Starfish, Asterina pectinifera (cover picture), is easily found in shallow waters of the Pacific and the Japan Sea. The pretty five armed shape is familiar to almost everyone. Fishermen, however, hate the animal because it attacks and consumes scallops or clams.
In the breeding season (from April to December, depending on the area), starfish have several million fully grown oocytes. At this stage, oocytes are “immature” and have a large nucleus called a germinal vesicle (GV). They are arrested in prophase of meiosis I. Meiosis is reinitiated by the hormone, 1-methyladenine (1-MA), which is released from surrounding follicle cells (Kanatani et al., 1969). Germinal vesicle breakdown (GVBD) is the first easily observable event in meiosis reinitiation after hormonal stimulation. After GVBD, oocytes are called “mature”. Then, they are spawned and fertilization occurs in the seawater.
Starfish are ideal animals for the study of meiosis reinitiation. They are easy to keep in marine aquaria. Also, oocyte maturation is easily induced by experimental application of 1-MA to the isolated ovary or oocytes. Usually synchronous GVBD of Asterina pectinifera occurs within 30 min after 1-MA addition, in which period protein synthesis is not required. The first and second meiotic cycles are completed without arrest of meiosis. In this review, I would like to summarize the pathway of 1-MA signal transduction.
Properties of 1-MA receptors
Intracellularly injected 1-MA does not cause GVBD (Kanatani and Hiramoto, 1970), suggesting that 1-MA receptors exist on the plasma membrane. Indeed, 1-MA specifically binds to the oocyte surface (Yoshikuni et al., 1988).
The surface receptors coupling to guanosine nucleotide binding (G) proteins have two different affinities for the agonists (Gilman, 1987). The high-affinity receptors are converted to the low-affinity ones by GTP or GTPγS which can activate the G protein. To determine whether 1-MA receptors couple to G proteins, we measured the specific binding of 3H-radiolabeled 1-MA to a oocyte membrane with or without GTPγS. As expected, the binding activity decreased, when GTPγS was added. Scatchard plot analysis showed that there were two affinities for 1-MA binding with apparent Kd's of approximately 30 nM and more than 1 μM. The lower value of Kd appeared to be reasonably high as a character of G protein-coupled membrane receptors for the agonist. The high Kd value of micromolar range was consistent with the value reported by Yoshikuni et al. (1988). In the presence of GTPγS, the high-affinity fraction was completely abolished. The conversion of the binding activity from high-affinity to low-affinity was not induced by ATP or its analogue (Tadenuma et al., 1992). These results suggest that starfish oocyte membranes have 1-MA receptors coupling to G proteins. The next step of this study is to identify the receptor.
PTX-sensitive G protein in starfish oocytes
G proteins mediate signal transduction from a receptor to an effector enzyme in many systems. In mammals, pertussis toxin (PTX) ADP-ribosylates the α subunit of Gi-type G protein and prevents the α subunits from accepting the signal of the receptor (Gilman, 1987; Ui and Katada, 1990). When we injected PTX into starfish oocytes, 1-MA-induced GVBD was completely inhibited (Shilling et al., 1989). These results suggest that 1-MA receptors interact with PTX-sensitive G proteins. To demonstrate whether PTX ADP-ribosylates the G protein in vivo, [32P]NAD and PTX were injected into immature oocytes. Then, the oocytes were analyzed by SDS-PAGE and autoradiography. As shown in Fig. 1, a 39-kDa protein was radiolabeled. When [32P]NAD without PTX was injected, the band was not found (Fig. 1). Thus, it is likely that microinjected PTX ADP-ribosylated the 39-kDa G protein α subunit in the oocytes. Indeed, the purified G protein from oocyte plasma membranes had an αβγ-trimeric structure consisting of 39-kDa α, 37-kDa β, and 8-kDa γ subunits. During purification, a PTX-substrate G protein (39-kDa α) was always recovered as a single peak, indicating that there is the only one class of PTX substrate in starfish oocyte membranes, in contrast to mammalian G protein having three distinct forms of α subunit, Gi-1, Gi-2 and Gi-3 (Tadenuma et al., 1991).
The α subunit of the starfish G protein was purified and digested with trypsin. The resulting peptides were fractionated by HPLC and purified peptides were microsequenced, which revealed their high degree of identity with mammalian Gi-α. Thus we screened a cDNA library constructed from the starfish ovary with rat Giα cDNA. A positive cDNA clone contained an open reading frame of 1,062 bases. The deduced amino acid sequence contained microsequences, indicating that the isolated cDNA clone encodes the α subunit of the PTX-sensitive G protein in oocytes (Chiba et al., 1992). Identity of starfish Gα was 89% with rat Gi-1α, 85% with rat Gi-2α, 86% with rat Gi-3α, and 72% with rat Goα. Similarly, high identity of starfish Gβ with mammalian one was suggested, since the 37-kDa β subunit strongly reacted to an antibody against mammalian 36-/35-kDa β subunits (Tadenuma et al., 1991).
Usually, cholera toxin (CTX) cannot ADP-ribosylate the PTX-substrate G proteins. However, when the agonist stimulates the receptor and G proteins, CTX can ADP-ribosylate the PTX-sensitive G proteins (Iiri et al., 1991). Thus, coupling of the PTX-sensitive G protein with the receptor can be demonstrated using CTX-dependent ADP-ribosylation. As expected, the starfish 39-kDa G protein became a CTX substrate in the presence of 1-MA. We thus concluded that the G protein in starfish oocyte couples to the 1-MA receptors (Tadenuma et al., 1992).
PTX did not block dithiothreitol-induced maturation
Dithiothreitol is known to induce GVBD of starfish oocytes (Kishimoto and Kanatai 1973), presumably due to its ability to reduce disulfide bonds (Kishimoto et al., 1976: Mita et al., 1987). Dithiothreitol-treated oocytes, as well as the 1-MA-treated, form a maturation-promoting factor (MPF: see below) in the cytoplasm (Kishimoto et al., 1976), yet the primary target of dithiothreitol is an open question. If a target molecule of dithiothreitol is similar to that of 1-MA, PTX should inhibit dithiothreitol-induced maturation as well. However, dithiothreitol induced GVBD of oocytes preinjected with PTX. It is therefore concluded that dithiothreitol acts on a downstream pathway of G protein (Chiba et al., 1992).
Induction of GVBD by microinjected starfish G protein βγ-subunit
The binding of 1-MA to the receptor should induce dissociation of Gα and Gβγ. Thus, 1-MA-induced GVBD is likely to be mediated by released Gα and/or Gβγ. Jaffe et al. (1993) found that mammalian Gβγ purified from brain or retina induced GVBD when they were microinjected into starfish oocytes. Similarly, when we purified starfish Gβγ from oocyte plasma membrane and microinjected it into starfish oocytes, GVBD was occurred (Chiba et al., 1993). Starfish Gβγ seems to be more effective than mammalian one, since it induced GVBD faster. Oocyte maturation induced by starfish Gβγ was quite similar to that induced by 1-MA: MPF or high activity of cdc2 kinase were found in the cytoplasm of Gβγ-injected oocytes. Maturing oocytes formed fertilization envelope after the penetration of a spermatozoon, expelled polar bodies, and proceeded to cleavage. These results indicate that βγ-injection mimicked 1-MA treatment in both cortical maturation (Meijer and Guerrier, 1984; Chiba and Hoshi, 1989; Chiba et al., 1990) and nuclear maturation. We concluded that 1-MA induces GVBD through an action of dissociated Gβγ from Gα.
Localization G protein βγ subunit in starfish oocytes
We raised the monoclonal antibody which cross-reacts with purified β subunit from starfish oocytes. The antibody recognized a single band in Western blots of immature and mature oocyte lysate, indicating that it is specific for β subunit.
When isolated plasma membranes of immature and mature oocytes were labeled with the antibody, there was faint punctate staining of β subunit throughout the plasma membrane. Thus, as had been expected, G proteins interacting with the 1-MA receptor exist on the plasma membrane (Chiba et al., 1995). To our surprise, whole mount preparation of immature oocytes labeled with the antibody revealed a reticular network throughout the cytoplasm. The fiber exhibited complex striation with a periodicity of 0.7 μm. The thickness of the fibers were about 0.4 μm. The identical staining patterns were obtained, when the oocyte was stained with anti-γ subunit antibodies. Similar networks of filaments in immature oocytes of sea urchin (Boyle and Ernst, 1989) as well as starfish (Schroeder and Otto, 1991) were stained with antibodies against cytokeratin. Indeed, when we double-stained the oocytes with anti-Gβ and anti-cytokeratin antibodies, the networks labeled with both antibodies shared a common feature of looping and branching. However, the pattern of stained striation alternated. These results suggest that G protein βγ sub-unit distributes segmentally on the continuous cytokeratin filaments (Chiba et al., 1995). The role of the cytoplasmic Gβγ remains unclear.
PI3K and protein phosphorylation
In mammalian cells, Gβγ directly activates some members of the phosphatidylinositol 3-kinase (PI3K) family (Kurosu et al., 1995, 1997; Leopoldt et al., 1998; Stephens et al., 1997; Stoyanov et al., 1995; Thomason et al., 1994). PI3K phosphorylates inositides at the D-3 position of the inositol ring to generate such lipid messengers as phosphatidylinositol 3,4-bisphosphate and phosphatidylinositol 3,4,5-triphosphate. Sadler and Ruderman (1998) reported that 1-MA-induced oocyte maturation in starfish Asterina miniata is blocked by the PI3K inhibitors, wortmannin and LY294002, suggesting that PI3K acts downstream of the 1-MA receptor. We showed that these lipid kinase inhibitors blocked Gβγ-induced oocyte maturation in starfish Asterina pectinifera (Nakano et al., 1999). These results support the hypothesis that PI3K is the target of the Gβγ. Cytoplasmic distribution of the subsequent effector for Gβγ (Chiba et al., 1993) is also consistent with the above hypothesis, since PI3K is demonstrated to exist in the cytoplasm (Leopoldt et al., 1998).
To investigate targets for PI3K, we examined endogenous protein phosphorylation (Nakano et al., 1999). When starfish oocytes were microinjected with [γ-32P]ATP, the significant enhancement of phosphorylation of a 62-kDa protein from 1-MA-treated oocytes was observed. The stimulation of the phosphorylation was a very early event in the 1-MA signaling pathway, since phosphorylation of the 62-kDa protein started within the first 4 min of 1-MA stimulation. In the cell-free preparations (Chiba et al., 1999), the 62-kDa protein was also phosphorylated on serine residue(s) immediately after addition of 1-MA or Gβγ (Nakano et al., 1999). If the PI3K is an essential component of the pathway linking the Gβγ to the serine kinase activation, inhibitors of the lipid kinase should block the phosphorylation of the 62-kDa protein. As expected, when we added Gβγ to the immature oocyte supernatant preincubated with wortmannin or LY294002, the phosphorylation of the 62-kDa protein was inhibited. These results indicate that the 62-kDa protein functions downstream of Gβγ and PI3K in the early signaling pathway of 1-MA-induced starfish oocyte maturation. Although it is much more likely that PI3K activation sets off a signal cascade that includes the protein kinase, there is a possibility that the PI3K has a protein kinase activity, since some PI3K family members have protein kinase activity (for review see Hunter, 1995). In particular, DNA-dependent protein kinase (DNA-PK), which lacks detectable lipid kinase activity, is inhibited by wortmannin or LY294002 (Christodoulopoulos et al., 1998; Gu et al., 1998). If this is the case, the 62-kDa protein may be a direct target of the PI3K.
MPF activity of maturing oocytes
MPF is activated in the cytoplasm about 20 min after the phosphorylation of 62- kDa protein. MPF, which is composed of cdc2 protein kinase and cyclin B, not only induces starfish oocyte maturation, but also regulates G2- to M phase transition of mitosis from yeast to mammalian cells (Kishimoto et al., 1982; Tachibana et al., 1987; Lee, M. G. and Nurse, P., 1987). The cdc2 kinase in starfish oocytes is activated by dephosphorylation of cdc2 kinase/cyclin complex, which is mediated by cdc25 protein phosphatase (for review see Kishimoto, 1999). Interestingly in starfish oocytes, Okumura et al. (1996) indicated that cdc25 phosphatase (90-kDa) is activated by an unknown kinase called “initial kinase”. The serine kinase activated by PI3K is likely to be involved in the activation of the “initial kinase”. Or the phosphorylation of the 62-kDa protein may be required for the activation of the kinase. Studies on the identification and characterization of the 62-kDa protein and the serine kinase are now in progress.
In summary as shown in Fig. 2, the receptor of 1-MA on the plasma membrane of starfish oocytes couples to the αβγ trimeric G protein. Microinjected pertussis toxin (PTX) catalyzes ADP-ribosylation (ADPR) of Gα. This modification prevents the interaction between the receptor and G protein, and blocks the 1-MA signal transduction. The hormonal stimulation dissociates Gβγ from GTP-bound Gα, and the dissociated Gβγ activates PI3 kinase (PI3K).The PI3K participates in the activation of the protein kinase that phosphorylates the serine residue of the 62-kDa protein, which results in the activation of MPF (cdc2 kinase/cyclin B). MPF eventually induces GVBD.
This study was supported in part by grants from the Ministry of Education, Culture, Sports and Sciences of Japan.