Substituted benzothiadizine inhibitors of Hepatitis C virus polymerase

Article abstract of DOI:10.1016/j.bmcl.2009.05.091

The synthesis and optimisation of HCV NS5B polymerase inhibitors with improved potency versus the existing compound 1 is described. Substitution in the benzothiadiazine portion of the molecule, furnishing improvement in potency in the high protein Replicon assay, is highlighted, culminating in the discovery of 12h, a highly potent oxyacetamide derivative.

Full text of DOI:10.1016/j.bmcl.2009.05.091

Bioorganic & Medicinal Chemistry Letters 19 (2009) 4350–4353
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Substituted benzothiadizine inhibitors of Hepatitis C virus polymerase
a,
c
a
d
Antony N. Shaw a, , Rosanna Tedesco , Ramesh Bambal , Deping Chai , Nestor O. Concha ,
Michael G. Darcy a, Dashyant Dhanak a, Kevin J. Duffy a, Duke M. Fitch a, Adam Gates b,
Victor K. Johnston b, Richard M. Keenan a, Juili Lin-Goerke b, Nannan Liu a, Robert T. Sarisky b,
Kenneth J. Wiggall a, Michael N. Zimmerman a
*
*
a Department of Medicinal Chemistry, GlaxoSmithKline, Collegeville, PA 19426, USA
b Department of Virology, GlaxoSmithKline, Collegeville, PA 19426, USA
c Department of Drug Metabolism and Pharmacokinetics, GlaxoSmithKline, Collegeville, PA 19426, USA
d Department of Computational, Analytical and Structural Sciences, GlaxoSmithKline, Collegeville, PA 19426, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The synthesis and optimisation of HCV NS5B polymerase inhibitors with improved potency versus the
existing compound 1 is described. Substitution in the benzothiadiazine portion of the molecule, furnish-
ing improvement in potency in the high protein Replicon assay, is highlighted, culminating in the discov-
ery of 12h, a highly potent oxyacetamide derivative.
Received 19 March 2009
Revised 18 May 2009
Accepted 20 May 2009
Available online 28 May 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
HCV
Hepatitis C virus
NS5B polymerase inhibitors
Benzothiadiazines
Replicon assay
High protein replicon assay
Hepatitis C virus (HCV) is responsible for a variety of clinical
conditions ranging from acute viral hepatitis to chronic liver
disease and cirrhosis. It is the major cause of liver cancer and about
two thirds of all liver transplants are a result of HCV infection.
There are an estimated 170 million people worldwide chronically
infected and 3–4 million new infections annually.2 In the United
States, approximately 19,000 new infections occurred in 2006
and an estimated 3–4 million carry the disease, many of whom
were infected through blood transfusions with contaminated blood
prior to the start of screening for the virus in 1992.3 There is
currently no vaccine against the virus.
O
O
S
O O
S
OH
N
OH
N
R
R2
N
N
R
H
H
N
O
N
O
Alk
Alk
2
1
cellular potency, particularly under high plasma-protein condi-
tions more similar to those encountered physiologically. The
preceding Letter describes an alternative approach to modulation
of physical properties using related heterocyclic isosteres.7
The compounds were synthesized according to the methods
shown in the Schemes 1 and 2 (experimental details supplied in
the Supplementary data).
Novel 2-Aminobenzenesulfonamides 4 were synthesized by
reaction of anilines 3 with chlorosulfonyl isocyanate followed by
cyclization with aluminium chloride and hydrolysis,8 then acylated
with ethyl malonyl chloride or diethyl malonate, and cyclised with
phosphorus oxychloride or aqueous sodium carbonate to give the
benzothiadiazine-3-acetic acid ester 5. Ester 5 was condensed with
In a previous publication,1 we describe the development of a
series of 3-(benzothiadiazin-3-yl)quinolinone inhibitors 1 of the
non-structural protein NS5B, an RNA dependent RNA polymerase
encoded by the viral genome and known to be essential for viral
replication.4 Related benzothiadiazine inhibitors of NS5B have
been described by our group5 and others.6
This Letter describes the structure–activity relationships of substi-
tution on the benzothiadizinyl ring as in compound 2 with the aim
of improving the developability of the compounds and their
N-alkyl isatoic anhydrides
6 and cyclised to the required
compound 2 (Scheme 1).1 In an alternative route, the isatoic anhy-
drides 6 were converted to quinoline-3-nitriles 7 by condensation
* Corresponding authors. Tel.: +1 610 917 4715 (R.T.).
E-mail address: Rosanna_2_Tedesco@gsk.com (R. Tedesco).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.05.091

A. N. Shaw et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4350–4353
4351
O
O
H2NO2S
H2N
S
c, d
a, b
N
R
R
R
EtO2C
N
H
H2N
3
4
5
e
O
O
S
OH
OH
N
O
R
R2
CN
O
R2
R2
N
H
O
f
g
N
N
O
N
O
R1
R1
R1
7
2
6
Scheme 1. Reagents and conditions: (a) ClSO2NCO, AlCl3, EtNO2; (b) 50% H2SO4,
reflux; (c) EtO2CCH2CO2Et, DMF, or Cl(O)CCH2CO2Et, Et3N, DMAP, THF; (d) aq
D
Na2CO3 or POCl3, reflux; (e) 6, NaH, then AcOH, THF, reflux or DBU, then AcOH,
DMF; (f) EtO2CCH2CN, NaH, DMF, heat; (g) 4, Me3Al, 1,4-dioxane; then NaOH, reflux.
Figure 1. X-ray of 10a bound to NS5B (genotype 1b, BK strain,
D
21 construct, PDB
deposition code 2FVC).
O
O
S
O O
OH
S
OR7
functionality at these positions was poorly tolerated or syntheti-
cally less accessible (data not shown).
OH
N
OH
N
R2
R2
N
H
N
H
a
The first compounds 10 with an N-isoamyl quinolinone ring
(Table 1) showed that substitution is well tolerated at the 7
position with a wide variety of substituents. Generally, smaller,
hydrophilic groups are preferred. Fluorine maintains the activity
in the NS5B enzymatic assay,14 while activity drops as the halogen
gets larger. Both electron donating (e.g., OMe, compound 10k) and
electron-withdrawing substituents (e.g., CN, compound 10f) are
tolerated. Hydrophilic OH (compound 10l), NH2 (compound 10i)
substituents gave compounds with activity similar to the parent.
Introduction of a carbonyl function is well tolerated. Ester 10n
and acid 10o have similarly potencies, while the 7-carboxamide
derivative 10p is about fivefold more potent than the unsubstitut-
ed parent 10a. The more bulky tertiary amide group (compound
10q) resulted in significantly lower enzymatic activity.
N
O
N
O
R1
R1
8
9
Scheme 2. Reagents and Conditions: (a) NaH or K2CO3, R7Br, R7Cl or R7OTs, DMF or
aq DME, heat.
with ethyl cyanoacetate and cyclization. The nitriles 7 were
coupled with 2-aminobenzenesulfonamides 4 to give, after cycliza-
tion with sodium hydroxide, the required compound 2.
Benzothiadiazines (10f, m–q) were made after the heterocycle
had been constructed, either by palladium-catalyzed substitution
of a halide,9,10 or by nitrile hydrolysis/methanolysis. 7-Amino com-
pounds 10i and 10j were made by nitration of the parent heterocy-
cle followed by hydrogenation, then reductive alkylkation.
Hydroxy compound 10l was made by demethylation of the meth-
oxy compound 10k, while 12b was made as in Scheme 1 with the
appropriate benzothiadiazine-3-acetic acid ester 5. Carbon-linked
compounds 12i and 12j were made by palladium-catalysed cou-
pling of the corresponding 7-iodo compound with acrylamide,11
followed by hydrogenation. Hydroxymethyl compound 12d was
made from the corresponding ester by reduction with lithium alu-
minum hydride and converted to the carbamate derivative 12k by
reaction with trichloroacetyl isocyanate followed by hydrolysis.12
More elaborate alkoxy compounds 9 were made by alkylation of
the hydroxy derivatives 8 using the appropriate alkyl bromide,
chloride or tosylate as in Scheme 2, followed, if necessary, by func-
tional group manipulation. Compounds 12n and 12o were made by
alkylation with the appropriate homochiral 2-tosyloxypropiona-
mide without loss of chiral integrity.13
An X-ray structure of compound 10a bound to the NS5B poly-
merase in the palm/thumb domain region showed a tight fit of
the compound in the binding site (Fig. 1).1 It was apparent from
the structure that, while most substitutable positions around the
quinolinone ring were close to the protein surface, there was space
available around the benzothiadiazine ring for further substitution.
An examination of the protein structure showed a number of
hydrophilic residues in close proximity to the benzothiadiazine
7-position and prompted us to explore substitution in the thiadi-
azine with an emphasis on polar substituents.
In general, activity in the Replicon assay15 tracks fairly well
with enzymatic activity. The hydroxy compound 10l, however, is
Table 1
Polymerase inhibition,14 type 1b Replicon assay data15 under high and low protein
conditions for compounds 10
O
O
S
R7
OH
N
N
H
N
O
Ami
10
Compd R7
NS5Ba IC50
(nM)
Replicon IC50
(nM)
PA-Repliconb IC50
(nM)
10a
10b
10c
10d
10e
10f
H
F
Cl
Br
32
26
76
173
>10,000
57
417
334
2321
5141
NT
NTc
NT
>50,000
NT
NT
NT
I
CN
486
10g
10h
10i
10j
10k
10l
10m
10n
10o
10p
10q
Me
50
238
13
142
55
30
127
65
45
1784
>20,000
218
1247
1394
>20,000
5195
2625
>20,000
383
NT
NT
45,000
NT
NT
NT
NT
NT
NT
NO2
NH2
NMe2
OMe
OH
C(O)Me
CO2Me
CO2H
CONH2
6
NT
NT
The goal of this strategy was not only to improve potency by
making additional H-bonding interactions with the protein, but
also to reduce the highly lipophilic nature of our earlier
compounds1 in an attempt to lower protein binding and improve
activity in cell-based assays and in vivo settings. Substitution at
other positions 5, 6 and 8 was also explored, but bulky polar
CONMe2 102
>20,000
a
D
21-truncated NS5B used in the assay.
b
PA-Replicon = protein-attenuated Replicon assay, run in the presence of 45 mg/
mL HSA and 1 mg/mL AAG.
c
NT = not tested.

4352
A. N. Shaw et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4350–4353
Table 2
(e.g., compound 10c, Table 1). An X-ray structure of 11d bound
to the type 1b NS5B was solved and shows the oxyacetamide group
filling out the binding pocket and making a favorable hydrogen
bond to the side-chain amide of Asn291 (Fig. 2).
The optimal 6-fluoro and 1-2-(cyclopropyl)ethyl substituents
on the quinolinone ring,1 were next combined with a panel of
the best substituents from the benzothiadiazine SAR studies
described above to maximize the activity of the substituted thiadi-
azines both enzymatically and in the cellular assay. The activity of
these compounds 12 is shown in Table 3.
Polymerase inhibition, type 1b Replicon assay data under high and low protein
conditions for compounds 11
O
O
S
OR7
OH
N
R2
N
H
N
O
Ami
11
Compd
R2
R7
NS5B
IC50 (nM)
Replicon
IC50 (nM)
PA-Replicon
IC50 (nM)
The unsubstituted thiadiazine compound 12a, our first develop-
ment candidate,1 is potent against the isolated enzyme and in the
standard Replicon assay, but much weaker in the protein-attenu-
ated Replicon assay in the presence of physiologically relevant
concentrations of HSA and AAG. The 7-hydroxy and 7-methoxy
compounds 12b and 12c are similar to the parent, but the longer
side-chain analogs again showed significant improvements in
potency. The oxyacetic acid 12e is potent but weak in cells, proba-
bly due to high polarity and therefore poor membrane permeabil-
ity. The corresponding tetrazole 12g offers no improvement in
cells, nor does the 1,1-dimethyloxyacetic acid 12p. The oxyaceto-
nitrile 12f shows high potency and low micromolar activity in
the protein-attenuated Replicon assay, but the oxyacetamide 12h
is more potent still, mirroring the SAR trends seen in the previously
described isoamyl series. Interestingly, an all carbon-linked side-
chain, either saturated or unsaturated, (compounds 12j and 12i)
abolishes the gain in the high-protein Replicon assay, and the
isomeric carbamate 12k of amide 12h is weaker in both primary
assays. Confident that the oxyacetamide imparted the best activity
of the compounds made at that time, we looked at simple substi-
tution on this substituent for further improvement. The N-methyl
secondary amide 12l is slightly less potent in both Replicon assays
and the N-dimethyl tertiary amide 12m significantly weaker, pre-
sumably due to the steric perturbation of the interaction with
Asn291. Introduction of a methyl group into the oxyacetamide
side-chain gives one enantiomer ((R) 12n) with activity compara-
ble to 12h in enzymatic and high protein cellular assays. Interest-
ingly, the (S)-enantiomer 12o is weaker biochemically, as was
11a
11b
11c
11d
11e
11f
H
H
H
H
H
H
F
CH2CN
7
23
8
81
NT
NT
NT
2646
NT
NT
8000
3097
CH2CO2CH3
CH2CO2H
CH2CONH2
(CH2)2NH2
(CH2)2N(CH3)2
CH2CN
815
1604
15
568
NT
6
142
685
5
11g
11h
29
13
F
CH2CONH2
18
very weak in the cellular assay and the biochemically potent 7-car-
boxamide derivative 10p is equipotent with the parent 10a,
despite its higher biochemical activity.
The 7-hydroxyl group was a useful handle for introducing fur-
ther functionality through alkylation, and allowed us to reach out
further into the binding pocket. Table 2 shows such alkylated
derivatives, 11.
The oxyacetonitrile substituted compound 11a is very potent in
the enzymatic assay and also shows higher potency in the Replicon
assay than all previous compounds 10a–q. The corresponding oxy-
acetic acid 11c and its methyl ester 11b are also potent, but weaker
in the cellular assay. The oxyacetamide 11d stands out as it has
very high potency in both the enzymatic and the cellular assay,
being about 25-fold more potent in the latter assay than the parent
compound 10a. The 7-aminoethoxy (compound 11e) or 7-dimeth-
ylaminoethoxy (compound 11f) derivatives are weaker NS5B
inhibitors. Incorporation of a fluoro substituent at the 6 position
of the quinolinone ring has previously been shown to improve
cellular activity1 and exhibits a similar effect here, as exemplified
by the 7-oxyacetonitrile (compound 11g) and 7-oxyacetamide
(compound 11h) derivatives. Indeed, the 7-oxyacetamide analog
11h was, at the time, the most potent compound tested in the cel-
lular Replicon system. Importantly, compounds 11d, 11g, 11h have
low micromolar inhibitory activity in the type 1b Replicon assay
when run in the presence of 45 mg/mL HSA and 1 mg/mL alpha
acid glycoprotein (AAG), representing a significant improvement
over earlier compounds with smaller, lipophilic C-7 substituents
Table 3
Polymerase inhibition, type 1b Replicon assay data under high and low protein
conditions for compounds 12
O
O
S
R7
OH
N
N
F
N
H
O
12
Compd
R7
NS5B
IC50 (nM)
Replicon
IC50 (nM)
PA-Replicon
IC50 (nM)
12a1
12b
12c
12d
12e
12f
12g
12h
12i
12j
12k
12l
12m
12n
12o
12p
12q
H
OH
OMe
10
<5
24
17
<5
<5
6
<5
202
5
36
<5
12
<5
12
22
9
38
57
24,461
50,000
>50,000
NT
NT
2388
NT
1,484
50,000
>50,000
NT
4,737
NT
1017
50,000
NT
342
233
2,295
8
2000
2
39
208
48
CH2OH
OCH2CO2H
OCH2CN
OCH2(tetrazole)
OCH2CONH2
(E)-CH@CHCONH2
(CH2)2CONH2
CH2OCONH2
OCH2CONHCH3
OCH2CON(CH3)2
(R)-OCH(CH3)CONH2
(S)-OCH(CH3)CONH2
OC(CH3)2CO2H
OC(CH3)2CONH2
9
211
3
4
3,742
272
Figure 2. X-ray of 11d bound to NS5B (genotype 1b, BK strain,
D21 construct, PDB
5358
deposition code 3HHK).

A. N. Shaw et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4350–4353
4353
Table 4
Inhibitor 12h showed the best overall profile of the compounds
examined and was selected as an improved follow-up develop-
ment candidate for the treatment of HCV infection.
Comparison of selected properties of HCV NS5B inhibitors
Property
12a
12h
12n
Type 1b NS5B IC50 (nM)
10
<5
<5
Type 1b Replicon IC50 (nM)
38
2
3
Supplementary data
Protein-attenuated 1b Replicon IC50 (nM)
Equilibrium dialysis% in human plasma
Rat PKa: T1/2 (h), F(%)
24,461
99.9
3, 44
0.84
1484
99.9
1, 33
5.4
1017
99.8
0.7, 35
1.7
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2009.05.091.
CYP2C9 IC50 (lM)
a
Low dose (ꢀ4 mg/kg) oral and iv pharmacokinetics in rat. T1/2 = terminal half-
life, F = bioavailability.
References and notes
1. Tedesco, R.; Shaw, A. N.; Bambal, R.; Chai, D.; Concha, N. O.; Darcy, M. G.;
Dhanak, D.; Fitch, D. M.; Gates, A.; Gerhardt, W. G.; Halegoua, D. L.; Han, C.;
Hofmann, G. A.; Johnston, V. K.; Kaura, A. C.; Liu, N.; Keenan, R. M.; Lin-Goerke,
J.; Sarisky, R. T.; Wiggall, K. J.; Zimmerman, M. N.; Duffy, K. J. J. Med. Chem. 2006,
49, 971.
predicted by prior modeling studies showing a small pocket for a
group in the (R)-configuration but not for the one in the (S)-config-
uration (data not shown). Consistently, dimethylation of the
oxyacetamide side-chain also gives a compound (12q) with lower
biochemical potency than 12n, although the activity in the protein-
attenuated Replicon assay, surprisingly, is only about fivefold
worse.
Most of the compounds described in this Letter were tested in a
large panel of developability assays, including pharmacokinetics,
human serum albumin (HSA) column retention and cytochrome
P450 assays. The data for the most promising compounds 12h
and 12n are compared to those of the first development candidate
12a in Table 5‘. CYP2C9 inhibition was routinely measured, since,
historically, the benzothiadiazine inhibitors show the most potent
inhibition against this isoform. Inhibitory potencies against other
P450 enzymes (data not shown) are significantly lower than for
2C9 for these compounds.
While the human serum protein binding, measured by equilib-
rium dialysis, ultimately was not significantly reduced for 12h and
12n, their potencies in the key assay, the protein-attenuated Rep-
licon assay, were significantly improved over that of 12a. Protein
binding, as measured by HSA column retention, was not in fact sig-
nificantly reduced for any of the compounds described here (data
not shown).
2. World Health Organisation; Hepatitis C Factsheet, www.who.int/mediacentre/
factsheets/fs164/en/index.html, 2000.
3. Centers for Disease Control; Hepatitis
www.cdc.gov/hepatitis/Statistics.htm, 2008.
C Statistics and Surveillance,
4. Behrens, S. E.; Tomei, L.; de Francesco, R. EMBO J. 1996, 15, 12.
5. Fitch, D. M.; Evans, K. A.; Chai, D.; Duffy, K. J. Org. Lett. 2005, 7, 5521.
6. Rockway, T. W.; Zhang, R.; Liu, D.; Betebenner, D. A.; McDaniel, K. F.; Pratt, J. K.;
Beno, D.; Montgomery, D.; Jiang, W. W.; Masse, S.; Kati, W. M.; Middleton, T.;
Molla, A.; Maring, C. J.; Kempf, D. J. Bioorg. Med. Chem. Lett. 2006, 16, 3833.
7. Tedesco, R.; Chai, D.; Darcy, M. G.; Dhanak, D.; Fitch, D. M.; Gates, A.; Johnston,
V. K.; Keenan, R. M.; Lin-Goerke, J.; Sarisky, R. T.; Wiggall, K. J.; Zimmerman, M.
N.; Duffy, K. J.; Shaw, A. N.; Valko, K. L. Bioorg. Med. Chem. Lett. 2009, 19, 4354.
8. Girard, Y.; Atkinson, J. G.; Rokach, J. J. Chem. Soc., Perkin Trans. 1 1979, 1043.
9. Tschaen, D. M.; Desmond, R.; King, A. O.; Fortin, M. C.; Pipik, B.; King, S.;
Verhoeven, T. R. Synth. Commun. 1994, 24, 887.
10. Xu, L.; Chen, W.; Xiao, J. J. Mol. Catal. A: Chem. 2002, 187, 189.
11. Zhang, Z.; Pan, Y.; Hu, H.; Kao, T.-Y. Synthesis 1991, 539. Pd(OAc)2 used as
catalyst.
12. Guindon, Y.; Murtagh, L.; Caron, V.; Landry, S. R.; Jung, G.; Bencheqroun, M.;
Faucher, A.; Guérin, B. J. Org. Chem. 2001, 66, 5427.
13. Chiral HPLC, Chiralpak AD 20 Â 250 mm column, 6/1/1 Hexanes/MeOH/EtOH
containing 0.1% i-PrNH2.
14. Dhanak, D.; Duffy, K. J.; Johnston, V. K.; Lin-Goerke, J.; Darcy, M.; Shaw, A. N.;
Gu, B.; Silverman, C.; Gates, A. T.; Nonnemacher, M. R.; Earnshaw, D. L.; Casper,
D. J.; Kaura, A.; Baker, A.; Greenwood, C.; Gutshall, L. L.; Maley, D.; DelVecchio,
A.; Macarron, R.; Hofmann, G. A.; Alnoah, Z.; Cheng, H. Y.; Chan, G.; Khandekar,
S.; Keenan, R. M.; Sarisky, R. T. J. Biol. Chem. 2002, 277, 38322.
15. Bartenschlager, R. J. Hepatol. 2005, 43, 210.