Introduction
Roses are cultivated in commercial greenhouses for year-round production, and they account for about 30% of cut flower production in Korea. They have various phenotype traits, such as flower shape, type, and size; color of petal and stamen; scent type or intensity; and number of petals (Yeon and Kim 2017;Yeon and Kim 2020a). Domestic cut rose export to Japan has accounted for 99% and climate changes in Korea have lowered the quality, which has resulted in underestimation of the quality of Korean flowers in overseas markets (Kim et al. 2021;Lee et al. 2018;Yeon and Kim 2017). Consumers always demand new cultivars to satisfy their preferences for ornamental traits, and the frequency of cultivar replacement for cut rose production in commercial greenhouses was reduced from 4-5 to 2-3 years (Cheong et al. 2020).
Most domestic rose cultivars are bred by using traditional breeding techniques. This takes more than 5 years and requires many processes, such as pollination for crossing, hip harvesting, seed sowing, selection, and testing; but trend in consumer preferences changes more quickly in flower markets (Koh and Ahn 2008;Roberts et al. 2003), necessitating continued breeding programs.
To select seedlings at the early selection stage after crossing, breeders choose individual rose plants in accordance to major characteristics, such as color, size, and type of flowers; number of petal; and level of prickle expression (Roberts et al. 2003). Although estimating the flower quality using only the first shoot of seedlings in cut roses is fast, it may not be appropriate because floral characteristics change along with the developmental stage and environmental conditions (Yeon and Kim 2017;Yeon and Kim 2020a;Salehi et al. 2005). To analyze the efficiency in the selection of seedlings at an early stage during breeding for new cultivars of cut roses for export, the major floral characteristics were analyzed in our own breeding program.
Materials and Methods
Plant materials
Seedlings of two cross combinations in cut roses were produced by the National Institute of Horticultural and Herbal Science (NIHHS) in May 2019 and then sowed in March 2020. R168 combination (C1, Rosa hybrida ‘Yellow King’ x ‘Aladin’) had 146 lines and SR33 combination (C2, R. hybrida ‘Evelien’ x ‘Sunny Isle’) had 73 lines (Fig. 1). After the first selection of seedlings at NIHHS in March 2020, the seedlings were moved to the experimental greenhouse at the University of Seoul (UOS) and transplanted into round pots (2 L) containing commercial soil for horticultural crops (Baroker, Seoul Bio, Korea) and perlite (2:1). They were cultivated with a hydronic system by irrigating nutrient solution for rose development from UOS (EC 1.0 - 1.1 dS・m-1, pH 5.8).
Flower characteristics
We investigated some ornamental traits that have marketability in cut rose flowers at the fully open stage, such as the number of petals (5 ea: 5 - 9; 10 ea: 10 - 14; 15 ea: 15 - 19; 20 ea: 20 - 24, etc.), flower size (height and width, cm), petal color, prickle expression level on stem (1 to 9), petal size (height x width, cm), scent intensity (1 to 3, sensory test), floral shoot height (cm), and stem diameter (mm). We set a standard for petal color, prickle expression level, and scent intensity by following the characterization for examining new cultivars in Rosa spp. (KSVS 2007). The growth stages of the seedlings were divided into young (six-month-old plants in fall 2020) and mature (13-month-old plants in spring 2021).
Statistical analysis
Student’s t-test, analysis of variance (ANOVA), correlation analysis, and central tendency were performed in SAS 9.4 (SAS Institute, Cray, NC, USA).
Results and Discussion
Phenotypic variation of seedlings in cut roses
The preference for flower forms, doubling petals, size, types for shape, and colors varies by nationality, gender, age, and personality (Kim et al. 2021). The results of the variations in flower forms, types for shape, and colors showed that the growth stage affected the phenotypes of seedlings in cut roses (Fig. 2 and 3). The ratio of flower forms, standard (ST, one flower per stem) and spray (SP, more than 5 flowers per stem), of seedlings was distinguished at the mature plant stage (MPS) (Fig. 2A and B), which resulted from the different vegetative growth between the two growth stages, MPS and young plant stage (YPS). The flower shoot height was about 10 - 15 cm at YPS in the two cross combinations, but it was 33.8 ± 9.0 cm for C1 and 39.2 ± 12.6 cm for C2 (data not shown) at MPS. The flower forms of the parent plants also affected the phenotype of the seedlings, and C2 showed a 5 times higher proportion for SP than C1 (3.4%) because both Rosa hybrida ‘Evelien’ and ‘Sunny Isle’ are spray type (Fig. 2B).
Roses have some phenotypes for the flower shape: classic (typical shape of modern roses), english (full rosette or cup shape, English roses), and tomato (one of phyllody appearance showing a sepal-like structure for stamen and pistil) (Sim et al. 2004). Regardless of the flower shape in the parent plants, tomato-shaped flowers appeared at an occurrence rate of 4% at YPS and then decreased to 0.7 - 1.4% at MPS (Fig. 2C and D). This is because phyllody in roses is caused by an imbalance of phytohormones, environmental stress, damage from insects and viruses, and genetic mutations; and, in general, younger plants have less tolerance for various stress conditions than mature plants (Sim et al. 2004).
Petal color is an important selection target trait for producing marketable new cultivars (Yeon and Kim 2020b). Unlike flower form and the type of shape, the phenotypic variation of petal color was large, although pink and pink-mixed color accounted for 70 - 80% of the total (Fig. 2E and F). During the growth of seedlings in both C1 and C2, which was affected by environmental or endogenous conditions (Yeon and Kim 2016;Yeon and Kim 2020c), observed changes in the proportion of petal colors were greater than expected. Therefore, it is not efficient to select some lines for petal color at YPS because changes in the composition of floral pigments are expected at MPS.
Petal doubling is a significant floral trait in roses. It could affect the attractive appearance and is associated with some genes that form floral organs, such as APETALA1, APETALA2, PISTILLATA, AGAMOUS, and SEPALLATA1 (Ma et al. 2015). KSVS (2007) suggested a standard for petal doubling: single-petal flowers have 5 - 7 petals (ea), semi-double ones have 8 - 19 petals, and double flowers have more than 20 petals. In C1 and C2, 34 - 38.4% of seedlings showed double flowers at YPS, and this increased by 19.8 - 21.9% at MPS (Fig. 3A and B). The variation for petal number in C1 at YPS was higher than at MPS, and the number of petals increased during growth (Table 1). Nevertheless, the median and mode were higher at MPS, with more double-petal flowers, than at YPS. Seedlings of C2 had a lower variation for petal doublings at MPS, but there was no significant difference in the dispersion between the growth stages (Table 2).
The flower size of seedlings became 2.3 times bigger in C1 (Fig. 3C) and 1.2 times bigger in C2 (Fig. 3D) during growth and was similar to the trait of the parent plants. In C2, the difference of coefficient of variance (CV) for flower width between YPS and MPS was smaller than that of C1 (Table 1 and 2) with the parent plants R. hybrida ‘Evelien’ and ‘Sunny Isle’ of C2 having a diameter of 5 - 6 cm (Fig. 3D).
The C1 and C2 showed little scent or scentless flowers (scent intensity: level 1) for 94 - 96% at YPS, and they lost scented characteristics at MPS from 3 - 9% to 0 - 1% (data not shown). Parent plants of C1, R. hybrida ‘Yellow King’ and ‘Aladin’, had distinct scent intensity (level 2), and R. hybrida ‘Evelien’ and ‘Sunny Isle’ of C2 emitted little floral scent. To produce the scented flowers (level 3) by breeding cut roses, breeders should use the strongly scented ones as parent plants and understand the transmitted process of scent traits to their hybrid progeny.
Petal numbers, flower size, and prickle expression on stems are major ornamental factors in roses, but a high prickle expression level makes it difficult to manage the flowers during cultivation and distribution to markets. Although there is demand for a glabrous stem, information on the molecular mechanisms and developments is lacking (Zhou et al. 2020). Unlike the petal number and flower size, regardless of the parent plants the prickle expression of the phenotype at a specific class (level 5) was the highest at YPS, over 67% in C1 and C2 (Fig. 1 and 2). Also, the modes for prickle expression showed no changes in C1 and C2 (Table 1 and 2). Nevertheless, CV and kurtosis were lower, but variation was higher at MPS than at YPS, indicating that the phenotypes for prickle expression were eventually segregated at MPS. Therefore, to select rose plants with prickle-free stems, breeders should care for the seedlings until they reach an appropriate vegetative growth stage.
Correlation for flowering traits between young and mature seedlings in cut roses
We analyzed the correlation of flower traits in accordance with the growth stage to increase efficiency in selecting seedlings at an early stage. Some flower traits, as an essential tool for selecting seedlings, showed various tendencies at an early stage that correlated with important traits at MPS in C1 (Table 3) and C2 (Table 4). The petal numbers of seedlings at YPS showed a strong positive correlation with those at MPS in C1 (r = 0.84***) and C2 (r = 0.73***), and petal size at YPM also had a positive correlation with that at MPS (r = 0.29*** - 0.39***). On the other hand, petal numbers at YPM were negatively correlated with petal numbers at MPS in C1 (r = -0.16) and C2 (r = -0.23**). Prickle expression level at YPM showed a positive correlation with that at MPS in C1 (r = 0.52***) and C2 (r = 0.56***) and with the size of floral shoots at MPS in C1 (r = 0.23**). In conclusion, to select lines for producing new cultivars in roses, breeders should decide the trait target in accordance with the growth stage.