Cell, Vol. 62, 225-237,

July 27, 1990, Copyright

0 1990 by Cell Press

Gl-Specific Cyclins of S. cerevisiae: Cell Cycle Periodicity, Regulation by Mating Pheromone, and Association with the p34CDC28Protein Kinase Curt Wittenberg, Katsunori Sugimoto, and Steven I. Reed Department of Molecular Biology, MB-7 Research Institute of Scripps Clinic 10666 North Torrey Pines Road La Jolla, California 92037

Summary The S. cerevisiae CLN genes encode cyclin homologs essential for progression from Gl to S phase. The CLNP gene encodes a 62 kd polypeptide that accumulates periodically, peaking during Gl and decreasing rapidly thereafter, and is rapidly lost following exposure of cells to mating pheromone. Cln2 abundance can be explained by the Gl-specific accumulation of the CM2 transcript coupled with instability of the Cln2 protein. The abundance of the CLNT and CM2 transcripts increases greater than L-fold during the Gl interval, decreasing dramatically as cells enter S phase. Both transcripts decrease in cells responding to mating pheromone. Finally, we demonstrate that the Cln2 polypeptide interacts with p34cDc2e to form an active protein kinase complex. This physical interaction is consistent with the genetic interaction between the CLN genes and CCC28 and suggests that Cln proteins are an essential component of the active protein kinase complex required for the Gl to S transition. Introduction The cell cycle is composed of a series of ordered events that lead ultimately to the duplication of all cellular components and their segregation to progeny cells. The key regulatory events in this process are thought to depend, in part, on the controlled accumulation of molecules with inductive activities. A family of proteins known as cyclins has recently been identified as likely candidates for these cell cycle inducers. Cyclins are characterized by periodic accumulation and degradation during the cell cycle, attaining maximal levels at the G2+M transition and then decreasing abruptly during M phase. Similar kinetics of accumulation have been demonstrated in eggs and embryos (Evans et al., 1963; Pines and Hunt, 1967; Standart et al., 1967; Swenson et al., 1966) as well as in somatic cells of a number of evolutionarily divergent organisms (Booher et al., 1969; Lehner and C’Farrell, 1969; Pines and Hunt, 1967; Pines and Hunter, 1969). Evidence that these molecules provide inductive activities that are essential for progression from G2 into mitosis is provided by the observation that cyclin mRNA can induce maturation when introduced into oocytes (Pines and Hunt, 1967; Standart et al., 1967) and that cyclin synthesis alone is sufficient to explain the dependence of the maturation process on protein synthesis (Minshull et al., 1969; Murray and Kirschner, 1969; Murray

et al., 1969). Although this conclusion is corroborated by evidence that fission yeast cyclin B, encoded by the c&73+ gene, is essential for progression from G2 into M phase, cyclin is not rate limiting for the induction of that transition (Booher and Beach, 1966; Hagan et al., 1966). A similar observation has been made using Drosphila embryos (Lehner and O’Farrell, 1969). The role of cyclins in promoting cell cycle progression is thought to be mediated through their direct interaction with members of the p34ccc2~c2+ protein kinase family, including maturation-promoting factor and the growthassociated histone Hl kinase, and in several cases they have been demonstrated to be components of the active protein kinase complexes (Arion et al., 1966; Draetta et al., 1969; Labbe et al., 1969; Pondaven et al., 1990; Meijer et al., 1969; Gautier et al., 1990; Booher et al., 1969; Pines and Hunter, 1969). While passage from interphase into mitosis is the principal cell cycle control point in oocytes and early embryos, the primary control of cell cycle progression in most nonembryonic systems is exerted during the Gl interval (reviewed by Pardee, 1969). Although significant effort has been directed toward elucidating the role of cyclins and members of the p34cDc28/cdc2+ protein kinase family in the initiation of mitosis, very little is understood concerning their role in the Gl to S transition. Evidence that this kinase is required at both the G1-S and G2+M transition is derived from studies using temperature-sensitive mutants of the CDC28 gene of Saccharomyces cerevisiae (Reid and Hartwell, 1977; Reed, 1960; Piggott et al., 1962; Reed and Wittenberg, 1990) as well as of the c&,2+ gene of Schizosaccharomyces pombe (Booher and Beach, 1966; Nurse and Bissett, 1961). Whereas cdc2 mutants of S. pombe arrest in both Gl and G2 at the nonpermissive temperature, it is primarily Gl arrest that is observed using cdc28 mutants. The Gl bias of cell cycle arrest conferred by cdc28 mutations under normal circumstances has enabled us to use genetic analysis to investigate the regulation of the p34CoQ8 protein kinase during the Gl interval. One approach has been to isolate wild-type genes that suppress the phenotype of cdc28 mutants when present in the cell in multiple copies (dosage suppressors). Two such genes, CLNl and CLN2, encode proteins that, while highly homologous to each other, have limited but significant similarity to mitotic cyclins (Hadwiger et al., 1969a). They also show limited similarity to another S. cerevisiae gene, CLN3, which had been identified in earlier studies and known variously as W/ill (Carter and Sudbery, 1960; Nash et al., 1966; Sudbery et al., 1960) and OAF7 (Cross, 1966). The products of genes CLNl, CLNP, and CLN3 have been shown to perform an essential overlapping function required for passage from Gl into S phase (Richardson et al., 1969). That is, inactivation of all three CLN genes by insertional mutation results in failure to pass from Gl into S phase, whereas cells carrying inactivated alleles of any

Cell 226

one or two of these genes are viable. Furthermore, passage through the Gl-S phase transition is conditional in cln- cells that express any one of the three genes under control of a regulatable promoter (Richardson et al., 1989; C. W., F Cross, and S. I. R., unpublished data). In contrast to mitotic cyclins, studies using dominant hyperactivated alleles of CLNP (Hadwiger et al., 1989a) and C&V3 (Cross, 1988; Nash et al., 1988) suggest that the Gl-S phase function provided by these genes is rate limiting. This is supported by the observation that these mutations advance the Gl -S phase transition, causing a decrease in cell size. Furthermore, both of these mutants are defective to varying degrees in their response to the two known environmental regulators of cell cycle initiation, nutrient limitation and exposure to mating pheromone. Based on the method of isolation of CLNl and CLNP as well as the phenotypes associated with dominant and recessive CLN mutations and knowledge of the association between mitotic cyclins and p34cDc2tidc2+ in other systems, we have hypothesized that the essentiality of the CLN gene products results from their role as activators of the Gl function of the Cdc28 protein kinase. In addition, we have proposed that they accumulate periodically during the cell cycle and, in that sense, are true cyclins (Hadwiger et al., 1989a; Richardson et al., 1989). We show here that the Cln2 polypeptide behaves in just such a manner, appearing during late Gl and disappearing shortly thereafter. Furthermore, we provide evidence that the periodic accumulation can be explained based on the regulated accumulation of CLNP mRNA in conjunction with an intrinsic instability of the Cln2 polypeptide. Finally, we demonstrate an association between the Cln2 polypeptide and p34CDC28 to form an active protein kinase complex. Results CLNP Encodes a 62 kd Polypeptide The CLNP gene encodes a putative polypeptide of 62 kd that shares limited but significant homology with the family of proteins known as cyclins (Hadwiger et al., 1989a). To study the protein product produced in yeast, antiserum was prepared by inoculating rabbits with Cln2 polypeptide produced by expression of the CLNP gene in Escherichia coli (see Experimental Procedures). lmmunoblots prepared using affinity-purified Cln2 antiserum identified the immunogenic polypeptide prepared from bacterial lysates (Figure 1, lane 1) as well as a number of polypeptide species in total protein extracts from wild-type S. cerevisiae cells (lane 3) and from cells overexpressing the CLNP gene under control of the yeast GAL7 promoter (lane 2) but not from cells carrying a disruption (insertional mutatioa) of the CLNP coding region (lane 4). These species, varying in apparent molecular size from approximately 83 kd to 77 kd, are all enhanced in the strain overexpressing CLNP. In addition, extracts from that strain contain a number of higher mobility species that are specifically recognized by the antiserum. In contrast, the Cln2 polypeptide produced in E. coli (Figure 1, lane 1) migrates with the mobility predicted for a polypeptide of a molecular size of 83 kd. This agrees well with the size of 81,857 daltons calcu-

-97.4 Cln2-

-66.5 -43.0

-29.0 Figure 1. Antiserum peptide Recognizes

Directed against Bacterially Produced Cln2 Polythe Cln2 Polypeptide in Yeast Extracts

Extracts were prepared in SDS-PAGE sample buffer from a yeast strain overproducing Cln2 (CWYlSl; lane 2) a wild-type strain (15Dau; lane 3). or a strain with an inactivated CLN2 gene (JHY359; lane 4) and separated by electrophoresis on a 4%/10% SDS-polyacrylamide gel followed by immunoblotting with Cln2 antiserum. Lane 1 represents 0.05 ug of electrophoretically purified, bacterially produced Cln2 protein. Molecular weight markers (Bethesda Reserch Laboratories) were rabbit phosphorylase B, bovine serum albumin, ovalbumin, and carbonic anhydrase, respectively.

lated for the primary translation product of the CLNP gene (Hadwiger et al., 1989a). While it is likely that the species observed in the Cln2overproducing strain that have a greater mobility than the bacterially produced protein represent degradation products, it is probable that the species with a decreased mobility have undergone one or more covalent modifications. While other modifications have not been eliminated, increased mobility of Cln2 following alkaline phosphatase treatment suggests that this mobility shift is due, at least in part, to protein phosphorylation (data not shown). The extent of this modification is not appreciably altered by overexpression of the CLNP gene, suggesting that the modification system involved is present in excess of the wild-type level of the protein. A number of bands that are not Cln2 specific, as demonstrated by comparison to extracts derived from the CLN2 disruption mutant (Figure 1, lane 4), are visible in all of the extracts. Bands at approximately 65 kd and 75 kd, barely visible in this experiment, are present to varying degrees, depending on the lot of affinity-purified serum used and the preparation of specific samples (see Figure 2). The Cln2 Polypeptide Accumulates during Gl and Disappears during the G1-4 Wansition To assess the relative abundance of the Cln2 protein during the cell cycle, a yeast culture was synchronized during

Gl-Specific 227

Cyclins

of S. cerevisiae

Time (min)

Tlme (mm)

Figure 2. The Cln2 Polypeptide Cell Division Cycle

Accumulates

Periodically

during

the

(A) Mating pheromone release synchrony. barl- yeast cells (GCYll) were synchronized in Gl by treatment with 40 nglml a factor for 200 min followed by release into medium without a factor at 0 min. Samples were taken at 10 min intervals for 4 hr. Unbudded cells (top, solid squares) and cells bearing small buds (open circles) were quantitated. Protein samples were prepared from each sample and analyzed by immunoblotting with either Cln2 (center) or ~34~~~~ antiserum (bottom). ClnZ-specific bands are indicated by arrows, while nonspecific bands are indicated by asterisks. Molecular weight standards were the same as in Figure 1. (B) Elutriation synchrony. Small unbudded cells prepared by centrifugal elutriation were allowed to proceed synchronously through the cell cycle. Samples were taken at 20 min intervals for 280 min and examined for the presence of budded cells (top, closed circles) and the proportion of cells undergoing DNA duplication as determined by flow cytometry (closed squares). Protein samples were prepared from each sample and analyzed by immunoblotting with either Cln2 (center) or pW”zs antiserum (bottom). Extracts from cells overproducing Cln2 (YCpG2::CLNZ) or carrying a disruption of CLN2 (cln2::LElJZ) were subjected to the same analysis. ClnP-specific bands are indicated by arrows, while nonspecific bands are indicated by asterisks. Overlap between asterisks and arrows indicates a ClnP-specific increase in signal occurs at that position. Different preparations of affinity-purified antiCln2 serum were used for the experiments in (A) and (8).

the Gl interval by treatment with mating pheromone (see Experimental Procedures), released from pheromone-induced arrest by inoculation into fresh medium, and then sampled at 10 min intervals over the next 4 hr. The cells contained in each aliquot were then examined morphologically to determine cell cycle position (Figure 2A, top), and total protein extracts were prepared and examined for the presence of the Cln2 polypeptide by immunoblotting (Figure 2A, middle). At least four ClnPspecific bands (indicated by arrows) are undetectable at the time of release from mating pheromone but accumulate abruptly at approximately 50 min. The majority of cells in this population are in late Gl just prior to the appearance of buds. In S. cerevisiae, budding occurs coincident with entry into S phase (Hereford et al., 1981) and is, therefore, a good morphological marker for that event. The accumulation of Cln2 polypeptide continues until approximately 80 min following release and then decreases, returning to its initial level by 110 min, coincident with the disappearance of unbudded cells from the culture. Finally, the Cln2 polypeptide begins to accumulate again at approximately 160 min, just prior to the second cycle of bud emergence. While the abundance of the Cln2 polypeptide varies at least 5-fold over this time course, the Cdc28 polypeptide (Figure 2A, bottom) remains constant (Mendenhall et al., 1987; Wittenberg and Reed, 1988). Similar behavior is observed for the nonspecific 65 kd band recognized by the anti-Cln2 antiserum (indicated by the asterisk in Figure 2A, middle). To exclude potential artifacts associated with induction of synchrony by mating pheromone, it was necessary to monitor Cln2 abundance in cells that had not been previously arrested. Cells were therefore synchronized by selecting the smallest unbudded cells in an asynchronous culture by centrifugal elutriation (Hereford et al., 1981). This population, representing mostly newly abscised daughter cells, was allowed to grow synchronously for 280 min. Aliquots for immunoblotting, morphological examination, and flow cytometric analysis were taken at 20 min intervals (Figure 26). Although the degree of synchrony obtained in this experiment was inferior to that achieved using mating pheromone, the appearance of Cln2-specific bands (indicated by arrows in Figure 26, middle) was observed with similar timing. Again, the Cln2 polypeptide begins to accumulate just prior to the Gl+S phase transition and then rapidly decays. Determination of the extent of this increase is problematic due to the prominence of two relatively constant non-ClnBspecific bands of 65 kd and 75 kd (indicated by asterisks). These bands are present, albeit at lower intensity, in control extracts carrying an inactivated c/n2 gene (c/n2::LEU2; Figure 26, middle) as well as in the experiment shown above (Figure 2A). The variability in the intensity of these bands between experiments correlates with the batch of antiserum used and the preparation of the extracts. From these experiments it can be concluded that the maximal accumulation of Cln2 polypeptide occurs in cells prior to budding and before the initiation of S phase, as determined by flow cytometry. This is consistent with Cln gene products being required for cells to progress from Gl into S phase (Richardson et al., 1989).

Cell 226

A

cdc28 ” Time

0

at 35’(hr)

cd&

2

ts 2

0

Cln2j

B

cdc28 ts a-factor Temp Time

(hr)

-

-

23’

35’35’

o

2

+

4

-

-

35’

23’

4

4

0

30

60

90

Figure

3. The Cln2 Polypeptide

Accumulates

during

the Gl Interval

(A) Arrest by cdc2P and cdc8’s mutations. Yeast carrying a temperature-sensitive mutation in either the cdc28 gene (D4a) or the cdc8 gene (198dl) were analyzed for the presence of the Cln2 polypeptide by immunoblotting. Samples were prepared from cells growing asynchronously (0 hr) or following arrest by shift to the restrictive temperature (35OC) for 2 hr. (6) c. factor treatment of Gl-arrested c&28@ mutant. cdc28 mutant cells (D4a) were analyzed for the presence of Cln2 by immunoblotting either prior to arrest at the restrictive temperature (23’%), following arrest at the restrictive temperature for 2 hr (3VC, 2 hr). following an additional 2 hr at the restrictive temperature with and without the addition of 40 rig/ml a factor (36X, 4 hr), or following an additional 4 hr of growth at the permissive temperature (23% 4 hr).

The experiments described above depend to some degree on morphological markers of cell cycle progression to determine the timing of Cln2 accumulation. To confirm the conclusion that accumulation precedes the G1-S transition, the abundance of the Cln polypeptide was examined by immunoblotting in extracts from cells arrested either before the G1-6 phase transition by a cdc28 mutation or after entry into S phase using a cc/c8 mutation (Figure 3A). CDC8 encodes thymidylate kinase, which is essential for DNA replication (Sclafani and Fangman, 1984). In agreement with the cell cycle synchrony experiments that place Cln2 accumulation prior to the initiation of budding and S phase, the extracts from cells arrested at the preinitiation c&28 block point contain high levels of Cln2 protein, whereas Cln2 is undetectable in extracts from cc/c8 mutants arrested in S phase. Mating Pheromone Induces the Rapid Loss of the Cln2 Polypeptide One of the primary consequences of the response to mating pheromone is arrest of yeast cells in the Gl interval of the cell cycle (Biicking-Throm et al., 1973; Pringle and Hartwell, 1981). We have hypothesized that this response is mediated through the elimination or inactivation of the CLN products and hence inability to initiate a new cell cycle (Richardson et al., 1989). The availability of antiserum specific for the Cln2 polypeptide allowed us to

120

150

ISO

30

60

90

120

t -UF Time (mini

t +d

+aF Time(minj’0

Figure 4. Accumulation Responding to Mating

-aF 15 30 60 160'15 30 6090 1 20

of the Cln2 Polypeptide Pheromone

Is Repressed

in Cells

a-factor arrest and release. An asynchronous culture of barl- yeast cells (GCYll) was treated with 40 nglml a factor, and samples were taken for quantitation of budded cells (top) and immunoblotting using Cln2 antiserum (bottom) at the times indicated. After 3 hr in mating pheromone, cells were released by washing with fresh medium without a factor and analyzed for the next 2 hr.

test that hypothesis. Cln2 polypeptide was monitored over a time course of a factor-induced cell cycle arrest and subsequent synchronous recovery following inoculation into fresh medium without a factor (Figure 4). Asynchronous cells treated in this way arrest within a single cell cycle as unbudded Gl cells and then synchronously reenter the cell cycle after the removal of a factor (Figure 4, top). Cln2 polypeptide is rapidly lost from these cells (Figure 4, bottom), becoming undetectable within 30 min following addition of pheromone. As cells undergo synchronous progression from Gl into S phase, the Cln2 polypeptide reappears and accumulates to high levels, as anticipated. The loss of Cln2 observed in response to a factor could be either a direct result of pheromone action or an indirect effect mediated by synchronization at a point in the cell cycle prior to the window of Cln2 accumulation. Therefore, cells were synchronized in Gl using a c&z28 mutation, and then the arrested cells were treated with mating pheromone. Extracts of these cells were analyzed for the presence of Cln2 polypeptide prior to or following arrest at the restrictive temperature (Figure 3B). As demonstrated in Figure 3A, the Cln2 polypeptide persists in the c&28-arrested cells. However, when these cells were treated with mating pheromone while being held at the re-

Gl-Specific 229

Cyclins

A

of S. cerevisiae

h-3% E2; .

iJ e x

C Templ’C) CLN2-

CDC+cdc28 cdc7 cdc9 CDC+ -II--23 37 23 37 2337 23 37 HU Nz

*i

CLNlLEU2- CLN2-

0 15 30456012015304560

LEU2-

(23’X) or following arrest by shift to the restrictive temperature for 2 hr (PC) or from or by nocodazole (Nz), a microtuble inhibitor. Ten micrograms of RNA from each analyzed by Northern blot analysis using a CLN2epecific probe. The same blot (0) Analysis of CLNI, CLN2, and CLN3 transcripts in cells responding to a factor. at the indicated times following addition of 40 rig/ml a factor and during subsequent for the proportion of budded cells, and 10 pg of total RNA prepared from each sample by Northern blot analysis as in Figure 56.

strictive temperature for an additional 2 hr, the bulk of the Cln2 polypeptide was lost. In contrast, control cells held at the restrictive temperature without mating pheromone treatment maintained a high level of the Cln2 polypeptide. Thus, mating pheromone is capable of inducing loss of the Cln2 polypeptide independent of cell cycle progression or cell cycle position. Accumulation of Gl Cyclin mRNAs Is Periodic during the Cell Cycle The periodic accumulation of a polypeptide during the cell cycle could be achieved by regulating either its synthesis or its turnover, or by a combination of the two. To examine whether regulated accumulation of the CM2 transcript contributes to the regulation of Cln2 polypeptide abundance, Northern blot analysis was performed. Since we had previously demonstrated that the CM7 and CLN3 genes perform an essential function that overlaps with the function of CLN2, an analysis of those transcripts was carried out in parallel. All three of the CLN genes produce an mRNA transcript of approximately 2.1 kb, as determined by comparison to the LEU2 transcript and the 25s and 18s ribosomal RNAs. These transcripts are observed in haploid cells of both mating types as well as in a/a diploids (Figure 5A, lanes 4-8), consistent with the observation

Figure 5. The Accumulation of the CLN7 and CLN2 mRNAs Is Periodic during the Cell Cycle and Is Repressed in Cells Responding to Mating Pheromone

(A) Accumulation of CLNI, CLM, and CLN3 transcripts occurs in all ceil types independent of mating type. Ten micrograms of total RNA prepared from yeast carrying a disruption of each of the three CLN genes and from wildtype cells of the a and a mating types and a/a diploids was run on a 1% formaldehyde agarose gel and analyzed by Northern blot analysis using CLNI-, CLN2; and CLN3-specific probes. (9) Analysis of CLNl, CLN2, and CLN3 transcripts in cells undergoing synchronous cell cycle progression following release from mating pheromone-induced arrest. Ten micrograms of total RNA prepared from samples of cells taken at 20 min intervals from the mating pheromone release synchrony presented in Figure 2A was 90120 separated on 1% formaldehyde-agarose gels and analyzed by Northern blot analysis using CLNI-, CLNB, and CLN3epecific probes. The same blots were analyzed using a LEU2 probe as a control (the results of one such blot are shown). The top panel presents the same data shown in Figure 2A. (C) Analysis of CLN2 transcripts in cells arrested by cdc mutations or by metabolic inhibitors. Total RNA was prepared from cultures of wild-type cells (WC’) or temperature-sensitive cdc mutants growing either asynchronously cells arrested by hydroxyurea (HU), an inhibitor of DNA synthesis, sample was separated on 1% formaldehyde-agarose gels and was analyzed using a LEU2 probe as a control. Samples were taken from a culture of &Iyeast cells (GCYll) release into medium without pheromone. Samples were analyzed was separated on 1% formaldehyde-agarose gels and analyzed

that the requirement for CLNgene products is not cell type specific (Richardson et al., 1989). Evidence that the transcript observed in each case is the authentic CLN mRNA was derived from analysis of RNA from haploid yeast strains in which one of the endogenous CLN genes had been disrupted by insertional mutation. Insertional mutations of CLN7 and CLN3 result in an alteration of the size of the transcript observed (Figure 5A, lanes 1 and 3, respectively), while in the case of the CLN2 mutation, no intact transcript is observed (Figure 5A, lane 2). Although the CLN3 mutation results in the accumulation of transcripts of two distinct sizes, both differ from that produced by the wild-type gene. CLN7, CLN2, and CLN3 transcript levels were analyzed by Northern blot analysis of total RNA prepared from samples taken at 20 min intervals during the mating pheromone release synchrony experiment presented in Figure 2A. These experiments revealed that the abundance of both CLN7 and CLN2 transcripts varies at least 8-fold during the cell division cycle (Figure 58). Both transcripts, while undetectable at the beginning of the time course, increase rapidly between 20 and 40 min following release from mating pheromone arrest. This accumulation occurs approximately 40 min prior to the appearance of budded cells and at least 10 min prior to the appearance of Cln2

Cell 230

polypeptide in the cells taken from the same experiment (Figure 2A). Furthermore, both transcripts begin to decrease approximately 80 min after mating pheromone release and reach a minimum at 120 min. These times correlate quite well with the kinetics of Cln2 loss. Both CLNl and CLNP transcripts accumulate again at 140 min, preceding both the second cycle of bud emergence and the second increase in Cln2 polypeptide. Since it is difficult to determine the precise coordinates of cell cycle position based on morphology, it was necessary to use cell cycle mutants to establish the dependence of transcript accumulation on the completion of cell cycle events. The abundance of CLNl and CLNP transcripts was examined in several cdc mutant strains following arrest at the restrictive temperature. Analysis of the LEUP transcript on the same RNA blot is presented as a control. In contrast to wild-type controls (Figure 5C, lanes 1 and 2) the CLNP trancript accumulates in cells arrested during Gl by the temperature-sensitive mutation c&28-4 (Figure 5C, lanes 3 and 4). This accumulation is even more striking when compared with the LEUP control, which decreases noticeably under these conditions, perhaps due to death of a portion of the cells. This increase in the CLNP transcript is consistent with the persistence of Cln2 protein in cells arrested under the same conditions (Figure 3). However, while the CLNP transcript accumulates at the arrest point, the abundance of the Cln2 polypeptide remains constant. This might be explained by either an increase in the rate of protein turnover in the arrested cells or a relative decrease in the rate of translation of the mRNAs. In contrast to cells arrested in Gl, the CLNP transcript is not observed when cells are arrested in S phase as the result of either a temperature-sensitive mutation of the gene CDC8 (Figure 5C, lanes 7 and 8) (Hartwell, 1976) or by treatment with an inhibitor of DNA synthesis, hydroxyurea (lane 9) (Slater, 1983). Likewise, when cells are arrested in M phase by treatment with a microtubule inhibitor, nocodazole (lane 10) (Pilus and Solomon, 1986; Thomas et al., 1985) CLN2 mRNA fails to accumulate. Finally, CLN2 transcripts appear to be partially reduced in cells arrested late in Gl at the cdc7arrest point (lanes 5 and 6). This mutation arrests cells subsequent to the block point defined by mating pheromone or cdc28 mutations, known as Start (Reid and Hartwell, 1977) but prior to the hydroxyureaand CDCB-dependent events. These experiments restrict the window of accumulation of the CLN2 transcript to a period late in the Gl interval, either prior to or coincident with Start, an event associated with control of the Gl-S phase transition. A similar analysis of the CLNl transcript yielded essentially identical results (data not shown). In striking contrast to the CLNl and CLN2 mRNAs, the abundance of the CLN3 transcript is relatively invariant. This observation is in agreement with the results of Nash et al. (1988) who observed little difference during the cell cycle in samples obtained by sequential elutriation. Considering the similarity of their functions, the expression of the CLN genes might be expected to be subject to similar regulation. However, a difference in the regulation of CLN3 relative to CLNl and CLNP is not entirely surprising in light of a number of distinctive properties of the respective genes (discussed below).

Since mating pheromone treatment was shown to result in a decrease in the abundance of Cln2 polypeptide (Figure 4) we examined the affect of a factor treatment on the accumulation of CLN mRNAs. An asynchronous culture was treated with a factor and analyzed for the presence of CLN transcripts at the designated intervals over the next 2 hr (Figure 5D). Those same cells were then released from arrest by removal of mating pheromone and followed for an additional 2 hr. As was observed for the Cln2 polypeptide (Figure 4) the levels of both the CLNl and CLN2 transcripts decline in the presence of a factor. However, the kinetics of their loss are quite different. While the CLNl transcript is undetectable within 30 min following addition of a factor, the CLN2 transcript decreases slowly and is still barely detectable after 2 hr. This is in contrast to the Cln2 polypeptide, which is undetectable within 30 min following mating pheromone treatment (Figure 38). The discrepancy between the kinetics of decay of the CLNP transcript and that observed for the protein product can be explained if the decrease observed in the level of CLNP mRNA is sufficient to slow Cln2 synthesis to a rate that is significantly lower than the constitutive rate of Cln2 turnover (discussed below). Finally, as predicted by the mating pheromone release synchrony experiments shown in Figure 5B, both the CLNl and CLNP transcripts reaccumulate approximately 30 min following release from a factor arrest. The behavior of the CLN3 transcript in response to mating pheromone treatment, like that observed during the cell cycle, is quite different from that of the other two Gl cyclins. In agreement with previously published results (Nash et al., 1988) the transcript increases approximately 2-fold in response to mating pheromone treatment (Figure 5D). This accumulation is maximal within 30 min following addition of a factor and returns to the basal level within 15 min following its removal. Again, the pattern of CLN3 mRNA accumulation is inverted with respect to that of the CLNl and CLNP transcripts. The Cln2 Polypeptide Is Unstable To explain the rapid decrease in the Cln2 polypeptide observed both during the cell cycle and in response to mating pheromone, either constitutive or regulated instability of the protein must be invoked. Instability of the Clnl protein is suggested by the studies of cells in which the only source of Cln protein was a CLNl gene under control of the yeast GAL1 promoter (Richardson et al., 1989). These cells arrest within a single cell cycle when expression from the GAL1 promoter is repressed. First cycle arrest is observed despite the fact that CLN mRNA accumulates in the induced cells to greater than 5 times the wild-type level, suggesting that, even when present at elevated levels, the Clnl polypeptide can be inactivated in less than the duration of one cell cycle. To determine whether the Cln2 protein is unstable, we monitored its abundance in a strain identical to that studied by Richardson et al. (1989) except that CLNP, rather than CLNl, was under regulation of the GAL1 promoter. This strain has a constitutive level of the Cln2 polypeptide that is greater than 20fold higher than that of the wild-type strain (Figure 1). The cells were grown under inducing conditions (2% galac-

Gt-Specific 231

Cyclins

of S. cerevisiae

Time (mini after Glc addition

0 15 30 60 90 120 180

P34CDC28 Figure

6. The Cln2 Polypeptide

Is Very Unstable

Relative

to ~34~~28

Samples of cells carrying, as their sole source of either Cln2 (CWYlCl) or p34cccss (MMY149), the respective genes under control of the galactose-inducible promoter of the yeast GAL7 gene were taken at the times indicated following a shift from inducing (growth in 2% galactose) to repressing conditions (transfer to 2% glucose). Extracts were prepared and analyzed for the presence of Cln2 or ~34~~~~ by immunoblotting.

B

a-ClnZtose) and then shifted to repressing conditions (2% glucose), and the level of the Cln2 polypeptide was monitored by immunoblotting over the next 3 hr (Figure 6A, top). As a control, the level of the p34cDc28 polypeptide was followed in a strain with a galactose-regulated CDC28 gene as the only source of the ~34~~~8 protein (Figure 6A, bottom). The results of this experiment demonstrate that the Cln2 polypeptide is highly unstable relative to the ~34~~~8 polypeptide. Cln2 is lost with a half-life of less than 15 min, as determined by densitometry, whereas the half-life of ~34~~~~ exceeds the duration of this experiment. Thus, as predicted by the observation that newly synthesized Cln2 protein is required for initiation of each new cell cycle, the Cln2 protein produced from the GAL7 promoter can be completely degraded in less than one cell cycle. However, this experiment cannot distinguish between a discrete period of degradation during the cell cycle and constitutive degradation. Furthermore, it is unclear whether the instability is an inherent property of the protein or mediated by a system specifically targeted to this protein or class of proteins. It should be noted that all three of the Gl cyclins contain PEST sequences (Hadwiger et al., 1989a; Nash et al., 1988) which have been suggested to signal protein turnover based on their appearance in highly unstable proteins (Rogers et al., 1986). In any case, these results do suggest that the capacity of the degradation machinery substantially exceeds that necessary for degradation of the protein at wild-type levels. Cln2 Forms a Complex with p34cDc28 By analogy to mitotic cyclins, which are thought to perform their cell cycle regulatory function through association with p34ccc2~dc2+ (see Introduction), we have proposed that an interaction occurs between p34cDc28 and the Cln proteins (Hadwiger et al., 1989a; Richardson et al., 1989). Consistent with this hypothesis, high levels of expression of the CLN gene products suppress the temperature sensitivity of c&28 mutants (Hadwiger et al., 1989a). To determine whether a physical association in fact occurs, we examined the ability of antiserum directed against the Cln2 polypeptide to coimmunoprecipitate ~34~~~~ (Figure 7A). AntiCln2 immunoprecipitates, prepared under nondena-

]-Histone

H1

]-Histone

H1

Figure 7. The Cln2 Polypeptide Forms a Complex Phosphorylates Histone Hl In Vitro

with p34coQs

That

(A) Coimmunoprecipitation of p34cocz8 and Cln2. lmmunoprecipitates were prepared using affinity-purified Cln2 antiserum (see Experimental Procedures) from extracts of cells carrying an inactivated c/r12 gene (cln2::LEU2, JHY359), wild-type cells (CLNP, lLDau), wild-type cells carrying the plasmid YCpGP without an inserted gene (CWYl45), or cells overproducing Cln2 protein from the yeast GAL7 promoter (YCpG2::CLN2, CWY151). The immune complexes were separated by SDS-PAGE and analyzed for the presence of ~34~~~s by immunoblotting with ~34~~s~ antiserum. Purified ~34~~~~ monomer run on the same gel is shown as a control. (B) lmmunoprecipitation of histone HI kinase activity by antiCIn and anti-p34CocZ8 serum. Immune complexes prepared using either Cln2 antiserum (top) or p34coc28 antiserum (bottom) from a strain overproducing Cln2 (YCpG2::CLN2, CWY151), wild-type cells carrying the plasmid YCpGP without an inserted gene (CWY145), a strain carrying an inactivated chromosomal c/n2 gene (cln2::LEU2, CWYl51), or a strain carrying the wild-type CLN2 gene but with a temperaturesensitive mutation of cdc28 (CLN2 cdc28-4) were analyzed for histone Hl kinase activity. The autoradiograph for the top panel was for 4.5 hr with a Cronex Lightning intensifying screen, while that for the bottom panel was for 80 min without an intensifying screen.

turing conditions (see Experimental Procedures), were separated by SDS-PAGE and analyzed by immunoblotting for the presence of ~34~~~~. In each case, ~34~~28 coprecipitated with the Cln2 polypeptide (Figure 7A, lanes 2-4), except when the extract was derived from a strain that is unable to produce intact Cln2 polypeptide due to an insertional mutation of CLN2 (Figure 7A, cln2::LEU2). Furthermore, the amount of ~34~~~~ in the immune complexes was proportional to the abundance of the Cln2 polypeptide in the extract. When the extract was prepared from a strain overproducing the Cln2 polypeptide, the amount of ~34~~~~ observed in the immune complexes increased. This is consistent with the observation that overproduction of the Cln2 polypeptide results in several phenotypes that have been attributed to the inappropriate or excessive activation of the ~34~~~~ ,protein kinase (see Discussion). While we have not directly quantitated the extent of this association, only a small proportion of

C@ll 232

the total ~34~~~~ is involved. This result was expected since only a small proportion of the cells in these asynchronous cultures reside in the Gl interval, during which Cln2 is abundant. To assess whether the complexes containing Cln2 and ~34~~~~ possess protein kinase activity, anti-Cln2 immune complexes were evaluated for their ability to phosphorylate histone Hl. lmmunoprecipitation was performed using either Cln2 antiserum (Figure 78, top) or ~34~~~~ antiserum (Figure 7B, bottom), and the resulting immune complexes were used for histone Hl kinase assays (see Experimental Procedures). As expected, anti-p34Ccc28 immune complexes from all cells carrying a wild-type allele of CDC28 catalyzed the phosphorylation of histone Hl (Figure 78, bottom), while extracts from the temperaturesensitive mutant, cdc28-4, were inactive (CLN2 cdc28-4). In contrast, Cln2 antiserum immunoprecipitated histone Hl kinase activity from cells producing either wild-type (YCpGP) or elevated levels (YCpG2::CLN2) of Cln2. However, this activity was dependent on the presence of both the Cln2 protein (cln2::LEU2) and functional ~34~~~~ (CLNP cdc28-4). That is, when the extracts were prepared from yeast carrying an insertional mutation in the Cln2 gene or a thermolabile ~34~~~~ protein, histone Hl phosphorylation was not observed. Furthermore, the ability to immunoprecipitate Cdc28 protein kinase activity was enhanced in cells overproducing Cln2 (YCpG2::CLN2). These results correlate precisely with the presence of p34cDc28 in the antiCln2 immune complexes observed in Figure 7A. It should be noted that, while histone Hl is an excellent substrate for the p34cDc28 complexes that accumulate outside of Gl phase (Reed and Wittenberg, 1990), it may not be an optimal substrate for this protein kinase during Gl. In fact, we have previously shown that histone Hl kinase activity observed with immune complexes formed using antiserum directed against p34cDc28 is very low during the Gl interval (Reed and Wittenberg, 1990). Consistent with this observation and the fact that the majority of cells in these populations reside outside of the Gl interval, the relative histone Hl kinase activity 30observed with p34cDc28 antiserum is approximately fold higher than that observed with antiserum directed against Cln2. Discussion Gl Cyclins versus Mitotic Cyclins The results presented here establish that the Cln2 protein is an authentic cyclin with respect to several of its properties. As with mitotic cyclins, Cln2 abundance is periodic with respect to the cell cycle. However, in this case the maximal levels of Cln2 protein are attained during the Gl interval. Furthermore, the Cln2 protein is unstable. However, it is unclear whether Cln2 instability is constitutive or, as with mitotic cyclins, is limited to a discrete period of the cell cycle. The analogy to mitotic cyclins is extended by the observation that Cln2 associates with ~34~~~~ to form an active protein kinase complex. Finally, we have previously demonstrated that Cln protein, whether Cln2 or

one of the products of the other CLN genes, is essential for cell cycle progression (Richardson et al., 1989). Although these two classes of cyclins share a number of similarities, the properties of Gl cyclins and mitotic cyclins differ in several ways. First, while the abundance of previously described cyclins peaks at mitosis, Cln2 accumulation peaks during the Gl interval, consistent with the Gl requirement for CLN gene products (Richardson et al., 1989). Second, the accumulation of CLN7 and CLN2 RNA transcripts is periodic with respect to the cell cycle and limited to the Gl interval. Thus, regulation of the abundance of Cln2 polypeptide during the cell cycle could be explained on the basis of periodic accumulation of the CLNP transcript coupled with constitutive instability of the protein. Mitotic cyclins, on the other hand, appear to be synthesized throughout the cell cycle in oocytes and early embryos (Evans et al., 1983; Swenson et al., 1986; Westendorf et al., 1989) and from S through G2 phase in somatic cells (Pines and Hunter, 1989). However, the stability of mitotic cyclins varies with respect to position during the cell cycle, with rapid proteolysis occuring at or around the metaphase to anaphase transition (Evans et al., 1983; Pines and Hunter, 1989). Finally, Gl cyclins appear to be rate limiting for cell cycle initiation. This conclusion is based on the observation that dominant mutations of CLN2 and CLN3, which are thought to result in hyperactivity and/or hyperstability of the gene products, confer a number of phenotypes consistent with premature execution of the Gl -S phase transition, including small cell size and the inability to respond to Gl regulatory signals (Cross, 1988; Hadwiger et al., 1989a; Nash et al., 1988). Furthermore, a similar phenotype is observed in cells expressing high levels of the Clnl and Cln2 proteins (C. W. and S. I. R., unpublished data). In contrast, studies with mitotic cyclins have shown that they are necessary but not sufficient for mitotic induction. For example, in S. pombe the cdc73 gene product, a homolog of cyclin B (Solomon et al., 1988; Goebl and Byers, 1988) is essential but not rate limiting for mitotic induction (Booher and Beach, 1986; Hagan et al., 1988). The action of the mitotic inducer cdc25+, however, appears to be rate limiting for mitotic activation of the p34 kinase (Moreno et al., 1990; Russell and Nurse, 1986). Additionally, a prerequisite for tyrosine dephosphorylation of ~34~~~~ has been clearly demonstrated (Gould and Nurse, 1989). Parallel conclusions have been drawn concerning cyclin levels in early embryos of Drosophila, where the timing of entry into mitosis is independent of the rate of cyclin accumulation (Lehner and O’Farrell, 1989). Furthermore, although preliminary studies using a cyclindependent in vitro Xenopus oocyte system suggested that cyclins were rate limiting for mitotic induction (Murray et al., 1989) a cyclin-independent lag is ObseNed, consistent with a requirement for other components or activities. Recent studies have provided a possible biochemical basis for understanding these observations. While the association of cyclin with p34 results in a complex with the potential for activation as a histone Hl kinase, that kinase remains inactive unless further modifications of these pro-

G&Specific

Cyclins

of S. cerevisiae

teins occur (Fondaven et al., 1990). Thus, the addition of cyclin alone is insufficient to activate M phase-specific histone Hl kinase. In contrast, the genetic data derived from studies of the CLN genes predict that activation of the Gl-specific form of the kinase requires only Cln protein accumulation or activation. Regulation of CLNI and CLNP Wanscript Accumulation during the Cell Cycle and by Mating Pheromone We have previously proposed that the abundance of CLN gene products is regulated both as a function of the extent of cell growth and in response to environmental factors that control cell cycle progression during Gl. These factors include both nutrient limitation and response to mating pheromones. The results presented here demonstrate that the abundance of the CLN7 and CLNP transcripts is regulated as a function of position in the cell cycle and in response to mating pheromone. There are several examples of genes that are transcriptionally regulated as a function of cell cycle progression (reviewed by Breeden, 1988). However, in each of these cases, the expression is dependent on the execution of the CDCPB-dependent cell cycle event known as Start (Hartwell et al., 1974). Such cell cycle-dependent regulation has been observed for several genes whose products are involved in DNA replication (POL7, CDCB, CDCS, and CDC27), chromosome structure (HTA7 and HT67), and DNA repair (RADG and RADS) as well as the HO (Nasmyth, 1985) and SW15 (Nasmyth et al., 1987) genes, which are required for mating-type switching observed in homothallic strains (reviewed by Klar et al., 1984; Stern et al., 1984). However, the expression of all of these genes begins following Start during either late Gl or S phase. Thus, the CLN7 and CLN2 transcripts are, to our knowledge, the first observed to accumulate prior to Start. The timing of accumulation of these transcripts as well as the accumulation of the Cln2 polypeptide is consistent with the proposed role of these proteins in regulating the Gl function of the p34coc2* protein kinase. The negative effect of mating pheromone on the abundance of CLN7 and CLNP transcripts distinguishes them from most other pheromone-regulated genes that have been studied. While mating pheromone has been demonstrated to induce the transcription of a number of genes (reviewed by Cross et al., 1988) we are aware of only one example in which a negative effect on transcript accumulation has been demonstrated (Stetler and Thorner, 1984) and in that case the identity of the encoded protein is unknown. Based on its,size, this transcript is unlikely to encode any of the Cln proteins. The regulation observed for the CLN7 and CLNP genes could be effected via pheromone-induced repression of transcription or through the inactivation of a positive cell cycle regulatory signal. It is notable that the upstream region of the CLNP gene contains at least one perfect consensus pheromone response element (TGAAACA) .and several others with six of the seven nucleotides. This DNA element is found in the upstream regions of several genes that are under positive

pheromone regulation and Thorner, 1987).

(Kronstad

et al., 1987; Van Arsdell

Cln3 Is Regulated Differently Than the Other Gl Cycllns The analysis of the CLN transcripts reported here in conjunction with the studies of Nash et al. (1988) reveals a clear distinction between the regulation of CLN3 and the other CLN genes. In contrast to the apparent transcriptional regulation of CLN7 and CLN2, CLN3 transcript accumulation is, at most, only slightly periodic during the cell cycle and is induced rather than repressed by mating pheromone. Since any of the three CLN genes is sufficient to support cell cycle progression, it is likely that the products of all three of these genes must be regulated to control entry into S phase. This must be true to prevent cells that are growing vegetatively from initiating a new cell cycle prematurely or in response to mating pheromone. Thus, Cln3 activity must be regulated posttranscriptionally under both of these conditions. Consistent with this, preliminary evidence indicates that the level of the Cln3 protein is periodic during the cell cycle (C. W., G. Cole, and S. I. Ft., unpublished data). It has been previously suggested that such regulation might be imposed at the level of protein turnover (Hadwiger et al., 1989a; Nash et al., 1988; Richardson et al., 1989). This distinction between the CLN genes in terms of transcriptional control could explain the dramatic differences in pheromone resistance observed between the dominant mutants of CLN2 (CLNB7) and CLN3 (CLNS-7 and CLN3-2; formerly WH/7-7 and DAFI-7, respectively). While the single-copy dominant mutants of CLN3 are approximately 10 times more resistant to mating pheromone than wild-type cells, CLNP7 mutants are only a few-fold more resistant than wild-type cells. Since the CLNP transcript is rapidly lost in response to pheromone, the continued accumulation of the putative hyperstabilized and/or hyperactivated CLNP-7 gene product would not be expected to occur. However, under the same conditions the CLN3-7 transcript is actually induced, promoting the accumulation of its protein product and, as a result, cell cycle progression. On the other hand, expression of CLN7 and CLNP under control of the GAL7 promoter, which abrogates pheromone-responsive transcriptional control, results in high levels of pheromone resistance (unpublished data). Gl Cycllns Are Unstable Regulators of Cell Cycle Initiation The existence of an unstable molecule that regulates initiation of a new cell cycle has been hypothesized for a number of years based primarily on theoretical considerations (Hanic-Joyce et al., 1987; Rossow et al., 1979; Shilo et al., 1978, 1979). Such a molecule might act through its regulated accumulation to integrate information from a number of internal and external signaling systems relevant to cell cycle control. As described above, the proteins encoded by the CLN2 gene and, presumably, its functional homologs CLN7 and CLN3 have several of the properties predicted for this putative regulatory molecule. Our data

Cell 234

---r------------------, I Malllwd

msr-e

I

Mmmn, Lmtatan

, ,

? 1

Figure 8. A Model of the Regulation of Cln Proteins

and Role

Depicted is our current hypothesis of the regulation and role of the Cln proteins. We propose that the abundance of the Clnl and Cln2 proteins is regulated during the cell cycle by periodic transcription of the CLN7 and CLNZ genes coupled with the constitutive instability of the proteins. We suggest that transcription is induced, and therefore the proteins accumulate, during the late Gl interval and that they are lost as cells progress from Gi into S phase. The model also depicts the demonstrated effect of mating pheromone on the abundance of the CLN7 and CLN2 transcripts and posttranscriptional regulation of Cln protein accumulation by nutrient limitation. The dotted lines denote the possibility that these environmental factors affect the synthesis or stability of the protein. Furthermore, although the accumulation of Cln3 is thought to be periodic during the cell cycle, little if any transcriptional regulation has been demonstrated, suggesting that its accumulation or activation must be regulated posttranscriptionally. We suggest that these Gl cyclins play an essential role as activators of the p-&pC28 protein kinase toward Gl-specific substrates and that the phosphorylalion of those substrates is essential for progression into S phase, While the Gl cyclins are thought to be phosphorylated, it is unclear whether that phosphorylation is associated with their activation or ability to associate with p%~cccss, or whether it plays a role in targeting for degradation. This model also predicts the existence of M phase-specific cyclins that are essential for the passage of cells through the G2+M transition and that, like the mitotic cyclins in other systems, they act through their association with p&tcoc?s, activating that kinase toward mitosis-specific substrates. CLN oene E~essD”

1

4

CLN -

suggest that the accumulation of these unstable proteins could integrate Gl control of cell cycle initiation by controlling the activation of the ~34~~~~ protein kinase. Figure 8 presents a model that summarizes our current thinking concerning the regulation and role of the CLN gene products during the cell cycle. We propose that the function of the CLN gene products is restricted to the Gl interval and is essential for the Gl activity of the Cdc28 protein kinase. We suggest that their abundance is negatively regulated by both mating pheromone and nutrient limitation. While mating pheromone transcriptionally regulates CLN7 and CLNP, we suggest that nutrient limitation acts through global inhibition of protein synthesis. Since the Cln proteins are inherently unstable, they would be expected to be lost preferentially under such conditions. We have shown that the accumulation of CLN transcripts is also regulated as a function of position in the cell cycle, with expression restricted to the Gl interval. Our understanding of regulation of Cln3 polypeptide abundance is insufficient to allow its placement on this model. While genetic studies suggest that the activity of the CLN3 product must be subject to cell cycle- and pheromone-dependent regulation, it is not known whether that regulation occurs at the level of protein abundance or activity. The possibility that the other Gl cyclins may also be regulated at these levels is indicated by the dashed lines. The lower portion of Figure 8 arises from our demonstration that the Cdc28 protein kinase is also essential for mitotic induction (Reed and Wittenberg, 1990). We propose, by analogy to other systems, that this activity requires mitotic cyclins. This proposal is strengthened by our recent identification

of a cyclin B homolog from S. cerevisiae and S. I. Ft., unpublished data).

(l-l. Richardson

Gl Cyclins in Other Organisms? Although Gl cyclins have not been demonstrated in other organisms, the high degree of conservation of cell cycle regulatory elements suggests that they will be ubiquitous. First, histone Hl kinases of the p34ccc2mdc2 family are ubiquitous among eukaryotes. Furthermore, the genes encoding the p34 catalytic subunits are interchangeable between the fission and budding yeasts (Beach et al., 1982; Booher and Beach, 1986) as well as between yeasts and humans (Lee and Nurse, 1987; Wittenberg and Reed, 1989). In addition to the p34 kinase, a number of its associated components are conserved among organisms. An accessory subunit of p34 kinases, known as ~13, has been shown to interact with heterologous p34 subunits both in vitro and in vivo (Brizuela et al., 1987; Dunphy et al., 1988; Hadwiger et al., 1989b; Richardson et al., submitted). Furthermore, the known mitotic regulators of these kinases, including mitotic cyclins and the mitotic inducer cdc25, have been shown to be conserved from yeast to humans (Edgar and O’Farrell, 1989; Russell et al., 1989; Sadhu et al., 1990). While the mitotic function of the p34 protein kinases is clearly universal, less evidence exists for the conservation of their Gl function. In fact, direct demonstration of a requirement for ~34~~~~~ during the Gl interval has been accessible only through genetic analysis. Thus, evidence for distinct Gl*S and G2+M requirements has been established only in budding and fission yeast (Nurse

Gl-Specific 235

Cyclins

of S. cerevisiae

and Bissett, 1981; Reed and Wittenberg, 1990). However, the high degree of conservation observed between cell cycle regulatory elements despite the distant evolutionary relationship between these two yeasts suggests that similar Gl control mechanisms will be conserved throughout the eukaryotes. Furthermore, the ability of the human p34 protein to substitute for the budding yeast ~34~~~~ suggests that it can perform Gl-specific functions fulfilled by that kinase and that it can interact appropriately with Gl regulatory elements in yeast (Wittenberg and Reed, 1989). Those observations, along with the demonstration that the primary controls of cell proliferation in animal cells are exerted during the Gl interval (reviewed by Pardee, 1989), suggest that regulation analogous to the Clndependent system described here exists in those cells. Therefore, the paucity of evidence is more likely due to experimental limitations, such as the lack of Gl-specific substrates of the kinase or instability of the Gl kinase complexes, rather than a lack of conservation of Gl-specific mechanisms of cell cycle control involving p34 protein kinases. Experlmental

Procedures

Strains and Plasmlds All yeast strains used in this study were derivatives of BF264-150, a e&l his2 leu2-3,112 frppl-la (Reed et al., 1965) except as noted. The genotypes of the strains used in this study were D4, a c&28-4; D4a. a cdc28-4; CWY135, D4a (YCpGP); MMY149, BF264-15D with the chromosomal CDC28 gene replaced with GAL7P::CDC28 and TRW (Wittenberg et al., 1967); IBDau, MAR ufa3Ans derivative of BF264-15D (Cole et al., 1990); 15Dau diploid, 15Dau diploidized by introduction of a plasmid carrying the HO gene; CWY151, a wa3Ans clnl::TRPl cln2::LEU2 ch3::ura3 (YCpG2[CLN2]); JHY357, BF264-15D c/nl::TRPl; JHY359, BF264-15D c/n2::LEU2; YFCIIP, 15Dau c/n3::URA3; GCYII, 15Dau bar7::LEUP; YH129,15Dau (YCpGP), CWY145,15Dau (YCpGP[CLN2]); 196d1, ala adel/ade/l ade%de2 cdcWdc8 gall~all his77 his7 /ys,?//ys2 (C. S. McLaughlin): CG7a6, 3816 MATa cdc7-4 ade8. All cultures were grown in YEPD (1% yeast extract, 2% Bacto peptone, 2% glucose) except when selecting for plasmids and for growth of cells for elutriation. In those cases, cultures were grown in 2% galactose or 2% sucrose minimal medium supplemented with amino acids and adenine but lacking uracil or leucine as necessary. All reagents were from Sigma Chemicals or Fisher Scientific except where noted. Preparetlon of Antlserum Cln2 antiserum was prepared using Cln2 protein produced in E. coli according to the method of Rosenberg et al. (1967). Briefly, the CLN2 gene was mutagenized to introduce an Ndel site at position -3 to +3 and a BamHl site 56 nucleotides downstream of the translation termination codon using the polymerase chain reaction (Saiki et al., 1968). The CLN2 gene was placed under control of the bacteriophage T7 promoter by introduction of the resulting Ndel-BamHI fragment into the vector pRKl7l cut with the same two enzymes. The plasmid pRKl7l[CLN2] was transformed into the E. coli strain BL21(DE3) and the expression of the gene induced by induction with 0.4 mM IPTG at a culture density of Ass5 = 0.4. The Cln2 protein was enriched by preparation of an insoluble fraction as described by Kleid et al. (1961) and purified by preparative SDS-PAGE followed by electroelution. The Cln2 antiserum was affinity purified either by chromatography on bacterially produced Cln2 bound to CNBr-Sepharose CL-4B as previously described (Wittenberg et al., 1967) or by absorbtion to and elution from nitrocellulose strips containing bacterially produced Cln2 protein (Harlow and Lane, 1988). Columns and nitrocellulose strips were eluted with 0.1 M glycine (pH 2.5) and immediately neutralized with one-tenth volume of 1 M Tris-HCI (pH 6.0).

The preparation and use of p34cccza carboxy-terminal tiserum (anti-CDC28rxt-r& were as previously described et al., 1967).

peptide an(Wittenberg

Gel Electrophoresls and lmmunoblottlng Gel electrophoresis and immunoblotting were performed as previously described (Wittenberg and Reed, 1966). Except where noted, 0.6 Aaes units of cell extract was used for each sample. Northern Blotting Yeast RNA was prepared as described (Elder et al., 1963) and separated on 1% agarose gels containing formaldehyde. The RNA was then transferred to BIDTRANS (ICN) (Maniatis et al., 1962) and hybridization performed as recommended by the manufacturer. Probes were radiolabeled with [c$*P]dCTP using a random primed DNA labeling kit (Boehringer Mannheim) according to the manufacturer’s instructions. The DNA fragments used as probes were as follows: CLNl, the 0.70 kb Ncol-EcoRI fragment internal to the open reading frame (Hadwiger et al., 19893; CLN2, the 0.86 kb Xhol-Hindlll fragment internal to the open reading frame (Hadwiger et al., 19693; CLN3, the 0.45 kb Xbal-Hindlll fragment internal to the open reading frame (Cross, 1966); LEUP, the 1.42 kb Clal-Sal1 fragment derived from the plasmid YEpl3 (Nasmyth and Tatchell, 1960). Cell Cycle Synchronlzatlon by Mating Pheromone T)satment and Elutriatlon Synchronization by mating pheromone was accomplished by growing 2 liters of a bar7 yeast strain to a density of & = 0.5 in YEPD medium followed by addition of 40 nglml a factor for 200 min at 23OC. Cells were then collected and washed one time in fresh medium without mating pheromone and then inoculated into 1 liter of the same medium. Samples were taken prior to and at 10 min intervals following removal of mating pheromone for analysis of budding index and for preparation of protein and RNA samples. When necessary, pheromone treatment of BARC yeast strains was achieved by the addition of 10 ug/ml a factor. Cells used for elutriation were grown to a density of Asm = 0.5 in sucrose minimal medium to enrich the population for small Gl cells while maintaining a relatively rapid doubling time. Two liters of cells was loaded into the cell of a Beckman JE5.0 elutriation rotor at 4500 rpm with a flow rate of 75-100 mllmin at 4O C. The smallest 10% of cells in this population were collected into 1 liter of the original growth medium by decreasing the speed to 2500 rpm with a flow rate of approximately 35 mllmin, monitoring for the purity of unbudded cells by microscopic examination. Cell samples (50-75 ml) were taken at 20 min intervals for 260 min for analysis of Cln2 protein levels. Smaller samples were fixed in formaldehyde for analysis of budding index and in 70% ethanol for analysis of DNA content. Cells were analyzed for DNAcontent by flow cytometry following staining with propidium iodide as described previously (Hutter and Eipel, 1979) using a Becton Dickinson FACS IV analyzer. The proportion of S phase cells was estimated by determining the proportion of cells with a DNA content intermediate between those having a In (Gl cells) and a 2n (G2*M cells) DNA content. lmmunopmclpltatlons and Protein Klnaae Aaseys Protein kinase assays were performed as described previously (Wittenberg and Reed, 1966) except that immunoprecipitations were performed using affinity-purified Cln2 antiserum in place of anti-CDC28,-,* where indicated and all reactions were performed without the addition of 1 mM ZnC12. lmmunoprecipitations for analysis of association between Cln2 and p34coc28 were performed as described for protein kinase assays except that 20 Aso, units (instead of 6 As,,,, units) of cells was used for each immunoprecipitation. Acknowledgments The research reported here was supported by grants GM43467 to C. W. and GM38326 to S. I. R. We would like to thank Helena Richardson, Jeff Hadwiger, and Fred Cross for several of the plasmids and strains used in these studies and Gary Cole, Miguel de Barros Lopes,

Cell 236

Connie Stueland, and David Stone for helpful discussion and comments concerning the manuscript. We would also like to acknowledge the expert technical assistance of Julian Thomas.

Hadwiger, J. A., Wittenberg, C., Mendenhall, M. D., and Reed, S. I. (1989b). The Saccharomyces cerevisiae CKS7 gene, a homolog of the Schizosaccharomyces pombe sucl’ gene, encodes a subunit of the Cdc28 protein kinase complex. Mol. Cell. Biol. 9, 2034-2041.

Received

Hagan, I., Hayles, J., and Nurse, f? (1988). Cloning and sequencing of the cyclin-related cdc73+ gene and a cytological study of its role in fission yeast mitosis. J. Cell Sci. 97. 587-595.

April 2, 1990; revised

May 31, 1990

References Arion, D., Meijer, L., Brizuela, L., and Beach, D. (1988). cdc2 is a component of the M phase-specific histone Hl kinase: evidence for identity with MPF. Cell 55, 37t-378. Beach, D., Durkacz, B., and Nurse, P (1982). Functionally homologous cell cycle control genes in fission yeast and budding yeast. Nature 300, 706-709. Booher, Ft., and Beach, D. (1988). Site-specific mutagenesis of cd&, a cell cycle control gene of the fission yeast Schizosacoharornyces pombe. Mol. Cell. Biol. 6, 3523-2530. Booher, R., and Beach, D. (1988). Involvement of cdc73+and in mitotic control in Schizosacchammyces pombe: possible interaction of the gene product with microtubules. EMBO J. 6, 3441-3447. Booher, R. N., Alfa, C. E., Hyams, J. S., and Beach, D. H. (1989). The fission yeast cdc2/cdcl3/sucl protein kinase: regulation of catalytic activity and nuclear localization. Cell 58, 485-497. Breeden, L. (1988). Cell cycle-regulated Trends Genet. 4, 249-253.

promoters

in budding

yeast.

Brizuela, L., Draetta, G., and Beach, D. (1987). pl3/sucl acts in the fission yeast cell division cycle as a component of the p34/cdc2 protein kinase. EMBO J. 6, 3507-3515. Bucking-Throm, E., Duntze, W., Hartwell, L. H., and Manney, T. R. (1973). Reversible arrest of haploid yeast cells at the initiation of DNA synthesis by a diffusible sex factor. Exp. Cell Res. 76, 99-110. Carter, B. L. A., and Sudbery, P E. (1980). Small sized mutants chammyces cemvisiae. Genetics 96, 561-586.

of Sac-

Hanic-Joyce, P J., Johnston, lated arrest of cell proliferation Cell Res 772, 134-145.

G. C., and Singer, R. C. (1987). Regumediated by yeast prt7 mutations. Exp.

Harlow, E., and Lane, D. (1988). Antibodies: A Laboratory Manual (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory). Hartwell, L. H. (1976). Sequential function of gene products relative DNA synthesis in the yeast cell cycle. J. Mol. Biol. 704, 803-817. Hartwell, COntrOl

to

L. H.. Culotti, J., Pringle, J. R., and Reid, B. J. (1974). Genetic of the cell division cycle in yeast. Science 783, 46-51.

Hereford, L. M., Osley, M. A.. Ludwig, J. R.. Ill, and McLaughlin, C. S. (1981). Cell cycle regulation of yeast histone mRNA. Cell 24, 367-375. Hutter, K. J., and Eipel, H. E. (1979). Microbial DNAdeterminations flow cytometry. J. Gen. Microbial. 773, 369-375.

by

Klar, A. J. S., Strathern, J. N.. and Hicks, J. B (1984). Developmental pathways in yeast. In Microbial Development, R. Losickand L. Shapiro, eds. (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory), pp. 151-195. Kleid, D. G., Yonsura, D., Small, B., Dowbenkjo, D.. Moore, D. M., Grubman, M. J., McKercher, P D., Morgan, D. O., Robertson, B. H., and Bachrach, H. L. (1981). Cloned viral protein vaccine for foot-andmouth disease: responses in cattle and swine. Science 274, 11251129. Kronstad, J. W., Holly, J. A., and MacKay, V. L. (1987). A yeast operator overlaps an upstream activation site. Cell 50, 369-377.

Cole, G. M., Stone, D. E., and Reed, S. I. (1990). Stoichiometry of G protein subunits affects the Sacchammyres cefevisiae mating pheromone signal transduction pathway. Mol. Cell. Biol. 70, 510-517.

Labbe, J. C., Capony, J.-P, Caput, D., Cavadore, J. C.. Derancourt, J., Kaghad, M., Lelias, J.-M., Picard. A., and Doree, M. (1989). MPF from starfish oocytes at first meiotic metaphase is a heterodimer containing one molecule of cdc2 and one molecule of cyclin B. EMBO J. 8, 3053-3058.

Cross, F. (1988). DAF7, a mutant gene affecting size control, pheromone arrest and cell cycle kinetics of S. cerevisiae. Mol. Cell. Biol. 8, 4675-4884.

Lee, M. G.. and Nurse, f? (1987). Complementation used to clone a human homologue of the fission yeast cell cycle control gene cdc2 from humans. Nature 292, 558-580.

Cross, F., Hartwell, L. H., Jackson, C., and Konopka, J. B. (1988). Conjugation in Sacchammyces cerevisiae. Annu. Rev. Cell Biol. 4, 429457.

Lehner, C. F., and O’Farrsll, P H. (1989). Expression and function of Drosophila cyclin A during embryonic cell cycle progression, Cell 56, 957-968.

Draetta, G., Luca, F., Westendorf, J., Brizuela, L., Ruderman, J., and Beach, D. (1989). cdc2 protein kinase is complexed with both cyclin A and 8: evidence for proteolytic inactivation of MPF. Cell 56, 829-838.

Maniatis, T., Fritsch, E. F., and Sambrook. J. (1982). Molecular Cloning: A Laboratory Manual (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory).

Dunphy, W. G., Brizuela, L., Beach, Xenopus cdc2 protein is a component of mitosis. Cell 54, 423-431.

Meijer, L., Arion. D., Golsteyn, R., Pines, J., Brizuela, L.. Hunt, T., and Beach, D. (1989). Cyclin is a component of the sea urchin M-phase specific histone HI kinase. EMBO J. 8, 2275-2282.

Edgar, B. A., and O’Farrell, patterns in the Drosophila

D., and Newport, J. (1988). The of MPF, a cytoplasmic regulator

P H. (1989). Genetic control embryo. Cell 57, 177-187.

of cell division

Elder, R. T., Loh, E. Y.. and Davis, R. W. (1983). RNA from the yeast transposable element Tyl has both ends in the direct repeats, a structure similar to retrovirus RNA. Proc. Natl. Acad. Sci. USA 80, 24322436.

Mendenhall, M. D., Jones, C. A., and Reed, S. I. (1987). Dual regulation of the yeast CDC28-p40 protein kinase complex: cell cycle, pheromone, and nutrient limitation effects. Cell 50, 927-935. Minshull, J., Blow, J. J., and Hunt, T. (1989). Translation of cyclin mRNA is necessary for extracts of activated Xenopus eggs to enter mitosis. Cell 56, 947-956.

Evans, T, Rosenthal, E. T, Youngblom, J., Distel, D., and Hunt, T. (1983). Cyclin: a protein specified by maternal mRNA in sea urchin eggs that is destroyed at each cleavage division. Cell 33, 389-396.

Moreno, S., Nurse, cyclic accumulation ture, in press.

P., and Russell, l? (1990). Regulation of mitosis by of p8OCdC*s mitotic inducer in fission yeast. Na-

Gautier, J., Minshull, J., Lohka, M., Glotzer, M., Hunt, T., and Maller, J. L. (1990). Cyclin is a component of maturation-promoting factor from Xenopus. Cell 60, 487-494.

Murray, A., and Kirschner, M. (1989). Cyclin embryonic cell cycle. Nature 339, 275-280.

synthesis

drives

the early

Cell 54,

Murray, A., Solomon, M. J.. and Kirschner, M. W. (1989). The role of cy clin synthesis and degradation in the control of MPF activity. Nature 339, 280-286.

Gould, K., and Nurse, f? (1989). Tyrosine phosphorylation of the fission yeast cdc.?’ protein kinase regulates entry into mitosis. Nature 342, 39-46.

Nash, R., Tokiwa, G., Anand, S., Erickson, K., and Futcher, A. 8. (1988). The WHI7+ gene of S. cerevisiae tethers cell division to cell size and is a cyclin homolog. EMBO J: 7, 4335-4346.

Hadwiger, J. A., Wittenberg. C., de Barros Lopes, M. A., Richardson, H. E., and Reed, S. I. (1989a). A family of cyclin homologs that control G, phase in yeast. Proc. Natt. Acad. Sci. USA 86, 6255-6259.

Nasmyth, K. (1985). At least 1400 base pairs of 5’-flanking DNA is required for the correct expression of the HO gene in yeast. Cell 42, 213-223.

Goebl, M., and 739-740.

Byers,

B. (1988).

Cyclin

in fission

yeast.

Gl-Specific 237

Cyclins

of S. cerevisiae

Nasmyth, K. A., and Tatchell, K. (1980). The structure yeast mating type loci. Cell 19, 753-764.

of transposable

Slater, M. L. (1983). Effect of reversible acid synthesis on the yeast cell cycle.

Nasmyth, K., Seddon, A., and Ammerer, G. (1987). Cell cycle regulation of SW15 is required for mother-cell-specific HO transcription in yeast. Cell 49, 549-558. Nurse, P., and Bissett, Y. (1981). Gene required in Gl for commitment to the cell cycle and in G2 for control of mitosis in fission yeast. Nature 292, 558-560. Pardee, Science

A. B. (1989). Gl 246, 603-608.

events

and regulation

of cell proliferation.

Solomon, in fission

M., Booher, R., Kirschner, yeast. Cell 54, 738-739.

inhibition of deoxyribonucleic J. Bacterial. 173, 283-270.

M., and Beach,

D. (1988). Cyclin

Standart, N., Minshull, J., Pines, J., and Hunt, T. (1987). Cyclin synthesis, modification and destruction during meiotic maturation of the starfish oocyte. Dev. Biol. 724, 248-258. Stern, M., Jensen, R., and Herskowitz, I. (1984). Five SW/ genes are required for expression of the HO gene in yeast. J. Mol. Biol. 778, 853-888.

Piggott, J. Ft., Rai, R., and Carter, B. L. A. (1982). A bifunctional gene product involved in two phases of the yeast cell cycle. Nature 298, 391-393.

Stetler, G. L., and Thorner, J. (1984). Molecular cloning of hormoneresponsive genes from the yeast Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 81, 1144-1148.

Pilus, L., and Solomon, tures in Saccheromyces 2468-2472.

Sudbery, P E., Goodey, control cell proliferation 288, 401-404.

F. (1988). Components of microtubular struccerevisiae. Proc. Natl. Acad. Sci. USA 83,

Pines, J., and Hunt, T. (1987). Molecular cloning and characterization of the mRNA for cyclin from sea urchin eggs. EMBO J. 6, 2987-2995. Pines, J., and Hunter, T. (1989). Isolation of a human cyclin cDNA: evidence for cyclin mRNA and protein regulation in the cell cycle and for interaction with p34cdc*. Cell 58, 833-848. Pondaven, P., Meijer, L., and Beach, D. (1990). Activation of M-phase specific histone HI kinase by modification of the phosphorylation of its p34cdc2 and cyclin components. Genes Dev. 4, 9-17. Pringle, J. P., and Hartwell, L. H. (1981). The Saccharomyces cerevisiae cell cycle. In The Molecular Biology of the Yeast Saccharomyces, J. D. Strathern, E. W. Jones, and J. R. Broach, eds. (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory), pp. 97-142. Reed, S. I. (1980). The selection of S. cerevisiae mutants the start event of cell division. Genetics 95, 561-577. Reed, S. I., and Wittenberg, tein kinase of S. cerevisiae.

defective

in

C. (1990). A mitotic role for the Cdc28 proProc. Natl. Acad. Sci. USA, in press.

Reed, S. I., Hadwiger, J. A.. and Lorincz, A. T. (1985). Protein kinase activity associated with the product of the yeast cell division cycle gene CDCPB. Proc. Natl. Acad. Sci. USA 82, 4055-4059. Reid, 8. J., and Hartwell, cycle of Saccharomyces

L. H. (1977). Regulation of mating in the cell cerevisiae. J. Cell Biol. 75, 355-385.

Richardson, H. E., Wittenberg, C., Cross, F., and Reed, S. I, (1989). An essential Gl function for cyclin-like proteins in yeast. Cell 59, 11271133. Rogers, quences Science

S., Wells, R., and Rechsteiner, M. (1988). Amino acid secommon to rapidly degraded proteins: the PEST hypothesis. 234, 364-368.

Rosenberg, A. H., Lade, B. N., Chui, D. S., Lin, S. W., Dunn, J. J., and Studier, F. W. (1987). Vectors for selective expression of cloned DNAs by T7 RNA polymerase. Gene 56, 125-135. Rossow, P W., Riddle, V. G. H., and Pardee, A. B. (1979). Synthesis of labile, serum dependent protein in early Gl controls animal cell growth. Proc. Natl. Acad. Sci. USA 76, 4448-4450. Russell, P., and Nurse, l? (1986). c&25+ functions as an inducer the mitotic control of fission yeasts. Cell 45, 145-153. Russell, controls

P., Moreno, S., and Reed, S. I. (1989). Conservation in fission and budding yeast. Cell 57, 295-303.

in

of mitotic

Sadhu, K., Reed, S. I., Richardson, H. E., and Russell, P. (1990). cdc25Hs. a putative mitotic inducer gene in human cells, is predominantly expressed in G2. Proc. Natl. Acad. Sci. USA, in press. Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T, Mullis, K. B., and Erlich, H. A. (1988). Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239, 487-491. Sclafani, R. A., and Fangman, W. L. (1984). Yeast gene CDCB encodes thymidylate kinase and is complemented by the herpes thymidine kinase gene TK. Proc. Natl. Acad. Sci. USA 87, 5821-5825. Shilo, 8.. Simchen, G., and Pardee, A. B. (1978). Regulation of cellcycle initiation in yeast by nutrients and protein synthesis. J. Cell. Physiol. 97, 177-188. Shilo, B., Riddle, V. G. H., and Pardee, A. B. (1979). Protein turnover and cell-cycle initiation in yeast. Exp. Cell Res. 123, 221-227.

A. R., and Carter, B. L. A. (1980). Genes which in the yeast Sacchatumes cefevisiae. Nature

Swenson, K. I., Farrell, K. M., and Ruderman, J. V. (1986). The clam embryo protein cyclin A induces entry into M phase and the resump tion of meiosis in Xenopus oocytes. Cell 47, 861-870. Thomas, J. H., Neff, N. F., and Botstein, acterization of mutations in the 6-tubulin ics 772, 7t5-734.

D. (1985). Isolation and chargene of S. cerevisiae. Genet-

Van Arsdell, S. W., and Thorner, J. (1987). Hormonal regulation of gene expression in yeast. In Transcriptional Control Mechanisms, G. D. Granner, M. G. Rosenfeld, and S. Cheng, eds. (New York: Alan R. Liss), pp. 325-332. Westendorf, J. M., Swenson, K. I., and Ruderman, J. V. (1989). role of cyclin B in meiosis I. J. Cell Biol. 708, 1431-1444. Wittenberg, C., and Reed, S. I. (1988). is associated with assembly/disassembly complex. Cell 54, 1081-1072.

The

Control of the yeast cell cycle of the Cdc28 protein kinase

Wittenberg, C., and Reed, S. I. (1989). Conservation of function and regulation within the Cdc281cdc2 protein kinase family: characterization of the human Cdc2Hs protein kinase in Saccharomyces cerevisiae. Mol. Cell. Biol. 9, 4064-4068. Wittenberg, C., Richardson, S. L., and Reed, S. I. (1987). Subcellular localization of a protein kinase required for cell cycle initiation in Saccharomyces cefevisiae: evidence for an association between the CDC28 gene product and the insoluble cytoplasmic matrix. J. Cell Biol. 705, 1527-1538.