Quantitative and Qualitative Analysis of Flavonoids and Phenolic Acids in Snow Chrysanthemum (Coreopsis tinctoria Nutt.) by HPLC-DAD and UPLC-ESI-QTOF-MS

A simple, accurate and reliable high performance liquid chromatography coupled with photodiode array detection (HPLC-DAD) method was developed and then successfully applied for simultaneous quantitative analysis of eight compounds, including chlorogenic acid (1), (R/S)-flavanomarein (2), butin-7-O-β-d-glucopyranoside (3), isookanin (4), taxifolin (5), 5,7,3′,5′-tetrahydroxyflavanone-7-O-β-d-glucopyranoside (6), marein (7) and okanin (8), in 23 batches of snow chrysanthemum of different seed provenance and from various habitats. The results showed total contents of the eight compounds in the samples with seed provenance from Keliyang (Xinjiang, China), are higher than in samples from the other five provenances by 52.47%, 15.53%, 19.78%, 21.17% and 5.06%, respectively, which demonstrated that provenance has a great influence on the constituents in snow chrysanthemum. Meanwhile, an ultra performance liquid chromatography coupled with electrospray ionization and quadrupole time-of-flight-mass spectrometry (UPLC-ESI-QTOF-MS) was also employed to rapidly separate and identify flavonoids and phenolic acids in snow chrysanthemum from Keliyang. As a result, a total of 30 constituents, including 26 flavonoids and four phenolic acids, were identified or tentatively identified based on the exact mass information, the fragmentation characteristics, and retention times of eight reference standards. This work may provide an efficient approach to comprehensively evaluate the quality of snow chrysanthemum.


Introduction
Coreopsis tinctoria Nutt. (Asteraceae), native to North America, is widely cultivated around the world [1]. The capitulum of Coreopsis tinctoria is called "snow chrysanthemum" (Xue Ju) or "Kunlun snow chrysanthemum" (Kun Lun Xue Ju) in China [2]. The whole plant has been employed in China for hundreds of years as a folk herb for heat-clearing and detoxifying , while the capitulum of C. tinctoria was taken as a drink to deter diabetes in Portugal [3]. Since the beginning of this century, it has been known that snow chrysanthemum was used by the Uyghur nationality in Xinjiang (China) to prevent cardiovascular and cerebrovascular diseases, such as hypoglycemia, hypolipidemia, and good antioxidant activities [4][5][6]. Phytochemical studies have shown that snow chrysanthemum mainly contains flavonoids, phenolic acids, and terpenoids [7]. Flavanomarein and marein, the predominant flavonoids, were proved to have good hypoglycemic activities [3]. Meanwhile, satisfactory antioxidant activites were also reported for some of the other flavonoids, such as flavanokanin, okanin, coreopsin and flavanocorepsin [8][9][10].
The effects on cardiovascular and cerebrovascular diseases was first discovered in C. tinctoria from Keliyang, Xinjiang [11]. Since then, a large crop area has been dedicated to the cultivation of snow chrysanthemum in Keliyang (Xinjiang) and it is considered a Xinjiang-specific medicinal material. The good hypoglycemic and hypolipidemic effects of C. tinctoria have attracted wide attention in recent years [12][13][14], but the excessive planting has brought a series of problems. The influx of seeds from different sources, including inter-mating of seeds from Keliyang, as well as imported seeds, and the expansion to different areas of Xinjiang have led to unevenness of the quality of snow chrysanthemum on the market. In general, the quality of C. tinctoria was traditionally evaluated by the color and size of the flowers, and seeds from Keliyang were considered as superior quality, and the quality of snow chrysanthemum was better with the higher altitude.
In recent years there was some research to verify this notion. There were reports showing that the chemical composition of snow chrysanthemum was related to the ecological environment [15,16], however due to the limited samples studied and the lack of major compounds this idea was not convincing. A recently reported study about the antioxidant activities and chemical characteristics of different parts from snow chrysanthemum made the point that the contents of flavonoids and phenolic acids of snow chrysanthemum collected from Mt. Kunlun were higher than those from other regions [8,17]. To our knowledge, there is no report of relationship between seed provenance of C. tinctoria and its quality.

Optimization of the Extraction Method
To optimize the extraction procedure, samples (0.5 g) were extracted for 30 min using six different extraction solvents (30%, 70%, and 100% methanol and 30%, 70%, and 95% ethanol). Comparing the peak areas of the HPLC chromatograms, 30% ethanol gave poor extraction efficiency for the eight target compounds, while 70% methanol showed the best extraction effiency for all of them. Besides, ultrasonic extraction and heat reflux were also compared and the results indicated that these two methods had similar performance. Additionally, different extraction times (30,45, and 60 min) and extraction cycles (1-3 times) were tested, too. As a result, optimal extraction condition for the active ingredients of snow chrysanthemum was determined to be one cycle of 30 mins of ultrasonic extraction with 70% methanol [18,19].

Validation of the Method Calibration Curves, Limits of Detection and Limits of Quantification
The calibration curves were generated from six different concentrations of the standard working solutions. Each concentration was analyzed in triplicate. The limits of detection (LOD) and limits of quantification (LOQ) were measured on the basis of the corresponding concentrations at signal-to-noise ratios of 3 and 10, respectively. All analytes showed good linearity of the correlation curves with correlation coefficients better that 0.9990. The LOD and LOQ values of the eight analytes varied from 3.55 to 9.25 ng and 12.46 to 35.88 ng, respectively, indicating a good sensitivity (Table 1). Precision, Stability, Repeatability and Accuracy The intra-day and inter-day precision were determined by analyzing the six replicates of the eight target standard solutions on the same day and three consecutive days, measured by calculating the relative standard deviation (RSDs) of the contents of all analytes. The RSDs of the precision of intra-day and inter-day analyses ranged from 1.76% to 4.18% and 1.63% to 4.26%, respectively (Table 2).
For the stability test, the same sample was analyzed at 0, 2, 4, 8, 12, and 24 h at room temperature. The RSDs of the peak areas of the eight constituents in the same sample ranged from 1.24% to 3.22%, indicating that the sample was stable under the experimental conditions (Table 2).
In order to examine the recovery of the extraction method, high amounts of the standards were added into a sample that was analyzed as described in Sections 3.2.1 and 3.3.1. Recovery was calculated as follows: The results showed that the mean recovery ranged from 95.75% to 103.50% with RSDs less than 4.88% for the eight compounds (Table 2).
For the stability test, the same sample was analyzed at 0, 2, 4, 8, 12, and 24 h at room temperature. The RSDs of the peak areas of the eight constituents in the same sample ranged from 1.24% to 3.22%, indicating that the sample was stable under the experimental conditions (Table 2).
In order to examine the recovery of the extraction method, high amounts of the standards were added into a sample that was analyzed as described in Sections 3.2.1 and 3.3.1. Recovery was calculated as follows: The results showed that the mean recovery ranged from 95.75% to 103.50% with RSDs less than 4.88% for the eight compounds (Table 2).
Among the four flavone glycosides, flavanomarein and marein, which both exhibited various good biological activities in vitro, accounted for a large percentage of the flavonoids ( Table 3). As shown in Table 3, through the 23 batches of samples, the content of flavanomarein or marein was mostly highest in all samples with a seed source from Keliyang, and the the contents of both compounds represented more than 52% of the investigated compounds. Thus, flavanomarein and marein could plays a key role in the quality assessment of snow chrysanthemum. Among the four flavone glycosides, flavanomarein and marein, which both exhibited various good biological activities in vitro, accounted for a large percentage of the flavonoids ( Table 3). As shown in Table 3, through the 23 batches of samples, the content of flavanomarein or marein was mostly highest in all samples with a seed source from Keliyang, and the the contents of both compounds represented more than 52% of the investigated compounds. Thus, flavanomarein and marein could plays a key role in the quality assessment of snow chrysanthemum.
Due to the different seed provenance and habitat, there were remarkable differences between the samples in terms of the total contents of the investigated compounds which ranged from 73.99 mg/g-136.38 mg/g, as shown in Table 3. In addition, the contents of four flavone glycosides, including (R/S)-flavanomarein (2), butin-7-O-β-D-glucopyranoside (3), 5,7,3 ,5 -tetrahydroxyflavanone-7-O-β-D-glucopyranoside (6), and marein (7), accounted 74.46%-83.69% of the eight investigated compounds in all tested samples. As shown in Figure 3A, there was significant difference in the total contents of the eight compounds between the six snow chrysanthemum seed sources. According to the folk tradition, snow chrysanthemum with seed provenance from Keliyang is considered to be the genuine herbal medicine, which is supported by the fact that the total contents of the eight compounds in snow chrysanthemum with Keliyang provenance seed were much higher than that of the four other seed source samples (p < 0.05). There was no difference in the total contents of the eight compounds between Keliyang-1 and Keliyang-2 provenance. Meanwhile, the contents of eight compounds in three other seed sources of snow chrysanthemum (seeds from Ningxia, Jiangsu, and Xinjiang) also showed no difference. As the major contributors of the components, the contents of flavone glycosides followed the same rule as the total components ( Figure 3B).  As shown in Figure 3A, there was significant difference in the total contents of the eight compounds between the six snow chrysanthemum seed sources. According to the folk tradition, snow chrysanthemum with seed provenance from Keliyang is considered to be the genuine herbal medicine, which is supported by the fact that the total contents of the eight compounds in snow chrysanthemum with Keliyang provenance seed were much higher than that of the four other seed source samples (p < 0.05). There was no difference in the total contents of the eight compounds between Keliyang-1 and Keliyang-2 provenance. Meanwhile, the contents of eight compounds in three other seed sources of snow chrysanthemum (seeds from Ningxia, Jiangsu, and Xinjiang) also showed no difference. As the major contributors of the components, the contents of flavone glycosides followed the same rule as the total components ( Figure 3B).  On the other hand, the total contents of the eight compounds in snow chrysanthemum showed no significant regularity in planting in the four different habitats. The total contents of eight compounds, flavonoid glycosides and two major flavonoid glycosides in samples from the habitat of Minfeng (Xinjiang) with higher altitude were not higher than those from other three habitats. Also the total contents of the eight compounds in samples from the habitat of IMPLAD (Beijing) with lower altitude were also not lower than those from three other higher altitude habitats ( Figure 3C). Therefore, the idea that the higher the attitude, the better the quality remains to be confirmed. What is interesting, however, is that in Figure 3D, the total contents of flavones aglycones, including isookanin (4), taxifolin (5), and okanin (8) in snow chrysanthemum planted in Minfeng (Xinjiang) were typically higher than that in samples from IMPLAD (Beijing, China) (p < 0.01). That may be caused by the different habitats, but there were less differences between the samples planted in Minfeng, Xinjiang and Sihai, Beijing (p > 0.05).
Besides, there were two batches of seeds from inter-mating from seeds of Keliyang. One was amphichrome, and the other had red flowers. In most snow chrysanthemums, the outside of the ligulate flower is yellow, and the center is red, so the flowers are basically amphichrome. The phenomomenon that the whole of flower is red is also seen in a few individuals.
For consumption as a tea, the completely red flowers are considered the most desirable, not the amphichrome ones. In the study, the red flower samples were selected from the autologous breeding of the same batch of amphichrome plants. The similar genetic background made them comparable. According to report, the red parts of flower are due to the presence of anthocyanins which have the same upstream metabolic pathways as flavonoids [20]. As shown in Figure 4, the contents of seven flavonoids in amphichrome samples (S5, S11, and S17) were comparatively higher than those in red flower ones (S6, S12, and S18) in this study. This may be explained by the fact that more anthocyanins were synthesized in red flowers and less flavonoids were synthesized in the biosynthetic pathway. On the other hand, the total contents of the eight compounds in snow chrysanthemum showed no significant regularity in planting in the four different habitats. The total contents of eight compounds, flavonoid glycosides and two major flavonoid glycosides in samples from the habitat of Minfeng (Xinjiang) with higher altitude were not higher than those from other three habitats. Also the total contents of the eight compounds in samples from the habitat of IMPLAD (Beijing) with lower altitude were also not lower than those from three other higher altitude habitats ( Figure 3C). Therefore, the idea that the higher the attitude, the better the quality remains to be confirmed. What is interesting, however, is that in Figure 3D, the total contents of flavones aglycones, including isookanin (4), taxifolin (5), and okanin (8) in snow chrysanthemum planted in Minfeng (Xinjiang) were typically higher than that in samples from IMPLAD (Beijing, China) (p < 0.01). That may be caused by the different habitats, but there were less differences between the samples planted in Minfeng, Xinjiang and Sihai, Beijing (p > 0.05).
Besides, there were two batches of seeds from inter-mating from seeds of Keliyang. One was amphichrome, and the other had red flowers. In most snow chrysanthemums, the outside of the ligulate flower is yellow, and the center is red, so the flowers are basically amphichrome. The phenomomenon that the whole of flower is red is also seen in a few individuals.
For consumption as a tea, the completely red flowers are considered the most desirable, not the amphichrome ones. In the study, the red flower samples were selected from the autologous breeding of the same batch of amphichrome plants. The similar genetic background made them comparable. According to report, the red parts of flower are due to the presence of anthocyanins which have the same upstream metabolic pathways as flavonoids [20]. As shown in Figure 4, the contents of seven flavonoids in amphichrome samples (S5, S11, and S17) were comparatively higher than those in red flower ones (S6, S12, and S18) in this study. This may be explained by the fact that more anthocyanins were synthesized in red flowers and less flavonoids were synthesized in the biosynthetic pathway. S6 S11 S12 S17 S18

Identification of Flavonoids and Phenolic Acid in Snow Chrysanthemum by UPLC-MS
Optimized chromatographic conditions, including mobile phase, elution program and spectrometer conditions were determned through trials in this study. As a result, a linear gradient elution with acetonitrile and water containing formic acid as the mobile phase gave the best peak resolution. The parameters of flow rate of gas, gas pressure, spray voltage, capillary temperature and voltage of entrance were optimized to obtain appropriate ionization. Figure 5 presents a typical chromatogram of a sample with mass spectrometric detection in negative ion mode using the optimum condition. A total of 30 compounds (Figures 5 and 7), including four phenolic acids, 11 flavanones, six chalcones, two flavones, four flavonols and three aurones were identified in the snow chrysanthemum samples. Among them, eight compounds were unambiguously identified by comparing their retention time, the accurate masses and fragment ions with those of reference

Identification of Flavonoids and Phenolic Acid in Snow Chrysanthemum by UPLC-MS
Optimized chromatographic conditions, including mobile phase, elution program and spectrometer conditions were determned through trials in this study. As a result, a linear gradient elution with acetonitrile and water containing formic acid as the mobile phase gave the best peak resolution. The parameters of flow rate of gas, gas pressure, spray voltage, capillary temperature and voltage of entrance were optimized to obtain appropriate ionization. Figure 5 presents a typical chromatogram of a sample with mass spectrometric detection in negative ion mode using the optimum condition. A total of 30 compounds (Figures 5 and 7), including four phenolic acids, 11 flavanones, six chalcones, two flavones, four flavonols and three aurones were identified in the snow chrysanthemum samples. Among them, eight compounds were unambiguously identified by comparing their retention time, the accurate masses and fragment ions with those of reference compounds and 22 compounds were tentatively assigned by matching the empirical molecular formula with that of the known compounds previously reported in the literature (Table 4). compounds and 22 compounds were tentatively assigned by matching the empirical molecular formula with that of the known compounds previously reported in the literature (Table 4).   [21], peaks 18, 19 and 22 belonged to caffeoyl quinic acids and were tentatively identified as 1,3-dicaffeoylquinic acid, 3,5-dicaffeoylquinic acid, 3,4-dicaffeoylquinic acid, respectively.  compounds and 22 compounds were tentatively assigned by matching the empirical molecular formula with that of the known compounds previously reported in the literature (Table 4).   [21], peaks 18, 19 and 22 belonged to caffeoyl quinic acids and were tentatively identified as 1,3-dicaffeoylquinic acid, 3,5-dicaffeoylquinic acid, 3,4-dicaffeoylquinic acid, respectively.

Identification of Flavonoids in Snow Chrysanthemum by ESI-Q-TOF-MS
The Fragmentation Rules of Standards The ion-fragmentation pathways of flavonoids often show a retro-Diels-Alder rearrangement in ring C, and the loss of neutral molecules of H 2 O, sugar and carbonyl groups. Among the seven flavonoid standards, there were two major aglycones, flavanones (flavanomarein, butin-7-O-β-D-glucopyranoside, isookanin, taxifolin, 5,7,3 ,5 -tetrahydroxyflavanone-7-O-β-D-glucopyranoside) and chalcones (marein, okanin). The two aglycones have hydroxyls at C-3 , 4 , or 5 in the B ring and the loss of H 2 O caused by the 3 ,4 -dihydroxyl disposition. In addition, these two aglycones have a hydroxyl and a sugar at C-5 and C-7 in the A ring, usually identified as being glucose.  Figure 6. Okanin (peak 24) gave an [M − H] − ion at m/z 287.0558, which was 162 Da less than marein, proving the loss of a sugar from the C-7 position compared with marein. In addition, okanin showed the similar fragmentation rules as marein.
In summary, these fragmentation behaviors of standards could be used to obtain the fragmentation rules of different types of flavonoids and the 3-OH, 5-OH or 3 ,4 -dihydroxyl, 3 ,5 -dihydroxyl substitution patterns could be identified by the anundance of characteristic fragments.

Identification of Flavanones in Snow Chrysanthemum by ESI-Q-TOF-MS
Eleven flavanones have been identified from snow chrysanthemum, six of which (peaks 4, 5, 6, 9, 11 and 12) were unambiguously identified by comparison with the reference standards. Likewise, the fragment ion at m/z 151.0042 was produced by the characteristic RDA cleavage and neutral loss of C 8 H 6 O from the fragment ion at m/z 269.0334. Therefore, peak 7 was tentatively identified as luteolin 7-O-β-D-sophoroside [22]. Similarly, peak 10 exhibited an ion [M − H] − at m/z 433.1135 (C 21 H 22 O 10 ), which produced a major fragment ion at m/z 271.0605 by loss of glucose, indicating the presence of a glucose unit in its structure. A fragment ion at m/z 135.0424 resulted from the loss of C 8 H 8 O 2 from the fragment ion at m/z 271.0605. That indicated the existence of a 3 ,5 -dihydroxy disposition in peak 10 without the loss of water, so peak 10 was identified as 7,3 ,5 -trihydroxyflavanone-7-O-β-D-glucopyranoside by comparison with an authentic standard. Peak 17 gave an [M − H] − at m/z 271.0605, corresponding to the fragment ion of peak 10. According to the fragment ions, peak 17 was identified as 7,3 ,5 -trihydroxyflavanone. Similarly, by comparison with the available data, peak 25 was also identified having a 3 ,5 -dihydroxyl susbtituion pattern and the major fragment ions at m/z 287.0534 and 151.0042 displayed the existence of a 5-OH group, so peak 25 was tentatively assigned as 5,7,3 ,5 -tetrahydroxyflavanone.

Identification of Chalcones in Snow Chrysanthemum by ESI-Q-TOF-MS
Chalcones, one of the major type of flavonoids in the snow chrysanthemum, have similar fragment rules as flavanones, but there is a tremendous difference in chromatographic behavior between chalcones and flavanones in that chalcones have longer retention times than flavanones [18]. Peaks 14 and 24 were unambiguously identified as marien and okanin by comparison with reference standards.

HPLC-DAD Analysis Conditions
HPLC-DAD analysis was performed on a Flexar system (Perkin Elmer, Waltham, MA, USA) equipped with a pump, a diode array detector (DAD), and a Totalchrom chromatographic workstation,

Precision, Repeatability, Stability and Accuracy
Intra-day and inter-day variations were chosen to determine the precision of the method. They were determined by analyzing mixed standards solution six times within one day and on three consecutive days. Then the relative standard deviation (RSD) was taken as a measure of precision. To assess the repeatability of the method, a sample was extracted and analyzed six times parallelly. Then calculate the RSDs of the contents of eight compounds. The stability test was performed by analysis of the sample at 0, 2, 4, 6, 8, 12 and 24 h, and calculating the RSDs for peak area ratio of each analyte. The recovery test was used to evaluate the accuracy of this method. Accurate amounts of eight standards were added to the known amounts of eight compounds in samples. The above-prepared samples (n = 6) were extracted and analyzed as described as in Sections 3.2.1 and 3.3.1. The average recoveries were determined by the formula: Recovery (%) = (amount found − original amount)/spiked amount × 100%.

Conclusions
In this study, a high-performance liquid chromatography coupled with diode array detection (HPLC-DAD) method was established and validated in terms of linarity, sensitivity, precision, accuracy and stability, and then successfully applied for simultaneous quantitative analysis of eight characteristic compounds in snow chrysanthemum. The results of the 23 batches of samples demonstrate the influence of seed provenance and habitat. Flavanomarein and marein are found to be abundant in 23 batches of samples and could be used as suitable markers for quality control of snow chrysanthemum. In addition, a high resolution UPLC-ESI-QTOF-MS/MS method was developed for the systemic analysis of the flavonoids and phenolic acids in snow chrysanthemum. Based on their exact mass, fragmentation behaviors and retention timee, a total of 30 ingredients were identified in the crude extract of snow chrysanthemum, providing a deeper understanding of the chemical constituents of snow chrysanthemum.