Genome of Diaporthe sp. provides insights into the potential inter-phylum transfer of a fungal sesquiterpenoid biosynthetic pathway

Abstract

Fungi have highly active secondary metabolic pathways which enable them to produce a wealth of sesquiterpenoids that are bioactive. One example is Δ6-protoilludene, the precursor to the cytotoxic illudins, which are pharmaceutically relevant as anticancer therapeutics. To date, this valuable sesquiterpene has only been identified in members of the fungal division Basidiomycota. To explore the untapped potential of fungi belonging to the division Ascomycota in producing Δ6-protoilludene, we isolated a fungal endophyte Diaporthe sp. BR109 and show that it produces a diversity of terpenoids including Δ6-protoilludene. Using a genome sequencing and mining approach 17 putative novel sesquiterpene synthases were identified in Diaporthe sp. BR109. A phylogenetic approach was used to predict which gene encodes Δ6-protoilludene synthase, which was then confirmed experimentally. These analyses reveal that the sesquiterpene synthase and its putative sesquiterpene scaffold modifying cytochrome P450(s) may have been acquired by inter-phylum horizontal gene transfer from Basidiomycota to Ascomycota. Bioinformatic analyses indicate that inter-phylum transfer of these minimal sequiterpenoid secondary metabolic pathways may have occurred in other fungi. This work provides insights into the evolution of fungal sesquiterpenoid secondary metabolic pathways in the production of pharmaceutically relevant bioactive natural products.

Keywords: Anticancer; Ascomycota; Biosynthesis; Evolution; Natural products; Sesquiterpene synthase.

Fig 1. Phylogenetic analyses of fungal isolate BR109

ITS sequencing shows that BR109 is most closely related to Diaporthe anacardii and Diaporthe foeniculacea.

Fig 2. GC/MS analysis of the terpenes/terpenoids produced by fungal isolate BR109

Peaks that could be assigned based on comparison of mass spectra to known libraries of terpenes are labelled with numbers. Numbering of peaks on the chromatogram correspond to the structures shown in Fig 3. Peaks that could not be assigned, but had m/z parent ions likely related to modified sesquiterpenoids, are labelled with an asterisk *.

Fig 3. Chemical structures of terpenoids produced by fungal isolate BR109

Identified terpenes/oids are numbered according to the peak labelling in Fig 2. Relative stereochemistries of structures are shown.

Fig 4. Phylogenetic analyses of the 17 putative sesquiterpene synthases Dia1–17 identified in the genome of fungal isolate BR109

Putative sesquiterpene synthases Dia1–17 cluster in distinct clades alongside other characterized sesquiterpene synthases from Basidiomycota and Ascomycota, apparently according to cyclization mechanism. Clusters that contain sesquiterpene synthases previously characterized by our group with known cyclization mechanisms are highlighted with a shaded oval. The sesquiterpene synthase that we characterize in this study, Dia1, is highlighted with a star. Dia1 clusters closely with other previously characterized Δ6-protoilludene synthases from Basidiomycota. Bootstrap values (1000 replicates) are shown for the first internal nodes (see Supporting Information).

Fig 5. Intron splicing patterns of Basidiomycota and Ascomycota genes encoding 1,11 cyclizing sesquiterpene synthases differ

(A) The gene architecture of Dia1 is compared to characterized Basidiomycota genes encoding 1,11 cyclizing sesquiterpene synthases Omp6, Omp7, Stehi25180, Stehi64702 and Stehi73029; as well as characterized Ascomycota genes encoding 1,11 cyclizing sesquiterpene synthases Ffsc4 and BcBOT2. The splicing patterns between characterized Basidiomycota and Ascomycota genes are different; Dia1 most closely matches the splicing pattern of Basidiomycota genes. (B) The gene architecture of Dia1 is compared to putative Ascomycota genes that may encode sesquiterpene synthases as determined by sequence comparisons. Terpene synthase gene names refer to those indicated in Fig 7. Exons are indicated by black rounded rectangles, introns are indicated by grey lines. The scale bar represents 200 bp.

Fig 6. GC/MS analysis of the volatile sesquiterpenes produced by E. coli overexpressing Dia1

Headspace analysis of E. coli cells overexpressing Dia1 reveals Δ6-protoilludene 6 as a major peak on the chromatogram, as well as minor peaks corresponding to pentalenene 17, 2-sterpurene 18, african-1-ene 19 and brasila-5(10),6-diene 20, which are all derived from a 1,11 cyclization of (2E,6E)-FPP. β elemene 8 is a heat induced rearrangment product of a 1,10 cyclization of (2E,6E)-FPP. Peaks that could be assigned based on comparison of mass spectra to known libraries of terpenes are labelled with numbers, and chemical structures are shown in the grey box. Indole, highlighted with an asterisk (*), is a breakdown product of tryptophan naturally produced by E. coli, and serves as an internal standard.

Fig 7. Putative sesquiterpene synthases and biosynthetic gene clusters associated with Δ6-protoilludene synthase Dia1 and its closely related Ascomycota homologs, and classification of the P450 most closely associated with these putative terpene synthases

Homologs of Δ6-protoilludene synthase Dia1 were identified by BLAST searches of Ascomycota genomes in JGI using a cutoff value of 1E-60. Putative terpene synthases are indicated by a red arrow, with the JGI identifier and species name provided above the arrow. Potential biosynthetic genes clustered around the terpene synthase were manually annotated using AUGUSTUS and BLAST. Identified were genes putatively encoding transporters, enzymes involved in scaffold modification, transcription factors, cytochrome P450 monooxygenases and oxidoreductases (colored arrows show their putative functions). Genes with unknown function or annotated as hypothetical are indicated by grey arrows. The numbers within arrows refer to gene model prediction from AUGUSTUS outputs. The P450 most closely associated with the putative terpene synthases (bold arrow outline) was categorized according to CYP family using Dr. Nelson’s P450 database and BLAST server (http://drnelson.uthsc.edu/CytochromeP450.html). Dia_g10 (3e-167), Diaam_g6 (1e-128), Melsp_g2 (2e-109), Thiar_g2 (5e-165) belong to CYP family 5032 (black arrow outline) associated with Basidiomycota, while the other P450s are most closely related to Ascomycota CYP families: Chagl_g1 (2e-143), Stagr_g6 (9e-172) belong to CYP family 65 (orange arrow outline); Conli_g2 (0), Ophpc_g1 (0) belong to CYP 582 (rose arrow outline); and Oidma_g6 (2e-52) to CYP 682 (pink arrow outline). E-values for top P450 hits are shown in parentheses as an indicator of significance of CYP family assignments (see Supporting Information for sequences).