Cyclic amide bioisosterism: Strategic application to the design and synthesis of HCV NS5B polymerase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2010.06.008

Conformational modeling has been successfully applied to the design of cyclic bioisosteres used to replace a conformationally rigid amide bond in a series of thiophene carboxylate inhibitors of HCV NS5B polymerase. Select compounds were equipotent with th

Full text of DOI:10.1016/j.bmcl.2010.06.008

Bioorganic & Medicinal Chemistry Letters 20 (2010) 4614–4619
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Cyclic amide bioisosterism: Strategic application to the design and synthesis
of HCV NS5B polymerase inhibitors
a,
a
a
a,
Hanbiao Yang a, , Robert T. Hendricks , Nidhi Arora , Dov Nitzan , Calvin Yee , Matthew C. Lucas a,
Yanli Yang b, Amy Fung b, Sonal Rajyaguru b, Seth F. Harris c,à, Vincent J. P. Leveque b, Julie Q. Hang b,
Sophie Le Pogam b, Deborah Reuter a, Gisele A. Tavares c,§
,
*
*
a Medicinal Chemistry Department, Roche Palo Alto, LLC, 3431 Hillview Avenue, Palo Alto, CA 94304, United States
b Viral Diseases Discovery and Translational Area, Roche Palo Alto, LLC, 3431 Hillview Avenue, Palo Alto, CA 94304, United States
c Discovery Technologies X-ray Crystallography Group, Roche Palo Alto, LLC, 3431 Hillview Avenue, Palo Alto, CA 94304, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Conformational modeling has been successfully applied to the design of cyclic bioisosteres used to
replace a conformationally rigid amide bond in a series of thiophene carboxylate inhibitors of HCV
NS5B polymerase. Select compounds were equipotent with the original amide series. Single-point mutant
binding studies, in combination with inhibition structure–activity relationships, suggest this new series
interacts at the Thumb-II domain of NS5B. Inhibitor binding at the Thumb-II site was ultimately con-
firmed by solving a crystal structure of 8b complexed with NS5B.
Received 4 April 2010
Revised 26 May 2010
Accepted 1 June 2010
Available online 8 June 2010
Keywords:
Amide bioisostere
Modeling
Ó 2010 Elsevier Ltd. All rights reserved.
Synthesis
HCV
NS5B polymerase
Thumb
Mutant
3mt5
X-ray structure
PDB
Ruthenium catalyst
Hepatitis C virus (HCV) is a major threat to public health world-
wide as an estimated 3% of the world’s population is infected.
Approximately 15–20% of infected individuals clear the virus with-
out treatment, while the remaining 75–85% become chronically in-
fected.1 Chronic HCV infection has become a leading cause of
cirrhosis, hepatocellular carcinoma and liver transplantation in
the US.2 The current standard of care treatment for HCV infection
US. In addition, patient compliance is often limited by severe side
effects. Based on these limitations, there is still a significant unmet
clinical need for safer and more effective anti-HCV therapies.
HCV contains a positive-sense, single-strand RNA molecule of
9.6 kb with one open reading frame coding for a large polyprotein
of ꢀ3000 amino acids. Intracellular processing of the polyprotein
yields at least ten individual functional proteins, namely C, E1,
E2, P7, NS2, NS3, NS4A, NS4B, NS5A, and NS5B. Although the best
path for HCV treatment has yet to be defined,3 it will likely require
a combination therapy approach targeting multiple viral proteins
and/or pathways in the HCV lifecycle in order to effectively sup-
press resistance.
consists of a combination dose of pegylated interferon-
a (PEG-
IFN ) along with the broad spectrum antiviral agent Ribavirin for
a
up to 48 weeks. However, only a moderate sustained virological re-
sponse (SVR) of 42–46% is achieved in patients infected with geno-
type-1 virus, the genotype which accounts for ꢀ70% of cases in the
One particular HCV target protein of interest is NS5B, an RNA-
dependent RNA polymerase (RdRp) that is essential for the synthe-
sis and replication of viral RNA. It has a three-dimensional shape
similar to other known RNA polymerases and is often described
as a right-handed motif, with distinct Palm, Finger, and Thumb do-
mains. Several inhibitors targeting various known binding sites
within HCV NS5B polymerase have demonstrated useful efficacy
in clinical trials.4
* Corresponding authors. Tel.: +1 650 361 1595.
E-mail addresses: Hanbiaoyang@gmail.com (H. Yang), RTH.Hendricks@gmail.-
com (R.T. Hendricks).
Present address: Hoffmann-La Roche Inc., 340 Kingsland Street, Nutley, NJ 07110,
United States.
à
Present address: Roche Group, Genentech, Inc., 1 DNA Way, South San Francisco,
CA 94080, United States.
§
Present address: Novartis Pharma AG, NIBR, CHBS, WSJ-210.5.13, Switzerland.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.06.008

H. Yang et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4614–4619
4615
At Roche Palo Alto we initiated a program to identify new inhib-
O
O
S
2
S
Ph
itors targeting the Thumb-II binding site, which is located near the
base of the Thumb domain and is approximately 35 Å away from
the active site.4,5 Several NS5B inhibitor chemo types, including
phenylalanine 1,6 thiophenecarboxylic acid 2,5a,b pyranoindole 3,7
4-hydroxydihydropyranone 4,8 and recently 1H-benzo[de]isoquin-
oline-1,3-(2H)-dione 59 (Fig. 1), have been shown to bind to this
hydrophobic pocket. In this Letter, we report the design, synthesis
and biological evaluation of a new series of thiophenecarboxylic
acid analogs derived from the prototypical compound 2.
Ph
Bioisosteric replacement
OH
OH
3
4N
R
5
R
O
aryl or heteroaryl
Figure 2. Bioisosteric amide bond replacement strategy.
It has been shown that both the amide linkage and the isopro-
pyl group of 2 are important structural elements for enzymatic and
replicon potency.5a We previously reported the X-ray structure of 2
bound to HCV NS5B polymerase5c which shows the amide carbonyl
is involved in a single water-mediated hydrogen bond to Trp528,
while the isopropyl group is fully exposed to solvent. However, this
isopropyl group is believed to play two important conformational
roles: (1) orienting the amide bond in a cis-like (i.e., relative to
the des-isopropyl analog) or E-configuration; (2) positioning the
plane of the amide bond nearly orthogonal relative to the thio-
phene ring. A similar conformation has been observed for the cor-
responding aromatic amides 6 and 7 when complexed to NS5B.5b
In this conformation, the trans-4-methylcyclohexyl group is favor-
ably positioned in a deep hydrophobic pocket and makes extensive
interactions with residues Arg422, Met423, Leu474, His475,
Tyr477, Lys501, and Trp528.
Bioisosteric replacement is an important tool in medicinal
chemistry that can be used to improve the potency, selectivity
and pharmacokinetics of key compounds or to generate new intel-
lectual property (IP).10 In particular, replacement of amide bonds
with suitable bioisosteres that maintain similar geometric, elec-
tronic, or hydrogen bonding properties have been extensively uti-
lized.11 Based on our structural analyses of the NS5B complexes
with inhibitors 2, 6, and 7,12 our strategy centered on replacement
of the amide bond with a bioisosteric cyclic ring system (Fig. 2).
The primary small molecule conformational elements considered
were the dihedral angle between the thiophene ring and the amide
plane (as defined by atoms C2–C3–N4–C5), and the exit vector for
the hydrophobic ‘R’ group.
Table 1
Dihedral angle comparison: X-ray data versus modeling analysis
S
S
S
2
3
CO2H
CO2H
CO2H
Ph
Ph
Ph
N
4
N
N
5
O
O
O
Cl
2
Cl
6
7
S
2
S
CO2H
Ph
S
2
2
CO2H
Ph
CO2H
Ph
3
N4
3
3
H
N
4
4
5
5N
5
R
N
O
R
R
N
8
10
9
Compd
R-group
X-raya
Modeleda,b
2
6
7
8b
9
93
75
80
—
63
49
59
96
47
70
79
4-Me-cyclohexyl
Tolyl
—
10b
10d
Tolyl
4-Me-cyclohexyl
—
—
a
Measured dihedral angle: C2–C3–N4–C5.
Angles noted are for each Thumb-II binding site-relevant lowest energy con-
b
formation within 5 kcal/mol of the observed global minimum.
We performed a small molecule conformational analysis on
several proposed analogs (8–10; Table 1) in order to evaluate the
impact of aryl or heteroaryl ring substitution on both the key dihe-
dral angle and the positioning of the hydrophobic R-group. Confor-
mational analysis was performed in Maestro13 using MMFFs
parameters in water with a distance dependant dielectric constant.
Torsional sampling was performed using 5000 steps of Monte Car-
lo. We narrowed our analysis to include only those observed con-
formations which were (a) within 5 kcal/mol of the global
minimum conformation; and (b) relevant to binding in the NS5B
Thumb-II site. Our results reasonably predicted the observed (via
X-ray) dihedral angles for molecules 2, 6 and 7 to within 30° (Ta-
ble 1). Compared to 6 and 7, the cyclohexyl group of 2 led to a
wider distribution of observed relevant conformations. The mod-
eled results for 6 and 7 match closely with what has been previ-
ously reported5a and we observed
a
relatively narrow
distribution of relevant conformations. Our analysis of the 4-
methyl-cyclohexylphenyl analog 8b showed a very narrow distri-
bution of conformations which strongly suggest a nearly orthogo-
nal relationship (96°) between the phenyl and thiophene rings is to
be expected. Examination of the N-linked pyrazalone 9 predicts an
angle that is close to that observed for compound 6; however, this
particular conformation is 4.7 kcal/mol higher in energy than the
observed global minimum for this structure. Favorable electro-
static interactions between the carboxylate and the pyrazalone-
NH in 9 appear to limit the number of binding site-relevant confor-
mations that are accessible in this molecule. The triazole results
(10b, d) match well with comparable analogs in the amide series
and 10d appears to have a somewhat smaller distribution of avail-
able relevant conformations. In all cases the general alignment of
the R-group exit vectors matched very well with the X-ray data
for 2 (Fig. 3).
O
O
N
CN
HO
S
O
Cl
O
O
OH
OH
N
N
H
Cl
O
CF3
3
2
1
OH
O
O
N
N
N
N
N
Br
O
HO
N
N
O
Encouraged by our initial conformational analysis, we prepared
two six-membered ring analogs in which the amide group was re-
placed by a simple phenyl ring as a preliminary test of our hypoth-
esis (Table 2). Analogs 8a and 8b were prepared in a similar fashion
H
4
5
Figure 1. Known Thumb-II domain NS5B inhibitors.

4616
H. Yang et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4614–4619
Figure 3. (a) Overlay of 8b lowest energy conformation (yellow) with X-ray bound structure of 2 (NS5B bound conformation in blue; red spheres are water); (b) overlay of
lowest energy relevant conformation of 10b (orange) with NS5B bound structure of 2.
Table 2
Sixmembered ringa amide replacement structure–activity relationships and inhibi-
tory activities against HCV NS5B
HO
O
B
HO
Br
a, b
c, d
O
O
3-TC
3-TC
3-TC
3-TC
N
R
NH
R
R
O
R
11
N
N N
H
14
O
O
S
8
13
15
Ph
S
OH
Ph
f, g
+
11
NS5Bb
Clog P
OH
Compd
R-group
2
8a
0.028
16
0.042
0.022
0.17
92
7.55
5.8
X
O-n-Butyl
X = NH2
8b
13a
13b
14
15
trans-4-Me-cyclohexyl
trans-4-Me-cyclohexyl
4-Me-cyclohex-1-enyl
Phenyl
7.67
6.22
6.21
2.97
3.47
e
X = Br (12
)
8b
Scheme 1. Reagents and conditions: (a) n-BuLi (2 equiv), À78 °C, Et2O, then trans-
4-methylcyclohexanone, À78 °C to À5 °C, 1 h (49%); (b) Et3SiH, TFA, rt, DCM (79%);
(c) Tf2O, pyridine/DCM (89%); (d) bis(pinacolato)diboron (1.1 equiv), PdCl2dppf
(0.08 equiv), dppf (0.16 equiv), KOAc (3 equiv), DMF, 100 °C (15%); (e) t-BuONO,
CuBr2, ACN, 0 °C, 2 h; (f) Pd(PPh3)4, Na2CO3, Bu4NBr, tol/H2O, 90 °C (70%); (g) LiOH,
THF/MeOH/H2O.
4-Tolyl
3.3
a
3-TC = 3-substituted 5-phenylthiophene-2-carboxylic acid.
GT-1b NS5B inhibition (IC50, lM).
b
by coupling bromide 12 with the appropriate aryl boron reagent
under Suzuki coupling conditions. The synthesis of 8b is outlined
in Scheme 1. It should be noted that the choice of solvent in the
preparation of 4-methylcyclohexyl phenyl boronate 11 was crucial.
A useful, albeit low, yield was obtained in DMF, whereas no prod-
uct was detected in dioxane. The assay results using GT-1b NS5B
(20 nM enzyme, poly-A RNA template)14 are summarized in Ta-
ble 2. Although the flexible n-butyl ether analog 8a was inactive,
we were quite pleased to see that 4-methyl-cyclohexyl substituted
8b showed inhibitory potency that was nearly equal to that of the
prototypical amide 2, and provided us with a convincing proof of
concept for our primary hypothesis.
In addition to using phenyl as a conformational mimic of the
amide linkage in 2, we were also interested in exploring the impor-
tance of additional polar group interactions. In particular, specific
analogs were designed to mimic the water-mediated hydrogen
bond of the amide carbonyl to Trp528, observed in the NS5B X-
ray bound complex with 2 (Fig. 3a).12 Along with the prospect of
improved potency, the incorporation of heteroatoms into the link-
age provides a method of reducing Clog P, thus improving the over-
all drug-like properties of this series. Towards this end we first
examined a set of six-membered heterocyclic analogs (13–15; Ta-
ble 2) capable of mimicking the hydrogen bond acceptor properties
of the amide carbonyl. These compounds were prepared using an
alternative coupling route highlighted in Scheme 2. Attempts to
prepare either the 2-(4-methyl-cyclohexyl)pyridyl-3-boronate or
3-boronic acid precursor to 13a were unsuccessful, however, we
were able to successfully convert bromide 12 into the requisite
boronate 17 and boronic acid 18 via palladium catalyzed boryla-
tion. Conversion of the unsaturated SEM intermediate 16 to the
corresponding triflate, followed by coupling with 17 led to 13b in
good yield. Palladium catalyzed coupling of SEM-protected 3,4-di-
chloro-6-pyridazinone with 18, followed by coupling with phenyl-
boronic acid provided 14. By reversing the order of coupling (i.e., 4-
tolylboronic acid first, then 18), we were able to prepare the iso-
meric pyridazinone 15 in similar overall yield.
Encouragingly, the 3-pyridyl analog 13a was equipotent with 2,
and was ꢀ1.3 log units less lipophilic based on Clog P calcula-
tions16 (Table 2). On the other hand, the corresponding olefin ana-
log 13b was eight-fold less active, and the 4-, 5-, and 6-pyridyl
variants17 (data not shown) all had roughly similar potency to
13b. The notably more polar pyridazinones 14 and 15 were signif-
icantly less active. We note that 14 lacks 4-substitution, which has
been shown to limit potency in related series.5a The isomeric
pyridazinone 15 has a 4-tolyl group, but is still two orders of mag-
nitude less active relative to pyridyl 13a.

H. Yang et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4614–4619
4617
S
O
Ph
CO2H
SEMO
TfO
SEMO
c, d, e
a, b
f, g
+
N
N
Br
N
S
N
16
13a
O
S
Ph
CO2H
SEM
S
j, k, l, g
h
N
N
Ph
CO2Me
i
Ph
CO2Me
O
+
12
O
Cl
B(OH)2
B
O
Cl
N
H
N
18
17
14
Scheme 2. Reagents and conditions: (a) n-BuLi, THF, À78 °C (50%); (b) Burgess reagent,15 benzene, reflux, 3 h (72%); (c) H2 (1 atm), PtO2, HOAc/EtOAc (1:6) (75%); (d) dil.
H2SO4, MeOH, rt, 24 h; (e) Tf2O, pyridine/DCM (78% for d + e); (f) 17, Pd(PPh3)4, K3PO4 (2 equiv), dioxane, 85 °C (89%); (g) 2 M LiOH, MeOH/THF, reflux, 1 h (90%; 70% for 14);
(h) Pd(PPh3)4, Et3N, THF, pinacolborane, 40 °C, 2 h; (i) AcOH/H2O/THF (63%); (j) Pd(PPh3)4, 2 M Na2CO3 (aq), MeOH, reflux (65%); (k) phenylboronic acid, Pd(PPh3)4, 2 M
Na2CO3 (aq), MeOH, reflux (39%); (l) HCl/MeOH, reflux, 24 h.
Our preliminary small molecule conformational analysis sug-
gested the various five-membered ring exit vectors, though gener-
ally similar to those of the six-membered ring analogs, offered
subtle variations in the positioning of the hydrophobic R-group.
Furthermore, five-membered heterocycles provide additional op-
tions for mimicking the amide carbonyl. We were further encour-
aged by recent results demonstrating a successful bioisosteric
exchange of an amide by a triazole ring.11c We chose to prepare
a limited set of five-membered ring analogs (9, 10, 26; Table 3)
in order to test this hypothesis.
The N-linked 3-pyrazalone 9 was prepared from the corre-
sponding 3-aminothiophene carboxylic ester as summarized in
Scheme 3. The mixed anhydride intermediate was readily prepared
starting from 4-methyl acetophenone and dimethyl malonate. Mild
acid catalyzed cyclization of the coupled hydrazide 19 gave 9 in
modest yield following saponification. In comparison to 15 (the
six-membered ring homolog), 9 displayed a five-fold enhancement
in potency (Table 3, IC50 = 0.62 lM).
The triazole and isoxazole analogs (10, 25; Table 3) were syn-
thesized in a convergent fashion via a Ru-catalyzed coupling.18
Specifically, the cycloaddition of alkyne 20 with alkyl or aryl azides
21a–d (see Table 3 for descriptions of a–d) gave the desired cyclo-
addition products 22a–d in low to moderate yields as a single reg-
ioisomer (Scheme 4). In accordance with the observation in the
literature, the less reactive alkyl azides 21c and 21d required the
Table 3
Fivemembered ring amide replacements:a structure–activity relationships and
inhibitory activities against HCV NS5B
*
use of the more reactive Cp Ru(COD) catalyst.
3-TC
3-TC
3-TC
H
In contrast, yields for the isoxazole synthesis shown in Scheme 5
were much higher (55–98%), except in the case of 24d (13% yield
plus 8% of the 1,4-regioisomer 26). We attributed the low yield
and poor selectivity in this case to the impure nature of chloro-
oxime 23d, which was prepared in five steps without purification
(Scheme 6). We suspect impurities present in 23d suppressed the
catalyzed process by poisoning the catalyst, therefore allowing a
background reaction leading to the 1,4-isoxazole regioisomer 26
to become a competing process.
The structure–activity relationships for the triazole and isoxaz-
ole series are consistent with what has been observed in the proto-
typical thiophene amide series: increasing steric bulk of the R-
group leads to increased inhibitory potency (Table 3; 10a, 25a vs
10b, 25b–e). Simply adding a 4-fluoro substituent (25f) gave a
modest five-fold improvement over the phenyl analog (25a). How-
ever, in contrast to the amide-linked series the addition of a trans-
4-methyl group to the cyclohexyl ring provided only a three-fold
improvement in the triazole series (10d), and showed essentially
no change in potency in the isoxazole series (25c,d). Both 4-meth-
N N
O
N
N
R
R
O
R
N
10
N
25
9
Compd
R-group
NS5Bb
Clog P
2
9
0.028
0.62
17
0.11
0.34
0.10
29
7.55
5.66
4.62
5.12
4.43
4.95
4.36
4.86
5.18
5.70
5.07
4.50
6.21
4-Tolyl
Phenyl
4-Tolyl
Cyclohexyl
10a
10b
10c
10d
25a
25b
25c
25d
25e
25f
27
trans-4-Me-cyclohexyl
Phenyl
4-Tolyl
Cyclohexyl
2.5
0.074
0.11
2.6
trans-4-Me-cyclohexyl
4-Chlorophenyl
4-Fluorophenyl
5.7
32
a
3-TC = 3-substituted 5-phenylthiophene-2-carboxylic acid.
b
GT-1b NS5B inhibition (IC50, lM).
O
O
S
O
O
Ph
CO2Me
S
Ph
CO2H
N
EtO
O
S
Ph
CO2Me
d, e
N
N
N
O H
c
NHR1
R1 = H
R1 = NH2
O
O
a, b
19
9
Scheme 3. Reagents and conditions: (a) NaNO2, HCl; (b) SnCl2, H2O (60%); (c) DCM, rt, then HCl (aq)/THF (43%); (d) p-TsOH/benzene, 60 °C; (e) LiOH, THF/MeOH/H2O, rt
(36%).

4618
H. Yang et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4614–4619
CO2R2
S
S
CO2Me
R1
Ph
Ph
c
a
12
R-N3
21
N
N
R1 = TMS
R
N
b
R1 = H (20)
R2 = Me (22)
R2 = H (10)
d
Scheme 4. Reagents and conditions: (a) trimethylsilyl acetylene, PdCl2(PPh3)2
(10 mol %), i-Pr2NEt, CuI, THF, 85 °C, overnight; (b) K2CO3, MeOH, rt, (90% for a + b);
(c) condition 1: Cp*RuCl(PPh3)2 (5 mol %), benzene, 60 °C, overnight, 22a (17%); 22b
(19%); Condition 2: Cp*Ru(COD) (2 mol %), toluene, rt, 16 h, 22c (28%); 22d (38%);
(d) KOH, MeOH or LiOH, THF/H2O.
CO2R1
S
Ph
CO2R1
S
Ph
a
20
+
OH
Cl
O
N
O
N
R
N
R
23
R1 = Me (24)
R1 = H (25)
R1 = Me (26)
R1 = H (27)
b
b
*
Scheme 5. Reagents and conditions: (a) Cp Ru(COD), Et3N, DCE, rt, 16 h; (b) KOH,
MeOH or LiOH, THF/H2O.
Figure 4. NS5B co-crystal complex with compound 8b; PDB code 3MF5.
O
NOH
NOH
Cl
c
d
(dihedral angle: 96.7°), nearly identical to our conformational anal-
ysis prediction. The bridging water observed in the NS5B structure
with amide 2 has been displaced by the linking aryl ring. A new
water is observed near the carboxyl binding region (3.0 Å to the
nearest carboxyl oxygen). This water is also near His475, but the
closest distance is 3.8 Å, suggesting the His475 interaction may
be only weakly stabilized through hydrogen bonding. A second
water is observed deeper in the lipophilic binding pocket that is in-
volved in a stabilizing hydrogen bond (2.8 Å) with Lys533.
R
d
H
R = OH
R = Cl
a
23d
R = H
Scheme 6. Reagents and conditions: (a) oxalyl chloride, DMF/DCM; (b) 10% Pd/C,
2,6-lutidine; (c) hydroxylamine–HCl, NaOAc, H2O, rt; (d) NCS, DMF, cat. HCl, rt.
ylcyclohexyl analogs were only four-fold less potent than the pro-
totypical compound 2. Not surprisingly, the 3,5-isoxazole isomer
27 showed essentially no inhibitory potency.
Amide 2 has been reported to show good inhibition in a cell-
l
M)5a and we have independently
In order to confirm that this series of NS5B inhibitors were in-
deed binding at the putative Thumb-II binding site, we performed
a structure–activity analysis based on binding affinity with 10b
and 10d using two well characterized NS5B single-point mutants:
NS5B L419M and NS5B M414T.19,5c The L419M mutation displays
resistance against inhibitors that bind in the Thumb-II region,
whereas the M414T mutation displays resistance against inhibitors
that bind in the Palm region. The amide-linked analog 2 was used
as a positive control. The data shown in Table 4 clearly demon-
strate that only the Thumb region mutation significantly impacted
binding: both triazole analogs showed ꢀfive-fold reduction in
binding affinity relative to WT binding and showed no change in
affinity towards the Palm region mutant. The amide control 2
showed a 10-fold loss in affinity for L419M and is similarly un-
changed by the M414T mutation.
based replicon assay (EC50 = 0.6
confirmed similar potency in our own GT-1b replicon assay14
(EC50 = 0.23 M; CC50 = 48 M). Regardless of the generally good
l
l
predicted permeability and a wide range of estimated polarity
(i.e., Clog P) across our series of NS5B-active amide replacements,
selected compounds tested in a GT-1b replicon assay showed gen-
erally weak replicon activity and significant cytotoxicity, with only
low selectivity indices (CC50/EC50 <2). For example, compound 8b
showed an EC50 of 12
entiated from toxicity (CC50 = 20
(EC50 and CC50 >30 M) and 10d showed only weak activity
(EC50 = 15 M; CC50 >50 M). It has been suggested that within
lM, but this could not be reasonably differ-
lM). Triazole 10b was inactive
l
l
l
an acidic series, replicon potency is reduced with decreasing
pKa.21 The measured pKa of 2 was found to be 3.8, whereas the
pKa of 10d was 3.6, suggesting that pKa alone is not a significant
contributing factor in explaining the limited replicon potency for
this series.
In summary, we describe our efforts towards successful bioisos-
teric replacements of the rigidly oriented amide bond in a previ-
ously reported series of HCV NS5B inhibitors. We performed a
preliminary conformational analysis on proposed structures de-
signed to mimic the shape and electronic nature of the amide link-
age and prepared a series of ortho-substituted aryl and heteroaryl
analogs. Members of the series possess excellent inhibitory po-
tency against HCV NS5B and demonstrated structure–activity rela-
tionships consistent with the prototypical amide series. Data from
both a single-point NS5B mutant binding analysis and an X-ray co-
crystal structure confirm that these cyclic amide mimics bind at
the NS5B Thumb-II domain.
Further confirmation of our binding hypothesis was provided by
a co-crystal structure of the 4-methyl-cyclohexylphenyl analog 8b
complexed to NS5B (Fig. 4), obtained through a crystal soaking
experiment.5c,20 The structure of the protein-inhibitor complex
clearly shows 8b bound at the Thumb-II site with the phenyl ring
positioned orthogonal relative to the plane of the thiophene ring
Table 4
Thumb versus Palm region binding affinity (KD, lM)
Compd
NS5B WT
L419M
M414T
2
0.041 0.011
4.32 0.555
0.86 0.365
0.427 0.111
23.09 2.87
4.48 0.581
0.033 0.008
3.994 0.91
0.597 0.218
10b
10d

H. Yang et al. / Bioorg. Med. Chem. Lett. 20 (2010) 4614–4619
4619
Carroll, S. S.; Olsen, D. B.; De Francesco, R.; Altamura, S.; Migliaccio, G.; Carfì, A.
J. Med. Chem. 2009, 52, 5217.
Acknowledgments
10. Patani, G. A.; LaVoie, E. J. Chem. Rev. 1996, 96, 3147.
The authors would like to thank Steve Swallow, Connie Oshiro
and George Adjabeng for contributions to the design and synthesis
of these analogs, and Dave Rotstein and Francisco Talamas for their
helpful discussions.
11. For recent examples, see: (a) Black, W. C.; Bayly, C. I.; Davis, D. E.; Desmarais,
S.; Falgueyret, J.-P.; Léger, S.; Li, C. S.; Massé, F.; McKay, D. J.; Palmer, J. T.;
Percival, M. D.; Robichaud, J.; Tsou, N.; Zamboni, R. Bioorg. Med. Chem. Lett.
2005, 15, 4741; (b) Appendino, G.; Bacchiega, S.; Minassi, A.; Cascio, M. G.;
Petrocellis, L. D.; Di Marzo, V. Angew. Chem., Int. Ed. 2007, 46, 9312; (c) Horne,
W. S.; Olsen, C. A.; Beierle, J. M.; Montero, A.; Ghadiri, M. R. Angew. Chem., Int.
Ed. 2009, 48, 4718.
References and notes
12. PDB accession codes: 2: 2GIR; 6: 1YVX; 7: 1YVZ.
13. Maestro, version 8.5, Schrödinger LLC., New York, NY, 2009
(www.schrodinger.com).
1. http://www.cdc.gov/hepatitis/HCV/index.htm.
2. Brown, R. S. Nature 2005, 436, 973.
3. Manns, M. P.; Foster, G. R.; Rockstroh, J. K.; Zeuzen, S.; Zoulim, F.; Houghton, M.
P. Nat. Rev. Drug Disc. 2007, 6, 1991.
4. Beaulieu, P. L. Curr. Opin. Invest. Drugs 2007, 8, 614.
14. (a) Klumpp, K.; Leveque, V.; Le Pogam, S.; Ma, H.; Jiang, W.-R.; Kang, H.;
Granycome, C.; Singer, M.; Laxton, C.; Hang, J. Q.; Sarma, K.; Smith, D. B.;
Heindl, D.; Hobbs, C. J.; Merrett, J. H.; Symons, J.; Cammack, N.; Martin, J. A.;
Devos, R.; Najera, I. J. Biol. Chem. 2006, 281, 3793; (b) Ma, H.; Leveque, V.; De
Witte, A.; Li, W.; Hendricks, T.; Clausen, S. M.; Cammack, N.; Klumpp, K.
Virology 2005, 332, 8.
15. Atkins, G. M., Jr.; Burgess, E. M. J. Am. Chem. Soc. 1968, 90, 4744.
16. Clog P and predicted permeability values calculated using proprietary Roche
tools.
17. Attempts to hydrogenate the olefin in the precursors needed to prepare the 4-,
5, and 6-pyridyl analogs of 13a were unsuccessful, giving either no reaction
(<10%; inseparable) or over-reduction of the pyridine ring.
18. (a) Boren, B. C.; Narayan, S.; Rasmussen, L. K.; Zhang, L.; Zhao, H.; Lin, Z.; Jia, G.;
Fokin, V. V. J. Am. Chem. Soc. 2008, 130, 8923; (b) Grecian, S.; Fokin, V. V. Angew.
Chem., Int. Ed. 2008, 47, 8285.
19. Hang, J. Q.; Yang, Y.; Harris, S. F.; Leveque, V.; Whittington, H. J.; Rajyaguru,
S.; Ao-Ieong, G.; McCown, M. F.; Wong, A.; Giannetti, A. M.; Le Pogam, S.;
Talamas, F.; Cammack, N.; Najera, I.; Klumpp, K. J. Biol. Chem. 2009, 284,
15517.
20. The crystal structure coordinates and structure factor data have been deposited
in the PDB: accession code 3MF5.
21. Li, H.; Linton, A.; Tatlock, J.; Gonzalez, J.; Borchardt, A.; Abreo, M.; Jewell, T.;
Patel, L.; Drowns, M.; Ludlum, S.; Goble, M.; Yang, M.; Blazel, J.; Rahavendran,
R.; Skor, H.; Shi, S.; Lewis, C.; Fuhrman, S. J. Med. Chem. 2007, 50, 3969.
5. (a) Chan, L.; Pereira, O.; Reddy, T. J.; Das, S. K.; Poisson, C.; Courchesne, M.;
Proulx, M.; Siddiqui, A.; Yannopoulos, C. G.; Nguyen-Ba, N.; Roy, C.; Nasturica,
D.; Moinet, C.; Bethell, R.; Hamel, M.; L’Heureux, L.; David, M.; Nicolas, O.;
Courtemanche-Asselin, P.; Brunette, S.; Bilimoria, D.; Bédard, J. Bioorg. Med.
Chem. Lett. 2004, 14, 797; (b) Biswal, B. K.; Cherney, M. M.; Wang, M.; Chan, L.;
Yannopoulos, C. G.; Bilimoria, D.; Nicolas, O.; Bedard, J.; James, M. N. G. J. Biol.
Chem. 2005, 280, 18202; (c) Le Pogam, S.; Kang, H.; Harris, S. F.; Leveque, V.;
Giannetti, A. M.; Ali, S.; Jiang, W.-R.; Rajyaguru, S.; Tavares, G.; Oshiro, C.;
Hendricks, T.; Klumpp, K.; Symons, J.; Browner, M. F.; Cammack, N.; Nájera, I. J.
Virol. 2006, 80, 6146.
6. Wang, M.; Ng, K. N. S.; Cherney, M. M.; Chan, L.; Yannopoulos, C. G.; Bédard, J.;
Morin, N.; Nguyen-Ba, N.; Bethell, R. C.; Alaoui-Ismaili, M. H.; James, M. N. G. J.
Biol. Chem. 2003, 278, 9489.
7. Howe, A. Y. M.; Cheng, H.; Thompson, I.; Chunduru, S. K.; Herrmann, S.;
O’Connell, J.; Agarwal, A.; Chopra, R.; Vecchio, A. M. D. Antimicrob. Agents
Chemother. 2006, 50, 4103.
8. Li, H.; Tatlock, J.; Linton, A.; Gonzalez, J.; Jewell, T.; Patel, L.; Ludlum, S.; Drowns,
M.; Rahavendran, S. V.; Skor, H.; Hunter, R.; Shi, S. T.; Herlihy, K. T.; Parge, H.;
Hicket, M.; Yu, X.; Chau, F.; Nonomiya, J.; Lewis, C. J. Med. Chem. 2009, 52, 1225.
9. Ontoria, J. M.; Rydberg, E. H.; Di Marco, S.; Tomei, L.; Attenni, B.; Malancona, S.;
Hernando, J. I. M.; Gennari, N.; Koch, U.; Narjes, F.; Rowley, M.; Summa, V.;