Non-nucleoside inhibitors of the hepatitis C virus NS5B polymerase: Discovery of benzimidazole 5-carboxylic amide derivatives with low-nanomolar potency

Article abstract of DOI:10.1016/j.bmcl.2003.12.032

Optimization of benzimidazole 5-carboxamide derivatives previously identified as specific inhibitors of the NS5B polymerase of the hepatitis C virus (HCV) has led to the discovery of potent analogues that inhibit the enzyme at low-nanomolar concentrations. Greater than 800-fold improvement in potency from the original lead structure was achieved through the combined effects of conformational rigidification, molecular size extension and the identification of previously unexploited interactions. Furthermore, these inhibitors retain specificity for HCV polymerase relative to other viral and mammalian RNA polymerases.

Full text of DOI:10.1016/j.bmcl.2003.12.032

Bioorganic & Medicinal Chemistry Letters 14 (2004) 967–971
Non-nucleoside inhibitors of the hepatitis C virus NS5B
polymerase: discovery of benzimidazole 5-carboxylic amide
derivatives with low-nanomolar potency
Pierre L. Beaulieu,a,* Michael Bos,a Yves Bousquet,a Patrick DeRoy,a,y Gulrez Fazal,a
Jean Gauthier,a James Gillard,a Sylvie Goulet,a Ginette McKercher,b
Marc-Andre Poupart,a Serge Valoisa and George Kukoljb
aDepartment of Chemistry, Boehringer Ingelheim (Canada) Ltd, Research and Development,
´
2100 Cunard Street, Laval, Quebec, Canada H7S2G5
bDepartment of Biological Sciences, Boehringer Ingelheim (Canada) Ltd, Research and Development,
´
2100 Cunard Street, Laval, Quebec, Canada H7S2G5
Received 15 September 2003; accepted 1 December 2003
Abstract—Optimization of benzimidazole 5-carboxamide derivatives previously identified as specific inhibitors of the NS5B
polymerase of the hepatitis C virus (HCV) has led to the discovery of potent analogues that inhibit the enzyme at low-nanomolar
concentrations. Greater than 800-fold improvement in potency from the original lead structure was achieved through the combined
effects of conformational rigidification, molecular size extension and the identification of previously unexploited interactions.
Furthermore, these inhibitors retain specificity for HCV polymerase relative to other viral and mammalian RNA polymerases.
# 2003 Elsevier Ltd. All rights reserved.
The hepatitis C virus has infected an estimated 1–3% of
the world population, exceeding 170 million indivi-
duals.1,2 In up to 70–80% of incidences the infection is
chronic and may progress to conditions that require
orthotopic liver transplantation as
a
life-saving
measure.2 Current therapies based on combinations of
pegylated interferons and the broad spectrumantiviral
ribavirin are ineffective in a significant proportion of
cases and are associated with the occurrence of severe
side effects.3 A growing patient population is in urgent
need for novel therapies that could address a clearly
unmet medical need.
Figure 1. Benzimidazole-based inhibitor of HCV NS5B polymerase.
5-carboxylic acid derivatives which are specific inhibi-
tors of HCV polymerase.5 Compound 1, a representa-
tive of this novel class of non-nucleoside inhibitors, had
an IC50 in the low-micromolar range and delineates the
minimum core for activity against the enzyme (Fig. 1).
The HCV NS5B RNA-dependent RNA polymerase is a
central enzyme in the replication of the virus and has
become a target of choice for the screening and design
of small molecule inhibitors which, in principle, should
interfere with viral replication.4 We recently described
the discovery and initial optimization of benzimidazole
This class of inhibitors was shown to inhibit polymerase
activity in a non-competitive fashion with respect to
nucleotide incorporation. Based on evidence obtained
from several biochemical experiments, the most likely
mode of action is believed to result from interference
with productive binding of the RNA substrate to the
enzyme.6 This highlights the potential of a novel
mechanism for inhibition of the polymerase, distinct
fromactive site-directed inhibitors such as nucleoside
* Corresponding author. Fax: +1-450-682-8434; e-mail: pbeaulieu@
lav.boehringer-ingelheim.com
y
Undergraduate chemistry COOP student (Universite de Sherbrooke,
Quebec, Canada).
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.12.032

968
P. L. Beaulieu et al. / Bioorg. Med. Chem. Lett. 14 (2004) 967–971
analogues. The existence of allosteric binding sites to
which small molecules can bind in a productive fashion
represents an attractive avenue for the discovery of
specific and novel antiviral agents.4 Unfortunately,
compound 1 did not inhibit replication of HCV repli-
cons in a cell-based assay.7 Factors responsible for this
lack of efficacy had to be addressed in order to advance
further towards the development of HCV therapeutics.
The modest inhibition by these compounds in our in
vitro enzymatic assay was likely an important liability
and our objectives focused on improving potency.
addition of a hydroxyl group to the aromatic ring
(tyrosine analogue 7) provided a 4- to 5-fold increase in
potency over a hydrogen atom( 6 versus 7). Extending
the aromatic surface of the amino acid from a simple
phenyl ring (6) to an indole gave tryptophan analogue 8
which was also 3- to 4-fold more potent. The 5-
hydroxytryptophan analogue 9 was then prepared with
the expectation that the beneficial effect of the hydroxyl
group (seen with 6 versus 7) was reproducible in this
derivative. Indeed the two effects, expanding the size of
the aromatic surface and addition of a hydroxyl group
were additive. Thus, the first inhibitor in this class to
inhibit the polymerase activity at sub-micromolar
concentrations was identified. Compound 9 had an
In our initial studies leading to the discovery of
compound 1,5 substitution of the benzimidazole scaffold
at N1 and C-2 by a cyclohexyl ring and a 3-furyl moiety
respectively, were optimal. SAR had revealed the
former to be rather intolerant to manipulation whereas
Table 1. Benzimidazole 5-carboxamide SAR
small aromatic heterocycles with
a
variety of
substitutions were favored at the C-2 position. Since the
minimum core for inhibitory activity, as exemplified by
compound 1, offered no obvious possibilities for further
improvements, the amide derivative 2 (Table 1), which
was also identified as an inhibitor (IC50=7 mM) during
preliminary hit-to-lead activities,5 was selected as the
starting point for further optimization. Taking
advantage, once again, of parallel synthesis techniques
established in our previous work,5 we describe modific-
ations to the right-hand side of the molecule leading to
800-fold increases in potency.8
HNR
HCV NS5B IC50 (mM)a
HET=2-pyridyl HET=3-furyl
2
3
7Æ1.5
12Æ5
Addition of a methyl group at the benzylic position of 2
generated a pair of enantiomers (compounds 3 and 4,
Table 1). The (R)-enantiomer was approximately 4
times more potent than the (S)-isomer and about 2-fold
more potent than unsubstituted benzylic amide 2. The
increased potency could be due to a productive interac-
tion of the methyl group with the protein, but is more
likely due to an orienting effect that helped rigidify the
4
5
6
7
8
9
3Æ1
2Æ0.8
10
11
12
13
14
15
16
17
18
0.4Æ0.2
4Æ1
7Æ3
free-state of the inhibitor towards
a bound-like
conformation.9 Support for the latter proposal derives
from replacement of the neutral methyl group with an
ionized carboxyl group of identical configuration that
resulted in an analogue of similar potency (5), despite
large differences in electronic character between the two
groups. One advantage of incorporating the a-carboxyl
group was that it also improved aqueous solubility of
this compound class.10
1.5Æ0.7
1.9Æ0.5
0.14Æ0.05
0.4Æ0.1
0.4Æ0.1
0.05Æ0.01
0.7Æ0.08
0.16Æ0.02
0.46Æ0.18
1.5Æ0.3
Compound 5 being the coupling product between the
5-benzimidazole carboxylic acid derivative and
a
phenylglycine analogue, a set of commercially available
a-amino acids with diverse side chains were coupled to
N-cyclohexyl-2-(2-pyridyl) benzimidazole 5-carboxylic
acid to further expand the SAR in this new direction.
Most compounds inhibited the polymerase with IC50
values in the 2-35 mM range (examples included: Ala,
Leu, Gly, Ile, Val, Asp, Glu, Ser, Thr, His, phenylglycine,
tert-butylglycine; results not shown).
The most improved results in this survey were obtained
with amino acid derivatives bearing aromatic side
chains. For example, whereas phenylalanine amide 6
was 3-fold less potent than phenylglycine analogue 5,
a Values are the mean of duplicate experiments with two separate
weighings.

P. L. Beaulieu et al. / Bioorg. Med. Chem. Lett. 14 (2004) 967–971
969
IC50=0.14 mM, and represented a 50-fold increase in
potency fromthe starting benzylic amide 2.
Optimization, so far, was carried out on a benzimi-
dazole scaffold carrying a 2-pyridyl substituent at C-2.
In our previous work,5 we found that a 3-furyl group at
C-2 provided a 2- to 3-fold improvement in potency. A
similar effect was anticipated in the current benzi-
midazole amide series. The outcome of this modification
is also shown in Table 1. In all cases, replacement of the
2-pyridyl group by a 3-furyl group increased potency an
additional 2- to 5-fold (compare 5–9 and 10–14). The
most potent analogue in this series (14) had an IC50=50
nM. The parallel SAR displayed by the two C-2
heterocycles is an indication that despite large structural
differences between the original benzimidazole 5-
carboxylic acid series5 and the current amide series, the
two classes of inhibitors have similar modes of action
and bind similarly to the polymerase. Interestingly,
potent inhibitors appear to be more sensitive to config-
urational changes at the asymmetric carbon center.
Whereas a 4-fold difference in potency was previously
observed between enantiomers 3 and 4, the gap increased
to 14-fold in the case of the more potent 5-hydroxy-
tryptophan derivatives 14 and 15. This result suggests that
complementarities between the ligands and the protein
binding site were improved, and in turn, increased
compound sensitivity to conformational effects. Methyla-
tion of the indole nitrogen or the C-2 indolic carbon of
the tryptophan residue resulted in a 3- and 9-fold decrease
in potency, respectively. In the latter case, methylation
affects the free-state conformation. The importance of
conformational contributions to compound potency
was evident fromthe 30-fold loss in inhibitory activity
(compound 18) when the a-carboxyl group of 14 was
removed. Although, at this stage, we could not exclude
minor binding contributions from the carboxyl group.
Scheme 1. General synthesis of inhibitors.
potency resulting fromaddition of hydrogen-bond
donors to the 5-position of the tryptophan ring. How-
ever, such conclusions await further confirmation from
structural studies (e.g., X-ray crystallography).
Potent analogues in Table 2 (14 and 23–32) were tested
for specificity against another RNA-dependent RNA
polymerase from polio virus and a DNA-dependent
RNA polymerase isolated from calf-thymus.5,8
Specificity windows of >1500-fold were seen for all
compounds.
Several inhibitors generated during this optimization
study were also tested in an HCV cell-based replicon
assay of RNA replication.7 Compound 31, despite good
intrinsic potency, had only marginal efficacy (i.e., 63%
inhibition at 50 mM) which may reflect its inability to
permeate cells effectively and reach the intracellular
enzymatic target. The presence of a highly ionized car-
boxyl group at physiological pH is presumably a major
cause for this lack of efficacy. Further support for this
hypothesis is provided by the measured log D values for
compounds 14 and 31, which were found to be 0.45 and
Table 2. SAR at the 5-position of the tryptophan indole nucleus
Following the discovery of benzimidazole–tryptophan
conjugates with inhibitory potency in the 50 nM range,
further improvements were examined through SAR at
the easily accessible 5-position of the tryptophan indole
nucleus of compounds such as 13. As discussed above,
the introduction of a phenolic hydroxyl group at the
5-position of the tryptophan indole nucleus increased
potency 8-fold (compound 14). Groups not capable of
providing hydrogen bonds such as NO2, Me, F and
OMe were all significantly less potent (19–22). The ben-
eficial effect of an OH group added to the indole is not
likely due to changes in electronic properties of the ring
system, as the OMe derivative (22) was equipotent to
13. All compounds with functional groups capable of
donating hydrogen bonds, on the other hand, had
IC50s<150 nM, independent of their electronic proper-
ties or ionization state: basic groups (aniline 23), neutral
sulfonamides (24 and 25), anilides (26), carboxamides
(29) and acidic functions (27, 28, and 30–32) all behaved
similarly. Oxalic amide (27) was the m ost potent analogue
identified in this series with an IC50 <10 nM.
Compds
X
HCV NS5B Poliovirus and CT RNA
IC50 (mM)a
polymerases IC50 (mM)
13
14
19
20
21
22
23
24
25
26
27
28
29
30
31
32
H
OH
NO2
Me (dl)
F (dl)
OMe (dl)
NH2
NHSO2CH3
NHSO2CF3
NHAc
NHCOCOOH
COOH
CONH2
C5-tetrazole
OCH2COOH
OC(Me)2COOH
0.4Æ0.1
0.05Æ0.01
2.0Æ1
200, 316
79, >200
2.2Æ0.3
2.9Æ1.9
0.4Æ0.1
0.09Æ0.04
0.12Æ0.04
0.15Æ0.05
0.06Æ0.01
0.008Æ0.002
0.02Æ0.008
0.037Æ0.01
0.015Æ0.005
0.019Æ0.005
0.10Æ0.02
>500, >250
>500, –
>250, >250
>125, >250
>250, >250
>200, >250
>200, >250
>200, >250
100, >200
>200, —
It is tempting to invoke the identification of a new
pharmacophore on the right-hand side of the molecules
that may be responsible for the observed increases in
a Values are the mean from at least duplicate experiments on two
separate weighings.

970
P. L. Beaulieu et al. / Bioorg. Med. Chem. Lett. 14 (2004) 967–971
Scheme 2. Synthesis of functionalized tryptophan derivatives: (a) MeOH/SOCl2.. (b) Boc2O. (c) BrCH2COOMe or BrC(CH3)2COOMe/Cs2CO3 (2
equiv)/acetone/50 ꢀC. (d) HCl-dioxane. (e) 5-benzimidazole carboxylic acid derivative (e.g., 1)/TBTU/iPr2EtN/DMSO or DMF. (f) NaOH then
AcOH or TFA. (g) SnCl2 (5 equiv)/3:1 DMF–water/60 ꢀC. (h) Acid chloride, sulfonyl chloride or anhydride/pyridine/DMSO or DMF.
À1.62, respectively.11 Future efforts in this series,
focused on addressing the issue of cell permeability and
activity in the replicon, will be reported in due course.
5-carboxyl group and the corresponding primary amide
in a 2:1 ratio respectively. Following esterification of
carboxyl groups (MeOH/SOCl2), the mixture of trypto-
phan derivatives was coupled to the left-hand side scaf-
fold and following deprotection, 28 and 29 were
separated by preparative HPLC. 5-Tetrazolyl analogue
30 was prepared by condensation of 5-cyano-Trp-
OMe16 with the left-hand side benzimidazole scaffold
and the 5-cyano group converted to a tetrazole ring
using tri-n-butyltin azide in DMSO at 80 ꢀC.17
Saponification of the protected inhibitor in the usual
manner gave 30.
Inhibitors were prepared in solution-phase by
condensation of 1,2-disusbtituted benzimidazole 5-car-
boxylic acids5,12 with amine derivatives, using standard
amide bond forming reagents (Scheme 1).13 Inhibitors
were purified by reversed-phase HPLC to >90%
homogeneity (isolated as TFA salts in most cases).
Compounds gave spectral data (electrospray MS and
1H NMR) consistent with their assigned structures.
Amines were obtained from commercial sources or
prepared using standard protocols as described in the
literature. Amino acid derivatives were coupled as their
methyl esters (prepared using MeOH/SOCl2) and
required subsequent hydrolysis of ester functionalities
using hydroxide.14 The synthesis of some ring-function-
alized tryptophan derivatives is illustrated in Scheme 2.
Commercially available (S)-5-hydroxytryptophan 33
was esterified to the methyl ester and N-protected to
give carbamate 34. Alkylation of the phenolic hydroxyl
group with bromoacetate derivatives and cleavage of
the Boc protecting group under acidic conditions gave
ethers 35 (R¼H or Me), precursors to inhibitors 31 and
32. For the synthesis of inhibitors 24–27, (S)-5-nitro-
tryptophan 3615 was esterified to the methyl ester
hydrochloride and coupled in the usual manner to
benzimidazole 1. Reduction of the nitro group then
gave 5-aminotryptophan derivative 37. Compound 37
was reacted with acid chlorides, sulfonyl chlorides or
anhydrides to provide protected derivatives 38, which
upon saponification of the methyl esters provided the
desired inhibitors. 5-Carboxy and 5-carboxyamide
analogues 28 and 29 were prepared from5-cyano-
tryptophan16 by hydrolysis under acid conditions
(concd HCl/80 ꢀC), which yielded a mixture of the free
This study describes the optimization of a previously
discovered series of substituted 5-carboxybenzimidazole
inhibitors of HCV NS5B polymerase. An 800-fold
increase in intrinsic potency relative to the initial hit
compound was achieved, leading to the discovery of
inhibitors with low nanomolar potency in our in vitro
enzymatic assay. This result was achieved through the
combined effects of conformational rigidification and
the discovery of a potential new pharmacophore on the
right-hand side of the molecule, favoring hydrogen-
bond donors. Potent compounds in this series retain
specificity for inhibition of HCV polymerase. Additional
studies are required to address the issue of cellular per-
meation with this class of inhibitors and improve their
efficacy in a cell-based HCV-replicon assay.
Acknowledgements
The authors wish to thank Lukxmi Balathasan, Sylvain
Lefebvre, Leslie Magtanong, Martin Marquis, Charles
Pellerin and Louise Thauvette for their support provid-
ing protein and running biological assays. The
contributions of Samuel April and Charles Banville who
were involved in the synthesis of inhibitors is also

P. L. Beaulieu et al. / Bioorg. Med. Chem. Lett. 14 (2004) 967–971
971
acknowledged. We are grateful to Colette Boucher,
Sylvain Bordeleau and Michael Little for analytical
support.
7. Lohmann, V.; Korner, F.; Koch, J. O.; Herian, U.;
Theilmann, L.; Bartenschlager, R. Science 1999, 285, 110.
8. The in vitro enzymatic assay used to evaluate the poten-
tial of compounds as inhibitors of HCV polymerase is
described in reference 6 and US patents 6,448,281 B1 and
6,479,508 B1 (2002) to Boehringer Ingelheim(Canada)
Ltd.
References and notes
9. NMR studies on the bioactive conformation of this class of
inhibitors when bound to NS5B will be reported separately:
LaPlante, S. et al. (manuscript in preparation).
10. The solubility of compounds 4 and 10 was measured in
aqueous pH 7. 2 phosphate buffer and were found to be 5
mg/mL and >991 mg/mL, respectively. The low solubility
of neutral molecules such as 2–4 and others (not shown)
was considered a liability at this stage, in terms of estab-
lishing solid SAR patterns. For this reason, optimization
studies were pursued in the more soluble carboxylic acid
class.
11. The methyl ester of compound 14 was also prepared and
tested in the replicon assay. Despite its good inhibitory
potency against the enzyme (IC50=0.11 mM) and a sub-
stantial increase in lipophilicity resulting fromthe
removal of ionizable functions (log D=3.07, solubi-
lity=0. 6 mg/mL in pH 7.2 phosphate buffer), only 30%
inhibition was measured at the highest non-cytotoxic
concentration (3.4 mM). No hydrolysis of the methyl ester
was detected under assay conditions.
12. Beaulieu, P. L.; Hache, B.; von Moos, E. Synthesis 2003,
1683.
13. TBTU was used as coupling reagent: Knorr, R.; Trzeciak,
A.; Bannwarth, W.; Gillessen, D. Tetrahedron Lett. 1989,
30, 1927.
14. Saponification of the a-ester group was found to proceed
with minimal racemization as determined by HPLC
analysis on a chiral support.
15. Irie, K.; Ishida, A.; Nakamura, T.; Oh-Ishi, T. Chem.
Pharm. Bull. 1984, 32, 2126.
1. (a) Di Bisceglie, A. M. Lancet 1998, 351, 351. (b) Choo,
Q.-L.; Kuo, G.; Weiner, A. J.; Overby, L. R.; Bradley,
D. W.; Houghton, M. Science 1989, 244, 359. (c) World
Health Organization; Hepatitis C. Seroprevalance of
hepatitis C virus (HCV) in a population sample, Wkly.
Epidemiol. Rec. 1996, 71, 346.
2. Wasley, A.; Alter, M. J. Semin. Liver Dis. 2000, 20, 1.
3. (a) Di Bisceglie, A. M.; Hoofnagle, J. H. Hepatology 2002,
36, S121. (b) Chandler, G. Hepatology 2002, 36, S135.
4. (a) Beaulieu, P. L.; Llinas-Brunet, M. Curr. Med. Chem.-
Anti-Infective Agents 2002, 1, 163. (b) For recent dis-
closures of other non-nucleoside HCV NS5B polymerase
inhibitors, see: Dhanak, D.; Duffy, K. J.; Johnston, V. K.;
Lin-Goerke, J.; Darcy, M.; Shaw, A. N.; Gu, B.;
Silverman, C.; Gates, A. T.; Nonnemacher, M. R.;
Earnshaw, D. L.; Casper, D. J.; Kaura, A.; Baker, A.;
Greenwood, C.; Gutshall, L. L.; Maley, D.; DelVecchio,
A.; Macarron, R.; Hofmann, G. A.; Alnoah, Z.; Cheng,
H.-Y.; Chan, G.; Khandekar, S.; Keenan, R. M.; Sarisky,
R. T. J. Biol. Chem. 2002, 277, 38322. (c) Chan, L.;
Reddy, T. J.; Proulx, M.; Das, S. K.; Pereira, O.; Wang,
W.; Siddiqui, A.; Yannopoulos, C. G.; Poisson, C.;
Turcotte, N.; Drouin, A.; Alaoui-Ismaili, M. H.; Bethell,
R.; Hamel, M.; L’Heureux, L.; Bilimoria, D.; Nguyen-Ba,
N. J. Med. Chem. 2003, 46, 1283.
5. Beaulieu, P. L.; Bos, M.; Bousquet, Y.; Fazal, G.;
Gauthier, J.; Gillard, J.; Goulet, S.; LaPlante, S.; Poupart,
M.-A.; Lefebvre, S.; McKercher, G.; Pellerin, C.; Austel,
V.; Kukolj, G. Bioorg. Med. Chem. Lett. 2004, 14, 119.
6. McKercher, G.; Beaulieu, P. L.; Lamarre, D.; LaPlante,
S.; Lefebvre, S.; Pellerin, C.; Thauvette, L.; Kukolj, G.
Nucleic Acid Res. 2004, 32, in press.
16. Dua, R. K.; Phillips, R. S. Tetrahedron Lett. 1992, 33, 29.
17. Sisido, K.; Nabiko, K.; Isida, T. Organomet. Chem. 1971,
33, 337.