Resistance Gene-Directed Genome Mining of 50 Aspergillus Species

Abstract

Fungal secondary metabolites are a rich source of valuable natural products, and genome sequencing has revealed a proliferation of predicted biosynthetic gene clusters in the genomes. However, it is currently an unfeasible task to characterize all biosynthetic gene clusters and to identify possible uses of the compounds. Therefore, a rational approach is needed to identify a short list of gene clusters responsible for producing valuable compounds. To this end, several bioactive clusters include a resistance gene, which is a paralog of the target gene inhibited by the compound. This mechanism can be used to identify these clusters. We have developed the FRIGG (fungal resistance gene-directed genome mining) pipeline for identifying this type of biosynthetic gene cluster based on homology patterns of the cluster genes. In this work, the FRIGG pipeline was run using 51 Aspergillus and Penicillium genomes, identifying 72 unique families of putative resistance genes. The pipeline also identified the previously characterized resistance gene inpE from the fellutamide B cluster, thereby validating the approach. We have successfully developed an approach to identify putative valuable bioactive clusters based on a specific resistance mechanism. This approach will be highly useful as an ever-increasing amount of genomic data becomes available; the art of identifying and selecting the right clusters producing novel valuable compounds will only become more crucial. IMPORTANCE Species belonging to the Aspergillus genus are known to produce a large number of secondary metabolites; some of these compounds are used as pharmaceuticals, such as penicillin, cyclosporine, and statin. With whole-genome sequencing, it became apparent that the genetic potential for secondary metabolite production is much larger than expected. As an increasing number of species are whole-genome sequenced, thousands of secondary metabolite genes are predicted, and the question of how to selectively identify novel bioactive compounds from this information arises. To address this question, we have created a pipeline to predict genes involved in the production of bioactive compounds based on a resistance gene hypothesis approach.

Keywords: Aspergillus; bioactive compounds; comparative genomics; fungal; genome mining; genomes; natural products; resistance; secondary metabolism.

FIG 1

Resistance mechanism. (A) Mycophenolic acid chemical structure and biosynthetic cluster with the known resistance gene mpaF (highlighted in red), which is an IMP dehydrogenase (IMPDH) inhibitor. (B) Chemical structure of fellutamide B and overview of the biosynthetic cluster, including the resistance gene inpE (highlighted in red), which is a proteasome inhibitor. (C) Illustration of the resistance mechanism used by some toxin producers. The secondary metabolite is a toxin which inhibits an essential enzyme, the target of the compound. Within the cluster responsible for producing the toxin, a copy of the target gene is found; this version is still functioning despite the compound’s presence and hence makes the organism self-resistant.

FIG 2

Overview of the pipeline illustrating the initial data of predicted secondary metabolite gene clusters and homologous protein families. In step 1, the initial data are combined to generate counts of how many homologs are found in each organisms and how many are within predicted clusters. In step 2, clusters are selected based on specific homology count patterns, either “strict,” where only one gene can have a homolog outside the cluster, or “alternative,” where genes belonging to large protein families are allowed. Step 3 is filtering of the selected clusters based on other clusters having a homologous resistance gene; this step is optional. In step 4, filtering based on the majority of the organisms should provide a homolog of the resistance/target genes. Step 5 is filtering based on the number of organism that should have only a single homolog.

FIG 3

Phylogenetic tree and principal-component analysis of protein family 596635, including the fellutamide B resistance gene. (A) Phylogenetic tree of the protein family (596635) containing the fellutamide B resistance gene (protein identification 2700) from A. nidulans. The phylogenetic tree is a maximum likelihood tree and was created with 500 bootstraps. (B) Principal-component analysis of protein family 596635 containing the fellutamide B resistance gene (protein identification 2700) from A. nidulans. The panels to the left are colored based on the sections to which the proteins belong, while the panels to the right are colored based on if the protein is a putative resistance gene (StrictClust), in a predicted cluster (Clust), a target not in a cluster (0), and the homolog to the putative resistance gene (putative target) (outsideSC).