In Xenopus, cdc2 tyrosine phosphorylation is detected in the first 60-75 minute cell cycle but not in the next eleven cell cycles (cycles 2-12) which are only 30 minutes long. Here we report that the wee1/cdc25 ratio increases before the first mitotic interphase. We show that the Xe-wee1 protein is absent in stage VI oocytes and is expressed from meiosis II until gastrulation. A dominant negative form of Xe-wee1 (KM wee1) reduced the level cdc2 tyrosine phosphorylation and length of the first cycle. However, the ratio of wee1/cdc25 did not decrease after the first cycle and therefore did not explain the lack of cdc2 tyrosine phosphorylation in, nor the rapidity of, cycles 2-12. Furthermore, there was no evidence for a wee1/myt1 inhibitor in cycles 2-12. We examined the role of Mos in the first cycle because it is present during the first 20 minutes of this cycle. We arrested the rapid embryonic cell cycle (cycle 2 or 3) with Mos and restarted the cell cycle with calcium ionophore; the 30 minute cycle was converted into a 60 minute cycle, with cdc2 tyrosine phosphorylation. In addition, the injection of a non-degradable Mos (MBP-Mos) into the first cycle resulted in a dramatic elongation of this cycle (to 140 minutes). MBP-Mos did not delay DNA replication or the translation of cyclins A or B; it did, however, result in the marked accumulation of tyrosine phosphorylated cdc2. Thus, while the wee1/cdc25 ratio changes during development, these changes may not be responsible for the variety of cell cycles observed during early Xenopus embryogenesis. Our experiments indicate that Mos/MAPK can also contribute to cell cycle length.
In Xenopus, the standard somatic cell cycle is modified during oocyte maturation and early embryogenesis (Kirschner and Gerhart, 1981; Newport and Kirschner, 1982, 1984). Oocyte maturation consists of two consecutive M-phases without an intervening S-phase and results in an egg arrested at metaphase of meiosis II. Fertilization releases the arrest and initiates the first mitotic cycle, which has two atypical ‘gap’ phases and is 60-75 minutes long. The next eleven cycles are composed of alternating S- and M-phases and are only 30 minutes long. The length of the first embryonic cell cycle is curious since the vast stores of nutrients and the lack of transcriptional requirements eliminate the need for growth or gap phases during early embryogenesis (Gerhart, 1980). Here we begin to examine these early embryonic cell cycles in vivo.
The meiotic and mitotic cycles of Xenopus are similar in that entry into these cycles is regulated by Maturation Promoting Factor (MPF), a complex of cdc2 and cyclin B. The catalytic activity of cdc2 is regulated by oscillating levels of cyclin B and phosphorylation on three critical residues (for review see Murray and Kirschner, 1989; Norbury and Nurse, 1992). Phosphorylation on Thr 161 is required for full activation of the kinase, however, this phosphorylation does not appear to be regulated during the cell cycle (for review see Morgan, 1995). In contrast, the phosphorylation of cdc2 at Tyr 15 and Thr 14 is inhibitory and tightly regulated during the cell cycle (for review see Dunphy, 1994). Immature Xenopus oocytes contain inactive stores of cyclin B complexed to cdc2 (pre-MPF), the absence of biological or enzymatic activity is maintained by these inhibitory phosphorylations (Dunphy and Newport, 1989; Gautier and Maller, 1991). Entry into M-phase is regulated, in part, by the dephosphorylation of Tyr 15 and Thr 14 executed by the cdc25 phosphatase (for review see Dunphy, 1994). In Drosophila, these phosphatases (string and twine) are developmentally regulated. The appropriate expression of each isoform is critical for meiosis and the transition from maternal to zygotic control (for review see Edgar, 1995; Orr-Weaver, 1994).
The inhibitory phosphorylations on Tyr 15 and Thr 14 are rendered by at least two kinases; wee1 and myt1. All wee1 homologues isolated to date phosphorylate cdc2 exclusively on Tyr 15 (Booher et al., 1993; Campbell et al., 1995; Featherstone and Russell, 1991; Igarashi et al., 1991; Lee et al., 1994a; McGowan and Russell, 1993; Mueller et al., 1995a; Parker et al., 1991, 1992; Parker and Piwnica-Worms, 1992; Russell and Nurse, 1987). The phosphorylation of Thr 14 (as well as Tyr 15) is mediated by myt1, a membrane associated kinase that is a distinct member of the wee1 family (Atherton-Fessler et al., 1994; Kornbluth et al., 1994; Liu et al., 1997; Mueller et al., 1995b). The regulation of these inhibitory kinases occurs primarily at the level of phosphorylation and is tightly controlled during the cell cycle. The activity is high in interphase and low in mitosis (McGowan and Russell, 1995; Parker et al., 1995; Watanabe et al., 1995). In Xenopus, the negative regulation of these kinases is mediated, in part, by cyclin B/cdc2 (Mueller et al., 1995a,b; Tang et al., 1993).
In addition to MPF, oocyte maturation in Xenopus also requires the Mos protooncogene. In Xenopus, Mos is necessary for entry into meiosis I (Sagata et al., 1988) and meiosis II (Daar et al., 1991; Kanki and Donoghue, 1991). Mos is also involved in the suppression of DNA synthesis during meiosis (Furuno et al., 1994) and it is a key component of cytostatic factor (CSF), which maintains the metaphase arrest at meiosis II (Sagata et al., 1989). Fertilization of the egg causes the destruction of Mos which follows the destruction of MPF by approx. 20 minutes (Watanabe et al., 1991).
In this report we examine the regulation of the embryonic cell cycles in vivo. We show that the level of Xe-wee1 protein increases during oocyte maturation, increasing the wee1/cdc25 ratio before the first mitotic interphase. A dominant negative wee1 (KM wee1) reduced the level of cdc2 tyrosine phosphorylation and shortened the length of the cell cycle indicating that Xe-wee1 does, in part, contribute to the length of this cycle. However, the changes in the wee1/cdc25 ratio cannot mediate all of the cell cycle modifications which occur during early Xenopus embryogenesis. We show that the activation of the Mos/MAPK pathway in the early embryonic cell cycle can result in a longer cell cycle.
MATERIALS AND METHODS
Cloning and mutagenesis
A Xenopus wee1 homologue was isolated by screening an ovarian cDNA library (Stratagene, La Jolla) at low stringency (30% formamide) with the partial human wee1 cDNA (generously provided by H. Okayama). At the amino acid level, Xe-wee1 shares 26% identity with both S. pombe and S. cerevisiae wee1, 57% identity with the human wee1 and 92% identity with the Xenopus wee1 isolated by Mueller (1995a). The GenBank ascension number is AF035443.
The lysine at position 242, in the ATP binding motif, was mutated to an isoleucine by PCR-based site directed mutagenesis. Two mutagenized fragments were generated using the external vector primers (T3 and T7 from Promega) and the following mutagenic oligonucleotide (sense and antisense) – the mutated nucleotide is in bold and underlined; 5′GGTTGTTTCTATGTCATCATACGCTC-CAAGAAGCCATTGGC 3′. 200 pmol of each oligonucleotide was used with 1 μg target DNA, with 25 cycles of 45 seconds at 94°C, 45 seconds at 50°C and 2 minutes at 72°C. The mutagenized fragments were assembled in a second PCR reaction using only the T3 and T7 vector primers with the two mutagenized fragments as template. The resultant DNA was digested with BstEII and BamHI, this fragment was used to replace the corresponding wild-type fragment. The entire region derived from the PCR mutagenesis procedure was sequenced to verify the point mutation.
Xe-wee1 kinase assays
In vitro translated (TNT, Promega) Xe-wee1 was incubated with preimmune or immune sera (Ab 725 or 1532 – 50 μl TNT reaction: 2 μl antibody: 4 μg peptide where indicated). The Xe-wee1 immune complexes were collected on Protein-A Sepharose (Pharmacia) and washed 3 times with Wash buffer (50 mM Hepes pH 7.5, 150 mM NaCl, 5 mM EDTA, 5 mM EGTA) and once with Kinase buffer (20 mM Hepes pH 7.5, 150 mM NaCl, 5 mM MgCl2, 5 mM MnCl2, 10 μM ATP, 1 mM DTT, 10% sucrose). The Xe-wee1 immune complexes were incubated in Kinase buffer supplemented with 10 μCi [γ-32P]ATP (3000 Ci/mmol, Amersham) and kinase-inactive cyclinB/cdc2 complex (2 μl; generously provided by Margaret S. Lee and H. Piwnica-Worms) in a 25 μl reaction for 20 minutes at room temperature (RT). For the phosphoamino acid analysis, the kinase assays were resolved by SDS-PAGE and transferred to immobilon membranes (Millipore). The band corresponding to cdc2 was subjected to acid hydrolysis and 2D electrophoresis as described by Boyle et al (1991). The effect of Xe-wee1 on cyclin B/cdc2 histone H1 kinase activity was analyzed as described above except that the Xe-wee1 immune complexes were incubated in Kinase buffer with ‘kinase active’ MPF (5 ng, UBI) for 15 minutes at RT, after which time 2 μg histone H1, 10 μCi [γ-32P]ATP and PKA inhibitor peptide (5 μM final) was added and the reaction was continued for another 15 minutes at RT.
Histone H1 kinase assays
Oocyte/embryo lysates were prepared in modified EB (80 mM β-glycerophosphate, 20 mM Hepes pH 7.5, 20 mM EGTA, 15 mM MgCl2, 1 mM sodium vanadate, 50 mM NaF, 20 mM sodium pyrophosphate, 2 mM DTT, 1 mM PMSF, 10 μg/ml aprotinin, 50 μg/ml leupeptin, 2 μg/ml pepstatin, 1 μM microcystin, 2.5 μM okadaic acid) at 10 μl/oocyte or embryo. The lysates were clarified by centrifugation at 15,000 r.p.m. for 15 minutes at 4ºC. 5 μl of the resulting supernatant was added to 20 μl of H1 kinase reaction mix (final concentrations, 20 mM Hepes pH 7.5, 5 mM EGTA, 10 mM MgCl2, 0.1 mM ATP, 5 μM PKA inhibitor peptide, 2 μg histone H1 and 10 μCi [γ-32P]ATP). The reactions were incubated at room temperature for 15 minutes then resolved by SDS-PAGE.
DNA replication assays
In vitro matured oocytes were injected with 1 μCi [α-32P]dCTP at time 0 minutes. At the appropriate interval, oocytes/eggs were collected, solubilized in 1% SDS, 10 mM Tris, pH 7.0 and 10 mM EDTA. Lysates were incubated for 1 hour with 10 μg/ml proteinase K, then the nucleic acids were extracted with phenol and precipitated with ethanol. Samples were resolved on a 0.8% TAE agarose gel which was subsequently dried and exposed for autoradiography.
Oocyte and embryo manipulation
Stage VI oocytes were isolated by manual dissection in 1× Modified Barth’s Saline (Specialty Media, Lavallette, NJ) from unprimed females (Xenopus I, Ann Arbor MI). Oocytes were cultured at 16°C in Ooycte Culture Medium (60% Leibovitz-15 medium, 0.4 mM L-glutamine, 1% penicillin-streptomycin (Gibco) and 0.04% BSA). Maturation was induced with 10 μg/ml progesterone. GVBD was scored by the appearance of a white spot in the animal hemisphere and confirmed after fixation in 10% TCA. Nuclear and cytoplasmic fractions were prepared following manual dissection of the germinal vesicle in isolation medium (83 mM KCl, 17 mM NaCl, 6.5 mM Na2HPO4, 3.5 mM KH2PO4). All fractions were resuspended in EB (10 μl/oocyte, nuclei or cytoplasm) and clarified by centrifugation. Injections into stage VI oocytes were done in OCM. MBP-Mos fusion protein was prepared as described by Yew (1992) and injected to a final concentration of approximately 9 ng/oocyte. RNAs were transcribed and capped in vitro using the mMessage machine kit (Ambion), resuspended in DEPC treated H2O and injected at approx. 40 ng/oocyte except as indicated.
Unfertilized eggs were isolated either by in vitro maturation of stage VI oocytes, or by injecting female frogs with 800 U Human Chorionic Gonadotropin 12-16 hours prior to isolation. For the generation of embryos, unfertilized eggs were washed twice in 1× MMR and fertilized with a macerated testis. 15-30 minutes after fertilization the embryos were dejellied (in 2% cysteine/0.3× MMR) and injected with RNA at the one or two cell stage (in 1× MMR/5% Ficoll). In some experiments, the embryos were allowed to develop normally and the stages were determined according to Nieuwkoop and Faber (1967).
Eggs were dejellied prior to activation with ionophore. In vitro matured oocytes, eggs and embryos were washed twice in 1× MMR, then incubated in A23187 calcium ionophore for 5-10 minutes at RT (Boehringer Mannheim; 0.5 μg/ml in 1× MMR for eggs and in vitro matured oocytes; 0.3 μg/ml A23187 in 0.3× MMR for embryos), then washed 3× in 0.1× MMR. Embryos were monitored carefully and transferred to 0.1× MMR as soon as a change in cell shape was evident.
Western blotting and immunoprecipitation
Lysates were prepared in EB and 1-1.5 oocytes or embryo equivalents were loaded per lane. For the analysis of Xe-wee1, two antibodies were generated by immunizing rabbits with the following peptides; KAAELQKQLNVEKFKTAMLERELQAAK (Ab 725) and CRGRKRLVGAKNARSLSFT (Ab 1532). Antibodies 1532 and 725 were used at a dilution of 1:25 for immunoprecipitation and antibody 725 was used at a dilution of 1:1000 for immunoblotting. Mos (SC-86 @ 1:200), MAPK (SC-154 @1:200) and cdc2 (SC-54@1:200) antibodies were purchased from Santa Cruz Biotechnology. Cyclin A1, B1 and B2 antisera (1:750) were a generous gift from J. Maller (University of Colorado). Immunoblotting was completed with an HRP-conjugated goat anti-rabbit or sheep secondary antibody (Boehringer Mannheim) and ECL (Amersham) detection.
For the analysis of cdc2 tyrosine phosphorylation, lysates from 10-15 oocytes or embryos were incubated with 30-50 μl of p13/suc1 beads (Oncogene Science) overnight at 4ºC. Although the presence of endogenous Xenopus suc1 might interfere with the complete collection of cdc2, the use of p13/suc1 beads has been reported to efficiently deplete extracts of cdc2 (Dunphy and Newport, 1989; Izumi and Maller, 1995). The beads were washed 3× in Wash buffer, subjected to SDS-PAGE and transferred to Immobilon membranes which were incubated with the anti-phosphotyrosine antibody 4G10 (1:1000, generously provided by D. Morrison). The analysis was completed with an HRP-conjugated goat anti-mouse secondary antibody and ECL detection.
A Xenopus homologue of wee1 was isolated from an ovarian cDNA library (GenBank accession AF035443). As it shared 92% identity, at the amino acid level, with the Xenopus wee1 homologue isolated by Mueller et al. (1995a), we consider it to be another isoform rather than a distinct family member. We confirmed that this cDNA encoded a bona fide wee1 by several criteria. The in vitro translated Xe-wee1 cDNA was used in an immune complex kinase assay with kinase-inactive cyclin B/cdc2 as a substrate. Xe-wee1 efficiently phosphorylated cdc2 (Fig. 1A, lane 2) and phosphoamino acid analysis demonstrated that cdc2 was phosphorylated exclusively on tyrosine (Fig. 1B). This phosphorylation was functional as it inhibited the histone H1 kinase activity of active cyclinB/cdc2 complexes (Fig. 1C, lanes 3 and 6). We analyzed the biological activity of Xe-wee1 by injecting in vitro transcribed RNA into stage VI oocytes. Following incubation for 16 hours at 18°C, progesterone was added and oocyte maturation was monitored. Injection of 12.5 ng of RNA, transcribed in the sense direction, efficiently inhibited oocyte maturation, while injection of 50 ng RNA, transcribed in the antisense direction, had virtually no effect (Fig. 1D). Thus, Xe-wee1 inhibits entry into meiosis as well as mitosis (Mueller et al., 1995a).
Northern analysis of endogenous Xe-wee1 RNA in adult tissues revealed a single transcript exclusive to the ovary (not shown). The endogenous Xe-wee1 protein was characterized by western analysis using antibody 725, which detected two proteins in egg lysates of approximately 75and 85×103Mr (Fig. 1E, lane 3). Only the lower band represents Xe-wee1: the lower band was not present in uninjected stage VI oocytes, but appeared upon injection of Xe-wee1 RNA (Fig. 1E, lanes 1 and 2; Fig. 1F); the lower band was immunodepleted and immunoprecipitated with two different Xe-wee1 specific antisera (Fig. 1E, lanes 5-10); and the lower band was recognized by a third antibody generated against the entire kinase domain (data not shown). Interestingly, western analysis indicated that Xe-wee1 was not present in stage VI oocytes, but it was readily detected after injection of full length Xe-wee1 RNA (Fig. 1F). Following injection of Xe-wee1 RNA, the protein product was found to be equally distributed between the nucleus and cytoplasm (Fig. 1F, lanes 4-6).
Analysis of endogenous Xe-wee1 during oocyte maturation revealed that the protein product appeared at the onset of meiosis II (Fig. 2A). Using unfertilized eggs activated with ionophore, we also found that the levels of Xe-wee1 did not change during the first cell cycle but that the electrophoretic mobility increased during interphase and decreased in M-phase (Fig. 2B and Mueller et al., 1995a). During early embryogenesis, the Xe-wee1 protein was detected in embryos at fairly constant levels until mid-gastrulation, however, it was undetectable after neurulation (Fig. 2C). The modest increase at the medium cell blastula stage was not consistently observed. Thus, Xe-wee1 is synthesized at the onset of meiosis II and is present at constant levels until gastrulation. The expression pattern of Xe-wee1 differs significantly from that of the opposing phosphatase, Xe-cdc25, which is present at constant levels from immature stage VI oocytes through gastrulation (Hartley et al., 1996; Izumi et al., 1992).
In Xenopus, cdc2 is tyrosine phosphorylated at three points during early development: immature stage VI oocytes; interphase of the first cell cycle and after the mid-blastula transition (MBT; Ferrell et al., 1991). As the levels of cdc25 do not vary during oocyte maturation or early embryogenesis (Hartley et al., 1996; Izumi et al., 1992), the synthesis of Xe-wee1 during oocyte maturation results in an increase in the wee1/cdc25 ratio before the first mitotic interphase. Could this increase in the wee1/cdc25 ratio mediate the cdc2 tyrosine phosphorylation observed in the first cell cycle? We tested this possibility by generating a dominant negative form of Xe-wee1; the lysine at position 242 was changed to an isoleucine. The autophosphorylation activity of the mutated Xe-wee1 (KM wee1) was significantly reduced relative to the wild-type Xe-wee1 when the in vitro translated gene products were assayed in an immune complex kinase assay (data not shown). The biological activity of KM wee1 was assessed by monitoring the effect on oocyte maturation. Following the injection of RNA into stage VI oocytes, the expression of wild-type and KM wee1 was verified by western analysis (Fig. 3A legend). In contrast to wild-type Xe-wee1 which inhibited progesterone induced GVBD (Fig. 1D and Fig. 3A), KM wee1 consistently accelerated the rate of maturation (Fig. 3A), however, it did not induce oocyte maturation in the absence of progesterone.
We examined the effect of KM-wee1 on the first embryonic cycle by injecting stage VI oocytes with KM-wee1 RNA. The expression of KM-wee1 had no effect on the progression from meiosis I to meiosis II (data not shown). The oocytes were matured in vitro with progesterone, then treated with ionophore. The activated eggs did not undergo cleavage since they were not fertilized, however, the profile of H1 kinase activity was similar to that of the first cell cycle in a fertilized embryo (Gerhart et al., 1984; Hartley et al., 1996). By western analysis, the level of KM-wee1 was approx. 2 fold higher than endogenous levels (data not shown). These results were also observed when p13/suc1 precipitates were used to monitor the level of endogenous and KM-wee1 expression (Fig. 3C). The same p13/suc1 precipitate was used to analyze the tyrosine phosphorylation state of cdc2 with the anti-phosphotyrosine antibody 4G10. In control eggs, cdc2 tyrosine phosphorylation was detected 30 minutes following entry into the first cell cycle and was absent by 80 minutes (Fig. 3D, lower pair of panels). We observed that the expression of KM-wee1 significantly reduced the level of cdc2 tyrosine phosphorylation in the first cell cycle (Fig. 3D, upper pair of panels). The absence of cdc2 tyrosine phosphorylation was not due to lower levels of cdc2 protein in the precipitates (Fig. 3D, lower blot in each pair). We determined the effect of KM-wee1 on the length of the cell cycle and found that the cell cycle was always shortened by approx. 10 minutes (Fig. 3B). These data suggest that Xe-wee1 contributes to the cdc2 tyrosine phosphorylation and the length of this cycle.
We were puzzled by the absence of cdc2 tyrosine phosphorylation in cycles 2-12, as the wee1/cdc25 ratio does not decrease after the first cycle. The levels of Xe-wee1 do not decline (Fig. 2C) and the levels of Xe-cdc25 do not increase after cycle 1 (Hartley et al., 1996; Izumi et al., 1992). While it is possible that the levels of myt1 decrease after the first cycle, we tested for the presence of a wee1/myt1 inhibitor in cycles 2-12. We lengthened the rapid embryonic cycles by injecting p21/Cip, an inhibitor of cyclin E/cdk2 (El-Deiry et al., 1993), into one blastomere of a 2-cell embryo. A wee1/myt1 inhibitor would prevent cdc2 tyrosine phosphorylation when the rapid cell cycles were lengthened. The effects of p21/Cip injection are shown in Fig. 4A; the injected cells are larger, reflecting the fact that they have undergone fewer cell divisions. We analyzed the state of cdc2 tyrosine phosphorylation by immunoblotting p13/suc1 precipitates with an anti-phosphotyrosine antibody. As previously reported, we detected tyrosine phosphorylated cdc2 in stage VI oocytes, during the interphase of the first cell cycle, and in gastrula and neurula stage embryos (Fig. 4B), (Ferrell et al., 1991). We did not detect cdc2 tyrosine phosphorylation in unfertilized eggs, p21/Cip antisense injected embryos, uninjected early blastula stage embryos or embryos arrested in M-phase with Mos protein (Fig. 4B). In contrast, we did detect tyrosine phosphorylated cdc2 in embryos where one blastomere was injected with p21/Cip sense RNA (Fig. 4B, p21 sense). Thus, cdc2 tyrosine phosphorylation was observed when the rapid cell cycles of the embryos were lengthened with p21/Cip, arguing against the presence of a wee1/myt1 inhibitor in these cycles.
We postulated that there could be an additional component in the first cell cycle, that extends the length of this cycle and results in cdc2 tyrosine phosphorylation. A cdk inhibitor such as p21/Cip would be an ideal candidate as it would inhibit DNA replication while the egg completes meiosis II. However, the Xenopus p21/p27 homologue is highly expressed at gastrulation and very little is found during cycles 1-12 (Shou and Dunphy, 1996). Alternatively, it is possible that the Mos oncogene could function to lengthen the first cell cycle; Mos inhibits DNA synthesis during meiosis (Furuno et al., 1994) and it is present during the first 20-30 minutes of the first cycle (Watanabe et al., 1991). To test the role of Mos in the regulation of cell cycle length, we injected Mos RNA into one cell embryos thereby arresting the embryos at the M-phase of cycle 2 or 3 (Sagata et al., 1989). To restart the cell cycle, the Mos-arrested embryos were treated with calcium ionophore which dramatically altered the cell shape and caused the pigment to migrate to the animal hemisphere (Fig. 5A). We monitored the histone H1 kinase activity following activation; the exit from the M-phase arrest was identical in both the Mos-arrested embryo and unfertilized egg (Fig. 5B, 0-10 minutes). Strikingly, the profile of H1 kinase activity following the activation of the Mos-arrested embryo was also identical to that observed in the unfertilized egg (Fig. 5B, 0-70 minutes), both display an approx. 60 minute cell cycle. Furthermore, our results show that the tyrosine phosphorylation of cdc2 which is normally detected during the first interphase (Fig. 5C, left panel) is also detected in the interphase following the activation of the Mos-arrested embryo (Fig. 5C, right panel). These results demonstrate that the 30 minutes embryonic cycle (with no detectable cdc2 tyrosine phosphorylation) can be converted into a 60 minute cycle (with cdc2 tyrosine phosphorylation) following the arrest with Mos and reactivation with ionophore.
To further assess the role of Mos in the regulation of cell cycle length, we examined the consequence of constitutive Mos expression in the first cell cycle. We injected stage VI oocytes with either the wild type (WT) or kinase mutant (KM) versions of the MBP-Mos fusion protein. Following maturation in vitro with progesterone, we activated the eggs and monitored the level of endogenous Mos and exogenous MBP-Mos. While the endogenous Mos was degraded after 20-30 minutes (Fig. 6A, upper panel), the MBP-Mos fusion proteins were not degraded following activation (Fig. 6A, middle and lower panels). We assessed the tyrosine phosphorylation state of MAP kinase (MAPK) which reflects the activation of this signaling cascade by Mos (Nebreda and Hunt, 1993; Posada et al., 1993; Shibuya and Ruderman, 1993; for review see Kosako et al., 1994). In control eggs, the pattern of MAPK tyrosine phosphorylation closely follows the levels of endogenous Mos: both are present for the first 20 minutes of the cell cycle (Fig. 6B, upper and middle pair of panels; Roy et al., 1996). However, when the eggs were injected with the wild-type MBP-Mos fusion protein, MAPK remained tyrosine phosphorylated during the entire time course (Fig. 6B, lower pair of panels). These data demonstrate that the MBP Mos fusion proteins are not degraded following activation and that the wild-type MBP-Mos fusion protein remains active throughout the entire time course.
The effect of constitutive wild-type Mos expression was striking, while the kinase inactive MBP-Mos had no effect on the length of the cell cycle, the wild type MBP-Mos extended the length of the first cycle to 140 minutes (from 80 minutes) (Fig. 6C). These results indicate that the elongation of the cell cycle is not simply due to an influx of calcium or the recovery from an M-phase arrest. Moreover, the effect of Mos must be mediated through the kinase, as the kinase inactive MBP-Mos has no effect (Fig. 6C).
We examined the mechanism by which Mos mediated this M-phase delay. We assessed the effect of constitutive Mos expression on the accumulation of cyclin A1 and B1/2. Western analysis shows that wild type MBP-Mos did not markedly alter the appearance of either cyclin A1 or B1/2 (Fig. 6D, 0-60 minutes), although it dramatically inhibited the degradation of these cyclins (Fig. 6D, 80-100 minutes). Therefore, the delayed activation of MPF does not appear to be mediated through the slower translation of either cyclin A1 or B1/2. We examined the effect of constitutive Mos expression on DNA replication. Previous studies have shown that Mos is involved in the suppression of DNA replication during meiosis (Furuno et al., 1994). Furthermore, the inhibition of replication in the mitotic cycle by p21/Cip 1 in Xenopus extracts (Jackson et al., 1995; Strausfeld et al., 1994) blocks the subsequent entry into M-phase (Guadagno and Newport, 1996). However, we observed no differences between the wild type MBP-Mos injected, kinase mutant MBP-Mos injected or uninjected eggs with respect to the timing of the onset of DNA replication (Fig. 6E). These data indicate that Mos does not appear to be solely responsible for the inhibition of replication, and the Mos mediated elongation of the cell cycle does not appear to be mediated through the inhibition of S-phase.
Lastly, we examined the state of cdc2 tyrosine phosphorylation since neither cyclin synthesis nor DNA replication was markedly affected in the presence of wild type MBP-Mos (Fig. 6D,E). While the initial appearance of cdc2 tyrosine phosphorylation appears to be unchanged in the presence of the wild type MBP-Mos fusion protein, the levels and duration of tyrosine phosphorylated cdc2 increase significantly (Fig. 6F, bottom panel). These differences were not due to changes in the levels of cdc2 found in the p13/suc1 precipitates (data not shown) and the presence of the KM Xe-wee1 had no significant effect on the levels of cdc2 tyrosine phosphorylation or cell cycle length (data not shown). Collectively, these data indicate that the activation of the Mos/MAPK pathway in the mitotic cell cycle results in a longer cell cycle and the accumulation of tyrosine phosphorylated cdc2.
In Xenopus, the cycles which precede the MBT are minimal cell cycles consisting of alternating S- and M-phases (Newport and Kirschner, 1982). The first cell cycle is unusual in that it appears to have gap phases, yet there are no transcriptional or growth requirements at this point in development. In this cycle, the time required for the completion of meiosis II appears to account for the ‘gap’ before S-phase, while the time required for the fusion of the sperm and egg nuclei may account for the ‘gap’ after S-phase (Gerhart et al., 1984; Gerhart, 1980; Kirschner et al., 1981) (Fig. 7). This cycle is also unique in that cdc2 is tyrosine phosphorylated in the interphase of this cycle but not in the meiotic ‘interphase’ or the interphases of cycles 2-12. We observed that Xe-wee1 protein was absent in immature stage VI oocytes and was synthesized at the onset of meiosis II. This pattern of expression results in an increase of the ratio of wee1 to cdc25 before the interphase of the first cell cycle, indicating that Xe-wee1 might contribute to the cdc2 tyrosine phosphorylation observed in this cycle. Interestingly, these findings also indicate that Xe-wee1 is not involved in the generation or regulation of pre-MPF in immature stage VI oocytes. We examined the role of Xe-wee1 in the first cell cycle using a dominant negative (KM-wee1) which modestly decreased the level of cdc2 tyrosine phosphorylation and the cell cycle length (Fig. 3). While these results indicate that Xe-wee1 may, in part, contribute to the length of this cycle, the pattern of Xe-wee1 expression did not explain the lack of cdc2 tyrosine phosphorylation, nor the rapidity of the following eleven cycles. The wee1/cdc25 ratio does not decrease during cycles 2-12. In fact, we found that the levels of Xe-wee1 declined after gastrulation (Fig. 2C) when cdc2 tyrosine phosphorylation and cell cycle length are greater.
What prevents the appearance of cdc2 tyrosine phosphorylation in cycles 2-12? We tested for a wee1/myt1 inhibitor (or cdc25 activator) in cycles 2-12 by slowing down the rapid cycles (cycles 2-12) then monitoring the level of cdc2 tyrosine phosphorylation. We hypothesized that the presence of a wee1/myt1 inhibitor would prevent the appearance of tyrosine phosphorylated cdc2 when the cell cycles were slower. However, the injection of p21/Cip into cleaving embryos (which resulted in slower cell divisions) resulted in the detection of tyrosine phosphorylated cdc2 (Fig. 4). These results argue against the presence of a wee1/myt1 inhibitor (or cdc25 activator) in cycles 2-12.
We then considered the possibility that the first cell cycle might have an additional component (that is degraded before the second cycle), which extends the length of the cycle and results in cdc2 tyrosine phosphorylation. A Xenopus homologue of p21/Cip would be an ideal candidate, however, the expression of the Xenopus p21/p27 homologue is very low in eggs and early embryos (Shou and Dunphy, 1996). Although it is possible that other inhibitors may be specific to cycle 1, this pattern of expression has not been described (Kumagai and Dunphy, 1995; Lee and Kirschner, 1996; Lee et al., 1994b; Su et al., 1995). We tested the role of Mos in the first cell cycle because western analysis showed that Mos was present during the first 20-30 minutes of the first cycle (Fig. 6A and Roy et al., 1996; Watanabe et al., 1991). The injection of Mos into fertilized eggs arrested the embryos at M-phase of cycle 2 or 3. After the cell cycle was restarted by ionophore treatment, the resultant cell cycle was 60 minutes long, twice the usual length (Fig. 5). We also detected cdc2 tyrosine phosphorylation in these Mos-activated/ionophore-activated embryos. It is unlikely that these effects are the result of the calcium influx; experiments done with calcium chelators indicate that calcium promotes H1 kinase activation (Lindsay et al., 1995; Snow and Nuccitelli, 1993). Another possibility was that our results were simply the consequence of arresting and restarting the cell cycle. We further characterized the role of Mos on cell cycle length by injecting the eggs with a non-degradable form of Mos (MBP-Mos). The constitutive presence of the wild type MBP-Mos fusion protein in the first mitotic cycle resulted in a significant delay in the activation of MPF (approx. 140 minute cell cycle) (Fig. 6C). The delay of MPF activation was not attributable to slower cyclin A or B translation and the onset of DNA replication was not affected (Fig. 6D and E). However, the levels and duration of cdc2 tyrosine phosphorylation were significantly greater in the presence of the MBP-Mos fusion protein (Fig. 6F). While these results are similar to those of Pan et al. (1994) and Picard et al. (1996) who showed that the activation of MAPK in interphase results in a G2 arrest, we observe a M-phase delay rather than a G2 arrest. Collectively, these results suggest that the activation of the MAPK pathway can lead to a longer embryonic cell cycle.
The consequence of Mos/MAPK activation varies as the oocyte progresses through maturation to embryogenesis. Work from our lab and others has shown that Mos promotes entry into meiosis, prevents exit out of meiosis II (for review see Sagata, 1996), and delays entry into mitosis (this paper). These various effects of Mos/MAPK appear to be contradictory, however, we note that Mos mediated entry into meiosis requires several hours (Yew et al., 1992) and that the cast of cell cycle proteins changes significantly during the course of meiotic maturation (i.e., cyclin E and Xe-wee1; Rempel et al., 1995; and Fig. 2). In addition, the timing of Mos/MAPK activation is critical; CSF arrest requires the activation of the Mos/MAPK pathway just before, but not during M-phase (Abrieu et al., 1996), and Mos does not mediate a M-phase arrest in meiosis I (Yew et al., 1992). We propose that the activation of the Mos/MAPK pathway during early interphase delays the onset of M-phase, while the activation of Mos/MAPK in late interphase results in CSF arrest (Abrieu et al., 1996).
At this point the downstream targets of Mos/MAPK are unknown. Both Mos and cdc25 are present in the meiotic interphase yet there is no cdc2 tyrosine phosphorylation (Ohsumi et al., 1994), thus, it seems unlikely that Mos functions to inhibit cdc25. On the other hand, it is possible that Mos may function to activate Xe-wee1, as this kinase is not present until meiosis II (Fig. 2A). However, we have not been able to detect activation of Xe-wee1 by Mos, MEK or MAPK, and the dominant negative Xe-wee1 does not reverse the Mos mediated M-phase delay (data not shown). We have been able to detect both the endogenous Mos and MBP-Mos fusion protein in p13/suc 1 precipitates (data not shown and Zhou et al., 1992). This observation is notable since recent experiments have shown that the immunodepletion of the Xenopus suc1 homologue (p9) at interphase prevents entry into M-phase and leads to the accumulation of tyrosine phosphorylated cdc2. Moreover, the depletion of p9 during M-phase prevents the extracts from activating the destruction of cyclin B (Patra and Dunphy, 1996). The effects of p9 depletion are similar to the effects of Mos; the constitutive presence of Mos in the mitotic cycle delays the activation of MPF with the accumulation of tyrosine phosphorylated cdc2 (Fig. 6C and F), and Mos inhibits the degradation of cyclin in embryos and extracts (Abrieu et al., 1996; Jones and Smythe, 1996; Sagata et al., 1989 and Fig. 6D). One possible explanation of our data would be that the presence of Mos at the beginning of the first cycle affects some modification that is similar to the depletion of p9.
Meiosis consists of two consecutive M-phases and results in an egg arrested at metaphase of meiosis II. Fertilization releases this arrest and marks the beginning of the first cycle. Consequently, the end of meiosis (polar body extrusion), the decondensation of the sperm nucleus and the fusion of the two pronuclei are events which are unique to the first cycle (Fig. 7 and Graham, 1966). Furthermore, the S-phases of cycles 2-12 occur during the cytokinesis of the preceding cycle (Fig. 7). Significantly, the first S-phase cannot occur during meiosis (formation of second polar body) thus, there are two S-phases between fertilization and the first cytokinesis (Fig. 7; Graham, 1966; Graham and Morgan, 1966; Miake-Lye et al., 1983; see also Hartley et al., 1996, showing two S-phase peaks of cyclin E/cdk2 activity before cytokinesis). While these events justify the unusual length of the first cycle, we examined some of the cell cycle components which might regulate the length of this cycle. We found that the wee1/cdc25 ratio increased prior to the first mitotic interphase. While this could explain the presence of cdc2 tyrosine phosphorylation in, and the length of, this first cycle, we believe that other components contribute to the length of this cycle: the ratio of wee1/cdc25 remains the same after the first cycle, and we found no evidence of a wee1/myt 1 inhibitor in cycles 2-12. We propose that the activation of the Mos/MAPK pathway during early interphase contributes to the generation of the atypical gap phases in the first cycle.
We are grateful to H. Okayama for the human wee1 cDNA, M. S. Lee and H. Piwnica-Worms for the purified baculovirus cyclin B/cdc2 complex, B. Vogelstein for the p21/Waf1/Cip vector, D. Morrison for the 4G10 antibody, J. Maller for cyclin antibodies, and W. Dunphy for Xe-wee1 cDNA and antibody. We are indebted to I. Daar, M. Dasso, L. Lock, W. Matten, D. Morrison, M. Oskarsson, R. Rempel, L. Roy, and T. Stukenberg for valuable advice, discussion and critical reading of the manuscript, and we thank J. Newport for advice regarding the activation of arrested embryos. We also thank Ave Cline for expert manuscript preparation. Research sponsored by the National Cancer Institute, DHHS, under contract with ABL.