The effect of cropping practices on the chemical properties
The chemical properties of B. chinense rhizosphere soil in different treatments were shown in Table 2. Compared with intercropping and crop rotation, soil pH and the contents of NO3--N and Ava-K decreased after continuous planting of B. chinense, but the Ava-P content increased. The chemical parameters of rhizosphere soil differed significantly among the treatments, except the NO3--N content.
Amplicon sequencing and rarefaction curves
To characterize the microbiome in the B. chinense rhizosphere soil in different cropping practices, nine samples were sequenced by Illumina MiSeq. The amplicon sequencing resulted in 450,038 effective reads of bacterial 16S rRNA genes and 437,141 effective reads of fungal ITS region. Based on 97% similarity, the OTUs of microbial community in the rhizosphere soil were obtained. The results were shown in supplementary Table 1.
To construct rarefaction curves, the dataset was flattened according to the minimum number of sample sequences. The rarefaction curves of nine rhizosphere soil samples were constructed based on the number of OTUs observed (Supplementary Fig 1). The rarefaction curves showed that the number of OTUs rose sharply and then gradually flattened out, indicating that the sequencing library reached saturation. Therefore, it could be used for analyzing the diversity of microorganism in the rhizosphere soil of B. chinense.
Alpha diversity of bacterial and fungal communities
The alpha diversity represents the measurement of within-community microbial diversity (Table 3). Theoretically, the larger the Shannon index or the smaller the Simpson index, the higher the community diversity. According to the Shannon index, the bacterial richness was highest (6.513) in the rhizosphere soil of the rotation of B. chinense and corn, followed by continuous monocropping (6.421) and intercropping of B. chinense and corn (6.328). The Simpson index analysis confirmed the above-mentioned diversity analysis. The Shannon index and Simpson index values for fungal communities in the rhizosphere of B. chinense-corn intercropping were 4.401 and 0.029, respectively, followed by those of rotation with corn (4.250 and 0.033, respectively), and the lowest diversity values were in B. chinense monocropping (4.201 and 0.049, respectively). The results showed that the rotation and intercropping of B. chinense with corn were the main factors affecting the diversity of, respectively, bacteria and fungi in the rhizosphere. In summary, the cropping practices had an important effect on the diversity of rhizosphere microorganisms.
Beta diversity of bacterial and fungal communities
In order to shed more light on the differences in microbial community structure, NMDS analysis was performed based on the Weighted UniFrac distance (Fig 1). When the sample distance was above 0.2 (e.g. 0.3 for fungi), the samples could be divided into three groups, according to the species composition in the B. chinense rhizosphere (Fig 2). There were similarities in the structure of microbial communities within the groups and significant differences in the structure among the groups, which indicated that the cropping practices in the same field was an important factor affecting the composition microbial communities in the B. chinense rhizosphere.
The composition and structure of the bacterial community
In order to clarify the microbial community structure in the B. chinense rhizosphere, two taxonomic levels (phylum and genus) were analyzed.
As shown in Fig 3A, 13 bacterial phyla were detected in the soil from different cropping practices. The dominant bacterial phyla in the B. chinense rhizosphere soil were Proteobacteria (27%), followed by Actinomycetes (22%), Acidobacteria (20%), and Chloroflexi (9%). The relative abundances of Proteobacteria and Actinobacteria increased significantly after 3 years of continuous cropping of B. chinense, but decreased after intercropping or rotation with corn. The relative abundance of Acidobacteria and Chloroflexi decreased after continuous cropping, but increased after intercropping and rotation.
At the genus level (Fig 4A), 34 bacterial genera were detected in the soil from different cropping practices. The dominant genera were Pseudarthrobacter (4%), Microvirga (2%), Gaiella (1%), Nitrospira (1%), and Pirellula (1%). Compared with the intercropping and rotation of B. chinense and corn, the relative abundance of Pseudarthrobacter and Gaiella increased after continuous cropping. However, the relative abundance of Microvirga and Nitrospira showed a downward trend after continuous cropping and intercropping.
The composition and structure of the fungal community
As shown in Fig 3B, five fungal phyla were detected in the soil from different cropping practices. The dominant fungal phyla were Ascomycota (78%), Basidiomycota (6%) and Zygomycota (5%). The relative abundance of Ascomycota decreased after continuous cropping and intercropping, but increased after rotation with corn. The relative abundance of Basidiomycetes increased after continuous cropping, but decreased after intercropping and rotation.
At the genus level (Fig 4B), 44 fungal genera were detected in the soil from different cropping practices. The dominant fungal genera in the rhizosphere soil were Gibberella (5%), Cercophora (5%), Fusarium (4%), Chaetomium (4%), Mortierella (4%), Preussia (4%), Cryptococcus (3%), Alternaria (3%), unclassified_Ascobolaceae (2%), Cladorhinum (2%) %), Paraphoma (1%), Knufia (1%), and Cladosporium (1%). After 3 years of continuous cultivation of B. chinense, the relative abundance of Cercophora, Cryptococcus, Alternaria, Paraphoma, and Cladosporium increased, but the relative abundance of Chaetomium, Mortierella, Preussia, and Cladorrhinum significantly decreased.
Correlation analysis of dominant bacteria and soil properties
Soil chemical properties were important explanatory factors that determined the clustering patterns of soil microbial communities in different cropping treatments [21]. The chemical properties of the B. chinense rhizosphere soil were significantly different under different cropping practices (Table 2). Hence, redundancy analysis (RDA) was conducted on the relative abundance of dominant bacteria and fungi at the genus level and soil chemical factors (Fig 5). The results showed that the cumulative variation explained by the soil chemical properties was 87.84% and 59.31% for bacteria and fungi, respectively, indicating that explanatory variables had a significant influence on the structure of microbial communities. The effects of soil chemical properties on bacteria and fungi were in the order of NH4+-N>SOM>Ava-K>pH>NO3--N>Ava-P and NH4+-N>SOM> Ava-K >Ava-P >pH> NO3--N, respectively (Fig 5). In conclusion, NH4+-N, SOM and Ava-K were the main chemical properties that affected the microbial abundance and composition in the Bupleurum rhizosphere soil.
Biomarker analysis
In order to identify the dominant microbial biomarkers in the B. chinese rhizosphere soil under different cropping practices, the linear discriminant analysis (LDA) effect size (LEfSe) was carried out (Fig 6). The LDA results identified 30, 33 and 55 bacterial biomarkers in continuous monocropping, intercropping and rotation with corn, respectively (Fig 6A). The most abundant bacterial family was Comamonadaceae from B. chinense continuous monocropping soil. Rhizobium giardinii, Desulfurellaceae and Burkholderiales were abundant in the rhizosphere of B. chinense intercropped with maize, whereas Methylobacteriaceae and Microvirga were significantly enriched in the rhizosphere of B. chinense in rotation with corn.
For the fungal community, we identified 92, 57 and 34 fungal biomarkers in continuous monocropping, intercropping and rotation with corn, respectively (Fig 6B). The relatively abundant biomarker fungal taxa included Dothideomycetes and Pleosporales in the B. chinense continuous monocropping, Chaetomiaceae, Mortierellaceae and Zygomycota in B. chinense intercropped with corn, and Nectriaceae, Chytridiomycetes and Rhizophlyctidales in B. chinense in rotation with corn.