A Novel Yeast Screen for Mitotic Arrest Mutants Identifies DOC1, a New Gene Involved in Cyclin Proteolysis

Abstract

B-type cyclins are rapidly degraded at the transition between metaphase and anaphase and their ubiquitin-mediated proteolysis is required for cells to exit mitosis. We used a novel enrichment to isolate new budding mutants that arrest the cell cycle in mitosis. Most of these mutants lie in the CDC16, CDC23, and CDC27 genes, which have already been shown to play a role in cyclin proteolysis and encode components of a 20S complex (called the cyclosome or anaphase promoting complex) that ubiquitinates mitotic cyclins. We show that mutations in CDC26 and a novel gene, DOC1, also prevent mitotic cyclin proteolysis. Mutants in either gene arrest as large budded cells with high levels of the major mitotic cyclin (Clb2) protein at 37°C and cannot degrade Clb2 in G₁-arrested cells. Cdc26 associates in vivo with Doc1, Cdc16, Cdc23, and Cdc27. In addition, the majority of Doc1 cosediments at 20S with Cdc27 in a sucrose gradient, indicating that Cdc26 and Doc1 are components of the anaphase promoting complex.

INTRODUCTION

Oscillations in the activity of cyclin-dependent kinases (Cdk) control the eukaryotic cell cycle (reviewed in Morgan, 1995). Multicellular eukaryotes express a number of Cdks with different activities through the cell cycle. Yeasts, however, express a single Cdk, called Cdc28 in budding yeast (Saccharomyces cerevisiae) and Cdc2 in fission yeast (Schizosaccharomyces pombe). The oscillation of Cdk activity in all eukaryotes depends on phosphorylations of the kinase subunit and the cyclic expression and degradation of cyclins, the regulatory subunits of Cdks. Entry into mitosis requires the expression of B-type (mitotic) cyclins, and exit from mitosis requires their degradation (reviewed in Murray and Kirschner, 1989; Nurse, 1990). The Cdk activity associated with B-type cyclins is required for chromosome condensation and proper formation of the mitotic spindle. The degradation of B-type cyclins, which leads to inactivation of the mitotic Cdk, results in chromosome decondensation, breakdown of the mitotic spindle, and cell division.

A short N-terminal sequence found in A- and B-type cyclins, termed the destruction box, is required for their proteolysis (Glotzer et al., 1991). Deletion or mutation of this sequence produces a nondegradable cyclin, and cells or extracts expressing these nondegradable B-type cyclins are unable to exit mitosis and remain arrested in anaphase (Murray et al., 1989; Holloway et al., 1993; Surana et al., 1993; Sigrist et al., 1995). In contrast, inhibiting the machinery that is required for cyclin proteolysis arrests cells in metaphase with unseparated sister chromatids (Holloway et al., 1993). This observation suggests that additional proteins, whose destruction is required for the separation of sister chromatids, must also be targets of the machinery that degrades cyclins. Candidates for these regulators of sister cohesion have recently been identified (Funabiki et al., 1996; Yamamoto et al., 1996).

The degradation of cyclin is regulated by the assembly of multiubiquitin chains onto the cyclin substrate that targets it for proteolysis by the proteasome. In frog egg extracts, mutations in cyclins that block their degradation also block ubiquitination (Glotzer et al., 1991; King et al., 1995), and methylated ubiquitin, an inhibitor of multiubiquitin chain formation, inhibits the degradation of cyclins in clam egg extracts (Hershko et al., 1991). In yeast, a number of mutants that block cyclin degradation also block ubiquitination (Zachariae and Nasmyth, 1996).

Ubiquitination of proteins begins with the activation of ubiquitin by a ubiquitin-activating enzyme (E1). The E1 forms a thiol ester with ubiquitin and then transfers that ubiquitin onto a ubiquitin-carrier protein (E2). Some substrates are directly ubiquitinated by an E2. Other substrates, however, also require a ubiquitin-protein ligase (E3), which appears to provide a higher level of specificity to the degradation system. Finally, the 26S proteasome recognizes multiubiquitinated substrates and degrades them (reviewed in Ciechanover, 1994).

In budding yeast, clam, and frog, a nonspecific E1 is capable of activating ubiquitin. A large family of E2s (UBCs) exist in yeast, and the E2 required for Clb degradation has not yet been definitively identified. E2-C (clam), homologues of the budding yeast UBC4, and a novel UBC, UBCX (frog) are capable of acting as E2s for cyclin B in reconstituted cyclin ubiquitination systems derived from clam and frog egg extracts (King et al., 1995; Sudakin et al., 1995). The E3 for B-type cyclins was first identified as a 20S complex in clam and frog egg extracts, known as the cyclosome or anaphase-promoting complex (APC; King et al., 1995; Sudakin et al., 1995). In frog egg extracts, yeast homologues of CDC16 and CDC27 were identified as components of this E3 complex, which appears to be active only in mitosis in both clams and frogs.

CDC16, CDC23, and CDC27 are all members of a family of tetratricopeptide repeat proteins and form a complex in yeast that is required for the ubiquitination and degradation of Clb2, the major mitotic cyclin (Lamb et al., 1994; Irniger et al., 1995; Zachariae and Nasmyth, 1996). Temperature-sensitive mutants in any of these genes cause a metaphase arrest with high levels of Clb2 and unseparated sister chromosomes. Another gene, CSE1, has also been implicated in cyclin ubiquitination (Irniger et al., 1995). Mutations in CDC16, CDC23, and CSE1 all result in increased chromosome loss, suggesting that proper regulation of the cyclin proteolysis machinery is required for faithful chromosome segregation (Hartwell and Smith, 1985; Xiao et al., 1993).

We developed an enrichment for cell cycle arrest mutants in budding yeast and used this to screen for new mitotic arrest mutants. We isolated a large number of alleles of previously identified cdc (cell division cycle) mutants, and identified a novel mutant, doc1 (destruction of cyclin B). CDC26 and DOC1 are involved in the degradation of Clb2 and the products of both genes associate with the yeast APC.

MATERIALS AND METHODS

Yeast Strains, Media, and Genetic Methods

All yeast strains are derivatives of W303, except those used for complementation analysis. All strains used in this work are listed in Table 1. Yeast media and genetic manipulations were as described (Sherman et al., 1974). LH103 (mec1–1) was made by backcrossing AFS85 (MATa, mec1–1, GAL-CLN3::URA3::cln3Δ, ssd1-v1::LEU2, leu2–3,112, ura3, his3–11,15,ade2–1, trp1–1) four times into W303 and subsequently crossing in mad1Δ::HIS3 (from KH123, Hardwick and Murray, 1995) and bar1 (from AFS92, kindly provided by Aaron Straight, University of California, San Francisco). BglII-cut pLH17 (CLB2-lacZ) was integrated at CLB2, recombination and loss of the URA3 marker were selected for on 5-fluoroorotic acid medium, and retention of CLB2-lacZ was screened for by using β-galactosidase (β-gal) plate assays. Strains containing GAL-CLB2 were made by integrating pDK27 (kindly provided by Doug Kellogg, University of California, Santa Cruz) at URA3.

Table 1.

Yeast strains

Strain	Relevant genotype	Source
LH103	MATa mec1-1 mad1Δ::HIS3 clb2::CLB2-LacZ bar1	This work
	YCp50/MAD1
AFS34	MATa ade2-1 can1-100 ura3-1 leu2-3,112 his3-11,15 trp1-1	R. Rothstein
(W303-1a)
YPH218	MATα cdc16-1	P. Hieter
YPH221	MATα cdc23-1	P. Hieter
LH127	MATα cdc20-1 ura3-1 leu2-3,112 trp1-1	This work
H160-3-3	MATα cdc27-1	L. Hartwell
H152-4-2	MATα cdc26-1	L. Hartwell
ELW65-93	MATa cdc31-2 YCp50/CDC31	M. Winey
CMY763	MATα cim3-1	C. Mann
CMY765	MATα cim5-1	C. Mann
LH202	MATa cim5-1 bar1	This work
LH103-15	MATa doc1-1 mec101 mad1Δ::HIS3 clb2::CLB2-lacZ bar1	This work
	YCp50/MAD1
LH103-66	MATa cdc26-100 mec1-1 mad1Δ::HIS3 clb2::CLB2-lacZ bar1	This work
	YCp50/MAD1
LH225	MATa cdc26-100 clb2::CLB2-lacZ bar1	This work
LH226	MATa doc1-1 clb2::CLB2-lacZ bar1	This work
ADR58	MATa pDK27 (GAL-CLB2)	This work
ADR103	MATa cdc16-1 pDK27	This work
LH227	MATa cdc26-100 bar1 pDK27	This work
LH228	MATa doc1-1 bar1 pDK27	This work
LH229	MATa cdc26-100 bar1 ura3-1::CDC26myc-URA3 pDK27	This work
LH230	MATa doc1-1 bar1 ura3-1::DOC1myc-URA3 pDK27	This work
LH209	MATa bar1 pUb-R-βgal	This work
LH210	MATa bar1 pUbV76-V-eΔK-βgal	This work
LH211	MATa doc1-1 bar1 pUb-R-βgal	This work
LH212	MATa doc1-1 bar1 pUbV76-V-eΔK-βgal	This work
LH213	MATa cdc26-100 bar1 pUb-R-βgal	This work
LH214	MATa cdc26-100 bar1 pUbV76-V-eΔK-βgal	This work
LH215	MATa cim3-1 bar1 pUb-R-βgal	This work
LH216	MATa cim3-1 bar1 pUbV76-V-eΔK-βgal	This work
LH231	MATa bar1 pWAM10 (CDC16HA)	This work
LH232	MATa bar1 pRS239 (CDC23HA)	This work
LH233	MATa bar1 pJL25 (CDC27HA)	This work
LH234	MATa bar1 pLH32 (DOC1HA)	This work
LH235	MATa bar1 cdc26Δ::URA3 pWAM10	This work
LH236	MATa bar1 cdc26Δ::URA3 pRS239	This work
LH237	MATa bar1 cdc26Δ::URA3 pJL25	This work
LH238	MATa bar1 cdc26Δ::URA3 pLH32	This work
LH297	MATa bar1 doc1Δ::URA3 leu2-3,112::DOC13XHA-LEU2	This work
	pJL25

All strains from this work are W303 (R. Rothstein). 

Hydroxyurea (Sigma, St. Louis, MO) was used at 10 mg/ml, final concentration, in medium. α-Factor (Bio-synthesis, Lewisville, TX) was used at 1 μg/ml for bar1⁻ strains and 10 μg/ml for BAR1 strains, from a stock solution at 10 mg/ml in dimethyl sulfoxide (Aldrich, Milwaukee, WI). Nocodazole (Sigma) was used at 15 μg/ml from a stock solution of 10 mg/ml in dimethyl sulfoxide. Cycloheximide (Sigma) was used at 10 μg/ml from a stock solution of 10 mg/ml. Nocodazole treatments were carried out at 23°C. The other treatments were performed at various temperatures.

Plasmid Constructions

For CLB2-lacZ (pLH17), CLB2 (including the open reading frame [ORF] and 301 bp upstream) was amplified by polymerase chain reaction (PCR) using oligomers containing BamHI and SalI sites at 5′ and 3′ ends, respectively. This was ligated into pRS304 (Sikorski and Hieter, 1989) and then cut with SalI and KpnI. The KpnI site was blunted with T4 polymerase. lacZ was cut from pAFS35 (kindly provided by Aaron Straight) by using SalI and BamHI. The BamHI site was blunted by using the Klenow fragment of DNA polymerase I. The lacZ fragment was then ligated into the CLB2 construct. A second BamHI site was created 3′ of lacZ by ligation of the blunted BamHI and KpnI sites. A 270-bp fragment of sequence immediately 3′ of the CLB2 ORF was amplified by PCR with BamHI and SalI sites 5′ and 3′, respectively. This was ligated into pRS306 (Sikorski and Hieter, 1989). The BamHI fragment containing CLB2-lacZ was then ligated into the pRS306 construct containing the 3′ flanking region of CLB2. The Clb2-LacZ protein is functional and capable of acting as the sole mitotic cyclin since pLH17 rescues the temperature sensitivity of a strain deleted for CLB1, CLB3, and CLB4 with CLB2 replaced with a temperature sensitive allele, K3080 (Amon et al., 1993).

pLH25 was made by PCR amplifying the ORF of DOC1 and 508 bp upstream with a SalI site 5′ and an NcoI site 3′ to the ORF. The PCR product was ligated into pKH511 (kindly provided by Kevin Hardwick, University of Edinburgh), to place a single myc tag at the C terminus of the protein. To make pLH23, pLH25 was cut with EcoRV and BamHI and the 1.0-kb fragment was ligated into pDK20 (kindly provided by Doug Kellogg) cut with SmaI and BamHI, to place the DOC1 ORF behind the GAL1–10 promoter. pLH24 was made by cutting pLH25 with SalI and BamHI and ligating the 1.5-kb fragment into YIplac211, cut with the same enzymes. pLH26–1B was made by cutting T1 (the original rescuing plasmid from the YCp50 library) with NdeI and blunting with Klenow. For pLH26–1A, URA3 was cut from pLH3 with XhoI and XbaI and also blunted with Klenow. The URA3 fragment was ligated into the blunted NdeI site in T1.

To make pLH32, the ORF of DOC1 with 508 bp upstream was amplified by PCR with a SalI site 5′ and a KpnI 3′ to the ORF. The KpnI site was designed so that the stop codon of DOC1 was excluded and the 3′ end of DOC1 would be in-frame with a triple hemaglutinin (3×HA) tag in YCplac111–3×HA. The DOC1 PCR product was cut with SalI, blunted with Klenow, and then cut with KpnI. YCplac111–3×HA was cut with EcoRI, blunted with Klenow, and then cut with KpnI. The PCR fragment was then ligated into YCplac111–3×HA. pLH59 (DOC1–3×HA-LEU2) was made by cutting pLH32 with SphI and SpeI and ligating the 1.8-kb fragment into YIplac128 cut with SphI and XbaI. All enzymes were from New England Biolabs (Beverly, MA) and used according to the manufacturer’s specifications.

Plasmids expressing Ub-R-βgal and Ub^V76-V-e^ΔK-βgal were kindly provided by Dr. Erica Johnson (Rockefeller University, NY). Plasmids containing HA-tagged CDC16 (pWAM10), CDC23 (pRS239), and CDC27 (pRS248) were kindly provided by Dr. Phillip Heiter (Johns Hopkins University, MD).

Mutant Isolation

Strain LH103 (8 × 10⁶ cells) was mutagenized with ethyl methanesulfonate (Sigma) to 50% killing. The mutagenized cells were diluted 1:25 and allowed to recover for 12 h at room temperature in YPD. They were shifted to 37°C, to prearrest potential G₁ and mitotic arrest mutants, for 2 h, then hydroxyurea was added to 10 mg/ml, and cells were incubated at 37°C for an additional 5 h. This culture was then plated on YPD plates and incubated at 23°C. Surviving colonies were replica plated onto YPD plates in duplicate with one set at 23°C and the other at 37°C to test for temperature sensitivity.

Clb2-LacZ and Visual Screen

Temperature-sensitive mutants were patched onto YPD plates and allowed to grow overnight at 23°C. They were then replica plated to YPD and placed at 37°C for 4 h, replica plated again to Whatman filters (VWR, San Francisco, CA) on YPD plates that contained 1 μg/ml α-factor, and returned to 37°C for 5 h. The filters were assayed for β-gal activity by freezing them in liquid nitrogen, thawing them, and incubating them on Whatman paper soaked in Z buffer plus 5-bromo-4-chloro-3-indolyl β-d-galactoside, US Biological, Swampscott, MA; 60 mM Na₂HPO₄, 40 mM NaH₂PO₄, 10 mM KCl, 1 mM MgSO₄, 0.03% 5-bromo-4-chloro-3-indolyl β-d-galactoside) at 30°C overnight.

Strains that produced bright blue patches were further analyzed by microscopy. The strains were grown in liquid medium to logarithmic phase at 23°C and then shifted to 37°C. Cells were taken and fixed in 3.7% formaldehyde (Fisher, Santa Clara, CA) after 3 and 6 h at 37°C. They were stained with 4,6-diamidino-2-phenylindole (1 μg/ml) and anti-tubulin (YOL1/34, Accurate Chemical and Scientific, Westbury, NY) diluted 1:200, and their arrest phenotype was determined by light microscopy.

Mitotic arrest mutants were tested for complementation against a collection of known mitotic mutants. Only two mutants were not identified as previously isolated mutants. One was cloned and named DOC1, and the other (two alleles) has not been cloned and appears to arrest in mitosis with pleiotropic defects. One allele of cdc14 was also isolated; however, it was isolated as a double mutant with a cdc16 mutant. CDC14 codes for a phosphatase, and mutants in this gene arrest in anaphase of the cell cycle (Wan et al., 1992).

Cloning of DOC1

A YCp50 library (described in Hardwick and Murray, 1995) was transformed into doc1–1 by lithium acetate transformation and plated on −Ura medium at 37°C. Plasmids that were capable of rescuing the temperature sensitivity were recovered and the inserts were sequenced at both 5′ and 3′ ends. These sequences were used to probe the yeast genome database.

Preparation of Antibodies against Cdc26, Western Blot Analysis, and Immunoprecipitations

The ORF of CDC26 was amplified by PCR with oligomers containing BamHI and EcoRI sites 5′ and 3′ to the ORF, respectively. This fragment was cloned into pGEX-1 cut with BamHI and EcoRI (Smith and Johnson, 1988), forming a glutathione S-transferase (GST) fusion construct (pLH29). pLH29 was transformed into Escherichia coli strain TG1 (Maniatis et al., 1982) and induced for expression with 0.1 mM isopropyl β-d-thiogalactoside (US Biologicals) for 2 h at 37°C. Cells were pelleted, washed with PBS (140 mM Na₂HPO₄, 1.8 mM KH₂PO₄, 138 mM NaCl, 2.7 mM KCl, pH 7.2), pelleted again, and frozen in liquid nitrogen. The frozen pellet was resuspended in five volumes of PBS containing 1 mM EDTA, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, and 200 μg/ml lysozyme and then briefly sonicated. Triton X-100 was added to 0.5%, dithiothreitol was added to 15 mM, and cells were sonicated again. The lysate was spun at 15,000 rpm in an SS34 rotor (Sorvall, Burbank, CA) for 30 min, and then the supernatant was transferred to a new tube and spun again for 20 min. The cleared supernatant was then loaded onto a 5-ml glutathione-agarose column (Sigma) that was then washed with PBS containing 0.5 M NaCl, 1 mM EDTA, and 1 mM EGTA. The Cdc26-GST fusion protein was eluted with 5 mM reduced glutathione (Sigma) in 50 mM Tris(hydroxymethyl)aminomethane, pH 8.1. The peak fractions were pooled and dialyzed into 50 mM HEPES, pH 7.6, 50 mM KCl, and 30% glycerol. The protein was sent to Berkeley Antibody (Berkeley, CA) where it was used to immunize a rabbit. The rabbit serum was passed over a 50-ml column of GST protein coupled to Affi-Gel 10 (Bio-Rad Labs, Hercules, CA) to remove anti-GST antibodies and then affinity purified by using a 5-ml column of the Cdc26-GST fusion protein coupled to Affi-Gel 10 (Bio-Rad).

Yeast extracts for immunoblotting and immunoprecipitations were made by bead beating cells for two 90 s periods, in lysis buffer A (50 mM HEPES, pH 7.6, 75 mM KCl, 1 mM MgCl₂, 1 mM EGTA, 0.1% Nonidet P-40, 50 mM NaF, 100 μM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride, and LPC (10 μg/ml each leupeptin, pepstatin, and chymostatin; Boehringer, Indianapolis, IN). The lysates were spun briefly to separate beads from the lysate and then cleared by centrifugation for 5 min in an Eppendorf microcentrifuge at 4°C. This step was repeated three times. Protein concentrations of extracts were determined with the Bradford assay using bovine serum albumin as a standard (Bradford, 1976). For immunoblotting, the extracts were adjusted to the same protein concentration and diluted with 2× SDS sample buffer (1× SDS sample buffer: 80 mM Tris(hydroxymethyl)aminomethane, pH 6.8, 2% SDS, 10% glycerol, 10 mM EDTA, 0.0013% bromophenol blue, 5% 2-mercaptoethanol). Standard methods were used for SDS-PAGE and protein transfer to nitrocellulose (Harlow and Lane, 1988). Blots were stained with Ponceau S (Fisher) to confirm transfer and equal protein loading and then blocked for 30 min with blotto (4% dried milk, PBS, 0.2% Tween 20). Antibodies were used at a 1:1200 dilution for anti-Clb2, a 1:1000 dilution for 12CA5 (BABCO, Berkeley, CA), a 1:1000 dilution for anti-β-gal (Cappel, Durham, NC), and a 1:1000 dilution for anti-Cdc26 at either room temperature for 1 h or 4°C overnight. Blots were washed three times 10 min in PBS with 0.2% Tween 20 (PBST) and then incubated with horseradish peroxidase-conjugated anti-rabbit or anti-mouse antibodies (Amersham, Buckinghamshire, United Kingdom) at a 1:5000 dilution in PBST. They were washed again and developed by using Amersham ECL detection reagents following manufacturer’s instructions. Cdc26 protein is not detectable by Western blot, likely due to an inability to bind to nitrocellulose.

For immunoprecipitations, antibodies were used at a 1:33 dilution for 12CA5 and a 1:50 dilution for Cdc26. Lysates with primary antibody were rotated for 1 h at 4°C, then transferred to a tube containing protein A-Sepharose beads (Pharmacia, Pleasant Hill, CA), and rotated at 4°C for an additional hour. The beads were washed twice with lysis buffer A, transferred to a new tube, washed again with PBST, and then resuspended in SDS sample buffer.

Sucrose Gradient Analysis

Solutions with lysis buffer A were prepared containing either 5% or 40% sucrose. Additional solutions with 13.75%, 22.5%, and 31.25% sucrose were made by combining various amounts of the 5% and 40% solutions. Nine hundred fifty microliters of each solution were layered into Beckman Ultra-Clear Centrifuge tubes (Beckman, Palo Alto, CA), size 0.5 × 2 inches. The gradient was incubated 12–16 h at 4°C. Yeast extract (100 μl), made as described above, was layered onto the gradient and spun at 50,000 rpm in a SW-55 rotor (Sorvall) for 4 h at 4°C. Sixteen 300-μl fractions were taken, the fractions were precipitated with 10% trichloroacetic acid, loaded onto SDS-Page gels, and analyzed by immunoblotting. Molecular weight markers were analyzed by Coomassie blue staining (Fisher).

RESULTS

Screening for New Mitotic cdc Mutants

We designed a strategy that enabled us to enrich for temperature-sensitive budding yeast mutants that arrest in G₁ or mitosis (Figure 1). We used mec1–1, a cell cycle checkpoint mutant, to devise conditions under which cycling cells would die and arrested ones would survive. mec1–1 mutants lack the checkpoint that detects unreplicated DNA and DNA damage and cannot arrest their cell cycle in the presence of unreplicated DNA (Weinert, 1992). As a result, treating cycling mec1–1 cells with DNA replication inhibitors, such as hydroxyurea, induces rapid cell death. In contrast, treating G₁ or mitotically arrested mec1–1 cells with hydroxyurea will not kill them. We exploited this difference to enrich for mutants that caused a temperature-sensitive arrest in G₁ or mitosis. This protocol does not enrich for most mutants arrested in S or G₂ phase, because their arrest requires the MEC1-dependent checkpoint. A mec1–1 strain was mutagenized and the mutagenized cells were shifted to 37°C to allow potential cell division cycle mutants to prearrest and then treated with hydroxyurea to kill the majority of cells that continued to progress through the cell cycle. Cells that arrest in either G₁ or mitosis remain viable during the hydroxyurea treatment and form colonies on plating at the permissive temperature of 23°C.

Figure 1.

Schematic representation of the mutant enrichment strategy. Hydroxyurea, a drug that inhibits DNA synthesis, arrests wild-type cells in S phase. Mutants that are defective in the checkpoint gene MEC1 are unable to arrest in response to unreplicated DNA and die in the presence of hydroxyurea. If mec1–1 cells are prearrested in G₁ or mitosis, they remain viable after exposure to hydroxyurea.

We mutagenized 8 × 10⁶ cells to 50% survival with ethyl methanesulfonate, and 40,000 survived the mec1–1 enrichment. Of those, 1968 (5%) were temperature sensitive. To distinguish mutants that arrested in mitosis from those that arrested in G₁, we exploited the observation that mitotic cyclins are strongly expressed in mitosis but rapidly degraded as cells exit mitosis and enter G₁ (Amon et al., 1994). The temperature-sensitive mutants were tested for expression of Clb2-lacZ, a protein fusion between the major mitotic cyclin (Clb2) and β-gal, whose activity can be easily monitored by exposing cells to a chromogenic substrate. Cells from mutants that expressed Clb2-lacZ strongly at 37°C (see MATERIALS AND METHODS) were screened visually for their arrest phenotypes. We found 32 mutants that arrested with large budded cells at 37°C, a phenotype that is consistent with a mitotic arrest. These mutants fell into 11 complementation groups. Each group was tested against known mitotic mutants by complementation of temperature sensitivity and for rescue of temperature sensitivity by transformation with plasmids containing wild-type copies of genes known to function in mitosis.

We isolated mitotic arrest mutations in nine genes (Table 2). One, CDC31 (one allele) is involved in spindle pole body duplication and cdc31^ts mutants arrest in mitosis with only one spindle pole body (Byers, 1981; Baum et al., 1986). Another, PRP22 (two alleles) codes for an RNA helicase involved in mRNA splicing (Company et al., 1991). The alleles of PRP22 isolated in this screen arrests in mitosis with no microtubules (our unpublished results). TUB1 and TUB3, the genes that code for α-tubulin, each contain an intron. It is likely that the tubulin defect seen in these mutants is due to a failure to splice the mRNAs from these genes that leads to a deficiency in α-tubulin.

Table 2.

Mitotic-arrest mutants isolated by mec1-1 enrichment

Gene	No. of alleles isolated	Biochemical function	Reference
CDC16	16	APC component	Irniger et al. (1995), King et al. (1995)
CDC23	3	APC component	Irniger et al. (1995)
CDC27	4	APC component	King et al. (1995)
CDC26	1	APC component	This work
DOC1	1	APC associated	This work
CIM3	1	Proteasome component	Ghislain et al. (1993)
CDC20	1	Unknown	Sethi et al. (1991)
CDC31	1	SPB component	Baum et al. (1986)
PRP22	2	Splicing	Company et al. (1991)
Uncloned	2	Unknown

Mutants were identified based on complementation of temperature sensitivity with previously identified mutants. They were also tested for rescue of temperature sensitivity by wild-type copies of known genes on plasmids. 

Five of the genes we identified are involved in the proteolysis of cyclin that triggers the exit from mitosis. CDC16 (16 alleles), CDC23 (3 alleles), and CDC27 (4 alleles) encode proteins that are involved in cyclin proteolysis and have been shown to form a multiprotein complex in yeast, frogs, and clams (Lamb et al., 1994; Irniger et al., 1995; King et al., 1995; Sudakin et al., 1995). CDC20 (1 allele) has also been implicated in cyclin proteolysis. Its homologue in Drosophila, fizzy, appears to be required for the proteolysis of cyclin A and B (Dawson et al., 1995; Sigrist et al., 1995). In yeast, cdc20 mutants arrest in metaphase and appear to have microtubule abnormalities (Byers and Goetsch, 1974; Palmer et al., 1989; Sethi et al., 1991). CIM3 (1 allele) is a subunit of the 26S proteasome, the multiprotein complex that degrades ubiquitinated proteins (Ghislain et al., 1993).

Finally, we isolated one allele each of CDC26 and a novel gene that we named DOC1 (destruction of cyclin B). CDC26 is a mitotic arrest mutant, identified in the original Hartwell screen for cdc mutants (Hartwell et al., 1970). The CDC26 gene is essential only at 37°C; cdc26Δ cells grow well at 23°C (Araki et al.., 1992), a finding that we confirmed in the W303 strain background.

CDC26 and DOC1 Are Required for Clb2 Proteolysis

Because the majority of mutants isolated in our screen identify genes involved in mitotic cyclin proteolysis, we tested the cdc26 and doc1 mutants for defects in this process. To determine whether CDC26 and DOC1 were involved in cyclin proteolysis, we exploited the observation that cyclin proteolysis begins in late mitosis and persists well into G₁ (Amon et al., 1994). Haploid yeast cells can be arrested in G₁ by treatment with mating pheromone. Thus in MATa cells arrested in G₁ by α-factor treatment, the B-type cyclin proteolysis machinery is active and ectopic expression of the CLB2 gene does not lead to Clb2 protein accumulation. In a cdc16–1 mutant, however, Clb2 protein does accumulate in G₁-arrested cells as a result of a defect in the ubiquitination of Clb2 (Irniger et al., 1995; Zachariae and Nasmyth, 1996).

To determine the half-life of mitotic cyclins, we integrated a copy of CLB2 under the control of the inducible GAL promoter into wild-type and mutant strains. Wild-type and the doc1–1 and cdc26–100 mutants isolated in our screen were arrested with α-factor and shifted to 37°C for 20 min, and CLB2 expression was induced with galactose. Cycloheximide and glucose were added after a 30-min induction to stop Clb2 protein synthesis and CLB2 transcription. Clb2 protein levels were monitored by Western blotting using anti-Clb2 antibodies. Figure 2A shows that the half-life of Clb2 in wild-type cells is less than 1 min. In doc1–1 and cdc26–100 mutants, the half-life of Clb2 protein increases to greater than 20 min. This defect is rescued in doc1–1 and cdc26–100 by the integration of a wild-type copy of the corresponding gene at the LEU2 locus (our unpublished results). Others have also identified CDC26 as playing a role in cyclin proteolysis (Zachariae et al., 1996).

Figure 2.

Phenotype of cdc26 and doc1 mutants. (A) Stability of Clb2 in proteolysis mutants. The indicated strains contained an integrated copy of pGAL-CLB2, were grown in YEP + 2% raffinose medium, and arrested in G₁ by exposure to α-factor for 3 h at 23°C. The cells were shifted to 37°C for 20 min before CLB2 expression was induced by the addition of 2% galactose while maintaining the α-factor arrest. Cycloheximide (10 μg/ml) and glucose (2%) were added after a 30-min induction and time points were taken at 0, 1, 2, 5, and 20 min. An exposure of a Western blot probed with polyclonal anti-Clb2 antibodies is shown. Wild-type (lanes 1–5), doc1–1 (lanes 6–10), and cdc26–100 (lanes 11–15). (B) Stability of β-gal derivatives in proteolysis mutants. Wild-type, doc1–1, cdc26–100, and cim3–1 strains containing plasmids expressing either Ub-R-β-gal (lanes 1–4) or Ub^V76-V-e^ΔK-βgal (Ub^V76-βgal; lanes 5–8) under the control of the GAL promoter were grown in YEP + 2% raffinose and then arrested in G₁ with α-factor for 3 h. Expression from the plasmids was induced by the addition of 2% galactose for 2 h. At time zero, cycloheximide was added to 10 μg/ml and samples were taken at 15, 30, and 45 min. An exposure of a Western blot probed with anti-β-gal antibodies is shown.

The Clb2 destruction defect in cdc26–100 and doc1–1 appears to be specific to Clb2 and not a general defect in the proteolysis of ubiquitinated substrates. Figure 2B shows that cdc26–100 and doc1–1 cells have no defect in the degradation of two substrates that require ubiquitination for destruction: Ub-R-βgal, which generates β-gal with an N-terminal arginine, a substrate that is degraded by the N-end rule (Figure 2B, Ub-R-βgal; Bachmair et al., 1986), and Ub^V76-V-e^ΔK-βgal, a fusion with a noncleavable N-terminal ubiquitin (Figure 2B, UB^V76-βgal; Johnson et al., 1992). Cycloheximide was added to cells expressing Ub-R-βgal or Ub^V76-V-e^ΔK-βgal to inhibit protein synthesis, and stability of the substrates was determined by Western blotting with anti-β-gal antibodies at different times after inhibiting protein synthesis. cim3–1, a mutant in a subunit of the 26S proteasome (Ghislain et al., 1993) that degrades all ubiquitinated proteins, is defective in the degradation of both substrates (Figure 2B, lanes 1–8).

To see if the mutant phenotype of cdc26–100 and doc1–1 cells reflected defects in mitotic proteolysis, we overexpressed a mitotic cyclin in these mutants. cdc26–100 and doc1–1 strains containing an integrated copy of GAL-CLB2 are unable to grow on galactose at their permissive temperature (data not shown). We suggest that these strains are already crippled in their Clb2 destruction machinery at the permissive temperature and that excess Clb2 overloads this machinery and results in a cell cycle arrest.

Cloning and Disruption of DOC1

Like previously identified mitotic cyclin proteolysis mutants, doc1–1 mutants arrest as large budded cells at 37°C with a single nucleus and a short spindle (Figure 3A and Table 3).

Figure 3.

(A) doc1–1 arrest phenotype. The top two large-budded cells arrested with short spindles and a single DNA mass. The bottom two adjacent cells are large-budded. One has arrested with a short spindle and single DNA mass, the other appears to have undergone anaphase. doc1–1 cells were grown to logarithmic phase at room temperature and then shifted to 37°C for 3 h. The cells were fixed and stained with 4,6-diamidino-2-phenylindole and anti-tubulin. (B) Structure of the DOC1 locus and complementation of doc1–1 temperature sensitivity with various plasmids. The shaded region represents the DOC1 ORF and the hatched region represents the 5′ end of the CSE1 ORF. The arrows show the direction of transcription. The DOC1 ORF with 508 bp upstream is sufficient to complement the temperature sensitivity of doc1–1 at 37°C. TI, the rescuing plasmid from a YCp50-based library, contains approximately 640 bp upstream of the ORF of DOC1 and approximately 1150 bp of the CSE1 ORF. T1 was cut with NdeI and blunted with the Klenow fragment of DNA polymerase I (New England Biolabs), producing a frameshift in the DOC1 ORF (pLH26–1B). The frameshifted construct was unable to rescue doc1–1. pLH26–1A, in which URA3 is inserted into the NdeI-blunted site, was also unable to rescue doc1–1. (C) Amino acid sequence of DOC1. This sequence corresponds to ORF YGL240W in the yeast genome database (the GenBank accession number for DOC1 is Z72762). The DOC1 amino acid sequence is aligned with rat and mouse expressed sequence tag sequences (accession numbers H33761 and W49295, respectively). Identities are boxed. The doc1–1 mutation is indicated above the sequence. The mutation in doc1–1 (S137 → F137) is indicated in boldface type.

Table 3.

doc1-1 arrest phenotype

	Strain	% of total
	Wild type	75	10	11	4
	doc1-1	16	1	79	4

Wild-type and doc1-1 strains were grown to logarithmic phase and then shifted to 37°C for 3 h. The cells were fixed and then stained for DNA using 4,6-diamino-2-phenylindole and tubulin by using immunofluorescence. Percentage of total cell morphologies of unbudded with a single nucleus and a G₁ microtubule array, small budded with a single nucleus and short spindle, large budded with a single nucleus and short spindle, and large budded with separated DNA masses and long spindle are indicated from left to right. Two hundred cells were counted for each strain.  

DOC1 was cloned by complementing the temperature sensitivity of doc1–1 with a genomic yeast library on a centromeric vector (Figure 3B, plasmid T1). The minimal complementing region contains a single ORF coding for a protein of 283 amino acids. A construct containing the ORF of DOC1 and 508 bp upstream, with a single myc tag on the C terminus and no sequences 3′ to the ORF, complements the temperature sensitivity and Clb2 proteolysis defect of the doc1–1 mutant (Figure 3B, plasmid pLH24). In addition, a construct containing a frameshift at an NdeI site 297 bp into the DOC1 ORF, will not rescue the temperature sensitivity of doc1–1 (Figure 3B, plasmid pLH26–1B). Disruption of DOC1 by insertion of the URA3 gene at the NdeI site (Figure 3B, plasmid pLH26–1A) results in viable cells that grow poorly at 23°C, forming colonies in which most of the cells have large buds and cannot grow at 37°C. Coincidentally, DOC1 lies directly upstream of the CSE1 gene, which is also implicated in cyclin proteolysis (Xiao et al., 1993; Irniger et al., 1995). These genes have distinct functions in proteolysis because plasmids that contain only the CSE1 gene do not complement doc1–1. The Doc1 amino acid sequence is shown in Figure 3C. The Doc1 sequence lacks obvious motifs and shows no homology to proteins of known function; however, it does show homology to rat (Lee et al., 1995) and mouse expressed sequence tag sequences (GenBank accession numbers H33761 and W49295, respectively). The doc1–1 mutation was recovered and sequenced. The mutation changes a conserved serine (Ser-137) to phenylalanine.

Cdc26 and Doc1 Associate with APC Components (Cdc16, Cdc23, and Cdc27)

Cdc16, Cdc23, and Cdc27 form a complex in yeast (Lamb et al., 1994) and their homologues are components of a 20S complex in frog egg extracts (King et al., 1995). This complex behaves as a ubiquitin ligase for B-type cyclins, and CDC16, CDC23, and CDC27 are required for the ubiquitination of Clb2 in yeast (Zachariae and Nasmyth, 1996). We therefore tested whether Cdc26 or Doc1 associated with Cdc16, Cdc23, or Cdc27. Extracts were made from exponentially growing wild-type or cdc26Δ strains that had been transformed with centromeric plasmids containing HA-tagged versions of CDC16, CDC23, CDC27, and DOC1. Cdc26 was immunoprecipitated with polyclonal antibodies against Cdc26. The bound proteins were probed by Western blotting with mouse monoclonal antibodies against HA (12CA5) to test whether the tagged proteins coimmunoprecipitate with Cdc26.

Figure 4 shows that immunoprecipitation of Cdc26 in a wild-type strain results in the coimmunoprecipitation of HA-tagged Cdc16, Cdc23, Cdc27, and Doc1. No HA-tagged Cdc16, Cdc23, Cdc27, or Doc1 was detected when immunoprecipitations were performed on extracts made from a cdc26Δ strain. These experiments suggest that Cdc26 is a functional component of the APC, and this suggestion is strengthened by the observation that the Xenopus homologue of Cdc26 is a component of the purified APC (M. Kirschner, personal communication).

Figure 4.

Coimmunoprecipitation of Cdc26 with epitope-tagged forms of Cdc16, Cdc23, Cdc27, and Doc1 (Cdc16^HA, Cdc23^HA, Cdc27^HA, and Doc1^HA). Wild-type (+) and cdc26Δ (Δ) strains containing centromeric plasmids expressing Cdc16^HA (pWAM10; lanes 1 and 2), Cdc23^HA (pRS239; lanes 3 and 4), Cdc27^HA (pRS248; lanes 5 and 6), or Doc1^HA (pLH32; lanes 7 and 8) were grown to logarithmic phase in selective medium. Cdc26 was immunoprecipitated from those cell lysates with polyclonal anti-Cdc26 antibodies. Immunoprecipitates were analyzed by immunoblotting with 12CA5. The positions of molecular mass markers and Cdc16^HA, Cdc23^HA, Cdc27^HA, and Doc1^HA are indicated in kilodaltons.

The association of Doc1 with Cdc26 could be interpreted in two ways. Either Doc1 is a component of the APC or there are two pools of Cdc26, one associated with the APC and one bound to Doc1. To distinguish these possibilities, we analyzed the sedimentation of Doc1 and Cdc27 in a sucrose gradient. Extracts from exponentially growing cells expressing both HA-tagged Doc1 and HA-tagged Cdc27 were separated on a 5–40% sucrose gradient (Figure 5). The majority of Doc1 sediments in a peak at approximately 20S, coinciding with the sedimentation of Cdc27. The sedimentation coefficient of the frog and clam APC is 20S (King et al.., 1995; Sudakin et al., 1995).

Figure 5.

Cosedimentation of Doc1 with Cdc27. A lysate from exponentially growing cells expressing both Doc1 and Cdc27 tagged with the HA epitope (Doc1^HA and Cdc27^HA) was separated through a 5–40% sucrose gradient. Fractions were analyzed by immunoblotting with 12CA5.

Genetic Interactions among CDC26, DOC1, and Other Mitotic Mutants

Genetic interactions between mutants often indicate that they function in the same pathway. We found that double mutants of cdc26–100 and doc1–1 are synthetically lethal (Table 4), suggesting that the functions of Cdc26 and Doc1 overlap. In addition, cdc26–100 is synthetically lethal with cdc16–1 and cdc23–1, and cdc26–100 cdc27–1 double mutants have a lower permissive temperature than either single mutant. doc1–1 double mutants with cdc16–1 or cdc23–1 have a lower permissive temperature than any of the single mutants. The interactions of cdc26–100 and doc1–1 with mutants involved in the ubiquitination of mitotic cyclins do not simply reflect synthetic interactions between any pair of mitotic mutants. Double mutants between cdc26–100 or doc1–1 and cim5–1 (a subunit of the 26S proteasome, required for the destruction of ubiquitinated proteins; Ghislain et al., 1993) have the same nonpermissive temperature as the cdc26–100 or doc1–1 single mutants.

Table 4.

Genetic interactions between cdc26, doc1, and other mitotic mutants

	cdc26-100	doc1-1	cdc16-1	cdc23-1	cdc27-1	cim5-1
Maximum permissive temperature (°C)	33	33	30	30	33	37
Maximum permissive temperature of double mutant with cdc26-100 (°C)	33	Dead	Dead	Dead	30	33
Maximum permissive temperature of double mutant with doc1-1 (°C)	Dead	33	23	23	30	33

Synthetic interactions among cdc26-100, doc1-1, and various mitotic mutants. cdc26-100 and doc1-1 were mated to each other or to cdc16-1, cdc23-1, cdc27-1, or cim5-1. The resultant diploids were sporulated and tetrads were dissected and germinated at 23°C. They were then tested for growth at 23°C, 30°C, 33°C, 35°C, and 37°C. The genotype of viable spores was determined by complementation testing with appropriate mutant tester strains. The maximum permissive temperatures of the cdc26-100 and doc1-1 mutants are shown in boldface type. 

DISCUSSION

We have identified two new components of the cyclin proteolysis machinery, Cdc26 and Doc1, and show that Cdc26 and Doc1 are physically associated with the APC/cyclosome, the multiprotein complex that acts as an E3 for cyclin ubiquitination (King et al., 1995; Sudakin et al., 1995; Zachariae and Nasmyth, 1996). Similar results were recently obtained for CDC26 by Zachariae et al. (1996).

We isolated temperature-sensitive mitotic mutants by using an enrichment that kills the majority of cells passing through S phase at the nonpermissive temperature. Of the 32 temperature-sensitive mitotic arrest mutations we isolated, 27 are in genes involved in the cyclin proteolysis machinery; 1 is in CDC31, a gene required for spindle pole body duplication; and 2 are in PRP22, which encodes a splicing function that is probably required to ensure adequate levels of α-tubulin synthesis. Why do proteolysis mutants dominate the metaphase arrest mutants isolated in both this screen and the original Hartwell screen for cdc mutants? We believe that three factors account for this observation. First, although defects in the mitotic spindle can arrest cells in mitosis by transiently activating the spindle assembly checkpoint, the checkpoint eventually adapts to persistent defects thus allowing cells to exit mitosis (Hardwick, Rudner, Wiess, Winey, and Murray, unpublished results). Because cells that adapt and leave mitosis with a defective spindle are likely to suffer lethal errors in chromosome segregation, mutants that cause spindle defects may have died during the prolonged incubation at 37°C that was used during the mutant enrichment. Second, there appears to be considerable functional overlap among the components that assemble the mitotic spindle, so that mutations that inactivate single components do not cause a mitotic arrest. For example, strains lacking the Kar3 microtubule motor are viable, although mitosis in these cells is clearly abnormal (Meluh and Rose, 1990). Finally, the integrity of some spindle functions, such as the protein kinase Mps1 (Hardwick et al., 1996), are required for the spindle assembly checkpoint to detect defects in the spindle. As a result, lesions in these components would not arrest cells in mitosis.

cdc26 and doc1 mutants arrest in metaphase and cannot destroy mitotic cyclins that are expressed in G₁-arrested cells. Unlike mutations in the proteasome, cdc26 and doc1 mutants do not suffer general defects in ubiquitin-mediated proteolysis. A substrate that is recognized by a destabilizing N-terminal amino acid or a noncleavable N-terminal ubiquitin is degraded normally in cdc26 and doc1 mutants. Thus, these observations suggest that the mitotic arrest of cdc26 and doc1 reflects their inability to degrade mitotic cyclins and proteins, such as Cut2 and Pds1, whose destruction is required for sister chromatid separation (Funabiki et al., 1996; Yamamoto et al., 1996). This conclusion is supported by both genetic and biochemical evidence for interactions of Cdc26 and Doc1 with the APC. Immunoprecipitates prepared with anti-Cdc26 antibodies contain Cdc16, Cdc23, and Cdc27, all of which are characterized components of the APC, as well as Doc1. In addition, a population of Doc1 cosediments with Cdc27. This evidence suggests that both Cdc26 and Doc1 are associated with the APC. Because we cannot easily measure the amounts of Cdc26 and Doc1 relative to other components of the APC, it is unclear whether Cdc26 and Doc1 are stoichiometric components of the APC. The observation that neither Cdc26 or Doc1 show any homology to cloned subunits of the biochemically purified Xenopus APC suggests that Cdc26 and Doc1 cannot be tightly associated with all APC complexes (King et al., 1995). The observation that doc1–1 is synthetically lethal with cdc26–100, but not with mutants in other components of the APC, suggests that Doc1 and Cdc26 may play partially overlapping roles in cyclin proteolysis. Although cdc26Δ mutants grow well at 23°C, they are synthetically lethal with cdc16–1 and cdc23–1, suggesting that Cdc26 may play a role in stabilizing the APC and that this role is dispensable under optimal conditions but becomes essential under environmental stress or in the presence of defects in other APC components.

Defects in the mitotic spindle activate a spindle assembly checkpoint that arrests eukaryotic cells in mitosis (Hoyt et al., 1991; Li and Murray, 1991). Activation of the checkpoint prevents the destruction of mitotic cyclins and sister chromatid separation, suggesting that the principal action of the checkpoint is to inhibit the proteolysis of mitotic cyclins and proteins that play a role in maintaining the linkage between sisters. It is currently unclear whether this inhibition is due to inactivation of the APC itself or to reactions that protect specific substrates from a normally active APC in checkpoint-arrested cells. Further biochemical comparison between APC components in anaphase- and checkpoint-arrested extracts should help to resolve this issue.