Impact of Root Rot Disease of Zanthoxylum armatum on Rhizosphere Soil Microbes and Screening of Antagonistic Bacteria

: Zanthoxylum armatum , a signiﬁcant forest plant in southwestern China, is crucial for preserving soil and water resources. However, the presence of root rot disease has led to plant death, impacting the pepper sector. Effective control measures for this disease are still lacking. Rhizosphere microorganisms play a vital role in plant health by inhibiting plant pathogens and inducing plant resistance. This research aimed to isolate and characterize the pathogen responsible for root rot disease in Z. armatum . Comparative analysis of fungal and bacterial communities in the rhizosphere soil of healthy and diseased plants revealed Fusarium solani as the pathogenic fungus causing root rot disease. Diseased plants had a higher occurrence of Fusarium spp., while disease-free plants had a higher abundance of ecologically beneﬁcial microbial communities that could potentially serve as biocontrol agents. Three bacterial strains ( Bacillus subtilis , Bacillus amyloliquefaciens , and Bacillus siamensis ) were identiﬁed as effective biocontrol agents, inhibiting the growth of the pathogenic fungus F. solani both in vivo and in vitro. This study deepens our understanding of the rhizosphere soil microbial community differences between diseased and healthy Z. armatum , providing potential biocontrol bacteria to enhance plant resistance against root rot disease.


Introduction
Zanthoxylum armatum, a small-sized tree or shrub, is predominantly found in China, Japan, Korea, India, and Pakistan. In China, it is primarily distributed in the southwestern region (21.37 • N, 85.121 • E) [1]. This species holds great significance in afforestation, rural industrial restructuring, and improving farmers' livelihoods, owing to its capabilities in soil and water conservation [2]. Unfortunately, root rot disease poses a significant threat to Z. armatum production, resulting in substantial economic losses in the agricultural and forestry sectors worldwide [3]. Root rot disease of Zanthoxylum spp. primarily affects the root system, leading to discoloration and decay of the affected roots. Infected roots emit a foul smell, and the root epidermis easily detaches from the woody tissues. In severe cases, the woody tissues become blackened, spreading along the main root towards the root-stem junction. Above-ground parts of the affected plants show symptoms such as reduced leaf size, chlorosis (yellowing), stunted fruit development, and incomplete branch growth. Ultimately, this condition results in wilting and the eventual death of the plants. Conventional chemical pesticides have proven ineffective in controlling soil-borne plants. Furthermore, we used a culture-based approach to isolate and screen biocontrol bacteria from the rhizosphere soil, followed by testing the strains for their disease-control and growth-promoting effects. By understanding the impact of root rot disease on rhizosphere microbiota, we aim to gain better insight into the mechanisms of plant disease occurrence and develop appropriate measures for disease control and treatment.

Experimental Design
Siwei Village, Taihe Town, Meishan City, Sichuan Province, China (30°04′33″ N; 103°50′54″ E) is where the sample location is situated. This region had an average annual temperature of 17.9 degrees and 855.5 mm in typical annual precipitation in 2021. It is one of China's main centers for the production of Sichuan peppers. The most significant variety among these peppers is Z. armatum, and August is the season when it is most likely to contract the illness known as root rot. Twenty sampling locations were established in two 30 m × 30 m Z. armatum plant growth areas. Among the sampling locations, ten were found to be free from root rot (Figure 1a,c), while ten others exhibited signs of root rot ( Figure  1b,d). The five-point sampling method was employed to collect samples in each plot, with a distance of 500 m between each sampling site. A total of 20 samples were collected, including 10 healthy samples (labeled as HR1-10, HRS) and 10 samples affected by root rot (labeled as DR1-10, DRS). Prior to sample collection, the ground was cleared of dead branches, fallen leaves, and debris. Samples were then collected within a circular area of 0.5 m in diameter, centered around the main stem of the plant. A gentle brushing was performed to remove 2 mm thick soil adhering to the roots, which served as rhizosphere soil samples. The root depth ranged from 0 to 20 cm. The collected samples are placed in 50 mL centrifuge tubes, numbered, sealed, and brought back to the laboratory for storage at −20 °C. The root tissues were stored in a refrigerator at −4 °C.

Comparing the Physicochemical Characteristics of Diseased and Healthy Plants' Rhizosphere Soil
Soil pH value was determined by the electrode method. The potassium dichromate oxidation spectrophotometric method was employed to determine soil organic matter (SOM) [25]. The Kjeldahl method was used to measure the concentration of total nitrogen (TN) in the soil [26]. The molybdenum antimony blue colorimetric method was used to detect the concentration of total phosphorus (TP) in the soil [27]. The ammonium acetate method was used to measure the concentration of total potassium (TK) in the soil [28].

Isolating and Characterizing the Pathogen Causing Root Rot Disease in Z. armatum
Pathogen Isolation and Purification: Following Fang's plant pathology research method [29], ten bamboo leaf pepper root systems displaying symptoms of root rot were collected for the isolation of the fungal pathogens, with slight modifications (Supplementary Materials).
Pathogenicity Test: To confirm the pathogenicity of the isolated fungal strains, a modified version of the bare-root inoculation test was conducted [30]. Twenty healthy Z. armatum plants (1 year old) were carefully washed with tap water to clean their root systems, which were then surface sterilized. The pathogenic isolate HJGFB-1, grown on PDA for 7 days, was used. The spore suspension was adjusted to 1 × 10 7 spores/mL. Sterile filter paper strips (approximately 1 cm wide × 8 cm long) were immersed in the spore suspension for 30 s. Meanwhile, control paper strips were soaked in sterile water for 30 s. The treated paper strips were placed at the top and middle parts of the roots (with 4 replicates for each treatment). The middle part of the root was incised with a sterile surgical blade (approximately 1 cm long and 1 mm deep). Each inoculated root was then wrapped with four layers of sterile, moist gauze and placed in a folded plastic bag. The bags were incubated at 22 ± 1 • C in darkness for 3 weeks. Subsequently, the pathogen was recovered from symptomatic root portions and identified through ITS sequence analysis.
Pathogen Identification: Fungal genomic DNA was extracted from the pathogenic isolates using the Fungi DNA Extraction Kit (TIANDZ, Beijing, China). PCR was performed to amplify the internal transcribed spacer (ITS), DNA-dependent RNA polymerase II subunit (RPB2), β-tubulin (TUB), and translation elongation factor 1-α (TEF1) genes, following the protocol by O'Donnell (1996) [31]. After gel electrophoresis of the PCR products, the amplified PCR products of the antagonistic strains were sent to a specific company for sequencing. The obtained sequences were subjected to BLAST analysis in NCBI GenBank (http://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 10 December 2020) to download highly similar reference gene sequences. The MEGA software version 5.0 was used to manually align and remove redundant sequences as required, followed by the construction of a phylogenetic tree containing multiple genes using PhyloSuite (jushengwu.com, accessed on 25 December 2020). The phylogenetic status of the antagonistic bacteria was determined based on the analysis of the gene phylogeny. The nucleotide sequences were deposited in the GenBank database and assigned accession numbers MW350047, MW350046, MW350047, and MW446809, respectively.

Characterization of Rhizosphere Soil Microbial Microbiome
Microbial DNA was extracted from each sample using the standard protocol of the PowerSoil DNA Isolation Kit (MO BIO Laboratories, Inc., Carlsbad, CA, USA) [15]. The quality and quantity of DNA were evaluated using the 260 nm/280 nm and 260 nm/230 nm ratios. The V3-V4 region of the bacterial 16S rRNA gene was amplified using the forward primer 338F (5 -ACTCCTACGGGAGGCAGCA-3 ) and the reverse primer 806R (5 -GGACTACHVGGGTWTCTAAT-3 ) [32]. The ITS region of fungal DNA was amplified using the forward primer ITS1 (5 -CTTGGTCATTTAGAGGAAGTAA-3 ) and the reverse primer ITS2 (5 -GCTGCGTTCTTCATCGATGC-3 ) [33]. The PCR reactions were performed using the Phusion ® High-Fidelity PCR Master Mix with GC Buffer from New England Biolabs [34,35]. The PCR amplification was carried out with efficient and high-fidelity enzymes to ensure amplification efficiency and accuracy. The resulting PCR products were checked using 2% agarose gel electrophoresis (at a voltage of 120 V for 30 min) and then sent to Novogene Biotech Co., Ltd. (Beijing, China) for paired-end sequencing (2 × 250) on the Illumina HiSeq 2500 platform.

Bioinformatics Analysis
Microbial community analysis was performed using the QIIME (Quantitative Insights Into Microbial Ecology) version 1.7.0 [36]. A total of 142,378 high-quality bacterial sequences and 2,021,972 high-quality fungal sequences, with average lengths of 1447 bp and 241 bp, respectively, were recovered after quality filtering, denoising, merging, and chimeric filtering. Sequences relevant to chloroplasts and mitochondria were eliminated since bacterial 16S rDNA primers can also target those organelles' DNA. Additionally, OTUs were grouped using UPARSE (version 7.1) with a 97% similarity criterion. The UNITE database (version 18.11) and USEARCH database (version 8.0) consensus taxonomy classifier against the SILVA 16S rRNA database (version 1.3.2) were used to classify representative sequences of each fungal and bacterial OUT [37,38]. The acquired sequences of the root zone bacteria and the rhizosphere soil fungus were then uploaded to the NCBI Sequence Read Archive (SRA) with the accession codes PRJNA858954 and PRJNA863757, respectively.

Isolation of Cultivable Bacterial and Fungal Strains from Rhizosphere Soil
Considering the notable disparity in the relative abundance of Fusarium between the rhizosphere soil of healthy and diseased plants, we extracted potential antagonistic microorganisms from the rhizosphere soil of healthy Z. armatum plants. The primary process involved the following steps: We transferred 10 g of rhizosphere soil into a 250 mL conical flask, supplemented with 90 mL of sterile water. The soil sample solution was further diluted to 10 −1 , 10 −2 , 10 −3 , 10 −4 , 10 −5 , 10 −6 , and 10 −7 using sterile water. Subsequently, 100 µL of the solution from the 10 −4 , 10 −5 , 10 −6 , and 10 −7 dilutions was evenly spread on nutrient agar (NA) plates for bacteria (n = 3 plates) and potato dextrose Agar (PDA) plates for fungi (n = 3 plates). The NA plates were incubated at 37 • C in a bacterial incubator for 24 h, while the PDA plates were incubated at 25 • C in a constant temperature incubator for 5 days. For purification, single colonies with observable morphological differences were chosen and subcultured on NA or PDA medium. Separately inoculated on NA and PDA medium, the purified bacterial and fungal strains were then placed in storage at 4 • C for later use [39].

Screening and Determination of Antagonistic Effects of Antagonistic Bacteria
Modifying the research methodology based on the study conducted by Lin et al., we employed the plate confrontation method to investigate the antagonistic effects of several microbial strains on the growth and development of the pathogenic fungus F. solani. Single colonies exhibiting antagonistic activity were selected and inoculated into 50 mL/250 mL of LB broth, followed by continuous shaking (180 rpm) for 24 h at 28 • C. The screening of antagonistic bacteria involved placing the pathogenic fungus (diameter d = 8 mm) on one side of a PDA plate and making a sterile hole with a diameter of 6 mm, 3 cm away from the fungus on the other side of the plate. Then, 100 µL of bacterial culture medium was injected into the sterile hole. The control plate contained only the pathogenic fungus and sterile water, with sterile water being injected into the sterile hole on the PDA plate. The pathogenic fungus (diameter d = 8 mm) was put on one side of a PDA plate for the purpose of screening antagonistic fungi, and a sterile 8 mm hole was constructed 3 cm distant from the fungus. Then, 2 µL of bacterial fermentation broth was dropped into the sterile hole. In accordance with the method outlined by Ritpitakphong et al. [40] the sizes of the lesions caused by the pathogenic fungus were assessed following a 7-day cultivation period at 28 • C. Each antagonistic strain was subjected to three repetitions to ensure accuracy and reliability of the measurements. In potted plants, the antagonistic bacteria's capacity to prevent Z. armatum plant root rot was assessed. Each planting bag included 500 g of the substrate (soil:vermiculite:peat = 2:1:2, V/V) before being sterilized under high pressure. Each bag was then sealed for 7 days of cultivation after being injected with 10 mL of hostile bacterial fermentation broth. With five replicates for each treatment, sterile water injected into the soil medium at an identical volume served as the control. A sterile syringe needle was used to puncture the roots of healthy one-year-old Z. armatum plants after they had been cleaned with sterile water for 30 min, dried, and punctured. The wounded plants were transplanted into the aforementioned substrate and grown continuously for 7 days before being inoculated with a 10 mL suspension of F. solani spores (1 × 10 7 spores/mL). All plants were cultured in a greenhouse with a photoperiod of 16 h light and 8 h dark, a humidity of 60%, and a temperature of 25 • C. The occurrence rate of root rot and the mortality rate of plants were observed and assessed every 5 days for up to 20 days, and the disease index [41] of the plants was calculated.

Identification of Purified Microbial Strains with Potential Antagonistic Activity
The DNA of antagonistic bacterial strains was extracted using the Tiangen Bacterial Genome Kit. The extracted genomic DNA of the tested bacterial strains was used as a template for amplifying the 16S rDNA, gyrA, and gyrB genes of the strains using universal bacterial primers [42,43]. After gel electrophoresis to confirm the PCR products, the PCR products of the antagonistic bacterial strains were sent to a sequencing company for sequencing. The obtained sequences were subjected to BLAST analysis in the NCBI GenBank (http://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 29 August 2022), and highly similar and relevant gene sequences were downloaded. Based on the BLAST results, the MEGA software (version 5.0) was used to manually align and remove unnecessary sequences as needed. The IQTREE software was then used to construct a phylogenetic tree using multiple genes. The phylogenetic analysis of the gene evolution was performed to determine the phylogenetic status of the antagonistic bacteria. The nucleotide sequences were stored in the GenBank database, and the accession numbers were obtained ( Table 1). As no antagonistic fungal strains were obtained in this experiment, no identification of antifungal strains was performed.

Statistical Analysis
The R environment (version: V3.6.0, http://www.r-project.org/, accessed on 1 May 2022) was used to conduct all statistical analyses. The "vegan" package in R was used to investigate alpha diversity indices such as the Shannon, Simpson, and Chao1 indices [42]. ANOVA was used to investigate how distinct samples' alpha diversity indices varied from one another. The Kruskal-Wallis test was used to assess differences between the abundance and enrichment levels of bacterial and fungal taxa, and the "ggplot2" package in R was used for display [39]. A generalized linear model (GLM) technique was employed in conjunction with the EdgeR program to identify operational taxonomic units (OTUs) that have differences in abundance [44].
Using principal coordinate analysis (PCoA) based on Bray-Curtis dissimilarity, the beta diversity of rhizosphere soil microbial communities was assessed. The beta diversity of soil microbial communities as determined by PCoA was used to analyze the size and importance of the impacts of season and illness [32]. To evaluate the association between disease severity markers (% of Fusarium lesion area and relative abundance) and the compositional distribution of microbial communities, redundancy analysis (RDA) was used. Using FUNGuild and PICRUSt2, functional prediction of bacterial and fungal OTUs was carried out [38,45,46].

Isolation and Identification of the Pathogen Causing Root Rot in Z. armatum
Eight strains were obtained from diseased roots of Z. armatum, and pathogenicity testing identified a highly pathogenic strain, HJGFB-1. The morphological characteristics of this strain on PDA medium included smooth, branched hyphae with septa, transparent appearance, and light brown color. The conidia produced on conidiophores were simple to branches. The conidia consisted of one or two cells and were elliptical or kidney-shaped, varying in size from 3.4 to 5.6 × 7.5 to 12.3 µm (n = 50 conidia). The macroconidia were crescent-shaped with 3-4 septa and a characteristic foot-shaped basal cell, ranging in size from 3.3 to 7.9 × 21.2 to 47.1 µm (n = 50 conidia). The chlamydospores produced intercalarily in the hyphae had thick walls and were produced in chains, measuring 8.5 to 11 × 7.6 to 9 µm ( Figure 2).
Eight strains were obtained from diseased roots of Z. armatum, and pathogenicity testing identified a highly pathogenic strain, HJGFB-1. The morphological characteristics of this strain on PDA medium included smooth, branched hyphae with septa, transparent appearance, and light brown color. The conidia produced on conidiophores were simple to branches. The conidia consisted of one or two cells and were elliptical or kidneyshaped, varying in size from 3.4 to 5.6 × 7.5 to 12.3 µm (n = 50 conidia). The macroconidia were crescent-shaped with 3-4 septa and a characteristic foot-shaped basal cell, ranging in size from 3.3 to 7.9 × 21.2 to 47.1 µm (n = 50 conidia). The chlamydospores produced intercalarily in the hyphae had thick walls and were produced in chains, measuring 8.5 to 11 × 7.6 to 9 µm ( Figure 2). DNA sequencing and BLAST analysis of the ITS, RPB2, TUB, and TEF1 gene sequences of this strain showed 99% identity with F. solani sequences (GenBank accession numbers KU321553, MK968890, MH888086, and MN652888). A phylogenetic tree was inferred from a combined dataset (ITS, TEF1, TUB, and PRB2 genes) of Fusarium complex members analyzed in this study. The phylogenetic tree demonstrated 100% branch support, indicating a match between the isolated strain and F. solani ( Figure S1). Based on the morphological features and multi-gene phylogenetic analysis, the isolate was identified as F. solani. The root systems of the inoculated plants exhibited brown discoloration and softening of lateral roots, while the roots of the control plants remained healthy with welldeveloped lateral root systems. The experiment was repeated three times.
In summary, F. solani was successfully re-isolated from diseased tissues, and its identity was confirmed based on morphological and sequencing analysis, thus fulfilling Koch's postulates. DNA sequencing and BLAST analysis of the ITS, RPB2, TUB, and TEF1 gene sequences of this strain showed 99% identity with F. solani sequences (GenBank accession numbers KU321553, MK968890, MH888086, and MN652888). A phylogenetic tree was inferred from a combined dataset (ITS, TEF1, TUB, and PRB2 genes) of Fusarium complex members analyzed in this study. The phylogenetic tree demonstrated 100% branch support, indicating a match between the isolated strain and F. solani ( Figure S1). Based on the morphological features and multi-gene phylogenetic analysis, the isolate was identified as F. solani. The root systems of the inoculated plants exhibited brown discoloration and softening of lateral roots, while the roots of the control plants remained healthy with well-developed lateral root systems. The experiment was repeated three times. In summary, F. solani was successfully re-isolated from diseased tissues, and its identity was confirmed based on morphological and sequencing analysis, thus fulfilling Koch's postulates.

Impact of Fusarium Abundance on the Overall Taxonomic Distribution of Rhizosphere Soil Microbial Communities
The relative abundance of Fusarium in the rhizosphere soil of diseased plants was significantly higher compared to that in the rhizosphere soil of healthy plants ( Figure 3). Therefore, we analyzed the overall taxonomic classification and distribution of rhizosphere soil microbial communities in response to the increased abundance of Fusarium. The results showed that a total of 3689 bacterial operational taxonomic units (OUTs) and 1771 fungal OUTs were obtained from the rhizosphere microbiota of healthy and diseased Z. armatum plants, with lower fungal diversity compared to bacterial diversity. The fungal community OTUs were associated with 13 phyla, 32 classes, 65 orders, 127 families, 226 genera, and 293 species, while the bacterial community OTUs were associated with 36 phyla, 48

Impact of Fusarium Abundance on the Overall Taxonomic Distribution of Rhizosphere Soil Microbial Communities
The relative abundance of Fusarium in the rhizosphere soil of diseased plants was significantly higher compared to that in the rhizosphere soil of healthy plants ( Figure 3). Therefore, we analyzed the overall taxonomic classification and distribution of rhizosphere soil microbial communities in response to the increased abundance of Fusarium. The results showed that a total of 3689 bacterial operational taxonomic units (OUTs) and 1771 fungal OUTs were obtained from the rhizosphere microbiota of healthy and diseased Z. armatum plants, with lower fungal diversity compared to bacterial diversity. The fungal community OTUs were associated with 13 phyla, 32 classes, 65 orders, 127 families, 226 genera, and 293 species, while the bacterial community OTUs were associated with 36 phyla, 48

Study on the Diversity of Rhizosphere Soil Microbial Communities
The α-diversity of fungal and bacterial communities in the rhizosphere soil of healthy and diseased Z. armatum plants was assessed using Shannon, Chao1, Simpson, and goodcoverage indices. The fungal Shannon, Chao1, and Simpson indices showed significant differences in diversity between diseased and healthy rhizosphere soil (Wilcoxon test, P< 0.05). The results indicated that the presence of diseased plants increased the diversity of rhizosphere soil microbial communities (Figure 4a,b). Regarding the α-diversity of

Study on the Diversity of Rhizosphere Soil Microbial Communities
The α-diversity of fungal and bacterial communities in the rhizosphere soil of healthy and diseased Z. armatum plants was assessed using Shannon, Chao1, Simpson, and goodcoverage indices. The fungal Shannon, Chao1, and Simpson indices showed significant differences in diversity between diseased and healthy rhizosphere soil (Wilcoxon test, p < 0.05). The results indicated that the presence of diseased plants increased the diversity of rhizosphere soil microbial communities (Figure 4a,b). Regarding the α-diversity of bacterial communities, there was also a significant difference in microbial diversity between the rhizosphere soil of diseased and healthy plants (Wilcoxon test, p < 0.05), with the diseased plants exhibiting a 6.41% increase on the Shannon index and a 47.69% increase on the Chao1 index compared to healthy plants (Figure 4e,f). crease on the Chao1 index compared to healthy plants (Figure 4e,f).
Principal coordinate analysis (PCoA) was performed to assess the β-diversity of soil microbial communities, and distinct separation was observed between fungal and bacterial communities derived from the rhizosphere soil of healthy and diseased plants. According to the PCoA results, the first two axes explained 54.48% and 6.01% of the total variance in the bacterial community and 52.82% and 12.35% of the total variation in the fungal community, respectively (Figure 4i,j). Non-metric multidimensional scaling (NMDS) analysis of pairwise distances variance between bacteria (R 2 = 0.5419, p < 0.001) and fungi (R 2 = 0.5083, p < 0.001) showed substantial differences in the composition of rhizosphere microbial communities among different health conditions. This indicates that root rot significantly influences the species and composition of bacterial and fungal populations.  Principal coordinate analysis (PCoA) was performed to assess the β-diversity of soil microbial communities, and distinct separation was observed between fungal and bacterial communities derived from the rhizosphere soil of healthy and diseased plants. According to the PCoA results, the first two axes explained 54.48% and 6.01% of the total variance in the bacterial community and 52.82% and 12.35% of the total variation in the fungal community, respectively (Figure 4i,j). Non-metric multidimensional scaling (NMDS) analysis of pairwise distances variance between bacteria (R 2 = 0.5419, p < 0.001) and fungi (R 2 = 0.5083, p < 0.001) showed substantial differences in the composition of rhizosphere microbial communities among different health conditions. This indicates that root rot significantly influences the species and composition of bacterial and fungal populations.

Analysis of Rhizosphere Soil Microbial Community Abundance in Z. armatum
The relative abundance and composition of fungal and bacterial communities in the rhizosphere soil of healthy and diseased Z. armatum plants were differentiated based on their most common categories. The richest fungal categories in the rhizosphere soil fungal community were Sordariomycetes (45%), Agaricomycetes (18%), and Eurotiomycetes (12%). Among them, the relative abundance of the fungal class in the rhizosphere soil of diseased plants was 6.88% higher than that in the rhizosphere soil of healthy plants. On the other hand, the most abundant bacterial categories in both healthy and diseased Z. armatum rhizosphere soil were Gammaproteobacteria (24%), Acidobacteriia (22%), Alphaproteobacteria (14%), Bacteroidia (9%), and Deltaproteobacteria (6%) (Figure 5a,b). each box plot indicate significant differences among treatments (Wilcoxon test, p < 0.05). Principal coordinate analysis (PCoA) based on the Bray-Curtis distance matrix demonstrates the separation between soil bacterial and fungal communities under different treatments. PCoA for bacterial (i) and fungal (j) communities.

Analysis of Rhizosphere Soil Microbial Community Abundance in Z. armatum
The relative abundance and composition of fungal and bacterial communities in the rhizosphere soil of healthy and diseased Z. armatum plants were differentiated based on their most common categories. The richest fungal categories in the rhizosphere soil fungal community were Sordariomycetes (45%), Agaricomycetes (18%), and Eurotiomycetes (12%). Among them, the relative abundance of the fungal class in the rhizosphere soil of diseased plants was 6.88% higher than that in the rhizosphere soil of healthy plants. On the other hand, the most abundant bacterial categories in both healthy and diseased Z. armatum rhizosphere soil were Gammaproteobacteria (24%), Acidobacteriia (22%), Alphaproteobacteria (14%), Bacteroidia (9%), and Deltaproteobacteria (6%) (Figure 5a,b).
A clustering heatmap was used to examine the differences in the top 30 abundant fungal and bacterial genera in the rhizosphere soil of healthy and diseased Z. armatum plants (Figure 5c,d)  A clustering heatmap was used to examine the differences in the top 30 abundant fungal and bacterial genera in the rhizosphere soil of healthy and diseased Z. armatum plants (Figure 5c,d). In the rhizosphere soil of healthy Z. armatum plants, the following fungal genera were more abundant: Pyrenochaetopsis, Chloridium, Cylindrocarpon, Dictyochaeta, Talaromyces, Thozetella, Kurtzmaniella, Penicllifera. The rhizosphere soil of diseased Z. armatum plants had higher abundance of the fungal genera Trichoderma, Fusarium, Exophiala, Colletotrichum, Arthrobotrys, Metarhizium, Phaeosphaeria, Pseudeurotium, Zopfiella, Apiosordaria, Neosetophoma, Coprinellus, Rhodotorula, Pilidium, and Boothiomyces.
The clustering heatmap analysis of bacterial genera in the rhizosphere soil of Z. armatum plants showed that healthy rhizosphere soil had a higher relative abundance of genera such as Chujaibacter, Sinomonas, Proteus, Tumebacillus, Acidibacter, Granulicella, Dyella, Candi-datus_Udaeobacter, Acidipila, and Bacillus. On the other hand, the rhizosphere soil of diseased plants had a higher relative abundance of genera such as Telmatospirillum, Flavobacterium, Niveispirillum, Hydrogenispora, Geobacter, Sideroxydans, Paludibacter, and Sulfuricurvum.

Relationship between Soil Physicochemical Properties and Microbial Community Composition
The relationship between soil physicochemical properties and microbial community composition was investigated. The main physicochemical characteristics of diseased and healthy rhizospheric soils were determined according to the experimental design. The Mantel test indicated significant variations (p < 0.05) in soil physicochemical properties, including pH, SOM, TP, TK, and TN, between different health conditions. RDA was used to test and visualize the associations between soil properties and microbial community composition (Figure 6a,b). The results showed that the first two axes explained more than 90% of the fungal and bacterial community structure, with RDA1 explaining the majority of the variation. Based on the length of the vectors, pH, SOM, TP, TK, and TN were identified as the main factors influencing the differences in fungal and bacterial communities. Furthermore, a Spearman correlation heatmap revealed significant positive correlations between TN, SOM, pH, and fungi such as Fusarium, Pseudeurotium, Boothiomyces, Trichoderma, Exophiala, Neosetophoma, Apiosordaria, and Zopfiella, while negative correlations were observed with TK and TP. The soil physicochemical properties TK and TP exhibited positive correlations with bacterial genera such as Acidibacter, Dyella, Proteus, Chujaibacter, Occallatibacter, Acidipila, Granulicella, Bacillus, and Bryobacter (Figure 6c,d).

Functional Prediction of Fungal and Bacterial Taxa
Functional predictions of fungal communities in diseased and healthy plant rhizospheric soils were analyzed using FUNGuild (Figure 7a). The results showed significant differences (p < 0.05) in fungal functional profiles between diseased and healthy rhizospheric soils, including animal pathogen, soil saprotroph, animal endosymbiont-undefined saprotroph, dung saprotroph-undefined saprotroph, and dung saprotroph. Additionally, functional predictions of bacterial communities in diseased and healthy rhizospheric soils were performed based on the KEGG database (https://www.kegg.jp/, accessed on 25 August 2022) using PICRUSt2 (Figure 7b). The results revealed 30 secondary functional categories in the bacterial communities, including sorting and degradation, aging, amino acid metabolism, biosynthesis of other secondary metabolites, cancer: overview, cancer: specific types, carbohydrate metabolism, cell growth and death, cell motility, cellular community-eukaryotes, cellular community-prokaryotes, circulatory system, cevelopment and regeneration, digestive system, drug resistance: antimicrobial, drug resistance: antineoplastic, endocrine and metabolic disease, endocrine system, energy metabolism, environmental adaptation, excretory system, folding, glycan biosynthesis and metabolism, immune disease, immune system, infectious disease: bacterial, infectious disease: parasitic, infectious disease: viral, and lipid metabolism. Among these, there were significant differences (p < 0.05) in the cellular community-eukaryotes function between bacterial communities in diseased and healthy plant rhizospheric soils.

Functional Prediction of Fungal and Bacterial Taxa
Functional predictions of fungal communities in diseased and healthy plant rhizospheric soils were analyzed using FUNGuild (Figure 7a). The results showed significant differences (p < 0.05) in fungal functional profiles between diseased and healthy rhizospheric soils, including animal pathogen, soil saprotroph, animal endosymbiont-undefined saprotroph, dung saprotroph-undefined saprotroph, and dung saprotroph. Additionally, functional predictions of bacterial communities in diseased and healthy rhizospheric soils were performed based on the KEGG database (https://www.kegg.jp/, accessed on 25 August 2022) using PICRUSt2 (Figure 7b). The results revealed 30 secondary functional categories in the bacterial communities, including sorting and degradation, aging, amino acid metabolism, biosynthesis of other secondary metabolites, cancer: overview, cancer: specific types, carbohydrate metabolism, cell growth and death, cell motility, cellular community-eukaryotes, cellular community-prokaryotes, circulatory system, cevelopment and regeneration, digestive system, drug resistance: antimicrobial, drug resistance: antineoplastic, endocrine and metabolic disease, endocrine system, energy metabolism, environmental adaptation, excretory system, folding, glycan biosynthesis and

Screening of Antagonistic Bacteria and Determination of Their Antagonistic Effects
A total of 48 bacterial strains were isolated and purified from healthy rhizospheric soil. The antagonistic activity of these strains was screened using the plate confrontation method, and three bacterial strains showed significant inhibitory effects against the pathogenic fungus F. solani. Specifically, the lesion size of F. solani grown on the culture medium mixed with the bacterial strains (T1, T2, T3) was compared to the lesion size on the control plate. It was observed that T1, T2, and T3 exhibited an average reduction of 52.94%, 46.06%, and 48.23%, respectively (Figure 8a). disease: parasitic, infectious disease: viral, and lipid metabolism. Among these, there were significant differences (p < 0.05) in the cellular community-eukaryotes function between bacterial communities in diseased and healthy plant rhizospheric soils.

Screening of Antagonistic Bacteria and Determination of Their Antagonistic Effects
A total of 48 bacterial strains were isolated and purified from healthy rhizospheric soil. The antagonistic activity of these strains was screened using the plate confrontation method, and three bacterial strains showed significant inhibitory effects against the pathogenic fungus F. solani. Specifically, the lesion size of F. solani grown on the culture medium mixed with the bacterial strains (T1, T2, T3) was compared to the lesion size on the control plate. It was observed that T1, T2, and T3 exhibited an average reduction of 52.94%, 46.06%, and 48.23%, respectively (Figure 8a). Moreover, the three biocontrol bacterial strains were evaluated on Z. armatum plants that were infected with F. solani. The outcomes demonstrated a progressive increase in the disease index of the positive control plants, which were solely inoculated with F. solani spore suspension. In contrast, the disease index of Z. armatum plants treated with spore suspensions of the antagonistic bacterial strains (T1, T2, T3) initially increased but subsequently declined, in comparison to the positive control plants. These findings provide evidence of the potent inhibitory effects exerted by the three antagonistic bacterial strains on the growth and pathogenicity of F. solani, both in vivo and within vitro (Figure 8b). that were infected with F. solani. The outcomes demonstrated a progressive increase in the disease index of the positive control plants, which were solely inoculated with F. solani spore suspension. In contrast, the disease index of Z. armatum plants treated with spore suspensions of the antagonistic bacterial strains (T1, T2, T3) initially increased but subsequently declined, in comparison to the positive control plants. These findings provide evidence of the potent inhibitory effects exerted by the three antagonistic bacterial strains on the growth and pathogenicity of F. solani, both in vivo and within vitro (Figure 8b).

Identification of Antagonistic Bacteria
The morphological characteristics of the three antagonistic bacterial strains are shown in Figure 9. Strain T1 has a wrinkled, irregular, white colony surface and exhibits chain-shaped rod cells. Strain T2 appears as a light yellow, opaque colony with a rough surface, protrusions, and irregular edges on LB agar medium. Strain T3 has a flat, smooth, moist, opaque colony surface with a light-yellow color. Gram staining of all three bacterial strains showed positive results.
Based on bacterial molecular biology identification, sequencing of the 16S rRNA, gyrA, and gyrB genes was performed. The obtained sequences were subjected to BLAST nucleotide sequence alignment in the NCBI database. The results confirmed that all three biocontrol strains belong to the genus Bacillus. The alignment of 16S rRNA, gyrA, and gyrB sequences showed that T1 exhibited 100% homology with Bacillus subtilis, T2 exhibited 99% homology with Bacillus amyloliquefaciens, and T3 exhibited 99% homology with Bacillus siamensis (Figures 10 and 11).

Identification of Antagonistic Bacteria
The morphological characteristics of the three antagonistic bacterial strains are shown in Figure 9. Strain T1 has a wrinkled, irregular, white colony surface and exhibits chainshaped rod cells. Strain T2 appears as a light yellow, opaque colony with a rough surface, protrusions, and irregular edges on LB agar medium. Strain T3 has a flat, smooth, moist, opaque colony surface with a light-yellow color. Gram staining of all three bacterial strains showed positive results.
Based on bacterial molecular biology identification, sequencing of the 16S rRNA, gyrA, and gyrB genes was performed. The obtained sequences were subjected to BLAST nucleotide sequence alignment in the NCBI database. The results confirmed that all three biocontrol strains belong to the genus Bacillus. The alignment of 16S rRNA, gyrA, and gyrB sequences showed that T1 exhibited 100% homology with Bacillus subtilis, T2 exhibited 99% homology with Bacillus amyloliquefaciens, and T3 exhibited 99% homology with Bacillus siamensis (Figures 10 and 11).
Maximum likelihood and Bayesian posterior probability analyses were conducted using IQtree v.1.6.8 and MrBayes v.3.2.6 to infer the phylogenetic relationships of T1, T2, and T3 with related taxonomic groups based on concatenated sequences of the 16S rRNA and gyrA genes. The nodes in the phylogenetic tree represent the RAxML bootstrap support values (ML ≥ 50) and Bayesian posterior probabilities (PP ≥ 90) (ML/PP).   Bootstrap support values greater than 0.5 are indicated above/below the nodes. The ex-type (ex-epitype) and voucher stains are in red font. Confirmed that T1 is Bacillus subtilis and T2 is Bacillus amyloliquefaciens.   Bootstrap support values greater than 0.5 are indicated above/below the nodes. The ex-type (ex-epitype) and voucher stains are in red font. Confirmed that T1 is Bacillus subtilis and T2 is Bacillus amyloliquefaciens. Bootstrap support values greater than 0.5 are indicated above/below the nodes. The ex-type (ex-epitype) and voucher stains are in red font. Confirmed that T3 is B. siamensis.

Discussion
In this study, we isolated and identified the pathogenic fungus causing root rot in the Z. armatum plant. Based on morphological characteristics and multi-gene phylogenetic analysis, we confirmed that the pathogen responsible for Z. armatum root rot disease is F. solani (Figures 1 and S1) [47]. Soil microbial characteristics can to some extent reflect soil quality and health. When soil-borne diseases occur, there are often significant differences in microbial diversity and intergroup dynamics between healthy and diseased plant rhizosphere soils [48]. Higher microbial community diversity indices indicate more complex community structures and greater stability, thus enhancing the ability to cope with environmental changes and pathogen invasion [49,50]. In this study, we found that the infection of Z. armatum plant roots with root rot led to an increase in the diversity of fungi and bacteria in the rhizosphere soil. The Chao 1 and Shannon indices of the rhizosphere soil of diseased plants were significantly higher than those of healthy plants. This result is inconsistent with the changes in microbial diversity observed in common soil-borne diseases, but a similar phenomenon has been found in the rhizosphere soil of Panax notoginseng infected with root rot [11]. The differential response may be due to changes in the rhizosphere soil ecological conditions of Z. armatum plants caused by the occurrence of root rot disease, promoting the growth and reproduction of certain microorganisms and thereby increasing the species diversity of rhizosphere soil microorganisms. In this study, we compared the relative abundance of Fusarium in the rhizosphere soil of healthy and diseased Z. armatum plants during the peak period of root rot. The results showed that the amount of Fusarium was significantly higher in the rhizosphere soil of diseased plants compared to healthy plants. We also found significant differences in soil physicochemical properties between healthy and diseased rhizosphere soils ( Figure 6). Soil physicochemical properties can indirectly affect the composition and relative abundance of rhizosphere microorganisms by influencing plant physiology. Our study found that the content of SOM, pH, and TK in the rhizosphere soil of diseased Z. armatum plants was higher than that in healthy plants, which may be attributed to the high content of humus in the rhizosphere soil caused by root rot infection. There is increasing evidence that the structure of rhizosphere-associated soil microbial communities differs greatly between healthy and Figure 11. Phylogram generated from Bayesian inference analysis based on combined sequences of 16S and gyrA from 24 recognized Bacillus species. Bootstrap support values greater than 0.5 are indicated above/below the nodes. The ex-type (ex-epitype) and voucher stains are in red font. Confirmed that T3 is B. siamensis.

Discussion
In this study, we isolated and identified the pathogenic fungus causing root rot in the Z. armatum plant. Based on morphological characteristics and multi-gene phylogenetic analysis, we confirmed that the pathogen responsible for Z. armatum root rot disease is F. solani (Figure 1 and Figure S1) [47]. Soil microbial characteristics can to some extent reflect soil quality and health. When soil-borne diseases occur, there are often significant differences in microbial diversity and intergroup dynamics between healthy and diseased plant rhizosphere soils [48]. Higher microbial community diversity indices indicate more complex community structures and greater stability, thus enhancing the ability to cope with environmental changes and pathogen invasion [49,50]. In this study, we found that the infection of Z. armatum plant roots with root rot led to an increase in the diversity of fungi and bacteria in the rhizosphere soil. The Chao 1 and Shannon indices of the rhizosphere soil of diseased plants were significantly higher than those of healthy plants. This result is inconsistent with the changes in microbial diversity observed in common soilborne diseases, but a similar phenomenon has been found in the rhizosphere soil of Panax notoginseng infected with root rot [11]. The differential response may be due to changes in the rhizosphere soil ecological conditions of Z. armatum plants caused by the occurrence of root rot disease, promoting the growth and reproduction of certain microorganisms and thereby increasing the species diversity of rhizosphere soil microorganisms. In this study, we compared the relative abundance of Fusarium in the rhizosphere soil of healthy and diseased Z. armatum plants during the peak period of root rot. The results showed that the amount of Fusarium was significantly higher in the rhizosphere soil of diseased plants compared to healthy plants. We also found significant differences in soil physicochemical properties between healthy and diseased rhizosphere soils ( Figure 6). Soil physicochemical properties can indirectly affect the composition and relative abundance of rhizosphere microorganisms by influencing plant physiology. Our study found that the content of SOM, pH, and TK in the rhizosphere soil of diseased Z. armatum plants was higher than that in healthy plants, which may be attributed to the high content of humus in the rhizosphere soil caused by root rot infection. There is increasing evidence that the structure of rhizosphereassociated soil microbial communities differs greatly between healthy and diseased plants, indicating that plants can recruit beneficial microorganisms in their rhizosphere to suppress soil-borne pathogens.
Therefore, we further analyzed the compositional differences of bacterial and fungal taxa in the rhizosphere of healthy and diseased Z. armatum plants and found that the rhizosphere soil of Z. armatum plants harbored a rich abundance of ecologically beneficial bacterial genera, with the Bacillus sp. being higher in healthy plant rhizosphere microbiota compared to diseased plant rhizosphere microbiota. Currently, Bacillus-based biocontrol agents have been widely developed and applied in various fields such as agriculture, forestry, livestock, and fisheries [51]. Previous studies have demonstrated that several species within the genera Bacillus possess strong biocontrol capabilities by producing a wide range of antagonistic compounds. Additionally, Bacillus spp. can assist plants in acquiring nutrients such as nitrogen and phosphorus, promote plant hormone synthesis, enhance the host plant's nutrient utilization efficiency, and improve its tolerance to biotic and abiotic stresses. Bacillus-based biocontrol agents have been applied in the management of Fusarium crop diseases and agricultural environmental pollution [52,53]. For example, B. velezensis has been used to control root rot disease in Z. armatum [54]; B. amyloliquefaciens has shown biocontrol activity against Xanthomonas sp. and the potential as a biocontrol agent for bacterial diseases in rice [55]. Additionally, research has shown that Bacillus spp. has the ability to suppress the growth of Colletotrichum gloeosporioides and F. oxysporum [5].
We isolated cultivatable bacterial strains from the soil near healthy plants' rhizomes and tested them in vivo and in vitro for their ability to suppress the pathogenic fungus F. solani. This further confirms the suppressive impact of rhizosphere microbial communities on soil-borne pathogens. Results showed that three bacterial strains, B. subtilis, B. amyloliquefaciens, and B. siamensis, isolated from soil in the rhizosphere had strong inhibitory effects on F. solani. In addition, we found that healthy plants had a greater variety of Bacillus spp. in their rhizosphere soil than did sick plants. In conclusion, we have shown that healthy Z. armatum plants are able to prevent infection with the soil-borne pathogen F. solani by recruiting hostile bacteria in the soil of the rhizosphere.

Conclusions
This research revealed that F. solani is the main culprit responsible for Z. armatum root rot, and that plants infected with F. solani demonstrated a high Fusarium abundance in their rhizosphere soil. Furthermore, microbial community research revealed that healthy Z. armatum plants possess a greater quantity of beneficial microbial groups in the rhizosphere soil, suggesting that these microbes may function as biocontrol agents. The rhizosphere soil of healthy plants was sampled, and three bacterial strains were identified as potential biocontrol agents. Both in vitro and in vivo testing demonstrated their antifungal activity against F. solani. This study greatly expands our understanding of the structural and functional differences between rhizosphere soil microbial communities that contribute to the spread of soil-borne diseases. Moreover, it offers possible biocontrol strains that might boost plant resilience to root rot. It's worth noting that these biocontrol strains showed signs of boosting plant development as well.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/f14081561/s1, Figure S1: Construction of Phylogenetic Tree by Combining Multiple Genes of HJGFB-1 strain. Phylogenetic tree inferred from the combined datasets (ITS, TEF, TUB, and RBP2) from members of the F. solani species analyzed. The evolutionary history was inferred by using the Bayesian Inference (BI) analyses, using Fusarium plagianthi NRRL22632 and Fusarium vinguliforme NRRL 32392 as the outgroup taxon.