Prolylisoxazoles: Potent inhibitors of prolyloligopeptidase with antitrypanosomal activity

Article abstract of DOI:10.1016/S0960-894X(03)00579-1

Prolylprolylisoxazoles and prolylprolylisoxazolines were synthesized through a 1,3-dipolar cycloaddition reaction. These compounds are potent inhibitors of human and trypanosomal prolyloligopeptidase. They were shown to inhibit Trypanosoma cruzi and Trypanosoma b. brucei in in vitro systems with ED₅₀'s in the lower μM range.

Full text of DOI:10.1016/S0960-894X(03)00579-1

Bioorganic & Medicinal Chemistry Letters 13 (2003) 2875–2878
Prolylisoxazoles: Potent Inhibitors of Prolyloligopeptidase with
Antitrypanosomal Activity
Gunther Bal,a Pieter Van der Veken,a Dimitri Antonov,a Anne-Marie Lambeir,b
Philippe Grellier,c Simon L. Croft,d Koen Augustynsa and Achiel Haemersa,*
aDepartment of Medicinal Chemistry, University of Antwerp, Belgium
bDepartment of Medical Biochemistry, University of Antwerp, Belgium
Museum National d’Histoire Naturelle, USM 0504 Biologie Fonctionnelle des Protozoaires, Paris, France
c
´
dDepartment of Infectious and Tropical Diseases, London School of Hygiene & Tropical Medicine, London, UK
Received 24 March 2003; revised 3 June 2003; accepted 3 June 2003
Abstract—Prolylprolylisoxazoles and prolylprolylisoxazolines were synthesized through a 1,3-dipolar cycloaddition reaction. These
compounds are potent inhibitors of human and trypanosomal prolyloligopeptidase. They were shown to inhibit Trypanosoma cruzi
and Trypanosoma b. brucei in in vitro systems with ED50’s in the lower mM range.
# 2003 Elsevier Ltd. All rights reserved.
Prolyl oligopeptidase (PO, POP, prolylendopeptidase,
PEP, EC 3.4.21.26) is an endopeptidase with high sub-
strate specificity for proline residues in peptides. It is a
serine protease, cleaving peptide bonds on the C-term-
inal side of prolyl residues within peptides. PO is widely
distributed between organisms and has been isolated
from several sources, including plants, micro-organisms,
invertebrates and various mammalian tissues and body
fluids.1
selectivity index (IC50 human PO/IC50 T. cruzi PO)
reported is around 80.
Most PO inhibitors are substrate analogues. Some
important inhibitors are illustrated in Figure 1. They are
often slow binding inhibitors, with transition state
mimicking properties. They are characterized by a pro-
line residue where the carboxylic group is omitted or
substituted for an electrophilic group such as a carbo-
nitrile, an aldehyde or a ketone. The latter groups scav-
enge the serine alcoholate in the active site of the
enzyme. Sulphonyl-activated Michael substrates are
also used as an electrophilic group. The P2 position is
mostly proline (1) or a cyclic amino acid such as Tic (2)
and the P3 position a carbamate (e.g., benzyloxy-
carbonyl) or amide [e.g., 4-(phenyl)butanoyl] group.
An 80-kDa proteinase (Tc80) secreted by the infective
trypomastigotes of Trypanosoma cruzi was character-
ized as a member of the PO family.2 This enzyme
hydrolyses collagen (types I and IV) and might be
involved in the invasion of the parasite in the mamma-
lian cells by degrading extracellular matrix components.
Hence PO inhibitors may prevent the invasion of
T. cruzi trypomastigotes into mammalian cells. This
protozoan parasite is the causative agent of Chagas’
disease. Several inhibitors have been synthesized3 and
were found to effectively inhibit this invasion in vitro.
Inhibition of human and T. cruzi PO has been compar-
ed.3d Trypanosomal PO is usually more sensitive to
inhibitors but selectivity is limited and the maximum
*Corresponding author. Tel.: +32-38202717; fax: +32-38202739;
e-mail: achiel.haemers@ua.ac.be
Figure 1. PO inhibitors.
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00579-1

2876
G. Bal et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2875–2878
Several reports have described the use of electron-with-
drawing heterocycles to activate the reactive P1
carbonyl group in the development of protease (e.g.,
chymase and human neutrophil elastase) inhibitors.4
A typical approach to synthesize five-membered rings, is
a 1,3-dipolar cycloaddition. A [3+2] cycloaddition8 of a
nitrile oxide to a triple bond will provide an isoxazole
ring. The corresponding isoxazolines are readily pre-
pared by the same [3+2] cycloaddition with the
corresponding alkene.
Tsutsumi et al.5 synthesized PO inhibitors in which a-
ketoheterocycles are introduced at the C-terminal end of
peptides with PO substrate activity. Many of these pro-
ducts exhibit IC50 values at nM levels against PO. A
thiazole as illustrated in Figure 1 shows an IC50 of 8.5
nM against human PO. Structure–activity studies of the
C-terminal heterocyclic groups indicate the importance
of an sp2 nitrogen atom at the b-position of the adjoin-
ing carbonyl group. This heterocyclic nitrogen atom
would provide a critical hydrogen-bond interaction with
the histidine residue of the catalytic triad in PO.5 This
hydrogen bond was also observed in the X-ray crystal
structure of porcine pancreatic elastase with a peptidyl
a-ketobenzoxazole.4a Other ketoheterocycles4c used in
serine protease inhibitor design are pyrroles, oxazoles,
imidazoles, tetrazoles, pyridines, pyrimidines, benzimid-
azoles, benzothiazoles and oxazolopyridines.
The key compound in this synthesis is a nitrile oxide.
Nitrile oxides however are not stable and are generally
formed in situ from nitro compounds.9 Either prolyl
nitro-alcohols or prolyl nitroketones can be used to
prepare the nitrile oxide.
In a first approach nitro-alcohol (5) was synthesized by
a Henry reaction10 between Boc-prolinal (4)11 and
nitromethane using tetrabutylammonium fluoride as a
base.12 The nitro-alcohol was protected by a tri-
methylsilyl group with trimethylsilylchloride and tri-
ethylamine.13 This TMS-protected 2-nitro-alcohol (6)
was dehydrated in situ by phenylisocyanate and a cata-
lytic amount of triethylamine in refluxing benzene to the
desired nitrile oxide (7) which immediately reacted fur-
ther with the alkyne or alkene to the isoxazole or isox-
azoline (8a–i).8,13 Starting materials were commercially
available or were prepared by benzylation of the com-
mercial propargylalcohol and allylalcohol. Deprotection
with trifluoroacetic acid (TFA/H2O) and coupling with
Z-Pro or N-phenylbutanoyl-Pro with TBTU resulted in
the alcohols 9a–i. Final compounds 3a–i were obtained
by a Swern oxidation (Scheme 1).8,14
We were particularly interested in the 3-acylisoxazole
group as inhibitory moiety in PO inhibitor design. This
heterocycle was until now unexplored in this field.
Moreover, it is well known that 3-acylisoxazoles
undergo a ring-opening upon reaction with alcoholates,
giving rise to formylacetonitrile and the corresponding
ester.6 We assumed that the same reaction could happen
in the active site of the enzyme, where the catalytic
serine acts as a nucleophile.
In a second approach the [3+2] cycloaddition was per-
formed on a nitroketone. Boc-proline (10) was activated
as carbonylimidazole (11) using 1,10-carbonyldiimid-
azole (CDI) in dry tetrahydrofurane (THF).15 The acti-
vated proline was not isolated but immediately used in
the next step. Reaction with nitromethane and NaH in
dry THF resulted in the desired 2-nitroketone (12). As
described before, dehydration of this nitroketone using
phenylisocyanate and a catalytic amount of triethyla-
mine, followed by a [3+2] cycloaddition with the cor-
responding alkyne/alkene, gave the isoxazole/
isoxazolines (13a–i). Deprotection with TFA/dichloro-
methane (DCM) and coupling with Z-Pro or N-phe-
nylbutanoyl-Pro with TBTU resulted in the final
compounds 3a–i. The advantage of this pathway is the
reduction of the number of steps from nine to five but
We were also interested in the corresponding iso-
xazolines. We prepared a set (Table 1) of substituted
prolylisoxazoles 3a–3e and prolylisoxazolines 3f–3i as
potential PO inhibiting compounds.
Z-Proline or N-(4-(phenyl)butanoyl)proline was chosen
for the P2–P3 position. These N-substitutions have been
shown to fit well in the non-polar environment of the S3
site of the enzyme.7
Table 1. PO inhibitory activity of prolylisoxazoles 3a–3e and prolyl-
isoxazolines 3f–3i
R1
R2
R3 PO humand PO Tc80d
3a
3b
3c
3d
3e
3f
–H
–H
–H
–COOCH3
–H
–H
Z
Z
Z
Z
Z
Xc
Z
Z
Z
Z
—
36
4
—
15
19
0.34
6
NTf
38
10
16
72
0.28
—
–CH2OCH2C6H5
–COOCH3
–C6H5
e
e
1.87
0.45
NT
1.25
3.30
7.50
0.21
0.26
–Si(CH3)3
–CH2OCH2C6H5
–H
–H
–H
–H
–H
—
3g1a –CH2OCH2C6H5
3g2a –CH2OCH2C6H5
3h
3i
Refb
–CN
–C6H5
—
aDiastereomers.
b2-(Z-Prolyl-prolyl)thiazole.
c4-(Phenyl)butanoyl.
dKi (nM).
eModerately active: IC50 >1 mM.
fNot tested.
Scheme 1. (a) CH3NO2, N(Bu)+4 Fꢀ; (b) TMSCl, Et3N; (c) (i) PhNCO,
Et3N, alkyne/alkene; (ii) AcOH/H2O; (d) (i) TFA/H2O; (ii) Z-ProOH
or N-4-(phenyl)butanoyl-ProOH, TBTU, Et3N; (e) (COCl)2, DMSO.

G. Bal et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2875–2878
2877
All compounds were active against PO in the nM range.
Allowing for some variations for the assays used in
different laboratories were more effective against T.
cruzi PO than against human PO (selectivity about 50).
Even with the small number of compounds tested in this
study, it is clear that the type of substitutions on the
isoxazole/isoxazoline ring is important. Compounds
such as 3c with an electron withdrawing substituent
were less active. There is only a slight difference in
activity between the isoxazoles and the isoxazolines.
Hence, the alcoholate sensitivity of ketoisoxazoles
appears not to be predictive for the inhibitory potency.
Scheme 2. (a) CDI/THF; (b) NaH, CH3NO2; (c) PhCNO, alkyne/
alkene; (f) (i) TFA/DCM; (ii) Z-ProOH or N-(4-(phenyl)butanoyl)-
ProOH, TBTU, Et3N.
the overall yield is lower due to the low yields of the
cycloaddition and of the final coupling step (max 37%)
(Scheme 2).
The most active product against trypanosomal PO is the
5-phenyl isoxazoline 3i.
During the formation of the heterocycle two regio-
isomers can be formed. Indeed, the substituent on the
isoxazole/isoxazoline ring can be in position 4 or 5
depending on the regioselectivity of the 1,3-dipolar
cycloaddition. We could isolate only one regioisomer
for the isoxazoles as well as for the isoxazolines. 1H
NMR data proved that the substituents were located on
position 5 for both the isoxazoles and isoxazolines.16,17
The keto function is essential for this high potency. The
corresponding alcohol of compound 3b (9b) gave a 100-
fold reduction. This confirmed the necessity of an elec-
trophilic moiety at the P1 position.
These results prompted us to test some promising com-
pounds such as 3d and 3f for in vitro antitrypanosomal
activity against Trypanosoma b. rhodesiense, T. cruzi
and Leishmania donovani.21
This regioselectivity is in good correlation with the lit-
erature. Monosubstituted olefins and acetylenes show
high regioselectivity and give 5-substituted derivatives
for both electron donating and electron withdrawing
groups. Only very strong electron withdrawing groups,
such as –SO2R, give predominantly 4-substituted deri-
vatives.18 This selectivity can be explained by a Frontier
orbital model of the 1,3-dipolar cycloaddition and the
application of the perturbation theory.19
The results are summarized in Table 2. Both com-
pounds show ED50 values in the low mM range. T. cruzi
and T. b. rhodesiense appears the most sensitive organ-
ism and 3f the most interesting compound. Indeed the
toxicity (ED50) of this compound against KB cells
(nasopharyngeal cancer cell line) is >300 mg/mL.
We can conclude that these prolylisoxazole derived PO
inhibitors are very interesting compounds in the search
for new antitrypanosomal compounds. It was already
known that PO inhibitors decrease the ability of T. cruzi
trypomastigotes to invade in mammalian cells.3e These
ketoisoxazoles show also substantial antitrypanosomal
activity against L. donovani, T. cruzi and T. b. rhode-
siense in vitro and low toxicity against KB cells. Further
research will be required to investigate the detailed
mechanism of action of these compounds and to estab-
lish a structure–activity relationship. Although 3f
appears non-toxic against KB cells, the compounds
This cycloaddition introduces chirality at position 5 of
the heterocycle. For compound 3g the two diastereo-
mers (3g1 and 3g2) were separated by column
chromatography but not characterized.
The peptidyl a-ketoisoxazoles/-isooxazolines have been
tested as potential inhibitors of human PO20 as well as
for PO Tc 80 from T. cruzi.2 Table 1 shows the inhibi-
tory potencies of the tested compounds on both
enzymes. 2-(Z-Prolyl-prolyl)thiazole (Fig. 1)5 was used
as reference compound.
Table 2. Antiprotozoal activity of compounds 3d and 3f
Parasite
% Inhibition (mg/mL)
ED50 (mg/mL)
Toxicitya
ED50 (mg/mL)
30
10
3
1
3d
3f
L. donovani
T. cruzi
T. b. rhodesiense
L. donovanib
T. cruzic
T. b. rhodesiensed
100 (T)
86.7 (T)
99.9
100 (T)e
86.7(T)
99.9
1.8
59.9
99.8
6.9
69.4
95.5
0
29
83.4
0
42.6
6.9
—
10.8
7.7
—
16
0.9
13.5
6.7
2.2
13.1
4.7
5.5
23
>300
aKB cell toxicity—control drug: podophyllotoxin—ED50: 0.008 mg/mL.
bL. donovani L82—control drug: pentostam—ED50: 9.7 mg/mL.
cT. cruzi Lac Z—control drug: benznidazol—ED50: 0.23 mg/mL.
dT. b. rhodesiense STIB900—control drug: pentamidine–ED50: 0.00125 mg/mL.
e(T): toxic to host macrophages.

2878
G. Bal et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2875–2878
described show only limited selectivity against human
PO. PO degrades proline-containing neuropeptides such
as vasopressin, substance P and thyrotropin-releasing
hormone, which are involved in the process of learning and
memory22 and PO inhibitors are therefore considered as
potential cognition-enhancing drugs. To prevent side
effects due to interference with PO function in mammals
compounds with pronounced selectivity will be necessary.
Tsutsumi, S.; Okonogi, T.; Shibahara, S.; Ohuchi, S.; Hat-
sushiba, E.; Patchett, A. A.; Christensen, B. G. J. Med. Chem.
1994, 37, 3492.
6. Kochetkov, N. K.; Sokolov, S. D. Recent Developments in
Isoxazole Chemistry. In Advances in Heterocyclic Chemistry;
Katritzky, A. R., Ed.; Academic: New York, 1963; p 398.
7. Fulop, V.; Bocksei, Z.; Polgar, L. Cell 1998, 94, 161.
¨
¨
¨
8. Kantorowski, E. J.; Kurth, M. J. J. Org. Chem. 1997, 62,
6797.
9. Mukaiyama, T.; Hoshino, T. J. Am. Chem. Soc. 1960, 82,
5339.
10. Luzzio, F. A. Tetrahedron 2001, 57, 915.
Acknowledgements
11. Fehrentz, J.-A.; Castro, B. Synthesis 1983, 676.
12. Hanessian, S.; Devasthale, P. V. Tetrahedron Lett. 1996,
37, 987.
13. Beebe, X.; Schore, N. E.; Kurth, M. J. J. Org. Chem.
1995, 60, 4196.
14. Tidwell, T. T. Synthesis 1990, 857.
15. Baker, D. C.; Putt, S. R. Synthesis 1978, 478.
16. In the unsubstituted isoxazole (3a), the chemical shifts of
the isoxazole-protons were determined at d 6.73–6.75 ppm and
d 8.45–8.47 ppm. Because of the electron-withdrawing oxygen
next to the C-5 proton, the doublets at d 8.45–8.47 ppm were
attributed to the C-5 proton. Based on this knowledge we were
able to assign the proton signals for the other isoxazoles. The
proton signals were all situated in the range d 6.63–6.88 ppm.
The isoxazolines showed always two protons in the area d
3.08–3.85 ppm, corresponding with the C-4 protons, and only
one proton at d 5.22–6.00 ppm corresponding with the C-5
proton. These chemical shift values are in accordance with
examples from literature.17
This work received support from The Fund for Scientific
Research, Flanders, Belgium (FWO). G.B. is research
assistant with the FWO. P.V.d.V. is a fellow of the Insti-
tute for the Promotion of Innovation by Science and
Technology in Flanders (IWT). S.L.C. received financial
support from UNDP/World/Bank/WHO Special Pro-
gramme for Research and Training in Tropical Diseases.
References and Notes
1. Cunningham, D. F.; O’Connor, B. Biochim. Biophys. Acta
1997, 1343, 160.
2. Santana, J. M.; Grellier, P.; Schrevel, J.; Teixeira, R. L.
Biochem. J. 1997, 324, 129.
3. (a) Vendeville, S.; Buisine, E.; Williard, X.; Schrevel, J.;
Grellier, P. P.; Santana, J.; Sergheraert, C. Chem. Pharm. Bull.
1999, 47, 194. (b) Vendeville, S.; Bourel, L.; Davioud-Charvet,
E.; Grellier, P.; Deprez, B.; Sergheraert, C. Bioorg. Med.
Chem. Lett. 1999, 437. (c) Joyeau, R.; Moulida, C.; Guillet, C.;
Frappier, F.; Teixeira, A. R. L.; Schrevel, J.; Santana, J.;
Grellier, P. Eur. J. Med. Chem. 2000, 35, 257. (d) Vendeville,
S.; Goossens, F.; Debreu-Fontaine, M.-A.; Landry, V.;
Davioud-Charvet, E.; Grellier, P.; Scharpe, S.; Sergheraert, C.
Bioorg. Med. Chem. 2002, 10, 1719. (e) Grellier, P.; Vendeville,
S.; Joyeau, R.; Bastos, I. M. D.; Drobecq, H.; Frappier, F.;
Teixeira, A. R. L.; Schrevel, J.; Davioud-Charvet, E.; Ser-
gheraert, C.; Santana, J. M. J. Biol. Chem. 2001, 276, 47078.
4. (a) Edwards, P. D.; Meyer, E. F., Jr.; Vijayalakshmi, J.;
Tuthill, P. A.; Andisik, D. A.; Gomes, B.; Strimpler, A. J. Am.
Chem. Soc. 1992, 114, 1854. (b) Akahoshi, F.; Ashimori, A.;
Sakashita, H.; Yoshimura, T.; Imada, T.; Nakajima, M.; Mit-
sutomi, N.; Kuwahara, S.; Ohtsuka, T.; Fukaya, C.; Miyazaki,
M.; Nakamura, N. J. Med. Chem. 2001, 44, 1286. (c) Edwards,
P. D.; Wolanin, D. J.; Andisik, D. W.; Davis, M. W. J. Med.
Chem. 1995, 38, 76.
17. (a) Mirta, L.; D’Accorso, N. B.; D’Accorso, F. J. Hetero-
cyclic. Chem. 1996, 33, 1573. (b) Sakamoto, T.; Kondo, Y.;
Uchiyama, D.; Yamanaka, H. Tetrahedron 1991, 47, 5111. (c)
Nasu, T.; Tagawa, T.; Imafuku, K. J. Heterocyclic Chem.
1998, 35, 389.
18. Torssell, K. B. G. In Nitrile Oxides, Nitrones, and Nitro-
nates in Organic Synthesis, Novel Strategies in Synthesis; Feuer
H., Ed.; VCH: New York, 1988; p 25.
19. Houk, K. N.; Sims, J.; Duke, R. E., Jr.; Strozier, R. W.;
George, J. K. J. Am. Chem. Soc. 1973, 95, 7287. Houk, K. N.;
Sims, J.; Watts, C. R.; Luskus, L. J. J. Am. Chem. Soc. 1973,
7301.
20. Goossens, F.; Van Hoof, G.; De Meester, I.; Augustyns,
K.; Borloo, M.; Tourwe, D.; Haemers, A.; Scharpe, S. Eur. J.
Biochem. 1997, 250, 177.
21. (a) Neal, R. A.; Croft, S. L. J. Antimicrob. Chemother.
1984, 14, 463. (b) Buckner, F.; Verlinde, C. L. M. J.; La
Flamme, A. C.; Van Voorhis, W. C. Antimicrob. Agents Chemo-
ther. 1996, 40, 2529. (c) Raz, B.; Iten, M.; GreterBuhler, Y.;
Brun, R. Acta Trop. 1997, 139.
5. (a) Tsutsumi, S.; Okonogi, T.; Shibahara, S.; Patchett,
A. A.; Christensen, B. G. Bioorg. Med. Lett. 1994, 4, 831. (b)
22. De Nanteuil, G.; Portevin, B.; Lepagnol, J. Drugs Future
1998, 23, 167.