Fig 1.
Biosynthetic pathways for the trichothecene analogs deoxynivalenol and T-2 toxin in Fusarium graminearum and F. sporotrichioides, respectively.
The inset at the top right shows the structure and numbering systems for trichodiene and 12,13-epoxytrichothec-9-ene (EPT).
Fig 2.
Representative diversity of trichothecene structures in fungi other than Fusarium.
See Fig 1 for examples of Fusarium trichothecenes.
Table 1.
Functions of trichothecene biosynthetic genes.
Table 2.
Information on fungal strains and genome sequences examined in the current study.
Fig 3.
Gene content and arrangement at TRI loci in fungi examined in this study.
The phylogenetic tree to the left was inferred by maximum likelihood analysis of concatenated nucleotide sequences of TRI3, TRI5, and TRI14, the only three TRI genes common to all the fungi. Numbers near branch nodes are bootstrap values based on 1000 pseudoreplicates. The four colored blocks (labeled A, B, C and D) indicate four lineages of TRI genes. The diagrams to the right show the content and arrangement of genes at TRI loci. Green arrows represent homologs of previously described TRI genes, and numbers within arrows indicate TRI gene designations (e.g., 14 indicates TRI14). TRI22 was originally described as TRI11 in Trichoderma, but here we consider TRI11 and TRI22 to be functionally and phylogenetically distinct genes (S1 Fig). For the purposes of this study, paralogs are indicated by the lowercase letters a, b, and c; e.g., TRI6 paralogs are indicated as 6a, 6b, and 6c. Gray arrows represent genes present in the TRI cluster of only one genus; orange arrows represent genes unique to Beauveria and Cordyceps; and purple arrows represent genes unique to Stachybotrys and Myrothecium (S2 Fig). Arrows overlaid on the same line indicate genes on the same contig, whereas arrows overlaid on different lines indicate genes on different contigs. For Stachybotrys, the numbers 7711 and 40293 above TRI paralogs indicate strains in which the paralogs occur. TRI3b in S. chartarum strain 40293 is truncated relative to other TRI3 homologs and, as a result, is likely nonfunctional.
Table 3.
TRI gene content of fungi examined in the current study.
A gray box indicates that a TRI gene is present in the genome of a fungus, while a white box indicates the gene was not detected. Numbers within boxes indicate the number of paralogs. The Greek letter ψ indicates that a large portion of the gene is present but that it is a pseudogene.
Fig 4.
Functional analysis of TRI3 in Trichoderma arundinaceum.
(A-C) High performance liquid chromatograms showing harzianum A (HA) production by: (A) wild-type progenitor strain Ta37; (B) tri3 mutant strain tri3.1; and (C) strain tri3.1.C1, tri3 mutant strain tri3.1 complemented with a wild-type copy of TRI3. (D) Quantification of HA production in the wild type and tri3 deletion mutant strains tri3.1, tri3.30, tri3.33, and tri3.48. (E) Quantification of HA production in the wild type, tri3 mutant strain tri3.1, and three tri3.1-derived strains that were complemented with a wild-type copy of TRI3 (strains tri3.1.C1, tri3.1.C4 and tri3.1.C5). (F and G) Total ion chromatograms from gas chromatography-mass spectrometry analysis of culture extracts of the (F) wild-type strain and (G) tri3 mutant strain tri3.1. The peaks labeled 4OH and ISD are for trichodermol (4-hydroxy EPT) and isotrichodiol, respectively.
Fig 5.
Functional analysis of TRI17 in Trichoderma arundinaceum.
(A) Quantitation of harzianum A production in the wild-type strain Ta37, tri17 mutant strain tri17.139, and three tri17.139-derived strains (tri17.C1, tri17.C4 and tri17.C10) that were complemented with a wild-type copy of T. arundinaceum TRI17. (B and C) Total ion chromatograms of culture extracts of (B) wild-type progenitor strain Ta37 and (C) tri17 deletion mutant tri17.139. The trichothecene biosynthetic intermediate trichodermol (4-hydroxy EPT) is indicated at 5.268 min. 4OH indicates trichodermol (4-hydroxy EPT).
Fig 6.
Total ion chromatograms from GC-MS analysis of cultures from heterologous expression of TRI genes from Beauveria bassiana strain ARSEF 2860.
(A) wild-type Fusarium verticillioides grown on YEPD medium containing trichodiene (TD); (B) F. verticillioides expressing the B. bassiana TRI4 grown on YEPD medium containing trichodiene; (C) ayt1 mutant of Saccharomyces cerevisiae grown on YG medium containing isotrichodermol (3-hydroxy EPT, 3OH); (D) ayt1 mutant of S. cerevisiae expressing the B. bassiana TRI101 grown on YG medium containing isotrichodermol; (E) wild-type F. verticillioides grown on YEPD medium containing isotrichodermin (3-O-acetyl EPT, 3OAc); (F) F. verticillioides expressing the B. bassiana TRI22 grown on YEPD medium containing isotrichodermin. In each chromatogram, the Y-axis is total ion abundance, and the X-axis is time in minutes. The peak labeled 3OAc,4OH indicates 4-hydroxy isotrichodermin (3-O-acetyl-4-hydroxy EPT).
Fig 7.
Maximum likelihood trees inferred from sequences of selected TRI genes: TRI3, TRI4, TRI5, TRI6, TRI14, and TRI22.
Trees for TRI10, TRI12 and TRI18 are shown in the Supporting Information (S8 Fig). Numbers near branch nodes are bootstrap values based on 1000 pseudoreplicates.
Fig 8.
Maximum likelihood trees inferred from predicted amino acid sequences of Tri17 (top) and Tri101 (bottom) and related homologs from trichothecene-nonproducing fungi.
The chemical structures shown to the right of the Tri17 tree are predicted structures of polyketides synthesized by the different Tri17 homologs. Asterisks (*) indicate species/strains that produce trichothecenes or are predicted to produce trichothecene based the presence of TRI genes. In the TRI101 tree, the gray boxes indicate the strains/species that have a TRI101 gene that functions or is likely to function in trichothecene biosynthesis. Numbers near branch nodes are bootstrap values based on 1000 pseudoreplicates. Strain designations are shown only for species with two or three strains included in a tree.
Fig 9.
Comparison phylogenetic tree inferred from concatenated sequences of TRI3, TRI5 and TRI14 (left) and a species phylogeny inferred from concatenated sequences of 20 housekeeping genes (right).
Numbers near branch nodes are bootstrap values from 1000 pseudoreplicates. Only bootstrap values greater than 70% are shown. The housekeeping genes used in this analysis are listed in S3 File.
Fig 10.
Scenarios for changes in trichothecene biosynthesis during evolutionary divergence of TRI cluster homologs.
The scenarios presume the ancestral TRI cluster and trichothecene pathway presented in Fig 12. In each scenario, the tree is a simplified version of the tree inferred from concatenated sequences of TRI3, TRI5, and TRI14 (Fig 3). A change in biosynthesis from the ancestral state is indicated by a change in color of a branch from black to blue. A. acquisition of C8 oxygen by gain of Tri1 in Fusarium and an unknown enzyme(s) in Microcyclospora and Trichothecium; B. acquisition of C15 oxygen by gain of Tri11 in Fusarium, Beauveria and Cordyceps, and an unknown enzyme in Myrothecium and Stachybotrys; C. acquisition of macrolide ring by gain of unknown enzymes in Myrothecium and Stachybotrys; D. change in C4 hydroxylation enzyme from Tri22 to Tri13 during divergence of Fusarium TRI cluster; E. change in Tri3 function from C4 acylation to C15 acetylation during divergence of Fusarium TRI cluster; and F. change in Tri4 function from 3 to 4 oxygenations during divergence of TRI cluster lineage in Fusarium, Beauveria, and Cordyceps.
Fig 11.
(A) Variation in the structure of polyketide-derived substituents in selected trichothecene analogs. The polyketide-derived substituent of each structure is highlighted with a gray background. Names below and/or to the right of highlighted areas are chemical or trivial names of the polyketide-derived substituents. Trichothecene names are indicated above each structure. (B) Proposed schemes for synthesis of octatrienedioate and 6,7-dihydroxy-2,4-octadienoate, the polyketide-derived substituents that occur in the structures of harzianum A and the macrocyclic trichothecene roridin A, respectively. In the proposed schemes, the parent compound of both substituents is 2,4,6-octatrienoate, an eight-carbon polyketide with three carbon-carbon double (enoyl) bonds. To form octatrienedioate, the polyketide would undergo a hydroxylation reaction followed by two oxidation reactions, first to form an aldehyde and second to form a carboxylic acid group. To form 6,7-dihydroxy-2,4-octadienoate, the polyketide would undergo two hydroxylation reactions during which an enoyl bond is lost. Formation of hexa-2,4-dienedioate, the polyketide-derived substituent in verrucarin A is not included in the figure, but would involve the same reactions as for synthesis of octatrienedioate, except that the parent polyketide would be a six-carbon polyketide with two enoyl bonds. It is not known whether the structural modifications of the polyketides occur before or after esterification to EPT. Thus, in the proposed schemes, R could be C4 of EPT, Co-enzyme A, or a hydrogen atom.
Fig 12.
(A) Proposed ancestral TRI cluster based on the distribution of TRI genes among fungi examined in the current study. (B) Proposed ancestral trichothecene biosynthetic pathway based on inferred gene content of the ancestral TRI cluster and knowledge of TRI gene functions. (C) The presence of two acetyl/acyl transferase genes in the proposed ancestral cluster raises the possibility of a branch at the end of the ancestral pathway that would result in the formation of two trichothecene analogs: 4-acetyl EPT and 4-butenoyl EPT.