Abstract
In mature soils, plant species and soil type determine the selection of root microbiota. Which of these two factors drives rhizosphere selection in barren substrates of developing desert soils has, however, not yet been established. Chronosequences of glacier forelands provide ideal natural environments to identify primary rhizosphere selection factors along the changing edaphic conditions of a developing soil. Here, we analyze changes in bacterial diversity in bulk soils and rhizospheres of a pioneer plant across a High Arctic glacier chronosequence. We show that the developmental stage of soil strongly modulates rhizosphere community assembly, even though plant-induced selection buffers the effect of changing edaphic factors. Bulk and rhizosphere soils host distinct bacterial communities that differentially vary along the chronosequence. Cation exchange capacity, exchangeable potassium, and metabolite concentration in the soil account for the rhizosphere bacterial diversity. Although the soil fraction (bulk soil and rhizosphere) explains up to 17.2% of the variation in bacterial microbiota, the soil developmental stage explains up to 47.7% of this variation. In addition, the operational taxonomic unit (OTU) co-occurrence network of the rhizosphere, whose complexity increases along the chronosequence, is loosely structured in barren compared with mature soils, corroborating our hypothesis that soil development tunes the rhizosphere effect.
Original language | English (US) |
---|---|
Pages (from-to) | 1188-1198 |
Number of pages | 11 |
Journal | The ISME Journal |
Volume | 12 |
Issue number | 5 |
DOIs | |
State | Published - Jan 15 2018 |
Bibliographical note
KAUST Repository Item: Exported on 2020-10-01Acknowledgements: We thank CNR-DTA for the use of the CNR Arctic Station Dirigibile Italia in Ny-Ålesund, Svalbard. ER was supported by Università degli Studi di Milano (DeFENS), European Social Found and Regione Lombardia (contract 'Dote Ricerca'). This work was partially supported by funding from King Abdullah University of Science and Technology (KAUST). We are particularly grateful to Karoline Faust for essential support on the network analysis.