Meiosis reinitiation in the mollusc Patella vulgata. Regulation of MPF, CSF and chromosome condensation activity by intracellular pH, protein synthesis and phosphorylation

GV prophase-arrested oocytes of most animals are triggered to resume meiotic divisions in response to various stimuli (Masui & Clarke, 1979; Maller & Krebs, 1980; Guerrier et al. 1982; Meijer & Guerrier, 1984). In the prosobranch mollusc Patella vulgata, it has been shown that meiosis reinitiation proceeds in two steps and is regulated by ionic factors (Guerrier et al. 1986a). The first step, which leads to GVBD, chromosome condensation and formation of the metaphase-1 spindle, involves an increase in the intracellular pH. The second step, which releases the oocyte from the metaphase-1 block depends directly upon a rise in intracellular calcium. Although, at present, we ignore how these signals act intracellularly, more generally it has been established by cell fusion and heterologous transfers of cytoplasm that a universal nonspecies-specific factor, known as M-phase or maturation promoting factor (MPF), controls both the release of oocytes from the prophase block and the progression of somatic cells from G₂ to mitosis (Masui & Markert, 1971; Reynhout & Smith, 1974; Balakier & Czolowska, 1977; Balakier, 1978; Wasserman & Smith, 1978; Sunkara et al. 1979; Nelkin et al. 1980; Tarkowski & Balakier, 1980; Weintraub et al. 1982; Kishimoto et al. 1982, 1984; Schalt et al. 1983; Gerhart et al. 1984). It is generally agreed that this factor, which promotes GVBD, nuclear envelope disruption and chromosome condensation, may be a protein kinase or a phosphoprotein, if not a phosphoprotein phosphatase inhibitor (Pondaven & Meijer, 1986). Indeed, MPF activity appears always to be associated with a high level of endogenous protein phosphorylation, as observed during meiosis reinitiation of the amphibians (Maller et al. 1977), mammals (Wassarman et al. 1979; Crosby et al. 1984), starfishes (Guerrier et al. 1977, Dorée et al. 1983; Guerrier et al. 1988), echiuroids (Meijer et al. 1982), polychaetes (Peaucellier et al. 1981, 1982) and bivalves (Dubé et al. 1987; Eckberg et al. 1987). This correlation between phosphorylation and MPF activity has been observed also after activation of the amphibian oocyte (Capony et al. 1986), during early synchronous cleavage of the sea urchin (Schatt et al. 1983) or division of somatic mammalian cells (Westwood et al. 1985).

In addition to MPF, a second factor, CSF, has been described also, which arrests the cell cycle in metaphase, preserves MPF activity (Gurdon, 1968; Masui & Markert, 1971; Masui, 1974, 1985; Masui et al. 1980; Ryabova, 1983; Gerhart et al. 1984) but is sensitive to protein synthesis inhibitors (Clarke & Masui, 1983, 1985; Newport & Kirschner, 1984) or to the injection of lysolecithin-treated sperm nuclei (Shibuya & Masui, 1982).

In amphibians, the appearance of MPF prior to GVBD requires the synthesis of a protein initiator, while its autocatalytic amplification following cytoplasm microinjection into a recipient unstimulated GV-blocked oocyte does not depend on protein synthesis (Wasserman & Masui, 1975; Schuetz & Samson, 1979; Gerhart et al. 1984). The same situation prevails for ovine oocytes (Osborn & Moor, 1983; Moor & Crosby, 1986), whereas no new transcription or protein synthesis is required for triggering GVBD and chromosome condensation in the bivalve Spisula (Swenson et al. 1986), the mouse (Jagiello, 1969; Stern et al. 1972; Golbus & Stein, 1976; Balak-ier & Czolowska, 1977; Crozet & Szollosi, 1980) and the starfish (Guerrier & Dorée, 1975; Dorée, 1982). This implies that oocytes of these animals contain a stockpile of MPF precursor molecules which are only activated at the posttranslational level. However, after the first meiotic cycle, as after each mitotic cycle, protein synthesis inhibitors have been shown to block subsequent cleavage (Wilt et al. 1967, Wagenaar & Mazia, 1978; Westwood & Wagenaar, 1983) and to preclude MPF reappearance (Gerhart et al. 1984; Picard et al. 1985). Such an absolute requirement for protein synthesis suggests a potential role for those proteins, known as cyclins, which are destroyed at the metaphase-anaphase transition and must be restored at each cell cycle (Evans et al. 1983; Standart et al. 1987). Recently, it has even been shown that a mRNA from Spisula, coding for cyclin A, could reinitiate meiosis of GV-arrested Xenopus oocytes, even when cyclin A was not present or needed to trigger GVBD in Spisula (Swenson et al. 1986). Obviously, more information is required to understand better the metabolic and ionic conditions that are likely to regulate those cytoplasmic factors responsible for driving on the cell cycle. We therefore decided to study this question further, using the oocyte of Patella vulgata as a model system which may contain both MPF and CSF activities.

In this paper, we show that MPF activation and amplification in Patella are pH-dependent but protein-synthesis-independent events that actually result in an increased phosphorylation of endogenous proteins. Instead, cytostatic factor and chromosome condensation activity seem to depend on a continuous supply of new proteins which may be required to protect MPF against dephosphorylation or to activate relevant protein kinase.

Adult limpets, collected from September to March in the vicinity of Roscoff, were maintained at 10°C under a shallow flow of running sea water. Prophase-blocked oocytes were obtained by cutting and agitating the gonad either in filtered (FSW) or in artificial sea water (ASW) prepared according to MBL formula (Shapiro, 1941; Cavanaugh, 1975) and adjusted to pH 8·2–8·5 following addition of 6mm-Tris. They were washed repeatedly, filtered through cheese cloth and resuspended in FSW. Under these conditions, only a few batches of oocytes (2–5 %) had to be discarded for undergoing spontaneous maturation. All the experiments reported here were performed on follicle-enclosed oocyte populations which did not present a percentage of spontaneous maturation higher than 5%. Meiosis reinitiation was triggered by adding to the oocyte suspensions 5–20 mm-NH₄Cl from a 2 M-stock solution containing lümm-Tris, pH 8·5. Maturation also occurred up to metaphase I in some batches of oocytes which were shaken for 30s in a 10ml glass tube covered with Parafilm (Guerrier et al. 1986a).

Intracellular microinjections of cytoplasm were carried out according to the method Hiramoto (1974), as precisely described by Meijer et al. (1984) and Kishimoto (1986). The tip of the micropipette was broken to a diameter wider than 5 μm in order to allow free motion of the huge cytoplasmic inclusions from Patella. Recipient oocytes were suspended in 80% sea water.

For incorporation studies, the cultures were continuously stirred at 60 revs min^-1 with a glass paddle.

For continuous or preloading experiments [¹⁴Cjarginine (140mCimM^-1) or [³H]leucine (50CimM^-1) from CEA (France) were added to an oocyte suspension containing about 10⁴ oocytes ml^-1. Pulse experiments were performed using 1 ml samples. Sampling was performed at the desired times by centrifuging 1 ml of the oocyte suspension through a preocooled isotonic 1· 1 m-sucrose/ASW cushion (30–70 %). The eggs were then washed four times with cold ASW and the fina] pellet resuspended into 1 ml NaOH, 0·5 N. The same washing procedure was used for preloading experiments to remove the remaining amino acids. In this case, manipulations were done at room temperature. Total uptake was determined by counting 50 μl of the NaOH solution diluted with distilled water up to 0·5 ml in 8 ml ACS liquid scintillant (Amersham). For measuring incorporation into proteins, remaining samples were reprecipitated overnight with t.c.a., washed again twice with t.c.a. and redissolved in 0·5 ml NaOH. Protein concentration was determined using the method of Lowry et al. (1951). All determinations were made in duplicates. Final results arc expressed in ctsmin^-1 mg^-1 protein.

Oocytes were pre loaded with 10 μCi ml^-1 carrier-free ³²PO₄ (aqueous solution, PBS 13, Amersham) for 4 h, washed free of external label and incubated in various conditions. 1 ml samples were injected in 4 ml ice-cold 10% t.c.a.. washed twice in the same medium, dissolved in 1 ml NaOH 0·5 N and reprecipitated overnight. The final pellet, washed again with 5ml t.c.a., was dissolved in NaOH 0·5N. Protein concentration was determined and cts min^-1 recorded from 0·5 ml samples suspended in 8 ml of ACS. All experiments were performed at the constant temperature of 20°C.

For electrophoresis and autoradiography, oocytes (10⁴ cellsml^-1) were incubated with 5μptCi ml^-1 [³⁵S]methionine (SJ 1515, Amersham). 0·2 ml samples were withdrawn from the culture, spun down in a microfuge, washed and dissolved into 0·4 ml of the SDS gel sample buffer, which contained 60mm-Tris-HCl, pH 6·8, 2% SDS, 5% β-mer-captoethanol, 10 % glycerol and a drop of bromophenol blue. Samples were vortexed and heated for 5 min in boiling water. They were deposited on top of 15% or 10–15% linear gradient SDS-PAGE according to Laemmli (1970). Gels were stained with Coomassie blue R 250 (0·25 % in 50% (v/v) methanol and 10% (v/v) acetic acid or t.c.a., destained in 40 % (v/v) methanol and 7 % (v/v) acetic acid. After drying in a Hoefer Scientific Instruments slab gel dryer, gels were processed for autoradiography, using Kodak X-OM AR film or Amersham hyperfilm β-max, in the presence of an intensifying screen.

GVBD, formation of the metaphase spindle or resting nuclei were easily observed in vivo after the eggs were mounted and flattened by removing excess fluid at the edge of the coverslip. These observations were routinely confirmed after the oocytes had been fixed for 10 min in Carnoy’s fluid, stained in some drops of acetic carmine and glycerol and mounted between slide and coverslip. For ultrastructural observations, specimens were fixed in 4 % glutaraldehyde in 0·1 M-cacodyiate pH7·4 for 1h at (0°C, washed overnight in 0·2 M-cacodylate, 2 % NaCl and postfixed 1 h at 0°C in 1 % osmium tetroxide prepared in 0·1 M-cacodylate, 2·5 % NaCl. Dehydrated samples were embedded in Spurt’s resin. Ultrathin sections were stained with 0·5% uranyl acetate and 2% lead citrate. Observations were performed on a JEOL JEM HX) CX transmission electron microscope at 80 kV.

All chemicals were obtained from Sigma, except 2,4-dinitrophenol (DNP) which was purchased from Prolabo and prepared in 10mm-Tris, pH 8·2.

When exposed to 5–10mm-NH₄Cl in SW, pH 8·2, GV-arrested oocytes of Patella vulgata undergo GVBD within 7min. They further condense their chromosomes and form a stable metaphase-1 spindle by 30–45 min (Fig. 1A), while the follicle cells detach from the plasma membrane and gather in a small clump at the animal pole (Guerrier et al. 1986a). Since all these features are usually related to MPF appearance, at least in the starfish (Kishimoto et al. 1984), we decided to check for the presence of this factor by injecting cytoplasm from such maturing oocytes into recipient GV-arrested oocytes. Unfortunately, these experiments, which involved the injection of up to 180 pl of cytoplasm (6 % of the oocyte volume) and were performed on more than 10 different batches of oocytes, always failed to produce GVBD in the recipient oocytes. This could signify either that a too small amount of MPF was injected or, at least, that MPF amplification could not occur at the original low intracellular pH that characterized the unstimulated Patella oocyte. To answer this question, we performed heterologous microinjection experiments, taking as recipients GV-arrested starfish oocytes, which are known to exhibit a higher intracellular pH, close to neutrality (Moody & Hagiwara, 1982; Picard & Dorée, 1983). Even when only a few experiments of this type could be performed, due to a rather bad overlap of the breeding seasons for both Patella vulgata and Asterias rubens, these clearly confirmed that Patella maturing oocytes did actually produce amplifiable active MPF. Indeed, in three independent assays, which involved both donor and recipient oocytes issued from different animals, we found that the microinjection of about 100pl of Patella maturing cytoplasm was sufficient to produce GVBD in the recipient starfish oocytes. Similar results were also obtained after microinjection of cytoplasm taken from a batch of G V-disrupted shaken oocytes that proved able to reach the metaphase-1 stage in the absence of previous ammonia stimulation (Guerrier et al. 1986). In any case, it was found (1) that GVBD always occurred between 17 and 21 min after microinjection of the Patella donor cytoplasm, i.e. according to the same kinetics as observed following microinjection of starfish MPF (Guerrier & Néant, 1986a); (2) that the metaphase-1-blocked oocyte preserved its MPF activity for at least 2·5 h following GVBD; (3) that cytoplasm from GV-blocked oocytes of Patella never produced maturation of the recipient starfish oocytes.

NH₄Cl-induced meiosis reinitiation was not inhibited when the oocytes had been previously treated for more than 2h with 200 μM-forskolin, a drug that has been shown to stimulate adenylate cyclase and to delay or inhibit maturation in amphibian (Schorderet-Slatkin & Baulieu, 1982), mammalian (Eppig et al. 1983; Schultz et al. 1983) and starfish oocytes (Meijer & Zarutskie, 1987). The same absence of effect was observed following 30 min incubation in the presence of either 1 mm-dibutyryl-cyclic AMP (dbc-AMP), cyclic GMP (c-GMP) or the phosphodiesterase inhibitor 3-isobutyl-l-methyl-xanthine (IBMX) used separately or in association. GVBD and spindle formation also occurred normally, following incubation of the oocytes with 20 μg ml^-1 actinomycin D or with 90 μM-emetine.

In contrast, ammonia-induced maturation was inhibited by the uncoupler DNP and the puromycin analogue 6-dimethylaminopurine (6-DMAP) which blocked GVBD at 200–500 μM and 1·5–2 MM, respectively. Al lower concentrations, these drugs only delayed GVBD which, for example, took some 30min to occur in the presence of 120 μM-DNP or 1 MM-6-DMAP. Under these conditions, the metaphase-1 spindles formed normally and the chromosomes condensed, giving rise to nine typical tetrads which, however, did not persist as long as observed in control oocytes. Indeed, while cyclic nucleotides had no specific cytological effect upon oocyte maturation, except that IBMX alone could block follicle cell detachment, all the other treatments listed were found to severely affect the mitotic apparatus and chromosome structure. Thus, both 6-DMAP (600 μM) and emetine (75 μM) rapidly induce disappearance of the metaphase-1 spindle. Within 40 min, the chromosomes disperse and decondense to produce a set of resting nuclei which exhibits well-formed nucleoli and a typical nuclear envelope with nuclear pores (Fig. 1A–F). This process is only delayed when lower concentrations of emetine (18μM) or 6-DMAP (150 μm) are used. In any case, it does not result from activation and true mitosis since no spindle cycle is observed under these conditions. Moreover, the number of nuclei that appear simultaneously in the treated oocytes is never in great excess over the number of tetrads originally present in the metaphase plate. It is worth noting that this evolution is also triggered when emetine or 6-DMAP are only applied 1 h after ammonia stimulation when the metaphase-1 spindle is already attached to the animal pole cortex. The effects of emetine and 6-DMAP only differ in one respect, their reversibility. Indeed, whereas emetine-induced nuclei persist after washing out the inhibitor, in 6-DMAP-depleted oocytes new asters are rapidly produced, nuclei break down and the chromosomes recondense (Fig. 2A–D). This rescue process also occurs in the presence of 1 in M-eyelie nucleotides and IBMX, in contrast to the situation found in puromycin-treated mouse oocytes (Clarke & Masui, 1985).

Finally, it was found that the application of DNP to the metaphase-l-blocked oocyte also induced chromosome decondensation which, however, did not usually result in the production of multiple resting nuclei. These figures were only present in some instances, after the asters have completely disappeared and the oocytes had been washed out from the inhibitor.

In preliminary experiments, we found that ammonia stimulated by 100% the uptake and incorporation of amino acids into proteins to reach a steady-state level within 1 h. Pulse experiments also indicated that the rate of protein synthesis might increase during the course of maturation and that emetine (90 μM) efficiently blocked amino acid incorporation (data not shown). Preloading experiments were also performed which showed that ammonia did actually stimulate this incorporation independently of the uptake. They also confirmed that emetine was fully effective within 5 min of its application, in great contrast to the effect of 6-DMAP (Fig. 3), a drug that has been already shown to block the sea urchin cell cycle without inhibiting protein synthesis (Rebhun et al. 1973). These important conclusions are corroborated by the data produced in Fig. 4, where the rates of incorporation, as expressed in percentage of the uptake, have been determined in pulse experiments performed within 1 h after ammonia stimulation.

Qualitative differences were also recorded in the patterns of protein synthesis observed after [³⁵S]methionine incorporation, SDS-PAGE and autoradiography. In particular, a new protein of about 48×10³M_r was found to appear within 20min after ammonia stimulation, which was not synthesized by the unstimulated GV-blocked oocyte and remained stable following an 8min treatment with emetine (Fig. 5). Since this protein is also produced in the presence of inhibiting concentrations of 6-DMAP (150 μM-1 min), it is thus clear that the emetine-like cytological effects produced by 6-DMAP cannot be explained by a failure of the protein synthesis machinery.

Within 2 min following ammonia addition, the rate of endogenous protein phosphorylation increases by about 20-fold, showing a small inflexion around GVBD (Fig. 6). As shown on Fig. 7, plateau level is reached within 1 h, at which time the metaphase spindle has already attached to the animal pole. This figure also illustrates the fact that a 30 min incubation with 90 μM-emetine neither blocks meiosis reinitiation nor the phosphorylation burst induced by ammonia, which persists at least for 30 min and allows the formation of a typical metaphase spindle bearing condensed chromosomes. At this stage, however, a rapid dephosphorylation process occurs, while resting nuclei begin to appear in the cytoplasm after destruction of the spindle. The same features are also triggered when emetine is only added to the meta-phase-l oocytes, 1 h after ammonia stimulation.

Phosphorylation was also reduced following a 30min incubation with 100 μM-DNP, a concentration that allowed GVBD to occur but blocked spindle formation, and definitively suppressed with 400 μM, a concentration that precluded GVBD. Addition of the same concentrations of DNP to metaphase-1 oocytes also produced a 50% reduction in the extent of phosphorylation (data not shown). The same progressive reduction in phosphorylation capacity has been recorded following the application of various concentrations of 6-DMAP, which also correlated perfectly with the cytological effects of this compound. Data presented in Fig. 8 demonstrate, moreover, that this drug, which does not affect protein synthesis, is quite effective in reducing endogenous protein phosphorylation, whenever it is applied. However, as illustrated in Fig. 9, this effect is perfectly reversible, an observation supported by the fact that resting nuclei rapidly disappearand that chromosomes recondense under these conditions. In contrast, phosphorylation did not start again after washing out such inhibitors as emetine or DNP, which permanently affected the oocyte (data not shown).

The effect of 6-DMAP upon protein phosphorylation can be explained in different ways, i.e. by a stimulation of the protease or phosphoprotein phosphatase activities, by an inhibition of the protein kinase activity or by a reduction in the availability of the ATP nucleotide pool. To investigate these possibilities, we first tried to characterize the components that may affect dephosphorylation in vitro. Results from such experiments demonstrate that 6-DMAP (750 μM) as well as emetine (90μM) have no influence on the rate of déphosphorylation observed under these conditions (Fig. 10). Moreover, when used in association with phosphoprotein phosphatase or protease inhibitors, these drugs did not influence their protective effect. This indicates that they do not directly activate proteolysis or the phosphoprotein phosphatases that were present in our assay and that we could inhibit using α-naphtyl phosphate (Pondaven & Meijer, 1986). At present, we cannot eliminate the possibility that 6-DMAP might have affected the ATP pool level. However, the rapidity with which oocytes recover from this inhibitor and the fact that its cytological effects are closer to those produced by emetine than by DNP makes this proposition rather unlikely. We are thus inclined to think that 6-DMAP may directly affect protein kinase activity as we have recently shown to be the case in starfish ooctyes, where 6-DMAP and emetine exert the same biochemical and cytological effects as observed in Patella (Guerrier & Néant, 1986b).

We have shown previously that GV-blocked oocytes of Patella vulgata reinitiate meiosis up to metaphase 1 when their intracellular pH is increased by about 0·6 units (Guerrier et al. 1986). Shaking experiments also demonstrated that the threshold pH required for chromosome condensation and metaphase-1 spindle formation was slightly lower than that needed to obtain GVBD. In the present study, we confirm that these treatments induce the appearance of a stable MPF activity which can be amplified following heterologous cytoplasm transfers using GV-arrested starfish oocytes as recipients. In contrast, homologous transfers performed from metaphase-1 to GV-arrested Patella oocytes failed to trigger GVBD. This indicates that, in Patella, MPF amplification as well as MPF formation depend on an elevated intracellular pH that may approach the values reported to occur in GV-arrested oocytes of the starfish, i.e. 7· 09 ±0·8, according to Moody & Hagiwara (1982) or from 6·8 to 7· 04, according to Picard & Dorée (1983).

Our data also reveal that ammonia-induced MPF formation is accompanied by a dramatic increase in protein phosphorylation up to a plateau level characteristic of the metaphase-1 stage. The inhibition of protein synthesis by emetine allows development of the phosphorylation burst, GVBD and formation of the metaphase spindle but soon results in protein dephosphorylation. Simultaneously, the metaphase-1 spindle disappears and the chromosomes disperse and decondense, giving rise to typical resting nuclei. This suggests that MPF formation in Patella does not require the synthesis of an initiator but that CSF, which maintains phosphorylation, chromosome condensation and metaphase-1 spindle structure depends on a continuous supply of short-lived proteins. These may be new MPF molecules differing in some way to the previously activated MPF precursor molecules and/or regulatory factors which may activate protein kinases or counteract dephosphorylation. Indeed, one of the most important findings made during this study was the observation that chromosome condensation activity was sensitive to the déphosphorylations induced by emetine, DNP or the drug 6-DMAP which had no direct effect upon protein synthesis or phosphoprotein phosphatase activity.

Thanks are due to Mrs C. Guerrier for excellent technical assistance throughout this work and for preparing the illustrations, to Mrs N. Guyard for typing the text, to G. Sinquin, who is responsible for the electron microscope in Brest and to the fishermen of the Biological station at Roscoff who provided us with animals throughout the breeding season. This work was partly supported by ATP 84C0939 from the ministry of Research and Industry and by the ARC (fellowship to I.N.).