Background information. The actin cytoskeleton forms distinct actin arrays which fulfil their functions during cell cycle progression. Reorganization of the actin cytoskeleton occurs during transition from one actin array to another. Although actin arrays have been well described during cell cycle progression, the dynamic organization of the actin cytoskeleton during actin array transition remains to be dissected.
Results. In the present study, a GFP (green fluorescent protein)-mTalin (mouse talin) fusion gene was introduced into suspension-cultured tobacco BY-2 (Nicotiana tabacum L. cv Bright Yellow) cells by a calli-cocultivation transformation method to visualize the reorganization of the actin cytoskeleton in vivo during the progression of the cell cycle. Typical actin structures were indicated by GFP—mTalin, such as the pre-prophase actin band, mitotic spindle actin filament cage and phragmoplast actin arrays. In addition, dynamic organization of actin filaments was observed during the progression of the cell from metaphase to anaphase. In late metaphase, spindle actin filaments gradually shrank to the equatorial plane along both the long and short axes. Soon after the separation of sister chromosomes, actin filaments aligned in parallel at the cell division plane, forming a cylinder-like structure. During the formation of the cell plate, one cylinder-like structure changed into two cylinder-like structures: the typical actin arrays of the phragmoplast. However, the two actin arrays remained overlapping at the margin of the centrally growing cell plate, forming an actin wreath. When the cell plate matured further, an actin filament network attached to the cell plate was formed.
Conclusions. Our results clearly describe the dynamic organization of the actin cytoskeleton during mitosis and cytokinesis of a plant cell. This demonstrates that GFP—mTalin-transformed tobacco BY-2 cells are a valuable tool to study actin cytoskeleton functions in the plant cell cycle.
The actin cytoskeleton is a widely distributed structure in cells and plays important roles in many cellular processes, including the cell cycle in plants. The organization of plant-specific arrays of actin filaments and their functions in the cell cycle have been subjects for detailed investigation. By using various techniques, the distinct microfilament arrays during cell cycle progression have been well described. In most of the plant cells investigated, actin arrays in the interphase cell are a fine meshwork consisting of cortical microfilaments and bundles of endoplasmic microfilaments connecting the nucleus to the cell periphery (Cho and Wick, 1990; McCurdy and Gunning, 1990), which are involved in pre-mitotic nuclear movement (Kennard and Cleary, 1997; Grolig, 1998) and nuclear movement following division (Palevitz, 1988). During the G2 to early M-phase transition, cortical actin filaments co-align with the microtubule pre-prophase band to form an actin pre-prophase band (Sonobe and Shibaoka, 1989), which is required for correct narrowing of the pre-prophase band as it develops and for precise positioning of the division site (Mineyuki and Palevitz, 1990; Eleftheriou and Palevitz, 1992). In vacuolated cells undergoing mitosis, a spindle-associated actin basket is found in both endosperm cells and walled cells, which could serve as an anchor and thus control lateral spindle movement (Hepler et al., 1993; Kengen et al., 1995). In late anaphase, during phragmoplast formation, actin filaments have been initially described as a poorly defined mass of microfilaments between the incipient daughter nuclei, but soon develop into two antiparallel arrays at both sides of the division plane, with the barbed ends at the central region (Kakimoto and Shibaoka, 1988; Cleary et al., 1992). Studies using F-actin depolymerizing drugs, such as cytochalasin B or D, have shown that the actin cytoskeleton has a role in directing the location of cell plate (Cleary et al., 1992; Smith, 1999).
Although the arrays of actin cytoskeleton during cell cycle progression have already been described, the precise functions of actin arrays and how the dynamic organization of actin arrays takes place are not yet fully understood. Due to the extreme dynamics that actin filaments undergo, concerns have been raised as to whether or not chemical fixation sufficiently preserves the dynamic F-actin populations in plant cells. The development of new approaches to observe actin organization during the cell cycle in vivo will be indispensable for the further investigation of actin functions in higher plant cells. The discovery of GFP (green fluorescent protein) has greatly assisted researchers in plant cellular and molecular biology studies. It has dramatically improved our ability to visualize the organization of actin cytoskeleton in plant cells. Kost et al. (1998) first constructed and introduced the GFP—mTalin (mouse talin) fusion gene into plant cells to indicate the positioning of F-actin, and observed F-actin structures in plant cells of various tissues. Wang et al. (2004) produced several truncated fimbrins fused to GFP, which were expressed in Arabidopsis roots and found to decorate the networks of dynamic actin filaments. Sheahan et al. (2004) generated a fusion protein of GFP with the second actin-binding domain of Arabidopsis fimbrin (fABD2), which allowed the labelling of highly dynamic and dense actin networks in diverse species and various cell types, and indicated that reorganization of actin cytoskeleton occurred during the transition of a cell from expansion to division. Voigt et al. (2005) also found that, with the help of the fusion construct of GFP and the C-terminal half of Arabidopsis fimbrin 1, the dynamic arrays of F-actins can be observed in many types of cells in stably transformed Arabidopsis seedlings. However, these methods have not yet been used to study dynamic organization of the actin cytoskeleton during the complete progession of the cell cycle. This might provide new insights into the functions of the actin cytoskeleton during the cell cycle.
In the present study, we introduced the GFP—mTalin fusion construct into suspension-cultured tobacco BY-2 (Nicotiana tabacum L. cv Bright Yellow) cells using Agrobacterium-mediated transformation of BY-2 calli. Observation of the living BY-2 cells transformed with GFP—mTalin revealed that GFP—mTalin allowed in vivo visualization of dynamic actin cytoskeleton at different phases of the cell cycle. Therefore, a description of the dynamic organization of actin filaments has been made according to these results. Our study demonstrated that BY-2 cells expressing GFP—mTalin are a useful tool for research into the plant cell cycle.
Transformation of BY-2 calli with GFP—mTalin
In order to investigate the dynamic organization of actin arrays during plant cell cycle, BY-2 cell lines stably expressing the GFP—mTalin fusion protein were generated. Both calli and suspension-cultured BY-2 cells were assessed as transformation material. Results showed that transformed BY-2 cells were easily obtained when the calli were used, but it was difficult to produce transformed BY-2 cells from suspension-cultured cells. The successfully transformed BY-2 calli were bright yellow and easily distinguished from the non-transformed ones, which were brown, on selective agar (Figure 1A). The transformation efficiency was dependent on the growing time of calli, and the highest transformation efficiency was achieved with 3-week-old calli (Figure 1B). To find out whether the transfected gene affected the growth of BY-2 cells or not, SCV (settled cell volume) and MI (mitotic index) values were used for comparison between wild-type and GFP—mTalin-expressing BY-2 cells. Our data (Supplementary Figure 1, http:www.biolcell.orgboc098boc0980295add.htm) suggested that although the expression of GFP—mTalin decreased the growth rate of the transformed cells, it had no significant effect on the cell cycle of BY-2 cells, and hence the transformed BY-2 cells can be used to observe the dynamic organization of the actin cytoskeleton.
Organization of the actin cytoskeleton during cell cycle progression
Stable expression of GFP—mTalin by BY-2 cells allowed the organization of the actin cytoskeleton during cell cycle progression to be observed. In interphase, the cortical F-actin network was found in the cytoplasm next to the cell membrane (Figure 2A), and the nucleus was on one side of the cell and encased by large mass of cytoplasm within which the F-actin ‘basket’ formed (Figures 2B and 2C). When the cell starts to enter mitosis, the nucleus translocates to the centre of the cell (Lloyd, 1991), with long actin bundles radiating from the nucleus to the periphery of the cell where the actin forms a dense network (Figures 2D–2F). In pre-prophase, the amount of dense interphase cortical microfilament arrays decreased, but a cortical band of actin filaments was found to be located at the future equatorial plane where the microtubule pre-prophase band is known to be localized (Figure 2G). During this stage, there were a large number of F-actin bundles around the centrally located nucleus and vesicles (Figures 2H and 2I).
After the cell entered metaphase, the number of actin filaments in the cortex decreased, but they were focused at the mid-plane to form a mitotic spindlelike structure (Figures 3A–3D). Chromosomes were located at the equatorial plane (stained red; Figure 3B), and actin bundles (green) formed a spindle structure at the mid-plane of the cell. Figure 3(C) shows the co-localization of the chromosome and actin staining (yellow). In anaphase, when the two sets of daughter chromosomes separate, two layers of actin filaments between daughter chromosomes at the mid-plane of the cell were observed, but in the cortex the two sets of actin filaments overlapped (Figures 3E–3G). At this stage, membrane vesicles accumulated around the phragmoplast actin arrays, as described above (Figure 3H). In telophase, actin filaments were mostly absent from the cortex, and the structure of the phragmoplast was replaced by the cell plate with actin networks surrounding it, and the perinuclear (or perichromosomes) F-actin shells formed at this stage (Figures 3I–3L) (Supplementary Figure 2 shows the actin filament networks attaching to the cell plates, http:www.biolcell.orgboc098boc0980295add.htm). After cytokinesis, the dense cortical actin network reformed and the new daughter cell membranes produced a strong fluorescence signal (Figures 3M–3P).
Actin dynamics during mitosis and cytokinesis
In order to understand how the actin cytoskeleton changed, the dynamic process of cell division and the reorganization of the actin cytoskeleton were observed. As shown in Figure 4, actin filaments were found in both the periphery and the mid-plane of the spindle when the cell was in mid-metaphase, and an actin network formed in the cell cortex where there was a mass of vesicles collected around the spindle (Figures 4A–4C). After 30 min, the spindle actin filaments shrank towards the equatorial plane both in the long and short axes (Figures 4D and 4E), and the vesicles also moved towards the equatorial plane (Figure 4F). After 60 min, the spindle actin filaments shrank further and some vesicles reached the equatorial plane (Figures 4G–4I).
Approx. 5 min after the sister chromosomes separated from the equatorial plane, short actin filaments aligned in parallel at the mid-region of the interzone, forming a cylinder-like structure (Figures 5A–5D). After 15 min, one cylinder-like structure changed into the typical phragmoplast actin filament structure—two layers of actin arrays—when the cell plate started to form from the mid-region of the interzone. However, the two layers of actin arrays overlapped with each other, forming an actin ‘wreath’ at the margin of the centrally growing cell plate (Figures 5E–5H). After 30 min, the two layers of the overlapping actin wreath separated to form the typical phragmoplast actin arrays, while the cell plate expanded further into the wall of the mother cell (Figures 5I–5L). Another telophase cell (Supplementary Figure 3, http:www.biolcell.orgboc098boc0980295add.htm) dislayed the overlapping fine short actin filaments in the cortex and discontinuous membrane structures of the cell plate at the cell division plane.
As the cell cycle progressed on to late cytokinesis, the two daughter cells were separated from each other by a newly formed cell wall, and the actin filaments appeared to attach to the daughter membranes (Figure 6). At the beginning of this experiment, the fusion of the cell plate to the edge of the cell appeared to be incomplete, because the adjacent two daughter cells were not separated completely and the cell wall was not as smooth as the other areas (Figures 6A–6C). The two cells carried on separating over the next 60 min (Figures 6D–6F), and the gap between the two daughter cells in the cortex, indicated by the non-fluorescent zone, increased in size by 120 min (Figures 6G–6I). During this process, a dense cortical actin network returned to the daughter cells.
The results of the present study have allowed us to propose a model for dynamic organization of the actin cytoskeleton in cell cycle progression from metaphase to cytokinesis (Figure 7).
Calli transformation is an efficient method for BY-2 cell gene transfection
There are successful examples for gene transformation of BY-2 cells; for example, transformation by Agrobactium tumefaciens (An, 1985; Granger and Cyr, 2000) and gene gun (Kost et al., 1998; Yang et al., 2001). Suspension-cultured cells are generally used as the transformation material. In the present study, we have transformed the calli of BY-2 cells, and obtained much higher transformation efficiency compared with suspension-cultured cells.
The calli transformation method has several advantages over the transformation of suspension-cultured cells. Firstly, the washing and centrifugation steps, during which many suspension cells can be lost and damaged, are avoided. Secondly, the use of a shaker, which promotes the growth of bacteria, is avoided, so that uncontaminated calli are easily obtained. Thirdly, the steps in the process are simplified.
It has been reported that expression of GFP—mTalin in plant cells results in the termination of growth, cell death or artificial aggregation of actin networks (Ketelaar et al., 2004; Sheahan et al., 2004). In our experiments, the growth rates of wild-type and GFP—mTalin-transformed BY-2 cells were compared in order to investigate the effect of GFP—mTalin expression on the cells. Our results showed that the expression of GFP—mTalin in BY-2 cells did not cause large changes in cell morphology or cell death. The growth rate of the transformed cells was slightly lower than that of the wild-type cells, but the transformed cells were not affected significantly.
GFP—mTalin fusion protein is a useful tool for in vivo visualization of the dynamic actin cytoskeleton to study the cell cycle
GFP—fABD2 has been used to visualize actin organization during the cell cycle in Arabidopsis root-tip cells (Sheahan et al., 2004). However, detailed structures could not be visualized, probably because the cells are too small in size. Using the GFP—mTalin transgenic cell line, we investigated the reorganization of actin arrays during cell cycle progression and detected almost all the actin filament structures that have been reported previously by other techniques.
The cortical F-actin network is persistent during mitosis (Kakimoto and Shibaoka, 1988; Schmit and Lambert, 1990). Nevertheless, we found that the cortical F-actin network reorganized periodically. During interphase cortical actin filaments formed a compact network and the number of filaments was large. After the cell entered mitosis, most of the actin was involved in the organization of mitosis apparatus and only a few actin filaments were observed in the cortical protoplasm. At late cytokinesis, actin filaments returned to the cortical protoplasm and were reorganized into a dense network. The typical actin filament structures have also been observed in the present study. During the G2 to early M phase transition of BY-2 cells, cortical actin filaments formed an actin pre-prophase band which has been described previously (Sonobe and Shibaoka, 1989). When the BY-2 cells entered mitosis, the spindle-associated actin basket appeared to serve as an anchor and thus controlled lateral spindle movement, which has been described previously by different techniques (Hepler et al., 1993; Kengen et al., 1995). In late anaphase, a phragmoplast actin filament structure with two antiparallel arrays began to form between the incipient daughter nuclei, as described previously (Kakimoto and Shibaoka, 1988; Cleary et al., 1992). In addition, our present study is the first to observe the formation of a fine transient actin structure in the1 interzone between the separating chromosomes, just before the formation of phragmoplast actin arrays. This structure has previously been described as a poorly defined mass of microfilaments between the incipient daughter nuclei (Kakimoto and Shibaoka, 1988; Cleary et al., 1992). More recently, Sano et al. (2005) observed MFTP (microfilament twin peaks) during the mitosis of BY-2 cells expressing GFP—fABD2, which differed from the spindle actin filament structure found in our present study. The structure of the actin wreath around the cell plate detected in our present study was not described by Sano et al. (2005), although they showed that two populations of longitudinal microfilaments appeared temporarily between the two sets of chromosomes. The discrepancies between the studies discussed above and our present results are probably because of the difference in affinity of mTalin and fABD2 for the various populations of actin filaments. It indicates that the details of dynamic actin reorganization in living cells should be determined by combining the results from different labelling methods. Altogether, it is demonstrated that GFP—mTalin-transformed BY-2 cells provide an excellent experimental system for the study of actin cytoskeleton rearrangement and the role of actin during cell cycle progression.
Dynamic actin reorganization is involved in plant cell mitosis and cytokinesis
Actin filaments are recognized as widespread components in plant cells; however, data on the existence and role of the actin cytoskeleton during mitotic apparatus formation are controversial. Although the absence of actin filaments in the mitotic spindle has been reported (McCurdy and Gunning, 1990; Cho and Wick, 1991), other studies have indicated that actins are components of the plant cell spindle (Schmit and Lambert, 1988; Van Lammeren et al., 1989). Recently, Yasuda et al. (2005) reported the existence of cytochalasin D-insensitive actin filaments on the mitotic apparatus and the phragmoplast in BY-2 cells by rhodamine—phalloidin staining and immunofluorescence analysis using an anti-actin antibody. Our immunofluorescence analysis also showed that antiactin antibody recognized spindle mitotic apparatus in BY-2 cells expressing GFP—mTalin, demonstrating that our results were not due to the over-expression of the exogenous gene (see Supplementary Figure 4, http:www.biolcell.orgboc098boc0980295add.htm). In addition, our results showed that actin filaments were found not only in the periphery, but also in the mid-plane of the spindle in mid-metaphase cells, and subsequently the spindle actin filaments shrank in both long and short axes, which appeared to correlate with the localization of the microtubular arrays (Sider et al., 1999; Foe et al., 2000). This suggests that the spindle actin filaments may serve as an anchor and thus control lateral spindle movement. This is consistent with a recent study by Lenart et al. (2005) on the function of the actin cytoskeleton in starfish oocytes, in which it was observed that contraction of the actin network delivered chromosomes to the microtubule spindle during the first meiotic division.
Actin filaments are a major structural component of the cytokinetic apparatus, and inhibitory studies have shown that they play a crucial role in the proper execution of cytokinesis (Palevitz, 1988; Valster et al., 1997; Molchan et al., 2002, Yoneda et al., 2004). A number of studies have attempted to describe the dynamic transitions of the actin structure during the initiation and completion of the cell plate by microinjecting a low concentration of fluorescent phalloidin into the cell (Cleary et al., 1992; Zhang et al., 1993; Valster et al., 1997; Molchan et al., 2002). However, the details of the initiation of phragmoplast actin arrays are still unclear. In our present study, we found that membrane vesicles accumulated at the division plane soon after the two sister chromosomes separated, and at this time, short actin filaments overlapped at the mid-region of the interzone, which were probably the remnants of spindle actin arrays. However, during the formation of the cell plate, the remaining overlapping actin arrays separated and formed the two layers of the phagmoplast actin arrays in the region where the cell plate formed. These two overlapping layers of actin arrays then grew centrifugally to the edge of the growing cell plate, forming a wreath-like structure around the cell plate to the periphery of the cell, which had not been described before the present study. However, the actin wreath disappeared when the cell plate developed further and expanded into the periphery. This implies that actin may be involved in the organization and formation of the cell plate, which agrees with the observations that treatment with cytochalasin B or D distorts the growing cell plate (Cleary et al., 1992; Smith, 1999). At telophase, F-actin sheets have been found to be located between the two nuclei (Supplementary Figure 2, http:www.biolcell.orgboc098boc0980295add.htm). These results imply that the actin structure which co-localizes with the cell plate may play a role in driving the maturation of the cell plate, supporting the hypothesis that actin filaments participate in the alignment of cell plate (Valster et al., 1997). This theory is also supported by the finding that myosin VIII has been observed to concentrate along the newly formed cell walls of cress and maize root cells where F-actin is attached perpendicularly, and may be part of the system for anchoring or directing the cytoplasmic F-actin cables (Reichelt et al., 1999). In addition, Sano et al. (2005) also found that the MFTP had a role in cell plate guidance.
In conclusion, dynamic actin organization is involved in the progression of plant cell mitosis and cytokinesis, and the GFP—mTalin-transformed BY-2 cells are a valuable addition to the range of tools currently available to study the cytoskeleton rearrangements during cell cycle progression.
Materials and methods
Maintenance of cell cultures
For suspension culture, BY-2 cells were maintained in modified Murashige and Skoog liquid medium (3% sucrose, 4.3 g/l Murashige and Skoog salts, 100 mg/l inositol, 370 mg/ml KH2PO4, 1 mg/l aneurin, 0.2 mg/l 2,4-dichlorophenoxyacetic acid, pH 5.8) on a rotary shaker at 26°C in the dark. Cells were subcultured in new medium (1:20) weekly. BY-2 calli were maintained on solid medium (plus 0.8% agar) and replated every 3–4 weeks. Transgenic lines had 100 mg/l sterile kanamycin added to the medium.
Transformation of BY-2 calli
A. tumefaciens (strain GV3101) with the binary vector pBI121 encoding the GFP—mTalin protein were cultivated overnight in YEP (yeast extract peptone) liquid medium at 28°C. The bacterial cells were then collected by centrifugation at 3024 g for 5 min (Beckman Avanti J-25), and then suspended in modified Murashige and Skoog liquid medium. The D600 of the bacterial suspension was adjusted to approx. 0.5.
Small pieces of well-grown BY-2 calli were dipped into the pre-prepared A. tumefaciens suspension for 20 min, and the calli were placed on to agar medium for 2 days after aspirating the remaining A. tumefaciens liquid. Then the calli were washed with sterile water 4 times and were dipped into sterile water containing 500 mg/l ampicillin for 60 min to kill the bacteria, and then put on to agar medium containing 500 mg/l ampicillin and 100 mg/l kanamycin. Once the kanamycin-resistant calli had grown, they were transferred on to new agar containing kanamycin only and the calli were suspended in liquid medium to allow observations to take place.
Transformation of suspension-cultured BY-2 cells
A. tumefaciens cells (100 μl) at a D600 of approx. 0.5 were added to 4 ml of 3-day-old suspension-cultured BY-2 cells. They were co-cultivated on the rotary shaker at 26°C for 3 days. BY-2 cells were then collected by centrifugation at 500 g for 2 min, washed with liquid medium containing 500 mg/l ampicillin four times, and then put on to agar medium containing 500 mg/l ampicillin and 100 mg/l kanamycin. When kanamycin-resistant calli grew, they were transferred on to new agar containing only kanamycin and resuspended in liquid medium to allow observations to take place.
To visualize the nuclei, live BY-2 cells were suspended with 50 mM Pipes (pH 6.9) containing 0.1% DMSO and 2 μg/ml propidium iodide.
Samples were examined using a ×60 oil immersion objective [Plan Apochromat, ×60, NA (numerical aperture) 1.40] under an Olympus confocal microscope (Olympus FV300-IX 70) equipped with an argon laser (FV5-LA-AR) and helium—neon green laser (FV5-LA-HEG). GFP was excited at a wavelength of 488 nm by the argon laser and the emitted fluorescence was collected through a 505/525 band-pass filter (FVX-BA505-525). Propidium iodide was excited at a wavelength of 543 nm by the helium—neon green laser, and the emitted fluorescence was collected through a 610 nm long-pass filter (FVX-610IF). Serial confocal optical sections were taken at a step size of 0.5 μm. Transmitted light reference images were captured using DIC (differential interference contrast) optics (DIC prism IX-DPO 60). Images were presented as single sections or stacks of neighbouring sections, as indicated in the Figure legends.
This work was supported by the National Natural Science Foundation for Distinguished Young Scholars (grant no. 30325005), the National Natural Science Foundation of China (grant no. 30470176) and the National Basic Research Programme of China (grant no. 2006CB100100) to H.R.