Endopolyploidy in Vanda Miss Joaquim (Orchidaceae)


Author for correspondence: Chiang Shiong Loh Tel: +65 68742916 Fax: +65 67795671 Email: dbslohcs@nus.edu.sg


  • • Occurrence of endopolyploidy in somatic tissues of the hybrid orchid Vanda Miss Joaquim (Vanda hookeriana × Vanda teres) was investigated with respect to tissue type and developmental stage. Effects of naphthaleneacetic acid (NAA) and gibberellic acid (GA3) on endopolyploidy during embryo development were also studied.
  • • For the study of endopolyploidy, flow cytometric analysis was employed to determine nuclear DNA content of cells of somatic tissues.
  • • Multiploid cells were observed in leaves, roots and column, but not in shoot apex, stem, perianth and pedicel. Furthermore, differential distribution of multiploid cells was found among different parts of leaves and roots. The degree of endopolyploidy in embryos increased with development. NAA was shown to induce endoreduplication in germinating embryos to a much larger extent than GA3.
  • • The pattern of endopolyploidy was characteristic of tissue type and developmental stage. The implications of endopolyploidy during differentiation and development, as well as the relevance of endopolyploidy to somaclonal variation, are discussed.


Apart from germ line polyploidy in gametic cells, plants often exhibit endopolyploidy – genome variation in which multiple ploidy levels are found in somatic cells (Nagl, 1978; De Rocher et al., 1990; Joubès & Chevalier, 2000). Endoreduplication, the process of endonuclear chromosome duplication without intervening segregation and cytokinesis, is considered the most common mode of cell polyploidization in plants, being present in > 90% of angiosperms (Joubès & Chevalier, 2000). Endopolyploidy has been reported in many plant tissue types, especially those with cells of large size and high specificity. For orchids in particular, it has been observed in leaves (Jones & Kuehnle, 1998), root tips (Jones & Kuehnle, 1998), perianth (Mishiba & Mii, 2000) and growing embryos (Alvarez, 1968; Nagl, 1972).

Systemic control of endopolyploidy has been shown in plant species such as ice plant Mesembryanthemum crystallinum (De Rocher et al., 1990), Arabidopsis thaliana (Galbraith et al., 1991), cucumber (Gilissen et al., 1993) and tomato (Smulders et al., 1994). Increasing evidence has revealed that the pattern of endopolyploidy is characteristic of tissue type and developmental stage, indicating that endoreduplication is spatially and temporally regulated. In many plant species, older tissues showed higher levels of ploidy than younger ones (De Rocher et al., 1990; Galbraith et al., 1991; Smulders et al., 1994). It is suggested that endoreduplication is involved in developmental events such as cell expansion and differentiation (Kondorosi et al., 2000; Larkins et al., 2001). Furthermore, there is growing evidence that cell size is linked to the degree of endoreduplication (Melaragno et al., 1993; Cebolla et al., 1999; Kondorosi et al., 2000).

Specific combination of hormone, nutrient and light, which are major regulators of cell cycle and growth, might possibly initiate the signals that transform the mitotic cycle to endoreduplication during cell differentiation (Kondorosi et al., 2000). However, reports on the effects of plant growth regulators on endopolyploidy are equivocal and vary with plant species and tissue type investigated. Auxin treatment induced endoreduplication in tobacco single cells (Valente et al., 1998), but not in cactus Mammilaria san-angelensis cultured regenerants (Palomino et al., 1999). Enhancing effect of gibberellins on endoreduplication in Pisum sativa was cultivar-dependent (Mohamed & Bopp, 1980; Callebaut et al., 1982). Further investigation on the effects of auxins and gibberellins on endopolyploidy might provide more understanding of the role of these plant growth regulators in such a phenomenon.

Orchids are widely propagated for commercial purposes via tissue culture, with a variety of plant parts as explants (Arditti, 1996). A better insight to the plant genetic profile could be useful in tissue culture, since cytological differences inherent in explants or induced in regenerants during culture, might be some possible factors contributing to somaclonal variation. The objectives of this present study were to determine the occurrence of multiploid cells in different tissues of orchid Vanda Miss Joaquim and investigate whether the pattern of endopolyploidy could be affected by development. Effects of NAA and GA3 on endopolyploidy in the plant subject during embryo development were also investigated.

Materials and Methods

Plant materials

Hybrid orchid Vanda Miss Joaquim (Vanda hookeriana ×Vanda teres) is widely found in tropical regions (Hew et al., 2002). The orchid plants used for experiments were grown in pots in the planthouse.

Plant organs were excised and used immediately for analysis, while seeds were germinated in vitro. Mature seed pods were washed with detergent, surface-sterilized in 20% (v/v) CloroxTM solution plus a drop of Tween 20 for 15 min, with constant agitation. The seed pods were then rinsed twice with autoclaved water, before being cut into two halves. Seeds were transferred to 60-mm wide plastic Petri dishes containing 12 ml Murashige and Skoog (MS) medium (Murashige & Skoog, 1962) supplemented with 2% (w/v) sucrose, 10% (v/v) coconut water and solidified with 0.8% (w/v) agar, at pH 5.3. The cultures were maintained at 25°C under 24 h illumination of intensity 33 µE m−2 s−1. For experiments involving the effects of plant growth regulators on endopolyploidy during embryo development, GA3 or NAA of varying concentrations was added to the culture medium described earlier.

For flow cytometric analysis of leaves, the tip, middle and base sections of each harvested leaf, all 1.5 cm in length, were sampled. 5-cm long first leaf, 9-cm long first leaf, second leaf (from plant with 9-cm long first leaf), ninth leaf and 16th leaf were tested. Leaves were numbered from the shoot apex downwards, with the youngest leaf as first leaf (Fig. 1). The shoot apex (section right below first leaf) and stem section from fourth internode, both 1-cm long, were also being analysed for endopolyploidy. In the case of roots, the tip, middle and base sections of 8-cm and 15-cm long aerial roots (with green tips) were chosen for sample preparation, whereas the tip and base sections of 5-cm long aerial roots (with green tips) and terrestrial roots were used. Root sections used were all 0.8 cm long for terrestrial roots and 1.2 cm long for aerial roots. As for floral tissues, the column (with pollenia removed), pedicel, labellum, lateral petals, as well as dorsal and lateral sepals, were analysed. Floral stages investigated were 1 d before anthesis, 4 and 24 d after anthesis. Anthesis was taken as the time at which the flower bud opened. Analyses involving embryo development were carried out using whole embryos or protocorms.

Figure 1.

Vanda Miss Joaquim shoot, showing position of first leaf (9-cm long). Shoot on left showing part of the first leaf being covered by leaf sheaths of leaves below. Bar, 5 cm. Close-up of shoot on right, but with second and third leaves removed to show the whole of first leaf. Bar, 2.4 cm.

Nuclei isolation for flow cytometric analyses

Plant parts were taken from two different plant sources for each sample, whereas embryos from seed pod of one plant source constituted one sample. All samples, excluding embryos, were washed and dried before nuclei isolation. All steps of nuclei isolation were carried out on ice. Each sample was fixed in 2–3 ml fixation buffer (10 mM Tris, 10 mM Na2EDTA, 100 mM NaCl and 0.1% (v/v) Triton X-100, pH 7.5) for 15 min for perianth parts and 30 min for other sample types. The sample was then washed once with 1 ml extraction buffer (MgSO4 buffer with 1 mg ml−1 dithiothreitol and 2.5 mg ml−1 Triton X-100 added) prepared in accordance to Arumuganathan & Earle (1991). Tissue samples were thinly sliced with a sterile blade in 500 µl extraction buffer. However, for embryos cultured for 0, 2 and 4 weeks, they were gently pounded with a clean glass rod for nuclei release instead of chopping. After incubation for 1 h, the nuclei suspension was filtered through a 40-µm nylon cell strainer and kept on ice until flow cytometric analysis was carried out.

Flow cytometric analyses

Nuclei staining was carried out on each sample of nuclei suspension by incubating with 25 µl of 5 mg ml−1 propidium iodide and 50 µl of 5 µg ml−1 DNase-free RNase for 30 min at room temperature. Flow cytometric analyses were then performed using a Coulter EPICS® Elite ESP flow cytometer (Coulter, Miami, Florida, USA) with laser at 488 nm excitation. For each sample, 10 000 nuclei were analysed and at least four replicates, obtained from different plants, were subjected to analyses, where C is the nuclear DNA content of haploid genome. The 2C peak of fluorescence produced by nuclei extracted from the tip of ninth leaf was used to estimate the standard peak position of 2C nuclei. Data analyses were carried out using software WINMDI (version 2.8). To determine whether significant differences were present between the nuclei frequencies, t-test (if two means were involved) or one-way anova (if more than two means were involved) was employed at P= 0.05 unless stated otherwise.


Pattern of endopolyploidy in different organs during development

Results for the distribution of nuclei of various ploidy levels in different parts of leaves at different developmental stages are shown in Table 1. For 5-cm long 1st leaf, the tip contained nuclei of 2C and 4C DNA content, at mean frequencies of about 93.24% and 6.76%, respectively. Only nuclei with 2C DNA content were detected in the middle part and base of such a leaf. As the 1st leaf grew in size, nuclei of higher C-levels started to appear and their frequencies increased rapidly. For 9-cm long 1st leaf, nuclei of DNA content up to 16C were observed in the tip, whereas 2C, 4C and 8C nuclei accounted for nuclei present in the middle part; only 2C and 4C nuclei were found in the base. Nuclei with DNA content up to 16C were observed in all parts of the second leaf. Frequencies of nuclei of higher ploidy levels, namely 8C and 16C, in the base of 1st and second leaves never exceeded those in other leaf parts. However, the difference in the distribution of multiploid cells became smaller among the tip, middle part and base of second leaf, as compared to 1st leaf. For 9-cm long 1st leaf, mean frequencies of 2C, 4C and 8C nuclei were about 46.36%, 31.54% and 19.58%, respectively, in the tip and about 90.97%, 9.03% and 0.00%, respectively, in the base; however, for its corresponding second leaf, they were about 21.35%, 25.39% and 33.65%, respectively, in the tip and about 25.52%, 30.90% and 27.10%, respectively, in the base. As for ninth leaf, nuclei up to 16C DNA content were consistently found in all leaf parts. Occasional presence of a minor nuclei population with 32C DNA content was observed in the tip and middle part of the ninth leaf, but not in the base. The mean frequency of 8C nuclei ranged from about 34.20% to 39.50% for the tip, middle part and base of ninth leaf, while no significant difference in mean frequencies of 2C, 4C and 16C nuclei was found among all three leaf parts of the ninth leaf. The mean frequencies of 2C, 4C and 8C nuclei in the tip and middle part of the ninth leaf were similar to those from the same parts of the second leaf. For 16th leaf, no significant difference in the proportions of nuclei with 2C, 4C, 8C and 16C DNA content were observed among the tip, middle part and base, and these mean nuclei frequencies were not significantly different from those of the ninth leaf. In addition, all three parts of 16th leaf contained 32C nuclei for all replicates tested. The mean frequency of 32C nuclei increased progressively from about 1.11% at the base to about 2.76% at the tip and was higher than that of all the three leaf parts of the ninth leaf.

Table 1.  Pattern of endopolyploidy in different organs of Vanda Miss Joaquim during development, based on flow cytometric analyses of 10 000 nuclei per replicate
Organ part # Proportion of nuclei present (%)
  1. # Mean ± sd. DBA, day(s) before anthesis; DAA, day(s) after anthesis.

1st leaf (5-cm long)Tip 93.24 ± 11.71 6.76 ± 11.71 0.0 0.0 0.0
Middle100.0 0.0 0.0 0.0 0.0
Base100.0 0.0 0.0 0.0 0.0
1st leaf (9-cm long)Tip 53.83 ± 30.8329.17 ± 15.4214.39 ± 13.85 2.61 ± 3.37 0.0
Middle 88.95 ± 19.0811.05 ± 12.79 0.0 0.0 0.0
Base100.0 0.0 0.0 0.0 0.0
2nd leafTip 21.35 ± 4.1425.39 ± 3.0333.65 ± 2.4710.75 ± 2.32 0.0
Middle 28.98 ± 10.8228.90 ± 8.1527.93 ± 12.52 5.23 ± 4.33 0.0
Base 25.52 ± 7.7630.90 ± 6.6527.10 ± 9.69 4.82 ± 4.03 0.0
9th leafTip 21.92 ± 2.3822.04 ± 1.3934.20 ± 1.1920.60 ± 2.81 1.24 ± 0.40
Middle 18.42 ± 2.4222.76 ± 3.8236.89 ± 2.3920.95 ± 3.23 0.98 ± 0.63
Base 19.27 ± 2.4322.54 ± 2.5039.50 ± 2.4218.69 ± 2.57 0.0
16th leafTip 21.89 ± 2.7720.46 ± 2.7531.90 ± 2.5123.00 ± 4.74 2.76 ± 1.00
Middle 20.82 ± 3.7121.68 ± 4.2332.30 ± 3.2722.88 ± 5.33 2.31 ± 0.92
Base 22.69 ± 2.7122.30 ± 3.5334.72 ± 4.3419.18 ± 5.69 1.11 ± 0.38
Shoot apex 100.0 0.0 0.0 0.0 0.0
Stem 100.0 0.0 0.0 0.0 0.0
Aerial root (5-cm long)Tip 55.20 ± 15.1225.23 ± 8.2814.16 ± 4.70 5.41 ± 2.97 0.0
Base100.0 0.0 0.0 0.0 0.0
Aerial root (8-cm long)Tip 62.32 ± 26.3323.98 ± 15.3511.13 ± 8.97 2.56 ± 3.90 0.0
Middle100.0 0.0 0.0 0.0 0.0
Base100.0 0.0 0.0 0.0 0.0
Aerial root (15-cm long)Tip 60.87 ± 11.2022.39 ± 7.3212.22 ± 4.23 4.53 ± 1.11 0.0
Middle100.0 0.0 0.0 0.0 0.0
Base100.0 0.0 0.0 0.0 0.0
Terrestrial root (5-cm long)Tip 10.13 ± 2.8815.53 ± 2.6542.87 ± 1.1418.55 ± 1.7712.92 ± 2.32
Base 18.50 ± 1.7643.32 ± 2.0519.65 ± 2.2413.77 ± 0.99 4.76 ± 1.00
Petals and sepals1 DBA100.0 0.0 0.0 0.0 0.0
4 DAA100.0 0.0 0.0 0.0 0.0
24 DAA100.0 0.0 0.0 0.0 0.0
Pedicel1 DBA100.0 0.0 0.0 0.0 0.0
4 DAA 79.41 ± 25.7920.59 ± 25.79 0.0 0.0 0.0
24 DAA100.0 0.0 0.0 0.0 0.0
Column1 DBA100.0 0.0 0.0 0.0 0.0
4 DAA 13.32 ± 4.9215.22 ± 4.5571.47 ± 10.12 (intermediate of 8C and 16C) 
24 DAA 45.16 ± 34.3110.08 ± 3.7244.76 ± 30.76 (intermediate of 8C and 16C) 

However, for young leaves, the frequencies of nuclei of each ploidy level were highly variable among replicates of leaf parts of the same kind, but the degree of variability, as reflected by the standard deviations, decreased with leaf development (Table 1).

Nuclei isolated from the shoot apex and the fourth internode contained solely 2C DNA content (Table 1). No endopolyploidy was detected in these two plant parts.

The tips of aerial roots were made up of a mixture of multiploid nuclei, ranging from 2C to 16C DNA content. This held true for 5-cm, 8-cm and 15-cm long aerial roots (Table 1). In the tip of aerial roots, mean nuclei frequencies of the respective ploidy levels did not change significantly with the increase in root length and the majority of the nuclei present were of 2C DNA content. Endopolyploidy was absent in the middle part and base of aerial roots and remained so, as the roots grew longer (Table 1). By contrast to the aerial roots, endopolyploidy was observed in the entire 5-cm long terrestrial roots, which contained multiploid nuclei with DNA content up to 32C. The proportion of nuclei of higher ploidy levels, namely 8C, 16C and 32C, was larger in the tip (about 74%) than in the base (about 38%) (Table 1). The 5-cm long aerial and terrestrial roots were too short for the allocation of middle part to be deemed necessary.

One day before anthesis, only nuclei with 2C DNA content were revealed in flow cytometric analyses for sepals, petals, column and pedicel (Table 1). A column is a modified organ derived from the fusion of anther, stigma and style (Hew & Yong, 1997). Four days after anthesis, nuclei isolated from the column produced three peaks of fluorescence in flow cytometric analyses. They corresponded to DNA content of 2C, 4C and intermediate of 8C and 16C, with mean frequencies of about 13.32%, 15.22% and 71.47%, respectively (Table 1). For the pedicel, nuclei isolated either had solely 2C DNA content for most replicates, or 2C and 4C DNA content for remaining replicates. Only 2C nuclei were present in petals and sepals (Table 1). Twenty-four days after anthesis, the proportion of 2C nuclei of the column increased by about 32% (Table 1). As for other floral tissues, all nuclei sampled maintained the 2C DNA content (Table 1).

Pattern of endopolyploidy during embryo development

2C nuclei accounted for all the nuclei present in embryos isolated from seed pods (Table 2; Fig. 2). Endopolyploidy occurred progressively during embryo development (Table 2; Fig. 2). After 2 weeks of culture, nuclei of four ploidy levels corresponding to 2C, 4C, 8C and 16C were present in the germinating embryos, at mean frequencies of 79.17%, 14.86%, 4.27% and 1.70%, respectively (Table 2). After 4 weeks of culture, 32C nuclei appeared, while the mean frequencies of 4C, 8C and 16C nuclei increased (Table 2). The embryos had developed into small protocorms by 6 weeks of culture and the proportion of 8C, 16C and 32C nuclei became higher at the same time (Table 2). After 8 weeks of culture, nuclei of 64C DNA content appeared in the protocorms (Table 2). The mean frequencies of 2C and 4C nuclei decreased considerably whereas those of 8C, 16C and 32C nuclei showed a further increase (Table 2). The distribution of multiploid cells obtained after 10 weeks of culture was not significantly different from that obtained 2 weeks before (Table 2).

Table 2.  Pattern of endopolyploidy in embryos of Vanda Miss Joaquim during development, based on flow cytometric analyses of 10 000 nuclei per replicate
Weeks of culture# Proportion of nuclei (%)
  1. # Mean ± sd.

0100.0 0.0 0.0 0.0 0.00.0
2 79.17 ± 16.2214.86 ± 11.79 4.27 ± 3.23 1.70 ± 1.27 0.00.0
4 46.48 ± 5.6734.91 ± 1.8511.72 ± 2.11 5.34 ± 1.16 1.61 ± 0.490.0
6 27.88 ± 2.9237.50 ± 1.0918.55 ± 2.5110.18 ± 0.62 5.89 ± 0.710.0
8 12.70 ± 2.97 6.23 ± 2.5531.72 ± 7.9326.85 ± 2.1713.08 ± 0.899.43 ± 0.65
10  7.53 ± 1.97 5.08 ± 0.7938.07 ± 10.0226.17 ± 3.9813.92 ± 2.119.22 ± 1.16
Figure 2.

Flow cytometric histograms obtained from Vanda Miss Joaquim embryos cultured on basal MS medium for varying time durations, with peaks of fluorescence representing nuclei of various ploidy levels or nuclear DNA content. Number of weeks indicated on each histogram represents the number of weeks of culture for the embryos.

Effect of NAA on the pattern of endopolyploidy during embryo development

After 2 weeks of culture, embryos from media containing 10−7 M and 10−5 M NAA were observed to have 32C nuclei and a significant increase in the proportion of 4C and 8C nuclei, as compared to those in NAA-free medium (Table 3). However, the mean frequencies of 4C, 8C and 32C nuclei were not significantly different between the two NAA concentrations. After 4 weeks of culture, in media with 10−7 M and 10−5 M NAA, a significant increase in the proportion 16C and 32C nuclei was noticed, with the increases induced by 10−5 M NAA being more substantial (Table 3). For embryos cultured with 10−5 M NAA, the proportion of 4C nuclei was significantly lower than that with no or 10−7 M NAA (Table 3). Meanwhile, mean frequencies of nuclei of 2C and 8C DNA content for both NAA concentrations remained similar to those for the control (Table 3).

Table 3.  Effect of naphthaleneacetic acid (NAA) on the pattern of endopolyploidy in embryos of Vanda Miss Joaquim during development, based on flow cytometric analyses of 10 000 nuclei per replicate
Weeks of cultureConcentration (M)# Proportion of nuclei (%)
  1. # Mean ± sd. *, Significantly different from control (0 m) of same week, in the mean frequency for same C-level at P = 0.05 by t-test; **, significantly different from control (0 m) of same week, in the mean frequency for same C-level at P = 0.01 by t-test.

0100.0 0.0 0.0 0.00.0
20 49.23 ± 8.9925.06 ± 1.8815.94 ± 3.28 9.77 ± 3.830.0
10−7 30.84 ± 7.31*20.02 ± 3.60*28.46 ± 5.93**14.20 ± 2.70*6.48 ± 3.53*
10−5 40.75 ± 1.6219.61 ± 2.63*24.59 ± 1.51**10.97 ± 0.854.08 ± 1.68*
40 24.81 ± 7.7136.91 ± 5.1926.73 ± 4.58 7.10 ± 2.594.44 ± 0.89
10−7 24.68 ± 6.7230.99 ± 5.0425.36 ± 7.0612.77 ± 2.64*6.19 ± 1.11*
10−5 25.88 ± 9.1617.83 ± 2.65**27.30 ± 6.1520.17 ± 3.90**8.83 ± 0.74**

Effect of GA3 on the pattern of endopolyploidy during embryo development

After 2 weeks of culture, the distribution of multiploid cells in embryos cultured with 10−7 M and 10−5 M GA3 was not significantly different from the control cultured in GA3-free medium (Table 4). All of them consisted of nuclei up to 16C DNA content, with 2C nuclei being the majority at about 50% (Table 4). After 4 weeks of culture, mean nuclei frequencies for embryos cultured with 10−7 M GA3 remained similar to those of the control (Table 4). However, for embryos cultured with 10−5 M GA3, the proportion of 2C nuclei was significantly lower and the proportion of 4C and 8C nuclei was significantly higher than the control (Table 4).

Table 4.  Effect of GA3 (gibberellic acid) on the pattern of endopolyploidy in embryos of Vanda Miss Joaquim during development, based on flow cytometric analyses of 10 000 nuclei per replicate
Weeks of cultureConcentration (M)# Proportion of nuclei (%)
  1. # Mean ± sd. *, Significantly different from control (0 m) of same week, in the mean frequency for same C-level at P = 0.05 by t-test.

0100.0 0.0
20 48.53 ± 7.3235.90 ± 3.6710.87 ± 2.434.70 ± 0.980.0
10−7 52.09 ± 1.4530.44 ± 1.7711.75 ± 2.305.62 ± 0.800.0
10−5 56.58 ± 8.9928.07 ± 5.3611.02 ± 3.004.33 ± 0.700.0
40 36.70 ± 13.5031.49 ± 2.2720.56 ± 8.597.70 ± 3.433.56 ± 0.90
10−7 39.51 ± 9.8332.80 ± 5.1918.41 ± 8.496.11 ± 2.383.17 ± 0.61
10−5 18.22 ± 6.51*37.68 ± 3.17*31.89 ± 3.16*9.02 ± 1.553.19 ± 0.47


Leaves of Vanda Miss Joaquim, mainly made up of large mesophyll cells, exhibited high endopolyploidy possibly to support the cell volume. De Rocher et al. (1990) had suggested that cells expected to reach high ploidy levels were large-volume cells such as water-storing mesophyll cells. Distribution of multiploid cells in Vanda Miss Joaquim leaves was coupled with development and often varied with leaf parts. Endopolyploidy seemed to be originally absent in the leaves and might have been initiated during the early stages of leaf development (Table 1). The endoreduplication process was apparently most rapid in growing young leaves, since nuclei of 9-cm long first leaf could possess DNA content up to 16C, as compared to the absence of endopolyploidy in 5-cm long first leaf (Table 1). Analyses on ninth and 16th leaves suggested that older leaf tissues exhibited higher ploidy levels than younger ones (Table 1), which were similar to the results from ice plant (De Rocher et al., 1990), tomato (Smulders et al., 1994) and Arabidopsis thaliana (Galbraith et al., 1991). Cell expansion could occur with age, thus a higher DNA content fulfilled by endoreduplication might be needed to sustain the larger cell volume. It supports the suggestion that endoreduplication is involved in cell development and differentiation (Kondorosi et al., 2000; Larkins et al., 2001). Besides, new rounds of endoreduplication might have started from leaf tip and progressed towards the base, since nuclei of higher C-levels in leaves always appeared first in the tip and last in the base during development (Table 1). It will be interesting to find out in future, the relative times leaf cells at the tip, middle and base switch from division to expansion. In algae Alaria esculenta sporangia, there is a series of four mitotic divisions in which there is only one round of DNA synthesis (Garbary & Clarke, 2002).

Observed absence of endopolyploidy in the shoot apex of Vanda Miss Joaquim (Table 1) further supported the proposal that endoreduplication had started during early leaf development, after the shoot stage. Endopolyploidy was absent in shoot tip of Brassica rapa and B. oleracea too and it was suggested that repression of endoreduplication at the apical meristem might be one mechanism to ensure the genetic stability of the germ line (Kudo & Kimura, 2001a). Also, cells at the apical meristem being small might not need a high nuclear DNA level to support their function.

In Vanda Miss Joaquim, only the tip of the aerial root exhibited endopolyploidy and its degree did not change significantly with root growth (Table 1). The pattern at the tip might not be due to the presence of large cells, as other root parts also have a substantial number of large cells. A more intricate differentiation control might be involved, since the aerial root tip is morphologically different from the other parts; for example, it is photosynthetic and not covered by veleman (Hew & Yong, 1997). By contrast, extensive endopolyploidy occurred in the whole of terrestrial roots (Table 1). In tomato roots, the proportion of 8C nuclei present increased upon arbuscular mycorrhizal colonization (Berta et al., 2000). Mycorrhizal infection is often associated with nuclear hypertrophy (Goh et al., 1992; Berta et al., 2000), which depends on chromatin decondensation and possibly endoreduplication (Berta et al., 2000). Terrestrial roots used for the present study were only mildly infected, thus fungal infection could not account for much of the extensive endopolyploidy. Besides, there is no direct evidence linking mycorrhizal infection and endoreduplication at present. Endoreduplication has been suggested to be involved in cell differentiation (Kondorosi et al., 2000; Larkins et al., 2001). Aerial and terrestrial roots of Vanda Miss Joaquim have distinctly different cell composition, morphology and growth characteristics, and they experience diverse environmental conditions. Whether these factors contribute to their different patterns of endopolyploidy remains to be elucidated.

Our findings indicated that endoreduplication was not a common feature in the cell growth and development of Vanda Miss Joaquim flowers and did not seem to couple with tissue differentiation in the flowers, with the column as an exception (Table 1). Endopolyploidy was also not detected in floral organs of Arabidopsis thaliana and it was suggested that the exclusion of multiploidy from somatic cells of floral structures might avoid the potential production of polyploid gametophytes (Galbraith et al., 1991). However, endoreduplication in perianth was reported in cabbage (Kudo & Kimura, 2001b) and portulaca (Mishiba & Mii, 2000). The column displayed an atypical pattern of endopolyploidy after anthesis with a substantial nuclei population with DNA content intermediate of 8C and 16C (Table 1), possibly contributed by its structural modifications. Confocal microscopic analysis of the column showed that it contained many cell types of diverse sizes and the nuclear DNA content was not highly correlated to cell size (W. L. Lim , unpublished data).

Endoreduplication occurred during embryo development (Table 2). Cytological studies on embryos of another orchid, Vanda sanderiana, revealed that while nuclei in the meristematic cells remained 2C throughout development, endopolyploidy was exhibited in the parenchymal cells (Alvarez, 1968). Endoreduplication in parenchymal cells of growing embryos had also been implicated in Cymbedium (Nagl, 1972) and Spathoglottis plicata (Raghaven & Goh, 1994). Confocal microscopic analysis of Vanda Miss Joaquim protocorm showed that parenchymal cells had very much larger cell and nuclear areas than meristematic cells, indicating that endopolyploidy might be a very prominent feature in the former (W. L. Lim, unpublished data). Orchid seeds lack endosperm and cells present were heavily packed with food reserves (Arditti, 1992). The parenchymatous region of orchid embryo appears to be analogous to the endosperm of other angiosperms, in that both tissues function in the nutrition of the embryo (Alvarez, 1968). During maize endosperm development, endoreduplication occurred to drive massive synthesis of storage proteins and starch (Lur & Setter, 1993). The degree of endopolyploidy increased with embryo development of Vanda Miss Joaquim (Table 2), possibly therefore to support the nutrient requirements of growing protocorms.

Exogenous NAA might increase the endopolyploidy levels in embryos of Vanda Miss Joaquim (Table 3), though it was observed that the rate of embryo growth decreased with increasing NAA level. Auxin was reported to affect endosperm development in maize, which involved endoreduplication (Lur & Setter, 1993). At present, auxin has induced endoreduplication in culture for most of the studies at varying extents (Valente et al., 1998; Gendreau et al., 1999; Mishiba et al., 2001), though in Mammillaria cactus, no such enhancing effect of auxin was observed (Palomino et al., 1999). This suggested the possible involvement of auxins in the regulation of endoreduplication. Evidence indicated that inhibition of the activity of p34cdc2, a cell cycle regulator, might result in the switch of mitotic cycles to endoreduplication and this inhibition might involve auxin (John et al., 1993; Valente et al., 1998).

Only 10−5 M GA3 caused a small increase in the amount of endoreduplication for the embryos after 4 weeks of culture (Table 4). It was also observed that only 10−5 M GA3 led to a mild induction in embryo growth. This indicated that GA3 might induce endoreduplication and growth in Vanda Miss Joaquim embryos by a small extent, with its effect being dependent on concentration. GA treatment increased the endopolyploidy levels in GA-deficient mutants of Arabidopsis (Gendreau et al., 1999). In addition, the enhancing effect of GA on endopolyploidy appeared to be cultivar-dependent in Pisum sativa (Mohamed & Bopp, 1980; Callebaut et al., 1982) and Triticum durum (Cavallini et al., 1995).

There is growing evidence that, in context of a given genome, the size and morphology of cells are linked to their DNA content (Kondorosi et al., 2000). Cells that have undergone endoreduplication are larger than comparable cells that have not (Larkins et al., 2001). Studies on Triticum durum and Arabidopsis thaliana epidermal cells also showed the correlation of cell size and nuclear DNA content (Cionini et al., 1983; Melaragno et al., 1993). Recent work had provided molecular evidence for cell size regulation by endopolyploidy. One gene involved in the transformation of mitotic cycles to endoreduplication, ccs52, was identified in alfalfa. Reduction in ccs52 transcript level correlated with a decrease in endopolyploidy and cell size in Medicago roots, cotyledons and hypocotyls (Cebolla et al., 1999). Positive correlation between cell volume and degree of endopolyploidy in most cell types indicates that nuclear DNA content might play a key role in regulating cell volume. This could also apply to Vanda Miss Joaquim.

In conclusion, systemic control of endopolyploidy was displayed where multiploid cells were found in many tissues of Vanda Miss Joaquim with varying patterns. Our results showed that embryos of Vanda Miss Joaquim were originally diploid and that extensive endoreduplication occurred during embryo development (Table 2; Fig. 2). Such a high degree of endopolyploidy was replaced by the differential distribution of multiploid cells in organs of the mature plant (Table 1). The pattern of endopolyploidy was found to change with development for embryos, leaves and flowers (Tables 1 and 2). These findings are in line with the general observation that the distribution of endopolyploid cells is characteristic with tissue type and developmental stage. Organ-specificity in the pattern of endopolyploidy in Vanda Miss Joaquim might possibly reflect its integral involvement in cell expansion and differentiation. Endoreduplication was suggested to cope with the gene expression required for cell expansion and specialization, rapid cell growth, high metabolic activity, coordination with organelle genomes, or adaptation to environmental changes (Galbraith et al., 1991; Joubès & Chevalier, 2000; Larkins et al., 2001).

Somaclonal variation can arise from genetic changes such as gene mutation (Dennis et al., 1987), increase in chromosome number (Lewis-Smith et al., 1990) and DNA methylation (Jaligot et al., 2000). Apart from elucidating the gene expression during differentiation and development, studies on endopolyploidy could also give understanding to the nature of tissues used for clonal propagation, thus providing some possible explanations to the occurrence of somaclonal variation. Plant regeneration is based on the concept of cell totipotency. An explant exhibiting endopolyploidy would contain a mixture of cells of varying ploidy levels, therefore cell heterogeneity in DNA content within the explant might contribute to variation in gene expression and phenotype of somaclones. Vanda Miss Joaquim could be micropropagated using shoot tip, stem, axillary buds, leaf and root tip explants (Arditti, 1996) and the last two have been shown to exhibit a high degree of endopolyploidy (Table 1). If genetically uniform plants were to be micropropagated, explants could be obtained from organs in which endopolyploidy is absent, such as shoot tip for Vanda Miss Joaquim, so to minimize variation. Apart from the endopolyploid nature of original explant, the presence of plant growth regulators in culture might induce or inhibit endoreduplication and alter gene expression of cultured plants, as seen in Vanda Miss Joaquim (Tables 3 and 4). Hence, knowledge on the effect of plant growth regulators on endoreduplication would be helpful in the maintenance of ploidy levels of cultured plants.