Skip to main content
NCBI home page
Current Research in Food Science logoLink to Current Research in Food Science
. 2024 Oct 29;9:100910. doi: 10.1016/j.crfs.2024.100910

New insights into the role of lipids in aroma formation during black tea processing revealed by integrated lipidomics and volatolomics

Shan Zhang a,1, Le Chen a,b,1, Linchi Niu a, Haibo Yuan a, Xujiang Shan a, Qianting Zhang a,c, Yuning Feng a,b, Qinghua Zhou b, Yongwen Jiang a,⁎⁎, Jia Li a,⁎
PMCID: PMC11577132  PMID: 39569005

Abstract

Lipids are important tea aroma precursors. Due to the complexity of black tea processing involving both enzymatic and thermal reactions, the role of lipids in black tea aroma formation remained unclear. Herein, the dynamic changes of lipids and volatiles during black tea processing were simultaneously analyzed by lipidomics and volatolomics using ultra-high-performance liquid chromatography coupled to Q-Exactive Orbitrap mass spectrometry (UHPLC-Q-Exactive) and gas chromatography-tandem mass spectrometry (GC-MS/MS). The lipidomics method was validated in linearity, reproducibility, and recovery, which showed a high reliability. A total of 374 lipids and 88 volatiles were detected. Among them, 362 lipids and 29 fatty acid-derived volatiles (FADVs) were significantly altered depending on different processing stages. During the enzyme-driven stages of black tea processing (withering, rolling and fermentation), monogalactosyldiacylglycerol (MGDG), phosphatidylcholine (PC), and phosphatidylethanolamine (PE) were largely downregulated (<0.33 folds). Instead, in the non-enzymatic drying steps of black tea processing, triacylglycerol (TG), diacylglycerol (DG), and phosphatidic acid (PA) were mainly degraded (<0.24 folds). MS/MS fragmentation revealed that these most prominently degraded lipids were structurally enriched with fatty acyl (FA) 18:2 and 18:3 residues, such as MGDG (18:2/18:3), PC (18:2/18:2), PE (18:1/18:2), TG (18:3/18:3/18:3), DG (18:3/18:3), PA (18:3/18:3). Correlation analysis showed significant negative correlation between these lipids and FADVs such as aliphatic aldehydes, alcohols, ketones, and esters, etc. These most prominently degraded lipids were highlighted as the key potential aroma precursors during black tea processing, which were possibly oxidized and degraded into volatiles through enzyme- and thermal-driven pathways at different processing stages.

Keywords: Black tea, Lipids, Lipidomics, Volatiles, LC-MS, GC-MS

Graphical abstract

Image 1

Open in a new tab

Highlights

  • •

    Integrated lipidomics and volatolomics were applied.

  • •

    The changes of 374 lipids and 88 volatiles were quantified in black tea processing.

  • •

    The lipids and volatiles showed stage-specific variations in black tea processing.

  • •

    MGDG, PC, PE with FA 18:2 and 18:3 were mainly degraded in enzymatic reaction stage.

  • •

    TG, DG, PA, PI with FA 18:2 and 18:3 were mainly degraded in thermal reaction stage.

1. Introduction

Tea (Camellia sinensis L.) is an important flavored beverage. According to the post-harvest processing and fermentation degree, tea can be divided into six classes: un-fermented green tea, partially fermented white tea and yellow tea, semi-fermented oolong tea, fully fermented black tea, and post-fermented dark tea (Chen, 1981; Shevchuk et al., 2020). Black tea has gained the highest popularity among consumers, accounting for approximately 78% of global tea consumption (Li et al., 2017; Shevchuk et al., 2020), owing to its appealing flavor and multiple health benefits (Li et al., 2019; Sharma and Rao, 2009). The unique organoleptic flavor of black tea includes a rich and fragrant aroma, sweet mellow taste, and bright amber-red liquor color (Li et al., 2019; Zhang et al., 2023a). Aroma is a crucial contributor to the quality of tea and is an essential index for tea quality evaluation (Yang et al., 2013). Previous studies have reported that only a few aromatic compounds are present in fresh tea leaves dominated by grass and green notes; however, more than 700 volatiles have been identified in finished black tea (Wan, 2003; Zeng et al., 2018a). These volatiles are mainly produced from aroma precursors via a series of transformations during black tea processing (withering, rolling, fermentation, and drying), and shape the elegant black tea aroma quality with sweet, fresh, fruity, and floral notes (Feng et al., 2019; Zeng et al., 2018a). Thus, elucidating the metabolic fates of key aroma precursors in black tea processing is of particular significance for advancing our understanding about the black tea flavor formation.

Many studies have focused on dynamic changes in aromatic substances and their precursors during tea processing (Chen et al., 2019, 2022; Feng et al., 2019). Typical aromatic precursors in plants include lipids, amino acids, geranyl/farnesyl pyrophosphate, and carotenoids, which can be transformed into volatile compounds through metabolic pathways of lipid oxidation, amino acid degradation, synthesis of terpenoids, and carotenoid degradation, respectively (Chen et al., 2022; Dudareva et al., 2013). Accordingly, the volatile compounds in tea can be classified as fatty acid-derived volatiles (FADVs), amino acid-derived volatiles (AADVs), volatile terpenoids (VTs), and carotenoid-derived volatiles (CDVs) (Chen et al., 2022). Lipid degradation contributes substantially to the black tea aroma formation. Evidences have demonstrated that unsaturated fatty acids, such as linolenic acid and linoleic acid, undergo lipoxygenase (LOX)-catalyzed oxidation and decomposition during tea processing, generating a number of aliphatic aldehydes, alcohols, ketones, and esters which impart various scents (Chen et al., 2022; Ho et al., 2015; Liu et al., 2023; Yang et al., 2013). For example, hexanal and (E,E)-2,4-heptadienal (green, fresh notes), cis-jasmone (sweet, floral notes) are typical FADVs in tea (Ho et al., 2015; Yang et al., 2013; Zeng et al., 2018b). Feng et al. (2019) reported that FADVs contributed approximately 38% of the total volatiles in black tea aroma during processing. Therefore, the conversion of lipids and their correlation with volatile compounds during tea processing need to be systematically elucidated.

Some studies have been conducted in this context. Zhou et al. (2022a) showed that fatty acid composition and LOX activity were affected by oolong tea processing. Chen et al. (2022) reported that six major fatty acids were extensively reduced, whereas FADVs were significantly altered during black tea fermentation. In our previous studies, we demonstrated the high chemical diversity of lipids in tea leaves (Li et al., 2017, 2021, 2023). Some lipids, especially monogalactosyldiacylglycerol (MGDG) (18:3/18:3), were found to be significantly decreased in the rolling and fermentation of black tea processing (Li et al., 2017). However, to date, no studies have simultaneously analyzed the dynamic variations in lipid species and volatiles, and the correlation network between them remains unclear. Moreover, some issues remain far from being completely clarified. In view of the highly diverse lipid composition in tea leaves, the key lipid species functioning as aroma precursors remain unclear. Furthermore, the processing steps for black tea are rather complicated and involve both enzymatic (such as fermentation stage) and thermal (such as drying stage) reactions (Feng et al., 2019; Zhang et al., 2023b). It poses challenges towards an in-depth and comprehensive understanding of the “metabolic” pathways of lipid conversion and volatile formation in black tea processing.

Lipidomics allows a large-scale depiction of lipids on the scale of individual molecular species and has been proven to be a versatile tool in the foodomics field (Chen et al., 2023, 2024a; Li et al., 2020; Xu et al., 2020). In our previous studies, a tea lipidomics profiling method was established based on liquid chromatography coupled to mass spectrometry (LC-MS) (Li et al., 2017, 2021, 2023). Volatolomics using gas chromatography-mass spectrometry (GC-MS) combined with headspace solid-phase microextraction (HS-SPME) has emerged as a promising tool for the comprehensive analysis of volatile compounds and has been widely employed in the food flavor research including tea (Feng et al., 2019; Yang et al., 2013). Integrated characterization of lipids and volatiles has demonstrated their promising roles in the elucidation of flavor formation mechanisms in various foods (Li et al., 2022; Yu et al., 2023).

The aims of the present study were: to (1) simultaneously track the dynamic changes of lipids and volatiles during black tea manufacturing and their correlations, and (2) probe the potential key aroma precursors and metabolic pathways. To address these questions, a multi-omics strategy was employed by integrating lipidomics and volatolomics using UHPLC-Q-Exactive and GC-MS/MS.

2. Materials and methods

2.1. Chemicals and reagents

Liquid chromatographic-grade solvents including acetonitrile, isopropanol, methanol (MeOH), methyl tert-butyl ether (MTBE), ammonium acetate, dichloromethane (CH2Cl2), and ethanol were purchased from Merck (Darmstadt, Germany). Internal standards of 13C18-linoleic acid (isotopic purity 99 atom% purity), lyso-phospatidylcholine (LPC) (19:0) (> 99% purity), and phosphatidylcholine (PC) (19:0/19:0) (> 99% purity) were obtained from Merck (Darmstadt, Germany) or Avanti Polar Lipids (Alabaster, Alabama, USA). Chromatographically pure ethyl decanoate was purchased from the Aladdin Biochemical Technology Company (Shanghai, China). Stock solutions of lipid standards were prepared in CH2Cl2/MeOH (2:1, v/v). Ethyl decanoate was dissolved in ethanol. All the stock solutions were stored at −20 °C before using. Ultrapure water was obtained from a Milli-Q system (Millipore, MA, USA).

2.2. Black tea manufacturing

The fresh leaves (one bud with two leaves) of tea variety Camellia sinensis var. Jiukeng were harvested in April 2022 from the Shengzhou Experiment Base, Tea Research Institute, Chinese Academy of Agricultural Sciences (Shengzhou, Zhejiang, China). Black tea was manufactured according to a previous study (Wu et al., 2022), following the procedures of withering, rolling, fermentation, first drying, and second drying. First, fresh tea leaves were spread and withered at 28 °C and 70% relative humidity (RH) for approximately 15 h until the moisture content of the withered leaves dropped to about 64%. The leaves were then rolled using a roller machine (6CRN-35, Xiangfeng Machinery Co., Ltd., China) for 75 min. And then, the rolled leaves were fermented in a climatic cabinet at 30 °C and 90% RH for 3 h. Finally, the fermented tea leaves were dried for 25 min at 110 °C followed by drying at 90 °C for 35 min. The tea leaves were collected at different time points during black tea manufacturing, including fresh leaves (FL) and withered leaves (WL, after withering); after rolling for 15 min (R1), 45 min (R2), and 75 min (R3); after fermentation for 1 h (F1), 2 h (F2), and 3 h (F3); after the first drying (D1); and after the second drying (D2). The samples were immediately snap-frozen in liquid nitrogen and freeze-dried. All samples were stored at − 80 °C before analysis.

2.3. Lipidomics analysis

2.3.1. Total lipid extraction

The total lipid extraction was performed using the MTBE method (Matyash et al., 2008; Li et al., 2017; Chen et al., 2024b). Briefly, 300 μL of MeOH spiked with mixed exogenous internal standards was added to 20 mg of finely grounded tea powder, followed by addition of 1 mL of MTBE. After vortexing, 300 μL of ultrapure water was added. The two phases were separated by centrifugation. The total hydrophobic metabolites (lipids) were obtained from the organic phase of the supernatant. The internal standards were 13C18-linoleic acid, LPC (19:0), and PC (19:0/19:0), and their concentrations in MeOH solvent were 6.7 μg/mL, 2.0 μg/mL, and 1.3 μg/mL, respectively. Lipid extraction was performed in triplicates for each sample.

2.3.2. UHPLC-Q-Exactive analysis

UHPLC-Q-Exactive (Thermo Fisher, CA, USA) was used for lipidomic analysis, as previously reported with slight modifications (Li et al., 2017, 2021, 2023). LC-MS was calibrated using fresh positive and negative ion calibration solutions before running, by direct infusion through a syringe pump to ensure mass accuracy. The detailed information is provided in the Supplementary Information.

LC separation was conducted on an ACQUITY UPLC BEH C8 column (2.1 × 100 mm, 1.7 μm, Waters, MA, USA) at 55 °C using mobile phase A (acetonitrile:water = 6:4, v/v, containing 10 mM ammonium acetate) and phase B (isopropanol:acetonitrile = 9:1, v/v, containing 10 mM ammonium acetate) and operated in a flow rate of 0.26 mL/min. The binary gradient elution was adopted: 0–1.5 min, maintaining at 32% B; 1.5–15.5 min, linear increase of B to 85%; 15.5–15.6 min, further increase of B to 97%; 15.6–18.0, maintaining at 97% B; 18.0–18.1, return to 32% B; and 18.1–21 min, reconditioning at 32% B. The LC effluent was transferred to a heated ESI interface for ionization. The ESI settings were as follows: the sheath gas and auxiliary gas flow rates were 45 arb and 10 arb, respectively; the capillary temperature was 300 °C; the probe heater temperature was 350 °C; the spray voltage was 3.50 kV and 3.0 kV for ESI (+) and ESI (−), respectively. A complete MS scan was conducted at a mass resolution of 70,000 with a mass-to-charge ratio (m/z) ranging from 150 to 1500. Stepped collisional energies of 25, 35, and 45 eV were used for the MS/MS fragmentation. Quality control (QC) samples obtained by mixing equal aliquots from all samples were analyzed during the entire run. Structural annotation of lipid species was accomplished based on accurate mass measurements (<5 ppm), MS/MS fragments, retention time (RT), and database queries, including LIPIDMAPS (https://www.lipidmaps.org/, HMDB (https://hmdb.ca/), and Metlin (https://metlin.scripps.edu/). The detailed information of lipid identification is provided in the Supplementary Information. The diagnostic MS/MS fragments for lipid identification are shown in Table S1.

2.3.3. Lipidomics data processing

The raw LC-MS data were processed using the XCMS 3.4.1 software for ion feature extraction and time alignment. All detected ions were normalized to the total peak areas, and missing values were treated using the 80% rule (Shan et al., 2024; Zhang et al., 2023b). For quantification, the peak areas were obtained using the Xcalibur software (Thermo Fisher, San Jose, USA), and the quantification of individual lipid species was achieved by the one-point quantification method using corresponding internal standards, as previously described (Li et al., 2023).

The calculation formula is as follows:

Clipid = Cis ∗ Alipid / Ais

Where Clipid represents the target lipid concentration (nmol/g), Cis represents the internal standard concentration (nmol/g), Alipid indicates the peak area of target lipid, and Ais indicates the peak area of internal standard.

Phosphoglycerolipids, acylglycerolipids, sphingolipids, and sterols were quantified using the internal standard PC (19:0/19:0). Lysophosphoglycerolipids and glycoglycerolipids were quantified using LPC (19:0/19:0). Free fatty acids (FFA) were quantified using 13C18-linoleic acid. Ions or lipid species with relative standard deviations (RSD) less than 20% in all QC samples were retained for further statistical analysis.

2.3.4. Lipidomics method validation

The analytical performance of lipidomics method was evaluated regarding linearity, reproducibility, and recovery.

Linearity Linearity was determined by the mass spectrometric intensity of three non-endogenous lipids (13C18-linoleic acid, LPC (19:0), and PC (19:0/19:0)) at a series of concentration levels, i.e., 0.0005, 0.002, 0.01, 0.04, 0.2, 1.0, 4.0, and 10.0 μg/mL. Calibration curves were constructed by linear regression for each lipid.

Reproducibility Reproducibility was evaluated by intraday precision. Five replicates were performed following the same lipid extraction and LC-MS analysis. All detected endogenous lipids were semi-quantified by one-point quantification method using corresponding internal standards. And the RSD% of each lipid among five replicates was investigated.

Recovery Recovery was evaluated by comparing mass spectrometric intensity of three non-endogenous lipids (13C18-linoleic acid, LPC (19:0), and PC (19:0/19:0)) spiked prior to and after tea lipids extraction at low (0.25 μg/mL), median (1.0 μg/mL) and high (4.0 μg/mL) levels. The same tea sample was used for extraction. At each spiking level, three replicates were performed.

2.3.5. Lipid nomenclature and abbreviations

Lipid annotation in this manuscript followed the lipid nomenclature system of LIPID MAPS (Fahy et al., 2009; Li et al., 2023). Abbreviations of lipid names were used as following: PC, phosphatidylcholine; PE, phosphatidylethanolamine; PA, phosphatidic acid; PG, phosphatidylglycerol acid; PI, phosphatidylinositol; PS, phosphatidylserine; LPA, lyso-phosphatidic acid; LPC, lyso-phosphatidylcholine; LPE, lyso-phosphatidylethanolamine; LPG, lyso-phosphatidylglycerol; LPI, lyso-phosphatidylinositol; TG, triacylglycerol; DG, diacylglycerol; MGDG, monogalactosyldiacylglycerol; DGDG, digalactosyldiacylglycerol; SQDG, sulfoquinovosyldiacylglycerol; Cer, ceramide; HexCer, hexcosylceramide; FFA, free fatty acid.

2.4. Volatolomics analysis

2.4.1. HS-SPME-GC-MS/MS analysis

The tea volatiles were extracted using HS-SPME equipped with a divinylbenzene/carboxen/polydimethylsiloxane (DVB/CAR/PDMS) fiber (50/30 μm, Supelco, PA, USA) as previous studies (Wang et al., 2021; Zhang et al., 2020; Xie et al., 2024). Briefly, tea samples (0.5 g) were weighed accurately and transferred to a 20 mL headspace vial. Ten microliters of ethyl decanoate (20 μg/mL) was added as internal standard, followed by 1 mL of ultra-pure water. The headspace vial was then capped with a PTFE/silicone septum (18 mm) to isolate the air and the septum was immediately penetrated using the SPME fiber. The headspace vial was incubated in water bath at 60 °C and the SPME fiber was exposed to the headspace to absorb the volatile compounds for 60 min. Further, the adsorbed head was inserted into the GC-MS inlet for thermal desorption and kept at 250 °C for 5 min. All samples were prepared in triplicates.

The volatile compounds were analyzed using a GC-MS/MS system (7890B-7000C; Agilent Technologies; CA, USA) equipped with an HP-5 MS capillary column (30 m × 250 μm × 0.25 μm; Agilent Technologies; CA, USA). The column temperature was initiated at 40 °C which was held for 3 min, linearly increased to 240 °C with a rate of 5 °C/min and maintained for 3 min, and then continually increased to 250 °C with the rate of 10 °C/min and maintained for 3 min. Helium (>99.999% purity) was used as the carrier gas at a flow rate of 1.0 mL/min. The ionization energy was set to 70 eV. The temperatures of the ion source and quadrupole detector were 230 °C and 150 °C, respectively. The mass spectral scanning range was 40–450 m/z. Each injection lasted 60 min.

2.4.2. Volatolomics data processing

The GC-MS data were processed using Agilent MassHunter Workstation Software. First, the volatile metabolites were screened by matching their mass spectra in the NIST 11 database with a similarity of over 80%. Second, the volatile compounds were further identified by matching the retention index (RI) value to the theoretical RI value reported in the standard reference databases (https://www.chemspider.com/, https://webbook.nist.gov/chemistry/, https://www.flavornet.org/f_kovats.html). The RI values of volatiles were calculated using the linear equations of n-alkanes (C7-C40) performed on the same conditions. The theoretical RI values reported in NIST database were generated using the same column. The concentrations of the volatile compounds were semi-quantified using the internal standard as previous studies (Wang et al., 2022a; Zhang et al., 2020).

The calculation formula is as follows:

Cvolatile = Cis ∗ Avolatile / Ais

Where Cvolatile represents the target compound concentration (μg/g), Cis indicates the internal standard concentration (μg/g), Avolatile represents the peak area of target compound, and Ais is the peak area of internal standard.

2.5. Statistical analysis

Significant differences were analyzed by one-way ANOVA using SPSS software (version 26.0.0, IBM, NY, USA). Principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA) were performed using SIMCA-P (version 14.1, Umetrics, Umeå, Sweden). Heatmaps were visualized using TBtools-II (v1.120, Toolbox for Biologists, China). The hierarchical clustering was performed based on the Euclidean distance. Sankey diagram was performed using OmicShare Tools (https://www.omicshare.com/). Volcano plots were conducted on the online website of Sangerbox (http://www.sangerbox.com/). Bar graphs were generated using the GraphPad Prism software (GraphPad Software, San Diego, CA, USA). The chord diagram of correlation analysis was achieved using Bioinformatics (https://www.bioinformatics.com.cn/). Correlation heatmap was conducted using Chiplot (https://www.chiplot.online/). The pathway was mapped by referring to the databases of LIPIDMAPS (https://www.lipidmaps.org/) and KEGG (https://www.genome.jp/kegg/).

3. Results and discussion

3.1. Lipidomics analysis of tea samples

The UHPLC-Q-Exactive based lipidomics analysis of black tea samples generated 3739 and 3338 ions in ESI (+) and ESI (−) modes, respectively. The typical total ion chromatograms (TICs) in ESI (+) and ESI (−) modes of QC samples are displayed in Fig. S1. The analytical performance of the lipidomics method was evaluated in terms of its linearity, reproducibility, and recovery.

Linearity Wide dynamic ranges covering 3.0–4.3 orders of magnitude were obtained for three non-endogenous lipids (13C18-linoleic acid, LPC (19:0), and PC (19:0/19:0)), with regression coefficients R2 > 0.99 (Table S2).

Reproducibility Five replicates were performed to assess the intraday precision. A median RSD of 10.2% was achieved for all the detected lipids. Approximately 81% of the detected lipids had RSD <20%, contributing to 98.2% of the total peak area (Fig. S2). Moreover, in the non-supervised PCA score plots (Fig. S3), all QC samples were clustered closely, which also indicated satisfactory reproducibility of the present study.

Recovery At three different spiking levels, the recovery was 87.3%–95.5% for 13C18-linoleic acid, 79.9%–83.9% for LPC (19:0), and 88.7%–99.8% for PC (19:0/19:0). Mean recoveries of 90.9%, 82.2%, and 93.4% were achieved, respectively, for 13C18-linoleic acid, LPC (19:0) and PC (19:0/19:0) (Table S3).

In general, the method validation results were satisfactory, demonstrating the reliability of the lipidomics approach.

3.2. Dynamic changes of lipid molecular species in black tea processing

The global lipidomic alterations during black tea processing using total ions with RSD <20% are shown in Fig. 1. The supervised PLS-DA score plots of lipid fingerprints that acquired from the ESI (+) (R2X = 0.911, R2Y = 0.981, Q2 = 0.874) and ESI (−) (R2X = 0.895, R2Y = 0.983, Q2 = 0.806) modes demonstrated a pronounced stagewise change along with the manufacturing steps of black tea (Fig. 1A and B). Tea samples were separated into two groups in principal component 1 (PC1). Tea samples of fresh leaves (FL), withering leaves (WL), rolled leaves (R1-R3), and fermented leaves (F1-F3), were mainly located in the right half of PC1. While tea samples after the first and second drying steps were placed in the left half of PC1. Cross-validation with 200 permutations suggested a good predictive performance of the PLS-DA models (Fig. S4).

Fig. 1.

Fig. 1

Open in a new tab

(A, B) PLS-DA score plots of lipid profiles acquired from the ESI (+) and ESI (−) mode. (C) The (sub)classification of 362 identified lipid species in black tea samples. (D) Heatmap visualization of the dynamic variations of the lipid subclasses during black tea processing. (E) The content proportions of different (sub)classes of lipid molecules and their changes during black tea manufacturing as visualized by the Sankey diagram. The total contents of each lipid subclass are calculated by summing up all individual lipids belonging to the same subclass. The data shown in the heatmap are the average of three replicates after unit variance (uv) scaling, red or blue colors indicate lipids occurred at higher or lower level than the average. FL, fresh tea leaves; WL, withered tea leaves; R1, R2 and R3, tea leaves rolled after 15 min, 45 min, 75 min; F1, F2 and F3, tea leaves fermented after 1 h, 2 h, 3 h; D1 and D2, tea leaves after first-drying and second-drying. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

High chemical complexity and diversity were observed in the black tea lipidome (Fig. 1C). A total of 374 lipid molecular species were annotated, spanning over six major lipid classes: 124 phosphoglycerolipids (33.16%), 104 acylglycerolipids (27.81%), 47 glycoglycerolipids (12.57%), 45 sphingolipids (12.03%), 33 fatty acyls (8.82%), and 21 sterol lipids (5.61%). These lipid molecules can be further categorized into 22 subclasses, comprising 34 PC, 24 phosphatidylethanolamines (PE), 13 phosphatidic acids (PA), 11 phosphatidylglycerols (PG), 11 phosphatidylinositols (PI), eight phosphatidylserines (PS), three lyso-phosphatidic acids (LPA), 13 LPC, four lyso-phosphatidylethanolamines (LPE), two lyso-phosphatidylglycerols (LPG), one lyso-phosphatidylinositols (LPI), 84 triacylglycerols (TG), 20 diacylglycerol (DG), 13 MGDG, 18 digalactosyldiacylglycerols (DGDG), 16 sulfoquinovosyldiacylglycerols (SQDG), 15 ceramides (Cer), 30 hexosylceramides (HexCer), eight acyl-Glc-sitosterols, 10 acyl-(Glc)-stigmasterols, three acyl-Glc-campesterols, and 33 FFA. This lipid species distribution was generally in line with our previous study (Li et al., 2017). Among them, 362 lipid molecular species showed significant alterations (P < 0.05), accounting for 96.8% of all lipids, indicating that lipids were extensively influenced by manufacturing processes. Detailed information on the 362 differentially expressed lipid molecules is provided in Supplementary Information Table S4.

The dynamic changes in the contents of the 22 lipid subclasses during black tea processing are visualized as heatmap and all showed statistical differences (P < 0.05) (Fig. 1D). The heatmap revealed obvious stage-specific variations along with the time-course of black tea processing, i.e., most lipids were generally in higher abundances in the time frame of “FL-WL-R-F” and were drastically declined in the stages of “D1-D2”. These two periods were largely dominated by enzymatic and thermal reactions. The lipid subclasses were clustered into three groups. The subclasses in cluster II, including glycoglycerolipids (MGDG, DGDG, and SQDG) and phosphoglycerolipids (PE, PC, and PS), demonstrated a gradual decreasing trend and reached their lowest abundances during drying. Conversely, lysophospholipids (LPA, LPI, and LPG), classified as cluster I, exhibited a reversed pattern. They successively increased during black tea processing and accumulated during the second drying. Cluster III comprised subclasses including (lyso-) phospholipids (PG, PA, PI, LPC, and LPE), acylglycerolipids (TG and DG), sphingolipids (Cer and HexCer), fatty acyls (FFA), and sterol lipids, which first increased and then decreased. Their contents increased during the fermentation stage and then sharply decreased during the drying stage. The overall changing patterns in the different lipid subclasses were in agreement with those of previous studies (Li et al., 2017; Zhang et al., 2024). The Sankey diagram illustrates the proportions of different lipid subclasses and their changes during the enzyme-driven steps (FL-F3) and thermal reaction steps (D1 and D2), where the height of the bar indicates the lipid content (Fig. 1E). The results demonstrated that phosphoglycerolipids, glycoglycerolipids, acylglycerolipids, and fatty acyls (PC, PE, PG, PI, PA, MGDG, FFA, and TG) were the dominant lipids in black tea samples and the lipid contents declined sharply after drying.

During tea processing, enzymatic oxidation and thermochemical reactions were recognized as the two major routes responsible for flavor formation (Zeng et al., 2018a; Zhang et al., 2019a). Black tea processing can be divided into two stages: enzyme-mediated oxidation (withering, rolling, and fermentation) and thermal oxidation (drying) (Feng et al., 2019; Zeng et al., 2018a). Lipids, as the aroma precursors, are vulnerable to enzymatic hydrolysis, as well as thermal degradation (Chen et al., 2023; Huang et al., 2024; Li et al., 2021; Shahidi and Hossain, 2022; Zhang et al., 2019b; Zhou et al., 2015). Our results indicated a stage-dependent evolution of lipid profiles during black tea processing and implied that enzymatic and thermal reactions may exert distinct effects on lipid degradation.

3.3. Screening and identification of key lipid molecules as potential aroma precursors

To probe the key lipid molecules as potential aroma precursors in the enzymatic reaction phase (FL, WL, R, and F) and thermal reaction phase (D1 and D2), the major altered lipid species were screened according to folds changes and lipid concentrations. In the enzymatic reaction stage, the lipid molecules with more than 2-folds decrease or increase (folds change < 0.5 or > 2, F3 vs. FL) were screened, among which the molecules with abundances ranking in top 50 (at time point of FL) were further selected as the major altered lipids (Table S5). Similarly, in the thermal reaction phase, the lipids with more than 2.5-folds change (folds change < 0.4 or > 2.5, D2 vs. F3) was set as a criterion. Among them, those lipids with concentrations ranking in top 50 (at time point of F3, i.e., the beginning of drying phase) were further screened (Table S6). Glycoglycerolipids and phosphoglycerolipids were the major downregulated lipid molecules in the enzyme-active stage, including 23 PC, nine PE, eight MGDG, four DGDG, four PG, one PS, and one SQDG, whereas the major upregulated lipids were 19 TG, nine PA, five LPC, three LPA, three DG, three PI, three sterol lipids, two LPE, two FFA, and one LPG (Table S5). Instead, acylglycerolipids and phosphoglycerolipids were emphasized as the major downregulated lipids in the thermal reaction stage, comprising 11 TG, five DG, seven PI, six PA, six PC, six PG, two MGDG, three DGDG, one PE, one SQDG, and two HexCer, while the major upregulated lipids were one LPI and one LPG (Table S6). Among the major downregulated lipids, only 12 lipids were shared in the two stages (Fig. S5). The results suggested that the lipid metabolic fates were dependent on different processing stages, and different lipid molecules were differently susceptible to enzymatic or thermochemical reactions.

Subsequently, the detailed fatty acyl compositions of the major downregulated lipid molecules were determined using MS/MS. The MS/MS spectrum are exemplarily shown in Fig. 2A–H for MGDG (18:2/18:3), PC (18:2/18:2), PE (18:1/18:2), TG (18:3/18:3/18:3), DG (18:3/18:3), PA (18:3/18:3), PG (16:1/18:3), and PI (18:3/18:3). As an example of the glycoglycerolipid MGDG (18:2/18:3) (Fig. 2A), the ion at 799.53 represented the precursor ion [M+Na]+. The product ions of 519.29 and 521.31 represented the neutral loss (NL) of 280 and 278, demonstrating the fatty acyl residues of linoleic acid (FA 18:2) and linolenic acid (FA 18:3), respectively. As another example of phosphoglycerolipid PI (18:3/18:3) (Fig. 2B), the product ion at 241.01 represented the polar head group of inositol, whereas the fragment at 277.22 suggested the fatty acyl residues of linolenic acid (FA 18:3). Similarly, fatty acyl residues were identified for other lipids, such as PC (18:2/18:2), PE (18:1/18:2), PA (18:3/18:3), PG (16:1/18:3), TG (18:3/18:3/18:3), and DG (18:3/18:3) (Fig. 2C-H). The detailed fatty acyl compositions of the significantly downregulated lipids in the enzyme-driven and thermal-reaction stages of black tea processing are provided in Supplementary Information Tables S5 and S6.

Fig. 2.

Fig. 2

Open in a new tab

(A–H) Identification of acyl chain compositions of lipid molecules, illustrated for MGDG (18:2/18:3), PI (18:3/18:3), PC (18:2/18:3), PE (18:1/18:2), PA (18:3/18:3), PG (16:1/18:3), TG (18:3/18:3/18:3), and DG (18:3/18:3) as examples, by MS/MS spectrum and fragmentation patterns.

The representative major altered lipids with both remarkably down- and up-regulated folds and high concentrations are labeled in the volcano plots (Fig. 3A and B). They were located far from the center of the volcano plot. Quantitative variation results for representative lipid molecules are presented in Fig. 4A. Quantitative estimation of the fatty acyl compositions of these lipids is shown in Fig. 4B and C.

Fig. 3.

Fig. 3

Open in a new tab

(A, B) Representative lipids with remarkable down-/upregulated folds and high concentrations involved in the enzymatic oxidation stage (A) and thermal reaction stage (B) labeled on the volcano plots. The volcano plots comprised 362 statistically changed lipid molecules during black tea processing. Green triangles represent the down-regulated lipid molecules with more than 2-folds (enzymatic reaction stage, F3 vs. FL) or 2.5-folds (enzymatic reaction stage, D2 vs. F3) decrease. Red triangles represent the up-regulated lipid molecules with more than 2-folds (enzymatic reaction stage, F3 vs. FL) or 2.5-folds (thermal reaction stage, D2 vs. F3) increase. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4.

Fig. 4

Open in a new tab

(A) Content changes of representative major down/up-regulated lipid species in the two stages of black tea processing. (B, C) Variations of fatty acyls of the down-regulated lipids during enzymatic reaction stage (B) and thermal reaction stage (C). The abundances of the fatty acyls were calculated based on the contents of all down-regulated lipids and their corresponding acyl chain compositions.

In enzymatic reactions (from fresh leaves to fermentation), MGDG, PC, and PE were the predominantly downregulated lipids (Fig. 3A). For example, MGDG (18:3/18:2), MGDG (18:3/18:3), PC (18:2/18:2), PC (18:2/18:3), PE (16:0/18:2), and PE (18:2/18:2) exhibited reduction with folds change value less than 0.33 after fermentation (Fig. 4A). Regarding the fatty acyl residues, linoleic acid (FA 18:2) was the major esterized fatty acid chain during the enzyme-active steps, followed by linolenic acid (FA 18:3), palmitic acid (FA 16:0), and linoleic acid (FA 18:1) (Fig. 4B). The abundances of 18:2 and 18:3 fatty acyls were obviously diminished, with folds changes of 0.34 and 0.35 (F3 vs. FL), demonstrating their great degradation in response to enzyme-induced reactions (Fig. 4B). Conversely, the lysophospholipids LPA, LPC, LPE, and LPG were the principally upregulated lipids, including LPA (18:2), LPC (18:3), LPE (16:0), and LPG (16:0) with approximately 2.96–5.78 times increase (Fig. 4A). MGDG, a subclass of glycoglycerolipids, is an important component of the thylakoid membrane in the chloroplasts of higher plants (Hölzl and Dörmann, 2007; Li et al., 2021). Glycerophospholipids, such as PC and PE, are the major components of the cytoplasmic membrane in tea plant cells (Chen et al., 2021; Li et al., 2017), and also serve as the precursors for lyso-phospholipids and FFA (Chen et al., 2023). The post-harvest tea manufacturing is essentially a process of cell dehydration and destruction (Zeng et al., 2018a; Zhang et al., 2023b). Previous studies have reported that lipase catalyzes lipid degradation, promoting the generation of polyunsaturated fatty acids and lysophospholipids (e.g., LPC, LPE, LPA, LPG), under stress conditions (Hou et al., 2016; Li et al., 2017), such as dehydration (withering) and cell disruption (rolling). And lipids have been reported to be the important precursors of volatile compounds in foods including tea, driven by LOX-catalyzed reactions (Chen et al., 2023; Ho et al., 2015; Ramaswamy and Ramaswamy, 2000; Huang et al., 2024). In the present study, the significantly declined amount of esterified FA 18:2, FA 18:3 in MGDG, PC, and PE after fermentation was inferred to be transformed into lyso-phospholipids and free polyunsaturated fatty acids, which further transformed into aliphatic aromas.

In the stage of thermal reactions (drying), the acylglycerolipids of TG, DG, and phosphoglycerolipids of PA and PI were found to be the pronounced down-regulated lipids (Fig. 3B). For instance, TG (18:3/18:2/18:2), TG (18:3/18:3/18:3), DG (18:2/18:2), DG (18:3/18:3), PA (18:3/18:2), PA (18:3/18:3), PI (18:3/18:2), and PI (18:3/18:3) decreased sharply after drying (folds change ranging from 0.10 to 0.24) (Fig. 4A). Besides, PC also presented downward trend during drying but less prominent (folds change > 0.31) (Table S6). Regarding the fatty acyl composition, FA 18:3 was the dominant esterized fatty acid, followed by FA 18:2 and other acyl residues, and their abundances manifested 0.28-folds and 0.26-folds (D2 vs. F3), respectively, after drying (Fig. 4C). Contrarily, the lysophospholipids LPG and LPI were recognized as the major upregulated lipids during drying, such as LPG (16:1) and LPI (18:3), which accumulated 4.96-folds and 7.91-folds, respectively, after drying (Fig. 4A). Lipids undergo decomposition via a free-radical chain mechanism during thermal oxidation to produce a wide range of volatile compounds (Shahidi and Hossain, 2022; Zhou et al., 2015). Nonpolar TG and DG are important energy storage units in eukaryotes (Chen et al., 2023; Wang et al., 2022b). Zhou et al. (2015) reported that TG species appeared more susceptible to thermal oxidation than PC species. PA has been reported to be sharply declined in the fixation stage of green tea processing at high temperature (Jia Li et al., 2021). The polar head group of PC has been reported to enhance its thermal stability and delay thermal oxidation (Zhou et al., 2022b), which explains the smaller decrease in PC during drying. Herein, TG, DG, PA, and PI harboring the 18:2 and 18:3 fatty acids were identified as the principal degraded lipids and may serve as major aroma precursors during the drying phase of black tea processing.

3.4. Volatile fingerprint evolution during black tea processing

A total of 88 volatile components were tentatively identified and semi-quantified in the tea samples, all of which were significantly altered during black tea processing (P < 0.05). Detailed information on the volatile compounds is provided in the Supplementary Information Table S7. A PLS-DA score plot (R2X = 0.98, R2Y = 0.96) was generated using all volatile compounds, where the fresh leaves (FL), withered leaves (WL), and initially rolled leaves (R1) were situated in the lower-right quadrant, the other rolled leaves and fermented leaves located in the left quadrants, and the dried leaves were situated in the upper right quadrant (Fig. 5A). The results demonstrate a stepwise alteration in the volatile profiles during black tea processing. The cross-validation with 200 times permutations suggested a good performance (R2 = 0.23, Q2 = − 0.87) of the PLS-DA model (Fig. 5B).

Fig. 5.

Fig. 5

Open in a new tab

(A) PLS-DA score plot of volatile metabolites during black tea processing. (B) The cross-validation of PLS-DA with 200 times permutations. (C) The classification of 88 identified volatile compounds classified as biochemical characters. (D) The dynamic changes of volatile compositions in black tea processing. (E) The classification of 88 identified volatile compounds classified as their precursors. (F) Heatmap of the FADVs. The data in heatmap are the average values of three replicates after uv scaling.

Based on their chemical structures, 88 volatile metabolites were classified into 13 aldehydes, 20 alcohols, six ketones, five acids, 10 alkanes, five terpenes, five furans, 18 esters, and six other compounds (Fig. 5C). The volatile composition changed dynamically during the black tea preparation procedure, as shown in Fig. 5D. Among them, the classes of aldehydes, alcohols, and esters accounted for the largest proportion of the total volatile abundance, as calculated in terms of the content percentage (Fig. 5D). Furthermore, 88 volatile compounds were annotated based on their precursors and synthetic pathways (Chen et al., 2022; Dudareva et al., 2013; Yang et al., 2013). They included 29 FADVs, 17 AADVs, 13 TVs, five CDVs, and 24 other volatile molecules (Fig. 5E). Notably, FADVs accounted for the largest proportion (approximately 32.95%) of the total volatiles, including five alcohols, six aldehydes, ten esters, six alkanes, one ketone, and one furan (Fig. 5E).

3.5. Dynamic changes of FADVs during black tea processing

The content changes of 29 FADVs during the time course of black tea processing are shown in a heatmap (Fig. 5F). The red box represents a higher level, while the blue box indicates a lower content compared to the average. The FADVs also exhibited distinct changes in the time periods of enzymatic and thermal reaction stages, which were similar as the lipidomic changing patterns (Fig. 1, Fig. 5D). Overall, FADVs were assigned to three groups based on hierarchical clustering. Quantitative changes in the contents of representative volatile molecules in each group are presented in Fig. 6.

Fig. 6.

Fig. 6

Open in a new tab

Content changes of the representative FADVs during black tea processing.

The volatiles in cluster II generally decreased. They comprised nine volatile compounds, including three alkanes, one aldehyde, one alcohol, and four esters (Fig. 5F). Alkanes such as dodecane presented highest levels in fresh leaves and generally declined (Fig. 6). Fatty esters, such as trans-2-hexenyl isovalerate and cis-3-hexenyl iso-butyrate, mostly accumulated during the withering and rolling steps (Fig. 6). In addition, decanal and 1-heptanol were largely elevated during withering, however fluctuating declined after rolling (Fig. 6). These results were consistent with those of previous studies (Chen et al., 2022; Liu et al., 2023). The volatiles in cluster III showed a firstly increasing and then decreasing trend, reaching a maximum amount during the rolling and fermentation steps of black tea processing (Fig. 5F). They included 12 volatile compounds, mainly aliphatic alcohols and esters such as (E)-3-hexen-1-ol, 1-nonanol, (Z)-2-penten-1-ol, (E)-hexanoic acid 2-hexenyl ester, and (Z)-hexanoic acid 3-hexenyl ester (Fig. 5, Fig. 6). Conversely, the volatiles in Cluster I showed a gradually increasing trend during black tea processing (Fig. 5F). They included fatty aldehydes and esters as well as ketones and furans such as hexanal, heptanal, (E,E)-2,4-heptadienal, hexadecanoic acid methyl ester, cis-3-hexenyl salicylate, cis-jasmone, and 2-pentyl-furan. They were generally increased, reaching their content peaks in the final stage of fermentation or drying (Fig. 5, Fig. 6).

Previous studies have reported that C6-C9 alcohols and aldehydes derived from unsaturated fatty acids, such as hexanal, (E)-2-hexenal, 1-nonanol, (Z)-2-penten-1-ol, and (E)-3-hexen-1-ol, are important contributors to the characteristic fresh and greenish odors (Ho et al., 2015; Yang et al., 2013). Heptanal and (E,E)-2,4-heptadienal play important roles in the fresh and floral flavors of tea (Liu et al., 2021; Yang et al., 2020). The aliphatic esters generally exhibit a pleasant fresh, sweet, and floral aroma in tea (Chen et al., 2022; Liu et al., 2023). Cis-jasmone is a representative floral and fruity odor aroma derived form α-linolenic acid (Yang et al., 2013). Moreover, 2-pentyl-furan is regarded to be generated by the cyclization of the oxidation products of linoleic acid, which contributes to the fruity and bean-like aroma of black tea (Liu et al., 2021, 2023).

Overall, the content of FADVs showed significant changes during black tea processing, which contributed to a complex scent of black tea. Stage-specific variations in FADVs are largely related to enzymatic-/thermally induced lipid decomposition during tea manufacturing.

3.6. Correlation analysis between changes in lipid molecules and FADVs

To evaluate the interactions between the key degraded lipid molecules and the significantly altered FADVs at different steps of black tea processing, chord diagrams showing the correlation network were constructed, as shown in Fig. S6. In the enzymatic oxidation stage (Fig. S6A), the glycoglycerolipid MGDG and glycerophospholipids PC and PE showed significant negative correlation with FADVs. During the thermal oxidation stage, lipid species belonging to TG, DG, PA, and PI were significantly negatively correlated with FADVs (Fig. S6B). Moreover, Pearson correlation between the individual lipid molecules (folds change <0.37 and 0.26 in enzymatic and thermal reaction stages, respectively) and FADVs compounds were further employed, as visualized in the correlation heatmap (Fig. 7). In the enzymatic reaction stage (Fig. 7A), MGDG (18:3/18:3, 18:3/18:2, 18:1/18:3, and 18:0/18:3) showed significant negative correlations with volatiles, such as (E)-2-hexenal, (E)-3-hexen-1-ol, cis-3-hexenyl salicylate, 2-pentyl-furan, and cis-jasmone. PC (18:0/18:3, 14:0/18:3, and 18:2/18:2) and PE (16:0/18:2, 18:3/18:2, and 18:2/18:2) also showed negative correlations with cis-jasmone, cis-3-hexenyl salicylate, (E)-2-decenal, and (E)-3-hexen-1-ol. In the thermal reaction stage, TG (16:0/18:2/18:3, 18:3/18:2/18:3, 18:3/18:3/18:3), DG (18:3/18:3), PA (18:3/18:2, 18:3/18:3), and PI (18:3/18:2, 18:3/18:3) presented obviously negative correlations with FADVs, such as fatty aldehydes (decenal, heptanal), alcohols (1-heptanol, 1-nonanol), esters (cis-3-hexenyl salicylate, hexanoic acid hexyl ester), alkanes (dodecane), and furan (2-pentyl-furan) (Fig. 7B).

Fig. 7.

Fig. 7

Open in a new tab

(A, B) Pearson correlation analysis of predominately down-regulated individual lipid molecules and FADVs in the enzymatic reaction stage (A) and thermal reaction stage (B) of black tea processing. Red and blue colors represent positive and negative correlations, respectively. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

By far, studies reporting the direct evidences of the relationship between individual lipids and volatiles are limited. For instance, Sekiya et al. (1984) verified that hexanal and (E)-2-hexenal were generated from linoleic and linolenic acids in tea fermentation. Zhou et al. (2014, 2015) reported that volatiles such as pentanal, hexanal, (E)-2-hexenal, (E,E)-2,4-decadienal, 1-octen-3-one, 2-nonenal, 2-pentyl-furan, and 1-pentanol were formed by the oxidative degradation of PC (18:0/18:2), TG (18:0/18:2), and PE (18:0/18:2) during thermal processing using a model system. Due to the complexity in black tea processing step and lipids composition, in-depth validation is needed to clarify the relationship between the individual lipid molecules and volatile compounds in tea processing.

3.7. Mapping the potential pathway of lipid degradation and aroma formation in black tea processing

As mentioned above, both tea lipids and their derived volatiles underwent stage-specific variations during black tea processing. Based on previous studies (Ho et al., 2015; Huang et al., 2024; Zhang et al., 2019b), it was supposed that the FADVs were mainly generated through two potential pathways in black tea processing, i.e., enzyme-driven and thermal-induced lipid oxidative degradation. The potential mechanism of lipid degradation and aroma formation in the two stages of black tea processing is depicted in Fig. 8.

Fig. 8.

Fig. 8

Open in a new tab

Potential pathway of lipid oxidative degradation and aroma formation in black tea processing through enzymatic-/thermal-induced oxidation.

In the enzyme-driven stages, the FADVs were proposed to be mainly formed through LOX-catalyzed reaction (Fig. 8). The predominant lipid precursors of MGDG, PC, and PE containing poly-unsaturated FA 18:2, FA 18:3, such as MGDG (18:3/18:2), PC (18:2/18:2), and PE (18:1/18:2) were possibly prone to be decomposed by lipase, resulting the release of FFA, LPC, and LPE, which was supported by the result of their changes in Table S5 and Fig. 4A. The released linoleic acid (FFA 18:2) and linolenic acid (FFA 18:3) were prone to be oxidized by 9-LOX and 13-LOX, respectively, to produce 9-hydroperoxide and 13-hydroperoxide (Ho et al., 2015; Yang et al., 2013; Huang et al., 2024). The hydroperoxides were further decomposed into volatiles by a series of enzymatic reactions (Chen et al., 2022; Liu et al., 2023).

During the thermal processing stages, FADVs were proposed to be mainly formed through a series of thermally induced free-radical chain reactions (Fig. 8) (Huang et al., 2024). In this study, unsaturated lipids such as TG (18:3/18:3/18:3), DG (18:3/18:3), PA (18:3/18:3), and PI (18:3/18:3) were identified as the major potential precursors for the generation of FADVs in the drying stage. These unsaturated lipids were prone to be oxidized into hydroperoxide radicals at the double bond sites (Zhou et al., 2022b). The hydroperoxide radicals were possibly decomposed by homolytic cleavage to produce alkoxyl radical, and the latter could be further degraded by the carbon-carbon scission (α-scission and β-scission), resulting the generation of volatile compounds (Zhou et al., 2014; Xia and Budge, 2017), such as aldehydes, alcohols, ketones, alkanes, alkenes, and furans, etc.

In addition, it has been reported that lipid degradation also participated in the Maillard rection via lipid-Maillard interaction to produce a myriad of volatile components, such as 2-pentylpyridine, 2,5-dimethyl-3-pentylpyrazine (Zhang et al., 2019b; Shahidi and Hossain, 2022). These compounds have been reported to contribute to food flavor such as cooked meat (Shahidi and Hossain, 2022). However, the role of lipid-Maillard interaction in black tea aroma formation has been rarely explored and needs to be further investigated.

4. Conclusion

To summarize, we present the first comprehensive report on the simultaneous characterization of dynamic changes in lipids and volatiles as well as their correlations during black tea processing using a multi-omics approach. Novel findings of stage-specific variations in lipids and FADVs in the enzymatic/thermal reaction phases, as well as the key potential aroma precursors in these two stages, were revealed for the first time. During the enzyme-driven processing steps, MGDG, PC, and PE-containing polyunsaturated FA 18:2 and FA 18:3 were identified as the major potential aroma contributors, which were possibly decomposed into aromatic compounds through LOX-catalyzed reactions. While in drying steps, TG, DG, PA, and PI species enriched in FA 18:2 and FA 18:3 were screened as the predominant aroma precursors, which were potentially transformed into volatile products by thermal degradative reactions. This study provides novel insights into lipids’ contribution to black tea flavor formation. In the future, more in-depth validation is needed to bridge the individual lipid molecules and volatile compounds by applying tools such as isotope labeling.

CRediT authorship contribution statement

Shan Zhang: Investigation, Formal analysis, Visualization, Writing – original draft. Le Chen: Investigation, Formal analysis, Visualization, Resources. Linchi Niu: Investigation, Resources. Haibo Yuan: Investigation, Methodology, Data curation. Xujiang Shan: Formal analysis, Visualization. Qianting Zhang: Investigation, Resources. Yuning Feng: Formal analysis, Validation. Qinghua Zhou: Methodology, Software, Writing – review & editing. Yongwen Jiang: Project administration, Supervision, Funding acquisition. Jia Li: Conceptualization, Investigation, Formal analysis, Supervision, Writing – review & editing.

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 work was supported by the Zhejiang Provincial Natural Science Foundation of China (No. ZCLZ24C1602), the Basal Research Fund of Tea Research Institute Chinese Academy of Agricultural Sciences (No. 1610212024011), the China Agriculture Research System of MOF and MARA (CARS-19), the Science and Technology Innovation Project of the Chinese Academy of Agricultural Sciences (No. CAAS-ASTIP-TRICAAS), and the Natural Science Foundation of China (No. 42277275).

Handling Editor: Professor A.G. Marangoni

Footnotes

Appendix A

Supplementary data to this article can be found online at https://doi.org/10.1016/j.crfs.2024.100910.

Contributor Information

Yongwen Jiang, Email: jiangyw@tricaas.com.

Jia Li, Email: jiali1986@tricaas.com.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Multimedia component 1
mmc1.docx (1.1MB, docx)

Data availability

Data will be made available on request.

References

  1. Chen C. Tea classification in theory and practice. J. Tea Business. 1981;1(2):329–344. [Google Scholar]
  2. Chen J., Zhang Y., Huang X., Wang H., Dong X., Zhu B., Qin L. Analysis of lipid molecule profiling and conversion pathway in Mandarin Fish (Siniperca chuatsi) during fermentation via untargeted lipidomics. J. Agric. Food Chem. 2023;71(22):8673–8684. doi: 10.1021/acs.jafc.3c00769. [DOI] [PubMed] [Google Scholar]
  3. Chen Q., Zhu Y., Dai W., Lv H., Mu B., Li P., Tan J., Ni D., Lin Z. Aroma formation and dynamic changes during white tea processing. Food Chem. 2019;274:915–924. doi: 10.1016/j.foodchem.2018.09.072. [DOI] [PubMed] [Google Scholar]
  4. Chen Q., Liu M., Li R., Jiang B., Liu K., Xiao Y., Wang Q., Wang T., Zhao L., Wang W., Liu Z., Chen L., Ma Y., Zhao M. Changes in lipids and medium-and long-chain fatty acids during the spontaneous fermentation of ripened pu-erh tea. Curr. Res. Food Sci. 2024;9 doi: 10.1016/j.crfs.2024.100831. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chen L., Zhang S., Feng Y., Jiang Y., Yuan Y., Shan X., Zhang Q., Niu L., Wang S., Zhou Q., Li J. Seasonal variation in non-volatile flavor substances of fresh tea leaves (Camellia sinensis) by integrated lipidomics and metabolomics using UHPLC-Q-Exactive mass spectrometry. Food Chem. 2024;462 doi: 10.1016/j.foodchem.2024.140986. Article 140986. [DOI] [PubMed] [Google Scholar]
  6. Chen X., Wang P., Wei M., Lin X., Gu M., Fang W., Zheng Y., Zhao F., Jin S., Ye N. Lipidomics analysis unravels changes from flavor precursors in different processing treatments of purple-leaf tea. J. Sci. Food Agric. 2021;102(9):3730–3741. doi: 10.1002/jsfa.11721. [DOI] [PubMed] [Google Scholar]
  7. Chen Q., Zhu Y., Liu Y., Liu Y., Dong C., Lin Z., Teng J. Black tea aroma formation during the fermentation period. Food Chem. 2022;374 doi: 10.1016/j.foodchem.2021.131640. Article 131640. [DOI] [PubMed] [Google Scholar]
  8. Dudareva N., Klempien A., Muhlemann J.K., Kaplan I. Biosynthesis, function and metabolic engineering of plant volatile organic compounds. New Phytol. 2013;198(1):16–32. doi: 10.1111/nph.12145. [DOI] [PubMed] [Google Scholar]
  9. Fahy E., Subramaniam S., Murphy R.C., Nishijima M., Raetz C.R.H., Shimizu T., Spener F., van Meer G., Wakelam M.J.O., Dennis E.A. Update of the LIPID MAPS comprehensive classification system for lipids. J. Lipid Res. 2009;50:9–14. doi: 10.1194/jlr.R800095-JLR200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Feng Z., Li Y., Li M., Wang Y., Zhang L., Wan X., Yang X. Tea aroma formation from six model manufacturing processes. Food Chem. 2019;285:347–354. doi: 10.1016/j.foodchem.2019.01.174. [DOI] [PubMed] [Google Scholar]
  11. Ho C.-T., Zheng X., Li S. Tea aroma formation. Food Sci. Hum. Wellness. 2015;4(1):9–27. doi: 10.1016/j.fshw.2015.04.001. [DOI] [Google Scholar]
  12. Hölzl G., Dörmann P. Structure and function of glycoglycerolipids in plants and bacteria. Prog. Lipid Res. 2007;46(5):225–243. doi: 10.1016/j.plipres.2007.05.001. [DOI] [PubMed] [Google Scholar]
  13. Hou Q., Ufer G., Bartels D. Lipid signalling in plant responses to abiotic stress. Plant Cell Environ. 2016;39(5):1029–1048. doi: 10.1111/pce.12666. [DOI] [PubMed] [Google Scholar]
  14. Huang F.-F., Yang P.-D., Bai S.-L., Liu Z.-H., Li J., Huang J.-A., Xiong L.-G. Lipids: a noteworthy role in better tea quality. Food Chem. 2024;431 doi: 10.1016/j.foodchem.2023.137071. [DOI] [PubMed] [Google Scholar]
  15. Li C., Li X., Huang Q., Zhou Y., Xu B., Wang Z. Influence of salt content used for dry-during on lipidomic profiles during the processing of water-boiled salted duck. J. Agric. Food Chem. 2020;68(13):4017–4026. doi: 10.1021/acs.jafc.0c01513. [DOI] [PubMed] [Google Scholar]
  16. Li J., Hua J., Yuan H., Deng Y., Zhou Q., Yang Y., Dong C., Zeng J., Jiang Y. Investigation on green tea lipids and their metabolic variations during manufacturing by nontargeted lipidomics. Food Chem. 2021;339 doi: 10.1016/j.foodchem.2020.128114. Article 128114. [DOI] [PubMed] [Google Scholar]
  17. Li J., Hua J., Zhou Q., Dong C., Wang J., Deng Y., Yuan H., Jiang Y. Comprehensive lipidome-wide profiling reveals dynamic changes of tea lipids during manufacturing process of black tea. J. Agric. Food Chem. 2017;65(46):10131–10140. doi: 10.1021/acs.jafc.7b03875. [DOI] [PubMed] [Google Scholar]
  18. Li J., Yang Y., Tang C., Yue S., Zhao Q., Li F., Zhang J. Changes in lipids and aroma compounds in intramuscular fat from Hu sheep. Food Chem. 2022;383 doi: 10.1016/j.foodchem.2022.132611. Article 132611. [DOI] [PubMed] [Google Scholar]
  19. Li J., Yao Y., Wang J., Hua J., Wang J., Yang Y., Dong C., Zhou Q., Jiang Y., Deng Y., Yuan H. Rutin, gamma-aminobutyric acid, gallic acid, and caffeine negatively affect the sweet-mellow taste of congou black tea infusions. Molecules. 2019;24(23):4221. doi: 10.3390/molecules24234221. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Li J., Yuan H., Rong Y., Qian M.C., Liu F., Hua J., Zhou Q., Deng Y., Zeng J., Jiang Y. Lipid metabolic characteristics and marker compounds of ripened Pu-erh tea during pile fermentation revealed by LC-MS-based lipidomics. Food Chem. 2023;404 doi: 10.1016/j.foodchem.2022.134665. [DOI] [PubMed] [Google Scholar]
  21. Liu H., Xu Y., Wu J., Wen J., Yu Y., An K., Zou B. GC-IMS and olfactometry analysis on the tea aroma of Yingde black teas harvested in different seasons. Food Res. Int. 2021;150 doi: 10.1016/j.foodres.2021.110784. [DOI] [PubMed] [Google Scholar]
  22. Liu Y., Chen Q., Liu D., Yang L., Hu W., Kuang L., Huang Y., Teng J., Liu Y. Multi-omics and enzyme activity analysis of flavour substances formation: major metabolic pathways alteration during Congou black tea processing. Food Chem. 2023;403 doi: 10.1016/j.foodchem.2022.134263. [DOI] [PubMed] [Google Scholar]
  23. Matyash V., Liebisch G., Kurzchalia T.V., Shevchenko A., Schwudke D. Lipid extraction by methyl-tert-butyl ether for high-throughput lipidomics. J. Lipid Res. 2008;49(5):1137–1146. doi: 10.1194/jlr.D700041-JLR200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Ramaswamy R., Ramaswamy P. Lipid occurrence, distribution and degradation to flavour volatiles during tea processing. Food Chem. 2000;68:7–13. doi: 10.1016/S0308-8146(99)00143-0. [DOI] [Google Scholar]
  25. Sekiya J., Kajiwara T., Hatanaka A. Seasonal changes in activities of enzymes responsible for the formation of C6-aldehydes and C6-alcohols in tea leaves, and the effects of environmental temperatures on the enzyme activities. Plant Cell Physiol. 1984;25:269–280. doi: 10.1093/oxfordjournals.pcp.a076711. [DOI] [Google Scholar]
  26. Shahidi F., Hossain A. Role of lipids in food flavor generation. Molecules. 2022;27(15):5014. doi: 10.3390/molecules27155014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Shan X., Deng Y., Niu L., Chen L., Zhang S., Jiang Y., Yuan H., Wang Y., Li J. The influence of fixation temperature on Longjing tea taste profile and the underlying non-volatile metabolites changes unraveled by combined analyses of metabolomics and E-tongue. LWT--Food Sci. Technol. 2024;191 doi: 10.1016/j.lwt.2023.115560. [DOI] [Google Scholar]
  28. Sharma V., Rao L.J. A thought on the biological activities of black tea. Crit. Rev. Food Sci. Nutr. 2009;49(5):379–404. doi: 10.1080/10408390802068066. [DOI] [PubMed] [Google Scholar]
  29. Shevchuk A., Megías-Pérez R., Zemedie Y., Kuhnert N. Evaluation of carbohydrates and quality parameters in six types of commercial teas by targeted statistical analysis. Food Res. Int. 2020;133 doi: 10.1016/j.foodres.2020.109122. [DOI] [PubMed] [Google Scholar]
  30. Wan X. third ed. ed. China Agriculture Publishing House; Beijing: 2003. Biochemistry of Tea. [Google Scholar]
  31. Wang H., Hua J., Yu Q., Li J., Wang J., Deng Y., Yuan H., Jiang Y. Widely targeted metabolomic analysis reveals dynamic changes in non-volatile and volatile metabolites during green tea processing. Food Chem. 2021;363 doi: 10.1016/j.foodchem.2021.130131. [DOI] [PubMed] [Google Scholar]
  32. Wang H., Ouyang W., Yu Y., Wang J., Yuan H., Hua J., Jiang Y. Analysis of non-volatile and volatile metabolites reveals the influence of second-drying heat transfer methods on green tea quality. Food Chem. X. 2022;14 doi: 10.1016/j.fochx.2022.100354. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Wang H., Wu Y., Xiang H., Sun-Waterhouse D., Zhao Y., Chen S., Li L., Wang Y. UHPLC-Q-Exactive Orbitrap MS/MS-based untargeted lipidomics reveals molecular mechanisms and metabolic pathways of lipid changes during golden pomfret (Trachinotus ovatus) fermentation. Food Chem. 2022;396 doi: 10.1016/j.foodchem.2022.133676. Article 133676. [DOI] [PubMed] [Google Scholar]
  34. Wu S., Yu Q., Shen S., Shan X., Hua J., Zhu J., Qiu J., Deng Y., Zhou Q., Jiang Y., Yuan H., Li J. Non-targeted metabolomics and electronic tongue analysis reveal the effect of rolling time on the sensory quality and nonvolatile metabolites of congou black tea. LWT--Food Sci. Technol. 2022;169 doi: 10.1016/j.lwt.2022.113971. Article 113971. [DOI] [Google Scholar]
  35. Xia W., Budge S.M. Techniques for the analysis of minor lipid oxidation products derived from triacylglycerols: epoxides, alcohols, and ketones. Compr. Rev. Food Sci. Food Saf. 2017;16(4):735–758. doi: 10.1111/1541-4337.12276. [DOI] [PubMed] [Google Scholar]
  36. Xie J., Wang Q., Cui H., Wang L., Deng Y., Yuan H., Zhu J., Yang Y., Jiang Y. Characterization of Gardenia tea based on aroma profiles using GC-E-Nose, GC-O-MS and GC × GC-TOFMS combined with chemometrics. Beverage Plant Res. 2024;4(1) doi: 10.48130/bpr-0023-0034. [DOI] [Google Scholar]
  37. Xu T., Hu C., Xuan Q., Xu G. Recent advances in analytical strategies for mass spectrometry-based lipidomics. Anal. Chim. Acta. 2020;1137:156–169. doi: 10.1016/j.aca.2020.09.060. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Yang Z., Baldermann S., Watanabe N. Recent studies of the volatile compounds in tea. Food Res. Int. 2013;53(2):585–599. doi: 10.1016/j.foodres.2013.02.011. [DOI] [Google Scholar]
  39. Yang Y., Hua J., Deng Y., Jiang Y., Qian M.C., Wang J., Li J., Zhang M., Dong C., Yuan H. Aroma dynamic characteristics during the process of variable-temperature final firing of Congou black tea by electronic nose and comprehensive two-dimensional gas chromatography coupled to time-of-flight mass spectrometry. Food Res. Int. 2020;137 doi: 10.1016/j.foodres.2020.109656. Article 109656. [DOI] [PubMed] [Google Scholar]
  40. Yu Z., Ye L., He Y., Lu X., Chen L., Dong S., Xiang X. Study on the formation pathways of characteristic volatiles in preserved egg yolk caused by lipid species during pickling. Food Chem. 2023;424 doi: 10.1016/j.foodchem.2023.136310. Article 136310. [DOI] [PubMed] [Google Scholar]
  41. Zeng L., Watanabe N., Yang Z. Understanding the biosynthesis and stress response mechanisms of aroma compounds in tea (Camellia sinensis) to safely and effectively improve tea aroma. Crit. Rev. Food Sci. Nutr. 2018;59(14):2321–2334. doi: 10.1080/10408398.2018.1506907. [DOI] [PubMed] [Google Scholar]
  42. Zeng L., Zhou Y., Fu X., Liao Y., Yuan Y., Jia Y., Dong F., Yang Z. Biosynthesis of jasmine lactone in tea (Camellia sinensis) leaves and its formation in response to multiple stresses. J. Agric. Food Chem. 2018;66(15):3899–3909. doi: 10.1021/acs.jafc.8b00515. [DOI] [PubMed] [Google Scholar]
  43. Zhang G., Wei J., Li L., Cui D. Lipidomics, transcription analysis, and hormone profiling unveil the role of CsLOX6 in MeJA biosynthesis during black tea processing. Horticul. Res. 2024;11(3) doi: 10.1093/hr/uhae032. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
  44. Zhang L., Ho C.T., Zhou J., Santos J.S., Armstrong L., Granato D. Chemistry and biological activities of processed Camellia sinensis teas: a comprehensive review. Compr. Rev. Food Sci. Food Saf. 2019;18(5):1474–1495. doi: 10.1111/1541-4337.12479. [DOI] [PubMed] [Google Scholar]
  45. Zhang M., Yang Y., Yuan H., Hua J., Deng Y., Jiang Y., Wang J. Contribution of addition theanine/sucrose on the formation of chestnut-like aroma of green tea. LWT--Food Sci. Technol. 2020;129 doi: 10.1016/j.lwt.2020.109512. [DOI] [Google Scholar]
  46. Zhang S., Shan X., Niu L., Chen L., Wang J., Zhou Q., Yuan H., Li J., Wu T. The integration of metabolomics, electronic tongue, and chromatic difference reveals the correlations between the critical compounds and flavor characteristics of two grades of high-quality Dianhong congou black tea. Metabolites. 2023;13(7):864. doi: 10.3390/metabo13070864. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Zhang S., Wu S., Yu Q., Shan X., Chen L., Deng Y., Hua J., Zhu J., Zhou Q., Jiang Y., Yuan H., Li J. The influence of rolling pressure on the changes in non-volatile compounds and sensory quality of congou black tea: the combination of metabolomics, E-tongue, and chromatic differences analyses. Food Chem. 2023;X 20 doi: 10.1016/j.fochx.2023.100989. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Zhang W., Cao X., Liu S. Aroma modulation of vegetable oils—a review. Crit. Rev. Food Sci. Nutr. 2019;60(9):1538–1551. doi: 10.1080/10408398.2019.1579703. [DOI] [PubMed] [Google Scholar]
  49. Zhou L., Zhao M., Bindler F., Marchioni E. Comparison of the volatiles formed by oxidation of phosphatidylcholine to triglyceride in model systems. J. Agric. Food Chem. 2014;62(33):8295–8301. doi: 10.1021/jf501934w. [DOI] [PubMed] [Google Scholar]
  50. Zhou L., Zhao M., Bindler F., Marchioni E. Identification of oxidation compounds of 1-stearoyl-2-linoleoyl-sn-glycero-3-phosphoethanolamine during thermal oxidation. J. Agric. Food Chem. 2015;63(43):9615–9620. doi: 10.1021/acs.jafc.5b03753. [DOI] [PubMed] [Google Scholar]
  51. Zhou Z., Wu Q., Yang Y., Hu Q., Wu Z., Huang H., Lin H., Lai Z., Sun Y. The dynamic change in fatty acids during the postharvest process of oolong tea production. Molecules. 2022;27(13):4298. doi: 10.3390/molecules27134298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Zhou Z., Li Y.-L., Zhao F., Xin R., Huang X.-H., Zhang Y.-Y., Zhou D., Qin L. Unraveling the thermal oxidation products and peroxidation mechanisms of different chemical structures of lipids: an example of molecules containing oleic acid. J. Agric. Food Chem. 2022;70(51):16410–16423. doi: 10.1021/acs.jafc.2c06221. [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Multimedia component 1
mmc1.docx (1.1MB, docx)

Data Availability Statement

Data will be made available on request.

Articles from Current Research in Food Science are provided here courtesy of Elsevier

RESOURCES