Enzymatic synthesis of artemisinin from natural and synthetic precursors

Article abstract of DOI:10.1021/np970024s

To investigate the biosynthetic pathway of artemisinin, an assay system for the determination of activity of the enzymes involved in its synthesis has been developed. Results from these experiments have shown that HEPES provides a better buffer system than Tris-HCl. The enzyme(s) requires Mg²⁺ and/or Mn²⁺, and the addition of ATP and NADPH+H⁺ significantly enhances the enzyme activity. A new substrate, dihydroarteannuin B, has been synthesized that can easily be radiolabeled with high specific activity. It is utilized by the enzyme system and is converted to artemisinin with the same efficiency as the natural substrates. This can be conveniently used as a precursor for elucidation of the pathway for artemisinin biosynthesis.

Full text of DOI:10.1021/np970024s

J . Nat. Prod. 1998, 61, 633-636
633
Notes
En zym a tic Syn th esis of Ar tem isin in fr om Na tu r a l a n d Syn th etic P r ecu r sor s
†
Shalini Bharel, Anamika Gulati, M. Z. Abdin, P. S. Srivastava, R. A. Vishwakarma, and S. K. J ain*
Faculty of Science, Hamdard University, New Delhi, India, and National Institute of Immunology, New Delhi, India
Received December 27, 1996
To investigate the biosynthetic pathway of artemisinin, an assay system for the determination
of activity of the enzymes involved in its synthesis has been developed. Results from these
experiments have shown that HEPES provides a better buffer system than Tris-HCl. The
enzyme(s) requires Mg2 and/or Mn , and the addition of ATP and NADPH+H significantly
enhances the enzyme activity. A new substrate, dihydroarteannuin B, has been synthesized
that can easily be radiolabeled with high specific activity. It is utilized by the enzyme system
and is converted to artemisinin with the same efficiency as the natural substrates. This can
be conveniently used as a precursor for elucidation of the pathway for artemisinin biosynthesis.
+
2+
+
Malaria continues to be a major health concern in
tropical countries. Various chemotherapeutic agents
currently in use have become ineffective, as Plasmo-
dium falciparum has developed resistance to these
drugs. It is, therefore, quite imperative to look for new
antimalarial drugs. Artemisinin (Qinghaosu) (1), a
unique sesquiterpene lactone with an endoperoxide
bridge, obtained from the above-ground parts of Arte-
misia annua L. (Asteraceae), has been found to be active
against both, the chloroquine-resistant as well chloro-
quine-sensitive strains of Plasmodium falciparum and
also against cerebral malaria.1
-5
Currently, the primary method of obtaining artemisi-
nin is by extraction of intact plants. However, endog-
enous bioproduction of artemisinin is relatively low
(
ranging from 0.01 to 0.5%), thereby limiting its clinical
6
utilization. To enhance its bioproduction, it is perti-
nent to understand the complete biosynthetic pathway,
which is only partially known as yet. As reported by
7
8
Nair et al., arteannuin B is one of the immediate
precursors of artemisinin. There have been several
reports suggesting that artemisinic acid is converted to
arteannuin B and to artemisinin.9,10 However, none of
these reports provide direct proof whether arteannuin
B is the precursor of artemisinin or not. To study the
biosynthetic pathway of artemisinin, we have chemically
synthesized dihydroarteannuin B by the reduction of
arteannuin B and have used it as an alternate precursor
in our assay system. In this report, we present data to
show that dihydroarteannuin B is enzymatically con-
verted to artemisinin and can be used as a convenient
precursor for the biosynthesis of artemisinin in an in
vitro system.
(
buffer 1) and Tris-HCl (buffer 2), which were most
effective, have been shown. It has been found that
buffer 2 (see the Experimental Section) is a better
system as far as the rate of conversion of arteannuin B
to artemisinin is concerned. However, the preference
for substrates is different between the two buffers. The
substrate specficity of the enzyme in buffer 1 was
dihydroarteannuin B > artemisinic acid > arteannuin
B, whereas in buffer 2 it was on the order of arte-
misinic acid > arteannuin B > dihydroarteannuin B
A number of different buffers have been reported for
the preparation of leaf homogenate. We used a number
of these systems for standardization of the enzyme
assay. The results with two systems, namely HEPES
(Table 1). This is probably because Tris-HCl is an ionic
buffer that might affect the coupling of enzyme and
substrate. On the other hand, HEPES has better
buffering capacity and is less prone to variation in pH
with the change in temperature. The overall rate of
*
To whom correspondence should be addressed. Tel.: 0091-11-
6
984685. Fax: 0091-11-6988874. E-mail: root@hamduni.ren.nic.in.
S0163-3864(97)00024-4 CCC: $15.00
© 1998 American Chemical Society and American Society of Pharmacognosy
Published on Web 04/17/1998

6
34 J ournal of Natural Products, 1998, Vol. 61, No. 5
Notes
Ta ble 1. Effect of Various Buffers and Substrates on in Vitro
Biosynthesis of Artemisinin
derivatives of arteannuin B that could act as suitable
precursors for the biosynthetic pathway studies. We
have found that the reduction of arteannuin B at C-13
with tritiated sodium borohydrate (NaBT4) formed its
tritiated dihydro derivative with high specific activity.
The compound maintains the structural integrity of
arteannuin B and is used by the enzyme(s) as substrate
for its efficient conversion to artemisinin. Radiolabeled
artemisinic
acid
arteannuin
B
dihydroarteannuin
B
substrate
Buffer
Tris
HEPES
Artemisinin biosynthesized (µmol/mg of protein)
0.003
0.017
0.002
0.014
0.010
0.011
a
Leaves of A. annua (80-100 days old) were homogenized in
buffer 1 or 2 (see Experimental Section for details) and incubated
with various substrates for 150 min. The amount of artemisinin
produced was estimated as described in the text. The control
values representing endogenous levels were deducted, and the
amount of artemisinin produced has been expressed as µmol/mg
protein.
5
dihydroarteannuin B (3.2 × 10 cpm, specific activity
9
-
1.29 × 10 cpm/mmol) was used as the substrate, and
its conversion to artemisinin was followed. The arte-
misinin purified through preparative TLC gave 1.0 ×
4
1
0 cpm/mg of protein. The authenticity of the radio-
labeled artemisinin obtained was established by cocrys-
tallizing it with cold artemisinin as carrier. Though a
number of phytochemical studies have been done on A.
annua, dihydroarteannuin B has not been isolated as a
naturally occurring chemical constituent of the plants.
It suggests that probably dihydroarteannuin B does not
exist in vivo. Alternatively, it might be getting con-
verted to artemisinin very fast.
As shown in Table 1, the rate of conversion of
dihydroarteannuin B to artemisinin is high. It can
therefore, serve as an alternate substrate for assaying
the enzyme activity in an in vitro system. It can also
serve as a convenient radiolabeled precursor for study
of the metabolic pathway of artemisinin and its break-
down to the products such as artemisitene.
To purify the enzyme(s), the leaf homogenate was
separated into different fractions by ammonium sulfate
precipitation, and each fraction was assayed for enzyme
activity. The majority of the proteins were precipitated
in the 0-25% fraction, which included most of the
cellular proteins and the pigments. The highest enzyme
activity, however, was obtained in the 75-100% (NH4)2-
SO4 fraction. This fraction had 3.1% of the total leaf
proteins but ∼25% of the total enzyme activity. There
was, thus, an 8-fold enrichment of the enzyme activity
conversion for all three substrates was higher in HEPES
buffer than in Tris-HCl buffer. For conversion of
dihydroarteannuin B to artemisinin, both buffer 1 and
buffer 2 were equally effective.
The time course studies show that conversion of
artemisinic acid or arteannuin B to artemisinin is at a
maximum after 150 min of incubation (Figure 1). All
the enzyme assays were therefore, carried out for 150
min at 30 °C. The decline in the level of artemisinin at
longer incubations may probably be due to its chemical
instability in the aqueous environment of the cell free
system. This is in concordance with the findings of
1
1
Kudakessril et al., who reported that artemisinin is
relatively unstable in aqueous environment. The pos-
sibility that upon longer incubations the artemisinin is
getting converted to its metabolites, however, cannot be
ruled out.
The highest enzyme activity was found during the full
vegetative growth, i.e., 80-100 days after sowing (data
not shown). All the analyses reported here were carried
out by harvesting the leaf material from the plants of
this age. In initial experiments, no cofactors were
added, and the rate of conversion was found to be very
low. A number of cofactors in different combinations
and in different concentrations were added, and their
effect on the conversion of various precursors to arte-
misinin was observed. As shown in Table 2, it was
(
Table 3). The SDS-PAGE of the 75-100% fraction
showed two prominent bands at molecular range of 12.5
and 14 kD (data not shown). Experiments are in
progress to further purify these enzymes.
+
found that a combination of ATP (0.1 mM), NADPH+H
(
0.1 mM), MgSO4 (1 mM) and MnSO4 (1 mM) added
together resulted in the maximum conversion. It is well
Exp er im en ta l Section
2
+
established that divalent cations ions such as Mg and
Mn2 are essential for the terpene biosynthesis. We
+
12
Gen er a l Exp er im en ta l P r oced u r es. Analytical
HPLC was performed on a Perkin-Elmer instument
using a C18 column (4 i.d. × 125 mm) (mobile phase,
100 mM PO4 buffer/MeOH 60:40) and UV detection at
260 nm. Electrospray mass spectra were obtained using
a VG platform (Fisons Instruments, UK quadrapole
mass spectrometer equipped with Masslynx data sys-
tem, Dynolite detector system and pneumatic nebulizer-
assisted electrospray LC/MS interface). The molecular
2
+
2+
have, therefore, included both Mg and Mn in our
assay system.
The HPLC analysis does not show the presence of any
other intermediate compound during the conversion of
arteannuin B to artemisinin, suggesting that artean-
nuin B is the immediate precursor of artemisinin, which
8
is in agreement with Nair et al. It has been reported
previously as well as detected by us through HPLC that
the relative yield of arteannuin B is high in plants.
Arteannuin B can serve as a suitable precursor for in
vitro bioproduction of artemisinin using an immobilized
enzyme system.
To study the biosynthetic pathway of artemisinin the
radiolabeled precursors are essential, so that their
conversion can be followed conveniently. It is tedious
to synthesize labeled precursors by chemical means, and
their biosynthesis in vivo results in poor incorporation
of radioactivity leading to very low specific activity. We
used a novel approach to synthesize various radiolabeled
+
ions were detected as total ion current in ES mode.
1
13
The H, C, and 2D-NMR spectra were obtained with
a Bruker-DRX (300 MHz) spectrometer operating at 300
1
13
MHz for H and 75 MHz for C, respectively; chemical
shifts are reported on (scale and coupling constants are
H
in Hz). HMQC (J CH ) 145 Hz) and HMBC ( J CH-
optimized for 8 Hz) experiments were performed using
standard Bruker UXNMR pulse sequences.
P la n t Ma ter ia l. A. annua L. plants were raised from
the seeds obtained from Walter Reed Army Institute of
Research, Washington, D.C. The plants were main-

Notes
J ournal of Natural Products, 1998, Vol. 61, No. 5 635
Ta ble 2. Effect of Various Cofactors on in Vitro Biosynthesis of Artemisinina
artemisinin
biosynthesized
(µmol/mg protein)
Mg2
+
Mn2+
(1 mM)
ATP
(0.1 mM)
NADPH+
(0.1 mM)
FAD
(0.1 mM)
Fe2+
(1 mM)
R-ketoglutarate
cofactors
(1 mM)
(2.5 mM)
1
2
3
4
5
6
-
+
+
+
+
+
-
+
+
+
+
+
-
+
+
+
+
+
-
-
-
-
+
+
-
+
-
-
-
-
-
+
-
+
+
-
-
-
+
-
+
-
0.003
0.012
0.012
0.009
0.020
0.026
a
The leaf material was homogenized in buffer 2 and incubated with artemisinic acid for 150 min in the presence of various cofactors.
Optimum concentration of the each cofactor was used. The amount of artemisinin was estimated as described in the text.
droarteannuin B were dissolved in 10% DMSO. The
concentration of all the stock solutions was 1 mg/mL.
In cu ba tion . Leaf homogenate (equivalent to 1 g
fresh leaf material) was incubated with 100 µL of
substrate stock (100 µg, and the following cofactors were
+
added: 0.1 mM ATP + 0.1 mM NADPH + H + 1 mM
MgSO4 + 1 mM MnSO4. Controls received all the
constituents except the substrate. The controls thus
show an endogenous level of artemisinin. These values
were deducted from the values obtained that have been
expressed as µmol artemisinin produced per mg of
protein. The reaction vials were stoppered and incu-
bated at 30 °C for 150 min in a shaking water bath.
Extr a ction a n d Detection of En d P r od u cts. After
misinic acid to artemisinin. Leaf material (80-100 days old)
F igu r e 1. Time kinetics for enzymatic conversion of arte-
incubation, the reaction mixture was extracted with
diethyl ether (3×) and dried over Na2SO4. Artemisinin
was detected on TLC plates by cochromatography with
standard artemisinin using benzene/EtOAc (9:1) as the
solvent system. Artemisinin gave a characteristic yel-
low spot when the plates were sprayed with anisalde-
hyde reagent.
was homogenized in HEPES-â-mercaptoethanol buffer and
incubated with 100 µg artemisinic acid for various time
intervals along with the required cofactors. Artemisinin
synthesized was quantitated by HPLC. See text for details.
Ta ble 3. Distribution of the Enzyme Activity and Proteins in
Various Ammonium Sulfate-Precipitated Fractions (Enzyme
Unit: 1 µg/min of Artemisinin Produced)a
Qu a n tita tion . The synthesized artemisinin was
protein
(mg/mL)
enzyme activity
(unit/mL)
specific
activity
1
3
fraction (%)
quantitated using the method of Zhao et al. with
certain modifications. Radiolabeled artemisinin was
purified through preparative TLC and was monitored
in a liquid scintillation counter (Pharmacia LKB).
Syn th esis of Dih yd r oa r tea n n u in B. Dihydroarte-
annuin B was synthesized by the reduction of artean-
nuin B at the C-13 position using the catalyst-based
reduction protocol described for other compounds by
crude homogenate
-25
5-50
0-75
5-100
1.310
1.000
0.205
0.065
0.040
0.0270
0.0002
0.0002
0.0004
0.0064
0.0207
0.0002
0.0010
0.0059
0.1620
0
2
5
7
a
Leaf homogenate was fractionated by precipitation with
various concentrations of ammonium sulfate and desalted, and the
related enzyme activity and protein content were measured.
1
4
Zhou.
For this, arteannuin B (0.04 mmol) was dis-
tained in the herbal garden of J amia Hamdard following
the standard agronomic practices. The leaf material for
enzyme assay was harvested at preflowering stages of
plant growth starting from the age of 80-100 days.
En zym e Assa y. A cell-free system obtained from the
leaf homogenate was used as enzyme source. To
prepare the cell-free system, two different homogenizing
buffers were used: (1) Tris-HCl (50 mM) containing
EDTA (3 mM), ascorbic acid (5 mM), DTT (2 mM),
sucrose (300 mM), PVP (0.02 g/mL); final pH 7.5; (2)
HEPES (50 mM) containing â-mercaptoethanol (2 mM),
final pH 7.2. Young leaves were washed with deionized
water and ground with sand in a prechilled mortar and
pestle with 2 volumes of chilled buffer. The ratio of
biological material to buffer was kept as 1:2. The
homogenate was centrifuged at 20000g for 15 min, and
the resulting supernatant was used as the enzyme
source.
solved in 10 mL of MeOH and cooled to -15 °C, and
NiCl2‚6H2O (0.42 mmol) was added. It was followed by
addition of NaBH4 (3.9 mmol), added slowly with
constant stirring). The reaction mixture was incubated
for 30 min in an ice bath with constant stirring.
Thereafter, the reaction was stopped by diluting it with
10 mL of H2O and its acidification to pH 6.0 with 3 N
HCl. It was later extracted with EtOAc, the solvent was
evaporated, and the residue was crystallized with
EtOAc and hexane. The complete conversion of arte-
annuin B to its dihydro derivative was determined by
the analysis on TLC (40% EtOAc in hexane was used
as the solvent system), followed by spraying with
vanillin/H2SO4 reagent. Arteannuin B gave a blue color
while dihydroarteannuin B gave a turquoise blue color
(Rf 0.8). The reduction was further confirmed by MS
and NMR studies. Dihydroarteannuin B was obtained
1
as crystals (EtOH and hexane): mp 168-170 °C; H
Su bstr a te. Artemisinic acid (2), arteannuin B (3),
and dihydroarteannuin B (4) were used as substrates
for the enzyme assay. Artemisinic acid was dissolved
in 10% Triton X 100, and arteannuin B and dihy-
NMR (CDCl3, 300 MHz) (2D-COSY and HMQC) assign-
ments δ 2.82 (1H, s, H-5), 2.64 (1H, m, J ) 7.5 Hz,
H-11), 1.79 (2H, m, H-3), 1.69 (2H, m, H-2), 1.80 (2H,
m, H-9), 1.78 (2H, m, H-7),1.34 (3H, s, Me-15), 1.37 (1H,

6
36 J ournal of Natural Products, 1998, Vol. 61, No. 5
Notes
m, H-10), 1.32 (1H, m, H-1), 0.94 (3H, d, J ) 7.5 Hz,
Ack n ow led gm en t. These studies were supported by
a research grant (BT/TF/09/09/91) from DBT, Govern-
ment of India. S.B. and A.G. thank CSIR and UGC
respectively, for the research fellowships.
1
3
Me-14), 1.20 (3H, d, J ) 7.5 Hz, Me ) 13). C NMR
(CDCl3); assignment from HMQC experiment δ 56.47
(C-5), 53.02 (C-7), 42.96 (C-1), 36.06 (C-11), 33.19 (C-
9
1
), 33.19 (C-10), 20.83 (C-3), 21.69 (C-15),16.52 (C-14),
Refer en ces a n d Notes
4.51 (C-2), 11.06 (C-13), 22.84 (C-8). Electrospray MS
+
(
2
CH3CN/H2O 50:50) m/z M + H 251.0 (43.72), M + H2O
(1) J ung, M. Curr. Med. Chem. 1984, 1, 35-49.
+
+
(2) Zhou, W. S.; Xu, X. X. Acc. Chem. Res. 1994, 27, 211-216.
68.0 (100), M + Na 273.0 (48.91), M + K 289.0 (5.19).
(3) Klayman, D. L. Science 1985, 288, 1049-1055.
3
Similarly, H-labeled dihydroarteannuin B (specific
(4) Luo X. D.; Shen, C. C. Med. Res. Rev. 1987, 7, 29-52.
(5) Trigg, P. I. In Economic and Medicinal Plant Research; Wagner
H., Hakino H., Farnsworth N. R., Eds.; Academic Press: New
York, 1989; Vol. 3, pp 19-55.
9
activity 1.29 × 10 cpm/mmol) was synthesized when
NaBT4 was used in place of NaBH4.
P a r tia l P u r ifica tion of En zym es. Efforts were
made to partially purify the enzymes. In the first set
of experiments, 5 g of leaf material was homogenized
in HEPES-â-mercaptoethanol buffer, and then the
proteins were precipitated using ammonium sulfate. A
total of five fractions of 0-25%, 25-50%, 50-75%, and
(6) Fourth Meeting of the Scientific Working Group on the Chemo-
therapy of Malaria, Beijing. People’s Republic of China, 1981.
WHO Report TDR/Chemal-SWG (4)/QHS/81.3.
(
7) Akhila, A.; Thakur, R. S.; Popli, S. P. Phytochemistry 1987, 26,
1927-1930.
(
8) Nair, M. S.R.; Basile D. V. J . Nat. Prod. 1993, 56, 1559-1566.
9) Wang, Y. U.; Xia, Z. Q.; Zhou, Y.; Wu, Y. L.; Huang J . J .; Wang
Z. Z. Chin. Chem. 1993, 11, 452-463.
(
(
(
(
(
10) Sangwan, R. S.; Agarwal, K.; Luthra, R.; Thakur, R. S.; Sang-
wan, N. S. Phytochemistry 1993, 34, 1301-1302.
11) KudaKasseril, E. J .; Lam, L.; Staba, E. J . Planta Med. 1987,
53, 280-284.
7
5-100% (NH4)2SO4 saturation and the final superna-
tant were obtained and then desalted through a Sepha-
dex G -25 column. The protein concentration of each
desalted fraction was estimated by using the Bradford
method15 using bovine serum albumin as the standard.
The enzyme assay was carried out using 1 mL of each
of the fraction with dihydroarteannuin B as the sub-
strate.
12) V o¨ geli, U.; Freeman, J . W.; Chappell, J . Plant Physiol. 1990,
9
3, 182.
13) Zhao, S. S.; Zeng, M. Y. Planta Med. 1985, 51, 233-237.
(14) Zhou, W. S. Pure Appl. Chem. 1986, 58, 817-824.
(15) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254.
NP970024S