Soil Organic Matter, Aggregates, and Microbial Characteristics of Intercropping Soybean under Straw Incorporation and N Input

Abstract

Soil organic matter (SOM), soil aggregates, and soil microbes play key roles in agriculture soil fertility. In intercropping systems, the influences of straw incorporation and N input on the dynamics of soil physicochemical and microbial properties and their relationships are still unclear. We explore the changes in soil physicochemical and microbial properties with two straw managements, i.e., wheat straw incorporation (SI) and straw removal (SR), and four N supply rates for intercropped soybean, i.e., 60 (N60), 30 (N30), 15 (N15), and 0 (N0) kg N ha⁻¹, in the wheat–maize–soybean relay strip intercropping systems. The results showed that SOM and SOM fractions contents, soil macroaggregate stability, and microbial and fungal α-diversity, e.g., Chao1 and Shannon indices, increased through straw incorporation and N input. The α-diversity was significantly positively correlated with soil physicochemical characteristics. Compared with SR, the relative abundance of Actinobacteria and Mortierellomycota in SI increased, but the relative abundance of Proteobacteria, Acidobacteria, and Ascomycota in SI decreased. In SI treatment, soil physicochemical characteristics and microbial diversity improved through N input, but that difference was not significant between N60 and N30. In conclusion, SI+N30 was the most effective way to maintain soil fertility and reduce the N fertilizer input in the wheat–maize–soybean relay strip intercropping.

1. Introduction

Food security has become a severe issue with the continuous growth of the world’s population and food consumption. Food production must grow sustainability to meet the world’s food security and sustainability needs [1]. Intercropping is recognized as a sustainable agricultural development model in the world [2,3]. It has multiple ecological benefits, e.g., yield advantages over single crops, the efficient use of fertilizers, and maintaining soil fertility [4]. The wheat–maize–soybean relay strip intercropping system can take full use of farmland on temporal and spatial scales. Intercropping systems, including legumes, can increase soil N input and avoid farmland degradation. In China, considerable wheat straws were burned or removed in the wheat–maize–soybean relay strip intercropping system. Meanwhile, the overuse of N fertilizer for soybean in the system results in resource waste and environmental pollution. A sustainable straw and fertilizer management strategy for the wheat–maize–soybean relay strip intercropping system is urgently needed.

Straw incorporation improves soil fertility and maintains farmland quality by improving soil structure and increasing soil organic matter (SOM) [5,6,7]. Understanding the SOM dynamics is essential for the sustainability of the agroecosystem. However, the short and medium changes of SOM are difficult to detect because of their high temporal and spatial variability [7]. The labile SOM fractions, e.g., dissolved organic matter (DOM), light fraction organic matter (LFOM), and microbial biomass [7,8], can be considered as fine indicators of soil quality. The labile SOM influences soil function in specific ways and it is sensitive to changes in soil management practice [9,10]. The application of mineral N fertilizer alone does not improve soil organic carbon (SOC) content [11,12] while combining mineral N fertilizer with organic manures can increase SOM and SOM fractions contents [13]. Similarly, Moran et al. [14] observed that straw incorporation coupled with mineral N input can promote soil SOM formation. Huang et al. [15] found that mineral N incorporated with straw can increase SOC stock at 0–60 cm soil depth. In intercropping, straw incorporation, e.g., straw covering, can improve water use efficiency and alleviate greenhouse gas emissions in arid areas [16,17]. However, the response of SOM and SOM fractions to straw incorporation and N input in the intercropping system is not well understood.

The SOM and its fractions alter affected by soil biological and physicochemical properties. Soil microbes play crucial roles in soil SOM formation, nutrient availability, and C and N cycling in the ecosystem [18,19,20]. Soil microbial community diversity and composition are shaped by straw management, N input, tillage methods, etc. [10,21,22]. During the progress of straw decomposition and SOM formation, bacteria dominate in the incipient stage, and fungi dominate in the later stage due to their specific functions. However, the straw decomposition is limited by cereal straw’s high C: N ratio [23]. A reasonable N input contributes to balancing the C: N ratio and promotes residual decomposition by soil microbes [24,25]. Furthermore, the soil-specific microbes, e.g., Rhizobium sp. and Azotobacter vinelandii, can produce extracellular compounds to increase the soil aggregate stability [26]. Soil aggregates are critical for soil nutrient storage, stabilizing soil’s biological, physical, and chemical properties [19,27,28]. Soil macroaggregates (diameter > 250 μm) are used to assess land use’s effect on organic matter turnover [29,30]. Besides, soil aggregates can harbor microbes and improve microbial community structures [31]. Tillage, straw management, fertilizer application, and cropping systems directly affect soil aggregate formation and its stability [32,33,34,35]. Therefore, understanding the effects of straw management practices and N input on soil physical properties and microbial community diversity in intercropping systems is important for sustainable agricultural development.

Our early studies showed that maize–soybean strip intercropping with reduced N input could increase soil fertility and farmland productivity [36,37] by strengthening the soybean biological N fixation capacity [2] and reducing N loss [38]. Straw incorporation practices provide suitable approaches to improving resource use efficiency, increasing soil fertility and avoiding environmental pollution during agricultural production [16,17,39,40]. However, the effects of straw incorporation coupled with N input in the SOM, SOM fractions, and the soil microbial properties of soybean in the wheat–maize–soybean relay strip intercropping system are still unclear. Hence, the objectives of this study were to (1) evaluate the effects of straw incorporation coupled with N input on soil SOM and SOM fractions, soil aggregates stability, and soil microbial properties of soybean, and (2) explore the relationships between soil physicochemical properties and soil microbial activity in the wheat–maize–soybean relay strip intercropping. We hypothesized that straw incorporation combined with N input would increase the SOM and its fractions contents, increase soil aggregation stability, and enhance the soil microbial activity of soybean in the wheat–maize–soybean strip intercropping system. This study will provide a new perspective on efficiently used resources and the sustainable development of intercropping systems.

2. Materials and Methods

2.1. Site Description

The field experiments were conducted in Renshou County, Sichuan Province (30°16′ N, 104°00′ E), Southwest China. The climate of the experimental site was subtropical monsoon humidity, with an average annual precipitation of 1110.7 mm and a temperature of 17.9 °C. The precipitation and temperature of the soybean cropping season are shown in Figure 1. The local soil is anthrosol with a clay loam texture, and the physical and chemical properties are shown in

Table 1. The soil’s physical and chemical properties.

Indexes	Organic Matter (g kg−1)	Total N (g kg−1)	Total P (g kg−1)	Total K (g kg−1)	pH
Content	7.85	0.61	0.84	22.66	8.21

.

2.2. Experimental Design and Field Management

During the 2018 to 2020 cropping seasons, the field experiments were carried out with a two-factor split-plot design with three replications. The main factor was straw management, i.e., full straw incorporation (SI) and full straw removal (SR). The sub-factor was the N input of intercropped soybean, i.e., conventional N employed by local farmers (N60, 60 kg N ha⁻¹), reduced N by 50% (N30, 30 kg N ha⁻¹), reduced N by 75% (N15, 15 kg N ha⁻¹), and zero N (N0). The present study contains 24 plots (two straw treatments × four N input levels × three replications). In the SI treatment, all of the wheat straw was crushed into pieces (0.05 m) and incorporated into the soil by rotary tillage (about 20 cm depth) after wheat harvest every year (Figure 2). In the SR treatment, crop stubble was less than 5 cm in height, and all of the wheat straw was removed from the field (Figure 2). The wheat straw used was 2801 kg ha⁻¹ in 2019 and 2363 kg ha⁻¹ in 2020.

In the wheat–maize–soybean relay strip intercropping system, a wide-narrow row planting was adopted (1.6 m and 0.4 m for wide and narrow rows), and the total ratio of wheat-to-maize-to-soybean rows was 4:2:2. Wheat was planted in the wide rows with row spacings of 0.25 m (Figure 2). After the wheat harvest, two rows of soybean were sown in the wheat strips with row spacings of 0.4 m. Intercropped maize was sown in narrow rows with row spacings of 0.4 m. The distance was 0.425 m between wheat and maize rows, which was 0.6 m between maize and soybean (Figure 2). All plots were 6.0 m in length and 5.0 m in width. The planting density of wheat was 2,000,000 plants ha⁻¹, 58,863 plants ha⁻¹ for maize, and 117,726 plants ha⁻¹ for soybean.

Urea (46% N), superphosphate (12% P₂O₅), and potassium chloride (60% K₂O) were used as N, P, and K fertilizers. The P and K fertilizers were applied as base fertilizers for each crop, i.e., 36 kg P₂O₅ ha⁻¹ and 54 kg K₂O ha⁻¹ for wheat, 120 kg P₂O₅ ha^−1, and 120 kg K₂O ha⁻¹ for maize, and 60 kg P₂O₅ ha⁻¹ and 52.5 kg K₂O ha⁻¹ for soybean. N fertilizer for wheat (150 kg N ha⁻¹) and soybean was applied as basal fertilizer, while N for maize was divided into two parts, i.e., 120 kg N ha⁻¹ as basal fertilizer and 120 kg N ha⁻¹ as topdressing. Wheat fertilizers were broadcast in the planting strips and rotationally plowed into the soil before wheat sowing. Maize fertilizers were strip placed at a distance of 20 cm away from maize rows. Soybean fertilizers were hole placed at a distance of 10 cm away from the soybean (Figure 2).

Wheat was sown on 15 November 2018 and 14 November 2019 and harvested on 14 May 2019 and 8 May 2020. Maize was sown on 9 April 2019 and 5 April 2020 and harvested on 27 July 2019 and 29 July 2020. Soybean was sown on 8 June 2019 and 9 June 2020 and harvested on 3 November 2019 and 28 October 2020. The cultivars of wheat (Triticum aestivum L.), maize (Zea mays L.), and soybean (Glycine max L. Merr.) were Zhongkemai-138, Denghai-605, and Nandou-25, respectively.

2.3. Soil Sampling and Storage

Soil samples (0–20 cm depth) of soybean were collected with soil anger (2 cm diameter and 20 cm depth) at the fifth trifoliolate stage (V5, 17 July 2019 and 21 July 2020), the beginning seed stage (R5, 29 August 2019 and 8 September 2020), and the full-maturity stage (R8, 3 November 2019 and 28 October 2020) of soybean [41]. Three individual samples were collected per plot (Figure 2), then thoroughly mixed and sieved through a 2 mm mesh to remove plant tissues, roots, and rocks. Finally, the soil samples were divided into three parts: one subsample was air-dried for SOM, LFOC, LFN, HFOC, and HFN determination; one subsample was stored at 4 °C for DOC, DON, MBC, and MBN measurement; and the other subsample was stored at −80 °C. In 2019, the soil samples at the R8 stage of soybean were used for real-time PCR quantification and Illumina MiSeq sequencing analysis to evaluate the soil microbial diversity.

2.4. Soil Aggregates and Bulk Density Analysis

The soil samples for the soil aggregates and bulk density (BD) assessment were collected after the soybean harvest. Soil clods were collected at 0–20 cm soil depth. Within each plot, five individual soil samples were collected. The fresh soil was gently stripped into 10–12 mm soil clods along the natural planes of weakness and then air-dried for soil aggregation analysis. Soil aggregate separation was performed according to Guo et al. [32]. The MWD and GMD were calculated according to Kihara et al. [34]. Two undisturbed soil cores from each plot at 0–10 cm and 10–20 cm depths with a volume of 100 cm³ were collected for soil BD measurement. The soil samples were oven-dried at 105 °C for 24 h, long enough to reach constant weight for weighting and BD calculation [5].

2.5. Soil Organic Carbon and Nitrogen Analysis

According to Chen et al. [22], the SOC content was determined. Then, SOM was calculated by multiplying the SOC percentage by a coefficient of 1.72 [12]. The soil LFOC, LFN, HFOC, and HFN were calculated according to Zhao et al. [9]. The soil DOC determination according to Chen et al. [6]. Soil DON, MBC, and MBN were determined according to Dong et al. [42].

2.6. Amplicon Sequencing

Total microbial genomic DNA samples were extracted from 0.5 g of fresh soil using the Faster DNA SPIN extraction kits (MP Biomedicals, Santa Ana, CA, USA), following the manufacturer’s instructions, and stored at −20 °C before further processing analysis. Primers 338F (ACTCCTACGGGAGGCAGCA) and 806R (TCGGACTACHVGGGTWTCTAAT) were used to amplify 16S rRNA from the V3-V4 region of bacteria [43], and primers ITS5F (GGAAGTAAAAGTCGTAACAAGG) and ITS2R (GCTGCGTTCTTCATCGATGC) were used to amplify ITS rRNA from the V1 region of fungi [44]. Sample-specific 7-bp barcodes were incorporated into the primers for multiplex sequencing. PCR was performed in 25 μL reactions containing 5 μL of Q5 reaction buffer (5×), 5 μL of Q5 High-Fidelity GC buffer (5×), 0.25 μL of Q5 High-Fidelity DNA Polymerase (5U μL⁻¹), 2 μL (2.5 mM) of dNTPs, 1 μL (10 uM) of each Forward and Reverse primer, 2 μL of DNA Template, and 8.75 μL of ddH₂O.

For bacteria, thermal cycling consisted of initial denaturation at 98 °C for 5 min, followed by 24 cycles consisting of denaturation at 98 °C for 30 s, annealing at 55 °C for 30 s, and extension at 72 °C for 45 s, with a final extension of 5 min at 72 °C. For fungi, thermal cycling consisted of initial denaturation at 98 °C for 5 min, followed by 28 cycles consisting of denaturation at 98 °C for 30 s, annealing at 55 °C for 45 s, and extension at 72 °C for 45 s, with a final extension of 5 min at 72 °C. PCR amplicons were purified with Agencourt AMPure Beads (Beckman Coulter, Indianapolis, IN, USA) and quantified using the PicoGreen dsDNA Assay Kit (Invitrogen, Carlsbad, CA, USA). PCR amplicons were purified with Agencourt AMPure beads (Beckman Coulter, Indianapolis, IN, USA) and quantified using the PicoGreen dsDNA Assay Kit (Invitrogen, Carlsbad, CA, USA). After the individual quantification step, the amplicons were pooled in equal amounts. Then, pair-end 2 × 300 bp sequencing was performed using the Illumina MiSeq platform with MiSeq Reagent Kit v3 at Shanghai Personal Biotechnology Co., Ltd. (Shanghai, China).

2.7. Bioinformatics Analysis

Sequence data analyses were mainly performed using QIIME and R packages (v3.2.0). OTU-level alpha diversity indices, such as the Chao1 richness estimator and Shannon diversity index, were calculated using the OTU table in QIIME. OTU-level ranked abundance curves were generated to compare the richness and evenness of OTUs among samples. The taxonomy compositions and abundances were visualized using MEGAN [45] and GraPhlAn [46].

2.8. Statistical Analysis

Statistical significance was performed with the two-way analysis of variance (ANOVA) and Fisher’s least significant difference method (LSD, α = 0.05) for SOM, SOM fractions, MWD, GMD, and soil microbial diversity index. The analyses were performed with SPSS v.22 (IBM Corp., Armon, NY, USA) and Microsoft Excel. Pearson correlation analysis was conducted to assess the relationship between soil physicochemical properties and soil microbial diversity. The redundancy analysis (RDA) was performed with the genescloud tools (a free online platform for data analysis, https://www.genescloud.cn, accessed on 10 August 2022) to explore the relationships between the composition of bacteria and fungi communities and soil physicochemical properties. The Mantel test was conducted to reveal the relationships between the soil microbial features and soil properties. The Mantel test was performed with the mantel_test() function of the ggcor package in R (v.4.0.5) (https://github.com/houyunhuang/ggcor, accessed on 10 August 2022). The correlations between soil characteristics were evaluated with the corrplot package in R (https://github.com/taiyun/corrplot, accessed on 10 August 2022). SigmaPlot v.14.0 (Systat Software Inc., San Jose, CA, USA) and Origin 2017 (OriginLab Corporation, Venice, FL, USA) were used to draw the figures.

3. Results

3.1. Soil Organic Matter

The SOM content of N60 and N30 in SI were higher by, respectively, 8.4~13.2% and 9.7~16.6% than in SR at the R5 and R8 stages (Figure 3). In SI, the SOM content of treatments with a N input exceeding 30 kg N ha⁻¹ was considerably higher than other treatments, while no significant differences were observed between N60 and N30. In contrast, the SOM content was independent of N input under SR. In N30, the SOM content was significantly increased by 11.7~16.6% in SI compared with SR at the R8 stage (Figure 3C,F).

3.2. Soil Light Fraction Organic C and N

The soil LFOC content was independent of N input under SR treatment (Figure 4A–F). In contrast, the soil LFN content significantly decreased with the decrease in N input in SR treatment (Figure 4G–L). In SI, the soil LFOC and LFN contents of treatments with a N input exceeding 30 kg N ha⁻¹ were significantly higher than other treatments (Figure 4B–F,G–L), while no differences were observed between N60 and N30 (Figure 4). When the N input exceeded 30 kg N ha⁻¹, the soil LFOC and LFN contents increased significantly in SI compared with SR (Figure 4B–F,G–L). On average, the soil LFOC and LFN content in SI was greater by, respectively, 43.3% and 16.1%, 46.5%, and 26.1% than in SR under the N60 and N30 treatments at the R8 stage.

3.3. Soil Heavy Fraction Organic C and N

In SR, the soil HFOC and HFN contents were independent of N input (Figure 5A–C,G–I). On average, the soil HFOC and HFN content were greater in SI than in SR, by 3.2% and 6.0% at the V5 stage, by 4.2% and 2.8% at the R5 stage, by 3.5% and 4.9% at the R8 stage, respectively. In SI, the HFOC of treatments with a N input exceeding 30 kg N ha⁻¹ was increased significantly compared to other treatments in 2020 at different growth stages (Figure 5D–F). In SI, although the HFN content decreased with the decrease in N input at the V5 stage, the HFN content was independent of N input in 2020 at the R5 and R8 stages (Figure 5J–L).

3.4. Soil Dissolved Organic C and N

With the decrease in N input, the soil DOC and DON contents were remarkably reduced under both SI and SR treatments (Figure 6). On average, the DOC and DON content in SI were higher by, respectively, 3.4% and 28.1%, 9.0% and 17.4%, 9.5% and 10.8% than in SR at the V5, R5, and R8 stages. In SR, the soil DOC and DON contents were significantly decreased in N30 compared with N60 in 2020 (Figure 6D–F,J–L). Compared with SR, the soil DOC and DON contents of treatments with a N input exceeding 30 kg N ha⁻¹ were significantly enhanced in SI. However, the differences in soil DOC and DON contents in SI were not significant between N60 and N30 (Figure 6).

3.5. Soil Microbial Biomass C and N

The averaged soil MBC and MBN content was higher in SI than in SR, by 20.7% and 32.9% at the V5 stage, by 14.1% and 26.1% at the R5 stage, and by 18.4% and 18.7% at the R8 stage, respectively. In SI, the soil MBC and MBN contents of treatments with a N input exceeding 30 kg N ha⁻¹ were significantly greater than other treatments, while no significant differences were observed between N60 and N30. In SR, with the decrease in the N input, the soil MBC and MBN contents were significantly decreased (Figure 7).

3.6. Soil Aggregates Stability and Bulk Density

There were no significant differences in the soil physical properties between different N application rates under both SI and SR treatments (Figure 8). The soil aggregate MWD and GMD in SI were higher by, respectively, 6.0% and 9.1% than in SR (Figure 8A,B,D,E). On the contrary, the soil BD in SI was decreased by −4.6% compared with SR treatment (Figure 8C,F).

3.7. Soil Microbial Community α-Diversity and Community Composition

The Chao1 index of soil bacteria and fungi was greater by, respectively, 4.0~31.1% and 7.8~186.1% in SI than in SR (Figure 9A,B). Compared with SR, the Shannon index of soil bacteria and fungi was increased by, respectively, 1.2~3.0% and 7.7~20.3% (Figure 9C,D). There were no differences in the Chao1 index of bacteria and fungi between N60 and N30 under SI treatment (Figure 9A,B). However, the Chao1 index of bacteria and fungi was significantly decreased in N30, N15, and N0 compared with N60 under SR treatment (Figure 9A,B). Moreover, the Shannon index of soil bacteria and fungi was independent with N input under the SI treatment, but it was significantly decreased in N0 under the SR treatment (Figure 9C,D).

The dominant bacteria phyla consisted of Proteobacteria (35.3%), Actinobacteria (21.0%), and Acidobacteria (16.4%) (Figure 9E), and the dominant fungi phyla consisted of Ascomycota (58.8%), Mortierellomycota (15.6%), Basidiomycota (11.6%), and Mucoromycota (0.5%) (Figure 9F). The relative abundances of Actinobacteria, Chloroflexi, Gemmatimonadetes, Rokubacteria, Latescibacteria, Mortierellomycota, and Mucoromycota were higher by, respectively, 12.6%, 32.2%, 0.7%, 19.0%, 28.6%, 20.5%, and 11.0% in SI than in SR. However, the relative abundances of Proteobacteria, Acidobacteria, Bacteroidetes, Ascomycota, and Basidiomycota were lower by 10.5%, 5.6%, 21.1%, 4.9%, and 18.8% in SI than in SR, respectively. With the decrease in N input, the relative abundances of Actinobacteria, Chloroflexi, Rokubacteria, Ascomycota, Basidiomycota, and Mucoromycota in SI were increased in contrast to SR. At the genus level, the others of bacterial and fungal populations accounted for, respectively, 70.5% and 58.5%, indicating that there are still a large number of bacterial and fungal species to be explored in the soil (Figure 9G,H). The dominant bacteria genera consisted of Subgroup_6 (9.7%), Sphingomonas (4.5%), and Lysobacter (2.9%) (Figure 9G). The dominant fungal genera consisted of Mortierella (15.6%), Botryotrichum (4.3%), and Aspergillus (4.1%) (Figure 9H). The relative abundance of Subgroup_6, Sphingomonas, Lysobacter, and Aspergillus was greater in SR than in SI.

3.8. Relationship between the Soil Environmental Factors and Soil Microbial Community

The results of the RDA analysis indicated that the first two axes explained 47.9% and 5.5% variations for the bacteria phyla community (Figure 10A). The soil HFN, MBC, MBN, DOC, MWD, GMD, and BD were the main factors affecting the soil bacteria community (Figure 10A). Notably, the soil DOC, DON, MBC, and MBN were positively correlated to the relative abundance of Actinobacteria; in contrast, those soil properties were negatively correlated to the relative abundance of Proteobacteria (Figure 10A). The soil HFN and BD were positively correlated to the relative abundance of Proteobacteria, Acidobacteria, and Bacteroidetes, whereas the abundances of other bacteria were negatively correlated with soil HFN and BD (Figure 10A). The first two axes explained 31.0% and 7.6% variations for fungi phyla community, and the soil DON, LFN, MBN, and DOC contents were the main factors affecting the soil fungi community (Figure 10B). The relative abundances of Mortierellomycota were positively related to all soil properties. In contrast, only BD positively correlated with Ascomycota (Figure 10B).

The results of the Mantel test and correlation analysis showed that soil LFN and MBC contents were significantly positively correlated to the α-diversity of bacteria (Figure 10C). The α-diversity of fungus was significantly positively correlated to the soil LFN, DOC, DON, MBC, and MBN content. Only soil MBC was significantly positively correlated with bacteria community composition, while no significant correlation was found between fungal community composition and soil environmental factors. The soil LFN content was notably positively correlated with DOC, DON, MBC, MBN, MWD, and GMD, respectively. Meanwhile, soil MBC was remarkably positively correlated with SOM, LFOC, LFN, DOC, DON, MBN, MWD, and GMD, respectively.

4. Discussion

4.1. Response of SOM and SOM Fractions to Straw Incorporation and N Application

The SOM was independent of N input in SR, while the SOM significantly increased in SI when the N input exceeded 30 kg N ha⁻¹. Although plant straw is one of the sources of SOM, plant straw characterized by a high carbon-to-nitrogen ratio has low decomposition rates [47]. A reasonable extra N input can optimize the C:N ratio and promote crop straw decomposition and SOM formation [14]. Although the MBC, MBN, DOC, and DON increased with N input in SR, those labile SOM fractions were greater in SI than in SR. This indicates strengthened microbe activities [7]. During the processes of straw decomposition, soil microbes transform soil carbon and nitrogen into MBN and MBC [42,48]. This was consistent in this study, that the MBC and MBN contents increased with N input, and SI achieved greater MBC and MBN contents than SR. In SI, the MBC and MBN contents were significantly greater in treatments with the N input exceeding 30 kg N ha⁻¹ than in the other. This indicated that 30 kg N ha⁻¹ was the potential threshold value that affected the straw decomposition in the current condition.

However, the improved microbial activity through N input did not lead to greater LFOC, HFOC, and HFN in SR. In IS, the LFOC, LFN, DOC, and DON were greater in treatments with the N input exceeding 30 kg N ha⁻¹ than in the others. On the one hand, nitrogen addition promotes plant root growth, and greater root biomass promotes the release of root exudates [49]. Then, the increase in the below-ground investment strengthens soil microbial progress [50], promotes residue decomposition [7], and increases nutrient cycling. On the other hand, the high stability of soil macroaggregate help to prevent labile SOM fractions from decomposing [51], and thus achieve greater LFOC, LFN, DOC, and DON. Indeed, the MWD and GMD of macroaggregates significantly increased in SI compared with SR. The MWD and GMD of the macroaggregates were significantly positively correlated to LFOC, LFN, DOC, and DON. The MWD and GMD were significantly negatively correlated to BD. Straw incorporation decreases soil BD by protecting the surface soil from compaction and promoting soil aggregate formation [34,52]. Moreover, the HFOC and HFN content were not different between SI and SR in 2019, while it was more significant in SI than in SR in 2020. The LFOC and LFN mainly consist of microbial debris and animal and plant residues [53]. Parts of the LFOC and LFN combined with the soil microbe’s polysaccharides and extracellular compounds, then sequestered within soil aggregates and format stable HFOC and HFN [9]. Therefore, the HFOC and HFN primarily contain high-density organo-mineral complexes and are more stable than LFOC and LFN [54].

4.2. Response of Soil Aggregate Stability to Straw Incorporation and N Application

Soil aggregates’ size distribution and stability are important indicators of soil physical quality, which impacts soil microbial activity, SOM dynamics, nutrient cycling, and crop production [19,34,55]. Our results showed that SI significantly increased soil macroaggregates’ stability. Independent of N input, the MWD and GMD of soil macroaggregates in SI were significantly higher than those in SR. This is consistent with previous studies [33,34]. Straw incorporation significantly increases the soil macroaggregates stability may be resulted from the accumulation and degradation of soil organic matter. Jastrow et al. [56] found that the inputs of organic residues are decomposed relatively rapidly into particles or colloids that are physically protected, slowing decomposition and promoting the development of stable microaggregates within macroaggregates. Moreover, new organic matter is responsible for the stability of soil macroaggregates [55]. Indeed, our results showed that the soil SOM and most SOM fraction contents were significantly positively correlated to the MWD and GMD of soil aggregates. Although the MWD and GMD of soil macroaggregates were greater in SI than in SR, the MWD and GMD were independent of N input. However, Stainsby et al. [33] found that the MWD increased by 21% when in the reduced N fertilizer input. The differences in the results may result from the different returned crop straws, straw return management, and land-use systems [34,52].

Previous studies have shown that straw incorporation significantly decreases soil BD [5,57]. These findings are in coincidence with our present study, that SI significantly decreased soil BD compared with SR. Firstly, a low soil BD result from the increase in soil SOM [57]. Indeed, we observed a negative correlation between soil BD and SOM and fractions. Secondly, straw incorporation reduces rainfall scour on the soil, protects the surface soil from compaction, promotes soil particles together, and forms aggregates [34,52]. In our study, SI significantly increased the soil MWD and GMD of soil aggregate compared with SR. In SI, the soil BD was significantly negatively correlated with MWD and GMD, suggesting that straw incorporation is conducive to soil aggregate stability.

4.3. Response of Soil Microbial Activity to Straw Incorporation and N Application

The Chao1 index of soil bacteria and fungi was remarkably enhanced in SI compared with SR. The Shannon index of bacteria was considerably higher in SI than in SR, whereas that of fungi did not differ between SI and SR. Those were consistent with Huang et al. [24]. However, the effects of straw incorporation on soil microbial diversity are still under debate. Li et al. [25] pointed out that straw incorporation significantly decreased the bacterial Chao1 index and does not affect the bacterial Shannon index. Similarly, Fu et al. [58] reported that straw mulching enhanced fungal richness and diversity while not affecting bacteria richness and diversity. The different effects are probably due to the variation in the straw management strategies, fertilizer input, tillage practices, and planting systems.

The Chao1 and Shannon indices had no differences between N60 and N30 in SI, while those indices in SR were significantly decreased in N30 compared with N60. The increased microbial richness and diversity may be due to the increased soil C and N content, and the soil physical properties’ stability [25,34]. Indeed, the bacterial Chao1 and Shannon index and the fungal Chao1 index were significantly correlated with soil organic C and N content and the MWD and GMD of soil macroaggregates. The Mantel test indicated that the LFN and MBC were notably correlated with the soil bacteria α-diversity, and the LFN, DOC, DON, MBC, and MBN were remarkably correlated with the soil fungal α-diversity. These results indicate that straw incorporation combined with N input can improve soil microbial diversity by changing the soil C and N content.

The most abundant bacterial phylum in the soil was Proteobacteria, followed by Actinobacteria and Acidobacteria, and the most abundant dominant fungal phyla in the soil was Ascomycota, followed by Mortierellomycota and Basidiomycota. Compared with SR, the relative abundances of Actinobacteria and Mortierellomycota increased in SI, but the relative abundances of Proteobacteria, Acidobacteria, Ascomycota, and Basidiomycota were decreased in SI. Those are confirmed by Cong et al. [10]. In contrast, straw mulching increased the relative abundances of Proteobacteria and Ascomycota but decreased the relative abundance of Actinobacteria compared with no straw mulching [58]. With the decrease in N input, the relative abundance of Proteobacteria was increased in SI, and Actinobacteria and Acidobacteria were decreased in SR. Besides, the relative abundances of Ascomycota, Basidiomycota, and Mucoromycota were increased with N input, while Mortierellomycota was decreased under different straw management. The changes in the soil microbial community composition are probably due to the changed soil physicochemical properties by straw incorporation and N application treatments, and most of these microbes were in close relationships with the soil C and N transformation [10,24,58].

The Mantel test indicated that the soil MBC content was notably correlated with soil bacteria community composition, whereas the fungal community composition was independent of the soil’s physicochemical properties. As a dominant bacteria flora, the phylum Proteobacteria was positively correlated with HFN and BD, while the phylum Actinobacteria was positively correlated with most of the SOM fractions. Kausar et al. [59] documented that Actinobacteria play a key role in straw decomposition. As the dominant fungal flora, the phylum Ascomycota was negatively correlated with SOM fractions, while Mortierellomycota was positively correlated with SOM fractions. Those indicated that Mortierellomycota is more important than Actinobacteria in its response to increased N input in SI. Therefore, short-term straw incorporation combined with a reasonable N input can enhance microbial diversity and optimize the community composition by increasing soil physical stability and the contents of the SOM and SOM fractions.

5. Conclusions

In this study, we have shown that wheat straw incorporation could increase the soil SOM and SOM fractions, improve soil physical structure, and enhance the soil microbial community diversity of intercropped soybean in contrast to straw removal in the wheat–maize–soybean relay strip intercropping system. N application significantly impacted the soil SOM and SOM fraction and microbial community diversity but had no significant impact on the soil’s physical structure under both straw incorporation and straw removal treatments. The labile SOM fractions, e.g., LFOC and LFN were the most sensitive indicator to respond to straw incorporation, and the DOC and DON were the most sensitive indicators to respond to N application. There were no significant differences in soil SOM and SOM fractions and microbial diversity between 60 kg N ha⁻¹ and 30 kg N ha⁻¹ under the straw incorporation treatment. However, those were significantly decreased in 30 kg N ha⁻¹ in contrast to 60 kg N ha⁻¹ under the straw removal treatment. The relative abundance of Actinobacteria and Mortierellomycota were positively correlated with SOM fractions, suggesting that an increase in SOM fractions may occur with changes in relative abundance. The Mantel test showed that the soil LFN, MBC, MBN, DOC, and DON were remarkably correlated with the soil microbes community. In conclusion, straw incorporation combined with 30 kg N ha⁻¹ has the best effects on improving soil SOM, physical structure, and microbial activity in the wheat–maize–soybean relay strip intercropping system.