Analysis of bacterial community structure of Fuzhuan tea with different processing techniques

2.1 Tea samples

Tea samples were collected from a production workshop of Yiyang Guan-Longyu Dark Tea Development Co., Ltd, Hunan Province, China, on 5 March 2021. The samples included raw dark tea (G0), loose Fuzhuan tea (G1), lightly-pressed Fuzhuan tea (G2), hand-made Fuzhuan tea (G3), and machine-pressed Fuzhuan tea (G4) and were derived from the same batch of raw dark tea material. Among these, G1 was loose tea, while G2, G3, and G4 were pressed teas, with the pressing degree gradually rising from G2 to G4. All the samples were analysed in three biological replicates.

2.2 PCR amplification

Genomic DNA was extracted from the samples, using the Soil DNA kit (OMEGA BioTek, USA), with the purity and concentration of the extracted DNA determined using agarose gel electrophoresis. Aliquots of the purified DNA were diluted to a concentration of 1 ng/μL with sterile water and stored at −80°C for later use [13]. The total DNA of the samples was quantified using a NanoDrop 2000 ultramicroprotein analyser (Thermo Fisher Scientific, USA). The DNA mass concentration and the OD 260/OD 280 ratio were read for comparative analysis. The determination of DNA purity was based on the OD 260/OD 280 ratio, which ranged from 1.6 to 1.8 for highly purified DNA. The values below and above this range indicated excessive protein content and RNA, respectively, in the sample.

Using DNA extracted from G0, G1, G2, G3, and G4 samples as templates, we used the primer pair 338F (5′-TCCGTAGGTGAACCTGCGG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) to amplify the V3–V4 variable regions of bacterial 16S rRNA gene sequences.

The PCR reaction mixtures contained 10 ng of DNA template, 2.5 μM upstream and downstream primers, and 2.5 μM DNA polymerase (Hieff® Taq DNA Polymerase, 5 U/μL), made up to a final volume of 30 μL with PCR buffer. The PCR amplification conditions were as follows: pre-denaturation at 98°C for 60 s, followed by 30 cycles of denaturation at 98°C for 10 s, annealing at 50°C for 30 s, and extension at 72°C for 30 s, with a final extension at 72°C for 5 min. PCR products were detected using 2% agarose gel electrophoresis. The samples were mixed at equal concentrations according to the PCR product concentration, and the PCR products were detected using 2% agarose gel electrophoresis after thorough mixing.

2.3 Library construction and sequencing

Having qualified the constructed library based on Qubit quantification and library detection, we obtained an original data file by using the Illumina (MiSeq pe300) platform for sequencing, with the data being converted to sequenced readings after base calling. The results were stored in FASTQ format, containing information on readings and corresponding sequencing quality.

2.4 Bioinformatics analysis

The original data were spliced and filtered to obtain clean data. Based on the valid data, we performed sequence denoising or operational taxonomic unit (OTU) clustering, using the process of QIIME2 dada2 analysis [14] or Vsearch analysis [15], followed by species classification analysis. Based on denoising or clustering results, each sequence was annotated to obtain the corresponding species information and species-based abundance distributions.

Amplicon sequence variants (ASVs) or OTUs were analysed to obtain information on within-sample species richness and evenness [16]. Multi-sequence alignment of the ASVs or OTUs was performed, based on which, we constructed phylogenetic trees. Performing dimensionality reduction analysis, such as principal coordinates analysis (PCoA) and principal component analysis, we examined differences in the community structures of the different samples or groups [17].

The number of ASVs/OTUs per sample was obtained using QIIME2, and the numbers of common and unique ASVs/OTUs among the samples were displayed using Venn diagrams [18]. The sample alpha diversity index was assessed using the diversity alpha-rarefaction command of QIIME2 software [19].

3.1 Sequencing data statistics

The number of valid tags in the five groups of the samples was more than 53,000, with the Q20 and Q30 values greater than 95.54 and 88.58, respectively. More than 29,000 high-quality sequences were obtained. For each sample, we obtained more than 270 ASV/OTU sequences, with an average data quality of 35 or higher. For all the samples, the sequencing quality met the requirements for subsequent analysis.

3.2 ASV/OTU-Venn statistics and classification analysis

The Venn map was constructed to depict the similarity in the bacterial populations of the different samples (Figure 1). Among the five assessed tea groups, the lowest and highest numbers of unique sequences were detected in G3 (513) and G4 (946), respectively. We detected only 51 consensus sequences among the five groups. This in turn indicated that the different processing techniques exert a significant influence on the microbial community structure in Fuzhuan tea. Group G4 with the highest number of unique sequences may be attributable to the destructive nature of machine pressing as this process may lead to the pronounced releases of contents from disrupted tea tissues which may be conducive to the growth of different bacteria.

Figure 1

ASV/OTU Venn map. Note: Overlap area: the number of ASV/OTU shared between different samples. Non-overlapping areas: the number of ASV/OTU specific of each sample.

3.3 Alpha diversity index analysis

Values obtained for the Alpha diversity Chao1, Ace, and Simpson indices of each sample are shown in Table 1 and Figure 2. These results showed that the Chao1 and Ace indices of the samples from the five groups varied between 274 and 444, with the highest values obtained for group G4. We detected significant differences between groups G0 and G2, G1 and G2, G1 and G4, G2 and G4, and G3 and G4. Group G4 was found to significantly differ from all the other groups, thus indicating that the machine pressing process is conducive to microbial growth and higher bacterial abundance.

Figure 2

Alpha diversity index (a) ACE index, (b) Chao1 index, and (c) Simpson index. Note: The horizontal lines in the figure represent the two groups with differences (p < 0.05) and are marked with *.

The Simpson index is mainly used as a measure of species diversity. The lower the Simpson index value is, the higher the species diversity of a sample is. As shown in Table 1, the Simpson index values of the five groups ranged from 0.8589 to 0.9873, with the lowest value for group G4, thus indicating that the bacterial community in the machine-pressed tea is characterized by the highest bacterial diversity. We detected significant differences between group G4 and the other three processed teas with respect to the Simpson index, again indicating that the machine-pressed tea led to the highest bacterial diversity, as was consistent with the results from the ASV/OTU-Venn diagram analysis.

3.4 Rarefaction curve

Figure 3 shows the sample dilution curve. The 15 samples of the five groups were all within 2,500 sequences. With the increased number of sequencing samples, the curve showed a sharp rise, indicating that a large number of species was found in the sample community. When the number of sequences exceeded 2,500, the curve flattened, indicating that the number of species in this sample did not rise with the increased number of sequences. The number of samples sequenced was more than 20,000, indicating that the sequencing volume of each sample was sufficient. From the perspective of the number of new ASVs, compared to group G0, we detected a reduction in the final number of new ASVs in G1, G2, and G3 and an increase in the final number of new ASVs in G4. Therefore, these findings indicate that the processing technique exerts a significant impact on the growth of bacteria and that the machine pressing process was more conducive to increasing bacterial populations than the other processes.

Figure 3

Curve of dilution.

3.5 PCoA analysis

The results of PCoA analyses of the five groups are shown in Figure 4. The larger distance between G4 and G0 than between the other groups showed the distinct difference in the species diversity of the G4 bacterial community. As can be seen in Figure 5, according to principal component 1, G4 was significantly different from G1, G2, and G3 (Figure 5a). However, according to principal component 2, G1 was significantly different from G2, while G2 was significantly different from G3 and G4 (Figure 5b). Therefore, these findings indicate that the different processing techniques significantly influence the compositions of the bacterial communities in the samples. The composition of the bacterial community in G4 was affected by the processing and differed from G1, G2, and G3 significantly. This was consistent with the findings of the ASV/OTU-Venn diagram, Chao1, Ace, Simpson indices, and dilution curve analyses.

Figure 4

PCoA analysis. Note: 1, 2, and 3 after the sample numbers represent three parallel assays.

Figure 5

PCoA principal component analysis: (a) first principal component and (b) second principal component. Note: The horizontal lines in the figure represent the two groups with differences (p < 0.05) and are marked with *.

3.6 Species annotation analysis

The results of the taxonomic annotations are shown in Table 2. The highest number of annotated ASV/OTU sequences in the five groups was seven phyla, 23 classes, 41 orders, 154 families, 532 genera, and 231 species, with significant differences identified among the groups. Compared to the other groups, the number of ASV/OTU sequences annotated to each level in G4 was highest, indicating that the machine pressing process most significantly affected the composition of the bacterial communities in Fuzhuan tea, as was also consistent with the previous analysis results.

At the phylum level of classification, we annotated 17 phyla with clear status among the Fuzhuan tea samples subjected to the different processing techniques; namely, Actinobacteria, Firmicutes, Proteobacteria, Cyanobacteria, Bacteroidetes, Verrucomicrobia, Fusobacteria, Synergistetes, Deinococcus-Thermus, Planctomycetes, Patescibacteria, Tenericutes, Chloroflexi, Spirochaetes, Nitrospirae, Gemmatimonadetes, and Acidobacteria, the top ten of which are shown in Table 3. Among the five groups, Actinobacteria was identified as the predominant bacterial phylum, followed by Firmicutes, Proteobacteria, and Cyanobacteria. This was consistent with the results of MiSeq sequencing analysis of dark tea products, such as Fuzhuan tea and pu-erh tea, reported by Fu et al. [20], who found Firmicutes and Actinobacteria to be the dominant phyla.

With respect to the top five predominant phyla, the numbers of Actinobacteria, Firmicutes, and Bacteroidetes in G1 rose by varying degrees when compared to those in G0. The numbers of Proteobacteria and Cyanobacteria fell in G1 compared to those in G0. Compared to those in G0, the numbers of Actinobacteria, Firmicutes, Proteobacteria, and Bacteroidetes in G2 increased by a different degree, whereas there was a significant reduction in the number of Cyanobacteria. The pattern of phyla was found to be essentially the same between G3 and G1. Only the number of Cyanobacteria significantly increased (1.48 times), while the numbers of Actinobacteria, Firmicutes, Proteobacteria, and Bacteroidetes decreased in G4 when compared to G0. These observations verified that the different processing techniques significantly influenced the structures of the bacterial communities in Fuzhuan tea.

At the genus level, we annotated 532 genera with clear status in the different samples, among which, Corynebacterium, Methylobacterium, Bifidobacterium, Faecalibacterium, Escherichia–Shigella, Actinomyces, Bacteroides, Collinsella, Sphingomonas, Prevotella, Candidatus Xiphinematobacter, Caldicoprobacter, Aminobacterium, uncultured Acidothermaceae, Leptotrichia, Coprostanoligenes group, and Veillonella were detected commonly. The number of species in each group at the genus level is shown in Table 4. Among the five groups, Corynebacterium was most predominant, followed by Methylobacterium and Bifidobacterium. Compared to G0, a varying degree of increased numbers of Corynebacterium, Faecalibacterium, Escherichia–Shigella, Actinomyces, and Collinsella was detected in G1, whereas a reduction was observed for all the other genera, among which, a significant (approximately 10%) reduction in the number of Methylobacterium was found. Increases to a different extent in the numbers of Corynebacterium, Faecalibacterium, Escherichia–Shigella, and Bacteroides were observed for G2 when compared to G0, whereas the numbers of all the other genera declined during the processing. The varying degrees of increases in the numbers of Corynebacterium, Faecalibacterium, and Sphingomonas were found in G3 when compared to G0, with the numbers of the remaining genera declining, among which the number of Methylobacterium significantly fell by approximately 10%. Only the number of Collinsella increased in G4 compared to G0, with the numbers of all the other genera decreasing, in particular, the numbers of the unidentified genera. Collectively, these findings highlight the significant effects of the Fuzhuan tea processing techniques on the bacterial community structure, with important implications for the final tea products.

To further study the phylogenetic relationship among the species at the genus level, we used MEGA 5 to obtain representative sequences of the top 100 genera based on multiple sequence alignment [21]. The results shown in Figure 6 point to the extremely rich bacterial diversity among the differently processed Fuzhuan teas. However, the relationships among these microbial populations and their interaction effects on the product quality need to be further assessed.

Figure 6

Phylogenetic tree of top 100 genera in different sample groups.