Trichoderma reesei Contains a Biosynthetic Gene Cluster That Encodes the Antifungal Agent Ilicicolin H

The trili biosynthetic gene cluster (BGC) from the well-studied organism Trichoderma reesei was studied by heterologous expression in the fungal host Aspergillus oryzae. Coexpression of triliA and triliB produces two new acyl tetramic acids. Addition of the ring-expanding cytochrome P450 encoded by triliC then yields a known pyridone intermediate to ilicicolin H and a new chain-truncated shunt metabolite. Finally, addition of the intramolecular Diels-Alderase encoded by triliD affords a mixture of 8-epi ilicicolin H and ilicicolin H itself, showing that the T. reesei trili BGC encodes biosynthesis of this potent antifungal agent. Unexpected A. oryzae shunt pathways are responsible for the production of the new compounds, emphasising the role of fungal hosts in catalysing diversification reactions.


Introduction
Trichoderma reesei (Hypocrea jecorina) is a filamentous fungus widely used for the industrial production of cellulase [1]. The organism can grow to high cell density and can grow robustly on a variety of low-quality substrates. It has been subject to a wide variety of biotechnological improvements since the discovery of the wild-type QM6A strain in the 1940s, and many molecular tools are available for its genetic manipulation. Many Trichoderma species have been used in the remediation of fungal diseases. For example, Foster and coworkers have shown that some Trichoderma spp. can effectively treat Armillaria mellea (honey fungus) that causes root rot in many tree species [2]. T. reesei is also wellknown for the production of a number of secondary metabolites such as sorbicillinol 1 [3,4], and numerous multimeric derivatives such as bisvertinolone 2 [5].
However, despite decades of study, T. reesei itself has not been reported to produce any PKS-NRPS compounds. Since we have a broad interest in fungal PKS-NRPS compounds and we recently made detailed studies on the biosynthesis of sorbicillinoids in T. reesei [3,4], we set out to determine if this organism can also produce PKS-NRPS compounds.

Materials and Methods
All chemicals and media ingredients used in this work were purchased from Duchefa Biochemie, Roth, VWR, Fisher scientific, Sigma Aldrich, abcr and Formedium. All media, buffers, solutions and antibiotics were prepared with double-distilled water (ddH2O), where deionized water was further purified by GenPure Pro UV/UF Milipore device from Thermo scientific. All growth media and sterile solutions were autoclaved at 121 °C for 15 min using Systec VX150 or a classic Prestige Medical 2100 autoclave. Antibiotic solutions were sterilized using a sterile syringe filter (0.45 µm pore size, Roth). See ESI for details of all growth media, buffers and solutions, enzymes and antibiotics.

Microbiological Methods
Bacterial and fungal strains and all microbiological methods used in this work are summarized in the Electronic Supplementary Information (ESI).

Materials and Methods
All chemicals and media ingredients used in this work were purchased from Duchefa Biochemie, Roth, VWR, Fisher scientific, Sigma Aldrich, abcr and Formedium. All media, buffers, solutions and antibiotics were prepared with double-distilled water (ddH 2 O), where deionized water was further purified by GenPure Pro UV/UF Milipore device from Thermo scientific. All growth media and sterile solutions were autoclaved at 121 • C for 15 min using Systec VX150 or a classic Prestige Medical 2100 autoclave. Antibiotic solutions were sterilized using a sterile syringe filter (0.45 µm pore size, Roth). See ESI for details of all growth media, buffers and solutions, enzymes and antibiotics.

Microbiological Methods
Bacterial and fungal strains and all microbiological methods used in this work are summarized in the Electronic Supplementary Information (ESI).

Construction of Heterologous Expression Vectors
The genomic DNA (gDNA) of T. reesei QM6a ∆tmus53 ∆pyr4 [18] was used as a template to amplify genes for yeast homologous recombination (without removal of introns). The vector pEYA was used to assemble the PKS-NRPS (TRIREDRAFT_58285, triliA) gene (~12 kb) in four different DNA fragments of around 3 kb each using yeast homologous recombination. This was followed by LR recombination between the entry plasmid pEYA·triliA (MSIII139) and the expression vector pTYGS·arg to yield vector pTYGS·argB·triliA (MSIII144). The enoyl reductase gene (TRIREDRAFT_58289, triliB) was also amplified from the gDNA of T. reesei using primers with overhang to P gpdA and T gpdA . Using yeast homologous recombination, the final vector pTYGS·argB·triliA·triliB (MSIII152) was constructed which contains the PKS-NRPS (triliA)-and ER (triliB)-encoding genes from the specified cluster. A vector for the expression of the first three genes of the cluster; PKS/NRPS (TRIREDRAFT_58285, triliA), ER (TRIREDRAFT_58289, triliB), P450 RE (TRIRE-DRAFT_58953, triliC) was constructed using yeast homologous recombination using gDNA of T. reesei QM6a·∆tmus53·∆Pyr4 as a template for the amplification of the genes. A vector for the expression of the first four genes of the cluster; PKS/NRPS (TRIREDRAFT_58285, triliA), ER (TRIREDRAFT_58289, triliB), P450 RE (TRIREDRAFT_58953, triliC) and Diels-Alderase (triliD) was constructed via yeast homologous recombination using gDNA of T. reesei QM6a·∆tmus53·∆Pyr4 as template for the amplification of the genes.

Chemical Analysis
All the chemicals and materials were purchased from one of the following companies: Bio-Rad (München, Germany), New England Biolabs (Beverly, MA, USA), Roth (Karlsruhe, Germany), Sigma Aldrich (Steinheim, Germany), and Thermo Fisher Scientific (Waltham, MA, USA).
Single compounds were purified from the raw organic extracts by a Waters massdirected autopurification system. It comprises a Waters 2767 autosampler, a Waters 2545 binary gradient module system and a Phenomenex Kinetex Axia column (5 µm, C 18 , 100 Å, 21.2 mm × 250 mm) equipped with a Phenomenex Security Guard precolumn (Luna, C 5 , 300 Å). A water/acetonitrile gradient was run over 15 min with a flowrate of 20 mL/min and a post-column flow split of 100:1. The minority flow was applied for simultaneous analysis by a diode array detector (Waters 2998) in the range 210 to 1600 nm, an evaporative light-scattering detector (ELSD; Waters 2424) and a Waters SQD-2 mass detector, operating in ES+ and ES− modes between 100 and 1000 m/z. Selected peaks were collected into test tubes and solvent was evaporated under reduced pressure. Compounds were dissolved in methanol (1 mg/mL). High-Resolution Mass Spectrometry was performed on a Q-Tof Premier mass spectrometer (Waters) coupled to an Acquity UPLC-domain (Waters). Electron spray Ionisation (EsI) mass spectroscopy was measured in positive or negative mode depending on the compound.

Nuclear Magnetic Resonance (NMR) Analysis
Bruker Ascend 400 MHz, Bruker DRX 500 MHz or a Bruker Ascend 600 MHz Spectrometer (Bruker) were used for NMR measurements of the samples. Raw data were then analyzed using the software Bruker TopSpin 3.5. Chemical shifts are expressed in parts per million (ppm) in comparison to the Tetramethylsilane (TMS) standard and are referenced to the deuterated solvent.

Extraction of Fungal Cultures
For A. oryzae transformants, a small cell sample was dried by Büchner filtration or gravity filtration and used for gDNA analysis. Mycelia were ground using a hand blender and removed from the culture supernatant by Büchner filtration. Supernatant was acidified with 2 M HCL to pH 3-4 and extracted twice with ethyl acetate. Combined organic layers were dried over MgSO 4 and solvents were removed under vacuum. Organic residue was dissolved in methanol or acetonitrile to a concentration of 5-10 mg/mL (analytical) or 50 mg/mL (preparative) and filtered over glass wool.

Results
The genome of T. reesei QM6A, the wild-type isolate, was sequenced in 2008 [19]. We examined this genome sequence using a combination of automated AntiSMASH [20] and manual screening, with a particular focus on PKS and PKS-NRPS encoding genes. Bioinformatic analysis of all PKS and PKS-NRPS genes in T. reesei QM6a revealed the presence of 11 PKS and two hybrid PKS-NRPS encoding genes (see ESI for details). Of these, only the two PKS genes responsible for the sorbicillins and related compounds have been identified and confirmed (TRIREDRAFT_73618 and 73621) [4,21,22]. The HR-PKS TRIREDRAFT_65116 shows a very high similarity to the recently investigated PKS from Trichoderma virens, which is encoded in the biosynthetic gene cluster (BGC) that directs the production of virensols and trichoxide [23]. However, the other PKS and PKS-NRPS BGCs could not be linked to the production of any specific compound.
Although the T. reesei genome contains two hybrid PKS-NRPS encoding genes, no PKS-NRPS secondary metabolites have been reported from this fungus and these BGCs are likely to be silent under laboratory conditions. The two PKS-NRPS themselves show similarity (>60%) to proteins involved in fusarin C 8 biosynthesis in Fusarium moniliforme [13] and illicicolin H 10 biosynthesis in Penicillium variabile (= Talaromyces variabilis) [16]. Ilicicolin H 10 is an interesting antifungal agent that targets the cytochrome bc1 respiration complex in the 3-5 nM range [24], and it is conceivable that ilicicolin H 10 production may be responsible for the observed antifungal effects of Trichoderma species in general [2] and T. reesei in particular [25]. We therefore selected the putative T. reesei ilicicolin H BGC (trili) for further investigation.
Based on the NCBI prediction, the trili BGC ( Figure 2) consists of a hybrid PKS/NRPS (triliA), an ER (triliB), and a putative cytochrome P450 (triliC). Reannotation of the cluster by FungiSMASH and Clinker [26] and Artemis comparison to a closely related BGC in T. citrinoviride resulted in the identification of a putative S-adenosylmethionine (SAM)dependent Diels-Alderase in the cluster (triliD), that shows high similarity to iccD from the illicicolin-H BGC in Talaromyces variabilis [16]. Finally, triliE encodes a putative TIMlike epimerase homologous to IccE, known to epimerise epi-ilicicolin H in T. variabilis. Multigene analysis using cblaster [27] identified highly similar ili BGCs in other strains of T. reesei (RUT C-30) and other Trichoderma species, including T. parareesei, T. virens Gv29-8, T. citrinoviride TUCIM6016 and T. longibrachiatum ATCC 18648 (see ESI).  [4,21,22]. The HR-PKS TRIRE-DRAFT_65116 shows a very high similarity to the recently investigated PKS from Trichoderma virens, which is encoded in the biosynthetic gene cluster (BGC) that directs the production of virensols and trichoxide [23]. However, the other PKS and PKS-NRPS BGCs could not be linked to the production of any specific compound. Although the T. reesei genome contains two hybrid PKS-NRPS encoding genes, no PKS-NRPS secondary metabolites have been reported from this fungus and these BGCs are likely to be silent under laboratory conditions. The two PKS-NRPS themselves show similarity (>60%) to proteins involved in fusarin C 8 biosynthesis in Fusarium moniliforme [13] and illicicolin H 10 biosynthesis in Penicillium variabile (= Talaromyces variabilis) [16]. Ilicicolin H 10 is an interesting antifungal agent that targets the cytochrome bc1 respiration complex in the 3-5 nM range [24], and it is conceivable that ilicicolin H 10 production may be responsible for the observed antifungal effects of Trichoderma species in general [2] and T. reesei in particular [25]. We therefore selected the putative T. reesei ilicicolin H BGC (trili) for further investigation.
Based on the NCBI prediction, the trili BGC ( Figure 2) consists of a hybrid PKS/NRPS (triliA), an ER (triliB), and a putative cytochrome P450 (triliC). Reannotation of the cluster by FungiSMASH and Clinker [26] and Artemis comparison to a closely related BGC in T. citrinoviride resulted in the identification of a putative S-adenosylmethionine (SAM)-dependent Diels-Alderase in the cluster (triliD), that shows high similarity to iccD from the illicicolin-H BGC in Talaromyces variabilis [16]. Finally, triliE encodes a putative TIM-like epimerase homologous to IccE, known to epimerise epi-ilicicolin H in T. variabilis. Multigene analysis using cblaster [27]  In an initial heterologous expression experiment, triliA and triliB were cloned into the vector pTYGS·argB [28] [17]; Icc proteins are from the T. variabilis ilicicolin H biosynthetic pathway [16].
In an initial heterologous expression experiment, triliA and triliB were cloned into the vector pTYGS·argB [28] using rapid recombination of genome-derived PCR fragments in yeast and in vitro (Gateway) recombination. The PKS-NRPS-encoding gene triliA was placed downstream of the inducible A. oryzae amyB promoter (P amyB ), while triliB was placed under the control of the Aspergillus gpdA promoter (P gpdA ). The resulting vector pTYGS·argB·triliA·triliB was transformed into the fungal host Aspergillus oryzae NSAR1 [29]. Comparison of the organic extracts of transformants with those of untransformed A. oryzae NSAR1 revealed the production of new compounds by LC-MS analysis eluting at 6.7 min, 8.0 min and 11.0 min (Figure 3). [29]. Comparison of the organic extracts of transformants with those of untransformed A. oryzae NSAR1 revealed the production of new compounds by LC-MS analysis eluting at 6.7 min, 8.0 min and 11.0 min (Figure 3).
Compound 11 (RT 6.7 min, 30 mg) was isolated by preparative LC-MS, and the structure was elucidated using 1D and 2D NMR.  NMR analysis of these compounds (see ESI for full details) showed that 11 is a linear acyl tetramic acid with a C12 sidechain terminating in an unusual (at this position) carboxylic acid. Minor compound 12 was shown to be a related acyl tetramic acid featuring a decalin, with an unusual 3-propenoic acid substituent. Although the minor compound 13 at 11.0 min could not be fully identified by NMR, mass spectroscopic and UV analysis suggested that it is likely to be a third related tetramic acid with a C16 sidechain, first reported by Tang and coworkers during biosynthetic studies of ilicicolin H (Scheme 1) [16].
In the next round of experiments, triliC, encoding a cytochrome P450 oxidase, was added to the A. oryzae strains, driven by Padh. TriliC is related to ring-expanding cytochrome P450 oxidases first characterised from the tenellin pathway [15]. Coexpression of triliA, triliB and triliC resulted in production of two new compounds in comparison to the previous transformants and the untransformed A. oryzae (Figure 4), eluting at 6.7 (14), and 10.8 (15) min. Full NMR analysis of the RT 6.7 min compound showed it to be the pyridone NMR analysis of these compounds (see ESI for full details) showed that 11 is a linear acyl tetramic acid with a C 12 sidechain terminating in an unusual (at this position) carboxylic acid. Minor compound 12 was shown to be a related acyl tetramic acid featuring a decalin, with an unusual 3-propenoic acid substituent. Although the minor compound 13 at 11.0 min could not be fully identified by NMR, mass spectroscopic and UV analysis suggested that it is likely to be a third related tetramic acid with a C 16 sidechain, first reported by Tang and coworkers during biosynthetic studies of ilicicolin H (Scheme 1) [16]. In the next round of experiments, triliC, encoding a cytochrome P450 oxidase, was added to the A. oryzae strains, driven by P adh . TriliC is related to ring-expanding cytochrome P450 oxidases first characterised from the tenellin pathway [15]. Coexpression of triliA, triliB and triliC resulted in production of two new compounds in comparison to the previous transformants and the untransformed A. oryzae (Figure 4), eluting at 6.7 (14), and 10.8 (15) min. Full NMR analysis of the RT 6.7 min compound showed it to be the pyridone homolog of 11. Similarly, the 10.8 min compound was shown to be the pyridone homolog of 13.

PEER REVIEW
6 of 10 homolog of 11. Similarly, the 10.8 min compound was shown to be the pyridone homolog of 13. Finally, a gene homologous to Diels-Alderase iccD (triliD) was added to the expression system downstream of Peno and transformed into A. oryzae. This experiment again produced 11 (RT 6.7 min). Additionally, two new peaks in comparison to the previous transformants and the untransformed A. oryzae were also observed by LC-MS ( Figure 5) at RT 10.76 and 11.0 min. The minor compound at 11.0 min was again assigned as 13 after analysis of partial data. NMR and HRMS analysis (see ESI for full details) showed the peak at 10.76 min consists of an almost coeleuting mixture of the known 8-epi ilicicolin H 16 and ilicicolin H 10 itself. Finally, a gene homologous to Diels-Alderase iccD (triliD) was added to the expression system downstream of P eno and transformed into A. oryzae. This experiment again produced 11 (RT 6.7 min). Additionally, two new peaks in comparison to the previous transformants and the untransformed A. oryzae were also observed by LC-MS ( Figure 5) at RT 10.76 and 11.0 min. The minor compound at 11.0 min was again assigned as 13 after analysis of partial data. NMR and HRMS analysis (see ESI for full details) showed the peak at 10.76 min consists of an almost coeleuting mixture of the known 8-epi ilicicolin H 16 and ilicicolin H 10 itself.

Discussion
PKS-NRPS compounds have been reported before from the genus Trichoderma, but T. reesei QM6A itself has not been reported to produce these compounds. Genomic analysis, however, reveals two typical fungal PKS-NRPS BGCs. Heterologous expression experiments reveal that the trili BGC is fully functional, but apparently silent in T. reesei QM6A under laboratory conditions. Initial expression of the PKS-NRPS (triliA) and trans-ER (triliB) genes in A. oryzae led to production of two new acyl tetramic acids (11 and 12) and a third compound 13 that is likely to be a precursor of illicicolin H, recently described by Tang and coworkers [16] from Nectria sp. B13 and T. variabilis and Gao and coworkers [17] in Neonectria sp. DH2. Decalin 12 appears to have been oxidised to a carboxylic acid at its terminal polyketide methyl group. Addition of the ring-expanding P450 oxidase then produces the corresponding 2-pyridones 14 and 15. Compound 15 is identical to an ilicicolin H precursor previously identified by Tang and coworkers [16], but 14 is new. Finally, addition of the Diels-Alderase encoded by triliD produces a mixture of 8-epi ilicicolin H 16 and ilicicolin H 10 itself.
The T. variabilis icc [16] and Neonectra DH2 ili [17] BGCs were previously investigated by expression in A. nidulans. Our results are highly similar to those obtained from these studies, except that A. oryzae as a host appears to catalyse some unexpected steps. For example, in A. oryzae the initial tetramic acid precursor 13 appears to be oxidatively degraded to remove the terminal four polyketide carbons and leaves a carboxylic acid in the structure of 11. Previous heterologous expression studies on A. oryzae also indicate that this host has a propensity for the oxygenation of polyenes [30,31] and oxidative cleavage of olefins [32]. Likewise, oxygenation of the terminal methyl of 12 to a carboxylic acid is newly observed in A. oryzae. This is similar to the BueE-catalysed oxidation recently observed by Piggott and coworkers in the case of the fungal polyketide decalin burnettiene A in Aspergillus burnettii [33]. It also appears that a spontaneous (or A. oryzae-catalysed) Diels-Alder reaction can occur to produce a decalin that is then oxidised to a carboxylic acid at the terminal methyl to generate new compound 12. In previous studies on ilicicolin H biosynthesis, using A. nidulans as the host, a specific Diels-Alder catalyst was required for this step. We also observed the chain-truncated pyridone 14-again a new natural product. In A. nidulans it appears that 8-epi ilicicolin H 16 is a stable compound, requiring expression of iccE for the conversion to ilicicolin H. However, in A. oryzae, observable amounts of ilicicolin H 10 are produced in the absence of this epimerase catalyst.
Our results therefore illustrate that T. reesei QM6A has a fully functional PKS-NRPS BGC that encodes the biosynthesis of ilicicolin H, a potent antifungal agent. The ability of T. reesei QM6A to make this compound has never before been observed [34], despite nearly 70 years of laboratory-based investigations. It is likely that this BGC can be activated under unknown environmental conditions in its native habitat and may contribute to its utility as an agricultural antifungal agent, but it appears to be silent under laboratory conditions. Finally, our results also show that different fungal heterologous hosts can introduce unexpected shunt pathways which can diversify known secondary metabolites and produce new-to-nature compounds.