Lesions in Many Different Spindle Components Activate the Spindle Checkpoint in the Budding Yeast Saccharomyces cerevisiae

Corresponding author: Andrew W. Murray, Physiology Box 0444, UCSF, Parnassus Ave., San Francisco, CA 94143-0444., amurray{at}socrates.ucsf.edu (E-mail)

Communicating editor: D. BOTSTEIN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The spindle checkpoint arrests cells in mitosis in response to defects in the assembly of the mitotic spindle or errors in chromosome alignment. We determined which spindle defects the checkpoint can detect by examining the interaction of mutations that compromise the checkpoint (mad1, mad2, and mad3) with those that damage various structural components of the spindle. Defects in microtubule polymerization, spindle pole body duplication, microtubule motors, and kinetochore components all activate the MAD-dependent checkpoint. In contrast, the cell cycle arrest caused by mutations that induce DNA damage (cdc13), inactivate the cyclin proteolysis machinery (cdc16 and cdc23), or arrest cells in anaphase (cdc15) is independent of the spindle checkpoint.

MITOSIS produces two daughter nuclei with identical genetic contents. To achieve this feat, cells have to delay chromosome segregation until all their chromosomes are correctly aligned on a bipolar spindle. Spindle assembly depends on the dynamic properties of microtubules nucleated by microtubule organizing centers (the spindle pole bodies in yeast), microtubule motors (reviewed in HOYT and GEISER 1996 ; STEARNS 1997 ), and the attachment of spindle microtubules to kinetochores [the complex of centromeric DNA and proteins that attaches chromosomes to microtubules (Figure 1), reviewed in HYMAN and SORGER 1995 ; ALLSHIRE 1997 ]. Mutations or drugs that inhibit these functions lead to spindle defects that can cause aberrant chromosome segregation (reviewed in HOYT and GEISER 1996 ). The spindle checkpoint guards against such errors by delaying the onset of chromosome segregation in cells whose spindles are defective (reviewed in RUDNER and MURRAY 1996 ; WELLS 1996 ).

View larger version (14K): In this window In a new window Download PPT slide   Figure 1. Mitosis in budding yeast. Cells are shown in metaphase, anaphase, and G1. The mutants used in this article that compromise various structures and reactions are indicated.

The spindle checkpoint has been analyzed genetically in budding yeast. The mitotic arrest deficient (mad; LI and MURRAY 1991 ), budding uninhibited by benzimidazole (bub; HOYT et al. 1991 ), mps1 (HARDWICK et al. 1996 ; WEISS and WINEY 1996 ), and cdc55 (MINSHULL et al. 1996 ) mutations inactivate the checkpoint and allow cells whose microtubules have been depolymerized to pass through mitosis. Experiments in yeasts and frog egg extracts suggest that the spindle checkpoint acts by preventing the activation of the cyclin proteolysis machinery (MINSHULL et al. 1994 , MINSHULL et al. 1996 ; HARDWICK and MURRAY 1995 ; CHEN et al. 1996 ; LI and BENEZRA 1996 ; FANG et al. 1998A ; HWANG et al. 1998 ; KIM et al. 1998 ). This protein degradation system induces the exit from mitosis by activating a multiprotein complex [the cyclosome or anaphase-promoting complex (APC)] that catalyzes the ubiquitination of critical proteins, thus targeting them for degradation by the proteasome (IRNIGER et al. 1995 ; KING et al. 1995 ; SUDAKIN et al. 1995 ). The key targets for ubiquitination are the B-type mitotic cyclins that associate with Cdc2/Cdc28 and proteins that are required for sister chromatid separation (HOLLOWAY et al. 1993 ; FUNABIKI et al. 1996 ; YAMAMOTO et al. 1996 ; CIOSK et al. 1998 ; KUMADA et al. 1998 ).

To determine which lesions in the spindle the MAD-dependent spindle checkpoint detects, we combined the mad1, mad2, and mad3 mutations with other mutations that affect spindle or function. These included mutations causing defects in microtubule polymerization (HUFFAKER et al. 1988 ; HOYT et al. 1990 ; STEARNS et al. 1990 ), spindle pole body duplication (WINEY et al. 1991 ), kinetochore components (CAI and DAVIS 1990 ), or microtubule motors (MELUH and ROSE 1990 ; HOYT et al. 1992 ). These studies show that defects in microtubule polymerization, spindle pole bodies, microtubule motors, and kinetochores all arrest cells in mitosis by activating the spindle checkpoint.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Media and strains:
Media for yeast cultures and sporulation were as described (SHERMAN et al. 1974 ). Strains are listed in Table 1.

Genetic techniques:
Standard techniques were used for yeast mating, sporulation, tetrad analysis, and -factor treatment (SHERMAN et al. 1974 ).

Synthetic lethality:
These experiments were performed in one of two ways, either by crossing a haploid lacking a microtubule motor to a mad strain, or by using gene disruption to create heterozygosity for loss of a microtubule motor gene in a strain already heterozygous for a mad mutation. The latter method is more cumbersome, but it eliminates the possibility that slowly growing strains, such as kar3, will have acquired suppressor mutations.

Cell death assays:
The rate of cell death in mad cdc (cell division cycle) double mutants was determined as follows. A saturated culture was grown in YPD at 23°, and 0.1 ml was plated on a YPD plate and incubated for 2 days at 23°. Cells were collected by washing the plate with liquid YPD, diluted, and incubated at the nonpermissive temperature for the mutant. At the indicated times, appropriate dilutions of cultures were plated on YPD plates that were incubated for 3–5 days at room temperature before colony counting. This method was used because some of the mutants double so slowly at 23° that other methods of synchronizing the cells were ineffective. All figures are representative of experiments that were repeated at least three times.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We investigated the types of mitotic lesions that activate the MAD-dependent checkpoint by studying the interaction of mad mutations with mutations that cause defects in the execution of particular steps of mitosis (Figure 1). These include mutations that inhibit microtubule depolymerization, prevent spindle pole body duplication, remove centromere-binding proteins, and cdc mutations that arrest cells in mitosis at 37°.

We tested mitotic mutations in two classes, conditional lethal mutations and null mutations in nonessential genes. Figure 2 shows the rationale for these experiments. To determine the interaction of mad mutations with cdc mutations that cause arrest in G2 or mitosis, the properties of the mad cdc double mutant were compared with that of the cdc mutant alone. If the cell cycle arrest of the cdc mutant is independent of the spindle checkpoint, the cdc mad double behaves exactly like the cdc mutant alone: the cell cycle arrests at the nonpermissive temperature, and the single and double mutants lose viability at the same rate when incubated at the nonpermissive temperature. In contrast, if the arrest of the cdc mutant depends on the spindle checkpoint, inactivation of the checkpoint will allow the mad cdc double mutants to pass through mitosis at the restrictive temperature even though the spindle is defective. In a haploid strain, initiating anaphase in cells that have not aligned their chromosomes properly on a bipolar mitotic spindle will lead to chromosome loss and generate dead cells. Thus, combining such cdc mutations with mad mutations will produce double mutants that fail to arrest in mitosis and lose viability faster than the cdc mutant alone when the strains are incubated at the nonpermissive temperature (Figure 2). A second test for whether the spindle checkpoint detects a defect is to examine the cell cycle progress. If the cell cycle arrest of a cdc mutant is suppressed in the mad cdc double mutant, the spindle checkpoint must be required for the arrest. This analysis is only meaningful if most of the cells in the starting population are viable and the degree of synchrony in the population is high. Unfortunately, for several of the mutations that do show genetic interactions with the mad mutations, the double mutants are difficult to synchronize, and populations of these strains contain many dead cells, even at the permissive temperature.

View larger version (27K): In this window In a new window Download PPT slide   Figure 2. Interactions between mad mutations and mutations that cause spindle defects. The effect of inactivating the spindle checkpoint on cell division in wild-type cells, cdc mutants that affect spindle structure, and cells with nonlethal spindle defects. See text for further details.

A number of components of the spindle, such as individual microtubule motors, are not essential for cell viability (reviewed in HOYT and GEISER 1996 ). Although the loss of such components may compromise spindle function, cells can survive without them if the spindle checkpoint delays anaphase until alternative mechanisms have correctly aligned the chromosomes on a bipolar spindle. This model predicts that the combined loss of a nonessential spindle component and the spindle checkpoint would be lethal, since the double mutant cells would no longer be able to delay anaphase to allow backup mechanisms to assemble a functional spindle (Figure 2). This prediction has been confirmed by the isolation of several checkpoint mutations (bub2, bub3, mps1, and a dominant CDC20 allele, CDC20-50) in a screen for mutations that kill cells lacking Cin8, a kinesin with a role in spindle assembly (GEISER et al. 1997 ).

Tubulin and spindle pole body mutants activate the spindle checkpoint:
The mad mutants were originally isolated as mutants that had increased sensitivity to benomyl, an inhibitor of microtubule polymerization. The mutants failed to arrest in mitosis and died rapidly when treated with benomyl (LI and MURRAY 1991 ; HARDWICK and MURRAY 1995 ), suggesting that the defects in microtubule polymerization activated the spindle checkpoint. To confirm that this phenotype resulted from the effect of benomyl on microtubule polymerization, we tested the interaction of the mad mutations with a mutation in TUB2, the yeast ß-tubulin gene (HUFFAKER et al. 1988 ). When cells of a strain carrying the cold-sensitive tub2-403 mutation are incubated in rich medium at 14°, they arrest in mitosis as large-budded cells that lack polymerized microtubules and remain viable for >6 hr. In contrast, tub2-403 mad1, tub2-403 mad2, and tub2-403 mad3 double mutants lose viability during incubation at 14°, with <25% of the cells remaining viable at 6 hr (Figure 3A). In this experiment, cells were shifted to the restrictive temperature from a block at the G1/S phase boundary created by hydroxyurea treatment because release from a G1 arrest at 14° is very slow. These experiments show that the combination of the mad mutations with the tub2-403 mutation increases the rate at which cells that cannot polymerize microtubules die in the cold. These observations demonstrate that the MAD-dependent spindle checkpoint is required to arrest cells that cannot assemble microtubules.

View larger version (45K): In this window In a new window Download PPT slide   Figure 3. mad mutations increase the rate at which tubulin and spindle pole body mutants die. The indicated strains were arrested in G1 by nutrient starvation at 23° and released into rich medium at the nonpermissive temperature for (A) tub2-403 (14°) or (B) mps2-1 (37°). Samples were taken at the indicated time points, diluted, and plated for viability on rich medium at 23°. Values are expressed relative to the number of viable cells at time zero.

Mutations that prevent spindle pole body duplication have no effect on microtubule polymerization, but they cause cells to assemble monopolar rather than the normal bipolar spindles. Two mutants, mps2 and cdc31, that have this phenotype arrest in mitosis, whereas a third, mps1, fails to arrest in mitosis despite the absence of a bipolar spindle (BAUM et al. 1986 ; WEISS and WINEY 1996 ). The observation that mps1 mps2 double mutants fail to arrest in mitosis led to the suggestion that mps1 mutants lacked a checkpoint for spindle pole body duplication. Subsequent analysis showed that mps1 mutants, like the mad mutants, could not respond to spindle depolymerization, and demonstrated that Mps1 functions in the Mad-dependent spindle checkpoint (HARDWICK et al. 1996 ; WEISS and WINEY 1996 ). The inability of cdc31-2 mps1-1 double mutants to arrest in mitosis suggested that the spindle checkpoint was required to arrest cells with monopolar spindles in mitosis (WEISS and WINEY 1996 ), as did the observation that the cdc31-2 and mps2-1 mutations both led to the hyperphosphorylation of Mad1, a biochemical marker for activation of the checkpoint (HARDWICK et al. 1996 ).

We tested whether Mad proteins were needed to detect monopolar spindle mutations by combining the temperature-sensitive mps2-1 mutation with the mad1, mad2, or mad3 mutations and monitoring the rate of death at 37° as a measure of exit from mitosis. The single mps2-1 mutant arrests in mitosis and remains fully viable for at least 7 hr, when G1 stationary phase cells are inoculated into rich medium at 37°. All the mps2-1 mad double mutants, however, die rapidly, with 50% of the cells becoming inviable within 3–4 hr of incubation at 37° and 25% of the cells remaining viable at 7 hr (Figure 3B). This observation demonstrates that all the Mad proteins are required for cells to detect a monopolar spindle and arrest in mitosis as viable cells.

Microtubule motor defects activate the spindle checkpoint:
Microtubule motors move along the surface of microtubules, and genetic and cell biological experiments have implicated motors in many aspects of mitosis and meiosis in a wide variety of organisms. These include formation of a bipolar spindle (ENOS and MORRIS 1990 ; HAGAN and YANAGIDA 1990 , HAGAN and YANAGIDA 1992 ), the regulation of spindle length (HOYT et al. 1992 ; ROOF et al. 1992 ), the movement of chromosomes along microtubules (HYMAN and MITCHISON 1991 ; LOMBILLO et al. 1995 ), the regulation of microtubule stability (WALCZAK et al. 1996 ), and distributive chromosome segregation in meiotic cells (ASHFAR et al. 1995 ). Despite these many roles, no single microtubule motor in yeast is essential for viability at 30° (HOYT and GEISER 1996 ). This observation can be explained in two ways: either there is full overlap between the function of different motors so that loss of one motor has no effect on the robustness or fidelity of mitosis, or there is partial overlap, and the spindle checkpoint delays the onset of anaphase until cells with motor defects have aligned their chromosomes correctly. A variety of observations support the latter view. Detailed analysis of the dynamics of mitosis reveals that defects in different motors affect different phases of mitosis (STRAIGHT et al. 1998 ); cells that lack the minus-end-directed motor, Kar3, delay in mitosis (MELUH and ROSE 1990 ), and cells that lack the plus-end-directed motor, Cin8, show elevated frequencies of chromosome loss (HOYT et al. 1992 ). In addition, the inability to recover mad2-1 kar3 double mutants (ROOF et al. 1991 ) suggests that the proliferation of cells lacking Kar3 depends on a functional spindle checkpoint, and the recovery of mad and bub mutations as synthetic lethal mutations with cin8 suggests that the lack of the Cin8 motor causes defects that are lethal in the absence of the spindle checkpoint (GEISER et al. 1997 ).

We tested the interaction of microtubule motors and the spindle checkpoint by combining the kar3 and cin8 mutations with the mad1, mad2, and mad3 mutations. Diploids that were heterozygous for a mad mutation and a defect in a microtubule motor were sporulated and subjected to tetrad dissection. Because both mutations were marked by the insertion of standard genetic markers, the genotype of every spore could be unambiguously deduced from its pattern of growth on selective media. The results of this analysis are presented in Table 2. In all but one of the crosses (that between mad3 and cin8), fewer double-mutant spores were recovered than any other genotype. We tested the statistical significance of these findings in two ways. First, we tested whether the results we obtained were significantly different from those expected if all four genotypes were equally likely to survive. Second, we tested whether the viability of the double-mutant spores was less than the viability of the single mad mutant spores' viability multiplied by the viability of the spores lacking only the microtubule motor. Both tests indicated that the viability of the kar3 mad1, kar3 mad2, cin8 mad1, cin8 mad2, and cin8 mad3 double mutants was significantly less than expected, showing that viability of mutants lacking Kar3 or Cin8 is dependent on the integrity of the spindle checkpoint. For the kar3 mad3 double mutant, the low spore viability of the kar3 MAD3 spores makes it impossible to determine whether the viability of kar3 cells depends on Mad3.

The absence of a kinetochore component activates the spindle checkpoint:
Many observations suggest that interactions between the kinetochore and the spindle microtubules play an important part in sensing the structure of the mitotic and meiotic spindles (reviewed in RUDNER and MURRAY 1996 ; WELLS 1996 ). These suggest that the spindle checkpoint monitors spindle integrity by detecting kinetochores that have not attached to microtubules or by measuring the tension exerted by microtubules on the kinetochore.

We assayed the interaction between kinetochore mutations and mad mutations. In our hands, none of the mutations in the essential components of the kinetochore (Ndc10, Ctf13, and Cbf3/Cep3; reviewed in HYMAN and SORGER 1995 ) gave a robust cell cycle arrest in the W303 strain background, making it impossible to assess whether the phenotype associated with any of these mutations was altered by loss of the spindle checkpoint. Instead, we assayed the interaction between the mad mutations and cbf1, which eliminates a nonessential protein that binds to the CDEI element of the centromeric DNA (CAI and DAVIS 1990 ). We constructed diploids that were heterozygous for genetically marked cbf1 and mad deletions, sporulated the diploids, and subjected them to tetrad dissection. Very few cbf1 mad1 or cbf1 mad2 spores were recovered from these crosses, demonstrating that the viability of cbf1 cells depends on the integrity of the spindle checkpoint (Table 3). Like kar3, the cbf1 mutation was viable when combined with the mad3 mutation. This observation suggests that the mad3 mutant is, in some sense, less defective in the spindle checkpoint than mad1 or mad2 mutants. Others have shown spindle checkpoint-dependent arrests in response to ctf13 mutations and mutations in the kinetochore DNA in the S288C strain background (WANG and BURKE 1995 ; PANGILINAN and SPENCER 1996 ).

The arrest of proteolysis-defective mutants is checkpoint independent:
Experiments in budding yeast and frog egg extracts support the idea that the spindle checkpoint prevents anaphase by inhibiting the activation of the cyclin proteolysis machinery (HOYT et al. 1991 ; LI and MURRAY 1991 ; MINSHULL et al. 1994 ; HARDWICK and MURRAY 1995 ). If the critical event for inducing anaphase is the activation of this machinery, mutations that block the proteolysis of mitotic cyclins should arrest the cell cycle in mitosis, even in cells that lack the spindle checkpoint. We tested this hypothesis by examining the effect of mad mutations on the arrest phenotype of cdc15 and cdc23, two mutations with defects in the proteolysis of mitotic cyclins. The cdc23-1 mutation affects a component of the anaphase-promoting complex, the multiprotein complex that catalyzes the ubiquitination of cyclin B and Pds1, a protein that regulates the cohesion of sister chromatids (IRNIGER et al. 1995 ). In contrast, Cdc15 is a protein kinase whose inactivation arrests cells in anaphase, apparently because they are unable to completely degrade cyclin B (SCHWEITZER and PHILIPPSEN 1991 ; ZACHARIAE and NASMYTH 1996 ; JASPERSEN et al. 1998 ).

We combined cdc23-1 and cdc15 with mad mutations and examined the phenotypes of the double mutants at the nonpermissive temperature (Figure 4A and Figure B). In all cases, the results were the same: the arrest phenotype of the double mutants was indistinguishable from that of the single proteolysis mutants.

View larger version (16K): In this window In a new window Download PPT slide   Figure 4. mad mutations do not increase the rate of cell death in mutations that impair mitotic protein degradation. The indicated strains were arrested in G1 by nutrient starvation at 23° and released into rich medium at 37° for (A) cdc23-1, (B) cdc15-2, or (C) cdc20-1 strains. Samples were taken at the indicated time points, diluted, and plated for viability on rich medium at 23°. Values are expressed relative to the number of viable cells at time zero. All experiments were repeated three times. Although the results of individual experiments with cdc20-1 show some variability, there is no consistent difference between the rate at which the cdc20-1 and cdc20-1 mad double mutants die.

We also examined the interaction of the mad mutations with cdc20-1, which results in arrest in mitosis with increased numbers of microtubules. Cdc20 is a member of a conserved family of proteins that binds to the APC and plays an essential but poorly defined role in APC-mediated proteolysis (FANG et al. 1998B ; KALLIO et al. 1998 ). Evidence in budding yeast suggests that Cdc20 preferentially targets Pds1 rather than cyclin B for destruction, whereas Hct1/Cdh1, another member of this family, shows a preference for ubiquitination of cyclin B and is responsible for the destruction of cyclin B in G1-arrested cells (SCHWAB et al. 1997 ; VISINTIN et al. 1997 ). Drosophila mutations in fizzy, the homolog of CDC20, arrest the embryonic cell cycle in metaphase and block the destruction of both cyclins A and B (DAWSON et al. 1995 ; SIGRIST et al. 1995 ). We compared the phenotypes of cdc20-1 mad double mutants with cdc20-1 mutants (Figure 4C). Like cdc20-1 mutants, the cdc20-1 mad double mutants arrested in mitosis as large-budded cells (data not shown) and remained viable, demonstrating that this arrest is independent of the spindle checkpoint.

A specific checkpoint arrests the cell cycles of cells whose DNA has been damaged (WEINERT and HARTWELL 1988 ). The cdc13-1 mutation leads to the accumulation of single-stranded DNA at telomeres, and it arrests cells with a G2 DNA content and a short spindle (GARVIK et al. 1995 ). This arrest is completely suppressed by mutations in the DNA damage checkpoint (LYDALL and WEINERT 1995 ). To test whether DNA damage could activate the spindle checkpoint, we examined the interaction between cdc13 and mad mutations (Figure 5). By this test, the cdc13-1 arrest was independent of the spindle checkpoint, since cdc13-1 mad double mutants, like the cdc13-1 single mutant, remained arrested as large-budded cells when shifted to 37° (data not shown), and the single and double mutants die at similar rates during prolonged incubation at 37°. These findings show that the components of the spindle checkpoint are not needed for the DNA damage checkpoint. WEISS and WINEY 1996 reached similar conclusions to ours by showing that the spindle checkpoint mutation mps1-1 failed to overcome the cell cycle arrest caused by DNA damage in the cdc9-1 and cdc13-1 mutations or the arrest caused by defects in APC-mediated proteolysis in cdc16-1, cdc20-1, and cdc23-1 mutations.

View larger version (23K): In this window In a new window Download PPT slide   Figure 5. mad mutations do not affect the behavior of a mutation that activates the DNA damage checkpoint. The indicated cdc13 strains were arrested in G1 by nutrient starvation at 23° and released into rich medium at 37°. Samples were taken at the indicated time points, diluted, and plated for viability on rich medium at 23°. Values are expressed relative to the number of viable cells at time zero.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We have analyzed the defects that the spindle checkpoint detects. Cells with defects in microtubule polymerization or spindle pole body duplication die rapidly at the nonpermissive temperature if they lack the spindle checkpoint. Spores with defects in the checkpoint and microtubule motors or a kinetochore component survive at a greatly reduced frequency, suggesting that the survival of cells with these defects depends on the integrity of the spindle checkpoint. In contrast, mutations that inhibit the anaphase-promoting complex or activate the DNA damage checkpoint arrest the cell cycle independently of the spindle checkpoint. These observations confirm the original conclusion that the spindle checkpoint is restricted to monitoring the function of the mitotic spindle (HOYT et al. 1991 ; LI and MURRAY 1991 ).

Benomyl and nocodazole treatments and the tub2-403 mutation all reduce the stability of yeast microtubules in both the spindle and the cytoplasm (HUFFAKER et al. 1988 ; JACOBS et al. 1988 ; HOYT et al. 1990 ; STEARNS et al. 1990 ). These perturbations all induce a cell cycle arrest or delay that is abolished by mad (LI and MURRAY 1991 ) and bub mutations (HOYT et al. 1991 ). Tubulin mutants with defects that are restricted to cytoplasmic microtubules pass through mitosis and produce two daughter nuclei (SULLIVAN and HUFFAKER 1992 ). Although these cells initiate anaphase normally, they are defective in the positioning of the nucleus within the cell, and as a result, both daughter nuclei are often found in one of the two progeny cells (SULLIVAN and HUFFAKER 1992 ). This observation suggests that the spindle checkpoint only monitors the intranuclear spindle and does not respond to defects in nuclear positioning during mitosis. Examination of mutations that affect spindle positioning suggests that there is an independent checkpoint that delays the completion of nuclear division and cytokinesis in cells that have undergone anaphase but have not succeeded in placing one set of chromosomes in the bud (YEH et al. 1995 ; LI et al. 1998 ).

In contrast to the tub2-403 mutant or nocodazole-treated cells, the number of spindle microtubules in the cdc20-1 mutant is greater than that in wild-type cells (O'TOOLE et al. 1997 ). The mitotic arrest in cdc20-1 cells is not eliminated in mad mutants, raising the possibility that the Mad-dependent checkpoint is not involved in detecting spindle lesions caused by excess microtubules. We favor two alternative possibilities, either that cdc20 mutations have direct effects on the anaphase-promoting complex and indirect effects on microtubule distribution, or that the cdc20-1 allele is closely linked to another mutation that affects the number of spindle microtubules. The latter idea is supported by the observations that mutations in fizzy, the Drosophila homolog of Cdc20, do not have obvious effects on spindle morphology, but arrest cells in mitosis and prevent the proteolysis of both cyclins A and B (DAWSON et al. 1995 ; SIGRIST et al. 1995 ). Since cyclin A is not protected from proteolysis by the spindle checkpoint in frog (MINSHULL et al. 1994 ) and clam (HUNT et al. 1992 ) eggs, the phenotype of the fly mutant suggests that the arrest of cdc20/fizzy mutants reflects a defect in some step required for cyclin proteolysis rather than a defect in activation of the spindle checkpoint. This suggestion is strengthened by the observations that Cdc20 and its relative, Hct1, appear to act as specificity factors for APC-mediated protein degradation (SCHWAB et al. 1997 ; VISINTIN et al. 1997 ), and that Cdc20 and its homolog in fission yeast are targets that the spindle checkpoint inhibits to produce a mitotic arrest (HWANG et al. 1998 ; KIM et al. 1998 ). Finally, independently isolated, temperature-sensitive cdc20 alleles or extensively backcrossed cdc20-1 strains show a mitotic arrest without increased antimicrotubule staining compared to other metaphase-arrested mutants (A. AMON, personal communication; L. HWANG and A. W. MURRAY, unpublished data).

Mutations that affect spindle pole body function can also activate the MAD-dependent checkpoint. mps2-1, a monopolar spindle mutant, arrests in mitosis, and mad mps2-1 double mutants die faster than mps1 mutants. Our attempts to test the interaction of the checkpoint with other spindle pole body mutations, such as cdc31 and ndc1, have been frustrated by the rapid diploidization of these mutant strains, which hampers genetic analysis. Mutations in Mps1, a protein kinase whose activity is required for spindle pole body duplication, also inactivate the spindle checkpoint (HARDWICK et al. 1996 ). mps1-1 mutants fail to arrest in response to the spindle pole body defects or microtubule depolymerization (HARDWICK et al. 1996 ; WEISS and WINEY 1996 ). This analysis suggests that like the mad mutants, mps1 mutants fail to respond to several different defects in the spindle. Although the role of Mps1 in spindle pole body duplication makes it tempting to speculate that the spindle pole is involved in the checkpoint, there is no evidence that Mps1 is localized to the spindle pole body and that spindle pole body duplication and checkpoint functions of Mps1 clearly occur at different points in the cell cycle (WINEY et al. 1991 ; HARDWICK et al. 1996 ; WEISS and WINEY 1996 ).

Preventing microtubule polymerization or spindle pole body duplication leads to gross defects in spindle structure that clearly make normal chromosome segregation impossible. As a result, it has been possible to isolate conditional lethal alleles in genes involved in spindle pole body duplication and microtubule polymerization. In contrast, mutants that lack single microtubule motors are viable, suggesting that there is considerable overlap between the functions of different motors involved in spindle and chromosome segregation (HOYT et al. 1992 ; ROOF et al. 1992 ). Such overlap may reflect the formidable task of assembling an enormous macromolecular assembly that will capture, align, and then accurately segregate the chromosomes. The spindle checkpoint may help spindles function correctly in the presence of environmental perturbations, stochastic defects in chromosome alignment, and mutant alleles of genes involved in spindle function. The interaction of motor mutants with the spindle checkpoint supports this idea. All double-mutant combinations of mad mutations with cin8 or kar3, except mad3 kar3, show greatly reduced spore viability, indicating that the survival of strains with nonlethal defects in the spindle depends on the action of the spindle checkpoint.

The final spindle component we examined was the kinetochore, the specialized structure that attaches chromosomes to microtubules. Evidence in insect spermatoctyes (LI and NICKLAS 1995 ), mammalian tissue culture cells (RIEDER et al. 1995 ), and budding yeast (WELLS and MURRAY 1996 ) indicates that normal kinetochores that are not attached to or improperly aligned on the spindle can activate the spindle checkpoint. In addition, a subset of mutations in the centromeric DNA that impair kinetochore function (SPENCER and HIETER 1992 ), as well as a temperature-sensitive mutation in CTF13, an essential kinetochore function, both induce a mitotic delay that depends on the spindle checkpoint (WANG and BURKE 1995 ; PANGILINAN and SPENCER 1996 ). We tested whether elimination of a nonessential kinetochore component engaged the checkpoint. The Cbf1 protein binds to the CDEI element of centromeric DNA, and although this interaction is not essential for viability, disrupting it increases the frequency of chromosome loss (CAI and DAVIS 1990 ). The mad1 cbf1 and mad2 cbf1 double mutants show greatly reduced spore viability, suggesting that the survival of cbf1 mutants depends on the ability of the spindle checkpoint to delay anaphase until chromosomes with compromised kinetochore function have aligned correctly on the spindle.

How many different aspects of the spindle does the spindle checkpoint monitor? If there were multiple detection systems that each monitored a single feature of the spindle, different checkpoint mutants could inactivate different detection systems, with the result that the different mutants would fail to respond to different subsets of spindle defects. For example, a mutation that inactivated surveillance of the spindle pole body but had no effect on kinetochore monitoring would not arrest the cell cycle in cells with unduplicated spindle pole bodies, but it would still arrest in response to defects at the kinetochore. We have not observed this type of qualitative specificity. The mad1, mad2, and mad3 mutations show similar interactions with a wide range of spindle defects. The only exception to this conclusion is the observation that mad3 mutations allow the survival of kar3 or cbf1 mutants, whereas the mad1 and mad2 mutations do not. This result is not easy to interpret. In a quantitative assay, interactions of different strengths can easily be compared with each other. In a qualitative assay, however, such as the assessment of whether two mutations are synthetically lethal with each other, a continuous variable (the strength of the interaction between two mutations) is converted into an all-or-none output. Thus, there are two possible interpretations of the observation that mad3 is not synthetically lethal with kar3 or cbf1: (i) compared to wild-type cells, mad3 reduces the delay that these mutations cause, but the reduced delay is still sufficient to allow a colony to grow; or (ii) the delay these mutations cause is the same as that in wild-type cells. Distinguishing between these possibilities is likely to require conditional alleles of checkpoint genes.

Although we cannot rule out the future discovery of mutants that fail to respond to a specific spindle defect, we favor the idea that the spindle checkpoint monitors only one aspect of the spindle, the interaction between kinetochores and microtubules. Three arguments support this hypothesis. The first is that defects in microtubule polymerization, formation of a bipolar spindle, microtubule motors, and the kinetochore will all impair the ability of the kinetochore to capture and move along microtubules. The second is that the ndc10-1 mutation, which disrupts kinetochore function, abolishes the ability of nocodazole to arrest the budding yeast cell cycle (TAVORMINA et al. 1997 ). The third is that many components of the spindle checkpoint have been found at the kinetochore and respond to changes in kinetochore-microtubule interactions. One is an unidentified phosphoepitope, detected with the 3F3/2 antibody, which is present only on those kinetochores that are not both attached to microtubules and under tension (GORBSKY and RICKETTS 1993 ; NICKLAS et al. 1995 ). The others are the vertebrate homologs of the budding yeast Mad1 (CHEN et al. 1998 ; JIN et al. 1998 ), Mad2 (CHEN et al. 1996 ; LI and BENEZRA 1996 ), Bub1 (TAYLOR and MCKEON 1997 ; CHAN et al. 1998 ; JABLONSKI et al. 1998 ), and Bub3 (BASU et al. 1998 ; TAYLOR et al. 1998 ) checkpoint proteins. All these proteins are preferentially recruited to those kinetochores that have not attached to microtubules, suggesting that their recruitment plays an important role in monitoring the interaction between kinetochores and microtubules. Future studies should provide molecular details of how cells monitor the interaction between kinetochores and microtubules and then use this information to regulate the progress of the cell cycle.

	FOOTNOTES

¹ Present address: Institute of Cellular and Molecular Biology, University of Edinburgh, Edinburgh EH9 3JR, Scotland.
² Present address: Cell Biology, Harvard Medical School, Boston, MA 02115.
³ Present address: Biochemistry, Molecular and Cell Biology, Cornell University, Ithaca, NY 14853.

	ACKNOWLEDGMENTS

We thank Lee Hartwell, Dan Burke, Mark Winey, Eric Weiss, Tim Stearns, Pamela Forman, and members of the Murray lab for supplying strains, plasmids, and helpful discussions. We are grateful for the helpful comments of two anonymous reviewers. This work was supported by grants from the National Institutes of Health, the Human Frontiers in Science Program, the Markey Charitable Trust, the Packard Foundation, and the March of Dimes.

Manuscript received August 28, 1998; Accepted for publication March 11, 1999.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ALLSHIRE, R. C., 1997  Centromeres, checkpoints and chromatid cohesion. Curr. Opin. Genet. Dev. 7:264-273[Medline].

ASHFAR, K., N. R. BARTON, R. S. HAWLEY, and L. S. B. GOLDSTEIN, 1995  DNA binding and meiotic chromosomal localization of the Drosophila Nod kinesin-like protein. Cell 81:129-138[Medline].

BASU, J., E. LOGARINHO, S. HERRMANN, H. BOUSBA, and Z. LI et al., 1998  Localization of the Drosophila checkpoint control protein bub3 to the kinetochore requires bub1 but not Zw10 or Rod. Chromosoma 107:376-385[Medline].

BAUM, P., C. FURLONG, and B. BYERS, 1986  Yeast gene required for spindle pole body duplication: homology of its product with Ca²⁺-binding proteins. Proc. Natl. Acad. Sci. USA 83:5512-5516[Abstract/Free Full Text].

CAI, M. and R. W. DAVIS, 1990  Yeast centromere binding protein CBF1, of the helix-loop-helix protein family, is required for chromosome stability and methionine prototrophy. Cell 61:437-446[Medline].

CHAN, G. K., B. T. SCHAAR, and T. J. YEN, 1998  Characterization of the kinetochore binding domain of CENP-E reveals interactions with the kinetochore proteins CENP-F and hBUBR1. J. Cell Biol. 143:49-63[Abstract/Free Full Text].

CHEN, R.-H., J. C. WATERS, E. D. SALMON, and A. W. MURRAY, 1996  Association of spindle assembly checkpoint component XMAD2 with unattached kinetochores. Science 274:242-246[Abstract/Free Full Text].

CHEN, R.-H., A. SHEVCHENKO, M. MANN, and A. W. MURRAY, 1998  Metaphase arrest induced by an excess of the spindle checkpoint protein Xmad2 is independent of Xmad1. J. Cell Biol. 143:283-285[Abstract/Free Full Text].

CIOSK, R., W. ZACHARIAE, C. MICHAELIS, A. SHEVCHENKO, and M. MANN et al., 1998  An ESP1/PDS1 complex regulates loss of sister chromatid metaphase to anaphase transition in yeast. Cell 93:1067-1076[Medline].

DAWSON, I. A., S. ROTH, and S. ARTAVANIS-TSAKONAS, 1995  The Drosophila cell cycle gene fizzy is required for normal degradation of cyclins A and B during mitosis and has homology to the CDC20 gene of Saccharomyces cerevisiae.. J. Cell Biol. 129:725-737[Abstract/Free Full Text].

ENOS, A. P. and N. R. MORRIS, 1990  Mutation of a gene that encodes a kinesin-like protein blocks nuclear division in A. nidulans. Cell 60:1019-1027[Medline].

FANG, G., H. YU, and M. W. KIRSCHNER, 1998a  The checkpoint protein MAD2 and the mitotic regulator CDC20 form a ternary complex with the anaphase-promoting complex to control anaphase initiation. Genes Dev. 12:1871-1883[Abstract/Free Full Text].

FANG, G., H. YU, and M. W. KIRSCHNER, 1998b  Direct binding of CDC20 protein family members activates the anaphase-promoting complex in mitosis and G1. Mol. Cell 2:163-171[Medline].

FUNABIKI, H., H. YAMANO, K. KUMADA, K. NAGAO, and T. HUNT et al., 1996  Cut2 proteolysis required for sister-chromatid separation in fission yeast. Nature 381:438-441[Medline].

GARVIK, B., M. CARSON, and L. HARTWELL, 1995  Single-stranded DNA arising at telomeres in cdc13 mutants may constitute a specific signal for the RAD9 checkpoint. Mol. Cell. Biol. 15:6128-6138. (erratum: Mol. Biol. Cell. 16: 457).[Abstract].

GEISER, J. R., E. J. SCHOTT, T. J. KINGSBURY, N. B. COLE, and L. J. TOTIS et al., 1997  Saccharomyces cerevisiae genes required in the absence of the CIN8-encoded spindle motor act in functionally diverse mitotic pathways. Mol. Biol. Cell. 8:1035-1050[Abstract].

GORBSKY, G. J. and W. A. RICKETTS, 1993  Differential expression of a phosphoepitope at the kinetochores of moving chromosomes. J. Cell Biol. 122:1311-1321[Abstract/Free Full Text].

HAGAN, I. and M. YANAGIDA, 1990  Novel potential mitotic motor protein encoded by the fission yeast cut7+ gene. Nature 347:563-566[Medline].

HAGAN, I. and M. YANAGIDA, 1992  Kinesin-related cut7 protein associates with mitotic and meiotic spindles in fission yeast. Nature 356:74-76[Medline].

HARDWICK, K. and A. W. MURRAY, 1995  Mad1p, a phosphoprotein component of the spindle assembly checkpoint in budding yeast. J. Cell Biol. 131:709-720[Abstract/Free Full Text].

HARDWICK, K. G., E. WEISS, F. C. LUCA, M. WINEY, and A. W. MURRAY, 1996  Activation of the budding yeast spindle assembly checkpoint without mitotic spindle disruption. Science 273:953-956[Abstract].

HOLLOWAY, S. L., M. GLOTZER, R. W. KING, and A. W. MURRAY, 1993  Anaphase is initiated by proteolysis rather than by the inactivation of MPF. Cell 73:1393-1402[Medline].

HOYT, M. A. and J. R. GEISER, 1996  Genetic analysis of the mitotic spindle. Annu. Rev. Genet. 30:7-33[Medline].

HOYT, M. A., T. STEARNS, and D. BOTSTEIN, 1990  Chromosome instability mutants of Saccharomyces cerevisiae that are defective in microtubule-mediated processes. Mol. Cell. Biol. 10:223-234[Abstract/Free Full Text].

HOYT, M. A., L. TROTIS, and B. T. ROBERTS, 1991  S. cerevisiae genes required for cell cycle arrest in response to loss of microtubule function. Cell 66:507-517[Medline].

HOYT, M. A., L. HE, K. K. LOO, and W. S. SAUNDERS, 1992  Two Saccharomyces cerevisiae kinesin-related gene products required for mitotic spindle assembly. J. Cell Biol. 118:109-120[Abstract/Free Full Text].

HUFFAKER, T. C., J. H. THOMAS, and D. BOTSTEIN, 1988  Diverse effects of ß-tubulin mutations on microtubule formation and function. J. Cell Biol. 106:1997-2010[Abstract/Free Full Text].

HUNT, T., F. C. LUCA, and J. V. RUDERMAN, 1992  The requirements for protein synthesis and degradation, and the control of the destruction of cyclin A and cyclin B in the meiotic and mitotic cell cycles of the clam embryo. J. Cell Biol. 116:707-724[Abstract/Free Full Text].

HWANG, L. H., L. F. LAU, D. L. SMITH, C. A. MISTROT, and K. G. HARDWICK et al., 1998  Budding yeast Cdc20: a target of the spindle checkpoint. Science 279:1041-1044[Abstract/Free Full Text].

HYMAN, A. A. and T. J. MITCHISON, 1991  Two different microtubule based motor activities with opposite polarities in kinetochores. Nature 351:206-211[Medline].

HYMAN, A. A. and P. K. SORGER, 1995  Structure and function of kinetochores in budding yeast. Annu. Rev. Cell Dev. Biol. 11:471-495[Medline].

IRNIGER, S., S. PIATTI, C. MICHAELIS, and K. NASMYTH, 1995  Genes involved in sister chromatid separation are needed for B-type cyclin proteolysis in budding yeast. Cell 77:1037-1050.

JABLONSKI, S. A., G. K. T. CHAN, C. A. COOKE, W. C. EARNSHAW, and T. J. YEN, 1998  The hBUB1 and hBUBR1 kinases sequentially assemble onto kinetochores during prophase with hBUBR1 concentrating at the kinetochore plates in mitosis. Chromosoma 107:386-396[Medline].

JACOBS, C. W., A. E. M. ADAMS, P. J. SZANISZIO, and J. R. PRINGLE, 1988  Functions of microtubules in Saccharomyces cerevisiae cell cycle. J. Cell Biol. 107:1409-1426[Abstract/Free Full Text].

JASPERSEN, S. L., J. F. CHARLES, R. L. TINKER-KULBERG, and D. O. MORGAN, 1998  A late mitotic regulatory network controlling cyclin destruction in Saccharomyces cerevisiae. Mol. Biol. Cell 9:2803-2817[Abstract/Free Full Text].

JIN, D. Y., F. SPENCER, and K. T. JEANG, 1998  Human T cell leukemia virus type 1 oncoprotein Tax targets the human mitotic checkpoint protein MAD1. Cell 93:81-91[Medline].

KALLIO, M., J. WEINSTEIN, J. R. DAUM, D. J. BURKE, and G. J. GORBSKY, 1998  Mammalian p55CDC mediates association of the spindle checkpoint protein Mad2 with the cyclosome/anaphase-promoting complex, and is involved in regulating anaphase onset and late mitotic events. J. Cell Biol. 141:1393-1406[Abstract/Free Full Text].

KIM, S. H., D. P. LIN, S. MATSUMOTO, A. KITAZANO, and T. MATSUMOTO, 1998  Fission yeast slp1+ encodes the effector of the Mad2-dependent spindle checkpoint. Science 279:1045-1047[Abstract/Free Full Text].

KING, R. W., J. M. PETERS, S. TUGENDREICH, M. ROLFE, and P. HIETER et al., 1995  A 20S complex containing CDC27 and CDC16 catalyzes the mitosis-specific conjugation of ubiquitin to cyclin B. Cell 81:279-288[Medline].

KUMADA, K., T. NAKAMURA, K. NAGAO, H. FUNABIKI, and T. NAKAGAWA et al., 1998  Cut1 is loaded onto the spindle by binding to Cut2 and promotes anaphase spindle movement upon Cut2 proteolysis. Curr. Biol. 8:633-641[Medline].

LI, Y. and R. BENEZRA, 1996  Identification of a human mitotic checkpoint gene: hsMAD2. Science 274:246-248[Abstract/Free Full Text].

LI, R. and A. W. MURRAY, 1991  Feedback control of mitosis in budding yeast. Cell 66:519-531[Medline].

LI, X. and R. B. NICKLAS, 1995  Mitotic forces control a cell cycle checkpoint. Nature 373:630-632[Medline].

LI, M. H., N. R. ADAMES, M. D. MURPHY, C. R. SHIELDS, and J. A. COOPER, 1998  A cytokinesis checkpoint requiring the yeast homologue of an APC binding protein. Nature 393:487-491[Medline].

LOMBILLO, V. A., C. NISLOW, T. J. YEN, V. I. GELFAND, and J. R. MCINTOSH, 1995  Antibodies to the kinesin motor domain and CENP-E inhibit microtubule depolymerization-dependent motion of chromosomes in vitro [see comments]. J. Cell Biol. 128:107-115[Abstract/Free Full Text].

LYDALL, D. and T. WEINERT, 1995  Yeast checkpoint genes in DNA damage processing: implications for repair and arrest. Science 270:1488-1491[Abstract/Free Full Text].

MELUH, P. B. and M. D. ROSE, 1990  KAR3, a kinesin-related gene required for yeast nuclear fusion. Cell 60:1029-1041[Medline].

MINSHULL, J., H. SUN, N. K. TONKS, and A. W. MURRAY, 1994  MAP-kinase dependent mitotic feedback arrest in Xenopus egg extracts. Cell 79:475-486[Medline].

MINSHULL, J., A. STRAIGHT, A. RUDNER, A. DEMBURG, and A. BELMONT et al., 1996  Protein phosphatase 2A regulates MPF activity and sister chromatid cohesion in budding yeast. Curr. Biol. 6:1609-1620[Medline].

NICKLAS, R. B., S. C. WARD, and G. J. GORBSKY, 1995  Kinetochore chemistry is sensitive to tension and may link mitotic forces to a cell cycle checkpoint. J. Cell Biol. 130:929-939[Abstract/Free Full Text].

O'TOOLE, E. T., D. N. MASTRONARDE, T. H. GIDDINGS, JR., M. WINEY, and D. J. BURKE et al., 1997  Three-dimensional analysis and ultrastructural design of mitotic spindles from the cdc20 mutant of Saccharomyces cerevisiae. Mol. Biol. Cell 8:1-11[Abstract].

PANGILINAN, F. and F. SPENCER, 1996  Abnormal kinetochore structure activates the spindle assembly checkpoint in budding yeast. Mol. Biol. Cell. 7:1195-1208[Abstract].

RIEDER, C. L., R. W. COLE, A. KHODJAKOV, and G. SLUDER, 1995  The checkpoint delaying anaphase in response to chromosome monoorientation is mediated by an inhibitory signal produced by unattached kinetochores. J. Cell Biol. 130:941-948[Abstract/Free Full Text].

ROOF, D. M., P. B. MELUH, and M. D. ROSE, 1991  Multiple kinesin-related proteins in yeast mitosis. Cold Spring Harbor Symp. Quant. Biol. 56:693-703[Abstract/Free Full Text].

ROOF, D. M., P. B. MELUH, and M. D. ROSE, 1992  Kinesin-related proteins required for assembly of the mitotic spindle. J. Cell Biol. 118:95-108[Abstract/Free Full Text].

RUDNER, A. D. and A. W. MURRAY, 1996  The spindle assembly checkpoint. Curr. Opin. Cell Biol. 8:773-780[Medline].

SCHWAB, M., A. S. LUTUM, and W. SEUFERT, 1997  Yeast Hct1 is a regulator of Clb2 cyclin proteolysis. Cell 90:683-693[Medline].

SCHWEITZER, B. and P. PHILIPPSEN, 1991  CDC15, an essential cell cycle gene in Saccharomyces cerevisiae, encodes a protein kinase domain. Yeast 7:265-273[Medline].

SHERMAN, F., G. FINK and C. LAWRENCE, 1974 Methods in Yeast Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SIGRIST, S., H. JACOBS, R. STRATMANN, and C. F. LEHNER, 1995  Exit from mitosis is regulated by Drosophila fizzy and the sequential destruction of cyclins A, B and B3. EMBO J. 14:4827-4838[Medline].

SPENCER, F. and P. HIETER, 1992  Centromere DNA mutations induce a mitotic delay in Saccharomyces cerevisiae.. Proc. Natl. Acad. Sci. USA 89:8908-8912[Abstract/Free Full Text].

STEARNS, T., 1997  Motoring to the finish: kinesin and dynein work together to orient the yeast mitotic spindle [commentary]. J. Cell Biol. 138:957-960[Free Full Text].

STEARNS, T., M. A. HOYT, and D. BOTSTEIN, 1990  Yeast mutants sensitive to antimicrotubule drugs define three genes that affect microtubule function. Genetics 124:251-262[Abstract].

STRAIGHT, A. F., J. W. SEDAT, and A. W. MURRAY, 1998  Time-lapse microscopy reveals unique roles for kinesins during anaphase in budding yeast. J. Cell Biol. 143:687-694[Abstract/Free Full Text].

SUDAKIN, V., D. GANOTH, A. DAHAN, H. KELLER, and J. HERSKO et al., 1995  The cyclosome, a large complex containing cyclin-selective ubiquitination ligase activity, targets cyclins for destruction at the end of mitosis. Mol. Biol. Cell 6:185-198[Abstract].

SULLIVAN, D. S. and T. C. HUFFAKER, 1992  Astral microtubules are not required for anaphase B in Saccharomyces cerevisiae.. J. Cell Biol. 119:379-388[Abstract/Free Full Text].

TAVORMINA, P. A., Y. WANG, and D. J. BURKE, 1997  Differential requirements for DNA replication in the activation of mitotic checkpoints in Saccharomyces cerevisiae. Mol. Cell. Biol. 17:3315-3322[Abstract].

TAYLOR, S. S. and F. MCKEON, 1997  Kinetochore localization of murine Bub1 is required for normal mitotic timing and checkpoint response to spindle damage. Cell 89:727-735[Medline].

TAYLOR, S. S., E. HA, and F. MCKEON, 1998  The human homologue of Bub3 is required for kinetochore localization of Bub1 and a Mad3/Bub1-related protein kinase. J. Cell Biol. 142:1-11[Abstract/Free Full Text].

VISINTIN, R., S. PRINZ, and A. AMON, 1997  CDC20 and CDH1: a family of substrate-specific activators of APC-dependent proteolysis. Science 278:460-463[Abstract/Free Full Text].

WALCZAK, C