Long-term effects of straw and straw-derived biochar on soil aggregation and fungal community in a rice–wheat rotation system

Experimental design

An in situ 6-year field study was operated with a typical rice–wheat rotation system in a sandy loam soil at Zhuanghang Field Station, Shanghai Academy of Agricultural Sciences, China (30°53′N, 121°23′E). Treatments were denoted in triplicate as follows: (i) CK (no fertilizer application), (ii) CF (conventional inorganic fertilizer application), (iii) CS (the same total pure nitrogen application as in CF plus three t ha⁻¹ straw returning to the field), and (iv) CB (the same total pure nitrogen application as in CF plus one t ha⁻¹ straw-derived biochar returning to the field). The experimental plots were arranged in a completely randomized block with each area being 60 m².

The application amounts of pure N, P, and potassium (K) for rice were 225, 112.5, and 255 kg ha⁻¹, respectively. For wheat season, the corresponding amounts were 180, 90, and 204 kg ha⁻¹, respectively. The N deficiencies in CS and CB were complemented by inorganic N fertilizer. The N, P, and K fertilizers used in the experiment were urea, calcium superphosphate, and potassium sulfate, respectively. Straw or biochar was added to the soil surface and thoroughly mixed with a depth of approximately 15 cm. Rice/wheat straw and the derived biochar were continuously sent back to the land as amendments in wheat/rice season. Pyrolysis was anaerobically performed at 500–600 °C in a vertical charcoal furnace (ECO-5000; Wuneng Environment Co., Ltd, Zhejiang, China) for biochar preparation. The basic characteristics of the straw and biochar are listed in Table S1.

Soil sample collection and analysis

In June 2016, after wheat harvest, samples were aseptically taken from the 0–20 and 20–40 cm soil layers using a five-point sampling method. The samples were stored in polyethylene bags at low temperature and immediately returned to the laboratory. Then, the samples were sieved to remove roots, rocks, and litter. Approximately 100 g of sieved soil was subjected to different aggregate classifications and subsequently frozen at −80 °C for fungal community analysis. The residual portions of soil samples were further air-dried for physicochemical property determination.

Soil pH was measured by potentiometry with a soil-water ratio of 1:2.5 (Mettler Toledo FE20 Plus, Shanghai, China). Gravimetric moisture content was determined by weighing the soil, followed by drying the soil samples at 105 °C for 24 h and weighing the soil again. NH₄⁺–N and NO₃⁻–N were assayed with Nessler’s reagent and the phenol disulfonic acid colorimetric methods, respectively. SOC and total nitrogen (TN) contents were determined with an elemental analyzer (Elementar, Langenselbold, Germany) and the KDN-08C automatic nitrogen determination instrument (Xinrui, Shanghai, China), respectively.

Aggregate size distribution

Soil water-stable aggregates were separated by an agglomerate analyzer according to the wet sieving method with some modifications (Elliott, 1986). Briefly, a series of sieves was used to obtain three aggregate classes: macroaggregates (>0.25 mm), microaggregates (0.053−0.25 mm), and silt clay (<0.053 mm). First, the soil samples were slowly wetted for 5 min using sterile water. The analyzer was then shaken up and down at 20 times min⁻¹ for 5 min with the column submerged in water throughout the whole process. The fractions remaining on each sieve were collected at the end of sieving. Aggregates in the beakers were stored at −80 °C for microbial analysis, and the remaining portions were weighed to analyze the relative indices after being oven-dried.

The mean weight diameter (MWD) represents aggregate stability and was calculated using Eq. (1): (1) $$\text{MWD}={\displaystyle \sum _{i=1}^n}{X}_i{W}_i$$

The geometric mean diameter (GMD) was calculated using Eq. (2): (2) $$\text{GMD}=\text{exp}\left(\frac{{\sum }_{i=1}^n{W}_i\ln X_i}{{\sum }_{i=1}^n{W}_i}\right)$$ where i corresponds to each collected fraction, and X_i and W_i refer to the mean diameter and the dry weight of i-size aggregates, respectively.

The fractal dimension (D) was obtained by regression analysis using Eq. (3): (3) $$\text{(}3-D\text{)l}g\left(\frac{d_i}{d_{\text{max}}}\right)=\text{l}g\left[\frac{W_{\left(\delta \le d_i\right)}}{W_0}\right]$$ where d_i, d_max, W_{(δ ≤ di)}, and W₀ refer to the average diameter of an i-size aggregate, the maximum diameter of all aggregates tested, the weight of aggregates with particle size < d_i, and the total weight of aggregates, respectively. Finally, D was obtained by regression analysis.

Fungal DNA extraction and high-throughput pyrosequencing

Aggregate DNA was extracted using the MoBio PowerSoil Soil® DNA Isolation Kit (12888) according to the manufacturer’s instructions. An aliquot of the DNA extracted from each sample was used as a template for amplification after its quality was confirmed by spectrophotometer (RS232G; Eppendorf, Hamburg, Germany). The internal transcribed spacer (ITS) regions were amplified with the primers ITS1F (5′-CTTGGTCATTTAGAGGAAGTAA-3′) and ITS1R (5′-GCTGCGTTCTTCATCGATGC-3′). PCR amplification was conducted as follows: 30 s of initial denaturation at 98 °C, followed by 35 cycles of denaturation at 98 °C for 15 s, annealing at 50 °C for 30 s, extension at 72 °C for 30 s, and a final extension at 72 °C for 5 min. Sequencing was performed on a MiSeq PE300 platform at Majorbio Biotechnology Co. Ltd (Shanghai, China).

The original data fragments were quality-controlled using Trimmomatic software, merged with FLASH and analyzed using Quantitative Insights into Microbial Ecology software. Operational taxonomic units (OTUs) with 97% similarity were clustered using USEARCH 7.0 and analyzed against the UNITE database (Release 6.0 http://unite.ut.ee/index.php) with a confidence threshold of 70%. All sequences have been deposited in the NCBI Short Reads Archive database (Accession Number: SRP161464).

Statistical analysis

The data were analyzed on the free online Majorbio I-Sanger Cloud Platform (www.i-sanger.com). Principal coordinates analysis (PCoA) was carried out to investigate the fungal community differences between four different treatments based on OTU results using R software. Significant differences between the trial treatments were determined by ANOVA analysis using SPSS 16.0 software (SPSS Corp., Chicago, IL, USA). Significant differences between means were determined with the least significant difference (LSD) test at the p = 0.05 level. Figures were prepared using ORIGIN 8.0 (OriginLab Corporation, Northampton, MA, USA).

Soil agrochemical properties

Soil nutrient contents were measured to investigate the influence of different fertilization treatments on soil physiochemical properties (Table 1). For the 0–20 cm soil depth, CB had the highest SOC content compared to CF and CK (p < 0.05); the difference with CS was not significant. Compared to CF, CS and CB both significantly increased the SOC and TN contents in 20–40 cm soil (p < 0.05). CS had a lower soil pH than CK in the 0–20 cm layer (p < 0.05), which may be due to more anions being released during straw decomposition. In 20–40 cm soil, CB had a higher electrical conductivity (EC) value than CS (p < 0.05).

Water-stable aggregate distribution and stability analysis

Macroaggregates were the dominant aggregate size class in both 0–20 and 20–40 cm layers (Fig. 1). CS and CB, compared with CF, dramatically increased the macroaggregate content by 17.77–18.87% and 33.55–50.87% in the 0–20 and 20–40 cm soil layers, respectively, with CB performing more remarkably (p < 0.05). In terms of microaggregates, compared to CF, CS decreased its mass proportion in the 0–20 cm soil layer by 11.61%; CB reduced its content in 20–40 cm soil layer by 25.84% (p < 0.05). For silt clay content, compared to CF, CS and CB showed dramatic decreases of 39.46–66.12% and 40.40–50.99% in the surface soil and deep soil, respectively (p < 0.05).

Aggregate stability is an influential factor governing soil erodibility (Ahmadi et al., 2011). MWD and GMD have commonly been used as indicators to reflect agglomeration aggregation and the stability of soil aggregates. Compared to CF, CS and CB increased MWD and GMD values at soil depths of 0–20 cm (16.92–29.23%, 29.41–47.06%) and 20–40 cm (23.915–30.43%, 39.13–43.48%), respectively; CB exhibited a more prominent effect than CS (p < 0.05) (Table 2), indicating that organic amendment facilitates aggregate stability. D was used to indicate the size, distribution, and homogeneity of the soil particles; the higher the value, the poorer the soil permeability (Ahmadi et al., 2011). CS and CB remarkably reduced the D-value by 1.22–1.49% and 0.44–0.60% at the 0–20 and 20–40 cm soil depths, respectively, which verified that straw and straw biochar promoted the uniform distribution of soil aggregates.

Sequence data and alpha diversity

A total of 4,245,052 sequences with an average length of 258 bp were obtained from all 72 samples. Fertilization affected soil fungal abundance in different aggregates (Fig. S1). In the 0–20 cm soil layer, the number of fungal OTUs was inversely related to the aggregate size; CB harbored the smallest number of OTUs of the treatments (636, 831, and 872 in macroaggregate, microaggregate, and silt clay, respectively), which was inconsistent with previous reports that the proportion of fungal groups in macroaggregates is high, indicating a major role in their formation. In the 20–40 cm soil, no such remarkable tendency was observed.

Alpha analysis of the fungal community under different treatments at different soil depths was estimated using Student’s t-test (Fig. 2). The Chao1 estimator, abundance-based and coverage estimator (ACE), and observed richness (Sobs) were employed as community richness indices (Liu et al., 2015). Significant variances were observed among the four treatments for macroaggregates in surface soil. The Chao 1 and ACE values of CS were higher than those of CK (p < 0.05, p < 0.01, respectively); higher Sobs values were observed in CS than in CK, CF, and CB (p < 0.05). As for both microaggregates and silt clay, no significant difference was observed between the four treatments in the 0–20 cm soil layer, suggesting that fungi played a more important role in large aggregate formation in CS (Li et al., 2018). The Shannon and Simpsoneven indices, calculated by richness and species abundance, were used as diversity and evenness indices, respectively, within each individual sample. Compared to CS, CB decreased the Shannon index in macroaggregates in 0–20 cm soil (p < 0.05); no difference was observed with the other treatments. In deep-layer soil, microaggregates in CF had a dramatically lower Simpsoneven value than CK (p < 0.05), which suggested that long-term chemical fertilization probably impairs the evenness of the fungal community.

Comparison of fungal community structure

PCoA was used to investigate the variations in fungal community structure caused by long-term fertilization and aggregate distribution (Fig. 3). For the surface soil, samples of CB were separated with the other three treatments by PC1 (35.48%); CS samples were gathered at the bottom left corner of the graph, apart from CK and CF, by PC2 (19.86%). CK and CF showed similar fungal community structures, as they were located in the upper part of PC2. With respect to the 20–40 cm soil layer, CF and CB were separated from CS and CK by PC2 (19.66%); PC1 (30.39%) approximately differentiated CB from CK and CF. Non-metric multidimensional scaling analysis based on Pearson distances, in accordance with the PCoA results, presented stress values of 0.098 and 0.091 for top soil (0–20 cm) and deep soil (20–40 cm), respectively (Fig. S2). The fungal community structures in CS and CB were significantly different from those in CF and CK at the 97% OTU level, and that of CS was significantly different from that of CB.

Fungal community composition

Fertilization, agglomerate levels and sampling depth all can largely alter the community composition of soil fungi. In the top layer of soil, the phyla Ascomycota, unclassified_k_Fungi, and Basidiomycota constituted over 97% of sequences (Fig. S3). Ascomycota was the most prevalent phylum in the soil; its members are known for their ability to degrade lignin, but it also includes many sugar fungi that utilize simple substrates (Chen et al., 2013). The other less dominant phyla were Zygomycota, Glomeromycota, and Chytridiomycota. At the class level, fertilization management played an important role in the relative abundance of fungi (Figs. 4 and 5). Figures 5A–5J showed the statistically differences of the representative classes in the aggregates in 0–20 and 20–40 cm, respectively. In the 0–20 cm soil layer, CB, compared to CF, decreased Dothideomycetes in all three aggregates and Glomeromycetes in silt clay (p < 0.05). Leotiomycetes contents in macroaggregates and silt clay in the 0–20 cm soil were significantly higher in CB than CF (p < 0.05). A similar phenomenon was observed in the deep-layer (20–40 cm) soil to variable extents. Fertilization decreased and increased the relative abundances of Dothideomycetes (in both macroaggregates and silt clay) and Sordariomycetes (in macroaggregates), respectively, with CB significantly differing from CK (p < 0.05). In microaggregates, fertilizer application resulted in trends of decreased Chytridiomycetes and unclassified_k_Fungi. CF possessed higher Glomeromycetes content in silt clay than CK and CB (p < 0.05).

The statistical analysis results for the relevant abundances of the dominant fungal genera are shown in Table 3 (0–20 cm) and Table 4 (20–40 cm). At 0–20 cm, CB, compared to CF, reduced the contents of Monographella (macroaggregate), Pyrenochaetopsis (microaggregate and silt clay), and Gibberella (microaggregate) and increased Articulospora content (macroaggregate) (p < 0.05). CS also lowered the relative abundances of Pyrenochaetopsis (silt clay) and Gibberella (microaggregate) relative to CF (p < 0.05). As for the 20–40 cm soil, CB, compared with CF, decreased Pyrenochaetopsis (microaggregate and silt clay) and increased Neurospora (macroaggregate) (p < 0.05). CS had lower Pyrenochaetopsis (microaggregate) and higher Neosetophoma (silt clay) contents than CF (p < 0.05). Additionally, a small number of “unclassified” sequences were identified in this study, which means that these soils contained several unidentified fungi.