Studies on the behavior of mitochondrial DNA: Synthesis of mitochondrial dna occurs actively in a specific region just above the quiescent center in the root meristem of Pelargonium zonale

The sizes of mitochondrial genomes in higher plants are known to range from 200 to 2400 kilobase pairs (kbp; Ward et al., 1981). However, because of the detection by electron microscopy of much smaller mtDNA molecules (approximately 30 kbp) than would be expected from the size of the genome (Sparks and Dale, 1980; Wong and Wildman, 1972), the large sizes of the mt-genomes of higher plants have been attributed to heterogeneity of the mtDNA molecules. On the basis of the physical maps of mtDNA from Brassica (Palmer and Schield, 1984) and maize (Lonsdale et al., 1984), it has been proposed that these small mtDNA molecules are derived from one large, circular molecule called the ‘master genomic circle’ of 218 kbp in Brassica and 570 kbp in maize. By contrast, it has been revealed by conventional microscopy combined with photometry that each mitochondrion squashed out from cells of Allium cepa and Nitellaflexilis contains one mt-nucleus, and the intensity of the fluorescence from each mt-nucleus is approximately 40% (70 kbp) of that of T4 phage (170 kbp; Kuroiwa, 1982; Nishibayashi et al., 1987). Therefore, there appears to be a discrepancy between the large size of the mitochondrial genome and the small amount of DNA per individual mitochondrion.

In the past ten years, studies of the molecular biology of the mtDNA of higher plants have made remarkable progress, and the presence of a DNA-protein complex called the ‘mt-nucleus’ in the mitochondria of Physarum polycephalum has also been firmly established (Kuroiwa, 1982; 1991). Two distinct events appear to be involved in the division of mitochondria in P. polycephalum-. the duplication of electron-dense mt-nuclei that contain replicated mtDNA and the division of the matrix and cristae, which is referred to as ‘mitochon-driokinesis’. With respect to the function of organelle DNA in meristematic cells of higher plants, the most direct method for evaluating its function involves the visualization of mt-nuclei in sections, and the assessment of the amount of mtDNA in mt-nuclei and the frequency of labeled mt-nuclei in various regions of the meristem. However, it has been difficult to visualize mt-nuclei that contain extremely small amounts of DNA in thin sections. Therefore, no direct information has been available to date on the mode of profiferation of mitochondria in higher plants. Recently, it has become possible to observe mt-nuclei in sections prepared in a new resin with good permeability to solutions of DAPI (Kuroiwa et al., 1990). This method has allowed visualization of the large mt-nuclei that contain 2300 kbp of DNA in the cells of ovules of Pelargonium zonale (Kuroiwa et al., 1990).

We wished to explain the discrepancies associated with the sizes of mitochondrial genomes and how it is that differentiated cells contain mitochondria with only a small amount of mtDNA, and to reveal the mode of proliferation of mitochondria in roots that contain a typical meristematic tissue. We therefore examined simultaneously the number of mitochondria per cell, the number of mt-nuclei per mitochondrion, the amount of DNA per individual mitochondrion and the labeling index of mitochondria in the root meristem of P. zonale, after embedding samples in Technovit 7100 resin and staining them with DAPI, in combination with light-microscopic autoradiography and microphotometry.

Seeds of P. zonale were germinated on moist filter paper in Petri dishes. Seedlings with primary roots of approximately 4 mm length were used for each experiment. After immersion in an aqueous solution of 7.4 × l(r Bq/ml (spec. act. 37 × 10⁹ Bq/mmol) [³H]thymidine (Daiichi Pure Chemical Company, Tokyo, Japan) for 30 min, 60 min, 2 h, 4 h, 5 h, 8 h, 12 h and 24 h, the root tips were fixed in 4% formaldehyde, buffered with cacodylate at pH 7.2, for 12 h at 20°C, and then dehydrated through an ethanol series. Then the roots were embedded in Technovit 7100 resin in accordance with the manufacturer’s instructions (Kulzer and Co., GmbH, Wehrheim, Federal Republic of Germany; Kuroiwa et al., 1990). Thin serial sections, 0.5 μm and 1 μm thick, were cut with a Diatome diamond knife on an MT-6000 XL Ultramicrotome (RMC-Eiko Corp., Kawasaki, Japan).

Sections of root tips were placed in a drop of distilled water on a glass slide and air-dried. The samples were coated with Sakura NR-H2 liquid emulsion (Konica Ltd., Tokyo) by dipping in a darkroom. Samples were exposed for one or two months. After exposure, the samples were developed in Sakura Konidole Fine and fixed with Sakura Konifix (Konica Ltd.).

The meristem was examined by superimposing five longitudinal median sections of roots and drawing the sites of mitoses and labeled cell nuclei on the same tracing paper. Silver grains, within an arbitrary unit area of cytoplasm after treatment with [³H]thymidine for 60 min and exposure for 2 months, were counted and the labeled areas were classified according to the number of silver grains per unit area of cytoplasm: heavily labeled cytoplasm (more than 51 grains/ unit area), less heavily labeled cytoplasm (26–50 grains/unit area) and lightly labeled cytoplasm (10–25 grain/unit area), respectively.

The frequencies of labeled nuclei and labeled mitoses were examined in longitudinal median sections of root meristem fixed at various times after initiation of continuous labeling with [³H]thymidine. On the basis of the data obtained, the duration of a total cycle and of the M, G₁, S and G₂ phases was calculated.

For observation of DNA, sections were stained with DAPI dissolved at a concentration of 1 μg/ml in S buffer (0.25 M sucrose, 1 mM EDTA, 0.6 mM spermidine, 0.05% 2-mercaptoethanol, 10 mM Tris-HCl, pH 7.6; Kuroiwa and Suzuki, 1980). For observation of DNA and RNA, they were first stained with DAPI and then with acridine orange at a concentration of 2 μg/ml in S buffer. Then all samples were observed under an Olympus BHS-RFK epifluorescence microscope (Kuroiwa et al., 1986). Photographs were taken at magnifications of ×330 and ×500 on 35 mm Fuji Neopan 400 film and on 35 mm Kodak Ektachrome 400 film for color slides. For observation of both the fluorescence microscopic image and the electron microscopic image in the same field, sections of root tips 0.3 μm and 0.5 μm thick were cut and placed on a drop of distilled water on a nickel grid and airdried. The samples were stained with DAPI, observed by fluorescence microscopy and photographed according to the method described above. After the cover slip was carefully removed, the grid was washed in distilled water and stained with both uranyl acetate and lead citrate for electron microscopy.

Longitudinal median sections 1 μm thick were stained with both DAPI and acridine orange and then the numbers of mitochondria present in those cells that contained the section of nuclei passing through nucleoli were counted and the numbers integrated according to the thickness (15 μm -35 μm) of cells in regions at various distances from the tip of the root. The approximate numbers of mitochondria per cell during the S phase were estimated.

The longitudinal median sections of root meristems were stained with DAPI at a concentration of 1 μg/ml. The amounts of DNA in the mt-nuclei and cell nuclei during the S phase were examined directly with VIMPICS (Hamamatsu Photonics Ltd., Hamamatsu, Japan) connected to an Olympus epifluorescence microscope BHS-RFK, as described previously (Kuroiwa et al., 1986). VIMPICS allows fluorimetry of extremely small amounts of DNA. Intensities of fluorescence emitted from the background and the mt-nuclei can easily be measured by moving the measuring square or lines on the TV monitor. A 490 nm short-pass filter was used for quantification of DAPI fluorescence. For standarization of the fluorescence intensity of specimens, T4 phage (about 170 kbp; Freifelder, 1970) was employed. The fluorescence intensities of the samples were expressed as approximate multiples of that of T4 phage (170 kbp), with the arbitrary unit T used for the expression of DNA content (Kuroiwa et al., 1986) because the GC content of mtDNA of P. zonale was similar to that of T4 phage (Suzuki and Kuroiwa, unpublished data). G ₁ cells were easily distinguished from G₂ cells by VIMPICS. Intensities of fluorescence from sections of G₁ nuclei passed through the nucleoli were similar to those of late telophase cells, while intensities from G₂ nuclei were similar to those of early prophase cells. Often early G₁ daughter cells made a pair.

After treatment with [³H]thymidine for 60 min and exposure for 2 months by the method described above, the number of labeled mitochondria per cell at various phases during the mitotic cycle and at various distances from the tip of the root, in longitudinal median sections of the root, were counted and the percentage of labeled mitochondria was calculated.

The root tips were fixed in 2% glutaraldehyde, buffered with cacodylate, pH 7.2, for 2 h at 23°C, washed in cacodylate buffer and post-fixed in 2% OsO₄ in the same buffer for 2 h at 23°C. The fixed materials were then dehydrated in a graded series of ethanol and propylene oxide (30 min at each step). The samples were embedded in Spurr’s resin. Ultra thin sections, 0.07 μm in thickness, were cut with the Diatome diamond knife on the MT-6000 XL Ultramicrotome and mounted on Formvar-coated grids. The sections were stained with 2% uranyl acetate solution for 30 min and post-stained with lead citrate for 7 min (Nishibayashi et al., 1987). These sections were examined with a JEM-1200 EX II electron microscope (JEOL, Tokyo, Japan).

Fig. 1A,B shows the distribution of mitoses and labeled nuclei in the root apex of P. zonale after treatment with [³H]thymidine for 60 min. The mitoses and labeled cell nuclei are distributed in the region between 150 μm and 700 μm from the tip of longitudinal median sections of roots. We define this region here as the root meristem, because the region at the root apex called the ‘meristem’ is different for each investigator (Buvat, 1988; Clowes, 1961). By contrast, the cells with heavily labeled cytoplasm are located in the region just above the quiescent center (QC) and pericycle of the central cylinder (Fig. 1C,D). The results suggest that the synthesis of organelle DNA, such as mtDNA and plastid (pt-) DNA, does not occur in the entire region but only in a limited region of the root meristem.

Since the profiles of about 432 cells appeared in a cross section 300 μm from the tip of the root, it seems likely that the root meristem consists of about 432 rows of cells that are arranged in a longitudinal direction. The cells in the rows have a variety of shapes -cubic, cuboidal, etc. The pattern of labeled cytoplasm and the structures of such root cells show that the tissues of the root are not uniform. Therefore, to simplify matters, we shall deal mainly with the cells of the cortex during the S phase as a model for cell proliferation, because these cells are relatively cubic and the cell nuclei are spherules or ellipsoids throughout the meristem, unlike the cells of the central cylinder and epidermis and their nuclei.

Since three sides of the cubes of cortex cells during the S phase were between 10 μm and 35 μm in length (unpublished data), a row of cells in the cortex consisted of a linear gradient of about 40 sequentially older cells, with the oldest cells near the elongation zone and the youngest ones nearest to the QC. In one column, cells just above the QC should reach the meristematic region just below the elongation zone after at least five cell divisions.

Fig. 2A shows the large numbers of mt-nuclei-like and pt-nuclei-like fluorescence spots around the cell nuclei in the zone just above the QC after staining with DAPI. The mt-nuclei-like spots can be distinguished from the pt-nuclei-like spots in thin sections of the root, since the mt-nuclei-like spots emit stronger blue-white fluorescence than that emitted by the pt-nuclei-like spots, as described previously (Kuroiwa et al., 1990). To reveal clearly the distribution of mt-nuclei-like and pt-nuclei-like spots, the root meristem was stained with both DAPI and acridine orange (Fig. 2B). The mt-nuclei-like spots are included in rod-shaped or small spherical structures, while pt-nuclei-like spots are in more voluminous ovoid structures.

To confirm that the mt-nuclei-like spots and ptnuclei-like spots are mt-nuclei and pt-nuclei, respectively, mitochondria and plastids in ultra-thin sections embedded in Spurr’s resin were observed by electron microscopy, and then their fluorescence microscopic image due to staining with DAPI was compared with the electron microscopic image in the same field as sections embedded in Technovit 7000 resin. A cell from a root tip is shown in Fig. 3A. Since the cells at this region have not elongated and contain almost all of the organelles, they are an example of plant cell structure as seen by electron microscopy. The mitochondria, spherical or filamentous organelles present in the cytoplasm around the cell nucleus, are characterized by cristae and somewhat electron-dense, fibrous mt-nuclei. DNA-like fibers are observed around the mt-nuclei (Fig. 3B). Plastids (leucoplasts) are ovoid, doughnut- or cupshaped structures, containing electron-dense matrix, electron-dense inclusions and a few thylakoid membranes, and they occur in relatively small numbers (Fig. 3A,B). After staining with DAPI, mt-nuclei-like spots emit stronger fluorescence than the pt-nuclei-like spots (Fig. 3C). When mt-nuclei-like spots and pt-nuclei-like spots were examined in the same field by electron microscopy, they corresponded to spherical or filamentous structures and voluminous plastids, respectively (Figs. 3C,D). At higher magnification, cristae-like structures were observed in the spherical or filamentous structures (Fig. 3E). In addition, when the thin, 0.3 μm section of the cell was slightly swollen on the surface of distilled water, electron-dense cores, consisting of DNA-like fibers, become visible in the central area of the spherical or filamentous structures, the mitochondria (Fig. 3F,G). In other words, the electron-dense fibrous cores are equivalent to those spots identified as mt-nuclei by fluorescence microscopy. On the basis of these observations, we confirm that mt-nuclei-like spots and pt-nuclei-like spots are equivalent to mt-nuclei and pt-nuclei, respectively. Mitochondria and plastids containing organelle nuclei could be observed in cells at various phases of the mitotic cycle, such as mitosis and interphase (Fig. 3B). The mitochondria in the cortex cells are smaller than those of the central cylinder and pericycle and, in general, each mitochondrion contains one mt-nucleus, except that some large mitochondria in the cells close to the QC each contain a few mt-nuclei. The number of pt-nuclei per plastid is indistinct. There are approximately 150 to 500 mitochondria per cell during the S phase, and this does not change markedly within the column of root meristem cells (Fig. 4A). The results suggest that the mitochondria can divide regularly in all the cortex cells of the meristem.

Mt-nuclei in the cells in the region just above the QC emitted stronger blue-white fluorescence than those of meristematic cells in the upper region of the root meristem (Figs 4A, 5C,F). To confirm these observations quantitatively, levels of mtDNA in individual mt-nuclei in mitochondria of cells during S phase, (selected as a representative phase) were measured by VIMPICS (Fig. 4B). The amount of mtDNA in the mt-nuclei is very low in mitochondria of the root cap. It is markedly higher in mitochondria just above the QC, and reaches approximately 3000 kbp (20 T), but soon decreases rapidly as the distance between the mitochondria and the QC increases. The amount of mtDNA falls to less than 170 kbp (T) in the mitochondria located near the elongation zone (Fig. 4B). Similar tendencies are observed in the mitochondria of cells in the G ₁, G₂ and mitotic phases (Figs 5 and 6). The results suggest that, since mitochondrial division occurs in all the cortex cells of the meristem but the synthesis of mtDNA ceases in all cells except those near the QC, the amount of DNA in mitochondria must be diluted stepwise by continuous mitochondrial divisions. Furthermore, the results imply that the absolute values of mtDNA in mt-nuclei are a consequence of increases due to the synthesis of mtDNA and reductions due to divisions of mitochondria. Such tendencies shown by cells during S phase are also associated with mitotic cells (Figs 5 and 6). This hypothesis can be verified directly by studying the synthesis of mtDNA in the meristem.

To clarify the distribution of silver grains in root meristems, organelle nuclei were stained with DAPI, or with both DAPI and acridine orange, and observed under an epifluorescence microscope equipped with VIMPICS. Two or three separate images of cells in the root meristem, namely fluorescence, bright-field and the combined bright-field/fluorescence images, in the same field, were photographed on the TV monitor of VIMPICS. Fig. 5 shows representative video-frame photographs of autoradiograms of mt-nuclei and cell nuclei of cells in S phase, just above the QC (Fig. 5A-C) and near the elongation zone (Fig. 5D-F), after incorporation of [³H]thymidine into roots for 60 min. The mt-nuclei of cells during the S and G₁ phase, just above QC, emit strong fluorescence (Fig. 5A) and are also heavily labeled (Fig. 5B,C), while mt-nuclei of cells during S phase near to the elongation zone are visible only as tiny spots (Fig. 5D) and are not labeled (Fig. 5E,F). To confirm these observations quantitatively, labeled mitochondria were examined through the cortex from the zone just above the QC to the region near the elongation zone (Fig. 4C). The results indicate that, among those located near the QC, more than 95% of mt-nuclei in the cells during S and G₁ phases are labeled, but the frequency of labeled mt-nuclei decreases rapidly as the distance of cells from the region near the QC increases. The heavy deposition of label is also evident on the mt-nuclei in cells just above the QC during other phases, such as prophase and the G₂ phase, (Fig. 6A,B). The metaphase cells are also amenable to observations of such trends. Fig. 6C-H shows metaphase cells in the cortex at 250 μm (C,D), 400 μm (E,F) and 550 μm (G,H) from the tip of the root. Fluorescence intensities of individual mt-nuclei around the metaphase chromosomes in the equatorial plane decrease rapidly, with a remarkable reduction in label as the metaphase cells become more distanced from the QC. Accordingly, the results show that the mitochondria in the area near the QC can synthesize their DNA throughout the mitotic cycle.

The duration of the generation time and of the G₁, G₂, S and mitotic phases of the meristematic cells can be estimated as 17 h, 4.6 h, 6.0 h, 5.0 h and 1.4 h, respectively, by continuous labeling with [³H]thymi-dine, and the mitotic index is about 8% (Table 1). The patterns of labeling of cells in metaphase were examined at about 300 μm from the tip after incorporation of [³H]thymidine into roots. After 6 h, 8 h 10 h and 12 h, the metaphase chromosomes are labeled but the mt-nuclei are not labeled. Even after 24 h, the mitochondria are weakly labeled (Fig. 6l,J). Such findings support our results and strongly suggest that, in the cells just above the QC, mitochondria can divide after synthesis of their DNA, but when cells are at some distance from the QC, the division of mitochondria takes place in the absence of the synthesis of mtDNA.

Essentially similar trends to those observed in the cortex of the meristem can be observed in the epidermis and the central cylinder. However, they are not found in the pericycle of the central cylinder, where mt-nuclei are large and the synthesis of mtDNA continues in the entire meristem, even though the frequency of labeled mt-nuclei decreases gradually (Fig. 1D).

In the particular region in which mtDNA is synthesized actively, we examined whether or not the division of mitochondria and the synthesis of mtDNA occur during a specific phase of the mitotic cycle (Table 2). The number of mitochondria present per cell increases continuously throughout the mitotic cycle after cell division, suggesting that some mitochondria are capable of division independently of the mitotic cycle. In addition, approximately 90% of mitochondria in cells at various phases of the mitotic cycle are labeled after treatment with [³H]thymidine for 60 min. The label is found in mitochondria of cells in mitosis and in the G₁ phase. Thus, the synthesis of mtDNA occurs both prior to and during mitosis. Although the synthesis of mtDNA tends to decrease during mitosis, it occurs quite independently of the mitotic cycle (Table 2). However, as shown in Fig. 4C, the percentage of labeled mt-nuclei in mitotic cells decreases earlier than in other interphase cells. Therefore, the data suggest that the division of mitochondria and the synthesis of mtDNA in the region just above the QC occur at a relatively uniform rate throughout the mitotic cycle. Moreover, mitochondria in mitotic cells first cease the synthesis of mtDNA as the cells become separated from the region just above the QC.

In meristematic cells, mitochondria and plastids have approximately the same morphology, so that, as described by Buvat (1988) it is difficult to discriminate between them with a light microscope. Since the mitochondria of higher plants do not contain electrondense mt-nuclei in which mtDNA is organized in tight complexes with proteins, as found in mitochondria of P. polycephalum (Kuroiwa et al., 1977), it has been difficult to determine the behavior of mt-nuclei in plant tissues during the cell division cycle by means of light and electron microscopy.

In our present experiments, it has been possible to focus on particular regions of the root meristem of P. zonale and to observe the behavior of mt-nuclei containing extremely small amounts of DNA, in thin sections of the root meristem. The observations were made possible by embedding the samples in Technovit 7100 resin and double-staining with DAPI and acridine orange, in combination with light-microscopic autoradiography and microphotometry. In the root meristem of P. zonale, synthesis of cell-nuclear DNA and cell division occur actively in the entire root meristem 150 μm to 700 μm from the tip, as is typical of root meristems of dicotyledons. By contrast, the mtDNAs are synthesized in a specific region just above the QC, 150 μm from the tip. Meristematic cells contain 150-500 mitochondria and the division of mitochondria occurs throughout the meristem. These observations suggest that the amount of mtDNA per mitochondrion is reduced stepwise after each mitochondrial division. Moreover, the cytophotometric studies reveal that each mitochondrion in the cells just above the QC contains approximately 3000 kbp of DNA, while the amount of mtDNA per mitochondrion in the meristematic cells just below the elongation zone is less than 150 kbp. The specific region, which covers the QC, may correspond to the region in which so-called T (or Y) divisions occur more conspicuously (Clowes, 1961) and has an inverted cup-shaped profile. The specific region generates, respectively, the cortex and the central cylinder.

On the basis of experiments with restriction enzymes, the sizes of the mt-genomes of many higher plants have been estimated to range from 200 to 600 kbp (Ward et al., 1981). The size of the mt-genome of P. zonale is probably within this range, since the amount of mtDNA per mitochondrion in the differentiated cells of P. zonale is less than 170 kbp, and is very similar to that in other plants such as A. cepa and N. flexilis (Nishibayashi and Kuroiwa, 1985). By contrast, because of the detection by electron microscopy of much smaller mtDNA molecules (approximately 30 kbp) than would be expected from the size of the genome (Sparks and Dale, 1980; Wong and Wildman, 1972), the large sizes of the mt-genomes of higher plants have been attributed to heterogeneity of the mtDNA molecules. On the basis of the physical maps of mtDNA from Brassica (Palmer and Schield, 1984) and maize (Lonsdale et al., 1984), it has been proposed that the small mtDNA molecules are derived from one large, circular molecule called the ‘master genomic circle’ which is 218 kbp long in Brassica and 570 kbp long in maize.

If there is a master genomic circle (of about 300 kbp) of mtDNA in the mitochondria of many higher plants, including P. zonale, mitochondria may be highly polyploid with respect to the number of copies of mtDNA (12 or more copies per mitochondrion). In P. zonale, the mitochondria with the highest ploidy must be localized in the region just above the QC. By contrast, the mt-nuclei containing a small amount of DNA (i.e., less than that in the master genomic circle) are observed in the root cap and elongation zone. These phenomena can be explained as follows. In the meristematic cells, except those in the specific region just above the QC, the division of mitochondria takes place continuously in the absence of synthesis of mtDNA and, as a result, the number of copies of mtDNA must decrease to less than one master circle after about five mitochondrial divisions. In the meristematic region near the elongation zone, the master genomic circle probably divides into subcircles of smaller sizes, as observed by electron microscopy, and such subcircles are separated into daughter mitochondria. Therefore, mitochondria in differentiated cells contain a smaller amount of mtDNA than the master genomic circle. Similar small mt-nuclei, containing smaller amounts of DNA, were revealed in studies using a Zeiss microscope photometer, in which each mitochondrion squashed out from epidermal cells of A. cepa and from intemodal cells of N. flexilis contained one mt-nucleus, and the intensity of the fluorescence from each mt-nucleus was approximately 40% (70 kbp) of that of T4 phage (Nishibayashi and Kuroiwa, 1985). Such data and the present results support the idea that mitochondria in differentiated tissues tend to contain a small amount of mtDNA which corresponds to a few subcircles (Kuroiwa et al., 1990).

The division of mitochondria and of plastids appears to be synchronized in normal tissues of higher plants (Leach and Pyke, 1988) and in unstressed slime mould (Kuroiwa et al., 1977).

Kawano et al. (1983) showed that, during differentiation from syncytium plasmodium to sclerotium of the slime mold P. polycephalum, the electron-dense mt-nuclei, containing 32 copies (1920 kbp) of the mt-genome, could divide in the absence of synthesis of mtDNA and, thus, the number of copies per mitochondrion was reduced to 4 copies (240 kbp) by continued divisions. In P. polycephalum, the mt-genome is not subdivided into two or three separate circles. The division of organelles in the absence of DNA synthesis occurs in the chloroplast of the red alga Cyanidium caldarium (Kuroiwa et al., 1989). During the first 40 h after the initiation of a synchronous culture of young cells, the chloroplasts increase markedly in size, and DNA content per pt-nucleus increases to approximately 16 times the value in 16-endospore cells. After four endospore divisions, the volume of each cell and the amount of ptDNA is reduced stepwise after each endospore division until finally, at the 16-endospore stage, they reach approximately 1/16 of the original values for the mother cells.

Possingham et al. (1988) suggested, on the basis of observations of chloroplasts in cultured leaf discs of spinach, that the chloroplasts of young leaves divide at random, with respect both to time and to each other, so that rates of division are similar throughout the 24 hour cell cycle. Therefore, when the cells are activated, as appears to occur in part of the meristem and after explanting of leaf discs, division of organelles and the synthesis of mtDNA and ptDNA may occur independently of the mitotic cycle.

Fig. 7 provides a summary of our observations and some speculations. We strongly emphasize that, for example, cells during the mitotic phases, which appear in various regions from the tip of the root, are very similar to each other with respect to only the cell-nuclear division cycle, but differ from each other with respect to both the cell-nuclear division cycle and the organelle division cycle.

To date, the differentiation of cells from the meristem to differentiated tissues in plants has been discussed in morphological terms, with emphasis on the evolution of the vacuolar apparatus, mitochondria, plastids and the DNA content of cell nuclei (Buvat, 1988). As shown by the present results, the structure of organelle nuclei and their DNA content should also be recognized as important parameters in cell differentiation.

This work was supported by grants nos. 02304007 and 63440003, by a grant for Special Research on Priority Area (project no. 032576105, Molecular Structure of Chromosome) from the Japanese Ministry of Education, Science and Culture.