The Ascidiacea, the invertebrate chordates, includes three orders; the Stolidobranchia is the most complex. Until the present study, the onset of oocyte maturation (germinal vesicle breakdown) had been investigated in only a single pyurid (Halocynthia roretzi), in which germinal vesicle breakdown (GVBD) begins when the oocyte contacts seawater (SW); nothing was known about internal events. This study strongly suggests the importance of protein phosphorylation in this process. Herdmania pallida (Pyuridae) functions like H. roretzi; GVBD occurs in SW. Oocytes of Cnemidocarpa irene (Styelidae) do not spontaneously undergo GVBD in SW but must be activated. Herdmania oocytes are inhibited from GVBD by pH 4 SW and subsequently activated by mastoparan (G-protein activator), A23187 (Ca2+ ionophore) or dimethylbenzanthracene (tyrosine kinase activator). This requires maturation promoting factor (MPF) activity; cyclin-dependent kinase inhibitors roscovitine and olomoucine are inhibitory. It also entails dephosphorylation as demonstrated by the ability of the phosphatase inhibitor vitamin K3 to inhibit GVBD. GVBD is also inhibited by the tyrosine kinase inhibitors tyrphostin A23 and genistein, and LY-294002, a phosphatidylinositol-3-kinase inhibitor previously shown to inhibit starfish GVBD. LY-294002 inhibits strongly when activation is by mastoparan or ionophore but not when activated by dimethylbenzanthracene (DMBA). The DMBA is hypothesized to phosphorylate a phosphatase directly or indirectly causing secondary activation, bypassing inhibition.
Development of an egg capable of fertilization involves many synthetic processes. These are supported by the large nucleus termed the germinal vesicle (GV) blocked in the late first prophase of meiosis. In most animals, the GV envelope disappears shortly after ovulation (germinal vesicle breakdown or GVBD). The process of meiosis then proceeds to a stop point to await fertilization and the onset of development. This stop point varies from organism to organism. The eggs of most vertebrates proceed to the second metaphase before fertilization. Starfish eggs resume meiosis when ovulated and the egg completes meiosis and pronucleus formation, but fertilization is possible any time after the first metaphase.
The control of GVBD has been investigated most strenuously in eggs of deuterostomes (echinoderms and chordates). GVBD is initiated by a complex kinase comprised of the cyclin-dependent kinase (CDK) p34cdc2 and its activator cyclin B. This complex is activated by its phosphorylation state. In some organisms, cyclin binds to the kinase at the reinitiation of meiosis and in others the cyclin has already complexed with the kinase before GVBD but remains inactive because of inhibitory phosphate groups. Active p34cdc2/cyclin B induces maturation of the egg and is termed maturation promoting factor (MPF) (Eckberg 1988; Ferrell 1999; Yamashita et al. 2000 for reviews). MPF is activated by a factor produced by follicle cells surrounding the oocyte. In starfish, the factor is 1 methyladenine released in response to a peptide produced by the radial nerves (Borgne & Meijer 1996; Kishimoto 1998, 1999; Borgne et al. 1999). In vertebrates the follicle cells produce a steroid, which varies from group to group (Yamashita 1998; Byskov et al. 1999; Ferrell 1999, Yamashita et al. 2000 for reviews).
Ascidians are deuterostomes and members of the phylum Chordata. They lack a backbone and segmentation but have all the other features of the phylum and are thus classified as invertebrate chordates. They include three orders: Aplousobranchia, Phlebobranchia and Stolidobranchia (Lahille 1886). In contrast to other chordates, which form a polar body prior to fertilization, their eggs block meiosis at the first metaphase until they are fertilized. The aplousobranchs are all colonial with small zooids and internal fertilization and there is little known about their eggs at fertilization. The eggs of phlebobranchs such as Ascidia, Phallusia, and Ciona have been well studied in terms of fertilization and the restarting of meiosis; they store eggs that have completed GVBD in their ovaries or oviducts before spawning. Following fertilization, the eggs rapidly reinitiate meiosis (McDougall et al. 1995) and move on to the mitotic divisions of the zygote under control of MPF (Russo et al. 1996; McDougall & Levasseur 1998). The situation is less well known in stolidobranchs, but one report indicates that eggs are spawned with intact germinal vesicles in Styela canopus (formerly Cynthia partita, Family Styelidae) which breaks down rapidly before fertilization (Conklin 1905). However, Styela gibbsii spawns oocytes in which GVBD has already occurred prior to spawning (Lambert, unpubl. data). In the pyurid Halocynthia, GVBD occurs shortly after ovulation and is accompanied by expansion of the perivitelline space and release of the test cells (Sakairi & Shirai 1991; Fuke & Numakunai 1996; Fuke & Numakunai 1999). Sakairi & Shirai (1991) have suggested that a protease released by the follicle cells may be the trigger of GVBD in Halocynthia. An ovarian inhibitor of meiosis may also be found in this species (Numakunai 2001).
The oocytes of two tropical stolidobranch ascidians Herdmania pallida (Pyuridae) and Cnemidocarpa irene (Styelidae) have been examined here and found to have substantial differences in how the ovulated eggs, stored in the ovary with intact germinal vesicles, respond to sea water (SW). Both species respond similarly to activators and inhibitors of protein phosphorylation, suggesting that the internal changes may be similar, with phosphorylation of a phosphatase being a key event.
Materials and Methods
Animals and oocytes
The large solitary ascidians H. pallida and C. irene were collected from buoys in Apra Harbor, Guam. A few individuals of H. pallida were also collected from buoys, barges and floats in Malakal Harbor, Palau. Following collection, they were held in running SW under dim illumination. For many years, only a single species of Herdmania was recognized, Herdmania momus; however, several species are now accepted (Monniot & Monniot 2001; Nishikawa 2002). The most numerous Herdmania species in Guam is H. pallida, which has a thick, light colored tunic and single sperm duct. Recognizing that nomenclature may change, only a single morphotype was used in the work reported here. The oocytes were 250 µm or greater in diameter excluding extra-ovarian coverings (measurements were taken without a cover slip). They were pigmented dark pink to orange with multivacuolate follicle cells and test cells with numerous red granular inclusions. Individuals were bisected through the siphons and each half was laid out in a Petri dish, the gonads removed and chopped with scissors and the oocytes and sperm separated from debris by filtration through a 330 µm Nytex screen. Sperm was removed by washing, using a 200 µm Nytex screen to retain the oocytes. Herdmania oocytes were collected in pH 4 SW containing 0.05 m sodium citrate as a buffer and thoroughly washed to remove sperm. The Cnemidocarpa oocytes were collected in normal SW pH 8.2.
Oocytes were pre-incubated in specific inhibitors for 1 h or as indicated before meiosis was activated by the appropriate agents or by raising the pH to 8.2. The oocytes were placed on microscope slides in 30 µL drops and gently flattened under a cover slip. This mild flattening allowed visualization of the GV in these fairly opaque oocytes.
All chemicals were purchased from Sigma (St. Louis, MO, USA) except the roscovitine (Biomol, Plymouth Meeting, PA, USA). Stock solutions of A-23187 (3.82 mm), vitamin K3 (10 mm) and dimethylbenzanthracene (DMBA 10 mm) were dissolved in dimethylformamide. Genistein (20 mm), tyrphostin A23 (20 mm), roscovitine (2.7 mm), LY-29004 (10 mm), and olomoucine (20 mm) were dissolved in dimethylsulfoxide. Mastoparan (10 mm) and soybean trypsin inhibitor (10 mg/mL) were prepared in distilled water.
Patterns of GVBD in ascidian oocytes
Ascidian oocytes grow in ovarian follicles surrounded by test cells, a thick vitelline coat and two layers of follicle cells; following ovulation the outer follicle cells are left behind (Burighel & Cloney 1997 for review). The inner follicle cells are subsequently referred to as the follicle cells. Ovaries of H. pallida contain numerous oocytes that have been ovulated and are completely devoid of outer follicle cells. On several occasions, individuals spawned oocytes from the oviduct after dissection. These had all completed GVBD, released the test cells and had an elevated vitelline coat. Freshly dissected oocytes of Herdmania initially have intact GV and a small perivitelline space, and the test cells are tightly associated with the oocyte surface. In clean SW the GV rapidly break down (Fig. 1) and a wide perivitelline space forms, which is soon filled with the released test cells.
This contrasts with another pyurid, Halocynthia roretzi (Sakairi & Shirai 1991), in which nearly all of the oocytes remain enclosed by the outer follicle cells when the ovarian follicles are dissected from the ovary but undergo ovulation and GVBD within 2 h after removal. However, in another study (Fuke & Numakunai 1999), different illumination and temperature regimes caused H. roretzi adults to ovulate oocytes in a high percentage of which GVBD occurred within the ovary before spawning. This issue is complex as there exist three races of H. roretzi, each with a distinct reproductive behavior (Numakunai & Hoshino 1980).
In C. irene (Family Styelidae) ovaries, the oocytes have a wide perivitelline space containing very large multivacuolate test cells and a complete covering of conical inner follicle cells but are completely free of outer follicle cells. The diameter of the uncompressed oocyte is 160 µm. These oocytes have a large GV which maintains its structure for a long period after removal from the ovary. Even storage overnight does not result in spontaneous GVBD, in contrast to H. pallida and H. roretzi in which the oocytes ‘spontaneously’ undergo GVBD after they are removed from the ovary.
Activation of GVBD in ascidian oocytes
Acidic SW at pH 5 that reversibly prevents GVBD in H. roretzi oocytes (Sakairi & Shirai 1991) fails to suppress GVBD in Herdmania oocytes but pH 4 (0.05 m sodium citrate) SW depresses GVBD in Herdmania with the advantage of allowing several chemicals including mastoparan (a G-protein activator), A23187 (a calcium ionophore), and DMBA (a tyrosine kinase activator) (Archuleta et al. 1993) to trigger GVBD (Table 1). Swelling of the perivitelline space or release of the test cells does not accompany this GVBD.
Table 1. Activation and inhibition of GVBD in Herdmania pallida oocytes*
*Oocytes in pH 4 seawater (SW) were activated by mastoparan (0.1 mm), A23187 (0.013 mm) or dimethylbenzanthracene (DMBA) (0.1 mm). The inhibitors used were LY-294002 (0.5 mm), Vit. K3 (0.5 mm) and roscovitine (0.0134 mm). In all cases the percent spontaneous germinal vesicle breakdown (GVBD) (average 20%) has been subtracted from the indicated values before calculating the percentage inhibition. The numbers in parentheses indicate the number of independent trials each using oocytes from a different individual. Negative means indicate that GVBD was less than that which occurred spontaneously in untreated pH 4 oocytes.
41 ± 12 (3)
−2.5 ± 4.5 (2)
−6 ± 3.0 (3)
8.3 ± 13 (3)
31.5 ± 7 (4)
−12 ± 4 (3)
−4 ± 10 (3)
−3.75 ± 6.6 (4)
27 ± 8.2 (4)
24.3 ± 8.2 (3)
−12 ± 5 (3)
−10.75 ± 3.5 (4)
Cnemidocarpa oocytes at pH 8.2 can be induced to undergo GVBD with 0.1 mm of mastoparan (75.5% GVBD, four duplicate experiments), 0.1 mm of DMBA (81.5% GVBD, six duplicate experiments) or 0.032 mm of A23187 (100% GVBD, one duplicate experiment).
Trypsin and GVBD in ascidian oocytes
Trypsin can induce GVBD in pH 5 treated Halocynthia oocytes (Sakairi & Shirai 1991) and in pH 4 Herdmania oocytes (not shown). It fails to induce GVBD in Cnemidocarpa oocytes at pH 8.2 even when used long enough to cause complete removal of the vitelline coat. Similarly, soybean trypsin inhibitor (SBTI) has no effect upon GVBD (data not shown). SBTI blocks GVBD in a dose-dependent manner when pH 4 Herdmania oocytes are raised to pH 8.2 in normal SW (Fig. 2).
The protease inhibitor is a very effective blocking agent in oocytes that have been activated by the increase in pH but without effect on those activated by DMBA. This suggests that DMBA acts on events interior to the cell surface, while the pH shift is working at the cell surface (Fig. 3).
MPF and GVBD in Herdmania oocytes
In oocytes of starfish, amphibians and many other animals, GVBD is initiated by active MPF (Yamashita et al. 2000 for review). Here, we see that inhibition of MPF activity by roscovitine (Meijer et al. 1999) severely blocks GVBD (Table 1). Thus, the ascidian oocyte, like others, controls meiosis with MPF. Blockage by another cyclin-dependent kinase (CDK) inhibitor, olomoucine (Fig. 4), also supports this view.
Phosphatases, kinases and GVBD in ascidian oocytes
The activation state of MPF can be controlled by many mechanisms including the synthesis and binding of cyclin to the CDK or by the dephosphorylation of a pre-existing pre-MPF, in which cyclin is already coupled with the kinase but with several inhibitory phosphate groups attached. Given the short time course of GVBD in ascidian oocytes and the finding that protein synthesis inhibitors do not inhibit GVBD (Sakairi & Shirai 1991) it is probable that they contain pre-MPF. Phosphatases such as Cdc25 are thought to dephosphorylate select phosphate residues activating the CDK (Borgne & Meijer 1996; Borgne et al. 1999). The phosphatase inhibitor vitamin K3 blocks Cdc25 activity and GVBD in starfish oocytes (Borgne & Meijer 1996). It also blocks GVBD in ascidian oocytes with all three activators (Table 1, Fig. 5), supporting the view that there is a phosphatase step in the maturation process that cannot be bypassed by any of the activators. Direct biochemical evidence would be necessary to conclusively prove this point.
LY-294002 inhibits phosphatidylinositol 3 kinase (PI-3 kinase) activity and GVBD in starfish oocytes (Sadler & Ruderman 1998; Nakano et al. 1999). It has a dramatic inhibitory effect on GVBD in ascidian oocytes when activated by mastoparan or A23187 but fails to inhibit activation by DMBA (Table 1, Fig. 5). This implies that a tyrosine phosphorylation is involved in the LY-294002 inhibition and downstream of the steps activated by mastoparan or A23187, but upstream of the step activated by DMBA. In other systems, PI-3 kinase occupies an early step in the pathway (Sadler & Ruderman 1998; Ferrell 1999, Kishimoto 1999). Perhaps this step is accomplished directly by the DMBA bypassing the inhibition. Alternatively, a hypothetical intermediate could be phosphorylated to secondarily activate the phosphatase (Fig. 6).
To test further the possibility of a crucial tyrosine phosphorylation step in the activation process, two tyrosine kinase inhibitors were tested on DMBA- or increased pH-induced GVBD in Herdmania oocytes. Genistein and tyrphostin A23 inhibit different phases of tyrosine kinase activity and inhibit the release of a glycosidase from ascidian follicle cells (Robert et al. 1999). In the present study, both inhibitors block the activation by DMBA (Fig. 4) or increased pH (Table 2). This further supports the involvement of a tyrosine phosphorylation in GVBD. Protein phosphorylation seems to be a general feature of GVBD in all animals studied (Peaucellier et al. 1982; Eckberg 1997 for reviews).
Table 2. Effect of kinase and phosphatase inhibitors on GVBD in Herdmania pallida oocytes experiencing a change in pH from 4 to 8.2*
Oocytes were pre-incubated for an hour at pH 4 in the inhibitors before changing to 8.2 in the presence of the phosphorylation inhibitors. Controls (lacking inhibitors) were transferred to pH 8.2 SW after one h in pH 4 SW. The concentration of inhibitors was 0.2 mm. Means and SEM are shown. Six independent replicate experiments were performed using oocytes from three individuals. Negative means indicate that GVBD was less than that which occurred spontaneously in untreated pH 4 oocytes.
48.4 ± 3.73
−8.3 ± 8.2
11.4 ± 8.8
−23.5 ± 6.3
−2.6 ± 5.0
These results show that genistein and tyrphostin A23 both inhibit GVBD even when not activated by the tyrosine kinase activator DMBA. They also show that activation of PI-3 kinase and a phosphatase such as Cdc25 are essential events in ascidian GVBD.
Independence of GVBD and surface changes during maturation of Herdmania oocytes
The Herdmania oocytes that were induced to undergo GVBD by an increase in pH or trypsin also showed vitelline coat swelling to produce a large perivitelline space and released the test cells into the space. Those oocytes induced to undergo GVBD in pH 4 SW by mastoparan, ionophore A23187 or DMBA never formed the perivitelline space or released the test cells. The independence of GVBD from vitelline coat swelling and test cell release was also noticed in Halocynthia oocytes (Sakairi & Shirai 1991; Numakunai 2001). In C. irene the perivitelline space was already formed in freshly dissected oocytes.
There are two distinct patterns in activation of ascidian oocytes from two different stolidobranch families. Oocytes from the pyurid H. pallida undergo GVBD spontaneously when removed from the ovary. This agrees with data from another pyurid, H. roretzi (Sakairi & Shirai 1991). In contrast, oocytes from the styelid C. irene removed from the ovary do not initiate meiosis unless stimulated by artificial means. Starfish also exhibit two disparate pathways; Asterias amurensis oocytes undergo ‘spontaneous’ GVBD when the oocytes with follicle cells are suspended in SW but Asterina pectinifera oocytes do not (Shirai & Walker 1988). Many more ascidian species must be investigated to see how widespread these differences are.
Maturation of ascidian oocytes involves GVBD and two other events that may be under the same or different controls: vitelline envelope swelling and release of the test cells into the perivitelline space. Sakairi & Shirai (1991) reported that Halocynthia oocytes inhibited from GVBD in pH 5 SW, but induced to undergo GVBD with trypsin treatment in the acidic SW, would do so without the usual swelling of the perivitelline space or release of the test cells. Numakunai (2001) confirmed this finding. Herdmania oocytes stimulated to undergo GVBD in pH 4 SW by ionophore, mastoparan or DMBA fail to release the follicle cells or swell the vitelline coat; however, pH 4 oocytes stimulated with trypsin do swell the vitelline coat. Careful examination of the oocytes disclosed a thin membrane covering the test cells, which must rise or disperse to release these cells. In the few instances of spontaneous spawning from the oviduct in dissected Herdmania, the GV had already broken down and there was a substantial perivitelline space enclosing numerous free test cells.
Oocytes of Cnemidocarpa (Styelidae) already have a substantial perivitelline space filled with test cells when removed from the ovary. As stated above, this space forms coincident with GVBD in the pyurids. In the phlebobranch Phallusia, the perivitelline space arises over an hour after soaking dissected oviducal eggs in pH 8.2 SW (Honegger 1986; Eisenhut & Honegger 1997). In Ascidia (another phlebobranch), swelling of the vitelline coat is pH-dependent (Koch & Lambert, unpubl. data).
Both the pyurid and styelid oocytes responded to the G-protein activator mastoparan and the tyrosine kinase agonist DMBA by initiating meiosis. This implies that maturation involves a G-protein activation and protein phosphorylation within the oocyte, though direct biochemical evidence is lacking. That mastoparan stimulates maturation suggests that either the release of an αi/αo or release of the βγ subunits is responsible. That release of an αi/αo subunit stimulates GVBD in ascidian oocytes agrees with a similar finding in starfish oocytes (Kalinowski et al. 2003). Injection of free βγ subunits also causes GVBD in starfish oocytes (Chiba et al. 1993; Jaffe et al. 1993; Nakano et al. 1999). In nemertean oocytes mastoparan neither stimulates nor inhibits GVBD but melittin, an inhibitor of Gs, blocks GVBD in response to serotonin (Stricker & Smythe 2001). However, the cell surface events in ascidians are not known. Trypsin could be the natural maturation-inducing substance (MIS) in some pyurids such as Halocynthia or Herdmania, but it definitely is not in the styelid tested in these experiments. A protease also stimulates GVBD in polychaete oocytes (Peaucellier 1977). The follicle cells are the source of the MIS in starfish and vertebrates, and Halocynthia oocytes deprived of their follicle cells fail to undergo GVBD (Sakairi & Shirai 1991). Ascidian follicle cells have also been implicated in enzyme release (Lambert et al. 1997). Thus, it is very likely that the follicle cells are the source of the MIS. Possibly it is the protease suggested by Sakairi & Shirai (1991). However, data from this paper and that of Sakairi & Shirai (1991) also support the hypothesis that the protease may activate another molecule, which is the true MIS. Alternatively, it could modify cell surface receptor molecules enabling the actual MIS to activate. Protease activity activates the onset of self-sterility in Halocynthia (Fuke & Numakunai 1999) and Ciona (Marino et al. 1999) oocytes without itself being the responsible molecule. Other possible candidates for MIS in ascidians include insulin and insulin-like growth factor, which is present in follicle cells (Reinecke et al. 1999) and is a known MIS in vertebrate oocytes (Adashi 1994). Another possibility is an inhibitory substance in the ovary, which would be diluted in spawning (Numakunai 2001). Much more work is required to clarify the events at the surface of the oocyte and the proximal portion of the pathway. The different ascidian families may have different means for accomplishing the same tasks.
More is known about activation events within the oocyte. Both Herdmania and Cnemidocarpa oocytes are induced to undergo GVBD by the G-protein activator mastoparan. This suggests that they are like all other known oocytes in which the MIS receptor activates a trimeric G protein early in the initiation process (Nakano et al. 1999 for review). They both are also activated by DMBA, a tyrosine kinase agonist. That tyrosine phosphorylation is activated by DMBA and crucial for GVBD is supported by the inhibition of GVBD by genistein and tyrphostin A23 when the oocytes are activated by DMBA. The experiments in which the same tyrosine kinase inhibitors also inhibit GVBD induced by the pH shift from 4 to 8.2 imply that phosphorylation is a part of the natural activation process. Ascidian blood maintains an acidic pH (Michibata et al. 1991); it is feasible that the pH of the oviduct is higher than that of the ovarian stroma. Thus, the low pH of the ovarian stroma could be mimicked by low pH SW to explain its inhibitory effect on maturation of Herdmania and Halocynthia oocytes. The Ca2+ ionophore A23187 also induces GVBD in ascidian oocytes (this study; Sakairi & Shirai 1991) as it does in the oocytes of many animals including both deuterostomes and protostomes (Duesbery & Masui 1996 for review). Although the connection between intracellular Ca2+, G proteins and protein phosphorylation has not been investigated in ascidians, the similarity in response to inhibitors exhibited by mastoparan and A23187 activation suggests that Ca2+ may be involved early in the pathway. However, a recent report shows that an increase in intracellular Ca2+ inhibits the onset of GVBD in Xenopus oocytes (Sun & Machaca 2004). In addition, A23187 causes an increase in intracellular pH in bivalve mollusc oocytes (Deguchi & Osani 1994). Furthermore, intracellular pH has a profound role in the control of GVBD in several systems (Lee & Steinhardt 1981; Dube & Guerrier 1982; Dube 1988; Deguchi & Osani 1994) and is known to activate cdc-25 (Dunphy & Kumagi 1991); it could have the same effect here (Fig. 6).
Alternatively, A23187 could artifactually cause GVBD in ascidian oocytes as it does in those from polychaetes (Eckberg 1997). The fact that pH 4 oocytes of H. pallida undergo GVBD when appropriately stimulated demonstrates that it is the control of GVBD rather than the process itself, which is pH dependent.
The PI-3 kinase inhibitor LY-294002 inhibits G-protein-coupled activation of GVBD in starfish oocytes (Sadler & Ruderman 1998; Nakano et al. 1999). Activation of PI-3 kinase in starfish oocytes is stimulated by free βγ subunits of a G protein (Sadler & Ruderman 1998). The step catalyzed by PI-3 kinase in ascidian oocytes apparently involves tyrosine phosphorylation, as LY-294002 inhibits GVBD when induction is by A23187 or mastoparan but not DMBA. Possibly the DMBA directly phosphorylates cdc-25 or some intermediate of the pathway resulting in a secondary phosphorylation (Fig. 6).
I thank Gustav Paulay (University of Guam) and Patrick and Lori Colin (Coral Reef Research Foundation, Koror, Palau) for providing research facilities. My thanks also go to Gretchen Lambert, Gustav Paulay, Lisa Kirkendale, Patrick Colin, Chris Meyer and John Stamner for collecting all of the animals used in these studies. I am deeply indebted to Gretchen Lambert for identification of the ascidians and a critical reading of the manuscript. The manuscript was also improved by the constructive criticism of two anonymous reviewers. We are both grateful to Claude Monniot for verifying the identity of H. pallida. This study was supported in part by a grant from Cal ERFA and travel funds from Sea Grant to Gustav Paulay and the National Cancer Institute to Patrick Colin.