How does urbanization impact connections within the soil microbiome?

Kathryn Atherton, Graduate Program in Bioinformatics, Boston University, Boston, MA, Kathryn Atherton, Graduate Program in Urban Biogeoscience and Environmental Health, Boston University, Boston, MA, Chikae Tatsumi, Graduate School of Agriculture, Kyoto University, Kyoto, Japan, Chikae Tatsumi, Research Faculty of Agriculture, Hokkaido University, Hokkaido, Japan, Chikae Tatsumi, Pamela Templer and Jennifer M. Bhatnagar, Department of Biology, Boston University, Boston, MA, Sarah M. Garvey and Lucy R. Hutyra, Department of Earth & Environment, Boston University, Boston, MA

Background/Question/Methods
Urbanization and forest fragmentation are increasing globally, shifting soil microbial community composition and depleting biodiversity. Interactions between microorganisms are vital to soil community function, as they determine community diversity, stability, and biogeochemical functions, yet the combined impact of urbanization and habitat fragmentation on microbial interactions is unknown. We hypothesized fragmentation would interact with urbanization, such that soil community networks at the forest edge would display greater disruption with urbanization than networks at the forest interior. To test this hypothesis, we collected soil samples from eight sites along an urban-to-rural gradient from Boston into Western Massachusetts and edge-interior gradients of forests at each site in 2018 and 2019. We extracted DNA, sequenced fungal ITS amplicons using Illumina MiSeq, and assigned taxonomy to amplicon sequence variants. We then created interaction networks of soil fungal communities at the edge and interior of the forest for each site. For each network, we calculated assortativity; average shortest path length between taxa; betweenness, closeness, and degree centrality; mean clustering coefficient; cohesion; complexity; connectance; number of edges; and edge density. We used a two-way factorial ANOVA to determine whether these metrics were impacted by distance from forest edge, distance from Boston, and the interaction between these factors.
Results/Conclusions
We found an interaction between edge and urbanization effects on soil microbial community networks, where edge communities were more disrupted in urban forests than rural forests. Average shortest path length between taxa decreased with distance from edge in urban forests, but increased in rural forests (edge*urbanization interaction: F(1,15) = 6.268, p = 0.0277). Additionally, average shortest path length between taxa significantly decreased as distance from Boston increased (F(1,15) = 8.818, p = 0.0117), indicating greater connectedness in rural forests. Contrastingly, the number of connections between microbes within the networks was significantly affected by urbanization alone. Network complexity (F(1,15) = 50.789, p = 1.2e-5), degree centrality (F(1,15) = 12.550, p = 0.00405), and number of edges (F(1,15) = 48.814, p = 1.46e-5) increased significantly with distance from Boston. These results suggest that as microbiomes become more rural, the number of associations between taxa increases. None of the network metrics were significantly associated with distance from forest edge. These vital connections between microbial taxa in complex soil communities may help rural microbiomes retain more biodiversity than urban microbiomes. Future research will aim to identify urban ecosystem properties that interrupt microbial interactions.