Proline-based macrocyclic inhibitors of the hepatitis C virus: Stereoselective synthesis and biological activity

Article abstract of DOI:10.1002/anie.200501553

Life-saving donut: Macrocyclization through a Mitsunobu reaction was used to synthesize a 17-membered macrocycle. The bicyclic acetal core was prepared completely diastereoselectively. The macrocyclic peptidomimetic surrogate of the P2-P3 dipeptide moiety was designed to function as a hepatitis C virus (HCV) NS3 serine protease inhibitor, and the pentapeptide α-ketoamides derived from the macrocycle were shown to be potent HCV inhibitors. (Figure Presented).

Full text of DOI:10.1002/anie.200501553

Angewandte
Chemie
DOI: 10.1002/anie.200501553
Communications
A peptidomimetic surrogate of the P2–P3
dipeptide moiety acts as a potent hepatitis Cvirus (HCV)
inhibitor. This proline-based macrocycle binds tightly with the Ala156
methyl group of HCV NS3 protease while the P2’ phenyl and P1
n-propyl groups form a C-shaped clamp around the Lys136 side chain.
For more information, see the Communication by K. X. Chen et al. on
the following pages.
7024
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7024 – 7028

Angewandte
Chemie
Protease Inhibitors
acetal) and 3,4-dihydroxyproline, and the latter through aryl
alkyl ether formation under Mitsunobu reaction conditions.[7]
The synthesis started with osmium-catalyzed dihydroxy-
lation of commercially available N-Boc-3,4-dehydroproline
methyl ester (1; Scheme 1), to give an inseparable mixture of
DOI: 10.1002/anie.200501553
Proline-Based Macrocyclic Inhibitors of the
Hepatitis C Virus: Stereoselective Synthesis and
Biological Activity
Kevin X. Chen,* F. George Njoroge, Bancha Vibulbhan,
Andrew Prongay, John Pichardo, VincentMadison,
Alexei Buevich, and Tze-Ming Chan
The hepatitis C virus (HCV) is the leading cause of chronic
liver disease and has infected more than 170 million people
worldwide. It has emerged as a major public health threat.[1]
Currently, the standard treatment is a-interferon that is
attached to polyethylene glycol, in combination with the
antiviral agent ribavirin.[2] This therapy is effective only in
approximately 50% of patients and has considerable associ-
ated side effects. Given the prevalence of HCV infections,
there is an urgent need to develop more-effective, less-toxic,
and orally bioavailable small-molecule antiviral agents.
HCV is a single-stranded RNA genome that encodes a
polyprotein of about 3000 amino acids. The polyprotein is
processed into structural and nonstructural (NS) proteins by
host peptidases and virally encoded proteases. The NS3
protease, located towards the N-terminal side of the NS3
protein, is a chymotrypsin-like serine protease.[3] As a result
of its essential role in HCV replication, it has been an
attractive target for intensive research for new anti-HCV
therapy.[4] However, the fact that the NS3 protease has a
shallow and solvent-exposed substrate-binding region makes
it a formidable task to develop apropriate small-molecule
inhibitors.[5]
Scheme 1. a) OsO4, NMO, tBuOH/H2O/acetone/THF, RT, 77%;
b) BnO(CH2)3CHO, p-TsOH, MgSO4, CH2Cl2, RT, 91%; c) HCl (2m,
dioxane/EtOAc, RT, quant. NMO=N-methylmorpholine N-oxide,
Boc=tert-butyloxycarbonyl, p-TsOH= p-toluenesulfonic acid.
During the course of our research in potency optimization
and structure depeptization of peptide-based substrates of
HCV NS3 protease,[6] we were interested in the incorporation
of macrocyclic structures into our target molecules. The
bicyclic-acetal proline-based macrocycle 9 (Scheme 2) was
designed as a peptidomimetic surrogate for the P2–P3
dipeptide moiety on the substrate. We envisioned that the
additional contact of the macrocycle with the Ala156 methyl
group of the enzyme backbone would enhance binding of the
inhibitors. We anticipated two major challenges in the syn-
thesis: first, the stereoselective construction of 3,4-proline
cyclic acetal core 3 (Scheme 1), and second, macrocyclization
to the 17-membered ring 9. The former could be achieved
through an acid-catalyzed reaction between an aldehyde (or
3S,4R-diol 2 (major) and 3R,4S-diol (minor, structure not
shown).[8] Both diols were then treated with 4-benzyloxy-
butyraldehyde in the presence of a catalytic amount of p-
TsOH.[9] The product 3, and its isomer were isolated as a
mixture in excellent yield (91%) when 2 equivalents of the
aldehyde were used. The two isomers could be separated in a
ratio of 2:1, respectively, by column chromatography. 2D
1H NMR spectroscopy experiments were conducted to deter-
mine the absolute configuration of the products and the
stereochemical outcome of the acetal formation. To avoid
interference from rotamers in 3, it was necessary to perform
1
the H NMR spectroscopic analysis on amine 4 instead. The
Boc protecting group was removed selectively in the presence
of acid-sensitive acetal functionality when 3 was treated with
hydrochloric acid (2m). NOSEY experiments on compound 4
revealed the presence of NOE interactions between 6-H and
4-H, 7-H and 2-H, and 7-H and 5-Htrans. However, there was
no observed coupling between protons 2-H and 3-H, which
therefore indicates a trans relationship. This evidence led to
the structural assignment of 4, with 4A as its preferred
conformation. On the other hand, the amine derived from the
minor isomer demonstrated NOE signals between 7-H and 5-
[*] Dr. K. X. Chen, Dr. F. G. Njoroge, B. Vibulbhan, Dr. A. Prongay,
J. Pichardo, Dr. V. Madison, Dr. A. Buevich, Dr. T.-M. Chan
Schering-Plough Research Institute
2015 Galloping Hill Road, K-15-3/3545, Kenilworth, NJ 07033 (USA)
Fax : (+1)908-740-7152
E-mail: kevin.chen@spcorp.com
H
cis as well as 6-H and 3-H, but not between 7-H and 2-H. The
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
coupling constant between 2-H and 3-H (J = 5.2 Hz) suggests
Angew. Chem. Int. Ed. 2005, 44, 7024 –7028
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7025

Communications
a cis relationship. As a result, the structure 5 and its preferred
conformation 5Awas assigned. Significantly, the formation of
the acetal with both 3S,4R- and 3R,4S-dihydroxyprolines led
exclusively to the endo products.[10]
Once the stereochemistry was established, amine 4 was
then coupled to N-Boc-cyclohexylglycine under standard
conditions to give dipeptide 6 (Scheme 2). Selective removal
Scheme 3. a) respective amine, DhBtOH, ClCH2CH2Cl, NMM, DMF/
CH2Cl2, ꢀ208C, 65–95%; b) H2, Pd-C, EtOH, RT, 98%; c) 2m HCl,
dioxane/EtOAc, RT, quant. DhBtOH=3,4-dihydro-3-hydroxy-4-oxo-
1,2,3-benzotriazine, NMM=N-methylmorpholine, DMF=N,N-di-
methylformamide.
Scheme 2. a) N-Boc-Chg-OH, HATU, iPr2NEt, CH2Cl2, 08C!RT, 76%;
b) HCl (2m), dioxane/EtOAc, RT, quant.; c) 3-OH-PhCO2H, HATU,
iPr2NEt, CH2Cl2, 08C to RT, 98%; d) H2, Pd C 10%, EtOH/EtOAc, RT,
quant.; e) PPh3, ADDP, CH2Cl2, RT, 72%. Chg-OH= cyclohexylglycine,
HATU= O-(7-azabenzotriazol-1-yl)-N,N,N’N’-tetramethyluronium
hexafluorophosphate, ADDP=1,1’-(azodicarbonyl)dipiperidine.
corresponding a-ketoamides by using Dess–Martin period-
inane.[13] The products were mixtures of two diastereomers at
the P1–Nva/Nle a center. The tert-butyl ester isomers 18 and
21 and the dimethyl amides 20 and 23, were separated.
Treatment of 18 and 21 with TFA gave carboxylic acids 19 and
22, respectively. The allyl amide 24 and the P1–Nle analogue
25 were an inseparable mixture of diastereomers.
of the Boc protecting group and subsequent coupling with 3-
hydroxyphenylacetic acid provided compound 7. Cleavage of
the benzyl ether gave the phenol alcohol 8, which is the
substrate for the key Mitsunobu macrocyclization reaction.[11]
Treatment of 8 with triphenylphosphane and 1,1’-(azodicar-
bonyl)dipiperidine furnished the desired macrocycle 9 in
excellent yield (72%).
The “right-hand” segment (P1–P2’) of the molecule was
prepared through a series of peptide couplings and depro-
tections, starting with a-hydroxy-b-amino acid norvaline
(Nva) and norleucine (Nle) derivatives[12] (10 and 15,
respectively; Scheme 3). The hydrochloride intermediates of
the tripeptide amines 12, 13, 16, and allyl amide amine 14
were obtained as mixtures of four diastereomers.
All macrocyclic inhibitors 18–25 were tested in an HCV
continuous assay[14] (Table 1). The inhibition constants (Ki )
*
were determined by using NS4A-tethered single-strain NS3
serine protease.[15] The Ki values for compounds 18, 19, and
*
20 were 0.81, 0.36, and 0.21 mm, respectively. However, the
opposite diastereomers 21, 22, and 23 were much more potent
*
(Ki = 0.038, 0.038, and 0.044 mm, respectively). Based on the
structure–activity relationship results from our acyclic pep-
tide HCV protease inhibitors,[16] the more potent compounds
21–23 were assigned the S configuration at P1 and the less
potent compounds 18–20 the R configuration. The P1–Nle
analogue 25 was equally potent (0.076 mm) considering it was
a mixture of two isomers, whereas the smaller tripeptide
inhibitor 24 was only moderately active (3.1 mm).
The selectivity of these inhibitors against human neutra-
phil elastase (HNE), a structurally closely related serine
protease, was also measured (Table 1). Compounds 18–23
were moderately selective against elastase (HNE/HCV=
Finally, the macrocyclic “left-hand” segment and the a-
hydroxyamide “right-hand” moiety were combined
(Scheme 4). The methyl ester 9 was hydrolyzed to the
carboxylic acid 17, which was then coupled with amines 12–
14 and 16. The resulting hydroxyamides were oxidized to the
7026 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7024 –7028

Angewandte
Chemie
Figure 1. The X-ray crystal structure of compound 22 bound to HCV
protease.
the acetal center and the P1 a center. As expected a reversible
covalent-bond formation occurred between the carbonyl
group of the ketone and the hydroxyl group of Ser139.
Most importantly, the 17-membered macrocycle formed a
perfect “donut” that surrounded the Ala156 methyl group.
The P1 n-propyl side chain was buried in the S1 pocket and a
“C”-shaped clamp was formed around the side chain of
Lys136 along with the P2’ phenyl group. The P2 bicyclic
praline derivitive, the P3 cyclohexyl ring and the phenyl group
in the macrocycle all had hydrophobic interactions with the
protein surface. Furthermore, there were several hydrogen
bonds between the inhibitor peptide chain and that of the
enzyme backbone. All these interactions contributed to the
excellent potency of 22.
Scheme 4. a) LiOH, MeOH/THF/H2O, RT, 93%; b) 12–14 or 16,
DhBtOH, EDC, NMM, DMF/CH2Cl2, ꢀ208C, 50–85%; c) Dess–Martin
periodinane, CH2Cl2, RT, 51–89%; d) TFA, CH2Cl2, RT, quant. TFA=tri-
fluoroacetic acid
In summary, a concise and highly stereoselective synthesis
of the 17-membered macrocycle 9 was completed and the
endo product was shown to form exclusively at the acetal
center of the bicyclic proline core. The key macrocyclization
provided excellent yields through a Mitsunobu protocol. The
resultant macrocyclic inhibitors were active against HCV NS3
serine protease with the (S)-P1 isomers being the more
potent. The X-ray crystal structure of compound 22 bound to
the enzyme revealed that the macrocycle formed good
contact with the enzyme by adopting a “donut”-shaped
conformation above the methyl group of Ala156. Further
evaluation of these compounds and the design of inhibitors
with better potency and PK profiles are currently in progress.
Table 1: Potency and selectivity of compounds 18–25 against HCV
protease.
*
Compound
R
P1
(R)-Nva
(R)-Nva
(R)-NVa
(S)-Nva
(S)-Nva
(S)-Nva
(R,S)-Nva
(R,S)-Nle
Ki [mm]
HNE/HCV
18
19
20
21
22
23
24
25
OtBu
OH
NMe2
OtBu
OH
NMe2
–
NMe2
0.81ꢁ0.04
0.36ꢁ0.03
0.21ꢁ0.04
0.038ꢁ0.001
0.038ꢁ0.001
0.044ꢁ0.009
3.1ꢁ0.3
6
11
32
3
4
5
–
383
0.076ꢁ0.006
3:32), whereas the P1–Nle derivative 25 was highly selective
(383-fold). This indicates that the P1 residue plays an
important role in selectivity. Pharmacokinetic (PK) studies
in rats showed that compound 23 had a low oral AUC (area
under the curve) value (0.1 mm h@3 mpk) and bioavailability
(0.3%), whereas compound 24 gave a respectable AUC value
(2.6 mm h). Presumably, the smaller size and fewer amide
bonds of 24 contributed to the improved PK profile. The
solubility of 18–23 in pH 7.4 buffer was in the range of 5–
80 mm, with 19 and 20 at the higher end and 18, 21, 22, and 23
at the lower end.
Received: May 6, 2005
Revised: August 14, 2005
Published online: October 7, 2005
Keywords: antiviral agents · inhibitors · macrocycles · peptides ·
.
proteases
[1] J. Cohen, Science 1999, 285, 26.
[2] a) A. U. Neumann, N. P. Lam, H. Dahari, D. R. Gretch, T. E.
Wiley, T. J. Layden, A. S. Perelson, Science 1998, 282, 103;
b) A. M. Di Bisceglie, J. McHutchison, C. M. Rice, Hepatology
2002, 35, 224.
The X-ray crystallography structure of compound 22
bound to the active site of HCV protease was obtained
(Figure 1). It confirmed the absolute configurations at both
[3] A. M. Lesk, W. D. Fordham, J. Mol. Biol. 1996, 258, 501.
Angew. Chem. Int. Ed. 2005, 44, 7024 –7028
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 7027

Communications
[4] a) S.-L. Tan, Y. He, Y. Huang, M. Gale, Jr., Curr. Opin.
Pharmacol. 2004, 4, 465; b) C. Steinkuhler, U. Koch, F. Narjes,
V. G. Matassa, Curr. Med. Chem. 2001, 8, 919; c) S. Colarrusso,
U. Koch, B. Gerlach, C. Steinkuhler, R. De Francesco, S.
Atlamura, V. G. Matassa, F. Narjes, J. Med. Chem. 2003, 46,
345; d) M. Llinas-Brunet, M. D. Bailey, E. Ghiro, V. Gorys, T.
Halmos, M. Poirier, J. Rancourt, N. Goudreau, J. Med. Chem.
2004, 47, 6584; e) Y. S. Tsantrizos, G. Bolger, P. Bonneau, D. R.
Cameron, N. Gordreau, G. Kukolj, S. R. LaPlante, M. Llinas-
Brunet, H. Nar, D. Lamarre, Angew. Chem. 2003, 115, 1394;
Angew. Chem. Int. Ed. 2003, 42, 1355; f) E. S. Priestley, I.
De Lucca, B. Ghavimi, S. Erickson-Viitanen, Bioorg. Med.
Chem. Lett. 2002, 12, 3199.
[5] A. D. Kwong, J. L. Kim, G. Rao, D. Lipovsek, S. A. Raybuck,
Antiviral Res. 1998, 40, 1.
[6] M. Llinas-Brunet, M. D. Bailey, G. Fazal, S. Goulet, T. Halmos,
S. R. LePlante, R. Maurice, M. Poirier, M.-A. Poupart, D.
Thibeault, D. Wernic, D. Lamarre, Bioorg. Med. Chem. Lett.
1998, 8, 1713.
[7] a) D. L. Hughes, Org. React. 1992, 42, 335; b) O. Mitsunobu,
Synthesis 1981, 1.
[8] G. Guillerm, M. Varkados, S. Auvin, F. Le Goffic, Tetrahedron
Lett. 1987, 28, 535.
[9] M. H. Hopkins, L. E. Overman, G. M. Rishton, J. Am. Chem.
Soc. 1991, 113, 5354.
[10] a) M. Garcia-Valverde, R. Pedrosa, M. Vincente, Tetrahedron:
Asymmetry 1995, 6, 1787; b) X. Ariza, A. M. Costa, M. Faja, O.
Pineda, J. Vilarrasa, Org. Lett. 2000, 2, 2809; c) P. T. Kaye, W. E.
Molema, Synth. Commun. 1999, 29, 1889.
[11] K. X. Chen, F. G. Njoroge, B. Vibulbhan, A. Buevich, T.-M.
Chan, V. Girijavallahan, J. Org. Chem. 2002, 67, 2730.
[12] S. L. Harbeson, S. M. Abelleira, A. Akiyama, R. Barrett, R. M.
Carroll, J. A. Straub, J. N. Tkacz, C. Wu, G. F. Musso, J. Med.
Chem. 1994, 37, 2918.
[13] D. B. Dess, J. C. Martin, J. Am. Chem. Soc. 1991. 113, 7277.
[14] R. Zhang, B. M. Beyer, J. Durkin, R. Ingram, F. G. Njoroge,
W. T. Windsor, B. A. Malcolm, Anal. Biochem. 1999, 270, 268.
[15] S. S. Taremi, B. Beyer, M. Maher, N. Yao, W. Prosise, P. C.
Weber, B. A. Malcolm, Protein Sci. 1998, 7, 2143.
[16] Unpublished results from the Schering-Plough Research Insti-
tute.
7028 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 7024 –7028