Association between Soil Physicochemical Properties and Bacterial Community Structure in Diverse Forest Ecosystems

Although the importance of the soil bacterial community for ecosystem functions has long been recognized, there is still a limited understanding of the associations between its community composition, structure, co-occurrence patterns, and soil physicochemical properties. The objectives of the present study were to explore the association between soil physicochemical properties and the composition, diversity, co-occurrence network topological features, and assembly mechanisms of the soil bacterial community. Four typical forest types from Liziping Nature Reserve, representing evergreen coniferous forest, deciduous coniferous forest, mixed conifer-broadleaf forest, and its secondary forest, were selected for this study. The soil bacterial community was analyzed using Illumina MiSeq sequencing of 16S rRNA genes. Nonmetric multidimensional scaling was used to illustrate the clustering of different samples based on Bray-Curtis distances. The associations between soil physicochemical properties and bacterial community structure were analyzed using the Mantel test. The interactions among bacterial taxa were visualized with a co-occurrence network, and the community assembly processes were quantified using the Beta Nearest Taxon Index (Beta-NTI). The dominant bacterial phyla across all forest soils were Proteobacteria (45.17%), Acidobacteria (21.73%), Actinobacteria (8.75%), and Chloroflexi (5.06%). Chao1 estimator of richness, observed ASVs, faith-phylogenetic diversity (faith-PD) index, and community composition were distinguishing features of the examined four forest types. The first two principal components of redundancy analysis explained 41.33% of the variation in the soil bacterial community, with total soil organic carbon, soil moisture, pH, total nitrogen, carbon/nitrogen (C/N), carbon/phosphorous (C/P), and nitrogen/phosphorous (N/P) being the main soil physicochemical properties shaping soil bacterial communities. The co-occurrence network structure in the mixed forest was more complex compared to that in pure forests. The Beta-NTI indicated that the bacterial community assembly of the four examined forest types was collaboratively influenced by deterministic and stochastic ecological processes.