COS 189-5 - Community structure of plant-associated bacteria suggests stringent host and environment filters

Abstract: Consistent with ecological theory, assembly of plant microbiomes results from environmental filters, host filters, and neutral processes (e.g., dispersal). However, studies differ on whether those microbes are sourced from soils or plants. Based on our previous results, we hypothesized that soils are not the main reservoir for plant microbiomes and tested this by asking if root and leaf microbial communities nest within soil microbial communities. Given their proximity, we anticipated root-associated bacteria would be nested in soils, whereas leaf bacteria would not, instead exhibiting a higher degree of specialization. We also expected the degree of nestedness to vary with environmental conditions and their variability, which could affect availability of bacterial taxa. To test these ideas, we used 16S rRNA to characterize bacterial communities of switchgrass (Panicum virgatum L.) leaves, roots, and soils within and across 14 stands spanning mountain to coastal ecoregions of North Carolina, USA. We analyzed network structure of leaf, root, and soil bacterial communities at each site using bipartite networks and graph-based permutation tests. We examined correlations between nestedness, environmental heterogeneity, and environmental variables dimensionally reduced with PCA to represent soil properties, plant traits, and climate conditions.

We observed minimal nesting of plant-associated microbiomes (WNODF< 15, where 100 is perfectly nested). Nevertheless, when considered separately by plant compartment, root bacteria were more nested in soil (WNODF=9-18) than leaf bacteria (WNODF< 0.1). Leaf and root networks were both specialized (H2' >0.6), meaning taxa tended to co-occur within plant compartment as opposed to between compartments. Further supporting specialization, graph-based tests revealed samples grouped significantly by plant compartment (P< 0.5). Root bacteria nestedness in soil was correlated (R=-0.57, P< 0.05) with the soil properties PC (representing soil moisture, nitrogen, bulk density, and dissolved organic carbon), whereas leaf bacteria nestedness in soil was correlated (R=-0.65, P< 0.05) with the climate PC (representing mean annual temperature). This is consistent with different environmental filters that affect local availability of bacterial taxa above- and belowground. Within-site environmental heterogeneity did not predict network structure (P≥0.05), providing additional support for site-level variation as the primary filter. These results imply roots are more likely to recruit bacteria from soil microbial communities, with selection depending on soil characteristics. Conversely, leaf microbiomes do not appear reliant on soil as a major reservoir, but possibly on plant-to-plant dispersal, with selection depending on temperature. Future work will need to directly measure potential inputs to better identify relevant species pools and filters.