Thiazolone-acylsulfonamides as novel HCV NS5B polymerase allosteric inhibitors: Convergence of structure-based drug design and X-ray crystallographic study

Article abstract of DOI:10.1016/j.bmcl.2007.01.024

A novel series of thiazolone-acylsulfonamides were designed as HCV NS5B polymerase allosteric inhibitors. The structure based drug designs (SBDD) were guided by docking results that revealed the potential to explore an additional pocket in the allosteric site. In particular, the designed molecules contain moieties of previously described thiazolone and a newly designed acylsulfonamide linker that is in turn connected with a substituted aromatic ring. The selected compounds were synthesized and demonstrated low μM activity. The X-ray complex structure was determined at a 2.2 A resolution and converged with the SBDD principle.

Full text of DOI:10.1016/j.bmcl.2007.01.024

Bioorganic & Medicinal Chemistry Letters 17 (2007) 1991–1995
Thiazolone-acylsulfonamides as novel HCV NS5B polymerase
allosteric inhibitors: Convergence of structure-based
drug design and X-ray crystallographic study
Shunqi Yan,*, Todd Appleby, Gary Larson, Jim Z. Wu, Robert K. Hamatake,
Zhi Hong and Nanhua Yao*
Valeant Pharmaceutical Research & Development, 3300 Hyland Ave., Costa Mesa, CA 92626, USA
Received 6 December 2006; revised 4 January 2007; accepted 8 January 2007
Available online 19 January 2007
Abstract—A novel series of thiazolone-acylsulfonamides were designed as HCV NS5B polymerase allosteric inhibitors. The struc-
ture based drug designs (SBDD) were guided by docking results that revealed the potential to explore an additional pocket in the
allosteric site. In particular, the designed molecules contain moieties of previously described thiazolone and a newly designed acyl-
sulfonamide linker that is in turn connected with a substituted aromatic ring. The selected compounds were synthesized and dem-
˚
onstrated low lM activity. The X-ray complex structure was determined at a 2.2 A resolution and converged with the SBDD
principle.
Ó 2007 Elsevier Ltd. All rights reserved.
Hepatitis C virus (HCV) is a small positive-stranded
RNA virus whose viral genome encodes a single poly-
protein.1 NS5B polymerase—one of the non-structural
proteins released by the proteolytic processing of this
polyprotein—is responsible for the replication and tran-
scription of the viral genome and is critical for viral
infectivity.2,3 Thus, a small molecule that inhibits
NS5B polymerase may prevent HCV from replicating
and consequently stop viral infection. Novel anti-HCV
therapies utilizing a small molecule inhibitor with such
a mechanism of action are not yet available, but would
be ultimately desired for the treatment of the growing
HCV patient population worldwide.
mode of these inhibitors, that is, they bind to the site
on the thumb subdomain distant from the active site.
Inhibitors of different scaffolds interacting with this
same binding pocket have also been reported by oth-
ers.6–8 All of these disclosed scaffolds, though structural-
ly diverse, hydrogen-bond with Ser476, Tyr477, and
Arg501, either directly or via a water molecule, and in
the same region, make similar hydrophobic contacts
with the side-chains of Leu419, Met423, Ile482,
Val485, Leu489, Leu497, and Trp528 (Fig. 1a). Close
examination of the inhibitor–protein interactions
revealed that in the vicinity of these inhibitor-binding
interactions there seems to be an additional pocket that
has yet to be fully explored by the reported inhibitors.9
Notably, this added pocket is flanked by two basic res-
idues, His475 and Lys533, near its entrance (Fig. 1a).
An inhibitor with a suitable moiety to establish interac-
tions specifically with these two residues will represent a
new molecular platform for the exploration of this extra
pocket. More importantly, a corresponding X-ray com-
plex structure of such a novel inhibitor with NS5B will
provide further structural insights for the future design
of novel and specific HCV NS5B inhibitors.
Our structure-based drug design (SBDD) effort focusing
on an allosteric site of NS5B had resulted in the discov-
ery of thiazolone-carboxylic acid (A) and its isosteric tet-
razole derivatives as inhibitors of this enzyme.4,5 X-ray
complex structures established the predicted binding
Keywords: HCV; NS5B; NS5B inhibitor; HCV NS5B polymerase;
HCV NS5B inhibitors; Thiazolone; Acylsulfonamide; Structure-based
drug design; GOLD docking; Computer-modeling; Computational
chemistry.
Therefore as a continued SBDD support for our in-
house anti-HCV program targeting the above-men-
tioned NS5B allosteric site, we envisioned that a design
of such a new scaffold would constitute three iterative
*
Corresponding authors. Tel.: +1 714 729 5560; fax: +1 714 729 5577
(S.Y.); e-mail addresses: yshunqi@yahoo.com; nhyao@yahoo.com
Present address: Ardea Biosciences, Inc., 3300 Hyland Ave., Costa
Mesa, CA 92626, USA.
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.01.024

1992
S. Yan et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1991–1995
Figure 1. (a) Surface representation of X-ray structure of an analog of A; (b) Predicted binding mode of a new design molecule from docking.
R2
O
O
Linker:
N
N
O
O
O
R1
S
R1
SBDD
S
O
O
Linker
HN
HN
OH
O
S
1
R2
R2
B
O
O
O
S
2
A
N
H
New Design
Scheme 1. SBDD of novel acylsulfonamides as NS5B inhibitors.
steps (Scheme 1). First, a fragment template, that is, part
B in Scheme 1, was selected as an anchor to the binding
site for the purpose of establishing the same interactions
with the protein as previously described for A.4 Second,
an acylsulfonamide (1) or its reverse form (2) that has a
comparable pKa with –COOH was identified as a candi-
date linker with B in order to hydrogen-bond with the
basic side-chain of Lys533. The design was completed
by connecting an aromatic moiety to the other side
of the linker in order to pick up a seemingly aromatic
p–p stacking with His475.
one exemplary compound with linker 2, was synthe-
sized. Previous SAR on scaffold A showed that the opti-
mal groups for R11/R12 are F/H or Cl/Cl4,5 and the same
R11/R12 groups were therefore adopted. General proce-
dures for the synthesis of the desired compounds are
shown in Scheme 2 and Scheme 3.
The synthesis of thiazolone-acylsulfonamides 8–13, sym-
bolized by C, was started with commercially available
amino-acids (3) (Scheme 2). Protection of amino acid
3 gave 4, a compound which underwent coupling reac-
tion with sulfonamides upon treatment with EDC to af-
ford acylsulfonamide intermediate 5. De-protection of 5
was carried out under acidic conditions to provide ami-
no acylsulfonamide 6, which was coupled with 7 to af-
ford desired products using previously described
procedures.4
The predicted binding mode by GOLD docking for an
exemplary molecule agreed with the initial design princi-
ple (Fig. 1b).10 Namely, the thiazolone moiety interacts
with NS5B in the same way, as anticipated, as A.4 The
acylsulfonamide moiety hydrogen-bonds extensively
with the basic side-chains of Arg501 and more impor-
tantly Lys533 as it was designed to do. Additionally,
the added phenyl ring connected with the linker on the
model inhibitor appears to engage in an anticipated p
stacking with His475. It is worthy noting that GOLD
docking seems to favor the face-edge stacking
(Fig. 1b), but cursory assessment of Protein Data Bank
(PDB) revealed that the other possibility, the face-face
stacking, is equally plausible.11
Compound 21 was synthesized in nine steps from amino
acid 14 with reasonable overall yield (Scheme 3). Amino
acid 14 was first Boc-protected and its COOH was then
reduced to produce amino alcohol 15.12 Bromination of
15 with CBr4 in the presence of PPh3 readily formed
16.13 Treatment of 16 with KSAc, followed by chlorina-
tion and amination, provided sulfonamide intermediate
19.14–16 Coupling reaction of 19 with PhCOCl under
DIEA/DMAP, followed by de-protection under acidic
condition, gave 20, which was allowed to react with 7
to furnish the final compound 21.4
To test the validity of the design principle, a small
focused set of thiazolone-acylsulfonamides, as well as

S. Yan et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1991–1995
1993
O
Boc
O
O
Boc
O
O
H2N
OH
R12
HN
S
HN
a
NH
R12
OH
R12
c
b
R2
R11
R11
R11
3
4
5
O
O
O
O
O
H2N
R12
S
NH
Ph-R2
d
N
O
O
S
S
O
NH
R12
N
O
H
O
R11
N
S
R2
SMe
R11
6
7
C
Scheme 2. General procedures for the synthesis of compounds 8–13. Reagents and conditions: (a) NaOH(1N, aq), dioxane, 0°C ! rt, 14 h, 73%; (b)
R2–Ph–SO2NH2, EDC, DMAP, DCM, 0 °C ! rt, 48 h, 76%; (c) HCl in dioxane (4N), rt, 14 h, 100%; (d) 7, DIEA, EtOH, Microwave, 10min, 80 °C,
67%.
O
H
H
H
S
H2N
CO2H
a,b
N
N
N
OH
Boc
Boc
S
Boc
Br
c
d
14
15
16
17
Boc
HN
Boc
HN
O
O
H2N
NH
O
S
O
O
S
S
O
Ph
g,h
Cl
f
NH2
O
20
19
18
O
S
i
+ 7
N
HN
O
S
NH
O
S
O
O
21
Scheme 3. General procedures for the synthesis of reverse acylsulfonamide 21. Reagents and conditions: (a) (Boc)2O, NaOH, dioxane, 14 h 90%; (b)
˚
BH3 Å THF, THF, 0 °C ! rt, 3.5 h, 61%; (c) CBr4, Ph3P, DCM, 0 °C, 10 min ! rt, 2.5 h, 26%; (d) potassium thioacetate, DMF, 4 A mol sieves, rt,
18 h, 100%; (e) Cl2(gas), H2O, 0 °C, 40 min, then extraction with DCM; (f) NH3(gas), THF, 0 °C, 30 min ! rt, 16 h, 98%; (g) Ph-COCl, DMAP,
DIEA, DCM, rt, 16 h, 89%; (h) TFA, DCM, 1.5 h, 100%; (i) 7, DIEA, EtOH, 80 °C, 2 h, 70%.
The resultant compounds 8–13 and 21 were evaluated
for the inhibition against BK strain of HCV NS5B poly-
merase using a previously described protocol.17 As
shown in Table 1, these compounds inhibited NS5B
with potencies in the low lM range. It seems, from the
limited SAR available, that para-substituted R2 (8, 9,
and 13) resulted in better potency than the meta-substi-
tuted ones (10, 11, and 12). Interestingly, compound 21,
with a reverse acylsulfonamide linker 2, had an IC50 of
16.0 lM. This inhibitory activity is comparable to those
of the most potent compounds (8, 9, and 13) with acyl-
sulfonamide linker 1, suggesting that both linkers might
be interchangeable, in general, for future designs of nov-
el inhibitors against other disease targets.
Encouraged by the moderate potency and more impor-
tantly, aiming to substantiate the SBDD principle for
this new series of inhibitors, we turned our attention
to understanding the mechanism of action structurally.
As such, an X-ray complex structure of NS5B with com-
˚
pound 13 was obtained at a resolution of 2.2 A from
soaked crystals.18 The resulting structure revealed that
the compound was bound to an allosteric site at the
thumb domain as anticipated (Fig. 2a and b). This

1994
S. Yan et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1991–1995
Table 1. IC50 values against HCV NS5B polymerase
(Fig. 2c and d), whereas the ethyl-furan extends to a
hydrophobic channel defined by residues of Met423,
Ile482, Val485, Leu489, and Leu497. In the same region,
the 4-F-phenyl of 13 is positioned nearly parallel with
the indole-ring of Trp528 in a p–p stacking, as indicated
clearly by the electron-density map, and further contact-
ed by hydrophobic side-chains of Leu419, Met423, and
Tyr477.
a
Compound
R
11/R12
R2
IC50 (lM)
8
9
F/H
Cl/Cl
F/H
F/H
F/H
F/H
4-Me
4-NO2
3-NO2
9.3
6.6
15
10
11
12
13
21
3-COOMe
3-COOH
4-NO2
14
12
7.0
16
a IC50 values are means of three experiments, standard deviation is less
than 10% of the IC50 values.
We were pleased to find that the newly incorporated
moieties were bound in the predicted mode. In particu-
lar, the designed linker (1) makes extensive hydrogen-
bonding with the protein. The deprotonated N and
one of the O atoms on the sulfonamide engage in hydro-
gen-bonding with the basic side-chain of Arg501, and
both O atoms on –SO2NHꢀ form hydrogen-bonds with
˚
binding site is about 30 A away from the active site that
includes a conserved catalytic GDD motif located in the
palm domain. The electron density was evident for
inhibitor 13 and the surrounding amino acids, and the
inhibitor structure fits adequately into the electron den-
sity map (Fig. 2c). In particular, the electron density
map of the side-chain of Lys533 is unambiguous, and
it was the first time we saw such well-defined density
for its long linear side-chain, indicating a favorable sta-
bilizing force in that area between the linker atoms of
the inhibitor 13 and NS5B.
þ
ꢀNH3 of Lys533 (Fig. 2d). Interestingly, the 4-NO2-Ph
involves a possible face–face p stacking with His 475,
which is in contrast with edge-face stacking predicted
initially by GOLD docking (Fig. 1b). Because both
stacking patterns are well represented in the PDB data-
base and in the absence of a higher resolution X-ray
complex structure, a definitive assignment of a stacking
pattern would be elusive. Furthermore, a superimposi-
tion of the bound structure of 13 and the top-scored
docking pose of a model structure afforded a RMSD
The overall binding mode of 13 with NS5B established
by the X-ray crystallographic study concurs with previ-
ously determined binding conformation and that pre-
dicted by GOLD docking. The determined complex
structure showed, as expected, that the thiazolone moi-
ety of 13 binds to the pocket of NS5B similarly as the
compound described before.4 Namely, the C@ O and
lone-pair N on the thiazolone ring hydrogen-bond with
–NHs of backbone Tyr477 and Ser476, respectively
˚
of less than 1.0 A (Fig. 3). Such a small RMSD value re-
veals a remarkable convergence of the initial structure-
based drug design and the subsequently established
complex structure by X-ray cryptography.22
In conclusion, a novel series of thiazolone-acylsulfona-
mides were designed as HCV NS5B polymerase
˚
Figure 2. X-ray complex structure of HCV NS5B with 13 at a resolution of 2.2 A. (a) Molecular surface representation of NS5B with bound inhibitor
13, electron density of the inhibitor in solid-shape with green color; (b) ribbon representation; GDD motif represents the activated site of NS5B. (c)
electron-density map contoured at a 2r level; (d) binding mode of 13 with removal of electron-density map for clarity. The pictures were generated by
VMD1.8 and rendered by POV-Ray3.5.19–21

S. Yan et al. / Bioorg. Med. Chem. Lett. 17 (2007) 1991–1995
1995
8. Pfefferkorn, J. A.; Nugent, R.; Gross, R. J.; Greene, M.;
Mitchell, M. A.; Reding, M. T.; Funk, L. A.; Anderson,
R.; Wells, P. A.; Shelly, J. A.; Anstadt, R.; Finzel, B. C.;
Harris, M. S.; Kilkuskie, R. E.; Kopta, L. A.; Schwende,
F. J. Bioorg. Med. Chem. Lett. 2005, 15, 2812.
9. X-ray studies of NS5B inhibitors having a binding to this
additional pocket have been reported by Love et al. in Ref
6. But at the time when this work was initiated, to our best
knowledge, no such inhibitors were publicly known.
10. GOLD v2.1 docking software from: www.cdc.cam.ac.uk,
Cambridge, UK.
11. Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.;
Bhat, T. N.; Weissig, H.; Shindyalov, I. N.; Bourne, P. E.
Nucleic Acids Res. 2000, 28, 235.
12. Tucci, F. C.; Zhu, Y. F.; Guo, Z.; Gross, T. D.; Connors,
P. J., Jr.; Gao, Y.; Rowbottom, M. W.; Struthers, R. S.;
Reinhart, G. J.; Xie, Q.; Chen, T. K.; Bozigian, H.; Killam
Bonneville, A. L.; Fisher, A.; Jin, L.; Saunders, J.; Chen,
C. J. Med. Chem. 2004, 47, 3483.
13. Mewshaw, R. E.; Zhou, D.; Zhou, P.; Shi, X.; Hornby, G.;
Spangler, T.; Scerni, R.; Smith, D.; Schechter, L. E.;
Andree, T. H. J. Med. Chem. 2004, 47, 3823.
Figure 3. Superimposition of X-ray structure of bound inhibitor 13
and the top-scored GOLD docking pose of the model structure; atom-
type coloring scheme is for 13, and yellow color is for the model.
allosteric inhibitors. The structure-based drug designs
were guided by docking results that revealed the poten-
tial to explore an additional pocket in the allosteric site.
In particular, such designed molecules contain moieties
of previously described thiazolone, and a newly designed
acylsulfonamide linker that is in turn connected with a
substituted aromatic ring. The selected compounds were
synthesized and demonstrated to be active with low lM
potency. The X-ray complex structure was established at
14. Bhattacharya, S.; Srivastava, A. Proc. Indian Acad. Sci.
(Chem. Sci.) 2003, 115, 613.
15. Tomlinson, I. D.; Grey, J. L.; Rosenthal, S. J. Molecules
2002, 7.
16. Cherney, R. J.; Mo, R.; Meyer, D. T.; Hardman, K. D.;
Liu, R. Q.; Covington, M. B.; Qian, M.; Wasserman, Z.
R.; Christ, D. D.; Trzaskos, J. M.; Newton, R. C.;
Decicco, C. P. J. Med. Chem. 2004, 47, 2981.
17. Shim, J.; Larson, G.; Lai, V.; Naim, S.; Wu, J. Z. Antiviral
Res. 2003, 58, 243.
˚
a 2.2 A resolution and converged with the initial SBDD
principle.
18. Coordinates for the complex structure have bee deposited
at Protein Data Bank (www.rcsb.org) under the Code#:
2O5D.
19. Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graph.
1996, 14, 27.
References and notes
20. VMD 1.8 is available from: http://www.ks.uiuc.edu/Re-
search/vmd/.
1. Choo, Q. L.; Kuo, G.; Weiner, A. J.; Overby, L. R.;
Bradley, D. W.; Houghton, M. Science 1989, 244, 359.
2. Ali, N.; Tardif, K. D.; Siddiqui, A. J. Virol. 2002, 76,
12001.
3. Kolykhalov, A. A.; Mihalik, K.; Feinstone, S. M.; Rice, C.
M. J. Virol. 2000, 74, 2046.
4. Yan, S.; Appleby, T.; Larson, G.; Wu, J. Z.; Hamatake,
R.; Hong, Z.; Yao, N. Bioorg. Med. Chem. Lett. 2006, 16,
5888.
5. Yan, S.; Larson, G.; Wu, J. Z.; Appleby, T.; Ding, Y.;
Hamatake, R.; Hong, Z.; Yao, N. Bioorg. Med. Chem.
Lett. 2007, 17, 63.
6. 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.
7. 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.
21. Pov-Ray is available from: http://www.povray.org.
22. It is worth noting that the described molecules are very
rigid, and may possess entropic disadvantage for a surface
binding site like this which normally requires subtle
structural adjustment of both ligand and engaging resi-
dues of protein for a favorable binding. Furthermore, the
newly designed moiety targets hydrophilic residues of Lys,
Arg, and His on the surface which are presumably
solvated in apo state of the enzyme. Any direct interaction
from a ligand to these residues would have to overcome
desolvation first. The desolvation process is energetically
unfavorable and could offset the additional interaction
observed by X-ray structure. This may account for the
flatness of the SAR for this scaffold in this region of the
binding site despite the predicted/observed additional
protein-ligand interactions. Future direction should be
directed toward the relaxation of the rigidity of the
inhibitors, taking into consideration the desolvation, and
at the same time retaining the similar protein-ligand
interaction network.