Identification of Chinese green tea (Camellia sinensis) marker metabolites using GC/MS and UPLC-QTOF/MS

Abstract

Tea is one of the most widely consumed aromatic beverages in the world because of its taste and flavor, as well as due to many potential health beneficial properties. Metabolomics focuses on an in-depth analysis of all metabolites in living organisms. In this study, 29 primary metabolites and 25 secondary metabolites were identified using GC/MS and UPLC-QTOF/MS, respectively. Further, PCA analysis showed conspicuous discrimination for the ten varieties of green tea with metabolite profiling. Among them, organic acids, amino acids, flavan-3-ols, and flavonol glycosides varied greatly through checking the VIP values of the PLS-DA model. Moreover, the intrinsic and/or extrinsic factors characterizing each type of green tea were also discussed. The chemical component marker derived here should be used as an important detection index, while evaluating the tea quality, as well as while establishing the tea quality standard.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10068-021-00970-4.

Introduction

China is the largest tea producer, consumer, and exporter in the world, accounting for more than 40% of the total global production (Xiao et al., 2018). The annual output of tea in China was 298.6 million tons in 2020 according to the official statistics of the China Tea Marketing Association, and it has continued the tendency of rapid growth. Green tea, white tea, yellow tea, oolong tea, black tea, and dark tea are the six well-known categories of tea in China. Among them, green tea is the most widely produced form of tea in China, which accounts for more than 60% of the tea production, followed by black tea, dark tea, and oolong tea. Green tea, a type of unfermented tea, can retain maximized primary metabolites of tea by quick application of heating or steaming called green-killing of polyphenol oxidase and oxidase, which are inactivated during this process (Yi et al., 2015).

Tea contains several active components, such as polyphenols, alkaloids, and polysaccharides, as well as volatile oils and amino acid. These chemical metabolites determine the tea quality and are regarded as the most active ingredients that contribute to the potential health benefits of tea (Frei & Higdon, 2003; Sharangi, 2009). Metabolomics focuses on an in-depth analysis of all metabolites in living organisms, which can help to elucidate the relationship between genotypes and phenotypes (Fiehn, 2002; Weckwerth & Fiehn, 2002). Plant metabolomics has been widely used in the chemical analysis of plant samples with differences involving geographical factors, species, and processing treatments (Sumner et al., 2003). It has been reported that green teas from China, South Korea, and Japan could be recognized by UHPLC-HRMS based metabolomics (Navratilova et al., 2019). Wang et al. distinguished volatiles and sensory metabolites in different varieties of Dianhong tea using HS–SPME–GC–MS, and 40 key compounds were identified as biomarkers among different tea cultivars (Wang et al., 2017). Different tea categories could be clearly classified according to their degree of fermentation, which was based on the contents of their major constituents and/or UHPLC-Q-TOF/MS derived metabolites (P. Li et al., 2018; Yi et al., 2015). Green teas from South and East Asian countries was subjected to different growth environments, and they could be discriminated using UV–Visible, FTIR, and HPLC techniques coupled with chemometric analysis (Aboulwafa et al., 2018). Green teas from Korea, China, and Japan were processed using their respective conventional methods, and the concentrations of catechins and caffeine varied considerably (Kim et al., 2009). In general, classified tea samples in the above studies came from areas situated far from each other, e.g. different countries or belonging to different tea categories. The wide difference caused by genetic strain or fermentation process between different tea samples led to a great variance in the chemical composition. However, a systematic metabolomics analysis of Chinese green tea derived from different varieties and geographical origins has not been studied to date.

In China, Xihu Longjing (LJ), Luan Guapian (GP), Biluochun (BLC), Anji Baicha (BC), Taiping Houkui (HK), Huangshan Maofeng (MF), Yuhuacha (YH), Zhuyeqing (ZYQ), Maojian (MJ), and Yunwu (YW) were identified as the top ten most popular green tea varieties. They differ essentially due to variable growing plantation, planting area, harvesting time, processing procedure, and climate (Su et al., 2019; Wen et al., 2019). There are large differences in both quality and price among these broad varieties. Motivated by interest, some unscrupulous vendors deceitfully label their product to show as if it is coming from genuine areas to make big profits (Huo et al., 2014). This behavior harms the interests of consumers, as well as the interests of tea producers. Thus, methods for identifying the chemical markers for each type of tea are significant to provide security for both tea producers and consumers.

In this study, the synergistic combination of UPLC-QTOF/MS and GC/MS provided global metabolites, including catechins, alkaloids, flavan-3-ols, flavonols, phenolic acid, organic acid, and amino acid, from ten China green tea varieties. Then, the obtained metabolic profiles were processed with the multivariate data analysis methods, including PCA and PLS-DA model. Furthermore, the identification of chemical markers affecting the quality and characterization each type of tea were conceptualized and exploited to establish the green tea quality standard and distinguish fake products.

Materials and methods

Chemicals and materials

LC/MS grade acetonitrile (Optima® LC–MS grade), methanol (Duksan Pure Chemical Co., Ltd., Korea), and water (Lichrosolv, Merck) were purchased for the UPLC-QTOF/MS analysis. Methoxyamine hydrochloride (MH, sigma aldrich, cat. No. 226904), and N-methyl-N-(trimethylsily) Trifluoroacetamide (MSTFA, sigma aldrich, cat. No. 69479) were used for derivatization in the GC–MS analysis. Hexane (JT Baker, Phillipsburg, NJ, USA) and pyridine (Sigma-aldrich) were used for dilution and derivatization in GC–MS analysis, respectively.

Collection of tea samples

The green tea samples Longjing (Hangzhou city, LJ1-10), Biluochun (Suzhou city, BLC1-10), Taiping Houkui (Huangshan city, HK1-10), Yuhua (Nanjing city, YH1-10), Luan Guapian (Luan city, GP1-10), Anji Baicha (Huzhou city, BC1-10), Zhuyeqing (Emei city, ZYQ1-10), Xinyang Maojian (Xinyang city, MJ1-10), Huangshan Maofeng (Huangshan city, MF1-10), and Lushan Yunwu (Lushan city, YW1-10) were collected from the local tea factory and company. All areas producing these ten samples were located at a latitude of 29 ~ 32 degrees north. The regional climate in these areas was endowed with a mild weather, which was moist and rainy. The daily average temperature ranged from 15.3 °C to 17.0 °C, annual precipitation ranged from 1385 to 1980 mm, and annual sunshine hours ranged from 1675 to 2180 h in these tea producing locations (the climate data was collected from the China Meteorological Data Service Center, http://data.cma.cn/en). A favorable climate provides a good environment for tea growth (Hajiboland, 2017). To avoid the possible influence of individual tea production process on its metabolite variation, 10 samples were collected from different manufacturers per locality to make up 100 green tea samples, and detailed manufacturer names are listed in Table 1s in the appendix file.

Green tea extract preparation

The extraction method was conducted as previously reported (Das et al., 2019), with a modification. Briefly, tea leaves were grounded into powder, and then 200 mg of each sample were extracted with distilled water (90 °C) and placed in an ultrasonic bath (Power sonic 520, Hwashin Co., Korea) for 20 min. Then the tea extract was filtered and dried with a chemical freeze dryer (Operon freeze-dryer, Operon Co. Ltd., Korea). Further, 50% MeOH was used to dissolve the tea extract (100 μg/ml), and the solution was filtered through a 0.2 μm filter before being injected into UPLC-QTOF/MS.

Prior to the GC–MS analysis, the tea extract was derivatized using a two-step process as previously reported (Jung et al., 2017; Li et al., 2019). Four micrograms of tea extract was precisely weighed and oximated using 100 μL methoxyamine hydrochloride in pyridine (20 mg/mL) at 30 °C for 90 min, and silylated with 100 μL N-Methyl-N-(trimethylsilyl) trifluoroacetamide (MSTFA) at 38 °C for 30 min. Then 800 μL of n-hexane was added. Finally, the solution was diluted to 500 μg/mL before injection.

GC–MS analysis

GC–MS analyses were conducted on Agilent 7890B GC equipped with a 7010 Triple Quadrupole mass spectrometer coated with an HP-5 MS capillary column (Agilent J & W; GC Columns, USA, 30 m × 250 μm, 0.25 μm). The analysis method was optimized based on the previously published technique (Ray et al., 2017), with a modification. High-purity helium was used as carrier gas at a flow velocity of 2.25 mL/min, nitrogen was used as collision gas at a flow rate of 1.5 mL/min, and one microliter (1 μl) of sample was injected into GC–MS. The column temperature was maintained at 60 °C for 3 min isothermally, and increased to 179 °C at a rate of 7 °C/min. Then the temperature was increased to 290 °C at a rate of 5 °C/min and maintained for 5 min isothermally. The ion source temperature was 230 °C, and the interface temperature was set at 290 °C. The MS was scanned at the range of 50–900 atomic mass units with an electron-impact mode at 70 eV.

UPLC-QTOF/MS analysis

UPLC-QTOF/MS was performed with a UPLC system (Waters Acquity® FTN, Acquity BSM) coupled with a Waters Vion™ IMS-Qtof mass spectrometer, operated in a positive ion electrospray mode with mobility-enabled non-targeted HDMS^E scan methods with a range of 50–1000 Da using 0.1 s scan time. Chromatographic separation was achieved with gradient elution at a flow rate of 0.35 mL/min using 0.1% formic acid in water as mobile phase A and methanol/acetonitrile (60:40, v/v) with 0.1% formic acid as mobile phase B. The elution condition was set as follows: 0–3 min (5%B), 3–18 min (5–19%B), 18–28 min (19–40%B), 28–29 min (40–95%B), 29–31.5 min (95%B), 31.5–32 min (95–5%B) and 32–36 min (5%B). The low collision energy and high collision energy of the mass spectrometer were 6 eV and 20–40 eV, respectively. Nitrogen was used as the drying gas with a flow rate of 800 L/h, and the cone gas flow was maintained at 50 L/h. The de-solvation temperature was 400 °C and the source temperature was 150 °C. The observed capillary and sampling cone voltages were 3.5 kV and 30 V, respectively. The reference capillary voltage was set at 2.0 kV. Data acquisition and processing was conducted by using the UNIFI v1.8.1 software.

Data analysis

Identification of peaks from the total ion chromatography of GC/MS was performed by matching the mass spectra with the NIST library, and the peaks with a similarity index more than 80% were assigned the name of the corresponding compound. For UPLC-QTOF/MS, the UNIFI traditional medicine library (UTML) offered a simple and effective way to quickly identify the compounds through a database. Information of compounds identified by the UTML included names, molecular formulas, and chemical structures, as well as the exact mass of main compounds (Tang et al., 2016). Relative percentage amounts of the identified compounds were calculated based on the percentage of their relative peak areas.

Multivariate analysis

After peak alignment, and data normalization, the relative percentage amounts of the identified compounds were standardized before being exported to the SIMCA-P software (version 14.1, Umetrics, Umeå, Sweden). Then PCA analysis was performed after the data had undergone Pareto-scaling. Furthermore, PLS-DA was applied to maximize the separation of fed and fasted individuals.

Results and discussion

Primary metabolite profiling and multivariate statistical analysis based on GC–MS

Identification of primary metabolites was performed on the basis of the analytical ion chromatogram from the measurement of tea samples using GC–MS (Fig. 1s). To avoid analyzing duplicate records from UPLC-QTOF/MS, secondary metabolites were excluded from the statistics. In total, 29 common metabolites were identified based on the NIST library (Table 1), and the corresponding fragment patterns showed good agreement with those in the libraries. They included 11 organic acids, 10 amino acids, 2 sugars, 2 amines, 2 silanes, 1 alcohol, and 1 nitrile. Main metabolic profiling derived here was in agreement with the previously reported profiling (Pongsuwan et al., 2007).

Table 1.

Identified compounds by GC/MS

No.	Compound Name	RT*(min)	Mass spectra
1	Methoxydi(tert-buty) silane (MS)	6.88	174,117,89,75,59
2	Lactic acid (LA), 2TMS	8.28	234,191,147,117,73
3	l-Alanine (Ala), 2TMS	9.21	233,147,116,73
4	Hydroxylamine (HX), 3TMS	9.41	249,146,133,119,100,86,73
5	Oxalic acid (OA), 2TMS	9.92	234,147,73
6	l-Valine (Val), 2TMS	11.68	261,218,144,100,73
7	l-Leucine (Leu), 2TMS	12.85	275,158,147,102,73
8	Silanol, trimethyl-phosphate(3:1) (Sil)	12.94	314,299,211,147,133,73
9	l-Isoleucine (Isol), 2TMS	13.28	275,218,158,147,100,73
10	Butanedioic acid (BA), 2TMS	13.62	262,247,147,73
11	Serine (Ser), 3TMS	14.66	321,306,278,218,204,188,147,100,73
12	l-Threonine (Thr), 3TMS	15.17	355,291,218,147,117,101,73
13	Malic acid (MA), 3TMS	17.01	350,335,245,233,190,147,73
14	l-5-Oxoproline (Oxo), 2TMS	17.47	273,258,230,156,147,73
15	l-Aspartic acid (Asp), 3TMS	17.54	349,306,232,218,147,101,73
16	4-Aminobutanoic acid (AA), 3TMS	17.60	319,304,174,147,73
17	2,3,4-Trihydroxybutyric acid tetrakis (TAT)	18.38	424,292,220,205,147,117,103,73
18	l-Glutamic acid (LGA), 3TMS	19.18	363,348,246,156,147,128,73
19	l-Glutamine (Glut), 3TMS	21.64	362,347,245,156,73
20	2-keto-l-gluconic acid (KLGA)	21.80	554,307,292,217,205,189,147,107,73
21	Shikimic acid (SA), 4TMS	22.36	462,255,204,147,133,73
22	Citric acid (CA), 4TMS	22.56	480,363,347,273,147,73
23	Quinic acid (QA), 5TMS	23.37	552,345,255,205,191,147,133,73
24	D-Glucose (Glu)	24.04	569,319,217,205,160,147,103,73
25	Myristic acid (MyA), TMS	25.87	300,285,145,132,129,117,73
26	Oleanitrile (ON)	26.40	263,136,122,83,55
27	Myo-Inositol (Myo), 6TMS	27.17	612,318,305,265,217,191,147,73
28	Oleamide (OM), TMS	31.80	353,338,144,128,116,75
29	Sucrose (Suc), 8TMS	36.09	437,361,271,217,147,73

*RT: Retention time

To obtain an overview of the difference in the primary metabolites of different green tea samples, PCA was employed to analyze the relationship of tea samples from different sources. As shown in Fig. 1A, ten varieties of green teas were clearly classified in the PCA score plot, and the first two principal components PC1 (33.6%) and PC2 (25.8%) explained 59.4% of the total variance. The PCA model indicated good predictive ability, and the Q² was 0.75. Thus, we drew a conclusion that primary metabolites varied greatly among different varieties of green tea.

Fig. 1.

PCA score plot (A), loadings plot (B), PLS-DA score plot (C), and VIP values (D) of the ten varieties of green teas based on the GC/MS data. VIP values were derived from the PLS-DA model. Variables with VIP values larger than 1, variables with VIP values smaller than 1. Variables with VIP values larger than 1 are regarded as the marker components

Secondary metabolite profiling and multivariate statistical analysis based on UPLC-QTOF/MS

Identification of secondary metabolites was performed on the basis of the analytical ion chromatogram from the measurement of tea samples using UPLC-QTOF/MS (Fig. 2s). Identification of compounds was compared the mass spectra with the UNIFI traditional medicine library, as well as the previous literature. Twenty-five secondary metabolites were identified, as shown in Table 2. We found that flavanols, alkaloids, and polyphenols, as well as glycosylated flavonols were the major secondary metabolites in green tea, which was in agreement with the previously published findings (Das et al., 2019). Parent ions of ions at m/z 303 and 287 were assigned to quercetin and kaempferol, which stands for their respective protonated aglycon moieties (Liu et al., 2014).

Table 2.

Identified compounds by UPLC-QTOF/MS

No.	Compounds name	RT* (min)	[M + H+] (m/z)	Mass error (ppm)	Formula	MS2(m/z)
1	Gallic acid (GA)	3.23	171.03797	1.2	C7H6O5	–
2	5-Galloylquinic acid (5GQA)	4.24	345.10021	0.8	C14H16O10	367, 327, 153, 125
3	3-Galloylquinic acid (3GQA)	4.63	345.10021	0.8	C14H16O10	367, 327, 153, 125
4	(− )-Gallocatechin (GC)	6.74	307.09676	0.3	C15H14O7	289, 139
5	Theobromine (TB)	7.3	181.08117	1.2	C7H8N4O2	163
6	Apigenin-6,8-C-diglucoside (AD)	7.65	595.1501	1.3	C27H30O15	617, 279
7	(− )-epigallocatechin (EGC)	12.03	307.09676	1.0	C15H14O7	329, 289, 139
8	Catechin (Cat)	12.27	291.10278	0.7	C15H14O6	313, 273, 165, 139, 123
9	Caffeine (Caf)	13.81	195.09853	1.7	C8H10N4O2	138
10	Procyanidin dimer (PD)	15.64	579.16653	0.4	C30H26O12	601, 441, 291, 279, 195
11	(− )-epicatechin (EC)	16.94	291.10278	1.2	C15H14O6	313, 273, 165, 139, 123
12	(− )-Epigallocatechin gallate (EGCG)	17.81	459.11716	0.8	C22H18O11	481, 289, 151, 139
13	Theaflavin 3-O-(3-O- methyl)gallate (TG)	19.91	731.20173	1.6	C37H30O16	619, 443, 279, 153
14	(− )-Gallocatechin gallate (GCG)	20.65	459.11716	1.4	C22H18O11	481, 289, 151, 139
15	(− )-Epicatechin gallate (ECG)	22.4	443.12125	1.1	C22H18O10	465, 291, 273, 151, 139, 123
16	( +)-Catechin gallate (CG)	23.18	443.12125	1.1	C22H18O10	465, 291, 273, 151, 139, 123
17	Quercetin-3-O- Gallactosylrutinoside (QGAR)	24.05	773.25681	0.5	C33H40O21	795, 611, 465, 303
18	Querctdin-3-O- glucosylrutinoside (QGLR)	24.47	773.25681	0.5	C33H40O21	795, 611, 465, 303
19	Quercetin 3-O-rhamnosylgalactoside (QRG)	24.82	611.18527	− 0.9	C27H30O16	633, 465, 433, 303
20	Rutin	25.07	611.19575	− 0.9	C27H30O16	633, 595,465, 433, 303
21	Kaempferol-3-O-β-D- glucopyranoside (KG)	25.28	449.13261	− 0.5	C21H20O11	471, 303
22	Kaempferol 3-O- glucosylrutinoside (KGR)	26.09	757.26091	0.7	C33H40O20	595, 449, 287
23	Kaempferol 3-O- rhamnosylrutinoside (KRR)	26.4	741.26606	− 0.5	C33H40O19	763, 595, 449, 287
24	Kaempferol-3-O-rutinoside(KR)	26.86	595.19937	0.3	C27H30O15	617, 449, 287
25	Quercetin-3-O-α-L-rhamnoside (QR)	27	449.13261	0.5	C21H20O11	471, 303, 287

*RT: Retention time

To compare the difference in secondary metabolites between different tea varieties, the PCA analysis was conducted. As shown in Fig. 2A, the 10 varieties of green tea were separated clearly in the sore plot. The first two principal components PC1 (68.2%) and PC2 (12.8%) explained 81.0% of the total system variance. PC1 was the key component for sample separation. The Q² for the PCA model was 0.85, which showed good predictive ability for the ten green varieties. In addition to caffeine, EGCG, GCG, and ECG, compounds KGR and 5-GQA also contributed greatly for the projection of PC1 with positive values. On the other hand, flavonol glycosides, such as KGR, QGLR, QGAR, and QRG showed great importance for PC2 (Fig. 2B).

Fig. 2.

PCA score plot (A), loadings plot (B), PLS-DA score plot (C), and VIP values (D) of the ten varieties of green teas based on the UPLC-QTOF/MS data. VIP values were derived from the PLS-DA model. Variables with VIP values larger than 1, variables with VIP values smaller than 1. Variables with VIP values larger than 1 are regarded as the marker components

Identification of marker compounds resulting in the classification of green tea varieties

To demonstrate that metabolites from green tea extract could provide sufficient relevant information on different tea varieties, PLS-DA was applied to classify and predict the hundred tea samples from different cultivated regions. As shown in Fig. 1C and Fig. 2C, both the data sets derived from the GC–MS and UPLC-QTOF/MS provided a conspicuous discrimination of the ten varieties of green tea in the PLS-DA model. The Q² values were 0.89 and 0.91 for the PLS-DA model of GC–MS and UPLC-QTOF/MS derived data set, respectively, which showed superior prediction capability. To screen the important composition for the discrimination model, we checked the VIP values, which reflected the importance for projection. Variables with VIP values larger than 1 were regarded as marker components. As shown in Fig. 1D, oxalic acid, sucrose, quinic acid, L-5-oxoproline, D-glucose, oleamide, L-aspartic acid, myo-inositol, L-glutamine, citric acid, and malic acid were identified as distinct primary metabolites among different green tea varieties. Meanwhile, KGR, caffeine, KRR, KR, QGLR, GCG, ECG, QR, EGCG, QGAR, 5-GQA, KG, and rutin were considered as important substances in the UPLC-QTOF/MS derived PLS-DA model (Fig. 2D).

Overall, organic acid, amino acid, and sugars should be regarded as marker components as they varied greatly among different green tea varieties. Meanwhile, caffeine and flavan-3-ols, as well as some glycosylated flavonols were major distinct secondary metabolites among the ten varieties of green tea. As mentioned thus far, these compounds with a large variance should be used as an important index while evaluating the different green tea varieties.

Identification of characteristic metabolites for the ten varieties of green tea

The comparison of major components from different tea varieties is of great significance for tea quality control and human health. To identify the metabolite markers for each variety of green tea, the coefficient values for the PLS-DA model were checked to uncover the correlativity between the metabolites and different green tea varieties. The characteristic metabolites for the ten varieties of green tea are shown in Table 3.

Table 3.

Characteristic metabolite for ten varieties of green tea

Green tea variety	Primary metabolites		Secondary metabolites
	+	−	+	−
LJ	Ala, Oxo, Ami, Glut	QA, MyA, Asp	GA, 5-GQA, CG, KRR	TB, EGC, TG, QRG
GP	MA, TAT, KLGA, CA	HX, Leu, Isol, Val, Thr, Oxo	QGAR, QRG, QGLR, KG, QR	Cat, PD, EC, KRR, ECG
BLC	Ala, OA, LGA, Myo	KLGA, MyA, Asp	TB, EGC, PD, EC, TG, Rutin, KR	GA, 5-GQA, GC, QGAR, GCG
HK	SA, Glu, MyA	Ala, LGA, Ami, Myo	GC, EGC, Cat, GCG, EC,	5-GQA, EGCG, KGR, QGAR, Rutin,
BC	Ala, Ami, Myo, Suc	OA, Oxo, Asp, TAT, KLGA, MyA	GC, Cat, QRG, QGAR, QGLR	TB, KGR
YH	OA, Oxo, Asp, MyA	Ala, LGA, CA, QA, Glu, Suc	GC, KGR, QR	TG, QGAR, QRG, KR, Rutin,
ZYQ	Leu, Isol, Thr, LGA, QA, Asp, Glu	LA, OA, MA, Ami, TAT, KLGA	TB, PD, TG, QRG	GA, GC, Cat, EC, KRR
MJ	MS, Ami, MyA, OM	LA, OA, Glut, QA	GC, EGCG, GCG,ECG	PD, TG, KRR
MF	MS, LA, Glut, KLGA	Isol, Ami, MyA	GCG, Rutin, KR	GC, TB, EGC, EC, KG, QR
YW	HX, OA, Val, Isol, Thr, MyA, Asp, Suc	LGA, KLGA, ON	TB, EGC, PD, TG, Caf	GA, 5-GQA, QR

All the data were obtained by checking the coefficient values for each class in the PLS-DA model. + , compounds were abundantly detected in green tea. − , compounds were scantily detected in green tea

Compared with the secondary metabolites, the chemical properties of amino acids were more stable in tea plants (Ananingsih et al., 2013). It was reported that the contents of amino acids were negatively related with the climate condition, such as rainfall, temperature, and sun exposure time (Lee et al., 2010). Both ZYQ and YW were cloud teas grown in mountains at an altitude of more than 800 m, and they were found to be rich in amino acids, such as L-Valine, L-Isoleucine, L-Threonine, L-Leucine, and L-Aspartic acid. We deduced that an environment without sunshine might promote the accumulation of amino acids. Meanwhile, Liu et al. (2016) revealed that the contents of amino acids decreased in the green tea leaf grown in the late spring season. Consistently, we also found that GP had low levels of amino acid, which was attributed to the fact that they were harvested until the fourth leaf grew out in late April. Organic acids, such as oxalic acid, quinic acid, and critic acid, varied greatly among different green tea samples, and we speculated that the tea variety and edaphic factor might be the precipitating factors for the difference in content. The contents of Sucrose and D-Glucose were high in YW, ZYQ, BC, and HK, which accounted for the sweet taste of tea infusion.

Geographic factors and climate conditions are dominant external factors that influence the metabolic diversity of green tea plant (Deng et al., 2020; Stilo et al., 2020). Lee et al. (Lee et al., 2010) had shown that tea plants grown in low temperature and short sun exposure time contained high contents of flavanol. The place of origin of MJ was located in the northernmost region of China among the ten green tea samples and the daily average temperature was the lowest; this led to high contents of EGCG, GCG, ECG, and GC in MJ (Table 3). The synthesis of caffeine was reported to be stimulated by relative humidity (Wang et al., 2011). Here, we also found that YW had the highest levels of caffeine due to the cloud environment, as well as due to the rainy climate all year round in Lushan Mountain. Tea species with a large leaf having high levels of quercetin glycosides was regarded as a suitable raw material for black tea manufacturing (Jiang et al., 2015). GP and BC also had high contents of quercetin glycosides, such as QGAR, QGLR, and QRG, which indicated that they might be served as potential candidates for black tea processing.

To reveal the difference in metabolomics among teas from different regions was of vital significance for the quality control of tea plant and products. In this study, a set of high throughput, holistic analysis technologies was developed with the application of metabolomics based on GC/MS and UPLC-QTOF/MS that could be used to analyze the complete metabolites in tea leaves. Both the effects of variety and geographical origins on tea componential phenotypes were revealed with the application of metabolomic analysis coupled with multivariate statistical analysis. The method could be applied in detection of food adulteration to identify the ten varieties of famous green tea from China. Furthermore, the identified markers varied greatly between different tea groups and should serve as important index components in regard to quality control and standard setting regarding the tea quality.

Supplementary Information

Below is the link to the electronic supplementary material.

Declarations

Conflicts of interest

The authors declare that there are no conflicts of interest.