Antimalarial activity of compounds interfering with Plasmodium falciparum phospholipid metabolism: Comparison between mono- and bisquaternary ammonium salts

Article abstract of DOI:10.1021/jm9911027

On the basis of a previous structure - activity relationship study, we identified some essential parameters, e.g. electronegativity and lipophilicity, required for polar head analogues to inhibit Plasmodium falciparum phospholipid metabolism, leading to parasite death. To improve the in vitro antimalarial activity, 36 cationic choline analogues consisting of mono-, bis-, and triquaternary ammonium salts with distinct substituents of increasing lipophilicity were synthesized. For monoquaternary ammonium salts, an increase in the lipophilicity around nitrogen was beneficial for antimalarial activity: IC₅₀ decreased by I order of magnitude from trimethyl to tripropyl substituents. Irrespective of the polar head substitution (methyl, ethyl, hydroxyethyl, pyrrolidinium), increasing the alkyl chain length from 6 to 12 methylene groups always led to increased activity. The highest activity was obtained for the N,N,N-tripropyl-N-dodecyl substitution of nitrogen (IC₅₀ 33 nM). Beyond 12 methylene groups, the antimalarial activities of the compounds decreased slightly. The structural requirements for bisquaternary ammonium salts in antimalarial activity were very similar to those of monoquaternary ammonium salts, i.e. polar head steric hindrance and lipophilicity around nitrogen (methyl, hydroxyethyl, ethyl, pyrrolidinium, etc.). In contrast, with bisquaternary ammonium salts, increasing the lipophilicity of the alkyl chain between the two nitrogen atoms (from 5 to 21 methylene groups) constantly and dramatically increased the activity. Most of these duplicated molecules had activity around 1 nM, and the most lipophilic compound synthesized exhibited an IC₅₀ as low as 3 pM (21 methylene groups). Globally, this oriented synthesis produced 28 compounds out of 36 with an IC₅₀ lower than 1 μM, and 9 of them had an IC₅₀ in the nanomolar range, with I compound in the picomolar range. This indicates that developing a pharmacological model for antimalarial compounds through choline analognes is a promising strategy.

Full text of DOI:10.1021/jm9911027

J . Med. Chem. 2000, 43, 505-516
505
An tim a la r ia l Activity of Com p ou n d s In ter fer in g w ith P la sm od iu m fa lcipa r u m
P h osp h olip id Meta bolism : Com p a r ison betw een Mon o- a n d Bisqu a ter n a r y
Am m on iu m Sa lts‡
,
†
#
†
†
†
Mich e` le Calas,* Marie L. Ancelin, G e´ rard Cordina, Philippe Portefaix, Gilles Piquet,
†
#
Val e´ rie Vidal-Sailhan, and Henri Vial
Laboratoire des Aminoacides, Peptides et Proteines, CNRS, UMR 5810, CP 22, Universit e´ de Montpellier II,
Place E. Bataillon, 34095 Montpellier Cedex 5, France, and CNRS, UMR 5539, CP 107, Universit e´ de Montpellier II,
Place E. Bataillon, 34095 Montpellier Cedex 5, France
Received J une 22, 1999
On the basis of a previous structure-activity relationship study, we identified some essential
parameters, e.g. electronegativity and lipophilicity, required for polar head analogues to inhibit
Plasmodium falciparum phospholipid metabolism, leading to parasite death. To improve the
in vitro antimalarial activity, 36 cationic choline analogues consisting of mono-, bis-, and
triquaternary ammonium salts with distinct substituents of increasing lipophilicity were
synthesized. For monoquaternary ammonium salts, an increase in the lipophilicity around
nitrogen was beneficial for antimalarial activity: IC50 decreased by 1 order of magnitude from
trimethyl to tripropyl substituents. Irrespective of the polar head substitution (methyl, ethyl,
hydroxyethyl, pyrrolidinium), increasing the alkyl chain length from 6 to 12 methylene groups
always led to increased activity. The highest activity was obtained for the N,N,N-tripropyl-
N-dodecyl substitution of nitrogen (IC50 33 nM). Beyond 12 methylene groups, the antimalarial
activities of the compounds decreased slightly. The structural requirements for bisquaternary
ammonium salts in antimalarial activity were very similar to those of monoquaternary
ammonium salts, i.e. polar head steric hindrance and lipophilicity around nitrogen (methyl,
hydroxyethyl, ethyl, pyrrolidinium, etc.). In contrast, with bisquaternary ammonium salts,
increasing the lipophilicity of the alkyl chain between the two nitrogen atoms (from 5 to 21
methylene groups) constantly and dramatically increased the activity. Most of these dupli-
cated molecules had activity around 1 nM, and the most lipophilic compound synthesized
exhibited an IC50 as low as 3 pM (21 methylene groups). Globally, this oriented synthesis
produced 28 compounds out of 36 with an IC50 lower than 1 µM, and 9 of them had an IC50 in
the nanomolar range, with 1 compound in the picomolar range. This indicates that developing
a pharmacological model for antimalarial compounds through choline analogues is a promising
strategy.
In tr od u ction
biogenesis and parasite growth. More particularly, phos-
phatidylcholine (PC) constitutes the major PL of infected
erythrocytes (about 50% of total PL). PC is synthesized
from polar heads, mainly choline, drawn from plasma,
and using the parasite enzymatic machinery.4
Malaria has reached epidemic proportions because of
the increasing population at risk of the disease, difficul-
ties in eradicating the mosquito vector in the tropics,
and increased resistance to the most commonly used
antimalarial drugs.1,2 The high parasite cross-resistance
to compounds that are not even structurally or phar-
macologically related complicates the problem. New
drugs with novel mechanisms of action are thus ur-
-6
Systematic screening of 80 polar head analogues
allowed us to demonstrate that interference with de
novo PC biosynthesis from choline is lethal to the
parasite, even against chloroquine-resistant parasites.
In contrast, these compounds were weakly active in
vitro against other eukaryotic cells, suggesting high
3
gently required.
In this context, only the identification of original
target(s) could lead to novel antimalarial compounds
which would a priori not allow cross-resistance with
preexisting antimalarials. The phospholipid (PL) me-
tabolism of infected erythrocytes is an effective phar-
macological target for a new chemotherapy approach
because of its specificity and importance for membrane
6
parasite selectivity. The probable target for these
analogues is choline carrier which is a rate-limiting step
7
in PC biosynthesis. This is notably shown by a specific
inhibition of de novo PC biosynthesis, as indicated by
the early effect on PC biosynthesis and the very close
correlation between PL antimetabolic and antimalarial
activities. The compounds are specific to mature para-
sites (trophozoites), i.e. the most intense phase of PL
*
To whom correspondence should be addressed. Tel: 04 67 14 38
1
7. Fax: 04 67 14 48 66. E-mail: mcalas@univ-montp2.fr.
5,6
†
biosynthesis during the erythrocytic cycle.
CNRS, UMR 5810, CP 22.
CNRS, UMR 5539, CP 107.
Abbreviations: PC, phosphatidylcholine; PL, phospholipid; (Q)-
#
Kinetic properties of the choline carrier have been
investigated. It is asymmetric and functions according
to a cyclic model with Kt for choline around 10 µM and
‡
SAR, (quantitative) structure-activity relationship; IC50, concentration
resulting in 50% inhibition of parasite growth.
1
0.1021/jm9911027 CCC: $19.00 © 2000 American Chemical Society
Published on Web 01/26/2000

5
06 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 3
Calas et al.
Ta ble 1. Structures and Antimalarial Activities of Monoquaternary Ammonium Salts
compd
R1
R2
R3
n
R
X
IC50 (µM)
E2a
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
C2H5
C2H5
C2H5
C2H5
C3H7
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
C2H5
CH3
C2H5
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
C2H5
n C3H7
n C4H9
8
H
Φ
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
Br
5
0.62
0.5
E70
E6*
E8*
E20*
E21*
E22*
E23*
E25
E24
E27
E26
E60
E10*
E11
E12
E13*
E35
E36
E34
E37
E38
E30*
F 2a
8
12
16
12
12
12
12
6
12
14
16
6
12
16
18
12
12
12
12
12
12
12
8
Cl
Br
Br
Br
I
0.8
0.11
0.15
0.26
0.21
26
0.22
0.27
0.42
9.7
0.064
0.14
0.27
0.033
0.27
0.36
0.23
0.14
0.38
0.70
3.85
0.48
0.81
115
I
CH2-CH2-Br
-(CH2)4-
-(CH2)4-
-(CH2)4-
-(CH2)4-
C2H5
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
I
OTs
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
C2H5
C2H5
C2H5
C2H5
C3H7
CH3
CH3
CH3
CH2CHdCH2
CH2-CtCH
CH3
CH3
CH3
CH3
CH3
CH3
CH3
C2H5
CH2-CH2-OH
CH2-CH2-OH
C2H5
C2H5
C2H5
C3H7
CH2-CHdCH2
CH2-CtCH
CH2-CH2-CtCH
CH2-CHdCH2
CH2-CtCH
(CH2)11-CH3
CH2-CH2-OH
CH2-CH2-OH
CH2-CH2-OH
CH2-CH(CH3)OH
CH2-CH(CH3)OH
CH(CH3)-CH(OH)Φ
CH2-CH2-OH
(CH2)11-CH3
F 4*
F 25
F 32
F 30
F 50
F 8*
F 7*
F 11*
12
16
5
12
12
12
12
12
0.43
0.67
0.34
0.84
1.5
(CH2)11-CH3
*
The activity of these compounds has already been reported.8 The bold entries are those compounds possessing one or two dodecyl
substituents.
a Vm which is 10-fold increased after infection compared
to the noninfected erythrocyte. At the present time,
information concerning protein sequence, structural
nature, or subcellular localization is not available yet.
Resu lts
A SAR study previously performed with 80 PL polar
7
head analogues highlighted some essential parameters
for in vitro antimalarial activity, i.e. essentially elec-
Ongoing structure-activity relationship (SAR) stud-
ies using autocorrelation vectors and multidimensional
analysis indicated that among the relevant parameters
for the antimalarial activity of these polar head ana-
logues, the electronegativity and lipophilicity distribu-
6,8
tronegativity and lipophilicity. In the present work,
to improve antimalarial activity, we thus designed,
synthesized, and tested for in vitro antimalarial activity
against P. falciparum 36 new quaternary ammonium
salts with various degrees of lipophilicity. Nineteen
compounds were monoquaternary ammonium salts,
whose nitrogen atom bears one lipophilic chain (5-18
methylene groups) and three other small alkyl substit-
uents (1-4 carbon atoms), saturated or unsaturated
8
tion in the molecular space were prominent. N-Tri-
methyl compounds possessing a cationic nitrogen bear-
ing a long lipophilic chain with 10-12 methylene groups
were found to be active against Plasmodium falciparum
in vitro (IC50 in the micromolar range).
(series E, Table 1), or two small alkyls and one hydroxy-
The present studies were aimed at improving the
antimalarial activity, notably by changing the lipophi-
licity and electronegativity of cationic compounds. For
this purpose, 36 new compounds consisting of mono- or
bisquaternary ammonium salts and one tri-ammonium
with various lipophilic substituents around nitrogen or
in the alkyl chain have been synthesized. Ten analogues
out of 36 exhibited very potent in vitro antimalarial
activity (IC50 in the nanomolar range), and one of them
with high lipophilicity was active in the picomolar
range. The structure of the most active compounds
allowed us to develop a more accurate topographic mod-
el of the ligand binding site of the choline transporter,
the probable target of these drugs, which provides the
choline precursor for PC biosynthesis to the parasite.
alkyl group (series F, Table 1).
Sixteen compounds were symmetrical bisquaternary
ammonium salts, whose nitrogen atoms were linked by
a chain of 3-21 methylene groups (Table 2). For 11 of
them, other nitrogen substituents were alkyl (and/or

Compounds Interfering with P. falciparum Metabolism
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 3 507
Ta ble 2. Structures and Antimalarial Activities of Bis- and Triquaternary Ammonium Salts
compd
R1
R2
R3
n
X
IC50 (µM)
G1*
G4*
G5
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
C2H5
C2H5
C2H5
C2H5
CH3
CH3
C2H5
C2H5
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
CH3
-(CH2)4-
-(CH2)4-
-(CH2)4-
-(CH2)4-
-(CH2)4-
C2H5
6
Br
Br
Br
C4H5O6
Br
Br
Br
Br
Br
Br
Br
Br
Br
OTs
Br
Br
Br
Br
700
0.09
0.004
650
0.15
0.013
0.0011
0.00064
0.54
0.045
0.0016
3 × 10
0.0076
0.044
0.0053
0.0049
12
16
5
10
12
14
16
8
12
16
21
12
12
12
16
3
G20*
G23*
G24
G22
G25
G12
G14
G15
G19
G94
G74
G84
H5
C2H5
C2H5
C2H5
C2H5
CH3
C2H5
C2H5
C2H5
-
6
CH2-CtCH
CH2-CH2-CtCH
CH2-CHdCH2
CH2-CH2-OH
(CH2)11-CH3
(CH2)11-CH3
(CH2)11-CH3
(CH2)11-CH3
CH3
C2H5
C2H5
CH3
CH3
CH3
G40
G41
G44
G45
TA1
10.8
6
12
16
0.22
0.18
1.8
Br
I
Br
CH3
+
+
+
(C2H5)3N -(CH2)6-N (C2H5)2-(CH2)6-N (C2H5)3
800
*
The activity of these compounds has already been reported.8 The bold entries are the most active compounds in this series.
alkenyl) groups lower than 5 carbon atoms (G5-G19,
G24, G25, G22, G74-G94). For four other compounds,
each nitrogen atom bears a dodecyl substituent and two
methyl groups (G40-G45), and for another each nitro-
gen atom bears a hydroxyethyl and two ethyl groups
E), and 2-hydroxypropyldimethylammonium (plot F).
IC50 values are plotted on a logarithmic scale against
the lipophilic chain length on the nitrogen atom in
Figure 1 for each polar head. Interestingly, irrespective
of the cationic head, the plots showed a very close
pattern when the alkyl chain length increased (from 2
to 18 carbon atoms), with a sharp decrease in the IC50
and maximal inhibitory effect occurring at a carbon
chain length of 12, for all compounds. At this chain
length, the IC50 values of these compounds ranged from
(H5).
5
00 nM (for trimethylammonium, E6) to 64 nM (for
triethylammonium, E10). Beyond a chain length of 12
methylene groups, the antimalarial activities of the
compounds reached a plateau or slightly decreased,
irrespective of the cationic head.
We also synthesized compound E70 which possesses
an electron-rich phenyl ring at the end of an octyl chain,
making the substituent roughly the same length as a
dodecyl. Interestingly, the antimalarial activity of this
compound was very similar to that of the dodecyl
analogue (E70, 0.62 µM/E6, 0.5 µM).
One compound is a triammonium salt (TA1, Table 2).
Mon oqu a ter n a r y Am m on iu m Sa lts. The electro-
negativity and lipophilicity distribution in the molecular
space was modified in several ways, i.e. by lengthening
one or several alkyl chains on the nitrogen atom, by
increasing the bulk of the cationic head, by introducing
various electronegative groups (bromide or hydroxyl),
or by including nitrogen substituents with interesting
electronic density, such as unsaturations.
All of the newly synthesized monoquaternary am-
monium salts were lethal to P. falciparum in vitro, in a
dose-dependent manner with IC50 values ranging from
(
b ) Mod ifica t ion of t h e Bu lk of t h e Ca t ion ic
Hea d . We also studied the effects of varying the
nitrogen substituents on antimalarial activity. First,
these was a marked similarity in the activity patterns
of compounds shown in Figure 1 when other substitu-
ents were trimethyl (plot A), dimethylhydroxyethyl (D),
diethylhydroxyethyl (E), or dimethyl-2-hydroxypropyl
(F). With an N-methylpyrrolidinium head (C), activities
1
15 µM (F 32) to 0.14 µM (E11, E37) (Table 1).
a ) Ch a in -Len gth Mod ifica tion . The effects of
were slightly improved (at least for 12, 14, or 16
methylenes in the alkyl chain), whereas triethyl sub-
stitution (B) significantly increased the activity by 6-8-
fold as compared to the trimethyl substitution.
(
chain-length modification were studied via stepwise
increases in methylene number (from 2 to 18) in one
alkyl chain borne by the nitrogen atom. This was
performed for six different polar heads: trimethylam-
monium (plot A), triethylammonium (plot B), N-meth-
ylpyrrolidinium (plot C), 2-hydroxyethyldimethylam-
monium (plot D), 2-hydroxyethyldiethylammonium (plot
We then synthesized various analogues of monoam-
monium salts possessing one N-dodecyl substituent
which was found to be optimal for antimalarial activity
(as indicated above). The three other substituents were

5
08 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 3
Calas et al.
F igu r e 2. Antimalarial activity of N-dodecylammonium salts
(plot A), N-tetradecylammonium salts (plot B), N-hexadecyl-
F igu r e 1. Antimalarial activity of alkyltrimethylammonium
salts (plot A), alkyltriethylammonium salts (plot B), N-alkyl-
N-methylpyrrolidinium salts (plot C), alkyl-2-hydroxyeth-
yldimethylammonium salts (plot D), alkyl-2-hydroxyethyldi-
ethylammonium salts (plot E), and alkyl-2-hydroxypropyl-
dimethylammonium salts (plot F) as a function of alkyl chain
length. IC50 values against in vitro growth of P. falciparum,
expressed as µM, are from Table 1 and ref 8.
ammonium salts (plot C), N-octadecylammonium salts (plot
D), 1,12-dodecamethylenebisammonim salts (plot E), and
1,16-hexadecamethylenebisammonim salts (plot F) as a func-
tion of the polar head volume. The polar head volume was
calculated using a molecular modeling software program
(TSAR, Oxford Molecular). The best volume fitting the active
site was between 200 and 350 Å . The IC50 values against in
vitro growth of P. falciparum, expressed as µM, are from
Tables 1 and 2 and ref 8.
3
either alkyl (E20-E22, E13) and/or alkenyl (E35, E37)
or alkynyl (E36, E34, E38) of various sizes (1-4 carbon
atoms). We also synthesized bulkier ammonium salts
bearing two dodecyl groups, with the two other substit-
uents being two methyl groups (E30) or one hydroxy-
ethyl group and one short alkyl group (methyl, F 7, or
ethyl, F 11), or bearing one dodecyl group and one
arylalkyl group, with the two other substituents being
two methyl groups (F 50). All of these compounds are
shown in bold in Table 1. The IC50 values ranged from
(
E20, 0.11 µM) or a propyl (E21, 0.15 µM) group slightly
improved the activity. This was also the case when two
of the methyl groups were substituted by a cyclic
tetramethylene (E24, 0.22 µM), which sterically roughly
corresponds to the diethyl analogue. However, beyond
a certain limit, increasing the nitrogen substituent
length decreased the antimalarial activity [0.26 µM for
butyl (E22), 0.67 µM for F 50, which roughly corresponds
to a hexyl, 0.7 µM for a dodecyl substituent (E30); see
also F 7, 0.84 µM/F 4, 0.48 µM]. A further increase in
the steric hindrance at the nitrogen level also led to an
increase in the IC50 (see F 11, 1.5 µM/F 7, 0.84 µM).
Hence, for N-dodecyl-substituted monoquaternary
ammonium salts, the presence of alkyl groups bulkier
than methyl on the nitrogen atom was favorable for the
activity up to three methylenes (E24, E10, E13, or E37
1
.5 µM (F 11) to 33 nM (E13) and closely depended on
the size of the polar head. It should be noted that among
these ammonium salts bearing a dodecyl group the
activity was not linearly correlated with the volume of
the polar head, since both dodecyltrimethylammonium
(E6, 0.5 µM), bearing the smallest substituents, and
compounds bearing the bulkiest substituents (two dode-
cyl groups for E30: 0.7 µM, F 7: 0.84 µM, and F 11: 1.5
µM; and a phenyl cycle for F 50: 0.67 µM) were the least
active molecules.
8
compared to E6). This was also the case for tetradecyl,
hexadecyl (E8/E11), and octadecyl substituted mono-
quaternary ammonium salts.
8
Simultaneous substitution of the three methyl groups
It thus appears that the volume of the cationic head
meets very strict requirements to adapt to the active
site. Interestingly, plotting IC50 as a function of the polar
head volume (Figure 2) revealed a very similar pattern
with dodecyl (plot A), tetradecyl (plot B), hexadecyl (plot
C), or octadecyl (plot D) trialkylammonium. There was
(
(
(
E6, 0.5 µM) by three ethyl (E10, 64 nM) or three propyl
E13, 33 nM) groups markedly improved the activity
10-fold), with the latter being the most active in the
monoquaternary ammonium series. Progressive substi-
tution of one methyl group (E6, 0.5 µM) by an ethyl

Compounds Interfering with P. falciparum Metabolism
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 3 509
a sharp decrease in the IC50 between 100 and 250 Å3,
and then, the IC50 increased again. Clearly, in this
context, an additional dodecyl chain could not fit within
the active site and was thus excluded (see Discussion).
This active site can then be schematized as a sphere
whose volume is similar to that of an N-tripropyl head
3
(
i.e. ∼350 Å ) and therefore whose radius is ∼4 Å.
(
c) In tr od u ction of Electr on ega tive or Electr on -
Rich N-Su bstitu tion s. Finally, we modified the lipo-
philicity and electronic density distribution of the
N-dodecyl-substituted compounds by introducing un-
saturations or electronegative groups (hydroxyl or bro-
mine atom). Both modifications decreased the lipophi-
licity and modified the polarity of compounds relative
to the corresponding saturated carbon groups.
The presence of unsaturations in N-substituents,
consisting of short alkenyl or alkynyl groups (i.e. con-
taining 3 or 4 carbon atoms) did not significantly change
the activity (by less than 3-fold), regardless of whether
they were present in one or two nitrogen substituents
(compare E21/E35/E36, E37/E38, or E22/E34).
Introduction of a hydroxyl function in one N-substitu-
ent decreased the activity by 3-5-fold when dealing
with compounds whose polar heads were weakly hin-
dered (up to propyl) (see E20, 0.11 µM/F 4, 0.48 µM; E10,
6
4 nM/F 8, 0.34 µM; E21, 0.15 µM/F 30, 0.43 µM). In
contrast, this had no significant effect for compounds
possessing two long N-alkyl chains (e.g. dodecyl, see
E30/F 7). In this latter case, it seemed that the deleteri-
ous effect of the second dodecyl chain hid other possible
adverse effects on the activity.
Finally, in the alkyl chain, the presence of a bromine
atom that was less electronegative than a hydroxyl
group was only slightly detrimental (see E23/E21).
Bisqu a ter n a r y Am m on iu m Sa lts. The pharmaco-
logical concept of molecule duplication is often an
efficient way to further improve the biological activity
of active compounds. We thus designed twin drugs
containing two identical pharmacophoric groups, i.e. two
quaternary ammonium heads combined covalently in a
single molecule.
F igu r e 3. Antimalarial activity of alkylbis(trimethylammo-
nium) salts (plot A), alkylbis(triethylammonium) salts (plot
B), alkylbis(N-methylpyrrolidinium) salts (plot C), alkylbis(2-
hydroxyethyldiethylammonium) salts (plot D), and alkylbis-
(dodecyldimethylammonium) salts (plot E) as a function of the
methylene group number between the two cationic heads. IC50
values against P. falciparum, expressed as µM, are from Table
2
and ref 8.
(trimethylammonium), B (triethylammonium), C (N-
methylpyrrolidinium), and D (2-hydroxyethyldiethylam-
monium) show a significant and continuous steep
decrease in IC50 values, when the methylene group
number between the two nitrogen atoms increased. This
indicated that the antimalarial activity was closely
correlated with the chain length separating the two
ammonium groups. Even up to 21 carbon atoms, these
plots did not exhibit any plateau, contrary to the
corresponding monoquaternary ammoniums (see Figure
1). In fact, with up to 12 carbon atoms in the lipophilic
chain, the mono- and bisammonium activity plots
merged, but above, the IC50 values of monoammonium
salts reached a plateau, while the bisammonium IC50
values continued to decrease constantly, reaching 3 ×
Bisquaternary ammonium salts, whose chemical struc-
tures and IC50 values against P. falciparum are reported
in Table 2, exhibited a wide range of antiplasmodial
-
4
activities, with IC50 values ranging from 7 × 10
M
-
12
(
G1) to as low as 3 × 10
M (G19), i.e. up to 4 orders
of magnitude lower than the monoquaternary am-
monium salts (see above). As for monoquaternary am-
monium salts, we investigated the possible beneficial
effects of various modifications such as lipophilicity by
increasing the inter-nitrogen chain length or by adding
electronegative or electron-rich N-substitutions, i.e.
introducing unsaturations or hydroxyl groups on the
N-substituents of cationic heads.
-
12
10
M for 21 carbon atoms, i.e. for N,N,N,N′,N′,N′-
hexaethyl-1,21-heneicosanediaminium (G19).
(
a ) Ch a in -Len gth Mod ifica tion . As for monoqua-
On the other hand, more hindered bisammonium
compounds whose nitrogen atoms each bear a second
long lipophilic chain (dodecyl) (G40-G45, Figure 3, plot
E) behaved differently. Up to 6 carbon atoms between
the two cationic heads, they were more active than other
analogues by 3 orders of magnitude (G41, 1.22 µM/G1,
700 µM). From 12 carbon atoms in the inter-nitrogen
alkyl chain, the activity decreased, whereas the activity
of compounds with three small substituents on the
nitrogen greatly increased (G45, 1.8 µM/G5, 4 nM).
ternary ammonium salts, we synthesized five series of
compounds: trimethylammonium, triethylammonium,
N-methylpyrrolidinium, 2-hydroxyethyldiethylammo-
nium, and dimethyldodecylammonium. These polar
heads were linked by an alkyl chain of 3-21 methyl-
enes. Figure 3 summarizes the present results (Table
), supplemented with other data obtained previously
with compounds with short alkyl chains (less than 12
methylenes) and weaker antimalarial activity. Plots A
2
8

5
10 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 3
Calas et al.
Hence, for bisdodecyl compounds, the IC50 was minimal
roughly between 6 and 12 carbon atoms in the inter-
chain, followed by a slight increase, instead of a decrease
in the nanomolar range of activity. In fact, for short
polymethylene interchains (n ) 6, G41), the compounds
roughly behaved as the corresponding monoquaternary
ammonium (E6, trimethyldodecylammonium). This was
probably also the case for shorter lengths (n ) 3, G40),
but in this case the IC50 was slightly higher than that
of E6, probably due to mutual steric hindrance of the
polar heads which were too close.
Finally, we tested the antimalarial activity of one tri-
quaternary ammonium salt (TA1) in which each nitro-
gen atom was separated by 6 methylene groups. It was
found to be far less active (IC50 800 µM) than the
bisquaternary ammonium which contained 12 meth-
ylenes (G4, 0.09 µM). In contrast, the TA1 IC50 was
very close to that of the bisquaternary ammonium
whose interchain between the two nitrogens possessed
exact saturated counterpart with similar polar head size
was not synthesized. Note also that the propargyl
substituent (G94, 7.6 nM) led to a 6-fold lower activity
than butynyl (G74, 44 nM), probably due to a higher
steric hindrance at the nitrogen level in the latter.
Nevertheless, this effect was not observed for monoam-
monium salts (compare E36 and E34, IC50 0.36 and 0.23
µM, respectively).
Discu ssion
Extensive fundamental research on intraerythrocytic
Plasmodium PL metabolism revealed potential targets
for chemotherapeutic interference. The most effective
interference now appears to be blockage of the choline
carrier, a limiting step which gives the parasite the
precursor required for the synthesis of PC, the major
4,5
PL of Plasmodium. This pharmacological interference
was previously validated with a very close correlation
between antimalarial activity and inhibition of PL
6
methylenes (G1, 700 µM).
6
metabolism.
(
b ) Mod ifica t ion of t h e Bu lk of t h e Ca t ion ic
QSAR analyses have indicated that lipophilicity is a
Hea d . When we compare bisammonium salts with the
same inter-nitrogen atom but different cationic heads,
i.e. trimethyl, methylpyrrolidinium, triethyl, diethylhy-
droxyethyl, or dimethyldodecyl (G4/G24/G14/G44 or
G5/G25/G15/H5/G45), we noticed a similar correlation
between antimalarial activity and cationic head size, as
observed with monoammonium salts (Table 1), but the
order changed according to the polar head. For the same
inter-nitrogen atom chain (either C12 or C16), compounds
with a triethylammonium head were about twice as
active as their trimethyl analogues (G14/G4 and G15/
G5). This factor was around 7 for monoammonium salts
8
major relevant parameter for antimalarial activity. In
the present study, 36 original compounds, mono- or
biscationic compounds, with various lipophilicities were
synthesized to improve in vitro activity. The present
data revealed a dramatic improvement in IC50, from
-8
-12
1
0
to 10
M.
The most active monoquaternary monoammonium
salts were triethyldodecylammonium bromide (E10, 6.4
-8
×
10
M) and tripropyldodecylammonium bromide
-8
(
E13, 3.3 × 10 M). Bisquaternary ammonium salts
exhibited far higher activity than monoammoniums,
especially those with a long intercationic chain of more
(E10/E6 and E11/E8, see also Figures 1 and 3). An
-9
-12
than 12 carbon atoms (10 -10
M). It is noteworthy
N-methylpyrrolidinium head instead of a trimethylam-
monium improved the activity by 6-7-fold (G24/G4 and
G25/G5), whereas it slightly increased the activity (by
that our oriented chemical syntheses led to eight novel
compounds (out of 36 newly synthesized) whose IC50
values were lower than 10 nM. They are therefore more
active than most current antimalarial compounds, i.e.
antifolic/antifolinic or lysosomotropic compounds (e.g.
2
-fold) for monoammonium (E24/E6 or E26/E8).
The presence of an additional lipophilic chain (C12H25)
on each nitrogen atom slightly decreased the activity
G4/G44), as observed with monoammonium salts. In
9
9
quinine or chloroquine). Only halofantrine and arte-
(
10
misinin analogues exhibited IC50 values in a low
contrast, when the interchain had 16 methylenes, the
presence of C12H25 caused a dramatic decrease in
activity (by 450-fold, G5/G45). With 1,12-dodecane and
nanomolar range. Furthermore, some of our analogues
in the monoammonium (e.g. E10, E30, F 8) and bisam-
monium (e.g. G5 and G25) series were tested against
various chemoresistant P. falciparum strains or isolates.
They showed the same high activity as against sensitive
strains. More specifically, G25 was equally active
against the two chloroquine-resistant strains FCR3 and
L1 (IC50 of chloroquine is 0.8 and 0.9 µM, respectively)
with IC50 of 3.6 and 1.4 nM, respectively. G25 was
also very active (nanomolar range) against various
human isolates, as they are cycloguanil-, chloroquine-,
or mefloquine-resistant (data not shown).
This study also provided additional information re-
garding the active site of the pharmacological tar-
get. Our previous study suggested that potent antima-
larials are cationic substances with a “small” (not pre-
cisely defined) quaternary ammonium head and a long
hydrophobic chain (roughly corresponding to 10-12
methylenes), which could combine with an anionic site
and a long hydrophobic adjacent region of the target.8
The results of the present study extend these data and
specify the optimum volume of the anionic pocket in
which the quaternary ammonium polar head can fit and
1
,16-hexadecane diaminium, plots of IC50 as a function
of the polar head volume (Figure 2) were similar to those
obtained with monoammonium salts, i.e. minimums
3
between 200 and 350 Å .
(
c) In tr od u ction of Hyd r oxyl Gr ou p s a n d Un -
sa tu r a tion s in N-Su bstitu en ts. We also investigated
the effects of introducing a hydroxyl function or unsat-
urations in the nitrogen substituents of these duplicated
molecules on the antimalarial activity. The presence of
a hydroxyl group on one short nitrogen substituent led
to a 3-fold reduction of antimalarial activity (G15/H5,
see also Figure 3), as with monoammonium salts (E10/
F 8, see also Figure 1).
The presence of an unsaturated group (allyl in G84,
propargyl in G94, or but-3-ynyl in G74) on the nitrogen
atom led to very active compounds (IC50 in the nano-
molar range). A slight improvement in the activity was
noted in comparison to trimethyl (G4, 90 nM) or triethyl
(G14, 45 nM), but in this case the degree of improve-
ment was difficult to accurately determine since the

Compounds Interfering with P. falciparum Metabolism
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 3 511
F igu r e 4. Schematic description of steps involved in the binding of a bivalent ligand to vicinal recognition sites: (i) two unoccupied
vicinal sites, (ii) univalent binding of a bivalent ligand, (iii) occupation of vicinal recognition sites by individual bivalent ligands,
(iv) bridging of vicinal sites. The sites may or may not be identical. This scheme is adapted from ref 17. For the choline carrier,
the active site is schematized by a sphere whose radius is 4 ( 0.4 Å (capable of accommodating up to 3 propyl groups) and the
length of the hydrophobic domain (hatched) between the two sites is 20 Å (corresponding to 14 methylene groups).
the nature of the hydrophobic domain. The possible
nature of the target is also discussed in the light of data
obtained with the most active bisquaternary ammonium
salts, thus providing a more accurate definition of the
active site.
notable for bisquaternary ammonium with a hexade-
cylmethylene inter-nitrogen chain, for which the pres-
ence of a dodecyl substituent decreased by 450-fold the
activity (G45/G25).
Hydrogen bonds are probably not required to enable
the cationic portion of active molecules to come within
the anionic pocket, since long active molecules devoid
of a hydroxyl group have the same or better activity
than molecules possessing one.
1
. Active Site(s) Accom m od a te Ca tion ic Liga n d s
(
Ch olin e or Qu a ter n a r y Am m on iu m ). Concerning
the strict requirement of a cationic headgroup of our
8
compounds, it is likely that the target has a negative
subsite that could involve either carboxylate anions of
Asp or Glu or aromatic rings of some amino acids. The
presence of aromatic side chains (also called “aromatic
basket”) in the binding sites for quaternary ammonium
ligands appears to be a common property shared by
several acetylcholine binding proteins, acetylcholinest-
erase, and acetylcholine muscarinic and nicotinic recep-
tors and also in the potassium channels which bind
tetraethylammonium.11
In both mono- and bisquaternary ammonium salts, a
moderate increase in lipophilicity around nitrogen was
quite beneficial for antimalarial potency, indicating that
this pocket could be large enough to accommodate a
polar head bigger than a trimethylammonium, e.g.
bearing up to three propyl groups (E13), corresponding
In fact, the affinities of corresponding trimethyl and
triethyl (or N-methylpyrrolinium) analogues (between
2 and 16 methylenes in the lipophilic chain) almost
always differed throughout both series. In the monoam-
monium series, the improvement was constantly higher
when the polar head was an N-methylpyrrolidinium
1
(
2.5-fold compared with trimethyl) and particularly for
a triethyl (8-fold increase) (see Figure 1). This consis-
tency indicated a highly specific binding site and that
all the analogues are bound at this site. Affinities are
thus not simply determined by the hydrophobic chain
length of the analogues, and inhibitors containing the
same number of methylene groups may have distinct
affinities. For the bisammonium series, a significant
difference was also noted, but in this series N-meth-
ylpyrrolidinium derivatives were constantly more active
(11-fold higher compared to trimethyl), whereas N-
triethyl analogues were only 2-fold more active (see
Figure 3). The reason the target better accommodates
triethyl groups for monoammonium salts and pyrroli-
dinium groups for bisammonium salts remains unclear
but probably reflects the presence of other constraints
imposed by additional interactions of bisquaternary
ammonium salts with the target.
to a globular pocket whose volume ranges from 200 to
3
3
(
50 Å and therefore whose mean radius is ∼4 ( 0.4 Å
see Figure 4).
When steric hindrance was further increased, e.g.
presence of a second long lipophilic group (butyl and
especially equivalent to hexyl or dodecyl) on the nitrogen
atom, one of the chains is probably forced to extend out
of the pocket which is energetically more unfavorable,
thus leading to decreased affinity. This was especially

5
12 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 3
Calas et al.
The distinct pattern of activity as a function of
square of the monovalent ligand affinity, i.e. the maxi-
mal expected value (see above). This suggests that there
was almost optimal docking of the free pharmacophore
at this spacer length. It also indicates that this spacer
(containing 21 methylene groups) permits bridging of
vicinal sites without imposing any translational con-
polymethylene chain length of mono- and bisquaternary
ammonium compounds, along with the high potency
improvement, may indicate the presence of two active
anionic (or electron-rich) pockets in the target (see
below). These two active sites could be identical or
distinct and located either in two regions of a single
protein or on two separate protein subunits. Indeed,
some divalent pharmacological ligands can recognize
distinct regions in a target,12 as demonstrated with
symmetrical polymethylene tetraamines, with respect
to the M2 muscarinic receptor12 or to the acetylcholine
nicotinic receptor protein with quaternary ammonium
salts.1
1
7
straints upon the pharmacophore.
Consequently, the relationship between antimalarial
potency improvement and the spacer length of bivalent
ligands noted for P. falciparum-infected erythrocytes
might provide a rough indication of the distance be-
tween proximal sites in the bound state. These results
are consistent with the bridging of proximal sites when
the spanner has a linear length of 28 Å (equivalent to
roughly 21 methylenes). The linear spacer length is
probably greater than the intersite distance because,
after proceeding to state (ii), sufficient translational
mobility of the free pharmacophore might be required
for docking the proximal recognition site.
3. Lip op h ilic Dom a in . The nature and length of the
hydrophobic/lipophilic domain adjacent to the anionic/
electron-rich active site(s) of the target could also be
determined. The TA1 compound including three qua-
ternary ammoniums separated by six methylenes was
not very active (IC50 800 µM). In contrast, a bisammo-
nium analogue with a similar total length, but without
any cationic charge between the two nitrogen atoms
(G14, 0.045 µM), showed very potent antimalarial
activity. This suggests that the lipophilic region of the
target probably does not bear a negative charge, thus
enhancing the ability to bind this domain. Moreover,
E70 which possesses a phenyl ring at the end of an octyl
chain had the same activity as its saturated dodecyl
analogue (E6). This electron-rich phenyl ring therefore
does not form additional specific bonds. This suggests
that there is no positive charge in the long domain
adjacent to the anionic (or electron-rich) pocket. This
area is probably neutral and purely hydrophobic.
The length of this domain between two active sites is
probably higher than 20 Å, i.e. the length of an alkyl
chain containing 14 methylene groups (see above).
Figure 4 summarizes the main topographic features of
the target in malaria-infected red blood cells, regarding
the size of the active sites and the length of the lipophilic
domain bridging two active sites.
4. Sp ecificit y of Qu a t er n a r y Am m on iu m t o P .
fa lcipa r u m -In fected Er yth r ocytes. We have already
discussed the specificity of structural requirements for
inhibiting choline entry into infected erythrocytes rela-
tive to other cholinergic systems [notably the high-
affinity choline transport (HACT) in synaptosomes].6
Briefly, discrepancies with other systems concern the
steric fit of analogues at the active site, the absence of
stereospecificity of choline stereoisomers, and the ab-
sence of dramatic antimalarial activity of various well-
known cholinergic compounds (e.g. hemicholinium, which
is a potent HACT inhibitor).
1,13
2
. Br id gin g Tw o P r oxim a l Active Sites. Dimer-
ization of the quaternary ammonium produced effectors
with enhanced antimalarial activity relative to the
monomeric congeners, indicating increased affinity. This
is a result of the presence of the second nitrogen atom
in the molecule, not a nonspecific effect of the hydro-
phobic spacer.
This huge increase in activity is not unusual for
bivalent ligands able to bind adjacent target proteins
or subsites,1
4-17
regardless of whether they are identical
or not (see above). Binding of a bivalent ligand to vicinal
recognition sites may involve distinct steps: (i) two
unoccupied vicinal sites, (ii) univalent binding of a
bivalent ligand via only one pharmacophore, (iii) oc-
cupation of vicinal recognition sites by individual biva-
lent ligands, or (iv) bridging, i.e. simultaneous occupa-
tion of vicinal sites (which may or may not be identical)
by one bivalent ligand1
6-18
(see Figure 4). Proceeding
from state (ii) to state (iv) would be more likely than
univalent binding of a second bivalent ligand (iii), if the
spacer length permits bridging.17 In this case, a bivalent
ligand should, at most, possess an affinity constant
equal to the product of the binding contributions of the
individual recognition units,16 i.e. for symmetrical biva-
lent compounds, the square of the affinity constant of
the monoligand if the two sites are identical.
Our results with bisammonium salts are consistent
with bridging between proximal sites at a specific spacer
length. Bivalent ligands containing 10-12 methylenes
had antimalarial potency equivalent to monovalent
analogues with the same number of methylene groups,
-
7
-8
i.e. around 10 -10 M (see Tables 1 and 2). With
successive additions of methylene units to the spacer
(
from 14 methylenes), the potency increased markely
for bis forms but not for mono forms. Ligands with
carbon chains longer than 14 methylene groups (e.g. 20
Å) had the optimal distance to allow simultaneous
interaction with the two target binding sites. They could
also be folded so as to enable better hydrophobic
interactions and greater rigidity than a monoammo-
nium whose carbon chain is barely long enough to span
the distance between the two presumed anionic do-
mains.
The highest potency ratio between bis and mono
analogues was obtained for 21 methylenes, reaching ca.
The present study extends these results. Indeed, the
absence of a plateau for antimalarial activity inhibition
as a function of the polymethylene chain length of
bisammonium (up to 21) (see above and Figure 3)
strongly contrasts with the patterns noted for nicotinic
receptors or HACT in synaptosomes, for which inhibi-
tion was maximal for 8-12 and 16-18 methylenes,
5
6
1
0 -10 times that of the monovalent analogue (see
G19, 3 pM/monoammonium analogue, whose IC50 can
be extrapolated from Figure 1 as ≈1 µM). Hence,
assuming that the IC50 reflects the affinity for the
target, the bivalent G19 affinity is roughly equal to the

Compounds Interfering with P. falciparum Metabolism
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 3 513
respectively, followed by a decrease.18,19 This indicates
that in both cases the intersite distance is much shorter
than in infected erythrocytes.
1 nM) was obtained for bisquaternary ammonium salts
containing more than 14 methylene groups in the inter-
nitrogen spacer. Moreover the longer the hydrophobic
Assuming that, in Plasmodium, the IC50 is directly
alkyl interchain, the better the antimalarial activity (up
6
-12
related to the affinity constant K for the target (likely
to 21 methylene groups, G19, IC50 3 × 10
M).
7
the choline carrier of infected erythrocytes ), the incre-
N-Methylpyrrolidinium was the polar head that prob-
ably fits best into the active site of the target molecule.
These results suggest the presence of two anionic sites
(with a radius of the globular pocket of 4 ( 0.4 Å) in
the target (likely the choline carrier). Between these
sites, there is a long hydrophobic domain corresponding
to a length of at least 14 methylene groups (higher than
20 Å).
ment in affinity can be estimated in association with
the addition of a CH2 group to the molecule and
compared with other cholinergic models. For the striatal
synaptosome HACT, the affinity increased by 1 order
of magnitude when the polymethylene chain of the
1
9
bisammonium salt was increased from 10 to 18 units,
as compared to more than 3 orders of magnitude for the
activity against Plasmodium-infected erythrocytes (see
Figure 3). Regarding cholinesterase activity, both ab-
solute values of affinity constants K for mono- and
bisquaternary ammonium were 2-fold lower2 than in
Plasmodium, also indicating a lower affinity for choline
esterase than for the malarial choline carrier.
Exp er im en ta l Section
Ch em ist r y. Gen er a l P r oced u r e for t h e Syn t h esis of
Am m on iu m Qu a ter n a r y Sa lts. Most of the quaternary
ammonium salts indicated in Table 1 were synthesized by
alkylation of a tertiary amine with an alkyl halide in a polar
solvent or without solvent. In this way, using the appropriate
alkyl bromide, trimethylamine gave E2a and E70, N-meth-
ylpyrrolidine led to E24-E27, triethylamine gave E60, E11,
and E12, 3-dimethylaminopropyne provided E36, 1-dimeth-
ylamino-2-propanol gave F 30 and F 32, and dimethylamino-
ethanol gave F 2a and F 25. 1-Dimethylaminododecane led to
E35 by action of allyl bromide and to E34 by reaction with
0
Finally, regarding the steric fit at the active site,
bistrimethyl analogues were less active in comparison
to other more hindered compounds (bistriethyl, etc.)
against Plasmodium-infected erythrocytes than against
nicotinic receptor.18
Tubocurarine, a cholinergic cyclic bisquaternary am-
monium salt, was also shown to be the most potent
compound possessing nicotinic curare-like activity rela-
tive to other bivalent ligands (such as decametho-
4
-tosylbutyne. E37 and E38 were produced by reaction, under
basic conditions, of N-methyldodecylamine with an excess of
3-bromopropene and 3-bromopropyne, respectively.
Similarly, most of the bisquaternary ammonium salts (Table
2) were obtained by alkylation of a tertiary amine with an alkyl
dihalide. Action of the suitable R,ω-dihaloalkane in ethanol
with an excess of (a) trimethylamine yielded G5, (b) trieth-
ylamine led to G12, G14, G15, and G19, (c) N-methylpyrro-
lidine gave G23-G25, (d) 3-dimethylaminopropyne led to G94,
(e) 2-diethylaminoethanol gave H5, and (f) 1-dimethylamino-
dodecane yielded G40, G41, G44, and G45. Finally, G74 was
synthesized by reaction of N,N,N′,N′-tetramethyldiaminodode-
cane with an excess of 4-tosylbutyne, and G84 was obtained
by alkylation of N,N,N′,N′-tetraethyldiaminododecane with
allyl bromide.
1
8
nium). In contrast, tubocurarine exhibited very low
antimalarial activity (IC50 ) 100 µM; data not shown,
i.e. 100-fold higher than decamethonium).
The marked differences in response, in terms of
pattern and relative affinity as a function of polymeth-
ylene chain length and of steric fit at the active site,
relative to the nitrogen substitution bulk between
antimalarial effect and other cholinergic models (e.g.
HACT, nicotinic or muscarinic cholinergic receptor,
cholinesterase, etc.) reflect considerable differences in
these different targets (see above).
Pentamethylenebis(1-methylpyrrolidinium) oxalate (G20)
and N-dodecyl-N-methylephedrinium bromide (F 50) were ob-
tained from Sigma-Aldrich Chemical Co. (St. Louis, MO). 1,-
In summary, with these compounds, a suitable very
high level of activity has now been reached in vitro
against P. falciparum (eight molecules with an IC50
lower than 10 nM). When tested (e.g. G25), they were
equally potent against various resistant isolates. The
compounds also exerted potent antimalarial activity in
vivo both against murine malaria in the rodent model
and against human malaria in monkeys (manuscript in
preparation). The potent in vitro antimalarial activity
noted in the present study and the probable specificity
of the choline carrier of infected erythrocytes (compared
to other cholinergic systems such as HACT in synap-
tosomes) suggest in vivo specificity. This is supported
by the high in vitro therapeutic index against various
1
4-Tetradecamethylenebis(N-methylpyrrolidinium) dibromide
(G22) was synthesized and provided by Virbac laboratories.
Melting points were determined in capillary tubes on a
1
Thomas-Hoover apparatus and were uncorrected. H NMR
spectra were recorded on a Brucker AC 250 (250 MHz).
Chemical shifts (δ) are expressed in ppm downfield from the
internal standard tetramethylsilane (TMS). Elemental analy-
ses were within (0.4 of calculated values. Yields, melting
points, and H NMR data of products are reported in Table 3.
N,N,N-Tr im eth yl-1-octa n a m in iu m Br om id e (E2a ),
N,N,N-Tr im et h yl-1-(8-p h en yl)oct a n a m in iu m Ch lor id e
1
(
E70), an d N,N,N,N′,N′,N′-Hexam eth yl-1,16-h exadecan edi-
a m in iu m Dibr om id e (G5). 10 mmol of halogenated deriva-
tive (1-bromooctane for E2a , 1-choro-8-phenyloctane for E70)
or dihalogenated derivative (1,16-dibromohexadecane for G5)
was added to 25 mL solution of methylamine in ethanol (33%).
The mixture was heated under reflux for 12 h. The solvent
was removed and the salt was recrystallized from a mixture
of acetone and diethyl ether. E2a : Anal. (C11
N. E70: Anal. (C17
C, H, N.
N-Meth yl-N-h exylp yr r olid in iu m Br om id e (E25), N-
Meth yl-N-d od ecylp yr r olid in iu m Br om id e (E24), N-
Meth yl-N-tetr a d ecylp yr r olid in iu m Br om id e (E27), N-
Meth yl-N-h exa d ecylp yr r olid in iu m Br om id e (E26), 1,12-
Dod eca m eth ylen ebis(N-m eth ylp yr r olid in iu m ) Dibr o-
m id e (G24), a n d 1,16-Hexa d eca m eth ylen ebis(N-m eth yl-
p yr r olid in iu m ) Dibr om id e (G25). 5 g (60 mmol) of N-
6
eukaryotic cells of other analogues in these series. This
also concerns some bis analogues of the present study.
We have shown that they were much less toxic in vitro
against mammalian cells (macrophages, lymphoblastoid
cells) than against P. falciparum with a differential of
H
26NBr) C, H,
H
30NCl) C, H, N. G5: Anal. (C22 Br )
H
50
N
2
2
6
activity ranging between 300 (G12) and 1.2 × 10 (G19)
(data not shown). Altogether this indicates a promising
future for the development of this new pharmacological
model against malaria.
This study provides a rough topographic model of the
ligand binding site. The highest activity (IC50 lower than

5
14 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 3
Calas et al.
Ta ble 3. Physicochemical and Spectral Data for Synthesized Compounds
compd
yield (%)
mp (°C)
201
215-220
74
186
204
209
104
155
160
63
1H NMR (δ in ppm)
0.90 (t,3H); 1.05 (m,10H); 1.5 (m,2H); 3.2 (s,9H); 3.3 (m,2H)a
E2a
E70
E25
E24
E27
E26
E60
E11
E12
E35
E36
E34
48
86
70
52
61
72
51
61
38
62
65
61
a
1.2 (m,8H); 1.5 (m,2H); 1.6 (m,2H); 2.5 (t,2H); 3.4 (s,9H); 3.5 (m,2H); 7.1 (m,3H); 7.2 (m,2H)
a
0.70 (t,3H); 1.1 (m,6H); 1.5 (m,2H); 2.0 (m,4H); 3.1 (s,3H); 3.35 (m,2H); 3.5 (m,4H)
0.80 (t,3H); 1.2 (m,18H); 1.65 (m,2H); 2.2 (m,4H); 3.2 (s,3H); 3.6 (m,2H); 3.7 (m,4H)a
a
0.65 (t,3H); 1.05 (m,22H); 1.55 (m,2H); 2.05 (m,4H); 3.05 (s,3H); 3.45 (m,2H); 3.6 (m,4H)
a
0.78 (t,3H); 1.15 (m,26H); 1.7 (m,2H); 2.23 (m,4H); 3.21 (s,3H); 3.55 (m,2H); 3.75 (m,4H)
0.8 (t,3H); 1.4 (m,15H); 1.6 (m,2H); 3.3 (m,8H)a
a
0.8 (t,3H); 1.4 (m,35H); 1.7 (m,2H); 3.3 (m,8H)
0.8 (t,3H); 1.3 (m,39H); 1.6 (m,2H); 3.2 (m,8H)a
a
0.9 (t,3H); 1.3 (m,18H); 1.8 (m,2H); 3.4 (s,6H); 3.5 (m,2H); 4.4 (d,2H); 5.8 (m,2H); 6,0 (m,1H)
70
105
0.8 (t,3H); 1.2 (m,18H); 1.7 (m,2H); 3.0 (t,1H);3.35 (s,6H); 4.0 (t,2H); 4.7 (d,2H)a
0.8 (t,3H); 1.15 (m,18H); 1.55 (m,2H); 2.15 (t,1H); 2.3 (s,3H); 2.7 (td,2H); 3.2 (s,6H); 3.3 (m,2H);
.5 (t,2H); 7.1 (d,2H); 7.7 (d,2H)a
3
a
a
E37
E38
F 2a
F 25
65
70
70
68
20
0.8 (t,3H); 1.25 (m,18H); 1.7 (m,2H); 3.2 (s,3H); 3.3 (m,2H); 4.25 (m,4H); 5.7 (m,4H); 5.9 (m,2H)
49-51
116-118
203-204
0.9 (t,3H); 1.3 (m,18H); 1.9 (m,2H); 3.0 (t,2H); 3.5 (s,3H); 3.7 (m,2H); 4.7 (m,4H)a
0.80 (t,3H); 1.2 (m,10H); 1.6 (m,2H); 3.2 (s,6H); 3.4 (m,2H); 3.55 (m,2H); 3.9 (m,2H); 4.85 (t,1H)
0.90 (t,3H); 1.32 (m,26H); 1.75 (m,2H); 3.41 (s,6H); 3.60 (m,2H); 3.80 (m,2H); 4.15 (m,2H);
c
5
.10 (m,1H)
F 32
F 30
43
73
hygroscopic
60
0.8 (t,3H); 1.2 (d,3H); 1.3 (m,4H); 1.7 (m,2H); 3.3 (s,6H); 3.4 (m,2H); 3.8 (m,2H); 4.4 (m,1H);
a
4
.9 (m,1H)
0.8 (t,3H); 1.2 (d,3H); 1.3 (m,18H); 1.7 (m,2H); 3.3 (s,6H) 3.4 (m,2H);3.8 (m,2H); 4.4 (m,1H);
a
5
.0 (m,1H)
b
G5
61
88
52
70
60
67
63
50
10
227-230
130 (dec)
162
198
205
204
180
145
170
1.4 (m,24H); 1.8 (m,4H); 3.2 (s,18H); 3.3 (t,4H)
b
G12
G14
G15
G19
G24
G25
G94
G74
1.4 (m,18H); 1.5 (m,8H); 1.85 (m,4H); 3.7 (m,16H)
1.3 (m,16H); 1.4 (t,18H); 1.7 (m,4H); 3.5 (m,16H)b
b
1.25 (m,18H); 1.38 (m,24H); 1.82 (m,4H); 3.32 (m,16H)
1.2 (s,34H); 1.35 (t,18H); 1.5 (m,4H); 3.4 (m,16H)
a
1.40 (m,16H); 1.85 (m,4H); 2.2 (m,8H); 3.15 (s,6H); 3.55 (m,12H)b
1.45 (m,24H); 1.93 (m,4H); 2.38 (m,8H); 3.20 (s,6H); 3.63 (m,12H)
b
1.3 (m,16H); 1.7 (m,4H); 3.15 (s,12H); 3.40 (m,4H); 4.1 (t,2H); 4.5 (d,4H)a
1.3 (m,16H); 1.65 (m,4H); 2.3 (s,6H); 2.75 (td,4H); 3.1 (t,2H); 3.25 (m,4H); 3.45 (t,4H); 7.1 (d,2H);
a
7
.5 (d,2H)
a
G84
H5
75
60
70
80
80
76
45
pasty
120
195
235
235
128
262
1.2 (t,12H); 1.3 (m,16H); 1.6 (m,4H); 3.15 (m,4H); 3.25 (q,8H); 4.0 (d,4H); 5.6 (m,4H); 6.0 (m,2H)
1.45 (m,36H); 1.85 (m,4H); 3.5 (m,16H); 4.15 (t,4H)b
0.8 (t,6H); 1.25 (m,36H); 1.75 (m,4H); 2.75 (m,2H); 3.5 (s,12H); 3.9 (m,8H)
a
G40
G41
G44
G45
TA1
a
0.9 (t,6H); 1.3 (m,40H); 1.65 (m,4H); 1.75 (m,4H); 3.5 (s,12H); 3.7 (m,8H)
0.8 (t,6H); 1.3 (m,52H); 1.7 (m,4H); 1.8 (m,4H); 3.5 (s,12H); 3.7 (m,8H)a
a
0.9 (t,6H); 1.3 (m,60H); 1.65 (m,4H); 1.75 (m,4H); 3.6 (s,12H); 3.7 (m,8H)
1.5 (m,40H); 3.3 (m,24H)b
a
In CDCl3. b In D2O. c In DMSO-d6.
methylpyrrolidine was added to 1.1 equiv of halogenated
derivative (1-bromohexane for E25, 1-bromododecane for E24,
reflux for 18 h. The solvent was removed and the salt
recrystallized from a mixture of methanol and diethyl ether
or ethanol and diethyl ether or acetone and diethyl ether.
1
-bromotetradecane for E27, 1-bromohexadecane for E26) or
to 0.55 equiv of dihalogenated derivative (1,12-dibromodode-
cane for G24, 1,16-dibromohexadecane for G25) in 100 mL of
ethanol. The solution was heated under reflux for 24 h. The
solvent was removed and the salt was recrystallized from a
G12: Anal. (C20
C, H, N. G15: Anal. (C28
Br ) C, H, N.
N,N-Dim eth yl-N-(2-p r op yn yl)-1-d od eca n a m in iu m Io-
d id e (E36). 20 mmol of 1-iodododecane was added to a solution
of 15 mmol of 3-dimethylaminopropyne in 60 mL of anhydrous
acetone. The solution was heated under reflux for 24 h; the
solvent was removed and the salt recrystallized from acetone.
H
46
N
2
Br
2
) C, H, N. G14: Anal. (C24
54 2 2
H N Br )
H
62
N
2
2
Br ) C, H, N. G19: Anal.
(C32
H
72
N
2
2
mixture of methanol and diethyl ether. E25: Anal. (C11
NBr) C, H, N. E24: Anal. (C17 36NBr) C, H, N. E27: Anal.
40NBr) C, H, N. E26: Anal. (C21 44NBr) C, H, N. G24:
) C, H, N. G25: Anal. (C26 Br ) C, H,
24
H -
H
(C
19
H
H
Anal. (C22
N.
H
46
N
2
Br
2
H
54
N
2
2
Anal. (C17H34NI) C, H, N.
N,N,N-Tr ieth yl-1-h exan am in iu m Br om ide (E60), N,N,N-
Tr ieth yl-1-h exadecan am in iu m Br om ide (E11), an d N,N,N-
Tr ieth yl-1-octa d eca n a m in iu m Br om id e (E12). 5 g (50
mmol) of triethylamine was added to 1.1 equiv of halogenated
derivative (1-bromohexane for E60, 1-bromohexadecane for
E11, 1-bromooctadecane for E12) in 100 mL of ethanol. The
solution was heated under reflux for 18 h. The solvent was
removed and the salt was recrystallized from a mixture of
methanol and diethyl ether or ethanol and diethyl ether.
N,N-Dim et h yl-N-(2-h yd r oxyp r op yl)-1-p en t a n a m in i-
u m Br om id e (F 32) a n d N,N-Dim eth yl-N-(2-h yd r oxyp r o-
p yl)-1-d od eca n a m in iu m Br om id e (F 30). 20 mmol of halo-
genated derivative (1-bromopentane for F 32, 1-bromododecane
for F 30) was added to a solution of 15 mmol of 1-dimethy-
lamino-2-propanol in 50 mL of ethanol. The solution was
heated under reflux for 12 h; the solvent was removed and
the salt recrystallized from ethanol-diethyl ether. F 32: Anal.
(C10H24NOBr) C, H, N. F 30: Anal. (C17H38NOBr) C, H, N.
E60: Anal. (C12
H, N. E12: Anal. (C24
N,N,N,N′,N′,N′-Hexa eth yl-1,8-octa n ed ia m in iu m Dibr o-
m id e (G12), N,N,N,N′,N′,N′-Hexa eth yl-1,12-d od eca n ed i-
a m in iu m Dibr om id e (G14), N,N,N,N′,N′,N′-Hexa eth yl-
H
28NBr) C, H, N. E11: Anal. (C22
52NBr) C, H, N.
H
48NBr) C,
N-(2-Hydr oxyeth yl)-N,N-dim eth yl-1-octan am in iu m Br o-
m id e (F 2a ) a n d N-(2-H yd r oxyet h yl)-N,N-d im et h yl-1-
h exa d eca n a m in iu m Br om id e (F 25). 0.12 mol of 2-(di-
methylamino)ethanol and 0.11 mol of the corresponding bro-
minated derivative (1-bromooctane for F 2a , 1-bromohexade-
cane for F 25) were added to 100 mL of methanol and heated
under reflux conditions for 20 h. The solvent was removed and
the crude product recrystallized from a mixture of methanol
H
1
,16-h exa d eca n ed ia m in iu m Dibr om id e (G15), a n d
N,N,N,N′,N′,N′-Hexaeth yl-1,21-h en eicosan ediam in iu m Di-
br om id e (G19). 10 g (0.1 mol) of triethylamine was added to
a solution of 25 mmol of dihalogenated derivative (1,8-
dibromooctadecane for G12, 1,12-dibromododecane for G14,
and diethyl ether. F 2a : Anal. (C12
Anal. (C16 36NOBr) C, H, N.
H28NOBr) C, H, N. F 25:
H
1
,16-dibromohexadecane for G15, 1,21-dibromoheneicosane21
N,N-Dim eth yl-N-(2-p r op en yl)-1-d od eca n a m in iu m Br o-
m id e (E35). 10 mmol of 3-bromopropene was added to a
for G19) in 100 mL of ethanol. The solution was heated under

Compounds Interfering with P. falciparum Metabolism
J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 3 515
solution of 15 mmol of dimethyldodecylamine in 50 mL of
ethanol. The solution was heated under reflux for 24 h; the
solvent was removed and the salt recrystallized from acetone-
under reduced pressure and the salt recrystallized from
ethanol-diethyl ether. Anal. (C28
62 2 2 2
H N O Br ) C, H, N.
N,N,N′,N′-Tet r a m et h yl-N,N′-d id od ecyl-1,3-p r op a n ed i
a m in iu m Dibr om id e (G40), N,N,N′,N′-Tetr a m eth yl-
N,N′-d id od ecyl-1,6-h exa n ed ia m in iu m Dibr om id e (G41),
N,N,N′,N′-Tetr a m eth yl-N,N′-d id od ecyl-1,12-d od eca n ed i-
a m in iu m Dibr om id e (G44), a n d N,N,N′,N′-Tetr a m eth yl-
diethyl ether. Anal. (C17H36NBr) C, H, N.
N,N-Dim eth yl-N-(3-bu tyn yl)-1-d od eca n a m in iu m Tosy-
la te (E34). 16.3 g (85.5 mmol) of 4-methylbenzenesulfonyl
chloride was added to a solution of 3 g (43 mmol) of 3-butyn-
1
-ol in 135 mL of anhydrous pyridine at 0 °C. The reaction
N,N′-d id od ecyl-1,16-h exa d eca n ed ia m in iu m
Diiod id e
mixture was kept at 0 °C for 12 h and poured in 250 mL of
cold water. The mixture was extracted with three portions of
diethyl ether, washed with diluted HCl then with water, dried
(G45). 1 equiv of the appropriate dihalogenated derivative (1,3-
dibromopropane for G40, 1,6-dibromohexane for G41, 1,12-
dibromododecane for G44, 1,16-diiodohexadecane for G45) was
refluxed for 48 h with 3 equiv of dimethyldodecylamine in
ethanol. The solvent was evaporated and the salt recrystallized
over MgSO
-butynyltosylate (57%) was obtained as an oil, which was used
4
, and evaporated under reduced pressure. 5.5 g of
3
in the next step. The previously obtained tosylate was added
to 1.5 equiv of dimethyldodecylamine in 50 mL of acetone. The
solution was heated under reflux for 24 h; the solvent was
removed and the salt recrystallized from acetone. Anal.
from dichloromethane-diethyl ether. G40: Anal. (C31
Br ) C, H, N. G41: Anal. (C34 Br ) C, H, N. G44: Anal.
Br ) C, H, N. G45: Anal. (C44 ) C, H, N.
N,N,N,N′,N′,N′′,N′′,N′′-Oct a et h ylb is(h exa m et h ylen e)-
tr ia m in iu m Tr ibr om id e (TA1). 5 g (23.3 mmol) of bis-
(hexamethylene)triamine, 25 mL (340 mmol) of diisopropyl-
ethylamine, and 15.3 g (140 mmol) of bromoethane were
refluxed for 72 h in 25 mL of ethanol. The solvent and the
excess bromoethane were evaporated, and the salt was recrys-
68 2
H N -
2
H
74
N
2
2
(C40
H
86
N
2
2
94 2 2
H N I
(C
25
H
43NO
3
S) C, H, N.
N,N-Di(2-p r op en yl)-N-m eth yl-1-d od eca n a m in iu m Br o-
m id e (E37) a n d N,N-Di(2-p r op yn yl)-N-m eth yl-1-d od e-
ca n a m in iu m Br om id e (E38). 30 mmol of appropriate halo-
genated derivative (3-bromopropene for E37, 3-bromopropyne
for E38) was added to a solution of 10 mmol of methyldode-
tallized from dichloromethane-diethyl ether. Anal. (C28
Br ) C, H, N.
Biologica l Activity. Human blood or AB human serum
64 3
H N -
cylamine and excess Na
2
CO
3
in 50 mL of acetonitrile. The
3
solution was heated under reflux for 24 h. The reaction mixture
was cooled; the precipitate was filtered and the filtrate
evaporated. The residue was recrystallized from a mixture of
came from the local blood bank. The Nigerian strain of P.
falciparum22 was maintained at 37 °C by serial passages in
human erythrocytes suspended in Hepes-buffered RPMI 1640
(Gibco, France) supplemented with 10% AB human serum
using the Petri dish candle jar method.23
dichloromethane-diethyl ether. E37: Anal. (C19
H38NBr) C, H,
N. E38: Anal. (C19 34NBr) C, H, N.
H
N,N,N′,N′-Tet r a et h yl-N,N′-d i(2-p r op en yl)-1,12-d od e-
ca n ed ia m in iu m Dibr om id e (G94). A mixture of 2 g (6
mmol) of 1,12-dibromododecane and 1.5 g (18.3 mmol) of
The antimalarial activity of the compound was measured
in microtiter plates according to the method of Desjardins et
2
4
1
-dimethylamino-2-propyne in 15 mL of acetone was heated
under reflux for 48 h. The white precipitate was filtered. Anal.
Br ) C, H, N.
N,N,N′,N′-Tet r a m et h yl-N,N′-d i(3-b u t yn yl)-1,12-d od e-
al. The final volume of the incubation medium in each well
was 200 µL and the hematocrit of the P. falciparum-infected
erythrocyte suspension was 1-2%, with an initial parasitemia
(i.e. the percentage of infected erythrocytes) of 0.3-0.5%. In
some cases, the drugs were dissolved in DMSO and then
further diluted in culture medium so that the final DMSO
concentration never exceeded 0.25%. After contact of the drug
(C
22
H
42
N
2
2
ca n ed ia m in iu m Ditosyla te (G74). A mixture of 3 g (15
mmol) of 1,12-diaminododecane and 7.5 g (250 mmol) of
paraformaldehyde in 250 mL of methanol was heated under
3
reflux for 1 h. After cooling, 3.6 g (95 mmol) of NaBH
4
was
with the parasite for one parasite cycle (48 h), [ H]hypoxan-
added, and the mixture was stirred for 1 h at room tempera-
ture; then 100 mL of water was added. The excess paraform-
aldehyde was filtered and the filtrate extracted with three
portions of dichloromethane, washed with water, dried over
thine (Amersham, France) was added for 18 h to assess
parasite viability. The results are expressed as the drug
concentration resulting in 50% inhibition (IC50) of parasite
growth and are means of at least two independent experiments
performed in triplicate using different drug stock solutions.
Inside each experiment, standard deviation was always lower
than 20% of the mean (intraexperiment) and when comparing
the two mean values for each experiment (interexperiment)
they did not differ by more than 50% (mean values of IC50 do
not differ significantly using Student’s t test). In the few cases
for which the two means differed by more than 50%, a third
experiment was performed to ascertain the value.
4
MgSO , and evaporated under reduced pressure. The oily
residue was diluted with 50 mL of diethyl ether, in which HCl
gas was bubbled. The precipitate of 1,12-bis(dimethylamino)-
dodecane dichlorhydrate was filtered and recrystallized from
acetone. Yield: 10%. Mp:169 °C.
1
,12-Bis(dimethylamino)dodecane was regenerated by treat-
ment in basic conditions, and 0.3 g (1 mmol) of this was
refluxed with 0.57 g (2.5 mmol) of 3-butynyl tosylate in 10 mL
of acetone for 12 h; the solvent was removed and the salt
recrystallized from dichloromethane-diethyl ether. Anal.
Ack n ow led gm en t. This work was supported by the
UNDP/World Bank/WHO special program for Research
and Training in Tropical Diseases (Grant 950165), the
Commission of the European Communities (INCO-DC,
PL-950529: IC18-CT960056), CNRS (GDR 1077, Etude
des parasites pathog e` nes), and Minist e` re de l′Enseigne-
ment Sup e´ rieur et de la Recherche, DSPT N˚5 and
MENESR-DGA/DSP and AUPELF-UREF ARC n˚X/
7.10.04/Palu 95.
(C
38
H
60
N
2
O
6
S
2
) C, H, N.
N,N,N′,N′-Tetr a m eth yl-N,N′-d i(2-p r op en yl)-1,12-d od e-
ca n ed ia m in iu m Dibr om id e (G84). A mixture of 5 g (15
mmol) of 1,12-dibromododecane and 11.1 g (150 mmol) of
diethylamine in 75 mL of ethanol was heated under reflux for
1
1
2 h. The solvent was removed and the residue diluted with
00 mL of diethyl ether, in which HCl gas was added. The
precipitate of 1,12-bis(diethylamino)dodecane dichlorhydrate
was filtered and recrystallized from acetone. Yield: 55%. Mp:
1
70 °C.
1
,12-Bis(diethylamino)dodecane was regenerated by treat-
Refer en ces
ment under basic conditions, and 1.4 g (4.4 mmol) of this was
refluxed with 1.62 g (13.3 mmol) of 3-bromopropene in 15
mL of acetone for 48 h; the solvent was removed and the
pasty residue could not be recrystallized. Anal. (C26
C, H, N.
N,N,N′,N′-Tetr aeth yl-N,N′-di(2-h ydr oxyeth yl)-1,16-h exa-
d eca n ed ia m in iu m Dibr om id e (H5). 1 g (2.6 mmol) of
,16-dibromohexadecane was refluxed with 2 mL of 2-dieth-
(
1) Wernsdorfer, W. H.; Payne, D. The dynamics of drug resistance
in Plasmodium falciparum. Pharmacol. Ther. 1991, 50, 95-121.
2) White, N. J .; Olliaro, P. L. Strategies for the prevention of
antimalarial drug resistance: rationale for combination chemo-
therapy for malaria. Parasitol. Today 1996, 12, 399-401.
3) Butler, D.; Maurice, J .; O′Brien, C. Briefing malaria. Nature
(
54 2 2
H N Br )
(
(
1
997, 386, 535-541.
4) Vial, H. J .; Ancelin, M. L. Malarial lipids. In Subcellular
Biochemistry; Avila, J . L., Harris, J . R., Eds.; Plenum Press:
New York, 1992; pp 259-306.
1
ylaminoethanol for 3 h; the excess amine was evaporated

5
16 J ournal of Medicinal Chemistry, 2000, Vol. 43, No. 3
Calas et al.
(
5) Vial, H. J .; Ancelin, M. L. Malarial lipids. In Malaria: Parasite
Biology, Pathogenesis, Protection; Sherman, I. W., Ed.; ASM
Press: Washington, D. C., 1998; Chapter 11, pp 159-175.
6) Ancelin, M. L.; Calas, M.; Bompart, J .; Cordina, G.; Martin, D.;
Ben Bari, M.; J ei, T.; Druilhe, P.; Vial, H. Antimalarial activity
of 77 phospholipid polar head analogues: close correlation
between inhibition of phospholipid metabolism and in vitro
Plasmodium falciparum growth. Blood 1998, 91, 1426-1437.
7) Ancelin, M. L.; Parant, M.; Thuet, M. J .; Philippot, J . R.; Vial
H. J . Increased permeability to choline in simian erythrocytes
after Plasmodium knowlesi infection. Biochem. J . 1991, 273,
(15) Portoghese, P. S.; Ronsisvalle, G.; Larson, DL.; Yim, C. B.; Sayre,
L. M.; Takemori, A. E. Opioid agonist and antagonist bivalent
ligands as receptor probes. Life Sci. 1982, 31, 1283-1286.
(
(
(
(
16) Portoghese, P. S. Bivalent ligands and the message-address
concept in the design of selective opioid receptor antagonists.
Trends Pharmacol. Sci. 1989, 10, 230-235.
17) Portoghese, P. S. Bivalent ligands in the development of selective
opioid receptor antagonists. In Trends in medicinal Chemistry;
Mutschler, E., Winterfeldt, E., Eds.; 1987, pp 327-336.
18) Barrow, R. B.; Ing, H. R. Curare-like action of polymethylene
Bis-quaternary ammonium salts. Br. J . Pharmacol. 1948, 3,
(
(
7
01-709.
2
98-304.
(19) Roberts, E.; Tamaru, M. The ligand binding site of the synap-
tosomal choline transporter: provisional model based on
8) Calas, M.; Cordina, G.; Bompart, J .; Ben Bari, M.; J ei, T.;
Ancelin, M. L.; Vial, H. Antimalarial activity of molecules
interfering with Plasmodium falciparum phospholipid metabo-
lism. Structure-activity relationship analysis. J . Med. Chem.
a
inhibition studies. Neurochem. Res. 1992, 17, 509-528.
(20) Bergmann, F.; Segal, R. The relationship of quaternary am-
monium salts to the anionic sites of true and pseudo cholinest-
erase. Biochem. J . 1954, 58, 692-698.
1
997, 40, 3557-3566.
(
9) Antimalarial drugs I. Biological background, experimental
methods and drug resistance; Peters, W., Richards, W. H. G.,
Eds.; Springer-Verlag: Berlin, 1984; Vol. 68/I, p 484.
(21) Friedman, L.; Arnon, S. Halopolycarbon homologation. J . Am.
Chem. Soc. 1974, 96, 7101.
(
10) De Vries, P. J .; Dien, T. K. Clinical pharmacology and thera-
peutic potential of artemisinin and its derivatives in the treat-
ment of malaria. Drugs 1996, 52, 818-836.
11) Changeux, J . P.; Galzi, J . L.; Devillers-Thi e´ ry, A.; Bertrand, D.
The functional architecture of the acetylcholine nicotinic receptor
explored by affinity labeling and site-directed mutagenesis.
Quart. Rev. Biophys. 1992, 25, 395-432.
12) Melchiorre, C.; Minarini, A.; Angeli, P.; Giardina, D.; Gulini, U.;
Quaglia, W. Polymethylene tetramines as muscarinic receptor
probes. TIPS 1989, December, 55-59.
13) Kosower, E. M. A structural and dynamic model for the nicotinic
acetylcholine receptor. Eur. J . Biochem. 1987, 168, 431-449.
14) Costa, T.; W u¨ ster, M.; Herz, A.; Shimohigashi, Y.; Chen, H. C.;
Rodbard, D. Receptor binding and biological activity of bivalent
enkephalins. Biochem. Pharmacol. 1985, 34, 25-30.
(
22) Richards, W. H.; Maples, B. K. Studies on Plasmodium falci-
parum in continuous culture. 1. The effects of chloroquine and
pyrimethamine on parasite growth. Ann. Trop. Med. Parasitol.
(
1
977, 73, 99-108.
(
23) J ensen, J . B.; Trager, W. Plasmodium falciparum in culture:
use of outdated erythrocytes and description of the candle-jar
method. J . Parasitol. 1977, 63, 883-886.
(
(24) Desjardins, R. E.; Canfield, C. J .; Haynes, J . D.; Chulay, J . D.
Quantitative assessment of antimalarial activity in vitro by a
semiautomated microdilution technique. Antimicrob. Agents
Chemother. 1979, 16, 710-718.
(
(
J M9911027