A designed P1 cysteine mimetic for covalent and non-covalent inhibitors of HCV NS3 protease.

Article abstract of DOI:10.1016/S0960-894X(01)00842-3

The difluoromethyl group was designed by computational chemistry methods as a mimetic of the canonical P1 cysteine thiol for inhibitors of the hepatitis C virus NS3 protease. This modification led to the development of competitive, non-covalent inhibitor 4 (K(i) 30 nM) and reversible covalent inhibitors (6, K(i) 0.5 nM; and 8 K*(i) 10 pM).

Full text of DOI:10.1016/S0960-894X(01)00842-3

Bioorganic & Medicinal Chemistry Letters 12 (2002) 701–704
A Designed P1 Cysteine Mimetic for Covalent and Non-Covalent
Inhibitors of HCV NS3 Protease
Frank Narjes,a,* Konrad F. Koehler,a Uwe Koch,a Benjamin Gerlach,a
Stefania Colarusso,a Christian Steinkuhler,b Mirko Brunetti,b Sergio Altamura,b
Raffaele De Francescob and Victor G. Matassaa
aDepartment of Chemistry, IRBM, MRL Rome, Via Pontina Km 30.600, Pomezia, 00040 Rome, Italy
bDepartment of Biochemistry, IRBM, MRL Rome, Via Pontina Km 30.600, Pomezia, 00040 Rome, Italy
Received 3 October 2001; revised 4 December 2001; accepted 12 December 2001
Abstract—The difluoromethyl group was designed by computational chemistry methods as a mimetic of the canonical P1 cysteine
thiol for inhibitors of the hepatitis C virus NS3 protease. This modification led to the development of competitive, non-covalent
inhibitor 4 (Ki 30 nM) and reversible covalent inhibitors (6, Ki 0.5 nM; and 8 Ki* 10 pM). # 2002 Elsevier Science Ltd. All rights
reserved.
The hepatitis C virus (HCV) is the major causative
agent of parenterally-transmitted and sporadic non-A,
non-B hepatitis (NANB-H),1 and is believed to affect
some 200 million people worldwide.2 Infection by the
virus can result in chronic hepatitis and cirrhosis of the
liver, and may lead to hepatocellular carcinoma. There
is neither a vaccine nor an established therapy for HCV,
although partial success has been achieved by treatment
with recombinant a-interferon, either alone or in com-
bination with ribavirin.2,3
acids. The reversible competitive inhibitor 1 (Ki 40 nM,
Table 1) emerged from optimization of an initial lead
identified in biochemical studies. Cysteine was estab-
lished as the preferred P1 amino acid for intermolecular
cleavage of substrate by NS3 and for product-based
inhibitors.5 This structural characteristic complicates
the ready preparation of serine-trap inhibitors due to
the juxtaposition of mutually-incompatible nucleophilic
(i.e., thiol) and electrophilic (i.e., serine-trap) function-
ality. As the thiol is a potential Achilles’ heel for any
drug that derives from this series of molecules, we
sought a surrogate for the P1 cysteine sulfhydryl to
facilitate the preparation of stable biochemical tools and
to progress our drug discovery efforts. We report here:
(i) the design by computational chemistry methods of a
P1 thiol replacement; (ii) its successful introduction into
product-based inhibitors; and (iii) its utility in preparing
potent serine-trap inhibitors of NS3. In the accom-
panying paper we will present the evolution of these
inhibitors to small and potent tripeptides.9
A serine protease located in the N-terminal domain of
the NS3 protein is one of the prime targets for the
development of antiviral drugs, since it is responsible
for proteolytic maturation of most of the non-structural
region of the viral polyprotein. The NS3 protease
belongs to the trypsin superfamily, and is unique in
requiring a (non-catalytic) structural zinc atom and a
co-factor protein (NS4A).4,5 Crystal structures of the
protease4,5 reveal that the S (non-prime)6 site of the
substrate-binding channel is strongly cationic, solvent-
exposed and relatively featureless, explaining why the
minimum substrate is a polycarboxy decapeptide.
Design of a P1 Thiol Replacement
The NS3 protease is subject to product inhibition by
N-terminal hexapeptide7 and tetrapeptide8 carboxylic
The S1 specificity pocket of the NS3 protease is small
and lipophilic, comprising Leu135, Phe154 and Ala157.
Substrate mutagenesis studies4,5 and product inhibitor
optimization7 concur that cysteine is preferred at P1.
The cysteine sulfhydryl is lipophilic, but has lone-pairs
*Corresponding author. Tel.: +39-06-910-93-625; fax: +39-06-910-
93-654; e-mail: frank_narjes@merck.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00842-3

702
F. Narjes et al. / Bioorg. Med. Chem. Lett. 12 (2002) 701–704
of electrons and a polarized S–H bond which confer the
potential for electrostatic non-covalent interactions.10
In attempting to discover a novel P1 substituent, an
analysis was therefore made of the steric and electro-
static properties of the thiol.11 Subsequent comparison
with a selection of small fluoroalkyl and fluoroalkenyl
groups suggested interesting parallels for CF2H. The
steric and electrostatic parameters for CF2H and SH
were evaluated on 1,1-difluoroethane and methanethiol,
respectively. The volume of 1,1-difluoroethane, 46.7 A3,
is similar to that of methanethiol (47.1 A3) (Fig. 1). The
MEP surfaces are also similar (Fig. 2), displaying a
substantial build-up of negative potential around the
sulfur lone pairs and the two fluorine atoms, and a
positive potential around the S- and CF2-bound pro-
tons. The CF2H and SH groups thus emerged as having
similar calculated steric and electrostatic features which
encouraged an investigation of CF2H in the P1 position
of NS3 protease inhibitors.
Thus, (Æ)-4,4-difluoro-2-aminobutyric acid (difluoro-
Abu) (13)13 was converted to the aldehyde 15 in three
steps. Acetalization and removal of the Boc group gave
amine 16. Coupling with the pentapeptide AcAs-
p(OtBu)Glu(OtBu)DifGlu(OtBu)-ChaOH (17) was fol-
lowed by complete deprotection and purification with
separation of the diastereomers by HPLC to give the
aldehyde 6.14 Product inhibitors 4 and 5 were syn-
thesized as described12 using chiral 13 prepared from l-
or d-aspartic acid.15 Using allylglycine instead of
difluoroAbu gave inhibitors 3 and 7.16
Scheme 2 shows the synthesis of a-ketoamide 10 and the
a-ketoheterocycles 11–12, as exemplified for benzothia-
zole 11a. Conversion of aldehyde 15 to the a-hydroxy-
amide 18 was achieved in five steps, while reaction of
Weinreb amide 14 with 2-lithio-benzothiazole gave 19
after reduction and deprotection. Both compounds were
then converted to the final products as described above,
except that oxidation of the hydroxy groups was carried
out prior to the final deprotection step.
Synthesis
The synthesis of 1, 2, 8 and 9 have been described.7,12
The aldehyde 6 was prepared as outlined in Scheme 1.
Results and Discussion
Substitution of the P1 cysteine by (S)-difluoroAbu was
first investigated in hexapeptide product acids. Inhibi-
tors were evaluated on the NS3/4A complex described
Figure 1. Orthogonal views of the van der Waals surfaces of meth-
anethiol (cyan) and 1,1-difluoroethane (red). The surfaces are Z-clipped
for clarity.
Scheme 1. (a) Boc2O; (b) MeNH(OMe) HCl, EDCI, HOBt, iPr2NEt;
(c) DIBAL, THF, À78 ꢀC; (d) HC(OMe)3, TsOH; (e) HCl, MeOH,
0 ꢀC–rt; (f) 17, EDCI, HOBt, iPr2NEt, DCM; (g) TFA, H2O, DCM;
(h) RP-HPLC.
Figure 2. Orthogonal views of electrostatic potentials displayed over
the electron density surface of methanethiol (top) and 1,1-difluoro-
ethane (bottom). Stick diagrams of the respective compounds are
displayed to the left and color-coded scales of electrostatic potential
are displayed on the right for reference.
Scheme 2. (a) Me2C(OH)CN, TEA, DCM; (b) HCl, H2O, dioxane; (c)
Boc2O; (d) BnNH2, EDCI, HOBt; (e) HCl, EtOAc; (f) 17, EDCI,
HOBt, iPr2NEt, DCM; (g) Dess–Martin, DCM, t-BuOH; (h) TFA,
H2O, DCM; (i) RP-HPLC; (j) benzothiazole, BuLi, THF, À78 ꢀC; (k)
NaBH4, EtOH, 0 ꢀC.

F. Narjes et al. / Bioorg. Med. Chem. Lett. 12 (2002) 701–704
703
previously.17 The difluoro analogue 4 (Ki 30 nM) is at
least as potent as its cysteine counterpart 1 (Ki 40 nM)7
and substantially more active than the Abu and the allyl
analogues 2 and 3 (Table 1). Compound 5 (Ki 640 nM),
the P1 diastereomer of 4, was less active.
Simplistically, ketoacid 8 may be considered as an inhi-
bitor that combines the binding energy contributed by a
covalent bond (as in aldehyde 6) with the electrostatic
stabilization of the charged inhibitor by Lys136 (as in
acid 4).12,18 This is also evident from two close analo-
gues, a-ketoester 9 and a-ketoamide 10. Although both
are potent inhibitors of NS3, as was observed previously
in related series,19 they do not achieve the potency of 8.
The identification of this non-reactive and equipotent
thiol surrogate facilitated the synthesis and evaluation
of serine trap-based inhibitors. Of these the a-ketoacid 8
(Ki* 10 pM) proved to be particularly effective. Kinetic
investigations showed that the potency of 8 is due to the
extremely slow dissociation rate of the covalent
adduct.12 The aldehyde 6 (Ki 0.5 nM) is a reversible
inhibitor whose affinity is intermediate between that of
the product acid 4 and the ketoacid 8. Allyl aldehyde 7
was significantly less active than 6.
Inhibitors incorporating the a-keto heterocycle moiety
like 11 and 12 proved to be less active than 4, 7, or 8.
These heterocyclic groups give potent inhibitors with
the related proteases elastase, chymase and thrombin.20
Presumably, their lack of activity is due to some steric
restriction in the active site of the enzyme, which pre-
vents optimal binding.
Table 1. Inhibitors of NS3/4A protease
From our results and previous studies, it is evident that
the substitution of the P1 cysteine using conventional
amino acids in inhibitors and in substrates failed to find
an effective replacement,4,5,17 suggesting that the
difluoromethyl group has specific properties which
mimic the thiol. The CF2H group has a similar volume
and shape to SH (Fig. 1). Also, SH and CF2H are lipo-
philic, a feature that is presumably of importance to
binding, since the P1 pocket is substantially hydro-
phobic.
Compd
R
Ki (nM)
1
40
Electrostatic properties could play a role, however. The
P1 cysteine has non-covalent contacts with the phenyl
ring of the S1 Phe1544,5 (Fig. 3). The thiol could interact
with the phenyl ring either as a proton donor, through
interaction of the SH proton with the electron-rich pi-
cloud, or as a proton acceptor, through interaction of
the sulfur lone electron pairs with a ring proton, the
latter having partial positive character. Electrostatic
interactions involving the pi-cloud and protons of a
phenyl ring are well documented.21 Directional sulfur-
aromatic interactions are commonly observed in protein
crystal structures,22 and the thiol group can be a
hydrogen bond donor or acceptor in biological sys-
tems.23,24 The cysteine sulfhydryl thus could in principle
act in either capacity in the NS3 protease P1–S1 inter-
action. Despite its calculated electron-richness, in prac-
tice aliphatic fluorine rarely functions as a hydrogen
2
3
700
100
4
5
6
30
640
0.5
7
7
8
0.01*
2.0
9
10
1.5
11a (X=S)
11b (X=O)
150
600
12
1500
Figure 3. Interactions of the P1 difluoroAbu in the P1 specificity
pocket according to the crystal structure of the NS3-protease with the
ketoacid covalently bound to the catalytic serine 139.

704
F. Narjes et al. / Bioorg. Med. Chem. Lett. 12 (2002) 701–704
bond acceptor, probably because of its high electro-
negativity and low polarizability.25,26 In contrast, CF2H
has been documented as a proton donor in hydrogen
bonding interactions.27 Recent structural studies28 of
tripeptide analogues of 6 co-crystallized with NS3/4A
confirm this design. The CF2H moiety is almost com-
pletely buried in the P1-specificity pocket creating a
large lipophilic contact surface area. In this structure
the flexible, lipophilic part of the Lys136 sidechain is
oriented so that it covers the CF2H moiety, while the
ammonium group is close to the ketoacid carboxylate
(Fig. 3). In the rather apolar environment of the P1-
specificity pocket electrostatic complementarity becomes
important, even for relatively unpolar groups. Accord-
ingly, the CF2H is oriented with its proton close to the
carbonyl oxygen of Lys136 and one fluorine atom in
contact with the para hydrogen of Phe154. Thus, steric
and electrostatic similarity of the Cys sidechain and
difluoroAbu allows a similar mode of binding in the P1-
specificity pocket with almost complete burial of the
sidechain. The observation that difluoroAbu works
under such highly specific conditions as a Cys mimetic
underlines its general applicability as a non-reactive Cys
surrogate.
10. Burley, S. K.; Petsko, G. A. In Weakly Polar Interactions
in Proteins. In: Adv. Prot. Chem. 39; Anfinsen, C. B., Edsall, J.
T., Richards, F. M., Eisenberg, D. S., Eds.; Academic: San
Diego, USA, 1988; p 125.
11. Calculations were performed on a Silicon Graphics
Extreme workstation (R4000 CPU). Ab initio calculations
were performed using Spartan version 5.02, with default set-
tings. Starting geometries of methanethiol and 1,1-difluoro-
ethane were built using the Spartan molecular editor and were
fully optimized using the 6-31G* basis set. Molecular electro-
static potentials (MEP) were plotted over the total electron
density surface using the 6-31G* wave function. Ab initio
optimized geometries were imported into Sybyl version 6.3
and the surfaces were generated using the Van der Waals dot
option.
12. Narjes, F.; Brunetti, M.; Colarusso, S.; Gerlach, B.; Koch,
U.; Biasiol, G.; Fattori, D.; De Francesco, R.; Matassa, V. G.;
Steinkuhler, C. Biochemistry 2000, 39, 1849.
13. Analogous to the trifluoro analogue: Tsushima, T.;
Kawada, K.; Ishihara, S.; Uchida, N.; Shiratori, O.; Higaki,
J.; Hirata, M. Tetrahedron 1988, 44, 5375.
14. The more active diastereomer was assigned the l-config-
uration at P1, in accord with the results from 4 and 5 and
recently published structures of enzyme inhibitor complexes
(see ref 28).
15. Winkler, D.; Burger, K. Synthesis 1999, 1419.
16. Intermediates and final compounds have been character-
1
ized by H and 13C NMR, MS and RP-HPLC.
17. Urbani, A.; Bianchi, E.; Narjes, F.; Tramontano, A.; De
Francesco, R.; Steinkuhler, C.; Pessi, A. J. Biol. Chem. 1997,
272, 9204.
Acknowledgements
The authors thank K. Prendergast for valuable discus-
sions, B. Wang for early contributions, A. Pessi for a
sample of 2, S. Pesci for NMR spectra, and F. Naimo
and F. Bonelli for Mass spectra.
18. Koch, U.; Biasiol, G.; Brunetti, M.; Fattori, D.; Brunetti,
M.; Palloro, M.; Steinkuhler, C. Biochemistry 2001, 40, 631.
19. (a) Bennett, J. M.; Campbell, A. D.; Campbell, A. J.;
Carr, M. G.; Dunsdon, R. M.; Greening, J. R.; Hurst, D. N.;
Jennings, N. S.; Jones, P. S.; Jordan, S.; Kay, P. B.; O’Brien,
M. A.; King-Underwood, J.; Raynham, T. M.; Wilinson, C. S.;
Wilinson, T. S. I.; Wilson, F. X. L. Bioorg. Med. Chem. Lett.
2001, 11, 355. (b) Han, W.; Hu, Z.; Jiang, X.; Decicco, C. P.
Bioorg. Med. Chem. Lett. 2000, 10, 711.
20. Akahoshi, F.; Ashimori, A.; Sakashita, H.; Yoshimura,
T.; Imada, T.; Nakajima, M.; Mitsutomi, N.; Kuwahara, S.;
Ohtsuka, T.; Fukaya, C.; Miyazaki, M.; Nakamura, N. J.
Med. Chem. 2001, 44, 1286 and references cited therein.
21. Ma, J. C.; Dougherty, D. A. Chem. Rev. 1997, 97, 1303
and references cited therein.
22. Reid, K. S. C.; Lindley, P. F.; Thornton, J. M. FEBS
1985, 190, 209.
23. Platts, J. A.; Howard, S. T.; Bracke, B. R. F. J. Am.
Chem. Soc. 1996, 118, 2726.
24. Erlanson, D. A.; Verdine, G. L. Chem. Biol. 1994, 1, 79.
25. Dunitz, J. D.; Taylor, R. Chem. Eur. J. 1997, 3, 89.
26. Howard, J. A. K.; Hay, V. J.; O’Hagan, D.; Smith, G. T.
Tetrahedron 1996, 52, 12613.
27. Erickson, J. A.; McLoughlin, J. J. Org. Chem. 1995, 60,
1626.
References and Notes
1. Review: Clarke, B. J. Gen. Virol. 1997, 78, 2397.
2. Cohen, J. Science 1999, 285, 26.
3. Dymock, B. W.; Jones, P. S.; Wilson, F. X. Antiviral Chem.
Chemother. 2000, 11, 79.
4. Review: Kwong, A. D.; Kim, J. L.; Rao, G.; Lipovsek, D.;
Raybuck, S. A. Antiviral Res. 1998, 40, 1.
5. Review: De Francesco, R.; Steinkuhler, C. Curr. Top. Micr.
Immunol. 2000, 242, 149.
6. Schechter, I.; Berger, A. Biochem. Biophys. Res. Commun.
1967, 27, 157.
7. Ingallinella, P.; Altamura, S.; Bianchi, E.; Taliani, M.;
Ingenito, R.; Cortese, R.; De Francesco, R.; Steinkuhler, C.;
Pessi, A. Biochemistry 1998, 37, 8906.
8. Llinas-Brunet, M.; Bailey, M.; Fazal, G.; Ghiro, E.; Gorys,
V.; Goulet, S.; Halmos, T.; Maurice, R.; Poirier, M.; Poupart,
M.-A.; Rancourt, J.; Thibeault, D.; Wernic, D.; Lamarre, D.
Bioorg. Med. Chem. Lett. 2000, 10, 2267.
9. Colarusso, S.; Gerlach, B.; Koch, U.; Muraglia, E.; Conte,
I.; Stansfield, I.; Matassa, V. G.; Narjes, F. Bioorg. Med.
Chem. Lett. 2002, 12, 705.
28. Di Marco, S.; Rizzi, M.; Volpari, C.; Walsh, M. A.;
Narjes, F.; Colarusso, S.; De Francesco, R.; Matassa, V. G.;
Sollazzo, M. J. Biol. Chem. 2000, 275, 7152.