Changes in sensory characteristics, chemical composition and microbial succession during fermentation of ancient plants Pu-erh tea

Highlights • The soluble sugar content of the fermented ancient plants Pu-erh tea was 7.55%.• Microbial succession during the fermentation was analyzed.• The relative levels of 113 metabolites changed significantly during fermentation.• The microbial genera responsible for compositional changes were identified.


Introduction
The tender leaves of the tea plant [Camellia sinensis var.sinensis, or var.assamica (L.) O. Kuntze] are used to produce tea, the world's second most popular beverage after water (Wang et al., 2020).In 2021, the area of tea cultivation was 8,636,270 ha, spread over 48 countries (http:// www.fao.org/faostat).In modern tea plantations, tea plants are generally pruned to a height of 70 to 80 cm to improve yields and harvesting efficiency, which dominates worldwide.
Yunnan Province in southwest China is a major area of tea plant cultivation; there are more than 460,000 ha of tea plantations, including for the famous pu-erh tea and black tea.There are many ancient tea plants in the cities of Yunnan Province, Xishuangbanna, Pu'er, and Lincang (Zheng et al., 2019;Zi, Jiang, Cheng, Li, & Huang, 2020).According to the Forestry Industry Standards of the People's Republic of China (LY/T 3311-2022), "ancienttea plants" are defined as tea trees > 100 years old, or with a trunk diameter > 25 cm (Lan et al., 2022).In ancient tea plantations, there is no modern cultivation management, for example, tea plants are not pruned, so they can grow to a height of several meters and receive little or no fertilizer; there are at least 220,000 ha ancient tea plantation in Yunnan province.The fresh leaves of ancient plants are manufactured to sun-dried green tea, and then fermented into high-quality, valuable pu-erh tea, named ancient plants pu-erh tea (APPT) (Ge et al., 2021).APPT possess thicker leaves with Abbreviations: APPT, Ancient plants Pu-erh tea; HPLC, High-performance liquid chromatography; LC-MS, Liquid chromatography mass spectrometry; RM, Raw material; F1, First turning in fermentation; F2, Second turning in fermentation; F3, Third turning in fermentation; F4, Fourth turning in fermentation; WE, Water extracts; TPs, Tea polyphenols; SS, Soluble sugars; TF, Theaflavins; TR, Thearubigins; TB, Theabrownins; PCA, Principal Components Analysis; RLs, Relative levels.
* Corresponding authors at: College of Tea Science & College of Food Science and Technology, Yunnan Agricultural University, Kunming 650201, Yunnan, China (Y.Ma; M. Zhao).
E-mail addresses: 565652744@qq.com(Y.Ma), zhaoming02292002@aliyun.com (M.Zhao). 1 Contributed equally to this work.more leather substances, clearer veins; the flavor is deeper, thicker and could retain for a longer time; the taste is complex mixing rough, fine, rich and smoot; the aftertaste is sweetness (Mei, 2016).Due to the unique and desirable flavor, APPT commands a relatively high price compared with normal tea from young plants (Liang et al., 2006).
Pu-erh tea is a famous traditional Chinese tea, which classified as raw or ripen pu-erh tea manufactured by compression or microbial fermentation using sun-dried green tea leaves as raw material (Wang et al., 2018).Ripen pu-erh tea has unique sensory characteristics, including a brownish-red tea infusion, a mellow taste, and a stale flavor, which are distinct from other types of tea.In addition, it also has various health beneficial effects, such as anti-hyperlipidemia, anti-obesity, antioxidant, anti-tumor (Lee & Foo, 2013).Ripen pu-erh tea is popular in Southeast Asia and is becoming increasingly popular in Western countries.The natural microbial fermentation is an essential part of ripen puerh tea manufacture; sun-dried tea leaves are pile-fermented under conditions of high humidity and a temperature of 40-60 • C for a few weeks, which results in oxidation, degradation, and condensation of the chemical components of the leaves (Zhu et al., 2020).Both culturedependent and independent methods showed the importance of microbial diversity and succession during fermentation, which strongly influence the sensory quality and chemical composition of tea leaves; the major microbial genera include Aspergillus, Penicillium, Rhizopus, and Saccharomyces (Ma et al., 2017).In a previous report, the microbial succession, association, activity and metabolite changes were determined during the fermentation of pu-erh tea using sun-dried green tea leaves from modern tea plants as raw material (Zhao et al., 2019).
Differences in raw material source and quality, and pile-fermentation processes also affect the composition and succession of fermenting microorganisms, which in turn, affects the quality of the resulting tea leaves (Li, Feng, Luo, Yao, Zhang, & Zhang, 2018).Therefore, fermentation of pu-erh tea using ancient plants sun-dried green tea as raw material may involve a microbiome distinct from other tea fermentations and result in distinct chemical changes.However, little is known about the metabolism of tea chemical compounds and microbial succession during APPT fermentation.
In this study, changes in sensory characteristics, metabolism of chemical compounds and microbial composition were investigated during APPT fermentation, by sensory evaluation, metabolomic analysis, and high-throughput sequencing.In addition, the correlation between the dominant microorganisms and changes in chemical composition was analyzed, to improve understanding of the relationships between them.

Fermentation of pu-erh tea and sample collection
Fresh leaves with one bud and three leaves were plucked from "Ancient tea plants" at Laobanzhang Village (100 • 29′52.638''N,21 • 43′43.6584''E) in Xishuangbanna Autonomous Region, Yunnan Province, China.Fresh leaves were spreading for 4-5 h, and fixed in a hotpan with about 280 • C for 30 min.Which were rolled, then sun-dried to moisture less than 12%, and this manufactured tea leaves were sundried green tea.
The fermentation of APPT was developed using the traditional method at Menghai Dazhuo Tea Co., Ltd., Yunnan, China, from June to July 2022.The raw materials, water, containers and the environment were not sterilized and no fermentation starter culture was used, so microorganisms from the tea leaves performed the fermentation.The tea leaves were sprayed with water to achieve a moisture content of ~40%, then formed into piles ~1 M high.The tea piles were turned over and reformed into piles four times (Table S1), depending on the internal temperature, and the fermentations were terminated after 36 days.During each turning, 2 kg samples of fermenting tea leaves were collected and divided into two parts, one of which was freeze dried and subjected to sensory evaluation and compositional analysis.The other part was stored at − 80 • C for high throughput sequencing analysis.All samples were analyzed by sensory evaluation, HPLC, spectrophotometry and metabarcoding analysis of microbial DNA.In addition, the raw material and samples from on tea pile were subjected to metabolomics analysis.Detailed sample information is provided in Table S1.

Sensory evaluation of tea leaves
The sensory evaluation of tea leaves was performed according the Chinese standard (GB/T 23776-2018; Gong et al., 2018).The sensory evaluation was carried out in a purpose-built sensory review room at 25 • C and relative humidity < 70%.Briefly, each sample was evenly spread out on bamboo dividing trays and the appearance of these teas, including streaking, clarity, integrity and color were assessed visually and graded.Secondly, dry tea leaves (3 g) were added to boiling water (150 mL) for 5 min, then nine trained sensory testers (five males and four females, aged 24 to 48 years) assessed the infused tea for color, aroma and flavor.In addition, taste profiles (bitterness, astringency, thickness, sweetness, sourness and umami) were assessed using quantitative descriptive analysis (Fan et al., 2021).The intensity level of the flavor attributes was assessed using a scale ranging from 0 (undetectable) to 10 (very strong).The CIELAB color parameters were determined, i.e., L* -lightness, a* -redness (+) and greenness (− ) and b*yellowness (+) and blueness (− ).
The mass spectrometric data were acquired using a Thermo UHPLC-Q-Exactive HF-X Benchtop Orbitrap mass spectrometer, equipped with a heated-electrospray ionization (HESI) source, operating in both positive and negative ion mode.The analysis conditions were set as follows: Sheath gas flow rate 60 psi; aux gas flow rate 20 psi; aux gas heater temperature 370 • C; ion-spray voltage floating (ISVF) at − 3000 V in negative mode and at +3000 V in positive mode; normalized collision energy, 20-40-60 V rolling for MS/MS.Data acquisition was performed in data dependent acquisition (DDA) mode, over a mass range of 200-2000 m/z.

Data analysis
The pretreatment of LC/MS raw data was performed by Progenesis QI (Waters Corporation, Milford, USA) software, and a threedimensional data matrix in CSV format was exported.The information in this three-dimensional matrix included: sample information, metabolite name and mass spectral response intensity.Internal standard peaks, as well as any known false positive peaks (including noise, column bleed, and derivatized reagent peaks), were removed from the data matrix, deredundant and peak pooled.At the same time, the metabolites were identified by searching database, and the main databases were the HMDB (https://www.hmdb.ca/),Metlin (https://metlin.scripps.edu/).

Statistical analysis
The experimental data are presented as the mean ± standard deviation (SD) from at least triplicate experiments for each condition.Oneway analysis of variance (ANOVA) with Dunnett's multiple comparisons test was used to distinguish the significance level of differences between samples; p < 0.05 was considered to be statistically significant.Heatmaps were plotted using the TBtools software (Toolbox for Biologists; Version 1.082, China).Linear discriminant analysis Effect Size (LEfSe) was performed using the Galaxy tool (https://huttenhower.sph.harvard.edu/galaxy/).The correlation between dominant genera and chemical composition was calculated by Spearman's test and visualized by Python (version 3.6.6).

Changes in sensory characteristics
The infusion of the green tea raw material (RM) was yellow-green in color, bitter and thick in taste, with a fresh aroma.During the initial stage of fermentation (sample F1), the tea infusion was yellow with little sour taste and at the end of fermentation (sample F4), the tea infusion became reddish brown, mellow in taste, with a stale aroma (Fig. 1A).These changes in sensory attributes are similar to those generally observed in pu-erh tea fermentations (Zhao et al, 2019).Compared with the RM, the sweetness (6.97), thickness (7.23), sourness (0.73), a* (34.91) and b* (78.77), of sample F4 were higher (p < 0.05), whereas the bitterness (3.73), astringency (3.57), freshness (5.17) and L* (56.32), were lower (p < 0.05; Fig. 1B-C).Overall, during the fermentation, the astringency and bitterness decreased, the infusion turned from yellow to reddish-brown, a stale aroma formed and the resulting APPT had an outstanding sweetness score.
Water extracts are an important indicator of the quality of tea leaves and contain the main flavoring substances in tea infusion (Li et al, 2017).The water extracts increased significantly in F1 and then decreased (Fig. 2), in agreement with a previous report (Gong, Zhou, Zhang, Song, & An, 2005), that the water extracts decreased during fermentation.
Tea polyphenols are the main components that contribute to the flavor and health benefits of tea.Tea polyphenols in the RM were 21.59% significantly decreased to 10.58% in F4 (Fig. 2), consistent with a previous report (Zhao et al, 2019).In addition, the contents of tea polyphenols in other dark teas (e.g., Pu-erh, Qingzhuan, and Fu brick) were reduced by microbial fermentation (p < 0.05) (Gong et al., 2005;Cheng et al., 2020;Qin, Li, Tu, Ma, & Zhang, 2012).This probably results from the oxidation, condensation and polymerization of tea polyphenols, which produces large amounts of pigments or polymers, which bind to proteins (Zhu et al., 2020).
Soluble sugars reduce the bitterness of tea, from compounds such as theobromine and polyphenols, which are the main flavor compounds of tea (Hu et al., 2021).Notably, the soluble sugar content increased from 3.51% in RM to 7.55% in F4 (p < 0.05; Fig. 3), consistent with a similar trend in a previous report (Li et al., 2022).Interestingly, Xu et al (2022) suggested that the increase in soluble sugars after fermentation was due to the secreted cellulase and pectinase enzymes by the fungus that can degrade cellulose, pectin and other polysaccharides in pu-erh tea.The high soluble sugar content in the fermented tea leaves correlated with the outstanding sweetness sensory score of APPT infusion.
Theaflavins, thearubigins and theabrownins are the oxidation products of catechins, and are the main pigments responsible for the color of tea infusion.During fermentation, the content of both theaflavins and thearubigins decreased (p < 0.05), which contributed to the reduced bitterness and astringency of the tea infusion (Fig. 2).However, theabrownins, which are mainly responsible for the characteristic brown color of tea infusion, increased from 2.13% to 12.28%.Theabrownins have various health-beneficial effects, such as antioxidant, antihypercholesterolemic (Huang et al., 2019) and anti-tumor effects (Xu et al., 2020).Theabrownins are negatively associated with bitterness and astringency (Chen et al., 2022) and make a large contribution to the formation of the mellow flavor and distinctive color of tea (Cheng et al.,  ).Therefore, it appears that the continuous increase in theabrownin content in APPT during fermentation contributes to the unique quality characteristics of APPT.Consequently, theabrownins can be used as a marker of APPT maturity, enabling real-time monitoring of the fermentation process.In summary, the contents of catechins (epicatechin gallate, epigallocatechin-3-gallate, gallocatechin gallate, catechin, catechin gallate), taxifolin, luteolin, kaempferol, quercetin, thearubigins, tea polyphenols, theaflavins and theophylline decreased (p < 0.05), whereas those of soluble sugars and theabrownins increased (p < 0.05) during fermentation.Overall, these chemical compositional changes endowed APPT with a sweet, mellow taste and a reddish-brown infusion.

Changes in microbial diversity during fermentation
A total of 632,598 valid bacterial sequences was identified, then clustered into 332 bacterial OTUs, assigned to 10 phyla, 16 classes, 56 orders, 100 families, 181 genera, and 221 species of bacteria.A total of 842,926 valid fungal sequences was identified and clustered into 436 fungal OTUs, assigned to 5 phyla, 22 orders, 65 families, 131 families, 198 genera, and 249 species of fungi.

Alpha diversity
Alpha diversity is the observed richness (number of taxa), or evenness (the relative abundances of those taxa) of a microbiome.The microbial diversity and richness/abundance of the tea leaf microbiome were assessed using the Shannon and ACE indices, respectively (Fig. 4).All the sample coverage indices were greater than 0.980 (Fig. 4C, F), indicating that the sequencing depth was sufficient and accurately reflects the microbiome composition (Shi et al., 2021).The bacterial diversity (Shannon index) initially decreased markedly (RM to F1), then stayed constant apart from a small increase in F2 (Fig. 4A).The bacterial abundance (ACE index) gradually increased from RM to F3, then decreased markedly (Fig. 4E).The fungal diversity decreased markedly from RM to F1, then gradually increased, whereas the fungal abundance also decreased markedly from RM to F1, then remained relatively constant (Fig. 4D, E).The high ACE index/fungal abundance in the RM (Fig. 4E), is consistent with the industrial fermentation of pu-erh tea and indicates that the initially very abundant fungi in RM cannot adapt to the hot and humid fermentation environment (Li et al., 2022).

Beta diversity
Beta-diversity is the variability in microbial composition (the identity of taxa observed) in a microbiome.Principal components analysis (PCA) was performed on the OTU data, based on the Bray-Curtis distance matrix.PC1 and PC2 together explained 72.76% of the variation in bacterial Beta diversity and 69.62% of the variation in fungal Beta diversity (Fig. S1A, B), indicating that PCA is a good reflection of the factors influencing the differences in bacterial and fungal composition among different tea samples.It is notable that both bacterial and fungal microbiomes can be classified into three groups, i.e., group I (RM), group II (F1), and group III (F2, F3 and F4; Fig. S1A, B).In summary, the microbiome structure changed markedly in the early stage of fermentation, then stabilized in the middle and late stages.
At the genus level, four dominant genera (relative abundance > 1%) were identified as Staphylococcus, Achromobacter, unclassified_k_nor-ank_d_Bacteria, and Sphingomonas (Fig. 5B).The major genera in RM were Achromobacter, Staphylococcus, and Sphingomonas with relative abundances of 36.63,16.80 and 13.45%, respectively.Staphylococcus was the dominant genus in the middle and late stages of fermentation (F2, F3, F4) with an average relative abundance of 88.93%.Similarly, Staphylococcus multiplied markedly to 82.97% in the later stages of fermentation (Li et al., 2022).Further research is needed to fully elucidate the relative contributions of different phyla and genera to tea fermentation.

Fig. 3.
Changes in relative levels (RLs) of metabolites during pile fermentation of tea leaves.Change trend of 17 metabolites identified both by HPLC and metabolomics analysis (A); PCA analysis of peak areas (B); changes and classification of all 113 differential metabolites (C).

Fungal composition
Ascomycota was the strongly dominant phylum throughout the fermentation; its relative abundance increased from 93.32 to 99.86% and plateaued in the middle of fermentation (Fig. 5C), consistent with a previous report (Zhang et al., 2017).At the genus level (Fig. 5D), five dominant fungal genera (relative abundance > 1%) were detected, including Aspergillus, Rasamsonia, Blastobotrys, Cladosporium, and Thermomyces.Of these, Aspergillus is considered to be the most abundant and dominant genus in microbial fermentation and commercially ripen puerh tea products (Zhu et al., 2020).The relative abundance of Aspergillus increased to 92.85% in F1, then decreased to 63.47% (F2), 22.28% (F3), 29.12% (F4).Similarly, the dominant microorganism at the genus level was Aspergillus (42.10%) in the early stage of pu-erh tea fermentation; its relative abundance decreased substantially to 6.51% after fermentation (Zhang, Zhang, Zhou, Ling, & Wan, 2013).In summary, the dominant fungi in young plant pu-erh tea fermentation are similar to those found in APPT fermentation, in this study.It appears that there is a relationship between the chemical composition and nutrient resources in the fermented tea and differences in microbial composition between raw and fermented tea (Li et al., 2022).Different raw materials and fermentation environments also appear to influence differences in the microbial composition in normal pu-erh tea fermentation (Li et al., 2018).The microbial diversity of the final product is related to the raw material, fermentation environment, and storage conditions.This study found Achromobacter and Sphingomonas to be dominant microorganisms for the first time in ripen pu-er tea, providing new possibilities for process modification and improvement.

Microbial co-occurrence networks
The microbiome composition changed markedly during the fermentation and would be expected to strongly influence the chemical composition of the tea.To investigate the microbial interactions and growth-regulatory mechanisms during the fermentation, a cooccurrence network diagram of the top 20 genera in relative  S1).
The co-occurrence network diagram at the bacterial genus level (Fig. S3A) had 16 hubs and 70 linkages, of which 94.29% were positive and 5.71% were positively correlated.Staphylococcus, Achromobacter, Sphingomonas, and Acinetobacter negatively correlated (p < 0.05); it is noteworthy that the relative abundance of three of these genera decreased as the fermentation progressed, except for Staphylococcus.Staphylococcus had an average relative abundance of 88.93% in the middle and late stages of fermentation (F2, F3 and F4), and was the dominant genus during these stages.The Variovorax and Sphingobacterium genera were negatively correlated with each other (p < 0.05), and the other genera were positively correlated.The co-occurrence network diagram at the fungal genus level had 19 hubs and 238 linkages, all of which were positive, i.e., all the genera were significantly, positively correlated with each other (Fig. S3B).In summary, the growthregulatory mechanisms of most of the microbial genera in the APPT fermentation significantly positively correlated.
Fungi are mainly involved in forming the characteristic qualities of ancient plants pu-erh tea by secreting abundant extracellular enzymes, which catalyze various biochemical reactions.Thermomyces, Rasamsonia and Blastobotrys were positively correlated (p < 0.05) with theabrownins and negatively correlated (p < 0.05) with most of the characteristic components (Fig. 6B).Cladosporium was negatively correlated with theabrownins and positively correlated with caffeine, taxifolin, luteolin, kaempferol, thearubigins, quercetin, theaflavins, polyphenols, soluble sugars, theophylline, and four catechins (epigallocatechin-3-gallate, catechin, epicatechin gallate, catechin gallate) (p < 0.05).Notably, Aspergillus had the maximum relative abundance throughout the fermentation, but only showed significant positive correlations with rutin, gallic acid, epigallocatechin, gallocatechin, water extracts, epicatechin, and myricetin (p < 0.05).The metabolites produced by Aspergillus may be used by other microorganisms (Huang et al., 2022).During dark tea fermentation, Aspergillus produces CAZymes, which contribute to the mellow taste of tea infusion (Ma et al., 2021).Aspergillus can also secrete glycoside hydrolase, glycosyltransferases and vanillyl-alcohol oxidases during fermentation to catalyze the hydrolysis, oxidization, transformation, and biodegradation of phenolic compounds, thus decreasing the content of tea polyphenols and increasing the content of theabrownins and gallic acid (Liu et al., 2020).
This study clarified the influence of eight dominant genera on 24 characteristic components.Pile fermentation is responsible for the changes the sensory characteristics and composition, which contribute to the high quality of APPT.The interactions between the dominant microorganisms and the chemical composition strongly influences the sensory quality of APPT, reducing its astringency and bitterness, forming a reddish-brown color, as well as the stale aroma.However, the correlation analysis revealed only the phylum and genus level relationships between microorganisms and chemical composition.Therefore, future research should include isolation and characterization of the key microbial species and determine their individual contributions to the chemical composition of APPT.

Conclusion
This study investigated changes in sensory characteristics, metabolism of chemical compounds and microbial succession during the fermentation of ancient plants pu-erh tea (APPT).After fermentation, the astringency and bitterness of the fermented tea leaves were reduced and the soluble sugar content increased, resulting in an outstandingly sweet tea infusion.The relative levels of 113 chemical components changed significantly.Staphylococcus, Achromobacter, Sphingomonas, Thermomyces, Rasamsonia, Blastobotrys, Aspergillus and Cladosporium were the main microbial genera in the fermenting tea leaves; their relative abundances (RLs) and changes in their RLs correlated with changes in chemical composition.

Fig. 2 .
Fig. 2. Content changes of 24 characteristic compounds and other compositional quality characteristics in tea leaves during fermentation.

Fig. 6 .
Fig. 6.Heatmap of correlation analysis between characteristic tea compositional factors and dominant bacterial (A) and fungal (B) genera, using Spearman's correlation coefficients.Each column in the graph represents a characteristic tea compositional factor, each row represents a genus, and each square represents the Spearman's correlation coefficient between a genus and a metabolite.p < 0.05 is marked with *.Positive values (green) indicate a positive correlation and negative values (orange), a negative correlation.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)