Heterologous expression of highly reducing polyketide synthase involved in betaenone biosynthesis

Article abstract of DOI:10.1039/c4cc09512j

A unique highly reducing polyketide synthase (HR-PKS) with a reductase domain was identified in a betaenone biosynthetic gene cluster. Successful heterologous expression and characterization of the HR-PKS and trans-acting enoyl reductase (ER) provide insights into the core structure formation with a decalin scaffold and allow reconstitution of the betaenone biosynthetic machinery.

Full text of DOI:10.1039/c4cc09512j

ChemComm
View Article Online
View Journal
COMMUNICATION
Heterologous expression of highly reducing
polyketide synthase involved in betaenone
Cite this:DOI: 10.1039/c4cc09512j
biosynthesis†
Received 28th November 2014,
Accepted 15th December 2014
Takahiro Ugai,a Atsushi Minami,a Ryuya Fujii,a Mizuki Tanaka,b Hiroki Oguri,a
Katsuya Gomib and Hideaki Oikawa*a
DOI: 10.1039/c4cc09512j
www.rsc.org/chemcomm
A unique highly reducing polyketide synthase (HR-PKS) with a tricycle[6.2.2.0]dodecane skeleton derived from an intra-
reductase domain was identified in a betaenone biosynthetic gene molecular aldol reaction of 3, which harbors a b-ketoaldehyde
cluster. Successful heterologous expression and characterization of side chain attached to a decalin scaffold. Feeding experiments
the HR-PKS and trans-acting enoyl reductase (ER) provide insights with isotopically labeled precursors suggested that the carbon
into the core structure formation with a decalin scaffold and allow skeleton is constructed in a common polyketide pathway
reconstitution of the betaenone biosynthetic machinery.
catalyzed by HR-PKS.8 Treatment with a cytochrome P450
inhibitor, ancymidol, enabled us to isolate probetaenone I (4).9
Incorporation of isotopically labeled 4 into 2 showed that 4 is
a biosynthetic intermediate in betaenone B biosynthesis.10
Taking together these results, the biosynthetic pathway of 1 was
proposed as shown in Scheme 1. Of particular interest in the
skeletal construction are the plausible direct reductive cleavage of
the linear polyketide chain to give the b-ketoaldehyde moiety found
in dehydroprobetaenone I (5), a putative biosynthetic intermediate
in the production of 1, and [4+2] cycloaddition to construct
the trans-decalin skeleton. In this communication, we report the
identification of a betaenone biosynthetic gene cluster and the
heterologous expression of HR-PKS in Aspergillus oryzae to elucidate
the characteristic chain-release mechanism and decalin scaffold
formation. The successful functional characterization of HR-PKS
allows us to reconstitute the biosynthetic machinery of betaenone B.
Isolation of probetaenone I and betaenone C as biosynthetic
intermediates showed an involvement of the reductase (R)
domain of HR-PKS, a typical domain that catalyzes the reduc-
tive release of the aromatic polyketide chain in the NR–PKS
system, and oxidative modification enzymes. The draft genome
sequence of P. betae revealed that the genome contains six
HR-PKS genes, of which two HR-PKSs have a terminal reductase
(R) domain. Of these two HR-PKS-encoding open reading
frames, we assumed that the HR-PKS gene in the gene cluster
also containing a cytochrome P450 gene is responsible for
betaenone biosynthesis. This putative biosynthetic gene cluster
(bet) encoded four genes, as shown in Fig. 1 and Table S1 (ESI†).
Bet1 exhibits similarity to HR-PKS and is composed in a
KS-AT-DH-MT-ER0-KR-ACP-R domain order, which is identical
in domain organization to those of typical HR-PKSs, except for
the C-terminal R domain and inactive ER domain (see below).
Fungal polyketides are structurally diverse secondary metabolites
exhibiting a wide range of biological activities.1 Representative
examples are the Ras farnesylation inhibitor phomoidride2 and the
cholesterol-lowering agents lovastatin3 and zaragozic acid.4 Assembly
of the carbon skeleta is catalyzed by type I polyketide synthases
(PKSs).5,6 Unlike the bacterial multi-modular type I PKSs, fungal type
I PKSs have a single modular architecture and iteratively use a single
set of active sites through multiple catalytic cycles.5,6 According to the
domain organization, fungal type I PKSs are classified into non-
reducing (NR), partially reducing (PR), and highly reducing (HR)
PKSs. Among them, HR-PKSs contain three fundamental chain
extension domains, b-ketoacyl synthase (KS), acyltransferase (AT),
and the acyl carrier protein (ACP), and additional b-keto-processing
domains, ketoreductase (KR), dehydratase (DH), and enoyl reductase
(ER). Branched methyl groups on the polyketide chain are intro-
duced by the catalysis of methyltransferase (MT). Intriguingly, most
HR-PKSs lack a chain-release domain, and limited information is
available on the chain-release mechanism. This is one of the bottle-
necks in the analysis of the function of HR-PKSs and the reconstitu-
tion of the biosynthetic machinery.
Betaenones A–C (1–3) are phytotoxic polyketides isolated
from Phoma betae Fr., the causal fungus of leaf spot disease
in sugar beets (Scheme 1).7 Among betaenones, 3 exhibits the
highest phytotoxic activity causing wilting of the host plant.7
The unique structural feature of 1 is the highly substituted
a Division of Chemistry, Graduate School of Science, Hokkaido University, Japan.
E-mail: hoik@sci.hokudai.ac.jp
b Graduate School of Agricultural Science, Tohoku University, Japan
† Electronic supplementary information (ESI) available: Experimental proce-
dures, LC-MS data, multiple sequence alignment, phylogenetic analysis, confor-
mational search of 5, and 1H and 13C NMR spectra. See DOI: 10.1039/c4cc09512j Bet2 is a cytochrome P450, a typical oxidation enzyme that
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun.

View Article Online
Communication
ChemComm
Scheme 1 Proposed biosynthetic pathway of betaenone A. Methyl groups derived from SAM are marked with an asterisk.
metabolite (188 mg per kg of rice) was observed only in
the extracts obtained from bet1/3 transformants (Fig. 2(C) and
Fig. S2(A), ESI†). Based on HR-MS data (m/z 319.2682 [M + H]+),
the molecular formula of the new metabolite was determined as
C21H34O2, which is consistent with that of dehydroprobetaenone
Fig. 1 Genetic organization of the betaenone biosynthetic gene cluster.
1
activates an inert C–H bond. Bet3 exhibits similarity to ER and
might act as a functional alternative to the inactive ER domain
of Bet1. Bet4 exhibits homology to short-chain dehydrogenase,
possibly involved in a modification reaction (Table S1, ESI†).
Two genes, orf1 and orf2, are located upstream of bet4 and
encode a putative FAD-dependent oxidase and dehydrogenase,
respectively (Table S1, ESI†).
To characterize the function of individual biosynthetic enzymes,
we applied a heterologous expression system in A. oryzae, a
promising method to elucidate the biosynthetic machinery of
fungal metabolites. Representative examples are found in the total
biosynthesis of tenellin,11 aphidicolin,12 paxilline,13 aspyridone,14
aflatrem,15 and anditomin (11 + 1 genes).16 For skeletal construc-
tion of 1 in A. oryzae, bet1 and bet3 genes were subcloned into
fungal expression vectors, pUARA2 (argB marker) and pUSA2
(sC marker),15 to construct pUARA2-bet1 and pUSA2-bet3, which were
then introduced into A. oryzae NSAR1 to construct both a bet1 single
transformant and a bet1/3 double transformant (Fig. S1(A), ESI†).
The metabolite profile of each transformant cultured on a solid
medium was examined by LC-MS (Fig. 2(A)–(C)). One new
I (5). In addition, the H-NMR spectrum of the new metabolite
was similar to that of 4 except for characteristic signals of an enol
form of the b-ketoaldehyde moiety at 15.5 (s), 7.39 (m), and 5.41
(d, J = 4.9 Hz) ppm (see ESI†), suggesting production of 5 in the
bet1/3 transformant. Treatment of the metabolite with NaBH4
resulted in reduction of the aldehyde moiety (Scheme S1, ESI†).
The 1H-NMR spectrum of the product completely agreed with
that of 4,9 allowing us to determine the structure of the observed
metabolite as 5. These results showed that the collaborative
action of Bet1 and Bet3 plays a key role in the skeletal construc-
tion of 5 in betaenone biosynthesis.
After identifying the observed metabolite of the bet1/3
transformant, we then focused on the subsequent reduction
step. Since Bet4, which is homologous to short-chain dehydro-
genase, is likely to be involved in the reduction of 5, bet4 was
subcloned into the NheI site of the pAdeA2 vector (adeA marker;
Fig. S3, ESI†) to construct pAdeA2-bet4. The bet1/3 double
transformant was then transformed with pAdeA2-bet4 to create
a bet1/3/4 transformant (Fig. S1(B), ESI†). The triple trans-
formant produced a new metabolite (17.4 mg kgÀ1; Fig. 2(D)
and Fig. S2(B), ESI†) having a retention time and the MS
spectrum identical to those of authentic 4,17 demonstrating that
Bet4 catalyzes reduction of 5 to 4 in betaenone biosynthesis.
Next, we turned our attention to the oxidative modification
of the decalin scaffold. The plausible candidate bet2, a cyto-
chrome P450 gene, was subcloned into the NheI site of the
pAdeA2 vector to construct pAdeA2-bet2, which was used for
transformation of the bet1/3 double transformant, creating a
bet1/2/3 transformant (Fig. S1(C), ESI†). LC-MS analysis of the
extracts showed production of a new metabolite (153 mg kgÀ1
)
(Fig. 2(E) and Fig. S2(C), ESI†). The retention time and MS
spectrum were identical to those of authentic betaenone B
(Fig. S2(D), ESI†), thus revealing that this new metabolite 2 is
betaenone B. The structure of 2 was further confirmed by
1H-NMR analysis (see ESI†). These results suggested that Bet2
Fig. 2 LC-MS profiles of metabolites produced by transformants.
(A) A. oryzae NSAR1. (B) bet1 transformant. (C) bet1/3 transformant.
(D) bet1/3/4 transformant. (E) bet1/2/3 transformant.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2014

View Article Online
ChemComm
Communication
catalyzes multistep oxidations from 5 to 3 and that an enzyme
existing in A. oryzae unexpectedly mediates the reduction of the
b-ketoaldehyde moiety of 3 to give 2. Introduction of orf1 or orf2
in the bet1/2/3 transformant did not give a new metabolite,
suggesting that functionally unassigned genes adjacent to the bet
gene cluster can be involved in the aldol reaction of 3 to afford 1.
In the present transformation studies, the frequency rate (the ratio
of functionally active clones to obtained clones) is summarized as
follows: bet1/3 (2/5), bet1/3/4 (4/6), and bet1/2/3 (5/5). Notably,
in all cases, we obtained the functionally active clone in a single
transformation, demonstrating that heterologous expression
in A. oryzae is a reliable method to characterize the function
of biosynthetic genes.
Currently, to our knowledge, polyketide chain-release
mechanisms of HR-PKSs are mainly classified into the following
three mechanisms (Table S2, ESI†): (1) spontaneous cleavage
accompanying pyrone formation found in solanapyrone18 and
alternapyrone19 biosyntheses, (2) polyketide chain transfer from
HR-PKS to the starter unit:ACP transacylase (SAT) domain of
NR–PKS in several HR-PKS–NR–PKS systems,5,6 and (3) enzymatic
cleavage of the polyketide chain catalyzed by thioesterase and
acyltransferase reported in lovastatin biosynthesis.20,21 Intriguingly,
in these cases, no inherent chain-release domain is installed on the
HR-PKS. By contrast, the production of 5 in the bet1/3 transformant
indicated that the C-terminal R domain of Bet1 catalyzes the
reductive release of the polyketide chain, differentiating the func-
tion of Bet1 from that of other HR-PKSs. To reveal the function of
the conserved SYK catalytic triad (S2705/Y2737/K2741)22 found in
the Bet1-R domain (Fig. S4, ESI†), mutational analysis was then
conducted. In contrast to the bet1/3 transformant, no detectable
level of 5 was produced by double transformants with a bet1
mutant gene (S2705A, Y2737F, and K2741A), suggesting the
importance of the catalytic triad in reductive cleavage of
the polyketide chain (Fig. S5(D)–(F), ESI†). Unfortunately, the
chemical release of the PKS-bound mature polyketide chain by
alkaline treatment23 failed.
Scheme 2 Proposed mechanism of decalin scaffold formation in (A)
solanapyrone A, (B) lovastatin, and (C) betaenone biosyntheses.
A previous synthetic study of betaenone,17 in which a linear
triene gave modified probetaenone I as the sole product sug-
gested that existing stereogenic centers of the triene precursor
control diastereoselectivity, and the role of the enzyme is
Fig. 3 Lowest energy conformer of 5. Red and black balls show oxygen
and carbon atoms.
acceleration of the reaction rate in the [4+2] cycloaddition search using MacroModel. All low-energy conformers obtained
possibly by stabilizing the transition state and activating the (o8 kJ molÀ1) were essentially the same as a decalin moiety
dienophile. Currently, enzymatic decalin scaffold formation by and just differed in the orientation of secondary butyl and
[4+2] cycloaddition is firmly established in the biosyntheses of b-hydroxyenone groups (Fig. 3 and Fig. S6, ESI†). These results were
the two fungal polyketides, solanapyrone18,24 and lovastatin23 supported by 1H-NMR data. The highly congested carbonyl group in
(Scheme 2(A) and (B)). In the latter case, the HR-PKS LovB the conformers of 5 and probably in intermediate II-2 likely shows
catalyzes a Diels–Alder reaction of intermediate I during the chain resistance to nucleophilic attack of NaBH4 and enolate from the
extension process (Scheme 2(B)). Our experimental evidence that malonyl unit, respectively. In the previous report, we pointed out
the bet1/3 transformant gave a [4+2] cycloadduct suggested that that prosolanapyrone synthase (PSS) and LovB/MlcA have a similar
the HR-PKS Bet1 may mediate a Diels–Alder reaction of inter- domain architecture and synthesize a common hexaketide pre-
mediate II-1 followed by additional chain extension of inter- cursor, but PSS did not catalyze the [4+2] cycloaddition, whereas
mediate II-2 to afford the core structure of 5 (Scheme 2(C); LovB/MlcA did.18 Based on these data, the [4+2] cycloaddition seems
pathway 1). However, the unusual observation that only the to occur after further chain elongation. This suggests two alternative
terminal carbonyl group of 5 was reduced by the small reducing triene intermediates II-3 and II-4 for the cycloaddition in betaenone
agent NaBH4 questioned this scenario. To examine the reactivity biosynthesis (Scheme 2(C); pathway 2 and 3). Further experimental
of the carbonyl group in 5, we performed a conformational data are required to determine the actual biosynthetic pathway.
This journal is ©The Royal Society of Chemistry 2014
Chem. Commun.

View Article Online
Communication
ChemComm
(b) Fungal Metabolites, ed. W. B. Turner, Academic Press Inc, 1982,
vol. 2.
2 (a) T. T. Dabrah, H. J. Harwood, L. H. Huang, N. D. Jankovich,
T. Kaneko, J.-C. Li, S. Lindsey, P. M. Moshier, T. A. Subashi,
M. Therrien and P. C. Watts, J. Antibiot., 1997, 50, 9933;
(b) T. T. Dabrah, T. Kaneko, W. Massefski and E. B. Whipple,
J. Am. Chem. Soc., 1997, 119, 1594.
3 A. W. Alberts, J. Chen, G. Kuron, V. Hunt, J. Huff, C. Hoffman,
J. Rothrock, M. Lopez, H. Joshua, E. Harris, A. Patchett, R. Monaghan,
S. Currie, E. Stapley, G. Albers-Schonberg, O. Hensens, J. Hirshfield,
K. Hoogsteen, J. Liesch and J. Springer, Proc. Natl. Acad. Sci. U. S. A.,
1980, 77, 3957.
4 M. J. Dawson, J. E. Farthing, P. S. Marshall, R. F. Middleton,
M. J. O’Neill, A. Shuttleworth, C. Stylli, R. M. Tait, P. M. Taylor,
H. G. Wildman, A. D. Buss, D. Langly and M. V. Hayes, J. Antibiot.,
1992, 45, 639.
5 Y. H. Chooi and Y. Tang, J. Org. Chem., 2012, 77, 9933.
6 R. J. Cox and T. J. Simpson, in Comprehensive Natural Products II:
Chemistry and Biology, 10 Volume Set, ed. L. Mander and H. W. Liu,
Elsevier, Oxford, 2010, vol. 1, pp. 347–383.
7 (a) A. Ichihara, H. Oikawa, K. Hayashi and S. Sakamura, J. Am. Chem.
Soc., 1983, 105, 2907; (b) A. Ichihara, H. Oikawa, M. Hashimoto,
S. Sakamura, T. Haraguchi and H. Nagano, Agric. Biol. Chem., 1983,
47, 2965.
8 H. Oikawa, A. Ichihara and S. Sakamura, J. Chem. Soc., Chem.
Commun., 1984, 814.
9 H. Oikawa, A. Ichihara and S. Sakamura, J. Chem. Soc., Chem.
Commun., 1988, 600.
10 H. Oikawa, A. Ichihara and S. Sakamura, J. Chem. Soc., Chem.
Commun., 1990, 908.
11 M. N. Heneghan, A. A. Yakasai, L. M. Halo, Z. Song, A. M. Bailey,
T. J. Simpson, R. J. Cox and C. M. Lazarus, ChemBioChem, 2010,
11, 1508.
12 R. Fujii, A. Minami, T. Tsukagoshi, N. Sato, T. Sahara, S. Ohgiya,
K. Gomi and H. Oikawa, Biosci., Biotechnol., Biochem., 2011, 75, 1813.
13 K. Tagami, C. Liu, A. Minami, M. Noike, T. Isaka, S. Fueki,
Y. Shichijo, H. Toshima, K. Gomi and H. Oikawa, J. Am. Chem.
Soc., 2013, 135, 1260.
14 Z. Wasil, K. A. K. Pahirulzaman, C. Butts, T. J. Simpson, C. M. Lazarus
and R. J. Cox, Chem. Sci., 2013, 4, 3845.
Most HR-PKSs possess a functionally active ER domain
(cis-ER) (Table S2, ESI†). By contrast, a prominent feature of
the Bet1/3 system is that the cis-ER of Bet1 is inactive and Bet3
participates in the polyketide chain construction as a trans-
acting ER (trans-ER). A similar collaborative action of trans-ER
is frequently found in PKS-NRPS hybrids.5,25 A common feature
of Bet1 and PKS-NRPSs is that both PKSs produce a linear
polyketide chain having a b-ketoacyl group. The exceptional
case was reported for the LovB/C system (Table S2, ESI†), which
produces polyketide with a b-hydroxy group.23 Previous structural
analysis of LovC provided sufficient data to propose the catalytic
mechanism.26 A phylogenetic analysis including 8 trans-ERs
found in HR-PKS–PKS-NRPS systems and 22 ER domains
(cis-ER) of HR-PKS–PKS-NRPSs suggested that ERs could be
divided into 3 groups (Fig. S7, ESI†). trans-ERs, including Bet1
and LovC, form Clade I, while functionally active and inactive
cis-ERs are classified into Clade II and Clade III, respectively.
This classification is likely general because the uncharacterized
HR-PKS system, comprising FSL1 (HR-PKS) and FSL5 (trans-ER)
involved in fusarielin biosynthesis,27 belongs to the same group
as Bet1/3 and LovB/C. Multiple sequence alignment revealed
a new ‘‘fingerprint’’ region to distinguish functionally active
and inactive cis-ERs (Fig. S8, ESI†). In addition, a previously
proposed point mutation in the NADPH-binding motif of
cis-ERs28 was also found in the case of Acrts2, which is
classified into Clade II. The inactivation of the ER domain
corresponds to the polyketide structure of ACR-toxin.29 Notably,
sequence and phylogenetic analyses of ERs revealed that
Bet1 and LovB, having a C-terminal reductase domain and a
condensation domain, respectively, are classified into the same
clade with PKS-NRPS, suggesting an evolution of ancestral
PKS-NRPS towards Bet1 and LovB by partial deletion of char-
acteristic domains for NRPS.
15 K. Tagami, A. Minami, R. Fujii, C. Liu, M. Tanaka, K. Gomi, T. Dairi
and H. Oikawa, ChemBioChem, 2014, 15, 2076.
16 Y. Matsuda, T. Wakimoto, T. Mori, T. Awakawa and I. Abe, J. Am.
Chem. Soc., 2014, 136, 15326.
In summary, we have identified and characterized the
betaenone biosynthetic gene cluster. Through heterologous
expression studies, we have established that Bet1 and Bet3
are key enzymes for skeletal construction in betaenone bio-
synthesis. To our knowledge, this is the first functional char-
acterization of a fungal HR-PKS harboring an R domain, which
catalyzes the reductive release of the polyketide chain. Based on
the chemical reactivity and conformational search of dehydro-
probetaenone I, we proposed the biosynthetic hypothesis that
reductive cleavage of the linear polyketide chain followed
by a Diels–Alder reaction gives the trans-decalin skeleton.
Co-expression of bet4 and bet2 with bet1/3 in A. oryzae also
revealed subsequent reductive and oxidative modifications in
betaenone biosynthesis.
17 S. Miki, Y. Sato, H. Tabuchi, H. Oikawa, A. Ichihara and
S. Sakamura, J. Chem. Soc., Perkin Trans. 1, 1990, 1228.
18 K. Kasahara, T. Miyamoto, T. Fujimoto, H. Oguri, T. Tokiwano,
H. Oikawa, T. Ebizuka and I. Fujii, ChemBioChem, 2010, 11, 1245.
19 I. Fujii, N. Yoshida, S. Shimomaki, H. Oikawa and Y. Ebizuka, Chem.
Biol., 2005, 12, 1301.
20 W. Xu, Y. H. Chooi, J. W. Choi, S. Li, J. C. Vederas, N. A. Da Silva and
Y. Tang, Angew. Chem., Int. Ed., 2013, 52, 6472.
21 X. Xie, M. J. Meehan, W. Xu, P. C. Dorrestein and Y. Tang, J. Am.
Chem. Soc., 2009, 131, 8388.
22 B. Manavalan, S. K. Murugapiran, G. Lee and S. Choi, BMC Struct.
Biol., 2010, 10, 1.
23 S. M. Ma, J. W. H. Li, J. W. Choi, H. Zhou, K. K. M. Lee,
V. A. Moorthie, X. Xie, J. T. Kealey, N. A. Da Shilva, J. C. Vederas
and Y. Tang, Science, 2009, 326, 589.
24 (a) H. Oikawa, Y. Suzuki, A. Naya, K. Katayama and A. Ichihara,
J. Am. Chem. Soc., 1994, 116, 3605; (b) H. Oikawa, K. Katayama,
Y. Suzuki and A. Ichihara, J. Chem. Soc., Chem. Commun., 1995, 908.
25 D. Boettger and C. Hertweck, ChemBioChem, 2013, 14, 28.
26 B. D. Ames, C. Nguyen, J. Bruegger, P. Smith, W. Xu, S. Ma, E. Wong,
S. Wong, X. Xie, J. W.-H. Li, J. C. Vederas, Y. Tang and S. Tsai, Proc.
Natl. Acad. Sci. U. S. A., 2012, 109, 11144.
27 J. L. Sorensen, F. T. Hansen, T. E. Sondergaard, D. Staerk,
L. G. Klitgaard, S. Purup, H. Giese and R. J. N. Frandsen, Environ.
Microbiol., 2012, 14, 1159.
We are grateful to Prof. Masanori Arita of the National
Institute of Genetics for meaningful discussions about the
phylogenetic analysis of the KS domain of PKS. This work
was financially supported by a MEXT research grant on innova-
tion area 22108002 to H. Oikawa.
28 K. Kasahara, I. Fujii, H. Oikawa and Y. Ebizuka, ChemBioChem,
2006, 7, 920.
29 Y. Izumi, K. Ohtani, Y. Miyamoto, A. Masunaka, T. Fukumoto,
K. Gomi, Y. Tada, K. Ichimura, T. L. Peever and K. Akimitsu, Mol.
Plant-Microbe Interact., 2012, 25, 1419.
Notes and references
1 (a) Handbook of Secondary Fungal Metabolites, 3-Volume Set, ed.
R. J. Cole, B. B. Jarvis and M. A. Schweikert, Elsevier, 2003;
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2014