Variation of soil nutrients and bacterial community diversity of different land utilization types in Yangtze River Basin, Chongqing Municipality

Study site and soil sampling

The selected study sites belong to the Yangtze River Basin of Chongqing Municipality, with a mild climate and a subtropical monsoon humid climate. The annual mean temperature in the region is 16−18 °C and the normal year annual rainfall is 1,000–1,350 mm, more than 70% of which occurs from May to September. The types of landforms in the Yangtze River Basin are complex and diverse. The main soil types are purple soily, ellow loam, paddy soil, red soil, new soil, and mountain meadow soil. The area is part of the key environmental protection areas of the Yangtze River Basin because of the complex natural landforms, abundant precipitation, and erosive soils.

Based on, the characteristics of the Yangtze River Basin and the principle of vegetation division, the land types in the whole basin are divided into pure forest, mixed forest, grass, and bare land. The soil samples were collected randomly from different land-use types in the study area (N29°27^′24.54^″, E106°31^′35.55^″) in September 2018. Soil type in the research area is “red soil”. A total of 12 soil samples (three random samples per environment) were taken at four locations from pure forest (P1, Subtropical evergreen broad-leaved forest), mixed forest (M2, Subtropical mountain evergreen and deciduous broad-leaved mixed forest), grassland (G3), and bare land (B4). In each plot, the mixed soil samples (removing the surface litter layer) were collected from a depth of 0–20 cm by five-point sampling method and placed in individual sterile brown paper bags and stored at 4 °C for further studies. The fresh soil samples were divided into two parts, one part was passed through a 40 mesh sieve to determine the soil culturable bacterial number and diversity, while the other was naturally air dried to determine the soil nutrient.

Soil properties

The physical and chemical characteristics of soils, including pH, moisture content, available nitrogen (AN), available phosphorus (AP) and available potassium (AK) were determined as described. Briefly, soil water content was estimated by drying the sample at 105 °C for 24 h (Lu, 2000). Soil pH was measured in 1:2.5 soil water suspensions (Lu, 2000). Available nitrogen was determined using the alkaline diffusion method (Page, Miller & Keeney, 1982). Soil available phosphorus was extracted with 0.5 mol/L NaHCO₃ at a pH of 8.5, and P content was estimated colorimetrically by the molybdate method (Olsen, Cole & Watanable, 1954). Soil available potasssium was determined by flame spectrometry method (Lu, 2000). Soil enzyme activities were analyzed using standard methods (Chen et al., 2015). Catalase enzyme activity was determined by chemical titration. Soil urease activity was determined by estimating the amount of ammonia and carbonic acid produced during urea hydrolysis. Soil sucrase activity was determined using methods of 3,5-dinitrosalicylic acid colorimetry method.

Determination of the viable count of culturable bacteria

The culturable bacterial counts were determined by plate counting method (Li et al., 2017a; Li et al., 2017b; Li et al., 2017c). Briefly, 10 g of soil sample was suspended in 90 mL of sterile distilled water and shaken for 30 min on a Thermostat oscillator. 0.1 mL of serial dilutions of the soil sample solutions were spread on Luria-Bertani (LB) agar plates and incubated at 30 °C for 24 h. Bacterial count was expressed as colony forming units per fresh weight (cfu/g) of soil. All experiments were carried out in triplicates.

Metagenomic DNA isolation and community 16S rDNA gene amplification

DNA from all soil samples was extracted and purified using the E. Z. N. A.™ Mag-Bind Soil DNA Kit (D5625-01, Omega, USA) using 0.2 g of soil as per the manufacturer’s instructions and stored at −80 °C until further analysis. For bacterial 16S rRNA, PCR amplification was performed using the primers 16S 515F (5′-GTGCCAGCMGCCGCGGTAA-3′) and 16S 806R (5′-GGACTACHVGGGTWTCTAAT-3′) with genomic DNA as template. The PCR reaction conditions were as follows: initial denaturation at 94 °C for 2 min; denaturation at 94 °C for 30 s, annealing at 55 °C for 20 s, extension at 72 °C for 60 s, for 32 cycles; and a final extension at 72 °C for 10 min. PCR products were stored at −20 °C and analyzedon 1% gel using DL2000 marker for quality examination.

Metagenomic sequencing

High-throughput sequencing was performed using the Illumina MiSeq platform at the Sangon Biotech Engineering Technology & Services Co., Ltd., Shanghai, China. The original sequence of MiSeq sequencing contains the barcode sequence primers and linker sequences. In order to ensure that the results of information analysis are more accurate and reliable, the raw reads were filtered using Cutadapt (V1.9.1, http://cutadapt.readthedocs.io/en/stable/) (Louca, Parfrey & Doebeli, 2016), and then compared with the species annotation database (http://github.com/torognes/vsearch/) (Martin, 2011) to delete the chimeric sequences and obtain clear reads (Rognes et al., 2016). Clean reads from all samples were clustered into OTUs (Operational Taxonomic Units) with 97% identity using Uparse software (Uparse v7.0.1001, http://www.drive5.com/uparse/) (Haas et al., 2011). The OTU sequences were annotated, using Mothur method and SSUrRNA database for species annotation analysis to obtain taxonomic information at each classification level (Edgar, 2013; Wang et al., 2007). After obtaining the sequencing results and calculation of operational taxonomic units (OTUs) matrix, statistical analysis was performed using alpha indices (Shannon, Simpson, Chao 1 and ACE), heatmap of genera, principal coordinate analysis (PCoA) and UPGMA. The alpha diversity index and UPGMA were calculated by using the Qiime software (Version 1.9.1) and Principal coordinate analysis (PCoA) was performed using the R package phyloseq.

Statistical analyses

Statistical analysis of data was carried out using Origin 8.0 software (OriginLab Corp., USA). ANOVA with least-significant-difference (LSD) tests, cluster analysis, correlation analysis and principal component analysis were carried out using the SPSS 19.0 software (ver. 19.0; SPSS Inc., USA). For all analyses, the results were considered to be significant at p < 0.05.

Soil physical and chemical properties

The basic physicochemical characteristics of soil samples from different land-use types are shown in Table 1. The soil pH ranged from 6.74–7.42, and with the M2 sample having the lowest pH. The moisture content ranged from 12.89% in sample B4 to 18.25% in sample M2 (p < 0.05). Significant differences in soil nutrients were observed in different land-use types. The soil available P (47.00 mg/kg), available K (112.61 mg/kg), available N (94.83 mg/kg) contents in M2 were significantly higher than those in P1, G3 and B4 soils. The soil catalase and sucrase activities were similar in the four samples, but the soil urease activity varied greatly, ranging from 0.7816 to 3.0474 mg/g (p < 0.05). The M2 soil sample had higher urease activity when compared with P1, G3 and B4.

Number of soil cultivable bacteria in different land-use types

The number of culturable soil bacteria varied among different land-use types (Table 2). The culturable bacterial counts in P1, M2, G3 and B4 were 59. 3 ×10⁵, 85. 3 ×10⁵, 51. 3 ×10⁵ and 10. 0 ×10⁵ cfu/g respectively. The highest number of culturable bacteria was detected in M2 sample, while the lowest was found in B4.

Abundance and diversity of members of the bacterial microbiota

Sequences with 97% or more similarity were classified as an operational classification unit (operational taxonomic units, OTU). As shown in Fig. 1, the sequences were grouped into 3861 OTUs, which reduced to 1837 after removing singletons, there was a similar trend in the P1, M2, G3 and B4 samples. Bacterial OTUs were 2679, 2814, 3008 and 3086 in P1, M2, G3 and B4 respectively.

The Good’ s coverage of P1, M2, G3 and B4 samples reached 99.1%, 98.7%, 99.1% and 99.0% respectively, that captured the majority of microbial diversity (Table 3). The Chao1/ACE index reflects the species richness information of the samples and the Shannon/Simpson index reflects the species diversity of microbes in a sample. The Chao1 and ACE scores ranged from 1867.768 to 2819.313 and from 1900.661 to 2542.059, respectively. The Shannon and Simpson scores ranged from 8.496 to 9.696 and 0.974 to 0.997, respectively (Table 3). The results showed that P1 had the largest Shannon index and the smallest Simpson index. One the other hand, M2 had maximum values for both the ACE index and the Chao1 index.

Comparison of soil bacterial communities in different land-use types

Nine bacterial phyla, Firmicutes, Proteobacteria, Actinobacteria, Bacteroidetes, Acidobacteria, Gemmatimonadetes, Chloroflexi, Fusobacteria and Tenericutes were identified in different land-use types analyzed (Fig. 2A). Among these Proteobacteria and Actinobacteria were the predominant phyla in P1, M2, G3 and B4 (Fig. 3). The vast majority of bacteria were classified as Proteobacteria and Actinobacteria in P1, and a similar pattern was observed in G3 and B4. The Firmicutes phylum was found to be abundantly distributed in P1 and M2, but was less abundant in G3 and B4 samples. Tenericutes were less abundant in P1 samples but not in M2, G3 and B4 samples. Heatmaps revealed that the bacterial community of P1 sample was more diverse than those of other samples (Fig. 3).

At the genus level, the diversity of the bacterial microbiota in P1, M2, G3 and B4 samples was higher, especially in the M2 samples (Figs. 2B, 3). Twenty-four bacterial genera were detected with higher relative abundance in P1, M2, G3 and B4. Dongia, Aeromonas, Candidatus_Bacilloplasma, Cetobacterium, Lysobacter and Agromyces were the predominant genera in P1 samples. In M2 samples, majority of the bacterial community belonged to eight genera, Romboutsia, Blautia, Fusicatenibacter, Colwellia, Faecalibacterium, Agathobacter, Reyranella and Bifidobacterium. Haliangium and Gaiella were the predominant genera in G3 samples, while Acidibacter and Polycyclovorans were the predominant genera in B4 samples. Other major genera were Bryobacter, Streptococcus, Streptomyces, Marmoricola, Pseudomonas, lamia, Sphingomonas and Bacteroides. Dongia was less abundant in M2, G3 and B4, and was found to be higher in P1 samples. Faecalibacterium and Agathobacter were highly abundant in M2 samples but were rarely detected in P1, G3 and B4 samples. The bacterial microbial community thus varied greatly, and the abundance of each genus varied among soil samples from different land-use types.

PCoA and UPGMA analyses were performed to evaluate similarities in the bacterial communities of P1, M2, G3 and B4. One weighted (PC1 = 71.44%, PC2 = 16.04%) and another weighted (PC1 = 32.67%, PC2 = 20.18%) PCoA of bacterial communities was performed to analyze the community structure similarity (Fig. 4). The UPGMA clustering tree was built by using unweighted group averaging method (Fig. 5). The branch length of the samples reflected their similarity. The results showed that there was higher similarity among the community structures of all soils, although some differences existed in the study region of different land-use types.

Correlations between microbial diversity and environmental factors

To further estimate the effects of environmental factors on microbial quantity and diversity, soil environmental factors such as moisture content, pH, available P, available N, available K, catalase, urease and sucrase in the samples were analyzed by correlation analysis using SPSS 19.0. These analyses demonstrated that environmental factors were strongly correlated with the microbial quantity and diversity (p < 0.01). The canonical correlation of the eight soil properties with the culturable bacterial numbers and α-diversity of the P1 soil are listed in Table 4. The bacterial numbers were significantly positively correlated with catalase and sucrase (P < 0.05), with correlation coefficients of 0.823 and 0.993 respectively. The Shannon index was negatively correlated (−0.806) and Simpson index was positively correlated (0.992) with pH. The Chao1 index and ACE index were significantly positively correlated with catalase, with correlation coefficients of 1.000 and 0.999 respectively (P < 0.01). The data in Table 5 indicate that moisture content (−0,926), catalase (0.957) and urease (0.851) had significant effects on microbial quantity. And available K (−1.000) and available N (−0.999) were negatively correlated with bacterial diversity and abundance in the M2 sample. Available K and moisture content are dominant factors that affect the number of culturable bacteria, and the α-diversity index was primarily positively affected by available K and urease in the G3 sample (Table 6). The culturable bacterial quantity and α-diversity index showed a significantly negative correlation with sucrase (−0.971) and available P (−0.953) in the B4 sample (Table 7).