Mouse Cyclin-dependent Kinase (Cdk) 5 Is a Functional Homologue of a Yeast Cdk, Pho85 Kinase*

Mouse cyclin-dependent kinase (Cdk) 5 and yeast Pho85 kinase share similarities in structure as well as in the regulation of their activity. We found that mouse Cdk5 kinase produced in pho85Δ mutant cells could suppress some ofpho85Δ mutant phenotypes including failure to grow on nonfermentable carbon sources, morphological defects, and growth defect caused by Pho4 or Clb2 overproduction. We also demonstrated that Cdk5 coimmunoprecipitated with Pho85-cyclins including Pcl1, Pcl2, Pcl6, Pcl9, and Pho80, and that the immunocomplex could phosphorylate Pho4, a native substrate of Pho85 kinase. Thus mouse Cdk5 is a functional homologue of yeast Pho85 kinase.

Polymerase chain reaction Cloning and Construction of Plasmids-Cloning of PCL1, PCL2, CLB2, and PHO80 was described previously (19). DNA fragments encoding PCL6 or PCL9 were similarly cloned (19): the primers were synthesized to incorporate an NcoI site at the start codon of PCL6 (5Ј-AAATAGCGGCGCCATGGCTATCAAAGG) or an EcoRI site immediately downstream of the start codon of PCL9 (5Ј-CACAAAGAGATGAATTCTGACTACGAT) and a BglII site at the 3Ј-end of each ORF (5Ј-CATATTACGCATTTAGATCTGCCCGTAAC-TAG for PCL6 and 5Ј-GGCGAGTAACTTAAGATCTTTGCTTGAAAA-ACG for PCL9). These fragments were incorporated into pMF906 (19), together with a TRP1 marker, to produce HA-cyclins under the control of the GAL10 promoter. To disrupt genomic loci of these cyclins, a TRP1 fragment was used to replace an EcoRV-SalI fragment of PCL1, an SspI-SspI fragment of PCL2, a BamHI-XbaI fragment of PCL6, and an NcoI-EcoRV fragment of PCL9, and a LEU2 fragment was to replace a ClaI-XbaI fragment of PHO80. Successful disruption was confirmed by polymerase chain reaction. A cDNA clone encoding mouse Cdk5 kinase (26) and a URA3 marker were cloned into pMF906 plasmid to generate pMF1086. The cDNA fragment was also incorporated into pMF568 (URA3) to generate pMF1057 in which the kinase was produced under the control of the PYK1 promoter (27). Plasmid pMF1079 consists of a BamHI-SalI fragment containing the promoter and the ORF of PHO85 2 and a URA3 marker. To overproduce Pho4 and Clb2 proteins, DNA fragments encoding each protein were cloned into pMF906, together with a TRP1 or a LEU2 marker to generate pMF869 (GAL10-PHO4-TRP1), pMF1084 (GAL10-PHO4-LEU2), pMF922 (GAL10-CLB2-TRP1), and pMF1085 (GAL10-CLB2-LEU2).
Analysis of Mutant Phenotypes and Its Suppression-Utilization of nonfermentable carbon source, production of acid phosphatase, and accumulation of glycogen were assayed as described (11,12,28). Suppression of growth arrest of a cln1 cln2 pho85 triple mutant was tested with strain MFY151 as described (19). Morphological defects were microscopically observed with overnight culture, and cell number was counted with a hemocytometer. Suppression of growth defect caused by overproduction of Pho4 or Clb2 proteins was analyzed by streaking yeast transformants on SD and SGal media supplemented with nutrients but lacking uracil and leucine. For spotting cell suspension, over-* This work was supported by grant-in-aids for scientific research from Monbu-sho of Japan (to M. N., and A. T.) and by Research Grants for Life Sciences and Medicine from the Keio University Medical Science Fund (to M. N.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Immunoprecipitation and Kinase Assay-Preparation of cell extracts, immunoprecipitation, and kinase assay were as described (19). Briefly, after the induction of the production of HA-cyclin and Cdk5 in SGal medium at 30°C for 3 h, the extracts were prepared from cells suspended in lysis buffer by vortexing with glass beads (0.45-mm diameter). To 50 l of the extract containing ϳ200 g of protein, 1 l of anti-HA monoclonal antibody (HAmAb; clone 16B12, Berkeley Antibody Company) or 1 l of anti-GST antibody (GST Ab) was added, and the mixture was incubated on ice for 1 h. The immunocomplex was recovered by absorbing with Protein A-Sepharose 6B (Amersham Pharmacia Biotech) for 1 h at 4°C, washed three times with radioimmunoprecipitation assay buffer (lysis buffer supplemented with 150 mM NaCl) and twice with kinase assay buffer, and subjected to kinase assay using Pho4, Sic1, or histone H1 as substrate. Immunoblotting of Cdk5 and HA-cyclins were performed essentially as described (19) using anti-Cdk5 monoclonal antibody (a gift of G. Patrick and L.-H. Tsai) at 1:500 dilution and HAmAb at 1:1000 dilution, respectively.

RESULTS AND DISCUSSION
Suppression of pho85⌬ Mutant Phenotypes by Cdk5-Among pleiotropic pho85⌬ mutant phenotypes, we tested constitutive expression of acid phosphatase (11), growth arrest of a cln1 cln2 pho85 triple mutant (16,17), accumulation of glycogen (12,13), failure to grow on nonfermentable carbon sources (13,28), abnormal morphology (29), and growth arrest caused by Pho4 or Clb2 overproduction (this work). Overproduction of mouse Cdk5 kinase failed to suppress the first three phenotypes ( Fig.  1) but could suppress the other mutant phenotypes (Figs. 2 and 3). Cdk5 could restore growth of pho85⌬ cells on glycerol ϩ lactate medium ( Fig. 2A). Overproduction of Pho4 or Clb2 proteins directed by the GAL10 promoter caused a growth defect in the absence of Pho85 kinase, which was suppressed by overproduction of Cdk5 kinase (Fig. 2B). Overproduction of Cdk5 itself did not affect the growth of pho85⌬ cells (Fig. 2B). pho85⌬ mutant cells become large (29) and show an apparent defect in separation of daughter cells, resulting in multiple-budded cells (Fig. 3A). They did not separate after extensive sonication under which conditions the wild-type cells did. The ratio of cells with morphological defects to the total cells reached ϳ33% in pho85⌬ cells, which was decreased to 1.8% by expression of PHO85 and to 3.8% by that of CDK5 (Fig. 3A). Thus Cdk5 could function in yeast to suppress some of pho85⌬ mutant phenotypes.
Cdk5 Kinase Can Phosphorylate Pho4 in Vitro-Ten cyclinlike proteins are known to interact with Pho85, and among them Pho85 complexed with Pcl1, Pcl2, Pcl9, Pcl10, or Pho80 was shown to phosphorylate Pho4 in vitro (16 -18, 30, 31). These facts led us to a hypothesis that Cdk5 may associate with Pho85-cyclin(s) to phosphorylate Pho4, resulting in suppres-sion of the growth defect caused by Pho4 overproduction. To test this idea, we first immunoprecipitated Cdk5 from yeast extracts and assayed its kinase activity on Pho4, Sic1, and histone H1. As shown in Fig. 4A, immunoprecipitated Cdk5 could phosphorylate Pho4 and histone H1, with more efficient phosphorylation of the latter (lanes 1 and 3), but failed to phosphorylate Sic1 to a detectable level (lane 2). Pho85 kinase, on the other hand, could phosphorylate Pho4 and Sic1 but not histone H1 (lanes 6 -8). Control experiments (lanes 4 and 5) did not give a detectable level of Pho4 phosphorylation. These results demonstrate that Cdk5 kinase can phosphorylate Pho4 as well as histone H1 in vitro.
We next studied an interaction of Cdk5 with Pho85-cyclins by coimmunoprecipitation from cell extracts prepared from pho85⌬ cells overproducing Cdk5 and either HA-Pcl1, -Pcl2, -Pcl6, -Pcl9, or -Pho80. Any combination of the Pho85-cyclins and Cdk5 kinase could phosphorylate Pho4, and Pcl2-Cdk5 appeared most efficient (Fig. 4B, lane 10, top panel). The observation that Cdk5 was detected in the immunocomplexes with Pcl1, Pcl2, Pcl6, Pcl9, or Pho80 suggests that the kinase can interact with these Pho85-cyclins in yeast cells (Fig. 4B,  lanes 9 -13, bottom panel). Western blotting analysis of the cell extracts demonstrated that there were no significant differences in the amount of HA-cyclins in the extracts (Fig. 4C).
With respect to morphological defects of the double mutants, Cdk5 kinase could reduce the ratio of cells with abnormal shape, but not as efficiently as in pho85⌬ single mutant cells, and a deletion of PCL1 or PCL9 appeared to result in less efficient suppression than did that of the other cyclin genes (Fig. 3B). Expression of Pho85 kinase in the double mutants gave similar effect. These results suggest that cyclin(s) other than those tested may be required for Pho85 and Cdk5 to affect cell morphology or that cyclins interacting with either kinase to regulate cell morphology are redundant although individual cyclin may contribute to the suppression to a different extent.
Next we tested suppression of the growth defect caused by overproduction of Pho4 or Clb2 in the double mutants. As shown in Fig. 5, overproduction of these proteins caused the growth defect in the double mutants as well as in pho85⌬ single mutant, and even in the wild-type cells (Fig. 5, panels B1 and  B2). These single or double mutants could grow normally on galactose medium (panel B3). Overproduction of Cdk5 together with Pho4 could restore the growth of pho85⌬ and the wild-type cells (Fig. 5, panel C1). The suppression by Cdk5 appeared dependent on a specific Pho85-cyclin: the absence of Pcl1, Pcl6, Pcl9, or Pho80 did not appear to affect the suppression efficiency, whereas that of Pcl2 resulted in almost no suppression by Cdk5 kinase (panel C1), suggesting that the Pcl2-Cdk5 complex may be most crucial to overcome the growth defect. To the contrary, Pho85 kinase did not show a specific requirement of cyclin to counteract the effect of Pho4 overproduction: the double and pho85⌬ single mutants showed similar level of growth recovery by Pho85 (panel D1).
In the case of the growth defect caused by Clb2 overproduction, Cdk5 showed very weak suppression in the pho85⌬ pcl6⌬ mutant, whereas, in the other double mutants, it could suppress the defect as efficiently as in pho85⌬ single mutant (panel C2), suggesting that Cdk5 may be highly dependent on Pcl6 to counteract the overproduction effect of Clb2. This is in clear contrast to the function of Pho85 where PCL6 was dispensable (panel D2), whereas the other Pho85-cyclins tested might individually contribute to the Pho85 function to a certain extent (panel D2). Overproduction of Cdk5 or Pho85 alone did not affect the growth of strains tested (panels C3 and D3). Taken together, these results suggest that the suppression by Cdk5 of the growth defect caused by Pho4 or Clb2 was dependent on a specific Pho85-cyclin.
In this paper we demonstrated that mouse Cdk5 kinase could suppress some of pho85⌬ mutant phenotypes. This limited function could stem from defective interaction with some Pho85-cyclins and/or from failure to phosphorylate appropriate substrates to a sufficient level. We could detect, by coimmunoprecipitation, an interaction of Cdk5 with Pcl1, Pcl2, Pcl6, Pcl9, or Pho80 (Fig. 4B), but Cdk5 function in vivo appeared to be dependent on specific Pho85-cyclins, as observed in the suppression of the growth defect caused by Pho4 or Clb2 overproduction (Fig. 5). Thus it is formally possible that Cdk5 fails to interact with specific cyclin(s) in vivo, resulting in failure to suppress certain pho85⌬ mutant phenotypes. Alternatively, the affinity of Cdk5 for certain Pho85-cyclins may not be strong enough to form a complex of full activity, and/or complexes of Cdk5 and Pho85-cyclins may exhibit altered or loosened substrate specificity compared with that exerted by a combination of Pho85-cyclins and its native kinase. Either of these could result in weak kinase activity unable to phosphorylate appropriate substrates to a sufficient level. This idea can explain why Cdk5 failed to suppress growth arrest of a cln1 cln2 pho85 mutant (Fig. 1). We previously demonstrated that Pcl1-Pho85 can phosphorylate Sic1, targeting it to degradation, which is a major role of Pho85 kinase in the absence of Cln1, 2-Cdc28 kinase activity for cells to proceed through G 1 (19). Although Cdk5 could associate with Pcl1, it failed to phosphorylate Sic1 to a detectable level (Fig. 4, A and B). This idea can also be applied to the argument why Cdk5 that could phosphorylate Pho4 failed to suppress constitutive expression of PHO5 (Fig.  1). Phosphorylation of Pho4 regulates both its export from and import into the nucleus (32,33), and a recent report demonstrated that the phosphorylation status affects the efficiency of translocation of the transcription factor (34). We imagine that phosphorylation of Pho4 by Cdk5 may decrease the level of the transcription factor in the nucleus sufficient to inactivate transcription of genes responsible for the growth arrest but not to the level enough to repress PHO5 expression. In other words, Pho4-responsive genes may have a different threshold of the transcription factor in the nucleus.
Since the discovery that yeast Cdc28 kinase is a functional homologue of mammalian Cdc2 and other Cdks functioning in the cell cycle, the yeast system has been providing a convenient tool to study the regulation mechanism of the cell cycle. In this paper, we demonstrated that the mammalian Cdk family has another functional homologue of the yeast Cdks. This discovery will lead to further understanding of the function of Cdk5 and Pho85 kinases. The yeast system will provide a tool to identify yet unknown factors that associate with p35-Cdk5 to regulate events involved in neuronal developments. Conversely, Cdk5 kinase can be used to search for yeast proteins interacting with cyclin-Pho85 complex to regulate cell-cycle progression and cell morphology. These studies will provide more insights into regulatory mechanisms of Cdks functioning in the events both related and unrelated to the cell cycle.