Impact of litter quantity on the soil bacteria community during the decomposition of Quercus wutaishanica litter

Site description

Soil and litter samples were collected from the Fuxian Observatory for Soil Erosion and Eco-environment, a secondary forest region. Quercus wutaishanica was the predominant community, playing an important role in maintaining the stability of the system in this area (Fan, Wang & Guo, 2006; Guo et al., 2010). Therefore, understanding the effects of Quercus wutaishanica leaf litter decomposition provides insights into the carbon and nitrogen recycling in the soil-plant system. We established three plots in Quercus wutaishanica forests with similar topographical conditions to investigate the annual litter fall using the method described by Ukonmaanaho & Starr (2001). Over two years of observations, the annual litter fall of Quercus wutaishanica was approximately 200 g/m²/yr.

Soil and litter sampling

Soil samples from 0–20 cm were obtained in September 2015 when most of the leaves had fallen. All roots, stones, small animals and other debris were removed from the soil samples by hand, and the samples were sieved through a 2 mm screen. The mixed soils were used to conduct the litter decomposition experiment in the laboratory. The soil organic carbon and total nitrogen contents were 18.26 g/kg and 1.60 g/kg, respectively. Fresh litter was collected with a litter collector. To avoid damaging the litter structure, the leaves were air-dried for more than two weeks at room temperature to a consistent weight.

Litter decomposition experiment

Litter decomposition experiment was conducted using the nylon mesh bag technique. There were three treatments, including normal quantity (200 g/(m²/yr)) litter, double quantity (400 g/(m²/yr)) litter, and control (no litter) (Fig. 1). The litter bags (10 cm × 20 cm size) were constructed out of 1 mm nylon mesh. First, we placed 200 g fresh soils in a 1 L plastic basin and then placed a litter bag (5 g, normal quantity; 10 g, double quantity) on the surface. Each treatment had three replicates. We also conducted a control experiment without litter bags. All basins were incubated at 25 °C in an incubator. The soil water content was adjusted using the weighting method every week at a relative humidity of 20%. After 90 days, we collected the soil sample layer below the litter bags to analyze the soil properties and bacterial communities. After harvest, each soil sample was mixed and separated into two parts. One part was air-dried for the evaluation of the soil properties. The other part was frozen at −80 °C (using liquid nitrogen) for subsequent sequencing analysis.

Analysis of the soil properties

The soil moisture was determined gravimetrically with fresh soils at 105 °C for 24 h, and the water content was expressed as a percentage of the dry weight. The fumigation-extraction method was used to determine microbial biomass carbon (MBC) and microbial nitrogen (MBN) (Vance, Brookes & Jenkinson, 1987). The dissolved carbon (DOC) and dissolved nitrogen (DON) in the soil were determined by extracting the samples in 0.5 mol/L K₂SO₄. The soil total N (STN), soil organic carbon (SOC), soil nitrate nitrogen (NO${}_3^{-}$-N) and soil ammonia nitrogen (NH${}_4^{+}$-N) were analyzed using the method described by Zeng et al. (2016).

Soil NDA extraction and PCR amplification

The DNA of the soil was extracted from a 0.5 g soil sample using the CTAB method. The concentration and purity of the DNA were monitored using 1% agarose gels. According to the concentration, the DNA samples were diluted to 1 ng/µL with sterile water to reduce the effects of the PCR inhibitors. The V4 gene of the 16S rRNA was amplified using the 515F/806R primer sets (Bergmann et al., 2011; Zeng, An & Liu, 2017). All PCR reactions were carried out with Phusion^® High-Fidelity PCR Master Mix (New England Biolabs, Ipswich, MA, USA). The same volume of 1 × loading buffer (contained SYB green) was mixed with the PCR products, and electrophoresis was conducted on 2% agarose gels for detection. The samples with a bright strip between 400–450 bp were chosen for further experiments. The PCR products were mixed at equal density ratios. Then, the PCR mixtures were purified using a Qiagen Gel Extraction Kit (Qiagen, Hilden, North Rhine-Westphalia, Germany).

Illumina Miseq sequencing

Sequencing libraries were generated using a TruSeq^® DNA PCR-Free Sample Preparation Kit (Illumina, Hayward, CA, USA) following the manufacturer’s recommendations and index codes were added. The library quality was assessed on the Qubit@ 2.0 Fluorometer (Thermo Scientific, Waltham, MA, USA) and Agilent Bioanalyzer 2100 system. Finally, the library was sequenced on an Illumina HiSeq 2,500 platform and 250 bp paired-end reads were generated. The 16S rRNA gene amplicon sequencing was conducted at Novogene Bioinformatics Technology Co., Ltd., Beijing, China. The raw sequence data in FASTQ format are accessible from the NCBI SRA with the number of SRP107086.

Statistical and bioinformatics analysis

QIIME software was used to analyze the sequences data (Caporaso et al., 2010). The sequencing data yielded 569,171 raw reads, with 71,146 raw reads per sample. After removing the low quality reads and trimming the barcodes and primers, there were 545,740 valid reads (average length 253 bp). Clustering sequences at 97% similarity levels were assigned to the same OTUs (Stackebrandt & Goebel, 1994). After the removal of chimeric sequences, a total of 4,833 different OTUs were recorded. Taxonomy was assigned to each OTU via the Ribosomal Database Project (RDP) classifier (Cole et al., 2009). The representative sequence for each OTU was screened for further annotation. The abundance of OTUs information was normalized using a standard sequence number corresponding to the sample with the fewest sequences. The alpha diversity was applied to analyze the complexity of the species diversity of each sample, including the observed-species index and the Shannon index. All indices in our samples were calculated with QIIME (Version1.7.0) and displayed with R software (Version 2.15.3).

The similarities between treatments were measured using a principal coordinate analysis (PCoA) plot. The PCoA was analyzed using the WGCNA, stat and ggplot2 packages in R software (Version 2.15.3). One-way ANOSIM and SIMPER analysis were used to compare the differences in the bacterial composition among the different treatments using the Bray-Curtis method (PRIMER software v 7) (Zeng, An & Liu, 2017). A higher R value in ANOSIM indicated a higher separation between the treatments. The linear discriminant analysis effect size (LEfSe) method was used to determine the difference between the normal and the double litter amount treatments (Segata et al., 2011). One-way ANOVA was performed to explore the differences between the soil properties and the soil bacterial compositions under the different treatments (SPSS version 20.0 for Windows), and the Student-Newman-Keuls (SKN) method was used for the comparison (P = 0.05). The relationships between soil bacterial composition and the environmental factors were tested using Pearson correlation analyses using SPSS 20.0 for Windows.

Soil chemical properties and the response of microbial biomass to litter decomposition

The soil nitrogen fractions, carbon fractions and soil moisture were significantly altered by the addition of litter (Fig. 2). The soil moisture showed a significant decline in the normal treatment and an increase in double treatment. No significant differences were observed among the treatments for soil NH${}_4^{+}$-N, which ranged from 5.39 to 5.73 mg/kg. The MBN content was significantly higher in the normal treatment and ranged from 43.50 to 124.14 mg/kg, and was in the order of normal>double >control. The DON showed the opposite trend to the MBN, with the highest value measured in the control treatment. The soil nitrate nitrogen ranged from 21.98 to 27.90 mg/kg, and there was no significant difference between the normal and the control treatments. The control treatment had the highest MBC and the lowest DOC, and was significantly different from the double treatment. With the increase of litter quantity, the soil nitrate nitrogen, soil moisture, MBC, DOC and DON showed significant reductions in the normal treatment, and a significant increase was observed in the MBN.

Response of the soil bacterial community activity to litter decomposition

The bacterial diversity indices showed no significant changes between the different treatments (Table 1), but the soil bacterial community compositions demonstrated significant structuring in response to litter addition. The most dominant groups across all soil samples were Proteobacteria (38–42%), Actinobacteria (11–21%), Acidobacteria (18–20%), Gemmatimonadetes (5%), Bacteroidetes (4–6%), Chloroflexi (3%), Firmicutes (1–2%), Verrucomicrobia (2–4%), Planctomycetes (3–4%) and Nitrospirae (2%) (Fig. 3). The relative abundance of Actinobacteria, Bacteroidetes, Planctomycetes, Firmicutes and Nitrospirae in the normal treatment was significantly higher than in the double and control treatments (Fig. 3A).

To explore the dynamics of the major microbial taxa under different mounts of litter treatment, we found that Alpha, Beta, Gamma, and Delta-proteobacteria were the main members of Proteobacteria. Only Alpha-proteobacteria showed no significant differences among the different treatments, and ranged from 15.50 to 17.82%. With the increase of litter quantity, the relative abundance of Bet, Gamma, and Deltaproteobacteria showed a decrease in the normal treatment, and an increase in the double treatment. The Beta, Gamma, and Deltaproteobacteria occupied 5.75%, 6.00%, and 6.93%, respectively, in the normal treatment, which significantly differed from the double and control treatment (Fig. 3B). At the order level, Subgroup_6 and Subgroup_4 were the dominant taxa in the Acidobacteria phylum, and showed no significant changes with the increase of litter quantity. Rhizobiales was the dominant taxa of Alpha-proteobacteria, and ranged from 7.01 to 8.75%, and showed similar variation to those of the Alpha-proteobacteria. Solirubrobacterales, Xanthomonadales, Sphingobacteriales, Myxococcales and Gaiellales had significant differences among the litter addition treatments (Fig. 4). These differences were only detected between the normal treatment and the double or the control treatment. The cluster analysis and PCoA also indicted these changes (Figs. 3 and 5). More specifically, the bacterial community profiles in normal treatment trended to group together and were separated from those in the double and control treatments. A t-test showed that the soil bacterial taxa were significantly different between the normal and the double treatments, including Proteobacteria (Xanthomonadales, Salinisphaerales, Legionellales, Chromatiales, Syntrophobacterales, Sh765B-TzT-29, Myxococcales, SC-I-84, Sneathiellales, DB1-14 andCaulobacterales), Planctomycetes (WD2101_soil_group, Phycisphaerales, CCM11a), and Actinobacteria (Micrococcales, Solirubrobacterales, Rubrobacterales and Acidimicrobiales) (Fig. 5).

The ANOSIM based on the OTUs of the 16S rRNA gene sequences indicated that the differences were significant between the different litter addition treatments (ANOSIM Global R = 0.761, P = 0.01). SIMPER analysis revealed that bacterial communities were 76–81% similar between the normal, double and control treatments. The LEfSe analyses identified the significant difference in the abundant taxa between the different litter quantity treatments. Using the LEfSe, we found that Bacteroidetes, Myxococcales and Deltaproteobacteria were primarily different in the high-litter treatment (double). The green color in Fig. 6 indicates the significantly different taxa in the normal treatment, and these species could potentially be used as biomarkers in the normal quantity treatment (Fig. 6).

Pearson correlation analysis showed that soil moisture, DON and MBN were the factors that mainly contributed to the significant correlation with bacterial taxa (Table 2). DON was significantly correlated with the relative abundance of Actinobacteria, Bacteroidetes, Verrucomicrobia, Verrucomicrobia, Firmicutes and Nitrospirae, with coefficients of −0.684, 0.812, 0.679, 0.669, −0.804 and 0.715, respectively. The SM and MBN were similarly correlated with the bacterial community composition (Table 2). There were no significant correlations with the relative abundance of Acidobacteria, Gemmatimonadetes and Chloroflexi, as the abundance of these taxa was stable among the different treatments.