The potting soil was obtained from the 0-15cm layer of the farmland around Hunan Agricultural University in Changsha, China (113.80°E, 28.18°N). The soil was naturally air-dried, sieved to remove plant residues and stones, and then passed through a 2 mm sieve for the potting test. The initial physicochemical properties of the soil were as follows: the pH was 5.61, and the contents of organic matter (OM), total nitrogen (TN), total phosphorus (TP), total potassium (TK), available phosphorus (AP), and available potassium (AK) were 5.740 g/kg, 0.820 g/kg, 0.130 g/kg, 14.780 g/kg, 1.570 mg/kg, and 96.340 mg/kg, respectively. The microbial fertilizers tested were yeast (Pichia sp.) microbial fertilizer J1, PGPB (Klebsiella oxytoca) microbial fertilizer ZS4, and Pichia sp. and Klebsiella oxytoca composite microbial fertilizer JZ. Fermentation Conditions: J1 microbial fertilizer: The carbon source is bran, and the nitrogen source is soybean meal, with a mass ratio 6:4. The solid-to-liquid ratio is 1:1.75. The inoculation of Pichia sp. J1 is 5%, with fermentation conducted at 30°C for nine days. The resulting effective viable cell count is 3.6 × 10⁹ CFU/g. ZS4 microbial fertilizer: The carbon source is bran, and the nitrogen source is soybean meal, with a mass ratio of 4:6. The solid-to-liquid ratio is 1:1.25. The inoculation of Klebsiella oxytoca ZS4 is 3%, with fermentation performed at 35°C for seven days. The resulting effective viable cell count is 1.1 × 10¹¹ CFU/g. JZ composite microbial fertilizer: This blend of J1 and ZS4 microbial fertilizer in a fresh weight ratio of 1:1. These strains have not been used in commercial formulations, just for research purposes. They were provided by the Microbial Conservation Laboratory, School of Mineral Processing and Bioengineering (Central South University, China). J1, ZS4, and JZ microbial fertilizers were not commercially available and were used only for this study. The variety of tobacco (Nicotiana tabacum L.) used in the pot experiment was “Yunyan 203”, provided by Changsha Tobacco Bureau Company of Hunan Province (China).

Four treatment groups were established for the pot experiment: control group without microbial fertilizer (CK group); applied J1 microbial fertilizer (J1 group); applied ZS4 microbial fertilizer (ZS4 group); and applied JZ composite microbial fertilizer (JZ group). Six replicates were set up for each treatment group, resulting in a total of 24 pots for the four treatment groups. When N. tabacum seedlings reached 3 - 5 true leaves, seedlings of similar size and health condition were selected and transplanted in pots (20.5 cm × 23.5 cm) containing 1.8 kg of soil each, one plant per pot. We mixed microbial fertilizer and soil thoroughly at 1:100 and applied it evenly around the roots of the tobacco plants. The control group applied a mixture of bran and soybean meal without microbial inoculation, and no other chemical fertilizers were added in the experiment to ensure consistency in each experimental condition. Subsequently, the tobacco plants were placed outdoors for natural growth and were watered with 300 mL every three days. we determined the growth indexes of tobacco. Soil samples (5.0 g) were collected near the roots (2-4 cm), and visible plant residues and stones in the soil were removed, then air-dried and crushed to a particle size of less than 0.147 mm (Xiao et al., 2018). One part of the soil samples was used to determine the physicochemical properties of the soil, and the other was stored in a -80°C freezer for microbiological studies.

The field experiment site was located in Daoyi Science and Technology Park, Huayuan County, Xiangxi Tujia and Miao Autonomous Prefecture, Hunan Province (25°15′38.8″N, 109°15′32.5″E, Altitude: 530 m). The climate of the region is subtropical monsoon, with an average temperature of 13.9°C, an annual sunshine length of 1600-1800 h, and an annual rainfall of 1400-1800 mm. Two treatment groups were set up for the experiment: the control group without microbial fertilizer (CK) and the group with mixed microbial fertilizer (JZ). Three blocks were set up for each treatment group, and the blocks were designed in a randomized arrangement. The area of each block was about 150 m², and about 200 tobacco plants were planted in each block. The planting density of tobacco was 125 cm between rows and 60 cm between plants, and protected rows were set around. At the time of transplanting, the tobacco seedlings in the JZ group were treated with microbial fertilizer by hole application, while the CK group was treated with a mixture of bran and soybean meal without microbial inoculation; at the time of the doughnut stage, the JZ group was supplemented with microbial solution, while the CK group was supplemented with fresh water. The number of colonies of microbial fertilizer applied to each tobacco seedling was controlled at 10×10⁹ CFU/g. During the tobacco growing period, both groups regularly applied chemical fertilizers, with irrigation methods consisting of rainfed irrigation and supplementary irrigation. Measurement of tobacco growth indexes at maturity and disease surveys for TBW and TBS.

Soil and ddH₂O were mixed at a ratio of 1:1.25, and soil pH was measured using a pH meter. (PHS-3E, Shanghai Yidian Scientific Instrument Co., Ltd., China). The soil OM content was determined by the potassium dichromate oxidation method, the soil TN content was determined by the Kjeller method, and the soil AK was extracted with ammonium acetate and determined by a flame photometer (Li et al., 2017). The soil TP content was determined by molybdenum-antimony colorimetry after H₂SO₄-HClO₄ digestion (Ding et al., 2024). The soil AP content was determined by the ammonium fluoride-hydrochloric acid extraction method (Zhang et al., 2021a). The soil TK content was determined by the NaOH alkali melting method and flame photometer (Xu et al., 2018).

After the tobacco was harvested, the number of leaves per plant was recorded; the largest leaf of each plant was selected, leaf length and width were measured with a tape measure, and the maximum leaf area of the tobacco was calculated; the distance from the bottom of the stem to the top growing point of the plant was measured with a tape measure as plant height, and the circumference of the base of the stem was measured as stem circumference; stems and leaves were weighed on an electronic balance as aboveground fresh weight (Shang et al., 2023). The tobacco bacterial wilt and black shank disease index of each plant were determined based on the Chinese national standard GB/T 23222-2008, as described by (Gao et al., 2015).

The extraction and high-throughput sequencing of rhizosphere soil microbial DNA were performed by Beijing Baimark Biotechnology Co. The V3-V4 region of the 16S rRNA gene was amplified with primers 338F (5’-ACTCCTACGGGGAGGCAGCA-3’) and 806R (5’-GGACTACHVGGGGTWTCTAAT-3’). The ITS1 region of the 18S RNA gene was amplified with primers ITS1 (5’-CTTGGTCATTTAGAGGAAGTAA-3’) and ITS2 (5’-GCTGCGTTTTAGGAAGTAA-3’). PCR reaction conditions were as follows: initial denaturation at 95°C for 3 min, denaturation at 95°C for 30 s, annealing at 55°C for 30 s, extension at 72°C for 45 s for 27 cycles, and a final extension at 72°C for 10 min. The sequencing raw data has been submitted to the SRA database on the NCBI website, and the accession numbers are PRJNA1153728 and PRJNA1153733.

Raw Reads were filtered using Trimmomatic v0.33 to detect and remove sequences with quality scores below 20. Then, the sequences were denoised and chimerically filtered using UCHIME v4.2 to obtain valid reads, which were further clustered into Operational Taxonomic Units (OTUs) using an identity threshold of 97%. Finally, the OTUs are assigned a classification using the RDP classifier to ensure a confidence level of at least 50%.

In this study, we utilized the relative abundance of bacterial and fungal OTUs in soil to construct a phylogenetic molecular ecological network (PMEN) through a random matrix theory (RMT)–based approach (Zhang et al., 2023a). This network construction method automatically identifies key OTUs and determines the network’s topological properties. The low-frequency OTUs in the samples are filtered out, and the high-frequency OTUs are retained. In our study, five or more OTUs in six sample replicates were retained for network construction. Finally, the visualization of network graphs and calculating network topology parameters were implemented using Gephi 0.10.1. Network parameters such as network density, average path length, average clustering coefficient, and modularity index were shown. Molecular ecological networks consist of many nodes (OTUs) and edges, and multiple nodes can form a module. The roles of nodes in a module are characterized by Within-module connectivity (Zi) and Participation Coefficient (Pi) (Tao et al., 2018). Based on Zi and Pi, nodes in a network can be categorized into four roles: peripheral (Zi ≤ 2.5 and Pi ≤ 0.62); connector (Zi ≤ 2.5 and Pi > 0.62); module hub (Zi > 2.5 and Pi ≤ 0.62); and network hub (Zi > 2.5 and Pi > 0.62). Nodes that play the roles of connectors, module hubs, and network hubs can serve as key taxa for maintaining network structure and functionality. We utilized ZiPi diagrams to find key species in microbial association networks. ZiPi diagrams were completed in Origin2022 and the Wekemo Bioincloud (https://www.bioincloud.tech).

PICRUSt2 combines existing open-source tools to predict the genome of 16S rRNA gene sequences of environmental samples. Predicted functions of the metagenome in the samples based on several gene family databases, including Kyoto Encyclopedia of Genes and Genomes (KEGG) homologs (KOs) and Enzyme Commission Numbers (EC numbers) using PICRUSt2. The PICRUSt2 methodology consisted of phylogenetic localization, hidden state prediction, and sample gene and pathway abundance tables. The OTU sequences and abundances were used as inputs, and gene/enzyme and KEGG pathway abundances were used as predicted outputs. The accuracy of metagenome predictions could be verified by comparing the results of differential abundance tests of 16s rRNA-predicted metagenome with MGS data (Douglas et al., 2020).

A completely randomized experimental design was used with six replicates per treatment. All data except microbiological data were statistically analyzed using IBM SPSS Statistics 26. Where Analysis of variance (ANOVA) and comparison of means were performed using Tukey’s multiple range test at p < 0.05. Venn plots were utilized to demonstrate the relationship between the number of OTUs in the different treatment groups. Differences in the structural composition of microbial communities were shown using a Non-metric multidimensional scaling (NMDS) analysis based on the Bray-Curtis distance. The species diversity and evenness in the microbial community were assessed using Shannon and Pielou indexes, and the Chao1 and ACE indexes were calculated to assess the species richness in the community. Analysis of species composition in microbial communities using species abundance histograms and chord diagrams. LDA Effect Size (LEfSe) analysis at the genus level was performed using LDA threshold > 3, p < 0.05 as a screening criterion to identify significantly different microorganisms between groups. Random Forest Analysis was utilized to assess the extent of the contribution of key microbes in rhizosphere soils. The relationship between microorganisms and soil physicochemical properties was described using redundancy analysis (RDA). Correlation analysis was performed to show Spearman’s correlation coefficients between soil physicochemical properties, tobacco growth indexes, and microbial communities to characterize the relationship between these variables. Analysis of Similarities (ANOSIM) was utilized to characterize the structural differences of the predicted functional genes. The above mapping was completed in Origin 2022 and the Wekemo Bioincloud (https://www.bioincloud.tech), utilizing Adobe Illustrator 2022 for graphic editing.

The physicochemical properties of the rhizosphere soil of tobacco were shown in Table 1 . The content of OM (5.89-7.29 g/kg), TN (0.75-0.95 g/kg), and TP (0.09-0.31 g/kg) differed significantly (p < 0.05) among the different treatment groups, and significantly (p < 0.05) higher in JZ and ZS4 than in CK. The soil content of OM, TN, and TP in the JZ group significantly (p < 0.05) increased by 23.77%, 26.67%, and 244.44%, respectively, compared to the CK group. There was no significant (p < 0.05) difference in AP (1.55-2.29 mg/kg), AK (95.40-106.86 mg/kg), and TK (15.14-16.52 g/kg) contents among the four treatment groups, but J1, ZS4, and JZ were higher than CK. Also, soil pH was significantly (p < 0.05) lower in all three groups than CK.

The tobacco growth indexes of the four treatment groups were shown in Table 1 . The picture of the growth of the tobacco plant was shown in Supplementary Figure S1 . The growth indexes of tobacco were higher in the JZ, ZS4, and J1 groups than in the CK group, with the JZ group having the highest indexes. The plant height, number of leaves, aboveground fresh weight, and maximum leaf area of tobacco in the JZ group were significantly (p < 0.05) increased by 89.64, 29.79, 197.73, and 83.04%, respectively, and those of tobacco in the ZS4 group by 22.44, 26.32, 171.87, and 80.85%, respectively, when compared with those of CK.

A field experiment was conducted to validate further the effect of no microbial fertilizer application and application of mixed microbial fertilizer groups on tobacco growth. The photographs of tobacco plant growth in the field experiment were shown in Supplementary Figure S2 . The growth and disease indexes of tobacco in the two treatment groups of the field experiment were presented in Table 4 . The plant height (113.27 cm), stem girth (11.02 cm), number of leaves (14.67 pc), maximum leaf length (82.93 cm), maximum leaf width (30.25 cm), maximum leaf area (1766.58 cm²) of tobacco were significantly (p < 0.05) increased in JZ group as compared to CK. The incidence rate (13.33%) and disease index (1.48) of tobacco bacterial wilt and incidence rate (3.33%) and disease index (0.37) of tobacco black shank in the JZ group were significantly (p < 0.05) lower than the CK group. The relative efficacy of tobacco bacterial wilt and black shank in the JZ group was 85.71% and 66.67%, respectively, greater than that in the CK group.

Microbial fertilizers have been shown to promote the growth of tobacco by increasing soil fertility and reducing the incidence of soil-borne diseases, and composite microbial fertilizers have shown superior effects compared to individual microbial fertilizers, which is in accordance with the results of the present study (Ruan et al., 2020). The results of the pot experiment showed that the tobacco growth indexes of J1, ZS4, and JZ groups with microbial fertilizers were higher than that of the CK group without microbial fertilizers, and the effect of composite microbial fertilizers JZ was significantly greater than that of individual microbial fertilizers J1, suggesting that Pichia sp. J1 had the potential to augment the growth-promoting effects of PGPB on tobacco. In addition, further field experiments showed that tobacco in the JZ group with compound fertilizer was significantly (p < 0.05) higher than that in the J1 group with individual microbial fertilizer in growth promotion effect and disease resistance. These results may be related to microbial fertilizer-induced changes in soil physicochemical properties, microbial community structure, and promotion of microbial metabolic activities (Zhai et al., 2020). Therefore, it is necessary to study soil properties and soil rhizosphere microbial communities in soil-microbe-tobacco systems.

The present study showed that OM, AP, AK, TN, TP, and TK contents were increased in J1, ZS4, and JZ groups added with microbial fertilizers compared to the CK group without microbial fertilizers ( Table 1 ), which could be because the microbial fertilizers increased the nutrient content of the soil, thus promoting the growth of tobacco (Kour et al., 2020). Pichia sp. has an ammonia-oxidizing capacity that converts ammonia in the soil and increases plant uptakeable N (Zhang et al., 2021b). Klebsiella oxytoca has a phosphate solubilizing capacity and nitrogen fixation that increases the soil N and P content (Khalifa and Aldayel, 2022). Notably, we did not detect Klebsiella oxytoca in the later stages of the experiment. This may be because, as the microbial community evolved, Klebsiella oxytoca gradually disappeared, and even though it died out in the later stages, its early effects were still present (Xu et al., 2019). In addition, the contents of OM, AP, TN, and TP in the JZ group were higher than those in the J1 and ZS4 groups, which indicated that the enhancement of nutrients by composite microbial fertilizers was better than that of individual microbial fertilizers, and this also explained that the promotion effect of composite microbial fertilizers was higher than that of individual microbial fertilizers on tobacco. Correlation analysis showed that the content of OM, TN, and TP in the soil was positively correlated with aboveground fresh weight, number of leaves, maximum leaf area, and plant height of tobacco ( Figure 5C ). This might be due to the enrichment of beneficial microorganisms in the soil after the application of microbial fertilizers, which could fix nitrogen, dissolve phosphorus and potassium, and decompose organic matter, thus promoting the uptake of plant nutrients (Wei et al., 2024). As shown in Table 1 , pH was significantly (p < 0.05) lower in the J1, ZS4, and JZ groups compared to the CK group. Soil pH was significantly (p < 0.05) negatively correlated with tobacco aboveground fresh weight and plant height ( Figure 5C ). This phenomenon could be caused by soil microorganisms such as Klebsiella oxytoca, which carry out metabolic activities to secrete acids such as organic acids (Park et al., 2013; Zúñiga-Silgado et al., 2020). A decrease in soil pH can increase the mobility of essential heavy metals such as iron and magnesium in the soil, which promotes the effective uptake of these heavy metals by plants and further promotes plant growth (Zeng et al., 2011).

Our results showed that applying microbial fertilizers led to changes in the structure and diversity of soil microbial communities in the tobacco rhizosphere. From Figure 1 , soil microbial community diversity and richness decreased in J1, ZS4, and JZ groups applied with microbial fertilizers compared to CK. This may be due to the effect of Pichia sp. J1 and Klebsiella oxytoca in the soil, which makes microbial fertilizers alter the physicochemical properties of the soil and affect the microbial environment, resulting in the inability of certain microorganisms to adapt to the new conditions and reducing microbial diversity (Zhang et al., 2017; Wang et al., 2018). This study showed that the relative abundance of Proteobacteria, Sphingomonas, and Mortierella PGBRs was higher in J1, ZS4, and JZ groups than in the CK group ( Figure 2 ). PGBR competed for resources with pre-existing microorganisms in the soil, which led to reduced numbers of inter-root pathogens and other harmful microorganisms. They decreased microbial diversity, promoting plant growth (Podile and Kishore, 2007). Competitive, synergistic, or mutually beneficial relationships may occur between different microorganisms, enhancing or diminishing microbial communities’ network complexity (Selim and Zayed, 2017). As revealed by the co-occurrence network analysis of microorganisms ( Figure 3 ), the structural complexity of the soil microbial community network was increased in the J1, ZS4, and JZ group compared to CK, suggesting that microbial fertilizers may enhance microbial synergistic effects, which in turn promotes the growth of tobacco (Wu et al., 2024).

Through Spearman correlation analysis, we revealed the relationship between key microorganisms, soil factors, and tobacco growth indicators. Key bacteria such as Sphingomonas and Granulicella significantly (p < 0.05) enhanced aboveground fresh weight, leaf number, maximum leaf area, and plant height of tobacco. They were significantly (p < 0.05) positively correlated with TN and TP content in soil ( Figure 5D ). Previous studies have shown that Sphingomonas can fix nitrogen, ammoniate, and solubilize phosphorus, which helps to increase the N and P content of the soil, which in turn promotes plant growth (Wang et al., 2024). It also regulates phytohormones such as abscisic acid, jasmonic acid, and salicylic acid, which are essential for plant development and disease defense (Mazoyon et al., 2023). Granulicella promotes plant growth by synthesizing IAA and extracellular polysaccharides and inhibits pathogen growth by competitively adsorbing iron from the soil, reducing the source of iron available to pathogens (Kielak et al., 2016). Our findings were consistent with these reports. In addition, Kitasatospora and Nitrosospira were significantly (p < 0.05) positively correlated with aboveground fresh weight and maximum leaf area of tobacco, and Nitrosospira was significantly (p < 0.05) positively correlated with soil TN ( Figure 5D ). This correlation could be attributed to Nitrosospira’s role as a nitrifying bacterium involved in nitrification, which promoted plant growth by modifying soil N morphology and increasing plant AN (Ayiti et al., 2022). Kitasatospora acts as a PGPB that produces high concentrations of IAA (Arunachalam Palaniyandi et al., 2013) or promotes plant growth by inhibiting plant pathogens through the production of antimicrobials such as propioxatins, terpentecin, and setamicin (Arunachalam Palaniyandi et al., 2013). Our study also identified some key fungi affecting tobacco growth, such as Mortierella, Trichoderma, Condenascus, and Waitea, that positively correlated with growth indicators of tobacco ( Figure 5E ). It reported that Mortierella and Trichoderma belong to plant growth-promoting fungi. They could produce phytohormone and ACC deaminase to promote the growth of tobacco, which could protect the plant from pathogens and reduce the incidence of TBW (Deng et al., 2018; Li et al., 2020). In addition, Mortierella improved P content in soil, which was consistent with the results of the present study, showing a significant (p < 0.05) positive correlation between Mortierella and TP (Ozimek and Hanaka, 2021).

This study investigated the role of rhizosphere microbial metabolic activities in promoting plant growth. As seen from Table 2 , the abundance of Sucrose-phosphate synthase (SPS), ACC Deaminase, Nitrogenase, and Alkaline Phosphatase was significantly (p < 0.05) higher in J1, ZS4 and JZ groups with added microbial fertilizers than in the CK group (without microbial fertilizers). We speculated that these enzymes facilitate tobacco growth by participating in the cycling of carbon, nitrogen, and phosphorus in the soil and regulating the uptake of tobacco nutrients. Specifically, ACC deaminase promotes plant growth by cleaving ACC, providing essential carbon and nitrogen sources for plants and microorganisms, and reducing ethylene levels in plants (Arshad et al., 2007). Nitrogenase, encoded by the Nif gene, can fix atmospheric nitrogen (N₂) to provide plants with a needed nitrogen source (Zhang et al., 2023b). Alkaline phosphatase catalyzes the hydrolysis of organophosphorus compounds, releasing inorganic phosphate, which can be correlated to the key microbes Sphingomonas and Mortierella for phosphate solubilization. In addition, this study found that adding microbial fertilizer significantly enhanced the activities of antioxidant enzymes in plants. The abundance of Catalase peroxidase and Peroxidase was significantly (p < 0.05) higher in the J1, ZS4, and JZ groups than in the CK group. These antioxidant enzymes could effectively scavenge reactive oxygen species (ROS) and reduce oxidation-induced plant cell damage, promoting tobacco growth (Khan et al., 2023). The abundance of β-glucosidase, Chitinase, and Nonribosomal peptide synthetase was higher in the treatment group with microbial fertilizer addition than in the CK group. Chitinase can degrade chitin, a significant component of the cell wall of phytopathogenic fungi, thereby defending against fungal diseases while improving plant growth and yield (Kumar et al., 2018). Nonribosomal peptide synthetase is responsible for the synthesis of nonribosomal peptides, which are peptide natural products with properties such as toxins, iron carriers, and antibiotics (Martínez-Núñez and López, 2016). MATE transporter proteins and ATP-binding cassette transporter proteins (ABC transporter proteins) had the highest abundance in the ZS4 and J1 groups, respectively. We concluded that these enzymes function in inhibiting tobacco diseases and promoting tobacco growth. MATE transporter proteins are involved in the transport of organic acids, phytohormones, and secondary metabolites and regulate plant disease resistance as well as metal detoxification; ABC transporter proteins have been shown to participate in membrane transport of endogenous secondary metabolites in plants and play an essential role in antifungal diterpene secretion and heavy metal detoxification (Liu et al., 2023). Overall, adding microbial fertilizers increased the levels of some enzymes and genes related to plant growth promotion and resistance to pathogens, thereby promoting nutrient uptake, disease resistance, and growth of tobacco.

Analysis of the predicted metabolic pathways revealed that the abundance of biosynthetic pathways of secondary metabolites of some microorganisms was higher in the treatment group where microbial fertilizers were applied than in the CK group ( Table 3 ). Tetracycline biosynthesis was significantly (p < 0.05) higher in the J1, ZS4, and JZ groups than in the CK group. Tetracycline has been shown to have antipathogenic activity (Daghrir and Drogui, 2013). Moreover, Tetracycline plays a role in inhibiting TBS and TBW (Tang et al., 2023). Alkaloids such as Tropane, Piperidine, Pyridine, and Isoquinoline Alkaloids and flavonoids such as Flavone and flavonol exhibit a wide range of biological activities, including antiviral, antiparasitic, antimicrobial, and antioxidant properties (Sireesha et al., 2019; Zandavar et al., 2023). Phenylpropanoid and Brassinosteroid biosynthesis abundance was remarkably (p < 0.05) higher in the J1, ZS4, and JZ groups than in the CK group. The biosynthesis of Phenylpropanoid has been shown to inhibit Cercospora nicotianae and Phytophthora nicotianae, thereby enhancing disease resistance in tobacco (Yadav et al., 2020). Brassinolide (BL) is the essential brassinosteroid, which plays a crucial role in plant growth and development, and BL-treated wild-type tobacco was significantly more resistant to tobacco mosaic virus (TMV), Pseudomonas syringae pv. tabaci (Pst), and Oidium sp (Nakashita et al., 2003). This study found that the application of J1, ZS4, and JZ microbial fertilizers enhanced the activity of the above-mentioned antimicrobial compounds in tobacco, which reduce tobacco Soil-borne diseases by inhibiting the activity of pathogenic bacteria, thus promoting their growth, which could also explain the low incidence of TBW and TBS in the field experiment.