Long-term application of nitrogen and phosphorus fertilizers changes the process of community construction by affecting keystone species of crop rhizosphere microorganisms

Keystone species of microbial communities play a very important role in community structure and ecosystem function; however, the effect of long-term nitrogen (N) and phosphorus (P) fertilizers on key taxa and the mechanisms of community construction of rhizosphere microbial communities remain unclear. In this study, the effect of nine fertilization treatments (N0P0, N0P1, N0P2, N1P0, N1P1, N1P2, N2P0, N2P1, and N2P2) on soil microbial community diversity, keystone species, and construction methods in the crop rhizosphere were studied in a loess hilly area after 26 years of fertilization. The results showed that fertilization significantly increased the nutrient contents of the rhizospheric soil and root system and significantly affected microbial community composition (based on the Bray-Curtis distance) and community construction process (β-nearest taxon index: βNTI). The decrease in the abundance of oligotrophic bacteria (from phyla Acidobacteriota and Chloroflexi) in the keystone species of bacterial communities shifted the community construction process from homogenizing dispersal to variable selection process and was significantly regulated by soil factors (total P and carbon-N ratio). However, the decrease in the abundance of keystone species (from phylum Basidiomycota) in the fungal communities did not have a significant effect on community construction, which was mainly affected by root characteristics (root N content and soluble sugar). This study found that long-term N and P fertilization changed the keystone species composition of bacterial communities by affecting the nutrient content of the rhizospheric soil, such as total P, so that the construction mode of communities changed from a stochastic to a deterministic process, and the N2 fertilization, especially the N1P2 treatment was better for increasing network stability (modularity and clustering coefficient).