Genomics-driven discovery of the pneumocandin biosynthetic gene cluster in the fungus Glarea lozoyensis

Abstract

Background: The antifungal therapy caspofungin is a semi-synthetic derivative of pneumocandin B0, a lipohexapeptide produced by the fungus Glarea lozoyensis, and was the first member of the echinocandin class approved for human therapy. The nonribosomal peptide synthetase (NRPS)-polyketide synthases (PKS) gene cluster responsible for pneumocandin biosynthesis from G. lozoyensis has not been elucidated to date. In this study, we report the elucidation of the pneumocandin biosynthetic gene cluster by whole genome sequencing of the G. lozoyensis wild-type strain ATCC 20868.

Results: The pneumocandin biosynthetic gene cluster contains a NRPS (GLNRPS4) and a PKS (GLPKS4) arranged in tandem, two cytochrome P450 monooxygenases, seven other modifying enzymes, and genes for L-homotyrosine biosynthesis, a component of the peptide core. Thus, the pneumocandin biosynthetic gene cluster is significantly more autonomous and organized than that of the recently characterized echinocandin B gene cluster. Disruption mutants of GLNRPS4 and GLPKS4 no longer produced the pneumocandins (A0 and B0), and the Δglnrps4 and Δglpks4 mutants lost antifungal activity against the human pathogenic fungus Candida albicans. In addition to pneumocandins, the G. lozoyensis genome encodes a rich repertoire of natural product-encoding genes including 24 PKSs, six NRPSs, five PKS-NRPS hybrids, two dimethylallyl tryptophan synthases, and 14 terpene synthases.

Conclusions: Characterization of the gene cluster provides a blueprint for engineering new pneumocandin derivatives with improved pharmacological properties. Whole genome estimation of the secondary metabolite-encoding genes from G. lozoyensis provides yet another example of the huge potential for drug discovery from natural products from the fungal kingdom.

Figure 1

Pneumocandin structures and morphology of Glarea lozoyensis. (a) Chemical structures of pneumocandins. (b) Colony of G. lozoyensis on malt yeast agar (left panel); conidiophores and conidia of G. lozoyensis (right panels).

Figure 2

Genome features of Glarea lozoyensis. (a) General genome features of G. lozoyensis. I, 22 scaffolds (> 2 kb); II, gene density represented as number of genes per 100 kb; III, percentage of coverage of repetitive sequences; IV, GC content was estimated by the percent G + C in 100 kb. (b) Functional classificaton of proteins in the G. lozoyensis genome based on InterproScan analysis.

Figure 3

Phylogenetic analysis of G. lozoyensis using ITS sequences (brackets on the left) and genome protein sequences (brackets on the right). Left tree: The topology was estimated using neighbor-joining method based on the ITS sequence data from Peláez et al., 2011 [26] and the selected fungi on the right-side of the graphic. Right tree: A maximum likelihood phylogenomic tree showing evolutionary relationship of G. lozoyensis with selected ascomycete fungal species. The tree was constructed from the concatenated amino acid sequences of 878 common orthologous genes (Additional file 2: Table S2). The phylogenetic position of G. lozoyensis wild-type strain ATCC 20868 is marked in red. Branch nodes with greater than 60% support from 1000 bootstrapped pseudoreplicates are indicted with red dots in both trees. Both trees were rooted with S. cerevisiae.

Figure 4

KEGG functional classification of proteins in the G. lozoyensis genome. Distribution of the predicted proteins were assigned by the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. The top four categories in the KEGG functional classification were Carbon Metabolism, Energy Metabolism, Amino Acid Metabolism, and Infectious Diseases.

Figure 5

CAZymes (carbohydrate-active enzymes) analysis in the G. lozoyensis genome and other fungi. (a) Total number of CAZymes in different fungi (Glarea lozoyensis, Neurospora crassa, Aspergillus nidulans, Trichoderma reesei, Fusarium graminearum, Aspergillus oryzae, Verticillium albo-atrum, Sclerotinia sclerotiorum, Glomerella graminicola, Ascocoryne sarcoides, Epichloë festucae, Tuber melanosporum, Laccaria bicolor, Piriformospora indica, and Saccharomyces cerevisiae). (b) Number of different family of CAZymes in different fungi. PL: polysaccharide lyase, GH: glycoside hydrolase, CBM: carbohydrate-binding module, CE: carbohydrate esterases, and GT: glycosyltransferase. See Additional file 2: Table S2 for a detailed tabular summary.

Figure 6

Domain prediction and phylogenetic analysis of polyketide synthases (PKSs) and polyketide synthases-nonribosomal peptide synthetase hybrids (PKS-NRPS hybrids) in G. lozoyensis and other characterized fungal PKSs. PKS and PKS-NRPS domains from G. lozoyensis were annotated by SMURF, anti-SMASH and SWISS-MODEL tools. SAT, starter unit acyltransferase domain; KS, ketosynthase domain; AT, acyltransferase domain; PT, product template domain; DH, dehydratase domain; ER, enoylreductase domain; KR, β-ketoacylreductase domain; MT, methyltransferase domain; ACP, acyl carrier protein; TE, thioesterase domain; A, adenylation domain; T, thiolation domain; C, condensation domain; R, reductive domain. Genealogy of PKSs and PKS-NRPSs was inferred by neighbor-joining analysis of the aligned amino acid sequences of the KS domains. Classification of PKSs and PKS-NRPSs sharing a common domain organization are highlighted by gray shading. Branch length indicates number of inferred amino acid changes. Red dots indicate branch nodes with >60% support. PKSs from G. lozoyensis are marked in red. See in Additional file 2: Table S3 for details of gene designations and their corresponding metabolites and references.

Figure 7

Schematic representations of the functional domains in nonribosomal peptide synthetase (NRPS) and polyketide synthase-nonribosomal peptide synthetase hybrid (PKS-NRPS hybrid) proteins in G. lozoyensis. A, adenylation domain; T, thiolation domain; C, condensation domain; E, epimerization domain; KS, ketosynthase domain; AT, acyltransferase domain; DH, dehydratase domain; ER, enoylreductase domain; KR, β-ketoacylreductase domain; MT, methyltransferase domain; ACP, acyl carrier protein; R, reductive domain. The pneumocandin-encoding NRPS (GLAREA10035 GLNPRS4) with six A-T-C modules is outlined.

Figure 8

Schematic representation of the pneumocandins and echinocandin B biosynthetic gene cluster. (a) Pneumocandins biosynthetic gene cluster. (b) Echinocandin B biosynthetic gene clusters (including L-homotyrosine biosynthetic gene cluster) based on [8,24].

Figure 9

Chemical and functional analysis of pneumocandins produced by wild-type and mutant strains of G. lozoyensis. (a) HPLC-MS profiles of chemical extracts from the wild-type (WT), glnrps4 deletion mutant, and glpks4 deletion mutant of G. lozoyensis. Full-scan + mode spectra were acquired in over a scan range of m/z 80–1,200. When grown in FGY broth, pneumocandin B₀ (peak 1 m/z = 1065) and pneumocandin A₀ (peak 2 m/z = 1079) were detected in the WT strain. Deletion of glnrps4 and glpks4 abolished pneumocandin B₀ and pneumocandin A₀ production in the mutant strains. (b) Antifungal activity of culture extracts from the WT and corresponding inactive extracts from glnrps4 and glpks4 mutants of G. lozoyensis. Purified pneumocandin B₀ (5 mg mL^-1) and DMSO (100%) were used as positive and negative controls, respectively.