Casein kinase II is required for cell cycle progression during G1 and G2/M in Saccharomyces cerevisiae.

The catalytic subunit of Saccharomyces cerevisiae casein kinase II (Sc CKII) is encoded by the CKA1 and CKA2 genes, which together are essential for viability. Five independent temperature-sensitive alleles of the CKA2 gene were isolated and used to analyze the function of CKII during the cell cycle. Following a shift to the nonpermissive temperature, cka2ts strains arrested within a single cell cycle and exhibited a dual arrest phenotype consisting of 50% unbudded and 50% large-budded cells. The unbudded half of the arrested population contained a single nucleus and a single focus of microtubule staining, consistent with arrest in G1. Most of the large-budded fraction contained segregated chromatin and an extended spindle, indicative of arrest in anaphase, though a fraction contained an undivided nucleus with a short thick intranuclear spindle, indicative of arrest in G2 and/or metaphase. Flow cytometry of pheromone-synchronized cells confirmed that CKII is required in G1, at a point which must lie at or beyond Start but prior to DNA synthesis. Similar analysis of hydroxyurea-synchronized cells indicated that CKII is not required for completion of previously initiated DNA replication but confirmed that the enzyme is again required for cell cycle progression in G2 and/or mitosis. These results establish a role for CKII in regulation and/or execution of the eukaryotic cell cycle.

The catalytic subunit of Saccharomyces cerevisiae casein kinase II (Sc CKII) is encoded by the CKA1 and CKA2 genes, which together are essential for viability. Five independent temperature-sensitive alleles of the CKA2 gene were isolated and used to analyze the function of CKII during the cell cycle. Following a shift to the nonpermissive temperature, cka2 ts strains arrested within a single cell cycle and exhibited a dual arrest phenotype consisting of 50% unbudded and 50% largebudded cells. The unbudded half of the arrested population contained a single nucleus and a single focus of microtubule staining, consistent with arrest in G 1 . Most of the large-budded fraction contained segregated chromatin and an extended spindle, indicative of arrest in anaphase, though a fraction contained an undivided nucleus with a short thick intranuclear spindle, indicative of arrest in G 2 and/or metaphase. Flow cytometry of pheromone-synchronized cells confirmed that CKII is required in G 1 , at a point which must lie at or beyond Start but prior to DNA synthesis. Similar analysis of hydroxyurea-synchronized cells indicated that CKII is not required for completion of previously initiated DNA replication but confirmed that the enzyme is again required for cell cycle progression in G 2 and/or mitosis. These results establish a role for CKII in regulation and/or execution of the eukaryotic cell cycle.
Casein kinase II (CKII) 1 is a serine/threonine protein kinase which is ubiquitous among eukaryotic organisms (for review, see Issinger, 1993;Pinna, 1990;Tuazon and Traugh, 1991). The enzyme is composed of a catalytic ␣ and regulatory ␤ subunit that combine to form a native ␣ 2 ␤ 2 holoenzyme which is constitutively active in vitro. How (and indeed whether) the enzyme is regulated in vivo is unknown, though regulation via allosteric effectors (e.g. polyamines), covalent modification, cellular redistribution, and substrate-directed effects have all been proposed. CKII recognizes a Ser of Thr residue followed by a series of acidic residues and phosphorylates a broad and intriguing spectrum of both nuclear and cytoplasmic substrates.
Although the physiological role of CKII is not known, several lines of evidence suggest a role for the enzyme in cell prolifer-ation. First, CKII activity is elevated in rapidly dividing cells, both normal and transformed (reviewed in Issinger, 1993). CKII activity has also been reported to increase in response to stimulation of cells by mitogenic stimuli, though these effects have been difficult to reproduce (see Litchfield et al., 1994, for discussion). Second, CKII phosphorylates a number of proteins known to play crucial roles in cell proliferation (Meisner and Czech, 1991), including the nuclear oncogene proteins c-Myc, c-Myb, and c-Jun, the tumor suppressor protein p53, and the cyclin-dependent protein kinase p34 cdc2 (Russo et al., 1992). Cyclin-dependent protein kinases (CDKs) constitute the engine of the eukaryotic cell cycle and are essential for Start as well as the G 1 /S and G 2 /M transitions (Morgan, 1995). Third, CKII activity has been found to fluctuate in a cell cycle-dependent manner in mammalian cells (Carroll and Marshak, 1989;De-Benedette and Snow, 1991). Fourth and most importantly, experimental manipulation of CKII activity in vivo can have profound effects on cell proliferation. Disruption of CKII is lethal in Saccharomyces cerevisiae (Padmanabha et al., 1990) and probably also in Dictyostelium discoideum (Kikkawa et al., 1992) and Schizosaccharomyces pombe (Snell and Nurse, 1994). In mammalian cells, inhibition of CKII activity by microinjection of antibodies directed against the ␤ subunit into G 0 -synchronized human fibroblasts has been shown to inhibit cell cycle progression in response to serum stimulation (Pepperkok et al., 1994). Microinjection of nanomolar quantities of active CKII into Xenopus oocytes accelerates meiotic maturation induced by mitosis promoting actor (Mulner-Lorillon et al., 1988). Perhaps most remarkably, the catalytic subunit of CKII functions as an oncogene when expressed as a transgene in lymphocytes of transgenic mice, leading to the stochastic production of lymphomas when expressed alone and to the production of acute lymphocytic leukemia when co-expressed with a c-Myc transgene (Seldin and Leder, 1995).
We have exploited molecular and genetic methods available in the budding yeast S. cerevisiae (Sc) to explore the physiological role of CKII in this organism. Sc CKII consists of two catalytic subunits, ␣ and ␣Ј, and two regulatory subunits, ␤ and ␤Ј, which are encoded by the CKA1, CKA2, CKB1, and CKB2 genes, respectively (for review, see Glover et al., 1994). Deletion of either catalytic subunit gene alone has few if any phenotypic consequences, but simultaneous disruption of both CKA1 and 2 is lethal (Padmanabha et al., 1990). Cells arrested by depletion of CKII arrest as a population of unbudded and budded cells, with a significant proportion of the latter exhibiting an elongated bud characteristic of several cell division cycle mutants (Hartwell et al., 1973). Arrested cells continue to grow, suggesting that metabolism continues in the absence of cell division.
To further explore the role of CKII during growth and division, S. cerevisiae mutant strains temperature-sensitive for CKII have been constructed. In contrast to previously constructed strains that deplete kinase activity by attrition (Padmanabha et al., 1990), these mutants allow CKII activity to be inactivated within a time short with respect to the length of the cell cycle. Analysis of these temperature-sensitive strains reveals that CKII is required for cell cycle progression at two points in the cell cycle, one in G 1 and in the other in G 2 and/or mitosis. Furthermore, the data indicate that CKII activity is not required for ongoing DNA replication during S phase. Strains-S. cerevisiae strains used in this study are listed in Table I. Yeast strains were grown in rich glucose medium (YPD: 1% yeast extract, 2% peptone, and 2% glucose) or in supplemented minimal medium (SMM) (Rose et al., 1990). Temperature-sensitive strains were incubated at 25°C (permissive temperature) or either 37, 38, or 38.5°C (restrictive temperature). Escherichia coli strain DH5␣ (Clontech) was grown in Luria broth containing 50 g/ml ampicillin as necessary (Ausubel et al., 1987).

Materials-Chemicals
Isolation of Temperature-sensitive Mutants-A LEU2-marked plasmid carrying the wild-type CKA2 gene was constructed by subcloning the CKA2-containing HindIII/BamHI fragment from pRS316 -1 (Padmanabha et al., 1990) into HindIII/BamHI cut pRS315 (Sikorski and Hieter, 1989). The resulting plasmid, named pDH6, was subjected to hydroxylamine mutagenesis as described (Busby et al., 1982). Progress of the mutagenesis with time was monitored by the drop in transformation frequency of DH5␣, a 90% reduction in frequency being considered optimum. Strain RPG41-1a (Table I) was transformed with the mutagenized plasmid by the LiCl method (Rose et al., 1990) and plated on SMM lacking leucine at 25°C. Two thousand transformants were patched on SMM lacking leucine, grown at 25°C, and then replicaplated in duplicate to SMM containing 5-FOA plus uracil in order to select against the URA3-marked plasmid. One plate was incubated at the permissive temperature (25°C) and the other at the restrictive temperature (37°C). After 6 days of growth, temperature-sensitive strains were colony purified at 25°C on SMM lacking leucine, and uracil auxotrophy was confirmed by patching on SMM with and without uracil. A total of seven independent temperature-sensitive alleles were isolated in this manner, five of which could be recovered in DH5␣. To confirm linkage of temperature sensitivity with the plasmid, the five plasmids were retransformed into RPG41-1a. Ten independent transformants per plasmid were isolated and tested for the ability to form colonies at 37°C both before and after eviction of the URA3-marked plasmid as described above. In each case, all 10 transformants were able to form colonies at 37°C prior to eviction of the wild-type CKA2 gene but not afterward. These retransformed strains (cured of the URA3-marked plasmid) were used for all the experiments in this work and were named YDH7, YDH8, YDH11, YDH12, and YDH13. Each carries a derivative of plasmid pDH6 (i.e. strain YDH7 carries plasmid pDH7, and so on).
Nucleotide Sequencing-The complete protein coding region of the temperature-sensitive cka2 allele present in each of the five plasmids, pDH7, 8, 11, 12, and 13, was sequenced by the dideoxy chain termination method using five antisense oligodeoxynucleotide primers (5Ј-CAATACATCTACCGCTG-3Ј, GGAAGTTTGAATGTAGG, CTTGTGGT-AACGCGAAG, ATTCAGATGGTAAGTGC, and AGAATTGCCTTGC-TAAG). Primers were synthesized at the Molecular Genetics Instrumentation Facility, University of Georgia, and double-stranded DNA sequencing was carried out using [␣-35 S]dATP and the Sequenase kit (United States Biochemical Corp.) as described by the manufacturer.
Cell Growth and Synchronization-For growth studies, an overnight culture grown in YPD at 25°C was diluted into fresh YPD at a starting density of 2-8 ϫ 10 5 cells/ml and grown at 25°C with vigorous shaking. Shifts to restrictive temperature were carried out in log phase. For determining cell culture densities and budding profiles, aliquots of the culture were removed at the indicated times and fixed by the addition of 37% formaldehyde to a final concentration of 3.7%. The fixed samples were diluted as necessary, introduced into a hemacytometer, and examined at ϫ400 with an Olympus BH-2 phase contrast microscope. For determining cell viability, total cell number was determined as described, and viable cells were determined by plating appropriate dilutions of the culture on YPD plates at 25°C. Percent viability was calculated as 100 ϫ (number of colonies)/(total cell number).
For stationary phase G 1 synchronization, cultures were inoculated into YPD at 8 ϫ 10 5 cells/ml and grown at 25°C until stationary phase (approximately 60 h; 90 -95% unbudded cells). Cells were released from stationary phase arrest by dilution to 8 ϫ 10 5 cells/ml in fresh YPD at 25°C or the appropriate nonpermissive temperature. For arrest in G 1 in response to mating pheromone or in early S phase in response to hydroxyurea, cells were inoculated into YPD at 2-8 ϫ 10 5 cells/ml, grown at 25°C for 6 h (log phase), and then treated either with ␣-factor at 2.5 g/ml for 4.5 h or with hydroxyurea (added directly to the culture as a powder) at 0.1 M for 5 h. Synchronized cultures were released by filtering through a 2.5-cm HA 0.45-m filter (Millipore) mounted in a Swinnex 25 filter apparatus, followed by washing and resuspending the cells at the original density in fresh YPD at 25°C or the appropriate nonpermissive temperature.
Fluorescence Microscopy-For visualizing nuclear and spindle morphology, a sample of culture containing approximately 8 ϫ 10 6 cells was fixed by the addition of 37% formaldehyde to a final concentration of 3.7% followed by incubation at 23°C for 3-24 h. Fixed cells were collected by centrifugation, washed twice with sterile water, and resuspended in 1 ml of Buffer A (0.1 M K 2 HPO 4 , pH 7.0, 1.2 M sorbitol). Fifty l of Zymolyase 60,000 (ICN) at 4 units/ml in Buffer A was added, and the cells were incubated at 37°C for 1 h. Cells were washed twice in Buffer A and resuspended in 0.2 ml of Buffer A. Ten l of this suspension was spotted into a polylysine-coated well (polylysine 300K, 0.1% in H 2 O) of an eight-well Teflon-coated slide (Carlson Scientific Inc.) and incubated for 10 min. The solution was aspirated and the adherent cells washed once with PBS (0.1 M K 2 HPO 4 , pH 7.0, 0.15 M NaCl, 0.1% NaN 3 ), twice with 0.1% Nonidet P-40 in PBS, then again with PBS (5 min each). Ten l of primary antibody (monoclonal anti-tubulin antibody YOL 1/34, ascites fraction, Accurate Scientific, diluted 1:100 in PBS with 0.1% bovine serum albumin) was spotted into each well, and the slide was incubated at 25°C for 1-2 h in a humidified chamber. The primary antibody was aspirated and the cells washed as described above. Ten l of secondary antibody (Amersham fluorescein isothiocyanate-conjugated goat anti-rabbit IgG, diluted 1:25 in PBS with 0.1% bovine serum albumin) was spotted into each well and incubated at 25°C for 1-2 h in a humidified chamber. The secondary antibody was aspirated and the cells washed as described above. The wells were mounted in 90% glycerol, 0.02% p-phenylenediamine, 0.2 g/ml 4Ј,6Јdiamidino-2-phenylindole (DAPI) in PBS, and coverslips were sealed with nail polish. Cells were observed using a Zeiss IM 35 epifluorescence microscope equipped with a ϫ100 objective and either DAPI or fluorescein filters. Photographs were made with Kodak Technical Pan film.
RNA and Protein Synthesis-The following procedures were adapted from Elliott and McLaughlin (1978). For RNA labeling, cells were grown in SMM to 1-2 ϫ 10 7 cells/ml at 25°C and then either maintained for an additional 4 h at 25°C or shifted to 38°C. For each time point, a 4-ml aliquot was removed and analyzed. Cells were collected by filtration through an HA membrane as described above, resuspended in SMM containing one-tenth the normal amount of uracil, and pulselabeled with [ 14 C]uracil at 0.25 Ci/ml for 10 min. Incorporation was stopped by adding an equal volume of 20% trichloroacetic acid. The precipitate was collected on a Whatman GF/A filter and washed with 25 ml of 5% trichloroacetic acid followed by 25 ml of 95% ethanol. Filters were air dried under a heat lamp and counted in a Beckman LS6800 scintillation counter. Protein labeling followed the same procedure except that cells were grown to 2-4 ϫ 10 6 cells/ml and were pulse-labeled in the medium in which they were grown. For each time point, a 2-ml aliquot was removed, and [ 35 S]methionine was added to a final concentration of 0.2 Ci/ml for 10 min. The sample was then treated as above to determine total [ 35 S]methionine incorporation. As a positive control for inhibition of protein synthesis, cycloheximide (1 mg/ml stock in H 2 O) was added to a final concentration of 10 g/ml immediately after the first (zero) time point was taken. In both labeling experiments, incorporation was normalized to cell number, determined with a hemacytometer after fixation of an appropriate dilution in 3.7% formaldehyde.

RESULTS
Construction of Temperature-sensitive Strains-Strains temperature sensitive for CKII activity were constructed by plasmid shuffling (Boeke et al., 1987). Because of the functional redundancy of the CKA1 and 2 genes (Padmanabha et al., 1990), either could have been selected for mutagenesis, and the CKA2 gene was chosen arbitrarily. A LEU2 CEN/ARS plasmid bearing the wild-type CKA2 gene was mutagenized in vitro with hydroxylamine and then transformed into RPG41-1a, a cka1 cka2 strain rescued by the wild-type CKA2 gene on a URA3 CEN/ARS plasmid (Padmanabha et al., 1990). Twothousand independent transformants were plated in duplicate on minimal medium containing 5-FOA (to select against the URA3-marked plasmid) and incubated at both 25 and 37°C. Seven independent transformants that exhibited temperaturesensitive growth were identified and colony-purified. The LEU2-marked plasmids from five of these transformants were recovered in E. coli. Linkage between the mutagenized plasmid and the temperature-sensitive phenotype was confirmed by retransformation of RPG41-1a followed by plating on 5-FOA. None of the five strains was temperature sensitive prior to eviction of the CKA2 gene, confirming that all five alleles are recessive.
The nucleotide substitution(s) and corresponding amino acid replacement(s) present in each temperature-sensitive cka2 allele were determined by sequencing the protein coding region. As expected for hydroxylamine, which deaminates cytosines, all mutations could be explained by C 3 T transitions (see Table II). One allele (cka2-13) contained a single substitution while the remaining four (cka2-7, -8, -11, and -12) contained two, and every substitution resulted in an amino acid replacement. The temperature sensitivity of the single mutant (cka2-13) is presumably explained by the sole amino acid replacement in this mutant (D225N). The affected residue, which corresponds to D220 of cAMPdPK (Table II), is invariant in the protein kinase family (Hanks and Quinn, 1991) and is postulated to function in stabilizing the catalytic loop (Knighton et al., 1991). The D225N replacement also occurs as one of the two mutations present in the cka2-11 allele. The A190T and A190V replacements of cka2-7 and cka2-12, respectively, affect a residue which is moderately variable among protein kinases generally but is invariant among known CKIIs. Replacements in the other two alleles do not affect strongly conserved residues either in protein kinases generally or in CKII (Hanks and Quinn, 1991).
The maximum permissive and minimum restrictive temperatures of the five mutant strains are shown in Table III. The lower temperature transition of YDH11 relative to YDH13 implies that the E299K replacement in the cka2-11 allele is also a destabilizing mutation, at least in the context of the D225N mutation.
All five strains exhibited some phenotypic defect at permissive temperature. As shown in Table III, all five strains exhibited some increase in flocculation. This effect was severe in those strains showing the greatest temperature sensitivity, most notably YDH12, which grew essentially as a pellet at 25°C in liquid medium. Strains displaying the greatest temperature sensitivity also exhibited a modest slow growth phenotype at permissive temperature, as assessed by colony size on plates (data not shown). Because flocculation and slow growth are both characteristic of CKII depletion (Padmanabha et al., 1990), these results imply a reduction in CKII activity in these mutants at permissive temperature. Based on the above results, YDH8 was selected for the bulk of the studies described below because it has the lowest transition temperature consistent with a near normal growth rate (Fig. 1A) and a tractable level of flocculation (Table III) at permissive temperature.
Terminal Phenotype at the Restrictive Temperature-When shifted from 25 to 37°C during log phase growth, YDH8 exhibited arrest within one cell doubling, consistent with arrest during the first cell cycle following the shift (Fig. 1B). In contrast, the wild-type control strain YDH6 exhibited only a transient heat shock-induced depression in growth rate and then grew to saturation normally. As anticipated for cells depleted of CKII activity (Padmanabha et al., 1990), YDH8 exhibited complete flocculation within 1-2 h following the shift to restrictive temperature (data not shown). Similar behavior was observed with the other four temperature-sensitive alleles.
Budding profiles for the experiment shown in Fig. 1B are a Substitutions are shown in the context of the affected codon. All substitutions are explainable on the basis of a C to T transition (on either the coding or noncoding strand), as expected for hydroxylamine mutagenesis.
b Residues in the Cka2 polypeptide are numbered as in Padmanabha et al. (1990), i.e., relative to the initiating methionine as residue 1.
c Residues in cAMPdPK are numbered as in Fig. 1 of Hanks et al. (1988). d E299K and H294Y lie within an insertion between subdomains X and XI (Hanks and Quinn, 1991). A comparable insertion is not present in cAMPdPK.
presented in Table IV. Both YDH6 and YDH8 exhibited a transient, heat shock-induced increase in the proportion of unbudded cells during the first 1-2 h following the shift to restrictive temperature. YDH8 cells reached a stable state (terminal phenotype) by 4 -6 h after the shift. The arrested population consisted of approximately equal numbers of unbudded and large-budded cells, with small-to-medium-budded cells being present at very low levels. In contrast, YDH6 gradually returned to the roughly equal mixture of unbudded and smallto-medium-budded cells characteristic of log phase growth. Flow cytometry of YDH8 cells 4 -6 h after the shift indicated a mixture of cells containing 1 and 2 N DNA complement (data not shown). These results imply the existence of at least two distinct arrest points in the mutant. Because YDH8 cells do not arrest with a single terminal morphology, cka2-8 does not qualify as a classical cell division cycle mutation (Hartwell et al., 1973).
To determine whether the dual arrest phenotype was specific to YDH8, the budding profile of two other temperature-sensitive strains was examined at the restrictive temperature. Strains YDH11 and 13, whose cka2 alleles bear replacement(s) distinct from those of cka2-8 (Table II), also arrested as a 50:50 mixture of unbudded and large-budded cells (data not shown). This outcome indicated that the dual arrest phenotype is not an allele-specific response.
To better define the points of arrest, YDH8 cells incubated at the restrictive temperature for 5 h were double-stained with the DNA-binding dye DAPI (to visualize nuclear morphology) and an antitubulin monoclonal antibody (to visualize the tubulin cytoskelton, including the spindle). Stained cells were analyzed by immunofluorescence microscopy (Fig. 2B). The unbudded half of the arrested population uniformly displayed a single mass of DAPI-stainable material, indicative of an undivided nucleus. The majority of these cells contained an array of cytoplasmic microtubules radiating from a single focus, a morphology typical of cells arrested in G 1 . The large-budded half of the arrested population was heterogeneous. The majority of these cells (approximately two-thirds) contained two lobes of DAPI-stainable material and an elongated spindle. Cells in which the two lobes remained joined through the bud aperture and cells in which the two nuclei appeared to be fully separated were both observed, at approximately equal frequency. Both phenotypes are indicative of arrest in anaphase (Surana et al., 1993). The remaining third of budded cells contained a single, round nucleus traversed by a short, thick intranuclear spindle (not shown). The latter phenotype is characteristic of arrest in G 2 or metaphase (Irniger et al., 1995;Surana et al., 1991). At the level of resolution achieved by immunocytochemistry, the morphology of the spindle did not appear to be abnormal in any of these arrested cells. The mutant grown at 25°C ( Fig. 2A) and the wild-type strain grown at either 25 or 37°C (not shown) exhibited the expected array of nuclear and cytoskeletal morphologies typical of logarithmically growing cells.
The results obtained with asynchronous cultures suggested that CKII is required for cell cycle progression in G 1 as well as in G 2 and/or M. In order to probe these arrest points independently, we analyzed cell cycle progression of synchronized mutant and wild-type cultures. Cells were synchronized in G 1 by two different protocols, nutrient limitation and exposure to mating pheromone, and at the G 1 /S boundary by treatment with hydroxyurea.
CKII Is Required for Cell Cycle Progression in G 1 -S. cerevisiae respond to nutrient limitation by arresting cell division (and growth) early in the G 1 phase of the cell cycle. In order to synchronize cells at this point in the cycle, YDH6 and YDH8 were grown to late stationary phase in YPD at 25°C (90 -95% unbudded cells). The synchronized cells were then released into fresh medium prewarmed to 37°C (Fig. 3). In contrast to the control, which grew to saturation normally at this temperature, YDH8 exhibited no increase in cell number, clearly indicating arrest in the first cell cycle following release (Fig. 3A). Moreover, analysis of the budding profile indicated a marked defect in G 1 progression (Fig. 3B). Relative to the control, which budded synchronously in the first few hours after release, YDH8 cells budded with slower kinetics, and at least 80% of the initially unbudded population failed to bud at all. The small percentage of cells which did escape G 1 arrest subsequently arrested as large-budded cells (not shown). a Maximum permissive temperature is the highest temperature which allows essentially normal colony formation on YPD plates.
b Minimum restrictive temperature is the lowest temperature which prevents macroscopically detectable growth on YPD plates.
c Percent flocculation at 25°C was assayed as described under "Experimental Procedures." The mean and standard error for three independent cultures of each strain is shown.
d Data for the parental strain YPH250 and the isogenic control strain YDH6 are shown for comparison. The slight phenotype of YDH6 relative to YPH250 presumably reflects the absence of a functional CKA1 gene in YDH6 and/or occasional loss of the CEN plasmid carrying the rescuing CKA2 gene. The failure to obtain a quantitative G 1 arrest could be explained by the finite time required to inactivate the cka2-8 allele (or to dephosphorylate the relevant substrates) or by residual activity of the enzyme at the restrictive temperature. Preincubation of YDH8 at 37°C for either 1 or 2 h prior to release had little if any effect on the proportion of cells able to pass the G 1 block. In contrast, preincubation for 1 h at 38.5°C (the maximum permissive temperature of the control strain; Table III) and subsequent release at that temperature resulted in a complete G 1 block (data not shown). This result suggests some leakiness of the cka2-8 allele at 37°C.
Wild-type cells exposed to the mating pheromone ␣-factor arrest cell division, but not growth, at a later stage in G 1 than cells in stationary phase. To determine whether the point of G 1 arrest in response to CKII depletion lies before or after the point of ␣-factor arrest, YDH6 and YDH8 were grown to log phase, synchronized in G 1 with ␣-factor and then released at 25, 37, or 38.5°C. As shown in Fig. 4A, YDH8 responded to and recovered from ␣-factor with kinetics identical to those of the wild-type at 25°C. At 37°C (without preincubation) YDH8 exhibited only a slight G 1 defect (data not shown). The strain budded with near normal efficiency, albeit approximately 1 h slower than wild-type, and did not arrest until the G 2 and/or M block (75-80% budded cells, the vast majority large-budded). In contrast, at 38.5°C (and with a 1-h preincubation) the mutant exhibited a complete G 1 arrest (Fig. 4, B and C). Again, cell number did not increase (Fig. 4B), and in this case virtually all cells remained unbudded (Fig. 4C). At 38.5°C the control strain, YDH6, exhibited an approximately 50% reduction in growth rate and also reached stationary phase at a slightly lower cell density (Fig. 4B). Nevertheless, this strain was clearly able to exit G 1 and complete multiple rounds of cell division at this temperature.
upon prolonged incubation at 38.5°C. Such drifts have been noted before and are due to increased autofluorescence from cell enlargement (Reed and Wittenberg, 1990). This was consistent with the observed increase in the average size of the arrested cells throughout the course of the experiment. These experiments with pheromone-synchronized cells confirm that CKII is required for cell cycle progression in G 1 , at a point which must lie between the point of ␣-factor arrest and the onset of S phase.
In order to determine whether arrest at the G 1 block is reversible, G 1 -arrested YDH8 cells were shifted back to 25°C after 4 h at 38.5°C. Approximately 3-4 h after being returned to 25°C these cells simultaneously initiated bud formation (Fig. 4C) and entered S phase (Fig. 5). By 5 h of recovery approximately 80% of the cells were budded and contained a 2 N DNA complement. However, few if any cells appeared to be competent to divide at this point, and nearly all of these recovered cells formed an aberrant, elongated bud. Cells with the latter morphology have been observed previously following gradual depletion of CKII activity in a null background (Padmanabha et al., 1990) and are also prominent in cka2 ts strains incubated at a semipermissive temperature. 2 We speculate that this phenotype is associated with intermediate levels of CKII activity during G 2 /M, such that some CKII-dependent functions are completed but not others.
CKII Is Required for Cell Cycle Progression in G 2 /M-In order to probe the requirement for CKII at later points in the cell cycle, YDH6 and YDH8 were synchronized after the CKII G 1 block using hydroxyurea, an inhibitor of ribonucleotide reductase which arrests cells in early S phase with the largebudded morphology. Cells were released from hydroxurea arrest at 25 or 38°C (the latter after a 1-h preincubation). As shown in Fig. 6A, both strains exhibited a normal response to and recovery from hydroxyurea at the permissive temperature. At 38°C, the control strain YDH6 recovered and grew to saturation, though with the slower kinetics noted earlier (Fig. 6B). In contrast, YDH8 exhibited no increase in cell number at 38°C. This result implies the existence of at least one additional arrest point in the mutant. As with the first arrest point in G 1 , arrest at this second block also occurs in the first cycle following release from the synchronization block.
In order to define the position of the second arrest point, cells were analyzed by flow cytometry (Fig. 7). Prior to release from hydroxyurea, the mutant at 25 and 38°C and the wild-type at 38°C all exhibited a DNA content between 1 and 2 N, indicative of S phase arrest. Within the first 2 h after release, all three 2 D. E. Hanna and C. V. C. Glover, unpublished observation.

FIG. 4. Growth curves and budding profiles of CKA2 and
cka2-8 strains following release from mating pheromone-induced G 1 arrest. Cultures were inoculated into YPD, grown for 6 h at 25°C, and then synchronized in G 1 by exposure to ␣-factor. Cells were released from ␣-factor arrest at either 25 or 38.5°C (in the latter case, cultures were shifted to the nonpermissive temperature 1 h prior to the removal of ␣-factor). A, release at 25°C. B, release at 38.5°C. छ, YDH6 (CKA2); ࡗ, YDH8 (cka2-8). C, budding profile from an experiment identical to that in panels A and B, except that an additional YDH8 culture, released at 38.5°C, was returned to 25°C 4 h after release. f, YDH8 released at 25°C; छ, YDH6 released at 38.5°C; ࡗ, YDH8 released at 38.5°C; F, YDH8 released at 38.5°C and then returned to 25°C after 4 h. cultures resumed DNA synthesis, completed S, and acquired a 2 N DNA content. This result established that the mutant is able to complete previously initiated DNA synthesis at the restrictive temperature. At later time points, the two control cultures continued to cycle and ultimately lost synchrony, whereas the mutant at 38°C remained arrested with a 2 N DNA complement for up to 9 h. The flow cytometry data thus positioned the second cell cycle block in G 2 and/or M. The G 2 /M-arrested cells remained large-budded and, when examined for their nuclear and spindle morphologies, displayed the same characteristic range of large-budded phenotypes seen in asynchronous cultures at 37°C (data not shown). This suggested that the G 2 /M block defined using hydroxyurea synchronization is the same as that observed in the asynchronous cultures.
The reversibility of the G 2 /M block was assessed by returning a culture to 25°C after 4 h at 38°C. YDH8 remained largebudded after such a shift and did not resume normal cell cycle progression (data not shown), consistent with the results obtained with G 1 -arrested cells allowed to recover at 25°C (see above).
Because of potential artifacts associated with hydroxyurea treatment, we confirmed the G 2 /M block using a second synchronization protocol. Wild-type and mutant cells were synchronized in G 1 with ␣-factor and then released into fresh medium at 25°C. By 100 min after release approximately 80% of the mutant cells had initiated bud formation and DNA synthesis, indicating that they had passed the CKII G 1 block. When shifted to 38°C 100 min after release, the vast majority of these cells failed to complete division but arrested as large-budded cells with a 2 N DNA complement. These arrested cells displayed the same range of nuclear morphologies seen with hydroxyurea synchronization (data not shown). We conclude that the G 2 /M arrest is not an artifact of hydroxyurea treatment.
Viability at the Nonpermissive Temperature-The inability of YDH8 to resume normal cell cycle progression following release from CKII arrest at either the G 1 or G 2 /M block implied that YDH8 becomes inviable at the nonpermissive temperature. Direct measurement of percent viability (Fig. 8) confirmed that the viability of YDH8 declines precipitously between 2 and 4 h following a shift to 38.5°C, a time frame which coincides with cessation of cell cycle progression. The fact that few if any cells in the population retained viability at the nonpermissive temperature confirmed that cells arrested at either CKII block are inviable. Essentially identical results were obtained following a shift to 37°C (data not shown).
RNA and Protein Synthesis-The rapid loss of viability of YDH8 at the nonpermissive temperature raised the possibility that cell cycle arrest might be secondary to a general depression of cell metabolism or macromolecular biosynthesis. To address this issue, we compared the rates of total RNA and protein synthesis in YDH6 and YDH8 at both 25 and 38°C. These rates were estimated by measuring uptake of [ 14 C]uracil and [ 35 S]methionine, respectively, into acid-precipitable material during a 10-min pulse. As shown in Fig. 9A, the wild-type and mutant strains displayed comparable rates of RNA synthesis during log phase growth at 25°C. Both strains also displayed a similar heat shock-induced depression in RNA synthesis following a shift to 38°C. While the latter effect was transient in the wild-type, the rate of total RNA synthesis in the mutant failed to recover and remained approximately 4-fold lower than that of the controls for at least 4 h. Whether FIG. 6. Growth curves of CKA2 and cka2-8 strains following release from hydroxyurea-induced S phase arrest. Cultures were inoculated into YPD, grown for 6 h at 25°C, and then synchronized in early S phase by exposure to hydroxyurea. Cells were released from hydroxyurea arrest at either 25 or 38°C (in the latter case, cultures were shifted to the nonpermissive temperature 1 h prior to the removal of hydroxyurea). A, release at 25°C. B, release at 38°C. छ, YDH6 (CKA2); ࡗ, YDH8 (cka2-8).
FIG. 7. Flow cytometry of CKA2 and cka2-8 strains following release from hydroxyurea-induced S phase arrest. Data shown are from an experiment identical to that in Fig. 6, except that cells were exposed to hydroxyurea for 5.5 rather than 5 h. Cellular DNA content was determined by flow cytometry of propidium iodide-stained cells as described under "Experimental Procedures." The left peak in each profile reflects cells with 1 N DNA content, the right peak, cells with 2 N DNA content. Times shown refer to time after release from hydroxyurea. this reduction in the rate of RNA synthesis reflects a decrease in synthesis of rRNA, mRNA, or both is unknown. However, CKII phosphorylates a number of nucleolar proteins, including the mUBF transcription factor (Voit et al., 1992) and has been shown to activate rRNA transcription in isolated nuclei (Belenguer et al., 1989).
The reduction in the rate of total RNA synthesis was not reflected in a comparable reduction in the rate of total protein synthesis (Fig. 9B). While uptake of [ 35 S]methionine in the mutant was consistently lower than that of the wild-type, the magnitude of this effect (approximately 2-fold) was no greater at 38°C than at 25°C. By contrast, treatment with the protein synthesis inhibitor, cycloheximide, inhibited [ 35 S]methionine incorporation more than 20-fold for up to 4 h (Fig. 9B). The arrest of growth in the mutant strain, therefore, does not appear to be due to an overall inhibition of protein synthesis, although an effect on the synthesis of specific messages limiting for cell cycle progression cannot be ruled out by these experiments. Also, because of the energetic expense of protein synthesis, cell cycle arrest is unlikely to be due to a general decline in metabolic activity.

DISCUSSION
Conditional CKII Alleles-We have used a plasmid shuffling technique to isolate five independent temperature-sensitive alleles of the cka2 gene. All of these alleles are recessive to wild-type and confer a phenotype similar to that of a null (including flocculation, loss of viability, and arrest as a mixture of budded and unbudded cells), suggesting that they are lossof-function mutations. Consistent with this, extracts of YDH8 (grown at either 25 or 37°C) contain very low levels of CKII activity , 2 and the phosphorylation of at least two well characterized CKII substrates, topoisomerase II (Cardenas et al., 1992) and eIF2␣ (Feng et al., 1994), is temperature-sensitive in vivo in YDH8. The rapidity with which cells arrest following a shift to the nonpermissive temperature suggests that these mutants are temperature sensitive for activity per se rather than for enzyme synthesis or assembly (Hartwell et al., 1973). However, we have been unable to confirm this directly because of the low activity present in extracts of cells grown at permissive temperature.
We have used these temperature-sensitive alleles to define the requirement for CKII activity during the cell cycle in S. cerevisiae. As in any experiment employing a temperaturesensitive mutation, an important caveat is that the observed effects may be specific to heat-shocked cells. This concern is exacerbated in this case because of the higher temperatures required to obtain a tight arrest. While we cannot eliminate this caveat, we note that temperatures identical to those employed here have been effectively used to analyze cell cycle mutants in S. cerevisiae. For example, a temperature of 38°C was required to identify the G 2 /M function of Cdc28 (Reed and Wittenberg, 1990). Similarly, Tang and Reed (1993) used 38.5°C to obtain a tight G 1 arrest of a cks1 allele.
Requirement for CKII in G 1 -The results presented here demonstrate that CKII is required for cell cycle progression in G 1 . Studies with pheromone-synchronized cells indicate that CKII is required either at Start itself (defined as the point of ␣-factor arrest) or between Start and the initiation of DNA synthesis (the G 1 /S transition). We emphasize that the data do not exclude the possibility of additional arrest points in G 1 . Indeed, one apparent paradox in our results is that pheromonesynchronized cells do not arrest in G 1 when released at 37°C, whereas approximately half of an asynchronous population arrests in G 1 at this temperature. Although this result may simply reflect the shorter time available to inactivate the kinase and/or dephosphorylate the relevant substrates, it is possible instead that it implies an additional CKII-sensitive point in G 1 (prior to Start) for which 37°C is sufficient for arrest. Consistent with this, cells arrested in G 1 following release from stationary phase arrest at 37°C do not respond to ␣-factor (data not shown), indicating that they are not arrested at Start. Microinjection of antibodies against the ␤ subunit of CKII has been shown to inhibit mammalian cell cycle progression at multiple points in G 1 , specifically the G 0 /G 1 transition, early G 1 , and the G 1 /S transition (Pepperkok et al., 1994). A general requirement for CKII activity during G 1 is consistent with data FIG. 9. Rates of RNA and protein synthesis in CKA2 and cka2-8 strains at 25 and 38°C. Cells were grown in SMM to mid-log phase at 25°C and then either maintained at 25°C or shifted to 38°C. Aliquots were removed at the indicated times after the shift and pulselabeled for 10 min with either [ 14 C]uracil (A) or [ 35 S]methionine (B). Incorportation into acid-precipitable material was measured as described under "Experimental Procedures." The data are expressed as counts/min incorporated/cell. E, YDH6 (CKA2) at 25°C; q, YDH8 (cka2-8) at 25°C; छ, YDH6 at 38°C; ࡗ, YDH8 at 38°C; Ⅺ, YDH6 at 25°C in the presence of 10 g/ml cycloheximide (added immediately after the zero time point).
FIG. 8. Viability of asynchronously growing CKA2 and cka2-8 strains following a shift to nonpermissive temperature. Strains were inoculated into YPD at a starting density of 2 ϫ 10 5 cells/ml, grown for 8 h at 25°C, and then shifted to 38.5°C. Aliquots were removed at the indicated times after the shift, and percent viability was determined as described under "Experimental Procedures." छ, YDH6 (CKA2); ࡗ, YDH8 (cka2-8).
from mammalian systems indicating elevated levels of CKII activity in G 1 , particularly just prior to the G 1 /S transition (Carroll and Marshak, 1989;DeBenedette and Snow, 1991).
The targets of CKII which are essential for G 1 progression remain to be identified. In S. cerevisiae, progression through Start is regulated by a series of complexes between the Cdc28 protein kinase and five different cyclins, Cln1, Cln2, Cln3, Clb5, and Clb6 (Schwob and Nasmyth, 1993). Russo et al. (1992) have shown that p34 cdc2 is phosphorylated in vitro by purified CKII at Ser 39 and that this site is phosphorylated in vivo during the G 1 phase of the cell cycle. Whether Cdc28 is subject to the same modification is not known, but the relevant CKII phosphorylation site is conserved in Cdc28. Other proteins required for Start also represent potential targets. In an effort to identify genes which interact genetically with CKII, we have carried out a screen for multicopy suppressors of the cka2-13 allele. 3 Among the genes identified in this screen is CDC37, a previously characterized gene required for Start (Reed, 1980). We have found that Cdc37 is a physiological substrate of CKII and that mutation of the CKII recognition site impairs CDC37 function in vivo. 4 Although the biochemical function of Cdc37 is not known, failure to phosphorylate Cdc37 could explain the G 1 arrest of cka2 mutants.
Our analysis of cells released from hydroxyurea arrest suggests that CKII activity is not required for completion of S phase, consistent with the low level of CKII activity detected during S phase in mammalian systems (Carroll and Marshak, 1989;DeBenedette and Snow, 1991). However, we emphasize that this conclusion rests upon the dual assumption that the enzyme is fully inactivated by a 1-h preincubation at 38°C and that all relevant substrates become dephosphorylated. While the former appears likely in view of our studies of G 1 arrest, there is no information concerning the latter. At least one well characterized substrate of CKII is involved in DNA replication (DNA ligase I), and the activity of this protein is increased in response to CKII phosphorylation (Prigent et al., 1992). We also emphasize that our results do not preclude a requirement for CKII at the G 1 /S transition, as hydroxyurea arrest occurs after this point in the cycle.
Requirement for CKII in G 2 /M-The post-S phase arrest of CKII mutants is heterogeneous, with two-thirds of the arrested population exhibiting an anaphase morphology and the remainder a morphology typical of arrest in G 2 and/or metaphase. One interpretation of these data is that CKII is required at multiple points in G 2 /M. Alternatively, there may be a single requirement for CKII, but the resulting phenotype may be either leaky or intrinsically heterogeneous. We cannot at present distinguish among these possibilities. A post-S phase function for CKII is consistent with work in Xenopus laevis, where it has been demonstrated that microinjection of purified CKII into oocytes potentiates mitosis promoting factor-induced maturation (Mulner-Lorillon et al., 1988). Interestingly, the ␤ subunit of Xenopus CKII is phosphorylated in vitro by mitotic p34 cdc2 , and this phosphorylation results in increased CKII activity (Mulner-Lorillon et al., 1990). Both the ␣ and ␤ subunits of mammalian CKII are also phosphorylated in vitro by p34 cdc2 , and phosphorylation at these sites increases dramatically in human or chicken cells arrested in mitosis (Litchfield et al., 1991(Litchfield et al., , 1992. Although no changes in activity were reported in the latter studies, collectively these results suggest that CKII may function downstream of mitotic CDK activity. Although the targets of CKII relevant to arrest in G 2 /M remain to be identified, a promising candidate is the well characterized CKII substrate, topoisomerase II. This enzyme is a major structural component of the metaphase chromosome scaffold and is essential both for chromosome condensation in metaphase and for sister chromatid segregation during anaphase (for review, see . S. cerevisiae topoisomerase II is an excellent substrate of CKII in vitro, and phosphorylation strongly stimulates enzyme activity . The protein is phosphorylated in vivo and becomes hyperphosphorylated during mitosis (Cardenas et al., 1992). The excellent correlation between the in vivo and in vitro sites suggests that topoisomerase II is a physiological target of CKII in yeast. Consistent with this, phosphorylation of the protein is temperature sensitive in the YDH8 strain (Cardenas et al., 1992). Temperature-sensitive topoisomerase II (top2 ts ) mutants arrest in mitosis with an elongated spindle and a nucleus which is stretched through the bud aperture (Holm et al., 1985), a phenotype similar to that of anaphase-arrested cka2 ts cells. Moreover, like cka2 ts mutants, top2 ts mutants become inviable at the nonpermissive temperature, apparently during a failed attempt to segregate daughter chromosomes (Holm et al., 1985). Collectively, these data suggest that the anaphase arrest of cka2 ts strains may result from failure to phosphorylate and activate topoisomerase II. It appears unlikely that failure to activate topoisomerase II can account for those cka2 ts cells which arrest with a G 2 /metaphase phenotype, suggesting that additional substrates must be involved.
Conclusions-The results described here indicate a requirement for CKII in both the G 1 and G 2 /M phases of the cell cycle. An important unresolved issue is whether CKII is required only for the mechanics of cell cycle progression or whether it exerts a regulatory function in G 1 and/or G 2 /M. Our results argue against the trivial possibility that CKII-mediated cell cycle arrest is the result of a general decline in metabolic activity or biosynthetic capacity, since CKII-arrested cells continue to grow after cell cycle progression is halted (Padmanabha et al., 1990) and retain a significant capacity for protein as well as DNA synthesis. Based on arguments outlined above, we speculate that CKII acts downstream of mitotic CDK activity, perhaps as part of the interacting network of protein kinases activated at mitosis (King et al., 1994). Because CKII arrest is partially reversible in G 1 but irreversible in G 2 /M, an intriguing possibility is that a mechanism has evolved to monitor CKII activity in G 1 (perhaps via Cdc37) in order to ensure that adequate levels of activity are available to phosphorylate the structural and enzymatic proteins (such as topoisomerase II) needed during mitosis. The failure of cka2 ts mutants to recover completely from G 1 arrest may be an artifact associated with the necessity of resynthesizing CKII after the return to permissive temperature.
The properties of the CKII mutations described here differ in several respects from those reported recently for the orb5 allele of the S. pombe cka1 ϩ gene (Snell and Nurse, 1994). Following a shift to the nonpermissive temperature, orb5 cells undergo several cell divisions and ultimately die as small spherical cells. No cell cycle defects were noted in this mutant, which was interpreted as a morphological mutant defective in reinitialization of polarized growth following cytokinesis (Snell and Nurse, 1994). At face value, the different behavior of temperaturesensitive CKII mutations in the two organisms suggests significant differences in the physiological role of CKII in S. pombe and S. cerevisiae. However, we have recently isolated two temperature-sensitive alleles of the Sc CKA1 gene (encoding the CKII ␣ subunit) and find that these exhibit a behavior strikingly similar to that of the orb5 mutation, including the absence of first cycle arrest and adoption of a highly spherical morphology. 5 This result implies some functional specialization of CKA1 and CKA2 in S. cerevisiae and argues that the function of CKII in the two yeasts may be similar. Additional studies in both organisms will be required to clarify and correlate the physiological role of CKII in the two species.