Soil Bacterial Community Structure and Co-occurrence Pattern during Vegetation Restoration in Karst Rocky Desertification Area

Abstract

Vegetation restoration has been widely used in karst rocky desertification (KRD) areas of southwestern China, but the response of microbial community to revegetation has not been well characterized. We investigated the diversity, structure, and co-occurrence patterns of bacterial communities in soils of five vegetation types (grassland, shrubbery, secondary forest, pure plantation and mixed plantation) in KRD area using high-throughput sequencing of the 16S rRNA gene. Bray-Curtis dissimilarity analysis revealed that 15 bacterial community samples were clustered into five groups that corresponded very well to the five vegetation types. Shannon diversity was positively correlated with pH and Ca²⁺ content but negatively correlated with organic carbon, total nitrogen, and soil moisture. Redundancy analysis indicated that soil pH, Ca²⁺ content, organic carbon, total nitrogen, and soil moisture jointly influenced bacterial community structure. Co-occurrence network analysis revealed non-random assembly patterns of bacterial composition in the soils. Bryobacter, GR-WP33-30, and Rhizomicrobium were identified as keystone genera in co-occurrence network. These results indicate that diverse soil physicochemical properties and potential interactions among taxa during vegetation restoration may jointly affect the bacterial community structure in KRD regions.

Keywords: bacterial communities; co-occurrence; karst; rocky desertification; vegetation restoration.

FIGURE 1

Clustering analysis of bacterial communities under five vegetation types based on OTU abundance-based Bray-Curtis similarity coefficients.

FIGURE 2

Venn diagram of exclusive and shared bacterial OTUs (at the 3% evolutionary distance) under five vegetation types.

FIGURE 3

A heatmap diagram of the dominant 50 OTUs under five vegetation types.

FIGURE 4

Redundancy analysis to show correlation between the bacterial communities and physicochemical properties under five vegetation types.

FIGURE 5

Co-occurring network of bacterial communities across five vegetation types based on correlation analysis. The nodes in network (A) are colored by phylum. The nodes in network (B) are colored by modularity class. The connections stands for a strong (spearman’s ρ > 0.6) and significant (P < 0.01) correlations. The size of each node is proportional to the relative abundance of specific genus. The thickness of each edge is proportional to the ρ.