Quantitative analysis of the total main volatile organic compounds in tested species/cultivars by the internal standard method. (A) Wild Rosa species. (B) Chinese old garden roses. (C) Modern roses.
The relative contents of the three groups of volatile compounds in the tested Rosa species/cultivars. (A) Wild Rosa species. (B) Chinese old garden roses. (C) Modern roses.
Diagram showing the emission rates of the main volatile compounds from the tested Rosa species/cultivars. (A) Wild Rosa species. (B) Chinese old garden roses. (C) Modern roses.
Dendrogram resulting from hierarchical cluster analysis and principal component analysis (PCA) of wild Rosa species, Chinese old garden roses, and modern roses. (A) Hierarchical clustering of 59 samples. (B) Two-dimensional score plots of the PCA results.
Comparison of the main components of wild Rosa species, Chinese old garden roses, and modern roses.
Click on author name to view affiliation information
Rose (Rosa L.) is an economically important ornamental genus that has been cultivated for its scent for the perfume industry since antiquity. However, most modern roses have lost their fragrance during the later stages of the breeding process. Here, 59 species of Rosa, including 24 wild Rosa species, 20 Chinese old garden roses, and 15 modern roses, were examined by headspace solid-phase microextraction and gas chromatography–mass spectrometry. Fifty-three volatile organic compounds (VOCs), including terpenoids, benzenoids/phenylpropanoids, and fatty acid derivatives, were detected with qualitative and quantitative analyses. Thirteen common components, including geraniol, citronellol, 2-phenylethanol, 3,5-dimethoxytoluene, 1,3,5-trimethoxybenzene, germacrene D, and cis-3-hexenyl acetate, were found. Furthermore, different wild species or cultivars showed different characteristic compounds. 3,5-Dimethoxytoluene and 1,3,5-trimethoxybenzene were the main compounds in Rosa odorata and Rosa chinensis, which are the original parents of modern roses. 2-Phenylethanol, citronellol, and geraniol were the main aromatic compounds in Rosa damascene and Rosa centifolia. Methyl salicylate, eugenol, methyl eugenol, and benzyl acetate were lost during domestication and breeding of wild Rosa species to Chinese old garden roses and then to modern cultivars. Geranyl acetate, neryl acetate, and dihydro-β-ionol were gained during this time and showed higher amounts across the rose breeding process. Natural and breeding selection may have caused volatile compound gains and losses. These findings provide a platform for mining scent-related genes and for breeding improved ornamental plants with enhanced flower characteristics to develop new essential oil–producing plants.
Roses are used as ornamental plants in gardens, as cut flowers, and as potted flowers but also have economic value as sources of essential oils for perfumes and cosmetics (Yan et al., 2014; Zhou et al., 2020a). The genus Rosa comprises ≈200 species, and just 10–15 species have contributed to the creation of more than 35,000 complex hybrid rose cultivars (Channelière et al., 2002), resulting in a relatively narrow genetic background for these roses (Qiu et al., 2013). Among the original parents of those rose cultivars, 10 species originated from China, including Rosa chinensis var. chinensis, Rosa odorata var. odorata, Rosa rugosa, Rosa multiflora var. multiflora, and Rosa moyesii (Chen, 2001). Breeding efforts have focused on the development of rose cultivars with cold and disease resistance, certain flower forms, and long vase life. However, scent traits have largely been lost during the later stages of the breeding process (Channelière et al., 2002).
Floral fragrance is an evolutionary adaptation of plants to attract pollinators and resist herbivores; it also enhances the aesthetic value of ornamental plants (Dudareva et al., 2004; Scalliet et al., 2008). At present, more than 1700 floral volatile components have been identified from plants, and these components are mainly classified into three categories: terpenes, phenylpropanes, and fatty acid derivatives (Feng et al., 2021; Piechulla and Pott, 2003). Previous studies on the floral scents of roses have mainly focused on oil-bearing and modern roses. R. damascena Mill. was reported to have more than 127 components, mainly including 2-phenylethanol, citronellol, geraniol, and nerol (Kovats, 1987). Recently, the scent compositions and emissions of numerous rose cultivars were reported (Shalit et al., 2004). Aromatic alcohols, monoterpenes, sesquiterpenes, and various esters were found in Fragrant Cloud rose (Guterman et al., 2002). The scented modern rose ‘Jinyindao’ and its spontaneous nonscented rose offspring were evaluated by solid-phase microextraction followed by gas chromatography–mass spectrometry (GC/MS), and the results showed that the relative contents of 1-ethenyl-4-methoxy-benzene and benzothiazole were significantly different between the two roses (Yan et al., 2011). R. odorata and R. chinensis germplasms contained benzodiazepines and sesquiterpenoids as their characteristic substances, whereas 1,3,5-trimethoxybenzene was significantly higher in R. chinensis than in R. odorata, and 3,5-dimethoxytoluene and phenylethanol were not found in R. chinensis (Zhou et al., 2020b).
At present, there are few studies on how wild germplasms and Chinese old garden roses have contributed to the floral scent trait of modern varieties. Here, we present a detailed biochemical analysis of the major volatiles produced from 24 Rosa germplasm resources, 20 old garden cultivars, and 15 modern roses with strong scents. The results provide a theoretical basis for the further research and development of new aromatic varieties and the discovery of floral genes. Characterization of the marked differences in volatile components among these species allowed us to examine the mechanisms by which the metabolic diversity of rose flavor components may have been gained and lost.
Fifty-nine Rosa germplasm resources were used in this study, including 24 wild Rosa species belonging to 9 sections of the subgroup of Rosa (Table 1), 20 Chinese old garden roses, and 15 modern roses with distinct aromatic odors (Supplemental Table 1). The Rosa germplasm resources were collected from their original places and grafted onto R. ‘Natal Briar’ during 2005–13 (Table 1). Then, they were cultivated in the rose germplasm garden of the Flower Research Institute, Yunnan Academy of Agricultural Sciences, Kunming, Yunnan, PR China (Table 1).
Petals from different flowers on the middle branches of adult plants at the blooming stage (petals were spread out flat, but anthers had not released pollen) were collected from 10:00 to 11:00 am from March to August due to the different flowering periods of the resources (Feng et al., 2021). Samples were collected and placed in collection tubes in an ice box and brought to the laboratory for volatile analysis, with three replicates for each variety.
The contents of petal volatile compounds were assayed quantitatively using headspace solid-phase microextraction (HS-SPME) and GC/MS (Yan et al., 2018). One gram of petals was placed into a glass vial for volatile extraction with 10 μl internal standard, and a mixture of 2 μL ethyl caprate (CAS Number 110-38-3, 0.865 μg·μL−1) was dissolved in 1998 μL n-hexane AR (CAS Number 110-54-3). The extraction process lasted for 35 min at 35 °C with a 50-μm divinylbenzene/carboxen/polydimethylsiloxane solid-phase microextraction (SPME) fiber (Supelco, Bellefonte, PA), and the product was detected for 30 min in a 7890B GC system with an 5977A MSD (Agilent Technologies, Santa Clara, CA). The NIST2014 library and Kovats’ retention indices were used to identify the volatiles. The following formula was used: the emission rate of the volatile compound (μg⋅g−1) = (the area of the component/the area of the internal standard) × the concentration of the internal standard (μg⋅μL−1) × the volume of the internal standard (μL)/sample weight (g).
According to the total VOC emission analysis of different stages of rose development using qualitative and quantitative analyses, the results showed a higher emission rate at the blooming stage in section (sect.) Chinensis, sect. Cinnamomeae, sect. Rosa, and sect. Pimpinellifoliae. Lower amounts of VOCs were found for sect. Synstylae, sect. Banksianae, sect. Microphyllae, sect. Laevigatae, and sect. Bracteatae (Fig. 1A). A total of 47 floral volatile compounds were detected by HS-SPME–GC/MS, including 27 terpenoids, 15 aromatic hydrocarbons, and five fatty acid derivatives, which accounted for more than 95% of the total compounds detected (Fig. 2A, Table 1). Thirty-five major components were screened from the 47 volatiles and compared with the standard; these components had a content that exceeded 3 μg·g−1 in more than 3 of the 24 wild species (Fig. 3A).
Citation: HortScience 57, 5; 10.21273/HORTSCI16473-22
Citation: HortScience 57, 5; 10.21273/HORTSCI16473-22
Citation: HortScience 57, 5; 10.21273/HORTSCI16473-22
Each section of the wild species showed obviously different volatiles (Fig. 3A, Supplemental Table 2). For sect. Chinensis, including the R. odorata and R. chinensis complexes, the results showed that the main aromatic components of R. odorata var. odorata were alcohols, and the content of phenylethanol was the highest (Fig. 3A). The main volatile component of R. odorata var. gigatea was aromatic hydrocarbons, and the emission rate of 3,5-dimethoxytoluene was the highest in this species (4274 μg·g−1) (Table 2). The main volatile components of R. odorata var. pseudoindica were aromatic hydrocarbons and terpenes, and the content of 1,3,5-trimethoxybenzene was the highest in this species. For the R. chinensis complex, flowers mainly produced aromatic hydrocarbons, terpenoids, and fatty acid derivatives, of which 1,3,5-trimethoxybenzene was the characteristic aromatic component. There were some common volatile compounds between R. odorata and R. chinensis, such as linalool, methyl eugenol, and ylangene. Furthermore, the main characteristic components were strikingly different among these groups. R. odorata exhibited the unique compounds 3,5-dimethoxytoluene at 4273 μg·g−1 and dihydro-β-ionone at 247 μg·g−1 (Fig. 3A). The emission rates of 1,3,5-trimethoxybenzene and germacrene D were 331 and 1015 μg·g−1 in R. chinensis Jacq. var. spontanea, respectively (Table 2).
R. damascene and R. centifolia play important roles in sect. Rosa. In this study, the results showed that 2-phenylethanol, β-citronellol, and geraniol were the main aromatic components in R. damascene and R. centifolia (Fig. 3A). However, 2-phenylethanol accounted for 49.95% and 27.44% of the total volatiles in the two species, respectively. In addition, the emission rate was nearly 906 μg·g−1 in R. damascene and 899 μg·g−1 in R. centifolia. However, the rate of β-citronellol emission was strikingly different in R. damascene and R. centifolia, at 909 and 421 μg·g−1, respectively (Table 2). Regarding sect. Cinnamomeae, three wild species—R. rugosa, R. beggeriana, and R. willmottiae Hemsl var. glandulifera Yü et Ku.—were examined by SPME-GC/MS analysis. Twenty-five floral components were identified in these species, including 24 major compounds with greater than 3 μg per gram of petals (Fig. 3A). The main compounds were lower alcohols and terpenes in R. beggeriana and R. willmottiae Hemsl var. glandulifera Yü et Ku., whereas R. rugosa showed higher terpenoids, including β-citronellol, geraniol, and 2-phenylethanol, at 4154, 671, and 482 μg·g−1, respectively (Fig. 3A).
Chinese old garden roses have been cultivated by humans since as early as 5000 years ago in China (Wang, 2007) and are used as important rose germplasm resources for breeding cultivars. Old garden roses existed before 1867, the year of the creation of the first modern rose ‘La France’ (Liorzou et al., 2016). In this study, 20 varieties of Chinese old garden roses were examined for their volatile compound profiles by SPME-GC/MS. A total of 46 VOCs were detected, among which 21 were major compounds, including 14 terpenoids, 4 phenylpropanoids, and 3 fatty acid derivatives (Table 3, Fig. 3B). Terpenoids and benzenoids/phenylpropanoids accounted for more than 90% of the total VOCs, and terpenoids were more abundant than benzenoids/phenylpropanoids, showing different compositions from those of wild Rosa species (Fig. 2B). The main floral components identified were geraniol, β-citronellol, germacrene D, dihydro-β-ionol, 3,5-dimethoxytoluene (DMT), 1,3,5-trimethoxybenzene (TMB), phenylethanol, phenethyl acetate, and transrose oxide (Fig. 4B).
Citation: HortScience 57, 5; 10.21273/HORTSCI16473-22
‘Zixiangrong’, ‘Huzhongyue’, ‘Mudanyueji’, ‘Yipinzhuyi’, and ‘Huzhongyue’ contained high amounts of geraniol at 392 μg·g−1, DMT at 76 μg·g−1, citronellol at 63 μg·g−1, 2-phenylethanol at 76 μg·g−1, and germacrene D at 173 μg·g−1 (Supplemental Table 3). The release of total VOCs was in the following order: ‘Zixiangrong’ > ‘Huzhongyue > ‘Mudanyueji’ > ‘Ruanhongxiang’ > ‘Jinfenlian’ > ‘Yushizhuang’ > ‘Old Blush’ (Fig. 1B). Furthermore, the results showed that 85% of the Chinese old garden roses examined were strikingly fragrant, suggesting that Chinese old garden roses are a significant germplasm resource for scent breeding of rose.
Modern roses are used as cut flowers, potted plants, and garden plants, at an annual value of more than $10 billion, and high economic value also lies in the use of their petals as a source of natural fragrances and flavorings. However, most modern cut-flower varieties of rose lack distinct fragrances (Yan et al., 2018), and less than 10% of the 35,000 modern rose cultivars are scented (Yan et al., 2021). In this study, 15 cultivars with strong scents were chosen for evaluation by GC/MS. Benzenoids/phenylpropanoids, terpenoids, and fatty acid derivatives were identified in these roses (Supplemental Table 4). Alcohols represented 52.67% of the total volatiles in ‘Yunxiang’, which contained higher amounts of 2-phenylethanol at 190 μg·g−1 and β-citronellol at 78 μg·g−1 (Fig. 1C). A higher release of phenethyl acetate was detected in ‘Yunshi No. 1’ and ‘Yunxiang’, at 43 and 30 μg·g−1, respectively. Terpenes were the main components in ‘Hongshuangxi’, including citronellol at 54 μg·g−1, geraniol at 43 μg·g−1, and germacrene D at 22 μg·g−1 (Fig. 3C). The three dominant volatile compounds were 3,5-dimethoxybenzene, phenethyl alcohol, and 1,3,5-trimethoxybenzene in ‘Jinyindao’ and ‘Mitang’ (Fig. 2C). 3,5-Dimethoxybezene accounted for 87.29% of the total volatiles in ‘Mitang’, which contained high amounts of this substance at 837 μg·g−1 (Supplemental Table 5).
Hierarchical cluster analysis (HCA) and principal component analysis (PCA) were performed to analyze the correlation between 51 volatile compounds of wild Rosa species, Chinese old garden roses, and modern roses. HCA, using the Pearson correlation as the measurement standard, was used to cluster samples into groups by applying the intergroup connection. The Z score method was used to standardize the related variables to obtain the clustering diagram. The results of HCA are shown in Fig. 4A. HCA clustering showed that there were significant differences among wild Rosa species in VOC composition and quantity. R3 and R9, two wild Rosa species of R. odorata var. gigantea and R. rugosa Thumb., grouped differently based on their remarkably different VOCs. The modern rose R48, ‘Mitang’, was separated into another cluster alone because its VOCs may have been inherited from other wild species or old garden roses. The modern roses R47 and R49, wild Rosa species R11 and R12, and Chinese old garden rose R27 were grouped together. The modern rose R52, wild rose R5, and Chinese old garden rose R26 were grouped together. The modern roses R46, R50, R51, R56, and R58; wild rose R2; and Chinese old garden roses R28, R35, R38, and R41 were grouped together. The modern roses R45, R53, R54, R55, R57, and R59; wild roses R13, R14, R15, and R23; and Chinese old garden roses R29, R34, R37, R40, and R44 were grouped together. The cultivars that clustered into the same group had the same VOCs both in composition and quantity.
In the PCA, two principal components were constructed and explained 19.66% and 13% of the variability; the cumulative contribution was 32.66%. The wild Rosa species were widely dispersed, whereas the Chinese old garden roses and modern roses were relatively concentrated, which was consistent with the HCA results (Fig. 4B).
In the present study, the volatile compounds produced by wild Rosa species, Chinese old garden roses, and modern roses were compared by GC/MS. Thirteen common compounds were detected, including geraniol, β-citronellol, 2-phenylethanol, δ-cadinene, 3,5-dimethoxytoluene, 1,3,5-trimethoxybenzene, phenethyl acetate, and germacrene D. However, some common components were present in much lower amounts in modern roses than in wild Rosa species and Chinese old garden roses (Table 2), such as 3,5-dimethoxytoluene at 4273 and 688 μg·g−1 in R. odorata var. gigantea and pseudoindica, respectively. However, the compound 3,5-dimethoxytoluene is quite low in modern roses at 33.5–139.8 μg·g−1. In addition, phenylpropanoid-related compounds such as 2-phenylethanol contributed greatly to the typical rose scent and were significantly different among the tested roses. These compounds were found at levels above 500 μg·g−1 in R. damascene, R. centifolia, and R. rugosa Thunb. and at only 7.9–245 μg·g−1 in most modern roses.
Wild Rosa species exhibited 17 main unique compounds, including methyl salicylate, eugenol, methyleugenol, and benzyl acetate, but these compounds were lost in modern roses. On the other hand, geranyl acetate, neryl acetate, and dihydro-β-ionol showed higher levels in old garden roses and modern roses than in wild species, suggesting that they were gained in the process of rose cultivar breeding (Fig. 5).
Citation: HortScience 57, 5; 10.21273/HORTSCI16473-22
The genus Rosa contains ≈200 species, and 95 species of wild Rosa are widely distributed in China (Ku and Robertson, 2003; Wissemann and Ritz, 2005). Yunnan is rich in biodiversity and is known for its plants (Jian et al., 2013). Fifty-eight species/varieties of Rosa occur in Yunnan (Brichet, 2003; Ku and Robertson, 2003). Wild Rosa species from Yunnan played very important roles in the creation of modern roses by providing useful traits, especially tea scents (Scalliet et al., 2008). In the present study, the VOC compositions and quantities of wild Rosa species were first detected. Compared with the scent composition of Chinese old garden roses and modern cultivars, the results showed that sect. Chinensis, sect. Rosa, and sect. Cinnamomeae contain higher amounts of fragrant compounds, including DMT, germacrene D, TMB, and phenylethanol, suggesting that these three sections contributed scent traits to old garden and modern cultivars. Furthermore, the VOC results are in agreement with the genetic relationships found in a previous study (Ku and Robertson, 2003; Qiu et al., 2013).
3,5-Dimethoxytoluene and 1,3,5-trimethoxybenzene are the major scent compounds of R. odorata var. gigantea and R. chinensis Jacq. var. spontanea. The two compounds contribute to the characteristic “tea scent.” DMT can represent up to 90% of the total flower volatiles in R. gigantea, and TMB represents up to 56% of the volatiles in R. chinensis Jacq. var. spontanea; these roses have fragrances with earthy and spicy notes reminiscent of black tea (Scalliet et al., 2008; Zhou et al., 2020a). The study consistently confirmed that Sect. Chinensis played important roles in the history of modern rose breeding, especially regarding fragrance traits (Zhou et al., 2020b). In sect. Rosa, R. damascene, the damask rose, is the most important species used to produce rose water, essence of rose, and essential oils in the perfume industry (Knudsen et al., 2006). In this study, the content of 2-phenylethanol was the highest in R. damascena and represents up to more than half of the total volatiles, suggesting that R. damascena contributed the typical scent characteristic to modern rose. In addition, R. rugosa has been shown to have much higher amounts of 2-phenylethanol, geraniol, nerol, citronellol, and their derivatives. R. rugosa has a long history of cultivation for perfume production since at least the early 1800s in China (Feng et al., 2010). In the breeding process, R. rugosa contributed to the typical fragrance of roses due to its unique characteristic terpenoids citronellol and geraniol.
For flowers, scent volatiles attract pollinators and are used as defense compounds against animals and microorganisms (Atkinson, 2018; Prasad et al., 2004). In this study, we compared the volatile compositions among wild Rosa species, Chinese old garden roses, and modern cultivars and found strikingly different compounds. Many researchers have reported the large diversity of secondary metabolites that are emitted by plants and their differences between wild species and cultivars (Hanson et al., 1996; Hartmann, 1996; Schwab, 2003). Aharoni et al. (2004) identified markedly different volatile terpenoid flavor components between wild and cultivated strawberry species. Tadmor et al. (2002) found that different volatile compounds negatively affect tomato fruit aroma and that these compounds were selected against during wild species domestication. Hartmann (1996) holds the view that secondary metabolites evolved under the selection pressure of a competitive environment. In this study, the loss of methyl salicylate, eugenol, and methyl eugenol in modern roses might be related to a shift from wild environments and suggested that wild species released specific compounds to survive in adverse environments, to aid in plant reproduction, and to prevent attacks from pests and diseases. Sensitivities to diseases and insects, as well as a lack of fragrance, are some typical shortcomings of modern rose cultivars (Tang et al., 2008, Zhang et al., 2009). The results indicated that the shortcomings of modern cultivars might be due to the loss of these secondary metabolites. Furthermore, the pleasant fragrances geranyl acetate, neryl acetate, and dihydro-β-ionol showed higher levels in these cultivars. The main reason for this result might be that one of the cultivars selected by breeders had VOCs with mild and pleasant aromas that were beneficial to humans. Taken together, our data, along with those in the literature, suggest that the gain and loss of volatile compounds of secondary metabolites are the result of both natural and breeding selection.
In this study, the volatile compounds produced by wild Rosa species, Chinese old garden roses, and modern roses were analyzed by GC/MS. The results showed that different wild species/cultivars exhibited characteristic compounds. Methyl salicylate, eugenol, methyl eugenol, and benzyl acetate were lost during domestication and breeding from wild Rosa species to Chinese old garden roses and modern cultivars. Geranyl acetate, neryl acetate, citronellyl acetate, and dihydro-β-ionol were gained and produced higher amounts during the rose breeding process. We suggest that natural and breeding selection may have caused volatile compound gains and losses.
Quantitative analysis of the total main volatile organic compounds in tested species/cultivars by the internal standard method. (A) Wild Rosa species. (B) Chinese old garden roses. (C) Modern roses.
The relative contents of the three groups of volatile compounds in the tested Rosa species/cultivars. (A) Wild Rosa species. (B) Chinese old garden roses. (C) Modern roses.
Diagram showing the emission rates of the main volatile compounds from the tested Rosa species/cultivars. (A) Wild Rosa species. (B) Chinese old garden roses. (C) Modern roses.
Dendrogram resulting from hierarchical cluster analysis and principal component analysis (PCA) of wild Rosa species, Chinese old garden roses, and modern roses. (A) Hierarchical clustering of 59 samples. (B) Two-dimensional score plots of the PCA results.
Comparison of the main components of wild Rosa species, Chinese old garden roses, and modern roses.
Contributor Notes
This study was funded by the National Key Research and Development Program of China (Grant Number 2018YFD1000400) and the Natural Science Foundation of China (Grant Numbers 32060693 and 31872144), Major Agricultural Science and Technology Projects in Yunnan Province (Grant Number 202102AE090052), and the Green Food Brand-Build a Special Project (Floriculture) supported by Science and Technology (Grant Number 530000210000000013742). We thank the English Language Department for correcting and proofreading this document.
K.T. and H.Y. are the corresponding authors. E-mail: kxtang@hotmail.com or hjyan8203@126.com.
Quantitative analysis of the total main volatile organic compounds in tested species/cultivars by the internal standard method. (A) Wild Rosa species. (B) Chinese old garden roses. (C) Modern roses.
The relative contents of the three groups of volatile compounds in the tested Rosa species/cultivars. (A) Wild Rosa species. (B) Chinese old garden roses. (C) Modern roses.
Diagram showing the emission rates of the main volatile compounds from the tested Rosa species/cultivars. (A) Wild Rosa species. (B) Chinese old garden roses. (C) Modern roses.
Dendrogram resulting from hierarchical cluster analysis and principal component analysis (PCA) of wild Rosa species, Chinese old garden roses, and modern roses. (A) Hierarchical clustering of 59 samples. (B) Two-dimensional score plots of the PCA results.
Comparison of the main components of wild Rosa species, Chinese old garden roses, and modern roses.
