Requirement of Cks2 for the First Metaphase/Anaphase Transition of Mammalian Meiosis

See allHide authors and affiliations

Science  25 Apr 2003:
Vol. 300, Issue 5619, pp. 647-650
DOI: 10.1126/science.1084149


We generated mice lacking Cks2, one of two mammalian homologs of the yeast Cdk1-binding proteins, Suc1 and Cks1, and found them to be viable but sterile in both sexes. Sterility is due to failure of both male and female germ cells to progress past the first meiotic metaphase. The chromosomal events up through the end of prophase I are normal in both CKS2–/– males and females, suggesting that the phenotype is due directly to failure to enter anaphase and not a consequence of a checkpoint-mediated metaphase I arrest.

Like the mitotic cell cycle, the meiotic cell cycle is controlled by regulating the activity of maturation promoting factor (MPF), the complex of cyclin B and Cdk1. But the need to produce a haploid cell has necessitated unique modifications to the cell cycle. Two aspects of meiosis that distinguish it from mitosis are the behavior of sister chromatids in meiosis I and the transition from one metaphase (metaphase I) to a second (metaphase II) without intervening DNA synthesis (13). Changes in cell-cycle regulation observed in meiosis are brought about in part by modifications of MPF activity, most likely through interaction with regulatory proteins. Of the panoply of Cdk1-interacting proteins, among the least well understood are the Cks homologs. In both fission (Schizosaccharomyces pombe) and budding (Saccharomyces cerevisiae) yeast, depletion of Cks homologs leads to mitotic arrest (4, 5). Immunodepletion of Xe-p9, a Xenopus Cks homolog, from egg extracts prevents both entry into and exit from mitosis, depending on the experimental design (6). However, no conclusive evidence linking these mitotic phenotypes to a specific molecular defect has been presented. An additional unresolved issue is the existence of two closely related orthologs in mammalian species and possibly other vertebrates (7). Cks1 and Cks2 have 81% identical amino acids and yet have been maintained evolutionarily, which implies distinct functions. The recent finding that Cks1 has a specific role in the proteolysis of the Cdk inhibitor, p27Kip1, not shared by Cks2, supports this idea (8, 9). To elucidate ortholog-specific functions and the functions, in general, of Cks proteins in mammals, we constructed targeted disruptions for both CKS1 and CKS2 in the mouse. The murine CKS2 locus consists of three exons spanning ∼5 kb (fig. S1). The targeting vector was designed to insert a neomycin resistance cassette into Exon 1 of the CKS2 gene to cause deletion of 92% of the CKS2 coding region.

CKS2–/– mice were histologically indistinguishable from wild-type littermates for all tissues examined with the exception of the testes, which were significantly atrophied. Neither male nor female CKS2–/– mice mated with wild-type animals produced progeny. In sections of testes from wild-type mice, all stages of spermatogenesis were observed (10) (Fig. 1A), and mature sperm were located in the epididymydes (fig. S2B). However, in sections of testes from CKS2–/– mice no staging of the seminiferous epithelium could be detected due to an arrest of primary spermatocytes of the first wave of spermatogenesis at metaphase I (Fig. 1B). Very few, or no, anaphase figures were ever observed in the testes of juvenile or adult CKS2–/– mice (Fig. 1B, fig. S2A). In some tubules, aberrant multinucleated spermatids were detected (fig. S2D, arrowheads). Consistent with these observations, sections of CKS2–/– epididymydes did not contain sperm (10) (fig. S2C). Levels of several cell-cycle proteins thought to be expressed before and during meiosis were analyzed by immunoblotting extracts prepared from testis (10) (Fig. 1C). The amounts of cyclin A and Cdk2 were indistinguishable when wild-type and CKS2–/– testis were compared. However, cyclin B was significantly elevated in the mutant testis, consistent with a metaphase block. Taken together, the lack of meiotic anaphase figures and the elevated level of cyclin B strongly suggest that the CKS2–/– phenotype is due to arrest at metaphase I of meiosis. In addition, the most slowly migrating, presumably most hyperphosphorylated, species of the Cdc27 subunit of the mitotic protein–ubiquitin ligase, anaphase-promoting complex/cyclosome (APC/C), was absent in CKS2–/– testes. It is noteworthy that, in Xenopus eggs, Cdc27 has been shown to be a cyclin B/Cdk1 substrate that depends on the Cks2 homolog, Xe-p9, for efficient phosphorylation (11). Testes from adult CKS2–/– mice exhibited high levels of apoptosis (fig. S2, E and G) by comparison with wild-type animals (fig. S2, E and F), which suggests that spermatocytes that could not mature eventually lost viability and were cleared by apoptotic pathways. In situ hybridization with sections of wild-type testes (10) indicated that Cks1 is not expressed in seminiferous tubules of the mature testis, whereas Cks2 is abundantly expressed particularly in spermatocytes undergoing meiosis (Fig. 1, D and E). Therefore the sterility of CKS2–/– males appears to be due to spermatocyte arrest at metaphase I of meiosis. Apparently, the lack of expression of Cks1 in the male germ line prevents possible compensatory rescue of the Cks2 loss-offunction phenotype.

Fig. 1.

Metaphase I arrest in spermatocytes of CKS2–/– mice. (A) Testis sections from adult wild-type mice were stainedwith hematoxylin andeosin (10) and germ cell types contained within the different stages of the seminiferous epithelium were identified histologically. A tubule in stage XII of the seminiferous epithelium is shown. Spermatocytes in metaphase (redarrowheads) andin anaphase/telophase (green arrowheads) can be seen. (B) Testis section from an adult CKS2–/– mouse showing representative tubules arrestedat the spermatocyte stage. Spermatocytes are arrested in metaphase (redarrowheads). Original magnification, ×400. (C) Immunoblot analysis showing elevatedlevels of cyclin B and hypophosphorylated Cdc27 in CKS2–/– testes. Equivalent amounts of total protein were loaded per lane. (D and E) In situ hybridization of wild-type adult mouse testes with probes for Cks2 (D) or Cks1 (E). Original magnification, ×40.

At all ages, CKS2–/– ovaries were indistinguishable histologically from comparable wild-type ovaries (10) (fig. S3). Although mature, ovulated wild-type oocytes were arrested normally at metaphase II of meiosis (Fig. 2, A and C; Fig. 3A, first row), CKS2–/– oocytes were arrested at metaphase I. Chromosomes were aligned at metaphase I on a characteristic barrel-shaped metaphase I spindle (Fig. 2, B and D), and the oocytes had not extruded a polar body (Fig. 3A, second row). A small percentage of oocytes arrested at metaphase I with a phenotype of chromosome congression failure (Fig. 2B, inset). Therefore, the observed female sterility in CKS2–/– mice appears to be caused by oocyte arrest at metaphase I. Analysis of ovarian sections by in situ hybridization indicated that Cks1 and Cks2 mRNA were clearly expressed in follicles at all stages of development (Fig. 2, E and F). By analysis of adjacent sections it could be seen that Cks2 mRNA was strongly expressed in the developing oocytes (Fig. 2F), whereas Cks1 mRNA appeared to be primarily expressed in the surrounding granulosa cells (Fig. 2E), although a low level of expression could not be excluded in the oocytes themselves. Thus, Cks1 mRNA appears to be excluded or not expressed in the oocyte at sufficient levels to perform essential meiotic functions.

Fig. 2.

Metaphase I arrest in CKS2–/– oocytes. (A) DAPI staining of DNA in an ovulatedoocyte isolatedfrom a wild-type mouse. Chromosomes are alignedat the metaphase II plate. (B) DAPI staining of DNA in an ovulatedoocyte isolatedfrom a CKS2–/– mouse. Chromosomes are alignedat the metaphase I plate. Inset: An example of an ovulatedoocyte showing normal chromosome condensation and chiasmata formation but with a phenotype of congression failure in which chromosomes fail to align correctly at the metaphase plate. (C) β-Tubulin immunostaining of the metaphase II spindle in the same oocyte as in (A). (D) β-Tubulin immunostaining of the metaphase I spindle in the same oocyte as in (B). Note the typical barrel-shaped metaphase I spindle is distinct from the metaphase II spindle seen in wild-type oocytes. Original magnification, ×600. (E and F) In situ hybridization of wild-type adult mouse ovaries with probes for Cks1 (E) or Cks2 (F). Arrows indicate position of oocyte-containing follicles. A magnification of an individual oocyte is shown in the inset (designatedby *). Original magnification, ×40.

Fig. 3.

Cks1 and Cks2 can rescue the metaphase I arrest of CKS2–/– oocytes. (A) Left column, Nomarski optics. Right column, DAPI staining of DNA (green) and β-tubulin antibody (red). Extruded polar bodies can be observedby Nomarski optics and DAPI staining (green). Example of a wild-type mouse oocyte recoveredfrom the ovary at GV stage (see below) andinducedto progress to metaphase II in vitro (first row). CKS2–/– oocyte recoveredfrom the ovary at GV stage andinduced to mature in vitro (secondrow). Maturation is not complete andthe oocyte is arrestedat metaphase I (note the lack of polar body extrusion). Rescue of CKS2–/– metaphase I–arrestedoocytes by microinjection of either Cks2 (thirdrow) or Cks1 (fourth row) mRNA, but not by mutant Cks2 mRNA (fifth row). (B) Quantification of microinjection experiments. At least 70 oocytes were injected with each mRNA. The results of one representative experiment are shown. GV, oocytes remained arrested at the germinal vesicle stage. GVBD, oocytes undergo germinal vesicle breakdown but arrest at metaphase I. PBE, polar body extrusion representing progression to second meiotic division.

Metaphase I arrest could occur either because of an inability to enter anaphase or because the events of meiotic prophase, including synaptonemal complex formation and chromosome recombination, occur abnormally in CKS2–/– spermatocytes and oocytes and so induce a checkpoint-mediated metaphase I arrest. To test this, we analyzed meiotic prophase in both spermatocytes and oocytes immunolabeled with antibody against Cor1 (Scp3) to examine synaptonemal complex formation, Rad51-specific antibody to examine DNA double-strand breaks and chromosome recombination, and CREST serum to examine kinetochore formation and fusion (10). In addition, we analyzed chromosome spreads of diplotene/metaphase I spermatocytes stained with Giemsa and oocytes stained with DAPI (4′,6′-diamidino-2-phenylindole) for chromosome condensation and chiasmata formation (10). These analyses confirmed that all events of early meiotic progression and synapsis were normal in CKS2–/– spermatocytes (Fig. 4, A to D, I, and J) and oocytes (Fig. 2B, inset; Fig. 4, E to H). Thus, metaphase I arrest in CKS2–/– oocytes and spermatocytes is most likely caused by an inability to enter anaphase rather than a checkpoint-mediated metaphase I arrest.

Fig. 4.

Normal prophase I in CKS2–/– spermatocytes andoocytes. Early pachytene chromosome preparations were stainedfor axial elements of the synaptonemal complex present along each homolog with anti-Cor1 (Scp3) (FITC, green), kinetochore proteins with CREST antiserum (TRITC, red), and double-strand break foci with anti-Rad51 (AMCA, white). Shown are pairedimages of pachytene-stage spermatocytes from adult wild-type (A and C) or CKS2–/– (B and D) mouse testis or pachytene-stage oocytes from 18.5 days postcoitum wild-type (E and G) or CKS2–/– (F and H) embryos. Continuous Cor1 staining demonstrates that, at pachytene, synaptonemal complexes are fully formedin wild-type and CKS2–/– germ cells. The low numbers of Rad51 foci along the chromosomes indicate that strandexchange andrecombination is occurring normally in both wild-type and CKS2–/– spermatocytes andoocytes. Kinetochore fusion is indicated by the appearance of 20 CREST foci in red. Air-dried, Giemsa-stained preparations of diplotene/metaphase I chromosomes show proper chromosome condensation and chiasmata formation in wild-type (I) and CKS2–/– spermatocytes (J). Original magnification, ×1000.

The observed failure of germ cell development in CKS2–/– mice most likely results from a lack of Cks1 expression that is necessary to compensate for Cks2 loss of function in the germ line rather than from a germline–specific function of Cks2. To test this hypothesis, prophase-arrested, germinal vesicle (GV) stage CKS2–/– oocytes were microinjected with either Cks1 or Cks2 mRNA and then allowed to undergo GV breakdown (10). Both Cks1 and Cks2 mRNA could rescue the metaphase I arrest of CKS2–/– oocytes (Fig. 3, A and B), consistent with the hypothesis that germ line exclusion of Cks1 expression renders Cks2 essential. However, a mutant form of Cks2, defective in Cdk interaction, could not overcome the M1 arrest (Fig. 3, A and B), which indicates that the required function of Cks2 in oocytes involves interactions with a Cdk.

The CKS2–/– phenotype described here is distinct from most of the previous mouse mutants described to date. The majority of mouse mutations with meiotic phenotypes cause arrest in prophase (1215), mostly associated with defects in meiotic recombination or synaptonemal complex formation. Three notable exceptions are the MLH1 and MLH3 knockout mice and the LT/Sv strain of mice, which produce a high percentage of metaphase I–arrested oocytes (1619). In LT/Sv mice, the block is not complete, and many metaphase II–arrested oocytes are formed, which appear normal and can be fertilized. In mice lacking the DNA mismatch repair protein Mlh1, males and females are sterile and all spermatocytes and the majority of oocytes arrest at metaphase I with achiasmatic chromosomes (17, 18). Spermatocytes from mice lacking the related Mlh3 mismatch repair protein fail to maintain crossovers and arrest at metaphase I (19). Our data demonstrating that CKS2–/– mice also arrest at metaphase I suggest that Cks2 may act as a critical nexus between the proteins required for meiotic chromosome recombination and the proteins required for homologous chromosome segregation at meiotic anaphase.

Supporting Online Material

Materials and Methods

Figs. S1 to S3

References and Notes

  • * These authors contributedequally to this work.

References and Notes

View Abstract

Navigate This Article