Skip to main content
NCBI home page
Food Chemistry: X logoLink to Food Chemistry: X
. 2025 Nov 29;32:103345. doi: 10.1016/j.fochx.2025.103345

The impact of the unique two-stage bunch-drying on the formation of the floral aroma of Niangniang tea

Xinyi Sun 1, Ceyu Wang 1, Xingyu Long 1, Shumei Han 1, Wei Xu 1, Taolin Chen 1, Jianjun Liu 1,⁎, Beibei Wen 1,⁎, Meifeng Li 1,⁎
PMCID: PMC12721074  PMID: 41438550

Abstract

Niangniang tea is typically processed using a two-stage bunch-drying process, which gives it distinct floral aroma characteristics. This study investigated the impact of the two-stage bunch-drying process on the formation of Niangniang tea's floral aroma and compared it with green tea and yellow tea. Sensory evaluation, metabolomics, aroma recombination and omission tests, and molecular docking results revealed that Niangniang tea's key aroma compounds were linalool, hotrienol, p-anisaldehyde, 5-nonanone, S-methyl thiohexanoate, 1-nonen-3-ol, and 2-isopropyl-3-methoxypyrazine. p-Anisaldehyde, linalool, and hotrienol were responsible for Niangniang tea's intense floral aroma, and p-anisaldehyde was the marker of the floral aroma of NNT. This study identified the key floral compounds in Niangniang tea and revealed how the two-stage bunch-drying process influenced aroma formation. These results provide a theoretical basis for optimizing the bunch-drying process parameters in future studies.

Keywords: Niangniang tea, Two-stage bunch-drying, Floral aroma, Key compounds, P-Anisaldehyde

Graphical abstract

Unlabelled Image

Open in a new tab

Highlights

  • •

    The two-stage bunch-drying is characteristic of Niangniang tea (NNT) processing.

  • •

    NNT has a distinctively stronger floral aroma compared to green tea and yellow tea.

  • •

    Hydrothermal condition is necessary for the conversion of the key compounds of NNT.

  • •

    p-Anisaldehyde, linalool, and hotrienol are key to the floral aroma of NNT.

  • •

    p-Anisaldehyde serves as a distinguishing marker for the floral aroma profile of NNT.

1. Introduction

The quality of tea significantly influences its market value and consumer acceptance. During the sensory evaluation of tea quality, taste and aroma are the most important factors, with taste accounting for 30–35 % of the evaluation system and aroma accounting for 25–30 % (Lu, Wang, et al., 2024). Although volatile compounds of tea account for only about 0.01 % of tea's dry weight, they contribute 25–40 % to its quality. These compounds form a unique aroma fingerprint that influences the flavor characteristics of tea and also play a role in tea grading, origin identification, and consumer preferences (Guo et al., 2021).

The formation of tea aroma compounds is closely related to the way tea is processed. During withering, tea leaves undergo stress responses due to dehydration, which reduces the content of low-boiling-point volatile compounds, such as grassy odors. It also promotes the synthesis and accumulation of high-boiling-point volatile compounds, which are characterized by fresh, sweet, floral, fruity, and woody aromas (Qi et al., 2024). Wang et al. (2020) investigated how different fixation conditions affected key volatile compounds in green tea and found that rolling-hot air treatment produced strong, persistent chestnut and floral aromas. Yang, Wang, et al. (2024) found that sweet and floral aromas in black tea arose from glycoside hydrolysis, which released substances such as linalool and β-ocimene during the shaking and rolling stages, and carotenoid degradation during fermentation, which formed substances such as β-ionone and β-cyclocitral. Tu et al. (2025) discovered that processing green tea using shaking and piling promoted the formation and accumulation of floral compounds, such as linalool and β-ionone, which enhanced green tea's floral, fruity, and sweet aromas. The Maillard reaction occurs during tea drying and produces a series of heterocyclic compounds, including furans, pyrroles, and thiophenes, which exhibit roasted, nutty, and caramel-like aromas (Ho et al., 2015). Processing techniques such as withering, shaking, fermentation, and drying directly or indirectly influence the types and contents of volatile compounds in tea leaves by regulating pathways such as glycoside hydrolysis, enzymatic reactions, lipid oxidation, and carotenoid degradation (Cui et al., 2024). Thus, these techniques shape the unique aromatic qualities of tea.

The aroma of tea is influenced by a combination of different volatile compounds, their concentration, and their effect on olfactory nerves. The human nose has approximately 400 olfactory receptors that can identify and distinguish at least 10,000 different scents. One receptor can identify multiple odor molecules, and conversely, one odor molecule can be identified by multiple receptors (Oka et al., 2004). Olfactory receptors are the starting point for the formation of olfactory effects during the identification and analysis of odor components. OR1A1, OR1D2, OR1G1, OR2W1, and OR52D1 are broad-tuning receptor proteins in humans that can detect a wide range of odor molecules. They primarily convert odor molecules into neural signals that are transmitted to the brain for odor identification. This broad recognition capability enables humans to perceive and distinguish various complex odors (Xiao et al., 2024). Receptors can interact with floral and sweet volatile compounds, such as geraniol, p-anisaldehyde, linalool, 2-phenylethanol, and ethyl cinnamate, through various intermolecular interactions (Sun et al., 2024; Wang et al., 2025). This provides an ideal model for understanding interaction mechanisms between olfactory receptors and volatile compounds. Molecular docking is a computational simulation method that can be used to predict interactions between molecules and target receptors such as proteins. It can also be used for the in-depth analysis of the binding characteristics between olfactory receptors and volatile compounds (Paggi et al., 2024).

Niangniang tea (NNT) is processed from the landrace tea plant populations in Zhenfeng County, Guizhou Province via spreading, fixation, rolling, bundling, and two bunch-drying steps. NNT's bunch-drying process is different from both the yellowing process used for yellow tea (YT) and the drying process used for green tea (GT) (Wei, Liu, et al., 2025). Our team previously conducted physiology and anatomical structure research, as well as volatile metabolomics studies on the fresh leaves of the local Zhenfeng variety. The results showed that it was a high-aroma type variety rich in terpenoids (Li et al., 2022; Wei, Mu, et al., 2025). During the processing of NNT, we discovered that its unique bunch-drying technique resulted in finished tea with intense floral and sweet aromas. This study determined whether the unique floral and sweet aromas of NNT originated from the tea plant variety or the special processing method. We used the same fresh leaves as the raw material and used the same spreading, fixation, and rolling conditions to compare the aroma of finished teas of GT (high-temperature quick drying), YT (yellowing to dry), and NNT (two-stage bunch-drying). Sensory quantitative description analysis, headspace solid-phase microextraction gas chromatography–mass spectrometry (HS-SPME-GC–MS), multivariate statistical analysis, aroma recombination and omission tests, and molecular docking methods were used to identify the key compounds responsible for the floral aroma of NNT and clarify the floral aroma formation mechanism.

2. Materials and methods

2.1. Materials and reagents

All tea samples were harvested and processed in July 2022 in Po Liu Village, Longchang Town, Zhenfeng County, Guizhou Province, China, using the method described by Wei et al. (Wei, Liu, et al., 2025). The processing procedure is shown in Fig. 1. Three biological replicates were used for each sample and stored at −80 °C until analysis.

Fig. 1.

Fig. 1

Open in a new tab

Processing flowchart for three types of tea samples.

Sodium chloride (analytical grade) was purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Hexane (chromatographic grade) was purchased from Merck (Darmstadt, Germany). All chemical standards used for gas chromatography–mass spectrometry (GC–MS) were of chromatographic grade and were purchased from Sigma-Aldrich (St. Louis, Missouri, USA) or BioBioPha (Yunnan, China). The aroma standards used for quantitative descriptive analysis (QDA) and aroma reconstruction and omission tests, including delta-hexalactone, methyl benzoate, trans-2-undecenal, 1-nonen-3-ol, 2-isopropyl-3-methoxypyrazine, cis-3-hexen-1-ol, hexanal, and 3-methylbutyraldehyde, were purchased from McLean (Shanghai, China). Furaneol, cedrol, 5-nonanone, linalool, S-methyl thiohexanoate, and p-anisaldehyde were purchased from Aladdin (Shanghai, China). p-Cymene was purchased from Shanghai Yuanye Biotechnology Co., Ltd. (Shanghai, China), and hotrienol was purchased from Wuhan Lanabai Pharmaceutical & Chemical Co., Ltd. (Hubei, China).

2.2. Quantitative descriptive analysis of three types of tea samples

This study employed the method of Ni et al. (Ni et al., 2021) to perform quantitative descriptive analysis (QDA) on the three types of tea samples. The evaluation panel consisted of nine members (two males and seven females) between the ages of 23 and 40 from the Tea College of Guizhou University in Guizhou, China. According to GB/T 23776–2018 (“Sensory Evaluation Methods for Tea,” 2018), sensory evaluation was conducted in a clean environment at 25 ± 2 °C.

Before conducting the formal QDA, the nine group members evaluated the tea infusion individually and independently recorded descriptive terms reflecting the aromatic characteristics of the tea. The evaluation group leader organized and categorized all terms and then held a discussion with all evaluation group members to confirm and validate the descriptive terms. Five descriptive terms that best represented the aromatic attributes of the three types of tea were adopted as indicators for sensory evaluation. Samples were randomly assigned three-digit codes for scoring using a 5-point scale (0–5), where 0 indicates imperceptible, 1 indicates weak, 2 indicates moderately weak, 3 indicates moderate, 4 indicates moderately strong, and 5 indicates strong.

2.3. Detection of volatile compounds in three types of tea samples

Using the methods of Zhang et al. (Zhang et al., 2025), three types of tea samples (NNT, GT, and YT) were collected using different processing methods, and three biological replicates were established for each sample. First, three tea samples were extracted using solid-phase microextraction (SPME Arrow, CTC Analytics AG, Zwingen, Switzerland). Then, volatile compounds in three types of tea samples were identified using an Agilent Model 8890 GC and a 7000D mass spectrometer (Agilent, Santa Clara, CA, USA). During mass spectrometry, an electron bombardment ion source was used with an ion source temperature of 230 °C, a quadrupole temperature of 150 °C, and a mass spectrometry interface temperature of 280 °C. The electron energy was set to 70 eV, and scanning was performed in selected ion detection (SIM) mode. Downstream raw data were processed using Mass Hunter software from Agilent Qualitative Systems. Metabolites were qualitatively and quantitatively analyzed by mass spectrometry using a self-constructed Metware GC database. Qualitative and quantitative analysis of the raw data obtained via GC–MS was performed using MassHunter software.

2.4. Quantification and calculation of relative odor activity values

This study used relative odor activity values (rOAV) to analyze the contribution of each volatile compound to the overall aroma characteristics of three tea samples, as shown in Eq. (1):

rOAVi=CiTi (1)

where Ci (μg/g) indicates the relative content of compound i, and Ti (μg/g) indicates the threshold of compound i.

2.5. Aroma recombination and omission tests

Aroma recombination and omission tests were conducted according to Wei et al. (Wei et al., 2024). Three recombinant models, including a recombinant model of NNT (Re-NNT), a recombinant model of GT (Re-GT), and a recombinant model of YT (Re-YT), were prepared by mixing 12 key volatile compounds with rOAV ≥ 10 in water according to their concentrations in three tea samples. The recombinant model and the original tea infusion samples were labeled and then placed in 50 mL brown triangular bottles with caps in a 40 °C water bath. Using the method described in Section 2.2, we conducted an evaluation panel QDA on the recombination models and original tea infusion samples.

We conducted omission tests after the recombination experiments by first preparing different omission models by sequentially omitting one or a class of volatile compounds from the NNT recombination model. The evaluation panel was then used to assess the differences between each omission model and the recombination model using a triangle test. If six panel members answered correctly, the result was considered significant (0.01 < p ≤ 0.05). If seven to eight panel members answered correctly, the result was considered highly significant (0.001 < p ≤ 0.01). If nine panel members answered correctly, the result was considered extremely significant (p ≤ 0.001).

2.6. Binding interactions between volatile compounds and olfactory receptors with molecular docking

The SDF-format ligand files for the seven key volatile compounds in NNT were downloaded from PubChem (https://pubchem.ncbi.nlm.nih.gov/). These compounds were linalool, hotrienol, p-anisaldehyde, 5-nonanone, S-methyl thiohexanoate, 1-nonen-3-ol, and 2-isopropyl-3-methoxypyrazine. Then, the 3D structures of five broad-tuning receptor proteins were downloaded from the UniProt database (https://www.uniprot.org/), including OR1A1 (UniProt ID: Q9P1Q5), OR1D2 (UniProt ID: P34982), OR1G1 (UniProt ID: Q15617), OR2W1 (UniProt ID: Q9Y3N9), and OR52D1 (UniProt ID: Q9H346). Molecular docking was performed using AutoDock Vina 1.5.7, and the binding energy (kcal/mol) was used to evaluate the results. Protein-ligand interaction forces were analyzed using the PLIP online tool (https://plip-tool.biotec.tu-dresden.de/) and visualized using PyMOL 3.1.3.

2.7. Statistical analysis

Data processing and flavor wheel creation were completed using Microsoft Excel 2021. Radar and bar graphs were created using Origin 2024. Pie charts and Venn diagrams were created using the online platform, https://www.bioinformatics.com.cn. PCA graphs, heat maps, differential analysis, and KEGG analysis were created and performed using Metware Cloud (https://cloud.metware.cn) and the online platform, https://www.chiplot.online/. Inkscape was used for graphic adjustments. All analyses were repeated three times.

3. Results

3.1. Sensory evaluation of three types of tea samples

Before the formal QDA, the evaluation panel engaged in extensive discussions and eventually agreed upon the following descriptive terms for the three tea aromas: floral, sweet, chestnut-like, fresh, and green. After each team member was familiar with the characteristics and intensity of the aromas, QDA was conducted on the three tea samples. The results in Fig. 2A-B show that NNT had a strong floral aroma, accompanied by a relatively strong sweet aroma, a moderate fresh aroma, a weak chestnut-like aroma, and a very weak green aroma. GT had strong fresh, chestnut-like, and green aromas, a relatively strong floral aroma, and a very weak sweet aroma. YT had a relatively strong sweet aroma, weak chestnut-like and green aromas, and very weak floral and fresh aromas. These results suggest that the NNT processing method contributed to the floral aroma of NNT tea infusion.

Fig. 2.

Fig. 2

Open in a new tab

Sensory evaluation of three types of tea samples. Appearance of tea leaves and tea infusions from three types of tea samples (A). Characteristic aroma radar chart of the three tea samples (B). GT: green tea. YT: yellow tea. NNT: Niangniang tea. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.2. Characterization of volatile compounds in three types of tea samples

The volatile compounds in the three types of tea were identified using HS-SPME-GC–MS (Fig. S1). Principal component analysis (PCA) showed that the first two principal components accounted for more than 70 % of the total variance in the data (Fig. 3A). The three types of tea showed clear separation, indicating differences in their volatile compounds. The three biological replicates of each sample also clustered together, indicating the reliability of the results. Similar PCA results were obtained through hierarchical clustering of the three types of tea (Fig. S2), suggesting that different processing methods significantly impacted the volatile compounds in tea.

Fig. 3.

Fig. 3

Open in a new tab

Analysis of volatile compounds in three types of tea samples. PCA score plot of three tea samples (A). Pie charts of volatile compounds in NNT (B). Pie charts of volatile compounds in GT (C). Pie charts of volatile compounds in YT (D). The chemical structures in each category correspond to the compounds with the highest relative content.

A total of 386 volatile compounds were detected, including 370 in NNT, 334 in GT, and 331 in YT. The compounds were classified into 13 categories based on their structures, including esters, alcohols, aldehydes, terpenoids, aromatics, heterocyclic compounds, amines, phenols, acids, hydrocarbons, ethers, nitrogen compounds, and ketones (Fig. 3B-D). Terpenoids (91 and 81) were the most abundant compounds in NNT and GT and accounted for 24.59 % and 24.25 % of the total compounds, respectively. (+)-alpha-Terpineol (floral aroma) and hotrienol (floral aroma) were the terpenoids with the highest relative contents in NNT and GT, respectively. Esters (88 and 80) accounted for 23.78 % and 23.95 % of the total compounds in NNT and GT, respectively. Butyl 2-methylcrotonate (fruity aroma) and 3-octenoic acid, ethyl ester (fruity aroma) were the esters with the highest relative contents in NNT and GT, respectively. The substance categories with the highest relative contents in YT were esters (78) and terpenoids (76), which respectively accounted for 23.56 % and 22.96 % of the total compounds. Representative substances included 3-octenoic acid, ethyl ester (fruity aroma), and hotrienol (floral aroma). GT processing primarily retained terpenoids, YT processing primarily retained esters, and NNT processing significantly increased both of their contents.

3.3. Differential analysis of volatile compounds in three types of tea samples

Partial least squares-discriminant analysis (PLS-DA) was performed to maximize intergroup differences, and permutation tests were conducted to validate the reliability of the model (Fig. S3). Then we introduced variable importance in projection (VIP) values and fold change (FC). A total of 172 volatile compounds (VIP > 1, FC ≥ 1.5, and FC ≤ 0.67, p < 0.05) were selected as the differential compounds for distinguishing between the three tea samples. These compounds were mainly composed of terpenoids, esters, and ketones, and their relative contents in NNT were significantly higher than in YT and GT (Fig. 4A). A Venn diagram was used to compare the differential volatile compounds of comparison group NNT vs. GT, NNT vs. YT, and GT vs. YT (Fig. 4B). There were 134 differential compounds (94 up-regulated and 40 down-regulated) in comparison group NNT vs. GT, 120 differential compounds (95 up-regulated and 25 down-regulated) in comparison group NNT vs. YT, and 53 differential compounds (44 up-regulated and 9 down-regulated) in comparison group GT vs. YT. Among the different substances, we found that 31, including dihydrocarvyl acetate and trans-nerolidol with a floral aroma, and isogeraniol and linalyl acetate with a sweet aroma, were present only in NNT. These results suggest that these compounds may be major contributors to NNT's unique aroma, which may have been produced due to NNT's processing method. Additionally, there were 7 compounds that were not detected in NNT, including 2,6-dimethylcyclohexanol (alcohol odor), ethylpyrrole (burnt odor), and furfuryl heptanoate (green odor). This may be why NNT has a stronger floral aroma than the other two types of tea. For YT, we found that seven substances, including cis-3,7-dimethyl-1,3,6-octatriene and 2,6-dimethyl-2,4,6-octatriene with a floral aroma, and 2-buten-1-ol,3-methyl-,benzoate and 1,5-dimethyl-1-vinylhex-4-enyl valerate with a fruity aroma, were not detected in YT, but they were present in NNT and GT. This might also be why the floral aroma of YT was weaker than those of NNT and GT.

Fig. 4.

Fig. 4

Open in a new tab

Differential metabolite analysis of three types of tea samples. Heat map of differential volatile compounds in three tea samples (A). Venn diagram and bar chart of different comparison groups (B). KEGG pathway enrichment of different comparison groups (C-E).

KEGG analysis revealed enriched metabolic pathways in the three comparison groups (Fig. 4C-E). The enriched pathways in the NNT vs. GT included sulfur metabolism, biosynthesis of monoterpenoids, diterpenoids, sesquiterpenoids, and triterpenoids, as well as cysteine and methionine metabolism. These results indicate that the main differences in volatile compounds caused by the NNT and GT processing methods were terpenoid synthesis and sulfur-containing amino acid pathways. The enriched pathways in NNT vs. YT included sulfur metabolism, cysteine and methionine metabolism, sesquiterpenoid and triterpenoid biosynthesis, and the biosynthesis of secondary metabolites. This indicates that the differences in volatile compounds resulting from the NNT and YT processing methods were primarily due to the sesquiterpenoid and triterpenoid metabolic pathways, as well as a series of secondary metabolic pathways and sulfur-containing amino acids. The GT vs. YT enrichment pathways included the biosynthesis of various secondary metabolites, phenylpropanoids, various alkaloids, monoterpenoids, sesquiterpenoids, and triterpenoids. This indicates that differences in volatile compounds resulting from the GT and YT processing methods were due not only to terpenoid synthesis, but also to phenylalanine metabolic pathways and other pathways. KEGG metabolic pathway analysis revealed that the processing methods of the three different tea samples primarily influenced the synthesis of terpenoids, the metabolism of sulfur-containing amino acids, and the secondary metabolite pathways. These factors caused significant variations in the composition of volatile compounds and their aromas among different types of tea.

3.4. Analysis of key aroma compounds in three types of tea samples

Odor researchers have suggested that volatile compounds with rOAV >1 contribute to the overall aroma composition of an analyzed sample (Chen et al., 2024). Referring to Yang et al. (Yang et al., 2024) and considering the number of volatile compounds in this study, we screened 172 differential substances (VIP > 1, FC ≥ 1.5, and FC ≤ 0.67, p < 0.05) and identified 18 compounds with rOAV ≥10. These were then subjected to correlation analysis with QDA results (r ≥ 0.7, p < 0.05) (Fig. 5A), which showed that most compounds positively correlated with sweet and floral aromas were negatively correlated with chestnut-like, green, and fresh aromas. 7 of these compounds were significantly positively correlated with floral aroma, including 1-nonen-3-ol, S-methyl thiohexanoate, 5-nonanone, linalool, non-8-enal, 2-isopropyl-3-methoxypyrazine, and p-anisaldehyde. Compounds significantly positively correlated with sweet aroma included p-anisaldehyde and butyl 2-methylcrotonate, while 7 compounds were significantly negatively correlated with sweet aroma, including delta-hexalactone, 1,2-dihydro-1,1,6-trimethylnaphthalene, 2-acetyl-3,4,5,6-tetrahydropyridine, hotrienol, p-cymene, trans-2-undecenal and cedrol. Ultimately, 16 key aroma compounds were detected in the three types of tea samples, and the results were presented as a Venn diagram (Fig. 5B). These 16 compounds included 2-isopropyl-3-methoxypyrazine, non-8-enal, S-methyl thiohexanoate, 1-nonen-3-ol, butyl 2-methylcrotonate, linalool, methyl benzoate, cedrol, p-anisaldehyde, p-cymene, 5-nonanone, trans-2-undecenal, hotrienol, delta-hexalactone, 1,2-dihydro-1,1,6-trimethylnaphthalene, and 2-acetyl-3,4,5,6-tetrahydropyridine. Aroma wheels were constructed using these compounds for NNT, GT, and YT to determine the olfactory characteristics of each sample. As shown in Fig. 5C, the aroma profiles of the three samples were significantly different. NNT dominated the floral aroma, and the rOAV of non-8-enal was much higher than that of the other floral compounds, reaching 5167.0 in NNT. p-Anisaldehyde, the main compound contributing to NNT's intense floral aroma, was only present in NNT, despite its low rOAV of 24.4. In addition, among the 16 compounds, 2-isopropyl-3-methoxypyrazine had the highest rOAV, and among the three tea samples, it also had the highest value in NNT. In GT, fruity and fresh aromas were more prominent. The rOAVs of p-cymene and trans-2-undecenal were significantly higher in GT than in NNT and YT. 2-Acetyl-3,4,5,6-tetrahydropyridine, which has a caramel odor, had an rOAV of 42 in GT, which was about 4-fold higher than that in YT. This substance was not detected in NNT. Correlation analysis revealed a significant positive correlation between this substance and chestnut-like aroma. Therefore, we speculate that it also contributed to the chestnut aroma in GT. Although there were no prominent volatile compounds in YT, butyl 2-methylcrotonate (sweet odor) showed an rOAV of 31.2 in YT, second only to NNT and 1.8-fold higher than that of GT. 2-Acetyl-3,4,5,6-tetrahydropyridine (caramel odor) had an rOAV of 10.6 in YT, second only to GT, and it was not detected in NNT. These compounds were also key contributors to YT's sweet aroma.

Fig. 5.

Fig. 5

Open in a new tab

Key volatile compounds in three types of tea samples. Heat map of the correlations between the main aroma compounds and aroma attributes (A). * significant, ** highly significant, and *** very highly significant. The tea group with the highest rOAV for each substance is also marked. Venn diagram of screened key volatile compounds (B). Aroma wheel for 16 key volatile compounds in NNT, GT, and YT (C). The aroma wheel is divided into two main parts: an inner circle with the names, structural formulas, aroma classifications, and characteristics of the 16 compounds; an outer circle heat map that displays the rOAV of each compound in the corresponding samples.

3.5. Aroma recombination and omission tests

3.5.1. Aroma recombination

An aroma recombinant model was established using 12 volatile compounds with rOAV ≥ 10 among the key aroma compounds in the three types of tea, excepting for four substances that were not available for purchase including non-8-enal, butyl 2-methylcrotonate, 1,2-dihydro-1,1,6-trimethylnaphthalene, and 2-acetyl-3,4,5,6-tetrahydropyridine. Then, the aroma characteristics of the recombinant models were compared through sensory evaluation (Fig. 6). The recombinant model of NNT showed strong floral and sweet aromas, slightly weaker fresh aroma, and the weakest green aroma. The recombinant model of GT exhibited stronger fresh, chestnut, and green aromas, slightly weaker floral aroma, and the weakest sweet aroma. The recombinant model of YT exhibited stronger sweet aroma and slightly weaker green aroma, and the weakest floral and fresh aromas. Both the NNT and YT recombinant models exhibited an extremely weak chestnut-like aroma. The recombinant model results closely resembled the actual samples in terms of floral, sweet, and fresh aroma attributes, further confirming its ability to identify and quantify volatile compounds.

Fig. 6.

Fig. 6

Open in a new tab

Radar chart of the aroma recombination. R-NNT: Recombinant model of NNT. R-GT: Recombinant model of GT. R-YT: Recombinant model of YT. GT: green tea. YT: yellow tea. NNT: Niangniang tea. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.5.2. Omission tests

To evaluate the contribution of 12 key aroma compounds with rOAV ≥ 10 to the aroma characteristics of NNT, 15 omission models were prepared by sequentially omitting one or a class of aroma compounds from the recombinant model of NNT. Both the recombinant model of NNT and the omission models were submitted to group members for evaluation using a triangle test. As shown in Table 1, models omitting all terpenoids, esters, and aldehydes respectively exhibited highly significant differences (M1, M2, M3; p ≤ 0.001). Models omitting all terpenoids and aldehydes exhibited a lower floral aroma, and the model omitting all esters exhibited a lower sweet aroma, suggesting that terpenoids, esters, and aldehydes were major contributors to the floral and sweet aromas of NNT. Compared with the recombinant model, all nine panel members recognized that the model omitting 2-isopropyl-3-methoxypyrazine showed highly significant differences (p ≤ 0.001) and had a more pronounced floral fragrance. When omitting linalool, p-anisaldehyde, 5-nonanone, and 1-nonen-3-ol, the differences between the omission model and the recombinant model were highly significant (0.001 < p ≤ 0.01). The floral fragrance was lower in the models that omitted linalool, p-anisaldehyde, and 5-nonanone, while the model omitting 1-nonen-3-ol showed a reduction in unpleasant odors and enhanced floral and sweet aromas. When hotrienol and S-methyl thiohexanoate were absent, the differences were significant (0.01 < p ≤ 0.05). The absence of hotrienol reduced the floral fragrance of the model, while the absence of S-methyl thiohexanoate reduced the unpleasant odors and enhanced the floral fragrance. The five omitted compounds associated with chestnut-like aroma, including cedrol, p-cymene, methyl benzoate, delta-hexalactone, and trans-2-undecenal were not significant. This may be related to the masking effect of other aroma compounds. These results indicate that there were seven compounds were key volatile compounds in NNT, among which linalool, p-anisaldehyde, 5-nonanone, and hotrienol enhanced the floral aroma of NNT. Another three substances, 2-isopropyl-3-methoxypyrazine, 1-nonen-3-ol, and S-methyl thiohexanoate, had negative effects on the aroma of NNT.

Table 1.

Omission tests from the recombination model of NNT.

Model Compounds CAS Right of selectiona Significanceb Sensory evaluationc
M1 All terpenoids – 9 *** Floral fragrance ↓
M1–1 Linalool 78–70-6 8 ** Floral fragrance ↓
M1–2 Cedrol 77–53–2 2 – –
M1–3 p-Cymene 99–87-6 4 – –
M1–4 Hotrienol 20,053–88-7 6 * Floral fragrance ↓
M2 All esters – 9 *** Sweet fragrance ↓
M2–1 Methyl benzoate 93–58-3 2 – –
M2–2 S-Methyl thiohexanoate 2432-77-1 6 * Fragrance ↑
M2–3 delta-Hexalactone 823–22–3 4 – –
M3 All Aldehydes – 9 *** Floral fragrance ↓
M3–1 p-Anisaldehyde 123–11-5 8 ** Floral fragrance ↓
M3–2 trans-2-Undecenal 53,448–07-0 5 – –
M4 5-Nonanone 502–56-7 7 ** Floral fragrance ↓
M5 1-Nonen-3-ol 21,964–44-3 8 ** Fragrance ↑
M6 2-Isopropyl-3-methoxypyrazine 25,773–40-4 9 *** Fragrance ↑
Open in a new tab

a Number of correct judgments in omission groups from 9 assessors.

b Significance: * significant (0.01 < p ≤ 0.05); ** highly significant (0.001 < p ≤ 0.01); *** very highly significant (p ≤ 0.001); − Not significant.

c Comparison of the omission group and sensory evaluation of tea samples. ↑ Improved aroma; ↓ Deteriorated aroma; − No obvious difference in aroma.

3.6. Verification of interactions between key volatile compounds in NNT and olfactory receptors

Interactions between aroma molecules and olfactory receptor molecules typically occur spontaneously and are related to molecular recognition and binding. As shown in Table 2, the binding energy, interaction regions, and amino acid residues of different receptors and volatile compounds showed significant variations. The binding energy ranged from −6.004 kcal/mol to −4.163 kcal/mol, indicating that olfactory receptors could spontaneously interact with these volatile compounds. The average binding energies of linalool, hotrienol, p-anisaldehyde, 5-nonanone, S-methyl thiohexanoate, 1-nonen-3-ol, and 2-isopropyl-3-methoxypyrazine were −5.398 kcal/mol, −5.281 kcal/mol, −5.453 kcal/mol, −4.692 kcal/mol, −4.386 kcal/mol, −4.842 kcal/mol, and −5.067 kcal/mol, respectively. The binding energies of OR1A1 with linalool, hotrienol, p-anisaldehyde, 5-nonanone, S-methyl thiohexanoate, 1-nonen-3-ol, and 2-isopropyl-3-methoxypyrazine were −5.374 kcal/mol, −5.419 kcal/mol, −5.835 kcal/mol, −4.698 kcal/mol, −4.261 kcal/mol, −5.018 kcal/mol, and −5.213 kcal/mol, respectively, with a maximum energy difference of 36.9 %. Similarly, the maximum differences in binding energies between OR1D2, OR1G1, OR2W1, and OR52D1 and various volatile compounds were 19.7 %, 31.7 %, 26.4 %, and 23.2 %, respectively, indicating the diversity and selectivity of the interactions between olfactory receptors and ligands. Fig. 7 shows the molecular docking simulation results for five olfactory receptors and their volatile compounds with the lowest binding energy: OR1A1 and p-anisaldehyde, OR1D2 and p-anisaldehyde, OR1G1 and linalool, OR2W1 and hotrienol, and OR52D1 and p-anisaldehyde. p-Anisaldehyde, which has a floral aroma, exhibited the lowest binding energy for three olfactory receptors. The results indicate that p-anisaldehyde, linalool, and hotrienol were key floral aroma compounds in NNT.

Table 2.

Summary of the binding energy, interaction types, and key residues between ORs and ligands.

Receptora Ligandb Binding energyc (kcal/mol) Hydrogen bondsd Hydrophobic interactionse π-π Stackingf
OR1A1 p-Anisaldehyde −5.835 ASN109, ASN155 ILE105, VAL254, TYR258 PHE206
Hotrienol −5.419 SER242 TYR60, LEU63, THR239, TYR288, ASN292 –
Linalool −5.374 THR263 PRO183, THR257, LEU262, TYR265, ASP269 –
2-Isopropyl-3-methoxypyrazine −5.213 THR257, THR263 THR182, THR257, TYR265, ASP269 –
1-Nonen-3-ol −5.018 THR263 THR182, PRO183, LEU262, TYR265, ASP269 –
5-Nonanone −4.698 – THR182, PRO183, LYS186, THR257, ARG260, TYR265 –
S-Methyl hexanethioate −4.261 THR263 PRO183, LYS186, TYR265 –
OR1D2 p-Anisaldehyde −6.044 TYR155 PHE207, LEU255, TYR259 PHE207
Linalool −5.803 – LEU199, PHE207, TYR252, LEU255, TYR259 –
Hotrienol −5.754 LEU199 LEU199, PHE207, LEU255, TYR259 –
2-Isopropyl-3-methoxypyrazine −5.357 – LEU199, PHE207, TYR259 PHE207
1-Nonen-3-ol −5.346 – LEU199, PHE207, LEU208, LEU255, TYR259 –
5-Nonanone −5.268 – LEU199, PHE207, TYR252, LEU255, TYR259 –
S-Methyl hexanethioate −5.049 – PH207, TYR252, LEU255, TYR259 –
OR1G1 Linalool −5.481 ALA108 LEU105, ALA108, ILE109, PHE251, TYR259, TYR278 –
p-Anisaldehyde −5.219 – ILE207 TYR259
Hotrienol −5.016 SER64 TRP50, PHE68, ILE145 –
2-Isopropyl-3-methoxypyrazine −4.981 ALA108, TYR259 LEU105, ILE109, TYR259 –
1-Nonen-3-ol −4.573 – ALA108, ASN206, ILE207, PHE251, TYR259 –
5-Nonanone −4.526 – LEU105, ILE109, ILE207, PHE251, ALA255, TYR259, TYR278 –
S-Methyl hexanethioate −4.163 – PHE104, ALA108, PHE251, ALA255, TYR259, TYR278 –
OR2W1 Hotrienol −5.272 – VAL185,GLU196,VAL199,PHE200 –
Linalool −5.216 TYR259 LEU181, GLU196, VAL199, PHE200, ASN264 –
2-Isopropyl-3-methoxypyrazine −5.019 GLU196, GLN261 LEU181, VAL199 –
p-Anisaldehyde −4.89 GLU196 GLU196, VAL199 TYR260
1-Nonen-3-ol −4.686 SER217 ALA116, VAL117, TYR120, PHE123, LEU213, ILE216, TYR220 –
5-Nonanone −4.416 – LEU181, PRO182, VAL185, GLU196 –
S-Methyl hexanethioate −4.172 – PRO58, PHE61, PRO138, –
OR52D1 p-Anisaldehyde −5.277 GLU183 ALA258, TYR284 TYR111
Linalool −5.115 SER261 TYR111, ALA206, ALA258, PHE259, PHE262 –
Hotrienol −4.946 HIS108 TYR111, ALA258, PHE259, PHE262 –
2-Isopropyl-3-methoxypyrazine −4.763 GLU183 TYR111, ALA258 –
1-Nonen-3-ol −4.585 – TYR111, LEU207, MET210, ALA258, PHE259, PHE262, –
5-Nonanone −4.553 – ALA206, LEU207, ALA258, PHE259, PHE262 –
S-Methyl hexanethioate −4.285 – ALA206, MET210, ALA258, PHE259, PHE262 –
Open in a new tab

a Five proteins were selected as the olfactory receptors.

b Seven key compounds in NNT were selected as ligands.

c Binding energy between olfactory receptors and ligands.

d Key residues of OR that form hydrogen bonds.

e Key residues of the OR that form hydrophobic interactions.

f Key residues of the OR that form π-π stacking.

Fig. 7.

Fig. 7

Open in a new tab

Molecular docking diagram of the NNT key volatile compounds with the highest binding energy to five olfactory receptors. p-Anisaldehyde-OR1A1 (A). p-Anisaldehyde-OR1D2 (B). Linalool-OR1G1 (C). Hotrienol-OR2W1 (D). p-Anisaldehyde-OR52D1 (E).

Hydrophobicity drives interactions between volatile compounds and olfactory receptors, and by analyzing the binding sites, we discovered that all five olfactory receptors possessed amino acid residues that bound to NNT volatile compounds via hydrophobic interactions. Hotrienol was stabilized in OR2W1 via hydrophobic interactions with VAL185 (3.86 Å), GLU196 (3.65 Å, 3.52 Å), VAL199 (3.74 Å), and PHE200 (3.60 Å, 3.66 Å). Linalool interacted with LEU105 (3.69 Å), ALA107 (3.90 Å), ILE109 (3.65 Å), PHE251 (3.93 Å), TYR259 (3.7 Å, 3.70 Å, 3.58 Å), and TYR278 (3.90 Å) in OR1G1. p-Anisaldehyde interacted with OR1A1 at ILE105 (3.72 Å, 3.99 Å), VAL254 (3.69 Å), and TYR258 (3.46 Å); with OR1D2 via PHE207 (3.74 Å), LEU255 (3.61 Å), and TYR259 (3.47 Å, 3.62 Å, 3.84 Å); and with OR52D1 via ALA258 (3.75 Å) and TYR284 (3.70 Å). In addition to hydrophobic interactions, linalool could form hydrogen bonds with OR1G1 via ALA108 (2.97 Å). p-Anisaldehyde formed hydrogen bonds with OR1A1 via ASN109 (2.09 Å) and ASN155 (2.4 Å); with OR1D2 via TYR155 (2.62 Å); and with OR52D1 via GLU183 (2.87 Å). Due to its benzene ring, p-anisaldehyde also engaged in π-π stacking with OR1A1, OR1D2, and OR52D1, which enhanced its binding to olfactory receptors. These results suggest that p-anisaldehyde, linalool, and hotrienol bound to olfactory receptors via hydrophobic interactions, hydrogen bonds, and π-π stacking, imparting a floral aroma to NNT.

4. Discussion

4.1. Two-stage bunch-drying of NNT promotes the formation and release of floral aroma compounds

One of the key factors determining the quality of tea is its aroma, which is significantly affected by the processing method, particularly drying, which is the final processing step. Different drying temperatures can change the content and composition of the aromatic compounds in green tea, creating unique profiles with fresh, floral, chestnut-like, and bean-like aromas (Wang, Qu, et al., 2022). Yellowing reduces the fresh and green aromas of the tea while enhancing sweet, nutty, baked, and crispy-rice-like aromas (Yin et al., 2023). This occurs because a high temperature and high humidity cause compounds to undergo a series of non-enzymatic reactions, such as degradation, oxidation, and hydrolysis (Fan et al., 2022). Yellowing is a key step in the formation of the aroma of YT and is the main reason its aroma is significantly different from GT. Compared with GT's rapid high-temperature drying and YT's yellowing process, NNT undergoes two rounds of bunch-drying. The first bunch-drying round creates a humid and hot environment, while the temperature and humidity of the second bunch-drying are lower. Similar to yellowing, we speculate that the chemical changes caused by the high temperature and humidity environment during the first bunch-drying round are more important than those during the second bunch-drying round. This promotes the non-enzymatic reactions of compounds, which release more floral, sweet aroma compounds, in agreement with Wei, Liu, et al. (2025) regarding the key steps in the formation of NNT flavor. The QDA results of the three types of tea samples showed that NNT had a dominant floral aroma accompanied by a strong sweet aroma, GT had the strongest chestnut-like aroma, accompanied by a floral aroma, and YT had a stronger sweet aroma (Fig. 2B). This indicates that the two bunch-drying rounds of NNT were crucial to the formation of its characteristic aroma. Because this process combines the GT and YT processes, NNT has aroma characteristics of both GT and YT, resulting in a rich floral and sweet fragrance.

GC–MS was used to identify the volatile compounds in the three types of tea samples to further explain the reason for the aroma differences between NNT, GT, and YT, as well as the material basis for the floral aroma produced by NNT. The combined proportion of terpenoids, esters, and ketones exceeded half of the total volatile compounds, accounting for 58.64 %, 59.28 %, and 57.7 % in NNT, GT, and YT, respectively (Fig. 3B-D). Most terpenoids have floral or fruity aromas (Liu et al., 2023), most esters have mild fruity aromas (Shi et al., 2019), and ketones have a low odor threshold and unique floral and woody aromas (Lv et al., 2014). This suggests that all three types of tea sample made from the Zhenfeng variety have pleasant aromas. We speculate that the Zhenfeng variety itself has a high-quality aroma, consistent with Li et al. (2022) and Wei, Mu, et al. (2025). Wen et al. (2023) investigated the aromas of green tea and yellow tea and found that the main volatile compounds during the processing of green tea and yellow tea were ketones and heterocyclic compounds. Hong et al. studied four different aromatic types of yellow tea and found that the main volatile compounds were alcohols and esters. The main volatile compounds differed among the different aromatic types of yellow tea (Hong et al., 2022), which may be related to the raw materials and their drying method. Because GT, YT, and NNT are formed from the same fresh tea leaves, spreading, fixation, and rolling conditions, the significant changes in the content and composition of their aroma precursor substances and their derivatives likely occurred during the final processing step. This ultimately formed the distinctive aroma characteristics of the three different tea samples. Therefore, the strong floral aroma of NNT is related to the tea variety, as well as the processing technique, and the bunch-drying process may have been responsible for producing the floral and sweet aromas.

Previous studies have shown that the hot and humid condition during the yellowing process can promote a series of non-enzymatic thermal degradation reactions and Strecker degradation (Dong et al., 2023; Shi et al., 2021). High-humidity treatment also contributes the formation of key floral compounds such as 1-octanol and linalool (Wang, Feng, Min, Yin and Jiang, 2024). Wei et al. further indicate that the optimized heating and humidification process accelerates the non-enzymatic release of glycoside precursors, promoting the accumulation of linalool and geraniol, thereby enhancing the sweet and floral aromas of yellow tea (Wei et al., 2024). Compared to GT, the two-stage bunch-drying of NNT creates a hot and humid condition within the tea leaves. Although YT is also dried under humid condition, the higher drying temperature in NNT may significantly promotes the hydrolysis of glycoside. This may also explain why the content of 1-octanol, linalool and isogeraniol in NNT is much higher than that in GT and YT. Based on the above discussion, we speculate that compared to the drying methods of GT and YT, the two-stage bunch-drying of NNT, especially the first bunch-drying with higher humidity, can further promote the hydrolysis and release of glycoside precursors, generating more floral compounds.

4.2. p-Anisaldehyde, linalool, and hotrienol are key floral compounds in NNT

The aromatic characteristics of tea are determined by the concentration of volatile compounds, as well as their odor thresholds. The odor thresholds of different volatile compounds vary significantly, making it necessary to consider both the concentration and odor threshold of a compound when identifying key volatile compounds. We selected 16 compounds with rOAV ≥ 10 and r ≥ 0.7 from the differential compounds as key volatile compounds to distinguish the three types of tea samples (Fig. 5B-C), followed by aroma recombination and omission tests.

The aroma recombination tests confirmed the identification and quantification of volatile compounds in the three types of tea samples, but the chestnut-like aroma attribute could not be adequately reproduced. In the correlation analysis, 1,2-dihydro-1,1,6-trimethylnaphthalene and 2-acetyl-3,4,5,6-tetrahydropyridine were significantly and positively correlated with the chestnut aroma, but they could not be purchased. Their absence may partially explain the low chestnut-like aroma score. This study also found that models omitting cedrol and delta-hexalactone, respectively, showed no significant differences from the recombinant model in terms of sensory evaluation. Both compounds were significantly positively correlated with chestnut-like aroma. This suggests that cedrol and delta-hexalactone were not the key substances responsible for the chestnut-like aroma in tea, inconsistent with the conclusion of Wang et al. (2020) who identified cedrol as a key volatile compound responsible for the chestnut-like aroma. This indicates that in order to more reliably identify key aroma compounds, it is necessary to combine data analysis methods such as rOAV analysis and correlation analysis with sensory verification methods such as aroma recombination and omission tests to comprehensively evaluate the actual contribution of compounds.

The omission tests identified that linalool, hotrienol, p-anisaldehyde, and 5-nonanone were the key compounds responsible for a floral aroma in NNT. Of these, linalool and hotrienol are closely associated with floral and sweet aromas and are present in many flowers, tea, and honey (Jia et al., 2025; Li et al., 2024; Wang et al., 2022). 5-Nonanone also has floral and fruity aromas (Lu, Shi, et al., 2024). Wang et al. (2025) found that an increase in p-anisaldehyde enhanced the floral aroma of tea. Combining sensory experiments, Castro-Vazquez et al. identified the sinensal isomers, which were only present in citrus honey, as the best floral marker compounds for citrus honey (Castro-Vázquez et al., 2007). In this study, one of the four floral compounds, p-anisaldehyde, was found exclusively in NNT. This suggests that NNT's two-stage bunch-drying process contributes to the formation of it. Conversely, p-anisaldehyde can also serve as a potential floral marker compound to distinguish NNT from other tea. The other three key compounds included S-methyl thiohexanoate (cabbage odor), 1-nonen-3-ol (mushroom odor), and 2-isopropyl-3-methoxypyrazine (green pepper odor). In this study, 2-isopropyl-3-methoxypyrazine negatively impacted the aroma of NNT, similar to Coetzee et al., who reported that 3-isobutyl-2-methoxypyrazine (IBMP) typically interacts with volatile phenols and volatile thiols such as 3-mercaptohexan-1-ol. These compounds are produced during wine fermentation and aging and enhance undesirable aromas and mask positive flavors (Coetzee et al., 2015). Berger et al. conducted sensory evaluations of individual components (C2-C18) in an S-methyl thioester library and found that low concentrations of S-methyl thiohexanoate created a floral and green aroma. It is the only compound in the S-methyl thioester library that is considered to have a sweet aroma (Berger et al., 1999). However, this sulfide was negatively correlated with floral and sweet aromas in the omission test of this study, possibly due to its excessive concentration in NNT. Overall, among the seven key aromatic compounds, linalool, hotrienol, p-anisaldehyde, and 5-nonanone contributed significantly to the floral, fruity, and sweet aromas of NNT.

Molecular docking is used to explore interactions between volatile compounds and olfactory receptors to understand their binding mechanisms and evaluate their sensory effects (Xiao et al., 2024). In this study, molecular docking was used to study the binding energies and interactions between the 7 key NNT volatile compounds and 5 broad-tuning receptor proteins (Table 2). The lower the binding energy, the stronger the affinity of the volatile compound for the receptor. p-Anisaldehyde, linalool, and hotrienol exhibited the strongest affinity with the protein receptors, while S-methyl thiohexanoate exhibited the weakest. Molecular docking analysis indicates that hydrophobic interactions provide the fundamental stability for binding, while hydrogen bonds and π-π stacking represent the two key forces governing the interaction (Zeng et al., 2023). In this study, the interactions between the 7 key aroma compounds and the 5 receptor proteins were mainly formed by hydrophobic interactions and hydrogen bonds. However, among them, S-methyl thiohexanoate only binds to the receptor protein through hydrophobic interaction. In contrast, p-anisaldehyde can form two or more interactions with all 5 receptor proteins. Moreover, it is the only compound capable of forming π-π stacking interactions with all five receptors. This explains why S-methyl thiohexanoate was the least significant among the 7 key aroma compounds during the omission test, while p-anisaldehyde not only showed significant in the omission test but was also confirmed as an NNT floral marker compound. Wang et al. believe that compared with the linear structure, the cyclic structure can provide more action sites for amino acid residues (Wang et al., 2025). In this study, S-methyl thiohexanoate, which has the lowest affinity with the five receptors, shows a linear structure, while p-anisaldehyde, which has a stronger affinity, has a benzene ring structure. This result further confirms the theory put forward by Wang et al. This selective binding mechanism, combined with steric hindrance effects, enables p-anisaldehyde to efficiently activate a broader range of olfactory neurons even at relatively low concentrations (rOAV of 24.4). This is similar to the results reported by Luan et al., who showed that the olfactory receptor OR75 in the odorants of a mixture only responded to β-ionone, undecanal, cinnamaldehyde, and p-anisaldehyde. It showed the highest sensitivity to p-anisaldehyde (Luan et al., 2025). Overall, the stronger binding affinities of p-anisaldehyde, linalool, and hotrienol allowed them to be more quickly recognized and activated by olfactory receptors, thereby releasing their floral aromas.

The rOAV of 2-isopropyl-3-methoxypyrazine is significantly higher than that of several other compounds, but it was not the compound with the highest binding affinity to the five olfactory receptors examined in this study. There are nearly 400 types of human olfactory proteins, but this study only examined five broad-tuned receptor proteins that can simultaneously bind to multiple compounds. Therefore, these 5 proteins were not the most specific with the highest binding affinity for 2-isopropyl-3-methoxypyrazine. Wang et al. (2024) identified PxylOR29, PxylOR31, and PxylOR46 as potential molecular targets for 2,3-dimethyl-6-(1-hydroxy)-pyrazine. Molecular docking revealed that PxylOR31 exhibited the most stable binding to this compound. Further research is needed to identify the optimal receptor protein for 2-isopropyl-3-methoxypyrazine.

4.3. Synthesis mechanism of key aroma compounds

The aroma produced during tea drying mainly originates from a series of non-enzymatic thermochemical reactions, including the hydrolysis of glycosides, the oxidation and degradation of lipids, and the Maillard reaction (Ho et al., 2015; Yang et al., 2013). In the processing of the three types of tea samples in this study, only the drying conditions differed. Therefore, we speculate that the hydrothermal conditions created by NNT bunch-drying caused the conversion of NNT's aromatic substances (Fig. 8). Terpenoids are an important class of plant secondary metabolites that play a significant role in the aroma of tea. Linalool is abundant in tea leaves and exists in two forms: free type and glycoside type (Yan et al., 2024). Although linalool in its glycoside-bound state is non-volatile and odorless, the enzymatic hydrolysis and heat treatment during tea processing may release it from its glycoside precursor, thereby enhancing the aromatic quality of tea (Chen & Quek, 2023). In this study, the linalool content in NNT was significantly higher than that in GT and YT. Because the preliminary processing steps for the three types of tea were identical, we speculate that the bunch-drying process of NNT may have promoted the hydrolysis of linalool glycosides, allowing more of it to be released from the glycoside precursors. Wang, Liu, et al. (2024) found that more linalool was retained at a drying temperature of 85 °C. As the second drying temperature of NNT was also 85 °C, this may have promoted the retention of linalool. Hotrienol is generated during heat-drying or baking (Zeng et al., 2022) and is a particularly unstable terpenoid characterized by floral and sweet aromas (Zhu et al., 2021). Under hydrothermal conditions, hotrienol can be formed by dehydration of 2,6-dimethylocta-3,7-diene-2,6-diol, while 2,6-dimethylocta-3,7-diene-2,6-diol is derived from the transformation of linalool (Jerkovic & Kus, 2014). In this study, the content of hotrienol in GT was significantly higher than that in NNT and YT. Compared with the yellowing and drying of YT and the bunch-drying of NNT, the drying environment of GT had a higher temperature and lower humidity, which was more conducive to the formation of hotrienol.

Fig. 8.

Fig. 8

Open in a new tab

Synthesis mechanism of key aroma compounds of NNT. Different colors represent different reaction mechanisms. Solid arrows indicate reactions that have been confirmed by literature, while dashed arrows indicate speculated reactions. The three boxes represent three types of tea samples (left to right: NNT, GT, and YT), with red and blue representing high and low levels of volatile compounds, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Key volatile compounds produced during lipid degradation include C6-C10 fatty aldehydes, epoxydecene aldehydes, ketones, alcohols, and esters (Yin et al., 2022). Compared with GT and YT, the content of 1-nonen-3-ol (mushroom aroma) and 5-nonanone (floral aroma) was highest in NNT. Previous studies have shown that lipids are prone to thermal oxidation (Ho et al., 2015), which may lead to the decomposition of saturated fatty acids into short-chain compounds. This forms various aldehydes, ketones, and unsaturated alcohols (Lan et al., 2025). 1-Octen-3-ol, which has a similar odor and structure to 1-nonen-3-ol, is a product of the oxidation and cracking of linoleic acid (Li et al., 2018). 5-Nonenone may primarily originate from the lipid oxidation of unsaturated fatty acids in tea leaves under thermal conditions (Lu, Shi, et al., 2024). The first bunch-drying step of the NNT process created a thermal and humid environment, which may have promoted the accumulation of 1-nonen-3-ol and 5-nonanone.

p-Anisaldehyde has a sweet and floral aroma and is present in fennel (Acimovic et al., 2025; Maikhunthod & Marriott, 2013). It is typically produced via the methylation and oxidation of p-cresol or the oxidation of cis-anethol (Lin et al., 2022). In this study, both p-cresol and p-anisaldehyde were detected only in NNT, which showed that the humid and hot environment may be crucial for the synthesis of these two substances. The thermochemical conversion of lignin produces various compounds, including bio-oils containing phenolic compounds (Cesari, Mutelet and Canabady-Rochelle, 2019). Previous studies have found that lignin can be oxidatively degraded into lignocresol, whose main component is p-cresol (Yoshida et al., 2005). Therefore, we speculate that under the hydrothermal conditions during the bunch-drying step of NNT, lignin underwent thermal degradation and was converted into p-cresol (Cesari, Canabady-Rochelle and Mutelet, 2019), which was then further oxidized to p-anisaldehyde.

During the processing of tea, especially the drying process, the Maillard reaction occurs and produces 3 main types of compounds: sulfides, Strecker aldehydes, and many heterocyclic compounds such as pyrazines, furans, pyrroles, thiophene, and their derivatives (Qu et al., 2023; Yin et al., 2022). S-Methyl thiohexanoate has a fresh and fruity aroma (Ren et al., 2025), but its sulfur atoms give it a dual nature. At high concentrations, it may exhibit a chemical odor similar to that of sulfur (Du et al., 2024). 2-Isopropyl-3-methoxypyrazine (IPMP) is a methoxypyrazine, a class of compounds that contribute to green, earthy, bean, and sweet pepper aromas (Sala et al., 2002). G.J. Pickering et al. believe that although IPMP is not usually the main methoxypyrazine, its low threshold and high concentration mean that it may play an important role in the flavor of wine (Pickering et al., 2007). IPMP has also been detected in high-end green tea (Flaig et al., 2020). In this study, the contents of S-methyl thiohexanoate and IPMP were the highest in NNT. Methionine is the main precursor of sulfur compounds with unpleasant odors, and its Strecker aldehyde (methional) can be converted into methanethiol (Gijs et al., 2000). As a direct precursor of many sulfur compounds, we speculate that hydrothermal conditions may have caused methanethiol to react with hexanoic acid via a non-enzymatic esterification reaction to form S-methyl thiohexanoate (Ho et al., 2015). Amino ketones are the main precursors for the formation of most pyrazines (Cherniienko et al., 2022). During the two-stage bunch-drying process of NNT, valine may first undergo amidation, followed by Strecker degradation to yield isobutyraldehyde and amino ketone. Subsequently, two molecules of amino ketone undergo oxidation and condensation to form 2-isopropylpyrazine, which then undergoes methoxylation to produce IPMP (Lei et al., 2018). In summary, the Maillard reaction occurring in the humid, high-temperature environment of the NNT two-stage bunch-drying process promoted the accumulation of these two negative aroma compounds. Li et al. found that the content of heterocyclic compounds increased with the increase of drying temperature and drying time, while the content of dimethyl sulfide first increased and then decreased with the increase of drying time and drying temperature (Li et al., 2025). Furthermore, it was discovered that a roasting time over 25 min had a positive impact on the concentration of all methylpyrazines (Selmat et al., 1998). In this study, both NNT and GT were dried at 90 °C. However, NNT accumulated more sulfur compounds and heterocyclic compounds than GT. Therefore, we speculate that the increase in these two negative aroma compounds might be related to the excessively long drying time at high temperatures. In conclusion, future research will focus on specifically inhibiting the accumulation of such negative aroma compounds by optimizing the bunch-drying process such as appropriately shortening the bunch-drying time, and further optimizing the aroma of NNT.

5. Conclusion

This study identified 7 key aroma compounds in NNT, including linalool, hotrienol, p-anisaldehyde, 5-nonanone, S-methyl thiohexanoate, 1-nonen-3-ol, and 2-isopropyl-3-methoxypyrazine. Among these, p-anisaldehyde, linalool, and hotrienol conferred rich floral aroma, with p-anisaldehyde being the marker of the floral aroma of NNT. This study also reports the mechanisms by which hydrothermal conditions during the two-stage bunch-drying process in NNT promoted glycoside hydrolysis, lipid oxidative degradation, lignin degradation and the Maillard reaction, which contributed to NNT's floral aroma. The Maillard reaction during the two-stage bunch-drying process of NNT caused the accumulation of negative compounds such as S-methyl thiohexanoate and 2-isopropyl-3-methoxypyrazine. Future research should focus on optimizing the parameters of the two-stage bunch-drying process (such as appropriately shortening the bunch-drying time) to reduce the accumulation of these negative compounds and establish a standardized process to enhance the floral aroma of green tea.

CRediT authorship contribution statement

Xinyi Sun: Writing – original draft, Software, Methodology, Investigation. Ceyu Wang: Validation, Formal analysis, Conceptualization. Xingyu Long: Methodology, Investigation. Shumei Han: Validation, Software. Wei Xu: Methodology, Data curation. Taolin Chen: Validation, Investigation. Jianjun Liu: Validation, Conceptualization. Beibei Wen: Methodology, Investigation. Meifeng Li: Writing – review & editing, Project administration, Methodology, Investigation.

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.

Acknowledgments

This study was supported by the financial support of the National Key Research and Development Program (2022YFD1600802), the National Natural Science Foundation of China (32260786, 32460783), the Basic Research Program of Guizhou University (GDJC [2023] 05), the Science and Technology Plan Project of Guizhou Province ([2024] 581), and the Science and Technology Plan Project of Qiannan City ([2023] 12).

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.fochx.2025.103345.

Contributor Information

Jianjun Liu, Email: junjian.liu@163.com.

Beibei Wen, Email: bbwen@gzu.edu.cn.

Meifeng Li, Email: iamlimeifeng@126.com.

Appendix A. Supplementary data

Supplementary material 1: Fig. S1. Total ion current diagram of mixed sample mass spectrometry analysis. Fig. S2. Cluster tree diagram of three tea samples. GT: green tea. YT: yellow tea. NNT: Niangniang tea. Fig. S3. PLS-DA cross-validation findings utilizing a permutation test. R²: Coefficient of Determination. Q²: Predictive Ability.

mmc1.docx (692.2KB, docx)

Supplementary material 2

mmc2.xlsx (51.7KB, xlsx)

Data availability

Data will be made available on request.

References

  1. Acimovic M., Jeremic J.S., Cvetkovic M., Loncar B., Pezo L. Comparative analysis of essential oils and hydrolates from aniseed, coriander and fennel fruits. Chemistry & Biodiversity. 2025;10, Article 202500022 doi: 10.1002/cbdv.202500022. [DOI] [PubMed] [Google Scholar]
  2. Berger C., Martin N., Collin S., Gijs L., Khan J.A., Piraprez G.…Vulfson E.N. Combinatorial approach to flavor analysis. 2. Olfactory investigation of a library of S-methyl thioesters and sensory evaluation of selected component. Journal of Agricultural and Food Chemistry. 1999;47(8):3274–3279. doi: 10.1021/jf990205v. [DOI] [PubMed] [Google Scholar]
  3. Castro-Vázquez L., Díaz-Maroto M.C., Pérez-Coello M.S. Aroma composition and new chemical markers of Spanish citrus honeys. Food Chemistry. 2007;103(2):601–606. doi: 10.1016/j.foodchem.2006.08.031. [DOI] [Google Scholar]
  4. Cesari L., Canabady-Rochelle L., Mutelet F. Separation of phenols from lignin pyrolysis oil using ionic liquid. Separation and Purification Technology. 2019;209:528–534. doi: 10.1016/j.seppur.2018.07.083. [DOI] [Google Scholar]
  5. Cesari L., Mutelet F., Canabady-Rochelle L. Antioxidant properties of phenolic surrogates of lignin depolymerisation. Industrial Crops and Products. 2019;129:480–487. doi: 10.1016/j.indcrop.2018.12.010. [DOI] [Google Scholar]
  6. Chen G., Zhu G., Xie H., Zhang J., Huang J., Liu Z., Wang C. Characterization of the key differential aroma compounds in five dark teas from different geographical regions integrating GC-MS, ROAV and chemometrics approaches. Food Research International. 2024;194, Article 114928 doi: 10.1016/j.foodres.2024.114928. [DOI] [PubMed] [Google Scholar]
  7. Chen X., Quek S.Y. Free and glycosidically bound aroma compounds in fruit: Biosynthesis, transformation, and practical control. Critical Reviews in Food Science and Nutrition. 2023;63(28):9052–9073. doi: 10.1080/10408398.2022.2064422. [DOI] [PubMed] [Google Scholar]
  8. Cherniienko A., Pawelczyk A., Zaprutko L. Antimicrobial and odour qualities of Alkylpyrazines occurring in chocolate and cocoa products. Applied Sciences-Basel. 2022;12(22):27. doi: 10.3390/app122211361. [DOI] [Google Scholar]
  9. Coetzee C., Brand J., Emerton G., Jacobson D., Ferreira A.C.S., du Toit W.J. Sensory interaction between 3-mercaptohexan-1-ol, 3-isobutyl-2 -methoxypyrazine and oxidation-related compounds. Australian Journal of Grape and Wine Research. 2015;21(2):179–188. doi: 10.1111/ajgw.12133. [DOI] [Google Scholar]
  10. Cui L., Wang X., He C., Liu Z., Liang J. Effect of puffing treatment on volatile components of green tea explored by gas chromatography-mass spectrometry and gas chromatography-olfactometry. Food Chemistry-X. 2024;23, Article 101746 doi: 10.1016/j.fochx.2024.101746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dong R., Sheng X., Xie Q., Huang X., Yan F., Liu S. Aroma formation and transformation during sealed yellowing process of Pingyang yellow tea. Food Research International, Article. 2023;165 doi: 10.1016/j.foodres.2023.112535. [DOI] [PubMed] [Google Scholar]
  12. Du L., Ye Y., Shao J., Wang Y., Zhu G., Jiang H., Liu H., Liu Z. Characterization of primary aroma compounds in Pu-erh raw tea sourced from various regions using gas chromatography-mass spectrometry and headspace solid-phase microextraction. Journal of Food Science. 2024;89(12):9198–9213. doi: 10.1111/1750-3841.17562. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Fan F., Zhou S., Qian H., Zong B., Huang C., Zhu R., Guo H., Gong S. Effect of yellowing duration on the chemical profile of yellow tea and the associations with sensory traits. Molecules. 2022;27(3):940. doi: 10.3390/molecules27030940. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Flaig M., Qi S., Wei G., Yang X., Schieberle P. Characterization of the key odorants in a high-grade Chinese green tea beverage (Camellia sinensis; Jingshan cha) by means of the sensomics approach and elucidation of odorant changes in tea leaves caused by the tea manufacturing process. Journal of Agricultural and Food Chemistry. 2020;68(18):5168–5179. doi: 10.1021/acs.jafc.0c01300. [DOI] [PubMed] [Google Scholar]
  15. Gijs L., Perpete P., Timmermans A., Collin S. 3-methylthiopropionaldehyde as precursor of dimethyl trisulfide in aged beers. Journal of Agricultural and Food Chemistry. 2000;48(12):6196–6199. doi: 10.1021/jf0007380. [DOI] [PubMed] [Google Scholar]
  16. Guo X., Schwab W., Ho C., Song C., Wan X. Characterization of the aroma profiles of oolong tea made from three tea cultivars by both GC-MS and GC-IMS. Food Chemistry. 2021;376, Article 131933 doi: 10.1016/j.foodchem.2021.131933. [DOI] [PubMed] [Google Scholar]
  17. 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]
  18. Hong X., Wang C., Jiang R., Hu T., Zheng X., Huang J., Liu Z., Li Q. Characterization of the key aroma compounds in different aroma types of Chinese yellow tea. Foods. 2022;12(1):27. doi: 10.3390/foods12010027. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Jerkovic I., Kus P.M. Terpenes in honey: occurrence, origin and their role as chemical biomarkers. RSC Advances. 2014;4(60):31710–31728. doi: 10.1039/c4ra04791e. [DOI] [Google Scholar]
  20. Jia X., Gao Y., Xi H., Cui C., Yang X., He B., Xu C., Gao M., Li T. A flavor imitation method for Osmanthus aroma based on molecular docking screening and odor activity value analysis. Lwt-food. Science and Technology. 2025;223, Article 117697 doi: 10.1016/j.lwt.2025.117697. [DOI] [Google Scholar]
  21. Lan Q., Zhao X., Qin J., Tang J., Zeng R., Liu Y., Du X., Zhu C., Laghi L. Dynamic changes and correlations of physicochemical, volatile metabolites, and bacterial communities during fermentation of sour meat fermented from goose meat. Lwt-food. Science and Technology. 2025;224:117864. doi: 10.1016/j.lwt.2025.117864. [DOI] [Google Scholar]
  22. Lei Y., Xie S., Guan X., Song C., Zhang Z., Meng J. Methoxypyrazines biosynthesis and metabolism in grape : A review. Food Chemistry. 2018;245:1141–1147. doi: 10.1016/j.foodchem.2017.11.056. [DOI] [PubMed] [Google Scholar]
  23. Li H., Liu Z., Shuai M., Song M., Qiao D., Peng W., Chen L. Characterization of Evodia rutaecarpa (Juss) Benth honey: Volatile profile, odor-active compounds and odor properties. Journal of the Science of Food and Agriculture. 2024;104(4):2038–2048. doi: 10.1002/jsfa.13088. [DOI] [PubMed] [Google Scholar]
  24. Li L., Zan J., Chen W., Zong X., Yuan H., Jiang Y., Zhu H. Maillard reaction inducing amino acids degradation can adjust the flavour characteristic of black tea. Food Research International. 2025;201, Article 115685 doi: 10.1016/j.foodres.2025.115685. [DOI] [PubMed] [Google Scholar]
  25. Li M.F., Peng Y., Wen B.B., Liu Y., Liu J.J. Analysis of the morphological characteristics and anatomical structure of wild tea leaves in Zhenfeng. Southern Agricultural Science Journal. 2022;53(06):1674–1684. doi: 10.3969/j.issn.2095-1191.2022.06.021. [DOI] [Google Scholar]
  26. Li W., Chen Y., Tong S., Guo Y., Zhang Y., Ge M. Kinetic study of the gas-phase reaction of O 3 with three unsaturated alcohols. Journal of Environmental Sciences. 2018;71:292–299. doi: 10.1016/j.jes.2018.04.009. [DOI] [PubMed] [Google Scholar]
  27. Lin Q., Li J., Ling X., Zhang X. Cloning and expression of a novel trans - anethole oxygenase gene from Paraburkholderia sp. MR185. Journal of General and Applied Microbiology. 2022;68(3):163–167. doi: 10.2323/jgam.2021.10.001. [DOI] [PubMed] [Google Scholar]
  28. Liu X., Han Y., Luo L., Pan H., Cheng T., Zhang Q. Multiomics analysis reveals the mechanisms underlying the different floral colors and fragrances of Rosa hybrida cultivars. Plant Physiology and Biochemistry. 2023;195:101–113. doi: 10.1016/j.plaphy.2022.12.028. [DOI] [PubMed] [Google Scholar]
  29. Lu L., Wang L., Liu R., Zhang Y., Zheng X., Lu J., Wang X., Ye J. An efficient artificial intelligence algorithm for predicting the sensory quality of green and black teas based on the key chemical indices. Food Chemistry. 2024;441, Article 138341 doi: 10.1016/j.foodchem.2023.138341. [DOI] [PubMed] [Google Scholar]
  30. Lu X., Shi M., Liu L., Chen Z., Xu X., Feng G., Zeng M. Enhancement of flavor quality in oyster hydrolysate through fermentation with oyster -derived lactic acid bacteria. Food Bioscience. 2024;62, Article 105231 doi: 10.1016/j.fbio.2024.105231. [DOI] [Google Scholar]
  31. Luan X., Zhang X., Wei Z., Guo J., Obiero G.F.O., Getahun M.N.…Dong S. Odorant receptor 75 is essential for attractive response to plant volatile p-anisaldehyde in Western flower thrip. Pesticide Biochemistry and Physiology. 2025;211, Article 106421 doi: 10.1016/j.pestbp.2025.106421. [DOI] [PubMed] [Google Scholar]
  32. Lv S., Wu Y., Li C., Xu Y., Liu L., Meng Q. Comparative analysis of Pu-erh and Fuzhuan teas by fully automatic headspace solid-phase microextraction coupled with gas chromatography-mass spectrometry and chemometric methods. Journal of Agricultural and Food Chemistry. 2014;62(8):1810–1818. doi: 10.1021/jf405237u. [DOI] [PubMed] [Google Scholar]
  33. Maikhunthod B., Marriott P.J. Aroma-impact compounds in dried spice as a quality index using solid phase microextraction with olfactometry and comprehensive two-dimensional gas chromatography. Food Chemistry. 2013;141(4):4324–4332. doi: 10.1016/j.foodchem.2013.05.156. [DOI] [PubMed] [Google Scholar]
  34. Ni H., Jiang Q., Lin Q., Ma Q., Wang L., Weng S., Huang G., Li L., Chen F. Enzymatic hydrolysis and auto-isomerization during beta-glucosidase treatment improve the aroma of instant white tea infusion. Food Chemistry. 2021;342, Article 128565 doi: 10.1016/j.foodchem.2020.128565. [DOI] [PubMed] [Google Scholar]
  35. Oka Y., Omura M., Kataoka H., Touhara K. Olfactory receptor antagonism between odorants. EMBO Journal. 2004;23(1):120–126. doi: 10.1038/sj.emboj.7600032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Paggi J.M., Pandit A., Dror R.O. The art and science of molecular docking. Annual Review of Biochemistry. 2024;93(1):389–410. doi: 10.1146/annurev-biochem-030222-120000. [DOI] [PubMed] [Google Scholar]
  37. Pickering G.J., Karthik A., Inglis D., Sears M., Ker K. Determination of ortho- and retronasal detection thresholds for2-isopropyl-3-methoxypyrazine in wine. Journal of Food Science. 2007;72(7):S468–S472. doi: 10.1111/j.1750-3841.2007.00439.x. [DOI] [PubMed] [Google Scholar]
  38. Qi D., Shi Y., Lu M., Ma C., Dong C. Effect of withering/spreading on the physical and chemical properties of tea: A review. Comprehensive Reviews in Food Science and Food Safety. 2024;23(5) doi: 10.1111/1541-4337.70010. [DOI] [PubMed] [Google Scholar]
  39. Qu F., Li X., Wang P., Han Y., Wu Y., Hu J., Zhang X. Effect of thermal process on the key aroma components of green tea with chestnut-like aroma. Journal of the Science of Food and Agriculture. 2023;103(2):657–665. doi: 10.1002/jsfa.12177. [DOI] [PubMed] [Google Scholar]
  40. Ren D., Ren C., Ren J., Li S., Yang X., Li F. Changes in functional activities and volatile flavor compounds of fermented mung beans, cowpeas, and quinoa started with bacillus amyloliquefaciens SY07. Food Research International. 2025;201, Article 115636 doi: 10.1016/j.foodres.2024.115636. [DOI] [PubMed] [Google Scholar]
  41. Sala C., Mestres M., Marti M.P., Busto O., Guasch J. Headspace solid-phase microextraction analysis of 3-alkyl-2-methoxypyrazines in wines. Journal of Chromatography A. 2002;953(1–2):1–6. doi: 10.1016/s0021-9673(02)00123-1. [DOI] [PubMed] [Google Scholar]
  42. Selmat J., Rosli W.I.W., Russly A.R., Nordin L.M. Effect of roasting time and temperature on volatile component profile during nib roasting of cocoa beans (Theobroma cacao) Journal of the Science of Food and Agriculture. 1998 doi: 10.1002/(sici)1097-0010(199808)77:4&#x0003c;441::aid-jsfa46&#x0003e;3.0.co;2-#. [DOI] [Google Scholar]
  43. Shi J., Xie D., Qi D., Peng Q., Chen Z., Schreiner M., Lin Z., Baldermann S. Methyl Jasmonate-induced changes of flavor profiles during the processing of green, oolong, and black tea. Frontiers. Plant Science. 2019;10, Article 781 doi: 10.3389/fpls.2019.00781. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Shi Y., Wang M., Dong Z., Zhu Y., Shi J., Ma W., Lin Z., Lv H. Volatile components and key odorants of Chinese yellow tea (Camellia sinensis). Lwt-food. Science and Technology. 2021;146, Article 111512 doi: 10.1016/j.lwt.2021.111512. [DOI] [Google Scholar]
  45. Sun Z., Lin Y., Yang H., Zhao R., Zhu J., Wang F. Characterization of honey-like characteristic aroma compounds in Zunyi black tea and their molecular mechanisms of interaction with olfactory receptors using molecular docking. Lwt-food. Science and Technology. 2024;191, Article 115640 doi: 10.1016/j.lwt.2023.115640. [DOI] [Google Scholar]
  46. Tu Z., Li S., Tao M., He W., Shu Z., Wang S., Liu Z. Effect of shaking and piling processing on improving the aroma quality of green tea. Food Research International. 2025;201, Article 115624 doi: 10.1016/j.foodres.2024.115624. [DOI] [PubMed] [Google Scholar]
  47. Wang B., Qu F., Wang P., Zhao L., Wang Z., Han Y., Zhang X. Characterization analysis of flavor compounds in green teas at different drying temperature. Lwt-food. Science and Technology. 2022;161, Article 113394 doi: 10.1016/j.lwt.2022.113394. [DOI] [Google Scholar]
  48. Wang B., Zhang Y., Wei Y., Liao M., Cao H., Gao Q. Functional analysis of three odorant receptors in Plutella xylostella response to repellent activity of 2,3-dimethyl-6-(1-hydroxy)-pyrazine. Pesticide Biochemistry and Physiology. 2024;201, Article 105856 doi: 10.1016/j.pestbp.2024.105856. [DOI] [PubMed] [Google Scholar]
  49. Wang H., Hua J., Jiang Y., Yang Y., Wang J., Yuan H. Influence of fixation methods on the chestnut-like aroma of green tea and dynamics of key aroma substances. Food Research International. 2020;136, Article 109479 doi: 10.1016/j.foodres.2020.109479. [DOI] [PubMed] [Google Scholar]
  50. Wang M., Li J., Liu X., Liu C., Qian J., Yang J., Zhou X., Jia Y., Tang J., Zeng L. Characterization of key odorants in Lingtou Dancong oolong tea and their differences induced by environmental conditions from different altitudes. Metabolites. 2022;12(11):1063. doi: 10.3390/metabo12111063. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Wang T., Li R., Bo N., Guan Y., Yang D., Sha G., Chen Q., Liu S., Wang Z., Zhao M., Ma Y. Improvement of the floral aroma of ripened pu-erh tea via inoculation of Saccharomyces cerevisiae in industrial-level fermentation. Lwt-food. Science and Technology. 2025;223, Article 117776 doi: 10.1016/j.lwt.2025.117776. [DOI] [Google Scholar]
  52. Wang W., Feng Z., Min R., Yin J., Jiang H. The effect of temperature and humidity on yellow tea volatile compounds during yellowing process. Foods. 2024;13(20) doi: 10.3390/foods13203283. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Wang Y., Liu N., Yu T., Gao J., Fan Y., Wang W., Wang J., Wu Y., Zhang J., Ning J. The enhancement of flowery-like aroma in green tea under optimized processing conditions by sensory-directed flavor analysis. Food Chemistry-X. 2024;22, Article 101427 doi: 10.1016/j.fochx.2024.101427. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Wei J., Mu X., Wang S., Wei Q., Zhu L., Zhang X., Zhang J., Liu X., Wen B., Li M., Liu J. Integrated metabolome and transcriptome analysis provides insights into themechanisms of terpenoid biosynthesis in tea plants (Camellia sinensis) Food Research International. 2025;201, Article 115542 doi: 10.1016/j.foodres.2024.115542. [DOI] [PubMed] [Google Scholar]
  55. Wei Q., Liu X., Long S., Zhu L., Yang H., Song X., Peng Y., Chen T., Liu J., Wen B., Li M. Exploration of the substances and key processing steps related to the sweetness of Niangniang tea. Food Chemistry-X. 2025;28, Article 102556 doi: 10.1016/j.fochx.2025.102556. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Wei Y., Zhang J., Li T., Zhao M., Song Z., Wang Y., Ning J. GC-MS, GC-O, and sensomics analysis reveals the key odorants underlying the improvement of yellow tea aroma after optimized yellowing. Food Chemistry. 2024;431, Article 137139 doi: 10.1016/j.foodchem.2023.137139. [DOI] [PubMed] [Google Scholar]
  57. Wen S., Sun L., Zhang S., Chen Z., Chen R., Li Z., Lai X., Zhang Z., Cao J., Li Q., Sun S., Lai Z., Li Q. The formation mechanism of aroma quality of green and yellow teas based on GC-MS/MS metabolomics. Food Research International. 2023;172, Article 113137 doi: 10.1016/j.foodres.2023.113137. [DOI] [PubMed] [Google Scholar]
  58. Xiao Z., Shen T., Niu Y., Ke Q., Sun Z., Yang E., Liu X., Zhang J., Zhu J. Unraveling the characteristic aroma compounds in Longjing tea and their interaction mechanisms with S-curve and broad-spectrum olfactory receptors using molecular docking. Food Bioscience. 2024;60, Article 104423 doi: 10.1016/j.fbio.2024.104423. [DOI] [Google Scholar]
  59. Yan H., Lin Z., Li W., Gao J., Li P., Chen Q., Lv H., Zhang Y., Dai W., Lin Z., Zhu Y. Unraveling the enantiomeric distribution of Glycosidically bound linalool in teas (Camellia sinensis) and their acidolysis characteristics and pyrolysis mechanism. Journal of Agricultural and Food Chemistry. 2024;72(16):9337–9350. doi: 10.1021/acs.jafc.4c00037. [DOI] [PubMed] [Google Scholar]
  60. Yang Y., Chen J., Zheng F., Lin B., Wu F., Verma K.K., Chen G. Assessment of characteristic flavor and taste quality of sugarcane wine fermented with different cultivars of sugarcane. Fermentation-Basel. 2024;10(12), Article 628 doi: 10.3390/fermentation10120628. [DOI] [Google Scholar]
  61. Yang Y., Wang Q., Xie J., Deng Y., Zhu J., Xie Z., Yuan H., Jiang Y. Uncovering the dynamic alterations of volatile components in sweet and floral aroma black tea during processing. Foods. 2024;13(5), Article 728 doi: 10.3390/foods13050728. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Yang Z., Baldermann S., Watanabe N. Recent studies of the volatile compounds in tea. Food Research International. 2013;53(2):585–599. doi: 10.1016/j.foodres.2013.02.011. [DOI] [Google Scholar]
  63. Yin P., Kong Y., Liu P., Wang J., Zhu Y., Wang G., Sun M., Chen Y., Guo G., Liu Z. A critical review of key odorants in green tea : Identification and biochemical formation pathway. Trends in Food Science & Technology. 2022;129:221–232. doi: 10.1016/j.tifs.2022.09.013. [DOI] [Google Scholar]
  64. Yin X., Wei Y., Li T., Zhang J., Zou L., Cui Q., Lu C., Ning J. Heterocyclic compounds formation in large-leaf yellow tea induced by the Maillard reaction at different roasting temperatures. Lwt-food. Science and Technology. 2023;182, Article 1114856 doi: 10.1016/j.lwt.2023.114856. [DOI] [Google Scholar]
  65. Yoshida T., Xia Z., Takeda K., Katsuta T., Sugimoto K., Funaoka M. Peroxidase-catalyzed polymerization and copolymerization of lignin-based macromonomer (lignocresol) having high content of p-cresol and thermal properties of the resulting polymers. Polymers for Advanced Technologies. 2005;16(11−12):783–788. doi: 10.1002/pat.621. [DOI] [Google Scholar]
  66. Zeng L., Fu Y., Huang J., Wang J., Jin S., Yin J., Xu Y. 2022. Comparative analysis of volatile compounds in Tieguanyin with different types based on HS-SPME-GC-MS. foods, 11(11), article 1530. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Zeng S., Zhang L., Li P., Pu D., Fu Y., Zheng R., Xi H., Qiao K., Wang D., Sun B., Sun S., Zhang Y. Molecular mechanisms of caramel-like odorant-olfactory receptor interactions based on a computational chemistry approach. Food Research International. 2023;171, Article 113063 doi: 10.1016/j.foodres.2023.113063. [DOI] [PubMed] [Google Scholar]
  68. Zhang Y., Gu M., Yang S., Fan W., Lin H., Jin S., Wang P., Ye N. Dynamic aroma characteristics of jasmine tea scented with single-petal jasmine "Bijian": A comparative study with traditional double-petal jasmine. Food Chemistry. 2025;464, Article 141735 doi: 10.1016/j.foodchem.2024.141735. [DOI] [PubMed] [Google Scholar]
  69. Zhu Y., Dong J., Jin J., Liu J., Zheng X., Lu J., Liang Y., Ye J. Roasting process shaping the chemical profile of roasted green tea and the association with aroma features. Food Chemistry. 2021;353, Article 129428 doi: 10.1016/j.foodchem.2021.129428. [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: Fig. S1. Total ion current diagram of mixed sample mass spectrometry analysis. Fig. S2. Cluster tree diagram of three tea samples. GT: green tea. YT: yellow tea. NNT: Niangniang tea. Fig. S3. PLS-DA cross-validation findings utilizing a permutation test. R²: Coefficient of Determination. Q²: Predictive Ability.

mmc1.docx (692.2KB, docx)

Supplementary material 2

mmc2.xlsx (51.7KB, xlsx)

Data Availability Statement

Data will be made available on request.

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

RESOURCES