Design, synthesis and antimalarial activity of novel, quinoline-based, zinc metallo-aminopeptidase inhibitors

Article abstract of DOI:10.1016/S0960-894X(03)00550-X

PfA-M1, a neutral zinc aminopeptidase of Plasmodium falciparum, is a new potential target for the discovery of antimalarials. The design and synthesis of a library of 45 quinoline-based inhibitors of PfA-M1 is reported. The best inhibitor displays an IC₅₀ of 854 nM. The antimalarial activity on a CQ-resistant strain and the specificity towards mammalian aminopeptidase N are also discussed.

Full text of DOI:10.1016/S0960-894X(03)00550-X

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2659–2662
Design, Synthesis and Antimalarial Activity of Novel,
Quinoline-Based, Zinc Metallo-Aminopeptidase Inhibitors
Marian Flipo,a Isabelle Florent,b Philippe Grellier,b Christian Sergheraerta
and Rebecca Deprez-Poulaina,*
a
´
UMR CNRS 8525, Institut Pasteur et Institut de Biologie de Lille, Universite de Lille 2, Lille, France
b
´
´
´
USM 0504‘Biologie fonctionnelle des protozoaires’, De partement ‘Regulations, Developpement,
´
Diversite Moleculaire’, Museum National d’Histoire Naturelle, FR CNRS 63 Paris, France
´
´
Received 11 March 2003; revised 8 May 2003; accepted 30 May 2003
Abstract—PfA-M1, a neutral zinc aminopeptidase of Plasmodium falciparum, is a new potential target for the discovery of anti-
malarials. The design and synthesis of a library of 45 quinoline-based inhibitors of PfA-M1 is reported. The best inhibitor displays
an IC50 of 854 nM. The antimalarial activity on a CQ-resistant strain and the specificity towards mammalian aminopeptidase N are
also discussed.
# 2003 Elsevier Ltd. All rights reserved.
Introduction
(1) to hydrolyze the small peptides generated in the food
vacuole, during hemoglobin digestion by acidic endo-
peptidases (aspartyl, cysteine and metallo-endopepti-
dases), into amino-acids at the level of the cytoplasm, and:
(2) to play a role in the erythrocyte re-invasion by the
parasite.8 PfA-M1 shares a maximal homology in the
active site region (44%) with two human proteases: the
human aminopeptidase-N and leukotriene A4 hydrolase.9
Malaria, caused by the parasite Plasmodium falciparum,
is still prevalent and lethal in many countries.1 Chloro-
quine (CQ), exerting its antimalarial activity by inhibit-
ing haemozoin formation in the food vacuole of the
parasite, has been the standard drug for many decades.
Unfortunately, the spread of resistant P. falciparum to
this molecule has created an urgent need to develop new
antimalarial treatments.2 In a project to find new quin-
oline-based antimalarials that would not induce resis-
tance, unlike CQ, we have designed several compounds
displaying a piperazine linker. Several of these com-
pounds were active on the CQ-resistant strain FcB1.3
It is the only aminopeptidase of P. falciparum that has
been purified and biochemically characterized. This
enables the design and testing of potential inhibitors.
Such compounds would help understanding the role of
this enzyme and validate it as a therapeutic target.
The recent publication of the genome of the parasite
opens the opportunities of a better understanding of the
biology of the parasite.4 This will hopefully result in the
discovery of new specific targets to tackle with drugs.
Proteases expressed in the erythrocytic stage of P. falci-
parum, for example, can be considered as interesting
targets for the design of antimalarials.5,6
In a project aiming at decreasing the potential for resis-
tance occurrence of our quinoline-based antimalarials,
we have designed dual inhibitors that would inhibit
both the haemozoin formation and one of the proteases
potentially involved in globin digestion. Such a strategy
has been developed by Avery for the design of malarial
cysteine proteases inhibitors based on mefloquine and
chloroquine.10 The screening of several of our in-house
quinoline-based compounds allowed us to identify
compound 1 as an interesting inhibitor of PfA-M1, that
served as a starting point for analoguing (Fig. 1).
In particular, PfA-M1,7 a neutral zinc-aminopeptidase
(M1 family), inhibited by the classical aminopeptidase
inhibitor bestatin, has been proposed:
We report here the design and parallel synthesis of a
library of 45 non-peptidic analogues of 1, and its biological
*Corresponding author. Tel.:+33-3-2087-1217; fax:+33-3-2087-1233;
e-mail: rebecca.poulain@ibl.fr.
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00550-X

2660
M. Flipo et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2659–2662
Figure 1.
Scheme 2. (a) 1 equiv carboxylic acid, 0.7 equiv BF3–Et2O in t-Bu-
COOCH3, 4 h, rt; (b) 1.1 equiv amine 2, 1.1 equiv DIEA, EtOH, 15 h,
evaluation on PfA-M1. Preliminary results on the anti-
malarial activity on FcB1 strain of the most active enzyme
inhibitors, and preliminary specificity data on mammalian
aminopeptidase N, are presented and discussed.
.
reflux; (c) 1.2equiv H N–OtBu HCl, 2.4 equiv pyridine, THF, 1 h, rt;
2
(d) 1.1 HCOCH(CH3)2, molecular sieves, 0.4 equiv DIEA, 6 equiv
NaBH4, MeOH, overnight, rt.
reductive amination of an aldehyde or N-alkylation
with the appropriate halide (Scheme 2).
Chemistry
As for the carboxylic acid precursors, we synthesized
malonic derivatives (24–28) according to the procedure
described in Scheme 3. These moieties have been pre-
viously described in inhibitors of matrix-metallopro-
teases (MMPs).13 We also added to these in-house
precursors, some commercially available N-protected
amino-acids Lys, Leu, Phe, Asp (29–32, Fig. 3).
PfA-M1 is a zinc-metalloprotease, of the M1 family,
preferably cleaving basic, hydrophobic, as well as aro-
matic amino-acids.8 Zn-chelating groups are critical for
the inhibition of such enzymes.11 To fulfill this function,
we incorporated at least a carboxylic acid or an
hydroxamate group in our molecules. In order to mimic
the side chain of a Leucine residue, an isobutyl group
(Ibu), analogue of the methyl-cyclopropyl moiety in 1,
was also added in our potential inhibitors.
The 45 analogues (5 aminesꢀ9 carboxylic acids) have
been synthesized at a 5-mmol scale, in parallel, accord-
ing to the two-step procedure described in Scheme 1.
The first step consists in the formation of the amide
bond using PyBrop as the activator of the carboxylic
acid followed by deprotection of the Boc and tBu
groups of amines, hydroxamates and carboxylic acids.
The synthesis of this full combinatorial library required
specific protected amines and carboxylic acids pre-
cursors to be synthesized prior to the coupling step.
Amines 3–6 (Fig. 2) all derived from compound 2 pre-
viously obtained from 4,7-dichloro-quinoline and 1,4-
bis-(3-aminopropyl)-piperazine by nucleophilic aro-
matic substitution.12 They were synthesized either by
Scheme 3. (a) 1 equiv KOH, EtOH, 2h, rt; (b) 1.2equiv t-Bu-O-
.
NH2 HCl, 1.2equiv DIEA, 1.1 equiv EDCI and 1.1 equiv HOBt,
THF, overnight, rt or 5 equiv t-Bu-OH, 4 equiv DIEA, 2equiv 2-
chloro-1-methylpyridium iodide, acetonitrile, 2h, rt; (c) 3 equiv KOH,
EtOH, overnight, rt.
Scheme 1. (a) 1.1 equiv PyBrop, 2equiv DIEA in DCM, overnight, rt;
(b) TFA/DCM (20/80), 72 h, rt.
Figure 2. Amine precursors selected and synthesized for the library.
Figure 3. Carboxylic acids selected for the library.

M. Flipo et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2659–2662
2661
Analytical control
Each crude product was tested for purity using HPLC
and identity using MALDI-TOF. In all cases, the purity
exceeded 80% and the mass spectrum was consistent
with the anticipated product structure.
Figure 5. Inactive analogues 36, 37 from the screening and reference
compound 38.
Biological Assays
The analogues and the deprotected precursors were
screened for their ability to inhibit PfA-M1 at 10 mM.14
Crude products displaying an inhibition percentage
above 50% were selected for re-synthesis and IC50
determination on fully purified and controlled samples.
The selected compounds were then tested for their ability
to inhibit FcB1 parasite growth.15 The specificity against
other aminopeptidases was evaluated in a model of
mammalian aminopeptidase-N from porcine kidney.16
the analogues of 33–35 without this moiety (derivatives
of 2) were inactive.
The relative positions of the Zn-chelating group and the
isobutyl group are critical for the activity. For example,
compounds 36–37 (Fig. 5) are inactive while displaying
the same groups in a different 3D arrangement as
respectively in 34 and 33. None of the analogues derived
from the amino-acids (37 for instance) supposed to be
recognized by the enzyme, was active, suggesting that
the Zn-chelating group should be supported by the car-
boxylic acid precursor.
Structure–activity relationships on PfA-M1
Out of the 45 compounds tested at 10 mM, three com-
pounds (33–35, Fig. 4),17 displayed a percentage of
inhibition above threshold. Interestingly, they all
derived from the same amine 6. After resynthesis and
testing, they displayed IC50s from 0.85 to 1.54 mM on
PfA-M1, that is, a 10-25 times increase in potency
comparing to compound 1 (Table 1).
The central position in malonic derivatives seems toler-
ant since it can accommodate with basic (NH2) or
hydrophobic (isobutyl) groups. Nevertheless, the best
analogue, compound 34, is not substituted in that posi-
tion. In particular, this suggests that an amino-group is
not essential for the recognition by the enzyme although
it is an aminopeptidase.
The presence of a Zn-chelating group is essential for the
activity. Compound 33, the closest analogue of 1, bear-
ing an hydroxamate is 10 times more active on the
enzyme. In contrast, amine 6 is inactive. The hydro-
xamate group is more efficient than the carboxylic acid
group. Carboxylic acids analogues of 33 and 34 were
inactive. The isobutyl group is essential for activity, since
Antimalarial activity and specificity
The three PfA-M1 inhibitors are sub-micromolar inhi-
bitors of the parasite growth. Their antimalarial activity
is due to the inhibition of both PfA-M1 and haem
detoxification.18 Interestingly, reference compound 38,
displaying a COOH moiety at the same position as the
hydroxamate OH in compound 34 (and a cyclopropyl-
methyl group instead of an isobutyl group), is weakly
active on PfA-M1, and less active than 34 on parasite
growth inhibition. Specificity for compounds 33–35, as
expressed with the ratio of the IC50s on the two targets,
varies from 0.9 to 30.5. The size of the substituant on
the malonic methylene appears to be critical.
Figure 4. Best hits from the library on PfA-M1.
Table 1. Inhibition of PfA-M1 and of parasite growth, and specificity
towards mammalian aminopeptidase-N of compounds 33–35
Conclusion
We have successfully designed non-peptidic inhibitors
of PfA-M1, an aminopeptidase of P. falciparum. The
quinoline moiety allows these compounds also to be
inhibitors of haem detoxification, making these com-
pounds potentially dual inhibitors of hemoglobin diges-
tion. Yet the relative part of each mode of action needs
to be determined. The three selected inhibitors 33–35
are active against CQ-resistant strain, thus validating
their ability to cross membranes. The activity on PfA-
M1, and the specificity against other aminopeptidases
such as mammalian aminopeptidase-N need to be
improved. Such issues are under investigation and
results will be reported in due course. Antimalarial
activities and rough SARs emerging from this study,
Compd PfA-M1 Parasite growth Aminopeptidase Ratio IC50
inhibition
IC50, mM
inhibition
IC50, mMa
N (APN)
inhibition
IC50, mM
PfA-M1/IC50
APN
CQ
Bestatin
1
33
34
35
38
NDb
0.284 29
21
2.503
0.854
1.540
>10c
0.126 (0.026)
NDb
2.700
28
1.345
0.028
1.628
NDb
—
0.1
0.7
1.9
30.5
0.9
—
0.398 (0.056)
0.178 (0.037)
0.317 (0.063)
0.151 (0.013)
1.380 (0.050)
aValues are means of three experiments, standard deviation is given in
parentheses. Strain is FcB1.
bND, not determined.
c22% inhib (10 mM).

2662
M. Flipo et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2659–2662
have encouraged us to pursue the optimization of PfA-
M1 inhibitors. Especially, we are now focusing on
inhibitors devoided of the quinoline moiety.
trols that represent Vmax. For the IC50 values determination,
initial velocities were plotted as a function of inhibitor con-
centration.
15. P. falciparum strains were maintained continuously in
culture on human erythrocytes as described by Trager and
Jensen (Science 1976, 193, 673). In vitro antiplasmodial activ-
ity was determined using a modification of the semi-auto-
mated micro-dilution technique of Desjardins et al.
(Antimicrob. Agents Chemother. 1979, 16, 710). P. falciparum
CQ-resistant FcB1 (Colombia) strain was used. Stock solu-
tions of CQ-diphosphate and test compounds were prepared
in sterile distilled water and DMSO, respectively. Drug solu-
tions were serially diluted with culture medium and introduced
to asynchronous parasite cultures (0.5% parasitemia and 1%
final hematocrite) on 96-well plates for 24 h at 37 ꢂC prior to
the addition of 0.5 mCi of [3H]hypoxanthine (1–5 Ci/mmol;
Amersham, France) per well, for 24 h. The growth inhibition
of each drug concentration was determined by comparison of
the radioactivity incorporated into the treated culture with
that in the control culture (without drug) maintained on the
same plate. The IC50, as obtained from the drug concen-
tration–response curve were expressed as meanꢃstandard
deviations, determined from three independent experiments.
The DMSO concentration never exceeded 0.1% and did not
inhibit the parasite growth.
16. Microsomal aminopeptidase N from porcine kidney (EC
3.4.11.2) was purchased from Sigma-Aldrich France, as a sus-
pension in 3.5 M (NH4)2SO4 solution. Before use, it was dilu-
ted 600 times in Tris–HCl buffer (25 mM; pH 7.4. The assays
were set up in 96-well plates and proceeded like for PfA-M1
with the same substrate Leu-pNA (Km=0.13 mM). The Z0
factor of the test was 0.80 allowing activities to be determined
with a single point with a 95% confidence. The reference
inhibitor was bestatin.
References and Notes
1. WHO. World Health Report; WHO: Geneva, 2002.
2. Wellems, T. E. Science 2002, 298, 124.
3. Ryckebusch, A.; Deprez-Poulain, R.; Maes, L.; Debreu-
Fontaine, M.-A.; Mouray, E.; Grellier, P.; Sergheraert, C.
J. Med. Chem. 2003, 46, 542.
4. Gardner, M. J.; Hall, N.; Fung, E.; White, O.; Berriman,
M.; Hyman, R. W. et al. Nature 2002, 419, 498.
5. Werbovetz, K. A. Curr Med. Chem. 2000, 7, 835.
6. For reviews see: (a) Rosenthal, P. J. Curr. Opin. Hematol.
2002, 9, 140. (b) Blackman, M. J. Curr. Drug Targets 2000, 1, 59.
7. Swissprot accession number of PfA-M1: O96935.
8. Allary, M.; Schrevel, J.; Florent, I. Parasitology 2002, 125, 1.
9. Florent, I.; Derhy, Z.; Allary, M.; Monsigny, M.; Mayer,
R.; Schrevel, J. Mol. Biochem. Parasitol. 1998, 117, 37.
10. Lima, P. d. C.; Barreiro, E. J.; Avery, M. A. Abstracts of
Papers, 223rd ACS National Meeting, Orlando, FL, USA,
April 7–11, 2002; American Chemical Society: Washington,
DC, 2002; MEDI-176.
11. (a) Chan, W. W.; Dennis, P.; Demmer, W.; Brand, K.
J. Biol. Chem. 1982, 257, 7955. (b) Hernandez, J. F.; Soleilhac,
J. M.; Roques, B. P.; Fournie-Zaluski, M. C. J. Med. Chem.
1988, 31, 1825.
12. Ryckebusch, A.; Deprez-Poulain, R.; Debreu-Fontaine,
M.-A.; Vandaele, R.; Mouray, E.; Grellier, P.; Sergheraert, C.
Bioorg. Med. Chem. Lett. 2002, 12, 2595.
17. 33. (3TFA): NMR (DMSO-d6) 1H d ppm 0.80–0.84 (m;
14H); 1.10–1.26 (m; 5H); 1.27–1.90 (m; 6H); 2.75–3.58 (m;
14H); 6.89 (d; J=7.1 Hz; 1H); 7.76 (dd; J=9.0 Hz; J=1.1 Hz;
1H); 7.95 (s; 1H); 8.50–8.56 (m; 2H); 9.30 (br s; 1H; 2NH+);
9.73 (br s; 0.5H; NHO); MALDI-TOF: 534.2(M+H +). 34.
13. Graf von Roedern, E.; Grams, F.; Brandstetter, H.; Mor-
oder, L. J. Med. Chem. 1998, 41, 339.
14. Native PfA-M1 was purified according to the procedure
described in ref 9, and diluted 10 times in Tris–HCl buffer (25
mM; pH 7.4) before use. The assays were set up in 96-well
plates. The compounds were tested at a final concentration of
10 mM. 33 mL of purified PfA-M1 were pre-incubated 10 min
at rt with 33 mL of the inhibitor (30 mM in Tris–HCl buffer,
0.3% DMSO). 33 mL of the substrate Leu-pNA (Km=0.1
mM) 0.3 mM in Tris–HCl buffer were then added. The reac-
tion kinetics was followed on a UV-microplate reader Multi-
skanRC (Labsystems, Finland) at 405 nm. The control
activity was determined by incubating the enzyme in the same
conditions without inhibitor. Bestatin was used as the refer-
ence inhibitor. The statistical Z0 factor for the test (Zhang, J.
H.; Chung, T. D.; Oldenburg, K. R. J. Biomol. Screen. 1999,
4, 67) was 0.82, allowing activities to be determined with a
single point with a 95% confidence. Initial velocities are
expressed in mmol. minꢁ1. Data were normalized to the con-
1
(2TFA): NMR (CD3OD) H d ppm 0.80 (d; J=6.6 Hz; 2H);
0.86 (d; J=6.6 Hz; 4H); 1.86–1.97 (m; 5H); 2.72 (br t; J=6.5
Hz; 2H); 2.76–2.95 (m; 4H); 2.96–3.21 (m; 10H); 3.43 (br t;
J=6.6 Hz; 2H); 3.62 (br t; J=6.7 Hz ; 2H) ; 6.83 (d; J=6.8
Hz; 1H); 7.60 (dd; J=1.7 Hz; J=9.1 Hz; 1H); 7.79 (br s; 1H);
8.25–8.33 (m; 2H); MALDI-TOF: 519.2 (M+H+). 35.
(2TFA): NMR (DMSO-d6) 1H d ppm 0.80–0.84 (m; 14H);
1.10–1.26 (m; 5H); 1.27–1.90 (m; 6H); 2.75–3.58 (m; 14H);
6.90 (d; J=7.1 Hz; 1H); 7.78 (dd; J=9.0 Hz; J=1.9 Hz; 1H);
7.95 (d; J=1.8 Hz; 1H); 8.51 (d; J=9.0 Hz; 1H); 8.56 (d;
J=7.0 Hz; 1H); 8.91 (br s; 0.6H; NHO); 9.34 (br s; 1.3H;
2NH+); 10.75 (br s; 0.6H; OH); MALDI-TOF: 575.3
(M+H+).
18. In-vitro inhibition of haem polymerization induced by
lipids, unpublished data (see ref 3 for protocol).