1. Introduction
Rice feeds more than half of the world’s population and is an important staple food. In China, rice cultivation covers more than 25% of arable land and comprises 40% of national food production [
1]. However, excessive fertilization with (or overdependence on) mineral fertilizers is popular in China. As a result, the soil pH in central and southern Chinese paddy fields has decreased by 0.29 and 0.58, respectively, over the last 39 years. Furthermore, excessive nitrogen input and decreases in soil organic matter contributed 32.77% and 29.95% to paddy soil acidification, respectively [
2]. The excessive application of N fertilizers decreases N-use efficiency, pollutes ground water [
3,
4], and does not increase crop yield [
5]; however, reducing fertilization rates and partially substituting chemical N with manure can increase N-use efficiency, markedly reduce N losses [
6,
7], and enhance crop yield in upland [
8] and paddy soil [
1,
9]. Therefore, sustainable fertilizer regimes are needed in intensive cultivation systems to maintain high crop yields while ensuring sustainable soil health.
Fertilizer regimes, especially under long-term conditions, can profoundly affect bacterial communities in upland [
1,
10] and paddy soils [
1,
9,
11,
12]. Mineral fertilizers considerably decrease bacterial alpha diversity, whereas combining mineral fertilizers with manure restores bacterial alpha diversity to levels comparable to non-fertilization controls [
1,
10]. Bacterial community structure (beta diversity) is also affected by fertilizer regimes. For example, manure application considerably alters beta diversity compared to mineral fertilizer [
13,
14]. Furthermore, continuous applications of manure stimulate the growth of beneficial bacteria [
9]. However, applying large amounts of organic fertilizer (compost, sewage sludge, and cattle manure) did not cause major changes in alpha and beta diversity because the soil microbiota is very robust [
15]. Different fertilization regimes (N, NP, NPK, compost, and NPK+ compost) over a 45-year period also had no considerable effect on bacterial community structure [
11]. These contradictory data suggest that further studies are needed to explain how fertilizer regimes affect the soil microbial community, allowing us to ultimately predict microbe-mediated nutrient dynamics in agricultural systems.
Aside from manure and mineral fertilizer, controlled-release N [
16,
17], silicon [
18], and zinc [
19] also have the capabilities to considerably affect microbial biomass [
20], bacterial alpha diversity [
21,
22], and community structures [
22,
23]. However, little is known about how combined applications of organic fertilizer controlled-release N and micronutrient fertilizers affect rice yield, soil properties, and bacterial structures in paddy soil. Therefore, we studied changes in soil fertility, crop yield, and bacterial community structure after 4-year fertilizer treatments. Based on previous studies in paddy soils, we hypothesized that reducing mineral N applications and partially substituting chemical N with manure and controlled-release N will have positive effects on soil properties, bacterial alpha diversity, and bacterial community structures in the paddy soil.
2. Materials and Methods
2.1. Study Site
The study site is located in Gao’an City (28°15′26″ N, 115°7′33″ E), Jiangxi Province, south China. This region has a temperate continental monsoon climate, with an average annual temperature of 17.2 °C and precipitation of 1680 mm. The soil is red paddy soil (Eutric Cambisol, FAO), and its basic properties are as follows: pH, 5.28; soil organic matter, 23.62 g kg−1; total N, 1.69 g kg−1; available nitrogen, phosphate, and potassium—184.31, 44.54, and 178.33 mg kg−1, respectively.
2.2. Field Experiment and Sample Collection
The field experiment was carried out over four years (2012–2016) using four fertilizer treatments: (1) FP: local farmers’ practice with 100% urea N; (2) T1: 64% urea N + 16% manure N; (3) T2: T1 + Si, Zn, and S fertilizers; (4) T3: 40% urea N + 24% controlled-release urea + 16% manure N + Si, Zn, and S fertilizers. The FP nitrogen fertilizer application rates were 165 kg N ha
−1 and 195kg N ha
−1 for early and late rice, respectively (
Table 1). The experiment was carried out by randomized block design with four replications. Each plot covered a 6 × 7 m
2 area. The rice varieties were Zhongjia 17 and Wufengyou T025 for early and late rice, respectively. The early rice was sown in late March and harvested in middle July, and the late rice were transplanted in early August and harvested in late October. Rice straw was returned to the field after being sliced into 3–5 cm pieces. Soil samples were collected per season after the rice was harvested. Five cores were collected in each plot to form a compensate sample (0–20 cm). The samples were transported to the laboratory in an ice box. Each soil sample was separated into two parts after the fresh soil was passed through a 2 mm mesh sieve. One part was air dried for chemical analysis; the second was kept at −20 °C for microbial biomass analysis. Only the last soil samples were separated into three parts, and the third part of each sample was kept at −80 °C for DNA extraction. Finally, rice yield and above-ground biomass per season were measured.
2.3. Soil Chemical Properties
Soil pH was determined by a pH meter (weight: volume ratio 1:5). Soil organic matter and total N were determined by dry combustion ((Elementar, Elementar UK Ltd., South Manchester, UK). Soil microbial biomass was determined by the fumigation extraction method. Soil available phosphate was extracted with an 0.5 mol L−1 sodium bicarbonate buffer (pH 8.5) and determined by the Mo-Sb colorimetric method. Soil available potassium was extracted with 1 mol L−1 ammonium acetate buffer (pH 7.0) and determined by flame photometry. Soil available Si was extracted with an acetate-acetic buffer (pH 4.0) and determined by ICP-OES (Optima 7300 DV, Perkin Elementa, Boston, MA, USA).
2.4. Soil DNA Extraction
Total bacterial DNA was extracted from 0.25 g fresh soil using a PowerSoil® DNA Isolation Kit (Mo Bio Laboratories Inc., Carlsbad, CA, USA) following the manufacturer’s protocol. The DNA quality was determined by a NanoDrop ND-2000 spectrophotometer (NanoDrop, ND 2000, Thermo Scientific, Wilmington, DE, USA) based on absorbance at A260/A280.
2.4.1. 16S rRNA Gene Amplification and Sequencing
High throughput sequencing of the 16S rRNA gene was used to analyze the bacterial composition. The V4 region of the 16S rRNA gene was amplified by the PCR method (95 °C for 3 min, followed by 27 cycles at 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s, with a final extension at 72 °C for 10 min), using primers 338F (5′ACTCCTACGGG A GGCAGCAG-3′) and 806R (5′GGACTACHVGGGTWTCTAAT-3′). PCR was performed in triplicate using 20-μL mixtures containing 4 μL of 5 X Fast Pfu Buffer, 2 μL of 2.5 mM dNTPs, 0.8 μL of each primer (5 μM), 0.4 μL of Fast Pfu DNA Polymerase, and 10 ng of template DNA. The amplicons were then separated and extracted from 2% agarose gels, purified with an AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, USA) following the manufacturer’s instructions, and quantified by the QuantiFluor™ dsDNA system (Promega, USA). The purified amplicons were pooled in equimolar amounts and paired-end sequenced (2 × 250) on an Illumina MiSeq Platform, using the standard protocol. The raw reads were deposited in the National Center for Biotechnology Information-Sequence Read Archive database (accession number: SRP093212).
2.4.2. Bioinformation
Raw FASTQ files were demultiplexed and quality filtered by QIIME (version 1.17). Briefly, the 300 bp reads were truncated at any site with an average quality score < 20 over a 50 bp sliding window. Any truncated reads shorter than 50 bp were discarded. Barcode matching was strict, with only two-nucleotide mismatches allowed during primer matching. The reads containing ambiguous characters were removed. Only sequences that overlapped by more than 10 bp were assembled according to their overlapping sequences. Reads that could not be assembled were discarded. The operational taxonomic units (OTUs) were clustered with a similarity cut-off of 97% using UPARSE (version 7.1
http://drive5.com/uparse/, accessed on 14 July 2017). Chimeric sequences were identified and removed by UCHI-ME. The taxonomy of each 16S rRNA gene sequence was analyzed using an RDP Classifier (
http://rdp.cme.msu.edu/, accessed on 14 July 2017) against the silva (SSU115) 16S rRNA database, with a confidence threshold of 70%.
2.5. Statistical Analysis
The differences of the soil properties among the treatments and the variations of rice yield between treatments and years were tested by one-way and two-way ANVOA, respectively, using SigmaPlot12.5 (Systat Software, Inc. California, USA). For the samples from sequencing, each sample was rarefied to 84,208 and 111,158 reads for both alpha and beta diversity, respectively. All changes in the soil microbial communities were evaluated based on the OTU matrix. Mothur software was used to estimate bacterial α-diversity indices (Shannon’s diversity index, Simpson’s richness and diversity index, and the Chao1 estimator of richness). Principal coordinates analysis (PCoA) with CANOCO 5.0 (Microcomputer Power, NY, USA) was used to visualize the bacterial structure among the samples based on the Bray–Curtis distance. Additionally, the significance of Bray–Curtis dissimilarity was tested by PERMANOVA (1000 permutations). Finally, the relationship between soil properties and bacterial community structure was analyzed with redundancy analysis (RDA).