Importance of ligand bioactive conformation in the discovery of potent indole-diamide inhibitors of the hepatitis C virus NS5B

Article abstract of DOI:10.1021/ja101358s

Significant advances have led to receptor induced-fit and conformational selection models for describing bimolecular recognition, but a more comprehensive view must evolve to also include ligand shape and conformational changes. Here, we describe an example where a ligand's "structural hinge" influences potency by inducing an "L-shape" bioactive conformation, and due to its solvent exposure in the complex, reasonable conformation-activity-relationships can be qualitatively attributed. From a ligand design perspective, this feature was exploited by successful linker hopping to an alternate "structural hinge" that led to a new and promising chemical series which matched the ligand bioactive conformation and the pocket bioactive space. Using a combination of X-ray crystallography, NMR and modeling with support from binding-site resistance mutant studies and photoaffinity labeling experiments, we were able to derive inhibitor-polymerase complexes for various chemical series.

Full text of DOI:10.1021/ja101358s

Published on Web 10/13/2010
Importance of Ligand Bioactive Conformation in the
Discovery of Potent Indole-Diamide Inhibitors of the Hepatitis
C Virus NS5B
Steven R. LaPlante,* James R. Gillard, Araz Jakalian, Norman Aubry,
Rene´ Coulombe, Christian Brochu, Youla S. Tsantrizos, Martin Poirier,
George Kukolj, and Pierre L. Beaulieu
Departments of Chemistry and Biological Sciences, Boehringer Ingelheim (Canada) Ltd.,
2100 Cunard St., LaVal, Quebec, Canada, H7S2G5
Received February 24, 2010; E-mail: steven.laplante@boehringer-ingelheim.com
Abstract: Significant advances have led to receptor induced-fit and conformational selection models for
describing bimolecular recognition, but a more comprehensive view must evolve to also include ligand
shape and conformational changes. Here, we describe an example where a ligand’s “structural hinge”
influences potency by inducing an “L-shape” bioactive conformation, and due to its solvent exposure in the
complex, reasonable conformation-activity-relationships can be qualitatively attributed. From a ligand design
perspective, this feature was exploited by successful linker hopping to an alternate “structural hinge” that
led to a new and promising chemical series which matched the ligand bioactive conformation and the
pocket bioactive space. Using a combination of X-ray crystallography, NMR and modeling with support
from binding-site resistance mutant studies and photoaffinity labeling experiments, we were able to derive
inhibitor-polymerase complexes for various chemical series.
and conformational changes.1,2 Binding may be viewed as the
collision of two flexible objects, where their transient shapes
Introduction
The binding of small ligands to macromolecules involves
many events that commonly include the establishment of
hydrogen-bonds, van der Waals surface contacts, lipophilic
interactions, and ionic attractions.1 Equally important are some
less characterized events that contribute to ligand-receptor
interactions. These may include, for example, desolvation effects
and conformational changes of the ligand and receptor. How-
ever, optimization of ligand binding, one of the primary goals
of medicinal chemistry, is often confounded by the reality that
each complex entails a unique combination of the above events.
Unfortunately, the deconvolution and importance of the indi-
vidual energetic contributions have often proven to be very
difficult or impossible.
must have sufficient complementarity to form the encounter
complex, that is followed by mutual adaptations to stabilize the
interaction. Deciphering the details of the collision and binding
events (i.e., adaptations and energetic contributions) remains
largely elusive and impractical in drug discovery.
Here, we describe an example where a ligand’s “structural
hinge” influences potency by inducing an “L-shape” bioactive
conformation, and due to its solvent exposure in the complex,
reasonable conformation-activity-relationships can be qualita-
tively attributed. This feature was exploited where successful
linker hopping to an alternate “structural hinge” led to a new
and promising chemical series which matches the ligand
bioactive conformation and the pocket bioactive space. Reveal-
ing complexes with HCV polymerase are disclosed. It is also
noteworthy that the combination of strategies employed for
monitoring ligand structure and dynamics properties, in concert
with medicinal chemistry as described here, should have general
drug design utility.
In addressing conformational changes, significant advances
have led to receptor-induced fit and conformational selection
models for describing bimolecular recognition, but a more
comprehensive view must evolve to also include ligand shape
There is certainly an urgent need for drugs that specifically
target HCV as it is estimated that 1-3% of of the world
population are infected. HCV infections may lead to serious
and often fatal liver diseases such as cirrosis and hepatocellular
carcinomas. The current treatment option consists of pegylated
interferon-ribavirin combinations; however, the treatment has
limited efficacy against the major genotype 1 of HCV and can
have severe side effects. In industrialized nations, HCV infection
has become the major cause for liver transplantation.3a,b NS5B
(1) (a) Teague, S. Nature ReV. Drug Disc. 2003, 2, 527–541. (b) Weikl,
T. R.; von Deuster, C. Proteins: Structure, Function, and Bioinfor-
matics 2009, 75, 104–110. (c) Hammes, G. G.; Chang, Y. C.; Oas,
T. G. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 13737–13741. (d)
Bakan, A.; Bahar, I. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 14349–
14354. (e) Lakomek, N. A.; Lange, O. F.; Oliver, F.; Walter, K. F.;
Korvin, F. A.; Fares, C.; Egger, D.; Lunkenheimer, P.; Meiler, J.;
Grubmueller, H.; Becker, S.; de Groot, B. L.; Griesinger, C. Biochem.
Soc. Trans. 2008, 36, 1433–1437. (f) Lange, O. F.; Lakomek, N. A.;
Fares, C.; Schroeder, G. F.; Gunnar, G.; Walter, K. F. A.; Becker, S.;
Meiler, J.; Grubmueller, H.; Griesinger, C.; de Groot, B. L. Science
2008, 320, 1471–1475. (g) Kern, D.; Zuiderweg, E. R. Curr. Opin.
Struct. Biol. 2003, 13, 748–757. (h) Mittermaier, A.; Kay, L. E. Science
2006, 312, 224–228.
(2) Koshland, D. AdV. Enzymol. Relat. Subj. Biochem. 1960, 22, 45–97.
9
15204 J. AM. CHEM. SOC. 2010, 132, 15204–15212
10.1021/ja101358s  2010 American Chemical Society

Bioactive Conformation in Bimolecular Recognition
A R T I C L E S
Figure 1. Compounds derived from three HCV polymerase drug discovery series are shown. The structures are colored as described in the text. Methods
for determining activities against HCV NS5B polymerase (genotype 1b) are also described in the Supporting Information and elsewhere.9g The activity
values are the mean from multiple determinations with the following IC50 and EC50 standard deviations, respectively: 1, ( 2.6; 2, ( 0.006, ( 25; 3, ( 0.12,
( 0.48; 4, ( 0.08, ( 0.009; 5, ( 0.07, ( 20a; 6, ( 0.05, ( 0.2; 7, ( 0.042a, ( 0.03; 8, ( 0.003; 9, ( 0.07a, ( 0.12. aStandard deviations were overestimated
at 50% of the activity values but are typically much less.
RNA-dependent RNA polymerase is already promising to be a
valid target for the development of novel anti-HCV therapeutics.
Recently, nucleoside analogs and non-nucleoside allosteric
inhibitors of the enzyme have demonstrated efficacy in the clinic,
particularly in combination with interferon-based regimens.3c-e
This work discloses novel compounds that are among the most
potent reported for allosteric “thumb pocket I” inhibitors of the
hepatitis C virus (HCV) NS5B polymerase, and close analogues
of compound 7 (Figure 1) are now reported in patents and in
clinical trials.3f,g
Results and Discussion
Even though constraining the conformation of inhibitors has
been recognized as an important strategy in drug design,4
relatively few reports have extensively characterized ligand free-
state flexibility and conformational changes upon binding to
macromolecules.5 The importance of an inhibitor’s shape and
(3) (a) Choo, Q.-L.; Kuo, G.; Weiner, A. J.; Overby, L. R.; Bradley, D. W.;
Houghton, M. Science 1989, 244, 359. (b) Lavanchy, D. LiVer Int.
2009, 29, 74. (c) O’Brien, C.; Godofsky, E.; Rodriguez-Torres, M.;
Afdhal, N.; Pappas, S. C.; Pockros, P.; Lawitz, E.; Bzowej, N.; Rustgi,
V.; Sulkowski, M.; Sherman, K.; Jacobson, I.; Chao, G.; Knox, S.;
Pietropaolo, K.; Brown, N. A. Hepatology 2005, 42, 234A. (d) Brown,
N. Expert. Opin. InVest. Drugs 2009, 18, 709. (e) Huang, Z.; Murray,
M. G.; Secrist, J. A., III. AntiViral Res. 2006, 71, 351. (f) Erhardt, A.;
Deterding, K.; Benhamou, Y.; Reiser, M.; Forns, X.; Pol, S.; Calleja,
J. L.; Ross, S.; Spangenberg, H. C.; Garcia-Samaniego, J.; Fuchs, M.;
Enriquez, J.; Wiegand, J.; Stern, J.; Wu, K.; Kukolj, G.; Marquis, M.;
Beaulieu, P.; Nehmiz, G.; Steffgen, J. AntiViral Therapy 2009, 14,
23–32. (g) Beaulieu, P. L. Expert Opin. Ther. Pat. 2009, 19, 145–
164.
(5) (a) Kramer, B.; Rarey, M.; Lengauer, T. Proteins 1999, 37, 228. (b)
Verdonk, M. L.; Cole, J. C.; Hartshorn, M. J.; Murray, C. W.; Taylor,
R. D. Proteins 2003, 52, 609. (c) Goodsell, D. S. CSH Protoc 2009,
2009 (pdb prot5200); (d) Morris, G. M.; Huey, R.; Olson, A. J. Curr.
Protoc. Bioinformatics 2008, chapter 8, unit 814; (e) Ewing, T. J.;
Makino, S.; Skillman, A. G.; Kuntz, I. D. J. Comput.-Aided Mol. Des.
2001, 15, 411. (f) McGann, M. R.; Almond, H. R.; Nicholls, A.; Grant,
J. A.; Brown, F. K. Biopolymers 2003, 68, 76. (g) Goto, J.; Kataoka,
R.; Muta, H.; Hirayama, N. J. Chem. Inf. Model. 2008, 48, 583. (h)
Tondel, K.; Anderssen, E.; Drablos, F. J. Comput.-Aided Mol. Des.
2006, 20, 131. (i) Jackson, S.; DeGrado, W.; Dwivedi, A.; Parthasa-
rathy, A.; Higley, A.; Krywko, J.; Rockwell, A.; Markwalder, J.; Wells,
G. J. Am. Chem. Soc. 1994, 116, 3220–3230. (j) Miller, C. P.; Collini,
M. D.; Harris, H. A. Bioorg. Med. Chem. Lett. 2003, 13, 2399–2403.
Udugamasooriya, D. G.; Spaller, M. R. Biopolymers 2008, 89, 653–
667. (k) Sharma, R.; Lee, J.; Wang, S.; Milne, G. W. A.; Lewin, N. E.;
Blumberg, P. M.; Marquez, V. E. J. Med. Chem. 1996, 39, 19–28. (l)
Martin, S. F. Pure Appl. Chem. 2007, 79, 193–200. (m) Harris,
P. W. R.; Hu¨gel, H. M.; Nurlawis, F. Mol. Simul. 2002, 28, 889–902.
(n) Giebel, L. B.; Cass, R.; Milligan, D. L.; Young, D.; Arze, R.;
Johnson, C. Biochemistry 1995, 34, 15430–15435.
(4) (a) Peng, J. W. J. Am. Chem. Soc. 2003, 125, 11116–11130. (b)
LaPlante, S. R.; Aubry, N.; Deziel, R.; Ni, F.; Xu, P. J. Am. Chem.
Soc. 2000, 122, 12530. (c) Peng, W.; Wilson, B. D.; Namanja, A. T.
J. Biomolecular NMR 2009, 45, 1–2. (d) Peng, J. W. Structure 2009,
17, 319–320. (e) Mauldin, R. V.; Carroll, M. J.; Lee, A. L. Structure
2009, 17, 386–394. (f) Taylor, R. E.; Chen, Y.; Beatty, A.; Myles,
D. C.; Zhou, Y. J. Am. Chem. Soc. 2003, 125, 26–27. (g) Thoma, G.;
Magnani, J. L.; Patton, J. T.; Ernst, B.; Jahnke, W. Angew. Chem.,
Int. Ed. 2001, 125, 1941–1945.
9
J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010 15205

A R T I C L E S
LaPlante et al.
Figure 2. Shown are compounds (colored as white, blue, and red) bound to the allosteric “thumb pocket I” of HCV polymerase (colored green) which were
derived from a combination of X-ray, NMR, and docking. (A) The complex involving compound 2 is displayed, and was determined by docking 2 to the
pocket using using the X-ray structure shown in Figure 2C as a guide which included overlaying the common benzimidazole cores and cyclohexanes. A
representative, low-energy pose is shown which is also consistent with the bound structure reported from tr-NOESY NMR data.7 Binding of compound 2
to HCV polymerase was confirmed by 600 MHz 1H NMR resonance broadening of 2 upon incremental addition of 0.5 mM stock polymerase. Also, an NMR
competition experiment demonstrated that the resonance broadening observed for compound 8 upon addition of polymerase was lost in the presence of
compound 2. (B) The complex involving compound 10 is displayed. It was derived in a similar manner as described in (A) with support from tr-NOESY
data (see Supporting Information) and NMR binding and competition experiments. (C) Shown is the X-ray complex11a of compound 8 along with NMR
binding data.
flexibility in the free and bound states were previously reported
in our drug discovery programs aimed at targeting the human
cyctomegalovirus protease,6 the HCV NS3 protease7a and NS5B
polymerase.7b As a result, this knowledge contributed to the
design of drug candidates in the later two cases that effectively
suppressed HCV viral load in human clinical trials.3f,8
Our HCV polymerase drug discovery program is a good
illustration of these concepts,3f,7b,9 and began with a structure
and dynamics analysis early during the validation of compound
1 (Figure 1) as an attractive hit from screening a large compound
collection. In combination with structure-activity relationships
(SAR), we gained a qualitative understanding of the binding
roles of the molecular subgroups such as the central benzimid-
azole scaffold, the structural hinge (blue colored phenethyl-
amide), the critical cyclohexyl, and the more tolerant left-hand
side. This evolved into three independent chemical series (Series
A-C in Figure 1).
For example, efforts along Series A (Figure 1) revealed that
the red-colored carboxylic acid of compound 2 provided potency
by rigidifying the free-state conformation to resemble an NMR-
derived “L” shape when bound.7b,9d The bound model of its
complex with HCV polymerase shown here (Figure 2A) and
SAR9d clearly showed that this acid (red) was exposed to solvent
and had no direct interaction with the polymerase. The 48-fold
(9) (a) Beaulieu, P. L.; Bo¨s, M.; Bousquet, Y.; Fazal, G.; Gauthier, J.;
Gillard, J.; Goulet, S.; LaPlante, S. L.; Poupart, M. A.; Lefebvre, S.;
McKercher, G.; Pellerin, C.; Austel, V.; Kukolj, G. Bioorg. Med.
Chem. Lett. 2004, 14, 119. (b) Beaulieu, P. L.; Bo¨s, M.; Bousquet,
Y.; DeRoy, P.; Fazal, G.; Gauthier, J.; Gillard, J.; Goulet, S.; Poupart,
M.-A.; Valois, S.; McKercher, G.; Kukolj, G. Bioorg. Med. Chem.
Lett. 2004, 14, 967. (c) Beaulieu, P. L.; Gillard, J.; Bykowski, D.;
Brochu, C.; Dansereau, N.; Duceppe, J. S.; Hache´, B.; Jakalian, A.;
Lagace´, L.; LaPlante, S. R.; McKercher, G.; Moreau, E.; Perreault,
S.; Stammers, T.; Thauvette, L.; Warrington, J.; Kukolj, G. Bioorg.
Med. Chem. Lett. 2006, 16, 4987. (d) Beaulieu, P. L.; Bousquet, Y.;
Gauthier, J.; Gillard, J.; Marquis, M.; McKercher, G.; Pellerin, C.;
Valois, S.; Kukolj, G. J. Med. Chem. 2004, 47, 6884. (e) Kukolj, G.;
McGibbon, G.; McKercher, G.; Marquis, M.; Lefebvre, S.; Thauvette,
L.; Gauthier, J.; Goulet, S.; Poupart, M. A.; Beaulieu, P. L. J. Biol.
Chem. 2005, 280, 39260–39267. (f) Goulet, S.; Poupart, M. A.; Gillard,
J.; Poirier, M.; Kukolj, G.; Beaulieu, P. L. Bioorg. Med. Chem. Lett.
2010, 20, 196–200. (g) McKercher, G.; Beaulieu, P. L.; Lamarre, D.;
LaPlante, S. R.; LeFebvre, S.; Pellerin, C.; Thauvette, L.; Kukolj, G.
Nucleic Acids Res. 2004, 32, 422.
(6) LaPlante, S. R.; Bonneau, R.; Aubry, N.; Cameron, D. R.; De´ziel, R.;
Grand-Maˆıtre, C.; Plouffe, C.; Tong, L.; Kawai, S. H. J. Am. Chem.
Soc. 1999, 121, 2974–2986.
(7) (a) LaPlante, S. R.; Llina`s-Brunet, M. Curr. Med. Chem.: Antiinfect.
Agents 2005, 4, 111–132. (b) LaPlante, S. R.; Jakalian, A.; Aubry,
N.; Bousquet, Y.; Ferland, J. M.; Gillard, J.; Lefebvre, S.; Poirier,
M.; Tsantrizos, Y. D.; Kukolj, G.; Beaulieu, P. L. Angew. Chem., Int.
Ed. 2004, 43, 4306–4311.
(8) Lamarre, D.; Anderson, P. C.; Bailey, M.; Beaulieu, P. L.; Bolger,
G.; Bonneau, P.; Bo¨s, M.; Cameron, D. R.; Cartier, M.; Cordingley,
M. G.; Faucher, A. M.; Goudreau, N.; Kawai, S. H.; Kukolj, G.;
Lagace´, L.; LaPlante, S. R.; Narjes, H.; Poupart, M. A.; Rancourt, J.;
Sentjens, R. E.; St George, R.; Simoneau, B.; Stelnmann, G.; Thibeault,
D.; Tsantrizos, Y. S.; Weldon, S. M.; Yong, C. Y.; Linas-Brunet, M.
Nature 2003, 426, 186.
9
15206 J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010

Bioactive Conformation in Bimolecular Recognition
A R T I C L E S
Figure 3. (A) A pocket view and full display of the X-ray structure of apo HCV polymerase is shown.11a Encircled is the closed form of the “λ1 finger-
loop”. The pocket view is rotated 180° along the vertical axis with respect to the full view for visualization purposes. (B) A pocket view of the complex of
compound 10 (colored as white, red and blue) with HCV polymerase (colored green) is displayed on the left side, and the full view of the X-ray structure
of the complex with compound 8 is shown on the right. Encircled is the unobserved, open form of the “λ1 finger-loop”. The pocket view is rotated 180°
along the vertical axis with respect to the full view for visualization purposes. (C) A bottom-view of the complex involving 10 is shown to emphasize the
importance of multiple weak CH-π interactions involving HCV polymerase Val 494, Pro 495, and Pro 496.
difference in binding affinity gained by addition of this acid
demonstrated the benefits of reducing the entropic costs of ligand
binding.7b Furthermore, knowledge of the acid’s role as a
“rigidifier” led to its replacement with a heterocyclic isosteres
(e.g., thiazole compound 3 in Figure 1). Despite somewhat
reduced intrinsic potency, the increase in lipophilicity of
compound 3 provided permeability and inhibition of subgenomic
HCV RNA replication in a cell-based assay (EC50 ) 1.7 µM).
Although the subsequent replacement of the benzimidazole
scaffold with a more lipophilic indole (compound 4) led to
further improvements in cell-culture activity (EC50 ) 0.05 µM),
intractable physicochemical properties compromised any further
development as a drug candidate.9c,f Given these liabilities,
alternate but related series of compounds were pursued.
The truncated compounds of Series C (Figure 1) had attractive
features. Similar intrinsic potency could be achieved with
multiple changes that exploited alternate interactions with the
polymerase pocket. The carboxylic acid in series C forms a
better interaction with Arg 503 (Vide infra, Figure 2C and 3B).
Replacement of the benzimidazole with a more lipophilic indole
core allows the introduction of an additional appendage on the
indole nitrogen that can engage in CsH to carbonyl interactions
with Leu 492 and Gly 493 (see Supporting Information). Such
interactions could also account for the potency gain in series A
between compounds 3 and 4.
However, this series suffered from poor cell culture activity
and other properties required to advance competitively to the
clinic.10 Regardless, one of the benefits with this series was that
X-ray structures were solved for compounds 811a (as described
here) and 911b bound to HCV polymerase (Figure 2C) which
also served as the starting points for creating NMR-supported
docking models of the complexes shown in Figure 2A,B (vide
infra).
Given that compounds from series A and C had some
undesirable properties with regard to advancement as clinical
candidates, a third series of compounds was sought and
designed; the new series is shown as series B9f in Figure 1.
(10) (a) Beaulieu, P. L. Curr. Opin. Drug Disc.DeV. 2006, 9, 618–626. (b)
Harper, S.; Avolio, S.; Pacini, B.; Di Filippo, M.; Altamura, S.; Tomei,
L.; Paonessa, G.; Di Marco, S.; Carfi, A.; Guiliano, C.; Padron, J.;
Bonelli, F.; Miglaccio, G.; De Francesco, R.; Laufer, R.; Rowley, M.;
Narjes, F. J. Med. Chem. 2005, 48, 4547–4557. (c) Beaulieu, P. L.;
Jolicoeur, E.; Gillard, G.; Brochu, C.; Coulombe, R.; Dansereau, N.;
Duan, J.; Garneau, M.; Jakalian, A.; Kuhn, P.; Lagace, L.; LaPlante,
S. R.; McKercher, G.; Perrault, S.; Poirier, M.; Poupart, M.-A.;
Stammers, T.; Thauvette, L.; Thavonekham, B.; Kukolj, G. Bioorg.
Med. Chem. Lett. 2010, 20, 857–861.
9
J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010 15207

A R T I C L E S
LaPlante et al.
Overall, the design was based on the knowledge of the binding
mode and role of each substituent from compound 2 of series
A.7b,9f It was thought that opportunities for new chemical entities
could be identified by replacing the blue-colored “structural
hinge” and the red-colored acid “rigidifier” of 2. For this
purpose, a rational and novel parallel synthesis strategy9f was
developed where synthetically feasible replacements were
identified that were presumed to access the same spatial and
conformational roles. From several small libraries, the diamide
cinnamic acid moiety shown in blue in compounds 5 and 6
(Figure 1) was suitable as it allowed us to keep the spatial
arrangements of the critical left- and right-side pharmacophores
(colored in black on compounds 2,5,6). The nearly 6-fold
potency improvement in IC50 observed upon R,R-disubstitution
(compound 5 versus 6, Figure 1) made this new series attractive.
19F NMR, kinetics and mechanistic studies, etc) are consistent
with specific binding to the HCV polymerase.7b,9,10
The bioactive conformation of compounds 27b and 10 were
then determined by NMR transferred NOESY (tr-NOESY)
experiments12 (see also Supporting Information) when bound
to HCV polymerase. Modeling calculations then helped to dock
these compounds into “thumb pocket I” in a fashion similar to
that found in the X-ray structure of compound 811a (Figure 2C);
which included overlaying the common benzimidazole cores
and cyclohexanes. Again, this was also corroborated by binding-
site resistance mutant studies which suggested that inhibitors
bound close to Pro 495 and Pro 496.9e The transferred NOESY
distance information extracted from the bound-states were also
consistent with the bioactive “L” shape of compounds 2 and
10 in Figure 2A,B.
Deriving polymerase complexes with compounds from each
series shown in Figure 2 was achieved using a combination of
X-ray crystallography, NMR and modeling with support from
binding-site resistance mutant studies and photoaffinity labeling
experiments. The structure of 8 bound to “thumb pocket I” was
determined by X-ray crystallography and is shown in Figure
2C, and when compared to the X-ray structure of apo poly-
merase, it became clear that 8 required the displacement of the
λ1 finger loop from the thumb domain for binding. See the apo
and complex views shown in Figure 3.
The NMR-consistent models involving compounds 2 and 10
revealed notable features. For example, the role of the blue-
colored structural hinge became apparent as its curved shape
positioned the right-side aromatic rings over the ledge formed
by Val 494 and Pro 495 onto a plateau formed by Pro 495 and
Pro 496 (Figure 3C), which were earlier identified as part of
the binding site by a resistant mutant study.9e The bioactive
“L” conformation of compound 10 is clearly stabilized by a
significant network of weak interactions involving CsH’s from
Val 494, Pro 495, and Pro 496 to the aromatic rings of the
benzimidazole scaffold and the right-side moiety as shown in
Figure 3C. The guanidinium of Arg 503 further stabilizes this
complex by hydrogen bonding with the left-side amide carbonyl
of 10 as shown in Figure 3B.
The complex involving compound 10 in Figure 2 also
revealed that the red-colored piperidine is completely solvent
exposed and does not form any direct interactions with the
polymerase pocket. When compared to the glycine analogue 5,
addition of the piperidine group resulted in a 6.5-fold improve-
ment in intrinsic potency. Given that this group is solvent-
exposed in the bound-state, it may contribute to potency by
promoting the bioactive conformation in the free-state; this could
also explain the potency improvements for the other R,R-
disubstituted analogues 6, 10, 13, and 15 relative to the glycine
analogue 5.
Molecular dynamics calculations are a practical means of
exploring the free-state, conformational properties of ligands,13
and our simulations also supported the above interpretation. It
was evident in the data in Figure 4 that the naked glycine
analogue 5 sampled a wider range of S1 and S2 torsional angles
over time (solid bars in Figure 4B,C), indicative of it sampling
a variety of conformations, whereas only a narrow range of the
more favorable, bound-like conformations was sampled for 6
and localized around 80° and 70° for S1 and S2 in Figure 4B,C,
respectively. Clearly, the more active cyclobutyl analogue 6
prefers to adapt a narrower range of conformations in the free-
state (diamond lines and points in Figure 4) that are a better
match to the bound-state. In a sense, the cyclobutyl helps to
Attempts to obtain bound crystal structures of compounds
from series A and B proved to be difficult due the crstal cracking
as described in the Materials and Methods and in the Supporting
Information. In short, the crystal quality is likely affected by
the dynamic properties of the loop and the conformational
changes upon ligand binding. Therefore, binding site determina-
tions had to be secured by NMR binding and competition
experiments. As a control, it was first demonstrated that
compound 8 experienced NMR line-broadening (DLB) upon
incremental addition of HCV polymerase (red), as a result of
reversible fast-exchange binding. The data are displayed at the
bottom of Series C in Figure 2, where the resonance of free 8
(blue) broadened (red) upon incremental addition of small aounts
of stock 0.5 mM HCV polymerase. This showed that compound
8 binds to HCV polymerase, as also demonstrated by the X-ray
structure shown in Figure 2 Series C. Likewise, NMR binding
studies confirmed that compounds 2 and 10 also bind to HCV
polymerase (DLB data shown at the bottom of Series A and B
in Figure 2). Subsequent NMR competition experiments on
samples containing compound 8 (as a probe or spy molecule)
with HCV polymerase showed that its resonance (colored black
at the bottom of Series A and B) no longer broadens upon
addition of 2 and 10. In fact, it resembles that of the blue
resonance of 8 of the free state, suggesting that it is prevented
from binding to the pocket.
Overall, the ensemble of data is consistent with compounds
2 and 10 binding at the same “thumb pocket 1” site as 8. Our
confidence that these compounds bind to the same site was also
corroborated by resistant mutant and photoaffinity labeling
studies. We were able to locate the binding region using a
photoaffinity label on an analogue of compound 2 from series
A, and resistant mutants were selected at Pro 495 and Pro 496
in the presence of compound 6 of Series B.9e A close-up view
at Figure 3C shows that these amino acids are located at “thumb
pocket 1” and lie in close proximity to the critical L-shape
created by compounds 2 and 6 in the complex. In summary,
the ensemble of data acquired here and described elsewhere
(including appropriate inhibition curves, counterscreen assays,
(11) (a) PDB accession codes 3MWV for the apo HCV polymerase, and
3MWW for the complex with inhibitor 8. (b) PDB code 2BRK: Di
Marco, S.; Volpari, C.; Tomei, L.; Altamura, S.; Harper, S.; Narjes,
F.; Koch, U.; Rowley, M.; De Francesco, R.; Migliaccio, G.; Carf´ı,
A. J. Biol. Chem. 2005, 280, 29765–29770.
(12) (a) Ni, F. Prog. NMR Spect. 1994, 26, 517. (b) LaPlante, S. R.; Aubry,
N.; Bonneau, P.; Kukolj, G.; Lamarre, D.; Lefebvre, S.; Li, H.; Llina`s-
Brunet, M.; Plouffe, C.; Cameron, D. R. Bioorg. Med. Chem. Lett.
2000, 10, 2271.
(13) Hao, M.-H.; Muegge, I. J. Chem. Inf. Model. 2007, 47, 2242–2252.
9
15208 J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010

Bioactive Conformation in Bimolecular Recognition
A R T I C L E S
Figure 5. Analogues of series B are shown. See the Supporting Information
for synthetic schemes and characterization. Methods for determining
activities against HCV NS5B polymerase (genotype 1b) are also described
in the Supporting Information and elsewhere.9g The activity values are the
mean from multiple determinations with the following IC50 and EC50
standard deviations, respectively: 5, ( 0.07, ( 20a; 11, ( 0.08, ( 8; 12, (
0.06, ( 1.6; 13, ( 0.04, ( 0.4; 14, ( 0.03, ( 0.8; 6, ( 0.05, ( 0.2; 15,
( 0.03, ( 0.12; 16, ( 0.013, ( 0.04; 10, ( 0.01, ( 2.5a; 17, ( 0.23a, (
a
1a. Standard deviations were overestimated at 50% of the activity values
Figure 4. (A) Compounds 5 and 6 are shown along with free-state 13C
NMR (125 MHz) NT1 relaxation times (seconds). The NT1 data are displayed
beside each carbon atom for compound 5, and data for compound 6 are
given within square brackets. The samples were prepared in DMSO-d6
solvent and data were acquired as described elsewhere.4b,7b In cases where
two hydrogens are attached to a single carbon, NT1 data are given (the
product of the number of attached protons and the longitudinal relaxation
time). Note that due to the symmetry of the phenyl, only two relaxation
data points are observed and reported. (B) Molecular dynamics histograms
are shown for the number of angular occurrences for each torsion angles
of S1 for compounds 5 and 6 in the free-state. See Figure 4A for a definition
of the S1 torsion angle. (C) Molecular dynamics histograms are shown for
the number of angular occurrences for each torsion angles of S2 for
compounds 5 and 6 in the free-state. See Figure 4A for the definition of
the S2 torsion angle.
but are typically much less.
Conformational selection and an inhibitor’s flexibility in the
free-state likely influence many drug discovery programs. It may
account for the frequently observed phenomenon of nonlinear
SAR trends, which can be analyzed using “box analyses”.14 For
example, one would not have expected the 7-fold improvement
in potency of the gem-dimethyl analogue 13 relative to the naked
glycine 5 (Figure 5), considering that the (R) alanine analogue
(compounds 12) resulted in only a 2.5-fold improvement over
5, whereas the (S)-distomer had comparable activity to 5.9f The
potencies of the two alanine analogues simply are not additive
based on the improvement observed for the gem-dimethyl
analogue, and highlights the importance of conformational
selection and entropy in drug discovery.
“fine-tune” the structural hinge to resemble the bioactive
conformation.
The important role of the diamide segment described herein
can also help to decipher the SAR impact of multiple substit-
uents shown in Figure 5. For example, the cyclopropyl moiety
in compound 14 shown in Figure 5 resulted in a 3-fold loss in
potency compared to the closely related dimethyl 13 because
the cyclopropyl imparted an unfavorable angle involving the
RC center.15 The larger ring sizes of compounds 15 and 16
perturb the free-state population away from the bioactive
13C spin-lattice relaxation (13C T1) experiments were em-
ployed as an NMR parameter for monitoring the relative
flexibility of CsH vectors. As a method which is sensitive to
motions on the pico- to nanosecond time scales, shorter
relaxation times are typically indicative of relatively slower
segmental motion or flexibility. The T1 times reported here result
from the collection of a single set of experiments, and an error
range of (12% is anticipated (see Supporting Information).
Overall, much longer T1 times were not observed for either the
glycine 5 or the cyclobutyl 6 analogues, as observed with
flexible compounds in other programs,6,7 suggesting that the
diamide hinge effectively reduces the flexibility of the inhibitors.
It may be noteworthy in Figure 4A that the glycine analogue 5
has slightly longer T1 times in general (but close to the
anticipated error range) as compared to the cylcobutyl compound
6, which could be reflective of an limited increase in flexibility
that also goes beyond the local torsion angles S1 and S2.
(14) (a) Kawai, S. H.; Bailey, M. D.; Halmos, T.; Forgione, P.; LaPlante,
S. R.; Llina`s-Brunet, M.; Naud, J.; Goudreau, N. ChemMedChem 2008,
3, 1654–1657. (b) Cockroft, S. L.; Hunter, C. A. Chem. Soc. ReV.
2006, 36, 172–188. (c) Adams, H.; Carver, F. J.; Hunter, C. A.;
Morales, J. C.; Seward, E. M. Angew. Chem., Int. Ed. 1996, 35, 1542–
1544. (d) Fischer, F. R.; Schweizer, W. B.; Diederich, F. Angew.
Chem., Int. Ed. 2007, 46, 8270–8273.
(15) (a) Pirrung, M. C. J. Org. Chem. 1987, 52, 4179. (b) Valle, G.; Crisma,
M.; Toniolo, C.; Holt, E. M.; Tamura, M.; Bland, J. Int. J. Pept. Protein
Res. 1989, 34, 56. (c) Di Blasio, B.; Pavone, V.; Pedone, C. Cryst.
Struct. Commun. 1977, 6, 745.
9
J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010 15209

A R T I C L E S
LaPlante et al.
to the wells and the plate was left at room temperature for one
hour before quantifying the radioactive UMP incorporated onto the
biotinylated primer by counting for 60 s on a TopCount (Packard).
From serial dilution of the test compound, the percent inhibition
was plotted against the compound concentration and a nonlinear
curve was fitted (Hill model) to the % inhibition-concentration data.
The calculated % inhibition values were then used to determine
the median inhibitory concentration IC50, slope factor (n) and
maximum inhibition (Imax) by the nonlinear regression procedure
of SAS (Statistical Software System, SAS Institute Inc., Cary, N.C.)
conformation by the introduction of various chair/boat puckering
that affect the S1 and S2 torsion angles (defined in Figure 4).
This is more evident when one compares the 10-fold activity
difference between 16 and 10, which also differ significantly
in S1 and S2 energy torsion profiles and free-state conformations
as determined by NMR. The incorporation of a nitrogen as
shown in 10 favors the bioactive free-state population along
torsion angles S1 and S2 as compared to the carbocycle 16 (see
Supporting Information). Finally, the subsequent addition of a
large BOC group as in compound 17 again alters the free-state
chair/boat and S1/S2 free-state equilibrium, resulting in a 10-
fold loss in potency compared to 10. The reasonable activity of
compound 17 is nonetheless informative in that the addition of
the bulky BOC group on the nitrogen is consistent with the
hinge being exposed to solvent in the complex.
The polymerase “thumb pocket I” undergoes major changes
upon ligand binding. An X-ray structure of apo HCV polymerase
is displayed in Figure 3A, which reveals that the “λ1 finger-
loop” naturally occupies this polymerase pocket as deduced from
well-defined electron density. However, compound 8 occupies
this pocket in the complex, as shown in Figure 3B, and there is
a lack of electron density for the loop residues. Overall, this
suggests that the loop adopts a disengaged conformation (i.e.,
loop residues are displaced, exposed to solvent, and flexible).
The precise mechanism that results in the release of the λ1-
loop from the pocket is unclear, however, the disengaged state
likely exists at a low-level population in the apo form and
reversible ligand binding shifts the equilibrium toward this
alternate complex state. The displacement and increased flex-
ibility of the finger loop was also consistent with 15N TROSY
NMR experiments, where a new subset of sharper crosspeaks
appeared upon the addition of compound 8 as compared to the
15N TROSY spectra of apo polymerase (see Supporting Infor-
mation).
Protein for NMR experiments were prepared from constructs
expressing soluble versions that lack the C-terminal 21 amino acids
and encode a C-terminal hexa-histidine tag,9g purified as described
above, exchanged with deuterated buffer components (20 mM Tris
d-11 pH 7.5, 10% glycerol d-8, 2 mM DTT d-10, 1 mM EDTA D-16,
300 mM NaCl) and concentrated. The genotype 1b J4 strain NS5B
lacking the C-terminal 21 amino acids and encoding a C-terminal
hexa-histidine tag (HCVNS5B1bJ4-∆C21-H6) was used for X-ray
crystallography and also expressed in E. coli in 15N labeled bacterial
growth medium and purified for use in TROSY experiments. The
HCVNS5B1bJ4-∆C21-H6 used for X-ray crystallography was
purified as described above, further concentrated on Hi-Trap SP
column and then subjected to size exclusion chromatography on a
Superdex-200 column to obtain a homogeneous, monomeric and
highly purified preparation.
Determination of X-ray Structures of Apo HCV Polymerase
and Inhibitor Complex. Crystallization of apo protein proceeded
as follows. Crystal of HCVNS5B1bJ4-∆C21-H6 were grown by
the hanging drop technique using Nextal’s plates (Qiagen). The
well solution consisted of 100 mM MES pH 5.4, 21% PEG5000
mme, 400 mM ammonium sulfate and 10% glycerol. The crystal-
lization drops were made of 1 µL of the well solution and 1 µL of
protein solution at a concentration of 7.7 mg/mL in the final
purification buffer (20 mM tris pH 7.5, 300 mM NaCl, 10% glycerol
and 5 mM DTT). The crystallization plate was then incubated at
11 degree. Large crystals were observed within a week.
Notably, we recently reported9e that the L30R and L30S
mutant constructs at the tip of the λ1 finger-loop resulted in a
∼100-fold improvement in affinity for compound 8, suggesting
that the mutant constructs further shift the equilibrium to the
disengaged state relative to wild-type. These results are also
consistent with a receptor conformation selection mechanism
for inhibitor binding, and the λ1 loop conformational changes
are critical for compound binding and pharmacological inhibition
of HCV polymerase function.
Inhibitor soaking proceeded as follows. The NS5B crystals were
very sensitive to solution transfer. Eventually, crystals were
successfully transferred in a 5 µL drop of a soaking solution copying
the content of crystallization drop at the end of the vapor diffusion
experiment. The pH of the buffer was raised to be more similar to
assay conditions. The soaking solution was made of 14%
PEG5Kmme, 14 mM tris pH 7.5, 70 mM TES pH 7.0, 14%glycerol,
210 mM NaCl, 10 mg/mL lysosyme, and 280 mM ammonium
sulfate. The inhibitor molecule was dissolved in DMSO to a
concentration of 25 mM and 0.2 µL of this inhibitor stock solution
was added to the 5 µL soaking drop. The crystallization support
was screwed back on top of the well solution and let untouched
for 5 h at 11 °C. After incubation, the crystal was transfer in a
cryoloop (Hampton Research, California, USA), plunged and stored
in liquid nitrogen prior diffraction data collection.
Materials and Methods
HCV NS5B Polymerase Constructs for IC50 Activity Assays
and NMR/X-ray Crystallography. HCV genotype 1b NS5B strain
1b-40 and J4 were used to clone into baculovirus and E. coli
expression vectors.9g The full length HCV NS5B polymerase used
for compound IC50 determination was expressed as a hexa-histidine
fusion protein in baculovirus infected cells and purified by metal
chelating chromatography on Ni2+ column followed by DEAE-
sepharose and heparin-sepharose chromatography. The preparation
was homogeneous by SDS-PAGE analysis and used at 10 nM in
the standard HCV polymerase assay. This scintillation proximity
assay uses 0.5 µCi of [3H]-UTP, 1 µM UTP, 250 nM 5′-biotinylated
oligo(rU12), 10 µg/mL poly(rA) in 20 mM Tris-HCl pH 7.5, 5 mM
MgCl2, 1 mM EDTA, 1 mM DTT, 0.2 U/µL of RNasin, 5% DMSO,
3% glycerol, 30 mM NaCl, 0.33% dodecyl-ꢀ-D-maltoside, 0.01%
IGEPAL. The 60-µL reaction was terminated after 90 min at 22
°C by the addition of 20 µL of stop solution (150 µg/mL of tRNA
in 0.5 M EDTA) and 30 µL of streptavidin-coated beads (8 mg/
mL in 20 mM Tris-HCl pH 7.5, 25 mM KCl, 0.025% (w/v) sodium
azide) for Scintillation Proximity Assay (Perkin-Elmer). After 30
min at room temperature, 75 µL of 5 M cesium chloride were added
It was found that upon inhibitor soaking the lattice of the crystal
was affected such as the resolution of the diffraction pattern is
diminished proportionally to soaking time. Although an apo crystal
could be soaked in the described soaking solution for more than
24 h and still diffract to 2.2 Å resolution, a same soaking time in
presence of an inhibitor molecule resulted in a reduced resolution
limit to worst than 3 Å. Optimal e-density maps were obtained
with a soaking time of 5 h. This time was enough to obtain sufficient
occupancy of the inhibitor molecule while keeping the diffraction
to a resolution that allow clear interpretation.
For the structure determinations and refinements, the diffraction
data of the apo NS5B were phased by Molecular Replacement (MR)
using publicly available structure of HCV NS5B (pdb code: 1C2P).
Rotation and translation search were done using the program CNX
(Accelrys). Model building was carried out with the software O
9
15210 J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010

Bioactive Conformation in Bimolecular Recognition
A R T I C L E S
(Alwyn Jones, Upsala University, Sweden),16 and model refinement
was performed with the CNX software (Accelrys). The new model
was improved by a cycling procedure of model rebuilding and
refinement steps. The final model include two molecules of NS5B
(residues A1-A149, A154 to A563, B1 to B147, and B153 to B563).
The tip of the Λ2 finger loop (150-152) appeared to be flexible
and was not modeled. See the Table S1 in the Supporting
Information for the refinement statistics.
The structure of the complex involving compound 8 was solved
by MR using the apo structure as a model. On the second cycle,
the residues B18 to B35 were removed as the initial density map
was indicating a clear conformation change of that loop. The
subsequent difference e-density maps revealed a clear density for
the inhibitor molecule on the newly exposed surface of the thumb
domain. The final model includes two molecules of NS5B (residue
A1 to A149, A154 to A563, B1 to B17, B36 to B148, and B153 to
B563) and one molecule of compound 8 associated with NS5B
chain B. See Table S1 in the Supporting Information for refinement
statistics.
All data sets were processed using the TOPSPIN or WinNMR
software (Bruker Canada, Milton, Ontario) and software. Data sets
were typically zero-filled to yield 2048 (f2) × 1024 (f1) points
after Fourier transformation using a shifted sine-bell window
function. 1H and 13C spectra were chemical shift calibrated relative
to the standard values attributed to the DMSO peaks (2.49 and 39.51
ppm, respectively).
Molecular Dynamics Simulations and Docking. Molecular
dynamics (MD) simulations were performed for compounds 5 and
6 to generate free-state conformational histograms for angles S1
and S2 as shown in Figure 4. Each compound was first submitted
to a conformation sampling to generate 10 diverse conformations
that were then, each, submitted to 1 ns. MD run. The MD
simulations were run with MOE (versions 2008.10) under constant
volume and constant temperature conditions using MMFF94 (as
implemented in MOE version 2008.10). Each MD simulation
consisted of a heating phase that gradually increased the system
temperature from 0 K to 298 K in 10 ps, followed by a equilibration
phase at the target temperature of 298 K for 10 ps. The production
phase (data accumulation phase) consisted of 1 ns and snapshots
of the system were recorded every 0.5 ps. The nonbonded cutoff
radius was set to infinity and an implicit solvation model was used
to mimic an aqueous environment. The production phase trajectories
of the 10 MD runs were then combined to analyze the S1 and S2
angles with the Conformation Geometries module of MOE.
Compounds were docked to the allosteric “thumb pocket I” of
HCV polymerase using Glide (version 2007) in XP mode and with
the use of experimentally derived structures from X-ray crystal-
lography and bound-state information from NMR binding and
competition studies.
Despite the medium resolution of the data, most of the inhibitor
molecule was well-defined in the e-density maps with the exception
of the (N-acetyl substitution) which should be considered as flexible.
Also, the exact orientation of the furan group is not clearly defined
by the data as no H-bond interaction was found to differentiate the
oxygen atoms from the carbon of the ring.
X-ray data collection and refinement statistics are provided in
the Supporting Information. Also, both structures have been
deposited in the Protein Data Bank (accession code 3MWV and
3MWW for the apo and the inhibitor complex respectively).
NMR Binding and Competition Experiments. NMR experiments
were performed on Bruker Avance spectrometers (400 and/or 600
1
MHz H frequency) at 27 °C, unless otherwise stated.
Conclusions
Samples to be used for NMR binding and competition experi-
ments were prepared by adding concentrated inhibitor in DMSO-
d6 (typically 0.3 mgs. of compound) to an aqueous buffer composed
of 20 mM Tris-d11, 2 mM DTT-d10, 1 mM EDTA-d12, 300 mM
NaCl, and 10% (v/v) D2O spiked with TSP-2,2,3,3-d4. Buffer was
added to a final volume of 600 µL, and the pH was adjusted to 6.0
with diluted HCl. Spectra of free compound at 0.2 mM were then
acquired. A concentrated stock solution of HCV polymerase was
then added at less than stochiometric amounts to the NMR tube,
and spectra were again acquired at each increment such that the
inhibitor to polymerase ratio ranged from 40:1 to 10:1. The
polymerase stock solution contained 0.5 mM HCV polymerase
(NS5B∆21C-His6) in buffer consisting of 20 mM Tris-d11, 2 mM
DTT-d10, 1 mM EDTA-d12, 300 mM NaCl, and 10% (v/v) glycerol-
d8. NOESY spectra were also collected on these samples. For
competition experiments, concentrated stock solutions of other
compounds were added to the NMR tube followed by the
acquisition of spectra. TROSY NMR experiments were run using
26 µM NS5B∆21C-His6 (1b/J4) labeled with 15N in Tris-d11 25
mM, pH 7.5, NaCl 150 mM, glycerol-d8 5%, TCEP-d16 1 mM.
For spectra acquired on samples in buffer, suppression of the
solvent signal was achieved by the use of presaturation or by
inserting a 3-9-19 WATERGATE module prior to data acquisi-
tion. NOESY experiments on sample tubes containing no poly-
merase resulted in the observation of no significant cross-peaks.
However, NOESY spectra on sample tubes containing polymerase
resulted in the observation of many cross-peaks which contained
the valuable interhydrogen distance information of the compound
when bound to HCV polymerase. Thus, a series of NOESY spectra
were typically acquired on these latter samples which included the
following mixing times 50, 100, 200, and 300 ms (600 MHz).
NOESY crosspeak volumes were scaled and converted to apparent
interproton distances. The derived distances were normalized via
comparisons with the volumes from an internal fixed interhydrogen
distance.
In the context of deriving novel inhibitors of HCV poly-
merase, the bimolecular recognition we described here may be
represented by the collision of two flexible objects. Recognition
involves a disengaged “finger loop” from an allosteric site of
the thumb domain of the polymerase, and binding of ligands in
the bioactive “L” conformation (series A and B). From a ligand
design perspective, the work described here presents a good
example of how a ligand’s “structural hinge” influences potency,
and due to its solvent exposure in the complex, reasonable
structure-activity-relationships can be interpreted. Although the
other segments of the inhibitors in series A, B, and C are also
important for potency, such as the cyclohexyl and the lipophilic
benzimidazole/indole cores, the structural hinge can impart
significant potency that can be “fine tuned”. Moreover, variants
of the hinge also significantly influence cell culture activity and
other properties, which ultimately allowed us to successfully
transition from compounds that had poor physicochemical and
pharmacodynamic properties (series A) to a more promising
series (series B). As expected, the improvements in cell culture
potency observed in series A upon replacement of the benz-
imidazole by an indole core (compounds 3 to 4) were also
translated in the new series B (compounds 6 to 7), and likely
arise from an increase in lipophilic character (and cell membrane
permeability).
Overall, compound 7 maintains good enzymatic potency and
has an impressive cell-culture potency (IC50 ) 0.095 µM, EC50
) 0.072 µM,). Unlike compound 4, it also has beneficial
physicochemical and pharmacokinetic properties that promise
new opportunities for progress toward the development of novel
HCV therapeutics in the clinic. Indeed, a close analogue of
compound 7 has shown antiviral effects in clinic studies.3f
Acknowledgment. We are grateful to M.-A. Poupart for his
contributions, and to S. Lefebvre and G. McKercher for materials,
(16) Jones, T. A.; Zou, J.-Y.; Cowan, S. W.; Kjeldgaarrd, M. Acta
Crystallogr. 1991, A47, 110–119.
9
J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010 15211

A R T I C L E S
LaPlante et al.
M. Little and S. Bordeleau for HRMS assistance, along with P.
Bonneau, P. Edwards, R. Bethell, and M. Cordingley for support.
comparison of compounds 10 and 16 in the free state, NMR
data on compound 10 binding to HCV polymerase in solution,
and details regarding inhibitor synthesis and characterization.
This material is available free of charge via the Internet at http://
pubs.acs.org.
Supporting Information Available: Additional materials and
methods information are provided that include: X-ray data
collection and refinement statistics for the apo and inhibitor
complex, NMR materials and methods, more information on
competition and binding site determinations, a comformational
JA101358S
9
15212 J. AM. CHEM. SOC. VOL. 132, NO. 43, 2010