R. solanacearum and B. amyloliquefaciens ZM9 are preserved by our laboratory, among which B. amyloliquefaciens ZM9 is an antagonistic bacteria of R. solanacearum.

Tobacco were transplanted to peat soils, and then 15 days later, 1 ml ofR. solanacearum (T1), B. amyloliquefaciens ZM9 (T2) (1.0 × 109 CUF/mL) and ddH₂O (CK). The tobacco root exudates were collected according to the previous methods [22], and then he collected root exudates were extracted with ethyl acetate. The acidic, basic and neutral ethyl acetate extracts were concentrated under reduced pressure to 1 mL (45˚C), filtered with a 0.45 μm filter membrane, and the organic The phases are combined into a rotary evaporator, and the product is rotary evaporated to dryness at 50˚C, and the wall is washed with 1 mL ultrapure water, and stored at low temperature (−80˚C) for later use.

Add 40 µL of 20 mg/mL methoxyamine salt pyridine and 10 µL of 5% phenethyl acetate mixture to the freeze-dried root exudates (Phenyl acetate as internal standard), and after incubation in a 60˚C incubator for 60 min, Then add 50 µL of silylation reagent N,O-bis(trimethylsiyl)acetamide (BSTFA), keep it at 70˚C for 60 min, and cool to room temperature for later use.

Add 40 µL of 20 mg/mL methoxyamine salt pyridine and 10 µL of 5% phenethyl acetate/pyridine mixture (Phenethyl acetate as internal standard) to the freeze-dried root exudates sample, and place it at a constant temperature of 60˚C Keep the temperature in the box for 60 minutes. Then add 50 µL of silanization reagent N,O-bis(trimethylsilyl)acetamide (BSTFA), keep it at 70˚C for 60 min, cool naturally to room temperature, GC-MS detection adopts the American Thermo Company gas chromatography/mass spectrometer (DSQII) for analysis. GC conditions: The chromatographic column is TG-5MS 30 m × 0.25 mm × 0.25 μm, the inlet temperature is 250˚C, and the split ratio is 50:1. Column temperature rising program: increase from 50˚C to 250˚C at a rate of 3˚C/min, continue to increase to 300˚C at a rate of 25˚C/min, keep for 4 minutes; carrier gas is He (99.9%), flow rate is 20 psi, the injection volume is 1 μL. MS conditions: EI source, ion energy of 70 eV, ion source temperature of 300˚C, scanning range of 30 - 600 m/z, transmission line temperature of 280˚C.

Soil DNA was extracted from 0.5 g rhizosphere soil using the FastDNA Spin Kit (MP Biomedicals, USA) following the manufacture’s protocol. The integrity of DNA samples were determinde by 1% agarose gel electrophoresis. Then the concentration and purity of the DNA were determined using a Nanodrop ND-1000 Spectrophotmeter (Nanodrop Technologies, Wilmington, DE, USA). The extracted soil DNA was used as template to amplify 16S rRNA gene. The V4 region of 16S rRNA gene were amplified with the primers 515F (5’-GTGCCAGCMGCCGCGGTAA-3’) and 806R (5’-GGACTACHVGGGTWTCTAAT-3’)5. All PCR reactions were performed on Illumina HiSeq platforms (Illumina Inc., USA) at Novogene Bioinformatics Technology Co., Ltd. (Beijing, China). The library quality was assessed on the Qubit@ 2.0 Fluorometer (Termo Scientifc) and Agilent Bioanalyzer 2100 system. Sequences analysis was performed by Uparse software (Version 7.0.1001, http://drive5.com/uparse/). Sequences with ≥97% similarity were assigned to the same OTUs. Each representative sequence was screened for further taxonomic information annotation. FASTX Toolkit 0.0.13 software package was used to preliminarily filtrate the raw sequence data with removing the low mass base at the tail of the sequence (Q value less than 20) and the sequences with lengths less than 35 bp. The sequence quality was statistically analyzed by CASAVA1.8, and finally, the length of the valid reads was approximately 250 bp. The operational taxonomic units (OTUs), observed-species (Sobs), Shannon, Simpson and Chao1 were calculated to evaluate richness and diversity of soil microbial community.

The data were analyzed with Microsoft Excel 2007 and SPSS version 18.0 (IBM, USA). Differences between treatments were assessed by one-way analysis of variance (ANOVA) and least significant difference (LSD) test (p < 0.05). Correlation analysis was conducted by Pearson (2-tailed). Principal components analyzed (PCA) with the weighted UniFrac distance and canonical correspondence analysis (CCA) were carried out using the vegan package in R (Version 2.15.3).

GC-MS analyzed the composition changes of the root exudates of Yunyan 87 under different treatments. As shown in Figure 1, in the GC-MS total ion chromatogram after the bacterial wilt pathogen treatment, we can see 37 peaks, indicating that there are 37 types of root exudates. The composition changes of root exudates under different treatments were compared by GC-MS analysis, and it was found that different microorganisms have certain influence on the types of tobacco root exudates.

GC-MS analyzed the composition changes of the root exudates of Yunyan 87 under different treatments. Excluding the components whose types and contents did not change under the three conditions, under the action of R. solanacearum, acids, alcohols, amines and esters accounted for 39.02%, 4.75%, 29.91%, and 9.10%, respectively. Under the action of its antagonistic B. amyloliquefaciens ZM9, acids, alcohols, amines and esters accounted for 16.58%, 5.35%, 10.23%, and 19.65%, respectively. After removing the components whose types and contents did not change under the three conditions, under the action of R. solanacearum, the types of root exudates of Yunyan 87 increased by 19 components and decreased by 3 components compared with the control group. The increased components mainly include 3 kinds of amines, 6 kinds of alcohols, 8 kinds of acids, 1 kind of sugars, 1 kind of esters; the reduced components are diethylene glycol, phenethyl alcohol and triethylene glycol (Figure 2(A)). After adding B. amyloliquefaciens ZM9, 24 kinds of ingredients are added, 6 of which are newly added; 6 kinds of ingredients are reduced, including 6 kinds of thiodiglycol, 1-butanol, sebacic acid, benzimidate hydrochloride, isophthalic acid, and cyclopropane Completely disappear. Compared with the types of root exudates in the T1 group and the T2 group, there were 2 more types and 6 less components. Principal coordinate analyses (PCoA) were carried out using weighted UniFrac distance in different treatments, and PC1 and PC2 explained 74.96% of root exudates. The CK groups in different periods are clustered together, and the root exudates of the corresponding T1 and T2 groups in different periods are clustered together, and the root exudates under different treatments in the same period are separated from each other. This result shows that the change of root exudates has little relationship with time, and is mainly affected by microorganism (Figure 2(B)). The heat map analysis of the top 53 root exudates with hierarchical clustering is used to discuss the composition changes of root exudates under different treatments. There are differences in the composition of root

exudates between different treatments in the heat map. The treatment of R. solanacearum significantly increased the abundance of 3-hydroxybutyricacid, sucrose, cyclopropane, hydroabietic acid, l-butanol, benzimidate, hydrochloride, 3,4-dimethoxybenzoic, glycerol monostearate, and dodecanoic acid. In contrast, the treatment of B. amyloliquefaciens ZM9 increased terephthalic acid, d-Mannitol, myo-Inositol, 2,3-Butanediol, propyl heptanoate, Silanol, 1-monopalmitin, glycerol, benzoic acid, lactic acid, acetic acid, octanoic acid, urea, d-Pinitol, octanoic acid these abundance. Results indicate that the treatment of different microorganisms has a great influence on the composition of tobacco root exudates (Figure 2(C)). Comprehensive analysis found that stearic acid, propionic acid, oxalic acid, octanoic acid, myristic acid, itaconic acid, isophthalic acid, dodecanoic acid, benzoic acid, 3-hydroxybutyric acid and 3,4-dimethoxybenzoic acid fluctuates the most (Figure 2(D)).

In total 679,451 high-quality raw sequences with the average length of 252 bps for bacteria were obtained from rhizospherial soil samples after quality filtering. The OTUs, Chao1 and Shannon index were used to evaluate and compare the richness and diversity of bacterial community among different treatments (Table 1). It can be seen from the data that the OTUs, Chao1 and Shannon indexes in the rhizosphere soil treated by R. solanacearum were significantly lower than those in the CK and T2 groups. Compared with the CK group, the OTUs and Chao1 indexes of the B. amyloliquefaciens ZM9 treatment were slightly higher. The results indicate that R. solanacearum treatment affects the richness and diversity of soil bacterial communities to a certain extent, and its antagonistic antibacterial treatment can improve the richness and diversity of soil bacterial communities.

In tobacco soil samples of different treatments, Protebacteria accounted for 42% - 51.2% of the top 10 bacteria. The second was Acidoobacteria (18.35% - 23.64%), Actinobacteria (4.41% - 7.89%), Gemmatimonadetes (3.41% - 5.32%), Bacteroidetes (2.21% - 3.11%), Chloroflex (1.41% - 2.51%), Thaumarchaeota (1.21% - 2.01%), Verrucomicrobia (0.51% - 1.71%), Firmicutes and Cyanobacteria (0.71% - 1.41%). The relative abundance of Proteobacteria, Acidobacteria, Actinomycetes and Bacteroidetes was more than 4% (Figure 3(A)). Principal coordinate analysis (PCoA) was performed using weighted UniFrac distance in different treatments. PC1 and PC2 accounted for 66.76% of the total bacterial community. The bacterial communities in different time periods of CK group were clustered together, the corresponding bacterial communities in different time periods of

The observed OTUs, Shannon diversity and Chao1 richness of bacterial communities in the soil are expressed as mean ± SE.CK: ddH₂O treatment, T1: R. solanacearum treatment; T2: B. amyloliquefaciens ZM9 treatment.

T1 and T2 groups were clustered together, and the bacterial communities in different treatment groups were separated from each other (Figure 3(B)). There were differences in bacterial community structure among different treatments in heat map. The abundance of Acidibacter, Bryobacter, Geobacter, Arthrobacter, Flavobacterium, Ralstonia, Aeromonas and Flexibacter increased with the treatment of R. solanacearum; Devosia, Streptomyces, Acinetobacter, Sphingomonas, Bacillus, Pseudomonas, Sphingobiuml and Rhodococcus were added to the treatment of B. amyloliquefaciens ZM9. These results indicate that the application of different microorganisms has an effect on the structure of bacterial community (Figure 3(C)). The distribution of the first 30 abundance bacterial genera in the three treatments is shown in the figure. Among them, Sphingobium, Pseudomonas, Haliangium and Bryobacter were dominant in the treatment of B. amyloliquefaciens ZM9, while the proportion in T1 was lower (Figure 3(D)).

In total 679,451 high-quality raw sequences with the average length of 252 bps for fungus were obtained from rhizospherial soil samples after quality filtering. The OTUs, Chao1 and Shannon indexes were used to evaluate and compare the richness and diversity of fungal communities under different treatments (Table 2). It can be seen from the data that the indexes of OTUs and Shannon in the rhizosphere soil treated by R. solanacearum were slightly lower than those of the CK and T2 groups. Compared with CK group, Shannon and Chao1 index slightly lower B. amyloliquefaciens ZM9 process. The results show that the addition of different microorganisms affects the richness and diversity of soil fungal communities to a certain extent.

The observed OTUs, Shannon diversity and Chao1 richness of Fungal communities in the soil are expressed as mean ± SE.CK: ddH₂O treatment; T1: R. solanacearum treatment; T2: B. amyloliquefaciens ZM9 treatment.

In tobacco soil samples with different treatments, Ascomycota accounted for 35.12% - 62.8% of the top 10 fungi phyla. Followed by Aphelidiomycota (6.52% - 12.64%), Basidiomycota (4.34% - 8.9%), Mortierelomycota (2.12% - 3.61%), Olpidiomycota (1.02% - 1.23%), Rozellomycata, Glomeromycota, Mucoromycota, Monoblepharomycota and Aphelidiomycota. Among the 10 dominant bacteria, the relative abundance of Ascomycota, Aphelidiomycota and Basidiomycota exceeded 4% (Figure 4(A)). Using the weighted UniFrac distance for principal coordinate analysis (PCoA) in different treatments, PC1 and PC2 explained 66.76% of the total fungal community. The fungal communities of the CK group at different time periods gathered together, the corresponding fungal communities of the T1 and T2 groups at different time periods gathered together, and the fungal communities of different treatment groups were separated from each other (Figure 4(B)). There are differences in the fungal community structure between different treatments in the heat map. The treatment of R. solanacearum did not make much difference in the fungal community; while the treatment of B. amyloliquefaciens ZM9 increased the abundance of Ascobolus, Serendipita, Rhizoctonia, Conlarium, Papulasporiella, Candida, Entoloma, Entoloma, Ceratobasidium and Chaetomium. The two treatments are very different from the CK group. These results indicate that different microorganisms have a greater impact on the structure of the fungal community (Figure 4(C)). Further analysis was carried out at the genus level, and the different distributions of the first 30 abundance fungal genera in the three treatments are shown in the figure. There

were significant differences in 4 fungal genera among the 3 treatments, namely Trichoderma, Fusarium, Aspergillus and Alternaria (Figure 4(D)).

Redundant analysis (RDA) was used to analyze the relationship between root exudates and microbial community structure under different treatments. The results showed that the community variation of bacteria and fungi was 39.85% and 43.96%, respectively (Figure 5). Benzoic acid, propionic acid, sebacic acid, caprylic acid, palmitic acid, oxalic acid, myristic acid, and 3-hydroxybutyric acid are positively correlated with the bacteria Ralstonia, and negatively correlated with the bacteria Sphingobium, Sphingomonas, Haliangium, Flavobacterium, Paenibacillus, and Bacillus (Figure 5(A)). Sphingosine, oxalic acid, benzoic acid, sebacic acid, propionic acid, myristic acid, 3-hydroxybutyric acid, palmitic acid are positively correlated with the fungi Fusarium and Alternaria, and negatively

correlated with the fungi Aspergillus and Trichoderma (Figure 5(B)). Redundancy analysis shows that the microbial community has a greater impact on soil root exudates.