was created immediately upstream of the ATG start codon by PCR. The BamHI-BglII fragment from pUS529 was then replaced by the BamHI-BglII-digested 0.97-kb PCR product. The complete NDD1 gene was then excised with NcoI-HindIII and blunt-end ligated to the GAL1 promoter in a URA3- or TRP1-selectable CEN vector.

NDD1^Q

Random spore analysis. To determine the terminal phenotype of cells deficient in NDD1, a diploid strain heterozygous for ndd1::LEU2 (US241) was sporulated in liquid medium. Spores purified by centrifugation (7) were inoculated into medium lacking leucine. Samples were withdrawn 6 h after germination and were used for in situ immunofluorescence and photomicroscopy.

Gel mobility shift assay. Yeast protein extracts were prepared from protease-deficient (pep4) strains essentially as described by Company et al. (12). Double-stranded DNA probes were prepared by using the Klenow enzyme to fill in the protruding 5' end of a restriction site in the presence of radioactive [-³²P]dATP. The labelled probes were purified from the unincorporated nucleotides by using NucTrap probe purification columns (Stratagene). DNA binding reactions were carried out in 1.5-ml Eppendorf tubes. Generally, 10,000 to 20,000 cpm (Cerenkov) of probe was used for each reaction. Each reaction contained 4 µl of 5× BS buffer (100 mM Tris-HCl [pH 7.5], 250 mM NaCl, 15 mM MgCl₂, 5 mM dithiothreitol, 25 mM spermidine, 250 µg of bovine serum albumin per ml, 100 mM EDTA), 2 µl of 50% glycerol, 2.5 µl of a 1-mg/ml concentration of poly(dI-dC), radiolabelled probe, water, and 20 µg of protein extract to give a final volume of 20 µl. Protein extracts were always added last, and the reactions were incubated at 24°C for 15 min followed by a further 15 min of incubation at 4°C. Conditions for competition experiments were exactly the same except that cold competitor DNAs were added to the reaction mixture before the addition of radiolabelled probe. For antibody supershift assays, the appropriate antibodies were added 6 min after the addition of the protein extracts and processed as described above. The gel used for resolving the protein-DNA complexes was 4% (20:1) acrylamide-bisacrylamide gel in 0.5× Tris-borate-EDTA (TBE). The gels were prerun at 4°C for 2 h at 200 V (~20 mA) in 0.5× TBE. After incubation, the reaction mixtures were loaded onto the gel and run at 4°C for 3 h at 200 V (~8 mA). The gels were dried and autoradiographed with an intensifying screen at 70°C.

Other techniques. The lithium acetate method was used for all yeast transformations. Total RNA was isolated as described by Cross and Tinkelenberg (13), and Northern (RNA) blot analyses were performed as described by Price et al. (40). The method of Kilmartin and Adams (21) was used for immunofluorescence and photomicroscopy. DNA distribution analysis by flow cytometry was performed as described by Lim et al. (27).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

NDD1 is an allele-specific suppressor of the cdc28-1N mutation. Most CDC28 mutants (i.e., cdc28-1, cdc28-4, and cdc28-13) are predominantly defective in traversing START and therefore arrest in late G₁ phase as unbudded cells with 1N DNA content (20, 42). Some of them also exhibit a defect in progression through mitosis (39, 41). The cdc28-1N mutant is unique in that it progresses through START and S phase normally but fails to undergo nuclear division at the nonpermissive temperature, despite the high Clb2-associated histone H1 kinase activity (39, 53). We isolated NDD1 in a genetic screen designed to identify genes whose overexpression can suppress the mitotic defect of cdc28-1N mutant. To determine whether NDD1 overexpression causes suppression of other cdc28 mutations, a high-copy-number (2µm) vector carrying NDD1 was introduced into both cdc28-4 and cdc28-1N mutants, and the transformants were tested for growth at 37°C. While NDD1 allowed growth of cdc28-1N cells at 37°C, it failed to suppress the cdc28-4 mutation (Fig. 1A). Overexpression of NDD1 also failed to suppress the cdc28-4 mutation when tested at 31 and 35°C (data not included). Hence, NDD1 is an allele-specific suppressor of the cdc28-1N mutation. This behavior is identical to that of the mitotic cyclin genes CLB1 and CLB2, which were isolated in the same screening (53). This suggests that the role of NDD1 is related to the mitotic function of CDC28.

View larger version (131K): [in this window] [in a new window]   FIG. 1.   (A) NDD1 is an allele-specific suppressor of the cdc28-1N mutation. cdc28-4 and cdc28-1N cells were transformed with YEp13 (2µm) or the vector containing NDD1 as the 2.7-kb PvuII fragment. The transformants were plated on leucine-deficient plates and incubated at 24 or 37°C for 48 h. (B) Nucleotide sequence and the predicted amino acid sequence of NDD1. The polyglutamine stretch is shown in boldface, and the three well-conserved cdc2/CDC28 histone H1 phosphorylation sites are underlined.

NDD1 is an essential gene whose function is required during mitosis. To determine whether NDD1 is necessary for cell viability, the ORF in one of the two copies of NDD1 in a wild-type diploid was replaced by the LEU2 gene (Fig. 2A). The resulting heterozygous mutant (US241) was allowed to sporulate, and the tetrads were dissected on rich medium. Only two spores in each tetrad survived and gave rise to normal colonies that failed to grow when replica plated on leucine-deficient medium. By inference, the absence of viable Leu⁺ segregants in these dissections suggests that all segregants that were Leu⁺ and therefore lacked intact NDD1 ORF were nonviable. Microscopic examination of these segregants revealed that the spores had germinated but that most had arrested as large-budded cells, while others had undergone one or two divisions before cessation of growth. This suggests that NDD1 is essential for vegetative growth.

View larger version (38K): [in this window] [in a new window]   FIG. 2.   NDD1 is an essential gene. (A) Partial restriction map of the 2.7-kb PvuII fragment containing NDD1. The 0.45-kb EcoRV fragment in the ORF was replaced by the LEU2 gene to generate the disruption mutant ndd1::LEU2. The scale bar is in base pairs. Diploid cells heterozygous for ndd1::LEU2 disruption were sporulated and dissected on rich medium. Tetrads show a 2:2 segregation pattern for viability. All survivors were Leu. (B) Cell cycle arrest phenotype of ndd1. Spores from a wild-type diploid strain or an ndd1::LEU2 heterozygous diploid strain were purified and germinated in either complex medium or medium lacking leucine, respectively. Samples were withdrawn 6 h after germination and used for differential-interference-contrast (DIC) microscopy, staining with anti-tubulin antibodies, and DAPI (4',6-diamidino-2-phenylindole) staining. The lower panel shows DNA distribution profiles of asynchronously growing wild-type and Ndd1-depleted cells.

Expression and localization of Ndd1 during the cell cycle. Since NDD1 appeared to be necessary for the nuclear division process, we asked whether its expression is restricted to this stage of the cell cycle. Exponentially growing wild-type cells were synchronized in G₁ by -factor treatment for 3 h at 25°C and then allowed to resume cell cycle progression in -factor-free medium. Total RNA was isolated, and NDD1-specific transcript was detected by Northern blot analysis. The NDD1 RNA, first detectable above the basal level at 30 min after the release from G₁ arrest, peaks at 50 to 60 min and then declines ca. 30 min before the majority of cells reach anaphase (Fig. 3A). The transcript appears again in the next cycle when the proportion of cells with anaphase spindle has reached its minimum. The window of NDD1 expression lies between the peaks of CLN1 and CLB2 transcription (Fig. 3A), a period that corresponds to the S phase. It is not yet clear as to what restricts NDD1 transcription to the S phase. So far we have not found any known cis-regulatory elements within the 2-kb region upstream of the ORF except one loosely conserved SCB (Swi4/Swi6 cell cycle box) in reverse orientation.

View larger version (116K): [in this window] [in a new window]   FIG. 3.   (A) NDD1 gene expression during the cell cycle. Wild-type cells lacking BAR1 gene were arrested in G1 by -factor treatment and then released into growth medium without -factor. Samples were withdrawn at 10-min intervals, and total RNA was analyzed by Northern blotting with an internal EcoRI fragment of NDD1 as a probe. CLN1 and CLB2 RNA transcripts are also shown for comparison. The amount of total RNA loaded (30 µg) was visualized by methylene blue staining of the rRNAs. The graph at the bottom shows the percentage of cells with anaphase spindles after release. (B) Nuclear localization of the Ndd1-lacZp. Wild-type cells carrying pGAL1-NDD1-lacZ were grown in glucose or galactose medium and then used for both indirect immunofluorescence staining with anti--galactosidase antibodies (upper panels) and nuclear staining with DAPI (lower panels).

Overexpression of Ndd1 enhances the expression of CLB1, CLB2, and SWI5 genes. It is known that elevated levels of the B-type cyclins Clb1, Clb2, and Clb4 can suppress the nuclear division defect of the cdc28-1N mutant (53). We reasoned that any gene whose overexpression can cause the accumulation of these cyclins, either by upregulation of transcription or by increasing their stabilization, would also be identified by our genetic screen. Therefore, we tested NDD1 for its ability to enhance the transcription of CLB1 and CLB2. Wild-type cells carrying GAL-NDD1 on a CEN plasmid (US737) were grown in raffinose medium (in which GAL1 promoter is not active) for 3 h and then one-half of the culture was induced to express Ndd1 by the addition of galactose. Total RNA was prepared at various time points and analyzed for the presence of CLB1, CLB2, CLB3, SWI5, CDC20, and NDD1 transcripts. Of these, the CLB1, CLB2, SWI5, and CDC20 genes are transcribed in late S/G₂ phase, whereas CLB3 is expressed slightly earlier (22). The transcription of CLB1, CLB2, and SWI5 continues throughout mitosis and is switched off as the cells exit from mitosis (29, 53). URA3 served as an internal control for equal RNA loading since it is transcribed constitutively during the cell cycle. The overexpression of NDD1 did not cause any change in the level of the CLB3, CDC20, and URA3 transcripts. However, there was a three- to fivefold increase in the steady-state levels of CLB1 and CLB2 RNA (Fig. 4A). The SWI5 expression was also noticeably enhanced (Fig. 4A). CDC5 and ACE2 are two other genes that are expressed in late-S/G₂ phase, but their expression is not affected by the overexpression of NDD1 (data not shown). Thus, NDD1 modulates expression of only a subset of the late-S/G₂ genes. These data also suggest that the ability of NDD1 to suppress the cdc28-1N mutation is perhaps linked to its capacity to augment the expression of the mitotic cyclins CLB1 and CLB2.

View larger version (52K): [in this window] [in a new window]   FIG. 4.   (A) Overexpression of NDD1 results in enhanced expression of CLB1, CLB2, and SWI5. A wild-type strain containing pGAL1-NDD1 was grown to mid-log phase in raffinose medium. While one-half of the culture continued to grow in raffinose medium (left lanes), the other half was induced by the addition of galactose to express Ndd1 (right lanes). Total RNA was prepared from samples withdrawn at 2-h intervals and analyzed by Northern blotting for CLB1, CLB2, and SWI5 transcripts. CLB3 and CDC20 RNA transcripts are also shown for comparison. (B) NDD1 overexpression does not abrogate the cell-cycle-regulated transcription of CLB2. cdc15 cells carrying either a control plasmid (upper panel) or a plasmid harboring pGAL1-NDD1 (lower panel) were synchronized in telophase by growth in raffinose medium at 37°C for 2 h. NDD1 expression was induced for 2 h by the addition of galactose before the cells were released into galactose medium at 25°C. Samples were withdrawn at 15-min intervals and then analyzed by Northern blotting with CLB2-specific probe. The amount of total RNA loaded (30 µg) was visualized by methylene blue staining of the rRNAs.

NDD1 is essential for CLB1, CLB2, and SWI5 expression. Since Ndd1 can modulate the expression of CLB1, CLB2, and SWI5, we asked whether NDD1 is essential for the transcription of these genes. Since our efforts to isolate a "tight" temperature-sensitive allele of NDD1 were not successful, we used for these experiments an ndd1 strain kept alive by a CEN plasmid carrying GAL-NDD1 (US262). When transferred to glucose medium, this strain exhibits an arrest phenotype within 4 to 5 h. The cells were first grown in galactose until they reached log phase and were then shifted to glucose medium to switch off the NDD1 expression. Total RNA was isolated from samples withdrawn at various time intervals and analyzed for the presence of CLB1, CLB2, SWI5, CLB3, CDC20, NDD1, and URA3 RNAs by Northern blotting. The cells began to exhibit their characteristic terminal phenotype after 5 h and thereafter remained arrested with 2N content DNA, a short mitotic spindle, and an undivided nucleus. While the CLB3, CDC20, and URA3 transcripts were present throughout the course of the experiment, the CLB1, CLB2, and SWI5 RNAs became undetectable within 4 h of shifting to glucose medium and remained so for the remainder of the time course (Fig. 5A). As expected, NDD1 was not transcribed during this period. These results suggest that NDD1 is essential for the transcriptional activation of CLB1, CLB2, and SWI5 genes but not for the expression of CLB3 or CDC20.

View larger version (44K): [in this window] [in a new window]   FIG. 5.   (A) NDD1 is required for the normal expression of CLB1, CLB2, and SWI5. ndd1::LEU2 cells carrying pGAL1-NDD1 were grown to log phase in galactose medium. After 3 h, NDD1 expression from the GAL1 promoter was terminated by transferring the cells into glucose medium. RNA samples were taken after 4, 6, 8, and 10 h of incubation in glucose and then analyzed by Northern blotting with CLB1-, CLB2-, and SWI5-specific probes. CLB3 and CDC20 RNA transcripts are also shown for comparison. (B) Overexpression of CLB2 does not suppress the ndd1::LEU2 null mutation. ndd1::LEU2 cells, whose growth is dependent on wild-type NDD1 on a URA3 plasmid, were transformed with either TRP1-based control vector or TRP1 vector carrying either GAL1-CLB2 or GAL1-NDD1. Transformants were streaked onto galactose medium containing 5-FOA to counter-select for the URA3-based plasmid and to induce NDD1 expression from the GAL1 promoter. After incubation at 24°C for 3 days, only the cells carrying pGAL1-NDD1 survived.

Transcriptional activation by NDD1. Since our results raise the possibility that Ndd1 may be a transcription activator, we tested its ability to activate transcription in a heterologous context by using a yeast one-hybrid assay. The NDD1 coding region, fused in frame with the DNA-binding domain of the transcription factor GAL4, was put under the control of the ADH promoter and was transformed into a tester strain that carried the lacZ and HIS3 genes as reporters, both driven by three copies of a 17-mer Gal4-binding sequence. The Gal4-Ndd1 fusion was able to activate transcription of the reporter constructs, resulting in both the appearance of His⁺ colonies and the expression of -galactosidase (data not shown).

Normal periodic transcription of CLB2 requires a 240-bp promoter sequence. Maher et al. (30) have described a 55-bp UAS within the CLB2 promoter which contains one consensus Mcm1 binding site [DCCY(A/T)(A/T)(T/A)NN(G/A)G; D  C, Y = T or C] (25) and is both necessary and sufficient for the cell-cycle-regulated expression of CLB2. This 55-bp UAS is similar to a 55-bp element in the SWI5 promoter in both its sequence and its ability to bind a Mcm1-containing ternary complex (30).

View larger version (124K): [in this window] [in a new window]   FIG. 6.   (A) The 340-bp region of the CLB2 promoter. The 5' flanking region of CLB2 gene from 870 to 520 bp is shown. The boxed region indicates the presumptive 55-bp core promoter sequence. Mcm1 binding sites (MCE I, II, III, IV, and V) are indicated by arrows. Sff binding sites (SFRE) are indicated by dashed lines. The endogenous XhoI site is also shown. (B) Cell-cycle-regulated CLB2 transcription requires 240-bp (863 to 627) CLB2 UAS. To test the UAS activity at different stages of the cell cycle, cells carrying a single, integrated copy of the various UAS-ubiY lacZ reporter fusions were arrested in -factor (F), HU, or NOC. RNA samples were taken and analyzed by Northern blotting for lacZ and CLB2 transcripts. The amount of total RNA loaded (30 µg) was visualized by methylene blue staining of the rRNAs. The designations are as follows: dead, ubiY lacZ reporter without UAS sequences; 55 bp, CLB2 UAS[698 to 643]-ubiY lacZ; 240 bp, CLB2 UAS[863 to 627]-ubiY lacZ; 340 bp, CLB2 UAS[863 to 523]-ubiY lacZ; and SWI5, 55-bp SWI5 UAS-ubiY lacZ. (C) Pattern of gene expression from the 240-bp CLB2 UAS and the endogenous CLB2 promoter. Cells carrying one integrated copy of 240-bp CLB2 UAS-ubiY lacZ reporter gene were first arrested in G1 with -factor treatment (1 µg/ml) at 24°C and then allowed to resume cell cycle progression in medium without -factor. RNA samples were collected at 10-min intervals and analyzed by Northern blotting for lacZ and endogenous CLB2 RNA expression. The amount of total RNA loaded (30 µg) was visualized by methylene blue staining of the rRNAs. The degree of synchrony was measured in terms of the proportion of cells with anaphase spindles (lower panel).

The 240-bp fragment mediates potentiation of CLB2 expression by NDD1 overexpression. Since the 240-bp sequence is sufficient to impose cell cycle regulation on CLB2 transcription, we asked whether the enhancement of CLB2 expression by excess Ndd1 is effected through this promoter fragment. Therefore, we compared the efficacy of the 55-, 240-, and 340-bp fragments to elicit the upregulation of CLB2 transcription. A CEN vector carrying GAL-NDD1 was introduced into wild-type strains containing ubiY lacZ reporters driven by the various promoter fragments (described in the preceding section) at the URA3 locus. The cells were first grown in raffinose medium and then induced to express NDD1 for 2 h by the addition of galactose. The total RNA was analyzed for the presence of lacZ and the endogenous CLB2 transcripts. Whereas no lacZ RNA was detected when the reporter construct was under control of the 55-bp fragment, both the 240- and the 340-bp-fragment-driven reporters were transcribed to a readily detectable level (Fig. 7). The transcription from both the 240- and the 340-bp constructs was upregulated in response to the overexpression of NDD1 (Fig. 7). All three strains transcribed the endogenous CLB2 gene and, as expected, the expression was enhanced when Ndd1 was overexpressed. Hence, Ndd1 exerts its effect on CLB2 transcription via the 240-bp promoter sequence.

View larger version (70K): [in this window] [in a new window]   FIG. 7.   Enhancement of CLB2 transcription by excess Ndd1 requires the 240-bp CLB2 UAS. Strains containing the 55-, 240-, or 340-bp CLB2 UAS-ubiY lacZ reporter were transformed with pGAL1-NDD1 and grown in raffinose medium to mid-log phase. Galactose was then added to induce overexpression of NDD1 from the GAL1 promoter. RNA samples were withdrawn after galactose induction (at 0 and 2 h, respectively) and analyzed by Northern blotting with lacZ- and CLB2-specific probes. The amount of total RNA loaded (30 µg) was visualized by methylene blue staining of the rRNAs.

Ndd1 prevents formation of protein complexes on the 240-bp sequence, whose absence correlated with the expression of Clb2. Our results thus far strongly suggest that the 240-bp promoter fragment mediates both the cell-cycle-regulated expression of CLB2 and its potentiation by excess Ndd1. To determine whether Ndd1 is recruited to the protein complexes assembled on this cis-regulatory sequence, a radioactively labelled 240-bp fragment and the protein extract from cells expressing native-promoter-driven, fully functional HA₃-NDD1 were analyzed in a gel mobility shift assay. The assay yielded four specific complexes (designated I, II, III, and IV) (Fig. 8A, left panel, lane 2), all of which disappeared when challenged with a molar excess of nonradioactive probe (Fig. 8A, left panel, lane 6). However, when a molar excess of nonradioactive 55-bp fragment was used to challenge the binding, complex IV (and to some extent complexes I and II) was abrogated. This suggests that this particular complex is assembled on the core 55-bp region of the 240-bp fragment (Fig. 8A, right panel, lane 4). Complexes I, II, and IV contained Mcm1 because inclusion of anti-Mcm1 antibodies in the binding reactions abolished their formation (Fig. 8A, left panel, lane 5). Mcm1 is perhaps also present in complex III since anti-Mcm1 antibodies diminish, though do not abolish, its formation. However, anti-HA antibodies failed to supershift any of the complexes, indicating that Ndd1 is not present in these complexes (Fig. 8A, left panel, lanes 3 and 4). The failure of anti-HA antibodies to elicit any supershift in our experiments is not due to their general ineffectiveness in gel retardation assays, since the same antibodies (12CA5) have been shown to cause supershifts in a variety of different contexts (37a).

View larger version (58K): [in this window] [in a new window]   FIG. 8.   (A) DNA-protein complexes formed on 240-bp CLB2 UAS do not contain Ndd1. Radioactively labelled double-stranded 240-bp CLB2 UAS probe was used for gel retardation assays. The probe was mixed with 20 µg of protein extracts prepared from pep4 cells carrying native-promoter-driven HA3-NDD1. The designations are as follows: I, II, III, and IV, specific DNA-protein complexes I, II, III, IV; NS, nonspecific band; F, free probe. Antibody supershift reactions were performed with HA- or Mcm1-specific antibodies (left panel, lanes 3, 4, and 5). For competition, excess nonradioactive 240- or 55-bp competitor probe was added to the reaction prior to the addition of radiolabelled probe (left panel, lane 6; right panel, lanes 3 and 4). (B) Appearance of N complex on 240-bp CLB2 UAS coincides with the absence of CLB2 expression. Exponentially growing (cyc) pep4 cells were synchronized either in G1 phase, S phase, or mitosis by -factor (F), HU, or NOC treatments, respectively. Protein extracts made from these cultures were incubated with radiolabelled 240-bp CLB2 UAS probe (upper panel). Designations for the major, specific DNA-protein complexes are as in panel A. The antibody supershift reaction was performed with Mcm1-specific antibodies (lane 7). RNA samples from the cultures were analyzed by Northern blotting with CLB2-specific probe. The amount of total RNA loaded (30 µg) was visualized by methylene blue staining of the rRNAs. (C) HA-Ndd1 is unstable in G1. cdc15 cells carrying pGAL1-HA3-NDD1 were first synchronized in telophase by growth at 37°C and then released into 25°C to resume cell cycle progression as described in the text. Samples were withdrawn at 15-min intervals, and the stability of HA-Ndd1 was monitored by Western blotting with anti-HA antibodies (12CA5). Clb2 protein was detected with anti-Clb2 antibodies. The amount of Cdc28 protein, whose level remains constant throughout the cell cycle, shows equal loading of samples in all lanes. (D) Conditional overexpression of Ndd1 in G1 prevents formation of N complex. pep4 cells carrying three integrated copies of GAL1-NDD1 were arrested in G1 by -factor treatment in raffinose medium at 24°C. After 2 h, glucose was added to one-half of the culture, while the other half was induced for 2 h to express Ndd1 by the addition of galactose. Protein extracts made from these cells were used for gel retardation assay by employing radiolabelled 240-bp CLB2 UAS as a probe. Antibody supershift was performed with Mcm1-specific antibodies (lane 5). For competition, excess nonradioactive 240-bp competitor probe was added to the reaction prior to the addition of radiolabelled probe (lane 6).

A deletion analysis of Ndd1. Based on the requirement of Ndd1 for CLB1, CLB2 and SWI5 expression and its ability to activate transcription in a yeast one-hybrid assay, we had raised the possibility that Ndd1 may act as an activator of transcription. Since transcription activation in a heterologous context such as a one-hybrid assay can be artifactual, we asked whether the biological activity of Ndd1 (such as the suppression of cdc28-1N and mcm1 mutations) is related to its ability to activate transcription in a one-hybrid assay. Therefore, we made various constructs in which the DNA-binding domain of the GAL4 transcription factor was fused in frame with different parts of the NDD1 ORF and then tested these constructs for both their biological activity and their ability to activate transcription. Such a deletion analysis was also expected to identify the regions important for Ndd1 function.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The activities of many of the cell cycle effectors are regulated by protein modification, as well as by the controls operating at the level of transcription. Such multitiered regulation is presumably necessary in order to sharpen the timing of various cellular events during the progression through the division cycle. The cis-acting DNA sequences and the factors that activate the transcription of genes at START or at the onset of the S phase have been described in some detail (23, 28, 37, 38). However, the transcriptional regulation of the late S/G₂-phase-specific genes is poorly understood. So far, Mcm1 and another putative transcription factor, Sff, have been shown to participate in the control of gene expression at this stage of the cell cycle (1, 29). However, since Mcm1 is also required for the regulation of a number of other genes whose transcription is not periodic (31, 33, 51), it is unlikely that Mcm1 itself is the major target for cell cycle regulation. This notion is consistent with the finding that Mcm1 occupies both SWI5 and CLB2 promoters at all times, although their expression is cyclic (1).

We have identified and characterized a new gene, NDD1, that plays an essential role in the expression of the late-S-phase genes CLB1, CLB2, and SWI5, such that none of these genes are transcribed in its absence. Thus, it is a new player in the regulation of gene expression during late S phase. Ndd1 is not only essential for the transcription of CLB1 and CLB2, but its excess enhances the level of their expression. This observation may provide an explanation for the ability of Ndd1 to suppress the cdc28-1N mutation. The requirement of NDD1 for CLB1 and CLB2 expression may argue that the lethality of the ndd1 mutant is solely due to a lack of these mitotic cyclins. However, overexpression of CLB2 from the GAL1 promoter fails to suppress the nuclear-division defect of NDD1-deficient cells (Fig. 5B), implying that NDD1 is necessary for the expression of additional genes that participate in the process of nuclear division. NDD1 was also independently identified as a gene whose overexpression could suppress a crippling mutation in the SWI5 UAS, which abolishes SWI5 transcription (29a).

NDD1 is expressed in a cell-cycle-stage-dependent manner such that its expression peaks in S phase just prior to the expression of its target gene CLB2 (Fig. 3A). This may implicate NDD1 as a critical factor in the timing of CLB2 expression. Our observation that the constitutive expression of Ndd1 does not alter the cell-cycle-regulated pattern of CLB2 transcription (Fig. 4B) suggests that NDD1, though necessary for the activation of a subset of late-S-phase genes, is not a crucial determinant in the temporal control of their transcription. However, since the constitutive expression of Ndd1 leads to inefficient silencing of CLB2 transcription in late telophase (unpublished observations), the turning off of CLB2 transcription in late telophase may involve regulation of the Ndd1 protein. It is conceivable that the inactivation of the Ndd1 protein might be one of the steps required to switch off CLB2 expression at the end of mitosis. The instability of Ndd1 due to rapid degradation during mitotic exit is consistent with this notion (Fig. 8C). The mechanism, which renders Ndd1 unstable in G₁, remains unclear. We found four destruction-box-like sequences in the middle region of the protein, but their removal did not affect its stability (unpublished results). Another important factor in silencing CLB2 expression may be the Cdc28-Clb kinase itself. It has been previously shown that Cdc28-Clb kinase complex stimulates CLB2 expression via a positive feedback loop (4). Consequently, inactivation of the kinase by abrupt proteolysis of Clb proteins during the M/G₁ transition could lead to a rapid decline in CLB2 expression.

The genetic interaction with MCM1 and its ability to enhance CLB1, CLB2, and SWI5 transcription lead us to suspect that NDD1 may influence gene expression by modulating their promoter activity. During the course of our studies, we discovered that the previously reported 55-bp UAS (30) in the CLB2 promoter, which contains a pair of regulatory sites comprised of one MCE and one SFRE, is unable to drive the expression of a reporter gene in our assay system (see Materials and Methods). In an identical assay, the 55-bp UAS of SWI5 gene not only elicited transcription but also showed the expected pattern of expression during the cell cycle. This suggests that the inactivity of the 55-bp CLB2 UAS in our experiments is not due to a faulty assay system. By further investigations, we have identified a 240-bp fragment that is sufficient for both the expression and the cell cycle regulation of CLB2. The 240-bp sequence contains within it the 55-bp presumptive UAS and another pair of MCE and SFRE flanked by two additional MCE (Fig. 6A) (reference 1 and this study). This configuration is also capable of mediating the transcriptional enhancement caused by the overexpression of Ndd1; the 55-bp UAS alone, on the other hand, remains unresponsive (Fig. 7). The CLB1 promoter also harbors two pairs of MCE and SFRE but their spatial arrangement is not identical to that in the CLB2 promoter (1). It is intriguing that while one set of MCE and SFRE can appropriately regulate SWI5 expression, it is not sufficient for the activation of CLB2 transcription. Perhaps this is due to the differences in both the regulatory sequences themselves and in the promoter sequence within which these elements are embedded.

The 240-bp promoter fragment promotes the assembly of four prominent protein complexes. While the exact composition of these complexes remains unknown, they all appear to contain Mcm1 (Fig. 8A). Extracts prepared from cells traversing synchronously through the cell cycle are capable of assembling, albeit to various extents, all four complexes on this CLB2 promoter fragment (data not shown). Interestingly, Ndd1 is not recruited to any of these protein complexes, as is suggested by the gel mobility shift assay (Fig. 8A), despite its requirement for CLB1, CLB2, and SWI5 transcription and its ability to enhance gene expression. In immunoprecipitation experiments, Ndd1 neither associates with Mcm1 nor with any of the major components of the general transcription machinery such as yeast TBP, TAF145, or TAF90 (unpublished data). These observations suggest that Ndd1 cooperates indirectly with the transcriptional apparatus, possibly through intermediary proteins, to regulate the expression of late-S-phase genes. How then, does Ndd1 exert its effect on gene expression? One possibility is that Ndd1 is an effector of Cdc28-Clb kinase activity and that it affects CLB2 transcription by modifying the efficacy of the positive-feedback loop. However, this seems unlikely, because the Ndd1-depleted cells during NOC-induced arrest contain significant levels of the mitotic kinase activity but yet do not transcribe CLB2 (unpublished results).

Alternatively, Ndd1 may prevent a repressor from binding to the CLB2 promoter, thus allowing the activation of transcription. The finding that Mcm1, an activator of transcription, occupies the CLB2 promoter throughout the cell cycle (1) strengthens the possibility that a repressor may be, in part, responsible for the termination of CLB2 transcription in G₁. In the context of Ndd1 function, this repressor model is consistent with some of our findings: (i) a new protein complex N is detected on the 240-bp promoter fragment during G₁ when CLB2 transcription is abruptly switched off; (ii) this complex is undetectable in HU- or NOC-arrested cells where CLB2 is actively transcribed; (iii) overexpression of Ndd1 in G₁ abolishes the formation of the N complex; and (iv) in wild-type cells, Ndd1 is rapidly degraded in G₁, concomitant with the silencing of CLB2 expression. Complex III behaves like complex N in that its abundance is enhanced during G₁, dramatically reduced during NOC-induced arrest when CLB2 is maximally transcribed, and negatively influenced by the overexpression of Ndd1. Thus, complexes N and III could be repressor assemblages capable of terminating CLB2 expression. Although the presence of Ndd1 abolishes the formation of the presumptive repressor complexes, its overexpression does not elicit CLB2 expression in G₁. This implies that a lack of Ndd1 protein may not be the sole reason for the termination of CLB2 transcription; a progressive weakening of the positive-feedback loop upon proteolysis of Clb proteins may also be critical. The nature of the hypothetical repressor is so far unknown. To identify such a repressor, we have embarked upon a genetic screen to isolate mutations that will lead to inappropriate CLB2 expression in G₁.

With the identification of NDD1, we have added a new element in the regulation of gene expression in late S phase. However, further investigations will be required to uncover the mechanism that ensures the correct timing of the onset and termination of the expression of these genes.