Discovery of naturally occurring aurones that are potent allosteric inhibitors of hepatitis C virus RNA-dependent RNA polymerase

Article abstract of DOI:10.1021/jm200242p

We have identified naturally occurring 2-benzylidenebenzofuran-3-ones (aurones) as new templates for non-nucleoside hepatitis C virus (HCV) RNA-dependent RNA polymerase (RdRp) inhibitors. The aurone target site, identified by site-directed mutagenesis, is located in thumb pocket I of HCV RdRp. The RdRp inhibitory activity of 42 aurones was rationally explored in an enzyme assay. Molecular docking studies were used to determine how aurones bind to HCV RdRp and to predict their range of inhibitory activity. Seven aurone derivatives were found to have potent inhibitory effects on HCV RdRp, with IC₅₀ below 5 μM and excellent selectivity index (inhibition activity versus cellular cytotoxicity). The most active aurone analogue was (Z)-2-((1-butyl-1H-indol-3-yl)methylene)-4,6-dihydroxybenzofuran-3(2H)-one (compound 51), with an IC₅₀ of 2.2 μM. Their potent RdRp inhibitory activity and their low toxicity make these molecules attractive candidates as direct-acting anti-HCV agents.

Full text of DOI:10.1021/jm200242p

ARTICLE
pubs.acs.org/jmc
Discovery of Naturally Occurring Aurones That Are Potent Allosteric
Inhibitors of Hepatitis C Virus RNA-Dependent RNA Polymerase
Romain Haudecoeur,†,# Abdelhakim Ahmed-Belkacem,‡,# Wei Yi,† Antoine Fortunꢀe,† Rozenn Brillet,‡
||
Catherine Belle,§ Edwige Nicolle,† Coralie Pallier,‡, Jean-Michel Pawlotsky,‡,^,∞ and
Ahcꢁene Boumendjel*,†,∞
†Dꢀepartement de Pharmacochimie Molꢀeculaire, Universitꢀe de Grenoble, CNRS, UMR 5063, BP 53, 38041 Grenoble Cedex 9, France
‡H^opital Henri Mondor, INSERM U955, 51 Avenue du Marꢀechal de Lattre de Tassigny, 94010 Crꢀeteil, France
§Dꢀepartement de Chimie Molꢀeculaire, Universitꢀe de Grenoble, CNRS, UMR 5250, BP 53, 38041 Grenoble Cedex 9, France
Department of Virology, H^opital de Bic^etre, Le Kremlin-Bic^etre, France
^Department of Virology, H^opital Henri Mondor, National Reference Center for Viral Hepatitis B, C and Delta, Universitꢀe Paris-Est,
51 Avenue du Marꢀechal de Lattre de Tassigny, 94010 Crꢀeteil, France
S Supporting Information
b
ABSTRACT: We have identified naturally occurring 2-benzy-
lidenebenzofuran-3-ones (aurones) as new templates for non-
nucleoside hepatitis C virus (HCV) RNA-dependent RNA
polymerase (RdRp) inhibitors. The aurone target site, identified
by site-directed mutagenesis, is located in thumb pocket I of
HCV RdRp. The RdRp inhibitory activity of 42 aurones was
rationally explored in an enzyme assay. Molecular docking
studies were used to determine how aurones bind to HCV
RdRp and to predict their range of inhibitory activity. Seven
aurone derivatives were found to have potent inhibitory effects
on HCV RdRp, with IC50 below 5 μM and excellent selectivity
index (inhibition activity versus cellular cytotoxicity). The most
active aurone analogue was (Z)-2-((1-butyl-1H-indol-3-yl)methylene)-4,6-dihydroxybenzofuran-3(2H)-one (compound 51), with
an IC50 of 2.2 μM. Their potent RdRp inhibitory activity and their low toxicity make these molecules attractive candidates as direct-
acting anti-HCV agents.
’ INTRODUCTION
ribavirin has failed to eradicate the infection. Thus, as a pre-
ventive vaccine will not be available for several years, there is an
urgent need for more effective anti-HCV drugs.
Hepatitis C virus (HCV) infection is a global public health
problem. The World Health Organization (WHO) estimates
that 120ꢀ140 million people are infected worldwide (2ꢀ3% of
the world population), while 3ꢀ4 million are newly infected each
year. Chronic HCV infection is associated with chronic liver
inflammation and fibrosis. In the absence of antiviral treatment,
HCV liver disease progresses to cirrhosis in 20ꢀ30% of chroni-
cally infected patients, and patients with cirrhosis have a 1ꢀ4%
annual risk of developing hepatocellular carcinoma.1 HCV
infection is curable. The current standard of care is a combination
of pegylated interferon R (peg-IFN) and ribavirin (RBV).2ꢀ5
Unfortunately, the response rate is low, especially among
patients infected by HCV genotype 1, the most frequent geno-
type in the U.S., Europe, Asia, and Latin America (60ꢀ75% of
cases).6 In 2012, a new treatment based on a combination of peg-
IFN, ribavirin, and a direct-acting antiviral drug that inhibits
HCV serine protease activity (telaprevir or boceprevir) will be
available for patients with HCV genotype 1 infection. Yet these
new therapies will fail in 20ꢀ30% of treatment-naive patients and
in about 50% of patients in whom a first course of peg-IFN and
The HCV RNA-dependent RNA polymerase (RdRp) is a
particularly attractive target because of its key role in viral replication
and the fact that this enzyme has no functional equivalent in
mammalian cells. A number of HCV RdRp inhibitors have been
identified, including nucleoside/nucleotide analogues that target the
active site, and non-nucleoside allosteric inhibitors.7,8 The HCV
RdRp structure resembles a right hand. Four allosteric binding sites
have been identified so far, including “thumb” pockets I and II and
“palm” pockets I and II.8,9 To date, two classes of molecules sharing a
number of similarities have been reported to bind to thumb pocket I,
including 2-aryl-1-cyclohexylbenzimidazole-5-carboxylic acid deriva-
tives and their indole-based equivalents (Figure 1).10ꢀ17 Some of
these compounds are currently being tested in HCV-infected patients.
For example, a phase I clinical trial of MK-3281, an indole-based
thumb pocket I inhibitor,17 has recently been completed.18
Received: March 1, 2011
Published: June 23, 2011
r
2011 American Chemical Society
5395
dx.doi.org/10.1021/jm200242p J. Med. Chem. 2011, 54, 5395–5402
|

Journal of Medicinal Chemistry
ARTICLE
Scheme 2. Synthesis of Benzofuran-3(2H)-one Derivatives
2ꢀ9 as the Key Starting Blocksa
Figure 1. Two examples of known thumb pocket I inhibitors (top),16,17
general structure of the aurone backbone (bottom left), and structure of
naturally occurring aureusidin (compound 1, bottom right).
Scheme 1. Synthesis of Aurone Derivatives, Starting from
Building Blocks 2ꢀ9a
a Reagents and conditions: (a) ClCH2CN, HCl, ZnCl2, Et2O, then HCl,
H2O, 100 ꢀC, 1 h; (b) MeONa, MeOH, 60 ꢀC, 1.5 h; (c) trimethyl-
phenylammonium tribromide, THF, room temp, 6 h; (d) KOH, MeOH,
60 ꢀC, 3 h; (e) Me2SO4, K2CO3, dimethoxyethane, 80 ꢀC, 3 h; (f) BnBr,
K2CO3, dimethoxyethane, room temp, 3 days; (g) chloroacetic acid,
NaOH, H2O, 90 ꢀC, 18 h; (h) SOCl2, PhMe, 110 ꢀC, 2 h, then AlCl3,
CH2Cl2, 40 ꢀC, 18 h; (i) CuBr2, EtOAc, CHCl3, 60 ꢀC, 18 h; (j) K2CO3,
DMF, 0 ꢀC, 1.5 h.
a Reagents and conditions. (a) For procedure A: KOH/MeOH, 60 ꢀC,
1 ꢀ 18 h. For procedure B: KOH/EtOH, 80 ꢀC, 2ꢀ5 h. For procedure
C: Al2O3, CH2Cl2, room temp, 16 h. (b) BBr3, CH2Cl2, 0 ꢀC to room
temp, 24ꢀ72 h.
obtained with a softer method using neutral alumina, as described
by Varma et al., in order to avoid nucleophilic substitutions by
in situ formed potassium methoxide.22 4,6-Dihydroxyaurones
were synthesized by preparing the corresponding 4,6-dimethoxy
analogues, followed by deprotection with boron tribromide. In
some cases it was possible to obtain these derivatives directly by
simple basic condensation in KOH/EtOH (Scheme 1).
These pathways required the synthesis of benzofuran-
3(2H)-ones 2ꢀ9 by various methods, as described elsewhere
(Scheme 2). Compounds 2 and 3 were synthesized from
resorcinol and phloroglucinol, respectively. By use of chloroace-
tonitrile in HCl/Et2O, electrophilic addition in the ortho posi-
tion from one of the hydroxyl groups allowed the recovery of the
chloroacetophenone. Cyclization then readily occurred in
NaOMe/MeOH.23 Further reaction of 3 with methyl sulfate
(Me2SO4) or benzyl bromide (BnBr) in the presence of K2CO3
yielded 6 or 7, respectively.24 Compounds 4 and 5 were obtained
from the corresponding diacetoxyacetophenones after bromina-
tion of the R-position of the ketone with triphenylmethylammo-
nium tribromide and then one-pot deprotection and cyclization
Aurones (2-benzylidenebenzofuran-3-ones) are naturally oc-
curring small molecules belonging to the flavonoid family, the
pharmacological potential of which we reported for the first time
several years ago.19 As part of an ongoing program investigating
the therapeutic potential of aurones, we discovered that aureusidin
(30,4,40,6-tetrahydroxyaurone, 1, Figure 1),20,21 a naturally occur-
ring aurone, potently inhibits HCV RdRp with an IC50 of 5.2 μM.
Interestingly, the cellular toxicity of aurone 1 against mammalian
cells occurs at very higher concentration (>400 μM). Mutagenic
studies and molecular modeling identified thumb pocket I as the
auronebinding site. Onthebasis ofthese preliminarydata, weused
the aurone scaffold as a template for the design of aurone
analogues with potent inhibitory activity on HCV RdRp.
’ CHEMISTRY
The aurone derivatives described in this study were synthesized
by means of aldol condensation of a substituted benzofuran-
3(2H)-one with a benzaldehyde derivative, in basic KOH/MeOH
conditions under reflux. Aurones bearing fluoro groups were
5396
dx.doi.org/10.1021/jm200242p |J. Med. Chem. 2011, 54, 5395–5402

Journal of Medicinal Chemistry
ARTICLE
Table 1. Initial Array of NS5B Inhibition by Aureusidin Analogues (Compounds 10ꢀ20)
in KOH/MeOH.25 The same approach was used for compound
9 but with direct bromination of the 2-fluoro-6-hydroxyaceto-
phenone with CuBr2, then cyclization in a K2CO3/DMF
medium. Compound 8 was obtained from 3,5-difluorophenol,
which was alkylated with chloroacetic acid. Subsequent chlora-
tion of the acid, followed by FriedelꢀCrafts cyclization, yielded
the desired compound.27
Benzaldehyde derivatives were purchased from commercial
sources or synthesized using published approaches. 4-Cyclohex-
ylbenzaldehyde was synthesized from cyclohexylbenzene in a
formylation reaction, using hexamethylenetetramine in TFA.27
1-Butyl-1H-indole-3-carbaldehyde was obtained from indole-3-
carbaldehyde and butyl bromide, using NaH in DMF.
B-ring, considering the 4-hydroxy derivatives, position 40 ap-
peared to be the most beneficial for activity (IC50 = 5.9 μM for
11), since the 20-hydroxylated analogue led to a 2-fold loss of
activity (IC50 = 11.1 μM for 18) and the 30-hydroxylated deri-
vative was almost inactive (IC50 > 20 μM for 15). These
preliminary results designed a substitution pattern in which the
positions 4, 6, and 40 constituted the most promising ones for
further explorations.
In order to check the antiviral potential of aurones, we
evaluated the inhibition effect of aurones 11 and 12 on the
HCV replication in JFH1 (HCV) cell based assays. Aurones 11
and 12 showed IC50 of 18 ( 7 and 40 ( 4 μM, respectively.
The next step was to identify the aurone binding site in HCV
RdRp. Four sites, thumb pockets I and II and palm pockets I and
II, have been identified through X-ray structure of crystallized
complexes as binding sites for non-nucleoside allosteric inhibi-
tors of HCV RdRp (Figure 2). Amino acid substitutions at these
sites have been shown to reduce susceptibility to the correspond-
ing inhibitors. In order to determine if aurones target one of these
sites, we examined whether these substitutions reduced HCV
RdRp susceptibility to aurones in our cell-free enzyme assay. For
this purpose we engineered HCV-NS5BΔ21 carrying the P495L,
M423T, C316Y, or H95Q substitutions that are known to confer
resistance to reported inhibitors.29 The experiments were carried
out with (Z)-2-(4-hydroxybenzylidene)-4-hydroxybenzofuran-
3(2H)-one (11). The IC50 values of 11 for each HCV-NS5BΔ21
mutant were compared with the IC50 for wild-type HCV-
NS5BΔ21, and the results were recorded as fold differences
(Figure 3). A benzimidazole, 2-[4-[[4-(acetylamino)-40-chloro-
[1,10-biphenyl]-2-yl]methoxy]phenyl]-1-cyclohexyl-1H-benzi-
midazole-5-carboxylic acid, an inhibitor of HCV RdRp known to
target thumb pocket I, was used as a control.30 As expected, this
latter compound was 10-fold less potent against HCV-NS5BΔ21-
P495L than against HCV-NS5BΔ21-Wt, whereas its activity
was not affected by the other amino acid substitutions. The
inhibitory activity of aurone 11 was therefore only affected by
the P495L substitution (as a 6- to 7-fold loss of activity),
indirectly suggesting that this compound binds to (or close to)
thumb pocket I. However, the lack of X-ray structure of an
’ RESULTS AND DISCUSSION
An enzyme assay was used to assess the inhibition of the
polymerase activity of a purified RdRp (nonstructural 5B
[NS5B] protein), deleted of its 21 C-terminal amino acids in
order to ensure solubility (HCV-NS5BΔ21). The J4 genotype 1b
reference strain was used. This assay measures the amount of
double-stranded RNA synthesized in the presence of HCV-
NS5BΔ21, a homopolymeric RNA template and ATP, as pre-
viously described.28 Initial screening was performed at an aurone
concentration of 20 μM. The naturally occurring aureusidin
(1, Figure 1) was found to be a potent inhibitor of HCV-
NS5BΔ21 enzyme activity (85% inhibition at 20 μM, IC50
=
5.4 ( 1.9 μM). This compound served as a “hit” for our
subsequent investigations. As most naturally occurring aurones
are hydroxylated at positions 4, 6, 20, 30, and 40 (see Figure 1), we
first investigated aurones bearing two or three hydroxy groups at
the above-mentioned positions (Table 1). At the A-ring, the
presence of a hydroxyl group at position 4 seemed to be crucial,
as 4-hydroxy and 4,6-dihydroxy derivatives 11 and 12 showed a
one-digit micromolar range IC50 (respectively 5.9 and 9.7 μM)
and as deletion of this 4-hydroxy group provided an inactive
compound (IC50 > 20 μM for 10). Replacement of the hydroxy
substituents at positions 4 and 6 by methoxy groups also resulted
in an important drop of activity (IC50 > 20 μM for 13). At the
5397
dx.doi.org/10.1021/jm200242p |J. Med. Chem. 2011, 54, 5395–5402

Journal of Medicinal Chemistry
ARTICLE
9.7 μM for 12; IC50 of 2.6 μM for 22 versus 5.9 μM for 11),
suggesting that 20,40-dihydroxy and 30,40-dihydroxy substitutions
were favorable. The position 5 was also investigated, but the 40-
hydroxy derivative 49 did not exhibit significant activity.
On the basis of the reported X-ray crystallographic structure of
genotype 1b HCV RdRp complexed with a tetracyclic indole-
based inhibitor binding to thumb pocket I (PDB code 2dxs),30
docking calculations were used to rationalize the observed
activities. The best poses generated with Autodock software for
compound 1 showed that arginine 503 anchored the molecule,
forming two hydrogen bonds with the 4-hydroxyl group and the
ketone of the aurone. Glycine 493 allowed the formation of a
third weaker hydrogen bond with the 6-hydroxy group. These
findings could potentially explain the slightly better activities of
4,6-dihydroxyaurones compared to 4-hydroxyaurones, except for
22 which showed a 2.3-fold better activity than the correspond-
ing 4,6-dihydroxylated 34. At the B-ring, the 30- and 40-hydroxyl
groups of 1 allowed the formation of two additional chelating
hydrogen bonds with the carbonyl group of leucine 392 inside
the hydrophobic pocket, as shown in Figure 4A. These findings
could explain the strong impact of hydroxy groups at these
positions. However, as stated above, 30-hydroxylated aurones
were almost inactive, whereas 40-hydroxylated ones often pro-
vided some of the best observed activities, indicating a stronger
influence of the 40-position, not explained by this binding model.
In order to study the significance of the hydrophilicity and
hydrogen bond donor potential of hydroxy substituents in
aurones, we synthesized several compounds in which hydroxy
groups were replaced by isosteric hydrophobic fluorine atoms.
Compounds bearing 4-fluoro or 4,6-difluoro substituents were
found as totally inactive molecules (47 and 48, fluoro analogues
of 39 and 22). This highlighted the crucial importance of the
A-ring hydroxy substitution pattern as a hydrophilic and hydro-
gen bond donor moiety and seemed to confirm the docking
indications about the key anchor role of this part of the inhibitors.
On the other hand, the introduction of a fluorine atom in
position 40 of the aurones affected very slightly the biological
activity of the compound (IC50 of 11.2 μM for 42 versus 9.7
for 12), suggesting the possibility of introducing hydrophobic
substituents at this part of the molecule.
Figure 2. Structure of HCV RdRp with its three subdomains (fingers in
red, palm in blue, and thumb in green) and its four allosteric binding sites
for known non-nucleoside inhibitors. Key substitution sites of adjacent
amino acids are indicated in bold.
Given the strongly hydrophobic environment of thumb
pocket I and the design of our structureꢀactivity relationship,
we considered testing the effect of hydrophobic substituents at
the B-ring of aurones, especially at position 40. Molecular docking
based on the above known crystallographic structure (PDB code
2dxs) was used throughout the decision process as an indicator of
the possibly relevant substituents and their positions. With the
help of these tools, it appeared that the 40-butyl and 40-cyclohexyl
moieties were wise choices. Indeed, the 40-butyl and 40-cyclohex-
yl aurones were the most active compounds of the 4,6-dihydroxy
series (IC50 of 2.3 and 2.5 μM for 40 and 39), but the 40-ethyl or
40-isopropyl groups provided derivatives in the same range of
activity (IC50 of 3.1 and 3.6 μM for 38 and 37). Interestingly, the
4-hydroxy analogues were 10-fold less active (IC50 > 20 μM
for 27, 28, and 29) except for the 20,40-dimethyl derivative 23
(IC50 = 5.8 μM) and for the 40-N-methylpiperazinyl derivative 31
(IC50 = 3.8 μM). Methoxy substituted derivatives did not reach
inhibition values as well (IC50 = 9.7 μM for the best one 24, and
IC50 > 20 μM for 25, 26, and 36).
Figure 3. Fold increase in IC50 relative to wild-type NS5BΔ21 obtained
with aurone 11 and P495L, H95Q, C316Y, and M423T mutants. These
mutations are known to reduce susceptibility to thumb pocket I, palm
pocket II, palm pocket I, and thumb pocket II inhibitors, respectively.
auroneꢀNS5B complex, which would be a solid proof for the
binding site of aurones, suggests caution in the interpretation
of these results.
With the findings reported in Table 1 and the mutagenesis
experiments in hand, we explored further the structureꢀactivity
relationship of NS5B inhibition by aurones (Table 2). At the
B-ring, complete deletion of hydroxyl groups was followed by a
1.5- to 2.5-fold loss of activity (IC50 of 14.6 μM for 21 versus
5.9 μM for 11; IC50 of 15.1 μM for 33 versus 9.7 μM for 12). In
contrast, the introduction of additional hydroxy groups at the
B-ring led to 1.5- to 2.5-fold more active compounds (IC50 of
5.4 μM for the already described 1 and 5.8 μM for 34 versus
Additionally, among a large number of virtual alternative
structures, molecular modeling identified the replacement of
the B-ring by substituted indole cores as potential candidates for
5398
dx.doi.org/10.1021/jm200242p |J. Med. Chem. 2011, 54, 5395–5402

Journal of Medicinal Chemistry
ARTICLE
Table 2. Inhibition Activity of Compounds 1 and 21ꢀ51
0
0
0
0
R6
compd
R4
R5
R6
R2
R3
R4
inhibition (20 μM, %)
IC50 (μM)
1
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OMe
OMe
OMe
H
H
OH
H
H
OH
H
OH
H
H
85 ( 3
57 ( 3
94 ( 2
64 ( 1
75 ( 3
47 ( 2
52 ( 1
46 ( 2
30 ( 2
42 ( 1
61 ( 2
94 ( 3
58 ( 5
59 ( 4
81 ( 6
74 ( 3
50 ( 2
79 ( 7
78 ( 4
93 ( 4
92 ( 2
66 ( 1
80 ( 2
14 ( 3
3 ( 2
5.4 ( 0.3
14.6 ( 0.6
2.6 ( 0.1
5.8 ( 0.2
9.7 ( 0.2
nd
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
OH
H
H
H
H
H
H
H
OH
Me
OMe
OMe
OMe
H
H
OH
Me
H
H
H
H
OMe
H
H
H
OMe
OMe
H
H
H
OMe
Et
nd
H
H
nd
H
H
H
i-Pr
n-Bu
t-Bu
H
nd
H
H
H
H
nd
H
H
H
H
nd
H
H
H
N-Me-Ppz
H
H
3.8
H
F
H
H
nd
OH
OH
OH
OH
OH
OH
OH
OH
OH
OH
OMe
OMe
OMe
H
H
H
H
H
15.1 ( 0.7
5.8 ( 0.3
nd
OH
OMe
OMe
H
H
OH
H
OMe
OMe
H
H
H
OMe
Et
H
nd
H
3.6 ( 0.2
3.1 ( 0.1
2.5 ( 0.1
2.3 ( 0.2
11.4 ( 0.3
11.2 ( 0.2
nd
H
H
i-Pr
H
H
H
n-Bu
c-Hex
morpholinyl
F
H
H
H
H
H
H
H
H
H
H
Me
H
H
Me
H
H
n-Bu
c-Hex
n-Bu
n-Bu
F
H
nd
H
H
H
2 ( 1
nd
H
H
H
ꢀ34 ( 3
ꢀ33 ( 4
10 ( 2
46 ( 2
3 ( 3
nd
F
F
H
H
H
nd
F
H
F
H
H
nd
H
H
H
H
OH
H
nd
OBn
OBn
H
H
c-Hex
H
nd
95 ( 1
2.2 ( 0.2
the improvement of the activity. In models built with GOLD and
Autodock software, the best poses were obtained for N-alkylindole
analogues ((Z)-2-(N-alkylindolyl)methylene)-4,6-dihydroxyben-
zofuran-3(2H)-one). Compound 51, which belongs to this group
of indole derivatives, had a highly discriminant Goldscore of 52.4
(36.9ꢀ43.6 for compounds 23, 24, 27, 28, 29, 39, and 40).
Arginine 503 was found to anchor the molecule in a similar way
compared to compound 1, whereas hydrophobic residues (especially
leucine 492, valine 37, alanine 393, leucine 392, alanine 396,
alanine 395, phenylalanine 429, and isoleucine 424), forming
two pockets, maintained both the indole part and the N-alkyl
substituent of the ligand. The fixation mode of compound 51,
which had the strongest inhibitory activity (IC50 = 2.2 μM), is
shown in Figure 4B. Although these findings were a useful
indication of the binding mode and confirmed the inhibitory
potential of this group of indole derivatives, the slight gain of
biological activity (IC50 of 2.2 μM versus 2.5 μM for 39 or 2.3 μM
for 40) did not reflect completely the good discrimination in silico
between 51 and most of the other micromolar range inhibitors.
Overall, the docking results obtained with weak inhibitors
were, as expected, difficult to interpret, but the geometries of
poses obtained with more efficient compounds were often
homogeneous and relevant. The strong activities of both B-hy-
droxylated and B-hydrophobic compounds could be explained
5399
dx.doi.org/10.1021/jm200242p |J. Med. Chem. 2011, 54, 5395–5402

Journal of Medicinal Chemistry
ARTICLE
Figure 4. Surface and ribbon views of predicted binding modes in RdRp thumb pocket I of a reported X-ray structure (PDB code 2dxs) from docking
studies with Autodock. (A) For compound 1, involved in five hydrogen bonds: two with arginine 503 (1.9 Å and 2.5 Å), one with glycine 493 (1.8 Å), and
two with leucine 392 (2.0 Å and 2.0 Å). (B) For compound 51, involved in three hydrogen bonds: two with arginine 503 (1.8 Å and 2.0 Å) and one with
glycine 493 (2.2 Å). The pictures were built with Pymol software.
by the good fit of these two subclasses of aurones in the pocket.
Compounds 1 and 51 were representative of this hypothetic dual
mode of aurone binding to thumb pocket I of HCV RdRp.
Finally, in brief, we identified the following pharmacophoric
elements as crucial in the inhibitory activity of aurones on HCV
RdRp: (a) a hydroxy group at position 4 or a dihydroxy group at
positions 4 and 6 and (b) a 20,40- or 30,40-dihydroxylated B-ring or
a hydrophobic and bulky substituent or alternative core at the
B-ring.
All the reported compounds were evaluated for cytotoxicity,
based on the MTT assay with two human cell lines (HuH7 and
HEK293). Huh7 and HEK293 cells were incubated in the
presence of increasing concentrations of the tested compound,
ranging from 0 to 512 μM. All reported compounds were not
cytotoxic even at the highest concentration (data not shown).
shifts (δ) are reported in ppm relative to Me4Si (internal standard).
Electrospray ionization ESI mass spectra were acquired by the Analytical
Department of Grenoble University, France, on an Esquire 300 Plus
Bruker Daltonis instrument with a nanospray inlet. Combustion ana-
lyses were performed at the Analytical Department of Grenoble Uni-
versity, and all tested compounds have a purity of at least 95%. Thin-
layer chromatography (TLC) used Merck silica gel F-254 plates
(thickness 0.25 mm). Flash chromatography used Merck silica gel 60,
200ꢀ400 mesh. All solvents were distilled prior to use. Unless otherwise
stated, reagents were obtained from commercial sources and were used
without further purification. Compounds 2ꢀ9 were prepared according
to published procedures.23ꢀ27
General Procedure A for the Preparation of (Z)-2-Benzy-
lidenebenzofuran-3(2H)-one Derivatives. To a solution of a
benzofuran-3(2H)-one derivative in methanol (15 mL/mmol) were
added an aqueous solution of potassium hydroxide (50%, 1.5 mL/
mmol) and a benzaldehyde derivative (1.5 equiv). The solution was then
refluxed until TLC showed complete disappearance of the starting
materials (1ꢀ18 h). After cooling, the mixture was concentrated under
reduced pressure. Then the residue was diluted in water (50 mL/mmol),
and an aqueous solution of hydrochloric acid (1 N) was added to adjust
the pH to 2ꢀ3. The mixture was extracted with ethyl acetate or
dichloromethane and washed with water and brine. The combined
organic layers were dried over magnesium sulfate and filtered, and the
filtrate was concentrated under reduced pressure to give the crude
compound.
’ CONCLUSION
We investigated the capacity of aurones, nontoxic natural com-
pounds, to inhibit HCV RdRp. Aureusidin potently inhibited the
activity of this enzyme. By using classical medicinal chemistry, we
identified positions and substituents that are crucial for the inhibitory
activity of aurones on RdRp. A combination of site-directed muta-
genesis and molecular modeling was successfully used to understand
and optimize aurones’ anti-HCV activity. Our findings constitute a
new starting point in the development of new non-nucleoside
inhibitors of HCV RdRp. Studies are underway to synthesize more
active compounds warranting preclinical evaluation.
General Procedure B for the Preparation of (Z)-2-Benzy-
lidenebenzofuran-3(2H)-one Derivatives. To a solution of a
benzofuran-3(2H)-one derivative in ethanol (4 mL/mmol) were added
an aqueous solution of potassium hydroxide (50%, 5 mL/mmol) and a
benzaldehyde derivative (2 equiv). The solution was then refluxed until
TLC showed complete disappearance of the starting materials (2ꢀ5 h).
After cooling, the mixture was concentrated under reduced pressure.
Then the residue was diluted in water (50 mL/mmol), and an aqueous
’ EXPERIMENTAL SECTION
1
Chemistry. H and 13C NMR spectra were recorded on a Br€uker
AC-400 instrument (400 MHz for 1H and 100 MHz for 13C). Chemical
5400
dx.doi.org/10.1021/jm200242p |J. Med. Chem. 2011, 54, 5395–5402

Journal of Medicinal Chemistry
ARTICLE
solution of hydrochloric acid (1 N) was added to adjust the pH to 2ꢀ3.
The mixture was extracted with ethyl acetate or dichloromethane and
washed with water and brine. The combined organic layers were dried
over magnesium sulfate and filtered, and the filtrate was concentrated
under reduced pressure to give the crude compound.
used to assess the effect of amino acid substitutions conferring resistance
to other specific HCV RdRp inhibitors. The tested substitutions in-
cluded M423T, C316Y, H95Q, and P495L. The constructs were verified
by DNA sequencing, and the corresponding proteins were purified as
described above. Each experiment was performed in triplicate.
Cytotoxicity Assay. Exponentially growing HuH7 and HEK293
cells were trypsinized and plated (2000 and 1000 cells per well,
respectively) in 96-well microplates and allowed to attach for 96 h at
37 ꢀC under 5% CO2. A 200 μL aliquot of drug solution, diluted in
medium with 1% DMSO (final concentration), was added and the cells
were incubated for 72 h at 37 ꢀC under 5% CO2. The drug was evaluated
at different concentrations ranging from 0 to 512 μM. Cytotoxicity was
evaluated with a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium
bromide colorimetric assay (MTT assay). Each experiment was per-
formed in triplicate.
General Procedure C for the Preparation of (Z)-2-Ben-
zylidenebenzofuran-3(2H)-one Derivatives. To a solution
of a benzofuran-3(2H)-one derivative in anhydrous dichloromethane
(20 mL/mmol) were added a benzaldehyde derivative (1.5 equiv) and
aluminum oxide (4000 mg/mmol). The suspension was stirred under
argon overnight, and the solid was removed by filtration. The filtrate was
distilled under reduced pressure to give the crude compound.
General Procedure D for the Preparation of (Z)-2-Benzy-
lidenebenzofuran-3(2H)-one Derivatives. To a solution of a
(Z)-2-benzylidenebenzofuran-3(2H)-one derivative in anhydrous di-
chloromethane (10 mL/mmol) was added pure boron tribromide (20
equiv) at 0 ꢀC. The solution was then stirred at room temperature until
TLC showed complete disappearance of the starting materials (24ꢀ72 h).
Cold water was then added, and the suspension was extracted three
times with ethyl acetate and washed with water and brine. The combined
organic layers were dried over magnesium sulfate and filtered, and the
filtrate was concentrated under reduced pressure to give the crude
compound.
’ ASSOCIATED CONTENT
S
Supporting Information. Additional experimental de-
b
tails and 1H NMR, 13C NMR, and MS data for all compounds.
This material is available free of charge via the Internet at http://
pubs.acs.org.
Molecular Modeling. Docking studies used the reported X-ray
crystallographic structure of HCV genotype 1b RdRp complexed with a
known inhibitor binding to thumb pocket I of the enzyme (PDB code
2dxs).31 The protein structure was extracted, and water molecules were
deleted. The complexed ligand was used to define the binding site for
docking and was successfully docked in a similar conformation in the
pocket, using the same protocol as for aurones. The genetic algorithms
of GOLD and Autodock4 docking software performed flexible docking
with small molecules while keeping the protein structure rigid. All
compounds were built and their energies minimized with Sybyl, using a
conjugate gradient method and MMFF94 force field and charges. For
each compound, 20 poses were generated by GOLD and ranked
according to the Goldscore scoring function, while 100 poses were
generated by Autodock and clustered according to their spatial proxi-
mity and calculated free energy of binding. Molecular surfaces were
generated with MOLCAD, and the docking solutions were analyzed
with Sybyl.
Biology. Expression and Purification of Recombinant HCV
NS5BΔ21. RdRp (NS5B protein) from the HCV J4 genotype 1b
reference strain, truncated of its 21 C-terminal amino acids to ensure
solubility (NS5BΔ21) and carrying a hexahistidine tag (His-Tag) at its
N-terminus, was expressed in Escherichia coli C41(DE3) and purified.
Chromatography was performed on Ni-NTA columns (Qiagen, Valen-
cia, CA). The columns were washed with a buffer containing 50 mM
NaH2PO4 (pH 8.0), 500 mM NaCl, and 20 mM imidazole. The bound
protein was eluted in 1 mL fractions with a buffer containing 50 mM
NaH2PO4 (pH 8.0), 500 mM NaCl, and 250 mM imidazole. NS5BΔ21-
enriched fractions were selected with a Bradford colorimetric assay, and
NS5BΔ21 purity was determined by SDSꢀPAGE analysis with Coo-
massie staining. Purified NS5BΔ21 fractions were pooled and dialyzed
against a buffer containing 5 mM Tris (pH 7.5), 0.2 M sodium acetate,
1 mM DTT, 1 mM EDTA, and 10% glycerol. NS5BΔ21 purity was
>98%.
’ AUTHOR INFORMATION
Corresponding Author
*Phone: (33) 4 76 63 53 11. Fax: (33) 4 76 63 52 98. E-mail:
Ahcene.Boumendjel@ujf-grenoble.fr.
Author Contributions
#These authors made equal contributions to this work.
Author Contributions
∞These authors are equal senior investigators in this study.
’ ACKNOWLEDGMENT
R.H. is a recipient of a fellowship grant from the Ministꢁere de
l’Enseignement et de la Recherche, for which he is very thankful.
A part of the project related to this article is funded by l’Agence
Nationale de la Recherche (ANR). W.Y. is a recipient of a
fellowship from the University of Sun Yat-Sen (China), for
which he is thankful.
’ ABBREVIATIONS USED
DMF, dimethylformamide; DMSO, dimethylsulfoxide; DTT, di-
thiothreitol; EDTA, ethylenediaminetetraacetic acid; HCV, he-
patitis C virus; MMFF94, Merck molecular force field 94;MTT,
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide;
NS5B, nonstructural 5B protein; PDB, Protein Data Bank; peg-
IFN, pegylated interferon; RBV, ribavirin; RdRp, RNA-depen-
dent RNA polymerase; TFA, trifluoroacetic acid; Tris, tris(hydro-
xymethyl)aminomethane; WHO, World Health Organization;
Wt, wild-type
HCV-NS5BΔ21 Polymerase Assay. An enzyme assay was used
to measure inhibition of HCV-NS5BΔ21 polymerase activity. The cell-
free HCV-NS5BΔ21 polymerase assay is based on the amount of
double-stranded RNA synthesized in the presence of HCV-NS5BΔ21,
a homopolymeric RNA template (Poly U, GE Healthcare, Chalfont St.
Giles, U.K.) and ATP, measured with an intercalating agent (SYBR
Green, Applied Biosystems), as previously described.32 The Quik-
Change site-directed mutagenesis kit (Stratagene, La Jolla, CA) was
’ REFERENCES
(1) Wong, T.; Lee, S. S. Hepatitis C: a review for primary care
physicians. Can. Med. Assoc. J. 2006, 174, 649–659.
(2) Pawlotsky, J.-M.; McHutchison, J. G. Hepatitis C. Development
of new drugs and clinical trials: promises and pitfalls. Hepatology 2004,
39, 554–567.
5401
dx.doi.org/10.1021/jm200242p |J. Med. Chem. 2011, 54, 5395–5402

Journal of Medicinal Chemistry
ARTICLE
(3) Strader, D. B.; Wright, T.; Thomas, D. L.; Seeff, L. B. Diagnosis,
management, and treatment of hepatitis C. Hepatology 2004, 39, 1147–
1171.
(4) Dixit, N. M.; Layden-Almer, J. E.; Layden, T. J.; Perelson, A. S.
Modelling how ribavirin improves interferon response rate in hepatitis C
virus infection. Nature 2004, 432, 922–924.
(5) Feld, J. J.; Hoofnagle, J. H. Mechanism of action of interferon and
ribavirin treatment of hepatitis C. Nature 2005, 436, 967–972.
(6) Zein, N. N. Clinical significance of hepatitis C virus genotypes.
Clin. Microbiol. Rev. 2000, 223–225.
(7) Carroll, S. S.; Olsen, D. B. Nucleoside analog inhibitors of
hepatitis C virus replication. Infect. Disord.: Drug Targets 2006, 6, 17–29.
(8) Beaulieu, P. L. Recent advances in the development of NS5B
polymerase inhibitors for the treatment of hepatitis C virus infection.
Expert Opin. Ther. Pat. 2009, 19, 145–164.
(19) Boumendjel, A. Aurones: a subclass of flavones with promising
biological potential. Curr. Med. Chem. 2003, 10, 2621–2630.
(20) Seikel, M. K.; Geissman, T. A. Anthochlor pigments. VII. The
pigments of yellow Antirrhinum majus. J. Am. Chem. Soc. 1950, 72,
5725–5730.
(21) Nakayama, T.; Yonekura-Sakakibara, K.; Sato, T.; Kikuchi, S.;
Fukui, Y.; Fukuchi-Mizutani, M.; Ueda, T.; Nakao, M.; Tanaka, Y.
Aureusidin synthase: a polyphenol oxidase homolog responsible for
flower coloration. Science 2000, 290, 1163–1166.
(22) Varma, S.; Varma, M. Alumina-mediated condensation. A
simple synthesis of aurones. Tetrahedron Lett. 1992, 33, 5937–5940.
(23) Beney, C.; Mariotte, A.-M.; Boumendjel, A. An efficient synth-
esis of 4,6-dimethoxy aurones. Heterocycles 2001, 55, 967–972.
(24) B€uchi, G.; Weinreb, M. Total syntheses of aflatoxins M1 and G1
and an improved synthesis of aflatoxin B1. J. Am. Chem. Soc. 1971,
93, 746–752.
(25) Okombi, S.; Rival, D.; Bonnet, S.; Mariotte, A.-M.; Perrier, E.;
Boumendjel, A. Discovery of benzylidenebenzofuran-3(2H)-one (aurones)
as inhibitors of tyrosinase derived from human melanocytes. J. Med.
Chem. 2006, 49, 329–333.
(9) Pauwels, F.; Mostmans, W.; Quirynen, L. M. M.; van der Helm,
L.; Boutton, C. W.; Rueff, A.-S.; Cleiren, E.; Raboisson, P.; Surleraux, D.;
Nyanguile, O.; Simmen, K. A. Binding-site identification and genotype
profiling of hepatitis C virus polymerase inhibitors. J. Virol. 2007,
81, 6909–6919.
(26) Brehm, C.; Zimmermann, P. J.; Zemolka, S.; Palmer, A.; Buhr,
W.; Postius, S.; Simon, W.-A.; Herrmann, M. Pharmaceutically Active
Dihydrobenzofurane-Substituted Benzimidazole Derivatives. Patent
WO 2008/084067, July 17, 2008.
(27) Kodra, J. T.; Madsen, P.; Lau, J.; Jorgensen, A. S.; Christensen,
I. T. Novel Glucagon Antagonists. Patent WO 2003/048109, June 12,
2003.
(28) Ahmed-Belkacem, A.; Ahnou, N.; Barbotte, L.; Wychowski, C.;
Pallier, C.; Brillet, R.; Pohl, R.-T.; Pawlotsky, J.-M. Silibinin and related
compounds are direct inhibitors of hepatitis C virus RNA-dependant
RNA polymerase. Gastroenterology 2010, 138, 1112–1122.
(29) Kwong, A. D.; McNair, L.; Jacobson, I.; George, S. Recent
(10) Beaulieu, P. L.; B€os, M.; Bousquet, Y.; Fazal, G.; Gauthier, J.;
Gillard, J.; Goulet, S.; LaPlante, S.; Poupart, M.-A.; Lebebvre, S.;
McKercher, G.; Pellerin, C.; Austel, V.; Kukolj, G. Non-nucleoside
inhibitors of the hepatitis C virus NS5B polymerase: discovery and
preliminary SAR of benzimidazole derivatives. Bioorg. Med. Chem. Lett.
2004, 14, 119–124.
(11) Harper, S.; Pacini, B.; Avolio, S.; Di Filippo, M.; Migliaccio, G.;
Laufer, R.; De Francesco, R.; Rowley, M.; Narjes, F. Development and
preliminary optimization of indole-N-acetamide inhibitors of hepatitis C
virus NS5B polymerase. J. Med. Chem. 2005, 48, 1314–1317.
(12) Stansfield, I.; Ercolani, C.; Mackay, A.; Conte, I.; Pompei, M.;
Koch, U.; Gennari, N.; Giuliano, C.; Rowley, M.; Narjes, F. Tetracyclic
indole inhibitors of hepatitis C virus NS5B-polymerase. Bioorg. Med.
Chem. Lett. 2009, 19, 627–632.
progress in the development of selected hepatitis C virus NS3 4A
3
protease and NS5B polymerase inhibitors. Curr. Opin. Pharmacol. 2008,
8, 522–531.
(13) Habermann, J.; Capito, E.; del Rosario Rico Ferreira, M.; Koch,
U.; Narjes, F. Discovery of pentacyclic compounds as potent inhibitors
of hepatitis C virus NS5B RNA polymerase. Bioorg. Med. Chem. Lett.
2009, 19, 633–638.
(30) Hashimoto, H.; Mizutani, K.; Yoshida, A. Fused Cyclic Com-
pounds and Medicinal Use Thereof. Patent WO 2003/000254, January
3, 2003.
(31) Ikegashira, K.; Oka, T.; Hirashima, S.; Noji, S.; Yamanaka, H.;
Hara, Y.; Adachi, T.; Tsuruha, J.-I.; Doi, S.; Hase, Y.; Noguchi, T.; Ando,
I.; Ogura, N.; Ikeda, S.; Hashimoto, H. Discovery of conformationally
constrained tetracyclic compounds as potent hepatitis C virus NS5B
RNA polymerase inhibitors. J. Med. Chem. 2006, 49, 6950–6953.
(32) Wakita, T.; Pietschmann, T.; Kato, T.; Date, T.; Miyamoto, M.;
Zhao, Z.; Murthy, K.; Habermann, A.; Krausslich, H. G.; Mizokami, M.;
Bartenschlager, R.; Liang, T. J. Production of infectious hepatitis C virus
in tissue culture from a cloned viral genome. Nat. Med. 2005, 11,
791–796.
(14) Goulet, S.; Poupart, M.-A.; Gillard, J.; Poirier, M.; Kukolj, G.;
Beaulieu, P. L. Discovery of benzimidazole-diamide finger loop (thumb
pocket I) allosteric inhibitors of HCV NS5B polymerase: implementing
parallel synthesis for rapid linker optimization. Bioorg. Med. Chem. Lett.
2010, 20, 196–200.
(15) Beaulieu, P. L.; Jolicoeur, E.; Gillard, J.; Brochu, C.; Coulombe,
R.; Dansereau, N.; Duan, J.; Garneau, M.; Jakalian, A.; K€uhn, P.; Lagacꢀe,
L.; LaPlante, S.; McKercher, G.; Perrault, S.; Poirier, M.; Poupart, M.-A.;
Stammers, T.; Thauvette, L.; Thavonekham, B.; Kukolj, G. N-Acetami-
deindolecarboxylic acid allosteric “finger-loop” inhibitors of the hepatitis
C virus NS5B polymerase: discovery and initial optimization studies.
Bioorg. Med. Chem. Lett. 2010, 20, 857–861.
(16) Beaulieu, P. L.; Dansereau, N.; Duan, J.; Garneau, M.; Gillard,
J.; McKercher, G.; LaPlante, S.; Lagacꢀee, L.; Thauvette, L.; Kukolj, G.
Benzimidazole thumb pocket I finger-loop inhibitors of HCV NS5B
polymerase: improved drug-like properties through C-2 SAR in three
sub-series. Bioorg. Med. Chem. Lett. 2010, 20, 1825–1829.
(17) Narjes, F.; Crescenzi, B.; Ferrara, M.; Habermann, J.; Colarusso,
S.; del Rosario Rico Ferreira, M.; Stansfield, I.; Mackay, A. C.; Conte, I.;
Ercolani, C.; Zaramella, S.; Palumbi, M.-C.; Meuleman, P.; Leroux-
Roels, G.; Giuliano, C.; Fiore, F.; Di Marco, S.; Baiocco, P.; Koch, U.;
Migliaccio, G.; Altamura, S.; Laufer, R.; De Francesco, R.; Rowley, M.
Discovery of (7R)-14-cyclohexyl-7-{[2-(dimethylamino)ethyl](methyl)-
amino}-7,8-dihydro-6H-indolo[1,2-e][1,5]benzoxazocine-11-carboxylic
acid (MK-3281), a potent and orally bioavailable finger-loop inhibitor of
the hepatitis C virus NS5B polymerase. J. Med. Chem. 2011, 54, 289–301.
(18) Study To Evaluate the Safety, Tolerability and Pharmacokinetics of
MK-3281 in Healthy and Hepatitis C Infected Male Patients; Trial
NCT00635804; National Institutes of Health, Bethesda, MD, U.S.,
2008; http://clinicaltrials.gov/ct2/show/NCT00635804.
5402
dx.doi.org/10.1021/jm200242p |J. Med. Chem. 2011, 54, 5395–5402