Rice soil was collected from paddy fields of Taizhou, Jiangsu, China. The soil was derived from the fluvial deposit (Wang et al., 2021). After collection, the soil was air-dried, sieved (2 mm mesh), and homogenized for soil property analyses and rice growth. Soil properties were determined, including pH (6.79, 1:5 of soil-to-water ratio), total organic carbon (TOC, 23.7 g kg^–1), total nitrogen (TN, 1.15 g kg^–1), alkali-hydro nitrogen (AN, 36.7 mg kg^–1), available phosphorus (AP, 27.4 mg kg^–1), and available potassium (AK, 260 mg kg^–1) (Wang et al., 2021). Wheat straw was collected from a wheat production field during the harvest season. After air-drying, the straw was crushed to powder form and analyzed for contents of carbon (C, 46.5%) and N (0.48%). The ratio of C/N in the straw was 96.9.

To assess the influences of straw return on soil microbial community, three soil treatments were performed in triplicates, namely, (1) control treatment without straw amendment, (2) straw treatment by adding 1% w/w wheat straw (dry weight basis, dw), and (3) straw treatment by adding 2% w/w wheat straw (dw). Straw was added at 1 and 2% to represent the straw return of ∼12 and 24 t ha^–1 under field conditions.

To allow for rice cultivation in a greenhouse, soils without or with straw addition were filled into pots using rhizo-bags separating bulk and rhizosphere. Initially, soils without or with straw amendment were added with fertilizers, including urea (250 mg N kg^–1 dry soil), calcium superphosphate (60 mg P kg^–1 dry soil), and potassium chloride (100 mg K kg^–1 dry soil). Then, 1.5 kg dry weight (dw) of soil was transferred into rhizo-bags (30 μm nylon mesh, diameter 7.5 cm × height 23 cm), which were then placed in the center of polyvinyl chloride pots (diameter 15 cm × height 23 cm). Then, another 4.5 kg dw of soil was weighed into the pot to fill the space outside of the rhizo-bags. The rhizo-bags allow small molecular substrates to penetrate but prohibit roots to penetrate, thereby being a good way to divide rhizosphere from bulk soils (Nie et al., 2015). Soils were flooded for 3 days prior to the transplantation of rice seedlings (cv. Nanjing 9108). Uniform seedlings 3 days after germination were transplanted, with one seedling in each pot. Following transplantation, rice was daily flooded to maintain ∼2 cm overlying water during the period from seedling to flowering, while at the filling stage, rice was under the alternation of wetting and drying conditions to improve rice yield. At the maturity stage, i.e., 122 days after transplantation, rice was moved out from the rhizo-bags. The rhizosphere soils inside the rhizo-bags were collected and sieved (2 mm) to remove roots. The bulk soils outside the rhizo-bags were also collected and sieved. The rhizosphere and bulk soils were used for the analyses of soil properties, soil bacterial community structure, and N-related functional genes.

The dried soils were mixed with CO₂-removed water at a soil:water ratio of 1:5 (w/v) prior to measurements of pH and electrical conductivity (EC) using corresponding electrodes. The soil was oxidized with potassium dichromate and measured for TOC using titration with ferrous ammonium sulfate. Soil TN was measured by an elemental analyzer (Vario MAX CNS, Elementar, Germany). AN was measured using the alkali solution diffusion method (Lu, 2000). Concentrations of soil ammonium (NH₄⁺) and nitrate (NO₃^–) were analyzed using ion chromatography (ICS-3000, Dionex, United States) following extraction with 2 M KCl. AK and AP were measured using an inductively coupled plasma optical emission spectrometer (Optima 5300DV, PerkinElmer, United States).

A total of 18 soil samples, collected from bulk and rhizosphere of control and straw treatments, were subjected to soil bacterial community diversity and composition analysis using Illumina sequencing. Genomic DNA in the soil samples was extracted using a FastDNA SPIN Kit. The bacterial 16S rRNA gene was amplified using primers [515F (5′–GTGCCAGCMGCCGCGG–3′), 907R (5′–CCGTCAATTCMTTTRAGTTT–3′)] that target V4–V5 region. The thermal profile of PCR included an initial denaturation at 95°C for 3 min, 30-cycle denaturing at 95°C for 30, annealing at 58°C for 1 min, extension at 72°C for 45 s, and a final extension step at 72°C for 10 min. Equal amounts of PCR products from different samples (barcoded) were mixed, purified, and quantified prior to Illumina sequencing at Shanghai Lingen Genomics Institute, China (Wang et al., 2015).

The sequence data were processed using the Quantitative Insights into Microbial Ecology toolkit (QIIME, version 1.7.0) the data presented in the study are deposited in the GenBank Data repository, accession number PRJNA847362. After removing the low quality or ambiguous reads, the identified sequences were then clustered into operational taxonomic units (OTUs) at a 97% similarity level, and the representative sequence (i.e., the most abundant sequence) was assigned to taxonomy by an RDP classifier (version 2.2; Wang et al., 2007). Based on the OTU matrix, principal coordinate analysis (PCoA) was performed to assess the differences in overall community composition between bulk and rhizosphere with different amounts of the straw amendment (Lozupone et al., 2005). Moreover, heatmaps based on genus level and redundancy analysis (RDA) were conducted using R software (version 2.14.0) and the community ecology package vegan (2.0–4) to identify soil properties (including pH, EC, TOC, TN, AN, NH₄⁺–N, NO₃^––N, AP, and AK) contributing to the altered bacterial community (Oksanen et al., 2013). Envfit function (999 permutations) analyses were used to remove environmental variables that insignificantly contributed to the total soil bacterial community variance. Significant differences in the major genus of bacterial community composition between rhizosphere and bulk soils or between control and straw treatments were assessed by Student’s t-test (p < 0.05).

Primer pairs of nifHF/nifHR (Rosch et al., 2002), Arch-amoA-F/Arch-amoA-R (Francis et al., 2005), amoA-1F/amoA-1R (Rotthauwe et al., 1997), nirK-F1aCu/nirK-R3Cu (Throbäck et al., 2004), nirS-cd3aF/nirS-R3cd (Hallin and Lindgren, 1999), and nosZ-F/nosZ-1622R (Throbäck et al., 2004) were used to quantify nifH, archaeal and bacterial amoA (AOA and AOB), nirK, nirS, and nosZ genes in soil samples using real-time PCR system (ABI 7500, American), respectively (Supplementary Table 2). Following DNA extraction, each gene was amplified using a 20 μl system, i.e., 10 μl SYBR^® Premix Ex Taq™ II (2×), 0.4 μl ROX reference Dye II (50×), 0.8 μl forward primer (10 μM), 0.8 μl reverse primer (10 μM), 6 μl ddH₂O, and 2 μl template DNA (20 ng). Real-time PCR conditions for target genes are listed in Supplementary Table 2. Gene standard curves were constructed based on gradient dilutions of standard plasmids containing each gene with known copy numbers. To ensure correct amplification, DNA extracts were highly diluted to eliminate the inhibition. To ensure no contamination during qPCR, negative controls without DNA templates were included in each patch of gene amplification. In addition, gene quantification of each sample was performed in three parallel real-time PCR reactions, with reactions of efficiencies > 90% and correlation coefficients r² > 0.99 being accepted. Target gene copy numbers of each sample were calculated from the standard curves and expressed on a basis of soil dw (copies g^–1 dw soil).

For control treatment without straw return, there were significant differences in soil properties between rhizosphere and bulk (Table 1), indicating the sole rhizosphere effects. Soil TN (1.04 g kg^–1) and NO₃^––N (21.6 mg kg^–1) concentrations were significantly higher in the rhizosphere than in bulk soil (0.98 g kg^–1 and 9.65 mg kg^–1). In contrast, the rhizosphere showed significantly lower NH₄⁺–N concentration (6.18 mg kg^–1) than bulk soil (7.91 mg kg^–1). The pH of the rhizosphere was 7.09, significantly higher than that of bulk soil (6.97). There were insignificant differences in other soil properties.

For bulk soil, changes in soil properties with straw addition were observed (Table 1), indicating the sole straw effects. Excluding EC, AN, and AP, significant increases were observed in bulk soil pH (from 6.97 to 7.11), TOC (from 8.85 to 10.5 g kg^–1), TN (from 0.98 to 1.14 g kg^–1), NH₄⁺–N (from 7.91 to 10.0 mg kg^–1), NO₃^––N (from 9.65 to 14.5 mg kg^–1), and AK (from 70.0 to 146 mg kg^–1) with straw addition.

Changes in rhizosphere soil properties were also observed following straw addition (Table 1), indicating the interactions of rhizosphere and straw. Like bulk soil, EC, AN, and AP insignificantly changed with straw addition. In addition, NH₄⁺–N concentrations (6.22–6.31 mg kg^–1) of straw-treaded rhizosphere soils were also not significantly different from that of control rhizosphere soil (6.18 mg kg^–1). In comparison, straw addition significantly increased rhizosphere soil pH (from 7.09 to 7.22), TOC (from 8.43 to 11.1 g kg^–1), TN (from 1.04 to 1.23 g kg^–1), NO₃^––N (from 21.6 to 34.4 mg kg^–1), and AK (from 47.3 to 140 mg kg^–1).

Nitrogen content in the root, shoot, and rice from paddy soils with different amounts of straw addition are listed in Supplementary Table 1, showing that straw addition significantly increased the N content in the root, shoot, and rice from paddy soils.

Rarefaction of observed species showed that even at a sequencing depth of 290,704, the diversity of soil bacteria continued increasing rapidly with increasing sequencing depth (Figure 1A), suggesting the high diversity of soil bacteria. Insignificant differences in Chao, Shannon, Simpson, and Coverage index at the sequencing depth were observed between bulk and rhizosphere soils or between control and straw treatments (Table 2), suggesting that rhizosphere, straw, and their interactions had no effects on internal bacterial community diversity.

The major bacterial community compositions at class level were Actinobacteria (3.53–11.1%), Alphaproteobacteria (3.84–11.1%), Anaerolineae (6.22–22.0%), Bacilli (13.4–22.7%), Clostridia (6.99–20.4%), Deltaproteobacteria (3.21–5.19%), and Gammaproteobacteria (4.19–10.7%) (Figure 1B). The major bacterial community compositions at genus level were Anaerolinea (1.12–7.57%), Bacillus (9.67–13.7%), Clostridium (0.71–3.97%), Defluviicoccus (1.04–2.54%), Fonticella (0.58–1.20%), Nocardioides (0.87–2.84%), Tumebacillus (1.16–2.60%), with Bacillus being the dominant genera (Figure 1C).

The PCoA showed that the bacterial community compositions of the rhizosphere soils were significantly different from that of the bulk soils for control treatment or straw treatment at 1% (Figure 2A). However, for straw addition at 2%, the bacterial community compositions of bulk soils were overlapped with those of rhizosphere according to Adonis analysis (p > 0.05).

Regardless of bulk and rhizosphere soils, the bacterial community compositions of soils with straw addition were significantly separated from those of soils without straw addition (p < 0.05). This indicated that straw addition significantly changed the bacterial community composition of both bulk and rhizosphere soils, while root growth significantly altered the bacterial community composition for soils without and with a lower rate of straw addition. At higher straw application, the relatively stronger influence of straw might have obscured the effect of the rhizosphere on bacterial community compositions.

By envfit function (999 permutations), soil pH, TN, and TOC were found to be significantly correlated with the bacterial community composition at the OTU level (Figure 2B). RDA showed that variation in these three factors together explained 41.5% of soil bacterial community composition variation in bulk and rhizosphere soils with or without straw addition.

The genera whose relative abundance significantly varied with straw, rhizosphere, or their interaction effects were identified (Figure 3). By comparing bulk and rhizosphere soils without straw addition, a total of nine genera showed significant (p < 0.05) differences in relative abundances. However, by comparing bulk and rhizosphere soils treated with 1 or 2% straw, three or four genera were identified (Figures 3A–C), suggesting that the rhizosphere effect might be weakened when straw was returned to the soil.

For the straw effect, the number of differentially abundant taxa between bulk soils without and with the addition of straw at 1% was 13, while the number was 23 when straw was added at 2% (Figures 3D,E). These results suggested that compared to the rhizosphere effect, the straw effect affected more genera, highlighting the possible stronger impacts of straw amendments on soil bacterial community composition.

For the interaction effect of rhizosphere and straw, we observed 16 and 22 differentially abundant genera in rhizosphere soils when comparing control and 1% straw treatments and comparing control and 2% straw treatments (Figures 3F,G).

Bulk and rhizosphere soils without or with straw amendment varied in N-cycling functional genes (Figure 4). For control soils without straw addition, the abundances of archaeal amoA, nirK, and nosZ genes were significantly higher in the rhizosphere, but the abundance of bacterial amoA gene was lower compared to bulk soils. There were insignificant differences in the abundances of nifH and nirS genes between bulk and rhizosphere soils without straw addition. In comparison, for soils with 1% straw application, the gene abundances of archaeal amoA and bacterial amoA were lower in the rhizosphere than in bulk, while other genes were similar between rhizosphere and bulk. For soils with 2% straw application, nifH gene abundance was significantly higher in the rhizosphere compared to bulk soils.

For bulk soils, straw amendment significantly decreased the gene abundances of archaeal amoA and bacterial amoA, which decreased at higher extents at a higher straw application rate (Figure 4). The gene abundances of nirS and nosZ were also significantly decreased in bulk soils with straw addition, but unaffected by the straw application rate. No obvious difference in nifH and nirK gene abundances was observed in bulk soils with the straw amendment.

Like bulk soils, the straw amendment also significantly decreased the gene abundance of archaeal amoA, bacterial amoA, nirK, nirS, and nosZ in rhizosphere soils (Figure 4). At a higher rate of straw addition, the gene abundance of archaeal amoA was decreased to a higher extent, while the gene abundance of bacterial amoA, nirK, nirS, and nosZ was decreased at similar extents when 1 and 2% straw were amended. In comparison, the gene abundance of nifH was increased by 2% straw addition, while it was not affected by 1% straw addition.