Meiotic spindle stability depends on MAPK-interacting and spindle-stabilizing protein (MISS), a new MAPK substrate

Meiotic maturation in vertebrate oocytes corresponds to the period including the exit from the G2 block to the arrest in metaphase of the second meiotic division (metaphase II [MII]^*). Meiotic maturation starts with the activation of maturation-promoting factor (MPF), which triggers germinal vesicle breakdown (GVBD), chromosome condensation, and formation of the first meiotic spindle. This spindle migrates to the cortex and after extrusion of the first polar body, the second meiotic spindle forms below the cortex, where it remains stable for hours with chromosomes aligned on the metaphase plate.

The cascade initiated by c-mos, which activates mitogen-activated protein kinases (MAPK) (the Mos/…/MAPK pathway), controls many important events of meiotic maturation (for an extensive review see Abrieu et al., 2001). This pathway controls the migration of the germinal vesicle (GV) to the cortex in Xenopus oocytes and the migration of the first meiotic spindle to the cortex in mouse oocytes (Choi et al., 1996; Gavin et al., 1999; Verlhac et al., 2000a). It is therefore essential for the establishment of asymmetric divisions during meiosis. This pathway also controls microtubule organization in mouse oocytes (Verlhac et al., 1994, 1996). Moreover, the Mos/…/MAPK pathway promotes the cytostatic factor (CSF; Masui and Markert, 1971) arrest of vertebrate oocytes (Haccard et al., 1993; Colledge et al., 1994; Hashimoto et al., 1994; Verlhac et al., 1996). Indeed, vertebrate oocytes are arrested at metaphase of the second meiosis by an egg cytoplasmic substance called CSF. Then fertilization will release the egg from the arrest and allow it to complete meiosis and proceed to mitosis. The CSF arrest of the cell cycle is affected by the Mos/…/MAPK cascade, which results in the activation of the ribosomal S6 kinase p90^rsk (Bhatt and Ferrell, 1999; Gross et al., 1999, 2000). However, the mechanism of the cell cycle arrest in MII has not been fully clarified. Particularly puzzling is the fact that the cell cycle is not arrested at metaphase I (MI), whereas all the kinases mentioned above have already been activated before MI. So molecules activated only in MII and targets of the Mos/…/MAPK pathway could also be involved in the CSF arrest.

To look for physiological effectors of the Mos/…/MAPK pathway in meiosis, we generated a mouse immature oocyte cDNA library that we screened by two hybrid using MAPK as a bait. We isolated a new protein, MAP kinase–interacting and spindle-stabilizing protein (MISS), which is rich in proline residues and has four potential MAPK phosphorylation sites, a MAPK docking site, a PEST sequence, and a bipartite nuclear localization signal (NLS). Apart from a potential human homologue, we did not find any homologies with other proteins in the databases. In mouse oocytes, MISS is a potential in vivo substrate for the Mos/…/MAPK pathway. The endogenous protein accumulates during mouse meiotic maturation and localizes as discrete dots on the MII spindle. Also, a MISS–GFP construct localizes to kinetochore microtubules in MII oocytes. The stability of MISS is regulated during mouse meiotic maturation. It is unstable in MI, where its degradation is microtubule dependent, and becomes stable in MII. MISS is again unstable after parthenogenetic activation and the endogenous protein is no longer detected in early embryos. By microinjecting both antisense (as) RNA, double-stranded (ds) RNA, and morpholino oligonucleotides (Summerton and Weller, 1997) directed against MISS endogenous mRNA, we show that it is involved in maintaining MII spindle integrity. Mouse oocytes microinjected with any of these three different molecules, and not control molecules, display disorganized MII spindles and numerous asters in their cytoplasm. Moreover, the spindle defects induced after microinjection of morpholino oligonucleotides directed against 25 bp upstream of the ATG can be rescued by the overexpression of Myc–MISS from RNA corresponding to the full-length ORF devoid of MISS 5′ UTR. The injection of dsRNA against MISS has yet no effect on the first two mitotic divisions, suggesting a specific role for MISS in meiosis. MISS is the first example of a physiological MAPK substrate that is stabilized in MII and specifically regulates MII spindle integrity during the CSF arrest.

We have prepared a two-hybrid cDNA library from 1,000 mouse immature oocytes. This library was screened using rat extracellular regulated kinase 2 (ERK2, one of the two mammalian MAPK) as a bait. From this screen, six clones were isolated that encoded a partial cDNA of ∼500 bp. We confirmed the specificity of the two-hybrid interaction between this partial cDNA product and ERK2 (Fig. 1 A).

A mouse EST showing 95 overlapping base pairs with our partial cDNA enabled us to obtain the full-length ORF from the databases (Fig. 2 A). A full-length clone, identical to ours, has recently been pulled out from an mCAP-enriched cDNA library (Kawai et al., 2001). The sequence of the MISS mRNA and protein is presented in Fig. 2 B. This protein is very rich in proline residues in its NH₂-terminal end and harbors a PEST sequence, a bipartite NLS that contains a MAPK docking site (Tanoue et al., 2000), and potential MAPK phosphorylation sites (Fig. 2 C).

We confirmed the two-hybrid interaction by showing that MISS coimmunoprecipitates with endogenous MAPK from Xenopus oocyte extracts (Fig. 1 B).

To check if the full-length RNA encoding MISS was present in mouse oocytes, we performed RT-PCR using primers specific for the MISS mRNA. By RT-PCR, we could amplify a specific band of ∼800 bp corresponding to the full-length MISS ORF (Fig. 1 C). As shown in Fig. 1 C, MISS mRNA is expressed in mouse ovaries, mouse immature oocytes (at the GV stage), and two cell stage mouse embryos. However, both by immunoblotting (Fig. 1 D, lane 1) and immunofluorescence (Fig. 3 A) using an affinity-purified anti-MISS peptide antibody, we could not detect the presence of the endogenous protein in immature oocytes. MISS was present in mature oocytes (Fig. 1 D, lane 2). The apparent molecular weight being around 45 kD suggests that MISS undergoes posttranslational modifications. By immunofluorescence, MISS was undetectable in GV oocytes (Fig. 3 A), was hardly detectable in the vicinity of the chromosomes in late MI oocytes (Fig. 3 B), and accumulated strongly and became localized in MII-arrested oocytes (Fig. 3, C and D). Both the cytoplasmic and the MII spindle staining are specific because they are no longer present when the purified antibody is preincubated with the immunogenic peptide (Fig. 3 F). Interestingly, we could not detect any staining in early mouse embryos (Fig. 3 E), suggesting that MISS accumulates specifically in meiosis. A MISS–GFP construct also behaves like the endogenous protein during meiotic maturation. Indeed, the overexpressed MISS–GFP protein is detected by immunoblotting in MII oocytes (Fig. 4 F), and no localization of the MISS–GFP protein is detected in MI oocytes (Fig. 4 A). In MII oocytes, the MISS–GFP localizes in patches along fibers that evoke kinetochore fibers emanating from the chromosomes (Fig. 4, B, F, and G). The MISS–GFP protein also localizes along the spindle in mitotic dividing embryos (Fig. 4 D). After activation and in interphase, the MISS–GFP protein is completely excluded from the cytoplasm and present only in the nucleus, where small dots are observed (Fig. 4, C and E). This result suggests that the NLS present in the protein is functional and involved in a tight regulation of MISS localization.

Our anti-MISS antibody was not very sensitive by immunoblotting, as it required 2,000 MII oocytes (one mouse gives ∼40 oocytes). To analyze more easily the behavior of MISS during mouse meiotic maturation, we microinjected Myc–MISS RNA into mouse immature oocytes and followed the tagged protein during meiotic maturation.

First we checked that the Myc–MISS behaves like the endogenous MISS. Indeed, the exogenous protein could not be detected in immature oocytes, whereas it was present in mature oocytes (Fig. 5 A, lanes 1 and 2). We microinjected chimeric RNA that contained the MISS ORF flanked by the 5′ and 3′ UTR of the Xenopus β-globin gene, thus the translation of the Myc–MISS encoding RNA was not submitted to translational regulations by MISS endogenous 5′ and 3′ UTR. Using the same construct, we typically observed similar amounts of translation of the exogenous RNA both in immature and mature oocytes when allowed to overexpress for the same amount of time (Fig. 5 A; the levels of expression of Myc–ERK2 are similar in immature and mature oocytes).

Then, we injected a pool of GV oocytes with RNA coding for Myc–MISS and cultured them for the same total amount of time (14 h). We collected them at different stages after meiosis resumption: early MI (2–6 h after GVBD); late MI (8 h after GVBD); or MII (12 h after GVBD). This kind of experiment is possible in mouse oocytes by releasing oocytes after a different length of time spent in medium supplemented with dbcAMP. Meiosis resumption of mouse oocytes occurs spontaneously by removing the oocytes from the ovaries. However, they can be maintained at the GV stage by keeping them in a dbcAMP-containing medium (the dbcAMP mimics a high concentration of AMPc, which has an inhibitory effect on meiosis resumption). In this way, each sample belongs to a pool of oocytes that were injected at the same time and that had expressed the protein for the same amount of time. We could not detect Myc–MISS protein in early MI oocytes (Fig. 5 B). Myc–MISS accumulates only late during meiotic maturation, around the first polar body extrusion (Fig. 5 B). It suggests that the protein is unstable in MI. We show here that Myc–MISS accumulates rapidly around the first polar body extrusion.

Because the kinetics of Myc–MISS accumulation are similar in wild-type compared with mos^−/− oocytes (Fig. 5 C), which do not have MAPK activity (Verlhac et al., 1996), the Mos/…/MAPK pathway does not regulate MISS stability.

To demonstrate that Myc–MISS is indeed unstable in MI due to a high turnover of the protein, we examined its stability in oocytes incubated in nocodazole. In mouse oocytes, cyclin B1 is stabilized by nocodazole treatment, suggesting that its degradation requires intact microtubules (Kubiak et al., 1993). We treated Myc–MISS-injected oocytes with nocodazole for 3 h. As shown in Fig. 5 D, this treatment induces Myc–MISS and cyclin B1 accumulation before polar body extrusion. This shows that MISS is unstable in MI and is degraded by a microtubule-dependent mechanism.

To analyze MISS behavior after parthenogenetic activation, we microinjected Myc–MISS RNA into MII-arrested oocytes. Half of the injected oocytes were activated by ethanol treatment, whereas the other half was not treated. As shown in Fig. 5 E, Myc–MISS accumulates during the MII arrest (lanes 3 and 4), whereas it is unstable after activation (lanes 1 and 2).

As we observed by immunofluorescence on the endogenous protein, Myc–MISS protein accumulated late in MI. Based on the nocodazole experiment and the behavior of other Myc-tagged proteins in mouse oocytes (i.e., Myc–ERK2; Fig. 5 A; Verlhac et al., 2000b), we conclude that MISS is stabilized only during the MII stage.

The protein sequence of MISS displays four potential ERK phosphorylation sites and a MAPK docking site. We have shown that active MAPK can phosphorylate MISS in vitro (unpublished data). To determine whether the protein is an in vivo substrate of MAPK, we checked for its electrophoretic mobility in oocytes from mos^−/− oocytes. In these oocytes, the migration of Myc–MISS is faster compared with its electrophoretic mobility in wild-type oocytes, reflecting a posttranslational modification of Myc–MISS in wild-type oocytes (Fig. 6 A). Among the posttranslational modifications, we show that at least some of them are associated with phosphorylation. Indeed, treatment of the protein isolated from MII-arrested oocytes with λ-phosphatase induced a downshift of its electrophoretic mobility (Fig. 6 B, compare lanes 1 and 2).

Moreover, the migration of Myc–MISS in 2D gel electrophoresis is shifted to the acidic pole (H⁺) in wild-type oocytes compared with its migration in mos^−/− oocytes (Fig. 6 C). Because phosphorylations are associated with negative charges, this result suggests that the protein is more phosphorylated in wild-type oocytes compared with mos^−/− oocytes.

Altogether our results show that Myc–MISS is an in vivo substrate of the Mos/…/MAPK pathway.

To determine the potential role of MISS during mouse meiotic maturation, we decided to interfere with endogenous mRNA expression. Three different types of molecules were used. First, asRNA (asMISS) or dsRNA (dsMISS) directed against the full-length ORF were microinjected into immature oocytes. The microinjection of either asRNA or dsRNA directed against endogenous MISS mRNA had no effect on GVBD or first polar body extrusion (data not shown). However, both injections induced severe spindle disorganization in MII-arrested oocytes. As shown in Fig. 7, B and D, the MII spindle loses its bipolarity and numerous asters appear in the cytoplasm. This phenotype is visible right after the MII spindle is formed (13 h after GVBD). It is never observed in noninjected oocytes or in oocytes injected with an irrelevant dsRNA at the same concentration and directed against Xenopus frizzled 7 (Fig. 7, A and C). As expected, the injection of dsMISS induced a disappearance of MISS endogenous protein in MII oocytes (Fig. 7, compare E and F). Moreover, this phenotype was highly reproducible, as it was observed in >90% of the 96 asMISS- or dsMISS-injected oocytes (Fig. 7 E). Interestingly, this phenotype is similar to the one observed in mos^−/− oocytes, which arrest in metaphase III (MIII) with monopolar spindles after second polar body extrusion (Fig. 8 F; Verlhac et al., 1996).

To prove that the spindle phenotype we observed was specific, we then microinjected morpholino oligonucleotides directed against 25 bp upstream of the first ATG (Summerton and Weller, 1997). Again we observed the same phenotype after injection of morpholino directed against MISS endogenous mRNA (Fig. 8 C), but not after injection of the same morpholino presenting five mismatches, as recommended by the manufacturer (Fig. 8 A). The spindle defects observed after injection of the morpholino oligonucleotide were reproducible and were observed in 63% of the cases (Fig. 8 G). We could further show that these defects were specific by rescuing them with an RNA encoding Myc–MISS, which does not contain the 5′ UTR sequence and is therefore not targeted by the morpholino. The same oocytes were first injected with control morpholino or anti-MISS morpholino, and then were injected with Myc–MISS-encoding RNA. As shown in Fig. 8, B and D, the injection of Myc–MISS alone or the coinjection of the control morpholino together with Myc–MISS had no effect on spindle organization. However, the injection of Myc–MISS restored normal MII spindle organization (Fig. 8, E and G). The coinjection of anti-MISS morpholino with Myc–MISS RNA induced ∼2.5 times less oocytes with MII spindle defects than injection of the anti-MISS morpholino alone. MII spindle destabilization in the treated oocytes was not due to cyclin B1 degradation or global drop in histone H1 kinase activity, as shown in Fig. 8 H.

The fact that we observed the same phenotype using three different types of molecules directed against endogenous MISS mRNA, and not with control molecules, is a strong argument in favor of the specificity of the observed spindle defects. This is further demonstrated by the rescue experiment using Myc–MISS RNA in morpholino-injected oocytes.

Altogether our results using three different approaches demonstrate that MISS maintains a bipolar MII spindle during the CSF arrest of mouse oocytes.

Considering the severe phenotypes observed on MII spindles, we can assume that if such spindle defects are observed after injection of dsMISS into early embryos, they should block cell cycle progression. Indeed, similar defects are observed in mos^−/− oocytes (Fig. 8 F) and induce an MIII arrest. To test this hypothesis and therefore see whether MISS also stabilizes mitotic spindles, we microinjected dsMISS into either zygotes or two cell embryos and checked whether cell cycle progression was inhibited. The injection of dsMISS into zygotes does not block the division from the one to two cell stage: 99% (n = 38) of dsMISS-injected zygotes undergo the one to two cell division like the control (n = 36) dsFz-injected ones. Also, dsMISS injection does not block the two to four cell division: 79% (n = 68) of dsMISS-injected two cell blastomeres divided to four cell blastomeres, comparable to 81% (n = 70) of dsFz-injected blastomeres. We checked that our dsMISS were indeed efficient by coinjecting them together with MISS–GFP RNA (Fig. 9). As for zygotes in interphase (Fig. 4 C), the MISS–GFP protein accumulated in the nuclei of the injected blastomeres (Fig. 9 B). However, the MISS–GFP accumulation was very weak when coinjected with dsMISS, compared with its coinjection with dsFz (Fig. 9, B and D). We confirmed this by immunoblotting with an anti-GFP antibody (Fig. 9 E). The MISS–GFP protein accumulated in dsFz- but not in dsMISS-injected blastomeres.

These experiments demonstrate the specificity and high efficiency of our dsMISS; it can almost completely abolish expression of a protein from an abundant exogenous and normally overexpressed RNA. These experiments are consistent with the lack of endogenous MISS staining observed in early mouse embryos and suggest that MISS does not regulate early mitotic divisions.

When we started this project, two two-hybrid screens had been performed using MAPK as a bait (Cook et al., 1996; Waskiewicz et al., 1997). However, no screen had been performed with a mammalian library prepared from synchronized cells where MAPK is known to play an important role. We therefore chose mouse immature oocytes for construction of our cDNA library. These oocytes do not contain mRNA involved in the support of the whole development, as is the case for Xenopus oocytes, but they present only mRNA essential for supporting meiotic maturation as well as the first zygotic division. Despite the low amount of RNA present in a mouse oocyte (∼1 pg of polyA per oocyte), we were able to construct a cDNA library. To our knowledge our cDNA library is the sole library of this kind available. It provides an opportunity for the functional screening and discovery of mouse genes involved in meiotic maturation as well as early development.

From the screen performed using ERK2 as bait, we isolated a novel protein, MISS, to which, except for a potential human homologue, we cannot find any obvious homologues in the databases. MISS is very rich in proline residues in its NH₂-terminal end and has four potential ERK2 phosphorylation sites, a perfect PEST sequence, and a bipartite NLS that contains a potential MAPK docking site. We confirmed biochemically the two-hybrid interaction by showing that MISS coimmunoprecipitates with endogenous MAPK.

We show by RT-PCR that the mRNA is present in immature oocytes as well as early mouse embryos. We also show by immunoblotting and immunofluorescence using an affinity-purified antipeptide antibody that MISS protein is present in mouse mature oocytes.

Both by immunoblotting and immunofluorescence, we show that although the mRNAs are present in immature oocytes, MISS protein is absent from these oocytes. The endogenous protein starts to accumulate late in MI, around the first polar body extrusion. RNA encoding MISS, fused either to a Myc epitope in an NH₂-terminal position or to a GFP epitope in a COOH-terminal position, cannot induce MISS accumulation in GV or early MI oocytes, whereas RNA encoding Myc–ERK2 is typically associated with similar accumulation of Myc–ERK2 in GV and MII oocytes. Moreover, treatment of injected oocytes with nocodazole induces accumulation of the protein in MI, where it normally does not accumulate. Altogether, these data show that MISS is unstable in MI. The stabilization of the protein occurs in MII, after the first polar body extrusion. After parthenogenetic activation, the Myc–MISS protein does not accumulate and is, therefore, probably degraded. This observation is consistent with the fact that we cannot detect endogenous MISS in early embryos by immunofluorescence. The regulation of MISS stability during meiosis is in agreement with the presence of a true PEST sequence in the protein.

The activity that is involved in MISS degradation during MI and disappears in MII does not depend on the Mos/…/MAPK pathway because Myc–MISS also accumulates in mos^−/− oocytes.

From studies in Xenopus oocytes, we know that MISS is phosphorylated by active MAPK in vitro (unpublished data), and we show that in vivo, MISS is partly phosphorylated by the Mos/…/MAPK pathway. First, we show that MISS is hyperphosphorylated in MII oocytes. Indeed, both endogenous and Myc-tagged MISS migrate at an apparent molecular weight much higher than the calculated one (45 instead of 29 kD), suggesting the presence of posttranslational modifications. Second, Myc–MISS protein migrates as discrete spots in 2D gel electrophoresis, which suggests different phosphorylation states. Third, treatment of Myc–MISS from MII oocytes with λ-phosphatase induces a severe downshift of the protein. Eventually, we show by 1D and 2D electrophoresis that Myc–MISS is in part phosphorylated by the Mos/…/MAPK pathway.

As MISS is stable only in MII, it is reasonable to assume that the protein is regulated in two steps: (1) stabilization, which is independent of the Mos/…/MAPK pathway but might depend on phosphorylation events by other kinases, and, most likely, (2) full activation through phosphorylation by the Mos/…/MAPK pathway. This implies that the protein is mainly active in MII, when CSF activity is present (Masui and Markert, 1971).

The endogenous MISS accumulates in dots on the MII spindles. Also, a MISS–GFP fusion protein shows a strong localization of the protein as bigger patches along fibers that evoke kinetochore microtubules of the MII spindles. In interphase, although we cannot detect the endogenous MISS protein, the MISS–GFP protein concentrates into dots inside the nucleus and is excluded from the cytoplasm. This suggests that the NLS present in the protein is functional.

We cannot explain why the MISS–GFP protein accumulates in early embryos when the endogenous protein does not. One explanation could be that the activity that degrades MISS in GV, early MI, and activated oocytes is no longer present in early embryos, and therefore allows accumulation of an exogenous MISS–GFP protein, whereas it does not allow its accumulation in GV or early MI oocytes. This would imply that a mechanism specific to meiotic maturation controls MISS stability, and this mechanism disappears in early embryos. Alternatively, MISS could undergo other posttranslational modifications during early development that prevent its recognition by the antipeptide antibody.

Nonetheless, we believe that MISS is no longer synthesized during early development. First, we showed that Myc–MISS, which behaves like endogenous MISS and can rescue the depletion of the endogenous protein, is no longer accumulated after parthenogenetic activation. Second, we showed that dsMISS has no effect on the first two mitotic divisions, although we demonstrated its activity on abundant exogenous MISS–GFP RNA.

The localization of MISS and MISS–GFP to MII spindles is consistent with the phenotypes obtained after asRNA, dsRNA, and morpholino injection. Indeed, interfering with the endogenous mRNA by three different means induces severe spindle defects only in MII oocytes, where the protein is present, stable, and probably active. Moreover, we show that the phenotype is associated with a lack of accumulation of endogenous MISS protein in MII. Oocytes that do not possess MISS protein show disorganized spindles often lacking one pole, or have long microtubules emanating from both poles and numerous cytoplasmic asters, which are likely nucleated by cytoplasmic microtubule organizing centers (Maro et al., 1985). The phenotype is specific, because it is rescued by reintroducing Myc–MISS-encoding RNA into anti-MISS morpholino–injected oocytes. This phenotype is also very consistent with the microtubule defects observed in mos^−/− oocytes and with the localization of MAPK at spindle poles of mouse oocytes (Verlhac et al., 1994, 1996). The fact that mos^−/− oocytes present some spindle defects could be due to MISS, which is not fully phosphorylated in this strain and probably not active.

Our results suggest that MISS is a spindle-associated protein that is involved in the stabilization of the MII spindle during the CSF arrest. As mentioned in the introduction, it is puzzling that although the Mos/…/MAPK cascade is active in MI, it does not induce CSF arrest in MI. MISS could be an important target of the MAPK cascade that is only stable in MII (Fig. 10). Because we do not observe a significant drop in histone H1 kinase activity as well as a degradation of cyclin B1 in oocytes treated with dsMISS, we think that MISS plays a role in controlling microtubule dynamics, through an interaction with a microtubule associated protein, rather than a direct role in CSF arrest (Fig. 10).

The discovery of this new protein brings some very important insights concerning the CSF arrest in vertebrate oocytes: some substrate(s) of the Mos/…/MAPK pathway is only stable during MII and the Mos/…/MAPK pathway acts not only on the stabilization of maturation-promoting factor (MPF) activity but also on spindle stability. It will be interesting to analyze the relationships between MII spindle stability and CSF arrest.

Immature oocytes arrested at prophase I of meiosis were obtained by removing ovaries from 11-week-old OF1(WT) and mos^−/− female mice. Oocytes were removed and cultured as previously described (Verlhac et al., 1996). Immunofluorescence was performed as previously described (Polanski et al., 1998). Nocodazole was diluted in the culture medium at 10 μM. Activation by ethanol was performed as previously described (Kubiak, 1989).

Immature oocytes were removed as previously described (Verlhac et al., 1996), collected in sterile PBS, frozen immediately in dry ice, and stored at −80°C. Samples corresponding to 1,200 oocytes were lysed by freezing and thawing. RNA was prepared using the Rneasy Mini Kit (QIAGEN). The genomic DNA was digested by adding 2 U of RQ1 Dnase (Promega) and treating with 40 U of RNase inhibitor (Promega). The RNA was precipitated with ethanol and resuspended in water treated with 0.1% diethyl-pyrocarbonate (H₂O-DEPC). Double-stranded cDNA was synthesized from RNA using a commercial cDNA synthesis kit (SMART PCR cDNA Synthesis Kit; CLONTECH Laboratories, Inc.), as recommended by the manufacturer, and using 5′-AAGCAGTGGTAACAACGCAGAGTGAATTCGCGGG and 5′-GAGAGAGAGAGAGAGAGAGAACTAGTCTCGAG(T)₁₈ primers for the first strand cDNA synthesis. PCR amplification was performed as described in the cDNA synthesis kit using 5′-AAGCAGTGGTAACAACGCAGAGTGAATTCGCGGG and 5′-GAGAGAGAGAGAGAGAGAGAACTAGTCTCGAGTT primers. The cDNAs between 500 pb and 1 kb were gel purified and cloned into pYESTrp (Invitrogen) at EcoRI and XhoI sites. We estimated that the cDNA library consisted of ∼3 × 10⁶ transformants.

A two-hybrid screen was performed using the Hybrid Hunter Kit (Invitrogen). The bait constructs were generated by PCR amplification from plexA-ERK2WT and plexA-ERK2KD (Waskiewicz et al., 1997) using 5′-TGACTAGGATCCGTATGGCGGCGGCGGCG and 5′-TGCATCGAGTTAAGATCTGTATCCTGG primers and cloned as a BamH1/XhoI fragment into pEG202. plexA-Su(Fu) (mouse suppressor of fused) was a gift of A. Plessis (Institut Jacques Monod, Paris, France). We thank J. Cooper for the gift of the pLexA-ERKWT and plexA-ERKKD plasmids.

The pEG202-lexA-ERK2WT construct was introduced into the yeast EGY48 (MATα ura3 trp1 his3 6lexAop-LEU2) strain transformed with pSH18-34. pEG202-lexA-ERK2WT was tested for spontaneous activation of the LEU2 and lacZ reporter genes. The EGY48 strain transformed with pEG202-lexA-ERK2WT and pSH18-34 was further transformed with the cDNA library cloned into pYESTrp. We screened 4 × 10⁶ transformants.

The two-hybrid tests of specificity were performed using different baits in the strain RFY206 mated with the strain EGY48 containing the potential positive clones.

RT-PCR was processed as previously described (Ledan et al., 2001). PCR amplification was performed using 5′-ATGTATCCCTTCATCCCTCCA and 5′-TCATTTAGACCCTTCACTTTC primers.

pRN3Myc–MISS was constructed by RT-PCR amplification from mouse ovaries. Total RNA was extracted using the Rneasy Mini Kit (QIAGEN). The reverse transcription was performed on 500 ng of RNA. The PCR amplification was done on 50 ng of RNA/DNA using 5′-ATGCGAATTCAGAATGTATCCCTTCATCCCT and 5′-ATGCGCGGCCGCTTAAATTAAACTGTTTATTA primers. The 814-bp PCR product was then cloned into pRN3Myc₂. pRN3MISSGFP was obtained by PCR subcloning at XhoI/EcoRI sites using 5′-ATGCCTCGAGATGTATCCCTTCATCCCTCC and 5′-ATGCGAATTCCTTTAGACCCTTCACTTTCA primers. The in vitro synthesis of capped RNA was performed as previously described (Verlhac et al., 2000b).

Microinjection of in vitro–transcribed RNA was performed as previously described (Verlhac et al., 2000b). The sequence of the control morpholino oligonucleotide is 5′-CATACTTCTTGCTGCTGCCCCGTAG and the sequence of the anti-Miss morpholino oligonucleotide is 5′-CATTCTTGTTGGTGCTGCTCCGAAG.

Xenopus oocyte microinjection as well as extraction were performed as previously described (Gavin et al., 1999). For the immunoprecipitation of xp42^mpk1, we used an anti-ERK2 antibody conjugated to agarose (Santa Cruz Biotechnology, Inc.). Xenopus oocyte extracts (300 μl) were precleared with 20 μl of protein A coupled to agarose beads for 30 min at 4°C. The cleared extracts were then incubated for 2 h at 4°C with the antibody coupled to agarose beads. The beads were washed four times in lysis buffer supplemented with 150 mM NaCl and then processed for immunoblotting.

Just after collection and lysis, oocytes were incubated with or without 400 U of λ-phosphatase (New England BioLabs, Inc.) in λ-phosphatase buffer at 37°C for 1 h and then processed for immunoblotting.

Oocytes were stored at −80°C in water containing 1 μM okadaic acid (LC Laboratories) and 80 μM β-glycerophosphate. They were then processed as previously described (Louvet-Vallee et al., 2001).

Oocytes at the appropriate stage of maturation were collected in sample buffer (Laemmli, 1970) and heated for 3 min at 100°C. We used the following antibodies: an affinity-purified anti-MISS antibody directed against the peptide TMLHVNSQHESEGSK; the 9E10 monoclonal antibody (sc-40; Santa Cruz Biotechnology, Inc.) for Myc-tagged MISS; the monoclonal antibody (no. 1814460, Boehringer) against GFP-tagged MISS; the monoclonal antibody (no. M3530; Dako) against cyclin B1; and the anti-ERK antibody (sc-94; Santa Cruz Biotechnology, Inc.).

We thank all the members of the Maro laboratory for their support during the course of this work. We thank P. Hossenlopp (Université Paris 6, Paris, France) for his help with the design of the immunogenic peptide. We thank R. Schwartzmann (Université Paris 6) for his help with confocal microscopy. We thank A. Plessis and members of the C. Isnard (Institut Jacques Monod) laboratory for giving yeast strains and plasmids and for their precious advice with the two-hybrid screen.

This work was supported by grants from the Association pour la Recherche sur le Cancer (9913 and 5792 to M.H. Verlhac). C. Lefebvre and M.E. Terret are recipients of fellowships from the Ministère de l'Education et de la Recherche Technologique.