Discovery of the first C -nucleoside HCV polymerase inhibitor (GS-6620) with demonstrated antiviral response in HCV infected patients

Article abstract of DOI:10.1021/jm400201a

Hepatitis C virus (HCV) infection presents an unmet medical need requiring more effective treatment options. Nucleoside inhibitors (NI) of HCV polymerase (NS5B) have demonstrated pan-genotypic activity and durable antiviral response in the clinic, and they are likely to become a key component of future treatment regimens. NI candidates that have entered clinical development thus far have all been N-nucleoside derivatives. Herein, we report the discovery of a C-nucleoside class of NS5B inhibitors. Exploration of adenosine analogs in this class identified 1′-cyano-2′-C-methyl 4-aza-7,9-dideaza adenosine as a potent and selective inhibitor of NS5B. A monophosphate prodrug approach afforded a series of compounds showing submicromolar activity in HCV replicon assays. Further pharmacokinetic optimization for sufficient oral absorption and liver triphosphate loading led to identification of a clinical development candidate GS-6620. In a phase I clinical study, the potential for potent activity was demonstrated but with high intra- and interpatient pharmacokinetic and pharmacodynamic variability.

Full text of DOI:10.1021/jm400201a

Article
pubs.acs.org/jmc
Discovery of the First C‑Nucleoside HCV Polymerase Inhibitor (GS-
6620) with Demonstrated Antiviral Response in HCV Infected
Patients
Aesop Cho,* Lijun Zhang, Jie Xu, Rick Lee, Thomas Butler, Sammy Metobo, Vangelis Aktoudianakis,
Willard Lew, Hong Ye, Michael Clarke, Edward Doerffler, Daniel Byun, Ting Wang, Darius Babusis,
Anne C. Carey, Polina German, Dorothea Sauer, Weidong Zhong, Stephen Rossi, Martijn Fenaux,
John G. McHutchison, Jason Perry, Joy Feng, Adrian S. Ray, and Choung U. Kim
Gilead Sciences, 333 Lakeside Drive, Foster City, California 94044, United States
S
* Supporting Information
ABSTRACT: Hepatitis C virus (HCV) infection presents an unmet medical need requiring
more effective treatment options. Nucleoside inhibitors (NI) of HCV polymerase (NS5B)
have demonstrated pan-genotypic activity and durable antiviral response in the clinic, and
they are likely to become a key component of future treatment regimens. NI candidates that
have entered clinical development thus far have all been N-nucleoside derivatives. Herein, we
report the discovery of a C-nucleoside class of NS5B inhibitors. Exploration of adenosine
analogs in this class identified 1′-cyano-2′-C-methyl 4-aza-7,9-dideaza adenosine as a potent
and selective inhibitor of NS5B. A monophosphate prodrug approach afforded a series of
compounds showing submicromolar activity in HCV replicon assays. Further pharmacoki-
netic optimization for sufficient oral absorption and liver triphosphate loading led to
identification of a clinical development candidate GS-6620. In a phase I clinical study, the potential for potent activity was
demonstrated but with high intra- and interpatient pharmacokinetic and pharmacodynamic variability.
new drugs themselves.4 Therefore, there is still an urgent need
INTRODUCTION
■
for interferon-sparing regimens consisting of multiple DAAs for
Hepatitis C virus (HCV) infection presents a global heath
problem in need of more effective treatment options. The
World Health Organization estimates that 130−170 million
individuals have detectable antibodies to HCV worldwide,
higher efficacy and improved tolerability.5
A number of DAAs targeting various viral essential proteins
have demonstrated antiviral response in HCV infected
patients.3 Among them, nucleoside inhibitors (NIs) of NS5B
corresponding to 3% of the world’s population. Approximately
polymerase have a profile that suggests they will be critical
60−85% of such cases develop into chronic disease, leading to
components in future treatment regimens.6 Class related
liver cirrhosis (5−25%), end-stage liver disease, and hepatocel-
lular carcinoma (1−3%) over the course of the disease.1 The
properties that differentiate NI from other classes of HCV
inhibitors include a high genetic barrier to drug resistance,
allowing for durable antiviral suppression, and pan-genotype
anti-HCV activity affording universal treatment coverage.
However, the complex nature of the nucleoside pharmacology
poses a great difficulty in predicting efficacy and safety in
humans, resulting in high attrition during drug development.
Several nucleoside polymerase inhibitors have entered
clinical trials. Recent examples are shown in Figure 1.7,8
Structurally, all of these compounds are N-nucleosides
containing a 2′-C-Me branched sugar. Some of the 2′-C-Me
N-nucleosides were synthesized as early as the late 1960s prior
to the discovery of HCV, but their in vitro anti-HCV activity
was not discovered until the early 2000s.9,10 These inhibitors
are metabolized in cells to their respective nucleoside
current standard of care (SOC) includes administration of
pegylated interferon-α (Peg-IFN) and ribavirin (RBV), which
primarily relies on the patient’s ability to build up an immune
response against the virus. However, this SOC is suboptimal in
that it affords variable efficacy across the multiple genotypes of
HCV as well as poor safety and tolerability.2 Recently, there has
been an intense thrust for development of small molecule
therapeutic agents directly targeting the viral replication and
infectivity.3 The HCV genome (9.6 kb single (+)-strand RNA)
encodes both structural and nonstructural proteins (NS2 to
NS5), most of which have been exploited as drug targets. These
drug discovery efforts resulted in a number of direct-acting
antivirals (DAAs) entering clinical development over the past
several years. In 2011, two DAAs (NS3 protease inhibitors;
teleprevir and boceprevir) were approved for clinical use in
combination with Peg-IFN/RBV. Although these regimens
substantially improve response rates, they are limited to the
treatment of HCV genotype 1 infected patients, and they suffer
from the combined side effects of both Peg-IFN/RBV and the
Special Issue: HCV Therapies
Received: February 7, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 1. Nucleoside polymerase inhibitors and their clinical development status7,8
.
triphosphates, which bind to the active site of the NS5B
polymerase as a substrate and become incorporated into a
growing chain of viral RNA. Upon incorporation, they inhibit
further RNA elongation, resulting in inhibition of viral
replication. The 2′-C-Me group prevents an incoming nucleo-
side triphosphate from binding to the active site of NS5B.11
Previously, we reported the synthesis and structure−activity
relationship (SAR) of C-nucleosides containing the 2′-C-Me
group as NS5B inhibitors.12 In this earlier study, 2′-C-methyl 4-
aza-7,9-dideaza adenosine (1) was identified as a selective
inhibitor of HCV replication in cell cultures (EC50(1b) 1.98
μM, Table 1). Compound 1 can be viewed as a C-nucleoside
RNA synthesis in an enzyme assay (IC50 0.31 μM for 1, 0.30
μM for 29). This discrepancy was explained by a difference in
efficiency of triphosphate formation in replicon cells (Huh-7).
Compound 1 was metabolized to its biologically active
triphosphate species less efficiently than 29 in Huh-7 cells.
While less efficiently metabolized in Huh-7 cells, 1 was actually
more efficiently metabolized in primary hepatocytes in vitro
and more efficient at forming its active triphosphate in liver
following oral administration to rodents than 29. The
combined similar intrinsic potency and improved hepatic
delivery suggest that 1 would exhibit potent in vivo anti-HCV
activity rivaling that observed for 29.14
In the hopes of improving the cellular selectivity of 1, we
have pursued further structural modifications on both the base
and the sugar moieties. Previously, we also reported a series of
1′-substituted C-nucleosides displaying antiviral activity against
various RNA viruses, including HCV.15 Thus, we decided to
incorporate the 1′-substitution into 1. This strategy resulted in
the identification of 1′-cyano-2′-C-methyl 4-aza-7,9-dideaza
adenosine (2), which, as its triphosphate, showed a
substantially improved selectivity over mitochondrial RNA
polymerase (POLRMT) while maintaining the intrinsic NS5B
inhibitory activity. A monophosphate prodrug approach was
undertaken to achieve the whole cell replicon activity. Further
pharmacokinetic optimization led to the discovery of 18 (GS-
6620), the first C-nucleoside clinical candidate. In phase I
studies, the compound demonstrated the potential for potent
anti-HCV activity in HCV-infected patients, but with a high
PK/PD variability.16
Table 1. Comparison of N-Nucleoside and C-Nucleoside 7-
a
Deazaadenosine Analogs
29
1
EC50 (1b, μM)
0.08
>89
0.30
1.98
85
CC50 (Huh-7, μM)
IC50 as NTP (1b, μM)
NTP level (Cmax, pmol/million)
in Huh-7 cells
0.31
25
8
5
in human hepatocytes
22
RESULTS AND DISCUSSION
a
Summary of results reported previously.12
■
On the basis of the favorable properties previously reported for
1,12 toxicology studies were conducted in Sprague−Dawley
rats. Disappointingly, oral administration of 1 once a day for 7
days resulted in unscheduled deaths at 20 and 40 mg/(kg day).
The no-observable adverse effects level (NOAEL) was
determined to be less than 5 mg/(kg day), the lowest dose
tested in the studies. Adverse effects included necrosis in organs
such as the kidney, liver, and pancreas, as well as lymphocellular
variant of an N-nucleoside 2′-C-methyl-7-deaza adenosine 29
(MK-608), which was one of the most potent adenosine
analogs reported in the literature at the time we initiated this
work.13 Compared to 29, the replicon activity of 1 was rather
weak (EC50 0.08 μM for 29) despite its triphosphate being as
potent as the 29 triphosphate at inhibiting NS5B-mediated
B
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
depletion. This toxicity profile had a resemblance to amanitin
mushroom poisoning, which is primarily caused by inhibition of
host mRNA synthesis by RNA polymerase II (Pol II).17 To
understand the cause for the toxicity, the triphosphate of 1 was
tested against a series of host DNA and RNA polymerases.
While the triphosphate did not inhibit Pol II, it was a good
substrate for POLRMT. Since POLRMT is likely a key
mediator of ribonucleotide analog toxicity,18 we monitored
selectivity over this off-target enzyme in our discovery program.
In light of the finding that the triphosphate of 1 was a
substrate of POLRMT, a possible cause for the observed
toxicity in rats, further structural modification of 1 was sought
for improved selectivity. Applying chemistry that we have
developed,15 a series of 1′-substituted analogs of 1 were
prepared and evaluated (Table 2). Among those tested, 1′-
analogs (3, 4) had no appreciable NS5B inhibitory activity.
Overall, an addition of the CN group at the 1′ position of 1
maintained the intrinsic enzyme activity, while abolishing the
incorporation by POLRMT.
However, compound 2 was not active in the whole cell
replicon assay. This was not surprising, since previous studies
showed that the 1′-substitution limits the first of three
phosphorylation steps in metabolic conversion to the bio-
logically active nucleoside triphosphate.15 Accordingly, a
phosphoramidate-type monophosphate prodrug (5) was
prepared and tested in the HCV replicon assay. It was found
to exhibit respectable replicon activity (EC50 (1b) 1.05 μM). It
was also shown that a high level of the triphosphate (∼20 μM)
was formed when 10 μM of 5 was incubated for 24 h in Huh-7
cells that host the GT1b replicon. In comparison, less than 0.5
μM of the triphosphate was detected with the incubation at 10
μM of the parent nucleoside 2. These results indicate that the
first-step phosphorylation is slow, but the second and third
phosphorylations are efficient (Figure 2).
a
Table 2. 1′-Substituted Analogs of 1
On the basis of the potent enzyme activity, the remarkable
selectivity over POLRMT, and the whole cell replicon activity
as the monophosphate prodrug, compound 2 became our lead
nucleoside for prodrug optimization. Various monophosphate
prodrug approaches have been utilized to overcome the first-
step phosphorylation and enhance antiviral activity of the
parent nucleosides.19 Among them, the phosphoramidate-type
prodrug approach has recently demonstrated clinical utility in
the treatment of HCV infection. For example, a phosphor-
amidate prodrug of 2′-C-Me-2′-F-uridine (sofosbuvir) has
demonstrated high sustained virologic response (SVR) rates
when administered in combination with other antiviral agents
for the treatment of chronic HCV infection and has advanced
to phase III clinical trials.20 A similar chemical approach was
adopted in our work, initially focusing on SAR study around the
compd
R
EC50 (1b, μM) IC50 (1b, μM) mtRNA SNI (%)
1
2
3
4
H
1.98
>89
>89
>89
0.31
0.29
55
21
cyano
0.03
0.08
ND
methyl
ethynyl
>200
a
See the Experimental Section for detailed assay protocols.
cyano-2′-C-methyl 4-aza-7,9-dideaza adenosine (2) was the
most active as its triphosphate in the NS5B enzyme assay and,
importantly, not a substrate for POLRMT. Other 1′-substituted
Figure 2. Monophosphate prodrug activation starts with an enzymatic ester hydrolysis, which leads to formation of the 5′-monophosphate (2-MP),
which is further phosphorylated to the biologically active 5′-triphosphate (2-TP). Nucleoside 2 itself is poorly phosphorylated to 2-MP.
C
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
phosphoramidate portion of 5. Since the enzymatic ester
hydrolysis is known to be the first step in prodrug activation,
phosphoramidate prodrugs with various ester groups were
prepared and their anti-HCV activity tested (Table 3).
properties relevant to oral absorption were evaluated first
(Table 4). Water solubility was acceptable (>100 μg/mL) for
prodrugs with alkyl esters containing less than five carbons.
However, as the lipophilicity (log D) was increased, the
solubility dropped sharply. With a couple of exceptions (8 and
14), Caco-2 permeability was low with high efflux potential.
Intestinal absorption turned out to be a major issue with this
series of prodrugs (see below for further discussion). Stability in
both serum and intestinal S9 was excellent for secondary alkyl
esters. On the other hand, a majority of these prodrugs were
efficiently metabolized by hepatic S9, suggesting liver-specific
activation. To assess the liver triphosphate loading potential,
compounds were incubated in primary human hepatocytes and
the resulting triphosphate levels were measured. All compounds
except 10 produced high levels of triphosphate, where the
relative order of the level of the triphosphate tracked well with
the relative replicon activity.
Table 3. HCV Replicon Activity of Phosphoramidate
a
Prodrugs
compd
R1
EC50 (1b, μM)
CC50 (Huh-7, μM)
6
Me
Et
1.37
1.45
1.05
0.23
0.98
>89
>89
>89
>89
>89
>89
>89
>89
85
7
5
iso-Pr
To profile in vivo characteristics of these prodrugs,
pharmacokinetic studies assessing liver triphosphate levels
were performed in hamsters. Following intravenous (iv) and
oral administrations (po), levels of the triphosphate metabolite
in the liver were determined (Table 5). The iv dosing resulted
8
(S)-sec-Bu
(R)-sec-Bu
t-Bu
9
10
11
12
13
14
cyc-pentyl
neo-pentyl
2-EtBu
Bn
0.45
0.18
0.44
0.33
35
Table 5. In Vivo PK Properties of Phosphoramidate
Prodrugs
63
a
See the Experimental Section for detailed assay protocols.
hamsters, liver triphosphate
dogs, plasma prodrug
AUC0‑inf
Prodrugs with primary and secondary alcohol esters (5−9,
11−14) were all active in the replicon assay, but the tertiary
alcohol ester prodrug (10) was inactive, presumably due to the
lack of enzymatic ester hydrolysis. Bigger and more lipophilic
alkyl esters generally showed higher potency; however,
cytotoxicity was observed in the case of cyclo-pentyl (12), 2-
ethylbutyl (13), and benzyl (14) esters.
The ultimate goal for the program was to identify an orally
active agent. To be orally active, the monophosphate prodrug
must be well absorbed upon oral administration and travel to
the liver without degradation. The intact prodrug should then
be efficiently metabolized and converted to the triphosphate in
the liver, the major site for HCV replication. Thus, at the
outset, we aimed at achieving both high oral absorption and
efficient liver triphosphate loading. To assess the oral
absorption potential for these compounds, in vitro DMPK
a
c
d
Cmax (μM)
iv
%
(μM·h)
%
b
compd
po
0.29
”oral efficiency”
portal jagular
Fa
6
6.35
4.48
5.65
ND
ND
2.80
5.18
8.33
ND
0.9
2.5
1.9
7
0.57
0.53
1.27
5
0.99
0.80
0.46
0.16
0.12
0.14
10
16
8
9
1,87
8.1
11
12
13
14
0.84
1.44
2.26
0.86
6.0
5.6
5.4
0.25
0.21
0.09
0.02
13
7.9
a
b
Dosed at 1 mg-equiv/kg for iv, 5 mg-equiv/kg for po. “oral
c
d
efficiency” = po Cmax ÷ 5/iv Cmax × 100%. Dosed at 0.5 mg/kg. See
the Experimental Section for details.
a
Table 4. Physicochemical Properties, Permeability and Stability in Human Extracts of Select Phosphoramidate Prodrugs
Caco-2
permeability
b
(×10−6 cm/s)
c
d
compd
log D
solubility (μg/mL)
A to B
B to A
intestinal S9 stability (t1/2, min)
hepatic S9 (% extraction)
TP formation in hepatocytes
6
1.3
1.6
1.8
2.1
1.9
2.2
2.2
2.3
2.7
2.2
ND
108
306
200
151
380
10
0.41
0.17
0.16
1.47
0.75
0.67
0.37
0.69
0.36
3.35
4.64
2.98
8.92
17.9
37.5
31.1
5.86
29.1
7.20
25.3
190
69
90
89
90
93
94
84
95
94
97
93
+
7
+
5
>612
592
461
>612
30
+
8
++
+
9
10
11
12
13
14
NT
++
+++
+++
+++
115
1.8
85
74
0.05
129
a
b
c
d
See the Experimental Section for detailed assay protocols. Permeability through Caco-2 cells determined at 10 μM. Measured at pH 7. Relative
levels of triphosphate measured upon 1 h of incubation (10 μM) with primary human hepatocytes; +++, very high (>200 pmol/million); ++, high; +,
medium (50−100 pmol/million).
D
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
in relatively high levels of the triphosphate in all cases,
suggesting highly efficient liver loading. However, triphosphate
levels were substantially lower following oral administration,
suggesting inefficient oral absorption. The “oral efficiency” of
liver triphosphate loading based on dose-normalized po
triphosphate Cmax relative to iv triphosphate Cmax was poor
(<10%). These results suggest that only a small fraction of the
prodrugs were delivered to the liver by the oral route due to
inefficient prodrug absorption. To gain further insight, oral
absorption for selected prodrugs was determined in portal vein
cannulated dogs where both absorption and hepatic extraction
can be measured. In line with the poor oral delivery in
hamsters, oral absorption in dogs was poor (<20%) in all cases
evaluated. The in vivo results from these two different animal
studies strongly suggested this class of compounds has
limitations on oral bioavailability.
permeability was also substantially improved, corroborating the
in vivo results (Figure 3). In particular, the efflux effect was
dramatically reduced at a high incubation concentration (efflux
ratio of <2 at 100 μM). In addition, 16 delivered a much higher
level of the triphosphate in primary human hepatocytes. Thus,
16 achieved our targeted preclinical profile. Encouraged by
these results, the liver pharmacokinetic profile in dogs was
determined (Figure 4). Upon oral administration at 5 mg/kg,
We then sought out other ways of improving the oral
absorption potential. Strategies such as replacement of L-alanine
with other L-amino acids (e.g., glycine, L-leucine, L-phenyl-
alanine) and substitutions on the phenyl ring (e.g., 2-F, 2,4-
diMe, 3,4-methylenedioxy) were explored, but they proved
unfruitful. All these prodrugs, for example, suffered from poor
permeability with high efflux ratio, which will be summarized in
a further publication. Next, a double prodrug approach was
sought where one or two hydroxyl groups of the nucleoside
sugar were protected. Accordingly, compounds with 2′,3′-O-
carbonate (15) and the 3′-O-isobutyrate (16) were prepared21
and evaluated in a series of both in vitro and in vivo assays
(Table 6). Remarkably, compound 16 achieved substantially
Figure 4. Mean liver versus time profile of the triphosphate metabolite
following oral administration of 16 to beagle dogs (5 mg/kg; solution)
(mean SD; n = 2/time point for dog).
the triphosphate levels in the liver exceeded the IC90 over 24 h,
suggesting the potential to be effective in treating patients
infected with HCV at a reasonable dose.
a
Table 6. Comparison of Selected Prodrugs
Nonclinical toxicology studies were then conducted with
compound 16. Rats and dogs were dosed for 7 days at up to
300 and 75 mg/(kg day), respectively. No test-article related
toxicities were observed from these studies. Consequently, the
NOAEL was established at >300 mg/(kg day) for rats and >75
mg/(kg day) for dogs. The results for 16 were in striking
contrast to those described earlier for the 1′-unsubstitued C-
nucleoside 1, establishing a correlation between the in vitro
observation of incorporation by POLRMT and in vivo toxicity.
Subsequent toxicology studies further established the toler-
ability of this nucleotide prodrug. The NOAEL values were the
highest doses tested following chronic administration with 18
(a pure diastereoisomer of 16, see below) for 26 weeks in rats
and 39 weeks in dogs (1,000 and 60 mg/(kg day),
respectively).
5
15
1.41
16
0.61
EC50 (1b, μM)
CC50 (Huh-7, μM)
log D
1.05
>89
1.8
>89
47
2.6
2.9
hamsters, liver TP Cmax (μM); iv/po 5.63/0.53
1.92/0.59
29
4.50/3.67
88
dogs, % oral absorption
10
82
human hepatocytes TP (Cmax
pmol/million)
,
60
274
a
See the Experimental Section for detailed assay protocols.
higher liver triphosphate loading in hamsters compared to the
3′-unsubstitued precursor prodrug 5, when administered orally
(liver TP Cmax 3.67 μM vs 0.53 μM). Furthermore, oral
absorption in dogs for 16 was substantially high (88%). Caco-2
Figure 3. Prodrug 16 has high permeability and minimal efflux when tested at 100 μM in caco-2 cells. At lower concentrations, reduced forward
permeability and increased efflux ratio is observed, suggesting 16 is subject to saturable efflux transport. The apical to basolateral permeability of 16 is
increased 10-fold over that observed for the des-3′-isobutyryl prodrug 5. The blue bar is for forward permeability and the red bar for backward
permeability.
E
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Compound 16 is a diastereomeric mixture at the phosphorus
center of the phosphoramidate pro-moiety. To select the
optimal isomer for clinical development, the two single
diastereomers (18 and 20) were profiled separately and their
properties compared. Compared to the Rp isomer 20, the Sp
isomer 18 was 6-fold more potent in the replicon assay and two
times more efficient in the triphosphate formation in primary
human hepatocytes, and it achieved 3-fold higher levels of liver
triphosphate following oral administration to hamsters. Thus,
18 was the preferred isomer and was selected as a clinical
development candidate.
The antiviral characteristics of 18 or its diastereomeric
mixture 16 were more fully assessed in studies presented in the
Supporting Information. In an expanded panel of the replicon
assay using stable replicons GT1b, 1a, and 2a and chimeric
replicons 2b, 3a, 4a, 5a, and 6a, 18 displayed potent inhibition
against all the genotypes tested with EC50 ranging between
0.068 μM and 0.43 μM (Table S1). These results were in
contrast to those of a non-nucleoside NS5B inhibitor tested in
the same assays that showed EC50 values varying over 600-fold
depending on genotype. As observed with other 2′-C-Me
nucleos(t)ides, such as 29 and 2′-C-Me adenosine, 16 had a
significantly reduced activity (>95 fold) against the replicon
encoding the S282T mutation in NS5B. No resistance was
observed for the N142T mutation selected by the nucleoside
analog 4′-azido cytidine (Table S2). Combinations of 16 with
anti-HCV agents from the non-nucleoside NS5B, NS5A
inhibitor, and NS3 protease inhibitor classes in an Huh-7 cell
line carrying GT 1a replicon showed additive to minor synergy
in antiviral activity (Table S3).
fully functionalized heteroaromatic bases. This reaction leads to
various 1′-hydroxy C-nucleosides, which are versatile inter-
mediates for a number of both 1′-unsubstituted and substituted
C-nucleosides. Applying this chemistry, a series of 1′-
substituted 2′-C-Me nucleosides (2−4) were prepared. For
preparation of 2, the 1′-hydroxy intermediate 2112 was
subjected to an acid-promoted replacement of the 1′-hydroxyl
group by the 1′-CN. This reaction was highly stereoselective,
affording a single stereoisomer 22 as the sole isolated product.
This can be explained by blockage of the β-face approach of the
CN nucleophile toward the acid-induced five-membered-ring
oxocarbenium ion due to steric congestion exerted by the
neighboring 2′-β-Me. Subsequent debenzylation was conven-
iently achieved by boron trichloride (Scheme 1). Although the
a
Scheme 1.
The antiviral activity, safety, and tolerability of 18 were
evaluated in a randomized, double-blind, placebo controlled,
̈
multiple-dose, dose-escalation study in treatment-naive subjects
chronically infected with HCV genotype 1.16 Nine cohorts
(consisting of eight active and two placebo) assessing doses
between 50 mg once daily and 900 mg twice daily (BID)
administered under fed or fasted conditions for 5 days were
studied. Compound 18 was demonstrated to be safe and well
tolerated at up to 900 mg BID for 5 days. The greatest antiviral
activity was observed in BID cohorts with median viral load
reductions of 1.73 and 1.63 log10 IU/mL from baseline with
900 mg administered as a tablet or 450 mg administered as a
solution, respectively. The potential for potent antiviral activity
was evident in two patients in the 900 mg BID cohort who had
their viral loads reduced to below the lower limit of
quantification (representing respective viral load reductions of
greater than 3.37 log10 IU/mL and 4.37 log10 IU/mL) and one
patient in the 450 mg solution cohort achieving a viral load
reduction of 4.06 log10 IU/mL. However, substantial intra- and
interpatient PK/PD variability was observed in all cohorts. For
example, in the 900 mg BID cohort 18, exposure on day 5
varied by more than an order of magnitude (ranging from 41.6
ng h/mL to 760 ng h/mL) and maximal viral load reductions
varied by more than 3 orders of magnitude (ranging from 1.14
log10 IU/mL to greater than 4.37 log10 IU/mL). Notably,
plasma levels of 18 were markedly lower than those observed
preclinically in dogs, suggesting a species difference in prodrug
absorption or metabolism.
a
Reagents and conditions: (a) 20, 1,1,4,4-tetramethyl-1,4-dichlorodi-
silylethylene (1.0 equiv) and n-BuLi (3.3 equiv), THF, −78 °C
followed by addition of 19, 1 h, 65−80%; (b) TMSCN (6 equiv),
TMSOTf (4 equiv), CH2Cl2, −15 to 0 °C, 2 h, 85−95%; (c) BCl3 (3
equiv), CH2Cl2, 0 °C, 1 h, 90%.
debenzylation reaction was clean as long as the boron reagent
(1 M BCl3 in dichloromethane) was fresh, adequate
neutralization of the crude reaction mixture using a base such
as sodium methoxide or triethylamine during workup was
necessary to avoid unwanted hydrolysis of the 1′-CN group.
A number of phosphoramidate prodrugs derived from
nucleoside 2 were prepared initially following the literature
methods, which conventionally afforded the products as a
mixture of two diastereomers at the phosphorus center (5−
14).8,23 For example, compound 5 was prepared by reacting 2
in the presence of N-methylimidazole with the chloridate 24,
which was freshly prepared by treatment of L-alanine isopropyl
ester with phenyl phosphorus dichloridate (Scheme 2).
Trimethyl phosphate was used as a cosolvent to improve the
solubility of 2 and maintain a homogeneous solution
throughout completion of the reaction. Since the chloridate
25 was produced as an even mixture of the two diastereomers
at the phosphorus, the subsequent reaction product 5 was also
a diastereomeric mixture. Later in our program, some analogs
(e.g., compounds 8 and 9) were prepared by a modified
method where the p-nitrophenolates as phosphorylating
CHEMISTRY
■
Previously, we reported a convergent synthesis of C-nucleo-
sides.12,15,22 The key step of synthesis involves a carbon−
carbon bond formation between protected ribonolactones and
F
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 2.
a
Reagents and conditions: (a) triethylamine, CH2Cl2, −78 °C and then warm to rt, 16 h; (b) N-methyl imidazole, trimethyl phosphate and THF, 0
°C to rt, 2 h, overall yield 59%.
a
Scheme 3.
a
Reagents and conditions: (a) CDI, CH3CN, rt, 6 h, 86%; (b) N,N -dimethylformamide dimethyl acetal, CH3CN, rt, 1 h; (c) isobutyric acid, DCC,
CH3CN, rt, 48 h; (d) water-TFA, 0 °C to rt, 64 h, overall yield 73%.
reagents and tert-butyl magnesium chloride as a base were used
(as in Scheme 5).
chloridate 25 was sought. The single isomeric phosphorylating
reagent should be stable, readily separable from the
diastereomeric mixture, and react with a 5′-hydroxyl group of
the nucleoside 2 stereospecifically to afford the phosphate ester
formation. After brief screening, the p-nitrophenolate 28 was
identified as the one that met these criteria.24 Accordingly, the
diastereomeric mixture 28 was prepared and subjected to chiral
column chromatography or fractional crystallization to obtain
the single isomer (Sp)-28 (Scheme 4). Its absolute stereo-
chemistry was determined by single crystal X-ray analysis
(CDCD 929320). Nucleoside 2 was treated with 1.5 equiv of
tert-butyl magnesium chloride in THF, and subsequent
treatment of the resulting suspension with (Sp)-28 at 50 °C
for 2−3 h afforded the (Sp)-isomer 18 (Scheme 5), of which
the absolute stereochemistry was also determined by X-ray
crystallography (CDCD 929321, Figure 5).
For the 2′,3′-di-O-carbonate prodrug 15, compound 5 was
simply treated with CDI . For the 3′-O-butyrate prodrug 16,
however, protection of the 6-amino group was required. The 6-
amino group of 5 was transiently protected as a formamidine
using N,N-dimethylformamide dimethyl acetal, and then the 3′-
OH was acylated by treatment of isobutyric acid and DCC.
Deprotection in aqueous TFA led to 16 in high overall yield
from this three-step process (Scheme 3).
Initially, the diastereomeric mixture 16 was separated by
chiral column chromatography (Chiralcel OD-H; heptanes and
isopropyl alcohol) to obtain the two single diastereomers (18
and 20). However, preparation of 18 in larger quantities
required a more practical method. Thus, stereospecific
synthesis using a single diastereomeric alternative to the
G
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
studies, 18 displayed proof of concept and the potential for
potent antiviral activity, but its clinical utility was limited by
high PK/PD variability.
Scheme 4.
EXPERIMENTAL SECTION
■
Nuclear magnetic resonance (NMR) spectra were recorded on Varian
Mercury Plus 400 MHz and a Varian Unity Plus 500 MHz
spectrometers at room temperature, with tetramethylsilane as an
internal standard. Chemical shifts (δ) are reported in parts per million
(ppm), and signals are reported as s (singlet), d (doublet), t (triplet),
q (quartet), m (multiplet), or br s (broad singlet). Purities of the final
compounds were determined by HPLC and were greater than 95%.
HPLC conditions to assess purity were as follows: Agilent HPLC,
Phenominex Kinetex C18, 4.6 mm × 100 mm (2.6 μm); gradient of
0.1% FTA water (A) and 0.1% TFA acetonitrile (B); flow rate, 1.5
mL/min; acquire time, 9 min; wavelength, UV 214 and 254 nm; oven.
The preparative HPLC system includes two sets of Gilson 332 pumps,
a Gilson 156 UV/vis detector, and a Gilson 215 injector and fraction
collector, with Trilution LC software. A Synergi 150 mm × 30 mm 4u
Hydro-RP 80 A column was used. The mobile phase was a HPLC
grade water (A) and HPLC grade acetonitrile (B) system. LC/MS was
conducted on a Thermo Finnigan MSQ Std using electrospray both
positive and negative [M + H+ and M − H+], and a Dionex Summit
HPLC System (model: P680A HPG) equipped with a Gemini 5 u C18
110A column (30 mm × 4.60 mm), eluting with 0.05% formic acid in
1% acetonitrile/water (solvent A) and 0.05% formic acid in 99%
acetonitrile/water (solvent B). High-resolution mass spectra were
recorded on a Xevo G2 Q-Tof mass spectrometer with an ESI source.
2′,3′,5′-O-Tribenzyl-1′-hydroxy-2′-C-methyl-4-aza-7,9-di-
deazaadenosine (21). To a dry, argon purged 1 L round-bottom
flask with stir bar at room temperature were charged the aryl bromide
20 (8.0 g, 38 mmol) and anhydrous THF (500 mL). The suspension
was gently stirred for 10 min. At this point, 1,2-bis-
(chlorodimethylsilyl)ethane (8.2 g, 38 mmol) was added as a solid,
and the sides of the vessel were rinsed with a small volume of THF.
Within a few minutes, the mixture became cloudy. The mixture was
allowed to stir for 10 min. At this point, solid NaH (60% in mineral
a
Reagents and conditions: (a) triethylamine, CH2Cl2, −20 °C to rt,
77%; (b) Chiralpak IC, 70% heptanes and 30% isopropanol, or
crystallization from diethyl ether and hexanes for (Sp)-28.
CONCLUSION
■
Incorporation of the cyano group at the 1′ position of the
previously disclosed 2′-C-methyl-7-deaza adenosine C-nucleo-
side (1) resulted in a novel class of nucleoside 2, which
displayed the desired level of NS5B enzyme inhibitory activity
and selectivity. Phosphoramidate prodrugs such as 5 afforded
potent replicon activity and high triphosphate loading in vitro
but suffered from low oral absorption in animals. This problem
was solved by 3′-O-acylation. From these efforts, compound 18
was identified, becoming the first C-nucleoside class of NS5B
inhibitors to enter clinical development. In the first-in-man
a
Scheme 5.
a
Reagents and conditions: (a) 20, t-BuMgCl (1.5 equiv), THF, rt, 10 min and then (Sp) or (Rp)-28, 50 °C, 3 h, 70−90%; (b), (c), (d) same as in
Scheme 3.
H
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 5. X-ray crystal structure of 18.
oil; 3.8 g, 95 mmol) was added in three portions. The resulting
mixture was stirred for 5 min and then cooled to −78 °C over 45 min.
Separately, an oven-dried addition funnel purged with argon was
charged with BuLi (1.6 M in hexanes, 78 mL, 125 mmol). Then, BuLi
was added dropwise to the mixture at −78 °C over 45 min. The
reaction progress was monitored by HPLC, and the starting material
20 was almost completely consumed by the time the addition was
complete. Another oven-dried argon purged addition funnel was
charged with a solution of the lactone 19 (19 g, 44 mmol) in
anhydrous THF (∼70 mL). This solution was then added dropwise to
the −78 °C solution of the lithio species over 45 min and allowed to
stir under argon at the same temperature for 2 h. The cold bath was
removed to allow the reaction mixture to warm slightly (30 min). The
cold reaction mixture was slowly poured into a large rapidly stirring
flask containing 1 M citric acid (800 mL) and EtOAc (500 mL). The
mixture was allowed to stir for 15 min and was transferred to a
separatory funnel (2 L). The organic phase was separated. The
aqueous phase was extracted with EtOAc (2 × 400 mL). The
combined organic extracts were washed sequentially with water (1 ×
400 mL), saturated NaHCO3 (2 × 400 mL), water (1 × 400 mL), and
brine (1 × 400 mL), dried over sodium sulfate, and then concentrated
to dryness. The crude was purified by silica gel flash column
chromatography (50% hexanes + 50% EtOAc to 100% EtOAc),
affording 21 as a glassy solid obtained (17.2 g, 81%). MS, m/e 567 (M
+ 1)+. 1H NMR (300 MHz, CDC13) δ 7.85 (m, 1H), 7.27 (m, 15H),
7.01 (m, 1H), 6.51 (m, 1H), 4.66 (m, 8H), 4.40 (m, 2H), 3.79 (m,
3H), 1.62 (s, 2′-CH3 from one isomer), 1.18 (s, 2′-CH3 from the other
isomer).
CH2Cl2 (100 mL). The organic phase was separated, washed with
brine (150 mL), dried over Na2SO4, and concentrated. The residue
was purified by silica gel column chromatography (eluted with
hexanes−EtOAc 0 to 100%) to give the desired product 22 (6.9 g,
93%). MS, m/e 576 (M + 1)+. 1H NMR (300 MHz, CDCl3) δ 7.98 (s,
1H), 7.2−7.5 (m, 15H), 7.11 (d, 1H), 6.63 (d, 1H), 5.78 (brs, 2H),
4.99 (ABq, 2H), 4.4−4.8 (m, 5H), 4.04 (d, 1H), 3.96 (dd, 1H), 3.75
(dd, 1H), 1.16 (s, 3H).
1′-Cyano-2′-C-methyl-4-aza-7,9-dideazaadenosine (2). To a
solution of compound 22 (20 g, 35 mmol) in CH2Cl2 (200 mL) at 0
°C was added BCl3 (105 mL, 1 M in CH2Cl2) over 10 min. The
reaction mixture was stirred at the same temperature for 1 h. An ice-
cold sodium methoxide in methanol (25%, 50 mL) was dropwise
added over 3 min. The mixture was stirred at 0 °C for 30 min and then
allowed to warm up to room temperature for 30 min. The mixture was
concentrated, and the residue was redissolved in methanol (200 mL)
and concentrated again. This process was repeated twice. The residue
was dissolved in methanol. Silica gel (∼30 g) was added and
concentrated. The residue was purified by silica gel column (0−20%
1
methanol/CH2Cl2) to give the desired product 2 (9.5 g, 90%). H
NMR (300 MHz, D2O): δ 7.74 (s 1H), 6.76 (d, J = 5 Hz, 1H), 6.73
(d, J = 5 Hz, 1H), 4.1 (m, 1H), 3.9 (m, 1H), 3.8 (m, 2H), 0.84 (s,
3H). MS, m/e 306 (M + 1)+. HRMS, calculated for C13H15N5O4 (M +
1)+, 306.1197; found, 306.1203.
(2S)-Isopropyl 2-((((2R,3R,4R,5R)-5-(4-Aminopyrrolo[1,2-f ]-
[1,2,4]triazin-7-yl)-5-cyano-3,4-dihydroxy-4-methyltetrahydro-
furan-2-yl)methoxy)(phenoxy)phosphorylamino)propanoate
(5). Phenyl phosphorodichloridate (4.99 g, 23.8 mmol) was dissolved
in dichloromethane (100 mL), and L-alanine isopropyl ester
hydrochloride (3.98 g, 23.8 mmol) was added. The resulting clear
solution was cooled to −78 °C over 30 min. Triethylamine (6.63 mL,
47.5 mmol) was added dropwise over 15 min. The mixture was then
allowed to warm to room temperature. After 6 h, the solvent was
removed under an argon stream. The residue was redissolved in
MTBE (25 mL), and the insoluble portion was removed by filtration
2′,3′,5′-O-Tribenzyl-1′-cyano-2′-C-methyl-4-aza-7,9-didea-
zaadenosine (22). To a solution of compound 21 (azeotropically
dried using toluene, 6.8 g, 12 mmol) in CH2Cl2 (150 mL) at −15 °C
was added TMSOTf (56 mmol) dropwise. Then TMSCN (73 mmol)
was added dropwise. The reaction mixture was stirred at −15 °C for
1.5 h and warmed to 0 °C for an additional 0.5 h. The reaction was
quenched with saturated NaHCO3 (75 mL) at 0 °C and diluted with
I
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
under argon. The filtrate was then condensed with an argon stream,
and the crude product 25 was used for the next reaction without
further purification. 1H NMR (300 MHz, CDC13) δ 7.1−7.4 (m, 5H),
5.1 (m, 1H), 4.35 (m, 1H), 4.15 (m, IH), 1.5 (d, 3H), 1.2 (m, 6H). 31P
NMR (121.4 MHz, CDC13) δ 7.8, 8.4 (s).
(m, 1H), 4.2−4.5 (m, 3H), 3.93 (m, 2H), 1.80 (m, 2H), 1.5−1.7 (m,
6H), 1.37 (d, 3H), 0.86 (s, 3H). 31P NMR (121.4 MHz, CDC13) δ
2.89, 2.92 (s). MS, m/e 601 (M + 1)+.
(2S)-neo-Pentyl 2-((((2R,3R,4R,5R)-5-(4-Aminopyrrolo[1,2-f ]-
[1,2,4]triazin-7-yl)-5-cyano-3,4-dihydroxy-4-methyltetrahydro-
furan-2-yl)methoxy)(phenoxy)phosphorylamino)propanoate
(12). 1H NMR (300 MHz, CD3CN) δ 7.98 (s, 1H), 7.2−7.4 (m, 5H),
6.82 (m, 2H), 6.44 (brs, 2H), 5.27 (s, 1H), 4.3−4.5 (m, 4H), 4.03 (m,
1H), 3.93 (m, 1H), 3.80 (m, 3H), 1.38 (d, 3H), 0.86 (m, 12H). 31P
NMR (121.4 MHz, CD3CN) δ 3.14 (s). MS, m/e 603 (M + 1)+.
HRMS, calculated for C27H36N6O8P (M + 1)+, 603.2332; found,
603.2331.
(2S)-(2-Ethylbutyl) 2-((((2R,3R,4R,5R)-5-(4-Aminopyrrolo[1,2-
f ][1,2,4]triazin-7-yl)-5-cyano-3,4-dihydroxy-4-methyltetrahy-
drofuran-2-yl)methoxy)(phenoxy)phosphorylamino)-
propanoate (13). 1H NMR (300 MHz, CDC13) δ 8.02 (s, 1H), 7.1−
7.3 (m, 5H), 6.78 (2d, 1H), 6.65 (2d, 1H), 6.35 (brs, 1H), 6.02 (brs,
2H), 4.60 (m, 1H), 4.40 (m, 2H), 3.9−4.1 (m, 5H), 3.75 (brd, 1H),
1.50 (m, 1H), 1.3−1.4 (m, 7H), 0.86 (s, 9H). 31P NMR (121.4 MHz,
CDC13) δ 2.74, 2.75 (s). MS, m/e 617 (M + 1)+. HRMS, calculated
for C28H39N6O8P (M + 1)+, 617.2489; found, 617.2482.
(2S)-Benzyl 2-((((2R,3R,4R,5R)-5-(4-Aminopyrrolo[1,2-f ]-
[1,2,4]triazin-7-yl)-5-cyano-3,4-dihydroxy-4-methyltetrahydro-
furan-2-yl)methoxy)(phenoxy)phosphorylamino)propanoate
(14). 1H NMR (300 MHz, DMSO-d6) δ 7.91 (s, 1H), 7.85 (brs, 2H),
7.1−7.4 (m, 10H), 6.88 (2d, 1H), 6.77 (2d, 1H), 6.14 (m, 1H, NH),
5.95 (2s, 1H, 2′-OH), 5.41 (2d, 1H, 3′-OH), 5.02 (m, 2H, PhCH2),
4.38 (m, 1H, 5′-CH2), 4.21 (m, 1H, 5′-CH2), 4.18 (m, 1H, 4′-CH),
3.92 (m, 1H, NHCH), 3.65 (t, 1H, 3′-CH), 1.22 (q, 3H, CHCH3),
0.86 (s, 3H). 31P NMR (121.4 MHz, DMSO-d6) δ 3.84, 3.97 (s). MS,
m/e 623 (M + 1)+. HRMS, calculated for C29H32N6O8P (M + 1)+,
623.2019; found, 623.2017.
(2S)-Isopropyl 2-((((3aR,4R,6R,6aR)-6-(4-Aminopyrrolo[1,2-
f ][1,2,4]triazin-7-yl)-6-cyano-6a-methyl-2-oxotetrahydrofuro-
[3,4-d][1,3]dioxol-4-yl)methoxy)(phenoxy)phosphorylamino)-
propanoate (15). To a solution of compound 5 (100 mg, 0.174
mmol) in acetonitrile (10 mL) was added carbodiimidazole (60 mg,
0.370 mmol), and the reaction mixture was stirred at room
temperature for 6 h. Water (0.2 mL) was added, and the reaction
mixture was stirred for 0.5 h. The mixture was concentrated and
partitioned between ethyl acetate and water. The organic layer was
concentrated, and the residue was purified by silica gel column
chromatography (ethyl acetate−5% ethanol in ethyl acetate), affording
the product 15 (90 mg, 86%) as a solid. 1H NMR (300 MHz, CDC13)
δ 8.08 (s, 1H), 7.1−7.4 (m, 5H), 6.9 (2d, 1H), 6.7 (2d, 1H), 5.84 (brs,
2H), 5.01 (m, 1H), 4.8 (2d, 1H), 4.4−4.6 (m, 2H), 3.9−4.1 (m, 3H),
1.41 (d, 3H), 1.3 (m, 6H), 1.1 (2s, 3H). 31P NMR (121.4 MHz,
CDC13) δ 2.74, 2.88 (s). MS, m/e 601 (M + 1)+.
(2S)-Isopropyl 2-((((2R,3R,4R,5R)-3-Isobutyroxy-5-(4-
aminopyrrolo[1,2-f ][1,2,4]triazin-7-yl)-5-cyano-4-hydroxy-4-
methyltetrahydrofuran-2-yl)methoxy)(phenoxy)-
phosphorylamino)propanoate (16). To a solution of compound 5
(175 mg, 0.305 mmol) in acetonitrile (2 mL) was added N,N
-dimethyformamide dimethyl acetal (41 μL, 0.34 mmol, 1.1 equiv),
and the reaction mixture was stirred at room temperature for 1 h. The
reaction was complete (by LCMS). The mixture was then
concentrated to dryness. To the residue was added DCC (250 mg,
1.21 mmol, 4 equiv), acetonitrile (5 mL), and isobutyric acid (55 mg,
58 μL, 2 equiv). The mixture was stirred at room temperature for 48 h.
Water (0.2 mL) and trifluoroacetic acid (0.1 mL) were added at 0 °C,
and the reaction mixture was stirred at room temperature for 64 h.
Sodium bicarbonate (500 mg) was added at 0 °C. The mixture was
stirred at room temperature for 0.5 h and filtered. The filtrate was
concentrated, and the residue was purified by silica gel column
chromatography (5% methanol/dichloromethane), affording com-
pound 16 as a 1:1 diastereomeric mixture at phosphorus (144 mg,
73%). 1H NMR (300 MHz, CDC13) δ 8.00 (s, 1H), 7.1−7.4 (m, 5H),
6.83 (d, 1H), 6.71 (2d, 1H), 5.97 (brs, 2H), 5.94 (d, 1H), 5.07 (2d,
1H), 5.01 (m, 1H), 4.68 (m, 1H), 4.4 (m, 2H), 4.0 (m, 2H), 2.74 (m,
1H), 1.4 (2d, 3H), 1.2−1.3 (12H), 0.98 and 0.99 (2s, 3H). 31P NMR
To a solution of compound 2 (1.03 g, 3.37 mmol) in trimethyl
phosphate (2.0 mL) and THF (20 mL) was added N-methyl imidazole
(1.5 g, 18.3 mmol) at 0 °C. A solution of compound 25 (2.5 g, 8.18
mmol) in THF (3 mL) was dropwise added. The resulting mixture
was allowed to warm to room temperature over 1.5 h. The mixture was
partitioned between ethyl acetate and water. The ethyl acetate layer
was concentrated, and the residue was purified by silica gel
chromatography (ethyl acetate to 10% ethanol/ethyl acetate),
affording 1.15 g (59%) of compound 5 as a 1:1 diastereomeric
1
mixture at phosphorus. H NMR (300 MHz, CDC13) δ 8.02 (s, 1H),
7.1−7.4 (m, 5H), 6.8 (2d, 1H), 6.7 (2d, 1H), 6.08 (brs, 2H), 5.03 (m,
1H), 4.6 (m, 1H), 4.4 (m, 2H), 3.9−4.1 (m, 3H), 1.31 (d, 3H), 1.2
(m, 6H), 0.83 (s, 3H). 31P NMR (121.4 MHz, CDC13) δ 2.78 (s).
MS, m/e 575 (M + 1)+.
(2S)-Methyl 2-((((2R,3R,4R,5R)-5-(4-Aminopyrrolo[1,2-f ]-
[1,2,4]triazin-7-yl)-5-cyano-3,4-dihydroxy-4-methyltetrahydro-
furan-2-yl)methoxy)(phenoxy)phosphorylamino)propanoate
1
(6). H NMR (300 MHz, CDC13) δ 8.98 (s, 1H), 7.1−7.3 (m, 5H),
6.75 (2d, 1H), 6.60 (2d, 1H), 5.75 (brs, 2H), 4.55 (m, 1H), 4.35 (m,
2H), 3.6−4.1 (m, 3H), 3.62 (s, 3H), 1.31 (d, 3H), 0.83 (s, 3H). MS,
m/e 547 (M + 1)+.
(2S)-Ethyl 2-((((2R,3R,4R,5R)-5-(4-Aminopyrrolo[1,2-f ][1,2,4]-
triazin-7-yl)-5-cyano-3,4-dihydroxy-4-methyltetrahydrofuran-
1
2-yl)methoxy)(phenoxy)phosphorylamino)propanoate (7). H
NMR (500 MHz, DMSO-d6) δ 7.91 (s, 1H), 7.85 (brs, 2H), 7.1−7.4
(m, 5H), 6.90 (2d, 1H), 6.79 (2d, 1H), 6.11 (2dd, 1H, NH), 5.98 (s,
1H, 2′-OH), 5.45 (2d, 1H, 3′-OH), 4.41 (m, 1H, 5′-CH2), 4.24 (m,
1H, 5′-CH2), 4.20 (m, 1H, 4′-CH), 4.05 (m, 1H, 2H, OCH2), 3.86
(m, 1H, NHCH), 3.70 (t, 1H, 3′-CH), 1.25 (q, 3H, CHCH3), 1.14
(m, 3H, CH2CH3), 0.92 (s, 3H). 31P NMR (202.3 MHz, DMSO-d6) δ
3.94, 3.94 (s). MS, m/e 561 (M + 1)+. HRMS, calculated for
C24H30N6O8P (M + 1)+, 561.1863; found, 561.1871.
(2S)-(S)-sec-Butyl 2-((((2R,3R,4R,5R)-5-(4-Aminopyrrolo[1,2-
f ][1,2,4]triazin-7-yl)-5-cyano-3,4-dihydroxy-4-methyltetrahy-
drofuran-2-yl)methoxy)(phenoxy)phosphorylamino)-
propanoate (8). 1H NMR (300 MHz, CD3CN) δ 7.98 (s, 1H), 7.2−
7.4 (m, 5H), 6.82 (d, 2H), 6.45 (s, 2H), 5.28 (d, 1H, OH), 4.81 (m,
1H, -OCH), 4.3−4.5 (m, 4H), 3.93 (m, 2H), 3.82 (t, 1H, 3′-CH), 1.56
(m, 2H), 1.35 (d, 3H), 1.17 (q, 3H, CHCH3), 0.92 (s, 3H), 0.88 (m,
3H). 31P NMR (121.4 MHz, CD3CN) δ 3.19 (s). MS, m/e 589 (M +
1)+.
(2S)-(R)-sec-Butyl 2-((((2R,3R,4R,5R)-5-(4-Aminopyrrolo[1,2-
f ][1,2,4]triazin-7-yl)-5-cyano-3,4-dihydroxy-4-methyltetrahy-
drofuran-2-yl)methoxy)(phenoxy)phosphorylamino)-
1
propanoate (9). H NMR (300 MHz, DMSO-d6) δ 7.89 (s, 1H),
7.85 (brs, 2H), 7.1−7.4 (m, 5H), 6.87 (2d, 1H), 6.78 (2d, 1H), 6.04
(m, 1H, NH), 5.97 (s, 1H), 5.42 (brt, 1H), 4.68 (m, 1H, -OCH), 4.1−
4.5 (m, 3H), 3.82 (m, 1H), 3.65 (t, 1H, 3′-CH), 1.45 (m, 2H), 1.28 (d,
3H), 1.10 (d, 3H, CHCH3), 0.89 (s, 3H), 0.77 (m, 3H). 31P NMR
(121.4 MHz, DMSO-d6) δ 3.82, 4.01 (s). MS, m/e 589 (M + 1)+.
HRMS, calculated for C26H34N6O8P (M + 1)+, 589.2176; found,
589.2165.
(2S)-(tert-Butyl) 2-((((2R,3R,4R,5R)-5-(4-Aminopyrrolo[1,2-f ]-
[1,2,4]triazin-7-yl)-5-cyano-3,4-dihydroxy-4-methyltetrahydro-
furan-2-yl)methoxy)(phenoxy)phosphorylamino)propanoate
(10). 1H NMR (300 MHz, CDC13) δ 8.02 (s, 1H), 7.1−7.3 (m, 5H),
6.80 (2d, 1H), 6.70 (2d, 1H), 6.01 (brs, 2H), 4.60 (m, 1H), 4.40 (m,
2H), 3.93 (m, 3H), 3.62 (s, 3H), 1.43 (2s, 9H), 1.37 (d, 3H), 0.90 (2s,
3H). 31P NMR (121.4 MHz, CDC13) δ 2.89, 2.92 (s). MS, m/e 589
(M + 1)+. HRMS, calculated for C26H34N6O8P (M + 1)+, 589.2176;
found, 589.2180.
(2S)-cyclo-Pentyl 2-((((2R,3R,4R,5R)-5-(4-Aminopyrrolo[1,2-
f ][1,2,4]triazin-7-yl)-5-cyano-3,4-dihydroxy-4-methyltetrahy-
drofuran-2-yl)methoxy)(phenoxy)phosphorylamino)-
propanoate (11). 1H NMR (300 MHz, CDC13) δ 7.97 (s, 1H), 7.1−
7.3 (m, 5H), 6.72 (m, 2H), 6.40 (brs, 2H), 5.14 (m, 1H, OCH), 4.58
J
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
1
(121.4 MHz, CDC13) δ 2.56, 2.65 (s). MS, m/e 645 (M + 1)+. HRMS,
calculated for C29H38N6O9P (M + 1)+, 645.2432; found, 645.2434.
(2S)-Isopropyl 2-((4-Nitrophenoxy)(phenoxy)-
phosphorylamino)propanoate (28). To a suspension of L-alanine
isopropyl ester hydrochloride (7.95 g, 47.4 mmol) in dichloromethane
(100 mL) was added phenyl phosphorodichloridate (10 g, 47.4
mmol). After cooling to −20 °C, triethylamine (13.2 mL, 95 mmol)
was dropwise added over a period of 15 min (internal reaction
temperature; −10 °C ∼ −3 °C). When the reaction was almost
complete (by phosphorus NMR), p-nitrophenol (6.29 g, 45.0 mmol)
was added as a solid in one portion. Additional triethylamine (6.28
mL, 45 mmol) was added over a period of 15 min. The mixture was
then warmed up to room temperature. When the reaction was
complete, MTBE (100 mL) was added. The white precipitate was
removed by filtration. The filter cake was washed with MTBE (3 × 50
mL). The filtrate and the washings were combined and concentrated.
The residue was purified by silica gel column chromatography (0−50%
ethyl acetate/hexanes), affording compound 28 as a 1:1 ratio of a
material instead of 5. H NMR (300 MHz, CDC13) δ 8.00 (s, IH),
7.1−7.3 (m, 5H), 6.83 (d, IH), 6.71 (d, IH), 6.09 (brs, 2H), 5.95 (s,
IH), 5.04 (m, 2H), 4.67 (q, IH), 4.35−4.52 (m, 2H), 4.00 (m, 2H),
2.74 (m, IH), 1.40 (d, 3H), 1.2−1.3 (12H), 0.98 (s, 3H). 31P NMR
(121.4 MHz, CDC13) δ 2.72 (s). HRMS, calculated for C29H38N6O9P
(M + 1)+, 645.2432; found, 645.2434. The compound was
subsequently recrystallized from MTBE for X-ray quality crystals.
HCV Replicon Antiviral and Cytotoxicity Assays. All
compounds were supplied in 100% DMSO. Compound serial dilutions
were performed in 100% DMSO. For EC50 and CC50 determinations,
test compounds were serially diluted in ten steps of 1:3 dilutions in
384-well plates. All serial dilutions were performed in four replicates
per compound within the same 384-well plate. An HCV protease
inhibitor ITMN-191 at 100 μM was added as a control of 100%
inhibition of HCV replication while puromycin at 10 mM was included
as a control of 100% cytotoxicity. To each well of a black polystyrene
384-well plate (Greiner Bio-one, Monroe, NC), 90 μL of cell culture
medium (without Geneticin) containing 2000 suspended HCV
replicon cells was added with a Biotek μFlow workstation. For
compound transfer into cell culture plates, 0.4 μL of compound
solution from the compound serial dilution plate was transferred to the
cell culture plate on a Biomek FX workstation. The DMSO
concentration in the final assay wells was 0.44%. The plates were
incubated for 3 days at 37 °C with 5% CO2 and 85% humidity. The
HCV replicon assay was a multiplex assay, able to assess both
cytotoxicity and antireplicon activity from the same well. The CC50
assay was performed first. The media in the 384-well cell culture plate
was aspirated, and the wells were washed four times with 100 μL of
PBS each, using a Biotek ELX405 plate washer. A volume of 50 μL of a
solution containing 400 nM calcein AM (Anaspec, Fremont, CA) in 1
× PBS was added to each well of the plate with a Biotek μFlow
workstation. The plate was incubated for 30 min at room temperature
before the fluorescence signal (excitation 490 nm, emission 520 nm)
was measured with a Perkin-Elmer Envision plate reader.
1
diasteromeric mixture (14.1 g, 77%). H NMR (300 MHz, CDCl3): δ
8.22 (2d, 2H), 7.2−7.4 (m, 7H), 5.0 (m, 1H), 4.09 (m, 1H), 3.96 (m,
1H), 1.39 (2d, 3H), 1.22 (m, 6H). MS, m/e 409 (M + 1)+.
Separation of the Two Diastereomers of 28. The two
diastereomers were separated by chiral column chromatography
under the following conditions: column, Chiralpak IC, 2 cm × 25
cm; solvent system, 70% heptane and 30% isopropanol (IPA); flow
rate, 6 mL/min; loading volume per run, 1.0 mL; concentration of
loading sample, 150 mg/mL in 70% hepane and 30% IPA. (Sp)-28:
retention time 43 min. 31P NMR (161.9 MHz, CDCl3): δ −2.99 (s).
(Rp)-28: retention time 62 min. 31P NMR (161.9 MHz, CDCl3): δ
−3.02 (s).
Alternatively, the two diasteromers were separated by crystallization
under the following procedures:
Diastereomeric mixture 28 was dissolved in diethyl ether (∼10 mL/
g). While stirring, hexanes was then added until the solution became
turbid. Seed crystals (∼10 mg/g of compound C) were added to
promote crystallization. The resulting suspension was gently stirred for
16 h, cooled to ∼0 °C, stirred for an additional 2 h, and filtered to
collect the crystalline material (recovery yield of the crystalline
material ∼35%). The crystalline material contains ∼95% of (Sp)-28
and ∼5% of (Rp)-28. Recrystallization afforded >99% diastereomeri-
cally pure (Sp)-28.
The EC50 assay was performed in the same wells as the CC50 assay.
The calcein-PBS solution in the 384-well cell culture plate was
aspirated with a Biotek ELX405 plate washer. A volume of 20 μL of
Dual-Glo luciferase buffer (Promega, Madison, WI) was added to each
well of the plate with a Biotek μFlow Workstation. The plate was
incubated for 10 min at room temperature. A volume of 20 μL of a
solution containing a 1:100 mixture of Dual-Glo Stop & Glo substrate
(Promega, Madison, WI) and Dual-Glo Stop & Glo buffer (Promega,
Madison, WI) was added to each well of the plate with a Biotek μFlow
Workstation. The plate was then incubated at room temperature for 10
min before the luminescence signal was measured with a Perkin-Elmer
Envision Plate Reader.
(Sp)-Isopropyl 2-((S)-(((2R,3R,4R,5R)-5-(4-aminopyrrolo[1,2-
f ][1,2,4]triazin-7-yl)-5-cyano-3,4-dihydroxy-4-methyltetrahy-
drofuran-2-yl)methoxy)(phenoxy)phosphorylamino)-
propanoate (17). To a dry, argon purged round-bottom flask were
added compound 2 (1.0 g, 3.28 mmol) and anhydrous THF (15 mL).
The slurry was stirred for 10 min, and the flask was place in a water
bath at room temperature. t-Butylmagnesium chloride in THF (1.0 M,
4.91 mL) was dropwise added, and the mixture was stirred for an
additional 10 min. A solution of (Sp)-28 (2.68 g, 6.55 mmol) in THF
(10 mL) was then added. The flask was place in a heating oil bath
preset at 50 °C, and the mixture was stirred until compound 2 was
almost consumed. After ∼2.5 h, the reaction mixture was cooled to
room temperature, and methanol (5 mL) was added. Solvents were
removed under reduced pressure, and the residue was purified by silica
gel column chromatography (70% ethyl acetare/hexanes to remove
less polar impurities, 10% methanol/dichloromethane to elute the
Data Analysis. The cytotoxicity effect was determined by calcein
AM conversion to fluorescent product. The percent cytotoxicity was
calculated by eq 1:
⎛
⎞
⎟
⎠
XC − MB
MD − MB
%cytotoxicity or %inhibition = 100 × 1 −
⎜
⎝
(1)
where XC is the fluorescence signal from compound treated well; MB is
the average fluorescence signal from puromycin treated wells; and MD
is the average fluorescence signal from DMSO treated wells. The
percent anti-HCV replication activity was determined by the
luminescence signal generated from the renilla luciferase reporter of
the HCV replicon. The percent inhibition of HCV replicon was
calculated using eq 1, where XC is the luminescence signal from the
compound treated well; MB is the average luminescence signal from
ITMN-191 treated wells; and MD is the average luminescence signal
from DMSO treated wells.
The CC50 values were determined as the testing compound
concentration that caused a 50% decrease of cell viability. The EC50
values were determined as the testing compound concentration that
caused a 50% decrease in HCV replication. Both CC50 and EC50 values
were obtained using the Pipeline Pilot 5.0 software package (Accelrys,
San Diego, CA) by nonlinear regression fitting of experimental data to
eq 2:
1
product), affording 17 as an off-white solid (1.45 g, 77%). H NMR
(400 MHz, DMSO-d6): δ 7.89 (s, 1H), 7.84 (brs, 2H), 7.36 (t, 2H),
7.23 (d, 2H), 7.17 (t, 1H), 6.87 (d, J = 4.4 Hz, 1H), 6.74 (d, J = 4.4
Hz, 1H), 6.02 (dd, 1H), 5.96 (s, 1H), 5.41 (d, 1H), 4.82 (m, 1H), 4.38
(dd, 1H), 4.22 (q, 1H), 4.16 (m, 1H), 3.81 (m, 1H), 3.67 (dd, 1H),
1.22 (d, 3H), 1.11 (dd, 6H), 0.89 (s, 3H). 31P NMR (161.9 MHz,
DMSO-d6): δ 3.99 (s). MS, m/e 575 (M + 1)+, 573 (M − 1)−. HRMS,
calculated for C25H32N6O8P (M + 1)+, 575.2014; found, 575.2022.
(Sp)-Isopropyl 2-((S)-(((2R,3R,4R,5R)-3-isobutyroxy-5-(4-
aminopyrrolo[1,2-f ][1,2,4]triazin-7-yl)-5-cyano-4-hydroxy-4-
methyltetrahydrofuran-2-yl)methoxy)(phenoxy)-
phosphorylamino)propanoate (18). Compound 18 was prepared
by following the procedure for 16 except using 17 as the starting
K
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a − d
during compound incubation and at 0, 0.5, 1, 3, 6, 8, and 23 h
following compound removal. Cells were collected, and intracellular
metabolites were analyzed as described for replicon cells.
y = d +
b
⎡
⎤
x
1 +
⎢
⎣
(c )
⎥
⎦
(2)
Caco-2 Permeability. Bidirectional permeability studies were
done using confluent conolayers of the human colon carcinoma cell
line (caco-2). Caco-2 cells were grown to confluence over at least 21
days on 24 well PET (polyethylene-terephthalate) plates (BD
Biosciences). After an initial equilibration with transport buffer,
TEER values were read to test membrane integrity. The experiment
was started by the addition of buffers containing test compound, and
the samples were taken at 1 and 2 h from the receiver compartment.
All samples were immediately precipitated by 4 volumes of 100%
acetonitrile containing internal standard and analyzed by LC/MS/MS.
The compound was tested in 2 separate replicate wells for each
condition, and the permeability of control compounds (atenolol,
propranolol, and vinblastine) was determined to meet acceptance
criteria for each batch of assay plates. Cells were dosed on the apical
(A) or basolateral (B) side to determine forward (A to B) and reverse
(B to A) permeability. Permeability through a cell free trans-well was
also determined as a measure of cellular permeability through the
membrane. To test for nonspecific binding and compound instability,
the total amount of drug was quantified at the end of the experiment
and compared to the material present in the original dosing solution as
a percent recovery.
where y is the observed %inhibition of the HCV replicon at x
concentration of compound; d is the estimated response at zero
compound concentration; a is the estimated response at infinite
compound concentration; c is the midrange concentration (CC50 or
EC50); and b is the Hill slope factor.
HCV Polymerase (NS5B) Enzyme Assay. Inhibition of NS5B
was studied using GT1b NS5B recombinant protein. All concen-
trations are final concentrations. In a solution containing 50 mM Tris-
HCl (pH 7.5), 10 mM KCl, 1 mM DTT, 5 mM MgCl2, 0.2 unit/μL
RNAsin (Promega), and 0.01% BSA, 75 nM NS5B was preincubated
with 4 ng/μL (50 nM) heteropolymer RNA template (sshRNA)
(reference: Hung M.; Gibbs, C. S.; Tsiang, M. Biochemical
characterization of rhinovirus RNA-dependent RNA polymerase.
Antiviral Res. 2002, 56 (2), 99−114) at RT for 5 min, followed by
the addition and incubation of the inhibitors at RT for 5 min. The
reaction was started by addition of a mixture of ATP (2.5 μM), CTP
(2.5 μM), UTP (2.5 μM), 0.03 μM 33P-labeled GTP (Perkin-Elmer,
NEG603H, 3000 Ci/mmol, 3.3 μM), and 1.25 μM GTP. After 90 min
of incubation at 30 °C, the reaction mixture was spotted on DE-81
filter paper and washed with 0.125 mM Na2HPO4 (3×), distilled water
(1×), and reagent ethanol (1×). The filter paper was air-dried and
exposed to phosphoimager (Typhooh, GE), and the data was analyzed
using ImageQuant software (GE).
The apparent permeability, Papp, and % recovery were calculated as
follows:
P
= (dR/dt)V /(AD0)
app
r
Single Nucleotide Incorporation by Mitochondrial RNA
Polymerase (POLRMT). A mixture of MTCN buffer (50 mM
MES, 25 mM Tris-HCl, 25 mM CAPS, and 50 mM NaCl, pH 7.5),
200 nM 5′-32P-R12/D18 (reference Arnold 2012 PLoSPathogen), 10
mM MgCl2, 1 mM DTT, and 376 nM POLRMT (Enzymax,
Lexington, KY) was preincubated at 30 °C for 1 min. The reaction
was started by addition of 500 μM (final) natural NTP or NTP
analogs. At selected time points, the reaction mixture was removed and
quenched with gel loading buffer containing 100 mM EDTA, 80%
formamide, and bromophenol blue, and heated at 65 °C for 5 min.
The samples were run on a 20% polyacrylamide gel (8 M urea), and
the product formation was quantified using Typhoon Trio Imager and
Image Quant TL Software (GE Healthcare, Piscataway, NJ). The rate
of single nucleotide incorporation by mt RNA pol was calculated by
fitting the product formation using the single exponential equation:
[R13] = A(1 − e−kt), where [R13] represents the amount (in nM) of
the elongated product formed, t represents the reaction time, k
represents the observed rate, and A represents the amplitude of the
exponential.
Measurements of Triphosphate Levels in Replicon Cells.
Replicon cells (Huh-7 cells) were maintained in Dulbecco’s modified
eagle medium containing glutamax supplemented with 10% heat
inactivated fetal bovine serum, penicillin-streptomycin, and G418
disulphate salt solution. Cells were transferred to 12-well tissue culture
treated plates by trypsonization and grown to confluency (0.88 × 106
cells/well). Cells were treated for 24 h with a test compound (10 μM).
Following 24 h, cells were washed twice with 2.0 mL of ice cold 0.9%
sodium chloride saline. Cells were then scraped into 0.5 mL of 70%
methanol and frozen overnight to facilitate the extraction of nucleotide
metabolites. Extracted cell material in 70% methanol was transferred
into tubes and dried. After drying, samples were resuspended in 1 mM
ammonium phosphate pH 8.5. TP levels were quantified using liquid
chromatography coupled to triple quadrapole mass spectrometry by
methods similar to those previously reported in Durand-Gasselin, L.;
Van Rompay, K. K.; Vela, J. E.; Henne, I. N.; Lee, W. A.; Rhodes, G. R.
Mol. Pharm. 2009, 6, 1145.
%Recovery = 100((VR120) + (V D120))/(V D0)
r
d
d
where dR/dt is the slope of the cumulative concentration in the
receiver compartment versus time in μM/s based on receiver
concentrations measured at 60 and 120 min, Vr and Vd are the
volume in the receiver and donor compartments in cm3, respectively,
A is the area of the cell monolayer (0.33 cm2), D0 and D120 are the
measured donor concentrations at the beginning and end of the
experiment, respectively, and R120 is the receiver concentration at the
end of the experiment (120 min).
Intestinal and Hepatic S9 Stability. The compounds were
incubated at 2 μM in human intestinal and hepatic S9 fractions
(obtained from In Vitro Technologies, Baltimore, MD) for 90 min at
37 °C in the presence of NADPH and UDPGA (phase I and phase II
cofactor, Sigma-Aldrich). At specified time points following compound
addition, samples were quenched with nine volumes of an aqueous
solution containing internal standard, 50% acetonitrile, and 25%
methanol. Sample plates were centrifuged at 3000g for 30 min, and 10
μL of the resulting solution was analyzed by LC/MS/MS. Data
(sample to internal standard peak area ratio) were plotted on a semi
log scale and fitted using an exponential fit. Assuming first-order
kinetics, the half-life and rate of metabolism were determined.
Predicted hepatic extraction was calculated from the half-life by
reported methods using the well-stirred model for hepatic clearance.
LC/MS/MS Instrumentation. S9 stability and caco-2 assay: Liquid
chromatography was performed using an Agilent 1200-series
quaternary pump system (Agilent Technologies, Santa Clara, CA)
with a 2 μm 20 × 2.1 mm Mercury RP C18 (Phenomenex, Torrance,
CA) for S9 stability assay samples, and a 1.7 μm 50 × 2.1 mm Acquity
UPLC BEH C18 (Waters, Milford, MA) for caco-2 assay samples. A
HPLC system was coupled to a Quattro Premier triple-quadrupole
mass spectrometer (Waters, Milford, MA). Mass spectrometry was
performed in positive-ion mode and multiple reaction monitoring
modes using a Quattro Premier (Waters, Milford, MA). The test
compound was eluted with a mobile phase consisting of 0.2% formic
acid and a linear gradient from 0 to 95% acetonitrile over 2 min.
log D Measurement. A solution of test compound in DMSO (2
μL of 10 mM) was added into a 96-well plate containing 198 μL of 1:1
acetonitrile/water. The sample plate was shaken for 30 min, and 10 μL
of the resulting solution was analyzed by HPLC/UV. The
instrumentation used was an Alliance 2795 HPLC coupled with a
photodiode array detector 2996 (Waters, Milford, MA) using a Waters
Measurement of Triphosphate Levels in Primary Human
Hepatocytes. Primary human hepatocytes were maintained in
Williams E medium containing Cellz Direct’s proprietary supplement
cocktail. Cells were purchased from Cellz Direct as a twelve well plate
format grown to confluency (0.88 × 106 cells/well). Cells were treated
for 1 h with a test compound (10 μM) followed by a 23 h treatment
with nondrug containing media. Cells were collected at 0, 0.5, and 1 h
L
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
XTerra 3.5 μm 4.6 mm × 50 mm C18 column. The mobile phase
consisted of solvent A (20 mM ammonium acetate aqueous solution)
and solvent B (100% acetonitrile). Elution was performed using a
linear gradient of solvent B from 0% to 100% in 8 min. The log D
value was calculated using the retention time of test compound
compared to the reference compounds. The range of log D values for
reference compounds is approximately between 0.3 and 5.7.
Aqueous Solubility. The aqueous solubility of the test
compounds was determined at ambient temperature using pH 2
hydrochloric acid and pH 7 phosphate buffer. Excess solid drug
substance was used to saturate the solution media. The resultant
suspensions were sonicated using a sonication bath (VWR model
250T, VWR International LLC, Radnor, PA, or Branson model 2510,
Branson Ultrasonic Corporation, Danbury, CT) and were then stirred
for at least 2 h at ambient temperature using a stir plate (IKAMAG,
IKA Works, Inc., Wilmington, NC) or a Thermomixer (Eppendorf
AG, Hamburg, Germany). The samples were subsequently centrifuged
(Eppendorf Centrifuge, Eppendorf AG, Hamburg, Germany), and the
supernatant was assayed for test compound using a standard reverse-
phase UPLC-UV method with an ACE C18, 50 mm × 2.1 mm, 3 μm
particle column (Advanced Chromatography Technologies, Aberdeen,
U.K.).
oral bioavailability
%F = AUCpo × Doseiv/AUCiv × Dosepo
Liver triphosphate levels were determined by a method similar to that
described in hamster PK analysis.
ASSOCIATED CONTENT
* Supporting Information
■
S
Tables for pan-genotypic activity of 18 (Table S1), resistant
profile of 16 (Table S2), and in vitro activity of 16 in
combination with other classes of anti-HCV agents (Table S3).
This material is available free of charge via the Internet at
http://pubs.acs.org.
Accession Codes
The Cambridge Crystallographic Data Center (CCDC)
numbers for X-ray structures of compounds (Sp)-28 and 18
are 929320 and 929321, respectively.
AUTHOR INFORMATION
Corresponding Author
*Phone: 650-522-5177. Fax: 650-5225899. E-mail: Aesop.
Cho@gilead.com.
■
Hamster Liver Triphosphate Analysis. Groups of 12 Golden
Syrian hamsters were administered a test compound intravenously at 1
mg-equiv/kg and orally at 5 mg-equiv/kg. Formulation consisted of
EtOH, PEG400, and water and was adjusted to pH 4 by HCl (0.5 mg-
equiv/mL) for iv dosing and buffered with 50 mM pH 4 citrate (2.5
mg-equiv/mL) for oral dosing. At 1, 4, 8, and 12 h, livers were
harvested under isoflurane anesthesia. Collected livers were wrapped
in aluminum foil and snap-frozen in liquid nitrogen immediately
following removal to avoid sample dephosphorylation. Livers were
processed by sectioning into smaller pieces with a razor blade and
collecting into preweighed 15 mL conical tubes kept on dry ice. Four
volumes of ice-cold extraction buffer (0.1% KOH and 67 mM EDTA
in 70% MeOH, containing 0.25 μM Cl-ATP) were added, and samples
were promptly homogenized using an Omni-Tip TH with disposable,
hard tissue homogenizer probes (Omni International). Aliquots of
homogenate were then centrifuged (20,000g at 4 °C for 10 min).
Aliquots of supernatant were transferred to clean tubes, evaporated to
dryness in a heated centrifugal evaporator, and reconstituted with an
equal volume of 1 mM ammonium phosphate (pH 7.0). Intracellular
triphosphate levels were measured by LC/MS/MS as described for
samples generated in vitro. Levels of endogenous adenosine
triphosphate were also determined to ensure the sample stability.
Dog in Vivo Pharmacokinetic Analysis. Groups of three non-
Notes
The authors declare the following competing financial
interest(s): The authors are employees of Gilead Sciences
(except W. Zhong, S. Rossi, M. Fenaux, and C. Kim, who were
employed at Gilead Sciences during this work). All authors are
shareholders in Gilead Sciences.
ACKNOWLEDGMENTS
■
We thank Amy Kowk for assisting with chiral column
separation, Tim S. Lane for obtaining HRMS, and Arnold
Rheingold (UCSD) for X-ray crystal structures.
ABBREVIATIONS USED
■
HCV, hepatitis C virus; NI, nucleoside inhibitor; NS5B,
nonstructural protein 5B; SOC, standard of care; Peg-IFN,
pegylated interferon-α; RBV, ribavirin; DAA, direct-acting
antiviral; NOAEL, no-observable adverse effects level;
POLRMT, mitochondrial RNA polymerase; SVR, sustained
virologic response; DMPK, drug metabolism and pharmacoki-
netics; GT, genotype; NS5A, nonstructural protein 5A; NS3,
nonstructural protein 3; BID, twice daily; PK/PD, pharmaco-
kinetic/pharmacodynamic; MTCN, MES-TRIS-CAPS-NaCl;
MES, 2-(N-morpholino)ethanesulfonic acid; TRIS, tris-
(hydroxymethyl)aminomethane; CAPS, N-cyclohexyl-3-amino-
propanesulfonic acid; NTP, nucleoside triphosphate; NADPH,
nicotinamide adenine dinucleotide phosphate; UDPGA, uridine
5′-diphospho-glucuronic acid; Cl-ATP, 2-choloroadenosine
triphosphate
̈
naive male beagle dogs were administered a single dose of a test
compound by intravenous infusion at 0.5 mg/kg or orally at 5 mg/kg.
The intravenous dosing was done in 5/20/75 (EtOH/PEG 400/water,
pH 4 w/HCl), and the oral formulation was made up in 5/40/55
(EtOH/PEG 400/50 mM citrate buffer, pH 4). Blood samples were
collected over a 24 h period postdose into Vacutainer tubes containing
EDTA-K3 (BD Biosciences), and plasma was isolated by the
manufacturer’s suggested protocol. Plasma samples were processed
via protein precipitation by adding acetonitrile to a final concentration
of 60%. Following filtration to remove precipitated proteins, samples
were dried and reconstituted in 20% acetonitrile in water. Samples
were then analyzed by reversed phase liquid chromatography coupled
to a triple quadrapole mass spectrometer (LC/MS/MS) in positive ion
and multiple reaction monitoring modes. Percent oral absorption (%
Fa) was determined as follows:
REFERENCES
■
(1) Lavanchy, D. Evolving epidemiology of hepatitis C virus. Clin.
Microbiol. Infect. 2011, 17, 107−115.
(2) Zeuzem, S.; Berg, T.; Moeller, B.; Hinrichsen, H.; Mauss, S.;
Wedemeyer, H.; Sarrazin, C.; Hueppe, D.; Zehnter, E.; Manns, M. P.
Expert opinion on the treatment of patients with chronic hepatitis C. J.
Viral Hepatitis 2009, 16, 75−90.
% oral absorption
%F = %F/(1 − EH)
a
(3) (a) Vachon, M.-L.; Dieterich, D. T. The era of direct-acting
antivirals has begun: The beginning of the end for HCV. Semin. Liver
Dis. 2011, 31, 399−409. (b) Kronenberger, B.; Zeuzem, S. New
developments in HCV therapy. J. Viral Hepatitis 2012, 19, 48−51.
hepatic extraction
EH = (AUCportal − AUCsystemic)/AUCportal
M
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(4) Manns, M. P.; Foster, G. R.; Rockstroh, J. K.; Zeuzem, S.;
Zoulim, F.; Houghton, M. The way forward in HCV treatment -
finding the right path. Nat. Rev. Drug Discovery 2007, 6, 991−1000.
(5) Stedman, C. A. M. Current prospects for interferon-free
treatment of hepatitis C in 2012. J. Gastroenterol. Hepatology 2013,
28 (1), 38−45.
(15) Cho, A.; Saunders, O. L.; Butler, T.; Zhang, L.; Xu, J.; Vela, J. E.;
Feng, J. Y.; Ray, A. S.; Kim, C. U. Synthesis and antiviral activity of a
series of 1′-substituted 4-aza-7,9-dideazaadenosine C-nucleosides.
Bioorg. Med. Chem. Lett. 2012, 22, 2705−2707.
(16) Lawitz, E.; Hill, J.; Marbury, T.; Hazan, L.; Gruener, D.;
Webster, L.; Majuskas, R.; Morrison, R.; DiMicco, M.; German, P.;
Stefanidis, D., Svaroskaia, E.; Jain, A.; Arterburn, S.; Ray, A.; Rossi, S.;
McHutchison, J.; Rodriguez-Torres, M. GS-6620, A Liver-Targeted
Nucleotide Prodrug, Exhibits Antiviral Activity and Favorable Safety
(6) Sofia, M. J. Nucleotide prodrugs for HCV therapy. Antiviral
Chem. Chemother. 2011, 22, 23−49.
(7) (a) Mericitabine (R7128): Ali, S.; Leveque, V.; Pogam, S. L.; Ma,
H.; Philipp, F.; Inocencio, N.; Smith, M.; Alker, A.; Kang, H.; Najera,
I.; Klumpp, K.; Symons, J.; Cammack, N.; Jiang, W.-R. Selected
Replicon Variants with Low-Level In Vitro Resistance to the Hepatitis
C Virus NS5B Polymerase Inhibitor PSI-6130 Lack Cross-Resistance
with R1479. Antimicrob. Agents Chemother. 2008, 52, 4356−4369. (b)
BMS-986094 (INX-08189): McGuigan, C.; Madela, K.; Aljarah, M.;
Gilles, A.; Brancale, A.; Zonta, N.; Chamberlain, S.; Vernachio, J.;
Hutchins, J.; Hall, A.; Ames, B.; Gorovits, E.; Ganguly, B.; Kolykhalov,
A.; Wang, J.; Muhammad, J.; Patti, J. M.; Henson, G. Design, synthesis
and evaluation of a novel double pro-drug: INX-08189. A new clinical
candidate for hepatitis C virus. Bioorg. Med. Chem. Lett. 2010, 20,
4850−4854. (c) IDX-184: Zhou, X.-J.; Pietropaolo, K.; Chen, J.;
Khan, S.; Sullivan-Bolyai, J.; Mayers, D. Safety and Pharmacokinetics
of IDX184, a Liver-Targeted Nucleotide Polymerase Inhibitor of
Hepatitis C Virus in Healthy Subjects. Antimicrob. Agents Chemother.
2011, 55, 76−81.
(8) GS-7977 (PSI-7977): Sofia, M. J.; Bao, D.; Chang, W.; Du, J.;
Nagarathnam, D.; Rachakonda, S.; Reddy, P. G.; Ross, B. S.; Wang, P.;
Zhang, H.-R.; Bansal, S.; Espiritu, C.; Keilman, M.; Lam, A. M.; Steuer,
M.; Niu, C.; Otto, M. J.; Furman, P. S. Discovery of a β-D-2′-Deoxy-
2′-α-fluoro-2′-β-C-methyluridine Nucleotide Prodrug (PSI-7977) for
the Treatment of Hepatitis C Virus. J. Med. Chem. 2010, 53, 7202−
7218.
(9) (a) Jenkins, S. R.; Arison, B.; Walton, E. Branched-chain sugar
nucleosides. J. Org. Chem. 1968, 33, 2490−2494. (b) Walton, E.;
Jenkins, S. R.; Nutt, R. F.; Holly, F. W. Branched-chain sugar
nucleosides. V. Synthesis and antiviral properties of several branched-
chain sugar nucleosides. J. Med. Chem. 1969, 12, 306−309.
(10) Carroll, S. S.; Tomassini, J. E.; Bosserman, M.; Getty, K.;
Stahlhut, M. W.; Eldrup, A. B.; Bhat, B.; Hall, D.; Simcoe, A. L.;
LaFemina, R.; Rutkowski, C. A.; Wolanski, B.; Yang, Z.; Migliaccio, G.;
Francesco, R. D.; Kuo, L. C.; MacCoss, M.; Olsen, D. B. Inhibition of
Hepatitis C Virus RNA Replication by 2′-Modified Nucleoside
Analogs. J. Biol. Chem. 2003, 278, 11979−11984.
(11) Migliaccio, G.; Tomassini, J. E.; Carroll, S. S.; Tomei, L.;
Altamura, S.; Bhat, B.; Bartholomew, L.; Bosserman, M. R.; Ceccacci,
A.; Colwell, L. F.; Cortese, R.; Francesco, R. D.; Eldrup, A. B.; Getty,
K. L.; Hou, Z. S.; LaFemina, R. L.; Ludmerer, S. W.; MacCoss, M.;
McMasters, D. R.; Stahlhut, M. W.; Olsen, D. B.; Hazuda, D. J.; Flores,
O. A. Characterization of Resistance to Non-obligate Chain-
terminating Ribonucleoside Analogs That Inhibit Hepatitis C Virus
Replication in Vitro. J. Biol. Chem. 2003, 278, 49164−49170.
(12) Cho, A.; Zhang, L.; Xu, J.; Babusis, D.; Butler, T.; Lee, R.;
Saunders, O. L.; Wang, T.; Parrish, J.; Perry, J.; Feng, J. Y.; Ray, A. S.;
Kim, C. U. Synthesis and characterization of 2’-C-Me branched C-
nucleosides as HCV polymerase inhibitors. Bioorg. Med. Chem. Lett.
2012, 22, 4127−4132.
̈
Profile over 5 days in Treatment Naive Chronic HCV Genotype 1
Subjects. EASL, 2012; poster 1188.
(17) Bushnell, D. A.; Cramer, P.; Kornberg, R. D. Structural basis of
transcription: α-Amanitin−RNA polymerase II cocrystal at 2.8 Å
resolution. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 1218−1222.
(18) Arnold, J. J.; Sharma, S. D.; Feng, J. Y.; Ray, A. S.; Smidansky, E.
D.; Kireeva, M. L.; Cho, A.; Perry, J.; Vela, J. E.; Park, Y.; Xu, Y.; Tian,
Y.; Babusis, D.; Barauskus, O.; Peterson, B. R.; Gnatt, A.; Kashlev, M.;
Zhong, W.; Cameron, C. E. Sensitivity of mitochondrial transcription
and resistance of RNA polymerase II dependent nuclear transcription
to antiviral ribonucleosides. PLoS Pathog. 2012, 8 (11), e1003030.
(19) Ray, A. S; Karl Y. Hostetler, K. Y. Application of kinase bypass
strategies to nucleoside antivirals. Antiviral Res. 2011, 92, 277−291.
(20) Gane, E. J.; Stedman, C. A.; Hyland, R. H.; Ding, X.;
Svarovskaia, E.; Symonds, W. T.; Hindes, R. G.; Berrey, M. M.
Nucleotide polymerase inhibitor sofosbuvir plus ribavirin for hepatitis
C. N. Engl. J. Med. 2013, 368 (1), 34−44.
(21) A similar approach was reported in: Hecker, S. J.; Reddy, K. R.;
van Poelje, P. D.; Sun, Z.; Huang, W.; Varkhedkar, V.; Reddy, M. V.;
Fujitaki, J. M.; Olsen, D. B.; Koeplinger, K. A.; Boyer, S. H.;
Linemeyer, D. L.; MacCoss, M.; Erion, M. D. Liver-Targeted Prodrugs
of 2′-C-Methyladenosine for Therapy of Hepatitis C Virus Infection. J.
Med. Chem. 2007, 50, 3891−3896.
(22) For our early work on this convergent synthesis, see: Metobo, S.
E.; Xu, J.; Saunders, O. L.; Butler, T.; Aktoudianakis, E.; Cho, A.; Kim,
C. U. Practical synthesis of 1’-substituted Tubercidin C-nucleoside
analogs. Tetrahedron Lett. 2012, 53, 484.
(23) Gardelli, C.; Attenni, B.; Donghi, M.; Meppen, M.; Pacini, B.;
Harper, S.; Di Marco, A.; Fiore, F.; Giuliano, C.; Pucci, V.; Laufer, R.;
Gennari, N.; Marcucci, I.; Leone, J. F.; Olsen, D. B.; MacCoss, M.;
Rowley, M.; Narjes, F. Phosphoramidate Prodrugs of 2′-C-
Methylcytidine for Therapy of Hepatitis C Virus Infection. J. Med.
Chem. 2009, 52, 5394−5407.
(24) Independent from our work, a similar approach was recently
published Ross, B. S.; Reddy, P. G.; Zhang, H.-R.; Rachakonda, S.;
Sofia, M. S. Synthesis of Diastereomerically Pure Nucleotide
Phosphoramidates. J. Org. Chem. 2011, 76, 8311−8319.
(13) Olsen, D. B.; Eldrup, A. B.; Bartholomew, L.; Bhat, B.;
Bosserman, M. R.; Ceccacci, A.; Colwell, L. F.; Fay, J. F.; Flores, O. A.;
Getty, K. L.; Grobler, J. A.; LaFemina, R. L.; Markel, E. J.; Migliaccio,
G.; Prhavc, M.; Stahlhut, M. W.; Tomassini, J. E.; MacCoss, M.;
Hazuda, D. J.; Carroll, S. S. A 7-Deaza-Adenosine Analog Is a Potent
and Selective Inhibitor of Hepatitis C Virus Replication with Excellent
Pharmacokinetic Properties. Antimicrob. Agents Chemother. 2004, 48,
3944−3953.
(14) Carroll, S. S.; Ludmerer, S.; Handt, L.; Koeplinger, K.; Zhang,
N. R.; Graham, D.; Davies, M.-E.; MacCoss, M.; Hazuda, D.; Olsen, D.
B. Robust Antiviral Efficacy upon Administration of a Nucleoside
Analog to Hepatitis C Virus-Infected Chimpanzees. Antimicrob. Agents
Chemother. 2009, 53, 926.
N
dx.doi.org/10.1021/jm400201a | J. Med. Chem. XXXX, XXX, XXX−XXX