Drivers of genetic diversity in secondary metabolic gene clusters within a fungal species

Abstract

Filamentous fungi produce a diverse array of secondary metabolites (SMs) critical for defense, virulence, and communication. The metabolic pathways that produce SMs are found in contiguous gene clusters in fungal genomes, an atypical arrangement for metabolic pathways in other eukaryotes. Comparative studies of filamentous fungal species have shown that SM gene clusters are often either highly divergent or uniquely present in one or a handful of species, hampering efforts to determine the genetic basis and evolutionary drivers of SM gene cluster divergence. Here, we examined SM variation in 66 cosmopolitan strains of a single species, the opportunistic human pathogen Aspergillus fumigatus. Investigation of genome-wide within-species variation revealed 5 general types of variation in SM gene clusters: nonfunctional gene polymorphisms; gene gain and loss polymorphisms; whole cluster gain and loss polymorphisms; allelic polymorphisms, in which different alleles corresponded to distinct, nonhomologous clusters; and location polymorphisms, in which a cluster was found to differ in its genomic location across strains. These polymorphisms affect the function of representative A. fumigatus SM gene clusters, such as those involved in the production of gliotoxin, fumigaclavine, and helvolic acid as well as the function of clusters with undefined products. In addition to enabling the identification of polymorphisms, the detection of which requires extensive genome-wide synteny conservation (e.g., mobile gene clusters and nonhomologous cluster alleles), our approach also implicated multiple underlying genetic drivers, including point mutations, recombination, and genomic deletion and insertion events as well as horizontal gene transfer from distant fungi. Finally, most of the variants that we uncover within A. fumigatus have been previously hypothesized to contribute to SM gene cluster diversity across entire fungal classes and phyla. We suggest that the drivers of genetic diversity operating within a fungal species shown here are sufficient to explain SM cluster macroevolutionary patterns.

Fig 1. Genetic diversity of secondary metabolic gene clusters within a fungal species. The phylogeny was constructed using 15,274 biallelic SNPs with no missing data.

The tree is midpoint rooted and all branches with bootstrap support less than 80% are collapsed. This phylogeny does not include strains Af10, Af210, Z5, or RP-2014, as short-read data were not available. Superfixes following strain names indicate publications associated with DNA sequencing. * indicates strains sequenced in this study, † indicates strains sequenced at JCVI with no associated publication, and ‡ indicates strains sequenced by Rikenwith no associated publication. Heat maps show presence, absence, and polymorphisms in SM gene clusters. Black indicates the cluster is present in a strain with no polymorphisms, aside from missense variants, light gray indicates 1 or more genes in the cluster are pseudogenized, and dark gray indicates the cluster is partially or entirely absent (see Fig 2). Colors for cluster 4 indicate which pseudogenizing variants are present (see Fig 3) and colors for cluster 10 indicate which allele of the cluster is present (see Fig 4). Chromosomal location of clusters 1 and 33 are indicated. If more than one type of polymorphism is present within a cluster in a strain, only 1 is depicted. Types of polymorphisms found in each cluster are summarized below the cluster heat map. DHN, dihydroxynaphthalene; JCVI, J. Craig Venter Institute; NRPS, nonribosomal peptide synthase; PKS, polyketide synthase; SM, secondary metabolite; SNP, single nucleotide polymorphism.

Fig 2. Gene gains and deletions in SM gene clusters.

(A) Differences in gene content in SM gene cluster 14 in Aspergillus fumigatus strains and closely related species. Four A. fumigatus strains lack an 11-kb region in this cluster, including an NRPS backbone gene, highlighted in yellow. Regions upstream and downstream of this cluster are syntenic. LMB35Aa also contains a large inversion that moves a transcription factor, oxidoreductase, and hypothetical protein 275 kb away from the cluster. A. fischeri and A. lentulus, close relatives of A. fumigatus, contain a cluster lacking the 11-kb region. (B) SM gene cluster found in most A. fumigatus strains but absent from the Af293 reference and from the F7763 strain. EOS, end of scaffold; MFS, major facilitator superfamily; NRPS, nonribosomal peptide synthase; SM, secondary metabolite.

Fig 3. Pseudogenization in the fusarielin-like SM gene cluster.

(A) Positions of frameshift variants and nonsense variants in the fusarielin-like SM gene cluster 4. (B) Locus phylogeny of the fusarielin-like SM gene cluster based on a nucleotide alignment of the entire gene cluster, including intergenic and noncoding regions. The phylogeny is midpoint rooted and branches with bootstrap support <80% are collapsed. Two branches were shortened for visualization purposes. Strains with pseudogenizing variants are indicated with colored boxes. Colors correspond to variants shown in (A). PKS, polyketide synthase; SM, secondary metabolite; SNV, single nucleotide variant.

Fig 4. Six alleles of an idiomorphic SM gene cluster.

(A) Alleles of SM gene cluster 10 on chromosome 3. Red boxes denote transposable elements. Green arrows denote backbone genes (PKS or NRPS). (B) Locus phylogeny of conserved downstream of the idiomorph cluster (highlighted in gray in [A]). Phylogeny was constructed using a 48-mb nucleotide alignment with the GTRGAMMA model and midpoint rooted. Branches with bootstrap support <80% were collapsed. HR, highly reducing; NR, nonreducing; NRPS, nonribosomal peptide synthase; PKS, polyketide synthase; SM, secondary metabolite.

Fig 5. Multiple genomic locations of 2 SM gene clusters.

(A) SM gene cluster 1 (Afu1g00970-01010) and flanking region are found in different genomic locations. The flanking regions contain transposon-derived open reading frames, including 2 putative reverse transcriptases. In one strain, SM gene cluster 1 is found adjacent to SM gene cluster 30. (B) Synteny of A. fumigatus SM gene cluster 1 with clusters in Phaeosphaeria nodorum, Pseudogymnoascus pannorum, Escovopsis weberi, and Hypoxylon sp. CI4A. All species contain nonsyntenic genes predicted by antiSMASH to be part of a biosynthetic gene cluster. (C) SM gene cluster 33 (Afu5g00100-00135) is found in different genomic locations in different strains. In one strain, the cluster is adjacent to SM gene cluster 34. Multiple transposable elements flank the cluster in each strain. EOS, end of scaffold; FAD, flavin adenine dinucleotide; MFS, major facilitator superfamily; ORF, open reading frame; PKS, polyketide synthase; SM, secondary metabolite.

Fig 6. Types and frequencies of all SM gene cluster variants within A. fumigatus.

SM, secondary metabolite.