Effects of long-term differential fertilization on eukaryotic microbial communities in an arable soil: a multiple barcoding approach

To understand the fine-scale effects of changes in nutrient availability on eukaryotic soil microorganisms communities, a multiple barcoding approach was used to analyse soil samples from four different treatments in a long-term fertilization experiment. We performed PCR amplification on soil DNA with primer pairs specifically targeting the 18S rRNA genes of all eukaryotes and three protist groups (Cercozoa, Chrysophyceae-Synurophyceae and Kinetoplastida) as well as the ITS gene of fungi and the 23S plastid rRNA gene of photoautotrophic microorganisms. Amplicons were pyrosequenced, and a total of 88706 quality filtered reads were clustered into 1232 operational taxonomic units (OTU) across the six data sets. Comparisons of the taxonomic coverage achieved based on overlapping assignment of OTUs revealed that half of the eukaryotic taxa identified were missed by the universal eukaryotic barcoding marker. There were only little differences in OTU richness observed between organic- (farmyard manure), mineral- and nonfertilized soils. However, the community compositions appeared to be strongly structured by organic fertilization in all data sets other than that generated using the universal eukaryotic 18S rRNA gene primers, whereas mineral fertilization had only a minor effect. In addition, a co-occurrence based network analysis revealed complex potential interaction patterns between OTUs from different trophic levels, for example between fungivorous flagellates and fungi. Our results demonstrate that changes in pH, moisture and organic nutrients availability caused shifts in the composition of eukaryotic microbial communities at multiple trophic levels.