Simultaneous Determination and Quantification of Triterpene Saponins from Camellia sinensis Seeds Using UPLC-PDA-QTOF-MS/MS

Saponins in the Camellia sinensis seeds have a broad spectrum of biological properties and application potentials. However, up to now, no chromatographic methods have been developed to provide full fingerprinting and quality assurance for these saponins. This research aimed to develop a novel method to tentatively identify and quantify saponins in C. sinensis seeds by ultra-high-performance liquid chromatography coupled with photo-diode array detector and quadrupole time-of-flight mass spectrometry (UPLC-PDA-QTOF-MS/MS), and compare it with the classic vanillin-sulfuric acid assay. Fifty-one triterpene saponins, including six potentially new compounds, were simultaneously detected by UPLC-PDA-MS/MS, and their chemical structures were speculated according to the retention behavior and fragmentation pattern. The total saponin content in the crude extract and the purified saponin fraction of C. sinensis seeds were quantified to be 19.57 ± 0.05% (wt %) and 41.68 ± 0.09% (wt %) respectively by UPLC-PDA at 210 nm, while the corresponding values were determined to be 43.11 ± 3.17% (wt %) and 56.60 ± 5.79% (wt %) respectively by the vanillin-sulfuric acid assay. The developed UPLC-PDA -MS/MS method could determine specified saponins, and is more reliable for quantifying the C. sinensis seed saponins than the classic spectrophotometric method. It is of great significance for the future investigations and applications of these saponins.


Introduction
Camellia sinensis (L.) O. Kuntze, belonging to the family of Theaceae, is a woody perennial tree with great economic importance [1]. Tea manufactured from the leaves of C. sinensis is second in popularity only to water as a beverage in the world, and possesses many benefits to human health [2]. In addition to tea leaves, the seeds of C. sinensis contain a large number of bioactive constituents including saponins, flavonoids, unsaturated fatty acids, and polysaccharides, and are considered as a valuable resource for food, agriculture, pharmaceuticals, aquaculture, and the cosmetic industry [3][4][5][6][7].
Oleanane-type triterpene saponins were determined as one of the major bioactive ingredients in C. sinensis seeds. So far, more than 50 saponin compounds have been isolated and identified from the seeds of C. sinensis cultivated in China, Japan, Sri Lanka, and India [8]. Pharmacological studies indicated that these saponins could protect gastric mucosa of rats induced by ethanol or indomethacin [9], inhibit gastric emptying and alcohol absorption in mice [10], suppress ascites cancer S180 cells, myelocytic saponins were simultaneously detected from the crude extract of C. sinensis seeds (2.0 mg/mL) in negative ion mode ( Figure 1A), and 40 corresponding peaks were obviously observed under a PDA detector at a wavelength of 210 nm ( Figure 1B). For the purified total saponin fraction (2.0 mg/mL), the UPLC/MS total ion chromatogram was similar to that of the crude extract (data not shown), and 50 saponins were detected at 210 nm ( Figure 1C). The majority of saponins of C. sinensis seeds possess no chromophores, and could be detected at the non-specific ultraviolet wavelengths around 210 nm. This structure feature led to lower detection sensitivity under the ultraviolet (UV) detector compared with MS, and explained that some minor peaks were not observed in the UV chromatogram of the samples. For the purified total saponin fraction, the peaks from the UV and MS total ion current chromatograms were almost all corresponding with each other except for peak 3, indicating UV detector is more applicable for detecting C. sinensis seed saponins at higher concentrations.

Compound Characterization of Saponins by UPLC-QTOF-MS/MS
In our system, the signal intensity of the saponins obtained in the negative ionization mode was superior to that in the positive ionization mode, which was consistent with the previous studies [24]. In the precursor ion full-scan spectra, the 51 saponin peaks were detected with parent ions [M − H] − at m/z 1157.5841-1315.6046 (Table 1)

Compound Characterization of Saponins by UPLC-QTOF-MS/MS
In our system, the signal intensity of the saponins obtained in the negative ionization mode was superior to that in the positive ionization mode, which was consistent with the previous studies [24]. In the precursor ion full-scan spectra, the 51 saponin peaks were detected with parent ions [M -H] − at m/z 1157.5841-1315.6046 (Table 1) Further identification was conducted based on the MS/MS spectra and fragmentation pathways, referring to previous reports on the saponins isolated from C. sinensis seeds [8]. were tentatively identified as theasaponin A 1 [30], theasaponin A 7 [31], floratheasaponin A [33], foliatheasaponin I/III [31,35], teaseedsaponin G, teaseedsaponin I, and teaseedsaponin J/K [32], respectively. In addition, the peak 18 was identified as theasaponin E 1 by comparing the retention time and MS/MS spectra with standard theasaponin E 1 .
No saponins with the same molecular weight as the peak 38, 42 and 46 have been reported. Their possible structures were try to speculated based on the known saponins from the seeds of the genus Camellia. . These data suggested that the molecular weight of its sapogenin was around 556, and the sugar moiety contained two pentoses, one hexose, and one hexosuronic acid. Among all the reported saponins from the Camellia seeds, only the sapogenin of oleiferasaponin D 1 in the seeds of Camellia oleifera [37], namely 22-O-angeloyl-camelliagenin A, has the corresponding molecular weight. The other structure possibilities of the sapogenin results from the angeloyl-linked positions at C-16 and C-28, and the tigloyl is another possible acyl group. The MS/MS fragmentation pattern of peak 38 were consistent with the sugar chain of 38 known saponins from C. sinensis seeds, including theasaponin A 1 -A 3 [30], A 8 , A 9 [12], B 5 [31], C 1 [29], E 1 -E 10 [26,29,33,36], F 1 -F 3 [30], G 1 , H 1 [29], G 2 [36], assamsaponin A-E [9], teaseedsaponin C, E-G, J-L [35], and floratheasaponin A, I, III [31,33,35], and the oligosaccharidic moiety of all the reported Camellia seed saponins is located at C-3. So the compound 38 was tentatively identified as 16 which was the oligosaccharidic moiety of theasaponin A 4 -A 7 [31], E 11 , E 12 [36], assamsaponin F-I [10], teaseedsaponin A, B, D, H, I [32], and camelliasaponin B 1 , C 1 [33] in C. sinensis seeds.  26,35,48), the amount of compounds in each group was equal to the number of known C. sinensis seed saponins with the similar molecular weight and corresponding fragments. In order to predict the elution order of isomers in each group, Log P value was calculated in SciFinder by launching a chemical structure from ChemBioDraw Ultra 14. Log P is defined as a logarithm of the distribution coefficient of the compound in the liquid/liquid extraction system of water and n-octanol. It provides direct information on hydrophobicity, and is frequently used to predict water solubility, soil (sediment)/water partition coefficient, absorption, distribution, metabolism, excretion, and toxicity (ADMET). The higher the Log P value, the more the compound is hydrophobic [38,39]. The group composed of peaks 15 and 19 was taken as an example to elucidate the identification of saponins with a similar molecular weight. The two peaks showed almost the same pseudomolecular ion [M − H] − at m/z 1173.58, and similar fragment ions at m/z 1041, m/z 1023, m/z 909, m/z 891, m/z 747, m/z 729, m/z 711, and m/z 571, which were consistent with theasaponin C1 [29] and theasaponin B5 [31]. The two compounds may have different sapogenins and the same oligosaccharidic moiety, and are undistinguishable from each other by MS/MS spectra. The Log P values of theasaponin C1 and B5 were estimated to be 2.0 ± 0.9 and 2.4 ± 0.9, respectively, suggesting that the former saponin was more hydrophilic, and should be eluted earlier than the latter one. So the peak 15 was tentatively determined as theasaponin C1, and the peak 19 was tentatively determined as theasaponin B5. Similarly, another four groups of isomers with ion [M − H] − at m/z 1231.5838-1231.5873, m/z 1261.5947-1261.5989, m/z 1273.5938-1273.5993, and m/z 1301.5922-1301.6239 were tentatively identified as assamsaponin D (peak 11) [9], theasaponinA9 (peak 14) [12], theasaponin A2 (peak 22) [30], theasaponin A6 (peak 9) [31], theasaponin A5 (peak 25) [29], theasaponin A3 (peak 17) [30], theasaponin E10 (peak 23) [36], theasaponin A8 (peak 32) [12], teaseedsaponin F (peak 41) [32], assamsaponin F(peak 26) [10], teaseedsaponin D (peak 35), [35] and theasaponin E11 (peak 48) [36].   26,35,48), the amount of compounds in each group was equal to the number of known C. sinensis seed saponins with the similar molecular weight and corresponding fragments. In order to predict the elution order of isomers in each group, Log P value was calculated in SciFinder by launching a chemical structure from ChemBioDraw Ultra 14. Log P is defined as a logarithm of the distribution coefficient of the compound in the liquid/liquid extraction system of water and n-octanol. It provides direct information on hydrophobicity, and is frequently used to predict water solubility, soil (sediment)/water partition coefficient, absorption, distribution, metabolism, excretion, and toxicity (ADMET). The higher the Log P value, the more the compound is hydrophobic [38,39]. The group composed of peaks 15 and 19 was taken as an example to elucidate the identification of saponins with a similar molecular weight. The two peaks showed almost the same pseudomolecular ion [M -H] − at m/z 1173.58, and similar fragment ions at m/z 1041, m/z 1023, m/z 909, m/z 891, m/z 747, m/z 729, m/z 711, and m/z 571, which were consistent with theasaponin C 1 [29] and theasaponin B 5 [31]. The two compounds may have different sapogenins and the same oligosaccharidic moiety, and are undistinguishable from each other by MS/MS spectra. The Log P values of theasaponin C 1 and B 5 were estimated to be 2.0 ± 0.9 and 2.4 ± 0.9, respectively, suggesting that the former saponin was more hydrophilic, and should be eluted earlier than the latter one. So the peak 15 was tentatively determined as theasaponin C 1, and the peak 19 was tentatively determined as theasaponin B 5 . Similarly, another four groups of isomers with ion [M -H] − at m/z 1231.5838-1231.5873, m/z 1261.5947-1261.5989, m/z 1273.5938-1273.5993, and m/z 1301.5922-1301.6239 were tentatively identified as assamsaponin D (peak 11) [9], theasaponinA 9 (peak 14) [12], theasaponin A 2 (peak 22) [30], theasaponin A 6 (peak 9) [31], theasaponin A 5 (peak 25) [29], theasaponin A 3 (peak 17) [30], theasaponin E 10 (peak 23) [36], theasaponin A 8 (peak 32) [12], teaseedsaponin F (peak 41) [32], assamsaponin F (peak 26) [10], teaseedsaponin D (peak 35), [35] and theasaponin E 11 (peak 48) [36]. . This compound has not been found in C. sinensis seeds before, and its mass spectrometry data is in accordance with camelliasaponin A 1 and A 2 from C. japonica seeds [34]. . Among all the reported C. sinensis seed saponins, only teaseedsaponin L corresponds to these data [35]. So either of the peaks 44 and 47 might be teaseedsaponin L, and the another one (probably unreported) in the seeds of C. sinensis.  Table 1. It was noticed that the elution order of the 51 peaks were not all in the sequence of computer-calculated Log P values. Although Log P is widely used in many quantitative structure-activity relationship (QSAR) models to estimate critical physico-chemical properties of compounds, most of these models exhibit only modest predictive ability due to the chemical diversity of the systems modeled [40]. In this work, Log P value only was used as a reference basis to predict the elution order of isomers. However, it should be noticed that these speculated results need to be further verified due to the limitation of Log P value and the different conformation of isomers which can influence the retention time.
Completely-accurate identification results should be obtained depending on the availability of all appropriate standards, or development of efficient on-line hyphenated techniques coupling separation and structural identification. The liquid chromatography (LC)-solid-phase extraction (SPE)-nuclear magnetic resonance (NMR)-mass spectrometry (MS) coupling can greatly simplify the structure identification of unknown compounds in the complex mixtures, without the need for purifying of the analytes [41]. This technique has become a valuable tool for pharmaceutical and natural product research, and is worth being investigated for the analysis of C. sinensis seed saponins in the future. Although this UPLC-QTOF-MS/MS method could only provide inference of the compound structures, it was the first time to establish a liquid chromatography method for simultaneously detecting 51 saponins from the C. sinensis seeds.

Quantitation of Saponins by UPLC-PDA
Although several chromatographic techniques have been developed for saponin analysis, LC with UV detection remains the routine method for the quality control of saponin in herbal products [18]. In this work, the saponins in C. sinensis seeds were quantified by their response to PDA, and theasaponin E 1 was used as a standard. The calibration curves were linear over the concentration range of 0-1000 µg/mL for theasaponin E 1. The linear regression equation was Y = 8233.9X + 43043, where X is the concentration and Y is the peak area. The 95% confidence intervals for the regression slope and y-intercept were calculated to range between 8032.5-8435.3 and −62879.8-148965.6. The high correlation coefficient value (R 2 = 0.9994) indicated a good correlation between standard concentrations and their peak areas within the test ranges. The LLOD (S/N = 3) and LLOQ (S/N = 10) of theasaponin E 1 were 0.0313 µg/mL and 0.1043 µg/mL, respectively. The intra-and inter-day of accuracy and precision data at three concentrations of theasaponin E 1 were summarized in Table 2. The accuracy for theasaponin E 1 ranged from −0.092% to 0.079%, and the precision was between 0.013% and 0.152%. These results met the Guidance for Industry: Bioanalytical Method Validation (2013) proffered by the US Food and Drug Administration (FDA), which recommended the mean value should be within 15% of the nominal value for accuracy, and the precision should not exceed 15% of coefficient of variation (CV). The developed UPLC-PDA method was applied to analyze 70% methanol extract and the total saponin fraction of C. sinensis seeds. The individual saponins were quantified, and the total saponin content of the crude extract and the total saponin fraction were calculated to be 19.57 ± 0.05% (wt %) and 41.68 ± 0.09% (wt %), respectively ( Table 3). The disadvantage of this method was that all saponin compounds were quantified by the same one standard substance, not by each their own standards. This is due to no commercial C. sinensis seed saponin compounds are available, and purify of all these saponins as standards is still difficult so far. In view of the structure similarity of these compounds, this UPLC-PDA method is still relative reliable for quantitation of C. sinensis seed saponins.

Comparison between Vanillin-Sulfuric Acid Assay and UPLC-PDA
Considering the vanillin-sulfuric acid assay is a classical and popular spectrophotometric method for quantifying saponins from various plant materials [18], the developed UPLC-PDA method was compared with vanillin-sulfuric acid assay in the present work. C. oleifera seed saponins have been determined by this colorimetric method, and the detection wavelength was set at 538 nm [19,20]. However, the maximum absorption of reaction mixtures for theasaponin E 1 , 70% methanol extract and the total saponin fraction of C. sinensis seeds were found at 600 nm ( Figure 3). The standard curve for theasaponin E 1 was Y = 1.166X -0.1221 (R 2 = 0.9982), where X was the concentration and Y was the absorbance. The 95% confidence intervals for the slope and y-intercept were estimated to be 1.051 to 1.271, and −0.190 to −0.047. The contents of saponins in the 70% methanol extract and the purified total saponin fraction were determined to be 43.11 ± 3.17% (wt %) and 56.60 ± 5.79% (wt %), respectively. These results were significantly higher than those determined by the UPLC-PDA method (p < 0.05). The former value was more than twice, and the latter was 1.36 folds as compared with the UPLC-PDA data (p < 0.05).
Molecules 2019, 24, x FOR PEER REVIEW 11 of 18 Considering the vanillin-sulfuric acid assay is a classical and popular spectrophotometric method for quantifying saponins from various plant materials [18], the developed UPLC-PDA method was compared with vanillin-sulfuric acid assay in the present work. C. oleifera seed saponins have been determined by this colorimetric method, and the detection wavelength was set at 538 nm [19,20]. However, the maximum absorption of reaction mixtures for theasaponin E1, 70% methanol extract and the total saponin fraction of C. sinensis seeds were found at 600 nm ( Figure 3). The standard curve for theasaponin E1 was Y = 1.166X − 0.1221 (R 2 = 0.9982), where X was the concentration and Y was the absorbance. The 95% confidence intervals for the slope and y-intercept were estimated to be 1.051 to 1.271, and −0.190 to −0.047. The contents of saponins in the 70% methanol extract and the purified total saponin fraction were determined to be 43.11 ± 3.17% (wt %) and 56.60 ± 5.79% (wt %), respectively. These results were significantly higher than those determined by the UPLC-PDA method (p < 0.05). The former value was more than twice, and the latter was 1.36 folds as compared with the UPLC-PDA data (p < 0.05). Differences in selection of reagent, condition of full color development, standard, and wavelength make it hard to compare the results from different studies using this spectrophotometric assay [18]. More importantly, the vanillin-acid reagent could give a color reaction under acidic conditions with tannins [42], flavonoids [43], proanthocyanidins [44], and steroids [45], thus provide nonspecific and misleading information [21]. As shown in Figure 3, the absorption peaks of the total saponin fraction and crude extract were not obvious, indicating the absorption value might be easily interfered by other impurities. Although several trace saponins were not detected in the crude extract and purified saponin fraction (2 mg/mL) by UPLC-PDA, we thought that the exorbitant detection result from vanillin-sulfuric acid was mainly caused by some interfering substances such as flavonoids. Given the convenience and low cost, the vanillin-sulfuric acid assay could be used to quantify purified C. sinensis seed saponins, and give the comparable result with the UPLC-PDA method. However, it is not suitable for determining the saponin content in the crude extract of C. sinensis seeds. Differences in selection of reagent, condition of full color development, standard, and wavelength make it hard to compare the results from different studies using this spectrophotometric assay [18]. More importantly, the vanillin-acid reagent could give a color reaction under acidic conditions with tannins [42], flavonoids [43], proanthocyanidins [44], and steroids [45], thus provide nonspecific and misleading information [21]. As shown in Figure 3, the absorption peaks of the total saponin fraction and crude extract were not obvious, indicating the absorption value might be easily interfered by other impurities. Although several trace saponins were not detected in the crude extract and purified saponin fraction (2 mg/mL) by UPLC-PDA, we thought that the exorbitant detection result from vanillin-sulfuric acid was mainly caused by some interfering substances such as flavonoids. Given the convenience and low cost, the vanillin-sulfuric acid assay could be used to quantify purified C. sinensis seed saponins, and give the comparable result with the UPLC-PDA method. However, it is not suitable for determining the saponin content in the crude extract of C. sinensis seeds.

Plant Materials and Chemicals
The full ripe fruits of tea (Camellia sinensis (L.) O. Kuntze, Theaceae) variety Fudingdabai were harvested from the Tea Resource Garden of China National Tea Museum (Hangzhou, China). The color of these fruits was brown, and their diameter was between 2.7 cm and 3.3 cm. Methanol and acetonitrile of HPLC grade were purchased from Tianjin Shield Company (Tianjin, China). Ethanol, methanol, formic acid, and vanillin of analytical grade were obtained from the National Medicine Chemical Reagent Co. Ltd. (Shanghai, China). The ultrapure water was prepared by a Milli-Q water purification system (Millipore, Bedford, MA, USA).

Preparation of Tea Seed Extract, Total Saponin Fraction and Standard Substance (Theasaponin E 1 )
Tea seeds were removed from shells, freeze-dried and milled into powder by a pulverizer. Then, 800 g powder was extracted under reflux with 12 L of 70% methanol aqueous solution at 70 • C for 5 h. The filtered decoction was concentrated by rotary evaporation, and then lyophilized by a freeze dryer (LGJ-10C, Beijing Sihuan Science Instrument Factory, Bejing, China). The yield of extract from tea seeds was 106.06 g (13.26%).
The crude tea seed extract was suspended in water and extracted successively with petroleum ether, ethyl acetate (EtOAc), and 1-butanol (n-BuOH) thrice, respectively. The n-BuOH fraction was concentrated under reduced pressure, subjected to D101 column chromatography, and then eluted consecutively with H 2 O, 30%, 50%, and 70% ethanol-aqueous solution (v/v) at the flow rate of 3 mL/min each for 2 bed volume (BV, 2.5 L). The 30%, 50%, and 70% ethanol-eluted fractions were collected, concentrated, and lyophilized to yield 9.04 g, 24.69 g and 3.09 g, respectively. Each fraction was detected by the UPLC-PDA-MS method as described in the Section 2.3. The 50% ethanol-eluted fraction had the similar saponin compositions with the crude tea seed extract, and was used as the total saponin fraction for the next work.

Separation and Identification of Saponins by UPLC-PDA-QTOF-MS/MS
The chromatographic separation was performed on an Acquity UPLC instrument equipped with a PDA detector (Waters, Milford, MA, USA) and a Waters Acquity UPLC HSS T3 column (1.8 µm, 150 mm × 2.1 mm i.d., Waters, Milford, MA, USA). The PDA detector was set to scan a range from 200 nm to 400 nm, and the specific wavelength of 210 nm was selected for detecting saponins. The mobile phase A was water containing 0.1% formic acid, and B was acetonitrile containing 0.1% formic acid. Gradient elution was performed as follows: 0-4 min, 35-37% B; 4-32 min, 37% B; 32-58 min, 37-45% B; 58-60 min, 35% B. The sample injection volume was 5 µL, the column was kept at 30 • C, and the flow rate was 0.2 mL/min.
The mass spectrometry analysis was carried out using a hybrid triple quadrupole time-of-flight mass spectrometer (Triple TOFTM 5600 plus System, AB SCIEX, Foster City, CA, USA) equipped with an electrospray ionization (ESI) source in negative ionization mode. The mass range was set at m/z 100-2000 for both Time-of-Flight (TOF)-MS and TOF-MS/MS scan. Nitrogen was used as the nebulizer and auxiliary gas. The optimized MS conditions were as follows: ion spray voltage, −4.5 kV; ion source temperature, 500 • C; nebulizing gas (Gas 1), 50 psi; Tis gas (Gas 2), 50 psi; curtain gas, 35 psi; declustering potential (DP), 100 V; collision energy (CE), 10 V. For the MS/MS acquisition mode, the sweeping collision energy was set at 60 ± 20 V for collision-induced dissociation (CID). Data acquisition was carried out using Peak View Software 1.2.0.3 software (AB Sciex, Foster City, CA, USA) in IDA (information-dependent acquisition) mode.

Quantification of Saponins by UPLC-PDA
Quantitative analysis of the individual saponins in the tea seeds was conducted using UPLC-PDA as described in Section 2.3. The method was validated according to Guidance for Industry Bioanalytical Method Validation (FDA) [27]. Theasaponin E 1 was used as a reference standard, and the calibration curve was constructed at seven concentrations in the range of 50-1000 µg/mL. The "intra-day" and "inter-day" accuracy and precision of the method were determined with three concentration levels (60, 100, and 500 µg/mL) of theasaponin E 1 on the same day (five replicates) or on five different days. The accuracy, expressed as the percent deviation (% DEV), was calculated as (mean measured concentration − nominal concentration/nominal concentration) × 100%. The precision, expressed as the relative standard deviation (RSD), was calculated as (standard deviation/mean measured concentration) × 100% The lower limits of detection (LLOD) and quantification (LLOQ) were defined as analyte concentrations with a signal-to-noise (S/N) ratio of 3 and 10, respectively, by injecting a series of dilute solutions with known concentrations.

Vanillin-Sulfuric Acid Assay
The total saponin content of tea seeds was determined following the procedures described by Chen et al. with a few modifications [19]. Briefly, 0.5 mL of sample solution was sequentially reacted with 0.5 mL of vanillin ethanol solution (8%, w/v) and 4 mL of diluted sulfuric acid (72%, w/v). After incubation at 60 • C for 10 min, the reaction mixture was cooled in an ice water bath for another 15 min, and scanned from 450 to 800 nm by an UV-VIS spectrophotometer (UV-2550, Shimadzu, Japan). Theasaponin E 1 (0.2-1.0 mg/mL) was used as a reference standard to establish the calibration curve.

Statistical Analysis
The data were presented as the mean ± standard deviation (SD). Statistical differences between the two sets of data were evaluated using the Student s t-test, and a p value < 0.05 was considered statistically significant. All statistical analysis of data was performed using the SPSS (Statistical Product and Service Solutions) statistics [26].

Conclusions
A novel UPLC-PDA-QTOF-MS/MS method was developed and validated for tentatively determining 51 triterpene saponins in C. sinensis seeds. Based on the retention behavior, fragmentation pattern and previous literatures, six potentially new saponins (1, 6, 38, 42, 46, 44, or 47) were detected. One saponin (36) identified as camelliasaponin A 1 or A 2 from the C. japonica seeds [34] was found in the C. sinensis seeds for the first time. Compared with the vanillin-sulfuric acid assay, the UPLC-PDA method could quantify individual saponins, and provide more reliable quantitative test results for the total saponin content. Its linearity, accuracy and precision were satisfactory under the requirements of the Guidance for Industry: Bioanalytical Method Validation (FDA, 2013). This work offered a possible routine method for the identification and quantification of C. sinensis seed saponins, which was of great significance for the future investigations and applications of these compounds. Moreover, this analytical method could be further developed and validated for the quantification of the saponins in different organs of C. sinensis and other Camellia species such as C. reticulata, C. japonica, and C. sasanqua, etc. & Institute of Agrobiology and Environmental Sciences, Zhejiang University.