Cell division and cell enlargement during potato tuber formation

Transcription

1 Journal of Experimental Botany, Vol. 49, No. 320, pp , March 1998 Cell division and cell enlargement during potato tuber formation Xin Xu1,2,3, Dick Vreugdenhil1 and André A.M. van Lammeren2,4 1 Department of Plant Physiology, Wageningen Agricultural University, Arboretumlaan 4, 6703 BD Wageningen, The Netherlands 2 Department of Plant Cytology and Morphology, Wageningen Agricultural University, Arboretumlaan 4, 6703 BD Wageningen, The Netherlands 3 Institute of Vegetables and Flowers, Chinese Academy of Agricultural Sciences, Beijing, PR China. Received 12 May 1997; Accepted 7 November 1997 Abstract Key words: Cell division, cell enlargement, DNA synthesis, in vitro culture, potato, tuber formation. Introduction Cell division and cell enlargement were studied to reveal the developmental mechanism of potato tuberization using both in vivo and in vitro culture systems. Distribution of cells in S-phase was visualized by immunolabelling of incorporated bromodeoxyuridine (BrdU). Mitosis was detected in DAPI (4,6-di-amidino- 2-phenylindole) or toluidine blue-stained sections. Timing and frequency of cell division were determined by daily cell counting, and cell enlargement was deduced from measurements of cell diameters. Under in vivo condition, lateral underground buds developed into stolons due to transverse cell divisions and cell elongation in the apical region of the buds. At the onset of tuber formation, the elongation of stolons stopped and cells in pith and cortex enlarged and divided longitudinally, resulting in the swelling of the stolon tip. When tubers had a diameter of 0.8 cm, longitudinal divisions had stopped but randomly oriented division and cell enlargement occurred in the perimedullary region and continued until tubers reached their final diameter. In vitro tubers were formed by axillary buds on single node cuttings cultured under tuber-inducing condi- tions. They stopped growing at a diameter of 0.8 cm. Pith and cortex were involved in tuberization such as that found during the early stage of in vivo tuberization (<0.8 cm in diameter). The larger size of in vivo tubers is, however, due to further development of the perimedullary region, which is lacking in in vitro conditions. The formation of potato tubers comprises two different aspects: (a) the morphological development of the tuber, and ( b) the biochemical changes resulting in the formation and storage of starch. The latter process has been extensively studied during the past decades, but the morpholo- gical aspects have received much less attention. At the morphological level, the process of tuber formation results from two separate steps, i.e. stolon develop- ment and tuberization at the stolon tip (Booth, 1963). It is generally assumed that the longitudinal growth of stolons stops as soon as the thickening of the stolon tip starts (Leshem and Clowes, 1972; Cutter, 1978; Peterson et al., 1985; Vreugdenhil and Struik, 1989). However, other observations showed that the continued expansion of the tuber depends on the production of new internodes from the apex (Artschwager, 1924; Goodwin, 1967). Because the literature is not equivocal in this matter, it was decided to investigate the two processes by visualizing mitosis and studying the expansion of the cells to clarify the switch from stolon elongation to radial growth. Both cell division and cell enlargement contribute to the development of tubers (Plaisted, 1957; Booth, 1963; Reeve et al., 1973b; Cutter, 1978; Peterson et al., 1985). It is not certain whether the initial radial expansion of stolons is brought about by cell divisions or by cell 4 To whom correspondence should be addressed. Fax: andre.vanlammeren@algem.pcm.wau.nl Abbreviations: BrdU bromodeoxyuridine, CLSM confocal laser scanning microscopy, DAPI 4,6-di-amidino-2-phenylindole, FITC fluorescein isothiocyanate, PI propidium iodide. Oxford University Press 1998

2 574 Xu et al. enlargement. Some workers observed that mitotic activity Stolons and tubers, formed from the axillary buds, were occurred before an increase of cell size was detected sampled under green safe light. In vivo potato plants (Solanum tuberosum cv. Bintje) were ( Reed, 1910; Artschwager, 1918, 1924; Reeve et al., 1969; grown from tubers. When tubers had formed 20 cm shoots, Duncan and Ewing, 1984). Other observations indicated whole plants were transplanted to a liquid culture system that the early radial expansion was caused by an increase (Helder et al., 1993) and grown in a greenhouse (day length: in cell diameter ( Booth, 1963; Cutter, 1978; Peterson and 18 h; temperature: 20/18 C day/night). Stolons grew in Barker, 1979; Peterson et al., 1985; Sanz et al., 1996). sand5perlite (151 v/v) medium. Harvest of young tubers started after 6 weeks of culture. Hence, the timing of cell division and cell enlargement is unclear. BrdU incubation During the development of tubers, several types of Single-node cuttings were used to analyse nuclear DNA tissues are involved in cell multiplication and cell enlarge- synthesis by the detection of incorporated bromodeoxyuridine ment. The early idea that the new thin-walled storage (BrdU ). Before day 5, five newly formed stolons together with tissue was derived from cambial activity (De Vries, 1878) some stem tissue were cut each day. From day 5 to day 14, five tubers with the upper parts of the stolons (in total ±1.5 cm) is no longer accepted. Most later workers ascribed tuber were excised each day. The explants were then cultured formation to the increase in number and size of paren- immediately on BrdU-containing medium (15500 (v/v) BrdU chymatous cells in pith, cortex and perimedullary tissue in liquid inducing medium containing 8% sucrose). A Petri dish (Artschwager, 1918, 1924; Hayward, 1938; Plaisted, 1957; (diameter=5 cm) containing the culture medium was covered Reeve et al., 1969; Leshem and Clowes, 1972; Cutter, with parafilm through which the cut stolon end was put in the medium. The tips of the stolons or the tubers were above the 1978; Peterson et al., 1981). However, no consistent parafilm to get better evaporation. The Petri dish was then put description of these processes of tuber formation exists. into a closed jar to avoid over-evaporation. After 20 h Some workers mentioned cell divisions in the pith during incubation at 20 C in the dark, samples were harvested the initial phase of tuberization ( Hayward, 1938; for fixation. Bradbury, 1953; Plaisted, 1957; Reeve et al., 1969). Meanwhile others stated that the parenchyma cells in the Immunolabelling of BrdU pith divided only occasionally, and that the frequency of Samples were fixed according to a protocol described by Sanz et al. (1996). Briefly, they were fixed in 4% paraformaldehyde such divisions was negligible ( Reeve et al., 1973b). (PFA) in microtubule stabilizing buffer (MSB: 0.1 M Peterson and Barker (1979) did not observe longitudinal 1,4-piperazinediethanesulphonic acid (PIPES), 1 mm ethylene cell divisions in in vitro tubers. It is not clear whether glycol bis(2-aminoethyl ether)-n,n,n,n -tetraacetic acid longitudinal cell divisions occurred in the pith or not. (EGTA), 1 mm MgCl, and 0.4% polyethyleneglycol (PEG 2 Although it is widely accepted that the growth of the 6000), ph 6.9) for 2 h, rinsed in MSB for 3 30 min, dehydrated in an ethanol series and embedded in PEG Some samples perimedullary zone produces the largest portion of the were embedded in 4% agar after the fixation and rinsing steps. tissue in mature tubers (Hayward, 1938; Booth, 1963; The PEG sections (14 mm) and agar sections (60 mm) were Reeve et al., 1969; Cutter, 1978; Peterson et al., 1985), stretched in phosphate buffer saline (PBS: 8 g NaCl, 0.2 g KCl, as far as we know, no analysis has been made of the 0.2 g KH PO and 1.41 g Na HPO.2H O in 1000 ml H O), contribution of the various tissues to the growth of the put on organosilane coated slides, and dried at room temper- ature before labelling. tuber at different stages of development. 151 (v/v) antibrdu (Amersham) in 0.1% acetylated bovine In this work, in vitro tuberization was studied using serum albumin (BSAc) in PBS was applied as the first antibody, single node cuttings. The uniform growth of these tubers and FITC conjugated goat anti-mouse antibody (15200, v/v) allowed a detailed investigation of the successive events as the second antibody. Before immunolabelling, sections were of tuberization in time and in space. Because the final incubated in 2 N HCl for 20 min to unscrew the double strand- DNA and make the incorporated BrdU accessible to the size of in vitro tubers is largely different from tubers antibrdu (Gunning and Sammut, 1990). Nuclear DNA was grown on the plant in soil, this study was extended to a stained with 1 mgml 1 propidium iodide (PI) for 5 min after comparison between these two types of tubers, in order labelling with the second antibody. to detect the cause of the difference in growth pattern. Immunolabelled sections were enclosed in Citifluor and analysed with a Bio-Rad MRC600 Confocal Laser Scanning Microscope (CLSM ). BrdU-labelled nuclei were distinguished Materials and methods from PI stained nuclei by applying dual channel mode. Plant material Four-week-old in vitro potato plants (Solanum tuberosum cv. Bintje), propagated using stem cuttings, were transplanted to soil in a growth chamber at 20 C. They were first treated with long day illumination (16 h) for 4 weeks and then with short day illumination (8 h) for 3 weeks. Single node cuttings without leaves were cultured (Hendriks et al., 1991) on an inducing medium (Murashige and Skoog, 1962) containing 8% (w/v) sucrose and 5 mm benzylaminopurine in darkness at 20 C. Technovit sectioning and cell counting Five pieces of stolon tips or tubers (about 1 mm thick) were sampled each day at day 0 8, 10 and 14 of in vitro culture and at the size of 0.2, 0.3, 0.8, 1, 2, and 3 cm in diameter of in vivo culture. Longitudinal and transversal sections of 1 mm thickness were fixed in 2.5% glutaraldehyde and 2.5% PFA in 0.1 M phosphate buffer (ph 7.2) for 2 h, washed in the buffer for 4 15 min and then in H O for 2 15 min, dehydrated in a 2 series of ethanol, and embedded in Technovit Both

3 Potato tuber cell division and enlargement 575 longitudinal and cross-sections were made. They were stained stopped and radial growth of the stolon tip started at day with 0.01% (w/v) Calcofluor White or 1 mgml 1 DAPI (4,6-di- 5. Tuber swelling was first observed around the first node amidino-2-phenylindole), or both, to visualize cell walls and nuclei. Other sections were stained with 1% (w/v) toluidine blue above the two basal leaves. The swelling part included (in 1% NaB O.10H O) and used for cell counting. the 2 mm long upper part of the first internode and a Several data were collected from the longitudinal sections of part of the second internode (Fig. 1C). When the basal in vitro tubers: (a) tuber diameter (D); (b) cell numbers along end of the swelling tuber was marked with ink, the the tuber diameter (n ); (c) cell numbers along the long axis of D position of the marked spot did not change with respect the whole developing bud (n ). Average cell widths were L calculated by dividing the tuber diameter by the cell numbers to the site of attachment of the stolon during the further along the tuber diameter (D/n ). period of tuber growth, indicating that the tuber D Data collected from cross-sections of in vivo tubers included developed acropetally. Tubers stopped growing between cell number and thickness of the following regions: (a) cortex; day 10 and day 15. At this stage, the average length and (b) perimedullary zone; (c) pith. Average cell widths were width of the tuber was 0.8 cm. It was formed from the deduced by dividing the thickness of a region by the cell 2 3 mm long upper part of the first internode forming numbers in that region. the 1/3 basal part of the tuber, the whole second internode forming the middle part of the tuber, and the third Results internode forming the uppermost part of the tuber. The Morphological observation of tuber formation internodes above the third one neither elongated nor swelled ( Fig. 1D). When single-node cuttings of potato plants were cultured During in vivo tuber development, the length of the in the dark on a medium with 8% sucrose and without stolon and the numbers of internodes along the elongating GA, buds developed into stolons with tubers. At day 0, stolon were much more variable. However, the swelling buds were less than 1 mm long. In the first 4 d, the buds was always observed in the stolon tip starting in the grew to about 1 cm long stolons with about eight inter- upper part of the last well elongated internode ( Fig. 2A, nodes ( Fig. 1B). The two basal leaves never grew. The B). At least eight further internodes had developed in the first internode (numbered from base to apex of the stolon tip before tuber swelling started, much like the developing bud) above the two basal leaves is the only development in vitro. Until the size of 0.8 cm in diameter, well elongated one to form the stolon. Stolon elongation the morphology of the in vivo tuber was the same as that Fig. 1. Diagram illustrating the morphology of growing axillary buds of potato cuttings cultured in vitro. The distribution of BrdU-labelled nuclei is shown at the left-hand side in the drawings and all nuclei at the right-hand side. Nuclei are indicated schematically by ink dots. Position of nodes are demonstrated schematically. The numbers indicate the nodes. Bar=1 mm. (A) Stolon at day 3 with three elongating internodes. BrdU-labelled nuclei were mainly observed in subapical region. The part indicated by box a/b is shown in Fig. 4a, b. (B) Stolon at day 4. More BrdU-labelled nuclei were in the basal part of the subapical region. (C) Stolon swelling around the first node at day 5. BrdU-labelled nuclei were only present in the swelling part. The parts indicated by boxes c and d are shown in Fig. 4c, d. (D) Almost full-grown tuber at day 8, containing the upper part of the 1st internode, the 2nd internode and the 3rd internode. BrdU-labelled nuclei were only observed in periderm and vascular tissue. The parts indicated by boxes e and f are shown in Fig. 4e, f.

4 576 Xu et al. Fig. 2. Diagram of longitudinal sections through in vivo grown potato tubers, showing the morphology of the stolon and tuber and the thickening of the perimedullary zone (dark-shaded area). Positions of nodes are indicated schematically. The numbers indicate the nodes. Bar=1 cm. (A) 0.2 cm stolon, showing the continuous vascular bundle. (B) 0.3 cm tuber, showing the growth of pith ( light-shaded area). (C) 0.8 cm tuber, showing the onset of growth of the perimedullary region. (D) 2.0 cm tuber, showing the thickening perimedullary region. of the in vitro tuber at day 10 (Fig. 2C). Up to the end shown) and the cell number along the long axis of the of tuber growth, more than half of the basal part of the stolon remained unchanged ( Fig. 3D). Cell width in the tuber was formed by the proximal part of the last subapical region increased from day 4 onward ( Fig. 3B), elongated internode and the following two internodes. which was 1 d before the increase of the cell number The further five to seven internodes, indicated by the along the radial axis at day 5 ( Fig. 3C). Tuber size position of the eyes, were concentrated in the uppermost increased quickly from day 5 to day 8 because of both part of the tuber ( Fig. 2D), showing that proximal inter- intensive cell widening ( Fig. 3B) and frequent longitudinal nodes contributed to in vivo tuber formation in a later cell divisions starting at day 5, reaching a peak at day 6 stage only. and ceasing around day 8 (Fig. 3C). The total cell number in pith and cortex diameters increased from 47 to 140, Development of in vitro tubers i.e. 3 times. After day 8, the number of cells along the Time-course of cell division and cell enlargement: Both cell tuber diameter remained constant ( Fig. 3C), but cells division and cell expansion are involved in stolon elongaresulting in steady tuber growth (Figs 1A, 3A). continued to enlarge gradually till about day 15 (Fig. 3B), tion and tuber growth. Cell numbers along the transversal axis of tubers and along the long axis of stolons were determined by cell counting. Cell enlargement contribut- Localization of DNA synthesis and cell division: The ing to the swelling of the tuber, was measured by cell distribution of nuclear DNA synthesis was investigated width. by the immunocytochemical detection of incorporated Buds elongated and formed stolons from day 0 to day BrdU. Cell division was visualized by DAPI and 4 of culture. During this period, the diameter of the Calcofluor White double staining. stolons remained constant, i.e. around 1.5 mm ( Fig. 3A), BrdU-labelled nuclei were found in the whole subapical and cell width in the subapical region did not change region of the stolon tip from day 1 to day 3 ( Fig. 1A). ( Fig. 3B). The number of cells along the stolon diameter They were small and close to each other. Sometimes pairs was similar in all regions of the stolon and longitudinal of labelled daughter nuclei were observed ( Fig. 4a, b). In cell divisions were not observed ( Fig. 3C). In contrast, DAPI stained sections, it was observed that many new cell numbers along the long axis of the buds increased 6 cells were arranged in meristem-like cell columns. All times from 40 to 260 between day 0 and day 5 (Fig. 3D). divisions were transversal and added new cells to the long Besides cell division, cell elongation also contributed to axis of the stolon ( Fig. 5a). stolon elongation (Sanz et al., 1996). At day 4, some stolons still had a few BrdU-labelled The development of the stolon suddenly switched to nuclei in the meristem-like cell columns. However, most tuber swelling at day 5 (Fig. 3A). The elongation of the stolons showed more labelled nuclei at the basal part of stolon stopped as soon as swelling started (data not the subapical region (Fig. 1B). From morphological

5 Potato tuber cell division and enlargement 577 observation, it was known that this region was the upper part of the first internode where tuber swelling started. At day 5, hardly any BrdU-labelled nuclei were detected in the meristem-like cells in the tip of the stolon (Figs 1C, 4c). They were only observed in the swelling tuber (Figs 1C, 4d). The position of new cell walls and the orientation of the chromosomes as visualised by Calcofluor White and DAPI respectively, revealed that the divisions were longitudinal and added new cells to the transversal axis of the tuber ( Fig. 5b). Such divisions only occurred in enlarged cells. From day 6 to day 7, BrdU-labelled nuclei and mitosis were only occasionally found, and after day 8, labelled nuclei were only found in the periderm region and around the vascular tissue (Figs 1D, 4e, f ). DAPI staining also showed mitosis in the periderm. The development of in vivo tubers Growth pattern: Before tuber formation, stolons had a normal stem structure with continuous vascular bundles along the long axis (Fig. 2A). At the beginning of tuber formation, the stolon tip swelled in the subapical region. The vascular bundles remained continuous along the long axis but became arc-shaped because of the swelling of the pith (Fig. 2B). In cross-sections, a circle of the xylem tissue was clearly observed with external phloem and internal phloem in the early stage (less than 0.8 cm in diameter) (Fig. 5c). After tubers had grown to about 0.8 cm in diameter, they became larger mainly because of increasing thickness of the perimedullary region including external phloem, xylem and internal phloem (Fig. 2C, D). The width of the cortex did not change during later stages of development (diameter >0.8 cm) (cf. Figs 2C, D, and 6C). With increasing thickness of the perimedullary zone, the vascular tissue became irregularly arranged and the xylem and phloem elements were scattered in the whole perimedullary region ( Fig. 5d). Quantification of cell division and cell enlargement in different tissues: Cortex, perimedullary region and pith are the three major tissues of a tuber. Cell numbers and cell widths were measured in these tissues in the successive stages of tuber formation. The total numbers of cortical cells and pith cells on cross-sections increased from 35 in the stolons to 92 (about 3 times) in the tubers of 0.8 cm in diameter and then remained constant ( Fig. 6A). Cell numbers in the perimedullary region, however, continuously increased during tuber development, but especially Fig. 3. Quantification of changes in tuber diameter, cell number and cell width during in vitro tuber formation in potato. (A) Tuber diameter. (B) Average cell width. (C) Cell numbers along the transversal axis of the stolon and swelling tuber. (D) Cell numbers along the long axis of the stolon. Data are means with standard deviations of five independent measurements.

6 578 Xu et al.

7 Potato tuber cell division and enlargement 579 after tubers had reached a diameter of 0.8 cm ( Fig. 6A). Clowes, 1972; Cutter, 1978; Peterson et al., 1985). In In all three tissues cell width steadily increased ( Fig. 6B). both in vitro and in vivo conditions, the first indication of Cortical cells were always the smallest and enlarged tuber formation is a thickening of the last well elongated slightly. The pith cells were larger than the perimedullary internode, which is often the eighth internode counted cells especially in the early stage when the perimedullary from the apical one as described by Cutter (1978). Three cells were not growing ( Fig. 6B). internodes (the sixth to the eighth counted from the apical Cell division and cell enlargement resulted in the one) form almost the whole in vitro tubers and more than enlargement of tubers. Distinguishing the three regions, 2/3 of the in vivo tubers. The expansion and the growth the thickness of the cortical region only increased slightly of the further internodes (the first to the fifth counted in the early stage; the thickness of the pith region gradually from the apical one) in the upper 1/3 of in vivo tubers are increased and the perimedullary region exhibited the caused by the subsequent growth of the perimedullary most obvious growth ( Fig. 6C). tissue in all directions. Both cell division and cell expansion are involved in Visualization of cell divisions in the perimedullary region: tuber development. In the literature there is a debate on In the early stage of in vivo tuber growth ( less than 0.8 cm the timing of these cellular events; some authors describe in diameter), the developmental progress is the same as that cell division precedes cell enlargement (Artschwager, that of in vitro tuberization. Then, immediately after the 1924; Plaisted, 1957; Reeve et al., 1969, 1973a; Duncan longitudinal cell division in pith and cortex had stopped, and Ewing, 1984), whereas others advocate the opposite groups of meristem-like cells formed in the perimedullary (Booth, 1963; Cutter, 1978). The data presented in this region (Fig. 5e). Unlike the elongated vascular cells, these paper clearly show that both the timing and the location cells were isodiametric as concluded from the observation of cell division and cell expansion are different in various of cross-sections and longitudinal sections. Each group regions of the developing stolon and tuber. Cell divisions consisted of various cell types in a concentric arrange- were first seen in the apical region of the stolon as also ment: small meristem-like cells generally with a few xylem observed by Duncan and Ewing (1984). Cells divided or phloem elements were in the centre; around them were transversally and then elongated. However, these processes several layers of the enlarging parenchyma cells which were only involved in stolon elongation. When lacked starch grains; the outermost layers consisted of stolon tips started to swell, transverse cell division in the mature parenchyma cells filled with starch grains apex had stopped. Meanwhile from the basal part of the ( Fig. 5e). Randomly oriented cell divisions were observed subapical region upward, cells enlarged and then the in the meristem-like cells, in the enlarging cells and also enlarging cells divided longitudinally. This finding is in in the mature cells with starch grains (Fig. 5f ). Besides, agreement with that of Sanz et al. (1996) who stated that cell divisions were also observed in the cortical cells which cell enlargement precedes cell division during the initiation divided along the tangential direction to allow further of radial tuber growth. expansion of the tuber ( Fig. 5g). However, cells in the It is generally agreed that the growth of the in vivo pith did not exhibit divisions (Fig. 5h). BrdU immunolab- tuber occurs initially in the pith and cortex, and then elling was also applied to investigate cell divisions in the predominantly in the perimedullary zone ( Hayward, 1938; perimedullary zone, but no clear labelling was obtained Booth, 1963; Reeve et al., 1969; Cutter, 1978; Peterson because of poor uptake of BrdU in the larger tubers. et al., 1985). The observations here support this view. However, it was noticed that the growth in pith and Discussion cortex and in the perimedullary zone occurred in different stages and with different patterns. In the early stage of in Potato tuber formation consists of two different morphogenetic vivo tuber formation ( less than 0.8 cm in diameter), cells steps: stolon elongation and tuber initiation in the pith and cortex first enlarged and then divided ( Booth, 1963; Vreugdenhil and Struik, 1989). Contrary longitudinally. Longitudinal cell divisions stopped when to Artschwager s (1924) and Goodwin s (1967) opinion, tubers reached the size of 0.8 cm in diameter, whereas it was observed that the growth of tubers depends on the cell enlargement in the pith and cortex continued through- expansion of internodes already present in the stolon, out the whole process of tuber growth. Although longitudinal and not on the formation of new ones (Leshem and cell divisions in the pith and cortex happened in a Fig. 4. Fluorescence micrographs of longitudinal sections of in vitro stolons and tubers of potato. Photograph (a) and the left hand sides of photographs (c f ) show the distribution of FITC labelled nuclei indicating BrdU incorporation during S-phase. Photograph (b) and the right hand sides of the photographs (c f ) show the PI stained nuclei in the same sections. Sections were imaged by CLSM in dual channel mode. The long arrows in the right upper corner indicate the long axis of stolons or tubers. For topography see Fig. 1. Bar=100 mm. (a, b) Subapical region of the stolon at day 3. Short arrow points to a pair of BrdU-labelled daughter cells which passed transversal division. (c) Stolon tip at day 5. Note that only a few nuclei are labelled. (d) Swelling tuber at day 5. Note increased DNA synthesis in tuber tissue. (e) Periderm at day 8 with DNA synthesis in the cork cambium region. (f ) DNA synthesis in the region of the vascular tissue at day 8.

8 580 Xu et al.

9 Fig. 6. Graphs showing cell numbers, cell widths and the thickness of different regions along the transversal axis of in vivo potato tubers. (A) Cell numbers. (B) Average cell width. (C) Thickness of different regions. Data are means with standard deviations of three independent measurements Potato tuber cell division and enlargement 581 experiments show that cell numbers increased about 3 times in the pith and cortex of young tubers (less than 0.8 cm in diameter). Unlike the description by Reeve et al. ( 1969), the initial cell divisions in the pith and cortex are always parallel to the long axis of the stolon and contribute to the swelling of the stolon tip mainly in a transverse direction. Although some cell divisions were incidentally observed in the perimedullary region in the early stage, the actual growth of perimedullary region, which included a lot of cell divisions and cell enlargement, started when the tuber reached a diameter of 0.8 cm and continued until the tuber had reached its final size (3 cm in diameter in these experiments). The planes of cell divisions in the perimedullary region are randomly oriented in both cross- and longitudinal sections, resulting in the enlargement of tubers in all directions. Thus the perimedullary region formed the major portion of the mature tuber. The morphology and the processes of cell division and cell enlargement of in vitro tubers are similar to those observed during the early stage of in vivo tuber formation (day 10 in vitro is equal to 0.8 cm in diameter in vivo). The youngest tubers have a diameter of 0.3 cm (=day 5 in vitro). But in vitro tubers stop growing at a size of 0.8 cm in diameter. Comparing the in vitro and in vivo tuber formation, it was concluded that the much larger final size of in vivo tubers is caused by the further cell division and cell enlargement in the perimedullary region in the later stage, which is lacking in the in vitro tubers. Differences between the growing conditions of in vivo and in vitro tubers, are the presence or absence of the intact plant and the supply of nutrients. This suggests that the mother plant might provide some products to the growing tuber in vivo, resulting in the initiation and continuation of cell divisions in the perimedullary region which is lacking in vitro. It is unlikely that the absence of cell divisions in the perimedullary region is due to the depletion of the medium, since transfer experiments of in vitro tubers to fresh medium did not lead to larger tubers (data not shown). However, it is highly possible that a factor, essential for the initiation of cell division in the perimedullary region is missing in the medium. Analysis of tubers grown on single-node cuttings, with the leaf still attached, revealed that these tubers did not show cell division in the perimedullary tissue either ( Xu and Ewing, unpublished data). Therefore, it is suggested that a regu- lating factor is derived from either the root or the shoot very short period, the results of this study did not agree that cell divisions in the pith and cortex are negligible as mentioned by Reeve et al. ( 1973b) and Peterson and Barker (1979). The in vitro and in vivo data from these Fig. 5. Micrographs showing cell division during in vitro (a, b) and in vivo (c h) tuber formation in potato. Bar=100 mm. (a, b) DAPI and Calcofluor White stained longitudinal sections of in vitro stolon and tuber. The long arrows on the right upper corner indicate the long axis of stolon and tuber. (c h) Toluidine blue stained cross-sections of in vivo tubers. (a) Transversal cell divisions in the subapical region at day 3. The short arrow indicates the mitosis in one cell. (b) Longitudinal cell divisions in the swelling tuber at day 5. (c) Vascular tissue in in vivo tuber of diameter=0.3 cm: ep, external phloem; x, xylem; ip, internal phloem. (d) Irregular arrangement of vascular tissue (see the short arrow) due to the growth of the perimedullary region. (e) Groups of concentrically arranged cells in the perimedullary region. (f ) Random cell divisions (see the short arrows) in a group of perimedullary cells. (g) Tangential cell divisions (see the short arrows) in cortical cells. pe, periderm. (h) Absence of cell division in the pith of larger tubers (diameter >0.8 cm).

10 582 Xu et al. of the intact plant, a factor which is missing in the in after DNA synthesis and are completed just before mitosis. vitro cultures. Further studies will be needed to test this The Plant Cell 2, Hayward HE The structure of economic plants. New hypothesis. York: Wiley. Helder H, van der Maarl A, Vreugdenhil D, Struik PC Stolon characteristics and tuber initiation in a wild potato species (Solanum demissum Lindl.). Potato Research 36, Acknowledgements Hendriks T, Vreugdenhil D, Stiekema WJ Patatin and The authors are indepted to Henk Kieft, Elly Koot-Gronsveld four serine proteinase inhibitor genes are differentially and Wilma Pons-Drexhage for technical advice and support, to expressed during potato tuber development. Plant Molecular the Royal Dutch Academy of Sciences ( KNAW ) for supporting Biology 17, the co-operation between China and The Netherlands, and to Leshem B, Clowes FAL Rates of mitosis in shoot apices the Wageningen Agricultural University for financial support. of potatoes at the beginning and end of dormancy. Annals of This work was partly funded by the European Union s Botany 36, BIOTECH programme, as part of the Project of Technical Murashige T, Skoog F A revised medium for rapid Priority. growth and bio-assays with tobacco tissue cultures. Physiologia Plantarum 15, Peterson R L, Barker WG Early tuber development from explanted stolon nodes of Solanum tuberosum var. Kennebec. Botanical Gazette 140(4), References Peterson CA, Peterson RL, Barker WG Observations on the structure and osmotic potentials of parenchyma associated Artschwager EF Anatomy of potato plant, with special with the internal phloem of potato tubers. American Potato reference to the ontogeny of the vascular system. Journal of Journal 58, Agricultural Research 27, Peterson RL, Barker WG, Howarth MJ Development Artschwager EF Studies on the potato tuber. Journal of and structure of tubers. In: Li PH, ed. Potato physiology. Agricultural Research 27, London: Academic Press, Booth A The growth substances in the development of Plaisted PH Growth of the potato tuber. Plant Physiology stolons. In: Ivins JD, Mithorpe FL, eds. The growth of the 32, potato. London: Butterworth, Reed T On the anatomy of some tubers. Annals of Bradbury D Division of starch-containing cells. American Botany 24, Journal of Botany 40, Reeve RM, Hautala E, Weaver ML Anatomy and Cutter EG Structure and development of the potato compositional variations within potatoes. I. Developmental plant. In: Harris PM, ed. The potato crop. The scientific basis histology of the tuber. American Potato Journal 46, for improvement. London: Chapman & Hall, Reeve RM, Timm H, Weaver ML. 1973a. Parenchyma cell De Vries H Beitrage zur speziellen Physiologie landwirts- growth in potato tubers. I. Different tuber regions. American chaftlicher Kulturpflanzen. V. Wachstumsgeschichte der Potato Journal 50, Kartoffelpflanze. Landwirtschaftliches Jahrbuch der Schweiz Reeve RM, Timm H, Weaver ML. 1973b. Parenchyma cell 1, growth in potato tubers, II. Cell divisions vs. cell enlargement. Duncan DA, Ewing EE Initial anatomical changes American Potato Journal 50, associated with tuber formation on single-node potato Sanz MJ, Mingo-Castel A, van Lammeren AAM, Vreugdenhil (Solanum tuberosum L.) cuttings. Annals of Botany 53, D Changes in the microtubular cytoskeleton precede Goodwin PB The control of branch growth on potato in vitro tuber formation in potato. Protoplasma 191, tubers. Journal of Experimental Botany 18, Vreugdenhil D, Struik PC An integrated view of the Gunning BES, Sammut M Rearrangement of microtubules hormonal regulation of tuber formation in potato (Solanum involved in establishing cell division planes start immediately tuberosum). Physiologia Plantarum 75,