• The role of actomyosin and microtubules (MTs) in division plane establishment and cell plate expansion was studied in vacuolated cells of Coleus.
• Fluorescence and conventional microscopy were applied on control material and after treatment with anti-MT and anti-actin filament (AF) drugs as well as myosin inhibitors.
• The division plane is predicted by a cortical cytoplasmic ring (CCR), rich in AFs and endoplasmic reticulum, formed at interphase. The nucleus migrates to the CCR before its entrance in the phragmosome. The latter consists of transvacuolar cytoplasmic strands connecting the perinuclear cytoplasm with the CCR. During preprophase, a preprophase MT band is organized in the CCR. AF disruption and myosin inhibition destroy the CCR and cytoplasmic strands, arrest migration of the nucleus and affect cell plate expansion. MT disorganization inhibits anchoring of the cytoplasmic strands into the division plane.
• These observations support that CCR establishment, formation of transvacuolar cytoplasmic strands, migration of the nucleus and lateral cell plate expansion depend on actomyosin. The MTs guide and anchor the cytoplasmic strands of the phragmosome into the predetermined division plane.
Highly vacuolated higher plant cells can be induced to divide after tissue wounding. The determination of the division plane in this kind of cell has been studied in several plants, with emphasis in the formation of the phragmosome and the role of the preprophase MT band (Lloyd, 1991b). Since the first such study by Sinnott & Bloch (1940), it has been considered that a continuous sheet of cytoplasm, the phragmosome, is formed at the division site to suspend the dividing nucleus in a central cell area and to supply the necessary pathway for the developing cell plate (Lloyd, 1991b). However, the continuity, as a cytoplasmic sheet, of the phragmosome has been disputed for the case of long cambial cells (Goosen-de Roo et al., 1984) and its presence in even larger cells has not been examined.
The data about events that precede phragmosome formation are limited. Premitotic preparation and phragmosome formation have been described in detail for Nautilocalyx epidermal cells that divide after wounding (Venverloo & Libbenga, 1987; Venverloo, 1990). According to the above authors, phragmosome formation is preceded by a traumatotactic migration of the nucleus. Besides, Goodbody & Lloyd (1990) postulated that, in Tradescantia epidermal cells, which divide after wounding, the first sign of premitotic preparation is a change in the orientation of the cortical AF array. Accordingly, a belt-like array of AFs is formed, which is an early indicator of the future division plane.
The presence of cytoskeletal elements such as MTs and AFs in the phragmosome, as well as their contribution in its formation, have already been studied (Goosen-de Roo et al., 1984; Venverloo & Libbenga, 1987; Flanders et al., 1990; Goodbody & Lloyd, 1990; Goodbody et al., 1991; Lloyd et al., 1992). Phragmosome formation involves dramatic reorganization of the protoplast and, subsequently, major translocation processes need to take place. AFs have been reported to be responsible for the traumatotactic migration of the nucleus, while MTs seem to play a role in the entrance of the nucleus in a central cell site (Lloyd, 1991b). Myosin should contribute in at least the traumatotactic migration of the nucleus, but data do not yet exist.
In large, highly vacuolated cells, the distance between the initial site of cell plate formation and its fusion site with the parental cell walls is enormously long compared to that of meristematic cells. The cell plate has to expand and be guided to the exact predetermined cortical sites, very far away from its original formation area. The question ‘How does the cell plate push its way through vacuoles that in many shoot cells can be dozens of microns wide?’ has already been put (Cutler & Ehrhardt, 2002). For this process actomyosin should be a good candidate, as it has been proved very important for cell plate guidance in meristematic cells (Hepler et al., 2002; Molchan et al., 2002).
In the present study we examine the establishment of the division site and the guiding mechanism of the cell plate in huge vacuolated cells of the stem pith of Coleus spp. Cell division is induced by wounding (Sinnott & Bloch, 1940, 1941). The orientation of the cell division plane and the exact position of the daughter cell walls in this material are highly predictable, making it ideal for the study of division site establishment. Oryzalin and cytochalasin B (CB) are applied as anti-MT and anti-AF drugs, respectively, the effects of each one being well known and proven (Seagull, 1989; Morejohn & Fosket, 1991). For interfering with myosins 2,3-butanedione monoxime (BDM), a general inhibitor of myosin ATPase (Herrmann et al., 1992), and 1-(5-iodonaphtalene-1-sulphonyl)-1H-hexahydro-1,4-diazepine (ML-7), a specific inhibitor of myosin light-chain kinase (Saitoh et al., 1987) were used. These inhibitors have a different mode of action with BDM assumingly being less specific than ML-7. The latter interferes with the regulation of myosin II, a class of myosin that seems not to occur in plants, as may be deduced from the Arabidopsis genome (The Arabidopsis Genome Initiative, 2000). However, plant myosin VIII is very closely related to the animal class II myosin (Reichelt & Kendrick-Jones, 2000) and may thus possess a similar regulation pathway. Besides, a putative myosin II homologue has been identified in Zea mays (Šamaj et al., 2000). It is according to this view that ML-7 has been repeatedly used for experimental examination of the role of myosins in plant cell processes (Hepler et al., 2002; Molchan et al., 2002; Komis et al., 2003).
The observations support that, in pith cells of Coleus, the division site is ‘marked’ long before a phragmosome – if present – is formed and that the various cytoskeletal elements play discrete roles in the control of cell division. Hard evidence is provided for the contribution of actomyosin in the whole process.
Materials and Methods
Plant material and drug treatments
Coleus blumei Benth and Coleus pumilus Blanco plants, cultivated in the laboratory, were wounded on the second, third or lower shoot internodes with razor blades as described by Sinnott & Bloch (1940, 1941). Whole wounded shoot explants were put in test tubes to be treated with the drugs at room temperature, the treatment starting immediately (only for oryzalin) or 24 h after wounding. In all the cases, the wound area was not immersed in the treatment solution. Aqueous solutions of oryzalin (anti-MT, 10 µm) and CB (anti-AF, 100 µm) were prepared and applied as previously mentioned (Panteris et al., 1992). BDM (10 mm) and ML-7 (10 µm) were used as anti-myosin drugs according to Molchan et al. (2002). The duration of treatment and the number of experiments carried out were as follows: CB 24 h (> 10 experiments), BDM 24 h (10 experiments), ML-7 24 h (5 experiments), oryzalin 24–48 h (2 experiments). In addition, hand-cut sections of wounded internodes were treated with CB (200 µm), BDM (30 mm) and a combination (CB 200 µm and BDM 30 mm) on microscope glass slides, while being observed under the microscope. Five experiments were carried out with every inhibitor combination. The duration of such treatments never exceeded 3 h. Because of the size of the cells, the sensitivity of the vacuole and the thickness of the tissue in hand-cut sections special problems were faced in the preparation procedures, especially in keeping the cell shape intact. Fortunately, the tissue architecture and every cell's position in it were preserved.
For tubulin immunostaining of control material, hand-cut sections of the wounded internodes, parallel to the shoot axis, were prefixed for 20 min in 1% paraformaldehyde (PFA) in MT stabilizing buffer (MSB: 50 mm PIPES, 5 mm MgSO4 and 10 mm EGTA, pH 6.8) and then fixed in 4% PFA in the same buffer for 60 min. All the stages before and following were performed at room temperature unless otherwise stated. After rinsing with MSB, the cell walls were digested with a mixture of 1% cellulase, 1% pectinase and 2% driselase in MSB, pH 5.6, for 10 min and then, after rinsing with MSB pH 6.8, the sections were extracted with 5% DMSO, 0.05% Triton X-100 and 0.05% Igepal in MSB, pH 6.8, for 15 min. After rinsing, the sections were incubated for 45 min in anti-atubulin (YOL 1/34, Harlan Seralab, Loughborough, UK) antibody 1 : 40 in MSB and secondly for 45 min in FITC-anti rat (Sigma, St Louis, MO, USA) 1 : 40 in MSB at 37°C. Finally, the sections were washed with MSB, pH 6.8, and mounted in antifade solution.
For AF localization, hand-cut sections of control material, parallel to the shoot axis, underwent AF stabilization with 100 mmm-maleimidobenzoyl-N-hydroxysuccinimide ester (MBS) in MSB with 0.1% Triton X-100 plus 2% DMSO. Afterwards, they were fixed in two stages, 20 min in 1% PFA and 1 h in 4% PFA in MSB. After rinsing with MSB, the cell walls were digested with 1% cellulase and 0.5% pectinase in the same buffer for 10 min. In the stabilizing solution, as well as in the fixative and the enzyme solutions 1% (v/v; diluted from a stock solution of 300 units in 1.5 ml methanol) Alexa-Fluor 488 phaloidin (Molecular Probes, Eugene, OR, USA) was added. After 30 min extraction with 1% Triton X-100 and 5% DMSO in MSB, the AFs were stained with 10% (v/v) Alexa-Fluor 488 phalloidin in MSB for 1 h.
Light and electron microscopy
Preparation, sectioning and staining of control material for light microscopy and transmission electron microscopy (TEM) were performed according to Galatis, 1988, except for fixation, which followed the Karnovsky procedure. After aldehyde and osmium tetroxide fixation, the material was dehydrated in an acetone series and embedded in Spurr's resin. The cell shape was distorted in certain cases because of embedding difficulties.
Examination of specimens
Hand-cut sections of living tissue from control and drug-affected plants, parallel or normal to the shoot axis, were viewed with a Zeiss Axioplan microscope equipped with DIC optics, or after vital staining of endoplasmic reticulum with a 0.3% aqueous solution of DiOC6 (Quader & Schnepf, 1986). Living cells, cells stained for AFs, and semi-thin sections of fixed tissue stained with 1% Toluidine blue in 1% sodium tetraborate solution, were examined with a Zeiss Axioplan microscope equipped with an epi-fluorescence system and photographed on Agfa APX 100 or Kodak T-Max 400 film rated at ISO 1600. Ultrathin sections were examined with a Philips 300 TEM. Immunofluorescent specimens were examined with a Confocal-Laser-Scanning-Microscope (CLSM, TCS-4D; Leica Microsystems, Bensheim, Germany). In many cases it was not possible to scan whole cells, but only parts of them, because of the cell size. Unless otherwise stated, the cells depicted in this study were almost fully scanned.
In general, every hand-cut section of the wound area comprises 30–60 cells close to the wound. More than 200 such sections taken from 50 wounded internodes were examined. The observations that are described in the results were uniformly repeated in all the cells that were examined.
The nondividing pith cells are large, vacuolated, with a flat nucleus resting anywhere in the cortical cytoplasm. The cell size depends on the distance of the cells from the apical meristem, the longer cell dimension varying between 150 and 250 µm, or even more. The first cell divisions occurred 24–72 h after wounding, depending on the distance of the wounded internode from the apical meristem. The new cell walls produced by these wound-induced cell divisions were strictly parallel to the surface of the wound (Fig. 1). Examination of at least 200 sections of the wounded tissue showed that the division plane for all the cells in the wound area was quite predictable since it was always parallel to the wound surface. At the same time, the cells adjacent to the wound grew towards the wound surface, so that the wound became less deep. By the end of this first regeneration step, about 6–14 d after wounding depending on the distance of the wounded internode from the shoot apex, new cell files were produced (Fig. 1), connecting the severed vascular bundle edges (Warren Wilson & Warren Wilson, 1984).
Cell division in control cells
In wound-induced dividing pith cells of Coleus the first structural step of preparation for cell division was the establishment of the division site. It was initially manifested by the formation of a fine cytoplasmic band between cell wall and the vacuole (Figs 2a–l and 3a), combined with a shallow constriction of the tonoplast (Fig. 3a). This band, rich in AFs (Fig. 2g,h) and endoplasmic reticulum elements (Figs 2e,f,i,j and 3a), surrounded the cell periphery as a ring (cortical cytoplasmic ring, CCR) at the future division plane. It was always parallel to the wound surface (Fig. 2a–l; note the arrangement of the newly formed cell wall shown by the arrowhead in Fig. 2c). The CCR was commonly observed in more than 200 living cells with DIC optics (Figs 2a–d,l and 11a,f) and after staining with DiOC6 (Fig. 2e,f,i,j) and could also be observed in sections of fixed material (Fig. 3a). Once established, the CCR seemed to surround the whole cell periphery (Fig. 2a–c) and persisted at its original site throughout cell division (Figs 9a,b and 11a,f). The CCR was formed during interphase and, while the cell advanced to mitosis, the space it occupied between the tonoplast and plasmalemma became larger (Fig. 2i,j).
After CCR establishment, the nucleus migrated slowly in the cortical cytoplasm until it arrived to the CCR (Figs 4a–c and 5). The fact that CCR formation preceded migration of the nucleus was testified by continuous following of 30 living pith cells with DIC optics for 10–15 h. In these cells, a CCR was already formed while migration of the nucleus started later. The observation according to which the nucleus was about to start migrating was a change in its shape. Instead of being flat, it became spherical as it looked with DIC optics (Fig. 4a–d; cf Fig. 4f–i). Examination of 20 cells with TEM showed that every nucleus that was migrating became more or less amoeboid (Fig. 4e; cf Fig. 4e inset). By the beginning of and during the migration of the nucleus, transvacuolar cytoplasmic strands formed (Figs 4f–i, 5 and 8a), some of them connecting the perinuclear cytoplasm with the CCR (Figs 2e,f and 5b). The cytoplasmic strands were generated like tubular rods, penetrating laterally the vacuole in a furrowing pattern, with one end to the perinuclear cytoplasm and the other to any opposite area of the cortical cytoplasm (Fig. 4f–i).
After the nucleus arrived to the CCR, more and more ‘antenna-like’ perinuclear cytoplasmic strands arose, radiating from the nucleus and connecting it with several sites of the cell cortex through the vacuole (Figs 2h and 5c). In all the above stages the cytoplasmic strands were moving laterally, anastomosing and interconnecting to each other and to the cortical cytoplasm, and forming lateral branches (Fig. 4f–i). Cytoplasmic streaming was prominent in the strands, where organelles and quantities of cytoplasm were circulating. The nucleus moved from the cell cortex, at the CCR, suspended by perinuclear cytoplasmic strands, that later would form the phragmosome, and entered into an inner cell location (Fig. 5d). The site occupied by the nucleus could be central or eccentric inside the division plane (Figs 2k, 4d, 5d, 6b and 7a). In the latter case, the nucleus divided at a site near to the cell periphery (Fig. 9a,b). The nucleus always started to quit the cortical cytoplasm after it arrived and settled in the CCR (Figs 4a–d and 5c,d). No nucleus was observed to leave the cortical cytoplasm from other sites but the CCR. In all the above stages, the cortical and perinuclear cytoplasm, as well as the cytoplasmic strands, contained MTs (Figs 7b and 8a), AF bundles (Figs 2g,h and 6a,b) and endoplasmic reticulum elements (Figs 2k and 7b).
During migration, the nucleus exhibited interphase organization (Fig. 4e), while the cortical MTs form an interphase array. In the cell shown in Fig. 8a the nucleus had already left the cell cortex and cortical MTs still show interphase organization. When the nucleus that had already left the cell cortex entered preprophase, a preprophase MT band (Fig. 8b) was organized inside the CCR (Fig. 3b). Preprophase MT band organization in cells with cortically located nucleus was never observed. Instead of this, several interphase cells with the nucleus centrally located have been observed (Fig. 7a). The preprophase band of MTs was present in all the preprophase/prophase cells that were examined (6 cells with TEM, 55 cells with CLSM).
The preprophase MT band was not uniformly organized, exhibiting one or more gaps, depending on the cell size (Fig. 8b,c). Only in small cells was the preprophase MT band more complete, although it exhibited differences, between the different cell faces, in the number of MTs that it comprised. It was broad and, although its width was gradually reduced (Fig. 8b; cf Fig. 8c), it did never become quite narrow, as is the rule for meristematic cells (Mineyuki, 1999), even in cells at advanced prophase (Fig. 8c). Initially, the preprophase MT band was broader than the CCR, but gradually, as it was becoming tighter, it was restricted in the width of the CCR (Figs 3b and 8c). In preprophase/prophase pith cells, AFs remained inside the CCR, forming a preprophase AF band (Fig. 6c,d). As preprophase/prophase proceeded, most of the perinuclear cytoplasmic strands were gradually confined into the division plane, as defined by the width of the preprophase MT band (Fig. 8b,c), and thus the phragmosome was organized. The MTs in the strands extended from the perinuclear cytoplasm to the preprophase MT band (Fig. 8b,c). Motility of the cytoplasmic strands was reduced during this stage, as they composed the phragmosome.
In preprophase as well as in mitotic and cytokinetic pith cells of Coleus, a phragmosome, like that described in other dividing vacuolated cells (see Introduction), was not observed. In the cells studied here, the phragmosome seemed to be represented by the CCR and the transvacuolar cytoplasmic strands around the nucleus and the spindle, which were aligned in the division plane (Figs 8b,c, 9a,b and 11a,b). Apart from the cytoplasmic strands inside the division plane, polar cytoplasmic strands could sometimes be observed as well (Figs 8b,c and 11b). During metaphase, anaphase and telophase/early cytokinesis, no MTs could be observed anywhere in the cell, apart from the mitotic spindle and the newly formed phragmoplast (Fig. 8d), while AFs and endoplasmic reticulum elements were present in the perinuclear cytoplasm and cytoplasmic strands.
Cytokinesis was carried out by a typical phragmoplast/cell plate system. MTs, AFs and endoplasmic reticulum elements were present in the phragmoplast (Figs 8d and 10a,b). As a result of the eccentric position of the dividing nucleus (see Fig. 9a,b), cell plate expansion was not symmetrically proceeding. The cell plate fused with the parental cell wall on one cell side, while an active phragmoplast could still be observed at the other side, several microns away from the parent cell wall (Figs 10c,d and 11g,n), as has been mentioned for other vacuolated dividing cells (Cutler & Ehrhardt, 2002). As the phragmoplast and cell plate were expanding, a continuous sheet of cytoplasm between cell plate margins and the cortical cytoplasm was not observed (Figs 9c,d and 10c,d) and the phragmosome still consisted of a few cytoplasmic strands connecting the margins of the cell plate with the CCR (Figs 10a,b and 11g). All the above cytoplasmic strands were co-aligned inside the division plane. Few or no MTs at all could be observed in the cytoplasmic strands (Fig. 8d), which contained endoplasmic reticulum elements and AFs (Fig. 10a). In large dividing cells, the daughter nuclei started to migrate to the new division sites, which were parallel to the previous ones, while cytokinesis was not yet accomplished (Fig. 10c,d).
Cell division in drug-affected cells
Short-term treatments Short-term treatments were carried out while the cells in the sections of wounded tissue were alive and observed under the microscope with DIC optics (see Materials and Methods). A prominent effect of CB, BDM or the combination of CB and BDM on the cells was the arrest of cytoplasmic streaming. The combined treatment with CB and BDM resulted in very fast (10–20 min) destruction of the CCR in cells dividing (Fig. 11c,e; cf Fig. 11a) or preparing for cell division. Treatment with BDM alone also resulted in CCR destruction but at a slower rate (60 min, Fig. 11h; cf Fig. 11f). The CCR cortical site exhibited a ‘bubble’ like image after demolition of the CCR (Fig. 11h). At the same time, the cytoplasmic strands were strongly affected (Fig. 11d; cf Fig. 11b, Fig. 11i; cf Fig. 11g). The above results were uniform for all the cells examined, with no exceptions within the same section of wounded tissue.
Cytokinesis was dramatically affected in cells dividing in the presence of BDM or CB. Within 0.5–2 h, the cytoplasmic strands diminished (Fig. 11i; cf Fig. 11g, Fig. 11l,m; cf Fig. 11j,k). At the same time, the cell plate expansion to the division site was disturbed. In cells with centrally positioned cell plate, the latter was curved and, after fusion of the edges, it was finally encaged inside the vacuole (Fig. 11j–m). In cells where the cell plate was initially eccentrically positioned, one edge of it fused with the closest parent cell wall, while the other either remained encaged inside the vacuole (Fig. 11n,o) or was strongly curved and directed to the closest parent cell wall (Fig. 11i). All of these observations show that anti-AF (CB) and anti-myosin (BDM) treatment affected normal cell plate expansion, strongly involving actomyosin in the progress of cytokinesis.
Long-term treatments After 24 h treatment with CB, no cytoplasmic streaming could be observed in the cells. In cells close to the wound a CCR and/or transvacuolar cytoplasmic strands were not observed, while their nuclei might be resting at any site of the cortical cytoplasm. CB did not inhibit cell division, but affected the shape and orientation of the daughter cell walls. In cells that had been dividing during CB treatment the daughter cell walls were deposited quite abnormally, not conforming to the expected division plane (Fig. 12a–c; cf Fig. 1). Incomplete, rounded, curved and branched daughter cell walls, as well as spherical daughter cell compartments inside parent cells, could be observed (Fig. 12a–c). The latter obviously occurred from distorted cell plates similar to that of Fig. 11j–m.
Cells treated with BDM for 24 h did not show any sign of cytoplasmic streaming. No CCR and/or any cytoplasmic strands could be observed in these cells 48 h after wounding (Fig. 12d). The majority of the cells did not divide at all. Their nuclei were immobile, lying at any cortical cell site (Fig. 12d). In few cells, which had been already dividing at the beginning of BDM treatment, abnormally deposited, curved daughter cell walls could be observed (Fig. 12e) not conforming to the expected division site (Fig. 12e; cf Fig. 1). Such cell walls were produced by distorted cell plates like that depicted in Fig. 11i.
The effect of ML-7, after a 24 h treatment, was similar to that of BDM. Most cells did not divide at all while those that had been dividing during treatment exhibited abnormally positioned daughter cell walls (Fig. 12f). Sometimes, one cell in the whole wound area had some immobile cytoplasmic strands.
Treatment with oryzalin for 24–48 h did not interfere with the advance of pith cells to divide, but affected mitosis and cytokinesis resulting to binucleate/multinucleate or polyploid cells (Fig. 12i). Oryzalin did not interfere with polar cell growth towards the wound surface. As a result, long undivided cells were formed in the wound area (Fig. 12i). The drug did not also interfere with the formation of the CCR and cytoplasmic strands as well as with the migration of the nucleus. Cells with a CCR (Fig. 12g,h), cytoplasmic strands and an almost centrally positioned nucleus could be observed in sections of shoots treated with oryzalin 24 h after wounding (Fig. 12j). However, these cytoplasmic strands never gathered into the division plane to make a phragmosome (Fig. 12j) but remained in lateral gliding motion, in any direction, faster than in control cells.
In certain cases, the wounded shoots were treated with oryzalin immediately after wounding (see Materials and Methods). In sections of these shoots, cells with a centrally located nucleus, surrounded by antenna-like cytoplasmic strands, could be observed (Fig. 11p). Cells undergoing abnormal mitosis could also be observed (Fig. 11q), in which, although during mitosis, the cytoplasmic strands that suspended the area occupied by the chromosomes were not aligned to form a phragmosome, but were radiating in various directions (Fig. 11q; cf Figs 9a and 11b).
Our studies on division site establishment in huge vacuolated cells of Coleus after wounding revealed striking differences, compared to previous studies in vacuolated cells of other plant species. The following novel observations have been made: (i) A CCR, rich in AFs and endoplasmic reticulum elements, is the first visual sign of division site establishment. The CCR is formed at interphase and predicts the future cell division plane, before phragmosome and preprophase MT band organization. It persists during the whole process of mitosis and cytokinesis. (ii) The nucleus migrates first to the CCR and, only afterwards, it leaves the cell cortex and gets into a more or less central cell area where it divides. (iii) The phragmosome, which is organized at the plane defined by the CCR during the onset to preprophase, consists of aligned cytoplasmic strands, which never completely fuse to form a single cytoplasmic sheet. (iv) The only contribution of the MTs to the establishment of the division site is to align the cytoplasmic strands of the phragmosome into the division plane. (v) Creation of the CCR, formation and motility of the cytoplasmic strands, migration of the nucleus and guidance of the cell plate depend on actomyosin.
Establishment and function of the division site seems to occur by the following steps: (i) The division site is determined at interphase, as manifested by the presence of the CCR, and, consequently, the site of future preprophase MT band organization is ‘marked’. This differs from the meristematic cells, where division site establishment occurs during preprophase/prophase (Gunning, 1982; Mineyuki, 1999). (ii) The cytoplasmic strands that radiate from the nucleus are gathered and aligned at the division plane via their connection with the preprophase MT band. (iii) The cytoplasmic strands, permanently restricted into the division plane, determine the pathway for cell plate expansion during cytokinesis.
Division site determination
According to previous studies, the division site of vacuolated plant cells is determined by the formation of the phragmosome and/or the preprophase MT band organization (see Introduction). The present study reveals that determination of the division site in Coleus pith cells occurs at interphase, far before the phragmosome or preprophase MT band are organized. The first structural sign for its establishment is the formation of the CCR, which seems to be the result of the local actomyosin organization, being extremely rich in AFs. CCR establishment strongly supports that division site determination is related to cortical cytoplasm and/or plasmalemma polarization (Gunning, 1982; Gunning & Wick, 1985; Wick, 1991; Mineyuki, 1999; Pickett-Heaps et al., 1999; Cleary, 2000).
A CCR is never observed in cells affected by CB and anti-myosin drugs. This observation suggests a possible role for actomyosin in CCR formation and, subsequently, in early division site establishment and its manifestation. Besides, in Tradescantia epidermal cells, induced to divide by wounding, cortical AFs are the first cytoskeletal element to respond to wounding (Goodbody & Lloyd, 1990). In these cells, the cortical AFs are reoriented, resulting in a belt-like array of fine cortical AFs, which is an early indicator of the future division plane (Goodbody & Lloyd, 1990). Although this observation is not directly related to CCR organization, it is consistent with the involvement of actomyosin in establishment of the division plane. The belt-like AF array might be considered as a first step for consequent CCR formation. Apart from this, formation of the CCR can be considered as an evolutionary homologue to the furrowing of lower plant cells, as discussed for AFs at the division site by Pickett-Heaps et al. (1999).
On the other hand, it seems that cortical MTs do not participate in division site establishment, as they do not alter their distribution and orientation throughout the pre-mitotic migration of the nucleus (see also Goodbody et al., 1991), until they organize into a preprophase band of MTs. Besides, MT disassembly by oryzalin does not affect CCR formation. According to our results and previous observations (Goodbody & Lloyd, 1990), actomyosin may play a pivotal role in division site establishment in vacuolated cells, while in meristematic cells the cortical MTs seem to play this role (Gunning, 1982; Gunning & Wick, 1985; Wick, 1991; Mineyuki, 1999; Galatis & Apostolakos, 2004). The possible factors that induce division site determination in vacuolated cells are not revealed in the present study and only speculations can be made. However, our observations point to a strongly localized interaction between the plasmalemma and the cortical AFs, which may be the result of a site-dependent gradient of calcium ions, due to mechanical stress (see Lintilhac & Vesecky, 1981; Goodbody & Lloyd, 1990), or to specific changes regarding plasmalemma lipid composition, e.g. changes in phospholipid distribution.
An interesting observation is that long-term treatment with BDM and ML-7 inhibits cell division in the wound area of Coleus stem. On the other hand, CB and oryzalin do not affect the advance of the cells to divide. CB only interferes with division site establishment and cell plate expansion, while oryzalin interferes with the processes of mitosis and cytokinesis. It seems that induction for cell division after wounding is affected by inhibition of myosin function. This observation suggests a role for the myosins in the transduction of the stimulus for cell division. It has been supported that cell division after wounding is induced by factors, such as hormones, transferred from the severed vascular bundle edges to the wound area (Warren Wilson & Warren Wilson, 1984). The possible role of myosins in the above procedure remains to be studied.
Cytoplasmic strand formation-nucleus migration
The response of the cells, regarding future cell division, is obviously different in wounded Coleus stem from what has been described in Tradescantia leaves (Schnepf & von Traitteur, 1973) and Nautilocalyx explants (Venverloo et al., 1980; Venverloo & Libbenga, 1987). The nucleus slides through the cortical cytoplasm to the CCR in Coleus pith cells (Fig. 5), before moving to an almost central cell position for mitosis, instead of migrating towards the wound surface.
AFs have been shown to participate in traumatotactic migration of the nucleus in Tradescantia leaves (Schnepf & von Traitteur, 1973; Goodbody & Lloyd, 1990) and Nautilocalyx explants (Venverloo & Libbenga, 1987). However, in the latter material migration of the nucleus to the centre of the cell seems to depend on MTs and not on AFs (Venverloo & Libbenga, 1987). Similar observations have also been made for BY-2 suspension cells (Katsuta et al., 1990). As in other dividing vacuolated cells (Goosen-de Roo et al., 1984; Goodbody & Lloyd, 1990), the cytoplasmic strands and perinuclear cytoplasm of Coleus pith cells contain AFs while preparing for cell division. In these cells CB, as well as anti-myosin drugs, affect the establishment of the division site, while anti-MT treatment (oryzalin) has no effect on it. In the absence of MTs the nucleus enters into a phragmosome-like structure (see Fig. 11p,q). In Coleus pith cells the whole premitotic preparation, i.e. the formation and motility of the cytoplasmic strands, the migration of the nucleus to the CCR and its entrance to the developing phragmosome, depend on actomyosin and not on MTs.
Apart from AFs, the CCR as well as the cytoplasmic strands are especially rich in endoplasmic reticulum elements. A possible relationship between actomyosin and the distribution of endoplasmic reticulum can be supported. It has already been suggested that AFs parallel to tubular endoplasmic reticulum elements (Hepler et al., 1990) contribute in the spatial organization of the endoplasmic reticulum in vacuolated cells (Quader et al., 1996). The possible role of the endoplasmic reticulum in the organization and stabilization of the CCR and cytoplasmic strands in Coleus pith cells remains unclear.
Phragmosome-preprophase band organization
In the dividing large pith cells of Coleus the cytoplasmic strands never form a continuous phragmosome, as has been mentioned for other dividing vacuolated cells (Sinnott & Bloch, 1940, 1941; Goosen-De Roo et al., 1984; Venverloo & Libbenga, 1987). The phragmosome-like structure of the cells studied here consists of several cytoplasmic strands, aligned within the division plane, which do not fuse to a single cytoplasmic sheet as has been described for Nautilocalyx epidermal cells (Venverloo & Libbenga, 1987). This difference may be due to the size of the pith cells of Coleus, as the enormous surface of the division plane does not allow the existence of a continuous cytoplasmic diaphragm like that described in other plant material (see Introduction).
The cell size may also be the reason for the organization of an incomplete preprophase MT band. Similar observations have been made in bryophyte cells (Apostolakos & Galatis, 1992) and, in that case, the cell size in relation to the amount of tubulin is considered to be responsible for the incompleteness of the preprophase MT band (Apostolakos & Galatis, 1985, 1992). This may be the case for the large pith cells of Coleus as well. In smaller vacuolated cells, like those of Nautilocalyx explants, the preprophase MT band is complete (Goodbody et al., 1991).
The preprophase MT band seems to be the last cytoskeletal structure that appears at the division site. Its position is coincident to the CCR. The role of the preprophase MT band seems to be restricted in the permanent co-aligning and anchoring of the cytoplasmic strands into the division plane to form a phragmosome: it is by this procedure that the preprophase MT band premitotically determines the plane of cell plate development during cytokinesis and the site where the daughter cell wall will fuse with the parent cell wall. Failure of the cytoplasmic strands that radiate from the nucleus to be restricted into the division plane, during oryzalin treatment, further supports such a role for the preprophase MT band. It is important that the preprophase Mt band plays the role mentioned above during its presence, apart from any possible role of its region during completion of cytokinesis (Mineyuki & Gunning, 1990; Mineyuki, 1999; Galatis & Apostolakos, 2004). MTs in the cytoplasmic strands, interconnecting the perinuclear cytoplasm with the preprophase MT band, have been reported to play a role in the ‘bunching’ process that creates the phragmosome (Flanders et al., 1990; Lloyd, 1991a; Lloyd et al., 1992). In Coleus pith cells, it seems that MTs in the cytoplasmic strands play a similar role. Apart from guiding the cytoplasmic strands, no other role can be deduced from the effects of oryzalin treatment. This is also supported by the observation that, during mitosis in control cells, MTs are restricted in the spindle while the cytoplasmic strands of the phragmosome are stabilized by the AF bundles that they contain. The ability of AFs alone to tether the cytoplasmic strands has been also shown in BY-2 suspension cells (Katsuta & Shibaoka, 1988).
Cytokinesis-cell plate expansion
Cytokinesis is accomplished, in Coleus pith cells, by a typical phragmoplast consisting of MTs, AFs and endoplasmic reticulum elements. According to Valster & Hepler (1997), cytokinesis can be distinguished into two stages, one of early cell plate formation between the daughter nuclei, and the other of so-called ‘late lateral expansion’ beyond the width of the nuclei to the parental cell wall. In the vacuolated cells of Coleus the second stage is long and the cell plate has to grow a great distance until it approaches the parent cell wall. The permanent restriction of the cytoplasmic strands into the division plane, because of their premitotic connection with the preprophase MT band, assures accurate guidance that is necessary for the cell plate to grow to the exact predetermined division site. No other plane for cell plate expansion is permitted but the one occupied by the transvacuolar cytoplasmic strands. Besides, in the large vacuolated cells of Coleus the phragmoplast/cell plate system requires a cytoplasmic ‘front’ to ‘pave the way’ for the expanding cell plate in order to advance through the vacuole towards the cell cortex. It could be well supposed that this cytoplasmic ‘front’ is the result of forces that gradually pull the tonoplast away from the borders of the cell plate (Fig. 13a,b). Creation and application of these forces as well as transport of cytoplasm through the cytoplasmic strands to the centrifugally expanding cytoplasmic ‘front’ can be attributed to actomyosin (Fig. 13a,b).
Actomyosin has already been shown, in meristematic cells, to be very important for guiding the cell plate to the division site (Pickett-Heaps et al., 1999; Cleary, 2000; Hepler et al., 2002; Molchan et al., 2002). In the cells studied here, the cytoplasmic strands that extend from the cell plate border may or may not contain MTs, but they always contain AFs. Apart from this, AFs connecting the phragmoplast to the cortical division site have already been observed (Valster & Hepler, 1997) and the cytoplasmic strands of vacuolated cells have been shown to be under tension (Goodbody et al., 1991). It might well be through this tension, due to the AFs inside the cytoplasmic strands, that the latter provide the space for cell plate expansion towards the division site. The borders of the tonoplast between cytoplasmic strands may be pulled towards the cell cortex by a ‘tug of war’-similar traction, created by actomyosin function (Fig. 13a,b). Because of actomyosin traction, the vacuole is continuously reorganizing. Reorganization of the vacuole during cell division has already been observed in living BY-2 cells (Kutsuna & Hasezawa, 2002).
Experimental inhibition of actomyosin function further supports the above opinion. Similarly to what has been observed in other plant material (Venverloo & Libbenga, 1987; Molchan et al., 2002), in Coleus pith cells disruption of AFs by CB and treatment with inhibitors of myosins result in very abnormally deposited daughter cell walls (Fig. 12a–f), underlining the significance of actomyosin for cell plate expansion. In the absence of AFs and/or by the arrest of myosin function, the phragmoplast/cell plate system has no ability to advance through the area occupied by the vacuole (Fig. 11f–o) and, subsequently, grows abnormally towards the closest parent cell wall site or is encaged inside the vacuole.
We thank Dr M. Zachariadis for his help in the digital processing of the images from the CLSM and preparing the plates of this article. This work was supported by a grant from the State Scholarship Foundation of Greece (program IKYDA-2002).