An Unusual Skeletal Rearrangement in the Biosynthesis of the Sesquiterpene Trichobrasilenol from Trichoderma

Abstract The skeletons of some classes of terpenoids are unusual in that they contain a larger number of Me groups (or their biosynthetic equivalents such as olefinic methylene groups, hydroxymethyl groups, aldehydes, or carboxylic acids and their derivatives) than provided by their oligoprenyl diphosphate precursor. This is sometimes the result of an oxidative ring‐opening reaction at a terpene‐cyclase‐derived molecule containing the regular number of Me group equivalents, as observed for picrotoxan sesquiterpenes. In this study a sesquiterpene cyclase from Trichoderma spp. is described that can convert farnesyl diphosphate (FPP) directly via a remarkable skeletal rearrangement into trichobrasilenol, a new brasilane sesquiterpene with one additional Me group equivalent compared to FPP. A mechanistic hypothesis for the formation of the brasilane skeleton is supported by extensive isotopic labelling studies.

. Phylogenetic tree of TaTC6 and 886 putative fungal terpene synthases. The TaTC6 sequence is shown in Figure S4 and is published under accession number LC484924. Characterised enzymes [1][2][3][4][5][6][7][8] and their closest homologues with putatively the same function are shown in green and red. The branch of trichobrasilenol synthases is shown in purple. The enzymes characterised in this study are indicated by red asterisks. The scale bar displays substitutions per site.

In vivo functional analysis of TaTC6
To perform in vivo functional analysis of TaTC6, we used a heterologous expression system with an A. oryzae M-2-3 host [9] and a pUARA2 vector. [10] Primers taTC6_KpnI_Fw and taTC6_KpnI_Rv (Table S1) were used to amplify the region from the predicted start codon of the taTC6 gene to about 300 bp downstream of the predicted stop codon. The amplified DNA fragment was cloned into a pT7Blue T-vector (Takara Bio Inc., Tokyo, Japan) yielding plasmid pT7Blue_taTC6 and the sequence of the cloned gene was confirmed by sequencing. Plasmid pT7Blue_taTC6 was digested with KpnI and the target DNA fragment containing the TaTC6 sequence was ligated with KpnI-digested pUARA2. The resulting plasmid pUARA2_taTC6 was introduced into the A. oryzae M-2-3 host using the protoplast-polyethylene glycol method. [11] The transformant harboring pUARA2_taTC6 (A. oryzae M-2-3/pUARA2_taTC6) was precultured in PDB medium (2.4% Potato Dextrose Broth (Difco)) at 30°C for 3 days before growing in a 500 mL Erlenmeyer flask containing 100 mL of CD + starch medium (0.3 % NaNO3, 0.2 % KCl, 0.1 % KH2PO4, 0.05 % MgSO4·7H2O, 0.002 % FeSO4·7H2O, 2.0 % glucose, 2.0 % soluble starch, 1.0 % polypeptone) and incubated on a rotary shaker (160 rpm) at 30 °C. After 3 days of cultivation, acetone was added to the culture broth (2:1, v/v), followed by filtration and concentration under reduced pressure until most of the acetone was removed. The resulting aqueous phase was extracted with ethyl acetate and the organic layer was evaporated to dryness. GC-MS analysis of the crude extract revealed that the transformant A. oryzae M-2-3/pUARA2_taTC6 produces a specific metabolite, which is not detectable in the broth of A. oryzae M-2-3/pUARA2 ( Figure S2).

Cloning of the T. atroviride TaTC6 cDNA from A. oryzae M-2-3/pUARA2_taTC6
The freeze-dried mycelium (50 mg) of A. oryzae M-2-3/pUARA2_taTC6 was ground with a mortar and pestle. Sepasol ® -RNA I Super G (nacalai tesque, Kyoto, Japan) was added while freezing with liquid nitrogen. After the mycelium was thawed, chloroform was added, mixed and then allowed to stand at room temperature for 2 minutes, followed by centrifugation. (12,000 rpm, 4°C, 10 min). The aqueous layer was purified using Direct-zol TM RNA MiniPrep (ZYMO RESEARCH, Irvine, USA) according to the manufacturer's instructions. The DNase I treatment was performed on-column; 50 l of RQ1 RNase-free DNase (Promega, Tokyo, Japan) was added followed by incubation for 1 h at 37°C. For the reverse transcription, the Prime Script ® RT-PCR Kit (Takara Bio Inc., Tokyo, Japan) and the Oligo dT Primer packed together with the kit were used. To prepare a total RNA reverse transcription product, 1 μg of total RNA was used. Primers (RT_taTC6_Nhis_Fw and RT_taTC6_Nhis_Rv) were used to amplify the region from the predicted start codon to the predicted stop codon of the taTC6 gene from the total RNA reverse transcription product of the A. oryzae M-2-3/pUARA2_taTC6 transformant by PCR. Homologous recombination between the taTC6 gene and BamHI-HindIII-digested pHis8 [12] by the SLiCE method [13] yielded the plasmid pHIS8_taTC6. The nucleotide sequence of the taTC6 gene has been deposited in the DDBJ/EMBL/GenBank nucleotide sequence databases with the accession number LC484924 and is shown in Figure S4.

Incubation experiments with recombinant TaTC6
Incubation experiments were performed using freshly prepared enzyme solutions and FPP (final concentration 0.5mg mL -1 ) dissolved in substrate buffer (25 mm NH4HCO3). The enzyme solution was diluted with an equal volume of incubation buffer (50 mM TRIS, 10 mM MgCl2, 20 % glycerol, pH 8.2). The mixtures were shaken at 28 °C for 3 h. Small scale incubations using 1 mg substrate were extracted with 200 µL hexane, the organic phase was directly subjected to GC-MS. Incubation experiments for the preparative production of bungoene using 60 mg FPP were extracted with hexane (3 x 150 mL), the combined organic layers were dried with MgSO4, concentrated under reduced pressure and subjected to column chromatography on SiO2 (pentane:Et2O, 2:1). The desired fractions were combined and evaporated to give trichobrasilenol as colourless oil.

Plasmid construction for expression in Aspergillus oryzae
Fungal expression plasmid pTAex3 and Aspergillus oryzae NSAR1 were kindly provided by Prof. K. Gomi (Graduate School of Agricultural Sciences, Tohoku University) and Prof. K. Kitamoto (Graduate School of Agricultural Sciences, The University of Tokyo). Primers used for plasmid construction are listed in Table S1. PCRs were performed with Q5 polymerase (New England Biolabs, Ipswich, MA, USA) and ligations were performed with In-Fusion ® HD cloning kit (Clontech, Saint-Germain-en-Laye, France) according to the manuals provided by the manufacturers. All plasmids were constructed in Escherichia coli Stellar (Clontech, Saint-Germain-en-Laye, France) that was grown on LB medium with ampicillin. Transformations of A. oryzae were performed with the protoplast-polyethylene glycol method, as reported previously. [15] Candidate transformants were stabilized in proper selective agar medium (M agar with arginine and methionine for A. oryzae NSAR1-FPPS-OE: 0.

Detection of trichobrasilenol (6) production
A. oryzae NSAR1-FPPS-OE-TrTS was cultivated in DPY medium and incubated for 3 days at 30 °C. The mycelia were separated and dried and extracted with acetone under sonication. The extracts were subjected to GC/MS analysis ( Figure S15).
(i) Start with the 1,3-hydride migration from C9 to C6. The hypothetical product cation E has a cationic centre at the bridgehead carbon C9, making this reaction very unlikely. Its calculated structure (using the MM2 force field function of ChemBio3D Ultra 13.0) demonstrates that all three substituents at the cationic centre C9 point into one hemisphere. Moreover, the calculated structure of E suggests that the hydride migration should undergo with a concerted ring contraction by migration of the C6-C7 to a C6-C9 bond, equal to steps (ii) and (iv). The hypothetical product F, a cyclopropyl cation, should undergo immediate ring opening to a more stable allyl cation at C8-C7-C9 that can be captured by water at C8 to form 6. This analysis of the C-to-6 conversion as a stepwise process involves very unlikely cationic intermediates, but make the overall concerted reaction, theoretically separable into the three discussed elementary steps, very plausible. (ii) Start with breaking of the C6-C7 bond is not a possible reaction. (iii) Start with breaking of the C8-C9 bond must proceed with simultaneous making of a C6-C9 bond, as there is no other possibility where the electron density could migrate to. By this process a primary cation at C8 would be formed that could be avoided by simultaneous capture with water, step (vi), to form 14 that is not observed as an enzyme product (but it may be present among the unidentified minor side products). Compound 14 is no longer a reactive intermediate and it is difficult to understand its further conversion into 6. In conclusion, starting with step (iii) does not result in a good explanation for the formation of 6. (iv) Start with making of the C6-C9 bond could proceed like described for (iii). Alternatively, this could be realised by ring expansion of the cyclopropane to G, a cation that could potentially be trapped by water to form a sesquiterpene alcohol that was not observed, but may be present among the minor side products of the enzyme. This reaction does not explain the formation of 6 in which a C7=C9 double bond is needed. A third possibility is deprotonation from C9 with formation of a highly strained bicyclobutane moiety in 15. Its reprotonation at C6 must proceed with reintroduction of the proton abstracted from C9 and would lead to F that can rearrange to the allyl cation at C8-C7-C9 discussed in (i), followed by capture with water to 6. (v) Start by making of the C7-C9  bond is not a possible reaction. (vi) Start with the attack of water to C8 is only possible with simultaneous migration of the electron density in the C7-C8 bond to form a C6-C7  bond. This reaction can form compound 16 which may be among the minor side products of the enzyme reaction, but it does not explain the formation of 6.
In conclusion, the concerted reaction from C via D ‡ to 6 gives the best explanation. The theoretical separation into three steps as discussed in (i) makes this reaction more plausible, but the intermediates along this line are instable species that should be avoided as in the concerted mechanism.