Metagenomics Shines Light on the Evolution of "Sunscreen" Pigment Metabolism in the Teloschistales (Lichen-Forming Ascomycota)

Abstract

Fungi produce a vast number of secondary metabolites that shape their interactions with other organisms and the environment. Characterizing the genes underpinning metabolite synthesis is therefore key to understanding fungal evolution and adaptation. Lichenized fungi represent almost one-third of Ascomycota diversity and boast impressive secondary metabolites repertoires. However, most lichen biosynthetic genes have not been linked to their metabolite products. Here we used metagenomic sequencing to survey gene families associated with production of anthraquinones, UV-protectant secondary metabolites present in various fungi, but especially abundant in a diverse order of lichens, the Teloschistales (class Lecanoromycetes, phylum Ascomycota). We successfully assembled 24 new, high-quality lichenized-fungal genomes de novo and combined them with publicly available Lecanoromycetes genomes from taxa with diverse secondary chemistry to produce a whole-genome tree. Secondary metabolite biosynthetic gene cluster (BGC) analysis showed that whilst lichen BGCs are numerous and highly dissimilar, core enzyme genes are generally conserved across taxa. This suggests metabolite diversification occurs via re-shuffling existing enzyme genes with novel accessory genes rather than BGC gains/losses or de novo gene evolution. We identified putative anthraquinone BGCs in our lichen dataset that appear homologous to anthraquinone clusters from non-lichenized fungi, suggesting these genes were present in the common ancestor of the subphylum Pezizomycotina. Finally, we identified unique transporter genes in Teloschistales anthraquinone BGCs that may explain why these metabolites are so abundant and ubiquitous in these lichens. Our results support the importance of metagenomics for understanding the secondary metabolism of non-model fungi such as lichens.

Keywords: Lecanoromycetes; ABC-transporter; anthraquinone; biosynthetic gene cluster; fungal evolution; lichenized fungi.

Fig. 1.

ML tree including representatives of the largest class of lichenized fungi, the Lecanoromycetes (Ascomycota), with special focus on the order Teloschistales, reconstructed with IQTree using a 2,214 single-copy gene combined dataset. All internodes, except for the one connecting Umbilicaria muehlenbergii and U. vellea and the red branches in the family Parmeliaceae, were significantly supported by both bootstrap and posterior probabilities. Red branches indicate a conflict with the ASTRAL tree. Support values before and after the forward slash indicate gene and site concordance factors, respectively. Dotted tips show taxa sequenced and assembled in this study. Families/order are labeled and highlighted by alternating white and gray blocks. Ost. = Ostropales.

Fig. 2.

Diversity and distribution of secondary metabolite BGCs in the Lecanoromycetes. (a) Topology represents the ML tree shown in figure 1 with branches colored by taxonomic order as per figure 2e legend. (b) Presence–absence matrix of BGCFs in each genome (rows), showing only PKSI class BGCFs for readability. Each matrix column represents a single BGCF, with columns ordered by presence–absence similarity profiles across genomes. Dendrogram above matrix displays the BGCF similarity profile clustering produced with hclust function in R. Horizontal, orange dashes delineate taxonomic families. Presence = black, absence = white. (c) Stacked bar plot showing the number of predicted BGCs per genome. Colors represent different BGC classes. (d) Histograms showing the number of genomes per BGCF. The top histogram shows the BiG-SCAPE results using a strict cutoff of 0.46 to group BGCs into BCGFs and the bottom histogram showing the results using a more relaxed cutoff of 1. Beta diversity statistics at both cutoffs are shown on each graph and are separated into their turnover and nestedness components. (e) PCoA of all BGCF presence–absence profiles shown in (b). Colors represent taxonomic orders with ellipses drawn around orders with three or more points.

Fig. 3.

Putative anthraquinone BGCFs. (a) BiG-SCAPE similarity network linking BGCs (nodes) that showed a pairwise similarity score below the 0.46 cutoff established in this study. Node color shows to which BGCF each BGC was assigned by affinity clustering. Square nodes indicate MIBiG reference gene clusters that have been experimentally linked to a compound via genetic manipulation. (b) Presence–absence matrix of the four putative anthraquinone BGCFs plotted against the phylogenomic tree shown in figure 1. Orange dashed lines delineate lichen taxonomic families. (c) Zoomed in clade of the ML gene tree for the core PKS genes including the BGCFs shown in A and B. Colored dots at tips show which BGCF the 39 core PKS genes belong to, gray dots are sequences that do not belong to the four putative anthraquinone BGCFs. Full tree including all 340 sequences found for this orthogroup is depicted in the inset, where the clade of interest is highlighted with a gray rectangle.

Fig. 4.

(a) BGC showing the four common genes in Teloschistales putative anthraquinone BGCs. From left to right: ABC-transporter, MβL-TE, PKS (with SAT-KS-AT-ACP domains), and gene containing EthD domain. BGC shown is from the Variospora aurantia genome. (b) ML phylogeny reconstructed using the PT domain (Pfam: PF14765) and a HMM alignment. Bold branches indicate bootstrap support ≥95%. Alternating dark and light grey blocks highlight the PT classification of Liu et al. (2015). All lichenized-fungal putative anthraquinone sequences from this study fall within group V. Putative Lecanoromycetes anthraquinone sequences are those with orange text and sequence ADI24953.1. Green text signifies sequences from genes experimentally confirmed to produce anthraquinones.