Effective Use of Heterologous Hosts for Characterization of Biosynthetic Enzymes Allows Production of Natural Products and Promotes New Natural Product Discovery

December 2014 Chem. Pharm. Bull. 62(12) 1153–1165 (2014) 1153 Review

Effective Use of Heterologous Hosts for Characterization of Biosynthetic Enzymes Allows Production of Natural Products and Promotes New Natural Product Discovery Kenji Watanabe Department of Pharmaceutical Sciences, University of Shizuoka; Shizuoka 422–8526, Japan. Received June 23, 2014

In the past few years, there has been impressive progress in elucidating the mechanism of biosynthesis of various natural products accomplished through the use of genetic, molecular biological and biochemical techniques. Here, we present a comprehensive overview of the current results from our studies on fungal natural product biosynthetic enzymes, including nonribosomal peptide synthetase and polyketide synthase– nonribosomal peptide synthetase hybrid synthetase, as well as auxiliary enzymes, such as methyltransferases and oxygenases. Specifically, biosynthesis of the following compounds is described in detail: (i) Sch210972, potentially involving a Diels–Alder reaction that may be catalyzed by CghA, a functionally unknown protein identified by targeted gene disruption in the wild type fungus; (ii) chaetoglobosin A, formed via multi-step oxidations catalyzed by three redox enzymes, one flavin-containing monooxygenase and two cytochrome P450 oxygenases as characterized by in vivo biotransformation of relevant intermediates in our engineered Saccharomyces cerevisiae; (iii) ( )-ditryptophenaline, formed by a cytochrome P450, revealing the dimerization mechanism for the biosynthesis of diketopiperazine alkaloids; (iv) pseurotins, whose variations in the C- and O-methylations and the degree of oxidation are introduced combinatorially by multiple redox enzymes; and (v) spirotryprostatins, whose spiro-carbon moiety is formed by a flavin-containing monooxygenase or a cytochrome P450 as determined by heterologous de novo production of the biosynthetic intermediates and final products in Aspergillus niger. We close our discussion by summarizing some of the key techniques that have facilitated the discovery of new natural products, production of their analogs and identification of biosynthetic mechanisms in our study. Key words natural product; biosynthesis; polyketide synthase; nonribosomal peptide synthetase; de novo production; reaction mechanism

1. Introduction practical to obtain those enzymes in their active form in large To date, numerous secondary metabolites from various quantities from the producing organisms to be used as cata- natural sources, such as bacteria, fungi, and plants, have been lysts, if we can isolate the genes that encode such enzymes isolated and characterized. Many of those metabolites are from the original host and express them in a more convenient important and valuable, because they often exhibit a broad strain or heterologous host, desired natural products can be spectrum of biological activities. Their remarkable properties synthesized more readily. To achieve this, the development can be attributed to their structural complexity and diversity, of expression systems for heterologous production of complex and enzymes are responsible for the formation and diversifica- natural products has been pursued with success in various tion of those compounds. Each biosynthetic enzyme catalyzes organisms, including Escherichia coli,1–4) yeast (Saccharomy- a reaction that constructs a certain chemical structure, and ces cerevisiae)5) and fungi.6) On the other hand, isolation of collection of such transformations along a biosynthetic path- new secondary metabolites from natural sources has become way results in the formation of a natural product. Chemists more difficult in recent times because numerous secondary who are interested in the study of biosynthesis are attracted metabolites have already been isolated from Streptomyces and to carbon–carbon bond-forming enzymes having the ability to other microorganisms. Moreover, culturability of the organ- produce the carbon skeleton of a molecule and auxiliary en- isms significantly restricts the available pool of microbes for zymes catalyzing various modifications to the skeleton, such natural product isolation. Therefore, if drug discovery contin- as C- and O-methylations and oxidations. While it is often im- ues to rely on such traditional methods, the isolation of new natural products may encounter significant difficulties. More The author declares no conflict of interest. recent technology has explored the culturing of anaerobic mi-

e-mail: [email protected] © 2014 The Pharmaceutical Society of Japan 1154 Vol. 62, No. 12 croorganisms for production of metabolites that are expected synthesis of mRNA from a target gene cluster. Once mRNA to be new products. However, the methodology is not new—it can be obtained, the corresponding cDNA can be prepared in still relies on traditional processes that involve culturing of a streamlined fashion even for a polyketide synthase (PKS) or the organisms and screening, isolating and characterizing the a nonribosomal peptide synthetase (NRPS) gene, which can compounds.7,8) Thus, the development of new approaches for exceed 20 kilobases (kb) in length.16) Availability of cDNA al- isolating novel compounds from various microorganisms has lows transfer of the gene into a convenient host, such as yeast, also intensified in recent years. to achieve heterologous production of a biosynthetic enzyme During the past decade, many secondary metabolite gene for detailed analysis of the reaction mechanism. If sufficient clusters, including polyketide biosynthetic genes, have been genes from the cluster can be transferred to yeast, it becomes discovered through fungal genome sequencing. While typi- possible to biosynthesize the otherwise inaccessible new com- cally 30 to 40 gene clusters are identified in a single Asper- pounds heterologously. gillus genome, much fewer polyketide, peptide and terpene In this review, we will focus on heterologous production products can be isolated from a fungal culture grown under of natural products in yeast and fungi systems that have been typical growth conditions.9) Based on transcriptome analysis developed in our laboratory. Using our own work as examples, of secondary metabolism gene expression, the transcription we will illustrate how those systems have allowed us to gain level of some gene clusters responsible for secondary me- insight into natural product biosynthetic mechanisms. We will tabolite biosynthesis has been observed at a very low level, also touch on how the system can be exploited for engineering existing essentially as silent gene clusters, under conventional and identifying natural product biosynthetic pathways to ad- culture conditions.10) Consequently, natural products that those dress the obstacles described above in our efforts toward the gene clusters can biosynthesize are not attainable using con- production of potentially valuable analogs and discovery of ventional methods. To circumvent these situations and mine new natural products. for new natural products, techniques to artificially activate those silent biosynthetic gene clusters in fungal chromosomes 2. Sch210972 Biosynthesis Involvement with Diels– have been devised. Currently, there are two approaches avail- Alder Reaction Catalyzing Enzyme Responsible for the able for achieving gene cluster activation using molecular bio- Formation of Decalin Core Structure logical techniques. One is to decrease or increase epigenetic Discovering novel enzymes along with deciphering their regulation. For secondary metabolites, decreasing or increas- mechanism of enzymatic reaction is an arduous undertaking ing epigenetic regulation can increase transcription from a requiring collaboration among biochemistry, natural product biosynthetic gene cluster, where the increase of transcription chemistry, molecular biology, and structural biology. However, is achieved at the chromatin level by reducing or inducing such enzymes can shed light on how biology accomplishes modifications of histones.11) The alternate approach is to over- complex chemical synthesis, providing potential hints for in- express or knock out a transcriptional regulator (TR) that is novation in our drug discovery efforts. In this chapter, we will associated with a silent biosynthetic gene cluster. When a TR discuss approaches sought to obtain proof for the existence is located away from its corresponding gene cluster, such a of a Diels–Alder reaction-catalyzing enzyme in nature. It has TR often represses transcription from the gene cluster and been proposed that many natural product biosyntheses involve suppresses the production of its associated natural products.12) the participation of a Diels–Alder reaction.17–20) However, On the other hand, overexpression of a TR located within a only five examples of biotransformation and their reaction gene cluster frequently enhances the biosynthesis of its re- mechanisms have been examined in detail using purified spective secondary metabolites. Lack of expression of this natural enzymes that catalyze a pericyclic reaction. Those type of regulator is typically the cause of the low expression five enzymes, SpnF in spinosyn biosynthesis,21) solanapyrone of biosynthetic genes that results in poor production of the synthase (Sol5),22) LovB from lovastatin biosynthesis,23,24) corresponding natural product.13,14) In our recent study,15) we macrophomate synthase (MPS),22,25) and riboflavin synthase showed that overexpression of a TR encoded within a silent, (RibC)26,27) were reported previously as enzymes that cata- putative polyketide biosynthetic gene cluster in Aspergillus lyze a Diels–Alder reaction based on their observed ability oryzae allowed straightforward activation of the gene clus- to promote a cycloaddition reaction (Chart 1). SpnF has been ter and isolation of natural products that helped broaden our described as the first enzyme for which specific acceleration knowledge of mechanisms of natural product biosynthesis. of a [4+2] cycloaddition reaction has been verified experi- Techniques described above can also be applied to induce the mentally as its only observable function.21) Although SpnF is

Kenji Watanabe received his Ph.D. in 2000 in biological chemistry from Hokkaido University, Japan. His postdoctoral fellowships included a year at University of Wisconsin-Madison and a two- year appointment at Stanford University. In 2003, he returned to Hokkaido University as an assistant professor. In 2004, he joined the department of pharmaceutical sciences at University of Southern California as a research assistant professor. In 2008, he received a tenure-track position at Oka yama University and was promoted to associate professor at University of Shizuoka in 2009. His research interest has been focused on developing a heterologous production system for natural products from PKS, NRPS and other important molecules of interest. Also, he has placed much effort into producing new molecules by rationally engineering the above system using conventional and Kenji Watanabe straightforward methods. December 2014 1155

(A) SpnF for spinosyn A biosynthesis. (B) Sol5. (C) LovB for lovastatin biosynthesis. (D) MPS. (E) RibC. Chart 1. Reaction Schemes for Proposed Enzymatic Catalysis of Diels–Alder Reaction a cyclase catalyzing the Diels–Alder reaction for an estimated glutamic acid 4, an unusual amino acid building block of 1 rate enhancement of approximately 500 fold, the cyclic sub- (Chart 2A), we confirmed the involvement of a polyketide strate contains both the diene and dienophile and can undergo synthase–nonribosomal peptide synthetase (PKS–NRPS) hy- the cyclization spontaneously to form the observed product. brid megasynthase CghG and a stand-alone enoyl reductase LovB, which is a fungal highly reducing PKS, has been pos- CghC in the formation of the core structure of 1 through gene tulated to produce the decalin core structure of lovastatin via disruption in C. globosum32) (Chart 2B). This enzymatic ar- an endo Diels–Alder reaction. In vitro study has indicated that rangement is reminiscent of the involvement of LovB (PKS) LovB catalyzes the formation of an endo Diels–Alder prod- and LovC (stand-alone enoyl reductase) for the lovastatin uct having the same stereochemistry as dihydromonacolin L biosynthesis. Also, as proposed for the biosynthesis of the lo- that is different from nonenzymatic endo or exo adducts.23) vastatin core structure,23) the core biosynthesis of 1 is thought Macrophomate synthase has been studied extensively by this to proceed via a Diels–Alder reaction. Thus, we hypothesize author and his colleagues as the first Diels–Alder-catalyzing that a Diels–Alder reaction is involved in the biosynthesis of enzyme.28) However, recent studies have suggested that the Sch210972. A similar mechanism has also been proposed for tandem Michael–Aldol reaction is a more plausible reaction the PKS–NRPS-mediated formation of the core scaffolding pathway for this transformation,29,30) Other multifunctional structure of prochaetoglobosin I (2), the precursor of chaeto- enzymes, Sol5 and RibC, have been proposed to participate globosin A (3) (see the next section for details) also from C. in hydroxyl oxidation22) and hydride transfer,27) respectively, globosum33,34) (Chart 3A), cytochalasin E from another fungus in addition to the [4+2] cycloaddition reactions. However, the Aspergillus clavatus NRRL 1 (Chart 3B), and equisetin in concertedness of how these enzymes catalyze the multiple re- Fusarium heterosporum.35) Most interestingly, targeted disrup- actions has yet to be elucidated. tion of cghA from the Sch210972 biosynthetic gene cluster in In our continued search for a bona fide Diels–Alderase, C. globosum led to the formation of an exo form of the prod- we have focused on a compound called Sch210972 (1) from uct 5 as determined by NMR spectroscopy and X-ray crystal- Chaetomium globosum (Chart 2). This compound has a unique lography, in addition to the endo product 1, which is the sole inhibitory activity against the cell surface receptor CCR-5.16, product isolated from the wild type C. globosum.32) The ob- a biological activity that can be exploited as a novel antiviral served formation of the exo product in the cghA mutant strain treatment against human immunodeficiency virus-1 (HIV-1) strongly indicates that the endo-selective cyclization observed infection by blocking viral cell entry.31) We have identified in the wild type strain is indeed the result of a Diels–Alder recently the gene cluster responsible for the biosynthesis of reaction, and suggests the involvement of CghA in steering 1 through genome mining. In addition to the enzyme CghB, the stereoselectivity of the cycloaddition reaction. Further in- predicted to take part in the formation of 4-hydroxy-4-methyl vestigation is currently ongoing in our laboratory. 1156 Vol. 62, No. 12

(A) Proposed biosynthetic steps of unusual amino acid, 4-hydroxy-4-methyl glutamic acid 4, and (B) core structure of Sch210972 (1). Abbreviations: KS, ketosynthase; MAT, malonyl-CoA acyltransferase; DH, dehydratase; MT, methyltransferase; KR, ketoreductase; ACP, acyl carrier protein; C, condensation; A, adenylation; T, thiolation; R, reductase; SAM, S-adenosyl-L-methionine. Chart 2. Proposed Sch210972 Biosynthetic Pathway

Chart 3. Enzymatic Intermolecular Diels–Alder Reaction Predicted to Be Involved in the Biosynthesis of (A) Chaetoglobosin A and (B) Cytochalasin E

3. Chaetoglobosin A Biosynthetic Pathway While it is frequently difficult to resolve the biosynthetic Fungal natural products often exhibit medicinally useful bi- pathway in which multiple enzymes are employed to gener- ological activities. For example, chaetoglobosin A (3)36–38) pos- ate various intermediates for the formation of a complex final sesses unique inhibitory activity against actin polymerization product, the use of knockout strains greatly facilitates identifi- in mammalian cells.39,40) The gene cluster responsible for the cation of missing intermediates for elucidation of the details of biosynthesis of 3 was predicted and identified using a small the pathway. By preparing multiple knockout strains, we were interfering RNA (siRNA) technique in Penicillium expansum, able to determine the function of three genes encoding for revealing the organization of the genes involved in the biosyn- redox enzymes, one flavin-containing monooxygenase (FMO) thesis of 3.41) As in the case of the lovastatin biosynthesis, a (CHGG_01242-2) and two cytochrome P450 oxygenases PKS–NRPS and a stand-alone enoyl reductase were assigned (P450s) (CHGG_01242-1 and CHGG_01243)34) as illustrated in to be responsible for the formation of the core structure of 3. Chart 4. We also identified four products, prochaetoglobosin Another interesting aspect of the biosynthetic pathway of 3 is IV (6),33,43) 20-dihydrochaetoglobosin A (7),44) cytoglobosin D that a wide variety of chaetoglobosin-type natural products (8),45) and chaetoglobosin J (9),37) as intermediates formed by are isolated from C. globosum.37) To understand the source the redox enzymes during the biosynthesis of 3 from 2 based of the observed structural diversity, we sought to character- on the chemical structures of those compounds. Through ize the roles played by the enzymes encoded in the chaeto- the study, we were able to gain insight into the origin of the globosin biosynthetic gene cluster in C. globosum. We have structural diversity present in this class of natural products. established a DNA ligase ligD-deficient strain of C. globosum Our study showed that the auxiliary redox enzymes have a that allows targeted gene disruption at high efficiency.42) Using considerable substrate tolerance that leads to the formation of this strain, we prepared a series of mutant strains having each various intermediates, and reiterates the notion46) that redox of the genes in the chaetoglobosin biosynthetic gene cluster enzymes play a central role in introducing complexity into knocked out through targeted homologous recombination. the chemical structures of natural products. Application of the December 2014 1157

Chart 4. Biosynthetic Pathway for the Transformation of 2 to 3 Involving Multiple Oxidation Steps

(A) Chemical structures of representative dimeric compounds. Compounds 15 and 16 are newly isolated and characterized in this study. (B) A. flavus A1421 ditryptophenaline biosynthetic gene cluster and predicted function of the genes in the cluster. (C) Proposed mechanism of dimerization in 10, 15 and 16 from 12 and 14 via a radical route catalyzed by DtpC. Chart 5. Proposed Biosynthesis of (−)-Ditryptophenaline gene-specific knockout method to a secondary metabolite bio- dimeric natural products have been reported to date, among synthetic gene cluster allowed us to translate the genomic se- which nearly 100 are dimeric diketopiperazines.47–60) Dimeric quence information into biochemical information in the form natural products exhibit various biological activities,49–51) and of encoded enzymes and ultimately to chemical structures of dimerization is thought to contribute toward conferring bioac- the resulting compounds. Installation of the ability to perform tivity to these products.61) In this section, we will outline our site-specific homologous recombination by disrupting a ligD study on the enzymes involved in the biosynthesis of 10 that gene homolog in the fungus of interest made it possible to in- is aimed at understanding the biosynthetic mechanism of the vestigate those fungal species for which established molecular formation of dimeric natural products. genetic methods are not yet available. To identify the genes involved in the biosynthesis of 10 in the producing host A. flavus, we again exploited the site-spe- 4. Functional Analysis of a Cytochrome P450 as Di- cific homologous recombination activity, which is also avail- merization Catalyst in (−)-Ditryptophenaline Biosyn- able in the mutant strain A. flavus A1421 carrying deletions of thesis the orotidine 5′-monophosphate decarboxylase gene pyrG and The dimeric diketopiperazine (–)-ditryptophenaline (10)47) the ATP-dependent DNA helicase II subunit gene ku70 in its was isolated from Aspergillus flavus as a representative genome.62) We deleted a set of genes from a three-gene cluster member of alkaloids isolated from this fungus. This com- predicted to be involved in the formation of a nonribosomal pound contains two hexahydropyrroloindole substructures peptide product (Chart 6). Analysis of metabolites produced that are connected at vicinal quaternary stereocenters to form by those mutants identified that DtpA, an NRPS encoded by a dimeric molecule48) (Chart 5A). Currently, more than 600 AFLA_005440, is indispensable for the formation of 10, al- 1158 Vol. 62, No. 12

The catalytic domains found within the enzymes are represented by the following single letter notations: A, adenylation; T, thiolation; and C, condensation. Chart 6. Scheme for a Complete Cycle of Nonribosomal Peptide Chain Elongation during the (−)-Ditryptophenaline Biosynthesis lowing assignment of the gene cluster as the ditryptophenaline shielded from bulk solvent and kept in close proximity with biosynthetic gene cluster and DtpA as the NRPS responsible each other to promote the specific formation of the C3(sp3)– for the formation of 11, the peptide core of 10 (Chart 5B). The C3′(sp3) bond. In this study, determination of the ditrypto- structure of 11 indicated that DtpA accepts L-tryptophan and phenaline biosynthetic pathway revealed DtpC to be a unique L-phenylalanine as its substrates and forms the phenylalanyl- P450 that performs a ring closure and C–C bond formation tryptophanyl cyclic dipeptide product (Chart 5C). Deletion consecutively for the biosynthesis of dimeric diketopiperazine of the second gene of the cluster, AFLA_005450, which was alkaloids. Our study has also identified DtpC to be an inher- predicted to encode an N-methyltransferase, led to an accu- ently highly promiscuous catalyst that can be useful for engi- mulation of 12, the N-methylated product of 11. This allowed neered biosynthesis of novel heterodimeric natural products. us to assign this gene product, which we named DtpB, to be responsible for the N-methylation of 11 to yield 12. These 5. Understanding Pseurotin Scaffold by Functional results left AFLA_005460, the third gene of the cluster that is Analysis of the Biosynthetic Enzyme in Vitro and in predicted to code for a cytochrome P450, as a likely candidate Vivo Reactions for catalyzing the cyclization and dimerization of 12 to gener- Pseurotins comprise a family of fungal secondary me- ate 10. Our in vitro analysis of DtpC, the translation product tabolites that exhibit various medicinally valuable biological of AFLA_005460, confirmed that DtpC accepts an unfused activities.66–69) Genome sequencing of Aspergillus fumigatus70) diketopiperazine 12 as its substrate and performs concurrent has revealed a number of gene clusters that are predicted to pyrroloindole formation and dimerization to form 10 (Chart encode for enzymes that constitute natural product biosynthet- 5C). Furthermore, structural similarity between 10 and other ic pathways. The gene cluster responsible for the biosynthesis dimeric diketopiperazine alkaloids, as well as a previous feed- of pseurotin A (19) was predicted and identified by deleting ing experiment on the WIN 64821 (13)-producing Aspergillus or overexpressing a PKS–NRPS gene psoA in A. fumigatus sp. that produced a series of dimeric diketopiperazines dif- Af29371) (Table 1). From this study, PsoA was considered to be fering in the non-tryptophan residue of the cyclic dipeptide responsible for the biosynthesis of the 1-oxa-7-azaspiro[4,4]- core,52) prompted us to examine in vitro if DtpC can accept non-2-ene-4,6-dione core of pseurotin, which contains five other diketopiperazines for dimerization. When DtpC was chiral centers. While detailed analysis of the pseurotin biosyn- provided with brevianamide F (14), a prolyltryptophanyl thetic gene cluster and its genetic organization has been con- diketopiperazine63) obtained heterologously from yeast,6) the ducted,72) the mechanism of the biosynthesis of pseurotin and corresponding homodimeric compound (−)-dibrevianamide F its related compounds has remained unclear. As discussed ear- (15) was obtained. In addition, when a mixture of 12 and 14 lier, structural diversity of natural products is likely generated was reacted with DtpC, a heterodimeric product we named during the post-PKS–NRPS modification steps. We expected (−)-tryprophenaline (16) was obtained. this to be the case with the pseurotin-type of natural products, Based on the results from our study and a previous report which includes biologically active molecules like azaspirene that strongly indicated a radical pathway for the C2–N10 ring- (20),73) synerazol (21)74–76) and cephalimysin A (23)69) that closure and dimer formation,64,65) we proposed that a radical vary in C- and O-methylations and the extent of oxidation formed at N10 by heme-catalyzed hydrogen atom abstraction (Chart 7). Thus, methylation and oxidation steps were inves- could migrate to C3 and undergo a radical-mediated coupling tigated using strains carrying deletion of the genes predicted with another C3 radical-bearing monomer generated within to encode methyltransferase, FMO and P450 found in the the same active site of DtpC to form the bridging bond in di- cluster, namely psoC, psoD, psoE and psoF.77) Our study also mers like 10, 15 and 16. While the radical-bearing monomers employed in vivo and in vitro assays that revealed several in- could leave the DtpC active site and form the dimer either teresting aspects of the pseurotin biosynthetic enzymes. Most in solution or in the active site of another copy of DtpC, the notably, we identified that PsoF, a bifunctional fusion protein strict absence of the formation of monomeric products, such has an epoxidase domain and a C-methyltransferase domain. as 17 and 18 (Chart 5C), either in vivo or in vitro, and relative- Interestingly, each domain catalyzes two completely separate ly short lifetime of radicals strongly favor our hypothesis of steps of the pseurotin biosynthesis. Furthermore, this enzyme radical-mediated ring closure and dimer formation occurring is the first report of a trans-acting methyltransferase that in- concomitantly within a single DtpC active site. Presumably, troduces a methyl branch to the growing polyketide carbon the proposed series of reactions can proceed efficiently within backbone. the active site of DtpC where radical-bearing substrates are For PsoE, a predicted glutathione S-transferase, its dis- December 2014 1159 ruption led to a substantial loss of 19, new formation of 24 32, suggesting PsoD is responsible for the oxidation of C17. (13E-configured isomer of 19), and noticeable accumulation of In vivo conversion of 20 by S. cerevisiae expressing the psoD 25 and 26 (Chart 7). NMR analysis allowed us to determine gene and in vitro assay of PsoD with 20 resulted in the forma- 25 and 26 to be geometric isomers with 25 and 26 being as- tion of 29, confirming that psoD codes for an oxidoreductase signed as the 11E,13E- and the 11Z,13E-configured isomer, that is responsible for the installation of the C17 ketone in 20 respectively. These two compounds are both 13E isomers and and, by extension, 30. However, PsoD did not react with 31 are predicted to be formed before 13Z-containing 21 that is and 32, indicating that PsoD is like PsoC in that it is sensitive formed by the epoxidation activity of PsoF. Thus, accumula- to modifications to the diene side chain. No formation of O8- tion of 25 and 26 in the ΔpsoE strain suggests that PsoE may methylated products by the ΔpsoD deletion strain showed that catalyze the trans-to-cis isomerization of the C13 olefin. Next, PsoC requires the C17 ketone group to exert its O8-methyla- PsoC was determined to be an O-methyltransferase that acts tion activity, suggesting that the PsoD-catalyzed step comes on the hydroxyl group at C8 position when deletion of psoC upstream of the PsoC-catalyzed step during the biosynthesis led to accumulation of O8-unmethylated products, such as the of pseurotins. Also, the proposed functions of PsoC and PsoD shunt products 27 and 28. Further biochemical analysis re- allowed us to hypothesize that the product 33 from the ΔpsoF vealed that PsoC does not convert 27 to 19, indicating that the strain is formed from 34, a C3-unmethylated analog of 20, by substrate specificity of PsoC is sensitive to modifications to the actions of PsoC and PsoD. the diene side chain of the azaspirene skeleton. Lastly, deletion In summary, our study identified that the biosynthesis of of psoD, a predicted P450, led to an accumulation of 31 and 19 starts from 22 and proceeds through the formation of 20 and 21 via multiple oxidation and methylation steps. Also, our Table 1. Functional Annotation of the Open Reading Frames in the Pseu- detailed characterizations of the enzymes and the products rotin Biosynthetic Gene Cluster in A. fumigatus A1159 formed revealed the relatively broad substrate promiscuity of Locus ID Gene name Functional annotation those modifying enzymes that permits assembly of a complex network of pathways that results in a combinatorial formation AFUA_8G00420 fapR C6 Finger transcription factor of different products. AFUA_8G00440 psoF Cyclohexanone 1,2-monooxygenase (flavin-containing monooxygenase/ methyltransferase) 6. Distinct Mechanisms for Spiro-Carbon Formation in Spirotryprostain Biosynthetic Pathway AFUA_8G00530 psoB α,β-Hydrolase Spirotryprostatin A (35) (Chart 8) is a member of the fumi- AFUA_8G00540 psoA PKS–NRPS hybrid tremorgin class of alkaloids produced by A. fumigatus.78) This AFUA_8G00550 psoC Methyltransferase compound and its close relative, spirotryprostatin B (36),79) AFUA_8G00560 psoD Cytochrome P450 oxidoreductase have gathered interest due to their potent anti-cancer activity AFUA_8G00570 N/A α,β-Hydrolase that stems from their ability to arrest cell cycle.80) However, AFUA_8G00580 psoE Glutathione S-transferase interest in those compounds also arises from their complex Broken line denotes the discontinuity in the gene cluster organization. Annotation given within parentheses is a revision to the original annotation based on the findings chemical structure, namely the spirocyclic core architecture. from the current study. N/A: not assigned as no change was observed in the metabo- Various compounds having spiro-ring moieties, such as spi- lite profile upon deletion of the open reading frame. rotryprostatins, have been targeted for chemical synthesis

Pathway enclosed within the broken-line box with thick arrows represents the proposed main pathway for the formation of 19. Chart 7. Biosynthetic Pathway for the Transformation of a Precursor 22 to Pseurotin A 19, Involving Methylation and Multiple Oxidation Steps 1160 Vol. 62, No. 12

Chart 8. Proposed Biosynthesis of Tryprostatins, Fumitremorgins and Spirotryprostatins in A. fumigatus

(A) Proposed mechanism of imidazoindolone ring formation in the fumiquinazoline biosynthesis. Initial epoxide-forming step is thought to be carried out by a flavin- containing monooxygenase (FMO) FqzB, followed by a nonribosomal peptide synthetase FqzC carrying out the cyclization. (B) Proposed mechanism of spiro-carbon formation in the notoamide C biosynthesis involving semipinacol-type rearrangement catalyzed by another FMO, NotB. Chart 9. Two Examples of Chemical Transformations Involving a Reaction Step That Is Facilitated by Epoxidation because of their challenging structure and frequent association remorgin biosynthetic cluster. Subsequent opening of the ep- with interesting biological activities. oxide ring in 38 initiated by donation of the methoxy oxygen Recently, we identified the spiro-carbon-forming enzymes lone pair facilitates to formation of 42 through a semipinacol- involved in the biosynthesis of spirotryprostatins.6) Clues were type rearrangement (Chart 10C). However, for our target com- obtained from reports of “biomimetic” approach for the syn- pound 35 to be formed, FqzB must perform its epoxidation thesis of spirotryprostatins, which employed N-bromosuccin- on a cyclic precursor like 43 (Chart 10A). Subsequent detailed imide to induce a semipinacol rearrangement via oxidation of in vivo and in vitro biochemical characterizations allowed us the C2–C3 bond of the indole ring to form the spirooxinidole to demonstrate that FqzB indeed catalyzes a stereoselective moiety of spirotryprostatins.81–83) Also, for the biosynthesis epoxidation of the indole ring of 43 and sets up the compound of notoamide C (41), formation of a β-2,3-epoxy moiety on for a semipinacol-type rearrangement that leads to the forma- an indole group has been proposed as a key step enabling the tion of the stereospecific spirooxindole system in 35. subsequent semipinacol-type rearrangement via epoxide open- The reaction mechanism proposed above can also explain ing through which the 2-prenyl group can undergo an α-face how the S- versus R-stereoconfiguration of the spiro-carbon [2,3]-shift with a retention of the stereochemistry to yield no- among different natural products is determined. For example, toamide C (41)84) (Chart 9B). Combining those previous stud- antipodal pairs of the previously mentioned notoamide B and ies and the proposed mechanism of an FMO FqzB-catalyzed related compound versicolamide B are known to exist in na- epoxidation of fumiquinazoline F (37)85) (Chart 9A) and the ture.86) A terrestrial fungus Aspergillus versicolor produces structural similarity among the fumitremorgin and fumiquin- (+)-notoamide B and (+)-versicolamide B having an S-con- azoline biosynthetic intermediates, FqzB can catalyze the figured spiro-carbon, whereas a marine fungus Aspergillus sp. epoxidation of tryprostatin A (39) to form the 2,3-epoxyindole MF297-2 generates (−)-notoamide B and (−)-versicolamide B intermediate 38 despite originating from the unrelated fumit- with a reverted R configuration at the spiro-carbon. This ste- December 2014 1161

(A) Proposed reaction scheme for the formation of the spiro-carbon of spirotryprostatin A (35) by FqzB from 43 through an epoxidated intermediate 44. (B) A proposed epoxidation-mediated transformation of 45 into 46 partially catalyzed by FqzB. (C) Proposed mechanism of the semipinacol rearrangement of 39 catalyzed by FqzB, leading to the formation of 42 having a tetrasubstituted carbon center, as well as two diol shunt products 40a and 40b. Chart 10. Characterization of FqzB from the Fumiquinazoline Biosynthetic Pathway for Its Role in Spiro-Carbon Formation

Chart 11. Proposed Mechanism of Spiro-Carbon Formation in Spirotryprostatin B (36) and Diol-Containing Shunt Product Biosynthesized from 47 through Radical Route, Catalyzed by the Cytochrome P450 FtmG reospecificity can be controlled by a stereoselective epoxida- formation of 36, the 6-demethoxy analog of 35, revealed that tion of their precursors. In the case of notoamide biosynthesis, FqzB does not transform 45 into a 42-like 2-oxindole spiro installation of an α-face epoxide will lead to a β-face migra- compound in vitro. The only product formed was a diol 46 tion of the prenyl group to yield an S-configured carbon at C-3 (Chart 10B). This inability of FqzB to catalyze a semipina- of 41.87) On the other hand, installation of a β-face epoxide col-type rearrangement of 45 may arise from the lack of a can result in an α-face migration of the prenyl group to form 6-methoxy group on the indole ring. The oxygen atom allows a C-3 R-configured 3-epi-notoamide C84) (Chart 9B). Thus, it the formation of a para-quinone methide-like intermediate can be hypothesized that the marine-derived Aspergillus car- that can promote the ensuring semipinacol-type rearrange- ries an FMO that performs a β-face epoxidation, while the ter- ment (Chart 10C). Subsequent search identified that it was a restrial Aspergillus has an α-face-specific FMO for a parallel P450 FtmG that performs the spiro-carbon formation on 47 to biosynthesis of enantiometrically opposite natural products. form 36 (Chart 11). Furthermore, this P450 also works on 43 Interestingly, our attempt to identify the mechanism of to form a new spirooxindole compound 48 (Chart 8). While 1162 Vol. 62, No. 12

A) Reverse transcription PCR for the synthesis of a first strand cDNA (dotted box) with an oligo(dT)20 primer (black arrow) from an mRNA (thick black line). B) Stan- dard PCR for an amplification of short (3 kb on average) cDNA fragments (three shades of gray) from the full-length cDNA. C) Overlap extension PCR for joining the small fragments into a full-length double stranded DNA (dsDNA) (white box) encoding for the target gene. The terminal primer contains a 25-bp segment (cross-hatched and hatched box) whose sequence is homologous to the sequence at the site in the vector where this fragment is to be integrated. D) Homologous recombination in S. cerevisiae for combining the linearized expression vector (thinner black line) and the full-length dsDNA into a single circular plasmid. Chart 12. A Schematic Diagram Showing the Construction of a Yeast Expression Vector Using the ExRec Method, Which Involves an Overlap Ex- tension PCR and Yeast Homologous Recombination Steps the FMO-mediated spiro-carbon formation proceeds through an example to illustrate the usefulness of the S. cerevisiae- an epoxide intermediate, the P450 pathway appears to involve based system.6) In this example, we examined the feasibility radical-mediated two-step hydroxylation followed by dehydra- of expressing a large (13 kb) A. fumigatus gene that codes for tion and semipinacol-type rearrangement (Chart 11). an NRPS in yeast. Difficulty of expressing such a large gene Through this study, we identified that A. fumigatus main- heterologously stems not only from the difficulty of obtaining tains two orthogonal spiro-ring formation pathways, one a full-length cDNA but also from host’s ability to handle pro- through epoxide formation by the FMO (FqzB) and another duction of an enormous protein. Initially, cDNA of the puta- through the P450 (FtmG)-catalyzed radical route. What is tive NRPS gene Afu6g12080 was synthesized by reverse tran- particularly interesting is that the FMO belongs to the fu- scription PCR from the total RNA isolated from A. fumigatus miquinazoline biosynthetic pathway that is unrelated to the (Chart 12A). Assembly of a full-length cDNA of the target fumitremorgin biosynthetic pathway, which is responsible for gene was simplified by joining multiple partial amplicons of the formation of the precursors of spirotryprostatins, trypros- the cDNA (Chart 12B) using the overlap extension PCR (Chart tatins. This apparently fortuitous crosstalk of the two biosyn- 12C). The resulting full-length gene was inserted into a recipi- thetic pathways leads to the formation of 35. Our study also ent vector of choice using the in vivo recombination activity revealed that the P450 FtmG is capable of recognizing two of yeast (Chart 12D), yielding pKW5022 (Afu6g12080), the different substrates for the generation of two different spiro vector for expressing the NRPS gene. This method was also compounds 36 and 48. Those findings reiterate the critical role applied to clone full-length cDNAs of other NRPS and PKS oxygenating enzymes, such as FMOs and P450s, play in intro- genes. This technique, which we termed “overlap-extension ducing chemical diversity into natural products.34,46) PCR–yeast homologous recombination (ExRec),” allowed reli- able preparation of a S. cerevisiae expression plasmid contain- 7. Engineered Biosynthesis of Complex Natural Prod- ing the full-length, intron-free open reading frame of a target ucts in Heterologous Host large fungal gene synthesized from a pool of total mRNA.16) In this review, we described our attempts to decipher a Next, the expression vector was introduced into an engineered number of fungal natural product biosyntheses, where the suc- S. cerevisiae strain SCKW5, whose genome is modified to cess of the study relied strongly on the use of yeast and fungi include a G418 resistance marker kanMX, a malonyl-CoA for heterologous expression of various biosynthetic genes and synthetase gene matB and a phosphopantetheinyl transferase de novo production of intermediates and target compounds. gene npgA.16) Expression of the foreign gene was checked by The techniques developed during the course of our studies Western blotting analysis, and biosynthesis of the correspond- have served as an important tool for determining the mecha- ing nonribosomal peptide compound was examined. The nism of how the enzymes biosynthesize the complex chemical culture of pKW5022 (Afu6g12080)/SCKW5 contained peptide structures. In this last section, we will describe the yeast and products 37 (Chart 9A) and 49 (Chart 13) with approximate fungi systems, two of the most heavily used platforms in our yields of 0.4 and 1.8 mg/L of culture, respectively. However, work. 49 was a compound never identified before. Using similar First, we will use our heterologous production of 37 as methods, we expressed five genes ftmA–E from the A. fu- December 2014 1163

own producer.88) It is thought that MlcD-like proteins may bind and sequester those compounds to furnish the producing host a non-destructive resistance against the inhibitors. Simi- larly, fumagillin, a meroterpenoid from A. fumigatus, is known for its anti-angiogenic activity, that is exerted through inhibition of human methionine aminopeptidase 2 (MetAP-2).89,90) Indeed, the fumagillin biosynthetic gene cluster contains fpaII, which encodes for a protein homologous to MetAP-2.91) Again, FpaII is thought to bind fumagillin and confer a non- Chart 13. Chemical Structure of the Fumiquinazoline Analog 49 Bio- destructive resistance to A. fumigatus. These examples of a synthesized in the Engineered Yeast resistance system suggest that homologs of the protein the natural product targets could furnish a self-resistance function migatus spirotryprostatin biosynthetic pathway successfully in that confers resistance by binding the compound. Therefore, if S. cerevisiae to produce 43 (Chart 8) and 47 (Chart 11) from the target of inhibition for the compound of interest is known, L-proline and L-tryptophan. More importantly, this yeast-based incorporating the target or its homolog into the surrogate host heterologous production system allowed us to prepare path- may be an effective way of building tolerance in the host for way intermediates that were essential for conducting detailed high-titer production of the target compound. biochemical studies to determine the enzymatic mechanism of the formation of spirotryprostatin and other complex natural 8. Closing Remarks products. In addition, use of a plasmid-based system provides In this review, we summarized the findings from our recent an advantage in terms of ease and speed of cloning the target studies, which were focused on identifying the mechanisms of genes. The use of plasmids simplifies the effort of engineering biosynthesis of complex fungal natural products and under- biosynthetic pathways for production of various analogs by standing the enzymatic reactions involved in those metabolic allowing quick shuffling of genes that make up the biosyn- processes. We also briefly touched on the engineering of het- thetic pathway of interest using simple molecular biological erologous hosts, primarily yeast and fungi, which enabled the techniques. Our studies described in this review have demon- bulk of the studies described in this review. Heterologous pro- strated the effectiveness of the use of our system in translating duction of these complex compounds has an advantage over uncharacterized fungal biosynthetic genes into structurally chemical synthesis in that it allows us to take advantage of the characterized compounds. We believe that our approach will superior capability of enzymes to form specific carbon–car- help accelerate the efforts in isolating novel natural products bon bonds and dictate a molecule’s stereochemistry. Through and rationally engineering biosynthetic pathways for produc- the use of the yeast-based platform, we have established a tion of analogs having desirable characteristics. number of procedures, such as the ExRec method, that al- While yeast is a robust and versatile platform, it is not suit- lowed fast and efficient transfer of genetic information from able for all genes and proteins. An example is the case with de original production hosts to more convenient yeast and fungal novo biosynthesis of spirotryprostatins that we discussed earli- hosts for heterologous production of enzymes and compounds. er. Yeast was able to produce those compounds only at a trace However, there is still room for improvement of these meth- level, which made our attempt to verify the genes involved ods. For example, so far we have exploited the convenience of in the biosynthesis of those compounds very difficult. Thus, plasmid-based expression systems for the quick assembly of we attempted to establish an Aspergillus niger system for ex- engineered biosynthetic pathways in the heterologous hosts. pressing relevant genes for production of desired compounds. For a fungal host, however, chromosomal incorporation of a Using this system, we were able to demonstrate that ftmA, cluster of heterologous biosynthetic genes under the control of ftmB, ftmE and ftmG were sufficient for the biosynthesis of 36, a strong promoter, where only a single copy of each gene will whereas addition of ftmB and ftmE and exchange of ftmG with be introduced to the host, may help stabilize the expression of fqzB were necessary for the formation of 35 de novo. Analysis those genes for further improvement in the heterologous bio- of the A. niger culture identified compounds 14, 45 and 47 synthetic performance of fungal hosts. (Chart 8), intermediates that were expected to be involved in Development of those methods, combined with steady se- the pathway leading to the formation of 36. We were also able quencing of microbial genomes, will continue to make iden- to obtain the target product 35 at a yield of 0.2 mg/L from this tification and characterization of natural product biosynthetic system.6) pathways considerably easier. The resulting accumulation of When introducing any exogenous biosynthetic pathway knowledge on a multitude of gene clusters responsible for the into a host, including fungus, the toxicity of the biosynthetic production of diverse and important natural products will help product can impair the host. This problem can be avoided by expand the value and refine the versatility of the heterologous incorporating a self-resistance mechanism into the host that production systems. We expect to see many more examples of confers non-destructive resistance against the target com- heterologous production of complex polyketides, nonribosomal pound. Analysis of the biosynthetic gene cluster of a related peptides and their analogs in the near future. Further efforts compound compactin identified the presence of a gene mlcD into enhancing the robustness of the system should improve within the cluster that is homologous to a hydroxymethylgluta- the product yield from those hosts. This will result in broad- ryl coenzyme A (HMG-CoA) reductase gene. Since lovastatin ened variety and increased yield of compounds that can be and compactin are inhibitors of HMG-CoA reductase, it was prepared, allowing researchers to generate substantial librar- proposed that this gene codes for a resistance-conferring fac- ies of potential pharmacophores whose characteristics are not tor that prevents those natural products from poisoning their necessarily restricted by the limited variation in the chemistry 1164 Vol. 62, No. 12 that enables the synthesis of such libraries. Ultimately, those (1997). techniques should facilitate future drug discovery and posi- 13) Flaherty J. E., Payne G. A., Appl. Environ. Microbiol., 63, 3995– tively impact the industrial practice of therapeutic research 4000 (1997). 14) Bergmann S., Schümann J., Scherlach K., Lange C., Brakhage A. and development. A., Hertweck C., Nat. Chem. Biol., 3, 213–217 (2007). 15) Nakazawa T., Ishiuchi K., Praseuth A., Noguchi H., Hotta K., Wata- Acknowledgments With much appreciation, I would nabe K., ChemBioChem, 13, 855–861 (2012). like to thank Professor Hiroshi Noguchi of the University 16) Ishiuchi K., Nakazawa T., Ookuma T., Sugimoto S., Sato M., Tsu- of Shizuoka for all his support and his nomination for the nematsu Y., Ishikawa N., Noguchi H., Hotta K., Moriya H., Wata- Award for Division Scientific Promotions in 2014. I would nabe K., ChemBioChem, 13, 846–854 (2012). like to express my appreciation to Professors Hisao Moriya of 17) Stocking E. M., Williams R. M., Angew. Chem. Int. Ed. Engl., 42, Okayama University, Hideaki Oikawa of Hokkaido University, 3078–3115 (2003). Chaitan S. Khosla of Stanford University, and Charles Richard 18) Oikawa H., Bull. Chem. Soc. Jpn., 78, 537–554 (2005). Hutchinson of the University of Wisconsin Madison for their 19) Kelly W. L., Org. Biomol. Chem., 6, 4483–4493 (2008). constructive remarks and suggestions. Professor Yi Tang of 20) Oikawa H., “Comprehensive Natural Products II,” Vol. VIII, Chap. 8, ed. by Liu, H.-W., Mander, L., Elsevier, Oxford, 2010, the University of California Los Angeles, our collaborator, pp. 277–314. has provided indispensable and insightful contributions to the 21) Kim H. J., Ruszczycky M. W., Choi S. H., Liu Y. N., Liu H. W., pseurotin project. Finally, I thank deeply the committee mem- Nature (London), 473, 109–112 (2011). bers of the Pharmaceutical Society of Japan for their encour- 22) Kasahara K., Miyamoto T., Fujimoto T., Oguri H., Tokiwano T., agement and their nomination for the Award. Oikawa H., Ebizuka Y., Fujii I., ChemBioChem, 11, 1245–1252 (2010). References 23) Auclair K., Sutherland A., Kennedy J., Witter D. J., Van den Heever 1) Pfeifer B. A., Admiraal S. J., Gramajo H., Cane D. E., Khosla C., J. P., Hutchinson C. R., Vederas J. C., J. Am. Chem. Soc., 122, Science, 291, 1790–1792 (2001). 11519–11520 (2000). 2) Pfeifer B. A., Wang C. C., Walsh C. T., Khosla C., Appl. Environ. 24) Ma S. M., Li J. W., Choi J. W., Zhou H., Lee K. K., Moorthie V. Microbiol., 69, 6698–6702 (2003). A., Xie X., Kealey J. T., Da Silva N. A., Vederas J. C., Tang Y., 3) Watanabe K., Rude M. A., Walsh C. T., Khosla C., Proc. Natl. Science, 326, 589–592 (2009). Acad. Sci. U.S.A., 100, 9774–9778 (2003). 25) Watanabe K., Mie T., Ichihara A., Oikawa H., Honma M., J. Biol. 4) Watanabe K., Hotta K., Praseuth A. P., Koketsu K., Migita A., Chem., 275, 38393–38401 (2000). Boddy C. N., Wang C. C., Oguri H., Oikawa H., Nat. Chem. Biol., 26) Eberhardt S., Zingler N., Kemter K., Richter G., Cushman M., 2, 423–428 (2006). Bacher A., Eur. J. Biochem., 268, 4315–4323 (2001). 5) Ro D. K., Paradise E. M., Ouellet M., Fisher K. J., Newman K. L., 27) Kim R. R., Illarionov B., Joshi M., Cushman M., Lee C. Y., Eisen- Ndungu J. M., Ho K. A., Eachus R. A., Ham T. S., Kirby J., Chang reich W., Fischer M., Bacher A., J. Am. Chem. Soc., 132, 2983–2990 M. C., Withers S. T., Shiba Y., Sarpong R., Keasling J. D., Nature (2010). (London), 440, 940–943 (2006). 28) Ose T., Watanabe K., Mie T., Honma M., Watanabe H., Yao M., 6) Tsunematsu Y., Ishikawa N., Wakana D., Goda Y., Noguchi H., Oikawa H., Tanaka I., Nature (London), 422, 185–189 (2003). Moriya H., Hotta K., Watanabe K., Nat. Chem. Biol., 9, 818–825 29) Guimarães C. R., Udier-Blagović M., Jorgensen W. L., J. Am. (2013). Chem. Soc., 127, 3577–3588 (2005). 7) Primm T. P., Franzblau S. G., Curr. Bioact. Compd., 3, 201–208 30) Serafimov J. M., Gillingham D., Kuster S., Hilvert D., J. Am. Chem. (2007). Soc., 130, 7798–7799 (2008). 8) Ezaki M., Muramatsu H., Takase S., Hashimoto M., Nagai K., J. 31) Yang S. W., Mierzwa R., Terracciano J., Patel M., Gullo V., Wagner Antibiot. (Tokyo), 61, 207–212 (2008). N., Baroudy B., Puar M., Chan T. M., McPhail A. T., Chu M., J. 9) Brakhage A. A., Schroeckh V., Fungal Genet. Biol., 48, 15–22 Nat. Prod., 69, 1025–1028 (2006). (2011). 32) Sato M., Sugiyama T., Yagishita F., Mino T., Noguchi H., Hotta K., 10) Pel H. J., de Winde J. H., Archer D. B., Dyer P. S., Hofmann G., Tang T., Watanabe K., manuscript in preparation. Schaap P. J., Turner G., de Vries R. P., Albang R., Albermann K., 33) Oikawa H., Murakami Y., Ichihara A., J. Chem. Soc., Perkin Trans. Andersen M. R., Bendtsen J. D., Benen J. A., van den Berg M., 1, 1992, 2949–2953 (1992). Breestraat S., Caddick M. X., Contreras R., Cornell M., Coutinho 34) Ishiuchi K., Nakazawa T., Yagishita F., Mino T., Noguchi H., Hotta P. M., Danchin E. G., Debets A. J., Dekker P., van Dijck P. W., van K., Watanabe K., J. Am. Chem. Soc., 135, 7371–7377 (2013). Dijk A., Dijkhuizen L., Driessen A. J., d’Enfert C., Geysens S., 35) Sims J. W., Schmidt E. W., J. Am. Chem. Soc., 130, 11149–11155 Goosen C., Groot G. S., de Groot P. W., Guillemette T., Henrissat (2008). B., Herweijer M., van den Hombergh J. P., van den Hondel C. A., 36) Sekita S., Yoshihira K., Natori S., Kuwano H., Tetrahedron Lett., van der Heijden R. T., van der Kaaij R. M., Klis F. M., Kools H. J., 14, 2109–2112 (1973). Kubicek C. P., van Kuyk P. A., Lauber J., Lu X., van der Maarel M. 37) Sekita S., Yoshihira K., Natori S., Kuwano H., Chem. Pharm. Bull., J., Meulenberg R., Menke H., Mortimer M. A., Nielsen J., Oliver 30, 1629–1638 (1982). S. G., Olsthoorn M., Pal K., van Peij N. N., Ram A. F., Rinas U., 38) Silverton J. V., Akiyama T., Kabuto C., Sekita S., Yoshihira K., Na- Roubos J. A., Sagt C. M., Schmoll M., Sun J., Ussery D., Varga J., tori S., Tetrahedron Lett., 17, 1349–1350 (1976). Vervecken W., van de Vondervoort P. J., Wedler H., Wosten H. A., 39) Scherlach K., Boettger D., Remme N., Hertweck C., Nat. Prod. Zeng A. P., van Ooyen A. J., Visser J., Stam H., Nat. Biotechnol., Rep., 27, 869–886 (2010). 25, 221–231 (2007). 40) Löw I., Jahn W., Wieland T., Sekita S., Yoshihira K., Natori S., 11) Bok J. W., Chiang Y.-M., Szewczyk E., Reyes-Dominguez Y., Anal. Biochem., 95, 14–18 (1979). Davidson A. D., Sanchez J. F., Lo H.-C., Watanabe K., Strauss J., 41) Schümann J., Hertweck C., J. Am. Chem. Soc., 129, 9564–9565 Oakley B. R., Wang C. C. C., Keller N. P., Nat. Chem. Biol., 5, (2007). 462–464 (2009). 42) Nakazawa T., Ishiuchi K., Sato M., Tsunematsu Y., Sugimoto S., 12) An Z., Mei B., Yuan W. M., Leong S. A., EMBO J., 16, 1742–1750 Gotanda Y., Noguchi H., Hotta K., Watanabe K., J. Am. Chem. Soc., December 2014 1165

135, 13446–13455 (2013). S., Jiménez J., Keller N., Khouri H., Kitamoto K., Kobayashi T., 43) Oikawa H., Murakami Y., Ichihara A., Tetrahedron Lett., 32, 4533– Konzack S., Kulkarni R., Kumagai T., Lafton A., Latgé J.-P., Li W., 4536 (1991). Lord A., Lu C., Majoros W. H., May G. S., Miller B. L., Mohamoud 44) Oikawa H., Murakami Y., Ichihara A., Biosci. Biotechnol. Bio- Y., Molina M., Monod M., Mouyna I., Mulligan S., Murphy L., chem., 57, 628–631 (1993). O’Neil S., Paulsen I., Peñalva M. A., Pertea M., Price C., Pritchard 45) Cui C. M., Li X. M., Li C. S., Proksch P., Wang B. G., J. Nat. Prod., B. L., Quail M. A., Rabbinowitsch E., Rawlins N., Rajandream 73, 729–733 (2010). M.-A., Reichard U., Renauld H., Robson G. D., de Córdoba S. R., 46) Wang P., Gao X., Tang Y., Curr. Opin. Chem. Biol., 16, 362–369 Rodríguez-Peña J. M., Ronning C. M., Rutter S., Salzberg S. L., (2012). Sanchez M., Sánchez-Ferrero J. C., Saunders D., Seeger K., Squares 47) Springer J. P., Büchi G., Kobbe B., Demain A. L., Clardy J., Tetra- R., Squares S., Takeuchi M., Tekaia F., Turner G., de Aldana C. R. hedron Lett., 18, 2403–2406 (1977). V., Weidman J., White O., Woodward J., Yu J.-H., Fraser C., Gala- 48) Lian G., Yu B., Chem. Biodivers., 7, 2660–2691 (2010). gan J. E., Asai K., Machida M., Hall N., Barrell B., Denning D. W., 49) Barrow C. J., Cai P., Snyder J. K., Sedlock D. M., Sun H. H., Coo- Nature (London), 438, 1151–1156 (2005). per R., J. Org. Chem., 58, 6016–6021 (1993). 71) Maiya S., Grundmann A., Li X., Li S. M., Turner G., 50) Katagiri K., Sato K., Hayakawa S., Matsushima T., Minato H., J. ChemBio Chem, 8, 1736–1743 (2007). Antibiot. (Tokyo), 23, 420–422 (1970). 72) Wiemann P., Guo C. J., Palmer J. M., Sekonyela R., Wang C. C., 51) Greiner D., Bonaldi T., Eskeland R., Roemer E., Imhof A., Nat. Keller N. P., Proc. Natl. Acad. Sci. U.S.A., 110, 17065–17070 (2013). Chem. Biol., 1, 143–145 (2005). 73) Asami Y., Kakeya H., Onose R., Yoshida A., Matsuzaki H., Osada 52) Popp J. L., Musza L. L., Barrow C. J., Rudewicz P. J., Houck D. R., H., Org. Lett., 4, 2845–2848 (2002). J. Antibiot. (Tokyo), 47, 411–419 (1994). 74) Ando O., Satake H., Nakajima M., Sato A., Nakamura T., Kinoshita 53) Li G. Y., Li B. G., Yang T., Yan J. F., Liu G. Y., Zhang G. L., J. Nat. T., Furuya K., Haneishi T., J. Antibiot. (Tokyo), 44, 382–389 (1991). Prod., 69, 1374–1376 (2006). 75) Igarashi Y., Yabuta Y., Sekine A., Fujii K., Harada K., Oikawa 54) Boyer N., Morrison K. C., Kim J., Hergenrother P. J., Movassaghi T., Sato M., Furumai T., Oki T., J. Antibiot. (Tokyo), 57, 748–754 M., Chem. Sci., 4, 1646–1657 (2013). (2004). 55) Bugni T. S., Ireland C. M., Nat. Prod. Rep., 21, 143–163 (2004). 76) Igarashi Y., Yabuta Y., Furumai T., J. Antibiot. (Tokyo), 57, 537–540 56) Ding G., Jiang L., Guo L., Chen X., Zhang H., Che Y., J. Nat. (2004). Prod., 71, 1861–1865 (2008). 77) Tsunematsu Y., Fukutomi M., Saruwatari T., Noguchi H., Hotta K., 57) Pérez-Balado C., Rodríguez-Graña P., de Lera Á. R., Chemistry, 15, Tang Y., Watanabe K., Angew. Chem. Int. Ed. Engl., 53, 8475–8479 9928–9937 (2009). (2014). 58) Gomes N. M., Dethoup T., Singburaudom N., Gales L., Silva A. M. 78) Allen J. D., van Loevezijn A., Lakhai J. M., van der Valk M., van S., Kijjoa A., Phytochem. Lett., 5, 717–720 (2012). Tellingen O., Reid G., Schellens J. H. M., Koomen G. J., Schinkel 59) Cai S., Kong X., Wang W., Zhou H., Zhu T., Li D., Gu Q., Tetrahe- A. H., Mol. Cancer Ther., 1, 417–425 (2002). dron Lett., 53, 2615–2617 (2012). 79) Cui C.-B., Kakeya H., Osada H., J. Antibiot. (Tokyo), 49, 832–835 60) Rank C., Klejnstrup M. L., Petersen L. M., Kildgaard S., Frisvad J. (1996). C., Gotfredsen C. H., Larsen T. O., Metabolites, 2, 39–56 (2012). 80) Cui C.-B., Kakeya H., Osada H., Tetrahedron, 52, 12651–12666 61) Bérubé G., Curr. Med. Chem., 13, 131–154 (2006). (1996). 62) Chang P. K., Scharfenstein L. L., Wei Q., Bhatnagar D., J. Micro- 81) Edmondson S. D., Danishefsky S. J., Angew. Chem. Int. Ed. Engl., biol. Methods, 81, 240–246 (2010). 37, 1138–1140 (1998). 63) Maiya S., Grundmann A., Li S. M., Turner G., ChemBioChem, 7, 82) Edmondson S., Danishefsky S. J., Sepp-Lorenzino L., Rosen N., J. 1062–1069 (2006). Am. Chem. Soc., 121, 2147–2155 (1999). 64) Movassaghi M., Schmidt M. A., Ashenhurst J. A., Angew. Chem. 83) Wang H., Ganesan A., J. Org. Chem., 65, 4685–4693 (2000). Int. Ed. Engl., 47, 1485–1487 (2008). 84) Li S., Finefield J. M., Sunderhaus J. D., McAfoos T. J., Williams R. 65) Tadano S., Mukaeda Y., Ishikawa H., Angew. Chem. Int. Ed. Engl., M., Sherman D. H., J. Am. Chem. Soc., 134, 788–791 (2012). 52, 7990–7994 (2013). 85) Ames B. D., Liu X., Walsh C. T., Biochemistry, 49, 8564–8576 66) Bloch P., Tamm C., Bollinger P., Petcher T. J., Weber H. P., Helv. (2010). Chim. Acta, 59, 133–137 (1976). 86) Kato H., Nakamura Y., Finefield J. M., Umaoka H., Nakahara T., 67) Bloch P., Tamm C., Helv. Chim. Acta, 64, 304–315 (1981). Williams R. M., Tsukamoto S., Tetrahedron Lett., 52, 6923–6926 68) Breitenstein W., Chexal K. K., Mohr P., Tamm C., Helv. Chim. (2011). Acta, 64, 379–388 (1981). 87) Grubbs A. W., Artman G. D. 3rd, Tsukamoto S., Williams R. M., 69) Yamada T., Imai E., Nakatuji K., Numata A., Tanaka R., Tetrahe- Angew. Chem. Int. Ed. Engl., 46, 2257–2261 (2007). dron Lett., 48, 6294–6296 (2007). 88) Abe Y., Suzuki T., Mizuno T., Ono C., Iwamoto K., Hosobuchi M., 70) Nierman W. C., Pain A., Anderson M. J., Wortman J. R., Kim H. Yoshikawa H., Mol. Genet. Genomics, 268, 130–137 (2002). S., Arroyo J., Berriman M., Abe K., Archer D. B., Bermejo C., 89) Sin N., Meng L., Wang M. Q. W., Wen J. J., Bornmann W. G., Bennett J., Bowyer P., Chen D., Collins M., Coulsen R., Davies Crews C. M., Proc. Natl. Acad. Sci. U.S.A., 94, 6099–6103 (1997). R., Dyer P. S., Farman M., Fedorova N., Fedorova N., Feldblyum 90) Kruger E. A., Figg W. D., Expert Opin. Investig. Drugs, 9, 1383– T. V., Fischer R., Fosker N., Fraser A., García J. L., García M. J., 1396 (2000). Goble A., Goldman G. H., Gomi K., Griffith-Jones S., Gwilliam 91) Lin H. C., Chooi Y. H., Dhingra S., Xu W., Calvo A. M., Tang Y., J. R., Haas B., Haas H., Harris D., Horiuchi H., Huang J., Humphray Am. Chem. Soc., 135, 4616–4619 (2013).