Method for analyzing soil bacterial flora, related soil bacterial flora and applications thereof

Abstract

The application provides a soil microbial community analysis method, a related soil microbial community and application thereof. Based on a long-term positioning test of soil, absolute quantitative methods are used for the first time to explore dynamic assembly rules of rhizosphere and intraradical microbial communities in a whole life cycle of plants, response characteristics of the rhizosphere microbial community to different fertilization treatments are researched, and by combining methods such as metagenomics, microbial separation and culture and synthetic communities, functional adaptation of the rhizosphere microorganism under different fertilization conditions and a mechanism for affecting growth of leguminous plants are revealed. The application also provides a soil microbial community for improving chemical properties of soil and a synthetic microbial community for promoting growth and biomass of leguminous plants.

Description

Technical Field

[0001] This invention belongs to the field of botany, and more specifically, this invention relates to methods for soil microbial community analysis, related soil microbial communities and their applications. Background Technology

[0002] Soil microbial communities are interacting communities living in the soil, composed of bacteria, archaea, fungi, viruses, and protists. The taxonomic diversity of these microbial communities and the abundance of individual members are referred to as community structure. Soil microbiomes are a key component of ecosystems and the most diverse community in the biosphere, accounting for more than a quarter of the Earth's total biodiversity.

[0003] Tens of millions of bacteria, archaea, fungi, viruses, and protozoa coexist underground, with hundreds of thousands described. Although advancements in technology, such as marker genes, genomics, metabolomics, metagenomics, and metatranscriptomics, have improved our understanding of the soil microbiome, the majority of soil microorganisms remain undescribed.

[0004] Soil microbial community structure is diverse across different ecosystems and microhabitats. A variety of soil factors can shape the composition of the microbial community, including pH, nutrient levels, organic matter content and quality, available water, and oxygen levels.

[0005] Soil microorganisms are key participants in soil structure formation and important indicators of soil quality. Soil structure is not simply a combination of soil particles and fertilizers, but a three-dimensional arrangement of organic compounds (aggregates) and pore spaces. As an essential component of soil, soil microorganisms, through metabolic activities involving oxygen and carbon dioxide exchange and the secretion of organic acids, contribute to the formation of soil aggregates. Arbuscular mycorrhizal fungi (AMF) play a crucial role in this process. The importance of soil microorganisms is widely recognized; they play vital roles in nutrient cycling, carbon fixation, environmental remediation, soil fertility, and plant health and productivity.

[0006] Living soil microorganisms are the main engine of terrestrial biogeochemical cycles, driving the turnover of soil organic matter through their primary metabolic activities. The cellular remains of dead microorganisms serve as fuel for this biogeochemical engine, as their chemical components contribute to the accumulation of soil organic matter, forming approximately 50% of the soil organic matter pool. Soil microorganisms are primarily responsible for the cycling of soil organic carbon (SOC) and other nutrients, playing a crucial role in climate feedback, including the production and consumption of greenhouse gases such as CO2, CH4, and N2O. Soil microorganisms participate in the mineralization and stabilization of organic carbon; the balance between these two processes controls the net flux of greenhouse gases into the atmosphere, influencing global climate. Soil microorganisms are game-changers in the restoration of degraded land. In arid soils, some algae can enhance soil water retention and transport capacity by disrupting hydrophobicity and increasing water retention, while fungal hyphae can penetrate air-filled soil pores to obtain water and nutrients. Different bacteria can also produce hydrophilic or hydrophobic substances as part of extracellular polymeric substances (EPSs), thus affecting soil moisture. For example, Bacillus sphaericus of the Firmicutes phylum is hydrophilic; some bacteria of the Actinobacteria and Proteobacteria phyla can degrade waxes, thereby reducing hydrophobicity. In saline-alkali soils, halophilic or halophilic plant growth-promoting rhizobacteria (PGPR) and AMF have been used for bioremediation of saline-alkali soils, which is a sustainable alternative to conventional physical and chemical remediation. Halophilic PGPR can directly remediate saline soils by improving soil nutrients, soil structure, organic matter, pH, electrical conductivity, and ion salt deposition. In contaminated soils, some indigenous microorganisms can degrade pollutants through metabolic activities, playing a role in bioremediation. Studies have reported that PGPR (Polymer Progenitor Cell Remediation) also has an indirect positive effect on pollutant removal. These microorganisms can enhance the ability of plants to remediate contaminated soil by stimulating plant growth and increasing the bioavailability of pollutants. Microorganisms can also assist in the remediation of soils contaminated with heavy metals, pesticides, or hydrocarbons, further promoting the restoration of degraded ecosystems. For example, when vegetation is reintroduced into contaminated soil, inoculating the soil with fungi not only allows for the absorption of heavy metals from the contaminated soil but also enables plants to successfully establish themselves in degraded soil, thereby improving soil quality and health.

[0007] Rhizosphere microorganisms and host plants have a close relationship. Plants, due to their unique genetic, physiological, and metabolic characteristics, influence the composition and structure of the rhizosphere microbial community. Conversely, the rhizosphere microbial community plays a crucial role in plant growth, development, health, and yield due to its complex interactions and functions.

[0008] To increase crop yields, large amounts of nitrogen, phosphorus, and potassium fertilizers are applied to agricultural production, leading not only to eutrophication of water bodies and air pollution but also severely damaging the structure and function of soil microorganisms. Plant rhizosphere microorganisms play a crucial role in soil nutrient cycling and crop nutrition and health. However, systematic research on the dynamic response characteristics of rhizosphere microbial communities under long-term different fertilization conditions and their interactions with crops remains lacking.

[0009] Soil microorganisms are crucial for nutrient cycling and crop health, playing a role in regulating plant growth and development, promoting nutrient uptake, and enhancing plant resistance to abiotic and biotic stresses. However, a comprehensive understanding of the structure and function of root-associated microbial communities under nutrient-deficient conditions throughout the plant life cycle remains lacking. The interactions between rhizosphere microorganisms and plants are complex, and although numerous studies have revealed some mechanisms, many questions remain unanswered. Therefore, further research is needed to develop effective analytical methods and explore more technical solutions for soil improvement, soil microbiome optimization, and yield increase. Elucidating the mechanisms of plant-rhizosphere microbiome interactions under different nutrient conditions is beneficial for achieving healthy crop growth and increased yields through improvements to the rhizosphere microbiome. Summary of the Invention

[0010] The purpose of this invention is to provide a method for soil microbial community analysis, related soil microbial communities, and their applications.

[0011] In a first aspect of the present invention, a method is provided for improving the soil properties and soil microbiota of fertilized soil, or promoting the symbiotic nitrogen fixation of rhizobia and leguminous plants, comprising: adjusting the application methods of nitrogen fertilizer, phosphorus fertilizer and potassium fertilizer, and reducing (including reducing to 0) the amount of nitrogen fertilizer applied.

[0012] In one or more embodiments, the improved soil properties include: reducing soil acidification (increasing soil pH, such as by 0.5 units or 1 unit or more), increasing the number and diameter of root nodules, and increasing rhizobium abundance (the amount of symbiotic rhizobium in the rhizosphere); preferably, in legumes, the developmental pattern of the microbial community changes, with its α-diversity increasing in the early stages of development, enhanced nitrogen metabolism function of the microbial community, and enhanced nitrogen mineralization capacity of the microbial community; preferably, in legumes, the complexity and stability of the rhizosphere microbial community co-occurrence network are improved.

[0013] In one or more embodiments, the soil includes: black soil.

[0014] In one or more embodiments, the black soil is black soil used for cultivating or planting plants.

[0015] In one or more embodiments, phosphate and potassium fertilizers are applied at normal rates.

[0016] In one or more embodiments, the method improves the soil properties and soil microbiota of the fertilized soil, or promotes symbiotic nitrogen fixation between rhizobia and leguminous plants, while maintaining (without reducing or affecting) or increasing plant yield.

[0017] In one or more embodiments, the method causes changes in the assembly of the rhizosphere and root microbiota, which change dynamically with plant development.

[0018] In one or more embodiments, the soil is used for planting leguminous plants, and the fertilization method is: 75 kg N ha -1 150kg P2O5ha -1 75kg K2O ha -1 The amount of N, P2O5 or K2O can fluctuate by 50% (preferably by 40%, 30%, 20%, 10% or 5%).

[0019] In one or more embodiments, the nitrogen fertilizer includes urea (containing active ingredient N, or with N as the main active ingredient), the phosphate fertilizer includes superphosphate (containing active ingredient P2O5, or with P2O5 as the main active ingredient) and diammonium phosphate (containing active ingredient P2O5, or with P2O5 as the main active ingredient), and the potassium fertilizer includes potassium sulfate (containing active ingredient K2O, or with K2O as the main active ingredient).

[0020] In one or more embodiments, the cultivation or planting is a soybean-corn-wheat rotation cultivation or planting.

[0021] In one or more embodiments, the fertilization method during the wheat and corn seasons is: 150 kg N ha -1 75kgP2O5 ha -1 75kg K2O ha -1 Preferably, the amount of each component can fluctuate by 50% (preferably by 40%, 30%, 20%, 10% or 5%).

[0022] In one or more embodiments, nitrogen, phosphorus, and potassium fertilizers are applied after the autumn crop harvest; preferably, 50% of the nitrogen fertilizer in the maize season is applied at the jointing stage.

[0023] In one or more embodiments, the method further includes: preparing a synthetic microbial community SynCom5 or SynCom7, applying it (e.g., inoculating) to legumes to promote the growth of legumes, increase biomass, increase growth hormone levels, and enhance ACC deaminase activity.

[0024] In one or more embodiments, the SynCom5 comprises the following strains: Rhodococcus sp., Lysobacter sp., Terrabacter sp., Arthrobacter sp., Phyllobacterium sp.;

[0025] In one or more embodiments, the SynCom7 includes: strains of the synthetic microbial community SynCom5, as well as Bosea sp. and Aeromicrobium sp.

[0026] In one or more embodiments, the increase in biomass includes: increasing the plant height of leguminous plants, increasing the plant weight of leguminous plants; or, the growth hormone includes: IAA.

[0027] In another aspect of the invention, an application of the method is provided for improving the soil properties and soil microbiota of fertilized soil, or promoting symbiotic nitrogen fixation between rhizobia and leguminous plants.

[0028] In another aspect of the invention, an application of the method is provided for promoting the growth of leguminous plants, increasing biomass, increasing growth hormone levels, and enhancing ACC deaminase activity; preferably, the increase in biomass includes increasing plant height and increasing plant weight; preferably, the growth hormone includes IAA.

[0029] In another aspect of the invention, an artificially synthesized microbial community is provided, selected from the synthetic microbial community SynCom5 or SynCom7; said SynCom5 includes the following strains: Rhodococcus sp., Lysobacter sp., Terrabacters sp., Arthrobacter sp., Phyllobacterium sp.; said SynCom7 includes: strains of the synthetic microbial community SynCom5, as well as Bosea sp. and Aeromicrobium sp.; preferably, the artificially synthesized microbial community is contained in the composition; more preferably, the composition is a solution or suspension culture medium.

[0030] In one or more embodiments, the accession number of Rhodococcus sp. at the China Center for Type Culture Collection (CCTCC) is CCTCC M 2024006, the accession number of Lysobacter sp. at the CCTCC is CCTCC M2024011, the accession number of Terrabacter sp. at the CCTCC is CCTCC M2024007, the accession number of Arthrobacter sp. at the CCTCC is CCTCC M2024009, the accession number of Phyllobecacterium sp. at the CCTCC is CCTCC M2024010, the accession number of Bosea sp. at the CCTCC is CCTCC M 2024008, and the accession number of Aeromicrobium sp. at the CCTCC is CCTCC M 2024005.

[0031] In another aspect of the invention, a composition is provided comprising the aforementioned synthetic microbial community, or a culture of the aforementioned synthetic microbial community; preferably, it further comprises an agriculturally acceptable carrier, or a culture medium component (nutrients suitable for strain culture).

[0032] In one or more embodiments, the composition comprises: a pesticide composition; preferably, the formulation of the composition includes (but is not limited to): a solution (such as a sprayable solution), a suspension, an emulsion, an oily or aqueous dispersion, a powder (such as a lyophilized powder), a granule, or a microcapsule; preferably, the composition is a culture medium containing the artificially synthesized microbial community.

[0033] In another aspect of the invention, the application of the artificially synthesized microbial community or the composition is provided for promoting the growth of leguminous plants, increasing biomass, increasing growth hormone levels, and improving ACC deaminase activity; or for preparing compositions or kits that promote the growth of leguminous plants, increase biomass, increase growth hormone levels, and improve ACC deaminase activity.

[0034] In another aspect of the present invention, a method is provided to promote the growth of legumes, increase biomass, increase growth hormone levels, and enhance ACC deaminase activity, comprising: applying (e.g., inoculating) the artificially synthesized microbial community or the composition to the legume.

[0035] In another aspect of the invention, a kit is provided for promoting the growth of leguminous plants, increasing biomass, increasing growth hormone levels, and enhancing ACC deaminase activity, comprising the artificially synthesized microbial community or the composition described above.

[0036] In another aspect of the present invention, a method is provided for analyzing the ecology (existence status, including determining the characteristics and functions (including potential functions) of soil microbiota (including rhizosphere microbiota) under different fertilization methods, comprising: (1) fertilizing plants (e.g., legumes) with different fertilization methods; (2) obtaining the abundance of core ASVs in treatment groups with different fertilization methods; (3) analyzing the differences in core ASV abundance among different treatment groups using an absolute quantitative method based on the core ASV abundance obtained in (2); identifying biomarker groups (core microbiota marker groups) in different fertilization treatment groups or different developmental stages (including different time points), preferably using linear discriminant analysis (LEfSe) and / or random forest method; (4) analyzing the function of microbiota using a microbiota function annotation database; preferably, the microbiota function annotation database includes the FAPROTAX database; (5) analyzing the differences in microbiota function among different fertilization treatment groups by metagenomic sequencing and / or assembly.

[0037] In one or more embodiments, in (3), linear discriminant analysis is used to analyze the differences in species between treatment groups with different fertilization methods or at different developmental stages for microbial communities (such as rhizosphere soil microbial communities); preferably, the analysis is performed on the microbial communities from phylum to species.

[0038] In one or more embodiments, in (3), random forest analysis is used to establish a random forest prediction model; preferably, for the microbial community, a random forest prediction model is established at the family, genus and / or species level.

[0039] In one or more embodiments, in (5), the metagenomic analysis includes (but is not limited to): sequencing, classification, gene (functional) annotation; annotation to databases (including KEGG, COG and CAZyme), calculation of functional diversity among treatment group samples with different fertilization methods, including α-diversity and β-diversity; and analysis of the abundance differences of nitrogen, phosphorus and potassium cycle-related functional genes in microorganisms of treatment groups with different fertilization methods.

[0040] In one or more embodiments, the method further includes: performing co-occurrence network analysis and / or ecological cluster function analysis to obtain the microbial communities that play a decisive role under different fertilization methods; preferably, it further includes isolating the microbial communities that play a decisive role.

[0041] In one or more embodiments, the co-occurrence network analysis includes: calculating the Spearman correlation between each ASV based on the absolute abundance of core ASVs in the rhizosphere and root; using data with correlation coefficient r>0.8 and significance P<0.05 to construct bacterial co-occurrence networks for treatment groups with different fertilization methods in the rhizosphere and root; and calculating basic network topological characteristics including the number of nodes, number of edges, average degree, and robustness of each network. Preferably, the robustness of the rhizosphere or root microbial community network of treatment groups with different fertilization methods is further analyzed by randomly removing 50% of the nodes in the network.

[0042] In one or more embodiments, the microbial community that plays a decisive role includes (but is not limited to): a microbial community that improves the soil properties and soil microbial community of the fertilized soil, a microbial community that promotes the symbiotic nitrogen fixation of rhizobia and leguminous plants, and a microbial community that promotes the growth of leguminous plants, increases biomass, increases growth hormone levels, and enhances ACC deaminase activity.

[0043] In one or more embodiments, the different fertilization methods include (but are not limited to): variations in the application methods of nitrogen, phosphorus, and potassium fertilizers; preferably, the variations include increasing or decreasing (including decreasing to 0) the amount of nitrogen, phosphorus, and / or potassium fertilizers used.

[0044] In one or more embodiments, the fertilization methods for nitrogen, phosphorus, and potassium fertilizers are varied to form different fertilization treatment groups.

[0045] In one or more embodiments, after obtaining the microbial community that plays a decisive role under different fertilization methods, the method further includes: artificially preparing the microbial community to form a synthetic community, and re-inoculating it to verify the function of the microbial community; preferably, after re-inoculation, the soil properties and soil microbial community of the fertilized soil are ultimately improved, the symbiotic nitrogen-fixing microbial community of rhizobia and leguminous plants is promoted, and / or the growth of leguminous plants and the microbial community that increases biomass are promoted.

[0046] In another aspect of the invention, the method is provided for use in analyzing the ecology (existence status, including determining the characteristics and functions (including potential functions) of soil microorganisms (including rhizosphere microorganisms) under different fertilization methods, thereby obtaining microbial communities that play a decisive role under different fertilization methods; preferably, the microbial communities include (but are not limited to): microbial communities that improve the soil properties and soil microorganisms of the fertilized soil, microbial communities that promote the symbiotic nitrogen fixation of rhizobia and leguminous plants, and microbial communities that promote the growth of leguminous plants, increase biomass, increase growth hormone levels, and improve ACC deaminase activity.

[0047] Other aspects of the invention will be apparent to those skilled in the art from the disclosure herein. Attached Figure Description

[0048] Figure 1 Design and use of internal standard plasmids.

[0049] Figure 2 Overall experimental design. (A) Randomized block design of the black soil fertility field experiment. (B) Schematic diagram of soybean growth and development stages. (C) Schematic diagram of soybean rhizosphere and sequencing items.

[0050] Figure 3 Soil physicochemical properties. (AF) represent the alkaline nitrogen content, available phosphorus content, available potassium content, organic matter content, soluble organic carbon content, and pH value of non-rhizosphere soil in each fertilization treatment. Error bars represent standard deviation (SD), and asterisks indicate significant differences compared to the control NPK treatment (*P<0.05, **P<0.01, ***P<0.001).

[0051] Figure 4 Soybean yield and nodule phenotype. (A and B) represent soybean yields in two consecutive planting seasons in 2017 and 2020, respectively. (C and D) represent the number and diameter of soybean nodules at the initial flowering stage (D42), respectively. Error bars represent standard deviation (SD), and asterisks indicate significant differences compared to the control NPK treatment (*P<0.05, **P<0.01, ***P<0.001).

[0052] Figure 5 The relationship between the amount of internal standard plasmid added and the number of sequences. (AC) represent the relationship between the amount of internal standard plasmid added and the number of internal standard sequences in the sequencing results in non-rhizosphere soil, rhizosphere soil and root samples, respectively.

[0053] Figure 6 Bacterial community diversity. All results were calculated based on absolute abundance. (AC) Shannon diversity indices for non-rhizosphere, rhizosphere, and rhizosphere bacterial communities, respectively. (DF) β-diversity difference analysis for non-rhizosphere, rhizosphere, and rhizosphere bacterial communities, respectively. (G) Relationship between temporal distance and Bray-Curtis distance between different samples in each fertilization treatment in rhizosphere (left) and rhizosphere (right) samples. (H) Relationship between development time of each sample and Bray-Curtis distance of PK, NK, and NP samples relative to NPK samples in rhizosphere (left) and rhizosphere (right) samples.

[0054] Figure 7 Bacterial community composition. (AC) represent the absolute abundance of bacteria (top) and the relative abundance of the nine major phyla (bottom) in non-rhizosphere soil, rhizosphere soil, and root samples, respectively. Lowercase letters indicate significant differences in absolute bacterial abundance between different developmental stages (P<0.05).

[0055] Figure 8 Composition of the rhizobium community. (AC) represent the absolute abundance of rhizobia in non-rhizosphere soil, rhizosphere soil, and rhizobium samples (top) and the relative abundance of the nine major rhizobia (bottom), respectively. Lowercase letters indicate significant differences in absolute abundance of rhizobia among different developmental stages (P<0.05).

[0056] Figure 9 Dynamic trends of bacterial abundance over time. (AB) show the dynamic trends of absolute and relative abundance of bacteria (including Proteobacteria, Actinobacteria, Bacteroidetes, Firmicutes, Acidobacteria, and Chlorconoids) in rhizosphere soil and root samples, respectively, as plants develop.

[0057] Figure 10 Dynamic changes of rhizosphere ASVs with developmental stages. (AD) represent the dynamic changes in relative abundance (left) and absolute abundance (right) of high-abundance rhizosphere bacteria ASVs with plant development in NPK, PK, NK, and NP treatments, respectively. (E) The proportions of ASVs with similar (green) and opposite (red) trends in the dynamic changes of relative and absolute abundance in each fertilization treatment.

[0058] Figure 11 Differences in rhizosphere core ASVs among different fertilization treatments. All results are calculated based on absolute abundance. (A) The proportion of rhizosphere core ASV species (left) and absolute abundance (right) in total rhizosphere ASVs. (B) The relationship between the abundance of rhizosphere core ASVs and the abundance of total rhizosphere ASVs in the four different fertilization treatments. (C) The absolute abundance of rhizosphere core ASVs in each fertilization treatment and at each developmental stage.

[0059] Figure 12 1. Rhizosphere differentially expressed core ASVs in each fertilization treatment. All results were calculated based on absolute abundance. (AC) represent the rhizosphere differentially expressed core ASVs between the PK, NK, and NP treatments and the normal fertilization treatment, respectively. Red indicates exclusion, blue indicates enrichment, and the color of the clade represents different bacterial phylum classifications. All ASVs in the figure showed significant differences (P<0.05).

[0060] Figure 13 Biomarker taxa analysis in different fertilization treatments. All results are calculated based on absolute abundance. The figure shows the biomarker taxa (family level) in different fertilization treatments, P<0.05, LDAscore>6.

[0061] Figure 14Biomarker taxa analysis at each developmental stage. All results were calculated based on absolute abundance. The figure shows biomarker taxa (family level) at different growth and development stages, P<0.05, LDA score>6.

[0062] Figure 15 Rhizosphere samples were analyzed in a random forest based on fertilization treatments. All results were calculated based on absolute abundance, and all data were analyzed at the family / taxonomic level. (A) The 16 biomarker groups are ranked from most important to least important. The bar color represents the phylum-level classification of different bacteria, and the inset shows the 10-fold cross-validation error rate. (B) Absolute abundance of the 16 biomarker groups under different fertilization treatments.

[0063] Figure 16 Rhizosphere samples were analyzed using a random forest approach based on developmental stages. All results were calculated based on absolute abundance, and all data were analyzed at the family / taxonomic level. (A) The 16 biomarker groups are ranked from most important to least important. The bar colors represent the phylum-level classification of different bacteria, and the inset shows the 10-fold cross-validation error rate. (B) Absolute abundance of the 16 biomarker groups at different developmental stages.

[0064] Figure 17 Functional differences of core rhizosphere ASVs among different fertilization treatments. All results were calculated based on functional abundance proportions. The figure shows the abundance differences of 16 bacteria with high abundance proportions among different rhizosphere fertilization treatments. An asterisk indicates a significant difference compared with the normal fertilization treatment (*P<0.05, **P<0.01, ***P<0.001).

[0065] Figure 18 Functional differences of rhizosphere core ASVs at different developmental stages. All results were calculated based on functional abundance proportions. The figure shows the abundance differences of 16 bacteria with high abundance proportions at different rhizosphere developmental stages. Lowercase letters represent significant differences in the relative abundance of bacterial functions at different developmental stages (P<0.05).

[0066] Figure 19 Evolutionary analysis and functional annotation of bacterial bins. Evolutionary analysis of 140 bacterial bins obtained from rhizosphere sample binning, and annotation of functional genes related to nitrogen, phosphorus, and potassium cycling. Clade colors represent phylum-level classification, and small square colors represent the functions of genes in each category.

[0067] Figure 20Diversity and differential analysis of bacterial function. All results were calculated based on relative abundance (TMP). (AC) α-diversity analysis (left) and β-diversity analysis (right) of KEGG, COG, and CAzy functions in each fertilization treatment, respectively. (D) Differential analysis of the abundance of functional genes between PK, NK, and NP fertilization treatments and normal fertilization treatment. An asterisk indicates a significant difference compared to the control NPK (*P<0.05, **P<0.01, ***P<0.001).

[0068] Figure 21 1. Co-occurrence network analysis of rhizosphere bacteria in each fertilization treatment. All results are calculated based on absolute abundance. (A) Co-occurrence networks and related network characteristics of rhizosphere soil bacteria in NPK, PK, NK, and NP fertilization treatments. The color of the dots represents the classification at the phylum level, and the size of the dots represents the degree. (B) Analysis of differences in robustness among different networks. (C) Analysis of differences in mean degree at the phylum level among different networks. Error bars represent standard deviation (SD). An asterisk indicates a significant difference compared to the control NPK (*P<0.05, **P<0.01, ***P<0.001).

[0069] Figure 22 Co-occurrence network analysis of root bacteria in each fertilization treatment. All results are calculated based on absolute abundance. Co-occurrence networks and related network characteristics of root bacteria in NPK, PK, NK, and NP fertilization treatments are shown. The color of the dots represents the classification at the phylum level, and the size of the dots represents the degree.

[0070] Figure 23 Co-occurrence network analysis of rhizosphere bacterial communities. All results are calculated based on absolute abundance. (A) Co-occurrence network of rhizosphere bacterial communities; color represents different modules, and size represents degree. (B) Difference analysis of cumulative absolute abundance among different fertilization treatments in each module. (C) Species composition of bacteria at the phylum level in each module. Error bars represent standard deviation (SD), and asterisks indicate significant differences compared to the control NPK (*P<0.05, **P<0.01, ***P<0.001).

[0071] Figure 24 Evolutionary analysis of purified bacteria. (A) Evolutionary analysis of the 16S sequences of 1011 monoclonal strains obtained through isolation and purification. The clade color represents the phylum-level classification. The strains indicated by the triangles are the strains used in the subsequent synthetic community re-inoculation experiment. The blue triangles represent the control synthetic community, and the yellow and white triangles represent the experimental synthetic communities. (B) Coverage of the isolated strains in the rhizosphere core ASVs.

[0072] Figure 251. Growth phenotypes of soybean after inoculation with synthetic communities. (A) Evolutionary analysis of bacterial 16S sequences in the co-occurrence network of rhizosphere bacterial communities. The bar colors represent different modules, and the pentagrams represent synthetic communities. Blue pentagrams represent the five control group synthetic strains in M1, yellow pentagrams represent the five experimental combination synthetic strains in M2, and white pentagrams represent the other two experimental combination synthetic strains in M2. (B) Growth phenotypes of soybean after 3 weeks of growth under nitrogen-added (left) and nitrogen-free (right) conditions, respectively. The scale bar is 10 cm. (C and D) Phenotypic differences in plant height (left) and dry weight (right) of soybean after inoculation with synthetic communities under nitrogen-added and nitrogen-free conditions, respectively. Error bars represent standard deviation (SD), and lowercase letters represent significant differences in soybean growth phenotypes between different treatments (P<0.05).

[0073] Figure 26 Key findings diagram. (A) Differences in rhizosphere bacterial abundance and diversity across different developmental stages. (B) Differences in soil properties, plant phenotypes, rhizosphere bacterial function, and network characteristics among different fertilization treatments. Compared to normal fertilization treatment, blue upward arrows represent increases, and orange downward arrows represent decreases.

[0074] Figure 27 Bacterial β-diversity based on absolute abundance. All results were calculated based on absolute abundance. (A) Differences in bacterial community β-diversity among all samples (fertilization treatments, ecological niches, and developmental stages). (B) Bray-Curtis distance between rhizosphere and rhizosphere bacteria at each developmental stage. (C) The degree of influence of fertilization on β-diversity of rhizosphere and rhizosphere bacteria at each developmental stage. Lowercase letters indicate significant differences between different growth and development stages (P<0.05).

[0075] Figure 28 Shannon-Wiener curves for bacterial 16S data. Sparse curves of 16S rRNA gene sequencing results based on the Shannon index.

[0076] Figure 29Bacterial diversity based on relative abundance. All results were calculated based on relative abundance. (AC) Shannon diversity indices for non-rhizosphere, rhizosphere, and rhizosphere bacterial communities, respectively. (D) Differences in β-diversity of bacterial communities among all samples (fertilization treatments, ecological niches, and developmental stages). (E) Bray-Curtis distance between rhizosphere and rhizosphere bacteria at each developmental stage. (F) The degree of influence of fertilization on β-diversity of rhizosphere and rhizosphere bacteria at each developmental stage. (G) Relationship between temporal distance and Bray-Curtis distance between different samples in each fertilization treatment in rhizosphere (left) and rhizosphere (right). (H) Relationship between developmental time of each sample and Bray-Curtis distance of PK, NK, and NP samples relative to NPK samples in rhizosphere (left) and rhizosphere (right). An asterisk indicates a significant difference between each fertilization treatment and the control (*P<0.05, **P<0.01, ***P<0.001), and a lowercase letter indicates a significant difference between different developmental stages (P<0.05).

[0077] Figure 30 Accuracy of the random forest model. All results were calculated based on absolute abundance, and all data were analyzed at the family taxonomic level. (A and B) represent the accuracy of the random forest model established based on rhizosphere soil samples with different fertilization treatments and different developmental stages, respectively.

[0078] Figure 31 Co-occurrence networks of rhizosphere and rhizosphere bacteria in each fertilization treatment. All results were analyzed on the MENA website based on absolute abundance. (A and B) represent the co-occurrence networks and related network characteristics of rhizosphere and rhizosphere bacteria in NPK, PK, NK and NP fertilization treatments, respectively. The color of the dots represents the phylum-level classification of bacteria, and the size of the dots represents the degree.

[0079] Figure 32 Verification of the growth-promoting function of the synthetic community. (A) Verification of the IAA production, ACC deaminase activity, inorganic phosphorus solubility, and nitrogen fixation functions of each strain in the synthetic community. (B and C) represent the IAA concentration and ACC deaminase activity produced by each synthetic community, respectively. Detailed Implementation

[0080] Through in-depth research and long-term soil field experiments, the inventors have, for the first time, utilized absolute quantitative methods to explore the dynamic assembly patterns of rhizosphere and root microbiomes throughout the plant's life cycle, and investigated the response characteristics of the rhizosphere microbiome to different fertilization treatments. Combining metagenomics, microbial isolation and culture, and synthetic community methods, the invention revealed the functional adaptation of rhizosphere microorganisms under different fertilization conditions and the mechanisms by which they affect leguminous plant growth. Based on this, the invention provides methods for improving soil properties and soil microbiota in fertilized soils, or for promoting symbiotic nitrogen fixation between rhizobia and leguminous plants; and also provides artificially synthesized microbial communities useful for promoting leguminous plant growth and increasing biomass.

[0081] As used herein, "legumes" refers to legumes with root nodules, including "nodule plants." "Nodule plants" include plants that can form nodules on their roots due to invasion and stimulation by rhizobia.

[0082] As used herein, “nodule plants” also include “nodule-like plants”, which refers to plants with nodule-like or root-nodule-like structures.

[0083] As used in this article, “improvement” refers to improving the properties of soil or plants.

[0084] As used herein, “soil improvement” includes improving the soil properties and soil microbiota of fertilized soil, including but not limited to: reducing soil acidification (increasing soil pH by 0.5 units or 1 unit or more), increasing the number and diameter of root nodules in the soil, and increasing the abundance of rhizobia in the soil.

[0085] As used herein, “fertilized soil” refers to soil used for planting plants that is fertilized or requires fertilization.

[0086] As used herein, “improved plants” includes, but is not limited to: increasing the number of root nodules in legumes, increasing the size of root nodules in legumes, promoting the nitrogen-fixing capacity of root nodules in legumes, promoting the utilization of nutrients by legumes, reducing the demand for nitrogen fertilizer by legumes, increasing the biomass of legumes, and increasing the yield of legumes.

[0087] As used herein, terms such as “improvement of plant traits,” “improved traits,” “improved plant traits,” and “trait improvement” can be used interchangeably and refer to the positive changes in the performance of plants modified by the technical solution of this invention compared to plants before improvement (such as wild-type plants).

[0088] As used in this article, the term "biomass" can refer to the nutrient biomass of a plant. "Biomass" can include metrics such as plant height and plant weight.

[0089] As used herein, the terms “enhancement,” “improvement,” “elevation,” “promotion,” or “enhancement” are interchangeable and should be interpreted in the sense of “enhancement,” “improvement,” “elevation,” “promotion,” or “enhancement” compared to the soil or plants before the regulation or modification, or compared to the control group, for example, 2% to 100%, such as at least 2%, 3%, 4%, 5%, 6%, 7%, 8%, 9%, or 10%, preferably at least 15% or 20%, more preferably 25%, 30%, 40%, 60%, 80%, 90%, or higher.

[0090] Soil nutrients, rhizosphere microorganisms, and plant development are closely interconnected, collectively determining plant health and yield, with microorganisms playing a crucial role throughout the process. This invention investigates the community structure and function of soil microorganisms under different fertilization conditions, the interactions between plants and rhizosphere microorganisms at different developmental stages, and the interaction networks among rhizosphere microorganisms. This has positive implications for land resource protection and increasing food yield, providing new avenues for exploring better agricultural ecosystem fertilization systems and discovering new beneficial plant microorganisms.

[0091] In recent years, the extensive use of fertilizers such as nitrogen, phosphorus, and potassium has been a major reason for increased crop yields. Intensive agricultural production relies excessively on chemical fertilizers and pesticides, which not only causes numerous environmental problems but also severely damages the structure and function of soil microorganisms, affecting the sustainable development of agricultural ecosystems. This invention studies the characteristics of rhizosphere microbial community responses to different soil fertility conditions under long-term different fertilization practices and their relationship with plant growth phenotypes, particularly under nitrogen, phosphorus, or potassium deficiency conditions.

[0092] In the specific implementation method, four different fertilization treatments were selected for research, including normal fertilization (NPK, nitrogen, phosphorus, and potassium fertilizer), no nitrogen fertilizer (PK, phosphorus and potassium fertilizer), no phosphorus fertilizer (NK, nitrogen and potassium fertilizer), and no potassium fertilizer (NP, nitrogen and phosphorus fertilizer). Based on the absolute quantitative method of unit sample mass, a series of microbiome analyses were performed on the collected non-rhizosphere soil, rhizosphere soil, root, and root nodule samples.

[0093] In this specific implementation, the soil physicochemical properties and yields under different fertilization treatments were first analyzed. Then, based on 16S rRNA and rpoB gene amplicon sequencing, the diversity and community composition of rhizosphere and root bacterial communities were analyzed. Furthermore, the dynamic response characteristics of bacterial communities to different fertilization treatments at various growth and development stages of legumes were explored, providing the first systematic description of the quantitative microbiome characteristics of legumes throughout their entire life cycle under different nutrient conditions. Simultaneously, the differences between traditional relative abundance and absolute abundance in characterizing microbial community features were comprehensively compared, and the impact of sequence concentration on amplification and sequencing was evaluated. Next, differential analysis was used to explore the changes in the abundance of core microbial communities under different fertilization treatments. LEfSe analysis and random forest prediction were used to identify marker taxa of microorganisms under different fertilization treatments and at different growth and development stages. Then, the FAPROTAX database was used to predict the function of bacterial communities. Finally, metagenomic assembly was used to study the abundance differences of rhizosphere microbial functional genes among different fertilization treatments. This comprehensive study explored the characteristics and functions of core rhizosphere microorganisms in legumes under different fertilization treatments and at different growth and development stages, revealing the functional adaptation of legume rhizosphere microorganisms under different nutrient conditions. Finally, co-occurrence network analysis was used to analyze the interactions among core rhizosphere bacterial communities under different fertilization treatments. Modular analysis identified core ASVs that may potentially play a growth-promoting role in ecological clusters specifically enriched under nitrogen-free conditions. Then, through the isolation and culture of rhizosphere bacteria, their growth-promoting function was verified under greenhouse conditions, further exploring the potential link between the functional adaptation of rhizosphere microbial communities under different fertilization treatments and the growth and development of legumes.

[0094] Soil physicochemical property analysis showed that, compared with soils treated with normal fertilization (NPK), soils without nitrogen fertilizer (PK) had a 35% lower available nitrogen content and a 1.1-unit higher pH value; soils without phosphorus fertilizer (NK) had a 95% lower available phosphorus content; and soils without potassium fertilizer (NP) had a 61% lower available potassium content. Based on yield statistics from the two most recent consecutive planting seasons, the absence of phosphorus fertilizer significantly impacted legume productivity, reducing yields by 25% and 29% in the two seasons, respectively; the absence of potassium fertilizer slightly reduced legume yield, but the effect was not significant; and there was no significant difference in legume yield between the nitrogen-fed and control groups. This indicates that legumes grown in Northeast China's black soil do not require additional nitrogen fertilizer, while available phosphorus content is the main limiting factor for legume productivity.

[0095] The results of absolute quantitative microbiome analysis showed that the abundance of rhizosphere and root bacteria continuously increased with plant development, and that ecological niche, plant developmental stage, and fertilization treatment all significantly affected the bacterial community structure. Specifically, in the PK treatment, the α-diversity of rhizosphere bacteria was higher than that in the normal fertilization treatment during the vegetative growth stage, and its β-diversity was significantly different from other treatments; while the NK treatment reduced the abundance of rhizosphere bacteria and their turnover rate with plant development, and its α-diversity was higher than that in the normal fertilization treatment during the reproductive growth stage. Proteobacteria and Actinobacteria were dominant in both the rhizosphere and roots, with Actinobacteria being dominant in the early development of leguminous plants, while Proteobacteria were dominant in the later development of leguminous plants. Bacteroides mainly accumulated in the later stages of leguminous plant growth, and its abundance in the NK treatment was significantly lower than that in the control. In addition, long-term lack of nitrogen fertilization significantly increased the abundance of rhizobia in the rhizosphere.

[0096] Furthermore, 573 core ASVs (Amplicon Sequence Variants) were identified in the rhizosphere. These ASVs accounted for only 2.4% of the rhizosphere bacterial species, but contributed 74.6% of the abundance. Differential analysis showed that compared with the normal fertilization treatment, the PK, NK, and NP treatments had 172, 67, and 6 differentially expressed core ASVs, respectively. Functional prediction of the core ASVs indicated that the nitrogen fixation function of bacteria was significantly enhanced in the PK treatment, while the carbon metabolism function of bacteria was reduced and the nitrogen metabolism function was enhanced in the NK treatment. Metagenomic functional analysis revealed significant differences in the functional diversity of rhizosphere bacteria among different fertilization treatments, with the functional diversity of bacteria in the PK treatment being significantly lower than that in the normal fertilization treatment. Simultaneously, the PK treatment showed enhanced nitrogen mineralization capacity and reduced nitrogen reduction and inorganic phosphorus dissolution functions; the NK treatment showed enhanced nitrification, denitrification, inorganic phosphorus dissolution, and phosphorus starvation response functions and reduced nitrogen reduction function; while the NP treatment showed enhanced potassium transport function, indicating that nutrient stress induced functional adaptation of microorganisms to the rhizosphere environment.

[0097] Compared with normal fertilization treatments, both PK and NP treatments significantly increased the complexity and stability of the rhizosphere bacterial co-occurrence network, with rhizobia and sphingomyelin playing important roles in the network. In contrast, NK treatment had no significant effect on bacterial network complexity. Furthermore, five bacteria (Rhodococcus sp., Lysobacter sp., Terrabacter sp., Arthrobacter sp., and Phyllobacterium sp.) with close connections in the rhizosphere network were identified in the PK-specific enriched ecological clusters. Pot experiments showed that, compared with the inoculated control synthetic community (SynCtrl), legumes in nitrogen-treated and nitrogen-free treatments (SynCom5) showed increases in plant height of 15% and 6%, and dry weight of 38% and 18%, respectively. This indicates that under environmental stress, the microbial clusters enriched in the plant rhizosphere can synergistically promote growth and have potential for further application.

[0098] Plant rhizosphere microorganisms have broad application prospects in biofertilizers, biopesticides, promoting nutrient absorption, stress resistance, and yield increase, and are one of the key foundations for sustainable agricultural development. Based on the new findings of this invention, this invention also provides a series of synthetic microbial communities, including SynCom5 or SynCom7. Applying these synthetic microbial communities to legumes can promote their growth and increase their biomass.

[0099] The present invention also provides a composition containing the said synthetic microbial community, comprising: an effective amount of the said synthetic microbial community or a culture thereof. More preferably, the composition further comprises a pesticide-acceptable carrier.

[0100] In this invention, the *Rhodococcus* sp., *Lysobacter* sp., *Terrabacter* sp., *Arthrobacter* sp., and *Phyllobacterium* sp. are preferably also included, and *Bosea* sp. and *Aeromicrobium* sp. are also mentioned. These strains are known in the art, and those skilled in the art can obtain the corresponding strains based on the information provided in this invention.

[0101] The compositions of the present invention (such as pesticide compositions) can be formulated into any suitable dosage form. For example, the forms of the compositions include (but are not limited to): powders, granules, oils, emulsions, wettable powders, microcapsules, and nonwoven fabric microbial agents. Suitable solid diluents for preparing the compositions include (but are not limited to): diatomaceous earth, corn husks, tricalcium phosphate, cork powder, clays such as kaolin, bentonite, or siliceous magnesium, or water-soluble polymers. Solid compositions may contain one or more compatible wetting agents, dispersants, emulsifiers, or pigments, which may also act as diluents in the solid state. Liquid compositions may be in the form of solutions, suspensions, or emulsions, and may be encapsulated in natural or synthetic polymers, and may contain wetting agents, dispersants, or emulsifiers. Such emulsions, suspensions, or solutions may be prepared using aqueous, organic, or water-organic diluents to prepare water-soluble polymers (and mixtures of the above diluents). Furthermore, the diluents may contain, for example, the ionic or nonionic wetting agents, dispersants, or emulsifiers described above, or mixtures thereof. The composition may be in concentrated or diluted form.

[0102] Suitable methods for applying the composition to growing plants include: solid application, foliar spraying, liquid irrigation, etc. A preferred application method is application close to the plant roots.

[0103] Those skilled in the art will recognize that the application of the composition may vary depending on external conditions such as temperature, humidity, the area to be treated, and the plants to be treated. Therefore, the application ratio should be able to vary over a wide range. The frequency of use of the composition of this invention can be selected by farmers, disease control professionals, or other individuals skilled in the art based on the desired effects.

[0104] The composition based on synthetic microbial communities of the present invention is an artificially prepared combination of naturally occurring bacterial strains, which is harmless to other plants and animals and is an environmentally friendly composition.

[0105] Based on an absolute quantitative method, this invention describes for the first time the microbiome assembly patterns and response characteristics of leguminous plants throughout their entire life cycle under different fertilization conditions, and proposes for the first time that rhizosphere-specifically enriched bacterial ecological clusters under low nitrogen conditions can significantly promote the development of leguminous plants. The use of an absolute quantitative analytical method combines absolute abundance with relative abundance, providing a more comprehensive and accurate reflection of the true quantity and differences of microorganisms.

[0106] This invention proposes a new approach to soil fertilization strategies. According to the analysis results of this invention, legumes planted in black soil do not require additional nitrogen fertilizer. Only appropriate amounts of phosphorus and potassium fertilizers are needed to ensure yield. This will help reduce the use of chemical fertilizers and promote the sustainable development of agriculture.

[0107] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Experimental methods in the following embodiments that do not specify specific conditions are generally performed according to conventional conditions such as those described in J. Sambrook et al., Molecular Cloning: A Laboratory Manual, 3rd Edition, Science Press, or according to the manufacturer's recommendations.

[0108] Materials and methods

[0109] plant materials

[0110] Soybean (Glycine max): Cultivar variety "Heinong 84".

[0111] Rotation maize (Zea mays): Cultivar "Tiannong 9".

[0112] Rotation wheat (Triticum aestivum): Cultivar "Longfumai 196".

[0113] In the field trials, all plant materials were grown under natural conditions, and all plots underwent uniform irrigation and weeding throughout the growth cycle. In the synthetic community reinoculation experiment, all plant materials were grown in an artificial climate chamber under conditions of 22-24°C, 16 hours of light, and 8 hours of darkness.

[0114] Microbial materials

[0115] 1) Internal control clonal strains:

[0116] Escherichia coli TOP10 strain.

[0117] 2) SynCtrl control group synthetic community strains: PK85 (Brevundimonas sp.); NK173 (Sediminibacterium sp.); NP149 (Mycobacterium sp.); NP94 (Herbaspirillum sp.); NK151 (Sphingomonas sp.).

[0118] 3) SynCom7 experimental group of strains: PK167 (Rhodococcus sp.); PK216 (Bosea sp.); PK260 (Lysobacter sp.); NPK195 (Terrabacter sp.); NK218 (Arthrobacter sp.); NPK112 (Aeromicrobium sp.); PK237 (Phyllobacterium sp.).

[0119] 4) SynCom5 experimental group of strains: PK167 (Rhodococcus sp.); PK260 (Lysobacter sp.); NPK195 (Terrabacter sp.); NK218 (Arthrobacter sp.); PK237 (Phyllobacterium sp.).

[0120] Experimental design and sample collection

[0121] The Northeast Black Soil Fertility Targeted Experiment began in 1979, with the experimental site located in Minzhu Township, Daowai District, Harbin City, Heilongjiang Province (126°51′E, 45°50′N). The experimental site included 24 different fertilization treatments, employing a randomized block design with three replicates per treatment. Each experimental plot was 36 square meters, and different plots were separated by cement walls (15 cm wide and 1.1 m deep), managed uniformly by the Heilongjiang Academy of Agricultural Sciences. These experimental plots consistently practiced a soybean, corn, and wheat rotation, planting only one crop per year. This invention selected four experimental treatments from the soybean season: normal fertilization (NPK, applying nitrogen, phosphorus, and potassium fertilizer), no nitrogen fertilizer (PK, applying phosphorus and potassium fertilizer), no phosphorus fertilizer (NK, applying nitrogen and potassium fertilizer), and no potassium fertilizer (NP, applying nitrogen and phosphorus fertilizer) (the same fertilization strategy was used for other planting seasons). The nitrogen fertilizer used is urea, the phosphate fertilizer is superphosphate and diammonium phosphate, and the potassium fertilizer is potassium sulfate. The specific application rates are as follows: 150 kg N ha during the wheat and corn seasons. -1 75kg P2O5 ha -1 75kg K2O ha -1 During the soybean season, the yield is 75 kg N ha. -1 150kg P2O5 ha -1 75kg K2O ha -1 Nitrogen, phosphorus, and potassium fertilizers are applied after the autumn crop harvest, with 50% of the nitrogen fertilizer applied during the corn season at the jointing stage.

[0122] Soybeans were planted in 2020, sown on May 11, 2020. Before sowing, three samples of non-rhizosphere soil were collected from each experimental plot as a control. Then, some air-dried non-rhizosphere soil samples were sent to the Institute of Environment and Safety, Wuhan Academy of Agricultural Sciences, for soil physicochemical property testing (including available nitrogen, available phosphorus, available potassium, organic matter, soluble organic carbon, and pH). The soil type is typical black soil; relevant physicochemical properties are shown in Table 1.

[0123] Table 1. Soil physicochemical properties

[0124] NPK-1 164 133 376 28.6 29 5.34 NPK-2 172 143 398 30.2 28.7 5.06 NPK-3 180 159 361 29.4 29.1 5.21 PK-1 109 161 421 28.1 22.6 6.28 PK-2 101 144 403 27.8 24.1 6.33 PK-3 125 146 388 28 22.5 6.39 NK-1 166 7.1 371 29.4 27.5 5.12 NK-2 133 5.4 356 29.6 28.5 5.34 NK-3 147 8.6 348 30.6 27.4 5.3 NP-1 146 104 147 28.1 27.1 5.64 NP-2 166 132 135 29.3 30.5 5.31 NP-3 132 146 156 29.3 28.2 5.47

[0125] Note: AHN is alkaline nitrogen, AP is available phosphorus, AK is available potassium, OM is organic matter, DOC is soluble organic carbon, and pH is the hydrogen ion concentration index.

[0126] Soybeans germinated completely on May 22, 2020. Sampling was conducted at eight stages after germination: 1 day, 4 days, 7 days, 14 days (beginning of nodulation), 28 days (early growth stage), 42 days (initial flowering stage), 60 days (full flowering stage), and 72 days (pod formation stage). At each sampling stage, three plants of similar growth were selected from the corresponding treatment plot, and their roots (5-15 cm) were dug up. The soil attached to the roots was then removed, and the roots, nodules, and soil attached to the rhizosphere were collected in 50 mL sterile centrifuge tubes and brought back to the laboratory for further processing. After soybean maturity, the grain yield of each plot was also measured, and the specific yields are shown in Table 2.

[0127] Table 2. Soybean Production (2017 production refers to the previous season's soybean production, and 2020 production refers to the current season's soybean production).

[0128]

[0129] Sample processing

[0130] The specific steps are as follows:

[0131] 1) Add 25 mL of PBS buffer to a 50 mL sterile tube containing the sample and vortex thoroughly;

[0132] 2) Use an ultrasonic cleaner (40Hz, 1min) to ultrasonically clean the sample to completely separate the rhizosphere soil from the root tissue;

[0133] 3) Transfer the rhizosphere soil suspension to a new 50mL sterile tube (remove as much plant tissue as possible), centrifuge (9000rpm, 5min) and remove the supernatant;

[0134] 4) Resuspend the rhizosphere soil precipitate in 1 mL PBS buffer into a 2 mL centrifuge tube, centrifuge again (12000 rpm, 2 min), and then remove the supernatant as thoroughly as possible, retaining the rhizosphere soil precipitate.

[0135] 5) Dry the rhizosphere soil completely at room temperature using a vacuum centrifuge (1400 rpm, 2 h);

[0136] 6) Homogenize the dried rhizosphere soil using a multi-sample tissue grinder (40Hz, 1min);

[0137] 7) Weigh approximately 100 mg of the dried and homogenized rhizosphere soil sample and store it for later use;

[0138] 8) Then, in a 50mL sterile tube, use PBS buffer to sonicate and shake the roots and root nodules twice.

[0139] 9) Disinfect the surface of roots and root nodules with 5% (w / v) NaClO solution for 3 min;

[0140] 10) Use an equal volume of PBS buffer to sonicate and shake the roots and root nodules three times.

[0141] 11) Weigh approximately 200 mg of root and root nodule respectively, absorb the moisture with sterile absorbent paper, and then dry them completely using a vacuum centrifuge (1400 rpm, 2 h).

[0142] 12) Weigh and record the dried root and root nodule samples, then grind and homogenize them under liquid nitrogen and store them for later use;

[0143] 13) After drying and homogenizing the non-rhizosphere soil using a similar method as described above, weigh approximately 100 mg as a control soil sample and store it for later use;

[0144] After all samples are processed, store them in a -80°C freezer, and compile a sample list and record the dry weight.

[0145] DNA extraction

[0146] Microbial genomic DNA was extracted from all samples using the FastDNA SPIN Kit for Soil according to its instructions.

[0147] Design and addition of internal standards

[0148] Following previous principles and methods (Tkacz et al., 2018; Wang et al., 2020), 13 internal control (synthetic spike-in, SynSpike) genes were designed. Twelve of these were different 16S SynSpike genes, designed based on the V5-V7 variable regions of the bacterial 16S rRNA gene. The remaining gene was an rpoB SynSpike, designed based on the rpoB gene of symbiotic rhizobia. These internal control genes were synthesized and fused with plasmid pUC57, ultimately transformed into *E. coli* TOP10 for propagation and cloning.

[0149] All internal standard plasmids were extracted using the TIANpure Midi Plasmid Kit according to its instructions. After extraction, the concentration and purity of the DNA were determined using NanoDrop. Then, all internal standards were combined at different concentrations to form SynSpike for Soil and Nodule, which was diluted 10-fold to obtain SynSpike for Root. The combinations and detailed concentrations of the internal standards are shown in Table 3. Finally, 30 μL of microbial DNA and 10 μL of the corresponding SynSpike DNA were transferred to new centrifuge tubes, thoroughly mixed, and stored at -80°C for subsequent DNA amplification, library construction, and sequencing. Finally, the absolute abundance of microorganisms per unit mass of sample was calculated using the Qiime2 workflow based on the copy number of the SynSpike plasmid. Figure 1 ).

[0150] Table 3. Length and dosage of each internal standard plasmid

[0151]

[0152] Note: Synthetic DNA Length refers to the length of each internal standard DNA, and Plasmid Length refers to the length of each internal standard DNA after fusion into the pUC19 plasmid backbone (2699 bp). Soil and Nodule refer to the amount of each internal standard plasmid added in the soil, rhizosphere soil, and root nodule samples, and Root refers to the amount of each internal standard plasmid added in the root sample.

[0153] DNA amplification

[0154] The V5-V7 variable region of the bacterial 16S rRNA gene was amplified using 799F (5'-AAC MGG ATT AGA TAC CCK G-3') and 1193R (5'-ACG TCA TCC CCACCT TCC-3') (Bai et al., 2015); the rpoB gene of symbiotic rhizobia was amplified using rpoB1479F (5'-GAT CGA RAC GCC GGA AGG-3') and rpoB1831R (5'-TGC ATG TTC GAR CCCAT-3') (Wang et al., 2020). 16S rRNA gene amplification included all soil and root samples, while rpoB gene amplification included all soil and root nodule samples.

[0155] After amplification, the quality of DNA amplification was detected by 1% agarose gel electrophoresis. The amplified DNA was stored in a -80°C freezer for subsequent library construction and sequencing.

[0156] High-throughput sequencing

[0157] 1) Amplicon sequencing

[0158] The amplified samples were library constructed and sequenced. Amplicon sequencing was performed using the MGISEQ platform with single-end 400bp reads (SE400). The sequencing sample list and raw data have been uploaded to the NCBI SRA database (project number: PRJNA922226; 16S rRNA amplicon data sample numbers: SAMN32647504-SAMN32648110; rpoB amplicon data sample numbers: SAMN32650334-SAMN32650836).

[0159] 2) Metagenomic sequencing

[0160] Metagenomic sequencing was performed on rhizosphere microbial genomic DNA collected at three representative time points (1 day, 42 days, and 72 days after soybean germination). To save costs, 10 μL of DNA from three replicate samples of the same region were mixed to form one sample, and then the metagenomic sequencing sample DNA was used for library construction and sequencing. Metagenomic sequencing was performed using the DNBSEQ platform with paired-end 100 bp reads (PE100). The sequencing sample list and raw data have been uploaded to the NCBI SRA database (Project No.: PRJNA922226; Sample No.: SAMN32727371-SAMN32727406).

[0161] Isolation and purification of bacteria

[0162] 1) Culture medium: LB and R2A agar medium.

[0163] 2) Pouring the culture dish. After the culture medium is sterilized, accurately measure 20 mL of agar medium into a 10 cm diameter round dish using a measuring cup in a sterile operating table. After the culture medium cools and solidifies, cover it for later use.

[0164] 3) Preparation of soil dilution. Under aseptic conditions, weigh 3g of fresh rhizosphere soil (including four treatments: NPK, PK, NK, and NP). Place the weighed rhizosphere soil into a 50mL sterile tube, add 30mL of PBS buffer, and then sonicate and shake for 30min to thoroughly mix the sample and solution. The resulting liquid is 10... -1 Dilute the soil suspension in a gradient. Pipette 1 mL of soil suspension into a sterile tube containing 9 mL of PBS buffer, mix thoroughly, and the resulting solution is 10⁻⁶. -2 Dilute the soil suspension in the gradient. Continue with this step to prepare 10... -3 10 -4 10 -5 10 -6Soil dilution solutions of different concentrations.

[0165] 4) Preparation of soil dilutions for dried samples. To better isolate and purify Bacillus, 3g of fresh rhizosphere soil was weighed again and stored in a desiccant container. After two weeks, the soil was removed and placed in a sterile centrifuge tube containing 30mL of PBS buffer. After thorough shaking and mixing, the mixture was heated at 80℃ for 30min. Finally, the soil gradient dilutions for dried samples were obtained following step 3.

[0166] 5) Spreading. First, make the corresponding marks in the petri dishes, then take 50 μL of soil dilution into agar medium (LB and R2A) respectively, then spread it evenly with a sterile spreader, and finally seal with sealing film.

[0167] 6) Incubation. Invert the coated petri dishes and incubate them in a 30℃ incubator for 48 hours, observing the growth of the colonies during this period.

[0168] 7) Pick single colonies. Use an inoculation loop to pick a single colony that is growing well on the medium, streak it on a fresh medium of the same type, and then invert it in a 30°C incubator to continue culturing for 48 hours for subsequent identification.

[0169] 8) Culture preservation. Pick a single colony from the streak plate into a culture preservation tube containing 0.4 mL of sterile 50% glycerol and 0.6 mL of LB liquid medium, label and mix well, and store in a -80°C freezer.

[0170] Bacterial identification

[0171] The purified bacteria were identified using universal primers 27F (5'-AGA GTT TGA TCC TGG CTC AG-3') and 1541R (5'-AAG GAG GTG ATC CAG CCG CA-3') for the bacterial 16S rRNA gene, and PCR reactions were performed.

[0172] After the PCR reaction, 4 μL of PCR product was mixed with 2 μL of loading buffer and then detected by 1% agarose gel electrophoresis (120V, 30 min). PCR amplification products with normal bands (20 μL of PCR stock solution) were selected for Sanger sequencing. The sequencing results were aligned to the 16S rRNA gene database Greengenes 13_8 99% OTUs (McDonald et al., 2012) using BLAST version 2.12.0 to obtain the final annotation information.

[0173] Amplicon analysis

[0174] The amplicon sequencing results were analyzed using QIIME 2 (Bolyen et al., 2019) version 2021.11, including the following:

[0175] 1) FastQC (https: / / www.bioinformatics.babraham.ac.uk / projects / fastqc) version 0.11.9 and MultiQC (Ewels et al., 2016) version 1.11 were used to perform quality control on the sequencing results and remove low-quality samples that did not meet the requirements;

[0176] 2) Use the q2-demux plugin to perform quality filtering on the raw sequence data, and then use the q2-cutadapt (Martin, 2011) plugin to remove the amplification primer sequences;

[0177] 3) Use the q2-DADA2 (Callahan et al., 2016) plugin to denoise the sequence, and truncate the sequence at a specified length according to the results of the previous quality filtering step to remove low-quality bases in the sequence (305 bp for 16S sequence and 304 bp for rpoB sequence), and finally obtain amplicon sequence variants (ASVs);

[0178] 4) The q2-feature-classifier (Bokulich et al., 2018) plugin was used to classify ASVs. For the reference sequence of bacterial 16S rRNA Greengenes 13_8 99% OTUs (McDonald et al., 2012), the classify-sklearn naive bayes classifier was used to classify 16S ASVs; for the reference sequence of the rhizobium rpoB gene from Wang et al. (Wang et al., 2020), the classify-consensus-blast classifier was used to classify rpoB ASVs (identity set to 0.977).

[0179] 5) Based on the classification and annotation results, the q2-taxa plugin was used to filter the feature tables and representative sequences according to the following rules: For 16S data, only data annotated to the phylum level were retained, and data annotated as Unassigned, Archaea, Chloroplast, mitochoondria, or SynSpike were filtered out. Furthermore, because root tissue samples may contain small, incompletely removed root nodules, which can lead to excessively high rhizobium abundance and thus obscure the true bacterial community composition, 16S root tissue sample data also required additional filtering to remove data annotated as Bradyrhizobiaceae. For rpoB data, data annotated as Unassigned or SynSpike were filtered out, retaining only data annotated to the species level and belonging to nitrogen-fixing rhizobia.

[0180] 6) Annotate the filtered representative sequence again as in step 4.

[0181] 7) Align ASVs using the q2-phylogeny plugin via MAFFT (Katoh et al., 2002), and then construct a phylogenetic tree using FastTree (Price et al., 2010);

[0182] 8) After flattening the 16S sample into 1030 sequences using q2-diversity, calculate its α-diversity index (Shannon index, etc.) and β-diversity index (Bray-Curtis dissimilarity, etc.).

[0183] Metagenome assembly

[0184] The metagenomic data were assembled according to the Easy Metagenome Metagenomic Analysis Workflow published by Liu et al. (Liu et al., 2021c), including:

[0185] 1) Use FastQC and MultiQC to perform quality control on sequencing results and remove low-quality samples that do not meet the requirements;

[0186] 2) Use KneadData (https: / / huttenhower.sph.harvard.edu / kneaddata) version 0.10.0 to filter low-quality data and remove sequences from the host plant soybean;

[0187] 3) Single-sample assembly was performed on the high-quality data after filtering 36 samples using the default parameters (k-mers = 21, 33, 55, 77, 99, 127) of MEGAHIT (Li et al., 2015) version 1.2.9;

[0188] 4) The quality of genome assembly was assessed using QUAST (Gurevich et al., 2013) version 5.0.2;

[0189] 5) Use Bowtie 2 (Langmead & Salzberg, 2012) version 2.4.4 to align all Reads back to the assembled Contig, and then calculate the sequence coverage for each Contig;

[0190] 6) Use SeqKit (Shen et al., 2016) version 2.1.0 to filter sequences longer than 1000bp as assembly results;

[0191] 7) Use Prodigal (Hyatt et al., 2010) version 2.6.3 to predict the open reading frame for each contig and extract the complete gene;

[0192] 8) Use CD-HIT (Li & Godzik, 2006) version 4.8.1 to cluster the predicted gene sequences of all samples and construct a non-redundant gene set;

[0193] 9) Salmon (Patro et al., 2017) version 1.6.0 was used to quantify the assembled non-redundant gene set to obtain raw Reads data and standardized TPM (Transcripts Per Kilobase Million) data;

[0194] 10) Use eggNOG-mapper (Cantalapiedra et al., 2021) version 2.1.6 to annotate non-redundant genes into databases such as GO, KEGG and COG, and use DIAMOND (Buchfink et al., 2021) version 2.0.13 to annotate the database into the CAZy database (Drula et al., 2022).

[0195] Metagenomics binning

[0196] Metagenomic data were binned and purified using MetaWRAP (Uritskiy et al., 2018) version 1.3.2. The specific steps are briefly described below:

[0197] 1) Use the two algorithms MaxBin 2 (Wu et al., 2016) and MetaBAT 2 (Kang et al., 2019) in the binning module to bin contigs longer than 2000bp;

[0198] 2) Use CheckM (Parks et al., 2015) version 1.0.12 in the bin_refinement module to purify the obtained bins, retaining bins with integrity greater than or equal to 70% and contamination level less than 10%;

[0199] 3) Merge all sample bins into the same directory, and then use dRep (Olm et al., 2017) version 3.2.2 to remove redundancy from all bins at the species level based on the average nucleotide similarity (ANI) (detailed parameters: -sa0.95-nc 0.10-comp 70-con 10), finally obtaining a non-redundant genome set;

[0200] 4) Use GTDB-Tk (Chaumeil et al., 2019) version 1.7.0 to annotate species and construct phylogenetic trees for non-redundant genome sets;

[0201] 5) Use Prokka (Seemann, 2014) version 1.13 to annotate the genes in each non-redundant genome set;

[0202] 6) Use Salmon to calculate the relative abundance of each bin in the sample.

[0203] Random Forest Analysis

[0204] First, ASVs with an average relative abundance of less than 0.1% in each data set are filtered out. Then, two samples from each block in each growth stage and fertilization treatment are used as the training set for the random forest model, and the remaining sample in each block is used as the test set. The randomForest function in R language version 4.1.0 (https: / / CRAN.R-project.org / package=randomForest) version 4.7 is used to train the random forest model. Next, the predict function is used to test the test set data. Finally, the rcfv function is used to perform 10× cross-validation on the error rate.

[0205] Co-occurrence network analysis

[0206] Co-occurrence networks were used to reveal potential interactions among microbial communities in various treatments, and differences in relevant network properties among different treatments were compared. The specific steps are as follows:

[0207] 1) Select ASVs with an average relative abundance of not less than 0.1% at each developmental time point for further network analysis;

[0208] 2) A total of approximately 576 samples (containing 2 niches, 4 treatments, 8 time points, and 9 replicates) of ASV absolute abundance data were used for network construction;

[0209] 3) Calculate the pairwise Spearman correlation matrix using the corr.test function in version 2.2.5 of the psych package (https: / / CRAN.R-project.org / package=psych) in R language version 4.1.0, and correct the p-value using the FDR (False Discovery Rate) method;

[0210] 4) Based on the correlation results, select datasets with P-values ​​less than 0.05 and correlations greater than 0.8 or less than -0.8 for network construction;

[0211] 5) Use version 1.2.6 of the igraph (Csardi & Nepusz, 2006) function package in R to calculate the relevant properties of networks and nodes in each network;

[0212] 6) Calculate the robustness of each network according to the method described by Yuan et al. (Yuan et al., 2021);

[0213] 7) Finally, Gephi (Bastian et al., 2009) version 0.9.5 was used to visualize the co-occurrence network relationships among microbial communities in each treatment. The size of the dots represents the degree, and the color of the dots represents the phylum-level classification or different modules of bacteria.

[0214] Synthetic community backlink

[0215] Based on the results from the rhizosphere microbial community network, ASV sequences from the three modules were aligned to previously isolated bacterial sequences using BLAST (identity > 95%). Seven ASVs were identified in module 2, which was specifically enriched by PK treatment, and these seven ASVs were then used as experimental synthetic communities. In addition, five ASVs were screened in module 1 as control synthetic communities (including strains: Brevundimonas sp., Sediminibacterium sp., Mycobacterium sp., Herbaspirillum sp., and Sphingomonas sp.). The synthetic community re-inoculation experiment was conducted in non-nutrient soil (vermiculite) and included two nutrient treatments (nitrogen-treated and non-nitrogen-treated) and four inoculation treatments (SynCom7, an experimental synthetic community composed of seven bacteria; SynCom5, an experimental synthetic community composed of five bacteria; SynCtrl, a control synthetic community composed of five control bacteria; and a sterile water control). The specific procedures are as follows:

[0216] 1) After sterilizing the vermiculite (121℃, 20min), place it in a 10cm×10cm square plant culture box, and then pour an appropriate amount of sterile water into the bottom tray to make all the soil evenly wet.

[0217] 2) Disinfect soybean seeds with 5% (w / v) NaClO solution for 5 minutes, then rinse them 5 times with sterile water before sowing;

[0218] 3) Sow 4 soybean seeds in each small square box, then cover the surface with a thin layer of vermiculite and wait for them to germinate;

[0219] 4) After the seeds have fully germinated, cut off the plants that are growing poorly or too well so that all seedlings are of the same size. Then take a picture to record the results, and wait for inoculation to form a community.

[0220] 5) Streak the selected strains stored at -80℃ onto solid LB medium and incubate upside down at 30℃ for 48 hours;

[0221] 6) Pick a single colony and transfer it to a 50 mL sterile centrifuge tube containing 20 mL of liquid LB medium for expansion culture (30℃, 48 h);

[0222] 7) After centrifugation (9000 rpm, 2 min), collect the bacterial sludge, then resuspend it in 30 mL of sterile water, and then measure the OD value using a spectrophotometer (OD600).

[0223] 8) Prepare 100 mL of mixed bacterial solution for each bacterium with a final concentration of OD = 0.1 according to the required synthetic community;

[0224] 10) Inoculate the roots of each seedling with 1 mL of the corresponding inoculum according to the design;

[0225] 11) Seven days after the first vaccination, administer a booster dose using the same method;

[0226] 12) During the entire experiment, watering was done with the corresponding nitrogen-deficient nutrient solution (-N MFP, without NH4NO3) and nitrogen-containing nutrient solution (MFP); the MFP formula (1L) was: 132.3mg CaCl2·2H2O, 123.2mg MgSO4·7H2O, 95.3mg KH2PO4, 113.6mg Na2HPO4·12H2O, 4.9mg FeC6H5O7, 40mg NH4NO3, 100μg MnCl2·2H2O, 100μg CuSO4·5H2O, 100μg ZnCl2, 100μg H3BO4, 100μg Na2MoO4.

[0227] 13) After the plants have grown for 3 weeks, take photos to record their growth and measure their phenotypic characteristics such as plant height and dry weight.

[0228] Data statistics

[0229] All statistical analyses were performed using R version 4.1.0 or Linux (CentOS version 7.6). For all statistical tests, p < 0.05 was considered significant, and the significance of each figure was specified in the figure caption.

[0230] Statistical analysis of soil physicochemical properties and soybean yield and root nodule phenotype

[0231] All data were first tested for normality using the Shapiro-Wilk test (Shapiro & Wilk, 1965) and for homogeneity of variance using the Bartlett test. For normally distributed data with homogeneous variance, one-way ANOVA was used, followed by Dunnett's post-hoc test for multiple comparisons. All non-normally distributed or heterogeneous variance data were tested using the Kruskal-Wallis rank-sum test, followed by Dunn's post-hoc test for multiple comparisons.

[0232] Statistical analysis of Alpha diversity

[0233] A three-way ANOVA was performed using the `lmer` function from version 1.1 of the `lme4` package (https: / / CRAN.R-project.org / package=lme4) to construct a linear mixture model. The effects of niche, fertilization treatment, developmental stage, and the interactions among these factors on bacterial α-diversity were statistically analyzed. The Kruskal-Wallis rank-sum test was used to analyze differences in α-diversity between different developmental stages and between different fertilization treatments, and Dunn's post-hoc test was used for multiple comparisons. Differences in α-diversity between different developmental stages are represented by lowercase letters, and differences between different treatments within the same developmental stage are represented by asterisks (compared only to the normal fertilization treatment).

[0234] Statistical analysis of Beta diversity

[0235] Permuted multivariate analysis of variance (PERMANOVA) was performed on bacterial β-diversity using the adonis2 function in vegan (https: / / CRAN.R-project.org / package=vegan) version 2.5. The explanatory power and significance of niche, fertilization treatment, developmental stage, and the interactions among these factors on bacterial β-diversity were statistically analyzed. Multiple comparisons were performed using the Kruskal-Wallis rank-sum test and the Dunn post-hoc test to analyze the differences in Bray-Curtis distance between rhizosphere and root samples at different developmental stages. Univariate regression analysis was used to examine the relationship between soybean growth time distance and Bray-Curtis distance among samples under different fertilization treatments, as well as the relationship between developmental stage and Bray-Curtis distance of PK, NK, and NP treatment samples relative to the normal fertilization treatment samples.

[0236] Analysis of differences in species abundance among samples

[0237] The relative and absolute abundance of species among all samples were analyzed using the same method. Abundance differences among different developmental stages and different fertilization treatments were analyzed using the Kruskal-Wallis rank-sum test, and multiple comparisons were performed using the Dunn post-hoc test.

[0238] LEfSe absolute abundance difference analysis

[0239] The identification of microbial marker taxa was primarily conducted using LEfSe software (version 1.0.8) to perform differential analysis of the absolute abundance of each sample (ASVs with an average relative abundance of less than 0.1% in each group were filtered out beforehand). First, a nonparametric Kruskal-Wallis rank-sum test was performed on each group within each taxonomic attribute. Then, Wilcoxon rank-sum tests were performed on each pair of species showing significant differences from the previous step to detect inter-group differences. Finally, a linear discriminant model was established to assess the influence of species with significant differences between different groups (LDA score).

[0240] Differences in the relative abundance of bacterial ASV functional genes

[0241] Differences in the relative abundance of bacterial ASV functional genes among different growth and development stages were analyzed using a Kruskal-Wallis test with a Dunn post-hoc test for multiple comparisons. For the analysis of differences in the relative abundance of functional genes among different fertilization treatments, the average relative abundance of bacterial ASV functional genes for each treatment and each growth stage was first calculated. Then, a pairwise test using the Wilcoxon rank-sum test was performed to examine the differences between each treatment and the normal fertilization treatment.

[0242] Analysis of differences in absolute abundance of core ASV in the rhizosphere among different fertilization treatments

[0243] First, the average absolute abundance of bacterial core ASVs in each treatment and at each growth stage was calculated. Then, the Wilcoxon rank-sum test was used to perform a pairwise test on the abundance of ASVs at each developmental stage to detect the differences between each treatment and the normal fertilization treatment. Finally, ASVs whose abundance in all developmental stages of each fertilization treatment was 1.5 times higher or lower than the control were considered as the differential ASVs between the corresponding fertilization treatments.

[0244] Difference Analysis of Basic Characteristics of Co-occurrence Networks

[0245] The robustness of each network, the difference in the mean degree of nodes at the target level, and the difference in the cumulative absolute abundance of samples in each module of the rhizosphere network were all analyzed using the Kruskal-Wallis test with Dunn post-hoc test to perform multiple comparison analysis, comparing the differences between each treatment network or module and the normal fertilization treatment network or module.

[0246] Metagenomic functional gene differential analysis

[0247] The α-diversity analysis of metagenomic functions among fertilization treatments was performed using the Kruskal-Wallis test with Dunn post-hoc test for multiple comparisons; the β-diversity analysis used canonical correspondence analysis (CCA) to calculate the explanatory power of fertilization treatments for β-diversity, and then ANOVA-like permutation tests were used to calculate significance.

[0248] Phenotypic analysis of synthetic community reinoculation experiments

[0249] All data were first tested for normality and homogeneity of variance using the Shapiro-Wilk test and Bartlett test. If the data were normally distributed and homogeneous in variance, ANOVA one-way ANOVA was used, followed by LSD post-hoc test for multiple comparisons. Otherwise, Kruskal-Wallis rank-sum test was used, followed by Dunn post-hoc test for multiple comparisons.

[0250] Example 1: Soil physicochemical properties and soybean growth phenotype

[0251] The experiment was conducted in a black soil experimental field in Harbin, Heilongjiang Province, and included four different fertilization treatments: normal fertilization (NPK, nitrogen, phosphorus, and potassium fertilizer), no nitrogen fertilizer (PK, phosphorus and potassium fertilizer), no phosphorus fertilizer (NK, nitrogen and potassium fertilizer), and no potassium fertilizer (NP, nitrogen and phosphorus fertilizer). Each treatment was repeated in triplicate, and a randomized block design was used. Figure 2 A). Samples were collected at eight time points after soybean germination: 1 day, 4 days, 7 days, 14 days (beginning of nodulation), 28 days (early growth stage), 42 days (initial flowering stage), 60 days (full flowering stage), and 72 days (pod formation stage). Figure 2 B) Three plants were collected from each plot, resulting in nine replicates. Amplicon analysis (including 16S rRNA and rpoB genes) based on relative and absolute abundance quantification, as well as metagenomic analysis, were then performed on the collected non-rhizosphere soil (36 samples), rhizosphere soil (287 samples), roots (284 samples), and root nodules (180 samples). Figure 2 C).

[0252] The physicochemical properties of soils under different fertilization conditions (including long-term lack of N, P, and K fertilizers) were analyzed, along with soybean yield and root nodule phenotype.

[0253] 1. Soil physicochemical properties

[0254] Non-rhizosphere soil samples were collected before sowing, and their physicochemical properties were measured. The contents of available nitrogen, available phosphorus, and available potassium in the normal fertilization treatment were 172 mg / kg.-1 145mg kg -1 and 378mg kg -1 Compared with the normal fertilization treatment, the alkaline nitrogen content in the PK treatment was reduced by 35%, to 112 mg / kg. -1 ( Figure 3 A) The available phosphorus in NK treatment decreased by 95% to 7 mg / kg. -1 ( Figure 3 B); The available potassium content in the NP treatment decreased by 61%, to 146 mg / kg. -1 ( Figure 3 C), indicating that the absence of chemical fertilizers during long-term cultivation reduced the content of corresponding readily available nutrients in the soil. However, the organic matter content did not change significantly in any of the treatments. Figure 3 D), but compared with the control, the content of dissolved organic carbon (DOC) in the PK treatment was significantly reduced. Figure 3 E). In addition, it was found that the soil pH in the PK treatment was 1.1 units higher than that in the normal fertilization treatment. Figure 3 F) indicates that long-term nitrogen application will cause soil acidification.

[0255] 2. Soybean yield and root nodule phenotype

[0256] After soybeans matured, the grain yield was statistically analyzed. Based on yield results from two consecutive soybean growing seasons, both the NP and NK treatments showed a decreasing trend in soybean yield compared to normal fertilization, with the NK treatment showing the most significant reduction, decreasing by 25% and 29% in 2017 and 2020, respectively. There was no significant difference in soybean yield between the PK treatment and the normal fertilization treatment. Figure 4 A and Figure 4 B). Furthermore, since rhizobia can form symbiotic nodules with soybeans to fix nitrogen, the nodule phenotype at the initial flowering stage (D42) of soybeans was also statistically analyzed. Similar to the yield results, the number and diameter of nodules in the NK treatment were significantly lower than those in the normal fertilization treatment, suggesting that low phosphorus conditions inhibit the symbiotic nitrogen fixation process in soybean nodules and may lead to a decrease in soybean yield. However, in the PK treatment, the number and diameter of nodules were significantly higher than the control, while the NP treatment showed no significant difference from the control. Figure 4 C and Figure 4 D) indicates that in the black soil of Northeast China, low nitrogen promotes nitrogen fixation in soybeans through symbiotic interaction with rhizobia, and has no significant impact on soybean yield.

[0257] Example 2: Dynamic response characteristics of bacterial communities during plant development

[0258] Based on amplicon sequencing, the dynamic response characteristics of rhizosphere and root bacterial communities under different fertilization treatments at various growth stages of soybean were investigated. The diversity and composition of rhizosphere and root bacterial communities were analyzed, and the similarities and differences between traditional relative abundance and absolute abundance in characterizing microbial community features were compared.

[0259] 1. Calculation of absolute abundance of bacterial communities

[0260] Conventional amplicon sequencing, when analyzing the microbial composition of a sample, only considers the relative abundance (i.e., proportion) of microorganisms. Therefore, it can only describe the community composition and diversity of microorganisms, but cannot characterize the true quantity of each microorganism in the sample or the true differences between samples, leading to erroneous interpretations of the actual changes in the microbial community between samples. To better characterize the specific quantity of the microbial community and obtain the true changes in the community between samples, the inventors used a previous method based on adding artificially synthesized spike-in plasmids to the sample DNA (Wang et al., 2020) to quantitatively detect the copy number of bacterial 16S rRNA genes per unit mass of sample. However, because only a single spike-in plasmid was added to the sample DNA previously, it was impossible to determine whether the concentration of the internal standard plasmid affected the amplification efficiency. Therefore, based on this, the inventors improved the previous method by adding 12 16S spike-in plasmids (16S SynSpike) to the sample DNA to test the effect of DNA concentration on amplification efficiency.

[0261] The amount of 16S SynSpike added to the DNA samples increased in a gradient (Table 3). Univariate regression analysis between the amount of 16S SynSpike added and the number of 16S SynSpike sequences in the sequencing results revealed that the relationship between the amount of SynSpike added and the number of sequences was fitted to a straight line passing through the origin in non-rhizosphere soil, rhizosphere soil, and root samples. This indicates that the number of SynSpike sequences in the sequencing results is directly proportional to the amount added, and that the sequence concentration has no significant effect on the amplification efficiency of the spike-in plasmid. Figure 5 Based on this result, the ratio between the total number of 16S SynSpike sequences added and the total number of sequences annotated as 16S SynSpike in each sample was used as the basis for calculating the absolute abundance of each ASV. Finally, the absolute abundance of each ASV was estimated by calculating the copy number of the internal standard plasmid.

[0262] 2. Bacterial community analysis

[0263] High-throughput sequencing using 16S amplicon sequencing yielded 12,349-801,472 sequences per root sample and 198,300-1,213,079 sequences per rhizosphere soil sample. After filtering based on taxonomic annotation, 1,030-502,193 sequences remained for the root samples and 96,164-577,954 sequences remained for the rhizosphere soil samples.

[0264] First, bacterial community diversity was analyzed based on absolute abundance. The Shannon diversity index results indicated that long-term different fertilization treatments had no significant effect on the α-diversity of bacteria in non-rhizosphere soil. Figure 6 A). Three-way ANOVA revealed that ecological niche, plant developmental stage, and different fertilization treatments significantly affected bacterial α-diversity (P<0.001), and there were significant interactions among these three factors (Table 4). For example, regarding ecological niche, bacterial α-diversity was significantly higher in the rhizosphere (6.1) than in the root zone (4.5). Throughout the growth cycle, rhizosphere bacterial α-diversity was highest 4 days after seed germination, and gradually decreased with plant growth, reaching its lowest levels at full bloom (D60) and pod formation (D72). In contrast, bacterial α-diversity remained relatively stable in the NK treatment. Figure 6 B). Furthermore, rhizobium exhibits the highest α-diversity initially, then declines, reaching its lowest point 14 days after seed germination, before rising again and remaining relatively stable. Figure 6 C) This may be because soybean roots accumulate a large number of rhizobia at the D14 stage (when nodulation begins), thus reducing the α-diversity of other bacteria. Among different fertilization treatments, the α-diversity of rhizosphere bacteria in the PK treatment was significantly higher than that in the early vegetative growth stages (D1, D4, and D14) than in the normal fertilization treatment, while the α-diversity of rhizosphere bacteria in the NK treatment was significantly higher than that in the later reproductive growth stages (D42, D60, and D72). Figure 6 B). Within the roots, except for a significant decrease in bacterial α-diversity in the NP treatment compared to the control at 72 days after seed germination, there were no significant changes in α-diversity among different fertilization treatments at other growth stages. Figure 6 C).

[0265] Table 4. Three-way ANOVA of bacterial α-diversity based on absolute abundance (the results are based on absolute abundance calculations)

[0266] Compartment 2385.3 1.9E-193 Stage 39.7 1.9E-44 Treatment 13.6 1.6E-08 Compartment × Stage 22 5.0E-26 Compartment × Treatment 8.2 2.3E-05 Stage×Treatment 5.7 5.4E-14 Compartment × Stage × Treatment 3 1.3E-05

[0267] PCoA (Principal Coordinates Analysis) based on Bray-Curtis distance for bacterial β-diversity showed significant differentiation between the bacterial community and the rhizosphere soil and roots. Figure 27A). PERMANOVA results indicate that niche (R) 2 =0.241, P<0.001) is the main driver of bacterial β-diversity, followed by plant developmental stage (R 2 =0.114, P<0.001) and fertilization treatment (R 2 =0.040, P<0.001), and the interactions among these factors also significantly affected bacterial β-diversity, especially the relationship between ecological niche and plant developmental stage (R = 0.040, P<0.001). 2 =0.103, P<0.001)(Table 5).

[0268] Table 5. Contribution of each factor to the difference in bacterial β-diversity based on absolute abundance (the results are calculated based on absolute abundance).

[0269] Compartment 0.241 312.6 0.001 Stage 0.114 21.0 0.001 Treatment 0.040 17.4 0.001 Compartment × Stage 0.103 19.1 0.001 Compartment × Treatment 0.032 13.9 0.001 Stage×Treatment 0.043 2.6 0.001 Compartment × Stage × Treatment 0.037 2.3 0.001

[0270] The Bray-Curtis distance between rhizosphere and root bacteria gradually decreases as the plant develops. Figure 27 A and Figure 27 B), and in all stages of plant development, fertilization had a significantly greater impact on rhizosphere bacterial β-diversity than root bacterial β-diversity (B). Figure 27 C). In non-rhizosphere soil, fertilization treatments explained 58.6% of the differences in bacterial β-diversity, with PK treatment showing a significant difference in bacterial β-diversity compared to the control. Figure 6 D). In the rhizosphere soil, except for the NK treatment, the bacterial β-diversity of each fertilization treatment showed significant differentiation over time (R). 2 =0.271), while the bacterial community structure of the PK treatment was significantly different from that of the control (R = 0.271). 2 =0.199)( Figure 6 E). Similarly, the β-diversity of root bacteria also differentiates over time (R). 2 =0.293), but the fertilization treatment had a relatively small effect on bacterial β-diversity (R = 0.293). 2 =0.042)( Figure 6 F). Time decay patterns indicated that NK treatment reduced the regression slope of rhizosphere bacteria from 0.0043 in the normal fertilization treatment to 0.0014, but there was no difference within the roots. Figure 6 (G and Table 6). Throughout plant development, the Bray-Curtis distance of rhizosphere bacteria remained relatively stable between the PK and normal fertilization treatments (Slope = 0.0005). The differences between the normal fertilization treatment and the NK (Slope = 0.0040) and NP (Slope = 0.0028) treatments increased significantly over time, while the Bray-Curtis distance between each root treatment and the control varied relatively little over time. Figure 6H and Table 7).

[0271] Table 6. Relationship between temporal distance and Bray-Curtis distance based on absolute abundance ( Figure 6 (Statistical results based on absolute abundance)

[0272]

[0273] Table 7. Relationship between development time and beta distance based on absolute abundance between treatments ( Figure 6 (H-statistical results, calculated based on absolute abundance)

[0274]

[0275] To compare bacterial diversity results calculated based on relative abundance and absolute abundance, bacterial diversity based on traditional relative abundance was also analyzed. In the relative abundance-based analysis, to reduce the impact of sequencing depth between different samples and to ensure comparability between root and rhizosphere samples, 1030 sequences were extracted from each sample for downstream analysis. Sparse curve results based on the Shannon index showed that this sampling depth effectively characterizes the Shannon index in each sample. Figure 28 The calculations based on this sampling depth revealed that the bacterial α-diversity in each ecological niche was very similar to the results calculated based on absolute abundance. Figure 29 The results also showed that niche, developmental stage and fertilization treatment could significantly affect the α-diversity of bacterial communities (P<0.001) (Table 8).

[0276] Table 8. Three-way ANOVA of bacterial α-diversity based on relative abundance (calculated based on relative abundance)

[0277] Compartment 1422.0 5.8E-149 Stage 36.6 2.2E-41 Treatment 12.0 1.4E-07 Compartment × Stage 18.1 1.6E-21 Compartment × Treatment 7.0 1.3E-04 Stage×Treatment 5.0 6.5E-12 Compartment × Stage × Treatment 2.6 1.5E-04

[0278] In β-diversity analysis, the same was found in niche (R 2 =0.136, P<0.001) and plant development stage (R 2 =0.133, P<0.001) is the main driver of bacterial β-diversity ( Figure 29 D and Table 9), while the turnover rate of rhizosphere and endodermal bacteria (D and Table 9) Figure 29 (G and Table 10) and the degree of change of Bray-Curtis distance between each fertilization treatment and the control over time ( Figure 29 Both H and Table 11) are similar to the results based on absolute abundance. However, unlike the results calculated based on absolute abundance, the Bray-Curtis distance between rhizosphere and root bacteria based on relative abundance did not show a continuously decreasing trend with plant development. Figure 29 Furthermore, fertilization did not significantly affect the β-diversity of rhizosphere and root bacteria based on relative abundance across different stages of plant development. Figure 29 F).

[0279] Table 9. Contribution of each factor to the difference in bacterial β-diversity based on relative abundance (calculated based on relative abundance)

[0280] Compartment 0.136 132.0 0.001 Stage 0.133 18.4 0.001 Treatment 0.045 14.5 0.001 Compartment × Stage 0.066 9.2 0.001 Compartment × Treatment 0.018 5.9 0.001 Stage×Treatment 0.049 2.3 0.001 Compartment × Stage × Treatment 0.033 1.5 0.001

[0281] Table 10. Relationship between temporal distance and Bray-Curtis distance based on relative abundance (calculated based on relative abundance)

[0282]

[0283] Table 11. Relationship between development time and Bray-Curtis distance based on relative abundance among treatments (calculated based on relative abundance).

[0284]

[0285] These results indicate that although calculating bacterial diversity based on absolute abundance and relative abundance yields partially similar results, significant differences still exist.

[0286] In summary, nitrogen fertilization induced a new developmental pattern in the soybean rhizosphere bacterial community, with increased α-diversity in the early stages of soybean development and significant differentiation in β-diversity compared to other fertilization treatments. Furthermore, phosphorus fertilization reduced the turnover rate of rhizosphere bacteria, which maintained high α-diversity in the later stages of soybean development.

[0287] 3. Composition of bacterial communities

[0288] The 16S sequences filtered by annotation included 30,122 amplified sequence variants (ASVs), of which 23,866 ASVs were found in the rhizosphere and 11,690 ASVs were found in the roots. Absolute quantitative analysis revealed that ecological niche, fertilization treatment, and developmental stage all significantly affected bacterial absolute abundance (Table 12). Bacterial abundance did not show significant differences in non-rhizosphere soils under different long-term fertilization treatments. Figure 7 A). As the plant develops, the abundance of bacteria in both the rhizosphere and root zone increases, remaining relatively stable during the early vegetative growth stage (D1-D14) (average 5.8 × 10⁻⁶ in the rhizosphere). 9 The average value within the root is 1.5 × 10⁻⁶. 8 Then it gradually increases, reaching its peak during the pod-forming stage (D72) (rhizosphere average 2.3 × 10⁻⁶). 10 The average value within the root is 4.7 × 10⁻⁶. 9However, in the rhizosphere soil, the abundance of NK-treated bacteria did not show a similar trend, and its abundance was significantly lower than that of the normal fertilization treatment, especially in the later stages of development. Figure 7 B) indicates that the soybean rhizosphere microbial community has a unique dynamic response pattern under low phosphorus conditions, which may be due to changes in secretions or developmental delays.

[0289] Table 12. Three-way ANOVA of bacterial absolute abundance (based on absolute abundance calculations)

[0290] Compartment 730.8 3.4E-100 Stage 71.9 8.0E-72 Treatment 21.8 2.8E-13 Compartment × Stage 25.8 3.6E-30 Compartment × Treatment 16.4 3.4E-10 Stage×Treatment 9.7 4.3E-26 Compartment × Stage × Treatment 4.7 7.3E-11

[0291] Sequence annotation revealed that the abundance of the nine major bacterial phyla in rhizosphere soil and root samples accounted for 98.39%-99.68% and 98.58%-99.98% of the total abundance, respectively. The most prevalent bacterial groups were Proteobacteria (30.8%-58.8% in rhizosphere and 15.7%-75.2% in roots), Actinobacteria (14.4%-43.7% in rhizosphere and 16.4%-77.1% in roots), and Bacteroidetes (1.6%-31.7% in rhizosphere and 0.3%-11.5% in roots). During soybean development, Actinobacteria dominated root-related microorganisms in the early stages and Proteobacteria in the later stages, a phenomenon most pronounced in the roots. Interestingly, the relative abundance of Bacteroidetes was low in the early stages of soybean growth, but reached 23.2% in the rhizosphere and 5.7% in the roots during the pod-setting stage (D72). Figure 7 (See Tables 13 and 14). Statistical analysis revealed that the relative abundance of Bacteroidetes in the rhizosphere and roots during the NK treatment was significantly lower than that during the pod-setting stage (D72) compared to the normal fertilization treatment. In the PK treatment, the relative abundance of Acidobacteria in the rhizosphere was significantly lower than that during the four developmental stages of soybean (D4, D14, D28, and D72) compared to the normal fertilization treatment, while the relative abundance of Firmicutes in the roots was significantly higher than that during the early stages of soybean growth (D4, D7, and D14) compared to the normal fertilization treatment. Figure 7 (and Table 15).

[0292] Table 13. Differences in relative abundance of bacteria in rhizosphere soil at different developmental stages (The numbers in the table represent the relative abundance (percentage) of bacteria in the rhizosphere soil samples, and the lowercase letters represent significant differences in relative abundance of bacteria at different developmental stages in the rhizosphere soil (P<0.05)).

[0293]

[0294] Table 14. Differences in relative abundance of bacteria in roots at different developmental stages (The numbers in the table represent the relative abundance (percentage) of bacteria in the root samples, and the lowercase letters represent significant differences in relative abundance of bacteria at different developmental stages in the roots (P<0.05)).

[0295]

[0296] Table 15. Differences in relative bacterial abundance among different fertilization treatments (The numbers in the table represent the percentage of relative bacterial abundance in rhizosphere soil (left) or root (right) samples, and the lowercase letters represent significant differences in relative bacterial abundance among different fertilization treatments in rhizosphere soil (left) or root (right) (P<0.05)).

[0297]

[0298]

[0299] These results indicate that the assembly patterns of rhizosphere and endophytic bacteria differ under different fertilization treatments and change dynamically with plant development.

[0300] 4. Composition of Rhizobium Community

[0301] To investigate the role of rhizobia in soybean growth, amplicon sequencing of the rpoB gene was performed on nodule and rhizosphere soil samples. The results showed 244,251-1,299,348 sequences for each nodule sample and 89,563-1,522,022 sequences for each rhizosphere soil sample. After taxonomic annotation filtering, nodule samples yielded 30,155-629,691 sequences, and rhizosphere soil samples yielded 132-157,398 sequences. The filtered sequences included a total of 235 ASVs, with 139 ASVs in nodules and 126 ASVs in rhizosphere soil.

[0302] Absolute quantitative analysis revealed that in non-rhizosphere soils under different long-term fertilization treatments, the abundance of rhizobia in the PK treatment was significantly higher than that in the normal fertilization treatment, while there were no significant differences among the other treatments. With plant development, the abundance of rhizobia in the rhizosphere soil showed an increasing trend, and its abundance remained relatively stable (5.5 × 10⁻⁶) during the early vegetative growth stage (D1-D14). 6 Then it gradually increases, reaching its peak (6.7 × 10⁻⁶) during the pod-forming stage (D72). 8 This trend is consistent with the changes in bacterial abundance. However, the abundance of rhizobia in root nodules did not show a clear pattern of change, remaining at 3.3 × 10⁻⁶ as the plant developed. 10 Up to 2.0×10 11 Dynamic changes within the range ( Figure 8 ).

[0303] Annotating ASV sequences to the species level of rhizobia revealed significant differences in the rhizobium composition between the PK treatment and the normal fertilization treatment in non-rhizosphere soil. Specifically, the relative abundance of *Sinorhizobium sp.* and *Bradyrhizobium jicamae* was significantly higher in the PK treatment than in the normal fertilization treatment, while the relative abundance of *Bradyrhizobium japonicum* and *Rhizobium anhuiense* was lower in the PK treatment. Figure 8 (A and Table 16). The nine most abundant rhizobium sequences in the rhizosphere soil samples accounted for 99.17%-100% of all sequences in the samples, with the main enriched rhizobia being Bradyrhizobium japonicum (35.5%-91.8%), Rhizobium anhuiense (5.2%-50.8%), and Bradyrhizobium jicamae (0.2%-23.5%). Figure 8 B). Almost all sequences in the root nodules were annotated as *Bradyrhizobium japonicum* (99.8%-100%), indicating that this rhizobium is a species that establishes a symbiotic relationship with soybean to form root nodules. Figure 8 C).

[0304] Table 16. Differences in relative abundance of rhizobia among different fertilization treatments (The numbers in the table represent the relative abundance (percentage) of rhizobia in the sample, and the lowercase letters represent significant differences in relative abundance of rhizobia among different fertilization treatments (P<0.05)).

[0305]

[0306]

[0307] Differential analysis revealed significant differences in the composition of the rhizobium community between the PK treatment and the normal fertilization treatment. In the rhizosphere soil, the relative abundance of *Bradyrhizobium jicamae* in the PK treatment was significantly higher than that in the normal fertilization treatment throughout the vegetative growth period, while the relative abundance of *Sinorhizobium sp.* was higher than the control in all three stages of the vegetative growth period (D4, D14, and D28). Conversely, the relative abundance of *Bradyrhizobium elkanii* was lower than the control at D1, D4, D7, D14, and D42. Figure 8 B and Table 16). Importantly, compared to the vegetative growth stage, the soybean slow-growing rhizobium (Bradyrhizobium japonicum) symbiotic with soybean accumulated a higher relative abundance during the reproductive growth stages (D42, D60, and D72). Figure 8(B and Table 17). These results indicate that long-term lack of nitrogen fertilizer application increases the abundance of rhizobia in the soil, and that more symbiotic rhizobia exist in the rhizosphere during the later stages of soybean growth.

[0308] Table 17. Differences in relative abundance of rhizobia in rhizosphere soil among different developmental stages (The numbers in the table represent the relative abundance (percentage) of rhizobia, and the lowercase letters represent significant differences in relative abundance of rhizobia among different developmental stages (P<0.05)).

[0309]

[0310] 5. Dynamic changes in bacterial abundance

[0311] Analysis of the dynamic changes in bacterial abundance during plant development revealed significant differences between the absolute and relative abundance trends, with the most pronounced difference observed in Actinobacteria. During the soybean growth cycle, the relative abundance of rhizosphere Actinobacteria decreased from 33.4% to 18.0%, while the absolute abundance increased from 1.9 × 10⁻⁶. 9 Increased to 3.7 × 10 9 The relative abundance of rhizobium actinomycetes decreased from 68.3% to 32.6%, while the absolute abundance increased from 8.5 × 10⁻⁶. 7 It increased significantly to 1.1×10 9 Furthermore, the relative abundance of Acidobacteria in the rhizosphere decreased slowly, but its absolute abundance remained relatively stable; the relative abundance of Firmicutes, Acidobacteria, and Chlorophycetes in the rhizosphere showed a decreasing trend, while their absolute abundance showed an increasing trend. The relative abundance and absolute abundance of Proteobacteria and Bacteroidetes in both the rhizosphere and roots exhibited similar dynamic changes, both showing a continuous increase. Figure 9 (Figures, Tables 13 and 14). These results indicate that relative abundance information cannot reflect the true number of microorganisms in a sample, which can lead to serious misjudgments of the true changes in the microbial community between samples, thus affecting the experimental conclusions.

[0312] To further demonstrate the significant difference between the dynamic trends of absolute and relative bacterial abundance, the dynamic trends of high-abundance rhizosphere bacteria with plant development were also analyzed at the ASV level. Cluster heatmap analysis revealed that in rhizosphere samples under four different fertilization treatments, the relative abundance of these high-abundance bacterial communities exhibited trends over time that were either similar to or the same as the absolute abundance (green clusters) or completely different from the absolute abundance (red clusters). Figure 10 Further statistical analysis revealed that, in the NPK, PK, NK, and NP treatments, the proportions of bacterial communities with completely opposite trends in relative and absolute abundance over time accounted for 38.1%, 46.5%, 44.9%, and 48.3% of the total bacterial community in each treatment, respectively. Figure 10 E). These results further demonstrate that absolute quantitative methods can provide more accurate experimental results when characterizing microbial community features, while analysis results based on relative abundance are somewhat "misleading." Therefore, the analysis methods based on absolute quantification will be mainly adopted in the following steps.

[0313] Example 3: Characteristics and functions of soybean rhizosphere bacteria

[0314] In this embodiment, the abundance differences of core ASVs among different fertilization treatments were analyzed mainly based on absolute quantitative methods. Biomarker groups in different fertilization treatments and different developmental stages were identified by LEfSe analysis and random forest analysis. Then, the function of bacterial communities was predicted and analyzed using the FAPROTAX database (Louca et al., 2016). Finally, the differences in rhizosphere bacterial function among different treatments were analyzed by metagenomic assembly, and the characteristics and potential functions of core rhizosphere bacteria in soybean were comprehensively explored.

[0315] 1. Differences in rhizosphere core ASV among different fertilization treatments

[0316] After filtering out low-abundance ASVs with an abundance percentage of less than 0.1% in each rhizosphere sample, a total of 573 ASVs were obtained. Although these ASVs accounted for only 2.4% of the total ASV species in the rhizosphere, they represented 74.6% of the total abundance. Figure 11 A). Further analysis revealed that the abundance of these 573 ASVs was highly correlated with the abundance of total rhizosphere ASVs in all fertilization treatments (R2>0.977). Figure 11 B), therefore it is defined as a core ASV.

[0317] Statistical analysis of the abundance of core rhizosphere ASVs in various fertilization treatments and developmental stages revealed that, except for the NK treatment where rhizosphere bacterial abundance showed relatively small changes with plant development, the abundance of rhizosphere bacteria in the other fertilization treatments showed a slow increasing trend with plant development. Compared with normal fertilization, the abundance of rhizosphere bacteria in the NK treatment significantly decreased in the later stages of soybean development (D42-D72), while the abundance of rhizosphere bacteria in the NP treatment was highest at the pod-setting stage (D72). Figure 11 C), which is consistent with previous findings in bacterial community composition. Figure 7 B). Hierarchical cluster analysis was further used to explore the changes of these core rhizosphere ASVs across different fertilization treatments. The results showed that many core ASVs exhibited unique patterns in PK treatments compared to normal fertilization treatments. Specifically, ASVs in cluster 10 were specifically excluded in PK treatments, while ASVs in cluster 15 were specifically enriched in PK treatments. Figure 11 C), similar to the results of the diversity analysis ( Figure 6(B, E, and H) This further demonstrates that long-term lack of nitrogen fertilizer significantly alters the developmental pattern of rhizosphere core ASVs in PK treatment.

[0318] Next, the differences in ASVs between each fertilization treatment and the normal fertilization treatment were compared. Figure 12 Paired comparisons of the absolute abundance of core ASVs in the rhizosphere at different developmental stages among different fertilization treatments revealed that the abundance of 172 core ASVs differed between the PK treatment and the normal fertilization treatment, with 69 ASVs significantly enriched, mainly including Sphingomonadales and Rubrobacterales. More ASVs from bacterial groups such as Actinomycetales, Burkholderiales, and Acidobacteriales were excluded by the PK treatment. Figure 12 A). In the NK treatment, 67 core ASVs were significantly different from those in the normal fertilization treatment, the vast majority of which (98.5%, 66) were significantly downregulated, including beneficial bacterial groups such as Sphingomonas, Rhizobium, and Burkholderia. Figure 12 B) indicates that long-term lack of phosphate fertilizer application disrupts the beneficial interactions between plants and rhizosphere bacteria, which may be a reason for the reduced soybean yield. In the NP treatment, there were fewer differentially expressed core ASVs compared to the control, including four ASVs belonging to the Flavobacteriales order that were excluded. Figure 12 C).

[0319] 2. Identification of core bacterial marker groups

[0320] To identify differentially abundant species among different fertilization treatments and different growth and development stages, LEfSe (Linear Discriminant Analysis Effect Size) analysis based on absolute abundance was performed on rhizosphere soil bacterial communities at six taxonomic levels from phylum to species. A large number of differentially abundant marker taxa were found among different fertilization treatments (LDA>6) and different growth and development stages (LDA>6). A total of 217 differentially abundant taxa were identified among different fertilization treatments, with 83, 83, 1, and 50 bacterial marker taxa identified in the NPK, PK, NK, and NP treatments, respectively. Specifically, the normal fertilization treatment identified 20 families, including Micrococcaceae and Burkholderiaceae; the PK treatment identified 15 families, including Rhizobiaceae and Rubrobacteraceae; and the NP treatment identified 13 families, including Weeksellaceae and Enterobacteriaceae. Figure 13 ).

[0321] Furthermore, 199 differential taxa were identified across different developmental stages. Two, eight, 13, 1, 16, 70, 3, and 86 bacterial marker taxa were identified at developmental stages D1, D4, D7, D14, D28, D42, D60, and D72, respectively. Specifically, Oxalobacteraceae and Alteromonadaceae families were identified at stage D4; Moraxellaceae and Chromatiaceae families at stage D7; Pseudomonadaceae and Bacillaceae families at stage D28; Caulobacteraceae and Streptomycetaceae families at stage D42; Enterococcaceae family at stage D60; and Weeksellaceae and Bradyrhizobiaceae families at stage D72. Figure 14 These results indicate that the rhizosphere bacterial community contains specifically enriched bacterial groups in various fertilization treatments and different developmental stages. Therefore, it is possible to use these biomarker groups to predict nutrient types and plant developmental stages in farmland soil.

[0322] 3. Random forest prediction of rhizosphere core bacterial marker groups

[0323] To further analyze and identify biomarker groups of rhizosphere core bacterial communities under different fertilization treatments and at different growth and development stages, random forest prediction models were constructed at the phylum, class, order, family, genus, and species levels using species composition characteristics of rhizosphere core bacterial communities based on absolute abundance. First, the error rate and accuracy of the random forest prediction models were evaluated at each taxonomic level. The results showed that the error rate was relatively low and the accuracy was high at the family, genus, and species levels (Table 18). To ensure comparability with LEfSe results, the random forest model was trained and predicted at the family taxonomic level.

[0324] Table 18. Error rate and accuracy of random forest prediction models (These results are calculated based on absolute abundance. They represent the error rate and accuracy of random forest prediction models constructed based on developmental stage or fertilization treatment data at the phylum, class, order, family, genus, and species levels, respectively).

[0325]

[0326] At the scientific level, the trained random forest model achieved a prediction accuracy of 79.2% across different fertilization treatments and as high as 90.6% across different developmental stages. Figure 30 Next, 10× cross-validation revealed that using 16 families as biomarkers among different fertilization treatments and different developmental stages had a low cross-validation error rate of 19.9%. Figure 15 A) and 14.1% ( Figure 16 A). Next, the absolute abundance of these 16 biomarker groups in different fertilization treatments and at different growth and development stages was analyzed. The results showed that the PK treatment was enriched with bacterial groups such as Actinospicaceae, Rubobacteraceae, Sinobacteraceae, Dolo_23, and Propionibacteriaceae in different fertilization treatments. Consistently, these biomarker groups also appeared in the corresponding LEfSe results. Figure 13 This indicates that the prediction model has high accuracy.

[0327] In addition, during different growth and development stages, Oxalobacteraceae was enriched at D4, Moraxellaceae at D7, Bacillaceae and Methylophilaceae at D2 and D8, Halomonadaceae, Caulobacteraceae and Kouleothrixaceae at D4 and D7, and Paenibacillaceae, Bradyrhizobiaceae, Rhizobiaceae, Weeksellaceae and Sphingobacteriaceae at D7 and D8. Figure 16 B). Similarly, these biomarker groups at each developmental stage are consistent with the corresponding LEfSe results ( Figure 14 The above results further demonstrate that, in the Northeast black soil farmland system, these biomarkers can be used to predict soil nutrient types and soybean development stages, and also provide a theoretical basis for improving the structure and composition of soybean rhizosphere microbial communities.

[0328] 4. Functional analysis of the core bacterial community

[0329] To explore the potential functions of core rhizosphere bacterial communities in rhizosphere soil, functional annotation of all core bacterial ASVs was performed using the FAPROTAX database. After functional prediction of 573 core bacterial ASVs, 45 related metabolic pathways were identified. Then, 16 bacterial functions with high abundance were selected, and their abundance differences across different fertilization treatments and developmental stages were compared. These mainly included 8 carbon metabolism-related pathways and 8 nitrogen metabolism-related pathways. Figure 17 and Figure 18 ).

[0330] To compare the differences in the potential functions of core bacterial ASVs among different fertilization treatments, a pairwise comparative analysis was performed on the abundance ratios of core bacteria at each developmental stage in different treatments. The results showed that the abundance ratios of genes related to methylotrophy (C-metabolism) and nitrogen fixation (N-metabolism) were significantly higher in the PK treatment than in the normal fertilization treatment, while the abundance ratios of genes related to aromatic compound degradation and lignin degradation (C-metabolism) and nitrate ammonification (N-metabolism) were lower in the PK treatment. This indicates that low nitrogen conditions significantly altered the C and N metabolism functions of soybean rhizosphere bacteria, further validating the unique developmental pattern induced by long-term nitrogen-free fertilization in soybean rhizosphere bacteria. The increased proportion of nitrogen-fixing bacteria is beneficial for replenishing soil nitrogen nutrient levels. In the NK treatment, compared with the normal fertilization treatment, the abundance proportions of C metabolism functional genes such as chemoheterotrophy, aerobic chemoheterotrophy, fermentation, and chitinolysis decreased, while the abundance proportions of most N metabolism functional genes significantly increased, including nitrite respiration, denitrification, nitrification, and nitrate ammoniation. This indicates that low phosphorus reduced C metabolism activity in soybean rhizosphere bacteria but significantly enhanced N metabolism-related functions. Furthermore, there were no significant changes in rhizosphere bacterial function between the NP treatment and the normal fertilization treatment. Figure 17 This indicates that low potassium conditions have a limited impact on rhizosphere bacteria.

[0331] Furthermore, differential analysis was used to further compare the potential functional differences of core bacterial ASVs across different growth and development stages. The inventors found that the abundance proportion of genes involved in aromatic compound degradation was significantly higher in stage D7 than in other developmental stages, and the abundance proportion of genes involved in fermentation was also significantly increased in stage D28. The abundance proportion of genes involved in chitin degradation was highest in stage D72, and these changes may be due to alterations in soybean secretions during development. Importantly, the inventors also found that the abundance of genes involved in nitrogen metabolism, such as nitrite respiration, denitrification, nitrification, and nitrate ammoniation, was significantly higher in the early growth stages of soybean (D1-D28) than in the later stages of reproductive growth (D60 and D72), while the abundance proportion of genes involved in nitrogen fixation and urea degradation continued to increase with soybean development. These results indicate that, throughout the soybean growth cycle, rhizosphere bacteria exhibit higher nitrogen metabolism activity in the early developmental stages compared to later stages. However, the abundance of nitrogen metabolism functions associated with increasing soil nitrogen nutrient levels (nitrogen fixation and urea degradation) continues to increase with plant development, suggesting that more nitrogen nutrients are required in later stages of plant development. Figure 18 ).

[0332] 5. Metagenomic functional analysis

[0333] To further investigate the function of rhizosphere bacterial communities under different fertilization treatments, metagenomic sequencing analysis was performed on rhizosphere soil samples from soybeans at the early growth stage (D1), initial flowering stage (D42), and pod-setting stage (D72). A total of 2166 GB of data was obtained from 36 rhizosphere soil samples after metagenomic sequencing. After quality control assembly, each sample yielded an average of 2,081,141 contigs (786,987-3,223,911), with an average length of 793.8 bp (608.2-873.2 bp) and an average N50 of 1320 (981-1650). Gene prediction yielded an average of 470,823 complete genes (82,032-959,495) per sample. After deredundancy removal, a total of 3,454,121 non-redundant genes were obtained. The contigs assembled from each sample were binned using metagenomic methods. After redundancy removal from the resulting 655 bins, 142 bins with contamination levels less than 10% and integrity greater than 70% were retained. Annotation revealed that 140 of these bins were bacterial. Figure 19 ), 2 bins are archaea.

[0334] First, the 140 bacterial bins obtained after binning were classified and functionally annotated. Similar to the species composition results obtained from previous 16S amplicon sequencing, Proteobacteria and Actinobacteria accounted for a high proportion in the rhizosphere, representing 55% (77) and 27.9% (39) of the total bins, respectively. Furthermore, it was found that most bacteria in the soybean rhizosphere possessed functional genes related to nitrogen mineralization, nitrogen limitation adaptation, organic phosphorus mineralization, inorganic phosphorus solubility, phosphorus translocation, phosphorus starvation response, and potassium translocation. Only a few bacteria possessed functional genes related to nitrogen reduction and denitrification, and a very small number possessed genes related to nitrification. Figure 19 The high nutrient content of the black soil in Northeast China may explain why soybean rhizosphere bacteria commonly possess genes responsible for the metabolism of nutrients such as nitrogen, phosphorus, and potassium. Furthermore, these bacteria may further enhance the content and availability of these nutrients in the soil.

[0335] Next, all non-redundant genes were annotated into the KEGG, COG, and CAZyme databases, and functional diversity among different fertilization treatments was calculated. The Shannon index showed that, compared to the normal fertilization treatment, the functional α-diversity of the PK treatment was significantly reduced across all databases (KEGG, COG, CAZyme), while the NK and NP treatments showed no significant difference compared to the control. Constrained Principal Coordinate Analysis (CPCoA) results indicated that, except for the high similarity in functional β-diversity between the normal fertilization and NP treatments, the β-diversity of the PK and NK treatments compared to the normal fertilization treatment in KEGG, COG, and CAZyme functions showed significant differences (P<0.05). Figure 20 These results indicate that long-term unbalanced fertilization significantly alters the functional diversity of rhizosphere bacteria, which may exacerbate their functional adaptation under different nutrient conditions.

[0336] Further analysis revealed differences in the abundance of functional genes related to microbial nitrogen, phosphorus, and potassium cycling among different fertilization treatments. Compared with the normal fertilization treatment, gene-specificity was reduced in functions related to nitrogen reduction (narB, narG, and narH), denitrification (nirK), inorganic phosphorus dissolution (gcd and ppa), and potassium transport (kdpA, ktrB, and kefB), while nitrogen mineralization-related functions were enhanced. This was consistent with 16S function prediction results (…). Figure 17Similarly, the nitrogen-fixing gene clusters (nifD, nifH, and nifK) were consistently increased in the PK treatment. These results indicate that although low-nitrogen conditions generally reduce the nutrient metabolism of rhizosphere microorganisms, they enhance their nitrogen-fixing function. In the NK treatment, the abundance of inorganic phosphorus-dissolving genes (gcd and ppa) and phosphorus starvation response genes (phoU) increased significantly, while the abundance of nitrification genes (hao and amoA), denitrification genes (nirK, nosZ, and nirS), and nitrogen limitation adaptation genes (glnA), which are involved in nitrogen metabolism-related functions, also increased significantly. Conversely, the abundance of nitrogen reduction-related genes (nirB, nirA, narB, and nasA) and potassium transport-related genes (kdpA and ktrB) showed a decreasing trend. These results indicate that low-phosphorus conditions promote the responsiveness and supply capacity of soybean rhizosphere microorganisms to phosphorus, while also enhancing nitrogen metabolism and reducing potassium transport. Compared with normal fertilization treatment, the abundance of nutrient cycling-related functional genes did not change significantly in the NP treatment, but the abundance of potassium transport-related genes increased, especially the Kup gene. This indicates that low potassium conditions enhanced potassium transport by soybean rhizosphere microorganisms. Figure 20 These results further demonstrate that the functions of soybean rhizosphere microorganisms differ significantly under different nutrient conditions, and that the functional adaptation of rhizosphere microorganisms in different environments enhances the plant's adaptability to various stresses, which is beneficial for the plant's absorption of limiting soil nutrients.

[0337] Example 4: Co-occurrence network analysis and functional analysis of bacterial ecological clusters

[0338] Based on absolute quantitative methods, the co-occurrence network of core bacteria in each fertilization treatment was analyzed. Through modular analysis and differential analysis, ecological clusters that may function under low nitrogen conditions were identified. Then, the growth-promoting function of bacteria was verified through bacterial isolation culture and synthetic community re-inoculation experiments. The potential relationship between the function of rhizosphere bacterial communities and soybean growth and development under different fertilization treatments was comprehensively explored.

[0339] 1. Co-occurrence network analysis among different fertilization treatments

[0340] To explore the direct and indirect interactions among bacteria in different samples, co-occurrence networks analysis was performed on samples from the rhizosphere and root fertilization treatments. Spearman correlations among the ASVs were calculated based on the absolute abundance of core ASVs in the rhizosphere and root fertilization treatments. Data with correlation coefficients (r > 0.8) and significance (P < 0.05) were used to construct bacterial co-occurrence networks for different fertilization treatments in the rhizosphere and root fertilization treatments. Basic network topological characteristics such as the number of nodes, number of edges, average degree, and robustness were also calculated for each network. In the rhizosphere samples, the normal fertilization treatment network had 293 nodes, 1096 edges, and an average degree of 7.48. The PK and NP networks had similar numbers of nodes (289 and 266, respectively) to the normal fertilization treatment network, but significantly more edges (3249 and 3341, respectively). This also resulted in higher average degrees for the PK and NP networks (22.48 and 25.12, respectively). The NK network has 270 nodes, comparable to the normal fertilization treatment network, with 1442 edges and an average degree of 10.68, slightly higher than the control. Figure 21 A). These results indicate that under low nitrogen and low potassium conditions, the soybean rhizosphere enhances bacterial interactions, leading to increased bacterial network complexity. This promotes metabolic cooperation among microorganisms and enhances their adaptability to nutrient-deficient environments.

[0341] The robustness of the rhizosphere bacterial network under each fertilization treatment was further analyzed by randomly removing 50% of the nodes in the network. The results showed that the stability of the PK, NK, and NP networks was significantly higher than that of the normal fertilization treatment network. Figure 21 B). Next, each network was subdivided according to the target level, and the average degree difference of bacterial groups with high degree among different fertilization treatments was calculated. Significance analysis showed that the average degree of Rhizobiales, Actinomycetales, Sphingomonadales, and Burkholderiales bacterial groups in the PK and NP networks was significantly higher than that in the normal fertilization treatment network. Furthermore, the average degree of Gaiellales bacterial group in the PK network was also higher than that of the control (…). Figure 21 C). In summary, these results indicate that the absence of nitrogen or potassium fertilizer significantly increases the complexity of rhizosphere bacterial networks, while rhizosphere bacterial networks lacking nitrogen, phosphorus, and potassium exhibit greater stability. Network complexity and stability are important characteristics of ecosystem function, and increased complexity and stability may be a potential reason for rhizosphere bacteria and plants to alleviate stress.

[0342] Simultaneously, the bacterial co-occurrence networks of various fertilization treatments within the root were analyzed. The network for the normal fertilization treatment had 198 nodes, 584 edges, and an average degree of 5.9. Similar to the results of the rhizosphere bacterial network, within the root, the average degree of the PK network was 10.4, significantly higher than that of the normal fertilization treatment network, while the average degrees of the NK and NP networks were 7.76 and 6.63, respectively, also slightly higher than those of the normal fertilization treatment network. Figure 22 ).

[0343] To further verify the accuracy of the calculated basic network features, the same dataset was used for repeated validation on the MENA (Molecular Ecological Network Analysis Pipeline) website. Similar results were also observed across different fertilization treatments in the co-occurrence network of bacteria in the rhizosphere and roots. Figure 31 ).

[0344] 2. Co-occurrence network analysis of rhizosphere bacterial communities

[0345] Considering the co-correlation among different fertilization treatments and the fact that rhizosphere bacteria are more sensitive to fertilization treatments ( Figure 27 Therefore, a total co-occurrence network was constructed using samples from all different fertilization treatments in the rhizosphere to compare the changes in ecological clusters (i.e., modularization) caused by different fertilization treatments. The results showed that the rhizosphere bacterial network contained three main modules ( Figure 23 A), where the absolute abundance of bacteria in the PK treatment was significantly higher than that in the normal fertilization treatment in module 2, but significantly lower than that in module 3. The absolute abundance of bacteria in the NK treatment was significantly lower than that in the control in all three modules, while the absolute abundance of bacteria in the NP treatment showed no significant difference from the control. Figure 23 B). Statistical analysis of bacterial phylum-level classification revealed that Modules 1 and 3 were mainly composed of Proteobacteria, while Module 2 was primarily composed of Actinobacteria, with a significantly reduced proportion of Acidobacteria. Figure 23 C). Since the preceding results indicate that not applying nitrogen fertilizer does not reduce soybean yield ( Figure 4 Furthermore, under low-nitrogen conditions, soybean rhizosphere bacteria developed a unique developmental pattern. Figure 6 , Figure 11 and Figure 17 Therefore, it is suggested that the core bacteria in the PK treatment-specific enrichment module (module 2) may affect the growth and development of soybeans.

[0346] 3. Isolation and purification of rhizosphere bacteria

[0347] To verify the above results, rhizosphere bacteria from different fertilization treatments were isolated and purified. A total of 1011 monoclonal bacteria were isolated. These monoclonal bacteria were then purified by streak plating, followed by PCR amplification and Sanger sequencing identification using universal primers for the 16S rRNA gene. After annotating the sequencing results into the Greengenes database, it was found that the isolated bacteria mainly belonged to the phyla Proteobacteria, Firmicutes, Actinobacteria, and Bacteroidetes. Proteobacteria were the most numerous, accounting for 45% (455) of the total isolated bacteria, followed by Firmicutes and Actinobacteria, accounting for 24.4% (277) and 21.2% (214) respectively. Bacteroidetes were the least numerous, accounting for only 4.6% (47). Figure 24 A). Next, the core ASVs of rhizosphere bacteria were sorted from highest to lowest abundance, and the sequences of the isolated strains were compared with the sequences of the core ASVs. The results showed that when the cumulative abundance of core ASVs reached 60% and 80%, the total abundance of the isolated strains accounted for 58% and 49% of the total abundance of core ASVs, respectively, indicating that the isolated strains had a high coverage rate among rhizosphere ASVs. Figure 24 B).

[0348] Furthermore, BLAST analysis was performed on the ASV sequences of each module obtained from the previous co-occurrence network analysis of rhizosphere bacterial communities and the sequences of isolated strains, selecting results with a similarity greater than 95%. Combining the modular analysis and differential analysis results of the rhizosphere bacterial co-occurrence network, seven isolated strains were finally selected from module 2 as experimental synthetic communities, and combined into two synthetic communities, SynCom7 (containing 7 strains) and Syncom5 (containing 5 strains), for inoculation experiments to verify the function of the ecological cluster in module 2. Simultaneously, five isolated strains were also selected from module 1 as control synthetic communities. Figure 24 A).

[0349] 4. Synthetic community reintroduction experiment

[0350] Based on the above experimental results and the isolation, purification, and identification of bacteria, the experimental strains used for synthetic community re-inoculation were finally selected. The synthetic community re-inoculation experiment was conducted under artificial climate laboratory conditions. Soybean seeds were disinfected and germinated uniformly in native soil. Three days after complete germination, plants with excessively fast or slow growth were removed, and then 1 mL of inoculum solution was inoculated at the root of each plant. The inoculum solution mainly included four types: sterile water (Control), a control synthetic community containing 5 strains from module 1 (SynCtrl), an experimental synthetic community containing 7 strains from module 2 (SynCom7), and an experimental synthetic community containing 5 strains from module 2 (SynCom5). Figure 25A). Seven days after inoculation, 1 mL of the corresponding inoculation solution was added to the roots of the plants. After three weeks of growth, the plant height and dry weight were recorded.

[0351] Statistical analysis revealed no significant difference in plant height and dry weight between soybeans inoculated with nitrogen-enriched vermiculite and those inoculated with sterile water, compared to the control synthetic community. Compared to the control synthetic community, inoculation with both SynCom5 and SynCom7 synthetic communities increased soybean plant height and dry weight, with SynCom5 showing the most significant effect, increasing plant height by 15% and dry weight by 38%. In vermiculite without nitrogen, soybean plant height after inoculation with the control synthetic community was not different from that after inoculation with sterile water, but dry weight was higher. Similarly, compared to the control synthetic community, inoculation with SynCom5 increased soybean plant height by 6% and dry weight by 18%; inoculation with SynCom7 significantly increased both plant height and dry weight compared to inoculation with sterile water. Figure 25 BD).

[0352] To further investigate the pathways by which the bacteria in these synthetic communities promote soybean growth, the inventors examined the physiological functions of these bacteria. The results showed that, compared to the bacteria in the control synthetic communities, the bacteria in the experimental synthetic communities (SynCom5 and SynCom7) produced higher concentrations of IAA and more active ACC deaminase, and the inorganic phosphorus solubilization function was only observed in the experimental synthetic communities. Figure 32 IAA, as a plant growth regulator, effectively promotes plant growth. ACC deaminase effectively promotes biomass.

[0353] In summary, the experimentally synthesized communities promoted soybean growth under pot conditions, with SynCom5 showing the most significant growth-promoting effect. The growth-promoting effect of the specifically enriched ecological clusters in the PK treatment indicates that, under low-nitrogen conditions, rhizosphere bacterial communities other than rhizobia also have a significant growth-promoting function for soybeans. This further enhances the adaptability of soybeans to low-nitrogen environments and may potentially promote soybean productivity.

[0354] The above findings indicate that as soybeans develop, the α-diversity of rhizosphere bacteria gradually decreases, while bacterial abundance gradually increases, primarily including Proteobacteria, Bacteroides, and Rhizobia. In the PK treatment, bacterial α-diversity was significantly higher than in the normal fertilization treatment during the early vegetative growth stage, while in the NK treatment, bacterial α-diversity was significantly higher than in the normal fertilization treatment during the reproductive growth stage. Figure 26A). In black soils with long-term fertilization, compared with normally fertilized soils, PK soils showed a 35% decrease in available nitrogen content, a 1.1-unit increase in pH, and an increase in the number and diameter of soybean root nodules; NK soils showed a 95% decrease in available phosphorus content, a decrease in the number and diameter of soybean root nodules, and a 25-29% decrease in soybean yield; NP soils showed a 61% decrease in available potassium content. In the rhizosphere, different fertilization treatments induced functional adaptations of microorganisms to the environment. The PK treatment enhanced nitrogen fixation and nitrogen mineralization functions of microorganisms, the NK treatment enhanced inorganic phosphorus dissolution and phosphorus starvation response functions, and the NP treatment enhanced potassium translocation. Furthermore, nutrient deficiencies such as N, P, and K enhanced the stability of the rhizosphere bacterial network, while the PK and NP treatments also enhanced the complexity of the rhizosphere bacterial network. Most importantly, the specifically enriched ecological clusters in the PK rhizosphere network promoted soybean growth, increasing soybean dry weight and plant height, which may be one of the key factors in maintaining yield. Figure 26 B).

[0355] discuss

[0356] Rhizosphere microbial communities are highly dynamic and strongly regulated by soil nutrient conditions and plant growth and development. Plants shaping specific rhizosphere microbial communities can enhance their adaptability to different developmental stages and environments. Rhizosphere microorganisms are the "second genome" of plants, playing a crucial role in regulating plant growth and development, promoting nutrient absorption, enhancing plant resistance, and maintaining plant health. However, the long-term irrational use of chemical fertilizers has not only exacerbated environmental pollution but also severely damaged the structure and function of soil microorganisms, hindering sustainable agricultural development. Based on a long-term fertilization field experiment, this study investigated the assembly dynamics and functional adaptation of the rhizosphere microbial community in soybean under different fertilization treatments and at different developmental stages. For the first time, it described the quantitative microbiome characteristics of soybean throughout its entire life cycle under different nutrient conditions, revealing the ecological mechanisms by which rhizosphere microorganisms promote crop growth.

[0357] 1. The effects of fertilization on soil physicochemical properties and crop yield

[0358] This invention found that, compared with normally fertilized soils, the content of available nitrogen in PK soils decreased by 35%, the content of available phosphorus in NK soils decreased by 95%, and the content of available potassium in NP soils decreased by 61%. Figure 3The second national soil survey classified soil nutrient content into six levels from high to low (Soil Survey Office, 1993). Soils with normal fertilization had the highest levels of available nitrogen, available phosphorus, and available potassium (Level 1, extremely high). Soils with PK soil had the highest levels of available nitrogen (upper-middle range) and NP soil had the highest levels of available potassium (Level 3, upper-middle range), while soils with NK soil had the highest level of available phosphorus (Level 4, lower-middle range). The inventors' research showed that in the soybean-corn-wheat rotation system, although long-term lack of nitrogen fertilizer reduced the corresponding soil nutrient content, soybean yield did not decrease; in fact, it increased by 18% in the 2020 planting season. Furthermore, soybean productivity in Northeast China was mainly limited by the available phosphorus content in the soil; soybean yields in the no-phosphate fertilizer treatment decreased by 25% and 29% in the 2020 and 2017 planting seasons, respectively. Figure 4 (A and B).

[0359] Contrary to the results of the NK treatment, the number of nodules and the diameter of root nodules in soybeans were significantly higher in the PK treatment than in the nitrogen application treatment. Figure 4 (C and D) and the content of available nitrogen in the soil remained at a high level. The higher nitrogen content in the soil without nitrogen fertilizer may be due to the fact that soybeans can simultaneously utilize nitrogen absorbed by the roots and nitrogen fixed biologically. On the one hand, low nitrogen promotes nodule formation, and the increase in the number of root nodules can increase the nitrogen content in the plant and the soil; on the other hand, under the condition of no exogenous nitrogen input, the soil and root-associated microbial community can also increase the mineralization of nitrogen in the soil. In addition, it was found that the soil pH value in the PK treatment was significantly higher than that in other treatments, indicating that long-term nitrogen application will cause soil acidification. The continuous acidification of the soil will lead to soil compaction, decreased fertility and reduced biological activity, which will seriously affect crop health and yield. Although the soluble organic carbon content in the soil in the PK treatment was lower than that in the normal fertilization treatment, the organic matter content did not change significantly.

[0360] 2. Temporal dynamics of soybean root-related microorganisms

[0361] Microorganisms are key regulators of soil organic matter decomposition and nutrient cycling, and their diversity and biomass are crucial for soil quality and health. The inventors found no significant difference in α-diversity of non-rhizosphere soil bacteria across different long-term fertilization treatments. Figure 6 A and Figure 29 A), while the α-diversity of rhizosphere bacterial communities is significantly affected by plant developmental stage and different fertilization treatments ( Figure 6 B and Figure 29 B). Throughout the growth cycle, the α-diversity of rhizosphere bacteria was significantly higher during the vegetative growth phase than during the reproductive growth phase. Figure 6 B and Figure 29B) suggests that the decrease in root exudates during the transition from vegetative to reproductive growth in plants affects the diversity of the rhizosphere bacterial community. It was also found that in the rhizosphere, the α-diversity of the PK treatment was significantly higher than that of the normal fertilization treatment in the early stages of vegetative growth, while the α-diversity of the NK treatment was higher than that of the normal fertilization treatment in the later stages of reproductive growth. Figure 6 B and Figure 29 B). Rhizosphere bacterial diversity was not reduced during the reproductive growth phase in NK treatment.

[0362] Plants recruit a large number of microorganisms throughout their life cycle, and the colonization patterns of these microorganisms are driven by specific root exudates during plant growth, resulting in a highly dynamic root-associated microbial community. The rhizosphere is a hotspot for soil microbial communities, and rhizosphere microorganisms are crucial for maintaining plant growth and productivity. Because the abundance of rhizosphere microbial groups varies significantly across different plant developmental stages, data interpretation based on relative abundance may be biased. The inventors also found a significant difference between the absolute and relative abundance of bacteria as soybeans develop, with the most pronounced difference observed in Actinobacteria. Figure 9 Furthermore, 38.1%–48.3% of the rhizosphere-rich ASVs in different fertilization treatments also showed different trends. Figure 10 Based on quantitative microbiome profiling (QMP) results, significant differences exist between rhizosphere and root bacteria in terms of abundance, diversity, and community composition. Figure 6 and Figure 7 Microorganisms need to penetrate cell walls and overcome host immunity to successfully colonize roots; therefore, endophytic microbial communities are mainly regulated by their host genotype. Except for the NK treatment, rhizosphere bacterial communities exhibited a high turnover rate during plant growth and development (absolute quantitative results: 0.43-0.54%; relative quantitative results: 0.28-0.34%), significantly higher than in non-rhizosphere soils. These results indicate that as bacterial abundance gradually increases with plant growth and development, the rhizosphere bacterial community also accelerates its turnover rate. In particular, the inventors found that the absolute abundance of bacteria remained relatively stable throughout the vegetative growth phase in all treatments, except for the NK treatment, where the absolute abundance increased sharply during the reproductive growth phase. Figure 7 and Figure 11 C) suggests that this is due to changes in root exudates and increased root deposits after soybeans enter the flowering stage.

[0363] Soybean root-related bacteria mainly consist of Actinobacteria, Proteobacteria, and Bacteroidetes. Actinobacteria dominate in the early stages of plant growth, while Proteobacteria and Bacteroidetes become dominant in later stages. Figure 7 and Figure 9 The dominant positions of Proteobacteria, Actinobacteria, and Bacteroides in plant roots have long been widely recognized in Arabidopsis thaliana and various crops. Previous studies have shown that actinomycetes can be selectively recruited by living plant cells but not by lignified cells. Many actinomycete groups can produce a range of antimicrobial secondary metabolites and enhance the host's adaptability to the environment. Considering that plants are more susceptible to pathogen infection in the early stages of growth, the inventors hypothesize that protective microorganisms in early plant growth are crucial for plant health and development. In contrast, Proteobacteria and Bacteroides are generally considered eutrophic bacteria with a typical R strategy, enabling them to grow rapidly in resource-rich environments. The abundance of these groups increases rapidly with plant growth, which is partly related to changes in the types and quantities of plant root exudates. Many groups belonging to Proteobacteria can promote plant growth through various mechanisms, including nitrogen fixation, phosphorus solubilization, auxin production, and siderophore secretion. The role of Bacteroides in the rhizosphere is not yet fully elucidated, but they have been reported to be important participants in soil nutrient cycling. Bacterial groups belonging to the phylum Bacteroidetes contain genes related to denitrification, and Bacteroides exhibit a high response to nitrogen and phosphorus cycling processes in the rhizosphere soil. In this invention, all fertilized soils were acidic (pH < 6.5). Figure 3 F), almost all sequences in the root nodules were annotated as Bradyrhizobium japonicum (F), a slow-growing rhizobium of soybean. Figure 8 C). The composition of rhizobia in soybean root nodules depends on soil conditions. *Rhizobium stenoptera* is the most abundant genus in naturally acidic or neutral soils, while *Sinapis sinensis* is dominant in root nodules from alkaline soils. These results suggest that plant development drives the assembly of root-associated microbiomes and exhibits close interaction with host environmental adaptation.

[0364] 3. Effects of fertilization on rhizosphere microbial community and function

[0365] The inventors discovered that NK treatment significantly reduced the time turnover of rhizosphere bacteria. Figure 6 G), and the abundance of rhizosphere bacteria was consistently lower than that of normal fertilization treatment throughout the plant's developmental stages, especially at maturity. Figure 7 B) indicates that phosphorus deficiency delayed the development of the rhizosphere bacterial community. Similarly, drought stress also significantly reduced the development of the root microbial community. Therefore, the inventors propose the concept that the delay in root microbial development is common when plants face nutrient deficiency or abiotic stresses such as drought. Furthermore, it was found that the absolute abundance of most core ASVs that showed significant differences in the NK treatment was lower than that in the normal fertilization treatment, especially those belonging to groups such as Actinomycetes, Burkholderia, Rhizobium, and Sphingomonases. Figure 11 and Figure 12Consistent with this, compared to the control, the number and diameter of root nodules were significantly reduced in the NK treatment, indicating that phosphorus deficiency reduced the symbiotic nitrogen-fixing capacity of soybeans. Sphingomonas and Burkholderia have close interactions with plants, possessing multiple functions including nitrogen fixation and pathogen inhibition, and can rapidly assimilate plant root exudates. Phosphorus deficiency led to a decrease in the abundance of these rhizosphere bacterial groups, indicating that long-term lack of phosphorus fertilizer application decoupled the beneficial interactions between plants and microorganisms. Therefore, maintaining a reasonable soil phosphorus nutrient level is crucial for ensuring the ecological function of black soil.

[0366] Compared to normal fertilization, soybean yield in the NK treatment was reduced by about a quarter. On the one hand, in soils without phosphate fertilizer, microbial functional genes encoding inorganic phosphorus dissolution and phosphorus starvation responses were specifically upregulated. Figure 20 D), and these genes are ubiquitous in metagenome-assembled genomes (MAGs). Figure 19 On the other hand, other symbiotic microorganisms, such as arbuscular mycorrhizal fungi, may also promote the uptake of phosphorus nutrients by soybeans. The lower available phosphorus content in the soil may be due to the rapid turnover of phosphorus nutrients from microbial dissolution to plant uptake. Furthermore, metagenomic and FARPROTAX functional annotation results indicate that N cycling processes, including nitrification and denitrification, are significantly enriched in NK treatments. Figure 17 and Figure 20 D). Due to the stoichiometric characteristics of plant nutrient uptake, NK treatment may lead to nitrogen excess, thereby increasing nitrogen transformation processes in the rhizosphere, consistent with reports based on soil results.

[0367] Unlike phosphate fertilizers, rhizosphere bacterial communities are more sensitive to nitrogen fertilizers, while potassium application has no significant effect on most bacterial groups and functional genes. Figure 17 and Figure 20 D). The rhizosphere bacterial beta distance between the PK and normal fertilization treatments showed a high difference in the early stages of plant growth and remained stable throughout plant development; while the beta distances between the NK and NP treatments and the normal fertilization treatment showed a continuous increasing trend with plant growth, especially in the NK treatment. Figure 6 These results indicate that although nitrogen fertilizer did not significantly affect the turnover rate of rhizosphere bacteria, it induced a unique developmental pattern in the rhizosphere bacterial community, possibly due to the strong influence of nitrogen on the biochemical processes of soils in different ecosystems. Compared with normal fertilization treatments, the functional α-diversity of soybean rhizosphere microorganisms was significantly reduced in the PK treatment, while there were no significant differences between the NK and NP treatments and the control. Figure 20Furthermore, CPCoA results showed that, except for the NP treatment, the functional diversity of rhizosphere microorganisms differed significantly between the other fertilization treatments and the control. Figure 20 AC). It was also found that the abundance of nitrogen-sustaining functional genes in rhizosphere bacteria increased in PK treatment, while the abundance of nitrogen-reduction-related genes decreased. Figure 20 D). The increased abundance of genes involved in nitrogen fixation and organic nitrogen mineralization indicates that the rhizosphere microbial community promotes functional adaptation between plants and microorganisms, thereby enhancing plant phenotypic plasticity. Similarly, compared to no fertilization, long-term application of nitrogen, phosphorus, and potassium fertilizers significantly reduced nitrogen fixation activity in both rhizosphere and non-rhizosphere soils. Furthermore, long-term nitrogen application also reduced nitrogen metabolism processes in soil archaea while increasing bacterial denitrification, nitrate reduction assimilation, and organic nitrogen metabolism. Interestingly, there were almost no significant differences in microbial taxa and functional genes between the NP treatment and the normal fertilization treatment, but genes involved in potassium transport were significantly enriched in the rhizosphere of the NP treatment, particularly the Kup gene (…). Figure 20 D) Kup is a protein located within the cell membrane used by microorganisms to take up potassium ions, indicating that rhizosphere microorganisms have a strong resilience to potassium supply. These results based on microbial community composition and function suggest that the rhizosphere microbial community can regulate multiple functions to promote the adaptation of plants and themselves in different environments.

[0368] 4. The role of key bacterial communities in maintaining soybean growth

[0369] Rhizosphere microorganisms maintain their own growth and metabolism through interactions, and enhance the host's adaptability to the environment. Similar to the bacterial community composition results, the rhizosphere bacterial co-occurrence network responded more to fertilization than the intraroot microbial network, and the complexity of the rhizosphere bacterial network was significantly higher in the PK and NP treatments than in the normal fertilization treatment. Figure 21 A), and the bacterial network complexity of the PK treatment in the roots was also higher than that of the control ( Figure 22 It has been reported that microbial communities may enhance interactions under environmental disturbances, such as nutrient deficiency and pathogen invasion. Under nitrogen-deficient conditions, microbial communities strengthened their interaction networks. A study on microbial interactions also showed a high degree of metabolic cooperation among microorganisms under low-nutrient conditions. The complexity and stability of networks are important characteristics of ecosystem function, and the increase in these characteristics in PK and NP treatments suggests that enhanced metabolic interactions among microorganisms may be a potential cause of stress mitigation. Although the abundance of Rhizobium and Sphingosine Monoclonal bacteria did not change significantly between NP and normal fertilization treatments, NP treatment significantly enhanced the connectivity of these nodes in the network. Figure 21C). This may be due to the close cooperation among potassium-solubilizing bacteria, thereby promoting the activation of soil potassium. Unlike the higher network complexity in the PK and NP treatments, the NK treatment did not significantly affect network complexity, but it did increase network stability. Figure 21 B). Given the low phosphorus availability in soil treated with NK, it is speculated that there may be strong competition between bacteria and plants for phosphorus assimilation, which to some extent limits microbial biomass and interactions between microorganisms.

[0370] The rhizosphere microbial community structure is the result of a series of interactions and feedbacks between plants and the soil environment, playing a vital role in plant growth, development, and health. Therefore, revealing plant-microbe interactions related to plant phenotypes is of great significance for sustainable agricultural development. The application of plant growth-promoting bacteria is considered an effective measure to improve agricultural productivity and soil and environmental health. Although significant progress has been made in the field of plant rhizosphere microbiome research, the use of microbiome technologies to improve crop growth and development is still in its early stages. Analysis of the main ecological clusters (network modules) within the core microbial groups revealed high positive correlations within these modules, but significant differences in abundance existed between different treatments. Figure 23 (A and B). Different ecological clusters may possess specific metabolic functions and ecological attributes, thus revealing key species in microbial networks can provide further guidance for the utilization of beneficial microorganisms. Key species refer to microorganisms that establish more connections with other microorganisms in a microbial network, some of which play a crucial role in the construction of the entire network. In this invention, bacterial co-occurrence network analysis was used, followed by bacterial isolation and purification, to ultimately design bacterial synthetic communities (SynComs) from different ecological clusters. Five bacteria were screened and identified from PK-enriched module 2 as experimental synthetic communities: Rhodococcus sp., Lysobacter sp., Terrabacter sp., Arthrobacter sp., and Phyllobacterium sp. ( Figure 23 The pot experiment results showed that the synthetic bacterial community in Module 2 significantly promoted soybean growth; however, the synthetic bacterial community in Module 1 had no significant effect on soybean growth. Figure 25 The growth-promoting effect of the ecological clusters enriched in the PK treatment indicates that the rhizosphere bacterial community, excluding rhizobia, also has a significant growth-promoting function for soybeans.

[0371] Preservation of biological materials

[0372] The representative cell line of this invention, PK167 (Rhodococcus sp.), has been deposited at the China Center for Type Culture Collection (CCTCC; Wuhan University, Wuhan, China) on January 2, 2024, with accession number CCTCC M2024006. It has been confirmed as viable by the collection center.

[0373] The representative cell line of this invention, PK260 (Lysobacter sp.), has been deposited at the China Center for Type Culture Collection (CCTCC; Wuhan University, Wuhan, China) on January 2, 2024, with accession number CCTCC M2024011. It has been confirmed as viable by the collection center.

[0374] The representative cell line of this invention, NPK195 (Terrabacter sp.), has been deposited at the China Center for Type Culture Collection (CCTCC; Wuhan University, Wuhan, China) on January 2, 2024, with accession number CCTCC M2024007. It has been confirmed as viable by the collection center.

[0375] The representative cell line of this invention, NK218 (Arthrobacter sp.), has been deposited at the China Center for Type Culture Collection (CCTCC; Wuhan University, Wuhan, China) on January 2, 2024, with accession number CCTCC M2024009. It has been confirmed as viable by the collection center.

[0376] The representative cell line of this invention, PK237 (Phyllobacterium sp.), has been deposited at the China Center for Type Culture Collection (CCTCC; Wuhan University, Wuhan, China) on January 2, 2024, with accession number CCTCCM 2024010. It has been confirmed as viable by the collection center.

[0377] The representative cell line of this invention, PK216 (Bosea sp.), has been deposited at the China Center for Type Culture Collection (CCTCC; Wuhan University, Wuhan, China) on January 2, 2024, with accession number CCTCC M2024008. It has been confirmed as viable by the collection center.

[0378] The representative cell line of this invention, NPK112 (Aeromicrobium sp.), has been deposited at the China Center for Type Culture Collection (CCTCC; Wuhan University, Wuhan, China) on January 2, 2024, with accession number CCTCCM 2024005. It has been confirmed as viable by the collection center.

[0379] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims. Furthermore, all documents mentioned in this invention are incorporated herein by reference as if each document were individually incorporated by reference.

Claims

1. Artificially synthesized microbial community, selected from synthetic microbial community SynCom5 or SynCom7; The SynCom5 strain includes the following strains: Rhodococcus sp., Lysobacter sp., Terrabacter sp., Arthrobacter sp., Phyllobacterium sp.; The SynCom7 includes: The strains of the synthetic microbial community SynCom5, and Bosea sp. and Aeromicrobium sp.; The Rhodococcus sp. has the accession number CCTCC M 2024006 at the China Center for Type Culture Collection. Lysobacter sp. has the accession number CCTCC M 2024011 at the China Center for Type Culture Collection. Terrabacter sp. has the accession number CCTCC M 2024007 at the China Center for Type Culture Collection. Arthrobacter sp. has the accession number CCTCC M 2024009 at the China Center for Type Culture Collection. Phyllobacterium sp. has the accession number CCTCC M 2024010 at the China Center for Type Culture Collection. Bosea sp. has the accession number CCTCC M 2024008 at the China Center for Type Culture Collection. Aeromicrobium sp. has the accession number CCTCC M 2024005 at the China Center for Type Culture Collection.

2. The artificially synthesized microbial community as described in claim 1, characterized in that, The artificially synthesized microbial community is contained in the composition.

3. The artificially synthesized microbial community as described in claim 2, characterized in that, The composition is a solution or suspension culture medium.

4. A composition comprising, or being a culture of, any one of the synthetic microbial communities according to claims 1-3.

5. The composition according to claim 4, characterized in that, The composition also includes an agriculturally acceptable carrier or culture medium component.

6. The composition according to claim 4, characterized in that, The composition includes: a pesticide composition.

7. The composition according to claim 4, characterized in that, The dosage forms of the composition include: solutions, suspensions, emulsions, oily or aqueous dispersions, powders, granules, and microcapsules.

8. The composition according to claim 4, characterized in that, The composition is a culture medium containing the artificially synthesized microbial community.

9. The application of the artificially synthesized microbial community according to claim 1, for promoting the growth of legumes, increasing biomass, increasing growth hormone levels, and improving ACC deaminase activity; or, for preparing compositions or kits that promote the growth of legumes, increase biomass, increase growth hormone levels, and improve ACC deaminase activity.

10. The use of the composition according to any one of claims 4-8 for promoting the growth of leguminous plants, increasing biomass, increasing growth hormone levels, and improving ACC deaminase activity; or for preparing a kit for promoting the growth of leguminous plants, increasing biomass, increasing growth hormone levels, and improving ACC deaminase activity.

11. A method for promoting the growth of leguminous plants, increasing biomass, increasing growth hormone levels, and enhancing ACC deaminase activity, comprising: The artificially synthesized microbial community according to any one of claims 1-3 or the composition according to any one of claims 4-8 is applied to a legume.

12. A kit for promoting the growth of leguminous plants, increasing biomass, increasing growth hormone levels, and enhancing ACC deaminase activity, comprising: The artificially synthesized microbial community according to claim 1 or the composition according to any one of claims 4-8.

Images