Reconstitution of biosynthetic machinery for indole-diterpene paxilline in Aspergillus oryzae

Article abstract of DOI:10.1021/ja3116636

Indole-diterpenes represented by paxilline share a common pentacyclic core skeleton derived from indole and geranylgeranyl diphosphate. To shed light on the detailed biosynthetic mechanism of the paspaline-type hexacyclic skeleton, we examined the reconstitution of paxilline biosynthetic machinery in Aspergillus oryzae NSAR1. Stepwise introduction of the six pax genes enabled us to isolate all biosynthetic intermediates and to synthesize paxilline. In vitro and in vivo studies on the key enzymes, prenyltransferase PaxC and cyclase PaxB, allowed us to elucidate actual substrates of these enzymes. Using the isolated and the synthesized epoxide substrates, the highly intriguing stepwide epoxidation/cyclization mechanism for the construction of core structure has been confirmed. In addition, we also demonstrated "tandem transformation" to simultaneously introduce two genes using a single vector (paxG/paxB, pAdeA; paxP/paxQ, pUNA). This may provide further option for the reconstitution strategy to synthesize more complex fungal metabolites.

Full text of DOI:10.1021/ja3116636

Communication
pubs.acs.org/JACS
Reconstitution of Biosynthetic Machinery for Indole-Diterpene
Paxilline in Aspergillus oryzae
Koichi Tagami,† Chengwei Liu,‡ Atsushi Minami,† Motoyoshi Noike,‡ Tetsuya Isaka,§ Shuhei Fueki,†
Yoshihiro Shichijo,† Hiroaki Toshima,§ Katsuya Gomi,∥ Tohru Dairi,‡ and Hideaki Oikawa*,†
†Division of Chemistry, Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan
‡Graduate School of Engineering, Hokkaido University, Sapporo 060-8628, Japan
§Department of Bioresource Science, College of Agriculture, Ibaraki University, Inashiki, Ibaraki 300-0393, Japan
∥Graduate School of Agricultural Science, Tohoku University, Sendai 981-8555, Japan
S
* Supporting Information
series of gene inactivation and gene transfer to paxilline
ABSTRACT: Indole-diterpenes represented by paxilline
share a common pentacyclic core skeleton derived from
indole and geranylgeranyl diphosphate. To shed light on
the detailed biosynthetic mechanism of the paspaline-type
hexacyclic skeleton, we examined the reconstitution of
paxilline biosynthetic machinery in Aspergillus oryzae
NSAR1. Stepwise introduction of the six pax genes
enabled us to isolate all biosynthetic intermediates and
to synthesize paxilline. In vitro and in vivo studies on the
key enzymes, prenyltransferase PaxC and cyclase PaxB,
allowed us to elucidate actual substrates of these enzymes.
Using the isolated and the synthesized epoxide substrates,
the highly intriguing stepwide epoxidation/cyclization
mechanism for the construction of core structure has
been confirmed. In addition, we also demonstrated
“tandem transformation” to simultaneously introduce
two genes using a single vector (paxG/paxB, pAdeA;
paxP/paxQ, pUNA). This may provide further option for
the reconstitution strategy to synthesize more complex
fungal metabolites.
negative deletion mutants confirmed that four pax genes, paxG,
paxC, paxM, and paxB, are required for the biosynthesis of 2
(Figure 1A).8,9 Further gene disruption studies of paxP and
paxQ, a cytochrome P450 gene, indicated that PaxP and PaxQ
catalyze multiple oxidation steps from paspaline (2) to paxilline
(1) via 13-desoxypaxilline (9).10 Recently, two more
biosynthetic gene clusters for aflatrem (atm) and lolitrem
(ltm) have been identified and added further information for
the biosynthesis of indole diterpenes.11,12 However, little is
known about the detailed function of these four gene products.
Recently, we carried out heterologous expression of
aphidicolin biosynthetic genes (four genes)13 in Aspergillus
oryzae NSAR1 strain, which accepts five expression vectors
(pTAex3, pUSA, pAdeA, pPTR1, pUNA).14−16 Transformation
of plasmids harboring each gene into A. oryzae in stepwise
manner enabled us to isolate all biosynthetic intermediates in
addition to the natural product, aphidicolin, from correspond-
ing transformants. Accumulated examples of heterologous
production of fungal metabolites suggest that the A. oryzae
expression system allows us to analyze the biosynthetic
machinery of complex natural products. In this study, we
applied this reliable expression system to analyze the functions
of paxG, paxC, paxM, and paxB genes in the biosynthesis of 2
and achieved the total biosynthesis of 1.
ndole-terpenes are fungal and bacterial secondary metabo-
Ilites with unique biological activities such as inhibitory
activity of calcium-activated potassium channels,1 tremorgenic
activity,2 potent and selective progesterone receptor agonistic
activity,3 and anti-MRSA activity.4 The wide range of biological
activities are dependent on their structural diversity. Indole-
diterpenes share a common core structure consisting of a cyclic
diterpene skeleton derived from indole and geranylgeranyl
diphosphate (GGPP). Further modifications, such as prenyla-
tion, hydroxylation, and chlorination, increase members of this
family to include compounds such as nodulisporic acid,
lolitrem, thiersinine, and penitrem.
Among the four genes involved in paspaline biosynthesis,
bioinformatics analysis indicated that the putative gene
products PaxG, a geranylgeranyl diphosphate synthase, and
PaxC, a prenyltransferase, participate in the construction of
GGI 4. To examine the functions of PaxG and PaxC, A. oryzae
NSAR1 strain was transformed with pAdeA-paxG and pTAex3-
paxC. Transformants, that efficiently produced novel com-
pounds compared with the wild type strain, were screened by
monitoring of their mycelial extracts by TLC or HPLC. HPLC
analysis showed that the double transformant produced a new
compound 4 eluting with the same retention time as authentic
GGI (production of 4: 7.7 mg/L) (Figure 1B). After
A series of feeding experiments with 13C-labeled acetate
revealed that the terpenoid core structure is derived from
GGPP.5,6 The structure of less modified indole diterpene
paspaline (2) suggested that 2 is a cyclization product of a
putative common intermediate 3-geranylgeranyl indole (GGI,
4). This was experimentally supported by incorporation of
deuterium labeled 4.7 In 2001, the first biosynthetic gene
cluster of this family for paxilline (pax) was identified.8 Then, a
1
purification, the structure was also confirmed by H NMR
analysis in comparison with authentic GGI. These experiments
proved that paxC encodes a geranylgeranyl transferase using
Received: November 29, 2012
Published: January 11, 2013
© 2013 American Chemical Society
1260
dx.doi.org/10.1021/ja3116636 | J. Am. Chem. Soc. 2013, 135, 1260−1263

Journal of the American Chemical Society
Communication
Figure 1. (A) Biosynthetic pathway of paxilline. (B) HPLC profiles of the extracts from A. oryzae transformant: (i) transformant harboring paxG,
paxC, paxM, paxB, paxP, and paxQ; (ii) transformant harboring paxG, paxC, paxM, and paxB; (iii) transformant harboring paxG, paxC, and paxM;
(iv) transformant harboring paxG and paxC; (v) wild type strain. *: impurity.
GGPP, and that, as suggested in our previous experiment,7 GGI
4 is an intermediate in paxilline biosynthesis.
simultaneous prenylation and decarboxylation in the menaqui-
none biosynthesis of Escherichia coli.18 PaxC shared a significant
sequence similarity with typical polyprenyl synthetases having
DDxxD motif but not with members of indole prenyltrans-
ferases belonging to the fungal tryptophan dimethylallyltrans-
ferase (DMATS) superfamily (Figure S7).19
On the basis of the results of previous feeding experiments
using tryptophan or anthranilic acid,6,17 it was proposed that
either tryptophan or its precursor, such as indole-3-glycerol
phosphate (IGP: 3), is a substrate in the GGI forming reaction.
To determine the substrate of this prenylation step, we then
carried out in vitro analysis using recombinant PaxC. PaxC
cDNA was cloned into the pQE30 vector and expressed as an
N-terminal His6-tagged fusion protein. Recombinant PaxC was
purified by Ni-NTA column chromatography (Figure S2). IGP
3 was subjected to PaxC reaction with GGPP and Mg2+.
Subsequent HPLC analysis of the reaction products showed the
prominent peak of GGI 4 dependent on PaxC and Mg2+
(Figure S3). Importantly, 4 was also detected in the PaxC
reaction with indole (Figure S4). The kinetic parameters
revealed that IGP is a preferred substrate for PaxC (kcat/KM for
3, 28.2 mM−1 s−1; indole, 3.6 mM−1 s−1) (Table S2). In
contrast, trace amount of 4 was detected in the PaxC reaction
with tryptophan (Figure S4). Further LC-MS analysis revealed
that PaxC reaction accompanied the formation of glyceralde-
hyde 3-phosphate as a byproduct of PaxC reaction (Figure S5).
These results indicated that PaxC catalyzed geranylgeranylation
at C3 position of 3 followed by elimination of glyceraldehyde 3-
phosphate (Figure S6). To our knowledge, the preceding
example is an octaprenyltransferase MenA which catalyzes
We then turned our attention to the subsequent epoxidation
and cyclization mechanism. As paxM showed a high similarity
to flavin-dependent monooxygenase, which catalyzes various
reactions including epoxidation,20 the double transformant was
then transformed with pUSA-paxM. Chromatograms of extracts
from the triple transformant showed two new peaks (5 and 10)
with typical absorption spectra of indole skeleton (production
of 5: 32 mg/L) that were not present in the double
1
transformant (Figure 1B, Figure S8). H NMR spectrum of 5
displayed characteristic signals corresponding to indole and
terpene moieties. In particular, the triplet at 2.71 ppm indicated
the presence of epoxide. HMBC and HSQC analysis revealed
that the epoxide was installed at the C10−C11 position of GGI.
Another metabolite 10 was confirmed as a diol analog of 5
based on a series of NMR analysis and LC-MS analysis (Figure
S8, S9). Hydrolysis of the epoxide moiety in a heterologous
host A. oryzae was reported previously.21,22 To shed light on the
final cyclization step, we then introduced the paxB gene
(pAdeA-paxB), an integral membrane protein, in addition to
the paxG gene (pAdeA-paxG) into double transformant (paxC
1261
dx.doi.org/10.1021/ja3116636 | J. Am. Chem. Soc. 2013, 135, 1260−1263

Journal of the American Chemical Society
Communication
+ paxM). HPLC analysis of the extracts from this quadruple
transformant showed a new peak 2 (production of 2: 57 mg/L)
(Figure 1B). The 1H NMR spectrum of 2 was identical to that
of authentic paspaline.
Isolation of 5 and 2 from the triple and the quadruple
transformants suggested that initial cyclization of 5 yields
emindole SB (6)23 and a second round of epoxidation/
cyclization affords 2 (Figure 1A). To prove this stepwise
epoxidation/cyclization mechanism, we carried out a bio-
transformation study using two epoxide substrates 5 and 8.
Bisepoxide 8 was synthesized by the sequential Sharpless
asymmetric dihydroxylations in a highly stereocontrolled
manner (9,10-epoxide, 90% d.e.; 13,14-epoxide, 98% d.e.) as
shown in Supporting Information. Next, A. oryzae NSAR1 wild
type strain was transformed with the expression plasmid
pTAex3-paxB. Single paxB transformant was screened by
semiquantitative RT-PCR (Figure S10),8 monitoring the
expression level of paxB at different growth times. This also
enabled us to determine optimal fermentation period as 2.5
days for the biotransformation study. Incubation with 5 led to
the formation of 6 as determined by HPLC analysis only in the
extract from the transformant (Figure 2). Importantly this
Figure 3. Biosynthetic model for various indole terpenes.
by two different types of cyclization (scaffold A and scaffold B).
The characteristic THP ring system in indole diterpenes
represented by paspaline is constructed by a second cycle of
epoxidation/cyclization catalyzed by the same epoxidase and
terpene cyclase. This adds further structural diversity to this
family.
Finally, we focused on the heterologous production of
paxilline (1) in A. oryzae. For this purpose, it would be
necessary to introduce six genes, although currently only five
vectors are available (pPTR1, pTAex3, pUNA, pUSA, pAdeA).
Therefore, we examined “tandem transformation” using a single
vector carrying two individual genes paxP and paxQ into the
quadruple transformant which produces 2. When selection was
carried out on minimal medium lacking ammonium sulfate, the
desired transformant harboring paxP and paxQ (four of seven
positive clones) in addition to paxP and paxQ transformants
(paxP, one clone; paxQ, two clones) were isolated. As
1
expected, HPLC and H NMR analysis showed that the paxP
and the paxP + paxQ transformants produced 9 and 1,
respectively (production of 1: 35 mg/L) (Figure 1B, Figure
S11).
Figure 2. HPLC profiles of the biotransformation experiments: (i)
authentic 2 and 6; (ii) with 8; (iii) authentic 8; (iv) with 5; (v) 5 from
triple transformant; (vi) without substrate.
In this study, we successfully demonstrated stepwise
reconstruction of biosynthetic machinery of the target natural
product 1 by introducing six genes into A. oryzae. This
methodology enabled us to characterize each gene and to
obtain sufficient amount of intermediates for in vitro enzymatic
transformation. The same strategy has been applied to
synthesize natural products, such as polyketide-peptide hybrid
tenellin (four genes),25 diterpene aphidicolin (four genes),13
and late biosynthetic intermediates, such as meroterpenoids
pyripyropene (two + three genes)21 and terretonin (five
genes).22 The synthesis of complex fungal metabolites involving
more than ten biosynthetic genes requires practical methods for
introduction of multiple genes. For this purpose, we performed
“tandem transformation” to simultaneously indroduce two
genes using a single vector (paxG/paxB, pAdeA; paxP/paxQ,
pUNA). This “tandem transformation” in combination with the
conventional method using single plasmids carrying two
different genes enabled us to solve the vector problem for
heterologous production of fungal metabolites.
transformant also converted 8 into 2. These results confirmed
the stepwise epoxidation and cyclization mechanism and
supported the involvement of the pentacyclic intermediate 7
(Figure 2). The core structure formation is different from that
of ionophore polyether lasalocid that is biosynthesized from
sequential epoxidation and subsequent cyclization.20
To date, biosynthetic gene clusters of three fungal indole
diterpenes such as paxilline,8 aflatrem,11 and lolitrem12 and one
bacterial indole sesquiterpene, xiamycin,24 were identified. Key
genes for epoxidation and cyclization, paxM like flavin-
dependent monooxygenase (aflatrem, atmM; lolitrem, ltmM;
xiamycin, xiaO) and paxB like membrane protein (aflatrem,
atmB; lolitrem, ltmB; xiamycin, xiaH) are highly conserved in
all of these gene clusters. Therefore, the stepwise epoxidation/
cyclization mechanism to construct the hexacyclic skeleton of
paxilline can be extended to account for the biosynthetic
machinery of these types of indole terpenes; scaffold A (C15,
indosespene; C20, emindole SA) and scaffold B (C15,
sespendole; C20, paspaline) (Figure 3). It should be noted
that each core structure is formed by the regioselective
epoxidation toward a third olefin from the indole ring followed
In summary, we elucidated the detailed reaction mechanism
of paspaline biosynthesis using stepwise introduction of
biosynthetic genes. Bioconversion of the resultant trans-
formants and in vitro study enabled determination of the
1262
dx.doi.org/10.1021/ja3116636 | J. Am. Chem. Soc. 2013, 135, 1260−1263

Journal of the American Chemical Society
Communication
(16) Yamada, O.; Na Nan, S.; Akao, T.; Tominaga, M.; Watanabe,
H.; Satoh, T.; Enei, H.; Akita, O. J. Biosci. Bioeng. 2003, 95, 82.
(17) Byrne, K. M.; Smith, S. K.; Ondeyka, J. G. J. Am. Chem. Soc.
2002, 124, 7055.
(18) Suvarna, K.; Stevenson, D.; Meganathan, R.; Hudspeth, M. E. S.
J. Bacteriol. 1998, 180, 2782.
actual intermediate of geranylgeranyl transferase PaxC and to
establish two rounds of epoxidation/cyclization for construc-
tion of the characteristic paspaline scaffold. Our strategy can be
applied to produce known or unknown natural products for
which biosynthetic genes are available.
(19) (a) Li, S.-M. Phytochemistry 2009, 70, 1746. (b) Li, S.-M. Nat.
Prod. Rep. 2010, 27, 57.
(20) Minami, A.; Shimaya, M.; Suzuki, G.; Migita, A.; Shinde, S. S.;
Sato, K.; Watanabe, K.; Tamura, T.; Oguri, H.; Oikawa, H. J. Am.
Chem. Soc. 2012, 134, 7246.
ASSOCIATED CONTENT
* Supporting Information
Experimental procedures, HR-MS, NMR data, spectra and LC-
MS data. This material is available free of charge via the
Internet at http://pubs.acs.org.
■
S
(21) Itoh, T.; Tokunaga, K.; Matsuda, Y.; Fujii, I.; Abe, I.; Ebizuka,
Y.; Kushiro, T. Nat. Chem. 2010, 2, 858.
(22) Matsuda, Y.; Awakawa, T.; Itoh, T.; Wakimoto, T.; Kushiro, T.;
Fujii, I.; Ebizuka, Y.; Abe, I. ChemBioChem 2012, 13, 1738.
(23) Nozawa, K.; Yuyama, M.; Nakajima, S.; Kawai, K.; Udagawa, S.
J. Chem. Soc., Perkin Trans. 1 1988, 2155.
AUTHOR INFORMATION
Corresponding Author
hoik@sci.hokudai.ac.jp
■
(24) Li, H.; Zhang, Q.; Li, S.; Shu, Y.; Zhang, G.; Zhang, H.; Tian, X.;
Zhang, S.; Ju, J.; Zhang, C. J. Am. Chem. Soc. 2012, 134, 8996.
(25) Heneghan, M. N.; Yakasai, A. A.; Halo, L. M.; Song, Z.; Bailey,
A. M.; Simpson, T. J.; Cox, R. J.; Lazarus, C. M. ChemBioChem 2010,
11, 1508.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We are grateful to Prof. Y. Ito, Shinshu University, for a
generous gift of a strain P. paxilli. This work was supported by
the Grants-in-Aid for Scientific Research [22108002 to H.O.,
23108101 to T.D.] from Japan Society for the Promotion of
Science (JSPS).
REFERENCES
■
(1) Knaus, H. −G.; McManus, O. B.; Lee, S. H.; Schmalhofer, W. A.;
Garcia-Calvo, M.; Helms, L. M. H.; Sanchez, M.; Giangiacomo, K.;
Ruben, J. P.; Smith, A. B., III.; Kaczorowski, G. J.; Garcia, M. L.
Biochemistry 1994, 33, 5819.
(2) Miles, C. O.; Wilkins, A. L.; Gallagher, R. T.; Hawkes, A. D.;
Munday, S. C.; Towers, N. R. J. Agric. Food Chem. 1992, 40, 234.
(3) Roll, D. M.; Barbieri, L. R.; Bigelis, R.; McDonald, L. A.; Arias, D.
A.; Chang, L. -P.; Singh, M. P.; Luckman, S. W.; Berrodin, T. J.; Yudt,
M. R. J. Nat. Prod. 2009, 72, 1944.
(4) Ogata, M.; Ueda, J.; Hoshi, M.; Hashimoto, J.; Nakashima, T.;
Anzai, K.; Takagi, M.; Shin-ya, K. J. Antibiot. 2007, 60, 645.
(5) Acklin, W.; Weibel, F.; Arigoni, D. Chimia 1977, 31, 63.
(6) de Jesus, A. E.; Gorst-Allman, C. P.; Steyn, P. S.; van Heerden, F.
R.; Vieggaar, R.; Wessels, P. L.; Hull, W. E. J. Chem. Soc., Perkin Trans.
1983, 1863.
(7) Fueki, S.; Tokiwano, T.; Toshima, H.; Oikawa, H. Org. Lett.
2004, 6, 2697.
(8) Young, C.; McMillan, L.; Telfer, E.; Scott, B. Mol. Microbiol.
2001, 39, 754.
(9) Saikia, S.; Parker, E. J.; Koulman, A.; Scott, B. FEBS Lett. 2006,
580, 1625.
(10) (a) McMillan, L. K.; Carr, R. L.; Young, C. A.; Astin, J. W.;
Lowe, R. G. T.; Parker, E. J.; Jameson, G. B.; Finch, S. C.; Miles, C. O.;
McManus, O. B.; Schmalhofer, W. A.; Garcia, M. L.; Kaczorowski, G.
J.; Goetz, M.; Tkacz, J. S.; Scott, B. Mol. Gen. Genomics 2003, 270, 9.
(b) Saikia, S.; Parker, E. J.; Koulman, A.; Scott, B. J. Biol. Chem. 2007,
282, 16829.
(11) Nicholson, M. J.; Koulman, A.; Monahan, B. J.; Pritchard, B. L.;
Payne, G. A.; Scott, B. Appl. Environ. Microbiol. 2009, 75, 7469.
(12) Young, C. A.; Bryant, M. K.; Christensen, M. J.; Tapper, B. A.;
Bryan, G. T.; Scott, B. Mol. Gen. Genomics 2005, 274, 13.
(13) (a) Fujii, R.; Minami, A.; Tsukagoshi, T.; Sato, N.; Sahara, T.;
Ohgiya, S.; Gomi, K.; Oikawa, H. Biosci. Biotechnol. Biochem. 2011, 75,
1813. (b) Efficient production of aphidicolin has observed in the
fermentation of quadruple transformant using solid medium instead of
normal liquid medium (>130 mg/kg rice).
(14) Jin, F. H.; Maruyama, J.; Juvvadi, P. R.; Akioka, M.; Kitamoto, K.
FEMS Microbiol. Lett. 2004, 239, 79.
(15) Fujii, T.; Yamaoka, H.; Gomi, K.; Kitamoto, K.; Kumagai, C.
Biosci. Biotechnol. Biochem. 1995, 59, 1869.
1263
dx.doi.org/10.1021/ja3116636 | J. Am. Chem. Soc. 2013, 135, 1260−1263