Metagenomic analysis of rhizosphere soil microbiota in wild and cultivated Notopterygium incisum, an umbelliferae medicinal herb

Abstract

Notopterygium incisum is not only a traditional Chinese medicine but also an endemic herb. Artificial domestication and large-scale cultivation are crucial for resolving the crisis of wild resources and the supply–demand imbalance of N. incisum, yet current techniques have failed to stably provide the herb medicine in good quality. Metagenomic analyses revealed significant differences in the rhizomicrobiota between wild and cultivated N. incisum, particularly in microbial composition, gene functions, and community assembly. The rhizomicrobiota of the wild N. incisum from 3 different sites with an altitude drop over 1400 m had a similar composition when being compared with the cultivated samples. The wild N. incisum had higher abundances of beneficial microbes, particularly Hyphomicrobiales (Rhizobiales) (21.27% on average). In contrast, the rhizosphere microbial communities of the cultivated N. incisum showed a high prevalence of functional genes involved in the pathways of DNA repair and recombination proteins, replication and repair, peptidases and inhibitors, DNA replication proteins, and transfer RNA biogenesis. The co-occurrence networks analysis indicated that the stability of the wild samples' network remained significantly more robust when nodes were proportionally removed, as the wild samples' network had approximately the same positive and negative links while the cultivated samples' network had nearly all positive links and many fewer connectors. This is related to the conclusion that wild N. incisum exhibits superior efficacy, as reported in previous studies. Additionally, it can be observed from the sampling images that the root surface of wild N. incisum has more pronounced tiny protrusions, which may be associated with rhizobial attachment, thereby enhancing the nitrogen fixation process. Importantly, the observed shifts in rhizosphere microbial communities, particularly the enrichment of beneficial rhizobia in wild plants, are closely linked to enhanced accumulation of bioactive secondary metabolites such as coumarins and volatile oils. These microbiome-driven differences are likely associated with the superior medicinal quality of wild N. incisum compared to cultivated counterparts, highlighting the pivotal role of rhizosphere microbes in shaping therapeutic efficacy.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12866-026-05269-0.

Importance

This study revealed a high abundance of Hyphomicrobiales in the rhizosphere of wild N. incisum, a feature rarely observed in non-leguminous medicinal plants. These findings provide new insights into the potential microbiome-related factors contributing to the inconsistent quality observed in artificially cultivated N. incisum. By highlighting the ecological and functional distinctions between wild and cultivated rhizosphere microbial communities, the study paves the way for the development of microbiome-based interventions—such as targeted microbial formulations—to support the sustainable cultivation and quality assurance of this medicinal herb.

Introduction

The rhizomes of Notopterygium incisum Ting ex H.T. Chang (Apiaceae), known as “Qiang huo” in Chinese, are a traditional Chinese medicine used to treat rheumatism paralysis, cold, headache, shoulder and back pain [48, 63]. The chemical constituents of N. incisum include coumarins, polyalkenes, sesquiterpenes, phenolic acids, steroids and flavonoids, which provide definite curative effects [2]. This medical herb mainly is primarily distributed in the Qinghai-Tibetan Plateau and adjacent alpine areas, including Gansu, Qinghai, Shanxi, Sichuan, and Tibet Provinces of China. As a perennial plant, it grows at a slow pace and has been excessively harvested over the past few decades, thereby posing a severe threat to the wild resources of N. incisum [71].

Artificial domestication and large-scale cultivation are crucial ways to resolve the resource crisis and the contradiction between supply and demand of N. incisum. In recent years, significant progress has been made in its artificial cultivation, particularly in key areas such as seed breeding, wild domestication, environmental ecology research, and cultivation in simulated wild environments [23, 44, 61]. However, the quality of medicinal materials is unstable, and in some cases, the content of certain index components falls below the standards set by the 2020 edition of the “Pharmacopoeia of the People's Republic of China”, posing potential quality risks for the industrial promotion and large-scale production of N. incisum. Therefore, it is crucial to identify effective methods to enhance the quality of N. incisum medicinal materials.

The quality of Chinese herbal medicine is influenced by multiple factors; however, it is primarily defined by its traditional therapeutic efficacy and toxicity profile. This encompasses the bioactive constituents and their quantities responsible for pharmacological effects, the anticipated physiological responses following administration, as well as parameters related to purity and safety [53, 72]. In natural environments, healthy plants invariably harbor a variety of microorganisms within the habitats they provide [52]. A plant along with its microbiota has even been regarded as a unified entity known as the "holobiont", signifying a 'unit of selection' where the interactions between the plant and the microbiome have likely co-evolved to uphold the host's functionality and fitness across ecological and evolutionary time frames [54, 74]. The concept of the “holobiont” remains a topic of debate, but it’s well-known that microbiota play crucial roles in promoting the productivity and accumulation of secondary metabolites in plants [6, 16, 24, 39]. Within the habitats provided by plants, the rhizosphere has been intensively studied as a rich and diverse microbial environment. It hosts a wide variety of microorganisms, many of which benefit plants by suppressing pathogenic invasions and facilitating nutrient acquisition from the soil [6, 9, 35, 41, 52]. The term ‘plant growth-promoting rhizobacteria’ (PGPR) has become confined to the bacterial strains that fulfill at least two of the three criteria i.e. aggressive colonization, plant growth promotion and bio-control [4, 22, 55].

The rhizosphere microbiome plays a crucial role in the growth, development, and enhancement of the quality attributes of herbs. A substantial body of research indicates that rhizospheric microorganisms influence plant quality through multiple physiological and biochemical mechanisms. Notably, they contribute significantly to nutrient cycling and uptake. For example, nitrogen-fixing bacteria are capable of converting atmospheric nitrogen into bioavailable forms that can be assimilated by plants [3], a process that is essential for the synthesis of proteins and other nitrogen-containing compounds in herbs. Moreover, rhizosphere microorganisms can influence the biosynthesis of secondary metabolites in medicinal herbs. The therapeutic properties of many medicinal herbs are primarily attributed to these secondary metabolites. Accumulating evidence suggests that specific bacterial or fungal taxa within the rhizosphere can activate or enhance plant metabolic pathways associated with the synthesis of these valuable compounds. For instance, in the case of ginseng, root exudates can selectively promote the proliferation of specific rhizosphere fungi, and a balanced composition of beneficial fungal species—such as Mortierella—can enhance rhizome development and elevate ginsenoside content [47]. The rhizosphere microbiome also contributes to plant resilience against both biotic and abiotic stresses. Certain pathogen-suppressive microorganisms confer protection against diseases, thereby supporting the normal growth and development of medicinal herbs. In the case of Atractylodes lancea (Thunb.) DC., the elevated abundance of key rhizosphere microbial taxa—particularly Penicillium and Streptomyces—in the Maoshan region, as compared to other regions, may represent a critical factor influencing its yield and quality attributes [50]. In the case of N. incisum, microbiome-driven regulation plays a particularly crucial role, as wild populations are generally considered to possess higher medicinal efficacy than their cultivated counterparts. This difference is strongly associated with variations in rhizosphere microbial composition—particularly the enrichment of beneficial rhizobia—that not only facilitate nutrient uptake but also promote the accumulation of essential bioactive compounds, such as coumarins and volatile oils. Therefore, examining microbiome differences between wild and cultivated N. incisum from a quality-oriented perspective offers valuable insights into how microbial community structure influences medicinal quality.

Cultivated Chinese herbal medicines are commonly perceived as having lower quality compared to their wild counterparts; however, this perception is not universally applicable. Key factors that influence the quality of cultivated Chinese herbal medicines include the growing environment, the biosynthesis and accumulation of secondary metabolites [25], the impact of soil microorganisms and mineral elements [28, 29], variations in growth cycles and harvesting strategies [38], genetic diversity and the phenomenon of variety degeneration [32], the presence of pesticide residues and environmental pollutants, as well as the influence of harvesting and post-harvest processing techniques are all critical factors that significantly influence the quality of cultivated Chinese herbal medicines. Previous studies have documented substantial differences in both the diversity and structure of rhizosphere microbial communities between wild and cultivated ginseng [13]. Additionally, earlier research has shown that wild N. incisum from higher altitudes exhibits superior quality for medicinal purposes [19, 31]. These findings provide valuable insights into enhancing the quality of cultivated herbs via microbial preparations [73].

In this study, we conducted a comparative analysis of the root morphology between cultivated and wild N. incisum, with a particular focus on how rhizosphere traits relate to medicinal quality. We employed metagenomic sequencing to characterize the diversity, composition, and functional potential of rhizosphere microbial communities across wild populations from different altitudes and cultivated samples. By integrating microbial community structure, functional pathways, and network stability, we sought to clarify how microbiome shifts under cultivation affect the biosynthesis of secondary metabolites linked to therapeutic efficacy. In particular, we analyzed the nitrogen-fixing genes within the highly abundant genus Bradyrhizobium in wild samples to evaluate their potential contributions to nutrient assimilation and the accumulation of quality-related metabolites. Collectively, this quality-focused framework enables a mechanistic understanding of how rhizosphere microbiota shape the superior medicinal properties of wild N. incisum and provides a scientific basis for improving the quality of cultivated material.

Materials and methods

Sampling

All the Notopterygium incisum Ting ex H.T. Chang samples were collected in August 2023. The voucher specimen of wild Notopterygium incisum Ting ex H.T. Chang has been deposited in Key Laboratory of Medicinal Animal and Plant Resources in Qinghai Tibet Plateau, Qinghai Province (Deposition number: QHZWYSIN2024091301; Appraiser: Qiaoyu Luo). The 3 sampling sites for wild N. incisum ranged in altitudes from 2428 to 3868 m above the sea level. The herb grew within a bush dominanted by Berberis julianae or in association with Sibiraea angustata or Rhododendron yushuense (Table 1). Based on the rhizome morphology, size, and the number of stem base scars, rhizosphere samples of the wild N. incisum with a growth period of 6 to 8 years were collected. The cultivated N. incisum was grown at Huangzhong County for 7 years, with an altitude of 2952 m. It was surrounded by no woody vegetation, but instead coexisted with the perennial herbaceous plant Potentilla chinensis Ser. To collect the rhizosphere soils, N. incisum seedlings were carefully excavated from the soil using a shovel, and the roots were gently shaken to remove loosely attached soil. The roots were collected and washed three times with phosphate-buffered saline (PBS). The washed-off soil was transferred to a 50 mL centrifuge tube, centrifuged, and stored at − 80 °C until DNA extraction. In total, three rhizosphere soil samples were obtained for each site.

Table 1.

Geospatial coordinates and dominant plants of the sampling sites

Sites	Longitude	Latitude	Altitude	Dominant woody plants
Wild 1 (W1)	104°1′57	35°13′24	2428 m	Berberis julianae
Wild 2 (W2)	99°52′4.891	38°16′49.272	3209 m	Berberis julianae, Sibiraea angustata (Rehder) Hand.-Mazz
Wild 3 (W3)	97°11′57.9228	32°54′47.7576	3868 m	Berberis julianae, Rhododendron yushuense
Cultivated (CU)	101°49′46.2978″	37°8′56.4339″	2952 m	-

DNA Extraction and metagenomic sequencing

Community DNA was extracted from the rhizosphere soil using the FastDNA Spin Kit for Soil (MP Biomedicals, Solon, OH, United States) according to the manufacturer's instructions. DNA purity and concentration were assessed using a NanoDrop 2000 spectrophotometer (Thermo Fisher, Waltham, MA, USA). Metagenomic sequencing of the total rhizosphere soil DNA was performed by Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China). The sequencing libraries were generated using the NEBNext Ultra DNA Library Prep Kit for Illumina (New England Biolabs, Ipswich, MA, United States) following the instructions of the manufacturer. Briefly, the DNA samples were fragmented by sonication to an average size of 350 bp, followed by end repair, A-tailing, and ligation with full-length adapters for Illumina sequencing with subsequent PCR amplification. After purification, the libraries were explored for size distribution using the Agilent 2,100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, United States) and quantified by real-time PCR. Index-coded samples were clustered using the cBot Cluster Generation System based on the guidance of the manufacturer (Illumina Inc., San Diego, CA, United States). Library preparations were sequenced on the Illumina NovaSeq platform (Illumina Inc.). The raw sequence reads would be deposited in the National Genomics Data Center (https://ngdc.cncb.ac.cn/) for public access.

Sequencing data treatment, gene prediction, and abundance analysis

Raw data were filtered by removing: (1) low-quality reads with a default quality threshold value < 38 over a certain portion (default length of 40 bp); (2) reads with N bases smaller than 10 bp; (3) reads with more than 15 bp overlapping with adapters. The clean data were assembled using MEGAHIT software (v1.0.4-beta) [27]. Mixed assembly was performed using SOAPdenovo (V2.04)/MEGAHIT (v1.0.4-beta). Additionally, fragments shorter than 500 bp in all scaftigs generated from both single and mixed assemblies were filtered out for statistical analysis [65]. Open reading frames (ORFs) were predicted from the assembled scaftigs (> 500 bp) using MetaGeneMark (V2.10, http://topaz.gatech.edu/GeneMark/) software, and sequences shorter than 100 nt were filtered out. CD-HIT (v4.5.8, http://www.bioinformatics.org/cd-hit) was used to remove redundancies and generate a unique initial gene catalog with the following parameter settings: -c 0.95, -M 0, -G 0 -aS 0.9, -g 1, -r 1, -d 0 [49, 65]. The clean data from each sample were mapped to the initial gene catalog using Bowtie 2.2.4 software, and the unigenes were acquired. The parameters were set in line with a previous description [55]. Based on the number of mapped reads and gene length, the abundance of each gene in each sample was calculated as follows: Gk = rkLk⋅1∑ni = 0riLi, where r represents the number of reads mapped to genes and L represents the gene length. The summary statistics, core-pan gene analysis, and Venn diagram of gene counts were based on the abundance of each gene in each sample.

Taxonomic predictions and functional annotations

The unigenes were aligned to sequences of the different kingdoms of Bacteria, Fungi, Archaea, and Viruses by DIAMOND software[7]. Subsequently, the aligned sequences were extracted from the Non-Redundant (NR) database of NCBI (version 2018–01–02, https://www.ncbi.nlm.nih.gov/). For the final aligned results for each sequence, the alignment with the smallest e-value [60] was used for classification via the lowest common ancestor (LCA) algorithm implemented in MEGAN [18] to verify the species annotation information. The number of genes and the abundance information for each sample at each taxonomy hierarchy were obtained. The abundance of a species in each sample was defined as the sum of gene abundances attributed to that species. The gene count for a species in a sample was determined by the number of genes with a nonzero abundance. DIAMOND software (version 0.9.9) was used with the settings –BLASTp and -e 1e-5 to align the unigenes to the KEGG (Kyoto Encyclopedia of Genes and Genomes) and eggNOG (Evolutionary Genealogy of Genes: Non-supervised Orthologous Groups, version 4.5) databases [21]. The relative abundance of each annotated functional hierarchy was subsequently determined. The functional annotation results and gene abundance data provided the number of genes associated with each taxonomic hierarchy was obtained for each sample. The relative abundance of each functional hierarchy was calculated as the sum of the relative abundances of genes annotated to that specific functional level. Based on the functional annotation results and gene abundance table, a gene count table for each sample at each taxonomy hierarchy was generated. The gene number for a function in a sample was equal to the number of genes annotated to the function with nonzero abundance.

Analysis of microbial co-occurrence networks

The co-occurrence network was constructed using the Sparse Correlations for Compositional (SparCC) data algorithm [14] implemented in FastSpar version 0.0.10 [59]. Visualization of the network was performed using Gephi 0.9.6 software under the following conditions: |r|> 0.6 and p value < 0.05. Topological indices, including node number, edge number, density, average degree, centrality, and modularity, were calculated. Within-module connectivity (Zi) and among-module connectivity (Pi) were calculated for the module hubs (nodes with Zi ≥ 2.5 and Pi < 0.62), connectors (nodes with Zi < 2.5 and Pi ≥ 0.62), and network hubs (Zi ≥ 2.5 and Pi ≥ 0.62).

Statistical analysis

The alpha diversity indices were calculated using the function diversity in the “vegan” package. One-way analysis of variance (ANOVA) and Scheffe's post-hoc test were conducted to assess significant differences between groups. Principal coordinate analysis (PCoA) based on Bray–Curtis distances was applied to determine beta diversity variation across samples. Unweighted UniFrac distance matrices were analyzed using the analysis of similarities (ANOSIM) for categorical variables. The metagenomic taxa with significant differences were statistically evaluated using linear discriminant analysis effect size (LEfSe) with LEfSe software [43]. The histograms were utilized to depict the relative abundance of these significantly differentiated metagenomic taxa.

Results

Root morphology of wild and cultivated N. incisum

Representative root images of wild and cultivated N. incisum distinctly illustrated significant morphological variations (Fig. 1), raising the possibility that root surface features may be associated with differences in medicinal quality. Wild roots exhibited notably darker pigmentation, pronounced surface ridges, and abundant tiny protrusions, whereas cultivated roots presented smoother textures and lighter coloration.

Fig. 1.

Images of wild and cultivated rhizomes of N. incisum

Sequencing data and taxonomic composition

In total, 1,080,426,648 raw sequence reads with 162,289,886,826 bp raw data were obtained from 12 shotgun metagenome libraries, ranging from 69,109,348 to 106,202,334 reads per sample. Following quality filtering, over 98.2% of the reads were retained, and the mean percentage of reads achieving a Q30 quality score exceeded 95%. Taxonomic classification revealed a total of 13,733 species, with each sample containing between 11,084 and 11,260 species. Venn diagrams revealed that wild samples contained a greater number of unique species compared to cultivated samples (Fig. 2A). However, upon separately analyzing the wild samples from three different sites and cultivated samples, each of the four groups showed a similar proportion of unique species. These unique species accounted for roughly 2.5% of the total species at each site (Fig. 2B). Compared to fungi and archaea, bacteria constituted the predominant components of the rhizosphere microbiota, representing more than 95% of the microbial community. The dominant phyla were Pseudomonadota and Actinomycetota (Fig. S1). At the lower taxonomic levels, the orders Hyphomicrobiales and Burkholderiales were the most predominant, accounting for more than 30% abundance in the majority of samples (Fig. S2), and the most abundant genera were Bradyrhizobium and Streptomyces, with mean proportion of 8.37% and 7.87%, respectively (Fig. S3).

Fig. 2.

Venn diagrams of the taxonomic species in different groups. A Wild group and cultivated group. B The four groups consisted of W1, W2, W3 and CU. W1, W2, and W3 refer to samples collected from 1, 2, and 3 wild N. incisum sampling sites, respectively, and CU refers to samples from the cultivated N. incisum site

Microbial community dissimilarities between the different groups

Among the different sampling sites, W2 exhibited the highest median species richness, W1 had the highest evenness (the lowest median value of Simpson index), and CU had the highest Shannon index median value. However, the differences in all the alpha-diversity indices were not significant (Fig. S4). Comparisons between wild and cultivated samples also revealed non-significant differences in species richness indices, including Observed, ACE, and Chao1, but Shannon and Simpson indices were significantly different (Fig. S5).

Principal coordinate analysis (PCoA) results indicated that the rhizosphere microbial communities could be distinguished based on sampling sites. Within-site differences among wild samples were smaller compared to the differences between wild and cultivated samples (Fig. 3). ANOSIM results confirmed this finding, with R values of 0.3457 (p = 0.008) for different sites and 0.867 (p = 0.005) for wild versus cultivated samples, respectively.

Fig. 3.

Principal coordinate analysis (PCoA) illustrating variations in the rhizosphere microbial community based on Bray–Curtis distances. W1, W2, and W3 refer to samples collected from 1, 2, and 3 wild N. incisum sampling sites, respectively, and CU refers to samples from the cultivated N. incisum site

Differences in specific microbes

The linear discriminant analysis effect size (LEfSe) results indicated there were no significant different phyla among the different sites. However, the relative abundances of the phyla Bacillota, Acidobacteriota, Nitrososphaerota, Chloroflexota, and Nitrospirota were higher in cultivated samples compared to wild samples (Fig. S6). At the order level, Hyphomicrobiales and Micrococcales were the only 2 orders that showed significant differences between the sites (Fig. S7). Nevertheless, a greater number of orders exhibited significant differences between wild and cultivated samples. For examples, Hyphomicrobiales, Micromonosporales, and Pseudonocardiales were significant higher abundant in wild samples, while cultivated samples harbor higher abundant of the orders Micrococcales, Pseudomonadales, Bacillales, Vicinamibacterales, Cytophagales, Nitrososphaerales, and Nitrospirales (Fig. S8). At the genus level, Bradyrhizobium, Actinoplanes, and Micromonospora had significantly higher abundances in wild samples, while Arthrobacter, Pseudarthrobacter, Microvirga, and Pseudomonas had higher abundances in cultivated samples (Fig. 4). When samples were grouped by site, significantly fewer genera exhibited differences in relative abundance (Fig. 4). At the species level, numerous members of the genus Bradyrhizobium showed significant differences both among sites and between the wild and cultivated samples (Figs. S9 and S10).

Fig. 4.

Histogram of differential abundance analysis for different groups at the genus level. A Wild group and cultivated group. B The four groups consisted of W1, W2, W3 and CU. W1, W2, and W3 refer to samples collected from 1, 2, and 3 wild N. incisum sampling sites, respectively, and CU refers to samples from the cultivated N. incisum site

Microbial co-occurrence networks

Two microbial ecological networks were constructed by processing SparCC correlation coefficients with the SpiecEasi method, aiming to elucidate the distinct assembly patterns of microbial communities in the rizhosphere of wild and cultivated N. incisum (Fig. 5). The network of the wild samples comprised 325 nodes connected by 2,668 links, with approximately half positive links (56.63%) and significant correlation forming 4 modules. The network of the cultivated comprised 335 nodes connected by 11,389 links, with nearly all positive links (99.98%) and significant correlation forming 3 modules. The network of the cultivated samples showed a higher average degree, but lower modularity compared to that of the wild samples (Table S1). Network stability was assessed using both the average degree and the natural connectivity index. The results indicated that the stability of the cultivated network decreased as the ratio of randomly removed nodes increased, whereas the stability of the wild network remained significantly more robust when nodes were proportionally removed (Fig. S11). The analysis of the topological roles of nodes in the networks revealed that the wild network contained significantly more key nodes (species) compared to the cultivated network. Specifically, the wild network included 9 module hubs and 188 connectors, whereas the cultivated network had only 2 module hubs and 3 connectors (Fig. 5C and D). Among the 9 module hubs in the wild network, 4 belonged to the genus Bradyrhizobium.

Fig. 5.

Networks and Zi-Pi plots for wild and cultivated samples. A Network plot of wild samples. B Network plot of cultivated samples. C Zi-Pi plot of wild samples. D Zi-Pi plot of cultivated samples. Node colors represent network modules, whereas red and green edges indicate positive and negative correlations, respectively

Differences in functional potential of the rhizosphere microbiota

PCoA results indicated that functional potential of rhizosphere microbial communities could not be well-distinguished by the sites within the wild N. incisum, but very different from the cultivated samples (Fig. 6). LEfSe results confirmed this trend and provided additional insights. At the KEGG level 3 pathway, ABC transporters, quorum sensing, degradation of benzoate and fatty acid, flagellar assembly, and metabolism of branched-chain amino acids (valine, leucine and isoleucine) displayed higher functional gene abundance in the wild samples compared to cultivated samples (Fig. 7). Reversely, the rhizosphere microbial communities of the cultivated N. incisum had a high abundance of functional genes in the pathways related to DNA repair and recombination proteins, replication and repair, peptidases and inhibitors, DNA replication proteins, transfer RNA biogenesis, and etc. (Fig. 7). LEfSe analysis of KEGG Orthology indicated that the wild samples had higher abundance for several genes related to branched-chain amino acid transport system substrate-binding protein (k01995, k01996, k01997, k01998, and k01999) (Fig. S12), which were related to the pathways of ABC transporters and quorum sensing. Considering the higher abundance of rhizobia, particularly members of the genus Bradyrhizobium, the profile of nitrogen fixation genes in wild and cultivated samples was mined. Results indicated that only 3 nitrogen fixation genes were present: nitrogenase molybdenum-iron protein beta chain nifK (K02591), nitrogenase molybdenum-iron protein alpha chain nifD (K02586) and nitrogenase iron protein NifH (K02588). The abundance of these 3 genes varied in different samples but was not significantly different between the sites (Fig. S13).

Fig. 6.

Principal coordinate analysis (PCoA) showing the variations in the rhizosphere microbial community based on the KEGG level 3 pathway

Fig. 7.

Histogram of differential abundance analysis of functional composition between the wild and cultivated at the KEGG level 3 pathway

Discussion

Correlation between root morphology, rhizosphere microbiota, and quality of wild and cultivated N. incisum

The rhizosphere morphology images revealed significant differences between wild and cultivated N. incisum (Fig. 1). Wild roots displayed notably darker pigmentation, more pronounced surface ridges, and a greater abundance of tiny protrusions compared to cultivated roots. These observed morphological disparities likely reflect differences in rhizosphere conditions, such as soil characteristics and microbial community composition. Our results further demonstrated that the rhizosphere of wild N. incisum harbored high abundances of beneficial rhizobia (Hyphomicrobiales), potentially enhancing root development through nitrogen fixation, phytohormone secretion, phosphorus solubilization, and iron chelation [34, 40]. Additionally, the numerous tiny protrusions on the wild root surface may serve as colonization sites for rhizobia, facilitating enhanced nitrogen uptake and nutrient absorption efficiency [20]. Consistent with these morphological features, wild N. incisum also exhibited higher concentrations of bioactive coumarins and volatile oils, correlating positively with increased microbial diversity and rhizobial abundance in the rhizosphere [23]. Collectively, these findings highlight the pivotal role of rhizosphere-associated microbiota, particularly rhizobia, in shaping root morphology and enhancing the medicinal quality of wild N. incisum.

Substantially different rhizomicrobiota between the wild and the cultivated N. incisum

The metagenomic sequencing analysis revealed that bacteria constituted the predominant component of the rhizomicrobiota associated with N. incisum, accounting for 99.13% and 99.02% of the total microbial community in the rhizosphere of wild and cultivated N. incisum, respectively. Our research demonstrated that the predominant bacterial phyla in the rhizosphere of both cultivated and wild N. incisum were Pseudomonadota and Actinomycetota. However, prior research indicated that the predominant bacterial phyla in the rhizosphere of both cultivated and wild ginseng were Proteobacteria and Acidobacteria, potentially attributable to species-specific differences [13]. LEfSe results revealed that the relative abundances of the phyla Bacillota, Acidobacteriota, Nitrososphaerota, Chloroflexota, and Nitrospirota were significantly higher in cultivated samples compared to wild samples (Fig. S6). This phenomenon can be attributed to two primary factors; Firstly, nitrogen fertilizer is an essential and indispensable component for cultivated soil management. Previous studies [5, 56] have demonstrated that the phyla Nitrososphaerota and Nitrospirota play crucial roles in the nitrogen cycling process within soil ecosystems. Consequently, these taxa exhibit higher abundances in cultivated soil samples where nitrogen fertilizers were frequently applied. Secondly, the use of nitrogen fertilizer contributes to soil acidification. Further research [26, 33] has identified Nitrososphaerota and Acidobacteriota as dominant microbial groups in acidic soil environments. This finding also elucidates their elevated abundances in cultivated soil samples. In addition, this study found that the relative abundance of Arthrobacter, Pseudarthrobacter, and Pseudomonas in cultivated samples was significantly higher than that in wild samples (Fig. 4). This phenomenon might be attributed to the pollution pressure resulting from the long-term use of chemical fertilizers and pesticides in cultivated soils. Previous studies have confirmed that these three genera possess the ability to degrade polycyclic aromatic hydrocarbons (PAHs) in soil [37, 67]. Consequently, in polluted cultivated environments, microorganisms with specific pollutant degradation capabilities might acquire a growth advantage by adapting to pollution stress, thereby exhibiting a higher relative abundance within the microbial community structure.

Among the wild samples, Rhodoplanes, Pseudolabrys, Ancylobacter, Breoghania, Blastochloris, Starkeya, and Phreatobacter exhibited the highest relative abundance in sample W3 (Fig. 4). Rhodoplanes, as a member of the phylum Proteobacteria, not only demonstrates strong environmental adaptability but also plays a central role in soil nitrogen metabolism [68]. Its presence has been documented in alpine meadow soils at elevations of up to 4,000 m on the Qinghai-Tibet Plateau [70]. Starkeya, also belonging to the phylum Proteobacteria, has been shown to play a crucial ecological role, particularly in carbon and sulfur metabolism [42]. Therefore, it is reasonable to infer that its exceptional environmental resilience may contribute to its high relative abundance in the high-altitude sample W3.

In this study, the rhizosphere soil of wild N. incisum exhibited a greater number of microbial community network modules compared to cultivated samples. This suggested that the microbial community in the rhizosphere of wild N. incisum might possess more complex ecological functions and interactions, with distinct modules working synergistically to maintain ecosystem stability. Previous studies highlighted that increased network modularity served as a critical strategy for sustaining community stability [11]. In addition, the results of the network stability analysis further indicated that the structure of the rhizosphere microbial community in wild N. incisum was significantly more stable compared to that in cultivated samples (Fig. S11). In contrast, the rhizosphere soil of cultivated N. incisum might experience alterations in its nutrient structure due to prolonged use of chemical fertilizers; such changes could lead to the proliferation of certain microorganisms adapted to specific niches while inhibiting others, ultimately affecting microbial diversity and the modular structure of the community. The node connections within the network structure of the cultivated samples were predominantly positive, providing strong support for the hypothesis regarding this possibility. In the network graph of cultivated samples, Arthrobacter and Agromyces exhibited the highest degree values. Both genera belong to the phylum Actinobacteria and display a positive correlation. Long-term agricultural fertilization not only results in nitrogen accumulation in soils but may also lead to the persistence of pesticides and heavy metal residues. Accumulating evidence has demonstrated that prolonged nitrogen application enhances the abundance of Actinobacteria [12]. Moreover, both Arthrobacter and Agromyces have been shown to degrade specific soil contaminants, indicating their potential roles in soil bioremediation [15, 69].

Studies have shown that in the rhizosphere soil microbial community of wild N. incisum, the abundance of ABC transporters, benzoic acid degradation, and arginine-proline metabolic functions tends to be notably higher compared to cultivated N. incisum. This difference seems to be closely connected to the distinct metabolic features of the wild variety. Wild N. incisum is known to synthesize and release a considerable amount of hydrophobic coumarin compounds—such as ligustilide and isopimpinilin—into the rhizosphere. In order for rhizosphere microorganisms to utilize these compounds as carbon sources, they naturally rely on ABC transporters to facilitate active transmembrane transport [64]. Moreover, genomic research has indicated that key enzymes involved in coumarin biosynthesis—like p-coumaroyl-CoA 2'-hydroxylase (C2′H)—are highly active in wild N. incisum, which supports the efficient production of coumarins [30]. As part of this metabolic process, byproducts such as benzoic acid are generated, gradually accumulating in the rhizosphere environment. These accumulated benzoic acid compounds serve as valuable metabolic resources for surrounding microorganisms. As a result, the abundance of benzoic acid degradation functions in the rhizosphere microbial community of wild N. incisum also tends to be significantly higher. Besides, previous studies have demonstrated that when certain medicinal plants are subjected to environmental stress, the levels of bioactive compounds in their secondary metabolites increase significantly [57]. From a physiological perspective, proline functions as a key osmotic regulator, effectively maintaining intracellular osmotic balance and providing a stable internal environment for normal cellular metabolism under stress conditions. Arginine metabolism, in contrast, is closely linked to the synthesis of various bioactive molecules—these compounds not only contribute to plant defense responses under adverse conditions but may also indirectly modulate secondary metabolic pathways, thereby influencing the accumulation of bioactive compounds. Collectively, these findings suggest that distinct rhizosphere microbiomes drive variations in quality formation in N. incisum, further supporting the observation that wild N. incisum typically exhibits superior quality compared to its cultivated N. incisum.

Significantly high abundance of Hyphomicrobiales in the rhizosphere of N. incisum

In the rhizosphere of wild N. incisum, the abundance of the order Hyphomicrobiales was 21.27% on average, which is notably high for a non-legume plant. On average, Hyphomicrobiales accounted for 9.2% of soil bacteria, with a relative abundance ranging from 3.7% to 18.2% (Jones 2015). According to the List of Prokaryotic Names with Standing in Nomenclature (accessed 2nd March 2025), the order Hyphomicrobiales within the class Alpha proteobacteria comprises 41 families encompassing over 150 valid genera. From 2005 to 2020, this order was known as Rhizobiales. All known alpha-rhizobia, which are alpha proteobacterial nitrogen-fixing legume symbionts, are found within the order Hyphomicrobiales [17]. In the rhizosphere of N. incisum, the predominant families within the order Hyphomicrobiales were Nitrobacteraceae (9.3%), Phyllobacteriaceae (2.4%), Rhizobiaceae (2.3%), and Methylobacteriaceae (1.4%). The predominant genera included Bradyrhizobium (8.4%), Mesorhizobium (1.7%), Rhizobium (1.00%), and Methylobacterium (0.7%). Root growth promotion is a conserved trait of commensal Rhizobiales, and members from these predominant families and genera have been frequently reported as beneficial plant-associated bacteria [34, 40].

Rhizobiales can promote the growth of non-legume crops in multiple ways, including the production of phytohormones, facilitating nutrient uptake, such as by solubilizing phosphorus, addressing iron deficiencies through siderophore production, reducing ethylene levels under drought conditions via the ACC deaminase enzyme, as well as indirectly through pathogen biocontrol and inducing systemic resistance in the host plants [10]. Inside the rhizosphere of N. incisum, Bradyrhizobium is the predominant genus and is primarily recognized for its role as an N2-fixing symbiont that nodulates leguminous plants. This function is supported by the presence of nif and nod gene clusters [36].

Moreover, member of Bradyrhizobium genus have been confirmed to perform various functions, including phosphate solubilization, phytohormones production or regulation, siderophores production, and pathogens suppression [46]. Several studies have demonstrated different aspects of Bradyrhizobium's functions capabilities. For instance, radish plants inoculated with Bradyrhizobium japonicum exhibited a 15% increase in dry matter yield [1]. Inoculating Bradyrhizobium into wild rice demonstrated nitrogen-fixing activity through the acetylene reduction assay [8]. The biosynthesis of indole-3-acetic acid (IAA) in rhizobia primarily involves the indole-3-pyruvate pathway (IPyA), which has also been identified in Bradyrhizobium [52]. Additionally, isolates of Bradyrhizobium japonicum has been reported to exhibit antagonistic effects against plant pathogens such as Fusarium solani, Macrophomina phaseolina, and R. solani [45].

Importantly, the enrichment of Hyphomicrobiales in the rhizosphere of wild N. incisum not only enhances nutrient uptake and stress tolerance but also fosters a favorable microenvironment that correlates with the biosynthesis of quality-defining secondary metabolites, including coumarins and volatile oils. These microbiome-mediated improvements offer a mechanistic basis for the observed superior medicinal efficacy of wild N. incisum compared to its cultivated counterparts. Therefore, the high abundance of Hyphomicrobiales serves as a key microbial indicator that connects rhizosphere ecology with the therapeutic quality of this traditional medicinal herb.

Significance regarding the modern cultivation and quality assurance of N. incisum

In current practices of Chinese herbal medicine cultivation, core quality challenges primarily arise from two aspects: low content of target active compounds, and excessive pesticide residues and heavy metals [58]. These issues directly constrain the safety and clinical efficacy of medicinal materials. The emergence of these problems is closely associated with multiple interrelated factors: first, continuous cropping obstacles are widespread and severe, leading to imbalances in rhizosphere soil microecology and the accumulation of pathogenic microbial communities [62]; second, the ecological cultivation technology system remains underdeveloped, with insufficient regulation of key environmental factors that influence the formation of medicinal material "authenticity"; third, soil nutrient management lacks precision, failing to meet the metabolic demands for secondary metabolite synthesis in medicinal plants.

Current cultivation technologies include the following: First, the application of intelligent monitoring systems enables real-time regulation of the growth environment and physiological status of medicinal plants, offering data-driven support for precision management. However, high initial equipment and maintenance costs hinder its widespread adoption, particularly in small- and medium-scale planting bases. Second, gene editing technology has demonstrated significant potential for the targeted improvement of genetic traits in medicinal plants, although its application remains largely confined to laboratory research [51]. In addition, wild-simulated planting models, which mimic the natural habitats of medicinal plants, have shown promise in enhancing both the "authenticity" and quality consistency of cultivated materials [66]. Meanwhile, rhizosphere microbiome regulation presents a novel and promising approach for mitigating continuous cropping obstacles and restoring soil microecological balance.

Artificial domestication and large-scale cultivation play a vital role in resolving the crisis of wild resources and the supply–demand imbalance for N. incisum. Research has suggested that a key factor contributing to the decline in quality of cultivated N. incisum is the notably lower levels of coumarin compounds compared to its wild N. incisum [30]. However, the current cultivation techniques are not yet capable of consistently producing high-quality rhizomes. Results from this study suggest that the solution could potentially be applied to the rhizomicrobiota, as wild N. incisum exhibited both stable and high-quality herb roots and similar microbial communities in the rhizosphere. In the future, it is worthwhile to isolate specific microbial strains, particularly those belonging to Rhizobiales, from wild samples and inoculate them into the cultivated N. incisum to evaluate their function. Subsequently, synthetic microbiota approaches could be employed to develop N. incisum specific microbial formulations, which would be helpful in modern cultivation practices and quality assurance of N. incisum.

Although this study revealed the differences in rhizosphere microbial communities between wild and cultivated N. incisum, several limitations arising from the research design and external constraints remain to be addressed in future investigations. First, potential confounding factors related to growth conditions and material sources were not fully controlled or characterized: wild and cultivated samples might differ in their original growth sites, and key soil environmental variables—such as soil physicochemical properties—were not concurrently collected during sampling. Additionally, the consistency of planting materials (e.g., seed provenance, seedling genotype) between wild and cultivated groups was not explicitly verified, which may also introduce variations affecting microbial community assembly. Second, the sample size was relatively limited, with only three replicates per treatment. Although this meets the minimum requirement for replication, it may compromise the statistical robustness of the results. Future research will aim to strictly control the consistency of growth sites and planting materials, incorporate more comprehensive environmental data (including soil and climatic parameters), and increase the sample size to improve analytical depth and result interpretability, thereby enabling a more precise understanding of the interactions between rhizosphere microbial communities.

Authors’ contributions

Yonggui Ma and Huichun Xie conceived and designed the experiment. Tianqing Feng, Jun Shang, and Yonggui Ma conducted the fieldwork, while Tianqing Feng, Yanrong Qin, Jitao Zhang, and Shiyan Cheng completed the laboratory analyses. Gaosen Zhang and Tianqing Feng were responsible for data analysis. The initial draft of the manuscript was prepared by Tianqing Feng, Jun Shang, and Juan Li. All authors contributed to critically revising the manuscript.

Funding

The work was supported by Key Research and Development and Transformation Project of Qinghai Province (2026-QY-203), and Assurance Project of Ecological Planting and Quality of Daodi Herbs (Science and Technology Department of the State Administration of Traditional Chinese Medicine [2020] No. 153).

Data availability

The raw sequence reads data have been deposited in the National Genomics Data Center (NGDC) under the GSA accession number CRA006850, which is publicly accessible at https://ngdc.cncb.ac.cn/gsa/browse/CRA006850.

Declarations

Ethics approval and consent to participate

All plant materials were collected following national and international standards and local laws and regulations. Wild N. incisum is neither classified as an endangered species nor subject to protection regulations, so no special permissions were required for its collection. The collection of N. incisum cultivation samples has been authorized by the Key Laboratory of Medicinal Plant and Animal Resources of the Qinghai-Tibet Plateau.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.