TY - JOUR AU - Pellissier, Loïc AU - Pinto, Eric AU - Niculita-Hirzel, Hélène AU - Moora, Mari AU - Villard, Lucas AU - Goudet, Jérome AU - Guex, Nicolas AU - Pagni, Marco AU - Xenarios, Ioannis AU - Sanders, Ian AU - Guisan, Antoine PY - 2013 M3 - Hypothesis and Theory TI - Plant species distributions along environmental gradients: do belowground interactions with fungi matter? JO - Frontiers in Plant Science UR - https://www.frontiersin.org/articles/10.3389/fpls.2013.00500 VL - 4 SN - 1664-462X N2 - The distribution of plants along environmental gradients is constrained by abiotic and biotic factors. Cumulative evidence attests of the impact of biotic factors on plant distributions, but only few studies discuss the role of belowground communities. Soil fungi, in particular, are thought to play an important role in how plant species assemble locally into communities. We first review existing evidence, and then test the effect of the number of soil fungal operational taxonomic units (OTUs) on plant species distributions using a recently collected dataset of plant and metagenomic information on soil fungi in the Western Swiss Alps. Using species distribution models (SDMs), we investigated whether the distribution of individual plant species is correlated to the number of OTUs of two important soil fungal classes known to interact with plants: the Glomeromycetes, that are obligatory symbionts of plants, and the Agaricomycetes, that may be facultative plant symbionts, pathogens, or wood decayers. We show that including the fungal richness information in the models of plant species distributions improves predictive accuracy. Number of fungal OTUs is especially correlated to the distribution of high elevation plant species. We suggest that high elevation soil show greater variation in fungal assemblages that may in turn impact plant turnover among communities. We finally discuss how to move beyond correlative analyses, through the design of field experiments manipulating plant and fungal communities along environmental gradients. ER -