Development of anti-coxsackievirus agents targeting 3C protease

Article abstract of DOI:10.1016/j.bmcl.2012.08.120

Peptidomimetic anti-viral agents against Coxsackievirus B3 (CVB3) were developed using a strategy involving the inhibition of 3C protease (CVB3 3C ^pro), a target for CVB3-mediated myocarditis or pericarditis. In an attempt to improve the inhibi

Full text of DOI:10.1016/j.bmcl.2012.08.120

Bioorganic & Medicinal Chemistry Letters 22 (2012) 6952–6956
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Development of anti-coxsackievirus agents targeting 3C protease
Bo-Kyoung Kim a, Jeong-Hyun Kim a, Na-Ri Kim a, Won-Gil Lee a, So-Deok Lee a, Soo-Hyeon Yun b,
a,
Eun-Seok Jeon b, , Yong-Chul Kim
⇑
⇑
a School of Life Sciences, Gwangju Institute of Science and Technology, 1 Oryong-dong, Buk-gu, Republic of Korea
b Division of Cardiology, Samsung Medical Center, Sungkyunkwan University, School of Medicine, 50 Irwon-dong, Gangnam-gu, Seoul, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 8 April 2012
Revised 16 July 2012
Accepted 30 August 2012
Available online 8 September 2012
Peptidomimetic anti-viral agents against Coxsackievirus B3 (CVB3) were developed using a strategy
involving the inhibition of 3C protease (CVB3 3Cpro), a target for CVB3-mediated myocarditis or pericar-
ditis. In an attempt to improve the inhibitory activity against CVB3, a variety of hetero-aromatic groups
were incorporated into the
a,b-unsaturated ester as Michael acceptor moiety, which is the position of
interaction with the cysteine moiety in the P1’ active site of CVB3 3Cpro. Among these hetero-aromatic
groups, the quinoline analogs 9c and 9e, with IC50 values of 250 and 130 nM as determined from an
enzyme assay, significantly inhibited the CVB3-mediated cell cytotoxicity, indicating parallel anti-viral
activities. A comparison of the binding modes of the potent inhibitor 9e and the relatively weak inhibitor
9n was explored in a molecular docking study, which revealed that compound 9n lacked hydrogen bonds
in its interactions with Gly129, 128, and 145.
Keywords:
Coxsackievirus B3
3C protease
Enzyme inhibitors
Ó 2012 Elsevier Ltd. All rights reserved.
Picornaviruses include important human pathogenic viruses,
such as enterovirus (EV), human poliovirus (HPV), hepatitis A virus
(HAV), human rhinovirus (HRV), and coxsackievirus (CV).1 Among
them, coxsackiviruses (CVs) are non-enveloped positive-sense
single-stranded RNA viruses2 and are the most common cause of
human viral myocarditis, which can progress to chronic myocardi-
tis and dilated cardiomyopathy.3,4
Upon infection, the positive-strand RNA genome of CVB3 is
translated into a large poly-protein precursor that is subject to suc-
cessive proteolytic cleavage processes to generate functional and
structural viral proteins.5 The proteolytic processing of poly-pro-
tein, essential for the production of new infectious virions, depends
on two virally encoded proteases 2A and 3C.6,7 In particular, the
picornavirus 3C protease is required for the majority of proteolytic
cleavage events during viral replication.8 These enzymes are
attractive targets for the development of therapeutic antiviral
agents because effective therapeutic strategies for the prevention
or treatment of diseases caused by CVB3 infection are not currently
available.9
Figure 1. Structure of compound 1 (AG7088, Rupintrivir).
Fig. 1), developed as a peptidomimetic inhibitor targeting HRV chy-
motrypsin protease (3Cpro), was evaluated in clinical trials12, but it
failed in a study of naturally infected patients.13 One of the draw-
backs of 1 was the need for multiple daily doses to achieve efficacy;
because of the ethyl ester moiety (Fig. 1) rapidly hydrolyzed to
yield an inactive acid form.14 The
a,b-unsaturated ester of 1 is an
Although there are no approved antiviral therapies for the treat-
ment of picornavirus infections in humans, numerous 3C protease
inhibitors have been developed as selective therapeutic agents for
picornavirus infections.10 For example, Pleconaril has been investi-
gated in clinical trials for the treatment of common cold symptoms
associated with the human rhinovirus (HRV), although it was not
important structural moiety and interacts with the Cys residue of
the P1’ active site in the 3C protease as a Michael acceptor to form
an irreversible covalent bond.8,15
Although CVB3 3Cpro is highly homologous to HRV 3Cpro in
terms of the amino acid sequence and crystal structure6, the P1’
pocket of the CVB3 3Cpro is larger and deeper than that of the
HRV 3Cpro. This structural difference was considered previously5,
in a study of potent and selective CVB3 3Cpro inhibitors.
In that work, the ethyl group was replaced with bulky groups,
such as phenyl, di-phenyl, or naphthyl groups.5 However, water
approved by the FDA.11 Compound
1 (AG7088, Rupintrivir,
⇑
Corresponding authors. Tel.: +82 62 715 2502; fax: +82 62 715 2484.
E-mail addresses: yongchul@gist.ac.kr (Y.-C. Kim), esjeon@skku.edu (E.-S. Jeon).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.08.120

B.-K. Kim et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6952–6956
6953
Scheme 1. Reagents and conditions: (a) NMP, 2 M DIEA, 16 h; (b) 20% piperidine; (c) NMP, 2 M HOBt and DIC, 3 h; (d) 20% piperidine; (e) NMP, 2 M HOBt and DIC; (f) 2% TIS,
DCM; * 87% overall yield for 6 steps.
solubility of the most potent analog, 9a, with a naphthyl group was
too poor for in-vivo animal experiments. To overcome the solubil-
ity problems and improve the inhibitory potency against CVB3
3Cpro and the virus, we substituted a variety of hetero-aromatic
groups in place of the naphthyl group, such as quinoline (9c, 9e,
9f), isoquinoline (9b, 9d), benzimidazole (9g, 9h, 9i), indazole (9j,
9k), methyl-imidazo-pyridine ( 9l, 9m), and oxadiazolo-pyridine
(9n). Herein, we report the development of new anti-coxsackievi-
rus agents, along with a structure–activity relationship study. A
CVB3 3Cpro enzyme assay, an anti-CVB3 cell based assay, and a
molecular docking study were performed to understand the mode
shown in Figure 1. Fmoc-L-4-fluoro-phenylalanine was reacted
with trityl-resin, 2 (2-chloro-trityl chloride resin), and the Fmoc
was removed by 20% piperidine. The amino acid-loaded resin 3
was reacted with Fmoc-L-valine in the presence of HOBt and DIC
in NMP, and the Fmoc was removed by 20% piperidine. After the
reaction of 5-methyl-isoxazole-3-carboxylic acid with the dipep-
tide-loaded resin, 4, cleavage from the resin using 2% triethyliso-
propylsilane (TIS) afforded compound 5, which was subsequently
purified by HPLC.
The Michael acceptor moiety was synthesized by reacting phos-
phonoacetate compounds with various hetero-aromatic groups.
Compounds 6a–n were reacted with compound 7 (Boc-Gln(Trt)-
aldehyde)5 under the Wittig–Horner reaction conditions. The sub-
sequent deprotection of the Boc group with 3 N HCl in 1,4-dioxane
resulted in compounds 8a–n (Scheme 2). The final compounds 9a–
n were synthesized from general coupling reactions of compound 5
of binding to the active site of CVB3 3Cpro
.
The general synthetic approach to prepare the hetero-aromatic
group-containing peptidomimetic inhibitors (9a–n) is described in
Schemes 1 and 2. The P2-P3 recognition motif, compound 5, was
synthesized using solid phase peptide synthesis methods, as
Scheme 2. Reagents and conditions: (g) Sodium bis(trimethylsilyl)amide, THF, –78 °C, 2.5 h, 65% yield; (h) 3 N HCl in 1,4-dioxane, rt, 1.5 h, 30–40% yield; (i) PyBOP
(1.5 equiv), DIPEA (1.2 equiv), CH2Cl2, rt, 1–8 h, 80% yield; (j) 25% TFA in CH2Cl2 25–30% yield. The R groups are shown in Table 1.

6954
B.-K. Kim et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6952–6956
Table 1
with compounds 8a–n using PyBOP in DCM followed by removal
of the trityl groups using 25% TFA.
IC50 values of compounds 9 (a–n) tested against CVB3 3Cpro
The inhibitory activity of compounds 5, 9a–n against CVB3
3Cpro was first tested in a FRET-based enzyme assay using a fluoro-
genic substrate NMA-EALFQGPPVK-DNP.16 Compound 5 and 9a
were used as a negative and positive control, respectively.17 The
structures and inhibitory activities of the peptidomimetic inhibi-
tors are summarized in Table 1. Initially, we investigated the rela-
tionship between the relative positions of the nitrogen atoms in
the quinoline-based hetero-aromatic rings because the inhibitory
activity may be affected by the position of the ester group in the
active site. In general, the quinoline ( 9b, 9d) and isoquinoline
(9c, 9e, 9f) derivatives showed higher inhibitory activities than
the parent compound 9a, having a naphthyl moiety, suggesting
that additional interactions or more stable binding to the quinoline
moiety in the active site of CVB3 3Cpro may improve the inhibitory
activity. Among the various quinoline and isoquinoline derivatives
tested with different nitrogen atom positions, the 4-yl quinoline
compound 9e showed the most potent activity with an IC50 value
of 130 nM.
The inhibitory activities of the benzimidazole analogs were sen-
sitive to the substitution positions. For example, the activities of
the chloroimidazole (9h) and 5-yl benzimidazole (9i) analogs were
significantly reduced compared with the 2-yl benzimidazole deriv-
ative, 9g.
Similarly, the position of substitution of the indazole moiety
greatly affected the inhibitory activity, as indicated by the 390
and 180 nM IC50 values for 9j (6-yl indazol) and 9k (5-yl indazol),
respectively.
compd
R
IC50 (lM)
9a
0.41 0.01a
0.24 0.02
9b
9c
0.25 0.02
0.27 0.04
9d
9e
0.13 0.01
0.28 0.02
The results indicated that the relative positions of the nitrogen
atoms on the heterocyclic moieties seemed to be important for the
recognition of the binding partner in the 3Cpro active site. In the
case of methyl imidazo-pyridine, two region-isomers, 9l and 9m,
showed comparable IC50 values in the range of 170–180 nM. An ex-
tended form of the oxadiazo-pyridine analog, 9n, showed weak
activity, with an IC50 value of 650 nM, indicating that the S1’ pock-
et of the active site may not be large or deep enough to accommo-
date longer derivatives, such as 9n. Overall, most of the derivatives
with hetero-aromatic groups showed comparable or better inhibi-
tory activities relative to the parent compound, 9a, except for the
chlorobenzimidazole (9h) and oxadiazo-pyridine (9n) analogs.
The binding modes of new inhibitors were investigated by per-
forming a molecular docking study of the most potent inhibitor, 9e,
and relatively weak inhibitor, 9n. Molecular docking of the inhibi-
tors into the active site of CVB3 3Cpro, based on its X-ray crystal
structure (PDB: 2ZU3), was performed using a CHARMm force
field-based docking tool, CDOKER, from Discovery Studio 2.5
(Accelrys Inc., San Diego, CA). The structure of the binding site
was held rigid, and the ligand conformations were randomly orien-
tated. The formation of a covalent C–S bond between the b-carbon
of the Michael acceptor of the ligand and the Cys147 residue of the
S1’ binding site was observed. Final poses of optimized docked
structures were sorted according to the CHARMm energy.
The docking study results are shown in Figure 2. Figure 2A
shows that the oxygen and nitrogen atoms of the isoxazole ring
9f
9g
9h
9i
0.24 0.02
0.65 0.02
0.47 0.05
0.39 0.04
0.18 0.02
9j
9k
9l
0.17 0.01
of compound 9e formed hydrogen bonds with Gly129 and Gly128
.
The 4-fluoro phenyl moiety fit well into the S2 pocket by forming
hydrophobic interactions with the neighboring amino acid resi-
dues. Each carbonyl-oxygen of the ligand formed a hydrogen bond
with the neighboring amino acids in the binding pocket, including
Gly164 and Gly145, and the 4-yl quinoline moiety fit well into the S1’
pocket.
The docking mode of the weak inhibitor 9e did not form hydro-
gen bonds in the active site (Fig. 2B), except in an interaction with
Gly164. A superimposition of the two docking modes (Fig. 2C) re-
vealed that the 4-fluorophenyl, valine, and glutamine moieties of
9m
9n
0.16 0.01
0.65 0.02
a
Compound 9a was previously reported in Ref. 5.

B.-K. Kim et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6952–6956
6955
Figure 2. Molecular docking study of compounds 9e (A) and 9n (B). The nitrogen, oxygen, and sulfur atoms are colored blue, red, and orange. Hydrogen bonds are displayed
as green dashed lines. The docking of 9e and 9n are comparatively docked in the active site (C). 9e and 9n are colored yellow and green, respectively.
Figure 3. Antiviral activity assay results. The compounds were tested at various concentrations (0.5–50 lM). The concentrations are color-coded.
9e and 9n were similarly accommodated. However, differences in
the binding of the isoxazole rings in 9e and 9n clearly affected
the hydrogen bonds with Gly129 and Gly128 in the active site. More-
over, no hydrogen bonds between the carbonyl-oxygen of the
inhibitor and Gly145 were observed.
Affairs, Republic of Korea; (A090125) and by a grant from the Insti-
tute of Medical System Engineering (iMSE) in the GIST, Korea.
References and notes
Finally, the antiviral activity of each compound was tested in a
cell viability assay using HeLa-UVM cells, which were infected with
the Coxsackievirus (CVB3-H3).18 The results of the cell based anti-
viral activities of the compounds are shown in Figure 3. All inhib-
1. Isak, Im; Lee, E. S.; Choi, S. J.; Lee, J.-Y.; Kim, Y. C. Bioorg. Med. Chem. Lett. 2009,
19, 3632.
2. De Palma, A. M.; Vliegen, I.; De Clerq, E.; Neyts, J. Med. Res. Rev. 2008, 28, 823.
3. Feldman, AM.; McNamara, D. NEJM 2000, 343, 1388.
4. D’Ambrosio, A.; Patti, G.; Manzoli, A., et al Heart 2001, 85, 499–504.
5. Lee, E. S.; Lee, W. G.; Yun, S. H.; Rho, S. H.; Im, I. S.; Yang, S. T.; Sellamuthu, S.; Lee,
Y. J.; Kwon, S. J.; Park, O. K.; Jeon, E. S.; Park, W. J.; Kim, Y. C. BBRC 2007, 358, 7.
6. Zell, R.; Markgraf, R.; Schmidtke, M.; Gorlach, M.; Stelzner, A.; Henke, A.;
Sigusch, S. S.; Gluck, B. Med. Microbiol. Immunol. 2004, 193, 91.
7. Tong, L. Chem. Rev. 2002, 102, 4609.
8. Binford, S. L.; Maldonado, F.; Brothers, M. A.; Weady, P. T.; Zalman, L. S.;
Meador, J. W.; Matthews, D. A.; Patrick, A. K. Antimicrob. Agents Chemother.
2005, 49, 619.
9. Lall, M. S.; Jain, R. P.; Vederas, J. C. Curr. Top. Med. Chem. 2004, 4, 1239.
10. Barnard, D. L. Curr. Pharm. Des. 2006, 12, 1379.
11. Senior, K. Lancet Infect. Dis. 2002, 2, 264.
12. Patick, A. K.; Binford, S. L.; Brothers, M. A.; Jackson, R. L.; Ford, C. E.; Diem, M.
D.; Maldonado, F.; Dragovich, P. S.; Zhou, R.; Prins, T. J.; Fuhrman, S. A.; Meador,
J. W.; Zalman, L. S.; Matthews, D. A.; Worland, S. T. Antimicrob. Agents
Chemother. 1999, 43, 2444.
itors displayed almost 50% cell viabilities at 1 lM against the CVB3
infected cells. Among them, compounds 9c and 9e showed partic-
ularly potent anti-viral activities, with low inhibitory concentra-
tions at 50% cell viability. The water solubility of an HCl salt
form of the novel potent heterocyclic compound 9k (5-yl indazol)
was determined by HPLC. The water solubility of this salt form was
a factor of 20 greater than that of compound 9a (data not shown).
The novel CVB 3Cpro inhibitors are currently being tested in an
in vivo anti-viral efficacy study.
In conclusion, we developed novel potent CVB 3Cpro inhibitors
based on heterocyclic moieties, including quinoline, benzimidazo-
line, and methyl imidazo-pyridine. The compounds were effective
in enzymatic assays and cell-based anti-viral assays, displaying
sub-micromolar IC50 values. The binding modes of the analogs in
the 3Cpro active site, determined by molecular docking studies,
suggested that the quinoline-substituted analog maintained sev-
eral important hydrogen bonds and hydrophobic interactions in
the active site.
13. Patick, A. K. Antiviral Res. 2006, 71, 391.
14. Hsyu, P.-H.; Pithavala, Y. K.; Gersten, M.; Penning, C. A.; Kerr, B. A. Antimicrob.
Agents Chemother. 2002, 46, 392.
15. Johnson, T. O.; Hua, Y.; Luu, H. T.; Brown, E. L.; Chan, F.; Chu, S. S.; Dragovich, P.
S.; Eastman, B. W.; Ferre, R. A.; Fuhrman, S. A.; Hendrickson, T. F.; Maldonado,
F. C.; Matthews, D. A.; Meador, J. W.; Patick, A. K.; Reich, S. H.; Skalitzky, D. J.;
Worland, S. T.; Yang, M.; Zalman, L. S. J. Med. Chem. 2002, 45, 2016.
16. The CVB3 3Cpro coding region was amplified by PCR. The pTYB12 expression
vector was used to subclone, and the gene was transformed into Escherichia coli
strain BL21(DE3). CVB3 3Cpro was purified using the IMPACT-CN system (New
England Biolabs, Beverly, MA). The protease stock solution was maintained in
20 mM HEPES, 100 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol (pH 7.9) and
glycerol at –20 °C. These enzyme assays were performed in 96-well microplates
Acknowledgments
This study was supported by a grant of the Korea Healthcare
Technology R&D Project, Ministry for Health, Welfare & Family
in reaction volumes of 100
lL with 50 lL reaction buffer, 44.5 lL diluted water,
1.5 L CVB3 3Cpro, 1 L of various concentrations of the test compounds (5, 9
l
l

6956
a–n), and 3
B.-K. Kim et al. / Bioorg. Med. Chem. Lett. 22 (2012) 6952–6956
l
L substrate peptide. The reaction buffer was composed of 80 mM
17. The % inhibition of compound 5 at 10
33.5%.
lM in the FRET enzyme assay was only
Tris–HCl, 2 mM dithiothreitol, and 0.2% IGEPAL (pH 8.0). Inhibitors and
enzymes were incubated for 5 min in the reaction buffer at room
temperature, and reactions were initiated by the addition of a fluorogenic
substrate, NMA-EALFQGPPVK-DNP for CVB3 3Cpro, which was purchased from
Peptron (Daejeon, Republic of Korea). The reactions were incubated at 30 °C, for
1 h. The fluorescence values were then measured at 340 nm (excitation) and
440 nm (emission) and were monitored by spectrofluorometer. All compounds
18. HeLa-UVM cells were infected with CVB3–H3 at a multiplicity of infection
(MOI) of 100. The CVB3 3Cpro inhibitors (9a–9n) were added to the cultures at
various concentrations (0.5–50
end of the incubation period, 15
the Cell Counting Kit 8 (Oojindo Lab, Tokyo, Japan), was added. Only viable cells
with intact mitochondria were stained by this reagent. The absorbance at
450 nm was measured using an ELISA plate reader. Non-infected HeLa-UVM
cells were used as controls, and proliferation of these cultures was defined as
100%. Data are presented as the mean SD of results from three independent
experiments.
l
l
M), followed by incubation for 6–24 h. At the
L of the proliferation dye reagent, supplied in
were initially screened at 1 lM. The mean % inhibition was calculated based on
triplicate or quadruplicate experimental results. The IC50 values of all inhibitors
were determined from the concentration-dependent curves. Three independent
measurements were performed at each concentration.