Inhibitors of HCV NS5B polymerase. Part 2: Evaluation of the northern region of (2Z)-2-benzoylamino-3-(4-phenoxy-phenyl)-acrylic acid

Article abstract of DOI:10.1016/j.bmcl.2005.03.106

A novel series of non-nucleoside HCV NS5B polymerase inhibitors was prepared from a (2Z)-2-benzoylamino-3-(4-phenoxy-phenyl)-acrylic acid template. Solution and solid phase analog synthesis focused on the northern region of the template combined with structure based design led to the discovery of several potent and orally bioavailable lead compounds.

Full text of DOI:10.1016/j.bmcl.2005.03.106

Bioorganic & Medicinal Chemistry Letters 15 (2005) 2812–2818
Inhibitors of HCV NS5B polymerase. Part 2:
Evaluation of the northern region of (2Z)-2-benzoylamino-3-
(4-phenoxy-phenyl)-acrylic acid
Jeffrey A. Pfefferkorn,a,* Richard Nugent,b Rebecca J. Gross,c Meredith Greene,c
Mark A. Mitchell,c Matthew T. Reding,d Lee A. Funk,e Rebecca Anderson,c
Peter A. Wells,e John A. Shelly,a Robert Anstadt,c Barry C. Finzel,a Melissa S. Harris,a
Robert E. Kilkuskie,b Laurice A. Koptab and Francis J. Schwendea
aPfizer Global Research & Development, Michigan Laboratories, 2800 Plymouth Road, Ann Arbor, MI 48105, USA
bPfizer Global Research & Development, Research Technology Center, 620 Memorial Drive, Cambridge, MA 02139, USA
cPfizer Global Research & Development, Michigan Laboratories, 301 Henrietta Street, Kalamazoo, MI 49007, USA
dPfizer Global Research & Development, 800 North Lindbergh Blvd., Creve Coeur, MO 63167, USA
ePfizer Global Research & Development, La Jolla Laboratories, 10628 Science Center Dr., San Diego, CA 92121, USA
Received 21 January 2005; revised 21 March 2005; accepted 25 March 2005
Available online 27 April 2005
Abstract—A novel series of non-nucleoside HCV NS5B polymerase inhibitors was prepared from a (2Z)-2-benzoylamino-3-(4-phen-
oxy-phenyl)-acrylic acid template. Solution and solid phase analog synthesis focused on the northern region of the template com-
bined with structure based design led to the discovery of several potent and orally bioavailable lead compounds.
Ó 2005 Elsevier Ltd. All rights reserved.
1. Introduction
Hepatitis C virus (HCV) is the major causative agent of
blood-borne non-A, non-B hepatitis, and it is estimated
to infect over 170 million people worldwide. Chronic
HCV infections have been associated with liver fibrosis,
liver cirrhosis, heptacellular carcinoma, and other forms
of liver dysfunction.1,2 Given the wide spread impact of
this disease, there is a substantial medical need for new
anti-HCV agents to compliment current therapies. In
this and the preceding article, we report on our efforts
to develop orally bioavailable, non-nucleoside inhibitors
of HCV NS5B RNA-dependent RNA polymerase.3
In the preceding report, we described the identification
of
(2Z)-2-(benzoylamino)-3-(5-phenyl-2-furyl)acrylic
acid (1, Fig. 1) as a novel non-nucleoside inhibitor of
HCV NS5B polymerase.3 Optimization of the southern
region of this compound led to analogs 2 and 3 as potent
inhibitors. However, despite the potent activity of these
*
Corresponding author. Tel.: +1 734 622 1471; fax: +1 734 622 3171;
e-mail: jeffrey.a.pfefferkorn@pfizer.com
Figure 1. Previous lead compounds (1–3) and SAR strategy (4).
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.03.106

J. A. Pfefferkorn et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2812–2818
2813
compounds against isolated NS5B polymerase, the com-
pounds exhibited only moderate activity in a cellular
subgenomic replicon assay (2: ED50 = 28 7 lM and
3: ED50 = 60 7 lM).3
Herein, we describe the optimization of the northern re-
gion of this template including modification of the benz-
amide (i.e., C-ring) and carboxylic acid motifs (4, Fig.
1). In undertaking these studies, we hoped to improve
the potency of these inhibitors against NS5B polymerase
and also establish an improved correlation between en-
zyme and cellular replicon activity relative to that ob-
served for previous lead compounds 2 and 3.
2. Chemical synthesis
Analogs for the current studies were prepared using a
combination of solution and solid phase synthetic
routes. Solution phase synthesis of C-ring amide analogs
(4, Fig. 1) was accomplished as outlined in Scheme 1. As
described in the preceding report, N-(Cbz)-a-phospho-
noglycine trimethyl ester 5 was first deprotected by
hydrogenolysis4 to afford an amine that was acylated
with various acid chlorides [R1C(O)Cl] in the presence
of triethylamine to provide, after recrystallization, com-
pounds 6 where R1 was aryl, heteroaryl, or aliphatic.
Intermediate 6 was then condensed with optimally
substituted 4-phenoxybenzaldehydes (7, R2 = Br or I)
to afford an enamide product 8 as a single Z-stereoiso-
mer.5 The synthesis was completed by saponification
of the methyl ester 8 with aqueous LiOH in dioxane to
afford the carboxylic acid 9.
As a compliment to this solution phase method, an anal-
ogous solid phase synthesis was developed to facilitate
rapid synthesis of additional analogs. Strategically, a
route was envisioned that would enable the installation
of multiple elements of diversity on solid support. As
shown in Scheme 2, the loading precursor was prepared
from tert-butyl P,P-dimethylphosphonoacetate 10 by
treatment with O-(diphenyl phosphinyl)hydroxylamine6
in the presence of NaH to affect a-amination.7 The
Scheme 2. Reagents and conditions: (a) NaH, Ph2P(O)ONH2,6 THF,
ꢀ78 ! 25 °C, 12 h, 74%; (b) FMOC-Cl, NaHCO3, dioxane:water,
25 °C, 12 h, 91%; (c) TFA, CH2Cl2, 25 °C, 5 h, 70%; (d) i. Wang resin
(1.0 mmol/g), DIC, DMAP, CH2Cl2, 25 °C, 12 h (92% loading
determined via FMOC analysis); ii. Ac2O, pyridine, CH2Cl2, 25 °C,
0.5 h; (e) piperidine, DMF, 25 °C, 2 h; (f) R1C(O)Cl, diisopropyleth-
ylamine, CH2Cl2, 25 °C, 12 h; (h) DBU, 7, CH2Cl2, 25 °C, 12 h; (g)
TFA, CH2Cl2, 25 °C, 0.5 h.
resulting amine 11 was protected as its FMOC deriva-
tive 12 and the tert-butyl ester hydrolyzed with TFA
to provide carboxylic acid 13 suitable for loading onto
solid support. In that event, a mixture of carboxylic acid
13 and Wang resin was treated with DIC and catalytic
DMAP for 12 h. After the coupling reaction, any unre-
acted loading sites were derivatized with acetic anhy-
dride. The FMOC protecting group was subsequently
removed by treatment with piperidine. The liberated re-
sin-bound amine 14 was derivatized with various acid
chlorides [R1C(O)Cl] in the presence of N,N-diisoprop-
ylethylamine to provide resin-bound amide 16.
Intermediate 16 was then condensed with an optimally
substituted 4-phenoxybenzaldehyde (7, R2 = Br or I) to
afford olefin 17. Finally, the product was cleaved from
solid support by treatment with TFA to afford carbox-
Scheme 1. Reagents and conditions: (a) 10% Pd–C, H2, MeOH, 25 °C,
8 h; (b) R1C(O)Cl, Et3N, CH2Cl2; (c) DBU, CH2Cl2, 25 °C, 2 h; (d)
LiOH (aq), 1,4-dioxane, 25 °C, 2 h.

2814
J. A. Pfefferkorn et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2812–2818
ylic acid 18. After cleavage, the final compounds were
purified by automated LC/MS. A third method for the
solution phase preparation of (2Z)-2-benzoylamino-3-
(4-phenoxy-phenyl)-acrylic acid analogs employed an
Erlenmeyer azalactone strategy as outlined in Scheme
3.8 Hence, a substituted hippuric acid 19 was treated
with 4-(2-bromophenoxy)benzaldehyde, acetic anhy-
dride, and sodium acetate to afford azalactone 20.8 In
a subsequent step, azalactone 20 was hydrolyzed to car-
boxylic acid analogs 21 or treated with an amine
(R4NH2) to afford amide 22. This method was utilized
to prepare selected analogs for which the requisite
substituted hippuric acids were commercially available
or readily prepared.
Several carboxylic acid derivatives were prepared as out-
lined in Scheme 4 where carboxylic acid 23 was con-
verted to the corresponding acid chloride 24 via
treatment with oxayl chloride and catalytic DMF. Acid
chloride 24 was then separately treated with ammonia,
hydrazine, and O-methylhydroxylamine to afford pri-
mary amide 25, hydrazide 26, and N-hydroxyl amide
27, respectively.
3. Results and discussion
To assess the effects of our structural modifications, all
analogs were evaluated against HCV polymerase in a
continuous read picogreenÒ assay using a genotype 1b
CD21 NS5B polymerase enzyme construct.9 Addition-
ally, compounds with IC50 <10 lM were also evaluated
in a cellular subgenomic replicon assay for which ED50
and CC50 values are reported.10 Our structure–activity
discussion will begin by examining the effects of C-ring
(4, Fig. 1) changes and then evaluate the effects of mod-
ifications to the carboxylate.
Analogs containing substituted aryl ring C-rings are pre-
sented in Table 1. The most potent of these analogs had
a benzamide C-ring (i.e., n = 0) with an oxygen substitu-
ent at the R1 position as illustrated by compounds 28
(R1 = OEt, IC50 = 0.05 lM) and 29 (R1 = OH, IC50
=
Table 1. Activity of substituted aryl amide analogs 28–45 against HCV
NS5B polymerase
R1
O
R2
R3
O
H
n
NH
HO
O
Br
Compd R1
R2 R3
n
NS5B IC50 ED50
(lM)
CC50
(lM)
(lM)
Scheme 3. Reagents and conditions: (a) 4-(2-bromophenoxy)benzal-
dehyde, NaOAc, Ac2O, 60 °C, 1 h; (b) NaOH, acetone, 70 °C, 1 h; (c)
R4NH2, EtOH, 70 °C, 2 h.
2
H
H
H
0
0.10
28
7
>100
28
29
30
31
32
33
34
35
36
OEt
OH
H
H
H
F
H
H
H
Cl
Br
H
F
0
0
0
0
0
0
0
0
0
0.05
0.08
0.09
0.11
0.12
0.18
0.23
0.23
0.25
70 >100
>100 >100
53 11 >100
7
H
H
H
H
H
F
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
H
F
F
H
F
Me
H
H
37
38
39
40
41
H
H
H
H
H
H
H
H
H
H
Br
Cl
l
1
1
1
1
1
0.82
0.94
1.00
1.00
2.2
>100
>100
>100
>100
>100
>100
>100
>100
18
CF3
F
0
>100
42
43
44
45
H
H
H
H
H
H
H
H
CF3
Cl
2
2
2
2
0.37
0.50
0.60
2.6
>100
>100
>100
>100
>100
>100
>100
>100
F
Me
Scheme 4. Reagents and conditions: (a) (COCl)2, DMF (cat), CH2Cl2,
45 °C, 2 h; (b) R2R3NH, triethylamine, CH2Cl2, 25 °C, 2 h.

J. A. Pfefferkorn et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2812–2818
2815
0.08 lM). The addition of small, non-polar substituents
to the benzamide ring was tolerated, but did not afford
increased potency relative to the unsubstituted case
(i.e. 2) as illustrated by compounds 30–36. Most other
mono- or poly-substituted benzamide derivatives (not
shown) were less active.
analogs of this type also had lower potency than benz-
amide 2. Finally, as highlighted in Table 1, despite rea-
sonable enzyme potency none of these C-ring aryl
amides (i.e., 28–45) demonstrated improved cellular
activity relative to lead compound 2.
Beyond the aromatic C-ring analogs described above,
various substituted alkyl and cycloalkyl derivatives were
also prepared with representative examples presented in
Table 2 (46–58). In the case of cycloalkyl substituents,
inhibition potency increased with increasing size such
that compound 46 (R1 = cyclopropyl) was least potent
and compound 49 (R1 = adamantyl) was most potent
(IC50 = 0.07 lM). Extended cycloalkyl analogs proved
less potent than their directly bonded counterparts
(e.g., 47 > 52 > 53). The introduction of polar function-
ality including alcohols (54–56) and carboxylic acids (57
and 58) at the C-ring position did not afford improved
activity. As was the case with the aryl amides described
above, most of the alkyl and cycloalkyl amides analogs
Additional C-ring analogs were prepared containing an
ꢀ
aliphatic linker [–CH2 or –ðCH2Þ ] between a substi-
2ꢀ
tuted aromatic ring and the carbonyl of the amide as
highlighted in Table 1. Overall, analogs of this type
proved less potent than their benzamide counterparts.
Examining the phenyl acetyl compounds (n = 1, 37–41)
revealed that substitution at the R3 position with halo-
gens provided optimal relative activity (i.e., compounds
37–39); however, these compounds were approximately
10-fold less active than the simple unsubstituted benz-
amide (2). Similarly, for phenyl ethyl amide compounds
(n = 2, 42–45) substitution of the R3 position with CF3,
Cl, and F provided optimal relative activity, but overall
Table 2. Activity of alkyl and cycloalkyl C-ring amide analogs 46–58

2816
J. A. Pfefferkorn et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2812–2818
Table 3. Activity of selected cyclopentylethyl C-ring analogs
lected to be soaked into crystalline CD21 NS5B such
that inhibitor bound crystal structures could be ob-
tained. The highly potent adamantyl derivative (com-
pound 49) was initially selected and its bound
structure is presented in Figure 2.11 As was observed
in the preceding report,3 this compound binds in the pri-
mer grip site and engages in a hydrogen bond between
its carboxylic acid and the hydroxyl of Ser 556. There
is a second hydrogen bond between the amide carbonyl
˚
to backbone amide of Tyr 448 (N–O distance is 2.7 A).
Compd R1
R2
—
NS5B IC50 ED50
(lM)
CC50 ED50/
(lM) IC50
The phenyl ether motif again occupies a deep hydropho-
bic pocket formed by residues Pro 197, Leu 384, Met
414, Tyr 415, and Tyr 448. Examination of the binding
pocket region surrounding the C-ring reveals a relatively
large cavity explaining the diversity of functionality that
is tolerated at that position. Additionally, the residues
surrounding the C-ring pocket are generally hydropho-
bic thereby explaining the preference for lipophilic sub-
stituents at this position.
(lM)
2
—
0.10
28
7
>50
280
51
59
60
61
62
63
I
H
H
H
H
F
7.0
15
12
5
4
5
2
0
9
2
>50
>100
>100
>100
>100
>50
2
2
Br
CF3
Et
Br
6.0a
6.3a
7.1
0.7
2
15
12
5
15a
0.8
0.4
CF3 NH2 14a
a IC50 values determined on the sodium salt of the carboxylic acid due
to solubility considerations.
We next attempted to obtain a crystal structure of com-
pound 51 (Table 3) bound to NS5B polymerase. The
reader will recall that this compound, containing a
cyclopentylethyl C-ring, exhibited consistent replicon
activity. Despite substantial efforts, we were unable to
obtain a crystal structure of 51 (or several related com-
pounds from Table 3) bound to NS5B polymerase. Intri-
gued by these results and seeking to better understand
the interaction of this analog with the polymerase, resis-
tance selection studies were undertaken in the replicon
system with compound 51. These studies revealed a
mutation (M423 ! T423) that rendered inhibitor 51
inactive in the cellular assay. By comparison, lead com-
pound 2 retained its activity in this resistant (M423T)
replicon line.12 Interestingly, M423 is not a part of the
primer grip binding site previously described (Fig. 2).
Based on the literature reports, M423 is located in the
allosteric binding site of the polymerase.13,14 Based on
these data and the fact that the SAR data surrounding
compound 51 are inconsistent with those of the other
analogs from this template, we are proposing that intro-
duction of the cyclopentylethyl group into this analog
resulted in its ability to interact with NS5B polymerase
at the allosteric site. Follow up studies on this poten-
tially new series of inhibitors will be reported
separately.15
of Table 2 did not have improved cellular activity (rela-
tive to compound 2) with the notable exception of com-
pound 51.
Interestingly, analog 51 (Table 2) exhibited consistent
replicon activity (ED50 = 15 4 lM). A structure–activ-
ity comparison of 51 versus 50 or 52 suggested that this
replicon activity was rather anomalous; nevertheless,
additional analogs of compound 51 were prepared as
highlighted in Table 3. Examination of these analogs re-
vealed that while only moderately potent against NS5B,
these inhibitors had consistent activity in the replicon as-
say. Compounds 60 and 63 exhibited the best activities
(ED50 = 5 2 lM for both). Moreover, the ratios of cel-
lular to enzyme activity (i.e., ED50/IC50) were improved
approaching unity for compounds 51 and 59–63,
whereas the same ratio for compound 2 was several hun-
dreds reflecting its substantial differential in enzyme ver-
sus cellular activity.
In order to better understand the activity of the cyclo-
alkyl amides of Table 2, several compounds were se-
The final component of our SAR program on the (2Z)-
2-benzoylamino-3-(4-phenoxy-phenyl)-acrylic acid tem-
plate involved modification of the carboxylic acid.
Given that the carboxylate motif might be contributing
to poor membrane permeability and/or protein binding,
it was anticipated that modification as an amide might
lead to improved cellular activity for compounds bound
to the primer grip site. As highlighted in Table 4, various
amide derivatives were prepared.
Examination of Table 4 illustrates that primary amide
25, hydrazide 26, and N-methoxyamide 27 were less ac-
tive than the parent carboxylic acid 2. Encouragingly,
ethanolamide 64 while having modest enzyme potency
(IC50 = 2.7 lM) demonstrated reproducible replicon
activity (ED50 = 8 2 lM). X-ray studies on a related
˚
Figure 2. X-ray crystal structure (2.8 A resolution) of compound 49
(Table 2) bound to the Primer Grip site of CD21 NS5B polymerase.

J. A. Pfefferkorn et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2812–2818
Table 4. Activity of analogs containing modifications to the carboxylate
2817
O
O
H
NH
R1
O
Br
Compound
R1
NS5B IC50 (lM)
ED50 (lM)
CC50 (lM)
ED50/IC50
2
OH
0.10
28
7
7
>50
280
25
26
27
NH2
NHNH2
NHOMe
0.5
1.0
0.8
20
45
7
—
—
—
>50
>50
>50
22
64
65
NH(CH2)2OH
NHCH2CO2H
2.7
8
2
>50
>50
3
0.06
>50
—
66
67
68
69
70
71
NHCH2Ph
NHPh
>100
>100
12
>50
>50
>50
15
>50
>50
22
—
—
—
—
—
—
NH[(4-OH)–Ph]
NH[(3-OH)–Ph]
NH[(2-OH)–Ph]
NH[(2-CO2H)–Ph]
0
0
1.9
15
0
25
>50
>50
>50
>50
0.4
ethanolamide analog (not shown) confirmed that com-
pounds of this type do in fact bind at the primer grip site
of NS5B in a manner identical to that previously de-
scribed for compound 49 (Fig. 2). In the case of the eth-
anolamide containing analogs, the position of Ser 556 is
perturbed slightly to accommodate a hydrogen bond to
the terminal alcohol of the amide. Preparation of the
carboxylic acid equivalent of compound 64 (i.e., glyci-
mide 65) afforded a dramatic increase in enzyme potency
(IC50 = 0.06 lM) with a concomitant reduction in cellu-
lar activity (ED50 >50 lM). Simple amides such as ani-
lide 66 or benzyl amide 67 were less active; however,
installation of hydroxyl or carboxylic acid functional
group onto the anilide (i.e., compounds 68–71) restored
activity against NS5B polymerase.
Table 5. Pharmacokinetic properties of selected compounds
Compound
%F
Cmax (lM)
t1/2 (h)
CL (Lh/kg)
2
64
21
76
9
1.5
1.5
1.0
0.20
3.2
parent carboxylic acid 2, and it also had higher rates
of clearance and a slightly lower half-life.
In conclusion, we have reported on the structure–activ-
ity evaluation of a novel series of non-nucleoside inhib-
itors of HCV NS5B polymerase. In the preceding report,
we described as to how the modification of the southern
region of (2Z)-2-(benzoylamino)-3-(5-phenyl-2-furyl)-
acrylic acid afforded potent inhibitors such as com-
pound 2 that interact with NS5B polymerase at the
primer grip site as determined by X-ray crystallography.
The modest cellular efficacy of such compounds
prompted us to undertake modifications to the northern
region of the template. These efforts resulted in the iden-
tification of 64 as a lead compound with reasonable
activity and PK properties; additional optimization
studies on this molecule will be reported in due course.
Finally, this work led to the serendipitous identification
of 51 and related analogs as novel polymerase inhibitors
that may interact with NS5B at the allosteric site.
Taken together, these results (along with the previously
described X-ray structure, Fig. 2) indicate that analogs
of this template require a functional group capable
hydrogen bonding at the carboxylate position. While
carboxylic acids at this position afford high inhibitory
potency (as evidenced by 2 and 65), their potentially
poor membrane permeability and/or serum protein
binding properties prevent them from achieving sub-
stantial efficacy in a cellular assay. Encouragingly, the
activity of 64 suggests that it may be possible to satisfy
the hydrogen bonding requirements of this position
using functional groups that offer improved permeabil-
ity and protein binding properties.16
Acknowledgements
Finally, in order to evaluate the potential of compounds
such as 64 as orally administered anti-HCV agents, a
pharmacokinetic profile of the compound was obtained
(Table 5). Compound 64 was dosed to rats (Sprague-
Dawley) at 5 mpk. As illustrated, amide 64 demon-
strated improved oral bioavailabilities relative to the
We would like to thank Stephen Secreast for formula-
tion of selected compounds for pharmacokinetic studies.
We would also like to acknowledge all members of the
former Pharmacia HCV team for helpful discussions.
Finally, we would like to thank our former collabora-

2818
J. A. Pfefferkorn et al. / Bioorg. Med. Chem. Lett. 15 (2005) 2812–2818
mann, L.; Bartenschlager, R. Science 1999, 285, 110, EC50
values are reported as arithmetic mean along with
standard deviation for 3–10 independent measurements.
11. For crystallization conditions, see: Finzel, B. C. PCT Int.
Appl. WO 2005001417, 2005. Coordinates for structure
have been deposited at PDB under file name 1Z4U. For a
discussion of the potential biochemical relevance of this
binding site, see Ref. 3.
tors at Chiron Corporation for providing us with CD21
HCV NS5B.
References and notes
1. For a review of HCV epidemiology, pathogenesis, and
clinical treatment, see: (a) Lauer, G. M.; Walker, B. D. N.
Engl. J. Med. 2001, 345, 41; (b) Di Bisceglie, A. M. Lancet
1998, 351, 351.
12. Efforts to similarly generate a replicon line resistant to
compound 2 were not successful.
13. For
a description of the HCV allosteric site: (a)
2. For a clinical overview of HCV, see: Strader, D. B.; Seeff,
L. B. ILAR J. 2001, 42, 107.
Love, R. A.; Parge, H. E.; Yu, H.; Hickey, M. J.; Diehl,
W.; Gao, W.; Wriggers, H.; Ekker, A.; Wang, L.;
Thomson, J. A.; Dragovich, P. S.; Fuhrman, S. A. J.
Virol. 2003, 77, 7575; (b) 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.
3. Pfefferkorn, J. A.; Greene, M. L.; Nugent, R. A.; Gross,
R. J.; Mitchell, M. A.; Finzel, B. C.; Harris, M. S.; Wells,
P. A.; Shelly, J. A.; Anstadt, R. A.; Kilkuskie, R. E.;
Kopta, L. A.; Schwende, F. J. Bioorg. Med. Chem. Lett.
2005, 15, 2481.
4. Horenstein, B. A.; Nakanishi, K. J. Am. Chem. Soc. 1989,
111, 6242.
14. At the present time, we do not have an explanation for our
inability to soak compound 51 into the allosteric site of
NS5B and solve the structure. A possible explanation is
that our crystal packing form (Ref. 11) constricts the
allosteric site.
15. Competitive binding studies between a labeled variant of
compound 51 and a published allosteric inhibitor (Ref.
13b) supported the conclusion that 51 interacts at the
allosteric site. As a control, a labeled variant of compound
2 did not compete for the allosteric site. Details of these
studies along with additional SAR around compound 51
will be reported separately.
16. Preliminary studies confirmed that analogs bearing car-
boxylic acids were highly protein bound and had low
membrane permeability. For example, compound 2 had
Kd = 0.35 lM for human serum albumin (HSA, fraction
V) and low Caco-2 permeability. In contrast, compound
64 had Kd = 10 lM for HSA and high Caco-2
permeability.
5. Schmidt, U.; Griesser, H.; Leitenberger, V.; Lieberknecht,
A.; Mangold, R.; Meyer, R.; Riedl, B. Synthesis 1992, 487.
6. Colvin, E. W.; Kirby, G. W.; Wilson, A. C. Tetrahedron
Lett. 1982, 23, 3835.
7. Stocksdale, M. G.; Ramurthy, S.; Miller, M. J. J. Org.
Chem. 1998, 63, 1221.
8. (a) Herbst, R. M.; Shemin, D. Org. Synth. Coll. Vol. II
1943, 1; (b) Shen, L. M. C.; Hartung, W. H. J. Org. Chem.
1958, 23, 96; (c) Cativiela, C.; Diaz-de-Villegas, M. D.;
Galvez, J. A. Tetrahedron Lett. 1999, 40, 1027.
9. All biological data were generated on analytically pure (1H
NMR, MS, and CHN) samples. IC50 values are reported
as the arithmetic mean for 2–5 independent measurements.
For complete assay details, see: Yagi, Y.; Sheets, M.;
Wells, P.; Shelly, J. A.; Poorman, R. A.; Epps, D. E. PCT
Int. Appl. WO 2004044228, 2004.
10. Cellular replicon assay was performed in Huh-7 cells in
10% fetal calf serum following general procedure: Loh-
mann, V.; Ko¨rner, F.; Koch, J.-O.; Herian, U.; Theil-