Are seeds suitable for flow cytometric estimation of plant genome size?


  • Elwira Sliwinska,

    Corresponding author
    1. Laboratory of Molecular Biology and Cytometry, Department of Genetics and Plant Breeding, University of Technology and Agriculture, Bydgoszcz, Poland
    • Department of Genetics and Plant Breeding, University of Technology and Agriculture, Al. Kaliskiego 7, 85-789 Bydgoszcz, Poland
    Search for more papers by this author
  • Elzbieta Zielinska,

    1. Laboratory of Molecular Biology and Cytometry, Department of Genetics and Plant Breeding, University of Technology and Agriculture, Bydgoszcz, Poland
    Current affiliation:
    1. Department of Biochemistry, Warsaw Agriculture University, Nowoursynowska St. 159, 02-776 Warsaw, Poland
    Search for more papers by this author
  • Iwona Jedrzejczyk

    1. Laboratory of Molecular Biology and Cytometry, Department of Genetics and Plant Breeding, University of Technology and Agriculture, Bydgoszcz, Poland
    Search for more papers by this author



Nuclear DNA content in plants is commonly estimated using flow cytometry (FCM). Plant material suitable for FCM measurement should contain the majority of its cells arrested in the G0/G1 phase of the cell cycle. Usually young, rapidly growing leaves are used for analysis. However, in some cases seeds would be more convenient because they can be easily transported and analyzed without the delays and additional costs required to raise seedlings. Using seeds would be particularly suitable for species that contain leaf cytosol compounds affecting fluorochrome accessibility to the DNA. Therefore, the usefulness of seeds or their specific tissues for FCM genome size estimation was investigated, and the results are presented here.


The genome size of six plant species was determined by FCM using intercalating fluorochrome propidium iodide for staining isolated nuclei. Young leaves and different seed tissues were used as experimental material. Pisum sativum cv. Set (2C = 9.11 pg) was used as an internal standard. For isolation of nuclei from species containing compounds that interfere with propidium iodide intercalation and/or fluorescence, buffers were used supplemented with reductants.


For Anethum graveolens, Beta vulgaris, and Zea mays, cytometrically estimated genome size was the same in seeds and leaves. For Helianthus annuus, different values for DNA amounts in seeds and in leaves were obtained when using all but one of four nuclei isolation buffers. For Brassica napus var. oleifera, none of the applied nuclei isolation buffers eliminated differences in genome size determined in the seeds and leaves.


The genome size of species that do not contain compounds that influence fluorochrome accessibility appears to be the same when estimated using specific seed tissues and young leaves. Seeds can be more suitable than leaves, especially for species containing staining inhibitors in the leaf cytosol. Thus, use of seeds for FCM nuclear DNA content estimation is recommended, although for some species a specific seed tissue (usually the radicle) should be used. Protocols for preparation of samples from endospermic and endospermless seeds have been developed. © 2005 Wiley-Liss, Inc.

Flow cytometry (FCM) is commonly used for plant genome size estimation, especially after Galbraith et al. (1) developed a rapid procedure for nuclei isolation from intact plant tissues (2–7). Despite a suggestion that all kinds of plant tissue are suitable for FCM analysis (2, 8, 9), usually young, rapidly growing leaves are used for 2C nuclear DNA content estimation (5). This material is easy to obtain and gives high resolution histograms. However, leaves must be used soon after harvesting, it is difficult to store them (freezing usually lowers resolution of the FCM histograms), and no prolonged period of time can be taken for their transportation. Moreover, time and growing space (usually a greenhouse) are necessary to produce seedling leaves, and in some species nucleus-cytosol interactions affect fluorochrome accessibility to DNA, causing a stoichiometric error in estimation of DNA content (10–13). In addition, developmental variations in cell-cycle behavior occur in leaves, especially in differentiated tissues (1, 14–16).

FCM analysis of ploidy of different tissues of mature seeds provides information about the reproductive behavior of plants (17). FCM is also a fast and accurate method for analysis of the cell cycle in seeds (18–23). It allows determination of the nuclear replication stages in developing, mature, and germinating seeds and makes it possible to follow the progress of seed enhancement treatments. In dry seeds a majority of, or in some species all of, the embryo cells are arrested in the G0/G1 phase of the cell cycle. This is suggestive of the usefulness of such seeds as material for genome size estimation. However, seeds of some species contain an endosperm in addition to the embryo. This is a storage tissue, which in diploid species is triploid (formed by the fusion of two haploid nuclei from the female gametophyte and one haploid nucleus from the male gametophyte), composed often of regularly divided cells and ones that are endoreduplicated (e.g., up to 192C in maize) (19, 21, 24). Additional peaks on FCM histograms, representing endosperm cells, can overlap the G0/G1 peak of the species used as a reference (internal) standard (species of known genome size processed simultaneously with a target species) and thus confuse genome size estimations. In addition, in the seeds of some species (e.g., Anethum graveolens) the endosperm represents a major part of the seed and the 3C endosperm peak appearing on the histogram can be mistaken for the 2C embryo peak. Such confusing observations may be the reason seeds are not favored for plant genome size estimation. However, because seeds are easy to transport and store and can be analyzed dry at any convenient time, it appears worthwhile to develop procedures to use them as material for FCM determination of absolute DNA content.

The aim of this study was to determine whether a seed or any specific tissue therein could be used for genome size estimation by FCM and, if so, to establish optimal preparation procedures. We estimated nuclear DNA content in endospermic seeds (A. graveolens and Zea mays) and in endospermless ones (Beta vulgaris, Brassica napus var. oleifera, Helianthus annuus, and Pisum sativum) and we compared the data with those from leaves. Hydrated seeds were occasionally used before to analyze genome size variation in Pisum (25); however, to the best of our knowledge, this is the first report on the estimation of the absolute 2C DNA content of dry seeds by FCM.


Plant Material

Seeds and seedlings of six plant species, Pisum sativum L. cv. Set (pea), Anethum graveolens L. cv. Lazuryt (dill), Zea mays L. cv. Limabest (maize), Beta vulgaris L. cv. Arthur (sugar beet; 2x), Helianthus annuus L. cv. Paskowany (sunflower), and Brassica napus L. var. oleifera Moench cv. Bor (rapeseed), were used for FCM analyses. Seedlings were raised in a growth chamber adjusted to a 16-h/8-h, 26°C/20°C day/night cycle. Pisum sativum cv. Set was used as an internal standard. Its DNA amount was estimated in seeds and leaves (9.11 ± 0.10 pg/2C and 9.11 ± 0.11 pg/2C, respectively) using male human leukocytes (2C = 7.0 pg) (26) as a reference standard.

Flow Cytometry

Samples were prepared according to the method of Galbraith et al. (1) with some modifications. For each species, three combinations of different tissues were prepared: (a) leaf of the target species and leaf of P. sativum (internal standard); (b) selected seed tissue of the target species and leaf of P. sativum; and (c) selected seed tissue of the target species and radicle of P. sativum. Whole seed (B. napus), half seed with embryo (A. graveolens), true seed (removed from the pericarp by dissection) without radicle (B. vulgaris), or radicle (Z. mays, H. annuus) were used for sample preparation. Selected plant tissues of the target species and of the internal standard were chopped simultaneously with a sharp razor blade in a plastic Petri dish with 1 ml standard nucleus-isolation buffer (0.1 M Tris-HCl, 2.5 mM MgCl2. 6H2O, 85 mM NaCl, 0.1% v/v Triton X-100; pH 7.0; buffer A), supplemented with propidium iodide (PI; 50 μg/ml) and ribonuclease A (50 μg/ml). In addition, for DNA content estimation of H. annuus and B. napus, three other buffers were used: buffer B consisted of buffer A supplemented with 1% w/v PVP-10 (polyvinylpyrrolidone); buffer C consisted of Galbraith's buffer (45 mM MgCl2, 30 mM sodium citrate, 20 mM 3-[N-morpholino] propanesulfonic acid, 0.1% v/v Triton X-100, pH 7.0; 1) supplemented with ribonuclease A (50 μg/ml) and 1% w/v PVP-10; PI (50 μg/ml) was added after 10 min of incubation in the isolation buffer; buffer D consisted of buffer LB01 (15 mM Tris-HCl, 2 mM Na2 ethylenediaminetetraacetic acid, 0.5 mM spermine tetrahydrochloride, 80 mM KCl, 20 mM NaCl, 0.1% v/v Triton X-100, 15 mM β-mercaptoethanol, 50 μg/ml PI, and 50 μg/ml ribonuclease A, pH 7.5) (27). After chopping, the suspension was passed through a 50-μm mesh nylon filter. Then the filtrate was analyzed immediately after preparation (leaves) or incubated for 30 min (combinations a and c, including dry seed tissue that, according to our previous observations, needs the time for an effective staining) at room temperature. For each sample, DNA content of 7,000 to 10,000 nuclei was measured using a Partec CCA (Münster, Germany) flow cytometer. Analyses were performed on 20 replicates of each combination. Histograms were analyzed using a DPAC 2.2 computer program. Nuclear DNA content was calculated using the linear relation between the ratio of the 2C peak positions, target species/Pisum, on histogram of fluorescence intensities (9). The results were estimated using a single-factor analysis of variance and Duncan's test.

A test was also performed for the presence of PI-staining inhibitors in the plant tissues, as recommended by Price et al. (12), using buffer A. Nuclei were released from leaf or seed tissues of the internal standard alone and stained with PI (sample 1). Sample 2 consisted of specific tissues of the target species and of the internal standard processed simultaneously (co-chopped). FCM measurement of sample 2 followed directly the measurement of sample 1. The experiment was replicated five times. Decreased fluorescence of internal standard nuclei in sample 2, compared with the fluorescence of its nuclei in sample 1, was taken as evidence for the presence of inhibitors in tissues of the target species.


Different species and plant tissues showed different profiles of nuclear DNA contents. FCM of young leaves of all species revealed two peaks, a major one corresponding to the nuclei arrested in the G0/G1 phase of the cell cycle (2C DNA content) and a minor one of nuclei in the G2 phase (4C DNA content; Fig. 1A). A similar distribution of nuclei of different DNA contents was observed when whole seeds of B. napus and H. annuus were used (Fig. 2A and 2B). However, because the radicle of H. annuus produced higher resolution histograms than the whole seed (probably because oil, which contains phenols, is stored in the cotyledons), this tissue was selected for analysis of the genome size of this species. For other species, suitable seed tissue, containing a majority of the cells with 2C DNA content, was also determined. In the whole seeds of P. sativum and Z. mays, in addition to the 2C peak, a larger 4C peak and an additional 8C peak, corresponding to the endoreduplicated embryo nuclei, were detected (Figs. 1B and 2C, respectively). In the endospermic seed of Z. mays, virtually no endosperm nuclei were observed when the whole seed was analyzed because the large storage cells of endosperm that are full of storage reserves do not contain nuclei (Fig. 2C). In P. sativum and Z. mays seeds, endoreduplicated embryo nuclei were located in the cotyledons (Fig. 1C). In these species, therefore, as in H. annuus, the most suitable tissue for use is that of the radicle (Figs. 1D and 2D).

Figure 1.

Histograms of relative propidium iodide fluorescence in nuclei of Pisum sativum isolated from different tissues. A: Young leaf. B: Whole seed. C: Cotyledon. D: Radicle.

Figure 2.

Histograms of relative propidium iodide fluorescence in nuclei of the seed tissues of different species. A:Brassica napus, whole seed. B:Helianthus annuus, whole seed. C:Zea mays, whole seed. D:Zea mays, radicle. E:Anethum graveolens, whole seed. F:Anethum graveolens, half seed with embryo.

In B. vulgaris and A. graveolens seeds, peaks corresponding to endosperm nuclei and those of the embryo were detected (Fig. 2E; for B. vulgaris see 20). B. vulgaris seed is regarded as being endospermless because only the radicle is surrounded by a single layer of endosperm. Hence, a true seed without its radicle and the endosperm remnants was used for DNA content estimation. In endospermic seed of A. graveolens the embryo is relatively small and surrounded by the endosperm and hard pericarp, which makes difficult to isolate the embryo/radicle. The intact seed is not the best material for genome estimation, however, because the highest peak on the FCM histogram is that corresponding to 3C endosperm nuclei (Fig. 2E). Instead, half of the seed with the embryo (together with the pericarp that, being a dead tissue, does not contain any cells with intact nuclei) was used to prepare samples for FCM. These still contained a considerable proportion of endosperm nuclei, but the G0/G1 peak of the embryo nuclei was easy to distinguish (Fig. 2F).

Genome size measurements for B. vulgaris, Z. mays, and A. graveolens gave similar values (no significant differences) whether estimated in leaves or seeds when using the standard procedure (buffer A) for sample preparations (Table 1, Fig. 3A–C). The 2C DNA amounts for these species were 1.44–1.45 pg, 5.03–5.07 pg, and 2.66–2.69 pg, respectively. However, significant differences in genome size were obtained between leaves and seeds of H. annuus and B. napus, species that contain phenolic and possibly other compounds influencing PI staining, even when reductants were added to the buffers (Table 1, Fig. 3D–F). A presence of such compounds was confirmed in leaves of H. annuus and seeds of B. napus by reduced ratio of PI fluorescence of nuclei of Pisum tissues simultaneously processed and stained with the target species tissues to the fluorescence of Pisum nuclei processed independently, using buffer A for nuclei isolation (Table 2). Only when buffer C was used for preparation of H. annuus samples, was the estimated 2C DNA content the same, no matter what tissue combination was analyzed. This value of 7.10–7.17 pg was lower than that obtained for seeds when the other buffers were used for nuclei isolation (about 7.3–7.4 pg). The 2C values obtained for B. napus seeds, regardless of the isolation buffer used, were significantly higher than those for leaves (2.4–2.6 pg and 2.2–2.3 pg, respectively). The higher estimation of DNA content in seeds than in leaves in these two species was not a result of using a prolonged staining time for seeds (data not shown).

Table 1. Nuclear DNA Content Estimated for Different Plant Tissues*
SpeciesBufferNuclear DNA content (pg/2C)
Plant tissue combinations (target species/internal standard)
  • *

    Using internal standard Pisum sativum 2C = 9.11 pg; for buffers composition, see Material and Methods. NS, no significant difference. Values for particular species and buffer (in lines) followed by the same letter are not significantly different at P = 0.05 (Duncan's test). Data are mean ± standard deviation.

Anethum graveolensA2.67 ± 0.06NS2.69 ± 0.082.66 ± 0.06
Zea maysA5.07 ± 0.08NS5.03 ± 0.085.04 ± 0.09
Beta vulgarisA1.44 ± 0.02NS1.45 ± 0.031.45 ± 0.02
Helianthus annuusA7.08 ± 0.07b7.34 ± 0.13a7.37 ± 0.18a
 B7.03 ± 0.06b7.44 ± 0.11a7.40 ± 0.10a
 C7.10 ± 0.05NS7.14 ± 0.097.17 ± 0.12
 D7.20 ± 0.08b7.33 ± 0.07a7.29 ± 0.10a
Brassica napus var. oleiferaA2.20 ± 0.04c2.43 ± 0.05b2.56 ± 0.09a
 B2.26 ± 0.07c2.43 ± 0.06b2.53 ± 0.08a
 C2.27 ± 0.06c2.40 ± 0.06b2.59 ± 0.13a
 D2.25 ± 0.03b2.39 ± 0.05a2.39 ± 0.08a
Figure 3.

Histograms of fluorescence intensities in nuclei of different tissues of Zea mays (A–C) and Brassica napus (D–F) stained simultaneously with the nuclei of Pisum sativum (internal standard; 2C DNA content 9.11 pg) using propidium iodide and isolation buffer A. A, D: Leaves of target species and internal standard. B, E: Seed tissue of target species and leaf of internal standard. C, F: Seed tissues of target species and internal standard. Peak 1, G0/G1 nuclei of target species; peak 2, G0/G1 nuclei of internal standard.

Table 2. Ratio of Propidium Iodide Fluorescence of Nuclei of Pisum Tissues Simultaneously Processed and Stained With Target Species Tissues (Sample 2) to Fluorescence of Nuclei From Independently Processed Pisum Tissues (Sample 1)*
SpeciesSample 2/sample 1 ratio
Plant tissue combinations in sample 2 (target species/Pisum)
  • *

    Values are mean ± standard deviation. NS, no significant difference. Values for particular species followed by the same letter are not significantly different at P = 0.05 (Duncan's test).

Anethum graveolens1.01 ± 0.02NS0.99 ± 0.021.00 ± 0.01
Zea mays1.00 ± 0.02NS1.02 ± 0.021.02 ± 0.00
Beta vulgaris1.01 ± 0.00NS1.01 ± 0.021.02 ± 0.01
Helianthus annuus0.93 ± 0.04b1.02 ± 0.02a1.01 ± 0.02a
Brassica napus var. oleifera1.02 ± 0.02a0.96 ± 0.01b0.92 ± 0.02c


Young leaves usually contain most of their cells in the G0/G1 phase of the cell cycle, and because the DNA content of such cells is considered a 2C value characteristic of a certain species (C is the DNA amount of a haploid genome), they are very suitable for FCM measurements of genome size (9). However, in older leaves, especially in their distal part and veins, endoreduplication often occurs, which can confuse FCM genome size estimation (1, 14, 15; Sliwinska, unpublished results). In addition, leaves of some species may contain chemicals in the cytosol, such as phenols, tannins, or caffeine, that modify fluorochrome accessibility to DNA. Cytosolic compounds can considerably bias nuclear DNA content estimations, and there is no method to completely overcome their effects on the measurements (10, 11, 13). Therefore, using other plant tissues, free of such inhibitors, is a better alternative.

The usefulness of sugar beet seeds for ploidy estimation has been shown (20), and they are routinely used in our laboratory for checking the ploidy level of commercial seed lots of hybrid cultivars. However, estimation of absolute DNA content using dry seeds has not been reported. Our results confirmed previous observations (19, 21) that tissues of different ploidy levels are present in dry seeds of some species (especially in endospermic ones), which makes FCM estimation of genome size difficult. However, the use of specific seed tissues, usually the radicles, which, like young leaves, contain most cells in the G0/G1 phase of the cell cycle, facilitates interpretation of the histograms, and makes this plant material suitable for DNA content measurement by FCM. Cotyledons (e.g., these of P. sativum) can be less useful because they may contain endoreduplicated cells. It is possible to select an appropriate seed part for FCM even when the embryo constitutes only a small portion of the seed and cannot be easily isolated (e.g., A. graveolens). However, such a sample still usually contains endosperm nuclei, and a proper internal standard has to be used, with 2C DNA content considerably different from the 2C and 3C of the target species.

There is no universal internal standard for plants, and each FCM laboratory uses its own. Often it is breeding material (e.g., inbred line of Z. mays) (28) that may not be easily available or a species with very small seeds that makes difficult the isolation of sufficient amounts of a specific tissue (e.g., Petunia hybrida) (4). In such cases there is more convenient to use leaves of internal standard rather than its seeds (from one seed a plant with many leaves can be raised, and one leaf is usually more than enough for FCM sample preparation). In the present study we used seeds and leaves of the internal standard (P. sativum) to estimate the DNA content of seeds of target species. Similar 2C values in seeds of the target species were found no matter which tissue, leaf, or seed of P. sativum was used. This suggests that, if more convenient, leaves of an internal standard species can be used as well as its seeds.

The genome size of the species studied here was previously estimated by FCM (using only leaves) or/and Feulgen microdensitometry (3–5, 7, 29). The value obtained for B. vulgaris in this study (1.45 pg) is slightly lower than that reported previously [1.57 pg (3), 1.6 pg (5), and 1.75 pg (4)], probably because different cultivars and internal standards were used. In different cultivars of Z. mays, the 2C DNA content varies from 4.42 pg to 6.75 pg (3–5, 7), so our estimation, about 5.0 pg, falls within this range. The genome size of A. graveolens obtained here (2.7 pg) is higher than that assessed by Feulgen microdensitometry (2.4 pg) (29). FCM was not used for this species before.

The published variation in DNA content is higher for species containing PI-staining inhibitors in their tissues, e.g., H. annuus (3, 5, 12, 30). These values depend not only on cultivar and internal standard used but also on the composition of the buffers and procedures used for sample preparation. It was suggested that in all FCM studies a test for the presence of compounds influencing PI intercalation and/or fluorescence should be used (12). The tests performed here detected compounds causing a decrease in fluorescence of nuclei of P. sativum leaves when they were processed simultaneously with leaves of H. annuus and seeds of B. napus; this confirmed previous reports concerning a presence of phenolic compounds and/or others staining inhibitors in these species (12, 31). The decrease was not observed, however, when radicles of H. annuus were used, presumably because of a lack of PI-staining inhibitors in this tissue, which makes them more suitable for genome size estimation than the leaves. The DNA content estimated in radicles (7.3–7.4 pg) is very close to the best obtained by Price et al. (12) (7.3 pg) when procedures to avoid PI intercalation and/or fluorescence inhibitors effect were tested, and higher than estimated in leaves (7.0–7.2 pg). Similarly, the DNA content in seeds of B. napus, even estimated after nuclei isolation in buffer with no reductants, was higher than in leaves and closer to those found by other investigators (2.6–2.8 pg) (3–5). The reason for this is not clear. Our results, however, question the common assumption that reductants such as polyvinylpyrrolidone and β-mercaptoethanol neutralize the inhibitory effect of phenols on PI accessibility (12, 15, 32–34). These data, therefore, support the findings of Price et al. (12) that such compounds have no measurable effect on reducing inhibition (however, in most of the cases they improve resolution of the histograms).

Our results demonstrate that seeds not only can be used for plant genome size estimation, but in some cases are better material for FCM analysis than leaves. Seeds of the target species can be analyzed with seeds and with leaves of internal standard. However, seeds of some species contain tissues with considerable amounts of cells with other than a 2C DNA content, and in such cases specific tissues (most often the radicle), rather than the intact seed, must be selected for sample preparation. Estimating the absolute DNA content in seeds cannot be recommended for research in taxonomy and population biology, especially of cross-pollinated species. Because the origin of male gametes is not known, and the phenotype of the analyzed genotype cannot be evaluated, there is a risk that hybrids are mistaken for true-to-type specimens.


The authors thank the Plant Breeding Station (Wiatrowo, Poland) for kindly providing the P. sativum seeds used as an internal standard and Prof. M. Jassem (University of Technology and Agriculture, Bydgoszcz, Poland) and Prof. J. D. Bewley (University of Guelph, Canada) for critical comments on the manuscript.