The effects of different agronomic practices, such as fertilization regimes, can be experimentally tested in long-term experiments (LTE). Here, we aimed to evaluate the effect of different nitrogen fertilizations on the bacterial microbiota in both rhizosphere and bulk soil of sugar beet, in the Giessen-LTE (Germany). Fertilization treatments included mineral-N, manure, mineral-N + manure and no N-amendment. Metabarcoding and co-occurrence analysis of 16S rRNA genes, qPCR of amoA, nirK, nirS, nosZ-I and nosZ-II genes and soil physico-chemical analyses were performed. The effect of the fertilization treatments was more evident in the bulk soil, involving 33.1% of the microbiota. Co-occurrence analysis showed a rhizosphere cluster, dominated by Proteobacteria, Actinobacteria and Verrucomicrobia (hub taxa: Betaproteobacteriales), and a bulk soil cluster, dominated by Acidobacteria, Gemmatominadetes and "Latescibacteria" (hub taxa: Acidobacteria). In the bulk soil, mineral N-fertilization reduced nirK, amoA, nosZ-I and nosZ-II genes. Thirteen Operational taxonomic units (OTUs) showed 23 negative correlations with gene relative abundances. These OTUs likely represent opportunistic species that profited from the amended mineral-N and outgrew the species carrying N-cycle genes. Our results indicate trajectories for future research on soil microbiome in LTE and add new experimental evidence that will be helpful for sustainable management of nitrogen fertilizations on arable soils.
Keywords: Long-Term Experiments (LTE); co-occurrence network; high-throughput sequencing; nitrogen cycle; nitrogen fertilization; quantitative PCR; soil microbiota.