Metaphase II-arrested mouse eggs are stimulated to complete meiosis by sperm-induced Ca2+ spiking. The Ca2+ signal causes activation of the E3 ligase anaphase-promoting complex/cyclosome (APC), leading to the destruction of key proteins necessary for meiotic exit. We show, using western blots of mouse eggs, the presence of both APC activators cdc20 and cdh1, which target D-box and D-box/KEN-box substrates, respectively, for proteolysis. We decided to examine the temporal activation of APCcdc20 and APCcdh1 by coupling APC substrates to GFP and examining their destruction in real-time following release from second meiotic division arrest. D-box substrates were degraded quickly after the initiation of sperm-induced Ca2+ spiking, such that their degradation was complete by the time of second polar body extrusion. By contrast, KEN-box-containing substrates were degraded when CDK1 activity was low, during the period between polar body extrusion and pronucleus formation. This observation of apparent APCcdh1 activity in meiosis II based on destruction of exogenous GFP-coupled substrates was then confirmed by observing destruction of endogenous APCcdh1 substrates. These data are consistent with a model of initial APCcdc20 activation on sperm-induced activation, followed by APCcdh1 activation after second polar body extrusion. Interestingly, therefore, we propose that mammalian eggs undergo meiosis II with both APCcdc20 and APCcdh1, whereas eggs of other species so far described have APCcdc20 activity only.
Destruction of key cell cycle proteins is coordinated by the activity of the anaphase-promoting complex/cyclosome (APC/C) (Morgan, 1999; Peters, 2002; Zachariae and Nasmyth, 1999). This E3 ligase complex degrades its substrates by ligating them with ubiquitin, thereby earmarking them for immediate proteolysis through the 26S proteasome. Its substrate specificity changes with the cell cycle, such that, for example, cyclin A is ubiquitinated during prometaphase (den Elzen and Pines, 2001; Geley et al., 2001), cyclin B1 and securin during the metaphase-anaphase transition (Clute and Pines, 1999; Hagting et al., 2002) and aurora A and cdc20/fizzy/p55cdc/slp1 during late mitosis and G1 (Hagting et al., 2002; Littlepage and Ruderman, 2002). Thus coordinated destruction of cell cycle proteins is achieved by making them APC/C substrates at specific points in the cell cycle.
The precise mechanisms that underlie why substrates are degraded at each specific phase of the cell cycle are being resolved. One key element is the association of the APC/C with one of two WD-40 repeat-containing proteins: Cdc20/fizzy/p55cdc/slp1 or cdh1/fizzy-related/hct1/srw1/ste9 (Peters, 2002; Pfleger et al., 2001; Visintin et al., 1997). Substrates of APCcdc20, including cyclin B1 and securin, contain a RXXL destruction box (D box) motif that is essential for ubiquitin ligation. Thus D-box mutants of cyclin B1 and securin are stable during metaphase (Clute and Pines, 1999; Hagting et al., 2002), while the D-box motif is portable and confers degradation on otherwise stable proteins (Glotzer et al., 1991). During exit from mitosis both cyclin B and securin destruction are driven by APCcdc20 activity. At this time cdh1 is inactivated by phosphorylation through M-phase (maturation)-promoting factor (MPF) (Blanco et al., 2000; Jaspersen et al., 1999; Zachariae et al., 1998). However, during late mitosis, MPF activity declines as its regulatory cyclin B subunit is degraded by APCcdc20, and at this time APCcdh1 activity can be observed. APCcdh1 ubiquitinates substrates that have either a D-box or a KEN-box motif (Pfleger and Kirschner, 2000; Zur and Brandeis, 2002). Like the D box, the KEN box is portable. One of the main functions of APCcdh1 is thought to be in maintaining low mitotic cyclin levels during G1; it also functions in making a G1/S checkpoint possible (Sudo et al., 2001) and regulates S-phase entry (Bashir et al., 2004; Wei et al., 2004).
Cdc20, but not cdh1, appears to be the mediator of meiotic exit in eggs of species so far described. This has been most directly shown in Xenopus (Lorca et al., 1998) where cdh1 protein is absent in frog eggs and is not expressed until the mid-blastula transition; and as such inhibition of cdc20 blocks the normal process of egg activation. In Drosophila, cdh1 is absent during the early syncytial stages of embryogenesis, functioning only after cellularisation (Raff et al., 2002; Sigrist and Lehner, 1997). And finally in C. elegans cdh1 RNAi embryos can develop into adults (Fay et al., 2002), but cdc20 RNAi embryos arrest at metaphase I (Kitagawa et al., 2002).
In mouse eggs the first reductional meiotic division follows a mid-cycle gonadotropin surge, while the second equational division follows a sperm-triggered Ca2+ signal. Our understanding of these processes in mouse, as a mammalian model system, lags behind Xenopus and other higher eukaryotes. This reflects the fact that mouse eggs are not available in the quantity necessary to facilitate many biochemical analyses and usually require specialized techniques to microinject, such as Piezo-driven or negative-capacitance-driven injection procedures. However mouse eggs are optically clear and are able to translate efficiently exogenous cRNA. Therefore we took advantage of our ability to microinject metaphase II (MII) eggs to examine the role of APCcdc20 and APCcdh1 during exit from MII arrest. We examined the destruction of GFP-coupled substrates of APCcdc20 and APCcdh1 in single eggs in real-time during the second meiotic division. In so doing we found that D-box-containing substrates of APCcdc20 are degraded quickly at fertilization. KEN-box substrates of APCcdh1 were also degraded during meiosis II exit, but degradation was confined to a period following extrusion of the second polar body. We conclude that both APCcdc20 and APCcdh1 play a role in mouse meiosis, in contrast to reports in other higher eukaryotes.
Materials and Methods
All chemicals were from Sigma-Aldrich (UK) unless stated otherwise, and of tissue culture or embryo-tested grade where appropriate. In the present studies outbred MFI mice (Harlan, UK) were used. Ovulated eggs were collected from the ampulla oviductal region, 12-13 hours after intraperitoneal injection of 5 IU human chorionic gonadotrophin in mice that had been primed with 7.5 IU pregnant mares' serum gonadotrophin 44-52 hours previously. Spermatozoa were collected from the caudal region of ex-breeding MF1 mice (Harlan) and capacitated for at least 2 hours in medium T6 at a concentration of 12×105 sperm/ml.
Preparation of cRNA
Full length human cdc20 was amplified by PCR from testis cDNA and cloned into a modified pRN3 vector, designed to produce mRNA transcript C-terminally coupled to GFP, with maximal stability conferred by the presence of 5′ globin UTR upstream and both 3′UTR and a poly(A)-encoding tract downstream of the gene (Levasseur and McDougall, 2000). Cyclin B1, securin and securindm were cloned as described previously (Herbert et al., 2003; Nixon et al., 2002) using the same vector. cRNA was synthesised using T3 mMESSAGE and mMACHINE (Ambion, UK), and dissolved in nuclease-free water to a concentration of 1 μg/μl prior to microinjection. The KEN box of cdc20 was mutated to AAA (cdc20km) using a GeneEditor in vitro mutagenesis kit (Promega, UK) and standard cloning procedures.
Egg microinjection and imaging
Microinjection of cRNA constructs and of Fura2 dextran were made into MII eggs as described previously (Jones and Nixon, 2000). A bolus injection of 0.1-0.3% total egg volume was made and we estimate that between 0.3-1 pg of cRNA was injected, dependent on the size of injection and the dilution made with Fura2. In vitro fertilization was performed on the stage on an inverted microscope fitted for epi-fluorescence (Nixon et al., 2002). Imaging of eggs for intracellular Ca2+ and bright-field imaging was performed with a Xe lamp and CCD camera.
MII eggs were probed for cdc20 and cdh1 protein using one anti-cdc20 polyclonal antibody (sc8358; Santa Cruz, CA, USA) and two monoclonal anti-cdh1 antibodies, AR38 (Cell Cycle Control Laboratory; Cancer Research UK) and ab3242 (AbCam, UK). SDS-PAGE was performed on 10% polyacrylamide gels and primary antibodies were detected using ECL techniques according to the manufacturer's instructions (Amersham, UK).
H1 kinase assays were performed on groups of four eggs at specific times during egg activation. Eggs were lysed in kinase buffer (80 mM B glycerophosphate, 25 mM Hepes, 5 mM EGTA, 10 mM MgCl2, 1 mM DTT, 0.2 mM 4-(2-aminoethyl)benzenesulphonylfluoride hydrochloride (AEBSF), 1 mM benzamidine, 100 μM NaVO4, 5 mM NaF, 10 μg/ml leupeptin, 10 μg/ml pepstatin A and 10 μg/ml aprotinin, pH 7.2). Lysed eggs were then incubated for 30 minutes at 30°C with 0.1 mM ATP, 0.5 mCi/ml [γ-32P]ATP (Amersham, UK), 10 μM cAMP protein kinase inhibitor, in kinase buffer containing 0.15 mg/ml histone H1 (type III-S, from calf thymus). Reactions were terminated by introduction of sample buffer and incubation at 95°C for 5 minutes. Proteins were separated by SDS-PAGE and enzyme activity was visualized using a Fujifilm BAS-1500 Bioimaging analyzer (Fuji).
Cdc20 and Cdh1 protein present in mouse eggs
In this study we were interested in examining how the APC/C is regulated in mouse eggs as they undergo the second meiotic division. This is interesting given the likely signal transduction pathway employed at fertilization, in which a sperm-triggered Ca2+ signal leads to APC/C-mediated destruction of cell cycle factors that are responsible for egg arrest (Jones, 2004; Tunquist and Maller, 2003).
Both of the two main APC/C activators, cdc20 and cdh1, were present in mouse eggs at MII (Fig. 1). We also observed these proteins in oocytes at all stages of the first meiotic division (not shown). The fact that mouse eggs contain cdc20 is not surprising, given the probable role of APCcdc20 in destruction of cyclin B1/securin during both the first and second meiotic divisions in mouse (Herbert et al., 2003; Hyslop et al., 2004; Madgwick et al., 2004). By contrast, we were surprised to find cdh1, given its apparent absence in eggs and embryos of other species so far described (see Introduction). Two different anti-cdh1 antibodies were used to verify the presence of cdh1 in MII eggs, one of which (AR38) has been used previously to probe for mouse cdh1 (Listovsky et al., 2004).
Degradation of APCcdc20 substrates begins minutes after the initiation of Ca2+ spiking
The data above show that cdh1 is present in mouse eggs, but does not address whether APCcdh1 is actually responsible for any protein degradation during meiosis II. This is what we set out to examine. Our approach was to simultaneously microinject MII mouse eggs with fura2-dextran and cRNA for substrates of both APCcdc20 and APCcdh1 that had been C-terminally coupled to EGFP. We would then be able to follow intracellular Ca2+ changes triggered by the fertilizing sperm and the effect of fertilization on the levels of expressed GFP fusion proteins. In these experiments eggs were inseminated with low numbers of sperm, so that extrusion of the second polar body (at 0.6-3 hours after the first Ca2+ spike) and pronucleus formation (at 4-6 hours) could be observed with some accuracy (Fig. 2).
Securin and cyclin B1 degradation should begin at a similar time given that both are APCcdc20 substrates. Therefore eggs were microinjected with either cyclin B1::GFP or securin::GFP and then inseminated. Their profile of degradation was compared relative to the sperm Ca2+ signal and to the morphological events of activation. The results showed that securin was degraded with identical dynamics to cyclin B1 (Fig. 3). Securin::GFP was quickly degraded when eggs were inseminated (Fig. 3A), and the process appeared complete before second polar body formation. When eggs were simultaneously imaged for GFP levels and Ca2+ changes, it was calculated that securin (Fig. 3B, n=16) and cyclin B1 (Fig. 3C, n=16) degradation began at the same time, just a few minutes after the first Ca2+ spike, 12±5 versus 13±5 minutes, respectively (mean ± s.d., not significantly different, P=0.47, t-test). Both cyclin B1 and securin were degraded by 80-90%, such that their minimum intracellular concentration was associated with the time of second polar body extrusion.
Degradation of APCcdh1 substrates begins at second polar body extrusion
The above data suggest that the Ca2+ signal at fertilization rapidly switches on APCcdc20 activity so that cyclin B1 and securin are degraded fully by the time of second polar body formation. Both cyclin B1 and securin are also substrates of APCcdh1, therefore, it is possible that APCcdh1 activity is also responsible for their degradation at this time. However, if there is similar control of meiotic and mitotic cdh1 then the high CDK1 activity present in MII eggs, which does not decline until about 30 minutes after the start of egg activation (Fig. 4), should phosphorylate cdh1 and in so doing prevent any cdh1 association with the APC/C.
Any potential role of APCcdh1 in degradation of cyclin B1 and securin in the time before second polar body formation can be assessed directly by removing the D box of securin and measuring its degradation during fertilization. Wild-type securin contains both a D box and a KEN box. Therefore mutation of the D box (RXXL to AXXA) generates a securin construct (securindm; securin D-box mutant) that is a substrate of APCcdh1 only. In mitotic somatic cells this construct generates a so called `cut' phenotype, in which disjunction of sister chromatids is prevented but cytokinesis is not (Hagting et al., 2002; Zur and Brandeis, 2001). Recently we reported that eggs expressing securindm, which were parthenogenetically activated with Sr2+, went on to extrude a second polar body, but failed to separate their sister chromatids (Madgwick et al., 2004), consistent with studies in mitotic cells. In this study we fertilized securindm-expressing eggs and we also observed the normal morphological events of meiosis II, second polar body extrusion and pronucleus formation. By monitoring the securindm::GFP levels in these eggs, it was found that securindm was degraded following fertilization consistent with APCcdh1 activity (Fig. 5, n=12). However, the timing of degradation was greatly delayed relative to the wild-type securin carrying the intact D box. Securindm degradation began 80±29 minutes after the first Ca2+ spike, this time is similar to when the second polar body is extruded (85±29 minutes). Degradation continued until pronuclei formed in the one-cell embryo. Although the timing of securin degradation was delayed considerably by the mutation of its D box, the actual rate of securindm degradation, as calculated by the decrease in GFP signal, was unaffected. Loss of securindm signal proceeded at 65±20 arbitrary units/hour; while that of wild-type securin was 70±25 arbitrary units/hour (not statistically significantly different).
The above findings indicate that the removal a functional D box from securin uncovers a KEN-box-mediated degradation pathway, which is active following second polar body extrusion. Such timing in the activation of APCcdh1 would be consistent with cdh1 being inactive at metaphase and suggests that meiotic cdh1 may be controlled similar to mitotic cdh1. Phosphorylation changes in cdh1 can be detected on western blots; phosphorylated cdh1 is retarded on gels. Therefore cdh1 protein was compared between eggs and oocytes at metaphase, both MI and MII, and in activated eggs at the pronucleate stage. A cdh1 bandshift was observed in pronucleate stage eggs as compared with those at metaphase, consistent with phosphorylation of cdh1 at the MII stage (Fig. 6).
Cdc20 is degraded in meiosis II
To confirm that APCcdh1 was indeed active during meiosis II the degradation profile of cdc20, a well-characterized APCcdh1 substrate, was examined. Cdc20 has no D box and its only known mechanism of degradation in late mitosis is through APCcdh1, a process that has been widely reported and extensively examined in many other cell types (Kallio et al., 2002; Pfleger and Kirschner, 2000; Raff et al., 2002; Zur and Brandeis, 2002).
Cdc20 was coupled to GFP at its C terminus and cRNA expressed in mouse eggs to act as a reporter for APCcdh1 activity (Fig. 1). In agreement with the temporal pattern of securindm degradation, cdc20::GFP was degraded when inseminated eggs had extruded their second polar body (Fig. 7A, n=7). The KEN box of cdc20 was mutated to AAA (cdc20km) to confirm the need of the KEN box in degradation, and by inference APCcdh1 activity. The KEN-box mutation rendered this protein completely stable in eggs following fertilization (Fig. 7B, n=10), there being no inflexion in the synthesis rate of the cdc20km::GFP construct following polar body extrusion. Although there was a defined period of cdc20 degradation after polar body extrusion (Fig. 7A) that was blocked by removal of the KEN box, the extent of cdc20::GFP degradation was not as great as that for securin. These data suggest that cdc20::GFP may be a poor substrate for APCcdh1, an assumption that has also recently been made in HeLa cells, in which fluorescently labelled cdc20 was poorly degraded by APCcdh1 (Lindon and Pines, 2004). Therefore, although the above data are clear in that there is a period of cdc20::GFP degradation, we decided to confirm that APCcdh1 is effective at degrading endogenous cdc20. Mouse eggs were activated for different periods of time: for 0 hours, as a control; for 2 hours, at which point second polar body extrusion had already occurred; for 6 hours, when pronuclei had formed. In keeping with a role for APCcdh1 during meiosis II, we observed cdc20 destruction in eggs that had extruded their second polar body and those that had pronuclei (Fig. 7C).
In the present study we found that the second meiotic division of mouse eggs was associated with a period of destruction in proteins that contained either a D- or KEN-box motif, consistent with activity of both APCcdc20 and APCcdh1 at fertilization. Following insemination, a sperm-triggered series of Ca2+ spikes quickly induced destruction of both securin and cyclin B1 in eggs. Their destruction was complete by the time CDK1 activity had declined and the second polar body had been extruded. Since cyclin B1 and securin both contain a D box, this degradation could potentially have been due to APCcdc20 and/or APCcdh1. However, in MII eggs we observed that cdh1 protein was phosphorylated, which in mitotic cells prevents it binding the APC/C. The inability of cdh1 to activate the APC/C at MII was confirmed directly by using substrates of APCcdh1 that were not recognized by APCcdc20. Thus we used cdc20 itself and a D-box-mutated form of securin that both contain a KEN box, to generate GFP-fusion proteins that were substrates of APCcdh1. Both of these constructs were initially stable in fertilized eggs, but their degradation began abruptly 1-2 hours after sperm-triggered Ca2+ spiking, at the time of polar body extrusion. Their degradation continued until meiosis II was complete and pronuclei formed in the 1-cell embryo. Therefore there was no immediate activation of APCcdh1 at fertilization, instead there appeared to be a switch from APCcdc20 to APCcdh1 activity at the time of second polar body extrusion (this is depicted in Fig. 8). All substrates of the APC/C appeared stable after pronucleus formation.
Activation of APCcdc20 by Ca2+
Sperm trigger a long-lasting series of Ca2+ spikes by releasing into the egg a sperm-specific member of the phospholipase C (PLC) family, PLCζ (Saunders et al., 2002). This PLC raises egg InsP3 levels, so resulting in a train of Ca2+ spikes (Jones and Nixon, 2000). Ca2+ spiking, in a process mediated by a calmodulin-dependent protein kinase II (Markoulaki et al., 2003; Markoulaki et al., 2004), is the necessary and sufficient trigger for resumption of meiosis following a MII arrest (Hyslop et al., 2004). D-box dependent proteolysis of cyclin B1 (Nixon et al., 2002) and securin (present data) begin at the same time; just a few minutes after the first Ca2+ spike. We interpret this to mean that activation of the E3 ligase APCcdc20 occurs shortly after spiking is initiated in the same way that APCcdc20 has shown to be turned on, and essential, during frog egg activation (Lorca et al., 1998). However, mouse eggs also express RFLP4, an E3 ligase whose substrates include cyclin B1 (Suzumori et al., 2003) that may fulfil the same function as APCcdc20; indeed novel meiosis variants of cdc20 have been described in yeast (Cooper et al., 2000) and Drosophila (Chu et al., 2001). The present data do not rule out a function for RFLP4 in mouse meiosis. However, we show here that mouse eggs do contain cdc20, and previously we have demonstrated that Ca2+-mediated cell cycle resumption and cyclin B1 degradation can be blocked by induction of a spindle checkpoint (Jones et al., 1995; Nixon et al., 2002). This would support a role for APCcdc20 since activation of the spindle checkpoint is a potent repressor of APCcdc20 activity (Yu, 2002). This effect is mediated by mad2, one of the spindle checkpoint proteins, which binds specifically and directly to cdc20 (Luo et al., 2000). Therefore, it is likely that RFLP4 may have a specific role during meiosis I, or alternatively contributes to securin/cyclin B1 degradation during meiosis II, once the APC/C has been activated.
In further support of APCcdc20 having a role during meiosis II is the observation that both D-box- and KEN-box-containing substrates are targeted for proteolysis when CDK1 activity drops at second polar body extrusion. This implies a switch from APCcdc20 to APCcdh1 activity at this time. This switch does not occur in frog eggs that lack cdh1 protein, but the timing of this switch would be consistent with that reported from mitotic studies in mammalian cells (Hagting et al., 2002; Zur and Brandeis, 2002) and the negative regulation of cdh1 by MPF.
At this time it is unknown how Ca2+ activation of calmodulin-dependent protein kinase II leads to a switching on of APCcdc20. Historically the term `Cytostatic factor' has been used to describe the activity in eggs which is responsible for MII arrest. Cytostatic factor activity is likely to constitute a repression of APCcdc20 activity. Indeed, some candidates that have cytostatic factor activity are inhibitors of APCcdc20. These include Emi1, which binds cdc20 and in so doing prevents cdc20 binding the APC/C (Reimann and Jackson, 2002), and various components of the spindle checkpoint (Schwab et al., 2001; Tunquist et al., 2003).
Why some eggs have cdh1 and others do not
Further studies are needed to determine both the function and regulation of APCcdh1 during mouse meiosis II, and in addition, since we have observed its presence at all stages of oocyte maturation, its role in the first meiotic division. It is possible that mouse APCcdh1 is switched off during meiosis I, since by analogy in budding yeast a meiosis-specific inhibitor of cdh1 prevents APCcdh1 activity during the first meiotic division (Bolte et al., 2002). The present work points to a role of CDK1 activity in regulating APCcdh1 during meiosis II, but nothing is known about the role of the phosphatase cdc14, which is necessary to activate cdh1, in mouse eggs. Thus future studies should address how cdh1 is regulated by CDK1/cdc14 during the two meiotic divisions, which temporally are separated by just a few hours. In the mitotic division, cdh1 activity appears readily dispensable, such that cells can still exit mitosis in the absence of cdh1. Therefore it is possible that cdh1 can be dispensed with for the process of completing meiosis II. However, APCcdh1 has been found to have specific functions later in the cell cycle at G1, it is responsible for making a G1-S checkpoint permissible (Sudo et al., 2001) and preventing premature entry into S phase (Bashir et al., 2004; Wei et al., 2004). APCcdh1 appears also to have functions outside the cell cycle, such as in post-mitotic neurons where it regulates the growth of axons (Konishi et al., 2004) and in some studies its presence has been found to be associated with the onset of cell differentiation (Blanco et al., 2000). It will therefore be interesting to determine what role maternal cdh1 plays in embryos up until the activation of the zygotic genome.
Most cell divisions, including those that occur in early mouse embryos and mammalian mitotic divisions have associated gap phases to their cell cycle and so there would be a need for cdh1 in G1 for regulating entry to S phase. However eggs of some species, exemplified by Xenopus, C. elegans and Drosophila, appear to have dispensed with gap phases. Based on the present finding of APCcdc20 and APCcdh1 activity during meiosis II in mouse eggs it is probably not the case that the embryonic cell cycle is fundamentally different to the cell cycle in adult cells. Rather, the lack of cdh1 activity in eggs and embryos of some species probably reflects the requirement to speed up the cell cycle in order to reach a motile stage and avoid predation.
We thank Dr Julian Gannon and Professor Tim Hunt, Cell Cycle Control Laboratory, Cancer Research UK, for the donation of AR38 cdh1 antibody. Drs Junyong Huang, Alex McDougall and Professor Michael Whitaker for critical reading of the manuscript. This work was supported by the Wellcome Trust in the form of a project grant (069236/Z/02/Z) and equipment grant (065354/Z/01/Z) to K.T.J.