Food Research International
journal homepage: www.elsevier.com/locate/foodres
Food Research International 147 (2021) 110472
Changes of fungal community and non-volatile metabolites during Image pile-fermentation of dark green tea
Shuai Hu 1, Chang He 1, Yuchuan Li , Zhi Yu, Yuqiong Chen, Yaomin Wang *, Dejiang Ni *
Key Laboratory of Horticulture Plant Biology, Ministry of Education, College of Horticulture & Forestry Sciences, Huazhong Agricultural University, Wuhan, Hubei 430070, People’s Republic of China
Key Laboratory of Urban Agriculture in Central China, Ministry of Agriculture, Wuhan, Hubei 430070, People’s Republic of China
A R T I C L E I N F O
Dark green tea Pile-fermentation
IIIumina MiSeq sequencing
Fungal community, non-volatile components flavor ingredients Caffeine (PubChem CID: 2519)
Epigallocatechin (PubChem CID: 72277) Epigallocatechin gallate (PubChem CID: 65064)
L-isoleucine (PubChem CID:6306)
L-theanine (PubChem CID: 439378) Epicatechin gallate (PubChem CID: 107905) L-phenylalanine (PubChem CID: 6140) Gallic acid (PubChem CID: 370)
L-valine (PubChem CID: 6287) Gallocatechin (PubChem CID: 65084)
A B S T R A C T
Fungal community and non-volatile metabolites changes during the pile-fermentation are key factors to organoleptic qualities of dark green tea. However, the correlation between fungal succession and non-volatile compounds has never been satisfactorily explained. The purpose of the present study was to investigate fungal succession and its correlation with flavor compounds by multi-omics. Illumina Miseq sequencing of ITS1 region was conducted to analyze the fungal succession, a total of 78 OTUs which consisted of one phyla, nine classes, 15 orders, 26 families, 37 genera were identified, with Ascomycota as dominant phyla. Cluster analysis and non- metric multidimensional scaling of samples demonstrated the distribution of OTUs in multi-dimensional space, the pile-fermentation process of dark green tea can be divided into four periods according to the gener- ated trajectory of fungal population, S0, S1-S3, S4-S5, and S6. Aspergillus is the dominant genus. Penicillium, Cyberlindnera, Debaryomyces, Candida, Thermomyces, Rasamsonia, Thermoascus, and Byssochlamys appear in different periods. three alkaloids, seven catechins, nine amino acids, five organic acids, five flavones and flavonoid glycosides were identified by UPLC-QTOF-MS/MS, and the contents were all decreasing. Caffeine, EGC, EGCG, L-theanine, kaempferitrin, L-phenylalanine, gallic acid, and myricetin-3-O-galactoside are important ingredients which contribute to the flavor of dark green tea.
This study demonstrated the fungal succession, non- volatile flavor compounds and their relationships during pile-fermentation of dark green tea, and provides new insights into evaluating pivotal role of fungal succession in the manufacturing process of dark green tea.
Tea is typically categorized into green, white, yellow, black, and dark tea according to the fermentation degree. Unlike the other five types of tea, dark tea is produced by solid-state fermentation of tea leaves, which causes its special sensory quality. Dark tea varieties can be classified into Fu brick tea, Pu-erh tea, Liubao tea, and Dark green tea due to the origin and processing technologies. The differences in the grade of raw mate- rials and pile-fermentation processes affect the composition of fer- menting microorganisms and subsequently the quality of dark tea varieties (Zhu et al., 2020). Dark tea is mainly produced in Hubei, Hunan, Sichuan, Shanxi, Yunnan, and Guangxi provinces in China. In the past, dark tea was mainly consumed by pastoralists in border regions such as Qinghai, Tibet, and Mongolia (Zhang, Zhang, Zhou, Ling, & Wan, 2013). In recent years, emerging research has recognized thecritical role of dark tea in weight loss and blood lipid reduction and has, anti-obesity, antibacterial, anti-oxidation, anti-tumor and other healtheffects. Its consumption has expanded to mainland China, Hong Kong, Macau, Taiwan, Russia, Europe, and many other countries (Cheng et al., 2015). In China, the quality of dark tea varies depending on the envi- ronment of the origin and the processing technology. Dark green tea is a kind of dark tea, also named “Qingzhuan Tea”, which is native to Chibi of Hubei province, is made using sun-dried leaves of Camellia sinensis var. sinensis. Compared with other types of dark tea, the raw materials for producing dark green tea are coarser, the pile-fermentation needs higher moisture and water content. In addition, the pile-fermentation needs higher temperature and longer time (Zhu et al., 2020). These differences in raw materials and the pile-fermentation affect the chem- ical composition, and ultimately contribute to the beneficial effect on reducing blood lipids, weight loss, and other health properties (Feng
* Corresponding authors.
E-mail addresses: w[email protected] (Y. Wang), [email protected] (D. Ni).
1 Co-first author.
Received 5 May 2020; Received in revised form 17 May 2021; Accepted 23 May 2021
Available online 31 May 2021
0963-9969/© 2021 Elsevier Ltd. All rights reserved.et al., 2020; Lv, Zhang, Shi, & Lin, 2017).
The formation of beneficial components in dark tea is related to the key process called pile-fermentation. Pile-fermentation refers to the process that uses loose tea as substrates. Under the humid conditions and the involved microorganisms, a series of reactions such as oxidation, hydrolysis, polymerization, and secondary metabolic transformation occur, which accelerates changes in tea ingredients and further form the unique quality of dark tea (Zhang et al., 2013; Zhu et al., 2020). Chemical ingredients experience complex changes during the pile- fermentation due to the microorganisms and humidity, and tea cate- chins are oxidatively polymerized to form catechin derivatives. For instance, catechin derivatives Fuzhuanin A-F and Puerins C-F are formed during the fermentation process of Fu brick tea and Pu-erh tea, respec- tively (Luo et al., 2013; Zhu et al., 2015). The newly formed components during pile-fermentation have potential health effects. Studies have demonstrated that teasperol isolated from Fu brick tea have anti- bacterial effects, while another new flavonoid, camellikaempferoside A, can inhibit the proliferation of cancer cells (Luo et al., 2013; Tian et al., 2016). These new derivatives produced during pile-fermentation play an important role in the health functions and unique organoleptic qualities of dark tea.
The research on pile-fermentation technologies mainly focuses on
the changes of microorganisms and the chemical components. In recent years, modern molecular biology technologies such as polymerase chain reaction (PCR), denaturing gradient gel electrophoresis (DGGE), and the 16sRNA clone library method have been applied to the study of dark tea (Zhao et al., 2019). Moreover, rapid development of high-throughput
inside part of dark green tea brick. The process of pile-fermentation was as follows: 15 tons of raw green tea was piled up with a length of 15 m, a width of 10 m and a height of 2 m pile in a fermentation room. Thermo hygrometers were placed in locations of (10, 20, 30, 40, 50, 80 cm deep) at the center of each side of the fermentation pile. The overall pile- fermentation lasted for 37 days, and samples were collected every 5 days and one day before the turning process. The day before the fermentation started was recorded as day 0, with the pile-fermentation period considered as S0, and S1 represented the second time when collecting sample. Thus, a total of 7 pile-fermentation periods marked as S0–S6 were inferred in turn, with time points of turning process were at S2, S3, S4, and S5, respectively. Samples were divided triplicate and stored in —80 ◦C refrigerator before use.
2.2. DNA extraction and Illumina MiSeq sequencing
Genomic DNA was extracted from tea samples using E.Z.N.A.® soil kit (Omega Bio-tek, USA) according to the instructions (Hu et al., 2021). DNA concentration and purity were detected using NanoDrop2000, and extraction quality was detected by 1% agarose gel electrophoresis for further amplicon generation. Forward primer (5′-CTTGGTCATTTA- GAGGAAGTAA-3′) and the revise primer (5′-GCTGCGTTCTTCATC- GATGC-3′) were used for amplification of ITS1 region of fungi. PCR products were recovered on a 2% agarose gel after amplification, then purified using AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA), eluted with Tris-HCl, and detected by 2% agarose electrophoresis. PCR products were quantified using QuantiFluor ™ -STsequencing (NGS) technology, metagenomics and metaproteomics
(Promega, USA) and mixed in proportion to the sequencing re-have contributed greatly to this growing area of research by exploring the dynamic changes of microorganisms and metabolite profiles including volatile and non-volatile compounds during the fermentation process of dark tea (Li et al., 2020; Mao, Wei, Teng, Huang, & Xia, 2017; Shi et al., 2021; Zhang et al., 2021). These technologies have been actively applied to study of the fungal community of dark tea (Rui et al., 2019). Zhao et al. applied metagenomics and metaproteomics to study the fungal community in the pile-fermentation of Pu-erh tea and found that Aspergillus was the dominant genus during the overall pile- fermentation process among the detected fungal microorganisms (Zhao et al., 2015). Research on the flavor related metabolites during pile-fermentation of dark tea is another hot topic (Long et al., 2020; Zhang et al., 2021). Metagenomic technology has been applied to mul- tiple areas of dark tea research. The changes of metabolites and sensory quality during Qingzhuan Tea processing was evaluated through LC-MS- based metabolomics (Cheng et al., 2020). The secondary metabolites changes during post-fermentation of Qingzhuan tea was determined. Up to now, few studies have focused on the correlation between fungal community and non-volatile compounds during pile-fermentation pro- cesses of dark tea.
Herein, dark green tea produced was used in this study. High-
throughput sequencing technology was used to evaluate the changes of the fungal succession during the pile-fermentation process in dark green tea. The changes of non-volatile components were identified by UPLC-QTOF-MS/MS and the effect of fungal community changes on flavor substances were revealed by the Pearson’s correlation analysis. This study offers some important insights into the pile-fermentation process and makes a major contribution to research on fungal commu- nities and non-volatile metabolites to the quality of dark green tea.
2. Material and methods
2.1. Sample collection
The experimental samples were collected from Hubei Chibi Zhaoli- qiao Tea Co., Ltd., and raw materials were classified as third grade tea, fresh mature leaves with mainly red stems was used for making dried tea, which is typically considered as the raw material for making the
quirements. The pooled PCR products were used to construct a PE 2 * 300 library from the purified amplified fragments according to Illumina MiSeq platform (Illumina, San Diego, USA) standard operating proced- ures. Sequencing was performed using Illumina’s Miseq PE300 platform (Shanghai Meiji Biomedical Technology Co., Ltd.).
2.3. Detection of non-volatile components by UPLC-QTOF-MS/MS
0.15 g of ground tea sample was first accurately weighted, then,
7.5 mL of 75% (v/v) methanol aqueous solution (containing 150 μL internal standard, ethoxyphylline 37.5 μg/mL) was added, Extraction was then performed at 70 ◦C in a water bath for 30 min, followed by cooling to room temperature and centrifuging at 4000g for 5 min. Finally, 1.5 mL liquid was filtered through a 0.22 μm membrane for determination. Quality control (QC) samples were prepared by mixing 100 μL of each sample together for evaluating the stability and repro- ducibility of the metabolomic analysis.
Non-volatile compounds were determined by the Agilent UPLC sys- tem equipped with UV detector. Chromatographic separation was con- ducted with a Zorbax SB-Aq column (2.1 mm 100 mm, 1.8 μm) maintained at 60 ◦C. The mobile phase consisted of solvent A (water with 0.1% formic acid, v/v) and solvent B (acetonitrile). A liner gradient at a flow rate of 0.3 mL/min was applied for chromatographic separa- tion: 0–2 min: 5% B, 2–15 min: 5% B-35% B, 15–20 min: 35% B-100% B,
20–25 min: 100% B, 25.01–33 min: 5% B. The injection volume was 5 μL. The MS positive ESI mode was conducted with the following pa- rameters: 3.5 kV (capillary voltage), 300 ◦C (drying gas temperature),
8.0 L/min (flow rate), 35 psi (nebulizer pressure), 350 ◦C (sheath gas temperature) and 11 L/min (flow rate), m/z 100–1200 (mass range). Non-volatile compounds were identified by automatic UPLC-MS/MS analysis at 10, 20, and 30 V (collision-induced dissociation voltage).
2.4. Identification of non-volatile components
Raw data obtained by UPLC-QTOF/MS/MS was converted by Proteo Wizard 3.0, XC-MS was then used to analyze the converted data to obtain the original matrix data table. The original data matrix includes sample name (name), mass-to-charge ratio (m/z), and retention time(RT). MetaboAnalyst (https://www.metaboanalyst.ca/) was used for normalization, and the normalized data matrix obtained was used for further analysis. Qualitative analysis was performed based on mass-to- charge ratio (m/z) and tea plant secondary metabolism standard data- base. Quantification of non-volatile components were used according reported methods (Dai et al., 2015; Yu et al., 2020).
The content of detected compounds was calculated as following:
Relative content (mg/gDW) = peak area(C)/peak area (S)*37.5/W
C: compound; S: internal stands; W: sample dry weight; the concentra- tion of internal standard: 37.5 μg/ml.
2.5. Determination of main chemical compositions in tea samples
Total polyphenol content of dark green tea was detected by colori- metric method using Folin-Ciocalteu reagent according to National Standard of the People’s Republic of China (Li et al., 2017). The content of total free amino acids was measured by ninhydrin assay according to National Standard of the People’s Republic of China (Yin et al., 2021). The water extract content of dark green samples and total organic acid were measured according to the reported protocol (Gong et al., 2020; Lai et al., 2016; Zhang et al., 2020). The content of total theaflavin, thear- ubigins and theabrownine was detected by a previously published method (Song et al., 2020). The content of soluble polysaccharides was measured using Anthrone assay (Fu, Xie, Nie, Zhou, & Wang, 2001).
2.6. Data analysis
Raw reads were demultiplexed and quality-filtered by Trimmomatic
software and spliced by FLASH software using the following criteria: (a) Set a window of 50 bp, if the average quality score is <20, truncate the back-end bases from the window and remove sequences <50 bp. (b) Barcode needs to match exactly, and primers allow mismatch of 2 bases
to remove fuzzy bases. (c) The sequences at both ends were spliced ac- cording to the overlap which was greater than 10 bp, and sequences that cannot be spliced were remove. UPARSE software (version 7.1 http:// drive5.com/uparse/) was used to classify sequences based on 97% similarity and classify operational taxonomic unit (OTU). UCHIME software was used to eliminate chimeras. RDP classifier (http://rdp.cme
.msu.edu/) was used to annotate the species classification of each sequence, and the Unite database was compared, with alignment threshold set at 70%.
Based on this OTU division, the diversity index analysis such as alpha diversity (Shannon and Simpson), richness (ACE and Chao1), Rarefaction curve, Shannon-Wiener curve, Venn diagram, and detection of sequencing depth were performed. In addition, the classification information of each level from phylum to genus was ob- tained from each classification operation unit. Based on the taxonomy information, the statistical analysis of the community structure could be performed at each taxonomic level. Statistical and visual analysis such as community structure were then performed. R package vegan (version 3.4.0, https://mirrors.tuna.tsinghua.edu.cn/CRAN/) was used to generate non-metric multidimensional scaling (NMDS) diagram and ANOSIM & ADONIS diagrams.
Principal component analysis (PCA) was used to reduce the dimen-
sion of the multivariate data matrix in unsupervised mode, and the similarities and differences between samples were determined by observing the degree of aggregation between sample points. The iden- tification of differential metabolites was assessed by projection on latent structure-discriminant analysis (PLS-DA) analysis. VIP values were used to discriminate differential metabolites, and then a Kruskal-Wallis test
was used to perform a one-dimensional test, p < 0.05. SPSS 19.0 pro- gram was used for one-way analysis of variance (ANOVA), Pearson’s
correlation coefficients and P value calculation, and significant differ- ences were evaluated on the Fisher’s least significant difference and Duncan’s multiple range test. All results were expressed as mean ± SEM,
and the value of p < 0.05 was considered to be significant.
3.1. Analysis of fungal community during different pile-fermentation processes
The overall pile-fermentation processes lasted for 37 days and can be divided into 7 periods (Fig. 1A). We first analyzed the fungal community of different pile-fermentation processes of dark green tea. Data obtained by Illumina MiSeq sequencing revealed that the fungal sequencing depth index coverage of each sample was greater than 99.90%. It is known from the fungal community dilution curve and Shannon-Wiener curve that as the amount of sequencing data increases, the number of OTU species increases first, then the curve gradually decreases, and finally flattens (Fig. 1B and C).
There were 78 OTUs in the fungal community, which were classified into 1 phyla, 9 classes, 15 orders, 26 families and 37 genera. Ascomycota was the predominant fungus at the phylum level in all experimental samples with relative abundance greater than 99.00%. Furthermore, 11
important species (relative abundance >1.00%) were identified among 37 genera of the fungal community through comparative analysis with
the Unite database.
Aspergillus was the absolute dominant genus during the overall pile-
fermentation process of dark green tea, however, the relative abundance decreased from 75.10% in S0 to 38.10% in S6 (p < 0.05). Meanwhile, the relative abundance of Byssochlamys was low during S0 to S2, and increased to 3.60% in S3 and then significantly increased to 57.20% in
S6. Rasamsonia appeared in S1 with a relative abundance of 33.30% and significantly decreased to 2.70% in S6. Similarly, the relative abundance of Thermomyces increased to 49.00% in S1 and significantly decreased to
<1.00% in S6 (p < 0.05). The fungi that significantly decreased from S0 to S6 were Penicillium (14.90% to <1.00%), Cyberlindnera (5.90% to
<1.00%), Debaryomyces (2.40% to <1.00%), Candida (1.40% to almost 0.00%) (p < 0.05). Relative abundance of Sordariomycetes-unclassified and Eurotiales-unclassified were about 1.00% throughout the pile-
fermentation process (Fig. 1D).
3.2. Analysis of fungal succession differences in different pile- fermentation processes
To evaluate the similarity of the fungal community composition at different periods of pile-fermentation of dark green tea, cluster analysis and non-metric multidimensional scaling (NMDS) of samples from different fermentation processes were used based on the Bray-Curtis algorithm. According to the distribution of OTUs in multi-dimensional space, the generated trajectory could be divided into four stages as the fermentation progresses: the first stage was S0, which was classified as group I, the second period was S1-S3 and considered as group II, the third period was S4-S5, which was group III, the fourth period was S6, which was group IV (Fig. 2A). ANOSIM/Adonis analysis based on the Bray-Curtis algorithm and a non-parametric test between different groups Adonis analysis showed the groupings were reasonable. The distribution of the distance between the groups and samples within the four different groups at the genus level showed that the sample distances within the group were all lower than the sample distance among the groups. This indicated that the groupings were reasonable, and the difference among the groups was greater than the difference within a group, which was statistically significant (Fig. 2B). We further used the PCoA method to reanalyze the fungal OTU levels. Similar to the previous results, PCoA demonstrated high inter-group differences in seven
different stages, indicating that the seven processing stages are clearly
distinguishable. According to the distribution of PCoA1 and PCOA2,=four different stages can be obtained. PCoA distribution of four groups (R 0.73, p < 0.05) indicate that the grouping has significant differ- ences among groups, which further indicate that the changes in fungal
Analysis of fungal succession during different pile-fermentation processes. (A) General diagram for pile-fermentation process of dark green tea. (B) Rare- faction curves of fungal community in dark green tea pile-fermentation. (C) Shannon-Wiener curves of fungal community in dark green tea pile-fermentation. (D) Distribution of fungal community in genus level during the pile-fermentation process of dark green tea. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
OTU levels can be divided into four different periods, S0, S1–S3, S4–S5, and S6 (Fig. 2 C and D).
Samples were further divided into four groups based on NMDS analysis. The results demonstrated 25 of the identified OTUs were shared among all these four groups. OTUs shared between the groups in group I and group II, group II and group III, and group III and group IV were 43, 45, and 34, respectively (Fig. 2E).
Thus, based on the above groupings, Aspergillus was present throughout the pile-fermentation process and was the dominant species. In group I, the relative abundance content of Aspergillus was 75.10%, which decreased to 37.10% in group II, then increased to 53.20% in group III, and finally decreased to 38.10% at the end of pile- fermentation. Penicillium, Cyberlindnera, and Debaryomyces only appeared in Group I, with relative abundances of 14.90%, 5.90% and Analysis of fungal succession differences in different pile-fermentation processes. (A) Analysis of non-metric multidimensional scaling in different fermen- tation stages of dark green tea. (B) Genus level distance between groups and intra-group with samples of four groups. (C) Principal coordinates analysis of fungal succession differences in different fermentation phases. (D) Principal coordinates analysis of fungal succession differences in different stage. R value is a statistic of ANOSIM; P value describes the significance of different groups. (E) Analysis of OTU Venn of different groups. (F) Distribution of fungal community in genus level during four different pile-fermentation periods of dark green tea. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
2.40%, respectively, and the content in the latter groups was<1.00%. A large number of Rasamsonia and Thermomyces appeared in Group II,
with abundances of 27.60% and 30.90%, Rasamsonia showed an overall decreasing trend as the fermentation progressed and decreased to 2.70% at the end of the fermentation, while Thermomyces abundance was
<1.00% in Group III and continued until the end of fermentation.
Byssochlamys appeared in Group II with a relative abundance of 1.20% and gradually became the dominant genus in the later periods of pile- fermentation, reaching 57.20% in Group IV. In Group II, due to changes of the inside fermentation conditions of the pile, a variety of fungal genera began to appear, and the content changed with the changes of different floras during pile-fermentation (Fig. 2F).
3.3. Changes of non-volatile components during pile-fermentation
UPLC-QTOF-MS/MS was used for non-target metabolic profiling analysis. During the pile-fermentation, 29 non-volatile components were detected, including three alkaloids, seven catechins, nine amino acids, five flavonoids and flavonoid glycosides, and five organic acids, the detailed information of these 29 non-volatile components were shown in Supplementary Table 2.
The contents of the three alkaloids in S0 were caffeine, theophylline,and theobromine. With the progress of pile-fermentation, a significant 37.50% decrease in alkaloids was observed at S6 compared with S0 (p < 0.05). Caffeine accounted for 97.00% of the total alkaloids content,
which decreased from 8835.67 ± 260.68 μg/g in S0 to5553.23 ± 87.98 μg/g in S6 (p < 0.05). The content of catechins was decreased significantly from 7969.98 ± 346.36 μg/g in S0 to
233.98 20.294 μg/g in S6 (p < 0.05). A 97.70% decrease of EGC was observed at S6 compared with S0 (p < 0.05). A 97.30% decrease of EGCG was observed at S6 compared with S0 (p < 0.05). The content of ECG was decreased significantly from 924.98 ± 42.45 μg/g in S0 to
25.58 2.44 μg/g in S6 (p < 0.05). Meanwhile, the content of GC, GCG, EC, and C all decreased significantly during the pile-fermentation (p < 0.05). The contain of amino acids gradually decreased in early S3 and then continued to decrease to the end of pile-fermentation. The content of theanine significantly decreased from 1194.28 ± 133.54 μg/g in S0 to 28.82 2.68 μg/g in S6 (p < 0.05) (Fig. 3C).
A 70.70% decrease in flavonoids and flavonoid glycosides was observed at S6 compared with S0 (p < 0.05) (Fig. 3B), among which myricetin-3-O-galactoside decreased from 172.29 ± 7.81 μg/g in S0 to
30.06 ± 3.86 μg/g in S6 (p < 0.05). Kaempferitrin decreased from
76.46 ± 3.13 μg/g in S0 to 19.24 ± 1.13 μg/g in S6 (p < 0.05). Quercetin
decreased from 35.06 1.77 μg/g in S0 to 2.44 0.18 μg/g in S6. A 46.60% decrease in vitexin and a 17.50% decrease in kaempferol werobserved at S6 compared to S0 (p < 0.05) (Fig. 3C).
The content of organic acid decreased significantly from 194.98 16.40 μg/g in S0 to 80.24 1.25 μg/g in S6 (p < 0.05). Gallic acid accounted for 80.10% of the total organic acids, and a 61.80% decrease of gallic acid was observed at S6 compared with S0 (p < 0.05) (Fig. 3B and C).
In this study, we also measured the levels of water extract, thea- flavin, thearubigins, theabrownine, free amino acids, total polyphenol, soluble polysaccharides, and total organic acids using spectrophotom-etry. We found that water extract began to significantly decrease at S2 and continued to decrease until the end of pile-fermentation (p < 0.05). The level of total theaflavin ranged from 0.13 0.01% to 0.16 0.02%, but no significant difference was observed. The level of total thear- ubigins decreased significantly from 4.16 0.42% in S0 to 1.97 0.09%in S6 (p < 0.05), while theabrownine slightly increased. The level of total free amino acids significantly decreased from 1.69 0.18% in S0 to0.63 0.07% in S6 (p < 0.05). The level of total polyphenol significant decreased from 9.53 0.29% in S0 to 2.81 0.22% in S6 (p < 0.05). Soluble polysaccharides significantly increased from 1.18 0.11% in S0 to 2.42 0.05% in S6 (p < 0.05). The level of organic acids ranged from
3.69 0.09% to 4.28 0.03%, but no significant difference was observed between S0 and S6 (Supplementary Table 1).
Changes of non-volatile components during pile-fermentation. (A) Changes of alkaloids, catechins, and amino acids in the pile-fermentation process of dark green tea. (B) Changes of flavonoid glycosides, and organic acids in the pile-fermentation process of dark green tea. (C) Changes of non-volatile components in different pile-fermentation process of dark green tea. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
3.4. Screening and analysis of representative non-volatile components during pile-fermentation
As mentioned above that fungal community can be grouped to group I – group IV based on NDMS analysis, S1–S3, S4–S5 have similar dis- tributions in the horizontal axis. Combining the vertical axis distribution and the sequence of samples collection time, we regard S0 as group I, S1–S3 as group II, S4–S5 as group III, S6 as group IV (Fig. 2A). To further elucidate the effect of fungal community on non-volatile compounds in pile-fermentation, PCA analysis was then performed to analyze the
classification and grouping the non-volatile compounds. Results showed that non-volatile compounds changed periodically, as shown in Fig. 4A, from the distribution of principal component 1 (x-axis), S0 (group I) is quite different from other samples, while from the distribution of prin- cipal component 2 (y-axis), S1–S3 (group II) also show differences with other groups, and S4–S5 (groupIII) and S6 (group Ⅳ) are significantly different in spatial distribution. Thus, the composition of non-volatile compounds could also be divided into 4 stages (group I–IV). The grouping of non-volatile components was consistent with the grouping of the number of fungal species, indicating that the possibility of
Screening and analysis of representative non- volatile components during pile-fermentation. (A) Comparison of the dynamic non-volatile compounds using PCA analysis in different pile-fermentation periods of dark green tea. (B) The PLS-DA score plot of the dynamic non-volatile compounds in different pile-fermentation periods of dark green tea.
(C) Validation of the PLS-DA model of the dynamic non-volatile compounds in different pile- fermentation periods of dark green tea. (D) Correla- tion analysis using Pearson’s correlation analysis between the fungal genera and flavor compounds (|r|
> 0.5). (For interpretation of the references to colourin this figure legend, the reader is referred to the web version of this article.)correlation between changes in non-volatile components and changes in fungal communities.
To verify the relevance of the two factors, PLS-DA analysis was performed. As shown in Fig. 4B, there is clear difference between groups in the PLS-DA score map, the model interpretation variance and pre- dictive ability are well (R2Y 0.95, Q2 0.90). In order to avoid false positive results (Over-fitting), we repeated the calculation for the veri- fied model for 200 times to obtain the prediction figure (Fig. 4C), R2 (0.00, 0.25), Q2 (0.00, 0.49). Base on the above re-analysis, we can confirm that this model is suitable to describe the differences between groups. Further, twenty non-volatile compounds were selected using the
Kruskal-Wallis test (VIP >1, p < 0.05). VIP score represents the vari-
able’s importance in PLS-DA analysis, which summarize the contribu- tion of a variable to the PLS-DA model. According to VIP scores, these were, from high to low, caffeine, EGC, EGCG, L-isoleucine, L-theanine, ECG, L-phenylalanine, gallic acid, L-valine, GC, GCG, myricetin-3-O- galactoside, theophylline, L-tyrosine, L-glutamic acid, theobromine, L- aspartic acid, EC, kaempferol, and C (Table 1). These ingredients changed dramatically during pile-fermentation, and may be the main ingredients that contribute most to the characteristic flavor of dark green tea.
3.5. Correlation analysis of major fungal succession and flavor compound during pile-fermentation
The relationship between the important fungal community and major non-volatile compounds during different pile-fermentation pre- cesses was evaluated by Pearson’s correlation analysis. The genus of
Aspergillus was significantly (p < 0.05) positively correlated with ECG,
Byssochlamys was significantly (p < 0.05 or p < 0.01) negtively corre- lated with kaempferitrin, myricetin-3-O-galactoside, caffeine, isoleu-
cine, aspartic acid, glutamine acid, valine, tyrosine, phenylalanine, gallic acid, GC, EGC, C, and GCG. Candida Cyberlindnera, Debaryomyces,
and Penicillium were all significantly (p < 0.01) correlated with
kaempferitrin, myricetin-3-O-galactosidase, theobromine, caffeine, theanine, aspartic acid, glutamic acid, GC, EGC, EC, C, EGCG, ECG.
Rasamsonia was significantly (p < 0.01) correlated with isoleucine,
tyrosine, and phenylalanine. Both Thermoascus and Thermomyces were significantly (p < 0.01) correlated with tyrosine, phenylalanine, and gallic acid. Sordariomycetes-unclassified was significantly (p < 0.01) correlated with theophylline. Eurotiales-unclassified was significantly
Summary of non-volatile compounds biomarkers derived from PLS-DA model.
Compound ID VIP value P value
Caffeine 14.40 0.000**
EGC 10.93 0.000**
EGCG 8.75 0.000**
L-Isoleucine 6.40 0.000**
L-Theamine 5.89 0.000**
ECG 5.37 0.000**
L-Phenylalanine 3.80 0.000**
Gallic acid 3.60 0.001**
L-Valine 3.34 0.000**
GC 2.50 0.000**
GCG 2.47 0.000**
Myricetin 3-O-galactoside 2.42 0.000**
Theophylline 2.38 0.000**
L-Tyrosine 2.02 0.000**
L-Glutamic acid 1.90 0.000**
Theobromine 1.73 0.000**
L-Aspartic acid 1.56 0.000**
EC 1.38 0.000**
Kaempferitrin 1.25 0.000**
C 1.16 0.000**
Note: **indicated a significant difference (P < 0.01). EGCG, epigallocatechin gallate; EGC, epigallocatechin; ECG, epicatechin gallate; GC, gallocatechin; GCG, gallocatechin gallate; EC, epicatechin; C, catechins.
(p < 0.05 or p < 0.01) negatively correlated with myricetin-3-O- galactoside, theophylline, aspartic acid, glutamic acid, gallic acid, GC, and C. Fungi that were highly positively related to flavonoids and
flavonoid glycosides, alkaloids, amino acids, and catechins were Candida, Cyberlindnera, Debaryomyces, and Penicillium. Chlamydomonas was negatively correlated with flavones, flavonoid glycosides, organic acids, amino acids, and catechins. Cyanobacteria, Thermoascus, and Thermomyces were positively correlated with amino acids (Fig. 4C and Table 2).
Dark green tea is generally processed by solid-state pile-fermentation involving the fungal community which modifies the non-volatile flavor compounds. By using Illumina Miseq sequencing, UPLC-QTOF-MS/MS, PLS-DA and Pearson analysis, we clarified, for the first time, the varia- tions of the fungal community and the major non-volatile flavor com- pounds during the pile-fermentation process. Further analysis demonstrates that there is a strong correlation between the fungal community and the main non-volatile compounds.
The fungal community variations are related to tea raw materials, processing technology and environmental conditions. The dominant fungal strain in pile-fermentation is Asperigllus, Penicillium, Candida, Debaryomyces, and Cyberlindnera also play important roles. Thermo- myces, Rasamsonia, and Thermoascus occur during the mid-stage pile- fermentation. Byssochlamys and Aspergillus became the predominant fungus at the end of the fermentation. Aspergillus could produce abun- dant hydrolytic enzymes that can hydrolyze the cellulose, pectin and protein substances in the cell wall of the tea leaves to form soluble sugar, amino acids, soluble pectin and other compounds (Wang, Peng, & Gong, 2011), which are conducive to the formation of the mellow and taste quality of dark green tea. Raw material, different processing techniques and changes in environmental conditions will cause microorganism changes. Our study demonstrated that the fungal community changed with the pile-fermentation progresses. Aspergillus was the dominant genus. In addition, Penicillium, Candida, Debaryomyces and Pseudowell Saccharomyces occurred during the pile-fermentation process. We also observed that Thermomyces, Rasamsonia, Thermoascus occurring when the humidity and temperature sharply increased during the early pe- riods of pile-fermentation. Byssochlamys and Aspergillus were the domi- nant fungi at the end of pile-fermentation. These variations of fungal
community during pile-fermentation may be associated with the raw
materials, processing technology and environmental conditions of dark green tea (Li, Chai, Li, Huang, Luo, Xiao, & Liu, 2018). Raw materials of dark green tea are mainly made by the mature leaves of small leaf species (Camellia sinensis.var.sinensis) and are quite different from the Yunnan large leaf species (Camellia sinensis.var. Assamica) usually used to make Pu-erh tea. In addition, the water content of the piled tea could be as high as 39.54% at the beginning of the process, the highest tem- perature of the piled tea could be as high as 63.20 ◦C during the pile- fermentation, both of which are also significantly different from other dark teas. Meanwhile, the tea plants for making dark green tea is grown in subtropical climates, which is different from other types of dark teas’ production conditions. All of these factors contribute to the variation of fungal community during pile-fermentation of dark green tea.
The fungal community mainly affected non-volatile compounds with secreted enzymes. In our research, eleven important microorganisms occurring during the pile-fermentation. Aspergillus produces abundant hydrolases, such as cellulase, hemicellulase, xylanase, pectinase, pro- tease (Bourdichon et al., 2012). These enzymes hydrolyze cellulose, pectin and protein substances to form soluble sugars, amino acids, sol- uble pectin and other components (Wang et al., 2011). Penicillium se- cretes a variety of hydrolases, such as amylase, cellulase, chitinase, β-glucosidase, xylanase, ligninase, protease, and lipase. These enzymes participate in the cleavage of phenol and the derivatives and catalyze the conversion of terpenes to ketones and alcohols (Tai et al., 2016).
Thermomyces and Thermoascus produce large amounts of heat-resistant
enzymes, such as xylanase, lipase, protease and chitin enzyme.
Note: GC, gallocatechin; EGC, epigallocatechin; EC, epicatechin; C, catechins, EGCG, epigallocatechin gallate; GCG, gallocatechin gallate; ECG, epicatechin gallate. *P ＜ 0.05, **P ＜ 0.01.
Byssochlamys is a low-pressure and high-temperature resistant genus and secretes a large amount of pectinase a, which may participate in the glycoside hydrolysis reaction during pile-fermentation (Morais et al., 2018). Importantly, in our study, Rasamsonia was detected for the first time in dark green tea. Up to now, far too little attention has been paid to the role of Rasamsonia in pile-fermentation of dark teas. Thus, the mi- croorganisms mainly secrete abundant extracellular enzymes and participate in the formation of the characteristic qualities of dark green tea through various biochemical reactions.
Changes in catechins during pile-fermentation are related to the occurrence of hydrolysis, oxidation, and polymerization. Seven cate- chins significantly decreased during pile-fermentation of dark green tea, our finding is consistent with other related studies of Qingzhuan tea (Cheng et al., 2020; Feng et al., 2020). The reasons for the decline of catechins may be as follows: First, the hydrolysis reaction of ester cat- echins under the action of humid conditions. Second, catechins are enzymatically oxidized, with the polyphenol oxidase and catalase in the enzymatic reaction may be coming from Aspergillus, Candida and Ther- mophilus. Moreover, three catalase enzymes secreted by Aspergillus have
the ability to catalyze the oxidation of catechins during pile-
Pearson’s correlation between the relative abundances of important genera and non-volatile compounds in pile-fermentation of dark green tea.
fermentation (Zhao et al., 2015).
Changes of amino acids during pile-fermentation are related to the occurrence of decarboxylation and deamination, coupling oxidation, and Maillard reaction. Our study demonstrated that the content of amino acids was significantly decreased during pile-fermentation which is consistent with other related studies of dark tea (Feng et al., 2020). Reasons for the decrease of amino acids were: First, amino acids undergo decarboxylation and deamination (Qu et al., 2019). Second, amino acids undergo a coupling oxidation reaction with oxidation products such as catechin and orthoquinone (Tan et al., 2016). Third, amino acids un- dergo Maillard reaction (Hofmann & Schieberle, 2000). Furthermore, amino acids are considered as stable carbon and nitrogen sources, providing nutrients for the large-scale reproduction of microorganisms such as Aspergillus (Moe, 2013).
Caffeine significantly decreased during the pile-fermentation process
of dark green tea. On the one hand, caffeine may react with catechin and its oxidation products, proteins and other substances to form complexes by hydrogen bonding (Fu et al., 2008). On the other hand, caffeine undergoes a demethylation reaction under the action of extracellular enzymes secreted by microorganisms such as Aspergillus, Penicillium and Yeast. Seven strains of Aspergillus and Penicillium can degrade caffeine by removing the methyl group at position 7 to form theophylline and then removing the methyl group at position 1 to obtain 3-methylxanthine, finally demethylating to xanthine (Hakil, Denis, Viniegra-Gonza´lez, & Augur, 1998). Moreover, Saccharomyces could reduce the content of caffeine during fermentation.
Myricetin 3-O-galactoside Theobromine Theophylline
L-Aspartic acid L-Glutamic acid L-Valine
L-Phenylalanine Gallic acid
GC EGC EC C
EGCG GCG ECG
Changes of flavonoids and flavonoid glycosides in the pile- fermentation are related to the occurrence of oxidation, polymeriza- tion and hydrolysis. We indicated that the flavonoids and flavonoid glycosides were significantly decreased during pile-fermentation. Fla- vonols and flavonoid glycosides are not very stable under natural con- ditions and are prone to oxidative degradation under humid conditions (Dou, Lee, Tzen, & Lee, 2007). Flavonols can also polymerize with an- thocyanins to produce co-pigments. In addition, flavonoids are hydro- lyzed by heat and glycoside hydrolase (Bloor & Falshaw, 2000). Flavonoids may also be hydrolyzed by enzymes secreted by some mi- croorganisms. Aspergillus, Penicillium, and Chlamydomonas can produce β-glucoside hydrolase. Chlamyces can also produce a large amount of pectinase, which may be involved in the hydrolysis of flavonoid glyco- sides, producing glucose, rhamnose, galactose (Zhu et al., 2020). As a result, the release of flavonoid in water from the deglycoside base is conducive to reducing the bitterness and acerbity taste of black brick tea.
Change of gallic acid during pile-fermentation is related to theoccurrence of hydrolysis, oxidative polymerization, and secondary re- actions. Gallic acid accounts for about 80% of the total organic acids (Lu, Lin, Gu, Guo, & Tan, 2007). The increase in gallic acid in the early pe- riods of pile-fermentation may be related to the hydrolysis of the ester catechin. Ester catechins can undergo hydrolysis reactions to form simple catechins and gallic acid (Qin, Li, Tu, Ma, & Zhang, 2012). Meanwhile, hydrolysis reaction can also occur under the catalysis of tannase. Aspergillus, Penicillium, Debaryomyces, and Candida can produce tannase (Aguilar et al., 2007). Tannase hydrolyzes the ester and phenol condensation bonds in gallic tannin to produce simple catechins and gallic acid, which can also effectively hydrolyze the acetal bonds of EGCG to produce gallic acid and EGC (Bajpai & Patil, 1997). Thus, the decline of gallic acid in the middle and late periods of the pile- fermentation may be related to its secondary metabolic reaction and the formation of gallic acid derivatives. A study reported that two new gallic acid derivatives were isolated from Pu-er Tea tea (Zhou, Zhang, Xu, & Yang, 2005). Up until now, there were still relatively few studies on gallic acid derivatives in tea (Yang et al., 2014). The genus Byssochlamys occurring in the middle and late periods of pile- fermentation of dark green tea, and the role in degradation of gallic acid needs to be further studied.
In conclusion, the present findings demonstrate that the pile-
fermentation processes of dark green tea can be divided into four pe- riods based on the fungal community change by Illumina Miseq sequencing. The fungal community plays an important role in the pile- fermentation. Aspergillus is the dominant genus during the overall pro- cess, and Penicillium, Candida, Debaryomyces, Cyberlindnera, and Byssochlamys also played pivotal roles. We also uncovered the non- volatile components by UPLC-QTOF-MS/MS and PLS-DA analysis to identify the main non-volatile flavor compounds and further demon- strated that these flavor compounds are affected by the fungal com- munity during the pile-fermentation process. Importantly, the findings of our research may be considered a promising prospect to better improve the quality of Dark green tea by optimizing the key pile- fermentation process.
CRediT authorship contribution statement
Shuai Hu: Methodology, Software, Writing - original draft, Data curation. Chang He: Methodology, Software, Data curation, Validation. Yuchuan Zhi Li Yu: Software, Data curation. Yuqiong Chen: . Yaomin Wang: Funding acquisition. Dejiang Ni: Conceptualization, Visualiza- tion, Funding acquisition.
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.
This work was supported by Subsequent Work Scientific Research Project for National Three Gorges (Project No: YZJ-2019-040), and the Fundamental Research Funds for the Central Universities (Project No: 2662018QD060), and the Natural Science Foundation of Hubei Province (Project No: 2019CFB187).
Appendix A. Supplementary material
Supplementary data to this article can be found online at https://doi. org/10.1016/j.foodres.2021.110472.
Aguilar, C. N., Rodríguez, R., Guti´errez Sa´nchez, G., Augur, C., Favela Torres, E., Prado Barragan, L. A., … Contreras Esquivel, J. C. (2007). Microbial tannases: Advances and perspectives. Applied Microbiology and Biotechnology, 76(1), 47–59. https://doi. org/10.1007/s00253-007-1000-2.
Bajpai, B., & Patil, S. (1997). Induction of tannin acyl hydrolase (EC 220.127.116.11) activity in Myricetin some members of fungi imperfecti. Enzyme and Microbial Technology, 20(8), 612–614. https://doi.org/10.1016/s0141-0229(96)00206-2.
Bloor, S. J., & Falshaw, R. (2000). Covalently linked anthocyanin–flavonol pigments from blue Agapanthus flowers. Phytochemistry, 53(5), 575–579. https://doi.org/ 10.1016/s0031-9422(99)00572-5.
Bourdichon, F., Casaregola, S., Farrokh, C., Frisvad, J. C., Gerds, M. L., Hammes, W. P.,
… Hansen, E. B. (2012). Food fermentations: Microorganisms with technological beneficial use. International Journal of Food Microbiology, 154(3), 87–97. https://doi. org/10.1016/j.ijfoodmicro.2011.12.030.
Cheng, L., Yang, Q., Chen, Z., Zhang, J., Chen, Q., Wang, Y., & Wei, X. (2020). Distinct changes of metabolic profile and sensory quality during Qingzhuan tea processing revealed by LC-MS-based metabolomics. Journal of Agricultural and Food Chemistry, 68(17), 4955–4965. https://doi.org/10.1021/acs.jafc.0c00581.
Cheng, Q., Cai, S., Ni, D., Wang, R., Zhou, F., Ji, B., & Chen, Y. (2015). In vitro antioxidant and pancreatic α-amylase inhibitory activity of isolated fractions from water extract of Qingzhuan tea. Journal of Food Science and Technology, 52(2), 928–935. https://doi.org/10.1007/s13197-013-1059-y.
Dai, W., Qi, D., Yang, T., Lv, H., Guo, L., Zhang, Y., … Lin, Z. (2015). Nontargeted analysis using ultraperformance liquid chromatography-quadrupole time-of-flight mass spectrometry uncovers the effects of harvest season on the metabolites and taste quality of tea (Camellia sinensis L.). Journal of Agricultural and Food Chemistry, 63(44), 9869–9878. https://doi.org/10.1021/acs.jafc.5b03967.
Dou, J., Lee, V. S. Y., Tzen, J. T. C., & Lee, M.-R. (2007). Identification and comparison of phenolic compounds in the preparation of oolong tea manufactured by semifermentation and drying processes. Journal of Agricultural and Food Chemistry, 55(18), 7462–7468. https://doi.org/10.1021/jf0718603.
Feng, L., Liu, P., Zheng, P., Zhang, L., Zhou, J., Gong, Z., … Wan, X. (2020). Chemical profile changes during pile fermentation of Qingzhuan tea affect inhibition of
α-amylase and lipase. Scientific Reports, 10(1). https://doi.org/10.1038/s41598-020-
Fu, B., Xie, M., Nie, S., Zhou, P., & Wang, Y. (2001). Determination of polysaccharide content in tea. Food Science, 11, 69–73.
Fu, D., Hao, F., Huang, J., Lei, Y., Liu, Z., & Wang, F. (2008). Variations of components of fuzhuan tea during processing. Food Science, 29(2), 64–67. http://www.spkx.net.cn/ EN/abstract/article_13891.shtml.
Gong, Z., Ouyang, J., Wu, X., Zhou, F., Lu, D., Zhao, C., … Zhu, M. (2020). Dark tea extracts: Chemical constituents and modulatory effect on gastrointestinal function. Biomedicine & Pharmacotherapy, 130, 110514. https://doi.org/10.1016/j. biopha.2020.110514.
Hakil, M., Denis, S., Viniegra-Gonza´lez, G., & Augur, C. (1998). Degradation and product analysis of caffeine and related dimethylxanthines by filamentous fungi. Enzyme and Microbial Technology, 22(5), 355–359. https://doi.org/10.1016/s0141-0229(97)
Hofmann, T., & Schieberle, P. (2000). Formation of aroma-active strecker-aldehydes by a direct oxidative degradation of amadori compounds. Journal of Agricultural and Food Chemistry, 48(9), 4301–4305. https://doi.org/10.1021/jf000076e.
Hu, S., He, C., Li, Y., Yu, Z., Chen, Y., Wang, Y., & Ni, D. (2021). The formation of aroma quality of dark tea during pile[HYPHEN]fermentation based on multi[HYPHEN] omics. LWT – Food Science and Technology, 147, Article 111491. https://doi.org/ 10.1016/j.lwt.2021.111491.
Lai, Y., Li, S., Tang, Q., Li, H., Chen, S., Li, P., … Guo, X. (2016). The dark-purple tea cultivar ‘Ziyan’ accumulates a large amount of delphinidin-related anthocyanins. Journal of Agricultural and Food Chemistry, 64(13), 2719–2726. https://doi.org/ 10.1021/acs.jafc.5b04036.
Li, Q., Chai, S., Li, Y., Huang, J., Luo, Y., Xiao, L., & Liu, Z. (2018). Biochemical components associated with microbial community shift during the pile-fermentation of primary dark tea. Frontiers in Microbiology, 9(1509). https://www.frontiersin. org/article/10.3389/fmicb.2018.01509.
Li, Q., Huang, J., Li, Y., Zhang, Y., Luo, Y., Chen, Y., … Liu, Z. (2017). Fungal community succession and major components change during manufacturing process of Fu brick tea. Scientific Reports, 7(1). https://doi.org/10.1038/s41598-017-07098-8.
Li, Q., Li, Y., Luo, Y., Xiao, L., Wang, K., Huang, J., & Liu, Z. (2020). Characterization of the key aroma compounds and microorganisms during the manufacturing process of Fu brick tea. LWT – Food Science and Technology, 127, 109355. https://doi.org/ 10.1016/j.lwt.2020.109355.
Long, P., Wen, M., Granato, D., Zhou, J., Wu, Y., Hou, Y., & Zhang, L. (2020). Untargeted and targeted metabolomics reveal the chemical characteristic of pu-erh tea (Camellia assamica) during pile-fermentation. Food Chemistry, 311, 125895. https://doi.org/ 10.1016/j.foodchem.2019.125895.
Lu, H., Lin, Z., Gu, J., Guo, L., & Tan, J. (2007). Study on the gallic acid in Pu-erh tea.
Journal of Tea Science, 27(2), 104–110. http://www.tea-science.com/CN/abstract
Luo, Z., Du, H., Li, L., An, M., Zhang, Z., Wan, X., … Ling, T. (2013). Fuzhuanins A and B: The B-ring fission lactones of Flavan-3-ols from Fuzhuan Brick-tea. Journal of Agricultural and Food Chemistry, 61(28), 6982–6990. https://doi.org/10.1021/ jf401724w.
Lv, H., Zhang, Y., Shi, J., & Lin, Z. (2017). Phytochemical profiles and antioxidant activities of Chinese dark teas obtained by different processing technologies. Food
Research International, 100, 486–493. https://doi.org/10.1016/j. foodres.2016.10.024.
Mao, Y., Wei, B., Teng, J., Huang, L., & Xia, N. (2017). Analyses of fungal community by Illumina MiSeq platforms and characterization of Eurotium species on Liupao tea, a distinctive post-fermented tea from China. Food Research International, 99, 641–649. https://doi.org/10.1016/j.foodres.2017.06.032.
Moe, L. A. (2013). Amino acids in the rhizosphere: From plants to microbes. American Journal of Botany, 100(9), 1692–1705. https://doi.org/10.3732/ajb.1300033.
Morais, T. P. D., Barbosa, P. M. G., Garcia, N. F. L., Rosa-Garzon, N. G. D., Fonseca, G. G., Paz, M. F. D., … Leite, R. S. R. (2018). Catalytic and thermodynamic properties of β-glucosidases produced by Lichtheimia corymbifera and Byssochlamys spectabilis. Preparative Biochemistry & Biotechnology, 48(9), 777–786. https://doi.org/10.1080/ 10826068.2018.1509083.
Qin, J., Li, N., Tu, P., Ma, Z., & Zhang, L. (2012). Change in tea polyphenol and purine alkaloid composition during solid-state fungal fermentation of postfermented tea. Journal of Agricultural and Food Chemistry, 60(5), 1213–1217. https://doi.org/ 10.1021/jf204844g.
Qu, F., Zhu, X., Ai, Z., Ai, Y., Qiu, F., & Ni, D. (2019). Effect of different drying methods on the sensory quality and chemical components of black tea. LWT – Food Science and Technology, 99, 112–118. https://doi.org/10.1016/j.lwt.2018.09.036.
Rui, Y., Wan, P., Chen, G., Xie, M., Sun, Y., Zeng, X., & Liu, Z. (2019). Analysis of bacterial and fungal communities by Illumina MiSeq platforms and characterization of Aspergillus cristatus in Fuzhuan brick tea. LWT – Food Science and Technology, 110, 168–174. https://doi.org/10.1016/j.lwt.2019.04.092.
Shi, J., Ma, W., Wang, C., Wu, W., Tian, J., Zhang, Y., … Lv, H. (2021). Impact of various microbial-fermented methods on the chemical profile of dark tea using a single raw tea material. Journal of Agricultural and Food Chemistry, 69(14), 4210–4222. https:// doi.org/10.1021/acs.jafc.1c00598.
Song, C., Fan, F., Gong, S., Guo, H., Li, C., & Zong, B. (2020). Taste characteristic and main contributing compounds of different origin black tea. Scientia Agricultura Sinica, 53(2), 383–394. https://www.chinaagrisci.com/CN/abstract/article_20862.shtml.
Tai, Y., Xu, M., Ren, J., Dong, M., Yang, Z., Pan, S., & Fan, G. (2016). Optimisation of α
-terpineol production by limonene biotransformation using Penicillium digitatum DSM 62840. Journal of the Science of Food and Agriculture, 96(3), 954–961. https:// doi.org/10.1002/jsfa.7171.
Tan, J., Dai, W., Lu, M., Lv, H., Guo, L., Zhang, Y., … Lin, Z. (2016). Study of the dynamic changes in the non-volatile chemical constituents of black tea during fermentation processing by a non-targeted metabolomics approach. Food Research International, 79, 106–113. https://doi.org/10.1016/j.foodres.2015.11.018.
Tian, Y., Liu, X., Liu, W., Wang, W., Long, Y., Zhang, L., … Ling, T. (2016). A new anti- proliferative acylated flavonol glycoside from Fuzhuan brick-tea. Natural Product Research, 30(23), 2637–2641. https://doi.org/10.1080/14786419.2015.1136911.
Wang, Q., Peng, C., & Gong, J. (2011). Effects of enzymatic action on the formation of theabrownin during solid state fermentation of Pu-erh tea. Journal of the Science of Food and Agriculture, 91(13), 2412–2418. https://doi.org/10.1002/jsfa.4480.
Yang, H., Gu, Q., Gao, T., Wang, X., Chue, P., Wu, Q., & Jia, X. (2014). Flavonols and derivatives of gallic acid from young leaves of Toona sinensis (A. Juss.) Roemer and evaluation of their anti-oxidant capacity by chemical methods. Pharmacognosy Magazine, 10(38), 185–190. https://pubmed.ncbi.nlm.nih.gov/24914286.
Yin, Y., Chen, Y., Jiao, Y., Hao, J., Yu, Z., & Ni, D. (2021). Effects of raw materials from different tea cultivars on green brick tea quality. Journal of Tea Science, 41(1), 48–57. http://www.tea-science.com/EN/10.13305/j.cnki.jts.2021.01.006.
Yu, X., Li, Y., He, C., Zhou, J., Chen, Y., Yu, Z., … Ni, D. (2020). Nonvolatile metabolism in postharvest tea (Camellia sinensis L.) leaves: Effects of different withering treatments on nonvolatile metabolites, gene expression levels, and enzyme activity. Food Chemistry, 327, 126992. https://doi.org/10.1016/j.foodchem.2020.126992.
Zhang, C., Guo, J., Zhang, Z., Tian, S., Liu, Z., & Shen, C. (2021). Biochemical components and fungal community dynamics during the flowering process of Moringa-Fu brick tea, a novel microbially fermented blended tea. LWT – Food Science and Technology, 140, 110822. https://doi.org/10.1016/j.lwt.2020.110822.
Zhang, L., Zhang, Z., Zhou, Y., Ling, T., & Wan, X. (2013). Chinese dark teas: Postfermentation, chemistry and biological activities. Food Research International, 53 (2), 600–607. https://doi.org/10.1016/j.foodres.2013.01.016.
Zhang, Y., Lu, J., Guo, G., Chen, Y., Yin, P., & Liu, W. (2020). Technology research and quality analysis on substitute tea of Robinia Pseudoacacia L. flowers. Food Science and Technology, 45(3), 121–126. https://d.wanfangdata.com.cn/periodical/Ch lQZXJpb2RpY2FsQ0hJTmV3UzIwMjEwNDI4Eg1zcGtqMjAyMDAzMDI0Gghldnh pMXZlYw%3D%3D.
Zhao, M., Su, X. Q., Nian, B., Chen, L. J., Zhang, D. L., Duan, S. M., Wang, L. Y., Shi, X. Y.,
Jiang, B., Jiang, W. W., Lv, C. Y., Wang, D. P., Shi, Y., Xiao, Y., Wu, J., Pan, Y. H., & Ma, Y. (2019). Integrated meta-omics approaches to understand the microbiome of spontaneous fermentation of traditional Chinese Pu-erh tea. mSystems, 4(6). https:// doi.org/10.1128/msystems.00680-19.
Zhao, M., Zhang, D., Su, X., Duan, S., Wan, J., Yuan, W., … Pan, Y. (2015). An integrated metagenomics/metaproteomics investigation of the microbial communities and enzymes in solid-state fermentation of Pu-erh tea. Scientific Reports, 5(1), 10117. https://doi.org/10.1038/srep10117.
Zhou, Z., Zhang, Y., Xu, M., & Yang, C. (2005). Puerins A and B, two new 8-C substituted Flavan-3-ols from Pu-er tea. Journal of Agricultural and Food Chemistry, 53(22), 8614–8617. https://doi.org/10.1021/jf051390h.
Zhu, M., Li, N., Zhou, F., Ouyang, J., Lu, D., Xu, W., … Wu, J. (2020). Microbial bioconversion of the chemical components in dark tea. Food Chemistry, 312, 126043. https://doi.org/10.1016/j.foodchem.2019.126043.
Zhu, Y., Chen, J., Ji, X., Hu, X., Ling, T., Zhang, Z., … Wan, X. (2015). Changes of major tea polyphenols and production of four new B-ring fission metabolites of catechins from post-fermented Jing-Wei Fu brick tea. Food Chemistry, 170, 110–117. https:// doi.org/10.1016/j.foodchem.2014.08.075.