Novel thiazolones as HCV NS5B polymerase allosteric inhibitors: Further designs, SAR, and X-ray complex structure

Article abstract of DOI:10.1016/j.bmcl.2006.09.095

Structure-activity relationships (SAR) of 1 against HCV NS5B polymerase were described. SAR explorations and further structure-based design led to the identifications of 2 and 3 as novel HCV NS5B inhibitors. X-ray structure of 3 in complex with NS5B polymerase was obtained at a resolution of 2.2 A, and confirmed the design.

Full text of DOI:10.1016/j.bmcl.2006.09.095

Bioorganic & Medicinal Chemistry Letters 17 (2007) 63–67
Novel thiazolones as HCV NS5B polymerase allosteric inhibitors:
Further designs, SAR, and X-ray complex structure
Shunqi Yan,* Gary Larson, Jim Z. Wu, Todd Appleby, Yili Ding, Robert Hamatake,
Zhi Hong and Nanhua Yao*
Valeant Pharmaceuticals Research and Development, 3300 Hyland Ave., Costa Mesa, CA 92626, USA
Received 16 August 2006; revised 29 September 2006; accepted 29 September 2006
Available online 4 October 2006
Abstract—Structure–activity relationships (SAR) of 1 against HCV NS5B polymerase were described. SAR explorations and further
structure-based design led to the identifications of 2 and 3 as novel HCV NS5B inhibitors. X-ray structure of 3 in complex with
˚
NS5B polymerase was obtained at a resolution of 2.2 A, and confirmed the design.
Ó 2006 Elsevier Ltd. All rights reserved.
Hepatitis C virus (HCV) was identified in 1989 as the
leading pathogen for non-A, non-B viral hepatitis.1
More than 170 million individuals worldwide including
4 million people in the United States are believed to
have been infected.2 The current FDA approved stan-
dard therapy is effective for about half of the patient
population.3 Therefore, newer and more efficacious
therapies directly targeting HCV are still an unmet med-
ical need.
Previously, we disclosed a successful structure-based de-
sign of a novel thiazolone scaffold represented by com-
pound 1 as an example of this type of HCV NS5B
inhibitor. It has an IC50 of 3.0 lM and was established
by X-ray crystallography as an allosteric inhibitor that
binds to the thumb subdomain.7 Further structure-based
design and SAR exploration of 1 led to the identification
of new scaffolds exemplified by compounds 2 and 3 as
novel inhibitors of HCV NS5B polymerase. In this pa-
per, we report a more detailed SAR analysis of 1, as well
as further designs, and synthesis of 2 and 3. The X-ray
complex structure of 3 with HCV NS5B is also described.
HCV NS5B is a virus-encoded RNA-directed RNA
polymerase (RdRp). It is responsible for HCV viral rep-
lication,4 and thus represents a valid target for antiviral
therapy against HCV.5 During polymerization, NS5B
sub-domains undergo conformational changes as it
binds to its substrates and performs the subsequent cat-
alytic reactions. Recently, there has been growing inter-
est in discovery of scaffolds that target an allosteric
binding site of NS5B located on the thumb subdomains
distant from the polymerase active site.6 Inhibitors
bound to such allosteric sites prevent elongation by
locking in the conformation of NS5B. Allosteric inhibi-
tors may have fewer off-target side-effects than nucleo-
side analogs due to the absence of binding to
homologous cellular enzymes.
The synthesis of thiazolone derivatives discussed in this
paper was carried out according to Schemes 1 and 2. For
compounds 1, 15–38, the synthesis started with commer-
cially available rhodanine 4 and aldehyde 5 (Scheme 1).7
Thus, 4 was condensed with 5 to form 6, which was sub-
sequently coupled with MeI to give 7. Treatment of 7
with a suitable amino-containing reagent (8) in the pres-
ence of DIEA afforded the desired products in a good
yield.
The tetrazole derivatives, 39–42, were prepared from the
known amino acids (9) over seven steps by way of isolated
intermediates 10–13 (Scheme 2). Thus, the amino group
of an amino-acid (9) was first protected by a Boc group
to give 10, which was treated with ethyl chloroformate
in the presence of TEA, followed by the introduction of
a stream of NH3 (gas) to afford amide 11.8 Treatment of
11 with TFA in the presence of TEA gave nitrile 12 in a
quantitative yield. The resulting nitrile 12 was allowed
Keywords: HCV; NS5B; NS5B inhibitor; HCV NS5B polymerase;
HCV NS5B inhibitors; Thiazolone; Tetrazole.
*
Corresponding authors. Tel.: +1 714 545 0100; fax: +1 714 668
3142; e-mail addresses: syan@valeant.com; nyao@valeant.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.09.095

64
S. Yan et al. / Bioorg. Med. Chem. Lett. 17 (2007) 63–67
O
O
O
N
N
O
N
O
O
O
N
N
S
S
S
O
NH
HN
HN
HN
N
OH
OH
F
F
F
1
2
3
O
O
O
a
b
NH
O
NH
S
O
N
O
CHO
S
S
S
S
S
7
4
6
5
O
n=0,1
N
O
c
n=0,1
H2N
R2
S
R2
n
R2: -OH
R2: -NHOH
R2: -NH2
HN
R1
O
d
e
n
R1
O
1, 15 - 38
8
Scheme 1. General procedures for the synthesis 1, 15–38. Reagents and conditions: (a) EtONa, HOAc, 95 °C, 12 h, 93%.; (b) MeI, EtOH, DIEA, rt,
10–16 h, 95%; (c) i—EtOH, DIEA, 80 °C, 21 h; ii—NaOH (aq, 1 N), 100 °C, 30 min; iii—HCl (aq, 1 N), 93%; (d) HOBt, PS-Carbodiimide, DMF,
NH2OHÆHCl, DMF, rt, 16 h, 70%; (e) i—11, HCl (4 N in dioxane), rt overnight, 90%; ii—(c), 7 + de-Boc-11, 63%.
n = 0, 1
H
H
H2N
OH
a
N
OH
b,c
N
NH2
d
n
O
Boc
n
O
Boc
n
O
R
R
R
9
10
11
O
O
N
H
H
S
N
e
N
N
f,g
Boc
n
Boc
n
HN
N
N
N
R
R
N
NH
N
R
N
N
H
12
13
39 - 41
Scheme 2. General procedure for ester synthesis of 39–41. Reagents and conditions: (a) Boc2O, NaOH (1 N, aq), dioxane, rt, 12 h, 67%; (b) TEA,
ClCO2Et, THF, À10 °C, 30 min; (c) NH3 (gas), THF, <0 °C to >12 h, 90%; (d) (CF3CO)2O, TEA, THF, rt, 14 h, 100%; (e) NaN3, Et3NÆHCl,
toluene, 95 °C, 16 h, 83%; (f) HCl (3 N), rt, 16 h, 85%; (g) 7, DIEA, EtOH, 75 °C, 12 h, 96%.
to react with NaN3 and TEAÆHCl in toluene to form
tetrazole 13 in a good yield.9 Subsequently, Boc group
of 13 was removed and the resulted intermediate was
coupled with 7 to afford the desired compound 39–41.
(Table 1). Since replacement of the 5-ethyl furan moiety
with other groups did not result in significant variation
of IC50 values (data not shown), we have focused on
the substitution patterns on R1 and R2. The (S)-config-
uration (1) is approximately 17-fold more potent than
the (R)-isomer (2) and about 18-fold more potent than
the unsubstituted compound (22). The preference
of the (S)-isomer 1 could be due to the overall favorable
protein–ligand interactions from the 4-F-phenyl and
carboxylic group of 1.7 The X-ray complex structure
of 1 described in the previous paper confirmed that the
4-F-phenyl is buried into a hydrophobic pocket defined
by Leu 419, Met 423, Tyr 477, and Trp 528, and that the
The synthesized compounds were then evaluated for
their ability to inhibit HCV NS5B polymerase by an
in vitro NS5B RdRp-catalyzed elongation assay follow-
ing the protocol previously described.7,10 The resulting
IC50 values are shown in Tables 1–3.
The most potent compounds of these a-amino acid
derivatives 1, 15–29 are in the low micro-molar range

S. Yan et al. / Bioorg. Med. Chem. Lett. 17 (2007) 63–67
Table 2. IC50 values (lM) of 30–38
65
Table 1. IC50 values (lM) of 1, 15–29
O
O
N
N
O
O
S
O
S
HN
HN
OH
R2
R1
R2 –OH
O
R
Compound
R1
Compound
R
IC50
44.0
IC50
3.0
30
1
F
F
31
32
33
34
35
36
37
38
>100.0
16.5
8.5
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
50.0
6.0
F
N
Br
Br
Cl
Cl
7.6
Cl
Cl
27.0
>100.0
59.0
13.0
9.0
Me
21.0
14.0
11.0
25.0
55.0
>100.0
24.0
68.0
12.0
43.0
5.5
F3C
NC
Me
F
Cl
Cl
Cl
F
nearly parallel in a p–p stack to the indole ring of Trp
528. This interaction presumably renders the (S)-isomer
more favorable than the (R)-configuration. It is also
conceivable that an aryl group, which is positioned in
a p–p stack with Trp 528, would be more favorable than
an alkyl group for the hydrophobic interactions (Fig. 1).
Support for these theories are that compounds with R1
as alkyl groups, for example, 21, 23 and benzyl group
such as 25, have weaker or no inhibitory activities
against HCV NS5B. We thus focused on the substituted
aryl, particularly phenyl, as R1. A substitution of 4-Br
(16), or 2,4-Cl (17) can be well tolerated, as reflected
by their IC50 values of 6.0 and 7.6 lM, respectively,
which are similar to 1. The weaker potency of 27 with
4-OH is possibly caused by the hydrophilic nature of
4-OH, while a combination of electronic and steric fac-
tors might contribute to the weaker activities of 21 and
24. Hydrophobicity may be more critical than steric fac-
tors as 2-naphthalene (19), 4-Me– (20), and 2,4-Me– (26)
are roughly 6-fold more potent than the less hydropho-
bic unsubstituted 22 but only slightly less potent than
the less bulky 1 and 17.
Me
Me
HO
16.0
carboxylic group hydrogen-bonds in an ionic fashion
with the basic side-chains of both Arg 501 and Lys
533 (Fig. 1).7 Noticeably, the 4-F-phenyl moiety of 1 is

66
S. Yan et al. / Bioorg. Med. Chem. Lett. 17 (2007) 63–67
Table 3. IC50 values (lM) of tetrazoles 39–42
O
N
O
S
N
HN
n
NH
N
N
R
Compound
R
IC50
19.0
n = 0
39
n = 0
40
41
42
14.0
9.7
Me
F
n = 0
Figure 2. Overlay of X-ray structures of 1 (colored in yellow stick) and
GOLD docking pose of tetrazole 41 (colored by atomtypes).
n = 1
24.0
lene group. As for the effect of different R on IC50s,
the activities of 3-pyridyl (31) or 4-cyano (35) are prob-
ably eliminated because of their hydrophilic nature.
When R is a substituted phenyl group, compounds bear-
ing 4-Br (32), 4-Me (34), 4-F (30) or 4-H (36) have IC50
values of 16.5, 27.0, 44.0, and 59.0 lM, respectively.
The rank order of these IC50 values agrees well with
the opposite order of the van der Waals radius
(Br > Me > F > H), suggesting that bulkier groups in
that position result in more favorable interactions.
As demonstrated in Tables 1 and 2, a certain degree of
flexibility, in addition to the presence of a –COOH
head-group and a suitable R1 or R group, is allowed
in order to maintain potency against HCV NS5B.
Encouraged by this insight, we turned our attention to
the replacement of the –COOH group. The binding
mode of 1 in the X-ray complex structure shows that
the pocket occupied by the carboxylic group –COOH
could accept a hetero-ring moiety, preferably hydrophil-
ic in nature, to accommodate the hydrophilic receptor
surroundings defined by His 475, Arg 501, and Lys
533 (Fig. 1).7 Intuitively, a rational design for the first
round would replace the –COOH group by a bio-isoster-
ic tetrazole moiety and GOLD docking predicted a tet-
razole replacement could indeed fit well into the pocket
(Fig. 2).11 To test this design, four tetrazole derivatives
39–42 were synthesized and their IC50 values are shown
in Table 3.
Figure 1. The binding mode of 1 in a complex with NS5B.
Table 1 also shows that the acidic moiety appears to be
critical for activities. For example, amide replacement
(29) of –COOH (1) leads to a 5-fold loss of potency com-
pared to 1. On the other hand, compound (28) bearing
–CONHOH (5.6 lM), which has similar pKa to –COOH,
is at least as equipotent as 17 (7.6 lM) (Table 1).
As expected, all four tetrazoles are moderately active.
The most active compound (41) produced an IC50 value
of 9.7 lM, which is approximately 3-fold less potent
than the corresponding carboxylic version (1), thus indi-
cating that the carboxylic acid is more suitable for
potency, whereas the increased bulkiness of a tetrazole
group may decrease activity. 4-Me (40) and 4-H (39)
are about 2-fold weaker than 4-F (41). Interestingly,
extending the tetrazole group by one –CH2– (42) does
not seem to result in any significant loss of activity in
comparison to 39.
Compounds with more flexibility, achieved by increas-
ing the distance of carboxylic acid from the R group
with one methylene (–CH2–) group, were synthesized
and examined. The IC50 values against HCV NS5B for
the corresponding compounds 30–38 are shown in Table
2. The most potent compound is 33 with an IC50 of
8.5 lM, which is as potent as its other enantiomer 38
and similar in activity to 17 (7.6 lM) (Table 1). The
comparable potency indicates that the binding site seems
to be able to accommodate the flexibility of the methy-

S. Yan et al. / Bioorg. Med. Chem. Lett. 17 (2007) 63–67
67
activities, while an (S)-isomer is preferred for potency.
Replacing the –COOH head group by an –NHOH
group with a similar pKa preserves the activity. Based
on the binding mode from a complex of 1, further de-
signs of 2 and 3 were proposed and corresponding com-
pounds were synthesized. Compounds from both series
retain comparable potency with 1. The X-ray complex
structure of a tetrazole 41 with HCV NS5B polymerase
˚
was solved at a resolution of 2.2 A and confirmed its
binding mode.
Acknowledgments
We thank Dr. I. Wayne Cheney for critically reading the
manuscript. We also thank Drs. Hassan El Abdellaoui,
Vicky Lai, and Weidong Zhong for insightful discussion.
Figure 3. Electron density map of X-ray structure of 41 in complex
˚
with HCV NS5B polymerase at a resolution of 2.2 A.
References and notes
To confirm the design rationale and further comprehend
protein-ligand interactions for future structure-based
designs, single crystals of HCV NS5B polymerase in
complex with 41 were successfully prepared by soaking
crystal. The X-ray crystal structure was established at
1. Choo, Q. L.; Kuo, G.; Weiner, A. J.; Overby, L. R.;
Bradley, D. W.; Houghton, M. Science 1989, 244, 359.
2. World Health Organization; Hepatitis C-Global preva-
lence (update). Weekly Epidemiol. Rec. 2000, 75, 18.
3. Manns, M. P.; McHutchison, J. G.; Gordon, S. C.; Rustgi,
V. K.; Shiffman, M. Lancet 2001, 358, 958.
˚
a resolution of 2.2 A The electron density was unambig-
4. Kolykhalov, A. A.; Mihalik, K.; Feinstone, S. M.; Rice, C.
M. J. Virol. 2000, 74, 2046.
5. (a) Wu, J. Z.; Zhi, H. Curr. Drug Targets Infect. Disord.
2003, 3, 207; (b) Clercq, E. D. Nat. Rev. Drug Disc. 2002,
1, 13.
uous for inhibitor 41 and its surrounding amino acid
residues. The structure of inhibitor 41 fits satisfactorily
into the electron density map (Fig. 3). Inhibitor 41 binds
to the ‘thumb’ sub-domain, which is distant from the en-
zyme’s catalytic center.12
6. (a) Marco, S. D.; Volpari, C.; Tomei, L.; Altamura, S.;
Harper, S.; Narjes, F.; Koch, U.; Rowley, M.; Francesco,
R. D.; Migliaccio, G.; Carfi, A. J. Biol. Chem. 2005, 280,
29765; (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) Love, R. A.; Parge, H. E.; Yu, X.; Hickey, M.
J.; Diehl, W.; Gao, J.; Wriggers, H.; Ekker, A.; Wang, L.;
Thomson, J. A.; Dragovich, P. S.; Fuhrman, S. A. J. Virol.
2003, 77, 7575; (d) Wang, M.; Ng, K. K.-S.; Cherney, M.
M.; Chan, L.; Yannopoulos, C. G.; Bedard, J.; Morin, N.;
Nguyen-Ba, N.; Alaoui-Ismaili, M. H.; Bethell, R. C.;
James, M. N. G. J. Biol. Chem. 2003, 278, 9489.
7. Yan, S.; Appleby, T.; Larson, G.; Hamatake, R.; Wu, Z.
J.; Hong, Z.; Yao, N. Bioorg. Med. Chem. Lett. 2006, 16,
5888.
The overall structure of the NS5B–41 complex is basi-
cally identical to the previously described complex struc-
ture of NS5B–1.7 The binding mode of tetrazole
derivative 41 from X-ray coincides with the predicted
one and is verymuch comparable with amino-acid deriv-
ative 1 (Fig. 2). Similarly to 1, the 4-F-phenyl of 41 is
buried within a small deep hydrophobic pocket defined
by the side chains of Leu 419, Met 423, Tyr 477, and
Trp 528. The ethyl-furan moiety of the inhibitors is
bound to the surface of another hydrophobic channel
defined by Leu 419, Met 423, Ile 482, Val 485, Leu
489, and Leu 497. The C@O moiety in the thiazolone
ring of 41 accepts a hydrogen-bond from the backbone
–NH of Tyr 477, while its lone-pair N makes another
hydrogen-bond to the backbone –NH of Ser 476. The
tetrazole group of 41 retains the same hydrogen bond
network as –COOH of 1. Specifically, one N atom on
the tetrazole makes a hydrogen bond with Arg 501,
while another picks up one more hydrogen bond with
Lys 533 (Figs. 2 and 3).
8. Donkor, I.; Han, J.; Zheng, X. J. Med. Chem. 2004, 47, 72.
9. Koguro, K.; Oga, T.; Mitsui, S.; Orita, R. Synthesis 1998,
910.
10. All compounds listed in this paper were fully characterized
by H NMR and MS with more than 95% of purity from
HPLC analysis.
1
11. GOLD v2.1 docking, www.ccdc.cam.ac.uk. The docking
was performed using the standard parameters. Structure 1
was used as shape template similarity constraint with a
constraint weight of 10.
12. Coordinates for the structure have been deposited at
Protein Data Bank (www.rcsb.org) under file name: 2I1R.
In conclusion, a SAR study has been performed for 1 as
a novel HCV NS5B polymerase inhibitor. Substitution
on para/ortho positions of R1 with halogens increases