Effects of cover crop and fertilizer incorporation on the structure and function of microbial communities in soils under long-term organic management

Abstract

Incorporation of organic materials, including cover crop residues and biological fertilizers, is an important element of organic farming practices. These amendments can affect the composition of soil microbial communities that carry out nutrient cycling and perform other functions crucial to crop health and growth. These changes may be detectable through use of 16S rDNA profiling of soil bacterial community composition. We conducted a field experiment at three southern Minnesota sites under long-term organic management to determine the effects of cover crop and organic fertilizer treatments applied in fall 2012 on soil bacterial community structure, nutrient cycling functions, and physicochemical properties during the 2013 growing season. Illumina sequencing of 16S rDNA revealed diverse communities encompassing 45 bacterial phyla in bulk and rhizosphere soil. Cover crop and fertilizer application tended to lower OTU richness and diversity in May 2013 samples; however, these treatments increased diversity in some July 2013 samples compared to a no-amendment control. Fertilizer treatments tended to decrease relative abundance of the ammonia-oxidizing family Nitrosomonadaceae, as did some cover crops. Fertilizers that produced large increases in soluble N levels also decreased relative abundance of Rhizobiales. Soil functional profiles were more strongly predicted by location than by treatment. Location differences were strongly associated with variance in soil physicochemical parameters. Both treatment and location effects were mediated by soil physicochemical properties including pH, moisture, soil organic matter, and nutrient levels. While cover crop and fertilizer application affected several soil functions, including increases in respiration and enzyme activities in cover crop treatment, these effects were not consistent across locations and sampling time points. Differences in soil function were better explained by using both soil physicochemical test values and community structure data than using soil physicochemical tests alone. Strong associations were observed in both bulk and rhizosphere soil between functional profiles and the families Cytophagaceae and Micrococcaceae. The Actinobacteria overall emerged as the phylum most significantly predictive of functional profiles. We also observed correlations between soil physicochemical parameters and nutrient cycling functions and bacterial family abundances, including strong correlations between pH and several phylogenetically disparate families including Xanthobacteraceae, Pseudonocardiodaceae, Propionibacteraceae, and Sinobacteraceae. Strong positive and negative associations with bacterial families were observed for β-glucosidase activity. Bayesian analysis inferred no directional relationships between specific functional activities and particular bacterial families or physicochemical parameters. However, concentrations of NO3-N, SO4-S, and OM were each inferred to have parent relationships to the abundance of at least one family. OTU richness (Chao1) had both parent and child relationships to bacterial family abundances. Moisture and pH, although generally highly predictive of bacterial community composition, did not show directional relationships with taxa abundances in the Bayesian analysis. Our findings indicate that cover crop and fertilizer application practices are capable of affecting the microbial environment encountered by subsequent crops. Our results also support the use of sequencing-based microbial community structure profiling as an important tool for increasing our understanding of community structure-function relationships, with the goal of improving our ability to predict the effects of agricultural management practices on important soil functions.