Synthesis of new fluoroquinolones and evaluation of their in vitro activity on Toxoplasma gondii and Plasmodium spp.

Article abstract of DOI:10.1016/j.bmcl.2004.03.070

The synthesis of four new computer-designed fluoroquinolones which have been predicted by QSAR analysis to be active against the protozoa Toxoplasma gondii is described. These compounds are inhibitory in vitro for T. gondii. One of these compounds has a remarkably high activity comparable to that of trovafloxacin. It combines the basic cyclopropyl-quinoline structure of gatifloxacin or moxifloxacin with the C-7 6-amino-3-azabicyclo[3.1.0]hexyl side chain of trovafloxacin. The four compounds are also inhibitory for blood stages of Plasmodium falciparum though at high concentration. These results confirm the potential of quinolones as anti-T. gondii and antimalarial drugs but also show that the QSAR models for T. gondii cannot be reliably extended for screening antimalarial activity.

Full text of DOI:10.1016/j.bmcl.2004.03.070

Bioorganic & Medicinal Chemistry Letters 14 (2004) 2773–2776
Synthesis of new fluoroquinolones and evaluation of their in vitro
activity on Toxoplasma gondii and Plasmodium spp.
G. Anquetin,a M. Rouquayrol,a N. Mahmoudi,b,c M. Santillana-Hayat,b R. Gozalbes,b,ꢀ
J. Greiner,a K. Farhati,c F. Derouin,b R. Guedja and P. Vierlinga,*
aLaboratoire de Chimie Bioorganique UMR-CNRS 6001, Universitꢀe de Nice-Sophia Antipolis, Parc Valrose,
06108 Nice Cꢀedex 2, France
bLaboratoire de Parasitologie et Mycologie, Universitꢀe Paris 7, Assistance Publique-H^opitaux de Paris, 15 rue de l’Ecole de Mꢀedecine,
75006 Paris, France
cINSERM 511, Immuno-biologie Cellulaire et Molꢀeculaire des Infections Parasitaires, CHU Pitiꢀe-Salpꢀetriꢁere,
Assistance Publique-H^opitaux de Paris, 75013 Paris, France
Received 28 January 2004; revised 19 March 2004; accepted 24 March 2004
Abstract—The synthesis of four new computer-designed fluoroquinolones which have been predicted by QSAR analysis to be active
against the protozoa Toxoplasma gondii is described. These compounds are inhibitory in vitro for T. gondii. One of these compounds
has a remarkably high activity comparable to that of trovafloxacin. It combines the basic cyclopropyl–quinoline structure of
gatifloxacin or moxifloxacin with the C-7 6-amino-3-azabicyclo[3.1.0]hexyl side chain of trovafloxacin. The four compounds are also
inhibitory for blood stages of Plasmodium falciparum though at high concentration. These results confirm the potential of quino-
lones as anti-T. gondii and antimalarial drugs but also show that the QSAR models for T. gondii cannot be reliably extended for
screening antimalarial activity.
Ó 2004 Elsevier Ltd. All rights reserved.
Fluoroquinolones have attracted much attention be-
cause of their broad spectrum of activity against various
bacteria, mycobacteria, and parasites.1–3 Indeed, the
presence of a prokaryotic-like organelle in apicom-
plexan parasites, which comprise major human patho-
gens such as Plasmodium spp., Toxoplasma gondii, and
other coccidia, may represent a unique target for anti-
biotics against these protozoa.4–6 In this context and in
order to discover new and potent antiparasitic fluoro-
quinolones derivatives, a QSAR analysis by molecular
connectivity of a series of quinolones active against T.
gondii was performed.7 This analysis led us to identify
their main active pharmacophores and to design new
virtual structures that should display higher or at least
comparable biological activities to those of already
known fluoroquinolone drugs. Among the computer-
designed fluoroquinolones, we selected the four fluoro-
quinolones 1–4 (Scheme 1) for which a high anti-Toxo-
plasma activity was predicted (see Table 1) and synthesis
seemed feasible.
O
O
O
N
O
F
F
OH
OH
N
N
R1
H
N
R
OMe
H2N
H
F
N
Trovafloxacin (TVFX): R1
(TVFX)
=
=
Gatifloxacin: R =
(GTFX)
HN
CH3
F
H
CP-TVFX: R1
N
Moxifloxacin:R =
(MXFX)
N
H
H
O
O
5
F
6
10
9
1: R1 = cyclopropyl; R2 = NH2
3
OH
2'
2: R1 = cyclopropyl; R2 = CH2NH2
3: R1 = 2,4-difluorophenyl; R2 = NH2
4: R1 = 2,4-difluorophenyl; R2 = CH2NH2
7
2
1'
H
N
N
8
Keywords: Fluoroquinolone; Protozoa; QSAR; Malaria; Parasites.
* Corresponding author. Tel.: +33-0-4-92-07-61-43; fax: +33-0-4-92-
07-61-51; e-mail: vierling@unice.fr
OMe R1
R2
4'
5'
H
6'
ꢀ
Present address: CEREP, Molecular Modeling Department, 128 rue
Danton, 92500 Rueil Malmaison, France.
Scheme 1. Chemical structure of the target fluoroquinolones 1–4 and
of known close analogs.
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.03.070

2774
G. Anquetin et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2773–2776
Table 1. Activities of fluoroquinolones against Toxoplasma gondii (RH) on MRC5 fibroblast tissue cultures,7 against hepatic stages of Plasmodium
yoleii yoleii (265 BY strain), and against blood stages of chloroquine-resistant strain of Plasmodium falciparum (NF54-R)17
Compound
T. gondii IC50 (lM) [predicted]
P. yoleii yoleii IC50 (lM) ( SD)
P. falciparumf IC50 (lM) ( SD)
1
2.1 [1.15]a
(1.6–3.1)b
66 [1.21]a
(54–87)b
32 [0.58]a
(26–42)b
37 [0.63]a
(30–48)b
0.96 [1.2]c
2.93d
73.1 (8.2)
92 (13)
2
Not active
Not active
Not active
74.0 (47.8)e
100 (1)
105 (20)
40 (2)
3
4
TVFX
22 (12)e
CP-TVFX
GTFX
MXFX
1.68 [2.3]c
12.7 [33.8]c
10.9 [6.2]c
––
Not activee
Not activee
––
29 (1)e
45 (18)e
a The predicted value was calculated using the protocol described in Ref. 7.
b 95% Confidence interval.
c Data from Ref. 7.
d Data from Ref. 8.
e Data from Ref. 17.
f For comparison, the IC50 of chloroquine is of 0.57 0.4 lM.17
It should be emphasized that the fluoroquinolones 1–4
contain the basic C8-methoxy quinoline structure of
gatifloxacin (GTFX) and moxifloxacin (MXFX), and
are also close analogs of trovafloxacin (TVFX) and
cyclopropyl-TVFX8 (Scheme 1), these four latter drugs
being among the most anti-Toxoplasma active drugs
known so far (see Table 1).7–9
O
F
F
COOH
F
F
F
C(O)Cl
F
F
F
CO2Et
NEt2
i)
F
path A
OH
OMe
OMe
14
12
13
ii),
iii)
path B
O
F
O
O
N
We describe here the synthesis of these fluoroquinolones
and also report on their in vitro activity against T. gondii
and against blood and hepatic stages of Plasmodium
spp. since these parasites might share common drug
targets to quinolones.5 The known amino-azabicyclo-
hexane 11 and its methylamino analog 8 (Scheme 2), as
well as the difluoroquinolone 16a10 and its new analog
16b (Scheme 3) constitute the key synthons to the target
F
F
F
CO2Et
OEt
F
OMe
OMe R1
16a: R1 = cyclopropyl
16b: R1 = 2,4-difluorophenyl
15
Scheme 3. Synthetic pathway to the difluoroquinolone 16a,b synthons.
Path A: (i) Et2NCH@CHCO2Et, NEt3, toluene, 90 °C, 5 h; (ii) R1NH2,
1:2 EtOH/Et2O, rt, 3 h; (iii) K2CO3, DMF, 100 °C, 5 h. Path B: (a)
[O2CCHCO2Et]2ꢀ 2Liþ, THF; (b) (EtO)3CH, Ac2O; (c) R1NH2,
EtOH, (d) NaH, THF.10
Bn
N
Bn
N
Bn
N
Bn
N
H
N
iv)
v)
O
O
O
O
i)
iii)
vii)
ii)
vi)
viii)
and original fluoroquinolones 1–4 whose syntheses are
shown in Scheme 4. The synthesis of the two heterobi-
cyclic derivatives 8 and 11 (Scheme 2) was performed
from commercial N-benzylmaleimide in eight steps
(about 10% overall yield) according to published pro-
cedures.11
H
H
H
H
CO2Et
H
H
H
H
CH2NHBoc
CH2OH
CH2NH2
5
6
8
7
ix), x),
xi)
Z
Z
N
H
N
N
xii)
xiii)
The synthetic approach we used for the two dif-
luoroquinolones 16a,b (path A) was adapted from the
method developed by Cecchetti et al., for the synthesis
of similar fluoroquinolones.12–14 Its key steps lie in the
transformation of 14 into 16a,b. This method is easier to
implement (milder conditions and three steps instead of
four) and more efficient than the method briefly depicted
in path B that has originally been applied for the N-
cyclopropyl 16a derivative.10;15 Standard literature pro-
cedures applied to 3-hydroxy-2,4,5-trifluorobenzoic acid
1210 led to the formation of the acid chloride 13. Its
reaction with ethyl 3-(diethylamino)acrylate in the
H
NHBoc
10
H
NHBoc
11
H
H
H
H
COOH
9
Scheme 2. Synthetic pathway to the N-Boc-protected amino-azabi-
cyclohexanes 8 and 11: (i) N2CHCO2Et, Et2O, rt, 23 h; (ii) 185 °C, 2 h;
(iii) LiAlH4, THF, reflux, 48 h; (iv) (ClCO)2, DMSO, NEt3, CH2Cl2,
)75 °C, 19 h, rt; (v) CH3COONa, HONH2 Æ HCl, EtOH, rt, 16 h; (vi)
LiAlH4, THF, reflux, 15 h; (vii) (Boc)2O, NEt3, 9:1 dioxane/H2O, rt,
4 h; (viii) HCOONH4 or H2, Pd/C, EtOH, rt, 1 h; (ix) H2/Pd/C, MeOH,
0 °C, 23 h; (x) K2CO3, ClC(O)OCH2Ph (ZCl), 1:1 dioxane/H2O, rt,
15 h; (xi) Jones reagent, acetone, 0 °C then rt, 2 h; (xii) t-BuOH, NEt3,
(PhO)2P(O)N3, reflux, 44 h; (xiii) H2/Pd/C, MeOH, 0 °C, 5 h.

G. Anquetin et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2773–2776
2775
Our results show that the QSAR models developed for
T. gondii cannot be reliably extended for screening
antimalarial activity. Indeed, the T. gondii inhibitory
H
8: R3 = CH2NHBoc
11: R3 = NHBoc
16a,b
N
i)
H
F
F
H
O
O
activity
decreased
along
the
series
TVFX
R3
8 or 11
B
F
O
O
ꢁ1>MXFXꢁGTFX>3ꢁ4>2 while the P. falciparum
inhibitory activity decreased along the series
TVFXꢁGTFX>4ꢁMXFX>1ꢁ2ꢁ3.
OH
F
F
O
H
N
N
ii), iii)
OMe R1
N
R2
OMe R1
H
Concerning SAR for anti-T-gondii and anti-Plasmodium
activity, our data show that combining the basic cyclo-
propyl–quinoline structure of GTFX or MXFX with the
C-7 6-amino-3-azabicyclo[3.1.0]hexyl side chain of
TVFX, as in 1, increased substantially the activity
against T-gondii (as well as the anti-P. yoelii yoelii
activity of GTFX or MXFX which are not active) but
decreased the anti-P. falciparum activity of GTFX and
MXFX. By contrast, a significant anti-T. (as well as an
anti-P. falciparum) activity decrease was observed when
this basic cyclopropyl–quinoline structure was com-
bined with the 6-methylamino-3-azabicyclo[3.1.0]hexyl
side chain as in 2. It thus appears that the contribution
of the C7 substituent to anti-T-gondii activity decreases
along the series: 6-amino-3-azabicyclohexane (as in
1)>30-methylpiperazine (as in GTFX) ꢁ piperidino-
pyrrolidine (as in MXFX)>6-methylamino-3-azabicy-
clohexane (as in 2). For anti-P. falciparum activity, the
tendency of the C7 substituent contribution is as follow:
30-methylpiperazine (as in GTFX)ꢁpiperidinopyrroli-
dine (as in MXFX)>6-methylamino-3-azabicyclohexane
(as in 2)>6-amino-3-azabicyclohexane (as in 1).
17a,b
1-4
Scheme 4. Synthetic pathway to the target fluoroquinolones 1–4: (i)
BF3 Æ Et2O, THF, reflux; (ii) CH3CN, 7 days, reflux; (iii) 1:1 CH2Cl2/
TFA (for R1 and R2 see Scheme 1).
presence of NEt3 gave 14. Transaminolyse of 14 with
cyclopropylamine or 2,4-difluoroaniline and successive
cyclization with potassium carbonate in DMF gave the
target esters 16a,b, respectively.
The fluoroquinolones 1–4 were then obtained by cou-
pling the difluoroquinolones 16a,b with the amines 8 or
11, as outlined in Scheme 4. This conjugation needed the
activation of the C-7 position of the ethyl esters 16a,b.12
This was performed by converting them into their
respective boron difluoride complexes 17a,b using boron
trifluoride etherate in refluxing THF. The successive but
nevertheless time-consuming (7 days) nucleophilic dis-
placement of the C-7 fluorine atom in 17a (resp. 17b)
with 2.5–3 equiv of the selected heterobicyclic amine 11
or 8 followed by N-Boc deprotection using an excess of
1:1 TFA/CH2Cl2 gave the desired fluoroquinolones 1 or
2 (resp. 3 or 4), as their TFA salts16 with 27% or 40%
(resp. 40 or 47%) overall yields from 16a,b.
In line with literature,18 our data confirm further that
replacing the cyclopropyl in 1 by the 2,4-difluorophenyl
as in 3 induces a marked anti-T-gondii and anti-P. yoelii
yoelii activity decrease. This replacement has almost no
incidence on anti-P. falciparum activity. When both the
6-methylamino-3-azabicyclo[3.1.0]hexyl side chain and
the 2,4-difluorophenyl were combined with the basic
quinoline structure GTFX or MXFX as in 4, a more
drastic anti-T. gondii activity decrease was observed
while the anti-P. falciparum activity was almost not af-
fected. Our data show further that replacing the basic
naphthyridine structure of CP-TVFX by the basic
quinoline one of 1, (which is equivalent to replacing the
intracyclic N atom by a C-methoxy group) had no effect
on the anti-T. gondii activities.
The four fluoroquinolones 1–4 were tested in vitro on
MRC5 fibroblast tissue cultures inoculated with the
virulent RH strain of T. gondii,7 and against blood
stages of the chloroquine-resistant NF54-R strain of P.
falciparum (NF54-R) and liver stages of P. yoelii yoelii
(265 BY strain) according to published procedures.17
The results are collected in Table 1. These derivatives,
predicted to be active by our molecular connectivity
QSAR analysis against T. gondii were indeed found to
be active, though only one of the four derivatives, that
is, 1, displayed a comparable inhibitory activity to that
predicted. Moreover, 1 exhibited an anti-T-gondii IC50
value close to and higher than that of TVFX and of
GTFX and MXFX, respectively, which are among the
three most anti-T-gondii active fluoroquinolone drugs
known. This derivative, like TVFX, was also active at a
high concentration against the hepatic stages of P. yoelii
yoelii (Table 1). Its inhibitory effect was also associated
with alteration of schizont morphology and significant
reduction of schizont sizes. While only 1 was active in
this Plasmodium model, the four fluoroquinolones 1–4
were active against blood stages of the chloroquine-
resistant strain of P. falciparum, though at much higher
concentrations than that of chloroquine (Table 1).
However, it is 4 of the four new fluoroquinolones which
displays the highest antimalarial activity. This activity is
close to that of TVFX, GTFX, and MXFX (Table 1).
These data confirm the potential of quinolones as anti-
malarial drugs.
Finally, our results show the usefulness and the limits of
our predictive QSAR models for T. gondii. Indeed, the
four computer-designed fluoroquinolones were active
but only one out of these derivatives was as active as
predicted, showing that the models remain of interest to
orient synthesis of active anti-T. gondii quinolones.
Concerning Plasmodium spp., further studies are in
progress to define more adapted and specific QSAR
models for the design of new antimalarial drugs.
References and notes
1. Owens, R. C., Jr.; Ambrose, P. G. Med. Clin. North Am.
2000, 84, 1447–1469.

2776
G. Anquetin et al. / Bioorg. Med. Chem. Lett. 14 (2004) 2773–2776
2. Aminimanizani, A.; Beringer, P.; Jelliffe, R. Clin. Phar-
macokinet. 2001, 40, 169–187.
10. Sanchez, J. P.; Gogliotti, R. D.; Domagala, J. M.;
Gracheck, S. J.; Huband, M. D.; Sesnie, J. A.; Cohen,
M. A.; Shapiro, M. A. J. Med. Chem. 1995, 38, 4478–4487
11. Brighty, K. E. U.S. Patent 5,229,396, 1992.
12. Cecchetti, V.; Fravolini, A.; Palumbo, M.; Sissi, C.;
Tabarrini, O.; Terni, P.; Xin, T. J. Med. Chem. 1996, 39,
4952–4957.
3. Hooper, D. C. Emerg. Infect. Dis. 2001, 7, 337–341.
4. Fichera, M. E.; Roos, D. S. Nature 1997, 390, 407–409.
5. Roos, D. S.; Crawford, M. J.; Donald, R. G.; Kissinger, J.
C.; Klimczak, L. J.; Striepen, B. Curr. Opin. Microbiol.
1999, 2, 426–432.
6. Ralph, S. A.; D’Ombrain, M. C.; McFadden, G. I. Drug
Resist. Updat. 2001, 4, 145–151.
13. Cecchetti, V.; Fravolini, A.; Lorenzini, M. C.; Tabarrini,
O.; Terni, P.; Xin, T. J. Med. Chem. 1996, 39, 436–
445.
7. Gozalbes, R.; Brun-Pascaud, M.; Garcia-Domenech, R.;
Galvez, J.; Girard, P. M.; Doucet, J. P.; Derouin, F.
Antimicrob. Agents Chemother. 2000, 44, 2771–2776.
Briefly, the virulent RH strain of T. gondii was maintained
in mice by intraperitoneal passage every two days. For each
experiment, tachyzoites were collected from the peritoneal
cavity of infected mice then resuspended in physiological
saline. Tissue cultures and drug tests were carried out using
MRC5 fibroblast tissue cultures. Confluent monolayers
prepared in 96-well tissue culture plates were inoculated
with 2000 fresh tachyzoites. After 4 h, drugs at various
concentrations were added into the culture medium and
culture plates were incubated for an additional 72 h. Each
culture plate comprised eight negative control (without T.
gondii) and eight positive control wells (without drug).
After incubation, the plates were examined microscopically
for cytopathic effects and thereafter fixed with cold
methanol for 5 min. Toxoplasma growth was assessed by
enzyme linked immunoassay (ELISA) performed directly
on the fixed cultures using a peroxydase labeled mono-
clonal antibody directed against the SAG-1 surface protein
of T. gondii. After addition of the substrate, spectropho-
tometric readings were performed at a wavelength of
405 nm with blanking on the negative control well. For
each well, the results were expressed as optical density
(OD) values. The effect of each drug at various concentra-
tions was described by plotting the ODvalues as a function
of the logarithm of the concentration and a linear
regression model was used to summarize the concentra-
tion-effect relationship and to determine the IC50.
14. Cecchetti, V.; Tabarrini, O.; Sabatini, S.; Miao, H.;
Filipponi, E.; Fravolini, A. Bioorg. Med. Chem. 1999, 7,
2465–2471.
15. Da Silva, A. D.; De Almeida, M. V.; De Souza, M. V. N.;
Couri, M. R. C. Curr. Med. Chem. 2003, 10, 21–
39.
16. The chemical structure of compounds 1–4 was unambig-
uously established by 1H, 13C-DEPT, HSQC 1H–13C
NMR, ESI-MS. Compound 1 (TFA salt): ESI-MS:
(M+H)þ 374.2, (M+Na)þ 396.2 (in agreement with the
mass calculated for M ¼ C19H20FN3O4). For atom num-
1
bering, see Scheme 1: H NMR (1:4 CD3CN/D2O): 0.93–
0
1.13 (m, 4H, CH2 cyclopropyl), 2.04 (s, 2H, H1 and H5 ),
0
0
2.54 (s, 1H, H6 ), 3.52 (s, 3H, OCH3), 3.59, 3.82 (d, d, 2H,
2
0
0
2H, H2 /H4 exo/endo, JH–H ¼ 9:8 Hz), 4.10 (m, 1H, CH,
cyclopropyl), 7.63 (d, 1H, H5,3JH–F ¼ 13:1 Hz), 8.75 (s,
1H, H2). 13C NMR (DMSO-d6): 9.3 (s, CH2 cPr), 20.9 (s,
0
0
0
C1 and C5 ), 29.9 (s, C6 ), 41.1 (s, CH cPr), 51.1/51.2 (2s,
0
0
C2 and C4 ), 62.4 (s, OCH3), 106.7 (d, C5,
2JC–F ¼ 24:1 Hz), 106.7 (s, C3), 117.4 (q, TFA, CF3,
1JC–F ¼ 299:0 Hz), 119.9 (d, C9, JC–F ¼ 8:0 Hz), 134.5 (s,
4
2
2
C10), 136.1 (d, C7, JC–F ¼ 11:4 Hz), 144.8 (d, C8, JC–F
¼
1
6:9 Hz), 150.8 (s, C2), 154.9 (d, C6, JC–F ¼ 250:1 Hz),
2
158.8 (q, TFA, COOH, JC–F ¼ 31:7 Hz), 166.3 (s, C4),
176.5 (s, COOH). 19F NMR (1:4 CD3CN/D2O): )74.4
(6.3F, TFA), )117.4 (1F, F6).
17. Mahmoudi, N.; Ciceron, L.; Franetich, J.-F.; Farhati, K.;
Silvie, O.; Eling, W.; Sauerwein, R.; Danis, M.; Mazier,
D.; Derouin, F. Antimicrob. Agents Chemother. 2003, 47,
2636–2639.
18. Renau, T. E.; Gage, J. W.; Dever, J. A.; Roland, G. E.;
Joannides, E. T.; Shapiro, M. A.; Sanchez, J. P.;
Gracheck, S. J.; Domagala, J. M.; Jacobs, M. R.;
Reynolds, R. C. Antimicrob. Agents Chemother. 1996,
40, 2363–2368.
8. Khan, A. A.; Araujo, F. G.; Brighty, K. E.; Gootz, T. D.;
Remington, J. S. Antimicrob. Agents Chemother. 1999, 43,
1783–1787.
9. Gozalbes, R.; Brun-Pascaud, M.; Garcia-Domenech, R.;
Galvez, J.; Girard, P. M.; Doucet, J. P.; Derouin, F.
Antimicrob. Agents Chemother. 2000, 44, 2764–2770.