The root microbiota plays a crucial role in assisting the plant host in combating various biotic and abiotic stresses, notably drought, which poses a significant threat to global food security. Despite extensive efforts to understand the shifts in rhizosphere and endosphere bacteriomes, there remains a gap in comprehending how drought stress influences the co-occurring network patterns within these compartments and their ecological functional potentials. To address this gap, a pot experiment was conducted with two treatments: continuous flooding as a control and drought treatment. Bulk soil, rhizosphere, and endosphere samples were collected and subjected to high-throughput sequencing and bioinformatics analysis. The results revealed that drought stress significantly reduced the rice biomass but increased the Shannon diversity index in both the rhizosphere and endosphere bacterial communities with no observable effect on richness across compartments. Additionally, drought treatment markedly altered the community structure and bacterial assemblages in these compartments, resulting in the specific enrichment of Actinobacteriota, Gemmatimonadetes, and Patescibacteria, while Bacteroidetes and Firmicutes were depleted in the rhizosphere and endosphere. Furthermore, drought heightened the complexity of the co-occurring networks and the proportions of positive links across all sampling compartments; this effect was accompanied by an increase in the number of connectors in the bulk soil and rhizosphere, as well as module hubs in the rhizosphere. Functional potential prediction indicated that drought stress significantly altered multiple potential ecological functions across all sampling compartments, particularly enriching functions related to the oxidation of sulfur, manganese, and hydrogen in the bulk soil, while functions associated with iron oxidation were significantly depleted in the rhizosphere. Overall, our results demonstrate that under drought stress, rice may specifically enrich certain bacterial taxa and enhance their positive interactions within its root system to improve adaptation.

Long-term fertilization can reshape soil microbial communities and alter soil ecosystem functioning. However, few studies have comprehensively investigated changes in microbial functional potential for carbon, nitrogen, phosphorus, and sulfur (CNPS) cycling in bulk and rhizosphere soils in response to long-term fertilization, or the relationships between microbial taxonomic shifts and functional responses. We combined 16S rRNA gene amplicon sequencing with high-throughput quantitative PCR targeting 71 CNPS cycling-related genes in a 32-year field experiment comprising an unfertilized control (CK), mineral fertilizer (NPK), manure (M), and NPK combined with manure (NPKM). The results showed that both mineral NPK fertilizer and manure increased soil nutrient availability but had contrasting effects on soil pH. Relative to CK, NPK, M, and NPKM increased rapeseed grain yield by 303.9%, 108.7%, and 397.8%, respectively, and biomass by 303.0%, 99.6%, and 410.5%, respectively. Fertilization also reshaped bacterial community composition, and the responses were similar between bulk and rhizosphere soils. In contrast, functional-gene abundances showed significant soil compartment-dependent responses to fertilization. NPK fertilization reduced most rhizosphere gene groups associated with C degradation, N transformation, P mobilization, and S cycling while increasing nitrification-related genes, whereas manure increased genes involved in carbohydrate degradation, methane metabolism, multiple N transformations, organic P mineralization, phosphate solubilization, and sulfate reduction. Network analysis showed that genes mediating similar processes clustered within common modules and were associated with shared bacterial taxa. Crop productivity was associated with variation in genes involved in N fixation and organic P mineralization in rhizosphere soil. Our results revealed associations between functional-gene abundances, soil nutrients, crop productivity, and bacterial community changes under long-term fertilization, highlighting the importance of quantitative functional-gene analysis as an indicator for evaluating soil multifunctionality.