Impact of 25 years of inorganic fertilization on diazotrophic abundance and community structure in an acidic soil in southern China

Soil Biology & Biochemistry 113 (2017) 240e249

Contents lists available at ScienceDirect

Soil Biology & Biochemistry journal homepage: www.elsevier.com/locate/soilbio

Impact of 25 years of inorganic fertilization on diazotrophic abundance and community structure in an acidic soil in southern China Chao Wang a, Manman Zheng a, c, Wenfeng Song a, Shilin Wen b, Boren Wang b, Chunquan Zhu a, c, Renfang Shen a, c, * a b c

State Key Laboratory of Soil and Sustainable Agriculture, Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, China Qiyang Red Soil Experimental Station, Chinese Academy of Agricultural Sciences, Qiyang 426182, China University of Chinese Academy of Sciences, Beijing 100049, China

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 March 2017 Received in revised form 8 June 2017 Accepted 17 June 2017

Although diazotrophs are important in the nitrogen (N)-cycle and contribute to the pool of plant available N, the population response to long-term inorganic fertilization is largely unknown. Here, we investigated the diazotrophic populations in both the bulk and rhizosphere soils of maize grown in an acidic farmland soil that experienced 25 years of inorganic fertilization. The fertilization regimes included unfertilized control, N fertilizer alone, N fertilizer with quicklime, phosphorus (P) and potassium (K) fertilizers, N þ P þ K fertilizers, and N þ P þ K fertilizers with quicklime. Quantitative PCR and high-throughput pyrosequencing of the nifH gene were used to analyze diazotrophic abundance and community composition. All of the fertilizer treatments improved soil nutrient availability, but those without quicklime caused soil acidification. Maize biomasses and nifH copy numbers were significantly lower under N and N þ P þ K treatments but increased under P þ K fertilization. Quicklime applications effectively alleviated the inhibitory effect of N input. Fertilization led to decreases in operational taxonomic unit richness and shifts in diazotrophic community composition. Soil pH and nutrient availability had a cooperative effect on diazotrophic abundance, while soil nutrient availability appeared to be the main factor shaping diazotrophic community structure. Rhizosphere effects increased the nifH gene copy number but did not obviously change the diazotrophic community composition on the current research scale. Overall, the long-term inorganic fertilization affected both diazotrophic abundance and community composition, and the fertilizer treatment had a greater influence than quicklime remediation or crop cultivation on community composition. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Crop cultivation Diazotroph Long-term inorganic fertilization Nutrient availability nifH Quicklime application

1. Introduction To meet the food demand of an ever-increasing world population, the resulting intensive agriculture practices have included applying an overload of inorganic fertilizers (Liu et al., 2015; Zeng et al., 2016). In China, nitrogen (N), phosphorus (P), and potassium (K)-based fertilizer consumption increased to 38.94 million tonnes in 2013 from 0.73 million tonnes in 1961 (http://faostat3. fao.org/home) and is expected to continue to increase over the

* Corresponding author. State Key Laboratory of Soil and Sustainable Agriculture, Institute of Soil Science, Chinese Academy of Sciences, Nanjing 210008, China. E-mail address: [email protected] (R. Shen). http://dx.doi.org/10.1016/j.soilbio.2017.06.019 0038-0717/© 2017 Elsevier Ltd. All rights reserved.

next few decades (Liu et al., 2015). The increased long-term input of inorganic fertilizers, especially N-based fertilizers, in agricultural ecosystem improved soil fertility and crop yields over the past decades (Liu et al., 2015; Zeng et al., 2016) but also had various negative effects such as soil acidification, metal toxicity, lower nutrient use efficiencies, increased greenhouse gas emissions, and groundwater contamination that threaten soil quality, crop growth, biodiversity, and environmental health (Guo et al., 2010; Zhong et al., 2015) and changed the biogeochemical cycles of soil nutrient elements (i.e., carbon (C), N, and P) (Geisseler and Scow, 2014). Thus, there is an increasing concern on how to safely enhance agricultural sustainability. The N cycle is an important nutrient cycle that influences the productivity and sustainability of the terrestrial ecosystem

C. Wang et al. / Soil Biology & Biochemistry 113 (2017) 240e249

(Geisseler and Scow, 2014; Levy-Booth et al., 2014). Microbes are key players in the soil N cycle, play an important role in regulating soil N availability and transformation, and are closely related to plant productivity. The intensive fertilization has resulted in an unprecedented disturbance in the soil N cycle (Geisseler and Scow, 2014), which involved shifts in a diverse range of the microbial communities and functions (Sun et al., 2015; Zhou et al., 2015). To date, most of the studies have focused on the influences of longterm fertilization on the ammonia oxidizers (He et al., 2007; Wang et al., 2015; Zhong et al., 2015) and denitrifiers (Yin et al., 2015) because they are directly associated with soil acidification, N availability to plants, and N loss (Levy-Booth et al., 2014). The Nfixing bacteria (diazotrophs) are responsible for biological N fixation, another functionally important biogeochemical N cycle (LevyBooth et al., 2014). Because this bioprocess can supply additional N sources to the ecosystem to improve soil fertility, which is beneficial to plant productivity, it is a potentially sustainable alternative to inorganic N fertilizer use in agricultural systems (Gupta et al., 2006). However, the response patterns of diazotrophic bacteria to long-term fertilization in agricultural ecosystems are not well understood, although some short-term micro-zone experiments (within one complete growing season) have been carried out (Rodríguez-Blanco et al., 2015; Simonsen et al., 2015). However, short-term experiments cannot supply some valuable information that is detected only under long-term fertilization conditions, because changes in soil quality are slow and microbial community structures need time to stabilize (Chakraborty et al., 2011; Reardon et al., 2014). Detailed information on diazotrophic community composition and abundance, as well as their relationships to the soil environment, would improve our understanding of the longterm ecological effects of fertilization and of the biological functions of diazotrophic bacteria in the agroecosystem, as well as reduce inorganic N fertilizer use. Fertilization practices strongly influence soil microbial community structure and abundance either directly by altering soil characteristics or indirectly by changing the plant feedback (Geisseler and Scow, 2014; Zeng et al., 2016). In the long-term fertilization regimes, changes in soil physicochemical properties, such as pH, nutrient availability, and carbon quantity and quality, have had positive or negative influences on soil microbial populations (Geisseler and Scow, 2014; Levy-Booth et al., 2014; Li et al., 2014). Likewise, diazotrophic bacteria are very sensitive to soil properties (Pereira e Silva et al., 2013; Rodríguez-Blanco et al., 2015). Additionally, plants, as another important driving factor, have a close relationship to the functions and activities of the rhizospheric microbiome (Berendsen et al., 2012). The growth of crops, such as maize, rice, soybean, and sorghum, leads to a unique diazotrophic communities in the rhizosphere, which are different from the communities in the bulk soil (Coelho et al., 2008; Li et al., 2012; Rodríguez-Blanco et al., 2015; Wang et al., 2012). Conversely, rhizosphere microbes directly impact plant growth through positive or negative traits (Berendsen et al., 2012). Thus, variations in plantemicrobe interactions in the rhizosphere are of great significance in agronomy. Previous studies on the impacts of the longterm fertilization practices on the soil microbial community were confined to the bulk soil scale (Geisseler and Scow, 2014; LevyBooth et al., 2014; Li et al., 2014), which resulted in a lack of a deep understanding of how the plant-associated rhizosphere's microbiome responds to fertilization (Ai et al., 2015). Soil acidification resulting from inorganic N input has become a major limiting factor for crop yields in the red soil region of southern China (Guo et al., 2010). In agricultural management, lime is applied as an effective way to remediate the acidic soils, maintain crop productivity and alter the microbial population (Xun et al., 2016a, 2016b). Xun et al. (2016b) reported that the addition of

241

lime can quickly increase the soil pH and bacterial diversity in acidic farmland soil. Likewise, our previous studies found that a lime application to an acidic soil significantly improved the abundances of overall bacteria (Wang et al., 2013) and ammonia oxidizers (Che et al., 2015). However, such information is not available for diazotrophic bacteria. For this study, samples were collected from a long-term fertilized area in southern China that was established in 1990. Some severe acidification regimes were employed in 2010 and part of the area was remediated by adding quicklime (Xun et al., 2016a, 2016b). Previous studies at this site have focused on the influences of longterm fertilization on crop production (Zhang et al., 2009), soil acidification (Cai et al., 2015), remediation approaches to soil acidification (Xun et al., 2016b), soil enzyme activities (Ahamadou et al., 2009), overall bacterial abundance (Ahamadou et al., 2009; Xun et al., 2016a, 2016b), and ammonia oxidizer abundance (He et al., 2007). However, the effects on diazotrophic bacteria are still unknown. Thus, we chose this experimental site, which has undergone 25 years of fertilization, to investigate the shift of soil diazotrophic populations under the different fertilization treatments. The objectives were to: (1) identify the responses of the diazotrophic abundance and community composition to long-term inorganic fertilization, as well as their relationships with soil properties; (2) assess the impact of long-term inorganic fertilization on the rhizosphere effect; and (3) test whether lime remediation can also improve soil diazotrophic populations in the acidic soil. To do so, we chose the dinitrogenase reductase subunit gene nifH because it is the most used marker gene for analyzing the abundance and diversity of diazotrophic bacteria (Levy-Booth et al., 2014). Both quantitative-PCR and high-throughput sequencing were applied to determine the abundance and community structure of the nifH gene. 2. Materials and methods 2.1. Experimental site and sampling The fertilization experiment started in 1990 at a site located in the Red Soil Experimental Station at Qiyang, Hunan Province, China (26 450 N, 11153’ E). The soil on site originated from the Quaternary red clay soil and is classified as Ferralic Cambisol. The size of each fertilizer treatment was 20 m  10 m and designed with two replicate plots (Cai et al., 2015; Xun et al., 2016a,b). The cropping system was an annual rotation of summer maize (Zea mays L. Yedan 13) and winter wheat (Triticum aestivum L. Xiangmai 4). The tillage operation was conventional tillage. The inorganic fertilizers were applied as urea (300 kg N ha1 year1), superphosphate (53 kg P ha1 year1) and potassium chloride (100 kg K ha1 year1). The original soil properties, fertilization methods, natural environment, and field management have been described previously (Cai et al., 2015). Before sowing, fertilizers were applied by banding at a depth of 10 cm. For annual input, 30% of fertilizers were applied for wheat and 70% for maize. As the plots receiving N and N þ P þ K (NPK) fertilizers showed severe soil acidification, these two plots were each divided into two parts in 2010. One part maintained the same fertilization as before, while the other also received 2550 kg ha1 of quicklime based on the same fertilization protocol, followed by the addition of 1500 kg ha1 of quicklime in 2014. In this trial, six different fertilizer treatments were chosen: unfertilized control (CK), inorganic N fertilizer alone (N), inorganic N fertilizer plus quicklime (NCa), inorganic P þ K fertilizers (PK), inorganic NPK fertilizers, and inorganic NPK fertilizers plus quicklime (NPKCa). Soil samples were collected on June 12, 2015 (at the maize flowering stage). Each fertilizer treatment was established

242

C. Wang et al. / Soil Biology & Biochemistry 113 (2017) 240e249

randomly in three 3 m  3 m subplots. Due to two replicate plots in each treatment, two subplots were randomly selected from one replicate plot, and another was from the other replicate. The rhizosphere soils (0e15 cm depth) from five maize plants in each subplot were collected, and these five samples were mixed into a single soil sample. The roots of each plant were removed gently from the farmland using a spade, and the soil attached to the roots was pooled as the rhizosphere soil. Bulk soil was obtained from the area that was not disturbed by the roots and did not contain the soil that fell off from the roots. Each bulk soil sample was also collected from five soil cores (0e15 cm depth and 2 cm diameter) in each subplot. Each collected sample was sieved through a 2-mm mesh, homogenized thoroughly, and divided into two parts. One part was used for the soil property analysis, and the other, for DNA extraction was stored immediately at 80  C. Shoots and roots of each plant were harvested simultaneously and were washed three times with distilled water to thoroughly remove the adsorbed soil particles. Then, they were separately placed into envelopes and oven dried at 70  C to obtain a plant dry weight. 2.2. Soil property analysis The air-dried soils were analyzed for soil pH, total carbon (TC), total nitrogen (TN), soil organic carbon (SOC), available P (AP), and available K (AK). Fresh soil samples were used to determine soil  ammonium (NHþ 4 -N), nitrate (NO3 -N), dissolved organic carbon (DOC), and organic nitrogen (DON). Soil pH was determined in deionized water (soil:water, 1:2.5) using a pH meter (Mettler Toledo FE20, Shanghai, China). SOC was measured using the sulfuric acidepotassium dichromate oxidation method (Sims and Haby, 1971). TC and TN were measured with a Vario MAX CNS elemental analyzer (Elementar, Hanau, Germany). AP was extracted with 0.03 M hydrochloric acideammonium fluoride and determined according to the Bray method (Bray and Kurtz, 1945). AK was extracted with 1.0 M ammonium acetate and determined by flame  photometry (FP640; Shanghai, China). Both NHþ 4 -N and NO3 -N were extracted with 2.0 M KCl and measured by a continuous flow analyzer (Sanþþ, Skalar, Holland). DOC and dissolved total N (DTN) were extracted by 0.5 M K2SO4 and determined using a total organic carbon analyzer (Multi N/C 3000; Analytik, Jena, Germany) and the Kjeldahl method (Watkins et al., 1987), respectively. DON was calculated as follows: DTN  NHþ 4 -N. The soil water content was determined gravimetrically. 2.3. Soil DNA extraction and quantification of nifH gene abundance Soil DNA was extracted from 0.5 g of soil (fresh weight) using a Fast®DNA SPIN Kit (MP Biomedicals, CA, USA) and was purified subsequently using a PowerClean®DNA Clean-up Kit (MoBio, CA, USA), according to the manufacturer's instructions. Each soil sample contained three successive DNA extractions, and the triplicates were pooled as a DNA sample. The concentration and quality of the extracted DNA were measured using a NanoDrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, USA). The nifH gene copy number was determined by real-time quantitative PCR (qPCR) on a LightCycler 480 real-time PCR system (Roche Diagnostics, Mannheim, Germany) with SYBR®Premix Ex Taq™ (TaKaRa Bio, Dalian, China). The nifH gene-specific primers PolF/PolR (Poly et al., 2001) were used for qPCR amplification. The qPCR program and reaction composition were described by Sun et al. (2015). Melting curves were used to analyze the specificity of the amplified products. To obtain the standard curve, a nifH gene fragment was cloned into the pMD19-T vector (TaKaRa Bio) and transferred subsequently into Escherichia coli DH5a competent

cells. The plasmids containing the correct fragment length were selected and verified. The extracted plasmid DNA using Plasmid Purification Kit (TaKaRa Bio) was used as the template to generate a standard curve. Each DNA sample was run in triplicate. Copy numbers and PCR amplification efficiencies were calculated according to the study of Mirza et al. (2014). The efficiency was 89.1% with an R2 value of 0.99. 2.4. nifH gene sequencing and bioinformatics analysis A nested PCR approach was used to amplify the nifH gene fragments for pyrosequencing. The primer sets PolF/PolR and RoeschF/RoeschR (Roesch et al., 2006) were used in the first and second PCR amplification, respectively. The PCR program and reaction composition were described by Pereira e Silva et al. (2013). In order to distinguish the sample, a sample-specific tag (7-bp barcode) was added to the forward primer in the second PCR amplification. Triplicate PCR amplifications for each sample were conducted and pooled as a PCR product. Then, the purified amplicons using Agarose Gel DNA purification kit (TaKaRa Bio) were quantified and combined in equimolar ratios. The sequencing of the amplicon libraries was carried out on an Illumina MiSeq platform with 300-bp paired-end reads. The sequencing data has been deposited in the NCBI Sequence Read Archive (SRA) database under accession number SRP092284. Pairs of reads from the raw data were first merged with FLASH version 1.2.7 (Magoc and Salzberg, 2011), in which forward and reverse reads had the overlapping base length  10 bp and did not allow the base mismatch. Sequencing reads were processed with Mothur version 1.31.1 (Schloss et al., 2009). The low-quality sequences that had a quality score <20, contained ambiguous nucleotides, or did not match the primer and barcode, were removed. After the sequences of samples were sorted according to the barcodes, the barcode and primer sequences were deleted. The remaining sequences were further converted to amino acid sequences using the FunGene Pipeline of the Ribosomal Database Project server (http://fungene.cme.msu.edu/FunGenePipeline/) (Mirza et al., 2014; Pereira e Silva et al., 2013). Sequences whose translated proteins did not match the nifH protein sequence or that contained termination codons were discarded. The remaining sequences were aligned against the nifH gene database (Gaby and Buckley, 2014), and the failed and chimeric sequences were also removed. The remaining high-quality sequences were grouped into operational taxonomic units (OTUs) at 90% identity following the studies of Huang et al. (2016) and Pereira e Silva et al. (2013). Representative sequences in the OTUs were taxonomically classified by the BLAST algorithm-based search within GenBank (http:// blast.ncbi.nlm.nih.gov/Blast.cgi). 2.5. Data analysis The statistical analyses were performed using SPSS version 18.0 for Windows (SPSS Inc., Chicago, IL, USA). Differences in maize biomasses, and OTU numbers and relative abundances, were determined among the fertilizer treatments using a one-way analysis of variance (ANOVA), and Duncan's test was performed for multiple comparisons (p < 0.05). A two-way ANOVA was employed to evaluate the interaction effects of fertilizer treatments and root effect (bulk versus rhizosphere) on soil property parameters and total gene abundance. When the interaction factor reached a significant difference (p < 0.05) among treatments, a oneway ANOVA or Student's t-test was employed to analyze the degree of difference. Pearson's correlation coefficients were used to test relationships among soil properties, total nifH gene copy numbers, and OTU numbers and relative abundances.

C. Wang et al. / Soil Biology & Biochemistry 113 (2017) 240e249

Fig. 1. Maize biomass from the 25-year (1990e2015) long-term fertilizer experiment. Data are means ± standard deviation (n ¼ 15). Different letters on the bars indicate significant differences among fertilizer treatments at p < 0.05.

To calculate community similarities, an OTU-based hierarchical cluster analysis with the unweighted pair group method of arithmetic means, a non-metric multidimensional scaling (NMDS) based on Bray-Curtis distance matrices, as well as an analysis of similarity (ANOSIM), were carried out using the vegan package of the R software (Version 3.1.2). A Mantel test was employed to determine the correlation between the community composition and each soil variable. Only variables that had significant effects (p < 0.05) by the Mantel test were further used to perform a redundancy analysis using R software. 3. Results 3.1. Maize biomass and soil properties Compared with the non-fertilized CK, the long-term N treatment significantly reduced the maize biomass (p < 0.05), while biomass increased under the PK treatment (Fig. 1). NPK and CK

243

produced comparable maize biomasses. The application of quicklime (NCa and NPKCa) markedly increased the biomass, and the maize biomass produced under NPKCa was greater than that produced under NCa (p < 0.05) (Fig. 1). Soil properties are shown in Table 1. A two-way ANOVA showed that the fertilizer treatments significantly affected all soil property parameters (p < 0.01), and root effect (bulk versus rhizosphere soils) also influenced these parameters (p < 0.01 or p < 0.05), except for pH, TC, and TN (Table S1). Compared with the bulk soil, the rhizosphere soil showed higher contents of SOC, DOC, and DON,  and lower contents of NHþ 4 -N, NO3 -N, AK, and AP. In the bulk soil, the pH values displayed no significant differences between CK (5.38), NCa (5.20), and NPKCa (5.12), which were higher than those of N (4.04), PK (4.95), and NPK (4.00) (p < 0.05), with the lowest occurring under both N and NPK treatments (Table 1). The contents of TC, SOC, and DOC under PK, NPK, and NPKCa exceeded those  under CK, N, and NCa. The contents of NHþ 4 -N, NO3 -N, and TN under the treatments containing N fertilizer (N, NCa, NPK, and NPKCa) were greater compared with the treatments without N fertilizer  (CK and PK), except for NHþ 4 -N and NO3 -N under NPKCa treatment. Similarly, the contents of both AK and AP were increased under treatments that supplied both P and K fertilizers. The C/N ratio in the N treatment was the lowest but was not significantly different from that of the NCa treatment. DON did not show a statistical difference under the CK, N, PK, and NPK treatments. 3.2. nifH gene copy number and its correlation with soil properties The two-way ANOVA analysis showed that both fertilizer treatment and root effect significantly influenced the nifH gene copy number (p < 0.01) (Fig. 2). Further, Student's t-test indicated that the nifH gene copy number in the rhizosphere was obviously higher than in the bulk soil under each treatment (p < 0.01 or p < 0.05). In the bulk soil, the average copy numbers of the six treatments ranged from 0.58  105 to 3.85  105 gene copies per g dry soil (Fig. 2). Compared with CK, PK had a higher nifH gene copy number, whereas those of N and NPK were significantly lower (p < 0.05). These inhibitory effects of N-based fertilizers were

Table 1 Soil physicochemical properties in the bulk and rhizosphere soils of maize cultured for the 25-year (1990e2015) long-term fertilizer experiment. Soil properties

Sampled site

CK

N

NCa

PK

NPK

NPKCa

pH

Bulk Rhizosphere Bulk soil Rhizosphere Bulk Rhizosphere Bulk Rhizosphere Bulk Rhizosphere Bulk Rhizosphere Bulk Rhizosphere Bulk Rhizosphere Bulk Rhizosphere Bulk Rhizosphere Bulk Rhizosphere

5.38 ± 0.11a 5.43 ± 0.04a 8.02 ± 0.32c 8.32 ± 0.40c 1.08 ± 0.06c 1.14 ± 0.06c 7.40 ± 0.14ab 7.30 ± 0.12b 5.52 ± 0.28b 5.97 ± 0.20c 0.19 ± 0.02b 0.21 ± 0.01b 1.08 ± 0.32c 0.45 ± 0.49c 0.85 ± 0.32c 0.66 ± 0.06c 7.12 ± 2.22ab 10.16 ± 2.66c 52.80 ± 3.77c 54.30 ± 7.40d 3.04 ± 0.35b 2.99 ± 0.24b

4.04 ± 0.33c 3.85 ± 0.14c 7.21 ± 1.12c 8.63 ± 0.78c 1.36 ± 0.50b 1.41 ± 0.03b 5.63 ± 1.42c 6.12 ± 0.53c 5.35 ± 0.74b* 6.77 ± 0.33b 0.17 ± 0.02c 0.21 ± 0.02b 371.05 ± 77.02a* 160.44 ± 42.41a 25.62 ± 11.60b 18.75 ± 4.53a 5.00 ± 2.28b 7.64 ± 1.85c 57.62 ± 10.48c 41.94 ± 4.46d 3.80 ± 0.61b 3.77 ± 0.59b

5.20 ± 0.27ab 5.28 ± 0.38a 8.27 ± 0.35c 8.78 ± 0.72c 1.39 ± 0.19b 1.20 ± 0.09c 6.06 ± 1.03bc 7.32 ± 0.17b 5.97 ± 0.09b* 6.94 ± 0.18b 0.16 ± 0.01c* 0.22 ± 0.02b 28.00 ± 12.36b* 9.50 ± 0.38b 45.24 ± 21.65a* 14.64 ± 1.91a 8.33 ± 2.37ab 12.83 ± 4.45bc 62.75 ± 14.74c 61.24 ± 6.91d 4.36 ± 0.27b 4.19 ± 0.16b

4.95 ± 0.04b 5.14 ± 0.08a 9.62 ± 0.34b 9.59 ± 1.06bc 1.27 ± 0.06c 1.24 ± 0.11c 7.61 ± 0.51a 7.74 ± 0.20ab 6.77 ± 0.49a 6.98 ± 0.05b 0.20 ± 0.01b 0.22 ± 0.01b 1.57 ± 0.35c* 0.70 ± 0.21c 1.54 ± 0.57c 0.94 ± 0.25c 7.56 ± 3.60ab 11.00 ± 1.90c 598.87 ± 92.24a 520.17 ± 26.08a 280.76 ± 29.21a* 184.21 ± 19.20a

4.00 ± 0.08c 4.21 ± 0.12b 11.77 ± 0.93a 12.13 ± 1.17a 1.64 ± 0.12a 1.58 ± 0.16a 7.20 ± 0.46ab 7.70 ± 0.07ab 7.22 ± 0.30a* 8.15 ± 0.76a 0.26 ± 0.02a 0.29 ± 0.01a 52.16 ± 12.06b* 3.09 ± 1.83c 19.30 ± 3.18b* 7.15 ± 3.14b 9.51 ± 4.13ab* 18.02 ± 0.32ab 535.85 ± 61.26a* 249.09 ± 15.39b 285.73 ± 45.70a* 191.49 ± 42.66a

5.12 ± 0.21ab 5.24 ± 0.20a 10.80 ± 0.80ab 10.79 ± 0.84ab 1.43 ± 0.10ab 1.37 ± 0.06b 7.56 ± 0.40ab 7.88 ± 0.27a 7.23 ± 0.14a* 8.44 ± 0.21a 0.25 ± 0.00a* 0.29 ± 0.02a 2.44 ± 1.75c 1.08 ± 0.54c 3.27 ± 1.43c 2.07 ± 0.83c 12.32 ± 2.10a* 22.50 ± 4.34a 313.32 ± 66.25b* 192.10 ± 19.45c 297.59 ± 87.16a 174.855 ± 24.28a

TC (g kg1) TN (g kg1) C/N ratio SOC (g kg1) DOC (g kg1) 1 NHþ ) 4 -N (mg kg 1 NO ) 3 -N (mg kg

DON (mg kg1) AK (mg kg1) AP (mg kg1)

Data are means ± standard deviation (n ¼ 3). Different letters in each row indicate significant differences among fertilizer treatments for either rhizosphere or bulk soil at p < 0.05. The asterisks in bold for each treatment indicate significant differences between bulk and rhizosphere soil at p < 0.05. TC: total carbon; TN: total nitrogen; C/N ratio:  total carbon/total nitrogen; SOC: soil organic carbon; DOC: dissolved organic carbon; NHþ 4 -N: ammonium nitrogen; NO3 -N: Nitrate nitrogen; DON: dissolved organic nitrogen; AK: available potassium; AP: available phosphorus.

244

C. Wang et al. / Soil Biology & Biochemistry 113 (2017) 240e249

using an ANOSIM (Table S3), which showed significant differences between N-free fertilizer treatments (CK and PK) and N-containing fertilizer treatments (N, NCa, NPK, and NPKCa) (p < 0.01), but no differences between bulk and rhizosphere samples. Thus, the influence of the fertilizer treatment on the community composition and structure was far more obvious than that of root effect and quicklime application. Next analyses focus on bulk soil samples due to the lesser effect of rhizosphere under different fertilizations. 3.4. OTU richness and relative abundance of diazotrophic taxa in bulk soil

Fig. 2. nifH gene copy number as quantified by real-time PCR in the bulk and rhizosphere soils of maize cultured during the 25-year (1990e2015) long-term fertilizer experiment. All of the data were analyzed using a two-way analysis of variance, T: fertilizer treatments; R: root effect (bulk and rhizosphere). Different lowercase letters above the white columns indicate significant differences among fertilizer treatments for bulk soils at p < 0.05; different capital letters above the gray columns indicate significant differences among fertilizer treatments for rhizosphere soils at p < 0.05. The asterisks in each fertilizer treatment indicate significant differences between bulk and rhizosphere soils at **p < 0.01 and *p < 0.05.

markedly alleviated by quicklime applications in NCa and NPKCa, and even the nifH gene copy number in NPKCa was significantly higher than in CK (p < 0.05). The similar results were obtained in the rhizosphere samples, except that there was no significant difference between CK and NCa (Fig. 2). Pearson's correlation analysis showed that in the bulk soil, the nifH gene copy number was positively correlated with soil pH, C/N, and AP and negatively  correlated with NHþ 4 -N and NO3 -N. The nifH gene copy number in the rhizosphere soil was positively correlated with pH, C/N, and AK  and negatively correlated with NHþ 4 -N and NO3 -N (Table 2). For all of the soil samples, including bulk and rhizosphere soils, the nifH gene copy number was positively correlated with pH, C/N, SOC,  DOC, and DON and negatively correlated with NHþ 4 -N and NO3 -N. 3.3. Community structure analyses After quality filtering and screening of amino acid sequences, 107,039 high-quality sequences with 870 to 5258 sequences per sample were obtained, and the numbers of OTUs (90% similarity) ranged from 42 to 109, depending on the sample (Table S2). To compare the similarities and differences among the community compositions, we performed an NMDS, hierarchical cluster analysis and ANOSIM, based on the OTU composition. The stress value for the NMDS was 0.09 (Fig. 3a). This analysis showed three different groups, as follows: bulk and rhizosphere samples of CK and PK; bulk and rhizosphere samples of N and NCa; and all of the samples of NPK and NPKCa. The hierarchical cluster analysis obtained a similar result (Fig. 3b), with the communities forming three clusters, all of the NPK and NPKCa samples, the treatments containing only N fertilizers (N and NCa), and the CK and PK samples. The dissimilarities between community compositions were analyzed

In bulk soil, all of the fertilizer treatments significantly reduced the number of OTUs (p < 0.05) (Fig. 4). There were no significant differences among N, NCa, NPK, and NPKCa (Fig. 4), suggesting that quicklime applications did not relieve the inhibitory effects of N fertilizer. A further correlation analysis showed that the number of OTUs was significantly negatively correlated with TN and NO 3 -N (p < 0.05) (Table S4). Using the BLAST algorithm in NCBI, OTUs were taxonomically classified into different genera across all of the treatments. Fig. 5 shows the nine most abundant genera that had relative abundances of more than 1% in the bulk soils of the six fertilizer treatments. The genus Bradyrhizobium was the most abundant, containing 47.7%e69.9% of the total nifH gene sequences in all of the bulk soil samples. Different fertilizer treatments changed the relative abundances of these main genera (Fig. 5). Compared with CK, both N and NPK significantly increased the relative abundance of the genus Methylosinus (p < 0.05), while reducing the relative abundances of the genera Bradyrhizobium, Rhodopseudomonas, Xanthobacter, and Caenispirillum. The addition of quicklime (NCa and NPKCa) relieved the decrease in the relative abundance of the genus Bradyrhizobium but had no significant effect on the other four genera. The PK treatment significantly decreased the relative abundances of the genera Bradyrhizobium and Azovibrio (p < 0.05), but increased the relative abundances of the genera Rhodopseudomonas, Caenispirillum, and Phaeospirillum. In addition, the genera Phaeospirillum and Azospirillum were the most abundant in NPK, and the addition of quicklime (NPKCa) reduced their relative abundances. 3.5. Correlation between diazotrophic community structure and soil variables The Mantel test revealed that the diazotrophic community structures in the bulk soils were closely correlated with multiple soil variables (p < 0.01 or p < 0.05), and the correlation coefficient followed the trend: C/N > NHþ > AK > NO 4 -N 3N > AP > TC > SOC > DOC > pH (Table S5). The effects of soil properties on the structures of diazotrophic communities were further analyzed using the redundancy analysis, based on the selected soil variables and OTU composition (Fig. 6). These soil variables explained 82.8% of the variation, and the first two axes explained 64.0% and 9.9% of the total variation. According to the vectors, the diazotrophic communities of both CK and PK treatment were associated with higher pH values, and the PK treatment was

Table 2 Pearson's correlation coefficients between total nifH gene copy number and various soil variables.

nifH gene copies in bulk soils nifH gene copies in rhizosphere soils nifH gene copies in all samples **p < 0.01; *p < 0.05.

pH

TC

TN

C/N

SOC

DOC

NHþ 4 -N

NO 3 -N

DON

AK

AP

0.567* 0.745** 0.620**

0.344 0.154 0.267

0.186 0.313 0.247

0.602** 0.718** 0.638**

0.364 0.171 0.437*

0.311 0.203 0.498*

0.595** 0.625** 0.659**

0.630** 0.752** 0.622**

0.311 0.406 0.523*

0.393 0.504* 0.276

0.565* 0.356 0.317

C. Wang et al. / Soil Biology & Biochemistry 113 (2017) 240e249

245

þ were positively affected by NO 3 -N and NH4 -N. These soil variables were also correlated with most of the abundant genera (Table S6).

4. Discussion 4.1. Long-term inorganic fertilization influenced soil diazotrophic abundance and OTU richness

Fig. 3. Non-metric multi-dimensional scaling analysis (A) and hierarchical cluster analysis (B) of diazotrophic community composition in sampled maize rhizosphere and bulk soils from the 25-year (1990e2015) long-term fertilizer experiment. B indicates bulk soil, and R indicates rhizosphere soil. The samples were analyzed in triplicate plots.

Fig. 4. Numbers of operational taxonomic units across all of the analyzed bulk samples from the 25-year (1990e2015) long-term fertilizer experiment. Data are means ± standard deviation (n ¼ 3). Different letters above columns indicate significant differences among fertilizer treatments at p < 0.05.

also positively related to C/N and AK. Additionally, the diazotrophic populations of NPK were positively affected by higher levels of TC, SOC, and DOC, while the diazotrophic populations of N and NCa

Long-term inorganic fertilization was frequently reported to decrease soil microbial abundance (Geisseler and Scow, 2014; Sun et al., 2015). Similarly, the input of N fertilizers (such as N and NPK treatments) reduced soil diazotrophic abundance (Fig. 2), probably because of negative feedback from soil acidification or/ and high N content. Such a negative response in microbial abundance to low soil pH has be well demonstrated in several long-term trials (Geisseler and Scow, 2014; Levy-Booth et al., 2014), resulting in soil pH being considered a decisive factor affecting soil microbial abundance. Thus, a consistent result was expected in the current trial (Table 2), in which 25 years of inorganic N fertilizer input resulted in intensified soil acidification (Table 1). Additionally, the significant increase in diazotrophic abundance after adding quicklime further supports this explanation (Fig. 2). An N source is essential for the growth and metabolism of organisms, but N enrichment generally causes a reduction in soil diazotrophic abundance (Pereira e Silva et al., 2013; Poly et al., 2001). Under  long-term N input, high soil NHþ 4 -N and NO3 -N contents significantly inhibited members of the diazotrophic community (Coelho et al., 2008; Roesch et al., 2006), showing significant negative correlations with diazotrophic abundance (Table 2). In contrast, in treatments without N (CK and PK), long-term N deficiencies resulted in greater C/N ratios (Table 1), resulting in a competitive advantage for free-living diazotrophs (Mirza et al., 2014). Furthermore, the greater diazotrophic abundance under the PK treatment relative to that under CK could be attributed to the positive role of P, which has potential for biological N fixation (Wurzburger et al., 2012). This explains the positive correlation between diazotrophic abundance and both the C/N and AP content (Table 2). Thus, a dramatic shift in diazotrophic abundance may be a comprehensive consequence of both direct effects of fertilizers as nutrients and indirect effects of an altered soil pH, which is the result of longterm inorganic fertilization. The increased input of N fertilizer results in a decrease in bacterial OTU richness (Zeng et al., 2016; Zhou et al., 2015) and diazotrophic diversity in some agricultural soils (Coelho et al., 2009; Roesch et al., 2006). Soil pH is generally considered as a major impact factor (Shen et al., 2016; Zhou et al., 2015). However, we observed the decreased diazotrophic OTU richness in all of the treatments (Fig. 4) and the markedly negative correlation between OTU richness and TN and NO 3 -N (Table S4), suggesting that soil nutrients are as important as pH in decreasing the diazotrophic OTU richness when subjected to long-term inorganic fertilization. Similarly, Zeng et al. (2016) reported that in response to high N input, soil nutrients were directly responsible for the decrease in the bacterial OTU richness of the surface soil of farmlands. The focus on soil pH was probably because the high correlation between the N input and soil pH masked the N effect. In this trial, no increase in diazotrophic OTU richness occurred after the increase in soil pH when quicklime was added (Fig. 4), which corroborated the idea that soil pH may not be the main factor reducing diazotrophic OTU richness. Thus, some of the species of diazotrophic bacteria eliminated by long-term inorganic fertilization were difficult to restore, even though the soil pH was briefly returned to the control level.

246

C. Wang et al. / Soil Biology & Biochemistry 113 (2017) 240e249

Fig. 5. Relative abundances (%) of the nine most abundant genera (>1%) across all of the analyzed bulk soil samples from the 25-year (1990e2015) long-term fertilizer experiment. Data are means ± standard deviation (n ¼ 3). Different letters above columns in each genus indicate significant differences among fertilizer treatments at p < 0.05.

Fig. 6. Redundancy analysis of the diazotrophic community structure across all of the analyzed bulk soil samples from the 25-year (1990e2015) long-term fertilizer experiment. The positions and lengths of the arrows indicate the directions and strengths, respectively, of the effects of variables on the diazotrophic communities. The samples were analyzed in triplicate plots.

4.2. Long-term inorganic fertilization changed the diazotrophic community structure Long-term applications of inorganic fertilizers, especially N fertilizer, changed the soil microbial community structure, including the overall and ammonia-oxidizing bacterial, fungal, and ammonia-oxidizing archaeal levels (Ai et al., 2015; He et al., 2007; Li et al., 2014; Shen et al., 2016; Xun et al., 2016a). As expected, a shift in the diazotrophic community composition in response to long-term inorganic fertilization was observed (Fig. 3). First, the diazotrophic community composition under treatments containing N was significantly different from those not receiving N (PK and CK), indicating that compared with both P and K, N has a greater influence on soil diazotrophic community structure. Similarly, Coelho et al. (2008) found that N fertilizer is a determining factor for the diazotrophic community structure. The formation of a microbial community's structure is a complex process and requires the convergence of several soil variables (Geisseler and Scow, 2014).

The high N availability, brought on by N input, directly influences microbial community structure (Li et al., 2014; Zeng et al., 2016).  Likewise, N-related soil properties, including NHþ 4 -N, NO3 -N, and C/ N, were the most dominant factors shaping diazotrophic community structure (Table S5 and Fig. 6), as observed in other studies (Zhong et al., 2015; Zhou et al., 2015). In addition, soil AP and AK were the other two important factors (Table S5) and increased the diazotrophic populations under both PK and NPK treatments (Fig. 6). Therefore, soil nutrient availability, which depends on the elements supplied by the fertilizers, is central to the formation of the diazotrophic community's structure. In addition to soil nutrient availability, long-term inorganic fertilization increased soil organic C contents (Table 1), which corroborated previous study results from this experimental area (Xun et al., 2016b; Zhang et al., 2009), as well as those of other longterm experiments (Ai et al., 2015; Wang et al., 2015; Zhou et al., 2015). In this trial, the soil available C pool also affected diazotrophic community structure (Table S5 and Fig. 6). Cusack et al. (2011) reported that the variations in the amount and quality of soil organic matter was related to the shift in the soil microbial community under different fertilization practices. Indeed, the balanced fertilization in agricultural management could maintain and increase soil C availability (TC, SOC, and DOC) (Table 1), which resulted in a difference between the diazotrophic communities under the NPK and N treatments (Fig. 6). For example, several dominant diazotrophic genera (Phaeospirillum and Azospirillum) showed positive TC, SOC, and DOC responses (Table S6). Thus, improved agricultural production through balanced fertilization in farmland ecosystems contributed to the relevant microbial community, which was induced by C sources (Zhong et al., 2015). Soil pH has been well investigated in a number of ecosystems and is frequently considered an important determinant of bacterial community structure (Geisseler and Scow, 2014). However, such a decisive role was not observed in the current study, in which soil pH was less important compared with the soil nutrient availability discussed above (Table S5). Similar results were observed in other long-term fertilization experiments (Huang et al., 2016; Zhao et al., 2016; Zhong et al., 2015). Furthermore, most of the abundant genera were closely correlated with soil available nutrients rather than soil pH (Table S6). Thus, the response mechanism of the diazotrophic community structure to long-term inorganic fertilization may be different from those of ammonium oxidizers and

C. Wang et al. / Soil Biology & Biochemistry 113 (2017) 240e249

nitrifying organisms, which are more susceptible to soil pH variations (Geisseler and Scow, 2014; Wang et al., 2015). This may be attributed to the differences in soil type and nutrients, as well as the microbial population (Daquiado et al., 2016). Recently, Xun et al. (2016b) reported a distinct change in the overall bacterial community composition under NPK treatments following quicklime application, whereas another study found that short-term quicklime additions did not significantly alter the bacterial communities in long-term fertilization trials (Xun et al., 2016a). Similarly, we did not observe separation difference in the diazotrophic communities under N treatments with and without quicklime at the current research scale (Fig. 3). This result further confirmed that soil nutrients resulting from long-term inorganic fertilization have a greater effect than soil pH on soil diazotrophic community structure. 4.3. Responses of dominant diazotrophic genera to long-term inorganic fertilization Although the taxonomy of diazotrophic bacteria has been investigated extensively (Indrasumunar et al., 2012), little is currently known of the influences of long-term fertilization conditions. The responses of specific diazotrophic genera to long-term inorganic fertilization varied considerably, depending strongly on the fertilizer type (Fig. 5). Because different diazotrophic genera have different abilities to adapt to various nutrients and soil pH levels (Table S6), fertilization could change their competitiveness. In the current trial, although Bradyrhizobium species, commonly known as symbiotic N2-fixing bacteria, had significantly reduced relative abundances under all of the treatments, they may play an important role in N fixation in acid soils (Indrasumunar et al., 2012) as they are the most abundant (Fig. 5). Similarly, the relative abundances of the genera Rhodopseudomonas, Xanthobacter, Caenispirillum, and Methylosinus responded distinctly to N fertilizer input rather than quicklime application, suggesting that they are very sensitive to N. These observations were consistent with previous studies, in which low contents of inorganic N sources inhibited the growth of Bradyrhizobium (Wang and Stacey, 1990), Rhodopseudomonas (Yagi et al., 1994), and Methylosinus (King and Schnell, 1994) species. However, the PK treatment selectively stimulated some diazotrophic genera, including Rhodopseudomonas, Caenispirillum, and Phaeospirillum (Fig. 5), which was found in other experiments (Bellenger et al., 2014; Videira et al., 2013). The different response of each genus to fertilizer type could be directly responsible for the shift in diazotrophic community structure. 4.4. Impact of long-term inorganic fertilization on the rhizosphere effect The rhizosphere is a complex environmental spot in which plant growth generally influences the soil microbiome by affecting the interactions with soil variables (Hinsinger et al., 2009). A diverse array of organic exudates released by plant roots could stimulate the growth of microbial populations, and thus, the corresponding microbial community composition may be formed in the rhizosphere (Berendsen et al., 2012; Hinsinger et al., 2009). The increased abundance of the diazotrophic population in the rhizosphere has been widely reported under various environmental conditions (Coelho et al., 2009; Li et al., 2012; Rodríguez-Blanco et al., 2015). As expected, we observed that the total diazotrophic abundance in the rhizosphere was greater than that in the bulk soil across all of the treatments (Fig. 2), indicating that there could be the higher levels of biological N fixation in the rhizosphere region. The greater available C contents and lower available N contents

247

were observed in rhizosphere soil relative to bulk soil in the current trial, with a significantly positive effect on diazotrophic abundance (Table 2). Similarly, the previous studies suggested that the release of root exudates and soil nutrient uptake, brought on by plant growth, stimulated the biological N fixation (Levy-Booth et al., 2014). The distinct differences in plant growth among fertilizer treatments may affect relationships between plants and soil microbial populations (Hinsinger et al., 2009). In the present trial, the root effect on the soil diazotrophic community produced inconsistent results in previous studies. Some papers reported that plants could drive the distinct shifts in the diazotrophic community within a given site or soil type (Coelho et al., 2008; Li et al., 2012; RodríguezBlanco et al., 2015; Wang et al., 2012), and we obtained a similar result in another short-term culture experiment (data not published). In contrast, Ai et al. (2015) suggested that the increased soil nutrient availability in long-term fertilization practices decreased the dependence of the rhizosphere microbiome on plants. In the current trial, compared with fertilizer treatments, root effect showed little impact on the diazotrophic community (Fig. 3 and Table S3). Likewise, Poly et al. (2001) and Reardon et al. (2014) found that diazotrophic community structures were influenced more by N fertilization than crop. Cotta et al. (2014) also found that the influence of soil type and associated characteristics on the soil diazotrophic community was far greater than root effect. Longterm fertilization may form highly specialized soil characteristics and a physically non-perturbed soil environment (Ai et al., 2015), which are not significantly affected by plants in a single growth season. This is supported by the soil property analysis in the current trial, in which fertilizer treatments had greater impacts than plant factors (Table 1). The detailed mechanisms of interactions between plants and microbes are complicated and need further study, especially under long-term exposure to different fertilizers. 5. Conclusions This study demonstrated that 25 years of inorganic fertilization resulted in strongly selective forces acting on the soil diazotrophic population. It is likely that N has a greater influence than P and K on diazotrophic bacteria. The lower diazotrophic abundance under N fertilizer treatments may diminish the capability of biological N fixation. Quicklime applications may improve biological N fixation in acidic soil by increasing soil diazotrophic abundance, although their influence on diazotrophic community composition and OTU richness was not significant. The different response patterns of diazotrophic abundance, community composition, and OTU richness to soil characteristics revealed a complicated mechanism behind the diazotrophic population's adaptation to long-term inorganic fertilization. Balanced fertilization and the control of soil acidity may play important roles in improving biological N fixation. For the overall community composition, fertilizer treatments showed a greater influence than the rhizosphere effect, although the rhizosphere significantly enhanced diazotrophic abundance. Future investigations will determine which specific root exudates are more effective in regulating the increase in diazotrophic abundance and if application could stimulate soil biological N fixation, it will reduce the required input of inorganic N fertilizer. Acknowledgements This work was financially supported by the “Strategic Priority Research Program” of the Chinese Academy of Sciences (No. XDB15030000), the National Key Basic Research Program of China (No. 2014CB441000), the Frontal Field Project of the Chinese

248

C. Wang et al. / Soil Biology & Biochemistry 113 (2017) 240e249

Academy of Sciences (No. ISSASIP1638) and the Natural Science Foundation of China (No. 41501328). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.soilbio.2017.06.019. References Ahamadou, B., Huang, Q.Y., Chen, W.L., Wen, S.L., Zhang, J.Y., Mohamed, I., Cai, P., Liang, W., 2009. Microcalorimetric assessment of microbial activity in longterm fertilization experimental soils of Southern China. FEMS Microbiology Ecology 70, 186e195. Ai, C., Liang, G.Q., Sun, J.W., Wang, X.B., He, P., Zhou, W., He, X.H., 2015. Reduced dependence of rhizosphere microbiome on plant-derived carbon in 32-year long-term inorganic and organic fertilized soils. Soil Biology & Biochemistry 80, 70e78. Bellenger, J.P., Xu, Y., Zhang, X., Morel, F.M.M., Kraepiel, A.M.L., 2014. Possible contribution of alternative nitrogenases to nitrogen fixation by asymbiotic N2fixing bacteria in soils. Soil Biology & Biochemistry 69, 413e420. Berendsen, R.L., Pieterse, C.M., Bakker, P.A., 2012. The rhizosphere microbiome and plant health. Trends in Plant Science 17, 478e486. Bray, R.H., Kurtz, L.T., 1945. Determination of total, organic, and available forms of phosphorus in soils. Soil Science 59, 39e46. Cai, Z.J., Wang, B.R., Xu, M.G., Zhang, H.M., He, X.H., Zhang, L., Gao, S.D., 2015. Intensified soil acidification from chemical N fertilization and prevention by manure in an 18-year field experiment in the red soil of southern China. Journal of Soils and Sediments 15, 260e270. Chakraborty, A., Chakrabarti, K., Chakraborty, A., Ghosh, S., 2011. Effect of long-term fertilizers and manure application on microbial biomass and microbial activity of a tropical agricultural soil. Biology and Fertility of Soils 47, 227e233. Che, J., Zhao, X.Q., Zhou, X., Jia, Z.J., Shen, R.F., 2015. High pH-enhanced soil nitrification was associated with ammonia-oxidizing bacteria rather than archaea in acidic soils. Applied Soil Ecology 85, 21e29. Coelho, M.R.R., de Vos, M., Carneiro, N.P., Marriel, I.E., Paiva, E., Seldin, L., 2008. Diversity of nifH gene pools in the rhizosphere of two cultivars of sorghum (Sorghum bicolor) treated with contrasting levels of nitrogen fertilizer. FEMS Microbiology Letters 279, 15e22. Coelho, M.R.R., Marriel, I.E., Jenkins, S.N., Lanyon, C.V., Seldin, L., O'Donnell, A.G., 2009. Molecular detection and quantification of nifH gene sequences in the rhizosphere of sorghum (Sorghum bicolor) sown with two levels of nitrogen fertilizer. Applied Soil Ecology 42, 48e53. Cotta, S.R., Dias, A.C.F., Marriel, I.E., Andreote, F.D., Seldin, L., van Elsas, J.D., 2014. Different effects of transgenic maize and nontransgenic maize on nitrogentransforming archaea and bacteria in tropical soils. Applied and Environmental Microbiology 80, 6437e6445. Cusack, D.F., Silver, W.L., Torn, M.S., Burton, S.D., Firestone, M.K., 2011. Changes in microbial community characteristics and soil organic matter with nitrogen additions in two tropical forests. Ecology 92, 621e632. Daquiado, A.R., Kuppusamy, S., Kim, S.Y., Kim, J.H., Yoon, Y.E., Kim, P.J., Oh, S.H., Kwak, Y.S., Lee, Y.B., 2016. Pyrosequencing analysis of bacterial community diversity in long-term fertilized paddy field soil. Applied Soil Ecology 108, 84e91. Gaby, J.C., Buckley, D.H., 2014. A comprehensive aligned nifH gene database: a multipurpose tool for studies of nitrogen-fixing bacteria. Database 2014, bau001. Geisseler, D., Scow, K.M., 2014. Long-term effects of mineral fertilizers on soil microorganisms - a review. Soil Biology & Biochemistry 75, 54e63. Guo, J.H., Liu, X.J., Zhang, Y., Shen, J.L., Han, W.X., Zhang, W.F., Christie, P., Goulding, K.W.T., Vitousek, P.M., Zhang, F.S., 2010. Significant acidification in major Chinese croplands. Science 327, 1008e1010. Gupta, V.V.S.R., Roper, M.M., Roget, D.K., 2006. Potential for nonsymbiotic N2-fixation in different agroecological zones of southern Australia. Australian Journal of Soil Research 44, 343e354. He, J.Z., Shen, J.P., Zhang, L.M., Zhu, Y.G., Zheng, Y.M., Xu, M.G., Di, H.J., 2007. Quantitative analyses of the abundance and composition of ammonia-oxidizing bacteria and ammonia-oxidizing archaea of a Chinese upland red soil under long-term fertilization practices. Environmental Microbiology 9, 3152. Hinsinger, P., Bengough, A.G., Vetterlein, D., Young, I.M., 2009. Rhizosphere: biophysics, biogeochemistry and ecological relevance. Plant and Soil 321, 117e152. Huang, J.X., Xu, X., Wang, M., Nie, M., Qiu, S.Y., Wang, Q., Quan, Z.X., Xiao, M., Li, B., 2016. Responses of soil nitrogen fixation to Spartina alterniflora invasion and nitrogen addition in a Chinese salt marsh. Scientific Reports 6, 20384. Indrasumunar, A., Menzies, N.W., Dart, P.J., 2012. Laboratory prescreening of Bradyrhizobium japonicum for low pH, Al and Mn tolerance can be used to predict their survival in acid soils. Soil Biology & Biochemistry 48, 135e141. King, G.M., Schnell, S., 1994. Ammonium and nitrite inhibition of methane oxidation by Methylobacter albus BG8 and Methylosinus trichosporium OB3b at low methane concentrations. Applied and Environmental Microbiology 60, 3508e3513. Levy-Booth, D.J., Prescott, C.E., Grayston, S.J., 2014. Microbial functional genes

involved in nitrogen fixation, nitrification and denitrification in forest ecosystems. Soil Biology & Biochemistry 75, 11e25. Li, C.H., Yan, K., Tang, L.S., Jia, Z.J., Li, Y., 2014. Change in deep soil microbial communities due to long-term fertilization. Soil Biology & Biochemistry 75, 264e272. Li, Y.C., Yang, T.Y., Zhang, P.P., Zou, A.L., Peng, X., Wang, L.L., Yang, R.W., Qi, J.L., Yang, Y.H., 2012. Differential responses of the diazotrophic community to aluminum-tolerant and aluminum-sensitive soybean genotypes in acidic soil. European Journal of Soil Biology 53, 76e85. Liu, Y.B., Pan, X.B., Li, J.S., 2015. A 1961-2010 record of fertilizer use, pesticide application and cereal yields: a review. Agronomy for Sustainable Development 35, 83e93. Magoc, T., Salzberg, S.L., 2011. FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinformatics 27, 2957e2963. Mirza, B.S., Potisap, C., Nusslein, K., Bohannan, B.J.M., Rodrigues, J.L.M., 2014. Response of free-living nitrogen-fixing microorganisms to land use change in the Amazon rainforest. Applied and Environmental Microbiology 80, 281e288. Pereira e Silva, M.C., Schloter-Hai, B., Schloter, M., van Elsas, J.D., Salles, J.F., 2013. Temporal dynamics of abundance and composition of nitrogen-fixing communities across agricultural soils. PLoS One 8, e74500. Poly, F., Ranjard, L., Nazaret, S., Gourbiere, F., Monrozier, L.J., 2001. Comparison of nifH gene pools in soils and soil microenvironments with contrasting properties. Applied and Environmental Microbiology 67, 2255e2262. Reardon, C.L., Gollany, H.T., Wuest, S.B., 2014. Diazotroph community structure and abundance in wheat-fallow and wheat-pea crop rotations. Soil Biology & Biochemistry 69, 406e412. Rodríguez-Blanco, A., Sicardi, M., Frioni, L., 2015. Plant genotype and nitrogen fertilization effects on abundance and diversity of diazotrophic bacteria associated with maize (Zea mays L.). Biology and Fertility of Soils 51, 391e402. Roesch, L.F.W., Olivares, F.L., Passaglia, L.M.P., Selbach, P.A., de Sa, E.L.S., de Camargo, F.A.O., 2006. Characterization of diazotrophic bacteria associated with maize: effect of plant genotype, ontogeny and nitrogen-supply. World Journal of Microbiology and Biotechnology 22, 967e974. Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., Lesniewski, R.A., Oakley, B.B., Parks, D.H., Robinson, C.J., Sahl, J.W., Stres, B., Thallinger, G.G., Van Horn, D.J., Weber, C.F., 2009. Introducing mothur: opensource, platform-independent, community-supported software for describing and comparing microbial communities. Applied and Environmental Microbiology 75, 7537e7541. Shen, W.S., Ni, Y.Y., Gao, N., Bian, B.Y., Zheng, S.N., Lin, X.G., Chu, H.Y., 2016. Bacterial community composition is shaped by soil secondary salinization and acidification brought on by high nitrogen fertilization rates. Applied Soil Ecology 108, 76e83. Simonsen, A.K., Han, S., Rekret, P., Rentschler, C.S., Heath, K.D., Stinchcombe, J.R., 2015. Short-term fertilizer application alters phenotypic traits of symbiotic nitrogen fixing bacteria. Peerj 3, e1291. Sims, J.R., Haby, V.A., 1971. Simplified colorimetric determination of soil organic matter. Soil Science 112, 137e141. Sun, R.B., Guo, X.S., Wang, D.Z., Chu, H.Y., 2015. Effects of long-term application of chemical and organic fertilizers on the abundance of microbial communities involved in the nitrogen cycle. Applied Soil Ecology 95, 171e178. Videira, S.S., Pereira e Silva, M.C., Galisa, P.S., Dias, A.C.F., Nissinen, R., Divan, V.L.B., van Elsas, J.D., Baldani, J.I., Salles, J.F., 2013. Culture-independent molecular approaches reveal a mostly unknown high diversity of active nitrogen-fixing bacteria associated with Pennisetum purpureum-a bioenergy crop. Plant and Soil 373, 737e754. Wang, C., Zhao, X.Q., Chen, R.F., Chu, H.Y., Shen, R.F., 2013. Aluminum tolerance of wheat does not induce changes in dominant bacterial community composition or abundance in an acidic soil. Plant and Soil 367, 275e284. Wang, S., Pablo, G.P., Ye, J., Huang, D.F., 2012. Abundance and diversity of nitrogenfixing bacteria in rhizosphere and bulk paddy soil under different duration of organic management. World Journal of Microbiology and Biotechnology 28, 493e503. Wang, S.P., Stacey, G., 1990. Ammonia regulation of nod genes in Bradyrhizobium japonicum. Molecular Genetics and Genomics 223, 329e331. Wang, X.L., Han, C., Zhang, J.B., Huang, Q.R., Deng, H., Deng, Y.C., Zhong, W.H., 2015. Long-term fertilization effects on active ammonia oxidizers in an acidic upland soil in China. Soil Biology & Biochemistry 84, 28e37. Watkins, K.L., Veum, T.L., Krause, G.F., 1987. Total nitrogen determination of various dample types: a comparison of the Hach, Kjeltec, and Kjeldahl methods. Journal-Association of Official Analytical Chemists 70, 410e412. Wurzburger, N., Bellenger, J.P., Kraepiel, A.M.L., Hedin, L.O., 2012. Molybdenum and phosphorus interact to constrain asymbiotic nitrogen fixation in tropical forests. PLoS One 7, e33710. Xun, W.B., Zhao, J., Xue, C., Zhang, G.S., Ran, W., Wang, B.R., Shen, Q.R., Zhang, R.F., 2016a. Significant alteration of soil bacterial communities and organic carbon decomposition by different long-term fertilization management conditions of extremely low-productivity arable soil in South China. Environmental Microbiology 18, 1907e1917. Xun, W.B., Xiong, W., Huang, T., Ran, W., Li, D.C., Shen, Q.R., Li, Q., Zhang, R.F., 2016b. Swine manure and quicklime have different impacts on chemical properties and composition of bacterial communities of an acidic soil. Applied Soil Ecology 100, 38e44. Yagi, K., Maeda, I., Idehara, K., Miura, Y., Akano, T., Fukatu, K., Ikuta, Y., Nakamura, H.K., 1994. Removal of inhibition by ammonium ion in nitrogenase-

C. Wang et al. / Soil Biology & Biochemistry 113 (2017) 240e249 dependent hydrogen evolution of a marine photosynthetic bacterium, Rhodopseudomonas sp. Strain W-1S. Applied Biochemistry and Biotechnology 45, 429e436. Yin, C., Fan, F.L., Song, A.L., Cui, P.Y., Li, T.Q., Liang, Y.C., 2015. Denitrification potential under different fertilization regimes is closely coupled with changes in the denitrifying community in a black soil. Applied Microbiology and Biotechnology 99, 5719e5729. Zeng, J., Liu, X.J., Song, L., Lin, X.G., Zhang, H.Y., Shen, C.C., Chu, H.Y., 2016. Nitrogen fertilization directly affects soil bacterial diversity and indirectly affects bacterial community composition. Soil Biology & Biochemistry 92, 41e49. Zhang, H.M., Wang, B.R., Xu, M.G., Fan, T.L., 2009. Crop yield and soil responses to long-term fertilization on a red soil in southern China. Pedosphere 19, 199e207.

249

Zhao, J., Ni, T., Li, J., Lu, Q., Fang, Z.Y., Huang, Q.W., Zhang, R.F., Li, R., Shen, B., Shen, Q.R., 2016. Effects of organic-inorganic compound fertilizer with reduced chemical fertilizer application on crop yields, soil biological activity and bacterial community structure in a rice-wheat cropping system. Applied Soil Ecology 99, 1e12. Zhong, Y.Q.W., Yan, W.M., Shangguan, Z.P., 2015. Impact of long-term N additions upon coupling between soil microbial community structure and activity, and nutrient-use efficiencies. Soil Biology & Biochemistry 91, 151e159. Zhou, J., Guan, D., Zhou, B., Zhao, B., Ma, M., Qin, J., Jiang, X., Chen, S., Cao, F., Shen, D., Li, J., 2015. Influence of 34-years of fertilization on bacterial communities in an intensively cultivated black soil in northeast China. Soil Biology & Biochemistry 90, 42e51.