J. Cell Sd. 80, 253-268 (1986) 253 Printed in Great Britain © The Company of Biologists Limited 1986 MITOSIS IN THE FISSION YEAST SCHIZOSACCHAROMYCES POMBE AS REVEALED BY FREEZE-SUBSTITUTION ELECTRON MICROSCOPY KENJI TANAKA* AND TOSHIO KANBE Laboratory of Medical Mycology, Research Institute for Disease Mechanism and Control, Nagoya University School of Medicine, Tsuruma 65, Shotua-ku, Nagoya 466, Japan SUMMARY Nuclear division in Schizosaccharomyces pombe has been studied in transmission electron micrographs of sections of cells fixed by a method of freeze-substitution. We have found cytoplasmic microtubules in the vicinity of the spindle pole bodies and two kinds of microtubules, short discontinuous ones and long, parallel ones in the intranuclear mitotic spindle. For most of the time taken by nuclear division the spindle pole bodies face each other squarely across the nuclear space but early in mitosis they briefly appear twisted out of alignment with each other, thereby imparting a sigmoidal shape to the bundle of spindle microtubules extending between them. This configuration is interpreted as indicating active participation of the spindle in the initial elongation of the dividing nucleus. It is proposed that mitosis is accompanied by the shortening of chromosomal microtubules simultaneously with the elongation of the central pole-to-pole bundle of microtubules of the intranuclear spindle. Daughter nuclei are separated by the sliding apart of interdigitating microtubules of the spindle at telophase. Some of the latter bear dense knobs at their ends. INTRODUCTION Mitosis in Schizosaccharomyces pombe has been studied with light and electron microscopy by McCully & Robinow (1971). Mitotic chromosomes were neither identified in the nuclei of living cells nor in electron micrographs, but light microscopy of stained preparations showed them grouped in two clusters at opposite poles of dividing nuclei. The authors believed the chromosomes to be directly attached to the spindle pole bodies (SPBs), which, following Girbardt (1968), they more specifically referred to as 'kinetochore equivalents' (KCEs). They attributed division and separation, and thus the segregation of the chromosomes, chiefly to expansion of the region of the nuclear envelope between the SPBs and opposite the nucleolus. They assigned only a passive, scaffolding role to the intranuclear spindle, which in most of their micrographs appears as a straight bundle of parallel microtubules (MTs). The behaviour of chromatin in S. pombe has been further studied by Toda et al. (1981) with the help of the DNA-binding fluorescent probe DAPI. The work of these authors confirms that the configuration of the chromatin compartment of the nucleus is that ascribed to it, on the basis of electron micrographs, by McCully & Robinow (1971). • Author for correspondence. Key words: mitosis, Schizosaccharomyces pombe, microtubules, freeze-substitution, electron microscopy. 254 K Tanaka and T. Kanbe A great advance in our understanding of mitosis in yeast Saccharomyces cerevisiae is due to Peterson & Ris (1976). These authors demonstrated convincingly, with the help of series of longitudinal sections and cross-sections, that in 5. cerevisiae the central pole-to-pole bundle of MTs of the intranuclear spindle is flanked by short, divergent MTs whose number in haploid cells as well as in diploids is in good agreement with the known number of linkage groups (w = 17) in S. cerevisiae. This finding suggested to the authors that, apart from the failure of the chromosomes to condense, mitosis in Saccharomyces follows a conventional course, with the chromosomes attached by MTs (i.e. not directly) to the spindle poles. The presence of wisps of electron-dense material between and around the ends of the short, divergent MTs further strengthened the authors' proposal. Mitosis infissionyeasts is unlikely to differ fundamentally from mitosis in budding yeasts. The example of the fruitful work of Peterson & Ris (1976) has encouraged us to think that scrutiny of serial sections of S. pombe nuclei, something that had not been carried out before, might increase the probability of finding the few chromosomal MTs to be expected in S. pombe, in which n = 3 (Kohli et al. 1977). To improve further our chances of not missing important details we have used the freeze-substitution method of fixation advocated by Howard & Aist (1979) for better than usual preservation of microtubules. Our investigation is still in progress. We have indeed found what we interpret as chromosomal MTs but so far only in longitudinal sections; cross-sections of dividing nuclei remain to be studied. The purpose of the present paper is threefold. To demonstrate the preservation of the natural shape of 5. pombe nuclei attainable by freeze-substitution, to demonstrate the presence of short, diverging, discontinuous, probably chromosomal MTs, and lastly to report that we have several times encountered sigmoidally twisted spindles in otherwise well-preserved nuclei. The latter observation suggests to us that the spindle in S. pombe may play a more active role in mitosis than that ascribed to it by McCully & Robinow (1971). MATERIALS AND METHODS Materials Sckizosacchantmyces pombe strain h90 was used. The strain was maintained at room temperature on MY agar, which contained 0-3 % malt extract (Difco), 0-3 % yeast extract (Difco), 0-5 % peptone (Difco), 1 % glucose and 2 % agar. Phase-contrast microscopy of living cells Cells grown in MY broth at 28 °C overnight were used as an inoculum for 'spreading drop' slide cultures according to Robinow (1975). Living cells growing in a thin film of MY medium containing 21 % gelatin were viewed with phase-contrast using Zeiss Ultraphot II microscope. Photographs were taken with X100 objective in conjunction with OPTOVAR in position 1-25X, which provided an initial magnification of 400x on 35 mm film (Fuji NEOPAN 400). Freeze-substitution electron microscopy For freeze-substitution cells were inoculated into MY broth and cultured overnight at 28 °C. A drop of the culture was spread on the surface of rectangular 5 mm X 7 mm pieces of cellulose Mitosis in S. pombe 255 tubing (Cellophane tubing-seamless, Union Carbide Co., U.S.A.) placed onto the MY agar and was incubated for about 6h at 28 °C. Minor modifications apart, freeze-substitution was carried out as described by Howard & Aist (1979). Pieces of cellulose tubing on which cells were growing were quickly taken off the agar and immediately plunged into melting Freon 23 cooled with liquid nitrogen. The frozen sample was transferred to the substitution fluid, anhydrous acetone containing 2 % OsC^ and 0-05% uranyl acetate, maintained at —79°C with solid COz/acetone, and left for 48 h. Then the samples were transferred to —20°C for 2h, to 4°C for 1—1 -5 h and finally to room temperature for 30min. They were rinsed four times with anhydrous acetone. The pieces of cellophane with cells attached were infiltrated with increasing concentrations of Epon-Araldite in anhydrous acetone and finally with 100 % Epon-Araldite. Resin-infiltrated cellophane pieces were sandwiched between Teflon-coated glass, polymerized at 70°C for 48 h, and were checked under phase-contrast to determine which cells were well-frozen. Well-frozen cells were trimmed, mounted on resin blocks and thin sections were cut with a diamond knife. Serial sections collected on Formvar-coated single-slot grids were stained with uranyl acetate and lead citrate. They were viewed in a JEOL 100 CX or HITACHI H-800 electron microscope operated at lOOkV. RESULTS Observations on the nuclei of living cells with phase-contrast microscopy The log-phase cells had many vacuoles scattered throughout the cytoplasm, which sometimes made it difficult to follow the behaviour of the nucleus. However, in MY—gelatin medium vacuoles disappeared in due course and the nucleus stood out in low density against the dark background of the cytoplasm. Fig. 1 shows a set of timelapse phase-contrast micrographs illustrating phases of nuclear behaviour during mitosis in two cells of 5. pombe. The beginning of mitosis was difficult to detect but a slight decrease in the density of the nucleolus indicated that the cell was going into division. A distinct feature of the mitotic nucleus was a change in nuclear morphology from spherical to more or less rectangular or ellipsoidal (Fig. 1: 6, 20), followed by elongation into a gourd form (Fig. 1: 12, 25). The nucleolus was stretched out in the interior of the elongated nucleus. After the dumbbell stage had been reached, separation into daughter nuclei had taken place (Fig. 1: 14). The divided nuclei were seen at the end of the cell for some time after separation (Fig. 1: 20, 32) and moved backward later at the time of septum formation (Fig. 1: 52). We tried to determine the time course of mitosis in 5. pombe, but it was found to be difficult to define the exact time from the time-lapse photomicrographs for each stage of mitosis, in particular the time when mitosis began. However, our preliminary studies on five sets of time-lapse observations indicated that about 5 min were required for the ellipsoidal nucleus to divide and separate into the two daughter nuclei and about another 10 min before septum formation began. Electron microscopy of mitosis in S. pombe Fig. 2 shows part of a section with one cell about to be bisected by a transverse septum and the nuclei of its two neighbours at interphase of mitosis. The nuclei in all three cells have circular profiles. They contain eccentrically placed nucleoli (as familiarity with the nuclei of living S. pombe would lead one to expect) whose substance contrasts more strongly with the chromatin portion of the nucleus than it does in sections of conventionally fixed 5. pombe. Vacuoles that appear to be filled 256 K. Tanaka and T. Kanbe Mitosis in S. pombe 257 with dense material in conventionally fixed cells have a spongy structure in S. pombe prepared by freeze-substitution (cf. McCully & Robinow, 1971;figs1-6, 20 and 21). Ribosomes also appear uncommonly well preserved. An SPB is located in a zone bounded by the nuclear envelope and a mitochondrion (Fig. 3), an association already noted by McCully & Robinow (1971). Beneath the SPB across the nuclear envelope there is some amorphous dense material that seems to contain very short MTs (arrowheads in Fig. 3). The SPB is also associated with a few cytoplasmic MTs. It has a dumbbell shape, with a long axis of 350 nm and short axis of HOnm (Figs 3, 4). At the start of mitosis the SPB divides, with spindle MTs developing between sister SPBs. In the earliest stages when the two SPBs are only a short distance apart the spindle is composed of short divergent and longer pole-topole MTs (Fig. 5). At this stage of mitosis one of the SPBs invariably occupies a position in the nuclear envelope close to the nucleolar region (Figs 5,6). The nuclei at this stage have either rounded or oval contours. The growing spindle gradually comes to consist of a slightly curved bundle of long, parallel, probably continuous MTs and short discontinuous MTs. One of these (arrowheads in Fig. 7B) is seen ending in a circumscribed area of dense material differing slightly in texture from the rest of the nuclear contents and perhaps representing part of a chromosome. In the series illustrated by Figs 8 and 9 the nucleus becomes ellipsoidal, with slightly pointed poles. In Fig. 8 the nucleolus has just begun to be stretched out along the spindle and in Fig. 9 nucleolar material invests the spindle along most of its length, recalling fig. 35 of McCully & Robinow (1971) and several sets of observations on dividing nuclei of living cells of other species of fission yeasts (e.g. see Robinow, 1980). In the nucleus of Fig. 8 the flat inner surfaces of the SPBs are no longer oriented parallel to each other and the spindle MTs that arise or terminate perpendicular to these surfaces compensate with a double twist for the changed alignment of the SPBs. We have collected eight sets of sections illustrating this remarkable, previously unrecorded configuration, which will be dealt with further in the Discussion. Three short cytoplasmic MTs are associated with SPBs in Fig. 8c. The process of constriction begun in the nucleus of Fig. 9 has been completed in Figs 10 and 11, where daughter nuclei are still connected by a narrow corridor containing a few MTs, three of which bear knobs of dense material at their ends (Fig. 11). Freeze-substitution frequently does not preserve membranes well and did not do so in this instance, but we know from sections of a nucleus at a corresponding stage of constriction (figs 37, 39 of McCully & Robinow, 1971) that a continuous envelope Fig. 1. Time-lapse phase-contrast micrographs of the two cells undergoing mitosis. Numbers at the top of each Figure give minutes elapsed since the first picture of the series was taken. X36O0. The interphase nucleus is seen in the lower cell until 12min. The beginning of mitosis is seen at Omin for the upper cell and 14min for the lower cell. The nucleus takes a more or less rectangular shape (6 min for the upper and 20 for the lower cell), and elongates into the gourd form (12min for the upper cell and 25 for the lower one). The intermediate is shown in elongated ellipsoidal shape (9 min for the upper cell). The two daughter nuclei produced by division move towards the ends of the cell (20 min for the upper cell and 32 for the lower one) and turn back to the centre of the daughter cell when the septum is formed (52min for the upper and the lower cells). 258 K. Tanaka and T. Kanbe surrounds daughter nuclei as well as the narrow channel that still connects them. Arrowheads in Fig. 10 point to profiles of the incipient transverse septum. A slightly later stage of cell and nuclear division is represented by the cell on the right in Fig. 2. Our observations and conjectures are summarized schematically in Fig. 12. So far we have examined only sections cut parallel to the plane of the dialysis tubing on which the yeasts were grown. We have therefore not yet been able to study Mitosis in S. pombe 259 the cross-sections of mitotic spindles, which would be required for reliable counts of the type and number of MTs involved. Investigations along this line are now in progress. DISCUSSION The smooth contours of the nuclei in our sections of cells preserved by freezesubstitution recall those of the nuclei in living cells of 5. pombe examined by phasecontrast microscopy (Fig. 1) and suggest that the wrinkled, indented countours of the nuclei in the electron micrographs published by McCully & Robinow (1971) are artefacts of the conventional fixation procedure followed by these authors. The ribosomes also seem particularly well defined in our sections. This is not, however, true of plasma membrane and nuclear envelope, which, on the whole, are rather indistinct in our material. Better preservation than that achieved by our predecessors, combined with a sufficient number of serial sections, has enabled us to demonstrate cytoplasmic MTs in close vicinity to the SPBs (Figs 3, 4, 8B,C, 9A,B) as well as short, discontinuous intranuclear MTs diverging from the spindle poles (Figs 5B,D, 7B,C). Freezesubstitution thus endows the mitotic spindle of S. pombe, at least in the early stages of nuclear division, with a close resemblance to the mitotic spindle of 5. cerevisiae as described by Peterson & Ris (1976). We regard the short intranuclear MTs as chromosomal ones and are confident that further work will show that designation of the SPBs as kinetochore equivalents (McCully & Robinow, 1971) was inappropriate. In the majority of published light and electron micrographs of dividing yeast nuclei the mitotic spindles, beyond a certain point of development, appear remarkably straight. Exceptions from this rule are provided by fig. 29 of McCully & Robinow (1971) and fig. 7 of Byers & Goetsch (1975), which show slightly curved spindles. In dividing nuclei we have repeatedly encountered spindles more strikingly curved and twisted than either of the earlier examples and this has led us to conclusions regarding the mechanism of mitosis in S. pombe that differ from those arrived at by previous students of this species, as well as of Saccharomyces. In 5. cerevisiae the spindle tends initially to be inclined at a large angle to the main axis Fig. 2. A section of three S. pombe cells preserved by freeze-substitution. The cytoplasm is packed with dense ribosomes. Mitochondria are seen as variously shaped profiles of low density lacking defined internal structure. The cytoplasm also contains many vacuoles filled with a spongework of dense matter. Nearly half of the volume of the nuclei is occupied by nucleolar material of labyrinthine organization with more or less transparent channels traversing a dense granular matrix. The third cell from the left has completed mitosis and is about to be divided by an ingrowing transverse septum (arrowheads). Bar, 5/im. Fig. 3. A section of an interphase SPB on the outer membrane of the nuclear envelope. A mitochondrion (m) is close by and between them there is a microtubule. An arrowhead shows the presence of very short MTs associated with the dense material across the nuclear envelope beneath the SPB. Bar, 0-5/an. Fig. 4. Glancing section of an interphase SPB of dumbbell shape. It is associated with several cytoplasmic microtubules. Bar, 0-5 /ttn. 260 K. Tanaka and T. Kanbe Mitosis in S. pombe 261 of the elongating nucleus (Robinow & Marak, 1966; Byers & Goetsch, 1974, 1975; Peterson & Ris, 1976; King, Hymans & Luba, 1982). As it gets longer the spindle becomes increasingly attenuated by the loss of more and more MTs and ends up in the maximally stretched nucleus as a very weak reed indeed (King et al. 1982). For these reasons it is commonly agreed that in Saccharomyces the intranuclear spindle neither initiates nor sustains the marked elongation of the dividing nucleus. To account for the work to be done in dividing the nucleus Peterson & Ris (1976) had recourse to the modus operandi invoked for nuclear division by McCully & Robinow (1971), namely differential expansion of the nuclear envelope. Our own observations on 5. pombe provide no support for the concept of membrane growth as the prime motor of nuclear division. We instead suggest that in 5. pombe nuclear division proceeds in three steps. A fleeting initial phase, when the spindle is actively elongating faster than the long axis of the dividing nucleus, is followed by one in which there is synchrony between the two processes and, finally, by constriction of the nucleus. The first of these postulated phases would account for our repeated finding of Sshaped spindles. The forces involved in the initial moving apart of the SPBs in the periphery of the nucleus remain unknown in S. pombe and other yeasts but plausible speculations based on in vivo experiments may now be made on the mechanics of the later phases of mitosis. Studies of the effects of microlaser burns on the course of mitosis in nuclei of Fusarium (Aist & Berns, 1981) have engendered the view that the nucleus is pulled, not pushed, apart by forces acting on the SPBs via 'astral', i.e. cytoplasmic bundles of MTs, and that the spindle has the task of regulating, by the rate of its extension, the strength of this pull and thus the speed of genome segregation. That such a mechanism of balanced pulling and yielding may also be at work in mitosis in budding yeasts has to be considered after the detection by Kilmartin & Adams (1984), using immunofluorescence microscopy, of broad, often very long tracts of MTs extending from the SPBs into the cytoplasm of Saccharomyces. We have found profiles of a few MTs on the cytoplasmic side of the Fig. 5. A-D. Four consecutive sections of a nucleus at an early stage of spindle formation. MTs radiate from SPBs arranged opposite each other. Several MTs are grouped together to form the continuous pole-to-pole spindle. Bar, 0"5 fiin. Fig. 6. Two SPBs are joined by a slightly curved bundle of MTs in an oval nucleus with eccentrically placed nucleolus. The SPBs appear as dense plaques within the nuclear envelope. One of them is closer to the nucleolus than the other one. At the left end of the spindle a few short diverging MTs can still be discerned. Bar, 1-0 /an. Fig. 7. Five serial sections of a nucleus at a stage of division similar to that shown in Fig. 6. Arrowheads in B point to ends of discontinuous diverging MTs. The one on the left seems to be in contact with granular matter that differs in texture from the rest of the nuclear contents and may represent part of a chromosome. Bar, 0-5 /an. Fig. 8. A-C. Serial sections of an ellipsoidal nucleus with S-shaped central spindle. The nucleolus has a loose texture and partly surrounds the spindle. The SPBs are pressed closely to the nuclear envelope and are associated with a few cytoplasmic ('astral') MTs (arrowheads in B,c). The length of the spindle is 4-0 /an. Bar, 1-0/an. Fig. 9. A-C. Serial sections of a gourd-shaped nucleus entering the phase of constriction. MTs of the spindle, still visible in C, are slightly curved. The nucleolus is more stretched out than the one in Fig. 8 and appears wrapped around the spindle. The spindle is 4-5 /an long. Bar, 1-0/an. 262 K. Tanaka and T. Kanbe Fig. 7. For legend see p. 261 Mitosis in S. pombe Fig. 8, For legend see p. 261 K. Tanaka and T. Kanbe 264 '• *%',**'* <: fr~ j'- Fig. 9. For legend see p. 261 Mitosis in S. pombe Fig. 10. A dividing nucleus of dumbbell shape. AITOWS point to the SPBs at either end. The nucleolar material has been divided between the incipient daughter nuclei, which are still joined by a narroj? spindle channel (a tube, in reality; see the text). The length of the spindle is 8-2/an. Arrowheads point to invaginations of the plasmalemma where a transverse septum is beginning to grow inward. Bar, 1-0 ^m. Fig. 11. Higher magnification of the telophase spindle channel of another cell. It contains a few MTs some of which seem in touch with the borders of the channel and bear dense knobs at their ends. Bar, 1-0 fan. 265 K. Tanaka and T. Kanbe 266 D Fig. 12. Diagrammatic representation of a conjugated course of mitosis in S. potnbe. A. Early stage of spindle formation. B. The stage at which chromosomal tubules have become polarized in opposite direction. C. The S-shaped spindle with chromosomal MTs much shortened. D. Elongation and constriction of the nucleus which now assumes gourd-shape. E. Further elongation of the nucleus and its constriction into a dumbbell shape. Incipient daughter nuclei are still connected by a slender spindle channel (tube). Mitosis in S. pombe 267 SPBs but it remains to be seen whether the Aist-Berns mechanism holds true also for fission yeasts. The picture of mitosis in 5. pombe that may be inferred from our sections is, so far, compatible with the concept of an initially actively pushing but later 'regulating', gradually yielding spindle. However, mitosis mechanics here as elsewhere can be satisfactorily explored only by experiment. Chromosomes cannot be identified reliably in our sections but if the ends of the short, diverging spindle MTs are taken as indicators of the disposition of the chromosomes then it will be seen that genome segregation in S. pombe occurs very early in the course of nuclear division. In this respect 5. pombe behaves like Saccharomyces. Separation of incipient daughter nuclei is achieved by the sliding apart of interdigitating MTs of the telophase spindle. The nature of the dense knobs at the ends of some of these MTs is unknown. We are grateful to Dr C. F. Robinow at the University of Western Ontario and Dr I. B. Heath at York University in Canada for their valuable discussions and for their help in the preparation of this paper. This work was supported by a grant for scientific research (no. 58480018) from the Ministry of Education, Science and Culture and by a research fund from the Toray Science Foundation. REFERENCES AiST, J. R. & BERNS, M. W. (1981). Mechanics of chromosome separation during mitosis in Fusarium (Fungi Imperfecti): New evidence from ultrastructural and laser microbeam experiments. J . Cell Biol. 91, 446-458. BYERS, B. & GOETSCH, L. (1974). Duplication of spindle plaques and integration of the yeast cell cycle. Cold Spring Harbor Symp. quant. Biol. 38, 123-131. BYBRS, B. & GOETSCH, L. (1975). Behaviour of spindles and spindle plaques in the cell cycle and conjugation of Saccharomyces cerevisiae.J. Bact. 124, 511-523. GIRBARDT, M. (1968). Ultrastructure and dynamics of the moving nucleus. In Aspects of Cell Motility, 22nd Symp. Soc. exp. Biol. (ed. P. L. Miller), pp. 249-259. Cambridge University Press. HOWARD, R. J. & AIST, J. R. (1979). Hyphal tip cell ultrastructure of the fungus Fusarium: Improved preservation by freeze substitution. J. Ultrastruct. Res. 66, 224—234. KILMARTTN, J. V. & ADAMS, A. E. M. (1984). Structural rearrangements of tubulin and actin during the cell cycle of the yeast Saccharomyces. J. Cell Biol. 98, 922-923. KING, S. M., HYAMS, J. S. & LUBA, A. (1982). Absence of microtubule sliding and an analysis of spindle formation and elongation in isolated mitotic spindles from the yeast Saccharomyces cerevisiae.J. Cell Biol. 94, 341-349. KOHLI, J., HOTTINGER, H., MUNZ, P., STRAUSS, A. & THURIAUX, P. (1977). Genetic mapping in Schizosaccharomyces pombe by mitotic and meiotic analysis and induced haploidization. Genetics 87, 471-489. MCCULLY, E. K. & ROBINOW, C. F. (1971). Mitosis in the fission yeast Schizosaccharomyces pombe: A comparative study with light and electron microscopy. J. Cell Sci. 9, 475-507. PETERSON, J. B. & Ris, H. (1976). Electron microscopic study of the spindle and chromosome movement in the yeast Saccharomyces cerevisiae.J. Cell Sci. 22, 219-242. ROBINOW, C. F. (1975). The preparation of yeasts for light microscopy. In Methods in Cell Biology (ed. D. M. Prescott), vol. 11, pp. 1-22. New York: Academic Press. ROBINOW, C. F. (1980). The view through the microscope. In Current Development in Yeast Research. Proc. Fifth Int. Symp. Yeasts (ed. G. G. Stewart & I. Russell), pp. 623-634. Toronto: Pergamon Press. ROBINOW, C. F. &MARAK, J. (1966). A fiber apparatus in the nucleus of the yeast cell. J. Cell Biol. 29, 129-151. 268 K. Tanaka and T. Kanbe TODA, T., YAMAMOTO, M. & YANAGIDA, M. (1981). Sequential alterations in the nuclear chromatin region during mitosis of the fission yeast Schizosaccharomyces pombe: Video fluorescence microscopy of synchronously growing wild-type and cold-sensitive cdc mutants by using a DNA-binding fluorescent probe. J. Cell Sri. 52, 271-287. {Received 10 July 1985 - Accepted 29 August 1985)