Maintenance of spindle attachment to the cortex and formation of the cleavage furrow around the protruded spindle are essential for polar body extrusion (PBE) during meiotic maturation of oocytes. Although spindle movement to the cortex has been well-studied, how the spindle is maintained at the cortex during PBE is unknown. Here, we show that activation of Diaphanous-related formin mediated by mitogen-activated protein kinase (MAPK) is required for tight spindle attachment to the cortex and cleavage furrow closure during PBE in starfish (Asterina pectinifera) oocytes. A. pectinifera Diaphanous-related formin (ApDia) had a distinct localization in immature oocytes and was localized to the cleavage furrow during PBE. Inhibition of the Mos–MAPK pathway or the actin nucleating activity of formin homology 2 domain prevented cleavage furrow closure and resulted in PBE failure. In MEK/MAPK-inhibited oocytes, activation of ApDia by relief of its intramolecular inhibition restored PBE. In summary, this study elucidates a link between the Mos–MAPK pathway and Diaphanous-related formins, that is responsible for maintaining tight spindle attachment to the cortex and cleavage furrow closure during PBE.
In most animals, fully grown oocytes are arrested at G2 phase of meiosis I (immature oocyte). In these immature oocytes the germinal vesicle (GV) is positioned at the center (mammals) or close to the animal pole (echinoderms, amphibians and fish) (Johnson et al., 2011; Lénárt et al., 2003). Upon hormonal stimulation, oocyte maturation begins, germinal vesicle breakdown (GVBD) occurs, and small polar bodies are extruded from large oocytes by asymmetric spindle localization [polar body extrusion (PBE)] (McCarthy and Goldstein, 2006). GVBD is a hallmark of the resumption of meiosis I and the first PBE is the hallmark of the completion of meiosis I. After GVBD, the spindle forms, moves and attaches to the cortex, and the cleavage furrow forms around the spindle, resulting in PBE (Kishimoto, 2003; Li and Albertini, 2013; Yi and Li, 2012). Movement of the spindle to the cortex has been well studied in mouse oocytes (Cowan, 2007; Dumont et al., 2007; Pfender et al., 2011; Verlhac and Dumont, 2008; Li and Albertini, 2013; Chaigne et al., 2013). However, the maintenance of spindle attachment to the cortex has not been studied and cleavage furrow formation during PBE needs further clarification.
Currently it is known that branched F-actin (dynamic F-actin) is nucleated by the Arp2/3 complex, and straight F-actin (stable F-actin) formation is mediated by spire and formin family proteins. The formin family is composed of seven subgroups that are characterized by having formin homology domains (FH1 and FH2 domains). Diaphanous-related formins (Dia) are a subgroup of formins and there are three Dia proteins in mammals: Dia1, Dia2 and Dia3 (Campellone and Welch., 2010). Dia proteins are required for proper spindle formation, actin tubulin organization, cytokinesis and microtubule–kinetochore attachment (Ishizaki et al., 2001; Watanabe et al., 2008; Bartolini et al., 2008; Cheng et al., 2011; Mao, 2011). The budding yeast Diaphanous-related formins, Bni1 and Bnr1, are important for unequal cell division (Sagot et al., 2002; Gao et al., 2010); however, roles of Dia proteins in PBE have not been studied. Dia proteins have a Dia autoregulatory domain (DAD) at the C-terminus and a DAD-interacting domain (DID) at the N-terminus, and DID overlaps with the GTPase-binding domain (GBD). DAD and DID interact with each other and thereby keep Dia inactive. Small-GTPase binding to GBD enhances Dia activation but is not sufficient to fully activate Dia in physiological conditions (Li and Higgs, 2003).
In starfish (Asterina pectinifera) oocytes, the Mos–MAPK–p90Rsk pathway is activated following GVBD, remains active during maturation, maintains the G1 arrest after maturation and is inactivated after fertilization (Tachibana et al., 1997; Tachibana et al., 2000; Mori et al., 2006). This pathway is crucial in oocytes. Inhibition of Mos synthesis halts oocyte maturation at the end of meiosis I and triggers parthenogenesis (Tachibana et al., 2000), inhibition of MAPK activity in immature oocytes by treatment with a MEK-specific inhibitor (U0126) alters the activities of cell cycle regulators during oocyte maturation (Okano-Uchida et al., 2003), and inhibition of MAPK or p90Rsk activities terminates the G1 arrest despite the absence of fertilization (Tachibana et al., 1997; Mori et al., 2006). Yet, the role of the Mos–MAPK–p90Rsk pathway during PBE has not been studied in detail in starfish oocytes. In other organisms, conflicting results were obtained about whether the Mos–MAPK pathway functions in PBE. Inhibition of Mos1 but not Mos2 ortholog in cnidarian oocytes by morpholino oligonucleotide injection at GV stage prevents PBE (Amiel et al., 2009). In contrast, Mos knockout (Mos-KO) mouse oocytes extrude abnormally big polar bodies and enter parthenogenesis (Choi et al., 1996; Verlhac et al. 2000), but injection of Mos antisense oligonucleotides (MosAS) (Paules et al., 1989) at GV stage prevents PBE in wild-type mouse oocytes. Despite these differences, inhibition of MEK activity by U0126 at GV stage prevents PBE consistently in cnidarian, mouse and pig oocytes (Lee et al., 2000; Lee et al., 2007; Amiel et al., 2009). However, variable results were obtained about MEK inhibition following GVBD in mouse oocytes: extrusion of normal polar bodies, large polar bodies, polar bodies without chromosomes or no polar bodies (Tong et al., 2003; Lee et al., 2007).
Here, we show the requirement of the Mos–MAPK–p90Rsk pathway for PBE and a connection between this pathway and Dia in starfish oocytes, using live-cell imaging. We propose that the Mos–MAPK pathway regulates Dia activity and ensures proper PBE.
Tight spindle attachment at the cortex and PBE are perturbed in Mos- or MAPK-inhibited oocytes
To examine the role of Mos in PBE in starfish oocytes, fully-grown (see below) immature oocytes were injected with control oligonucleotide or MosAS, which prevents Mos synthesis in starfish (Tachibana et al., 2000), and then meiotic maturation was induced by treatment with 1-methyladenine (1-MeAde; starfish maturation-inducing hormone). All oocytes injected with MosAS failed in PBE whereas control oocytes extruded PBs (Fig. 1A).
Live-cell imaging was used to examine the effect of Mos on spindle and chromosome dynamics. Fully-grown immature oocytes were microinjected with HiLyte-488-labeled tubulin (green) and Alexa-Fluor-568-labeled histone H1 (red) (Hara et al., 2012), together with control oligonucleotide or MosAS and then meiotic maturation was induced.
In control oocytes, the spindle was formed in the GV region under the cortex, moved slightly towards cortex (data not shown) and then the spindle attached to the cortex and rotated; meanwhile chromosomes were aligned at the metaphase plate (Fig. 1B, 56–66 minutes). Upon completion of spindle orientation, the spindle was pushed against the animal pole (spindle repositioning), which caused the animal pole to move outwards; meanwhile chromosome separation began, indicative of anaphase (Fig. 1B, 73 minutes). After completion of chromosome separation, the animal pole was pulled back, but the spindle was kept relatively distant, resulting in formation of a protrusion (Fig. 1B, 81 minutes, arrow; see below Fig. 3E, 52–61 minutes). Upon completion of chromosome separation, the cleavage furrow began to close in the region between the chromosomes, and then abscission of the cleavage furrow was completed, resulting in the first PBE (Fig. 1B, 81–85 minutes). Thereafter, the second meiotic spindle formed in the same region and the second polar body was extruded (Fig. 1B, 91 minutes; supplementary material Movie 1).
When the oocytes were injected with MosAS, although the spindle oriented under the cortex and spindle repositioning started similarly to that in control oocytes (Fig. 1C, 49–69 minutes), there were several defects: (1) the protrusion was larger than that of control oocytes (Fig. 1C, 72–78 minutes); (2) the spindle pole was not closely attached to the cortex (Fig. 1C, 75–78 minutes, two-sided arrows); and (3) chromosome separation was not properly organized (Fig. 1C, 69–78 minutes). Despite these defects, chromosomes separated completely (Fig. 1C, 75–96 minutes), but the cleavage furrow did not close (Fig. 1C, 103–122 minutes). As a result, PBE failed, the separated chromosomes gathered together and multipolar spindles formed (supplementary material Movie 2), after which parthenogenesis began as described previously (Tachibana et al., 2000).
Next, to examine the effects of MAPK activity on PBE, the MEK-specific inhibitor, U0126, was used. First, to determine the effective concentration of U0126 on MAPK activation, oocytes were collected at 60 minutes after addition of 1-MeAde to immature oocytes in the presence of different concentrations of U0126, and western blotting was performed for MAPK and p90RSK, a kinase downstream of MAPK (Mori et al., 2006). MAPK and p90Rsk from DMSO-treated oocytes at 60 minutes migrated at higher molecular mass (activated forms), but this was prevented by treatment with 2.5 µM and higher concentrations of U0126 (Fig. 2A). Next, oocytes were collected at 10 minute intervals after treatment with 1-MeAde in the absence or presence of 3.5 µM U0126. Consistent with previous results (Okano-Uchida et al., 2003), MAPK and p90RSK were activated at 30 minutes after 1-MeAde treatment, which was prevented by treatment with 3.5 µM U0126 (Fig. 2B). Finally, oocytes that were induced for meiotic maturation in the presence of different concentrations of U0126 were examined for PBE. It was observed that 3.5 µM U0126, but not 1 µM U0126, prevented PBE in 100% of oocytes (Fig. 2C). Higher levels of U0126 (10 µM) were required to prevent PBE when treatment was done after GVBD (data not shown). This phenomenon suggests that the differences in the reports of Tong et al., Lee et al. and Yu et al. might be dependent on the timing and concentration of U0126 treatment (Tong et al., 2003; Lee et al., 2000; Lee et al., 2007; Yu et al., 2007).
To observe the effect of MAPK inhibition on spindle dynamics and PBE, live-cell imaging was performed. DMSO-treated control oocytes succeeded in PBE (supplementary material Fig. S1A; Movie 3). When oocytes were treated with 3.5 µM U0126, there were defects during the spindle repositioning and these defects were similar to those of MosAS-injected oocytes: large protrusions (Fig. 2D), non-tight spindle pole attachment to the cortex, and incorrect chromosome separation (Fig. 2E, 47–54 minutes). Despite these defects, similar to MosAS-injected oocytes, chromosomes once separated completely (Fig. 2E, 58 and 62 minutes), but the cleavage furrow did not close (Fig. 2E, 51–62 minutes), and the separated chromosomes gathered together, and then multipolar spindles formed (Fig. 2E, 68–92 minutes; supplementary material Movie 4).
To exclude the possibility that U0126 may have side effects, oocytes were injected with constitutively active p90Rsk (CA-Rsk) or kinase dead p90Rsk (KD-p90Rsk) (Hara et al., 2009), treated with 20 µM U0126 (eight times higher than the minimum effective concentration of U0126) and then induced for maturation. It was observed that PBE was rescued in all oocytes that were injected with CA-p90Rsk but not in KD-p90Rsk-injected oocytes (Fig. 3A). When oocytes were examined by live-cell imaging it was observed that, in the presence of 20 µM U0126, CA-p90Rsk was able to recover all the defects caused by U0126 treatment and rescue PBE (Fig. 3B; supplementary material Movie 5) whereas KD-p90Rsk-injected oocytes had similar phenotypes to the non-injected oocytes (Fig. 3C; supplementary material Movie 6). These data indicate that U0126 treatment is specific to the Mos–MAPK pathway at least up to a concentration of 20 µM, and exclude the possibility of side effects being involved in the observed defects in 3.5 µM U0126-treated oocytes. The rescue of PBE by CA-p90Rsk seems to conflict with a previous study (Mori et al., 2006) which reported PBE in neutralizing anti-p90Rsk-injected oocytes, but we suspect that since the neutralization of p90Rsk activity was not 100%, a well localized high p90Rsk activity could account for PBE.
In summary, MosAS injection and U0126 treatments indicate that during PBE Mos–MAPK pathway activity is required for proper spindle attachment, chromosome separation and cleavage furrow closure in starfish oocytes.
A. pectinifera Dia localizes to the cleavage furrow during PBE
Although chromosomes separated in Mos- or MAPK-inhibited oocytes, the cleavage furrow failed to close. We further examined this main defect, which might be related to contractile ring formation and thus straight actin nucleators (Barr and Gruneberg, 2007; Fededa and Gerlich, 2012). To determine whether formin proteins are involved in cleavage furrow closure during PBE, oocytes were treated with 100 µM SMIFH2, a specific inhibitor of the formin homology (FH) 2 domain (Rizvi et al., 2009). To avoid perturbation of GVBD and to ensure correct spindle formation, SMIFH2 treatment was performed after GVBD (30 minutes after the treatment with 1-MeAde). In these oocytes, the spindle aligned properly under the cortex and chromosomes separated normally (supplementary material Movie 7; Fig. 4A, 52–72 minutes); however, the cleavage furrow did not close completely and separated chromosomes gathered together (Fig. 4A, 83–100 minutes). The phenotypes of SMIFH2-treated and Mos- or MAPK-inhibited oocytes were very similar regarding cleavage furrow closure, although the phenotype of SMIFH2-treated oocytes was relatively milder.
The above mentioned similarity in the phenotypes indicates there is a connection between the Mos–MAPK pathway and formins. SMIFH2 inhibits all FH2 domain-containing proteins. However, we chose to concentrate on Dia proteins because the cleavage furrow did not close completely in perturbed oocytes and mouse Dia1 and Dia2 are located at the cleavage furrow and are required for completion of cytokinesis in somatic cells (Watanabe et al., 2008; Watanabe et al., 2010). Moreover, Bni1 and Bnr1, Dia counterparts in yeast, are required for completion of budding, a well-studied type of unequal cell division (Evangelista et al., 1997; Yu et al., 2011; Chen et al., 2012).
Only a single Dia gene has been identified in the ongoing starfish genome project organized by this laboratory, despite higher organisms having three Dia genes. This is a common phenomenon when comparing genes in echinoderms and higher organisms (Abe et al., 2010). The A. pectinifera Dia (ApDia) gene was cloned from a cDNA library of immature A. pectinifera oocytes (supplementary material Fig. S2), indicating that it is expressed in immature oocytes.
To identify the localization of ApDia in oocytes, HA-GFP2-ApDia mRNA was injected into immature oocytes. Five hours later, GFP2–ApDia fluorescence in the immature oocytes was examined by confocal microscopy. Fluorescence was intense at the cortex, weak in the cytoplasm and absent from the GV region (Fig. 4B, upper plane). Interestingly, when a thinner confocal plane was used, GFP2–ApDia was absent from a region in the animal pole cortex (Fig. 4B, lower plane). Construction of a three-dimensional view of this region shows an area where GFP2–ApDia is absent (Fig. 3C). This region is the spindle attachment region at which PBE occurs. This data indicates that cortical organization and distinct distributions of cortical proteins are already established in immature starfish oocytes, unlike mouse oocytes (Azoury et al., 2009). Expression of HA–GFP2–ApDia at the expected molecular mass was confirmed by western blotting with an anti-HA antibody (Fig. 4D).
Next, to examine whether ApDia localizes to the cleavage furrow, maturing oocytes expressing HA–GFP2–ApDia was examined by confocal microscopy (supplementary material Movie 8). After GVBD, GFP2–ApDia fluorescence slightly and transiently increased in the GV area (data not shown). Following spindle formation, attachment and rotation, GFP2–ApDia concentrated around the spindle (data not shown). During anaphase GFP2–ApDia was absent from the cortex of the protrusion and concentrated at the cleavage furrow (Fig. 5A, 52–59 minutes) indicating that ApDia forms a ring at the cortex around the spindle. Consistent with this notion, GFP2–ApDia was enriched more at the neck of the protrusion and the size of the ApDia ring diminished together with the closure of the cleavage furrow (Fig. 5A, 60–65 minutes). This localization of ApDia is consistent with its role in cleavage furrow formation (Watanabe et al., 2008). By using a thicker confocal plane, GFP2–ApDia enrichment at the cleavage furrow during PBE was quantified (supplementary material Fig. S3). It was observed that GFP2–ApDia localization at the presumptive PBE region increased when the spindle attached to the cortex and peaked at the time of cleavage furrow formation (supplementary material Fig. S3B, see −17 minutes and 0 minutes, respectively). On average, the GFP2–ApDia signal at the cleavage furrow was 1.67 times higher than that in the control cortex (supplementary material Fig. S3C). To our knowledge, this is the first data reporting localization of a Dia protein at the cleavage furrow during meiotic division.
Next, the localization of HA–GFP2–ApDia protein in MAPK-inactivated oocytes was examined (supplementary material Movie 9). In maturing oocytes that were treated with 3.5 µM U0126, GFP2–ApDia localization was similar to that in control oocytes until spindle repositioning (data not shown). And then, in contrast to control oocytes, GFP2–ApDia became concentrated around the protrusion (compare Fig. 5A, 59–65 minutes and Fig. 5B, 63–95 minutes). In addition, although higher GFP2–ApDia signals were observed at certain points (possible cleavage furrow region) (Fig. 5B, 71–85 minutes, arrowheads), they were not concentrated, which might coincide with the failure of the cleavage furrow closure (Fig. 5B, 106 minutes). These data show that the fine localization pattern of GFP2–ApDia is controlled by MAPK.
The first PBE requires activation of ApDia by relief of the DID–DAD interaction
Because furrow closure was not observed in U0126-treated oocytes, even though GFP2–ApDia was present around the cleavage furrow, we suspected that activation of ApDia in addition to its localization might be controlled by MAPK, and that both localization and activation disturbances might be responsible for PBE failure. To test this idea, first, we aimed to activate endogenous ApDia protein in MAPK-inactivated oocytes. Previous reports showed that exogenous DAD expression competes with the DAD of endogenous Dia, prevents the intramolecular DID–DAD interaction and thereby activates endogenous Dia (Palazzo et al., 2001). Thus, we injected GST–ApDia-DAD protein (amino acids 1004–1105) into immature oocytes, and then induced meiotic maturation in the presence of 3.5 µM U0126. When examined by live-cell imaging, surprisingly, all three defects observed following Mos–MAPK inhibition were rescued in these oocytes: (1) the protrusion was not large; (2) the spindle was closely attached to the cortex; and (3) chromosomes aligned and separated properly (Fig. 6A, 59–72 minutes). Finally, the cleavage furrow closed completely, resulting in the first PBE (Fig. 6A, 75–97 minutes; supplementary material Movie 10), after which the oocyte entered parthenogenesis (data not shown).
Currently known formins that have DID–DAD interaction are FHOD (FH1/FH2 domain-containing protein), FRL (formin-like) and DAAM (Dishevelled-associated activator of morphogenesis) (Campellone and Welch., 2010) and the genes of these formins are found in the starfish genome also. When the protein sequences are compared, the DAD of ApDia shows low similarity to the DAD of these formins except DAAM, but in the literature DAAM is not reported to be involved in cytokinesis. PBE was also rescued when oocytes were injected with GST–ApDia-GBD/DID or GFP–mDia-GBD/DID (supplementary material Fig. S4A–C), which do not show any significant similarity to GBD/DID of other formins. All together these observations suggest that GST–DAD and GST–GBD peptides target Dia, interfere with the intramolecular DID–DAD interaction and specifically activate endogenous ApDia but not other formins.
To further verify that activation of ApDia, following injection of GST–ApDia-DAD, is responsible for the restoration of PBE in MAPK-inhibited oocytes, a mutant with two frame-shift mutations was used. These mutations were a single nucleotide insertion and a single nucleotide deletion at the beginning and end of the DAD, respectively, so that the sequence was changed from DQAGVMDNLLEALQSGTAFNRGDKGGRKR to DQAGGWIIYWRHCRVVRPSTEGDKGGRKR (Fig. 6B). The ApDia-DAD core sequence is underlined and the residues essential for the interaction with DID are in bold (Nezami et al., 2010; Otomo et al., 2010). In this mutant, the C-terminal basic residues (double underlined) are intact, which might be crucial for full ApDia activity (Gould et al., 2011). This mutant is called ApDia-frameshifted DAD (ApDia-fsDAD) and is expected to be constitutively active.
Immature oocytes were injected with HA-GFP2-ApDia-fsDAD mRNA. After 5 hours, the oocytes were induced for meiotic maturation in the presence of 5 µM U0126. Live-cell imaging of maturing oocytes showed that, although fine localization of GFP2–ApDia was not fully recovered, GFP2–ApDia was enriched at the cleavage furrow (Fig. 6C, 64–71 minutes, arrowheads), the cleavage furrow closed and the first PBE was restored in these oocytes (Fig. 6C, 64–77 minutes; supplementary material Movie 11) after which the oocytes entered parthenogenesis (Fig. 6C, 150 minutes).
To further confirm the involvement of actin nucleating activity of ApDia in the rescue of PBE, the I699 residue of ApDia was mutated to alanine (homologous to the I704A mutation in mouse mDia2; supplementary material Fig. S2) to generate ApDia that cannot bind and nucleate actin filaments (Bartolini et al., 2008). In the presence of U0126, expression of this mutant did not rescue the first PBE, whereas HA–GFP2–ApDia-fsDAD did, despite both proteins being expressed at a similar level (Fig. 6D,E; supplementary material Movie 12). These data indicate that the FH2 domain of ApDia functions in PBE and that activation of ApDia, by relief of the DID–DAD interaction, is regulated by MAPK.
In this study, we show that the Mos–MAPK pathway regulates the maintenance of tight spindle attachment to the cortex, proper chromosome separation and cleavage furrow closure during PBE and that these are mediated by Dia in starfish oocytes. ApDia localization was shown for the first time at the cleavage furrow during PBE. Our data suggest that fine adjustment of Dia localization and its activation are regulated by the Mos–MAPK pathway (Fig. 7).
First, live-cell imaging enabled us to observe the fate of single oocytes during maturation in terms of spindle dynamics and PBE. If oocytes are fixed it is not possible to tell whether a single oocyte retains the polar body or there is an occasional polar-body-like structure caused by uncontrolled contractions. By using live-cell imaging we have avoided this, and observed that inhibition of the Mos–MAPK pathway causes several defects that results in PBE failure. These data indicate that the Mos–MAPK pathway is involved in PBE in starfish oocytes. Rescue of PBE by CA-p90Rsk suggests that p90Rsk, which is downstream of MAPK, is also involved in PBE. As mentioned earlier, previous reports showed that Mos-KO mouse oocytes succeeded in PBE but inhibition of the Mos–MAPK pathway activity at the GV phase either with MosAS injection or U0126 treatment prevented PBE in pig, mouse and cnidarian oocytes. Our results are in accordance with the latter one. Although literature data cannot explain why Mos-KO mouse oocytes extruded the PBs, they indicate that Mos–MAPK pathway is involved in PBE not only in starfish oocytes but also in other organisms. Interestingly, we observed (unpublished data) that starfish oocytes that were not fully grown, extruded large PBs in both MosAS-injected and 20 µM U0126-treated oocytes, suggesting that starfish oocytes respond differently in different growth phases, which might be attributed to expression levels of certain proteins in positive- or negative-feedback loops of regulatory systems that are involved in PBE.
It was observed that GFP2–ApDia localized strongly at the cortex of immature oocytes, leaving a ‘hole-like’ empty region at the animal pole. The centrosomes are located close to this Dia-absent region in immature oocytes and during maturation the polar bodies are extruded at this region. Since artificial arrangements in the positioning of GV alters the position of PBE (Matsuura and Chiba, 2004), it is plausible that positioning of the Dia-absent region is also GV dependent. Contents of GV (mainly chromosomes) were previously shown to regulate cortical organization of actin through a Ran gradient and Rac1 activity (Deng et al., 2007; Halet and Carroll, 2007). In this context, it would be interesting to analyze whether regulation of the localization pattern of Dia involves Ran-gradient signals or small GTPases that originate from GV. Very strong F-actin signals at the cortex were observed by using an F-actin probe (GFP–UtrCH) in oocytes of starfish and mouse (Mori et al., 2011; Azoury et al., 2011; Chaigne et al., 2013). In connection with this, localization of Dia at the cortex might be responsible for the nucleation of F-actin at the cortex, which might account for the rigidity of the oocyte cortex.
During PBE, GFP2–ApDia was enriched at the cleavage furrow and formed a ring around the spindle. It was previously shown that three straight F-actin nucleators, formin-2, Spire1 and Spire2, are localized at the cleavage furrow and required for PBE in mouse oocytes (Dumont et al., 2007; Campellone and Welch, 2010; Pfender et al., 2011). DNA sequences coding these proteins are present in the starfish genome also and possibly function in starfish oocytes. Here we add a new F-actin nucleator, Dia, that is required for PBE, but that is different from the others; its localization and activation are under the regulation of the Mos–MAPK pathway. Localizing at the same region and functioning in the same process suggest that these F-actin nucleating proteins are functioning coordinately in PBE. Although the spire family and formin-2 are known to interact directly and function together (Pfender et al., 2011; Vizcarra et al., 2011; Kerkhoff, 2011), any interaction with Dia is yet to be discovered.
Our results suggest that activities of the Mos–MAPK pathway are mediated by activation of Dia. Although the precise mechanism is unknown, our preliminary data indicate that GST–DAD, but not GST–GBD/DID, is phosphorylated MAPK dependently in oocyte extracts. Because of this, we suspect that a regulatory phosphorylation close to the DAD region is involved in the activation of ApDia. It is highly anticipated that, downstream of MAPK, p90Rsk is involved in the regulation of ApDia. However, since there is no solid evidence that ApDia is downstream of p90Rsk, currently, we cannot rule out a possibility that p90Rsk and ApDia constitute two parallel pathways downstream of MAPK, which are able to compensate for each other's activities. Moreover active p90Rsk cannot directly phosphorylate DAD in vitro (data not shown), indicating that even if ApDia is downstream of p90Rsk, there should be at least another kinase that mediates DAD phosphorylation.
It is estimated that physiological concentrations of Rho may not fully activate Dia (Li and Higgs, 2003; Chesarone et al., 2010); therefore, it is possible that Rho and regulatory kinases act together to activate Diaphanous-related formins and that in oocytes these kinases are regulated by MAPK. However, the possibility that activation of ApDia by the Mos–MAPK pathway involves Rho should also be considered. Actually, the loss of the ApDia-absent region at the protrusion in U0126-treated oocytes (Fig. 5B) suggests that localization determinants are also disrupted. These localization determinants may include RhoA and Cdc42; it is suggested that during PBE, Cdc42 drives dynamic F-actin through the Arp2/3 complex, which causes membrane protrusion, and RhoA drives stable F-actin, which restricts this protrusion and exerts constriction at cleavage furrow (Zhang et al., 2008; Maddox et al., 2012). Activation of Arp2/3 during lamellipodial protrusion involves ERK–MAPK activity (Mendoza et al., 2011), and a recent study shows that the Mos–MAPK pathway triggers Arp2/3 activity during spindle movement to the cortex (Chaigne et al., 2013), thus regulation of Arp2/3 during PBE by the Mos–MAPK pathway is also possible. Since regulation of Arp2/3 and Dia might involve Cdc42 and RhoA, respectively, it would be interesting to examine whether active RhoA and active Cdc42 localizations during PBE are regulated by the Mos–MAPK pathway.
It should also be noted that, although the localization pattern of ApDia is not totally rescued by the active Dia (fsDAD mutant), cleavage furrow closure was accomplished and PBE was rescued. This suggests that activity of Dia may be more prominent than its localization. It should be also be considered that we do not know to what extend GFP2–ApDia-fsDAD is overexpressed; since Dia forms dimers and there are inactive endogenous ApDia proteins in U0126-treated oocytes, it is possible that these endogenous inactive proteins alter localization of GFP2–ApDia-fsDAD. This notion can also explain why expression of GFP2–ApDia-fsDAD could not rescue spindle attachment in U0126-treated oocytes, whereas activation of endogenous ApDia by GST–DAD or GST–GBD/DID could rescue all the defects caused by U0126 treatment.
Materials and Methods
Oocyte culture, extract preparation and drug treatment
Fully grown immature A. pectinifera oocytes and oocyte extracts for western blotting were prepared as described previously (Okano-Uchida et al., 1998). Oocytes were incubated in 80% artificial sea water, Sealife (Marinetech, Japan) and maturation was induced by treatment with 1 µM 1-MeAde (Sigma). Diameters of oocytes (doocyte) and GVs (dGV) were measured. Average diameters were, doocyte = 208.8 µm and dGV = 85.0 µm (doocyte/dGV = 0.41) in fully grown oocytes, and doocyte = 199.4 µm and dGV = 72.2 µm (doocyte/dGV = 0.36) in non-fully grown oocytes. To inhibit MEK activity, immature oocytes were treated with U0126 (Sigma). FH2 activity of formins was inhibited by treatment with 100 µM SMIFH2 (Calbiochem) 30 minutes after maturation induction.
Protein and mRNA preparation
ApDia was cloned from a cDNA library prepared from immature A. pectinifera oocytes. For the GST–ApDia-GBD and GST–ApDia-DAD peptides; amino acids 1–323 and 1004–1105 of ApDia were cloned into pGEX-4T-2, and bacterially expressed proteins were purified as described by the manufacturer (GE Healthcare). The GST–mDia-GBD (amino acids 1–300) was a gift from Prof. Hiroshi Itoh (Nara Institute of Science and Technology). mDia-GBD/DID was tagged with GFP and cloned into pET21a (Novagen) to prepare GFP–mDia-GBD/DID protein. To prepare mRNA, the GFP2 DNA sequence was obtained from Clontech, and full-length ApDia tagged with HA–GFP2 at the N-terminus was cloned into a modified pSP64 (Promega) which contained Xenopus globin 5′UTR and 3′UTR sequences. Cloning was done using the In-Fusion PCR cloning system (Clontech). Site-directed mutagenesis was performed with the QuikChange kit (Stratagene). The mutagenesis primers are listed in supplementary material Table S1. Plasmids were linearized with SmaI and mRNA was synthesized using the mMESSAGE mMACHINE SP6 kit (Invitrogen).
Microinjection and microscopy
Microinjection was performed as described previously (Kishimoto, 1986). For Mos experiments, immature oocytes were injected with 100 pg control oligonucleotide (5′-AT*GCCT*T*GCGACACGGC-3′) or 100 pg MosAS (5′-CGGCT*GT*CGCAAGGCAT-3′) (Tachibana et al., 2000) together with HiLyte-488-labeled tubulin (Cytoskeleton) and Alexa-Fluor-568-labeled histone H1 (Hara et al., 2012) and meiotic maturation was induced 30 minutes after injection. For inhibitor treatments, immature oocytes were injected with HiLyte-488-labeled tubulin and Alexa-Fluor-568-labeled histone H1 and meiotic maturation was induced 30 minutes after injection (U0126 was administered together with 1-MeAde but SMIFH2 was administered 30 minutes after 1-MeAde). To activate endogenous ApDia, immature oocytes were injected with 2.4 pg GST–ApDia-DAD, 10 pg GST–ApDia-GBD/DID or 10 pg GFP-mDia-GBD/DID proteins together with HiLyte-488-labeled tubulin and Alexa-Fluor-568-labeled histone H1, and then meiotic maturation was induced 30 minutes after injection. For mRNA injection experiments, immature oocytes were injected with 150 pg mRNA of HA-GFP2-ApDia, HA-GFP2-ApDia-fsDAD or HA-GFP2-ApDia-fsDAD-I699A together with Rhodamine-labeled tubulin (Cytoskeleton), oocytes were incubated for 5 hours at 19°C to allow proteins to be expressed, and then meiotic maturation was induced in the presence of DMSO or U0126. Live-cell imaging was performed with an Olympus Fluoview-1000 confocal microscope equipped with a 40× water immersion objective. During oocyte maturation, 10–24 Z-section images, spanning a depth of 7–15 µm were captured at approximately 1 minute intervals. Z-sections were then compiled to represent single time points and movies were prepared with the compiled images. All experiments were done at least three times with different animals.
Immunoblotting was performed as described previously (Okano-Uchida et al., 1998). Proteins were blotted to Immobilon-P membrane (Millipore) by semi-dry blotting. Membranes were blocked with 5% skimmed milk (for MAPK and p90Rsk) or 2% Advanced Blocking Solution (GE Healthcare; for HA tag) for 1 hour at room temperature. After three washes with TBS-T (Tris-buffered saline + 0.05% Tween-20), antibodies were administered on membranes overnight at 4°C. Antibody dilutions were as follows: anti-MAPK (Millipore) 1∶500 in TBS-T, anti-p90Rsk 1∶200 in TBS-T, anti-HA 1∶1000 in Hikari-A solution (GE Healthcare). After three washes with TBS-T, HRP-conjugated anti-rabbit IgG (MAPK and p90Rsk) or anti-rat IgG (HA) secondary antibodies (1∶5000 dilution each) were administered at room temperature for 1 hour. After washing with TBS-T three times, ECL western blotting substrate kit (GE Healthcare) was used for the detection of MAPK and p90Rsk, or ECL Advance Western Blotting Detection kit (GE Healthcare) for HA.
We thank Hiroshi Itoh for supplying the GST–GBD fragment of mDia and Eiichi Okumura for supplying Alexa-Fluor-568-labeled histone H1.
All experiments were planned and conducted by H.U. The manuscript was prepared by H.U., K.T. and T.K.
This study was supported by grants-in-aid from the Ministry of Education, Science and Culture, Japan [grant numbers 19057003 and 21247030 to T.K.].