Potent inhibitors of the hepatitis C virus NS3 protease: Design and synthesis of macrocyclic substrate-based β-strand mimics

Article abstract of DOI:10.1021/jo049288r

The virally encoded NS3 protease is essential to the life cycle of the hepatitis C virus (HCV), an important human pathogen causing chronic hepatitis, cirrhosis of the liver, and hepatocellular carcinoma. The design and synthesis of 15-membered ring β-strand mimics which are capable of inhibiting the interactions between the HCV NS3 protease enzyme and its polyprotein substrate will be described. The binding interactions between a macrocyclic ligand and the enzyme were explored by NMR and molecular dynamics, and a model of the ligand/enzyme complex was developed.

Full text of DOI:10.1021/jo049288r

P oten t In h ibitor s of th e Hep a titis C Vir u s NS3 P r otea se: Design
a n d Syn th esis of Ma cr ocyclic Su bstr a te-Ba sed â-Str a n d Mim ics
Nathalie Goudreau, Christian Brochu, Dale R. Cameron,† J ean-Simon Duceppe,
Anne-Marie Faucher, J ean-Marie Ferland, Chantal Grand-Maˆıtre, Martin Poirier,
Bruno Simoneau, and Youla S. Tsantrizos*
Departments of Chemistry, Boehringer Ingelheim (Canada) Ltd., Research and Development,
2100 Cunard Street, Laval (Quebec), Canada H7S 2G5
ytsantrizos@lav.boehringer-ingelheim.com
Received April 28, 2004
The virally encoded NS3 protease is essential to the life cycle of the hepatitis C virus (HCV), an
important human pathogen causing chronic hepatitis, cirrhosis of the liver, and hepatocellular
carcinoma. The design and synthesis of 15-membered ring â-strand mimics which are capable of
inhibiting the interactions between the HCV NS3 protease enzyme and its polyprotein substrate
will be described. The binding interactions between a macrocyclic ligand and the enzyme were
explored by NMR and molecular dynamics, and a model of the ligand/enzyme complex was
developed.
In tr od u ction
polyprotein by host enzymes, whereas the nonstructural
(NS) proteins are processed by two virally encoded
proteases, the NS2/NS3 and the NS3 proteases.4 The NS3
protein is a multifunctional enzyme which harbors pro-
tease activity at its N-terminal domain and RNA helicase/
ATPase activity at its C-terminal domain. The NS3
protease domain is responsible for cleavage at four sites
along the HCV polyprotein nonstructural region (a >2000
amino acid polyprotein substrate) and is essential for the
release of most HCV nonstructural proteins, including
the RNA-dependent RNA polymerase enzyme (NS5B).
Therefore, the NS3 protease is essential for viral replica-
tion in vivo, and consequently, it is an important target
for drug discovery efforts.4-6
Small molecules which can attenuate the function of
a biological process by mimicking the secondary structure
of proteins, or the critical molecular recognition elements
(structural hot spots) between polypeptides and proteins,
are of enormous scientific interest.1 Such interactions
play a key role in numerous cellular functions (e.g.,
protein glycosylation,2 intracellular recognition,3 pro-
teolysis, and others) that can be of particular relevance
in biological targets associated with important disease
areas, including viral infections. However, the key chal-
lenge in designing these small molecular mimics is that,
unlike the structurally unique peptides found at the
contact points between two large biomolecules, short
peptides are conformationally heterogeneous and exhibit
low affinity for their intended binding sites. Nonetheless,
small peptidic ligands can provide valuable research tools
in structure-based drug design and important leads in
drug discovery efforts that target therapeutically relevant
protein-protein or protein-polypeptide interactions.
In our quest for the discovery of a small molecule which
could block replication of the hepatitis C virus (HCV),
the critical interactions between the virally encoded NS3
serine protease and its polyprotein substrate(s) were
investigated. The HCV single-stranded RNA genome
encodes a polyprotein of approximately 3000 amino acid
residues, which must be proteolytically processed into at
least four structural and six nonstructural mature viral
proteins. The structural proteins are cleaved from the
Some of the recognition elements dictating substrate
specificity and the catalytic efficiency of the HCV NS3
serine protease were initially probed using a number of
synthetic dodecamer substrates and hexapeptide ligands
which resembled the N-terminal proteolysis products
formed upon cleavage of the natural HCV polyprotein.7,8
Hexapeptide ligands, such as compound 1, were found
to be competitive inhibitors of the enzyme and provided
the first lead structures for the design of peptidomimetic
inhibitors of the HCV NS3 protease.9-11 In addition, such
peptides served as tools in further exploring the inter-
actions between substrate-based ligands and the NS3
(4) (a) De Francesco, R.; Tomei, L.; Altamura, S.; Summa, V.;
Migliaccio, G. Antiviral Res. 2003, 58, 1. (b) Tan, S.-L.; Pause, A.; Shi,
Y.; Sonenberg, N. Nature Rev./ Drug Discovery 2002, 1, 867. (c) Reed,
K. E.; Rice, C. M. Curr. Top. Microbiol. Immunol. 2000, 242, 55.
(5) Kolykhalov, A. A.; Mihalik, K.; Feinstone, S. M.; Rice, C. M. J .
Virol. 2000, 74, 2046.
(6) Foy, E.; Li, K.; Wang, C.; Sumpter, R., J r.; Ikeda, M.; Lemon, S.
M.; Gale, M., J r. Science 2003, 300, 1145.
(7) Llina`s-Brunet, M.; Bailey, M.; Fazal, G.; Goulet, S.; Halmos, T.;
LaPlante, S.; Maurice, R.; Poirier, M.; Poupart, M.-A.; Thibeault, D.;
Wernic, D.; Lamarre, D. Bioorg. Med. Chem. Lett. 1998, 8, 1713.
(8) Steinku¨hler, C.; Biasiol, G.; Brunetti, M.; Urbani, A.; Koch, U.;
Cortese, R.; Pessi, A.; De Francesco, R. Biochemistry 1998, 37, 8899.
* To whom correspondence should be addressed.
† Present address: Micrologix Biotech, Inc., BC Research Complex,
3650 Wesbrook Mall, Vancouver BC V6S 2L2, Canada.
(1) (a) Cochran, A. G. Curr. Opin. Chem. Biol. 2001, 5, 654. (b)
Peczuh, M. W.; Hamilton, A. D. Chem. Rev. 2000, 100, 2479. (c)
Cochran, A. G. Chem. Biol. 2000, 7, R85.
(2) Weerapana E.; Imperiali, B. Org. Biomol. Chem. 2003, 1, 93.
(3) Loregian, A.; Marsden, H. S.; Palu`, G. Rev Med. Virol. 2002, 12,
239.
10.1021/jo049288r CCC: $27.50 © 2004 American Chemical Society
Published on Web 08/14/2004
J . Org. Chem. 2004, 69, 6185-6201
6185

Goudreau et al.
protease by NMR,12-15 molecular modeling, and crystal-
lography.16 However, the inherent conformational flex-
ibility of linear peptides, in addition to the solvent-
exposed and shallow nature of the NS3 protease active
site, made optimization of these structures into a druglike
compound an extremely challenging endeavor.
Recently, we reported on the design and synthesis of
macrocyclic inhibitors of the NS3 protease which are
orally absorbed and inhibit replication of HCV RNA in a
cell-based replicon assay.17 We demonstrated that these
compounds possess many of the desirable properties of
a druglike archetype, thus bridging the gap between the
inhibitors of HCV NS3 protease that have been reported
so far,7,9-11,16,18 and a clinically useful antiviral agent for
the treatment of hepatitis C in humans.17,19 In this paper,
the design and synthesis of the initial 15-membered ring
peptidomimetic scaffold, which led to the discovery of the
clinical antiviral agent BILN 2061 (2),19 will be discussed.
Some of the critical structural studies which guided the
design of the â-strand scaffold, mimicking the NS3-bound
conformation of the N-terminal cleavage products of the
HCV polyprotein substrate(s), will be presented. The
synthesis and preliminary evaluation of the structure-
activity relationship (SAR) of these inhibitors will also
be described. The results of this investigation illustrate
that even in the absence of an enzyme-ligand cocrystal
structure, it is possible to obtain valuable structural data
that can successfully guide a rational drug design pro-
gram.20
(9) Llina`s-Brunet, M.; Bailey, M.; De´ziel, R.; Fazal, G.; 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. 1998, 8, 2719.
(10) P. Ingallinella, P.; Altamura, S.; Bianchi, E.; Taliani, M.;
Ingenito, R.; Cortese, R.; De Francesco, R.; Steinku¨hler, C.; Pessi, A.
Biochemistry 1998, 37, 8906.
(11) Llina`s-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.
(12) LaPlante, S. R.; Cameron, D. R.; Aubry, N.; Lefebvre, S.; Kukolj,
G.; Maurice, R.; Thibeault, D.; Lamarre, D.; Llina`s-Brunet, M. J . Biol.
Chem. 1999, 274, 18618.
(13) Cicero, D. O.; Barbato, G.; Koch, U.; Ingallinella, P.; Bianchi,
E.; Nardi, M. C.; Steinku¨hler, C.; Cortese, R.; Matassa, V.; De
Francesco, R.; Pessi, A.; Bazzo, R. J . Mol. Biol. 1999, 289, 385.
(14) Barbato, G.; Cicero, D. O.; Cordier, F.; Narjes, F.; Gerlach, B.;
Sambucini, S.; Grzesiek, S.; Matassa, V. G.; De Francesco, R.; Bazzo,
R. EMBO J . 2000, 19, 1195.
(15) LaPlante, S. R.; Aubry, N.; Bonneau, P. R.; Kukolj, G.; Lamarre,
D.; Lefebvre, S.; Li, H.; Llina`s-Brunet, M.; Plouffe, C.; Cameron, D. R.
Bioorg. Med. Chem. Lett. 2000, 10, 2271.
(16) 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.
(17) (a) Tsantrizos, Y. S.; Bolger, G.; Bonneau, P.; Cameron, D. R.;
Goudreau, N.; Kukolj, G.; LaPlante, S. R.; Llina`s-Brunet, M.; Nar, H.;
Lamarre, D. Angew. Chem., Int. Ed. 2003, 42, 1356. (b) Tsantrizos, Y.
S.; Cameron, D. R.; Faucher, A.-M.; Ghiro, E.; Goudreau, N.; Halmos,
T.; Llina`s-Brunet, M. PTC International Application [Boehringer
Ingelheim (Canada) Ltd.] WO 00/59929; October 2000.
(18) (a) Perni, R. B.; Pitlik, J .; Britt, S. D.; Court, J . J .; Courtney,
L. F.; Deininger, D. D.; Farmer, L. J .; Gates, C. A.; Harbeson, S. L.;
Levin, R. B.; Lin, C.; Lin, K.; Moon, Y.-C.; Luong, Y.-P.; O′Malley, E.
T.; Rao, B. G.; Thomson, J . A.; Tung, R. D.; Van Drie, J . H.; Wei, Y.
Biorg. Med. Chem. Lett 2004, 14, 1441. (b) Lamar, J .; Victor, F.; Snyder,
N.; J ohnson, R. B.; Wang, Q. M.; Glass, J . I.; Chen, S.-H. Biorg. Med.
Chem. Lett 2004, 14, 263. (c) Victor, F.; Lamar, J .; Snyder, N.; Yip, Y.;
Guo, D.; Yumibe, N.; J ohnson, R. B.; Wang, Q. M.; Glass, J . I.; Chen,
S.-H. Biorg. Med. Chem. Lett 2004, 14, 257. (d) Yip, Y.; Victor, F.;
Lamar, J .; J ohnson, R.; Wang, Q. M.; Barket, D.; Glass, J .; J in, L.;
Liu, L.; Venable, D.; Wakulchik, M.; Xie, C.; Heinz, B.; Villarreal, E.;
Colacino, J .; Yumibe, N.; Tebbe, M.; Munroe, J .; Chen, S.-H. Biorg.
Med. Chem. Lett. 2004, 14, 251. (e) J ohansson, A.; Poliakov, A.;
Akerblom, E.; Wiklund, K.; Lindeberg, G.; Winiwarter, S.; Danielson,
U. H.; Samuelsson, B.; Hallberg, A. Bioorg. Med. Chem. 2003, 11, 2551.
(f) Hegde, V. R.; Pu, H.; Patel, M.; Das, P. R.; Butkiewicz, N.; Arreaza,
G.; Gullo, V. P.; Chan, T.-M. Biorg. Med. Chem. Lett. 2003, 13, 2925.
(g) Orvieto, F.; Koch, U. Matassa, V. G.; Muraglia, E. Biorg. Med.
Chem. Lett. 2003, 13, 2745. (h) Colarusso, S.; Koch, U.; Gerlach, B.;
Steinku¨hler, C.; De Francesco, R.; Altamura, S.; Matassa, V. G.; Narjes,
F. J . Med. Chem. 2003, 46, 345. (i) Orvieto, F.; Koch, U.; Matassa, V.
G.; Muraglia, E. Bioorg. Med. Chem. Lett. 2003, 13, 2745. (j) Andrews,
D. M.; Chaignot, H.; Coomber, B. A.; Good, A. C.; Hind, S. L.; J ohnson,
M. R.; J ones, P. S.; Mills, G.; Robinson, J . E.; Skarzynski, T.; Slater,
M. J .; Somers, D. O’N. Org. Lett 2002, 4, 4479. (k) Narjes, F.; Koehler,
K. F.; Koch, U.; Gerlach, B.; Colarusso, S.; Steinku¨hler, C.; Brunetti,
M.; Altamura, S.; De Francesco, R.; Matassa, V. G. Bioorg. Med. Chem.
Lett. 2002, 12, 701. (l) Colarusso, S.; Gerlach, B.; Koch, U.; Muraglia,
E.; Conte, I.; Stansfield, I.;; Matassa, V. G.; Narjes, F. Bioorg. Med.
Chem. Lett. 2002, 12, 705. (m) Han, W.; Hu, Z.; J iang X.; Decicco, C.
P. Bioorg. Med. Chem. Lett. 2000, 10, 711. (n) Dunsdon, R. M.;
Greening, J . R.; J ones, P. S.; J ordan, S.; Wilson, F. X. Bioorg. Med.
Chem. Lett. 2000, 10, 1577. (o) Marchetti, A.; Ontoria, J . M.; Matassa,
V. G. Synlett 1999, S1, 1000.
Resu lts a n d Discu ssion
Design of Su bstr a te-Ba sed , â-Str a n d Mim ics a s
P oten t In h ibitor s of th e HCV NS3 P r otea se. During
the past decade, considerable effort has been devoted to
the design and synthesis of small molecules which can
mimic the R-helical,21 â-turn,22 or â-strand23 secondary
structures of proteins. Many of the scaffolds that have
been designed are undoubtedly inspired by nature’s
nonribosomally formed secondary metabolites which are
(19) (a) Llina`s-Brunet, M.; Bailey, M. D.; Bolger, G.; Brochu, C.;
Faucher, A.-M.; Ferland, J . M.; Garneau, M.; Ghiro, E.; Gorys, V.;
Grand-Maitre, C.; Halmos, T.; Lapeyre-Paquette, N.; Liard, F.; Poirier,
M.; Rhe´aume, M.; Tsantrizos, Y. S.; Lamarre, D. J . Med. Chem. 2004,
47, 1605. (b) Lamarre, D.; Anderson, P. C.; Bailey, M.; Beaulieu, P.;
Bolger, G.; Bonneau, P.; Bo¨s, M.; Cameron, D.; Cartier, M.; Cordingley,
M. G.; Faucher, A.-M.; Goudreau, N.; Kawai, S. H.; Kukolj, G.; Lagace´,
L.; LaPlante, S. R.; Narjes, H.; Poupart, M.-A.; Rancourt, J .; Sentjens,
R. E.; St George, R.; Simoneau, B.; Steinmann, G.; Thibeault, D.;
Tsantrizos, Y. S.; Weldon, S. M.; Yong, C.-L.; Llina`s-Brunet, M. Nature
2003, 426, 186. (c) Rice, C. M. Nature 2003, 426, 129.
(20) Davis, A. M.; Teague, S. J .; Kleywegt, G. J . Angew. Chem., Int.
Ed. 2003, 42, 2718.
(21) (a) Kutzki, O.; Park, H. S.; Ernst, J . T.; Orner B. P.; Yin, H.;
Hamilton, A. D. J . Am. Chem. Soc. 2002, 124, 11838. (b) Banerjee, R.;
Basu, G.; Che`ne, P.; Roy, S. J . Peptide Res. 2002, 60, 88.
(22) (a) Lucke, A. J .; Tyndall, J . D. A.; Singh, Y.; Fairlie, D. P. J .
Mol. Graph. Mod. 2003, 21, 341. (b) Hanessian, S.; McNaughton-Smith,
G.; Lombart, H.-G.; Lubell, W. D. Tetrahedron 1997, 53, 12789.
6186 J . Org. Chem., Vol. 69, No. 19, 2004

Potent Inhibitors of the Hepatitis C Virus NS3 Protease
often composed of unusual amino acid fragments and are
constrained into cyclic peptides with unique bioactive
conformations. Introduction of nature’s preorganization
elements to synthetic scaffolds has provided a wealth of
peptidomimetic compounds with diverse structural, physi-
cal, and conformational properties that have been ex-
ploited in supramolecular and combinatorial chemistry24
as well as chemical biology and medicinal chemistry.1,25
On the basis of this knowledge, macrocyclic peptides
which adopt an extented â-strand conformation have also
been explored as substrate mimics of metallo, aspartic,
cysteine, and serine proteases;23,25 the HCV NS3 protease
is an example of the latter class of enzymes. However,
the HCV NS3 enzyme is fairly unique among serine
proteases in that it is activated by its structure-modifying
cofactor NS4A (a 54-residue peptide).26 The interactions
between the NS4A cofactor and the NS3 protease are
known to induce conformational changes that signifi-
cantly reduce (but not entirely eliminate) the plasticity
of the enzyme, thus enhancing its catalytic power.
Induced-fit conformational changes of the enzyme upon
binding of a product-based ligand have also been re-
ported.27
nocyclopropylcarboxylic acid (ACCA) moiety at P1,11,37a
confirmed that the backbone of shorter peptides also
adopts the extended â-strand conformation upon binding
to the NS3.15 In addition, the structural differences
between the free (in solution) and the NS3-bound state
of ligand 3 were explored by both NMR and computa-
tional studies.
The conformation of the NS3-bound tripeptide 3 was
found to be distinctly different from that of its free state.
Conformational analysis of an ionized and solvated model
compound of the P1 residue, such as compound 4, by
semiempirical quantum mechanics calculations identified
two low energy conformations for the P1 ethyl ACCA
moiety, 4a and 4b (φ of approximately +90° and -90°,
respectively), and calculated an energy difference of ∼1.1
kcal/mol in favor of 4a (with an estimated ratio of ∼4:1).
The ROESY NMR spectrum of ligand 3 (Figure 1a)
confirmed that in the free state the predominant confor-
mation of its P1 moiety corresponded to that of 4a , in
agreement with the computational studies. A strong NOE
cross-peak between the P1 amide NH and its H1 proton
and a weaker NOE between the NH and the H2 (pro-R)
proton were clearly observed, whereas NOE interaction
between the NH and the H3 (pro-S) proton could not be
detected. The combined NOE data was consistent with
a preferred φ angle of approximately +90° for the P1
residue of the free ligand 3, as shown in Figure 1a.
In the course of our initial investigations, NMR studies
of the NS3-bound hexapeptide 1 revealed that this ligand
binds in an extended, â-strand conformation and its
interactions with the enzyme involve predominately the
P1-P3 residues.12,28-30 Additional NMR studies with an
analogue of tripeptide 3, having a more optimized 1-ami-
(23) (a) Bartlett, P. A.; Rezac, M.; Olson, S.; Phillips, S. PTC
International Application (The Regents of the University of California)
WO 02/099045; December 2002. (b) Dumez, E.; Snaith, J ., S.; J ackson,
R., F., W.; McElroy, A. B.; Overington, J .; Wythes, M. J .; Withka, J .
M.; McLellan, T., J . J . Org. Chem. 2002, 67, 4882. (c) Reid, R. C.; Kelso,
M. J .; Scanion, M. J .; Fairlie, D. P. J . Am. Chem. Soc. 2002, 124, 5673.
(d) Glenn, M. P.; Pattenden, L. K.; Reid, R. C.; Tyssen, D. P.; Tyndall,
J . D. A.; Birch, C. J .; Fairlie, D. P. J . Med. Chem. 2002, 45, 371. (e)
Nowick, J . S.; Smith, E. M.; Ziller, J . W.; Shaka, A. J . Tetrahedron
2002, 58, 727. (f) Ripka, A., S.; Satyshur, K. A.; Bohacek, R. S.; Rich,
D. H. Org. Lett. 2001, 3, 2309. (g) Mak, C. C.; Le, V.-D.; Lin, Y.-C.;
Elder, J . H.; Wong, C.-H. Bioorg. Med. Chem. Lett. 2001, 11, 219. (h)
Tyndall, J . D. A.; Fairlie, D. P. Curr. Med. Chem. 2001, 8, 893. (i) Tian,
Z.-Q.; Brown, B. B.; Mack, D. P.; Hutton, C. A.; Bartlett, P. A. J . Org.
Chem. 1997, 62, 514. (j) Nowick, J . S.; Pairish, M.; Lee, I. Q.; Holmes,
D. L.; Ziller, J . W. J . Am. Chem. Soc. 1997, 119, 5413. (k) Boumendjel,
A.; Roberts, J . C.; Hu, E.; Pallai, P. V. J . Org. Chem. 1996, 61, 4434.
(l) Smith, A. B.; Hirschmann, R.; Pasternak, A.; Guzman, M. C.;
Yokoyama, A.; Sprengeler, P. A.; Darke, P. L.; Emini, E. A.; Schleif,
W. A. J . Am. Chem. Soc. 1995, 117, 11113. (m) Martin, S. F.; Oalmann,
C. J .; Linas, S. Tetrahedron 1993, 49, 3532.
(24) (a) Glenn, M. P.; Kelso, M. J .; Tyndall, J . D. A.; Fairlie, D. P.
J . Am. Chem. Soc. 2003, 125, 640. (b) Park, C.; Burgess, K. J . Comb.
Chem. 2001, 3, 257. (c) Singh, Y.; Sokolenko, N.; Kelso, M. J .; Gahan,
L. R.; Abbenante, G.; Fairlie, D. P. J . Am. Chem. Soc. 2001, 123, 333.
(d) MacDonald, M.; Aube´, J . Curr. Org. Chem. 2001, 5, 417. (e) Wipf,
P.; Miller, C. P.; Grant, C. M. Tetrahedron 2000, 56, 9143. (f) Mink,
D.; Mecozzi, S.; Rebek, J . Tetrahedron Lett. 1998, 39, 5709.
(25) (a) Reid, R. C.; Pattenden, L. K.; Tyndall, J . D. A.; Martin, J .
L.; Walsh, T.; Fairlie, D. P. J . Med. Chem. 2004, 47, 1641. (b) Fairlie,
D. P.; West, M. L.; Wong, A. K. Curr. Med. Chem. 1998, 5, 29. (c)
Fairlie, D. P.; Abbenante, G.; March, D. R. Curr. Med. Chem. 1995, 2,
654.
In contrast, transferred NOESY (TRNOESY)31 NMR
experiments using compound 3 in the presence of the
(31) Review: Ni, F. Prog. Nucl. Magn. Reson. Spectrosc. 1994, 26,
517.
(32) Yao, N.; Reichet, P.; Taremi, S. S.; Prosise, W. W.; Weber, P.
C. Structure 1999, 7, 1353.
(33) Tong, L.; Wengler, G.; Rossmann, M. G. J . Mol. Biol. 1993, 230,
228.
(34) J ames, M. N. G.; Sielecki, A. R.; Brayer, G. D.; Delbaere, L. T.
J . J . Mol. Biol. 1980, 144, 43.
(35) Khan, A. R.; Parrish, J . C.; Fraser, M. E.; Smith, W. W.;
Bartlett, P. A.; J ames, M. N. G. Biochemistry 1998, 37, 16839.
(36) (a) Burgess, K.; Ke, C.-Y. Synthesis 1996, 1463. (b) Burgess,
K.; Li W. Tetrahedron Lett. 1995, 36, 2725. (c) Burgess, K.; Ho, K.-K.
J . Org. Chem. 1992, 57, 5931.
(37) (a) Rancourt, J .; Cameron, D. R.; Gorys, V.; Lamarre, D.; Poirier,
M.; Thibeault, D.; Llina`s-Brunet, M. J . Med. Chem. 2004, 47, 2511.
(b) Bailey, M. D. United States Patent [Boehringer Ingelheim (Canada)
Ltd.] US 6,268,207 B1; J uly 31, 2001.
(26) Bartenschlager, R. J . Viral Hepat. 1999, 6, 165.
(27) Bianchi, E.; Orru`, S.; Piaz, F. D.; Ingenito, R.; Casbarra, A.;
Biasiol, G. Koch, U.; Pucci, P.; Pessi, A. Biochemistry 1999, 38, 13844.
(28) Protease subsite nonmenclature: Schechter, I.; Berger, A.
Biochem. Biophys. Res. Commun. 1967, 27, 157.
(29) LaPlante, S. R.; Aubry, N.; De´ziel, R.; Ni, F.; Xu, P. J . Am.
Chem. Soc. 2000, 122, 12530.
(30) The role of the P5 and P6 residues was also examined by the
IRBM/Merck group using pre-steady-state kinetics. Based on their
data, they proposed that the electrostatic surface potential of these
residues enhances the collision rates between the peptidic ligand and
the active site of NS3 protease; U. Koch, U.; Biasiol, G.; Brunetti, M.;
Fattori, D.; Pallaoro, M.; Steinku¨hler, C. Biochemistry 2001, 40, 631.
J . Org. Chem, Vol. 69, No. 19, 2004 6187

Goudreau et al.
F IGURE 1. Selected region of the NMR spectra of tripeptide 3 (P1 moiety): (a) ROESY NMR spectrum of 3 alone in aqueous
buffer; (b) transferred NOESY NMR spectrum of 3 in the presence of the HCV NS3 protease domain (the relative intensities of
NOE interactions are indicated as S (strong), M (medium), and W (weak)); (c) model of the P1 moiety of tripeptide 3 docked in the
active site of the apo form of NS3 protease. RModel compound 4: colors indicate H1 (purple), H2 (green), H3 (blue).
NS3 protease domain provided unambiguous assignment
of the bound conformation of the ethyl ACCA moiety
(Figure 1b). NOE interactions were observed between the
amide NH and all three â-protons (Figure 1b); strong
NOE with H2 (pro-R) and weaker NOEs with H1 and H3
(pro-S). The relative intensities of these NOEs suggested
a bound conformation for the P1 moiety with a φ angle
of approximately -90° (Figure 1b). Docking of the NMR-
derived NS3-bound conformation of 3 into the active site
of the protease was subsequently attempted, using as
guidance the previously published crystal structures of
the apo HCV NS3 protease32 and the serine proteases of
the Sindbis virus33 and Streptomyces griseus.34 This
higher energy conformer of 3 (analogous to 4b, Figure
1b) was consistent with the expected NS3-bound struc-
ture (based on the literature),31-33 allowing the carboxy-
late anion of P1 to interact with the oxyanion hole of the
active site and possibly the side chains of the catalytic
residues H57 and S139 (Figure 1c). Consequently, we
concluded that the NH-CR bond of the free ligand 3 had
to undergo a rotation of approximately 180° in order to
adopt its NS3-bound conformation (Figure 1a vs 1b). The
entropic penalty associated with this realignment of the
ligand would be expected to have a negative impact on
its binding energy.
would adopt the desired â-strand NS3-bound conforma-
tion. This rigid macrocyclic ligand was expected to pay a
lower entropic penalty for binding to the protease and,
consequently, have a higher affinity for the enzyme than
its corresponding acyclic precursor.17,35 Furthermore, our
NMR data of linear peptides indicated that in the NS3-
bound conformation the P3 side chain is mostly solvent-
exposed.12,15 In contrast, the hydrophobic linker moiety
of the macrocyclic scaffold was expected to be properly
aligned into the S1-S3 binding pocket so as to favorably
interact with the protease. Thus, we predicted that the
combined effects of a properly pre-organized rigid scaf-
fold, which closely mimics the enzyme-bound conforma-
tion, and a hydrophobic chain that can participate in
additional interactions with the active site, would lead
to a novel class of very potent peptidomimetic inhibitors
of the HCV NS3 protease.
Syn th esis of Bu ild in g Block s a n d In h ibitor s. To
evaluate the biological and conformational properties of
â-strand mimics, a variety of P1, P2, and P3 building
blocks were synthesized (Schemes 1-3). The allylglycine
methyl esters 5a and 5b were prepared by treating the
corresponding amino acids with anhydrous methanol in
the presence of thionyl chloride (Scheme 1a). The cysteine
derivatives 6a and 6b were prepared by first reacting
the corresponding thiols with allyl bromide in the pres-
ence of base, followed by protecting the amines as the
tert-butyl carbamates and then treating with di-
azomethane. After removal of the Boc protecting group
under standard acidic conditions, the acyclic P1 frag-
ments 5a ,b and 6a ,b were used directly in coupling
reactions with one of the P2 building blocks.
Therefore, we embarked on the design and synthesis
of a rigid scaffold which could restrain the P1 φ angle to
the higher energy conformation (i.e., 4b, Figure 1b) and
simultaneously preorganize the P1-P3 amide backbone
exclusively to the all-trans geometry (â-strand);17 the
latter is a significant problem with proline-containing
linear peptides which exist as mixtures of cis and trans
rotamers. We predicted that covalent linking of the P1
to the P3 side chain, creating a 14- to 16-membered ring
structure, could result in a scaffold which in the free state
Both the vinyl (9) and homoallyl ACCA (10) fragments
were synthesized using a slightly modified procedure
from that originally reported by Burgess and co-workers
6188 J . Org. Chem., Vol. 69, No. 19, 2004

Potent Inhibitors of the Hepatitis C Virus NS3 Protease
SCHEME 1. Syn th esis of P 1 F r a gm en tsa
whereas preparation of 15c required the synthesis of the
precursor 4-hydroxy-7-methoxy-2-phenylquinoline frag-
ment (13c).43 Preparation of quinoline 13c was previously
reported as a mixture with its 5-methoxy isomer 14 (3:2
ratio).43b Using a modified protocol,43a we observed that
the reaction of m-anisidine with ethyl benzoyl acetate in
the presence of either HCl or p-TsOH (2 mol %) gave a
mixture of the enamines 11 and 12 in an approximately
1.2:1 ratio (Scheme 2). Pyrolysis of each purified enamine,
11 and 12, gave the desired quinoline 13c in 44% and
12% yield, respectively. However, pyrolysis of the crude
mixture of enamines 11 and 12, followed by a simple
trituration of the products in CH2Cl2, allowed for the
isolation of the pure 7-methoxy isomer 13c (without any
trace of the 5-methoxy isomer 14) in 17% overall yield.
This modified protocol is particularly amenable to large
scale synthesis (>100 g) of the key intermediate 13c.
The two P3 fragments, (2S)-N-Boc-amino 6-heptenoic
and (2S)-N-Boc-amino 8-nonenoic acids (20a and 20b,
respectively), were prepared as shown in Scheme 3. The
synthesis of 20a was initiated from the commercially
available acid 16a , whereas the 8-nonenoic acid (16b) was
obtained after a Grignard reaction of the 8-bromo-1-
octene (16c) with CO2. Enolization of intermediates 17a
and 17b with potassium hexamethyldisilazide was fol-
lowed by asymmetric azidation at the CR position with
2,4,6-triisopropylbenzenesulfonyl azide (trisyl azide), as
previously reported by Evans.44 Reduction of the azide
intermediates 18a and 18b with SnCl2 in methanol gave
the corresponding free amines (Scheme 3). These prod-
ucts were then protected as the tert-butyl carbamates
(Boc) to give intermediates 19a and 19b in good overall
yields. The final P3 building blocks 20a and 20b were
obtained in good yields after cleavage of the chiral
auxiliary with LiOOH (Scheme 3).
a
Reagents and conditions: (a) synthesis of allylglycine methyl
ester P1 analogues (5a , 5b) (i) MeOH/SOCl2; (b) synthesis of
cysteine analogues (6a and 6b) (ii) allyl bromide, NH4OH, (iii)
Boc2O, K2CO3, (iv) CH2N2, (v) 4N HCl in dioxane; (c) synthesis of
the (1R,2R)/(1S,2S)-1-amino-2-homoallylcyclopropyl derivative 8
(vi) BnNEt3Cl, 50% NaOHaq, di-tert-butyl malonate, 23 °C (38%),
(vii) t-BuOK, H2O, Et2O, 0-23 °C (85%), (viii) (PhO)2P(O)N3, Et3N,
C6H6, reflux, then (CH3)3Si(CH2)2OH, C6H6, reflux (88%).
for the preparation of 2,3-methanomethionine.36 Prepara-
tion of the vinyl ACCA moiety (1R,2S)-9 was achieved
in high enantiomeric purity (>98% ee) as recently
reported.37 On the basis of a similar approach (Scheme
1c), the synthesis of the racemic homoallyl ACCA frag-
ment 10 was initiated by the dialkylation of di-tert-
butylmalonate with 1,2-dibromo-5-hexene,38 followed by
selective hydrolysis of the less hindered ester,39 to give
the monoester intermediate 7b. The racemic mixture of
8 (having the allyl side chain syn to the tert-butyl ester)
was subsequently prepared via Curtius rearrangement
which was followed by trapping the products as the
2-(trimethylsilyl)ethyl carbamate derivatives.40 Com-
pound 8 was used directly in the synthesis of diastere-
omeric P1-P2 fragments (Scheme 4), which were easily
separated by flash column chromatography; the absolute
configuration of the (1R, 2R)-10b building block was
eventually confirmed.41
For the preparation of each macrocyclic ligand, the
appropriate P1, P2 and P3 building blocks were first
assembled into an acyclic tripeptide, using standard
solution-phase peptide chemistry; as an example, the
preparation of the P1-P2 dipeptide 25a and its conver-
(41) After separation of each P1-P2 diastereomeric fragment by
chromatography, each dipeptide was carried (independently) to the
corresponding acyclic tripeptide diene and macrocyclic inhibitor.
Biological evaluation of the final macrocyclic inhibitors indicated that
those derived from the P1 fragment 10a were significantly less potent
than those derived from 10b, consistent with our previous observation
using the enantiomers of vinyl ACCA 9.17 For example, in an in vitro
enzymatic assay, the macrocyclic compound 26 (derived from dipeptide
24a ) was found to be 40-fold less potent than the corresponding
compound 33b, which was derived from dipeptide 23a . Furthermore,
it was determined that the less polar diastereomer in each dipeptide
mixture (i.e., 23a /24a or 23b/24b) corresponded to analogues derived
from the homoallyl ACCA (1R,2R)-10b. Assignment of the absolute
configuration of 10b was based on analytical and NMR data, as well
as the in vitro enzymatic potency the saturated compound 35 (com-
pound 35 is the common product formed from catalytic hydrogenations
of either 34a , 34b, or 36b).
The P2 fragments 15d -f were prepared from the
commercially available Boc-protected cis-hydroxyproline
methyl ester and each of the three 4-hydroxyquinoline
analogues via a Mitsunobu reaction,42 followed by sa-
ponification of the methyl esters under basic conditions
(Scheme 2). Compounds 15a and 15b were prepared from
the commercially available quinolines 13a and 13b ,
(42) Poupart, M.-A.; Cameron, D. R.; Chabot, C.; Ghiro, E.; Gou-
dreau, N.; Goulet, S.; Poirier, M.; Tsantrizos, Y. S. J . Org. Chem. 2001,
66, 4743.
(43) (a) Venturella, P.; Bellino, A. J . Heterocycl. Chem. 1975, 12,
669. (b) Giardina, G. A. M.; Sarau, H. M.; Farina, C.; Medhurst, A. D.;
Grugni, M.; Raveglia, L. F.; Schmidt, D. B.; Rigolio, R.; Luttmann, M.;
Vecchietti, V.; Hay, D. W. P. J . Med. Chem. 1997, 40, 1794.
(44) (a) Evans, D. A.; Evrard, D. A.; Rychnovsky, S. D.; Fru¨h, T.;
Whittingham, W. G.; DeVries, K. M. Tetrahedron Lett. 1992, 33, 1189.
(b) The % ee of 20a and 20b was not determined at this stage, however,
upon coupling of these compounds with the P1-P2 dipeptides only one
product could be observed by HPLC and NMR, confirming the high
enantiomeric purity of the P3 fragments.
(38) Baldwin, J . E.; Adlington, R. M.; Rawlings, B. J . Tetrahedron
Lett. 1985, 26, 481.
(39) Gassman, P. G.; Schenk, W. N. J . Org. Chem. 1977, 42, 918.
(40) Capson, T. L.; Poulter, C. D. Tetrahedron Lett. 1984, 25, 3515.
J . Org. Chem, Vol. 69, No. 19, 2004 6189

Goudreau et al.
SCHEME 2. Syn th esis of th e P 2 F r a gm en ts 15a
a
Reagents and conditions: (i) N-Boc-protected cis-4-hydroxyproline, PPh3, DIAD, THF, rt; (ii) LiOH, H2O/THF/MeOH, rt.
SCHEME 3. Syn th esis of P 3 Am in o Acid Lin k er s 20a a n d 20ba
a
Reagents and conditions: (i) Mg, CO2; (ii) Et3N, (CH3)CCOCl, THF, -78 to 0 °C; (iii) n-BuLi, 4(S)-4-(phenylmethyl)-2-oxazolidinone
lithium salt, THF, -78 °C (68%); (iv) KHMDS, THF, -78 °C; (v) trisyl azide, THF, -78 °C (67%); (vi) SnCl2, MeOH, 0 °C; (vii) Boc2O,
NaHCO3, dioxane, H2O (60%); (viii) H2O2, LiOH, THF, H2O, 0 °C (70%).
sion to the macrocyclic ligand 33b are shown in Schemes
4 and 5, respectively. Peptide coupling reactions were
typically carried out in CH2Cl2 or DMF, using HATU or
TBTU as the coupling agent and DIPEA or NMM as the
base. The crude hydrochloride salts of the P1 moieties 5,
6, and (1R,2S)-9 (obtained after hydrolysis of the Boc
protecting group with HCl in dioxane) were coupled
directly with one of the P2 fragments 15 to give the
N-Boc-protected dipeptides 21, 22, and 27, respectively,
with average yields of 75-85%. For the synthesis of
dipeptides 23 and 24, the racemic intermediate 8 was
first treated with TBAF and then coupled directly with
one of the P2 derivatives 15a or 15c (Scheme 4). In each
case, the diastereomeric mixture of the corresponding
N-Boc-protected tert-butyl esters 23a ,b and 24a ,b was
separated by flash column chromatography and then
treated with 4 N HCl in dioxane to simultaneously
remove the N-Boc protecting group and hydrolyze the
tert-butyl esters (Scheme 4).41 Each compound was then
treated with diazomethane to reprotect the carboxylic
6190 J . Org. Chem., Vol. 69, No. 19, 2004

Potent Inhibitors of the Hepatitis C Virus NS3 Protease
SCHEME 4. Syn th esis of P 1-P 2 Dip ep tid es fr om th e Ra cem ic Hom oa llyl ACCA Der iva tive 8a
a
Reagents and conditions: (i) TBAF, THF, rt; (ii) 15, HATU, DIPEA, CH2Cl2, rt; (iii) 4 N HCl dioxane; (iv) Boc2O, NaHCO3, dioxane/
H2O; (v) CH2N2, MeOH, CH2Cl2.
SCHEME 5. Syn th esis of Ma cr ocyclic In h ibitor
33ba
a
Reagents and conditions: (i) 4 N HCl-dioxane; (ii) 20a ,
HATU, NMM, CH2Cl2, rt; (iii) bis(tricyclehexylphosphine)ben-
zylidene ruthenium(IV) dichloride, CH2Cl2, reflux; (iv) LiOH, H2O,
MeOH, rt.
acid moiety as the methyl ester. The diastereomerically
pure dipeptides were then used independently for the
synthesis of macrocyclic inhibitors.41 For the purpose of
this study, only dipeptides 25a and 25b, which are
derived from the P1 moiety (1R, 2R)-10b, will be dis-
cussed in detail.41 However, the macrocyclic peptides
derived from 10a (e.g., compound 26) were also investi-
gated and found to be weak inhibitors for the HCV NS3
protease.41
Coupling of each P1-P2 dipeptide with one of the P3
linker moieties (20a or 20b) was carried out under
standard condition to give the acyclic, tripeptide dienes
28 (a ,b), 29 (a ,b), 30 (a ,c), and 31. Cyclization of each
diene under ring-closing metathesis (RCM) conditions,
catalyzed by Grubbs’ catalyst (PCy3)2Cl2RudCHPh,45 led
to the desired 15-membered ring macrocyclic peptides in
good to excellent yields (50-95% yield);46 a typical
example is shown in Scheme 5. For each compound, the
formation of the Z or E double bond, or a mixture of both,
1
was confirmed by H NMR. Diene 31 derived from the
J . Org. Chem, Vol. 69, No. 19, 2004 6191

Goudreau et al.
F IGURE 2. Structure-activity of HCV NS3 protease inhibitors: IC50 average values from a minimum of three independent in
vitro assays, using the full-length NS3-NS4A heterodimer protein of hepatitis C genotype 1b.
rigid vinyl ACCA moiety (1R, 2S)-9 formed exclusively
[>98% of the crude macrocyclic product(s)] the Z isomer
(compounds 36a ), whereas the conformationally flexible
allylglycine (28), allylcysteine (29) and homoallyl ACCA
(30) analogues gave mixtures of Z and E macrocyclic
products. The Z/E ratio depended mostly on the stereo-
chemistry and conformational rigidity of the P1 moiety
(although other factors contributed to a lesser extent46).
For example, the allylglycine derivative 28a gave a
mixture of Z/E products in a 1:1 ratio, whereas its
diastereomer 28b gave a ratio of approximately 1:14 (Z/
E, respectively). The homoallyl ACCA derivative 30c
(having the allyl side chain in the same orientation as
28a and the same P2 quinoline) gave a Z/E ratio of 1:3.
Finally, saponification of all ester derivatives (macrocyclic
products, as well as the key diene precursors required
for SAR studies) under basic conditions, followed by C18,
reversed-phase HPLC purification, led to the isolation
of all compounds in high purity. These final products
were evaluated for their ability to inhibit the HCV NS3
protease (Figure 2), using the in vitro assay described
previously.17 In addition, to evaluate the role of the
hydrocarbon linker connecting P1 to P3 and confirm the
stereochemistry of fragment 10b, the macrocyclic com-
pounds 34a , 34b, and 36b were independently reduced
under catalytic hydrogenation conditions. As expected,
all three of these starting materials led to the same
saturated product, compound 35, confirming that the
stereochemistry of the homoallyl ACCA fragment 10b
was identical to that of the vinyl ACCA (1R,2S)-9.37b,41
Finally, to fully validate the impact of the macrocyclic
scaffold, the linear tripeptides 37 and 42, which cor-
respond directly to inhibitors 36b and 41b (respectively)
with a single σ-bond disconnection, were prepared and
tested in the same enzymatic assay (Figure 2).
Str u ctu r e-Activity Rela tion sh ip (SAR). In vivo, a
P1 cysteine residue is the dominant determinant for
recognition and cleavage efficiency by the NS3 protease
at three out of the four cleavage sites along the HCV
polyprotein.26 Therefore, in vitro SAR evaluation was
initiated with the thia-macrocyclic ligand 32b containing
an S-cysteine residue at P1. Disappointingly, this ligand
(32b) was only 2-fold more potent than its acyclic
precursor 29d , whereas the latter was approximately
3-fold more potent than its unnatural epimer 29c (Figure
2).
(45) (a) Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18.
(b) Schwab, P.; Grubbs, R. H.; Ziller, J . W. J . Am. Chem. Soc. 1996,
118, 100. (c) Schwab, P.; France, M. B.; Ziller, J . W.; Grubbs, R. H.
Angew. Chem., Int. Ed. Engl. 1995, 34, 2039.
(46) Other catalysts of the RCM reaction were also evaluated; their
effects on yield and the stereochemistry of the resulting macrocyclic
product(s) will be published elsewhere.
However, consistent with previous SAR studies,11
macrocyclic derivatives having a cyclopropane moiety at
P1 were found to be significantly more potent than ligand
32b. Compound 33b, having the least optimum quinoline
6192 J . Org. Chem., Vol. 69, No. 19, 2004

Potent Inhibitors of the Hepatitis C Virus NS3 Protease
moiety X1 and an E double bond at the same position as
32b (i.e., at Cꢀ,η to the P1 acid) was 4-fold more potent
than 32b and 32-fold more potent than its acyclic
precursor 30b (Figure 2). Optimization of the P2 moiety
with the 2-phenyl-4-hydroxy-7-methoxyquinoline (P2 )
15f), as previously reported,17,42,47 led to inhibitor 34a
(also having an E double bond at Cꢀ,η to the P1 acid)
which was over 200-fold and 50-fold more potent than
32b and 33b, respectively. Interestingly, the E analogue
34a was 3-fold more potent than the Z isomer 34b and
both of these compounds (34a and 34b) were approxi-
mately 3-fold and 9-fold, respectively, less potent than
their common saturated analogue 35 (Figure 2). These
results suggested that the electronic properties (π-
character) and rigidification elements of the olefinic
moiety did not have a significant impact on the overall
stability of the NS3 protease-inhibitor complex. How-
ever, the conformation of the P1 to P3 linker moiety (as
dictated by an E vs Z π- or a σ-bond) played a critical
role in the overall beneficial effects of the macrocyclic
scaffold.
Previous SAR studies, on acyclic tetrapeptide inhibitors
of the HCV NS3 protease, revealed that substitution of
the P1 cyclopropyl moiety (ACCA) with a vinyl group
provided ∼5-fold increase in potency over the correspond-
ing ethyl ACCA derivative.37a Macrocyclic derivatives of
vinyl ACCA (1R,2S)-9, such as compound 36b, were also
found to be more potent than the corresponding ana-
logues derived from homoallyl ACCA 10b (i.e., 34a and
34b). However, the effect of the olefinic moiety at Cγ,δ
to the P1 was less significant in this class of compounds
than in the acyclic peptides (∼2-fold difference in potency
was observed between inhibitor 35 and 36b, Figure 2).
Finally, a drop in potency of approximately 36-fold was
observed upon cleavage of the linker moiety of inhibitor
36b to produce the corresponding open-chain analogue
37 (Figure 2), once again, highlighting the value of the
macrocyclic scaffold.
configuration at P1 as the natural substrate(s) of the NS3
protease, is 7-fold more potent that 39b. However,
compound 41b is almost equipotent to the corresponding
acyclic tripeptide 42, confirming that the difference in
potency between 39b and 41b is primarily due to the
stereochemistry at the CR of P1 and not due to the
macrocyclization. These results further validate the
essential role of the P1 cyclopropyl moiety and the specific
molecular requirements of the target enzyme for which
these inhibitors were designed.
En zym e-Bou n d Con for m a tion of Liga n d 33c a n d
Molecu la r Mod elin g of Its P oten tia l In ter a ction s
w ith th e Active Site of HCV NS3 P r otea se. The
significant potency differences observed between the
macrocyclic inhibitors and their linear precursors pro-
vided strong validation for the intuition which led to the
design of this novel scaffold. To gain some insight into
the NS3-bound conformation of this class of inhibitors,
transferred nuclear Overhauser effect (TRNOE) experi-
ments were used to study the structure of inhibitor 33c
in the presence and absence of the enzyme.31,48
The necessity of the cyclopropyl moiety in providing
the optimum molecular recognition and conformation
elements between this class of compounds and the HCV
NS3 protease was further confirmed by designing a 15-
membered ring compound which lacked the â′ carbon of
the saturated analogue 35. Compound 39b was found to
be ∼2000-fold less potent than 35, consistent with
previous observations with ligand 32b (Figure 2). It is
worth noting that compound 41b, having the same
In the absence of the enzyme, very few positive NOEs
could be observed in the 2D NOESY spectrum of 33c
(Supporting Information, Figure 1a), consistent with a
small molecule tumbling rapidly in solution. In contrast,
the NOESY spectrum recorded after the addition of the
NS3 protease revealed an extensive set of negative
(48) A key condition for studies on protein-ligand interactions using
TRNOE experiments is a rapid and reversible exchange between the
free and enzyme-bound state of the ligand (the exchange must be faster
than the spin-lattice relaxation rate).31 The strong binding affinity of
compound 33b for the HCV NS3 protease limited the use of this
inhibitor for the TRNOE experiments. This problem was overcome by
using the less potent acetate derivative 33c for the NMR studies.
(47) Goudreau, N.; Cameron, D. R.; Bonneau, P.; Gorys, V.; Plouffe,
C.; Poirier, M.; Lamarre, D.; Llinas-Brunet, M. J . Med. Chem. 2004,
47, 123.
J . Org. Chem, Vol. 69, No. 19, 2004 6193

Goudreau et al.
bound conformations. A stereoview of the final 10 lowest
energy TRNOE-consistent bound conformations of 33c
are superimposed in Figure 3a. The well-defined en-
semble of structures obtained for inhibitor 33c gave a
root-mean-square deviation for the peptidic backbone of
only 0.04 Å. Due to some overlapping resonances, leading
to some ambiguous TRNOE cross-peaks, the precise
bound orientation of the hydrocarbon linker moiety could
not be fully determined, however, a restricted range of
conformations was reliably proposed (Figure 3a). Con-
sistent with previous observations,12,15,47 a strong TRNOE
cross-peak was observed between the â and γ pro-R
protons of the proline ring suggesting that the quinoline
moiety adopted a pseudoaxial conformation relative to
the proline ring (Figure 3b). Interestingly, the ROESY
NMR data of several macrocyclic compounds (free in
solution) was also consistent with the conformation
shown in Figure 3, thus indicating that the predominant
conformation of the macrocyclic scaffold in the free state
was very similar to that of the enzyme-bound complex.
Finally, a model of the inhibitor/enzyme complex was
created by docking the NMR-derived structure of 33c into
the active site of the apo NS3 protease.50 This complex
was first energy-minimized and then submitted to a 1
ns molecular dynamic simulation in a water droplet. The
minimized model of the NS3/33c complex (Figure 4a)
suggested an extensive network of interactions between
the inhibitor and the active site of the protease, including
hydrogen-bond interactions between the P3 NH and CO
groups of 33c and A157,16 and between the P1 NH of 33c
and the carbonyl of residue R155.17 In addition, the
carboxylate anion was poised perfectly, allowing for
potential interactions with both the oxyanion hole of the
protease and H57, while the P1-P3 hydrocarbon linker
moiety was docked well within the S1-S3 binding pocket
and within van der Waals distance from V132. Based on
this model, the P1 cyclopropyl moiety of the ligand would
also be expected to form beneficial interactions with the
F154 residue at the floor of the S1 binding pocket (Figure
4a). The difference in the π-character between a P1 vinyl
ACCA moiety and a P1 ethyl ACCA moiety may account
for the modest potency difference observed between
analogues such as 36b and 35 (Figure 2). It is interesting
to note that although this model predated our cocrystal
structure of inhibitor 36b bound to the NS3/4Apeptide
protease domain,17 it provided valuable structural infor-
mation with remarkable accuracy. In Figure 4b, the
NMR-derived bound conformation of 33c and the crystal
structure of 36b are superimposed; a root-mean-square
deviation for their P1-P3 backbone atoms of 0.32 Å was
calculated.
TRNOE cross-peaks (Supporting Information, Figure 1b),
suggesting a reversible and fast exchange between the
free in solution and the NS3-bound inhibitor. Further-
more, addition of a more potent competitive inhibitor
(hexapeptide 43; IC50 < 1 nM)49 to the preformed NS3/
33c complex restored the NMR spectrum to that origi-
nally observed for the free ligand 33c (Supporting
Information, Figure 1c), confirming that the observed
TRNOE cross-peaks originated from the binding of 33c
to the enzyme. This competition experiment provided
strong evidence that 33c was binding specifically to HCV
NS3 protease (presumably in, or near, the active site)
and could be competitively displaced by the more potent
substrate-based inhibitor 43. Binding of the macrocyclic
ligands in the active site of the enzyme was later
confirmed by the X-ray crystal structure of analogue 36b
bound to the HCV NS3 protease.17
Con clu sion s
We have undertaken the challenge of trying to under-
stand the subtle details governing the interactions be-
tween the HCV NS3 serine protease and its polyprotein
The distance restraints derived from the volumes of
all the TRNOE cross-peaks were then used in a simulated
annealing protocol in order to generate an ensemble of
(50) (a) Yan, Y.; Li, Y.; Munshi, S.; Sardana, V.; Cole, J . L.; Sardana,
M.; Steinku¨ehler, C.; Tomei, L.; De Franscesco, R.; Kuo, L. C.; Chen,
Z. Protein Science 1998, 7, 837. (b) Kim, J . L.; Morgenstern, K. A.;
Lin, C.; Fox, T.; Dwyer, M. D.; Landro, J . A.; Chambers, S. P.;
Markland, W.; Lepre, C. A.; O′Malley, E. T.; Harbeson, S. L.; Rice, C.
M.; Murcko, M. A.; Caron, P. R.; Thomson, J . A. Cell 1996, 87, 343. (c)
Love, R. A.; Parge, H. E.; Wickersham, J . A.; Hostomsky, Z.; Habuka,
N.; Moomaw, E. W.; Adachi, T.; Hostomska, Z. Cell 1996, 87, 331.
(49) Pause, A.; Kukolj, G.; Bailey, M.; Brault, M.; Doˆ, F.; Halmos,
T.; Lagace´, L.; Maurice, R.; Marquis, M.; Mckercher, G.; Pellerin, C.;
Pilote, L.; Thibault, D.; Lamarre, D. J . Biol. Chem. 2003, 278, 20374.
6194 J . Org. Chem., Vol. 69, No. 19, 2004

Potent Inhibitors of the Hepatitis C Virus NS3 Protease
F IGURE 3. (a) Stereoview of the superposition (P1-P3 backbone atoms only) of the 10 best structures generated by restrained
simulated annealing and derived from the NMR data of inhibitor 33c when bound to the NS3 protease. The structures are colored
by atom type: oxygen in red, nitrogen in blue, carbon in green, and hydrogen in gray (most of the hydrogen atoms are not shown).
(b) Side view of the 10 best NMR-derived NS3-bound structures of 33c.
F IGURE 4. (a) Model of the NS3-bound inhibitor 33c; key interactions that may be implicated in hydrogen bonds with the
active-site residues (as suggested by the model) are indicated with dashed orange lines. Carbon atoms of the inhibitor are colored
in green, those of the NS3 protein in gray. All nitrogen atoms and oxygen atoms are colored in blue and red, respectively. The
sulfur atom of C159 is colored in yellow. (b) Comparison of the NMR-derived bound structure of 33c (molecule in green) to the
cocrystal structure of 36b (molecule in gray) bound to the active site of the NS3/NS4Apeptide protease domain. The two structures
were superimposed using the P1-P3 backbone atoms only with a root-mean-square deviation of 0.32 Å.
substrate. Some key molecular recognition elements of
substrate-based acyclic peptides were evaluated by NMR
and computational chemistry in order to design â-strand
mimics which are potent inhibitors of the NS3 protease.
Efficient protocols were developed, which allowed for the
rapid synthesis of macrocyclic analogues addressing the
role of the P1 moiety and the hydrocarbon linker con-
necting the P1-P3 side chains. Finally, we demonstrated
that the successful design of this unique 15-membered
ring scaffold plays a critical role in the strong affinity
that this novel class of ligands exhibit toward the HCV
NS3 serine protease.
Exp er im en ta l Section
Gen er a l P r otocol for Am id e Bon d F or m a tion . A solu-
tion containing the free acid and the free amine fragments
(ratio of ∼1:1 equiv), in the minimum volume of CH2Cl2 or
DMF (depending on the solubility of the compounds), HATU
(∼1.1 equiv) and DIPEA (3 equiv) was stirred at rt for 5-15 h
(completion of the reaction was monitored by analytical
J . Org. Chem, Vol. 69, No. 19, 2004 6195

Goudreau et al.
HPLC).51 The reaction mixture was then diluted with EtOAc,
washed with 5% aqueous NaHCO3 and brine, dried over
anhydrous MgSO4, concentrated under reduced pressure, and
purified by flash column chromatography.
further purification: 1H NMR (DMSO-d6) δ 2.66 (dd, J ) 8.6,
14.0 Hz, 1H), 2.95 (dd, J ) 3.7, 14.0 Hz, 1H), 3.12-3.21 (m,
2H), 3.28 (dd, J ) 3.8, 14.0 Hz, 1H, 5.08 (dd, J ) 1.6, 9.9 Hz,
1H), 5.17 (dd, J ) 1.6, 16.9 Hz, 1H), 5.71-5.81 (m, 1H), 7.50
(bs, 2H).
Gen er a l P r otocol for RCM. The diene starting material
was dissolved in CH2Cl2 at a concentration of 0.01 M, and the
solution was deoxygenated by bubbling argon (∼1 h for a
volume of 500 mL). A solution of Grubbs’ catalyst bis-
(tricyclohexylphosphine)benzylideneruthenium(IV) dichloride
(5 mmol % dissolved in a small amount of degassed CH2Cl2)
was added, and the reaction mixture was stirred under reflux
for ∼2 h (completion of the reaction was monitored by
analytical HPLC). The crude reaction mixture was subse-
quently concentrated under reduced pressure and filtered
through a short pad of silica gel, eluting first with CH2Cl2 (to
remove most of the catalyst) and then with EtOAc to isolate
the macrocyclic product(s) in reasonable purity (>80% homo-
geneity by HPLC). In some cases, the desired macrocyclic
compound(s) was further purified by flash column chromatog-
raphy to obtain the ester(s) in very high purity (>98%
homogeneity) and 50-95% isolated yield. However, most of
the time, the semipure macrocyclic ester was used directly in
the final hydrolysis reaction.
Gen er a l P r otocol for Ester Hyd r olysis a n d P u r ifica -
tion of In h ibitor s. Saponification of all esters was carried
out under basic conditions using LiOH (2.5 equiv) in a solution
of THF/MeOH/H2O and over a period of 16 h at rt. The reaction
mixture was then acidified to pH ) 4 using 1 N HCl and
evaporated to dryness under reduced pressure to obtain the
free acids. Intermediate building blocks were used in the
subsequent synthetic steps without further purification. How-
ever, all compounds used for evaluation in the enzymatic in
vitro assay were purified to >95% homogeneity by semi-
preparative C18 reversed-phase HPLC using a solvent gradient
from 5% to 100% aqueous CH3CN (all solvents contained 0.06%
TFA). The final homogeneity of each inhibitor was determined
by analytical HPLC using either method A or B. Method A:
Vydac C18 reversed-phase column (0.46 × 12.5 cm i.d., 5 µm,
300 Å) and a 35 min linear gradient from 0 to 100% aqueous
CH3CN (all solvents contained 0.06% TFA) at a flow rate of
1.5 mL/min with the UV detector set at λ ) 230 nm. Method
B: The UV detector was set simultaneously at λ ) 220 and
254 nm. Solvent A ) 5% aqueous CH3CN, solvent B ) 100
CH3CN, both solvents contained 0.06% TFA. YMC Combi-
Screen ODS-AQ C18 reversed-phase column (50 × 4.6 mm i.d.,
S-5µm, 12 nm) with linear gradient from solvent A to 1:1
mixture of solvents A/B at a flow rate of 3 mL/min for the first
6 min. After that period, the flow rate was increased to 3.5
mL/min, and the solvent gradient was changed from 1:1 A,B
to 100% B over a period of 4.5 min.
The 3-allylsulfanyl-2-aminoproprionic acid intermediate (72
mg, 0.45 mmol) was subsequently mixed with Boc2O (110 mg,
0.49 mmol) and K2CO3 (1 M, 560 µL, 0.56 mmol) in THF (5
mL) and stirred at rt for 20 h. H2O was added and the reaction
mixture was concentrated under reduced pressure to remove
most of the THF. The aqueous mixture was extracted with
Et2O (2 × 20 mL) to remove any unreacted Boc2O, acidified to
pH ∼3 with acetic acid and re-extracted with EtOAc (2 × 20
mL). The combined organic layers were washed with brine,
dried over anhydrous MgSO4, filtered, and evaporated to give
the Boc-protected 3-allylsulfanyl-2-aminoproprionic acid in-
termediate as a white solid (∼120 mg) which was used in the
following step without purification: 1H NMR (CDCl3) δ 1.46
(s, 9H), 2.87-3.02 (m, 2H), 3.16 (d, J ) 7.0 Hz, 2H), 4.54 (bs,
1H), 5.12 (s, 1H), 5.16 (d, J ) 4.8 Hz, 1H), 5.35 (bd, J ) 5.7
Hz, 1H), 5.71-5.82 (m, 1H).
The above N-Boc acid was treated with CH2N2 in CH2Cl2
until the yellow color of the diazomethane persisted. The
solution was evaporated to dryness to give the N-Boc-protected
derivative of 6a as a white solid (∼125 mg): 1H NMR (CDCl3)
δ 1.45 (s, 9H), 2.79-2.92 (m, 2H), 3.12-3.15 (m, 2H), 3.77 (s,
3H), 4.52 (bs, 1H), 5.11 (bs, 1H), 5.14 (bs, 1H), 5.30 (bs, J )
1H), 5.70-5.81 (m, 1H).
After hydrolysis of the Boc protecting groups with HCl in
dioxane (∼30 equiv) the P1 fragment d-S-allylcysteine methyl
ester hydrochloride (6a ) was isolated [ES+ MS m/z 176 (M +
H)+; ES- MS m/z 174 (M - H)-] and used immediately for the
synthesis of the dipeptide P1-P2 intermediate 20a .
In ter m ed ia te 7a . To a suspension of benzyltriethylammo-
nium chloride (5.08 g, 22.3 mmol) in 50% aqueous NaOH (50
mL), 1,2-dibromo-5-hexene (8.10 g, 33.46 mmol), and di-tert-
butylmalonate (4.82 g, 22.30 mmol) were added in succession.
The mixture was stirred vigorously at rt for 16 h, diluted with
H2O, and extracted with CH2Cl2 (3 × 50 mL). The combined
organic layers were further washed with H2O (2 × 50 mL) and
brine/H2O (2/1, 2 × 50 mL), dried over anhydrous MgSO4, and
evaporated to dryness under vacuum. The crude residue was
purified by flash column chromatography on silica gel, using
a gradient from 3% to 5% EtOAc in hexanes as the eluent to
obtain compound 7a in 38% yield (2.48 g): 1H NMR (CDCl3) δ
1.19 (bd, J ) 7.9 Hz, 2H), 1.25-1.33 (m, 1H), 1.46 (s, 9H),
1.48 (s, 9H), 1.47-1.60 (m, 1H), 1.75-1.82 (m, 1H), 2.14-2.22
(m, 2H), 4.93-5.50 (m, 1H), 4.96 (dm, J ) 10.2 Hz, 1H), 5.18
(dm, J ) 17.2 Hz, 1H); ES+ MS m/z 297 (M + H)+.
In ter m ed ia te 7b. To a suspension of potassium tert-
butoxide (5.75 g, 51.25 mmol) in anhydrous diethyl ether (150
mL) at 0 °C was added H2O (203 µL, 11.27 mmol), and the
reaction mixture was stirred at 0 °C for 10 min. An ether
solution of compound 7a (2.48 g in 10 mL diethyl ether, 10.25
mmol) was added, and the mixture was stirred at rt for 5 h.
The mixture was diluted with ice-cold H2O and extracted with
diethyl ether (3 × 200 mL). The aqueous layer was acidified
to pH ∼3.5 with ice-cold 10% aqueous citric acid and re-
extracted with EtOAc (3 × 200 mL). The combined EtOAc
layers were washed with H2O (2 × 100 mL) and brine (100
mL), dried over anhydrous MgSO4, and evaporated to give
compound 7b in 85% yield (based on some recovered starting
material): 1H NMR (400 MHz, CDCl3) δ 1.51 (s, 9H), 1.64-
1.68 (m, 1H), 1.68-1.75 (m, 1H), 1.77-1.88 (m, 1H), 1.96-
2.01 (m, 1H), 2.03-2.22 (m, 3H), 5.01 (dm, J ) 6.4 Hz, 1H),
5.03 (dm, J ) 14.9 Hz, 1H), 5.72-5.83 (m, 1H); ES+ MS m/z
241 (M + H)+.
D-Allylglycin e Meth yl Ester Hyd r och lor id e (5a ). To an
ice-cold solution of anhydrous methanol (14.5 mL) was added
thionyl chloride (1.9 mL, 26 mmol) dropwise. To this solution
was added d-allylglycine (970 mg, 8.43 mmol) in small por-
tions. The resulting solution was stirred at rt for 4 h and then
at reflux for 2.5 h. After cooling, the reaction mixture was
concentrated under reduced pressure to afford compound 5a
as a white solid (1.40 g, >99% yield): 1H NMR (DMSO-d6) δ
2.59 (dd, J ) 7 Hz, 2H), 3.71 (s, 3H), 4.08 (bs, 1H), 5.21-5.10
(m, 2H), 5.82-5.72 (m, 1H), 8.70 (bs, 3 H); ES+ MS m/z 130
(M + H)+.
D-S-Allylcystein e Meth yl Ester Hyd r och lor id e (6a ).
D-Cysteine hydrochloride (290 mg, 1.6 mmol) was mixed with
allyl bromide (300 mg, 220 µL, 2.5 mmol) in 2 M NH4OH (5
mL) and stirred at rt for 20 h. The reaction mixture was
concentrated to precipitate the product as a white solid. The
solid was filtered, washed with ethanol (3 × 5 mL), dried under
reduced pressure, and used in the following step without
In ter m ed ia te 8. To a solution of the acid 7b in anhydrous
benzene (1.14 g in 25 mL benzene, 4.74 mmol) was added Et3N
(800 µL, 5.68 mmol) followed by the addition of diphenylphos-
phoryl azide (1.13 mL, 5.21 mmol), and the mixture was heated
to reflux for 3.5 h. Subsequently, trimethylsilylethanol (1.36
(51) In most cases, the choice of the coupling reagent and the base
was not critical to the efficiency of coupling; replacement of HATU/
DIPEA with TBTU/DIPEA or HATU/NMM gave very similar results.
6196 J . Org. Chem., Vol. 69, No. 19, 2004

Potent Inhibitors of the Hepatitis C Virus NS3 Protease
mL, 9.48 mmol) was added, and stirring at reflux was
continued for an additional 4 h. The mixture was then cooled
to rt, evaporated to half of its original volume, diluted with
diethyl ether (30 mL), washed with 5% aqueous NaHCO3 (2
× 30 mL) and brine (50 mL), dried over anhydrous MgSO4,
and evaporated. The residual oil was chromatographed on
silica gel using 10% EtOAc in hexanes as the eluent to obtain
pure compound 8 in 88% yield (1.49 g): 1H NMR (CDCl3) δ
0.03 (s, 9H), 0.91-0.99 (m, 2H), 1.18-1.29 (m, 2H), 1.45 (bs,
11H), 1.56-1.72 (m, 2H), 2.02-2.18 (m, 2H), 4.12 (t, J ) 8.3
Hz, 2H), 4.93 (dm, J ) 10.2 Hz, 1H), 4.98 (dm, J ) 17.2 Hz,
1H), 5.07 (bs, 1H), 5.71-5.83 (m, 1H); ES+ MS m/z 356.4 (M
+ H)+.
mixture was extracted with EtOAc, and the aqueous layer was
acidified to pH ∼4. The free acid 15f precipitated from the
acidic aqueous layer as a white solid, which was filtered and
dried under vacuum before used for the synthesis of P1-P2
fragments:
(major rotamer) 1.36 (s, 9H), 2.32-2.43 (m, 1H), 2.63-2.73
(m, 1H), 3.76 (bs, 2H), 3.93 (s, 3H), 4.34-4.41 (m, 1H), 5.53-
5.59 (m, 1H), 7.17 (dd, J ) 9.2, 2.5 Hz, 1H), 7.39 (d, J ) 2.5
Hz, 1H), 7.45 (s, 1H), 7.48-7.56 (m, 3H), 8.00 (d, J ) 9.2 Hz,
1H), 8.27 (d, J ) 7.0 Hz, 2H); ES+ MS m/z 465.5 (M + H)+.
1H NMR (DMSO-d6, rotamers in ∼2:1 ratio) δ
8-Non en oic Acid (16b). To a stirred suspension of finely
cut Mg ribbons (0.55 g, 22.5 mmol) in dry THF (30 mL)
containing dibromoethane (0.1 mL) was added dropwise
8-bromo-1-octene (2.52 mL, 15 mmol) over a period of 15 min
(the reaction is slightly exothermic). After 30 min, the mixture
was heated to 38 °C for 1 h and then cooled to -78 °C before
it was added via a cannula onto an excess amount of solid CO2.
The mixture was diluted with diethyl ether (100 mL), and the
solution was washed with brine (2 × 50 mL), dried over
anhydrous MgSO4, and evaporated to dryness under vacuum.
The crude oil obtained was purified by flash column chroma-
tography on silica gel using 15% EtOAc in hexanes as the
eluent to give the carboxylic acid 16b in 62% yield (1.44 g):
1H NMR (CDCl3) δ 1.31-1.42 (m, 6H), 1.60-1.69 (m, 2H),
2.02-2.09 (m, 2H), 2.35 (t, J ) 8.3 Hz, 2H), 4.99 (dm, J )
10.0 Hz, 1H), 5.04 (dm, J ) 17.0 Hz, 1H), 5.75-5.86 (m, 1H).
In ter m ed ia te 17b. To a vigorously stirring solution of the
carboxylic acid 16b (1.36 g, 8.7 mmol) in anhydrous THF (70
mL) at -78 °C were added freshly distilled Et3N (1.6 mL; 11.3
mmol) and pivaloyl chloride (1.18 mL, 9.58 mmol) via a syringe
under anhydrous conditions. The mixture was stirred at -78
°C for 15 min and then at 0 °C for 45 min. The mixture was
cooled again to -78 °C and transferred via a cannula into an
anhydrous solution of (4S)-4-(phenylmethyl)-2-oxazolidinone
lithium salt in THF at -78 °C; the lithium salt of the
oxazolidinone reagent had been previously prepared by the
slow addition of n-BuLi (2.00 M in hexanes, 7.85 mL, 15.7
mmol) into a THF (20 mL) solution of the oxazolidinone (2.78
g, 15.7 mmol) in THF at -78 °C. The reaction mixture was
stirred at -78 °C for 15 min and then at rt for 1.5 h. Finally,
it was quenched with an aqueous solution of sodium bisulfate
(1 M, 100 mL), and the THF evaporated to 3/4 of its initial
volume under vacuum. The residue was extracted with EtOAc
(2 × 150 mL), the combined organic layers were washed with
5% NaHCO3 (3 × 50 mL), brined (2 × 50 mL), dried over
anhydrous MgSO4, and evaporated to dryness under vacuum.
The resulting crude oil was purified by flash column chroma-
tography (using 15% EtOAc in hexanes) to obtain compound
17b in 68% yield (1.88 g): 1H NMR (CDCl3) δ 1.35-1.47 (m,
6H), 1.67-1.74 (m, 2H), 2.02-2.09 (m, 2H), 2.65 (dd, J ) 13.4,
9.9 Hz, 1H), 2.84-3.02 (m, 2H), 3.31 (dd, J ) 13.4, 3.2 Hz,
1H), 4.13-4.22 (m, 2H), 4.62-4.71 (m, 1H), 4.93 (dm, J ) 10.2
Hz, 1H), 5.00 (dd, J ) 1.6, 17.2 Hz, 1H), 5.75-5.84 (m, 1H),
7.18-7.38 (m, 5H); ES+ MS m/z 316.4 (M + H)+.
Com p ou n d 18b. To a stirred solution of KHMDS (0.8 M
THF, 22 mL, 17.5 mmol) in dry THF (50 mL) at -78 °C was
added a solution of the acid derivative 17b (3.25 g, 10.30 mmol)
in dry THF (40 mL) at -78 °C via a cannula. The reaction
mixture was stirred at -78 °C for 45 min, and then a solution
of trisyl azide (3.67 g, 11.85 mmol, in 40 mL of dry THF) was
added at -78 °C. The mixture was stirred at -78 °C for 3 min
and then quenched with acetic acid (5 mL). The mixture was
stirred at rt for 2 h and finally at 40 °C for 15 min. Most of
the THF was then evaporated under vacuum, and the residue
was dissolved in EtOAc (100 mL). This was washed with H2O
(50 mL), 5% NaHCO3 (3 × 50 mL), and brine (50 mL), dried
over anhydrous MgSO4, and evaporated to dryness under
vacuum. The oil obtained was purified by flash column
chromatography (using 1:1 hexanes/CH2Cl2) to give compound
18b (2.47 g, yield 67%): 1H NMR (CDCl3) δ 1.32-1.45 (m, 5H),
1.45-1.8 (m, 1H), 1.65-1.73, 1.75-1.88 (2m, 2H, rotamers),
2.01-2.11 (m, 2H), 2.72-3.02 (m, 1H), 3.33 (dd, J ) 3.2, 13.4
4-Hyd r oxy-7-m eth oxy-2-p h en ylqu in olin e (13c). A solu-
tion of ethyl benzoyl acetate (100.0 g, 0.52 mol), m-anisidine
(128.1 g, 1.04 mol), and 4 N HCl/dioxane (5.2 mL) in toluene
(1.0 L) was refluxed for 6.25 h in a Dean-Stark apparatus.
The cooled toluene solution was successively washed with
aqueous 10% HCl (2 × 300 mL), 1 N NaOH (2 × 300 mL),
H2O (300 mL), and brine (150 mL). The toluene phase was
dried over anhydrous MgSO4, filtered, and concentrated under
vacuum to give a 1.2:1.0 mixture of enamine ester 11 and
enamine amide 12 (total 144.6 g of material, 45% and 38%
crude yield of 11 and 12, respectively) as a dark brown oil.
The crude oil was heated to 280 °C for 80 min while the EtOH
generated from the reaction was distilled. The cooled dark solid
obtained was triturated with CH2Cl2 (200 mL). The suspension
was filtered and the resulting solid washed with CH2Cl2 to
give the desired quinoline 13c (22.6 g, 17% overall yield from
m-anisidine) as a beige solid:
1
En a m in e ester 11: H NMR (DMSO-d6) δ 1.22 (t, J ) 7.0
Hz, 3H), 3.51 (s, 3H), 4.13 (q, J ) 7.0 Hz, 2H), 4.93 (s, 1H),
6.27-6.25 (m, 2H), 6.48 (dd, J ) 8.1, 1.9 Hz, 1H), 6.99 (t, J )
8.1 Hz, 1H), 7.41-7.35 (m, 5H), 10.13 (s, 1H); 13C NMR
(DMSO-d6) δ 14.4, 54.8, 59.0, 91.3, 107.4, 108.8, 114.1, 127.9,
128.7, 129.4, 129.7, 135.6, 141.3, 158.3, 159.4, 168.9; ES+ MS
m/z 298 (M + H)+.
En a m in e a m id e 12: 1H NMR (DMSO-d6) δ 3.52 (s, 3H),
3.72 (s, 3H), 5.22 (s, 1H), 6.19 (bs, 1H), 6.24 (bd, J ) 8.1 Hz,
1H), 6.43 (dd, J ) 8.1, 2.0 Hz, 1H), 6.59 (dd, J ) 8.0, 1.5 Hz,
1H), 6.99 (t, J ) 8.1 Hz, 1H), 7.12 (bd, J ) 8.0 Hz, 1H), 7.18
(t, J ) 8.0 Hz, 1H), 7.35 (bs, 1H), 7.42-7.36 (m, 5H), 9.84 (s,
1H), 10.83 (s, 1H); 13C NMR (DMSO-d6) δ 54.7, 54.9, 96.3,
104.6, 106.7, 108.0, 108.1, 111.2, 113.5, 127.6, 128.7, 129.4,
129.5, 136.4, 140.8, 142.0, 155.4, 159.4, 159.5, 167.7; ES+ MS
m/z 375 (M + H)+.
Qu in olin e 13c (7-m eth oxy qu in olin e isom er ): 1H NMR
(DMSO-d6) δ 3.86 (s, 3H), 6.25 (s, 1H), 6.93 (dd, J ) 2.0, 9.0
Hz, 1H), 7.20 (d, J ) 2.0 Hz, 1H), 7.59-7.55 (m, 3H), 7.83-
7.78 (m, 2H), 7.99 (d, J ) 9.0 Hz, 1H), 11.51 (s, 1H); 13C NMR
(DMSO-d6) δ 55.4, 99.7, 107.1, 113.2, 119.2, 126.5, 127.2, 129.0,
130.4, 134.2, 142.3, 149.6, 161.9, 176.5; ES+ MS m/z 252 (M +
H)+.
Syn th esis of th e P 2 Bu ild in g Block s 15f. A solution of
Boc-protected cis-4-hydroxyproline, the desired 4-hydroxyquin-
oline (13c, 1 equiv), and Ph3P (2 equiv) in anhydrous THF (∼70
mM concentration) was cooled to 0 °C. Diisopropyl azodicar-
boxylate (DIAD, 2 equiv) was added dropwise, and the reaction
mixture was stirred at rt for approximately 15 h. After that
period, the reaction mixture was evaporated to dryness under
vacuum, and the residue was purified by flash column chro-
matography to give the methyl ester 15c in typically 70-80%
yield and reasonably good purity. Complete removal of the Ph3-
PO by flash column chromatography was often difficult.
However, an easily removable phosphine reagent was recently
reported which allowed for the synthesis of intermediate of
such intermediates in higher yields and excellent purity.52 The
methyl ester 15c was hydrolyzed under the usual basic
conditions. The reaction suspension was poured into H2O, the
(52) Yoakim, C.; Guse, I.; O′Meara, J . A.; Thavonekham, B. Synlett
2003, 473.
J . Org. Chem, Vol. 69, No. 19, 2004 6197

Goudreau et al.
1
Hz, 1H), 4.10-4.28 (m, 2H), 4.62-4.72 (m, 1H), 4.90-5.05 (m,
3H), 5.73-5.88 (m, 1H), 7.17-7.38 (m, 5H); ES+ MS m/z 357.5
(M + H)+.
Com p ou n d s 23a : H NMR (CDCl3) δ 1.14-1.35 (m, 1H),
1.44 (s, 9H), 1.45 (s, 9H), 1.62-1.78 (m, 4H), 2.09-2.22 (m,
2H), ∼2.6 (bs, 1H), ∼2.95 (bs, 1H), 3.77 (bs, 1H), 3.88 (bs, 1H),
4.44-4.55 (m, 1H), 4.98 (d, J ) 10.2 Hz, 1H), 5.03 (dd, J )
1.6, 17.2 Hz, 1H), 5.24 (bs, 1H), 5.75-5.88 (m, 1H), 7.42-7.58
(m, 2H), 7.63-7.73 (m, 2H), 8.04 (d, J ) 8.3 Hz, 1H), 8.11 (d,
J ) 8.3 Hz, 1H), 8.74 (d, J ) 5.1 Hz, 1H); ES+ MS m/z 552.3
(M + H)+; ES- MS m/z 550.2 (M - H)-.
Com p ou n d 19b. To a stirred solution of anhydrous SnCl2
(2.61 g, 13.8 mmol) in anhydrous MeOH (80 mL) was cannu-
lated a solution of the azide 18b (2.45 g, 6.9 mmol) at 0 °C in
anhydrous MeOH (20 mL). The mixture was stirred at rt for
4 h. The MeOH was evaporated, the foamy material obtained
was taken into dioxane/H2O (100 mL/20 mL) and treated with
Boc2O (3.0 g, 13.8 mmol) and NaHCO3 (2.89 g, 34.5 mmol),
the pH was adjusted to 8 with more NaHCO3 (if needed), and
the mixture was stirred at rt for 16 h. Part of the dioxane was
evaporated (∼50%), and the residue was extracted with EtOAc
(2 × 100 mL). The combined organic layer was washed with
brine (2 × 50 mL), dried over anhydrous MgSO4, and evapo-
rated to dryness. The residue obtained was purified by flash
column chromatography (using 20-25% EtOAc in hexanes as
eluent) to give the compound 19b (1.75 g, yield 60%): 1H NMR
(CDCl3) δ 1.27-1.53 (m, 7H), 1.46 (s, 9H), 1.80 (m, 1H), 2.00-
2.08 (m, 2H), 2.80 (t, J ) 12.1 Hz, 1H), 3.34 (d, 14.3 Hz, 1H),
4.17-4.23 (m, 2H), 4.60-4.66 (m, 1H), 4.93 (d, J ) 10.2 Hz,
1H), 5.50 (dd, J ) 1.9, 17.2 Hz, 1H), 5.13 (bs, 1H), 5.38-5.43
(m, 1H), 5.74-5.84 (m, 1H), 7.22-7.36 (m, 5H); ES+ MS m/z
431.6 (M + H)+.
(2S)-N-Boc-2-a m in o-8-n on en oic Acid (20b). To a stirred
solution at 90 °C of the N-Boc derivative 19b (1.74 g, 4.04
mmol) in THF/H2O (75 mL/15 mL) were added H2O2 (30% v/w,
2.05 mL, 16.2 mmol) and LiOH‚H2O (0.34 g, 8.1 mmol), and
the solution was stirred at 0 °C for 1 h. The reaction was
quenched with Na2SO3 (2.24 g in H2O, 15 mL, 17.8 mmol).
The pH was adjusted to ∼4 with 10% aqueous citric acid and
the mixture diluted with EtOAc. The aqueous fraction was
extracted once more with EtOAc (50 mL), and the combined
EtOAc layer was washed twice with brine (50 mL), dried over
anhydrous MgSO4, and evaporated to dryness under vacuum.
The residue obtained was purified by flash column chroma-
tography (using 20% hexane in EtOAc as the eluent) to give
the free carboxylic acid 20b as a pale yellow oil (0.76 g, yield
70%): 1H NMR (400 MHz, DMSO-d6) δ 1.21-1.35 (m, 6H),
1.38 (s, 9H), 1.50-1.65 (m, 2H), 2.00 (q, J ) 6.9 Hz, 2H), 3.83
(m, 1H), 4.93 (dm, J ) 10.2 Hz, 1H), 5.00 (dm, J ) 17.0 Hz,
1H), 5.79 (tdd, J ) 6.7, 17.0, 10.2 Hz, 1H), 7.01 (d, J ) 8.0 Hz,
1H), 12.35 (br, 1H); ES+ MS m/z 272.4 (M + H)+.
Dip ep tid es 23a . To a solution of derivative 8 (1.19 g, 3.35
mmol, in 30 mL of THF) was added t-Bu4NF (6.7 mL of 1 M
in THF, 6.7 mmol), and the mixture was first stirred at rt for
16 h and subsequently heated to reflux for 15 min. The solvent
was carefully evaporated under low pressure (due to the high
volatility of the tert-butyl ester of 10, caution should be
exercised during the evaporation of the solvent). The crude
residue was redissolved in EtOAc (100 mL), washed with H2O
(2 × 50 mL) and brine (50 mL), and dried over MgSO4. After
careful evaporation of the solvent under vacuum, an aliquot
of this racemic compound (10)53 was coupled with the P2
derivative 15a following the standard protocol. The crude
products was purified by flash column chromatography (using
∼8-10% diethyl ether in EtOAc as the eluent) to obtain a pure
sample of the set of diastereomers 23a and 24a . Each di-
astereomerically pure product was isolated in ∼20% yield,
however, the absolute stereochemistry was not assigned at this
stage.
Com p ou n d s 24a :54 1H NMR (CDCl3) δ 1.18-1.35 (m, 1H),
1.45 (bs, 18H), 1.50-1.60 (m, 1H), 1.60-1.78 (m, 3H), 2.11-
2.22 (m, 2H), ∼2.3 (bs, 1H), 2.93 (m,1H), 3.75 (bs, 1H), 3.90
(bs, 1H), 4.45-4.58 (bs, 1H), 4.98 (d, J ) 10.2 Hz, 1H), 5.07
(dd, J ) 1.6, 17.2 Hz, 1H), 5.22 (bs, 1H), 5.78-5.88 (m, 1H),
7.45-7.58 (m, 2H), 7.65-7.75 (m, 2H), 8.05 (d, J ) 8.3 Hz,
1H), 8.11 (d, J ) 8.3 Hz, 1H), 8.74 (d, J ) 5.1 Hz, 1H); ES+
MS m/z 552.3 (M + H)+; ES- MS m/z 550.2 (M - H)-.
N-Boc-P r otected Dip ep tid e Meth yl Ester 25a . To a
solution of N-Boc-protected, tert-butyl ester of dipeptides 23a
(68 mg, 0.123 mmol) in dry CH2Cl2 was added a solution of
HCl in dioxane (4 M, 4 mL), and the mixture was stirred at rt
for 1.5 h. The solvent was then evaporated and the residue
dried under high vacuum. The residue was disolved in dioxane/
H2O (5 mL, 4:1 ratio), Boc2O (81 mg, 0.37 mmol) and NaHCO3
(73 mg, 0.86 mmol) were added, and the reaction mixture was
stirred at rt for 12 h. The mixture was diluted with H2O (10
mL) and extracted with EtOAc (50 mL). The aqueous layer
was acidified to pH 4 and re-extracted with CH2Cl2 (2 × 100
mL). The combined CH2Cl2 layers were dried over anhydrous
MgSO4 and evaporated to dryness under reduced pressure. The
residue was purified by flash column chromatography, using
3% EtOH in EtOAc as the eluting solvent, to obtain the pure
P2 N-Boc protected, P1 free acid dipeptide intermediate. This
intermediate was dissolved in diethyl ether/MeOH (3:2 mL,
respectively) and treated with a slight excess of diazomethane
dissolved in diethyl ether. After 30 min, the excess diazo-
methane was destroyed with the addition of HCl (a drop of 4
M in dioxane), and the mixture was evaporated to dryness to
obtain the N-Boc-protected methyl ester 25a : 1H NMR (CDCl3)
mixture of rotamers (∼1:9 ration) δ (major rotamer) 1.22-1.35
(m, 2H), 1.42 (s, 9H), 1.55-1.62 (m, 1H), 1.62-1.74 (m, 2H),
2.06-2.18 (m, 2H), 2.32 (bs, 1H), 2.94 (bs, 1H), 3.71 (s, 3H),
3.70-3.90 (m, 2H), 4.55 (bs, 1H), 4.96 (dd, J ) 0.95, 9.2 Hz,
1H), 5.03 (dm, J ) 17.2 Hz, 1H), 5.27 (bs, 1H), 5.75-5.83 (m,
1H), 6.77 (bs, 1H), 7.49 (ddd, J ) 1.3, 1.3, 8.3 Hz, 1H), 7.61
(bs, 1H), 7.70 (ddd, J ) 1.3, 1.3, 8.3 Hz, 1H), 8.04 (d, J ) 8.6
Hz, 1H), 8.15 (d, J ) 8.0 Hz, 1H), 8.75 (d, J ) 5.1 Hz, 1H);
ES+ MS m/z 510.3 (M + H)+; ES- MS m/z 508.2 (M - H)-.
Acyclic Dien e 30a . The dipeptides 25a was first treated
with HCl (4 N in dioxane, ∼20 equiv) over a period of 1.5 h at
rt. The reaction mixture was concentrated under reduced
pressure, and the resulting free amine (HCl salt form) was
coupled with the P3 linker 20a under the standard reaction
conditions for amide bond formation. The crude product was
purified by flash column chromatography (using 2% EtOH in
EtOAc) to afford the corresponding diene 30a as a white foam
(∼70% yield): 1H NMR (CDCl3, rotamers in ∼10:1 ratio) δ
(major rotamer) 1.21-1.27 (m, 1H), 1.36 (s, 9H), 1.45-1.81
(m, 8H), 2.02-2.20 (m, 4H), 2.25-2.35 (m, 1H), 2.92-3.02 (m,
1H), 3.66 (s, 3H), 3.94-3.98 (dd, J ) 5.1, 11 Hz, 1H), 4.29-
4.30 (bd, J ) 9.9 Hz, 1H), 4.44-4.52 (m, 1H), 4.82 (dd, J )
5.4, 8.3 Hz, 1H), 4.92-5.06 (m, 4H), 5.14 (d, J ) 8.3 Hz 1H),
5.34-5.40 (m, 1H), 5.70-5.84 (m, 2H), 6.82 (d, J ) 5.1 Hz,
1H), 7.47-7.55 (m, 2H), 7.71 (dt, J ) 1.3, 7.5 Hz, 1H), 8.03 (d,
J ) 8.2 Hz, 1H), 8.17 (d, J ) 8.0 Hz, 1H), 8.78 (d, J ) 5.1 Hz,
1H); ES+ MS m/z 635.5 (M + H)+; ES- MS m/z 633.2 (M -
H)-.
(53) The methyl-(1R,2S)/ (1S,2R)-1-amino-2-homoallylcyclopropyl
carboxylate (10) was also prepared by first treating compound 8 (1.74
g, 4.89 mmol) with a mixture of TFA/CH2Cl2 (20 mL, 1: 1 ratio), and
then with CH2N2. The unreacted CH2N2 was quenced with a few drops
of acetic acid, 4 N HCl in dioxane (1.5 mL) was added and the solvents
were evaporated to dryness to give the crude HCl salt of compound
10. Fairly pure compound 10 was obtained after trituration in diethyl
ether/hexane (3 × 5 mL, 1:1 ratio). 1H NMR (400 MHz, CDCl3) δ 1.45-
1.49 (m, 1H), 1.53-1.79 (m, 2H), 1.85-1.89 (m, 1H), 1.98-2.05 (m,
1H), 2.10-2.21 (m, 2H), 3.83 (s, 3H), 4.99 (d, J ) 10.2 Hz, 1H), 5.07
(dd, J ) 1.6, 17.2 Hz, 1H), 5.73-5.83 (m, 1H), 9.0 (bs, 2H).
Acyclic Liga n d 30b. A sample of the tripeptide diene 30a
was saponified and puried by semipreparative C18 reversed-
(54) Compound 24a was converted to the macrocyclic inhibitor 26
using the same procedure as for the preparation of its epimer compound
33b. In the enzymatic in vitro assay, compound 26 was found to be
40-fold less potent than 33b.
6198 J . Org. Chem., Vol. 69, No. 19, 2004

Potent Inhibitors of the Hepatitis C Virus NS3 Protease
phase HPLC (using the standard protocols) to give the free
acid ligand 30b as an amorphous white solid in >98.5%
homogeneity, as determined by analytical HPLC method A:
1H), 7.10 (d, J ) 5.7 Hz, 1H), 7.25 (brd, J ) 8.0 Hz, 1H), 7.56
(br, 1H), 7.71 (br, 4H), 8.16-8.21 (br, 3H), 8.28 (d, J ) 13.3
Hz, 1H). 1H NMR decoupling experiments, via selective
irradiation of the methylene protons adjacent to the olefinic
protons [5.22-5.29 (m, 1H, Hꢀ), 5.47-5.54 (m, 1H, Hη)],
resulted in J ) 15.2 Hz between Hꢀ and Hη, confirming a trans
geometry of the double bond (the details of the NMR spectra
are available in the Supporting Information): ES+ MS m/z
699.3 (M + H)+; ES- MS m/z 697.3 (M - H)-.
1
tR ) 19.2 min; H NMR (DMSO-d6, rotamers in ∼3.5:1 ratio)
δ (major rotamer) 1.23 (s, 9H), 1.2-1.5 (m, 6H), 1.52-1.64 (m,
3H), 1.92-2.02 (m, 2H), 2.02-2.18 (m, 2H), 2.26-2.33 (m, 1H),
2.52-2.62 (m, 1H), 3.98 (bd, J ) 9.2 Hz, 1H), 4.10 (bd, J ) 7.0
Hz, 1H), 4.34-4.49 (m, 2H), 4.93 (bd, J ) 8.9 Hz, 2H), 5.00 (d,
J ) 17.2 Hz, 2H), 5.64 (bs, 1H), 5.73-5.89 (m, 2H), 7.12 (d, J
) 7.3 Hz, 1H), 7.28 (s, 1H), 7.48 (bs, 1H), 7.71 (dd, J ) 7.6 Hz,
1H), 7.99 (dd, J ) 7.4, 7.9 Hz, 1H), 8.09 (d, J ) 8.3 Hz, 1H),
8.34 (d, J ) 8.0 Hz, 1H), 9.08 (bd, J ) 5.1 Hz, 1H); ES+ MS
m/z 621.2 (M + H)+, 643.2 (M + Na)+; ES- MS m/z 619.1 (M
- H)-.
Ma cr ocyclic in h ibitor 34b: white amorphous solid in
>92% homogeneity with respect to 34b (sample contains ∼17%
of 34a ), as determined by analytical HPLC method A; tR
)
20.1 min; 1H NMR (DMSO-d6) δ ∼1.0-1.2 (m, 3H), 1.24 (s,
9H), ∼1.2-1.4 (m, 2H), ∼1.4-1.6 (m, 1H), ∼1.6-1.7 (m, 1H),
∼1.7-1.8 (m, 3H), 1.99 (br, 1H), 2.20 (br, 1H), 2.25-2.35 (m,
1H), 2.40-2.50 (m, 1H), 2.50-2.68 (m, 1H), 3.96 (s, 3H), 4.05
(brd, J ) 9.2 Hz, 1H), 4.22-4.28 (m, 1H), 4.44-4.49 (m, 2H),
5.22-5.30 (m, 1H, Hꢀ), 5.42-5.50 (m, 1H, Hη), 5.79 (br, 1H),
6.81 (br, 1H), 7.24 (d, J ) 7.6 Hz, 1H), 7.52 (br, 1H), 7.62-
7.70 (br, 4H), 8.05 (s, 1H), 8.15-8.26 (br, 3H). 1H NMR
decoupling experiments, via selective irradiation of the
methylene protons adjacent to the olefinic protons [5.22-5.30
(m, 1H, Hꢀ), 5.42-5.50 (m, 1H, Hη)], resulted in J ) 10.4 Hz
between Hꢀ and Hη, confirming a cis geometry of the double
bond (the details of the NMR spectra are available in the
Supporting Information): ES+ MS m/z 699.4 (M + H)+; ES-
MS m/z 697.3 (M - H)-.
Ma cr ocyclic In h ibitor 33b. Tripeptide diene 30a was
cyclized using the stranderd RCM reaction conditions. After
completion of the reaction, the solvent was evaporated and the
crude product was purified by flash column chromatography
using 4% EtOH in EtOAc as the eluting solvent to give 50%
yield of macrocyclic ester 33a as a mixture of the E and Z
double bonds. After saponification of the ester, under standerd
conditions, the mixture was purified by semipreparative C18
reversed-phase HPLC, using a solvent gradient from 0 to 40%
aqueous CH3CN (0.06% TFA), to isolate the pure E-isomer,
compound 33b, as a white amorphous solid in >93% homo-
geneity, as determined by analytical HPLC method A: tR
)
1
16.6 min; H NMR (DMSO-d6) δ 1.12 (s, 9H), 1.20-1.24 (m,
2H), 1.32-1.40 (m, 3H), 1.58-1.62 (m, 2H), 1.68-1.78 (m, 3H),
1.95-2.02 (m, 1H), 2.08-2.18 (m, 2H), 2.42-2.59 (m, 2H),
3.97-3.40 (bd, J ) 9.8 Hz, 2H), 4.47 (t, J ) 8.6 Hz, 1H), 4.58
(d, J ) 11.8 Hz, 1H), 5.22-5.29 (m, 1H), 5.46-5.54 (m, 1H),
5.66 (s, 1H), 7.12 (d, J ) 6.0 Hz, 1H), 7.49 (d, J ) 3.5 Hz, 1H),
7.68 (t, J ) 7.3 Hz, 1H), 7.98 (t, J ) 7.0 Hz, 1H), 8.08 (d, J )
8.3 Hz, 1H), 8.21 (s, 1H), 8.35 (d, J ) 8.3 Hz, 1H), 9.08 (d, J
) 5 Hz, 1H); FAB HRMS m/z found 593.297620 (M + H)+,
calcd for C32H41N4O7 593.297525.
Ma cr ocyclic in h ibitor 35: white amorphous solid in >99%
homogeneity, as determined by analytical HPLC method A;
tR ) 20.5 min; 1H NMR (DMSO-d6) δ ∼0.9-1.7 (m, 17H), 1.16
(s, 9H), 2.30-2.43 (m, 1H), 2.55-2.65 (m, 1H), 3.86 (bd, J )
10.5 Hz, 1H), 3.97 (s, 3H), 4.11 (br, 1H), 4.51-4.55 (dd, J )
7.9, 8.9 Hz, 1H), 4.66 (d, J ) 13.7 Hz, 1H), 5.82 (br, 1H), 6.93
(d, J ) 5.7 Hz, 1H), 7.27 (br, 1H), 7.55 (s, 1H), 7.69 (bs, 4H),
8.17 (bs, 2H), 8.21 (d, J ) 9.5 Hz, 1H), 8.50 (s, 1H); ES+ HRMS
m/z found 701.354636 (M + H)+, calcd for C39H49N4O8
701.355040.
Ma cr ocyclic In h ibitor 33c. An sample of the methyl ester
derivative of compound 33a (20 mg, 0.033 mmol) in dry CH2-
Cl2 (1 mL) was stirred in the presence of 4 M HCl/dioxane (5
mL) for 1 h. The mixture was evaporated and dried carefully.
The residue was redissolved in CH2Cl2/DMF (3 mL/1 mL),
treated with NMM (14.5 µL, 0.132 mmol) and acetic anhydride
(7.0 µL, 0.073 mmol), and stirred at rt for 14 h. The mixture
was evaporated and dried under high vacuum. The residue
was then dissolved in a mixture of THF/MeOH/H2O (4 mL/2
mL/2 mL) and stirred overnight with LiOH‚2H2O (11 mg, 0.264
mmol). The residue isolated after acidification to pH ) 3 with
1 N ice-cold HCl was purified by C18 reversed-phase HPLC
using a solvent gradient from 0 to 40% aqueous CH3CN (0.06%
TFA) in order to isolated pure compound 33c as an amorphous
white solid in >90% homogeneity, as determined by analytical
HPLC method A: tR ) 12.0 min; 1H NMR (50 mM Na2PO4
buffer, pH ) 6.0, 600 MHz) δ 1.22-1.27 (m, 2H), 1.38-1.43
(m, 2H), 1.58-1.64 (m, 2H), 1.67-1.76 (m, 2H), 1.77-1.84 (m,
1H), 1.92-1.99 (m, 1H), 2.22-2.08 (m, 1H), 2.12-2.27 (m, 1H),
2.22-2.27 (m, 1H), 2.60-2.67 (m, 1H, Pro-â_), 2.83-2.89 (m,
1H, Pro-â), 4.32 (dd, J ) 12.1, 3.5 Hz, 1H, Pro-δ′), 4.41 (dd, J
) 12.1, 7.3 Hz, 1H), 4.56 (bd, J ) 8.0 Hz, 1H, Pro-δ), 4.62 (dd,
J ) 8.9 Hz, 1H, Pro-R), 5.40-5.46 (m, 1H), 5.55-5.61 (m, 1H),
5.73 (bs, 1H, Pro-γ), 7.41 (d, J ) 6.3 Hz, 1H), 7.64 (bs, 1H,
Acca-NH), 7.80 (dd, J ) 7.9 Hz, 1H), 8.03 (dd, J ) 8.0 Hz,
1H), 8.07 (d, J ) 9.5 Hz, 1H), 8.16 (d, J ) 7 Hz, 1H, AcNH),
8.36 (d, J ) 8.3 Hz, 1H), 8.90 (d, J ) 6.0 Hz, 1H); ES+ MS m/z
535.5 (M + H)+; ES- MS m/z 533.5 (M - H)-.
Ma cr ocyclic in h ibitor 36b: white amorphous solid in
>99% homogeneity, as determined by analytical HPLC method
A; tR ) 20.0 min; 1H NMR (400 MHz, DMSO-d6) δ 1.14 (s, 9H),
1.25-1.40 (bm, 7H), 1.45-1.51 (m, 2H), 1.65-1.75 (m, 2H),
2.13-2.19 (m, 1H), 2.33-2.42 (m, 1H), 2.56-2.70 (m, 2H), 3.87
(d, J ) 9.4 Hz, 1H), 3.97 (s, 3H), 4.00 (br, 1H), 4.46-4.50 (dd,
J ) 8.2 Hz, 1H), 4.72 (d, J ) 11.2 Hz, 1H), 5.24-5.29 (dd, J )
9.5 Hz, 1H), 5.49-5.56 (m, 1H), 5.82 (br, 1H), 7.12 (d, J ) 5.7
Hz, 1H), 7.24 (br, 1H), 7.53 (s, 1H), 7.68 (br, 4H), 8.15-8.25
(m, 3H), 8.67 (s, 1H); ES+ HRMS m/z found 699.339469 (M +
H)+, calcd for C39H47N4O8 699.339390.
Acyclic Liga n d 37. The P3 fragment 20a was saturated
under catalytic hydrogenation conditions. The Boc protecting
group of dipeptide 27c was removed under the usual acid
conditions, and the saturated derivative of 20a was coupled
to this dipeptide to produce the tripeptide ester of 37. After
saponification of the ester (under the usual basic conditions)
and purification of the acid products by C18 reversed-phase
HPLC, the pure ligand 37 was obtained as an amorphous
white solid in >99% homogeneity, as determined by analytical
1
HPLC method B: tR ) 5.5 min. H NMR (DMSO-d6, mixture
of rotamers in ∼1:7 ratio) δ (major rotamer) 0.86 (t, J ) 6.8
Hz, 3H), 1.05-1.12 (m, 1H), 1.22 (s, 9H), 1.15-1.55 (overlap-
ping m, 9H), 2.02 (dd, J ) 8.6, 17.1 Hz, 1H), 2.22-2.34 (m,
1H), 2.50-2.62 (m, 1H), 3.95 (s, 3H), 3.99 (bs, 1H), 4.13 (dd, J
) 8.3, 14.3 Hz, 1H), 4.35-4.47 (m, 2H), 5.07 (dd, J ) 1.2, 11.6
Hz, 1H), 5.20 (dd, J ) 1.2, 17.3 Hz, 1H), 5.63-5.75 (m, 2H),
7.04 (d, J ) 7.2 Hz, 1H), 7.18 (br, 1H), 7.46 (s, 1H), 7.50-7.68
(m, 4H), 8.13 (bs, 1H), 8.21 (bd, J ) 6.4 Hz, 2H), 8.45 (s, 1H);
ES+ HRMS m/z found (M + H)+ 701.354471, calcd for
Ma cr ocyclic in h ibitor 34a : white amorphous solid in
>97% homogeneity, as determined by analytical HPLC method
1
A; tR ) 19.7 min; H NMR (DMSO-d6) δ 1.17 (s, 9H), 1.16-
C
39H49N4O8 701.355040.
1.25 (m, 2H), 1.28-1.37 (m, 3H), 1.52-1.65 (m, 2H), 1.65-
1.83 (m, 3H), 1.92-2.05 (m, 1H), 2.05-2.20 (m, 2H), 2.39-
2.48 (m, 1H), 2.58-2.68 (m, 1H), 3.94 (br, 2H), 3.97 (s, 3H),
4.45-4.49 (dd, J ) 7.6, 9.5 Hz, 1H), 4.62 (brd, J ) 12.4 Hz,
1H), 5.22-5.29 (m, 1H, Hꢀ), 5.47-5.54 (m, 1H, Hη), 5.82 (br,
Ma cr ocyclic In h ibitor 39b. The saturated macrocyclic
methyl ester 39a was saponified under basic conditions and
purified by C18 reversed-phase HPLC to obtain the macrocyclic
inhibitor 39b as a white solid in >99% homogeneity, as
J . Org. Chem, Vol. 69, No. 19, 2004 6199

Goudreau et al.
1
determined by analytical HPLC method B: tR ) 5.1 min; H
a dielectric constant of 80. A total of 43 NMR-derived distance
restraints were generated from transferred-NOESY volume
buildup rates using the Assign module of Felix software
(Molecular Simulations, Inc., San Diego, CA). These NMR
distance restraints were applied as strong (1.8-2.5 Å), medium
(1.8-3.5 Å) or weak (1.8-5.0 Å) flat-bottomed potentials
having force constants of 15 kcal/mol‚Å2. Pseudoatoms defining
the centroids of the methyl groups were introduced in the
definition of the NOE restraints and the interproton distances
were corrected accordingly.56 A single, high-temperature un-
restrained dynamics run was performed at 900 K using a time
step of 1 fs, with 50 structures collected at 2 ps intervals to
generate a starting set of conformations. Each structure was
then retrieved, cooled, and minimized using the following
simulated annealing protocol. The temperature was initially
lowered to 750 K at a rate of 30 K/ps where only strong
restraints were applied. The remaining restraints were added
and additional cooling to first 500 K (25 K/ps) and then 300 K
(20 K/ps) was performed, followed by restrained minimization
(including cross-terms) to a final gradient of 0.005 kcal/mol.Å.
A total of 10 low energy, NMR-consistent, structures were
selected at the end of this protocol. These final 10 structures
are shown superimposed (P1-P3 backbone atoms only) in
Figure 3A. The root-mean-square deviation for the backbone
atoms of P1-P3 is 0.04 Å. None of these structures have
distance violations greater than 0.05 Å and the average total
restraint violation energy is 0.03 kcal/mol with a S.D. ) 0.01
kcal/mol.
NMR (DMSO-d6) δ 1.11 (s, 9H), 1.15-1.45 (m, 13H), 1.60-
1.90 (m, 3H), 2.35-2.41 (m, 1H), 2.64-2.68 (m, 1H), 3.96 (s,
3H), 4.05-4.11 (m, 1H), 4.14-4.20 (m, 1H), 4.31-4.35 (m, 2H),
4.50 (bd, J ) 10 Hz, 1H), 5.80 (bs, 1H), 7.18 (d, J ) 8 Hz, 1H),
7.26 (bd, J ) 9 Hz, 1H), 7.54 (bs, 1H), 7.68 (bs, 4H), 7.81 (d, J
) 8 Hz, 1H), 8.16 (d, J ) 4 Hz, 2H), 8.23 (bd, J ) 8 Hz, 1H);
FAB HRMS m/z found 689.355070 (M + H)+, calcd for
C
38H49N4O8 689.355040.
Com p u ta tion a l Ch em istr y Meth od s. A small fragment
(4) representing the P1 portion of our inhibitors was utilized
in quantum mechanics calculations to determine the rotational
minima on each potential energy surface. Semiempirical
calculations using Spartan 5.0 (Wave Function, Inc., Irvine,
CA) were performed using PM3-SM3. The φ and ψ angles of 4
were rotated systematically while the remaining parameters
were allowed to relax. Specifically, angles φ and ψ were varied
by 15° to generate 144 points. Symmetry was exploited when
possible. Each point was gradient minimized using the PM3-
SM3 and the energy was plotted as a function of angles φ and
ψ. Two energy minima (4a and 4b) were identified from the
above conformational energy profile. The global minimum
corresponded to conformation 4a and was characterized by a
φ angle of +98°. Another local minimum (4b), which was 1.1
kcal/mol higher in energy compared to 4a , was also identified
and was characterized by a φ angle of -98°.
NMR Sa m p le P r ep a r a tion . NMR samples were prepared
by adding 24 µL of concentrated solution of inhibitor 33c in
DMSO-d6 to an aqueous buffer composed of 50 mM Na2PO4,
300 mM NaCl, 3 mM dithiothreitol-d10, and 10% (v/v) D2O
(spiked with TSP) at pH 6.5. The final volume of these
solutions was 600 µL with inhibitor concentrations ranging
between 1 and 1.5 mM. Expression and purification of NS3
protease BK strain (spanning amino acids 1-180 and harbored
the C-terminal poly Lys solubilization motif ASKKKK8) which
was used in the NMR experiments was previously described.17
For the transferred NOESY experiments, a concentrated stock
solution of NS3 protease BK* poly Lys (9.2 mg/mL, ∼450 µM,
in a buffer identical to that previously described)17 was added
to these samples such that an inhibitor/protease ratio of 20:1
to 25:1 was typically achieved. For the competition experi-
ments, a stock solution of a potent hexapeptide 43 (IC50 < 1
nM)49 in DMSO-d6 (3.6 mM) was added to the samples
described above such that a 3-4-fold excess of potent inhibitor
vs protease was achieved.
NMR Meth od s. All spectra were acquired at 27 °C. Two-
dimensional (2D) double-quantum-filtered COSY (DQF-
COSY), TOCSY, NOESY, and ROESY spectra were acquired
using standard pulse sequences with time proportional phase
incrementation (TPPI) method at 600 MHz. Suppression of the
solvent signal was achieved by the use of presaturation or by
inserting a 3-9-19 WATERGATE module prior to data
acquisition. The 2D transferred-NOESY experiments were
recorded with mixing time of 70, 100, 150, and 200 ms. In these
experiments, a 25 ms spin-lock pulse was applied prior to t1
delay in order to eliminate protein background signals.55 The
ROESY experiment was recorded with a 300 ms spin-lock
period. The 2D data sets were typically acquired with 2048
points in t2, 350-512 points in t1, and 96-128 scans. The data
were processed and analyzed using XWinNMR and WinNMR
software (Bruker Canada, Milton, Ontario) and Felix software
(Molecular Simulations, Inc., San Diego, CA). Shifted sine-
bell apodization and zero-filling (2048 × 1048 real points) were
applied prior to Fourier transformation, and subsequent
baseline corrections were applied in one or both dimensions.
Restr a in ed Dyn a m ics. The NS3-bound conformation of
inhibitor 33c was modeled by a simulating annealing protocol
using Discover 98.0 and the CFF force field (Molecular
Simulations, Inc., San Diego, CA). The dynamics were per-
formed without cross-terms and nonbonded cutoffs and with
Dock in g P r otocol. The NS3 protease complex model of
inhibitor 33c was obtained from energy minimization and
molecular dynamics using Discover 98.0 and the CFF force
field (Molecular Simulations Inc., San Diego, CA). The dynam-
ics were performed without cross-terms and nonbonded cutoffs
and with a dielectric constant of 1. As a starting point for the
simulation, the NS3 protease bound conformation of compound
33c (as determined above) was docked into the substrate
binding region of the X-ray crystal structure of the apo NS3
protease.50a The inhibitor was appropriately oriented in order
to allow its P1 carboxylate group to make H-bonds with the
side chains of the catalytic triad residues H57 and S139, as
well as with the backbone NH of the oxyanion hole residues
S139 and G137 and in order to allow the formation of
additional hydrogen bonds between the backbone NH and CO
groups of the inhibitor P3 and the complementary backbone
groups of the protein A157 residue. The bound inhibitor was
then energy-minimized. The conformation of the protein, as
well as the P1 carboxylate anion of the inhibitor, were kept
fixed during this first minimization step. This initial complex
model was subsequently soaked in a 20 Å sphere of water. The
NS3 protein and the water molecules were then divided into
three zones and tethered to different extent during the
simulation. Zone 1: 0-10 Å away from the center of mass of
the inhibitor, Zone 2: 10-15 Å away from the inhibitor, Zone
3: >15 Å away from the inhibitor. The entire complex was
first submitted to energy-minimization and then to 1 ns of
dynamics at 298 K following a 200 ps equilibration step. Zone
3 was kept fixed during this entire simulation, while zone 2
and zone 1 were tethered using quadratic force constants of
10 and 1 kcal/mol‚Å2 respectively during the initial minimiza-
tion step. These tethering force constants were then gradually
reduced during the dynamic equilibration steps, and were
totally removed during the 1 ns dynamic run. In addition, the
inhibitor P1 carboxylate was also tethered using a quadratic
force constant of 1 kcal/mol‚Å2 during the initial steps. The
bond lengths and water molecule angles were kept fixed during
the entire dynamic simulation using rattle. At the end of the
dynamics the complex was further minimized (including cross-
terms) to a final gradient of 0.1 kcal/mol.Å.
(56) Wu¨thrich, K.; Billeter, M.; Braun, W.; Anglister, J . J . Biol. Mol.
1983, 169, 949.
(55) Scherf, T.; Anglister, J . A. Biophys. J . 1993, 64, 754.
6200 J . Org. Chem., Vol. 69, No. 19, 2004

Potent Inhibitors of the Hepatitis C Virus NS3 Protease
Ack n ow led gm en t. We thank Ms. Diane Thibeault
and Mr. Roger Maurice for the in vitro enzymatic assays
and Dr. Steven LaPlante and Mr. Norman Aubry for
assistance with some of the NMR data. We thank Dr.
Lynn Amon for valuable discussions regarding some of
the computational data. We thank Mr. Chris Bryan for
assistance in the preparation of some intermediates and
Ms. Vida Gorys for providing a sample of compound 3.
We also wish to thank Dr. Pierre Bonneau and Ms.
Ce´line Plouffe for purification of the NS3 protease
enzyme. We are also grateful to Dr. J ean Rancourt for
his advice on the synthesis of compound 10. We wish
to thank Drs. Pierre Beaulieu and J ean Rancourt for
assistance in proofreading the manuscript. Finally, we
thank Drs. Paul Anderson and Robert De´ziel for their
support of this project.
Su p p or tin g In for m a tion Ava ila ble: Data of intermedi-
ates 5b, 6b, 15a ,b, 20a , 21a ,b, 22a ,b, 23b, 25b, 27, 28a ,b,
29a -d , 30c, 31, 32b, 39a , 41a ,b, and 42. 1H NMR spectra
and decoupling experiments of the final inhibitors used in the
SAR studies (Scheme 2: 26, 29c,d , 30b, 32b, 33b,c, 34a ,b,
35, 36b, 37, 39b, 41b, and 42); the chemical shift assignments
(as indicated) were confirmed by multiple 2D NMR experi-
ments. 2D TRNOESY data relevant to the determination of
the NS3-bound conformation of ligand 33c, as well as the
competition study with hexapeptide 43 (Figure 1). This mate-
rial is available free of charge via the Internet at http://
pubs.acs.org.
J O049288R
J . Org. Chem, Vol. 69, No. 19, 2004 6201