Focused Genome Mining of Structurally Related Sesterterpenes

Nov 29, 2017 - Division of Chemistry, Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan. ‡ Elements Chemistry ... Faculty of ...
7 downloads 13 Views 674KB Size
Download PDF
Letter pubs.acs.org/OrgLett

Focused Genome Mining of Structurally Related Sesterterpenes: Enzymatic Formation of Enantiomeric and Diastereomeric Products Koji Narita,† Hajime Sato,‡,§,∥ Atsushi Minami,*,† Kosei Kudo,† Lei Gao,†,⊥ Chengwei Liu,† Taro Ozaki,† Motoichiro Kodama,# Xiaoguang Lei,⊥ Tohru Taniguchi,∇ Kenji Monde,∇ Mami Yamazaki,∥ Masanobu Uchiyama,‡,§ and Hideaki Oikawa*,† †

Division of Chemistry, Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan Elements Chemistry Laboratory, RIKEN, and RIKEN Center for Sustainable Resource Science (Wako campus), Saitama 351-0198, Japan § Graduate School of Pharmaceutical Sciences, The University of Tokyo, Tokyo 113-0033, Japan ∥ Graduate School of Pharmaceutical Sciences, Chiba University, Chiba 260-8675, Japan ⊥ Department of Chemical Biology, College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China # Faculty of Agriculture, Tottori University, Tottori 680-8553, Japan ∇ Faculty of Advanced Life Science, Frontier Research Center for Post-Genome Science and Technology, Hokkaido University, Sapporo 060-0810, Japan ‡

S Supporting Information *

ABSTRACT: Heterologous expression of four clade-A bifunctional terpene synthases (BFTSs), giving di/sesterterpenes with unique polycyclic carbon skeletons such as sesterfisherol, enabled the isolation of the sesterterpenes Bm1, Bm2, Bm3, and Pb1. Their structures suggested that formation of the products occurs via various diastereomeric carbocation intermediates, allowing the proposal that clade-A BFTSs catalyze three-step cyclizations using several stereofacial combinations of allylic cation−olefin pairs in the corresponding intermediates to generate various stereoisomers.

B

ifunctional terpene synthases1 (BFTSs), which have two catalytic domains, a C-terminal prenyltransferase (PT) domain and an N-terminal terpene synthase (TS) domain, are a unique family of terpene synthases found in various species of fungi. The PT domain provides either a C20- or a C25prenyldiphosphate, while the TS domain catalyzes diphosphatecleavage-initiated cascade reactions to produce unique molecular skeletons, including eight-membered rings as well as macrocyclic and polycyclic systems. The characteristic domain organization enabled us to identify the genes for ophiobolin F synthase (AcOS) and sesterfisherol synthase (NfSS) by genome mining (Figure 1),2,3 which has been successfully used to characterize the function of terpene synthases and has been further developed.4,5 Considering the differences in the cyclization mechanisms and the amino acid sequences of several functionally characterized BFTSs, such as NfSS (clade A),3 PaFS (B),1 PaPS (B),6 and AcOS (B),2 we proposed that the initial cyclization mode is defined by the amino acid sequence of the TS domain of the BFTS (Figure S1).3 NfSS catalyzes A-ring formation at the C1−C15/C14−C18 positions (C1−IV−V) to give a 5−15 ring system, while the latter three enzymes catalyze A-ring formation at the C1−C11/C10−C14 positions (C1−III−IV) to yield a 5−11 ring system. Recently, additional BFTSs, including EvAS (B)7 and EvSS (B),8 which give di/sesterterpenes, have been characterized (Figure 1). © 2017 American Chemical Society

Figure 1. Functionally characterized bifunctional terpene synthases found in fungi.

The clade-A TS product sesterfisherol (1) and its oxidized analogue sesterfisheric acid are sesterterpenes possessing a 5− Received: November 2, 2017 Published: November 29, 2017 6696

DOI: 10.1021/acs.orglett.7b03418 Org. Lett. 2017, 19, 6696−6699

Letter

Organic Letters 6−8−5 tetracyclic ring system.3 Similar carbon skeletons can be found in the structures of sesterterpenes not only from fungi (terpestacin9) but also from lichen (retigeranic acids10) and plants (nitidasin11). Given their broad range of biological activities, including antiangiogenic activity and the inhibition of glycosylphosphatidylinositol anchor biosynthesis, the sesterterpene family is an attractive target for genome mining. However, the biosynthetic genes and pathways, except for those of sesterfisheric acid, remain to be elucidated. In this study, we examined genome mining, focusing on clade-A BFTSs to explore sesterfisherol-related natural products. Four BFTSs (BmTS1, BmTS2, BmTS3, and PbTS1) were functionally characterized by heterologous expression in Aspergillus oryzae and Escherichia coli. Phylogenetic analysis of the TS domains of BFTSs found in the public database allowed us to categorize more than 100 BFTSs into six clades (A−F; Figure S1). Among them, clades A−C appeared to be the most abundant, accounting for more than 65% of the total. To explore terpene synthases producing sesterfisherol-related sesterterpenes, we selected four BFTS genes, BmTS1, BmTS2, and BmTS3 from Bipolaris maydis ATCC48331 and PbTS1 from Phoma betae PS-13, which showed moderate identities (35−38%) with NfSS. Applying the same strategy as for functional analysis of NfSS,3 we initially prepared the A. oryzae transformant AO-BmTS1/2/3 by introducing a pair of plasmids, pUARA2-BmTS1/2 and pUSA2-BmTS3/4 (see the Supporting Information). GC−MS analysis of the metabolites prepared from this transformant showed two peaks corresponding to Bm1 and Bm2 with molecular ion peaks M+ at m/z 340 and the peak of Bm3 with M+ at m/z 358 (Figures 2 and S2). To identify the genes

From the HRMS data for the four products, three were proposed to be sesterterpene hydrocarbons (Bm1, Bm2, and Pb1: C25H40) and the fourth to be a sesterterpene alcohol (Bm3: C25H42O). Their 1H NMR spectra showed characteristic C24,C25-doublet and C23-singlet methyl signals (Table S1), which exhibited strong correlations with the C19/C18 and C14/C15/C16-carbon signals, respectively, in their HMBC spectra. In addition, NOEs were observed between the C23-Me signal and the C25/C24-Me or H19 in Bm1/Bm2/Pb1 and between H18 and H14 in Bm1/Bm3 (Figure S4). These results showed that all four structures contain a trans-fused cyclopentane moiety (A ring) with an isopropyl group at C18 (Table S1), showing a trans−trans relationship at the C14, C15, and C18 chiral centers. This strongly suggested that the four cladeA BFTSs catalyzed cyclization using the same mechanism (C1−IV−V). Considering the A-ring characterization and the unsaturation degree (u.s.) of 5 for Bm3, together with the presence of six olefinic carbon signals, three allylic Me signals, and three olefinic proton signals in its 1H and 13C NMR spectra, we speculated that Bm3 is a simple macrocyclic alcohol with a trans-fused cyclopentane moiety and three trisubstituted olefins. Similarly, the NMR spectra of Bm1 and Pb1 showed that these products have a tricyclic system with three olefins (CH ×3, C ×3) on the basis of the u.s. of 6. On the other hand, in the NMR spectra of Bm2, four olefinic carbon signals were observed but the olefinic proton was not, indicating that Bm2 has a tetracyclic system with two tetrasubstituted olefins on the basis of the u.s. of 6. In addition to the C24,C25-doublet Me signals, other doublet Me signals were observed in the 1H NMR spectra of Bm1 and Pb1 (doublet Me ×1) and Bm2 (doublet Me ×2), implying that they contain a second cyclopentane moiety on the basis of the cyclization mechanism. Together with the data described above, extensive NMR analysis (Figure S4), including 2D NMR spectra, enabled us to determine the planar structures of Bm1, Bm3, and Pb1, as shown in Figure 3. On the basis of fairly rigid conformations originating from the conjugated diene system in Bm1 and the highly strained C5−

Figure 2. GC−MS profiles of metabolites from (A) A. oryzae wild-type strain, (B) AO-BmTS1/2/3, (C) AO-BmTS1/2/3, (D) AO-PbTS1, (E) E. coli wild-type strain, (F) EC-BmTS1, (G) EC-BmTS2, and (H) EC-BmTS3.

responsible for production of these sesterterpenes, we prepared E. coli transformants by introducing cDNA sequences from individual BFTSs prepared from AO-BmTS1/2/3. GC−MS analysis of the metabolites from each transformant showed that Bm1, Bm2, and Bm3 were produced by EC-BmTS1, ECBmTS2, and EC-BmTS3, respectively (Figures 2 and S3). We isolated the two major products, Bm1 and Bm3, from AOBmTS1/2/3. Because of the low yield of the minor product, Bm2, we obtained Bm2 from large-scale incubation of ECBmTS2. Similarly, we obtained another product, Pb1, from AOPbTS1 after introducing the plasmid pTAex3-PbTS1 (Figure 2).

Figure 3. (A) Structures of cyclization products from the reactions with various C25-terpene synthases and chemical conversions of 1 and Bm2. (B) Ozonolysis of the BFTS products. 6697

DOI: 10.1021/acs.orglett.7b03418 Org. Lett. 2017, 19, 6696−6699

Letter

Organic Letters Scheme 1. Proposed Cyclization Mechanism Forming the A Ring of the BFTSs Productsa

a

The structures in brackets are conformers of GFPP in the active site of BFTSs. Natural C25-terpenes having the corresponding moieties are depicted.

Scheme 2. Proposed Cyclization Mechanism of Various Clade-A Sesterterpene Synthases

and optical rotation values identical to those of the sample prepared from Pb1, suggesting that they have the same configuration. Conversion of Bm3 into terpestacin utilizing the A. oryzae transformant with three additional modification enzyme genes (Scheme S1) supported that the absolute configuration of Bm3 was 10S,14R,15S,18S.13 Taken together, these results established that the absolute configuration at the A ring of Bm1 is opposite to that of Bm3/Pb1. We recently reported the detailed cyclization mechanism of the clade-A BFTS NfSS based on DFT calculations of the formation of 1.14 The theoretically predicted mechanism revealed two thermodynamically and kinetically favorable cascade cyclization reactions (Path a: A ring−D ring−B/C rings; Path b: A ring−B ring−C/D rings) (Scheme S2). Point mutation of NfSS (F191A) showed aberrant products derived from the premature quenching of the carbocation present only in Path a. Isolation of four sesterterpenes possessing 5−15 bicyclic (Bm3), 5−12−5 tricyclic (Bm1 and Pb1), and 5−6− 8−5-tetracyclic (Bm2) ring systems suggested that these are produced via cyclization similar to that proposed in Path a. Intriguingly, the absolute configuration of the clade-A BFTS products revealed that NfSS-type BFTSs and BmTS3-type BFTSs catalyze the enantiofacially opposite cyclizations (Scheme 1 and Figure S8). These results suggested that the

C6 double-bond geometry in Pb1, their relative configurations were determined through the analysis of the corresponding NOESY spectra (Figure S4). During spectral analysis of Bm2, we found that a major constituent, Bm2A, was slowly converted into a hardly separable isomer, Bm2B, at room temperature (Figure S5). The similarity of the 1H NMR spectra of Bm2 and 1 indicates that Bm2 is a dehydration product of 1. Actually, treatment of 1 with p-TsOH in acetone gave a mixture of Bm2A and Bm2B. Finally, the structure of Bm2A was determined by NMR analysis of the corresponding epoxide 2, which was obtained by treatment with mCPBA (Figure S4). The structure of Bm2A suggested that the isomeric pair is interconvertible by a 1,5sigmatropic rearrangement. From the chemical correlations with 1, we deduced the absolute configuration of Bm2A, as shown in Figure 3. To determine the absolute configuration of the A ring securely, ozonolysis of Bm1 followed by reductive workup with NaBH4 was employed to give diols 3a and 3b (Figures 3 and S7). From the conformers of 3 proposed by DFT calculations, the absolute stereochemistry of 3 was determined by vibrational circular dichroism (VCD) (Figure S6),12 indicating that the absolute stereochemistry of Bm1 was 6R,7R,14R,15R,18R. Bm3 was ozonolyzed to give diol 4 (Figure 3), which had 1H NMR 6698

DOI: 10.1021/acs.orglett.7b03418 Org. Lett. 2017, 19, 6696−6699

Letter

Organic Letters Notes

former BFTSs might stabilize the transition state giving the 14S,15R,18R product while the latter might stabilize the transition state giving the enantiomeric product. We also found that all of the isolated products were closely related to carbocation intermediates in the NfSS cyclization pathway: the enantiomeric intermediate of IM4, ent-IM4, which may be involved in the production of Bm3; the diastereomeric intermediates ent-11Z-6-epi-IM6a for Pb1 and 6-epi-IM8a for Bm1; and the intact intermediate IM10a for Bm2 (Scheme 2). These results suggested that those BFTSs, including NfSS, follow Path a to construct the sesterterpene skeletons via stereoisomeric carbocation intermediates and eventually generate enantiomeric and diastereomeric products. Recently, plant sesterterpene synthases (PSTSs) from Arabidopsis thaliana have been reported to biosynthesize various molecular skeletons such as thalianatriene/arathanatriene and retigeranine B.15 The close structural relationship between plant and fungal sesterterpenes suggests that the cyclization mechanisms of PSTSs are similar to those of the NfSS reaction, although PSTSs do not have a PT domain and show low homology to fungal BFTSs. This was recently confirmed by the functional analysis of the PSTS AtTPS19.15b In addition to PSTS products, we can find many examples of natural diastereomeric sesterterpene metabolites. For the formation of stereoisomers at the A ring via Path a, a similar cyclization mechanism may be involved in the biosynthesis of retigeranic acid B (trans−trans),10b retigeranic acid A10a and aleurodiscal16 (diastereomeric trans−cis), and terpestacin9/ nitidasin11 (enantiomeric trans−trans) (Scheme 1). In this work, we conducted a genome mining study of four sesterterpenes by focused expression using the highly reliable A. oryzae expression system. Detailed structural analysis of the four products showed that the clade-A BFTSs generated structurally related products that originated from the diastereomeric intermediates in the NfSS reaction. These results suggested that the enantio- or diastereofacially selective three-step cyclizations forming bi-, tri-, and tetracyclic skeletons provide further structural diversity of sesterterpene homologues. Our data show that focused expression is an effective method for studying the cyclization mechanism of specific-clade BFTSs and for production of known/unknown sesterterpenes.

■

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was financially supported by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology, Japan (JSPS KAKENHI Grants JP15H01835 (H.O.), JP16H03277 (A.M.), JP16H06446 (A.M.), 17H05430 (M.U.), 17H06173 (M.U.), and JP16H06454 (M.Y.)) and partly supported by the JSPS A3 Foresight Program.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.orglett.7b03418. Experimental procedures, NMR data, LC−MS data, VCD data, and phylogenetic analysis (PDF)

■

REFERENCES

(1) Toyomasu, T.; Tsukahara, M.; Kaneko, A.; Niida, R.; Mitsuhashi, W.; Dairi, T.; Kato, N.; Sassa, T. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 3084−3088. (2) Chiba, R.; Minami, A.; Gomi, K.; Oikawa, H. Org. Lett. 2013, 15, 594−597. (3) Ye, Y.; Minami, A.; Mandi, A.; Liu, C.; Taniguchi, T.; Kuzuyama, T.; Monde, K.; Gomi, K.; Oikawa, H. J. Am. Chem. Soc. 2015, 137, 11846−11853. (4) Bian, G.; Han, Y.; Hou, A.; Yuan, Y.; Liu, X.; Deng, Z.; Liu, T. Metab. Eng. 2017, 42, 1−8. (5) Bian, G.; Deng, Z.; Liu, T. Curr. Opin. Biotechnol. 2017, 48, 234− 241. (6) (a) Toyomasu, T.; Kaneko, A.; Tokiwano, T.; Kanno, Y.; Kanno, Y.; Niida, R.; Miura, S.; Nishioka, T.; Ikeda, C.; Mitsuhashi, W.; Dairi, T.; Kawano, T.; Oikawa, H.; Kato, N.; Sassa, T. J. Org. Chem. 2009, 74, 1541−1548. (b) Shinde, S. S.; Minami, A.; Chen, Z.; Tokiwano, T.; Toyomasu, T.; Kato, N.; Sassa, T.; Oikawa, H. J. Antibiot. 2017, 70, 632−638. (7) Matsuda, Y.; Mitsuhashi, T.; Lee, S.; Hoshino, M.; Mori, T.; Okada, M.; Zhang, H.; Hayashi, F.; Fujita, M.; Abe, I. Angew. Chem., Int. Ed. 2016, 55, 5785−5788. (8) Matsuda, Y.; Mitsuhashi, T.; Quan, Z.; Abe, I. Org. Lett. 2015, 17, 4644−4647. (9) Oka, M.; Iimura, S.; Tenmyo, O.; Sawada, Y.; Sugawara, M.; Ohkusa, N.; Yamamoto, H.; Kawano, K.; Hu, S.-L.; Fukagawa, Y.; Oki, T. J. Antibiot. 1993, 46, 367−373. (10) (a) Kaneda, M.; Takahashi, R.; Iitaka, Y.; Shibata, S. Tetrahedron Lett. 1972, 13, 4609−4611. (b) Sugawara, H.; Kasuya, A.; Iitaka, Y.; Shibata, S. Chem. Pharm. Bull. 1991, 39, 3051−3054. (11) (a) Kawahara, N.; Nozawa, M.; Flores, D.; Bonilla, P.; Sekita, S.; Satake, M.; Kawai, K. Chem. Pharm. Bull. 1997, 45, 1717−1719. (b) Hog, D. T.; Huber, F. M.; Mayer, P.; Trauner, D. Angew. Chem., Int. Ed. 2014, 53, 8513−8517. (12) He, Y.; Bo, W.; Dukor, R. K.; Nafie, L. A. Appl. Spectrosc. 2011, 65, 699−723. (13) Narita, K.; Ye, Y.; Sato, H.; Kudo, K.; Gao, L.; Minami, A.; Taniguchi, T.; Kodama, M.; Gomi, K.; Monde, K.; Uchiyama, M.; Lei, X.; Oikawa, H. Symp. Chem. Nat. Prod.: Symp. Pap. 2017, 59, 273−278. (14) Sato, H.; Narita, K.; Minami, A.; Yamazaki, M.; Wang, C.; Suemune, H.; Nagano, S.; Tomita, T.; Oikawa, H.; Uchiyama, M. 2017, submitted. (15) (a) Shao, J.; Chen, Q.-W.; Lv, H.-J.; He, J.; Liu, Z.-F.; Lu, Y.-N.; Liu, H.-L.; Wang, G.-D.; Wang, Y. Org. Lett. 2017, 19, 1816−1819. (b) Huang, A. C.; Kautsar, S. A.; Hong, Y. J.; Medema, M. H.; Bond, A. D.; Tantillo, D. J.; Osbourn, A. Proc. Natl. Acad. Sci. U. S. A. 2017, 114, E6005−E6014. (16) Lauer, U.; Anke, T.; Sheldrick, W. S.; Scherer, A.; Steglich, W. J. Antibiot. 1989, 42, 875−882.

AUTHOR INFORMATION

Corresponding Authors

*[email protected] *[email protected] ORCID

Xiaoguang Lei: 0000-0002-0380-8035 Tohru Taniguchi: 0000-0002-4965-7383 Kenji Monde: 0000-0002-1424-1054 Masanobu Uchiyama: 0000-0001-6385-5944 Hideaki Oikawa: 0000-0002-6947-3397 6699

DOI: 10.1021/acs.orglett.7b03418 Org. Lett. 2017, 19, 6696−6699