Skip to main content
NCBI home page
Food Chemistry: X logoLink to Food Chemistry: X
. 2026 Apr 16;35:103867. doi: 10.1016/j.fochx.2026.103867

Study on bitterness reduction and aroma enrichment of Kejia green tea based on storage duration through UPLC-MS/GC–MS technology

Ping Wu a,1, He Ni a,1, Suwan Zhang b, Ruohong Chen b, Mengjiao Hao b, Lingli Sun b, Xingfei Lai b, Zhenbiao Zhang b, Shili Sun b,⁎, Qiuhua Li b,⁎
PMCID: PMC13103569  PMID: 42040320

Abstract

Kejia Green Tea (KJGT) exhibits particularly noticeable alterations in bitterness, astringency, and aroma during storage. However, the comprehensive metabolic mechanisms driving this evolution remain poorly understood. This study investigated the dynamic changes of flavor-related metabolites in KJGT during storage using UPLC-MS/GC–MS. Results demonstrated that storage significantly reduced bitter/astringent compounds (e.g., 5-hydroxymethylfurfural, proanthocyanidin B2) by modulating linoleic acid metabolism and quercetin glycoside biosynthesis. Concurrently, the roasted high-fire aroma significantly diminished (e.g., 2-ethoxy-3-methylpyrazine, 2-ethylbenzaldehyde-volatile) during storage, while nutty and woody undertones intensified, which significantly enhanced the tea's aroma. Notably, KJGT-10 exhibited the least bitterness and a more pleasant aroma compared to the other vintages. The study revealed the key material conversions and metabolic mechanisms in KJGT during storage, demonstrating that an appropriate storage duration markedly improved its flavor characteristics. It provided a theoretical basis for the scientific storage and quality regulation, and technical support for the sales of stocked KJGT.

Keywords: Kejia green tea, Storage duration, Flavor quality, Metabolomics

Highlights

  • •

    Kejia green tea stored for 10 years had the best flavor quality.

  • •

    Storage significantly reduced bitter and astringent substances.

  • •

    Seven key volatile metabolites regulated the aroma of aged tea.

  • •

    Storage significantly reduced the high-fire aroma, while enhancing the aroma.

1. Introduction

Green tea, which belongs to the non-fermented category, is one of the six major types of tea and a beverage that can promote health. According to the method of production, green tea is divided into four categories: stir-fried green tea, roasted green tea, steamed green tea, and sun-dried green tea. Among these, stir-fried green tea was the most popular type (Sarma et al., 2023). KJGT is a traditional Chinese stir-fried green tea, highly popular in the southeastern mountainous regions of China where Hakka communities reside, boasting distinct regional characteristics. Research indicated that KJGT offered multiple health benefits, such as reducing lipids and aiding weight loss, as well as improving gastrointestinal function (Chen et al., 2024; Zhou et al., 2022). The production technique of ‘prolonged, repeated high-temperature stir-frying’ imparts KJGT with strong bitterness and astringency, a robustly stimulating flavor profile, and the characteristic aroma profile of high-fired teas. Many consumers outside Hakka-speaking regions perceive KJGT's flavor profile as excessively bitter and pungent, rendering it unpalatable. This perception significantly constrains the product's market expansion.

Local residents' storage practices for KJGT found that aged KJGT developed a more distinctive flavor profile, becoming more mellow and silky-smooth in taste and enhancing in quality. This was similar to the storage characteristics of Pu'er tea. After storage, Pu'er tea exhibited diminished astringency, while its sweetness and mellowness increased (Xu et al., 2025). The chemical and biological reactions that occur during the storage of tea were the core mechanisms that enhance its quality (Liu et al., 2025). During storage, compounds in tea that were sensitive to environmental factors were transformed, leading to changes in aroma and taste, all of which could cause changes in tea quality (Xu et al., 2019). Prolonged storage had been shown to affect the flavor and quality of green tea (Liu, Zhang, et al., 2023). KJGT exhibited caramelization characteristics due to its “heavy roasting” technique, which imparted storage resistance and developed a distinctive “roasted rice aroma” (Zhu, Niu, & Xiao, 2021). Kejia green tea was stored through biological oxidation and weak oxygen fermentation, gradually forming a unique flavor such as aged, fruity and woody, with reduced bitterness and astringency, enhanced mellowness and warmth (Ocieczek et al., 2023). This characteristic makes it unique in the green tea category, with both cultural and economic attributes. However, the storage value of KJGT remains largely a folk tradition, with the material basis and molecular mechanisms underlying quality changes during storage still unclear, lacking theoretical underpinnings and in-depth analysis.

At present, research on the changes in tea components during long-term storage of tea has mainly focused on white and dark tea. During the storage of white tea, changes in flavonoids and amino acids promoted the development of aged aromas (Zhang et al., 2024). Storage facilitates the degradation of carbohydrates and their interaction with theanine, resulting in a reduction of gis-flavanols. This process leaded to the development of a mellow taste and aged aroma in Oolong tea (Peng et al., 2022). Raw pu-erh tea usually required long-temm aging under natural conditions to make it smoother and fuller-bodied (Ren et al., 2022a). A study of the composition of volatile compounds in black teas stored for 1 to 20 years showed an increase in methoxybenzene volatiles after 4 years of storage, which could lead to an aged flavor (Meng et al., 2021).

Ultra-Performance Liquid Chromatography-Mass Spectrometry (UPLC-MS) technology, with its high sensitivity and resolution, enables precise identification of complex flavor compounds in tea (Borrull et al., 2025). It serves as a key tool for in-depth research into the mechanisms underlying tea flavor formation. Gas chromatography-mass spectrometry (GC–MS) is widely employed in tea aroma research (Wen et al., 2023), and rOAV is a standard metric for quantifying the aroma contribution of volatiles (Liu et al., 2021). Therefore, this study employed a complementary approach combining UPLC-MS and GC–MS to analyze both volatile and non-volatile metabolites, comprehensively capturing the metabolic trajectories during tea storage. This revealed the material basis for quality differences among KJGT samples across three distinct storage years. Findings indicate that appropriate storage duration can significantly enhance KJGT's flavor profile. This provided a theoretical basis for scientific storage and quality regulation, evidence-based guidelines for consumer decision-making, and practical technical solutions for inventory management-effectively addressing market perception challenges and driving the industry's transition toward premium, sustainable development.

2. Materials and methods

2.1. Tea samples preparation

The Kejia green tea samples comprise three distinct storage periods: fresh tea (KJGT-0), aged for 10 years (KJGT-10), and aged for 20 years (KJGT-20) were prepared from Meizhou City, Jiaoling County, Guangdong Province, and provided by Meizhou Junbao Industrial Co., Ltd. All samples shared identical cultivation conditions, plantation location, and picking standards (one bud and two leaves). Traditional processing techniques were employed as follows. (1) fresh leaf picking, (2) spreading for withering, (3) fixation (in a heated wok at approximately 220–240 °C for 3–5 min), (4) rolling (by hand or mechanical roller for 15–20 min), (5) initial pan-firing, (6) repeated pan-firing, and (7) final drying (at 70–80 °C until the moisture content reached below 6%). All samples have been stored at 10–25 °C with a relative humidity of less than 50% and light-proof tea warehouse for 0, 10, and 20 years, respectively.

2.2. Sensory evaluation

Sensory evaluation of eleven samples followed Chinese National Standard GB/T 23776–2018. An eleven-member panel of certified tea experts conducted the assessments. Before the evaluation, consent was secured from all participants. The quality of KJGT samples is evaluated based on five aspects, scored on a 100-point scale: shape (20%), liquor color (10%), aroma (30%), taste (35%), and tea residue (5%). Human sensory evaluation was conducted in accordance with the Experimental Ethics Committee of the Tea Research Institute, Guangdong Academy of Agricultural Sciences, approval number: TRI-GDAAS-2025-000709. All human experiments adhered to the appropriate protocols for protecting the rights and privacy of all participants were utilized during the execution of the research. Informed consent was acquired from all participants in the tea review prior to the commencement of the sensory evaluation analysis. Participants could voluntarily withdraw from the evaluation at any moment without any explanation.

2.3. Evaluation of taste based on electronic tongue

Weigh 3.0 g of tea leaves in a review cup, fill 150 mL of boiling water with a lid, and brew for 5 min, filter the tea residue with gauze and cool it down to room temperature, take 100 mL of the filtrate, and then test it on the machine under the condition of room temperature of about 25 °C. An electronic tongue system (TS-5000Z, INSERT, Japan) equipped with an array of six chemical sensors including AAE (umami), CT0 (Saltiness), CA0 (sourness), C00 (bitterness), AE1 (astringency), GL1 (sweetness), 30 mM KCl and 0.3 mM tartaric acid as reference electrodes were used. Each tea sample was infused twice and measured three times per infusion. All data are absolute output values with artificial lipid membrane sensor (reference solution: 30 mM potassium chloride +0.3 mM tartaric acid) as the standard. The state of the artificial saliva in the electronic tongue test simulates the state of the human mouth when only saliva is present.

2.4. Broadly targeted metabolomic analysis of metabolites in KJGT by UPLC-MS

After freeze-drying and grinding the dry tea samples into powder, vortex extraction was performed at a ratio of 50 mg powder to 1200 μL pre-chilled methanol extract. Following centrifugation and filtration, the supernatant was obtained for UPLC-MS/MS analysis. The sample extracts were analyzed using a UPLC-ESI-MS/MS system (UPLC, ExionLC™ AD, https://sciex.com.cn/) and a Tandem mass spectrometry system (https://sciex.com.cn/). The analytical conditions were as follows, UPLC: column, Agilent SB-C18 (1.8 μm, 2.1 mm*100 mm); The mobile phase consisted of solvent A, pure water with 0.1% formic acid, and solvent B, acetonitrile with 0.1% formic acid. Sample measurements were performed with a gradient program that employed the starting conditions of 95% A, 5% B. Within 9 min, a linear gradient to 5% A, 95% B was programmed, and a composition of 5% A, 95% B was kept for 1 min. Subsequently, a composition of 95% A and 5.0% B was adjusted within 1.1 min and kept for 2.9 min. The flow velocity was set as 0.35 mL per minute; The column oven was set to 40 °C; The injection volume was 2 μL. The effluent was alternatively connected to an ESI-triple quadrupole-linear ion trap (QTRAP)-MS.

The ESI source operation parameters were as follows: source temperature 550 °C; ion spray voltage (IS) 5500 V (positive ion mode)/−4500 V (negative ion mode); ion source gas I (GSI), gas II (GSII), curtain gas (CUR) were set at 50, 60, and 25 psi, respectively; the collision-activated dissociation (CAD) was high. QQQ scans were acquired as MRM experiments with collision gas (nitrogen) set to medium. DP (declustering potential) and CE (collision energy) for individual MRM transitions were done with further DP and CE optimization. A specific set of MRM transitions was monitored for each period according to the metabolites eluted within this period.

2.5. Volatile metabolomics analysis of KJGT by GC–MS

Samples were removed from the −80 °C refrigerator for liquid nitrogen grinding. 500 mg (1 mL) of the powder was transferred immediately to a 20 mL headspace vial (Agilent, Palo Alto, CA, USA), containing NaCl saturated solution, to inhibit any enzyme reaction. The vials were sealed using crimp-top caps with TFE‑silicone headspace septa (Agilent). At the time of SPME analysis, each vial was placed at 60 °C for 5 min, then a 120 μm DVB/CWR/PDMS fibre (Agilent) was exposed to the headspace of the sample for 15 min at 60 °C.

After sampling, desorption of the VOCs from the fibre coating was carried out in the injection port of the GC apparatus (Model 8890; Agilent) at 250 °C for 5 min in the splitless mode. The identification and quantification of VOCs were carried out using an Agilent Model 8890 GC and a 7000E mass spectrometer (Agilent), equipped with a 30 m × 0.25 mm × 0.25 μm DB-5MS (5% phenyl-polymethylsiloxane) capillary column. Helium was used as the carrier gas at a linear velocity of 1.2 mL/min. The injector temperature was kept at 250 °C. The oven temperature was programmed from 40 °C (3.5 min), increasing at 10 °C/min to 100 °C, at 7 °C/min to 180 °C, at 25 °C/min to 280 °C, and held for 5 min. Mass spectra were recorded in electron impact (EI) ionization mode at 70 eV. The quadrupole mass detector, ion source, and transfer line temperatures were set, respectively, at 150, 230 and 280 °C. The MS was selected ion monitoring (SIM) mode was used for the identification and quantification of analytes.

2.6. The rOVA analysis method

The rOAV is the ratio of the concentration of volatile metabolites in aqueous solution to their odor thresholds. Following the relevant literature (Huang, Fang, et al., 2022; Xue et al., 2022), the rOAV is calculated as: rOAVi = Ci / Ti. where rOAVi is the relative odor-activity value of compound i, Ci is its relative content (ug/g or ug/mL), and Ti is its odor threshold (ug/g or ug/mL). The thresholds for the volatile organic compounds were taken from published compilations (Yang et al., 2022; Zhu, Niu, & Xiao, 2021), they represent the lowest concentration at which an odor becomes perceptible and are provided for reference only. rOAV is commonly used to evaluate the contribution of volatile metabolites to characteristic aroma. Setting rOAV ≥1 as the screening criterion allows identification of the key compounds responsible for sweet-aroma formation in different tea types.

2.7. Statistical analysis

Unsupervised PCA (principal component analysis) was performed by statistics function prcomp within R (www.r-project.org). The data was unit variance scaled before unsupervised PCA. The HCA (hierarchical cluster analysis) results of samples and metabolites were presented as heatmaps with dendrograms, while Pearson correlation coefficients (PCC) between samples were calculated by the cor function in R and presented as only heatmaps. Both HCA and PCC were carried out by the R package ComplexHeatmap. For HCA, normalized signal intensities of metabolites (unit variance scaling) are visualized as a color spectrum.

Identified metabolites were annotated using the KEGG Compound database (http://www.kegg.jp/kegg/compound/), and annotated metabolites were then mapped to the KEGG Pathway database (http://www.kegg.jp/kegg/pathway.html). Pathways with significantly regulated metabolites mapped were then fed into MSEA (metabolite sets enrichment analysis), their significance was determined by hypergeometric test's p-values.

3. Results and discussion

3.1. Sensory evaluation of KJGT with different storage years

As shown in the sensory evaluation results (Fig. 1A and Table S1), the new tea has a stronger fruity aroma and a more intense freshness taste. After storage, the Kejia green tea had an aged aroma, while the taste became mellow, and the color of the tea broth will change from yellow-green to orange-yellow. Compared with KJGT-20, the aroma of KJGT-10 was stronger and more persistent, and the taste was more mellow. When Kejia green tea was stored for different periods of time, the soup color, aroma and taste changed significantly (Liu, Zhuang, et al., 2023). Flavanols and sugars in tea formed flavanol glycosides, which were usually bright yellow, and their oxidation products were orange to brownish red, contributed to the color of green tea broths (Wang & Ruan, 2009).

Fig. 1.

Fig. 1

Open in a new tab

Sensory quality analysis of Kejia green tea with different storage years. (A) Results of the traditional sensory review; (B) Radar chart of the analyses of the individual flavors of KJGT. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

The electronic tongue e-tongue provides a more objective, acute, and digital description of taste than the human palate. All three tea samples had acidity levels below the no-taste point, which suggested that there was no effective acidity in the samples, and all other taste indicators were above the no-taste point, which is an effective taste indicator (Table S2). As shown in (Fig. 1B), KJGT-0 scored highest for bitterness, astringency, umami, saltiness, and richness among the three tea samples with varying storage years; the other tastes were no significant differences. KJGT-10 had the highest sweetness score and the lowest scores for bitterness and its aftertaste, astringency, and its aftertaste and richness. KJGT-20 had the highest aftertaste- astringency and aftertaste- bitterness scores and the lowest umami, saltiness, and sweetness scores.

The study found that compared to KJGT-0, KJGT-10 exhibited reduced bitterness and astringency, which enhanced its mellow flavor, while KJGT-20 showed diminished freshness. Bitterness and astringency were two important quality attributes of green tea (Deng et al., 2022).

3.2. Non-volatile metabolites analysis of KJGT of different storage years

3.2.1. Determination of total non-volatile metabolites among three different KJGTs

To gain insight into the non-volatile metabolite changes in KJGT of different storage years, we identified metabolites in the KJGT samples using a wide range of targeted metabolomics techniques on a UPLC-MS platform. The curve overlap of the total ion chromatogram (TIC) detected by the mass spectrometry of the quality control samples (QC) was high, indicating the high stability of UPLC-MS/MS (Fig. S1A and B). The proportion of substances with a coefficient of variation (CV value) of QC < 0.5 was higher than 85%, and the proportion of substances with a CV value <0.3 was higher than 75%, indicating that the experimental data was stable (Fig. S1C), and the Pearson correlation coefficient for the non-volatile metabolomic samples was above 0.9, which indicated that the samples were reproducible (Fig. S1D). A summary of 2228 non-volatile metabolites covering 13 categories were identified from three different storage years of KJGT by matching MS/MS spectral information in the established databases with the standards in MetWare's proprietary metabolite database (Fig. 2A).

Fig. 2.

Fig. 2

Open in a new tab

Overview of all non-volatile metabolites in KJGT of three storage years. (A)Total metabolite category composition ring chart: Each color represents a metabolite category, with the area of the color block indicating the proportion of that category. (B) Bar chart of relative percentage composition of various metabolites in KJGT samples. (C) Principal component analysis (PCA) plot. PC1 denotes the first principal component, PC2 denotes the second principal component, and the percentage denotes the explanation rate of that principal component for the dataset; each point in the plot denotes a sample, and samples from the same group are represented using the same color. (D) Clustering heat map. (E) Three groups of differential metabolites Venn diagram.

Based on the content and composition ratio of metabolites, it was found that the classes of flavonoids, phenolic acids, lipids, amino acids, alkaloids and derivatives were significantly represented by the highest content of total metabolites in the three different teas (64.95%) (Fig. 2B). Specifically, KJGT-0 exhibited a lower content of total non-volatile metabolites but a higher content of flavonoids, quinones, nucleotides and derivatives in comparison to the other two tea samples, indicating that the new tea possessed the most pronounced bitterness and astringency. Flavonoid and phenolic acid compound content has been shown to be positively correlated with the perceived bitterness of tea (Wang et al., 2023). The oxidation of polyphenols resulted in the formation of quinones, which underwent further oxidation to yield theaflavins, these compounds were important for the umami and crisp taste of tea (Long et al., 2023). KJGT-10 had the highest total volatile metabolite content and was more prominent in phenolic acids, lipids, alkaloids, lignans coumarins, and steroids. Both lignans and coumarins belong to the phenylpropane group of substances, similar to phenols, and were characterized by some bitterness and astringency (Fan et al., 2021). KJGT-20 had the highest content of terpenes, organic acids, tanninsamino acids, and derivatives. Amino acids, nucleotides, and derivatives are vital substances that contribute to the umami and crisp sensory experience associated with tea (Huang, Lu, et al., 2022).

A principal component analysis (PCA) was conducted to examine overall variances in non-volatile metabolites among three different storage years of KJGTs. The variances of the two main components (PC1 and PC2) were 50.35% and 25.12% respectively, indicating that the three groups of tea samples showed significant metabolomic differences mainly on the PC1 axis (Fig. 2C). KJGT-0 was located in the negative region of PC1, while KJGT-10 and KJGT-20 were located in the positive region of PC1, which showed that the storage process significantly affected the accumulation of non-volatile metabolites in KJGTs. KJGT-10 was located in the negative region of PC2, while KJGT-20 was located in the positive region of PC2, indicating that different storage times also significantly affected the non-volatile metabolites of KJGTs. Additionally, hierarchical cluster analysis (HCA) was also used to analyze the degree of similarity between classes. Different biological replicates clustered similarly, showing good homogeneity between biological replicates and high data reliability (Fig. 2D). The results demonstrated greater biological repeatability within groups, greater heterogeneity between groups and greater differences in metabolites, both PCA and HCA revealed noteworthy variations in non-volatile metabolites in all three different groups, with the greatest difference between KJGT-0 and the other groups.

3.2.2. Identify differential metabolites and analyze their relative variation

In order to clarify the effect of storage on the metabolites of KJGT. Score plots generated from pairwise analyses of KJGT-0, KJGT-10, and KJGT-20 using Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA) showed that Q2 was above 0.9 for all comparison groups, suggesting that the constructed model was appropriate. The identification of differential metabolites was based on variable importance in the project (VIP) ≥ 1 and fold change ≥2 or ≤ 0.5. 925 significant differential metabolites were screened out (Table S3).

There were 590 different non-volatile metabolites in the KJGT-10 versus KJGT-0 comparison (P < 0.05), with 307 different metabolites down-regulated and 283 different metabolites up-regulated, of which 31 substances had taste properties (Fig. 3A and Table 1). A total of 651 differential metabolites were screened in the comparison of KJGT-20 with KJGT-0 (P < 0.05), with 363 significantly down-regulated and 288 significantly up-regulated, including 34 substances had taste properties (Fig. 3B and Table 1). A total of 313 differential metabolites were screened in the comparison of KJGT-20 with KJGT-10 (P < 0.05), with 132 significantly down-regulated and 181 significantly up-regulated, including 16 substances had taste properties (Fig. 3C and Table 1). The Venn diagram analysis of differential metabolites revealed 52 metabolites shared among the three groups (Fig. 2E), indicating significant differences in the differential metabolites of KJGT across the three storage periods. Effective storage can significantly alter both the number and relative abundance of differential metabolites in KJGT.

Fig. 3.

Fig. 3

Open in a new tab

Screening of differential non-volatile metabolites among three different storage years of KJGT. (A-C) Differential metabolite volcano plot. (D—F) KEGG enrichment pathway analysis.

Table 1.

Differences flavor metabolites of Kejia green tea with three different storage years

NO. Compounds CAS taste
characteristics		 Recognition threshold
(mg/kg)		 KJGT-10 vs
KJGT-0		 KJGT-20 vs
KJGT-0		 KJGT-20 vs
KJGT-10
1 1,7-Dimethylxanthine 611–59-6 bitter 90–160 up up –
2 2′,4′,6′-Trihydroxyacetophenone 480–66-0 bitter 170–500 down – –
3 2-Picolinic acid 98–98-6 bitter 620–860 – down down
4 3,4-Dihydroxybenzoic acid (Protocatechuic acid) 99–50-3 astringent 31–32 up – –
5 3-Hydroxybutyric acid 300–85-6 bitter >10,400 up up –
6 4-Hydroxybenzoic acid 99–96-7 astringent 92 up up –
7 5-Hydroxymethylfurfural 67–47-0 bitter 630–1890 down down –
8 6-Aminocaproic acid 60–32-2 bitter 3670–4200 – – down
9 Azelaic acid 123–99-9 sour 188 up up –
10 Benzamide 55–21-0 bitter 97–121 down down –
11 Coumarin 91–64-5 bitter 29–73 – – down
12 Cyclic 3′,5′-Adenylic acid 60–92-4 umami 32,900 down down –
13 d-Fructose 6-Phosphate 643–13-0 sweet 570 – up –
14 D-Galacturonic acid 685–73-4 sour 125 up up –
15 Gallic acid 149–91-7 astringent 46.2–50 up up –
16 Gallocatechin 970–73-0 astringent 165 – down –
17 Guanine 73–40-5 bitter >760 up – –
18 Isonicotinic acid 55–22-1 bitter 3690–4920 – – down
19 l-Ascorbic acid
(Vitamin C)		 50–81-7 sour 120 up up –
20 l-Glutamine 56–85-9 salt 7300 down down –
21 L-Glycyl-L-isoleucine 19,461–38-2 bitter 410 down down –
22 L-Histidine 71–00-1 bitter 6980–7760 – down –
23 L-Tyramine 51–67-2 bitter 274–343 – up up
24 L-Tyrosine methyl ester 1080-06-4 bitter 27,900 down down –
25 Leu-Asp 32,949–40-9 bitter 1480 – – down
26 Malonic acid 141–82-2 sour 68 up up –
27 Methyl gallate 99–24-1 astringent 0.232 mM down – up
28 N-Glycyl-L-leucine 869–19-2 bitter 3600–4300 down down down
29 Nicotinamide 98–92-0 bitter 730–980 – down –
30 Nicotinic acid (Vitamin B3) 59–67-6 bitter 2460–3800 – – down
31 Phloroglucinol; 1,3,5-Benzenetriol 108–73-6 sweet 63–126 down down –
32 Piperidine 110–89-4 bitter 682–1022 – – down
33 Procyanidin B2 29,106–49-8 astringent 110–120 – down –
34 Procyanidin C1 37,064–30-5 astringent 260 down down –
35 Prolylproline 20,488–28-2 bitter 4.5 mM – – down
36 Pyrogallol 87–66-1 astringent 9.8 down down up
37 Pyrrole-2-carboxylic acid 634–97-9 bitter 3330–5550 down down up
38 Quercetin-3-O-rutinoside (Rutin) 153–18-4 astringent 1220–1830 down – –
39 Suberic Acid 505–48-6 sour 126 up up –
40 Terephthalic acid 100–21-0 sweet >6640 – – down
41 Theaflavic acid 30,407–93-3 astringent 0.024 mM up – –
42 Theaflavin 4670-05-7 astringent 750–1000 down down –
43 Theaflavin-3’-Gallate 28,543–07-9 astringent 0.015 mM down down –
44 Theaflavin-3,3′-di-O-gallate 30,462–35-2 astringent 0.013 mM down down –
45 Theaflavin-3-gallate 30,462–34-1 astringent 0.015 mM down down –
46 Theophylline 58–55-9 bitter 110–160 – up –
47 Thymine 65–71-4 bitter 440–630 up up –
48 Tryptamine 61–54-1 bitter 320–400 – down down
49 Xanthine 69–89-6 bitter >3040 – – down
Open in a new tab

In comparison to KJGT-0, KJGT-10 and KJGT-20 markedly changed 395 differential metabolites (P < 0.05). The top 20 metabolites with the highest VIP values in each comparison group were compared and analyzed (Fig. S2A and B). Differential metabolites significantly up-regulated in stored KJGTs (KJGT-10 and KJGT-20) included 3-hydroxy-3,7,11-trimethyldodeca-1,6E,10-trien-9-isobutyrate, 14,15-dehydrocrepenynic acid, azelaic acid, lysophosphatidylcholine 19: 1 and 9,10-dihydroxy-12,13-epoxyoctadecanoic acid, and significantly down-regulated differential metabolites included eriodictyol-7-O-(6″-acetyl)glucoside, acacetin-7-O-rutinoside (linarin), 5-hydroxymethylfurfural, and 1-(9Z-Octadecenoyl)-2-(9-oxo-nonanoyl)-sn-glycero-3-phosphocholine. Among these, 5-hydroxymethylfurfural could reduce organic compounds produced by sugar through the Meladic reaction and had a bitter taste that could be used as a chemical marker for tea processing and storage (Pérez-Burillo et al., 2019). Azelaic acid had sour taste, Dancong tea increased the acidic constituents such as azelaic acid and octanedioic acid in aged tea mainly by regulating phenylalanine metabolism and pyrimidine metabolism (Sun et al., 2023a). In addition, we analyzed the differences in non-volatile metabolites between KJGT-10 and KJGT-20, and found 313 non-volatile differential metabolites (P < 0.05), including 16 aroma metabolites), primarily phenolic acids, flavonoids, and lipids. The top 20 metabolites for VIP values were analyzed and only two substances had taste attributes, methyl gallate for astringency and 2-pyridinecarboxylic acid for bitterness (Fig. S2C). However, the flavor thresholds of the other substances remain uncertain and further analysis of the flavor metabolites is required.

All differential metabolites were referenced against flavor databases, identifying 49 metabolites exhibiting flavor attributes (Table 1). Among these, bitter and astringent compounds constituted 79.59%, indicating significant differences in bitterness and astringency among the three tea samples. The analysis focused on comparing non-volatile metabolites with high-fold differential changes. Compared to KJGT-0, L-tyrosine methyl ester, L-glycyl-L-isoleucine, benzamide, pyrrole-2-carboxylic acid, 5-hydroxymethylfurfural, and N-glycyl-L-leucine were significantly reduced in both KJGT-10 and KJGT-20. In KJGT-10, the reduction factors were 6.06, 2.14, 2.27, 10.59, 8.10, and 3.24-fold, respectively. In KJGT-20, the reduction factors were 6.06, 2.26, 3.88, 4.08, 8.10, and 19.35-fold, respectively. Furthermore, the tryptamine content in KJGT-20 is 5.69 -fold lower than that in KJGT-0. Among these, piperidine, prolylproline, Leu-asp, isonicotinic acid, nicotinic acid (vitamin B3), 6-aminocaproic acid, xanthine, coumarin, and 2-picolinic acid, all of which impart a bitter taste, were reduced by more than twofold compared to KJGT-10. The other compounds with high thresholds and low concentrations contributed less to the flavor of tea. The results indicate that KJGT storage significantly reduced the bitter taste present in KJGT-0, with KJGT-20 exhibiting the weakest bitterness. This finding was somewhat inconsistent with the electronic tongue results, as there might have been other bitter-tasting metabolites that remain unidentified. Bitterness in organic green tea decreased with years of storage (Wen et al., 2024). Aged Dancong tea reduced bitterness mainly by modulating the arginine and proline metabolic pathways to reduce bitter characterising substances such as benzamide (Sun et al., 2023b). Pu-erh tea had a high content of tryptamines and other biogenic amines that presented a bitter taste (Tan et al., 2024). Reduced levels of theaflavins, cis-catechins, etc. could reduce the bitterness of Longjing tea (Shan et al., 2024).

Additionally, compared to KJGT-0, the astringent compounds include theaflavin-3′-gallate (TF-3′-G), theaflavin-3-gallate (TF-3-G), theaflavin-3,3′-di-O-gallate, procyanidin C1, theaflavin (TF1), and pyrogallol. The relative reduction in KJGT-10 was 3.48, 3.34, 2.52, 2.27, 3.23, and 6.70-fold, respectively, while in KJGT-20 it was reduced by 4.24, 4.15, 2.40, 2.22, 5.08, and 3.34-fold, respectively. Compared to KJGT-0, other astringent compounds in KJGT-10-including quercetin-3-O-rutinoside (rutin) and methyl gallate-also exhibited significantly reduced levels, with decreases exceeding twofold. This indicates that KJGT storage markedly reduced the astringent taste present in KJGT-0, with KJGT-10 exhibiting the weakest astringency, consistent with electronic tongue results. The results indicated that hydrolysis of galloylated catechins was the key reason for astringency decrease during storage of Crude Pu-erh tea (Ren et al., 2022b). The introduction of black tea fermentation into Oolong tea processing significantly reduced the level of methyl gallate, thereby reducing astringency (Hao et al., 2024). Flavonoids and phenolic compounds were the main components of plant astringency, and when the proteins of the oral mucosa were coagulated, the astringent sensation was produced (Wei et al., 2023). The acids azelaic acid, D-galacturonic acid, ascorbic acid (vitamin C), malonic acid and octanedioic acid were significantly increased in both KJGT-10 and KJGT-20 compared to KJGT-0. This was consistent with the trend of acidic substances in Oolong tea in different storage years (Sun et al., 2023a). Additionally, cyclic 3′,5′-adenylic acid and l-glutamine, which have a fresh and savory flavor, were also reduced in the stored teas (KJGT-10 and KJGT-20), but the thresholds were too high to have a small effect on the flavor of the tea infusion. The sweetness characteristic component d-fructose-6-phosphate increased 1.54 and 2.54-fold in KJGT-10 and KJGT-20, respectively. Natural storage promoted raw Pu-erh tea (RaPT) aging along with chemical conversion. These transformations generated significantly decreased umami, bitterness, and astringency, while significantly increased sourness and mellowness (Ma et al., 2025). This was highly consistent with the changes observed in KJGT during storage.

In conclusion, the bitterness and astringency of stored Kejia green tea were significantly reduced. Stored teas (KJGT-10 and KJGT-20) significantly reduces the content of bitter substances such as 5-hydroxymethylfurfural, benzamide, nicotinamide and tryptamine, as well as the content of astringent substances of catechins such as gallocatechin, methyl gallate, proanthocyanidin B2 and quercetin-3-O-rutinoside (rutin), which contributes to the mellowness and smoothness of the tea broth. The acids azelaic acid, D-galacturonic acid, ascorbic acid (vitamin C), malonic acid and octanedioic acid and the sweet flavor component d-fructose-6-phosphate were significantly increased in KJGT-10 and KJGT-20. Another comparison between KJGT-10 and KJGT-20 showed that the astringency of KJGT-10 would be significantly lower than that of KJGT-20, but the content of some bitter compounds would be higher. This is broadly consistent with the results of sensory evaluation and electronic tongue testing.

3.2.3. KEGG functional annotation and enrichment of differential metabolites

Pathway enrichment analysis of differential metabolites of KJGTs from different storage years using KEGG databases to obtain the distribution of differential metabolites in different metabolites and biosynthetic pathways. A total of 10 MetMap pathways, 65 metabolic pathways, 1 genetic information processing pathway, and 1 environmental information processing pathway were enriched in the comparison of KJGT-10 and KJGT-0 (Fig. 3D). Among the significantly enriched pathways with greater than five differential metabolites included purine metabolism, linoleic acid metabolism, zeatin biosynthesis and biosynthesis of quercetin aglycones I. The purine metabolism pathway is critical for tea flavor formation, with caffeine, theobromine, and theophylline serving as the primary purine alkaloids (Zhou et al., 2020). The linoleic acid metabolism and flavonoid biosynthesis pathways are also significant. Studies indicate that differentially expressed genes and differentially accumulated metabolites associated with linolenic and linoleic acid metabolism are significantly enriched in tea plants (Liu, Zhang, et al., 2023). Comparing KJGT-20 and KJGT-0, a total of 10 MetMap pathways, 73 metabolic pathways, 1 genetic information processing pathway and 1 environmental information processing pathway were enriched (Fig. 3E). Among the significantly enriched pathways with greater than five differential metabolites included tyrosine metabolism, linoleic acid metabolism, flavonoid biosynthesis and biosynthesis of quercetin aglycones I. The biosynthesis pathways centered on the quercetin aglycone core, such as anthocyanin biosynthesis, play a crucial regulatory role in the accumulation of flavonoids in the tea plant (Wang et al., 2017). Comparison of KJGT-20 and KJGT-10 enriched a total of 9 MetMap pathways, 33 metabolic pathways (Fig. 3F). Among the significantly enriched pathways with greater than five differential metabolites included the biosynthesis of various plant secondary metabolites and the biosynthesis of secondary metabolites.

The significant pathways co-enriched by the three tea samples of different storage years were the linoleic acid metabolism and biosynthesis of quercetin aglycones I, suggesting that the differences in the relative contents of metabolites in KJGT during storage may be caused by the reactions of linoleic acid metabolism and biosynthesis of quercetin aglycones I (Fig. S3A—B). Unsaturated fatty acids, especially linolenic and linoleic acids, contribute to aroma formation through oxidation and decomposition to form precursors for aliphatic aroma compounds (Ho et al., 2015). The metabolite differences between KJGT-20 and KJGT-10 are mainly due to biosynthesis with secondary metabolism. Primary metabolites were essential for plant growth and development, and the biosynthesis of secondary metabolites began with basic pathways such as glycolysis or the manganic acid pathway (Patra et al., 2013).

3.3. Volatile metabolite in different storage years of KJGT

3.3.1. Analysis of volatile metabolites in KJGT of three storage years by GC–MS

To test the correlation between the aroma of KJGTs and different years of storage, we analyzed KJGT extracts using GC–MS. The curve overlap of the total ion chromatogram (TIC) detected by the mass spectrometry of the quality control samples (QC) was high, indicating the high stability of UPLC-MS/MS (Fig. S4 A). The proportion of substances with a coefficient of variation (CV value) of QC < 0.5 was higher than 85%, and the proportion of substances with a CV value <0.3 was higher than 75%, indicating that the experimental data was stable (Fig. S4B), and the Pearson correlation coefficient for the non-volatile metabolomic samples was above 0.9, which indicated that the samples were reproducible (Fig. S4C). Principal Component Analysis (PCA) was performed and the first principal component (PC1) accounted for 48.67% of the features in the original dataset, and the second principal component (PC2) accounted for 27.21% of the features in the original dataset, with the total contribution of these principal components being more than 70%, which effectively captured the overall variation in the dataset (Fig. 4A). As shown by the results of the principal component analysis, the separation of the three tea samples was obvious and the distribution areas were consistent with the non-volatile metabolites, indicating that there were significant differences in the metabolites in the tea samples from different years. The degree of similarity between classes was analyzed using HCA plots, where different biological replicates clustered similarly, indicating good homogeneity and high data reliability between biological replicates (Fig. 4B). Both cluster analysis and principal component analysis showed that the metabolites produced significant differences in the different groups.

Fig. 4.

Fig. 4

Open in a new tab

Overview of all volatile metabolites in KJGT of three different storage years. (A) Principal component analysis (PCA) plot. (B) Clustering heat map. (C) Volatile metabolite category composition ring chart. (D) Relative percentage composition of various metabolites in KJGT samples.

The aroma of tea was the result of the combined action of different concentrations of volatile substances (Qiu et al., 2023). In this research, using the GC–MS platform and database, a total of 839 metabolites were detected (Fig. 4C). An isotopic internal standard was selected and three replicate tests were conducted. The results indicated that terpenoids had the largest relative content (23.48%), followed by esters, heterocyclic compounds, ketones and aldehydes, which together accounted for more than 70% of the total volatile matter content. Among the three KJGTs with different storage years, KJGT-10 had significantly lower content of volatile metabolites compared to the other two teas, KJGT-0 had the highest contents of heterocyclic compounds, aldehydes, and phenols, and KJGT-20 had the highest contents of terpenes, ketones, alcohols and hydrocarbons (Fig. 4D). Terpene alcohols had ‘floral’ and ‘fruity’ aroma characteristics and exerted a great influence on the formation of tea aroma (Wang et al., 2022).

3.3.2. Changes in volatile differential metabolite

Differential metabolite analysis was performed to further investigate the effect of different storage years on KJGT volatiles. Screening for differential metabolites was based on fold change >2 and VIP > 1. A total of 456 differential metabolites were identified, including 14 types of metabolites, of which 109 terpenes, 80 esters, 68 heterocyclic compounds, 49 ketones, 34 aldehydes, 30 aromatics, 25 alcohols, 19 hydrocarbons, 11 amines, 10 phenols, 8 acids, 7 sulphur-containing compounds, 2 nitrogen-containing compounds (Table S4).

A total of 265 differential volatile metabolites were screened in the comparison between KJGT-10 and KJGT-0 (P < 0.05). 138 volatile metabolites were significantly down-regulated and 127 volatile metabolites were significantly up-regulated, of which 139 substances had aroma properties (Fig. 5A). In the comparison between KJGT-20 and KJGT-0, a total of 341 volatile metabolites were screened (P < 0.05), of which 100 metabolites were significantly down-regulated, 241 were significantly up-regulated and 165 substances had aroma properties (Fig. 5B). In addition, differential metabolites between KJGT-20 and KJGT-10 were compared and a total of 200 volatile metabolites were screened (P < 0.05), of which 33 metabolites were significantly down-regulated, 167 were significantly up-regulated, and 109 substances had aroma properties (Fig. 5C). It demonstrates that effective storage can significantly alter both the quantity and relative content of volatile metabolites in KJGT.

Fig. 5.

Fig. 5

Open in a new tab

Screening of differential volatile metabolites and flavoromics analysis in three different storage years of KJGT. (A-C) Volcano map. (D—F) Differential metabolite flavor wheel: Display the top 10 sensory flavors with the highest number of annotations.

Additionally, from the differential metabolites identified based on screening criteria within each comparison group and their annotated sensory flavor characteristics, a total of 230 differential metabolites with flavor attributes were selected (Table S4). The top 10 sensory flavors with the highest number of annotations were chosen for presentation on a flavor wheel (Fig. 5D-F). The sensory flavor differences across the three comparison groups encompassed odors such as green, fruity, sweet, woody, floral, herbal, and citrus. Among these, green, sweet, and fruity notes accounted for 53% of the total flavor metabolites, indicating that the primary aromatic shifts in KJGT from the three groups with differing storage durations were grassy, sweet, and fruity odors (Fig. S5A—C). Compared to KJGT-0, KJGT-10 exhibited a significant reduction in volatile metabolites contributing to the unpleasant green odor, with only 16.67% of compounds being upregulated. Among these, 2,6-nonadienol, 1,1-dimethoxyheptane, (Z)-3-hexenoate of n-valeric acid, and (Z)-6-nonanal. This markedly diminished the green odor in KJGT-10. Subsequently, the relative odor activity value (rOAV) of the tea sample was analyzed. The rOAV is a method for identifying key flavor compounds in foodstuffs, established by combining compound sensory thresholds. It serves to elucidate the contribution of each aroma compound to the overall aromatic profile of the sample. Seven differential metabolites with rOAV ≥1 in all three tea samples were identified as characteristic flavor volatile metabolites, including trans-beta-ocimene, (E)-4-nonenal, (E, E)-2,4-undecadienal, benzoic acid methyl ester, 2-ethyl-3,5-dimethyl-pyrazine, alpha-ionone and trans-nerolidol. Significant up-regulation of (E, E)-2,4-undecadienal, benzoic acid methyl ester, 2-ethyl-3,5-dimethyl pyrazine, trans-nerolidol and alpha-ionone were highly contributed in KJGT-10 and KJGT-20 compared to KJGT-0. Except for trans-nerolidol (2.45-fold and 3.09-fold), all others increased more than 5.37-fold. The highest increase in alpha-ionone was found in KJGT-10 (9.80-fold) and 2-ethyl-3,5-dimethyl pyrazine was found in KJGT-20 (10.69-fold). Trans-beta-ocimene, (E)-4-nonenal were significantly down-regulated and highly productive, which reduced KJGT-10 and KJGT-20 by 5.46 and 9.67-fold, respectively. 2-Ethyl-3,5-dimethyl-pyrazine plays a major role in the nutty aroma of sunflower seeds (Gracka et al., 2016). trans-Nerolidol was a naturally occurring sesquiterpene enol found in a variety of plants with floral fragrances (Chan et al., 2016). The high content of free monoterpenes such as trans-beta-ocimene in Muscat-type grapes was responsible for the high intensity of floral, rosy and sweet flavors (Xie et al., 2023).

In addition, differential metabolites between KJGT-20 and KJGT-10 were compared and a total of 200 volatile metabolites were screened (P < 0.05), of which 33 metabolites were significantly down-regulated, 167 were significantly up-regulated, and 109 substances had aroma properties (Fig. 5C). The top 10 sensory flavors with the most notes were green (30), fruity (29), sweet (24), herbal (16), woody (15), waxy (10), citrus (9), floral (9), fresh (9) and apple (8). These differential metabolites were all up-regulated in greater numbers than down-regulated, indicating that these 10 aroma attributes were significantly higher in KJGT-20 than in KJGT-10 (Fig. S4C). There were 45 differential metabolites identified with odor activity values, of which 37 were up-regulated and 8 down-regulated, and there were eight differential metabolites with rOAV ≥1 in both KJGT-10 and KJGT-20. Characteristic metabolites that were significantly up-regulated in KJGT-20 compared to KJGT-10 included phenylethyl alcohol, 1,3,8-p-menthatriene, 2-methylnaphthalene, 5-butyldihydro-2(3H)-Furanone, (Z)-6-nonanal, and 4-[2,2,6-trimethyl-7-oxabicyclo[4.1. 0]hept-1-yl]-3-buten-2-one, all of which had an increase of more than 2.29-fold, with (Z)-6-nonanal having the highest multiplicity of increase (6.32-fold). The 1-p-menthane-8-thiol and δ-cumulene were significantly down-regulated in KJGT-20, with 7.15-fold and 2.46-fold reductions, respectively. (Z)-6-Nominal was one of the cucumber aroma aldehydes with a refreshing cucumber aroma and flavor (Cho & Buescher, 2011). 1,3,8-p-Menthatriene was the main volatile substance of the aromatic plant parsley (source of essential or volatile oils) with essential oil, herbal and woody aromas (El-Zaeddi et al., 2016). δ-Juniperene was helium volatile from the fresh branches of camphor trees, which could be used as an industrial raw material for biomedicine and fragrances, with herbal and woody aroma (Tian et al., 2021).

Furthermore, a thorough analysis was conducted on the volatile metabolites associated with the distinctive ‘high-fire aroma’ characteristic of KJGT. Among these, the volatile metabolites 3-trimethylpyrazine, 2-acetyl-3,4,5,6-tetrahydropyridine, 2-ethoxy-3-methylpyrazine, and 2-ethylbenzaldehyde-volatile compounds imparting roasted and burnt odor notes-were significantly reduced in KJGT-10 and KJGT-20. Specifically, their concentrations decreased by 4.89, 5.64, 5.92, and 9.52-fold in KJGT-10, and by 3.27, 5.64, 5.92, and 4.58 times respectively in KJGT-20. This indicates that post-storage Kejia green tea effectively reduces the roasted-like aromas formed by the ‘heavy stir-frying’ process, allowing pleasant fruit and floral notes to become more prominent. During long-term storage, pyrazine-based aroma compounds formed through the Maillard reaction and Strecker degradation at high temperatures undergo slow chemical rearrangement or degradation (Karolkowski et al., 2021). Many pyrazine and short-chain aldehyde compounds possess high vapor pressures, leading to continuous volatilization losses from food matrices even under low-temperature storage conditions (Cui et al., 2024).

Research findings indicate that multiple key volatile metabolites govern the aroma transformation of KJGT during storage. Following storage, KJGT-0 exhibits significantly reduced levels of green, roasted, and burnt odors, thereby diminishing unpleasant aromas present in fresh tea. Sweet and fruity odors also diminish moderately during storage, though the overall profile retains these characteristics. Nutty and woody notes become more pronounced. This indicates that post-storage KJGT effectively mitigates undesirable notes such as greenness, roasted, and burnt odors in fresh tea, while accentuating desirable aromas. This enhances consumer appeal across different demographics and elevates the quality value of Kejia green tea.

3.3.3. KEGG functional annotation and volatile differential metabolite enrichment

These volatile differential metabolites were similarly analyzed for pathway enrichment using the KEGG database. 14 pathways were enriched in the comparison of KJGT-10 with KJGT-0, with 15 different metabolites with aroma properties enriched in these pathways (Fig. 6A). A total of 15 pathways were enriched when comparing KJGT-20 with KJGT-0, with 12 different metabolites with aroma properties being enriched in these pathways (Fig. 6B). There were 13 metabolic pathways enriched in the comparison of KJGT-20 with KJGT-10, with 9 differential metabolites with aroma properties enriched in these pathways (Fig. 6C). Pathway enrichment analysis based on differential metabolite results (Fig. 6D-F). Tea samples from three different storage years were collectively enriched for nine pathways, including sesquiterpene and triterpene biosynthesis, metabolic pathways, alpha-linolenic acid metabolism, biosynthesis of secondary metabolites, tyrosine metabolism, phenylalanine metabolism, tryptophan metabolism, tyrosine and tryptophan biosynthesis, phenylalanine and biosynthesis pathways for a variety of plant secondary metabolites, with notable enriched pathways being biosynthesis of secondary metabolites, sesquiterpene and triterpene biosynthesis. Terpenoids were a large class of plant-derived compounds, that are widely used as natural flavors and fragrances (Jiang & Wang, 2023). Many of the aroma compounds in tea were produced or broken down by secondary metabolic pathways (Zhang et al., 2023). The diversity and complexity of secondary metabolites in the tea plant contributed to the flavor and promoted the health benefits of tea (Zhao et al., 2020).

Fig. 6.

Fig. 6

Open in a new tab

KEGG functional annotation and enrichment analysis of volatile differential metabolites. (A-C) Classification of differential metabolite pathways. (D—F) KEGG enrichment pathway analysis.

4. Conclusions

Kejia green tea possesses high storage value, which is largely attributable to changes in its metabolic substances during storage. This study revealed the differences in the composition of volatile and non-volatile metabolites in Kejia green teas of three different storage years, and screened out the flavor components and key aroma substances that formed their unique quality differences. The results of the sensory review and e-tongue showed that the bitter and astringent flavor of Kejia green tea was reduced and the taste became mellow after storage, and KJGT-10 had the highest overall score. The contents of bitter substances and astringent substances were significantly reduced in the stored tea by altering linoleic acid metabolism and the biosynthetic pathway of quercetin glycoside I. Moreover, the aroma conversion of KJGT was driven by the content of volatile metabolites such as beta-rollerene, (E)-4-nonenal, (E, E)-2,4-undecadienal, methyl benzoate, 2-ethyl-3,5-dimethyl pyrazine, alpha-ionone, and trans-nerolidol. The storage process markedly reduced the green, roasted and burnt aromas, diminishing the high-fire aroma in fresh tea, while accentuating desirable fragrances such as sweetness, fruitiness, floral aromas and woody notes. KJGT-10 had the best flavor (lowest bitter and astringent) and aroma (prominent woody aroma). This study established a comprehensive framework for the KJGT tea industry by providing foundational theories for scientific storage and quality regulation, evidence-based guidelines for consumer decision-making, and practical technical solutions for inventory management. Future studies should employ targeted absolute quantitative methods to detect key metabolites, increase the number of sample storage time points, and conduct comprehensive nutritional assessments of tea samples to enhance the persuasiveness of results.

Author contribution

All authors' contributions as follow. Shili Sun and Qiuhua Li conceived and designed the experiments, and revised the manuscript. Ping Wu and He Ni performed the experiments, analyzed the data and wrote the paper. Suwan Zhang, Ruohong Chen, Mengjiao Hao, Lingli Sun, Xingfei Lai, Zhenbiao Zhang contributed reagents and funding, and investigation. All authors approved the final version of the manuscript.

CRediT authorship contribution statement

Ping Wu: Writing – original draft, Visualization, Methodology, Formal analysis, Data curation, Conceptualization. He Ni: Writing – original draft, Visualization, Funding acquisition, Conceptualization. Suwan Zhang: Validation, Investigation, Formal analysis. Ruohong Chen: Supervision, Resources, Funding acquisition. Mengjiao Hao: Resources, Project administration, Methodology. Lingli Sun: Resources, Project administration, Funding acquisition. Xingfei Lai: Resources, Investigation, Funding acquisition. Zhenbiao Zhang: Validation, Software, Resources. Shili Sun: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition, Formal analysis, Conceptualization. Qiuhua Li: Writing – review & editing, Supervision, Resources, Project administration, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This research was supported by Guangdong Foundation for Basic and Applied Research (Project Nos: 2022A1515110190, 2023A1515010264, 2024A1515010696), Guangdong Provincial Science and Technology Achievements Reach Townships Special Project (Project No.: 2025B0202010021), Guangdong Academy of Agricultural Sciences Special Fund Project to Introduce Science and Technology Talent (Project No.: R2024YJ-YB3005, R2021YJ-YB3014), the Qingyuan Science and Technology Plan Projects (grant number 2021ZDZX002), Youth S&T Talent Support Programme of Guangdong Provincial Association for Science and Technology (Project No.: SKXRC2025483), Special Funding for the “Jinying Talents” Project of Guangdong Academy of Agricultural Sciences (R2026PY-TJ027), Funds for the construction and operation of the Food Nutrition and Health Research Center of Guangdong Academy of Agricultural Sciences (XT202506), Special Funding for the Construction of the High-Level Academy of Agricultural Sciences (NYQS202612). The authors thanked Meizhou Junbao Industrial Co., Ltd. for providing tea samples for this study free of charge. Funders did not have any role in study design, data collection, and data analysis.

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.fochx.2026.103867.

Contributor Information

Ping Wu, Email: wuping202507@163.com.

He Ni, Email: 20131032@m.scnu.edu.cn.

Suwan Zhang, Email: swzhang1502@163.com.

Ruohong Chen, Email: chenruohong@tea.gdaas.cn.

Mengjiao Hao, Email: haomengjiao@tea.gdaas.cn.

Lingli Sun, Email: sunlingli@tea.gdaas.cn.

Xingfei Lai, Email: laixingfei@tea.gdaas.cn.

Zhenbiao Zhang, Email: zhangzhenbiao@tea.gdaas.cn.

Shili Sun, Email: sunshili@zju.edu.cn.

Qiuhua Li, Email: liqiuhua@tea.gdaas.cn.

Appendix A. Supplementary data

Supplementary material 1

mmc1.xlsx (295.9KB, xlsx)

Supplementary material 2

mmc2.pptx (2.8MB, pptx)

Data availability

No data was used for the research described in the article.

References

  1. Borrull J., Rambla-Alegre M., Diogène J., Alcaraz C. Evaluation of camellia oleifera seed saponins degradation in rice fields by UPLC-MS/MS method. Science of the Total Environment. 2025;958 doi: 10.1016/j.scitotenv.2024.177937. [DOI] [PubMed] [Google Scholar]
  2. Chan W.K., Tan L.T., Chan K.G., Lee L.H., Goh B.H. Nerolidol: A sesquiterpene alcohol with multi-faceted pharmacological and biological activities. Molecules. 2016;21(5) doi: 10.3390/molecules21050529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Chen Y., Gan S., Zhao X., Zhao L., Qiu T., Jiang Q., Dai Q. Rapid discrimination of xiaokeng green tea geographical origin using surface-enhanced raman spectroscopy coupled with chemometrics. Lwt. 2024;197 doi: 10.1016/j.lwt.2024.115867. [DOI] [Google Scholar]
  4. Cho M.J., Buescher R. Degradation of cucumber flavor aldehydes in juice. Food Research International. 2011;44(9):2975–2977. doi: 10.1016/j.foodres.2011.07.027. [DOI] [Google Scholar]
  5. Cui J., Wu B., Wang J., Jing T., Jin J., Zhao M., Song C. Effect of storage time on aroma profiles of wuyi rock tea. Lwt-food Science and Technology. 2024;203 doi: 10.1016/j.lwt.2024.116367. [DOI] [Google Scholar]
  6. Deng S., Zhang G., Olayemi Aluko O., Mo Z., Mao J., Zhang H., Liu H. Bitter and astringent substances in green tea: Composition, human perception mechanisms, evaluation methods and factors influencing their formation. Food Research International. 2022;157 doi: 10.1016/j.foodres.2022.111262. [DOI] [PubMed] [Google Scholar]
  7. El-Zaeddi H., Calín-Sánchez Á., Martínez-Tomé J., Noguera-Artiaga L., Burló F., Carbonell-Barrachina Á.A. Irrigation dose and plant density affect the essential oil content and sensory quality of parsley (petroselinum sativum) Scientia Horticulturae. 2016;206:1–6. doi: 10.1016/j.scienta.2016.04.028. [DOI] [Google Scholar]
  8. Fan F.Y., Huang C.S., Tong Y.L., Guo H.W., Zhou S.J., Ye J.H., Gong S.Y. Widely targeted metabolomics analysis of white peony teas with different storage time and association with sensory attributes. Food Chemistry. 2021;362 doi: 10.1016/j.foodchem.2021.130257. [DOI] [PubMed] [Google Scholar]
  9. Gracka A., Jeleń H.H., Majcher M., Siger A., Kaczmarek A. Flavoromics approach in monitoring changes in volatile compounds of virgin rapeseed oil caused by seed roasting. Journal of Chromatography A. 2016;1428:292–304. doi: 10.1016/j.chroma.2015.10.088. [DOI] [PubMed] [Google Scholar]
  10. Hao M., Lai X., Li Q., Cao J., Sun L., Chen R., Sun S. Widely targeted metabolomic analysis reveals metabolite changes induced by black tea fermentation techniques in oolong tea processing for quality improvement. Food Chemistry. 2024;459 doi: 10.1016/j.foodchem.2024.140433. [DOI] [PubMed] [Google Scholar]
  11. Ho C., Zheng X., Li S. Tea aroma formation. Food Science and Human Wellness. 2015;4(1):9–27. doi: 10.1016/j.fshw.2015.04.001. [DOI] [Google Scholar]
  12. Huang W., Fang S., Wang J., Zhuo C., Luo Y., Yu Y., Ning J. Sensomics analysis of the effect of the withering method on the aroma components of keemun black tea. Food Chemistry. 2022;395 doi: 10.1016/j.foodchem.2022.133549. [DOI] [PubMed] [Google Scholar]
  13. Huang W., Lu G., Deng W., Ning J. Effects of different withering methods on the taste of keemun black tea. LWT. 2022;166 doi: 10.1016/j.lwt.2022.113791. [DOI] [Google Scholar]
  14. Jiang H., Wang X. Biosynthesis of monoterpenoid and sesquiterpenoid as natural flavors and fragrances. Biotechnology Advances. 2023;65 doi: 10.1016/j.biotechadv.2023.108151. [DOI] [PubMed] [Google Scholar]
  15. Karolkowski A., Guichard E., Briand L., Salles C. Volatile compounds in pulses: A review. Foods. 2021;10(12) doi: 10.3390/foods10123140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Liu H., Xu Y., Wu J., Wen J., Yu Y., An K., Zou B. GC-IMS and olfactometry analysis on the tea aroma of Yingde black teas harvested in different seasons. Food Research International (Ottawa, Ont.) 2021;150(Pt A) doi: 10.1016/j.foodres.2021.110784. [DOI] [PubMed] [Google Scholar]
  17. Liu H., Zhuang S., Gu Y., Shen Y., Zhang W., Ma L., Zhong Y. Effect of storage time on the volatile compounds and taste quality of meixian green tea. LWT-Food Science and Technology. 2023;173 doi: 10.1016/j.lwt.2022.114320. [DOI] [Google Scholar]
  18. Liu S., Zhang S., He S., Qiao X., Runa A. Tea plant (camellia sinensis) lipid metabolism pathway modulated by tea field microbe (colletotrichum camelliae) to promote disease. Horticulture Research. 2023;10(4) doi: 10.1093/hr/uhad028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Liu Y., Wang Z., Zhang X., Chen H., Ruan D., Bai S., Liu Z. Characterization of volatile substances in pu-erh tea (raw tea) at different storage times. Foods. 2025;14(5) doi: 10.3390/foods14050840. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Long P., Rakariyatham K., Ho C., Zhang L. Thearubigins: Formation, structure, health benefit and sensory property. Trends in Food Science & Technology. 2023;133:37–48. doi: 10.1016/j.tifs.2023.01.013. [DOI] [Google Scholar]
  21. Ma B., Ma C., Zhou B., Chen X., Wang Y., Li Y., Li X. Quantitative descriptive analysis, non-targeted metabolomics and molecular docking reveal the dynamic aging and taste formation mechanism in raw pu-erh tea during the storage. Food Chemistry: X. 2025;25 doi: 10.1016/j.fochx.2025.102234. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Meng T., Zhipeng X., Ai H., Jiayu C., Tianji Y., Zhengquan L. Effect of 1–20 years storage on volatiles and aroma of keemun congou black tea by solvent extraction-solid phase extraction-gas chromatography-mass spectrometry. LWT. 2021;136 doi: 10.1016/j.lwt.2020.110278. [DOI] [Google Scholar]
  23. Ocieczek A., Pukszta T., Żyłka K., Kirieieva N. The influence of storage conditions on the stability of selected health-promoting properties of tea. LWT. 2023;184 doi: 10.1016/j.lwt.2023.115029. [DOI] [Google Scholar]
  24. Patra B., Schluttenhofer C., Wu Y., Pattanaik S., Yuan L. Transcriptional regulation of secondary metabolite biosynthesis in plants. Biochimica et Biophysica Acta. 2013;1829(11):1236–1247. doi: 10.1016/j.bbagrm.2013.09.006. [DOI] [PubMed] [Google Scholar]
  25. Peng J., Dai W., Lu M., Yan Y., Zhang Y., Chen D., Lin Z. New insights into the influences of baking and storage on the nonvolatile compounds in oolong tea: A nontargeted and targeted metabolomics study. Food Chemistry. 2022;375 doi: 10.1016/j.foodchem.2021.131872. [DOI] [PubMed] [Google Scholar]
  26. Pérez-Burillo S., Jiménez-Zamora A., Párraga J., Rufián-Henares J.A., Pastoriza S. Furosine and 5-hydroxymethylfurfural as chemical markers of tea processing and storage. Food Control. 2019;99:73–78. doi: 10.1016/j.foodcont.2018.12.029. [DOI] [Google Scholar]
  27. Qiu Z., Liao J., Chen J., Chen P., Sun B., Li A., Liu S. The cultivar effect on the taste and aroma substances of Hakka stir-fried green tea from Guangdong. Foods. 2023;12(10) doi: 10.3390/foods12102067. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Ren Y., Hou Y., Granato D., Zha M., Xu W., Zhang L. Metabolomics, sensory evaluation, and enzymatic hydrolysis reveal the effect of storage on the critical astringency-active components of crude pu-erh tea. Journal of Food Composition and Analysis. 2022;107 doi: 10.1016/j.jfca.2022.104387. [DOI] [Google Scholar]
  29. Ren Y., Hou Y., Granato D., Zha M., Xu W., Zhang L. Metabolomics, sensory evaluation, and enzymatic hydrolysis reveal the effect of storage on the critical astringency-active components of crude pu-erh tea. Journal of Food Composition and Analysis. 2022;107 doi: 10.1016/j.jfca.2022.104387. [DOI] [Google Scholar]
  30. Sarma A., Bania R., Das M.K. Green tea: Current trends and prospects in nutraceutical and pharmaceutical aspects. Journal of Herbal Medicine. 2023;41 doi: 10.1016/j.hermed.2023.100694. [DOI] [Google Scholar]
  31. Shan X., Deng Y., Niu L., Chen L., Zhang S., Jiang Y., Li J. The influence of fixation temperature on Longjing tea taste profile and the underlying non-volatile metabolites changes unraveled by combined analyses of metabolomics and e-tongue. LWT. 2024;191 doi: 10.1016/j.lwt.2023.115560. [DOI] [Google Scholar]
  32. Sun L., Zhang S., Li Q., Yuan E., Chen R., Yan F., Sun S. Metabolomics and electronic tongue reveal the effects of different storage years on metabolites and taste quality of oolong tea. Food Control. 2023;152 doi: 10.1016/j.foodcont.2023.109847. [DOI] [Google Scholar]
  33. Sun L., Zhang S., Li Q., Yuan E., Chen R., Yan F., Sun S. Metabolomics and electronic tongue reveal the effects of different storage years on metabolites and taste quality of oolong tea. Food Control. 2023;152 doi: 10.1016/j.foodcont.2023.109847. [DOI] [Google Scholar]
  34. Tan K.L., Zulkurnain M., Chua Y.S., Suah F.B.M., Leong Y., Wong Y.F. Occurrence, exposure estimation, and dietary risk assessment of biogenic amines in tea samples marketed in Malaysia. Food Control. 2024;164 doi: 10.1016/j.foodcont.2024.110593. [DOI] [Google Scholar]
  35. Tian Z., Luo Q., Zuo Z. Seasonal emission of monoterpenes from four chemotypes of cinnamomum camphora. Industrial Crops and Products. 2021;163 doi: 10.1016/j.indcrop.2021.113327. [DOI] [Google Scholar]
  36. Wang K., Ruan J. Analysis of chemical components in green tea in relation with perceived quality, a case study with Longjing teas. International Journal of Food Science and Technology. 2009;44(12):2476–2484. doi: 10.1111/j.1365-2621.2009.02040.x. [DOI] [Google Scholar]
  37. Wang L., Pan D., Liang M., Abubakar Y.S., Li J., Lin J., Chen W. Regulation of anthocyanin biosynthesis in purple leaves of zijuan tea (camellia sinensis var. kitamura) International Journal of Molecular Sciences. 2017;18(4) doi: 10.3390/ijms18040833. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Wang X., Xiong H., Wang S., Zhang Y., Song Z., Zhang X. Physicochemical analysis, sensorial evaluation, astringent component identification and aroma-active compounds of herbaceous peony (paeonia lactiflora Pall) black tea. Industrial Crops and Products. 2023;193 doi: 10.1016/j.indcrop.2022.116159. [DOI] [Google Scholar]
  39. Wang Y., Li C., Lin J., Sun Y., Wei S., Wu L. The impact of different withering approaches on the metabolism of flavor compounds in oolong tea leaves. Foods. 2022;11(22) doi: 10.3390/foods11223601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Wei F., Wang J., Luo L., Tayyab R.M., Zeng L. The perception and influencing factors of astringency, and health-promoting effects associated with phytochemicals: A comprehensive review. Food Research International. 2023;170 doi: 10.1016/j.foodres.2023.112994. [DOI] [PubMed] [Google Scholar]
  41. Wen S., Sun L., Zhang S., Chen Z., Chen R., Li Z., Li Q. The formation mechanism of aroma quality of green and yellow teas based on GC-MS/MS metabolomics. Food Research International (Ottawa, Ont.) 2023;172 doi: 10.1016/j.foodres.2023.113137. [DOI] [PubMed] [Google Scholar]
  42. Wen X., Han S., Wang J., Zhang Y., Tan L., Chen C., Wang M. The flavor characteristics, antioxidant capability, and storage year discrimination based on backpropagation neural network of organic green tea (camellia sinensis) during long-term storage. Foods. 2024;13(5) doi: 10.3390/foods13050753. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Xie S., Wu G., Ren R., Xie R., Yin H., Chen H., Ge M. Transcriptomic and metabolic analyses reveal differences in monoterpene profiles and the underlying molecular mechanisms in six grape varieties with different flavors. LWT. 2023;174 doi: 10.1016/j.lwt.2023.114442. [DOI] [Google Scholar]
  44. Xu J., Deng X., Wu Y., Zhou M., Du C., Wang Q., Wang B. Characteristic changes and potential markers of flavour in raw pu-erh tea with different ageing cycles analysed by HPLC, HS-SPME-GC-MS, and OAV. Foods. 2025;14(5) doi: 10.3390/foods14050829. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Xu S., Wang J.J., Wei Y., Deng W.W., Wan X., Bao G.H., Ning J. Metabolomics based on UHPLC-orbitrap-MS and global natural product social molecular networking reveals effects of time scale and environment of storage on the metabolites and taste quality of raw pu-erh tea. Journal of Agricultural and Food Chemistry. 2019;67(43):12084–12093. doi: 10.1021/acs.jafc.9b05314. [DOI] [PubMed] [Google Scholar]
  46. Xue J., Liu P., Yin J., Wang W., Zhang J., Wang W., Jiang H. Dynamic changes in volatile compounds of shaken black tea during its manufacture by GC × GC-TOFMS and multivariate data analysis. Foods. 2022;11(9) doi: 10.3390/foods11091228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Yang Y., Ai L., Mu Z., Liu H., Yan X., Ni L., Xia Y. Flavor compounds with high odor activity values (OAV > 1) dominate the aroma of aged chinese rice wine (huangjiu) by molecular association. Food Chemistry. 2022;383 doi: 10.1016/j.foodchem.2022.132370. [DOI] [PubMed] [Google Scholar]
  48. Zhang J., Xin W., Zou Y., Yan J., Tang W., Ji Y., Li W. Dynamic changes and correlation analysis of microorganisms and flavonoids/ amino acids during white tea storage. Food Chemistry. 2024;455 doi: 10.1016/j.foodchem.2024.139932. [DOI] [PubMed] [Google Scholar]
  49. Zhang S., Sun L., Wen S., Chen R., Sun S., Lai X., Cao J. Analysis of aroma quality changes of large-leaf black tea in different storage years based on HS-SPME and GC-MS. Food Chemistry-X. 2023;20 doi: 10.1016/j.fochx.2023.100991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Zhao J., Li P., Xia T., Wan X. Exploring plant metabolic genomics: Chemical diversity, metabolic complexity in the biosynthesis and transport of specialized metabolites with the tea plant as a model. Critical Reviews in Biotechnology. 2020;40(5):667–688. doi: 10.1080/07388551.2020.1752617. [DOI] [PubMed] [Google Scholar]
  51. Zhou B., Ma C., Xia T., Li X., Zheng C., Wu T., Liu X. Isolation, characterization and application of theophylline-degrading Aspergillus fungi. Microbial Cell Factories. 2020;19(1) doi: 10.1186/s12934-020-01333-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Zhou D.D., Mao Q.Q., Li B.Y., Saimaiti A., Huang S.Y., Xiong R.G., Li S. Effects of different green teas on obesity and non-alcoholic fatty liver disease induced by a high-fat diet in mice. Frontiers in Nutrition. 2022;9 doi: 10.3389/fnut.2022.929210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Zhu J., Niu Y., Xiao Z. Characterization of the key aroma compounds in Laoshan green teas by application of odour activity value (OAV), gas chromatography-mass spectrometry-olfactometry (GC-MS-o) and comprehensive two-dimensional gas chromatography mass spectrometry (GC×GC-qMS) Food Chemistry. 2021;339 doi: 10.1016/j.foodchem.2020.128136. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material 1

mmc1.xlsx (295.9KB, xlsx)

Supplementary material 2

mmc2.pptx (2.8MB, pptx)

Data Availability Statement

No data was used for the research described in the article.

Articles from Food Chemistry: X are provided here courtesy of Elsevier

RESOURCES