* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Three-Dimensional Reconstruction of Tubular Structure of Vacuolar
Survey
Document related concepts
Biochemical switches in the cell cycle wikipedia , lookup
Extracellular matrix wikipedia , lookup
Tissue engineering wikipedia , lookup
Cytoplasmic streaming wikipedia , lookup
Cell growth wikipedia , lookup
Cellular differentiation wikipedia , lookup
Cell encapsulation wikipedia , lookup
Cell culture wikipedia , lookup
Endomembrane system wikipedia , lookup
Organ-on-a-chip wikipedia , lookup
Cytokinesis wikipedia , lookup
Transcript
Plant Cell Physiol. 44(10): 1045–1054 (2003) JSPP © 2003 Three-Dimensional Reconstruction of Tubular Structure of Vacuolar Membrane Throughout Mitosis in Living Tobacco Cells Natsumaro Kutsuna 1, Fumi Kumagai 1, Masa H. Sato 2 and Seiichiro Hasezawa 1, 3 1 2 Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, Chiba, 277-8562 Japan Graduate School of Human and Environmental Studies, Kyoto University, Sakyo-ku, Kyoto, 606-8501 Japan ; Plant vacuoles are the largest of organelles, performing various functions in cellular metabolism, morphogenesis and cell division. Dynamic changes in vacuoles during mitosis were studied by monitoring tubular structure of vacuolar membrane (TVM) in living transgenic tobacco BY-2 cells stably expressing a GFP-AtVam3p fusion protein (BY-GV). Comprehensive images of the complicated TVM configurations were obtained by reconstructing three-dimensional (3-D) surface structures from sequential confocal sections, using newly developed software, SSR (stereo-structure reconstructor). Using the surface modeling technique, we succeeded for the first time in clarifying the development process of TVMs and the topological relationship between TVMs and large vacuoles. TVMs, initially organized from large vacuoles, elongated to encircle the spindle at metaphase. Subsequently, the TVMs invaded the equatorial region from anaphase to telophase, and then they were divided to the two daughter cells by the cell plate at cytokinesis. When the daughter nuclei were separating from the cell plate, some TVMs enlarged to form large vacuoles near the division site. Spatial analysis revealed that from anaphase until cytokinesis, TVMs connected the two large vacuoles and functioned as a route for inter-vacuolar transport. Furthermore, the experiments using the inhibitor for actin microfilaments indicated that the microfilaments were indispensable for the development and the maintenance of TVMs. Keywords: 3-D reconstruction — AtVam3p — GFP — Mitosis — Tobacco BY-2 cells — Vacuole. Abbreviations: 3-D, three-dimensional; BA, bistheonellide A; BY-GV, BY-2 cell line expressing GFP-AtVam3p fusion protein; CLSM, confocal laser scanning microscopy; GFP, green fluorescent protein; MF, microfilament; MT, microtubule; SSR, stereo-structure reconstructor; TVM, tubular structure of vacuolar membrane; VM, vacuolar membrane. Introduction Vacuoles constitute the largest organelles in plant cells, and serve as multifunctional compartments (Wink 1993, Marty 3 1999). The large central vacuole occupies a considerable part of the intracellular volume of most higher plant cells, and over 90% of the cell volume in mature tissues. The space-filling properties and solutes of vacuoles therefore play an important role in the growth and morphogenesis of higher plants. During the rapid growth of higher plants, for example, increased vacuolar volumes can promote cell expansion without the need for cell division. On the other hand, the large size of vacuoles affects the dynamic and physical characteristics of plant cells, and their structural behavior is thought to be important for cell division and morphogenesis. The morphological dynamics of vacuoles have been studied by light microscopy, fluorescence microscopy and confocal laser scanning microscopy (CLSM), and a greater understanding of their structural dynamics has been facilitated by labelling techniques using endogenous pigments (Palevitz and O’Kane 1981, Palevitz et al. 1981), vital staining dyes (Hillmer et al. 1989, Lazzaro and Thomson 1996, Emans et al. 2002) and green fluorescent protein (GFP)-fusion proteins (Di Sansebastiano et al. 1998, Cutler et al. 2000, Mitsuhashi et al. 2000, Hawes et al. 2001, Darnowski and Vodkin 2002, Saito et al. 2002, Uemura et al. 2002). To date, however, little progress has been made in characterizing the dynamic changes in vacuoles during the cell cycle, primarily because of the lack of appropriate experimental systems in living plant cells. In order to follow vacuolar dynamics during cell cycle progression, we initially performed vital staining of the vacuolar membrane (VM, tonoplast) of tobacco BY-2 cells harboring developed vacuoles, by pulse-labelling with the styryl dye, FM4-64 (Kutsuna and Hasezawa 2002). At interphase, the vacuoles were compartmentalized by cytoplasmic strands and, by late G2 phase, the cytoplasmic strands, which extended from the central nucleus, gathered to the mid-region of cell where they formed the cytoplasmic plate that supports the mitotic apparatus (phragmosome). Interestingly, the characteristic VM structures were found to surround the mitotic apparatus at M phase, and were recognized as small ellipses on each optical section. However, by observing serial optical-sections, it became evident that the VM structures were tubular in form, and were thus designated as TVM (tubular structure of vacuolar membrane) (Kutsuna and Hasezawa 2002). The use of FM4-64 for staining VMs is well suited for confirming the existence of thin structures, such as TVM or Corresponding author: E-mail, [email protected]; Fax, +81-4-7136-3706. 1045 1046 3-D structure of vacuoles in living BY-GV cells cytoplasmic strands, but is unsuitable for time-sequential or multi-focal observations due to fluorescence fading; thus making analysis of the complex structures and dynamics of TVMs extremely difficult. In this study, therefore, we intended to clarify the structure and dynamics of TVM in order to elucidate the functions of VM structures in plant cell division. We therefore established a BY-2 cell line stably expressing a fusion protein of GFP with AtVam3p/SYP22 (Kutsuna and Hasezawa 2002), which belongs to the syntaxin family of Arabidopsis thaliana proteins (Sato et al. 1997). Using BY-GV cells, we first followed the detailed changes in the vacuolar structures throughout mitosis. We subsequently developed software, termed SSR (stereo-structure reconstructor), to construct three-dimensional (3-D) images of the VM structures, especially TVMs, from each series of optical sections generated by CLSM. From the data obtained with this system, the structure and function of TVMs during mitosis could be analyzed. Additionally, visible effects of actin microfilament (MF) disruption on the structures of vacuoles were demonstrated stereologically. The doublelabeling experiments using GFP-AtVam3p and rhodaminephalloidin suggested that the MFs participated in the structural adjustment of TVMs. Results Transgenic BY-2 cells for observing vacuolar structures Tobacco BY-2 cells were transformed with Agrobacterium tumefaciens, harboring a 35S-GFP-AtVam3p construct, essentially as described by An (1985), and the transformed cell lines were designated BY-GV (BY-2 cells stably expressing GFP-AtVam3p) (Kutsuna and Hasezawa 2002). From several BY-GV cell lines we have chosen a line, designated BY-GV7, for the following observations. The line displayed bright GFPfluorescence, and it could be maintained by 95-fold dilutions at weekly intervals as in the case for the original BY-2 cells. In 2day-old BY-GV7 cells, 10–20% of cells were observed in the M phase. Superficially, the shapes and sizes of the BY-2 cells and BY-GV7 cells were indistinguishable. However, the BYGV7 cells emitted bright GFP-fluorescence that highlighted the VM structures, especially TVM gross morphology. In BYGV7 cells, the VMs were more clearly labelled with GFPfluorescence than FM4-64 shown in Fig. 1 and 4 of Kutsuna and Hasezawa (2002). Throughout the cell cycle, the VM structures observed in BY-GV7, including TVMs, were very similar to those observed in two other BY-GV lines whose GFP-fluorescence was slightly lower than that of BY-GV7 (data not shown). The VM structures of BY-GV7 cells were also similar to those of original BY-2 cells stained with FM464 (Fig. 1A), and in BY-GV7 cells the GFP-fluorescence perfectly overlapped with the FM4-64-fluorescence (Fig. 1C). The vacuolar morphology could be observed by staining the vacuolar lumen with Alexa 568 hydrazide (Fig. 1B) or 2¢,7¢bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM; Kutsuna and Hasezawa 2002), and the Fig. 1 Vital-staining of vacuoles in tobacco BY-2 cells and BY-GV7 cells at telophase. (A, B) BY-2 cells, treated with FM4-64 (A) and Alexa 568 hydrazide (B), were observed at telophase by CLSM. TVMs were developed around the daughter nuclei. (C) A BY-GV7 cell, treated with FM4-64, was observed at telophase by CLSM. The FM4-64 fluorescence of VMs (C-1) certainly overlapped with the fluorescence of GFP-AtVam3p (C-2). (D) A BY-GV7 cell, treated with Alexa 568 hydrazide, was observed at telophase. The Alexa-fluorescence emitted from vacuolar lumen (D-1) was encircled by the GFPfluorescence (D-2). n, nucleus. Bars = 10 mm. GFP-fluorescence encircled on the vacuoles was shown by Alexa dye (Fig. 1D). These observations indicated that the fluorescence of GFP-AtVam3p certainly localized on the VMs in BY-GV7 cells. Furthermore, based on the immunoblotting analyses by the antibody against AtVam3p, the amount of the GFP-AtVam3p fusion protein in BY-GV7 cells was not so high (data now shown). From these results, we concluded that the fused protein employed here was incorporated and located into VMs and the vacuolar morphology, including TVMs, was not artificially affected in BY-GV7 cells. Vacuolar dynamics throughout mitosis As described above, the BY-GV7 cells showed quite comparable growth rates to the original BY-2 cells, and they could be highly synchronized by aphidicolin. More importantly, their emitted GFP-fluorescence faithfully reflected the dynamic changes in VM structures from G2 phase to G1 phase. As the BY-GV7 observation system was superior, in terms of VM fluorescent brightness, slowness of fading and recovery from fading, to that using FM4-64 (Kutsuna and Hasezawa 2002), the 3-D structure of vacuoles in living BY-GV cells 1047 Fig. 2 Time-sequence observations of vacuolar structures throughout mitosis. In a living BY-GV7 cell, the dynamic changes in vacuolar structures were time-sequentially observed at the late G2 phase (0–40 min), prophase (60 min), metaphase (100 min), anaphase (160 min), telophase (180 min) and early G1 phase (220 min). At the late G2 phase, thick cytoplasmic strands gathered at the central region of the cell. In a single optical section of the central region, several vacuolar compartments appeared to be isolated from the large vacuole by cytoplasmic strands (20 min, arrowheads). They then changed their forms as TVMs at the G2/M interface (80 min, arrowheads). From anaphase to telophase, TVMs were also observed in the equatorial region (180 min, arrowheads), and tended to align in the longitudinal direction. At cytokinesis (200 min), the TVMs divided into the two daughter cells because of centrifugal expansion of the cell plate developing within the phragmoplast. After completion of cytokinesis, some TVMs between the cell plate and daughter nuclei (200 min, arrowheads) developed into large vacuoles (220 min, arrowheads) at the separating of the nuclei from the cell plate. n, nucleus. Bar = 10 mm. system was considered suitable for time-sequential or multifocal observations. We therefore aimed to follow the detailed changes in fluorescence during mitosis by CLSM, with special emphasis on TVM dynamics around the mitotic apparatus (Fig. 2). The observations were performed with one randomly chosen cell at 20-min intervals. Thick cytoplasmic strands began to gather near the central region of the cell at late G2 phase (Fig. 2, 20 min, arrowheads). At this time, in optical sections, some vacuolar compartments appeared sandwiched by the cytoplasmic strands and isolated from the large vacuole. From there, the TVMs seemed to develop at the G2/M interface, to surround the mitotic apparatus at metaphase (Fig. 2, 80 min, arrowheads), and then to invade the equatorial region from anaphase to telophase (Fig. 2, 180 min, arrowheads). The TVMs tended to elongate along the longitudinal axis of cell extension, and were then divided by extensive growth of the cell plate at cytokinesis (Fig. 2, 200 min, arrowheads). After cytokinesis, some of the TVMs that were located between the cell plate and nucleus developed into the large vacuole at the time when the nucleus was separating from the cell plate (Fig. 2, 220 min, arrowheads). These observations implied that the TVMs developed from the vacuolar compartments around the nucleus at late G2 phase, and were maintained until cytokinesis. However, we were unable to recognize each complete TVM from single optical sections because the TVMs were winding tubes running over a wide area around the mitotic apparatus. 3-D reconstruction of vacuolar structures To understand the spatial configuration of TVMs and their relationship with large vacuoles, we performed 3-D reconstructions of the vacuolar structures from a series of confocal images taken along the z-axis by CLSM. Due to the lack of an appropriate VM processing system, we developed a new software package, SSR (Kutsuna et al., in preparation), to perform 3-D modeling of a series of optical sections from the BY-GV7 cells. To overcome the complexity of vacuolar structures as well as anisotropic resolution, which is typical of serial sections obtained by CLSM, we employed parallel contour reconstruction by which surfaces of 3-D objects are reconstructed by connecting planar contours that represent cross-sections through the objects (Fuchs et al. 1977). The obtained vacuolar structures could be observed from any angle, and measurements of their volume and surface area could be obtained. The 3-D vacuolar structures, which were reconstructed by SSR at each stage of the cell cycle, are shown in Fig. 3. Each image was constructed from serial optical sections obtained by CLSM at 0.5–1.0 mm intervals. In the image at the late G2 phase, TVMs have not been developed yet, and the cytoplasmic strands run penetrating the large vacuole around central region of the cell (Fig. 3A). Although some vacuolar compartments appeared to be isolated from the large vacuoles on the single optical section (Fig. 2, 20 min), in fact these compartments connected with each other and with the large vacuole 1048 3-D structure of vacuoles in living BY-GV cells Fig. 3 3-D structures of vacuoles. The images were constructed from a series of optical sections taken at 0.5–1.0 mm intervals. Vacuolar structures were classified and distinguished by color according to the area of each optical section (B, C, D’). Hence, the large vacuoles, the smaller compartments and TVMs are presented in blue, sky-blue and green, respectively. (A) Vacuolar structures at late G2 phase. The large vacuole was penetrated by thick cytoplasmic strands near the central region of the cell. (B) Vacuolar structures at prophase. The TVMs began to elongate from the smaller compartment of large vacuole. At this time, the TVMs were short in length and few in number. (C) Vacuolar structures at metaphase. The large vacuoles were divided by the cytoplasm, including the mitotic apparatus. All TVMs were attached to the large vacuoles (arrowheads) and connected the two large vacuoles. (D, D’) Vacuolar structures at telophase. The division plane is represented by the orange line. TVMs appeared to be cut off by the cell plate. m represents the position of the mitotic apparatus. Bars = 10 mm. (Fig. 3A). In the image of vacuolar structures at prophase, the cross-sections of vacuoles are divided into three classes by the area size (Fig. 3B). The large sections, the middle ones and the small ones are shown in blue, sky-blue and green, respectively. This indicated that the TVMs, with small diameters and shown in green, extended from the sky-blue compartments, which were located at the periphery of the nucleus and in the midregion of the cell (Fig. 3B). The TVMs were few in number and short in length at this stage. In the image obtained at metaphase, the large vacuole was divided in two by the cytoplasm around the mitotic apparatus, proving that the TVMs elongated around the mitotic apparatus and connected the two large vacuoles (Fig. 3C). This finding is the first demonstration of the TVM dynamics using 3-D modeling. In the image at telophase, the cell plate was found to develop toward the edge of the cell (Fig. 3D). Numerous TVMs shown in green were distributed near the division plane (Fig. 3D’). These TVMs were therefore demonstrated to be cut off, when and where that part of the cell plate was completed. Lumenal connectivity of large vacuoles provided by TVMs From the 3-D modeling using SSR we therefore obtained a considerable amount of new information regarding TVMs, and from these findings we focused our subsequent studies on TVM cross-linking between two large vacuoles. In three optical sections, near the metaphase cell surface, obtained by CLSM at 0.7 mm intervals, the large vacuole was bisected by cytoplasm around the mitotic apparatus (Fig. 4). When the TVM was traced from left to right, from the point where it was connected to the large vacuole (Fig. 4, arrowheads), the TVM appeared to connect both large vacuoles (Fig. 4). When the vacuoles of metaphase BY-2 cells were stained with fluorescent dye, BCECF, large vacuoles with faint TVMs could be observed (Fig. 5A, left). Photobleaching was performed with strong excitation light only in the dashed area, after taking the photograph. After a minute, BCECF fading was observed both in the photobleached vacuole and in an additional vacuole (Fig. 5A, right). This not only suggested that both large vacuoles, which were bisected by the cytoplasmic plate, were topologi- 3-D structure of vacuoles in living BY-GV cells 1049 Fig. 4 Connections between the large vacuoles and TVMs at metaphase of a BY-GV7 cell. Three optical sections were obtained by scanning from the cell surface at 0.7 mm intervals. The large central vacuole was divided into two by the cytoplasm, including the mitotic apparatus, from where the TVMs were derived (arrowheads). The TVMs were observed to connect the two large vacuoles through three serial sections. l, large vacuole. Bar = 10 mm. cally connected by TVMs, but also that the vacuolar contents were always able to diffuse between the large vacuoles through the TVMs. Role of actin MFs on the morphology of TVMs To clarify the mechanism for development and maintenance of tubular structures characteristic of TVM, effects of cytoskeleton inhibitors on the morphology of vacuoles were observed using the BY-GV7 cells. We found that development of TVMs were inhibited by the treatment with 1 mM bistheonellide A (BA), a dimeric macrolide, that was recently reported to inhibit polymerization of G-actin (Saito et al. 1998, Hoshino et al. 2003). At the late G2 phase, BA treatment caused the formation of several spherical vacuoles around the nuclei (Fig. 6A-1, arrowheads). The 3-D image showed that the newly formed spherical vacuoles were not connected to the large vacuoles via VMs, and connectivity between two large vacuoles was disrupted (Fig. 6A-2). Moreover, the disruption of MFs inhibited the development and maintenance of TVMs at the metaphase (Fig. 6B). Instead of TVMs, the spherical vacuoles were observed (Fig. 6B-1). Treatments with 100 mM cytochalasin B and D, other actin inhibitors, caused similar changes in the structure of vacuoles (data not shown). The disappearance of TVMs, by the destruction of MFs, might cause the loss of lumenal connection between the two large vacuoles. Thus, the vacuolar lumen of BY-2 cells was vital-stained with BCECF, and then their MFs were destroyed with BA. Subsequently, the photobleaching was performed on Fig. 5 Connectivity of the lumen of two large vacuoles at metaphase. The vacuolar lumens of metaphase BY-2 cells were stained with BCECF, and the cells then observed by fluorescence microscopy. Photobleaching was performed by the irradiation with the strong excitatory light only to the dashed area. (A) The large central vacuole of the central cell was divided into two sections (prebleach). In addition to the large vacuoles, TVMs were also identified by faint fluorescence of BCECF (arrowheads). However, after photobleaching, the fading of the green fluorescence was not only limited to the left vacuole but also observed in the right vacuole (post-bleach). (B) A BA-treated cell. Instead of TVMs, some spherical vacuoles were observed around mitotic apparatus (pre-bleach). The fading, observed in the left large vacuole, did not spread to the right vacuole (postbleach). l, large vacuole. Bars = 10 mm. 1050 3-D structure of vacuoles in living BY-GV cells Fig. 6 Effect of BA on the organization of TVMs. BY-GV7 cells were treated with 1 mM BA which have immediate effect on actin-depolymerization. After 30 min of the treatment, structures of vacuoles were observed by the GFP fluorescence. The single optical sections (A-1, B-1) and the corresponding 3-D vacuolar structures (A-2, B-2) of living BY-GV7 cells were shown. Connections between the large vacuoles were lost and several spherical vacuoles were formed around the nuclei at the late G2 phase (A-1, arrowheads). In contrast to untreated cells, TVMs were not developed and the spherical vacuoles (B-1, arrowheads) were observed at the metaphase (B). n, nucleus. Bars = 10 mm. the metaphase cells, which were untreated or treated with BA (Fig. 5). In contrast to the untreated cell (Fig. 5A, right), the BCECF fading did not spread to another large vacuole in the BA-treated cell (Fig. 5B, right). As expected from the 3-D image (Fig. 6), the disruption of MFs actually caused the separation of vacuoles and the lumenal disconnection between the two large vacuoles, suggesting that the MFs may be necessary for the maintenance of TVMs as the lumenal connection. These experiments with the inhibitor indicated that the MFs were involved in the organization of TVMs. To gain insight into the spatial relationship of them, BY-GV7 cells were stained with rhodamine-phalloidin to visualize the MFs and were observed by CLSM (Fig. 7). Some TVMs seemed to be surrounded by or colocalized with the MFs during mitosis (Fig. 7A, B). On the other hand, when pretreated with BA for 30 min, such MFs were rarely observed (Fig. 7C), as described in Hoshino et al. (2003). Discussion In this study, we have established a cell line BY-GV7 that stably expresses a GFP-AtVam3p fusion protein, and examined the dynamics and 3-D structures of plant vacuoles during mitosis. GFP fluorescence in the VM of BY-GV7 cells was more intensive and durable than by staining with the fluorescent dye, FM4-64. In the latter case, the fluorescence gradually darkened, and some particulate artifacts appeared in the vacuoles within several days (Kutsuna and Hasezawa 2002). In con- Fig. 7 Double labelling of VMs and actin MFs in mitotic cells. To visualize VMs and MFs simultaneously, BY-GV7 cells were stained with rhodamine-phalloidin under condition that did not disturb VM structures. The green GFP fluorescence of VMs and the red rhodamine fluorescence of MFs were observed by CLSM. (A) A metaphase cell. A part of TVMs was surrounded in MFs (arrowhead). (B) A telophase cell. Some TVMs were colocalized with MFs around the cell plate (arrowhead). (C) A telophase cell pretreated with BA. Small spherical vacuoles were observed around the daughter nuclei. MFs were destroyed by BA treatment. n, nucleus. Bars = 10 mm. 3-D structure of vacuoles in living BY-GV cells 1051 Fig. 8 Schematic representation of vacuolar structures at mitosis in BY-2 cells. (A) The large central vacuole is compartmentalized around the nucleus by cytoplasmic strands at the G2 phase. Some of the cytoplasmic strands accumulate at the central region of the cell. (B) Some vacuolar compartments convert to TVMs at the G2/M interface. When the large vacuole is divided into two sections, some TVMs connect both vacuoles at metaphase. (C) After chromosomal segregation, TVMs invade into the equatorial region at telophase. The TVMs begin to be cut off by the cell plate that develops within the phragmoplast. (D) When cytokinesis is complete, all the TVMs are divided into the two daughter cells. Thereafter, some TVMs between the cell plate and daughter nuclei develop into large vacuoles at the early G1 phase. n, nucleus. red, VM; sky blue, vacuole; green, MT. trast, the BY-GV7 cells showed stable GFP fluorescence due to the constitutive expression of the GFP-AtVam3p fusion protein by the CaMV 35S promoter. From the observations of BY-GV7 cells double-labelled with GFP-AtVam3p and FM4-64 (Fig. 1C) or GFP-AtVam3p and Alexa hydrazide (Fig. 1D), the fluorescence of GFP-AtVam3p seemed to be located in the whole VMs and not to be located in the limited part of VMs. In this context, it has been previously demonstrated that the GFPAtVam3p was localized to the VMs and could be used as a fluorescent probe for the VM structures in transgenic A. thaliana (Uemura et al. 2002). Specific features of the BY-GV7 cells, notably fluorescence intensity, durability and sharpness, enabled us to perform not only studies on vacuolar morphology in living cells but also time-sequential and multi-focal observations. In addition, the shapes and sizes of the BY-GV7 cells were indistinguishable from the original BY-2 cells, their growth rates were essentially similar, and the cells could be synchronized by the same method described in Nagata and Kumagai (1999) (data not shown). Therefore, at the peak of mitotic index, a large number of cells at various mitotic and premitotic stages could be easily observed at the same time. The structure of VMs in living plant cells are dynamically modified within a brief time-span, and their configurations are sometimes very complicated and intricate. Spherical structures, designated as ‘bulbs’, have been reported within the lumen of vacuoles in rapidly expanding cells of A. thaliana (Saito et al. 2002), and cylindrical and sheet-like structures that invaginate into the vacuolar lumen have been observed in several tissues of transgenic A. thaliana expressing a GFPAtVam3p (Uemura et al. 2002). These structures were reported to exhibit considerable motility. Time-lapse observations and acquisition of 3-D images were indispensable for comprehending these structures. During mitosis, TVMs also displayed dynamic changes in their conformations, with entangled and complicated structures (Fig. 2), and 3-D image processing served as a powerful tool for understanding such structures. Of the numerous approaches to 3-D image processing, we selected parallel contour reconstruction, which performs surface reconstructions, and then developed specific computer software, SSR, as tool for reconstructing 3-D structures and analyzing the series of CLSM images. By employing the contour-based modeling system, we have succeeded in reconstructing and visualizing 3-D vacuolar structures containing TVMs (Fig. 3, 6). Compared to the single optical sections, the 3-D images reconstructed by SSR provided more comprehensive views of the morphological changes induced by the MF disruption (Fig. 6). In addition, the newly developed SSR software could be used to quantify 3-D structures, such as of the surface area and volume of both vacuoles and whole cells (data not shown). It is notable that the SSR software can be widely applied to various biological objects, especially for membranous structures even in large plant cells, in which the CLSM images are prone to anisotropic resolution and to greater topological differences between serial sections of whole cells. In addition, by combining SSR with a real-time observation system using CLSM, we think that we will be able to monitor the dynamic images of 3D structures over a long time-period. In such cases, SSR may be suitable for investigations of plant cell morphology. Our results on the dynamics of TVMs are summarized in Fig. 8. At the late G2 phase, the cytoplasmic strands converge on the central region of the cell, and compartmentalize the large central vacuole around the nucleus (Fig. 8A). The TVMs then elongate from some of these vacuolar compartments at the G2/M interface. Although the large vacuole is segregated into two parts by the cytoplasm and mitotic apparatus, they are connected to each other by the TVMs at metaphase (Fig. 8B). Some TVMs invade the equatorial area after chromosomal segregation at anaphase. When the cell plate is formed, the TVM bridges between the large vacuoles are cut off by the developing cell plate (Fig. 8C). After cytokinesis, some TVMs between the cell plate and daughter nuclei develop into large vacuoles at the early G1 phase (Fig. 8D). With respect to TVM behavior, the vacuolar-tubular continuum in the secretory trichomes of chickpea were reported to be continuous between the stalk cells 1052 3-D structure of vacuoles in living BY-GV cells and to function in the rapid delivery of solute from the trichome base to the secretory head cells (Lazzaro and Thomson 1996). In BY-GV7 cells, the TVMs connected the two large vacuoles and appeared to be cut by the cell plate during cell plate formation. It is possible that this TVM bridge is the precursor of the continuum observed in stalk cells of chickpea. Conversely, if TVMs accidentally escape being cut off at cytokinesis, the TVMs may penetrate into the new cell wall as in the case of the vacuolar-tubular continuum. The viscosity of the lumen of lytic vacuoles is considerably lower than that of the cytoplasm, a property that appears valuable for bulk transport. In fungal hyphae, motile tubular vacuoles have been implicated in longitudinal transport of nitrogen and phosphorus compounds (Ashford 1998). In our photobleaching experiment (Fig. 5A), we demonstrated that lumens of the two large vacuoles were connected by TVMs, and that lumenal solutes could probably diffuse freely. Compared to the result of BA treated cells in Fig. 5B, the BCECF fading in Fig. 5A was not due to the cellular damage caused by the excitation light of photobleaching. It is uncertain how solutes are mixed within the lumen and whether any mechanism of positive stirring exists. The time-course pattern of BCECF fading could be explained by passive diffusion. The connectivity of vacuolar lumens through the TVMs also suggests that specific conditions, for example osmotic pressure, pH, membrane potential, and electrolyte concentrations, were maintained at equal levels in the two large vacuoles. It is noteworthy that TVMs were found connected to the large vacuole throughout mitosis and cytokinesis, suggesting that they may function as branches of the large vacuole. Some TVMs invaded the equatorial region between anaphase and telophase (Fig. 2, 180 min), and subsequently co-localized with the developing phragmoplast within which the cell plate was forming (Kutsuna and Hasezawa 2002). For cell plate formation, large amounts of membranous components appear to be consumed by vesicle transport of the cell plate materials. In the contrast, the surface area and volume of the cell plates were decreased dramatically during their maturation (Otegui et al. 2001). Thus, the removal of excess membranes from the cell plates should occur accompanied by the maturation of cell plates. The appearance of multivesicular bodies and small vacuoles, adjacent to cell plates, has also been reported (Samuels et al. 1995, Otegui et al. 2001). These lytic compartments seemed to be involved in the recycling of membrane lipids as the downstream organelles of the cell plates. In our observations (Fig. 1, 2, 3D), some TVMs were located near the developing cell plates. Therefore, these TVMs may have the possible role in the recruitment of the membrane components. In addition, it is possible that TVMs fulfill other roles, for example in the protection or positioning of the mitotic apparatus and in preservation of VMs for vacuolar enlargement at the G1 phase. We previously demonstrated that microtubules (MTs) were indispensable to the organization of TVMs that encircle the chromosomal region (Kutsuna and Hasezawa 2002). Never- theless, disruption of MTs did not affect the tubular structure of TVMs themselves (Kutsuna and Hasezawa 2002). In this study, we found that development and maintenance of TVMs were directly or indirectly dependent on the actin MFs from inhibitor experiments (Fig. 5B, 6). The exact functions and mechanisms of MFs in the TVM development are hitherto unknown. In higher plant cells, MFs are involved in organization of the preprophase band (Eleftheriou and Palevitz 1992), phragmoplast (Zhang et al. 1993) and radial network of MTs (Miyake et al. 1997). Because the treatments for staining MFs by the fluorescent probes tend to deform VM structures, few studies demonstrating spatial relationships between MFs and VMs have been reported. Consequently, little is known about relationship between the vacuoles and MFs in higher plants. The dynamics of vacuoles in Allium stomatal cells require the actin MFs (Palevitz and O’Kane 1981). In epidermal cells of transgenic A. thaliana expressing a GFP-AtVam3p, cylindrical and sheet-like structures invaginated into the vacuolar lumen were observed (Uemura et al. 2002). While their morphology was not disrupted, the movement of them was blocked with the MF depolymerizing drug, cytochalasin D (Uemura et al. 2002). These reports also suggest the existence of MF-dependent machinery for deformation of vacuoles. In this article, the localizations of TVMs and MFs were shown at metaphase (Fig. 7A) and at telophase (Fig. 7B). The MFs were partially colocalized with the TVMs. These results implied that the VM dynamics was directly regulated through colocalization or contact with MFs. In summary, using our newly developed SSR software as a tool for modeling, analyzing and viewing 3-D biological image data, we have been able to demonstrate the following: (1) BYGV7 cell line is suitable for time-sequence observations of VMs; (2) SSR can clarify complicated cellular structures such as TVMs; (3) each TVM is a tubular structure and is derived from the periphery of the large vacuole; (4) TVMs actually connect two large vacuoles, and act as a pathway for the transport of soluble vacuolar contents during mitosis; and (5) development and maintenance of TVMs depend on MFs. Our future studies will focus on analyzing in detail the structure and function of plant vacuoles, especially in TVMs, using BY-GV7 cells and our SSR software. Materials and Methods Plant material At weekly intervals, suspension cultures of a tobacco BY-2 cell line, derived from seedlings of Nicotiana tabacum L. cv. Bright Yellow 2, were diluted 95-fold with a modified Linsmaier and Skoog (LS) medium (Linsmaier and Skoog 1965), as described (Nagata et al. 1992). The cell suspension was agitated on a rotary shaker at 130 rpm at 27°C in the dark. Transformation and establishment of the BY-2 cell line expressing the GFP-AtVam3p fusion protein Agrobacterium tumefaciens strain, C58C1, was transformed with the GFP(S65T)-AtVam3p construct. A 4-ml aliquot of 3-day-old BY-2 3-D structure of vacuoles in living BY-GV cells cells was inoculated with 100 ml of an overnight culture of the transformed A. tumefaciens as described (An 1985). After a 2-day incubation at 27°C, the cells were washed four times in 5 ml LS medium, and then plated onto solid LS medium containing 250 mg liter–1 kanamycin and 500 mg liter–1 carbenicillin. Calluses that appeared after 10 d were transferred onto new plates and cultured independently until they reached a diameter of about 1 cm, at which time they were transferred to 20 ml liquid LS medium in 100-ml Erlenmeyer flasks and agitated on a rotary shaker at 130 rpm at 27°C in the dark. After 1 month, a cell line suitable for VM structural observations was selected by identifying GFP-fluorescent cells by fluorescence microscopy. The selected cell line, designated BY-GV7 (BY-2 cell line expressing GFPAtVam3p fusion protein No.7), could be maintained by 95-fold dilutions at weekly intervals, and be synchronized similarly to the original BY-2 cells. Staining of vacuoles and vacuolar membranes To stain the inside of vacuoles, 2¢,7¢-bis-(2-carboxyethyl)-5-(and6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM, Molecular Probes Inc., Eugene, OR, U.S.A.) and Alexa 568 hydrazide (Molecular Probes) were used. BCECF-AM was added to the cell suspension at the final concentration of 10 mM. The cells were incubated for 1 h at 27°C, washed with fresh culture medium, and then incubated for an additional 6 h at 27°C. Alexa 568 hydrazide was used essentially according to Emans et al. (2002). To visualize the VM in BY-2 cells, N-(3-triethylammoniumpropyl)-4-(6-(4-(diethylamino)phenyl)hexatrienyl) pyridinium dibromide (FM4-64, Molecular Probes) was used (Kutsuna and Hasezawa 2002). Cells between 10 and 20 h after labelling were used to observe VMs. Staining of actin MFs For simultaneous observations of MFs and VMs, BY-GV7 cells were suspended in a solution containing 50 mM PIPES (pH 6.8), 1 mM MgSO4, 5 mM EGTA, 1% glycerol, 3% sucrose, 0.03% saponin (ICN Biomedicals Inc., Aurora, OH, U.S.A.), 250 mM m-maleimidobenzoyl-N-hydroxysulfosuccinimide ester (Sulfo-MBS; Pierce Chemical Co., Rockford, IL, U.S.A.) and 0.3 mM rhodamin-phalloidin (Molecular Probes). The cells were incubated for 5 min at room temperature, and then observed. Time-sequential and multi-focal observations For time-sequence observations, 2-day-old BY-GV7 cells were transferred into Æ 35 mm Petri dishes with Æ 14 mm coverslip windows at the bottom (Matsunami Glass Ind., Ltd., Osaka, Japan). The dishes were placed onto the inverted platform of a fluorescence microscope (IX, Olympus, Olympus Co. Ltd., Tokyo, Japan), equipped with a confocal laser scanning head and control systems (CLSM GB-200; Olympus). This observation system, which allows observation of VM dynamics in BY-GV7 cells with minimal damage, was found to be ideal for our studies. The 2-D images were processed digitally using Photoshop software (Adobe Systems Inc., San Jose, CA, U.S.A.). Reconstruction and computer visualization of 3-D structures of vacuoles Reconstruction and visualization of the 3-D structures of vacuoles were performed using SSR, an originally developed software for 3-D structure reconstruction (Kutsuna et al., in preparation), running under SerioWare GNU/Linux on an Athlon XP 2200+ personal computer (Advanced Micro Devices Inc., Sunnyvale, CA, U.S.A.) equipped with 2 gigabytes of random access memory. Serial optical sections taken at 0.5–1.0 mm intervals were displayed and analyzed with SSR. Once the 3-D surface data was reconstructed, vacuolar models were displayed either by surface rendering or by volume render- 1053 ing, using the Open Visualization Data Explorer (International Business Machines Corp., White Plains, NY, U.S.A.). Photobleaching of BCECF within vacuoles The BY-2 cells were stained with BCECF as described above, and monitored with a fluorescence microscope (BX, Olympus) equipped with a CCD camera system (DP70, Olympus). Following observation, and confirmation that BCECF fluorescence was limited to the vacuolar lumens, photobleaching was performed by irradiating only a small area of the cell with strong excitation light using an iris. Inhibitor treatment The BY-GV7 and BY-2 cells were treated with 1 mM bistheonellide A (BA, Wako Pure Chemical Ind., Osaka, Japan) or 100 mM cytochalasin B and D (Sigma, St. Louis, MO, U.S.A.) for disruption of MFs as described previously (Hoshino et al. 2003) and were then observed by the fluorescence microscopy and the CLSM. Acknowledgments This study was supported in part by a Grant-in-Aid for Scientific Research on Priority Areas (No. 10219201 and 15031209) from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and the Nissan Science Foundation (S. H). References An, G. (1985) High efficiency transformation of cultured tobacco cells. Plant Physiol. 79: 568–570. Ashford, A.E. (1998) Dynamic pleiomorphic vacuole systems: are they endosomes and transport compartments in fungal hyphae? Adv. Bot. Res. 28: 119– 159. Cutler, S.R., Ehrhardt, D.W., Griffitts, J.S. and Somerville, C.R. (2000) Random GFP::cDNA fusions enable visualization of subcellular structures in cells of Arabidopsis at a high frequency. Proc. Natl Acad. Sci. USA 97: 3718– 3723. Darnowski, D.W. and Vodkin, L.O. (2002) A soybean lectin-GFP fusion labels the vacuoles in developing Arabidopsis thaliana embryos. Plant Cell Rep. 20: 1033–1038. Di Sansebastiano, G.-P., Paris, N., Marc-Martin, S. and Neuhaus, J.-M. (1998) Specific accumulation of GFP in a non-acidic vacuolar compartment via a Cterminal propeptide-mediated sorting pathway. Plant J. 15: 449–457. Eleftheriou, E.P. and Palevitz, B.A. (1992) The effect of cytochalasin D on preprophase band organization in root tip cells of Allium. J. Cell Sci. 103: 989– 998. Emans, N., Zimmermann, S. and Fischer, R. (2002) Uptake of a fluorescent marker in plant cells is sensitive to brefeldin A and wortmannin. Plant Cell 14: 71–86. Fuchs, H., Kedem, M. and Uselton, S.P. (1977) Optimal surface reconstruction from planar contours. Comm. ACM 20: 693–702. Hawes, C., Saint-Jore, C.M., Brandizzi, F., Zheng, H., Andreeva, A.V. and Boevink, P. (2001) Cytoplasmic illuminations: in planta targeting of fluorescent proteins to cellular organelles. Protoplasma 215: 77–88. Hillmer, S., Quader, H., Robert-Nicoud, M. and Robinson, D.G. (1989) Lucifer yellow uptake in cells and protoplasts of Daucas carota visualized by laser scanning microscopy. J. Exp. Bot. 40: 417–423. Hoshino, H., Yoneda, A., Kumagai, F. and Hasezawa, S. (2003) Examining the roles of actin depleted zone (ADZ) and preprophase band (PPB) in determining the division site of higher plant cells using a BY-2 cell line expressing GFP-tubulin (BY-GT). Protoplasma in press. Kutsuna, N. and Hasezawa, S. (2002) Dynamic organization of vacuolar and microtubule structures during cell cycle progression in synchronized tobacco BY-2 cells. Plant Cell Physiol. 43: 965–973. Lazzaro, M.D. and Thomson, W.W. (1996) The vacuolar-tubular continuum in living trichomes of chickpea (Cicer arietinum) provides a rapid means of solute delivery from base to tip. Protoplasma 193: 181–190. 1054 3-D structure of vacuoles in living BY-GV cells Linsmaier, E.M. and Skoog, F. (1965) Organic growth factor requirements of tobacco tissue cultures. Physiol. Plant. 18: 100–127. Marty, F. (1999) Plant vacuoles. Plant Cell 11: 587–599. Mitsuhashi, N., Shimada, T., Mano, S., Nishimura, M. and Hara-Nishimura, I. (2000) Characterization of organelles in the vacuolar-sorting pathway by visualization with GFP in tobacco BY-2 cells. Plant Cell Physiol. 41: 993–1001. Miyake, T., Hasezawa, S. and Nagata, T. (1997) Role of cytoskeletal components in the migration of nuclei during the cell cycle transition from G1 phase to S phase of tobacco BY-2 cells. J. Plant Physiol. 150: 528–536. Nagata, T. and Kumagai, F. (1999) Plant cell biology through the window of the highly synchronized tobacco BY-2 cell line. Meth. Cell Sci. 21: 123–127. Nagata, T., Nemoto, Y. and Hasezawa, S. (1992) Tobacco BY-2 cell line as the “HeLa” cell in the cell biology of higher plants. Int. Rev. Cytol. 132: 1–30. Otegui, M.S., Mastronarde, D.N., Kang, B.-H., Bednarek, S.Y. and Staehelin, L.A. (2001) Three-dimensional analysis of syncytial-type cell plates during endosperm cellularization visualized by high resolution electron tomography. Plant Cell 13: 2033–2051. Palevitz, B.A. and O’Kane, D.J. (1981) Epifluorescence and video analysis of vacuole motility and development in stomatal cells of Allium. Science 214: 443–445. Palevitz, B.A., O’Kane, D.J., Kobres, R.E. and Raikhel, N.V. (1981) The vacuole system in stomatal cells of Allium. Vacuole movements and changes in morphology in differentiating cells as revealed by epifluorescence, video and electron microscopy. Protoplasma 109: 23–55. Saito, C., Ueda, T., Abe, H., Wada, Y., Kuroiwa, T., Hisada, A., Furuya, M. and Nakano, A. (2002) A complex and mobile structure forms a distinct subregion within the continuous vacuolar membrane in young cotyledons of Arabidopsis. Plant J. 29: 245–255. Saito, S., Watabe, S., Ozaki, H., Kobayashi, M., Suzuki, T., Kobayashi, H., Fusetani, N. and Karaki, H. (1998) Actin-depolymerizing effect of dimeric macrolides, bistheonellide A and swinhilide A. J. Biochem. 123: 571–578. Sato, M.H., Nakamura, N., Ohsumi, Y., Kouchi, H., Kondo, M., HaraNishimura, I., Nishimura, M. and Wada, Y. (1997) The AtVAM3 encodes a syntaxin-related molecule implicated in the vacuolar assembly in Arabidopsis thaliana. J. Biol. Chem. 272: 24530–24535. Samuels, A.L., Giddings, Jr., T.H. and Staehelin L.A. (1995) Cytokinesis in tobacco BY-2 and root tip cells: a new model of cell plate formation in higher plants. J. Cell Biol. 130: 1345–1357. Uemura, T., Yoshimura, S.H., Takeyasu, K. and Sato, M.H. (2002) Vacuolar membrane dynamics revealed by GFP-AtVam3p fusion protein. Gene. Cell. 7: 743–753. Wink, M. (1993) The plant vacuole: a multifunctional compartment. J. Exp. Bot. Suppl. 44: 231–246. Zhang, D., Wardsworth, P. and Hepler P.K. (1993) Dynamics of microfilaments are similar, but distinct from microtubules during cytokinesis in living, dividing plant cells. Cell Motil. Cytoskel. 24: 151–155. (Received April 18, 2003; Accepted July 30, 2003)