Analysis of Floral Volatile Components and Antioxidant Activity of Different Varieties of Chrysanthemum morifolium

This study investigated the volatile flavor compounds and antioxidant properties of the essential oil of chrysanthemums that was extracted from the fresh flowers of 10 taxa of Chrysanthemum morifolium from three species; namely Dendranthema morifolium (Ramat.) Yellow, Dendranthema morifolium (Ramat.) Red, Dendranthema morifolium (Ramat.) Pink, Dendranthema morifolium (Ramat.) White, Pericallis hybrid Blue, Pericallis hybrid Pink, Pericallis hybrid Purple, Bellis perennis Pink, Bellis perennis Yellow, and Bellis perennis White. The antioxidant capacity of the essential oil was assayed by spectrophotometric analysis. The volatile flavor compounds from the fresh flowers were collected using dynamic headspace collection, analyzed using auto thermal desorber–gas chromatography/mass spectrometry, and identified with quantification using the external standard method. The antioxidant activities of Chrysanthemum morifolium were evaluated by DPPH and FRAP assays, and the results showed that the antioxidant activity of each sample was not the same. The different varieties of fresh Chrysanthemum morifolium flowers were distinguished and classified by fingerprint similarity evaluation, principle component analysis (PCA), and cluster analysis. The results showed that the floral volatile component profiles were significantly different among the different Chrysanthemum morifolium varieties. A total of 36 volatile flavor compounds were identified with eight functional groups: hydrocarbons, terpenoids, aromatic compounds, alcohols, ketones, ethers, aldehydes, and esters. Moreover, the variability among Chrysanthemum morifolium in basis to the data, and the first three principal components (PC1, PC2, and PC3) accounted for 96.509% of the total variance (55.802%, 30.599%, and 10.108%, respectively). PCA indicated that there were marked differences among Chrysanthemum morifolium varieties. The cluster analysis confirmed the results of the PCA analysis. In conclusion, the results of this study provide a basis for breeding Chrysanthemum cultivars with desirable floral scents, and they further support the view that some plants are promising sources of natural antioxidants.


Introduction
Chrysanthemum (Chrysanthemum morifolium Ramat.) belongs to the Asteraceae family of leading ornamental species, second only to the rose in terms of its market value [1]. Dendranthema morifolium (Ramat.) and Pericallis hybrida, called Qiuju and Fuguiju in China, are widely used in South China, either as a fragrant floral tea or as an anti-inflammatory herb in Traditional Chinese Medicine. They have diseases, including inflammation [23] and aging and age-related diseases [25]. Importantly, free radical damage on the skin caused by ROS and UV-irradiation stress plays a key role in photoaging [26,27].
The present study was conducted to obtain clear inter-species classification by using multivariate analysis methods, such as PCA and cluster analysis, as well as by analyzing and comparing the volatile compounds of 10 taxa of three Chrysanthemum species. Moreover, the antioxidant activities of the three species were evaluated by DPPH and FRAP assay.

Optimization of ATD Parameters
The duration of desorption is an important parameter and its optimization can improve desorption efficiency and sensitivity. Therefore, three desorption times are selected for testing: 15, 30, and 45 min. As shown in Figure 1, a duration of desorption of 30 min, 18 compounds were detected and the content of hydrocarbons, aromatics, esters were relatively high. While for 45 min, 15 compounds were detected and the content of all compounds were low except the content of terpenoids. This phenomenon can be explained by the low volatility in addition to their thermal sensibility [28]. Therefore, a 30-min duration was confirmed as the optimization time condition. Tenax-TA and Tenax-GR are polymeric materials (styrene polyvinylbenzene) that are commonly used as an adsorbent for volatile compounds, especially for use with high temperatures. As can be seen in Figure 1, there is no distinctive difference, so the entire experiment path used Tenax-TA as an adsorbent for volatile compounds and the sampling time of 30 min. Importantly, free radical damage on the skin caused by ROS and UV-irradiation stress plays a key role in photoaging [26,27]. The present study was conducted to obtain clear inter-species classification by using multivariate analysis methods, such as PCA and cluster analysis, as well as by analyzing and comparing the volatile compounds of 10 taxa of three Chrysanthemum species. Moreover, the antioxidant activities of the three species were evaluated by DPPH and FRAP assay.

Optimization of ATD Parameters
The duration of desorption is an important parameter and its optimization can improve desorption efficiency and sensitivity. Therefore, three desorption times are selected for testing: 15, 30, and 45 min. As shown in Figure 1, a duration of desorption of 30 min, 18 compounds were detected and the content of hydrocarbons, aromatics, esters were relatively high. While for 45 min, 15 compounds were detected and the content of all compounds were low except the content of terpenoids. This phenomenon can be explained by the low volatility in addition to their thermal sensibility [28]. Therefore, a 30-min duration was confirmed as the optimization time condition. Tenax-TA and Tenax-GR are polymeric materials (styrene polyvinylbenzene) that are commonly used as an adsorbent for volatile compounds, especially for use with high temperatures. As can be seen in Figure 1, there is no distinctive difference, so the entire experiment path used Tenax-TA as an adsorbent for volatile compounds and the sampling time of 30 min.

Comparison of Antioxidant Activities
Two common in vitro assays were used to evaluate the antioxidant activities, and the results showed that the three species had significantly different antioxidant activity depending on the assay used (Table 1). In the DPPH assay, all of the extracts showed good radical scavenging activity. Samples of Dendranthema morifolium had a lower mean EC50 than that of Bellis perennis and a higher mean EC50 than that of Pericallis hybrida, but no significant difference was found (p > 0.05). In the FRAP assay, the antioxidant activity of Bellis perennis was significantly stronger than that of Dendranthema morifolium and Pericallis hybrida (p < 0.05) ( Table 1). The results indicated that essential oil extracted from Chrysanthemum morifolium exhibited DPPH free radical scavenging activity, and therefore it could be applied as an antioxidant agent.

Analysis of Volatile Components and Release Quantity in Different Varieties of Chrysanthemum
The presence, yield, and composition of secondary metabolites, such as volatile oils in plants, can be affected by several factors, including physiological variations, environmental conditions,

Comparison of Antioxidant Activities
Two common in vitro assays were used to evaluate the antioxidant activities, and the results showed that the three species had significantly different antioxidant activity depending on the assay used (Table 1). In the DPPH assay, all of the extracts showed good radical scavenging activity. Samples of Dendranthema morifolium had a lower mean EC50 than that of Bellis perennis and a higher mean EC50 than that of Pericallis hybrida, but no significant difference was found (p > 0.05). In the FRAP assay, the antioxidant activity of Bellis perennis was significantly stronger than that of Dendranthema morifolium and Pericallis hybrida (p < 0.05) ( Table 1). The results indicated that essential oil extracted from Chrysanthemum morifolium exhibited DPPH free radical scavenging activity, and therefore it could be applied as an antioxidant agent. DPPH IC50 (mgml −1 ) 0.108 ± 0.009 0.119 ± 0.0010 0.072 ± 0.003 0.080 ± 0.005 0.082 ± 0.001 FRAP C (Fe 2+ ) (mmolg −1 ) 3 3.090 ± 0.022 ** 3.209 ± 0.019 * 1.870 ± 0.009 * 1.990 ± 0.010 * 1.807 ± 0.007 * Note: Values for Chrysanthemum morifolium, which differ and were statistically significantly, are marked with "*". Analysis of variance was performed between various groups of Chrysanthemum morifolium. * p < 0.05. ** p < 0.01. Antioxidant activity expressed as mmol FeSO 4 equivalents per 1 g sample.

Analysis of Volatile Components and Release Quantity in Different Varieties of Chrysanthemum
The presence, yield, and composition of secondary metabolites, such as volatile oils in plants, can be affected by several factors, including physiological variations, environmental conditions, geographic variations, genetic factors, and evolution [29].
A typical GC-MS total ion chromatogram (TIC) of the volatile chemical profile of Chrysanthemum is shown in Figure 2.   The volatiles of the 10 taxa are listed in Table 2. GC-MS analysis detected 36 compounds in the floral scent of Chrysanthemum, which were identified by their mass spectra and RI. Based on the classification of aromatic compounds of Chrysanthemum, volatile compounds were categorized as: terpenoid compounds, hydrocarbons, aromatic compounds, alcohols compounds, and flavor compounds. Hydrocarbons and terpenoids predominated, while the other compounds were typically present in smaller amounts. The 36 detected compounds included 10 hydrocarbons, 10 terpenoids, 4 aromatics, 4 alcohols, 3 ketones, 3 esters, 1 aldehydes, and 1 ether (Table 2). Camphor, n-hexane, ethyl acetate, 1,3,5-cycloheptatriene, ethyl benzene, o-xylene, eucalyptol, nonane, alpha-Pinene, and pentadecane were the most common components, and camphor, n-hexane, o-xylene, and eucalyptol were present at particularly high levels. 5  The volatiles of the 10 taxa are listed in Table 2. GC-MS analysis detected 36 compounds in the floral scent of Chrysanthemum, which were identified by their mass spectra and RI. Based on the classification of aromatic compounds of Chrysanthemum, volatile compounds were categorized as: terpenoid compounds, hydrocarbons, aromatic compounds, alcohols compounds, and flavor compounds. Hydrocarbons and terpenoids predominated, while the other compounds were typically present in smaller amounts. The 36 detected compounds included 10 hydrocarbons, 10 terpenoids, 4 aromatics, 4 alcohols, 3 ketones, 3 esters, 1 aldehydes, and 1 ether ( Table 2). Camphor, n-hexane, ethyl acetate, 1,3,5-cycloheptatriene, ethyl benzene, o-xylene, eucalyptol, nonane, alpha-Pinene, and pentadecane were the most common components, and camphor, n-hexane, o-xylene, and eucalyptol were present at particularly high levels.
The volatile compound contents varied markedly among the 10 taxas. In Nos. 1, 4, 6, 8, 9, and 10, camphor (4.19 µg, 1.29 µg, 9.53 µg, 2.70 µg, 3.76 µg, and 4.06 µg) was the highest component. In No.2, endo-Borneol (6.01 µg) was the most abundant, and it was only detected in this cultivar. In No.6, the content of the volatile compound benzaldehyde (2.67 µg) was the highest component. Even though camphor was found to be the main volatile compound in the three species, the other volatile compounds were reported differently according to the taxa. There were seven same compounds in the ten taxa. As shown in Figure 3, the floral volatile components were significantly different among Chrysanthemum samples. The amount of terpenoids (51%) was the highest in No. 1 , and 10 were similar. This also showed that the content from a species of Chrysanthemum morifolium was similar.

PCA of Volatile Compounds
The main component analysis was used to explore the data matrix order; this can help determine the weight (importance) of parameters in the total variability through vector size and loads and respective percent contributions. The PCA results give three significant principal

PCA of Volatile Compounds
The main component analysis was used to explore the data matrix order; this can help determine the weight (importance) of parameters in the total variability through vector size and loads and respective percent contributions. The PCA results give three significant principal components (eigenvalues >1, represents the main variables), which explain 96.509% of the variation in the data (55.802%, 30.599%, and 10.108%, respectively).
The volatile compounds were evaluated and coded by each graded value to classify the 10 taxa of Chrysanthemum from the 3 species. Figure 4 shows the eigenvalues and their contribution through PCA using hydrocarbons, terpenoids, aromatic compounds, alcohols, ketone, ether, aldehyde, and ester volatile compounds of 10 taxa of Chrysanthemum. PCA is a method that can be used to identify patterns in a data set and to reduce dimensionality of multivariate data by removing inter-correlations among variables [30,31]. We obtained the eigenvalues of the matrix consisting of a component of the variance and covariance parameter. In the case of the first principal component (PC1), the eigenvalue of each of the characteristics was 5.580, thereby showing a 55.802% contribution to the total variation. The second and third PCs had eigenvalues of 3.060 and 1.011, respectively, and the degree of contribution to the total variation was 30.599% and 10.108%, respectively.

Cluster Analysis of Compositions of Volatile Compounds
To compare volatile compound compositions among the 10 taxa of Chrysanthemum from 3 species, we performed a hierarchical cluster analysis based on the contents of the 36 aromatic volatile compounds. We used Ward's method for between-group linkage and the squared Euclidean distance between clusters as a proximity measurement. The 10 Chrysanthemum formed three clusters in the dendrogram ( Figure 5  In addition, to estimate the characteristics of the various volatile compounds, correlations between the individual PCs were analyzed. PC1 (55.802% of variance) was accounted for by variables with the highest loads: hydrocarbons (1.676), terpenoids (1.535) compounds, ketones (−0.713), and aldehyde compounds (−0.674). PC2, which explained 30.599% of the variation, was accounted for by the following variables: terpenoids (1.535), alcohols (0.403), and hydrocarbons (−1.736). PC3, which explained 10.108% of the variation, was accounted for by the following variables: alcohols (1.863) and aldehyde compounds (−1.274). This can be interpreted as a gradient of variability among variables; positive signs mean that variables increase together, while a negative sign means that as one variable increases, the other variable decreases. It should also be noted that the sign of the terpenoids variable is the opposite of those of analysis doses because these are derived from the former (they are reciprocals).
As a result of arranging the values of the first three PCs on a 3-dimensional scatter diagram for 10 taxa of the Chrysanthemum species, these were categorized into three groups (Figure 4

Cluster Analysis of Compositions of Volatile Compounds
To compare volatile compound compositions among the 10 taxa of Chrysanthemum from 3 species, we performed a hierarchical cluster analysis based on the contents of the 36 aromatic volatile compounds. We used Ward's method for between-group linkage and the squared Euclidean distance between clusters as a proximity measurement. The 10 Chrysanthemum formed three clusters in the dendrogram (

Discussion
We investigated for the first time the volatile flavor compounds and antioxidant properties of essential oil that was extracted from fresh flowers, namely Dendranthema morifolium (Ramat.) Yellow, Dendranthema morifolium (Ramat.) Red, Dendranthema morifolium (Ramat.) Pink, Dendranthema morifolium (Ramat.) White, Pericallis hybrida Blue, Pericallis hybrida Pink, Pericallis hybrida Purple, Bellis perennis Pink, Bellis perennis Yellow, and Bellis perennis White. Dendranthema morifolium (Ramat.), Pericallis hybrid, and Bellis perennis. All had antioxidant activity, but our study had limitations so we were not able to further define the details of the bioactivity that is responsible for the observed effects. There were 36 volatile compounds that were identified with 8 functional groups: hydrocarbons, terpenoids, aromatics, alcohols, ketones, ethers, aldehydes, and esters. In Nos. 1, 4, 6, 8, 9, and 10, the content of camphor (4.19 µg, 1.29 µg, 9.53 µg, 2.70 µg, 3.76 µg, and 4.06 µg) was the highest component. In No.2, the content of endo-Borneol (6.01 µg) was most abundant, and it was only detected in this cultivar. In No.6, the content of the volatile compound benzaldehyde (2.67 µg) was the highest component. Thus, we determined that the chemotaxonomy technique using the compositions of the volatile compounds of the Chrysanthemum species would be useful in the classification and identification of the verified inter-species differences. Finally, this study provides the basic materials needed for species selection and cultivation of Chrysanthemum species, which are useful for food, cosmetics, and medicine, further supporting the view that the flowers are promising sources of natural antioxidants. The underlying antioxidant mechanisms and potential side effects of flowers also warrant further exploration and experimentation.

Plant Materials
The 10 varieties of Chrysanthemum morifolium used in this study were purchased from local markets and maintained by laboratory personnel in the nursery at Xinjiang Academy of Forestry, (Urumqi, China) ( Table 3). They were identified by Associate Professor Zhiyou Guo (Department of Life Sciences, Qiannan Normal University for Nationalities, Duyun, China). Alpha-Pinene standard solution (>99.9% purity) was diluted by methanol (HPLC purity) to the concentration range of 1 µg/mL-10 µg/mL. Then, a microliter injector (Anting micro injector factory, Shanghai, China) was used to get 0.5 µL standard solution injected into Tenax TA (PerkinElmer, Boston, MA, USA); it was plugged with glass cotton in order to stop the alpha-Pinene from evaporating. Each concentration was prepared 3 times for the test.

Extraction of Volatile Flavor Compounds
Simultaneous steam distillation (SDE) was performed following Schultz's extraction method to extract the volatile flavor compounds [32]. Two grams of the freeze-dried leaf sample and 500 mL of distilled water were added to one of the flasks, and heated-reflux was performed for 2 h on a 100 • C heating mantle. Meanwhile, the essential oils were extracted by adding 50 mL of pentanediethyl ether mixture in the other flask and heated-refluxing to 40 • C. The extracted volatile flavor compounds were dehydrated using anhydrous sodium sulfate. Afterwards, they were filtered using filter paper and concentrated using 99.9% nitrogen gas. Finally, the concentrated compounds were dissolved in 100 µL of acetone, and they were analyzed using gas chromatography/mass selective detector (GC/MSD) as below.

Sampling Collection
A single fresh flower from Chrysanthemum morifolium was collected in a sampling bag (355 mm × 355 mm, Reynolds, Jefferson, MO, USA). The sampling collection steps included: (a) an air sampler was used to remove the air in the sampling bag; (b) an atmosphere sampler was used to fill the air passed through the activated carbon into the sampling bag; (c) When the content of the air in the sampling bag was close to three-quarters filled, the emission volatile matter from Chrysanthemum morifolium was collected by active sampling onto the sorbent Tenax TA. Tenax-TA was chosen as it corresponds to a polymeric material (styrene polyvinylbenzene); it is commonly used as an adsorbent for semi-volatile compounds, especially for use with high temperature. The entire gas path used tasteless silicone tubing. The sampling time was fixed at 10:00-12:00 a.m. The sampling instrument volume flow rate was 100 mL min −1 , with different sampling times (15,30, and 45 min). Each variety was sampled 3 times during the same testing session. Empty bags served as the control samples.

Analytical Instrumentation
For the ATD, a two-stage desorption was adopted to reduce the component bandwidths and improve the efficiency of the chromatographic separation. Thermal primary desorption of the sampling tubes was carried out at 250 • C with a helium flow rate of 25 mL min −1 for 20 min in order to maintain conditions strictly similar to the online sampling. The outlet split was also fixed to 5 mL min −1 . The cold trap was maintained at −30 • C. Again, during secondary desorption, the cold trap was rapidly heated from −30 • C to 300 • C, and it was maintained at this temperature for 5 min. Analytes were then injected onto the capillary column via a transfer line that was heated to 250 • C. The outlet-split flow was adjusted for high-resolution capillary column to 10 mL/min, leading to a fraction of 50% of the sample entering the column [33].
ATD-GC-MS: automatic thermal desorption (ATD 350, PerkinElmer Corp., Norwalk, CT, USA) with gas chromatography (GC) coupled with a mass selective detector (MSD). The chromatograph was a 7890A Network GC System interfaced with a 5975C Network MSD (Agilent Technologies, Santa Clara, CA, USA), and the capillary column was a Thermo TR-5ms (60 m × 0.25 mm; film thickness of 0.25 µm). The MS detector provided acquisition in the full-scan mode or selected ion monitoring (SIM) mode. The ion energy of electron impact ionization was 70 eV, with the ion source temperature set to 250 • C. Helium was used as the carrier gas at 1.5 mL/min (with regulated constant flow). The GC oven was set at 40 • C for 2 min, followed by 6 • C/min to 280 • C, with a final extension for 3 min at 280 • C, and a total acquisition program of 45 min. Mass spectral data were acquired over a mass range of 29-500 amu for the full-scan mode. The qualitative identification of targeted compounds was based on retention times. Quantification of the extracted ions was performed using the external standard method.

Antioxidant Capacity Analysis
DPPH assay was performed according to the Brand-Williams method [34], which was slightly modified by Kim et al. [35]. In brief, 2 mL of Chrysanthemum morifolium extract solution and 2 mL of DPPH (200 µmol/L) were mixed and the absorbance was measured after 30 min at 517 nm at room temperature using an ultraviolet spectrophotometer and the effective concentration. The effective concentration (EC50) value was determined for the antioxidants. The FRAP assay was determined according to Li's method [36]. The FRAP assay measures the ability of the antioxidants in Chrysanthemum morifolium extracts to reduce the ferric tripyridyl-triazine (Fe 3+ -TPTZ) complex to the blue-colored ferrous form (Fe 2+ ), which absorbs light at 593 nm. Briefly, a sample extract (0.5 mL) was mixed with 3.0 mL of ferric-TPTZ reagent and added to 10 mL PV tubes. The tubes were incubated at 37 • C for the duration of the reaction. The absorbance was read at 593 nm at 75 min with an ultraviolet spectrophotometer (UV-2600/2700, Shimadzu, Japan).

Statistical Analysis
Based on the results of the GC/MSD analysis of the volatile compounds in the 10 taxa of Chrysanthemum species, ANOVA and multivariate analyses were performed using IBM's SPSS 19.0 software (IBM SPSS Institute, ver. 19.0, Chicago, IL, USA).