Anthropogenic actions have significantly increased biological nitrogen (N) availability on a global scale. In tallgrass prairies, this phenomenon is exacerbated by land management changes, such as fire suppression. Historically, tallgrass prairie fire removed N through volatilization, but fire suppression has contributed to increased soil N availability as well as woody encroachment. Because soil microbes respond to N availability and plant growth, these changes may alter microbial composition and important microbially-mediated functions. Grassland management affects the soil environment on multiple time scales including short (fertilization or fire event), seasonal (growing vs. non-growing season), and long-term (decadal plant turnover and nutrient accumulation), therefore my goal was to understand community variability at different time scales affecting the population and community dynamics of soil microbes. I predicted soil microbes would be sensitive to environmental changes at all time scales, seasonal variation would reflect increased plant rhizodeposit-supported populations during summer and decomposers during winter, and long-term fire suppression and chronic fertilization would drive soil microbial community turnover associated with accumulation of plant litter and N.
Soil microbial responses to short-term fire/fertilization events were minimal, while microbial population sizes fluctuate seasonally and synchronously, and microbial community composition varied more with management history than at shorter time scales. Bacterial populations increased 10x during growing-season plant rhizodeposition, while fungal populations were less dynamic, but decreased in fall, possibly reflecting a shift to subsistence on soil organic matter. In contrast, microbial community composition was seasonally stable, but distinct between long-term management treatments, which may indicate accumulation of niche-defining plant or soil properties over decades. Prokaryotic communities responded to altered N availability via both fertilization and loss due to fire, with the highest abundance of "copiotrophic" (r-selected) taxa in unburned, fertilized soils. Fungal communities responded to N fertilization with higher abundance of arbuscular mycorrhizal fungi, pathogens, and saprotrophs, possibly due to changes in nutrient stoichiometry and litter availability in fertilized plots. However, fungal response to fire was largely independent of N availability, and plant community differences were correlated with fungal, but not bacterial, community composition, highlighting the likely nutritional codependence of fungi and plants, and fungal competitive advantages for plant litter substrates. The timing of changes in soil microbial communities is critical for plant nutrition and nutrient cycling in prairies, and this novel dataset on the temporal resolution of microbial responses to environmental variability contributes to the broader understanding of ecosystem responses to global change.
To address these predictions, soils were collected from the Belowground Plot Experiment (BGPE) at Konza Prairie Biological Station: a 30-y factorial field manipulation of N fertilization and burning. Surface soils (0-15 cm) were sampled monthly between Nov 2014 – Dec 2015, including one week post-fire (April) and post-fertilization (June). Genomic DNA was extracted from each sample for qPCR and PCR for Illumina MiSeq library sequencing of the prokaryotic 16S rRNA gene and fungal ITS, to estimate population and community dynamics of soil microbes. Soil environmental characteristics and plant communities were measured in July 2015 to evaluate correlations between plant and microbial communities, and environmental variability.
Bacterial 16S rRNA gene and fungal ITS copy number:Population sizes of bacteria and fungi were estimated using quantitative PCR (qPCR) assays using a Bio-Rad CFX CONNECT system with Bio-Rad SsoAdvanced Universal SYBR Green Supermix (Bio-Rad Laboratories, Hercules, CA, USA). For 16S rRNA gene and ITS assays, respectively, ~1 ng (exact amount calculated on a per sample basis) template gDNA was added to a 10 μL PCR volume, with 0.02% and 0.04% final BSA concentration, 100 nM and 500 nM final primer concentrations, using established primer sequences and thermal cycler programs (Fierer et al. 2005, AEM 71: 4117-4120). Standard curves for 16S assays were prepared using E. Coli ATCC 25922 at 5x100 - 5x10-6 ng ul-1 DNA concentrations, and ITS standard curves were prepared using Candidia albicans SC5314 at 5x100 - 5x10-6 ng ul-1 concentrations; successful curves were accepted at 100 +/- 15% efficiency and R2 > 0.99. All assays included no-template controls and melting curves, which confirmed that only gene copies from template soil gDNA were quantified, and 3 technical replicates were run per sample. Soil gDNA yield (µg DNA g-1 dry soil) was used to normalize marker gene copy number per g dry soil.