Cyclization mechanism of amorpha-4,11-diene synthase, a key enzyme in artemisinin biosynthesis

Article abstract of DOI:10.1021/np050356u

Cyclization of farnesyl diphosphate into amorpha-4,11-diene by amorpha-4,11-diene synthase (ADS) initiates biosynthesis of artemisinin, a clinically important antimalarial drug precursor. Three possible ring-closure mechanisms, two involving a bisabolyl carbocation intermediate followed by either a 1,3-hydride shift or two successive 1,2-shifts, and one involving a germacrenyl carbocation, were proposed and tested by analyzing the fate of farnesyl diphosphate H-1 hydrogen atoms through ¹H and ²H NMR spectroscopy. Migration of one deuterium atom of [1,1-²H ₂]farnesyl diphosphate to H-10 of amorpha-4,11-diene singled out the bisabolyl carbocation mechanism with a 1,3-hydride shift. Further confirmation was obtained through enzyme reactions with (1R)- and (1S)-[1- ²H]farnesyl diphosphate. Results showed that deuterium of the 1R compound remained at H-6, whereas that of the 1S compound migrated to H-10 of amorpha-4,11-diene. Incorporation of one deuterium into amorphadiene in the cyclization process was observed when the reaction was performed in ²H₂O, as evidenced by an increase of 1 amu in the mass of the molecular ion.

Full text of DOI:10.1021/np050356u

758
J. Nat. Prod. 2006, 69, 758-762
Cyclization Mechanism of Amorpha-4,11-diene Synthase, a Key Enzyme in Artemisinin
Biosynthesis
Soon-Hee Kim,† Keon Heo,‡ Yung-Jin Chang,† Si-Hyung Park,§ Sang-Ki Rhee,† and Soo-Un Kim*,†,
School of Agricultural Biotechnology, Seoul National UniVersity, Seoul 151-921, Korea, Research and DeVelopment Center, Doosan
Cooperation, Yongin 449-795, Korea, Department of Pharmacognostic Resources, Mokpo National UniVersity, Muan 534-729, Korea, and
Plant Metabolism Research Center, Kyung Hee UniVersity, Yongin 449-701, Korea
ReceiVed September 22, 2005
Cyclization of farnesyl diphosphate into amorpha-4,11-diene by amorpha-4,11-diene synthase (ADS) initiates biosynthesis
of artemisinin, a clinically important antimalarial drug precursor. Three possible ring-closure mechanisms, two involving
a bisabolyl carbocation intermediate followed by either a 1,3-hydride shift or two successive 1,2-shifts, and one involving
a germacrenyl carbocation, were proposed and tested by analyzing the fate of farnesyl diphosphate H-1 hydrogen atoms
1
2
through H and H NMR spectroscopy. Migration of one deuterium atom of [1,1-2H2]farnesyl diphosphate to H-10 of
amorpha-4,11-diene singled out the bisabolyl carbocation mechanism with a 1,3-hydride shift. Further confirmation
was obtained through enzyme reactions with (1R)- and (1S)-[1-2H]farnesyl diphosphate. Results showed that deuterium
of the 1R compound remained at H-6, whereas that of the 1S compound migrated to H-10 of amorpha-4,11-diene.
Incorporation of one deuterium into amorphadiene in the cyclization process was observed when the reaction was
2
performed in H2O, as evidenced by an increase of 1 amu in the mass of the molecular ion.
The importance of terpenoids in human health care can never
be overemphasized, as evidenced by terpenoid-derived drugs such
as paclitaxel. The role of terpenoids in chemical communication
between species is also an important factor in ecology. In this
regard, the chemistry and biology of terpene biosynthesis have been
receiving much scientific and biotechnological attention. The recent
discovery of the 2-C-methyl-D-erythritol-4-phospahte (MEP) path-
way in bacteria1,2 and plants3 offers new opportunities to understand
and manipulate the biosynthesis of these vital molecules. Cloning
of terpene cyclases from various sources, on the other hand, enabled
us to gain deeper insight into the molecular action and biology of
terpene biosynthesis.4 In particular, the availability of large amounts
of the cyclase produced by heterologous expression in E. coli
offered an opportunity to directly probe the action mechanism of
the enzyme. Long before the unraveling of the three-dimensional
structures of terpene cyclases, various methods had been employed
to probe the cyclization process. For example, the use of deuterium-
labeled substrates enabled researchers to study the mechanism by
observing the labeling pattern of the enzymatic product.5-7 Labeled
substrate was also used to help identify the intermediate carbocation
during cyclization. Recently, elucidation of the three-dimensional
structure of epi-aristolochene synthase from tobacco (TEAS) opened
a new avenue into explaining the enzyme action in a more detailed
manner with respect to the putative mechanism.8
Amorpha-4,11-diene synthase (ADS) of Artemisia annua L. is
a sesquiterpene cyclase that catalyzes the conversion of farnesyl
diphosphate into amorpha-4,11-diene in the biosynthesis of arte-
misinin, a precursor of the clinically important antimalarial drug
artesunate.9 There has been intense interest in the unraveling of
the biosynthetic pathway leading to artemisinin and cloning of the
involved enzymes. The knowledge gained through these studies
could be directly applied to engineering of the plant toward higher
artemisinin production. A cDNA coding ADS, the only enzyme so
far identified to operate in the biosynthesis of artemisinin, was
cloned from A. annua and functionally identified through expression
in E. coli.10 The enzyme produces an amorphane-type ring, whereas
δ-cadinene synthase (CDS) in cotton11 makes the same carbon
skeleton but with a configuration and double bond geometry
different from those of amorpha-4,11-diene. Therefore, unraveling
the action mechanism of ADS would enable us to understand the
requirements involved in the bifurcation of the catalytic pathway
into the δ-cadinane or amorphadiene skeleton.
In the biosynthesis of amorpha-4,11-diene by ADS, two mech-
anisms involving a bisabolyl cation from the initial 1,6-closure
(pathways a and b, Figure 1) and one involving a germacryl cation
from the initial 1,10-closure (pathway c, Figure 1) are possible.
Among the bicyclic sesquiterpene synthases, CDS,11 TEAS,12 and
pentalenene synthase13 are known to produce a germacryl cation
as the initial cyclic intermediate. This prevalence of the germacryl
intermediate in the cyclase catalysis prompted us to consider the
viability of pathway c as an ADS action mechanism. However,
Picaud et al. proposed that ADS catalysis proceeds through the
bisabolyl intermediate (pathway a), a conclusion based on the
appearance of the reaction byproducts with a bisabolane skeleton.14
Alternatively, if one considers successive 1,2-hydride shifts instead
of one 1,3-hydride shift before the second annulation, a variant
bisabolyl mechanism (pathway b, Figure 1) would be possible.
Recently during manuscript preparation, Brodelius’ group arrived
at the conclusion that ADS catalyzes initial 1,6-ring closure followed
by a stereospecific 1,3-hydride shift based on the indirect observa-
tion of 2H by 1H NMR in amorpha-4,11-diene derived from
deuterium-labeled FPP.15
The present research was performed to establish the mechanism
of the ADS reaction by determining the migration of farnesyl
diphosphate H-1 atoms, in a different manner from Brodelius’
group, in the cyclization process. In this work, the deuterium-labeled
2
product amorpha-4,11-diene was analyzed by H NMR to locate
the position of the label. On the basis of these observations, we
confirmed the operation of a bisabolyl mechanism involving one
1,3-hydride shift (pathway a), and a mechanism explaining the
observed deuterium exchange with the medium was postulated.
Results and Discussion
* To whom correspondence should be addressed. Tel: 82-2-880-4642.
E-mail: soounkim@plaza.snu.ac.kr.
Protein Purification. The recombinant ADS protein tagged with
His6 was purified through a single-step nickel column chromatog-
raphy. The eluent was monitored by SDS-PAGE (Supporting
Information), and the purified enzyme with a MW of 61 was used
† School of Agricultural Biotechnology, SNU.
‡ Doosan Cooperation.
§ Department of Pharmacognostic Resources, MNU.
Plant Metabolism Research Center, KHU.
10.1021/np050356u CCC: $33.50
© 2006 American Chemical Society and American Society of Pharmacognosy
Published on Web 05/26/2006

Cyclization Mechanism of Amorpha-4,11-diene Synthase
Journal of Natural Products, 2006, Vol. 69, No. 5 759
Figure 1. Three possible mechanisms of ADS action. Pathways a and b involve one 1,3- and two 1,2-hydride shifts, respectively, after the
initial formation of a bisabolyl cation intermediate through 1,6-ring closure. Pathway c is characterized by initial formation of a germacryl
carbocation intermediate arising from 1,10-ring closure.
in the reaction with farnesyl diphosphate. The purity of the enzyme
used in the preparative reaction was estimated to be >95% by SDS-
PAGE.
Synthesis of Deuterium-Labeled Farnesyl Diphospahtes.
Asymmetric labeling at H-1 of farnesyl diphosphate was achieved
through the reduction of [1-2H]farnesal with (R)- or (S)-alpine
borane.17 To avoid ambiguity in the interpretation of amorphadiene
spectra, deuterium labeling of farnesyl diphosphate with high
enantiomeric purity was required at H-1. The reported enantiopurity
of [1-3H]isopentenyl diphosphate similarly prepared is 98% ee.18
The presence of the diphosphate group in the final products was
confirmed by 31P NMR (data not shown), and the resulting farnesyl
diphosphates were used as ammonium salts.
Position of Deuterium Label. Incubations were performed under
the established 1 mL assay condition to ensure high conversion of
farnesyl diphosphate into amorphadiene. Enzymatic cyclization with
the unlabeled substrate afforded amorphadiene (>95%) and some
byproducts, as shown by gas chromatography analysis (Supporting
Information).
Inspection of the mass fragmentation patterns (Supporting
Information) revealed the presence of deuterium in all labeled
amorpha-4,11-diene samples. Labeled amorphadiene obtained
through the incubation of singly and doubly labeled farnesyl
Figure 2. 2H NMR spectra of amorphadienes: (A) from [1,1-2H2]-
farnesyl diphosphate; (B) from (1R)-[1-2H]farnesyl diphosphate;
(C) from (1S)-[1-2H]farnesyl diphosphate.
diphosphates yielded molecular ions at m/z 205 and 206, respec-
tively. No significant differences were observed in the mass spectra
of H-6 and H-10 labeled amorphadienes. The deuterium position
was elucidated by 2H NMR (Figure 2) and also by 1H NMR (Figure
3) as follows. The chemical shift value of deuterium is, in general,
identical to the corresponding proton chemical shift; therefore, this
chemical shift relationship was used to observe the 2H label during
the cyclization of farnesyl diphosphate by ADS. Due to the low
intrinsic magnetic sensitivity of 2H nuclei and small sample
quantities, attaining a high signal-to-noise ratio value was difficult,
especially in the singly labeled compounds. As a compromise
between the magnetic field stability and measuring time, a signal-
to-noise ratio of 1.8, which was just adequate for positively judging
of the signals were comparable, indicating the presence of one
deuterium at each site.
To ascertain which deuterium of [1,1-2H2]farnesyl diphosphate
migrated to H-10 of amorphadiene, [1-2H]farnesyl diphosphates
asymmetrically labeled at H-1 were prepared. Incubation of (1R)-
[1-2H]farnesyl diphosphate with ADS yielded a singly labeled
2
product, whose H NMR signal appeared at δ 2.60 (Figure 2B).
This sample had small signals at δ 1.38 and 1.25, the former
possibly due to the incomplete induction of asymmetry on C-1 of
the labeled farnesyl diphosphate sample and the latter due to
impurity. The 1S isomer gave rise to amorphadiene with a signal
at δ 1.38. Again, the sample had an impurity signal at δ 1.25 (Figure
2C).
2
the presence of H, even in the least concentrated sample, was
adopted (Figure 2C). The amorphadiene derived from [1,1-2H2]-
farnesyl diphosphate displayed signals at δ 1.38 and 2.60 (Figure
2A), corresponding to the proton signals of H-10 and H-6 at δ 1.40
and 2.55, respectively. Broadening of the peak, unavoidable due
The position of the deuterium label in amorphadiene was
2
1
1
to the quadrupolar nature of the H nucleus and field drift during
confirmed by H NMR measurement. The H spectrum of amor-
phadiene from doubly labeled farnesyl diphosphate lacked signals
at both δ 1.40 and 2.55 as expected (Figure 3B), confirming the
the long accumulation period, resulted in chemical shifts slightly
different from those observed in the 1H NMR spectrum. Intensities

760 Journal of Natural Products, 2006, Vol. 69, No. 5
Kim et al.
result in amorphadiene with deuterium at C-6 and C-10. Stereo-
chemical considerations also favor pathway a over b, because only
the axial hydrogen atom is allowed to migrate in the process of
suprafacial hydride shift; therefore, the ring conformation of
intermediate 3 would dictate which hydrogen is to migrate. If the
hydrogen atom Ha were in an axial orientation, allowing 1,3-hydride
shift in the intermediate 4, the side chain would be at an equatorial
position (Figure 4). Model building demonstrated that this equatorial
disposition of the side chain enabled the intermediate to fold into
the appropriate conformation needed for the second ring closure.
On the other hand, if Hb were to migrate via 1,2-hydride shift, the
conformation of the intermediate 3a would be such that Hb and
the side chain point in the opposite axial directions. Molecular
modeling showed that the axial conformation of the side chain in
intermediate 4 is not favorable for the completion of the second
ring closure by p-orbital overlap. Therefore, the final labeling pattern
at H-1 and H-10 of amorphadiene derived from [1,1-2H2]farnesyl
diphosphate enables differentiation between pathways a and b, with
pathway a additionally supported by the conformational analysis
of the putative reaction intermediates. Indeed, one of the labels
was found at H-6 and another at H-10 in support of pathway a.
The above conclusion was supported through cryptochemistry
of pro-R (Hb) and pro-S (Ha), H-1 hydrogens that gave direct
support to the detailed cyclization mechanism (Figure 4). This work
reconfirmed the stereochemical course of the reaction presented
by Picaud et al., who also employed the chirally deuterated farnesyl
diphosphate to probe the mechanism.15 In the cyclization of farnesyl
diphosphate, isomerization of farnesyl diphosphate into nerolidyl
diphosphate (2, NDP) is a prerequisite for correct 1,6-ring closure.
An alternative to the isomerization is the rotation of the C-2-C-3
bond after the formation of an allylic farnesyl cation-pyrophosphate
anion ion pair, giving conformation 2 with Hb (pro-R H-1HR of 1)
cis to the main chain. Attack of C-1 at the si-face of C-6 positions
Ha-1 at the axial direction and the side chain in the desirable
equatorial position of 3. This axial disposition of Ha-1 enables
suprafacial transfer of the hydrogen to the empty p-orbital at C-7
of 3. The migration results in the R configuration at C-7 on 5, which
is the correct C-7 configuration of amorphadiene 6. The migration
of a hydrogen atom introduces the positive charge at C-1 with the
p-orbital in the axial direction (4). The final ring closure between
C-1 and C-10 in structure 4 leads to the desirable cis-decalin
configuration in compound 6, amorpha-4,11-diene.
Figure 3. 1H NMR spectra of amorphadienes: (A) authentic
amorpha-4,11-diene; (B) from [1,1-2H2]farnesyl diphosphate; (C)
from (1S)-[1-2H]farnesyl diphosphate; (D) from (1R)- [1-2H]farnesyl
diphosphate.
labeling at H-6 and H-10. Collapse of the H-15 signal at δ 0.88
2
into a singlet further supported the labeling of H-10 with H. In
amorphadiene derived from (1S)-[1-2H]farnesyl diphosphate, the
characteristic H-10 signal at δ 1.40 disappeared (Figure 3C). On
1
the other hand, the H NMR spectrum of amorphadiene derived
from (1R)-[1-2H]farnesyl diphosphate was devoid of a H-6 signal
at δ 2.55 (Figure 3D). These results unambiguously indicated that
HS-1 (Ha-1) of farnesyl diphosphate migrated to H-10 of amorpha-
diene, while HR-1 (Hb-1) remained at its position to label amor-
phadiene H-6.
Deduction of Cyclization Mechanism. Brodelius’ group pro-
posed a cyclization mechanism of ADS based on the structure of
ADS reaction byproducts; formation of compounds with a bisabo-
lane skeleton such as R-bisabolol, â-sesquiphellandrene, and
zingiberene as minor byproducts led to the proposal that the 1,6-
closure involving a bisabolyl carbocation intermediate precedes the
1,10-closure (pathway a, Figure 1).14,19 The group recently con-
cluded that 1,6-ring closure is the first event and 1,3-hydride shift
of the original HS-1 of farnesyl diphosphate is operating in the ADS
action,15 based on the experiment involving farnesyl diphosphate
asymmetrically labeled at C-1 with deuterium. Many sesquiterpene
cyclases such as CDS from cotton are known to operate through
the germacryl mechanism (pathway c, Figure 1).11
Deuterium Exchange in 2H2O. Amorphadiene derived from the
incubation of farnesyl diphosphate and ADS in deuterium oxide
showed a molecular ion at m/z 205, which indicated the incorpora-
tion of one deuterium atom from the reaction medium (Supporting
Information). The labeling pattern observed in the mass spectrum
was indistinguishable from those of amorphadienes derived from
singly labeled farnesyl diphosphates; therefore, the position of the
label could not be determined through mass spectrometric analysis.
The activity of ADS dramatically decreased when changing the
To establish the correct ADS mechanism from among the three
possibilities (Figure 1), migration of farnesyl diphosphate H-1
during cyclization was traced using deuterium-labeled farnesyl
diphosphates as substrate. The site of deuteration in the resulting
2
medium from H2O to H2O. Therefore, a large-scale preparation
2
2
labeled amorphadienes could be directly observed by H NMR
of amorphadiene in H2O for NMR analysis was not practical.
spectra of high signal-to-noise ratio (Figure 2). Even with only the
labeling pattern of amorphadiene from [1,1-2H2]farnesyl diphosphate
at hand, differentiation of the alternative mechanisms became
possible. Retention of two deuterium atoms at C-1 of amorphadiene
6 easily eliminated pathway c from the possibilities, because, with
pathway c in operation, one of the labels at farnesyl diphophate
H-1 would migrate to amorphadiene C-11 through 1,3-shift (7,
Figure 1) before being eliminated in the process of double-bond
formation at the isopropenyl side chain.
Bisabolyl carbocation intermediate 3 (Figure 1, pathway a)
arising from 1,6-closure would undergo hydride shift through either
one direct suprafacial 1,3-shift of axial Ha-1 to C-7 (4, pathway a,
Figure 1) or two suprafacial 1,2-hydride shifts, axial Hb-1 to C-6
and H-6 to C-7 (4a, pathway b), resulting in the correct cis-decalin
configuration at C-1 and C-6 of 6. Hence, only pathway a would
However, because amorphadiene obtained from deuterated
farnesyl diphosphate retains the deuterium label at H-6 and H-10
in H2O, the position of deuterium exchange would not be H-6 or
2
H-10. One of the possible mechanisms, though unprecedented,
compatible with deuterium incorporation from the reaction medium
is presented in Figure 4. In this mechanism, the intermediate 4 loses
allylic H-6 to generate 2,6,10-bisabolatriene 8, which would pick
up one deuterium from the 2H2O medium on return. A large kinetic
isotope effect is expected in the process. Indeed, deuterium
incorporation was not observed in the medium even with up to
2
75% H2O.
Experimental Section
General Experimental Procedures. Protein concentration was
determined using the Bio-Rad protein assay kit according to the

Cyclization Mechanism of Amorpha-4,11-diene Synthase
Journal of Natural Products, 2006, Vol. 69, No. 5 761
Figure 4. Cyclization mechanism of ADS.
manufacturer’s instruction (Bulletin LIT33 REV C, Bio-Rad Labora-
tories) with bovine serum albumin (Sigma) as a standard. The
electrophoresis with the standard molecular marker (Bio-Rad) was
carried out through Laemmli’s method using 10% polyacrylamide gel
on a vertical slab gel electrophoresis apparatus (Hoefer). Electrophoresis
run started at 75 V, and when the samples completely entered the
stacking gel, the voltage was raised to 150 V. The gels were stained
with 0.025% (w/v) Coomassie Brilliant Blue R-250 (Bio-Rad Labo-
ratories) and destained with 10% MeOH and 10% HOAc solution. 1H
mL of LB medium containing ampicillin in a flask, and the cells were
grown at 37 °C with continuous shaking until the optical density at
600 nm reached 0.5. At this point, isopropyl-â-D-thiogalactopyranoside
(IPTG) was added to give a final concentration of 0.4 mM, and the
culture was grown at 20 °C for an additional 6 h. The induced bacteria
were harvested by centrifugation at 7000g on a Beckman Avanti 30
centrifuge at 4 °C for 10 min. The cell pellets were then frozen in
liquid N2 and stored at -70 °C. Prior to purification, bacterial pellets
from 10 L culture were thawed and resuspended in 200 mL of lysis
buffer (100 mM Tris-HCl, pH 8.0, 10% glycerol, 1 mM PMSF, 5 mM
MgCl2), and the suspension was incubated on ice for 30 min. The lysate
was further homogenized by sonication with a 20 s pulse (VC-600,
Fisher), and the homogenate was centrifuged at 10000g at 4 °C for 20
min. The resulting supernatant was applied to Ni-NTA resin (Invitrogen)
that had been pre-equilibrated with the lysis buffer. The unbound
proteins were eluted from the column using 150 mM imidazol in 10
bed volumes of the washing buffer (50 mM NaH2PO4, pH 7.0, 300
mM NaCl). All purification procedures were carried out at 4 °C. Protein
content and SDS-PAGE pattern of the eluent were monitored to identify
ADS fractions.
2
(400 MHz), 31P (121 MHz), and H (61.25 MHz) NMR spectra were
2
recorded on a JEOL LA400 (JEOL, Japan). For measurement of H
NMR, CDCl3 in CHCl3 was used as a chemical shift standard. Before
processing the data, 10 000 transients were accumulated to ensure proper
signal-to-noise ratio. GC-MS analyses were performed on a Hewlett-
Packard 5890 series II (Palo Alto, CA) using a DB-5 column (30 m ×
0.25 µm × 0.25 mm) equipped with a JEOL AX505WA mass
spectrometer (JEOL, Japan) operating in the electron impact mode at
70 eV. Separation conditions were as follows: injection port 200 °C,
flame ionization detector 250 °C, split ratio 5:1, and helium flow 1.0
mL min-1. Temperature program for the enzyme reaction product was
as follows: 100 °C for 2 min, increase to 150 °C (5 °C min-1),
isothermal for 2 min, further increase to 250 °C (10 °C min-1), and
final isothermal for 2 min at 250 °C.
Synthesis of Deuterium-Labeled Farnesyl Diphosphates. The [1,1-
2H2]farnesyl diphosphate, (1R)-[1-2H]farnesyl diphosphate, and (1S)-
[1-2H]farnesyl diphosphate samples were synthesized according to the
modified procedure of Benedict et al.21 In short, farnesol was oxidized
into farnesal by MnO2 and reduced back to farnesol with NaB2H4. [1-2H]-
Farnesol was reoxidized to obtain [1-2H]farnesal. High kinetic isotope
effect ensured almost 100% deuterium retention on H-1. The labeled
farnesal was subsequently reduced with NaB2H4 to give [1,1-2H2]-
farnesol, or with (R)- and (S)-alpine-boranes yielding (1S)- and (1R)-
[1-2H]farnesol, respectively. Farnesol was chlorinated with N-chloro-
succinimide and finally pyrophosphorylated with tris(tetra-n-butyl)
ammonium hydrogen pyrophosphate. The resulting farnesyl diphos-
phates were isolated by cellulose column chromatography and stored
at -70 °C as ammonium salts before use. The synthesis of farnesyl
diphosphates was verified through 1H and 31P NMR (data not shown).
Construction of Expression Plasmid. E. coli BL21(DE3)pLysS
[hsdS, gal (λclts857, ind1, sam7, nin5, lacUV5-T7 gene1)] harboring
KCS12 was used as an ADS source.10 To isolate the ADS insert for
cloning into pRSET A vector (Invitrogen), the plasmid was digested
with BamHI and EcoRI, and the digests were purified using the PCR
purification kit (Takara). BamHI and EcoRI-digested pRSET A were
subjected to overnight ligation in the presence of T4 DNA ligase
(Promega). The ligated samples, which contained a nucleotide sequence
coding for six histidines placed upstream in frame with the pET-5a
ADS insert, were transformed into BL21(DE3) cells. A single colony
obtained on Luria-Bertani (LB)-ampicillin plates was selected and
grown for use in the glycerol stocks, and the sequence of the plasmid
was verified.
Amorpha-4,11-diene Production by ADS. ADS was used as tagged
with histidine, because a preliminary experiment showed that the
presence of the tag had little effect on the enzyme activity. The reaction
mixture for amorpha-4,11-diene production contained 100 mg of
purified ADS, 4 mM MgCl2, 30 mM HEPES, pH 8.0, 3 mM
â-mercaptoethanol, 4 mM DTT, and 100 µM farnesyl diphosphate in
a total volume of 1 mL. Each reaction mixture was then covered with
1 mL of hexane, sealed with Parafilm, and incubated at 25 °C for 6 h.
One hundred reactions were carried out for each farnesyl diphosphate.
Subsequently, the reaction mixture was extracted three times with 100
mL of hexane. The combined hexane extract was passed through a
silica gel column (1.5 cm × 5 mm i.d.) covered with anhydrous MgSO4
(1 cm) to yield the terpene hydrocarbon fraction. The hexane extract
was concentrated under nitrogen stream before GC-MS and NMR
analyses. For the enzyme reaction in 2H2O, the reaction medium without
2
the enzyme was lyophilized and suspended in H2O. The process was
repeated to ensure complete exchange with deuterium. The enzyme
was separately freeze-dried and resuspended in the reaction buffer
before the initiation of the reaction.
Authentic Amorpha-4,11-diene. Amorpha-4,11-diene was synthe-
sized through the reduction of artemisinic acid via artemisinol,10 and
the structure was confirmed through MS and NMR analyses for
unambiguous assignment of NMR signals and found consistent with
published data.10,15
Acknowledgment. This study was supported by the SRC program
of MOST/KOSEF (R11-2000-081) through the Plant Metabolism
Research Center, Kyung Hee University, and Brain Korea 21 Project
administered by Ministry of Education and Human Resources, Korea.
The measurement of deuterium spectra by Mr. P.-H. Kim, JEOL Korea,
and critical reading of the manuscript by Dr. D. E. Cane are appreciated.
Supporting Information Available: SDS polyacrylamide gel
electropherogram of ADS and mass spectra of the labeled amorpha-
diene. This material is available free of charge via the Internet at http://
pubs.acs.org.
Bacterial Expression and Purification of His6-Tagged ADS.
BL21(DE3) cells, which harbored a plasmid containing the ADS insert,
were grown overnight at 37 °C in 50 mL of LB broth supplemented
with ampicillin (50 µg/mL). The preculture was used to inoculate 500

762 Journal of Natural Products, 2006, Vol. 69, No. 5
Kim et al.
(14) Picaud, S.; Olofsson, L.; Brodelius, M.; Brodelius, P. E. Arch.
Biochem. Biophys. 2005, 436, 215-226.
(15) Picaud, S.; Mercke, P.; He, X.; Sterner, O.; Brodelius, M.; Cane, D.
E.; Brodelius, P. E. Arch. Biochem. Biophys. 2005, published on web.
DOI: 10.1016/j.abb.2005.07.015.
(16) Bohlmann, F.; Gerke, T.; Jakupovic, J.; King, R. M.; Robinson, H.
Phytochemistry 1984, 23, 1183-1184.
(17) Midland, M. M.; Greer, S.; Tramontano, A.; Zderic, S. A. J. Am.
Chem. Soc. 1979, 101, 2352-2355.
(18) Shibuya, M.; Chou, H. M.; Fountoulakis, M.; Hassam, S.; Kim, S.
U.; Kobayashi, K.; Otsuka, H.; Rogalska, E.; Cassady, J. M.; Floss,
H. G. J. Am. Chem. Soc. 1990, 112, 297-304.
(19) Mercke, P.; Bengtsson, M.; Bouwmeester, H. J.; Posthumus,
M. A.; Brodelius, P. E. Arch. Biochem. Biophys. 2000, 381, 173-
180.
(20) Chang, Y. J.; Jin, J.; Nam, H. Y.; Kim, S. U. Biotechnol. Lett. 2005,
27, 285-288.
(21) Benedict, C. R.; Alchanati, I.; Harvey, P. J.; Liu, J.; Stipanovic, R.
D.; Bell, A. A. Phytochemistry 1995, 39, 327-331.
References and Notes
(1) Flesch, G.; Rohmer, M. Eur. J. Biochem. 1988, 175, 405-411.
(2) Broers, S. T. J. Ph.D. Thesis, ETH, Zu¨rich, 1994.
(3) Schwarz, M. K. Ph.D. Thesis, ETH, Zu¨rich, 1994.
(4) Bohlmann, J.; Meyer-Gauen, G.; Croteau, R. Proc. Natl. Acad. Sci.
U.S.A. 1998, 95, 4126-4133.
(5) Cane, D. E.; Pawlak, J. L.; Horak, R. M. Biochemistry 1990, 29,
5476-5490.
(6) Schmidt, C. O.; Bouwmeester, H. J.; Bulow, N.; Konig, W. A. Arch.
Biochem. Biophys. 1999, 364, 167-177.
(7) He, X.; Cane, D. E. J. Am. Chem. Soc. 2004, 126, 2678-2679.
(8) Starks, C. M.; Back, K.; Noel, J. P.; Chappell, J. Science 1997, 277,
1815-1820.
(9) Klayman, D. L. Science 1985, 228, 1049-1055.
(10) Chang, Y. J.; Song, S. H.; Park, S. H.; Kim, S. U. Arch. Biochem.
Biophys. 2000, 383, 178-184.
(11) Benedict, C. R.; Lu, J. L.; Pettigrew, D. W.; Liu, J.; Stipanovic, R.
D.; Williams, H. J. Plant Physiol. 2001, 125, 1754-1765.
(12) Deligeorgopoulou, A.; Allemann, R. K. Biochemistry 2003, 42,
7741-7747.
(13) Cane, D. E.; Watt, R. M. Proc. Natl. Acad. Sci. U.S.A. 2003, 100,
1547-1551.
NP050356U