Cortical actin filaments at the division site of mitotic plant cells: a reconsideration of the ‘actin-depleted zone’


Author for correspondence:
Emmanuel Panteris
Tel: 2310 998908
Fax: 2310 998389


The preprophase bands of microtubules and F-actin are primary markers of the division site for most plant cells. After preprophase band breakdown, the division site has been considered to be ‘negatively’ memorized by the local absence of cortical actin filaments. However, there have been reports of cortical F-actin at the division site of mitotic plant cells, calling into question its distribution and possible role there. In this article, previous and recent data on this issue are reviewed. It is proposed that the division site of mitotic plant cells is not devoid of F-actin but is traversed by scarce cortical actin filaments. The description of the division site as an ‘actin exclusion zone’ might therefore be attributed to a failure to preserve and/or image the notoriously sensitive actin filaments. Accordingly, the ‘actin-depleted zone’ should be considered as a site with fewer actin filaments than the rest of the cortical cytoplasm. Taking into account recent molecular data on division site components, a possible role for the scarcity of cortical actin filaments in establishing a zone of minimum mobility is also proposed.


The determination of the division site (the cortical site of new cell wall insertion) in plant cells is of major importance, as tissue patterning and subsequent differentiation rely on proper insertion of new cell walls. The preprophase band of microtubules (for reviews see Mineyuki, 1999; Geelen & Inzé, 2006; Van Damme et al., 2007), actin filaments (for reviews see the same references) and, in certain cell types, endoplasmic reticulum (Zachariadis et al., 2001, 2003) are the main manifestations of division site establishment, marking the exact site of daughter wall insertion. However, by the end of prophase, the preprophase band disappears and there is no visible sign of what remains to mark the division site during the remaining cell division stages (for reviews see Mineyuki, 1999; Geelen & Inzé, 2006; Van Damme et al., 2007). Although mitotic plant cells lack cortical microtubules after preprophase band breakdown, F-actin has been shown to remain in the cortical cytoplasm throughout all the stages of cell division (among others, see Seagull et al., 1987; Traas et al., 1987; Panteris et al., 1992).

In 1992, the division site of mitotic plant cells was described as being devoid of F-actin (Cleary et al., 1992; Liu & Palevitz, 1992). According to these descriptions, a zone free of actin filaments, termed the ‘actin-depleted zone’ (Liu & Palevitz, 1992) or the ‘actin exclusion zone’ (Cleary et al., 1992), appears at the division site by late prophase/prometaphase and persists until telophase. It has been assumed that, because of its localization at the site of the former preprophase band of microtubules and F-actin, this zone is a ‘negative marker’ of the division site, possibly playing a role in the ‘memory’ of the predetermined division site.

Since then, this view has not been challenged and the presence of the actin-depleted zone was confirmed in a variety of cell types of various plant species (e.g. Cleary, 1995; Baluška et al., 1997; Vitha et al., 2000; Hoshino et al., 2003; Voigt et al., 2005). The view that such a zone really exists and marks the division site has already been included in several review articles (among others, see Mineyuki, 1999; Baluška et al., 2000; Mathur, 2004; Blancaflor et al., 2006; Geelen & Inzé, 2006; Van Damme et al., 2007), and this zone is generally accepted as a universal feature. However, the presence of such a zone was not mentioned in several articles (Panteris et al., 1992, 2004; Collings et al., 2001), while F-actin at the division site was clearly reported in certain cell types (Sano et al., 2005; Panteris et al., 2007).

In order to reconsider the presence and possible role of the actin-depleted zone, previous and recent data are comparatively reviewed and critically evaluated. The conclusion is that cortical actin filaments are scarce but still present at the division site throughout mitosis, while they are abundant adjacent to the spindle poles. Consequently, the total absence of F-actin from the cortical division site is considered as an artefact and the possible reasons for it are discussed. According to this view, ‘depletion’ of F-actin at the division site should be considered as a scarcity, but not as ‘exclusion’, of F-actin. A positive role for actin filaments at the division site of mitotic plant cells is also proposed.

F-actin preservation and imaging: what is the true image of actin filaments?

Before any other discussion about F-actin distribution in plant cells, one should focus on the notorious vulnerability of F-actin to imaging protocols (Vitha et al., 2000; Van Gestel et al., 2001). Aldehyde fixation is generally thought to damage F-actin (Sonobe & Shibaoka, 1989; Collings et al., 2001), although Vitha et al. (2000) reported that it is not aldehyde fixation that is to blame for F-actin deterioration. Phalloidin has been reported to miss or even destroy several classes of actin filaments in fixed cells (Meagher, 1991; Andersland & Parthasarathy, 1992; Cleary, 1995) and to be suitable for F-actin staining only in living cells (Tang et al., 1989; Vitha et al., 2000). However, in general, immunostaining reveals fewer F-actin arrays in fixed cells than fluorescent phalloidin. Actin filaments in the spindle, for example, have been found in several studies in both unfixed and fixed (Seagull et al., 1987; Traas et al., 1987; Panteris et al., 1992; Cleary, 1995) mitotic cells with phalloidin staining or a combination of phalloidin and immunostaining (Schmit & Lambert, 1987) but not with immunostaining alone. Either immunostaining misses some actin filaments or the F-actin patterns found with phalloidin are artefactual. In fact, phalloidin staining in living cells is suspected to produce over-polymerization of F-actin, which has been considered a disadvantage for live-cell imaging (Cleary et al., 1992; Cleary, 1995; Valster & Hepler, 1997; Voigt et al., 2005).

A reasonable compromise between the above options has been considered to be staining with phalloidin after fixation, in cells pre-treated with the stabilizing cross-linker m-maleimidobenzoyl N-hydroxysuccinimide ester (MBS) (Sonobe & Shibaoka, 1989). This practice has offered credible imaging of F-actin in a variety of cell types of various plant species (e.g. Panteris et al., 1992, 2007; Collings et al., 2001). However, MBS has occasionally been blamed for ‘over-stabilizing’ (Collings & Wasteneys, 2005), suggesting a shortening of the pre-treatment time or even its omission.

Live-cell imaging of F-actin by green fluorescent protein (GFP)-tagged actin-binding proteins seems to be a promising approach. Most surprisingly, however, even data from studies on plants transformed with the very same GFP-actin-binding protein reporter appear to be contradictory. For example, actin filaments in the preprophase band were prominent in BY-2 cells (Sano et al., 2005; Higaki et al., 2007) but were not found in Arabidopsis thaliana root cells (Voigt et al., 2005; Wang et al., 2008), although the same actin-binding domain of fimbrin was used as the reporter.

Escaping from the above labyrinth of technical options, one needs to define what a ‘fairly good’ actin image should be. In Vitha et al. (2000), for example, several protocols are described as providing equally good results. However, the actin images of this article are rather fuzzy (fig. 1 in Vitha et al., 2000), and are at least less sharp than those achieved with MBS and phalloidin protocols (compare with figures in Panteris et al., 1992, 2007; Collings et al., 2001; this article) or those obtained with GFP-fimbrin in BY-2 cells (Sano et al., 2005; Higaki et al., 2007). The major criterion for evaluating an actin-staining protocol is the quality of actin filaments depicted. In particular, the finer filaments should appear well resolved, neither bundled nor fragmented.

In order to obtain such ‘fairly good’ F-actin images, special protocols have to be designed for certain plant materials. The protocol described by Lovy-Wheeler et al. (2005), for example, revealed high-quality F-actin images in pollen tubes, although the same protocol was not better than that with MBS and fixation for studies in leaf tissue (Panteris et al., 2006, 2007). However, the application of the above cross-linkers, MBS or ethylene glycol bis[sulfosuccinimidylsuccinate] (sulfo-EGS; Lovy-Wheeler et al., 2005), suggests that even more complex and dynamic F-actin arrays may still be expected to be discovered. Accordingly, bearing in mind the images in the above articles, MBS/sulfo-EGS and phalloidin protocols seem to be suitable for both living and fixed plant cells, producing images comparable to the excellent F-actin images obtained in tobacco (Nicotiana tabacum) BY-2 cells with GFP-fimbrin constructs (Sano et al., 2005; Higaki et al., 2007). In addition, the possibility of F-actin over-stabilization by MBS (Collings & Wasteneys, 2005) seems not to be significant for studies in several cell types of various plant species (Panteris et al., 1992, 2007). F-actin distribution is the same with or without MBS, although the quality of the actin image may be better or worse.

Actin filaments in mitotic plant cells: a long story of contradiction

How is F-actin distributed in mitotic plant cells? As already mentioned, contradictory answers to this question have been suggested regarding distribution within the following cytoskeletal arrays: the preprophase band, the mitotic spindle, and the phragmoplast. Although a preprophase band of F-actin is thought to coincide with that of microtubules (Kakimoto & Shibaoka, 1987; Palevitz, 1987; Traas et al., 1987; Zachariadis et al., 2001, 2003), there have been reports of cortical actin filaments not always following the pattern of the preprophase microtubule band (Cho & Wick, 1990, 1991; Panteris et al., 1992, 2007; Collings & Wasteneys, 2005; Wang et al., 2008). Similarly, while actin filaments have been found in the mitotic spindle (Schmit & Lambert, 1987; Seagull et al., 1987; Traas et al., 1987; Lloyd & Traas, 1988; Panteris et al., 1992; Cleary, 1995; Collings et al., 2001), their absence from the spindle has also been reported (Liu & Palevitz, 1992; Baluška et al., 1997; Vitha et al., 2000; Voigt et al., 2005; Wang et al., 2008). Although the phragmoplast has generally been accepted, since 1985 (Clayton & Lloyd, 1985), to include F-actin parallel to microtubules (see Fig. 1b,c), the length of phragmoplast actin filaments has also been a point of contention (Collings & Wasteneys, 2005). Besides, the F-actin of the phragmoplast was found by microinjection not to reach the cell plate (Cleary et al., 1992; Zhang et al., 1993; Cleary, 1995; Valster & Hepler, 1997).

Figure 1.

Confocal laser scanning microscope (CLSM) micrographs of a Triticum turgidum subsidiary cell mother cell at cytokinesis (large asterisk in a, b). The cell is shown at surface view. Bar, 25 µm. (a, b) Cortical (a) and median (b) optical sections. The cortical cytoplasm of the cytokinetic cell (a) exhibits uniformly distributed actin filaments. (b) The small asterisk denotes the inducing guard cell mother cell and the arrow indicates the phragmoplast. (c) Projection of 35 optical sections of the above cell. Note the absence of thick F-actin bundles from the cytokinetic cell, in comparison with their presence in the neighboring interphase cells (one of which is indicated by the arrow).

Concerning the actin-depleted zone, there have been contradictory reports about: (a) its duration, (b) its presence in certain cell types and (c) its real nature. Although the common view is that an actin-depleted zone occurs from prometaphase until early telophase, according to some authors it may appear with preprophase band narrowing (Liu & Palevitz, 1992) and persist almost until completion of cytokinesis (Cleary & Smith, 1998).

Although the actin-depleted zone is generally accepted as universal among mitotic plant cells and has been reported in the majority of relevant articles, its presence in certain plant cell types was not universally confirmed. Specifically, in mitotic cells of Zea mays some authors described actin-depleted zones (Baluška et al., 1997, 2000; Cleary & Smith, 1998; see also Vitha et al., 2000) while Panteris et al. (2007) found actin filaments at the division site throughout mitosis and cytokinesis (see also Fig. 2a–j). What is more, in BY-2 tobacco cells, Hoshino et al. (2003) and Yoneda et al. (2005) described actin-depleted zones, while Sano et al. (2005) found a ‘valley’ of a few actin filaments between two ‘twin peaks’ of F-actin aggregation. According to the latter authors, it is the ‘twin peaks’ that mark the division site rather than the ‘valley’ between them. The ‘twin peaks’ observation was also confirmed by Higaki et al. (2007), providing excellent images of GFP-fimbrin-decorated actin filaments.

Figure 2.

Confocal laser scanning microscope (CLSM) micrographs of Zea mays (a–d) and Triticum aestivum (e–j) subsidiary cell mother cells at surface view, and the Z. mays basal coleoptile area (k), depicting actin filaments after 15 min of stabilization with 300 µµ m-maleimidobenzoyl N-hydroxysuccinimide ester (MBS), 1 h of fixation with 4% paraformaldehyde and phalloidin staining according to Panteris et al. (2006, 2007). The same staining protocol was applied for all the micrographs of this article. Median (a, c, e, g, i) and cortical (b, d, f, h, j) optical sections of mitotic cells are shown (the arrows in the median optical sections indicate the spindle). The arrowheads in the cortical optical sections (b, d, f, h, j) indicate the division plane of the cells, which in (b) and (d) are lens-shaped as the cells divide to produce the subsidiary cells (note also the F-actin patch at one spindle pole in (a) and (c)). Actin filaments are not absent from the cortical division site but line the whole surface of the periclinal walls (b, d, f, h, j) of mitotic cells. Bars: (a–d) 15 µm; (e–f) 15 µm; (g–h) 20 µm; (i–j) 20 µm. (k) Projection of 30 optical sections depicting a cell dividing longitudinally (upper right) and a long cell dividing vertically to its long axis (lower left). In both cells, the arrows indicate the division plane. The arrowheads at the lower cell indicate the ‘twin peaks’-like longitudinal wall F-actin aggregations. Bar, 30 µm.

Last but not least, what does ‘actin-depleted zone’ mean? Is it a zone with fewer actin filaments than the rest of the cortical cytoplasm, as depicted by Zachariadis et al. (2001, 2003), or a zone of complete F-actin absence (e.g. Cleary et al., 1992; Liu & Palevitz, 1992; Cleary, 1995; Hoshino et al., 2003)?

The actin-depleted zone: is there any F-actin at the division site ‘valley’?

Since its discovery, this zone has been considered as an ‘actin exclusion zone’ (Cleary et al., 1992), implying that the division site is ‘devoid of actin filaments’ (among others, see Geelen & Inzé, 2006; Van Damme et al., 2007). At one extreme, the images shown by Baluška et al. (1997) and Vitha et al. (2000) suggest that mitotic plant cells are more or less devoid of F-actin, apart from the cell faces proximal to the spindle poles. In contrast, the ‘actin exclusion zone’ found by phalloidin microinjection is narrow, limited to the exact cortical division site (Cleary et al., 1992; Cleary, 1995; Valster & Hepler, 1997).

However, the term ‘depleted’ that has prevailed does not mean an absolute absence and appears to be much closer to reality. In older studies (Seagull et al., 1987; Traas et al., 1987; Panteris et al., 1992) the absence of F-actin from the division site of mitotic plant cells was not noted, while in more recent studies (Collings et al., 2001; Panteris et al., 2007), phalloidin has been shown to stain F-actin at the division site of several dividing plant cell types. The microinjection of phalloidin into mitotic endosperm cells (Schmit & Lambert, 1990) did not reveal an actin-depleted zone; the cell system, however, is very different from that of the majority of dividing plant cells, as they do not form a preprophase band. However, the microinjection procedures applied in Tradescantia virginiana cells (Cleary et al., 1992; Cleary, 1995; Valster & Hepler, 1997) may miss the finest dynamic actin filaments at the division site. Whatever the reason might be, the fact that microinjected phalloidin fails to stain the whole actin filament population has also been confirmed for pollen tubes (P. K. Hepler, Biology Department, University of Massachusetts, USA; pers. comm.).

Studies with reporters based on GFP-fimbrin have also provided evidence for the presence of F-actin at the division site throughout the whole cell cycle. In living BY-2 cells (Sano et al., 2005; Higaki et al., 2007), a ‘valley’ of actin filaments has been shown to traverse the division site between the ‘twin peaks’ of F-actin aggregation. According to the above articles, a population of actin filaments is present at the division site of BY-2 tobacco cells, which is not devoid of F-actin as reported previously (Hoshino et al., 2003; Yoneda et al., 2005). Earlier studies with GFP-fimbrin-transformed Arabidopsis thaliana plants reported the presence of an actin-depleted zone (Voigt et al., 2005). However, the images of dividing cells were poor (fig. 1f–i in Voigt et al., 2005), probably because of the small size of the cells and the low-magnification lens (×40) that was used. In contrast, in the large BY-2 cells, the image of actin filaments is fairly good and there is no doubt that they are present at the division site of dividing BY-2 cells. In summary, the presence of F-actin at the division site has been undoubtedly confirmed by MBS/phalloidin protocols (Panteris et al., 2007; figures in this article) as well as by GFP-fimbrin (Sano et al., 2005).

It might be supported that the presence of an actin-depleted zone in mitotic cells is a matter of cell type and species. According to this view, the divergence that has been described in this article might be a consequence of the varying importance of F-actin for division site memory among various plant cell types. However, there are at least two arguments against such a view. First, given the importance of division site demarcation for proper cytokinesis, the mechanism used to achieve it should be universal, at least to the extent of the occurrence of the preprophase band. Secondly, observations in even the very same cell types, for example tobacco BY-2 cells, have been found to be contradictory (Hoshino et al., 2003; Yoneda et al., 2005; compare with Sano et al., 2005; Higaki et al., 2007). Consequently, any variation in the presence of F-actin at the division site of plant cells should be attributed to the imaging protocol rather than the cell type. Either the actin-depleted zone is universally present among plant cells or it is universally absent. In conclusion, it seems that the division site of mitotic plant cells is traversed by fine actin filaments during the transition from metaphase to cytokinesis. Accordingly, what has been considered as an ‘actin exclusion zone’ may result from the failure to observe the above-mentioned actin filaments.

Beyond ‘peaks’ and ‘valleys’, a role for cortical actin filaments at the division site

According to experiments with anti-actin drugs, what matters for the establishment and consolidation of the division site is the integrity of F-actin when the preprophase band is present rather than after its breakdown (Hoshino et al., 2003; Sano et al., 2005). Disruption of the actin-depleted zone or of the ‘twin peaks’ and ‘valley’ of BY-2 cells after prometaphase seems not to affect proper cytokinesis that much (Sano et al., 2005). The role of F-actin at the division site therefore remains to be re-evaluated.

In vacuolated cells, the presence and continuous motion of actin filaments in the subcortical cytoplasm and around the dividing nucleus/chromosomes is required for the maintenance of the cytoplasmic strands of the phragmosome (Traas et al., 1987; Lloyd & Traas, 1988; Flanders et al., 1990; Goodbody & Lloyd, 1990; Goodbody et al., 1991; Panteris et al., 2004) and rearrangement of the vacuolar compartments at the division plane (Kutsuna & Hasezawa, 2002; Kutsuna et al., 2003; Higaki et al., 2006). In addition, endoplasmic actin filaments connect the edges of the phragmoplast with the cortical cytoplasm, participating in its guidance (Lloyd & Traas, 1988; Valster & Hepler, 1997; Molchan et al., 2002; Panteris et al., 2004). Interaction of the cytoplasmic strands and endoplasmic actin filaments with the cortical division site may well be achieved through connections with the division-site F-actin.

In mitotic plant cells, cortical F-actin aggregations line the wall areas close to the spindle poles, probably participating in the anchoring and orientation of the spindle (Baluška et al., 2000; Panteris et al., 2007) through connections of the latter with the above aggregations (Lloyd & Traas, 1988). Depending on cell geometry and the orientation of the spindle, they may line the transverse or the longitudinal walls (Figs 2g,h,k and 3a,b). Especially in elongated cells, which divide vertically to their long axis, the above aggregations line the longitudinal walls close to the spindle poles (Figs 2k and 3c), somewhat resembling to the ‘twin peaks’ of BY-2 cells (Sano et al., 2005; Higaki et al., 2007). Although the ‘twin peaks’ have been proposed to participate in division site demarcation, F-actin aggregations under transverse or longitudinal walls and the ‘twin peaks’ share a common localization (Fig. 3), being closest to the mitotic spindle poles. This suggests that the ‘peaks’ may play a role similar to that of F-actin aggregations of other plant cell types (Baluška et al., 2000; Panteris et al., 2007) rather than demarcate the division site.

Figure 3.

Diagrammatic presentation of the distribution of actin filaments in three categories of dividing plant cells at cytokinesis. Cell walls and cell plates are shown in black, and F-actin in gray. The F-actin phragmoplast (central cell area in all cases) is connected to the division site by actin filaments. Thick gray lines denote F-actin aggregations, while the dotted gray lines denote sparsely distributed actin filaments. Compare the location of F-actin aggregations under the transverse (a) and longitudinal (b) walls, and the ‘twin peaks’ (c) configuration, depending on cell geometry and orientation of the division plane.

As for the division site ‘valley’, recent findings reveal not only what is excluded, such as the kinesin-like protein KCA1 (Vanstraelen et al., 2006), but also what remains there after the disappearance of the preprophase band (Van Damme et al., 2007; Walker et al., 2007). It seems that division site demarcation includes the local differentiation of the plasma membrane/cell wall, which prepares the division site for cell plate insertion (Lloyd & Buschmann, 2007; Van Damme et al., 2007; Van Damme & Geelen, 2008). The division site should harbor SNARE factors to act as vesicle fusion primers (Van Damme & Geelen, 2008). In addition, molecular factors such as TPLATE (Van Damme et al., 2006) and AIR9 (Buschmann et al., 2006; Lloyd & Buschmann, 2007) are exclusively found associated with the division site. In particular, the TANGLED protein has been found to be transported to the division site during preprophase and to remain there until completion of cytokinesis (Walker et al., 2007). Thus, this protein may create an ‘adhesive’ substrate that remains at the division site after preprophase band breakdown. In such a context, ‘depletion’, i.e. scarcity, of actin filaments at the division site ‘valley’ of mitotic plant cells may be related to a zone of immobility.

The stability of the division plane has already been attributed to the arrest of cytoplasmic streaming in mitotic cells (Lloyd & Traas, 1988). Indeed, as shown in Fig. 1c, the thick actin cables of interphase cells, which are involved in cytoplasmic streaming, are absent from dividing cells. Although the above suggestion (Lloyd & Traas, 1988) concerns mainly the anchoring of the nucleus at the division plane, this idea might well be applied to the cortical division site and its components. No mobility at all should be expected to occur around the site of future daughter wall insertion, to ensure that the local ‘adhesive’ plasma membrane and cell wall differentiation are not disturbed. Additionally, immobility at the division site would minimize disturbance for the edges of the cell plate, as the latter reaches the parent wall. If there was cytoplasmic mobility parallel to the plasma membrane at the division site, the local differentiation of the former might be displaced and the edges of the cell plate misguided, resulting in cell plate insertion out of the division site. Accordingly, scarcity of actin filaments at the division site (what has been considered as an ‘actin-depleted zone’) might be designed to provide optimum immobility. Therefore, the ‘valley’ between the ‘twin peaks’ might be considered as a ‘valley of serenity’, a peaceful cradle for the cell plate to rest in.


I am grateful to Professors Basil Galatis, Department of Botany, University of Athens, Greece, and Tobias Baskin, Biology Department, University of Massachusetts, USA, for critical reading of the manuscript and helpful suggestions. The help of our PhD student Ioannis-Dimosthenis Adamakis in preparing the figures for this article is gratefully acknowledged. The micrographs were taken with a Nikon D-Eclipse C1 CLSM, at the Faculty of Veterinary Medicine of our University, access to which was kindly provided by Dr Anastasia Tsingotjidou.