In vertebrates, M phase‐promoting factor (MPF), a universal G2/M regulator in eukaryotic cells, drives meiotic maturation of oocytes, while cytostatic factor (CSF) arrests mature oocytes at metaphase II until fertilization. Cdk2 kinase, a G1/S regulator in higher eukaryotic cells, is activated during meiotic maturation of Xenopus oocytes and, like Mos (an essential component of CSF), is proposed to be involved in metaphase II arrest in mature oocytes. In addition, cdk2 kinase has been shown recently to be essential for MPF activation in Xenopus embryonic mitosis. Here we report injection of Xenopus oocytes with the cdk2 kinase inhibitor p21Cip in order to (re)evaluate the role of cdk2 kinase in oocyte meiosis. Immature oocytes injected with p21Cip can enter both meiosis I and meiosis II normally, as evidenced by the typical fluctuations in MPF activity. Moreover, mature oocytes injected with p21Cip are retained normally in metaphase II for a prolonged period, whereas those injected with neutralizing anti‐Mos antibody are released readily from metaphase II arrest. These results argue strongly against a role for cdk2 kinase in MPF activation and its proposed role in metaphase II arrest, in Xenopus oocyte meiosis. We discuss the possibility that cdk2 kinase stored in oocytes may function, as a maternal protein, solely for early embryonic cell cycles.
In almost all animal species, immature oocytes are arrested naturally in the first meiotic prophase (prophase I) or the late G2 phase (Masui and Clarke, 1979; Murray and Hunt, 1993). This meiotic arrest is released usually by hormonal stimulation and is followed by activation of maturation‐ or M phase‐promoting factor (MPF) (Masui and Markert, 1971). MPF, which is a universal G2/M regulator in eukaryotic cells and consists of a complex of cdc2 kinase and cyclin B (Nurse, 1990), drives the oocyte to enter not only meiosis I but also meiosis II, with a transient decrease in its activity between the two meiotic divisions (Gerhart et al., 1984; Furuno et al., 1994).
In many vertebrate species, mature oocytes (or unfertilized eggs) are arrested again at the second meiotic metaphase or metaphase II until fertilization (Sagata, 1996). In progesterone‐induced Xenopus oocyte maturation, metaphase II arrest of mature oocytes is caused by an endogenous cytoplasmic factor called cytostatic factor (CSF) (Masui and Clarke, 1979). The c‐mos proto‐oncogene product Mos is shown to be an essential component of CSF (Sagata et al., 1989), and its CSF function most probably is mediated by the mitogen‐activated protein kinase (MAPK) (Haccard et al., 1993; Kosako et al., 1994). The Mos–MAPK pathway is thought to stabilize MPF, thereby enabling mature oocytes to arrest at metaphase II (reviewed in Sagata, 1996). However, some other new protein(s), besides Mos, has also been suggested to be required for metaphase II arrest in Xenopus oocytes (Yew et al., 1992; Furuno et al., 1994).
Cdk2 kinase, a cognate kinase of cdc2 kinase, functions in the G1/S transitions or for DNA replication in higher eukaryotic cells (Fang and Newport, 1991; reviewed in Sherr, 1993; Hunter and Pines, 1994). Interestingly, this kinase (and its associated partner cyclin E) has been shown to be synthesized de novo and accumulated several‐fold during Xenopus oocyte maturation (Gabrielli et al., 1992; Rempel et al., 1995), a process that lacks both G1 and S phases (Sagata, 1996). Moreover, inhibition of this synthesis during maturation by cdk2 antisense oligonucleotides has been shown to prevent metaphase II arrest in mature oocytes (Gabrielli et al., 1993). These results have led to the important proposal that cdk2 kinase (and cyclin E) must be the new protein that, besides Mos, is required for metaphase II arrest in mature Xenopus oocytes (Gabrielli et al., 1993; Rempel et al., 1995). For such an important proposal to be validated, however, yet another method to inhibit cdk2 kinase specifically may be required, since oligonucleotide injection often gives non‐specific effects in Xenopus oocytes (cf. Smith et al., 1990).
In addition to the proposed meiotic role and the well‐established role for DNA replication in the somatic cell cycle, cdk2 kinase has also recently been shown to be essential for MPF activation in Xenopus embryonic mitosis: specific inhibition or depletion of cdk2 kinase in cell‐free egg extracts prevented MPF activation, while readdition of it to the extracts allowed MPF activation (Guadagno and Newport, 1996). This direct linkage of cdk2 kinase (a G1/S regulator) to positive regulation of MPF (a G2/M regulator) may well be universal in mitosis in higher eukaryotic cells (see discussions in Guadagno and Newport, 1996, and in this paper). If so, it will be very important to know whether cdk2 kinase is also essential for MPF activation in meiotic maturation (or the meiotic cell cycle) in Xenopus oocytes.
In the present study, we took advantage of p21Cip, the cyclin‐dependent kinase inhibitor, to (re)examine any role for cdk2 kinase in the meiotic maturation of Xenopus oocytes. Surprisingly, our results argue strongly against a role for cdk2 kinase in MPF activation and its proposed role in metaphase II arrest, in Xenopus oocyte meiosis. Based on these and other results, we discuss the role of cdk2 kinase in oocyte meiosis and embryonic mitosis.
Effect of cdk2 kinase inhibition on MPF activation and initiation of maturation
To inhibit the total cdk2 kinase activity in oocytes, we utilized (human) p21Cip protein, an inhibitor of several cyclin‐dependent kinases such as cdk2 and cdc2 kinases (Elledge and Harper, 1994); at appropriate concentrations, this protein can specifically inhibit cdk2 kinase activity both in vivo and in vitro (Harper et al., 1995; Guadagno and Newport, 1996). We prepared glutathione S‐transferase (GST)–wild‐type p21Cip fusion protein (WTCip) and, as a control, GST–(N‐terminal half‐truncated) p21Cip fusion protein (ΔNCip); the latter is known to be unable to bind to and inhibit cdk2 kinase (Chen et al., 1995; Harper et al., 1995). First, we examined whether cdk2 kinase activity is required for MPF activation and initiation of maturation in progesterone‐treated oocytes. We injected prophase I (or G2)‐arrested immature oocytes with either ΔNCip or WTCip at an internal concentration of 100 nM, a concentration sufficient to inhibit cdk2 kinase activity completely (Guadagno and Newport, 1996; see also below). Upon progesterone treatment, uninjected control oocytes and those injected with either ΔNCip or WTCip (or GST alone; data not shown) all underwent germinal vesicle breakdown (GVBD; a hallmark of initiation of maturation), essentially with the same kinetics (Figure 1); they all also underwent MPF activation essentially normally, or shortly before GVBD (data not shown, but see the high MPF activities at GVBD in Figure 2B). Thus, cdk2 kinase activity was not required for either MPF activation or initiation of maturation in Xenopus oocytes. These results are surprising if we consider the essential role of cdk2 kinase in MPF activation in Xenopus embryonic mitosis (Guadagno and Newport, 1996) (but might not necessarily be surprising if we note its considerably low kinase activities around the time of MPF activation and GVBD during maturation; Gabrielli et al., 1992; Rempel et al., 1995; see also below).
Effect of cdk2 kinase inhibition on MPF reactivation and entry into meiosis II
We next examined whether cdk2 kinase activity is necessary for reactivation of MPF and entry into meiosis II during maturation, which normally occur 1–2 h after GVBD and approximately at the time of the large cdk2 kinase activation during maturation (Furuno et al., 1994; Rempel et al., 1995). To do this, we injected oocytes with either ΔNCip or WTCip, treated them with progesterone and harvested them every 1 h after GVBD; we then subjected their extracts to either total histone H1 kinase assays (for measuring MPF activity; Furuno et al., 1994) or H1 kinase assays of cdk2 kinase (after immunoprecipitation with anti‐cdk2 kinase antibody; for the specificity and efficiency of this immunoprecipitation, see Materials and methods). In ΔNCip‐injected oocytes (and uninjected control oocytes; not shown), cdk2 kinase activity increased progressively after 1 h of GVBD, attaining a 50‐fold activity at 6 h after GVBD (Figure 2A), essentially as reported previously (Gabrielli et al., 1992; Rempel et al., 1995). By contrast, in WTCip‐injected oocytes, cdk2 kinase activity remained extremely low throughout 6 h after GVBD; we could estimate that, in these oocytes, >98% of the cdk2 kinase activity was inhibited by WTCip (at 4–6 h after GVBD). In both types of oocytes (and the uninjected control oocytes; not shown), however, MPF activity showed essentially the same kinetics: it decreased transiently at 1 h, re‐increased greatly at 2–3 h (representing entry into meiosis II; Furuno et al., 1994) and then further increased gradually until 6 h after GVBD (Figure 2B). Thus, very interestingly, cdk2 kinase activity was also not required for MPF reactivation and entry into meiosis II in maturing oocytes, despite the fact that it was greatly activated during these periods. Taken together, the present results strongly indicate that, unlike that in embryonic mitosis (Guadagno and Newport, 1996), MPF activation in Xenopus oocyte meiosis does not depend on cdk2 kinase activity.
Effect of cdk2 kinase inhibition on metaphase II arrest
The high MPF activities, even at 4–6 h after GVBD, in WTCip‐injected oocytes (Figure 2B) suggest that cdk2 kinase activity may not be essential for metaphase II arrest in mature oocytes (usually, metaphase II arrest occurs from ∼3 h after GVBD; Furuno et al., 1994). This notion, if correct, contrasts with the previous proposal that cdk2 kinase is essential and functions (together with Mos) for metaphase II arrest in mature oocytes (Gabrielli et al., 1993; Rempel et al., 1995). [We initially attempted to confirm this proposal, by injecting oocytes with cdk2 antisense oligonucleotides as reported (Gabrielli et al., 1993), but we could not achieve the same results. Although the reported cdk2 antisense oligonucleotides (Gabrielli et al., 1993) and our newly designed antisense oligonucleotides (nucleotide positions 6–25, from the start codon; cf. Paris et al., 1991) both inhibited the synthesis of cdk2 kinase almost completely and (at most) 60% of its kinase activity (due to the presence of pre‐existing cdk2 kinase) during maturation, none of them could prevent mature oocytes from arresting at metaphase II, as judged both biochemically and cytologically (our unpublished data).] Therefore, this time, we injected mature oocytes (5 h after GVBD and arrested in metaphase II) with either WTCip or ΔNCip (under conditions where artificial activation of the mature oocytes does not occur by the action of injection; see Materials and methods). Cdk2 kinase activity in WTCip‐injected mature oocytes decreased drastically (to 6%) 1.5 h after the injection and, thereafter, to negligible levels (<2%) until 6 h after the injection, while that in ΔNCip‐injected oocytes remained high throughout 6 h after the injection (Figure 3A). By contrast, not only in ΔNCip‐injected but also in WTCip‐injected oocytes, MPF activity remained high (without any significant changes) throughout 6 h after the injection (Figure 3B), strongly suggesting the persistence of metaphase II arrest in the WTCip‐injected oocytes. To confirm this more directly, we cytologically examined mature oocytes 6 h after injection with either WTCip or ΔNCip. Both types of oocytes (and the oocyte before injection; not shown) showed a well‐formed bipolar spindle with highly condensed chromosomes, just under the animal cortical layer (Figure 3C), which is typical of metaphase II arrest in mature oocytes (Gard, 1992; Furuno et al., 1994). Thus, WTCip could not release mature oocytes from metaphase II arrest, despite its almost complete inhibition of cdk2 kinase activity. These results argue strongly that, in contrast to the previous proposal (Gabrielli et al., 1993; Rempel et al., 1995), cdk2 kinase is not required for metaphase II arrest in mature oocytes.
Release from metaphase II arrest in Mos‐ablated mature oocytes is independent of cdk2 kinase activity
Finally, we examined whether release from metaphase II arrest occurs in mature oocytes depleted of Mos, an essential component of CSF (Sagata et al., 1989), and whether this occurs independently of cdk2 kinase activity. To do this, we injected mature oocytes (5 h after GVBD) with either neutralizing anti‐Mos antibody (Furuno et al., 1994) or control rabbit IgG, and harvested them 1.5 and 3 h later to measure MPF and cdk2 kinase activities. As was expected (Sagata et al., 1989; Furuno et al., 1994), in anti‐Mos antibody‐injected oocytes, MPF activity dropped progressively to 35 and 10% at 1.5 and 3 h, respectively, after the injection, while in control IgG‐injected oocytes it remained high and constant (Figure 4B). When examined cytologically, the anti‐Mos antibody‐injected oocyte (3 h after the injection) exhibited a lobed nucleus somewhat deep in the cytoplasm (which is typical of an activated interphase egg; Hausen and Riebesell, 1991), while the control IgG‐injected oocyte retained a metaphase II spindle (Figure 4C); thus, Mos ablation released the mature oocyte from metaphase II arrest. In contrast to these results, it was found that cdk2 kinase activity remained high throughout 3 h after the injection in both Mos‐ablated and control oocytes (Figure 4A). Thus, very interestingly, metaphase II arrest was released independently of cdk2 kinase activity in Mos‐ablated oocytes. These results, together with those with WTCip injection (Figure 3), not only confirm the previously described role for Mos (Sagata et al., 1989; Furuno et al., 1994), but also argue against the proposed role for cdk2 kinase (Gabrielli et al., 1993; Rempel et al., 1995) in metaphase II arrest in mature Xenopus oocytes.
The present results, utilizing p21Cip, provide unexpected findings that cdk2 kinase activity is required for neither MPF activation (and reactivation) nor metaphase II arrest in the meiotic maturation of Xenopus oocytes. p21Cip, under the present conditions, inhibited >98% of the cdk2 kinase activity (but virtually no activity of cdc2 kinase, a catalytic component of MPF; see Figures 2A and 3B). p21Cip could also have inhibited other endogenous proteins, such as cdk4 kinase and PCNA (Elledge and Harper, 1994; Chen et al., 1995), that might be present in maturing oocytes. However, the apparently (both biochemically and cytologically) normal meiotic cell cycle in p21Cip‐injected oocytes strongly suggests that cdk2 kinase may not play any major role in this type of cell cycle.
Our results on the role of cdk2 kinase in metaphase II arrest in mature Xenopus oocytes clearly contradict the previous results and proposal (Gabrielli et al., 1993; Rempel et al., 1995). This contradiction, however, is unlikely to be due to the use of different cdk2 kinase inhibitors (p21Cip versus antisense oligonucleotides) (see Results). Currently, we do not know the reason(s) for the contradiction. However, the fact that metaphase II arrest persists in cdk2 kinase‐inhibited oocytes (Figure 3) while cdk2 kinase activity persists in Mos‐inhibited (and hence metaphase II‐released) oocytes (Figure 4) argues strongly that, unlike Mos (Sagata et al., 1989), cdk2 kinase is dispensable for metaphase II arrest in mature Xenopus oocytes. [The persistence of cdk2 kinase activity in Mos‐inhibited (and hence MPF‐inactivated) oocytes is interesting, however, since this indicates that cdk2 kinase activation is independent of MPF activity, as previously shown in cell‐free Xenopus egg extracts (Gabrielli et al., 1992).] In mice, Mos is also involved in metaphase II arrest in mature oocytes (O'Keefe et al., 1989; Colledge et al., 1994; Hashimoto et al., 1994); so far, however, neither the activity of cdk2 kinase in mouse oocyte maturation nor its involvement in metaphase II arrest has been known.
The apparent lack of a requirement for cdk2 kinase activity in MPF activation (Figure 1) and its reactivation (Figure 2) in Xenopus oocytes is very intriguing, because cdk2 kinase recently has been linked directly to positive regulation of MPF in (cell‐free) Xenopus embryonic mitosis (Guadagno and Newport, 1996) (we recently have confirmed this linkage by using intact embryos; our unpublished data). Xenopus embryonic mitosis is unique in its rapidity and independence of the preceding DNA replication (Newport and Dasso, 1989; Murray and Hunt, 1993). However, the regulation mechanism of mitosis is generally well conserved during evolution (Maller, 1990; Nurse, 1990; Nigg, 1995), and cdk2 kinase (associated A‐type cyclin) has been shown to have high activities in the G2/M transitions (as well as in the S phases) in mammalian somatic cells (Pagano et al., 1992, 1993; Rosenblatt et al., 1992; Tsai et al., 1993). Thus, the linkage of cdk2 kinase to MPF activation (found in Xenopus embryonic mitosis) may be universal in mitosis in higher eukaryotic cells. If so, our results suggest that the mechanism of MPF activation in oocyte meiosis may differ largely, at least in part, from that in mitosis. Indeed, MPF activation (and reactivation) in Xenopus oocytes, unlike that in mitosis, is known to depend, at least in part, on the meiosis‐specific protein Mos and its downstream kinase MAPK (reviewed in Sagata, 1997). In oocyte meiosis (in many species), therefore, some specific protein(s) might compensate for the cdk2 kinase function to regulate MPF activity positively (Figure 5).
If cdk2 kinase plays no major role in the meiotic cell cycle in (maturing) oocytes, what would its role be? In higher eukaryotic cells, cdk2 kinase plays a key role in DNA replication (Sherr, 1993; Hunter and Pines, 1994). In Xenopus, as in many other species, DNA replication does not occur during oocyte maturation (or the meiosis I–meiosis II transition), but occurs immediately after fertilization (Sagata, 1996). Cdk2 kinase stored in Xenopus oocytes is, in fact, very stable and persists long after fertilization (Hartley et al., 1996; Howe and Newport, 1996; our unpublished data) and, in each early embryonic cell cycle, cdk2 kinase has high activities in both S and M phases (Hartley et al., 1996). Thus, almost certainly, most of the cdk2 kinase stored in oocytes will contribute directly to DNA replication (Fang and Newport, 1991) (and MPF activation; Guadagno and Newport, 1996) in the early embryonic cell cycles (Figure 5). We cannot yet rigorously exclude a subtle role(s) for cdk2 kinase in oocyte maturation per se. However, as discussed here, its major role in oocytes may be best explained by the developmental context, or by preparation for early embryonic cell cycles.
Materials and methods
GST–p21Cip fusion proteins
A cDNA encoding human p21Cip (Harper et al., 1995) was subjected to PCR to yield a cDNA fragment encoding either the whole wild‐type p21Cip (WTCip) or N‐terminal half (amino acid positions 1–86)‐truncated p21Cip (ΔNCip); the 5′ primer used (containing an artificial BamHI site) was either 5′‐GCCGGATCCCCATGTCAGAACCGGCTGGGG‐3′ (for WTCip) or 5′‐GCCGGATCCCCATGGATGAGTTGGGAGGAGGCAGGC‐3′ (for ΔNCip), while the 3′ primer (also containing a BamHI site) was 5′‐GCCGGATCCTCTAGAGCTTCCTCTTGGAGAAGATCAGC‐3′ for both WTCip and ΔNCip. After digestion with BamHI, the PCR products were subcloned into a BamHI‐digested pGEX‐3X plasmid vector (encoding GST), transfected into Escherichia coli, and GST–WTCip and GST–ΔNCip fusion proteins were purified by glutathione–Sepharose chromatography (Pharmacia), as described previously (Furuno et al., 1994). The purified proteins were concentrated by Centricon‐30 (Amicon), buffer‐exchanged with 10 mM HEPES (pH 7.5)–50 mM NaCl, and the concentration was determined by Coomassie blue staining using bovine serum albumin (BSA) as a standard.
Polyclonal antibody was raised in rabbits against the synthetic peptide THPFFRDVSRPTPHLI, corresponding to the last 16 amino acids of Xenopus cdk2 kinase (formerly Eg1; Paris et al., 1991) and coupled to keyhole limpet hemocyanin, by standard methods. Upon immunoprecipitation and Western blotting analyses of Xenopus oocyte extracts, this antibody recognized a single band of 33 kDa or cdk2 kinase very specifically (or only in the absence of competing peptide antigens); moreover, the antibody was able to immunoprecipitate as much as 80–90% of the total cdk2 kinase present in oocytes that were either uninjected or injected with p21Cip, as revealed by Western blotting analysis of the immunoprecipitates with either anti‐PSTAIR antibody (Yamashita et al., 1991) or anti‐cdk2 kinase antibody (our unpublished data). Preparation of neutralizing anti‐Mos antibody was described previously (Furuno et al., 1994).
H1 kinase assays
Five oocytes were homogenized in EB buffer (10 μl/oocyte), centrifuged briefly, and 5 μl of the clear supernatant (45 μl in total) was subjected to histone H1 kinase assays to measure the total MPF activity, as described previously (Furuno et al., 1994). Ten μl of the remaining supernatant was diluted with 300 μl of RIPA buffer and then subjected to immunoprecipitation with anti‐cdk2 kinase antibody (in the presence or absence of competing peptide antigens), as described (Nishizawa et al., 1993). The immunoprecipitates were then assayed for H1 kinase activity by the addition of 25 μl of a cdk2 kinase‐assaying buffer (Gabrielli et al., 1993) supplemented with 200 μg/ml BSA, 200 μM ATP, 500 μg/ml histone H1 and 5 μCi of [γ‐32P]ATP, at 25°C for 30 min. Phosphorylated histone H1 was separated by SDS–PAGE, dried and fixed as described (Furuno et al., 1994), and then quantified for 32P incorporation by BAS‐1000 Image Analyzer (Fuji).
Microinjection and cytological examination
Preparation and culture of oocytes, induction of oocyte maturation by progesterone and cytological examination of fixed oocytes were all described previously (Sagata et al., 1989; Furuno et al., 1994). Forty nl of 2 μM test protein (WTCip or ΔNCip) was microinjected into oocytes (giving rise to an internal concentration of ∼100 nM of the protein), as described (Sagata et al., 1989). For microinjection of mature oocytes, test protein was co‐injected with 40 nl of 10 mM EGTA, to prevent artificial activation of the mature oocyte by the action of injection. Under these conditions, co‐injection of neutralizing anti‐Mos antibody (200 ng/oocyte) efficiently released mature oocytes from metaphase II arrest (after 1–2 h of the injection), while that of control rabbit IgG did not (see Results); these results not only show the validity of the present method but also confirm the previously described role for Mos in metaphase II arrest (see text).
We thank J.Wade Harper for the kind gift of human p21Cip cDNA, Yoshimasa Hayano for help in preparation of GST–p21Cip fusion proteins, and Miyuki Egashira for editing the manuscript. This work was supported by Grants‐in‐Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan.
- Copyright © 1997 European Molecular Biology Organization