Synthesis and structure-activity analysis of new phosphonium salts with potent activity against African trypanosomes

Article abstract of DOI:10.1021/jm2014259

A series of 73 bisphosphonium salts and 10 monophosphonium salt derivatives were synthesized and tested in vitro against several wild type and resistant lines of Trypanosoma brucei (T. b. rhodesiense STIB900, T. b. brucei strain 427, TbAT1-KO, and TbB48). More than half of the compounds tested showed a submicromolar EC₅₀ against these parasites. The compounds did not display any cross-resistance to existing diamidine therapies, such as pentamidine. In most cases, the compounds displayed a good selectivity index versus human cell lines. None of the known T. b. brucei drug transporters were required for trypanocidal activity, although some of the bisphosphonium compounds inhibited the low affinity pentamidine transporter. It was found that phosphonium drugs act slowly to clear a trypanosome population but that only a short exposure time is needed for irreversible damage to the cells. A comparative molecular field analysis model (CoMFA) was generated to gain insights into the SAR of this class of compounds, identifying key features for trypanocidal activity.

Full text of DOI:10.1021/jm2014259

Article
pubs.acs.org/jmc
Synthesis and Structure−Activity Analysis of New Phosphonium Salts
with Potent Activity against African Trypanosomes
Andrea Taladriz,†,# Alan Healy,†,# Eddysson J. Flores Perez,†,# Vanessa Herrero García,†
́
Carlos Ríos Martínez,† Abdulsalam A. M. Alkhaldi,‡,# Anthonius A. Eze,‡ Marcel Kaiser,§,∥
Harry P. de Koning,‡ Antonio Chana,⊥ and Christophe Dardonville*,†
†Instituto de Química Med
́
ica, IQM-CSIC, Juan de la Cierva 3, E-28006 Madrid, Spain
‡Institute of Infection, Immunity and Inflammation, College of Medical, Veterinary and Life Sciences, University of Glasgow,
Glasgow, United Kingdom
§Swiss Tropical and Public Health Institute, Socinstrasse, 57, CH-4002 Basel, Switzerland
∥University of Basel, Petersplatz 1, CH-4003 Basel, Switzerland
⊥Instituto de Química Física “Rocasolano”, IQFR-CSIC, Serrano 119, E-28006 Madrid, Spain
S
* Supporting Information
ABSTRACT: A series of 73 bisphosphonium salts and 10 mono-
phosphonium salt derivatives were synthesized and tested in vitro against
several wild type and resistant lines of Trypanosoma brucei (T. b. rhodesiense
STIB900, T. b. brucei strain 427, TbAT1-KO, and TbB48). More than half
of the compounds tested showed a submicromolar EC50 against these
parasites. The compounds did not display any cross-resistance to existing
diamidine therapies, such as pentamidine. In most cases, the compounds
displayed a good selectivity index versus human cell lines. None of the
known T. b. brucei drug transporters were required for trypanocidal
activity, although some of the bisphosphonium compounds inhibited the
low affinity pentamidine transporter. It was found that phosphonium drugs
act slowly to clear a trypanosome population but that only a short
exposure time is needed for irreversible damage to the cells. A comparative
molecular field analysis model (CoMFA) was generated to gain insights
into the SAR of this class of compounds, identifying key features for
trypanocidal activity.
literature on this subject in the past 30 years. We recently
regained interest in the antiparasitic activity of phosphonium
salts with the discovery of a series of benzophenone-derived
bisphosphonium salt derivatives that showed a marked
antileishmanial activity in vitro.5 In Leishmania, the best
compound of the series (i.e., 4,4′-bis((tri-n-
pentylphosphonium)methyl)benzophenone dibromide) was
found to target the mitochondria of the parasite, inhibiting
complex II of the respiratory chain.5 Since some of these
benzophenone compounds also showed interesting activity
against T. b. rhodesiense,6 we decided to prepare new derivatives
to systematically examine the structure−activity relationship of
antitrypanosomal phosphonium compounds. Hence, we
synthesized 60 new phosphonium salt derivatives with
variations in the following parts of the lead structure: (a)
linker type, (b) linker length, (c) number of cations, (d) nature
of the counterion, and (e) nature of the phosphonium groups
substituents R1, R2, R3 (Chart 1).
INTRODUCTION
■
Human African trypanosomiasis (HAT or sleeping sickness) is
a parasitic disease caused by infection with two subspecies of
trypanosomes: Trypanosoma brucei gambiense and T. b.
rhodesiense. HAT is endemic in 36 countries of sub-Saharan
Africa and, together with the corresponding condition in
domestic animals, is a major cause of suffering and poverty for
the affected population. If left untreated, the disease is usually
fatal.1 However, the treatment options are scarce, as few drugs
are available (i.e., suramin, pentamidine, melarsoprol, eflorni-
thine, and recently, the combination therapy nifurtimox−
eflornithine). These are far from being ideal because of
ineffectiveness to some trypanosome species or stages of the
infection, as well as toxicity, the parenteral mode of
administration, and the emergence of resistance.2,3
The antitrypanosomal activity of benzyltriphenylphospho-
nium salts against T. brucei was first reported in 1979 by
Kinnamon et al.4 Some compounds were curative in a murine
model of T. b. rhodesiense infection. However, the study of the
potential of phosphonium salts as antiprotozoal agents was not
followed up by any research group as shown by the lack of
Received: October 21, 2011
Published: March 6, 2012
© 2012 American Chemical Society
2606
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
Chart 1. General Structure of Benzophenone-Derived Bisphosphonium Salt Derivatives with Antileishmanial and
Antitrypanosomal Activity5 and New Series Being Studied
a
Scheme 1. Synthesis of the Phosphonium Salt Derivatives
a
Reagents and conditions: (i) R1R2R3P (excess), DMF or toluene, Δ. See Tables 1−4 for substituents pattern.
The compounds were tested for in vitro activity against T. b.
rhodesiense (strain STIB900) and on a standard panel of T. b.
brucei lines (s427, TbAT1-KO, and TbB48) with decreasing
sensitivity to most diamidine and arsenic-based drugs due to
loss of specific drug transporters (see Experimental Section for
details).7,8 This is important because, like diamidines, most of
the compounds described here are dications. In fact, the drug
transporters may contribute positively to the selectivity of the
compounds under development or, conversely, be a cause of
drug resistance.9 Resistance to first line diamidines such as
diminazene aceturate is linked to loss of these transporters10
and represents a genuine threat to the treatability of
trypanosomiasis.2 Thus, it is essential to ascertain that no
cross-resistance between new compounds and existing therapy
will arise. In addition, we established the dynamics of the action
of phosphonium salts on trypanosomes: trypanocidal or
trypanostatic, fast action or slow, minimum time of exposure,
etc. Finally, we sought to understand the SAR behind this class
of phosphonium compounds by performing CoMFA.
RESULTS
■
Chemistry. The synthesis of the target compounds is based
on the nucleophilic substitution of a (bis)halogenated precursor
(i.e., linker) with a trisubstituted phosphine giving rise to the
corresponding phosphonium salts (Scheme 1). Hence, the
(bis)halogenated precursors must hold the different structural
characteristics (i.e., linker type, linker length, number of cations,
type of counterion) one wants to introduce in the lead
structure.
The different halogenated linkers were synthetized as shown
in Scheme 2. The 4,4′-bisbromomethyl linkers 1a,6 1b,11 1d,12
and 1e13 were synthesized by N-bromosuccinimide (NBS)
bromination of the 4,4′-dimethylphenyl precursors as previously
reported. The linkers 1c and 2c were commercially available.
The 4,4′-bischloromethyl linker 3a was obtained by reaction of
2607
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
a
Scheme 2. Synthesis of the Linkers
a
t
Reagents and conditions. (i) NBS, BuOOH, CCl4, reflux, 20−52%; (ii) BiCl3, 1,2-dichloroethane, Δ, 60−72%; (iii) (1) oxalyl chloride, AlCl3,
CH2Cl2, −15 °C, 5 h; (2) chlorobenzene, reflux, 5 h, 80−84%; (iv) MeOH, rt, quantitative; (v) LiAlH4, THF, reflux, 18 h, 90−94%; (vi) thionyl
bromide, CH2Cl2, rt, 2 h, 82−95%; (vii) thionyl chloride, CH2Cl2, rt, 2 h, 68−97%; (viii) Et3SiH, CF3CO2H, 0 °C → 45 °C→ rt, 61−90%; (ix)
ClCOCH2Cl, AlCl3, CH2Cl2, reflux, 3 h, 94%.
the corresponding 4,4′-bisbromomethyl linker 1a with BiCl3 in
anhydrous 1,2-dichloroethane.14 The diphenylethane (10f, 12f)
and diphenylpropane (11g, 13g) linkers were synthesized as
shown in Scheme 2. Friedel−Crafts acylation of diphenylethane
and diphenylpropane (oxalyl chloride/AlCl3) followed by
decarbonylation in refluxing chlorobenzene gave 4 and 5,15
respectively. The acyl chlorides were stirred in methanol at
room temperature to yield the methyl esters 6 and 7
quantitatively. Lithium aluminum hydride reduction of 6 and
7 gave excellent yields of the diols 8 and 9. These were
converted to the dibromide (10f, 11g) and dichloride (12f,
13g) using thionyl bromide and thionyl chloride, respectively.
The synthesis of the 4,4′-bis(2-chloroethylphenyl) linkers 16f
and 17g started with Friedel−Crafts acylation of diphenyl-
ethane and diphenylpropane with chloroacetylchloride/AlCl3.
Reduction of 14 and 15 with triethylsilane in trifluoroacetic
acid yielded the expected products 16f and 17g, respectively.
The bisphosphonium salts 18−53(a−e) were synthesized in
good yields by reaction of the corresponding 4,4′-bischlor-
omethyl (3a, 12f, 13g), 4,4′-bisbromomethyl (1a−e, 10f, 11g),
or 4,4′-bischloroethyl linker (16f, 17g) with an excess of
commercially available trisubstituted phosphine in anhydrous
DMF at 100 °C (150 °C for 16f and 17g) (Scheme 1). The
monophosphonium salts 54, 55, 56a−63a, and 56c−57c were
prepared in the same way from 4-chloromethyltoluene, 4-
bromomethyltoluene, 4-bromomethylbenzophenone (2a), and
1-bromomethyl-4-phenoxybenzene (2c), respectively. A lower
reaction temperature (50 °C) was used with bis- and tris-2-
methoxyphenylphosphine (42a, 43a−c, and 61a) to avoid the
formation of several byproducts. The salts were isolated by
crystallization from the reaction mixture or by semipreparative
HPLC−MS for 64 and 65. Compounds 18a−24a, 26a, 28a−
32a, 36a−38a, 48a, and 49a were synthesized as reported
earlier.5
due to the slightly different assay procedures used in the two
laboratories involved in this work.16 Alternatively, they could
reflect the minor biochemical differences between the two
strains used.
In general, compounds with three short alkyl substituents
(Me, Et, or Pr) on the phosphonium cations were poorly active
against both trypanosome species (EC50 ≥ 10 μM), whatever
the linker was (Table 1, entries 1−6; Table 4, entries 7, 9). The
activity was enhanced with the increase in chain length (n-octyl
∼ n-hexyl > c-hexyl ∼ n-pentyl > n-Bu > i-Bu), reaching a
maximum for six- to eight-carbon substituents (Table 1, entries
7−18). Replacement of the alkyl substituents with phenyl rings
increased the activity in the order (R1, R2, R3 = alkyl) < (R1, R2
= alkyl; R3 = Ph) < (R1 = alkyl; R2, R3 = Ph) (Table 1, entries
19−32).
Since the best activities and selectivities were obtained with
phenyl substituents, we decided to study whether the presence
of substituents on the aromatic ring would influence the
antitrypanosomal activity. As shown in Table 2, the best results
were obtained with the 4-Me substituted phenyl rings (entries
16−25). Substitution pattern for the phenyl ring, in order of
decreasing activity, was 4-Me > 3-Me > H and 2-OMe > 4-
OMe > 4-Cl ∼ 4-F > 4-CF3. Within this series, the order of
decreasing activity as a function of the linker was O ∼ (CH2)3 >
(CH2)2 ∼ CH2 > CO > SO2 ∼ NAc.
Bioisosteric replacement of the phenyl substituents of the
phosphonium cation by 2-thienyl groups (Table 3, entries 1
and 2) or 1-naphthyl substituents (Table 3, entries 10−12)
maintained the activity. On the contrary, the introduction of an
oxygenated (2-furanyl) or amino (2-pyridyl) heterocycle
(entries 3 and 4) was detrimental to the activity, presumably
because these are more polar and the substituents can engage in
hydrogen bonds. Likewise, the replacement of one phenyl
susbtitutent by a more polar phenylsulphonate or pentafluor-
ophenyl ring abolished the trypanocidal activity (Table 3,
entries 6−9).
Interestingly, removal of one of the phosphonium cations did
not reduce the antitrypanosomal activity (Table 4). On the
contrary, lower EC50 values were observed for 56a, 56c, 57c,
59a, and 62a compared with their bisphosphonium salt
analogues (45a, 45c, 46c, 38a, and 25a, respectively).
In Vitro Antitrypanosomal Activity. The new com-
pounds were tested for in vitro activity against T. b. rhodesiense
(strain STIB900) and a standard panel of Trypanosoma brucei
lines (s427, TbAT1-KO, and T. b. B48) with decreasing
sensitivity to diamidine drugs. The small differences in
phosphonium compounds EC50 values observed between T.
b. rhodesiense STIB900 and T. b. brucei s427 are most probably
2608
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
2609
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
Regarding the substituents pattern, the same SAR as for the
bisphosphonium analogues was observed (i.e., p-tolyl > m-tolyl
and n-hex > i-Bu ≫ Et). In this series, replacement of the
carbonyl group by an oxygen linker enhanced the in vitro
activity 20- to 40-fold against T. b. rhodesiense (compare 56a vs
56c, 57a vs 57c). The diphenyl ether analogues 56c and 57c
were the most active compounds of all the series with EC50
values in the low nanomolar range against wild type and
resistant T. b. brucei lines (Table 4, entries 4 and 6),
respectively. Despite their submicromolar and low micromolar
cytotoxicity toward L6 cells and HEK cells, respectively, both
compounds still have a reasonable therapeutic window with a
selectivity index (SI) vs L6-cells of 131 and 62 and vs HEK-cells
of 1145 and 876, respectively.
As far as the counterion is concerned, no significant
difference in in vitro activity was observed between the
bromide and the chloride salt of the compounds (Table 2,
entries 16−22). In addition, a longer linker (i.e., ethylene
instead of methylene linker) between the phosphonium cation
and the central diphenyl core hardly affected the activity as
shown by the nanomolar EC50 values of compounds 65 (73 nM
vs 63 nM for 45g) and 64 (27 nM vs 50 nM for 45f) (Table 5).
In most cases, the compounds displayed an acceptable
selectivity index toward HEK cells (>100). However, higher
cytotoxicity was found with L6 cells, indicating a notable
difference in susceptibility to phosphonium compounds
between the two cell lines.
Cross-Resistance and Drug Transport. Importantly,
EC50 values were found to be statistically identical for the
s427, TbAT1-KO, and B48 strains in all cases (Student t-test, P
> 0.05). This indicates that there is no cross-resistance between
the phosphonium cations and the crucial diamidine and
melaminophenyl arsenical classes of trypanocides. However,
the resazurin assay employed to generate EC50 values provides
a single reading after an initial 48 h incubation period with test
compound followed by a further 24 h in the additional presence
of the dye. While this allows for high throughput and
reproducible EC50 values, it gives no information on how
speedily the compounds act on the cells. The relationship
between drug concentration and minimal exposure time is
important, as it has major implications for potential drug
development: should the drug be required to be present in
circulation for 48 h at a concentration greater than EC90, this
would inevitably require a far higher dosage, and more frequent
administration, than achieving such a dose for 2 h. We thus
conducted a series of experiments to assess differential action
on the cell lines by a small selection of phosphonium salts. We
used incubations with the viability reporter dye propidium
iodide (PI), which gives a fluorescent signal upon cell entry and
binding to nucleic acid.17 Figure 1 shows the effects on cellular
integrity by 24a, recorded in real time for a period of >8 h. This
revealed that for all three strains only concentrations of ≥3.3
μM affected viability over the course of the experiment, despite
EC50 values of ≤0.20 μM. Yet the speed at which the cell
membrane integrity was compromised appeared to be slightly
less in the diamidine resistant strains. Similar observations were
made for 25b, 25c, 26a, 35a, 36a, 38a, 43a, 43c, 45d, 45e, 47a,
and 55 (see Supporting Information).
Given the very similar sensitivities, it must be deemed highly
unlikely that the cellular entry of these cations depends
critically on either of the two diamidine transporters (i.e.,
TbAT1/P2 aminopurine transporter and the high affinity
pentamidine transporter, HAPT1) that are known to be, by
2610
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
2611
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
2612
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
2613
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
a
Table 5. Antitrypanosomal Activity of Phosphonium Salts with Ethylene Linker
i
EC50 (μM) (selectivity index)
compd
salt
m
n
cytox
b
a
c
d
e
f
e
g
h
form
T. b. rhod.
L6
T. b. brucei WT
TbAT1-KO
0.15 0.02
RF
T. b. B48
RF
cytox HEK
SI
45f
45g
64
Br−
Br−
Cl−
Cl−
2
3
2
3
1
1
2
2
0.084 (7.8)
0.081 (22)
0.027 (40.3)
0.656
1.80
0.15 0.06
1.0
1.7
0.8
0.6
0.15 0.06
1.0
1.8
0.7
0.6
32.4 2.4
25.8 1.7
12.8 0.9
5.6 0.2
213
897
103
73
0.029 0.005
0.125 0.027
0.049 0.007
0.094 0.023
0.043 0.009
0.053 0.003
0.089 0.017
1.09
65
0.073 (8.8)
0.645
0.077 0.022
0.048 0.007
a
b
c
d
T. b. rhodesiense STIB900 trypomastigotes. Rat skeletal myoblast L6 cells. T. b. brucei s427 trypomastigotes. T. b. brucei knockout strain lacking a
functional P2-transporter and resistant to diminazene aceturate.18 Resistance factor compared to WT. The B48 strain is a mutant derived from the
e
f
TbAT1-KO strain with a nonfunctional high affinity pentamidine transporter (HAPT). This strain is resistant to diminazene, pentamidine, and
g
h
i
melaminophenyl arsenicals.8 Human embryonic kidney (HEK) cells. Selectivity index = [EC50(HEK cells)/EC50(T. b .brucei WT)]. Selectivity
index = [EC50(L6-cells)/EC50(T. b. rhodesiense)].
their absence, the cause of resistance in the strain B48.8,18 We
can speculate that, unlike diamidines, the positive charge(s) of
the (bis)phosphonium compounds are highly dispersed and
shielded by hydrophobic substituents, allowing transmembrane
diffusion at an appreciable rate.19 It is noted, in this context,
that the antiparasitic activity seems to be proportionate with the
level of shielding, with methyl and ethyl substituents having
virtually no effect (18a, 19a). Yet some of the phosphonium
compounds, such as 24a and 45e, were good inhibitors of the
diamidines transporters, especially of the low affinity
pentamidine transporter (LAPT1),20 for which most of the
compounds displayed higher affinity than the original substrate
pentamidine (Table 6). As there was, however, no correlation
Table 6. Affinity of Selected Phosphonium Compounds on
the HAPT1 and LAPT1 Diamidine Transporters of T. b.
a
brucei
compd
Ki, HAPT1 ( SE) (μM)
Ki, LAPT1 ( SE) (μM)
b
pentamidine
18a
0.036 0.006
>250
56.2 8.3
>250
24a
5.2 0.9
2.2 0.4
c
25a
ND
25
6
c
25b
ND
9.5 0.5
7.2 4.0
25c
53 13
c
26a
ND
39
19
49
20
7
6
6
4
c
35a
ND
c
36a
ND
c
38b
ND
45e
9.2 1.3
3.6 0.4
25 11
c
55
ND
a
Experiments were performed in triplicate and on at least three
b
separate occasions, as described in refs 8 and 20. Taken from ref 20.
c
ND, not determined.
between inhibition of the transporters and the antiparasitic
activity, we conclude that the activity is not dependent on entry
through any of the known drug transporters, with diffusion
being the likely route of entry.
Characterization of the Action of Phosphonium Salts
on T. b. brucei. Figure 1 indicated that 24a only affected
cellular viability during an 8 h experiment at concentrations
>15-fold its EC50. In fact, numerous additional bisphosphonium
salts were tested over 8 h on T. b. brucei s427 and typically
affected parasite viability only at concentrations 70−120 ×
Figure 1. Real time monitoring of the effects of 24a on various strains
of T. b. brucei. The experiment was conducted with 5 × 105 cells/well
in a 96-well format, with 200 μL/well of HMI-9 medium containing
10% FBS and 9 μM PI. Fluorescence was determined every 2 min for a
total of 250 cycles. Fluorescence was read using 544 and 620 nm filters
for excitation and emission, respectively. Fluorescence values were
determined in parallel for all three strains, using a single 96-well plate
incubated in a Fluostar Optima fluorimeter at 37 °C and 5% CO2. The
slow increase in fluorescence in drug-free controls is attributable to the
slow entry of PI into live trypanosomes.17
2614
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
EC50 (see Supporting Information). The only exception was 55,
with only concentrations >700 EC50 affecting parasite survival
within this time frame. This indicates that monophosphonium
salts may affect trypanosomes more slowly than the
corresponding dications, probably because the single positive
charge provides less driving force to cross inside-negative
membranes. In fact, previous studies by Ross et al.19 have
shown that lipophilic triphenylphosphonium dications are
accumulated into mitochondria to a greater extent than their
monocationic counterpart.
It is thus clear that while monophosphonium and bi-
sphosphonium compounds act potently on trypanosomes, their
effect on cellular integrity is relatively slow. In order to assess
the timing of the antitrypanosomal effect of phosphonium
compounds, we conducted manual cell counts on cultures
incubated with various concentrations of test compound. We
found that growth rates of T. b. brucei bloodstream forms were
initially reduced upon incubation with phosphonium salts at
concentrations of EC50 or slightly over but typically recovered
after 36 h; only at concentrations ≥2 EC50 did the cells not
recover from the initial growth inhibition and started to die
after 12 h (Figure 2A).
30 min of incubation allowed for only very limited resumption
of growth, reaching a maximum density of 9.4 × 105 cells/mL
at 48 h (i.e., 15% of the drug-free control). Removal of drug
after 4 h did not lead to any resumption of growth at all, and
the trypanosome population slowly dwindled over the next 44
h. Continuous exposure to the drug resulted in sterilization of
the culture after approximately 30 h. Highly similar results were
obtained with a number of other compounds including 25b and
45c.
CoMFA Models. In order to gain insights into the SAR of
this class of phosphonium salts, we have derived two main
CoMFA models based on the EC50 values for T. b. rhodesiense
and T. b. brucei, respectively (Tables 1−5). Both models are, at
the least, able to discriminate activities, and they possess a
moderate to high predictive power. Nevertheless our
hypothesis that both end points have a similar behavior is
confirmed by the modeling step. The two models are so alike
that explaining both in detail becomes redundant, consistent
with identical modes of action for these closely related species.
Therefore, we are presenting here only the T. b. brucei model.
For further reading about the T. b. rhodesiense model see the
Supporting Information.
The training set (54 compounds) model statistics are fairly
good: slope of 0.8905, y intercept of 0.0466, r2 = 0.891, F =
99.7. Standard validations on the training set were conducted in
several ways: “leave one out” using SAMPLS algorithm21
(SAMple-distance partial least squares), q2 = 0.43, three
components partial least squares (PLS) fitting; 10 groups
“leave several out” (LSO), q2 = 0.379, three components PLS
fitting and bootstrapping, averaged r2 = 0.856, and averaged
standard error (SE) of 0.343. Cross-validation values were not
actually promising, so some suspicions of overfitting arose.
However, the test set statistics (Figure 3), a set never used in
Although trypanosomes died only slowly on incubation with
phosphonium compounds, only a relatively short exposure time
was required. Figure 2B shows that removal of 25c after a mere
Figure 3. CoMFA model for T. b. brucei. Plot of −log10[EC50
(μM)]experimental vs −log10[EC50 (μM)]predicted for T. b. brucei WT.
Outliers are not shown. Model statistics are commented in the text.
Figure 2. Effect of 25c on cultures of T. b. brucei bloodstream forms.
Cultures in the presence or absence of various concentrations of 25c
were set up at 105 cells/mL. Microscopic cell counts were performed
in triplicate using a hemocytometer, and average values are shown. (A)
the modeling step, refuted this point. For T. b. brucei (pEC50
experimental vs pEC50 theoretical), the model statistics are as
follows: slope of 0.7531, intercept of 0.0117, and r2 = 0.744
(excluding outliers) or r2 = 0.67 (including outliers). Such
values fulfill the rules for predictive models proposed by
□
Cells were incubated in the continuous presence of 0.05 μM ( ), 0.25
○
●
▲
μM ( ), or 1 μM 25c ( ), or no compound (control, ). Inset
shows the same data but with the full growth over 48 h on a different
scale. (B) Reversibility of growth inhibition by 0.25 μM 25c was
Tropsha et al.22 (i.e., r0 = 0.743) and can be considered as
2
clearly predictive.
■ □
determined by centrifugation 1100g) after either 30 min ( , ) or 4 h
( , ) and replacement with drug-free fresh medium (closed symbols,
Aside from the good behavior of this model, a number of
outliers were found within the test set: 48a, 52b, 52c, and 63a.
Compound 48a can be considered as an outlier because of its
similar structure to 47a, a rather active compound, while no
other related structures were included. This fact clearly biased
● ○
solid lines) or fresh medium with 0.25 μM 25c (open symbols, dashed
6
▲
lines). A control culture ( , no drug) reached a density of 6 × 10
cells/mL after 48 h, as in panel A.
2615
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
Figure 4. Steric and electrostatic CoMFA fields for T. b. brucei. (A) For steric fields, regions where increasing the volume of the molecule favors
antitrypanosomal activity are green. Regions where increasing the volume decreases antitrypanosomal activity are yellow. Compounds 18a (blue),
25a (magenta), and 41a (red) are shown. (B) Electrostatics field contours indicate an increase of anti-T. brucei activity with increasing positive
(blue) and negative (red) charge, respectively. Compound 43c is shown.
the model to predict a higher activity than it actually has.
However, 48a was still predicted as inactive.
compound 18a shown in blue), while they do not offer the
essential lipophilic shielding to the phosphonium cation.
One issue derived from the good activities obtained for
monocations is a remarkable loss of symmetry on the steric
fields. While one end of the molecule is nearly surrounded by
sterically unfavorable fields, the distal end is virtually the
opposite. Such assumption is coherent with the submicromolar
activities found for dications with symmetric aromatic
substitution (e.g., see Figure 4A, compound 41a shown in
red). Here the favorable effect of bulk in one molecular end is
compensated somehow by the same unfavorable bulk in the
distal end. It is indeed the flexibility of aliphatic chains, the
ability to avoid the restricted areas, and the predominant effect
on the nanomolar activities of compounds such as 25a−c (see
Figure 4A, compound 25a showed in magenta) compared with
26a.
Electrostatic fields show that positive charge dispersion over
the whole molecule favors the activity as we previously
supposed. Here we see again a loss on the symmetry on the
fields where a bit of negatively charged substituent, near the
phosphonium (e.g., OMe, F, ...) and close to the more sterically
impeded area, would favor the activity of these compounds.
This is the case of compounds 43b and 43c bearing 2-
methoxyphenyl substituents (see Figure 4B). In such cases this
is the predominant effect confronted with the mere bulk offered
52b and 52c were not included in the modeling step because
their activities were not determined at the time of generating
this model. These compounds are outliers because they are the
only ones with neutral charge (i.e., presence of the 3-
phenylsulfonate substituent) and no compound alike was
included in the model. The counterions are not taken into
account to generate the CoMFA models (see Experimental
Section). Finally, 63a has clearly a lower activity than would be
expected based on our model. This is probably due to the lack
of lipophilic shielding around the phosphonium cation,
reducing membrane permeation. For quantitative information
see Table S2 in Supporting Information.
CoMFA fields derived from this model denote that the steric
component is rather selective (Figure 4). Indeed the steric
contribution to the CoMFA is 0.641. Bulk is strictly forbidden
around the whole molecule except upon special positions on
the phosphorus substituent. Such positions coincide with para
and meta substitutions on six-membered aromatic rings which
can also be easily occupied by long enough aliphatic chains.
This fact explains why aliphatic chains are not active up to a
certain number of carbons (≤4); molecules with such small
substituents do not fill the preferred volume (sse Figure 4A,
2616
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
by only aromatic substitution, such as in 50a. Surprisingly
enough, there is no other negative favorable area in the model.
This is combined with a huge dispersion of the positive charge
upon the molecule even surrounding the negative surfaces. We
can deduce then a minor influence of localized electrostatic
interactions, such as hydrogen bonds, with the phosphonium
salts activities.
improved activity is in many cases also coupled to somewhat
increased toxicity against the mammalian cell lines employed.
The redundancy of the second phosphonium group is
consistent with the observation that longer, more flexible
linkers (i.e., ethylene instead of methylene linker) between the
phosphonium cation and the central diphenyl core hardly
affected the antitrypanosomal activity. This observation is
important because monophosphonium salts are smaller
molecules with physicochemical properties (i.e., lower molec-
ular weight, lower clogP) more likely to provide good
pharmacokinetic behavior in vivo compared with the bis-
phosphonium counterparts.
DISCUSSION
■
Phosphonium salts bearing hydrophobic substituents are
lipophilic cations with a delocalized charge. This affords
specific properties to these molecules such as the capacity to
cross biological membranes driven by electrical potential. Thus,
they tend to accumulate in organelles with high membrane
potential such as mitochondria.23 In Leishmania, several of the
benzophenone-derived bisphosphonium salts reported here
were found to target mitochondria of the parasite, a key to their
leishmanicidal action.5 Ongoing studies on its mechanism of
action in T. brucei also point to a mitochondrial target; for
instance, phosphonium compounds strongly affected the
mitochondrial membrane potential (unpublished results).
Interestingly, the SAR of the series of compounds presented
here shows some similarities with the SAR against Leishmania.
For instance, for compounds with homoalkyl groups, optimum
activity was observed with five to eight carbons, independent of
the linker (Tables 1 and 4). In the same way, the combination
of two phenyl groups and one alkyl substituent was enough to
get submicromolar activity whereas the combination of one
phenyl group and two short alkyl substituents (Me, Et) led to
micromolar EC50. These observations are clearly supported by
the CoMFA models that show that a minimum bulk around the
phosphonium cations is favorable to the activity, but this
volume is also limited in size.
In their 1979 paper,4 Kinnamon and Steck reported that the
monophosphonium compound 54 (chloride salt) displayed
100% curative activity (53 mg/kg sc) in a murine model of T. b.
rhodesiense infection. The bromide analogue 55 was found to be
less active in vivo than its chloride counterpart. However, no in
vitro data were reported. The results of our in vitro screening
are in agreement with these data, as 54 displayed between
submicromolar and nanomolar activities against T. b. brucei and
T. b. rhodesiense, respectively.
The excellent in vitro antitrypanosomal activity observed
with this set of phosphonium compounds confirms the
potential of this class of molecules. We show that phosphonium
compounds irreversibly inhibit trypanosome growth, leading to
certain clearance of the population after only a minimum
exposure time. This situation is similar to that of the highly
successful class of diamidine trypanocides, which also require
only a brief exposure time to ultimately kill the trypanosomes
after 1−2 days.26 This result is important for potential
therapeutic applications, as dosage can be kept relatively low
and a single administration may be sufficient, similar to the
single administration of the diamidine diminazene aceturate
that is the routine treatment for animal trypanosomiasis.2 In
addition, the positive correlation with lipophilicity and thus
membrane penetration should increase penetration into the
central nervous system, essential for treatment of late stage
trypanosomiasis.3
From these data, it is clear that a substantial level of lipophilic
shielding around the phosphonium cation(s) is necessary to get
high levels of activity against Trypanosoma and Leishmania
species, presumably by increasing membrane permeation. No
correlation was found between the antiparasitic activity and
inhibition of the known T. brucei drug transporters (P2,
HAPT1, and LAPT1). Although some bis-phosphonium
compounds displayed surprisingly high affinity for the
HAPT1 and LAPT1 transporters, most of the bis-phosphonium
compounds would be considered too large to be a substrate (as
opposed to be an inhibitor) of such transporters (e.g.,
molecular weights range between 800 and 1300 compared
with 340 for pentamidine). HAPT1 and LAPT1 were first
described to transport a range of diamidines,20,24 which, like the
bis-phosphonium compounds described here, are dications.
Our CoMFA models show the importance of considerable
positive charge dispersion over a large surface for increased
activity. On balance, we conclude that, similar to a series of
dicationic choline-derived compounds on which we recently
reported,25 the antitrypanosomal activity is not dependent on
entry through any of these transporters. In fact, it is highly
probable that diffusion is the route of entry. This interpretation
is much strengthened by the observation that substituents such
as furan, pyridine, and phenylsulfonate, which can engage in H-
bonding, display significantly reduced antitrypanosomal
activities, presumably because of impeded cellular penetration.
The data presented in Table 4 clearly show that mono-
phosphonium compounds have as good and often slightly
better activity against trypanosomes than the corresponding
symmetrical bis-phosphonium analogues. However, this
CONCLUSIONS
■
The in vitro screening of a set of 83 phosphonium salt
derivatives against African trypanosomes allowed the discovery
of several trypanocidal compounds with nanomolar activities
and adequate selectivities vs HEK and L6 cells. SAR studies
showed that bulky substituents around the phosphonium
cations (i.e., either C5−C8 homoalkyl chains or phenyl rings)
are necessary to get submicromolar activities, whereas the
linkers have less influence on the antitrypanosomal activity.
However, this statement is true as far as the bisphosphonium
salts are concerned. On the contrary, the diphenyl ether linker
seemed to be preferred in the case of monophosphonium
compounds. In fact, 56c and 57c were the most active
compounds of all the series with EC50 values in the low
nanomolar range against wild type and resistant T. brucei lines.
Phosphonium compounds displayed a very similar struc-
ture−activity relationship against T. brucei and Leishmania.5
This observation strongly suggests a similar mechanism of
action against the two parasite species and indicates that this
class may have broad antiprotozoal activity. This is the subject
of ongoing investigations in our laboratories.
Importantly, the screening reported here confirmed that
there is no cross-resistance between the new compounds and
2617
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
procedure with trihexylphosphine and 1a for 48 h at 80 °C. The
product was obtained as a hygroscopic oily solid after workup II.
Recrystallization from MeOH/Et2O yielded a yellowish hygroscopic
solid (27%). HPLC = 89% pure. 1H NMR (300 MHz, CDCl3) δ 7.63
(d, J = 7.7, 4H, ArH), 7.50 (d, J = 7.7, 4H, ArH), 4.67 (d, J = 16.1, 4H,
PCH2Ph), 2.29 (m, 12H, PCH2CH2), 1.6−1.1 (m, 48H, (CH2)4), 0.84
(m, 18H, CH3). 13C NMR (75 MHz, CDCl3) δ 194.9, 136.5 (d, J =
3.4), 134.3 (d, J = 8.7), 130.6 (d, J = 4.8), 129.8 (d, J = 79.7), 31.1,
30.6 (d, J = 14.7), 27.0 (d, J = 44.4), 22.4, 21.9 (d, J = 4.6), 19.2 (d, J =
46.1), 14.0. LRMS (ES+) m/z 779.86 [(M − H)+], 390.11 [M2+,
100%]. ESI-HRMS m/z 390.3233 [M2+] (C51H90OP2 requires
390.3228).
existing diamidines and arsenical trypanocides, an essential
condition for any new preclinical candidate.
Finally, the CoMFA models developed in this work should
help in the design of new custom-made phosphonium
derivatives with improved activity against African trypano-
somes.
EXPERIMENTAL SECTION
■
Chemistry. All dry solvents were purchased from Aldrich or Fluka
in Sure/Seal bottles. All reactions requiring anhydrous conditions or
an inert atmosphere were performed under argon atmosphere. All
reactions were monitored by thin layer chromatography (TLC) using
silica gel 60 F254 plates (Merck) or HPLC−MS. Chromatography was
performed with Isolute SI prepacked columns. 1H and 13C NMR
spectra were recorded on a Bruker Advance 300 or Varian Inova 400
spectromether. Chemical shifts of the 1H NMR spectra were internally
referenced to the residual proton resonance of the deuterated solvents:
CDCl3 (δ 7.26 ppm), D2O (δ 4.6 ppm), CD3OD (δ 3.49 ppm), and
DMSO (δ 2.49 ppm). J values are given in Hz. Melting points were
determined in open capillary tubes with a Stuart Scientific SMP3
apparatus or Mettler Toledo MP70 melting point system and are
uncorrected. All compounds are >95% pure by HPLC or combustion
analysis unless otherwise noted. Elemental analysis was performed on a
Heraeus CHN-O rapid analyzer. Analytical results were within 0.4%
of the theoretical values unless otherwise noted. Analytical HPLC−MS
was run with an Xbridge C18−3.5 μm (2.1 mm × 100 mm) column
on a Waters 2695 separation module coupled with a Waters
Micromass ZQ spectrometer using electrospray ionization (ESI+).
The following HPLC conditions were used: column temperature of 30
°C, gradient time of 5 min, H2O/CH3CN (10:90 → 90:10) (HCO2H,
0.1%), flow rate of 0.25 mL/min, UV detection using diode array (λ =
190−400 nm). Semipreparative HPLC−MS was run with a SunFire
Prep C18, 5 μm (19 mm × 150 mm) column on a Waters separation
module (Waters 2545/SFO/2767) coupled to a Waters 3100 mass
detector using ESI+. The fractions were collected with a Waters 2767
sampler manager.
4,4′-Bis((trihexylphosphonio)methyl)diphenylmethane Di-
bromide (25b). The reaction was carried out following the general
procedure with trihexylphosphine and 1b. The mixture was
concentrated to ∼1 mL to give a yellowish oil that was diluted with
toluene (10 mL). Addition of Et2O (10 mL) caused precipitation of an
oily residue. The flask was stored in the refrigerator overnight. The
supernatant was removed, and the oily precipitate was rinsed with
toluene. Et2O (10 mL) was added to the oily precipitate, which was
triturated with a spatula to yield 25b as a white solid (158.7 mg, 64%):
mp 111−112 °C with previous softening (DMF/toluene); HPLC >
1
94% pure. H NMR (300 MHz, CDCl3) δ 7.38 (dd, J = 8.3, 2.0, 4H,
ArH), 7.08 (d, J = 8.3, 4H, ArH), 4.23 (d, J = 15.1, 4H, PCH2), 3.89 (s,
2H, PhCH2Ph), 2.33 (br s, 12H, CH2), 1.42 (br s, 24H, CH2), 1.23 (br
s, 24H, CH2), 0.85 (t, J = 6.6, 18H, CH3). 13C NMR (75 MHz,
CDCl3) δ 140.8, 130.5 (d, J = 4.9), 130.0 (d, J = 2.8), 126.5 (d, J =
8.8), 41.1, 31.1, 30.6 (d, J = 14.6), 26.7 (d, J = 45.2), 22.4, 21.9 (d, J =
4.8), 19.1 (d, J = 46.3), 14.0. LRMS (ES+) m/z 765.87 [(M − H)+],
383.04 [M2+, 100%]. ESI-HRMS m/z 383.3357 [M2+] (C51H92P2
requires 383.3332).
4, 4′-Bis((phenyldimethylphosphonio)methyl)-
diphenylethane Dibromide (28f). The reaction was carried out
following the general procedure with phenyldimethylphosphine (68.94
mg, 0.497 mmol) and 10f (73.2 mg, 0.199 mmol). The mixture was
concentrated under vacuum until the formation of an oily solid which
was diluted with DMF. Et2O was added to precipitate the product as a
white hygroscopic solid. The flask was allowed to stand in the freezer
overnight. The solid was collected, rinsed with Et2O, and dried under
vacuum (106.5 mg, 98%). HPLC > 95% pure; mp >300 °C. 1H NMR
(300 MHz, CDCl3) δ 7.80 (dd, J = 12.4, 7.5, 4H, ArH), 7.66 (m, 2H,
ArH), 7.57 (m, 4H, ArH), 6.98 (dd, J = 3.0, 7.2, 4H, ArH), 6.79 (d, J =
7.5, 4H, ArH), 4.26 (d, J = 15.4, 4H, PCH2), 2.72 (s, 4H; CH2CH2),
2.30 (d, J = 13.8, 12H, CH3). 13C NMR (75 MHz, CDCl3) δ 141.4 (d,
J = 4.1), 134.6 (d, J = 1.8), 131.3 (d, J = 9.7), 130.2 (d, J = 5.2), 130.0
(d, J = 12.4), 129.4 (d, J = 2.9), 125.4 (d, J = 9.1), 119.8 (d, J = 83.6),
36.9, 31.2 (d, J = 48.0), 7.5 (d, J = 55.6). LRMS (ES+) m/z = 483.00
[(M − H)]+. ESI-HRMS m/z 242.1241 [M2+] (C32H38P2 requires
242.1219).
4, 4′-Bis((phenyldimethylphosphonio)methyl)-
diphenylpropane Dibromide (28g). The reaction was carried out
following the general procedure with phenyldimethylphosphine (66
mg, 0.478 mmol) and 11g (73.1 mg, 0.191 mmol). The mixture was
concentrated under vacuum until the formation of an oily solid which
was diluted with acetone. Et2O was added to produce precipitation of
the product as a dark brown hygroscopic solid (67.6 mg, 54%). HPLC
= 90% pure. 1H NMR (300 MHz, CDCl3) δ 7.86−7.66 (m, 7H, ArH),
7.66−7.57 (m, 3H, ArH), 6.98 (t, J = 6.9, 8H, ArH), 4.29 (d, J = 15.3,
4H, PCH2), 2.58−2.45 (m, 4H, PhCH2CH2), 2.37 (d, J = 13.7, 12H,
CH3), 1.80 (dd, J = 19.2, 11.6, 2H, PhCH2CH2). 13C NMR (75 MHz,
CDCl3) δ 142.4 (d, J = 4.3), 134.6 (d, J = 3.2), 131.9 (d, J = 9.6), 130.5
(d, J = 5.5), 130.1 (d, J = 12.2), 129.2 (d, J = 3.4), 125.3 (d, J = 9.4),
120.2 (d, J = 83.0), 35.1, 32.3, 31.3 (d, J = 48.1), 7.8 (d, J = 55.5).
LRMS (ES+) m/z = 497.48 [(M − H)]+, 248.95 [M2+, 100%]. ESI-
HRMS m/z 249.1316 [M2+] (C33H40P2 requires 249.1297).
1. General Procedure for the Synthesis of the Bisphospho-
nium Salts. A Kimax tube was charged with the appropriate bis-
halogenated precursor (100 mg, ∼0.28 mmol) and flushed with argon.
Anhydrous DMF (3 mL) was added followed by the phosphine (1.12
mmol, 4 equiv). The tube was flushed with argon, stopped, and the
reaction mixture was stirred at 100 °C for 20 h. A higher temperature
(150 °C) and longer reaction time were necessary with the 4,4′-
bischloroethyl linkers 16f and 17g. Different workup procedures were
used depending on whether the product precipitated from the reaction
mixture or not. For workup I, the mixture was allowed to cool to room
temperature and the precipitated product was collected by filtration,
rinsed successively with toluene and Et2O, and dried under vacuum.
For workup II, the reaction mixture was transferred to a flask. Then
toluene (10−20 mL) was added to precipitate the product. The flask
was stored in the refrigerator overnight. The supernatant was removed,
and the precipitate was rinsed with toluene. Et2O (10 mL) was added,
and the precipitate was triturated with a spatula. The solid was
collected, rinsed with Et2O, and dried under vacuum.
4,4′-Bis((triethylphosphonio)methyl)diphenylmethane Di-
bromide (19b). The reaction was carried out in toluene following
the general procedure with triethylphosphine and 1b. The product was
obtained as a white hygroscopic solid (107 mg, 65%) following
workup I procedure and recrystallization in DMF/toluene. HPLC =
91% pure; mp 246−248 °C. 1H NMR (300 MHz, CDCl3) δ 7.42 (dd,
J = 7.7, 2.1, 4H, ArH), 7.07 (d, J = 7.7, 4H, ArH), 4.23 (d, J = 15.2, 4H,
PCH2), 3.88 (s, 2H, PhCH2Ph), 2.46 (dq, J = 20.5, 7.7, 12H,
CH2CH3), 1.23 (dt, J = 17.9, 7.7, 18H, CH2CH3). 13C NMR (75 MHz,
CDCl3) δ 156.9, 131.9, 123.2, 119.9, 25.6, 25.0, 12.2 (J = 48.2), 6.2 (J
= 5.4). LRMS (ES+) m/z = 429.39 [(M − H)]+, 214.93 [M2+, 100%].
Anal. Calcd for C27H44Br2P2: C, 54.93; H, 7.51; Br, 26.91. Found: C,
54.75; H, 7.72; Br, 26.20.
4,4′-Bis((diphenyl-p-tolylphosphonio)methyl)diphenyl
Ether Dibromide (38c). The reaction was carried out following the
general procedure with diphenyl-p-tolylphosphine and 1c. The
product was obtained as a white solid (187 mg, 66%) following
workup II procedure: mp 230 °C (dec); HPLC > 97% pure. 1H NMR
4,4′-Bis((trihexylphosphonio)methyl)benzophenone Dibro-
mide (25a). The reaction was carried out following the general
2618
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
1
(300 MHz, CDCl3) δ 7.79−7.54 (m, 28H, ArH), 7.42 (dd, J = 7.9, 2.8,
4H, ArH), 7.08 (dd, J = 8.6, 2.4, 4H, ArH), 5.32 (d, J = 14.1, 4H,
PCH2), 2.45 (s, 6H, CH3). 13C NMR (75 MHz, CDCl3) δ 156.8 (d, J
= 4.0), 146.6 (d, J = 3.0), 135.0 (d, J = 2.7), 134.5 (d, J = 9.9), 133.2
(d, J = 5.4), 131.1 (d, J = 12.9), 130.2 (d, J = 12.5), 122.4 (d, J = 8.6),
119.1 (d, J = 3.0), 118.1 (d, J = 85.7), 114.0 (d, J = 88.1), 30.1 (d, J =
47.3), 22.0. LRMS (ES+) m/z 747.52 [(M − H)+], 373.94 [M2+,
100%]. ESI-HRMS m/z 374.1536 [M2+] (C52H46OP2 requires
374.1506).
softening (DMF/EtOH); HPLC > 93% pure. H NMR (300 MHz,
CDCl3) δ 7.54 (dd, J = 12.2, 8.2, 12H, ArH), 7.45−7.36(m, 12H,
ArH), 7.08 (dd, J = 8.4, 2.2, 4H, ArH), 6.66 (d, J = 8.4, 4H, ArH), 5.18
(d, J = 14.0, 4H, PCH2), 2.44 (s, 18H, CH3). 13C NMR (75 MHz,
CDCl3) δ 156.8 (d, J = 3.9), 146.3 (d, J = 3.2), 134.3 (d, J = 10.1),
133.1 (d, J = 5.4), 131.0 (d, J = 12.9), 122.6 (d, J = 8.5), 119.1 (d, J =
3.0), 114.7 (d, J = 88.4), 30.4 (d, J = 48.2), 22.0. LRMS (ES+) m/z
803.59 [(M − H)+], 401.93 [M2+, 100%]. ESI-HRMS m/z 402.1857
[M2+] (C56H54OP2 requires 402.1819).
4,4′-Bis((tri-4-methoxyphenylphosphonio)methyl)-
benzophenone Dibromide (41a). The reaction was carried out
following the general procedure with tris(4-methoxyphenyl)phosphine
and 1a for 22 h. The product was obtained as a white solid (214.5 mg,
74%) following workup I procedure: mp 244.2−245 °C; HPLC > 95%
4,4′-Bis((tri(p-tolyl)phosphonio)methyl)diphenylacetamide
Dibromide (45e). The reaction was carried out following the general
procedure at 100 °C with tri-p-tolylphosphine and 1e. The product
was obtained as a highly hygroscopic white solid (72 mg, 53%)
1
following workup II procedure. HPLC > 96% pure. H NMR (300
1
MHz, CDCl3) δ 7.63−7.47 (m, 14H, ArH), 7.41 (m, 10H, ArH), 7.15
(m, 4H, ArH), 6.94 (d, J = 7.4, 4H, ArH), 5.25 (d, J = 15.8, 4H,
PCH2), 2.45 (s, 18H, CH3), 1.94 (s, 3H, COCH3). 13C NMR (75
MHz, CDCl3) δ 170.4, 146.5 (d, J = 2.7), 142.7 (d, J = 3.0), 134.3 (d, J
= 10.2), 133.4−132.1 (br), 131.1 (d, J = 12.9), 114.5 (d, J = 88.6), 30.8
(d, J = 48.2), 24.0, 22.0. LRMS (ES+) m/z 844.61 [(M − H)+], 422.58
[M2+, 100%]. ESI-HRMS m/z 422.6984 [M2+] (C58H57NOP2 requires
422.6952).
pure. H NMR (300 MHz, CDCl3) δ 7.64 (dd, J = 11.9, 8.8, 12H,
ArH), 7.34 (d, J = 7.7, 4H, ArH), 7.25−7.19 (m, 4H, ArH), 7.07 (dd, J
= 8.8, 2.4, 12H, ArH), 5.38 (d, J = 15.2, 4H, PCH2), 3.88 (s, 18H,
CH3). 13C NMR (75 MHz, CDCl3) δ 195.3, 164.7 (d, J = 2.9), 136.5
(d, J = 11.5), 133.6 (d, J = 8.6), 131.7 (d, J = 5.3), 130.1 (d, J = 2.7),
115.9 (d, J = 13.8), 109.1, 107.9, 56.0, 31.6 (d, J = 48.7). LRMS (ES+)
m/z 911.46 [(M − H)+], 455.97 [M2+, 100%]. ESI-HRMS m/z
456.1683 [M2+] (C57H54O7P2 requires 456.1667).
4,4′-Bis((tri-2-thienylphosphonio)methyl)benzophenone Di-
bromide (47a). The reaction was carried out following the general
procedure with tri-2-thienylphosphine and 1a (110 mg, 0.3 mmol) for
18 h. The product was obtained as an off-white solid after workup II
and recrystallization from EtOH (200 mg, 72%). HPLC > 95% pure.
1H NMR (300 MHz, CDCl3) δ 8.12 (m, 12H, ArH), 7.41 (dt, J = 6.8,
6.2, 10H, ArH), 7.24 (dd, J = 8.2, 2.6, 4H, ArH), 5.26 (d, J = 15.4, 4H,
PCH2). 13C NMR (75 MHz, CDCl3) δ 195.3, 143.1 (d, J = 11.8),
140.3 (d, J = 5.4), 137.1 (d, J = 4.3), 132.0 (d, J = 9.5), 131.5 (d, J =
6.1), 131.0 (d, J = 16.1), 130.4 (d, J = 3.6), 117.1 (d, J = 109.4), 36.0
(d, J = 53.4). LRMS (ES+) m/z 767.1 [(M − H)+], 383.74 [M2+,
100%]. ESI-HRMS m/z 384.0085 [M2+] (C39H30OS6P2 requires
384.0043).
4 , 4 ′ - B i s ( ( t r i ( 1 - na p ht h y l ) p h os p ho n io ) m e t h yl ) -
diphenylmethane Dibromide (53b). The reaction was carried out
following the general procedure with tri-1-naphthylphosphine and 1b.
The product was obtained as a white solid (229.2 mg, 71%) following
workup I procedure and recrystallization from EtOH/Et2O: mp 230
°C (dec); HPLC > 95% pure. 1H NMR (300 MHz, CDCl3) δ 8.56 (d,
J = 11.5, 6H, ArH), 8.23 (d, J = 7.5, 6H, ArH), 7.89 (dd, J = 26.8, 8.0,
12H, ArH), 7.68 (br s, 6H, ArH), 7.41 (dt, J = 14.7, 7.1, 12H, ArH),
6.72 (d, J = 7.0, 4H, ArH), 6.25 (d, J = 7.0, 4H, ArH), 5.66 (d, J = 11.3,
4H, PCH2), 3.30 (s, 2H, CH2). 13C NMR (75 MHz, CDCl3) δ 140.5
(d, J = 2.6), 138.5 (d, J = 10.8), 137.1 (d, J = 2.8), 134.3 (d, J = 9.6),
132.6 (d, J = 8.4), 130.7 (d, J = 6.7), 130.5, 129.0, 128.8 (d, J = 2.1),
127.6, 126.7 (d, J = 7.5), 125.9−125.6 (m), 114.6 (d, J = 80.9), 40.6,
33.3 (d, J = 48.1). LRMS (ES+) m/z 1017.77 [(M − H)+], 508.96
[M2+, 100%]. ESI-HRMS m/z 509.1968 [M2+] (C75H56P2 requires
509.1923).
4,4′-Bis((trip-tolylphosphonio)ethyl)diphenylethane Di-
chloride (64). The reaction was carried out following the general
procedure with tri-p-tolylphosphine and 16f for 15 days. The mixture
was evaporated to dryness forming a yellowish oil which was purified
by semipreparative HPLC−MS. The following HPLC conditions were
used: column temperature of 25 °C, gradient time of 10 min, CH3CN/
H2O (30:70 → 100:0) (HCO2H, 0.1%), flow rate of 0.24 mL/min,
UV detection using diode array (λ = 230 nm). tR = 6.38 min. The
fractions containing the pure product were combined and lyophilized
to give a white hygroscopic solid (12.6 mg, 5%). HPLC > 99% pure.
1H NMR (300 MHz, CDCl3) δ 7.62 (dd, J = 12.3, 8.1, 12H, ArH),
4,4′-Bis((tri(2-methoxyphenyl)phosphonio)methyl)-
diphenylmethane Dibromide (43b). The reaction was carried out
following the general procedure at 50 °C instead of 100 °C with tri-2-
methoxyphenylphosphine and 1b. The product was obtained as a
white solid (259.7 mg, 93%) following workup II procedure: mp 210−
213 °C with previous softening (DMF/toluene); HPLC > 97% pure.
1H NMR (300 MHz, CDCl3) δ 7.73 (t, J = 7.9, 6H, ArH), 7.33−7.23
(m, 8H, ArH), 7.18−7.09 (m, 10H, ArH), 6.88 (d, J = 8.0, 4H, ArH),
6.79 (d, J = 8.0, 4H, ArH), 4.58 (d, J = 15.9, 4H, PCH2), 3.73 (s, 2H,
CH2), 3.63 (s, 18H, CH3). 13C NMR (75 MHz, CDCl3) δ 161.4 (d, J =
2.3), 140.5 (dd, J = 3.4, 1.2), 137.4 (d, J = 1.9), 135.2 (d, J = 8.2),
130.0 (d, J = 7.2), 129.1 (d, J = 2.3), 128.2 (d, J = 8.0), 122.1 (d, J =
12.8), 112.9 (d, J = 6.7), 105.7 (d, J = 91.8), 56.5, 40.9, 30.7 (d, J =
52.7). LRMS (ES+) m/z 897.60 [(M − H)+], 448.90 [M2+, 100%].
ESI-HRMS m/z 449.1816 [M2+] (C57H56O6P2 requires 449.1771).
4,4′-Bis((tri(2-methoxyphenyl)phosphonio)methyl)diphenyl
Ether Dibromide (43c). The reaction was carried out following the
general procedure at 50 °C instead of 100 °C with tri-2-
methoxyphenylphosphine and 1c. The product was obtained as a
white solid (231.3 mg, 85%) following workup II procedure: mp 186−
187 °C with previous softening (DMF/toluene); HPLC > 96% pure.
1H NMR (300 MHz, CDCl3) δ 7.74 (t, J = 7.8, 6H, ArH), 7.37−7.27
(m, 6H, ArH), 7.19−7.11 (m, 12H, ArH), 6.95 (dd, J = 17.0, 15.0, 4H,
ArH), 6.58 (d, J = 8.3, 4H, ArH), 4.62 (d, J = 15.6, 4H, PCH2), 3.70 (s,
18H, CH3). 13C NMR (75 MHz, CDCl3) δ 161.4 (d, J = 2.3), 156.4
(d, J = 3.5), 137.4 (d, J = 1.6), 135.2 (d, J = 8.2), 131.5 (d, J = 7.2),
125.4 (d, J = 7.8), 122.2 (d, J = 12.8), 119.0 (d, J = 2.2), 112.9 (d, J =
6.8), 105.8 (d, J = 91.9), 56.5, 30.3 (d, J = 52.6). LRMS (ES+) m/z
899.63 [(M − H)+], 449.88 [M2+, 100%]. ESI-HRMS m/z 450.1699
[M2+] (C56H54O7P2 requires 450.1667).
4,4′-Bis((tri-p-tolylphosphonio)methyl)benzophenone Di-
bromide (45a). The reaction was carried out following the general
procedure with tri-p-tolylphosphine and 1a. The product was obtained
as a white solid (213.9 mg, 78%) following workup II procedure: mp
1
308.8 °C (dec); HPLC > 95% pure. H NMR (300 MHz, CDCl3) δ
7.61 (dd, J = 12.4, 8.1, 12H, ArH), 7.40 (dd, J = 8.1, 3.1, 12H, ArH),
7.35 (d, J = 8.2, 4H, ArH), 7.25−7.20 (m, 4H, ArH), 5.50 (d, J = 15.2,
4H, PCH2), 2.45 (s, 18H, CH3). 13C NMR (75 MHz, CDCl3) δ 195.4,
146.3 (d, J = 3.0), 136.5 (d, J = 3.7), 134.5 (d, J = 10.3), 133.2 (d, J =
8.6), 131.8 (d, J = 5.4), 131.0 (d, J = 13.1), 130.1 (d, J = 2.8), 114.6 (d,
J = 88.7), 30.9 (d, J = 47.8), 22.0. LRMS (ES+) m/z 815.42 [(M −
H)+], 407.81 [M2+, 100%]. ESI-HRMS m/z 408.1852 [M2+]
(C57H54OP2 requires 408.1819).
4,4′-Bis((tri-p-tolylphosphonio)methyl)diphenyl Ether Di-
bromide (45c). The reaction was carried out following the general
procedure with tri-p-tolylphosphine and 1c. The product was obtained
as a white solid (149.3 mg, 57%) following workup II procedure and
recrystallization from EtOH/Et2O: mp >300 °C with previous
7.54−7.41 (m, 12H, ArH), 7.09 (d, J = 7.8, 4H, ArH), 6.95 (d, J = 7.8,
4H, ArH), 3.79−3.58 (m, 4H, PCH2), 2.99−2.84 (m, 4H,
PhCH2CH2P), 2.81 (s, 4H, CH2CH2), 2.48 (s, 18H, CH3). 13C
NMR (75 MHz, CDCl3) δ 146.8, 142.6 (d, J = 2.5), 136.7, 133.5 (dd, J
= 10.1, 2.4), 132.2 (d, J = 10.4), 129.3 (d, J = 12.6), 128.7, 114.6 (dd, J
= 88.1, 2.6), 37.6, 35.0, 28.1 (d, J = 1.4), 22.0. LRMS (ES+) m/z =
843.39 [(M − H)]+, 422.00 [M2+, 100%]. ESI-HRMS m/z 422.2181
[M2+] (C60H62P2 requires 422.2158).
2619
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
4,4′-Bis((trip-tolylphosphonio)ethyl)diphenylpropane Di-
chloride (65). The reaction was carried out following the general
procedure with tri-p-tolylphosphine and 17g for 19 days. The mixture
was evaporated to dryness forming a yellowish oil which was purified
by semipreparative HPLC−MS, eluting with H2O/CH3CN. The
following HPLC conditions were used: column temperature of 25 °C,
gradient time of 15 min, CH3CN/H2O (30:70 → 70:30) (HCO2H,
0.1%), flow rate of 0.24 mL/min, UV detection using diode array (λ =
240 nm). tR = 10.32 min. The fractions containing the pure product
were combined and lyophilized to give a white hygroscopic solid (17.3
mg, 17%). HPLC > 99% pure. 1H NMR (300 MHz, CDCl3) δ 7.72−
7.55 (m, 12H, ArH), 7.53−7.43 (m, 12H, ArH), 7.14 (d, J = 7.4, 4H,
ArH), 7.04 (d, J = 7.4, 4H, ArH), 3.78 (br m, 4H, PCH2), 2.95 (br m,
4H, CH2), 2.60−2.41 (m, 22H, CH2 and CH3), 1.86 (dt, J = 14.3, 7.4,
2H, CH2CH2CH2). 13C NMR (75 MHz, CDCl3) δ 146.4 (dd, J = 5.1,
3.0), 142.3 (d, J = 2.8), 135.5 (d, J = 24.9), 133.6 (dd, J = 10.4, 2.6),
132.2 (d, J = 10.3), 129.3 (d, J = 12.5), 128.6 (d, J = 7.2), 115.1 (dd, J
= 88.4, 7.2), 34.6, 33.0, 28.1 (d, J = 3.1), 24.8 (d, J = 50.1), 22.0.
LRMS (ES+) m/z = 857.88 [(M − H)]+, 429.21 [M2+, 100%]. ESI-
HRMS m/z 429.2277 [M2+] (C61H64P2 requires 429.2236).
2. General Procedure for the Synthesis of the Mono-
phosphonium Salts. The appropriate halogenated precursor (100
mg, ∼0.36 mmol) was added to a Kimax tube and dissolved in
anhydrous DMF (3 mL) under argon atmosphere. The phosphine was
then added (0.72 mmol, 2 equiv), the tube was stopped, and the
reaction mixture was stirred at 100 °C for 20 h. Next, the reaction
mixture was transferred to a flask, and DMF was evaporated under
vacuum. Subsequently, toluene (10−20 mL) was added to precipitate
the product and the flask was stored in the refrigerator overnight. The
supernatant was removed, and the precipitate was rinsed with toluene.
Et2O (10 mL) was added, and the precipitate was triturated with a
spatula. The solid was collected, rinsed with Et2O, and dried under
vacuum.
(4-Benzoylbenzyl)tri-p-tolylphosphonium Bromide (56a).
The reaction was carried out following the general procedure with
2a and tri-p-tolylphosphine. The product was obtained as a white solid
following workup II procedure (117.6 mg, 56%). HPLC > 95% pure;
mp 233 °C. 1H NMR (300 MHz, CDCl3) δ 7.68 (dd, J = 3.0, 6.5, 2H,
ArH), 7.62−7.50 (m, 9H, ArH), 7.47−7.32 (m, 8H, ArH), 7.28 (dd, J
= 2.3, 8.0, 2H, ArH), 5.44 (d, J = 15.0, 2H, PCH2), 2.42 (s, 9H, CH3).
13C NMR (75 MHz, CDCl3) δ 196.3, 146.4 (d, J = 3.0), 137.2 (d, J =
3.8), 137.1, 134.3 (d, J = 10.3), 132.8, 132.7, 131.7 (d, J = 5.4), 130.9
(d, J = 13.0), 130.2 (d, J = 3.1), 130.1, 128.4, 114.4 (d, J = 88.7), 31.1
(d, J = 48.3), 21.9. LRMS (ES+) m/z = 499.44 [M+]. ESI-HRMS m/z
499.2203 [M+] (C35H32OP requires 499.2185).
(4-Phenoxybenzyl)tri-p-tolylphosphonium Bromide (56c).
The reaction was carried out following the general procedure with
2c and tri-p-tolylphosphine. The product was obtained as a white solid
following workup II procedure (94.7 mg, 87%). HPLC > 95% pure;
mp 192 °C. 1H NMR (300 MHz, CDCl3) δ 7.56 (dd, J = 7.8, 12.2, 6H,
ArH), 7.47−7.00 (m, 11H, ArH), 6.93 (m, 2H, ArH), 6.76 (d, J = 8.1,
2H, ArH), 5.20 (d, J = 13.9, 2H, PCH2), 2.45 (s, 9H, CH3). 13C NMR
(75 MHz, CDCl3) δ 157.5 (d, J = 3.8), 156.7, 146.3 (d, J = 2.8), 134.4
(d, J = 10.1), 133.1 (d, J = 5.3), 131.0 (d, J = 12.9), 129.1, 123.8, 122.0
(d, J = 8.5), 119.1, 119.0 (d, J = 2.9), 114.8 (d, J = 88.3), 30.6 (d, J =
48.4), 22.0. LRMS (ES+) m/z = 487.40 [M+]. ESI-HRMS m/z
487.2206 [M+] (C34H32OP requires 487.2185).
(4-Benzoylbenzyl)tri-n-hexylphosphonium Bromide (62a).
The reaction was carried out following the general procedure with
2a and tri-n-hexylphosphine. The product was obtained as a white
solid following workup II procedure (84.3 mg, 52%). HPLC > 95%
pure; mp 109.1 °C. 1H NMR (300 MHz, CDCl3) δ 7.81−7.64 (m, 6H,
ArH), 7.63−7.55 (m, 1H, ArH), 7.47 (t, J = 7.5, 2H, ArH), 4.58 (d, J =
15.9, 2H, PhCH2), 2.51−2.29 (m, 6H, PCH2CH2), 1.42 (s, 12H, CH2),
1.33−1.12 (m, 12H, CH2), 0.84 (t, J = 6.7, 9H, CH3). 13C NMR (75
MHz, CDCl3) δ 195.9, 137.6 (d, J = 3.6), 137.0, 133.7 (d, J = 8.9),
133.0, 131.0 (d, J = 2.9), 130.4 (d, J = 4.9), 129.3 (d, J = 115.5), 31.1,
30.6 (d, J = 14.8), 22.4, 22.0 (d, J = 4.8), 19.2 (d, J = 46.1), 14.0.
LRMS (ES+) m/z = 482.36 [(M + H)]+. Anal. Calcd for C32H50BrOP:
C, 68.56; H, 8.81; Br, 14.25. Found: C, 68.59; H, 8.90; Br, 14.41.
Biology. In Vitro Antitrypanosomal Activity. The in vitro
trypanocidal and cytotoxic activities were determined using the Alamar
blue assay.27 Detailed experimental protocols for these assays with T. b.
rhodesiense STIB900 and L6 cells have been previously reported.28
A
slight modification of this protocol was used for the assays with wild
type and resistant T. b. brucei strains, as described.29 The TbAT1-KO
strain is derived from the wild type strain T. b. brucei Lister 427 (s427)
by deletion of the TbAT1 gene.18 The B48 strain is a mutant derived
from TbAT1-KO by in vitro selection to high levels of pentamidine
and does not express a functional high affinity pentamidine transporter
(HAPT1).8 Monitoring of trypanosome cellular integrity with
propidium iodide was performed exactly as described previously.25
Pentamidine Transport Assays. Uptake assays used [3H]-
pentamidine (Amersham) at either 30 nM final concentration for
the assessment of HAPT1-mediated transport or at 1 μM for the
assessment of LAPT1-mediated transport, exactly as described.20,30
Briefly, trypanosomes were grown in HMI-9 with 10% fetal bovine
serum (FBS) until late log phase, harvested, and washed into assay
buffer (33 mM HEPES, 98 mM NaCl, 4.6 mM KCl, 0.55 mM CaCl2,
0.07 mM MgSO4, 5.8 mM NaH2PO4, 0.3 mM MgCl2, 23 mM
NaHCO3, 14 mM glucose, pH 7.3) prior to use in the assay. Cells
were incubated with radiolabel for 60 and 120 s for HAPT1 and
LAPT1 assays, respectively. Incubations were stopped by the addition
of 2 mM ice-cold unlabeled pentamidine and immediate centrifugation
through oil. Quantification was through scintillation counting, and
curves were fitted to sigmoid curves using the Prism 5.0 software
package. All experiments were performed in triplicate and on at least
three fully independent occasions.
CoMFA Models. The set of compounds (78 in total) were
converted to 3D structures using LigPrep,31 considering pH 7.0 for the
protonation state. Nevertheless, only one compound within the set,
49a, showed an additional protonation at the fixed pH for the pyridine
moiety. Both states were modeled, and it turned out that the
protonated state was more consistent with the modeled data. We
removed all counterions, which is why compounds 45a, 45f, 45g, 64,
and 65 were not included in the modeling step. Atomic charges used
for the CoMFA models were calculated at the AM132 semiempirical
level using Gaussian 03.33
The whole set was divided into two subsets, training compounds
and test compounds, in order to properly validate the models. It was
considered as important to have a good balance between active and
inactive compounds within the test set for the sake of keeping a
predictive model. With this goal in mind the whole set was divided
into active, less active, and inactive compounds, doing the random
selection upon these three subsets and trying to keep a similar number
of compounds among the three subsets. However, during the
modeling step the activities (EC50) of some compounds were still to
be determined, so these compounds were included in the test set.
The molecular alignment was built with the flexible molecular
overlay method (50% steric, 50% electrostatic) as included in
Discovery Studio, Accelerys,34 using the selected conformations of
the more active compound in the training set as reference. Erroneously
fitted compounds were realigned manually defining pairs of atoms
ensuring a maximum molecular surface overlay. The resulting
conformations on the whole set were checked to verify that no
unusual energy penalty is paid for the resulting conformations.
(4-Phenoxybenzyl)tri-m-tolylphosphonium Bromide (57c).
The reaction was carried out following the general procedure with 2c
and tri-m-tolylphosphine. The product was obtained as a white solid
1
following workup II procedure (65.2 mg, 60%): mp 228.9 °C. H
NMR (300 MHz, CDCl3) δ 7.58−7.46 (m, 11H, ArH), 7.29 (dd, J =
14.5, 6.2, 3H, ArH), 7.04 (dd, J = 8.7, 2.4, 3H, ArH), 6.93 (d, J = 7.6,
2H, ArH), 6.75 (d, J = 8.2, 2H, ArH), 5.29 (d, J = 13.8, 2H, PCH2),
2.41 (s, 9H, CH3). 13C NMR (75 MHz, CDCl3) δ 157.7, 156.6, 140.6
(d, J = 12.4), 135.9 (d, J = 2.9), 134.6 (d, J = 9.7), 133.1 (d, J = 5.4),
131.6 (d, J = 9.7), 130.4 (d, J = 13.2), 128.9, 123.9, 121.9 (d, J = 8.6),
119.3, 118.8 (d, J = 3.1), 117.9 (d, J = 84.8), 30.4 (d, J = 47.1), 21.6.
LRMS (ES+) m/z = 487.40 [M]+. Anal. Calcd for C32H50BrOP: C,
71.96; H, 5.68; Br, 14.08. Found: C, 72.18; H, 5.71; Br, 13.93.
2620
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
Four different alignments were used within this work based on the
four more stable conformations, while keeping significantly different
root mean square deviations, found for the more active compound.
Here we only present the results on one alignment, the one
corresponding to the lowest energy conformations, given that it yields
the better results. In any case no remarkable differences were found for
the four calculated molecular overlays. CoMFA descriptors calculation
(standard CoMFA, HB, and index), PLS model fitting, region
focusing, and models validations were conducted in the standard
manner described within the SYBYL suite,35 and we found that
standard CoMFA descriptors were perfectly able to model the
proposed end point. Energetic cut-offs for CoMFA descriptors were
set to 30 kcal mol−1, which was more suitable for T. b. brucei than
other calculated values.
REFERENCES
■
(1) Working To Overcome the Global Impact of Neglected Tropical
Diseases: First WHO Report on Neglected Tropical Diseases; World
Health Organization: Geneva, 2010.
(2) Delespaux, V.; de Koning, H. P. Drugs and drug resistance in
African trypanosomiasis. Drug Resist. Updates 2007, 10, 30−50.
(3) Rodgers, J. Human African trypanosomiasis, chemotherapy and
CNS disease. J. Neuroimmunol. 2009, 211, 16−22.
(4) Kinnamon, K. E.; Steck, E. A.; Rane, D. S. A new chemical series
active against African trypanosomes: benzyltriphenylphosphonium
salts. J. Med. Chem. 1979, 22, 452−455.
(5) Luque-Ortega, J. R.; Reuther, P.; Rivas, L.; Dardonville, C. New
benzophenone-derived bisphosphonium salts as leishmanicidal leads
targeting mitochondria through inhibition of respiratory complex II. J.
Med. Chem. 2010, 53, 1788−1798.
(6) Dardonville, C.; Brun, R. Bisguanidine, bis(2-aminoimidazoline),
and polyamine derivatives as potent and selective chemotherapeutic
agents against Trypanosoma brucei rhodesiense. Synthesis and in vitro
evaluation. J. Med. Chem. 2004, 47, 2296−2307.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthesis and characterization of the new phosphonium salt
derivatives 19c, 23b−e, 25c, 28b,c, 33b,c, 34b,c, 35a, 38b, 39a,
40a, 42a, 43a, 44a−c, 45a,b,d,f,g, 46b,c, 50a, 51b,c, 52b,c,
53a,c, 57a−61a, 63a and the linkers 2a, 3a, 10f, 11g, 12f, 13g,
16f, and 17g; QSAR results and molecular overlay; effect of
compounds 25b,c, 26a, 35a, 36a, 38a, 43a,c, 45d,e, 47a, and 55
on parasite viability as determined with the propidium iodide
assay. This material is available free of charge via the Internet at
http://pubs.acs.org.
(7) de Koning, H. P. Ever-increasing complexities of diamidine and
arsenical crossresistance in African trypanosomes. Trends Parasitol.
2008, 24, 345−349.
(8) Bridges, D. J.; Gould, M. K.; Nerima, B.; Maser, P.; Burchmore,
̈
R. J. S.; De Koning, H. P. Loss of the high-affinity pentamidine
transporter is responsible for high levels of cross-resistance between
arsenical and diamidine drugs in african trypanosomes. Mol.
Pharmacol. 2007, 71, 1098−1108.
(9) Luscher, A.; de Koning, H. P.; Maser, P. Chemotherapeutic
̈
̈
strategies against Trypanosoma brucei: Drug targets vs. drug targeting.
Curr. Pharm. Des. 2007, 13, 555−567.
AUTHOR INFORMATION
■
(10) de Koning, H. P.; Anderson, L. F.; Stewart, M.; Burchmore, R.
J.; Wallace, L. J.; Barrett, M. P. The trypanocide diminazene aceturate
is accumulated predominantly through the TbAT1 purine transporter:
additional insights on diamidine resistance in african trypanosomes.
Antimicrob. Agents Chemother. 2004, 48, 1515−1519.
(11) Golden, J. H. Poly-p-xylylene and related compounds. J. Chem.
Soc. 1961, 1604−1610.
Corresponding Author
*Phone: +34 912587490. Fax: +34 915644853. E-mail:
dardonville@iqm.csic.es.
Author Contributions
#These authors contributed equally to this work.
(12) Jung, H. K.; Lee, J. K.; Kang, M. S.; Kim, S. W.; Kim, J. J.; Park,
S. Y. Synthesis and properties of poly(p-phenylenevinylene-co-
sulfonylene) for a blue light-emitting diode. Polym. Bull. 1999, 43,
13−20.
ACKNOWLEDGMENTS
■
This work was supported by grants from the Spanish Ministerio
de Ciencia e Innovacion (Grants SAF2006-04698 and
́
(13) Wakselman, M.; Dome,
halogenures benzyliques amines
mono- et bialkylants. Bull. Chem. Soc. Fr. 1975, 571−575.
(14) Boyer, B.; Keramane, E. M.; Arpin, S.; Montero, J.-L.; Roque, J.-
́
M. Alkylations en milieu aqueux par les
́
́
ou amides. 1. Synthese de reactifs
́
̀
́
SAF2009-10399) and the CSIC (Bilateral Project Grant
2008GB0021) to C.D. and by the UNDP/World Bank/
WHO Special Program for Research and Training in Tropical
Diseases (M.K.). C.R.M. was a recipient of a Ph.D. fellowship
from the Government of Panama (SENACYT Grant BIDP-
2008-030). E.J.F.P. was recipient of grants from the CSIC (JAE
INTRO 08/09). V.H.G. was recipient of a JAE-TEC contract
from the CSIC. H.P.d.K. acknowledges support from the Royal
Society (International Joint Project JP0872898). The UK
Commonwealth Office (Studentship to A.A.E.) and the
Government of Saudi Arabia (Studentship to A.A.M.A, Aljouf
University, Saudi Arabia) are also acknowledged for their
support. A.C. thanks the JAE-DOC program for his contract.
́
P. BiX3 as an efficient and selective reagent for the halogen exchange
reaction. Tetrahedron 1999, 55, 1971−1976.
(15) Grutzmacher, H. F.; Neumann, E.; Ebmeyer, F.; Albrecht, K.;
̈
Schelenz, P. [3.2]Paracyclophane-10-enes and [3.2.3.2]-
paracyclophane-10,27-dienes: a convenient synthesis by the McMurry
reaction and dynamic stereochemistry. Chem. Ber. 1989, 122, 2291−
2297.
(16) Berg, M.; Kohl, L.; Van der Veken, P.; Joossens, J.; Al-Salabi, M.
I.; Castagna, V.; Giannese, F.; Cos, P.; Versees, W.; Steyaert, J.;
Grellier, P.; Haemers, A.; Degano, M.; Maes, L.; de Koning, H. P.;
Augustyns, K. Evaluation of nucleoside hydrolase inhibitors for
treatment of African trypanosomiasis. Antimicrob. Agents Chemother.
2010, 54, 1900−1908.
(17) Gould, M. K.; Vu, X. L.; Seebeck, T.; de Koning, H. P.
Propidium iodide-based methods for monitoring drug action in the
kinetoplastidae: comparison with the Alamar Blue assay. Anal. Biochem.
2008, 382, 87−93.
ABBREVIATIONS USED
■
CoMFA, comparative molecular field analysis; EC50, 50%
effective concentration; HAPT, high affinity pentamidine
transporter; HAT, human African trypanosomiasis; HEK,
human embryonic kidney; LAPT, low affinity pentamidine
transporter; NBS, N-bromosuccinimide; PI, propidium iodide;
PLS, partial least squares; rmsd, root-mean-square deviation;
SAR, structure−activity relationship; SE, standard error; SI,
selectivity index
(18) Matovu, E.; Stewart, M. L.; Geiser, F.; Brun, R.; Maser, P.;
̈
Wallace, L. J. M.; Burchmore, R. J.; Enyaru, J. C. K.; Barrett, M. P.;
Kaminsky, R.; Seebeck, T.; De Koning, H. P. Mechanisms of arsenical
and diamidine uptake and resistance in Trypanosoma brucei. Eukaryotic
Cell 2003, 2, 1003−1008.
(19) Ross, M. F.; Da Ros, T.; Blaikie, F. H.; Prime, T. A.; Porteous,
C. M.; Severina, I. I.; Skulachev, V. P.; Kjaergaard, H. G.; Smith, R. A.
2621
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622

Journal of Medicinal Chemistry
Article
J.; Murphy, M. P. Accumulation of lipophilic dications by
mitochondria and cells. Biochem. J. 2006, 400, 199−208.
(20) de Koning, H. P. Uptake of pentamidine in Trypanosoma brucei
brucei is mediated by three distinct transporters: implications for cross-
resistance with arsenicals. Mol. Pharmacol. 2001, 59, 586−592.
(21) Bush, B. L.; Nachbar, R. B. Sample-distance partial least squares:
PLS optimized for many variables, with application to CoMFA. J.
Comput.-Aided Mol. Des. 1993, 587−619.
(35) SYBYL-X, version 1.2; Tripos Inc. (1699 South Hanley Road, St.
Louis, MO, 63144, U.S.).
(22) Golbraikh, A.; Tropsha, A. Beware of q2! J. Mol. Graphics.
Modell. 2002, 20, 269−276.
́
(23) Ross, M.; Kelso, G.; Blaikie, F.; James, A.; Cocheme, H.;
Filipovska, A.; Da Ros, T.; Hurd, T.; Smith, R.; Murphy, M. Lipophilic
triphenylphosphonium cations as tools in mitochondrial bioenergetics
and free radical biology. Biochemistry (Moscow) 2005, 70, 222−230.
(24) Teka, I. A.; Kazibwe, A. J. N.; El-Sabbagh, N.; Al-Salabi, M. I.;
Ward, C. P.; Eze, A. A.; Munday, J. C.; Maser, P.; Matovu, E.; Barrett,
̈
M. P.; De Koning, H. P. The diamidine diminazene aceturate is a
substrate for the high-affinity pentamidine transporter: implications for
the development of high resistance levels in trypanosomes. Mol.
Pharmacol. 2011, 80, 110−116.
(25) Ibrahim, H. M. S.; Al-Salabi, M. I.; Sabbagh, N. E.; Quashie, N.
B.; Alkhaldi, A. A. M.; Escale, R.; Smith, T. K.; Vial, H. J.; de Koning,
H. P. Symmetrical choline-derived dications display strong anti-
kinetoplastid activity. J. Antimicrob. Chemother. 2011, 66, 111−125.
(26) Ward, C. P.; Wong, P. E.; Burchmore, R. J.; De Koning, H. P.;
Barrett, M. P. Trypanocidal furamidine analogues: influence of
pyridine nitrogens on trypanocidal activity, transport kinetics, and
resistance patterns. Antimicrob. Agents Chemother. 2011, 55, 2352−
2361.
(27) Raz, B.; Iten, M.; Grether-Buhler, Y.; Kaminsky, R.; Brun, R.
̈
̈
The Alamar Blue assay to determine drug sensitivity of African
trypanosomes (T. b. rhodesiense and T. b. gambiense) in vitro. Acta
Trop. 1997, 68, 139−147.
(28) Dardonville, C.; Fernandez-Fernandez, C.; Gibbons, S. L.;
Jagerovic, N.; Nieto, L.; Ryan, G.; Kaiser, M.; Brun, R. Antiprotozoal
activity of 1-phenethyl-4-aminopiperidine derivatives. Antimicrob.
Agents Chemother. 2009, 53, 3815−3821.
(29) Rodenko, B.; Van Der Burg, A. M.; Wanner, M. J.; Kaiser, M.;
Brun, R.; Gould, M.; De Koning, H. P.; Koomen, G. J. 2,N6-
disubstituted adenosine analogs with antitrypanosomal and antima-
larial activities. Antimicrob. Agents Chemother. 2007, 51, 3796−3802.
(30) Wallace, L. J. M.; Candlish, D.; De Koning, H. P. Different
substrate recognition motifs of human and trypanosome nucleobase
transporters. Selective uptake of purine antimetabolites. J. Biol. Chem.
2002, 277, 26149−26156.
(31) LigPrep, version 2.5; Schrodinger, LLC: New York, NY, 2011.
̈
(32) Mittal, R. R.; Harris, L.; McKinnon, R. A.; Sorich, M. Partial
charge calculation method affects CoMFA QSAR prediction accuracy.
J. Chem. Inf. Model. 2009, 49, 704−709.
(33) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.;
Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.;
Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.;
Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.;
Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao,
O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J.
B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R.
E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;
Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J.
J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.;
Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman,
J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.;
Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen,
W.; Wong, M. W.; Gonzalez, C.; , Pople, J. A. Gaussian 03, revision C;
Gaussian, Inc.: Wallingford, CT, 2004.
(34) Discovery Studio; Accelrys Software Inc.: San Diego, CA, 2010.
2622
dx.doi.org/10.1021/jm2014259 | J. Med. Chem. 2012, 55, 2606−2622