Microbial Community and Functional Structure Significantly Varied among Distinct Types of Paddy Soils But Responded Differently along Gradients of Soil Depth Layers

Abstract

Paddy rice fields occupy broad agricultural area in China and cover diverse soil types. Microbial community in paddy soils is of great interest since many microorganisms are involved in soil functional processes. In the present study, Illumina Mi-Seq sequencing and functional gene array (GeoChip 4.2) techniques were combined to investigate soil microbial communities and functional gene patterns across the three soil types including an Inceptisol (Binhai), an Oxisol (Leizhou), and an Ultisol (Taoyuan) along four profile depths (up to 70 cm in depth) in mesocosm incubation columns. Detrended correspondence analysis revealed that distinctly differentiation in microbial community existed among soil types and profile depths, while the manifest variance in functional structure was only observed among soil types and two rice growth stages, but not across profile depths. Along the profile depth within each soil type, Acidobacteria, Chloroflexi, and Firmicutes increased whereas Cyanobacteria, β-proteobacteria, and Verrucomicrobia declined, suggesting their specific ecophysiological properties. Compared to bacterial community, the archaeal community showed a more contrasting pattern with the predominant groups within phyla Euryarchaeota, Thaumarchaeota, and Crenarchaeota largely varying among soil types and depths. Phylogenetic molecular ecological network (pMEN) analysis further indicated that the pattern of bacterial and archaeal communities interactions changed with soil depth and the highest modularity of microbial community occurred in top soils, implying a relatively higher system resistance to environmental change compared to communities in deeper soil layers. Meanwhile, microbial communities had higher connectivity in deeper soils in comparison with upper soils, suggesting less microbial interaction in surface soils. Structure equation models were developed and the models indicated that pH was the most representative characteristics of soil type and identified as the key driver in shaping both bacterial and archaeal community structure, but did not directly affect microbial functional structure. The distinctive pattern of microbial taxonomic and functional composition along soil profiles implied functional redundancy within these paddy soils.

Keywords: GeoChip; Mi-Seq sequencing; microbial community; network analysis; paddy soil; soil profile; soil type.

FIGURE 1

Detrended correspondence analysis (DCA) for microbial taxonomic and functional community. Results of (A) bacterial community, (B) archaeal community, and (C) all of the functional genes from four soil profiles (A, B, C, D) of three paddy soils (BH, LZ, TY).

FIGURE 2

The Z-P plot indicating categories of nodes according to Z_i (within-module connectivity) and P_i (among-module connectivity) of bacterial (A) and archaeal community (B) in four soil profile layers.

FIGURE 3

Dominant taxonomic groups of bacterial (A) and archaeal (B) groups in four soil profiles of three paddy soils. Cyan triangles denote that the relative abundance of the group uniformly decreased along depth layers within each soil type, cyan inverted triangles denote that the relative abundance of the group uniformly increased along depth layers.

FIGURE 4

Abundance of different of categories of microbial functional genes in three paddy soils. ^∗∗ Denotes that the abundance is significantly higher in BH and TY than in LZ soils.

FIGURE 5

Mutualistic networks of interaction between bacterial community and functions of the four soil profile layers (A–D), and archaeal community and functions (A′–D′). Edges in green colors represent influence of microbes to ecological functions, and edges in other colors represent influence of each functional category to microbes.

FIGURE 6

Effects of soil type (geographic distance), depth, soil HWC, pH and salinity (EC) on bacterial community and functional (A) and archaeal community and functional (B) structure. Geographical distance includes longitude and latitude of the soil sampling sites. Solid lines denote positive effects, and broken lines denote negative effects. Thickness of the arrows denotes significance and strength of the influence. R² represent percentage of explanation of the models on the chosen factors. Significant level: ^∗P < 0.05, ^∗∗P < 0.01, ^∗∗∗P < 0.001. Goodness-of-fit statistics are as following: (A) Chi-square = 0.000, degrees of freedom = 1, RMSEA = 0.000, AIC = 54, GFI = 1.000; (B) Chi-square = 0.000, degrees of freedom = 1, RMSEA = 0.000, AIC = 54, GFI = 1.000.