Rare and abundant taxa in Artemisia desertorum rhizosphere soils demonstrate disparate responses to drought stress

Abstract

The growth and adaptability of desert plants depend on their rhizosphere microbes, which consist of a few abundant taxa and numerically dominant rare taxa. However, the differences in diversity, community structure, and functions of abundant and rare taxa in the rhizosphere microbiome of the same plant in different environments remain unclear. This study focuses on the rhizosphere microbial communities of Artemisia desertorum, a quintessential desert sand-stabilizing plant, investigating the diversity patterns and assembly processes of rare and abundant taxa across four Chinese deserts: Mu Us, Kubuqi, Tengger, and Ulan Buh. The results show that climatic factors, especially aridity and mean annual precipitation (MAP), significantly influence bacterial community composition and microbial network complexity. The interactions between rare and non-rare taxa are non-random, forming a modular network in which rare taxa serve as central nodes, and their loss could destabilize the network. Rare taxa are primarily shaped by heterogeneous selection, whereas abundant taxa are mainly influenced by dispersal limitation. Functionally, abundant taxa exhibit higher metabolic potential, whereas rare taxa are more involved in processes such as cell motility, indicating distinct ecological roles. These results provide new insights into the ecological functions of rare and abundant taxa in desert rhizosphere communities and highlight the importance of microbial management for desert plant health.

Keywords: Artemisia desertorum; Community assembly process; Desert area; Network complexity; Rhizosphere bacteria.

Fig. 1

Composition and diversity of abundant and rare taxa in the rhizosphere of Artemisia desertorum. a, b Bacterial community composition in the rhizosphere of A. desertorum across different desert ecosystems. c, d Analysis of alpha-diversity (richness and Shannon diversity index) of bacterial communities. e, f Non-metric multidimensional scaling (NMDS) ordination of abundant and rare subcommunities. 95% confidence ellipses were shown around the samples grouped based on different desert systems. Group differences were tested based on ANOSIM analysis. Lowercase letters denote the differences among different deserts at P < 0.05

Fig. 2

Effects of environmental factors on abundant and rare bacterial subcommunities. a, c Distance-based redundancy analysis (dbRDA) showing the relationship between the structure of rare and abundant taxa and environmental factors. b, d Hierarchical partitioning of each environmental factor explaining the structure of rare and abundant subcommunities (P < 0.05). Full abbreviations: MAP, mean annual precipitation; MAT, mean annual temperature; SEC, soil electrical conductivity; TOC, total organic carbon

Fig. 3

Co-occurrence network of A. desertorum rhizosphere systems. a Co-occurrence network colored for phylum. b Module structure of the co-occurrence network. c Network properties colored by abundant or rare species, with external and internal connections among each subcommunity shown in the bottom right. d Composition of each module at phylum level. e Comparison of node-level topological characteristics between abundant and rare taxa. f Distribution of keystone nodes. NS, P > 0.05, **P < 0.01, *P ≤ 0.05 (Wilcoxon test, Benjamini Hochberg·correction)

Fig. 4

Linear regression models of subnetwork topological structure and climatic factors. a Linear regression analysis between subnetwork topology and aridity. b Linear regression analysis between subnetwork topology and mean annual precipitation (MAP)

Fig. 5

Mechanism of bacterial community assembly in A. desertorum rhizosphere. a Distribution of βNTI values for abundant and rare taxa. b The relatively explained ecological processes of abundant and rare subcommunities based on the values of βNTI and RC_bray. The relationships between bacterial phylogenetic turnover (βNTI) and c aridity and d mean annual precipitation (MAP). ***P < 0.001. Linear regressions models (blue lines for rare subcommunities and orange lines for abundant subcommunities) and associated correlation coefficients are provided on each panel

Fig. 6

Functional differences between abundant and rare bacterial subcommunities in the A. desertorum rhizosphere ecosystem. a–d Principal component analysis (PCA) shows the distinct metabolic pathways between rare and abundant taxa ground on KEGG Database level 1. e–h Comparison of the relative abundance of the top 25 functional profiles between abundant and rare taxa based on KEGG Database level 2 using statistical analysis of metagenomic profiles (STAMP). a, e Mu Us Desert, b, f Kubuqi Desert, c, g Tengger Desert, d, h Ulan Buh Desert