Sesquarterpenes (C35 terpenes) biosynthesized via the cyclization of a linear C35 isoprenoid by a tetraprenyl-β- curcumene synthase and a tetraprenyl-β-curcumene cyclase: Identification of a new terpene cyclase

Article abstract of DOI:10.1021/ja203779h

In this study, mono- and pentacyclic C₃₅ terpenes from Bacillus subtilis were biosynthesized via the cyclization of C₃₅ isoprenoid using purified enzymes, including the first identified new terpene cyclase that shows no sequence homology to any of the known terpene cyclases. On the basis of these findings, we propose that these C₃₅ terpenes should be called the new family of "sesquarterpenes."

Full text of DOI:10.1021/ja203779h

COMMUNICATION
pubs.acs.org/JACS
Sesquarterpenes (C35 Terpenes) Biosynthesized via the Cyclization
of a Linear C35 Isoprenoid by a Tetraprenyl-β-curcumene Synthase
and a Tetraprenyl-β-curcumene Cyclase: Identification of a New
Terpene Cyclase
Tsutomu Sato,*,† Satoru Yoshida,† Hiroko Hoshino,† Mizuki Tanno,† Mami Nakajima,‡ and
Tsutomu Hoshino†
†Department of Applied Biological Chemistry, Faculty of Agriculture, and Graduate School of Science and Technology,
Niigata University, 8050 Ikarashi-2, Niigata 950-2181, Japan
‡Center for Instrumental Analysis, Niigata University, 8050 Ikarashi-2, Niigata 950-2181, Japan
S Supporting Information
b
and “conserved hypothetical proteins” consisting of more than
ABSTRACT: In this study, mono- and pentacyclic C35
terpenes from Bacillus subtilis were biosynthesized via the
cyclization of C35 isoprenoid using purified enzymes, including
the first identified new terpene cyclase that shows no sequence
homology to any of the known terpene cyclases. On the
basis of these findings, we propose that these C35 terpenes
should be called the new family of “sesquarterpenes.”
180 and 240 amino acids, respectively, were selected as the
candidates because all known terpene cyclases have >240 amino
acids. Second, we analyzed the production of C35 terpenes in
several bacteria for which genomic analyses had been completed.
Four Bacillus species (B. amyloliquefaciens, B. licheniformis,
B. pumilus, and B. thuringiensis) biosynthesized 2 and 3, whereas
Cytophaga hutchinsonii, Flavobacterium johnsoniae, Pseudoaltero-
monas atlantica, and Pseudoalteromonas haloplanktis, which be-
long to the same genus as the marine bacteria that produce 2 and
3,5 unexpectedly did not. The genes found in the genomes of C35
terpene producers but not in those of the nonproducers were
selected as the candidates.
TLC and GCÀMS analyses of the lipid fractions of 49 gene-
disrupted strains revealed that only the ytpB disruptant produced
no detectable amounts of 2 and 3. To demonstrate that the YtpB
protein was TS, the ytpB gene was introduced into the pColdTF
vector, expressed as a soluble form in Esherichia coli, and purified
by nickel-chelating affinity column chromatography (Figure 1).
As a substrate for the TS assay, 1 was biosynthesized by the
incubation of farnesyl diphosphate (FPP) and isopentenyl
diphosphate (IPP) with HepS and HepT, which were purified
by a procedure similar to that for YtpB (Figure 2A). The purified
YtpB successfully converted 1 to 2 (Figure 2C). The ratio of 3 to
2 detected in the GCÀMS chromatogram of the reaction
products synthesized by the incubation of 1 with YtpB was
almost the same as that in the chromatogram for pure 2 isolated
from B. subtilis, suggesting that 3 could be formed by autooxida-
tion during the measurements (Figure 2B,C). The results from
the mutant strain and the in vitro enzymatic reaction indicate that
ytpB encodes TS. These results are the first to demonstrate at
both the gene and enzyme levels that C35 terpenes are bio-
synthesized via the cyclization of the linear C35 isoprenoid.
TS homologues exist as proteins of unknown function in
various bacterial species in addition to the Bacillus ones. Se-
quence alignments of TS with its homologues indicated the
presence of conserved amino acids. However, the active sites of
TS could not be determined because the primary structure of TS
erpenoids make up a very large family of natural products
T
containing more than 50 000 structurally diverse com-
pounds. They are categorized by the number of C5 isoprene
units as follows: hemi- (one C5 unit), mono- (C10; two C5 units),
sesqui- (C15; three C5 units), di- (C20; four C5 units), sester-
(C25; five C5 units), tri- (C30; six C5 units), and tetraterpenes
(C40; eight C5 units).1 In contrast, little is known about C35
terpenes. To the best of our knowledge, only 18 cyclic and 8
linear C35 terpenes have been identified to date,2À8 and no family
name has yet been assigned to this terpene class. Eight of the
compounds in this class, 1À8, have been found in Bacillus subtilis
(Scheme 1).2À4 Compound 1 is known to be synthesized by a
heterodimeric enzyme, heptaprenyl diphosphate synthase (HepS
and HepT) (Scheme 1).2 However, no enzymes have been
identified as producing the other seven compounds 2À8. In this
study, we have unveiled the biosynthetic pathway of C35 terpenes,
which was found to be catalyzed by two unique enzymes. One is a
new terpene cyclase, tetraprenyl-β-curcumene synthase (TS)
(Scheme 1). This enzyme shows no sequence homology with any
of the known terpene cyclases, and it cyclizes an unusually long
prenyl diphosphate chain (C35), whereas geranyl (C10), farnesyl
(C15), and geranylgeranyl (C20) diphosphate are the known
precursors for the diphosphate-triggered type of terpene cyclases.
A number of genes responsible for the biosynthesis of
terpenoids form clusters in microbial genomes.9 Although we
searched for ts among the genes around hepS, hepT, and a
putative tetraprenyl-β-curcumene cyclase (TC) gene (sqhC) in
the B. subtilis genome, we could not find a candidate. Thus, we
investigated 2514 gene-disrupted strains of B. subtilis provided by
the National BioResource Project (NIG, Japan) to isolate ts. The
candidates for ts were narrowed down to 49 strains by the
following two screenings. First, “functional unknown proteins”
Received: May 11, 2011
Published: June 01, 2011
r
2011 American Chemical Society
9734
dx.doi.org/10.1021/ja203779h J. Am. Chem. Soc. 2011, 133, 9734–9737
|

Journal of the American Chemical Society
COMMUNICATION
Scheme 1. Proposed Pathway for the Biosynthesis of C35 Terpenes in B. subtilisa
a The heterodimeric heptaprenyl diphosphate synthase (HepS and HepT), which converts FPP and IPP to 1, was identified by previous studies of
menaquinone biosynthesis.2 The nonenzymatic synthetic pathways 2 f 3 and 4 f 5 f 6À8 were identified by Takigawa et al.4
Figure 1. SDS-PAGE of the overproduced and purified recombinant
enzymes. The deduced molecular weights of HepS, HepT, YtpB, and
SqhC, which were expressed as fusion proteins with a trigger factor
(52 kDa), were 81.1, 91.5, 94.8, and 123.2 kDa, respectively. Lanes 1 and
11, molecular weight marker; lanes 2, 5, 8, and 12, crude extract from
BL21(DE3)/pColdTF-hepS, BL21(DE3)/pColdTF-hepT, BL21(DE3)/
pColdTF-ytpB, and BL21(DE3)/pColdTF-sqhC, respectively; lanes 3,
6, 9, and 13, total soluble proteins from BL21(DE3)/pColdTF-hepS,
BL21(DE3)/pColdTF-hepT, BL21(DE3)/pColdTF-ytpB, and BL21-
(DE3)/pColdTF-sqhC, respectively; lanes 4, 7, 10, and 14, purified
HepS, HepT, YtpB, and SqhC, respectively.
Figure 2. Analyses of products biosynthesized by purified heterodi-
meric heptaprenyl diphosphate synthase (HepS and HepT) and purified
TS (YtpB). (A) TLC analysis of radiolabeled products synthesized
by the incubation of FPP and [1-14C]IPP with HepS and HepT.
(B) GCÀMS analysis of pure 2 isolated from B. subtilis. (C) GCÀMS
analysis of reaction products synthesized by the incubation of 1 with TS
(YtpB). (D) GCÀMS analysis of reaction products synthesized by the
incubation of 1 with heat-denatured TS (YtpB).
has no similarity to that of the known terpene cyclases. Future
biochemical and structural characterizations of TS will elucidate
the catalytic mechanism. Previously, our study of a nonpatho-
genic Mycobacterium species revealed that the cell-free extracts
converted a Z-type C35 isoprenoid to a cyclic compound other
than 2À8.6 However, no TS homologue could be found in the
mycobacterial genome, indicating that at least one more novel
terpene cyclase remains unidentified.
A mutant B. subtilis strain lacking SqhC, which had similarity to
squalene-hopene cyclase,10 did not produce detectable amounts of
9735
dx.doi.org/10.1021/ja203779h |J. Am. Chem. Soc. 2011, 133, 9734–9737

Journal of the American Chemical Society
COMMUNICATION
of spores to reactive oxygen species.3 The ts gene (ytpB) forms
an operon with a lysophospholipase gene (ytpA), which is
responsible for the biosynthesis of the antibiotic bacilysocin in
B. subtilis,11 and many bacterial species with TS homologues
possess no TC homologue. Therefore, the role of 2 and 3 may
not be only to serve as the precursors of 4 and 5. On the other
hand, our previous studies of mycobacteria revealed that unique
enzymes, such as a bifunctional Z-prenyltransferase that prefer-
entially synthesizes C35 or C50 products and a prenyl reductase
that can reduce both E- and Z-prenyl residues, were also
responsible for the sesquarterpene biosynthesis in addition to
the terpene cyclase.6 Thus, biosynthetic studies of the sesquar-
terpenes could be one promising approach for the discovery of
novel enzymes. The studies of sesquarterpenes promise to be
an attractive field for expanding our understanding of the
terpene world.
Figure 3. HPLC analyses of products biosynthesized by purified TC
(SqhC). (A) Reaction products synthesized by the incubation of no
substrate with TC (SqhC). (B) Reaction products synthesized by the
incubation of 2 with TC (SqhC). (C) Reaction products synthesized by
the incubation of 3 with TC (SqhC).
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental details, sequence
b
data of enzymes, MS spectra, and NMR spectra. This material is
available free of charge via the Internet at http://pubs.acs.org.
6À8.3a On the basis of this result, it had been proposed that the
sqhC gene encodes TC.3a However, the enzymatic reaction has
not been clearly demonstrated in vitro. In this study, the sqhC
gene was expressed in E. coli using pColdTF (Figure 1), and the
structures of the enzymatic products 4 and 5, which were isolated
from the reaction mixture of 2 with the cell-free extracts that
included recombinant SqhC, were determined by NMR (1H,
13C, DEPT, COSY, HOHAHA, NOESY, HMQC, and HMBC)
and MS (ESI and EI) analysis. The molecular formula of 4 was
determined to be C35H58O on the basis of HR-ESI-MS. The
chemical shifts of 4 in the 13C NMR spectra assigned by us were
largely different from those by Takigawa et al.4 This may be a
result of mistyping by them, because our spectrum seems to be
same as theirs. In addition, the relative stereochemistry of the
tetracyclic skeleton in 4 was determined for the first time by
observing the NOE correlations (Me29/H5, H5/H9, H9/H14,
H14/H18, and Me32/Me33), as shown in Scheme 1. The
stereochemistry at C21 of 4 is assumed to be same as that of
substrate 2 (Scheme 1). The structure of 5 was also confirmed by
NMR and MS analysis. The purified SqhC (Figure 1) success-
fully converted 2 and 3 to 4 and 5, respectively, as confirmed by
HPLC (Figure 3) and GCÀMS (see the Supporting In-
formation). The amount of 5 formed by autoxidation of 4 was
negligible (Figure 3B), which may be the result of our careful
preparation. In particular, the sample was not evaporated to
complete dryness. Therefore, the cyclization of 2 to 4 was clearly
demonstrated by the in vitro reaction, indicating that the sqhC
gene encodes TC. In addition, the conversion of 3 to 5 by TC
may also have occurred in B. subtilis cells.
’ AUTHOR INFORMATION
Corresponding Author
satot@agr.niigata-u.ac.jp
’ ACKNOWLEDGMENT
This work was supported by a Grant-in-Aid for Young
Scientists (B) from the Japan Society for Promotion of Science
(JSPS) (21780109 to T.S.) and the Agricultural Chemical
Research Foundation (T.S.). We are grateful to Dr. Tanetoshi
Koyama and Dr. Seiji Takahashi of Tohoku University for the
standard samples of polyprenol, Dr. Hidetaka Hori of Niigata
University for B. thuringiensis HD-73, and Dr. Masahiro Fujihashi
of Kyoto University for helpful discussions.
’ REFERENCES
(1) (a) Buckingham, J. Dictionary of Natural Products, online Web
ed.; Chapman & Hall/CRC Press: London, 2002. (b) Christianson,
D. W. Science 2007, 316, 60–61.
(2) Zhang, Y.; Koyama, T.; Marecak, D. M.; Prestwich, G. D.; Maki,
Y.; Ogura, K. Biochemistry 1998, 37, 13411–13420.
(3) (a) Bosak, T.; Losick, R. M.; Pearson, A. Proc. Natl. Acad. Sci.
U.S.A. 2008, 105, 6725–6729. (b) Kontnik, R.; Bosak, T.; Butcher, R.;
Brocks, J. J.; Losick, R.; Clardy, J.; Pearson, A. Org. Lett. 2008, 10,
3551–3554.
(4) Takigawa, H.; Sugiyama, M.; Shibuya, Y. J. Nat. Prod. 2010, 73,
204–207.
In conclusion, the enzymatic biosynthetic pathway 1 f 2 f 4
was clearly shown using purified enzymes. Moreover, the results
indicate that the C35 terpenes 2À8 were biosynthesized via the
cyclization of C35 isoprenoid 1. We propose that these C35
terpenes should be called “sesquarterpenes”. The prefix sesquar
means half (semis in Latin) to four (quartus in Latin) (i.e., three
and a half), just as the prefix sester means half (semis in Latin) to
three (tertius in Latin). The sesquarterpenes may be widely
distributed in bacteria, as the TS homologues exist in various
bacterial species. The physiological functions of 2 and 3 remain
unknown, while it has been proposed that 6À8 (thermally
dehydrated compounds formed from 5) increase the resistance
(5) Boroczky, K.; Laatsch, H.; Wagner-Dobler, I.; Stritzke, K.;
Schulz, S. Chem. Biodiversity 2006, 3, 622–634.
(6) (a) Sato, T.; Kigawa, A.; Takagi, R.; Adachi, T.; Hoshino, T. Org.
Biomol. Chem. 2008, 6, 3788–3794. (b) Sato, T.; Takagi, R.; Orito, Y.;
Ono, E.; Hoshino, T. Biosci. Biotechnol. Biochem. 2010, 74, 147–151.
(c) Sato, T.; Takizawa, K.; Orito, Y.; Kudo, H.; Hoshino, T. ChemBioChem
2010, 11, 1874–1881.
(7) (a) Sporle, J.; Becker, H.; Gupta, M. P.; Veith, M.; Huch, V.
Tetrahedron 1989, 45, 5003–5014. (b) Sporle, J.; Becker, H.; Allen, N. S.;
Gupta, M. P. Phytochemistry 1991, 30, 3043–3047. (c) Hseih, C.; Tseng,
M.; Shao, Y.; Chang, J.; Kuo, C.; Chang, C.; Kuo, Y. J. Nat. Prod. 2006,
69, 1611–1613. (d) Chen, C.; Wu, J.; Yang, N.; Chang, J.; Kuo, C.;
Wang, S.; Kuo, Y. Org. Lett. 2010, 12, 2786–2789.
9736
dx.doi.org/10.1021/ja203779h |J. Am. Chem. Soc. 2011, 133, 9734–9737

Journal of the American Chemical Society
COMMUNICATION
(8) (a) Besra, G. S.; Sievert, T.; Lee, R. E.; Slayden, R. A.; Brennan,
P. J. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 12735–12739. (b) Hoefs,
M. J. L.; Sinninghe Damste, J. S.; de Leeuw, J. W. Org. Geochem. 1995,
23, 263–267. (c) Wolucka, B. A.; Hoffmann, E. Glycobiology 1998,
8, 955–1962. (d) Costantino, V.; Fattorusso, E.; Mangoni, A.; Rosa,
M. D.; Ianaro, A. Tetrahedron 2000, 56, 1393–1395. (e) Manquin, B. P.;
Morgan, J. A.; Ju, J.; Muller-Spath, T.; Clark, D. S. Extremophiles 2004,
8, 13–21.
(9) For examples, see: (a) You, Z.; Omura, S.; Ikeda, H.; Cane, D. E.
J. Am. Chem. Soc. 2006, 128, 6566–6567. (b) Kim, S.; Zhao, P.; Igarashi,
M.; Sawa, R.; Tomita, T.; Nishiyama, M.; Kuzuyama, T. Chem. Biol.
2009, 16, 736–743. (c) Itoh, T.; Tokunaga, K.; Matsuda, Y.; Fujii, I.;
Abe, I.; Ebizuka, Y.; Kushiro, T. Nat. Chem. 2010, 2, 858–864. (d) Ono,
Y.; Minami, A.; Noike, M.; Higuchi, Y.; Toyomasu, T.; Sassa, T.; Kato,
N.; Dairi, T. J. Am. Chem. Soc. 2011, 133, 2548–2555.
(10) Hoshino, T.; Sato, T. Chem. Commun. 2002, 291–301.
(11) Tamehiro, N.; Okamoto-Hosoya, Y.; Okamoto, S.; Ubukata,
M.; Hamada, M.; Naganawa, H.; Ochi, K. Antimicrob. Agents Chemother.
2002, 46, 315–320.
9737
dx.doi.org/10.1021/ja203779h |J. Am. Chem. Soc. 2011, 133, 9734–9737