Introduction

Ploidy level, the number of chromosome sets in a cell nucleus, is typically identified as an organism-level trait. In a diploid organism, there are two chromosome sets in a cell nucleus. However, polyploidy, with more than two chromosome sets in a nucleus, is relatively common or even prevalent in some plant lineages (Husband et al. 2013; Van de Peer et al. 2017). However, ploidy level may vary even within the same individual, e.g. most of the cells of the organism are diploid, but some tissues or group of cells may contain higher numbers of chromosome sets. This state of mixed ploidy tissues is known as endopolyploidy or somatic polyploidy, and it is a result of repeated genome multiplication without cytokinesis (D’Amato 1964; Nagl 1976; Smulders et al. 1994).

Endopolyploidy is generated through various processes, which include endomitosis (prevalent in animals), endoreduplication (dominant in plants) and progressively partial endoreplication (D’Amato 1989; Joubès and Chevalier 2000; Frawley and Orr-Weaver 2015; Trávníček et al. 2015; Leitch and Dodsworth 2017). During endomitosis, mitosis occurs with the absence of cytokinesis. The results of repeating endomitoses are multilobulated nuclei in cells or multinucleated cells (Abraham et al. 1965; Hasinoff et al. 2000; Frawley and Orr-Weaver 2015; Von Stetina et al. 2018; Kobayashi 2019). Opposite of endomitosis, endoreduplication occurs when a mitotic cell cycle shifts into an endocycle. DNA replication occurs without chromosome separation and cytokinesis, which leads to the formation of polytene chromosomes (Leitch and Dodsworth 2017; Kobayashi 2019). In progressively partial endoreplication, only a part of the genome is replicated during the synthetic S phase; however, this mechanism has only so far been documented in the Orchidaceae (Trávníček et al. 2015).

Whole organism endopolyploidy level (EL) varies among plant species and plant lineages, and while endopolyploidy may be absent in some plants, e.g. gymnosperms, it can be prevalent in others, e.g. bryophytes and angiosperms (Barow and Meister 2003; Bainard and Newmaster 2010; Bainard et al. 2011a; Bainard et al. 2012). Some estimates indicate that 90 % of angiosperms may be polysomatic, but the occurrence of polysomaty is limited to within certain groups or families (Nagl 1976; Barow and Meister 2003). Some plant families, e.g. the Brassicaceae or Fabaceae, are polysomatic, whereas others only exhibit polysomaty rarely, e.g. the Asteraceae (Joubès and Chevalier 2000; Larkins et al. 2001; Barow and Meister 2003; Jovtchev et al. 2007; Bainard et al. 2012). Endopolyploidy occurs irrespective of organism ploidy level and is present in both diploids and polyploids; however, it is not absolutely independent of organism cytotype, and it commonly occurs less frequently in polyploids than in diploids (Mishiba and Mii 2000; Jovtchev et al. 2007; Pacey et al. 2020a).

Determining how closely associated endopolyploidy is with physiological processes of plant growth and development remains an emerging field as technology improves. What has been thoroughly established is that cells of higher ploidy levels contain nuclei with higher DNA content. Furthermore, due to the physical constraint on minimum nuclei size, endopolyploidy is strongly correlated with nuclear and cellular size and with organ size in some plants (Melaragno et al. 1993; Cookson et al. 2006; Agulló-Antón et al. 2013; Kladnik 2015; Robinson et al. 2018). Increased cell size and gene expression are components related to increased growth, and endopolyploidy commonly arises in developmentally and metabolically active cells and tissues, such as vascular tissue, nutritive tissue, cotyledons and pericarp (Barow and Meister 2003; Kladnik et al. 2006; Bourdon et al. 2012; Sabelli et al. 2013; Bhosale et al. 2018; Kobayashi 2019), and in highly specialised cells, such as trichomes, antipodal cells, leaf bladder cells and tapetum (Melaragno et al. 1993; Weiss and Maluszynska 2001; Bartoli et al. 2017; Barkla et al. 2018). Conversely, endopolyploidy is an obstacle for the proper function of some specialised cells and is not commonly reported in meristematic, guard or gametic cells (Melaragno et al. 1993; Trávníček et al. 2015; Skaptsov et al. 2017).

An increase in the frequency of endopolyploidy has been documented as being associated with various biological processes and stress responses, but not necessarily related to development. Plant stress responses to UV-B or gamma irradiation can trigger a switch into an endocycle (Gegas et al. 2014; Zedek et al. 2016). Stress-associated endopolyploidy can also occur as part of a response to increased salinity (Ceccarelli et al. 2006; Bennici et al. 2008), whereas opposite effects have been observed in water deficiency experiments (Artlip et al. 1995; Cookson et al. 2006). Plant infection by nematodes can cause the formation of highly endopolyploid cells, where such cells serve as source for nematode nourishment (de Almeida Engler and Gheysen 2013; Smant et al. 2018). Plant root colonisation by arbuscular mycorrhizal fungi can also lead to the enhancement of EL (Lingua et al. 2001; Bainard et al. 2011b).

Generally, our knowledge of the functional roles of endopolyploidy is limited and has been mainly acquired from model polysomatic plants such as Arabidopsis thaliana (L.) Heynh., Solanum lycopersicum L. and Zea mays L. (Chevalier et al. 2014; Li et al. 2019; Musseau et al. 2020; Pacey et al. 2020a). Therefore, studies of different plants may reveal many new functions and roles of endopolyploidy, e.g. in Mesembryanthemum crystallinum L. (Barkla et al. 2018). Basic broad searches for endopolyploidy presence in various non-model plant groups are not common (e.g. Barow and Meister 2003; Bainard et al. 2012). Additionally, efforts to perform these studies are highly valuable, because EL may also be an important organismal feature in addition to the number of chromosomes or ploidy level.

Previous studies on endopolyploidy in solanaceous crops (Barow and Meister 2003; Galbraith 2014) revealed that Nicotiana tabacum L. and Capsicum annuum L. show polysomaty in some of their organs. According to these studies, petal, sepal, stamen, carpel, pedicel, upper leaf, lower leaf and cotyledon tissues of N. tabacum showed low and moderate EI (0.11 to 0.72). Furthermore, the fruit pericarp of C. annuum showed incredibly high levels (presence of up to 256 C cells) of endopolyploidisation in their cells, while the leaves of C. annuum were not polysomatic.

In this study, we investigated endopolyploidy within the Solanaceae, primarily focusing on N. tabacum and C. annuum. Moreover, we investigated EL in seven vegetative and two reproductive organs, most of which had not been previously analysed for C. annuum. We also studied four other taxa, C. baccatum L., C. chinense Jacq., N. rustica L. and N. sylvestris Speg., which have not had a survey for the presence of endopolyploidy published to date. Our results confirm that all of these investigated species are polysomatic. In addition, we have found that the EL is comparable among them and documented the association between EL and organism ploidy level in Nicotiana.

Materials and methods

Plant material

Plants used in the flow cytometry analyses (FCM) were cultivated from commercially available seeds (www.semenaonline.sk) of Capsicum annuum ‘Citrina’ (diploid, 2n = 24; Rice et al. 2015), C. baccatum ‘Aji’ (diploid, 2n = 24; Rice et al. 2015), C. chinense ‘Dorset Naga’ (diploid, 2n = 24; Rice et al. 2015), Nicotiana rustica (tetraploid, 2n = 48; Rice et al. 2015), N. sylvestris (diploid, 2n = 24; Rice et al. 2015) and N. tabacum (tetraploid, 2n = 48; Rice et al. 2015). Plants were sown in 15 cm pots with a common horticultural substrate and grown in the laboratory under the controlled conditions of 20°C, 60 % humidity and 12 h photoperiod cycle (12 h light/12 h darkness).

Analyses of endopolyploidy were performed on fully developed, mature organs of 2-month-old plants. Since various species required different conditions to reach maturity, not all species were investigated at anthesis. Individuals of N. sylvestris only reached the rosette stage, and they showed delayed development. Delays in development also occurred with some individuals of C. baccatum. Since not all tissue types were available for all investigated species, analyses were performed on three to nine organ tissues of each of the six investigated species (Table 1). The investigated organ tissues included root, stem, lower leaf petiole, lower leaf lamina, flower pedicel, calyx, corolla (tube and limb analysed separately in N. tabacum), pistil and stamen tissue.

Table 1 Survey of endopolyploidy level (EL) in species and organs investigated in this study
Full size table

Genome size analyses

The samples for genome size analyses were prepared from leaf tissue using a two-step procedure, which consisted of nuclear isolation and staining steps using propidium iodide as a DNA intercalator (Doležel and Göhde 1995; Loureiro et al. 2007). The internal standardisation method was used (Doležel et al. 2007) with the internal reference standards Zea mays ‘CE-777’ (2C = 5.43 pg; Lysák and Doležel 1998) for Capsicum L. species and Solanum pseudocapsicum L. (2C = 2.59 pg; Temsch et al. 2010) for Nicotiana L. species to minimise distortion in the data measurements.

Approximately 0.75 cm2 of young leaf tissue of each sample and the standard were chopped together with a razor blade in a Petri dish in 1 mL of ice-cold general-purpose buffer (GPB) [0.5 mM spermine × 4 HCl, 30 mM sodium citrate, 20 mM MOPS (4-morpholine propane sulfonate), 80 mM KCl, 20 mM NaCl, 0.5 % (v/v) Triton X-100, pH 7.0] (Loureiro et al. 2007) supplemented with 3 % PVP (polyvinylpyrrolidone). Each suspension was then filtered through 42 µm nylon filter, and each sample was subsequently incubated for approximately 0.5 h with β-mercaptoethanol (final concentration 2 µL/mL), RNase (30 µg/mL) and propidium iodide (30 µg/mL). Three individuals from each of the six species were measured three times on three different days for a total of nine measurements per species (Greilhuber and Obermayer 1997), except for the N. tabacum accession, where only two replicates of sufficient quality [coefficients of variation (CV) < 5.5 %] were measured.

Measurements for genome size were performed using a Partec CyFlow ML (Partec Gmbh, Münster, Germany) flow cytometer, which is housed at the Institute of Biological and Ecological Sciences, P. J. Šafárik University in Košice (Slovakia). The flow cytometry machine is equipped with a 532 nm (150 mW) green laser and 590 nm band-pass optical filter. FloMax ver. 2.70 software (Partec Gmbh, Münster, Germany) was used for flow cytometry analyses and final evaluation of all measurements. Histograms of the data were displayed on a linear scale (x-axis). At least 5 000 nuclei per measurement were collected, and the CV of the G0/G1 peaks of both the samples and the internal standards did not exceed 5.5 %. The estimation of the quantity of DNA in each sample was based on the value of the G0/G1 peak means calculated using the following equation from Doležel and Bartoš (2005):

$$ \mathrm{Amount}\ \mathrm{of}\ \mathrm{sample}\ \mathrm{DNA}=\mathrm{Amount}\ \mathrm{standard}\ \mathrm{DNA}\ \mathrm{used}\times \left[\frac{\left(\mathrm{sample}\ {\mathrm{G}}_0/{G}_1\ \mathrm{peak}\ \mathrm{mean}\right)}{\left(\mathrm{standard}\ {\mathrm{G}}_0/{G}_1\ \mathrm{peak}\ \mathrm{mean}\right)}\right] $$

Determining endopolyploidy using flow cytometry

Flow cytometry techniques were used to determine the nuclei frequency of ploidy classes, such as 2C, 4C, 8C, etc. The protocol for FCM preparation was used as described by Kocová et al. (2017). In brief, we applied a chopping technique using a razor blade to release nuclei from a given organ into 1 mL of ice-cold GPB + 3 % PVP. Nuclei filtering, staining and FCM measurements followed as described above for genome size determination. Log-scaled fluorescence intensity histograms were inspected for the symmetry of peaks and proportion of background debris. Flow cytometry records with skewed peaks or high amounts of background debris were excluded from final analyses.

Evaluation of endopolyploidy level

Endopolyploidy level was assessed through various approaches. The number of nuclei for each ploidy level class (2C, 4C, 8C, etc.) was recorded, and four different indices were calculated based on the recorded number of nuclei of each ploidy level class. Mean C value (Lemontey et al. 2000) also referred to as mean C-level (Jovtchev et al. 2007) represents the mean ploidy of cells of a tissue and is calculated according to the formula:

$$ \mathrm{MCV}=\frac{\left[\left(2\times \mathrm{n}2\mathrm{C}\right)+\left(4\times \mathrm{n}4\mathrm{C}\right)+\left(8\times \mathrm{n}8\mathrm{C}\right)+\left(16\times \mathrm{n}16\mathrm{C}\right)\right]}{\left(\mathrm{n}2\mathrm{C}+\mathrm{n}4\mathrm{C}+\mathrm{n}8\mathrm{C}+\mathrm{n}16\mathrm{C}\right)} $$

where n2C, n4C, n8C… are nuclei counts of corresponding ploidy level classes (2C, 4C, 8C…).

The endoreduplication index (Bainard et al. 2012) or cycle value (Barow and Meister 2003) indicates the mean number of endoreduplication cycles per nucleus of the cells and follows the equation:

$$\text{E}\text{I}= \frac{\left[(0\times \text{n}2\text{C}) +(1\times \text{n}4\text{C}) +(2 \times \text{n}8\text{C}) +(3\times \text{n}16\text{C})\dots \right] }{(\text{n}2\text{C} + \text{n}4\text{C} + \text{n}8\text{C} + \text{n}16\text{C}\dots )}$$

The parameter E4P is calculated according to the original parameter E6P proposed by Dilkes et al. (2002) for endospermal tissue and represents mean ploidy of endoreduplicated cells of a tissue using the equation:

$$\text{E}4\text{P}= \frac{\left[(4\times \text{n}4\text{C}) +(8 \times \text{n}8\text{C}) +(16\times \text{n}16\text{C})\dots \right] }{(\text{n}4\text{C} + \text{n}8\text{C} + \text{n}16\text{C}\dots )}$$

The proportion of cells that underwent at least one endocycle is simply the proportion of cells with > 2C level (given as a percentage). Again, it is a modification of the original parameter %E (proportion of > 3C cells) proposed by Dilkes et al. (2002) for endosperm using the equation:

$$\ge 4\text{C}=100\times \frac{(\text{n}4\text{C} + \text{n}8\text{C} + \text{n}16\text{C}\dots )}{(\text{n}2\text{C} + \text{n}4\text{C} + \text{n}8\text{C} + \text{n}16\text{C}\dots )}$$

In addition, the maximal number of endocycles in a sample that occurred during endopolyploidisation, as inferred from FCM histograms, is reported. In this study, when at least three peaks were recorded on a histogram, the maximal number of endocycles corresponded to n–1, where n is the maximal recorded number of peaks. However, when only two peaks of 2C and 4C nuclei were present, one endocycle was registered only if 4C nuclei exceeded 10 % in an organ that was fully developed, and mitotic activity was excluded or at minimal level.

Principal component analysis (PCA, performed as described in Kocová et al. 2017) and exploratory data analysis were performed to get an overview of the variation in ploidy classes and endopolyploidy parameters (mean and standard deviation are presented). Analysis of numerical variance (ANOVA) was used to test for significant differences between means of different species. Correlation analysis (Pearson correlation coefficient) was applied to test for relationships among various endopolyploidy indices and ploidy classes and between our data and published data (Barow and Meister 2003) for N. tabacum. Data analyses and graphs were produced using packages gplots ver. 3.0.1.1 (Warnes et al. 2019), ggplot2 ver. 2.2.1 (Wickham 2009), PerformanceAnalytics ver. 1.5.3 (Peterson and Carl 2019) and vegan ver. 2.5-4 (Oksanen et al. 2019) in R ver. 3.5.3 environment (R Core Team 2019).

Results

Genome size differences among the six species were statistically significant (ANOVA, F5, 17 = 2764, p < 0.001) and separated into five different groups, where C. annuum and C. chinense had an almost identical genome size and formed one group (Fig. 1). Genome size varied among Capsicum species from 6.99 pg in C. chinense to 7.41 in C. baccatum and among Nicotiana species from 5.51 pg in N. sylvestris to 10.43 pg in N. rustica.

Fig. 1
figure 1

Genome size variation (2C value) for investigated species of Capsicum and Nicotiana

Full size image

The presence of 2C – 16C nuclei (Figs. 2a and 3a; Table 1) were measured using flow cytometry in all the species investigated irrespective of genus assignment. Furthermore, 32C nuclei were recorded in two samples of calyces of N. rustica (0.19 and 0.27 % of 32C nuclei). 2C nuclei most frequently occurred in leaf lamina and corolla tissue (94.63 % on average in C. annuum corolla tissue) and less frequently in root tissues (21.64 % in C. baccatum root) and specifically the stamen tissue of N. tabacum (37.25 %). 2C nuclei frequency had a strong negative correlation with 4C nuclei frequency. This leads to observations such as 4C nuclei most frequently occurring in root tissues (72.49 % in C. baccatum root) and specifically the stamen tissue of N. tabacum and less frequently in corolla or leaf lamina tissues (5.37 % in C. annuum corolla tissue). Other ploidy classes were observed less frequently. Moreover, 8C was the highest in petiole tissue of N. tabacum (13.01 % on average), whereas 16C nuclei in petiole tissue of N. sylvestris were only 2.07 % on average.

Fig. 2
figure 2

Endopolyploidy analysis among samples of vegetative organs for six investigated species of Capsicum and Nicotiana. a Proportion of nuclei classes and endoreduplication index. Dashed line represents trashold value of EI = 0.1 (tissues with EI < 0.1 are not considered to be endopolyploid). NA, not applied. b PCA analysis of EI. Statistically homogeneous groups are denoted by lowercase letters

Full size image
Fig. 3
figure 3

Endopolyploidy analysis among samples of floral organs for four investigated species of Capsicum and Nicotiana. Endopolyploidy level in four species with details for corolla parts and fertile floral organs in N. tabacum. a Proportion of nuclei classes and endoreduplication index in sterile floral organs, note that corolla tube and corolla limb were analysed separately for N. tabacum; for C. chinense, only n = 2 flowers were analysed and only mean value (without error bars) is plotted. NA, not applied. b Variation in endoreduplication index of fertile floral organs, pistil and stamen of N. tabacum. Dashed line represents trashold value of EI = 0.1 (tissues with EI < 0.1 are not considered to be endopolyploid). c Representative histograms for corolla tube and corolla limb of N. tabacum

Full size image

Parameters EI, MCV and ≥4C were highly correlated (r > 0.90), which indicates that they likely reflect similar features of endopolyploidy. However, E4P was less correlated to the other three parameters (r = 0.31, 0.44 and 0.56, respectively). Further direct comparison of EI vs. E4P using biplot analysis allowed for the identification of a group of samples that differed in either EI or E4P. For instance, the E4P parameter emphasises number of endoreduplicated nuclei, also known as mean ploidy of endoreduplicated nuclei (Dilkes et al. 2002), and was able to differentiate between highly similar standard EI parameter samples, e.g. the petiole tissue of N. sylvestris from root tissue of C. annuum (Fig. 4).

Fig. 4
figure 4

Relationship between endoreduplication index and E4P parameter. a Biplot of per organ means for six species of EI vs. E4P parameters. b Examples of flow cytometry histograms for two cases that are identical in EI but differ in E4P parameters

Full size image

The standard parameter EI, which is preferred in comparative endopolyploidy studies, always reached < 1.0, e.g. a maximum of 0.96 for leaf petiole tissue of N. sylvestris and 0.84 on average for C. baccatum root (Table 1). The lowest values (EI < 0.10), which were commonly found in non-polysomatic organs such as the leaf lamina and corolla of Capsicum species with some exceptions (Figs. 2a and 3a).

Root, leaf petiole and leaf lamina vegetative organs were available for all the studied species and allowed for analyses to at least partially differentiate the species (Fig. 2). Statistically significant differences (at least p < 0.05) in EI values were identified among species for all investigated vegetative organs, including sterile floral organs (Figs. 2a and 3a); root (ANOVA, F5, 38 = 15.78, p < 0.001), stem (ANOVA, F4, 32 = 3.31, p < 0.05), leaf petiole (ANOVA, F5, 38 = 6.94, p < 0.001), leaf lamina (ANOVA, F5, 38 = 39.15, p < 0.001), pedicel (ANOVA, F2, 17 = 24.56, p < 0.001), calyx (ANOVA, F2, 17 = 7.59, p < 0.01) and corolla (ANOVA, F3, 21 = 89.01, p < 0.001). In corolla tissue, we determined that there was EI variability for both the corolla tube and limb in N. tabacum (Fig. 3a and 3c). Interestingly, there was no consistent pattern in EI values for specific organs; however, C. baccatum and N. sylvestris were clearly different from their congeners (Fig. 2b), as revealed in PCA analysis, which retrieved first two PC axes accounting together for almost 89 % of the total variation. A significant difference was also found between two fertile floral organs of N. tabacum, the stamen and pistil (t-test, t = − 9.37, df = 5.31, p < 0.001, Fig. 3b). Unfortunately, reproductive organs were unavailable for some species, limiting the extent of the conclusions that can be made with respect to species similarity in reproductive organs.

In comparison to other studies, correlation analysis revealed that there was a strong association between our data and data obtained by Barow and Meister (2003) for N. tabacum (r = 0.97, p < 0.01, n = 6), which increases our confidence in these findings and the conclusions drawn from the present data.

Discussion

Endopolyploidy is widespread among angiosperms (D’Amato 1984; Barow and Meister 2003; Bainard et al. 2012); it plays a prominent role in developmental processes (Kudo and Kimura 2001; Barow and Meister 2003; Kocová et al. 2014; Rewers and Sliwinska 2014; Straková et al. 2014; Skaptsov et al. 2017). It is commonly found in at least one cell or tissue type, usually with a specialised function, e.g. suspensor, tapetum, endosperm or leaf bladder cells (Nagl 1976; D’Amato 1984; Barow and Meister 2003). However, in addition to these specialised cells and tissues, endopolyploidy commonly occurs in various plant organs, particularly in vegetative organs, such as roots, stems and leaves. It has only been consistently documented in a few families that may be considered ‘polysomatic families’. Endopolyploidy in such families occurs at a high level, which means that 8C nuclei are commonly present, and the proportion of 2C nuclei is lower than that of polyploid nuclei. Many economically important plant families, such as the Brassicaceae, Fabaceae, Amaranthaceae and Cucurbitaceae, may be referred to as polysomatic families. Due to the impact that endopolyploidy has on plant physiology and crop plants, it is important to increase the extent of our knowledge of the overall presence of endopolyploidy among plants.

Endopolyploidy in Solanaceae

The Solanaceae family contains several economically important species. Solanaceous species are distributed worldwide and are widely known, e.g. potato (Solanum tuberosum L.), tomato (Solanum lycopersicum L.), bell pepper (C. annuum), eggplant (Solanum melongena L.) and tobacco (N. tabacum). Barow and Meister (2003) reported the Solanaceae as a polysomatic plant family, which was later confirmed in several studies. For instance, pericarp tissue of tomato is now known to be highly polysomatic, containing cells of ploidy levels up to 512C (Cheniclet et al. 2005; Chevalier et al. 2014). In addition, tubers of S. tuberosum and other related species were identified as highly polysomatic, reaching a maximum 64C in S. candolleanum Berthault (EI = 2.62, Laimbeer et al. 2017). It has also been found that mature leaf tissue of both diploid and tetraploid plants of Hyoscyamus niger L. is endopolyploid (EI ~ 0.4, Weber et al. 2008).

Only a few studies focusing on the endopolyploidisation of Capsicum and Nicotiana have been published. A study by Gilissen et al. (1996) reported the presence of endopolyploidy in stem cells of N. tabacum. Barow and Meister (2003) showed endoreduplication (EI values above 0.1) in four and eight organ types of Capsicum frutescens L. and N. tabacum, respectively. Ogawa et al. (2010) focused on revealing the degree of polysomaty in fruit tissues of Capsicum species, and they showed very high levels of endopolyploidisation in the pericarp tissue of C. annuum cv. ‘Édes alma’, reaching a maximum of 256C.

Several studies have identified endopolyploidy in various tissues of seeds of C. annuum during development (Lanteri et al. 1993, Lanteri et al. 1994, Portis et al. 1999, Rewers and Sliwinska 2014). In the developing embryo, cells with 4C ploidy levels were present in considerable amounts, while the mature embryo in dry seed possessed almost only 2C cells (4C < 10 %). Endosperm cells during development, as well as endosperm cells in mature seed, frequently had 6C ploidy level.

In this study, endopolyploidy patterns were evaluated across a sample of the Solanaceae. Endopolyploidy was analysed in six species: C. annuum, C. baccatum, C. chinense, N. rustica, N. sylvestris and N. tabacum. Our data support the regular inclusion of Solanaceae on the list of ‘polysomatic families’. The level of endopolyploidy appears to be moderate with a mean EI per organ always < 1.0.

This comprehensive study addressing endopolyploidy patterns and variation across several organ tissues of Capsicum and Nicotiana revealed previously unknown tissue-specific information. In the genus Capsicum, the lowest level of endopolyploidy in cells occurred in leaf lamina and corolla vegetative organs, while root tissues were identified as the vegetative organ with the highest EL.

Our results regarding EL in leaf tissue are consistent with previous studies wherein leaf lamina likely has minimal endopolyploid cells or a very low amount of 4C endopolyploid cells, while the leaf petiole is polysomatic (Lanteri et al. 1993, 1996; Barow and Meister 2003). We have recorded up to 8C nuclei in C. baccatum, and Barow and Meister (2003) reported a max EI of 0.13 in C. frutescens. In leaf petiole tissue, our results indicate a mean EI from 0.32 to 0.54 (depending on species), which is similar to previous reports (max EI = 0.51) for C. frutescens leaf tissue by Barow and Meister (2003).

With regard to species similarity, C. annuum and C. chinense showed nearly identical patterns, with the exception of sterile floral tissue. However, due to the low sample size in C. chinense, these results are not particularly conclusive. Notably, C. baccatum slightly deviated in EL from both C. annuum and C. chinense, but a significant differentiation in EL was only recorded in root tissue. These findings may stem from a difference in genetic relatedness (Carrizo García et al. 2016) and/or a larger genome size of C. baccatum (Fig. 1).

To date, EL reports for several organs have only been published for N. tabacum (Barow and Meister 2003), and these are highly congruent with our data. In this study, we have expanded these findings to include two novel Nicotiana species. In the genus Nicotiana, we found that the lowest EL occurred in leaf lamina and corolla tissue for all species, whereas the highest EL occurred in leaf petiole or root tissue depending on the species. In addition, Gilissen et al. (1996) reported that the EL for stem tissue of Nicotiana (N. tabacum) was 2C – 8C cells, which is consistent with this study. Such congruence among studies highlights the precision of FCM determination of EL. Furthermore, it opens the possibility for future comparisons of data across several different studies.

In N. tabacum, we identified that the EI was significantly lower in the corolla limb than in the corolla tube tissue. Despite only being demonstrated in one of six investigated species, this could still be similar to some EL patterns in other polysomatic species of Nicotiana. Furthermore, the EI might be significantly lower as a result of corolla tube elongation, which is often a result not only of cell proliferation (increasing number of cells) but also of cell size expansion. This growth model has been elucidated in Petunia Juss. of the Solanaceae (Stuurman et al. 2004) and in Lithospermum L. of the related family Boraginaceae (Cohen et al. 2012). Generally, cell elongation accompanied by endopolyploidisation is an important component of petal development in polysomatic plants (e.g. Kudo and Kimura 2002; Agulló-Antón et al. 2013; Ho et al. 2016). A cytohistologically oriented study coupled with FCM screening of several taxa should be performed to test this assumption in Nicotiana.

Reproductive organs, e.g. pistils and stamens, contain much more diverse tissues than vegetative organs, e.g. leaves, calices or corollas. However, most of the nuclei of the pistil were released from style tissues in Nicotiana and were found to be 2C – 8C. Certainly, ~ 25 % of the pistil cells had undergone one endocycle. Conversely, stamens also contained 2C – 8C cells, which indicated that at least a few cells were endopolyploid. More than 60 % of all stamen cells were found to be 4C, which resulted in a high EI value (mean EI = 0.63). However, this finding should not be interpreted as the presence of a high amount of endopolyploid cells in the case of an active tissue, such as developing sporogenous tissue or microspores inside the anther of stamen. Some of 2C cells in the stamen were assumed to be generative cells of microspores (pollen grains), which are cells arrested in the G2 phase of the cell cycle prior to post-pollination mitosis and formation of two sperm cells (D’Amato et al. 1965; Kron and Husband 2012). Thus, regarding EL determination, a lack of FCM differentiation between endopolyploid cells and cells arrested in the G2 phase of the cell cycle requires the differentiation between mitotically active and inactive tissues and careful interpretation of FCM data (Rewers and Sliwinska 2014).

Overall, the EL pattern in Nicotiana reveals that N. sylvestris possesses a higher EL than N. rustica or N. tabacum. This relationship likely reflects the diploid chromosome constitution of N. sylvestris, whereas both other species are tetraploids. The phylogenetic relatedness does not likely play a role in this case since both tetraploids, N. tabacum and N. rustica, have different origins, and the diploid N. sylvestris is the originating maternal species for N. tabacum only (Clarkson et al. 2017).

It remains unclear why polyploids tend to lose EL compared to diploids. Overall, there has been little research addressing this topic. However, Nagl (1976) hypothesised that endopolyploidy may be an evolutionary novelty appearing at a tissue and cellular level, which compensates for a low whole organism ploidy level. This hypothesis appears to be supported by tens of well-documented diploids and their synthetic (isogenic) autotetraploids in the model species Arabidopsis thaliana (Pacey et al. 2020a). A similar compensatory effect for small organism genome size was also suggested for other plant groups, e.g. geophytes (Kolarčik et al. 2020). However, EL is genotype specific, and observing endopolyploidy diversification among populations is likely an evolutionary response to natural selection driven by environmental differentiation (Pacey et al. 2020b).

Comparative exploratory analysis of endopolyploid level parameters

Various equations to calculate a value characterising EL from data on frequencies of different ploidy classes discerned from FCM histograms have been proposed and are inconsistently applied in endopolyploidy research. This is an obstacle in any comparative study. To date, however, no attempts have been made to investigate the significance of these equations or to propose which of them should be recommended for continued use. Different FCM profiles of nuclei ploidy frequency of two samples may result in the calculation of the same value for some of the parameters that have been proposed to express EL, e.g. EI or MCV. Thus, comparative inspection of these parameters between each other, as well as to frequency of 2C, 4C, etc., should be performed. Some of the EL parameters capture different aspects of EL. For instance, EI is used to estimate proportion of endocycles, while E4P is used to evaluate mean ploidy level of only endopolyploid cells.

In this study, we have shown that a simple reliance on a single parameter as performed in some important studies (Barow and Meister 2003; Jovtchev et al. 2007; Bainard et al. 2012) may lead to overlooking some of the real differences between samples, which is consistent with the findings of Kobayashi (2019). Moreover, we have shown that a low correlation was found between the E4P parameter and other parameters (EI, MCV and ≥4C). Identical values of EI may result from a high proportion of endopolyploid cells with one endocycle compared to a proportion of basic 2C nuclei and from a lower proportion of nuclei with one (4C) or more (8C, 16C) endocycles (Fig. 4). Therefore, a comparison between EI and E4P parameters can be very informative and allows for the identification of causes for organ or species differences in EL, such as varying proportions of 2C and various endopolyploid 4C – 32C cells. Therefore, common parameters and standardised analyses of all existing EL parameters in any study of endopolyploidy would be of great value and should be a goal in future studies.

Conclusions

Endopolyploidy plays a significant role in the development, differentiation and stress physiology of polysomatic plants. However, the current knowledge of endopolyploidy is lacking for most plant species, which is a considerable shortcoming in the interpretation of physiological or transcriptomic data. Surprisingly, ploidy surveys at cellular or tissue levels are rarely performed in such studies. For instance, Pirrello et al. (2018) suggested that a complete approach, which should be adopted, includes flow cytometry screening of endopolyploidy and nuclei ploidy classes sorting prior to transcriptome analysis. In this study, we provide foundational endopolyploidy information that will be instrumental in future robust transcriptomic analyses for solanaceous crops.

Furthermore, some studies have shown that endopolyploidy may be used in modern agriculture and plant breeding programmes for crop improvement. For instance, it is a reliable marker of seed development in many polysomatic species (Sliwinska 2009) and has also been shown to be indicative of crop quality (Kobayashi 2019). Endoreduplication also leads to doubling or multiplying of chromosomes in certain cell lines, which can be useful in the production of polyploid plants from tissue cultures or in vitro regeneration (Cheniclet et al. 2005; Chen et al. 2011). More RNA transcripts are produced through endoreduplication of DNA, and therefore, metabolite production increases (Lee et al. 2009; Scholes and Paige 2015). In this manner, it may be possible to either produce more bioactive metabolites in medicinal plants (Sliwinska 2018) or produce more metabolites for commercial use, such as increasing nicotine in N. tabacum or N. rustica or producing cytotoxic compounds and anti-herbivory toxins in plants that can be used as biopesticides (Leitch and Dodsworth 2017). In this study, we were able to identify tissues in Nicotiana with varying levels of endopolyploidy that, with further research, may be selected in the future to improve some of these desirable traits.