Physarum plasmodia do contain cytoplasmic microtubules!
ISABELLE SALLES-PASSADOR, ANDRE MOISAND, VIVIANE PLANQUES and MICHEL WRIGHT
Laboratoire de Pharmacologie et de Toxicologie Fondamentales, C.N.R.S., 205 route de Narbonne, 31077 Toulouse-Cedex, France
Summary It has been claimed that the plasmodium of the dense and complex three-dimensional network, dis- myxomycete Physarum polycephalum constitutes a tinct from the microfilamentous domains and from very unusual syncytium, devoid of cytoplasmic the nuclei. The orientation of the microtubule microtubules. In contrast, we have observed a network varies according to the plasmodial domain cytoplasmic microtubule network, by both electron examined. Generally microtubules show no special microscopy and immunofluorescence in standard orientation except in plasmodial veins where they synchronous plasmodia, either in semi-thin sections are oriented parallel to the long axis of the veins. or in smears, and in thin plasmodia, used as a Differences between our observations and those of convenient model. Cytoplasmic microtubules could previous workers who failed to find cytoplasmic be seen after immunofluorescent staining with three microtubules in plasmodia are discussed. We pro- different monospecific monoclonal anti-tubulin anti- pose that they reflect difficulties of observation bodies. The immunolabelling was strictly restricted mainly due to the fluorescent background. In con- to typical microtubules as shown by electron mi- trast with the previous view, the discovery of a croscopy. These cytoplasmic microtubules were en- microtubule cytoplasmic cytoskeleton in Physarum tirely and reversibly disassembled by cold treatment plasmodia raises several questions concerning its and by either of two microtubule poisons: methyl relationships with other cellular organelles and its benzimidazole carbamate and griseofulvin. The dynamics during different cell cycle events. microtubule network, present in all strains that have been studied, contains single microtubules and microtubule bundles composed of two to eight Key words: microtubule, microtubule disassembly, microtubule poisons, griseofulvine, methyl benzimidazole carbamate, microtubules. Cytoplasmic microtubules form a myxomycete, Physarum, syncytium.
Introduction electron microscopy (Dugas and Bath, 1962; Rhea, 1966; Porter et al. 1965; Havercroft and Gull, 1983) and The microtubular and microfilamentous cytoskeletonS immunofluorescence staining (Havercroft and Gull, 1983; have been extensively studied in the plasmodium of Solnica-Krezel et al. 1990). In contrast, several lines of Physarum, a large syncytium reaching several cm in evidence have suggested that microtubules might be diameter and containing up to 109 nuclei. This macro- present during interphase. First, although tubulin is scopic cell is able to move on its substratum and exhibits rapidly degraded after mitosis (Carrino and Laffler, 1985; vigorous cytoplasmic streaming (Sachsenmaier et al. 1972) Ducommun and Wright, 1989), the amount of tubulin in cytoplasmic differentiated veins; these movements reaches a plateau at the end of S-phase until the next involve an actomyosin network (Porter et al. 1965; Kessler, cyclic synthesis of tubulin occurring in late G2-phase 1982). Microtubules have been observed only in the (Carrino and Laffler, 1985; Ducommun and Wright, 1989). intranuclear spindle during mitosis (Guttes et al. 1968; Second, the heat-soluble 125 kDa polypeptide, a protein Goodman and Ritter, 1969; Ryser, 1970; Sakai and known to bind to microtubules in vitro, is phosphorylated Shigenaga, 1972; Wille and Steffens, 1979; Havercroft and twice during the cell cycle: at late S/early G2-phase and at Gull, 1983) and, with the exception of one unconvincing late G2/prophase stage (Albertini et al. 1990). The last report (McManus and Roth, 1967), it has been repeatedly phosphorylation event could possibly be correlated with claimed that plasmodia are devoid of cytoplasmic the formation of the mitotic spindle, but the former could microtubules (Green et al. 1987; Burland et al. 1983; Paul not be understood, given the apparent absence of micro- et al. 1987; Burland et al. 1988; Eon-Gerhardt et al. 1981; tubules during interphase. Electron microscopy and Gull and Trinci, 1974; Roobol et al. 1984; Birkett et al. immunofluorescence observations have unambiguously 1985a; Solnica-Krezel et al. 1988; Sauer, 1982; Burland, demonstrated that the intranuclear spindle is nucleated, 1985; Laffler and Tyson, 1985; Anderson et al. 1985; during early prophase, on an intranuclear spindle organiz- Solnica-Krezel et al. 1990). In this respect Physarum ing center (Sakai and Shigenaga, 1972; Blessing, 1972; plasmodia seem to constitute a very unusual eukaryotic Tanaka, 1973), and disappears in less than two minutes at cell. the end of telophase (Goodman and Ritter, 1969; Guttes et Despite these numerous reports, the absence of cytoplas- al. 1968). Thus it is likely that any putative interphase mic microtubules was based upon a few observations by microtubules would be located in the cytoplasm. Journal of Cell Science 100, 509-520 (1991) Printed in Great Britain © The Company of Biologists Limited 1991 509 In this report we demonstrate, using appropriate at —20°C. Then the plasmodial pieces were washed in PBS, conditions for both electron microscopy and immunolabel- embedded in Tissue-Tek (Miles) and frozen in liquid nitrogen. ling, that the cytoplasm of Physarum plasmodia contains a Semi-thin sections (4-6 fun) were obtained, placed on slides covered with 1 % gelatin and processed for immunofluorescence. three-dimensional network of microtubules that has no Third, observation of cytoplasmic microtubules in smears ob- obvious relationship with the nuclei. tained from standard synchronous plasmodia was performed in the following way: plasmodial pieces (2 cm x 2 cm) were cut, with the supporting filter, from standard plasmodia during the third Materials and methods cell cycle, and incubated for 15 min at 17 °C in 10 mM Pipes, pH6.5, 5mM EGTA, 10 mM magnesium chloride, 4 M glycerol, Physarum plasmodia 0.1% Triton X-100. Then, smears were fixed for 10 min in cold Microplasmodia (strains CL, M3CIV, TU291, CH713xCH957, methanol and dried or kept in PBS at 4°C before immunofluor- CH713XLU860) were grown at 22°C in agitated cultures (Daniel escence processing. and Rusch, 1961; Daniel and Baldwin, 1964) and used to prepare synchronous plasmodia (Guttes and Guttes, 1964). Microplas- Immunofluorescence modia were fused for 2h on the center of a nitrocellulose The rat monoclonal antibody YL 1/2 (Kilmartin et al. 1982) was membrane in the absence of nutrients. The resulting standard used for all immunofluorescence observations of microtubules plasmodium, grown on semi-defined agar medium (Wright and unless otherwise stated. Plasmodia were reacted for lh at 37 °C Tollon, 1978), was synchronous as judged by the observation of with anti-tubulin antibodies diluted in PBS containing 25 % fetal mitotic stages (Guttes and Guttes, 1964). Thin plasmodia were calf serum, then washed twice in PBS containing 0.05 % Tween-20 obtained from square plasmodial pieces (3 mm x 3 mm) cut, with and in PBS. Then plasmodia were reacted with goat anti-rat or the supporting membrane, from the growing edge of standard goat anti-mouse antibodies labelled with fluorescein, diluted in plasmodium during the third cell cycle following microplasmodial PBS containing 40 % fetal calf serum, for 30 min at 37 °C and then fusion. Plasmodial pieces were placed on a thin film (1 mm) of 1 % washed as before. Nuclei were stained with DAPI (0.2//g ml"1 in agar spread on a glass slide (Naib-Majani et al. 1983), covered PBS) for 15 min at 37 °C and washed twice for 5 min in PBS. with a cellophane membrane and sandwiched with another sheet Preparations were mounted in Mowiol (Planques et al. 1989) and of 1% agar (thickness 4 mm). After 6h the specimens reached a sealed with paraffin wax. Fluorescein (FITC) and DAPI staining diameter of 1-2 cm and were taken at intervals. A smear was were observed by epifluorescence using a Zeiss Axiophot observed by phase-contrast microscopy (Guttes and Guttes, 1964), microscope equipped with a x40 Plan-neofluar objective (NA, while the remaining part was fixed for immunofluorescence. The 0.90), a x63 Plan-apochromat objective (NA, 1.4) and X100 Plan- number of nuclei in thin plasmodia was determined by immuno- neofluar objective (NA, 1.30), an Optovar varying from xl.25 to fluorescence in a Malassez slide: the plasmodia were homogenized x2.5 and a x4 TV camera adaptor. Images, recorded with a in 0.25 ml 10 mM Tris-HCl (pH7.0), 0.25 M sucrose, 10 mix Nocticon camera (LH 4015 from Lhesa), were digitized with an calcium chloride and 0.1% Triton X-100 with a Dounce image processing system (Sapphire from Quantel). Except when homogenizer. The nuclear suspension was diluted in the same stated otherwise, 200 frames were averaged for each picture, an buffer, incubated in the presence of DAPI (4',6-diamidino-2- out-of-focus image was subtracted from the image to reduce phenylindole: 0.2 fig ml"1 in PBS) for 15min and counted by diffuse fluorescence and the final image was recalculated using epifluorescence. the histogram function. Images stored on an optical laser disk, were photographed on a black and white- monitor with a Nikon 35 mm camera (macro 50 mm lens) using Kodak T-Max film with Plasmodial treatments an exposure time varying from 0.25 to 0.5 s. Alternatively, images Thin sandwiched plasmodia incubated at 0°C were placed on of thin plasmodial strands have been observed with a laser scan crushed ice before subsequent treatment for immunofluorescence. microscope (Zeiss LSM 10) equipped with a x63 Plan-apochromat Treatment with the microtubule poisons methyl benzimidazole objective, a 488 nm argon laser, used at 10 % of its maximal power, carbamate (200/JM in 0.5% dimethyl sulfoxide, final concn) and and a X60 zoom. The fluorescence background was removed from griseofulvine (200 ; 510 /. Salles-Passador et al. Then it was incubated for lh in PHEM containing 2.5% likely that the elongated structures, which were immuno- ammonium chloride and washed extensively for 30 min in PHEM. labelled with three distinct monoclonal antibodies, corre- The plasmodium was incubated for 14 h at 4°C in the presence of sponded to cytoplasmic microtubules. However, these the antibody against tubulin (YL 1/2, diluted in 10 mM Tris-HCl, 150 mM sodium chloride, pH7.0 (TBS), containing 10% fetal calf observations did not permit the distinction between serum), and successively washed in TBS for 5, 10, 25 and 60min. isolated microtubules and small bundles composed of Then the plasmodium was incubated for 90 min in the presence of several microtubules parallel to each other. gold-labelled goat anti-rat IgG (Janssen, diluted in TBS contain- The presence of cytoplasmic microtubules in thin ing 25% fetal calf serum) and washed in TBS for 5, 10 and plasmodia was not an indirect consequence of the 120 min. After a second fixation for 1 h at 24°C in TBS containing flattening treatment. Cytoplasmic microtubules were also 1 % glutaraldehyde, the plasmodium was post-fixed for 30 min in observed after immunofluorescent labelling in semi-thin 2% osmium tetroxide and mounted for electron microscopic observation. sections (Fig. 1G) and in smears (Fig. 1H) obtained from standard synchronous macroplasmodia during the third synchronous cell cycle. Although plasmodial sections Results exhibited an apparent thickness of 4-6 /.an, which was similar to the thickness of thin plasmodia, the observation Immunofluorescent labelling of microtubules in the of microtubules by immunofluorescence was perturbed by cytoplasm of Physarum plasmodia the presence of large regions exhibiting an important The microfilamentous cytoskeleton of Physarum plas- fluorescence. The elongated structures that were immuno- modia has been most easily studied by immunofluor- labelled were generally thicker than those that were escence when thin and flat specimens are used (Kamiya observed in thin plasmodia. Moreover, except in some and Kuroda, 1965; Naib-Majani etal. 1982; Naib-Majani et limited areas (Fig. 1G, upper left), they did not form a al. 1983). Small plasmodial pieces, on a supporting complex meshwork. It is likely that the high fluorescence membrane were cut from large standard synchronous background limited the observations to the larger micro- macroplasmodia and sandwiched between two pieces of tubule bundles. In smears, the fluorescence background agar (Naib-Majani et al. 1983). This procedure flattened was reduced and more diffuse (Fig. 1H), but the pertur- the plasmodium without interfering with cell cycle events, bations induced by the shearing of the plasmodial pieces as shown by the occurrence and completion of mitosis. The prevented the study of the spatial organization of the agar-sandwich plasmodial preparations gave the clearest cytoplasmic microtubules. images and were therefore used most. However, the other The failure of previous workers to observe cytoplasmic plasmodial preparations obtained directly from standard microtubules is unlikely to be a consequence of the strains synchronous plasmodia also revealed microtubular el- used, since we found such microtubules in all plasmodial ements by immunofluorescence. strains that were tested: CL (Fig. IF and Fig. 5E-G), Three different monoclonal anti-tubulin antibodies M3CIV (Fig. 5H), TU291, CH957xCH713 and LU860x were used to reveal cytoplasmic microtubules by immuno- CH713 (not shown). fluorescence. Their specificity to tubulin polypeptides was checked by immunoblotting plasmodial extract submitted Sensitivity of plasmodial microtubules to cold treatment to one-dimensional polyacrylamide gel electrophoresis and microtubule poisons (not shown). The two monoclonal antibodies, TPH 4 and Cytoplasmic microtubules were not disassembled by a 1 h KMX 1, recognized only one polypeptide band correspond- treatment at 5°C but were entirely disassembled after ing to the a and /3-tubulin isotypes, respectively. But, in 15 min at 0°C (Fig. 2A (control) and Fig. 2B (treatment at the conditions used, no immunolabelling was observed 0°C for 15 min)). Microtubule disassembly was reversible. with the YL 1/2 monoclonal antibody, in contrast with When plasmodia treated for 1 h at 0°C were rewarmed to previous observations on Physarum purified tubulin 27 °C the cytoplasmic microtubule network reassembled. showing that this antibody was specific for tubulin «x and As expected from the absence of action of taxol (20 ;JM) on P2 isotypes terminated by an aromatic amino acid (Green Physarum mitosis in vivo (Wright et al. 1982), this et al. 1987; Burland et al. 1983; Gull et al. 1987). The YL microtubule poison did not stabilize cytoplasmic micro- 1/2 monoclonal antibody labelled numerous elongated tubules from cold disassembly. structures (Fig. 1A) in the cytoplasm of interphase Cytoplasmic microtubules were entirely disassembled plasmodia (Fig. IB) forming a dense meshwork without by a 30 min treatment with 200 JIM methyl benzimidazole any special orientation. When this antibody was deleted carbamate or 200 ^M griseofulvin, two microtubule poi- from the immunolabelling protocol only a diffuse back- sons already known to act on plasmodial mitosis (Wright et ground of fluorescence was observed (Fig. 1C). Similar al. 1976; Gull and Trinci, 1974; Hebert et al. 1980; Burland immunolabelling (Fig. ID) was obtained with the KMX 1 et al. 1984). No microtubules remained in the cytoplasm monoclonal antibody, which recognizes ft and ji2 -tubulin after 0.5 h in the presence of 200 /XM methyl benzimidazole isotypes (Birkett et al. 19856) and with the TPH 4 carbamate (Fig. 2C) or 200 fm griseofulvin (Fig. 2D). The monoclonal antibody specific for the a^-tubulin isotype initial disassembly of cytoplasmic microtubules by methyl (Planques et al. 1989) (Fig. IE and F, respectively). In both benzimidazole carbamate and by griseofulvin led to a cases, no elongated structures were observed when these complete absence of immunolabelling of tubulin as- antibodies were omitted from the immunolabelling proto- semblies both in the cytoplasm and in the nuclei. However, col (not shown). The fixation procedure used with the YL after 4 h of treatment with 200 ;i Physarum plasmodia contain cytoplasmic microtubules 511 Fig. 1. For legend see p. 514 512 /. Salles-Passador et al. Fig. 2. For legend see p. 514 Physarum plasmodia contain cytoplasmic microtubules 513 Fig. 1. Immunofluorescence labelling of cytoplasmic griseofulvin confirms that the elongated structures ob- microtubules in thin and standard plasmodia. (A-C) Specificity served by immunolabelling with antibodies against tubu- of the immunolabelling with the anti-tubulin antibody YL 1/2. lin are indeed microtubules. Moreover, the complete Overall view of thin plasmodia (50 ,um long) treated with (A) disappearance of the cytoplasmic microtubule network and without (C) the primary anti-tubulin antibody, respectively. The two images were recorded under identical suggests that these microtubules do not constitute a conditions and were not subjected to an image-enhancing heterogeneous class of organelle, in contrast to what has treatment. The dense and complex microtubule network is only been reported in other cells (Gundersen et al. 1984; Kreis, visible in the presence of the anti-tubulin antibody (A). Most 1987; Schulze and Kirchner, 1987; Piperno et al. 1987). microtubules are located in regions devoided of nuclei, as shown by comparing the same thin plasmodial part either Electron microscopic observation of plasmodial immunolabelled with the anti-tubulin antibody (A) or stained cytoplasmic microtubules with DAPI (B) in order to label nuclear and mitochondrial Both in standard synchronous plasmodia and thin plas- DNA. The diffuse limits of the nuclei suggest that the modia, cytoplasmic microtubules were observed by elec- procedure of fixation that has been used did not lead to a good preservation of the chromatin (see also Fig. 2G and H). X1500. tron microscopy (Fig. 3). They exhibited typical inner and (D-F) Immunolabelling of the microtubule network in thin outer diameters (Fig. 3A, B and C) and 13 protofilaments plasmodia with three different anti-tubulin antibodies (YL 1/2, in cross-section (not shown). Immunogold labelling ob- KMX 1 and TPH 4, respectively). In each case a microtubule tained with the anti-tubulin antibody YL 1/2 demon- network is immunolabelled. x 1500. (G) Immunolabelling, with strated that gold particles were observed in close associ- the anti-tubulin antibody YL 1/2, of cytoplasmic microtubules ation with cytoplasmic microtubules, while a complete in a semi-thin section obtained from a standard absence of gold particles was observed in the remaining macroplasmodium. X3650. (H) Immunolabelling, with the anti- cytoplasm (Fig. 3D). These observations confirm that the tubulin antibody YL 1/2, of cytoplasmic microtubules in a elongated structures observed in the cytoplasm by immu- smear made from a standard macroplasmodium. X1500. Fig. 2. Sensitivity of the cytoplasmic microtubule network to nofluorescent labelling corresponded to microtubules. cold and microtubule poisons. (A and B) Cold treatment. Thin Using serial thin sections, it has been possible to record plasmodia were immunolabelled with the anti-tubulin antibody single microtubules over 6 /an in length. The difficulties of YL 1/2 before (A) and after a 30min treatment at 0°C (B). The following microtubules over long distances in thin sections images have been recorded in identical conditions and have not account for the observation of smaller lengths by electron been subjected to an image enhancing treatment. X1500. (C-G) microscopy than by immunofluorescent staining. It was Treatment with microtubule poisons. (C-D) Immunolabelled thin clear from electron microscopic observations that the plasmodia treated for 1 h with 200 /m methyl benzimidazole cytoplasmic microtubules constituted a three-dimensional carbamate and 200 [a/i griseofulvin, respectively. In both cases, the cytoplasmic microtubule network has disappeared, leaving network. Although, observation of thin sections by fluorescent spots and a diffuse cytoplasmic labelling, respectively. electron microscopy did not easily permit the determi- (E-H) Thin plasmodia treated for 6h by methyl benzimidazole nation of the overall structure of the cytoplasmic micro- carbamate (E and G) and griseofulvin (F and H). Comparison of tubule network, it was clear that cytoplasmic micro- the immunolabelling, obtained with the antibody YL 1/2 (E and tubules had no preferential orientation relative to the F), and of the fluorescent staining of the nuclear and whole plasmodium. In numerous cases microtubules have mitochondrial DNA with DAPI (G and H) demonstrates that been observed underlying the plasma membrane. Numer- after a 6 h treatment with both 200 ,UM methyl benzimidazole ous images suggest that cytoplasmic microtubules can carbamate and 200 jtM griseofulvin no microtubules were present exhibit an overall orientation similar to that of the in the cytoplasm. However, intranuclear microtubules were microfilaments (Fig. 3E). It is also clear that portions of detected in all nuclei after a 6 h treatment by 200 ^M methyl benzimidazole carbamate, while only a few nuclei exhibited the microtubule network are independent of the microfila- intranuclear microtubules after a 6 h treatment with 200 /oi mentous cytoskeleton (Fig. 3E). griseofulvin. (C-H) X3600. Immunofluorescence did not permit distinction between single microtubules and bundles composed of a few microtubules (Inou6,1986). However, electron microscopic griseofulvin was completely reversible. A thin plas- observations showed that in more than 50 % of the cases, modium treated for 6h with 200 /IM griseofulvin and single microtubules were observed (Fig. 4, •). Groups of incubated for 0.5 h in the absence of the drug showed no 2—8 microtubules were frequent (Fig. 3A, B and C) and tubulin assemblies in the nuclei and a typical microtubule 75 % of assembled tubulin consisted of small microtubule network in the cytoplasm. In contrast, the action of methyl bundles rather than single microtubules (Fig. 3D). Half of benzimidazole carbamate was less readily reversible. the microtubules (58.7 %) were found in bundles of 2-5 After a 6h incubation in the presence of 100-200^M microtubules, while bundles of 6-8 microtubules were methyl benzimidazole carbamate, followed by a further observed with a frequency about 5 % (Fig. 4). The 1.5 h incubation in the absence of the drug, tubulin experimental distribution of the number of microtubules assemblies were absent from the nuclei, but no micro- per bundle (Fig. 4, •) did not coincide with a Poisson tubules were observed in the cytoplasm. A 6 h incubation distribution (Fig. 4, •). Less microtubule bundles con- in the presence of 50 ,UM methyl benzimidazole carbamate tained 2 microtubules and more microtubule bundles led to the complete disappearance of the cytoplasmic contained 4-8 microtubules than would be expected from microtubule network, without inducing the assembly of a Poisson distribution. Thus, it is unlikely that micro- tubulin in the nuclei. When this treatment was followed tubule bundles could result only from the random by 1.5 h incubation in the absence of the drug, the association of microtubules with one another, suggesting cytoplasmic microtubule network reappeared, showing the presence of microtubule nucleating centers in the that under these conditions the action of methyl benzimi- cytoplasm. The cytoplasmic microfilamentous and micro- dazole carbamate was fully reversible. tubule skeletons are quite distinct by their size (Fig. 3E). The absence of the cytoplasmic immunofluorescent Although microfilamentous domains occupy large cyto- network after incubation of plasmodia at 0°C or treatment plasmic regions, the complex cytoplasmic microtubule in the presence of methyl benzimidazole carbamate or network is composed mainly of single microtubules and 514 /. Salles-Passador et al. - *'**& Fig. 3. Electron-microscopic observation of cytoplasmic microtubules in standard and thin plasmodia. (A) Transverse section of two microtubule bundles close together in a standard macroplasmodium. One microtubule bundle is composed of two microtubules, while the other is composed of four microtubules. (B and C) Longitudinal section of microtubule bundles in a thin plasmodium (B) and in a standard macroplasmodium (C). (D) Longitudinal section of a thin plasmodium showing a microtubule immunolabelled with the anti-tubulin antibody YL 1/2. The gold particles (10 nm) are strictly limited to the microtubule. (E) Section of a standard macroplasmodium showing both the microtubule and the microfilament networks (*). Although some microtubules are more or less parallel to the microfilaments, other microtubules are either perpendicular to the microfilament network (arrowhead) or show no apparent interactions with the microfilaments (arrow). Bars, 0.2 ^m. bundles formed by a few microtubules. Even in oriented clear location of the structures playing a role in the thin sections obtained from standard macroplasmodia, the nucleation of cytoplasmic microtubules was found. overall structure of the cytoplasmic microtubule network strongly depended on the plasmodial area observed. But, Overall structure of the cytoplasmic microtubule network perhaps due to the complexity of the microtubule network In most areas, the elongated structures immunolabelled and the morphological plasticity of the plasmodium, no by the anti-tubulin antibodies constituted a very indis- Physarum plasmodia contain cytoplasmic microtubules 515 60 i i i i r i Fig. 5. Spatial organization of the cytoplasmic microtubule network in thin plasmodia. (A-D) Absence of relationship between the cytoplasmic microtubules and the nuclei. (A and B) Portion of a thin plasmodium showing a well- 0.5 developed and non-oriented cytoplasmic microtubule network revealed by immunofluorescence labelling with the anti-tubulin antibody YL 1/2 (A) in a region devoid of nuclei and containing mitochondria only (B, DAPI staining, right part of the plasmodial strand), x 1500. (C and D) Portion of a thin 0.4 plasmodium showing a microtubule bundle (C, immunofluorescent labelling with the anti-tubulin antibody YL 1/2) passing around a nucleus (D, DAPI staining). The nucleus is also faintly visible in the background of the enhanced image obtained with the anti-tubulin antibody (C). X3650. 0.3 (e-h) Preferential orientation of cytoplasmic microtubules according to the part of the thin plasmodia. (E) Undifferentiated vein of a thin plasmodium showing that most microtubule bundles are preferentially oriented parallel 20 - 0.2 to the long axis of the vein in contrast to the lack of orientation of microtubules in the cytoplasmic network observed in undifferentiated area. X3650. (F) Detail of the junction between two veins of the same plasmodium. Although most microtubule bundles show no preferential orientation, a 10 - 0.1 microtubule bundle follows the edge of the plasmodium. X3650. (G) Growing edge (50 jian long) of a thin plasmodium showing a non-oriented cytoplasmic microtubule meshwork. x 1500. (H) Part (200 fan long) of a large differentiated vein in a thin plasmodium. The internal microtubule network is oriented 12 3 4 5 6 7 8 preferentially parallel to the long axis of the vein, while the Number of microtubules microtubule bundles located at the two edges of the vein exhibit no preferential orientation. X372. Fig. 4. Distribution of microtubules in cytoplasmic microtubule bundles. (•) Percentage of isolated microtubules and bundles; (A) probability (P) of a cytoplasmic microtubule the microtubules located under the plasma membrane being either isolated or part of a microtubule bundle followed the membrane between the two interconnecting constituted of 2-8 microtubules; (•) theoretical percentage of veins (Fig. 5F), in agreement with electron-microscopic isolated microtubules and microtubule bundles, assuming a observations. In some large differentiated veins micro- Poisson distribution: y=mr.e~m:r\, y being the probability of tubules present in the central core of the vein were more or observing a bundle with 1 to r+1 microtubules, r varying from less parallel to the vein axis (Fig. 5H) and their length 0 to 7 and m=0.8 being the mean of the distribution. reached 60 fan. In contrast, microtubules present in the Microtubule bundles and isolated microtubules were recorded by electron-microscopic observation of thin sections obtained peripheral zone of the vein were shorter (25 fan) and from interphase standard macroplasmodia; 311 bundles presented no preferential orientation, although many corresponding to a total of 647 microtubules were recorded. were more or less oriented perpendicular to the edge of the Two microtubules were considered to be associated when the vein. distance between them was less than or equal to the external Observations performed by confocal laser microscopy diameter of a microtubule. demonstrated that cytoplasmic microtubules, although more abundant near the upper surface, were present in all the thickness of the plasmodium (Fig. 6). Moreover, they tinct meshwork (Fig. 1A and D, covering areas about 50 were not necessarily parallel to the substratum, and and 20 fjm long). There was no obvious relationship extended over several fan in depth. Thus, as shown by both between the density of cytoplasmic microtubules and the confocal microscopy and electron microscopy, the plas- presence or absence of nuclei; microtubules were abundant modial cytoplasmic microtubules constitute a three- in plasmodial areas containing nuclei (Fig. 1A and B) as well as in areas devoid of nuclei (Fig. 5 A and B). Although dimensional network. some cytoplasmic microtubules turned around a nucleus (Fig. 5C and D), we have never observed microtubules Discussion interacting directly with nuclei. Plasmodia exhibit different morphological areas such as Identification of cytoplasmic microtubules in Physarum large homogeneous parts, growing edges at the periphery plasmodia is supported, first by ultrastructural character- of the plasmodial mass and cytoplasmic veins of various istics such as length, inner and outer diameters, number of sizes (Sauer, 1982). In large homogeneous areas (Fig. ID protofilaments, immuno-gold labelling with antibody and e) and growing edges (Fig. 5G) cytoplasmic micro- specific for tubulin chains terminated by an aromatic tubules did not exhibit any particular orientation. They amino acid; second by immunofluorescence with anti- showed an apparent length varying from 2 to 10 /.an. In bodies showing different specificities towards a and /3- contrast, in most of the cytoplasmic veins, the micro- tubulin isotypes; and third by the reversible susceptibility tubules were generally oriented parallel to the long axis of of this network to cold treatment and to two chemically the veins (Fig. 5E) and they reach lengths of as much as distinct microtubule poisons. 20 /.an. In all cases some microtubules seemed to follow the The presence of microtubules in the cytoplasm of edges of the veins (Fig. 5E). When two veins were Physarum plasmodia contradicts previous observations interconnected, the region of connection showed micro- performed both by electron microscopy and immunofluor- tubules with no preferential orientation (Fig. 5E). Often escence (Havercroft and Gull, 1983; Solnica-Krezel et al. 516 I. Salles-Passador et al. 1990; Porter et al. 1965). It is likely that the absence of have been previously used, since we observed a microtubu- plasmodial cytoplasmic microtubules, repeatedly lar network in several distinct Physarum plasmodial reported, is not due to their absence in the strains that strains. The difficulties of preserving cytoplasmic micro- Physarum plasmodia contain cytoplasmic microtubules 517 Fig. 6. Three-dimensional reconstruction of the microtubule network in a thin plasmodial strand, (a) Stereo-view seen from the upper plasmodial surface, (b) Stereo-view seen from the lower surface in contact with the substratum. The plasmodial network labelled with an anti-tubulin antibody was recorded in 8 successive optical sections each 0.5 ,ian thick. It is apparent that some microtubules are more or less parallel to the upper and lower plasmodial surface, while other microtubules are either oblique or even perpendicular to the surface. The small dots of fluorescence that are visible among the microtubules most probably contribute to the fluorescence background, x 1280. tubules in large standard plasmodia (Dugas and Bath, in thickness with a scanning laser microscope as shown by 1962; Rhea, 1966) could explain the failure of previous the fluorescent dots that remained present in the electron-microscopic studies. The short permeabilization three-dimensional reconstructions despite the efficient of plasmodia prior to their fixation with glutaraldehyde threshold image treatment that was applied to each image seems a requisite for good preservation of microtubules before three-dimensional reconstruction (Fig. 6). Thus, we both in the cytoplasm and in mitotic nuclei. Using this suggest that the preparative procedure that we have used procedure, images of mitotic nuclei showed a larger for successful immunofluorescence labelling of cytoplas- number of microtubules than those obtained after fixation mic microtubules could reduce the intrinsic background without previous permeabilization (Moisand and Wright, fluorescence of plasmodia. However, it must be considered unpublished observations). The reported absence of cyto- that additional factors could interfere with the obser- plasmic microtubules by immunofluorescent staining with vation of cytoplasmic microtubules. Although cytoplasmic anti-tubulin antibodies is puzzling, since cytoplasmic microtubules have been repeatedly immunolabelled with microtubules have been carefully searched by different three distinct anti-tubulin monoclonal antibodies, the procedures (Havercroft and Gull, 1983). The disruption of efficiency of each antibody to label cytoplasmic micro- the microtubule network by mechanical shearing during tubules clearly depended on the fixation procedure applied preparation of smears did not prevent the observation of to thin plasmodia and to standard plasmodia. Although cytoplasmic microtubules in standard plasmodia immunolabelling with the YL 1/2 and KMX 1 monoclonal (Fig. 1H). The fixation with glutaraldehyde after gentle antibodies could be observed after a fixation in cold lysis in a microtubule-stabilizing medium permitted the methanol (Fig. ID and E), this procedure was not observation and the immunolabelling of cytoplasmic adequate for the observation of cytoplasmic microtubules microtubules by electron microscopy in both thin (Fig. 3B with the TPH 4 monoclonal antibody (not shown). With and D) and standard plasmodia (Fig. 3A, C and E). this anti-tubulin antibody, the best results were obtained However, fixation with aldehydes (1.5 % glutaraldehyde or when the fixation by cold methanol was preceded by a 3.7 % formaldehyde, for example) has not been successful gentle permeabilization in a microtubule-stabilizing me- in enabling us to observe the cytoplasmic microtubule dium (Fig. IF). network by immunofluorescence, although electron mi- croscopy showed that the same antibodies reacted with Cytoplasmic microtubules have been reported in the well-preserved microtubules after fixation with glutaral- multinucleate protoplasmodia of the myxomycete Echi- dehyde. We observed no cytoplasmic microtubules when nostelium minutum (Hinchee, 1976), which remains the fixation with aldehydes was either preceded by a microscopic throughout its entire existence (Gray and gentle lysis in various microtubule-stabilizing media or Alexopoulos, 1968), reaching a maximum size of 250 fan followed by a post-fixation with methanol. Cytoplasmic (Hinchee, 1976) and representing the most primitive type microtubules were repeatedly observed in semi-thin of plasmodium (Alexopoulos, 1960; Alexopoulos, 1962). In sections and in smears obtained from standard plasmodia, its early stages the plasmodium of Physarum is hardly but a high level of background fluorescence was always distinguishable from a protoplasmodium at the same stage present and prevented a clear observation of the micro- of development (Gray and Alexopoulos, 1968). Similarly, tubule network (Fig. 1G and H). Although more limited in in developing small Physarum plasmodia it has been thin plasmodia (Fig. ID and E), the same difficulty was recently reported from immunofluorescence studies that noticeable, especially when we followed the modified 'microtubules formed a very fine and uniform meshwork' protocol necessary for using the TPH 4 monoclonal that 'could be detected even in some plasmodia containing antibody (Fig. IF). This background of fluorescence was more than 100 nuclei' (Solnica-Krezel etal. 1990). Solnica- probably not due entirely to the diffuse fluorescence of out- Krezel et al. concluded that 'cytoplasmic microtubules of-focus immunolabelled microtubules. Diffuse fluor- disappear, so that in the mature plasmodium the only escence could be observed also in thin plasmodia of 4-6 /an microtubular structure detected is the closed mitotic spindle'. However, it cannot be argued that the thin 518 /. Salles-Passador et al. Physarum plasmodia that have been used in part of this protoplasmic streaming seems to be modified (Guttes and study are not relevant to the ultrastructure of the usual Guttes, 1963; Sachsenmaier et al. 1973; Kessler and synchronous giant plasmodia. First, in contrast with Lathwell, 1979; Nations et al. 1981; Kessler et al. 1981). protoplasmodia, thin plasmodia possessed the overall structure of phaneroplasmodia as shown by the presence of The support from 'L'Association pour la Recherche sur le differentiated protoplasmic veins (Fig. 5H). Second, thin Cancer' is gratefully acknowledged and the gifts of antibodies plasmodia were much larger than protoplasmodia and from Dr Kilmartin and Dr Birkett have been greatly appreciated. contained on average 0.8xl06 to 1.8 xlO6 nuclei. Simi- The confocal microscopy has been kindly performed by Mr Lago, larly, it could also be argued that the procedure used to from Zeiss, in the laboratory of Pierre Fabre Medicament. The flatten plasmodia between two pieces of agar could induce comments and the corrections of Dr M. R. Adelman were helpful some stress or lead to early plasmodial ageing. Thus it can in bringing the manuscript to its final form. be proposed that in these conditions the assembly of cytoplasmic microtubules corresponds to a pre-differen- tiated state leading to spherulation or sporulation. This is References probably not the case, since the procedure used to flatten ALBERTINI, C, AKHAVAN-NIAKI, H. AND WRIGHT, M. (1990). Polypeptides plasmodia mimicks the conditions that are encountered by from the myxomycete Physarum polycephalum interacting in vitro small plasmodia in decaying forest substrata. Moreover, with microtubules. Cell Motil. Cytoskel. 17, 267-275. no morphological signs support the hypothesis that ALEXOPOULOS, C. J. (1960). Gross morphology of the plasmodium and its flattened plasmodia were entering in a pre-differentiated possible significance in the relationships among the myxomycetes. Mycologia 52, 1-20. state. First, sandwiched plasmodia did not show the ALEXOPOULOS, C. J. (1962). Introductory Mycology. John Wiley & Sons compartimentalization that precedes the formation of Inc., New York. spherules. Second, all experiments were performed in the ANDERSON, R. W., DEE, J. AND GULL, K. (1985). Cellular transformations dark and did not permit the induction of sporulation. of myxamoebae. In The Molecular Biology of Physarum polycephalum Third, although a shortage of oxygen could act in (W. Dove, J. Dee, S. Hatano, F. Haugli and K. E. Wohlfarth- Bottermann, eds), vol. 106, pp. 111-130. NATO ASI series: Life preventing mitosis, the third mitosis occurred normally Sciences, Plenum Press, New York. and synchronously in thin plasmodia. Moreover, cytoplas- BIRKETT, C. R., FOSTER, K. E. AND GULL, K. (1985a). Evolution and mic microtubules were not restricted to thin plasmodia. patterns of expression of the Physarum multi-tubulin family analysed They were observed by immunofluorescence in standard by the use of monoclonal antibodies. In Molecular Genetics of Filamentous Fungi, pp. 265-275. Alan R. Liss, New York. plasmodia during the third cell cycle by two different BIRKETT, C. R., FOSTER, K. E., JOHNSON, L. AND GULL, K. (19856). Use of procedures: in thin sections (Fig. 1G) and in smears monoclonal antibodies to analyse the expression of a multi-tubulin (Fig. 1H). Although in both cases it was not possible to family. FEBS Lett. 187, 211-218. observe the overall disposition of the cytoplasmic micro- BLESSING, J. (1972). Ultrastructural changes of the nucleolus from the tubule network, cytoplasmic microtubules were visible, in end of G2-phase until prometaphase in Physarum polycephalum. Cytobiologie 6, 342-350. agreement with electron-microscopic observations made BURLAND, T. G. (1985). Genetic analysis in Physarum polycephalum. In in both standard (Fig. 3A, C and E) and flattened plas- The Molecular Biology of Physarum polycephalum (W. Dove, J. Dee, modial pieces (Fig. 3B and D). These observations demon- S. Hatano, F. Haugli and K. E. Wohlfarth-Bottermann, eds), vol. 106, strate that cytoplasmic microtubules are present both in pp. 19-38. NATO ASI Series: Life Sciences, Plenum Press, New York. BURLAND, T. G., GULL, K., SCHEDL, T., BOSTON, R. S. AND DOVE, W. F. standard plasmodia and in flattened plasmodial pieces. (1983). Cell type-dependent expression of tubulins in Physarum. J. Thus, thin plasmodia constitute a convenient model for Cell Biol. 97, 1852-1859. the observation of both the microtubule (this report) and BURLAND, T. G., PAUL, E. C. A., OETLIKER, M. AND DOVE, W. F. (1988). the microfilament network (Kamiya and Kuroda, 1965; A gene encoding the major beta tubulin of the mitotic spindle in Physarum polycephalum plasmodia. Molec. cell. Biol. 8, 1275-1281. Naib-Majani et al. 1983; Naib-Majani et al. 1982). BURLAND, T. G., SCHEDL, T., GULL, K. AND DOVE, W. F. (1984). Genetic analysis of resistance to benzimidazoles in Physarum: differential In contrast with the oversimplified view that has been expression of /3-tubulin genes. Genetics 108, 123-141. widely accepted, our demonstration of the presence of a CARRINO, J. J. AND LAFFLER, T. G. (1985). The effect of heat shock on cytoplasmic microtubule network in Physarum plasmodia the cell cycle regulation of tubulin expression in Physarum polycephalum. J. Cell Biol. 100, 642-647. raises several problems of general interest. First, the DANIEL, J. W. AND BALDWIN, H. H. (1964). Methods of culture for ultrastructural location of the cytoplasmic microtubules plasmodial Myxomycetes. In Methods in Cell Physiology (D. M. and microfilaments suggests that they form two distinct Prescott ed.), vol. 1, pp. 9-41. Academic Press, New York. cytoskeletal elements. Thus it is necessary to determine DANIEL, J. W. AND RUSCH, H. P. (1961). The pure culture of Physarum polycephalum on a partially defined soluble medium. J. gen. the relationships between them and to discover their Microbiol. 25, 47-59. respective roles in protoplasmic streaming and cell DUCOMMUN, B. AND WRIGHT, M. (1989). Variation of tubulin half-life organelle movement (Porter et al. 1965). Second, our during the cell cycle in the synchronous plasmodia of Physarum observations suggest the need to reinvestigate several polycephalum. Eur. J. Cell Biol. 50, 48-55. DUGAS, D. J. AND BATH, J. D. (1962). Electron microscopy of the slime questions that have been addressed to Physarum plas- mold Physarum polycephalum. Protoplasma 54, 421-431. modia, such as the use of the various tubulin isoforms EON-GERHARDT, R., TOLLON, Y., CHRAIBI, R. AND WRIGHT, M. (1981). (Burland et al. 1983; Paul et al. 1987; Burland et al. 1988), Regulation of thymidine kinase synthesis during the cell cycle of the transient synthesis of tubulin in late G2-phase (Laffler Physarum polycephalum: the effect of two microtubule inhibitors. et al. 1981; Carrino and Laffler, 1985), the arrest of tubulin Cytobios 32, 47-62. GOODMAN, E. M. AND RITTER, H. (1969). Plasmodial mitosis in Physarum degradation 3h after mitosis (Carrino and Laffler, 1985; polycephalum. Arch. Protistenk. Ill, 161-169. Ducommun and Wright, 1989), and the transient phos- GRAY, W. D. AND ALEXOPOULOS, C. J. (1968). Biology of the Myxomycetes. phorylation of a microtubule-binding protein in late S- The Ronald Press Company, New York. phase (Albertini et al. 1990). Third, the observation of GREEN, L. L., SOHROEDER, M. M., DIGGINS, M. AND DOVE, W. F. (1987). Developmental regulation and identification of an isotype encoded by cytoplasmic microtubules in Physarum plasmodia raises altB, an alpha-tubulin locus in Physarum polycephalum. Molec. cell. questions concerning the presence of cytoplasmic micro- Biol. 7, 3337-3340. tubule organizing centers, as well as the dynamics and GULL, K. AND TRINCI, A. P. J. (1974). Ultrastructural effects of turnover of microtubules in a constantly moving cyto- griseofulvin on the myxomycete Physarum polycephalum. Inhibition of plasm, both in interphase and in mitosis when the mitosis and the production of microtubule crystals. Protoplasma 81, 37-48. Physarum plasmodia contain cytoplasmic microtubules 519 GULL, K., WILOOX, M. AND BIRKETT, C. R. (1987). Immunological of tubulin isotype synthesis and usage during mitotic spindle evidence for a Physarum /3-tubulin polypeptide possessing an a- morphogenesis in Physarum. Cell Motil. Cytoskel. 7, 272-281. tubulin-like carboxyl terminus. FEBS Lett. 219, 31-36. PIPERNO, G., LEDIZET, M. AND CHANG, X. J. (1987). Microtubules GUNDERSEN, G. G., KALNOSKI, M. M. AND BULINSKI, J. C. (1984). containing acetylated n-tubulin in mammalian cells in culture. J. Cell Distinct populations of microtubules: tyrosinated and nontyrosinated Biol. 104, 289-302. alpha tubulin are distributed differently in vivo. Cell 38, 779-789. PLANQUES, V., DUCOMMUN, B., BERTRAND, M. A., TOLLON, Y. AND GUTTES, E. AND GUTTES, S. (1963). Arrest of plasmodial motility during WRIGHT, M. (1989). Variation of the immunolabelling of the 520 /. Salles-Passador et al.