Endoreduplication intensity as a marker of seed developmental stage in the Fabaceae


  • Monika Rewers,

    1. Laboratory of Molecular Biology and Cytometry, Department of Plant Genetics and Biotechnology, University of Technology and Life Sciences in Bydgoszcz, 85-789 Bydgoszcz, Poland
    Search for more papers by this author
  • Elwira Sliwinska

    Corresponding author
    1. Laboratory of Molecular Biology and Cytometry, Department of Plant Genetics and Biotechnology, University of Technology and Life Sciences in Bydgoszcz, 85-789 Bydgoszcz, Poland
    • Department of Plant Genetics and Biotechnology, University of Technology and Life Sciences in Bydgoszcz, Kaliskiego Ave. 7, 85-789 Bydgoszcz, Poland
    Search for more papers by this author


Flow cytometry (FCM) can be used to study cell cycle activity in developing, mature and germinating seeds. It provides information about a seed's physiological state and therefore can be used by seed growers for assessing optimal harvest times and presowing treatments. Because an augmented proportion of 4C nuclei usually is indicative of high mitotic activity, the 4C/2C ratio is commonly used to follow the progress of seed development and germination. However, its usefulness for polysomatic (i.e., containing cells with different DNA content) seeds is questioned. Changes in cell cycle/endoreduplication activity in developing seeds of five members of the Fabaceae were studied to determine a more suitable marker of seed developmental stages for polysomatic species based on FCM measurements. Seeds of Phaseolus vulgaris, Medicago sativa, Pisum sativum, Vicia sativa, and Vicia faba var. minor were collected 20, 30, 40, 50, and 60 days after flowering (DAF), embryos were isolated and the proportion of nuclei with different DNA contents in the embryo axis and cotyledon was established. The ratios 4C/2C and (Σ>2C)/2C were calculated. Dried seeds were subjected to laboratory germination tests following international seed testing association (ISTA) rules. Additionally, the absolute nuclear DNA content was estimated in the leaves of the studied species. During seed development nuclei with DNA contents from 2C to 128C were detected; the endopolyploidy pattern depended on the species, seed organ and developmental stage. The cell cycle/endoreduplication parameters correlated negatively with genome size. The (Σ>2C)/2C ratio in the cotyledons reflected the seed developmental stage and corresponded with seed germinability. Therefore, this ratio is recommended as a marker in polysomatic seed research and production instead of the 4C/2C ratio, which does not consider the occurrence of endopolyploid cells. © 2012 International Society for Advancement of Cytometry

In seeds, cell cycle activity corresponds with their physiological state and therefore can be used to follow seed development, maturation, and germination, as well as to control presowing priming treatments (1). Cell cycle progression is marked by changes in nuclear DNA content. At the G1 phase nuclei possess a 2C DNA content; some cells exit the cell cycle at this phase and enter a quiescent stage called G0, where they remain metabolically active but no longer proliferate (2). During the S phase doubling of DNA occurs, from 2C to 4C. When DNA synthesis is completed there is a second interval, called G2, during which the cell grows as synthesis of proteins that are necessary for mitosis takes place. During mitosis (the M phase) DNA is distributed to two daughter cells which consequently, after cellular division, possess 2C DNA. Because flow cytometry (FCM) is a fast and accurate method for estimation of DNA content in different plant tissues, it can be applied to the detection of nuclei at particular stages of the cell cycle in intact seeds and/or in different seed parts (e.g., embryo, endosperm; 1). Analysis of cell cycle activity by FCM can be an alternative to the determination of seed developmental stage by histological observations of embryo structure and cell division patterns in the embryo and the endosperm.

One of the most important issues in seed production is to establish an optimal harvest time. At that time the majority of a seed population should have attained physiological maturity, i.e., reached maximum germinability, viability, and vigor (2). Usually, seeds require some additional period on the mother plant (maturation drying) because at physiological maturity their moisture content is too high for safe handling and storage. Traditionally, harvest maturity is established visually (imprecisely) or by using germination tests. However, such tests usually take about 7–14 days; therefore, developmental markers that give faster information on the physiological state of a seed are desirable. FCM seems to be a suitable method to provide such a marker.

At the beginning of seed development cell cycle activity in the embryo increases, while it declines at the end of embryogenesis and becomes gradually arrested at maturation (3). While in developing embryos the proportion of 4C nuclei (being at the G2 phase of the cell cycle) is relatively high, mature embryos of desiccation-tolerant (orthodox) seeds possess predominantly nuclei with a 2C DNA content, indicating cell cycle arrest at the G0/G1 phase; in some species a small proportion of nuclei with a 4C DNA content is also present (1, 4–7). During germination or pre-sowing priming of seeds the proportion of the 4C nuclei increases again, marking the entrance of the cells into the G2 phase that precedes mitotic divisions, which typically occurs during seedling growth. Therefore, the 4C/2C ratio has been proposed as a marker for seed maturation and the advancement of seed germination (8, 9). However, there are seeds that contain 2C and 4C nuclei and also ones with a higher DNA content, which is the result of endoreduplication, a process during which the nuclei undergo repeated rounds of DNA replication without mitosis (endocycles; 10–14). Such seeds possess polysomatic organs, i.e., composed of somatic cells with different DNA contents, including endopolyploid ones. Here, it is hypothesized that for polysomatic seeds the 4C/2C ratio is not fully suitable as a marker of seed maturation, since it does not consider the presence of cells with a DNA content higher than 4C. To verify this hypothesis and to suggest another FCM marker, which reflects better the physiological state of the seed, five polysomatic species belonging to the Fabaceae family were analyzed using flow cytometry. The changes in DNA replication patterns in the embryo from 20 days after flowering (DAF) to harvest maturity, as well as after 4 months of storage, were followed. There are only a few reports on endoreduplication in the embryo during seed development and the significance of this process is still not clear (3, 14–18). The present study provides information on a relationship between endoreduplication intensity in different seed parts and genome size, and type of seedling establishment (epigeal and hypogeal). This is the first report containing FCM comparative data on DNA synthesis patterns in different polysomatic species of one family during seed development.

Materials and Methods

Plant Material

Plants of common bean (Phaseolus vulgaris L. ‘Tosca’), alfalfa (Medicago sativa L. ‘Alpha’), field pea (Pisum sativum L. ‘Brylant’), common vetch (Vicia sativa L. ‘Ina’), and horse bean (Vicia faba L. var. minor Harz. ‘Nadwiślański’), differing in seedling emergence type, life cycle, and genome size (Table 1) were grown in experimental plots of the Laboratory of Molecular Biology and Cytometry, University of Technology and Life Sciences in Bydgoszcz in spring and summer 2011. Developing pods were harvested at various times: 20, 30, 40, 50, and 60 days after flowering (DAF) (the last time point only for P. vulgaris and V. faba var. minor which are characterized by a longer seed development period). At their last harvest date the seeds were considered to be harvest mature. Seeds were extracted from the pods for laboratory germination tests and flow cytometry. Additionally, seeds harvested at the last date were stored for 4 months (until January 2012) at room temperature (S) to confirm that they retained their characteristics during storage, which is evidence for harvest maturity during collection.

Table 1. Chosen characteristics of five Fabaceae species
SpeciesSeedling emergence typeLife cycleGenome size (pg/2C; ± SD)Genome size categorya
  • a

    According to Soltis et al. (19).

Phaseolus vulgarisepigealannual1.28 ± 0.01very small
Medicago sativaepigealperennial3.56 ± 0.02small
Pisum sativumhypogealannual8.89 ± 0.04intermediate
Vicia sativahypogealannual3.59 ± 0.05small
Vicia faba var. minorhypogealannual26.87 ± 0.12intermediate

Laboratory Germination Tests

Seeds at different developmental stages were dried at room temperature and then exposed to laboratory germination tests, according to International Seed Testing Association (ISTA; 20). Four replicates of 50 seeds were placed in plastic boxes on two layers of filter paper moistened with distilled water to 65% humidity and incubated at 20°C. Seeds with radicle protrusion of at least 1 mm were regarded as having germinated. Germinated seeds were counted twice: (i) after 4–5 days of imbibition to establish germination energy and (ii) after 8–14 days of imbibition to establish germination capability (the day of counting was species-specific, following ISTA rules). The results were estimated using a one-way analysis of variance and a Duncan's test (P = 0.05). Data were transformed to arcsin square-root before statistical analysis, although the actual percentages are presented in Table 2.

Table 2. Germination energy and capacity of dried seeds of five Fabaceae species harvested on different days after flowering (DAF) and after 4 months of storage (S)
SpeciesTime of analysisGermination energy (%)Germination capacity (%)
  • Values are means ± SD.

  • a

    Values for the species and parameters (in columns) followed by different letters are significantly different at P = 0.05 (Duncan's test).

  • b

    Seeds at physiological maturity.

P. vulgaris20 DAF0 da0.63 ± 1.25c
30 DAF28.1 ± 2.39c33.1 ± 4.27b
40 DAFb96.8 ± 2.83b99.5 ± 1.00a
50 DAF99.5 ± 1.00a99.5 ± 1.00a
60 DAF100a100a
M. sativa20 DAF0c0d
30 DAF10.5 ± 1.92c16.0 ± 1.63c
40 DAF40.0 ± 7.48b54.5 ± 3.42b
50 DAFb85.5 ± 3.42a93.0 ± 5.77a
S84.0 ± 4.90a93.5 ± 4.73a
P. sativum20 DAF15.5 ± 4.44c35.5 ± 6.81c
30 DAF64.5 ± 5.74b89.5 ± 1.91b
40 DAFb81.0 ± 2.58a98.5 ± 1.91a
50 DAF80.5 ± 2.52a99.5 ± 1.00a
S80.5 ± 5.26a100a
V. sativa20 DAF67.0 ± 2.58d70.5 ± 5.26c
30 DAF91.0 ± 2.58c95.0 ± 2.58b
40 DAFb99.5 ± 1.00a99.5 ± 1.00a
50 DAF96.0 ± 1.63bc98.0 ± 2.31ab
S97.0 ± 2.58ab98.0 ± 1.63ab
V. faba var. minor20 DAF0d0d
30 DAF8.0 ± 5.16d47.0 ± 7.39c
40 DAF24.0 ± 3.27c93.0 ± 2.58b
50 DAFb58.0 ± 2.83b99.5 ± 1.00a
60 DAF66.0 ± 2.83a100a
S70.5 ± 5.00a99.5 ± 1.00a

Flow Cytometric Analysis of Endopolyploidy and Nuclear DNA Content

For analysis of endopolyploidy, directly after extraction a seed from the pod the embryo was dissected into embryo axis and cotyledons and samples of each part were prepared as previously described (14), using 4′,6-amidino-2-phenylindole (DAPI; 2 μg ml−1) for DNA staining. Analyses were performed on ten biological replicates using a Partec CCA (Partec GmbH, Münster, Germany) flow cytometer, equipped with an HBO lamp, KG1 heat protection filter, BG12 and UG1 short-pass filters, GG435 long-pass filter, and a dichroic mirror TK420, using a logarithmic amplification, with no gating. For each sample, fluorescence of 5,000–8,000 nuclei was analyzed. Histograms were collected as DYN files and evaluated manually using the DPAC v. 2.2 program (Partec GmbH, Münster, Germany). The proportion of nuclei with different DNA contents, the number of endocycles, the 4C/2C and (Σ>2C)/2C ratios were established. The 4C/2C ratio is the ratio between the number of nuclei with 4C DNA content to the number of nuclei with 2C DNA content. Similarly, the (Σ>2C)/2C ratio is the ratio between the number of nuclei with a DNA content higher than 2C (4C+8C+16C+32C+64C+128C) and the number of nuclei containing 2C DNA. The results were statistically analyzed using a one-way analysis of variance and a Duncan's test (P = 0.05).

In this work, nuclei having at least 8C DNA content were considered to be endopolyploid, since it is not possible to distinguish by FCM the 4C nuclei in cells that have just entered endoreduplication (i.e., being in the G1 phase of the first endocycle) from those within cells in the G2 phase of the mitotic cycle.

For genome size estimation, young-leaf samples were prepared as previously described (21), using propidium iodide (PI; 50 μg ml−1) for DNA staining. The following internal standards were used: Petunia hybrida ‘P × Pc6’ (2C = 2.85 pg; 22) for P. vulgaris; Zea mays ‘CE-777’ (2C = 5.43 pg; 23 for M. sativa, P. sativum, and V. sativa; and Allium cepa ‘Alice’ (2C = 34.89 pg; 24) for V. faba var. minor. For each sample, fluorescence was measured in at least 7,000 nuclei, using a CyFlow SL Green (Partec GmbH, Münster, Germany) flow cytometer, equipped with a high-grade solid-state laser with green light emission at 532 nm, long-pass filter RG 590 E, DM 560 A, as well as with side (SSC) and forward (FSC) scatters. Analyses were performed on ten biological replicates, using linear amplification. Histograms were collected as FCS files and evaluated using FloMax software (Partec GmbH, Münster, Germany), using manual individual gating, which assured that the G0/G1 peaks of a sample species and of an internal standard are included in a histogram. The coefficient of variation (CV) of the G0/G1 peak of Fabaceae species ranged between 1.86 and 5.30%. Nuclear DNA content was calculated using the linear relationship between the ratio of the 2C peak positions of Fabaceae species/internal standard on a histogram of fluorescence intensities. The Spearman's coefficient of rank correlation was calculated to quantify the relationship between the genome size and FCM markers in different seed parts.

For calibration of both cytometers, leaves of P. sativum were used. The original list-mode data files are available upon request at the Laboratory of Molecular Biology and Cytometry, UTLS in Bydgoszcz, Poland.


Germination Tests

Young green seeds of P. vulgaris, V. faba var. minor, and M. sativa, dissected from the pods at 20 DAF and dried, did not germinate (Table 2). At this stage, seeds of P. sativum expressed 36% germination and of V. sativa 70%. At 30 DAF, germination of these two species was already 90 and 95%, respectively. In P. vulgaris and V. faba var. minor germinability increased rapidly between 30 and 40 DAF, while M. sativa seeds exceeded 90% germination capacity first at 50 DAF. The highest germination percentage coincided with seed color change from green to yellow (or red in P. vulgaris), which is a visual sign of desiccation. After 4 months of storage germination did not change in any of the studied species.

Polysomaty in Different Seed Parts During Development

Flow cytometric analysis revealed the presence of endopolyploid nuclei in seeds of all studied species (Figs. 1 and 2). However, the intensity of endoreduplication differed between species, seed organs and developmental stages. The highest endopolyploidy was observed in the cotyledons of P. vulgaris, where nuclei with 2C, 4C, 8C, 16C, 32C, 64C, and 128C DNA content occurred. In contrast, the lowest endopolyploidy, only up to 8C, was detected in the cotyledons of M. sativa and embryo axes of most of the species; in the latter only in P. sativum were 16C nuclei also present, starting from 30 DAF. The proportion of the 2C nuclei was the highest in the cotyledons of M. sativa, being 80% at 20 DAF and 88% in the mature seeds, and the lowest in the cotyledons of P. vulgaris (30–40%).

Figure 1.

Proportions of nuclei with different DNA contents in cotyledons (A) and embryo axis (B) during development (DAF, days after flowering) and after 4-month storage (S) of five Fabaceae species.

Figure 2.

Selected histograms of DAPI fluorescence excited by an HBO lamp (366 nm; emission measured through a 435-nm long-pass filter, GG 435) in the cotyledon (A, B, E, F) and embryo axis (C, D, G, H) of P. vulgaris (A–D) and M. sativa (E–H) at 20 DAF (A, C, E, G), and at harvest maturity (B, D, F, H).

During P. vulgaris seed development the proportion of endopolyploid nuclei in the cotyledons increased from 28% at 20 DAF (8C–32C nuclei) to 40% at 60 DAF (8C–128C nuclei); the 64C nuclei appeared first at 30 DAF, and those containing 128C DNA at the last harvest date (Figs. 1A, 2A, and 2B). Afterstorage the proportion of endopolyploid nuclei was stillalmost as high as at harvest maturity (38%). In the embryo axis, only one endocycle occurred and the proportion of the 8C nuclei increased from 4 to 8% between 20 and 40 DAF (Figs. 1B, 2C, and 2D). At 60 DAF it decreased to 4% again.

The lowest endoreduplication intensity and difference in pattern of DNA synthesis from the other studied Fabaceae species, occurred in M. sativa seeds, where only nuclei with 2C, 4C, and 8C DNA content were detected in both the cotyledons and embryo axis (Figs. 1 and 2E–2H). In the cotyledons, the 8C nuclei were detected only up to 30 DAF and during further development only 2C and 4C nuclei were present. In contrast, in the embryo axes the 8C nuclei appeared first at 30 DAF and remained in almost the same proportion until the seeds reached harvest maturity and after storage.

In the cotyledons of P. sativum, nuclei with a DNA content up to 32C were detected at all developmental stages and the proportion of endopolyploid nuclei only slightly increased at the beginning of embryo development (from 19% at 20 DAF to 22% at 40 DAF), declining to 18% at harvest maturity (50 DAF; Fig. 1A). This species was the only one possessing 16C nuclei in the embryo axis, although in rather low proportion (1–2%; Fig. 1B).

In V. sativa cotyledons, the second highest (after P. vulgaris) proportion of endopolyploid nuclei (up to 64C) occurred, but in contrast to P. vulgaris it decreased during development; it was 36% at the first analyzed stage, 25% at 50 DAF and decreased further by 1% during storage (Fig. 1A). In the embryo axis of this species, there was almost no change in endoreduplication intensity during the whole of development, expressed as the presence of about 4–5% 8C nuclei (Fig. 1B).

Young cotyledons of V. faba var. minor contained nuclei with 2C, 4C, and 8C DNA content (Fig. 1A). A further two endocycles occurred in this organ at 30 and 40 DAF, respectively (the presence of 16C and 32C nuclei), which increased the initial proportion of endopolyploid nuclei of 3–21%. It decreased in the mature seeds to 13%. In the embryo axis, most of the nuclei were in the G0/G1 phase of the cell cycle (over 80% 2C nuclei), and only 1–2% of 8C nuclei represented endopolyploid ones (Fig. 1B).

Markers of Seed Developmental Stage and their Correlation with Genome Size

Both parameters, the 4C/2C ratio and (Σ>2C)/2C ratio were the highest in the cotyledons of the mature seeds of P. vulgaris that underwent five endocycles, and the lowest in the cotyledons of mature seeds of M. sativa, where there were no endopolyploid nuclei (Table 3). Endoreduplication intensity described by the (Σ>2C)/2C ratio corresponded with the number of endocycles and the proportion of endopolyploid nuclei, while the 4C/2C ratio reflected it less clearly. This was especially evident in cotyledons of V. faba var. minor, where the highest proportion of endoreduplicated nuclei (over 20%; Fig. 1A) was detected at 40 and 50 DAF and corresponded with the highest values of the (Σ>2C)/2C ratio (0.8–0.9; Table 3), while the highest 4C/2C ratio was observed at 20 DAF. Therefore, in further interpretation of the results the 4C/2C ratio was not taken into account. In the cotyledons of two species, M. sativa and V. sativa, the highest endoreduplication occurred at the early stages of seed development and in P. sativum and V. faba var. minor it was the highest at 40 DAF, and then decreased. In P. vulgaris cotyledons the (Σ>2C)/2C ratio increased during the whole of development, being the highest in harvest-mature seeds. In the embryo axis the intensity of endoreduplication was less variable. It was almost unchanged throughout development of the seeds of V. sativa and V. faba var. minor; in M. sativa it increased at 30 DAF and remained constant until the seeds were mature. In P. vulgaris it increased at 30 DAF but decreased at 50 DAF, and in P. sativum it was the highest in harvest-mature seeds.

Table 3. Markers of seed developmental stage in different embryo parts, during development and after 4 months of storage (S) of five Fabaceae species (DAF, days after flowering)
SpeciesSeed partTime of harvestNumber of endocycles4C/2C ratio(Σ>2C)/2C ratio
  • Values are means ± SD.

  • a

    Values for the species and parameters (in columns) followed by different letters are significantly different at P = 0.05 (Duncan's test); NS = no significant difference.

  • b

    Beginning of establishment of physiological maturity

  • c

    Physiological maturity.

P. vulgarisCotyledons20 DAF30.80 ± 0.10 da1.50 ± 0.16c
30 DAFb41.05 ± 0.08bc2.18 ± 0.27b
40 DAFc40.97 ± 0.05bc2.20 ± 0.20b
50 DAF40.96 ± 0.16c2.23 ± 0.37b
60 DAF51.10 ± 0.12b2.51 ± 0.42ab
S51.29 ± 0.20a2.71 ± 0.51a
Embryo axis20 DAF10.35 ± 0.06d0.40 ± 0.07d
30 DAFb10.80 ± 0.08b0.94 ± 0.09a
40 DAFc10.77 ± 0.11b0.93 ± 0.14a
50 DAF10.62 ± 0.05c0.74 ± 0.04b
60 DAF10.55 ± 0.08c0.62 ± 0.10c
S10.94 ± 0.06a1.01 ± 0.08a
M. sativaCotyledons20 DAF10.22 ± 0.03a0.25 ± 0.04a
30 DAF10.18 ± 0.03a0.21 ± 0.04b
40 DAFb00.21 ± 0.05a0.21 ± 0.05b
50 DAFc00.14 ± 0.05b0.14 ± 0.05c
S00.14 ± 0.02b0.14 ± 0.02c
Embryo axis20 DAF00.21 ± 0.07b0.21 ± 0.07b
30 DAF10.38 ± 0.14a0.42 ± 0.14a
40 DAFb10.38 ± 0.09a0.43 ± 0.10a
50 DAFc10.37 ± 0.11a0.39 ± 0.11a
S10.37 ± 0.10a0.39 ± 0.10a
P. sativumCotyledons20 DAF30.40 ± 0.03c0.74 ± 0.07c
30 DAFb30.50 ± 0.05b0.88 ± 0.11ab
40 DAFc30.55 ± 0.05a0.99 ± 0.13a
50 DAF30.48 ± 0.08c0.82 ± 0.15bc
S30.50 ± 0.07ab0.84 ± 0.16bc
Embryo axis20 DAF10.35 ± 0.09c0.40 ± 0.11b
30 DAFb20.43 ± 0.08b0.53 ± 0.10b
40 DAFc20.40 ± 0.08b0.52 ± 0.11b
50 DAF20.53 ± 0.08a0.68 ± 0.11a
S20.60 ± 0.16a0.78 ± 0.25a
V. sativaCotyledons20 DAF30.45 ± 0.03a1.29 ± 0.18a
30 DAFb40.46 ± 0.03a1.06 ± 0.06b
40 DAFc40.40 ± 0.05b1.01 ± 0.11b
50 DAF40.41 ± 0.04b0.88 ± 0.09c
S40.34 ± 0.07c0.75 ± 0.14c
Embryo axis20 DAF10.30 ± 0.02NS0.36 ± 0.04NS
30 DAFb10.33 ± 0.05 0.40 ± 0.06 
40 DAFc10.28 ± 0.02 0.35 ± 0.02 
50 DAF10.31 ± 0.06 0.39 ± 0.08 
S10.33 ± 0.14 0.39 ± 0.18 
V. faba var. MinorCotyledons20 DAF10.68 ± 0.28a0.73 ± 0.32bc
30 DAF20.38 ± 0.09b0.60 ± 0.15c
40 DAFb30.50 ± 0.07b0.95 ± 0.14a
50 DAFc30.41 ± 0.08b0.81 ± 0.15ab
60 DAF30.36 ± 0.04b0.56 ± 0.08c
S20.41 ± 0.08b0.65 ± 0.21bc
Embryo axis20 DAF10.22 ± 0.13a0.24 ± 0.13a
30 DAF10.14 ± 0.04b0.16 ± 0.04b
40 DAFb10.22 ± 0.05a0.25 ± 0.05a
50 DAFc10.18 ± 0.05ab0.20 ± 0.06ab
60 DAF10.15 ± 0.03b0.16 ± 0.04b
S10.18 ± 0.04ab0.19 ± 0.04ab

Genome size of the studied species varied between 1.28 and 26.87 pg/2C and applying the categorization proposed by Soltis et al. (19) it is very small, small or intermediate (Table 1, Fig. 3). As the Spearman's coefficient of rank correlation asserted, there is a negative correlation between nuclear DNA content and the markers of seed development stage (Table 4).

Figure 3.

Histograms of PI fluorescence excited by an high-grade solid-state laser (532 nm, emission measured through a 590-nm long-pass filter, RG 590 E), dot-plots of side scatter (SSC) vs. PI fluorescence, and forward scatter (FSC) vs. SSC of the nuclei isolated from leaves of (A) P. vulgaris and P. hybrida, (B) M. sativa and Z. mays, (C) P. sativum and Z. mays, (D) V. sativa and Z. mays, (E) V. faba var. minor and A. cepa. Histograms display gated (R1) nuclei of the G1/G0 peak of a sample (1) and of an internal standard (2).

Table 4. Correlations between markers of seed developmental stage and genome size
Seed partMarkerSpearman's coefficient of rank correlationStrength of correlation P
Genome size
Cotyledons4C/2C ratio−0.33weak0.020
(Σ>2C)/2C ratio−0.44average0.002
Embryo axis4C/2C ratio−0.56high0.000
(Σ>2C)/2C ratio−0.48high0.000


In seeds of Fabaceae species the endosperm is almost totally resorbed during embryo development and cotyledons become the major storage organ, which supports seedling growth before it establishes photosynthetic activity. At the beginning of embryo development, the cotyledon cells actively divide, but mitotic activity ceases during maturation (18, 25, 26). At this time and in mature seeds, in such Fabaceae species as P. vulgaris, P. sativum, V. faba, Glycine max, Lupinus albus, and Vigna radiata, high endopolyploidy was detected in the cotyledons (4, 16–18, 25, 26). Endoreduplication in this organ is directly related to storage protein and carbohydrate/oil synthesis, which confirms a suggestion that one of its roles is to increase the potential of gene activity in supplying multiple templates for transcription and translation (16, 27). In the present study endopolyploid nuclei were found in the cotyledons, supporting the hypothesis on endopolyploidization of storage cells. However, in M. sativa this occurred only at the beginning of seed development and as a low proportion. Evidently, in this species another mechanism must exist for the support of early seedling growth. Lack of endopolyploid nuclei can be related to a small size of the seeds of this species, a phenomenon that often correlates with endoreduplication intensity (25) and epigeal seedling development, which is characterized by the emergence of cotyledons above the soil surface; they become photosynthetic as the stored reserves are being depleted (2). In another species with epigeal seedling establishment, large-seeded P. vulgaris, however, endopolyploid nuclei up to 128C occurred, which suggests that the type of seedling emergence is not indicative of endoreduplication intensity. The type of cotyledons could be more determinative: M. sativa produces foliar ones, which have a greater initial allocation of photosynthetic tissue than the reserve-type ones of P. vulgaris and may respond quicker to light availability. The intensity of endoreduplication in the cotyledons of V. faba observed here was the same (32C) as previously detected (18), but in P. sativum it was lower than the earlier reported endopolyploidy level of up to 64C (26). However, variability in the maximum endopolyploidy in the cotyledons of this species (from 32C to 128C) is known, depending on seed size (25).

The pattern of endoreduplication in the cotyledons established in the present experiments depends on the species; nevertheless it can be considered as a marker of seed maturity. The beginning of establishment of physiological maturity was clearly marked in four out of the five species by the change in the number of endocycles: in P. vulgaris, V. sativa, and V. faba var. minor an additional endocycle occurred, and in M. sativa the endopolyploid 8C nuclei were no longer present (Table 3). When the seeds were dried at this time (at 30 DAF for P. vulgaris and V. sativa, and at 40 DAF for V. faba var. minor and M. sativa), they already had established some, but not maximal, germination energy and capacity (Table 2). After another 10 days they attained germinability close to 100%, related to full physiological maturity. Only in P. sativum cotyledons did endoreduplication intensity fail to correspond with the seed developmental stages, although physiological maturity was marked by the occurrence of an additional endocycle in the embryo axis at 30 DAF. The (Σ>2C)/2C ratio in the cotyledons also marked well seed physiological maturity and additionally indicated harvest maturity by the decrease in their value in all species but P. vulgaris. In the latter, the marker increased.

There is very little information on the pattern of DNA synthesis in the embryo axis of Fabaceae seeds. Bino et al. (4) observed some 8C nuclei in the radicle of the mature seed of P. vulgaris; however, in the developing embryo axis of G. max no endopolyploid cells were observed (16). In the species studied here, only in P. sativum did endoreduplication in the embryo axis reach 16C, while in the other species one endocycle occurred, resulting in the presence of 8C nuclei. The proportion of these nuclei in M. sativa, V. sativa, and V. faba var. minor did not change much during development, thus it did not correspond with the developmental stage of the embryo.

According to Barow and Meister (28), taxonomic position (family affiliation) is the major factor determining the degree of endopolyploidy within a species, while life cycle, genome size and organ type have a minor but also significant effect on endopolyploidization. This was not fully the case in seeds of species belonging to Fabaceae family. Although endoreduplication was detected in all studied species, its intensity varied significantly between them, with maximal endopolyploidy from 8C in M. sativa to 128C in P. vulgaris. As found previously, no endopolyploid nuclei occur in seeds of such perennial Fabaceae species as Olneya tesota and Parkinsonia aculeata (29). This instead confirms a suggestion that endoreduplication is negatively correlated with the length of the plant life cycle. Organ-specificity of endoreduplication was also clearly visible here and since the process was more intensive in the cotyledons its functional significance can be assumed. The negative correlation between endoreduplication and genome size, suggested by Nagl (27), was also confirmed.

In conclusion, in the Fabaceae, a marker based on the intensity of endoreduplication in the cotyledons, the (Σ>2C)/2C ratio, can be applied to follow seed development. Therefore, this is recommended to seed producers for establishing seed maturity and, consequently, optimal harvest time. The 4C/2C ratio should not be used for polysomatic species, because it can be misleading. Endopolyploidy occurrence in the cotyledons confirms that DNA amplification is necessary in the highly specialized storage cells, which play an important role in seedling development, unless another mechanism for supplying nutrients is provided. Further investigations, however, on a broader spectrum of species expressing epigeal and hypogeal seedling emergence are necessary to confirm this suggestion.

The seed embryo is a suitable model in which to follow plant tissue development and therefore the results obtained here can be used as part of a basis to study in vivo as well as in vitro cell/tissue differentiation. FCM, as a unique fast method for detecting cell cycle activity and endopolyploidy, can supplement or even replace more complex and time-consuming techniques, e.g. the establishment by microscopy of mitotic index and anatomy, as used to follow such processes as somatic embryogenesis or development of treachery elements. The 4C/2C ratio is recommended for tissues/organs with dividing cells and the (Σ>2C)/2C ratio for those also with endoreduplicating ones.


The authors thank Professor J. Derek Bewley (University of Guelph, Canada) for critical comments on the manuscript.