Differential responses of soil bacterial communities to

0 downloads 0 Views 1MB Size Report
Jul 5, 2018 - lected from a semi-arid steppe located in Inner Mongolia, were subject- .... Kit (MoBio Laboratories Inc., Carlsbad, CA, USA), according to the.

Geoderma 292 (2017) 25–33

Contents lists available at ScienceDirect

Geoderma journal homepage: www.elsevier.com/locate/geoderma

Differential responses of soil bacterial communities to long-term N and P inputs in a semi-arid steppe Ning Ling a, Dima Chen b, Hui Guo a, Jiaxin Wei a, Yongfei Bai b, Qirong Shen a, Shuijin Hu a,c,⁎ a b c

College of resources and environmental sciences, Nanjing Agricultural University, Nanjing 210095, China State Key Laboratory of Vegetation and Environmental Change, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China Department of Plant Pathology, North Carolina State University, Raleigh, NC 27695-7616, USA

a r t i c l e

i n f o

Article history: Received 16 September 2016 Accepted 9 January 2017 Available online 15 January 2017 Keywords: Semi-arid steppe N or P inputs Soil bacterial diversity Soil bacterial community structure Illumina Miseq sequencing

a b s t r a c t Both nitrogen (N) and phosphorus (P) may limit plant production in steppes and affect plant community structure. However, few studies have explored in detail the differences and similarities in the responses of belowground microbial communities to long-term N and P inputs. Using a high-throughput Illumina Miseq sequencing platform, we characterized the bacterial communities in a semi-arid steppe subjected to long-term N or P additions. Our results showed that both the Chao richness and Shannon's diversity were negatively correlated to N input rate, while only Chao richness was significantly and negatively correlated to P input rate. Also, both N and P additions altered the bacterial community structure. The bacterial community between plots of the same N or P input rate was much more dissimilar with the higher input level, indicating more severe niche differentiation in pots with higher N or P input. N Inputs significantly increased the relative abundance of the predicted copiotrophic groups (Proteobacteria and Firmicutes) but reduced the predicted oligotrophic groups (Acidobacteria, Nitrospirae, Chloroflexi), with the order Rhizobiales being most affected. P additions significantly affected only two phyla (Armatimonadetes and Chlorobi), which were positively correlated with P source. Results from the structural equation modelling (SEM) showed that N additions affected the bacterial community primarily by changing the pH, while P additions did so mainly by improving P availability. Our results suggest that the below-ground bacterial communities are more sensitive to N inputs, but P inputs can also play an important role in bacterial niche differentiation. These findings improve our understanding of bacterial responses to N and P inputs, and their impacts on bacterial-mediated processes, especially in the context of increasing anthropogenic nutrient inputs. © 2017 Elsevier B.V. All rights reserved.

1. Introduction The Inner Mongolian steppes comprise N 20% of the total grassland area in China, in which the plant communities play an important role in livestock farming and ecosystem services (Xu et al., 2014). To maximize biomass production, chronic nutrient additions, often in the form of chemical fertilizers (i.e., NPK) or animal manures, often occur in grasslands to maximize biomass production, this greatly contributes to the release of nutrients for plant uptake and growth (Zhou et al., 2015). Moreover, anthropogenic reactive nitrogen (N) inputs, which mainly originate from fossil-fuel burning and artificial fertilizer application, have increased three- to five-fold over the past century (Galloway et al., 2008) and is expected to continue to increase, especially in Asia (Chen et al., 2015a). In addition, it is generally believed that plant growth is limited by P in many grassland soils, especially calcareous ⁎ Correspondence author at: College of Resources and Environmental Sciences, Nanjing Agricultural University, Nanjing 210095, China. E-mail address: [email protected] (S. Hu).

http://dx.doi.org/10.1016/j.geoderma.2017.01.013 0016-7061/© 2017 Elsevier B.V. All rights reserved.

grasslands forming complexes with calcium (Elser et al., 2007; Ford et al., 2016), thereby the continually anthropogenic P inputs are inevitable for maintaining high grass biomass. All of these nutrient inputs may influence the structure of soil microbial communities, which play a pivotal role in regulating soil functioning and maintaining ecosystem sustainability (van der Heijden et al., 2008). Despite the widely acknowledged importance of soil microorganisms, how bacteria ultimately respond to long-term repeated inputs of chemical fertilizers remains poorly understood. Nitrogen (N) and phosphorus (P) should be the essential nutrients necessary to increase the grass biomass production in Inner Mongolian steppes. Their input often has multiple effects including changes in aboveground primary productivity, biodiversity, species composition, and ecosystem functioning (Bai et al., 2010; Li et al., 2010; Yang et al., 2015). Soil microbiota are recognized as key players in sustaining ecosystem functions and services. Therefore, elucidating the feedbacks of the soil microbiome to N and P inputs is fundamental to understanding the consequences of global changes on ecosystem processes regulated by soil biota. Recent studies on the Inner Mongolian steppe were mainly

26

N. Ling et al. / Geoderma 292 (2017) 25–33

concerned with N-induced soil microbial changes. Significant effects of N addition on bacterial communities were reported and were ascribed to N-induced pH changes, followed by N content and N forms (Zhang et al., 2014a; Chen et al., 2015b; Yang et al., 2015). Other soil chemical changes such as soil total C or P, impacted the microbial communities (Chen et al., 2013). Compared with studies on N addition, only a few studies focused on the effects of P addition on soil microbial biomass and the microbial community, and most of these studies were conducted in agricultural ecosystems (Beauregard et al., 2009; Shi et al., 2013; Tan et al., 2013). Understanding the effects of both N and P addition on soil microbial activities can improve our capacity to predict how the soil microbial community will respond to environmental changes in semi-arid steppe ecosystems. This information will help to develop effective strategies for the management and sustainability of ecosystems under nutrient additions. Although previous studies focused on the response of the soil microbial community to N addition, only a few studies compared the difference between the impacts of N and P inputs on the bacterial community in Inner Mongolian steppes, China. In this study, soils, collected from a semi-arid steppe located in Inner Mongolia, were subjected to a 16-year N or P addition with gradient rates. We used the highthroughput Illumina Miseq sequencing platform to characterize the bacterial communities' response to the two types of nutrients under varying simulated input rates. The following questions were specifically addressed: (1) Do the bacterial communities have differences and similarities in response to the gradient N and P inputs? and (2) What would be the primary factor resulting from the gradient N or P inputs to impact the soil bacterial communities in the steppes ecosystem? 2. Materials and methods 2.1. Long-term experiment site description The long-term N and P input experiments were conducted at the Inner Mongolia Grassland Ecosystem Research Station (IMGERS, 43°38′N, 116°42′E) of the Chinese Academy of Sciences, which is located in the Xilin River Basin of Inner Mongolia, China, at an altitude of approximately 1200 m a.s.l. (Bai et al., 2004). Before the experiment began, the site was dominated by Leymus chinensis, a widely distributed perennial C3 rhizome grass in the Eurasian steppe region (Yao et al., 2014). The mean annual precipitation in this area is about 340 mm, with c. 80% rainfall occurring during the growing season from May to September. The average monthly temperature ranges from − 21.6 °C in January to 19.0 °C in July. Precipitation mainly falls in the growing season (June–August), which is coincident with high temperatures. The site has a dark chestnut soil (Calcic Chernozem, according to the ISSS Working Group RB, 1998), with a loamy-sand texture (Bai et al., 2010). 2.2. Experimental design and soil sampling The N and P input experiments were established in the autumn of 1999. An experimental plot was 5 × 5 m in size, arranged in a randomized block design and separated by a 1-m buffer zone (Bai et al., 2010). The plots were not harvested for hay and not grazed. For the N input experiment, five levels of N input rates, 0 g·N·m− 2 yr−1, 1.75 g·N·m− 2 yr− 1, 5.25 g·N·m− 2 yr− 1, 10.5 g·N·m− 2 yr− 1 and 28 g·N·m−2 yr−1, with pelletized NH4NO3 fertilizer were applied to the soil. To ensure that N was the only limiting nutrient, all of the treatments were also supplied with the same amounts of P (10 g·P2O5·m−2yr−1), S (0.2 mg·m−2 yr−1) and trace elements (Zn, 190 mg·m−2 yr−1, Mn, 160 mg·m−2 yr−1 and B, 31 mg·m−2 yr−1) based on local soil census data. For the P input experiment, five levels of P input rates, 0 g·P2O5·m− 2 yr−1, 2 g·P2O5·m− 2 yr− 1, 4 g·P2O5·m− 2 yr−1, 8 g·P2O5·m−2 yr−1 and 32 g·P2O5·m−2 yr−1, were applied to the soil with superphosphate fertilizer. To ensure that P was the only limiting

nutrient, all of the treatments were also supplied with the same amounts of N (2.4 g·N·m−2 yr−1) and the same rates of S and trace elements as in the N input experiment. Nutrients were uniformly applied to each plot with manual broadcasting in the mid-growing season (1–5 July) every year, coinciding with high temperatures and precipitation. Soil samples were collected in the middle of July 2015. Five replicates for each treatment were collected from five individual plots. Six soil cores (2 cm diameter, 0–10 cm deep) were randomly collected from each plot and combined to form one composite soil sample per plot. Soil samples were transported to the laboratory from the experimental site in a constant temperature box with ice. After the soil was gently mixed and roots were removed, the moist soil was passed through a 2-mm-mesh sieve and separated into two parts. One part was directly subjected to DNA extraction. The second part was airdried for the determination of soil pH, total organic C (TC), total N (TN) and available P (AP). Soil TC and TN were measured with an elemental analyser (Vario MAX; Elementar, Germany), and AP was extracted with 0.5 M·NaHCO3 and determined using the ammonium molybdate ascorbic method. 2.3. DNA extraction and MiSeq sequencing of 16S rRNA gene amplicons Total genomic DNA was extracted using a PowerSoil DNA Isolation Kit (MoBio Laboratories Inc., Carlsbad, CA, USA), according to the manufacturer's instructions. The quality and quantity of the DNA samples was checked by a spectrophotometer (NanoDrop, ND2000, Thermo Scientific, Wilmington, DE, USA) after extraction. The composition and diversity of bacterial communities were assessed by Illumina MiSeq sequencing analysis of the 16S rRNA gene. The universal primers 515F (5′-GTGCCAGCMGCCGCGG-3′) and 907R (5′-CCGTCAATTCMTTTRAGTTT-3′) were selected for the PCR amplification of the V4 region. The reverse primer contains a 6-bp errorcorrecting bar code unique to each sample. Illumina MiSeq sequencing was conducted by the Major Biotechnology Co., Ltd. (Shanghai, China) using an Illumina MiSeq platform. Sequences were submitted to the NCBI database under the accession number SRR3452783 and SRR3452785 for the N input experiment and P input experiment, respectively. 2.4. Sequence data analysis The raw sequences obtained were processed using the Quantitative Insights Into Microbial Ecology (QIIME) toolkit (Caporaso et al., 2010) and UPARSE pipeline (Edgar, 2013). All sequence reads were trimmed and assigned to each sample based on their barcodes. Multiple steps were required to trim the sequences, such as the removal of sequences b220 bp. For the samples from the N and P input experiments, each sample was rarefied to 14,618 and 14,605 reads, respectively, from the high quality sequences for both alpha-diversity (Chao estimator of richness, observed species and Shannon's diversity index) and beta-diversity (nonmetric multidimensional scaling and NMDS) analyses. The UPARSE pipeline was used to pick the operational taxonomic units (OTUs) to obtain an OTU table at a 97% identity threshold. Taxonomy was assigned using the Ribosomal Database Project classifier (Wang et al., 2007). 2.5. Statistical analysis All of the changes in the soil microbial communities were evaluated based on the OTU matrix. NMDS plots were used to visualize the structure among the samples based on the Euclidean distance of the OTU matrix in PAST (http://folk.uio.no/ohammer/past/). The statistical significance among the datasets was assessed by PerMANOVA using the Euclidean distance matrix in PAST. The N or P gradient rates were natural logarithm transformed to evaluate the Pearson's correlation coefficients with the Chao estimator of richness, Shannon's diversity index

N. Ling et al. / Geoderma 292 (2017) 25–33

27

p b 0.01) between the samples with 0–5.25 g·N·m−2 yr−1 input and the samples with 10.5–28 g·N·m−2 yr−1 input (Fig. 2a). The bacterial community structure was also significantly affected by the P input (Fig. 2b). According to the perMANOVA tests, the highest P input (32 P2O5·m−2 yr−1) was significantly (p b 0.01) different than the other P input treatments. Moreover, the bacterial community structure similarities among the same input rates of N (Fig. 2c) or P was obtained based on the Euclidean distance. Overall, when all of the bacterial communities came from the same input level across the N or P gradient input treatments, their bacterial communities became significantly more dissimilar with increasing N or P input levels (p b 0.0001). The N gradient input explained 38% of the bacterial structure similarity variation within the same input rate, and the P gradient input explained 43% of the bacterial structure differentiation within the same input rate. The slope of the relationship between N gradient rates and the bacterial structure similarity distance (377.3) was much higher than that between the P gradient rates and the bacterial structure similarity distance (66.16), suggesting that N input could have a higher impact on the bacterial structure decay than P input. All of these results indicated that both the N and P inputs could change the bacterial structure. Even in plots with the same N or P input levels, their soil bacterial structures would be differentiated towards different directions, and the differentiation could more severe with the higher input level.

and the phyla, genera and community similarity (Euclidean distance) of the plots with the same input level. Each of the taxa must be present in at least two thirds of all of the samples (n = 25) in every input experiments to calculate the relationships with the N or P gradient rates. Cooccurrence networks between N gradient rates and genera and between P gradient rates and genera were constructed according to the Pearson's correlation coefficients (r N 0.6 or r b − 0.6) and p value (p b 0.05, Bonferroni corrected). The software Gephi (Bastian et al., 2009) was used to visualize the networks of co–occurring relationships. A structural equation model (SEM) was adopted to explore how the N input or P input influenced the bacterial richness, Shannon's diversity and structure by using AMOS software (IBM SPSS AMOS 20.0.0). The NMDS first axis scores of the samples were used as indicators of the bacterial community structure. Only soil variables that significantly correlated with bacterial richness, Shannon's diversity or structure were included in the model. We tested the fitness of the model with the data using the maximum likelihood (χ2) goodness-of-fit test, p value and root mean square error of approximation (RMSEA). 3. Results 3.1. Responses of soil properties to the N or P gradient input The soil pH value decreased with an increase of N input or P input (Table 1). In the N input experiment, the soil pH declined by 0.21– 2.38 units. In the P input experiment, the soil pH declined by 0.36– 0.6 units. P input significantly increased the soil available P (from 6.24 mg/kg to 422 mg/kg), which increased linearly with the increase of the P input rate. No significant impact was found with respect to the available P in the N input experiment. All of the other measured soil chemical properties (TC, TN and C/N) did not show increasing trends with increasing input N and P input rates, although there were some statistical differences between the treatments.

3.4. Responses of bacterial phylogenetic composition to the N or P gradient inputs The responsive differences of the phylogenetic compositions at the phylum and genus levels to the N or P gradient inputs were investigated according to Pearson's correlations. At the phylum level (Fig. 3), the N input significantly increased the relative abundances of Proteobacteria (r = 0.73, p b 0.0001), Firmicutes (r = 0.70, p = 0.0001) and Cyanobacteria (r = 0.67, p b 0.0003), while it significantly decreased the relative abundances of Acidobacteria (r = − 0.79, p b 0.0001), Planctomycetes (r = −0.68, p b 0.0002), Nitrospirae (r = −0.70, p = 0.0001), Elusimicrobia (r = − 0.76, p b 0.0001), Chloroflexi (r = −0.44, p b 0.05), Chlorobi (r = −0.44, p b 0.05) and candidate division WS3 (r = −0.70, p b 0.0001). Only two phyla were significantly correlated with the relative abundance vs. the P addition rate, Armatimonadetes (r = 0.61, p = 0.0012) and Chlorobi (r = 0.61, p = 0.0012) (Fig. 4). Intriguingly, both the N and P inputs could significantly affect the Chlorobi, but the responses of Chlorobi to the inputs were opposite. Moreover, at the genus level, more genera were significantly influenced by the N input than by the P input (Fig. 5). The N gradient rates significantly correlated with 64 genera (67% negative correlation and 33% positive correlation), while P gradient rates had only 15 correlations (40% negative correlation and 60% positive correlation) with the bacterial genera. For all of the genera that correlated with either N or P rates, most were classified as Rhizobiales (18.9%), followed by the Sphingobacteriales (10.8%), Xanthomonadales (10.8%) and then Bacillales

3.2. Responses of bacterial α-diversity to the N or P gradient inputs To understand the effect of the N and P inputs on bacterial α-diversity, the richness indicator (Chao) and the diversity indicator (Shannon index) were surveyed based on the OUT matrix (Fig. 1). The Chao richness and Shannon's diversity were both negatively correlated to the N input rates (p b 0.01), while only the Chao richness was significantly and negatively correlated to the P input rates (p b 0.05). There was no significant relationship between the bacterial Shannon's diversity and the P input rates. 3.3. Responses of bacterial β-diversity to the N or P gradient inputs The bacterial structure variations within the N or P gradient input experiments were evaluated by the NMDS (Fig. 2). The bacterial community structure was significantly different (perMANOVA tests,

Table 1 Soil properties at different N/P input rates. Treatment N input rates (g·N·m−2 yr−1)

P input rates (g·P2O5·m−2 yr−1)

pH 0 1.75 5.25 10.5 28 0 2 4 8 32

6.94 6.73 6.18 5.18 4.56 6.53 6.17 6.19 6.10 5.93

± ± ± ± ± ± ± ± ± ±

0.05a 0.24a 0.13ab 0.83bc 0.76c 0.27a 0.03b 0.07b 0.02bc 0.19c

TC (mg/kg)

TN (mg/kg)

C/N

AP (mg/kg)

23.82 22.18 24.00 23.32 22.68 22.10 23.42 23.16 22.46 23.08

2.28 2.16 2.32 2.24 2.32 2.20 2.26 2.56 2.18 2.20

10.45 ± 0.16a 10.28 ± 0.44a 10.35 ± 0.27a 10.43 ± 0.32a 9.78 ± 0.49b 10.04 ± 0.36a 10.37 ± 0.21a 9.45 ± 1.88a 10.29 ± 0.33a 10.48 ± 0.29a

141.9 ± 18.4ab 145.5 ± 37.5ab 127.1 ± 17.9b 163.2 ± 42.2ab 189.6 ± 22.8a 6.24 ± 0.52e 24.80 ± 1.83d 55.08 ± 7.07c 148.57 ± 23.18b 421.97 ± 93.57a

± ± ± ± ± ± ± ± ± ±

1.31ab 0.10b 2.23a 2.71ab 1.73b 1.24a 1.15a 2.01a 2.76a 2.08a

± ± ± ± ± ± ± ± ± ±

0.13ab 0.11b 0.22a 0.29ab 0.13ab 0.07a 0.11a 0.71a 0.22a 0.16a

TC: total organic C; TN: total N; AP: available P. Values with different letters in the column within N/P input rates are significantly different at p b 0.05 according to the Duncan's test. Values are means of five replicates ± SE.

28

N. Ling et al. / Geoderma 292 (2017) 25–33

Fig. 1. The natural logarithm transformed N or P input rates correlated with the bacterial evenness (Chao) and diversity (Shannon). The red line represents the trend line, and the zone between the blue lines indicates the 95% confidence interval.

(8.1%) at the order level. For the correlations with N rates, most of the genera belonged to Rhizobiales, and almost all of the relative abundances that were negatively impacted by the N rate increased, except for Methylovirgula and Rhizmicrobium. Only the genus Phyllobacterium, within the order Rhizobiales, was negatively impacted by both the N and P input rates. The orders Sphingobacteriales, Xanthomonadales and Bacillales each contained 5 genera that correlated with the N input

rate. In terms of relative abundances, all of the genera that belonged to Sphingobacteriales were negatively associated with the N rate, and all of the genera within Bacillales were positively related with the N rate. Nevertheless, most of the genera that correlated with P rates were in the Orders of Sphingobacteriales and Xanthomonadales (each contained 3 genera), and all of the genera that belonged to Sphingobacteriales were positively correlated with the P rates. Two-

Fig. 2. NMDS of bacterial communities based on OTUs for all samples from the N input experiment (a) and the P input experiment (b), and the correlation between bacterial community similarity (Euclidean distance) and logarithm transformed N input rates (c) or P input rates (d), the red line represents the trend line and the zone between the blue lines indicates the 95% confidence interval.

N. Ling et al. / Geoderma 292 (2017) 25–33

29

Fig. 3. Correlations with significance (p b 0.05) between natural logarithm transformed N input rates and the relative abundances of phyla. The red line represents the trend line, and the zone between the blue lines indicates the 95% confidence interval.

thirds of the genera belonging to Xanthomonadales had negative correlations with the P rates, with respect to their relative abundances. For all of the correlations, 7 genera, including Phenylobacterium, Ohtaekwangia, Lysobacter, Arenimonas, Phyllobacterium, Flavisolibacter and Tumebacillus, were shared in the N and P gradient

rates derived correlations. Intriguingly, the relative abundances of the genus Flavisolibacter were positively correlated with the P rate, but negatively correlated with the N rate, while the same influencing patterns of the N and P input rates were observed in all of the other shared genera.

30

N. Ling et al. / Geoderma 292 (2017) 25–33

Fig. 4. Correlations with significance (p b 0.05) between natural logarithm transformed P input rates and the relative abundances of phyla. The red line represents the trend line, and the zone between the blue lines indicates the 95% confidence interval.

3.5. Pathways determining soil bacterial communities in N and P gradient input experiments Some variables or categories examined in this study were correlated with one another, making this dataset appropriate for SEM analysis. SEM analyses were performed separately for each experiment (Fig. 6). In the N gradient input experiment, the N enrichment directly induced changes in the soil pH, thus impacting the soil bacterial community. This pathway explained 32%, 41% and 50% of the total variance in soil bacterial richness, diversity and structure, respectively. In the P gradient input experiment, P enrichment directly induced changes in soil P availability and pH, thus affecting the soil bacterial community. This pathway explained 17% and 52% of the total variance in soil bacterial richness and structure, respectively. Surprisingly, the effects of soil P availability and pH on the soil bacterial diversity were not significant, and the pH variation attributed to the P input could only impact the community structure but not the richness and diversity. Moreover, the P enrichment

could indirectly affect the community structure via mediating P availability and pH, but P availability could exhibit a much greater influence, according to the path coefficients. 4. Discussion Several studies reported that elevated N input profoundly impacted the soil microbial communities across different terrestrial ecosystems (He et al., 2013; Zhang et al., 2014b; Yang et al., 2015). However, few studies specifically investigated the effects of P addition on soil bacterial communities in a steppe ecosystem. Based on the long-term N and P input experiments in a steppe area of China, we observed the responses of bacterial diversity, composition and community structure to N and P inputs were all dependent on the input rates. The response patterns of some properties in the bacterial community to N addition at the level of 0–28 g·N·m−2 yr−1 and P addition at the level of 0–32 g·P2O5·m−2yr−1 clearly followed a linear pattern, an important feature in a steppe

Fig. 5. The co-occurrence between genera and N or P input rates. The nodes with different colours represent the order they belong to. The linkage marked by red and blue indicate the negative correlations and positive correlations, respectively.

N. Ling et al. / Geoderma 292 (2017) 25–33

31

Fig. 6. Structural equation model (SEM) analysis of the effects of N input and P input on the soil bacterial communities via pathways of major soil properties. For the N input pathway, χ2 = 13.48, p = 0.069 and df = 6, RMSEA = 0.194. For the P input pathway, χ2 = 19.26, p = 0.074, df = 7 and RMSEA = 0.127 (Note: a high p value associated with a χ2 test indicates a good fit of the model to data, i.e., no significant discrepancies). Solid and dashed arrows indicate significant (p b 0.05) and non-significant effects (p N 0.05), respectively. Additionally, r2 values associated with response variables indicate the proportion of variation explained by relationships with other variables. Values associated with solid arrows represent standardized path coefficients.

ecosystem that previous studies only partially revealed that N enrichment causes soil acidification and thus affects belowground communities (Yao et al., 2014; Chen et al., 2015b). Overall, our results intensively showed the apparent differences in the bacterial community responses between N and P gradient inputs in grassland soils, suggesting that both chronic N and P inputs could serve as driving forces influencing the diversification of bacteria. With respect to the bacterial α-diversity, both the bacterial richness and diversity were significantly correlated with the N input rates, which was consistent with previous studies that showed a decline in microbial diversity following N enrichment (Zeng et al., 2016), while only the richness was significantly correlated with the P input rates (Fig. 1). It is generally accepted that the loss of species diversity following N enrichment may threaten the ecosystem stability and affect interactions between above- and below-ground ecosystems (Tilman et al., 2006). Nevertheless, in our study, the bacterial richness was also significantly changed by the P input, even though the impact was not as strong as the N input, as revealed by the linear slope (148.4 from N input N 28.57 from P input). It was already well established that the N enrichment can significantly impact the bacterial community structure (Yao et al., 2014; Zhang et al., 2014b). In this study, we further found that not only N input but also P input could result in significant changes in the bacterial community structure in the surface soil (Fig. 2). The study a the secondary tropical forest of China also found that P addition not only increased the microbial biomass but also altered the soil microbial community composition (Li et al., 2014). Furthermore, we found that different plots with the same high level of N/P input rates frequently supported more dissimilar bacterial communities than different plots with inputs at the same low level of N/P (Fig. 2). This indicates there was little differentiation in the bacterial community structure at low N or P input rates. In contrast, the greater dissimilarity in bacterial communities at high N/P input rates suggest that the greater N/P inputs made the soil bacterial communities less diverse, i.e. inhomogeneous. The inhomogeneity effect on the communities was more severe with N input than with P input, as revealed by the slope of the linear relationship. This relationship was similar to that found in Fierer et al. (2012), who showed that phylogenetic and metagenomic shifts were most pronounced at the highest N input, while an intermediate level of N addition did not lead to a significant shift in the bacterial community. Plant richness in our investigated experiment was linearly decreased with the increase of N rates (Yao et al., 2014), and Schlatter et al. (2015) reported that plant richness was significantly and positively correlated with the similarity of bacterial community suggesting lower plant richness supported more dissimilar bacterial communities. Therefore, it is probably that lower nutrient inputs could result in greater plant

richness (Pallett et al., 2016) and thus generate a homogenizing effect on soil bacterial communities, and vice versa. The shifts in community assembly induced by N and P inputs may also be reflected in the trait differences of the dominant taxa. Our results confirmed the increase in predicted copiotrophic groups (Proteobacteria and Firmicutes) (Goldfarb et al., 2011; Ramirez et al., 2012; Yao et al., 2014) (Ramirez et al., 2012) and a decrease in predicted oligotrophic groups (Acidobacteria, Nitrospirae, Chloroflexi, etc.) (Zeng et al., 2016) with the addition of N. This was in line with the shifts in bacterial composition following N manipulation that were previously explained by the copiotrophic hypothesis (Fierer et al., 2007), in which copiotrophic groups that had fast growth rates were more likely to increase in nutrient-rich conditions, while oligotrophic groups that had slower growth rates would likely decline (Fierer et al., 2012). Copiotrophic taxa thrive in conditions of elevated C availability and exhibit relatively rapid growth rates, while oligotrophic with slower growth rates were able to metabolize nutrient poor and recalcitrant C substrates (Fierer et al., 2007). This community-level copiotroph–oligotroph variations across the N gradients may be a direct result of the increase in N content as copiotrophic taxa to have lower biomass C:N ratios and higher N demands than more oligotrophic taxa (Fontaine and Barot, 2005) or may be an indirect effect resulting from the increase in organic carbon availability associated with the increase in plant productivity as soils with N high input. The P gradients did not follow this rule to manipulate the community, and only Armatimonadetes and Chlorobi relative abundances were significantly and positively correlated with the P gradients. All currently available Armatimonadetes strains have a soil based ecological niche and are aerobic oligotrophs sensitive to nutrient-rich culture media, and its growth within a narrow pH range (Lee et al., 2014) corresponded to the narrow pH range variation that resulted from the P gradient inputs in our experiment. The N and P input rates were both able to affect the relative abundance of Chlorobi, a phylum of anoxygenic and phototrophic bacteria (Gupta and Lorenzini, 2007), but in the opposite manner (Fig. 3 and Fig. 4), indicating Chlorobi responded differently to the N and P inputs, Chlorobi could also live in the steppe ecosystem (Pan et al., 2014), as the inhomogeneous soil matrix still contained local hypoxic conditions. The above patterns of the relative abundance of some bacterial group shifts vs. N/P input rates were also observed at the genus level (Fig. 5). Many more genera were significantly impacted by N input rates rather than the P input rates. The order level can generally evaluate the potential function of bacterial groups, therefore, we found most of the genera impacted by N rates belonged to Rhizobiales, indicating that the N input rates largely and significantly altered the symbiotic N-fixing process with plant roots. Almost all of the relative abundances

32

N. Ling et al. / Geoderma 292 (2017) 25–33

of these Rhizobiales groups decreased with the increase of the N input rate, with the exception of Methylovirgula and Rhizomicrobium. These two genera were not so sensitive to the N input relative to the other genera belonging to Rhizobiales. It is well established that the N-fixing bacterial population can be reduced by high inorganic N input (DeLuca et al., 1996; Skogen et al., 2011), probably because long-term N input would change the plant community composition with the decline of legume population, thus decreasing the symbiotic N2 fixer abundance (Skogen et al., 2011); or because high concentrations of inorganic N was able to inhibit N2 fixation and, ultimately, the presence of these organisms (DeLuca et al., 1996). However, as the number of most Rhizobiales groups sharply decreased due to N input, consequently, the relative abundances of insensitive groups to N input could relatively increase. Interestingly, a large majority of the genus Phyllobacterium, within the order Rhizobiales, are plant-associated bacteria (Mantelin et al., 2006), and these were negatively impacted by both the N and P input rates, suggesting that this N2-fixing genus was not only sensitive to N addition but also to P status. All of the correlations between the N rate and relative abundances of the order Sphingobacteriales were negative, which was aligned with the report that Sphingobacteriales decreased in relative abundance in response to nitrogen addition (Amend et al., 2016). While our results also found that the correlation between P rate and the relative abundance of Sphingobacteriales were positive. These results indicated the opposite response of Sphingobacteriales to N and P inputs, and the genus Flavisolibacter (order Sphingobacteriales) was impacted by both N and P inputs. Similarly, Sapp et al. (2015) found that some members of the Sphingobacteriales were correlated with both N and P fertilizers. The N input rate always displayed a positive effect on Bacillales (Li et al., 2015), which might be because the gram-positive and endospores producing bacteria (e.g., Bacillales) was much more tolerant to environmental changes than others (Gontang et al., 2007), or might be related to the life strategy of these organisms that they have the highest number of ribosomal operons in their DNA (Tourova, 2003) and should have the fastest growth rates when their nutrient demand is covered (Ramirez et al., 2012; Yao et al., 2014). Soil is a strong ecological filter that influences the bacterial community structure and diversity. Chen et al. (2015a) reported that the soil acidification pathway, rather than the N availability pathway, explained most of the effects of long-term N enrichment on soil microorganisms. In addition to pH, changes in nutrient substrates (e.g., available carbon, TN and AP) are known to cause a shift of bacterial communities in soils subjected to short-term fertilization soil properties (Ramirez et al., 2010; Sun et al., 2015). According to the SEM analyses, N and P inputs both acidified the soil and caused changes in the bacterial community structure. However, P inputs resulting in pH variation only impacted the bacterial community structure. The N inputs resulting in a pH change impacted bacterial richness, diversity and community structure, which may be attributed to more severe acidic effects of N inputs on soils than P inputs. Zhao et al. (2015) suggested that soil pH had a greater impact than N fertilization on soil microbial community structure. Our results also showed that the bacterial community structure was much more sensitive to pH than the richness and diversity, which was in agreement with the finding that soil pH significantly affected the bacterial community structure but did not significantly impact the bacterial richness (Zeng et al., 2016). Apart from the pH value, the P inputs mainly depended on improving the P availability, which then affected the bacterial richness and community structure. Addition of P can increased soil microbial biomass (Griffiths et al., 2012; Liu et al., 2013), indicating that P availability is a limiting factor for microbial growth (Liu et al., 2013). Li et al. (2014) pointed out that phosphorus addition did not only increase the microbial biomass but also altered soil microbial community composition. P addition could lead to a decrease of light fraction soil C and total soil C (Liu et al., 2013) and thus regulate the bacterial community, as it is well established that variations in soil microbial community structure are always associated with alteration of soil C

storage and distribution (Ramirez et al., 2012). Moreover, N and P inputs may lead to increased plant growth rates, which would enhance plant C input into soils with the stimulation of soil microbial activities (Li et al., 2014). However, in this field study, it was difficult to determine whether microbial communities' shifts in the composition were a direct result of N/P addition or an indirect result of changes in the plant C inputs to soils. 5. Conclusion Collectively, both the N and P inputs impacted the bacterial community, but the bacterial community was much more sensitive to the N gradient inputs than to the P gradient inputs. The N input indirectly acted on the bacterial community, primarily by pH changing, while the P input indirectly exerted an effect on the community mainly by improving P availability. The genera impacted by N gradient input mostly belonged to the order Rhizobiales, and those affected by P gradient input were within diverse order which might be sensitive to the P availability. This study also implied that caution should be taken during the use of N/ P fertilization because the fertilization could result in bacterial community differentiation and diversity loss, which may reduce the stability of the Inner Mongolia grassland ecosystem. Further research requires the integration of focused experiments to better understand the impacts of N/P inputs on grassland ecosystem by linking changes in plant traits to the microbe-mediated processes. Acknowledgements This study was supported by the China Postdoctoral Science Foundation (2015M570460 and 2016T90473). Many graduate students and staffs involved in maintaining the long-term field experiments and collecting soil samples but not listed as coauthors are grateful. Reference Amend, A.S., Martiny, A.C., Allison, S.D., Berlemont, R., Goulden, M.L., Lu, Y., Treseder, K.K., Weihe, C., Martiny, J.B.H., 2016. Microbial response to simulated global change is phylogenetically conserved and linked with functional potential. ISME J. 10, 109–118. Bai, Y., Han, X., Wu, J., Chen, Z., Li, L., 2004. Ecosystem stability and compensatory effects in the Inner Mongolia grassland. Nature 431, 181–184. Bai, Y.F., Wu, J.G., Clark, C.M., Naeem, S., Pan, Q.M., Huang, J.H., Zhang, L.X., Han, X.G., 2010. Tradeoffs and thresholds in the effects of nitrogen addition on biodiversity and ecosystem functioning: evidence from Inner Mongolia grasslands. Glob. Chang. Biol. 16, 358–372. Bastian, M., Heymann, S., Jacomy, M., 2009. Gephi: an open source software for exploring and manipulating networks. Third International AAAI Conference on Weblogs and Social Media. AAAI Publications, Sn Jose, CA, pp. 361–362. Beauregard, M.S., Hamel, C., Atul-Nayyar, St-Arnaud, M., 2009. Long-term phosphorus fertilization impacts soil fungal and bacterial diversity but not AM fungal community in Alfalfa. Microb. Ecol. 59, 379–389. Caporaso, J.G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F.D., Costello, E.K., Fierer, N., Pena, A.G., Goodrich, J.K., Gordon, J.I., Huttley, G.A., Kelley, S.T., Knights, D., Koenig, J.E., Ley, R.E., Lozupone, C.A., McDonald, D., Muegge, B.D., Pirrung, M., Reeder, J., Sevinsky, J.R., Tumbaugh, P.J., Walters, W.A., Widmann, J., Yatsunenko, T., Zaneveld, J., Knight, R., 2010. QIIME allows analysis of high-throughput community sequencing data. Nat. Methods 7, 335–336. Chen, D., Li, J., Lan, Z., Hu, S., Bai, Y., 2015a. Soil acidification exerts a greater control on soil respiration than soil nitrogen availability in grasslands subjected to long-term nitrogen enrichment. Funct. Ecol. 30, 658–669. Chen, D., Zheng, S., Shan, Y., Taube, F., Bai, Y., 2013. Vertebrate herbivore-induced changes in plants and soils: linkages to ecosystem functioning in a semi-arid steppe. Funct. Ecol. 27, 273–281. Chen, D.M., Lan, Z.C., Hu, S.J., Bai, Y.F., 2015b. Effects of nitrogen enrichment on belowground communities in grassland: relative role of soil nitrogen availability vs. soil acidification. Soil Biol. Biochem. 89, 99–108. DeLuca, T.H., Drinkwater, L.E., Wiefling, B.A., DeNicola, D.M., 1996. Free-living nitrogenfixing bacteria in temperate cropping systems: influence of nitrogen source. Biol. Fertil. Soils 23, 140–144. Edgar, R.C., 2013. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat. Methods 10, 996. Elser, J.J., Bracken, M.E., Cleland, E.E., Gruner, D.S., Harpole, W.S., Hillebrand, H., Ngai, J.T., Seabloom, E.W., Shurin, J.B., Smith, J.E., 2007. Global analysis of nitrogen and phosphorus limitation of primary producers in freshwater, marine and terrestrial ecosystems. Ecol. Lett. 10, 1135–1142. Fierer, N., Bradford, M.A., Jackson, R.B., 2007. Toward an ecological classification of soil bacteria. Ecology 88, 1354–1364.

N. Ling et al. / Geoderma 292 (2017) 25–33 Fierer, N., Lauber, C.L., Ramirez, K.S., Zaneveld, J., Bradford, M.A., Knight, R., 2012. Comparative metagenomic, phylogenetic and physiological analyses of soil microbial communities across nitrogen gradients. ISME J. 6, 1007–1017. Fontaine, S., Barot, S., 2005. Size and functional diversity of microbe populations control plant persistence and long-term soil carbon accumulation. Ecol. Lett. 8, 1075–1087. Ford, H., Roberts, A., Jones, L., 2016. Nitrogen and phosphorus co-limitation and grazing moderate nitrogen impacts on plant growth and nutrient cycling in sand dune grassland. Sci. Total Environ. 542, 203–209. Galloway, J.N., Townsend, A.R., Erisman, J.W., Bekunda, M., Cai, Z.C., Freney, J.R., Martinelli, L.A., Seitzinger, S.P., Sutton, M.A., 2008. Transformation of the nitrogen cycle: recent trends, questions, and potential solutions. Science 320, 889–892. Goldfarb, K.C., Karaoz, U., Hanson, C.A., Santee, C.A., Bradford, M.A., Treseder, K.K., Wallenstein, M.D., Brodie, E.L., 2011. Differential growth responses of soil bacterial taxa to carbon substrates of varying chemical recalcitrance. Front. Microbiol. 2. Gontang, E.A., Fenical, W., Jensen, P.R., 2007. Phylogenetic diversity of gram-positive bacteria cultured from marine sediments. Appl. Environ. Microbiol. 73, 3272–3282. Griffiths, B.S., Spilles, A., Bonkowski, M., 2012. C:N:P stoichiometry and nutrient limitation of the soil microbial biomass in a grazed grassland site under experimental P limitation or excess. Ecol. Process. 1, 1–11. Gupta, R.S., Lorenzini, E., 2007. Phylogeny and molecular signatures (conserved proteins and indels) that are specific for the Bacteroidetes and Chlorobi species. BMC Evol. Biol. 7, 71. He, Y.T., Qi, Y.C., Dong, Y.S., Xiao, S.S., Peng, Q., Liu, X.C., Sun, L.J., 2013. Effects of nitrogen fertilization on soil microbial biomass and community functional diversity in temperate grassland in Inner Mongolia, China. Clean-Soil Air Water 41, 1216–1221. Lee, K.C.Y., Morgan, X.C., Dunfield, P.F., Tamas, I., McDonald, I.R., Stott, M.B., 2014. Genomic analysis of chthonomonas calidirosea, the first sequenced isolate of the phylum Armatimonadetes. ISME J. 8, 1522–1533. Li, C.-H., Tang, L.-S., Jia, Z.-J., Li, Y., 2015. Profile changes in the soil microbial community when desert becomes oasis. PLoS One 10, e0139626. Li, J., Li, Z., Wang, F., Zou, B., Chen, Y., Zhao, J., Mo, Q., Li, Y., Li, X., Xia, H., 2014. Effects of nitrogen and phosphorus addition on soil microbial community in a secondary tropical forest of China. Biol. Fertil. Soils 51, 207–215. Li, L.J., Zeng, D.H., Yu, Z.A., Fan, Z.P., Mao, R., 2010. Soil microbial properties under N and P additions in a semi-arid, sandy grassland. Biol. Fertil. Soils 46, 653–658. Liu, L., Zhang, T., Gilliam, F.S., Gundersen, P., Zhang, W., Chen, H., Mo, J.M., 2013. Interactive effects of nitrogen and phosphorus on soil microbial communities in a tropical forest. PLoS One 8. Mantelin, S., Saux, M.F.-L., Zakhia, F., Béna, G., Bonneau, S., Jeder, H., de Lajudie, P., CleyetMarel, J.-C., 2006. Emended description of the genus Phyllobacterium and description of four novel species associated with plant roots: phyllobacterium bourgognense sp. nov., phyllobacterium ifriqiyense sp. nov., phyllobacterium leguminum sp. nov. and phyllobacterium brassicacearum sp. nov. Int. J. Syst. Evol. Microbiol. 56, 827–839. Pallett, D.W., Pescott, O.L., Schäfer, S.M., 2016. Changes in plant species richness and productivity in response to decreased nitrogen inputs in grassland in southern England. Ecol. Indic. 68, 73–81. Pan, Y., Cassman, N., de Hollander, M., Mendes, L.W., Korevaar, H., Geerts, R.H.E.M., van Veen, J.A., Kuramae, E.E., 2014. Impact of long-term N, P, K, and NPK fertilization on the composition and potential functions of the bacterial community in grassland soil. FEMS Microbiol. Ecol. 90, 195–205. Ramirez, K.S., Craine, J.M., Fierer, N., 2012. Consistent effects of nitrogen amendments on soil microbial communities and processes across biomes. Glob. Chang. Biol. 18, 1918–1927. Ramirez, K.S., Lauber, C.L., Knight, R., Bradford, M.A., Fierer, N., 2010. Consistent effects of nitrogen fertilization on soil bacterial communities in contrasting systems. Ecology 91, 3463–3470 (discussion 3503-3414).

33

Schlatter, D.C., Bakker, M.G., Bradeen, J.M., Kinkel, L.L., 2015. Plant community richness and microbial interactions structure bacterial communities in soil. Ecology 96, 134–142. Shi, Y.C., Lalande, R., Hamel, C., Ziadi, N., Gagnon, B., Hu, Z.Y., 2013. Seasonal variation of microbial biomass, activity, and community structure in soil under different tillage and phosphorus management practices. Biol. Fertil. Soils 49, 803–818. Skogen, K.A., Holsinger, K.E., Cardon, Z.G., 2011. Nitrogen deposition, competition and the decline of a regionally threatened legume, Desmodium cuspidatum. Oecologia 165, 261–269. Sapp, M., Harrison, M., Hany, U., Charlton, A., Thwaites, R., 2015. Comparing the effect of digestate and chemical fertiliser on soil bacteria. Appl. Soil Ecol. 86, 1–9. Sun, R., Zhang, X.-X., Guo, X., Wang, D., Chu, H., 2015. Bacterial diversity in soils subjected to long-term chemical fertilization can be more stably maintained with the addition of livestock manure than wheat straw. Soil Biol. Biochem. 88, 9–18. Tan, H., Barret, M., Mooij, M.J., Rice, O., Morrissey, J.P., Dobson, A., Griffiths, B., O'Gara, F., 2013. Long-term phosphorus fertilisation increased the diversity of the total bacterial community and the phoD phosphorus mineraliser group in pasture soils. Biol. Fertil. Soils 49, 661–672. Tilman, D., Reich, P.B., Knops, J.M., 2006. Biodiversity and ecosystem stability in a decadelong grassland experiment. Nature 441, 629–632. Tourova, T.P., 2003. Copy number of ribosomal operons in prokaryotes and its effect on phylogenetic analyses. Microbiology 72, 389–402. van der Heijden, M.G.A., Bardgett, R.D., van Straalen, N.M., 2008. The unseen majority: soil microbes as drivers of plant diversity and productivity in terrestrial ecosystems. Ecol. Lett. 11, 296–310. Wang, Q., Garrity, G.M., Tiedje, J.M., Cole, J.R., 2007. Naive Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl. Environ. Microbiol. 73, 5261–5267. Xu, H., Su, H., Su, B.Y., Han, X.G., Biswas, D.K., Li, Y.G., 2014. Restoring the degraded grassland and improving sustainability of grassland ecosystem through chicken farming: a case study in northern China. Agric. Ecosyst. Environ. 186, 115–123. Yang, A., Liu, N.N., Tian, Q.Y., Bai, W.M., Williams, M., Wang, Q.B., Li, L.H., Zhang, W.H., 2015. Rhizosphere bacterial communities of dominant steppe plants shift in response to a gradient of simulated nitrogen deposition. Front. Microbiol. 6. Yao, M.J., Rui, J.P., Li, J.B., Dai, Y.M., Bai, Y.F., Hedenec, P., Wang, J.M., Zhang, S.H., Pei, K.Q., Liu, C., Wang, Y.F., He, Z.L., Frouz, J., Li, X.Z., 2014. Rate-specific responses of prokaryotic diversity and structure to nitrogen deposition in the Leymus chinensis steppe. Soil Biol. Biochem. 79, 81–90. Zhang, X., Wei, H., Chen, Q., Han, X., 2014a. The counteractive effects of nitrogen addition and watering on soil bacterial communities in a steppe ecosystem. Soil Biol. Biochem. 72, 26–34. Zhang, X.M., Wei, H.W., Chen, Q.S., Han, X.G., 2014b. The counteractive effects of nitrogen addition and watering on soil bacterial communities in a steppe ecosystem. Soil Biol. Biochem. 72, 26–34. Zhao, C., Fu, S., Mathew, R.P., Lawrence, K., Feng, Y., 2015. Soil microbial community structure and activity in a 100-year-old fertilization and crop rotation experiment. J. Plant Ecol. 8, 623–632. Zeng, J., Liu, X., Song, L., Lin, X., Zhang, H., Shen, C., Chu, H., 2016. Nitrogen fertilization directly affects soil bacterial diversity and indirectly affects bacterial community composition. Soil Biol. Biochem. 92, 41–49. Zhou, X., Fornara, D., Wasson, E.A., Wang, D.M., Ren, G.D., Christie, P., Jia, Z.J., 2015. Effects of 44 years of chronic nitrogen fertilization on the soil nitrifying community of permanent grassland. Soil Biol. Biochem. 91, 76–83.

Suggest Documents