Structure-activity relationships of the antimalarial agent artemisinin. 8. Design, synthesis, and CoMFA studies toward the development of artemisinin-based drugs against leishmaniasis and malaria

Article abstract of DOI:10.1021/jm030181q

Artemisinin (1) and its analogues have been well studied for their antimalarial activity. Here we present the antimalarial activity of some novel C-9-modified artemisinin analogues synthesized using artemisitene as the key intermediate. Further, antileish

Full text of DOI:10.1021/jm030181q

4244
J . Med. Chem. 2003, 46, 4244-4258
Articles
Str u ctu r e-Activity Rela tion sh ip s of th e An tim a la r ia l Agen t Ar tem isin in . 8.
Design , Syn th esis, a n d CoMF A Stu d ies tow a r d th e Develop m en t of
Ar tem isin in -Ba sed Dr u gs a ga in st Leish m a n ia sis a n d Ma la r ia †
Mitchell A. Avery,*,‡,§ Kannoth M. Muraleedharan,‡ Prashant V. Desai,‡ Achintya K. Bandyopadhyaya,‡
Marise M. Furtado,| and Babu L. Tekwani|
Department of Medicinal Chemistry, School of Pharmacy, National Center for Natural Products Research, and
Department of Chemistry, University of Mississippi, University, Mississippi 38677
Received April 17, 2003
Artemisinin (1) and its analogues have been well studied for their antimalarial activity. Here
we present the antimalarial activity of some novel C-9-modified artemisinin analogues
synthesized using artemisitene as the key intermediate. Further, antileishmanial activity of
more than 70 artemisinin derivatives against Leishmania donovani promastigotes is described
for the first time. A comprehensive structure-activity relationship study using CoMFA is
discussed. These analogues exhibited leishmanicidal activity in micromolar concentrations,
and the overall activity profile appears to be similar to that against malaria. Substitution at
the C-9â position was shown to improve the activity in both cases. The 10-deoxo derivatives
showed better activity compared to the corresponding lactones. In general, compounds with
C-9R substitution exhibited lower antimalarial as well as antileishmanial activities compared
to the corresponding C-9â analogues. The importance of the peroxide group for the observed
activity of these analogues against leishmania was evident from the fact that 1-deoxyartemisinin
analogues did not exhibit antileishmanial activity. The study suggests the possibility of
developing artemisinin analogues as potential drug candidates against both malaria and
leishmaniasis.
In tr od u ction
pharmaceutical profile.15-22 Artemisinin derivatives
such as artesunate and arteether, either alone or in
combination with other drugs, are being used as a
remedy for malaria in many endemic areas.23 Although
several C-10 ether derivatives of dihydroartemisinin are
in clinical use against malaria, poor oral bioavailability,
short plasma half-life, and concerns regarding neuro-
toxicity limit their use.24
Artemisinin (1), a sesquiterpene endoperoxide isolated
from the plant Artemisia annua, has emerged as an
important treatment option toward drug-resistant strains
of Plasmodium falciparum.1,2 After its isolation in 1972
by Chinese researchers and structure elucidation in
1979,3 artemisinin became the center of attention for
drug development against drug-resistant malaria. It has
been shown to quickly lower the parasite levels even in
severe cases of cerebral malaria. Artemisinin (1) has
demonstrated activity against drug-resistant strains of
P. falciparum such as W2-Indochina (chloroquine re-
sistant) and D6-Sierra Leone (mefloquine resistant)
clones. The mechanism of action of this unusual 1,2,4-
trioxane is believed to involve an initial interaction of
the endoperoxide moiety with heme, resulting in the
formation of free radical intermediates.4-12 The exact
mechanism by which these free radicals cause parasitic
death is still not completely understood.13,14 Recently
much attention has been given to the chemical modifi-
cation of artemisinin in order to improve its efficacy and
Artemisinin and its analogues have also been studied
for their anticancer,25-30 and antifungal action.31,32 Our
search for reports on the action of artemisinin against
other diseases caused by parasites related to P. falci-
parum, like trypanosomiasis and leishmaniasis, re-
sulted in only one relevant study in which its activity
against leishmania was described.33 Although it was a
preliminary finding, it did suggest a possible use of
artemisinin and related analogues against leishmania-
sis.
Human leishmaniasis comprises a heterogeneous
spectrum of diseases. Three major forms are generally
distinguished: cutaneous leishmaniasis, mucocutaneous
leishmaniasis, and visceral leishmaniasis, of which the
latter is potentially lethal. They are caused by various
species of the protozoan parasite Leishmania and
transmitted by female sandflies.34 The disease is cur-
rently proposed to affect some 12 million people in 88
countries.35-39 It is estimated that ∼350 million people
are exposed to infection by different species of leishma-
† This paper is dedicated to Dr. J effrey A. Vroman45 (deceased J an
2003).
* To whom correspondence should be addressed. Phone: (662)-915-
5879. Fax: (662)-915-5638. E-mail: mavery@olemiss.edu.
‡ Department of Medicinal Chemistry, University of Mississippi.
§ Department of Chemistry, University of Mississippi.
| National Center for Natural Product Research, University of
Mississippi.
10.1021/jm030181q CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/21/2003

SAR of the Antimalarial Agent Artemisinin
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4245
Sch em e 1
nia parasite. Leishmania/HIV coinfection is now con-
sidered as an “emerging disease” especially in southern
Europe, where 25-70% of adult visceral leishmaniasis
cases are related to HIV infection.
substitution at C-9â and the removal of the carbonyl
group to form the 10-deoxo system have been shown to
give optimal oral activity. On this basis, we have
modified the C-9 position of the parent skeleton with
various aliphatic and arylalkyl groups. The most prom-
ising candidates were those with a halogen-substituted
phenethyl group at the C-9 position. However, oral
bioavailability of most of these molecules was compara-
tively low.46a
In this paper we present the antimalarial activity of
some novel artemisinin analogues. Also, we report the
first comprehensive structure-activity relationship study
of more than 70 artemisinin analogues against leish-
mania promastigotes using comparative molecular field
analysis (CoMFA).
The current treatment for leishmaniasis involves
administration of pentavalent antimony complexed to
a carbohydrate in the form of sodium stibogluconate
(pentosam or Sb(V)) or meglumine antimony (glucan-
tine), which are the only antileishmanial chemothera-
peutic agents with a reasonable therapeutic index.40,41
The exact chemical structure and mode of action of
pentavalent antimonials is still uncertain but, as with
most metals, is thought to be multifactorial.41 Ampho-
tericin B and pentamidine are the second line of
antileishmanial agents, but they are reserved for non-
responding infections due to potential toxicity.42 The
development of lipid-associated amphotericin B provided
important clinical progress in this area.43,44 Even though
many promising leads exist, none of the above-men-
tioned drugs show promise as ideal treatment alterna-
tives for the developing world, due to their route of
administration (parenteral), high clinical failure rate
(40% for pentavalent antimonial agents), side effects,
and cost. Thus, the identification of an effective anti-
leishmanial agent for oral administration would be a
significant step toward reducing mortality due to leish-
maniasis.
Our ongoing program to develop artemisinin-based
antimalarial drugs, coupled with possible activity against
leishmania, prompted us to explore the possibility of
designing and developing artemisinin analogues having
both antimalarial and leishmanicidal activity.
Comparative molecular field analysis of over 200
artemisinin analogues from this laboratory has shown
important electrostatic and steric requirements that
could improve the antimalarial activity of the artemisi-
nin skeleton.45d Steric bulk in the form of phenethyl
Ch em istr y
The synthesis of new artemisinin derivatives, which
are expected to have improved bioavailability, is pre-
sented in Scheme 1. A selenium-free route to arte-
misitene (3) was employed, after which 3 was then
reacted with the appropriate arylalkyl bromide syn-
thons. For radical-induced Michael addition of arte-
misitene, a dilute solution of tributyltin hydride was
added slowly over 8 h, via syringe pump, to a refluxing
solution of artemisitene, AIBN, and the corresponding
bromide. The solvent was concentrated to 25 mL, stirred
with a saturated solution of KF for 10 h, filtered, and
concentrated. The products were separated by column
chromatography to afford the R- and â-epimers. Yield
of the â-isomer ranged from 14% to 32% and that of
R-isomer from 30% to 43%. The 9R-substituted ana-
logues were converted to the 9â-congeners by refluxing
with DBU in THF for 12 h. Table 1 shows the yields of
R- and â-isomers prior to equilibration.
As shown in Scheme 2, the bromide synthons used
(6a -c) were synthesized from the corresponding acids

4246 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Avery et al.
Sch em e 2
Ta ble 1. Percentage Yields of R- and â-Diastereomers Prior to
Equilibration as Described in Scheme 1
Ta ble 2. Antimalarial Activity of New Artemisinin Analoguesa
P. falciparum
P. falciparum
â-isomers
compd
R-isomers
IC50 (nM)
IC50 (nM)
no. D6 clone W2 clone
no.
D6 clone W2 clone
% yield
compd
% yield
14 (29)a
21 (24)a
17 (33)a
22c
15a
19
24a
26a
28
35
30
38
39
43
32
15
18
19
24
25
26
27
14.9
4.7
10.4
43.9
30.13
7.9
14.9
1.3
3.4
50.8
20
28
30
31
55
12.6
27.8
38.7
16.5
17.0
16.5
169.5
8.2
11.3
62.3
14.3
14.5
15
18b (R ) Ph)
24 (R ) SMe)
26b (R ) OTIPS)
27b (R ) 3,5-diCF3)
30a (R ) NMe2)
32 (45)a
30 (42)a
artemisinin
chloroquine
mefloquine
4.7
16
232.6
194.6
31
18.8
a
Numbers in parentheses represent the combined yield of C-9â
isomers after one round of DBU epimerization of R-diastereomer.
a
Activity was measured using the pLDH assay as described in
text.47 All the analogues were tested at least at six different
concentrations, each in duplicate. The mean values were used to
generate the growth inhibition curves and determination of IC50
values.
b
The analytical data of these compounds are described in ref 46a.
c Not subjected to equilibration.
Sch em e 3
and 7.04 to δ 7.28 and 7.65, respectively. The conversion
of sulfide (24) to sulfone (25) was accomplished in 94%
yield by treatment with m-CPBA at -78 °C in CH2Cl2.
Similarly, the silyl group in 26b could be removed
conveniently using tetrabutylammonium fluoride in
THF, at room temperature to afford 91% of 26. Most of
the newly synthesized artemisinin analogues exhibited
superior antimalarial activities compared to artemisi-
nin, as described in Table 2.
first by reduction to alcohols followed by bromination
using PPh3/CBr4 (for 6a ) or PPh3/Br2 (for 6b and 6c).
The synthesis of C-9 cyanoethyl derivative of arte-
misinin (15) is presented in Scheme 3. Here, bromo-
acetonitrile (6d ) was used in the Michael addition to
artemisitene to afford the â-isomer 15 and R-isomer 15a
in 14% and 35% yields, respectively. As described
previously, the R-isomer was converted to the â-form
in 41% yield, by refluxing with DBU in THF for 10 h.
The 10-deoxo derivative 55a was synthesized from
compound 30a in 35% yield using a two-step reaction
sequence involving two reductions, first affording the
lactol (30b) in 82% yield, followed by a second reduction
with triethylsilane and boron trifluoride etherate (Scheme
4). Due to the potential instability of the peroxide group
in the presence of the amino function,12b the N,N-
dimethylamino derivatives 30a and 55a were converted
to the corresponding hydrochloride salts by treatment
with equimolar amounts of HCl in ether. The formation
All the other compounds studied in this report and
listed in Tables 3 and 4 were retrieved from library
storage. Virtually all of the stored compounds were 10-
15 years old and were still of high purity and could be
used directly for bioassay. These compounds were (7-
13, 16-18, 20),46b (32-39),45c (40-43, 50, 51, 60-62),46c
(14, 21-23, 27, 29, 44-49, 52-54, 56-59),46a (63-
68),46d 69,46e 70,46f (71-72),46g and (73, 74).46h
To understand the importance of the peroxide entity
for the biological activity in this series, we synthesized
a cross section of various 1-deoxoartemisinin derivatives
(Scheme 5, 75-79) of the active compounds studied.
This was accomplished by hydrogenolysis of an ethyl
acetate solution of the corresponding peroxide com-
pounds in the presence of 10% Pd-C. During this
process, ring opening of the trioxane ring leads tran-
siently to a tricyclic diol, which readily undergoes loss
of water and reketalization to regenerate the noroxa-
tetracycles. Normally, this occurs during the first
purification by flash silica gel chromatography. Subse-
quent rigorous HPLC purification was conducted to
ensure that all traces of peroxide were removed.
1
of hydrochloride salts was evidenced by the shift of H
NMR signals corresponding to the NMe2 groups from δ
2.9 to 3.14 and that of aromatic doublets from δ 6.69

SAR of the Antimalarial Agent Artemisinin
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4247
Sch em e 4
In Vitr o An tim a la r ia l a n d An tileish m a n ia Assa y.
The antimalarial activity of the analogues was deter-
mined in vitro on chloroquine-sensitive (D6-Sierra
Leone) and -resistant (W2-Indochina) strains of P.
falciparum. The 96-well microplate assay is based on
evaluation of the effect of the compounds on the growth
of asynchronous cultures of P. falciparum, determined
by the assay of parasite lactate dehydrogenase (pLDH)
activity.47 The appropriate dilutions of the compounds
were prepared in DMSO or RPMI-1640 medium and
added to the cultures of P. falciparum (2% hematocrit,
2% parasitemia) set up in clear flat-bottomed 96-well
plates. The plates were placed into the humidified
chamber and flushed with a gas mixture of 90% N2, 5%
CO2, and 5% O2. The cultures were incubated at 37 °C
for 48 h. Growth of the parasite in each well was
determined by pLDH assay using Malstat reagent. The
medium and RBC controls were also setup in each plate.
The standard antimalarial agents, chloroquine and
artemisinin, were used as the positive controls while
DMSO was tested as the negative control.
Tables 3 and 4. Compounds for which exact IC50 values
were not available (28, 71, and 74) were not included
in the study. Diastereomeric mixtures and racemates
such as the dihydroartemisinin derivatives 40-49 were
also excluded from the analysis. A test set consisting of
11 molecules with diverse chemical structures and
activities was used for validation and was not included
in the model development stage. The remaining 45
molecules constituted the training set. The negative
logarithm of the molar IC50 values (pIC50) was used as
activity for the study in order to achieve a normal
distribution of activity values.
Using the reported crystal structure for artemisinin49
as a template, a database of both the test and training
set compounds was created. The molecules were mini-
mized with the Tripos force field and aligned by fitting
atoms C-1a, C-3, C-5a, and C-8a using the align
database command in Sybyl (Figure 1). A similar
alignment strategy was successfully used earlier to
derive CoMFA models for antimalarial activity of vari-
ous artemisinin analogues.45d
Antileishmanial activity of the compounds was tested
on a transgenic cell line of Leishmania donovani pro-
mastigotes expressing firefly luciferase.48a In a 96-well
microplate assay, compounds with appropriate dilution
were added to the leishmania promastigotes culture (2
× 106 cell/mL). The plates were incubated at 26 °C for
72 h and growth of leishmania promastigotes was
determined by luciferase assay with Steady Glo reagent
(Promega). Pentamidine and amphotericin B were used
as standard antileishmanial agents. The analogues were
simultaneously tested for cytotoxicity on VERO cells
(monkey kidney fibroblast) by Neutral Red assay.48b
All the analogues were tested at least at six different
concentrations, each in duplicate. The mean values were
used to generate the growth inhibition curves and
determination of IC50 values.
Com p a r a t ive Molecu la r F ield An a lysis (Co-
MF A). All calculations were carried out on SGI Octane2
machine, equipped with two parallel R12000 processors,
V6 graphics board, and 512 MB memory. Sybyl v. 6.9
(Tripos Inc., USA) was utilized for the molecular model-
ing studies including CoMFA. The CoMFA study was
performed on a total of 56 molecules out of 69 listed in
CoMFA involves the measurement of the electrostatic
and steric fields around a template molecule and the
relationship of these measurements to the molecule’s
biological activity. These measurements are taken at
regular intervals throughout the lattice using a probe
atom of designated size and charge. The steric (Lennard-
J ones) and electrostatic (Coulomb) interaction energies
between a probe atom and the molecules were calculated
within a region extending the molecular ensemble by 4
Å in each dimension (x, y, z) with a grid spacing of 2.0
Å. The effect of probe atoms was studied using three
different probes: proton, positively charged sp3 carbon
atom, or negatively charged sp3 oxygen atom. A distance-
dependent dielectric constant (1/r) was chosen, and the
maximum field values were truncated to 30 kcal/mol for
the steric and (30 kcal/mol for the electrostatic inter-
action energies.
There are numerous formalisms for calculating the
partial charges of atoms, and each method has a decisive
effect on the overall CoMFA model. To gauge the effect
of various charge assignment schemes on the Co-
MFA results, we calculated partial charges using the
Gasteiger-Hu¨ckel method,50 the Gasteiger-Marsili

4248 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Ta ble 3. Antileishmanial Activity of Artemisinin Analogues
Avery et al.
no.
R1
R2
R3
IC50 (µM)a
1
7
8
CH3 (artemisinin)
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
Et
H
H
H
CH3
H
H
H
H
124.0
242.3
60.7
101.2
16.8
17.8
12.3
8.5
9
CH3
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42c
43d
44
45d
46d
47d
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
nPr
CH2CHdCH2
nBu
(CH2)3CH(CH3)2
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
(CH2)3CF3
(CH2)2CN
CH2Ph
(CH2)2Ph
(CH2)3Ph
H
(CH2)4Ph
(CH2)3C6H4(m-F)
(CH2)3C6H4(p-Cl)
(CH2)3C6H4(p-OMe)
(CH2)3C6H4(p-SMe)
(CH2)3C6H4(p-SO2Me)
(CH2)3C6H4(p-OH)
(CH2)3C6H4(3,5-(CF3)2)
H
(CH2)3C6H4(3,5-F2)
(CH2)3C6H4(p-NMe2‚HCl)
H
H
H
H
H
H
H
H
H
79.3
38.9
15.3
40.3
11.6
45.3
7.5
H
(CH2)3Ph
H
H
H
H
H
H
H
H
3.7
11.9
12.0
1.4
53.8
44.7
14.4
NAb
49.7
1.9
88.3
177.1
40.5
257.7
31.0
12.6
11.1
35.5
4.5
44.4
70.3
44.8
382.9
26.3
2.9
(CH2)3C6H4(3,5-(CF3)2)
H
H
CH3
CH3
(CH2)3C6H4(p-NMe2‚HCl)
Et
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
nPr
nBu
(CH2)2CO2Et
(CH2)2Ph
(CH2)3C6H4(p-Cl)
Et
(CH2)4Ph
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
H
H
nBu
nBu
H
CH3
CH3
nBu
(CH2)3CF3
(CH2)3C6H4(p-CF3)
(CH2)3C6H4(m-Cl)
(CH2)3C6H4(p-Cl)
(CH2)3C6H4(3,5-F2)
(CH2)3C6H4(3,5-(CF3)2)
H
16.6
4.7
33.0
NAb
49.2
46.6
4.9
CH3
(CH2)3C6H4(p-Cl)
(CH2)3C6H4(m-Cl)
(CH2)3C6H4(3,4-Cl2)
(CH2)3C6H4(p-NMe2‚HCl)
(CH2)3C6H4(p-CF3)
(CH2)3C6H4(p-OMe)
(CH2)3C6H4(3,5-F2)
(CH2)3C6H4(3,5-(CF3)2)
H
CH3
CH3
CH3
CH3
CH3
CH3
3.7
2.3
36.0
1.8
17.4
11.0
0.3
70.8
151.8
37.7
0.01
1.35
CH3
nPr
nBu
H
H
(CH2)2Ph
amphotericin B
pentamidine
a
Activity was measured using luciferase assay as described in the text.48a Activity for 40-49 was measured for the racemic mixture.
All the analogues were tested at least at six different concentrations, each in duplicate. The mean values were used to generate the
growth inhibition curves and determination of IC50 values. Not active up to 125 µg/mL. c As arteether. These compounds showed
cytotoxicity in the concentration range of 2000-2700 µg/mL while all the other compounds did not exhibit observable cytotoxicity as
measured by the procedure described in the text.
b
d

SAR of the Antimalarial Agent Artemisinin
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4249
Ta ble 4
no.
IC50 (µM)a
no.
IC50 (µM)a
63
64
65
66
67
68
44.5
185.5
27.5
56.0
25.5
32.3
69
70
71
72
73
74
148.0
131.4
NAb
158.3
203.5
NAb
a
Activity was measured using luciferase assay as described in the text.48a All the analogues were tested at least at six different
concentrations, each in duplicate. The mean values were used to generate the growth inhibition curves and determination of IC50 values.
b
Not active up to 125 µg/mL.
Sch em e 5
optimal number of components and column filtration
values ranging from 0.5 to 2.0 to yield the final q2 value.
The standard CoMFA scaling was applied to both steric
and electrostatic energies. The non-cross-validated mod-
els were then calculated. Statistical validity of the
models was assessed by the variance (r2), standard error
of estimate (S), and the F-values.
The models were further improved by the q2-guided
region selection (q2-GRS) technique. The q2-GRS process
has been described in detail elsewhere54 and in recent
reviews.55 Conventional CoMFA routines apply equal
weight to data from each lattice point in a given field,
whereas the final result actually emphasizes the limited
areas of three-dimensional space as important for
biological activity. It is suggested that the deficiencies
of conventional CoMFA may be overcome by eliminating
from the analysis those areas of three-dimensional space
where changes in steric and electrostatic field do not
method,51 and AM1 calculations using MOPAC 6.052 as
implemented in Sybyl.
To determine the optimal number of components
corresponding to the lowest standard error of prediction
and the highest cross-validated r2 (q2), SAMPLS53 with
leave-one-out (LOO), no column filtering, and a maxi-
mum of 10 variables was carried out. This was followed
by cross-validation using the LOO method with the
F igu r e 1. Alignment of artemisinin analogues used for CoMFA model development.

4250 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Avery et al.
Ta ble 5. Statistical Results of CoMFA Analysis Using q2-GRS
optimum
component
d
model
chargea
probeb
CFc
q2
r2
S
F
r2
bs
1
2
3
4
5
6
7
8
9
GH
GH
GH
GM
GM
GM
AM1
AM1
AM1
C.3
O.3
H
C.3
O.3
H
C.3
O.3
H
6
6
6
6
6
6
4
4
4
2.0
1.5
1.0
0.5
1.5
1.5
0.5
0.5
0.5
0.803
0.805
0.774
0.712
0.732
0.751
0.683
0.670
0.679
0.966
0.969
0.972
0.958
0.956
0.964
0.964
0.965
0.956
0.131
0.125
0.118
0.145
0.148
0.134
0.132
0.128
0.144
179
196
220
144
138
170
264
279
218
0.977
0.980
0.980
0.974
0.971
0.975
0.960
0.960
0.959
a
b
GH, Gasteiger-Hu¨ckel; GM, Gasteiger-Marsili. H ) proton, C.3 ) sp3 carbon with +1 charge, O.3 ) sp3 oxygen with -1 charge.
c Column filtration value. Mean r2 values obtained by bootstrap analysis.
d
correlate with biological activity.54 The q2-GRS tech-
nique eliminates those areas from the analysis based
on low values of q2 obtained for such regions individu-
ally. Thus, it optimizes the region selection for the final
PLS analysis. In fact, in the present study, use of the
q2-GRS routine resulted in models with higher q2 and
better statistics.
To cross check for any chance correlations in the PLS
analysis, the PLS was repeated with complete random-
ization of the biological data and the q2 recalculated.
This was repeated 10 times. Also, to obtain statistical
confidence limits for the analysis, PLS analysis using
100 bootstrap runs56 with the optimum number of
components was performed.
Models were further evaluated on the basis of their
ability to predict activities of the test set molecules. The
predictive r2 for the test set was calculated using the
formula, predictive r2 ) 1- SSD/PRESS, where SSD is
the sum of squared deviations from the mean and
PRESS is the sum of squared differences between the
actual and the predicted values.
The 10-deoxo derivatives (50-62) were expected to be
more stable compared to the corresponding acetals,
which are relatively unstable due to possible solvolysis
or metabolism to a masked aldehyde carbonyl (lactols).
The 10-deoxo compounds also exhibited superior activity
to the corresponding lactones. The requirement of the
unique skeletal framework of artemisinin for its activity
is evident from the examples 69-74 (Table 4), where
one of the rings has been modified. These compounds
did not show any noticeable activity against leishmania.
In general, activity against P. falciparum appears to
be more pronounced (nanomolar range)45,46 compared
to that against L. donovani (micromolar range). How-
ever, in all the above cases, a similar profile of activity
was shown by the artemisinin derivatives against both
the organisms. The substitution at C-3 appears to be
an exception. Although, bulky substitution at C-3 was
shown to improve the antimalarial activity, in case of
antileishmanial potency this substitution does not seem
to make a significant contribution (Table 3), as exempli-
fied by 34 and 61. A more quantitative analysis of the
structure-activity relationship is presented below.
Str u ctu r e-Activity Rela tion sh ip . A comprehen-
sive QSAR study for antimalarial activity of 211 arte-
misinin analogues (which included most of the mol-
ecules presented in this paper) was reported recently
by our group45d and hence a QSAR study for this activity
would be redundant. To determine various factors
affecting the antileishmanial activity and understand
the relationship between structural requirements for
antimalarial vs antileishmanial activity, a CoMFA study
was carried out. Several CoMFA models were obtained
using different charges, probe atoms, and column filtra-
tion values. Statistical results of these models are
presented in Table 5.
Resu lts a n d Discu ssion
Antimalarial activities of the new analogues synthe-
sized are presented in Table 2. These compounds were
designed and synthesized to improve oral bioavailability
and to examine the effect of C-9R/C-9â substitution on
the antimalarial activity. The analogues exhibited
promising in vitro activity against P. falciparum and
thus offer an opportunity to develop better artemisinin-
based drugs.
To our surprise, most of the artemisinin analogues
tested were notably active against leishmania. The
activities of these molecules against leishmania are
presented in Tables 3 and 4. In general, these can be
classified into six different structural classes depending
upon the nature and sites of derivatization:2,45,46 (1)
alkyl or arylalkyl groups at the C-3 position, (2) alkyl
or arylalkyl groups at C-9, (3) analogues with substitu-
tion at both C-3 and C-9 positions, (4) 10-deoxo com-
pounds, (5) lactams, and (6) derivatives that lack one
or the other of the artemisinin ring systems (seco ring
systems).
For any given charge calculation method, the three
probes were shown to have little, if any, effect on the
overall statistical quality of the models (Table 5). The
most significant impact on the model quality was
observed with the use of various charge-calculating
methods. AM1 charges resulted in models with q2 values
ranging between 0.67 and 0.683. On the other hand,
with the use of Gasteiger-Hu¨ckel charges q2 values of
more than 0.8 could be obtained. In fact, Gasteiger-
Hu¨ckel charges are reported to generate reasonable
models in various CoMFA studies on artemisinin
derivatives.45d,57,58
An analysis of Table 3 clearly shows that substitution
at the C-9â position brings about a significant improve-
ment in the activity compared to artemisinin (1). It is
more pronounced in the case of phenethyl derivatives
with halogens (e.g. 21, 22 and 27), SMe (24), and NMe2
(30) groups on the aromatic ring. Even though the
lactols of this class of compounds were active, the
cytotoxicity exhibited by a few of them was discouraging.
The statistical validity of the models can be judged
by high q2 (more than 0.6) and r2 (more than 0.9) values
along with a low standard error of estimate (less than
0.15). Mean r2 values for bootstrap analysis of the

SAR of the Antimalarial Agent Artemisinin
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4251
would increase activity. As an example, 12 with its
n-butyl side chain at C-9â occupying the G1 region
exhibits better leishmanicidal potency compared to
artemisinin. However, the presence of a yellow region
(Y1, Figure 3) next to the G1 pocket indicates that a
longer side chain at C-9â may actually lead to reduction
in the activity. Thus, the length of the side chain at the
C-9â position appears to be critical for activity. The
second yellow region (Y2, Figure 3) is in proximity to
the C-10 position, while the third (Y3, Figure 3) is near
the C-3 side chain. The presence of Y3 suggests that a
bulky group at C-3 would decrease the activity. Al-
though molecules such as 34 and 61, having a butyl
substituent at C-3 and a relatively lower activity, are
in agreement with this observation, there are several
other molecules such as 36 and 37 that possess higher
activity despite having bulky side chains at C-3. Thus,
no clear relationship can be derived with respect to the
steric effect at the C-3 position. Perhaps, inclusion of
more diverse structures at the C-3 position would help
to arrive at a more clear relationship.
The important blue and red regions in the electro-
static maps are labeled as B1-B5 and R1-R2, respec-
tively, in Figure 4. Again, most of the contributing
regions in terms of electrostatic interactions are related
to substitution at the C-9â position. The B1 and B2
regions suggest that a positive charge at the terminal
end of a C-9â side chain should improve in vitro activity.
This is exemplified by a lower activity of 25 as compared
to that of 30. In the former, a negatively charged
sulfonyl group occupies B1 and B2 pockets, which is
replaced by positively charged quaternary nitrogen in
the latter.
F igu r e 2. Plot of predicted vs actual activities of compounds
in the training set using CoMFA model 2.
Ta ble 6. Residuals for Prediction of Test Set Molecules Using
Model 2
a
a
molecule residual (pIC50
)
molecule residual (pIC50)
14
19
21
24
26
29
-0.21
0.01
0.24
0.60
-0.13
-0.27
33
36
50
57
58
0.30
0.31
0.12
-0.35
-0.32
a
Predictive r2 ) 0.675.
It is interesting to note that 10-deoxo derivatives show
improved leishmanicidal activity. For example, com-
pounds such as artemisinin, 22, 27, and 34 have a
relatively lower potency compared to corresponding 10-
deoxo derivatives 51, 53, 59, and 61, respectively. This
may be explained by the presence of B4 close to the C-10
carbonyl oxygen, which indicates that a partial negative
charge in this vicinity would lead to a decrease in
antileishmanial activity.
models spanned a similar range as that of the r2 values
obtained by non-cross-validated analysis (Table 5). This
provided additional validation for the models.
Among the nine statistically significant models, model
2 with the highest q2 value was utilized for further
analysis. The plot of predicted vs actual activity for the
training set molecules (Figure 2) reveals an excellent
positive correlation between the two activities. To assess
the predictive capability of model 2, it was used to
predict the activity of a diverse test set of artemisinin
derivatives. The result of this prediction is presented
in Table 6. The activities predicted by the model are
highly consistent with the experimental data. All the
residual values for the prediction are less than 1 log
order of activity, the highest being 0.60. The mean
absolute residual value of 0.44 and the predictive r2
value of 0.675 suggest the outstanding predictive power
of the model.
Model 2 has about 68% contribution from steric field
values and 32% from electrostatic field values. The
CoMFA contour plots for the steric field are shown in
Figure 3 and those for the electrostatic field in Figure
4. The green and yellow areas represent regions where
steric bulk is favorable and unfavorable, respectively,
for binding. Analogously, red and blue areas represent
regions where a negative and a positive charge are
favorable, respectively.
The location of R1 and R2 indicates the role of
electron-withdrawing groups (negatively charged) on the
phenyl ring of the C-9â side chains of the 10-deoxo
derivatives. It suggests that negatively charged groups
occupying these regions would lead to increased activity.
For example, the relative order of activities of 53 < 54
< 56 < 59 may be explained by the observation that
the R1 and R2 pockets are occupied by one chloro, two
chloro, one CF3, and two CF3 groups, respectively, in
these molecules. Also, the relatively lower potency of
55 may be attributed to the presence of a quaternary
nitrogen on the terminal phenyl ring in proximity to
these regions.
It is interesting to compare the results of the current
QSAR study of the leishmanicidal activity with that of
a comprehensive QSAR study on the antimalarial
activity of artemisinin derivatives.45d Greater steric bulk
near the C-9â side chain was found to improve the
antimalarial activity of these derivatives. A similar
relationship was observed in the current study on
antileishmanial activity. However, in the case of the
antimalarial QSAR study, potency was found to have a
The steric map shows a green region (G1, Figure 3)
corresponding to the pocket occupied by the C-9â side
chain, suggesting that greater steric bulk in this area

4252 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Avery et al.
F igu r e 3. Steric contour map for CoMFA model 2 with 12 as an example. Green contours indicate areas where steric bulk is
predicted to increase antileishmanial activity, while yellow contours indicate regions where steric bulk is predicted to decrease
activity.
F igu r e 4. Electrostatic contour map for CoMFA model 2 with 59 as an example. Blue polyhedra indicate regions where partial
positive charge is correlated with improved antileishmanial activity, while red polyhedra suggest a relationship between partial
negative charge and activity.
positive correlation with increasing steric bulk near the
C-3 substituent, whereas no clear relationship was
observed between steric bulk at the C-3 position and
antileishmanial activity. With respect to the effect of
electrostatics on activity, removal of C-10 carbonyl was
observed to be favorable for both antimalarial as well
as antileishmanial activity.
As was the case for antimalarial activity, the presence
of the peroxide moiety appears to be essential for
leishmanicidal activity. This is evident from the fact
that 1-deoxyartemisinin derivatives 75-79 (Scheme 5)
do not exhibit leishmanicidal activity. It is also interest-
ing to note the effect of C-9R substitution on both
antimalarial and antileishmanial activity. In general,

SAR of the Antimalarial Agent Artemisinin
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4253
compounds with C-9R substitution appear to have lower
antimalarial (Table 2) as well as antileishmanial activ-
ity (Table 3). This is exemplified by compounds 19 and
31, which show relatively lower potency compared to
the corresponding C-9â-substituted compounds 18 and
30, respectively. This could be due to masking of the
peroxide group by the substituted aromatic ring when
C-9 is in the R orientation. However, it should be noted
that the difference in antileishmanial activity between
the C-9R- and C-9â-substituted compounds is more
pronounced than that for antimalarial activity. Com-
pound 28 with C-9R substitution appears to be an
exception in this case. While it shows no antileishmanial
activity, it has comparable activity to the corresponding
C-9â derivative 27 against P. falciparum.
P ossible Mod e of Action . As many of the artemisi-
nin derivatives studied were notably leishmanicidal, and
the presence of an endoperoxide has been shown to be
necessary for antimalarial action,2 it was logical to study
whether the endoperoxide moiety is a key structural
feature required for the observed antileishmanial activ-
ity. To test this idea, several 1-deoxo derivatives (75-
79) were prepared and tested for their influence on
parasitic growth. None of the nonperoxides showed
activity in the concentration range tested (up to 125 µg/
mL), which clearly emphasizes the role of the peroxide
group for the antileishmanial effect.
against malaria (nanomolar). Suitable structural modi-
fications based on the results of this study are underway
to achieve this goal.
Exp er im en ta l Section
All reactions were carried out under an argon atmosphere
with dry, freshly distilled solvents stored over 4 Å molecular
sieves under anhydrous conditions, unless otherwise stated.
Thin-layer chromatography (TLC) was performed on precoated
silica gel G and GP Uniplates from Analtech. The plates were
visualized with a 254 nm UV light, iodine chamber, or charring
with acid. Flash chromatography was carried out on silica gel
60 [Scientific Adsorbents Incorporated (SAI), particle size 32-
63 µm, pore size 60 Å]. Melting points were noted on a FP62
1
Mettler Toledo apparatus and are uncorrected. H NMR and
13C NMR spectra were recorded on
a Bruker DPX 300
operating at 400 and 100 MHz, respectively. The chemical
shifts are reported in parts per million (ppm) downfield from
tetramethylsilane, and J values are in hertz. IR spectra were
recorded using a Thermo Nicolet IR 300 FT/IR spectrometer
on a germanium crystal plate as neat solids or liquids. Low-
resolution mass spectra were recorded using a Waters Micro-
mass ZQ LC-Mass system or Thermo-Finnigan AQA by direct
injection in ESI + mode. GC/MS data were obtained with a
Hewlett-Packard 5890A gas chromatographic instrument. The
high-resolution mass spectra (HRMS) were recorded on a
Micromass Q-Tof Micro with lock spray source. Optical rota-
tions were measured on an Autopol IV automatic polarimeter
from Rudolph Research Analytical. Elemental analysis data
were obtained with a Perkin-Elmer series II 2400 CHNS/O
analyzer.
As described in the Introduction, the mechanism of
action of artemisinin against P. falciparum is believed
to involve free radical intermediates generated in a
reaction cascade activated by heme. The leishmania
parasite is also known to utilize heme and free iron-
(II)59 for its survival, which points toward a potentially
similar mode of action.
Finally, it should also be noted that the effect of
structural variation on the relative activity appears to
be similar for both antimalarial as well as antileishma-
nial compounds. All these facts together point toward
a common mode of action. However, this hypothesis
requires further experimental investigation, which is
currently being pursued.
Gen er a l P r oced u r e for th e P r ep a r a tion of Su bstitu ted
P h en eth yl Alcoh ols 5a a n d 5b. To a flame-dried, three-neck,
100 mL round-bottomed flask equipped with a condenser and
argon line was added the appropriate carboxylic acid (10 mmol)
dissolved in dry THF (25 mL). After cooling to 0 °C, 1 M BH3‚
THF (20 mmol) was added dropwise to the stirred mixture and
the reaction was then left at 0 °C for 5 h. The reaction mixture
was diluted with cold water, washed with a saturated solution
of NaHCO3, and extracted with ethyl acetate (3 × 25 mL). The
combined organic layers were washed with brine (1 × 25 mL),
dried over MgSO4, filtered, and concentrated under reduced
pressure. The residue was purified by flash chromatography
on silica gel, eluting with 25% ethyl acetate-hexanes to afford
the corresponding alcohols.
p-(N,N-Dim eth yla m in o)p h en eth yl Alcoh ol (5a ). Yield:
1
90%. Mp: 56-57 °C. H NMR (CDCl3): δ 2.77 (t, 2H, J ) 6.6
Hz), 2.92 (s, 6H, NMe2), 3.80 (t, 2H, J ) 6.5 Hz), 6.72 (d, 2H,
J ) 8.53 Hz), 7.10 (d, 2H, J ) 8.51 Hz). 13C NMR (CDCl3): δ
38.21, 40.93 (NMe2), 64.01, 113.20 (2C, Ar), 126.43 129.72 (2C,
Ar), 149.51. IR (neat) cm-1: 3284, 2855, 1617, 1527, 1352,
1049, 1021, 804. HRMS (ESI) m/z: calcd for C10H15NO [M +
H]+ 166.1232, found 166.1225 [M + H]+.
Con clu sion
Potent antileishmanial activity for a number of arte-
misinin analogues has been described for the first time.
A systematic structure-activity relationship for anti-
leishmanial activity of artemisinin analogues was per-
formed using CoMFA. Substitution at C-9â was shown
to improve antileishmanial potency compared to arte-
misinin. This effect was more pronounced with phen-
ethyl substitution. Also, 10-deoxo derivatives exhibited
better activity. A similar trend is already known for the
antimalarial activity of this class of compounds. The
lack of activity in the case of 1-deoxo analogues suggests
that the peroxide moiety is required for both antileish-
manial as well as antimalarial activity. This in turn
points toward a related mode of action for these ana-
logues against L. donovani and P. falciparum.
p-(Meth ylth io)p h en eth yl Alcoh ol (5b). Yield: 94%. Mp:
1
37-38 °C. H NMR (CDCl3): δ 1.82 (s, 1H), 2.48 (s, 3H), 2.83
(t, 2H, J ) 6.4 Hz), 3.83 (t, 2H, J ) 6.8 Hz), 7.16 (d, 2H, J )
8.0 Hz), 7.24 (d, 2H, J ) 8.0 Hz). 13C NMR (CDCl3): δ 16.17,
38.62, 63.57, 127.16 (2C, Ar), 129.57 (2C, Ar), 135.57, 136.20.
IR (neat) cm-1: 3408, 2949, 1495, 1429, 1037, 1021, 808 cm-1
.
GCMS (EI) m/z: calcd for C9H12OS (M+, 168), found 168 [M+].
Tr iisop r op yl Silyl E t h er P r ot ect ion of 4-H yd r oxy-
p h en yla cetic Acid Meth yl Ester (4c). To an ice-cooled,
stirred solution of the phenolic ester 4c (5.0 g, 30 mmol) and
imidazole (2.4 g, 35.25 mmol) in dry DMF (20 mL) in a dried
round-bottomed flask was added dropwise triisopropylsilyl
chloride (6.6 mL, 30.8 mmol), and the solution was stirred for
5 h, slowly warming to room temperature. The mixture was
then washed with water (1 × 25 mL) and extracted with ether
(3 × 30 mL). The combined organic layers were washed with
brine (1 × 25 mL), dried over MgSO4, and filtered. The solvents
were evaporated under reduced pressure. The crude product
was purified by flash chromatography over silica gel, eluting
with 5% ethyl acetate-hexanes to afford 9.5 g of the product
Thus, this study suggests the possibility of developing
artemisinin analogues as potential drug candidates
against both malaria and leishmaniasis. However,
considering bioavailability and potential toxicity issues,
it would be desirable to bring their antileishmanial
activity profile (currently micromolar) at par with that
1
4c′ as an oily liquid. Yield: 98%. H NMR (CDCl3): δ 1.13 (d,

4254 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Avery et al.
18H, J ) 7.2 Hz), 1.28 (m, 3H), 3.54 (s, 2H), 3.70 (s, 3H), 6.85
(d, 2H, J ) 8.4 Hz), 7.15 (d, 2H, J ) 8.4 Hz). 13C NMR
(CDCl3): δ 12.68, 17.91, 40.36, 51.90, 119.89, 126.34, 130.17,
155.17, 172.35. IR (neat) cm-1: 2945, 2868, 1740, 1511, 1266,
1156, 915, 686. GCMS (EI) m/z: calcd for C18H30O3Si (M+),
found 322 [M+].
Gen er a l P r oced u r e for th e Ra d ica l-In d u ced Mich a el
Ad d it ion of Va r iou s Ar yl Alk yl or Alk yl Br om id es t o
Ar tem isiten e. To a stirred solution of artemisitene46a (1
mmol) in dry benzene (30 mL) in a two-necked 100 mL round-
bottom flask equipped with a condenser and an argon line was
added a bromide derivative (6a -6d , 1.3 mmol) and AIBN (0.1
mmol). The reaction mixture was heated to reflux, and a
solution of tributyltin hydride (1.4 mmol) in benzene (20 mL)
was added to it via a syringe pump over a period of 8 h. After
the addition was completed, the reaction mixture was refluxed
for another 1 h. The flask was then cooled to room tempera-
ture, and the solvent was evaporated under reduced pressure.
The residue was diluted with ether (15 mL), and a saturated
solution of potassium fluoride (5 mL) was added. The mixture
was stirred at room temperature for 12 h and filtered and the
filtrate washed with water. The organic layer was dried over
anhydrous MgSO4 and filtered, and the combined solvents
were evaporated under reduced pressure to give a crude
residue, which was purified by flash column chromatography
on silica gel using 15% ethyl acetate-hexanes to afford the R-
and â-isomers. The R-isomer could be converted to the â-form
by refluxing with DBU in THF for 12 h.
P r ep a r a tion of 4-(Tr iisop r op ylsilyloxy)p h en yleth yl
Alcoh ol (5c). To an ice-cooled, stirred suspension of lithium
aluminum hydride (0.37 g, 9.6 mmol) in dry ether (12 mL) in
50 mL round-bottom flask was added dropwise a solution of
4c′ (1.55 g, 4.8 mmol) in ether (15 mL), and the mixture was
stirred for 1 h, the reaction being monitored by TLC until
completion. A saturated solution of sodium sulfate (10 mL) was
added to this dropwise at 0 °C. The solids were filtered and
washed with ether (3 × 30 mL). The organic layers were
combined, dried over MgSO4, and filtered. The combined
solvents were removed in vacuo. The residue was purified by
flash chromatography on a silica gel column using 15% ethyl
acetate-hexanes to afford 1.1 g of 5c as an oily liquid. Yield:
83%. 1H NMR (CDCl3): δ 1.1 (d, 18H, J ) 7.0 Hz), 1.25 (m,
3H), 2.79 (t, 2H, J ) 6.6 Hz), 3.81 (t, 2H, J ) 6.62 Hz), 6.82
(d, 2H, J ) 8.35 Hz), 7.05 (d, 2H, J ) 8.34 Hz). 13C NMR
(CDCl3): δ 12.69, 17.94, 38.40, 63.80, 119.94 (2C, Ar), 129.88
(2C, Ar), 130.71, 154.67. IR (neat) cm-1: 3330, 2953, 2868,
1605, 1511, 1266, 1046, 914. GCMS (m/ z): calcd for C17H30O2-
Si (M+, 294), found 294 [M+].
(+)-Octa h yd r o-3,6r-d im eth yl-3,12-ep oxy-9â-(2′-cya n o-
eth yl)-12H-pyr an o[4,3j]-1,2-ben zodioxepin -10(3H)-on e (15).
Yield: 14%. Mp: 171-172 °C. [R]24 +120.0 (CHCl3, c 0.5).
D
1H NMR (CDCl3): δ 1.0 (d, 3H, J ) 5.55 Hz), 1.12 (m, 2H),
1.42 (m, 3H), 1.45 (s, 3H), 1.61-1.64 (m, 2H), 1.79-1.88 (m,
2H), 2.01-2.10 (m, 2H), 2.20-3.30 (m, 1H), 2.40-2.48 (m, 1H),
2.57-2.63 (m, 1H), 2.77-2.85 (m, 1H), 3.30 (m, 1H), 5.87 (s,
1H). 13C NMR (CDCl3): δ 16.37, 19.76, 23.60, 24.58, 24.78,
25.12, 33.35, 35.81, 37.38, 37.45, 44.36, 49.89, 79.05, 93.87,
105.60, 119.27, 170.64. IR (neat) cm-1: 2933, 2864, 1732, 1442,
1384, 1115, 984. HRMS (ESI) m/z: calcd for C17H23NO5Na [M
P r ep a r a tion of p-(N,N-Dim eth yla m in o)p h en eth yl Br o-
m id e (6a ). To a dried 10 mL round-bottomed flash under
argon at 0 °C was added 5a (0.1 g, 0.6 mmol), CH2Cl2 (2 mL),
and carbon tetrabromide (0.3 g, 0.9 mmol), followed by
dropwise addition of a solution of triphenylphosphine (0.16 g,
0.6 mmol) in dry CH2Cl2 (1 mL). The mixture was warmed to
room temperature after 2 h, the reaction being monitored by
TLC for completion. The reaction mixture was then washed
with water (1 × 10 mL), and the aqueous layer was extracted
with CH2Cl2 (2 × 5 mL). The combined organic layers were
dried over MgSO4 and filtered. The solvents were evaporated
under reduced pressure. The residue was purified by flash
chromatography on silica gel (∼20 g), eluting with hexanes to
afford 0.86 g of 6a . Yield: 63%. Mp: 45-46 °C. 1H NMR
(CDCl3): δ 2.93 (s, 6H), 3.07 (t, 2H, J ) 7.92 Hz), 3.51 (t, 2H,
J ) 7.7 Hz), 6.70 (d, 2H, J ) 8.56 Hz), 7.09 (d, 2H, J ) 8.50
Hz). 13C NMR (CDCl3): δ 33.83, 38.74, 40.75 (NMe2), 112.82
(2C, Ar), 126.87, 129.41 (2C, Ar), 149.67. IR (neat) cm-1: 2900,
+ Na]+ 344.1474, found 344.1492 [M + Na]+. Anal. (C17H23
NO5): C, H.
-
(+)-Octa h yd r o-3,6r-d im eth yl-3,12-ep oxy-9r-(3′-p h en yl-
p r op yl)-12H -p yr a n o[4,3j]-1,2-b e n zod ioxe p in -10(3H )-
on e (19). This analogue was prepared by radical-induced
Michael addition of phenethyl iodide to artemisitene according
to the general procedure described above. Yield: 30%. [R]26
D
+92.0 (CHCl3, c 0.5). 1H NMR (CDCl3): δ 0.98 (d, 3H, J )
5.64 Hz), 1.01 (m, 1H), 1.40 (m, 4H), 1.45 (s, 3H), 1.75 (m, 6H),
2.08 (m, 4H), 2.40 (m, 1H), 2.62 (t, 2H, J ) 7.3 Hz), 5.90 (s,
1H), 7.17-7.29 (m, 5H). 13C NMR (CDCl3): δ 19.91, 24.75,
25.50, 29.70, 31.66, 34.02, 34.12, 35.87, 35.95, 37.57, 42.81,
45.06, 50.54, 80.31, 93.73, 105.29, 125.81, 128.36 (2C, Ar),
128.41 (2C, Ar), 142.22, 171.88. IR (neat) cm-1: 2929, 2851,
1740, 1458, 1380, 1106, 996. HRMS (ESI) m/ z: calcd for
1613, 1519, 1352, 1204, 808. HRMS (ESI) m/z: calcd for C10H14
-
BrN [M + H]+ 228.0388, found 228.0380 [M + H]+.
P r ep a r a tion p-(Meth ylth io)p h en eth yl Br om id e (6b)
a n d 4-(Tr iisop r op ylsilyloxy)p h en eth yl Br om id e (6c). In
a dried 25 mL round-bottomed flask under argon was taken a
solution of the appropriate alcohol (1 mmol) in ether/acetoni-
trile (3:1, 10 mL). Triphenylphosphine (3 equiv) was added to
this solution, followed by imidazole (3 equiv) and then bromine
(3 equiv), and the mixture was stirred at room temperature
for 2 h, the progress being monitored by TLC. The reaction
mixture was filtered and washed with ether (2 × 15 mL). The
organic layers were combined, washed with brine (1 × 25 mL),
dried over MgSO4, and filtered. The combined solvents were
evaporated under reduced pressure. The residue was purified
by flash chromatography on silica gel, eluting with 1% ethyl
acetate-hexanes to afford the corresponding bromides.
C
23H30O5Na [M + Na]+ 409.1991, found 409.1978 [M + Na]+.
Anal. (C23H30O5‚H2O): C, 68.64; H, 7.92.
(+)-Oc t a h yd r o-3,6r-d im e t h y l-3,12-e p oxy -9â-(3′-(p -
m et h ylt h iop h en yl)p r op yl)-12H -p yr a n o[4,3j]-1,2-b en zo-
d ioxep in -10(3H)-on e (24). Yield: 17%. Mp: 164-165 °C.
[R]25 +54.0 (CHCl3, c 0.5). 1H NMR (CDCl3): δ 0.99 (d, 3H, J
D
) 5.8 Hz), 1.04 (m, 2H), 1.38 (m, 3H), 1.44 (s, 3H), 1.56 (m,
3H), 1.74 (m, 3H), 2.05 (m, 3H), 2.41 (m, 1H), 2.46 (s, 3H),
2.57 (m, 1H), 2.63 (m, 1H), 3.21 (m, 1H), 5.83 (s, 1H), 7.10 (d,
2H, J ) 8.10 Hz), 7.20 (d, 2H, J ) 8.11 Hz). 13C NMR
(CDCl3): δ 16.69, 20.22, 23.66, 25.26, 25.58, 26.62, 29.20, 33.97,
35.52, 36.32, 37.89, 38.07, 43.27, 50.46, 79.60, 93.82, 105.74,
127.52 (2C, Ar), 129.33 (2C, Ar), 135.73, 139.44, 171.80. IR
(neat) cm-1: 2926, 2866, 1739, 1189, 1109, 1000. ESI MS m/z:
calcd for C24H32O5S [M + H]+ 433.2048, found 433.3 [M + H]+,
865.4 [2M + H]+, 887.4 [2M + Na]+. Anal. (C24H32O5S‚
0.5H2O): C, H.
6b. Yield: 92%. 1H NMR (CDCl3): δ 2.48 (s, 3H), 3.12 (t,
2H, J ) 7.6 Hz), 3.54 (t, 2H, J ) 7.6 Hz), 7.14 (d, 2H, J ) 8.11
Hz), 7.22 (d, 2H, J ) 8.12 Hz). 13C NMR (CDCl3): δ 16.42,
33.47, 39.25, 127.35 (2C, Ar), 129.62 (2C, Ar), 136.18, 137.40.
IR (neat) cm-1: 2921, 1495, 1437, 1095, 809. GCMS (EI) m/z:
calcd for C9H11BrS (M+, 230), found 230 [M+].
(+)-Oct a h yd r o-3,6r-d im et h yl-3,12-ep oxy-9â-(3′-(p -t r i-
isop r op ylsilyloxyp h en yl)p r op yl)-12H -p yr a n o[4,3j]-1,2-
1
ben zod ioxep in -10(3H)-on e (26b). Yield: 22%. [R]25 +84.0
6c. Yield: 75%. H NMR (CDCl3): δ 1.12 (d, 18H, J ) 7.6
D
1
Hz), 1.27 (m, 3H), 3.09 (t, 2H, J ) 8.0 Hz), 3.53 (t, 2H, J ) 7.6
Hz), 6.84 (d, 2H, J ) 8.40 Hz), 7.06 (d, 2H, J ) 8.40 Hz). 13C
NMR (CDCl3): δ 12.69 (3C, CHs), 17.95 (6 × CH3), 33.28,
38.82, 119.99 (2C, Ar), 129.58 (2C, Ar), 131.40, 155.02. IR
(neat) cm-1: 2946, 2864, 1736, 1511, 1266, 914. GCMS (EI)
m/z: calcd for C17H29BrOSi (M+, 356), found 356 [M+].
(CHCl3, c 0.5). H NMR (CDCl3): δ 0.98 (d, 3H, J ) 5.27 Hz),
1.02 (m, 2H), 1.08 (d, 18H, J ) 6.98 Hz), 1.22 (m, 6H), 1.38 (s,
3H), 1.70 (m, 6H), 2.04 (m, 3H), 2.50 (m, 3H), 3.21 (m, 1H),
5.82 (s, 1H), 6.77 (d, 2H, J ) 8.17 Hz), 7.0 (d, 2H, J ) 8.23
Hz). 13C NMR (CDCl3): δ 13.02 (3 × CH), 18.33 (6 × CH3),
20.20, 23.58, 25.25, 25.55, 26.49, 29.37, 33.97, 35.23, 36.31,

SAR of the Antimalarial Agent Artemisinin
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4255
37.88, 38.04, 43.13, 50.46, 79.58, 93.77, 105.68, 120.06 (2C,
(m, 1H), 2.75 (m, 2H), 3.04 (s, 3H), 3.21 (m, 1H), 5.83 (s, 1H),
7.37 (d, 2H, J ) 8.07 Hz), 7.84 (d, 2H, J ) 8.03 Hz). 13C NMR
(CDCl3): δ 14.47, 20.13, 22.96, 23.65, 25.16, 25.50, 26.78, 28.90,
31.90, 33.86, 35.98, 36.24, 37.78, 38.07, 43.46, 44.90, 50.37,
79.57, 93.85, 105.70, 127.82 (2C, Ar), 129.70 (2C, Ar), 138.54,
148.91, 171.66. IR (KBr) cm-1: 3017, 2926, 2869, 1735, 1302,
1191, 1146, 1110, 998. ESI MS m/z: calcd for C24H32O7S [M +
H]+ 465.1947, found 465.2 [M + H]+, 929.3 [2M + H]+, 951.3
[2M + Na]+. Anal. (C24H32O7S): C, H.
Ar), 129.53 (2C, Ar), 134.70, 154.43, 171.81. IR (neat) cm-1
:
2917, 2860, 1736, 1515, 1462, 1266, 1101, 1001, 914, 882, 681.
ESI MS m/z: calcd For C32H50O6Si [M + H]+ 559.3455, found
559.5 [M + H]+, 1116.9 [2M + H]+.
(+)-Octa h yd r o-3,6r-d im eth yl-3,12-ep oxy-9r-(3′-(3,5-bis-
(t r iflu or om et h yl)p h en yl)p r op yl)-12H -p yr a n o[4,3j]-1,2-
ben zod ioxep in -10(3H)-on e (28). This analogue was synthe-
sized as reported previously.46a Yield: 43%. [R]26 +92.0
D
(CHCl3, c 0.5). 1H NMR (CDCl3): δ 1.0 (d, 3H, J ) 5.6 Hz),
1.13 (m, 1H), 1.46 (s, 3H), 1.47-1.52 (m, 4H), 1.74 (m, 3H),
1.83 (m, 3H), 1.98 (m, 1H), 2.05-2.18 (m, 3H), 2.39 (m 1H),
2.78 (t, 2H, J ) 7.6 Hz), 5.92 (s, 1H), 7.65 (s, 2H), 7.71 (s, 1H).
13C NMR (CDCl3): δ 19.83, 24.70, 25.40, 29.13, 31.57, 33.86,
33.95, 35.52, 35.90, 37.52, 43.12, 44.95, 50.51, 80.13, 93.80,
105.33, 119.96, 122.08, 124.79, 128.57, 131.37, 131.70, 144.51
(2C), 171.51. IR (neat) cm-1: 2925, 2868, 1740, 1381, 1279,
1168, 1136, 1001, 890. HRMS (ESI) m/z: calcd for C25H28F6O5-
Na [M + Na]+ 545.1739, found 545.1740 [M + Na]+, 561.1382
[M + K]+. Anal. (C25H28F6O5): C, H.
(+)-Oct a h yd r o-3,6r-d im et h yl-3,12-ep oxy-9â-(3′-(p -h y-
dr oxyph en yl)pr opyl)-12H-pyr an o[4,3j]-1,2-ben zodioxepin -
10(3H)-on e (26). To a stirred solution of 26b (0.76 g, 1.36
mmol) in dry THF (12 mL) was added tetrabutylammonium
fluoride (0.34 mL, 1.0 M solution in THF) dropwise and the
mixture was stirred at room temperature for 2 h, the progress
of the reaction being monitored by TLC. The solvent was
evaporated under reduced pressure, and the residue was
purified by flash column chromatography on silica gel using
15% EtOAc-hexanes as the eluent to afford 0.5 g of 26.
1
Yield: 91%. Mp: 116-117 °C. [R]25 +78.0 (CHCl3, c 1.0). H
D
NMR (CDCl3): δ 0.98 (d, 3H, J ) 5.56 Hz), 1.03 (m, 2H), 1.34
(m, 4H), 1.43 (s, 3H), 1.45-1.79 (m, 6H), 2.05 (m, 3H), 2.40-
2.70 (m, 3H), 3.21 (m, 1H), 5.83 (s, 1H), 6.74 (d, 2H, J ) 8.42
Hz), 7.03 (d, 2H, J ) 8.31 Hz). 13C NMR (CDCl3): δ 20.19,
23.64, 25.24, 25.54, 26.51, 29.46, 33.94, 35.16, 36.31, 37.89,
38.14, 43.12, 50.43, 79.68, 94.03, 105.86, 115.74 (2C, Ar),
129.70 (2C, Ar), 133.92, 154.57, 172.60. IR (KBr) cm-1: 3386,
2928, 2872, 1733, 1715, 1613, 1514, 1445, 1378, 1202, 1113,
1000. ESI MS (m/z) calcd for C23H30O6 [M + H]+ 403.2, found
403.2 [M + H]+, 805.6 [2M + H]+, 827.5 [2M + Na]+. Anal.
(C23H30O6‚0.5H2O): C, H.
(+)-Octa h yd r o-3,6r-d im eth yl-3,12-ep oxy-9â-(3′-(p-N,N-
d im eth yla m in op h en yl)p r op yl)-12H-p yr a n o[4,3j]-1,2-ben -
zod ioxep in -10(3H)-on e (30a ). Yield: 30%. Mp: 167-169 °C.
[R]26 +68.0 (CHCl3, c 1.0). 1H NMR (CDCl3): δ 0.98 (d, 3H, J
D
) 5.73 Hz), 1.03 (m, 1H), 1.38 (m, 4H), 1.44 (s, 3H), 1.59-1.80
(m, 6H), 2.01-2.20 (m, 3H), 2.41-2.70 (m, 3H), 2.90 (s, 6H),
3.21 (ddd, 1H, J ) 5.33, 1.7, 5.2 Hz), 5.83 (s, 1H), 6.69 (d, 2H,
J ) 8.61 Hz), 7.05 (d, 2H, J ) 8.52 Hz). 13C NMR (CDCl3): δ
20.23, 23.65, 25.29, 25.59, 26.52, 29.44, 34.03, 35.02, 36.35,
37.96, 38.05, 41.32 (2C), 43.10, 50.54, 79.62, 93.80, 105.74,
113.45 (2C, Ar), 129.32 (2C, Ar), 130.58, 149.51, 171.93. IR
(neat) cm-1: 2921, 1748, 1613, 1523, 1180, 1119, 1033, 996,
804. HRMS (ESI) m/z: calcd for C25H35NO5 [M + H]+ 430.2593,
(+)-Octa h yd r o-3,6r-d im eth yl-3,12-ep oxy-9â-(3′-(p-N,N-
d im eth yla m in op h en yl)p r op yl)-12H-p yr a n o[4,3j]-1,2-ben -
zod ioxep in (55a ). To a stirred solution of 30a (0.41 g, 0.953
mmol) in dry CH2Cl2 (5 mL) at -78 °C was added 1 M
diisobutylaluminum hydride in CH2Cl2 (1.1 mL, 1.1 equiv), and
the reaction mixture was stirred for 2 h. It was then quenched
with saturated NaHCO3 (5 mL), diluted with CH2Cl2 (10 mL),
and allowed to warm to room temperature. The mixture was
then washed with 10% HCl/saturated NH4Cl (1:15 v/ v). The
organic layer was dried over Na2SO4 and filtered, and the
combined organic solvents were concentrated under reduced
pressure to afford a residue, which was chromatographed on
a short silica gel column using CH2Cl2-EtOAc (3:1) solvent
system to yield 0.333 g (82%) of the lactol as an isomeric
mixture, which was used for the next reduction.
found 430.2587 [M + H]+, 452.2424 [M + Na]+. Anal. (C25H35
NO5): C, H, N.
-
30 (Hydrochloride salt of 30a ). 1H NMR (CDCl3): δ 0.98 (d,
3H, J ) 5.52 Hz), 1.04 (m, 1H), 1.36 (m, 2H), 1.42 (s, 3H),
1.70 (m, 1H), 1.75 (m, 6H), 1.98 (m, 4H), 2.40 (m, 1H), 2.67
(m, 2H), 3.14 (s, 6H), 3.18 (m, 1H), 5.83 (s, 1H), 7.29 (d, 2H, J
) 8.16 Hz), 7.64 (d, 2H, J ) 8.04 Hz). IR (neat) cm-1: 2933,
1736, 1519, 1454, 1376, 1196, 1119, 1037, 1000.
(+)-Octa h yd r o-3,6r-d im eth yl-3,12-ep oxy-9r-(3′-(p-N,N-
d im eth yla m in op h en yl)p r op yl)-12H-p yr a n o[4,3j]-1,2-ben -
zod ioxep in -10(3H)-on e (31). Yield: 32%. Mp: 145-147 °C.
[R]26 +88.0 (CHCl3, c 0.5). 1H NMR (CDCl3): δ 0.98 (d, 3H, J
D
) 5.76 Hz), 1.37-1.44 (m, 4H), 1.45 (s, 3H), 1.64-1.77 (m, 7H),
2.04-2.13 (m, 4H), 2.37 (m, 1H), 2.53 (m, 2H), 2.90 (s, 6H),
5.89 (s, 1H), 6.69 (d, 2H, J ) 8.64 Hz), 7.06 (d, 2H, J ) 8.55
Hz). 13C NMR (CDCl3): δ 19.96, 24.76, 25.54, 30.12, 31.66,
34.0, 34.12, 34.84, 35.95, 37.56, 40.99 (NMe2), 42.63, 45.06,
50.51, 80.39, 93.72, 105.27, 113.07 (2C, Ar), 128.98 (2C, Ar),
130.51, 149.06, 172.05. IR (neat) cm-1: 2917, 2851, 1732, 1621,
1523, 1343, 1209, 1168, 1102. LCMS (ESI) m/z: calcd for
To a stirred solution of the lactol (30b) (0.33 g, 0.77 mmol)
in dry CH2Cl2 (6 mL) at -78 °C was added triethylsilane (0.49
mL, 3.06 mmol) and the mixture stirred for 10 min. To this
was added BF3‚OEt2 (0.5 mL, 1.15 mmol) and the mixture
stirred at that temperature for 7 h. It was then quenched with
pyridine and allowed to warm to room temperature. This
mixture was then poured into aqueous saturated NH4Cl (10
mL) and extracted with EtOAc (3 × 25 mL). The combined
organic layers were dried over anhydrous Na2SO4 and filtered.
The solvents were evaporated in vacuo. The residue was
chromatographed over silica gel using 15% EtOAc-hexanes
C
25H35NO5 [M + H]+ 430, found 430 [M + H]+, 447 [M +
NH4]+. Anal. (C25H35NO5‚EtOAc): C, 66.65; H, 8.20; N, 2.70.
(+)-Octah ydr o-3,6r-dim eth yl-3,12-epoxy-9â-(3′-(p-m eth -
a n esu lfon ylp h en yl)p r op yl)-12H-p yr a n o[4,3j]-1,2-ben zo-
d ioxep in -10(3H)-on e (25). To a stirred solution of 24 (0.42
g, 0.96 mmol) in dry CH2Cl2 (40 mL) at -78 °C was added a
solution of m-CPBA (0.49 g, 77%, 2.2 mmol) in CH2Cl2 (40 mL)
over a period of 2.5 h. The reaction mixture was stirred at that
temperature for another 1.5 h. To this mixture was added
water (10 mL), followed by solid NaHCO3 (7.0 g). The reaction
mixture was allowed to warm to room temperature, and the
product was extracted with CH2Cl2 (3 × 30 mL). The combined
organic layers were washed sequentially with 7% NaHCO3
solution, water (1 × 25 mL), brine (1 × 25 mL), and dried over
anhydrous Na2SO4. The solvents were removed under reduced
pressure to afford the crude product, which was purified by
flash column chromatography on silica gel using 3-4% EtOAc-
to afford 0.11 g of 55a . Yield: 35%. Mp: 131-132 °C. [R]26
D
+78.0 (CHCl3, c 0.5). 1H NMR (CDCl3): δ 0.95 (d, 3H, J )
6.13 Hz), 0.96-1.40 (m, 4H), 1.42 (s, 3H), 1.44-1.74 (m, 8H),
1.80-1.95 (m, 1H), 1.96-2.10 (m, 1H), 2.30-2.60(m, 4H), 2.91
(s, 6H), 3.42 (dd, 1H, J ) 11.69, 11.70 Hz), 3.79 (dd, 1H, J )
3.78, 3.95 Hz), 5.19 (s, 1H), 6.68 (d, 2H, J ) 8.53 Hz), 7.03 (d,
2H, J ) 8.52 Hz). 13C NMR (CDCl3): δ 20.71, 21.31, 25.16,
26.52, 27.63, 29.02, 33.34, 34.47, 35.25, 36.66, 37.74, 41.33 (2C,
NMe2), 43.67, 52.65, 65.66, 81.10, 92.82, 104.56, 113.40 (2C,
Ar), 129.30 (2C, Ar), 130.78, 150.0. IR (neat) cm-1: 2933, 1621,
1523, 1454, 1348, 1135, 1098, 1061, 874, 808. HRMS (ESI)
m/z: calcd for C25H37NO4 [M + H]+ 416.2801, found 416.2812
[M + H]+, 438.2639 [M + Na]+. Anal. (C25H37NO4): C, H, N.
CH2Cl2 as the eluent to yield 0.42 g of 25. Yield: 94%. Mp:
55 (hydrochloride). 1H NMR (CDCl3): δ 0.89 (d, 3H, J )
6.12 Hz), 0.98-1.29 (m, 3H), 1.35 (s, 3H), 1.38-1.66 (m, 9H),
1.80 (m, 1H), 1.84-1.97 (m, 1H), 2.27 (dd, 1H, J ) 3.20, 3.70
Hz), 2.38 (m, 1H), 2.55 (t, 2H, J ) 7.23 Hz), 3.12 (s, 6H), 3.39
1
157-158 °C. [R]25 +59.0 (CHCl3, c 1.0). H NMR (CDCl3): δ
D
0.99 (d, 3H, J ) 5.46 Hz), 1.08 (m, 2H), 1.30 (m, 4H), 1.39 (s,
3H), 1.57 (d, 1H, J ) 4.7 Hz), 1.74 (m, 4H), 2.02 (m, 3H), 2.41

4256 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Avery et al.
(dd, 1H, J ) 5.83, 11.64 Hz), 3.70 (dd, 1H, J ) 3.48, 3.57 Hz),
5.13 (s, 1H), 7.22 (d, 2H, J ) 7.7 Hz), 7.63 (d, 2H, J ) 7.86
Hz). 13C NMR (CDCl3): δ 20.65, 21.28, 25.05, 26.44, 27.50,
28.56, 33.26, 34.33, 35.69, 36.58, 37.64, 43.70, 47.03, 52.55,
53.88, 65.34, 81.00, 92.71, 104.55, 121.02 (2C, Ar), 130.66 (2C,
Ar), 141.10, 144.99, 162.73.
P r ep a r a tion of 1-Deoxo An a logu es 75-79. A solution
of the appropriate artemisinin analogue (0.35 mmol) in ethyl
acetate (6 mL) was mixed with palladium on charcoal (0.02 g
10% w/w) and hydrogenolyzed under a positive pressure of
hydrogen for 7 h, the progress of reaction being monitored by
TLC. The reaction mixture was filtered, the solvents evapo-
rated under reduced pressure, and the product purified by
chromatography on silica gel column, eluting with 10% ethyl
acetate-hexanes to afford the corresponding 1-deoxo com-
pounds. Subsequent preparative reverse-phase chromatogra-
phy on a Waters Prep-6000 system using a 7.8 × 150 mm
symmetry column (C-18) and eluting with H2O:CH3CN (1:1)
provided the highly pure products (>99.9% pure).
for C24H31F3O3Na [M + Na]+ 447.2123, found 447.2113 [M +
Na]+, 463.1857 [M + K]+. Anal. (C24H31F3O3): C, H.
(+)-Octa h ydr o-3,6r-d im eth yl-3,12-ep oxy-9â-(3′-(m-ch lo-
r op h en yl)p r op yl)-12H-p yr a n o[4,3j]-1,2-ben zoxep in (79)
Wa s P r ep a r ed fr om 53. Yield: 47%. [R]26 +30.0 (CHCl3, c
D
0.5). 1H NMR (CDCl3): δ 0.92 (d, 3H, J ) 5.6 Hz), 1.03 (m,
1H), 1.17-1.39 (m, 7H), 1.55 (s, 3H), 1.59-1.64 (m, 2H), 1.71
(m, 3H), 1.85 (m, 1H), 1.95 (m, 1H), 2.13 (m, 1H), 2.60 (t, 2H,
J ) 7.2 Hz), 3.37 (dd, 1H, J ) 4.4, 4.8 Hz), 3.94 (dd, 1H, J )
7.2, 6.8 Hz), 5.27 (s, 1H), 7.06 (d, 1H, J ) 6.8 Hz), 7.20 (m,
3H). 13C NMR (CDCl3): δ 18.87, 22.14, 23.87, 24.03, 29.05,
30.68, 31.66, 34.44, 34.47, 35.43, 35.72, 39.14, 45.78, 62.90,
82.40, 96.44, 107.34, 125.96, 126.58, 128.48, 129.56, 134.05,
144.37. IR (neat) cm-1: 2933, 1724, 1467, 1279, 1123, 1107,
1074, 1001, 882. HRMS (ESI) m/z: calcd for C23H31ClO3K [M
+ K]+ 429.1599, found 413.1873 [M + Na]+, 429.1596 [M +
K]+. Anal. (C23H31ClO3): C, H.
Ack n ow led gm en t. This work was supported by
CDC Cooperative agreements U50/CCU418839 and
UR3/CCU418652. Partial support from USDA under
cooperative agreement 58-6408-2-0009 to NCNPR is also
acknowledged. Authors are thankful to J ohn Trott for
the help in antimalarial testing and Dr. Blake Watkins
for helpful suggestions during the preparation of the
manuscript.
(+)-O c t a h y d r o -3,6r,9â-t r im e t h y l-3,12-e p o x y -12H -
p yr a n o[4,3j]-1,2-ben zoxep in -10(3H)-on e (75) Wa s P r e-
p a r ed fr om Ar tem isin in (1). Yield: 54%. Mp: 107-108 °C.
[R]26 -75.0 (CHCl3, c 1.0). 1H NMR (CDCl3): δ 0.94 (d, 3H, J
D
) 5.33 Hz), 1.01 (m, 1H), 1.19 (d, 3H, J ) 7.21 Hz), 1.25 (m,
4H), 1.52 (s, 3H), 1.59 (m, 1H), 1.76 (m, 2H), 1.92 (m, 2H), 2.0
(m, 1H), 3.19 (m, 1H), 5.69 (s, 1H). 13C NMR (CDCl3): δ 12.64,
18.60, 22.02, 23.53, 23.99, 32.76, 33.45, 33.97, 35.36, 42.41,
44.61, 82.42, 99.65, 109.23, 171.89. IR (neat) cm-1: 2941, 1748,
1384, 1139, 1009, 1001, 874. HRMS (ESI) m/ z: calcd for
15H22O4Na [M + Na]+ 289.1416, found 289.1424 [M + Na]+,
Refer en ces
C
305.1171 [M + K]+. Anal. (C15H22O4): C, H.
(1) Cumming, J . N.; Ploypradith, P.; Posner, G. H. Antimalarial
activity of artemisinin (qinghaosu) and related trioxanes: Mech-
anism(s) of Action. Adv. Pharmacol. (San Diego) 1997, 37, 253-
297.
(2) Avery, M. A.; Alvim-Gaston, M.; Woolfrey, J . R. Synthesis and
structure-activity relationships of peroxidic antimalarials based
on artemisinin. Adv. Med. Chem. 1999, 4, 125-217.
(3) Klayman, D. L. Qinghaosu (artemisinin): An antimalarial drug
from China. Science 1985, 228, 1049-1055.
(4) Robert, A.; Dechy-Cabaret, O.; Cazelles, J .; Meunier, B. From
mechanistic studies on artemisinin derivatives to new modular
antimalarial drugs. Acc. Chem. Res. 2002, 35, 167-174.
(5) Wu, Y. How might qinghaosu (artemisinin) and related com-
pounds kill the intraerythrocytic malaria parasite? A chemist’s
view. Acc. Chem. Res. 2002, 35, 255-259.
(6) Cazelles, J .; Robert, A.; Meunier, B. Alkylating capacity and
reaction products of antimalarial trioxanes after activation by
a heme model. J . Org. Chem. 2002, 67, 609-619.
(7) Wang, D.-Y.; Wu, Y.-L. A possible antimalarial action mode of
qinghaosu (artemisinin) series compounds. Alkylation of reduced
glutathione by C-centered primary radicals produced from
antimalarial compound qinghaosu and 12-(2,4-dimethoxyphen-
yl)-12-deoxoqinghaosu. J . Chem. Soc. Chem. Commun. 2000, 22,
2193-2194.
(8) Pandey A. V.; Tekwani, B. L.; Sing, R. L.; Chauhan, V. S.
Artemisinin, an endoperoxide antimalarial, disrupts the hemo-
globin catabolism and heme detoxification systems in malarial
parasite. J . Biol. Chem. 1999, 274, 19383-19388.
(9) Wu, W.-M.; Wu, Y.; Wu, Y.-L.; Yao Z.-J .; Zhou, C.-M.; Li, Y.;
Shan, F. Unified mechanistic framework for the Fe(II)-induced
cleavage of qinghaosu and derivatives/analogues. The first spin-
trapping evidence for the previously postulated secondary C-4
radical. J . Am. Chem. Soc. 1998, 120, 3316-3325.
(10) Paitayatat, S.; Tarmchompoo, B.; Thebtatranonth, Y.; Yuthavong,
Y. Correlation of antimalarial activity of artemisinin derivatives
with binding affinity with ferroprotoporphyrin IX. J . Med. Chem.
1997, 40, 633-638.
(11) Haynes, R. K.; Vonwiller, S. C. From qinghao, marvelous herb
of antiquity, to the antimalarial trioxane qinghaosusSome
remarkable new chemistry. Acc. Chem. Res. 1997, 30, 73-79.
(12) (a) Meshnick, S. R.; Yang, Y. Z.; Lima, V.; Cuypers, F.; Kam-
chonwongpaisan, S.; Yuthavong, Y. Iron-dependent free radical
generation from the antimalarial agent artemisinin (qinghaosu).
Antimicrob. Agents Chemother. 1993, 37, 1108-1114. (b) Haynes,
R. K.; Hung-On, P. H.; Voerste, A. Ring opening of artemisinin
(qinghaosu) and dihydroartemisinin and interception of the open
hydroperoxides with formation of N-oxidessA chemical model
for antimalarial action. Tetrahedron Lett. 1999, 40, 4715-4718.
(13) Bhisutthibhan, J .; Pan, X. Q.; Hossler, P. A.; Walker, D. J . The
Plasmodium falciparum translationally controlled tumor protein
homologue and its reaction with the antimalarial drug artemisi-
nin. J . Biol. Chem. 1998, 273, 16192-16198.
(+)-Octa h yd r o-3,6r-d im eth yl-3,12-ep oxy-9â-(3′-(3,5-bis-
(t r iflu or om et h yl)p h en yl)p r op yl)-12H -p yr a n o[4,3j]-1,2-
b en zoxep in -10(3H )-on e (76) w a s p r ep a r ed fr om 27.
1
Yield: 21%. [R]29 -32.0 (CHCl3, c 1.0). H NMR (CDCl3): δ
D
0.93 (d, 3H, J ) 4.97 Hz), 1.05 (m, 1H), 1.27 (m, 4H), 1.53 (s,
3H), 1.55-1.70 (m, 2H), 1.79 (m, 4H), 1.91 (m, 2H), 2.05 (m,
2H), 2.77 (m, 2H), 3.01 (m, 1H), 5.68 (s, 1H), 7.64 (s, 2H), 7.71
(s, 1H). 13C NMR (CDCl3): δ 18.55, 22.01, 23.59, 23.96, 26.75,
28.80, 33.39, 33.97, 35.34, 35.57, 37.68, 40.46, 44.65, 54.34,
82.17, 99.31, 109.34, 120.04 (2C), 122.06, 128.49 (2C), 131.43,
131.75, 144.37, 171.04. IR (neat) cm-1: 2921, 1744, 1384, 1287,
1164, 1127, 1013, 894. HRMS (ESI) m/z: calcd for C25H28F6O4-
Na [M + Na]+ 529.1789, found 529.1769 [M + Na]+, 545.1514
[M + K]+.
(+)-Octah ydr o-3,6r-dim eth yl-3,12-epoxy-9â-(3′-(p-m eth -
a n esu lfon ylp h en yl)p r op yl)-12H-p yr a n o[4,3j]-1,2-ben zox-
ep in -10(3H)-on e (77) Wa s P r ep a r ed fr om 25. Yield: 32%.
Mp: 160-161 °C. [R]25 +38.0 (CHCl3, c 0.5). 1H NMR
D
(CDCl3): δ 0.94 (d, 3H, J ) 4.4 Hz), 1.07 (m, 1H), 1.27-1.36
(m, 4H), 1.53 (s, 3H), 1.59-1.69 (m, 2H), 1.78 (m, 4H), 1.85-
1.93 (m, 2H), 2.04 (m, 2H), 2.75 (m, 2H), 2.99 (ddd, 1H, J )
4.4, 4.4, 4.4 Hz), 3.05 (s, 3H), 5.68 (s, 1H), 7.40 (d, 2H, J ) 8.0
Hz), 7.86 (d, 2H, J ) 8.4 Hz). 13C NMR (CDCl3): 18.56, 21.99,
23.57, 23.97, 26.62, 28.73, 33.39, 33.95, 35.32, 35.77, 37.69,
40.38, 44.60, 44.65, 82.16, 99.28, 109.30, 127.51 (2C, Ar),
129.36 (2C, Ar), 138.14, 148.63, 171.14. IR (neat) cm-1: 2929,
1748, 1392, 1307, 1143, 1106, 1029, 869. HRMS (ESI) m/z:
calcd for C24H32O6SNa [M + Na]+ 471.1817, found 471.1790
[M + Na]+, 487.1535 [M + K]+. Anal. (C24H32O6S‚0.5EtOAc):
C, H.
(+)-Oct a h yd r o-3,6r-d im et h yl-3,12-ep oxy-9â-(3′-(p -t r i-
flu or om eth ylp h en yl)p r op yl)-12H-p yr a n o[4,3j]-1,2-ben z-
oxep in (78) Wa s P r ep a r ed fr om 56. Yield: 42%. Mp: 82-
1
83 °C. [R]26 -7.0 (CHCl3, c 1.0). H NMR (CDCl3): δ 0.91 (d,
D
3H, J ) 5.2 Hz), 1.02 (m, 1H), 1.17-1.39 (m, 6H), 1.54 (s, 3H),
1.57-1.74 (m, 6H), 1.85 (m, 1H), 1.94 (m, 1H), 2.13 (m, 1H),
2.68 (t, 2H, J ) 7.2 Hz), 3.37 (dd, 1H, J ) 4.4, 4.4 Hz), 3.94
(dd, 1H, J ) 7.2, 7.2 Hz), 5.27 (s, 1H), 7.29 (d, 2H, J ) 7.6
Hz), 7.54 (d, 2H, J ) 8.0 Hz). 13C NMR (CDCl3): δ 18.85, 22.12,
23.88, 24.01, 29.02, 30.68, 31.65, 34.43, 34.46, 35.41, 35.87,
39.16, 45.77, 62.83, 82.38, 96.43, 107.34, 123.02, 125.22,
125.72, 127.98, 128.30, 128.66, 146.43. IR (neat) cm-1: 2941,
1323, 1160, 1127, 1066, 1001, 886. HRMS (ESI) m/z: calcd

SAR of the Antimalarial Agent Artemisinin
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4257
(14) Olliaro, P. L.; Haynes, R. K.; Meunier, B.; Yuthavong, Y. Possible
modes of action of the artemisinin-type compounds. Trends
Parasitol. 2001, 17, 122-126.
(15) Grellepois, F.; Chorki, F.; Crousse, B.; Ourevitch, M.; Bonnet-
Delpon, D.; J ean-Pierre, B. Anhydrodihydroartemisinin and its
10-trifluoromethyl analogue: Access to novel D-ring-contracted
artemisinin trifluoromethyl ketones. J . Org. Chem. 2002, 67,
1253-1260.
(34) Kern, P. Leishmaniasis. Antibiot. Chemother. 1981, 30, 203-
223.
(35) De Carvalho, P. B.; Arribas, M. A. D. G.; Ferreira, E. I.
Leishmaniasis: What do we know about its chemotherapy? Rev.
Bras. Cienc. Farm. 2000, 36, 69-96.
(36) Loiseau, P. M.; Bories, C. Recent strategies for the chemotherapy
of visceral leishmaniasis. Curr. Opin. Infect. Dis. 1999, 12, 559-
564.
(16) J ung, M.; Lee, K.; Kendrick, H.; Robinson, B. L.; Croft, S. L.
Synthesis, stability, and antimalarial activity of new hydrolyti-
cally stable and water-soluble (+)-Deoxoartelinic acid. J . Med.
Chem. 2002, 45, 4940-4944.
(17) Posner, G. H.; J eon, H. B.; Ploypradith, P.; Paik, I. H.; Borstnik,
K.; Xie, S.; Shapiro, T. A. Orally active, water-soluble antima-
larial 3-Aryltrioxanes: Short synthesis and preclinical efficacy
testing in rodents. J . Med. Chem. 2002, 45, 3824-3828.
(18) Hindley, S.; Ward, S. A.; Storr, R. C.; Searle, N. L.; Bray, P. G.;
Park, B. K.; Davies, J .; O’Neill, P. M. Mechanism-based design
of parasite-targeted artemisinin derivatives: Synthesis and
antimalarial activity of new diamine containing analogues. J .
Med. Chem. 2002, 45, 1052-1063.
(37) Selzer, P. M.; Pingel, S.; Hsieh, I.; Ugele, B.; Chan, V. J .; Engel,
J . C.; Bogyo, M.; Russell, D. G.; Sakanari, J . A.; McKerrow, J .
H. Cysteine protease inhibitors as chemotherapy: Lessons from
a parasite target. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 11015-
11022.
(38) Moreno, J .; Canavate, C.; Chamizo, C.; Laguna, F.; Alvar, J . HIV-
Leishmania infantum co-infection: Humoral and cellular im-
mune responses to the parasite after chemotherapy. Trans R.
Soc. Trop. Med. Hyg. 2000, 94, 328-332.
(39) Kennedy, M. L.; Cortes-Selva, F.; Perez-Victoria, J . M.; J imenez,
I. A.; Gonzalez, A. G.; Munoz, O. M.; Gamarro, F.; Castanys, S.;
Ravelo, A. G. Chemosensitization of a multidrug-resistant
Leishmania tropica line by new sesquiterpenes from Maytenus
magellanica and Maytenus chubutensis. J . Med. Chem. 2001,
44, 4668-4676.
(40) Berman, J . D.; Grog, I. M. Leishmania mexicana: Chemistry
and biochemistry of sodium stibogluconate (Pentostam). Exp.
Parasitol. 1988, 67, 96-103.
(19) Chorki, F.; Grellepois, F.; Crousse, B.; Hoang, V. D.; Hung, N.
V.; Bonnet-Delpon, D.; Begue, J . P. First synthesis of 10R-
(trifluoromethyl)deoxoartemisinin. Org. Lett. 2002, 4, 757-759.
(20) Liao, X.-B.; Han, J .-Y.; Li, Y. Michael addition of artemisitene.
Tetrahedron Lett. 2001, 42, 2843-2845.
(21) Chorki, F.; Grellepois, F.; Crousse, B.; Ourevitch, M.; Bonnet-
Delpon, D.; Begue, J . P. Fluoro artemisinins: Difluoromethylene
ketones. J . Org. Chem. 2001, 66, 7858-7863.
(41) Brost, P.; Ouellette, M. New mechanisms of drug resistance in
parasitic protozoa. Annu. Rev. Microbiol. 1995, 49, 427-460.
(42) Olliaro, P. L.; Bryceson, A. D. M. Practical progress and new
drugs for changing patterns of leishmaniasis. Parasitol. Today
1993, 9, 323-328.
(43) Thakur, C. P. Comparison of glucose versus fat emulsion in the
preparation of amphotericin B for use in kala-azar. Trans R.
Soc. Trop. Med. Hyg. 1994, 88, 689-699.
(44) Dietz, R.; Fagundes, S. M. S.; Brito, E. F.; Milan, E. P.; Fietosa,
T. F.; Suassuna, F. A. B.; Fonschiffrey, G.; Ksionski, G.; Dember,
J . Treatment of kala-azar in Brazil with Amphocil (amphotericin
B cholesterol dispersion) for 5 days. Trans R. Soc. Trop. Med.
Hyg. 1995, 89, 309-311.
(22) (a) Han, J .; Lee, J . G.; Min, S. S.; Park, S. H.; Angerhofer, C. K.;
Cordell, G. A.; Kim, S. U. Synthesis of new artemisinin ana-
logues from artemisinic acid modified at C-3 and C-13 and their
antimalarial activity. J . Nat. Prod. 2001, 64, 1201-1205. (b)
Ekthawatchai, S.; Kamchonwongpaisan, S.; Kongsaeree, P.;
Tarnchompoo, B.; Thebtaranonth, Y.; Yuthavong, Y. C-16 arte-
misinin derivatives and their antimalarial and cytotoxic activi-
ties: Syntheses of artemisinin monomers, dimers, trimers, and
tetramers by nucleophilic additions to artemisitene. J . Med.
Chem. 2001, 44, 4688-4695.
(23) Denis, M. B.; Davis, T. M. E.; Hewitt, S.; Incardona, S.; Nimol,
K.; Fandeur, T.; Poravuth, Y.; Lim, C.; Socheat, D. Efficacy and
safety of Dihydroartemisinin-Piperaquine (Artekin) in Cambo-
dian children and Adults with uncomplicated Falciparum ma-
laria. Clin. Infect. Dis. 2002, 35, 1469-1476, and references
therein.
(45) (a) Vroman, J . A.; Alvim-Gaston, M.; Avery, M. A. Current
progress in the chemistry, medicinal chemistry and drug design
of artemisinin based antimalarials. Curr. Pharm. Design 1999,
5, 101-138. (b) Vroman, J . A.; Elsohly, H. N.; Avery, M. A.
Conjugate addition of a cyano-Gilman cuprate to an acrylic
acid: Homologation of artemisinic acid and subsequent conver-
sion to 16-butylartemisinin. Synth. Commun. 1998, 28, 1555-
1562. (c) Avery, M. A.; Mehrotra, S.; Bonk, J . D.; Vroman, J . A.;
Goins, D. K.; Miller, R. Structure-activity relationships of the
antimalarial agent artemisinin. 4. Effect of substitution at C-3.
J . Med. Chem. 1996, 39, 2900-2906. (d) Avery, M. A.; Alvim-
Gaston, M.; Rodrigues, C. R.; Barreiro, E. J .; Cohen, F. E.;
Sabnis, Y. A.; Woolfrey, J . R. Structure-activity relationships
of the antimalarial agent artemisinin. 6. The development of
predictive in vitro potency models using CoMFA and HQSAR
methodologies. J . Med. Chem. 2002, 45, 292-303.
(46) (a) Avery, M. A.; Alvim-Gaston, M.; Vroman, J . A.; Wu, B.; Ager,
A.; Peters, W.; Robinson, B. L.; Charman, W. Structure-activity
relationships of the antimalarial agent artemisinin. 7. Direct
modification of (+)-artemisinin and in vivo antimalarial screen-
ing of new, potential preclinical antimalarial candidates. J . Med.
Chem. 2002, 45, 4321-4335. (b) Avery, M. A.; Gao, F.; Chong,
W. K. M.; Mehrotra, Sanjiv; Milhous, W. K., Structure-activity
relationships of the antimalarial agent artemisinin. 1. Synthesis
and comparative molecular field analysis of C-9 analogues of
artemisinin and 10-deoxoartemisinin. J . Med. Chem. 1993, 36,
26, 4264-75. (c) Avery, M. A.; Mehrotra, S.; Theresa, L. J .; Bonk,
J . D.; Vroman, J . A.; Miller R. Structure-activity relationships
of the antimalarial agent artemisinin. 5. Analogues of 10-
deoxoartemisinin substituted at C-3 and C-9. J . Med. Chem.
1996, 39, 4149-4155. (d) Avery, M. A.; Bonk, J . D.; Chong, W.
K. M.; Mehrotra, S.; Miller, R.; Milhous, W.; Goins, D. K.;
Venkatesan, S.; Wyandt, C.; Khan, I.; Avery, B. A. Structure-
activity relationships of the antimalarial agent artemisinin. 2.
Effect of heteroatom substitution at O-11: Synthesis and bio-
assay of N-alkyl-11-aza-9-desmethylartemisinins. J . Med. Chem.
1995, 38, 26, 5038-44. (e) Avery, M. A.; Chong, W. K. M.; Detre,
G. Synthesis of (+)-8a,9-secoartemisinin and related analogues.
Tetrahedron Lett. 1990, 31, 13, 1799-802. (f) Avery, M. A.; Fan,
P.; Karle, J ean, M.; Miller, R.; Goins, K. Replacement of the
nonperoxidic trioxane oxygen atom of artemisinin by carbon:
Total synthesis of (+)-13-carbaartemisinin and related struc-
tures. Tetrahedron Lett. 1995, 36, 23, 3965-8. (g) Avery, M. A.;
Chong, W. K. M.; Bupp, J . E. Tricyclic analogue of artemisinin:
Synthesis and antimalarial activity of (+)-4,5-secoartemisinin
and (-)-5-nor-4,5-secoartemisinin. J . Chem. Soc. Chem. Com-
mun. 1990, 21, 1487-9. (h) Avery, M. A.; Chong, W. K. M.; Bupp,
J . Preparation of antimalarial analogues of artemisinin. WO
9114689 A1 19911003, 1991. Chem. Abstr. 117, 48951.
(24) (a) Brossi, A.; Venugopalan, B.; Dominguez Gerpe, L.; Yeh, H.
J . C.; Flippen-Anderson, J . L.; Buchs, P.; Luo, X. D.; Milhous,
W.; Peters, W. Arteether, a new antimalarial drug: Synthesis
and antimalarial properties. J . Med. Chem. 1988, 31, 645-650.
(b) Brewer, T. G.; Grate, S. J .; Peggins, J . O.; Weina, P. J .;
Petras, J . M.; Levine, B. S.; Heiffer, M. H.; Schuster, B. G. Fatal
neurotoxicity of arteether and artemether. Am. J . Trop. Med.
Hyg. 1994, 51, 251-259. (c) Brewer, T. G.; Peggins, J . O.; Grate,
S. J .; Petras, J . M.; Levine, B. S.; Weina, P. J .; Swearengen, J .;
Heiffer, M. H.; Schuster, B. G. Neurotoxicity in animals due to
arteether and artemether. Trans. R. Soc. Trop. Med. Hyg. 1994,
88, 33-36. (d) Lee, I. S.; Hufford, C. D. Metabolism of antima-
larial sesquiterpene lactones. Pharm., Ther. 1990, 48, 345-355.
(25) Singh, N. P.; Lai, H. Selective toxicity of dihydroartemisinin and
holotransferrin toward human breast cancer cells. Life Sci. 2001,
70, 49-56.
(26) Thomas, E.; Heather, D.; Axel, S.; Hayato, M.; Chitambar, C.
R. The anti-malarial artesunate is also active against cancer.
Int. J . Oncol. 2001, 18, 767-773.
(27) Li, Y.; Shan, F.; Wu, J .-M.; Wu, G.-S.; Ding, J .; Xiao, D.; Yang,
W.-Y.; Atassi, G.; Leonce, S.; Caignard, D.-H.; Renard, P. Novel
antitumor artemisinin derivatives targeting G1 phase of the cell
cycle. Bioorg. Med. Chem. Lett. 2001, 11, 5-8.
(28) Posner, G. H.; Ploypradith, P.; Parker, M. H.; O’Dowd, H.; Woo,
S.-H.; Northrop, J .; Krasavin, M.; Dolan, P.; Kensler, T. W.; Xie,
S.; Shapiro, T. A. Antimalarial, antiproliferative, and antitumor
activities of artemisinin-derived, chemically robust, trioxane
dimers. J . Med. Chem. 1999, 42, 4275-4280.
(29) Beekman, A. C.; Wierenga, P. K.; Woerdenbag, H. J .; Van Uden,
W.; Pras, N.; Konings, A. W. T.; El-Feraly, F. S.; Galal, A. M.;
Wikstroem, H. V. Artemisinin-derived sesquiterpene lactones as
potential antitumour compounds: Cytotoxic action against bone
marrow and tumour cells. Planta Med. 1998, 64, 615-619.
(30) J ung, M. Synthesis and cytotoxicity of novel artemisinin ana-
logues. Bioorg. Med. Chem. Lett. 1997, 7, 1091-1094.
(31) Liu, C. H.; Zou, W. X.; Lu, H.; Tan, R. X. Antifungal activity of
Artemisia annua endophyte cultures against phytopathogenic
fungi. J . Biotechnol. 2001, 88, 277-282.
(32) Vikas, D.; Rao, P. S.; Lakshmi, N. M. Antimicrobial activity of
artemisinin and its precursors. Curr. Sci. 2000, 78, 709-713.
(33) Yang, D. M.; Liew, F. Y. Effects of qinghaosu (artemisinin) and
its derivatives on experimental cutaneous leishmaniasis. Para-
sitology 1993, 106, 7-11.

4258 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Avery et al.
(47) Makler, M. T.; Ries, J . M.; Williams, J . A.; Bancroft, J . E.; Piper,
R. C.; Gibbins, B. L.; Hinriches, D. J . Parasite lactate dehydro-
genase as an assay for Plasmodium falciparum drug sensitivity.
Am. J . Trop. Med. Hyg. 1993, 48, 739-741.
(48) (a) Sereno, D.; Roy, G.; Lemesre, J . L.; Papadopoulou, B.;
Ouellette, M. DNA transformation of Leishmania infantum
axenic amastigotes and their use in drug screening. Antimicrob.
Agents Chemother. 2001, 45, 1168-1173. (b) Babich, H.; Boren-
freund, E. Cytotoxicity of T2 toxin and its metabolites with the
neutral red cell viability assay. App. Envt Microbiol. 1991, 57,
2101-2103.
(49) Lisgarten, J . N.; Potter, B. S.; Bantuzeko, C.; Palmer, R. A.
Structure, absolute configuration, and conformation of the
antimalarial compound, artemisinin. J . Chem. Crystallogr. 1998,
28, 539-543.
(50) SYBYL 6.9, Force Field manual; Tripos Inc.: St Louis, MO, p
57.
(54) Cho, S. J .; Tropsha, A. Cross-validated r2-guided region selection
for comparative molecular field analysis: A simple method to
achieve consistent results. J . Med. Chem. 1995, 38, 1060-1066.
(55) Tropsha, A.; Cho, S. J . Cross-validated r2 guided region selection
for CoMFA studies. In 3D QSAR methodology: CoMFA and
related approaches; Kubinyi, H., Folkers, G., Martin, Y. C., Eds.;
Kluwer Academic Publishers: Dordrecht, The Netherlands,
1998; Vol. III, pp 57-69.
(56) Cramer, R. D., III; Bunce, J . D.; Patterson, D. E. Crossvalidation,
bootstrapping, and partial least squares Compared with multiple
regression in conventional QSAR studies. Quant. Struct.-Act.
Relat. 1988, 7, 18-25.
(57) Woolfrey, J . R.; Avery, M. A.; Doweyko, A. M. Comparison of
3D quantitative structure-activity relationship models: Analy-
sis of the in vitro antimalarial activity of 154 artemisinin
analogues by hypothetical active-site lattice and comparative
molecular field analysis. J . Comput-Aided. Mol. Des. 1998, 12,
165-181.
(51) Gasteiger, J .; Marsili, M. Iterative partial equalization of orbital
electronegativity: A rapid access to atomic charges. Tetrahedron
1980, 36, 3219-3222.
(52) Stewart, J . J . P. MOPAC 6.0: A general molecular orbital
package (QCPE 455); Frank J . Seiler Research Laboratory:
Colorado Springs, CO.
(53) Bush, B. L.; Nachbar, R. B., J r. Sample-distance partial least
squares: PLS optimized for many variables, with application
to CoMFA. J . Comput-Aided Mol. Des. 1993, 7, 587-619.
(58) Cheng, F.; Shen, J .; Luo, X.; Zhu, W.; Gu, J .; J i, R.; J iang, H.;
Chen, K. Molecular docking and 3-D QSAR studies on the
possible antimalarial mechanism of artemisinin analogues.
Bioorg. Med. Chem. 2002, 10, 2883-2891.
(59) Pal, J . K.; Purandare, M. J . Dose-dependent differential effect
of hemin on protein synthesis and cell proliferation in Leish-
mania donovani promastigots cultures in vitro. J . Biosci. (Indian
Acad. Sci.) 2001, 26, 225-231.
J M030181Q