Discovery of silyl proline containing HCV NS5A inhibitors with pan-genotype activity: SAR development

Article abstract of DOI:10.1016/j.bmcl.2016.01.050

HCV NS5A inhibitors have shown impressive in vitro potency profiles in HCV replicon assays thus making them attractive components for inclusion in an all oral fixed dose combination treatment regimen. Herein we describe the research efforts that led to the discovery of silyl proline containing HCV NS5A inhibitors such as 7e and 8a with pan-genotype activity profile and acceptable pharmacokinetic properties.

Full text of DOI:10.1016/j.bmcl.2016.01.050

Bioorganic & Medicinal Chemistry Letters 26 (2016) 1475–1479
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of silyl proline containing HCV NS5A inhibitors
with pan-genotype activity: SAR development
a,
⇑
a
a
a
a
a
Anilkumar G. Nair , Qingbei Zeng , Oleg Selyutin , Stuart B. Rosenblum , Yueheng Jiang , De-Yi Yang ,
a
a
a
a
a
Kerry Keertikar , Guowei Zhou , Michael P. Dwyer , Seong Heon Kim , Bandarpalle Shankar ,
Wensheng Yu , Ling Tong , Lei Chen , Robert Mazzola , John Caldwell , Haiqun Tang ,
a
a
a
a
a
a
b
b
b
b
Melissa L. Allard , Ronald N. Buckle , Polivina Jolicia F. Gauuan , Christian L. Holst ,
b
b
b
a
a
Gregory S. Martin , Kannan P. Naicker , Samuel Vellekoop , Sony Agrawal , Rong Liu ,
a
a
a
a
a
a
Rong Kong , Paul Ingravallo , Ellen Xia , Ying Zhai , Amin Nomeir , Joseph A. Kozlowski
a
Merck Research Laboratories, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
Albany Molecular Research Inc., 26 Corporate Circle, Albany, NY 12203, USA
b
a r t i c l e i n f o
a b s t r a c t
Article history:
HCV NS5A inhibitors have shown impressive in vitro potency profiles in HCV replicon assays thus making
them attractive components for inclusion in an all oral fixed dose combination treatment regimen. Herein
we describe the research efforts that led to the discovery of silyl proline containing HCV NS5A inhibitors
such as 7e and 8a with pan-genotype activity profile and acceptable pharmacokinetic properties.
Ó 2016 Elsevier Ltd. All rights reserved.
Received 16 December 2015
Revised 14 January 2016
Accepted 18 January 2016
Available online 19 January 2016
Keywords:
HCV
NS5a
Pan-genotype
Hepatitis C virus (HCV) is currently the leading cause of chronic
liver disease and liver transplantation, and has become the most
common blood-borne infection in developed countries with
the host intracellular immune response, thereby making it a target
4
–6
of high interest.
As a therapeutic target, it has been extensively
7,8
investigated and clinically validated with daclastavir. Some of
the recently approved promising HCV treatments such as Harvoni
and Viekira pak contain HCV NS5A inhibitors, further indicating
1
approximately 130–150 million people infected globally. The pre-
vious standard of care for HCV involved pegylated interferon-
a
9
,10
(
IFN) and ribavirin (RBV). The year-long treatment was accompa-
the utility of this mechanism.
nied by significant side effects and limited efficacy. Individuals
who managed to complete the year-long treatment, and were
infected with the most common virus, genotype (gt) 1b, had
response rates of less than 50%.2 Improved sustained virologic
response (SVR) outcomes became possible by combining IFN and
RBV with protease inhibitors such as boceprevir (Victrelis) and
Our initial efforts in this area identified MK-4882 (Fig. 1) as a
potent NS5A inhibitor for clinical development which has been
described previously. While MK-4882 demonstrated excellent
1
1
Val-Cap
O
Val-Cap
Central Core
3
O
telaprevir (Incivek). However, many patients still did not obtain
SVR leaving room for improvement of outcomes.
N
N
H
O
O
O
N
N
N
H
N
H
Given the medical need for improved outcomes for genotype 1
patients, as well as those infected with other genotypes, efforts
toward improved HCV treatment continued. Despite gaps in the
understanding of the role of NS5A in the virus replication cycle,
it is known that it plays an integral role in a number of pathogen
events such as viral replication and release of virus particles.
NS5A also interacts with cellular proteins and lipid factors required
for HCV replication as well as being important for interference with
N
O
N
H
O
Proline
MK-4882
Proline
Figure 1. MK-4882 in vitro potency profile.
Genotype, EC50 (nM)
Compd
MK-4882
1a
0.01
1b
0.001
2a/2b
0.4/286
3a
4
1aY93H
24
⇑
Corresponding author.
http://dx.doi.org/10.1016/j.bmcl.2016.01.050
960-894X/Ó 2016 Elsevier Ltd. All rights reserved.
0

1
476
A. G. Nair et al. / Bioorg. Med. Chem. Lett. 26 (2016) 1475–1479
gt1 replicon activity, it possessed relatively weak activity against
other genotypes, and also had the potential to be an inducer of
CYP3A4 drug interactions (PXR EC50 = 3.3 lM). In order to improve
N
N
O
N
N
O
N
N
O
a
Cl
Si
Cl
O
O
upon MK-4882, our efforts focused upon the identification of alter-
nate structures which could effectively inhibit NS5A, be pan-geno-
typic, and alleviate any concern around potential metabolic
liabilities. Herein, we describe our efforts to investigate the impact
of introducing silyl proline variations into different HCV NS5A inhi-
bitors with the goal of identifying novel inhibitors that possess
both pan genotype activity and improved pharmacokinetic (PK)
properties over the corresponding carbon analogs.
O
Si
Si
1A
1B
Cl
C
Cl
1
1C'
b
O
Boc
O
Boc
N
N
HO
c
O
Incorporation of a silicon atom into drug molecules provides a
source of chemical diversity in drug design that has not been rou-
tinely explored, but has the potential to affect both biological and
pharmaceutical properties.12 Silicon has a larger covalent radius
than carbon, and also an increased C–Si bond length compared to
the corresponding C–C bond. The increased atom size and bond
length corresponds to a change in the size and shape of the silicon
analog relative to the corresponding carbon analog. These subtle
changes in size and shape can impact the way in which silicon ana-
logs interact with proteins. One essential part of HCV NS5A inhibi-
tors, such as MK-4882, is the S-Pro-S-Val dipeptide subunit. We
were intrigued by the idea that introduction of silicon into the pro-
line moiety might impact the replicon potency as well as the phar-
macokinetic properties of this class of inhibitors.
Si
Si
1E
1D
Scheme 2. Synthesis of silaproline analog 1E, Reagents and conditions: (a) n-BuLi,
78 °C; (b) 1. aq HCl, MeOH, rt; 2. DIPEA, NaI, CH Cl /Et O, rt; 3. Boc O, rt; (c) 1 M
aq KOH, MeOH, rt.
ꢀ
2
2
2
2
derivative 1E bearing the desired stereochemistry. It should be
noted that starting from -valine, an amino acid of the -series
[(S)-configuration] is obtained, but in the case of -silaproline 1E,
it corresponds to the (R)-configuration, due to changed atom
priorities.
The replicon genotype and 1aY93H mutant profile is tabulated
in Table 1. Compound 1, the dimethyl silyl proline analog of MK-
4882, demonstrated slightly weaker replicon activity on the geno-
types and mutants tested compared to the parent compound. Con-
sidering the physical and chemical diversity the dimethyl silyl
proline offers compared to the carbon analog, we felt this result
warranted further exploration on the impact that silyl proline
introduction might have with different analogs.
D
L
L
Silyl proline analogs described in this communication were con-
structed by methods similar to the carbon analogs published pre-
1
1
viously. Scheme 1 represents a general retrosynthetic approach
toward the silyl proline analogs starting from a Boc protected
silaproline fragment. This approach and other variations to accom-
modate different central cores have been described previously in
1
3
the patent literature.
The silyl proline monomer synthesis was achieved as according
to a known method starting from the Schollkopf chiral auxiliary
To further explore the effect of silyl prolines in conjunction with
different central cores, we synthesized analogs with the classical
biphenyl central core (Table 2). To our surprise, the one-sided silyl
proline analog 2a showed replicon activity comparable to the car-
bon analog 2 (daclatasvir). In addition, the 1aY93H mutant activity
of 2a was comparable to the carbon analog in our assay. It was also
noted that the symmetric analog 2b bearing silyl proline motifs on
both sides showed significantly weaker replicon potency across
several genotypes and the 1aY93H mutant.
1
4
shown in Scheme 2. Deprotonation of the bis-ether 1A, prepared
from the -Val-Gly diketopiperazine, with n-BuLi afforded a cyclic
D
anion intermediate which reacted with bis-(chloromethyl)
0
dimethylsilane 1B to afford compounds 1C and 1C . The diastere-
omers were separated by silica gel column chromatography
whereupon the desired compound, 1C, was subjected to HCl pro-
moted hydrolysis, DIPEA and NaI mediated ring closure, followed
by Boc-protection to afford Boc-silaproline analog 1D. Hydrolysis
of ester derivative 1D with 1 M KOH yielded the silyl proline
To further explore the impact symmetry may have on silyl
proline analogs, we investigated central core systems such as the
Val-Cap
Val-Cap
Central Core
O
O
O
N
O
N
H
N
O
Ar
Ar
N
N
H
N
H
N
O
N
H
Si
Proline
O
O
O
O
N
N
N
H
(HO) B/X
Ar
N
2
N
H
O
Ar
X/B(OH)
O
N
2
N
N
H
N
H
Si
O
Boc
O
O
N
HO
Br
N
O
N
H
N
H
Si
Si
Scheme 1. General synthetic approach to silyl proline analogs.

A. G. Nair et al. / Bioorg. Med. Chem. Lett. 26 (2016) 1475–1479
1477
Table 1
1aY93H (>1
lM) activities. Replacement of central phenyl ring of
MK-4882 in vitro potency profile in comparison to silyl analog
3
a with a pyridine ring (analog 4d) was tolerated for gt 1a activity,
but showed a weaker 1aY93H mutant activity compared to phenyl
analog 3a. The fluoro substitution on the central phenyl ring (4e)
also showed a similar activity profile to pyridine analog 4d. Bulkier
substitutions such as methoxy, cyclopropyl, thiophene (4f–h) at
this position were detrimental to the replicon activity (Table 4).
Since the naphthyl imidazole structure as exemplified in 3b
provided the best 1aY93H mutant activity profile (70 nM), we
focused our efforts in this structural class to further explore the
limits of the silyl proline analogs. Expecting that the S-Val-Moc
cap was an essential component for good overall activity, and
replacing it might be difficult, we explored different amino
acid-Moc caps as replacements on either side of the lead molecule
O
O
N
N
O
O
O
N
H
N
N
H
N
H
O
N
H
N
O
Si
1
Compd
Genotype, EC50 (nM)
1a
1b
2a
3a
1aY93H
MK-4882
0.01
0.04
0.001
0.003
0.4
2.4
4
15
24
1
153
3
b. As was predicted, both right hand side amino acid variations
(Table 5) and left hand side amino acid variations (Table 6) in con-
junction with silyl proline in 3b led to no improvement in the over-
all activity profile of the resulting inhibitors compared to valine.
Based upon these results, we decided to focus on the S-Val-Moc
cap for further exploration in this series.
Table 2
Silyl analogs in bi-phenyl core, in vitro potency, EC50 (nM) profile
O
O
O
O
N
N
Table 4
N
H
O
N
Silyl analogs central core variations, in vitro potency, EC50 (nM) profile
N
O
N
H
N
H
N
H
X1
X2
O
O
O
O
N
Compd
X
1
/X
2
1a/1b
2a
3a
1aY93H
O
H
N
H
N
Central core
O
N
2
CH /CH
2
2
0.05/0.01
0.04/0.003
0.2/0.01
0.04
0.04
0.25
0.7
0.5
4.1
198
92
>1000
2
2
a
b
Me
Me
2
Si/CH
Si/Me
2
Si
2
2
Si
Compd
Central core
1a/1b
1aY93H
9917
N
Table 3
N
Silyl analogs in tricyclic central core, in vitro potency, EC50 (nM) profile
4a
N
H
>1/0.004
>1/0.006
N
H
O
O
O
O
O
S
N
N
N
N
H
O
N
H
N
4b
N
2333
921
N
H
N
H
N
N
H
HN
N
X1
X2
Compd
X
1
/X
2
1a/1b
2a
3a
1aY93H
4
4
c
S
N
H
0.14/0.004
3
CH
CH
Me
CH
2
/CH
2
0.01/0.004
0.014/0.003
0.003/0.004
0.06/0.005
0.02
0.02
0.004
0.01
0.07
0.05
0.26
0.2
119
95
35
N
NH
3a
3b
3c
2
/Me
Si/CH
/(CH
2
Si
2
2
2
2
)
4
Si
350
N
N
N
N
d
0.04/0.014
0.05/0.004
185
329
N
H
N
H
F
naphthyl imidazole ring system shown in Table 3. Syntheses of
1
5
N
N
these molecules was achieved as described in published patents.
4e
N
H
In this series, both silyl proline analog 3a and 3b showed compara-
ble replicon activity across genotypes 1a and 2a to the carbon ana-
log 3. Similar to the earlier series, comparable activity was
observed with 1aY93H mutant activity. The analog with silyl pro-
line on the left hand side 3b showed a ꢁ10ꢂ improvement in geno-
type 2a activity. In an effort to further explore the scope of
tolerations on silicon, the dimethyl was cyclized into a spirocyclic
silyl proline analog 3c, which showed genotype activities similar to
the dimethyl analog 3a, but offered no advantage in terms of
N
H
O
N
N
4
4
f
0.27/0.004
0.43/0.007
497
N
H
N
H
N
N
g
1598
1
aY93H mutant activity.
N
H
N
H
Even though these molecules provided excellent genotype
potency profiles, we investigated other constrained central core
variations as illustrated in Table 4 to further explore potential dif-
ferences in activity. Unfortunately, most of these inhibitors demon-
strated inferior replicon activity compared to the analogs in
Table 3. Examples 4a–c all showed weak gt 1a (>1 nM) and
S
4h
N
0.5/0.008
1148
N
H
N
H

1
478
A. G. Nair et al. / Bioorg. Med. Chem. Lett. 26 (2016) 1475–1479
Table 5
Silyl analogs in right side valine variations, in vitro potency, EC50 (nM) profile
Table 7
Silyl analogs in right side proline variations, in vitro potency, EC50 (nM) profile
O
O
N
N
O
O
N
R
N
O
N
N
H
N
N
N
H
N
H
O
H
N
N
H
N
H
R
Si
Si
Compd
R
1a
1aY93H
Compd
R
1a/1b
1aY93H
131
O
O
5
a
O
O
0.2
590
N
N
7a
O
0.09/0.01
N
H
O
H
O
O
O
5
5
b
c
0.08
0.07
260
225
N
H
O
O
O
N
H
O
N
7
b
0.04/0.002
0.1/0.009
204
728
O
O
F
N
H
F
O
O
O
O
N
H
5
5
d
e
O
O
0.15
0.14
640
579
O
N
7
7
c
N
H
O
OH
O
F
F
N
H
O
O
O
N
H
O
N
d
0.2/0.01
380
Table 6
Silyl analogs in left side valine variations, in vitro potency, EC50 (nM) profile
S
O2
O
O
O
O
N
N
R
N
N
H
N
H
O
N
O
7e
0.015/0.003
0.07/0.006
0.06/0.005
46
N
H
N
H
N
Si
F
F
Compd
R
1a
1aY93H
471
O
O
N
H
O
N
7
f
27
6
a
O
O
0.09
N
H
F
O
O
O
O
O
O
6
6
b
c
0.05
0.03
87
N
H
O
N
N
H
7g
44
O
O
OH
O
275
N
H
O
O
O
N
H
7h
N
O
0.015/0.003
0.005/0.002
90
30
6
6
d
e
O
O
O
0.08
0.6
279
N
H
O
O
O
OH
O
7i
N
H
O
>1000
N
N
H
O
An extensive SAR exploration to change or substitute the right
hand side proline was conducted in order to further understand
the broadness of the antiviral potential of this series. A selected
SAR representation is provided in Table 7. Most of the bicyclic vari-
ations (7a–b) showed good gt 1a wild type activity but with much
reduced 1aY93H mutant activity. The six membered variations, as

A. G. Nair et al. / Bioorg. Med. Chem. Lett. 26 (2016) 1475–1479
1479
Table 8
analogs 8a and 8b showed comparable activity profiles to 7e and
Silyl analogs in switched proline variations, in vitro potency, EC50 (nM) profile
7
f, respectively.
The pharmacokinetic properties of several lead compounds
O
O
were studied in Sprague-Dawley rats as shown in Table 9. The pos-
itive effect of silyl proline analogs on the pharmacokinetic profile
was shown to be an advantage for this series. The initial silyl pro-
line lead compound 3b showed comparable rat AUC (total drug
concentration), but with a significantly improved half-life com-
pared to the carbon analog 3 (Table 9). Prolonged half-life and
acceptable plasma exposure also were exhibited by lead silyl pro-
line analogs 7e and 8a.
In conclusion, we have demonstrated that the dimethyl silyl
proline moiety is well tolerated in several classes of NS5A inhibi-
tors compared to the carbon analogs and has an impact on com-
pound pharmacokinetic properties. A comparison of compounds
3 and 3b shows the replicon potency profile for the two analogs
was similar but there was a substantial improvement in the rat
half-life for the silicon analog, 3b. Further characterization and
ADME profile of these silicon-containing analogs will be reported
in a separate publication.
N
N
N
O
H
N
N
H
N
H
R
Si
Compd
R
1a/1b
1aY93H
65
O
O
N
H
O
N
8
a
0.009/0.002
F
F
O
O
N
H
O
N
8
b
0.02/0.003
10
F
References and notes
1.
(a) Lavanchy, D. Clin. Microbiol. Infect. 2011, 17, 107; (b) Liang, J. T.; Hoofnagle, J.
H.; Hepatitis, C. Biomedical Research Reports; Academic Press: London, UK,
Table 9
Lead silyl analogs activity/rat PK comparisons to carbon analog
2
000; pp 185–203.
2
3
4
5
.
.
.
.
Bobeck, D. R.; Schinazi, R. F.; Coats, S. J. Antivir. Ther. 2010, 15, 935.
Sheridan, C. Nat. Biotechnol. 2011, 29, 553.
Schinazi, R. F.; Bassit, L.; Gavegnano, C. J. Viral Hepat. 2010, 17, 77.
Katze, M. G.; Kwieciszewski, B.; Goodlett, D. R.; Blakely, C. M.; Neddermann, P.;
Tan, S. L.; Aebersold, R. Virology 2000, 278, 501.
Kim, J.; Lee, D.; Choe, J. Biochem. Biophys. Res. Commun. 1999, 257, 777.
(a) 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.; 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; (b) Nettles, R. E.; Gao, M.; Bifano, M.;
Chung, E.; Persson, A.; Marbury, T. C.; Goldwater, R.; DeMicco, M. P.; Rodriguez-
Torres, M.; Vutikullird, A.; Fuentes, E.; Lawitz, E.; Lopez-Talavera, J. C.; Grasela,
D. M. Hepatology 2011, 54, 1956.
Compd 1a EC50 2a EC50 3a EC50 1aY93H
Rat PO AUC
M h/mL)
Rat IV
(nM)
(nM)
(nM)
EC50 (nM)
(
l
T
½
(h)
3
3
7
8
0.01
0.01
0.015
0.009
0.02
0.1
0.2
0.25
0.2
119
70
46
7.3 (10 mpk) 0.8
3.2 (5 mpk)
5 (5 mpk)
6
7
.
.
b
e
a
0.005
0.003
0.002
6.6
7.6
6.7
65
7.6 (5 mpk)
exemplified by 7c–d, showed suboptimal activity profile for both
wild type and mutant genotypes. It was interesting to observe that
the difluoro and monofluoro substituted proline analogs 7e and 7f,
respectively, showed excellent genotype 1a and 1aY93H mutant
activity (46 nM and 27 nM, respectively). A polar hydroxyl substi-
tution was also well tolerated for the replicon activity as in the case
of 7g (Table 7). Among the lipophilic alkyl substitutions, the cyclo-
propane-fused analog 7i (1aY93H IC50 = 30 nM) showed a prefer-
ence over gem-dimethyl substituted analog 7h (1aY93H
IC50 = 90 nM).
At this point in the SAR exploration, a number of promising lead
compounds were identified in the silyl proline structural class. To
further expand the SAR efforts, we looked at key compounds in
which the positions of silyl proline and substituted proline are
switched (Table 8). Interestingly, both difluoro and monofluoro
8
.
.
(a) Lok, A. Clin. Liver Dis. 2013, 17, 111; (b) Pawlotsky, J.-M. J. Hepatol. 2013, 59,
375; (c) Belda, O.; Targett-Adams, P. Virus Res. 2012, 170, 1; (d) Cordek, D. G.;
Bechtel, J. T.; Maynard, A. T.; Kazmierski, W. M.; Cameron, C. E. Drugs Future
2
011, 36, 691.
9
Keating, G. M. Drugs 2015, 75, 675.
10. Deeks, E. D. Drugs 2015, 75, 1027.
1
1. Coburn, C. A.; Meinke, P. T.; Chang, W.; Fandozzi, C. A.; Graham, D. J.; Hu, B.;
Huang, Q.; Kargman, S.; Kozlowski, J.; Liu, R.; McCauley, J. A.; Nomeir, A. A.;
Soll, R. M.; Vacca, J. P.; Wang, D.; Wu, H.; Zhong, B.; Olsen, D. B.; Ludmerer, S.
W. ChemMedChem 2013, 8, 1930.
2. (a) Bains, W.; Tacke, R. Curr. Opin. Drug Discov. Devel. 2003, 6, 526; (b) Franz, A.
K.; Wilson, S. O. J. Med. Chem. 2013, 56, 388; (c) Meanwell, N. A. J. Med. Chem.
1
2
011, 54, 2529.
13. Refer to patent WO2013039876 and WO2011112429 for detailed procedures
and schemes.
4. Vivet, B.; Cavelier, F.; Martinez, J. Eur. J. Org. Chem. 2000, 807.
1
1
5. Refer to patent WO2011112429 for detailed procedures and schemes.