The Biogeography And Co-occurrence Patterns Of Root-associated Microbiomes In Soybean Fields Across China

Root-associated microbiomes have profound effects on plant nutrient uptake,growth promotion and disease suppression,but their biogeographic and co-occurrence patterns across large spatial scales remain poorly understood.In this study,we used high–throughput sequencing to investigate the microbial communities in the four soil–root compartments(bulk soil,rhizosphere,endosphere and nodule)across 51 soybean fields in China.We compared bacterial biogeographic patterns among the bulk soil,soybean rhizosphere and endosphere;and compared biogeographic patterns of rhizobial and non-rhizobial subcommunities in nodules.To delineate the co-occurrence patterns of bulk soil and rhizosphere microbiome and the geographic patterns of network topological features,we performed a network-based analysis using integrated bacterial and fungal community datasets.We also compared biogeographic and co-occurrence patterns of diazotrophic communities between bulk soil and soybean rhizosphere.The results are as follows:1.Both bacterial richness and community turnover were significantly correlated with different environmental and spatial variables among the bulk soil,soybean rhizoshere and endosphere.The richness of bulk soil bacteria was negatively correlated with latitude;the converse trend was true for rhizosphere bacteria,and the richness of endosphere bacteria did not have any latitudinal gradients.The spatial autocorrelation ranges for bacteria were also differed among the three soil-root compartments.Collectively,these results imply distinct bacterial biogeographic patterns existing in the three compartments.Environmental variables were more important than spatial variables,and edaphic variables were more important than climatic variables,for governing bacterial community turnover in each soil-root compartment.The distributions of nearest taxon index(NTI)showed that deterministic processes dominated local bacterial communities,while its importance decreased from the bulk soil to the endosphere.2.Dispersal probability was lower in NRE than in rhizobia,as revealed by a significant distance-decay relationship found in NRE,but not in rhizobia,in addition to a significant occupancy–abundance relationship in the entire community.Rhizobial and NRE subcommunities were significantly influenced by different environmental and spatial variables.Moreover,the rhizobial subcommunities were grouped into Ensifer-and Bradyrhizobium-dominated clusters that were significantly related to soil pH.The non-rhizobial subcommunities were grouped into Proteobacteria-and Firmicutes-dominated clusters that were more influenced by climatic than by edaphic factors.These results demonstrated that rhizobial and non-rhizobial subcommunities are characterized by distinct biogeographic patterns.Network analysis showed rhizobia and NRE as separately grouped and uncorrelated with each other,suggesting they did not share niche space in soybean nodules.3.Microbial networks differed between bulk soil and rhizosphere in terms of structure and composition.Compared with the bulk soil networks,the rhizosphere networks had fewer links between bacteria and fungi,lower modularity,and smaller average path length;the global,southern and northern networks of rhizosphere showed similar,higher and lower complexity,respectively.The southern-specific networks of both bulk soil and soybean rhizosphere had more links between bacteria and fungi compared with the northern-specific networks.Additionally,the geographic patterns of network topological features differed between bulk soil and rhizosphere habitats,northern and southern regions.Bacterial sub-networks of both bulk soil and rhizosphere were most influenced by soil pH;fungal sub-networks were related to fewer environmental factors and most influenced by soil Mg content.4.Diazotrophic communities in bulk soil and soybean rhizosphere displayed different biogeographic and co-occurrence patterns.The spatial turnover rate of diazotrophic communities in bulk soil was significantly faster than that in soybean rhizosphere.Environmental selection rather than stochastic processess determined the diazotrophic communities in both bulk soil and soybean rhizsophere.however,diazotrophic communities in these two habitats were significantly correlated with different environmental variables.Among the environmental variables measured,soil pH and MAP were the most important factors driving the diazotrophic community spatial turnover.Bulk soil and soybean rhizosphere were also differed in the structure and compostion of diazotrophic netwroks,suggesting different cooccurrence patterns of diazotrophs in these two habitats.In addition,diazotrophic OTUs from the same genera tend to co-occurr and form a single module,implying that deterministic processes at the genera level have an important effect on the diazotrophic communities.This study investigated the biogeographic and co-occurrence patterns of root-associated microbiomes at a continental scale,revealed the ecological factors that influence the community composition,diversity and potential interactions of root-associated microbiomes.Our results provide new insights into the assembly of root-associated micribial communities.