Microbial communities of Schisandra sphenanthera Rehd. et Wils. and the correlations between microbial community and the active secondary metabolites

Sampling and sample processing

Portions of this text were previously published as part of a preprint (https://www.researchsquare.com/article/rs-1937757/v1). Healthy plant materials (S. sphenanthera) were collected from Zhashui County, Shaanxi province of China, south of the Qinling Mountains, in August 2019. Rhizospheric soil, root, stem, leaf, and fruit of five individual plants lying approximately 10 m apart were sampled using sterile tools. The roots of similar thickness (about two mm) were collected at a depth of 10–20 cm below the ground, without damaging the taproot. The rhizospheric soil was soil particles adhered to the root system. The residual soil remaining on the root system was collected after gently shaking off the bulk soil attached to the root system (Jiao et al., 2022; Song et al., 2023). For stem and leaf, one complete branch was randomly collected from each plant, and all leaves were collected from the sampled offshoot. Fruits were collected from each individual. The plant materials were transported to the laboratory with the sterile bag in ice boxes, and stored at −80 °C. All voucher specimens of plant materials were stored in the College of Life Sciences, Shaanxi Normal University (Voucher number: SN-ZS-YP-ZJW 001-003).

Part of the sample was air-dried in the shade at room temperature, while the other part was cleared and sterilized from epiphytic bacteria (surface sterilization) according to the following methods. 5 g of root, stem, leaf, and fruit were divided into small pieces with a sterile scalpel, and soaked in 75% ethanol for 3 min. Then, the surface-sterilized samples were washed 3–5 times with sterile water to eliminate excess ethanol. To evaluate the efficiency of the surface sterilization, 100 µl of the past washed distilled water was plated on trypsin soybean agar (TSA) and potato dextrose agar (PDA) plates, and plates were incubated at 28 °C and 35 °C, respectively. Finally, samples corresponding to plates without bacterial and fungal growth were used for sequencing.

DNA extraction, PCR amplification, high-throughput sequencing, and sequencing data processing

The DNA of the root, stem, leaf, and fruit was extracted by the cetyltrimethylammonium bromide (CTAB) method. The DNA of rhizospheric soil microorganisms was extracted using Soil Kit (Thermo Fisher Scientific, Waltham, MA, USA) following the manufacturer’s instructions. Primers designed according to conserved regions (bacterial 16S rRNA gene primer (V4) and fungal ITS1-5F rRNA gene primer) were used for the community analysis of bacteria and fungi (Table S1). PCR reactions were performed with 10 ng of template DNA, 2 µM of primers, and 15 µL of Phusion High Fidelity PCR Master Mix (New England Biolabs, Ipswich, MA, USA). The cycling parameters of PCR were initiated at 98 °C for 1 min, followed by 30 cycles (98 °C for 10 s, 50 °C for 30 s, 72 °C for 30 s), and finally extended at 72 °C for 5 min. After the detection of PCR products with 2% agarose gel, PCR products were purified, quantified, and homogenized to form the microbial sequencing library.

The purified amplicon libraries were pooled in equal concentration and sequenced with a Miseq (Illumina, Foster City, CA, USA) system following the manipulation instructions at SAGENE Guangzhou (China). Sequences with ≥97% similarity were assigned to the same OTUs and were analyzed by Uparse software (Uparse v7.0.100). The SILVA database was used based on the Mothur method to annotate OTU sequences. To investigate the phylogenetic relationships of different OTUs, multiple sequence alignment was performed using MUSCLE software (Version 3.8.31). Subsequent analyses of alpha diversity and beta diversity were performed after the normalization of OTU abundance information. Alpha diversity (Chao1 index, ACE index, Shannon, and Simpson index) in our samples was calculated with QIIME (v1.9.1), which was used to analyze the complexity of species diversity of rhizospheric soil and four parts (root, stem, leaf, and fruit). Chao1 and ACE were used to evaluate the microbial species richness. Shannon and Simpson were used to evaluate the community diversity of microflora. Beta diversity analysis was also calculated with QIIME.

Extraction of essential oils and component identification

As too many root samples would cause damage to plants, the essential oil components were not extracted and analyzed from root samples. The dried stem, leaf, and fruit were comminuted to dried powders separately in a pharmaceutical disintegrator and sieved through a 100-mesh sieve. The essential oils of dried stem, leaf, and fruit were extracted with petroleum ether at a designed time, temperature, power, and solid–liquid ratio through the ultrasonic-assisted extraction method. The supernatant was collected by centrifugal extraction solution (1200 rpm for 3 min). Essential oils were analyzed on a Thermo GC-MS Trace U3000 system using a chromatographic column, TG-5MS (30 m × 0.25 mm × 0.25 µm). The oven column heating procedure had been slightly modified to maintain the initial temperature of 50 °C for two minutes and then increased to 180 °C at a rate of 6 °C/min (Wang et al., 2018). The mass spectra were obtained by automatic scanning at m/z 35–550 amu after injection of 1.0 µL sample. Components were identified based on matching their recorded mass spectra to the main library database, and their relative concentrations were calculated by comparing their GC peak areas to the total area.

The extraction of fruit essential oil was carried out according to the optimal extraction method (the solid–liquid ratio was 1:10, 30 °C, 240 W ultrasonic for 30 min) obtained from the orthogonal optimization experiment conducted in the laboratory earlier (Wang et al., 2018). The orthogonal L₉(3)⁴ design was used to extract essential oils from the stem and leaf, and the optimum extraction process was studied. Four factors with three variation levels were listed in Table S2. The experimental conditions and extraction rates for each test and the results of the analysis of variance were presented in Tables S3 and S4. The yield of essential oil (%) was calculated by dividing the essential oil content by the weight of the dried pretreated sample weight. The contents of essential oils extracted from the stem, leaf, and fruit by the optimum extracting technology reached 4.87%, 5.42%, and 6.33%, respectively, which were the average values of the triplicate experiments.

Statistical analysis

Principal component analysis (PCoA) and hierarchical cluster analysis were performed to evaluate differences of rhizospheric soil and four parts (root, stem, leaf, and fruit) in species complexity. The obtained OTU classification information was used to plot the structure and composition histogram of each sample and the visual heatmap. Finally, the metabolic and ecologically relevant functions of endophyte and rhizospheric soil microorganisms were annotated by Tax4Fun for 16S rDNA OTUs and FunGuild for ITS OTUs. Spearman correlation analysis was mainly used to reflect the relationship between the active secondary metabolites of the sample and the changes in the relative abundance of microorganisms (top 20) in different plant organs. The p-values were corrected by the BY method to adjust the error detection rate. All experiments were independently replicated at least three times, and all data were expressed as mean ± standard error. One-way analysis of variance (ANOVA) was used to analyze the differences among secondary metabolites with a significance of p < 0.05. All statistical analyses were carried out using R 4.3.1 (R Core Team, 2023).

OTU analysis and the diversity of bacteria and fungi

To obtain the microbial taxa distribution among plant organs (root, stem, leaf, and fruit), and rhizospheric soil, we analyzed the microbiomes by 16S/ITS amplicon sequencing analysis. The total OTUs of bacteria were significantly different from that of fungi, especially in rhizospheric soil (total OTUs of bacteria were 17 times more than fungi) (Fig. S1A). OTUs cluster analysis showed that the microorganisms in the above-ground parts (stem, leaf, and fruit) of S. sphenanthera were grouped into one group, and the microorganisms in the below-ground parts (rhizospheric soil and root) into another category (Figs. S1B and S1C). The Venn diagram results showed that there were differences in composition among rhizospheric soil and different parts of S. sphenanthera (Figs. S1D and S1E). There were 10 common OTUs of bacteria in rhizospheric soil and four parts (root, stem, leaf, and fruit) of S. sphenanthera (Fig. S1D), which belonged to three phyla (Proteobacteria, Actinobacteria, and Cyanobacteria) respectively. Among the fungi of S. sphenanthera, there were 6 common OTUs in rhizospheric soil and four parts (Fig. S1E) belonging to two phyla (Ascomycota and unclassified_Fungi).

Alpha diversity indices were applied for the estimating of endophytic community complexity based on the OTUs sequence, and the diversity of the community was further reflected (Table 1). Among the endophytic and rhizosphere bacteria identified, the diversity of the below-ground parts was higher than the above-ground parts, and the soil Shannon index was nearly two times as high as that of the endophytes index. It indicated that the abundance of microbial species in the rhizospheric soil was high and distributed evenly. Among the identified fungal microbial communities, the Simpson index was above 0.9, the Shannon indexes in the stem and leaf were higher, and the microbial diversity of the fungi was lower than that of the bacteria (Table 1). Chao1 also revealed that the diversity of root samples was higher than that in the leaf samples. On the contrary, fungi (Chao 1) showed an increasing trend of species richness from root to stem and leaf samples (Table 1). The diversity of endophytic microbial in different parts of the host and the specificity of the rhizospheric soil were fully reflected.

The Beta diversity of the microbial community structure of rhizospheric soil and four parts of S. sphenanthera was evaluated according to the principal coordinate analysis diagram, which could intuitively show the microbial community difference of rhizospheric soil and different parts (Fig. S2). The distribution of the three points representing stem, leaf, and fruit in the figure was not close to each other, and the separation distance was not very large, which illustrated that the endophyte community structures in the above-ground parts were different, but the differences were small. The part of above-ground parts (stem, leaf, and fruit) were far apart from the below-ground parts (root and rhizospheric soil), indicating that there were great differences in the composition of microbial between the above-ground parts and the below-ground parts (Fig. S2).

The microbial composition and community structure of bacteria and fungi

The histogram abundance map could analyze the composition and proportion of species between rhizospheric soil and different parts more intuitively, taking the level of the phylum as an example (Fig. 1). The endophytic bacteria were mainly composed of Cyanobacteria and Proteobacteria, and the bacteria of rhizospheric soil were mainly composed of Acidobacteria and Proteobacteria (Fig. 1A). The fungi in rhizospheric soil and different parts of S. sphenanthera were mainly composed of Ascomycota and Basidiomycota (Fig. 1B). Ascomycota had higher abundance in fruit, stem, leaf, and root, while Basidiomycota had higher abundance in rhizospheric soil (Fig. 1B). However, there were differences in the dominant flora of endophytes at lower levels in the root, stem, leaf, and fruit (Fig. 2, Figs. S3 and S4, Tables S5 and S6).

There were 12 common bacterial genera in the rhizospheric soil and four parts of S. sphenanthera (Fig. 2A). Achromobacter was the dominant genus of leaf and stem, and Methylobacterium was the dominant genus of leaf. In addition, Methylobacterium was also found in fruit and stem (Fig. 2A). Rhodoplanes was the dominant genus in root and rhizospheric soil, and only existed in below-ground parts (Table S5). There were 11 common fungal genera belonging to rhizospheric soil and four parts of S. sphenanthera (Fig. 2B). Among the common genera, Alternaria was found with high abundance in root, stem, leaf, and fruit, especially in the stem. Cercospora was found with high abundance in stem and leaf (Fig. 2B). In addition, Hebeloma (rhizospheric soil), unclassified_Helotiales (root), Stomiopeltis (stem), and Botrytis (fruit) ran to greater than 10% in their respective domains (Table S6).

The heat map of species level clustering showed that rhizospheric soil and root were still clustered as one group, indicating that microbes of rhizospheric soil and root were relatively close (Fig. S4). The clustering results of rhizospheric soil and four parts were consistent with OTUs clustering results (Fig. S1C). It indicated that rhizospheric soil and root microbes were relatively close, and the microbes of stem, leaf, and fruit were close (Fig. S4). According to the above results, there were significant differences in the composition and diversity of microflora in rhizospheric soil and different parts.

Functional annotations

To explore the function of microorganisms in different tissue parts of S. sphenanthera, Tax4Fun and FunGuild software were used to predict the functions of bacteria and fungi based on 16S rDNA and ITS sequences (Fig. 3). The results showed that there was no significant difference in the function of bacteria between rhizospheric soil and different parts of S. sphenanthera, and most of the bacteria were concentrated in metabolism, environmental information processing, cellular processes and so on (Fig. 3A). Among them, more than 70 percent of the bacteria were involved in metabolic function, and among metabolic functions, global and overview maps accounted for about half (Fig. 3A, and Fig. S5). Compared with other parts, the relative abundance of bacteria about two functions (xenobiotics biodegradation and metabolism (Fig. S5A), and microbial metabolism in diverse environments (Fig. S5B)) in rhizospheric soil and root was higher. In addition, the relative abundance of bacteria in root about lipid metabolism, metabolism of terpenoids and polyketides, and fatty acid metabolism was higher than that in other parts (Fig. S5).

However, there were differences in fungal functions between rhizospheric soil and different parts of S. sphenanthera (Fig. 3B). Besides the unidentified groups, the fungi in rhizospheric soil and different parts of S. sphenanthera included 11 ecological functional groups (Fig. 3B). The relative abundance of fungi about four functions (arbuscular mycorrhizal, bryophyte parasite, ectomycorrhizal, and lichenized) in below-ground parts were much higher than that in the other three parts. Among the four functions, only the relative abundance of fungi with arbuscular mycorrhizal function was greater in root than in rhizospheric soil. Endophytic fungi (root, stem, leaf, and fruit) were mainly pathogenic functional groups and saprophytes, among which the ecological functional groups with higher relative abundance were mainly plant pathogen, animal pathogen, wood saprotroph, and undefined saprotroph. Only endophytic fungi in stem had algal parasite function (0.25%) (Fig. 3B). The results also showed that the fungal groups in the samples were mainly predicted to have seven nutrient types, such as pathotroph, symbiotroph, saprotroph and pathotroph-saprotroph (Fig. 3C). In the rhizospheric soil, saprotroph-symbiotroph and symbiotroph were the main nutrient forms, accounting for more than 39%. Pathotroph was the main nutrient form in the leaf (6.70%) and fruit (29.08%), while the relative abundance of the fungal pathotroph-saprotroph trophic form in the leaf was also high (8.52%). Pathotroph-saprotroph-symbiotroph and saprotroph were the main trophic forms of stem endophytic fungi, accounting for more than 38%. And the highest trophic pattern of endophytic fungi in the root was pathotroph-saprotroph-symbiotroph (5.96%) (Fig. 3C).

Content of active secondary metabolites in S. sphenanthera

The optimum extraction method for the highest extraction rate of the stem (1:20 ratio, 30 min, 30 °C and 300 W) and leaf (1:15 ratio, 20 min, 35 °C, 240 W) were distinct (Tables S2 and S3). Except that ultrasonic temperature had a small impact on the leaf, other factors had a significant influence on the yield of essential oils from the stem and leaf (Table S4). The factors affecting the yield of essential oils from stem and leaf were listed in order as follows: ultrasonic time >ratio of material to solvent >ultrasonic temperature >ultrasonic power, and the ratio of material to solvent >ultrasonic time >ultrasonic power >ultrasonic temperature (Table S3).

It could be found that the components and content of essential oil varied greatly from stem, leaf, and fruit (Fig. S6, Table S7). The components of essential oils from all samples with low content (<0.1%) were not listed in Table S7. A total of 40 components were identified from all the samples, and 23 essential oil components were identified from the three parts (stem, leaf, and fruit) of S. sphenanthera, but the contents were different (Fig. 4 and Table S7). Each part had its high content of dominant components, but there were obvious differences in the dominant components of the three parts. For example, γ-muurolene, δ-cadinol, and trans-farnesol were characteristic components of the stem. α-Cadinol and neoisolongifolene-8-ol were characteristic components of leaf (Fig. 4). Isospathulenol, α- santalol, cedrenol, and longiverbenone were characteristic components of fruit (Table S7).

Correlation analysis

The correlation analysis between microbial communities (top 20 at genus level) and the common components (23) in different parts (Figs. S7 and S8) showed that 12 genera of bacteria and fungi had a significant correlation with the common components (p < 0.05) (Fig. 5). Of the 23 essential oil components, only 16 components were associated with bacteria and 10 components were related to fungi (Fig. 5). Staphylococcus and Hyphomicrobium had positive correlation with five common components (γ-muurolene, β-bisabolene, germacrene D-4-ol, trans-farnesol, and ledene oxide). In addition, Staphylococcus showed a positive correlation with δ-cadinene (p < 0.05), and Hyphomicrobium showed a significant positive correlation with β-bisabolene and ledene oxide (p < 0.01) (Fig. 5A). Methylobacterium and Propionibacterium were correlated with the contents of three components, respectively. Achromobacter, Burkholderia, and Planctomyces had positive effects on farnesol and negative effects on β-chamigrene, and there was a significant positive correlation among the three genera (p < 0.01). Except for the bacteria genera mentioned above, the remaining genera were only related to one component (Fig. 5A). Among the genera of fungi, only Stomiopeltis had the most influence on the essential oil components (farnesyl alcohol, δ-cadinene, and δ-cadinol). Botrytis, Cortinarius, and Zygophiala had negative effects on α-amorphene and positive effects on β-chamigrene, and there was a significant positive correlation relationship among the three genera (p < 0.01). Unclassified_Agaricales, unclassified_Thelephoraceae, and unclassified_Chaetothyriales had a negative correlation with α-bisabolene, and there was a positive correlation relationship between the three genera. In addition, except germacrene D, β-himachalene, α-cadinol, and α-bisabolol were affected by only one genus, and the remaining components were associated with two or three genera (Fig. 5B). These results suggested that bacterial communities played a greater role than fungal communities in the accumulation of active secondary metabolites.