Synthesis and Antimalarial Activity of Trioxaquine Derivatives

Article abstract of DOI:10.1002/chem.200305576

Trioxaquines are dual molecules that contain a trioxane motif linked to an aminoquinoline entity. Among the different compounds of this series, trioxaquine cis-15 (DU1302c), prepared from α-terpinene, a cheap natural product, showed efficient antimalarial activity in vitro on both sensitive and resistant strains of Plasmodium falciparum (IC₅₀ = 5-19 nM). A stereochemical description of this stable, nontoxic, and non-genotoxic antimalarial agent is detailed. Mice infected with P. vinckei were successfully treated with cis-15 in a four-day suppressive test. The doses required to decrease parasitemia by 50% (ED₅₀) were 5 and 18 mg kg ^-1d^-1 after intraperitoneal and oral administration, respectively. Parasitemia clearance was complete without recrudescence at an intraperitoneal dose of 20 mgkg^-1d^-1.

Full text of DOI:10.1002/chem.200305576

FULL PAPER
Synthesis and Antimalarial Activity of Trioxaquine Derivatives
Odile Dechy-Cabaret,[a] FranÁoise Benoit-Vical,[a, b] Christophe Loup,[a] Anne Robert,[a]
Heinz Gornitzka,[c] Anne Bonhoure,[d] Henri Vial,[d] Jean-FranÁois Magnaval,[b]
Jean-Paul Sÿguÿla,[b] and Bernard Meunier*[a]
Abstract: Trioxaquines are dual mole-
cules that contain a trioxane motif
linked to an aminoquinoline entity.
Among the different compounds of
of Plasmodium falciparum (IC50 =5
19 nm). A stereochemical description
of this stable, nontoxic, and non-geno-
toxic antimalarial agent is detailed.
Mice infected with P. vinckei were suc-
cessfully treated with cis-15 in a four-
day suppressive test. The doses re-
quired to decrease parasitemia by 50%
(ED50) were 5 and 18 mgkgꢀ1 dꢀ1 after
intraperitoneal and oral administration,
respectively. Parasitemia clearance was
complete without recrudescence at an
intraperitoneal dose of 20 mgkgꢀ1 dꢀ1.
this
series,
trioxaquine
cis-15
(DU1302c), prepared from a-terpi-
nene, a cheap natural product, showed
efficient antimalarial activity in vitro
on both sensitive and resistant strains
Keywords: aminoquinoline ¥ perox-
ides ¥ spiro compounds ¥ trioxane ¥
trioxaquine
Introduction
sistance has not been reported either in clinical isolates or
in laboratory experiments. These compounds are active
against the early form of the malarial blood stages and
induce a rapid clearance of the blood parasitemia. They also
act against gametocytes, which are the sexual stages of the
parasite that infect mosquitoes. As a result of a short half-
time of elimination and appearance of recrudescence, arte-
misinin derivatives are now combined with more slowly
eliminated drugs such as mefloquine or lumefantrine. This is
done both to increase the efficiency of the treatment and to
impede the development of resistance.[1,3] However, an er-
ratic supply of the natural parent compound and its short
half-life make it necessary to design cheap synthetic endo-
peroxide-based antimalarials.[4]
Malaria, the third most infectious cause of mortality, is pro-
lific in more than 40% of the world×s population and causes
more than one million deaths each year, especially in
Africa.[1] Given the spreading resistance of Plasmodium fal-
ciparum to most of the current antimalarial drugs (namely
chloroquine, mefloquine, pyrimethamine, and proguanil), as
well as the absence of a vaccine for protection against ma-
laria, there is an urgent need for new effective, safe, and af-
fordable drugs.[2] The use of artemisinin and related deriva-
tives is increasing since they are among the few molecules
active in the treatment of multidrug resistant P. falciparum
malaria. They are well tolerated, and to date, significant re-
The action of artemisinin arises from the presence of an
endoperoxide group that is able to produce radicals, which
in turn damage the parasite.[4a] Our group contributed to
demonstrating the alkylating activity of artemisinin deriva-
[a] Dr. O. Dechy-Cabaret, Dr. F. Benoit-Vical, C. Loup, Dr. A. Robert,
Dr. B. Meunier
7]
Laboratoire de Chimie de Coordination du CNRS
205 route de Narbonne, 31077 Toulouse Cedex 4 (France)
Fax : (+33)5-61-55-30-03
tives and other related antimalarial trioxanes.[5 Subse-
quently, we decided to develop new molecules based on the
™covalent bitherapy∫ concept in an effort to generate new
affordable drugs that would be less prone to the develop-
ment of resistance. These chimeric compounds, named triox-
aquines, are obtained by covalent attachment of a trioxane
entity, known to be responsible for the activity of artemisi-
nin, to an aminoquinoline entity, which is responsible for
the accumulation and activity of chloroquine in the parasite.
The synthesis and antimalarial activity of the first trioxa-
quine 5a, named DU1102 (Figure 1), was recently report-
ed.[8] Encouraged by the in vitro antimalarial activity of 5a
on laboratory strains,[8] as well as chloroquine- and pyri-
E-mail: bmeunier@lcc-toulouse.fr
[b] Dr. F. Benoit-Vical, Prof. J.-F. Magnaval, Prof. J.-P. Sÿguÿla
Service de Parasitologie-Mycologie
CHU Rangueil, 1 avenue Jean Poulhes, TSA 50032
31059 Toulouse Cedex 9 (France)
[c] Dr. H. Gornitzka
Laboratoire d×Hÿtÿrochimie Fondamentale et Appliquÿe
Universitÿ Paul Sabatier, B‚t. IIR1
118 route de Narbonne, 31062 Toulouse Cedex 4 (France)
[d] A. Bonhoure, Dr. H. Vial
UMR 5539, CNRS, Universitÿ Montpellier II
Place E. Bataillon, 34095 Montpellier Cedex 5 (France)
1625
Chem. Eur. J. 2004, 10, 1625 1636
DOI: 10.1002/chem.200305576
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

FULL PAPER
Figure 1. Structures of trioxaquine citrates 5a c, 6a, 6b, 10, cis-15, and 18. Trioxaquines 6a and 6b are racemates but only one enantiomer of each is de-
picted. Trioxaquines cis-15 and 18 are 50:50 mixtures of two diastereomeric racemates but only one of the four possible stereoisomers is depicted for
each trioxaquine.
methamine-resistant human isolates of P. falciparum,[9] we
aimed to synthesize a series of new trioxaquines (Figure 1)
in an effort to increase the antimalarial activity of this
series. In particular, we wanted to gain an understanding
into the influence of several structural parameters, such as:
1) the length of the tether between the 4-aminoquinoline
and the trioxane fragments [trioxaquines 5b (DU1106) and
5c (DU1108)]; 2) the use of different starting dienes [trioxa-
quines 10 (DU1402) and cis-15 (DU1302c), obtained from
1,3-cyclohexadiene and a-terpinene, respectively]; and 3)
the nature of the diketone used in the synthesis of the triox-
ane [trioxaquines 6a (DU1112) and 6b (DU1114) were ob-
tained from cis-bicyclo[3.3.0]octane-3,8-dione,[10] rather than
cyclohexane-1,4-dione, which has always been used previ-
ously]. Finally, trioxaquine 18 (DU2302c) was obtained
from primaquine, an antimalarial drug based on 8-aminoqui-
noline instead of the 4-aminoquinoline motif.
In vitro, trioxaquine cis-15 was found to be the most
active, its inhibitory concentration (IC50) being as low as
6 nm for the highly chloroquine-resistant P. falciparum strain
FcM29-Cameroon. Given these results, as well as taking into
consideration the feasibility of its synthesis, compound cis-
15 was chosen to be assessed for in vivo antimalarial activity
on mice (P. vinckei) in a four-day suppressive test. The doses
required to decrease parasitemia by 50% (ED50) were 5 and
18 mgkgꢀ1 dꢀ1 after intraperitoneal and oral administration,
respectively. Moreover, parasitemia clearance was complete
without recrudescence at an intraperitoneal dose of
20 mgkgꢀ1 dꢀ1. Mice treated at 120 mgkgꢀ1 dꢀ1 for four con-
secutive days (oral route) did not experience any toxic ef-
fects. Furthermore, it was found that trioxaquine cis-15 was
unable to induce the SOS response in E. coli even at a
20 mm concentration. This is 1000 times higher than its anti-
malarial IC50 values and indicates that in vitro cis-15 is not
genotoxic.
Results and Discussion
The parent trioxaquine 5a was obtained by the convergent
synthesis of building blocks that allowed for several modifi-
cations. Aminoquinoline 1a was prepared by condensation
of 1,4-diaminoethane with 4,7-dichloroquinoline. The triox-
ane ketone 3 was obtained by condensation of the 1,4-endo-
peroxide of 1,4-diphenyl-1,3-cyclopentadiene 2 with cyclo-
hexane-1,4-dione. Subsequent reductive amination of the tri-
oxane ketone 3 by the primary amine of aminoquinoline 1a
led to trioxaquine 4a. To enhance its solubility, compound
4a was protonated by citric acid to provide trioxaquine cit-
rate 5a (Scheme 1). As yet, a stereochemical study of com-
pound 5a has not been completed.
As previously documented for the condensation of endo-
peroxide 2 with cyclopentanone,[11] in trioxane ketone 3
(Scheme 1), both the trioxane and the cyclopentene are cis-
fused, and H5 and the phenyl C9 are contiguous. Therefore,
rather than a diastereomeric racemate (four stereoisomers),
as is theoretically expected for a molecule that has two
chiral carbons, compound 3 was obtained as a racemate.
Moreover, the reductive amination step in the trioxaquine
synthesis created a pseudoasymmetric center at the C12
spiro carbon. Therefore, two diastereomeric racemates were
obtained in which the amine and the endoperoxide were
1626
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 1625 1636

Synthesis of Trioxaquine Derivatives
1625 1636
the quinoline protons, 4b and 4c were also obtained (and
not separated) as the trans and cis diastereomeric racemates.
These were subsequently protonated with citric acid in
quantitative yields to give the corresponding trioxaquine di-
citrates 5b and 5c, respectively. The chemical shifts of the
quinoline and diaminoalkane tether protons indicate that in
a solution of DMSO, 5a c are monoprotonated trioxaquine
citrates. This has also recently been reported for trioxa-
quines 6a and 6b[10] (Figure 1).
Trioxaquine from 1,3-cyclohexadiene: Trioxaquines 4a c
were obtained from endoperoxide 2, which in turn is pre-
pared by photo-oxygenation of 1,4-diphenyl-1,3-cyclopenta-
diene. Unfortunately, this diene, which is synthesized in four
steps from the sodium salt of ethyl acetoacetate and phenac-
yl bromide,[11,15] is obtained in an overall yield of only 20%.
To reduce the number of steps and to enhance the total
yield of the synthesis, we decided to use a commercially
available diene. Therefore, as depicted in Scheme 2 and
Scheme 1. Synthesis of trioxaquines 5a c from 1,4-diphenyl-1,3-cyclopen-
tadiene: a) O2, hn, TPP, CH2Cl2, 58C; b) cyclohexane-1,4-dione, Me3-
SiOTf, CH2Cl2, ꢀ788C; c) H2N(CH2)nNH2, 858C; d) NaBH(OAc)3,
CH2Cl2; e) citric acid, acetone.
either trans or cis with respect to the mean plane of the cy-
clohexyl moiety. As a result, compound 4a (and consequent-
ly 5a) was isolated as an equimolecular mixture of two dia-
stereomeric racemates (four stereoisomers). This stereo-
chemical aspect is further discussed with respect to com-
pounds 12 and cis-14.
Trioxaquines synthesized from 4,7-dichloroquinoline
Scheme 2. Synthesis of trioxaquine 10 from 1,3-cyclohexadiene: a) O2,
hn, TPP, CH2Cl2, 58C; b) cyclohexane-1,4-dione, Me3SiOTf, CH2Cl2,
ꢀ788C; c) NaBH(OAc)3, 1a, CH2Cl2; d) citric acid, acetone.
Variations of the diaminoalkyl tether: Changes in the length
of the lateral chloroquine chain influence antimalarial activi-
ty, and are able, in some cases, to circumvent chloroquine-
resistance of Plasmodium falciparum.[12,13] Several trioxa-
quines with diaminoalkyl tethers of different lengths were
synthesized to determine the effect this had on antimalarial
activity. Substitution of 4,7-dichloroquinoline with different
a,w-diaminoalkanes is generally easy, but poor yields have
been reported in the case of long diaminoalkanes because of
the formation of bisquinoline derivatives.[14] Substitution of
4,7-dichloroquinoline with 1,3-diaminopropane and 1,4-dia-
minobutane gave the expected aminoquinolines 1b and 1c
in 75 and 60% yield, respectively (Scheme 1). The subse-
quent steps were then performed under conditions similar to
those used for 5a.[8] Finally, reductive amination of the triox-
ane ketone 3 by 1b and 1c produced trioxaquines 4b and
4c in 95 and 70% yield, respectively. As could be assessed
Scheme 3, respectively, trioxaquine 10 was prepared from
1,3-cyclohexadiene, while cis-15 was synthesized from a-ter-
pinene. Photo-oxygenation of 1,3-cyclohexadiene afforded a
mixture of the desired endoperoxide 7a (80% yield) and 3-
hydroperoxo-1,4-cyclohexadiene 7b (20% yield). Com-
1
pound 7a results from the [4+2]-cycloaddition of O2, while
1
7b arises from an ene reaction of O2. Attempts to purify 7a
by column chromatography were unsucessful. Subsequent
condensation of this mixture with cyclohexane-1,4-dione in
dichloromethane at ꢀ788C in the presence of Me3SiOTf as
catalyst gave, after silica-gel chromatography, the expected
trioxane ketone 8 in low yield (<10%). The low yield ob-
tained for 8 arises because the 1,4-endoperoxide does not
contain any substituents. Therefore, in contrast to 1,4-di-
phenyl-1,3-cyclopentadiene, when this endoperoxide is
opened by Me3SiOTf, an unstabilized carbocation is formed.
1
from the doubled-up H NMR signals assigned to H5 and
1627
Chem. Eur. J. 2004, 10, 1625 1636
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

B. Meunier et al.
FULL PAPER
In addition, in trioxane ketone 8 the trioxane and cyclohex-
ene ring can be either cis or trans fused; this gives rise to a
mixture of two diastereomeric racemates, which correspond
to four different possible configurations at C5 and C10. This
is not the case for trioxane ketone 3, as the short-ring con-
straint of cyclopentene does not allow a trans fusion. The
four stereoisomers of compound 8 could not be separated
(this problem will be further discussed for trioxane ketone
12 and trioxaquine cis-14, see below, Scheme 3). Reductive
amination of 8 by aminoquinoline 1a produced trioxaquine
9. The aminoquinoline fragment can be introduced at C13
on the same side as the peroxide bond with respect to the
mean plane of the cyclohexane ring, or on the opposite side.
Thus, this gives rise to two diastereomeric trioxaquines for
each starting trioxane ketone. Compound 9 (which refers to
the mixture of eight stereoisomers) was then further proto-
nated to give trioxaquine citrate 10 in 40% overall yield
from trioxane ketone 8. The use of 1,3-cyclohexadiene did
reduce the number of steps in the synthesis but did not im-
prove on the overall yield (7% for 5a from the sodium salt
of ethyl acetoacetate compared to 2% for 10 from 1,3-cyclo-
hexadiene). Moreover, the latter synthesis doubled the
number of stereoisomers obtained.
Trioxaquine from a-terpinene: Given the low yields ob-
tained when an unsubstituted diene such as 1,3-cyclohexa-
diene was used, we turned our attention to the cheap and
commercially available disubstituted diene a-terpinene (1-
methyl-4-isopropyl-1,3-cyclohexadiene). We hereby report a
detailed study that includes stereochemical considerations,
of the synthesis of trioxaquine cis-15. Photo-oxygenation of
a-terpinene afforded the desired endoperoxide 11. Com-
pound 11 is in fact the natural product ascaridole, a racemic
endoperoxide that has a (1R,4R) and (1S,4S) configuration,
and which exhibits weak antimalarial activity (Scheme 3).[16]
Condensation of cyclohexane-1,4-dione with 11 in dichloro-
methane at ꢀ788C in the presence of Me3SiOTf as catalyst
gave, after separation by silica gel chromatography, three
different trioxane ketones, namely trans-12, cis-12, and
trans-13. Both trans-12 and cis-12 have the expected triox-
ane ketone structure and consist of two racemic diastereom-
ers. In cis-12, the trioxane and cyclohexene rings are cis
fused, that is, H5 and the 11-methyl group are contiguous.
This corresponds to a 5S,10R and 5R,10S configuration.
Therefore, cis-12 is a racemate that should be noted as
(5RS,10SR)-12 according to CIP convention.[17] However, to
shorten this notation, we use the cis/trans descriptor because
it is also accepted for cyclanes. In trans-12, the trioxane and
cyclohexene rings are trans fused, that is, H5 and the 11-
Scheme 3. Synthesis of trioxaquine cis-15 from a-terpinene: a) O2, hn,
TPP, CH2Cl2, 58C; b) cyclohexane-1,4-dione, Me3SiOTf, CH2Cl2, ꢀ788C;
c) NaBH(OAc)3, 1a, CH2Cl2; d) citric acid, acetone; e) cis-14 is a 50:50
mixture of the two diastereomers trans,cis-14 and cis,cis-14; f) cis-15 is a
50:50 mixture of the two diastereomers trans,cis-15 and cis,cis-15.
slow evaporation of a solution of the corresponding com-
pound from
a
hexane/EtOAc (80:20, v/v) mixture
(Figure 2). It should be noted that in the case of trioxane
ketone 3, a trans fusion of the trioxane and cyclopentene
ring is not possible because of steric constraints.[11] There-
fore, the only stereoisomers obtained for 3 were enantiom-
ers that had a 5S,9S and 5R,9R configuration.
The structure of the third, minor trioxane ketone trans-13
was determined by NMR experiments. Compound 13 is an
analog of trans-12 in which the isopropyl substituent at C7
has been oxidized to an isopropenyl group (Scheme 3). In
the 1H NMR spectrum of trans-13, the signal for the 11-
methyl group appeared, as it did for trans-12, as a sharp
signal, whereas this resonance was a very broad singlet in
the case of cis-12. This analysis indicates that trans-13 has a
(5RS,10RS) configuration. In addition, NOE experiments
showed a spatial proximity between one of the H13 protons
and the two H8 protons; this is compatible with an s-trans
conformation of the diene.
The trioxane ketone cis-12 was the major product isolat-
ed for all the different experimental conditions used. For ex-
ample, when the reaction was carried out with 6 molar
equivalents of diketone for 2 h, the trans-12/cis-12/trans-13
ratio was 12:78:10, and cis-12 was isolated in 26% yield
with respect to the a-terpinene. Unfortunately, changes in
reaction time, temperature, quantity of catalyst, or quality
ꢀ
methyl group are on opposite planes of the C5 C10 bond.
In turn, this gives rise to the 5S,10S and 5R,10R configura-
tion (trans-12 should, therefore, be noted as (5RS,10RS)-12
according to CIP convention).
Although the formation of a trioxane that bears an angu-
lar isopropyl residue at C10 cannot in priciple be ruled out
during the condensation of ascaridole 11 with cyclohexane-
1,4-dione, this compound was not detected.
The structures of cis-12 and trans-12 were unambiguously
determined by X-ray diffraction of crystals obtained by the
1628
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 1625 1636

Synthesis of Trioxaquine Derivatives
1625 1636
ductive amination at C17 should afford compounds that
have the amine group of the aminoquinoline at C17 and the
peroxide at C3 either in an equatorial or axial position with
respect to the cyclohexane ring. Therefore, when a racemate
of the trioxane ketone cis-12 (5R,10S + 5S,10R) is used as
the starting material, reductive amination is expected to pro-
vide eight compounds. In particular, there should be four
stereoisomers, each of which exist as two conformers be-
cause of the conformational equilibrium of the cyclohexane
ring (Figure 3).
In B and C (and B’ and C’), the amine and the peroxide
substituents are trans with respect to the cyclohexane ring.
For this reason, the mixture of racemates BB’ + CC’ is
named trans,cis-14. Here trans refers to the relative position
of the aminoquinoline and peroxide with respect to the cy-
clohexane ring, while cis refers to the trioxane cyclohexene
ring junction. In A and D (and A’ and D’), the amine and
the peroxide substituents are cis, that is, on the same side of
the mean plane of the cyclohexane ring. As a result, this dia-
stereomeric racemate is named cis,cis-14.
To enable structure elucidation and biological evaluation,
the two diastereomers of cis-14 were partially separated by
selective precipitation (dichloromethane/cyclohexane, 1:8,
v/v); this led to an enriched trans,cis-14/cis,cis-14 mixture
(70:30). The enriched mixture was then subjected to column
chromatography on silica gel (EtOAc/triethylamine, 95:5,
v/v). The first diastereomeric racemate eluted was trans,cis-14.
Its structure was unambiguously determined [X-ray diffrac-
tion of a crystal obtained by the slow evaporation of an eth-
anol/EtOAc (50:50, v/v) solution (Figure 2)] to be conform-
er C (and C’), in which the aminoquinoline is in an equatori-
al position. Conformer B (and B’), which is less stable be-
cause the aminoquinoline residue is axial, was not observed.
The second diastereomeric racemate eluted was cis,cis-14.
This diastereomeric racemate was crystallized by the slow
diffusion of diethyl ether into a dichloromethane solution.
In this compound, the asymmetric unit contains both the A
and D conformer. It should be noted that a fast equilibrium
between conformers A (and A’) and D (and D’) exists in
solution, and therefore, did not allow the two to be distin-
guished by NMR spectroscopy at room temperature. More-
over, this conformational equilibrium makes the chromato-
graphic purification of cis,cis-14 tedious and erratic. The re-
covery of trans,cis-14 was much easier.
Each of the two diastereomeric racemates were protonat-
ed with citric acid to afford the corresponding trioxaquine
citrates trans,cis-15 and cis,cis-15 in quantitative yield. These
citrates were then tested for in vitro antimalarial activity.
Since trans,cis-14 and cis,cis-14 displayed similar in vitro
antimalarial activity (see below), they were not separated
any further. Compound cis-14, which consisted of a 50:50
Figure 2. Crystal structures of trioxanes cis-12 and trans-12, as well as tri-
oxaquines trans,cis-14 and cis,cis-14. In cis,cis-14, all conformers and
enantiomers co-crystallized but only A and D are represented.
1
of solvent (distilled or not) did not improve the yield of cis-
12. Thus, it was necessary to separate and purify the major
diastereomer cis-12 by chromatography.
mixture (quantified on the basis of its H NMR analysis) of
the two diastereomeric racemates trans,cis-14 and cis,cis-14
(four stereoisomers) was thereby protonated with citric acid
to afford cis-15. Subsequently, this mixture of trans,cis-15 and
cis,cis-15 (50:50 ratio) was used for in vivo antimalarial tests.
Reductive amination of cis-12 by aminoquinoline 1a af-
forded trioxaquine cis-14 in 60 80% yield (Scheme 3). From
1
analysis of the H NMR spectra, cis-14 was found to be a
50:50 mixture of two trioxaquine diastereomeric racemates
(i.e. four stereoisomers). In principle, non-stereoselective re-
Trioxaquine synthesized from primaquine: The quinoline
entity of all the trioxaquines described up to this point was
1629
Chem. Eur. J. 2004, 10, 1625 1636
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

B. Meunier et al.
FULL PAPER
reductive amination of a triox-
ane ketone. Reductive amina-
tion of trioxane cis-12 with pri-
maquine gave trioxaquine 17 in
63% yield (Scheme 4). Subse-
quent protonation was carried
out in acetone with 2.0 molar
equivalents of citric acid to
afford trioxaquine citrate 18
after evaporation of the solvent.
This new trioxaquine is highly
soluble in acetone as well as in
water. It should be noted that
the diastereomers from the cis
or trans disposition of the per-
oxide and primaquine substitu-
ent with respect to the cyclo-
hexane ring were not separated.
Compound 18 consisted of an
equimolecular mixture of eight
stereoisomers, just like cis-15
because it contains an addition-
al chiral center at C13’ of pri-
maquine.
Characteristic data for the tri-
oxaquine citrates: Table 1 pres-
ents several characteristic fea-
tures of the different trioxa-
quines: molecular weight, over-
all yield for the syntheses from
commercially available reac-
tants, number of steps, and sol-
ubility. The trioxaquine pre-
pared from primaquine is the
only one that is soluble in pure
water. All the other trioxa-
quines have to be previously
dissolved in DMSO prior to the
Figure 3. The eight possible structures of trioxaquine cis-14, in which the amine and the peroxide functions are
ꢀ
either axial or equatorial with respect to the cyclohexane residue. (Quin NH refers to the aminoquinoline 1a,
see Scheme 1).
similar to chloroquine, a 7-chloro-4-aminoquinoline, and
was used on the basis of many structure activity relationship
studies.[18] Primaquine 16 was considered as a suitable alter-
native as it has been widely used for eradication of the dor-
mant hepatic forms of the parasite responsible for relapses
in P. vivax or P. ovale infec-
addition of water, physiological serum, or culture medium.
The reported data indicate the maximum trioxaquine con-
centration in DMSO that can be diluted by an infinite
amount of physiological serum without precipitation.
Among the trioxaquines synthesized from 4-aminoquino-
tions,[19] and because it is also
active against the sexual stages
of the parasite involved in the
transmission of the disease.
Therefore, incorporation of a
primaquine moiety and a triox-
ane fragment into a single com-
pound could potentially lead to
new drugs that have both game-
tocidal and schizontocidal activ-
ities. Primaquine 16 (Scheme 4)
is an 8-aminoquinoline that
contains a primary amine that
Table 1. Molecular weights, yields, and solubilities of the different trioxaquines.
Trioxaquine
Molecular weight
[base (citrate form)]
Global
Number
Solubility[b]
yield[a] [%]
of steps[a]
[mgmLꢀ1
]
5a (DU1102)
5b (DU1106)
5c (DU1108)
6a (DU1112)
6b (DU1114)
10 (DU1402)
cis-15 (DU1302c)
18 (DU2302c)
568.2 (952.4)
582.2 (966.4)
596.2 (980.4)
594.2 (978.4)
594.2 (978.4)
430.0 (814.2)
485.1 (870.3)
523.0 (907.2)
7
5
5
2
2
2
20
20
9
9
9
9
9
5
5
4
1.0ꢁ0.2
1.0ꢁ0.2
1.0ꢁ0.2
0.6ꢁ0.2
0.6ꢁ0.2
70ꢁ10
50ꢁ10
soluble in water[c]
[a] Calculated with respect to the starting materials (sodium salt of ethyl acetoacetate for 5a c and 6a and 6b,
1,3-cyclohexadiene for 10, and a-terpinene for cis-15 and 18). [b] These trioxaquines are not directly soluble in
water or physiological serum. The reported solubility is the maximum concentration of the trioxaquine citrate
in DMSO that can be diluted with an infinite amount of physiological serum (0.9 wt% NaCl in water) without
can be directly utilized in the precipitation. [c] Solubility in water: 1 mgmLꢀ1
.
1630
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 1625 1636

Synthesis of Trioxaquine Derivatives
1625 1636
First of all, it should be highlighted that the antimalarial
activities are high, especially for trioxaquines 5a and cis-15,
which have IC50 values in the range of 22 27 nm and 5
19 nm, respectively. In addition, these drugs are as active on
chloroquine resistant as on chloroquine sensitive P. falcipa-
rum strains. For cis-15, the IC50 values are 7 and 17 nm for
F32-Tanzania and the Nigerian (sensitive) strains, and 15
and 6 nm for the FcB1 and FcM29 (resistant) strains, respec-
tively.
Variations in the length of the diaminoalkyl tether (-CH2-)n
between the trioxane and the 4-aminoquinoline moieties
(5a, n=2; 5b, n=3; 5c, n=4) did not significantly modify
the antimalarial activity. In some cases, lower activities were
observed when the diaminoalkyl chain was longer than two
carbon atom units: IC50 =181 and 41 nm for the Nigerian
and FcB1 strains, respectively, for 5b (n=3), while for 5a
(n=2) the values obtained were 22 and 27 nm, respectively.
As a result, a two-carbon atom tether was used for most of
the trioxaquines that contained a 4-aminoquinoline fragment.
The influence of certain stereochemical features was also
evaluated. Trioxaquines 6a and 6b (Figure 1) are diastereo-
meric racemates, which were prepared from cis-bicy-
clo[3.3.0]octane-3,8-dione. They differ only in the stereo-
chemistry at the C3 spiro junction. That is, the peroxide
function is either on the same side or on the opposite side
Scheme 4. Synthesis of trioxaquine 18 from primaquine: a) NaBH(OAc)3,
CH2Cl2; b) citric acid, acetone. cis-12 is a racemate but only one enan-
tiomer is depicted. Trioxaquines 17 and 18 are 50:50 mixtures of two dia-
stereomeric racemates but only one of the four possible stereoisomers is
depicted for each trioxaquine.
Table 2. IC50 values [nm] for the trioxaquine citrates against Plasmodium falciparum.
lines, those that contain a cyclo-
hexene (10) or terpinene
moiety (cis-15) had a solubility
of 70 and 50 mgmLꢀ1 in
DMSO, respectively. This is
much higher than that found
for trioxaquines 5 and 6, which
Trioxaquine
citrates
IC50 [nm][a]
Nigerian
CQS[b]
FcB1-Columbia
FcM29-Cameroon
CQR[b]
CQR+[b]
5a (DU1102)
5b (DU1106)
5c (DU1108)
6a (DU1112)
6b (DU1114)
10 (DU1402)
cis-15 (DU1302c)
trans,cis-15
22 (1)
181 (59)
87 (20)
112[c]
27 (3)
41 (4)
41 (4)
31 (17)
63 (2)
71 (52)
15 (2)
27 (10)
133 (102)
41 (14)
39 (22)
85 (7)
49 (0)
6 (1)
7 (2)
5 (1)
have
a
diphenylcyclopentene
108 (43)
36 (12)
residue (1 mgmLꢀ1). A molecu-
lar weight below 500 is one of
the ™five×s rule∫ points reported
by Lipiski et al. in order for a
drug to have good oral availa-
bility.[20] With a molecular mass
of 485 for its base form, and in
light of the high overall yield
(20%) obtained, cis-15 is the
most promising compound.
17(6); [7[c]
]
[d]
[d]
[9[c]
[5[c]
]
]
19 (2)
11 (4)
[d]
cis,cis-15
18 (DU2302c)
Chloroquine[e,f]
Artemisinin[e]
Primaquine[e,g]
176 (39)
62 (10)
6 (1)
112 (17)
116 (57)
5 (1)
108 (16)
174 (26)
8 (1)
875 (200)
1075 (17)
1400 (300)
[a] IC50 values are the mean of at least 2 5 independent experiments, except where noted (the SEM value is
indicated in parentheses, SEM=standard error of the mean). [b] CQS=chloroquine sensitive, CQR=chloro-
quine resistant, CQR+=highly chloroquine resistant. [c] Single experiment. [d] Measured on the chloroquine-
sensitive F32-Tanzania strain instead of the Nigerian strain. [e] Measured as reference. [f] Chloroquine diphos-
phate was used. [g] Primaquine diphosphate was used.
Biological activity
In vitro antimalarial activity:
of H11 and H15 with respect to the C10-C12-C14-C16
We tested the in vitro antimalarial activity of the eight syn-
thesized trioxaquine citrates to determine the influence that
the different structural modifications had on their biological
activity. Four different strains of Plasmodium falciparum
were used: a Nigerian strain and F32-Tanzania, which are
chloroquine sensitive (CQS); and FcB1-Columbia and
FcM29-Cameroon, which are chloroquine resistant (CQR)
and highly chloroquine resistant (CQR+), respectively. The
threshold IC50 for in vitro resistance to chloroquine is usual-
ly estimated to be 100 nm.[21] The results are summarized in
Table 2.
plane; this gives rise to the exo and endo racemates 6a and
6b, respectively. The structure of each diastereomer has
been proposed from a detailed NMR study.[10] The antima-
larial activity of these two diastereomers was not significant-
ly different. For example, IC50 values of 112 and 108 nm
were obtained for 6a and 6b, respectively, against the chlor-
oquine-sensitive Nigerian strain. For the chloroquine-resist-
ant strain a difference was observed (factor of 2), but it was
low and did not allow us to conclude that the C3 stereo-
chemistry of this trioxaquine has any real influence on its
antimalarial activity.
1631
Chem. Eur. J. 2004, 10, 1625 1636
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

B. Meunier et al.
FULL PAPER
Changes in the starting diene, that is, from 1,4-diphenyl-
1,3-cyclopentadiene in 5a to 1,3-cyclohexadiene in 10 or a-
terpinene in cis-15 also did not effect the antimalarial activi-
ties of these compounds and they too were quite active. Of
the two, trioxaquine cis-15 is the more active against both
the CQS and CQR strains, with IC50 values of 7 and 17 nm
for the sensitive F32-Tanzania and Nigerian strains, respec-
tively, and 15 and 6 nm for the resistant FcB1 and FcM29
strains, respectively (Table 2). These results indicate that tri-
oxaquine cis-15 is efficient on a wide range of strains. Fur-
thermore, each of the two diastereomers of cis-15 (trans,cis-
15 and cis,cis-15) were independently tested, and each were
found to have activities in the same range. For the highly re-
sistant strain FcM29, the IC50 values were 6(1) nM for the
50:50 mixture of diastereomers, and 7(2) and 5(1) nM for
trans,cis-15 and cis,cis-15, respectively. On the other hand,
for the chloroquine-sensitive F32 strain, the IC50 values were
9 nm for trans,cis-15 and 5 nm for cis,cis-15, while the value
for cis-15 was 7 nm. These IC50 values fall below the range
expected (10 20 nm) for efficient antimalarials based on a
trioxane moiety.[4,22,23] Since the two diastereomers of cis-15
displayed similar antimalarial activity in vitro, they were not
separated for further in vivo evaluation in mice. The high ef-
ficacy and solubility of trioxaquine cis-15, as well as the ease
with which it was synthesized, make cis-15 an attractive tri-
oxaquine.
A lower antimalarial activity was observed for the com-
pounds that contained 8-aminoquinoline as the quinoline
moiety (i.e. 18), and confirmed that the 4-aminoquinoline
entity is vital for activity. However, the results obtained for
18 also showed that covalent coupling of a trioxane moiety
with primaquine leads to a large increase in the intraery-
throcytic antimalarial activity in comparison to primaquine
itself, particularly for the CQR strains (FcB1 strain: IC50 =
112 and 1075 nm for trioxaquine 18 and primaquine, respec-
tively). This result indicates that 18, as a ™modified prima-
quine∫, may have potential usage against both sexual and
asexual erythrocytic stages. In addition, since primaquine is
the only drug active against hepatic malaria stages, it would
be of interest to test 18 on hepatic parasite stages.
be noted that there was an absence of in vivo toxicity at
doses of 120 mgkgꢀ1 dꢀ1. Infected mice treated at this dose
orally for four days did not experience any toxic effects over
60 days of being monitored, and weight loss was not ob-
served for cured mice. The absence of toxicity at high doses
administered orally has also been observed in non-infected
mice (100 mgkgꢀ1 dꢀ1 for three consecutive days).
In addition, it should be noted that the ED50 value for
oral treatment with cis-15 was only three to fourfold higher
than that obtained when treatment was intraperitoneal; this
suggests that this drug has a rather good bioavailability by
oral administration.
In vitro genotoxicity evaluation of cis-15 (DU1302c): The
genotoxicity of trioxaquine cis-15 was also evaluated.
Agents that damage DNA, such as alkylating or oxidative
agents, or UV radiation, also induce systems of DNA repair,
one of which involves a complex system of cellular changes
known as the SOS response. The SOS response has been ex-
haustively studied for Escherichia coli, and it has been
shown that a correlation exists between the ability of a drug
to induce the SOS response in E. coli and the genotoxicity
that it will display in humans.[26] To evaluate the potential
genotoxicity of cis-15, we measured its ability to induce the
SOS response in a modified strain of E. coli (GE 864). This
strain shows an increased expression of b-galactosidase
upon induction of the SOS response.[27] The mutagen drug
mitomycin C was used as a standard. The b-galactosidase ac-
tivity was measured through hydrolysis of the o-nitrophenyl-
b-d-galactopyranoside (ONPG), and was monitored spectro-
photometrically at 420 nm.[28] The ability to induce the SOS
response was then compared for trioxaquine cis-15 (5 mm in
DMSO) and mitomycin C (3 mm in H2O) (pure DMSO and
H2O were tested as blank experiments). The results are pre-
sented in Figure 4. Only baseline b-galactosidase activity
was detected when E. coli was treated with cis-15, either in
H2O or DMSO. Therefore, in contrast to mitomycin C, triox-
aquine cis-15 was unable to induce the SOS response. As a
result, higher concentrations of cis-15 were tested, and even
at a concentration of 20 mm, which is 1000 times more than
what was used to obtain the antimalarial IC50 values (results
not shown), cis-15 was unable to induce the SOS response.
In vivo antimalarial activity of cis-15 (DU1302c) against P.
vinckei: Compound cis-15 (DU1302c) was tested in mice in-
fected with P. vinckei according to a modified version of
Peters× four day suppressive test, and was evaluated both
after intraperitoneal and oral administration using the intra-
venous route for parasite inoculation. Parasitemia in control
mice generally reached as high as 20 40% on day 5, and
death occurred within the following four days. After once
daily intraperitoneal administration for four days, the first
dose being administered 1 day after parasite inoculation,
compound cis-15 exhibited potent antimalarial activity with
an ED50 of 5 mgkgꢀ1 dꢀ1. Moreover, parasitemia was com-
pletely cured without recrudescence over 60 days when 20
and 50 mgkgꢀ1 dꢀ1 doses were administered.
Stability of the trioxaquine cis-15: An antimalarial drug can-
didate should be stable under most storage conditions.
1
Indeed, as determined by H NMR spectroscopy, crystalline
trioxaquine cis-15 underwent less than 5% decomposition
upon storage in air at room temperature for more than six
months, or in acidic solution (HCl 0.01m) for three days.
The thermal stability of cis-15 is also very good. Upon heat-
ing the crystalline trioxaquine at 608C in air for 24 h, less
than 5% decomposition was observed (1H NMR spectrosco-
py).
By oral administration, an ED50 value of 18 mgkgꢀ1 dꢀ1
was obtained for cis-15; this corresponds to 21 mmolkgꢀ1 dꢀ1,
and is in the range reported for artemisinin antimalarial ac-
tivity (18 30 mmolkgꢀ1 dꢀ1 against P. berghei).[24,25] It should
Conclusions
A family of new antimalarial drugs has been prepared by
modification of the convergent synthesis of trioxaquine 5a.
1632
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 1625 1636

Synthesis of Trioxaquine Derivatives
1625 1636
Experimental Section
Chemical materials and methods: NMR spectra were recorded on
Bruker AM250, DPX300, or AMX400 spectrometers. Chemical shifts are
given with respect to external trimethylsilane (TMS). DCI mass spectra
were acquired on a Nermag R10-10H instrument and electrospray (ES)
mass spectra were obtained on an API 365 Sciex Perkin Elmer instru-
ment. Chromatography columns were conducted on silica gel 60 ACC
Chromagel, 70 200 mm granulometry (SDS, Peypin, France), and 60 F254
plates (Merck, Darmstadt, Germany) were used for thin-layer chroma-
tography. Chloroquine diphosphate, primaquine diphosphate, artemisinin,
and mitomycin C were purchased from Sigma Aldrich, France.
Crystallographic data
≈
cis-12: C16H24O4, Mr =280.35, triclinic, P1, a=5.9954(5), b=9.3036(8),
c=14.2689(12) ä, a=106.335(1), b=90.329(2), g=100.609(2)8, V=
744.32(11) ä3, Z=2, T=193(2) K, 5064 reflections (3617 independent,
R
int =0.0143), largest electron density residue: 0.410 eäꢀ3, R1 [for I>
2s(I)]=0.0518 and wR2 =0.1493 (all data) with R1 =ꢀkFo jꢀjFc k/ꢀjFo j
1
2
and wR2 =[ꢀw(F20ꢀF2c)2/ꢀw(F20)2] =
.
Figure 4. Genotoxicity test: trioxaquine cis-15-mediated induction of b-
≈
&
galactosidase expression by E. coli [5 mm in DMSO ( )]. Mitomycin C
trans-12: C16H24O4, Mr =280.35, triclinic, P1, a=6.056(2), b=9.230(3),
c=14.055(4) ä, a=104.992(6), b=90.626(6), g=100.614(6)8, V=
744.5(3) ä3, Z=2, T=193(2) K, 2431 reflections (1824 independent,
*
*
[3 mm in H2O ( )], H2O (&), and DMSO ( ) are given for comparison.
R
int =0.0201), largest electron density residue: 0.277 eäꢀ3, R1 [for I>
2s(I)]=0.0642 and wR2 =0.1905 (all data).
trans,cis-14: C27H36ClN3O3, Mr =485.04, monoclinic, P21/c, a=17.429(2),
b=11.143(1), c=12.932(1) ä, b=91.443(2)8, V=2510.7(5) ä3, Z=4, T=
153(2) K, 8518 reflections (3206 independent, Rint =0.0492), largest elec-
tron density residue: 0.255 eäꢀ3, R1 [for I>2s(I)]=0.0594 and wR2 =
0.1464 (all data).
Compound 5a was the first dual antimalarial drug to com-
bine both a trioxane entity that acts as a potential alkylating
agent, as in artemisinin, and a 4-aminoquinoline moiety sim-
ilar to chloroquine. Therefore, trioxaquines are composed of
two fragments that act on a common target, namely heme
released from parasite hemoglobin digestion, but do so
through two different mechanisms. This concept of ™cova-
lent bitherapy∫ is better than a simple combination of an
aminoquinoline with a trioxane, as it allows potent antima-
larial agents to be prepared, while at the same time consid-
erably reducing the risk of drug resistance.
Trioxanes functionalized by a ketone were synthesized by
reaction of a diketone with the photo-oxygenation product
of a diene, namely 1,4-diphenyl-1,3-cyclopentadiene or a-
terpinene. The described trioxaquines were obtained by re-
ductive amination of the resultant trioxane ketones with a
4-aminoquinoline that bears a primary amine side chain.
Synthetic modifications (quinoline, diamine, diene, dike-
tone) of the parent trioxaquine afforded compounds with
excellent in vitro antimalarial activities against different
CQS and CQR strains of P. falciparum. The majority of IC50
values obtained were in the range 5 40 nm.
Trioxaquine cis-15 (DU1302c) was selected for in vivo
evaluation. Mice infected with P. vinkei were successfully
treated in a four-day suppressive test with a once daily intra-
peritoneal or oral administration. The doses required to de-
crease parasitemia by 50% (ED50) were 5 and 18 mgkgꢀ1 dꢀ1
for intraperitoneal and oral administration, respectively. Par-
asitemia clearance was complete without recrudescence at
an intraperitoneal dose of 20 mgkgꢀ1 dꢀ1. In addition, trioxa-
quine cis-15 did not induce any toxic effects in mice treated
by the oral administration of 120 mgkgꢀ1 dꢀ1 over four con-
secutive days.
cis,cis-14: C27H36ClN3O3, Mr =485.04, orthorhombic, Pca21, a=13.079(4),
b=11.157(3), c=35.171(10) ä, V=5132(3) ä3, Z=8, T=193(2) K, 11630
reflections (5093 independent, Rint =0.1790), largest electron density resi-
due: 0.318 eäꢀ3, R1 [for I>2s(I)]=0.0921 and wR2 =0.2488 (all data).
Very small, weak diffracting crystals and high disorder problems account
for the high R values obtained, as well as a data/parameter ratio of 6.16.
Data for all the structures were collected at low temperatures using an
oil-coated shock-cooled crystal on a Bruker-AXS CCD 1000 diffractome-
ter with MoKa radiation (l=0.71073 ä). The structures were solved by
direct methods (SHELXS-97)[29] and all non-hydrogen atoms were re-
fined anisotropically using the least-squares method on F2.[30] CCDC-
219910 219913 contain the supplementary crystallographic data for this
paper. These data can be obtained free of charge from www.ccdc.cam.
ac.uk/conts/retrieving.html (or from the CCDC, 12 Union Road, Cam-
bridge, CB2 1EZ, UK; fax: (+44)1223-336033; e-mail: deposit@ccdc.
cam.ac.uk).
Synthesis of the aminoquinolines: In a typical procedure,[8] a mixture of
4,7-dichloroquinoline (5.0 g, 25 mmol) and a,w-diaminoalkane (5.0 mol
equiv) was stirred at 858C for 5 h. Sodium hydroxide (40 mmol) was then
added (1m aqueous solution), and the resultant solid was extracted with
EtOAc (250 mL) at 508C. The organic layer was washed with distilled
water, brine, distilled water again, dried over anhydrous sodium sulfate,
and the solvent was evaporated under vacuum to dryness. Aminoquino-
lines 1a c were used without further purification. Proton numbering is
reported in Figure 1 (the same numbering applies for both aminoquino-
lines 1a c and the aminoquinoline moiety of trioxaquines 5a c). The syn-
thesis of compounds 1a, 4a, and 5a was reported in ref. [8].
7-Chloro-4-[N-(3-aminopropyl)amino]quinoline
(1b):
Light-yellow
powder; yield: 75%; 1H NMR (250 MHz, CDCl3): d=1.58 (brs, 2H;
H2NC13’), 1.87 (m, 2H; H12’), 3.03 (t, 2H; H13’), 3.39 (m, 2H; H11’),
3
3
4
6.30 (d, J=5.5 Hz, 1H; H3’), 7.29 (dd, J=8.9 and J=2.2 Hz, 1H; H6’),
3
4
7.50 (brs, 1H; HNC4’), 7.70 (d, J=8.9 Hz, 1H; H5’), 7.90 (d, J=2.2 Hz,
1H; H8’), 8.48 ppm (d, J=5.5 Hz, 1H; H2’); MS (DCI/NH3+): m/z (%):
3
235 (2), 236 (100) [M+H]+, 237 (14), 238 (34), 239 (5).
7-Chloro-4-[N-(4-aminobutyl)amino]quinoline
(1c):
Light-yellow
In conclusion, trioxaquine cis-15 meets the main criteria
of pharmacological activity necessary for further biological
evaluation, namely the absence of genotoxicity, as well as
good chemical stability. Therefore, cis-15 can be considered
as a promising antimalarial drug candidate.
powder; yield: 60%; 1H NMR (250 MHz, CDCl3): d=1.45 (brs, 2H;
H2NC14’), 1.64 (m, 2H; H13’), 1.85 (m, 2H; H12’), 2.81 (t, 2H; H14’),
3.29 (m, 2H; H11’), 6.04 (brs, 1H; HNC4’), 6.36 (d, 3J=5.5 Hz, 1H;
3
4
3
H3’), 7.31 (dd, J=8.9 and J=2.2 Hz, 1H; H6’), 7.72 (d, J=8.9 Hz, 1H;
H5’), 7.92 (d, 4J=2.2 Hz, 1H; H8’), 8.50 ppm (d, 3J=5.5 Hz, 1H; H2’);
1633
Chem. Eur. J. 2004, 10, 1625 1636
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

B. Meunier et al.
FULL PAPER
MS (DCI/NH3+): m/z (%): 249 (2), 250 (100) [M+H]+, 251 (18), 252
(36), 253 (5).
(C13), 120.5 (C6), 142.1 (C7), 142.8 (C12), 211.2 ppm (C17); MS (DCI/
NH3+): m/z (%): 280 (100), 281 (22), 282 (8).
Synthesis of the trioxaquines: In a typical procedure,[8] the trioxane ke-
tone (1.0 molequiv, 50 mm) and aminoquinoline (1.25 equiv) were mixed
in dichloromethane. Sodium triacetoxyborohydride (1.25 equiv) was then
added. The reaction mixture was stirred at room temperature for 15 20 h
and was then washed twice with distilled water. The organic layer was
dried over anhydrous sodium sulfate and the solvent was evaporated
under vacuum to dryness. For the preparation of trioxaquine 17, the com-
mercially available primaquine diphosphate was initially deprotonated
with aqueous sodium hydroxide (1m), and the primaquine base was then
extracted with dichloromethane. This step was quantitative.
Synthesis of the trioxane ketones: In a typical procedure, the endoperox-
ide (25 mmol) was first obtained by photo-oxygenation (irradiation with
a 600 W visible lamp at 0 58C under bubbling oxygen with tetraphenyl-
porphyrin as photosensitizer) of the corresponding diene in dichlorome-
thane (50 mL). Formation of the endoperoxide was monitored by
1H NMR spectroscopy of the crude mixture (1 5 h from the starting
diene). When photo-oxygenation was complete, the crude solution was
cooled to ꢀ788C. Cyclohexane-1,4-dione (6.0 molequivwith respect to
the starting diene) and Me3SiOTf (0.4 equiv) were added, and the reac-
tion was stirred at ꢀ788C for 2 h. Triethylamine (1 equiv) was then
added and the solution was allowed to warm up to room temperature.
The organic layer was washed with distilled water, dried over anhydrous
sodium sulfate, and the solvent was evaporated. The light-brown powder
isolated was purified by column chromatography. Proton and carbon
atom numbering is reported in Scheme 2 and 3 for compounds 8 and 12,
respectively.
Proton and carbon atom numbering is reported in Figure 1 (the same
numbering is used for each trioxaquine and the corresponding citrate
salt, for instance for trioxaquine 4a and its citrate 5a, for 9 and 10, for
cis-14 and cis-15, as well as for 17 and 18).
Trioxaquine 4b: Light-yellow oil; yield: 95%; 1H NMR (250 MHz,
CDCl3): d=1.25 2.10 (m, 10H; H10, H11, H13, H14, and H12’), 2.42 (m,
1H; H12), 2.60 (m, 1H; HNC12), 2.95 (m, 3H; 1H H8 and 2H H13’),
3.29 (d, 1H; H8), 3.39 (q, 2H; H11’), 5.10 and 5.17 (2îbrs, 2î0.5H;
H5), 6.30 (m, 2H; H6 and H3’), 7.25 7.65 (m, 11H; phenyl and H6’),
7.75 and 7.83 (2îd, 3J=10.0 Hz, 2î0.5H; H5’), 7.93 (2îd, 4J=3.0 Hz,
2î0.5H; H8’), 8.01 (brs, 1H; HNC4’), 8.50 ppm (2îd, 3J=6.0 Hz, 2î
0.5H; H2’); MS (DCI/NH3+): m/z (%): 580 (5), 582 (100) [M+H]+, 583
(39), 584 (39), 585 (15).
Trioxaquine 4c: Light-yellow oil; yield: 69%; 1H NMR (250 MHz,
CDCl3): d=1.25 2.10 (m, 12H; H10, H11, H13, H14, H12’, and H13’),
2.46 (m, 1H; H12), 2.61 (m, 1H; HNC12), 2.74 (q, 2H; H14’), 3.00 (d,
1H; H8), 3.30 (m, 3H; H8 and H11’), 5.18 (2îbrs, 2î0.5H; H5), 5.96
(brs, 1H; HNC4’), 6.35 (m, 2H; H6 and H3’), 7.25 7.65 (m, 11H; phenyl
and H6’), 7.78 (2îd, 2î0.5H; H5’), 7.89 (2îd, 2î0.5H; H8’), 8.50 ppm
(2îd, 2î0.5H; H2’); MS (DCI/NH3+): m/z (%): 596 (100) [M+H]+.
When 1,3-cyclohexadiene was used, the photo-oxygenation product con-
sisted of a mixture of the desired 1,4-endoperoxide (80%) and 3-hydro-
peroxo-1,4-cyclohexadiene (20%). This mixture was then used without
purification.
Trioxane ketone 8: Chromatography: SiO2, hexane/EtOAc, 70:30, v/v;
white solid; yield: 7%. Two diastereomeric racemates 8a and 8b were
obtained and were not separated. Characterization was performed on the
1
mixture, but for clarity, the H NMR signals of each diastereomeric race-
mate are described separately. 1H NMR (250 MHz, CDCl3): 8a d=1.55
(m, 1H; H9), 1.92 (m, 1H; H9), 2.02 (m, 2H; cyclohexanone), 2.32 (m,
2H; H8), 2.41 (m, 5H; cyclohexanone), 2.68 (m, 1H; cyclohexanone),
4.25 (ddd, 3J=13.0, 8.5, and 3.5 Hz; 1H; H10), 4.43 (m, 1H; H5), 5.53
(m, 1H; H7), 5.70 ppm (m, 1H; H6); 8b d=1.70 2.70 (m, 12H; H9, H8,
and cyclohexanone), 4.33 (m, 1H; H10), 4.52 (m, 1H; H5), 5.70 (m, 1H;
H7), 5.92 ppm (m, 1H; H6); MS (DCI/NH3+): m/z (%): 242 (100)
[M+NH4]+, 243 (40), 244 (17).
Trioxaquine 9: Light-yellow oil; yield: 40%; the two diastereomeric race-
1
mates were not separated. H NMR (250 MHz, CDCl3): d=1.25 2.40 (m,
Trioxane ketone 12: Chromatography: SiO2, hexane/EtOAc, 70:30, v/v;
then SiO2, pentane/diethyl ether, 75:25, v/v. Three trioxane ketones were
separated. In order of elution they were trans-12, cis-12, and trans-13.
The ratio of purified compounds trans-12/cis-12/trans-13 was 12:78:10.
Yield of cis-12 was 26% with respect to a-terpinene.
trans-12: 1H NMR (400 MHz, CDCl3): d=1.00 (d, 3J=6.8 Hz, 6H; H13
and H14), 1.41 (s, 3H; H11), 1.63 (m, 2H; H9), 2.05 (m, 2H; cyclohexa-
none), 2.15 (m, 3J=6.8 Hz, 1H; H12), 2.24 (m, 2H; H8), 2.30 2.55 (m,
5H; cyclohexanone), 2.65 (m, 1H; cyclohexanone), 4.61 (m, 1H; H5),
5.29 ppm (m, 1H; H6); 13C NMR (100.6 MHz, CDCl3): d=15.3 (C11),
21.7 and 21.8 (C13 and C14), 26.2 (C8), 27.5 (CH2 cyclohexanone), 29.6
(C9), 34.2 (C12), 34.2, 36.8, and 37.1 (CH2 cyclohexanone), 73.1 (C5),
81.2 (C10), 103.0 (C3), 118.4 (C6), 145.3 (C7), 210.5 ppm (C17); MS
(DCI/NH3+): m/z (%): 298 (100) [M+NH4]+, 299 (27), 300 (11). Slow
evaporation of a solution of trans-12 in a mixture of hexane/EtOAc 80:20
v/v, afforded monocrystals that were subsequently analyzed by X-ray dif-
fraction.
cis-12: 1H NMR (400 MHz, CDCl3): d=1.02 and 1.04 (2îd, 3J=6.4 Hz,
6H; H13 and H14), 1.10 (brs, 3H; H11), 1.55 (m, 1H; H9), 2.03 (m, 2H;
cyclohexanone), 2.10 2.35 (m, 4H; H12, H8, and 1H cyclohexanone),
2.35 2.55 (m, 4H; cyclohexanone), 2.75 (m, 2H; 1H cyclohexanone and
1H C9), 4.06 (m, 1H; H5), 5.44 ppm (m, 1H; H6); 13C NMR (100.6 MHz,
CDCl3): d=19.8 (C11), 21.5 and 21.8 (C13 and C14), 25.7 (C9), 26.4
(C8), 27.8 (CH2 cyclohexanone), 34.2 (CH2 cyclohexanone), 35.0 (C12),
37.0 and 37.2 (CH2 cyclohexanone), 68.3 (C5), 79.3 (C10), 101.4 (C3),
116.4 (C6), 150.5 (C7), 210.8 ppm (C17); MS (DCI/NH3+): m/z (%): 298
(100) [M+NH4]+, 299 (22), 300 (7). Slow evaporation of a solution of
isomer cis-12 in a mixture of hexane/EtOAc 80:20 v/v, afforded mono-
crystals that were subsequently analyzed by X-ray diffraction.
trans-13: 1H NMR (400 MHz, CDCl3): d=1.35 (s, 3H; H11), 1.74 (m,
1H; H9), 1.94 (m, 4H; 1H H9 and H14), 2.01 and 2.08 (m, 4H; H19 and
H15), 2.16 (m, 1H; H8), 2.42 (m, 1H; H8), 2.51 (m, 4H; H18 and H16),
4.34 (d, 3J=4.4 Hz, 1H; H5), 5.03 (s, 1H; H13), 5.13 (s, 1H; H13),
5.89 ppm (d, 3J=4.4 Hz, 1H; H6); 13C NMR (100.6 MHz, CDCl3): d=
21.0 (C14), 24.1 (C8), 24.5 (C11), 32.9 (C9), 36.0 and 37.2 (C19 and C15),
38.6 and 38.8 (C18 and C16), 78.1 (C5), 79.0 (C10), 107.4 (C3), 114.2
12H; H10, H11, H13, H14, H8, and H9), 2.65 (m, 2H; HNC13 and H13),
3.00 (m, 2H; H12’), 3.27 (m, 2H; H11’), 4.05 4.25 (m, 1H; H10), 4.42
(m, 1H; H5), 5.45 5.90 (m, 2H; H6 and H7), 6.03 (brs, 1H; HNC4’),
6.32 (m, 1H; H3’), 7.30 (m, 1H; H6’), 7.66 (m, 1H; H5’), 7.89 (m, 1H;
H8’), 8.46 ppm (m, 1H; H2’); MS (DCI/NH3+): m/z (%): 430 (100)
[M+H]+, 431 (29), 432 (45).
Trioxaquine cis-14: Light-yellow oil; yield: 60%; MS (DCI/NH3+): m/z
(%): 486 (100) [M+H]+, 487 (36), 488 (42), 489 (12), 490 (2).
Separation of the two diastereomeric racemates of cis-14: The 50:50 mix-
ture of trans,cis-14 and cis,cis-14 was diluted (12 gLꢀ1) at room tempera-
ture in a mixture of dichloromethane/cyclohexane (1:8, v/v). Upon cool-
ing to ꢀ208C, a 70:30 mixture of trans,cis-14 and cis,cis-14 precipitated.
The precipitate was recovered by filtration. Further purification on silica
gel (EtOAc/triethylamine, 95:5, v/v) provided quantitative recovery of
the pure trans,cis-14, which eluted first. cis,cis-14 was then eluted, but
chromatography resulted in the loss of about 65% of this latter com-
pound. The supernatant solution, which contained a mixture of trans,cis-
14 and cis,cis-14 (30:70), was evaporated and the two diastereomeric
racemates were purified in the same manner as the precipitate. The over-
all yield of trioxaquine cis-14 isolated, as calculated from the commercial-
ly available a-terpinene starting material was 20%.
trans,cis-14: 1H NMR (400 MHz, CDCl3): d=1.02 (d, 3J=6.9 Hz, 3H;
H14), 1.04 (d, 3J=6.9 Hz, 3H; H13), 1.05 (brs, 3H; H11), 1.52 (m, 2H;
1H H9 and 1H cyclohexyl), 1.59 (m, 2H; cyclohexyl), 1.70 1.90 (m, 5H;
cyclohexyl), 2.10 2.30 (m, 4H; H8, 1H H12, and HNC17), 2.60 2.80 (m,
2H; H17 and 1H H9), 3.07 (m, 2H; H12’), 3.36 (m, 2H; H11’), 3.98 (brs,
1H; H5), 5.42 (d, 1H; H6), 6.20 (brs, 1H; HNC4’), 6.35 (d, 3J=5.4 Hz,
3
4
3
1H; H3’), 7.35 (dd, J=8.9 and J=2.1 Hz, 1H; H6’), 7.76 (d, J=8.9 Hz,
1H; H5’), 7.91 (d, 4J=2.1 Hz, 1H; H8’), 8.48 ppm (d, 3J=5.4 Hz, 1H;
H2’); 13C NMR (100.6 MHz, CDCl3): d=19.8 (C11), 21.5 (C13), 21.8
(C14), 25.7 (C9), 26.6 (C8), 28.0 35.0 (C15, C16, C18, and C19), 35.0
(C12), 42.8 (C11’), 45.1 (C12’), 55.3 (C17), 67.4 (C5), 79.0 (C10), 99.5
(C3’), 102.5 (C3), 116.8 (C6), 117.7 (C10’), 122.1 (C5’), 125.7 (C6’),
128.6 (C8’), 135.3 (C7’), 149.1 (C9’), 150.0 (C7), 150.4 (C4’), 152.1 ppm
(C2’).
1634
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 1625 1636

Synthesis of Trioxaquine Derivatives
1625 1636
Slow evaporation of a solution of trans,cis-14 in a mixture of EtOH/
EtOAC 50:50 v/v afforded monocrystals that were subsequently analyzed
by X-ray diffraction.
Trioxaquine citrate cis-15: White powder (quantitative yield from cis-14);
1H NMR (300 MHz, [D6]DMSO): d=0.97 (m, 9H; H11, H13, and H14),
2
1.10 2.50 (m, 13H; H8, H9, H12, H15, H16, H18, and H19), 2.57 (d, J=
15.2 Hz, 4H; citric acid/citrate), 2.67 (d, 2J=15.2 Hz, 4H; citric acid/cit-
rate), 3.24 (m, 3H; H12’ and HC17), 3.69 (m, 2H; H11’), 3.98 and 4.06
(2îbrs, 1H; H5), 5.32 (m, 1H; H6), 6.70 (d, 3J=5.8 Hz, 1H; H3’), 7.59
cis,cis-14: 1H NMR (300 MHz, CDCl3): d=0.98 (d, 3J=6.9 Hz, 3H;
H14), 0.99 (d, 3J=6.9 Hz, 3H; H13), 0.99 (brs, 3H; C11), 1.30 1.60 (m,
4H; 1H C9 and 3H cyclohexyl), 1.70 1.90 (m, 5H; cyclohexyl), 2.00 2.30
(m, 4H; H8, H12, and HNC17), 2.40 2.70 (m, 2H; 1H H9 and H17), 2.97
(m, 2H; H12’), 3.24 (m, 2H; H11’), 4.02 (brs, 1H; H5), 5.37 (d, 1H; H6),
6.12 (brs, 1H; HNC4’), 6.30 (d, 3J=5.4 Hz, 1H; H3’), 7.29 (dd, 3J=9.0
and 4J=2.1 Hz, 1H; H6’), 7.68 (d, 3J=9.0 Hz, 1H; H5’), 7.88 (d, 4J=
2.1 Hz, 1H; H8’), 8.44 ppm (d, 3J=5.4 Hz, 1H; H2’); 13C NMR
(75.4 MHz, CDCl3): d=20.5 (C11), 21.0 (C13), 21.2 (C14), 23.0 35.0 (C8,
C9, C15, C16, C18H2, and C19), 34.5 (C12), 42.4 (C11’), 44.6 (C12’), 55.4
(C17), 67.2 (C5), 78.0 (C10), 98.9 (C3’), 102.0 (C3), 116.3 (C6), 117.3
(C10’), 121.8 (C5’), 125.1 (C6’), 127.9 (C8’), 134.7 (C7’), 148.7 (C9’), 149.5
(C7), 150.1 (C4’), 151.6 ppm (C2’).
3
4
4
(dd, J=9.0 and J=2.1 Hz, 1H; H6’), 7.88 (d, J=2.1 Hz, 1H; H8’), 8.29
(d, 3J=9.0 Hz, 1H; H5’), 8.50 ppm (d, 3J=5.8 Hz, 1H; H2’); MS (ES+):
m/z 486.2 (monoprotonated base); elemental analysis calcd (%) for
(C39H52ClN3O17, 1H2O): C 52.73, H 6.13, N 4.73; found: C 52.66, H 6.30,
N 4.52.
Trioxaquine citrates trans,cis-15 and cis,cis-15 were prepared in the same
way, that is, from the separated trioxaquine diastereoisomers trans,cis-14
and cis,cis-14, respectively.
Trioxaquine citrate 18: Trioxaquine 17 (164 mg) was dissolved in acetone
(4 mL) and a solution of 2.0 equivalents of citric acid (120 mg) in acetone
(4 mL) was added with stirring. The trioxaquine citrate 18 was obtained
by evaporation of the solvent. Orange powder (quantitative yield from
17); 1H NMR (300 MHz, [D6]DMSO): d=0.96 (m, 9H; H11, H13, and
Slow diffusion of diethyl ether into a dichloromethane solution afforded
monocrystals that were subsequently analyzed by X-ray diffraction.
Trioxaquine 17: Column chromatography: SiO2, hexane/EtOAc/Et3N,
3
H14), 1.23 (d, J=6.3 Hz, 3H; H12’), 1.30 2.30 (m, 16H; H14’, H15’, H8,
60:35:5, v/v/v). Although several stereoisomers were present, they were
H9, H15, H16, H18, and H19), 2.57 (d, 2J=15.2 Hz, 4H; citric acid/cit-
rate), 2.66 (d, 2J=15.2 Hz, 4H; citric acid/citrate), 2.94 (m, 2H; H16’),
3.07 (m, 1H; H17), 3.37 (brs, 1H; HN), 3.70 (m, 1H; H13’), 3.83 (s, 3H;
H11’), 4.00 and 4.04 (2îbrs, 2î0.5H; H5), 5.33 (d, 1H; H6), 6.18 (d,
3J=8.9 Hz, 1H; HNC8’), 6.31 (d, 4J=2.4 Hz, 1H; H7’), 6.50 (d, 4J=
2.4 Hz, 1H; H5’), 7.44 (dd, 3J=8.3 and 3J=4.2 Hz, 1H; H3’), 8.09 (dd,
3J=8.3 and 4J=1.6 Hz, 1H; H4’), 8.30 (brs, HNC17), 8.55 ppm (dd, 3J=
4.2 and 4J=1.6 Hz, 1H; H2’); MS (ES+): m/z 524.4 (monoprotonated
base); elemental analysis calcd (%) for (C43H61N3O18, 1H2O, 1 C3H6O): C
56.15, H 7.07, N 4.27; found: C 55.74, H 6.78, N 4.12.
1
not separated. Light-yellow oil; yield: 63%; H NMR (250 MHz, CDCl3):
d=0.90 1.10 (m, 9H; H13, H14, and H11), 1.28 (2îd, 3J=6.3 Hz, 3H;
H12’), 1.30 1.90 (m, 13H; 1H H9, H15, H16, H18, H19, H14’, and H15’),
2.00 2.30 (m, 3H; H8 and H12), 2.40 2.80 (m, 5H; 1H H9, H16’, H17,
and HNC17), 3.60 (m, 1H; H13’), 3.86 (s, 3H; H11’), 3.98 (2îbrs, 1H;
H5), 5.39 (brs, 1H; H6), 6.03 (d, 3J=8.1 Hz, 1H; HNC8’), 6.27 (d, 4J=
2.4 Hz, 1H; H7’), 6.30 (d, 4J=2.4 Hz, 1H; H5’), 7.28 (dd, 3J=4.2 and
3
3
3J=8.3 Hz, 1H; H3’), 7.90 (brd, J=8.0 Hz, 1H; H4’), 8.50 ppm (dd, J=
4.2 and 4J=1.2 Hz, 1H; H2’); MS (DCI/NH3+): m/z (%): 524 (100)
[M+H]+, 525 (62), 526 (17), 527 (2).
Biological activity of the trioxaquines
Synthesis of the trioxaquine citrates: In a typical procedure,[8] the rele-
vant trioxaquine (20 200 mg) was solubilized in acetone (0.5 5 mL)
before a solution of anhydrous citric acid (2.5 mol equiv) in acetone
(0.5 5 mL) was added. The trioxaquine dicitrate spontaneously precipi-
tated, and after being centrifuged, the precipitate was washed with dieth-
yl ether and dried under vacuum. Elemental analysis and NMR spectro-
scopy confirmed the presence of two citric acid or citrate molecules per
trioxaquine. In DMSO solution, monoprotonation of the trioxaquine was
attested by the chemical shifts of the quinoline and diaminoalkane tether
protons. Proton numbering is reported in Figure 1.
Biological materials: Four strains of P. falciparum were cultured accord-
ing to a modified Trager and Jensen method in a 5% CO2 atmosphere at
378C.[31, 32] The chloroquine-sensitive strains, Nigerian (IC50 =62 nm) and
F32-Tanzania (IC50 =25 nm), and both chloroquine-resistant strains,
FcB1-Columbia (IC50 =116 nm) and FcM29-Cameroon (IC50 =174 nm),
were chosen for this study. The parasites were maintained in vitro in
human red-blood cells (Oꢁ) that were diluted to 1% hematocrit in
RPMI 1640 medium (BioMedia, Boussens, France) and supplemented
with 25 mm Hepes complemented with 5% human AB serum (Centre de
Transfusion Sanguine, Toulouse, France). Parasite cultures tested were
not synchronized in vitro. Male Swiss albino mice, which weighed 30
40 g, were obtained from C.E.R Janvier (France). P. vinckei petteri
(279BY) was provided by Dr. I. Landau (Museum National d’Histoire
Naturelle, Paris, France).[33]
Trioxaquine citrate 5b: White powder (quantitative yield from 4b);
1H NMR (250 MHz, [D6]DMSO): d=1.40 2.20 (m, 10H; H12’, H10,
2
H11, H13, and H14), 2.65 (d, J=15.1 Hz, 4H; citric acid/citrate), 2.76 (d,
2J=15.1 Hz, 4H; citric acid/citrate), 3.00 3.80 (m, 7H; H8, H11’, H13’,
and H12), 5.43 (2îbrs, 1H; H5), 6.61 (2îdd, 1H; H6), 6.72 (2îd, 1H;
H3’), 7.45 (m, 6H; phenyl), 7.65 (m, 5H; 4H phenyl and H6’), 7.93 (2îd,
1H; H8’), 8.42 (2îd, 1H; H5’), 8.60 ppm (2îd, 1H; H2’); MS (ES+): m/
In vitro antimalarial activity: The antiplasmodial activity of the trioxa-
quines was evaluated by the radioactive microdilution method described
by Desjardins et al. and modified as follows.[34,35] Drug dilutions were
tested several times in triplicate in 96-well plates (TPP, Switzerland) that
contained cultures at various stages of 1% parasitemia, and that had a
1% hematocrit.[35] For each test, the plates of parasite culture were incu-
bated with drugs at decreasing concentrations for 48 h, and radioactive
hypoxanthine was added to the medium 24 h after the beginning of incu-
bation. The stock solutions of the trioxaquines (5 mgmLꢀ1), primaquine,
and artemisinin (1 mgmLꢀ1) were prepared in DMSO (Acros Organics,
Belgium), while the stock solution of chloroquine was prepared in RPMI
1640 (1 mgmLꢀ1). All dilutions were done in RPMI 1640, and it was en-
sured that the trioxaquines did not reprecipitate under these conditions.
Parasite growth was estimated by [3H]hypoxanthine (Amersham Pharma-
cia Biotech, France) incorporation. Concentrations of trioxaquines that
inhibited 50% of the parasite growth (IC50) were graphically determined
by plotting the drug concentration versus percent of parasite growth in-
hibition at 48 h of incubation.[36] The IC50 values given in the text repre-
sent the mean value of 2 5 independent experiments (as mentioned in
the footnotes of Table 2). The chloroquine diphosphate, artemisinin, and
primaquine diphosphate (Sigma Aldrich, France) sensitivities for the four
strains were routinely tested.
z
582.3 (monoprotonated base); elemental analysis calcd (%) for
(C47H52ClN3O17, 1H2O): C 57.35, H 5.53, N 4.27; found: C 57.09, H 5.20,
N 4.24.
Trioxaquine citrate 5c: White powder (quantitative yield from 4c);
1H NMR (250 MHz, [D6]DMSO): d=1.50 2.10 (m, 12H; H12’, H13’,
H10, H11, H13, and H14), 2.65 (d, 4H; citric acid/citrate), 2.76 (d, 4H;
citric acid/citrate), 3.10 3.70 (m, 5H; H8, H14’, and H12), 3.50 (m, 2H,
H11’), 5.43 (2îbrs, 1H; H5), 6.61 (2îdd, 1H; H6), 6.75 (2îd, 1H; H3’),
7.53 (m, 6H; phenyl), 7.71 (m, 5H; 4H phenyl and H6’), 7.96 (m, 2H;
H8’ and HNC4’), 8.47 (2îd, 1H; H5’), 8.57 ppm (2îd, 1H; H2’); MS
(ES+): m/z 596.2 (monoprotonated base); elemental analysis calcd (%)
for (C48H54ClN3O17, 1H2O): C 57.74, H 5.65, N 4.21; found: C 57.91, H
5.59, N 4.33.
Trioxaquine citrate 10: White powder (quantitative yield from 9);
1H NMR (250 MHz, [D6]DMSO): d=1.50 2.60 (m, 12H; H8, H9, H11,
H12, H14, and H15), 2.68 (d, 4H; citric acid/citrate), 2.78 (d, 4H; citric
acid/citrate), 3.35 (m, 3H, H11’ and H13), 3.77 (m, 2H; H12’), 4.05 4.40
(m, 1H; H10), 4.50 4.65 (m, 1H; H5), 5.55 6.10 (m, 2H; H6 and H7),
6.78 (m, 1H; H3’), 7.70 (m, 1H; H6’), 7.98 (m, 1H; H8’), 8.38 (m, 1H;
H5’), 8.63 ppm (m, 1H; H2’); MS (ES+): m/z 430.3 (monoprotonated
base); elemental analysis calcd (%) for (C35H44ClN3O17, 1H2O): C 50.51,
H 5.57, N 5.05; found: C 50.87, H 5.20, N 5.09.
In vivo antimalarial activity: In vivo antimalarial activity was determined
against the rodent strain P. vinckei petteri according to a modified version
of the four-day Peters× suppressive test.[37] Swiss mice were inoculated in-
1635
Chem. Eur. J. 2004, 10, 1625 1636
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

B. Meunier et al.
FULL PAPER
travenously with 107 parasitized red-blood cells (resuspended in 0.2 mL
saline medium). Thereafter, the drug was administered to the animals
once daily in DMSO for four consecutive days. Drug treatment was initi-
ated 24 h after parasite inoculation to study the curative properties of the
compounds and was administered intraperitoneally (100 mL) or orally
(100 mL). Parasitemia levels were determined on the day following the
last treatment. ED50, which is the dose that leads to 50% parasite growth
inhibition in comparison to the control test (treated with an equal
volume of vehicle) was evaluated from a plot of activity (expressed as
% of control) versus the log dose. For each dose, three animals were
treated, while six control animals were considered. Treatment was consid-
ered as curative when no parasites were detected after 60 days.[37]
[5] A. Robert, B. Meunier, J. Am. Chem. Soc. 1997, 119, 5968 5969.
[6] A. Robert, O. Dechy-Cabaret, J. Cazelles, B. Meunier, Acc. Chem.
Res. 2002, 35, 167 174.
[7] A. Robert, J. Cazelles, B. Meunier, Angew. Chem. 2001, 113, 2008
2011; Angew. Chem. Int. Ed. 2001, 40, 1954 1957.
[8] O. Dechy-Cabaret, F. Benoit-Vical, A. Robert, B. Meunier, Chem-
BioChem 2000, 1, 281 283.
[9] L. K. Basco, O. Dechy-Cabaret, M. Ndounga, F. S. Meche, A.
Robert, B. Meunier, Antimicrob. Agents Chemother. 2001, 45, 1886
1888.
[10] O. Dechy-Cabaret, A. Robert, B. Meunier, C. R. Chim. 2002, 5,
297 302.
[11] C. W. Jefford, S. Kohmoto, D. Jaggi, G. Timari, J.-C. Rossier, M.
Rudaz, O. Barbuzzi, D. Gÿrard, U. Burger, P. Kamalaprija, J.
Mareda, G. Bernardinelli, I. Manzanares, C. J. Canfield, S. L. Fleck,
B. L. Robinson, W. Peters, Helv. Chim. Acta 1995, 78, 647 662.
[12] D. De, F. M. Krogstad, F. B. Cogswell, D. J. Krogstad, Am. J. Trop.
Med. Hyg. 1996, 55, 579 583.
[13] R. G. Ridley, W. Hofheinz, H. Matile, C. Jaquet, A. Dorn, R. Mas-
ciadri, S. Jolidon, W. F. Richter, A. Guenzi, M. A. Girometta, H. Ur-
wyler, W. Huber, S. Thaithong, W. Peters, Antimicrob. Agents Che-
mother. 1996, 40, 1846 1854.
[14] D. De, F. M. Krogstad, L. D. Byers, D. J. Krogstad, J. Med. Chem.
1998, 41, 4918 4926.
[15] N. L. Drake, J. R. Adams, J. Am. Chem. Soc. 1939, 61, 1326 1329.
[16] Y. Pollack, R. Segal, J. Golenser, Parasitol. Res. 1990, 76, 570 572.
[17] E. S. Eliel, S. H. Wilen, Stereochemistry of Organic Compounds,
Wiley, New-York, 1994, p. 119.
Genotoxicity evaluation: An exponential culture of the Escherichia coli
modified strain GE864 was grown in an M9 liquid medium (6 gLꢀ1
Na2HPO4, 3 gLꢀ1 KH2PO4, 0.5 gLꢀ1 NaCl, 1 gLꢀ1 NH4Cl autoclaved for
20 min at 1208C, then supplemented with 0.1 mm CaCl2, 1 mm MgSO4,
1 mm 0.2% glucose, 1 mgmLꢀ1 thiamine, and 0.1% ™casaminoacid∫). This
E. coli culture was treated with cis-15 at concentrations of 5 20 mm in
DMSO or with 3 mm mitomycin C in H2O. The culture was incubated at
378C and aliquots were withdrawn after 0, 30, 60, 90, and 120 min of in-
cubation. The absorbance of the aliquots was measured at 600 nm to
evaluate the quantity of bacteria. Bacteria were lysed with chloroform
and
a phosphate buffer that contained b-mercaptoethanol (0.06m
Na2HPO4, 0.04m NaH2PO4, 0.01m KCl, 0.001m MgSO4, 0.05m b-mercap-
toethanol). To measure the b-galactosidase activity the lysed cultures
were incubated at 288C and o-nitrophenyl-b-d-galactopyranoside
(ONPG) was added. After a determined time of incubation, Na2CO3
(1m) was added to quench the reaction, the bacterial fragments were cen-
trifuged, and absorbance of the supernatant at 420 nm was measured to
quantify the o-nitrophenate. The enzymatic activity u was calculated ac-
cording to the following formula and plotted against time: u=1000î
OD(420 nm)/reaction timeîreaction volumeîOD(600 nm).[28]
[18] P. M. O’Neill, P. G. Bray, S. R. Hawley, S. A. Ward, B. K. Park, Phar-
macol. Ther. 1998, 77, 29 58.
[19] B. Lell, J. F. Faucher, M. A. Missinou, S. Borrmann, O. Dangelmaier,
J. Horton, P. G. Kremsner, Lancet 2000, 355, 2041 2045.
[20] C. A. Lipinski, F. Lombardo, B. W. Dominy, P. J. Feeney, Adv. Drug
Delivery Rev. 2001, 46, 3 26.
[21] L. K. Basco, J. Bickii, P. Ringwald, Antimicrob. Agents Chemother.
1998, 42, 2347 2351.
[22] J. L. Vennerstrom, Y. Dong, S. L. Andersen, A. L. Ager Jr. , H. Fu,
R. E. Miller, D. L. Wesche, D. E. Kyle, L. Gerena, S. M. Walters,
J. K. Wood, G. Edwards, A. D. Holme, W. G. McLean, W. K. Mil-
hous, J. Med. Chem. 2000, 43, 2753 2758.
[23] C. Wongsrichanalai, T. Wimonwattrawatee, P. Sookto, A. Laoboon-
chai, D. G. Heppner, D. E. Kyle, W. H. Wernsdorfer, Bull. W. H. O.
1999, 77, 392 398.
[24] C. Jaquet, H. R. Stohler, J. Chollet, W. Peters, Trop. Med. Parasitol.
1994, 45, 266 271.
[25] G. H. Posner, J. N. Cumming, S. H. Woo, P. Ploypradith, S. Xie, T. A.
Shapiro, J. Med. Chem. 1998, 41, 940 951.
[26] M. Hofnung, P. Quillardet, Ann. N. Y. Acad. Sci. 1988, 534, 817
825.
[27] C. J. Kenyon, G. C. Walker, Proc. Natl. Acad. Sci. USA 1980, 77,
2819 2823.
[28] F. Larminat, C. Cazaux, M. Germanier, M. Defais, J. Bacteriol. 1992,
174, 6264 6269.
[29] G. M. Sheldrick, Acta Crystallogr. Sect. A 1990, 46, 467 473.
[30] G. M. Sheldrick, SHELXL-97, Program for Crystal Structure Re-
finement, University of Gˆttingen, Gˆttingen (Germany) 1997.
[31] W. Trager, J. B. Jensen, Science 1976, 193, 673 675.
[32] W. Van Huyssen, K. H. Rieckmann, Trop. Med. Parasitol. 1993, 44,
329 330.
[33] R. Carter, D. Walliker, Ann. Trop. Med. Parasitol. 1975, 69, 187
196.
[34] R. E. Desjardins, C. J. Canfield, J. D. Haynes, J. D. Chulay, Antimi-
crob. Agents Chemother. 1979, 16, 710 718.
[35] F. Benoit-Vical, A. Robert, B. Meunier, Antimicrob. Agents Chemo-
ther. 1999, 43, 2555 2558.
[36] F. Benoit, A. Valentin, Y. Pelissier, C. Marion, Z. Dakuyo, M.
Mallie, J. M. Bastide, Trans. R. Soc. Trop. Med. Hyg. 1995, 89, 217
218.
Acknowledgements
The authors are grateful to Dr. Martine Defais (IPBS-CNRS, Toulouse)
for making possible the SOS repair experiments in her laboratory and
also for the fruitful discussions on this topic. They are also grateful to
Prof. Jean Bernadou (LCC-CNRS) and Dr. Antoine Berry (CHU Ran-
gueil) for many fruitful discussions, to Dr. JÿrÙme Cazelles (LCC-CNRS
and, now, Palumed) and to Christine Salles (Palumed) for the prepara-
tion of several of the trioxaquine derivatives, to Dr. Yannick Coppel
(LCC-CNRS) for discussions about NMR analyses, and to Suzy Richelme
and Catherine Claparols (MS service FR14-LCC, Toulouse) for mass
spectrometry analyses. Jean-Claude Rives (CHU Rangueil), Angÿlique
Erraud, and Carine Duboe (Palumed) are acknowledged for technical as-
sistance with the in vivo and in vitro experiments. Dr. Bernard Malfroy
(Eukarion, Boston) is gratefully acknowledged for preliminary toxicologi-
cal evaluation of trioxaquine DU1302c on uninfected mice. The financial
support of Palumed is gratefully acknowledged. The trioxaquines are cur-
rently being studied by Palumed and Sanofi-Synthÿlabo. CNRS
(GDR1077) and DGA (contract no. 00CO040) are also gratefully ac-
knowledged for financial support.
[1] a) N. J. White, F. Nosten, S. Looareesuwan, W. M. Watkins, K.
Marsh, R. W. Snow, G. Kokwaro, J. Ouma, T. T. Hien, M. E. Moly-
neux, T. E. Taylor, C. I. Newbold, T. K. Ruebush, M. Danis, B. M.
Greenwood, R. M. Anderson, P. Olliaro, Lancet 1999, 353, 1965
1967; b) R. G. Ridley, Nature 2002, 415, 686 693.
[2] C. Wongsrichanalai, A. L. Pickard, W. H. Wernsdorfer, S. R. Mesh-
nick, Lancet Infect. Dis. 2002, 2, 209 218.
[3] a) F. Nosten, P. Brasseur, Drugs 2002, 62, 1315 1329; b) I. M. Hast-
ings, W. M Watkins, N. J. White, Philos. Trans. R. Soc. London, Ser.
B 2002, 357, 505 519.
[4] a) S. R. Meshnick, T. E. Taylor, S. Kamchonwongpaisan, Microbiol.
Rev. 1996, 60, 301 315; b) G. H. Posner, H. B. Jeon, M. H. Parker,
M. Krasavin, I. H. Paik, T. A. Shapiro, J. Med. Chem. 2001, 44,
3054 3058.
[37] W. Peters, J. H. Portus, B. L. Robinson, Ann. Trop. Med. Parasitol.
1975, 69, 155 171.
Received: September 25, 2003 [F5576]
1636
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 1625 1636