Novel, potent phenethylamide inhibitors of the hepatitis C virus (HCV) NS3 protease: Probing the role of P2 aryloxyprolines with hybrid structures

Article abstract of DOI:10.1016/S0960-894X(03)00536-5

Synthesis of hybrid HCV NS3 protease/NS4A inhibitors having the 4,4-difluoroaminobutyric acid (difluoroAbu) phenethylamides as P1-P1′ and quinolyloxyprolines as P2 fragments led to 7 (IC₅₀ 54 nM). Molecular modelling suggests that this potent tripeptide inhibitor utilizes interactions in the S1′, S1, S2, S3 and S4 sites of the protease.

Full text of DOI:10.1016/S0960-894X(03)00536-5

Bioorganic & Medicinal ChemistryLetters 13 (2003) 2745–2748
Novel, Potent Phenethylamide Inhibitors of the Hepatitis C Virus
(HCV) NS3 Protease: Probing the Role of P2 Aryloxyprolines
with Hybrid Structures
Federica Orvieto, Uwe Koch, Victor G. Matassay and Ester Muraglia*
Medicinal Chemistry Department, IRBM-MRL Rome, V. Pontina Km 30,600, 00040 Pomezia, Rome, Italy
Received 13 January2003; revised 25 February2003; accepted 16 May2003
Abstract—Synthesis of hybrid HCV NS3 protease/NS4A inhibitors having the 4,4-difluoroaminobutyric acid (difluoroAbu) phe-
nethylamides as P1–P10 and quinolyloxyprolines as P2 fragments led to 7 (IC50 54 nM). Molecular modelling suggests that this
potent tripeptide inhibitor utilizes interactions in the S10, S1, S2, S3 and S4 sites of the protease.
# 2003 Elsevier Ltd. All rights reserved.
Despite its high global prevalence, the onlycurrent
treatment for hepatitis C infection is a-interferon, either
alone or in combination with the antiviral ribavirin.1
These therapies are not of general efficacyand are
characterised bysevere side effects.
subsequent utilityin enabling the preparation of potent
5
monocarboxytripeptide ketoacid inhibitors. The crys-
tal structures of ternaryketoacid/protease/co-factor
complexes6 and solution NMR data7 provided molec-
ular-level understanding of the covalent and non-cova-
lent interactions involved in the inhibition process.
Several likelymolecular targets for future drug therapy
have been identified in recent years, among them the
NS3 protease, a unique serine protease that acts in
concert with a co-factor, the NS4A protein, to cleave
the viral gene product at several defined positions, deli-
vering mature viral peptides that are believed to be
necessaryfor the life-cycle of the virus. The NS3 pro-
tease is thus a primaryfocus for small-molecule drug
discovery.
We recentlyreported a new series of phenethyl amide
NS3 protease inhibitors that feature a phenyl ring in
P10.8 SAR showed that a carboxylic acid in the para
position was beneficial for activityand that introduction
of halogens meta to the carboxylate gave further
improvements (Fig. 1). Molecular modelling, mutagen-
esis studies, and subsequent SAR stronglysuggested
that the phenyl ring found a unique binding site along-
side Lys-136.8 Additional SAR in this series showed
that 3-substituted prolines were a preferred P2 group.9
2
The NS3/NS4A protease/cofactor is an inherently
daunting target for drug design, however, on account of
its overall architecture and the molecular features of
both the protease itself and its (minimal) polycarboxy
decapeptide substrates. NS3 is subject to product inhib-
ition: the thiol-containing, polycarboxy hexapeptide
Independently, workers at Boehringer produced an
interesting and novel series of highlypotent mono-
carboxytripeptide product-like inhibitors that had
unique aryl fragments (e.g., 4-quinolyloxy, (2-phenyl-7-
methoxy)-4-quinolyloxy) attached at C3 of a P2 proline
(1,6).10
3
4
product inhibitors reported byus and byothers, have
been used as a starting point in the search for drug-like
leads. We have described our progress in designing a P1
replacement for the ubiquitous cysteine and its
*Corresponding author. Tel.: +39-06-910-93-276; fax: +39-06-910-
93-225; e-mail: ester_muraglia@merck.com
yPresent address: GraffinityPharmaceuticals AG, Im Neuenheimer
Feld 518, Heidelberg, 69120, Germany.
Figure 1. Tripeptide phenethylamides HCV protease inhibitors.8
0960-894X/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00536-5

2746
F. Orvieto et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2745–2748
We were intrigued bythe possibilitythat the aryl rings
from our phenethyl amides and the Boehringer quino-
lines might occupythe same binding site on the pro-
tease. To probe this, we have made hybrid structures
between the two families of inhibitors. We envisaged
two positive outcomes:
Given that the P1 difluoroethyl substituent is compat-
ible with either a terminal carboxygroup or the p-car-
boxyphenethyl amide,8 the data in Table 1 suggest an
incompatibility between the P1 vinylcyclopropyl sub-
stituent and/or the P2 quinolyl group with the phenethyl
amide group in 3, 4.
a. If the two series shared a common aryl binding
site, novel hybrid series could be designed, for
example where the phenethyl amide ring could be
elaborated into substituted quinolines; or
b. If the aryl binding sites were discrete, then
alternative novel hybrid series could result
wherein both aryl groups could be present in the
same structure.
To discern whether the incompatibilitywas due to P1 or
P2, analogues were synthesized that had 4,4-difluoro-2-
aminobutyric acid (difluoroAbu) in P1.12 In addition,
we chose to use the potent Boehringer P2 (2-phenyl-7-
methoxyquinolyl proline10), together with the most
potent Merck phenethyl amide (difluoro-carboxy-
phenyl).13 The acid 5 was prepared as a mixture of P1
diastereomers (Table 2).
Our unexpected findings are the subject of this
communication.
Allowing for the presence of one (largely) inactive dia-
stereomer,12 (S)-5 (IC50 ꢀ90 nM) had an activitysome-
what less, but in the same range as that of 6 (IC50
ꢀ25 nM).10 This result demonstrated that the vinylcy-
clopropyl group could be successfully replaced by the
difluoroethyl group. However, by contrast with the
results of Table 1, making the carboxyphenethyl amide
of (S)-5 gave the potent hybrid 7 (IC50 54 nM) (Table 3).
To afford an initial comparison, the diastereomeric
phenethyl amides 3 and 4 were prepared from the par-
ent acids 1, 210 (Table 1). In contrast to 1 (IC50 330 nM)
which was, as reported,10 a potent inhibitor of the pro-
tease, the phenethyl amides 3, 4 were completelyinac-
tive up to a concentration of 100 mM. Inhibition data
were obtained on the full-length NS3/NS4A complex.11
With difluoroAbu in P1, we can therefore conclude that
the P10 phenethyl amide ring and the P2 quinoline (or
substituents) are binding at separate sites.
The comparison of 3, 4 with 7 suggests interference
between the phenethyl amide and the vinylcyclopropyl
group in the enzyme-bound conformation of 3/4.
According to our modelling, the vinylcyclopropyl group
causes a distinct change in the binding conformation of
the phenethyl amide, reducing favorable interactions in
the oxoanion cavityand between the phenyl ring and
Lys136 (Fig. 2).8 Consequently, the P10 SAR for the P1
vinylcyclopropyl series is different from the difluoroAbu
Table 1. P1 vinylcyclopropyl acids and corresponding phenethyl-
amides
Table 2. Comparison of P1 vinylcyclopropyl and difluoroethyl acids
Compd
P1
IC50 (mM)
a,b
1
0.33
2
3
naa,c
nac
Compd
P1
IC50 (mM)
5
0.180
4
nac
6
0.025a
aRef 10.
bIC50 <500 nM,10 confirmed bysnythesis and testing of the com-
pounds.
cna, not active.
aRef 10.

F. Orvieto et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2745–2748
Table 3. P1-Difluoroabu-phenethylamides
2747
character and molecular weight in these non-covalent
inhibitors (Table 3). Replacing the terminal urethane in
8 byan iso-butyl alcohol14 gave 9 that was some 25–30-
fold less active. Capping the proline as a t-butyl amide
abolished activity( 10: IC50 > >100 mM). Simplification
of 7 gave 11, which was essentiallyinactive.
Simplifications at P3 resulted in significant drop in
activity. The mode of binding and SAR of the P3 side
chain and the capping group, which essentiallybinds to
the S4 site, is similar to that observed for other phe-
nethylamides (Fig. 2).8 We observe close contacts
between the P1 and P3 and the P2 and capping group side
chains indicating a direction for further optimization.
Compd
P3
R
F
IC50 (mM)
a
7
0.054
0.4
8
9
H
H
Synthesis
All the final compounds were isolated bypreparative
RP-HPLC.15 1, 2 were synthesized according to the lit-
erature.10 3, 4 were prepared from the diastereomeric
mixture at P1 of acids 1, 2 bycoupling with 4-methoxy-
carbonylphenethylamine (HATU, DIPEA, DCM),
separation of the two diastereomers byflash chromato-
graphy, followed by hydrolysis (NaOH, MeOH/THF).
The synthesis of 5, 7–11 is summarized in Scheme 1.
Boc-protected proline 1210 was coupled with methyl 4,4-
difluoro-2-aminobutyrate, readily obtained from 4,4-
difluoro-2-aminobutyric acid,12 to give 13 that, after
Boc removal (14), was reacted with N-Boc-(l)-t-Bu-
GlyCOF, obtained by fluorination with tri-
fluorotriazine16 of N-Boc-(l)-t-Bu-Gly, and the crude
methyl ester was hydrolyzed to 5. Coupling of 12 with
the preformed phenethylamide 15 (2,5-difluoro-4-t-
butoxycarbonylphenethylamide of (S)-4,4-difluoro-2-
aminobutyric acid),13 produced 16, that underwent
simultaneous Boc and t-Bu ester deprotection, followed
bycoupling of the resulting aminoacid with an excess of
N-Boc-(l)-t-Bu-GlyOH, to give 7.
11
10
11
H
F
nab
Me
38,150c
aSingle diastereomer, S configuration at P1.
bna, not active.
cValues for the two diastereomers.
Phenethyl amides 8–10 were prepared from 13, which
was hydrolyzed to the corresponding carboxylic acid
and coupled with 4-methoxycarbonylphenethylamine to
give 17. 17 was Boc-deprotected and coupled with the
required carboxylic acid (8, 9) or acylated with pivaloyl
chloride (10) and finally hydrolyzed to the final com-
pounds. The two diastereomers of 11 were synthesized
from 14 byacetylation followed byseparation of the
two diastereomers at P1 (18) byflash chromatography.
Each of the two diastereomers was then separately
hydrolyzed, coupled with 2,5-difluoro-4-t-butoxy-
carbonylphenethylamine, and finally deprotected.
Figure 2. Modeling of 7 in the NS3/NS4A complex.
series. Interestingly, whereas addition of the phenethyl-
amide to tripeptides with a P2 Leu caused a significant
increase in activitywith respect to the corresponding
carboxylic acid,8 in our case of tripeptides with a P2
quinolinoxy-proline such a significant improvement was
not observed. The SAR trends are therefore not para-
llel, perhaps not surprisinglygiven the structural
differences at P2 between the two series.
In conclusion, through synthesizing hybrid inhibitors
that incorporated diverse P1, P2 and P10 groups we gen-
erated potent tripeptide inhibitors that utilize interac-
tions in the S10, S1, S2, S3 and S4 sites of the protease.
We established that the phenyl rings of the P10 phenethyl
amide and 2-phenyl-7-methoxyquinoline substituent of
P2 proline bind at different sites on the protease.
With potent phenethyl amides in hand, P3 SAR was
brieflystudied in an effort to further reduce peptide
With 4,4-difluoro-2-aminobutyric acid in P1, the phe-
nethyl amide group is compatible with the presence of

2748
F. Orvieto et al. / Bioorg. Med. Chem. Lett. 13 (2003) 2745–2748
Scheme 1. (a) Methyl 4,4-difluoro-2-aminobutyrate (1 equiv), HATU (1.2 equiv), DIPEA (3 equiv), DCM, 82%; (b) 4 N HCl in dioxane; (c) Boc-(l)-
tBuGlyCOF (1.2 equiv), TEA (3 equiv), DCM; (d) 1 N NaOH, MeOH, THF; (e) (S)-15 (1 equiv), HATU (1.2 equiv), DIPEA (2.5 equiv), DCM,
96%; (f) TFA–H2O; (g) Boc-(l)-tBuGlyOH (4 equiv), HATU (3.5 equiv), DIPEA (7 equiv), DCM; (h) 1 N NaOH, MeOH; (i) 4-methoxy-
carbonylphenethylamine (1 equiv), HOBt (2 equiv), DIPEA (3 equiv), EDC (1.5 equiv), DCM; (l) TFA; (m) RCO2H (1 equiv), HATU (1.2 equiv),
DIPEA (5 equiv), DCM; (n) tBuCOCl (2 equiv), DIPEA (5 equiv), DCM; (o) AcCl (1.5 equiv), Et3N (5 equiv), DCM; flash chromatography; (p) 2,5-
difluoro-4-t-butoxycarbonylphenethylamine (1.1 equiv), HATU (1.3 equiv), DIPEA (12 equiv), DCM.
I.; Stansfield, I.; Matassa, V.; Narjes, F. Bioorg. Med. Chem.
Lett. 2002, 12, 705.
6. 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.
7. Barbato, G.; Cicero, D. O.; Cordier, F.; Narjes, F.; Ger-
the phenyl-methoxyquinoline at P2. However, the P1
vinylcyclopropyl group is not compatible with the phe-
nethyl amide.
Acknowledgements
lach, B.; Sambucini, S.; Grzesiek, S.; Matassa, V. G.; De
Francesco, R.; Bazzo, R. EMBO J. 2000, 19, 1195.
The authors gratefullyacknowledge Christian Stein-
kuhler for the assaydata, Silvia Pesci for NMR spectra,
Michael Rowleyfor valuable discussions.
8. Colarusso, S.; Koch, U.; Gerlach, B.; Steinkuhler, C.; De
Francesco, R.; Altamura, S.; Matassa, V. G.; Narjes, F. J.
Med. Chem. 2003, 46, 345.
9. Unpublished results.
10. Llinas-Brunet, M.; Bailey, M. D.; Cameron, D.; Faucher,
A.-M.; Ghiro, E.; Goudreau, N.; Halmos, T.; Poupart, M.-A.;
Rancourt, J.; Tsantrizos, Y. S.; Wernic, D. M.; Simoneau, B.
PCT Int. Appl. 2000, WO 00/09543.
11. Gallinari, P.; Paolini, C.; Brennan, D.; Nardi, C.; Stein-
kuhler, C.; De Francesco, R. Biochemistry 1999, 38, 5620.
12. Narjes, F.; Koehler, K. F.; Koch, U.; Gerlach, B.; Colar-
usso, S.; Steinkuhler, C.; Brunetti, M.; Altamura, S.; De
Francesco, R.; Matassa, V. G. Bioorg. Med. Chem. Lett. 2002,
12, 701.
References and Notes
1. Beaulieu, P. L.; Llinas-Brunet, M. Curr. Med. Chem. Anti-
Infect. Agents 2002, 1, 163.
2. Steinkuhler, C.; Koch, U.; Narjes, F.; Matassa, V. G. Curr.
Med. Chem. 2001, 8, 919, and refs therein.
3. Ingallinella, P.; Altamura, S.; Bianchi, E.; Taliani, M.;
Ingenito, R.; Cortese, R.; De Francesco, R.; Steinkuhler, C.;
Pessi, A. Biochemistry 1998, 37, 8906.
4. Llinas-Brunet, M.; Bailey, M. D.; Fazal, G.; Goulet, S.;
Halmos, T.; Laplante, S.; Maurice, R.; Poirer, M.; Poupart,
M.-A.; Thibeault, D.; Wernic, D. M.; Lamarre, D. Bioorg.
Med. Chem. Lett. 1998, 8, 1713.
13. Narjes, F. et al. Manuscript in preparation.
14. For a-hydroxy acid capped dipeptide HCV inhibitors, see:
Nizi, E.; Koch, U.; Ponzi, S.; Matassa, V. G.; Gardelli, C.
Bioorg. Med. Chem. Lett. 2002, 12, 3325.
15. Final compounds have been characterized by 1H NMR,
MS and RP-HPLC.
16. Carpino, L. A.; Mansour, E.-S. M. E.; Sadat-Aalaee, D. J.
Org. Chem. 1991, 56, 2611.
5. Colarusso, S.; Gerlach, B.; Koch, U.; Muraglia, E.; Conte,