Report

APC-Mediated Proteolysis of Ase1 and the Morphogenesis of the Mitotic Spindle

See allHide authors and affiliations

Science  28 Feb 1997:
Vol. 275, Issue 5304, pp. 1311-1314
DOI: 10.1126/science.275.5304.1311

Abstract

The molecular mechanisms that link cell-cycle controls to the mitotic apparatus are poorly understood. A component of the Saccharomyces cerevisiae spindle, Ase1, was observed to undergo cell cycle-specific degradation mediated by the cyclosome, or anaphase promoting complex (APC). Ase1 was degraded when cells exited from mitosis and entered G1. Inappropriate expression of stable Ase1 during G1 produced a spindle defect that is sensed by the spindle assembly checkpoint. In addition, loss of ASE1 function destabilized telophase spindles, and expression of a nondegradable Ase1 mutant delayed spindle disassembly. APC-mediated proteolysis therefore appears to regulate both spindle assembly and disassembly.

Cell cycle-specific proteolysis was first discovered as a mechanism for inactivation of mitotic cyclins. B-type cyclins are degraded through the ubiquitin-proteasome pathway (13). The cell-cycle specificity of this process comes from the ubiquitination reaction and not from the degradation by the proteasome. Ubiquitination requires the activation of ubiquitin by an E1 enzyme and its subsequent transfer to one of a family of ubiquitin-conjugating enzymes (E2 enzymes). Often, a third activity, the E3, is also required and is a determinant of substrate specificity (1). The E3 for proteolysis of mitotic cyclins is a multiprotein complex termed the cyclosome or anaphase promoting complex (47). Ubiquitination of mitotic cyclins requires a sequence motif termed the destruction box that is thought to be recognized by the APC (2, 8).

Although the role of the APC in cyclin proteolysis is well established, there is now mounting evidence that the APC has other cell-cycle functions. A requirement for the APC during sister chromatid separation was deduced from experiments in Xenopus laevis egg extracts and in yeast (5, 9, 10). Recently, two APC substrates required for sister chromatid separation have been identified: Cut2, in fission yeast, and Pds1, in budding yeast (11). A role for the APC in some aspect of DNA replication has also been inferred from the finding that alleles of Saccharomyces cerevisiae CDC16 and CDC27 allow more than one complete round of DNA replication during a single cell cycle (12).

We found that the APC regulates the mitotic apparatus by targeting a component of the yeast spindle, Ase1, for degradation. ASE1 (for anaphase spindle elongation) encodes a yeast nonmotor microtubule-binding protein [MAP (13)] that localizes to the anaphase spindle midzone, where spindle fibers from opposite poles overlap. Together with another nonmotor MAP, Ase1 is required for anaphase B, the elongation of the spindle and separation of the spindle poles (13).

Consistent with our genetic analysis that suggests a function for ASE1 late in mitosis, the pattern of ASE1 mRNA expression during the cell cycle closely paralleled that of the mitotic cyclin, CLB2 (Fig. 1A) (14). The amount of Ase1 protein is also regulated during the cell cycle. Like Clb2, Ase1 is not detected during G1 and is present in largest amounts during mitosis. Although Clb2 appears to be degraded during mid-anaphase, Ase1 persists throughout mitosis and then is abruptly lost as cells undergo cytokinesis (13, 15). The general similarities between the pattern of expression of Ase1 and Clb2 prompted us to determine if Ase1 is rapidly degraded during G1. The ASE1 coding sequence was placed under the control of the inducible GAL1 promoter, and Ase1 turnover in logarithmically growing cells and in G1 cells was then compared. Ase1 is rapidly degraded in G1 cells but is stable in cycling cells, with half-lives of 5 and 50 min, respectively [Fig. 1, B and C (16)]. Therefore, like mitotic cyclins, Ase1 expression is regulated by both transcriptional and proteolytic mechanisms.

Fig. 1.

Transcription and cell cycle-specific proteolysis of Ase1. (A) Northern analysis of ASE1, CLB2, and ACT1 (actin) mRNA at intervals after release of cells from arrest in G1 with the mating pheromone, α-factor (25). (B) Stability of 35S-labeled Ase1 in cycling cells and in α-factor-arrested G1 cells (26). The relative amount of Ase1 is indicated below each lane. ND indicates a time point where Ase1 was not detected. (C) Flow cytometry of cells at the last time point from (B) (27).

At least four of the APC subunits, CDC16, CDC23, CDC27, and BIME/APC1, are highly conserved between yeast and vertebrate cells (57, 15, 17). In yeast, mutations in CDC16, CDC23, CDC26, or APC1 block the rapid degradation of Clb2 in G1 cells. To determine if the G1-specific degradation of Ase1 requires the yeast APC, we measured the half-life of Ase1 in G1-arrested CDC23 and temperature-sensitive cdc23 strains at the nonpermissive temperature. Ase1 was stable in the G1-arrested cdc23 strain but not in the CDC23 strain (Fig. 2, A and B). Flow cytometry confirmed that these strains remained arrested in the G1 phase throughout the experiment (Fig. 2B).

Fig. 2.

Requirement of both CDC23 and a destruction box for proteolysis of Ase1. (A) Congenic MATa ase1 Δ bar1 CDC23 and MATa ase1Δ bar1 cdc23-1 strains (5) were grown in the presence or absence of α-factor at 24°C, then shifted to 36°C for 30 min, and the half-life of Ase1 was determined (26, 28). (B) Flow cytometry of cells from the last time point in (A). (C) Stability of Ase1 and Ase1-db expressed from the GAL1 promoter in MATa bar1 ase1Δ cells arrested in G1 with α-factor (29). (D) Flow cytometry of cells from the last time point in (C).

The Ase1 polypeptide contains five sequences with similarity to the cyclin destruction box (2, 18). To determine if any of these sequences is a functional destruction box, all of the conserved residues in each of these sequences were mutated to alanine. Mutation of only one of these sequences (amino acids 760 to 768, Embedded Image-Gln-Leu-Embedded Image-Pro-Ile- Pro-Leu-Embedded Image, changed to Embedded Image-Gln-Leu-Embedded Image- Pro-Ile-Pro-Leu-Embedded Image) produced an Ase1 protein that was stable in G1 cells (Fig. 2, C and D). This mutant, hereafter referred to as Ase1-db, is apparently functional because it both complements ase1Δ mutant phenotypes and displays normal localization. The fact that Ase1 degradation requires both CDC23 and a destruction box sequence suggests that Ase1 is an APC substrate.

Because of the localization of Ase1 to the spindle midzone, we reasoned that Ase1 might have a role in maintaining the interaction between the two half-spindles and further, that Ase1 degradation might have a role in spindle disassembly. Therefore, we tested whether loss of Ase1 altered the structure of the spindle in a mutant, cdc15, that blocks cell-cycle progression at the end of mitosis. ASE1 cdc15 and ase1Δ cdc15 strains were arrested at the nonpermissive temperature, and spindle structures were visualized by tubulin immunofluorescence. The ASE1 cdc15 strain arrested in telophase, with long spindles and segregated chromosomes (94% of cells, Fig. 3A, a and c). In contrast, although chromosome segregation was completed, the spindles in the ase1Δ cdc15 strain fell apart, leaving structures resembling the astral microtubules observed in G1 cells (91% of cells, Fig. 3A, b and d). These ase1Δ cdc15 cells remained arrested and did not undergo cytokinesis. This experiment demonstrates that in the absence of Ase1, telophase spindles disassemble.

Fig. 3.

Effects of ASE1 mutations on spindle disassembly. (A) ASE1 cdc15-2 (a and c) and ase1Δ cdc15-2 (b and d) strains were arrested at 37°C for 3 hours. Tubulin immunofluorescence (a and b) and 4′,6′-amidino-2-phenylindole (DAPI) staining (c and d) were done as described (13). (B) An ASE1 cdc15-2 strain containing either GAL1::ASE1 or GAL1::ASE1-db integrated at the LEU2 locus was arrested for 3 hours at 37°C, expression from the GAL1 promoter was induced for 3 hours with galactose, cells were then released from the cdc15 block into medium containing α-factor, and samples were taken at intervals for tubulin immunofluorescence. Two hundred cells were counted at each time point. The ASE1-db-expressing cells had a significantly higher percentage of telophase spindles than did the ASE1-expressing cells at 40, 50, and 60 min, with respective P values of 0.005, 0.001, and <0.001 (using a binomial comparison of the proportions). Similar results were observed in four independent experiments. (C) Amount of Ase1 protein from (B). Methods are as described (13).

Because Ase1 is required for the stability of telophase spindles, we tested whether the expression of nondegradable Ase1 at the end of mitosis would block or delay spindle disassembly. The expression of Ase1 or Ase1-db was induced in cdc15-arrested cells, and spindle morphology was determined at intervals after release to the permissive temperature in medium containing the mating pheromone α-factor to produce a G1 block. Expression of Ase1-db caused a delay in spindle disassembly (Fig. 3B). GAL1-driven expression of Ase1 had no effect on spindle disassembly (19, 20). Protein immunoblotting confirmed that Ase1 but not Ase1-db was degraded after release from the cdc15 block (Fig. 3C).

Next, we studied the effect of expression of Ase1-db in G1, when it is normally not expressed. ase1Δ strains containing either GAL1::ASE1, GAL1::ASE1-db, or a control vector were arrested in G1, expression from the GAL1 promoter was transiently induced, and the cells were released from the G1 block. In comparison with the cells expressing the control vector or Ase1, expression of Ase1-db in G1 delayed cell-cycle progression through mitosis. The cells transiently expressing Ase1-db accumulated early in mitosis as large budded cells containing undivided nuclei, short mitotic spindles, and 2N DNA content [Fig. 4A (21)]. This phenotype is similar to that observed when either wild-type Ase1 or the destruction box mutant is induced for long periods of time in unsynchronized cells (19).

Fig. 4.

Activation of the spindle assembly checkpoint from transient expression of Ase1-db in G1 cells. (A) MATa ase1Δ bar1 strains containing GAL1::ASE1, GAL1::ASE1-db, or a 2μ control vector were arrested with α-factor, and GAL1 expression was induced for 3 hours with galactose. GAL1 expression was then repressed with glucose, the cells were released from the G1 block by washing into fresh medium, and samples were taken at intervals for DAPI staining and differential interference contrast microscopy. The percentage of cells that are budded and contain a single nucleus (early mitosis) is indicated. (B) The transient expression experiment from (A) was repeated for strains with the following genotypes: GAL1::ASE1 MAD1, GAL1::ASE1-db MAD1, GAL1::ASE1 mad1Δ, and GAL1::ASE1-db mad1Δ.

A checkpoint mechanism has been identified that blocks or delays mitosis in the presence of abnormal spindles (22). To determine if the mitotic delay induced by the expression of Ase1-db is due to a spindle defect, we expressed Ase1-db in a strain lacking a component of the checkpoint, MAD1. Expression of Ase1-db in G1-arrested mad1Δ cells did not cause a mitotic delay (Fig. 4B). However, the absence of this delay resulted in decreased viability. Only 20% of mad1Δ cells transiently expressing ASE1-db were viable, whereas transient expression of ASE1 or the control vector had no effect on viability. This demonstrates that the expression of Ase1-db during G1 leads to spindle damage when cells subsequently enter mitosis. These observations provide an additional rationale for the surprising finding that APC-mediated proteolysis remains active during G1. APC-mediated proteolysis during G1 may prevent inappropriate expression of Clb2 that would inhibit budding (5, 23) and may also prevent inappropriate expression of Ase1, which would interfere with spindle assembly.

Our findings suggest that the APC may have two important roles in regulating the function of the mitotic spindle: it may mediate the disassembly of the spindle and prevent proteins that are normally assembled onto the spindle late from accumulating too early. Because the nondegradable Ase1 protein delays but does not block spindle disassembly, we expect that other proteins that contribute to the stability of the spindle are targeted for degradation by the APC. Indeed, two components of the mammalian mitotic apparatus, CENP-E and CENP-F, are degraded at the end of mitosis (24). The localization of APC subunits to the spindle also supports a general role for the APC in regulating the mitotic apparatus (7).

Our findings also broaden the scope of cellular processes under APC control. In addition to controlling the abundance of mitotic cyclins, the APC regulates sister chromatid cohesion, the cellular DNA content, and the function of the mitotic spindle (5, 1012). The APC proteolytic system may therefore be a global cell-cycle regulator much like the cyclin-dependent kinases.

REFERENCES AND NOTES

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
    Additionally, Ase1 binds taxol-stabilized microtubules in vo (S. Schuyler and D. Pellman, unpublished results). ASE1 is nonessential, but in the strain background used in this study, the ase1Δ mutation causes a temperature-sensitive growth defect.
  14. 14.
    Also, the ASE1 promoter (−287 to −251) contains a close match to a transcriptional element, the Mcm1-SFF binding site, that drives the G2-specific expression of CLB2 [
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
View Abstract

Navigate This Article