ABSTRACT
Straw return is widely applied to increase soil fertility and soil organic carbon storage. However, its effect on N2O emissions from paddy soil and the associated microbial mechanisms are still unclear. In this study, wheat straw was amended to two paddy soils (2% w/w) from Taizhou (TZ) and Yixing (YX), China, which were flooded and incubated for 30 d. Real-time PCR and Illumina sequencing were used to characterize changes in denitrifying functional gene abundance and denitrifying bacterial communities. Compared to unamended controls, straw addition significantly decreased accumulated N2O emissions in both TZ (5071 to 96 mg kg–1) and YX (1501 to 112 mg kg–1). This was mainly due to reduced N2O production with decreased abundance of major genera of nirK and nirS-bacterial communities and reduced nirK and nirS gene abundances. Further analyses showed that nirK-, nirS- and nosZ-bacterial community composition shifted mainly along the easily oxidizable carbon (EOC) arrows following straw amendment among four different soil organic carbon fractions, suggesting that increased EOC was the main driver of alerted denitrifying bacterial community composition. This study revealed straw return suppressed N2O emission via altering denitrifying bacterial community compositions and highlighted the importance of EOC in controlling denitrifying bacterial communities.
INTRODUCTION
Agricultural soil is a significant anthropogenic source of nitrous oxide (N2O) emissions (Akiyama, Uchida and Yamamoto 2011; Cai and Yan 2011). Rice is a staple food for more than 50% of the global population, therefore contributing greatly to N2O emissions (Khan et al. 2013; Liu et al. 2016). Inevitably, increasing N2O emissions are expected to occur to meet increasing food demands (Zhang, Chen and Vitousek 2013; Song et al. 2018). In addition, increasing atmospheric carbon dioxide and temperature rises are expected to increase soil N2O emissions, which requires preventative measures to reduce greenhouse gas outputs from croplands (van Groenigen, Osenberg and Hungate 2011, 2013). Compared to carbon dioxide, N2O has a long half-life (∼150 years) and 298 times higher global warming potential (van Groenigen, van Kessel and Hungate 2013). Elevated N2O accumulation in the atmosphere are expected to have destructive impacts on ozone for centuries (Lashof and Ahuja 1990). To mitigate N2O emissions, cost-effective soil amendments are required, with straw amendments proposed as such a solution (Ma et al. 2009; Wang et al. 2018).
China has 13 million ha of rice–wheat cropping, inevitably producing enormous amounts of straw residue. To alleviate air pollution from open burning, efficient recycling is required (Qu et al. 2012), while straw return is advocated as such a measure (Zhao et al. 2016). Straw return has been widely accepted as a measure to increase soil fertility and soil organic carbon (SOC) storage (Liu et al. 2014; Chen et al. 2017). The practice of crop straw return provides a source of readily available carbon (C), which has been shown to reduce N2O emissions from rice paddies (Ma et al. 2009). However, not all studies have shown suppressed soil N2O emissions; some have measured increased N2O emissions due to enhanced microbial activity with straw addition, which accelerated decomposition (priming) of soil inherent organic carbon (Millar and Baggs 2004; Zhu et al. 2015).
The uncertain effects of straw addition are dependent on complex microbial processes controlling N2O emissions. Denitrification is the dominant N2O-producing mechanism in paddy soil due to flooded conditions (Bateman and Baggs 2005), involving nitrate reduction (catalyzed by reductase napA and narG), nitrite reduction (catalyzed by nirK and nirS), nitric oxide reduction (catalyzed by qnorB and cnorB) and N2O reduction (catalyzed by nosZ) (Ishii et al. 2011; Jones et al. 2013; Wei et al. 2015). The first three processes lead to N2O production with nitrite reduction being the limiting step, while N2O reduction to N2 is a N2O consumption process (Braker and Tiedje 2003; Henry et al. 2004; Wang et al. 2017). With balancing of production and consumption processes, effects of straw return on N2O emissions can vary considerably depending on its specific effects on nirK-, nirS- and nosZ-type bacteria. Reduced N2O emissions can be observed and attributed to three mechanisms, (i) decreased N2O production due to decreased abundance of nirK- or nirS-type denitrifying bacteria; (ii) increased N2O consumption by increased abundance of nosZ-type bacteria and (iii) higher increase in N2O consumption than N2O production (Xu et al. 2018). However, limited studies have discriminated the predominant microbial mechanisms of straw return influences on N2O emissions.
Providing a source of readily available carbon, straw return can alter bacterial community composition through increasing SOC and changing soil carbon/nitrogen ratios (Henderson et al. 2010). However, studies have shown that total organic carbon (TOC) is often not a sensitive measure of straw return due to high background concentration of SOC (Haynes 2000; Zhao et al. 2016). Compared to TOC, labile organic carbon fractions including dissolved organic carbon (DOC), easily oxidizable carbon (EOC) and particulate organic carbon (POC) are more sensitive indicators of soil quality changes (Yan, Wang and Yang 2007; Benbi et al. 2015; Chen et al. 2017). However, there is a lack of information comparing the roles of different carbon fractions with straw addition in altering denitrifying functional community compositions. Given that EOC represents the more readily available carbon fraction to microbes than TOC, it was hypothesized that EOC rather TOC drives the shift in denitrifying bacterial community composition and denitrifying functional genes when straw is added to paddy soil, thereby leading to changes in N2O emissions.
To test this hypothesis, wheat straw was incorporated into two types of flooded rice–wheat rotation soils (silty soil and clayed soil). Quantification of nirK, nirS, and nosZ genes in soils after flooding and incubation was performed using real-time PCR. The composition of soil nirK-, nirS- and nosZ-type denitrifying bacterial communities was analyzed using Illumina sequencing technology. The specific objectives were to (i) determine effects of straw return on N2O emissions from paddy soils; (ii) investigate the response of soil properties, carbon fractions, relative abundance of nirK, nirS, and nosZ genes and compositions of nirK, nirS, nosZ-type denitrifying bacterial community to straw addition and (iii) explore microbial mechanisms of straw-induced changes in N2O emissions.
MATERIALS AND METHODS
Soil and straw
Two types of rice–wheat rotation soils (silty and clayed soils) were collected from paddy fields of Taizhou (TZ) and Yixing (YX) cities, Jiangsu, China, respectively. YX clayed soil was derived from lacustrine sediment, while TZ silty soil was derived from fluvial deposit. After collection, soils were air-dried, sieved (2-mm mesh) and homogenized for physicochemical property analyses and flooding incubation experiment. Wheat straw, collected from wheat cultivation base of Jiangsu Academy of Agricultural Sciences, was air-dried, crushed, sieved (40-mm mesh) and homogenized prior to amendment into soils.
Straw addition and flooded incubation
For each soil, ∼1.8 kg dry weight (dw) of soil was added with deionized water at a soil: water ratio of 1:1 for pre-incubation in the dark at 25°C for 30 days. Each soil was then homogenized prior to the transfer of soil slurries (500 g fresh weight) into wide-mouthed glass bottles (diameter 10 cm × height 15 cm), after which 100 mg N kg–1 dry soil as urea was added to serve as nitrogen fertilizer. Then, two treatments were prepared in triplicate, i.e. control treatment without straw amendment and straw treatment by adding 2% w/w wheat straw (dw basis). Following wheat straw addition, deionized water was added to maintain an overlying water depth of ∼2 cm. Bottles remained open and were incubated in dark at 25°C over a 30-d period during which water was replenished daily.
At time intervals of 0, 4, 7, 14 and 30 d, gas samples were collected from bottles as described by Wang et al. (2018). Briefly, bottles were sealed 2 h prior to the collection of air using gas tight syringes. After gas sampling, bottles were opened again for further incubation. At the end of the incubation period, soil porewater was collected using porewater samplers (3S–10, Institute of Soil Science, Chinese Academy of Sciences, Nanjing) and immediately filtered (0.22 μm) for analyses. Following porewater collection, soils were destructively sampled for the analysis of soil properties and denitrifying bacterial communities.
Chemical analyses
Gas samples were analyzed for N2O concentrations using an Agilent Technologies 7890A GC system equipped with an electron capture detector within 1 d of sampling (Wang et al. 2015). Soil pH and electrical conductivity (EC) were measured following mixing soil with carbon dioxide removed water at a soil: water ratio of 1:5 (w/v), while total carbon (C) and nitrogen (N) were determined using an element analyzer (CHN-O-Rapid, Heraeus, German). Soil organic matter (SOM) was measured as weight loss following heating of dried samples at 550°C for 4 h (Dean 1974; Yang, Engstrom and Rose 2010). Soil porewater samples were analyzed for NH4+ and NO3– using ion chromatography (ICS-3000, Dionex, USA) (Mou, Wang and Sun 1993).
To determine the roles of different carbon fractions in N2O emissions, fractionation of SOC in control and straw-amended soils was assessed as described by Chen et al. (2017). Four fractions of SOC were determined including TOC, DOC, EOC and POC. Briefly, TOC was determined by oxidation with potassium dichromate and titration with ferrous ammonium sulfate (Lu 1999); EOC was determined as described by Blair, Lefroy and Lise (1995); POC was determined using the method of Cambardella and Elliott (1992); DOC in porewater samples was measured using a total organic carbon (TOC) analyzer (Shimadzu TOC-Vcph, Japan).
Real-time PCR of denitrifying functional genes and 16S rRNA illumina sequencing
Genomic DNA in soil collected at day 30 was extracted using FastDNA SPIN Kits (MP Biomedicals, Solon, OH, USA). Amplification of 16S rRNA and denitrifying bacterial nirK, nirS and nosZ genes was performed utilizing primer pairs of nirK-F1aCu/nirK-R3Cu, nirS-cd3aF/nirS-R3cd and nosZ-F/nosZ-1622R, respectively (Hallin and Lindgren 1999; Throbäck et al. 2004). Each sample was amplified using the following conditions in a final volume of 50 μL: 25 μL SYBR® Premix Ex TaqTM II (2 ×), 1 μL BSA (20 mg mL–1), 1 μL forward primer (10 μM), 1 μL reverse primer (10 μM), 18 μL ddH2O and 4 μL template DNA (100–200 ng). The thermal profile included an initial denaturation step at 95°C for 2 min, 40 cycles of 95°C for 30 s (denaturing), 60°C for 1 min (annealing), 72°C for 1 min (extension), followed by a final extension step for 5 min at 72°C. PCR products were purified, quantified using real-time PCR system (Light Cycler 480II, Switzerland) and sent to Shanghai Lingen Biotech Limited Company (Shanghai, China) for 16S rRNA Illumina sequencing. Details on the quantification of nirK, nirS and nosZ genes can be sourced from Wang et al. (2015).
Data processing and statistical analyses
Accumulated N2O production (mg kg–1 soil) during the 30-d incubation was calculated as the area under the N2O emission rate time curves using SigmaPlot 10.0.
Illumina sequence data were processed using the Quantitative Insights into Microbial Ecology toolkit (QIIME, version 1.7.0) and R software (version 2.14.0) (Caporaso et al.2010). After removing low quality or ambiguous reads, the qualified sequences were clustered into Operational Taxonomic Units (OTUs) at 97% similarity level by default using the Usearch (version 7.0) and then the most abundant sequence in each OTU was selected as the representative sequence and was assigned to taxonomy using a RDP classifier (version 2.2) on the QIIME platform. The indexes of Sobs, Simpsoneven and Shannon were calculated to compare alpha-diversity of denitrifying bacterial communities harboring nirK, nirS and nosZ genes. Beta-diversity was generated to assess the differences in community composition between control and straw treatments using principle co-ordinates analysis (PCoA) at OTU level based on the Bray–Curtis metric on the QIIME platform. To identify soil properties influencing denitrifying bacterial communities, canonical correlation analyses (CCA) or redundancy analyses (RDA) was conducted using R software. Variance test of significance was used to identify environmental variables contributing to the total soil denitrifying bacterial community variance using R software. To further identify effects of denitrifying bacterial communities on N2O emissions, spearman correlation heatmaps between N2O and denitrifying bacterial community composition at genus level were created using R software. The phylogenetic trees on Genus bar belonging to denitrifying bacterial community composition based on the maximum likelihood method were established using FastTree (version 2.1.3).
A one-way analysis of variance (ANOVA) and Student's t-test were conducted to test the significant (P < 0.05) difference in soil properties, N2O emissions and functional gene abundances between treatments. All statistical analyses utilized the IBM SPSS Statistics 23.0 and all data were processed using Microsoft Excel. Diagrams were created using SigmaPlot 10.0 and R software (version 2.14.0).
RESULTS
Soil properties
Soil pH, EC, the ratio of soil carbon to nitrogen (C/N), NH4+ and NO3– concentration in soil porewater, SOM and different carbon fractions in both TZ and YX soils were measured after a 30-d flooded incubation (Table 1). For control treatments, soil characteristics varied considerably between the two soils, with pH, EC and C/N in TZ being significantly (P < 0.05) higher than in YX, although NO3– concentrations were significantly (P < 0.05) lower. TZ soil showed significantly lower SOM (2.22% vs. 5.19%), TOC (16.4 vs. 21.9 g kg–1) and POC (0.82 vs. 2.98 g kg–1) than YX, but DOC (4.94 vs. 3.20 mg kg–1) and EOC (1.29 vs. 1.24 g kg–1) concentrations were similar between the two soils.
Soil Characteristics after a 30-d Incubation Period in Two Paddy Soils from Taizhou (TZ) and Yixing (YX), Jiangsu Province, China without (Control) and with Straw Addition (n = 3).
| TZ | YX | |||
|---|---|---|---|---|
| Treatments | Control | Straw | Control | Straw |
| pH | 6.59 ± 0.03b | 7.2 ± 0.08a | 6.36 ± 0.09c | 6.36 ± 0.08c |
| EC (μs cm−1) | 97.3 ± 19.2ab | 117.5 ± 10.5a | 58.3 ± 23.2b | 84.5 ± 29.9ab |
| C/N ratio | 15.0 ± 3.97b | 26.2 ± 0.35a | 7.93 ± 0.42c | 12.7 ± 0.25b |
| NO3− (mg L−1) | 2.16 ± 0.35b | 2.47 ± 0.59ab | 3.03 ± 0.47a | 2.92 ± 0.19ab |
| NH4+ (mg L−1) | 4.32 ± 1.10a | 4.29 ± 0.18a | 5.36 ± 0.32a | 5.12 ± 1.46a |
| Soil organic matter (SOM, %) | 2.22 ± 1.05c | 3.37 ± 1.37c | 5.19 ± 0.08b | 6.65 ± 0.17a |
| TOC (g kg−1) | 16.4 ± 1.13c | 26.4 ± 2.81a | 21.9 ± 1.13b | 28.3 ± 2.55a |
| DOC (mg kg−1) | 4.94 ± 0.89b | 22.0 ± 10.62a | 3.20 ± 0.97b | 9.40 ± 2.54b |
| EOC (g kg−1) | 1.29 ± 0.41c | 2.43 ± 0.83b | 1.24 ± 0.51c | 3.73 ± 0.92a |
| POC (g kg−1) | 0.82 ± 0.02d | 1.18 ± 0.04c | 2.98 ± 0.14b | 10.4 ± 0.88a |
| TZ | YX | |||
|---|---|---|---|---|
| Treatments | Control | Straw | Control | Straw |
| pH | 6.59 ± 0.03b | 7.2 ± 0.08a | 6.36 ± 0.09c | 6.36 ± 0.08c |
| EC (μs cm−1) | 97.3 ± 19.2ab | 117.5 ± 10.5a | 58.3 ± 23.2b | 84.5 ± 29.9ab |
| C/N ratio | 15.0 ± 3.97b | 26.2 ± 0.35a | 7.93 ± 0.42c | 12.7 ± 0.25b |
| NO3− (mg L−1) | 2.16 ± 0.35b | 2.47 ± 0.59ab | 3.03 ± 0.47a | 2.92 ± 0.19ab |
| NH4+ (mg L−1) | 4.32 ± 1.10a | 4.29 ± 0.18a | 5.36 ± 0.32a | 5.12 ± 1.46a |
| Soil organic matter (SOM, %) | 2.22 ± 1.05c | 3.37 ± 1.37c | 5.19 ± 0.08b | 6.65 ± 0.17a |
| TOC (g kg−1) | 16.4 ± 1.13c | 26.4 ± 2.81a | 21.9 ± 1.13b | 28.3 ± 2.55a |
| DOC (mg kg−1) | 4.94 ± 0.89b | 22.0 ± 10.62a | 3.20 ± 0.97b | 9.40 ± 2.54b |
| EOC (g kg−1) | 1.29 ± 0.41c | 2.43 ± 0.83b | 1.24 ± 0.51c | 3.73 ± 0.92a |
| POC (g kg−1) | 0.82 ± 0.02d | 1.18 ± 0.04c | 2.98 ± 0.14b | 10.4 ± 0.88a |
Different upper letters indicated significant (P < 0.05) differences in various treatments in both soils.
Soil Characteristics after a 30-d Incubation Period in Two Paddy Soils from Taizhou (TZ) and Yixing (YX), Jiangsu Province, China without (Control) and with Straw Addition (n = 3).
| TZ | YX | |||
|---|---|---|---|---|
| Treatments | Control | Straw | Control | Straw |
| pH | 6.59 ± 0.03b | 7.2 ± 0.08a | 6.36 ± 0.09c | 6.36 ± 0.08c |
| EC (μs cm−1) | 97.3 ± 19.2ab | 117.5 ± 10.5a | 58.3 ± 23.2b | 84.5 ± 29.9ab |
| C/N ratio | 15.0 ± 3.97b | 26.2 ± 0.35a | 7.93 ± 0.42c | 12.7 ± 0.25b |
| NO3− (mg L−1) | 2.16 ± 0.35b | 2.47 ± 0.59ab | 3.03 ± 0.47a | 2.92 ± 0.19ab |
| NH4+ (mg L−1) | 4.32 ± 1.10a | 4.29 ± 0.18a | 5.36 ± 0.32a | 5.12 ± 1.46a |
| Soil organic matter (SOM, %) | 2.22 ± 1.05c | 3.37 ± 1.37c | 5.19 ± 0.08b | 6.65 ± 0.17a |
| TOC (g kg−1) | 16.4 ± 1.13c | 26.4 ± 2.81a | 21.9 ± 1.13b | 28.3 ± 2.55a |
| DOC (mg kg−1) | 4.94 ± 0.89b | 22.0 ± 10.62a | 3.20 ± 0.97b | 9.40 ± 2.54b |
| EOC (g kg−1) | 1.29 ± 0.41c | 2.43 ± 0.83b | 1.24 ± 0.51c | 3.73 ± 0.92a |
| POC (g kg−1) | 0.82 ± 0.02d | 1.18 ± 0.04c | 2.98 ± 0.14b | 10.4 ± 0.88a |
| TZ | YX | |||
|---|---|---|---|---|
| Treatments | Control | Straw | Control | Straw |
| pH | 6.59 ± 0.03b | 7.2 ± 0.08a | 6.36 ± 0.09c | 6.36 ± 0.08c |
| EC (μs cm−1) | 97.3 ± 19.2ab | 117.5 ± 10.5a | 58.3 ± 23.2b | 84.5 ± 29.9ab |
| C/N ratio | 15.0 ± 3.97b | 26.2 ± 0.35a | 7.93 ± 0.42c | 12.7 ± 0.25b |
| NO3− (mg L−1) | 2.16 ± 0.35b | 2.47 ± 0.59ab | 3.03 ± 0.47a | 2.92 ± 0.19ab |
| NH4+ (mg L−1) | 4.32 ± 1.10a | 4.29 ± 0.18a | 5.36 ± 0.32a | 5.12 ± 1.46a |
| Soil organic matter (SOM, %) | 2.22 ± 1.05c | 3.37 ± 1.37c | 5.19 ± 0.08b | 6.65 ± 0.17a |
| TOC (g kg−1) | 16.4 ± 1.13c | 26.4 ± 2.81a | 21.9 ± 1.13b | 28.3 ± 2.55a |
| DOC (mg kg−1) | 4.94 ± 0.89b | 22.0 ± 10.62a | 3.20 ± 0.97b | 9.40 ± 2.54b |
| EOC (g kg−1) | 1.29 ± 0.41c | 2.43 ± 0.83b | 1.24 ± 0.51c | 3.73 ± 0.92a |
| POC (g kg−1) | 0.82 ± 0.02d | 1.18 ± 0.04c | 2.98 ± 0.14b | 10.4 ± 0.88a |
Different upper letters indicated significant (P < 0.05) differences in various treatments in both soils.
For both soils, compared to control treatments, straw application at 2% w/w significantly increased C/N (from 15.0 to 26.2 for TZ and from 7.93 to 12.7 for YX). In addition, pH, EC, NO3– and SOM also tended to increase with straw addition, although the increase was insignificant. No effect of straw addition on NH4+concentration was observed for either soil. The most obvious influence of straw addition was significant (P < 0.05) increases in TOC, DOC, EOC and POC in TZ (1.61-, 4.45-, 1.88- and 1.44-fold) and YX (1.29-, 2.94-, 3.01- and 3.49-fold higher). Increased C with straw addition was also reflected by increase in SOM in YX (from 5.19% ± 0.08% to 6.65% ± 0.07%) and TZ (2.22% ± 1.05% to 3.37% ± 1.37%).
N2O emissions
For control treatments without straw addition, N2O was emitted from YX immediately after flooding, reaching a peak (11.9 mg kg–1 h–1) at day 4 then sharply decreased to low level (0.24 mg kg–1 h–1) at day 14, lasting until day 30. In contrast, N2O emissions from TZ was delayed, starting at day 4 (0.14 mg kg–1 h–1), peaking at day 14 (17.7 mg kg–1 h–1) and decreasing to 0.04 mg kg–1 h–1 at day 30 (Fig. 1A and B). The N2O emission peak was significantly higher in TZ compared to YX; accumulated N2O production was 1501 and 5071 mg kg–1 from YX to TZ, respectively (Fig. 1C and D).
Temporal changes of N2O emissionrate (A, B), and accumulated N2O emissions (C, D), in two paddy soils from Taizhou (TZ) and Yixing (YX), China without (Control) and with wheat straw addition (Straw). Error bars represent the standard deviation of triplicate analyses.
For both soils, straw amendments significantly decreased N2O emissions. Compared to control treatments, straw amendments significantly decreased peak N2O emission rates from TZ (17.7 to 0.03 mg kg–1 h–1) and YX (11.9 to 0.03 mg kg–1 h–1) (Fig. 1A and B). As a consequence, accumulated N2O production was reduced to 96 and 112 mg kg–1 for TZ and YX with straw addition (Fig. 1C and D).
Functional gene abundance
Abundances of bacterial 16S rRNA and denitrifying functional genes were measured using real-time PCR in samples collected at day 30 (Fig. 2; Fig. S1, Supporting Information). Compared to control treatment, no obvious changes in the absolute abundance of bacteria were observed in soils with straw amendment. However, the abundance of bacteria was higher in TZ (∼1.96 × 1010 copies g−1 dry soil) than in YX (∼6.66 × 109 copies g−1 dry soil). To facilitate revealing the changes of denitrifiers with straw amendments, the relative abundance of denitrifying functional genes to 16S rRNA were measured. For control soils, the relative abundance of nirK to 16S rRNA, was ∼3.0-fold higher in TZ (0.59 ± 0.22) compared to YX (0.18 ± 0.00), while the relative abundance of nirS (0.29 ± 0.08 vs. 0.41 ± 0.11) and nosZ (0.27 ± 0.11 vs. 0.26 ± 0.12) were similar between the two soils. Straw amendment significantly decreased the relative abundance of nirK, nirS and nosZ to 0.17 ± 0.08, 0.07 ± 0.04, and 0.00 ± 0.00 in TZ and to 0.11 ± 0.00, 0.05 ± 0.00 and 0.02 ± 0.00 in YX, respectively.
Relative abundance (relative to 16S rRNA) of three denitrification genes (nirK, nirS and nosZ) in two paddy soils from Taizhou (TZ, A) and Yixing (YX, B), China collected at day 30 following incubation under flooded conditions with (Straw) and without straw addition (Control). Stars indicate significant (P < 0.05) differences between treatments by Student's t-test.
Denitrifying bacterial community composition
Soil samples collected at day 30 were subjected to denitrifying bacterial community composition analysis using Illumina sequencing. There were no significant (P > 0.05) differences in sobs, simpson even, and shannon indexes of nirK-, nirS- and nosZ-type denitrifying bacterial communities between control and straw amended soils (Fig. S2, Supporting Information), suggesting that straw addition had no effects on richness, evenness and diversity of denitrifying bacterial communities.
For both soils without straw addition, the major community contributions to nirK-denitrifying bacteria at class level were Alphaproteobacteria (34.3%–36.6%), Unclassified Bacteria (29.6%–54.0%) and Unclassified Proteobacteria (7.66%–24.7%). The major community contributions to nirS-denitrifying bacteria were Unclassified Bacteria (30.9%–61.2%), Unnamed Bacteria (15.7%–47.8%) from environmental samples and Unclassified Proteobacteria (10.5%–18.9%), while the major community contributions to nosZ-denitrifying bacteria were Unclassified Bacteria (19.9%–42.8%), Alphaproteobacteria (15.5%–24.1%), Betaproteobacteria (9.33%–14.5%) and Unclassified Proteobacteria (11.3%–17.3%) (Fig. 3). At genus level, the predominant composition ofnirK-, nirS- and nosZ-denitrifying bacteria were shared by Unclassified Bacteria, followed by Unnamed Bacteria and Unclassified Proteobacteria (Fig. S3, Supporting Information). These taxonomically unassigned sequences were observed to have a close genetic relationship to Alphaproteobacteria, Betaproteobacteria or Gammaproteobacteria (Fig. S4, Supporting Information).
Class distribution of nirK-, nirS- and nosZ-type bacterial community composition in two paddy soils from Taizhou (TZ) and Yixing (YX) without (Control) and with wheat straw addition (Straw). ‘Other’ includes the low abundance bacteria (<0.1%) at class level. Each column represents the mean value of triplicate analyses.
Although large proportions of the nirK-, nirS- and nosZ-type denitrifying bacterial were not exactly classified or named, principle co-ordinates analysis (PCoA) at OTU level showed that nirK-, nirS- and nosZ-type denitrifying bacterial community composition was significantly (P < 0.05) separated from control treatments when straw was amended (Fig. 4), suggesting an influence on denitrifying bacterial community composition with straw addition. At genus level, straw addition significantly decreased the relative abundance of the major genus (unclassified bacteria, 54.0% vs. 38.5%) belonging to nirK-type denitrifying bacterial community for TZ and the major genus (unnamed bacteria, 47.7% vs. 32.2%) belonging to nirS-type denitrifying bacterial community for YX (Fig. S3, Supporting Information).
Principle coordinateanalysis (PCoA) of nirK- (A), nirS- (B), and nosZ-denitrifying bacterial community (C), composition intwo paddy soils from Taizhou (TZ) and Yixing (YX) without (Control) and with wheat straw addition (Straw).
Correlation between N2O emissions and denitrifying bacteria
Spearman correlation heatmaps between N2O emissions and denitrifying bacterial community at genus level showed significant positive relationships between N2O emissions and major taxas of nirK-bacteria and nosZ-bacteria, while relationships between N2O emissions and major taxas of nirS-bacteria were insignificant (Fig. S5, Supporting Information).
Soil properties contributing to altered denitrifying bacterial community composition
By envfit function (999 permutations) and RDA analyses, soil SOM and EOC were identified as variables that significantly contributed to variance in denitrifying bacterial community composition, regardless of nirK-, nirS- or nosZ-types. However, shifts in the composition of the three denitrifying communities were associated with EOC not the SOM arrow, implying that EOC rather than SOM or other soil properties was the main driver for altered denitrifying bacterial community composition (Fig. 5).
Correlation analysis of nirK- (A), nirS- (B), and nosZ-denitrifying bacterial community (C) composition in two paddy soils from Taizhou (TZ) and Yixing (YX) without (Control) and with wheat straw addition (Straw). pH: soil pH value; EOC: easily oxidizable carbon; TOC: total organic carbon; SOM: soil organic matter; POC: particulate organic carbon; DOC: dissolved organic carbon; EC: electrical conductivity; C/N: carbon to nitrogen ratio.
DISCUSSION
N2O emissions from soil
In this study, we observed that N2O emissions from TZ were much delayed compared to YX in control treatments (i.e. no straw amendment) (Fig. 1). In paddy soil, reductive conditions are favorable for N2O production from anaerobic nitrate reduction (Shan et al. 2018). Compared to YX (clayed soil), TZ was a silty soil with higher porosity, facilitating penetration of oxygen below the soil–water surface. To form required reductive conditions for N2O production, a prolonged flooding period was required for TZ, thereby leading to delayed N2O production compared to YX.
Although delayed, accumulated N2O emissions from silty TZ were significantly higher than clayed YX (Fig. 1). This variability was related to differences in denitrifying bacterial community composition and abundance of functional genes. Compared to YX, the relative abundance of unclassified bacteria in nirK-type (54.0% vs. 29.6%) and nirS-denitrifying bacteria (61.2% vs. 30.9%) at class level was higher in TZ (Fig. 3). At genus level, the abundance of unclassified bacteria, the predominant nirK- and nirS-bearing bacteria, was ∼2-fold higher in TZ compared to YX (Fig. 3), thereby leading to greater N2O production and emission. Analysis of functional genes further showed that the abundance of nirK was ∼3-fold higher in TZ compared to YX, while the abundance of nirS and nosZ were similar between the two soils (Fig. 2), confirming that higher N2O emissions from TZ was mainly due to its higher abundance of nirK-type denitrifying bacteria.
The reason for the varied denitrifying bacteria community composition and functional gene abundance between the two soils can be further explained by differences in soil properties. Soil properties have been identified as variables that significantly contribute to variance in bacterial community composition (Deklein and Vanlogtestijn 1994; Blackmer and Bremner 1978; Chu et al. 2007; Baggs, Smales and Bateman 2010; Cuhel et al. 2010). In this study, in control treatments, we observed that nirK-, nirS-, nosZ-type denitrifying bacterial community composition was significantly separated between TZ and JX, shifting from TZ to YX mainly along the SOM arrow (Fig. 5). This indicated that among the tested soil properties (e.g. pH, EC, C/N, SOM, TOC, DOC, POC and EOC), SOM was the most important driver of variation in denitrifying bacterial community composition between soils without straw amendment. Compared to TZ, YX contained significantly higher SOM (5.19% vs. 2.22%) (Table 1), suggesting that higher inherent SOM content in YX inhibited N2O emissions. This was similar to the role of straw amendment into both TZ and YX in suppressing N2O emissions (Fig. 1), suggesting both high inherent and exotic inputs of organic matter could decrease N2O emissions from paddy soils.
Reduced nirK and nirS rather than increased nosZ contributed to decreased N2O emissions with straw amendment
In this study, straw amendment significantly suppressed N2O emissions (Fig. 1), similar to previous reports (Ma et al. 2009). However, studies on the effects of straw on N2O emissions are inconsistent, with increased N2O emissions also being reported (Millar and Baggs 2004; Zhu et al. 2015). The different effects on N2O emissions were likely related to different changes in microbial community composition with straw amendment. Since N2O can be the final or intermediate product of denitrification, straw addition could decrease the relative abundance of nirK or nirS-type denitrifying bacteria to suppress N2O production or increase the relative abundance of nosZ-type bacteria to promote N2O consumption (Mmm, Marchant and Kartal 2018). In this study, we observed that straw addition significantly altered nirK and nirS-bacterial community composition and decreased the relative abundance of major genera belong to the nirK- and nirS-type denitrifying bacterial community (Fig. S2, Supporting Information) that underpin N2O production, thus reducing N2O emissions. Positive correlation between N2O and major taxas of nirK-bacterial community composition further supports the hypothesis. In addition, we observed 1.63–3.46 and 4.05–8.37-fold decreases in the relative abundance of nirK and nirS genes for TZ and YX with straw addition (Fig. 2). The above suggests that the decreased abundance of nirK- and nirS-type denitrifying bacteria was the main reason for reduced N2O emissions with straw amendment.
In addition to suppressing N2O production via decreasing nirK- and nirS-bacteria, enhanced N2O consumption with straw addition maybe another factor in reducing N2O emissions. Straw addition could accelerate microbial C-turnover in soil resulting in reducing conditions, which could promote N2O reduction to N2 (Chen et al. 2017). Previous studies have shown that rice straw incorporation altered community composition and increased the abundance of nosZ-type denitrifying bacteria, which promoted N2O reduction (Chen et al. 2012). However, in this study, we observed that straw addition significantly decreased the relative abundance of nosZ gene (Fig. 2), suggesting that changes in nosZ-type denitrifying bacteria was not be the main factors contributing to reduced N2O emissions. These results implied that wheat straw return reduced N2O emissions from flooded paddy soils via the reduced abundance of nitrite reductase nirK and nirS-bearing communities.
EOC rather than TOC drives variation in denitrifying community composition with straw amendment
Soil characteristics have been reported to be an important factor in shaping bacterial community composition (Lüdemann, Arth and Liesack 2000; Högberg, Högberg and Myrold 2007; Shen et al. 2013). In this study, changes in the composition of three denitrifying soil communities with straw addition were influenced by different soil properties, particularly SOM and EOC (Fig. 5). SOM may be oxidized under flooded condition by microbes releasing SOC which may serve as an available substrate for denitrifying bacterial growth (Gardenas et al. 2011). However, previous studies have predominantly focused on the influence of TOC with a lack of analysis into the role of different carbon fractions in influencing bacterial community profiles (Hu et al. 2014).
In this study, EOC was the shared carbon fraction influencing the composition of each type of denitrifying community (Fig. 5). Although TOC or DOC and POC were important factors influencing the compositions of nirK- and nirS-containing denitrifying bacterial communities (Fig. 5A and B), changes in the composition of three denitrifying communities with straw addition were mainly along the EOC arrows. This implies that EOC rather TOC, DOC or POC was the main driving factor for variations in denitrifying bacterial community compositions in soil with straw addition (Fig. 5). Among four SOC fractions, DOC content in soil was low, accounting for < 0.1% of TOC in TZ and YX (Table 1). In anaerobic flooded paddy soil, higher concentrations of EOC are more readily available to microbes for oxidation compared to TOC and POC, thereby playing a significant role in altering bacterial community profiles.
In conclusion, based on laboratory experiments, we showed that wheat straw amendments to flooded paddy soil significantly reduced the abundance of major taxas of nirK- and nirS-denitrifying bacterial communities and decreased the abundance of nirK, nirS and nosZ functional genes, thereby significantly inhibiting N2O emissions. The decreased abundance of nitrite reductase-bearing bacteria was identified as the main reason accounting for reduced N2O emissions with straw amendment. Changes in denitrifying bacterial community composition were attributed to altered soil characteristics when straw was added, especially SOC. Among different carbon fractions, denitrifying bacterial community composition varied mainly along EOC rather than TOC, DOC and POC in soil amended with straw, highlighting the importance of EOC in controlling denitrifying bacterial communities.
ACKNOWLEDGEMENTS
This study was supported by the National Natural Science Foundation of China (41601261).
Conflict of interest None declared.
