HCV NS5A replication complex inhibitors. Part 5: Discovery of potent and pan-genotypic glycinamide cap derivatives

Article abstract of DOI:10.1016/j.bmcl.2013.05.040

The isoquinolinamide series of HCV NS5A inhibitors exemplified by compounds 2b and 2c provided the first dual genotype-1a/1b (GT-1a/1b) inhibitor class that demonstrated a significant improvement in potency toward GT-1a replicons compared to that of the initial program lead, stilbene 2a. Structure-activity relationship (SAR) studies that uncovered an alternate phenylglycine-based cap series that exhibit further improvements in virology profile, along with some insights into the pharmacophoric elements associated with the GT-1a potency, are described.

Full text of DOI:10.1016/j.bmcl.2013.05.040

Bioorganic & Medicinal Chemistry Letters 23 (2013) 4428–4435
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
HCV NS5A replication complex inhibitors. Part 5: Discovery of potent
and pan-genotypic glycinamide cap derivatives
a
a
a
a
Makonen Belema a, , Van N. Nguyen , Denis R. St. Laurent , Omar D. Lopez , Yuping Qiu ,
Andrew C. Good c, Peter T. Nower b, Lourdes Valera b, Donald R. O’Boyle II b, Jin-Hua Sun b,
Mengping Liu b, Robert A. Fridell b, Julie A. Lemm b, Min Gao b, Jay O. Knipe d, Nicholas A. Meanwell a,
Lawrence B. Snyder a
⇑
a Department of Medicinal Chemistry, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
b Department of Virology, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
c Department of Computer-Aided Drug Design, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
d Department of Metabolism and Pharmacokinetics, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The isoquinolinamide series of HCV NS5A inhibitors exemplified by compounds 2b and 2c provided the
first dual genotype-1a/1b (GT-1a/1b) inhibitor class that demonstrated a significant improvement in
potency toward GT-1a replicons compared to that of the initial program lead, stilbene 2a. Structure–
activity relationship (SAR) studies that uncovered an alternate phenylglycine-based cap series that exhi-
bit further improvements in virology profile, along with some insights into the pharmacophoric elements
associated with the GT-1a potency, are described.
Received 12 April 2013
Accepted 13 May 2013
Available online 23 May 2013
Keywords:
Daclatasvir
HCV
Ó 2013 Elsevier Ltd. All rights reserved.
NS5A
Mandelamide X-ray
Pan-genotype HCV replicon inhibition
Phenylglycine
Hepatitis C virus (HCV) infection is responsible for causing seri-
ous liver diseases, including cirrhosis, hepatocellular carcinoma
and liver failure. It is estimated that HCV has infected over 180 mil-
lion people worldwide and, according to a recent survey conducted
in the United States, about 50% of these individuals might be una-
ware of their infection.1 The significant heterogeneity of the viral
genome—six genotypes and over 100 subtypes have been identi-
fied thus far—coupled with its high mutation rate have hampered
the development of effective therapies.2 A combination of pegylat-
development of more effective, tolerable and interferon-free ther-
apies to cure HCV infection that rely upon direct-acting antiviral
agents is an ongoing enterprise, with promising initial results.
These regimens have examined combinations of inhibitors of NS3
protease, NS5B polymerase and NS5A.5
Clinical validation of the HCV NS5A protein as a viable target for
antiviral therapy was established with trials of daclatasvir (1).6 We
have previously described early aspects of the medicinal chemistry
effort that generated the NS5A inhibitor lead, stilbene 2a, and the
preliminary SAR studies that delineated some of the chemotype’s
pharmacophoric elements, including the identification of an iso-
quinoline class of derivatives that significantly enhanced its inhib-
itory potency toward a GT-1a replicon (compare isoquinolinamide
2c vs phenyl acetamide 2a).7 Although this improvement in GT-1a
inhibitory potency marked an important milestone for the pro-
gram, the polyaromatic makeup of the isoquinolinamide cap ser-
ies—as reflected by the low percent composition of sp3-carbons
for 2c (22%) compared to that of marketed drugs (47%)—was of
some concern, providing an impetus for additional structural
exploration.8 Herein, we report the SAR investigation that uncov-
ered a phenylglycine-based cap series exhibiting high GT-1a
and -1b inhibitory potency in in vitro replicon assays, a key
development in the optimization campaign that culminated in
the discovery of daclatasvir (1).
ed interferon-a (PEG-IFN-a) and ribavirin (RBV), which has served
as the standard of care for HCV for the past decade, has many
drawbacks, including poor tolerability and non-optimal efficacy to-
ward GT-1, the most dominant genotype across the world.3 In May
2011, two first-generation protease inhibitors that enhanced the
efficacy of the PEG-IFN-
a
/RBV regimen from ꢀ45% to ꢀ70% in
treatment-naïve GT-1 patients, while decreasing the therapy dura-
tion by up to 50%, were approved.4 Despite this marked progress,
however, the new triple combinations exhibit less than optimal
efficacy in treatment-experienced subjects and are associated with
more side effects than the original PEG-IFN-a/RBV regimen. The
⇑
Corresponding author. Tel.: +1 203 677 6928.
E-mail address: makonen.belema@bms.com (M. Belema).
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.05.040

M. Belema et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4428–4435
4429
the analogues discussed thus far indicated favorable properties.
For example, 3g.2 had a reduced activity in the GT-1b LV/YH resis-
tant replicon (EC50 = 620 nM), minimal activity in the BVDV assay,
and an excellent selectivity index.
MeO2CHN
H
N
O
N
N
N
N
N
H
O
NHCO2Me
It was not initially clear to us how the benzylic substituents
were influencing replicon potency. Although it was tempting to
hypothesize that the hydroxyl group of the mandelamide moiety
might be involved in some form of electrostatic interaction with
the proximal carbonyl group and, thus, perhaps populating a pre-
ferred conformational state, rationalization of the stereochemical
disparity between the relatively more active diastereomers (S)-
3g.2 and (R)-3b.1 was not feasible.11 Alternatively, the similar
GT-1a activities between alcohols 3b.1/3b.2 and methyl analogue
3d.1 in the GT-1a replicon indicated that a hydrogen-bond interac-
tion between the benzylic region and the NS5A protein was unli-
kely to be the sole driving factor for the observed inhibitory
effects.7b To gain additional insight into these SAR findings, the sin-
gle crystal X-ray structures of bromides 4a, (S)-4b, and (R)-4c,
which represent the monomeric structural fragments of 2a, (S)-
3g.2, and (R)-3b.1, respectively, were determined. Although fully
cognizant of the limitations that such an approach may have in
illuminating aspects of a dynamic process, interestingly this exer-
cise revealed a similar dihedral angle between the cap carbonyl
moiety and the methyl group of (S)-4b (4–22°) or the alcohol group
(R)-4c (4–16°), resulting in similar topological dispositions for their
respective phenyl groups and yet clearly different from that of 4a,
the substructure of the least GT-1a active analogue of the set (i.e.,
compound 2a) (see Fig. 2).12,13
At this juncture, it became apparent that the alkene linker of the
stilbene chemotype was susceptible to trans–cis isomerization
when compounds were stored as DMSO solutions under standard
fluorescent lighting.7b In order to obviate the complications that
such isomerization could create in building reliable SAR, the corre-
sponding alkyne-linked template was considered. Surveying a
select set of caps uncovered in the stilbene series on an alkyne-
linked core revealed a similar stereochemical preference, accompa-
nied with comparable or slightly decreased GT-1a inhibitory
potency (compare 3g vs 3h in Table 2; data for remaining subset
is not shown). Thus, the alkyne template was adopted for the next
phase of the study, the objective of which was to enhance GT-1a
potency further by examining glycine-based caps. For the majority
of the cases, both symmetrical diastereomeric analogues (i.e., ana-
logues with identical cap stereoconfigurations) were prepared and
assayed in a fashion similar to the SAR investigation illustrated in
Table 2. For each pair of symmetrical diastereomers tested, one
was consistently more active than its alternate isomer toward
GT-1a and/or GT-1b replicons in all of the cases where such a com-
parison could be made. In addition, at least for the arylglycinamide
subset where the absolute stereochemistry of the cap was known,
the (R)-isomer exhibited the better inhibitory potency (e.g., com-
pare 5b.1 vs 5b.2 in Table 3).
Daclatasvir (1)
Homodimeric cap analogues were prepared and screened for
inhibitory activity in GT-1a and -1b replicons and for cellular tox-
icity and target specificity in GT-1b and bovine viral diarrhea virus
(BVDV) replicons, respectively. On a selective basis, target specific-
ity was also assessed for some compounds in a GT-1b L31V/Y93H
double mutant resistant replicon, henceforth referred to as GT-1b
LV/YH resistant replicon.9
The design concept envisioned amides represented by the gen-
eric structure A in Figure 1 as a deannelated isostere of the isoquin-
olinamide 2b, which facilitated further SAR evolution by variation
of the benzylic substituent. The first analogue prepared in this
exercise, ketoamide 3a, exhibited a GT-1a inhibitory potency that
was comparable with that of 2b (see Table 2). Replacing the
benzylic carbonyl moiety of 3a with small functional groups had
neutral to detrimental impact on GT-1a inhibitory potency (see
3b–3d). In comparing the GT-1a/-1b activities of the diastereo-
meric pairs 3b.1 verses 3b.2 or 3d.1 verses 3d.2, the (R)-stereoiso-
mer appears to correlate with better GT-1a and/or GT-1b inhibitory
potency, although the magnitude of the effect was not consistently
compelling (e.g., the GT-1a EC50s of 3b.1 and 3b.2 were within the
margins of error of the assay, and the enhanced GT-1a potency of
3d.1 over that of its diastereomer 3d.2 was accompanied with
some increase in non-specific signal in the BVDV assay).10 Satura-
tion or homologation of the mandelamide cap generally resulted in
reduced potency, albeit it is noteworthy that in both cases the (R)-
stereoisomer was relatively more potent in the GT-1b replicon
(compare 3e.1 vs 3e.2 and 3f.1 vs 3f.2). A marked improvement
in GT-1a potency was secured with the introduction of a methyl
group at the benzylic position of the mandelamide (EC50 of
3g.2 = 84 nM), which was also accompanied by single digit picom-
olar inhibitory potency in the GT-1b replicon. Interestingly, con-
trary to the initial SAR trend, it was the (S)-stereoisomer that
was the more active analogue toward the GT-1a subtype for the
methyl carbinol. In general, assessment of target specificity for
N
N
O
O
O
O
N
X
2b
A
Figure 1. Deannelation design strategy.
Table 1
Activity and specificity of selected early leadsa
O
H
N
Ph
O
N
N
O
O
O
O
N
H
Ph
N
N
O
2a
2c
19
2b
Cl
GT-1a EC50 (nM)
GT-1b EC50 (nM)
GT-1b CC50 (nM)
BVDV EC50 (nM)
>104
0.086
>104
>104
390
22
7.1
>104
>104
>104
>104
a
Data represent mean values of at least two experiments with a maximum of threefold variation.

4430
M. Belema et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4428–4435
Table 2
Activity and specificity of mandelamides and related analoguesa,b
H
N
N
O
O
R
R
O
O
N
N
H
3h
3a 3g
-
Compd
R
Replicon activity
GT-1b CC50 (nM)
BVDV EC50 (nM)
GT-1a EC50 (nM)
GT-1b EC50 (nM)
O
3a
447
7.0, <4.6
>104
>104
>104
>104
>104
>104
Ph
Ph
Ph
Ph
OH
OH
OMe
OMe
3b.1
3b.2
3c.1
3c.2
640
1020
1290
1270
0.022
3.8, < 4.6
<4.6
>104
5110, >104
>104
<4.6
>104
Ph
Ph
Ph
3d.1c
3d.2c
772
<4.6
5.0
>104
>104
2500
>104
>104
OH
OH
3e.1
3e.2
3f.1
4650
970
<4.6
>104
>104
>104
>104
6740
7760
C-Hexyl
63
C-Hexyl
OH
1860
5.2, <4.6
Ph
OH
3f.2d
3g.1
3g.2
3h.1
3h.2
2320
6850
84
229
980, >104
>104
6390
NTe
Ph
OH
<4.6
0.006
<4.6
<4.6
Ph
HO
>104
>104
>104
>104
Ph
OH
6510
284
8380
Ph
HO
Ph
>104
a
b
c
Data represent mean values of at least two experiments with a maximum of threefold variation.
EC50 in GT-1b LV/YH resistant replicon: 3b.1 (2030 nM), 3g.2 (620 nM).
The inhibitory activities of 3d.1 and 3d.2 in GT-1b replicon were disclosed previously in Ref. 7b.
Sample was not retested to check on the difference in CC50 value between the two experiments.
NT, not tested.
d
e
The parental phenylglycine analogue 5a.1 exhibited a GT-1a
Although the benzylic position initially appeared to exhibit reason-
able tolerance for structural diversity, there was an unexpected
reduction in GT-1a inhibitory potency for the relatively less basic
morpholine analogue 5h (see Table 4 for measured pKa values for
a selected set of analogues). Equally noteworthy was the fact that
analogues containing less basic piperazine appendages, such as 5j
and 5k, also exhibited decreased activity toward the GT-1a
inhibitory activity that was comparable to that of the mandela-
mide 3b.1, and introduction of a simple dimethyl group (as in
5b.1) enhanced the GT-1a inhibitory potency by nearly a 100-fold.
Further structural refinement of this region afforded an additional
ꢀ10-fold potency gain toward the GT-1a replicon, resulting in a
low single digit nanomolar EC50 for compounds 5f, 5g.1 and 5i.1.

M. Belema et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4428–4435
4431
tion, homologation of the benzyl moiety of 5b.1 to afford 5m.1 re-
sulted in over a 100-fold potency loss toward the GT-1a replicon.
Some basic appendages are associated with relatively weak
cytotoxicity in the replicon assay and this may have contributed
to the observed weak BVDV inhibitory activities.14 Considering
that potent analogues have been prepared in this series with min-
imal cytotoxic signals (e.g., 5g.1: GT-1a EC50 = 0.82 nM, GT-1b
<0.13 nM, and CC50 >10 lM), it is unlikely that generic cytotoxicity
is making a meaningful contribution to the observed antiviral
activity. Furthermore, the potency of piperidine 5g.1 is >90-fold
weaker toward the GT-1b LV/YH resistant mutant than the GT-1b
wild type strain (EC50 = 12 nM vs <0.13 nM), which is consistent
with a mode of action targeting the NS5A protein.
Concurrent with the above SAR surveys, non-basic phenylgly-
cine cap derivatives with H-bond donating capability were also
investigated and a diverse set of analogues were identified that
exhibited sub-100 nM GT-1a EC50s and picomolar GT-1b EC50s, as
illustrated in Table 5. The stereochemical preference was deter-
mined for the pair of compound 6a.1 and 6a.2 and, in line with
the basic phenylglycine series discussed above, the (R)-stereoiso-
mer was the more active. Despite the similar stereochemical
requirement, however, there appeared to be a decreased steric tol-
erance around the benzylic region for this series, as indicated by
the ꢀ9-fold GT-1a potency difference between Boc analogue 6a.1
and its methyl carbamate variant 6e. It is noteworthy that methyl-
ation of the acetamide group in 6g, to afford 6h.1 or 6h.2, resulted
in a significant erosion in both GT-1a and GT-1b inhibitory activi-
ties, highlighting the likely role of the NH moiety of the cap in en-
abling favorable interactions with the NS5A protein. This
observation is in concordance with the deleterious effect that tem-
pering the basicity of the phenylglycine-based caps had on GT-1a
activity. Although the non-basic glycinamide derivatives illus-
trated in Table 5 generally had weaker GT-1a potency than the ba-
sic family compiled in Table 3—the pKa of a subset of which is
noted in Table 4—they did exhibit an improved cytotoxicity
profile.14
O
N
HN
Br
O
R2
R1
4a (R1 = H, R2 = H)
(R1 = OH, R2 = CH3)
(R1 = H, R2 = OH)
4b
4c
Figure. 2. The X-ray structures of prolinamides 4a–c.
Pharmacokinetic assessment of the selected set of compounds
3g.2, 5i.1 and 6g in a 4 h rat screen indicated poor systemic expo-
sure after oral dosing (see Table 6). Since all three analogues dem-
onstrated very good stability in rat liver microsomes (100%
remaining after 10 min of incubation) and that two of them had
moderate to low clearance after IV dosing, it is believed that the
lack of oral exposure most likely reflects low absorption due to
poor intestinal permeability.15
replicon when compared with the parental analogue 5i.1. Although
it is possible that the decreased potency of these latter two ana-
logues could be due to unfavorable interactions between their
acetyl (5j) or oxa (5k) functional groups and the NS5A protein,
the reduced potency associated with morpholine 5h could not be
rationalized similarly considering the relatively high inhibitory po-
tency exhibited by a range of functionally diverse analogues,
including dimethylamine 5b.1, hydroxypiperidine 5g.1 and piper-
azine 5i.1. In light of these observations, we postulate that caps
with benzylic substituents that are protonated at physiological
pH could be involved in a productive H-bond interaction with
the NS5A protein, particularly for the GT-1a subtype and, hence,
modifications that attenuate their basicity would negatively im-
pact such an interaction, translating into a weaker inhibitory effect.
It should be noted, however, that the ability of a compound to en-
gage in this type of H-bond interaction does not necessarily guar-
antee a desirable outcome as reflected in the weaker potency of
pyrrolidine 5d, which is actually more basic than the more potent
piperidine 5g.1. Moreover, although the alcohol group of mandela-
mide 3g.2 might be involved in a similar type of H-bond interac-
tion with the NS5A protein, in this scenario the vectorial
disposition of its phenyl group would likely be different from that
of the potent phenylglycine isomers in light of their differing
stereoconfigurations.
We reported previously that even though the isoquinolinamide
cap enhanced the GT-1a potency of the stilbene chemotype, the
outcome with other strains was mixed, with enhanced potency
for some genotypes (GT-2a NIH) and reduced potency for others
(GT-1b & GT-4a, compare 2c vs 2a in Tables 1 and 7).7c On the other
hand, compound 5g.1 exhibited an excellent pan-genotypic inhib-
itory effect that was significantly improved over that of the proto-
type 2c (see Table 7). In addition, 5g.1 showed encouraging
potency toward the GT-1b LV/YH resistant replicon (EC50 = 12 nM),
representing a >800-fold improvement over isoquinolinamide 2b
(EC50 >10 l
M).16,17 The virological profile of 5g.1 was critical to
the program because it demonstrated that potent and
pan-genotypic antiviral activity accompanied by favorable potency
toward resistant phenotypes, at least for GT-1b, could be achieved
in replicons by appropriate modification of the peripheral regions
of the lead chemotype. Beside the virology gain, the phenylglycina-
mide series attained an improved sp3-carbon percent composition
compared to the original isoquinolinamide series (e.g., 40% for 5g.1
vs 22% for 2c).8 Finally, while certain aspects of the SAR were
discrete and that GT-1a activity was more sensitive than GT-1b
activity to the structural and stereochemical factors considered,
it became apparent that the locus in the NS5A protein where these
An attempt to replicate the favorable effect that the introduc-
tion of an
a-methyl group had on the activity of mandelamide
3g.2 in the phenylglycinamide series proved detrimental to both
GT-1a activity and HCV specificity (compare 5b.1 vs 5l.1). In addi-

4432
M. Belema et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4428–4435
Table 3
Activity and specificity of phenylglycinamide analoguesa,b
O
O
O
N
R
N
R
HN
NH
O
Ph
Me
Ph
O
O
R
R
Ph
Ph
5l
5m
)
5a 5k
- )
(
)
(
(
Compd
R (stereochemistry)
Replicon activity
GT-1b CC50 (nM)
BVDV EC50 (nM)
GT-1a EC50 (nM)
GT-1b EC50 (nM)
NH2
(R)
NH2 (S)
NMe2
5a.1
5a.2
5b.1
5b.2
5c
910
<4.6
9800, >104
>104
3760
>104
1900
1200
2790
>9400
<4.6; 13
628
<4.6; 13
<4.6
(R)
5390
NMe2 (S)
<4.6
2300
NMeEt
7.9
<0.13; 0.13
4750
N
5d
5e
19
<4.6
0.5
3170
>104
1380
>104
(R)
OH
3.0
N
OH
5f
1.4
0.094
>104
3280
N
N
N
N
N
N
N
5g.1c
5g.2
5h
0.82
600
125
<4.6
2700
37
<0.13
>104
5000
>104
6020
2840
>104
6980
2970
>104
OH
OH
<4.6, 9.8
<4.6
O
N
N
(R)
5i.1
5i.2
5j
0.152
>3330
1360
>104
<4.6; 4.4
0.42
N
O
Ac
5k
388
<4.6
>104
NTd
N
NH
NMe2
NMe2
NMe2
5l.1
100
<4.6
<4.6
<4.6
500
3110
2130
>104
5820
697
5l.2
1080
1290
3240
769
(R)
5m.1
5m.2
1730
1490
NMe2 (S)
a
Except for 5d and 5h, both symmetrical diastereomers were prepared for all of the indicated analogues and that for a subset of the analogues (5c, 5e, 5f, 5j and 5k) only
the data for the more active stereoisomer are included. Where known, the stereochemistry of the benzylic center is noted in parentheses. Compounds 5g.2, 5i.2 and 5l.2 are
the alternate symmetrical diastereomers of 5g.1, 5i.1 and 5l.1, respectively.
b
Data represent mean values of at least two experiments with a maximum of threefold variation.
EC50 in GT-1b LV/YH resistant replicon = 12 nM.
NT, not tested.
c
d
Table 4
pKa of conjugate acida
Compounds were prepared as single stereoisomers in one of
two ways.18 In the first approach, pyrrolidines 8x or 8y, readily
Compd
5d
5g.1
5h
5k
synthesized from the commercially-available anilines 7x and 7y,
respectively, were coupled with the single stereoisomers of the
cap acid precursors to afford final products, as summarized in
Scheme 1. Deprotection of carbamate 6a afforded amine 5a,
which was elaborated further under standard acylation, sulfony-
lation or reductive amination protocols to provide additional ana-
logues.19 Compound 5m was prepared from N-Boc-phenylalanine
by adoption of the procedure outlined for 5b. When the requisite
acid precursors were not readily available in optically pure form,
pKa-1
pKa-2
8.6
8.1
7.6
6.8
5.7
4.6
3.4
2.5
a
pKa was measured in pH range 2–11 using potentiometric titration in varying
MeOH/H2O mixture.
inhibitors interact accommodates a wider range of functionalities,
an early and encouraging indication of the chemotype’s potential
for further optimization.

M. Belema et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4428–4435
4433
Table 5
a second approach was implemented where the diastereomeric
mixture exemplified by bromide 10 was prepared from the race-
mic acid, separated by reverse phase preparative HPLC methodol-
ogy and homocoupled to the alkyne linker under Stille conditions
to afford the two symmetrical diastereomers of the final products
(see Scheme 2). Compounds 6b–6e were prepared from the
appropriately derivatized (R)-phenylglycine under the homocou-
pling conditions shown in Scheme 2.20 The acid precursors uti-
lized in both routes were readily obtained either from
commercial sources or by adoption of well-established literature
protocols, as exemplified in Scheme 3. Assessment of a select set
of the monomeric synthetic precursors noted in Scheme 2 in the
replicon assay indicated significantly reduced antiviral activity.
For example, of the two diastereomeric bromides 10 that were
precursors to piperazine 5i.1 and 5i.2, one exhibited GT-1a/-1b
Activity and specificity of non-basic phenylglycinamide analoguesa,b,c
O
O
N
N
N
N
HN
NH
O
O
Ph
Ph
R2
Replicon activity
R2
R1
R1
Compd
R1NR2
GT-1a EC50 (nM)
GT-1b EC50 (nM)
6a.1
6a.2
6b
(R) BocNH
(S) BocNH
245
>104
137
88
114
26
70
24
2450
>5000
71
0.002, <0.046
<4.6
0.28
0.16
0.018
(R) i-PrOCONH
(R) n-PrOCONH
(R) EtOCONH
(R) MeOCONH
(R) i-PrCONH
(R) AcNH
AcNMe
AcNMe
(R) MeSO2NH
6c
6d
6ed
6f
0.013
EC50s of >10
l
M/0.45
lM while the second had no meaningful
0.002, <0.046
0.005, <0.046
57
632
<4.6
activity toward either genotypes, EC50s >10 lM. It is noteworthy
6g
that the precursor that exhibited measurable activity toward the
GT-1b replicon is the one that afforded piperazine 5i.1, the more
potent of the diastereomeric pair, suggesting discriminatory
interaction between the monomeric-pharmacophore element
and the NS5A protein.
6h.1e
6h.2e
6i
a
Where known, the stereoconfiguration of the benzylic center is noted in
parentheses.
In summary, an SAR campaign that was implemented to ad-
dress the potential structural liabilities of an early GT-1a/-1b
inhibiting lead culminated in the discovery of a new class of
phenylglycine cap derivatives that exhibited potent inhibitory
activity toward both GT-1a and GT-1b HCV replicons, with EC50
reaching single digit picomolar for GT-1b and subnanomolar for
GT-1a. In general, these compounds were relatively more active
toward GT-1b than GT-1a replicons and that the GT-1a inhibitory
activity was considerably more sensitive to the structural modi-
fications examined. Whereas significant steric tolerance was ob-
served for the benzylic position of the glycinamide series,
securing high inhibitory potency particularly toward GT-1a re-
quired a benzylic substituent with H-bonding capability and
(R)-stereoconfiguration. Supportive evidence for this pharmaco-
phore hypothesis was garnered, in part, from directional correla-
b
Data represent mean values of at least two experiments with a maximum of
threefold variation.
c
GT-1b CC50 and BVDV EC50 >10 lM.
EC50 in GT-1b LV/YH resistant replicon = 382 nM.
6h.1 and 6h.2 are symmetrical diastereomers.
d
e
Table 6
Data from 4 h rat PK screena
Compd
3g.2
5i.1
6g
0–4 h AUC (nM h)-PO
0–4 h AUC (nM h)-IV
CL (mL/min kg)-IV
166
5690
5.1
<LLQb
NDc
93
2110
13.3
NDc
a
PO/IV dose: 5/2 mg/kg; plasma was sampled 5 times (0.17, 0.5, 1, 2 and 4 h);
n = 2
b
c
Below level of quantitation.
ND, not determined (both rats died).
tion that was established between basicity and the GT-1a EC50
.
Unlike the phenylglycinamide series, the less potent mandela-
mide series exhibited a nuance in the SAR where stereochemical
preference was dependent on whether the substituent at the car-
binol carbon was hydrogen or a methyl group. The potent inhibi-
tion exhibited by 5g.1 toward a wider panel of HCV genotypes
was an early but crucial demonstration that potent and pan-
genotypic in vitro replicon inhibitory activity could be secured
by targeting the NS5A protein with evolved derivatives of the
stilbene chemotype. The subsequent phase of the SAR campaign
that dealt with the further optimization of these leads toward
the discovery of daclatasvir (1) will be the subject of future
communications.
Table 7
EC50 (nM) toward a panel of genotype strains20
Compd
2a
2c
5g.1
G-2a NIH
G-3a
G-3aYH
G-4a
260
0.30
>3000
0.16
<0.23
14
3.2
74.9
>10,000
3.0
0.083
0.037
5.4
0.042
0.017
G-5a
a
Data represent mean values of at least two experiments with a maximum of
threefold variation.
O
L
L
R
. HCl
O
L
O
i
H
ii
N
N
N
H
N
H
NH2
2
2
2
8x / 8y
3a - 3h, 5d, 5g, 5h, 6a
7x: L =
7y: L =
iv
iii
6a
5a
6f, 6g, 6i
v
5b
Scheme 1. Reagents and conditions: (i) (a) Boc-
L-proline, EEDQ, CH2Cl2; (b) 4 N HCl, dioxane; (ii) RCO2H, HATU, i-Pr2EtN, DMF; (iii) 25% TFA/ CH2Cl2; (iv) standard acylation or
sulfonylation conditions;18 (v) H2CO, HCO2H, CH2Cl2,
D
.

4434
M. Belema et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4428–4435
O
NH
ii
N
Br
5c 5e 5f 5i 5l
6h
and
,
,
,
-
R1
O
N
R2
i
Ph
10
O
O
Diastereomers separated
O
N
NH
R1
N
H
Br
NH
9
iii
O
N
R2
Ph
2
N
Br
NH
R1
i
ii
O
6b 6e
-
N
R2
4a 4c
-
Ph
11
Scheme 2. Reagents and conditions: (i) RCO2H, HATU, i-Pr2EtN, DMF; (ii) Me3SnC„CSnMe3, Pd(Ph3P)4, DMF, 80 °C; (iii) (a) Boc-(R)-phenylglycine, HATU, i-Pr2EtN, DMF; (b) 4
N HCl, dioxane; (c) ClCO2R, Et3N, THF
Ph
Ph
Ph
Ph
N
CO2H
ii
i
H2N
CO2H
N
CO2H
HO
CO2Bn
HO
15 (either enantiomer)
X
12
14
13a (X = CH2)
13b (X = OCH2)
Scheme 3. Reagents and conditions: (i) 1,4-dibromobutane or 1-bromo-2-(2-bromoethoxy)ethane, Na2CO3, EtOH, 100ꢀC; (ii) (a) p-TsCl, Et3N, DMAP, CH2Cl2; (b) 4-
hydroxypiperidine, i-Pr2EtN, THF,
D
; (c) enantiomers separated on Chiracel OJ column; (d) H2, 10% Pd/C, MeOH. Note: close monitoring is required to minimize the cleavage of
the benzylic C–N bond.
7. (a) Romine, J. L.; St. Laurent, D. R.; Lett, J. E.; Martin, S. W.; Serrano-Wu, M. H.;
Yang, F.; Gao, M.; O’Boyle, D. R., II; Lemm, J. A.; Sun, J.-H.; Nower, P. T.; Huang,
X.; Deshpande, M. S.; Meanwell, N. A.; Snyder, L. B. ACS Med. Chem. Lett. 2011, 2,
Acknowledgments
We thank members of our analytical group (Dieter Drexler, Qi
Gao, Xiaohua Huang, Alicia Ng, Gottfried Wenke, and Dedong
Wu) for conducting HRMS, 1H NMR, X-ray and pKa analyses.
224; (b) St. Laurent, D. R.; Belema, M.; Gao, M.; Goodrich, J.; Kakarla, R.; Knipe,
J. O.; Lemm, J. A.; Liu, M.; Lopez, O. D.; Nguyen, V. N.; Nower, P. T.; O’Boyle, D. R.,
II; Qiu, Y.; Romine, J. L.; Serrano-Wu, M. H.; Sun, J.-H.; Valera, L.; Yang, F.; Yang,
X.; Meanwell, N. A.; Snyder, L. B. Bioorg. Med. Chem. Lett. 2012, 22, 6063; (c)
Lopez, O. D.; Nguyen, V. N.; St. Laurent, D. R.; Belema, M.; Serrano-Wu, M. H.;
Goodrich, J. T.; Yang, F.; Qiu, Y.; Ripka, A. S.; Nower, P. T.; Valera, L.; Liu, M.;
O’Boyle, D. R., II; Sun, J.-H.; Fridell, R. A.; Lemm, J. A.; Gao, M.; Good, A. C.;
Meanwell, N. A.; Snyder, L. B. Bioorg. Med. Chem. Lett. 2013, 23, 779; (d) St.
Laurent, D. R.; Serrano-Wu, M. H.; Belema, M.; Ding, M.; Fang, H.; Gao, M.;
Goodrich, J. T.; Krause, R. G.; Lemm, J. A.; Liu, M.; Lopez, O. D.; Nguyen, V. N.;
Nower, P. T.; O’Boyle, D. R., II; Pearce, B. C.; Romine, J. L.; Valera, L.; Sun, J.-H.;
Wang, Y.-K.; Yang, F.; Yang, X.; Meanwell, N. A.; Snyder, L. B. J. Med. Chem. 2013.
http://dx.doi.org/10.1021/jm301796k.
References and notes
1. (a) Lavanchy, D. J. Viral Hepat. 1999, 6, 35; (b) Shepard, C. W.; Finelli, L.; Altar,
M. J. Lancet Infect. Dis. 2005, 5, 558; (c) Hanafiah, K. M.; Groeger, J.; Flaxman, A.
D.; Wiersma, S. T. Hepatology 2013. http://dx.doi.org/10.1002/hep.26141; (d)
Denniston, M. M.; Klevens, R. M.; McQuillan, G. M.; Jiles, R. B. Hepatology 2012,
55, 1652.
2. (a) Simmonds, P. J. Gen. Virol. 2004, 85, 3173; (b) Neumann, A. U.; Lam, N. P.;
Dahari, H.; Gretch, D. R.; Wiley, T. E.; Layden, T. J.; Perelson, A. S. Science 1998,
282, 103.
8. For discussions on how structural compositions of candidate molecules—
assessed in terms of aromatic count or sp3-carbon content—affects their
developability, see: (a) Lovering, F.; Bikker, J.; Humblet, C. J. Med. Chem. 2009,
52, 6752; (b) Ritchie, T. J.; Macdonald, S. J. F. Drug Discovery Today 2009, 14,
1011.
9. ELISA, FRET and luciferase assays were used to assess inhibitory activities
toward GT-1a, GT-1b and BVDV replicons, respectively, and Alamar blue assay
was used to determine cytotoxicity in GT-1b replicon. ELISA and FRET assays
afforded similar EC50 value for control compounds in GT-1a replicon. For
additional details on the replicon assays, see Ref. 7d and those cited therein.
10. The stereochemical designations in the SAR discussions pertain to the benzylic
centers of the cap moieties, unless noted otherwise.
3. Keam, S. J.; Cvetkovic, R. S. Drugs 2008, 68, 1273.
4. (a) Poordad, F.; McCone, J., Jr.; Bacon, B. R.; Bruno, S.; Manns, M. P.; Sulkowski,
M. S.; Jacobson, I. M.; Reddy, K. R.; Goodman, Z. D.; Boparai, N.; DiNubile, M. J.;
Sniukiene, V.; Brass, C. A.; Albrecht, J. K.; Bronowicki, J.-P. N. Eng. J. Med. 2011,
364, 1195; (b) Jacobson, I. M.; McHutchison, J. G.; Dusheiko, G.; Di Bisceglie, A.
M.; Reddy, K. R.; Bzowej, N. H.; Marcellin, P.; Muir, A. J.; Ferenci, P.; Flisiak, R.;
George, J.; Rizzetto, M.; Shouval, D.; Sola, R.; Terg, R. A.; Yoshida, E. M.; Adda,
N.; Bengtsson, L.; Sankoh, A. J.; Kieffer, T. L.; George, S.; Kauffman, R. S.;
Zeuzem, S. N. Eng. J. Med. 2011, 364, 2405; (c) Ghany, M. G.; Nelson, D. R.;
Strader, D. B.; Thomas, D. L.; Seeff, L. B. Hepatology 2011, 54, 1433.
5. (a) Lok, A. S.; Gardiner, D. F.; Lawitz, E.; Martorell, C.; Everson, G. T.; Ghalib, R.;
Reindollar, R.; Rustgi, V.; McPhee, F.; Wind-Rotolo, M.; Persson, A.; Zhu, K.;
Dimitrova, D. I.; Eley, T.; Guo, T.; Grasela, D.; Pasquinelli, C. N. Eng. J. Med. 2012,
366, 216; (b) Chayama, K.; Takahashi, S.; Toyota, J.; Karino, Y.; Ikeda, K.;
Ishikawa, H.; Watanabe, H.; McPhee, F.; Hughes, E.; Kumada, H. Hepatology
2012, 55, 742; (c) Gane, E. J.; Stedman, C. A.; Hyland, R. H.; Ding, X.;
Svarovskaia, E.; Symonds, W. T.; Hindes, R. G.; Berrey, M. M. N. Eng. J. Med.
2013, 368, 34; (d) Poordad, F.; Lawitz, E.; Kowdley, K. V.; Cohen, D. E.;
Podsadecki, T.; Siggelkow, S.; Heckaman, M.; Larsen, L.; Menon, R.; Koev, G.;
Tripathi, R.; Pilot-Matias, T.; Bernstein, B. N. Eng. J. Med. 2013, 368, 45; (e)
Suzuki, Y.; Ikeda, K.; Suzuki, F.; Toyota, J.; Karino, Y.; Chayama, K.; Kawakami,
Y.; Ishikawa, H.; Watanabe, H.; Hu, W.; Eley, T.; McPhee, F.; Hughes, E.;
Kumada, H. J. Hepatology 2013, 58, 655; (f) Asselah, T.; Marcellin, P. Liver Int.
2013, 33, 93.
11. Kuhn, B.; Mohr, P.; Stahl, M. J. Med. Chem. 2010, 53, 2601.
12. Number of conformers observed: 4a (1), 4b (2), and 4c (8). Cambridge
Structural Database deposition numbers are: CCDC 939741-939743.
13. It is noteworthy that 4b exhibited EC50s of >1.3 lM (G-1a) and 0.38 lM (G-1b),
which were significantly weaker than that of the parental dimeric analogue
3g.2.
14. For discussions on potential off-target liabilities of basic compounds, see: (a)
Azzaoui, K.; Hamon, J.; Faller, B.; Whitebread, S.; Jacoby, E.; Bender, A.; Jenkins,
J. L.; Urban, L. ChemMedChem 2007, 874; (b) Peters, J.-U.; Schnider, P.; Mattei,
P.; Kansy, M. ChemMedChem 2009, 4, 680; (c) Meanwell, N. A. Chem. Res.
Toxicol. 2011, 24, 1420; (d) Tarcsay, A.; Keseru, G. J. Med. Chem. 2013, 56, 1789.
15. All the three analogues had Caco-2 cell permeability <15 nm/s. The
susceptibility of compounds to PGP-efflux was not determined.
16. The GT-1b LV/YH EC50 of 2c was not determined.
17. The pan-genotype coverage assessment used hybrid replicons in a Con1 (GT-
1b) backbone where either the first 110 amino acid of NS5A (for G-3a, G-3aYH,
G-4a or G-5a) or the full length NS5A (for G-2a NIH) was replaced by that of the
respective genotype.
6. Gao, M.; Nettles, R. E.; Belema, M.; Snyder, L. B.; Nguyen, V. N.; Fridell, R. A.;
Serrano-Wu, M. H.; Langley, D. R.; Sun, J.-H.; O’Boyle, D. R., II; Lemm, J. A.;
Wang, C.; Knipe, J. O.; Chien, C.; Colonno, R. J.; Grasela, D. M.; Meanwell, N. A.;
Hamann, L. G. Nature 2010, 465, 96.

M. Belema et al. / Bioorg. Med. Chem. Lett. 23 (2013) 4428–4435
4435
18. Belema, M.; Nguyen, V. N.; Serrano-Wu, M. H.; St. Laurent, D. R.; Qiu, Y.; Ding,
M.; Meanwell, N. A.; Snyder, L. B. PCT Intl. Application 2010, WO2010-039793.
19. It is noteworthy that (R)-2-acetamido-2-phenylacetic acid was susceptible to
epimerization when coupled with pyrrolidine 8y under HATU/DIEA activation,
an observation with a literature precedent: Greene, T. W.; Wuts, P. G. M.
Protective Groups in Organic Synthesis, 3rd ed.; John Wiley & Sons, Inc: NY, New
York, 1999. p 503. Thus, 6g was prepared from amine 5a through acylation..
20. Although compounds 6b–6e could also be accessed readily from amine 5a,
they were prepared via the homocoupling route in order to assess the replicon
inhibitory activities of some of their monomeric synthetic precursors.