HCV NS5B polymerase inhibitors 1: Synthesis and in vitro activity of 2-(1,1-dioxo-2H-[1,2,4]benzothiadiazin-3-yl)-1-hydroxynaphthalene derivatives

Article abstract of DOI:10.1016/j.bmcl.2009.05.063

2-(1,1-Dioxo-2H-[1,2,4]benzothiadiazin-3-yl)-1-hydroxynaphthalene derivatives as potential anti-HCV drugs targeting NS5B polymerase have been investigated. Their synthesis, HCV NS5B polymerase inhibition and replicon activity are discussed.

Full text of DOI:10.1016/j.bmcl.2009.05.063

Bioorganic & Medicinal Chemistry Letters 19 (2009) 4476–4479
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
HCV NS5B polymerase inhibitors 1: Synthesis and in vitro activity of
2-(1,1-dioxo-2H-[1,2,4]benzothiadiazin-3-yl)-1-hydroxynaphthalene derivatives
b
b
b
a
a
Guangyi Wang a, , Yanzhen He , Jun Sun , Debasis Das , Mougang Hu , Jianhua Huang ,
*
Donald Ruhrmund a, Lisa Hooi a, Shawn Misialek a, P. T. Ravi Rajagopalan a, Antitsa Stoycheva a,
Brad O. Buckman a, Karl Kossen a, Scott D. Seiwert a, Leonid Beigelman a,
*
a Discovery Research, InterMune, Inc., Brisbane, CA 94005, USA
b Chemistry Department, WuXi AppTec Co., Ltd, Shanghai 200131, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
2-(1,1-Dioxo-2H-[1,2,4]benzothiadiazin-3-yl)-1-hydroxynaphthalene derivatives as potential anti-HCV
drugs targeting NS5B polymerase have been investigated. Their synthesis, HCV NS5B polymerase inhibi-
tion and replicon activity are discussed.
Received 6 March 2009
Revised 5 May 2009
Accepted 6 May 2009
Available online 23 May 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Benzothiadiazine
HCV NS5B polymerase inhibitors
WHO estimates that about 3% of the world’s population has
been infected with HCV and that some 170 million are chronic car-
riers at risk of developing liver cirrhosis and/or liver cancer.1 Hep-
atitis C is recognized as a major cause of end-stage liver disease as
well as the leading cause of liver transplantations in the developed
world.2 Thus far, there is no universally effective therapy for all
HCV genotypes. The current treatment for patients infected with
itors such as nucleoside or nucleotide inhibitors that mimic natural
polymerase substrates and allosteric inhibitors that bind to less
conserved sites outside the active site and impair the enzyme’s cat-
alytic efficiency.
Recently, several classes of non-nucleoside allosteric NS5B
inhibitors5–7 have been reported, some of which achieved low
nanomolar inhibition in both enzyme and replicon assays. Among
the most potent inhibitors are 1,1-dioxo-2H-benzothiadiazine
compounds, of which compound 1 shown below was one of the
first benzothiadiazine compounds found active against HCV NS5B
polymerase.8 Compound 1 binds in a site that is located in the palm
domain of the polymerase in the proximity of the active site (see
Ref. 8 for a crystal structure of a benzothiadiazine compound
bound to NS5B). Recent work from another institution identified
the low nanomolar benzothiadiazine inhibitor 2,9,10 which exhib-
ited a good DMPK profile achieving high plasma and liver concen-
tration and showed anti-HCV activity in chimpanzee model.11
In search of potent allosteric inhibitors of HCV polymerase NS5B
at InterMune, we explored several series of benzothiadiazine com-
pounds. In this report we describe synthesis and in vitro anti-HCV
activities of compounds 3, 4 and their analogs.
Target compound 3a was synthesized as shown in Scheme 1.
Compound 5 was alkylated with isopentyl iodide, followed by alka-
line hydrolysis, to give the acid 6. Treatment of 6 with thionyl chlo-
ride and subsequent reaction of the resulting acyl chloride with
diethyl malonate yielded 7. Cyclization of 7 in methanesulfonic
acid afforded the naphthalene derivative 8 in good yield. Methyla-
tion of 8 with TMS–diazomethane, followed by hydrolysis, afforded
genotype 1 HCV is 48 weeks of pegylated interferon-
a (Peg-IFN-
a) and ribavirin (RBV), and the success rate for achieving a sus-
tained viral response for genotype 1 patients in the US, Europe
and Japan is ꢀ40%.3 The long duration of treatment is difficult for
patients to tolerate due to side effects associated with Peg-IFN-
a
and RBV that include flu-like symptoms, fatigue, depression, gas-
trointestinal symptoms, pulmonary effects and others.3 These
limitations have led to intense interest in the discovery and devel-
opment of novel compounds that target the viral and host proteins.
HCV NS5B polymerase is an RNA dependent RNA polymerase
that resides at the C-terminal domain of a polypeptide of several
structural and nonstructural proteins and contains the catalytic
machinery responsible for synthesis and replication of the viral
RNA.4 NS5B is essential for viral replication and has been clinically
validated.6 Along with HCV protease NS3/4A, NS5B is recognized as
the most viable protein target for HCV drug discovery.5–7 Two clas-
ses of NS5B inhibitors have been well developed: active site inhib-
* Corresponding authors.
E-mail addresses: gyiwang@yahoo.com, gwang@intermune.com (G. Wang),
lbeigelman@intermune.com (L. Beigelman).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.05.063

G. Wang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4476–4479
4477
amination and subsequent hydrogenolysis. Sandmeyer-type reac-
tion of the amine 15, followed by hydrolysis, afforded the acid 16.
The desired benzothiadiazine compounds 3b–g were prepared
by the general procedure as shown in Scheme 3. Condensation of
the acids 10, 14 and 16 with the 2-aminobenezenesulfonamides
17 and 189 in polyphosphoric acid trimethylsilyl ester (PPSE)12
gave the benzothiadiazines 19a–e, respectively, in 50–65% yields.
While the preparation of the benzothiadiazines 3b–f and 20 was
readily achieved by demethylation of 19a–e under mild conditions,
removal of both methyl groups of 19e for preparation of 3g re-
quired a large excess of BBr3 and 2–3 days at 38 °C under argon
flow. Compounds 3a and 3g having two hydroxy groups could be
oxidized in the air, particularly 3g which was much more unstable
than 3a. In preparation of 3g, a fast flash chromatographic purifica-
tion on silica gel with EtOAc/DCM gave the desired compound, but
with significant loss in yield while 3a was obtained in good yield
by flash chromatography. When a solution of 3g in methanol and
water (1:1) was in an open flask for 10 days, 3g almost completely
disappeared and a major product was formed. This product was
purified and assigned by MS and NMR as the racemic 21.13
Although oxidative mechanism remains unclear at this time, it is
certain that the oxidant source is air. Compounds 3f and 20 as a
mixture prepared from 19e were readily separated by flash chro-
matography on silica gel. The regiochemical assignment of 3f and
20 was determined by NOESY NMR: the methoxy protons in 20
correlated with a naphthalene proton (H-8) and had no correlation
with the alpha-methylene while the methoxy protons in 3f had a
correlation with alpha-methylene protons and no correlation with
a naphthalene proton (H-8). In contrast to 3g, compounds 3f and
20 having only one hydroxy group were stable and remained intact
at room temperature after several months.
O
O
O
O
H
S
N
S
OH N
S
OH N
O O
N
H
N
H
O
N
O
1
2
IC50 = 32 nM (1b)
EC50 = 417 nM(1b)
IC50 = 0.33 nM (1b)
IC50 = 1.25 nM (1a)
EC50 = 3 nM(1b); 11 nM(1a)
O
O
O
O
H
R3
S
N
N
S
OH N
S
OH N
O O
N
H
R1
O
3
4
R1 = H, F, OH, OMe
R3 = H, NHS(O2)Me
the acid 10. Attempts to use 8 as starting material for condensation
with 2-aminobenzenesulfonamide under various conditions
yielded a complicated mixture, probably due to the instability of
the hydroxyl groups at high temperature. Condensation of 9 with
2-aminobenzenesulfonamide did not proceed under various condi-
tions, presumably owing to the unreactive nature of the carboxylic
ester. It appeared that a more reactive derivative of 9 was required
for a successful condensation. After treatment of 10 with thionyl
chloride, the resulting acyl chloride was condensed with 2-amino-
benzenesulfonamide in the presence of DMAP and triethylamine.
Interestingly, the acyl chloride resulting from 10 did not react with
2-amino group of 2-amniobenzenesulfonamide, instead, the acyl
chloride was coupled with the sulfonamide amino group to give
11. Treatment of 11 in aqueous potassium hydroxide at up to
140 °C failed to produce the desired benzothiadiazine 12. Finally,
12 was obtained by heating 11 neat at 200 °C. Demethylation of
12 with boron tribromide afforded the desired benzothiadiazine
3a.
H
N
O
O
O
O
H
S
S
N
OH N
S
O
N
S
O O
O O
N
H
N
H
O
OH
OH
20
21
The naphthalene-2-carboxylic acids 14 and 16 as the intermedi-
ates for 3b–e were synthesized according to Scheme 2. Compound
8 was selectively methylated and subsequently triflated. The meth-
ylation site was determined by NOESY NMR experiment of the
monomethoxy product, where the methoxy protons had a correla-
tion with a naphthalene proton (H-8). The resulting triflate 13 was
subjected to reduction and hydrolysis to give the acid 14. Com-
pound 13 was also converted to the amino derivative 15 via the
To find out whether a constrained conformation of the naphtha-
lene–benzothiadiazine compounds would enhance NS5B binding,
compound 4 was designed as a rigid analog of 3. Synthesis of 4
and its analogs is shown in Scheme 4. A mixture of 3f and 20
was treated with benzoyl chloride to protect both hydroxy and
methylsulfonamide functions. After separation by chromatogra-
OH O
O
O
O
O
CH(COOEt)2
OEt
OH
OEt
f
a,b
d
e
c
N
OH
O
5
6
7
8
9
O
O
O
O
S
S
O
O
NH2
O
O
O
OH N
O
N
O
g,h
S
i
j
N
H
OH
N
N
H
H
O
O
OH
O
11
12
3a
10
Scheme 1. Reagents and conditions: (a) i-pent-I, NaH, 0 °C, 30 min, 69%; (b) NaOH, H2O, EtOCH2CH2OH, reflux, overnight, 93%; (c) (i) SOCl2, ClCH2CH2Cl, reflux, 2 h; (ii)
diethyl malonate, MgCl2, TEA, CH3CN, rt, overnight; (d) CH3SO3H, 30 °C, overnight, 67% (three steps); (e) TMS–diazomethane, DIPEA, THF, MeOH, 30 °C, overnight, 93%; (f)
NaOH, H2O, 1,4-dioxane, reflux, 2 days; (g) SOCl2, ClCH2CH2Cl, 60 °C, 5 h; (h) 2-aminobenzenesulfonamide, DMAP, TEA, DMF, 45 °C, 2 h, 37% (three steps); (i) neat, 200 °C,
1.5 h, 40%; (j) BBr3, ClCH2CH2Cl, 45 °C, 24 h, 73%.

4478
G. Wang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4476–4479
O
O
3f
+
20
O
O
c,d
a,b
OH
OEt
a,b,c
8
OTf
O
O
H
O
O
H
S
N
S
N
OH N
S
OBzN
S
13
14
O O
O O
N
H
N
H
e,f
O
OBz
OH
O
O
O
F
23
22
g,h
OEt
NH2
OH
d,e
d,e
O
O
O
O
H
15
H
16
S
C
N
S
N
N
N
S
OH N
S
OH
D
O
O O
O O
Scheme 2. Reagents and conditions: (a) TMS–diazomethane, MeOH, rt, 4 h, 35%; (b)
Tf2O, pyridine, rt, 3 h, 46%; (c) Pd(Ac)2, PPh3, TEA, HCOOH, DMF, 65 °C, 24 h, 93%; (d)
NaOH, EtOH, H2O, rt; (e) benzylamine, Pd(Ac)2, BINAP, Cs2CO3, toluene, 80 °C, 20 h,
78%; (f) Pd/C, EtOH, rt, 30 h, 87%; (g) BF3-etherate, t-BuONO, 0 °C, 2 h, 18%; (h)
NaOH, EtOH, H2O, quant.
N
A
B
O
O
25
+
O
4
O
N
H
+
S
N
N
O
S
O
H
N
O O
S
N
O
N
S
phy, the two benzoyl-protected products were subjected to
demethylation with BBr3 and subsequent selective removal of ben-
zoyl group from the sulfonamide both under mild conditions to
give 22 and 23, respectively. Cyclization of 22 and 23 and subse-
quent removal of benzoyl group afforded two pairs (4 and 24, 25
and 26) of compounds, respectively. The structural assignments
for each pair of compounds are determined by NOESY NMR. For
the pair of 4 and 24, the isomer in which the methylene protons
correlate with H5 of the D ring in NOESY NMR was assigned as 4
and the other as 24. Similarly, for the pair of 25 and 26, the isomer
in which there was such a correlation in NOESY NMR was assigned
as 25 and the other as 26.
O O
OH
OH
26
24
Scheme 4. Reagents and conditions: (a) BzCl, DAMP, pyridine, rt, overnight, 82%;
(b) BBr3, DCM, ꢁ10 °C, 1 h, 29–48%; (c) NH3, MeOH, H2O, rt, 1 h, 83–96%; (d) CH2Br2,
Cs2CO3, DMF, 145 °C, overnight, 21–29%; (e) NH3, MeOH, H2O, rt, 39–71%.
19 lM; CC50: 37 and 12 lM). The enhanced potency and decreased
cytotoxicity achieved from the methylsulfonamide moiety at R3
position were likely due to both enhanced binding and selectivity
to NS5B. The replacement of the C1-hydroxy group on the naph-
thalene ring of 3 by methoxy group seems tolerated, as demon-
Benzothiadiazine compounds 3a–g, 4, 19a–b, 19e, 20, 21 and
24–26 were tested in HCV NS5B genotype 1b polymerase and rep-
licon assays.14,15 As can be seen from Table 1, compounds 3c, 3e
and 19b with methylsulfonamide moiety (IC50: 0.15, 0.031 and
strated by 19b, 20 and 19e (IC50: 0.048, 0.005 and 0.63 lM)
despite previous studies showed that the corresponding hydroxy
functionality of quinolinone–benzothiadiazine scaffold was essen-
tial for the binding to the enzyme.8 Fluoro and methoxy group at R1
position led to more potent inhibition than proton at R1 position as
revealed by IC50 values for 3f, 3e and 3c (0.037 and 0.031
0.15 M). The introduction of the constrain within the scaffold
0.048
without the moiety (IC50: 8.4, 0.11. and 27
zyme inhibition assay. In the cellular assay the compounds 3c
and 3e with the methylsulfonamide moiety (EC50 4.3 and
2.7 M, both CC50: >100 M) exhibited not only higher potency,
but also higher selectivity window over 3b and 3d (EC50: 26 and
l
M) were significantly more potent than 3b, 3d and 19a
l
M) in the NS5B en-
:
lM vs
l
l
l
proved to be detrimental, as shown by the weak activities of 4
and 24–26.
In the initial in vitro evaluation, compounds 3a and 3g showed
promising activity in both NS5B and replicon assays, with IC50
O
N
O
S
R3
O
O
O
R3
S
a
N
H
H2N
10, 14, 16
+
R1
H2N
Table 1
3
Inhibition of HCV NS5B polymerase genotype 1b and replicon by benzothiadiazine
compounds
17
R = H
18 R3= NHS(O2)Me
19a R1 = H, R3 = H
Compound
IC50
(
lM)
EC50
(lM)
CC50 (lM)
19b R1 = H, R3 = NHSO2Me
O
O
3b
3c
3d
3e
3f
19a
19b
19e
20
21
4
24
25
26
8.4
0.15
0.11
26
4.3
19
2.7
2.9
9.5
21
>100
1.1
0.019
>100
8.4
34
12
37
>100
12
>100
120
13
S
R3
OH N
19c R1 = F, R3 = H
1
3
19d
19e
R = F, R = NHSO2Me
b
N
H
1
3
0.031
0.037
27
0.048
0.63
0.005
<0.005
8.7
R = OMe, R = NHSO2Me
R1
3b R1 = H, R3= H
51
40
3c R1 = H, R3= NHSO2Me
3d R1 = F, R3 = H
110
>100
>100
33
>100
35
1
3
3e
3f
R = F, R = NHSO2Me
R1 = OMe, R3 = NHSO2Me
3g R1 = OH, R3 = NHSO2Me
0.49
0.53
1.3
Scheme 3. Reagents and conditions: (a) PPSE, 160 °C, 1–3 h; (b) 2 equiv BBr3,
ClCH2CH2Cl, argon, ꢁ10 °C, 1 h for 3b–f; 10 equiv BBr3, 38 °C, 2–3 days for 3g.

G. Wang et al. / Bioorg. Med. Chem. Lett. 19 (2009) 4476–4479
4479
10. Krueger, A. C.; Madigan, D. L.; Jiang, W. W.; Kati, W. M.; Liu, D.; Liu, Y.; Maring,
C. J.; Masse, S.; McDaniel, K. F.; Middleton, T.; Mo, H.; Molla, A.; Montgomery,
D.; Pratt, J. K.; Rockway, T. W.; Zhang, R.; Kempf, D. J. Bioorg. Med. Chem. Lett.
2006, 16, 3367.
11. Chen, C.-M.; He, Y.; Lu, L.; Lim, H. B.; Tripathi, R. L.; Middleton, T.; Hernandez, L.
E.; Beno, D. W. A.; Long, M. A.; Kati, W. M.; Bosse, T. D.; Larson, D. P.; Wagner,
R.; Lanford, R. E.; Kohlbrenner, W. E.; Kempf, D. J.; Pilot-Matias, T. J.; Molla, A.
Antimicrob. Agents Chemother. 2007, 51, 4290.
values of 44 nM (3a) and less than 5 nM (3g) and EC50 values of
6.1 M (3a) and 22 nM (3g). However, it was found later that 3a
l
and 3g were not stable under the assay conditions although 3a
was much more stable than 3g. A stability study demonstrated that
3g in phosphate buffer (pH 7.4) was partially converted to hydrox-
ylated derivative 21. Therefore, the inhibitory effects observed for
3g could result from both 3g and 21. Unlike 3a and 3g, compound
21 and other compounds in Table 1 showed good stability. Com-
pound 21 also exhibited excellent activities in both NS5B and rep-
licon assay, with IC50 <5 nM and EC50 = 19 nM. Although 21 has a
structural similarity to 2, the hydroxyl group may provide addi-
tional advantages concerning chemical and biological properties.
In summary, a series of substituted naphthalene–benzothiadi-
azine compounds have been synthesized. Several compounds
showed potent NS5B inhibition while compound 21 exhibited both
potent NS5B and replicon activity. The discovery of 21 as a potent
NS5B inhibitor provides a lead for further optimization.
12. Imai, Y.; Mochizuki, A.; Kakimoto, M. Synthesis 1983, 10, 851.
13. Thanks to this manuscript reviewer’s comments, the chemical synthesis of
compound 21 via
a different route was found in a patent application
(WO2006093801, e.g., 2F), but without activity data.
14. NS5B assay: A modified assay based on a published method (McKercher, G.;
Beaulieu, P. L.; Lamarre, D.; LaPlante, S.; Lefebvre, S.; Pellerin, C.; Thauvette, L.;
Kukolj, G. Nucleic Acids Res. 2004, 32, 422) was used. Assays were performed at
room temperature in 96-well (white, round bottom) plates in 20 mM Tris, pH
7.5 buffer containing 5 mM MgCl2, 5 mM KCl, 1 mM EDTA, 2 mM DTT, 0.01%
BSA. Appropriate serial dilutions of inhibitors in DMSO were prepared and
added to 5 nM NS5b
5 min of incubation, reactions were initiated by the addition of a buffered
substrate mix containing 250 nM 50-biotinylated-rU12 RNA primer, 1
g/mL
poly-rA RNA template, 1 M UTP and 0.625
Ci 5,6-3H-UTP. Total reaction
volumes were 100 L with 5% DMSO (v/v). The reaction was stopped after 2 h
by adding 20 L of 164 g/mL yeast RNA and 10 mg/mL streptavidin PVT SPA
beads in 0.5 M EDTA, pH 8.0. After 30 min, 80 L of 5 M CsCl was added and
D21 (genotype 1b, J4 strain) enzyme in above buffer. After
l
l
l
l
l
l
References and notes
l
incubated for 1 h. Plates were then read using a Wallac MicroBeta reader.
Inhibition data was plotted and fit to a 4-parameter logistic equation to extract
IC50 values. Z prime values under these conditions were >0.6.
1. WHO, Hepatitis C. Incidence/Epidemiology under Section Surveillance and
Control. 2002. http://www.who.int/csr/disease/hepatitis/whocdscsrlyo2003/
en/index.html.
15. HCV replicon assay (EC50, lM): A modified assay based on a published method
(Vrolijk, J. M.; Kaul, A.; Hansen, B. E.; Lohmann, V.; Haagmans, B. L.; Schalm, S.
W.; Bartenschlager, R. J. Virol. Methods 2003, 110, 201) was used. Exponentially
growing Huh-7 cells stably transcfected with luc/neo ET replicon were
maintained in DMEM media supplemented with 10% fetal calf serum and
seeded at a density of 5 ꢂ 103 cells/well in white 96-well plates. Compounds
were dissolved in DMSO, diluted in DMSO in a serial fashion to create an
appropriate range of concentrations, and added to cells approximately 24 h
after plating. The final DMSO concentration in the cell plate was 1%. After 46–
50 h exposure, the media was discarded from the assay plate and the cell
2. Shepard, C. W.; Finelli, L.; Alter, M. J. Lancet Infect. Dis. 2005, 5, 558.
3. Manns, M.; Wedemeyer, H.; Cornberg, M. Gut 2006, 55, 1350.
4. Lindenbach, B. D.; Rice, C. M. Nature 2005, 436, 933.
5. Beaulieu, P. L. Curr. Opin. Invest. Drugs 2007, 8, 614.
6. Rönn, R.; Sandström, A. Curr. Top. Med. Chem. 2008, 8, 533.
7. Tramontano, E. Mini-Rev. Med. Chem. 2008, 8, 1298.
8. 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.
9. Krueger, A. C.; Madigan, D. L.; Green, B. E.; Hutchinson, D. K.; Jiang, W. W.; Kati,
W. M.; Liu, Y.; Maring, C. J.; Masse, S. V.; McDaniel, K. F.; Middleton, T. R.; Mo,
H.; Molla, A.; Montgomery, D. A.; Ng, T. I.; Kempf, D. J. Bioorg. Med. Chem. Lett.
2007, 17, 2289.
monolayers were lysed by addition of 100 lL of either BrightGLO (Promega) or
ATPlite reagent (PerkinElmer) with incubation at 20 °C for 2 min on an orbital
shaker. Following incubation, luminescence was assessed on a SpectraMax M5
plate reader (Molecular Devices). Plots of luminescesce versus log compound
concentration were fit to a 4-parameter logistic equation to determine EC50
and CC50 values.