INTRODUCTION

Top Previous Next References

The entrance of eukaryotic cells into mitosis from G₂ is a highly regulated event which involves a complex series of interactions among an evolutionarily conserved set of kinases and phosphatases (reviewed in references 41 and 57). Genetic studies with the fission yeast Schizosaccharomyces pombe have identified molecules that participate in the regulation of the G₂/M transition through phosphorylation (34, 44-46), including cdc2⁺, cdc13⁺, cdc25⁺, wee1⁺, and mik1⁺. cdc2⁺ encodes the single essential cyclin-dependent kinase (CDK) required to drive fission yeast cells through the cell cycle. During G₂, cdc2p is associated with the mitotic B-type cyclin encoded by cdc13⁺. The activity of cdc2p/cdc13p complexes is inhibited by phosphorylation of cdc2p at tyrosine 15, an event which is catalyzed by the wee1p and mik1p protein tyrosine kinases. As fission yeast cells achieve a critical size required for entry into mitosis, the cdc25p protein tyrosine phosphatase activates cdc2p/cdc13p complexes by catalyzing the dephosphorylation of tyrosine 15.

In addition to genes which regulate cdc2p function directly, two other S. pombe genes, cdc5⁺ and cdc28⁺, are known to be required during G₂ for entry into mitosis. cdc28⁺ encodes a DEAH-box protein homologous to the Saccharomyces cerevisiae splicing factors Prp2p, Prp16p, and Prp22p, suggesting that RNA processing is in some way required to promote the G₂/M transition (33). The cdc5⁺ gene product contains a putative DNA-binding domain (DBD) that is most similar to those contained within Myb-related proteins and is the only putative transcription factor cloned in S. pombe that performs an essential role during G₂/M progression (49).

Recently identified Arabidopsis thaliana, Homo sapiens, and Xenopus laevis cDNAs encode proteins closely related to S. pombe cdc5p (4, 23, 58). Of these, the A. thaliana cDNA, AtCDC5, was found to be a functional homolog of cdc5⁺ since it was capable of complementing the temperature-sensitive growth defect of the fission yeast cdc5-120 mutant (23). A partial X. laevis Cdc5 cDNA clone was isolated in a screen designed to identify mitotic phosphoproteins that can be phosphorylated in vitro by cyclin B-Cdc2 (58). In addition, Bernstein and Coughlin (4) have reported that the human Cdc5 protein, PCDC5RP (pombe Cdc5-related protein), translocates from the cytoplasm to the nucleus of cultured mammalian cells stimulated with serum, implicating the human protein as a potential effector in a mitogen-activated signaling pathway.

To explore further the structural and functional conservation of cdc5p, we have isolated cDNAs and genes from S. cerevisiae, Caenorhabditis elegans, Drosophila melanogaster, and H. sapiens that encode homologs of S. pombe cdc5p and report here an analysis of these genes. Supporting a role for this subfamily of Myb-related proteins in G₂/M cell cycle progression, budding yeast cells lacking CEF1 (S. cerevisiae homolog of cdc5⁺) arrest growth during G₂/M. Furthermore, we demonstrate that human and fly Cdc5 proteins are functionally homologous to S. pombe cdc5p because they are able to complement the temperature-sensitive lethality of the S. pombe cdc5-120 mutant.

Aside from the sequence similarity of Cdc5-related proteins to the transcriptional activator c-Myb, two lines of evidence have implicated Cdc5 in DNA binding. First, the Myb repeats of S. pombe cdc5p bind DNA-cellulose (49). Second, the Myb repeats of AtCDC5 were found to be capable of preferentially binding to the DNA sequence CTCAGCG (23). In this report, we demonstrate by mutational analysis of the Myb repeats of Cef1p and localization of the cdc5-120 mutation to the Myb repeat R1 of cdc5p that the Myb repeats of these proteins are essential for their function in vivo. In the budding yeast S. cerevisiae, the Mcm1p transcription factor, in concert with an uncloned gene encoding a protein named Sff (SWI5 transcription factor), is required for transcription of at least six genes that normally exhibit a G₂/M-specific expression pattern, namely, CLB1, CLB2, CDC5, SWI5, ACE2, and ASE1 (1, 26, 35). On the basis of phenotypic analysis of S. cerevisiae cells genetically depleted of CEF1 activity and the ability of these cells to transcribe gene targets of Mcm1p/Sff, we conclude that CEF1 does not encode Sff. Our work demonstrates that fission yeast cdc5p is a member of a highly conserved subfamily of Myb-related proteins that plays a key role in cell cycle progression at the G₂/M boundary.

MATERIALS AND METHODS

Top Previous Next References

Strains, growth media, and genetic methods. The S. cerevisiae and S. pombe strains used in this study are listed in Table 1. Media used to grow S. pombe and general genetic manipulations of S. pombe were described previously (39). S. pombe transformations were performed by electroporation (50). Repression of transcription from the nmt1⁺ promoter (37) was achieved by addition of thiamine to a concentration of 2 µM.

Isolation and molecular cloning of CEF1 and the Cdc5Ce, Cdc5Dm, and Cdc5Hs. A 2,555-bp fragment encompassing the CEF1 gene was amplified from genomic DNA by using primers cef1.P5 (5'-CCCGGATCCGGCTTCTTATCTGTGGTC-3') and cef1.P6 (5'-CCCCTGCAGCCGGATGGTACTTCTTAG-3') (introduced restriction enzyme sites are italicized). cef1.P5 introduces a BamHI site onto the 5' end of the CEF1 gene, and cef1.P6 introduces a PstI site onto the 3' end of the CEF1 gene. The PCR product was cleaved with BamHI and PstI and ligated into pRS415 (9), which had been cut with BamHI and PstI to produce the plasmid pCEF1. The CEF1 open reading frame (ORF) was sequenced to ensure the absence of PCR-induced mutations. A URA3-CEN plasmid (pRS416) containing the CEF1 gene, which was used in all plasmid shuffle assays, was constructed by inserting the BamHI/HindIII CEF1 fragment from pCEF1 into pRS416, which had been cleaved with these enzymes. The 1,770-bp CEF1 ORF was amplified from S. cerevisiae genomic DNA by using primers SC5531 (5'-GCCGATATCTACTAGCATTTCAAGATGCCC-3') SC5351 (5'-GAAGATATCCTATACTAATGCTATATGGAA-3'). Both primers were designed to add EcoRV restriction sites to the ends of the coding region during amplification. The amplified fragment was cut with EcoRV and cloned into the EcoRV site of pSKREG351 to produce pSKREG351CEF1. pSKREG351 is pBS(SK⁺) (Stratagene, LaJolla, Calif.) containing 585 bp of nucleotide sequence immediately upstream of the S. pombe cdc5⁺ ATG. The cdc5⁺ promoter sequence was amplified by using primers 5REG351 (5'-GAAGATATCAACCCTGAC-3') and T7 and cloned into pBS(SK⁺) which had been linearized with EcoRV and treated with dTTP and Taq polymerase. The EcoRV fragment from pSKREG351CEF1 was subcloned into the SmaI site of the 2 µm LEU2-based shuttle vector pRS425GAL1 (42) to produce pGAL1-CEF1. pCEF1 and pGAL1-CEF1 both produce functional Cef1p since they are capable of restoring growth to a strain lacking the CEF1 gene.

Disruption of CEF1 and construction of KGY1120. Deletion of the CEF1 coding sequence was performed by replacing nearly the entire ORF with a DNA fragment containing the HIS3 gene. The cef1-1::HIS3 null allele was generated by PCR as described by Baudin et al. (3; see also reference 53), using the primers 5'-GCATTTCAAGATGCCCCCCGTACCAATATACGTGAGCAGATTGTACTGAG-3' and 5'-CATGGCTTGAAGACGCTTTCTTCTACGACCCTCAAGCGCTCCTTACGCATCTG-3' and pRS313 (55) as a template to amplify the HIS3 gene. The underlined portions of each primer correspond to the HIS3 flanking sequence. The resulting PCR product contains the complete HIS3 gene flanked by 35 bp of CEF1 5' sequence corresponding to nucleotides 10 to +25 and 35 bp of CEF1 3' sequence corresponding to nucleotides +1714 to +1748. When substituted at the CEF1 locus, this DNA fragment deletes the entire CEF1 gene except for the initial 8 codons and final 18 codons. The diploid strain YPH274 was transformed with this DNA fragment, and two independent transformants (KGY856 and KGY857) that contained a cef1-1::HIS3 allele substituted correctly at one of the two CEF1 loci were identified by Southern analysis. These strains were sporulated, and tetrads were dissected and analyzed.

Cytological methods and flow cytometry. To visualize nuclei, cells were fixed with 70% ethanol overnight at 4°C, washed twice with phosphate-buffered saline, and resuspended in 1 µg of DAPI per ml. For immunostaining of tubulin, cells were processed as outlined by Pringle et al. (51) and stained with TAT-1 monoclonal antibody (65) followed by a Texas red-conjugated goat-anti mouse immunoglobulin G secondary antibody. Cells were viewed with a Zeiss Axioskop20 with the appropriate set of filters, and images were captured with a Zeiss ZVS-47DEC image-capturing system. Yeast cells were processed for flow cytometric analysis as described previously (49).

Preparation of total RNA and Northern blot analysis. S. pombe wild-type (strain 972) and temperature-sensitive strains were grown to mid-log phase at 25°C in yeast extract medium. Aliquots of the cultures were collected while the remaining portions were shifted to 37°C. Cells at 37°C were collected following a 5-h incubation. S. cerevisiae KGY1120 cells were grown to mid-log phase in synthetic medium containing the appropriate supplements and 2% raffinose-2% galactose (SRG medium) as the carbon sources. A portion of the culture was collected for processing, while the remainder of the culture was harvested and resuspended in synthetic medium containing glucose (SD medium) and the appropriate supplements. Cells were collected for processing at 8, 10, 12, 14, and 16 h following the SRG-to-SD medium shift. Total RNA was prepared from S. cerevisiae and S. pombe cells by glass bead disruption as described by Moreno et al. (39) or by RNeasy columns (Qiagen, Chatsworth, Calif.). Twenty micrograms of total RNA was electrophoresed on 1% formaldehyde agarose gels, blotted onto Gene Screen Plus membranes (DuPont-NEN, Boston, Mass.), and fixed by incubation at 80°C under vacuum for 2 h. Hybridization was performed as described by Church and Gilbert (10). Templates for probes were prepared from genomic DNA by PCR or by releasing appropriate fragments from plasmids by restriction enzyme digestion. Probes for hybridization were prepared by using the Rediprime random primer labeling system (Amersham, Arlington Heights, Ill.).

RESULTS

Top Previous Next References

S. pombe cdc5p is conserved throughout evolution: isolation of and structural analysis of Cdc5 gene homologs. We previously reported that the S. pombe cdc5⁺ gene encodes a protein that has significant sequence similarity to the c-Myb family of proteins (49). More recently, genome sequencing efforts have revealed that gene products much more closely related to cdc5p than c-Myb family members exist in diverse eukaryotic organisms, including S. cerevisiae, C. elegans, and H. sapiens. Additionally, Hirayama and Shinozaki (23), Bernstein and Coughlin (4), and Stukenberg et al. (58) have reported the identification of cDNAs from A. thaliana, H. sapiens, and X. laevis, respectively, whose predicted proteins are closely related to cdc5p. To determine if the cell cycle function of S. pombe cdc5⁺ also has been conserved throughout evolution, we isolated presumptive homologs of cdc5⁺ from S. cerevisiae, C. elegans, D. melanogaster, and H. sapiens.

View larger version (83K): [in this window] [in a new window]   View larger version (21K): [in this window] [in a new window]   FIG. 1.   Isolation of homologs of S. pombe cdc5+. (A) Alignment of S. pombe cdc5p homologs by using ClustalW1.6 (14, 60). Residues found to be identical or similar to those of Cdc5Hs are highlighted in black or grey, respectively, to emphasize the evolutionary conservation of Cdc5 homologs. Roman numerals signify the beginning of a Myb repeat. Asterisks indicate the positions of the amino acid residues which are classically occupied by tryptophan residues in Myb repeats. Percent identities among the various Cdc5 family members are indicated at the bottom of this panel. (B) Alignment of MLR3 of Cdc5 family members with R3 of human c-, A-, and B-Myb and Drosophila Myb. Alignment was performed by using ClustalW1.6. Identical and similar residues found within the various proteins are highlighted in black and grey, respectively. Residues found to be identical between all proteins are indicated by arrows. Sp, S. pombe; Sc, S. cerevisiae; Hs, H. sapiens; Dm, D. melanogaster. (C) Phylogenetic analysis of Cdc5 and c-Myb family members.

The temperature-sensitive growth defect of the S. pombe cdc5-120 mutant is complemented by both human and fly Cdc5 cDNAs. To determine if CEF1 and the Cdc5Ce, Cdc5Dm, and Cdc5Hs cDNAs were capable of rescuing the temperature-sensitive lethality of the S. pombe cdc5-120 mutant (46, 49), we introduced these cDNAs into cdc5-120 cells under control of the thiamine-repressible nmt1⁺ promoter (37). Full-length cDNAs for CEF1 and the Cdc5Ce and Cdc5Hs cDNAs were incapable of rescuing growth of cdc5-120 cells at 36°C in both the presence and absence of thiamine (Fig. 2). In contrast, the full-length Cdc5Dm cDNA was capable of rescuing growth of cdc5-120 cells when induced (in the absence of thiamine) but not when repressed (in the presence of thiamine) (Fig. 2). Microscopic examination of cdc5-120 cells overexpressing CEF1 and the Cdc5Hs cDNA revealed that many of them were elongated at 25°C (data not shown). Overexpression of full-length CEF1 and the Cdc5Hs cDNA had the same phenotypic consequence in wild-type cells (data not shown).

View larger version (93K): [in this window] [in a new window]   FIG. 2.   The temperature-sensitive growth defect of cdc5-120 is complemented by human and fly Cdc5 proteins. Full-length cDNAs for CEF1 and the Cdc5Ce, Cdc5Dm, and Cdc5Hs cDNAs were placed under control of the nmt1+ promoter in pREP3X. In addition, sequences for products of these cDNAs were truncated at their C termini by 129 aa (CEF1dCT), 337 aa (Cdc5Ce dCT), 293 aa (Cdc5Dm dCT), 441 (EcoRI), and 196 (Acc651) aa (Cdc5Hs) and placed under control of the nmt1+ promoter in pREP3X. S. pombe cdc5+ is pIRT2 (39) harboring full-length cdc5+ cDNA under control of 585 bp of the cdc5+ promoter sequence. pREP81cdc5Sp is pREP81 (2) harboring full-length cdc5+ cDNA. KGY360 transformants carrying Cdc5Hs constructs were struck to minimal medium with or without thiamine and incubated for 18 h at 25°C to induce or repress expression of the inserts, respectively, and then shifted to 36°C for 3 days. KGY360 transformants carrying CEF1, Cdc5Ce cDNA, and Cdc5Dm cDNA constructs were struck to minimal medium with or without thiamine at 36°C and grown for 3.5 days.

View larger version (41K): [in this window] [in a new window]   FIG. 3.   Deletion of CEF1 blocks cell growth at G2/M. (A) KGY1120 fails to grow on medium containing glucose as a carbon source. KGY1120 and the isogenic wild-type strain, KGY1140, were struck onto solid medium containing either 2% glucose or 2% raffinose-2% galactose as a carbon source(s) and incubated at 30°C for 4 days. (B and C) Growth of KGY1140 (B) or KGY1120 (C) cells in liquid SD medium or SRG medium is shown. KGY1140 or KGY1120 cells were grown to mid-log phase in SRG medium at 32°C and then shifted to SD medium (), or maintained in SRG (). The medium shift was performed at time (t) 0 h. (D) Cells from experiments shown in panels B and C were collected every 2 h, stained with propidium iodide, and subjected to flow cytometric analysis as described in Materials and Methods. Linear fluorescence histograms show relative DNA content in arbitrary units on the horizontal axis and cell number on the vertical axis. Arrows indicate the positions of 1N and 2N DNA peaks. (E) Growth-arrested KGY1120 cells contain short intranuclear spindles. Microtubules were detected by indirect immunofluorescence using the -tubulin monoclonal antibody TAT-1 (65) followed by staining with Texas red-conjugated secondary antibodies. DNA was visualized by using DAPI. Arrowheads indicate cells that contain a single tubulin plaque instead of a short intranuclear spindle. Raf/Gal, medium containing 2% raffinose-2% galactose; Glu, medium containing 2% glucose.

Mutational analysis of Cef1p. The importance of the two Myb repeats and MLR3 of Cdc5 family members to their function is predicted by the high degree of sequence identity among them (Fig. 1A). As expected, Myb repeats R1 and R2 of Cef1p are essential for its function in vivo; when expressed under control of the GAL1 promoter, a truncated version of Cef1p lacking the R1 and R2 Myb repeats (pGAL1-CEF1CT) is incapable of rescuing the growth of the cef1-1::HIS3 mutant (Fig. 4). Furthermore, the mutation responsible for causing the temperature sensitivity of the S. pombe cdc5-120 mutant was sequenced and found to be present in the first Myb repeat (R1). In cdc5-120p, the second tryptophan of R1 (W29) is converted to an arginine residue as a result of a T:A to C:G mutation at the corresponding codon (TGG to CGG). Although these results underscore the importance of the Cef1p and cdc5p Myb repeats, a construct containing only R1 and R2 of Cef1p under control of the GAL1 promoter (pGAL1-CEF12R) is incapable of rescuing the growth of the cef1-1::HIS3 mutant, indicating that other regions of Cef1p are essential for function as well (Fig. 4).

View larger version (38K): [in this window] [in a new window]   FIG. 4.   Mutational analysis of Cef1p. Amino acids predicted to be critical for maintaining the structure of the Myb repeats of Cef1p were replaced with the indicated amino acid(s) as described in Materials and Methods. Substitution mutations were constructed in pCEF1, which is pRS415 (9) harboring a 2.5-kb genomic DNA fragment encompassing the CEF1 coding region. In addition to performing site-directed mutagenesis of the Cef1p gene, various domains of Cef1p were removed. pGAL1-CEF12R, Cef1p R1R2; pGAL1-CEF1CT, Cef1p CT including MLR3; pGAL1-CEF1dR1, Cef1p lacking R1; pGAL1-CEF1dR2, Cef1p lacking R2. These mutants were assayed for their abilities to rescue growth of the cef1-1::HIS3 mutant by tetrad analysis and/or plasmid shuffle. Plasmid shuffle with the pRS425GAL1-based plasmids was performed on 5-fluoroorotic acid medium containing raffinose and galactose as the carbon sources. ++, rescue or growth comparable to that of wild-type CEF1; +, reduced rescue of growth and colony formation; +/, poor rescue of growth; , failure to rescue growth of the cef1-1::HIS3 mutant or no growth; NA, not applicable.

cef1-1::HIS3

cef1-1::HIS3

cef1-1::HIS3

cef1-1::HIS3

cef1-1::HIS3

cef1-1::HIS3

CEF1 and S. pombe cdc5⁺ are not required for expression of genes which are known to be transcriptionally up-regulated during G₂/M. The G₂/M and G₂ growth arrest phenotypes of S. cerevisiae cells lacking CEF1 activity and S. pombe cdc5-120, respectively, and the similarity of Cdc5 family members to c-Myb prompted us to examine a potential involvement of these proteins in the transcriptional activation of genes known to be up-regulated during G₂/M. In S. cerevisiae, at least six genes show a G₂/M-specific pattern of transcription: SWI5, CLB1, CLB2, ACE2, CDC5, and ASE1 (1, 26, 35). The first five of these genes, and very likely the sixth as well, require the activity of the MADS-box transcription factor encoded by MCM1 and an uncloned protein termed Sff (1, 26, 35). To test whether Cef1p was required for transcriptional activation of these genes, total RNA was prepared from cells depleted of CEF1 activity and subjected to Northern blot analysis using portions of CLB1 and SWI5 ORFs as probes. As shown in Fig. 5A, CLB1 and SWI5 transcripts persist in cells depleted of CEF1 activity, suggesting that CEF1 is not required for expression of these genes.

View larger version (56K): [in this window] [in a new window]   FIG. 5.   CEF1 and cdc5+ are not required for transcriptional activation of genes known to be transcriptionally up-regulated during G2/M. (A) CEF1 activity is not required for transcription of SWI5 and CLB1. KGY1120 cells were grown under permissive (SRG medium) or restrictive (SD medium) nutritional conditions as described in the legend to Fig. 3 and Materials and Methods. Total RNA was isolated from cells at 8, 10, 12, 14, and 16 h and subjected to Northern blot analysis using portions of ACT1, SWI5 and CLB1 ORFs as probes. ACT1 transcript levels served as a loading control. WT, wild type; RG, synthetic medium containing 2% galactose and 2% raffinose as carbon sources. (B) cdc5-120 is not defective for transcription of cdc13+ and cdc25+. Logarithmically growing wild-type (strain 972) or temperature-sensitive (ts) strains cdc25-22 or cdc5-120 were collected during growth at 25°C or after incubation at 37°C for 5 h. Total RNA was prepared from these cells and subjected to Northern blot analysis using portions of ura4+, cdc13+, and cdc25+ ORFs as probes. ura4+ transcript levels served as a loading control.

In this study, we have reported the identification and characterization of cDNAs and genes from S. cerevisiae, C. elegans, D. melanogaster, and H. sapiens that encode proteins closely related to S. pombe cdc5p. Bernstein and Coughlin (4) have independently reported the isolation of human Cdc5 (PCDC5RP). In addition, Hirayama and Shinozaki (23) have reported the isolation and characterization of AtCDC5, a gene which encodes the A. thaliana homolog of fission yeast cdc5p. Stukenberg et al. (58) have identified an X. laevis relative, and genome sequencing efforts have revealed the existence of a Cdc5-related protein in mice (for an example, see the EST under GenBank accession no. AA269568). When subjected to phylogenetic analysis, all of the analyzed Cdc5 family members display significantly greater similarity to S. pombe cdc5p than to any other Myb-related protein (Fig. 1C). cdc5p is thus a member of a family of Myb-related proteins that has been highly conserved throughout the eukaryotic lineage. In S. cerevisiae, only one gene, CEF1, encodes a protein that bears any significant similarity to fission yeast cdc5p. This also appears to be the case in complex metazoan systems, since all of about 20 Cdc5Hs-related human sequences in dbEST appear to represent overlapping segments of a single gene which we have mapped to chromosome 6p21. Since mammalian Cdc5 activity is likely encoded by a single genetic locus, it is possible that inappropriate modulation of Cdc5 activity may be associated with the onset or progression of tumorigenesis. Although no obvious previously described cancer or disease genes have been mapped to this locus, knowledge of its chromosomal location provides an opportunity to investigate such a possibility directly.

Several lines of evidence suggest that the cell cycle function of S. pombe cdc5⁺ has also been conserved throughout the eukaryotic lineage. Like S. pombe cdc5⁺, S. cerevisiae CEF1 is essential for viability. Haploid cells lacking a functional CEF1 gene undergo growth arrest with a 2N DNA content and short intranuclear spindles, indicating that they are arrested during G₂/M. Interestingly, a noticeable population of cells lacking CEF1 activity display abnormal nuclear morphologies, indicating that Cef1p may be important for maintenance of nuclear integrity prior to or during the process of nuclear division. Similarly, in S. pombe, approximately 30% of fission yeast cells lacking the cdc5⁺ gene also display aberrant nuclear morphologies (49). Despite the sequence identity possessed by Cef1p and cdc5p, CEF1 is unable to complement the temperature-sensitive growth defect of S. pombe cdc5-120 at its restrictive temperature of 36°C. Similar to AtCDC5 (23), H. sapiens and S. cerevisiae Cdc5 cDNAs are unable to rescue growth of S. pombe cells harboring a deletion of cdc5⁺. Human and fission yeast Cdc5 proteins were also observed to be incapable of rescuing the lethality of S. cerevisiae cef1 cells. The inability of various Cdc5 family members to rescue growth of cef1 or cdc5 cells is not surprising given that the C termini of Cdc5 family members are less well conserved than the putative DBDs and that the C termini of S. pombe cdc5p and S. cerevisiae Cef1p have indispensable functions (49, 49a) (Fig. 4). Full-length Cdc5Dm and partially truncated human Cdc5, however, were capable of rescuing growth of cdc5-120 cells at 36°C, thus demonstrating that Cdc5Hs and Cdc5Dm are functional homologs of fission yeast cdc5p. The inability of the Cdc5Ce cDNA and S. cerevisiae CEF1 to rescue growth of cdc5-120 is most simply explained by a limited degree of sequence identity possessed by the Myb repeats of S. pombe cdc5p, Cdc5Dm, and Cdc5Hs (Fig. 1A). Recently, Bernstein and Coughlin (4) have reported that an epitope-tagged version of human Cdc5 translocates from the cytoplasm to the nuclei of serum-deprived cultured mammalian cells upon stimulation with serum. Based on these results, the authors speculated that Cdc5 may function in a mitogen-activated signaling pathway during the G₀/G₁ transition. Although a definitive role(s) for Cdc5 in the mammalian cell cycle requires further analysis, our data indicate that an essential, conserved function of Cdc5 in the eukaryotic cell cycle is at the G₂/M transition.

In S. cerevisiae, the Mcm1p transcription factor, in combination with the uncloned SWI5 factor (Sff), is required during G₂/M to activate transcription of at least six genes which are necessary for entry into and completion of mitosis: CLB1, CLB2, ACE2, SWI5, CDC5, and ASE1 (1, 26, 35). Sff is absolutely required for transcription of SWI5 because mutation of the Sff binding site results in the loss of SWI5 upstream activating sequence activity (35). S. cerevisiae cells lacking MCM1 activity undergo growth arrest with replicated DNA and elongated buds (1), a phenotype commonly associated with S. cerevisiae mutants which possess low Clb-associated Cdc28p (Cdk1) kinase activity (31, 52, 59). Cells lacking CEF1 undergo arrest with a round-budded phenotype, which indicates the presence of Clb-associated Cdk1 activity (31). The phenotypic difference between cells lacking MCM1 and CEF1 activities indicates that CEF1 is unlikely to encode Sff. We have confirmed this hypothesis by demonstrating that cells depleted of CEF1 activity are capable of accumulating transcripts of Mcm1p and Sff target genes.

In S. pombe, cdc13p and cdc25p levels oscillate in a cell cycle-dependent manner, with maximal levels observed near the G₂/M transition (12, 40). cdc25⁺ transcript is most abundant near the G₂/M transition, indicating that accumulation of cdc25p during G₂ is achieved, at least in part, through its transcriptional activation (40). The cdc5-120 mutant is not defective for transcription of these mitotic regulators. We conclude from these results and those obtained from Cef1p-depleted cells that Cdc5 does not regulate the G₂/M transition in eukaryotic cells at the level of transcription of these known mitotic regulators.

Although cdc5p and Cef1p do not appear to function through activating transcription of genes known to be up-regulated during G₂/M, it is still tempting to assign these proteins a role as DNA-binding proteins for the following reasons: (i) all Cdc5 family members analyzed thus far contain two highly conserved Myb repeats, (ii) the S. pombe cdc5-120 mutation converts the second tryptophan in R1 of cdc5p to an arginine residue, and (iii) the Myb repeats of AtCDC5 have been shown to preferentially bind a specific nucleotide sequence (23). For these reasons, we have tried to identify a preferential DNA-binding sequence for S. pombe cdc5p and S. cerevisiae Cef1p. Specifically, we employed the cyclic amplification and selection of target sequence (CASTing) approach with the binding (23, 62) and washing (16) conditions described previously. The N-terminal 201 amino acids of Cef1p, which contain R1, R2, and MLR3, were fused downstream of glutathione S-transferase (GST), and this fusion protein was used for CASTing. An oligonucleotide library containing a core of 20 random nucleotides was used as a probe. A similar approach was used to identify a target sequence of AtCDC5 (23). While it is clear that the Myb repeats of Cef1p and cdc5p have the ability to bind DNA, we have not been able to select a high-affinity binding sequence for these proteins although under the same conditions a fusion protein between GST and the DBD of AML was able to select the sequence TGT/cGGT, the known high-affinity target sequence of AML (38, 49, 49b). In a separate experiment, a hexahistidine-tagged version of the DBD of Cdc5Dm was used in an electrophoretic mobility shift assay with the AtCDC5 binding sequence as a probe. We reasoned that since the DBDs of these proteins have 80% identity, they may be capable of binding the same sequence. While a modest binding of Cdc5Dm to this probe was observed, the binding was abrogated in the presence of increasing amounts of nonspecific competitor, indicating that this interaction occurs with low affinity (37a). In this regard, it is noteworthy that AtCDC5 binding to the sequence CTCAGCG was also reduced with nonspecific competitor DNA (23). While these data may argue that Cdc5 does not bind a particular DNA sequence with high affinity, we cannot exclude the possibility that additional regions of these proteins are required for sequence-specific DNA binding or that they require additional factors to bind DNA with high affinity.

In this study, we have analyzed the in vivo consequences of altering residues within the Cef1p Myb repeats which are predicted to disturb the conformation of these domains. Of particular relevance to this work, mutational studies of c-Myb have demonstrated that hydrophobic residues forming the backbones of the DBD helices are essential for the DNA-binding activity of c-Myb in vitro and its trans-activation activity in reporter assays (27, 54). Surprisingly, we found that the Cef1p Myb repeats are quite resiliant to similar changes; single, double, or even certain triple mutant combinations in residues which form the hydrophobic backbones of the second or third helices in both R1 and R2 do not abolish Cef1p function in vivo, although many produce a temperature-sensitive growth defect. Although these results are unexpected, it is important to bear in mind that our assay determined the functional effects of altering helical backbone residues in vivo and is therefore not directly comparable to the DNA binding, transcriptional activation, and differentiation assays that have traditionally been used in mutational analyses of c-Myb. To our knowledge, systematic mutational analyses of Myb domains in other proteins, such as those presented here, have yet to be reported. Thus, it is possible that interaction of Cef1p with other proteins in vivo may stabilize the conformation of mutant Myb repeats, thus allowing Cef1p to bind DNA.

An intriguing, equally plausible interpretation is that in vivo, Cef1p may cooperate with a second DNA-binding protein whose activity compensates for a diminution of Cef1p DNA-binding activity. Evidence does exist for Myb-related proteins functioning in heterodimeric complexes. c-Myb itself cooperates with heat shock factor 3 (HSF3) to activate transcription from the hsp70 promoter (28, 29). In S. cerevisiae and Zea mays, the Myb-related proteins Bas1p and C1, respectively, are thought to act in concert with a second DNA-binding protein to activate transcription of target genes; Bas1p cooperates with the homeobox protein Bas2p, and C1 is thought to dimerize with basic helix-loop-helix proteins such as B, R, SN, and LC (13, 18, 36, 61). In the cases of C1 and c-Myb, heterodimerization has been shown to require the Myb domains (18, 28). The in vivo activity of Cef1p clearly requires both R1 and R2. Perhaps, as in the case of C1, R1R2 of Cef1p functions as both a protein-protein interaction domain and a DBD. Evidence also exists for heterodimeric transcription factor complexes that require only one active DNA-binding subunit. For example, mutational studies have demonstrated that elimination of the sequence-specific DNA-binding activity of the yeast homeodomain protein -2 does not destroy its ability to bind DNA in concert with a1 (64). Additionally, c-Myb, in concert with HSF3, stimulates transcription from the hsp70 promoter through the heat shock element without apparently binding to this sequence (28, 29).

Although the precise function of Cdc5 family members is not understood presently, it is clear that the product of the cdc5⁺ gene in S. pombe and its homolog in S. cerevisiae are required during G₂/M. At present, all available evidence indicates a role for this protein family in DNA binding. However, it is quite possible that Cdc5 family members do not function as transcriptional activators. Indeed, there is precedence for Myb-related proteins which play roles in biological processes other than activation of transcription. For example, human TRF1 and TRF2 and fission yeast taz1p are Myb-related proteins that negatively regulate telomere length (5, 8, 11, 63) and the Myb-related budding yeast protein Reb1p functions in transcriptional termination as well as transcriptional activation of rRNA genes (30). An approach utilizing a combination of genetic, biochemical, and cytological analyses of the isolated Cdc5 family members should lead to an understanding of how Cdc5 facilitates the G₂/M transition in eukaryotic cells.