Discriminating the effects of agricultural land management practices on soil fungal communities

The structure of fungal communities was examined in soil subjected to 5 years of different agricultural land management and tomato production practices. Length heterogeneity polymerase chain reaction (LH-PCR) of fungal rDNA internal transcribed spacer-1 (ITS-1) regions was used to create genomic fingerprints of the soil fungal communities. Three years after initiation of land management practices, univariate analysis of genetic diversity failed to detect differences among soil fungal communities in plots managed organically, conventionally or maintained free of vegetation by continuous tillage (disk fallow). Genetic diversity was significantly higher in plots maintained as a perennial pasture grass (Paspalum notatum var Argentine bahiagrass) or as an undisturbed weed fallow. The composition of soil fungal communities within organic, pasture grass or disk fallow plots were separated into unique clusters by non-parametric multivariate analysis of their Bray-Curtis similarity matrices, computed from the relative abundance of ITS-1 amplicons, while the composition of communities within disk fallow and conventional plots could not be distinguished from each other. Diversity of soil fungal communities was significantly reduced following the cultivation of tomato in year four when compared to the diversity in plots where tomato was not cultivated. Divergence in the composition of soil fungal communities was observed following the cultivation of tomato under all land management regimes except organic, where communities continued to remained clustered based upon similarities among their ITS-1 amplicons. Divergence in the composition of fungal communities became more pronounced following two major hurricanes (Francis and Jeanne, September 2004) except for communities in the organic and pasture grass plots. Following the completion of a second tomato crop in year 5, genetic diversity and richness was similar under all land management regimes except the pasture grass, where it remained significantly higher. By contrast, following two consecutive years of tomato production, unique but mutually similar compositions of fungal communities were detected only in plots subjected to the organic land management regime. This was supported by observations that fungal communities were dominated by a 341 bp rDNA amplicon fragment in all land management regimes except the organic. Cloning and sequencing indicated that the 341 bp fragment generated by LH-PCR had a sequencing size of 343 bp, which was most closely related to Fusarium oxysporum. Thus, land management practices that disturb or disrupt soil fungal communities will significantly reduce their diversity. However, the composition of soil fungal communities is more strongly influenced by land management practices and communities within an organically management system were more resistant to anthropogenic and meteorological disturbances. (c) 2007 Elsevier Ltd. All rights reserved.