Environmental Filtering of Microbial Communities in Agricultural Soil Shifts with Crop Growth

<div><p>Plant and soil properties cooperatively structure soil microbial communities, with implications for ecosystem functioning. However, the extent to which each factor contributes to community structuring is not fully understood. To quantify the influence of plants and soil properties on microbial diversity and composition in an agricultural context, we conducted an experiment within a corn-based annual cropping system and a perennial switchgrass cropping system across three topographic positions. We sequenced barcoded <i>16S</i> ribosomal RNA genes from whole soil three times throughout a single growing season and across two years in July. To target the belowground effects of plants, we also sampled rhizosphere soil in July. We hypothesized that microbial community α-diversity and composition (β-diversity) would be more sensitive to cropping system effects (annual vs. perennial inputs) than edaphic differences among topographic positions, with greater differences occurring in the rhizosphere compared to whole soil. We found that microbial community composition consistently varied with topographic position, and cropping system and the rhizosphere influenced α-diversity. In July, cropping system and rhizosphere structured a small but specific group of microbes implying a subset of microbial taxa, rather than broad shifts in community composition, may explain previously observed differences in resource cycling between treatments. Using rank abundance analysis, we detected enrichment of Saprospirales and Actinomycetales, including cellulose and chitin degraders, in the rhizosphere soil and enrichment of Nitrospirales, Syntrophobacterales, and MND1 in the whole soil. Overall, these findings support environmental filtering for the soil microbial community first by soil and second by the rhizosphere. Across cropping systems, plants selected for a general rhizosphere community with evidence for plant-specific effects related to time of sampling.</p>