Diversity and Functional Differences in Soil Bacterial Communities in Wind–Water Erosion Crisscross Region Driven by Microbial Agents

Abstract

Soil erosion-prone areas require effective microbial treatments to improve soil bacterial communities and functional traits. Understanding the driving effects of different microbial interventions on soil ecology is essential for restoration efforts. Single and combined microbial treatments were applied to soil. Bacterial community structure was analyzed via 16S IRNA high-throughput sequencing, and functional groups were predicted using FAPROTAX. Soil microbial carbon, nitrogen, metabolic entropy, and enzymatic activity were assessed. Microbial Carbon and Metabolic Activity: The Arbuscular mycorrhizal fungi (AMF) and Bacillus mucilaginosus (BM) (AMF.BM) treatment exhibited the highest microbial carbon content and the lowest metabolic entropy. The microbial carbon-to-nitrogen ratio ranged from 1.27 to 3.69 across all treatments. Bacterial Community Composition: The dominant bacterial phyla included Firmicutes, Proteobacteria, Acidobacteria, Bacteroidetes, and Actinobacteria. Diversity and Richness: The AMF and Trichoderma harzianum (TH) (AMF.TH) treatment significantly reduced diversity, richness, and phylogenetic diversity indices, while the AMF.BM treatment showed a significantly higher richness index (p < 0.05). Relative Abundance of Firmicutes: Compared to the control, the AMF, TH.BM, and TH treatments decreased the relative abundance of Firmicutes, whereas the AMF.TH treatment increased their relative abundance. Environmental Correlations: Redundancy and correlation analyses revealed significant correlations between soil organic matter, magnesium content, and sucrase activity and several major bacterial genera. Functional Prediction: The AMF.BM treatment enhanced the relative abundance and evenness of bacterial ecological functions, primarily driving nitrification, aerobic ammonia oxidation, and ureolysis. Microbial treatments differentially influence soil bacterial communities and functions. The AMF.BM combination shows the greatest potential for ecological restoration in erosion-prone soils.

1. Introduction

Soil remediation (“remediation” refers to targeted interventions to restore soil health by enhancing microbial activity, nutrient availability, and structural stability in erosion-degraded soils) through various biological agents, primarily plants and microorganisms, is considered among the most economical and safest remediation methods. Soil microbial communities play pivotal roles in maintaining ecosystem functions, including nutrient cycling and organic matter stabilization [1,2]. For example, AMF and Bacillus spp. enhance soil aggregation and carbon sequestration in erosion-vulnerable regions [3,4]. The response and remediation functions of highly complex microbial communities in soil ecology under extreme climatic conditions, such as drought, have been a focal point in soil microbiology research [5]. The semi-arid region of northwest Liaoning is characterized by a complex terrain, including both hydraulic and wind erosion, poor physical and chemical soil properties, a sparse microbial flora, and low microbial activity, all of which contribute to the formation of an “ecologically fragile zone” [6,7]. Early vegetation improvement techniques in this area primarily focused on adjusting soil structure and enhancing physical and chemical properties through the selection of dominant plants to achieve soil remediation [8]. However, natural degradation of overall soil quality under “double force” erosion, combined with anthropogenic factors such as overuse, leads to poor plant survival rates or physiological decline after planting, thereby reducing remediation effectiveness [9].

The appropriate use of microbial agents can enhance the structure and function of soil microbial communities; improve soil fertility and utilization; promote plant survival, growth, and development; and establish a positive mutual feedback and synergistic effect between soil, microbes, and plants. Arbuscular mycorrhizal fungi (AMF), Trichoderma harzianum (TH), and Bacillus mucilaginosus (BM) are commonly used agents in soil remediation [10]. Analyzing the effects of externally applied microbial active substances on soil microbial communities and functions serves as an essential foundation and prerequisite for microbial utilization. In dual hydraulic–wind erosion zones, elucidating the mechanisms driving microbial community structure and functional shifts—particularly under exogenous microbial inoculation—remains a critical research priority for advancing soil remediation strategies [11,12]. The external application of microbial agents can alter the structure and function of the soil microbial community. Studies have shown that the relative abundance of Actinomycetes, Chlorobacterium, Bacillariophyta, and Ascomycetes in soil increased after rice fertilization [13]. In contrast, the addition of microbial agents to salinized soils can significantly increase the relative abundance of Ascomycota and decrease the abundance of Basidiomycota, Chytridiomycota, Mortierellomycota, and other phyla [6]. Following inoculation with AMF, the functional diversity of soil microorganisms was significantly enhanced; the proportion of fungi increased, while the proportions of bacteria and actinomycetes decreased. Additionally, the utilization intensity of soil microorganisms for amines and polymers increased [14]. Trichoderma harzianum (TH) effectively improved the microbial community structure in the continuous cropping soil of Populus japonica, altered the dominant pattern of pathogenic bacteria, increased the biomass of bacteria and actinomycetes, and stabilized the soil micro-ecosystem [15]. While the above studies have made progress in understanding the effects of microbial agents on soil microorganisms, the driving effects of AMF, BM, and TH on soil remediation and microbial community structure and function, when applied with Populus simonii Carr in the “double erosion” zone, remain unclear.

Based on the existing knowledge gaps, this study tested two central hypotheses:

• The combined application of AMF, BM, and TH with Populus simonii Carr will synergistically improve soil microbial community structure and function in the wind–water erosion zone, leading to enhanced enzyme activity and microbial biomass.
• The inoculants will differentially modulate the relative abundance of key bacterial phyla (e.g., Firmicutes and Actinobacteria) and fungal taxa, thereby driving soil remediation.

The study was designed to achieve three primary objectives:

• To quantify the effects of AMF, BM, and TH treatments on soil microbial diversity, enzyme activity, and biomass carbon/nitrogen in eroded soils.
• To identify the dominant microbial taxa and functional shifts induced by each inoculant combination using 16S sequencing.
• To elucidate the mechanistic links between inoculant-driven microbial changes and soil remediation potential.

Thus, this study selected the wind–water erosion intersecting region of Weizigou Town, Changwu County, Fuxin City, Liaoning Province, and utilized Populus simonii Carr as the improved vegetation to measure soil enzyme activity, microbial biomass carbon, and nitrogen properties under the combined application of AMF, BM, and TH. 16S high-throughput sequencing ecological analysis technology was employed to examine the soil microbial community structure and functional differences of Populus simonii Carr in the improved area. The driving effects of various inoculant application schemes on soil microbial community structure and function were analyzed, and the mechanisms of inoculant action were discussed, providing a theoretical reference for the advancement of microbial inoculant technology and its application in the wind–water erosion intersecting region.

2. Materials and Methods

2.1. Study Area

The study area was in Yaolinggangzi Village (42°26′15″ N, 122°48′38″ E), Weizigou Town, Zhangwu County, Fuxin City, Liaoning Province, which is located in the southeast of Horqin Sands, a semi-arid region. In the same period of rain and heat, the spatial and temporal distribution is uneven. The average annual rainfall is more than 550 mm. The rainfall from June to August is concentrated and lasts for a long time, constituting heavy rain, accounting for more than 70% of the annual average precipitation (744 mm). The area is cold and dry in autumn and winter, with little rain and strong and concentrated winds, and there are frequent gales in autumn and winter; the area is also prone to sandstorms, with an average wind speed of 3.7~4.6 m/s. The annual evaporation is >1200 mm, the dryness is in the range of 1.2~2.0, and the relative humidity is in the range of 48~78% (defined as the ratio of annual potential evapotranspiration to precipitation [16]). Wind and water erosion are serious, the soil type is brown loam, the nutrient content of agricultural land is low, and desertification is serious. The plants in the region are typical sandy grassland plants, for example, Semen pruni Humilis, Parochetus communis Buch, and Festuca ovina Linn. The plants were removed manually before the trial began. The physical and chemical properties of the soil are shown in

Table 1. Initial physicochemical properties of experimental soils.

Soil Microaggregates	2–0.25 mm	0.25–0.05 mm	0.05–0.01 mm	0.01–0.005 mm	0.005–0.001 mm	˂0.001 mm
	2.48%	51.20%	39.90%	3.34%	0.20%	2.88%
Soil mechanical composition	2.0–1.0 mm	1.0–0.5 mm	0.5–0.25 mm	0.25–0.05 mm	0.05-0.02 mm	0.02–0.002 mm	˂0.002 mm
	0.63%	1.02%	0.50%	48.95%	14.29%	15.21%	19.39%
Soil chemical properties	SOM %	TN g·kg−1	TP g·kg−1	TK g·kg−1	Ca g·kg−1	Mg g·kg−1	UR mg·g−1
	2.49	1.04	0.27	20.43	65.78	8.27	7.72
	SU mg·g−1	DH mg·g−1	CA mg·g−1	AP mg·g−1	MBC mg·g−1	MBN mg·g−1
	41.33	223.31	0.59	8.56	305.81	93.39
Soil physical properties	SWC %	pH	EC mS cm−1
	20.51	6.83	114.63

, and the basic characteristics of the soil after treatment with different microbial agents are presented in

Table 2. Basic properties of soil treated with different microbial agents.

Sample	CK	AMF	TH	BM	AMF.BM	AMF.TH	BM.TH	AMF.BM.TH
SOM %	2.06 abc	2.91 d	2.07 abc	1.83 a	1.87 ab	2.18 c	2.23 c	2.09 c
TN g·kg−1	0.85 e	0.94 de	0.92 de	0.97 cde	1.04 cd	1.11 bc	1.35 a	1.21 ab
TP g·kg−1	0.22 cd	0.25 c	0.25 c	0.19 d	0.21 cd	0.29 b	0.42 a	0.20 d
TK g·kg−1	16.75 b	17.27 ab	18.32 ab	18.86 ab	17.58 ab	18.17 ab	19.65 a	17.75 ab
Ca g·kg−1	53.48 d	51.62 c	47.31 b	56.18 e	57.95 e	45.97 b	61.04 f	41.08 a
Mg g·kg−1	6.57 cd	5.68 bc	5.51 bc	5.72 bc	10.09 e	5.41 b	4.21 a	7.29 d
UR mg·g−1	6.33 a	6.68 b	6.99 c	7.74 d	8.53 e	7.04 c	7.04 c	7.71 d
SU mg·g−1	33.07 a	60.95 d	56.82 d	30.68 a	44.72 c	39.04 b	39.76 bc	28.92 a
DH mg·g−1	184.55 e	257.86 g	188.06 f	104.92 a	129.22 b	138.49 c	141.63 d	138.21 c
CA mg·g−1	0.48 ab	0.54 b	0.30 a	1.01 d	1.34 e	1.03 d	0.72 c	1.15 d
AP mg·g−1	6.69 a	16.83 b	16.82 b	19.21 b	33.71 c	16.88 b	9.95 ab	12.36 ab
MBC mg·kg−1	250.66 a	400.13 ab	690.46 c	439.17 ab	907.38 d	562.94 bc	339.91 ab	296.44 a
MBN mg·kg−1	77.19 a	183.45 b	370.00 d	346.79 cd	337.49 c	186.60 b	92.16 a	189.01 b
SWC %	16.67 abc	15.62 a	16.17 a	17.94 d	16.15 a	17.39 bcd	16.34 ab	17.76 cd
pH	6.71 c	6.93 ab	6.92 ab	6.79 bc	6.67 c	6.93 ab	6.82 bc	7.05 a
EC mS cm−1	105.28 e	64.56 a	66.03 a	94.71 d	131.27 g	74.15 b	126.36 f	76.98 c

Note: Different lowercase letters in the same line indicate significant differences among treatments (p < 0.05), and the same below. AMF = Arbuscular mycorrhizal fungi; TH = Trichoderma harzianum; BM = Bacillus mucilaginosus; SOM = Soil organic matter; TN = Total nitrogen; TP = Total phosphorus; TK = Total potassium; Ca = Calcium; Mg = Magnesium; UR = Urease activity; SU = Sucrase activity; DH = Dehydrogenase activity; CA = Catalase activity; AP = Alkaline phosphatase activity; MBC = Microbial biomass carbon; MBN = Microbial biomass nitrogen; SWC = Soil water content; pH = Soil pH; EC = Electrical conductivity.

.

2.2. Experimental Materials and Design

Arbuscular mycorrhizal fungi (AMF), Trichoderma hartz (TH), and Bacillus glia (BM) were selected as the test agents. TH was purchased from Green Long Biologicals at a concentration of 1.0 × 10¹⁰ CFU·g⁻¹; BM was purchased from Green Long Biologicals at a concentration of 2.0 × 10¹⁰ CFU·g⁻¹; and AMF was selected from Glomus mosseae, which has an excellent mycorrhizal effect, and was purchased from Shihezi University. As shown in Table 2, there were three types of experimental groups: a control group, microbial agent single application groups, and microbial agent mixed application groups. There were 8 treatments in total, and each treatment was repeated 3 times. Each replicate was set as 1 experimental plot with a total of 24 in a randomized group design. Each test plot was 2 m × 2 m in size, with 4 poplar plants planted within the plot, all spaced 1.0 m × 1.0 m. In order to prevent the migration of microbial agents in the plot, 2 m isolation belts were installed between the plots. According to the experimental scheme, two-year-old Populus simonii Carr (2–3 cm diameter) was planted in each plot in March 2020. After mixing the bacterial agent with soil 30 cm from the center of the planting site, it was buried and applied, and conventional artificial tending and management in the field were carried out in time to ensure the good and uniform growth of the plants.

2.3. Sample Collection

Microbial agents were applied in March 2020, and the soil was collected 90 days later. A five-point mixed sampling method was used in each plot; a 0–20 cm soil layer was collected, and sand, stone, plant, and animal debris were removed. The same treatment samples were thoroughly mixed, screened with a 100-mesh sieve (aperture: 1.5 mm), sealed and stored in the refrigerator, and brought back to the laboratory. One sample was stored at 4 °C, and the water content and enzyme activity were measured. One sample was frozen at −80 °C for DNA extraction and high-throughput sequencing. One sample was dried and ground to determine the soil’s physical and chemical properties.

2.4. Index

Determination of soil physical and chemical properties: Refer to “Soil Agrochemical Analysis” [17] for soil physical and chemical properties; organic matter was determined by the low-temperature external heat potassium dichromate oxidation–colorimetric method [18]. Soil total nitrogen was determined by a Kjeldahl nitrogen analyzer. Soil phosphorus, potassium, calcium, and magnesium indexes were determined by a plasma emission spectrometer (ICP-AES).

Soil enzyme activity [19]: Urease activity was determined by the sodium phenol colorimetric method. Sucrase activity was measured using the 3,5-dinitrosalicylic acid colorimetric method. Catalase activity was quantified by potassium permanganate titration. Dehydrogenase activity was assessed via the triphenyltetrazolium chloride (TTC) colorimetric method, where TTC is enzymatically reduced to formazan as an indicator of microbial metabolic activity.

Soil microbial biomass carbon and nitrogen were fumigated and unfumigated by the chloroform fumigation–extraction method, and 0.5 mol·L⁻¹ K₂SO₄ solution was used for extraction [20].

Microbial community diversity determination: Based on high-throughput sequencing technology, the composition and structure of the bacterial community was analyzed. Sample DNA was isolated and extracted by the CTAB method; the qualified samples were diluted to 1 ng/μL and identified with 16S V4 (not V3-V4) region primers (515F and 806R) for sequencing regions. PCR amplification was performed using barcode-specific primers with Phusion^® High-Fidelity PCR Master Mix with GC Buffer (New England Biolabs, Ipswich, MA, USA). Libraries were constructed using the TruSeq^® DNA PCR-Free Sample Preparation Kit (Illumina, San Diego, CA, USA). Following quantification using a Qubit fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) and quantitative PCR, libraries were sequenced on an Illumina NovaSeq 6000 platform (Illumina, San Diego, CA, USA).

2.4.1. Soil Physicochemical Properties

• Soil Organic Matter (SOM)
  Method: Modified Walkley–Black wet oxidation method [21].
  A quantity of 0.5 g air-dried soil was mixed with 5 mL 0.8 M K₂Cr₂O₇ and 5 mL concentrated H₂SO₄.
  The sample was heated at 135 °C for 30 min, cooled, and diluted with 10 mL deionized water.
• Total Nitrogen (TN)
  Method: Automated Kjeldahl digestion and distillation (FOSS Kjeltec 8400) [22].
  Digestion with H₂SO₄-H₂O₂ at 420 °C for 2 h.
  Distillation with 40% NaOH, absorbed in 2% boric acid, and titrated with 0.01 M HCl.
• Phosphorus, Potassium, Calcium, and Magnesium
  Method: Inductively Coupled Plasma Optical Emission Spectrometry [23].
  Soil digested with HNO₃-HF-HClO₄ (3:1:1) at 180 °C for 6 h.
  Wavelengths (nm): P 213.618, K 766.490, Ca 317.933, Mg 285.213.

2.4.2. Soil Enzyme Activities

• Urease
  Method: Phenol–sodium hypochlorite colorimetry [24].
  Incubation with 10% urea at 37 °C for 24 h.
  Color development with phenol (1.5%) and NaOCl (0.9%), measured at 635 nm.
• Sucrase
  Method: 3,5-Dinitrosalicylic acid (DNS) assay [25].
  Substrate: 8% sucrose solution, reacted at 37 °C for 24 h.
  Absorbance measured at 540 nm after DNS reagent addition.
• Dehydrogenase
  Method: Triphenyltetrazolium chloride (TTC) reduction [26].
  Soil incubated with 1% TTC (pH 7.4) at 37 °C for 6 h in darkness.
  Formazan extracted with methanol, measured at 485 nm. Activity expressed as μg Formazan·g⁻¹·h⁻¹.
• Phosphatase
  Method: p-Nitrophenyl phosphate (pNPP) hydrolysis [27].
  Substrate: 5 mM pNPP in acetate buffer (pH 5.5), incubated at 37 °C for 1 h.
  Reaction stopped with 0.5 M NaOH, measured at 405 nm.

2.4.3. Microbial Biomass Carbon and Nitrogen

Method: Chloroform fumigation–extraction (CFE) [28].

Fumigation: 25 g fresh soil fumigated with ethanol-free CHCl₃ for 24 h at 25 °C.

Extraction: 0.5 M K₂SO₄ (1:4 w/v), shaken at 200 rpm for 30 min.

Analysis:

Microbial biomass carbon (MBC): TOC analyzer (Shimadzu TOC-L), kc = 0.45.

Microbial biomass nitrogen (MBN): Micro-Kjeldahl method, kn = 0.54.

2.4.4. Microbial Community Diversity

NA Extraction

Method: CTAB-based extraction [29].

A quantity of 0.5 g soil was homogenized in CTAB buffer (2% CTAB, 100 mM Tris-HCl, 20 mM EDTA).

DNA purified using a QIAamp PowerSoil Pro Kit (Qiagen, Venlo, The Netherlands).

16S rRNA Sequencing

Primers: 515F (5′-GTGYCAGCMGCCGCGGTAA-3′) and 806R (5′-GGACTACNVGGGTWTCTAAT-3′) [30].

PCR: Phusion^® High-Fidelity PCR Master Mix (Thermo FisherScientific, Waltham, MA, USA), 30 cycles.

Sequencing: Illumina NovaSeq 6000 (2 × 250 bp).

Bioinformatics

Pipeline: QIIME2 v2023.2 [31].

Denoising with DADA2.

Taxonomic assignment using SILVA v138.1.

α-diversity (Shannon, Chao1) and β-diversity (PCoA, Bray–Curtis).

2.5. Statistics and Analysis

The experimental data were statistically collated using Excel 2013 and SPSS 26.0 software at a significant level of p < 0.05. The ratio of basal respiration intensity to microbial mass carbon was used to calculate soil microbial metabolic entropy (qCO₂). The velocity data were split, spliced, and filtered to obtain high-quality Clean Tags data, and chimaeras were removed to obtain the final Effective Tags data. Uparse v 7.0.1001 software was used to conduct OTU clustering for sequences with similarity ≥ 97%. The SSUrRNA of Mothur and SILVA 132 was used for species annotation. Shannon, Simpson, Chao1, ACE, and coverage indexes were obtained by **Qiime 1.9.1 (approximate version)**, and UPGMA clustering trees were constructed. The classification level was calculated using the RDP Classifier. The Spearman correlation coefficient was used to study the correlation between the physicochemical index and bacterial community. **FAPROTAX (v1.2.4, approximate version)** software was used to predict bacterial function.

3. Results

3.1. Effects of Different Bacterial Combinations on Soil Microbial Carbon and Nitrogen and Metabolic Entropy

The effects of different microbial agent treatments on microbial biomass carbon (MBC, mg·kg⁻¹ soil) and nitrogen (MBN, mg·kg⁻¹ soil) are shown in Figure 1. After TH, AMF.BM, and AMF.TH application, the MBC content in soil was significantly increased compared with CK, and TH application alone showed the highest content, which was 175.45% higher than that of CK. In mixed application, the content of the AMF.BM treatment was the highest, followed by that of the AMF.TH treatment, which were 261.99% and 124.58% higher than that of CK, respectively, and the difference was significant. Soil microbial biomass nitrogen (MBN) was positively affected by different bacterial agents. TH, BM, and AMF.BM treatments had the best improvement effects: 379.32%, 349.26%, and 337.21% higher than CK, respectively, while BM.TH had no significant improvement effect. The MBC/MBN values for all treatments (Figure 2) ranged from 1.27 to 3.69. Except for the BM.TH treatment, the other treatments showed significantly lower values than CK. The values, in descending order, were as follows: BM > AMF.BM.TH > TH > AMF > AMF.BM > BM.TH, with 60.92%, 51.69%, 42.46%, 32.92%, 17.23%, and 7.07%, respectively. The metabolic entropy statistics of the different microbial agent treatments showed that all the treatments except AMF.BM.TH had lower values than CK, and that of AMF.BM was the most significant, followed by TH, which decreased by 68.05% and 50.00% compared with CK, respectively.

3.2. Bacterial Sequencing Statistics and Rationality Analysis

Sequencing results revealed significant variation in high-quality sequence counts across treatments (Figure 3a), with AMF.BM yielding the highest (57,396 sequences) and BM.TH the lowest (36,639). All treatments exhibited a sequencing coverage > 0.98 (

Table 3. Experimental design and dosages of microbial agents.

Treatments	Sample Description	Application Rate (g·m−2)
Blank control	CK	—
Mono-microbial	AMF	50
	BM	10
	TH	20
Composite microbial	AMF.BM	50 + 20
	AMF.TH	50 + 10
	BM.TH	10 + 20
	AMF.BM.TH	50 + 10 + 20

Notes: “—” = No microbial agent application; CK = plant alone; AMF = plant + AMF; BM = plant + BM; TH = plant + TH; AMF + BM = plant + AMF + BM; AMF + TH = plant + AMF + TH; TH + BM = plant + TH + BM; AMF + TH + BM = plant + AMF + TH + BM.

), confirming near-complete representation of bacterial diversity. Rarefaction analysis (Figure 3b) showed OTU richness plateauing at ~2500 sequences per sample, indicating sufficient sequencing depth to capture community diversity.

3.3. Effects of Different Microbial Agents on Soil Microbial Community Structure

The top 10 species groups in terms of relative abundance (Figure 4a) were Firmicutes, Proteobacteria, Acidobacteria, Bacteroidetes, Actinobacteria, Verrucomicrobia, Thaumarchaeota, Gemmatimonadetes, Chloroflexi, and Nitrospirillum. In the dominant phyla with a relative abundance > 5%, the relative abundance of Firmicutes was significantly increased by AMF.TH treatment to 42.45%, which was significantly increased by 201.92% compared with CK. In Proteobacteria, the relative abundance of the AMF.BM treatment was significantly increased by 31.56% compared with CK. The relative abundance for the TH treatment was lower than that for CK, reducing by 64.41%. Among Acidobacteria, the increase in the AMF treatment was the highest, increasing by 33.72% compared with CK. The relative abundances in the AMF.BM, AMF.TH, and AMF.BM.TH treatments were lower than in CK. Among Bacteroidetes, the proportion of relative abundance in the AMF.TH treatment was the highest, which increased by 192.25% compared with CK. The results for AMF, AMF.BM, and BM.TH were lower than for CK. The abundance of Acidobacteria under different treatments ranged from 10.85% to 13.98%, except for the AMF and TH treatments, which only slightly increased compared to CK, by 0.30% and 5.67%, while all others were lower than CK. The most significant increase in the abundance of Verrucomicrobia was achieved by the BM.TH treatment with 11.67%, 98.80% higher than CK, followed by TH and AMF.BM.TH. Overall, the bacterial community composition was similar between treatments, but its abundance differed between fertilization treatments.

The relative abundance of species at the genus level is shown in Figure 4b. Apart from the combined abundance of <0.5% for others, the relative abundance at the TOP 30 genus level was higher than that of CK (26.02%) for only two mixed treatments, AMF.TH (70.08%) and AF.BM (32.91%), and the difference was significant, while all other treatments failed to play a driving role. The analysis of the top 10 dominant genera showed that AMF.TH treatment significantly improved the abundance of dominant species, including Lactobacillus, Bacteroide, and Bifidobacterium, accounting for 49.95%. Other non-dominant species, such as Collinsella, Romboutsia, and Turicibacter, accounted for 20.13%. The overall relative abundance increased by 2.24 times compared with CK. The six dominant species in AMF.BM, Lactobacillus, Candidatus_Udaeobacter, Panacagrimonas, Enhydrobacter, Acinetobacter, and Pseudomonas, were largely similar in abundance, with a combined percentage of 17.94%. The sum of non-dominant species, such as Bacteroides, Bifidobacterium, and Romboutsia, accounted for 14.25%. Compared with CK, the overall relative abundance increased by 1.26 times. Overall, the community composition of AMF.BM and AMF.TH was similar to that of other treatments at the genus level, but there were significant differences in the abundance proportions of the dominant genera.

3.4. Analysis of Soil Bacterial Community Diversity

As shown in the petal diagram below (Figure 5), the total number of OTUs in the eight treatments was 848, accounting for 40.51% of the total OTUs. The unique OTUs of each treatment ranged from 64 to 190, and the AMF.BM treatment had the highest number, accounting for 18.30% of the total. The treatment with the lowest number was AMF.TH, accounting for only 5.14%. The common OTUs of all treatments were significantly greater than the unique number, indicating that the dominant species of the bacterial community structure in the different treatments were basically similar.

An alpha diversity analysis is shown in

Table 4. Alpha diversity indices of soil bacteria under different treatments.

Treatments	Shannon	PD_Whole_Tree	Chao1	ACE	Coverage
CK	9.566 cd	197.551 bc	3007.171 bc	3062.13 bc	0.986 a
AMF	9.533 cd	232.647 d	2826.562 b	2878.51 b	0.988 a
BM	9.536 cd	199.09 bc	2898.360 bc	2956.51 bc	0.987 a
TH	9.650 d	193.08 b	2952.173 bc	2996.06 bc	0.987 a
AMF.BM	9.247 b	226.75 cd	3090.625 c	3075.32 c	0.982 a
AMF.TH	6.046 a	131.898 a	1894.853 a	2013.79 a	0.989 a
BM.TH	9.428 bc	222.012 bcd	3000.769 bc	3039.83 bc	0.983 a
AMF.BM.TH	9.558 cd	247.878 d	3006.216 bc	3063.75 bc	0.986 a

Note: Different lowercase letters in the same line indicate significant differences among treatments (p < 0.05).

; the Shannon index was the highest in the TH treatment and significantly different from that of CK (p < 0.05), and the bacterial species diversity in this treatment was higher. Both Chao1 and ACE indices were the highest in the AMF.BM treatment; the bacterial community diversity in this treatment was relatively richer. The PD_whole_tree index of the AMF treatment was the highest, and the phylogenetic evolutionary differences in this treatment were higher, while all indices of the AMF.TH treatment were significantly lower (p < 0.05) compared to CK, and the species diversity and richness were lower.

To investigate the similarity of the bacterial communities in the different treatments, the eight treatments were subjected to UPGMA cluster analysis (Figure 6). The bacterial community composition was divided into four categories when the Unifrac distance was 0.2. The bacterial community structures of the CK, BM, and AMF.BM.TH treatments were more similar and clustered into one category; AMF, TH.BM, and TH were more similar in structure and clustered into one category, and all the thick-walled phylum (Firmicutes) was lower than in CK; while AMF.BM and AMF.TH were independently clustered into one category, indicating that the bacterial community structure and similarity between the two treatments were not high, with the relative abundance of Proteobacteria in AMF.BM treatment being significantly higher than that in CK, while the relative abundance of Firmicutes in the AMF.TH treatment was the highest and that of Proteobacteria the lowest.

3.5. Correlation Analysis Between Community Structure of Soil Bacteria and Environmental Factors in Different Treatments

Redundancy analysis (dbRDA) (Figure 7) showed the relationship between soil bacterial communities at the phylum taxonomic level and soil nutrients under different treatments. Soil nutrient indicators explained 95.11% of the variation in soil bacterial community structure relationships, with the first and second ordination axes explaining 84.63% and 10.48%, respectively; Mg, TP, and TK were the key nutrient factors affecting the bacterial community. AMF, TH, and BM.TH were positively correlated with SOM and explained changes in Acidobacteria, Verrucomicrobia, and Gemmatimonadetes. AMF.TH mainly explained the changes in Firmicutes and Bacteroidetes. AMF.BM.TH was significantly correlated with TP and TN and explained the relationship between Thaumarchaeota and Firmicutes. AMF.BM significantly affected Mg content and explained the changes in Chloroflexi and Proteobacteria. TK and Ca mainly explained the species changes in Proteobacteria.

The results of the dbRDA analysis with microbial properties showed that the indicators explained 91.15% (80.94% for dbRDA1 and 10.21% for dbRDA2) of the variation in the soil bacterial community, with MBC, CA, and UR being the key factors affecting the bacterial community. The AMF.BM and CK treatments with SU and MBN were in the first phase quadrant, associated with Proteobacteria and Chloroflexi. UR, MBC, CA, and AP were in the second quadrant, associated with Firmicutes and Actinobacteria. The third phase quadrant included TH and AMF.BM.TH, which were associated with changes in the Bacteroidetes. The AMF, TH, and BM.TH treatments and DH were in the fourth quadrant and were associated with Acidobacteria, Verrucomicrobia, Bacillariophyta, and Gemmatimonadetes.

The data with significant correlations at the genus level of the top 35 abundance ratios were analyzed (Figure 8). SOM was negatively correlated with Chryseobacterium, Bacillus, unidentified_Enterobacteriaceae, and Pseudomonas. It was positively correlated with Staphylococcus. TP was negatively correlated with Gemmatimonas and Unidentified Gammaproteobacteria. TK was positively correlated with Brevundimonas, Kocuria, and Psychrobacter, but negatively correlated with Turicibacter. Mg was associated with Chryseobacterium, Unidentified Enterobacteriaceae, Bacillus, Pseudomonas, and Acinetobacter, and Enhydrobacter were significantly positively correlated, while Bifidobacterium was significantly negatively correlated.

For soil enzyme activity, SU was significantly positively correlated with duboisella, Gaiella, Bradyrhizobium, and Unidentified-Acidobacteria and negatively correlated with Romboutsia. DH was significantly positively correlated with Staphylococcus and negatively correlated with Lactobacillus. CA was significantly negatively correlated with Gemmatimonas and Unidentified-Gammaproteobacteria and significantly positively correlated with Lactobacillus. MBC was significantly negatively correlated with bryobacter, Microlunatus, and Candidatus-udaeobacter, while MBN was significantly positively correlated with Panacagrimonas. In addition, SWC was significantly positively correlated with Turicibacter.

Variance partitioning canonical correspondence analysis showed (Figure 9) that the variance contribution to the microbial bacterial diversity data for each type of environmental factor set for the different bacterial agent treatments was 100%, with the soil physical and chemical properties factor contributing 18.81% alone. The contribution of microbial properties alone was 35.45%, and the contribution of the joint effect of the two factors was 45.74%. These results indicated that both groups of environmental factors significantly affected soil microbial community diversity, and the contribution rate of soil microbial property factors was higher than that of soil physical and chemical factors, and the synergistic effect of the two factors was more significant.

The FAPROTAX tool was used to annotate community functions, and a total of 69 functional groupings were obtained, excluding species that had not yet been identified, with large differences in community functions between treatments (Figure 10). Only the AMF.TH and AMF.BM treatments had a higher relative abundance of functions than CK.

A total of 59 bacterial functions were annotated by AMF.BM treatment. The total of the top 10 community functions was 51.63%, which included chemoheterotrophy, aerobic-chemoheterotrophy, and fermentation functions. The sum of the abundance of the remaining 49 non-dominant functions was 12.83%. A total of 48 functions were annotated by AMF.TH treatment, and the total abundance of the top 10 treatments was 84.59%, in which chemoheterotrophy and fermentation had the highest proportions, 29.58% and 28.45%, respectively. The remaining abundance of 43 accounted for only 1.71%. Although the overall abundance was highest, the annotation had less functional information and poor uniformity in structure. In conclusion, the functional richness and evenness of AMF.BM processing were significantly better than those of AMF.TH processing.

The top 25 functions in terms of abundance and information on their abundance in each sample were heat-mapped and clustered at the level of functional differences (Figure 10). Eight treatments were shown to have different levels of ability to drive community function, with significant effects occurring in two treatments, AMF.TH and AMF.BM, and relatively diminished microbial community functions in the other treatments.

All the treatment functions can be classified into four categories. The first included fermentation, chemoheterotrophy, human gut, mammal gut, and six anaerobic functions, which were significantly affected by AMF.TH, and their abundance was significantly higher than in the other treatments. The second category included nitrification, aerobic–ammonia–oxidation, and ureolysis and five aerobic functions, which were mainly driven by AMF, BM.TH, and AMF.BM.TH. The third category of plant pathogen functions included nitrite respiration and nitrite ammonification, and the fourth category included nitrate respiration. A total of 13 bacterial functions, including nitrite_respiration, nitrite ammonification and class IV nitrate respiration, nitrate reduction, and nitrogen respiration, were subject to a greater functional driving effect of AMF.BM, and the relative abundance of the community was higher than in the other treatments. The clustering of the annotation function of soil bacterial communities among all treatments was significantly consistent with the UPGMA clustering analysis of bacterial communities (Figure 11).

4. Discussion

4.1. Effects of Combined Application of Different Microbial Agents on Soil Microbial Carbon and Nitrogen

Soil microbial biomass carbon and nitrogen were more sensitive to the addition of microbial agents, and MBC was an important reference indicator of soil microbial biomass [32]. In this study, the highest MBC content was found in the AMF.BM treatment, followed by the AMF.TH treatment, indicating that the soil fertility and microecology of the two treatments were good and that the soil microbial quantity was large, which was consistent with the sequence number results obtained from sequencing. MBN is a comprehensive reflection of the ability of microorganisms to transform nitrogen in soil, and it is also one of the important nitrogen pools in soil. It mainly comes from the decomposition of soil organic nitrogen by soil microorganisms and the nitrogen fixation of organisms with biological nitrogen fixation functions [33]. In this study, MBN was significantly correlated with TN, MBC, AP, and other indicators, which meant that MBN was closely related to various environmental factors such as fertility, microbial biomass, and enzymes related to nitrogen metabolism, which was consistent with previous research results [34]. In addition to the AMF.BM.TH treatment, other treatments significantly increased the MBN content, and the differences were significant. The combined application of the three bacterioides might make the total amount of application in the soil larger, resulting in intraspecific competition and high consumption of soil carbon and nitrogen nutrients. Therefore, in the process of adding bacterioides, the combination of different bacterioides and the dosage of bacterioides should be paid attention. To some extent, MBC/MBN can express the community structure and flora of soil microorganisms with a low ratio and a large bacterial dominance, with bacterial values ranging from 3 to 5 and fungal values ranging from 4 to 15 [35]. In this study, the MBC/MBN values of all samples were lower than 5, indicating that the bacteria in the soil of all the samples were dominant fungi. Microbial metabolic entropy effectively combined MBC with microbial activity, which could reflect the survival energy consumption of soil microorganisms and the utilization of the soil matrix. In general, the higher the soil physical and chemical properties or soil enzyme activity, the less energy is consumed by microorganisms to maintain the same biomass, and the soil metabolic entropy gradually decreases [36]. In this study, the microbial metabolic entropy of the AMF.BM, AMF.TH, and TH treatments was significantly lower than that of the control, with that of AMF.BM being the lowest. The results showed that improved AMF.BM had good effects on the recovery of soil enzyme activities and soil nutrients.

Microbial diversity plays a key role in maintaining soil ecological health [37]. This study showed that the combined application of various bacteria did not damage the composition of dominant phyla of the soil bacterial community, and the composition of the top 10 dominant phyla of the soil bacterial community was basically consistent with that of other similar areas studied [38,39]. However, it changed the relative abundance of each dominant phylum. Compared with CK, AMF.BM improved the relative abundance of Proteobacteria, which participated in the carbon and nitrogen cycle of the soil and had a good effect on carbon and nitrogen fixation, and also increased the solubility of phosphate in the soil, which had a positive effect on stabilizing and maintaining soil ecological environment balance.

However, the AMF.TH treatment significantly reduced the diversity, richness, and evenness of bacteria, specifically by reducing the abundance of Proteobacteria and increasing the relative abundance of Firmicutes. Most bacteria in firmicutes have the function of biological disease control [40], while AMF and TH have similar ecological functions, which may be related to their driving effects on Firmicutes [41,42]. Studies have shown that the application of BM in cultivation areas of the Chinese herbal medicine hemlock can increase the relative abundance of Proteobacteria in soil bacteria [43]. In this study, BM did not significantly improve their relative abundance, but they were significantly increased in the AMF.BM treatment, indicating a good synergistic effect between AMF and BM. Based on the similarity clustering of community composition, all treatments were roughly classified into four categories, and the bacterial community evolution trend was not consistent among all treatments, indicating that the bacterial community evolution trend and structure composition were significantly changed after the application of bacterioides.

4.2. Relationship Between Soil Bacterial Community Structure and Soil Physicochemical Properties and Enzyme Activities

As the most abundant and widely distributed group of soil microorganisms [44], bacteria provide an important driving force and regulate the formation of soil organic matter, nutrient transformation and cycling, etc. Meanwhile, soil nutrients will also affect the number of soil bacteria and community composition, creating a mutual feeding effect [45]. The AMF, TH, and BM.TH treatments were positively correlated with SOM and mainly affected Acidobacteria, Verrucomicrobia, and Gemmatimonadetes, all of which play an important role in carbon metabolism [46]. Blastomonas was involved in carbon and nitrogen fixation [47]. Acidobacteria could degrade plant residue polymers and photosynthesis, especially in cold northern soils, where other cellulose-degrading bacteria found it difficult to survive, and the degradation function of Acidobacteria was particularly important [48]. The studies showed that carbon bacteria such as Chryseobacterium and Bacillus were significantly negatively correlated with SOM, which may be related to their strong organic matter decomposition ability. These bacteria provide an important driving force for soil organic matter metabolism and cycling [38]. AMF.BM had a significant effect on magnesium (Mg); Mg²⁺ in soil is the activator of many enzymes in microorganisms, promoting microbial DNA synthesis. Appropriate Mg levels had a positive effect in terms of the improvement of the soil microbial environment and soil fertility [49]. The study showed that Chryseobacterium, Unidentified Enterobacteriaceae, Bacillus, Pseudomonas, Acinetobacter, Enhydrobacter, and other bacteria were significantly positively correlated with Mg. Gaiella and Bradyrhizobium are important rhizosphere growth-promoting bacteria, which play a key role in balancing soil microecology and promoting healthy plant growth [50]. Correlation analysis showed a significant positive correlation with urease (SU), which was consistent with previous studies on the effects of actinomyces and bradyrhizobia on sucrase [51,52].

4.3. Effects of Different Bacterial Agents on Ecological Functions of Soil Bacteria

Changes in soil microbial community structure affect their ecological functions [53]. In this study, the cluster analysis results for soil bacterial microbial community and function prediction showed a high degree of consistency. Under different bacterial agent treatments, the ecological functional composition of soil bacteria tended to be similar, mainly including chemoheterotrophy, fermentation, aerobic–chemoheterotrophy, animal parasites or symbionts, mammal-gut bacteria, human-gut bacteria, human pathogens (all), and nitrification, which was consistent with previous studies on the function of soil bacteria in sandy areas [54]. However, soil microecology was very sensitive and responded promptly and sensitively to subtle changes in the soil environment. The application of bacterial agents had an impact on plant root exudes, soil nutrients, soil pH, and air permeability, and thus indirectly changed the community composition and ecological function of soil bacteria [55,56]. Therefore, although the functional composition of annotations was similar under different combinations of inoculants, the relative abundance of components was significantly different. Only two groups had a higher proportion of functional bacterial community abundance than the control group, among which the relative abundance of AMF.TH was the highest, followed by AMF.BM. However, there were more functional groups of AMF.BM annotations, and the overall uniformity was better than that of the former. The heat map showed that AMF.TH had a significant driving effect on only 6 functional groups, whereas AMF.BM had a significant driving effect on 13 of the 25 dominant functional groups. Previous studies have found that this treatment had significant effects on soil nutrient and enzyme activities, which the authors believed was closely related to the higher relative abundance and evenness of bacterial community functions driven by this treatment.

5. Conclusions

The experimental findings demonstrate distinct effects of microbial inoculants on soil bacterial communities and their functional roles in soil remediation. Key conclusions are summarized as follows:

The AMF.BM treatment significantly enhanced soil microbial biomass carbon (MBC) by 261.99% compared to the control (CK), while reducing metabolic entropy (qCO₂) by 68.05%. This treatment also improved soil enzyme activity (e.g., urease activity increased by 175%) and nutrient recovery (organic matter content rose by 33.72%), indicating its potential for restoring degraded soils.

TH treatment exhibited the highest Shannon diversity index (9.650 ± 0.06), which was significantly greater than that of CK (9.566 ± 0.11; p < 0.05), reflecting improved species evenness.

AMF.BM treatment showed the highest bacterial richness indices (Chao1: 3090.625 ± 9.72; ACE: 3075.324 ± 28.26), driven by increased dominance of Firmicutes (+201.92% vs. CK) and Proteobacteria (+31.56% vs. CK).

Redundancy analysis (RDA) identified organic matter, magnesium, urease, and catalase as key environmental drivers (p < 0.01), with microbial properties contributing more significantly than physicochemical factors (PERMANOVA, R² = 0.62).

AMF.TH and AMF.BM treatments enhanced the relative abundance of metabolic functions (e.g., nitrogen cycling and carbohydrate degradation) compared to CK. Specifically, AMF.BM significantly influenced 13 out of 25 dominant functional groups (FAPROTAX, p < 0.05). Predictive functional profiling (PICRUSt2) aligned closely with 16S rRNA sequencing results (Mantel test, r = 0.84, p = 0.001), validating the structural and functional coherence of the bacterial community.

The AMF.BM treatment demonstrated synergistic effects, significantly enhancing soil remediation through both physiological metrics (e.g., dehydrogenase activity increased by 49.95%) and molecular mechanisms (e.g., enrichment of Lactobacillus to 17.94% of total genera). These findings provide a theoretical foundation for optimizing microbial consortia in erosion-prone soils and advancing sustainable land reclamation strategies.