Impacts of Soil Compaction and Phosphorus Levels on the Dynamics of Phosphate-Solubilizing and Nitrogen-Fixing Bacteria in the Peanut Rhizosphere

Abstract

Soil properties, including soil compaction and the nutrient content, influence the composition and functions of rhizosphere microbial communities. There is limited information on how soil compaction and phosphorus application affect phosphate-solubilizing (PSB) and nitrogen-fixing bacteria (NFB). This study aimed to examine the responses of PSB and NFB in the rhizosphere of peanut (Arachis hypogaea L.) plants under varying soil compaction and phosphorus application levels. To address this, pot experiments were conducted to assess the composition and assembly processes of rhizosphere PSB and NFB in peanut cultivar Hua Yu 22 under two soil compaction levels (T1, 1.25 g/cm³ compaction, and T2, 1.00 g/cm³ compaction) and two phosphorus (P) levels (P0, no P applied, and P1, 1.2 mM P/kg soil applied). The results showed that PSB community shifts were closely correlated with the content of soil available phosphorus, soil acid phosphatase activity, soil nitrogenase activity, and soil compaction. Additionally, the content of soil available phosphorus and soil compaction were correlated with changes in operational taxonomic units of NFB. A network analysis revealed that the complexities of PSB were significantly higher than those of NFB. A stronger negative relationship was identified among NFB communities. The assembly of PSB communities was primarily driven by drift processes, whereas NFB communities were influenced by a combination of homogenizing selection and drift. Both PSB and NFB community compositions were significantly affected by phosphorus limitations and soil compaction. These findings enhance our understanding of the impacts of soil compaction and phosphorus application on PSB and NFB communities, with implications for optimizing peanut crop production. Our results will provide reference for crop cultivation in compacted and low-phosphorus soils. The important phosphate-solubilizing and nitrogen-fixing bacteria screened in the interaction network in this study will become candidate microbial agents for alleviating soil compaction and low phosphorus levels.

1. Introduction

Phosphorus (P) is crucial for peanut growth, impacting metabolism, including respiration, photosynthesis, and protein synthesis [1,2,3,4]. Phosphorus in soil exists as available inorganic P for plant uptake and less accessible insoluble forms fixed by metal cations [5]. The soil mineral and organic compositions significantly influence P formation in the soil. Despite high total P level, the amount of soluble P available for plant uptake is often limited [6,7]. Moreover, both peanut root growth and pod development require moderate soil mechanical stimulation, soil moisture content, aeration, and nutrient transport, all of which can be hindered by unsuitable soil compaction (either too loose or too compact). Unsuitable soil compaction would slow down root and pod growth, leading to reduced crop yield even when nutrients and water are adequate. Additionally, long-term changes in soil properties due to phosphate fertilizer application and alterations in the oxygen content caused by soil compaction can result in shifts in the microbial community composition and diversity [8].

The root system is the most predominant and sensitive organ for detecting the soil P content and responding to P deficiency through various strategies to increase the source of available P (AP). Plants respond to P deficiency by modifying the root architecture, adjusting P transport, and secreting acid phosphatase to increase available P [9]. Additionally, plants establish symbiotic relationships with rhizobia to mobilize unavailable soil P and capture sufficient nutrients around the rhizosphere [10]. Oxygen isotopes (δ¹⁸O-PO₄) are used to trace phosphorus uptake due to their stability and abundance [11,12,13].

The rhizosphere is rich in nutrients due to root–microbe interactions. Rhizosphere microorganisms play a key role in plant nutrient absorption by the intricately linking with soil pore connectivity. Additionally, changing the microbial community composition would bring large-scale influences on soil physicochemical properties [14]. Phosphate-solubilizing bacteria (PSB) convert unavailable P into forms usable by plants, enhancing growth and reducing the need for P fertilizers [15,16,17]. Given that peanut is a typical legume crop with nodulation, it is essential to investigate the changes in the rhizosphere microorganisms of peanut under low phosphorus stress and unsuitable soil compaction. While numerous reports have tightly connected the relationship between plant roots and rhizo-microorganisms, there remain many unknown aspects regarding the alterations of rhizo-microorganisms in peanut roots under low P stress and unsuitable soil compaction. Various studies mainly determined the beneficial influences of PSB on crop yield, but few of them focus on the alleviating function of PSB on soil–PSB–plant nutrient cycle and peanut cultivation under stress. Therefore, determining detailed alterations in the PSB community and facility to the soil–root system, as well as the relationships with soil physicochemical properties and plant performance is very crucial for optimizing the soil microbial species composition and simultaneously reducing P fertilizer application to promote plant development and production [17].

Additionally, it has been demonstrated that sufficient phosphorus can enhance the activity of nitrate reductase in roots, leading to increased nitrogen (N) absorption by plants. Under low P stress, not only were the leaves deficient in P but also the N contents decreased significantly. After the reasonable application of P fertilizer, N absorption increased significantly [18]. The bioavailable nitrogen form, including nitrate and ammonium, depends on various conversion reactions which are carried out by versatile nitrogen-fixing microorganisms, especially nitrogen-fixing bacteria (NFB) [19]. Phosphorus enhances nitrate reductase activity, boosting nitrogen absorption. Symbiotic NFB convert atmospheric N₂ into ammonia, a process influenced by phosphorus availability [18,20]. Conversely, a lack of phosphorus nutrition can significantly decrease the nitrogen-fixing capacity of nitrogen-fixing microorganisms. Additionally, phosphate-solubilizing bacteria (PSB), as plant-friendly beneficial rhizo-microorganisms, can enhance the available P content in soil and potentially promote biological nitrogen fixation. Although the effects of P addition on soil nitrogen cycling have been extensively studied, research on the effects of P addition on two functional soil microorganisms, PSB and NFB, is scarce. This study aims to investigate the influences of soil conditions on these two functional microbial communities (PSB and NFB), which are closely related to plant phosphorus and nitrogen uptake.

In addition, unsuitable soil compaction is considered one of the major obstacles to crop growth and yield, as stated by the Food and Agriculture Organization (FAO). Unsuitable soil compaction is a major obstacle to crop growth and yield, and is influenced by mechanized agriculture and soil characteristics [21]. It affects the soil structure, including porosity, bulk density, water infiltration, oxygen diffusion, and root penetration, which in turn impacts N and P uptake and microbial community structures [22,23].

Peanut (Arachis hypogaea L.) is a legume rich in protein and oil, and its growth is limited by insufficient soluble P in the soil. Adequate P availability leads to higher yields and better quality [24]. Peanut root and pod growth require proper soil compaction, moisture, aeration, and nutrient transport. Unsuitable compaction can hinder growth and reduce yield, even with adequate nutrients and water. Long-term changes in soil properties due to phosphate fertilization and compaction can shift the microbial community composition and diversity [8]. Given the importance of peanuts and their nodulation, investigating PSB and NFB in the rhizosphere under low P stress and unsuitable soil compaction is crucial. Despite research on plant–root–microbe relationships, details on rhizosphere PSB and NFB under these conditions remain unclear. This study used a pot experiment to examine the effects of low P and soil compaction on peanut phenotypes and rhizosphere properties. We assessed the PSB and NFB composition and abundance using high-throughput sequencing and molecular analysis, and analyzed the relationships between microbial communities and soil characteristics.

2. Materials and Methods

2.1. Plant Materials and Experimental Layout

The cultivated variety of Arachis hypogaea L., Hua Yu 22 (HY22), which is one of the dominant high-yield domestic landraces in Shandong Province, China, was used in this study. Plump and uniformly sized peanut seeds were selected, surface-sterilized with 100 mL of 70% ethanol (v/v) for 20 s, treated with 0.1% mercury for 2 min, rinsed three times with distilled water, and then randomly planted in different treatment pots in the greenhouse. One day before planting, the soil was pre-moistened with 900 mL of water per pot. The soil was sourced from the 0–20 cm soil layer at the Laixi experimental farm in Qingdao, China (36°87′ N, 120°05′ E). The soil is characterized as brown soil with an initial pH of 6.42, total available phosphorus (AP) content of 24.35 mg/kg, available nitrogen (AN) content of 81.45 mg/kg, and available potassium (AK) content of 30.54 mg/kg. The soil was sifted through a 7 mm sieve and filled into pots with a 30 cm diameter, 24 cm high and 9 kg soil volume.

The treatments included two P levels: (i) P0 with no P application and (ii) P1 with 174.42 mg of KH₂PO₄ (1.2 mM P) per kilogram of soil. The same amounts of nitrogen (N) were applied to both treatments using 195.65 mg of urea (3.2 mM N) per kilogram of soil. The potassium (K) level was adjusted in P1 by adding KCl to match the same amount in P0, as KH₂PO₄ introduced K into the P1 treatment. Both P1 and P0 treatments received 2.5 mM K (190.21 mg of KCl in P0 and 97.87 mg of KCl in P1) per kilogram of soil. Soil compaction treatments were performed under dry conditions with two levels: (1) T1 treatments were compacted at a soil bulk density of 1.25 g/cm³, and (2) T2 treatments were compacted at a soil bulk density of 1.00 g/cm³. For the 1.25 g/cm³ soil bulk density, we first weighed a certain amount of soil and then compacted it to the corresponding volume, which we marked on the pots. For the 1.00 g/cm³ soil bulk density, we first weighed a certain amount of soil, put the soil into a counting cup, then added a certain volume of perlite, which basically would not increase the mixed soil weight and only increase the volume; last, we mixed it well and placed it into a pot.

After P application and soil compaction stress treatments, the pots were divided into four groups (T1P0, T1P1, T2P0, and T2P1) randomly, with each group conducted in triplicate, and sown with HY22 peanut seeds. We watered 300 mL every other day for each pot. Simultaneously, a phosphate oxygen isotope (δ¹⁸O-PO₄) tracing experiment was conducted in small pots (with a 9 cm diameter and 1 kg soil volume) by adding ¹⁸O-labeled KH₂PO₄ (99 atom% ¹⁸O, Sigma-Aldrich, St. Louis, MO, USA) to trace P uptake in peanuts (130.815 mg of KH₂PO₄ and 43.605 mg of ¹⁸O-labeled KH₂PO₄ (KH₂PO₄/¹⁸O-labeled KH₂PO₄ = 3:1) per kilogram of soil). All treatment pots were randomly placed in the greenhouse to control for location differences, and the experiment lasted for three months. The conditions were controlled at 22 °C to 28 °C, with a 12 h sunlight and 12 h night cycle, a relative humidity of 50%, and 1300 lux. Initially, two seeds were sown in one pot to ensure at least one surviving plant, and only one seedling was retained in each pot for detection.

2.2. Phenotype Detection

The plant morphology of the four treatment groups (T1P0, T1P1, T2P0, and T2P1) was assessed after cultivation in the greenhouse for 90 days, including the following parameters: main stem height (cm), main stem diameter (mm), first lateral branch length (cm), branch number, number of leaves on the main stem, and number of nodules. Roots, stems, leaves, and pods were harvested, separated, steamed at 100 °C for 30 min, dried at 80 °C to a constant weight in the oven, and weighed for root dry matter (g), stem dry matter (g), leaf dry matter (g), and pod dry matter (g) after samples had cooled to room temperature. The growth parameters of roots were investigated using a root scan system and WinRHIZO root analysis software pro 2009c (Regent instruments Inc., Quebec, QC, Canada) under standardized conditions, including the total root length (cm), total root surface area (cm²), total root volume (cm³), and total root tip number. Photosynthesis index detection was conducted with the third youngest fully expanded leaf on each plant using the CIRAS-3 portable photosynthesis system (PP SYSTEMS, MA, USA) according to standard instructions. This included the net photosynthesis rate (A, μmol CO₂/m²/s), intercellular CO₂ concentration (Ci, μmol CO₂/mol), transpiration rate (E, mmol H₂O/m²/s), and stomatal conduction degree (Gs, mmol H₂O/m²/s) in the greenhouse at the time from 9:00 to 11:00 in the morning at 60 days after planting.

To determine P and N contents in the roots, stems, leaves, and pods, tissues of three samples were taken from each treatment group, baked at 100 °C for 30 min, and dried at 80 °C to a constant weight. The molybdenum blue colorimetric method was used to detect the concentration of P in different samples, and the wavelength for the absorbance measurement was 700 nm. The samples were digested with sulfuric acid and hydrogen peroxide to convert various forms of P into orthophosphate, which reacted with the molybdenum antimony anti-chromogenic agent to produce blue phosphomolybdenum. The absorption value of the solution was read using an ELISA (enzyme-linked immunosorbent assay) reader. The Kjeldahl method was used for N concentration testing in different tissues (K1100, HANON, Jinan, China). After grinding into a fine powder and passing through a 40-mesh sieve, a total of 0.3 g of the sample was completely digested with concentrated sulfuric acid and tested with a nitrogen analyzer.

To determine the characteristics of rhizosphere soil, samples were collected from the peanut root rhizosphere of each plant 90 days after planting. After sampling, soil from the same treatment group was thoroughly mixed and divided into two parts for subsequent PSB and NFB composition analyses and soil nutrient analyses, respectively. The contents of NH₄⁺ and NO₃⁻ in soil were extracted with 50 mL of a 2 mol/L KCl solution, shaken at 200 rpm for 40 min, and the content of soil ammonium nitrogen was determined by the indophenol blue colorimetry method at a 625 nm wavelength using UV spectrophotometry. Meanwhile, the content of soil nitrate nitrogen was determined at a 410 nm wavelength using UV spectrophotometry [25]. The pH value of the soil was detected using a pH meter with 10 g of soil and 25 mL of water (ratio of 1:2.5 (w/v)). The content of soil AP was determined by leaching with ammonium fluoride and a hydrochloric acid solution, and the Mo-Sb colorimetric method according to the instruction of Acid Soil Rapid Phosphorus Test Kit (MDBio, Taipei, China). To estimate the activities of NITS and ACP in both peanut roots and rhizosphere soil in different treatment groups, an enzyme solution was extracted from 0.5 g of fresh roots or rhizosphere soil ground by a pestle in a mortar with 6 mL of 0.2 M ice pre-cooled potassium phosphate (KH₂PO₄ and K₂HPO₄) buffer (pH = 7.8). After centrifuging at 4000 × g for 20 min, the supernatant was used to determine NITS and ACP enzyme activities using an ELISA kit (MDBio, Taipei, China) following the standard protocols for each enzyme. All procedures were carried out at 4 °C. The δ¹⁸O-PO₄ contents in peanut roots, stems, leaves, and soils were analyzed by a MAT253 isotope ratio mass spectrometer (Thermo Electron, Bremen, Germany) and an elemental analyzer (TC-EA, Thermo Finnigan, Bremen, Germany), with a precision limit of δ¹⁸O-PO₄ of 0.1%. All δ¹⁸O-PO₄ analyses were undertaken at The Institute of Subtropical Agriculture Chinese Academy of Sciences, China.

2.3. High-Throughput Sequencing of PSB and NFB

Rhizosphere soil from each treatment was collected and stored at −80 °C for the preparation of PSB and NFB sequencing. Total DNA was extracted from 0.5 g of rhizosphere soil using the TGuide S96 Magnetic Soil/Stool DNA Kit (Tiangen Biotech, Beijing, China) following the manufacturer’s instructions. The concentration and quality of DNA were assessed using the Qubit dsDNA HS Assay Kit and Qubit 4.0 Fluorometer (Invitrogen, OR, USA) with A260/A280 ratio of 1.83. DNA integrity was assessed via gel electrophoresis.

The phoD gene, encoding an alkaline phosphatase gene with the highest abundance in rhizosphere soil, served as a crucial functional marker for monitoring phosphate-solubilizing bacteria [26]. The phoD-F733 (5′-TGGGAYGATCAYGARGT-3′) and phoD-R1083 (5′-CTGSGCSAKSACRTTCCA-3′) universal primers were employed to amplify the phoD gene fragment, enabling the identification of PSB from the genomic DNA extracted from rhizosphere soil. The nifH gene, encoding one of the components of nitrogenase and commonly harbored by most NFB, was frequently used as a marker gene for detecting NFB [20]. A pair of nifH-F (5′-TGCGAYCCSAARGCBGACTC-3′) and nifH-R (5′-ATSGCCATCATYTCRCCGGA-3′) primers were used to amplify the nifH gene, which is extensively used for identifying NFB in rhizosphere soil.

The PCR reaction mixture (10 μL) consisted of 5 μL of KOD FX Neo Buffer (2×), 1 μL (50 ng) of DNA template, 0.3 μL of each primer (10 μM), 2 μL of the dNTP solution (2 mM each), 0.2 μL of KOD FX Neo, and 1.2 μL of PCR-grade water. The PCR program included an initial denaturation at 95 °C for 5 min, followed by 30 cycles of denaturation at 95 °C for 30 s, annealing at 55 °C for 30 s, extension at 72 °C for 40 s, and a final extension step at 72 °C for 10 min. The PCR products (approximately 300 bp to 500 bp) were verified on a 1.2% (w/v) agarose gel, stained with Gelred (US Everbright INC., CA, USA), and visualized under UV light. The PCR amplicons were purified with Agencourt AMPure XP Beads (Beckman Coulter, IN, USA), quantified using the Qubit dsDNA HS Assay Kit and Qubit 4.0 Fluorometer (Invitrogen), and pooled in equal molar amounts to construct the library for sequencing. High-throughput sequencing was performed using the Illumina NovaSeq 6000 platform (Illumina, Santiago, CA, USA).

2.4. Bioinformatics

The bioinformatics analysis for this study was conducted on the BMK Cloud platform (Biomarker Technologies, Beijing, China). Initially, raw data underwent primary filtering to eliminate adapter and primer sequences using Trimmomatic v.0.33 [27] and Cutadapt v.1.9.1 [28], with the filtering criteria set to sequences longer than 140 bp and a quality score higher than 20. Clean data and paired-end reads obtained from these steps were assembled using USEARCH v.10 [29] and subjected to chimera removal by UCHIME v.4.2 [30]. High-quality sequences were aligned, and clusters were generated into operational taxonomic units (OTUs) with a similarity threshold of more than 97%. The taxonomic classification of OTUs was performed using the Naive Bayes Classifier in QIIME2 [31] based on the SILVA database with a confidence threshold of 70% [30]. The raw data for PSB and NFB have been deposited in the NCBI Sequence Read Archive (SRA) under the accession numbers PRJNA953208 and PRJNA954120, respectively.

2.5. Statistical Analysis

Two-way ANOVA was conducted to analyze significant differences in peanut phenotypic characteristics, soil properties, and relative microbial abundance among treatment groups. A completely randomized design (CRD) was employed for each test with at least three replicates per treatment. Statistical analyses of the data were performed using R project v.4.1.1 at a significance level of p ≤ 0.05 and an extremely significant level of p ≤ 0.01 with Bonferroni’s correction for multiple comparisons. The phylogenetic tree of all detected OTUs was generated on the BMK Cloud platform (Biomarker Technologies, Beijing, China). All detected OTUs were assigned at both the genus and phylum levels, and OTUs with significantly different relative abundances in soil compaction and low P treatment groups were marked using iTOL v6 [32].

To assess the sequencing depth, alpha diversity, including the rarefaction curve, Shannon curve, rank abundance curve, and species accumulation curve, was evaluated separately for T1P0, T1P1, T2P0, and T2P1 samples using QIIME2 and displayed using the R project. Analysis of variance (ANOVA) was employed to test the significance of changes between treatments. The beta diversity of treatments was calculated to determine the diversity of microbial communities between samples using QIIME2. Principal Coordinate Analysis (PCoA) and a heatmap based on Bray–Curtis distances were used to analyze beta diversity. The effects of soil compaction and P application on the community composition were detected by PERMANOVA/ANOSIM (permutational multivariate analysis of variance/analysis of similarities) using the “vegan” package. Differences in the relative abundance of PSB and NFB between groups were assessed using the Wilcoxon rank-sum test. DEseq2 in R project was used to determine the significant differences (enriched or depleted) in OTU relative abundance between treatments according to the criteria log₂fold change > 1.3 (p < 0.05), with a mean relative abundance greater than 0.0001%.

Detrended correspondence analysis (DCA), redundancy analysis (RDA), and the Mantel test were used to investigate the relationship between significantly different bacterial community compositions and soil environmental factors in the R project. Pearson’s correlation analysis was employed to detect the relationships between soil environmental factors using the “corrplot” package (4.1.3), and the results were visualized using the “ggcorrplot” package (4.1.3) in the R project. An envift test was conducted to identify the significant operational parameters with 1000 permutations using the “vegan” package (4.1.3).

To evaluate the community assembly processes of PSB and NFB under the treatments, a null model framework method was employed to shuffle species and abundances across phylogeny tips. The taxonomic and phylogenetic β-nearest taxon index (βNTI) and Bray–Curtis-based Raup–Crick (RCbray) were calculated with 1000 randomization repeats to quantify the null model expectation [33,34] on the BMK Cloud platform (Biomarker, Beijing, China). To detect the significant phylogenetic signal, which was a prerequisite for the βNTI analysis, the relationship between environmental differences and the phylogenetic distance of pairwise OTUs was tested by the Mantel test on the BMK Cloud platform (Biomarker, China). Euclidean distance was used to calculate the differences in the environment between pairwise OTUs. The phylogenetic distance of pairwise OTUs was obtained using “picante” in the R package (4.1.3). Positive and negative values of βNTI, which are the observed β-mean nearest taxon distance (βMNTDobs) more and less than the mean null distribution of phylogenetic turnover (βMNTDnull), respectively, characterized the deviation magnitude between βMNTDobs and βMNTDnull. The pairwise comparisons of βNTI > +2 and βNTI < −2 were used to evaluate the influence of shaping the community composition by variable selection and homogenizing selection (deterministic processes), respectively. Meanwhile, the pairwise comparisons of |βNTI| < 2 meant that stochastic processes governed the community turnover more, including drift acting alone, dispersal limitation acting alongside drift, or homogenizing dispersal, divided by the criterion of RC_bray > +0.95, RC_bray < −0.95, and |RC_bray| < 0.95, which were signified as dispersal limitation, homogenizing dispersal, and ecological drift, respectively.

A correlation network analysis was conducted to explore the connections and the coexistence relationships within and between PSB and NFB using the Spearman correlation algorithm. The filter rule was absolute correlation coefficients |R| greater than 0.85, with a significance of p < 0.05 for each network, and was visualized using Cytoscape 3.9.1 [35]. Both within-module connectivity (Zi) and among-module connectivity (Pi) were calculated according to the RMT-based molecular ecological network analysis method [36]. Nodes were divided into four categories, including peripheral nodes (zi  ≤  2.5, Pi  ≤  0.62), connectors (zi  ≤  2.5, Pi  >  0.62), module hubs (zi  >  2.5, Pi  ≤  0.62), and network hubs (zi  >  2.5, Pi  >  0.62) [37].

3. Results

3.1. Phenotypic Characteristics of Peanuts in Each Treatment Group

The phenotype of peanut plants in each treatment group was investigated to test the joint effects of soil compaction and P application on the growth and development of peanuts. In the –P treatment, obvious growth inhibition was identified both in morphology and physiology compared with the +P treatment in both T1 and T2 compaction levels. The main stem height and the first lateral branch length in T1P1 were significantly higher than in other groups (p < 0.01). The main stem diameter, branch number, and the number of leaves on the main stem in T1P1 were also higher between groups, although with no significance (Figure 1A). The dry weight (DW) of stems, leaves, and pods under T1P1 treatment were all heavier than that of other treatments (Figure 1B). To detect the P and N contents in different tissues, three samples were randomly taken from each treatment group. The results showed that the P content in roots, stems, and leaves under the +P treatment were significantly increased compared with that under the −P treatment (Figure 1C). In the T1P1 group, the N contents of the roots and stems/leaves, and the P content of the roots were the highest at 2.21, 2.89, and 0.89 g kg⁻¹ DW, respectively. The content of soil nitrate N in T1 was obviously lower than that in T2, while the content of soil ammonium N had a few differences that were not statistically significant between the four treatment groups. Both the content of soil available P and the number of nodules were higher in the +P treatment group than in the −P treatment group (Figure 1D). The values of total root length, total root surface area, total root volume, and total root tip number were increased in the P1 group than in the P0 group (Supplementary Figure S1). This result indicated that the joint effects of soil compaction and P application had a great impact on the whole plant and especially on the roots.

The net photosynthetic rate, A (Figure 1E); the intercellular carbon dioxide concentration, Ci (Figure 1F); the transpiration rate, E (Figure 1G); and the stomatal conductance, Gs (Figure 1H) under the +P treatment were highly significantly different from that under the –P treatment (p < 0.01). P application significantly increased the A value by 84.4% and 28.6%, the Ci value by 9.5% and 26.8%, the E value by 61.1% and 49.5%, and the Gs value by 138.3% and 12.2% in the plants with the T1 and T2 soil compaction levels, respectively. The activities of ACP and NITS in soil and roots were detected using ELISA methods (MDBio). P addition enormously decreased the activities of ACP and increased the activities of NITS in roots. However, in the soil, the trends for the two enzyme activities were similar (Figure 1I–L). Meanwhile, the δ¹⁸O-PO₄ values were analyzed in peanut roots (Figure 2A), stems and leaves (Figure 2B), and soils (Figure 2C) from all treatment groups. It was indicated that δ¹⁸O-PO₄ values were significantly increased in the +P treatment group than in the −P treatment group. Between different soil compaction treatments, δ¹⁸O-PO₄ values of peanut roots and soils in the T1P1 group were significantly higher than that in T2P1 group, while δ¹⁸O-PO₄ values of stems and leaves were lower in the T1P1 group (Figure 2). This result was consistent with the P contents of the root and stems/leaves detected in the above experiment.

3.2. Detection of Rhizosphere PSB in Each Treatment Group

The average length was 350 bp, and the average GC% was 67.29%. It is suggested that the sequences were highly accurate and reliable, with a mean value of 98.74% at the Q30 level (a 0.1% probability of an error) (Supplementary Table S1). In total, 1248 operational taxonomic units (OTUs) were identified from soil samples from the four treatment groups at the species level. Specifically, 171 and 147 OTUs were found between the T1 and T2 treatment groups, respectively, while 171 and 153 characteristic OTUs were found between the P0 and P1 treatment groups, respectively (Supplementary Figure S2A,B). OTUs were classified into 9 main genera, including Alphaproteobacteria, Sphingopyxis, Bradyrhizobium, Pseudomonas, Gammaproteobacteria, Gemmatirose, Actinobacteria, Streptomyces, and Proteobacteria, which were affiliated with four main phyla, including Actinobacteria, Proteobacteria, Sphingopyxis, and Gemmatimonadetes, according to the Naive Bayes Classifier (Supplementary Figure S2C, Supplementary Table S2). The major PSB derived from all treatments consisted of the phyla Proteobacteria, Sphingopyxis, Actinobacteria, and Gemmatimonadetes, and genera Pseudomonas, Streptomyces, and Gemmatirosa. The relative abundance heatmap of each PSB community revealed that the trends for changes in genera abundance of Pseudomonas and Streptomyces increased and clustered into one clade when the soil compaction decreased (Supplementary Figure S2D). This suggested a certain correlation between the abundance of Pseudomonas and Streptomyces and the compactness of the soil.

For the analysis of alpha diversity, the values of the ACE, Chao1, Shannon, and Simpson indexes were used to compare the diversity of the PSB community between different treatment groups (Supplementary Figure S3A). A total of five dominant genera, including Bradyrhizobium, Gemmatirosa, Pseudomonas, Sphingopyxis, and Streptomyces, comprised the major proportion of PSB communities with similar abundance trends across the soil samples. The genus accumulation curve, conducted according to the OTU numbers, had gained a plateau with stable species numbers (Supplementary Figure S3B).

Principal Coordinate Analysis (PCoA) of the PSB communities showed that PC1 (horizontal axis) and PC2 (vertical axis) accounted for 16.37% and 12.95% of the variation, respectively (Supplementary Figure S4A,B). Based on the similarities of the species composition at the genus level, the Bray–Curtis similarity distance between samples was gained to make a heat map, indicating that the difference in soil compaction has a great influence on the PSB composition (Supplementary Figure S4C). The PERMANOVA according to the Bray–Curtis distances showed a significant difference between treatment groups (permutational ANOVA, R = 0.439, p = 0.005) (Supplementary Figure S4D). To detect the influences of different soil compactions and P applications, the comparison of the relative abundance of PSB between treatments was conducted by DEseq2 according to the criteria log₂fold change > 1.3 (p < 0.05). The relative abundances of 58 up-regulated OTUs and 80 down-regulated OTUs were identified as significantly different in the T1 vs. T2 comparison (Figure 3), while 68 up-regulated OTUs and 55 down-regulated OTUs were obviously changed compared in the P0 vs. P1 comparison, all belonging to the most dominant 4 genera (Streptomyces, Pseudomonas, Gemmatirosa, and Bradyrhizobium) (Figure 3).

3.3. Diversity and Composition of the NFB Community in Each Treatment Group

The average length was 320 bp, and the average GC% was 65.22%. It is suggested that the sequences were highly accurate and reliable, with a mean value of 98.47% at the Q30 level (Supplementary Table S3). Accurate and reliable data are an important foundation for a subsequent bioinformatics analysis. In total, we obtained 957,866 high-quality clean reads of bacterial nifH gene sequences, which were well clustered into a total of 593 OTUs from rhizosphere soil samples at two compaction levels and two P application treatments. The NFB community composition at each taxonomic level is shown in Table S4. Specifically, 137 and 129 OTUs were found in the T1 vs. T2 comparison, respectively, while 122 and 141 characteristic OTUs were found in the P0 vs. P1 comparison, respectively (Supplementary Figure S5A,B). OTUs were classified into five phyla, including Cyanobacteria, Actinobacteria, Verrucomicrobia, Firmicutes, and Proteobacteria, and ten main genera, including Zoogloea, Pelomonas, Azoarcus, Azotobacter, Methylocystis, Azohydromonas, Geobacter, Azospirillum, Anaeromyxobacter, and Bradyrhizobium, by the Naive Bayes Classifier (Supplementary Figure S5C, Supplementary Table S4). The major NFB derived from all treatments consisted of the phylum Proteobacteria (97.76%) dominated by the genus Bradyrhizobium (17.29%), followed by Anaeromyxobacter (13.54%), Azospirillum (4.25%), Geobacter (4.09%), and Azohydromonas (3.95%). The relative abundance heatmap was also displayed at the genus level (Supplementary Figure S5D). These genera are important components of soil ecosystems and play a crucial role in soil nitrogen cycling.

The values of the ACE, Chao1, Shannon, and Simpson indexes were analyzed for alpha diversity (α diversity) to detect the NFB community diversity between treatment groups. The Simpson and Shannon values of the T1P1 treatment group were significantly higher than those of the T2P0 and T2P1 treatment groups (t-test, p < 0.05), meaning that the NFB diversity between T1 and T2 treatment groups was different (Supplementary Figure S6A). The Shannon index curve, rarefaction curve, and genus accumulation curve reaching saturation all showed that the obtained sequences were sufficient for OTU assembly with the current sequencing depth (Supplementary Figure S6B). Furthermore, the correlation analysis indicated that the features (OTU number) and Shannon, ACE, Chao1, and Simpson indexes were significantly positively correlated with the soil compaction level, while significantly negatively correlated with the content of nitrate N (NO₃⁻-N) (p < 0.05) by Spearman’s correlation test (

Table 1. Soil and root properties across soil compaction treatments and P application levels.

	Soil pH	Soil Nitrate N	Soil Ammonium N	Soil Available P	Soil ACP Activity	Soil NITS Activity	Root ACP Activity	Root NITS Activity	Root N Content	Root P Content	Nodule Number
T1P0	6.49 ± 0.04 a	38.14 ± 0.97 b	15.44 ± 0.82 a	43.50 ± 1.96 c	53.32 ± 1.37 a	200.42 ± 5.76 a	47.16 ± 1.02 ab	175.27 ± 1.66 c	1.84 ± 0.07 b	0.61 ± 0.02 c	63.67 ± 6.66 b
T1P1	6.16 ± 0.13 c	37.07 ± 2.37 b	15.10 ± 0.52 a	105.99 ± 2.49 a	50.18 ± 0.24 b	186.46 ± 3.03 b	44.64 ± 1.21 c	191.29 ± 1.19 b	2.21 ± 0.05 a	0.90 ± 0.03 a	113.33 ± 13.05 a
T2P0	6.43 ± 0.05 ab	48.53 ± 2.81 a	14.43 ± 0.70 a	53.27 ± 0.80 b	53.02 ± 1.10 a	175.96 ± 3.23 c	47.74 ± 1.40 a	177.51 ± 2.37 c	1.84 ± 0.06 b	0.81 ± 0.03 b	61 ± 9.17 b
T2P1	6.29 ± 0.05 bc	52.74 ± 2.41 a	15.17 ± 0.81 a	104.66 ± 2.51 a	53.55 ± 0.96 a	188.70 ± 2.55 b	45.48 ± 0.75 bc	197.14 ± 4.04 a	1.93 ± 0.09 b	0.88 ± 0.04 a	95 ± 11.53 a

Note: the data in the table represent the means ± standard errors. Samples with different lowercase letters were significantly different from each other (p < 0.05).

and

Table 2. Spearman’s correlation test for soil and root properties with PSB and NFB α diversity indexes.

	Soil pH	Soil Nitrate N	Soil Ammonium N	Soil Available P	Soil ACP Activity	Soil NITS Activity	Root ACP Activity	Root NITS Activity	Root N Content	Root P Content	Nodule Number	Soil Compaction	Applied P	Feature	ACE	Chao1	Simpson	Shannon
Soil pH	1
Soil nitrate N	0.126	1
Soil ammonium N	−0.0823	−0.28	1
Soil available P	−0.7928 **	0.2204	0.0127	1
Soil ACP activity	0.5685	0.4721	0.1136	−0.4439	1
Soil NITS activity	0.1451	−0.4054	0.478	−0.2836	0.1331	1
Root ACP activity	0.4848	0.0736	−0.2212	−0.7619 **	0.5428	−0.0536	1
Root NITS activity	−0.7138 **	0.2849	0.1096	0.9236 **	−0.2284	−0.1199	−0.6966 *	1
Root N content	−0.796 **	−0.3995	0.0444	0.6985 *	−0.7726 **	−0.0866	−0.6763 *	0.4911	1
Root P content	−0.6651 *	0.3543	−0.1223	0.8963 **	−0.4015	−0.5838 *	−0.5849 *	0.7549 **	0.5727	1
Nodule number	−0.7423 **	−0.1363	0.0375	0.8506 **	−0.609 *	−0.0127	−0.7953 **	0.7794 **	0.7758 **	0.6901 *	1
Soil compaction	−0.1292	−0.9387 **	0.3384	−0.1506	−0.4815	0.6012 *	−0.2286	−0.2149	0.4425	−0.3842	0.2234	1
Applied P	−0.8219 **	0.1128	0.1465	0.9611 **	−0.4093	−0.0326	−0.7723 **	0.9464 **	0.699 *	0.7608 **	0.89 **	0	1
PSB_Feature	0.2444	0.0353	0.2296	−0.3701	0.1653	0.2119	0.3736	−0.3592	−0.483	−0.3379	−0.2848	0.0201	−0.3221	1
PSB_ACE	0.2099	−0.4437	0.3042	−0.5163	−0.1727	0.5063	0.2632	−0.5312	−0.2096	−0.6015 *	−0.243	0.5295	−0.3957	0.7837 **	1
PSB_Chao1	0.4865	−0.2355	0.324	−0.6435 *	0.167	0.4723	0.3249	−0.5696	−0.4908	−0.7257 **	−0.396	0.331	−0.526	0.7866 **	0.8887 **	1
PSB_Simpson	0.2811	0.3854	−0.1938	−0.2011	0.5641	−0.0549	0.4008	−0.0071	−0.4572	−0.0163	−0.2708	−0.5257	−0.2114	−0.2892	−0.4868	−0.3547	1
PSB_Shannon	0.15	0.2537	0.0417	−0.151	0.4459	0.0738	0.4484	0.0234	−0.4305	−0.0022	−0.0987	−0.3727	−0.1067	0.1424	−0.1146	−0.0634	0.7961 **	1
NFB_Feature	−0.3227	−0.8951 **	0.3575	0.0146	−0.5339	0.4998	−0.2586	−0.0845	0.5794 *	−0.2144	0.3894	0.9509 **	0.1623	1
NFB_ACE	−0.0096	−0.698 *	0.6357 *	−0.2656	−0.0115	0.5536	−0.0438	−0.2438	0.1423	−0.4124	0.0921	0.7222 **	−0.0974	0.7586 **	1
NFB_Chao1	−0.118	−0.8237 **	0.4261	−0.3101	−0.1463	0.4501	0.122	−0.336	0.2318	−0.4514	0.0473	0.7947 **	−0.1581	0.8461 **	0.9235 **	1
NFB_Simpson	−0.1886	−0.6634 *	0.4099	−0.0479	−0.3655	0.3359	−0.1896	−0.2601	0.5554	−0.1583	0.1838	0.7466 **	0.028	0.8204 **	0.7008 *	0.7345 **	1
NFB_Shannon	−0.3877	−0.6966 *	0.3932	0.1206	−0.4799	0.3001	−0.2581	−0.0838	0.682 *	−0.0362	0.3451	0.7695 **	0.207	0.8837 **	0.6643 *	0.7414 **	0.9664 **	1

*, p < 0.05. **, p < 0.01.

).

PCoA of the NFB community indicated that PC1 and PC2 accounted for 20.43% and 11.82% of the variation, respectively, and indicated significant segregation between treatment groups (PERMANOVA, R = 0.774, p = 0.006, df = 1) (Supplementary Figure S7A,B). The result showed that soil compaction has a significant impact on NFB. Likewise, based on the Bray–Curtis similarity distances, the sample heat map showed that T1 and T2 groups were well-clustered according to the soil compaction level, indicating that the soil compaction treatment has an obvious influence on the shifts in the NFB community composition (Supplementary Figure S7C,D). The relative abundances of 41 OTUs, all belonging to the most dominant 9 genera, were significantly different in the T1 vs. T2 comparison. The relative abundances of 35 OTUs affiliated with 8 genera were obviously changed in the P0 vs. P1 comparison (Figure 4).

3.4. Effects of Soil Environmental and Root Nutritional Condition Variables on PSB and NFB Communities

To investigate the relationship between soil environmental traits and PSB/NFB, both a detrended correspondence analysis (DCA) and redundancy analysis (RDA) were performed across all treatment samples. The results showed that the first horizontal axis explained 21.95% of constrained variations in the PSB community, with the longitudinal axis explaining 16.08%. Meanwhile, the horizontal axis explained 19.99% and the longitudinal axis explained 17.92% of the variations in the NFB community (Figure 5A,B). An RDA associated with an envfit test was performed to test the significant effects of soil properties on the PSB and NFB community compositions using the Bray–Curtis beta diversity matrix of PSB and NFB at the genus level. The first two dimensions of RDA represent 38.03% of the variation, with the key soil factors initially linked to the PSB composition including the soil pH value (envfit analysis, R² = 0.54, p < 0.05) and the content of nitrate N (envfit analysis, R² = 0.56, p < 0.05). For the NFB community, the total of two dimensions of the RDA represents 37.91% of the variation, with the significant soil properties including the content of nitrate N (envfit analysis, R² = 0.65, p < 0.01) and the soil ACP activity (envfit analysis, R² = 0.57, p < 0.01). The results indicate that P and N nutrients play key roles in the PSB and NFB community compositions.

To detect the linkage between environmental factors and significantly changed (enriched or depleted) OTUs in the two treatment groups, a Mantel test was conducted. The results for PSB showed that significantly up-regulated OTUs were positively correlated with the soil nitrate N content, soil/root NITS activity, and soil/root ACP activity, while significantly down-regulated OTUs were positively correlated with the soil pH, soil AP content, soil ACP activity, soil NITS activity, and root P/N contents under soil compaction treatment. Between the two P application levels, up-regulated OTUs were positively correlated with the soil pH, soil AP content, soil/root ACP activity, root NITS activity, and root N contents, while significantly down-regulated OTUs were positively correlated with the soil AP content, soil ACP activity, soil NITS activity, and root P content (Figure 5C, Supplementary Table S5). For NFB, the significantly up-regulated OTUs were positively correlated with the soil AP content and root NITS activity, while significantly down-regulated OTUs were positively correlated with the soil pH, soil nitrate N content, soil AP content, soil ACP activity, and root N/P contents under soil compaction treatment. Between the two P application levels, the significantly up-regulated OTUs were positively correlated with the soil pH, soil nitrate N content, soil ACP activity, and soil N/P contents, while the significantly down-regulated OTUs were positively correlated with the nodule number (Figure 5D, Supplementary Table S6). Altogether, three environmental variables, the soil nitrate N content, soil ACP activity, and soil NITS activity, have effects on both up-regulated OTUs and down-regulated OTUs of PSB simultaneously under different soil compaction levels. The soil AP content, soil ACP activity, and root NITS activity affect the significantly changed OTUs of PSB concurrently under two P levels. Furthermore, the results indicate that soil ACP activity has an important effect on the significantly changed OTUs of PSB under both treatments. For NFB, the results showed that the soil AP content played an important role in both the increased and decreased OTUs at the two soil compaction levels. This indicated that the AP content has a significant impact on the NFB community.

The relationships between soil and root properties and PSB/NFB beta diversity were identified using the Mantel test. The results showed that soil ACP activity (R = 0.36, p < 0.01), the root N content (R = 0.50, p < 0.01), and the root P content (R = 0.50, p < 0.01) were significantly positively correlated with PSB beta diversity under T2 soil compaction, while no soil/root property was strongly correlated under T1 soil compaction. For NFB, the root N content (R = 0.55, p < 0.01), root P content (R = 0.53, p < 0.01), soil AP content (R = 0.22, p < 0.01), and soil ACP activity (R = 0.38, p < 0.01) were obviously correlated with NFB beta diversity under T2 soil compaction, while root NITS activity (R = 0.20, p < 0.05) and the soil nitrate N content (R = 0.22, p < 0.05) were correlated with that under T1 soil compaction (Supplementary Figure S8). The above results demonstrate the soil physicochemical properties that cause significant changes in the composition of PSB and NFB communities, providing a reference for subsequent soil ecological improvement.

The changes in soil and root properties were observed between different soil compaction and P application treatments. The soil compaction level significantly negatively affected the soil nitrate N content (R = −0.94, p < 0.01) and positively affected soil NITS activity (R = 0.60, p < 0.05), while the amount of phosphate fertilizer added was negatively correlated with the soil pH (R = −0.82, p < 0.01) and root ACP activity (R = −0.77, p < 0.01), and positively correlated with the soil AP content (R = 0.96, p < 0.01), root NITS activity (R = 0.95, p < 0.01), root N content (R = 0.69, p < 0.05), root P content (R = 0.76, p < 0.01), and nodule number (R = 0.89, p < 0.01). Soil pH was negatively correlated with the soil AP content (R = −0.79, p < 0.01), root NITS activity (R = −0.71, p < 0.01), root N content (R = −0.79, p < 0.01), root P content (R = −0.66, p < 0.05), and nodule number (R = −0.74, p < 0.01). The increased soil AP content could positively affect the root NITS activity (R = 0.92, p < 0.01), root N content (R = 0.69, p < 0.05), root P content (R = 0.86, p < 0.01), and nodule number (R = 0.85, p < 0.01), accompanied by decreased root ACP activity (R = −0.76, p < 0.01). Soil ACP activity was negatively correlated with the root N content (R = −0.77, p < 0.01) and nodule number (R = −0.61, p < 0.05), whereas soil NITS activity was negatively correlated with the root P content (R = −0.58, p < 0.05). Root ACP activity was significantly negatively correlated with the nodule number (R = −0.79, p < 0.01), while root NITS activity was strongly positively correlated with the nodule number (R = 0.77, p < 0.01). Meanwhile, root ACP activity was significantly negatively correlated with root NITS activity (R = −0.69, p < 0.01) (Figure 5B).

3.5. Effects of Soil Compaction and P Application on the Assembly Processes of PSB and NFB Communities

To identify the importance of assembly processes, βNTI values combined with RC_bray values were used for analysis. The majority of βNTI values of PSB were between −2 and +2, suggesting that PSB community assembly was dominated by stochastic processes (Figure 6A), whereas some βNTI values of NFB were between −4 and −2, indicating that NFB community assembly was greatly affected by homogenizing selection (deterministic processes) and stochastic processes (Figure 6B). Additionally, we quantified the relative importance (percentage) of five assembly processes for phylogenetic turnover in two treatment groups through a combined analysis with RC_bray values. The results showed that the PSB community was primarily governed by stochastic processes (drift processes from 84.8% to 96.9%) (Figure 6C), whereas the NFB community was dominated by both deterministic processes and stochastic processes in the two treatment groups (Figure 6D). To identify the law of process changes across treatment groups, the relative importance for each treatment group was fitted to quadratic models. The importance of drift processes in the PSB community was significantly fitted to quadratic models (R² = 0.90), suggesting that the decreased importance of drift processes with decreased soil compaction (Figure 6E). The importance of drift processes and homogenizing selection in the NFB community was fitted to quadratic models (R² = 0.57), indicating the increased importance of drift processes accompanied by the contrary trends of homogenizing selection importance and increased soil compaction on the NFB community (Figure 6F). To detect the environmental factors that affected community assembly processes, a Mantel test was conducted to identify the relationships between soil/root properties and βNTI (Supplementary Figure S9). In the T1P1 group, root NITS activity (R = 0.38, p < 0.01) mainly positively influenced the PSB assembly processes. For NFB, soil ACP activity (R = 0.59, p < 0.01) and the root N content (R = 0.51, p < 0.01) were significantly positively correlated with βNTI values, and soil pH (R = 0.41, p < 0.05) was positively correlated with that in the T2P0 group.

3.6. Effects of Soil Compaction and P Application on Networks of PSB and NFB Communities

Three networks were built and analyzed for PSB, NFB, and PSB-NFB, respectively, based on the significantly changed (enriched or depleted) OTUs of PSB and NFB in two treatment groups with the relative abundance. The correlation between the OTU relative abundances was analyzed by the Spearman method. The PSB network consisted of 483 edges and 172 nodes with an average degree of 5.949 (Figure 7A), while the NFB network included 35 edges and 42 nodes with an average degree of 2.0 (Figure 7B). It indicated that the PSB network was more complex and connected than that of the NFB network. The PSB network included 172 OTUs containing Proteobacteria (58 OTUs, 33.72%), Gemmatimonadetes (6 OTUs, 3.49%), and Actinobacteria (44 OTUs, 25.58%); meanwhile, the NFB network including 42 OTUs was only enriched with Proteobacteria (41 OTUs, 97.62%). Additionally, we analyzed the PSB-NFB networks, which consisted of 277 edges and 176 nodes with an average degree of 3.148 (Figure 7C). The rates of negatively correlated links within the NFB network (22.86%) were significantly higher than that within the PSB network (6.00%) and the interaction network between PSBs and NFBs (9.39%), suggesting competitive relationships between these OTUs. The ratio of down-regulated NFB OTUs between the groups without and with P application under T1 and T2 soil compaction, respectively, were 55.0% and 66.7%, while the ratio was 37.5% and 52.5% for PSB OTUs. Ratios under T2 soil compaction for both NFBs and PSBs were much higher compared to that under T1 soil compaction.

The OTUs within module degrees, within module connectivity, among module connectivity, and modularity were analyzed to define the importance of OTUs positioned within and among modules, providing a reference for the function of OTUs (genus). According to the topological roles of OTUs in the network, four PSB OTUs (OTU214 belonging to Bradyrhizobium, and OTU150, OTU87, and OTU52 being unclassified) (Figure 8A) and two NFB OTUs (OTU14 belonging to Azohydromonas and OTU3 being unclassified) were identified as connectors in two networks, respectively (Figure 8B). In the PSB-NFB interaction network, fifteen (including NFBOTU3 unclassified, NFBOTU40 belonging to Pelomonas, NFBOTU599 belonging to Azotobacter, PSBOTU215, PSBOTU125, PSBOTU254, and PSBOTU674 belonging to Streptomyces, PSBOTU491 and PSBOTU117 belonging to Gemmatirosa, PSBOTU340, PSBOTU627, PSBOTU777, PSBOTU84, PSBOTU496, and PSBOTU613 being unclassified) and seven OTUs (NFBOTU108 belonging to Azohydromonas, NFBOTU47, NFBOTU70, NFBOTU87 belonging to Anaeromyxobacter, NFBOTU12, NFBOTU317, and NFBOTU68 being unclassified) were considered to be connectors and module hubs, respectively. Additionally, no network hub was identified from the three networks (Figure 8C, Supplementary Table S7).

4. Discussion

Soil compaction caused an increase in the soil bulk density and changes in soil microorganism compositions, and thus influenced agriculture production. The rhizosphere microorganism compositions and their interactions with each other drive important roles in agricultural systems, including soil fertility, plant productivity, and stress tolerance [38]. Exogenous application of phosphorus fertilizer can induce agricultural system alterations to the rhizosphere microorganism compositions, interactions, soil structure and fertility, and root systems, and relieve the impact of stress on crop growth [14,39]. Meanwhile, PSB could enhance soil phosphorus cycling under ecological stresses [8,40]. In this study, we detected the community composition under different soil compaction and P application levels. The results indicated that soil compaction significantly altered the richness of the NFB community between different soil compaction levels by affecting soil conditions, especially influencing the shift in NFB community composition in T1P1 with the highest Simpson and Shannon values among treatments (Supplementary Figure S6). However, soil compaction had no significant effects on PSB community richness between treatment groups in this study (Supplementary Figure S3). Previous studies have reported that soil compaction has significant relationships with soil porosity, ventilation, and moisture [41,42]. Many studies have demonstrated that the rates of biological N₂ fixation (BNF) could be influenced by changes in the air content, especially oxygen pressure, which could negatively affect the efficiency of nitrogenase, the N₂ content, and soil moisture caused by different soil compaction levels and P nutrient contents, which are crucial for NFB growth and nitrogen fixation activity in leguminous crops [43,44].

Generally, species diversity is an important and useful positive indicator of community stability, and higher diversity would have stronger adaptability, which may be an essential determinant for ecosystem maintenance, balance, and recovery under environmental pressure [45]. The diversity of NFB was significantly higher in the T1 treatment group than that in the T2 treatment group (p < 0.01) (Supplementary Figure S6). The diversity of PSB was much higher than that of NFB in these studies. Meanwhile, the effects of two treatments on PSB were not significant (Supplementary Figure S3), indicating that the function of PSB in adopting various strategies to solubilize insoluble P into a soluble form was limited in impact by soil compaction and P fertilizer application.

The PSB obtained in the current study generally belonged to three phyla, including the two dominant phyla Actinobacteria and Proteobacteria, and another phylum, Gemmatimonadetes. Over half of OTUs were members affiliated with Proteobacteria, while interestingly, only one species belonged to the phylum Gemmatimonadetes, which has not been listed in the recently summarized 17 main genera of phosphate-solubilizing bacteria associated with alkaline phosphomonoesterase (phoD) gene (Supplementary Figure S2) [46,47]. The total relative abundances of PSB phylum Actinobacteria decreased significantly under both T1 soil compaction and P0 treatment, while that of phyla Gemmatimonadetes and Proteobacteria increased significantly under T1 soil compaction. Many studies have reported that PSB belonging to Actinobacteria phylum present evident positive effects on plant growth properties under different P application levels by the turnover of organic P in soil with extracellular enzymes, enhancing the utilization of soil P, and providing P nutrition for plants [48,49]. It was proven that PSB A. pittii could promote the phosphorus-cycling-related functions of the rhizosphere bacterial community [50]. PSB belonging to the genus Pseudomonas, which were considered to be plant growth-promoting bacteria (PGRB), were reported to play remarkable roles in integrated nutrient management for improving the nutrient acquisition of plants by their metabolic activities, in down-regulating fungal pathogens in stressed soil conditions by secreting antifungal metabolites [51,52], and in producing ascorbate peroxidase under salt stress conditions [53].

Regarding the results for NFB, the total relative abundances of the phylum Cyanobacteria, genus Cyanothece, and the phylum Firmicutes, order Bacillales, and genus Paenibacillus increased significantly in the T1 group, while they were not found in the T2 treatment group (Supplementary Figure S6). Cyanobacteria are crucial for the global nitrogen cycle and beneficial for modern agriculture as effective biofertilizers [54]. Firmicutes have been reported to play important roles in soil organic matter (SOM) cycling in anoxic soil [55]. Species of Paenibacillus, which are aerobic or facultative anaerobic, Gram-positive bacteria, interact with plants as plant growth-promoting rhizobacteria (PGPR) and have the potential to become biological fertilizers to increase root growth and crop productivity in agriculture [56]. The increased abundance of the phylum Cyanobacteria (genus Cyanothece) and Firmicutes in the T1 treatment group, especially in the T1P1 group, implies a more favorable growth environment for NFB, and the further moderation of the microbial environment promotes peanut root and plant growth, as observed in the phenotype of the T1 group.

The abundance of Azospirillum in the phylum Proteobacteria was significantly increased in T1 compared to that in T2 (Supplementary Figure S6). Azospirillum, a well-known associative nitrogen fixer affiliated with facultative endophytic diazotroph groups, has been reported to have close associations with many crops, like wheat and oats. It is able to participate in associative nitrogen fixation with host plants by colonizing both the root surface and interior. Azospirillum is considered the starter of the biological nitrogen fixation (BNF) process of non-legume plants and directly improves root growth by increasing the uptake of water and minerals [57]. The increase in nitrogen fixation by Azospirillum species could be mainly determined by low oxygen in the rhizosphere environment and sufficient host photosynthates. The increased abundance of Azospirillum in the T1 soil compaction group indicates that in a certain compaction environment, it is more conducive to the growth of both peanut roots and Azospirillum species, leading to an improvement in nitrogen fixation levels.

The results show that the abundance of the order Rhizobiales in the T1 treatment group was significantly lower than that in the T2 group. Rhizobia are the most famous and crucial symbiotic nitrogen fixers, with a beneficial relationship through intercellular symbioses with legume plants by inducing nodules and providing fixed nitrogen for plant growth. The abundance of Rhizobia can be greatly positively affected by an adequately low level of combined nitrogen [58]. The results indicate that the content of combined nitrogen might also be influenced by soil compaction. The family Bradyrhizobiaceae of the order Rhizobiales was also observed in the lowest abundance in the T1P1 group among different soil compaction and P application levels. The most important process of symbiotic nitrogen fixation is closely involved in the formation of nodules. Additionally, the nodule number could indicate a further symbiotic nitrogen fixation activity. Previous reports have shown that the number and ability of nodules can be promoted by P fertilizer application [59,60]. The results show that the nodule number was significantly higher in the group with P applied than in the group without P (Figure 1). Meanwhile, the nodule number in the T1P1 group was higher than that in the T2P1 group (Figure 1). The results suggest that when supplying the same amount of P fertilizer, the T1 level is more suitable for nodule formation, which also means that growing plants can obtain more nitrogen by symbiotic nitrogen fixation.

For our research, the abundances of OTUs (including PSB and NFB) that changed significantly in different groups were affected by soil compaction and P application levels (Figure 3 and Figure 4). The results showed that the content of soil available P was an obviously important environmental factor, and was mainly affected by the amount of phosphate fertilizer applied in agricultural production. Phosphorus is important not only because it is an indispensable nutrient element for plant growth but also because it affects plants’ nitrogen absorption [60,61]. Both rhizobium symbiotic nitrogen fixation and legume development, especially contributing to root growth, require sufficient P as a crucial energy source [61,62]. P deficiency could also significantly negatively affect the process of nodule formation and the ability of symbiotic nitrogen fixation, while sufficient P fertilizer application promotes a higher nodule number [63]. According to the results of the Mantel test, the nodule number was significantly positively correlated with the content of soil available P, and the contents of root N/P, which are consistent with previous reports. Root NITS activity was also positively correlated with the soil available P content, suggesting that a sufficient P supply could promote symbiotic nitrogen fixation efficiency not only by increasing the nodule number but also by enhancing the activity of NITS (Figure 5).

The influence of soil pH on PBS and NFB showed that it had a relationship with down-regulated OTUs under T1 soil compaction treatments and up-regulated OTUs under P0 levels (Figure 5). The lowest pH values were detected in the T1P1 group, which may correlate with and promote the content of soil available P, root NITS activity, and root N/P contents. From the results of the Mantel test, soil pH (in the appropriate range; in this study, soil pH values ranged from 6.01 to 6.52) was significantly negatively correlated with the soil available P content, root NITS activity, the contents of root N/P, and nodule number. The result was affirmed by previous reports that suggested a lower soil pH, ranging from 6.0 to 7.0, could benefit crop availability of P uptake [59]. The pH value, one of the most important soil characteristics, has a great relationship with nutrient availability, the microbial community composition and diversity, plant response, and enzyme activity. This is due to the function of amino acids, which could be sensitively altered at different pH levels in the soil [64]. The activity of ACP, predominantly identified in acidic soil, was considered a crucial reference for the soil pH status. In this study, there was no significant correlation between soil/root ACP activity and the soil pH value, maybe because the ratio of alkaline phosphatase (AlkP)/ACP was more relevant to soil pH. This ratio has been reported as a preferable and sensitive indicator for soil pH [65]. According to the results, both ACP activity in roots and soil were significantly negatively correlated with the content of root N and nodule number (Figure 5). Plants grown in P deficiency could be urged to increase the production of ACP in roots as well as nodules to hydrolyze organic P into available P, thus improving P uptake [66]. Furthermore, extracellular ACP secreted into the soil has been identified to be involved in the conversion of purines into ureides during symbiotic nitrogen fixation.

The results of the Mantel test also showed that with the increase in soil compaction, the content of nitrate N decreased and the activity of NITS increased (Figure 5). A higher nitrate N content could reduce the nodule respiration rate and NITS activity, inhibiting the synthesis and accumulation of plant signal molecules such as flavonoids. This inhibition could further hinder the recognition and infection of Rhizobia to plant roots [67,68]. NITS is known as an enzyme system that plays an important role in the biological nitrogen fixation process, catalyzing N₂ into NH₃, which requires a significant amount of energy. Oxygen sensitivity is one of the important properties of NITS, as it is unable to catalyze biological nitrogen fixation under a high pressure of O₂. T1 soil compaction reduces the oxygen content in the soil, making it more suitable for NITS catalytic action. Meanwhile, NITS activity was negatively correlated with P supply levels in both the T1 and T2 groups. Phosphate application significantly reduced the activity of NITS in the soil, consistent with many previous reports [69,70,71].

Furthermore, the results indicated that the soil ACP content and root N content were positively correlated with beta diversity in the T2P0 group (Supplementary Figure S8). Under P-supplied conditions, the soil AP content and root P content were highly correlated with the beta diversity of both PSB and NFB in the T2P1 group, suggesting that PSB and NFB assembly processes were mainly affected by N and P nutrient conditions in soil and roots via an adaptive adjustment of the community composition [72]. Under T2P1 treatment, P fertilizer application and the root P content alleviated the N limitation of microbes compared with no P supply conditions under the T2P0 treatment. Soil ACP activity and the root N content were the main factors limiting beta diversity. Meanwhile, under T1 treatment, there were no significant factors correlated with PSB beta diversity, while soil compaction was obviously correlated with NFB beta diversity in T1P1. Root NITS activity was correlated with NFB beta diversity when no P fertilizer was applied under the T1 treatment. The results suggested that the beta diversity of PSB and NFB were closely related to N and P limitations in the environment. N and P, the most critical nutrient elements sustaining crop yield, have been reported to have interactions at ecological, agronomic, physiological, and even molecular levels [73,74]. PSB increases the available P content in the soil and simultaneously enhances the nodule number and dry weight of peanuts [75]. The co-limitation and synergistic interaction between N and P are ubiquitous in soil ecosystems, as production under sufficient N and P conditions is better than adding any one alone [76]. Our results are consistent with previous reports that demonstrated that P deficiency negatively influences N assimilation, and N/P co-limitation (interaction) positively correlates with microbial beta diversity [77].

Understanding the processes shaping bacterial community assemblages is of increasing importance and is mainly influenced simultaneously by two major processes, including neutrality-based stochastic processes and traditional niche-based deterministic processes [78,79]. Stochastic processes assume that the community composition is independent of species and is driven by dispersal, immigration, or drift [80]. In contrast, based on deterministic processes, we hypothesize that the community assembly is imposed by deterministic factors, such as abiotic environmental factors (pH, salt, and moisture), species traits, root exudates, and interactions with antagonism and synergism, and focus on trade-offs among coexisting species to illustrate the abundance and distribution of microorganisms [81]. However, the relative contribution of stochastic and deterministic processes differs according to environmental and ecological conditions. Multiple deterministic factors, including various environmental and habitat conditions in hosts, nutrient resources, and interspecies interactions, are crucial for microbial community assembly [82,83]. Still, numerous observed microbial community assemblies are mainly caused by stochastic processes, such as contingency, ecological drift, and dispersal limitation, in nature [84]. Since natural selection usually occurs at the individual population level, populations within a microbial community would undergo different functional-trait-specific assembly processes with a phylogenetic signal. This signal could be undetectable in the overall community [85]. Therefore, in this study, we identified the community assembly processes by focusing on two functional bacterial communities, PSB and NFB.

PSB assembly processes were dominated by strong drift processes, while NFB assembly processes were dominated by complementary homogenizing selection and drift processes (Figure 6). These results were quite different from those in the eastern China agricultural ecosystem, where homogenous selection dominated more than 50% of bacterial community assemblies in the soil [86]. Of particular interest, almost all βNTI values of PSB were between −2 and +2, implying that stochastic processes (drift processes) play a crucial role (>84%) in shaping the PSB community under different P applications and soil compactions. This may be because the soil AP content was not low enough, causing abiotic environmental nutrient stress, as the AP content in the soil without P application even reached 72 mg/kg soil (P-deficient peanut-cultivated soil generally refers to soil with an AP content of less than 15 mg/kg soil). This may be related to the annual fertilization of our experimental base. Ecological drift, an unambiguous stochastic process, means stochastic changes along with the relative abundances of species within a community due to random birth, death, and reproduction [87]. Drift would alter the community composition when there is no selection or weak selection, suggesting the critical importance of drift in shaping the microbial community [88]. The influence of drift would also be enhanced by a larger number of species, as there is more probability of arising stochastic variation within the community. Furthermore, functional redundancy of populations that share a similar function would raise neutrality and make them susceptible to drift processes [85]. We speculated that the abundance and diversity of the PSB community would be significantly affected by drift processes in this long-term fertilized soil.

Generally speaking, stochastic and deterministic processes are two interdependent complementary ecological forces and combine with each other in shaping the community structure [89]. Although selection and drift typically change in contrary directions, the functions of these two fundamental processes in shaping the community also depend on each other [85]. The community structure can be viewed as the outcome of the dynamic balance and interaction between selection and drift [90]. For the NFB community assembly, both homogenizing selection processes and drift processes were dominant processes, with homogenizing selection accounting for more than half in the T1P1 group, while drift processes accounted for more than half in other treatment groups. Homogenous selection has been identified to be responsible for more than half of bacterial community assemblies in the soil [86,91,92]. The difference in the T1P1 group indicated the different ecosystem characteristics of this treatment, and the association of plant and bacterium reached an equilibrium state, significantly benefiting plants under T1P1 conditions according to the plant phenotype analysis. Additionally, we supposed that the interaction between root secretions and recruitment signals, and NFB nitrogen-fixing activities would cause the directional enrichment of NFB in rhizosphere soil, as the greatest nodule number was identified in the T1P1 treatment group. Compared with the PSB community, the importance of deterministic processes in the NFB community was significantly large, which may be the reason for the lower alpha diversity of NFB. The diversity and composition of symbiotic microorganisms have a great relationship with the strength of host selection, which leads to the adaption to plant immune systems, roots exudates associated with plant growth, and nutrient acquisition from resources [93,94,95]. Additionally, the filtered microorganism can be united and modulated by the host plant to increase host fitness [96,97]. The enhanced deterministic processes in the NFB community might be caused by peanut habitat filtering strategy. Among the dominant genera of peanut rhizosphere NFB, the relative abundances of Bradyrhizobium and Anaeromyxobacter were in the top two, which often acted as N nutrient resource providers for both microbial and plant consumers and were also the major host-dependent beneficial bacteria for peanuts. The relative abundance of Bradyrhizobium in T1P1 was the lowest among treatment groups, suggesting differences in peanut root exudates and physiological metabolism under different soil compactions and P applications. The increase in abundance of NFB in other treatment groups suggest that under stress conditions, soil ecology alleviates the inhibition of peanut root nutrient absorption and plant growth and development by increasing rhizosphere nitrogen cycling. The different importance of ecological forces in PSB and NFB indicated that the filtering strategies of peanuts differed between different functional communities. Furthermore, βNTI of PSB significantly positively correlated with root NITS activity, while that of NFB had a greater relationship with soil ACP activity, the soil pH value, and the root N content (Supplementary Figure S9). The results suggested that βNTI of PSB and NFB were closely related to N and P limitations in the environment. Soil pH was a crucial environmental factor, and at acidic pH, the NFB community structure tended to increase phylogenetic evenness [98]. As most NFB were enriched at an environment with a neutral pH [99], we supposed that the increased environmental stress and filtering in the T1P1 group, which had the lowest pH values among treatment groups, led to an increased dominance of homogenizing selection processes in the NFB community. Soil pH is an important physicochemical property of soil, which has a great effect on the availability of nutrients for peanut. Meantime, soil pH influenced the biological nitrogen fixation process and nitrogen cycling. Maintaining an optimal pH level in the soil is beneficial for soil microbial communities, soil ecology, and peanut health. Our findings suggest that reasonable P fertilizer application could effectively maintain the soil pH level. As the main bacteria that form nodules with nitrogen-fixing plants, and symbiotic nitrogen-fixing bacteria were the main component of NFB, the community structure of NFB was strongly selected by P and N nutrition of peanut roots and the rhizosphere. For peanuts and other leguminous crops, rhizobia are important symbiotic partners, especially under high environmental pressure. So, we believe that the main reason for the decrease in the abundance of rhizobia in the rhizosphere is the result of peanut roots selecting based on their own nutritional needs.

PSB networks consisted of 483 edges, which were more complex than those of NFB networks, including 35 edges (Figure 7). However, the negative correlations in the PSB-NFB network could be interpreted as a competitive relationship between species [100]. The network analysis of potential relationships within and between significantly changed OTUs of the PSB and NFB communities revealed a stronger antagonism between NFB OTUs, as the proportion of OTUs negatively associated within the NFB network was 22.86%, which was significantly higher than that of PSB-PSB and PSB-NFB networks. The results were consistent with previous reports, as NFB colonize the rhizosphere was a comprehensive balance process of limiting nutrients, competitive interactions, and predation pressure [101]. Additionally, the higher ratio of positive correlations in the PSB network indicated a stronger synergism within most changed PSB OTUs. Network complexities of PSB were significantly higher than NFB according to the number of edges, nodes, and degrees. From the PSB-NFB network, the results showed that NFBOTU389, NFBOTU70, NFBOTU583, NFBOTU58, NFBOTU276, NFBOTU366, and NFBOTU137 had close relationships with PSB OTUs based on the degrees of the interaction network. Both PSB and NFB were two main functional PGPB in rhizosphere soil, which could increase soil fertility and improve plant growth, nutrient uptake, and yield increase due to their abilities to improve the availability and biosynthesis of P and N, especially in assisting against nutrient deficiency stress [102,103]. The nitrogen fixation process was energy-consuming, supported not only by the function of microorganisms decomposing the nutrient resource pool in the rhizosphere but also by the supplement from root exudation of photosynthetic production. The nodule formation process mediated by the interaction between symbiotic nitrogen-fixing bacteria and plants suggested a closer synergistic relationship, indicating that root exudates from the peanut plant could greatly influence the competition. The increased availability of nutrients from roots might lessen competition between rhizosphere microorganisms [98]. This means that sufficient P fertilizer is more beneficial for root nodule formation and promotes root nitrogen absorption. These findings of potential interventions and adjustments in P fertilizer application may optimize nodule formation and nitrogen fixation. The increased ratio of downregulated NFB OTUs under the two P application levels in T2 soil compaction (66.7%) compared with that (55.0%) in the T1 level implied more competitive associations in the T2 level, which suggested decreased root exudates and changes in the growth state of plant roots under different soil compactions. The same phenomenon was also observed in the PSB community. The decreased rhizosphere soil compaction may have a more pronounced influence on both the state of root–bacteria interactions and the enhanced rhizosphere microorganism competition under looser soil compaction levels.

Rhizosphere microorganisms can not only interact with plants but also influence each other. For example, Pseudomonas fluorescens, a kind of PSB, and Rhizobium spp., which are affiliated with NFB, may have joint effects on plant adaptability with increased beneficial effects [104] and remain a steady microbial community for long-term coexistence during stress [105]. The results showed that the ratio of positive interactions was much higher than that of negative interactions in PSB-NFB networks (Figure 7). Positive associations generally mean three main relationships of bacteria, including mutualism, protocooperation, and commensalism, while negative associations signify suppression and competition [106]. Our results indicated that relationships between PSB and NFB include major synergism and minor competition. Additionally, PSBOTU644 has been identified belonging to the genus Bradyrhizobium, which also was an important NFB of peanut and confirmed by blasting at the National Center for Biotechnology Information (NCBI) with 100% query cover and 92.11% identity to B. diazoefficiens, which was the widely predominant nitrogen-fixing endosymbiont for soybean nodules in neutral to acidic soil [107]. The result indicated that PSBOTU644 could have the function of both phosphate solubilization and symbiotic nitrogen fixation. The mechanisms for B. diazoefficiens to solubilize insoluble P might depend on the existence of NH₄⁺ [108]. The finding enhanced our knowledge of PSB and NFB, and a better understanding of the mechanism for the use of biofertilizers by P solubilizing and concurrently symbiotic nitrogen fixation needs to be proved by further studies.

Network hubs were not identified from PSB, NFB, and PSB-NFB networks, while only seven module hubs were found in PSB-NFB networks, suggesting that the structure of networks was not hub-based (Figure 8). The topological features of networks varied between PSB and NFB, as the distribution of the degree for NFB and PSB-NFB networks followed power law distributions [109], which are often identified in biological networks [98], while the PSB network did not follow the power law distribution pattern, indicating that interactions between PSBOTUs were equally likely distributed among PSB [110]. That would mean neutral assembly and stochastic processes may play a more crucial role than that of deterministic processes for the PSB community (Figure 6A), which was consistent with previous bacterial community assemblage results [111]. For the PSB-NFB network, twelve of fifteen connectors were PSBOTUs (mainly derived from the phyla Actinobacteria and Gemmatimonadetes and the genera Streptomyces and Gemmatirosa), implying that PSB play a much more important function for the link among modules in the interaction network, while seven module hubs were all NFBOTUs indicating that PSB coexist with NFB to collaborate with mutualistic symbiosis. The network connectors of both PSB and NFB were derived from phylum Proteobacteria (genera Bradyrhizobium and Azohydromonas) indicating that these microbial taxa were crucial for interactions among modules in each network. Both the genera Streptomyces and Bradyrhizobium, which were especially crucial in the network of endophytic microbes of peanut roots, could be able to increase P nutrient accumulation by mobilizing P resources in soil and enhance resistance by producing defense hormones [94]. Streptomyces were reported to produce diverse bioactive compounds that could promote the soil ecological balance and crop adaptability in agriculture [112,113]. Bradyrhizobium is well known as a crucial nitrogen-fixing bacteria with nodulate legumes, which is beneficial for crop production, soil health, and the soil nitrogen cycle [114]. These two genera provide reference for screening agricultural microbial agents that will better serve agriculture. The subsequent identification of these two bacterial strains and their functions will be an important work.

5. Conclusions

We concluded that the soil compaction treatment significantly affected the shifts in phosphate-solubilizing bacteria (PSB) and nitrogen-fixing bacteria (NFB) communities and the properties of peanut roots and rhizosphere soil. The shifted PSB OTUs had a close relationship with the soil AP content, soil ACP activity, soil NITS activity, and soil compaction. Meanwhile, the altered NFB OTUs had a close relationship with the soil AP content and soil compaction. PSB community assembly processes were dominated by strong drift processes, while NFB community assembly processes were dominated by complementary homogenizing selection and drift processes. The network analysis revealed that the complexities of PSB were much higher than NFB, and a stronger antagonistic relationship was identified within NFB. A summary of the main results and predicted relationships among peanut, soil, and PSB/NFB is given in Figure 9 and Figure 10. Furthermore, our results will provide reference for crop cultivation in compacted and low-phosphorus soils. The important phosphate-solubilizing and nitrogen-fixing bacteria screened in the interaction network in this study will become candidate microbial agents for alleviating soil compaction and low phosphorus. Taken together, these results improve our knowledge and understanding of how soil compaction and P application alter peanut growth and the effects on the alterations of PSB and NFB communities in peanut rhizosphere soil. Our findings would suggest that beneficial phosphorus-solubilizing and nitrogen-fixing bacterial agents may provide new ideas and methods for crop cultivation in compact and low-phosphorus soils. The next step of isolating and identifying the relevant microbial functions will be particularly important.