PNNL

Abstract

Section Flavi encompasses both important harmful and beneficial Aspergillus species. The best known are A. oryzae used in food fermentation and enzyme production and A. flavus, a food and feed spoiler and mycotoxin producer. Here, we have de novo sequenced 19 genomes spanning the breadth of the Flavi section and made comparisons of 31 fungal genomes including 23 section Flavi species to investigate the genomic diversity of this section. Genome analysis shows a high diversity, and identifies islands of diversity in the genomes, including sub-telomeric regions, as the major loci of genetic diversity. Genome-scale phylogeny across the section shows that the closest relative of A. oryzae is not A. flavus, but A. minisclerotigenes or A. aflatoxiformans. Addressing food fermentation and industrial biotechnology, we have predicted 13,759 CAZymes (598/species) in section Flavi. The CAZyme content was compared with growth experiments on 35 media showing that the variations in gene content are not necessarily reflected in the growth patterns. Of particular interest is CAZyme family GH28, which is relevant to food fermentation and generally highly expanded in section Flavi, suggesting a part of the success of this section for traditional food production. Addressing fungal safety and potentials for discovery of bioactive compounds, we identified 1,606 secondary metabolite gene clusters in section Flavi, (73/species). This is a very high abundance and shows that section Flavi is much more prolific than comparable species such as Aspergillus section Nigri and Penicillium species. Each species has on average >5 clusters found in none of the other species; in total 130 gene clusters are only found in a single species. We have further assigned clusters to 20 types of compounds across 29 species. E.g. for the highly toxic aflatoxin, the gene cluster is highly conserved in 14 section Flavi species, a truncated version is found in A. caelatus, and a complete loss has happened in A. tamarii. By coupling the cluster predictions to chemical data, we have identified a putative miyakamide biosynthetic gene cluster demonstrating one of the versatile possibilities unlocked from this data. In summary, we have generated a highly valuable and versatile resource and used it to demonstrate the diversity and similarities within section Flavi both in terms of CAZyme content and secondary metabolite potential.