Report

Role of Polo-like Kinase CDC5 in Programming Meiosis I Chromosome Segregation

See allHide authors and affiliations

Science  18 Apr 2003:
Vol. 300, Issue 5618, pp. 482-486
DOI: 10.1126/science.1081846

Abstract

Meiosis is a specialized cell division in which two chromosome segregation phases follow a single DNA replication phase. The budding yeast Polo-like kinase Cdc5 was found to be instrumental in establishing the meiosis I chromosome segregation program. Cdc5 was required to phosphorylate and remove meiotic cohesin from chromosomes. Furthermore, in the absence of CDC5 kinetochores were bioriented during meiosis I, and Mam1, a protein essential for coorientation, failed to associate with kinetochores. Thus, sister-kinetochore coorientation and chromosome segregation during meiosis I are coupled through their dependence on CDC5.

Sexually reproducing diploid organisms rely on a specialized cell cycle, meiosis, for gamete formation. During meiosis, DNA replication is followed by two rounds of chromosome segregation, in which homologs segregate during the first and sister chromatids split in the second. For this unusual chromosome segregation program to occur accurately, several meiosis-specific events must take place (13). First, reciprocal recombination between homologs generates linkages called chiasmata, which ensure that homologs are accurately aligned on the metaphase I spindle. Second, sister kinetochores face the same spindle pole (coorientation) to facilitate sister-chromatid cosegregation during anaphase I. Before the second meiotic division, coorientation is lost, and sister kinetochores attach to microtubules from opposite spindle poles (biorientation), which then separate sister chromatids during anaphase II (4). Lastly, cohesin complexes that hold sister chromatids together are lost in a stepwise manner. Loss of arm cohesion allows for the resolution of chiasmata and segregation of homologs during meiosis I. Retention of centromeric cohesion ensures that sister chromatids properly align on the meiosis II spindle (5). Sister chromatids separate at the onset of anaphase II when centromeric cohesion is lost (4).

To gain insight into how the meiotic segregation pattern is established, we examined the role of Polo kinase (Cdc5 in budding yeast), which has been implicated in regulating meiosis in many organisms (69). We generated a meiosis-specific null allele of CDC5 by replacing the CDC5 promoter with the CLB2 promoter, which is expressed in mitosis but is repressed during meiosis (10) [pCLB2-CDC5 (11)]. Because Cdc5 is degraded during G1 (12, 13), it is absent during meiosis in pCLB2-CDC5 cells (Fig. 1A). Cdc5-depleted cells progressed through pre-meiotic S phase (Fig. 1B) and entered metaphase I as wild type but arrested in metaphase I (Fig. 1C).

Fig. 1.

Cdc5 or Cdc20 depletion causes a metaphase I arrest. (A) pCLB2-3HA-CDC5 (A5120) and pCLB2-3HA-CDC20 (A5128) cells were sporulated (11) and 3HA-Cdc5, 3HA-Cdc20, and Kar2 (loading control) were analyzed by Western blot analysis (31). (B, C, E) Wild-type (A4964), pCLB2-CDC5 (A5844), and pCLB2-CDC20 (A5823) cells carrying PDS1-18MYC and REC8-3HA fusions were sporulated (11). DNA content (B), two or more nuclei [MI; closed symbols in (C), left], three or four nuclei [MII; open symbols in (C), left], metaphase I spindles [(C), middle], and sum of anaphase I and meiosis II spindles [(C), right] were determined. Pds1-18MYC, Rec8-3HA, and Pgk1 (loading control) were monitored by Western blot analysis (E) (32). (D) Strains in Fig. 1B and a Pds1-18MYC strain (A3685, no tag) were collected 6 hours after induction of sporulation and subjected to CIP (calf intestinal alkaline phosphatase) treatment as described (11).

To investigate why cells lacking Cdc5 failed to enter anaphase I, we determined whether Pds1 was stabilized. Pds1 inhibits the protease Esp1, and its degradation leads to the activation of Esp1 (2). Esp1 triggers anaphase by cleaving a subunit of the cohesin complex (14, 15). Because CDC20 is required for Pds1 degradation during mitosis and meiosis (13, 16), we generated an allele of CDC20 that was depleted during meiosis in the same manner as Cdc5 (pCLB2-CDC20) (11). In wild-type cells, Pds1 protein levels peaked during metaphase I and declined during anaphase I. The second wave of Pds1 in meiosis II was not as evident, owing to the poor synchrony of meiotic cultures (Fig. 1E). Cdc20-depleted cells arrested in metaphase I (Fig. 1C) with high levels of Pds1 (Fig. 1E), indicating that CDC20 was required for Pds1 degradation in meiosis. In cells lacking Cdc5, Pds1 accumulated to higher levels and degradation was delayed, the extent of which varied between experiments (Fig. 1E; fig. S1A). This delay in Pds1 degradation, however, was not the sole reason for the metaphase I arrest observed in Cdc5-depleted cells because these pCLB2-CDC5 cells lacking PDS1 still largely arrested in metaphase I (fig. S1B). Thus, the delay in Pds1 degradation was not the sole reason for the metaphase I arrest in Cdc5-depleted cells.

During mitosis, Cdc5 phosphorylates the cohesin subunit Scc1/Mcd1, thereby promoting its efficient cleavage by Esp1 (17). We found that CDC5 was also required for the efficient phosphorylation of Rec8, which replaces Scc1/Mcd1 during meiosis (5). Rec8 phosphorylation was reduced but not completely abolished in cells lacking Cdc5 compared with Cdc20-depleted or wild-type cells (Fig. 1, D and E; fig. S1A). CDC5 was also required for Rec8 cleavage (Fig. 2, A and B), which is a prerequisite for anaphase I entry (15) and the dissociation of Rec8 from chromosomes (Fig. 2C; fig. S2). Thus, CDC20 and CDC5 were necessary for efficient Rec8 cleavage, but CDC5 and not CDC20 was required for complete Rec8 phosphorylation.

Fig. 2.

Rec8 cleavage and removal from chromosome arms depends on CDC5 and CDC20. (A and B) ubr1Δ (A2704), ubr1Δ pCLB2-CDC5 (A6143), and ubr1Δ pCLB2-CDC20 (A6144) cells carrying a REC8-3HA fusion were sporulated (11). Two or more nuclei [MI; closed symbols in (A), left], three or four nuclei [MII; open symbols in (A), left], metaphase I spindles [(A), middle], and sum of anaphase I and meiosis II spindles [(A), right] were determined. Kar2 and Rec8-3HA were analyzed by Western blotting (B). (C) Wild-type (A4758), pCLB2-CDC5 (A6136), and pCLB2-CDC20 (A6586) cells carrying REC8-3HA and NDC10-13MYC fusions (NDC10 encodes a kinetochore component) were sporulated (11). The percentage of cells with one nucleus (♢), two or more nuclei (◯), or three and four nuclei (Δ) was determined. The percentage of Ndc10-positive cells with Rec8 associated with the entire chromosome (♦), withRec8 absent from chromosome arms (⚫), or withno Rec8 staining (▲) was determined (33).

Removal of chromosome arm cohesion is essential for chiasmata resolution, which in turn is a prerequisite for homolog segregation (15). In the absence of chiasmata, cells undergo anaphase I with homologs segregated in a random manner even in the absence of Rec8 cleavage and separase activity (15). If the only role of CDC20 and CDC5 in promoting anaphase I entry were to induce cohesin removal, abolishing chiasmata formation should render CDC20 and CDC5 dispensable for anaphase I entry. To eliminate chiasmata, we deleted SPO11, a gene required to initiate meiotic recombination and chiasmata formation (18, 19). Deletion of SPO11 allowed Cdc20-depleted cells to progress through meiosis I and into metaphase II (fig. S3A), where they arrested because cohesin removal is essential for sister-chromatid segregation (15). Thus, the failure to remove cohesion distal to chiasmata was the sole reason why Cdc20-depleted cells did not enter anaphase I. Deletion of SPO11 did not allow Cdc5-depleted cells to enter anaphase I (fig. S3A), indicating that Rec8 removal from chromosomes was not the only defect preventing anaphase I entry in these cells.

Insight into the cause of the metaphase I arrest of spo11Δ pCLB2-3HA-CDC5 cells came from our analysis of these cells lacking REC8. Deletion of REC8 allowed spo11Δ pCLB2-3HA-CDC5 cells to undergo anaphase I (fig. S3B). The kinetics with which spo11Δ rec8Δ pCLB2-3HA-CDC5 cells entered anaphase I was the same as in spo11Δ rec8Δ cells, indicating that loss of CDC5 function did not interfere with kinetochore attachment or spindle elongation (fig. S3B). spo11Δ rec8Δ pCLB2-3HA-CDC5 cells then arrested in anaphase I with the protein phosphatase Cdc14 sequestered in the nucleolus (figs. S3B and S4), indicating that, as in mitosis, CDC5 was required for Cdc14 release from the nucleolus and anaphase spindle disassembly (2023).

Why does deletion of REC8 allow Cdc5-depleted cells to enter anaphase I but deletion of SPO11 does not? One explanation for this observation is that sister kinetochores are bioriented and not cooriented during metaphase I in Cdc5-depleted cells (Fig. 3A). Because loss of sister-chromatid cohesion is prevented in Cdc5-depleted cells (Fig. 2C), biorientation of sister kinetochores would prevent chromosome segregation even if the need to resolve chiasmata were eliminated by deleting SPO11 (Fig. 3A, second panel). In contrast, deleting REC8 would allow random chromosome segregation (Fig. 3A, third panel). To determine whether kinetochores were bioriented in Cdc5-depleted cells, we replaced REC8 with SCC1/MCD1 (pREC8-SCC1-3HA) (4). Removal of Scc1/Mcd1 from chromosomes does not absolutely require CDC5 (17), and it is lost along the entire length of the chromosome during meiosis I (4). Thus, replacing Rec8 with Scc1/Mcd1 should allow spo11Δ pCLB2-CDC5 cells to undergo anaphase I. Crucially, if sister kinetochores were bioriented, sister chromatids should segregate from each other in this division (Fig. 3A, fourth panel).

Fig. 3.

CDC5 is required for sister-kinetochore coorientation. (A) Predictions of a model in which sister kinetochores are bioriented in Cdc5-depleted cells (see text). (B) Control (A6245), pCLB2-CDC20 (A6321), pCLB2-CDC5 (A6342), and spo12Δ (A8279) cells containing spo11Δ rec8Δ pREC8-SCC1-3HA and heterozygous URA3 dot (11) were sporulated (11). The percentage of binucleates (◼), tri/tetranucleates (♦), homolog segregation with GFP label in only one of the two nuclei (△), and sister-chromatid segregation withGFP label in bothnuclei (◯) was determined. (C) pCLB2-CDC20 (◯, A7118), pCLB2-3HA-CDC5 (▲, A7151), and pCLB2-CDC20 mam1Δ (▢, A7316) cells withheterozygous CENV dots (11) were sporulated (11). Metaphase I spindles (left) and separation of CENV dots (right) were determined. (D) pCLB2-CDC20 (◯, A7117), pCLB2-3HA-CDC5 (▲, A5762), and pCLB2-CDC20 mam1Δ (▢, A7460) strains containing a heterozygous URA3 dot (11) were sporulated (11). The percentage of cells with metaphase I spindles (left) and the separation of URA3 dots (right) were determined.

To distinguish between sister chromatids and pairs of sister chromatids (homologs) segregating, we integrated an array of TET operators on one homolog of chromosome V [URA3 green fluorescent protein (GFP) dots] and expressed a Tet repressor–GFP fusion, which binds to these sites (5, 24). If sister chromatids segregated during anaphase I, each nucleus should contain one GFP dot. Random segregation of homologs would produce a GFP signal in only one nucleus because the sister chromatids remained paired. Control cells (spo11Δ pREC8-SCC1-3HA) and spo11Δ pREC8-SCC1-3HA cells depleted for Cdc20 segregated homologs as expected (Fig. 3B). In contrast, when Cdc5 was depleted in spo11Δ pREC8-SCC1-3HA cells, sister chromatids were segregated (Fig. 3B). We also examined the segregation pattern of chromosome V in spo11Δ pREC8-SCC1-3HA cells lacking SPO12. SPO12, like CDC5, is a component of the FEAR network, which is essential for preventing meiosis II chromosome segregation when meiosis I spindle disassembly is inhibited (22, 23) (see legend to fig. S5). spo11Δ pREC8-SCC1-3HA spo12Δ cells segregated homologs rather than sister chromatids (Fig. 3B). Thus, pCLB2-CDC5 but not pCLB2-CDC20 or spo12Δ cells attempted to segregate sister chromatids during the first meiotic division, indicating that sister kinetochores were bioriented instead of cooriented in Cdc5-depleted cells.

To visualize kinetochore orientation directly, we took advantage of the observation that in mitosis, bioriented sister centromeres separate as a result of the pulling force of the spindle without loss of cohesion (2527). During metaphase I, sister kinetochores are cooriented; their centromeres are not under tension and thus remain together. In wild-type metaphase I cells and Cdc20-depleted cells, chromosome V sister centromeres marked with Tet operator sequences 1.4 kb from the centromere (CENV dots) were not separated (Fig. 3C). When MAM1, a gene essential for kinetochore coorientation (4), was deleted in pCLB2-CDC20 cells, up to 50% of sister centromeres were separated (Fig. 3C; fig. S5). pCLB2-3HA-CDC5 cells separated sister kinetochores with the same kinetics as pCLB2-CDC20 mam1Δ cells (Fig. 3C; fig. S5B). Sister-centromere separation was also observed in pCLB2-3HA-CDC5 pCLB2-CDC20 cells but not in pCLB2-CDC20 spo12Δ or pCLB2-CDC20 cdc14-1 cells (fig. S5B), indicating that sister-centromere separation is not a FEAR network defect. The separation of CENV dots reflected sister-centromere separation and not loss of sister-chromatid cohesion because GFP dots located 35 kb away from CENV (URA3 dots) remained together (Fig. 3D). Thus, CDC5 but not the FEAR network was required for sister-kinetochore coorientation during meiosis I.

To determine the mechanism whereby CDC5 controlled sister-kinetochore orientation during meiosis I, we examined the localization of Mam1, the only known protein required for this process (4). In wild-type cells, Mam1 localized to the nucleus and to kinetochores in late prophase and metaphase I (Fig. 4) (4). In Cdc20-depleted cells, Mam1 localized to the nucleus and associated with and dissociated from kinetochores with wild-type kinetics, despite cells arresting in metaphase I, indicating that entry into anaphase I was not a prerequisite for loss of Mam1 from kinetochores (Fig. 4). Mam1 accumulated in the nucleus of Cdc5-depleted cells but did not associate with kinetochores (Fig. 4A). Instead, Mam1 was found in aggregates that associated with noncentromeric chromosomal regions (Fig. 4B). CDC5 was also necessary for efficient phosphorylation of Mam1 (fig. S6). Thus, CDC5 was required for Mam1 phosphorylation and localization to kinetochores.

Fig. 4.

CDC5 is required for Mam1 kinetochore association. (A and B) Wild-type (A7097), pCLB2-CDC5 (A7449), and pCLB2-CDC20 (A7450) cells carrying MAM1-9MYC and NDC10-6HA fusions were sporulated (11). Metaphase I spindles (◼), sum of anaphase I and meiosis II spindles (♢), Mam 1 localized to the nucleus (blue circles), and Mam1 colocalized withNdc10 by chromosome spreads (red triangles) were determined. (B) Ndc10-6HA (red), Mam1-9MYC (green), and a merge of the two with 4′,6-diamidino-2-phenylindole (DAPI) in wild-type (left) and pCLB2-CDC5 (right) cells 5 hours after the induction of sporulation.

Our data show that CDC5 is required for (i) cleavage and removal of cohesin from chromosome arms at the onset of anaphase I, (ii) sister-kinetochore coorientation during meiosis I, and (iii) Cdc14 release from the nucleolus and anaphase I spindle disassembly. Cdc5 is likely to phosphorylate Rec8, thereby targeting it for cleavage, and probably causes anaphase I spindle disassembly by promoting the release of Cdc14 from the nucleolus. Cdc5 controls sister-kinetochore coorientation by promoting the localization of Mam1 to kinetochores, perhaps in part through phosphorylating Mam1. Although our data show that CDC5 is a key regulator of meiosis I, the protein kinase may collaborate with meiosis-specific factors such as Spo13 to accomplish this task. Spo13 is required for sister-kinetochore coorientation and to inhibit cohesin cleavage (2830). Together, these two proteins may modulate the mitotic chromosome segregation machinery to bring about the specialized meiosis I chromosome segregation program.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1081846/DC1

Materials and Methods

Figs. S1 to S6

Table S1

References

References and Notes

View Abstract

Navigate This Article