The characteristics of gametogenesis may contribute to the partial female-fertility in aneuploid Lilium

Abstract

Unlike most aneuploid plants which are usually seedless, aneuploid lilies (Lilium), have partial female-fertility regardless of their male sterility. However, the underlying mechanisms remain unclear. ‘Chen Xi’, as an example of aneuploid lilies, was investigated for its gamete formation and fertility using modern cytogenetic methods. It had 44 chromosomes, consisting of 35 A-chromosomes and 9 L-chromosomes. Many irregular meiotic characters, such as trivalents, univalents, lagging chromosomes and micronuclei, were commonly observed. It produced functional eggs with variable chromosomes ranging from 19 to 48 among the analyzed nine progenies, indicating that it had partial female-fertility regardless of male sterility. However, the chromosome numbers in the secondary nuclei of embryo sacs produced by ‘Chen Xi’ were invariable twice those of its somatic cells. This was the first time to confirm the derived principle – Each aneuploid Lilium produces variable chromosomal eggs and invariable chromosomal secondary nuclei in its embryo sacs. The principle would be associated with the partial female-fertility regardless of male sterility of aneuploid lilies.

Key Message

Fluorescence in situ hybridization first confirmed the derived principle - Each aneuploid lily produces variably chromosomal eggs and invariably chromosomal secondary nuclei in its embryo sacs.

Introduction

Lilium spp. have giant genomes with 1 C values from 13.7 to 48.15 pg [1, 2]. They have been ideal plants to study chromosomes [3], and are becoming hotspots for DNA sequencing, genome assembly, etc [4–6]. They produce Fritillaria-type embryo sacs and large flowers, and they are also good plants to study sexual reproductive characters [7]. In addition, lilies are economically important flowers worldwide, and thus, it would be valuable to study their fertility and reproductive characters for breeding new varieties.

Embryo sacs, which are female gametophytes of angiosperms, are mainly classified into monosporic-, bisporic-, and tetrasporic-type [8]. Polygonum-type and Fritillaria-type sacs are most popular and representative types of monosporic- and tetrasporic-type embryo sacs respectively [9]. The essential difference between Polygonum-type and Fritillaria-type embryo sac is the ploidy levels or chromosome numbers of their secondary nuclei (Fig. 1). According to normal megasporogenesis [9], it is derived that the chromosome number of the secondary nucleus is twice that of the egg cell in each Polygonum-type embryo sac (Fig. 1A); however, the chromosome number of the secondary nucleus is twice that of the somatic cells in each Fritillaria-type embryo sac (Fig. 1B and C) [10–12].

Fig. 1.

Schematic comparison of megasporogeneses and secondary nuclei between different embryo sac plants. A In the Polygonum-type sacs of triploid plants, the amount of nuclear DNA in the secondary nucleus is twice that of the egg cell. B In Fritillaria-type embryo sacs of triploid lilies, the amount of nuclear DNA in the secondary nucleus is twice that of its original somatic cell, i.e., hexaploid (6x). C In any aneuploid lily, the amount of nuclear DNA in the secondary nucleus is twice that of its origin somatic cell

Numerous cases of 3x × 2x/4x hybridizations of Lilium indicate that these triploids could be female parents and produce aneuploid seeds though triploid lilies are male sterile [11–24]. The phenomenon is well explained with the derived principle - The chromosome number of the secondary nucleus is invariably hexaploidy in each embryo sac of triploid Lilium (Fig. 1B) [10–12].

It is challenging to confirm the derived principle. Flow cytometry does not work well for lily endosperm, though it is a fast method for Lilium leaves and tulip endosperm [25, 26]. Chromosome preparation is also hard because chromosomes of lily endosperm are so large and too many to be clearly counted. Fortunately, 5S rDNA can be an index for judging the genomes, as one genome of Lilium usually contains one chromosome bearing 5S rDNA loci [13, 27]. With this idea, the derived principle was finally confirmed for the first time [24].

Aneuploid lilies have a similar phenomenon to triploids regarding their fertility, and theoretically, the chromosome number of the secondary nucleus of each aneuploid lily is twice that of its somatic cells (Fig. 1C). However, it has not been confirmed with any aneuploid lily. To mainly prove the derived principle shown in Fig. 1C, an aneuploid lily ‘Chen Xi’ was used as an example, we investigated its microsporogenesis and pollen germination, made hybridization between aneuploid, diploid and tetraploid lilies, and analyzed their endosperm and progenies using modern cytogenetic techniques. Based on the results, the characteristics of aneuploid lily embryo sacs and reasons for success or failure in lily hybridizations were discussed.

Materials and methods

Plant materials

An aneuploid lily ‘Chen Xi’ (L*A*: 2n = 4x − 4 = 44), three Asiatic lilies, ‘Lollypop’ (AA: 2n = 2x = 24), Dan Feng’ and ‘Tresor’ (AAAA: 2n = 4x = 48), and a Longiflorum lily ‘White Fox’ (LL = 2x = 24) were the main materials. ‘Chen Xi’ and ‘Dan Feng’ were bred by our laboratory and registered by the Royal Horticultural Society (Surrey, UK). ‘Lollypop’, ‘Tresor’ and ‘White Fox’ were purchased from Hongyue Horticultural Co., Ltd. (Haining, Zhejiang, China). All plants were grown in a greenhouse under natural light with regular maintenance at Jiangxi Agricultural University, Nanchang, China.

Microsporogenesis

Microsporogenesis was observed according to Xiao et al. [28]. When the flower buds were about 22–26 mm, one anther was taken out and fixed in a solution containing three parts of ethanol and one part of acetic acid for at least 30 min, after that, a small part of anther was cut off, only pollen mother cells (PMCs) were put on the slide with a needle and mixed with 20 µl 2% Carbol Fuchsin (Phygene, Fuzhou, China) or 45% acetic acid (Xilong Scientific,‌Guangzhou, China), then, they were covered with a square coverslip. The slides prepared with Carbol Fuchsin were directly observed under a light microscope (ZEISS Scope. A1, Germany). The slides prepared with acetic acid were stored at -80 °C for about 1 h, then their covers were removed with a knife, and then immersed in pure ethanol for 2 min, air dried, and stored at -20 °C until used for GISH.

Pollen germination and pollination

Male fertility of all the materials was tested on a medium, containing 100 g l^− 1 sucrose (Xilong Scientific, Guangdong, China), 5 g l^− 1 bacteriological agar (Tianjin Damao Chemical Reagent Factory, Tianjin, China), 20 mg l^− 1 boric acid (Tianjin Damao Chemical Reagent Factory), 200 mg l^− 1 Ca(NO₃)₂ (Tianjin Damao Chemical Reagent Factory) and pH 5.8. The fresh pollen grains were scattered on the medium and cultured at 25 °C for about 2 h, and then observed under the microscope (ZEISS Scope. A1). Three fields under the microscope were viewed, and 300–400 pollen grains were counted for each cultivar. The rate of pollen germination (R) was equal to the number of germinated pollen grains (G) divided by total pollen grains (T), i.e., R = G/T. When the flowers of females were fully open, their stigmas were pollinated and wrapped with aluminum foil.

Chromosome preparation of endosperm

Approximately four to six weeks after pollinating, fruits were cut open with scalpel. Some seeds were collected and fixed in ethanol: acetic acid (3: 1) solution for 30 min. After their coats were carefully removed with knife and needle, they were softened with an enzyme mix, containing 1% (w/v) cellulase RS (Duchefa Biochemie, Haarlem, The Netherlands) and 1% (w/v) pectinase Y23 (Duchefa Biochemie), at 37 °C for about 40 min. Finally, some endosperms were thoroughly mixed with 20 µl 45% acetic acid on a glass slide, covered with a glass cover, placed on a 65 °C heating plate for 10 s, and then squashed. The slides were checked with a phase contrast microscope (ZEISS Scope. A1).

Embryo rescue

The fruits were harvested when they became soft or yellow. Each fruit was immersed in 75% ethanol for a few seconds and then taken out and flamed to sterilize surface, then, it was cut open on a sterilized dish. After seed coats were carefully removed, the seeds were put on a medium, containing 2.2 g l^− 1 MS (Duchefa Biochemie), 60 g l^− 1 sucrose (Xilong Scientific), 4 g l^− 1 gelrite (Duchefa Biochemie) and pH 5.8. They were cultured in dark boxes at 25 °C for 40–60 d and then transferred to lily propagation medium, containing 2.2 g l^− 1 MS, 50 g l^− 1 sucrose, 4 g l^− 1 gelrite and pH 5.8, and cultured at 25 °C under 2500 lx light intensity for about 10 weeks.

Mitotic chromosome preparation

The mitotic chromosomes were prepared according to Zhou et al. [29]. When lily roots were about 1 cm long, they were cut off and pretreated with 0.7 mM cycloheximide (Duchefa Biochemie) at room temperature for 4 h, and then fixed in ethanol: acetic acid (3: 1) solution for 12 h. The root tips were washed with distilled water and softened in an enzyme mix, containing 1% (w/v) cellulase RS (Duchefa Biochemie) and 1% (w/v) pectinase Y23 (Duchefa Biochemie), at 37 °C for about 45 min. Their meristem was mixed with 20 µl 45% acetic acid on a glass slide, covered with a glass coverslip, placed on a 65 °C heating plate for 10 s and then squashed. The chromosome preparation was checked with a phase contrast microscope (ZEISS Scope. A1).

Fluorescence in situ hybridization (FISH)

The protocol was according to Liu et al. [24]. One 59-nucleotide sequence (5ˊ-GGATGCGATCATACCAGCACTAATGCACCGGATCCCATCAGAACTCCGCAGTTAAGCGT-3ˊ) of 5S rDNA of Arabidopsis thaliana (AJ307350) was labeled with FAM (a green fluorescence dye) or TAMRA (a red fluorescence dye) by Tsingke Company (Beijing, China). The probe was diluted to 50 ng/µl.

The slides were pre-treated with 0.01 M HCl for 2 min, followed by 5 µg/ml pepsin at 37 °C for 10 min. They were then washed in Milli-Q water for 2 min, followed by two washes in 2 × SSC (0.3 M NaCl and 30 mM sodium citrate, pH 7.0) for 5 min each. Subsequently, the slides were incubated in 4% paraformaldehyde for 10 min at room temperature to firmly mount the chromosomes onto the slides, followed by one additional wash in 2 × SSC for 5 min. The cleaned chromosome slides were then dehydrated through a graded ethanol series (70%, 90%, and 100%) for 3 min each and air-dried.

For hybridization, a 40 µl mixture was added to each slide, containing 20 µl deionized formamide (Sangon Biotech, Shanghai, China), 8 µl 50% dextran sulfate (prepared by 0.5 g dextran sulfate with 1 ml distilled water; Solarbio Life Science, Beijing, China), 4 µl 20 × SSC (3 M NaCl and 0.3 M sodium citrate, pH 7.0), 1 µl 10% SDS solution (Phygene), 2 µl probe and 5 µl Milli-Q water. Then, after the chromosomes were denatured for 5 min at 80 °C, the slides were placed in a humid box at 37 °C overnight. Following hybridization, the slides were washed three times in 2 × SSC for 5 min each, and twice in 0.1 × SSC at 42 °C for 15 min each. Finally, a drop of Vectashield 1200 was added to each slide, which was then covered with a glass coverslip. Zeiss fluorescence microscope was used to examine the slides and capture images.

Genomic in situ hybridization (GISH)

The genomic DNA of ‘White Fox’ (LL) was extracted with universal plant genomic DNA kit (TransGen Biotech, China). The DNA was labeled with biotin-16-dUTP (Biotin-Nick translation Mix, Roche, 11745824910). GISH protocol was according to Zhou et al. [29]. Herring sperm (HS) DNA was dissolved in Milli - Q water and used as blocking DNA to reduce the combination of the probe and non-specifc DNA sequence. The hybridization mix (40 µl) contained 20 µl of deionized formamide, 8 µl 50% dextran sulfate, 4 µl 20 × SSC, 1 µl 10% SDS solution, 25–50 ng of probe DNA, and 1–5 µg of herring sperm. Biotin signal was detected and amplified with Streptavidin-CY3 (Invitrogen, Camarillo) and Biotinylated anti-Streptavidin (Vector Laboratories, Burlingame). Having been mounted with VECTASHIELD (H-1200, Vector Laboratories, Burlingame), the slides were observed under a fluorescence microscope (ZEISS Scope. A1).

Results

Genome composition and pollen germination

The flowers of the four lily cultivars were shown in Fig. 2A₁-D₁. ‘Dan Feng’ and ‘Lollypop’ were diploid (2n = 2x = 24). Each contained two sets of Asiatic (A) chromosomes, i.e., AA, and each had two chromosomes bearing 5S rDNA loci (Fig. 2A₂ and B₂). ‘Tresor’ was tetraploid (2n = 4x = 48), it comprised four sets of A-chromosomes, i.e., AAAA, and four chromosomes have 5S rDNA loci (Fig. 2C₂). The results indicated that one genome had one chromosome with 5S rDNA loci, suggesting that the 5S rDNA chromosomes could represent the lily genomes. ‘Chen Xi’ contained 9 Longiflorum (L) chromosomes and 35 A-chromosomes, and it had one L-chromosome and three A-chromosomes with 5S rDNA loci (Fig. 2D₂). The GISH results confirmed that ‘Chen Xi’ was an LA aneuploid.

Fig. 2.

Flowers, mitotic chromosomes painted with FISH or GISH, and pollen germination. A₁-₃) ‘Dan Feng’; B₁-₃) ‘Lollypop’; C₁-₃) ‘Tresor’; D₁−₃) ‘Chen Xi’. The white arrow heads indicate the 5S rDNA loci. Red chromosomes are L-chromosomes, and blue chromosomes are A-chromosomes. Scale bars are 2 mm in A₁-D₁, 20 μm in A₂-D₂ and 200 μm in A₃-D₃, respectively

The rate of pollen germination of ‘Dan Feng’, ‘Lollypop’ and ‘Tresor’ was 34.5%, 30.6% and 30.8%, respectively, implying that they were male fertile (Fig. 2A₃-C₃); However, that of ‘Chen Xi’ was 1.5%, indicating that it was highly male sterile (Fig. 2D₃).

Abnormal meiosis of ‘Chen Xi’

At metaphase I (Fig. 3A-B), chromosome associations were variable in different PMCs. About half of A-chromosomes were associated and formed trivalent (white “III”), other A-chromosomes formed bivalents (white “II”), tetravalent (white “IV”) or univalents (white “I”). Most L-chromosomes usually remained univalents (red “I”). A very few L-chromosomes were loosely associated with A-chromosomes (white “III” + red “I”, white “I” + red “I”). From anaphase I to telophase II (Fig. 3C-E), chromosomal bridges, lagging (free) chromosomes, or micronuclei were commonly observed. Its microspores or pollen grains (Fig. 3F-H) had variable chromosome numbers, ranging from 20 to 29 and contained L-, A-, and L/A recombinant chromosomes. GISH indicated that the micronuclei were formed from free or lagging chromosomes or fragments.

Fig. 3.

The representative phases of microsporogenesis of ‘Chen Xi’ painted with FISH, GISH or Carbol Fuchsin. A-B) metaphase I; C-D) anaphase I; E) telophase II; F-H) microspores. Note: In A and B, A-chromosomes are blue and L-chromosomes are red; “I”, “II”, “III” and “IV” means univalent, bivalent, trivalent and tetravalent of synapses, respectively; the white arrow heads indicate the 5S rDNA loci, the red arrow heads indicate the L/A intergenomic recombinant chromosomes. Bars = 20 μm

Number of chromosomes with 5 S rDNA loci of endosperm nuclei

The chromosomes with 5S rDNA loci at prophase or metaphase of endosperm nuclei from the three hybridizations were illustrated in Fig. 4. In ‘Dan Feng’ × ‘Lollypop’ (AA × AA), 36 out of 45 endosperm nuclei had five chromosomes with 5S rDNA loci (Table 1), and it was lucky to get a good metaphase in which 60 chromosomes could be clearly counted (Fig. 4A₁-A₂). For ‘Dan Feng’ × ‘Tresor’ (AA × AAAA), 46 of 62 endosperm nuclei had six chromosomes bearing 5S rDNA loci (Fig. 4B₁-B₃, Table 1). Since one lily genome contained one chromosome with 5S rDNA loci, it was concluded that the endosperm of ‘Dan Feng’ × ‘Lollypop’ (AA × AA) and ‘Dan Feng’ × ‘Tresor’ (AA × AAAA) was 5x and 6x respectively, and thus the secondary nuclei of embryo sacs of the diploid ‘Dan Feng’ were tetraploid.

Fig. 4.

The representative images of endosperm chromosomes painted with FISH and/or GISH. A₁−₂) ‘Dan Feng’ × ‘Lollypop’ (AA × AA); B₁−₃) ‘Dan Feng’ × ‘Tresor’ (AA × AAAA); C₁−₃) and D₁−₃) ‘Chen Xi’ × ‘Tresor’ (L*A* × AAAA). The chromosomes with 5S rDNA loci are pointed with arrow heads. The red chromosomes belong to L-genome. Bars = 20 μm

Table 1.

The numbers of chromosomes with 5S rDNA loci of endosperm nuclei (Ch. of E. N.) in the number of endosperm nuclei (No. of E. N.) observed and expected. According to the chi-square text (X² = ∑(O – E)²/E = 1.24, df = 2), the possibility value is between 50–90%, which is much greater than the cutoff value 5%, indicating that the observed results are compatible with the expected

		Observed								Expected
24,037	Ch. of E. N.	5	6	7	8	9	10	11	12	5
	No. of E. N.	36	4	3	2	0	0	0	0	45
24,095	Ch. of E. N.	5	6	7	8	9	10	11	12	6
	No. of E. N.	6	46	8	2	0	0	0	0	62
24,096	Ch. of E. N.	5	6	7	8	9	10	11	12	10
	No. of E. N.	0	0	0	3	5	68	8	1	85

As for ‘Chen Xi’ × ‘Tresor’ (L*A* × AAAA), 68 of 85 endosperm nuclei had ten chromosomes with 5S rDNA loci (Fig. 4C-D; Table 1) and 18 L-chromosomes (Fig. 4D₁-D₃). Since the tetraploid lily ‘Tresor’ had four chromosomes with 5S rDNA loci and contributed 2x pollen containing only A-chromosomes, thus, the secondary nuclei of ‘Chen Xi’ (L*A*) contained eight chromosomes with 5S rDNA loci and 18 L-chromosomes. Based on the results and the chi-square test, it was clearly confirmed that the chromosome numbers of the secondary nuclei of embryo sacs were also twice those of somatic cells of aneuploid lilies.

Genome composition of functional egg cells

From 10 pollinated flowers of ‘Chen Xi’ × ‘Tresor’ (L*A* × AAAA), 10 fruits were harvested, 213 seeds with developed endosperm could be isolated and cultured in growth room, finally, 92 seedlings were obtained.

To analyze the genome compositions of functional eggs of ‘Chen Xi’, 9 seedlings with vigorous roots were selected for FISH and GISH. The results were shown in Fig. 5; Table 2. All progeny from ‘Chen Xi’ × ‘Tresor’ (L*A* × AAAA) were aneuploid, with chromosome numbers ranging from 43 to 72 (on average 58.1). The number of 5S rDNA chromosomes ranged from three to six (on average 4.7). Four individuals had six chromosomes containing 5S rDNA loci; three of them had three 5S rDNA chromosomes; 24096-5 and 24096-14 had five and four 5S rDNA chromosomes, respectively. Since the male parent ‘Tresor’ (AAAA, 2n = 4x = 4 A = 48) contributed 24 A-chromosomes, the female parent ‘Chen Xi’ (L*A*, 2n = 4x − 4 = 9 L + 35 A = 44) must have contributed functional aneuploid egg cells with highly variable chromosome numbers, ranging from 19 to 48 (on average 34.1). Surprisingly, two functional egg cells for two seedlings, 24096-4 and − 8, contained more chromosomes than the somatic cells of ‘Chen Xi’ itself, implying that the meiosis of ‘Chen Xi’ was perplex.

Fig. 5.

The representative images of metaphase chromosomes of root tips of seedlings obtained from ‘Chen Xi’ × ‘Tresor’ (L*A* × AAAA). The L-chromosomes are red, and A-chromosomes are blue. The yellow arrow heads indicate the fused positions of L/A intergenomic chromosomes. A 24096-1; (B) 24096-3; (C) 24096-4; (D) 24096-5; (E) 24096-8; (F) 24096-14; (G) 24096-16; (H) 24096-21; (I) 24096-22. The red arrow heads indicate the loci of 5S rDNA. All their genome compositions refer to Table 2. Bars = 20 μm

Table 2.

Genome compositions of functional aneuploid eggs produced by aneuploid ‘Chen Xi’. They are the number of chromosomes (No. of Ch) of seedlings minus that of sperms based on analyzing 9 progenies using GISH. L^ch, A^ch, or L/A^ch means the number of L-chromosomes, A-chromosomes, or recombinant chromosomes, respectively

			Functional eggs produced by ‘Chen Xi’
Seedling Code	Seedling No. of Ch.	Sperm No. of Ch.	Egg No. of Ch.	Lch	Ach	L/Ach
24096-1	67	24	43	8	34	1
24096-3	45	24	21	2	17	2
24096-4	72	24	48	8	39	1
24096-5	68	24	44	9	34	1
24096-8	69	24	45	9	36	0
24096-14	46	24	22	5	17	0
24096-16	45	24	21	3	18	0
24096-21	68	24	44	7	34	3
24096-22	43	24	19	1	18	0
Average	58.1	24	34.1	5.8	27.4	0.9

Discussion

The present research not only confirms that the chromosome numbers of secondary nuclei in the embryo sac are also invariably twice those of somatic cells in aneuploid lilies but also could explain why the aneuploid lilies have partial female-fertility.

Aneuploids are usually low viability, highly sterile and seedless, because their additional or missing chromosomes cause gene-unbalanced and their meiosis is irregular [30–35]. A very limited cases are reported in seed crops except that some aneuploids are usually near euploid, like 2x ± 1 or 3x ± 1 [36–41]. However, the aneuploid lilies have variable chromosome numbers, and they are obtained from 3x × 2x/4x crosses [11–14, 19, 20, 23, 24]. In addition, the aneuploid lilies, such as ‘Chen Xi’ in the present research, can also be female to crosses with tetraploid or diploid, and produce aneuploid progenies. On this point, aneuploid lilies are similar to triploid lilies, but they are different from triploid or aneuploid polygonum-type plants because triploid watermelon and banana are seedless. Ratio of endosperm/embryo and endosperm balance number (EBN) are the theories for explaining the success or failure of hybridizations in Polygonum-type plants, however, they do not apply to the success or failure of lily hybridizations because Lilium are Fritillaria-type plants [10, 12]. Since angiosperms produce seeds through double fertilization - one sperm fuses with an egg cell and develops into an embryo while the other sperm fuses with the central cell and develops into endosperm [8, 42–44], then, the success or failure of a hybridization is determined by both embryo and endosperm [45]. Though many researchers use ratio of endosperm/embryo and EBN to explain the success or failure of hybridizations [46–52], there are too many exceptions and some reports suggest the genetic balance of endosperm itself is more important than the ratio of endosperm/embryo [53]. In some Arabidopsis mutants, endosperm could independently develop [54–56].

Considering the difference of viability between euploid (x/2x) gametes and aneuploid gametes, we have well explained why triploid lilies usually have partial female-fertility regardless of male sterility [10–12, 23, 24]. The main reason is that the chromosome number of the secondary nucleus is twice that of the egg cell in each Polygonum-type embryo sac, while the chromosome number of the secondary nucleus is twice that of the somatic cells in each Fritillaria-type embryo sac (Fig. 1). After double fertilization in 3x × 2x/4x hybridizations in Polygonum-type plants, both embryos and endosperm are aneuploid, then their seeds could not develop because neither embryo nor endosperm is functional (Fig. 6A); However, in Fritillaria-type plants, embryos are aneuploid while endosperm is euploid, thus most embryos could not develop while endosperm could develop (Fig. 6B).

Fig. 6.

Schematic comparison of prior and post double fertilization between Polygonum-type and Fritillaria-type plants. A In 3x × 2x/4x hybridizations of Polygonum-type plants, they are usually seedless. B In 3x × 2x/4x hybridizations of lilies, their seeds could develop at some extent. C In aneuploid × 2x/4x in lilies, their seeds also could develop at some extent

When aneuploid Lilium used as female in the present research (Fig. 6C), both embryos and endosperms of aneuploid × 2x/4x are aneuploid; however, the aneuploid Lilium has partial female-fertility rather than completely seedless. It seems to be confusing. How to explain the phenomenon? The aneuploid lilies are usually weak because of unbalanced chromosomes (or genes), however, among many aneuploid lilies with variable chromosomes, there are a few lines which grow vigorously. The vigorous aneuploid lilies, such as ‘Chen Xi’, should have balanced genes as euploid lilies. Therefore, secondary nuclei of such aneuploid × 2x/4x hybridizations have balanced genes because they have twice nuclear DNA as that of somatic cells. This would be the reason why the aneuploid lilies are partial female-fertile in the present research.

Tulips are also Fritillaria-type plants; however, the ratio of the ploidy levels of embryo and endosperm of crosses like 2x × 2x, 3x × 2x, or 2x × 4x may 1: 1 [26], implying that the amount of nuclear DNA in the secondary nucleus (central cell) is the same as that of the egg cell. Regarding the point, it is completely different in Lilium because the amount of nuclear DNA of the secondary nucleus (central cell) is twice that of somatic cells whether in diploid, triploid or aneuploid lilies.

Aneuploidy has diverse effects in plant species [57–59]. Some Lilium spp. are also edible or medicinal plants. Aneuploid lilies could also increase chances to select promising varieties for food or medicine.

Since characteristics of Fritillaria-type are the essential of partial-female fertility of triploid and aneuploid lilies, it is reasonable that the principle in Lilium might be applied to most other plants with tetrasporic embryo sac because most tetrasporic embryos sacs have the same characteristics - The amount of nuclear DNA in the secondary nucleus is twice that of its somatic cells in embryo sacs regardless of normal or abnormal megasporogenesis.

Conclusions

Based on present results and discussion, we conclude that aneuploid Lilium produces variable chromosomal eggs and invariable chromosomal secondary nuclei in its embryo sacs, and this is the reason why the aneuploid lilies have partial female-fertility regardless of its male sterility. This principle would also be applied to Fritillaria-type or other tetrasporic-type plants.

Authors’ contributions

JJC performed most experiments and wrote initial manuscript. LC, YLZ, LL and YNS assist in some related works. SJZ proposed the idea and revised the manuscript. All authors read and approved the final manuscript for submission.

Funding

This work was funded by the National Natural Science Foundation of China for financial support (Grant No. 32260783).

Data availability

All original contributions made in this study are contained within the article, and any additional queries may be addressed to the corresponding author.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.