The Ultrastructure of Mitosis in Plasmodiophora Brassicae (Plasmodiophorales)

J. Cell Set. 40, 89-110 (1979) Printed in Great Britain © Company of Biologists Limited igyg

THE ULTRASTRUCTURE OF MITOSIS IN PLASMODIOPHORA BRASSICAE (PLASMODIOPHORALES)

ROBERT C. GARBER AND JAMES R. AIST Department of Plant Pathology, Cornell University, Ithaca, New York 14853, U.S.A.

SUMMARY Mitosis was examined in plasmodia of Plasmodiopfwra brassicae within artificially inoculated cabbage roots, using light- and electron microscopy. Mitotic nuclear divisions are characterized by a persistent nucleolus, bipolar centrioles paired end-to-end, densely staining chromatin, and a complex array of membranes that surround and ramify through the spindle. Chromatin begins to condense in prophase, and is aligned at metaphase in a reticulate plate on the nuclear equator. The chromatin is not resolvable into distinct chromosomes at metaphase, and a chromosome count is not possible. Large amounts of membrane cisternae within the spindle are most clearly visible at metaphase, and apparently represent the remains of the nuclear envelope. The nuclear envelope is disrupted during prometaphase and may become entangled in the spindle when centriolar microtubules enter the nucleus. Several concentric sheets of perinuclear endoplasmic reticulum surround the spindle and give the mitotic nucleus the superficial appearance of having an intact nuclear envelope. This interpretation of the identity of nucleus-associated membranes differs from those previously reported for other protists, including members of the Plasmodiophorales.

INTRODUCTION Ultrastructural analyses of mitosis in a variety of lower eukaryotes have contributed in recent years to our appreciation of the great diversity that exists in the mitotic apparatus, and to an understanding of the basic processes responsible for nuclear division. The Plasmodiophorales is a small order of obligately parasitic, plasmodial protists whose phylogenetic affinities are uncertain (Olive, 1975). Plasmodiophora brassicae Wor., a member of the Plasmodiophorales, is an economically important plant pathogen (Karling, 1968) responsible for the clubroot disease, affecting cabbages and other crucifers. Interesting features of mitosis in the Plasmodiophorales, including a persistent nucleolus, an apparently persistent nuclear envelope, and, during metaphase, a dense equatorial plate of chromatin, were noted by light microscopy 80 years ago (Nawaschin, 1899). However, electron micrographs of mitotic nuclei have been published for only 2 members of the Plasmodiophorales: Polymyxa betae Keskin (Keskin, 1971) and Sorosphaera veronicae Schroet. (Braselton, Miller & Pechak, 1975; Dylewski, Braselton & Miller, 1978). The goals of this study were to document the mitotic process in P. brassicae and to attempt to resolve a long-standing controversy concerning its chromosome number (reviewed in Karling, 1968). 90 R. C. Garber andj. R. Aist

MATERIALS AND METHODS Preparation of infected plants Host plants were cabbages, Brassica oleracea L. var. capitata 'Jersey Queen', grown from seed in coarse, autoclaved sand in a controlled-environment growth chamber. The chamber was programmed for a 16-h photoperiod (19000 lux of mixed incandescent and fluorescent lights), a diurnal temperature of 22 °C day and i8°C night, and 65 % relative humidity. Inoculum in the form of resting sporangia of P. brassicae was prepared according to the protocol of Williams, Aist & Aist (1971) from firm, unrotted, field-infected clubbed roots of cabbages collected in New York State. Seedlings were inoculated 7-10 days after planting of seed. Roots were gently washed free of sand, dipped briefly in suspensions of resting sporangia containing approximately 10' sporangia/ml, and replanted in 75-mm diameter clay pots containing either autoclaved coarse sand or autoclaved organic soil. Clubbed roots were fixed for electron microscopy 4-5 weeks after inoculation, at which time the roots were spindle-shaped and 3-6 mm in diameter and contained numerous infected cells whose plasmodia were mostly in the vegetative stage of growth.

Preparation of specimens for electron microscopy Clubs were washed free of soil and immediately cut into 1-mm3 pieces in glutaraldehyde. Tissue cubes were transferred to vials containing fresh glutaraldehyde. Fixation, dehydration, and infiltration schedules are outlined below. All steps were performed at room temperature unless otherwise indicated. (A) (1) Primary fixation with 25 % glutaraldehyde in 0-05 M phosphate buffer, pH 7-2, 1-5-2 h; (2) Postfixation with 0-5 % OsO4 in 0-05 M phosphate buffer, pH 7-2, 2 h; (3) Soak in 0-5 % aqueous uranyl acetate, in the dark, 14 h; (4) Dehydration with acetone series, followed by 100% propylene oxide, total of 6-5 h; (5) Polymerization of plastic at 70 °C, 30 h. (B) (1) Primary fixation with 5 % glutaraldehyde in 008 M NaOH-PIPES buffer (Salema & Brandao, 1973), pH 80, 1-5 h; (2) Soak in 0-5 % aqueous uranyl acetate, in the dark, 2 h; (3) Postfixation with o-i % OsO4 in 02 M NaOH-PIPES buffer, pH 6'8, 1 h; (4) Dehydration with acetone series in 10—20% increments, total of 2-5 h; (5) Infiltration with Epon-Araldite epoxy in 10-20 % increments, total of 7 h; (6) Polymerization of plastic at 70°C, 48 h. (C) (1) Primary fixation with 2-5 % glutaraldehyde in o-i M phosphate buffer, pH 70, 1-5 h; (2) Postfixation with 03 % OsO« in 01 M phosphate buffer, pH 7-0, 2 h; (3) Soak in 0-5 % aqueous uranyl acetate, in the dark, 2 h; (4) Dehydration and embedding as in schedule B.

Light microscopy To improve the efficiency with which the infrequent dividing nuclei were located in fixed and embedded tissue, 2 separate procedures were adopted for preliminary light-microscopic examination. The first technique involved cutting sections 12-15 fim thick, mounting them on glass slides, flooding each section with immersion oil, and examining the unstained sections with a Zeiss Photomicroscope II using Nomarski differential interference-contrast optics. Cells appropriate for ultrastructural examination were photographed at X 128 with a Polaroid camera. Following removal of immersion oil by rinsing with amyl acetate, the entire section was removed from the glass slide, remounted on a fresh block of polymerized Epon-Araldite epoxy using cyanoacrylate glue, and trimmed to a final block face size of approximately 100 /an per side using the Polaroid micrograph as a 'map' to relocate the cell of interest with a stereo- microscope. Thus, the cells that were thin-sectioned and examined in the electron microscope were exactly the same cells that had been observed and selected by light microscopy. The second technique was similarly based on the use of light microscopy followed by electron microscopy of selected host cells. Sections 1 fim thick were cut with a glass knife on an ultramicrotome and mounted on glass slides. Immediately after each i-fim section was removed, a thick section of 8-10 fim was cut and mounted on a fresh epoxy block with cyanoacrylate glue. The i-fim sections were stained on a warming plate for 30 s with 0-05 % Mitosis in Plasmodiophora 91 toluidine blue containing 1 % sodium borate. Cells of interest were selected by bright-field light microscopy and photographed as described above, and the remounted thick sections were similarly trimmed for ultramicrotomy, using the Polaroid micrograph taken of the i-/im section. This procedure allowed more critical light-microscopic examination by the use of thinner, stained sections. However, it sacrificed the advantages of using the identical section for both light- and electron-microscopy.

Electron microscopy Serial thin sections (70-80 nm) were cut with a diamond knife on a Porter-Blum MT-2 ultramicrotome, collected on slot grids coated with Formvar, stained with uranyl acetate in 2 % methanol followed by aqueous lead citrate, and examined in a Philips 200 electron microscope. Sections cut at 200 nm were also used occasionally. The frequency with which each of the stages of nuclear division was observed in fixed material varied considerably. Interphase and metaphase nuclei were examined ultrastructurally in several dozen plasmodia each. Complete series of micrographs of consecutive thin sections were made of at least 14 nuclei at metaphase. Several nuclei at prophase were examined, within a single plasmodium. Prometaphase and telophase nuclei were observed in 2 plasmodia; nuclei in anaphase, in three. Micrographs of partial sets of serial sections were obtained for prometaphase, anaphase and telophase. All linear dimensions are mean values based on measurement of at least 6 examples.

OBSERVATIONS Interphase The great majority of P. brassicae nuclei observed within plasmodia during vegetative growth of the pathogen were in interphase. A large, centrally positioned nucleolus is the most prominent internal feature of the interphase nucleus, both in the light and electron microscopes (Fig. 1). The nuclear envelope is interrupted occasionally by nuclear pore complexes (Figs. 1, 4). In face view (Fig. 4), the pore complexes have the characteristic ring or annulus of subunits, surrounding a pore lumen with a central granule (Franke & Scheer, 1974). Centrioles lie at opposite poles of the nucleus, clearly outside the nuclear envelope. Centrioles in P. brassicae exhibit the typical 9x3 cartwheel configuration of microtubules (Fig. 3), are paired at each pole, and at interphase are closely aligned end-to-end (Fig. 5). This combination of centriole features was consistent in every interphase nucleus observed. Microtubules radiate into the cytoplasm from electron-dense material at the periphery of the centrioles. These cytoplasmic microtubules also extend along the sides of the nucleus but never pass through the nuclear envelope. Cisternae of endoplasmic reticulum (ER) are frequently aligned with the cytoplasmic microtubules (Fig. 1), and may also be closely associated with the nuclear envelope (Fig. 2). Dictyosomes (Fig. 4) are always located at both poles at interphase.

Prophase Mitotic prophase is characterized by the appearance of irregularly shaped masses of chromatin around the nucleolus (Fig. 6). In other respects, the few prophase nuclei observed did not appear to differ ultrastructurally from interphase nuclei. R. C. Garber andj. R. Aist Mitosis in Plasmodiophora 93

Prometaphase From the very different ultrastructural features of prophase and metaphase nuclei, it is possible to predict a number of characteristics that might be expected in an intermediate stage of division, prometaphase. In nuclei that are thus interpreted to be at prometaphase (Fig. 7), there is still no significant elongation along the polar axis. The chromatin bodies stain more densely than at prophase, and are arranged in a configuration that suggests movement toward an equatorial plate from their earlier scattered positions. The nuclear envelope is intact except at the poles, where it is discontinuous. The centrioles are no longer clearly outside the boundary of the envelope, and microtubules radiating from the centriolar region, also previously observed only outside the nuclear envelope, are now visible inside the nucleus and pass from the polar regions to the chromatin masses. All intranuclear microtubules are of the pole-to-chromatin type; no interzonal microtubules were observed. The nucleolus remains in its central position and is still approximately spherical. Dictyo- somes are located near the poles, and there is a substantial number of membrane profiles around the dictyosomes. Membrane cisternae are also present at the polar openings in the nuclear envelopes, as well as inside the nucleus among the microtubules of the forming spindle.

Abbreviations used in figures ce centriole mt micro tubule ch chromatin ne nuclear envelope d dictyosome np nuclear pore complex er endoplasmic reticulum •mi nucleolus I lipid body per perinuclear endoplasmic reticulum in mitochondrion spc spindle cisterna Scale bare represent 05 fim unless otherwise indicated in caption.

Figs. 1-5. Interphase nuclei and centrioles. Fig. 1. A nucleus with bipolar centrioles and a central nucleolus. Segments of endoplasmic reticulum (er) are aligned with some of the microtubules that radiate from the centrioles. Note the nuclear pore complexes. Fixation schedule A. x 23000. Inset: Photomicrograph showing an interphase nucleus with bipolar centrioles (arrows). Fixation schedule A. x 4200. Scale bar, ro/im. Fig. 2. A nucleus partially surrounded by ER. Fixation schedule A. x 19000. Fig. 3. A centriole, sectioned transversely. Members of all 9 microtubule triplets are clearly visible, as are the central hub and spokes. Fixation schedule A. x 90000. Scale bar, 02 fim. Fig. 4. Glancing section of a nucleus, showing face views of 4 nuclear pore complexes (arrows) in the nuclear envelope. The pore complexes have central granules surrounded by an annulus. Fixation schedule A. x 36000. Fig. 5. Centriole pair in longitudinal section, showing close end-to-end pairing at interphase. Fixation schedule A. x 62000. Scale bar, 0-2 /im.

CEL 40 R. C. Garber and J. R. Aist Mitosis in Plasmodiophora 95

Metaphase Metaphase was the most commonly recognized stage of mitosis. A number of trends first apparent in prometaphase are further developed at metaphase. These include movement of the chromatin to an equatorial alignment, enclosure of the centrioles within the outlines of the dividing nucleus, extensive rearrangement of nucleus-associated membranes, and formation of an intranuclear spindle of microtubules. Each of these phenomena is now described in more detail. The chromatin is aligned at the equator of the nucleus in a well defined plate (Figs. 8, 10, 13, 14) surrounding the nucleolus and perpendicular to the spindle axis. In longitudinal sections, the metaphase nucleus displays the characteristic 'Saturn' configuration noted in light-microscope observations (Cook, 1928). The chromatin plate is interrupted by spaces through which microtubules pass (Fig. 12). In nuclei sectioned transversely, through the equator (Fig. 8), the chromatin is seen to be aligned in an anastomosing reticulum surrounding the nucleolus, with numerous interruptions containing microtubules. Serial section analyses showed clearly that the chromatin mass at metaphase was not divided into distinct chromosomes, and that a chromosome count at this stage of division was not possible. The original nuclear envelope is disrupted extensively at both poles at metaphase, and the centrioles are bounded on their cytoplasmic sides by segments of membrane to such an extent that they are part of the spindle apparatus (Figs. 10, 11), in distinct contrast to the extranuclear position of the centrioles at interphase. A narrow gap now separates the members of the centriole pairs (Fig. 12). A striking proliferation and rearrangement of membranes within and around the nucleus is a prominent feature at metaphase. These membranes may be classified into 3 types: membranes within the spindle; large sheets of double membrane which make up an enveloping system around the spindle apparatus; and membranous vesicles at the poles, associated with dictyosomes. The membranes within the spindle are cisternae, and are either positioned near the chromatin plate (Fig. 15) or are scattered among the spindle microtubules (Fig. 9). A short segment of double membrane was consistently found adjacent to each centriole on its nuclear side (Figs. 10-12) and sometimes was apparently continuous with a spindle cisterna (Fig. 20). Examination of serial sections demonstrated the presence of nuclear pores on this membrane fragment and identified it as a piece of nuclear envelope. The 3-dimensional configurations of the spindle cisternae were difficult to trace in thin sections. Connexions between the cisternae and the innermost sheet of membrane which

Figs. 6, 7. Mitotic prophase and prometaphase nuclei. Fig. 6. Prophase, with scattered chromatin masses and extranuclear microtubules. Fixation schedule C. x 32(500. Fig. 7. Prometaphase. Chromatin masses are more dense than at prophase. Intra- nuclear microtubules extend from the poles to the chromatin. The nuclear envelope has become partially disrupted at the poles, and a few spindle cisternae are present. Fixation schedule A. x 29000. 7-2 i?. C. Garier awdj. i?. Aist Mitosis in Plasmodiophora 97 surrounded the spindle were occasionally seen (Fig. 15). Also, the spindle cisternae are aligned frequently with microtubules (Fig. 14) and can be traced toward the poles. The spindle cisternae sometimes show nuclear pore complexes which in morphology and dimensions are indistinguishable from the pore complexes found on nuclear envelopes at interphase (Fig. 9, inset). The membrane system which encircles the mitotic apparatus at metaphase is more highly developed than at earlier stages. Two or three layers of rough endoplasmic reticulum, or 'perinuclear endoplasmic reticulum' (Marchant & Pickett-Heaps, 1970) have replaced the single nuclear envelope which was present as recently as prometaphase (compare Fig. 7 with Figs. 8, 10, 13). The layers are generally separate from one another, although occasional connexions between them, confirmed by serial sections, may be observed (Fig. 20). When 2 dividing nuclei are closely juxtaposed, some of the perinuclear ER may be shared by the 2 nuclei (Fig. 13). Nuclear pore complexes are not discernible in any of the perinuclear ER, suggesting that it does not represent the original nuclear envelope. Two dictyosomes are consistently located at each nuclear pole during metaphase, outside all of the layers of perinuclear ER (Figs. 13, 16). In side view, each dictyosome is seen to consist of 6 or fewer closely stacked cisternae (Figs. 16, 17, 19). Most dictyosome layers are reticula of anastomosing tubules (Fig. I8A). Some layers are composed entirely of spherical, electron-dense vesicles (Fig. 18 B). The face of the dictyosome closer to the nucleus has numerous vesicles interposed between it and the polar ends of the perinuclear ER, which are typically dilated. Some vesicles show apparent continuity with the dilated ER (Fig. 19). The membranes of the spindle also tend to converge at the poles (Figs. 10, 14), but specific close associations or continuities between dictyosome vesicles and spindle cisternae were not found. The cisterna on the cytoplasmic face of the dictyosome is considerably dilated at the margins (Figs. 17, 20). This dilation is relatively electron-transparent. In sections made above the nucleus at the level of the centrioles, the dictyosomes follow the curved outline of a complex aster consisting of the centriole pair and radiating microtubules, as well as dilated perinuclear ER (Fig. 16). The ER nearly encircles the pole.

Figs. 8, 9. Mitotic metaphase nuclei and nuclear pore complexes. Fig. 8. Two neighbouring nuclei, sectioned transversely (left) and longitudinally (right). An equatorial reticulum of chromatin surrounds each nucleolus. Spindle cisternae are present, and perinuclear ER surrounds each spindle. A nuclear pore complex is labelled in a piece of spindle cisterna in the nucleus on the right (see enlarge- ment in bottom-right inset of Fig. 9). Fixation schedule A. x 18000. Fig. 9. Part of a transversely sectioned nucleus, with several spindle cisternae positioned among the spindle microtubules. Fixation schedule A. x 58000. Inset: Comparison of nuclear pore complexes from interphase and mitotic metaphase nuclei. Top: pore complexes in side view, from (left) the nuclear envelope of the interphase nucleus illustrated in Fig. 1 and (right) a spindle cisterna of the metaphase nucleus illustrated in Fig. 9. Bottom: pore complexes in face view, from (left) the nuclear envelope of the interphase nucleus illustrated in Fig. 4 and (right) a spindle cisterna of the metaphase nucleus illustrated in Fig. 8. All are Fixation schedule A. x 61 000. Scale bar, o-i /tm. R. C. Garber andj. R. Aist Mitosis in Plasmodiophora 99 The pole-to-chromatin microtubules, first observed during prometaphase, are further developed at metaphase into a microtubular spindle apparatus. Nearly all microtubules are intranuclear rather than cytoplasmic, and extend from the vicinity of the centrioles to the chromatin plate. Microtubules terminate at the chromatin plate in a variety of ways, some of which suggest kinetochores. Figs. 21 and 22, for example, both show a differentiated structure between the microtubules and the edge of the chromatin plate. These structures are different in nature and electron density from the adjacent chromatin, and resemble simple kinetochores. Such kinetochore-like microtuble terminations are not consistently observed, however. In Fig. 23 B, for example, the area between the microtubules and the chromatin contains material of a very diffuse nature. In Fig. 24B, there appears to be a widening and capping of the microtubule at its tip, without a well differentiated structure between the microtubule and the chromatin. Figs. 23A-C and 24A-C each demonstrate the location of the termination of the microtubule in question by showing the complete absence of the microtubule from the same area in adjacent serial sections. We considered the possibility of determining a 'chromosome number' by counting kinetochores using serial sections. This assumes, however, that in P. brassicae there is a specific number of kinetochores associated with a given mass of chromatin in a mitotic nucleus, and that each such chromatin mass corresponds to a single functional chromosome. Neither assumption is known to be true. Furthermore, due to the highly variable morphology of kinetochore-like structures at metaphase, we were frequently unable to determine whether a given microtubule terminated in a kinetochore. Thus, we believe that 'kinetochore counts' in P. brassicae are unlikely to yield reliable results.

Anaphasejtelophase During anaphase (Fig. 25) the chromatin divides into 2 plates, the chromatin plates separate, and the nucleolus elongates along the spindle axis. Microtubules are present in the interzonal space between the chromatin plates. Chromatin separation is due primarily to elongation of the spindle rather than shortening of the chromatin- to-pole distance. Even in late anaphase, the chromatin-to-pole distance is only slightly decreased from what it was at metaphase. By telophase (Fig. 26), the nucleus is extremely long (approximately 8 /tm). Near

Figs. 10-12. Mitotic metaphase nuclei. Fig. 10. Median longitudinal section, showing centrioles at both poles and pieces of the original nuclear envelope adjacent to each centriole. Fixation schedule A. x 28000. Fig. 11. Part of a nucleus, showing perinuclear ER dilated near the pole (arrows) and 'enclosing' the centriole pair within the division figure. Fixation schedule A. x 24000. Fig. 12. Part of a nucleus, showing interchromosomal microtubules (arrow) extending through a space in the chromatin plate. A narrow gap separates the members of the centriole pair. 200-nm section. Fixation schedule A. x 24000. 100 R. C. Garber and J. R. Aist

Fig. 13. Two mitotic metaphase nuclei, sharing perinuclear ER (arrows). The lower nucleus has 2 dictyosomes associated with 1 pole. Fixation schedule A. x 25 000. Mitosis in Plasmodiophora lit

Figs. 14, 15. Mitotic metaphase nuclei with perinuclear ER and spindle cisternae. Fig. 14. Extensive spindle cisternae are visible in this 200-nm section. Fixation schedule A. x 320500. Fig. 15. Part of a nucleus, showing spindle cisternae near the chromatin and a probable continuity (arrow) between the perinuclear ER and the spindle cisternae. Fixation schedule A. x 37000. 102 R. C. Garber and J. R. Aist

x^lfe?^

% |f r :vSS

^ Mitosis in Plasmodiophora 103 each pole is a densely packed mass of chromatin and nucleolar material, the nucleolus having finally broken apart. Pieces of double membrane which will become the daughter nuclear envelopes form an incomplete covering on the poleward surface of the chromatin. Pieces of envelope also cover the sides of the chromatin/nucleolus mass and trail into the interzonal region. At telophase, there is still a distance of approximately 1 /tm between the centriole and the leading chromatin edge. Dictyosomes are now located adjacent to the chromatin rather than to the centriole. Microtubules and small membrane cisternae are evident in the interzonal region.

DISCUSSION Mitosis in P. brassicae is summarized diagrammatically in Fig. 27. It is characterized by a persistent nucleolus, centrioles paired end-to-end at both poles, densely staining chromatin which aligns in a reticulate metaphase plate, and a complex array of membranes surrounding and ramifying through the spindle. This combination of features, some of which have also been reported for the Plasmodiophoralean species Sorosphaera veronkae (Braselton et al. 1975; Dylewski et al. 1978), differs from the mitotic apparatus known for any other group of organisms, although it has much in common with members of the protozoa, algae, mycetozoa, and phycomycetous fungi (Kubai, 1975). The behaviour of the nuclear envelope during mitosis is an important feature in distinguishing the process of nuclear division in lower eukaryotes from that which occurs in higher plants and animals (Pickett-Heaps, 1969). In contrast to the long- recognized dispersal of the nuclear envelope in prophase, as occurs in most higher eukaryotes, a number of organisms are known in which the nuclear envelope either remains completely closed through mitosis, or else fenestrates at the poles but remains otherwise intact (summarized in Kubai, 1975, and Heath, 1978). Spindle microtubules in the latter case are apparently initiated outside the original nucleus and enter it as the nuclear envelope is disrupted (McNitt, 1973; Whisler & Travland, 1973). The microtubules themselves may be responsible for the disruption of the

Figs. 16-19. An aster complex and dictyosomes associated with mitotic metaphase nuclei. Fig. 16. An aster complex at the pole of a metaphase nucleus. The centriole pair is surrounded by radiating microtubules, dilated perinuclear ER, and 2 dictyosomes. Fixation schedule A. x 43 000. Fig. 17. A dictyosome, showing dilation of terminal cistema (arrows) on the cytoplasmic face. Fixation schedule A. x 29000. Figs. 18 A, B. Adjacent serial sections through a dictyosome, showing a reticulate- tubular cisterna (A) and a layer composed entirely of vesicles (B). Fixation schedule A. x 33000. Fig. 19. A dictyosome and dilated perinuclear ER, showing vesicles interposed between the dictyosome and the ER. Two places (arrows) show apparent continuity between vesicles and the ER. Fixation schedule A. x 48000. R. C. Garber andj. R. Aist Mitosis in Plasmodiophora 105 envelope (Bajer & Mole'-Bajer, 1969). The evidence from P. brassicae suggests another possible fate of the nuclear envelope during mitosis. As the spindle microtubules enter the nucleus early in division, the original envelope would become partially disrupted and entangled among the microtubules. The old envelope would not disperse completely, but instead would remain recognizable throughout mitosis as intranuclear spindle cisternae. Late in division, these cisternae appear to function in reformation of daughter nuclear envelopes. The large quantity of spindle cisternae observed in P. brassicae mitoses makes it unlikely that such cisternae represent merely the remnants of the polar portions of the nuclear envelope, pushed into the spindle by microtubules. The perinuclear ER which surrounds the spindle would serve as a structural, though not necessarily functional, equivalent of the nuclear envelope during mitosis and would give the dividing nucleus the superficial appearance of having a persistent and largely intact nuclear envelope. Evidence for this interpretation comes from the presence of nuclear pore complexes in the spindle cisternae, the absence of recognizable nuclear pores and presence of ribosomes on both surfaces of the perinuclear ER, and the consistent presence of a fragment of isolated and apparently undisturbed original nuclear envelope on the nuclear side of the centrioles. This fragment occasionally showed continuity with the spindle cisternae, but never with the perinuclear ER. Such an interpretation has no close counterpart in the literature, and is different from that presented for mitosis in S. veronicae (Braselton et al. 1975; Dylewski et al. 1978), as well as other organisms where extensive intranuclear membranes (Powell, 1975) or perinuclear ER (Marchant & Pickett-Heaps, 1970; McDonald, 1972; McNitt, 1973; Whisler & Travland, 1973) have been found during mitosis. In S. veronicae, 'intranuclear membranous vesicles' were observed as early as prophase

Figs. 20-24. Nucleus-associated membranes and microtubule terminations in mitotic metaphase nuclei. Fig. 20. Part of a nucleus, showing continuity between 2 adjacent layers of perinuclear ER (single arrow) as well as apparent continuity between a short piece of nuclear envelope next to the centriole and a spindle cisterna (double arrow). Fixation schedule A. x 29000. Figs. 21, 22. Microtubules terminating in kinetochore-like structures (arrows) at chromatin plates. Micrographs are from different nuclei. Fixation schedule A. x 54000. Scale bar, 02 fim. Figs. 23 A—C. Consecutive serial sections, showing a microtubule in B terminating short of the chromatin in a diffuse zone (arrow) which does not show a clearly recognizable kinetochore-like structure. Arrows in A and C point to the same area as the arrow in B, demonstrating by the absence of the microtubule in question that it does in fact terminate where indicated in B. Fixation schedule A. x 54000. Scale bar, 0-2 /(m. Figs. 24A-C. Consecutive serial sections, showing a microtubule in B terminating short of the chromatin with a broadening and apparent 'capping'. There is no clearly identifiable kinetochore-like structure. A and C demonstrate the absence of the microtubule in question from the same area in adjacent sections. Fixation schedule A. x 54000. Scale bar, 0-2 /tm. io6 R. C. Garber and J. R. Aist

Figs. 25, 26. Mitotic anaphase and telophase. Fig. 25. An anaphase nucleus, showing the nucleolus stretched between the 2 separating chromatin plates. Arrows indicate some of the interzonal microtubules. Fixation schedule D. x 30000. Fig. 26. Part of a telophase nucleus. The nucleolus/chromatin mass is partially enclosed by the newly forming daughter nuclear envelope (arrows). Fixation schedule A. x 40000. Inset: Photomicrograph of a telophase nucleus. Only 1 of the 2 nucleolus/chromatin masses is contained in the section (arrow). Fixation schedule A. x 4200. Scale bar, 0-2 /»m. Mitosis in Plasmodiophora 107

Fig. 27. Diagrammatic interpretation of mitotic nuclear division in P. brassicae. A, interphase; B, prophase; C, prometaphase; D, metaphase; E, anaphase, F, telophase. and were interpreted as originating from the inner membrane of a persistent nuclear envelope. The distribution of nuclear pore complexes during division in S. veronicae was not indicated, nor was it evident in the micrographs published. The membrane system surrounding the spindle in S. veronicae appears to be studded with ribosomes, as it is in P. brassicae, and the fragment of envelope on the nuclear side of the centrioles is clearly present (see Figs. 3, 6-8 in Braselton et al. 1975). A close association between dictyosomes and either nuclear envelopes or perinuclear ER, especially at the poles of nuclei with centrioles, has frequently been illustrated in ultrastructural studies of protists (Heath & Greenwood, 1971; Hoch & Mitchell, 1972; Lunney & Bland, 1976; Marchant & Pickett-Heaps, 1970; Markey & Wilce, 1975; Massalski & Leedale, 1969). The occasional discussions of such associations in the literature have either stressed the necessity of having a mechanism for the equal distribution of dictyosomes to daughter cells (Heywood, 1978), or have suggested that nucleus-associated dictyosomes may package molecular precursors of micro- 108 R. C. Gather and J. R. Aist tubules for spindle formation (Heath & Greenwood, 1970, 1971). It is presumably advantageous for an organism such as P. brassicae, in which the entire vegetative thallus is converted into uninucleate spores that function in dissemination and survival, to have a mechanism which assures the consistent association of dictyosomes with every nucleus. The large increase in the amount of perinuclear ER that occurs between interphase and metaphase, and the evident passage of vesicles between the dictyosomes and the dilated ends of the ER are a combination of features of con- siderable interest in our efforts to understand the dynamics of mitosis in the Plasmodiophorales. Chromosome counts of P. brassicae have been attempted numerous times by light microscopy, usually at metaphase (Tommerup & Ingram, 1971; summary of earlier studies in Karling, 1968). The results have been uncertain and frequently contra- dictory. The ultrastructural demonstration, at metaphase of mitosis, of a reticulate mass of chromatin which is clearly not resolvable into discrete chromsomes, even by serial-section analysis, indicates that chromosome counts based on light microscopy were centred on the erroneous interpretation that the metaphase plate was composed of a definite number of separate chromosomes. Thus, the validity of such counts is very doubtful. Centriole replication and migration in vegetative plasmodia of P. brassicae remain unclear. Extensive serial section work demonstrated that paired, bipolar contrioles were present at interphase and metaphase - the only difference being a narrow gap which separated the centrioles of a pair at metaphase. Such observations are in agreement with the appearance of centrioles in S. veronicae. Although Dylewski states that there is only one centriole associated with the interphase nucleus in S. veronicae (Dylewski, 1976), he illustrates (fig. 1 in Dylewski et al. 1978) at both poles of an interphase nucleus the cytoplasmic microtubule arrays which we found from serial sectioning of P. brassicae to be consistently associated with a centriole pair. It therefore seems likely that the 2 organisms are alike in this regard. Although complete serial section sets were not made for mitotic stages other than metaphase in P. brassicae, examination of partial series from nuclei of all mitotic stages failed to show any evidence of replicating, unpaired, or migrating centrioles. Young daughter nuclei appear to have centrioles at only a single pole, as would be expected immediately after division. Since mature interphase nuclei consistently show bipolar centrioles, replication and migration presumably occur soon after telophase. The hope has been expressed several times that a thorough understanding of comparative features of nuclear division among protists can be coupled with other avenues of investigation to yield new insights into protist phylogenetics and the evolution of the mitotic spindle in eukaryotes (Pickett-Heaps, 1969, 1972; Heath, 1975; Fuller, 1976). Such syntheses are beginning to be made (Kubai, 1975, 1978), and the variations in nuclear divisions that emerge from ultrastructural studies of lower eukaryotes such as P. brassicae continue to add to the remarkably wide range of mitotic mechanisms known to have evolved in these organisms. Mitosis in Plasmodiophora 109 We thank H. W. Israel and M. V. Parthasarathy for providing facilitie8 for electron microscopy, M. V. Parthasarathy and H. C. Hoch for critical review, and S. Bucci and M. Strang for technical assistance.

REFERENCES BAJER, A. & MOLE-BAJER, J. (1969). Formation of spindle fibers, kinetochore orientation, and behavior of the nuclear envelope during mitosis in endosperm. Chromosoma 27, 448-484. BRASELTON, J. P., MILLER, C. E. & PECHAK, D. G. (1975). The ultrastructure of cruciform nuclear division in Sorosphaera veronicae (Plasmodiophoromycete). Am.J. Bot. 62, 349-358. COOK, W. R. I. (1928). The methods of nuclear division in the Plasmodiophorales. Ann. Bot. 42, 347-377- DYLEWSKI, D. P. (1976). Somatic Nuclear Division in Sorosphaera veronicae. M.S. Thesis, Ohio University, Athens, Ohio. DYLEWSKI, D. P., BRASELTON, J. P. & MILLER, C. E. (1978). Cruciform nuclear division in Sorosphaera veronicae. Am. J. Bot. 65, 258-267. FRANKS, W. W. & SCHEER, U. (1974). Structures and functions of the nuclear envelope. In The Cell Nucleus (ed. H. Busch), pp. 219-347. New York: Academic Press. FULLER, M. S. (1976). Mitosis in fungi. Int. Rev. Cytol. 45, 113-153. HEATH, I. B. (1975). The possible significance of variations in the mitotic systems of the aquatic fungi (Phycomycetes). Biosystems 7, 351-359. HEATH, I. B. (1978)- Experimental studies of mitosis in the fungi. In Nuclear Division in the Fungi (ed. I. B. Heath), pp. 89-176. New York: Academic Press. HEATH, I. B. & GREENWOOD, A. D. (1970). Centriole replication and nuclear division in Saprolegnia. J. gen. Microbiol. 62, 139-148. HEATH, I. B. & GREENWOOD, A. D. (1971). Ultrastructural observations on the kinetosomes and Golgi bodies during the asexual life cycle of Saprolegnia. Z. Zellforsch. mikrosk. Anat. 112, 37I-389- HEYWOOD, P. (1978). Ultrastructure of mitosis in the chloromonadophycean alga Vacuolaria virescens. J. Cell Sri. 31, 37-51. HOCH, H. C. & MITCHELL, J. E. (1972). The ultrastructure of zoospores of Aphanomyces evtiches and of their encystment and subsequent germination. Protoplasma 75, 113-138. KARLING, J. S. (1968). The Plasmodiophorales, 2nd edn. New York: Hafner. KESKIN, B. (1971). Beitrag zur Protomitose bei Polymyxa betae Keskin. Arch. Mikrobiol. 77, 344-348. KUBAI, D. F. (1975). The evolution of the mitotic spindle. Int. Rev. Cytol. 43, 167-227. KUBAI, D. F. (1978). Mitosis and fungal phylogeny. In Nuclear Division in the Fungi (ed. I. B. Heath), pp. 177-229. New York: Academic Press. LUNNEY, C. Z. & BLAND, C. E. (1976). An ultrastructural study of zoosporogenesis in Pythium proliferum de Bary. Protoplasma 88, 85-100. MARCHANT, H. J. & PICKETT-HEAPS, J. D. (1970). Ultrastructure and differentiation of Hydro- dictyon reticulum. I. Mitosis in the coenobium. Aust.J. biol. Sri. 23, 1173-1186. MARKEY, D. R. & WILCE, R. T. (1975). The ultrastructure of reproduction in the brown alga Pylaiella littoralis. I. Mitosis and cytokinesis in the plurilocular gametangia. Protoplasma 85, 219-241. MASSALSKI, A. & LEEDALE, G. F. (1969). Cytology and ultrastructure of the Xanthophyceae. I. Comparative morphology of the zoospores of Bumilleria sicula and Tribonema vulgare. Br. Phycol. J. 4, 159-180. MCDONALD, K. (1972). The ultrastructure of mitosis in the marine red alga Membranoptera platyphylla. J. Phycol. 8, 156-166. MCNITT, R. (1973). Mitosis in Phlyctochytrium irregulare. Can. J. Bot. 51, 2065-2074. NAWASCHIN, S. (1899). Beobachtungen uber den feineren Bau und Umwandlungen von Plasmodiophora brassicae Woron. im Laufe ihres intracellularen Lebens. Flora, Jena 86, 404-427. OLIVE, L. S. (1975). The Mycetozoans. New York: Academic Press. PICKETT-HEAPS, J. D. (1969). The evolution of the mitotic apparatus: an attempt at comparative ultrastructural cytology in dividing plant cells. Cytobios 1, 257-280.

8 CEL 40 no R. C. Garber and J. R. Aist PICKETT-HKAPS, J. D. (1972). Variation in mitosis and cytokinesis in plant cells: its significance in the phylogeny and evolution of ultrastructural systems. Cytobios 5, 59-77. POWELL, M. J. (1975). Ultrastructural changes in nuclear membranes and organelle associations during mitosis of the aquatic fungus Entophlyctis. Can. J. Bot. 53, 627-646. SALEMA, R. & BRANDAO, I. (1973). The use of PIPES buffer in the fixation of plant cells for electron microscopy. J. submicrosc. Cytol. 5, 79—96. TOMMERUP, I. C. & INGRAM, D. S. (1971). The life cycle of PlasmodiopJiora brassicae in Brassica tissue cultures and in intact roots. New Phytol. 70, 327-332. WHISLER, H. C. & TRAVLAND, L. B. (1973). Mitosis in Harpochytrium. Arch. Protistenk. «s. 69-74- WILLIAMS, P. H., AIST, S. J. & AIST, J. R. (1971). Response of cabbage root hairs to infection by Plasmodiophora brassicae. Can. J. Bot. 49, 41-47.

{Received 5 April 1979)