An investigation into the key volatile compounds that dominate the characteristic aroma and flavor of pomelo flower–green tea

Abstract

Pomelo flowers (PFs) enhance the flavor diversity of green tea (GT) and reduce resource wastage, but the aroma-enhancing mechanism was unclear. In this study, electronic-nose and electronic-tongue analyses demonstrated that pomelo flower-green tea (PT) had a richer aroma and taste than GT. Gas chromatography-mass spectrometry analysis revealed an increased total volatile content and enhanced floral notes in PT. Specifically, 23 compounds were upregulated and 10 were downregulated in PT. Major odorants (variable importance in projection >1) included linalool, nerolidol, methyl anthranilate, and indole. PT also showed elevated free amino acids, particularly theanine and glutamic acid, which intensified its umami and aftertaste. Seven volatile compounds that migrated from PFs to GT (compounds not originally present in GT) interacted with umami taste receptors via hydrogen bonding and hydrophobic interactions, thereby shaping the umami profile of PT. This study clarifies the flavor formation of PT and supports the development of scented tea.

Highlights

• •Pomelo flower could improve the aroma and taste of green tea.
• •Pomelo flower-green tea enhanced the umami by significantly increasing amino acids.
• •Linalool, nerolidol, methyl anthranilate and indole were the main volatiles of PT.

1. Introduction

Originating in China, tea is now enjoyed worldwide and is classified into various types based on color and processing methods. Green tea (GT) is a non-fermented tea produced by directly drying the leaves, thereby preserving its natural flavor being rich in phenolic compounds, primarily catechins. GT is known for its wide range of health benefits, including the prevention of cardiovascular diseases, cancer, and diabetes (Cheng et al., 2020; Tuo et al., 2024; Wan et al., 2024; Wong et al., 2022). However, despite being the most widely consumed type of tea, GT faces challenges such as limited product variety, flavor convergence, and a lack of taste differentiation. Although there are over 20 common types of GT, most products are based on traditional flavor profiles, leading to homogeneity in the market. Flower-scented teas can address this limitation by introducing floral aroma compounds through scenting, which synergize with the polyphenolic components of GT, creating a composite flower-tea aroma system that significantly enhances product distinctiveness.

Flower-scented teas are a popular extension of GT, in which finely processed tea leaves are scented with fragrant flowers to enhance flavor and introduce unique aroma profiles (Wang et al., 2020). As scent is a key aspect in tea tasting, flower-scented teas offer not only the pure taste of tea but also a fresh floral aroma (Yin et al., 2015). For example, jasmine tea contains up to 10 volatile compounds that are absent in GT alone. These compounds, including (Z)-3-hexenol, ethyl benzoate, and terpinene, contribute to its distinctive aroma (An et al., 2023; Shen et al., 2017). Similarly, citrus-flower green tea has a unique citrus flavor derived from volatiles such as β-pinene, (E)-β-ocimene, and γ-terpinene (Yang et al., 2024). Other flower-scented teas, such as rose, osmanthus, and lavender teas, introduce floral or herbal notes, enriching the tea-drinking experience (Wang et al., 2020).

After aroma, taste and aftertaste are the next most important aspects in tea evaluation. Flower-scented teas not only enhance aroma but also diversify the flavor profile of the tea infusion. Studies have shown that enzymatic reactions occurring during the scenting process can alter the chemical compounds in tea, including polyphenols, theophylline, and amino acids, resulting in a fresher and sweeter taste (Chen et al., 2017, 2023). Human perception of flavor is a complex process involving cross-modal interactions between taste and odor, which together influence the overall sensory experience (Eggink et al., 2012; Pu et al., 2024). Osmanthus flowers, which are rich in (E)-β-ionone, geraniol, linalool, and γ-decalactone, can enhance the sweetness of GT. Moreover, the addition of specific aroma compounds such as 1-octen-3-ol, phenylacetaldehyde, and β-ionone can amplify the sweetness and umaminess of GT (Deng et al., 2024; Wei et al., 2023). Together, these aromatic and taste-active compounds contribute to the characteristic flavor of flower-scented teas.

Pomelo (Citrus grandis L. Osbeck), a citrus species in the Rutaceae family, is native to Asia and is classified as a subtropical evergreen tree. Pomelo flowers (PFs) possess notable pharmacological properties and nutritional benefits (Moslemi et al., 2019). Pomelo trees produce a large number of flowers, approximately 60–80 % of which are either naturally shed or artificially thinned and discarded, resulting in a waste of resources (Ambrožič Turk et al., 2014; Cheng et al., 2024). As a means to enhance the utilization of PFs, in addition to their use in the production of PF essential oil, pomelo flower–green tea (PT) has gradually gained popularity in the market. It combines the rich aroma of PFs with the adsorptive properties of GT, resulting in a product that not only imparts a floral aroma to GT but also enriches its flavor and significantly increases the added value of PFs. However, while the volatile compounds in PFs have been investigated, little is known about the aroma profile and key odorants in PT.

In this study, PT was prepared using PFs and GT, and the effects of PF addition on the aroma and taste of PT were analyzed using electronic nose, electronic tongue, and gas chromatography–mass spectrometry (GC–MS) analyses. Additionally, molecular docking was employed to explore the interactions between aroma components and umami taste receptors in order to reveal the potential mechanism contributing to the formation of the PT flavor profile. This study contributes to an understanding of the formation and retention of characteristic flavor components in PT and provides a scientific basis for guiding the development of PT-based products.

2. Materials and methods

2.1. Raw materials and reagents

PFs were collected in early April 2024 from the germplasm resource nursery of the Citrus Research Institute at Southwest University. After collection, the flowers were immediately stored at −80 °C for subsequent scenting experiments. The GT used in this study was produced from tea leaves harvested from a plantation in Dazu County, Chongqing, China. The process of scenting GT leaves with PFs (Fig. 1) followed the patented method developed by Wang Jun (China Invention Patent Number: ZL202110841759.2). Fresh PFs were mixed with GT leaves, which had an initial moisture content of 6 %, at a 1:2 weight ratio, and the mixture was subjected to two rounds of scenting, with each lasting 12 h at temperatures between 35 °C and 50 °C. Although increasing the amount of PFs enhanced the flavor of PT, excessive PF addition resulted in an undesirable greasy mouthfeel. After the scenting process, the PFs were removed, and the scented GT leaves were subjected to two steps of drying. The first drying was performed at 85 °C for 15 min to adjust the moisture content of the tea to 6–8 %. The second drying was conducted at 80 °C for 10 min to stabilize the final moisture at approximately 8 %. The final product, the dried PT mixture, was packed in a waterproof and moisture-proof bag, sealed, and stored at 4 °C for future use. Unused PFs were immediately frozen in liquid nitrogen and transported to the laboratory using dry ice. Upon arrival, they were stored at −80 °C for future use.

Fig. 1.

Production of PT.

2.2. Sensory evaluation

Sensory evaluation was conducted by a panel of 10 professionals (5 females and 5 males). The sensory evaluation was performed according to China National Standard GB/T 23776-2018. First, the appearance of the tea was evaluated. Then, 3 g of each tea sample was weighed and placed in a tea cup, followed by the addition of 150 mL of boiling water (100 °C). The sample was brewed for 4 min, and the tea infusion was then filtered, leaving the tea leaves at the bottom of the cup. Evaluation was subsequently conducted in the following order: infusion color, infusion aroma, infusion taste, and leaf residue conditions. The evaluation criteria are illustrated in Table S1. The overall sensory score was calculated by multiplying the score for each evaluation indicator by its corresponding weighting factor.

2.3. Electronic nose analysis

For electronic nose analysis, 4 g of a tea sample was brewed with 150 mL of boiling water and steeped for 4 min. The resulting infusion was filtered through three layers of gauze. After cooling to room temperature, 5 g of the filtrate was transferred to a 100 mL beaker. The beaker was sealed with a double layer of preservative film and left undisturbed for 1 h to generate headspace gas. Headspace analysis was conducted using a PEN3 portable electronic nose (WNA Airsense Analysentechnik GmbH, Germany), which consists of 10 metal oxide semiconductor sensors. The performance of each sensor was provided in Table S2. The headspace gas of each sample was collected for 80 s at a constant flow rate of 400 mL/min. Data were recorded during the stable sensor response window between the 75th and 77th seconds, and the value at the 76th second was extracted for further analysis.

2.4. Electronic tongue analysis

Tea taste characteristics were assessed using a TS-5000Z Taste-Sensing System (Intelligent Sensor Technology Co., Ltd., Japan). The system was equipped with a sensor array consisting of a reference electrode and six electrodes for detecting umaminess (AAE), saltiness (CT0), sourness (CA0), bitterness (C00), astringency (AE1), and sweetness (GL1). Sample preparation followed the same procedure used for electronic nose analysis.

2.5. Volatile compound analysis

2.5.1. Headspace solid-phase microextraction (HS-SPME)

For volatile compound extraction, 0.5 g of a tea sample was infused with 5 mL of boiling water and transferred into a 20 mL headspace injection vial. After the solution cooled to room temperature, 1.5 μL of a 14.25 μg/mL cyclohexanone solution (used as an internal standard) was added, and the vial was sealed. Volatile compounds were extracted via HS-SPME using a 120 μm DVB/CAR-WR/PDMS fiber (Agilent Technologies, USA). The sample was equilibrated at 60 °C for 10 min and then adsorbed onto the fiber for 50 min. The desorption of volatiles from the fiber was performed at 225 °C for 5 min.

2.5.2. GC–MS conditions

Gas chromatography-mass spectrometry (GC–MS) analysis was performed using an Agilent 8890-7000E system equipped with an Agilent DB-WAX capillary column (30 m × 0.25 mm, 0.25 μm). The inlet temperature was set to 225 °C, and helium was used as the carrier gas at a flow rate of 1 mL/min. The oven temperature program was as follows: 1) maintained at 40 °C for 2 min, increased to 85 °C at 3 °C/min and maintained for 2 min; 2) increased to 110 °C at 2 °C/min and maintained for 2 min; 3) increased to 160 °C at 5 °C/min and maintained for 1 min; 4) increased to 225 °C at 5 °C/min and maintained for 5 min; 5) increased to 230 °C at 10 °C/min and maintained for 8 min. MS analysis was conducted using electron ionization at 70 eV with a scan range of m/z 50–650.

2.5.3. Qualitative and quantitative analyses of volatile compounds

Retention indices (RIs) were determined using C₅–C₂₅ n-alkanes (Shanghai Anpu Resplendent Standard Technical Service Co., Ltd., China). Volatile compounds were tentatively identified by comparing their RIs and mass spectra with data from the NIST 20 library. Quantification of identified volatile compounds was performed using the internal standard method.

2.5.4. Calculating ROAVs

Relative odor activity values (ROAVs) are a method used to identify key volatile compounds in food by integrating their sensory thresholds, which helps to determine the contribution of individual volatile compounds to the overall aroma profile of food. In order to evaluate the effect of different volatile compounds on PT, the ROAVs were calculated by using the compounds' sensory thresholds (Wang et al., 2022). The calculation formula is as follows: ${\mathit{ROAV}}_i\approx \frac{C_i}{T_i}\times \frac{T_{\mathit{max}}}{C_{\mathit{max}}}\times 100$, where $C_i$ represents the relative concentration (μg/kg) of VOCs, and $T_i$ represents the threshold concentration (μg/kg) of the compound in water. $C_{\mathit{max}}$ and $T_{\mathit{max}}$ represent the relative content and threshold of the compounds that contribute most to the overall flavor.

2.6. Qualitative and quantitative analyses of amino acids

Amino acid analysis was performed using an L-8900 automatic amino acid analyzer (Hitachi, Japan). A tea sample (0.1 g) was homogenized in a 10 % sulfosalicylic acid solution and centrifuged at 15,000×g for 30 min. The supernatant was filtered through a 0.22 μm membrane and analyzed for amino acids. Standards for serine (Ser), aspartic acid (Asp), glutamic acid (Glu), valine (Val), isoleucine (Ile), leucine (Leu), tyrosine (Tyr), phenylalanine (Phe), histidine (His), cysteine (Cys), arginine (Arg), alanine (Ala), glycine (Gly), threonine (Thr), methionine (Met), and lysine (Lys) were injected at a volume of 20 μL.

Theanine content was determined by high-performance liquid chromatography (HPLC) using an LC20AD system (Shimadzu, Japan) equipped with a C18 (ODS-C18) column (4.6 mm × 250 mm, 5 μm) and a UV detector set at 210 nm. Approximately 0.2 g of a tea sample was weighed into a 15 mL centrifuge tube, to which 15 mL of boiling ultrapure water was added. The mixture was shaken thoroughly and then incubated in a thermostatic water bath at 100 °C for 30 min. After cooling to room temperature, the infusion was centrifuged at 4000 rpm for 10 min. The supernatant was transferred to a new centrifuge tube, diluted with 20 mL of ultrapure water, mixed well, and filtered through a 0.45 μm membrane prior to HPLC analysis. The mobile phases consisted of water (A) and acetonitrile (B). The injection volume was 10 μL. Gradient elution was performed as follows: from 0 to 10 min, the mobile phase composition gradually changed from 0 % to 80 % B; from 10 to 22 min, it returned from 80 % to 0 % B. Quantification was performed through peak area comparisons between the sample and amino acid standards.

The taste activity value (TAV) was calculated as the ratio of its concentration determined in tea to its threshold value. The compounds were considered active in food taste if their TAV was greater than 1 (Wang et al., 2020).

2.7. Molecular docking between umami taste receptors and aroma compounds

The molecular docking of seven volatile compounds and two umami taste receptors (TAS1R1 (Uniport ID: Q8TE23) and TAS1R3 (Uniport ID: Q7RTX0)) was conducted according to the method described by Deng et al. (2024). The three-dimensional structures of β-ocimene (CID: 18756), β-pinene (CID: 14896), farnesol (CID: 3327), methyl anthranilate (CID: 8635), nerol (CID: 643820), (E)-β-farnesene (CID: 5281517), and nerolidol (CID: 5284507) were obtained from the Pubchem database (https://pubchem.ncbi.nlm.nih.gov/). Molecular docking results were analyzed and visualized using Discovery Studio 2019 software (Chuangteng Technology Co., Ltd., China).

2.8. Statistical analysis

Each experiment was conducted in triplicate. Data were analyzed for statistical significance (p < 0.05) using Excel 2016. Multivariate analysis and graphical visualization were performed using SIMCA-P 14.1 and Origin 9.0, respectively.

3. Results and discussion

3.1. Sensory evaluation of PT

Sensory evaluation was employed to analyze the taste and appearance of green tea (GT) and pomelo flower green tea (PT), as illustrated in Fig. 2A and B. The sensory scores for GT and PT were recorded at 86.55 and 92.00, respectively. In terms of appearance and liquor color, GT received a marginally higher score than PT. The addition of pomelo flowers (PF) did not result in visually discernible changes to the physical morphology of the tea leaves. Both GT and PT infusions exhibited yellowish-green and vibrant soup colors; however, the GT infusion was characterized by a more intense hue. We hypothesize that this difference is attributable to the oxidation of phenolic compounds facilitated by polyphenol oxidase (PPO) in GT, which enhances color depth (Thanaraj & Seshadri, 1990). Conversely, PT, which underwent two rounds of scenting and re-firing, likely experienced diminished PPO activity, thereby limiting phenolic oxidation and resulting in a lighter color. Notably, PT exhibited significantly higher sensory scores in aroma and taste compared to GT. The sweet floral notes of pomelo flowers infused into the green tea synergize with its inherent freshness, creating a distinctive aroma profile characterized by mellowness and freshness.

Fig. 2.

Comparisons of appearance, aroma, and taste between GT and PT. (A) Visual appearances of dried tea leaves, tea infusions, and spent tea leaves. (B) Sensory evaluation. (B) Radar charts and (D) a PCA biplot of electronic nose analysis results. (E) Radar charts and (F) a PCA biplot of electronic tongue analysis results.

The radar charts based on the electronic nose analysis results of PT and GT (Fig. 2C) displayed strong responses at sensors R2, R4, R6, R7, and R9, indicating that nitrogen oxides, methane, sulfides, and aromatic compounds were the predominant volatile compounds in both teas. Previous studies have identified nitrogen oxides, sulfides, and aromatic compounds as the key volatile compounds in GT (Yu et al., 2008; Zou et al., 2018). The main differences between GT and PT were observed in the response values of sensors R2, R7, and R9 (Table S3), with PT showing higher response values. Principal component analysis (PCA) was conducted to further investigate these differences (Fig. 2D). The first (PC1) and second (PC2) principal components explained 73.4 % and 23.2 % of the total variance, respectively, accounting for 96.6 % of the variation. The clear separation between PT and GT in the PCA biplot indicated a marked difference in their odor composition.

The radar charts based on the electronic tongue analysis results of PT and GT (Fig. 2E) indicated that both PT and GT infusions exhibited a combination of bitterness, sweetness, umaminess, and saltiness, along with a lingering aftertaste characterized by umaminess, bitterness, and astringency. These sensory characteristics are typical of the basic flavor profile of GT (Zou et al., 2018). After the addition of PFs, PT showed a decrease in sweetness but increases in other taste attributes, with significant differences observed for umaminess, sweetness, aftertaste-A (aftertaste-astringency), and saltiness (p < 0.05) (Table·S4).

The sweetness reduction perceived by the electronic tongue was attributed to the degradation of sugars or sugar alcohols during the scenting process, potentially via the Maillard reaction. A similar observation was reported by Deng et al. (2024), who found that the addition of osmanthus flowers to tea significantly reduces the level of soluble sugars and significantly increases the levels of arginine and catechin, both of which contribute to bitterness. Therefore, the increased bitterness in PT could be due to the accumulation of bitter compounds. In reality, however, human taste perception is more complex than taste detection using the electronic tongue, as it involves synergistic or inhibitory interactions between the olfactory and gustatory senses. For example, certain odorants can enhance the perception of saltiness or sweetness. It has been reported that the addition of geraniol and β-ionone can enhance the sweetness of tea infusions (Yu et al., 2021). Thus, while the electronic tongue can quantify basic taste attributes, it should be used in conjunction with sensory evaluation to identify synergistic or antagonistic flavor effects and subtle mouthfeel characteristics for a more comprehensive understanding of tea flavor. Together with the results of sensory evaluation, the findings from electronic nose and electronic tongue analyses suggested that the addition of PFs enhanced the complexity of the flavor profile of GT.

The PCA biplot of electronic tongue analysis results (Fig. 2F) showed that PC1 and PC2 contributed 72.78 % and 16.89 % of the total variance, respectively, cumulatively explaining 89.67 % of the total variation. This indicated that PC1 and PC2 effectively captured the majority of the flavor differences between GT and PT. The PT and GT samples were projected on opposite sides of the PC1 axis, suggesting significant differences in taste between the two teas. PT was positioned closer to the sensor responses for aftertaste-A, umaminess, and saltiness, while GT was more strongly associated with sweetness. These results suggest that PT exhibits enhanced umaminess and aftertaste-A, along with increased saltiness. It is well documented that umaminess and saltiness are closely related and can synergistically enhance each other's perception (Huang et al., 2024; Li et al., 2024).

In conclusion, the scenting process had a more pronounced impact on the aroma and taste of tea than on its appearance, consistent with previous findings (Chen et al., 2023). The addition of PFs to GT not only modified its aroma profile but also enhanced certain taste attributes, such as umaminess and aftertaste, while reducing sweetness.

3.2. Qualitative and quantitative analyses of volatile compounds

3.2.1. Qualitative analysis of volatile compounds

To further investigate the impact of PF addition on the aroma profile of GT, GC–MS was employed to qualitatively and quantitatively analyze the volatile compounds in PFs, GT, and PT, and total of 77 volatile flavor compounds were detected in PF, GT and PT infusion. However, given the emphasis on the aroma profiles of GT and PT, the analysis focused on the tea infusions. A total of 33 volatile compounds were detected in GT and PT infusions, including 13 terpenoids, 9 alcohols, 3 esters, 3 heterocyclic compounds, 2 aldehydes, 2 ketones, and 1 other compound (Fig. 3A). The volatile profile of PT was dominated by terpenoids and alcohols, followed by ketones and heterocyclic compounds (Fig. 3B). These results are consistent with findings from studies on osmanthus tea, in which alcohols and ketones were also identified as the primary volatile compounds (Meng, 2024). PCA revealed distinct differences in the volatile compound profiles of GT, PT, and PFs (Fig. 3C), demonstrating that PF addition significantly altered the aroma of GT.

Fig. 3.

Volatile compound profiles of PFs, GT, and PT. (A) Categories and quantities of volatile compounds. (B) The number of volatile compounds identified in each sample. (C) PCA of these volatile compound profiles. (D) The average relative content of each volatile compound in each category. (E) A heatmap of 33 differential volatile compounds between GT and PT.

3.2.2. Quantitative analysis of volatile compounds

The volatile compounds in PFs, GT, and PT were determined semi-quantitatively using cyclohexanone as an internal standard. The concentrations of all identified compounds are shown in Table S5 and Fig. 3D. The total relative content of terpenoids in PT was 203.02 μg/L, significantly higher than the 109.19 μg/L found in GT, suggesting that PF addition significantly increased terpenoid levels. PFs themselves contained an exceptionally high concentration of terpenoids, several times higher than that found in GT. Key terpenoids, such as β-ocimene, β-pinene, and (E)-β-farnesene, were notably elevated in PT (Fig. 3E), indicating their introduction through the scenting process. Previous studies have shown that volatile terpenoids contribute to the sweet and floral aromas of tea (Sheibani et al., 2016; Zhu et al., 2017), playing a vital role in enhancing tea quality. Thus, the elevated terpenoid content in PT likely contributed to its richer aroma.

In addition to an elevated level of terpenoids, a significant accumulation of alcohols was observed in PT. The total alcohol content in PT was 2271.85 μg/L, approximately three times higher than that in GT. Meanwhile, the total alcohol content in PFs was 2191.51 μg/L. The elevated alcohol content in PT could be attributed to the additive effect of aroma substances, whereby GT readily adsorbed and retained alcohol compounds from PFs during the scenting process. GT is a good porous solid adsorbent with an intricate microporous internal structure; this structural characteristic enhances its adsorption properties (Dai et al., 2024; Wang, Huang, et al., 2024). The temperature of the scenting process plays a critical role in balancing the volatilization and adsorption of alcohols. Due to their hydroxyl groups, alcohols readily form molecular interactions, such as hydrogen bonds, with tea polyphenols (Liu, Ran, et al., 2023), enhancing alcohol retention. Notably, in jasmine tea, the highest retention of linalool also occurs through adsorption (An et al., 2022).

GT showed a strong capacity to adsorb heterocyclic compounds, esters, and ketones, resulting in their concentrations being 4.7–10.7 times higher in PT than in GT. Ketones generally impart floral, fruity, and creamy aromas, while heterocyclic compounds are typically associated with nutty, caramel, and sweet aromas (Scalone et al., 2019). These findings suggested that the scenting process not only increased the concentrations of terpenoids and alcohols but also enhanced other aroma compounds, enriching the flavor profile of PT.

3.2.3. Formation of volatile compounds during the scenting process of PT

The addition of PFs to GT led to significant changes in the volatile compound profile of the resulting PT. These changes could be classified into three main categories: (1) flavor-enhancing, (2) flavor-reducing, and (3) flavor-balancing effects.

Flavor-enhancing effects: The addition of PFs led to the elevated concentrations of 23 volatile compounds in PT, including linalool, β-ocimene, farnesol, β-pinene, nerol, nerolidol, methyl anthranilate, and indole. Notably, linalool was significantly enriched in PT, with a total concentration of 1597.05 μg/L, approximately equal to the combined linalool concentrations in GT and PFs. A similar trend has also been observed in other citrus-scented teas, where the increase in linalool results from the additive effect of both citrus-derived and tea-derived linalool (Wang et al., 2020). Additionally, aroma compounds such as nerol (floral and green notes), nerolidol (floral, apple-like, and green notes), and methyl anthranilate (honey-like and floral notes), which were absent in GT, were detected in PT (Table S5), suggesting that PF addition significantly enhanced the aroma profile of GT.

Flavor-reducing effects: During the production of PT, the adsorption effect of the base tea increases the content of certain volatile compounds. However, the content of some other substances decreases. Ten substances were found to have decreased in PT compared to GT, including (E)-β-ocimene, DL-limonene, (Z)-ocimene, 1-nonanal, benzaldehyde, (E)-linalool oxide, linalool oxide C, methyl salicylate, 3,5-octadiene-2-one, dimethyl sulfide. Among these substances, DL-limonene, (Z)-ocimene and benzaldehyde accumulate in both PF and GT, but their concentrations decreased in PT after scenting. This suggests that these compounds in PF were not adsorbed by GT, and that some compounds in GT evaporated, degraded, or underwent transformation. Similar phenomena occurred in the processing of other scented teas, but the types and concentrations of substances depend on the raw materials and processing methods (Chen et al., 2017; Hou et al., 2024). These changes in compounds may involve chemical reactions in addition to evaporation. Benzaldehyde can be reduced through oxidation upon exposure to oxygen or via condensation with amino acids (Li et al., 2014). Reduction of this substance has been reported during the scenting process of jasmine, osmanthus, and michelia teas (Chen et al., 2017; Hou et al., 2024; Wang, Deng, et al., 2024). Methyl salicylate levels generally decrease during tea leaf processing, from the fresh leaf stage to the final product (Ao et al., 2024). It has been reported that methyl salicylate can be converted into salicylic acid via enzymatic reactions (Cao et al., 2019; Deng et al., 2017), which may be one of the pathways contributing to its lower content. However, the migration and changes of other substances during the scenting processing remain unclear and require further research and validation.

Flavor-balancing effects: The flavor-balancing effects of PT refer to the fact that some volatile compounds in the PFs were difficult to be adsorbed by the green tea leaves. This resulted in either a smaller increase or zero content of these volatile compounds in the PT. Therefore, these aroma compounds have a relatively small impact on the flavor. PF contained 40 volatile compounds, including β-caryophyllene, geraniol D, α-selinene, and (E)-citral, all of which were present in higher concentrations and none of which were found in GT or PT. Additionally, six substances, including β-myrcene, terpinene, γ-terpinene, terpinolene, phenethyl alcohol, and α-terpineol, exhibited a slight increase in PT compared to GT, though this increase was insignificant. This suggests that these compounds were not adsorbed by GT, were desorbed after adsorption, or underwent chemical reactions and transformed. A similar trend has been observed in the scenting of osmanthus tea, where more than 10 volatile compounds (including 3-methyl-2-butenal, α-terpineol, and γ-terpinene) do not migrate into OSGT (Wang, Deng, et al., 2024). Previous studies have shown that the adsorption of floral aromas by green tea involves physical adsorption via van der Waals forces, as well as chemical adsorption, throughout the scenting process (Wang, Huang, et al., 2024). Furthermore, studies have shown that tea processing encompasses a series of complex and intense thermochemical reactions, including decomposition, Maillard reactions, redox reactions, and isomerization (Wang, Chen, et al., 2024).

While this study sheds light on the characteristic aromatic components of yuzu flower green tea and their potential migration patterns, the changes to these components and how they interact with the green tea surface remain unclear. Further research is needed to clarify the changes in volatile substances during the scenting process and how they are adsorbed by green tea.

3.3. Transformation of key volatile compounds in PT

To investigate the contribution of specific volatile compounds to the flavor formation in PT, we conducted orthogonal partial least squares discriminant analysis (OPLS-DA) on the volatile compound profiles of GT and PT. These two sample groups were clearly distinguished (Fig. 4A), indicating the reliability of the model. A permutation test with 200 iterations confirmed the absence of overfitting (Fig. 4B). The model parameters exhibited a high explanatory variance (R²X = 0.987, R²Y = 0.998) and strong predictive power (Q² = 0.993). The R2 and Q2 values were 0.311 and − 0.574, respectively, suggesting that the OPLS-DA model was reliable and predictable. The analysis focused on four key volatile compounds that were significantly elevated in PT: linalool, nerolidol, methyl anthranilate, and indole. These compounds were selected based on the following criteria: p < 0.05 and Variable Importance in Projection (VIP) ≥ 1.0. Notably, they were nearly absent in GT but accumulated substantially in PT (Fig. 4C). Specifically, linalool was 2.7 times more concentrated in PT than in GT. The concentrations of these compounds in PFs were comparable to those in PT, ranging from 0.9 to 1.4 times their concentrations in PT, suggesting that they migrated from PFs to GT during the scenting process.

Fig. 4.

OPLS-DA of volatile compounds: (A) an OPLS-DA score plot, (B) results of a permutation test with 200 iterations, and (C) VIP values of differential compounds.

The contribution of volatile compounds to the overall aroma of PT and GT were further evaluated based on their ROAVs. The ROAVs data for the 32 volatile compounds identified among the 33 volatile compounds were shown in Table 1 (with no references to the odor threshold for β-bisabolene). Compounds with an ROAV >1 were considered to be the critical volatile aroma compounds of the sample. Those with an ROAV between 0.1 and 1 were considered to have a modifying effect on the flavor of the sample (Ren et al., 2025). There were 10 compounds in PT with ROAV >1, and Methyl anthranilate, geraniol, linalool, and nerol had ROAV >10. This indicated that these compounds made a major contribution to the overall flavor of PT. Ten substances in GT had ROAV >1. Eight of these were the volatile compounds (E)-β-Ocimene, β-Myrcene, (Z)-β-Ocimene, 3,5-octadiene-2-one, limonene, and 1-nonanal, Methyl salicylate and Dimethyl sulfide which differed from those in PT and contributed to a different flavor profile. (See Table 1.)

Table 1.

ROAVs of volatile compounds in GT and PT.

Volatile compounds	Odor threshold (μg/kg)	ROAV
		GT	PT
(E)-β-Ocimene	34	2.07
β-Pinene	140	–	0.02
β-Myrcene	15	2.74	0.43
α-Terpinene	85	0.13	0.04
Limonene	34	10.90	0.79
(Z)-β-Ocimene	55	2.58	–
γ-Terpinene	1000	0.03	0.01
β-Ocimene	34	–	0.55
Terpinolen	200	0.30	0.06
(E)-β-Farnesene	87	–	0.47
(Z,E)-α-Farnesene	1	–	14.85
α-Farnesene	87	–	0.37
Geraniol	1.1	49.42	49.49
Benzyl alcohol	100	–	0.02
Phenethyl alcohol	45	0.52	0.08
Citronellol	4	–	1.03
Nerolidol	250	–	1.83
Nerol	1.1	–	11.93
Farnesol	20	–	1.18
α-Terpineol	1200	0.05	0.01
Linalool	50	77.67	27.04
1-Nonanal	1.1	21.96	–
Benzaldehyde	300	0.23	0.01
(E)-Linalool oxide	100	0.72	0.05
Indole	40	0.58	3.67
Linalool oxide C	30	0.53	–
Methyl geranate	7	–	1.23
Methyl anthranilate	3	–	100.00
Methyl salicylate	40	4.37	0.14
3,5-Octadiene-2-one	0.15	100.00	–
6,10-Dimethyl-5,9-undecadien-2-one	60	–	0.15
Dimethyl sulfide	12	2.78	0.05

Odor threshold: the data from compilations of odor threshold values in air, water, and other media (second enlarged and revised edition). Beijing: Science Press. 2018.

Table 2.

Binding energies of selected volatile compounds and the umami taste receptors TAS1R1 and TAS1R3.

Ligands	Binding Energy (kcal/mol)
	TAS1R1	TAS1R3
β-Ocimene	−6.29	−5.65
β-Pinene	−6.13	−6.10
(E,E)-Farnesol	−5.75	−6.11
Methyl Anthranilate	−5.21	−5.65
Nerol	−6.04	−5.41
(E)-β-Farnesene	−5.62	−6.44
Nerolidol	−5.29	−7.55

The four substances, methyl anthranilate, indole, linalool and nerolidol, obtained through screening based on VIP values, had ROAVs at PT greater than one, demonstrating their substantial contribution to the overall aroma of the tea. Linalool, known for its floral and fruity aroma, was a key aromatic compound in both PFs and PT. Methyl anthranilate, with a honey-like floral note, and nerolidol, known for its floral and apple-like characteristics, were also identified as important contributors to the floral notes of PT.

Indole is a crucial floral component in jasmine and osmanthus teas (Wang, Deng, et al., 2024; Zhao et al., 2023). It is an aromatic compound with dual sensory characteristics, contributing a sweet floral odor at low concentrations (Zhang et al., 2022) and a feces odor at high concentrations (Rujirapong et al., 2022). In PT, indole exhibited an ROAV of 3.67, indicating its significant role in enhancing the floral aroma of PT. This finding suggested that the indole content in PT was optimally balanced, effectively contributing to its floral aroma. The resulting flavor profile of PT was similar to that of jasmine tea, thereby enhancing its commercial appeal and market potential.

In conclusion, the transformation of volatile compounds during the scenting process of PT played a critical role in enhancing its aroma profile. The increased levels of key compounds such as linalool, nerolidol, methyl anthranilate, and indole significantly contributed to the overall flavor improvement, demonstrating the potential of PF in enhancing the sensory characteristics of GT.

3.4. Differences in amino acid profiles between GT and PT

Free amino acids are important non-volatile components in tea and play a significant role in the taste profile of GT (Zhu et al., 2016). The total content of free amino acids was significantly higher in PT than in GT. A total of 14 free amino acids were identified, but Ser and Asp were not detected (Fig. 5A). All amino acids, except Ala, accumulated in PT, with theanine, Glu, His, Arg, and Ile showing significant accumulation (p < 0.05). Theanine, a typical water-soluble free amino acid in tea, imparts a sweet and umami taste and significantly contributes to tea aroma. The theanine content in GT ranged from 11.45 to 12.32 mg/g, consistent with previous reports (Li et al., 2021). To evaluate the contribution of each amino acid to the overall flavor, we calculated the TAVs for the amino acids that were detected (Table S6). Glu and theanine had TAVs greater than 1 in both PT and GT, indicating that they play a prominent role in shaping the taste of PT. It has been reported that the umami flavor in tea was primarily the result of three amino acids: theanine, Glu, and Asp (Zhang et al., 2020). Consequently, the significant increase in Glu and theanine content in PT intensifies its umami flavor. Additionally, Thr had a TAV greater than 1 in PT and is characterized as sweet. This suggests that increasing this substance may enhance the sweetness of the PT infusion. However, although the sweetness of the PT infusion detected by the electronic tongue decreased, sugars or sugar alcohols remain the primary substances responsible for sweetness. Therefore, the increase in Thr contributes little to the increase in sweetness. Other amino acids, such as Arg, Phe, Val, and Ile, which have a bitter taste, also increased in PT. However, their TAVs were all less than 1, indicating that they had a minimal impact on the taste profile of PT. The main source of free amino acids in PT was the base tea, but the scenting process also contributed to some extent. Previous studies have demonstrated dynamic changes in amino acid content at different stages of GT processing, potentially due to protein hydrolysis, the Maillard reaction, etc. (Yang et al., 2024).

Fig. 5.

(A) Differences in amino acid content between PT and GT. * represents p < 0.05, and * * represents p < 0.01. (B) A correlation network between taste attributes and amino acids. Only parameters with statistically significant correlations (p < 0.05) are shown. Correlations were analyzed using the Pearson method, with red indicating positive correlations and blue indicating negative correlations. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

3.5. Correlations of amino acids with taste attributes

Amino acids are known to contribute a wide range of taste attributes, including sweetness, bitterness, and umaminess ness. To investigate these contributions, we correlated the levels of amino acids with taste attributes assessed by the electronic tongue. Most amino acids, except Met and Ala, were significantly correlated with umaminess, saltiness, aftertaste-A, aftertaste-B (aftertaste-bitter), and bitterness (Fig. 5B). The 14 amino acids detected in this study were categorized into three groups based on their taste profiles: umami amino acids (e.g., theanine and Glu), bitter amino acids (e.g., Val, Ile, Leu, Tyr, Phe, His, Cys, and Arg), and sweet amino acids (e.g., Ala, Gly, Thr, Met, and Lys).

Approximately 70 % of the umami taste perception in tea infusion is derived from amino acids, particularly theanine, Asp, and Glu (Ye et al., 2018). Therefore, the significantly increased levels of theanine and Glu in PT likely enhanced its umami flavor. Other amino acids, including some bitter ones such as phenylalanine, also showed positive correlations with umaminess ness. This amino acid has been shown to enhance the umami flavor in tea (Shan et al., 2024).

Among bitter amino acids, Arg and Val were predominant in PT (Das et al., 2019). Although the arginine content in PT was significantly elevated, the overall increase in bitterness was not significant, suggesting that while these amino acids contributed to bitterness, other bitter components such as caffeine, theophylline, and tannins also played important roles (Chen et al., 2025). Moreover, interactions between amino acids and other compounds could reduce bitterness, contributing to a more balanced and complex flavor profile (Liu, Xiao, et al., 2023).

Notably, most sweet amino acids showed a non-significant negative correlation with sweetness, with only alanine exhibiting a positive correlation (p > 0.05). This suggested that alanine played a key role in the perceived sweetness of tea, consistent with the report by Wei et al. (2023) that L-alanine is the primary sweet compound in Camellia nanchuanica black tea.

In summary, the increased concentrations of amino acids in PT, especially theanine and glutamic acid, corresponded to the enhanced sensor responses for bitterness, astringency, and umaminess. The increased amino acid content in PT significantly enhanced its umaminess.

3.6. Contribution of volatile compounds to the umami taste perception in PT

The addition of PFs significantly improved the umami taste of GT. This improvement could be attributed not only to the increased amino acid content but also to the synergistic enhancement of umami taste by saltiness. In addition, volatile compounds also contributed to this flavor enhancement (Yu et al., 2021). To explore this contribution, we performed molecular docking to assess the interactions between seven volatile compounds that migrated from PFs to GT and the umami taste receptors TAS1R1 and TAS1R3 (See Table 2 and Fig 6). All these compounds could bind to the umami taste receptors, with binding energies ranging from −7.55 kcal/mol to −5.21 kcal/mol. These differences in binding energy could be attributed to the structural variation among the compounds and the conformational difference between the receptors (Zhang et al., 2021), with lower binding energies indicating a higher binding affinity between the compounds and the receptors (Herlina et al., 2024). Among the tested compounds, β-ocimene showed the lowest binding energy for TAS1R1, while nerolidol exhibited the lowest binding energy for TAS1R3. These findings indicated that β-ocimene and nerolidol had a strong binding affinity for the umami taste receptors and were likely to form stable complexes, thus contributing significantly to umami taste perception (Yao et al., 2024).

Fig. 6.

Molecular docking of seven volatile compounds with the umami taste receptors TAS1R1 and TAS1R3. (A)–(G) Docking models with TAS1R1 for (A) β-ocimene, (B) β-pinene, (C) farnesol, (D) methyl anthranilate, (E) nerol, (F) (E)- β-farnesene, and (G) nerolidol. (H–N) Docking models with TAS1R1 for the same compounds in the same order.

The interactions between the volatile compounds and the TAS1R1/TAS1R3 receptor occurred via hydrogen bonding and hydrophobic interactions, two mechanisms known to play a crucial role in stabilizing ligand–receptor binding (Shan et al., 2022). Although further experimental validation is required to elucidate the binding mechanisms, these molecular docking results suggested that the volatile compounds migrating from PFs to GT could interact with the amino acid residues of umami taste receptors through hydrogen bonding and hydrophobic interactions, thus influencing the flavor profile of PT.

4. Conclusion

This study aims to evaluate the enhancement of aroma and taste quality in GT following scenting with PFs. The findings demonstrated that both the electronic nose and electronic tongue effectively differentiated GT and PT based on their odor and taste profiles. Compared to GT, PT exhibited a significantly greater number and total content of volatile compounds. The addition of PFs increased the levels of 23 compounds, including linalool, nerol, nerolidol, and methyl anthranilate, while reducing the levels of 10 compounds, including DL-limonene, (Z)-ocimene, benzaldehyde, and methyl salicylate. These volatile compounds were key to the overall aroma profile of PT. OPLS-DA identified linalool, nerolidol, methyl anthranilate, and indole as differential aroma compounds between PT and GT, each exhibiting an ROAV >1 and significantly influencing the aroma profile of PT. Electronic tongue analysis revealed significant differences in umaminess, sweetness, saltiness, and aftertaste-A between the two teas. The enhanced umaminess in PT was attributed to increased amino acid levels (particularly theanine and glutamic acid, TAVs >1), synergistic effects between umaminess and saltiness, and aroma–taste interactions, which collectively intensified the umami taste perception.

Based on comprehensive aroma and taste analyses, it was concluded that GT absorbed aromatic compounds from PFs during the scenting process, leading to increased amino acid content in PT and effectively improving its sensory quality. These findings establish a theoretical basis for developing PF-based tea products and provide insights into the mechanism shaping the flavor quality of PT.

CRediT authorship contribution statement

Yanyan Ma: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Yiwen Hu: Writing – original draft, Visualization, Software, Investigation, Formal analysis, Data curation. Honggui Peng: Visualization, Formal analysis. Zhenni Yang: Validation, Software. Yongqiang Zheng: Writing – review & editing, Resources, Conceptualization. Jun Wang: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition, 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.

Appendix A. Supplementary data

Supplementary material. Table S1: Tea quality commentary and rating scale for each quality factor. Table S2. Electronic nose sensors and its corresponding representative sensitive compounds. Table S3 ANOVA and significance analysis of electronic nose. Table S4 ANOVA and significance analysis of electronic tongue. Table S5. Volatile compounds identified in the aroma concentrate of PF, GT and PT. Table S6. Amino acid taste thresholds and taste activity values (TAV) in GT and PT.

Data availability

Data will be made available on request.