I. Novel HCV NS5B polymerase inhibitors: Discovery of indole 2-carboxylic acids with C3-heterocycles

Article abstract of DOI:10.1016/j.bmcl.2011.07.021

SAR development of indole-based palm site inhibitors of HCV NS5B polymerase exemplified by initial indole lead 1 (NS5B IC₅₀ = 0.9 μM, replicon EC₅₀ >100 μM) is described. Structure-based drug design led to the incorporation of novel heterocyclic moieties at the indole C3-position which formed a bidentate interaction with the protein backbone. SAR development resulted in leads 7q (NS5B IC₅₀ = 0.032 μM, replicon EC₅₀ = 1.4 μM) and 7r (NS5B IC₅₀ = 0.017 μM, replicon EC₅₀ = 0.3 μM) with improved enzyme and replicon activity.

Full text of DOI:10.1016/j.bmcl.2011.07.021

Bioorganic & Medicinal Chemistry Letters 21 (2011) 5336–5341
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
I. Novel HCV NS5B polymerase inhibitors: Discovery of indole 2-carboxylic acids
with C3-heterocycles
⇑
Gopinadhan N. Anilkumar , Charles A. Lesburg, Oleg Selyutin, Stuart B. Rosenblum, Qingbei Zeng,
Yueheng Jiang, Tin-Yau Chan, Haiyan Pu, Henry Vaccaro, Li Wang, Frank Bennett, Kevin X. Chen,
Jose Duca, Stephen Gavalas, Yuhua Huang, Patrick Pinto, Mousumi Sannigrahi, Francisco Velazquez,
Srikanth Venkatraman, Bancha Vibulbhan, Sony Agrawal, Nancy Butkiewicz, Boris Feld, Eric Ferrari,
Zhiqing He, Chuan-kui Jiang, Robert E. Palermo, Patricia Mcmonagle, H.-C. Huang, Neng-Yang Shih,
George Njoroge, Joseph A. Kozlowski
Merck Research Laboratories, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
SAR development of indole-based palm site inhibitors of HCV NS5B polymerase exemplified by initial
Received 28 April 2011
Revised 30 June 2011
Accepted 6 July 2011
Available online 19 July 2011
indole lead 1 (NS5B IC50 = 0.9
to the incorporation of novel heterocyclic moieties at the indole C3-position which formed a bidentate
interaction with the protein backbone. SAR development resulted in leads 7q (NS5B IC50 = 0.032 M,
replicon EC50 = 1.4 M) and 7r (NS5B IC50 = 0.017 M, replicon EC50 = 0.3 M) with improved enzyme
and replicon activity.
lM, replicon EC50 >100 lM) is described. Structure-based drug design led
l
l
l
l
Keywords:
HCV
Ó 2011 Elsevier Ltd. All rights reserved.
NS5B polymerase
Hepatitis C virus (HCV) is a major health hazard affecting over
170 million individuals worldwide1a and its infection is a leading
cause of chronic liver cirrhosis and death from liver disease in
the United States.1b,c The current standard of care is treatment
with a combination of subcutaneous pegylated interferon adminis-
tration with oral dosing of the cytotoxic nucleoside drug ribavirin.2
The response rate is >75% for HCV patients with genotypes 2 and 3
after a 24 week treatment regimen while genotype 1 patients have
a response rate of less than 50% after 48 weeks of treatment.3 With
a clear opportunity to improve clinical outcomes, and given the
side effects associated with the current standard of care, it is valu-
able to discover potent inhibitors of HCV replication that will
improve outcomes and shorten treatment duration.
The HCV NS5B protein is an RNA-dependent RNA polymerase
critical for the synthesis of progeny viral genomes. The crystal
structure of HCV NS5B displayed an overall subdomain architec-
ture similar to other members of the PolI family,4 with a deep ac-
tive site cavity, located at the top of the ‘palm’ subdomain, and
sealed at its base by a unique b-loop. Furthermore, there was an
unexpected interaction identified between the tip of the ‘fingers’
subdomain and the ‘thumb’ subdomain to encircle the presumed
nucleoside triphosphate substrate entry trajectory (Fig. 1).
Figure 1. Overall protein molecular surface of NS5B with compound 1 (yellow
spheres) bound within the active site cavity at the palm site. The ‘fingers’, ‘palm’,
and ‘thumb’ domains are labeled and colored in blue, red, and green, respectively,
beneath the molecular surface.
Sequence variation analysis suggests that residues lining the
active site cavity (‘palm site’) more conserved than in other
regions, such as the ‘thumb-site’. This makes the palm site an
attractive target for inhibition of the viral polymerase, though
⇑
Corresponding author.
E-mail address: gopinadhan.anilkumar@merck.com (G.N. Anilkumar).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.07.021

G. N. Anilkumar et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5336–5341
5337
Table 1
Indole-N1 substitutions
3'
F
3
5
Cl
O
F
Cl
O
N
OH
N
OH
R1
Compds
R1
IC50 (l
M)7
1
Benzyl (lead)
H
Methyl
Acetyl
Benzoyl
Naphthalene-2-ylmethyl
3-Me-benzyl
3-CF3-benzyl
3-CF3O-benzyl
2,5-F-benzyl
2-NH2-benzyl
0.9
7.3
1.3
8.1
7.8
4.4
2.5
I, NS5B IC50 = 0.9 µM h
Replicon EC50 > 100 µM
1a
1b
1c
1d
1e
1f
1g
1h
1i
1j
1k
1l
1m
1n
1o
1p
Figure 2. X-ray structure of lead indole acid analog 1 bound within the active site
cavity of NS5B at the palm site. Residues presenting sidechain atoms within 5 Å of
the compound are shown, along with backbone atoms of Ile-447 and Tyr-448. An
unusual Fꢁ ꢁ ꢁHN interaction and a close CHꢁ ꢁ ꢁO@C approach are illustrated by dotted
lines.
>44
1.8
0.3
0.6
0.2
1.3
0.6
0.7
1.3
1.2
3-NH2-benzyl
not all residues lining this site are absolutely conserved. Clinical
efficacy has been demonstrated with non-nucleoside inhibitors
binding at the palm, thumb, and finger-loop subdomains.5
High throughput screening of HCV NS5B polymerase using a no-
vel trinucleotide-primed assay,6 designed to respond to both initi-
ation and elongation inhibitors, revealed a structurally-similar
cluster of weak indole inhibitors exemplified by initial lead 1 with
3-NH2, 4-Me-benzyl
3-AcNH-benzyl
3-BnNH-benzyl
Pyridin-3-ylmethyl
Pyridin-4-ylmethyl
ꢀ0.9
lM binding affinity. The X-ray structure of NS5B with initial
Ar
lead 1 (Fig. 2), confirms that compound 1 binds within the active
site cavity of the apoprotein at the ‘palm site’. The N-benzyl group
of inhibitor 1 stacks upon Cys-366. The C2 carboxylate does not
contact the protein directly. Hydrophobic interactions from the in-
dole core and the N1 phenyl ring dominated the contacts, and
make no specific polar interactions. The indole core and the C3
phenyl ring are in contact with the side chain of Met-414. Of par-
ticular note is the identification of an unusual Fꢁ ꢁ ꢁHN interaction
with a 2.6 Å distance between the C3 aromatic fluorine and the
protein backbone of Tyr-448. The phenyl 30-carbon adjacent to
the fluorine is only 2.9 Å from the backbone carbonyl oxygen of
Ile-447, well within the van der Waals contact distance. The indole
5-chlorine atom protrudes into a constricted tunnel formed pri-
marily by hydrophobic and aromatic residues such as Leu-384
and Tyr-415.
Br
R5
R5
R5
b
a
CO2R
CO2R
CO2R
N
H
3
N
H
N
H
4
5
Ar
Ar
R5
R5
c
d
CO2R
CO2H
N
N
R1
R1
6
1a-o
2a-k
Scheme 1. General synthesis of analogs. Reagents and conditions: (a) NBS, THF, rt;
(b) Ar-(B(OH)2, Pd(dppf)Cl2, K2CO3, H2O, DME, reflux; (c) R1-Br, Cs2CO3, DMF, rt; (d)
aq 1 N LiOH, TFH, reflux.
The initial efforts were focused on variation of the indole N1
substituent with selected SAR is presented in Table 1. Synthesis
of the analogs in Table 1 generally followed a route as outlined
in Scheme 1. Commercially available indole 2-carboxylic acid es-
ters (ethyl or methyl) were selectively brominated at the C3 posi-
tion by treatment with N-bromosuccinimide. Bromide 4 was
subjected to Suzuki coupling with aryl boronic acids using
Pd(dppf)Cl2 as the catalyst. N-alkylation of the indole was achieved
by treatment with an appropriate halide and Cs2CO3 as the base. It
should be noted that other bases such as sodium hydride also gave
satisfactory yields. In some examples the arylation and alkylation
steps were done in a reverse order. In the final step, the indole
C2 acids were prepared by hydrolysis of the corresponding ester
using aqueous 1 N LiOH. In cases were there was no substitution
at N1 (1a), the corresponding functionalization step was omitted.
The 5-chloro indole analog (1a) with an unsubstituted N1 was
approximately 10-fold less potent than 1, while introducing small
alkyl substitutions like methyl at the N1 position were tolerated
(1b). Both acetyl and benzoyl substitutions at the N1 position
showed eightfold reduced enzyme activity (1c,d). Bulky lipophilic
substitution such as naphthalene-2-ylmethyl was not tolerated
(1e). The observed N1 SAR trend, coupled with the fact that the
screening hits were part of a larger library with diverse N1 substi-
tutions (not described), suggested the critical nature of the N1
benzyl scaffold despite the lack of any obvious interaction with
the protein. The focus was shifted to substitutions around the
phenyl portion of the N1-benzyl group. The X-ray structure with
1, showed that the benzyl group was close enough to the protein
backbone that substitutions offered a chance to identify new
interactions. Lipophilic benzyl substituents such as 3-methyl
(1f), 3-trifluoromethyl (1g) and the 3-trifluoromethoxy analog
(1h) were deleterious relative to the initial lead. On the other
hand, 2,5-difluoro substitution of the benzyl group had a benefi-
cial effect on the enzyme activity (1i, IC50 = 0.3
3- amino substitutions also showed improved activity, with a reg-
iochemical preference (1j,k, IC50 = 0.6–0.2 M) for 3-amino substi-
lM). Both 2- and
l
tuted benzyl. A follow up focused library of ꢀ300 compounds
with functionalized 3-amino substitutions to probe additional
interactions with protein backbone did not identify any com-
pounds with improved activity. The analogs which showed com-
parable activity to 1k were the 3-acetyl and 3-amino benzyl
analogs, 1m and 1n. The aniline liability of 1k prompted prepara-
tion of pyridine–methyl analogs 1o and 1p which were equipotent
with initial lead 1.
The X-ray structure of the lead 1 revealed that the C3-fluoro-
phenyl group is close to the protein backbone. Efforts were
made to understand the potential interactions in this region as

5338
G. N. Anilkumar et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5336–5341
Table 2
while hydroxyl substitution (2c, IC50 = 1.4 lM) was tolerated. This
Variation of C3-phenyl substitutions
observation suggested that this region does not accommodate
bulky functionalities. Alternatively, an unsubstituted phenyl group
at C3 (2d) also did not improve the binding activity, which sug-
gested proper aromatic substitutions could have a positive impact
on the binding with the protein. A variety of substitutions were
examined at different positions around the C3-phenyl group as
exemplified by 2e–j. Overall there was limited success in improv-
ing the binding activity. Apart from the observation that substitu-
tions at 3- and 4- positions of the ring were not tolerated, it was
noticed that the position of F group in the phenyl ring was also crit-
ical. This further reinforced the observation in the X-ray structure
of 1 that ortho flourine substitution is in close proximity with the
Tyr-448 residue of the protein. The gain in activity from 2h
R3
Cl
O
N
OH
R1
R3
Compds
R1
hIC50
17
(l
M)7
2a
H
F
F
2b
2c
10
OMe
OH
(IC50 = 48
is greater than 40-fold. Interestingly the 20-fluoropyridine analog
(2k, IC50 = 0.3 M) showed promising binding activity, but the ob-
lM to 2j (IC50 = 1.1 l
M) by adding in the 20-F substituent
l
F
served chemical reactivity of this system raised questions about
the stability and selectivity which deprioritized further follow up.
Our initial efforts to explore the N1 and C3 positions of the lead
compound 1 resulted in many compounds with sub-micromolar
1.4
F
enzyme activity (IC50 = 0.2–0.3
showed significant cell based activity in the replicon assay
(EC50 >25 M). This prompted us to revisit the X-ray structure of
lM). None of these compounds
2d
2e
2f
4.2
Cl
Cl
Cl
Cl
Cl
Cl
Cl
l
Cl
the lead compound 1 and look for specific interactions with protein
which can be improved. The most obvious finding from the X-ray
structure as described above was the Fꢁ ꢁ ꢁHN interaction. We rea-
soned that this may be an interaction with the protein which can
be improved if we can replace the C-F with a carbonyl, as a
C@Oꢁ ꢁ ꢁHN would provide a more canonical hydrogen bonding
interaction. Thus analogs with carbonyl containing functionalities
were targeted. The synthesis of diverse N-linked and C-linked het-
erocycles at the indole C3 position were carried out as described in
Schemes 2–4.
Synthesis of the C-linked heterocyclic analogs pyridone and
pyrimidinedione in Table 3 were made as outlined in Scheme 2.
The indole 2-carboxylic acid esters (ethyl or methyl) were selec-
tively iodinated at the C3 position by treatment with N-iodosuccin-
imide. The 3-iodo intermediate 8 was subjected to Suzuki coupling
with either 2-methoxypyridine boronic acid or 2,4-dimethoxypyr-
imidine-5-boronic acid using Pd(dppf)Cl2 as the catalyst. The sub-
sequent N-alkylation of the indole was achieved with an
appropriate electrophile, and typically using Cs2CO3 as the base.
33
CN
30
O
O
2g
2h
2i
29
F
>48
>46
1.1
F
F
F
N
N
NH
R5
d,e
OMe
CO2R
R5
OMe
CO2R
R5
O
CO2H
c
2j
N
R1
10
N
H
9
N
R1
F
F
N
7 b,d,h-m, r
OMe
(HO)2B
N
I
R5
b
R5
a
2k
0.3
CO2R
CO2R
F
N
H
N
b
H
MeO
N
F
3
8
N
OMe
OMe
N
O
B(OH)2
N
OMe
N
HN
NH
N
R5
summarized in Table 2. Syntheses of series 2a–k compounds were
also achieved following the chemistry described in Scheme 1. Since
halogen-substituted benzyl groups were tolerated at the indole N1
OMe
CO2R
c
R5
O
CO2H
d,e
R5
OMe
CO2R
N
H
N
N
R1
R1
position, exemplified by 1i (IC50 = 0.3 lM), both chloro- and fluoro-
7 c,e
11
12
substituted benzyl analogs were installed at N1. The C3 unsubsti-
tuted indole analog (2a) was only a weak binder of NS5B.
Replacement of fluorine at the ortho position of C3 phenyl with a
Scheme 2. General synthesis of C3- C-linked hetero analogs. Reagents and
conditions: (a) NIS, CHCl3, rt; (b) Pd(dppf)Cl2, K2CO3, H2O, DME, reflux; (c) R1-Br,
Cs2CO3, DMF, rt; (d) 4 N HCl/dioxane, MeOH, 90 °C; (e) aq 1 N LiOH, TFH, reflux.
methoxyl group attenuated activity as well (2b, IC50 = 10 lM)

G. N. Anilkumar et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5336–5341
5339
NH2
CO2tBu
3-aminoindole 23 which was cyclized to 24 under basic condi-
tions by reaction with corresponding chloro alkyl sulfamoyl
chlorides.
Earlier SAR results at N1 demonstrated that either fluoro- or
amino- substituted benzyl groups were preferred. These particular
N1-benzyl groups were selected for further SAR explorations
(Table 3). Interestingly, the N-linked pyridone analog 7a showed
an improved enzyme potency with moderate replicon activity
R5
Cl
R5
CN
R5
CN
F
b
c
a
N
N
H
14
CO2tBu
O
Cbz
15
H2N CO2tBu
Bn
O
13
Ar
Ar
O
NH
O
C
MeO2C
R5
NH
O
O
N
N
R5
HN
d
R5
e
t
CO2tBu
CO2 Bu
CO2tBu
N
CO2Me
Ar
N
H2N
(IC50 = 0.15 lM, EC50 = 19 lM). The X-ray structure with lead com-
N
H
Cbz
H
16
18
pound 1 had also revealed a close contact between the hydrogen at
meta position of the C3 phenyl and the carbonyl group of Ile-447.
This suggested an opportunity to replace this interaction with a
N–Hꢁ ꢁ ꢁO@C, which could result in bidentate interaction of the
inhibitor with the protein backbone. To achieve such an interac-
tion, the C-linked pyridone derivative 7b was prepared which
could offer the desired C@O and NH functionalities needed to
interact with the polypeptide backbone of Tyr-448 and Ile-447.
Gratifyingly, 7b showed a substantial improvement in the enzyme
activity (IC50 = 35 nM) and moreover an improved replicon activity
17
Ar
Ar
Ar
NH
NBoc
NBoc
h
g
O
f
O
O
N
N
N
R5
O
R5
R5
O
O
CO2tBu
CO2tBu
CO2H
N
N
N
R1
7o,p,q
R1
H
20
19
Scheme 3. General synthesis of C3 N-linked pyrimidinediones. Reagents and
conditions: (a) DIEA, NMP, 110 °C; (b) KOtBu, THF, DMF, rt; (c) triphosgene, TEA,
Toluene, reflux; (d) trifluromethylbenzene, microwave, 150 °C; (e) NaOMe, MeOH,
microwave, 100 °C; (f) (Boc)2O, Dioxane, rt; (g) R1-Br, KOtBu, THF, DMF, rt; (h) TFA,
DCM, rt.
(EC50 = 10 lM). The discovery of the pyridone moiety at C3 which
anchored the inhibitor in the active site cavity of the enzyme
encouraged us to look for other similar functionalities in this re-
gion. A pyrimidinedione derivative 7c also displayed enzyme and
replicon activity comparable to the pyridone. The N1 aminobenzyl
group showed a positive impact on the binding activity from ear-
lier SAR (1l), but presented a potential aniline toxicology liability,
the corresponding aminopyridylmethyl analogs were prepared at
N1. Both C3 pyridone and pyrimidinedione analogs in this series
(7d,e) showed improved replicon activity while maintaining simi-
I
I
R5
R5
R5
a
b
CO2R
CO2R
CO2R
N
N
H
N
H
R1
3
21
8
e,f,g
R5
c
N
H
O
NH2
lar enzyme potency (EC50 = 2–4.8 lM). Five- and six-membered
sulfonyl urea analogs (7f,g) were both inactive, suggesting impro-
per alignment of the interacting groups.
N
R5
O
CO2R
N
d
R5
N
CO2H
O
R1
N
CO2R
R1
7a
The chloro group at R5 position protruded into a constricted
tunnel defined by primarily lipophilic residues as seen in the
X-ray. A limited effort to fine-tune the C5 functionality in the pyri-
done series (7h–m) showed intolerability to bulky groups while
small alkyls and halogens were preferred. The trifluoromethyl ana-
log at this position showed comparable binding affinity with an
23
N
n=1-2
R1
22
Cl
h,i
NHSO2Cl
n=1-2
n=1-2
NH
SO2
NH
N
j
R5
SO2
N
R5
CO2H
N
CO2R
R1
improved replicon activity (7m, EC50 = 4.8
lM) to the 5-chloro
N
R1
compound 7b (EC50 = 10 M). Further efforts to improve the cell
l
7f,g
24
based activity in this series were performed with a C5 trifluoro-
methyl group. The decreased enzyme activity of acyclic urea deriv-
ative 7n (IC50 = 0.35 lM) highlighted the importance of proper
alignment the amide pairs of the inhibitor and protein to evoke
the bidentate interaction to anchor the molecule. Efforts to con-
struct a stable cyclic structure which could align this urea moiety
properly were successful with the reverse pyrimidinedione analogs
7o,p. For example, the thienopyrimidinedione (7p, IC50 = 59 nM,
Scheme 4. General synthesis of C3 N-linked pyridone and sulfonyl urea. Reagents
and conditions: (a) NIS, CHCl3, rt; (b) R1-Br, Cs2CO3, DMF, rt; (c) CuI, K2CO3, 8-
hydroxyquinoline, DMSO, 100 °C; (d) aq LiOH, THF, reflux.; (e) nitric acid, (f)) R1-Br,
Cs2CO3, DMF, rt; (g) H2, Pd-C; (h) Et3N; (i) Cs2CO3, (j) aq LiOH, THF.
The reverse sequence of functionalizing N1 and C3 was also used in
some examples. The demethylation of intermediates 10 and 12
with 4 N HCl in dioxane liberated the pyridone and pyrimidinedi-
one moieties at C3 position. As above the final C2 ester hydrolysis
step was accomplished by use of aqueous LiOH.
Analogs with N-linked pyrimidinediones were synthesized as
described in Scheme 3 below. The 3-amino indole intermediate
15 was prepared in two steps from 2-fluorobenzonitrile 13. Inter-
mediate 15 was converted to the C3-isocyanate 16 with triphos-
gene, which was subsequently treated with appropriate aromatic
ortho-amino esters to form urea derivatives 17. Intermediate 17
was cyclized to the pyrimidinedione analog 18 by heating with so-
dium methoxide. Subsequent protection, N-alkylation and depro-
tection steps under standard conditions yielded the final
compounds 7o–q.
EC50 = 12 lM) showed the best activity profile. Since the C5 trifluo-
romethyl group showed improved replicon activity, we wanted to
examine the effect this could provide in combination with the N1
aminopyridylmethyl group. Gratifyingly both the thienopyrimidin-
edione (7q, IC50 = 32 nM, EC50 = 1.4
lM) and pyridone (7r,
IC50 = 17 nM, EC50 = 0.3 M) analogs in this series showed greatly
l
improved enzyme and cell based activity. The pyridone analog 7r
was the first compound in this series to afford sub-micromolar rep-
licon activity.
X-ray structures of representative inhibitors with C3-heterocy-
cles designed to have improved backbone interactions were ob-
served to bind to the active site of the enzyme ( Fig. 3). These
compounds exhibited the intended interactions with the backbone
atoms of Ile-447 and Tyr-448. Furthermore, the optimal aminopyr-
idylmethyl group was found to interact with the sidechain hydro-
xyl of Ser-367 and the C2-carboxlate group interacted with the
protein backbone at Gly-449 indirectly via a bridging water
molecule.
The syntheses of N-linked pyridone and cyclic sulfonyl urea
derivatives at C3 were accomplished as described in Scheme 4.
The key reaction step for synthesis of N-linked pyridone analog
was the copper iodide mediated N-arylation of intermediate 21
to 22. The cyclic sulfonyl urea derivatives were synthesized from

5340
G. N. Anilkumar et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5336–5341
Table 3
Variation of C3 hetero substitutions
R3
Cl
O
OH
N
R1
Compds
R5
Cl
R1
R3
IC50
(
l
M)7
EC50
19
(l
M)8
7a
0.15
N
O
F
NH
7b
7c
Cl
Cl
0.035
0.047
10
10
O
F
F
F
F
O
HN
NH
O
NH
7d
7e
7f
Cl
Cl
Cl
Cl
0.053
0.031
4.8
N
N
N
N
O
NH2
NH2
NH2
NH2
O
HN
NH
2.0
O
NH
S
O
>20
NA
N
O
NH
O
S
N
O
7g
>20
NA
NH
7h
7i
Br
0.043
7.6
O
F
F
F
F
F
F
F
F
F
F
F
F
NH
OMe
OEt
0.043
0.12
11
O
NH
7j
20
O
NH
7k
7l
0.56
>50
O
NH
Me
CF3
CF3
0.046
0.025
0.35
Cytotox@5
4.8
lM
O
NH
7m
7n
O
F
F
NH2
O
>10
HN

G. N. Anilkumar et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5336–5341
5341
Table 3 (continued)
Compds
R5
R1
R3
IC50
(
l
M)7
EC50
>10
(l
M)8
N
N
F
N
O
NH
7o
7p
7q
CF3
0.27
F
F
O
S
NH
CF3
CF3
CF3
0.059
0.032
0.017
12
F
O
O
N
O
S
NH
1.4
N
N
O
NH2
NH2
NH
O
7r
0.3
N
C3 pyridone 7r and pyrimidinedione 7q. Further optimization ef-
forts in this series will be reported in separate publications.
References and notes
1. (a) Cohen, J. Science 1999, 285, 26; (b) Alter, M. J.; Kruszon-Moran, D.; Nainan,
O. V.; McQuillan, G. M.; Gao, F.; Moyer, L. A.; Kaslow, R. A.; Margolis, H. S. N.
Engl. J. Med. 1999, 341, 556; (c) Cuthbert, J. A. Clin. Microbiol. Rev. 1994, 7, 505;
(d) Monto, A.; Wright, T. L. Semin. Oncol. 2001, 441.
2. (a) Neumann, A. U.; Lam, N. P.; Dahari, H.; Gretch, D. R.; Wiley, T. E.; Layden, T.
J.; Perelson, A. S. Science 1998, 282, 103; (b) Rosen, H. R.; Gretch, D. R. Mol. Med.
Today 1999, 5, 393; (c) Di Bisceglie, A. M.; McHutchison, J.; Rice, C. M.
Hepatology 2002, 35, 224.
3. Zeuzem, S.; Berg, T.; Moeller, B.; Hinrichsen, H.; Mauss, S.; Wedemeyer, H.;
Sarrazin, C.; Hueppe, D.; Zehnter, E.; Manns, M. P. J. Viral Hepat. 2009, 16, 75.
4. (a) Lesburg, C. A.; Cable, M. B.; Ferrari, E.; Hong, Z.; Mannarino, A. F.; Weber, P.
C. Nat. Struct. Biol. 1999, 6, 937; (b) Bressanelli, S.; Tomei, L.; Roussel, A.; Incitti,
I.; Vitale, R. L.; Mathieu, M.; De Francesco, R.; Rey, F. A. Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 13034.
5. Beaulieu, P. L. Expert Opin. Ther. Pat. 2009, 19, 145.
6. (a) Ferrari, E.; He, Z.; Palermo, R. E.; Huang, H. C. J. Biol. Chem. 2008, 283, 33893;
(b) Cheng, C. C.; Shipps, G. W., Jr.; Yang, Z.; Kawahata, N.; Lesburg, C. A.; Duca, J.
S.; Bandouveres, J.; Bracken, J. D.; Jiang, C. K.; Agrawal, S.; Ferrari, E.; Huang, H.
C. Bioorg. Med. Chem. Lett. 2010, 20, 2119.
7. HCV NS5B polymerase activity was measured in a radiolabeled nucleotide
incorporation assay as described [Cheng 2010], in a reaction buffer containing
20 mM HEPES (pH 7.3), 7.5 mM DTT, 20 U/mL RNasIN, 0.1
UTP, 60
Ci/ml [33P]-CTP supplemented to 20 nM CTP, 10 mM MgCl2, 60 mM
NaCl, 100 g/ml BSA, 100 nM heteropolymer RNA template, 0.25 mM
trinucleotide initiator and 30 nM NS5B ( 21) enzyme. Reaction was allowed
to proceed for 150 min at room temperature and terminated by EDTA. The
reaction mixture was washed on Millipore DE81 filter plate and the
incorporated labeled CTP quantitated by Packard TopCount. Compound IC50
values were calculated from experiments with 10 serial twofold dilutions of
the inhibitor in duplicate.
lM GTP, ATP and
l
l
D
Figure 3. X-ray structures of complexes with C3-heterocycles (orange: 7q;
magenta: 7r) superimposed upon the structure of NS5B complexed with 1 (yellow)
[9,10]. In the structures containing a C3 heterocycle, a water molecule was observed
(denoted ‘W’) which bridges the C2 carboxylate group and the backbone NH group
of Gly-449. The N1-aminopyridyl group interacts with the sidechain hydroxyl of
Ser-367.
8. To measure cell-based anti-HCV activity, replicon cells (1b-Con1) were seeded
at 5000 cells/well in 96-well plates one day prior to inhibitor treatment.
Various concentrations of an inhibitor in DMSO were added to the replicon
cells, with the final concentration of DMSO at 0.5% and fetal bovine serum at 5%
in the assay media. Cells were harvested 3 days post dosing. The replicon RNA
level was measured using real-time RT-PCR (Taqman assay) with GAPDH RNA
as endogenous control. EC50 values were calculated from experiments with 10
serial twofold dilutions of the inhibitor in duplicate.
9. Figures 1–3 were generated using the program PyMOL (The PyMOL Molecular
Graphics System, Version 1.2r1, Schrödinger, LLC.).
10. Crystal structures of HCV NS5B in complex with 1, 7r, and 7q have been
deposited in the Protein Data Bank with accession numbers 3SKH, 3SKE, 3SKA,
respectively.
In conclusion, we have described a structure-based drug design
assisted progression of a low affinity indole lead 1 with no detect-
able replicon activity to advanced leads with nanomolar NS5B en-
zyme potency and sub-micromolar replicon activity. The X-ray
structure of initial lead 1 in the palm site of NS5B suggested poten-
tial hydrogen bonding interactions with the protein backbone. Di-
rected SAR investigations to maximize the bidentate backbone
interactions led to the discovery of heterocyclic moieties such as