Cell cycle regulation of p34 ^cdc2 kinase activity in Physarum polycephalum

The M-phase Promoting Factor (MPF) has been characterized in a large variety of organisms from sea urchin eggs to human mitotic cells (reviewed by Hunt, 1989; Murray and Kirschner, 1989a) as an intracellular activity catalyzing the G2/M-phase transition (Smith and Ecker, 1971; Masui and Market, 1971). Characterization of the components of MPF (Dunphy et al. 1988; Gautier et al. 1988; Labbé et al. 1989; Arion et al. 1988) and studies of the genes involved in yeast mitotic control (Beach et al. 1982; Booher et al. 1989; Moreno et al. 1989) have identified in each species a protein homologous to the p34^cdc2Schizosaccharomyces pombe gene product, which is assumed to be the catalytic subunit of this mitotic protein kinase. To date, histone H1 is the best identified in vitro substrate of the cdc2 kinase (Brizuela et al. 1989). In the mature form of the enzyme, p34^cdc2 is stoichiometrically associated with cyclin (Booher and Beach, 1988; Draetta and Beach, 1988; Draetta et al. 1989; Pondaven et al. 1990; Brizuela et al. 1989), a cell cycle regulated protein, originally identified in marine invertebrates (Evans et al. 1983). At the G2/M-phase transition, p34^cdc2 is dephosphorylated on tyrosine and threonine residues, and the p34^cdc2/cyclin complex becomes fully activated (Dunphy and Newport, 1989; Moria et al. 1989; Pondaven et al. 1990; Jessus et al. 1990). Cyclin phosphorylation occurs in metaphase (Meijer et al. 1989; Draetta et al. 1989) and its destruction is probably responsible for enzyme inactivation. A third component of the complex is pl3 the product of the sucl⁺ gene in S. pombe (Brizuela et al. 1987). The exact role of this protein in the regulation of the enzyme activity is still not clear and subject to investigation (Dunphy and Newport, 1989; Jessus et al. 1990). Nevertheless, pl3^cdc2 binds to the cdc2 gene product in S. pombe in vivo and in vitro (Brizuela et al. 1987) and pl3 coupled to Sepharose can be used as an affinity matrix for the purification of p34^cdc2 homologues and associated proteins in every organism studied (Arion et al. 1988; Draetta and Beach, 1988; Booher et al. 1989; Dunphy et al. 1989; Pondaven et al. 1990). It is thus a convenient reagent for the assay of p34^cdc2 kinase activity in a wide range of species.

While the understanding of the activation mechanisms of the p34/cyclin complex is progressing (Dunphy and Newport, 1989; Gould and Nurse, 1989; Pondaven et al. 1990; Ducommun et al. 1990), the regulation of these events in relation to other cell cycle pathways has not been subjected to intensive investigation. The naturally synchronous plasmodium of the myxomycete Physarum polycephalum is a syncytium that can contain 10⁸to 10⁹ nuclei dividing every 10 h with perfect synchrony (Howard, 1932; Guttes and Guttes, 1964). This model offers a unique opportunity for following a biochemical event in a single cell during a normal cell cycle or after physical or pharmacological perturbations (Tyson, 1982). Previous studies using this organism have permitted the characterization of the pathways involved in the regulation of tubulin synthesis (Ducommun et al., unpublished data) and degradation (Ducommun and Wright, 1989), and the regulation of thymidine kinase synthesis (Wright and Tollon, 1979, 1988, 1989; Eon-Gerhardt et al. 1981a,b).

Physarum polycephalum was one of the first systems in which the oscillation of histone H1 kinase activities during the cell cycle were described (Bradbury et al. 1974; Hardie et al. 1976). Furthermore, it has been claimed that addition of partially purified Physarum histone H1 kinases or Physarum extracts on the external surface of an intact plasmodium advances mitosis (Bradbury et al. 1974; Inglis et al. 1976; Loidl and Grobner, 1981). More recently, Physarum pre-mitotic extracts have been shown to contain a MPF activity able to trigger prophase/metaphase transition after injection in prophase-arrested Xenopus oocytes, whereas post-mitotic extracts (S-phase, since there is no G₁-phase in Physarum) contain an inhibitory activity (Adlaka et al. 1988). Furthermore, a homolog of p34^cdc2 has recently been immunologically identified in Physarum (Shipley and Sauer, 1989), using antibodies against a consensus conserved sequence (called ‘PSTAIR’).

In this study, using the universal binding property of pl3^cdc2 to p34^cdc2, we affinity-purified the p34^cdc2 kinase activity from Physarum and we studied its regulation during the cell cycle. Evidence is presented here, first, for the existence of several ‘activating signals’ involving microtubular components and DNA synthesis in the activation of the mitotic kinase, and second, for the occurrence of a metaphase signal leading to the inactivation of this kinase.

Plasmodia (strain CL) were cultured and prepared as described by Ducommun and Wright (1989). Synchronous giant plasmodia (approximately 10cm diameter), grown at 22°C except when stated otherwise, were used before or after the third synchronous mitosis. Plasmodia! fragments were taken at intervals during the cell cycle and flash-frozen in liquid nitrogen. The timing of the mitotic events was followed on plasmodial smears taken at intervals and observed by phase-contrast microscopy (Zeiss Axio-phot with a ×63 objective, ×2 Optovar and a ×4 videolens). Images were recorded with a Lhesa camera (Pasecom) and treated with an image processing system (Sapphire from Quantel) by integrating 200 frames and applying histogram and stretch functions. Screen pictures were taken with a Nikon 35 mm camera (macro-50 mm lens).

pl3 (S. pombe sucl⁺ gene product) was purified from a bacterial expression system as described by Brizuela et al. (1987) and coupled to Sepharose (Pharmacia) following the manufacturer’s instructions with 5 mg of purified pl3 per ml of unpacked beads. Frozen fragments of Physarum. were thawed and sonicated three times for 10 s (Branson sonicator) in 0.2 ml of buffer I (25 mM Tris-HCl, pH 8.0, 10 mM MgCl₂, 15 mM EGTA (ethyleneglycolbis-N,N′,N′AT-tetraacetic acid), 0.1 mM sodium fluoride, 60 mM β-glycerophosphate, 15 mM para-nitrophenylphosphate, 0.1 mM sodium orthovanadate and 0.1% Triton X-100) kept in ice. Sodium deoxycholate and SDS were added to final concentrations of 0.5 and 0.1% (w/v), respectively, and after 5 min of incubation on ice, 800 μl of buffer I was added. The soluble protein fraction was recovered by centrifugation for 10 min at 11000 g. After a 30 min preincubation at 4 °C with 30 μl Sepharose CL6B (Pharmacia) and a 10 min centrifugation at 11000 g, to remove any nonspecific precipitate, specific precipitations were carried out by incubating the lysate with 30 μl of pl3-Sepharose for at least 4 h at 4 °C on a rotator. The complexes were brought down by a 3 min centrifugation at 2000 revs min⁻¹ in a Sorvall RT6000B table-top centrifuge. The pellets were then washed three times in buffer II (50 mM Tris-HCl, pH 7.4, 250 mM NaCl, 50 mM sodium fluoride, 5mM EDTA (ethylenediaminetetraacetic acid), 0.1 mM sodium orthovanadate and 0.1% Triton X-100) with the same conditions of centrifugation. The final pellets were subsequently treated as described below to assay the kinase activity.

Protein concentration was determined according to the method of Bradford (1976) using the Biorad reagent and bovine serum albumin (BSA) as standard. In most experiments, 500 μg of total protein was used for each determination, except for the experiment shown in Fig. 1 in which 2.5 mg of protein was used for each determination. The following inhibitors of proteases were added to buffers I, H and IH: 0.lmM-PMSF (phenylmethylsulfonyl fluoride), 1μgml⁻¹ leupeptin, 10μgml⁻¹ soybean trypsin inhibitor, lμgml⁻¹ aprotinin and 10 μg ml’¹ TPCK (tosyl phenylalanine chloromethyl ketone).

Proteins precipitated by pl3-Sepharose beads were washed (see below) and each sample was equilibrated by an extra wash in kinase assay buffer (50 mM Tris-HCl, pH 7.5,10 mM MgCl₂,1 mM DTT (dithiothreitol)). Pellets were resuspended in 30 id of the same buffer with or without 83 pg ml⁻¹ of histone H1 (Boerhinger-Mannheim). When simian virus 40 (SV40) large T antigen was used (kindly provided by Duncan McVey), 0.2 μg T antigen was added to each reaction. After 5 min of equilibration at assay temperature, 10 μl of kinase assay buffer containing 2 μM cold ATP and 5 μCi of [γ-³²P] ATP (NEN, 3000 Ci mmol ⁻¹) was added to each sample. The kinase reaction was carried out for 15 min for histone H1 kinase assays or for 30 min for phosphorylation of the associated proteins in the absence of histone Hl. The reactions were stopped by addition of 10 μl of Laemmli sample buffer (Laemmli, 1970) and was quantitated by spotting 15 μl of the reaction mixture on Whatmann 3MM paper followed by TCA (trichloracetic acid) precipitation. The paper was incubated first for 10 min in 10% (w/v) TCA containing 40 mM sodium pyrophosphate, then washed three times for 10 min in 5% (w/v) TCA and briefly rinsed in cold (-20°C) ethanol. Quantitation was done using the AMBIS beta scanner system but can also be done by liquid scintillation. When cdc2-associated protein phosphorylation was examined, the reaction was stopped with 30 μl Laemmli sample buffer, the samples were boiled for 3 min and submitted to electrophoresis according to the method of Laemmli (1970). The gels were then dried and autoradiographed with film X OMAT R (Kodak).

Synchronous plasmodia, prepared as described in Materials and methods, were collected at intervals during the cell cycle. The stage of the cell cycle was determined by monitoring the nuclear morphology and the occurrence of the different stages of mitosis by phase-contrast microscopy (Fig. 1A). Histone H1 kinase activity was quantitated after precipitation using pl3-Sepharose. The level of kinase activity was very low during interphase and only started to rise 30 min before metaphase; it was maximal in metaphase and then dropped very abruptly (Fig. 1B).

It has been demonstrated that phosphorylation of the SV40 large T antigen by p34^cdc2 kinase activates the initiation of viral DNA replication in vitro (McVey et al. 1989). Since, in Physarum, in the absence of G₁-phase, S-phase initiation occurs at the end of mitosis (3min after the metaphase/anaphase transition; Beach et al. 1980), we investigated the presence of a kinase activity phosphorylating the T antigen in mitotic extracts. Phosphorylation of large T antigen (Fig. 1C) was assayed after pl3 precipitation from the same Physarum extracts used for determining histone H1 kinase activity (Fig. 1B). A cell cycle variation of kinase activity was observed with a mitotic peak and little activity during interphase (Fig. 1C).

A more accurate determination of the timing of the histone H1 kinase activation during mitosis was obtained by harvesting parts of the same plasmodium at short intervals before and after mitosis and monitoring the stage of mitosis each time. Fig. 2 shows two independent determinations of histone H1 kinase variation in two different plasmodia after pl3 precipitation. The peak of histone H1 kinase activity is an early metaphase event, and the level of kinase activity starts to decrease consistently before the occurrence of anaphase (Fig. 2A and B). When the kinase was assayed in the same conditions in the absence of exogenous added substrate, a 60K (K= 10³M_r) band was heavily phosphorylated at the beginning of metaphase, suggesting that it could be a Physarum cyclin homolog (not shown). Because of the lack of antibodies recognizing cyclin in Physarum, we were not able to confirm this hypothesis.

Methyl benzimidazole carbamate (MBC) and griseofulvin, two chemically unrelated microtubular poisons, have been previously shown to perturb the mitotic microtubules of Physarum plasmodia (Wright et al. 1976; Gull and Trinci, 1974). When a synchronous plasmodium is transferred during G₂-phase to a medium containing 100μM MBC or 20μM griseofulvin, protein, DNA and RNA syntheses are virtually unaffected, while the occurrence of the following mitosis is delayed by several hours (Eon-Gerhardt et al. 1981a). Previous studies have shown that such treatments did not delay the occurrence of periodic tubulin synthesis. In a control plasmodium, tubulin synthesis starts 4h before mitosis and stops abruptly during this stage. In a treated plasmodium, tubulin synthesis begins as in the control plasmodium (Ducommun et al., unpublished data), even though mitosis is delayed. By contrast, the periodic synthesis of thymidine kinase is delayed (Eon-Gerhardt et al. 1981a).

The effects of microtubule poisons on p34^cdc2 kinase activity were investigated. The principle of such experiments consists of transferring a synchronous plasmodium to a medium containing the drug, keeping a small part on untreated medium, and harvesting portions of the treated plasmodium during the subsequent cell cycle. A synchronous plasmodium was transferred 7h before mitosis to a medium containing 200 μM griseofulvin (Fig. 3A) or 5.5 h before mitosis to a medium containing 100 μM MBC (Fig. 3B). Mitosis was delayed by 2.75 h and 4.5 h, respectively. In both cases, the histone H1 kinase activity was not triggered at the time of the control mitosis, but occurred only at the onset of the treated mitosis (Fig. 3A and B). Therefore, the activation of the histone H1 kinase activity seems to require an event involving the formation of the normal microtubular mitotic cytoskeleton. Upon treatment with MBC an abnormal mitosis occurred (Planques et al. 1989). Metaphase persisted for one to two hours with the formation of multipolar spindles, polyasters and condensed chromosomes (see stages d-e, Fig. 3B). The kinase activity decreased before the end of this abnormal metaphase stage, thus suggesting that total completion of metaphase is not necessary to turn off the activation of the p34^cdc2 kinase.

Aphidicolin, an inhibitor of DNA polymerases crand <5, has been shown to be active in Physarum polycephalum. When used at a concentration of 200μM it inhibits DNA synthesis (73%) whereas RNA and protein syntheses are not affected (Eon-Gerhardt et al. 19815). When a plasmodium is treated with aphidicolin up to 3h before metaphase (from S-phase to mid G₂-phase), mitosis is delayed several hours and the nuclei are blocked in early prophase (Eon-Gerhardt et al. 19815).

To investigate the effect of such treatment on the histone H1 kinase activity, a synchronous plasmodium was transferred 4.25 h before the third synchronous mitosis to a medium containing 200 μM aphidicolin (Fig. 4). The occurrence of mitosis in the treated plasmodium was delayed 6.3 h compared to the control (untreated) part of the same plasmodium. The peak of histone H1 kinase activity did not occur at the same time as that of the control mitosis, but was delayed until the occurrence of metaphase in the treated plasmodium.

We investigated the effect of protein synthesis inhibitors on the activation as well as on the inactivation of the histone H1 kinase activity. A synchronous plasmodium was transferred 45 min before the third metaphase on a medium containing 150 mM cycloheximide (Fig. 5A). As previously reported, this treatment completely inhibited protein synthesis (Cummins et al. 1965). The treated part of the plasmodium never underwent mitosis, and the histone H1 kinase activity stayed low. When a similar treatment was applied to a synchronous plasmodium 0.5 h before the third metaphase, the nuclei entered a metaphase-like stage (Fig. 5B), in which they remained blocked for 2-3 h (Ducommun and Wright, 1989). Under these conditions, the kinase activity was activated and remained at a plateau for 2 h before it declined progressively as nuclei were asynchronously progressing through metaphase and anaphase (Fig. 5B).

We have taken advantage of the absolute synchrony of Physarum plasmodia to determine accurately the timing of p34^ccfc2 kinase activation and inactivation during the cell cycle and to define their relationship with other cell cycle-regulated pathways.

In Physarum synchronous plasmodia, the activation of the kinase occurs before metaphase and requires protein synthesis as shown here (Fig. 5A) and as previously reported in different systems (Hunt, 1989). One hour before metaphase (Fig. 1B), the p34^cdc2 kinase activity is low and similar to the overall interphase level in late G₂-phase, then the kinase activity increases during prophase. The maximal activity is observed in early metaphase and decreases abruptly thereafter. Among the various cell cycle events (Tyson, 1982) that have been reported to occur in the synchronous plasmodia of Physarum, the cyclic increase of two nuclear histone H1 kinase activities has already received extensive interest (Bradbury et al. 1973; Hardie et al. 1976). It has been suggested that these nuclear histone H1 kinases could play a role in both histone H1 phosphorylation and chromosome condensation (Bradbury et al. 1974; Matthews, 1980). However, these two nuclear histone H1 kinase activities and the p34^cdc2 kinase activity described in this report do not show a similar timing during the cell cycle. In contrast to the overall variations of p34^cdc2 kinase activity in plasmodial extract, the two histone H1 kinase activities measured in plasmodial nuclei began to increase at least 3h before mitosis and reach their maximal values 1 and 2 h before mitosis, respectively, when the p34^cdc2 kinase was at its low basal level. The two nuclear histone kinase activities decreased thereafter and were very low during mitosis (Hardie et al. 1976). These differences raise the possibility that the nuclear histone H1 kinases that have been previously characterized and the p34^cdc2 histone H1 kinase activities could correspond to distinct kinases.

In order to determine whether the timing of the increase or decrease of p34^cdc2 kinase activity is correlated with a cell cycle event, we have applied several perturbations to the synçhronous plasmodia and determined their effects on the p34^cdc2 activity.

Treatment with aphidicolin delays mitosis and blocks the nucleus in early prophase (Eon-Gerhardt et al. 1981b). During this delay the histone H1 kinase is not activated (Fig. 4). This result demonstrates that entry of the nuclei into a very early prophase stage and H1 kinase activation are at least partially independently regulated, but that the activation of the enzyme between prophase and metaphase requires an activating signal indicating completion of DNA synthesis. The existence of such an effect of aphidicolin in G₂-phase has also been reported in fibroblasts (Fukuda and Ohashi, 1983), suggesting a requirement for DNA polymerase a- or <5 late in G₂-phase.

Among the different microtubular poisons used in various studies, some have been shown to act in vivo on tubulin by forming abnormal microtubule-like structures. In mammalian cells, vinblastine forms paracrystallin structures (Bryan, 1971) and, in Physarum, MBC and griseofulvin induce the assembly of tubulin into abnormal ‘macrotubules’ (Wright et al. 1976; Gull and Trinci, 1974). In both systems these drugs act by decreasing the level of free tubulin available and consequently increase tubulin synthesis (Ben-Ze’ev et al. 1979; Ducommun et al., unpublished data). On the other hand, drugs like nocodazole and colchicine act in vivo on mammalian cells by disassembling microtubules and raising the level of free tubulin, which induces a decrease in tubulin synthesis (Ben-Ze’ev et al. 1979; Caron and Kirschner, 1986).

In mammalian cells nocodazole induces mitotic arrest with a high level of cdc2 kinase activity (Draetta and Beach, 1988) and this property has been used as a method of increasing the yield of p34^cdc2 in enzyme purification (Brizuela et al. 1989). In Physarum (this study), griseofulvin and methyl benzimidazole carbamate delay both mitosis and the triggering of mitotic kinase activity. Thus, different drugs that interfere with microtubules have quite opposite effects, even though each is a potent inhibitor of in vitro Physarum tubulin assembly (Quinlan et al. 1981).

A resolution of this discrepancy might be that those drugs that inhibit cdc2 activation (e.g. MCB in Physarum) sequester tubulin, decreasing the level of functional protein. By contrast, in the case of treatment with nocodazole in mammalian cells, cdc2 is fully activated but the drugacts to disassemble microtubules and raise the level of free tubulin. The present results do suggest that interference with microtubules can inhibit activation of cdc2.

In untreated plasmodia the kinase activity is maximal in early metaphase (Fig. 1) and the histone H1 kinase starts to decline when nuclei are still in metaphase. When a plasmodium is treated with MBC, metaphase is abnormal and lengthened, but the kinase activity declines before completion of this stage. Thus, it is likely that both in untreated plasmodia and plasmodia treated with MBC the signal leading to the inactivation of histone H1 kinase occurs during metaphase and not after completion of metaphase. Cycloheximide treatment of plasmodia in late G₂-phase, blocks the nucleus in a ‘metaphase-like’ stage (Ducommun and Wright, 1989). During this period, the level of histone H1 kinase activity stays high, suggesting that in this organism the signal permitting kinase inactivation requires protein synthesis. It has been suggested that cyclin degradation triggers the inactivation of the kinase activity (Draetta et al. 1989; Murray and Kirschner, 1989b) and we presume that protein synthesis is required for cyclin degradation in Physarum. This study graphically illustrates the highly asymmetric nature of kinase activation and inactivation. One occurs gradually by post-translational modification of pre-existing components (cyclin/cdc2), whereas the other is due to proteolytic cyclin degradation and is very abrupt.

We thank Mark Adelman, Guilio Draetta and Karen Lundgren for their interest in this work and their comments on the manuscript. Phil Renna, Jim Duffy and Louis Donna are thanked for their photographic and artistic work. This work was supported by l’Association pour la Recherche contre le Cancer, and NIH grant GM39620 to D.B., who is an investigator of the Howard Hughes Medical Institute.