Title: The Polo-like kinase Plx1 interacts with and inhibits Myt1 after fertilization of Xenopus eggs
Abstract: Article3 February 2005free access The Polo-like kinase Plx1 interacts with and inhibits Myt1 after fertilization of Xenopus eggs Daigo Inoue Daigo Inoue Department of Biology, Graduate School of Sciences, Kyushu University, Hakozaki, Fukuoka, Japan Search for more papers by this author Noriyuki Sagata Corresponding Author Noriyuki Sagata Department of Biology, Graduate School of Sciences, Kyushu University, Hakozaki, Fukuoka, Japan CREST, Japan Science and Technology Agency, Kawaguchi, Saitama, Japan Search for more papers by this author Daigo Inoue Daigo Inoue Department of Biology, Graduate School of Sciences, Kyushu University, Hakozaki, Fukuoka, Japan Search for more papers by this author Noriyuki Sagata Corresponding Author Noriyuki Sagata Department of Biology, Graduate School of Sciences, Kyushu University, Hakozaki, Fukuoka, Japan CREST, Japan Science and Technology Agency, Kawaguchi, Saitama, Japan Search for more papers by this author Author Information Daigo Inoue1 and Noriyuki Sagata 1,2 1Department of Biology, Graduate School of Sciences, Kyushu University, Hakozaki, Fukuoka, Japan 2CREST, Japan Science and Technology Agency, Kawaguchi, Saitama, Japan *Corresponding author. Department of Biology, Graduate School of Sciences, Kyushu University, Hakozaki 6-10-1, Fukuoka 812-8581, Japan. Tel./Fax: +81 92 642 2617; E-mail: [email protected] The EMBO Journal (2005)24:1057-1067https://doi.org/10.1038/sj.emboj.7600567 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info During the meiotic cell cycle in Xenopus oocytes, p90rsk, the downstream kinase of the Mos–MAPK pathway, interacts with and inhibits the Cdc2 inhibitory kinase Myt1. However, p90rsk is inactivated after fertilization due to the degradation of Mos. Here we show that the Polo-like kinase Plx1, instead of p90rsk, interacts with and inhibits Myt1 after fertilization of Xenopus eggs. At the M phase of the embryonic cell cycle, Cdc2 phosphorylates Myt1 on Thr478 and thereby creates a docking site for Plx1. Plx1 can phosphorylate Myt1 and inhibit its kinase activity both in vitro and in vivo. The interaction between Myt1 and Plx1 is required, at least in part, for normal embryonic cell divisions. Finally, and interestingly, Myt1 is phosphorylated on Thr478 even during the meiotic cell cycle, but its interaction with Plx1 is largely inhibited by p90rsk-mediated phosphorylation. These results indicate a switchover in the Myt1 inhibition mechanism at fertilization of Xenopus eggs, and strongly suggest that Plx1 acts as a direct inhibitory kinase of Myt1 in the mitotic cell cycles in Xenopus. Introduction In most eukaryotic cells, the mitotic cell cycle consists of two alternating S and M phases with intervening G1 and G2 phases. The G2/M transition is a crucial point for progression through the cell cycle and is controlled by the Cdc2–cyclin B complex (reviewed by Nurse, 1990; Morgan, 1995). The Wee1 family of protein kinases phosphorylates and inhibits Cdc2 during interphase of the cell cycle, while the Cdc25 phosphatase dephosphorylates and activates it at entry into the M phase (reviewed by Coleman and Dunphy, 1994). The Wee1 family consists of Wee1 (present in all eukaryotes), Mik1 (in fission yeast), and Myt1 (in metazoans). While Wee1 and Mik1, nuclear kinases, phosphorylate Cdc2 exclusively on Tyr15 (Featherstone and Russell, 1991; Mueller et al, 1995a), Myt1, a membrane-associated kinase, phosphorylates it on both Thr14 and Tyr15 (Mueller et al, 1995b; Liu et al, 1997). In addition to being a negative regulator of Cdc2, Myt1 may also function to sequester Cdc2–cyclin B to the Golgi apparatus (Liu et al, 1999). Immature animal oocytes are generally arrested at the first meiotic prophase (prophase-I), a stage equivalent to the G2 phase of the mitotic cell cycle (reviewed by Sagata, 1996; Nebreda and Ferby, 2000). In several known species, Myt1, but not Wee1, is present in immature oocytes and functions to inhibit Cdc2, thereby causing prophase-I arrest (Nakajo et al, 2000; Okumura et al, 2002). Upon oocyte maturation induced by hormonal stimulation, however, Myt1 becomes phosphorylated and inhibited by Akt/PKB kinase in starfish (Okumura et al, 2002) and by p90rsk, the downstream kinase of the Mos–MAPK pathway (reviewed by Sagata, 1997; Ferrell, 1999), in Xenopus (Palmer et al, 1998). However, p90rsk is inactivated on fertilization of mature metaphase-II-arrested oocytes because of the degradation of Mos (Watanabe et al, 1991; Bhatt and Ferrell, 1999)—nevertheless, Myt1 activity can be downregulated by phosphorylation in fertilized eggs (Mueller et al, 1995b). Therefore, there must be another inhibitory kinase for Myt1 after fertilization or during embryonic mitosis in Xenopus. Myt1 is also hyperphosphorylated and inhibited at M phase of the somatic cell cycle in humans (Booher et al, 1997; Wells et al, 1999), but, even in this case, the responsible inhibitory kinase(s) is not known. Polo-like kinases (Plks) play key roles during multiple steps of mitosis, including prophase, metaphase, anaphase, and cytokinesis (reviewed by Barr et al, 2004), and recent studies demonstrate the essential role of their C-terminal Polo-box domain (PBD) in the phospho-dependent recognition of target proteins (Elia et al, 2003a, 2003b). In human cells, Plk1 is expressed primarily during late G2 and M phases (Golsteyn et al, 1995) and directly phosphorylates Cdc25C, perhaps to facilitate nuclear activation of Cdc2–cyclin B (Toyoshima-Morimoto et al, 2002). Moreover, Plx1, a Xenopus homolog of mammalian Plk1, can phosphorylate and activate Cdc25C in vitro (Kumagai and Dunphy, 1996) and is required for activation of Cdc2 during both oocyte maturation and embryonic mitosis (Abrieu et al, 1998; Qian et al, 1998; Karaiskou et al, 1999; Liu et al, 2004). As Myt1 is inactivated by phosphorylation at the M phase (Mueller et al, 1995b; Booher et al, 1997), it might also be regulated by Plk1/Plx1 at the M phase. In some support of this possibility, human Plk1 can phosphorylate Myt1 in vitro (Nakajima et al, 2003), and starfish Plk1 has recently been suggested to function somewhere upstream of Myt1 in eggs (Okano-Uchida et al, 2003). So far, however, there is no direct evidence that Plk1 interacts with and inhibits Myt1 directly. In this study, we have investigated the possible interaction between Myt1 and Plx1 during oocyte maturation and embryonic mitosis in Xenopus. We show strikingly that Myt1 forms a stable complex with Plx1 (but not with p90rsk) during embryonic mitosis, but not during oocyte maturation. Myt1 interacts with Plx1 via its Cdc2-phosphorylated Plx1-docking motif, and phosphorylation by Plx1 can inhibit Myt1 activity both in vivo and in vitro. Furthermore, although Myt1 is phosphorylated on the Plx1-docking motif during oocyte maturation, its interaction with Plx1 is inhibited by p90rsk-mediated phosphorylation. These results indicate that the mechanism of Myt1 inhibition switches over at fertilization of Xenopus eggs, and strongly suggest that Plk1/Plx1 is a direct inhibitory kinase of Myt1 during the mitotic cell cycle. In addition, our data implicate that recognition of target proteins by Plk1/Plx1 can be reversibly regulated by their phosphorylation status other than that at the Plk1-docking site. Results Physical interaction between Myt1 and Plx1 after egg activation (or fertilization) Myt1 undergoes a large electrophoretic mobility upshift (due to hyperphosphorylation) during progesterone (PG)-induced Xenopus oocyte maturation, while p90rsk and Plx1 undergo a small mobility shift (Figure 1A; see also Qian et al, 1998; Gross et al, 2000; Peter et al, 2002). p90rsk interacts with and phosphorylates Myt1 and thereby inhibits its kinase activity during oocyte maturation (Palmer et al, 1998). First we examined whether endogenous Plx1 could physically interact with Myt1 during oocyte maturation or after calcium ionophore-induced egg activation (which mimics fertilization; Watanabe et al, 1991). The kinase activity of Plx1 paralleled that of Cdc2–cyclin B, peaking at M phase both during oocyte maturation and after egg activation (Figure 1B; see also Qian et al, 1998). When first immunoprecipitated with anti-Myt1 antibody and then immunoblotted with anti-p90rsk antibody, Myt1 was shown to be bound to p90rsk in mature metaphase-II-arrested oocytes (MII oocytes, also called eggs), but not in activated eggs (Figure 1C). When the immunoprecipitates were blotted with anti-Plx1 antibody, however, Myt1 was found to be associated with Plx1 80 min after egg activation or at the first mitotic M phase (M1; see Figure 1B), but not appreciably in immature (IMO) or mature MII oocytes (Figure 1C). The interaction between Myt1 and Plx1 in M1 eggs, but not in MII oocytes, was also evident when it was analyzed by reciprocal immunoprecipitation (with anti-Plx1 antibody) and immunoblotting (IB; with anti-Myt1 antibody) (Figure 1D). Moreover, the Myt1–Plx1 interaction was also observed in normally fertilized eggs or embryos (see Figure 5). Thus, very interestingly, Plx1, instead of p90rsk, interacted with Myt1 at the mitotic M phase after egg activation or fertilization. Figure 1.Interaction of Plx1 and Myt1 at the mitotic but not meiotic M phase. (A) Immature oocytes (IMO, arrested at prophase-I) were treated with PG to induce maturation, and about 7 h later, fully mature oocytes (also called eggs) (MII, arrested at metaphase-II) were treated with calcium ionophore to induce activation (which mimics fertilization). The activated eggs were further incubated until the first mitotic M phase (M1; see (B)). (A similar in vitro maturation/activation system was employed below and in other experiments in this study.) Oocytes at the indicated stages and eggs at the indicated times (min after activation) were analyzed by IB for the indicated endogenous proteins. (B) At the indicated stages, kinase activities of endogenous Cdc2 and Plx1 were assayed in vitro by using histone H1 and α-casein as substrates, respectively. For details, see Materials and methods. (C) At the indicated stages, endogenous Myt1 was immunoprecipitated (IP) with anti-Myt1 antibody and then immunoblotted with either anti-Myt1, anti-p90rsk or anti-Plx1 antibodies. In control IP, anti-Myt1 antibody was pre-incubated with antigen peptides (+pep). Input, one oocyte or egg; IP, 10 oocytes or eggs. (D) IP and subsequent IB as in (C) were performed reciprocally by using anti-Plx1 and anti-Myt1 antibodies, respectively. Cont., immunoprecipitation with normal rabbit IgG. Download figure Download PowerPoint Figure 2.Inhibition of Myt1 activity by Plx1 in vitro. (A) Recombinant Myt1 protein was incubated with either wild-type (wt) or kinase-dead (K73M) Plx1 protein in the presence or absence of [γ-32P]ATP and analyzed by SDS–PAGE followed by autoradiography (top) or by IB (bottom). The arrowhead (on the top) denotes autophosphorylated Myt1. (B) Myt1 protein incubated with Plx1 as in (A) was analyzed by IB (top) or further incubated with KR Cdc2–cyclin B, which was then analyzed by IB for phospho-Thr14/Tyr15 (bottom). (C) Wild-type Myt1 and two consensus Plx1 phosphorylation site mutants of Myt1, S424A and 5A (S424A/S433A/S487A/T508A/T546A), were incubated with either wt or K73M Plx1, and their activities to phosphorylate kinase-dead Cdc2 were analyzed as in (B). For detailed methods in (A–C), see Materials and methods. Download figure Download PowerPoint Inhibition of Myt1 activity by Plx1 in vitro Human Plk1 can phosphorylate Myt1 in vitro (Nakajima et al, 2003), but it is not known whether this phosphorylation can regulate Myt1 activity. Given the (physical) interaction between Plx1 and Myt1 (Figure 1), however, Plx1 might naturally function to regulate Myt1 by phosphorylation. Recombinant wild-type Plx1 protein, but not its kinase-dead version (K73M), was able to phosphorylate Myt1 and cause a large mobility shift of it in vitro (Figure 2A). When first incubated with kinase-dead Plx1 and then with (kinase-dead) Cdc2–cyclin B, Myt1 was able to normally catalyze inhibitory phosphorylation of the Cdc2 protein on Thr14/Tyr15; notably, when preincubated with wild-type Plx1, however, Myt1 failed to phosphorylate Cdc2 (Figure 2B). Thus, these results demonstrate, for the first time, that Plk1/Plx1 can inhibit Myt1 activity in vitro. Figure 3.Cdc2-mediated phosphorylation of Myt1 on Thr478 at M phase. (A) Conservation of a Plk1-docking motif (SS/TP) in Myt1 kinases from various species. Potential phosphorylation sites are shown in red. In the following figures (B–E), maturing oocytes and activated eggs were derived from immature oocytes that had been injected with 1 ng of mRNA encoding Myt1-KD or its mutant (T478A/KD). (B) Either Myt1-KD protein or T478A/KD protein immunoprecipitated from five activated eggs (80 min after activation) was treated or not with λ-phosphatase (λ-PPase), and analyzed by IB with anti-Myt1 or anti-phospho-Thr478 antibodies. (C) Extracts from the indicated stages of oocytes or eggs (expressing Myt1-KD) were either immunoblotted as in (B) or assayed for Cdc2 activity. h, time after PG treatment; min, time after egg activation. MII, 7 h after PG treatment or the time of egg activation. (D, E) Immature oocytes (1 h before PG treatment) (D) or eggs 30 min after activation (E), both expressing Myt1-KD, were treated (+Ros) or not (−Ros) with roscovitine and, at the indicated times after PG treatment (D) or after egg activation (E), subjected to IB as in (B). (F) Recombinant Myt1-KD protein was incubated with either constitutively active (AF) or kinase-dead (KR) Cdc2–cyclin B complexes and then immunoblotted as in (B). For details, see Materials and methods. Download figure Download PowerPoint Human Myt1 can be phosphorylated at Ser426 by Plk1 in vitro (Nakajima et al, 2003). We also independently identified Ser424 (equivalent to Ser426 of human Myt1) as a major Plx1 phosphorylation site of Xenopus Myt1 (data not shown, but see Figure 4A). However, a Ser424 → Ala mutant (S424A) of Myt1, like wild-type Myt1, lost its kinase activity towards (kinase-dead) Cdc2–cyclin B after incubation with wild-type (but not kinase-dead) Plx1 (Figure 2C). Somewhat surprisingly, even a Myt1 mutant (5A)—in which four other serine or threonine residues lying in the consensus Plk1 phosphorylation motif (Asp/Glu-X-Ser/Thr; Nakajima et al, 2003) were also mutated to Ala—lost its activity towards Cdc2 after incubation with Plx1 (Figure 2C). Thus, although Myt1 could be phosphorylated and inhibited by Plx1, the inhibitory phosphorylation site(s) was not likely to reside in the proposed Plk1 phosphorylation motif, suggesting the occurrence of yet another Plx1 phosphorylation site(s) in Myt1. Consistent with this idea, the 5A (as well as S424A) mutant still underwent phosphorylation after treatment with Plx1 (see Supplementary Figure 1). Figure 4.Identification of phosphorylated Thr478 as the Plx1-docking site of Myt1. (A) Extracts from the indicated stages of oocytes or eggs expressing either Myt1-KD or the T478A/KD mutant as in Figure 3 were immunoblotted with anti-Myt1 or anti-phospho-Ser424 antibodies. Note that Ser424 phosphorylation of the T478A/KD mutant is undetectable at 80 min after egg activation. MI, maturing oocytes undergoing GVBD. (B) Extracts from MII oocytes or M1 eggs (expressing Myt1 constructs as in (A)) were first immunoprecipitated with either anti-Plx1 antibody (left) or anti-p90rsk antibody (right) and then immunoblotted with anti-Myt1 antibody (top). Input, one oocyte or egg; IP, 10 oocytes or eggs. (C) Either phospho-Thr478-containing peptides (pT478) or unphosphorylated peptides (T478), coupled to agarose beads, were incubated with either interphase (I, 30 min after egg activation) or M1 egg extracts, pulled down (PD), and immunoblotted with anti-Plx1 antibody. Input, one egg; PD, 10 eggs. For details, see Materials and methods. Download figure Download PowerPoint In vivo, efficient phosphorylation/inhibition of Myt1 by Plx1 seems to require a Cdc2-dependent docking of the two proteins (see below). In the above in vitro experiments (involving no active Cdc2), however, Plx1 probably phosphorylated (and inhibited) Myt1 directly without such a docking, similar to in vitro phosphorylation of Myt1 by human Plk1 (Nakajima et al, 2003). Phosphorylation of the potential Plx1-docking site Thr478 by Cdc2 Given the recent identification of a phospho-Plk1-docking motif (Ser-pSer/pThr-Pro) in a certain Plk1 target protein (Elia et al, 2003a), the interaction between Plx1 and Myt1 (observed after egg activation) might occur via a similar docking motif present in Myt1 protein. We noticed that Thr478 of Xenopus Myt1 lies in the Plk1-docking motif (STP) and is well conserved in Myt1 kinases from both vertebrate and invertebrate species (Figure 3A). Therefore, we first determined whether Thr478 of Myt1 would undergo any phosphorylation during oocyte maturation or after egg activation, by using an anti-phospho-Thr478-peptide antibody. By direct IB, this antibody could not appreciably detect endogenous Myt1 protein from a single egg (although it could detect this protein after immunoprecipitation from 20 eggs; see Supplementary Figure 2). Therefore, to facilitate the detection, we ectopically expressed (a kinase-dead (KD; N231A) form of) Myt1 in oocytes or eggs (so that it would not affect cell cycle progression; see Figure 5). The anti-phospho-Thr478 antibody was now able to recognize the Myt1-KD protein (but not its T478A mutant) expressed in activated eggs, in a Myt1-phosphorylation-dependent manner (Figure 3B). When monitored during oocyte maturation and after egg activation, Thr478 phosphorylation of Myt1-KD was found to occur in maturing (MI) and mature (MII) oocytes, to substantially decrease shortly after egg activation (or during interphase), and then to reincrease at the first mitotic M phase (M1); this pattern of Thr478 phosphorylation closely paralleled that of Cdc2–cyclin B activity during development (Figure 3C). These results, together with the fact that Thr478 also lies in the consensus Cdc2 phosphorylation motif (S/TP; Morgan, 1995), raise the possibility that Thr478 may be phosphorylated by Cdc2. Consistent with this possibility, inhibition of endogenous Cdc2 activity by the Cdc2 inhibitor roscovitine abrogated phosphorylation of Myt1-KD on Thr478 both during oocyte maturation (Figure 3D) and after egg activation (Figure 3E). Moreover, a constitutively active (AF; T14A/Y15F), but not kinase-dead (KR), form of recombinant Cdc2–cyclin B was able to strongly phosphorylate Myt1-KD on Thr478 in vitro (Figure 3F). Endogenous Plx1 was not likely to be significantly involved in Thr478 phosphorylation, since endogenous-level expression of a constitutively active form of Plx1 in oocytes or eggs could not appreciably restore Thr478 phosphorylation in the presence of roscovitine (data not shown). Thus, these results strongly suggest that Cdc2 phosphorylates Myt1 on Thr478 at M phase both during oocyte maturation and after egg activation (or fertilization). Figure 5.Inhibition of embryonic cell division by ectopic expression of the T478A mutant. (A) A total of 50 immature oocytes left uninjected (none) or injected with 2 ng of mRNA encoding either wild-type Myt1 or the T478A mutant were incubated for 12 h, treated with PG, and then cultured and scored for the percentage GVBD. Inset, immunoblot analysis of the Myt1 proteins (just before PG treatment) with anti-Myt1 antibody. (B) In all, 50 two-cell embryos were uninjected (none) or injected (at their one blastomere) with 1 ng of mRNA encoding either wild-type or T478A Myt1, cultured, and analyzed for the external morphology at stage 8 (top; mRNA injected at the right side of the embryos) and for the percentage of embryos that showed a cleavage delay (by 2–3 cycles, determined by counting the number of cells per a fixed area) at stage 8 (bottom; data from three independent experiments with means±s.e.m.). (C) One-cell embryos injected with 2 ng of the above-described mRNAs were cultured and, at the indicated stages, subjected to IB with either anti-Myt1 or anti-Cdc2 pT14/pY15 antibodies. (D) Extracts from stage 8 embryos prepared as in (C) were first immunoprecipitated with anti-Plx1 antibody and then immunoblotted with anti-Myt1 antibody. Input, one embryo; IP, 10 embryos. Download figure Download PowerPoint Phospho-Thr478-mediated interaction between Myt1 and Plx1 after egg activation We then addressed whether Thr478 phosphorylation was required for Myt1 to interact with Plx1 after egg activation. Before addressing this question, we examined the (putative) Plx1 phosphorylation status of Myt1-KD and its T478A mutant during oocyte maturation and after egg activation. Compared to Myt1-KD, the T478A/KD mutant showed a similar (if not identical) mobility upshift during oocyte maturation, but a considerably larger downshift after egg activation (Figure 4A, top), suggesting that the T478A/KD mutant might have failed to be phosphorylated by Plx1 after egg activation. Indeed, when analyzed with a phospho-specific antibody and compared to Myt1-KD, phosphorylation of Ser424, a major Plx1 phosphorylation site of Myt1 (see Figure 2C), was found to be severely impaired in the T478A/KD mutant after egg activation (Figure 4A, bottom). Somewhat surprisingly, Ser424 phosphorylation occurred even during oocyte maturation, but this phosphorylation was little affected in the T478A/KD mutant, suggesting that it was performed perhaps by another kinase during this period. These results would support the idea that Thr478 phosphorylation is required for Myt1 to interact with (and, thereby, to be phosphorylated by) Plx1 after egg activation. Consistently, co-immunoprecipitation analysis revealed that the T478A/KD mutant, unlike Myt1-KD, was unable to interact with (endogenous) Plx1 in activated M1 eggs (Figure 4B, left), whereas it was able to normally interact with p90rsk in mature MII oocytes (Figure 4B, right). Given these results (and the fact that Thr478 lies in the consensus Plk1/Plx1-docking motif), Thr478 phosphorylation may directly mediate the Myt1–Plx1 interaction in activated eggs. Consistent with this idea, short synthetic peptides (residues 470–481) containing phosphorylated Thr478, but not unphosphorylated peptides, could bind Plx1 in extracts from not only M1 but also interphase (I) eggs (Figure 4C). Thus, these results strongly suggest that Myt1 interacts with Plx1 via its Plx1-docking motif containing phospho-Thr478 after egg activation. These results, however, raise the question of why Myt1, although phosphorylated on Thr478 (Figure 3C), cannot interact with Plx1 during oocyte maturation, an issue that will be addressed below. Inhibition of Myt1 by Plx1 during the embryonic cell cycle We examined whether Thr478 phosphorylation (and, hence, Plx1 interaction) of Myt1 was required for progression through oocyte maturation or the embryonic cell cycle. For this, we expressed either the wild-type or T478A mutant forms of Myt1 in oocytes or in just fertilized eggs. The expression of the T478A mutant (about 20-fold over endogenous levels) decreased the rate of germinal vesicle breakdown (GVBD, a hallmark of oocyte maturation) of PG-treated oocytes, but essentially to the same extent as did the expression of wild-type Myt1 (Figure 5A). This result would be consistent with the undetectable physical interaction of Myt1 and Plx1 during oocyte maturation (Figures 1C, D and 4B). In contrast, the expression of the T478A mutant (about 15-fold over endogenous levels; see Figure 5C) in fertilized eggs caused an appreciable (or 2–3 cycles) delay in cell division in more than 65% of the stage 8 blastula embryos, whereas that of the wild-type Myt1 did so in less than 15% of the embryos (Figure 5B) (see Figure 5B legend for determination of the cleavage delay). Consistent with this result, the inhibitory Thr14/Tyr15 phosphorylation of endogenous Cdc2 was significantly stronger in the embryos expressing the T478A mutant than in the embryos expressing wild-type Myt1 (Figure 5C). Moreover, in these experiments, the T478A mutant, unlike wild-type Myt1, failed to interact with (endogenous) Plx1 (Figure 5D) and underwent a mobility downshift (Figures 5C and D), as expected. Thus, it seems clear that Plx1 interaction/phosphorylation via phosphorylated Thr478 can prevent Myt1 from phosphorylating and inhibiting Cdc2 in early embryos (consistent with Plx1 being capable of phosphorylating and inhibiting Myt1 in vitro; Figure 2). Indeed, in in vitro kinase assays, the T478A mutant from activated (M1) eggs had a significantly stronger kinase activity than wild-type Myt1 (see Supplementary Figure 3). Thus, these results show that inhibition of Myt1 by Plx1 plays a role, at least in part, in the mitotic cell cycle of early embryos. Inhibition of the Plx1–Myt1 interaction by the Mos–MAPK pathway Intriguingly, during oocyte maturation, Myt1 could not appreciably interact with Plx1 (Figures 1C, D and 4B), even though it was phosphorylated on Thr478 most likely by Cdc2 (Figures 3C and D). We suspected that the Mos–MAPK pathway, which plays key roles specifically in oocyte maturation (Sagata, 1997; Nebreda and Ferby, 2000), might act to inhibit the Myt1–Plx1 interaction during oocyte maturation. To test this possibility, first we ectopically expressed Mos in activated eggs (in which endogenous Mos was degraded) to reactivate the MAPK pathway (Figure 6A). As seen in Figure 6B, (ectopically expressed) Myt1-KD could not bind to Plx1 in the eggs expressing Mos, whereas it could do so in control eggs expressing no Mos. (The control eggs, about 2.5 h after egg activation, were in an M-like phase as determined by their Cdc2 activity; data not shown.) We next expressed a constitutively active (CA) form of p90rsk, the direct downstream kinase of MAPK (which itself can interact with Myt1 during oocyte maturation), in activated eggs (Figure 6A). Even in this case, Myt1-KD could not appreciably bind to Plx1 (Figure 6B). In contrast to these, Myt1-KD could bind to p90rsk (whether endogenous or exogenous) in both Mos- and p90rsk-CA-expressing eggs, but not in control eggs (Figure 6C). Thus, even in activated eggs, the presence of either Mos or activated p90rsk was able to reproduce the interaction of Myt1 with p90rsk, but not with Plx1, as in mature MII oocytes (Figures 6B and C). To determine whether Mos could inhibit the Myt1–Plx1 interaction independently of p90rsk (Mos can bind to Myt1 during oocyte maturation; Peter et al, 2002), we also coexpressed Mos and MKP-1, a specific inhibitory phosphatase of MAPK (Sun et al, 1993), in activated eggs. In this case, Myt1-KD could normally bind to Plx1 as in untreated control eggs (Figure 6D), indicating the dependence of Mos-inhibited Myt1–Plx1 interactions on p90rsk. These results suggest that p90rsk, the most downstream kinase of the Mos–MAPK pathway, inhibits either directly or indirectly the interaction between Myt1 and Plx1 during oocyte maturation. Consistent with this, the inhibition of MAPK during oocyte maturation allowed the Myt1–Plx1 interaction to occur at GVBD (see Supplementary Figure 4). Figure 6.Inhibition of the Plx1–Myt1 interaction by the Mos–MAPK pathway. Activated eggs expressing Myt1-KD were uninjected (Cont.), injected with either 4 ng of Mos mRNA or 40 ng of Myc-p90rsk-CA mRNA (A–C), or coinjected with 4 ng of Mos mRNA and 4 ng of MKP-1 mRNA (D), after 40 min of egg activation; 90 min later, they were immunoprecipitated with anti-Plx1 (B, D) or anti-p90rsk antibodies (C) and then immunoblotted with anti-Myt1 antibody (B–D). As another control, mature MII oocytes (expressing only Myt-KD as ectopic protein) were also used. In (A) and (D), pMAPK denotes phosphorylated MAPK. Input, one oocyte or egg; IP, 10 oocytes or eggs. Download figure Download PowerPoint Direct inhibition of the Plx1–Myt1 interaction by p90rsk Given the above results, p90rsk might cause some modification (e.g., phosphorylation) of either Myt1 or Plx1 to inhibit their interactions during oocyte maturation. To determine which of the two proteins was modified not to form a stable complex during oocyte maturation, we isolated ectopically expressed His-tagged Myt1 (KD or T478A/KD) and glutathione S-transferase (GST)-fused Plx1 from either mature MII oocytes or activated M1 eggs, mixed and incubated them (GST-Plx1 had been immobilized to glutathione-beads), and then tested for their binding by pulldown assays. GST-Plx1 isolated from either MII oocytes or M1 eggs was able to bind Myt1-KD (but not the T478A/KD mutant) isolated from M1 eggs (Figure 7A, lanes M1), indicating that Plx1 from MII oocytes had the ability to interact with Thr478-phosphorylated Myt1 (isolated from M1 eggs). In contrast, neither the T478A/KD mutant nor Myt1-KD isolated from MII oocytes could bind to GST-Plx1 isolated from either MII oocytes or M1 eggs (Figure 7A, lanes MII). Apparently, this was not due to the binding of (endogenous) p90rsk to the ectopic Myt1 proteins, which was relatively poor (see Figure 6C) and had been eliminated by stringent washing of the Myt1 proteins (see Materials and methods). Thus, these results suggest strongly that, during oocyte maturation or in MII oocytes, Myt1 but not Plx1 is somehow modified so that they cannot form a stable complex. We infer that the Myt1 modification (if any) probably occurred somewhere outside the Plx1-docking motif, because we did observe that short phospho-Thr478-containing peptides (see Figur