Effects of plant species and soil characteristics on rhizosphere microbial community structure in North Texas prairies

Background/Question/Methods

Microbes play a key role in promoting a healthy ecosystem in grassland prairies. However, the factors governing the structure of rhizosphere microbial communities are not fully understood.  In this study, we compared the bacterial and fungal community composition in the rhizosphere of a native grass (Andropogon gerardii) and a non-native grass (Sorghum halepense) from two North Texas prairies with dramatically different soil textures.  Soils were analyzed for pH, organic matter, moisture, total carbon, total nitrogen, and clay and sand percentages. Microbial community fingerprints were generated using terminal restriction fragment length polymorphism (TRFLP) analysis.  The overall structure of bacterial and fungal communities between the prairie sites and grass species were compared by analyzing binary data using the Additive Main Effects and Multiplicative Interaction (AMMI) model.  This technique combines analysis of variance to first partition the variation into main effects and interactions, and then applies principal component analysis (PCA) to the interactions to create interaction principal component axes.  Stepwise multiple regressions were used to examine the correlation between soil characteristics and the interaction principal component axes, which represent bacterial or fungal community structure.

Results/Conclusions

Variability in microbial community structure was completely captured by two principal components.  As evidenced by separation on the first principal component, the primary driver of bacterial and fungal communities was differences between the sites.  Site differences explained 80% of bacterial community variability and 77% of fungal community variability.  As anticipated by our experimental design, site separation of both bacterial and fungal community structures was significantly predicted by soil texture (% clay).  The remaining independent variables measured (pH, total carbon, total nitrogen, soil moisture, soil organic matter) increased with percent clay content along the first principal component axis.  Soil moisture, pH, and total nitrogen were significant predictors of bacterial communities while pH and soil organic matter were significant predictors of fungal communities.  Grass species appeared to be a secondary driver of bacterial and fungal community structure, separating along the second principal component.   Along this axis, soil pH and organic matter were significant predictors of the bacterial communities while none of the variables measured were good predictors of fungal communities.  Although plant type may influence rhizosphere microbial communities, this study indicates the majority of variability associated with Johnson grass and big bluestem microbial communities could be explained by soil characteristics.