A novel metabarcoding approach to investigate Fusarium species composition in soil and plant samples

The genus Fusarium contains more than 300 species, most of which are plant pathogens. Appropriate molecular tools for accurately and rapidly describing temporal and spatial shifts in Fusarium communities would be useful for the development of control strategies. Here, we present a new Fusarium-specific primer pair targeting the translation elongation factor 1-α (EF1α) gene with amplicons of ~430 bp, suitable for MiSeq metabarcoding sequencing. Mock Fusarium communities were used to evaluate its resolution and to optimize read filtering and downstream analyses. The use of the DADA2 pipeline coupled with operational taxonomic unit (OTU) picking at 98% similarity cut-off significantly increased the accuracy of read filtering. Building a phylogenetic tree using a manually curated database as a reference allowed taxonomic assignment at the species or species-complex level. This methodology was tested on soil and maize residue samples collected from crop fields. Up to 18 Fusarium OTUs, belonging to 17 species and 8 species complexes, were obtained, with F. oxysporum being the most abundant species in soil samples, while F. graminearum and F. avenaceum were the most abundant in maize residues. We demonstrated the high performance of this workflow which could be further used for profiling Fusarium species composition and dynamics during the cultivation cycle.

Combined Metabarcoding and Co-occurrence Network Analysis to Profile the Bacterial, Fungal and Fusarium Communities and Their Interactions in Maize Stalks

Fusarium Head Blight (FHB) is one of the most devastating diseases of cereals worldwide, threatening both crop production by affecting cereal grain development, and human and animal health by contaminating grains with mycotoxins. Despite that maize residues constitute the primary source of inoculum for Fusarium pathogenic species, the structure and diversity of Fusarium spp. and microbial communities in maize residues have received much less attention than in grains. In this study, a metabarcoding approach was used to study the bacterial, fungal and Fusarium communities encountered in maize stalks collected from 8 fields in Brittany, France, after maize harvest during fall 2015. Some predominant genera found in maize residues were cereal or maize pathogens, such as the fungal Fusarium, Acremonium, and Phoma genera, and the bacterial Pseudomonas and Erwinia genera. Furthermore, a high predominance of genera with previously reported biocontrol activity was found, including the bacterial Sphingomonas, Pedobacter, Flavobacterium, Pseudomonas, and Janthinobacterium genera; and the fungal Epicoccum, Articulospora, Exophiala, and Sarocladium genera. Among Fusarium spp., F. graminearum and F. avenaceum were dominant. We also found that the maize cultivar and previous crop could influence the structure of microbial communities. Using SparCC co-occurrence network analysis, significant negative correlations were obtained between Fusarium spp. responsible for FHB (including F. graminearum and F. avenaceum) and bacterial OTUs classified as Sphingomonas and fungal OTUs classified as Sarocladium and Epicoccum. Considering that isolates belonging to these taxa have already been associated with antagonist effect against different Fusarium spp. and/or other pathogenic microorganisms and due to their predominance and negative associations with Fusarium spp., they may be good candidates as biocontrol agents. Combining the use of Fusarium-specific primers with universal primers for bacteria and fungi allowed us to study the microbial communities, but also to track correlations between Fusarium spp. and other bacterial and fungal genera, using co-occurrence network analysis. Such approach could be a useful tool as part of a screening strategy for novel antagonist candidates against toxigenic Fusarium spp., allowing the selection of taxa of interest.