Directed evolution to bypass cyclin requirements for the Cdc28p cyclin-dependent kinase.

To identify cyclin-dependent kinase mutants with relaxed cyclin requirements, CDC28 alleles were selected that could rescue a yeast strain expressing as its only CLN G1 cyclin a mutant Cln2p (K129A,E183A) that is defective for Cdc28p binding. Rescue of this strain by mutant CDC28 was dependent upon the mutant cln2-KAEA, but additional mutagenesis and DNA shuffling yielded multiply mutant CDC28-BYC alleles (bypass of CLNs) that could support highly efficient cell cycle initiation in the complete absence of CLN genes. By gel filtration chromatography, one of the mutant Cdc28 proteins exhibited kinase activity associated with cyclin-free monomer. Thus, the mutants' CLN bypass activity might result from constitutive, cyclin-independent activity, suggesting that Cdk targeting by cyclins is not required for cell cycle initiation.     

Mutations in CDC14 result in high sensitivity to cyclin gene dosage in Saccharomyces cerevisiae

We screened for mutations that resulted in lethality when the G1 cyclin Cln2p was overexpressed throughout the cell cycle in Saccharomyces cerevisiae. Mutations in five complementation groups were found to give this phenotype, and three of the mutated genes were identified as MEC1, NUP170, and CDC14. Mutations in CDC14 may have been recovered in the screen because Cdc14p may reduce the cyclin B (Clb)-associated Cdc28 kinase activity in late mitosis, and Cln2p may normally activate Clb-Cdc28 kinase activity by related mechanisms. In agreement with the idea that cdc14 mutations elevate Clb-Cdc28 kinase activity, deletion of the gene for the Clb-Cdc28 inhibitor Sic1 caused synthetic lethality with cdc14-1, as did the deletion of HCT1, which is required for proteolysis of Clb2p. Surprisingly, deletion of the gene for the major B-type cyclin, CLB2, also caused synthetic lethality with the cdc14-1 mutation. The clb2 cdc14 strains arrested with replicated but unseparated DNA and unseparated spindle pole bodies; this phenotype is distinct from the late mitotic arrest of the sic1::TRP1 cdc14-1 and the cdc14-1 hct1::LEU2 double mutants and of the cdc14 CLN2 overexpressor. We found genetic interactions between CDC14 and the replication initiator gene CDC6, extending previous observations of interactions between the late mitotic function of Cdc14p and control of DNA replication. We also describe genetic interactions between CDC28 and CDC14. Received: 24 May 1999 / Accepted: 19 October 1999     

Mutations in <Emphasis Type="Italic">CDC14</Emphasis> result in high sensitivity to cyclin gene dosage in <Emphasis Type="Italic">Saccharomyces cerevisiae</Emphasis>

We screened for mutations that resulted in lethality when the G1 cyclin Cln2p was overexpressed throughout the cell cycle in Saccharomyces cerevisiae. Mutations in five complementation groups were found to give this phenotype, and three of the mutated genes were identified as MEC1, NUP170, and CDC14. Mutations in CDC14 may have been recovered in the screen because Cdc14p may reduce the cyclin B (Clb)-associated Cdc28 kinase activity in late mitosis, and Cln2p may normally activate Clb-Cdc28 kinase activity by related mechanisms. In agreement with the idea that cdc14 mutations elevate Clb-Cdc28 kinase activity, deletion of the gene for the Clb-Cdc28 inhibitor Sic1 caused synthetic lethality with cdc14-1, as did the deletion of HCT1, which is required for proteolysis of Clb2p. Surprisingly, deletion of the gene for the major B-type cyclin, CLB2, also caused synthetic lethality with the cdc14-1 mutation. The clb2 cdc14 strains arrested with replicated but unseparated DNA and unseparated spindle pole bodies; this phenotype is distinct from the late mitotic arrest of the sic1::TRP1 cdc14-1 and the cdc14-1 hct1::LEU2 double mutants and of the cdc14 CLN2 overexpressor. We found genetic interactions between CDC14 and the replication initiator gene CDC6, extending previous observations of interactions between the late mitotic function of Cdc14p and control of DNA replication. We also describe genetic interactions between CDC28 and CDC14.     

Ubiquitination of the G1 cyclin Cln2p by a Cdc34p-dependent pathway.

Recombinant G1 cyclin Cln2p can bind to and stimulate the protein kinase activity of p34CDC28 (Cdc28p) in an extract derived from cyclin-depleted and G1-arrested Saccharomyces cerevisiae cells. Upon activating Cdc28p, Cln2p is extensively phosphorylated and conjugated with multiubiquitin chains. Ubiquitination of Cln2p in vitro requires the Cdc34p ubiquitin-conjugating enzyme, Cdc28p, protein phosphorylation and unidentified factors in yeast extract. Ubiquitination of Cln2p by Cdc34p contributes to the instability of Cln2p in vivo, as the rate of Cln2p degradation is reduced in cdc34ts cells. These results provide a molecular framework for G1 cyclin instability and suggest that a multicomponent, regulated pathway specifies the selective ubiquitination of G1 cyclins.     

The yeast Cln3 protein is an unstable activator of Cdc28.

The Cln3 cyclin homolog of Saccharomyces cerevisiae functions to promote cell cycle START for only a short time following its synthesis. Cln3 protein is highly unstable and is stabilized by C-terminal truncation. Cln3 binds to Cdc28, a protein kinase catalytic subunit essential for cell cycle START, and Cln3 instability requires Cdc28 activity. The long functional lifetime and the hyperactivity of C-terminally truncated Cln3 (Cln3-2) relative to those of full-length Cln3 are affected by mutations in CDC28: the functional lifetime of Cln3-2 is drastically reduced by the cdc28-13 mutation at the permissive temperature, and the cdc28-4 mutation at the permissive temperature completely blocks the function of Cln3-2 while only partially reducing the function of full-length Cln3. Thus, sequences in the C-terminal third of Cln3 might help stabilize functional Cdc28-Cln3 association, as well as decreasing the lifetime of the Cln3 protein. These and other results strongly support the idea that Cln proteins function to activate Cdc28 at START.     

Molecular Evolution Allows Bypass of the Requirement for Activation Loop Phosphorylation of the Cdc28 Cyclin-Dependent Kinase

Many protein kinases are regulated by phosphorylation in the activation loop, which is required for enzymatic activity. Glutamic acid can substitute for phosphothreonine in some proteins activated by phosphorylation, but this substitution (T169E) at the site of activation loop phosphorylation in the Saccharomyces cerevisiae cyclin-dependent kinase (Cdk) Cdc28p blocks biological function and protein kinase activity. Using cycles of error-prone DNA amplification followed by selection for successively higher levels of function, we identified mutant versions of Cdc28p-T169E with high biological activity. The enzymatic and biological activity of the mutant Cdc28p was essentially normally regulated by cyclin, and the mutants supported normal cell cycle progression and regulation. Therefore, it is not a requirement for control of the yeast cell cycle that Cdc28p be cyclically phosphorylated and dephosphorylated. These CDC28 mutants allow viability in the absence of Cak1p, the essential kinase that phosphorylates Cdc28p-T169, demonstrating that T169 phosphorylation is the only essential function of Cak1p. Some growth defects remain in suppressed cak1 cdc28 strains carrying the mutant CDC28 genes, consistent with additional nonessential roles for CAK1.     

Dephosphorylation of threonine 169 of Cdc28 is not required for exit from mitosis but may be necessary for start in Saccharomyces cerevisiae.

Entry into mitosis requires activation of cdc2 kinase brought on by its association with cyclin B, phosphorylation of the conserved threonine (Thr-167 in Schizosaccharomyces pombe) in the T loop, and dephosphorylation of the tyrosine residue at position 15. Exit from mitosis, on the other hand, is induced by inactivation of cdc2 activity via cyclin destruction. It has been suggested that in addition to cyclin degradation, dephosphorylation of Thr-167 may also be required for exit from the M phase. Here we show that Saccharomyces cerevisiae cells expressing cdc28-E169 (a CDC28 allele in which the equivalent threonine, Thr-169, has been replaced by glutamic acid) are able to degrade mitotic cyclin Clb2, inactivate the Cdc28/Clb2 kinase, and disassemble the anaphase spindles, suggesting that they exit mitosis normally. The cdc28-E169 allele is active with respect to its mitotic functions, since it complements the mitosis-defective cdc28-1N allele. Whereas replacement of Thr-169 with serine affects neither Start nor the mitotic activity of Cdc28, replacement with glutamic acid or alanine renders Cdc28 inactive for Start-related functions. Coimmunoprecipitation experiments show that although Cdc28-E169 associates with mitotic cyclin Clb2, it fails to associate with the G1 cyclin Cln2. Thus, an unmodified threonine at position 169 in Cdc28 is important for interaction with G1 cyclins. We propose that in S. cerevisiae, dephosphorylation of Thr-169 is not required for exit from mitosis but may be necessary for commitment to the subsequent division cycle.     

Properties of Saccharomyces cerevisiae wee1 and its differential regulation of p34CDC28 in response to G1 and G2 cyclins.

Wee1 is a protein kinase that negatively regulates p34cdc2 kinase activity. We have identified a Saccharomyces cerevisiae wee1 homolog encoded by the SWE1 gene. SWE1 overexpression arrests cells in G2 with short spindles whereas deletion of SWE1 did not alter the cell cycle but did eliminate the G2 delay observed in mih1- mutants. Swe1 immunoprecipitates were capable of tyrosine phosphorylating and inactivating p34CDC28 complexed with Clb2, a G2-type cyclin, but not p34CDC28 complexed with Cln2, a G1-type cyclin, consistent with the inability of Swe1 overexpression to inhibit the G1/S transition. These results suggest that specific cyclin subunits target p34CDC28 for distinct regulatory controls which may be important for ensuring proper p34CDC28 function during the cell cycle.     

p34Cdc28-mediated control of Cln3 cyclin degradation.

Cln3 cyclin of the budding yeast Saccharomyces cerevisiae is a key regulator of Start, a cell cycle event in G1 phase at which cells become committed to division. The time of Start is sensitive to Cln3 levels, which in turn depend on the balance between synthesis and rapid degradation. Here we report that the breakdown of Cln3 is ubiquitin dependent and involves the ubiquitin-conjugating enzyme Cdc34 (Ubc3). The C-terminal tail of Cln3 functions as a transferable signal for degradation. Sequences important for Cln3 degradation are spread throughout the tail and consist largely of PEST elements, which have been previously suggested to target certain proteins for rapid turnover. The Cln3 tail also appears to contain multiple phosphorylation sites, and both phosphorylation and degradation of Cln3 are deficient in a cdc28ts mutant at the nonpermissive temperature. A point mutation at Ser-468, which lies within a Cdc28 kinase consensus site, causes approximately fivefold stabilization of a Cln3-beta-galactosidase fusion protein that contains a portion of the Cln3 tail and strongly reduces the phosphorylation of this protein. These data indicate that the degradation of Cln3 involves CDC28-dependent phosphorylation events.     

Genetic analysis of Cln/Cdc28 regulation of cell morphogenesis in budding yeast.

The CLN1, CLN2 and CLN3 gene family of G1-acting cyclin homologs of Saccharomyces cerevisiae is functionally redundant: any one of the three Cln proteins is sufficient for activation of Cdc28p protein kinase activity for cell cycle START. The START event leads to multiple processes (including DNA replication and bud emergence); how Cln/Cdc28 activity activates these processes remains unclear. CLN3 is substantially different in structure and regulation from CLN1 and CLN2, so its functional redundancy with CLN1 and CLN2 is also poorly understood. We have isolated mutations that alter this redundancy, making CLN3 insufficient for cell viability in the absence of CLN1 and CLN2 expression. Mutations causing phenotypes specific for the cell division cycle were analyzed in detail. Mutations in one gene result in complete failure of bud formation, leading to depolarized cell growth. This gene was identified as BUD2, previously described as a non-essential gene required for proper bud site selection but not required for budding and viability. Bud2p is probably the GTPase-activating protein for Rsr1p/Bud1p [Park, H., Chant, I. and Herskowitz, I. (1993) Nature, 365, 269-274]; we find that Rsr1p is required for the bud2 lethal phenotype. Mutations in two other genes (ERC10 and ERC19) result in a different morphogenetic defect: failure of cytokinesis resulting in the formation of long multinucleate tubes. These results suggest direct regulation of diverse aspects of bud morphogenesis by Cln/Cdc28p activity.     

A phosphorylation-independent role for the yeast cyclin-dependent kinase activating kinase Cak1

Cdc28 is the main cyclin-dependent kinase (CDK) directing the cell cycle in the budding yeast Saccharomyces cerevisiae. Besides cyclin binding, Cdc28 requires phosphorylation by the Cak1 kinase to achieve full activity. We have previously isolated carboxy-terminal cdc28^CST mutants that are temperature sensitive and exhibit high chromosome instability. Both phenotypes are suppressed by high copy Cak1 in a manner that is independent of its catalytic activity and conversely, combination of cdc28^CST and cak1 mutations results in synthetic lethality. Altogether, these results suggest that for the Cdc28 complexes to remain stable and active, an interaction with Cak1 is needed via the carboxyl terminus of Cdc28. We report two-hybrid assay data that support this model, and results that indicate that actively growing yeast cells require an optimum Cdc28:Cak1 ratio. While Cak1 is constitutively active and expressed, dividing cells tightly regulate Cak1 protein levels to ensure presence of adequate levels of Cdc28 CDK activity.     

Protein phosphatase 2A regulates MPF activity and sister chromatid cohesion in budding yeast

Background Mitosis is regulated by MPF (maturation promoting factor), the active form of Cdc2/28–cyclin B complexes. Increasing levels of cyclin B abundance and the loss of inhibitory phosphates from Cdc2/28 drives cells into mitosis, whereas cyclin B destruction inactivates MPF and drives cells out of mitosis. Cells with defective spindles are arrested in mitosis by the spindle-assembly checkpoint, which prevents the destruction of mitotic cyclins and the inactivation of MPF. We have investigated the relationship between the spindle-assembly checkpoint, cyclin destruction, inhibitory phosphorylation of Cdc2/28, and exit from mitosis.Results The previously characterized budding yeast mad mutants lack the spindle-assembly checkpoint. Spindle depolymerization does not arrest them in mitosis because they cannot stabilize cyclin B. In contrast, a newly isolated mutant in the budding yeast CDC55 gene, which encodes a protein phosphatase 2A (PP2A) regulatory subunit, shows a different checkpoint defect. In the presence of a defective spindle, these cells separate their sister chromatids and leave mitosis without inducing cyclin B destruction. Despite the persistence of B-type cyclins, cdc55 mutant cells inactivate MPF. Two experiments show that this inactivation is due to inhibitory phosphorylation on Cdc28: phosphotyrosine accumulates on Cdc28 in cdc55Δ cells whose spindles have been depolymerized, and a cdc28 mutant that lacks inhibitory phosphorylation sites on Cdc28 allows spindle defects to arrest cdc55 mutants in mitosis with active MPF and unseparated sister chromatids.Conclusions We conclude that perturbations of protein phosphatase activity allow MPF to be inactivated by inhibitory phosphorylation instead of by cyclin destruction. Under these conditions, sister chromatid separation appears to be regulated by MPF activity rather than by protein degradation. We discuss the role of PP2A and Cdc28 phosphorylation in cell-cycle control, and the possibility that the novel mitotic exit pathway plays a role in adaptation to prolonged activation of the spindle-assembly checkpoint.     

Enhanced Cell Polarity in Mutants of the Budding Yeast Cyclin-dependent Kinase Cdc28p

The yeast cyclin-dependent kinase Cdc28p regulates bud morphogenesis and cell cycle progression via the antagonistic activities of Cln and Clb cyclins. Cln G1 cyclins direct polarized growth and bud emergence, whereas Clb G2 cyclins promote isotropic growth of the bud and chromosome segregation. Using colony morphology as a screen to dissect regulation of polarity by Cdc28p, we identified nine point mutations that block the apical-isotropic switch while maintaining other functions. Like a clb2 Delta mutation, each confers tubular bud shape, apically polarized actin distribution, unipolar budding, and delayed anaphase. The mutations are all suppressed by CLB2 overexpression and are synthetically lethal with a CLB2 deletion. However, defects in multiple independent pathways may underlie their common phenotype, because the mutations are scattered throughout the CDC28 sequence, complement each other, and confer diverse biochemical properties. Glu12Gly, a mutation that alters a residue involved in Swe1p inhibition of Cdc28p, was unique in being suppressed by deficiency of SWE1 or CLN1. With wild-type CDC28, filament formation induced by CLN1 overexpression was markedly decreased in a SWE1 deletion. These results suggest that Swe1p, via inhibition of Clb2p/Cdc28p, may mediate much of the effect of Cln1p on filamentous morphogenesis.     

PP2ACdc55 regulates G1 cyclin stability

Maintaining accurate progression through the cell cycle requires the proper temporal expression and regulation of cyclins. The mammalian D-type cyclins promote G₁-S transition. D1 cyclin protein stability is regulated through its ubiquitylation and resulting proteolysis catalyzed by the SCF E3 ubiquitin ligase complex containing the F-box protein, Fbx4. SCF E3-ligase-dependent ubiquitylation of D1 is trigged by an increase in the phosphorylation status of the cyclin. As inhibition of ubiquitin-dependent D1 degradation is seen in many human cancers, we set out to uncover how D-type cyclin phosphorylation is regulated. Here we show that in S. cerevisiae, a heterotrimeric protein phosphatase 2A (PP2A^Cdc55) containing the mammalian PPP2R2/PR55 B subunit ortholog Cdc55 regulates the stability of the G₁ cyclin Cln2 by directly regulating its phosphorylation state. Cells lacking Cdc55 contain drastically reduced Cln2 levels caused by degradation due to cdk-dependent hyperphosphorylation, as a Cln2 mutant unable to be phosphorylated by the yeast cdk Cdc28 is highly stable in cdc55-null cells. Moreover, cdc55-null cells become inviable when the SCF^Grr1 activity known to regulate Cln2 levels is eliminated or when Cln2 is overexpressed, indicating a critical relationship between SCF and PP2A functions in regulating cell cycle progression through modulation of G₁-S cyclin degradation/stability. In sum, our results indicate that PP2A is absolutely required to maintain G₁-S cyclin levels through modulating their phosphorylation status, an event necessary to properly transit through the cell cycle.