Metabolomics Analysis Reveals the Accumulation Patterns of Flavonoids and Volatile Compounds in Camellia oleifera Petals with Different Color

To systematically and comprehensively investigate the metabolic characteristics of coloring substances and floral aroma substances in Camellia oleifera petals with different colors, ultrahigh-performance liquid chromatography–mass spectrometry (UPLC–MS/MS) and headspace solid phase microextraction and gas chromatography–mass spectrometry (HS–SPME–GC–MS) metabolomics methods were applied to determine the metabolic profiles of white, candy-pink and dark-red petals. The results revealed that 270 volatile organic compounds were detected, mainly terpenoids, heterocyclic, esters, hydrocarbons, aldehydes, and alcohols, in which phenylethyl alcohol, lilac alcohol, and butanoic acid, 1-methylhexyl ester, hotrienol, alpha-terpineol and 7-Octen-4-ol, 2-methyl-6-methylene-, (S)-, butanoic acid, 2-methyl-, 2-methylbutyl ester, 2,4-Octadienal, (E,E)- could act as the floral scent compounds. A total of 372 flavonoid compounds were identified, and luteolin, kaempferol, cyanidin and peonidin derivatives were considered as the main coloring substances for candy-pink and dark-red petal coloration. In conclusion, this study intuitively and quantitatively exhibited the variations in flower color and floral scent of C. oleifera petal with different colors caused by changes in variations of flavonoids and volatile organic compound composition, and provided useful data for improving the sensory quality and breeding of C. oleifera petals.


Introduction
Camellia oleifera, belonging to Camellia in Theaceae, is a unique woody oil plant in China with high oil content [1].C. oleifera flowers are complete flowers with high ornamental value [2].The petals are generally white, and a few are pink or red, with unique delicate fragrance and sweetness [3].C. oleifera petals are rich in phenolic compounds and volatile compounds, including phenolic acids, flavonoids, alcohols, ketones, aromatic hydrocarbons, esters and aldehydes, which are the most important secondary metabolites [4,5].In addition, C. oleifera petals are abundant in biochemical components such as amino acids, proteins, tea polysaccharides and saponins, which have significant antioxidant, lipid-lowering, anti-allergy, hypoglycemic, gastrointestinal protection and immunity-enhancement functions [6].
Petal color is the typical target of research in ornamental plants, as flowers attract pollinators, absorb ultraviolet (UV) radiation, and decorate the environment [7,8].Petals' color diversity is mainly caused by the pigment types and contents [9].Pigment compounds are widely distributed in plant tissues, and the three main types controlling petal color are flavonoids, carotenoids and betaines [10,11].Camellia oleifera petals have been the subject of recent studies due to the plant's status as an ornamental plant [12].The pigments in Camellia petals are mainly flavonoids (especially anthocyanins), which are contributed to the formation of pink and red petals [7,13].Cyanidin-core structure pigments (such as cyanidin 3,5-di-O-glucoside) have the potential to produce the most dominant phenotype among wild red-flowered Camellia species in China [14].Cyanidin-3-O-(6 -O-malonyl) glucoside was the main anthocyanin component in the Camellia japonica petals, while cyanidin-3-O-rutinoside, peonidin-3-O-glucoside, cyanidin-3-Oglucoside, and pelargonidin-3-O-glucoside were responsible for the color intensity of the C. japonica petals [7].Further research on these pigments will have important guiding significance for exploring the molecular regulation mechanism of flower color change in C. oleifera.
Petal color and fragrances in C. oleifera are very diverse, and flower color and scent have always been the important characters in the selection process [23,24].However, the research on C. oleifera is mainly concentrated in seedling-raising techniques, cultivation and management techniques, stress resistance study and so on.The investigations on characteristics of key coloring substances and VOCs in C. oleifera petals have not been reported.Here, ultrahigh-performance liquid chromatography-mass spectrometry (UPLC-MS/MS) and headspace solid phase microextraction and gas chromatography-mass spectrometry (HS-SPME-GC-MS) were employed to identify and quantify the differential flavonoids and VOCs in C. oleifera petals with different color.The complete chemical characterization of flavonoids and VOCs in C. oleifera petals was established to understand the petals' color and fragrance variations.The results laid a metabolic foundation for further revealing the flower color and the VOCs variations in C. oleifera petals, and provided valuable information for understanding the flower color and scent change mechanism and metabolic pathways in C. oleifera petals.

Analysis of Metabolite Profiling in C. oleifera Petals with Different Color
To further understand differences of flavonoids and volatile profiling in C. oleifera petals with different colors, the flavonoids and VOCs were determined by UPLC-MS/MS and HS-SPME-GC-MS.Pearson's correlation coefficient r was applied to evaluate the biological repeatability of each group of the petal samples.When r 2 was close to 1, the correlation between the two repeated samples was stronger, which indicated that the metabolite data had good homogeneity (Figure 1A).

Analysis of Metabolite Profiling in C. oleifera Petals with Different Color
To further understand differences of flavonoids and volatile profiling in C. oleifera petals with different colors, the flavonoids and VOCs were determined by UPLC-MS/MS and HS-SPME-GC-MS.Pearson's correlation coefficient r was applied to evaluate the biological repeatability of each group of the petal samples.When r 2 was close to 1, the correlation between the two repeated samples was stronger, which indicated that the metabolite data had good homogeneity (Figure 1A).Based on the quality evaluation, there were significant differences in the levels of VOCs among the W, CP, and DR petals (Figure 1B).A total of 270 VOCs were tentatively detected and identified, including 49 terpenoids, 47 heterocyclic compounds, 43 esters, 30 hydrocarbons, 27 aldehydes, 23 alcohols, 17 ketones, 8 amines, 6 phenols, 6 nitrogen compounds, 5 acids, 4 aromatics, 3 sulfur compounds, and 2 others (Table S1), which indicated that terpenoids, esters, heterocyclic compound, hydrocarbons, and alcohol were the main VOCs in C. oleifera petals.Moreover, the relative content of VOCs in W petals was significantly different from that in CP and DR petals.Cluster analysis of the detected petal samples and metabolites indicated that there were significant differences in VOC accumulation patterns in the three petals samples, indicating that there were significant differences in the composition of the three petals (Figure 1B).
Qualitative and quantitative analysis of flavonoid metabolites in W, CP, and DR petals by UPLC-MS/MS was carried out, and a total of 372 flavonoid metabolite species were detected and quantified, including 26 proanthocyanidins, 8 biflavones, 38 tannins, 22 flavanols, 83 flavonols, 85 flavonoids, 27 anthocyanins, 11 dihydroflavonols, 27 flavanones, 6 aurones, 23 chalcones (including C-glucosylquinochalcones), and 9 others (Table S2).Nine samples could be clearly divided into three groups based on the HCA heatmap (Figure 1C).Interestingly, a clear separation was found between W, CP, and DR petals, demonstrating that the accumulation of flavonoids in the three petals were obviously distinct.The above statistical results demonstrated that all petal samples had good repeatability, and the data of flavonoid metabolites were reliable, and were suitable for further qualitative and quantitative analysis.

PCA and OPLS-DA of the Three C. oleifera Petals with Different Colors
In the 2D-PCA score plot, the results exhibited that the flavonoid and VOCs metabolites in three petals displayed a clear separation trend between groups, and there was a tight aggregation trend within groups.It could be seen that the cumulative contribution rate of two principal components (PC1 66.28% × PC2 19.46%) reached 85.74% (Figure 2A).W, CP, and DR petals were obviously separated, and the biologically repeated same petals were closely grouped.These results indicated that the volatile profiling data had high repeatability, which was convenient for further analysis.
The OPLS-DA model was applied to compare the differential accumulation metabolites (DAMs) in the three C. oleifera petals with different colors.When high predictability (Q 2 ) > 0.9 and the goodness of fit was strong (R 2 X, R 2 Y was close to 1), the model was considered to be excellent and stable.Here, the high Q 2 , R 2 X, and R 2 Y were obtained to evaluate the validity of OPLS-DA model between the three C. oleifera petals.All the flavonoids and VOCs in three C. oleifera petals were accessed to determine the difference in the W_vs_CP comparison (Q 2 = 0.989, R 2 X = 0.874, and R 2 Y = 1.000; Figure 2B), CP_vs_DR comparison (Q 2 = 0.965, R 2 X = 0.787, and R 2 Y = 1.000; Figure 2C), and W_vs_DR comparison (Q 2 = 0.997, R 2 X = 0.848, and R 2 Y = 1.000; Figure 2D) based on the OPLS-DA model through pairwise comparison.The results of OPLS-DA analysis and cross validation demonstrated that the Q 2 of the three comparison groups was greater than 0.9, indicating that the model was accurate and stable, which explained the variations of flavonoids and VOCs in the three C. oleifera petals and could be performed to further screen the DAMs among different comparison groups using VIP values.

Differential Accumulation Metabolites of Volatile Organic Compounds Analysis of C. oleifera Petals 2.3.1. Differentially Accumulated Metabolites of Volatile Organic Compounds
To screen the differential accumulation metabolites of volatile organic compounds (VOCs-DAMs) among three C. oleifera petals, the 270 VOCs metabolites were selected based on the one-dimensional analysis t test (p < 0.05), multidimensional analysis of VIP value (VIP > 1) of OPLS-DA, and fold change of ≥2 or ≤0.5, and there were 180 VOCs-DAMs in the W_vs_CP_vs DR comparison (Table S3).There were 162 VOCs-DAMs (74 up-regulated, and 88 down-regulated) in the W_vs_CP comparison (Table S4), 149 VOCs-DAMs (79 up-regulated, and 70 down-regulated) in the W_vs_DR comparison (Table S5), 50 VOCs-DAMs (21 up-regulated, and 29 down-regulated) in the CP_vs_DR comparison (Table S6).Next, VOCs-DAMs in the three comparison groups (W_vs_CP, W_vs_DR, and CP_vs_DR) were classified into 11, 12, and 10 different categories, respectively, and the most common VOCs-DAMs were esters, hydrocarbons, terpenoids, and alcohols.In addition, 24 VOCs-DAMs were identified among the three groups, indicating that these 24 VOCs-DAMs were differentially accumulated among the three C. oleifera petals (Figure 3A).
Molecules 2023, 28, x FOR PEER REVIEW 6 of 17 CP_vs_DR) were classified into 11, 12, and 10 different categories, respectively, and the most common VOCs-DAMs were esters, hydrocarbons, terpenoids, and alcohols.In addition, 24 VOCs-DAMs were identified among the three groups, indicating that these 24 VOCs-DAMs were differentially accumulated among the three C. oleifera petals (Figure 3A).Enrichment analysis of VOCs-DAMs in three C. oleifera petals samples was performed by KEGG to obtain comprehensive functional information, and Most of the VOCs-DAMs were categorized into metabolic pathways, biosynthesis of secondary metabolites, alpha-Linolenic acid metabolism, biosynthesis of various alkaloids, and biosynthesis of various plant secondary metabolites (Figure 3B-D).
Overall, the contents of phenylethyl alcohol, lilac alcohol, and butanoic acid, 1-methylhexyl ester with rose, fruity, and honey aroma attributes in white C. oleifera petals were significantly higher than those in candy-pink and dark-red petals, which might make the white petals richer in aroma.The hotrienol, alpha-terpineol and 7-Octen-4-ol, 2-methyl-6-methylene-, (S)-with sweet and delicate fragrance flavor exhibited a higher accumulation level in candy-pink C. oleifera petals.The hotrienol, butanoic acid, 2-methyl-, 2-methylbutyl ester, and 2,4-Octadienal, (E,E)-with wood and fruity flavor were the primary VOCs in the dark-red C. oleifera petals.

Differential Flavonoid Compounds Analysis of C. oleifera Petals 2.4.1. Differentially Accumulated Metabolites of Flavonoid Compounds
C. oleifera petals were found to be rich in flavonoids, and their flavonoid contents were markedly influence by color genotypes.In our study, the differentially accumulated metabolites (DAMs) between pairwise comparisons among W_vs_CP, W_vs_DR, and CP_vs_DR were screened by the variable importance in projection values (VIP) ≥ 1 and fold change ≥ 2 or fold change ≤ 0.5.The W_vs_CP comparison and W_vs_DR comparison had the largest number of up-regulated and down-regulated DAMs (Figure 5).Among these comparisons, there were 194 DAMs (99 up-regulated and 95 down-regulated) in the W_vs_CP comparison (Figure 5A), 197 DAMs (100 up-regulated and 97 down-regulated) in the W_vs_DR comparison (Figure 5B), and 61 DAMs (14 up-regulated and 47 down-regulated) in the CP_vs_DR comparison (Figure 5C), respectively.KEGG pathway enrichment analyses were carried out to gain further insights into the biochemical pathway to which the DAMs belonged to.The top three enriched KEGG pathways between the three comparisons were anthocyanin biosynthesis, flavonoid biosynthesis, and flavone and flavonol biosynthesis (Figure 5D-F).Given the role of anthocyanins and flavonoids in petal coloration, we deduced that the DAMs in anthocyanin biosynthesis pathway and flavonoid biosynthesis pathway might be likely the key metabolites underlying the variations in C. oleifera petals.

Discussion
Camellia chekiangoleosa with a bright flower color and beautiful flower type, and Camellia yuhsienensis with a heavy floral scent were the excellent garden greening and woody oil tree species.Flower color and floral fragrance, as important ornamental traits of plants, are the main indicators to identify and distinguish different varieties, and affect the ornamental value and economic value of plants [8,25,26].However, the regulation mechanism on the formation of flower color and fragrance between different flower color varieties of C. oleifera is still unclear.
Flavonoids are one of the main pigment components involved in the formation of flower color [27].The difference of composition and anthocyanin content directly affects the flower color of plants [28].Previous studies have confirmed that cyanidin and its derivatives widely acted on red petals of plants [29].For example, it was found that the content of cyanidin in Rhododendron simsii Planch.with red flower colors was the highest.Jin et al. (2018) had found that cyanidin was the main flavonoid component in rosa crimson glory, and its content was significantly higher than other substances [30]; it has also been found that the main pigment in the Rosa rugosa × Rosa Sertata was cyanidin-3-glucose [31].Peonidin-3-O-glucoside and kaempferol derivatives were mainly detected from the red rapeseed petals [9].The white petals of rosa glaucca pourr contained only flavonoids, while the pink petals and purple petals of rosa glaucca pourr contained flavonoids and anthocyanins [32].White Rosa multiflora petal and white chrysanthemum petals only contain light yellow or near colorless pigments, such as flavonoids and flavonols [33].Therefore, the difference of flavonols, flavones, and anthocyanin content will direct influence on plant petals color.Here, the qualitative and quantitative analysis of flavonoids and anthocyanins in white, candy-pink, and dark-red color petals of three C. oleifera cultivars was identified and detected by UPLC-MS/MS in this study.In the white petals, luteolin-4 -Oglucoside, kaempferol-7-O-glucoside, luteolin-7-O-glucoside, quercetin-3-O-robinobioside, and kaempferol-3-O-galactoside were the main pigments in white petals, which was in accordance with the results that luteolin and kaempferol were the main substances that determined the white color of 'Rosa alba' [34].In the candy-pink and dark-red petals, cyanidin-3-O-glucoside, cyanidin-3-O-galactoside, cyanidin-3-O-rutinoside, cyanidin-3-O-(6 -O-malonyl)glucoside, and peonidin-3-O-glucoside displayed higher accumulation levels, which was consistent with this conclusion that those above pigments were the main anthocyanin components in red petals of C.japonica [35].Moreover, the combination of flavonoids and other related co-pigment compounds with anthocyanins exhibited a hyperchromic effect, and the co-color effects also increased with the increase in anthocyanin methylation and glycosylation, of which the flavonol and flavonoids are the most common co-pigments [36].Luteolin-7-O-gentiobioside and isosakuranetin-7-O-rutinoside were found at higher accumulation levels in candy-pink and dark-red petals, respectively, which might be combined with cyanidin and peonidin derivatives to form a co-pigmentation effect to make the petals pinker and redder.Therefore, luteolin, kaempferol, cyanidin, and peonidin derivatives were acted as the main pigments for C. oleifera petal coloration.
Floral scent is a mixture of chemical compounds emitted by plant tissues, which is a crucial factor that affects the overall aroma and consumer preference [37].The petal flavor is strongly influenced by the type and amount of aroma components present [38].VOCs are the main components of flowers, and the flowers of different plant varieties contain different VOCs [39].For example, the main VOCs in 'Gesang Lv' were linalool, caryophyllene, oxidized linalool, and pinene, the main VOCs in 'Gesang Hong' included macrophyllene D, methyl decanoate, hexol, and pinene, and the main VOCs in 'Gesang Huang' were linalool, caryophyllene, methyl decanoate, and germacrene [40].The volatile components of the flowers within the genus Chimonanthus were mainly terpenes and esters, and terpenoids and aromatic hydrocarbons were the main VOCs in rose [41].A total of 270 volatile organic compounds were identified using HS-SPME in conjunction with GC-MS.Esters, hydrocarbons, terpenoids, and alcohols were the dominant aroma components in three C. oleifera petals in terms of quantity and proportion.Phenylethyl alcohol was the main aroma component, widely existing in plant essential oils, which is an important aroma compound, with rose, fruity, and honey flavor [42].Here, phenylethyl alcohol was the main VOCs in the white petals, which was in accordance with the results that phenylethyl alcohol was the unique compound of 'Fragrant cloud' varieties and phenylethyl alcohol was the highest aroma components in Rosa.odorata var odorata [43].The hotrienol and alpha-terpineol were detected with a high content in candy-pink petals.In accordance with our results, Liu et al. ( 2016) identified hotrienol and alpha-terpineol as the important aroma compounds in red freesia flower [44].Zhang Ming et al. (2020) detected as main aroma constituent butanoic acid, 2-methyl-, 2-methylbutyl ester in Areca catechu flower [45], in which we detected a high accumulation of butanoic acid, 2-methyl-, 2-methylbutyl ester in the dark-red C. oleifera petal.Different volatile organic compounds and release amount were together contributed to the unique aroma of C. oleifera petals, which not only increased the ornamental value, but also had important significance for the commercial development of aromatic C. oleifera petals.In summary, phenylethyl alcohol, hotrienol, alpha-terpineol, and butanoic acid, 2-methyl-, 2-methylbutyl ester were generally present in white, candy-pink, and dark-red C. oleifera petal, which mainly contributed to the perfumery value.
Our understanding of the floral scent traits of C. oleifera petals with different colors is limited.A comprehensive knowledge of the coloring substances and volatile compounds in C. oleifera petals is the first step in cultivating colorful and aroma varieties.Esters, alcohols and terpenoids are ubiquitous in floral volatiles.Functional studies of genes involved in the biosynthesis of terpenoids would be very prospective.Finally, with the support of biotechnology tools, metabolic engineering methods for flower color and fragrancerelated products could be realized.Therefore, our research not only would be helpful to provide comprehensive information about the characteristic components of flower color and fragrance in C. oleifera petals with different colors, but also lay a fundamental reference for cultivating excellent C. oleifera varieties with specific flower color and fragrance.

Materials
Three C. oleifera varieties with different-colored petals were planted in the Shaanxi C. oleifera germplasm repository with similar soil conditions in Nanzheng District, Shaanxi Province, China.The petal colors of three C. oleifera varieties, namely "Camellia yuhsienensis" (White, W), "Camellia semiserrata" (Candy-pink, CP), and "Camellia chekiangoleosa" (Darkred, DR) were white, candy-pink and dark-red, respectively (Figure 7).Three C. oleifera petals samples were collected from fresh petals during full flowering, with three biological replicates on 8 March 2023.Then, the petals were immediately stripped, put into liquid nitrogen, and brought back to the laboratory for storage at −80 • C for the flavonoid and volatile organic compound determination.The mass spectrograms corresponding to each chromatographic peak were compared with NIST05 and NIST05s standard spectrum library (NIST Mass Spectral Database 2.2) to identify the volatile organic compounds in three C. oleifera petals.Each sample was repeated three times.The MassHunter quantitative software (version B.08.00, Agilent Technologies Inc., Santa Clara, CA, USA) was carried out to integrate and correct the chromatographic peaks, and peak area normalization method was applied to determine the relative content of each component after detection by GC-MS.The peak area of each chromatographic peak represented the relative content of the corresponding metabolite.

Qualitative and Quantitative Analyses of Metabolites
Orthogonal partial least squares discriminant (OPLS-DA) analysis was conducted on the total metabolites between samples of different groups and metabolites within samples of different groups to investigate the differences of flavonoid metabolites among the petals of three C. oleifera petals.The screening standards for differential metabolites were (1) Variable importance in projection (VIP), VIP > 1; (2) significance threshold, p < 0.05; (3) fold change ≥ 2 or fold change ≤ 0.5.Through searching the KEGG (Kyoto Encyclopedia of Genes and Genomes) database, functional annotation analysis and metabolic pathway enrichment analysis were performed on metabolites with significantly different contents obtained from metabolomics analysis.

Statistical Analysis
Chemometric analyses such as hierarchical cluster analysis (HCA), correlation analysis (CA) and orthogonal partial least squares discriminant analysis (OPLS-DA) were performed to systematically analyze the difference of the flavonoid compounds and volatile organic compounds in three C. oleifera petals, which were generated by using R (http://www.r-project.org/(accessed on 20 September 2023)).Variance of flavonoid compounds and volatile organic compounds among three petals was calculated and generated by using SPSS 26.0 for Windows (SPSS Inc., Chicago, IL, USA).The violin plots and histograms were drawn and generated by using the Origin Pro 2023 for statistics and computing (Origin Lab, Northampton, MA, USA).

Conclusions
In our study, the metabolic profiles of coloring substances and volatile organic compounds in three C. oleifera petals with different colors were systematically evaluated to explore the differences in metabolites based on the UPLC-MS/MS and HS-SPME-GC-MS approach.The results demonstrated that a total of 372 flavonoid compounds and 270 volatile organic compounds were detected in the C. oleifera petals, which visually revealed how changes in the composition of flavonoid compounds and volatile organic compounds affected the overall flower color and floral scent of C. oleifera petals.Among them, phenylethyl alcohol, lilac alcohol, and butanoic acid, and 1-methylhexyl ester were the major flavor substances in white C. oleifera petals; hotrienol, alpha-terpineol and 7-Octen-4-ol, 2-methyl-6-methylene-, (S)-were the main flavor substances in candy-pink C. oleifera petals; and hotrienol, butanoic acid, 2-methyl-, 2-methylbutyl ester, and 2,4-Octadienal, (E,E)-were the primary flavor substances in the dark-red C. oleifera petals.Luteolin and kaempferol derivatives might be the key coloring substances contributed to white color formation, and cyanidin and peonidin derivatives were considered as the main coloring substances for candy-pink and dark-red petal coloration.In summary, this study not only investigated the key metabolites that controlled the flower color and fragrance of C. oleifera petals, but also aided in evaluating the metabolic quality of C. oleifera petals by laying a solid foundation for further cultivating C. oleifera varieties with a specific color and strong fragrance.

Supplementary Materials:
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/molecules28217248/s1,Table S1: A total of 270 volatile organic compounds tentatively detected and identified from three C. oleifera petal; Table S2: A total of 372 flavonoid metabolite tentatively detected and identified from three C. oleifera petal; Table S3: Differential accumulation metabolites of volatile organic compounds in the W_vs_CP_vs DR comparison; Table S4: Differential accumulation metabolites of volatile organic compounds in the W_vs_CP comparison; Table S5: Differential accumulation metabolites of volatile organic compounds in the W_vs_DR comparison; Table S6: Differential accumulation metabolites of volatile organic compounds in the CP_vs_DR comparison.

Figure 1 .
Figure 1.Differential petals chemotype in C. oleifera petals with different color, W = white petals, CP = candy-pink petal, DR = dark-red petal.(A): Correlation analysis of three C. oleifera petals based on identified VOCs and flavonoids; (B): Heatmap of the 270 VOCs identified in three C. oleifera petals

Figure 1 .
Figure 1.Differential petals chemotype in C. oleifera petals with different color, W = white petals, CP = candy-pink petal, DR = dark-red petal.(A): Correlation analysis of three C. oleifera petals based on identified VOCs and flavonoids; (B): Heatmap of the 270 VOCs identified in three C. oleifera petals with three biological replicates; (C): Heatmap of flavonoids identified in three C. oleifera petals with three biological replicates.The relative content values of all metabolites were denoted with a unique color, among which red color indicated a high accumulation level, and green color indicated a low accumulation level.

Figure 2 .
Figure 2. The 2D-PCA plot and OPLS-DA plot of the three C. oleifera petals based on the relative content of flavonoids and VOCs.(A): The 2D-PCA score plot of W, CP, and DR petals; (B): Score plots of the OPLS-DA model for W_vs_CP; (C): Score plots of the OPLS-DA model for CP_vs_DR; (D): Score plots of the OPLS-DA model for W_vs_DR.W = white petals; CP = candy-pink petal; DR = dark-red petal.

Figure 2 .
Figure 2. The 2D-PCA plot and OPLS-DA plot of the three C. oleifera petals based on the relative content of flavonoids and VOCs.(A): The 2D-PCA score plot of W, CP, and DR petals; (B): Score plots of the OPLS-DA model for W_vs_CP; (C): Score plots of the OPLS-DA model for CP_vs_DR; (D): Score plots of the OPLS-DA model for W_vs_DR.W = white petals; CP = candy-pink petal; DR = dark-red petal.

Figure 4 .
Figure 4. Violin plots of peak areas values of 12 crucial differential VOCs identified in W, CP, and DR petals.W = white petals; CP = candy-pink petal; DR = dark-red petal.The distribution and probability density of the 12 VOCs-DAMs were represented by a combination of box plots and density plots.The outer shapes represented the density of the peak area values distribution, the black rectangular box in the middle represents the quartile range, and the white circle in the middle represents the median.

Figure 4 .
Figure 4. Violin plots of peak areas values of 12 crucial differential VOCs identified in W, CP, and DR petals.W = white petals; CP = candy-pink petal; DR = dark-red petal.The distribution and probability density of the 12 VOCs-DAMs were represented by a combination of box plots and density plots.The outer shapes represented the density of the peak area values distribution, the black rectangular box in the middle represents the quartile range, and the white circle in the middle represents the median.

4. 2 .
UPLC-MS/MS Analysis 4.2.1.Petal Preparation and Extraction Fresh petal samples were freeze-dried, and pulverized to powder with a grinder (MM 400, Retsch, Haan, Germany) at 30 Hz for 1.5 min.For each sample, 100 mg of powder was extracted in 1.0 mL of methanol.The extracts were vortexed once every 30 min for 30 s 6 times, and then put in the refrigerator at 4 °C overnight.After the sample