Introduction

Interspecific hybridization has long been a cornerstone in the breeding of ornamental plants, enabling the combination of desirable traits such as novel flower morphology, fragrance, and enhanced stress tolerance (Van Tuyl and Lim 2003;Kato and Mii 2012). In Lilium breeding, complex hybrids such as longiflorum-Asiatic (LA) and Oriental-Trumpet (OT) cultivars have been extensively developed through interspecific crosses between distantly related genome groups (Lim et al. 2008;Marasek-Ciolakowska et al. 2018;Van Tuyl et al. 2018). These hybrids, widely cultivated in the global floriculture market, represent significant breeding achievements that integrate diverse phenotypic and horticultural characteristics (Marasek-Ciolakowska et al. 2018).

The creation of interspecific hybrids often results in complex genomic compositions and unpredictable chromosomal behaviors, especially when involving parents from divergent genome groups (Wang et al. 2015;Rodionov et al. 2019). These hybrids frequently suffer from reduced fertility or meiotic irregularities due to chromosomal incompatibilities (Mao-Sen et al. 2021). In many cases, breeders overcome these obstacles by employing somatic chromosome doubling techniques to stabilize genome sets, enabling clonal propagation in otherwise sterile hybrids (Alix et al. 2017). As a result, numerous triploid cultivars are maintained commercially through vegetative methods such as bulb division or tissue culture, although their underlying chromosomal structure and genomic origin remain poorly documented.

Although homoeologous recombination can occur during meiosis, meiotic polyploidization provides a mechanism for retaining such recombinant chromosomes in viable gametes (Barba-Gonzalez et al. 2005;Pecinka et al. 2011). While these mechanisms have been well-documented in experimental hybrids or controlled breeding populations (Karlov et al. 1999;Lim et al. 2003, 2005), it remains unclear whether such recombination or structural dynamics are present in commercially developed cultivars. Despite the widespread the widespread commercialization of interspecific lily cultivars, the genomic structure and chromosomal behavior of these hybrids remain largely uncharacterized. Most breeding programs focus primarily on phenotypic selection and pedigree records, while the underlying genome compositions, such as parental origin and chromosomal stability, are often not cytogenetically verified. This lack of genomic profiling leaves a critical gap in our understanding of how these hybrids are maintained, stabilized, and propagated over generations.

Genomic in situ hybridization (GISH) offers a reliable and informative method to visualize the genome composition of interspecific hybrids by differentiating chromosomes derived from each parental genome (Marasek and Okazaki 2006;Van Laere et al. 2010). It has been widely applied to assess meiotic recombination, chromosomal introgression, and genome stability in synthetic and experimental hybrids. Although GISH has been widely used in experimental lily hybridizations and genome composition studies (Van Tuyl et al. 2002;Marasek et al. 2004;Xie et al. 2013), its application to multiple commercially cultivated cultivars for systematic cytogenetic profiling remains limited.

In this study, we applied GISH to a panel of triploid LA and OT lily cultivars to investigate their chromosomal composition, evaluate genome stability, and explore the utility of GISH as a practical tool for cultivar verification. Our goal was to assess whether the observed genomic structures reflect somatic polyploidization without recombination and to provide cytogenetic evidence supporting the clonal maintenance strategy of these commercially important hybrids.

Materials and Methods

Morphological Trait Analysis

Morphological traits were assessed using greenhousegrown plants at full flowering stage. For each cultivar, three representative individuals were selected for measurement and observation. Vegetative traits (Table 1) were recorded as follows: stem height was measured from the base of the bulb to the topmost flower using a ruler; presence or absence of stem hair and stem anthocyanin was assessed visually; total leaf number was counted manually; leaf arrangement and the shape of the leaf base were recorded based on visual examination. Leaf length and width were measured at the middle third of the stem using a digital caliper. Leaf gloss, yellow patch occurrence, and leaf burn symptoms were assessed visually under consistent lighting conditions.

Floral traits (Table 2) were recorded from fully opened flowers. The number of flowers per stem was counted, and flowering direction was categorized as “upward”, “outward”, or “side-facing” based on the orientation of the flower axis. Flower type was classified as “single” or “double.” Flower width was measured as the maximum diameter across fully opened tepals. Flower color and color type (unicolor or complex) were documented photographically and described visually. Presence, size, and location of flower spots were recorded, along with their contribution to overall floral coloration. Petal texture (e.g., wrinkling, bumps), waviness, and edge curvature were recorded based on close-up visual inspection. Stigma position relative to anthers was scored as “equal”, “higher”, or “lower” and fragrance was qualitatively rated as “weak”, “strong”, or “moderate” by at least two observers.

Plant Materials and Preparation of Somatic Chromosomes

Scales from three Lilium longiflorum-Asiatic (LA) cultivars (‘Eyeliner’, ‘Serengeti’ and ‘Bonsoir’) and five Oriental-Trumpet cultivars (‘Bowmore’, ‘Motown’, ’Scheherazade’, ‘Pink Palace’ and ‘Amarossi’) were collected and cultured in vitro on Murashige and Skoog (MS) medium (60 g·L^-1 sucrose, 4.4 g·L^-1 MS basal salts, and 8 g·L^-1 plant agar) adjusted to pH 5.8. Young root tips were treated with 1 M α-bromonaphthalene for 5 hrs at 16°C to arrest mitotic cells at metaphase, rinsed with distilled water, and fixed in a 3:1 (v/v) ethanol-acetic acid solution overnight. Fixed samples were stored in 70% ethanol at -20°C until use. For chromosome preparation, root tips were digested with a pectolytic enzyme solution (0.3% pectolyase Y23, 0.3% cellulase RS, and 0.3% cytohelicase) in 10 mM citric acid buffer at 37°C for approximately 1 hour. Root meristems were squashed in 60% acetic acid, air-dried, and stored until use. After air drying, the slides were observed under a microscope (Olympus, BX3 URA, Japan).

Ploidy Analysis

Flow cytometry was conducted using a Partec PA ploidy analyzer (Sysmex, Japan) to determine the ploidy levels of each cultivar. Approximately 1 cm² of young leaf tissue was finely chopped with a razor blade in nuclear extraction buffer (Sysmex, Japan), filtered through a 30 μm nylon mesh, and stained using staining buffer (Sysmex, Japan). The resulting nuclear suspensions were analyzed to determine ploidy level.

Genomic DNA Extraction and Probe Labeling

Genomic DNA was extracted from young leaves of L. longiflorum ‘Woori Tower’ and Oriental hybrid ‘Sunny Bahamas’ for use as GISH probes. Frozen leaf tissue was ground in liquid nitrogen and suspended in extraction buffer containing 1% N-lauryl sarcosine, 100 mM Tris-HCl (pH 8), 10 mM EDTA (pH 8), and 100 mM NaCl. After vortexing, an equal volume of phenol-chloroform-isoamyl alcohol (25:24:1) was added. Following centrifugation at 16,000 rpm for 3 min at room temperature, the aqueous phase was transferred to a clean tube, and genomic DNA was precipitated with 70 μL of 3 M sodium acetate and 700 μL of isopropanol. After centrifugation at 16,000 rpm for 10 min at 4°C, the DNA pellet was washed with 70% ethanol, air-dried, and dissolved in 20 μL distilled water. Genomic DNA was sonicated to 1-10 kb fragments and labeled with digoxigenin-11-dUTP using a nick translation kit (Roche, Germany). Probe size (300-500 bp) was confirmed via agarose gel electrophoresis. Herring sperm DNA (1 μg·μL⁻¹), autoclaved at 121°C for 10 min, was used as blocking DNA to reduce non-specific signals.

Genomic In Situ Hybridization (GISH)

GISH was performed following Lim et al. (2008) with minor modifications. Chromosome spreads were pretreated with RNase A (100 μg·mL^-1) at 37°C for 1 hr, incubated in 10 mM HCl for 2 min, and treated with 5 μg·mL^-1 pepsin at 37°C for 20 min. Slides were fixed in 4% formaldehyde for 10 min. Hybridization mixture contained 50% deionized formamide, 10% dextran sulfate, 2×SSC, 0.25% sodium dodecyl sulfate, 0.8-1.0 ng·μL^-1 per probe, and 20–50 ng·μL^-1 blocking DNA. The hybridization mixture was denatured at 80°C for 5 minutes and applied to the chromosome preparations. Slides were incubated overnight at 37°C in a moist chamber. After hybridization, slides were washed three times in 2× SSC at room temperature. The pre-denatured probes were incubated at 70-75°C for 10 minutes and then denatured at 80°C for 5 minutes before hybridization overnight in a moist chamber at 37°C. After hybridization, the probes were washed three times with 2×SSC for 5 minutes each at room temperature, followed by three washes with 0.1×SSC for 5 minutes each at 42°C. The slides were then hybridized overnight at 37°C. Once hybridization was complete, the probes were washed three more times with 2×SSC for 5 minutes each at room temperature, followed by a final wash with 0.1×SSC for 40 minutes at 42°C. To detect digoxigenin-labeled DNA, anti-digoxigenin-fluorescein antibodies raised in sheep (Boehringer, Mannheim, Germany) were used, followed by amplification with fluorescein anti-sheep immunoglobulin raised in rabbit (Vector Laboratories, Burlingame, USA). For the detection of digoxigenin-11-dUTP, FITC-conjugated anti-digoxygenin antibodies (Roche, Germany) were applied. The slides were then counterstained with 1 μg·mL⁻¹ 4,6-diamidino-2- phenylindole (DAPI, Sigma) and mounted with Vectashield (Vector Laboratories). At least five metaphase cells from each genotype were examined, and images were captured using an imaging photomicroscope (Olympus BX3-URA, Japan) equipped with epifluorescence illumination.

Results

Phenotypic Characteristics Analysis

Phenotypic differences between L. longiflorum-Asiatic (LA) and Oriental-Trumpet (OT) cultivars were evident across multiple vegetative and floral traits, including stem height, leaf dimensions, and flower morphology. OT cultivars generally exhibited greater stem heights, ranging from 70.38±1.18 cm in ‘Scheherazade’ to 120.09±1.42 cm in ‘Motown’, whereas LA cultivars such as ‘Serengeti’ and ‘Bonsoir’ ranged from 77.33±0.49 cm to 108.31±2.82 cm (Table 1). While there was some overlap in size, OT cultivars tended to be taller overall.

Leaf characteristics also differed between the groups. LA cultivars displayed relatively narrow leaves (2.01±0.07-3.32±0.10 cm), whereas OT cultivars, including ‘Pink Palace’, exhibited broader leaves up to 4.44±0.09 cm in width. Leaf length varied, with ‘Bowmore’ reaching 18.38±0.16 cm, the longest among all cultivars. Leaf gloss was observed in several OT cultivars (‘Motown’, ‘Scheherazade’, ‘Pink Palace’, and ‘Amarossi’) and in one LA cultivar (‘Bonsoir’), suggesting variation in surface texture. All cultivars shared alternate leaf arrangement and straight basal leaf orientation, except for ‘Eyeliner’, which displayed a slight fold-back.

OT cultivars generally exhibited larger and more structurally complex flowers, with ‘Bowmore’ and ‘Scheherazade’ showing average flower widths of 23.77±0.35 cm and 17.27 ± 0.06 cm, respectively. These cultivars displayed bicolored floral patterns, such as white-pink in ‘Bowmore’ and red-yellow in ‘Scheherazade’. In contrast, LA cultivars produced relatively smaller, unicolored flowers, white in ‘Eyeliner’ (16.87 ± 0.49 cm), red in ‘Serengeti’ (15.73 ± 0.15 cm), and orange in ‘Bonsoir’ (14.70 ± 0.10 cm).

Petal surface traits also varied among cultivars. Wrinkling was common across all cultivars, but bumps were more commonly observed in OT cultivars such as ‘Scheherazade’ and ‘Amarossi’. On the other hand, LA cultivars, particularly ‘Eyeliner’ and ‘Serengeti’, exhibited more frequent flower spots, indicating a group-specific pattern in surface ornamentation. Petal edges were generally curled in all cultivars, but wavy margins were more prevalent in OT types.

Bloom direction was another distinguishing trait, LA cultivars displayed upward or diagonal-upward floral orientation, while OT cultivars such as ‘Pink Palace’ and ‘Amarossi’ showed diagonal or downward-facing flowers. Fragrance was typically weak across most cultivars, though moderate scent was noted in ‘Scheherazade’ and ‘Pink Palace’. These phenotypic patterns reflect both the genomic backgrounds and breeding targets of LA and OT hybrids, with LA cultivars showing simpler, more uniform traits and OT cultivars expressing more diverse and elaborate floral features.

Ploidy Analysis

Ploidy levels of all analyzed cultivars were determined through flow cytometry and confirmed via somatic chromosome counting. Flow cytometry histograms exhibited clear peaks corresponding to triploid DNA content in all LA (‘Eyeliner’, ‘Serengeti’, and ‘Bonsoir’) and OT (‘Bowmore’, ‘Motown’, ‘Scheherazade’, ‘Pink Palace’, and ‘Amarossi’) cultivars (Fig. 1). Microscopic observation of root-tip metaphase cells further validated these findings, with each cultivar consistently possessing 36 chromosomes (2n = 3x = 36) (Fig. 2; Table 3). These results confirm the triploid status of both LA and OT hybrids, supporting their origin through somatic polyploidization and providing the cytogenetic basis for subsequent GISH analysis.

GISH and Karyotype Analysis

Genomic in situ hybridization (GISH) was used to assess the genomic constitution and chromosomal stability of the eight triploid Lilium cultivars. Using L. longiflorum and Oriental hybrids genomic DNA as probes, parental chromosomes were clearly distinguished in each hybrid background.

In the LA cultivars (‘Eyeliner’, ‘Serengeti’, and ‘Bonsoir’), GISH revealed 12 chromosomes derived from L. longiflorum (labeled with FITC, green) and 24 from Asiatic parents (counterstained with DAPI and visualized in red using pseudocolor rendering), totaling 2n = 3x = 36 (Fig. 3A-C). No evidence of intergenomic translocations or recombination was observed, indicating a non-recombinant genomic structure. These results support the hypothesis that the LA hybrids were generated via somatic polyploidization of F₁ interspecific hybrids, allowing preferential homologous chromosome pairing and suppressing homoeologous recombination.

Similarly, the five OT cultivars (‘Bowmore’, ‘Motown’, ‘Scheherazade’, ‘Pink Palace’, and ‘Amarossi’) exhibited the same genomic pattern, 12 chromosomes of Oriental hybrid origin and 24 of Trumpet hybrid origin (Fig. 3D-H). No recombination or structural rearrangements were detected in any OT hybrid, further confirming their triploid and non-recombinant nature. To visualize the chromosomal composition of each cultivar, idiograms were constructed based on GISH results (Fig. 4). All cultivars showed stable chromosomal configurations with clearly separated parental genomes and no chromosomal aberrations. These findings suggest that commercially cultivated triploid LA and OT hybrids exhibit consistent cytogenetic profiles, and that GISH is an effective tool for confirming genome structure and hybrid authenticity.

Discussion

The phenotypic and cytogenetic profiles of the eight triploid Lilium cultivars analyzed in this study revealed characteristic distinctions between the LA (L. longiflorum- Asiatic) and OT (Oriental-Trumpet) hybrid groups. Morphological traits such as stem height, leaf dimensions, and floral architecture reflected the influence of their parental genomic backgrounds (Table 1 and 2). LA cultivars tended to exhibit relatively smaller floral organs, narrower leaves, and more frequent flower spotting, while OT cultivars were characterized by larger, bi-colored flowers, broader petals, and floral features such as petal bumps and edge curling. Fragrance was also more pronounced in certain OT cultivars. These phenotypic patterns are consistent with known parental contributions and horticultural breeding goals of each hybrid group (Dhiman et al. 2022).

All eight cultivars were confirmed to be triploid (2n = 3x = 36) with clearly distinguishable parental chromosome sets and no evidence of intergenomic recombination, as revealed by GISH analysis (Fig. 3 and 4). This stable cytogenetic structure supports somatic polyploidization as the primary mechanism underlying the formation of these hybrids (Sattler et al. 2016;Rodionov et al. 2019). Somatic polyploidization, typically induced during early in vitro culture stages or via chemical treatment, causes genome duplication through mitosis without involving meiotic recombination (Manzoor et al. 2019;Miri 2020;Niazian and Nalousi 2020). The resulting triploid plants inherit full chromosome sets from both parents without chromosomal exchange, as consistently observed in the GISH patterns..

Compared to meiotic polyploidization, which often results in irregular pairing, unbalanced segregation, and chromosomal rearrangements (De Storme and Geelen 2013;Pelé et al. 2018), somatic duplication offers superior genomic stability. This method preserves desirable traits from parental lines, ensuring fixed combinations are maintained across generations (Ramanna et al. 2012;Eeckhaut et al. 2018;Datta 2022). The suppressed homoeologous recombination observed here further reinforces this stability, indicating that homologous rather than homoeologous chromosomes are involved in segregation, preventing the reshuffling of genetic material (Gleba et al. 1987;Jones and Pašakinskiene 2005).

While this stability benefits uniformity and cultivar maintenance, it limits the potential for introducing novel traits via recombination. This restricts the utility of triploid hybrids as breeding materials in conventional sexual hybridization. The lack of recombination also contributes to poor pollen viability due to irregular gamete formation, a common limitation in odd-ploidy plants (Van Tuyl and De Jeu 1997;Pelé et al. 2018). As such, these hybrids are more suited to clonal propagation, such as bulb division or in vitro tissue culture, for commercial use. Breeding approaches that aim to introduce genetic diversity, such as trait pyramiding or transgressive segregation, would require diploid or tetraploid lines with active recombination (MacKay et al. 2009;Xu and Crouch 2008).

In conclusion, the somatic origin and cytogenetic stability of the analyzed triploid Lilium cultivars underscore their suitability for vegetative propagation and commercial production. GISH analysis not only confirms their genomic integrity but also provides a valuable tool for cultivar verification and propagation planning in Lilium breeding programs.

Conclusion

This study demonstrates that the eight commercially cultivated Lilium cultivars, encompassing both LA (L. longiflorum-Asiatic) and OT (Oriental-Trumpet) hybrid groups, are cytogenetically stable triploids (2n = 3x = 36) derived through somatic polyploidization. GISH analysis revealed intact, non-recombinant parental chromosome sets in all cultivars, with no evidence of structural rearrangements or intergenomic recombination. These findings underscore the reliability of somatic polyploidization for producing stable triploid hybrids and confirm that homoeologous recombination is strongly suppressed in these genotypes. In addition to their genomic stability, the cultivars exhibited distinct morphological characteristics reflecting their hybrid backgrounds, including differences in stem height, leaf width, floral size, color patterns, and fragrance. Such traits are consistent with the known horticultural profiles of LA and OT hybrids and contribute to their commercial value.

Collectively, these results validate GISH as a practical tool for assessing genome composition and hybrid authenticity in Lilium breeding programs. The combination of cytogenetic profiling and morphological evaluation provides a comprehensive framework for cultivar verification, propagation planning, and the future registration of elite ornamental hybrids.