Histograms from flow cytometry of bougainvillea samples. The horizontal axis is exported fluorescence values representing peak positions, and the vertical axis is the number of nuclei. (Upper row) Tender leaves treated with different lysates: (A) woody plant buffer (WPB), (B) Galbraith’s buffer, and (C) Sysmex Partec Cystain ultraviolet Precise P. (Lower row) different tissues treated with WPB: (D) young leaves, (E) bracts, and (F) mature leaves.
DNA ploidy estimation by flow cytometry analysis for (A) Nicotiana benthamiana (the reference standard), (B) ‘Miss Manila’ (2n = 2x), (C) ‘Chitra Varigata’ (2n = 3x), (D) ‘Wajid Ali Shah’ (2n = 3x), (E) ‘Super Miss Manila’ (2n = 2x/4x), and (F) ‘Chitra’ (2n = 4x).
Statistical results of the ploidy number of different varieties.
Microphotographic observations of chromosomes in the root tip cells. (A) Bougainvillea ×buttiana ‘Miss Manila’ (2n = 2x = 34), (B) Bougainvillea ×buttiana ‘Formosa’ (2n = 2x = 34), (C) Bougainvillea ×buttiana ‘Wajid Ali Shah’ (2n = 3x = 51), and (D) Bougainvillea ×buttiana ‘Chitra’ (2n = 4x = 68). Different chromosomes from the mixoploid variety of Bougainvillea ×buttiana ‘Super Miss Manila’: (E) 2n = 4x = 68 and (F) 2n = 2x = 34. Bar = 100 μm.
Schematic diagram of 1C genome size from different bougainvillea groups.
Click on author name to view affiliation information
Bougainvillea Comm. ex Juss. (Nyctaginaceae; Bougainvillea) is a popular ornamental plant with vigorous growth, luxuriant blooming, colorful bracts, and a high tolerance to the stresses of temperature, drought, and soil pollution, and thus is widely cultivated in tropical and subtropical regions. However, the paucity of information on ploidy and the genomic constitution is a significant challenge to genome research and cultivar improvement. We present a flow cytometry method for ploidy detection in bougainvillea based on evaluating different lysates and tissues, identify the ploidy level of a batch of bougainvillea accessions, and infer the genome size of horticultural species Bougainvillea glabra, Bougainvillea spectobilis, and Bougainvillea peruviana. The results show that tender leaves and woody plant buffer (WPB) were optimal for flow cytometry analysis. The 2C nuclear DNA amounts in 176 bougainvillea accessions ranged from 4.66 ± 0.04 to 12.26 ± 0.1 pg, which represents 161 diploids, 13 triploids, 1 tetraploid, and 1 di-tetraploid mixoploid. For B. glabra, B. spectobilis, and B. peruviana, the mean 1C values were 3.201, 3.066, and 2.915 pg, respectively. The genome size of B. glabra was significantly larger than that of B. peruviana (P = 0.0004), but had no significant variation with that of B. spectobilis (P = 0.1061). These results reveal fundamental cytogenetic information for bougainvillea that are beneficial to whole-genome sequencing and hybrid breeding programs.
Bougainvillea Comm. ex Juss., belonging to Nyctaginaceae family, is one of the most popular garden ornamental plants, and is widely cultivated in tropical and subtropical regions of the world because of its properties of appealing bracts, long flowering period, mass blooming, and varied adaptability to various soil and climatic conditions (Kulshreshtha et al. 2009). Bougainvillea is also explored regarding medicinal uses for different therapeutic purpose in Panama, India, and Thailand (Abarca-Vargas and Petricevich 2018). Bougainvillea varieties with significant horticultural value are mainly domesticated from B. glabra Choisy, B. spectabilis Willdenow, and B. peruviana Humboldt & Bonpland, all of the three elemental species possessing greater ornamental value and were diploidy (2n = 34) (Ohri 2013; Zadoo et al. 1975). Although bougainvillea have a 150-year history of domestication and cultivation outside their natural habitat, less than 500 cultivated varieties are recorded with the International Bougainvillea Registration Authority. Many more varieties with various flower colors, bract types, flowering periods, and stress resistance are always needed for family potting and public landscaping.
Ploidy is an important factor for plant breeding, which influences cross compatibility, fertility, and the phenotype of the offspring (Adams and Wendel 2005; Touchell et al. 2020). Intraspecific and interspecific hybridization of Bougainvillea has been carried out extensively during past decades and has produced various horticulture cultivars. Most of them can be divided into three groups of B. glabra, B. peruviana, B. spectabilis, and their interspecific groups (B. ×buttiana or B. ×glabra-peruviana, B. ×specto-peruviana, and B. ×specto-glabra). However, few cultivars (<100) cultivated were generated through cross breeding, because the offspring, especially those derived from interspecific hybridization, are mostly sterile or partly sterile, which poses a bottleneck for further improvement (Datta 2021). Fortunately, the fertility in these hybrids cam be restored by induced polyploidy, and this expanded germplasm was used in crossing to produce novel hybrids with highly positive horticultural traits (Datta 2021; Sindhu et al. 2020). So, understanding the ploidy levels and genome size of Bougainvillea is very essential to design effective breeding strategies for genetic improvement. Although cytological work had been done and the ploidy level of some varieties of bougainvillea have been reported (Ohri and Khoshoo 1982; Zadoo et al. 1976), the information has been limited because of the number of samples and the lack of varieties developed in recent decades. Chromosome counting in the root tip is a common approach for identifying ploidy level, which has been applied in numerous species. However, this method is technically demanding, time-consuming, and laborious. Currently, flow cytometry (FCM) analysis has been widely used in estimating ploidy level and genome size because it is simpler and more efficient, which make it suitable for analyzing numerous samples (Bourge et al. 2018; Sliwinska 2018). The detection of plant ploidy and genome size with FCM is based on the comparison of DNA contents. Under the excitation of a high-pressure mercury lamp, fluorescence of a certain wavelength will be produced. The DNA content in a single cell is directly proportional to fluorescence intensity. It can be displayed in the form of peaks through FCM analysis. With the known chromosome ploidy level of a plant as a control, the chromosome ploidy of the plant to be tested can be determined by comparing the fluorescence values. A software package (Partec Gmb H, Münster, Germany) was used for calculating the mean fluorescence values of the peak and its coefficient of variation (CV).
Considering that high-throughput ploidy detection technology and cytological information of bougainvillea germplasms are still lacking in breeding practice, our aim is to establish a reliable and simple FCM analysis method for bougainvillea, and to determine the ploidy levels, genome sizes, and their variation among bougainvillea accessions with different genetic backgrounds.
A total of 176 accessions analyzed (Supplemental Table S1) were planted at the germplasm nursery of the Tropical Crop Genetic Resource Institute, Chinese Academy of Tropical Agricultural Science, Danzhou, China (lat. 19°35’N; long. 109°42’E). These accessions were obtained from germplasm collections worldwide, representing three elementary species and their interspecific hybrids cultivated widely. Those with ambiguity of genetic origin were classified into Bougainvilleasp. groups.
To establish the optimal FCM system for bougainvillea, different lysis solutions and plant tissues were evaluated first. The lysates include WPB, Sysmex Partec Cystain ultraviolet Precise P, and Galbraith’s buffer, which are tested on the young leaves of the known ploidy B. glabra ‘Formosa’. About 0.1 g of leaves were chopped in a petri dish with 1 mL lysis solution then filtered through a 50-μm mesh filter and stained with 50 μL 4, 6-diamino-2-phenylinode solution for 5 min. The CV of the peak cell DNA content as an index to screen the most suitable lysate was analyzed using Sysmex CyFlow Ploidy Analyzer (Partec GmbH, Münster, Germany). According to the FCM analysis results, the most suitable lysis solution was used to treat the tender leaves, bracts, and mature leaves of ‘Formosa’ to determine the sample tissues. Last, the optimal FCM method was used to evaluate ploidy levels of 176 bougainvillea varieties. A total of 3000 nuclei were analyzed and three independent replications were conducted for each tested sample. The average CV values closer to 5% were considered to be reliable results.
Nicotiana benthamiana, with a DNA content of 3.2 pg/C and an estimated genome size of 3.0 Gb (Goodin et al. 2008), was selected to be the reference standard in ploidy determination. By comparing the average florescence peaks of the tested sample with the reference standard, the experimental genome size and 2C value of each accessions were calculated using the following equations: and
Chromosome counting was performed on the varieties representing the diploids, triploids, tetraploids, and mixoploids using microscopy. The method for preparing root tip chromosome preparation is as follows: the young root tips, ∼0.5 to 1.0 cm in length, were sampled and cleaned, and then incubated with saturated p-dichlorobenzene for 6 h at room temperature in dark conditions. After rinsing with double distilled water, they were transferred into a freshly prepared Carnoy’s solution of absolute ethanol: glacial acetic acid (3:1, v/v) at 4 °C for 24 h. Next, they were placed in 1 mol/L hydrogen chloride at 60 °C for 10 min. Stained with modified carbol-fuchsin solution on a slide, the tissues were squashed with a coverslip and viewed using microscopy (Leica DM 2500, Wetzlar, Germany). Three root tips per genotype were treated, and at least three cells per root tip showing easily distinguishable chromosome were selected to count the numbers of chromosomes. The chromosome images were captured with a digital camera (DS-L1; Nikon, Tokyo, Japan).
The obtained FCM data were analyzed and flow histograms of tested samples were generated using FCS Express version 3. The ploidy levels and DNA contents were calculated using Excel version 2010 (Microsoft Corp., Redmond, WA). One-way analysis of variance (ANOVA) and Duncan’s multiple range test were performed using SPSS version 21 (SPSS Inc., Chicago, IL). The statistical result charts were drawn by using Origin version 2021.
FCM histograms show all the cell nucleus suspension extracted from three different lysates produced symmetric peaks with negligible background. However, the CV values of the WPB samples (2.90% ± 0.24%) were significantly less than those of Sysmex Partec Cystain ultraviolet Precise P (4.89% ± 0.80%, P = 0.0023) and Galbraith’s buffer (5.84% ± 0.28%, P = 0.0004) (Table 1). Moreover, the relative fluorescence intensity (RFI) peak from the WPB samples was significantly greater than those of the other two lysates (P = 0.0007 and P = 0.0005) (Fig. 1; Table 1). So, with the use of WPB buffer, more intact nuclei could be isolated successfully from bougainvillea cells, and were then selected to evaluate the various tissues of bracts, young leaves, and mature leaves with FCM.
Citation: HortScience 57, 12; 10.21273/HORTSCI16862-22
The histogram results show that both tender leaves and bracts exhibited a single peak, whereas mature leaves produced a state of miscellaneous peaks. The CV value of tender leaves (2.90% ± 0.24%) was significantly less than those of mature leaves (4.17% ± 0.42%, P = 0.0003), but there was no significant difference with bracts (3.06% ± 0.268%, P = 0.583) (Table 2). The RFI values of tender leaves were significantly greater than those of bracts (P = 0.0174) and mature leaves (P = 0.0000) (Fig. 1; Table 2). Based on our analysis, tender leaves are the optimal sampling tissue materials for FCM analysis of Bougainvillea (Fig. 2).
Citation: HortScience 57, 12; 10.21273/HORTSCI16862-22
The RFI values of germplasm samples varied greatly, ranging from 7793 ± 424 (B. peruviana ‘Thimma Special’) to 22,772 ± 465 (B. ×buttiana ‘Chitra’). For the known diploid ‘Fomosa’ (2n = 2x = 34), triploid ‘Wajid Ali Shah’ (2n = 3x = 51), and tetraploid ‘Chitra’ (2n = 4x = 68), the RFI values were 12,897 ± 52, 17,168 ± 636, and 22,772 ± 465, respectively (Supplemental Table S1). By comparing the fluorescence intensity of DNA contents with that of the known varieties, 161 diploids (91%), 13 triploids (7%), 1 tetraploid (1%), and 1 di-tetraploid mixoploid (1%) were detected in 176 accessions (Fig. 3; Supplemental Table S1). The fluorescence intensity of the nuclear DNA content of ‘Super Miss Manila’ has two peaks, which are at 12,500 and 25,000, respectively. This indicates that it was a mixoploid, including a diploid and a tetraploid (Supplemental Table S1).
Citation: HortScience 57, 12; 10.21273/HORTSCI16862-22
To confirm the ploidy determined with FCM, root squashes for chromosome counts were carried out on 4 diploid, 13 triploid, 1 tetraploid, and 1 mixoploid accession. Root squashes confirmed 34, 51, and 68 chromosomes for diploids, triploids, and tetraploids, respectively (Fig. 4). Regarding the di-tetraploid mixoploid ‘Super Miss Manila’, both karyotypes (including 34 and 68 chromosomes) were detected in different nuclei (Fig. 4E and F). No other euploid or aneuploid was found. The results of chromosome counting in root tip cells showed that the ploidy values assigned by FCM reflected the true ploidy levels.
Citation: HortScience 57, 12; 10.21273/HORTSCI16862-22
Estimated C values and calculated genome sizes are summarized in Supplemental Table S1 along with their assigned groups. Each species group and their interspecific groups showed a significant range of genome sizes (Fig. 5). The mean 1C values are 3.20 ± 0.13 pg in B. glabra, 3.07 ± 0.21 pg in B. spectabilis, and 2.92 ± 0.37 pg in B. peruviana, and are equivalent to genome sized of 3.06 Gb, 2.93 Gb, and 2.76 Gb, respectively (Fig. 5). The results of the one-way ANOVA showed there are significant differences between B. glabra and B. peruviana (P = 0.0004), but no significant differences between B. glabra and B. spectabilis (P = 0.1061). The average 1C values of the interspecific groups are generally between the average values of their two donor species.
Citation: HortScience 57, 12; 10.21273/HORTSCI16862-22
Although FCM has been widely used for ploidy analysis in many plants for decades, and various nuclei isolation buffers have been applied, none has worked well with all plant species because of the diversity in tissue type and genome size. So, successful preparation of a nuclear suspension is the first essential step in FCM for a new plant species. In our study, when compared with Sysmex Partec Cystain ultraviolet Precise P and Galbraith’s buffer, the use of WPB resulted in improved histogram quality with a minimal CV, symmetric peaks, and negligible background. As expected, the tissue of tender leaves resulted in lower CV and higher RFI values relative to bracts and mature leaves, which may be attributed to their differences in the content of some cytosolic compounds, such as phenolics, tannins, glycosides, terpenoids, alkaloids, and flavonoids, which have a negative effect on nuclei isolation, DNA stability, and stoichiometry. Therefore, WPB and tender leaves were considered the optimum combination and were subjected to FCM in Bougainvillea varieties. The results of FCM and chromosome counting were consistent, indicating the reliability of this method in ploidy identification.
It was reported previously that the 2C genome sizes of B. glabra, B. spectabilis, and B. peruviana were 7.0 pg, 8.91 pg, and 8.25 pg, respectively (Ohri and Khoshoo 1982). However, in our study, the genome size variation was common in intraspecific and interspecific groups, with mean 2C values of 6.40 pg (B. glabra), 6.14 pg (B. spectabilis), and 5.84 pg (B. peruviana). The genome size of B. glabra is significantly larger than that of B. peruviana, but not significantly different from B. spectabilis. This result supports the recent phylogenetic viewpoint that there is a close relationship between B. glabra and B. spectabilis, but B. peruviana is relatively far from them in terms of an evolutionary relationship (Bautista et al. 2020, 2022). These differences imply there may be chromosomal structural variations, such as chromosome rearrangements, sequence deletions or insertions, tandem repeats and transposable elements, or chromosomal number variations of polyploidization aneuploidy, as reported previously (Khoshoo and Zadoo 1969).
The comprehensive breeding program for bougainvillea is hampered by widespread pollen and/or seed sterility. Past studies on breeding of bougainvillea showed that tetraploids had high fertility and cross compatibility. However, among the germplasms tested in our study, only one famous tetraploid germplasm, ‘Chitra’, was confirmed, which was derived from a cross between two induced tetraploids (B. ×buttiana ‘Tetra Mrs. McClean’ × B. peruviana ‘Dr. B.P. Pal’). Ninety-one percent of the tested samples are diploid, especially the interspecific hybrids with high sterility, which implies that it is necessary to induce more polyploids to promote breeding projects.
In addition to the known triploids ‘Wajid Ali Shah’, ‘Dr. H. B. Singh’, and ‘Begum Sikander’, which derived from artificial crosses of diploids and tetraploids (Datta 2021, 2022), other triploids ‘Jamaica White’, ‘San Diego Red’, ‘Duck Feet’, ‘Rainbow’, ‘Chitra Variegata’, ‘Super Marie’, ‘Tang Porcelain’, ‘Pink Angel’, and ‘Purple Butterfly’ were confirmed for the first time. These autonomous triploids may result from bud sport or natural hybridization. Because many hybrids have been crossed over several generations, and natural mutations occur spontaneously throughout the world, it is difficult to identify their respective origins. Polyploidization has enriched the genetic and phenotypic diversity of bougainvillea resources. Although a few triploid varieties have been reported to carry out hybridization successfully (Khoshoo and Zadoo 1969), the crossing potential of these newly identified triploids in bougainvillea breeding still needs to be evaluated, because the high sterility of triploids is a common phenomenon in the plant kingdom.
Ploidy detection for bougainvillea by FCM was explored and optimized, and 176 Bougainvillea varieties were evaluated, which consisted of 161 diploids, 13 triploids, 1 tetraploid, and 1 di-tetraploid mixoploid. Moreover, the genome size of bougainvillea and its variation were also revealed. This work lays a foundation for the research of cytology, genomics, and hybrid breeding in bougainvillea.
Histograms from flow cytometry of bougainvillea samples. The horizontal axis is exported fluorescence values representing peak positions, and the vertical axis is the number of nuclei. (Upper row) Tender leaves treated with different lysates: (A) woody plant buffer (WPB), (B) Galbraith’s buffer, and (C) Sysmex Partec Cystain ultraviolet Precise P. (Lower row) different tissues treated with WPB: (D) young leaves, (E) bracts, and (F) mature leaves.
DNA ploidy estimation by flow cytometry analysis for (A) Nicotiana benthamiana (the reference standard), (B) ‘Miss Manila’ (2n = 2x), (C) ‘Chitra Varigata’ (2n = 3x), (D) ‘Wajid Ali Shah’ (2n = 3x), (E) ‘Super Miss Manila’ (2n = 2x/4x), and (F) ‘Chitra’ (2n = 4x).
Statistical results of the ploidy number of different varieties.
Microphotographic observations of chromosomes in the root tip cells. (A) Bougainvillea ×buttiana ‘Miss Manila’ (2n = 2x = 34), (B) Bougainvillea ×buttiana ‘Formosa’ (2n = 2x = 34), (C) Bougainvillea ×buttiana ‘Wajid Ali Shah’ (2n = 3x = 51), and (D) Bougainvillea ×buttiana ‘Chitra’ (2n = 4x = 68). Different chromosomes from the mixoploid variety of Bougainvillea ×buttiana ‘Super Miss Manila’: (E) 2n = 4x = 68 and (F) 2n = 2x = 34. Bar = 100 μm.
Schematic diagram of 1C genome size from different bougainvillea groups.
Contributor Notes
This project was supported by National Tropical Plants Germplasm Resource Center, the Key Research and Development Projects of Hainan Province (ZDYF2022XDNY267, ZDYF2021XDNY169), and the Special Basic Research Fund for Nonprofit Central Public Research Institutes (1630032022004).
J.N. is the corresponding author. E-mail: niujunhai2014@sina.com.
Histograms from flow cytometry of bougainvillea samples. The horizontal axis is exported fluorescence values representing peak positions, and the vertical axis is the number of nuclei. (Upper row) Tender leaves treated with different lysates: (A) woody plant buffer (WPB), (B) Galbraith’s buffer, and (C) Sysmex Partec Cystain ultraviolet Precise P. (Lower row) different tissues treated with WPB: (D) young leaves, (E) bracts, and (F) mature leaves.
DNA ploidy estimation by flow cytometry analysis for (A) Nicotiana benthamiana (the reference standard), (B) ‘Miss Manila’ (2n = 2x), (C) ‘Chitra Varigata’ (2n = 3x), (D) ‘Wajid Ali Shah’ (2n = 3x), (E) ‘Super Miss Manila’ (2n = 2x/4x), and (F) ‘Chitra’ (2n = 4x).
Statistical results of the ploidy number of different varieties.
Microphotographic observations of chromosomes in the root tip cells. (A) Bougainvillea ×buttiana ‘Miss Manila’ (2n = 2x = 34), (B) Bougainvillea ×buttiana ‘Formosa’ (2n = 2x = 34), (C) Bougainvillea ×buttiana ‘Wajid Ali Shah’ (2n = 3x = 51), and (D) Bougainvillea ×buttiana ‘Chitra’ (2n = 4x = 68). Different chromosomes from the mixoploid variety of Bougainvillea ×buttiana ‘Super Miss Manila’: (E) 2n = 4x = 68 and (F) 2n = 2x = 34. Bar = 100 μm.
Schematic diagram of 1C genome size from different bougainvillea groups.
