Composition of the endophytic and rhizospheric fungal community in the root
In this study, we collected rhizospheric, endophytic, and corresponding unplanted bulk soil samples from S. ningpoensis cultivars in 16 well-separated major regions of China. We sequenced the ITS1F-ITS2R region of the ITS rRNA genes in 96 samples (48 rhizospheric and 48 endophytic samples); after filtering out reads for quality, we obtained 6,055, 850 high-quality reads for subsequent analyses, i.e., an average of 63, 081 sequences per sample. After discarding operational taxonomic units (OTUs) that were non-fungal and low in abundance, we obtained 8,709 unique OTUs with 97% similarity in the entire S1-S2 region (Table 1).
The rarefaction curves observed using the Chao 1 index indicated that the sequencing depth was sufficient to cover the fungal diversity within individual samples (Fig. S1). The α-diversity decreased from the rhizospheric to endophytic fungi in different compartments (Wilcoxon signed-rank test, p < 0.001) (Fig. S2). A Venn diagram revealed that 2,574 OTUs, which represent 29.6% of the total number of OTUs, were common to all rhizospheric and endophytic fungal communities. In addition, a total of 5,780 OTUs (66.4% of the total number of OTUs) were unique to rhizosphere samples, whereas 355 OTUs (4.1% of total) were unique to endophytic samples (Fig. S3). These results suggested that the rhizosphere harbored most of the unique OTUs.
The fungal community consisted of 16 different phyla. The distribution of the rhizospheric and endophytic microbiota obtained from 16 regions at the phylum level has been shown in Fig. 1. The dominant phyla present among all the fungal communities were Ascomycota, Basidiomycota, Mortierellomycota, unclassified_k_fungi, Glomeromycota, and Rezollomycota. In the rhizospheric compartment, Ascomycota and Basidiomycota represented an average of 69.03% and 18.31% of all species, whereas they represented 71.31% and 26.24% of the species in the endophytic community. However, the relative abundance of Mortierellomycota was 9.18% higher in the rhizosphere and 9-fold higher among endophytic fungi. The genus Plectosphaerella was found to be highly abundant in areas such as Taiqiu, Zhuzhuang, Nanfeng, Luolong, Yangxi, Yuxi, Guoyang, Qiaocheng, Lixin, and Nanchuan, while the unclassified_f_Schizoporace and Exophiala genera were abundant in Zhiwuyuan. The Guehomyces, Ceratobasidium, and Cladosporium genera were found to be abundant in Panan, Linan, and Longdong, respectively. The genera found in abundance in Wulong included unclassified_p_Ascomycota and Mortierella.
The Chao richness and Shannon diversity values indicated that the fungal communities in the endophytic compartment were less rich and diverse than the rhizosphere (Table S3). Furthermore, it was apparent that the endophytic fungal community was more selective, as the community exhibited a lower level of richness and diversity than the rhizosphere. Additionally, higher levels of richness were observed in areas such as Longdong and Yongfu, while lower levels of richness were observed in Taiqiu and Zhuzhuang.
We performed a hierarchical clustering analysis involving the 50 most abundant species in the fungal community across rhizospheric and endophytic samples (Fig. S4). The analysis showed that the rhizospheric community was well separated from the endophytic compartment. Fungal communities occurring in the same province were clustered into one category; for example, the rhizospheric samples were clustered together in the RS_NF, RS_ZZ, and RS_TQ categories. Additionally, categories such as RS_WL, EP_WL, RS_PA, and EP_PA were used for the clustering of individual samples; this suggested that the rhizospheric fungal community was similar to the endophytic community. PCoA was performed based on the OTU composition; the results revealed that there were significant variations among the 96 rhizospheric and endophytic fungal samples. The first two axes (PC1 and PC2) showed that 22.63% and 8.57% of the total variance could be observed in the fungal OTUs of the rhizospheric and endophytic samples, respectively (Fig. 2). Based on the different communities, rhizospheric samples were clustered in two regions, while there were more variations in the endophytic communities observed between sampling locations compared to the rhizosphere. Overall, the results indicate a clear division between the endospheric and rhizospheric compartments.
Correlation between soil properties and fungal communities
The enrichment of microbial communities is influenced by biotic and abiotic factors. We estimated the chemical and physical properties of all soil samples (Table 2), and the results showed that the difference in TP in all bulk soil samples was the smallest; conversely, the difference in the Mn content was the largest. In addition, the samples exhibiting lower TP and Mn levels were found in Panan, while those exhibiting higher TP and Mn levels were found in Yangxi and Nanfeng. Meanwhile, we found that the correlation between the rhizospheric and soil physicochemical compartments was stronger than that observed with the endophytic compartment (Table S4, S5). In the rhizospheric community, the Plectosphaerella (PL) and Bulleromyces (BU) genera exhibited a significant negative correlation with OM (RPL2 = 0.320, RBU2 = 0.066, p ≤ 0.01), TOC (RPL2 = 0.319, RBU2 = 0.065, p ≤ 0.01), and TN (RPL2 = 0.191, RBU2 = 0.037, p ≤ 0.05), and were positively correlated with TK (RPL2 = 0.308, RBU2 = 0.007, p ≤ 0.05). Conversely, genera such as unclassified_p_Ascomycota (UPA) and Apiotrichum (AP) were observed to be positively correlated with OM (RUPA2 = 0.288, RAP2 = 0.030, p ≤ 0.05), TOC (RUPA2 = 0.286, RAP2 = 0.029, p ≤ 0.05), and TN (RUPA2 = 0.032, RAP2 = 0.027, p ≤ 0.05) (Table 4). In the endophytic fungal community, unclassified_f_Ceratobasidiaceae (UFC), Pyrenochaeta (PY), and unclassified_f_Melanommataceae (UFM) were positively correlated with OM (RUFC2 = 0.555, RPY2 = 0.41, RUFM2 = 0.000, p ≤ 0.05), TOC (RUFC2 = 0.556, RPY2 = 0.041, RUFM2 = 0.000, p ≤ 0.01), and TN (RUFC2 = 0.553, RPY2 = 0.066, p ≤ 0.05; RUFM2 = 0.000, p ≤ 0.01).
We then removed the redundant variables and performed detrended correspondence analysis (DCA) for the ten remaining environmental characteristics. The results of DCA showed that the responses of the rhizospheric fungal community composition to the soil properties fit a single peak model (Length = 2.75). Therefore, we further analyzed how fungal communities associated with soil physicochemical factors were based on RDA (Fig. 3a). The first two axes of RDA explain why 45.13% and 12.27% of the total variations were observed in the data. The results of RDA suggested that soil physicochemical factors such as TOM, TOC, TN, and Hg were positively correlated with each other, while TP, TK, pH, Mn, Cu, and Cr levels were positively correlated with each other. Moreover, TOM (p = 0.001), TOC (p = 0.001), and TK (p = 0.001) had a significant correlation with the rhizospheric fungal community structure. Meanwhile, we used the CCA method for the analysis of the association between the endophytic community and soil physicochemical factors (Length = 4.71). We found that the correlation between endophytic fungi and environmental factors was consistent with that between the rhizospheric fungi and environmental factors (Fig. 3b).
Correlation between the root fungal community and active ingredients of S. ningpoensis
There are many non-pathogenic microorganisms present inside and outside plants. These microorganisms coexist with plants for a long time. This has a great influence on the formation and medicinal ingredient content in plants (Tello, 2011). According to the 2020 edition of the Chinese pharmacopoeia, harpagide and harpagoside were used to control the quality of S. ningpoensis. We measured the levels of harpagide and harpagoside in 16 main producing regions (Table S6), and found that the levels of harpagide isolated from Longdong and Yuxi were higher than those isolated from other areas, while the content isolated from Qiaocheng was lower. The harpagoside content in Panan was higher than that in other areas, and lower levels of harpagoside were found in Yongfu. In addition, the total harpagide and harpagoside content isolated from Nanchuan was higher than that obtained from other areas, while that obtained from Zhiwuyuan was lower. However, the sum of the harpagide and harpagoside levels in Radix scrophulariae obtained from 16 areas was higher than 0.45 mg, which indicated that Radix scrophulariae obtained from all producing areas met the standards of the 2020 edition of the Chinese Pharmacopoeia.
We analyzed the relationship between the active constituents and the abundance of fungi (RF >1%), including the rhizospheric and endophytic compartments, using correlation and regression analysis (Table S7). The harpagide content (Y1), harpagoside content (Y2), the sum of the two (Y3), and the relative frequencies (X) of the fungal community of S. ningpoensis roots were analyzed in table S8. Equation 1 suggests that the highest effect on the harpagide content was observed in unclassified_o_Pezizales, followed by Cladosporium, unclassified_p_Basidiomycota, unclassified_f_Melanommataceae, Pyrenochaeta, and Aspergillus. The harpagide content was positively correlated with the abundance of Cladosporium, Pyrenochaeta, unclassified_o_Pezizales, unclassified_p_Basidiomycota, and Aspergillus; however, a negative correlation with unclassified_f_Melanommataceae was observed. The results obtained using equation 2 showed that the harpagoside content was positively correlated with Alternaria, Apiotrichum, Plectosphaerella, Minimelanolocus, Penicillium, and unclassified_f_Cordycipitaceae, and was negatively correlated with unclassified_f_Ceratobasidiaceae, Geotrichum, and Trichoderma. The effect on the harpagoside content was the highest in unclassified_f_Cordycipitaceae, followed by Alternaria, Minimelanolocus, Penicillium, Apiotrichum, unclassified_f_Ceratobasidiaceae, Geotrichum, Trichoderma, and Plectosphaerella. In addition, equation 3 suggests that the sum of the harpagide and harpagoside content was positively correlated with the abundance of unclassified_c_Sordariomycetes, and negatively correlated with the abundance of unclassified_f_Schizoporaceae and unclassified_f_Ceratobasidiaceae. The effect on the sum of harpagide and harpagoside content was the highest in unclassified_c_Sordariomycete, followed by unclassified_f_Ceratobasidiaceae and unclassified_f_Schizoporaceae.