Non-nucleoside inhibitors of HCV polymerase NS5B. Part 4: Structure-based design, synthesis, and biological evaluation of benzo[d]isothiazole-1,1-dioxides

Article abstract of DOI:10.1016/j.bmcl.2009.08.022

A series of benzo[d]isothiazole-1,1-dioxides were designed and evaluated as inhibitors of HCV polymerase NS5B. Structure-based design led to the incorporation of a high affinity methyl sulfonamide group. Structure-activity relationship (SAR) studies of th

Full text of DOI:10.1016/j.bmcl.2009.08.022

Bioorganic & Medicinal Chemistry Letters 19 (2009) 5652–5656
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Non-nucleoside inhibitors of HCV polymerase NS5B. Part 4: Structure-based
design, synthesis, and biological evaluation of benzo[d]isothiazole-1,1-dioxides
a
a
a
b
b
Javier de Vicente a, , Robert T. Hendricks , David B. Smith , Jay B. Fell , John Fischer , Stacey R. Spencer ,
Peter J. Stengel b, Peter Mohr b, John E. Robinson b, James F. Blake b, Ramona K. Hilgenkamp a, Calvin Yee a,
George Adjabeng a, Todd R. Elworthy a, Jim Li a, Beihan Wang a, Joe T. Bamberg a, Seth F. Harris a,
April Wong a, Vincent J. P. Leveque a, Isabel Najera a, Sophie Le Pogam a, Sonal Rajyaguru a,
Gloria Ao-Ieong a, Ludmila Alexandrova a, Susan Larrabee a, Michael Brandl a, Andrew Briggs a,
Sunil Sukhtankar a, Robert Farrell a
*
a Roche Palo Alto LLC, 3431 Hillview Avenue, Palo Alto, CA 94304, USA
b Array BioPharma Inc, 3200 Walnut Street, Boulder, CO 80301, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of benzo[d]isothiazole-1,1-dioxides were designed and evaluated as inhibitors of HCV polymer-
ase NS5B. Structure-based design led to the incorporation of a high affinity methyl sulfonamide group.
Structure–activity relationship (SAR) studies of this series revealed analogues with submicromolar poten-
cies in the HCV replicon assay and moderate pharmacokinetic properties. SAR studies combined with
structure based drug design focused on the sulfonamide region led to a novel and potent cyclic analogue.
Ó 2009 Elsevier Ltd. All rights reserved.
Received 7 June 2009
Revised 4 August 2009
Accepted 5 August 2009
Available online 8 August 2009
Keywords:
HCV
Polymerase
NS5B
Diels–Alder
DMPK
Hepatitis
C
virus (HCV) is the major causative agent of
Herein, we report a novel series of inhibitors of HCV polymerase
in which the benzothiazine fragment is replaced by a benzo[d]iso-
thiazole-1,1-dioxide.5,6 Our initial SAR studies on the tetramic acid
portion, which were conducted in parallel with the studies on the
benzothiazine-substituted tetramic acids, demonstrated that the
optimal substitutions on the pyrrolidinone ring are tert-butyl and
p-fluorobenzyl groups.7 An early lead compound (2, Fig. 1) exhib-
ited reasonable potency against the genotype 1 NS5B enzyme
blood-borne non-A, non-B hepatitis, and it is estimated to infect
over 170 million people worldwide.1 Chronic HCV infections have
been associated with liver fibrosis, liver cirrhosis, heptacellular
carcinoma, and other forms of liver dysfunction.2,3 Given the wide-
spread impact of this disease, there is a substantial medical need
for new anti-HCV agents to complement current therapies.4 In this
article we report on our efforts to develop orally bioavailable, non-
nucleoside inhibitors of the HCV RNA-dependent RNA polymerase.
In the previous Letter, we described the synthesis and SAR stud-
ies of a new class of benzothiazine-substituted tetramic acids as
non-nucleoside inhibitors of HCV NS5B polymerase.3 While these
compounds (1; Fig. 1) were found to be potent inhibitors of HCV
polymerase both in enzymatic and replicon assays, they displayed
low oral bioavailability and high clearance (Fig. 1). These findings
were attributed to the metabolic instability of the benzothiazine
ring, which also intervened in an undesired intramolecular hydro-
gen bond that contributed to their in poor permeability.
(IC50 = 0.069 lM) and HCV subgenomic replicon (EC50 = 0.648 lM).
Compound 2 provided a good starting point for further optimiza-
tion. Our early studies showed that the addition of a methyl sulfon-
amide group to the 7-position of the benzothiazine ring
dramatically improved the potency in that series.3 Installation of
the methyl sulfonamide group at the 7-position of the benzo[d]iso-
thiazole-1,1-dioxides (compound 3, Fig. 1) provided an excellent
boost in enzyme activity (IC50 = 0.005
con inhibition potency (EC50 = 4.924
l
M) but showed poor repli-
l
M).
An atom extension of the methyl sulfonamide group (com-
pound 5, Fig. 1) retained good potency against the enzyme
(IC50 = 0.003
potency (EC50 = 0.1
l
M), and showed dramatically improved replicon
* Corresponding author. Tel.: +1 6508555632; fax: +1 6508556908.
E-mail address: javier.devicente@roche.com (J. de Vicente).
lM). To test the optimal substitution on the
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.08.022

J. de Vicente et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5652–5656
5653
O
S
NH
O
S
O
O
O
O
S
N
HO
N
R
HO
N
R1
N
H
O
O
1a, R1, R2 = H
2, R = H
3, R = NHSO2Me
F
1b, R1, R2 = Me
F
R2
O
O
O
O
O
S
NH
O
S
S
N
S
O
O
HO
N
N
HO
N
H
N
O
O
F
4
F
5
Figure 1. Previous lead compounds 1a, 1b and scaffold hopping strategy.
benzo[d]isothiazole ring, the extended methyl sulfonamide group
was also installed at the 6-position (compound 4, Fig. 1). Substitu-
tion at this position was tolerated but did not provide a similar
inhibitors potency was evaluated in the HCV subgenomic replicon
cell based assay where 40% human serum (HS) is added. The results
of these assays are shown in Table 1. Replacement of the sulfon-
amide NH by a methylene linker had a mild detrimental effect in
replicon potency (compare 5 and 7). These compounds were likely
protein bound as the antiviral potency was reduced by addition of
HS. In order to increase the polarity of the molecules, the sulfon-
amide or sulfone group was exchanged by a sulfamide. Relative
to 5, sulfamides 10, 11 and 12 showed similar levels of replicon po-
tency and as predicted the serum shift was modestly improved.
Interestingly, the introduction of a single methyl group on the
NH sulfamide was very well tolerated (compound 12). This modi-
fication also reduced serum shift so it was incorporated into our
initial sulfonamide, leading to compounds 136 and 14. Cyclization
of the methyl sulfonamide onto the nitrogen (compound 15) did
not provide an improvement in potency.
boost in replicon potency (IC50 = 0.043 lM, EC50 = 3.844 lM).
X-ray analysis of compounds 1b and 5 demonstrates the opti-
mal location of the methyl sulfonamide group.8 The inclusion of
a CH2 spacer between the benzo[d]isothiazole-1,1-dioxide ring an
the sulfonamide moiety allows the functionality to appropriately
access the far right side of the palm binding pocket. As shown in
Figure 2, the sulfonamide oxygen atoms interact directly with
Asn291 and with Ser288 via a conserved water molecule. In con-
trast to the benzothiazine series, the sulfonamide NH does not
interact strongly with Asp318. Thus, we decided to explore sulfon-
amide analogues were this NH was substituted or replaced by C in
hopes of improving cell permeation and replicon inhibition
potency.4
A series of benzo[d]isothiazole-1,1-dioxides were evaluated for
in vitro potency against the NS5B enzyme9 and in an HCV subge-
nomic replicon (GT-1b)10,11 cell based assay in the presence of 5%
fetal bovine serum (FBS). The effect of protein binding on the
In vivo PK data were obtained for our most potent compounds
(11, 6 and 14, Table 2) following administration of a single oral
dose to rats or monkeys. Sulfonamide 11 displayed poor exposure
in rats due to low oral bioavailability and rapid clearance. This data
was supported by our in vitro data: poor Caco-2 permeability and
low rat liver microsomal stability. Compound 6 showed low clear-
ance in vivo and modest levels of exposure and oral bioavailability
in rat and monkey. The N-methylated sulfonamide 14 also showed
encouraging levels of exposure. In general, we observed improved
oral exposure in analogues bearing the 3-methyl-4-fluorobenzyl
lipophilic group.
Compound 6 showed promising levels of exposure but poor rep-
licon activity in presence of 40% human serum. It is estimated that
the minimum efficacious concentration of an HCV inhibitor should
be greater than the replicon EC90 in presence of 40% HS.12 In order
to enhance the potency of our benzo[d]isothiazole-1,1-dioxides
series, we turned our attention to the binding conformation and
interactions of sulfonamide 5.
Molecular orbital calculations with solvent corrections,13 sug-
gested the bound sulfonamide side chain conformation of 5 is ca.
2.1 kcal/mol less favored relative to the low energy conformation
containing an internal hydrogen bond to the SO2 of the benzo[d]iso-
thiazole ring. Interestingly, the small molecule X-ray structure of
N-benzyl methane-sulfonamide (CSD entry KOYVID0001) displays
a very similar overall conformation to the analogous side chain
portion of 5 in our bound X-ray structure (Fig. 2), which further sug-
gests the system is not highly strained.
We hypothesized that added potency could be gained by lock-
ing the sulfonamide group in the X-ray bound conformation by
connecting the sulfonamide nitrogen and the C6 position of the
Figure 2. X-ray crystal structure overlap of compounds 1b and 5 bound to the palm
site of CD21 NS5B polymerase.

5654
J. de Vicente et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5652–5656
Table 1
benzo[d]isothiazole ring with an ethylene linker (Fig. 3). Since sub-
stitution at these positions was tolerated (recall compound 4), we
embarked on the synthesis of fused ring analogue 16.
The synthesis of fused ring analogue 16 began with the con-
struction of the benzo[d]isothiazole-1,1-dioxide core. This tricyclic
system was readily assembled by a Diels–Alder reaction of diene
19 and known dienophile 20 (Scheme 1).14 Olefination of ketone
17 followed by standard manipulations led to aldehyde 18. Con-
Enzyme and replicon data for benzo[d]isothiazole 1,1-dioxide derivatives 5–15
O
O
HO
N
R2
S
N
R1
O
F
densation of
a,b-unsaturated aldehyde 18 with dimethyl amine
Compds R1
R2
NS5B IC50
(
l
M) Replicon EC50 (lM)
a
a,b
provided diene 19, which upon treatment with a stoichiometric
amount of dienophile 20 resulted in the formation of the desired
benzo[d]isothiazole 1,1-dioxide core 21 as well as compound 22.
The Diels–Alder cycloaddition of 19 and 20 is followed by a
double elimination and aromatization (Scheme 2, path a). Interest-
ingly, cycloadduct 22 was also isolated. This compound was possi-
bly obtained by a Diels–Alder reaction between rearranged diene
26 and dienophile 20 (Scheme 2, path b). A postulated mechanism
for the formation of diene 26 involves an unusual conrotatory ring
closure of diene 19 followed by an acid promoted double bond
isomerization and a conrotatory ring opening. The stereospecificity
of the above reactions are in agreement with the optimal orbital
overlap of dienes 19 and 26 (HOMO) with unsaturated amide 20
(LUMO).
c
5% FBS 40% HS (FS)
O
O
O
S
5
H
0.003
0.004
0.004
0.009
0.006
0.003
0.003
0.001
0.01
0.1
3.728 (37.3)
2.573 (31.3)
4.796 (17)
N
H
O
S
6
Me
H
0.082
0.282
0.235
0.228
0.105
0.076
0.272
0.108
0.129
0.202
N
H
O
O
7
S
O
O
S
8
H
4.662 (19.8)
2.896 (12.7)
1.299 (12.4)
1.166 (15.3)
1.272 (5.9)
3.038 (28.1)
1.067 (8.27)
4.66 (23.1)
O
O O
S
O
S
9
H
O
O
O
O
S
O
NH2
O
O
S
5
O
N
HO
N
5
N
S
N
HO
N
N
H
5
O
O
6
6
S
10
11
12
13
14
15
H
N
NH2
O
H
O
F
16
F
O
O
S
Me
H
N
H
NH2
Figure 3. Rational design of constrained analogue 16 from initial lead 5.
O O
S
boc
boc
boc
N
NH2
N
N
N
d
a-c
O O
S
O
19
N
O
Cl
Me
H
17
18
O
N
O
N
O
t-Bu
S
e
20
O O
S
O
0.005
0.012
Br
N
O
S
O
O
O
t-Bu
O
O
boc
S
N
S
N
N
t-Bu
N
+
N
boc
O
22
21
a
IC50 and EC50 values for inhibition calculated from two independent
determinations.
Scheme 1. Reagents and conditions: (a) EtO2CCH2PO(OMe)2, MeONa, THF, 0 °C; (b)
DIBAL, THF, À78 °C; (c) MnO2, DCM, 25 °C, 56% (three steps); (d) NHMe2, toluene,
90 °C, 100%; (e) TEA, MeCN, 70 °C, 21 (43%), 22 (12%).
b
CC50 > 50 M for all analogues.
c
IC50 fold-shift values (40% human serum/5% FBS) shown in parentheses.
Table 2
Mean pharmacokinetic parameter values for selected compounds 11, 6 and 14 following single dose administration to rats and monkeys
Compds
Caco-2 A to B (ER)a
Species
LMb
(lL/min/mg)
AUC (h ng/mL)
T1/2 (h)
Cl (mL/kg/min)
F(%)
11
6
0.29 (6.7)
0.09 (31)
Ratc
239
431
23
102
2750
1230
582
4.7
48
8.4
10
8.2
17
11
Ratc
4.18
1.75
3.37
Monkeyd
Ratc
14
0.21 (4.3)
939
34.4
24
a
b
c
Caco-2 assay run for 21 days. ER = (B-to-A)/(A-to-B).
LM: in vitro liver microsomal data.
Rat PK study: dose iv = 2 mg/Kg, dose po = 5 mg/Kg.
d
Cynomolgus monkey PK study: dose iv = 2 mg/Kg, dose po = 5 mg/Kg.

J. de Vicente et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5652–5656
5655
O
O
t-Bu
O
boc
S
N
N
Br
NEt3
O
N
23
Path a
19 + 20
Path b
O
S
O
t-Bu
O
boc
S
N
NEt3 (xs)
CH3CN, 70oC, 4h
N
t-Bu
N
+
N
O
boc
O
22, 12%
21, 43%
20
boc
boc
N
boc
N
N
Et3NHBr
N
N
N
26
24
25
Scheme 2. Proposed mechanism for the formation of tetrahydroisoquinolines 21 and 22.
The Diels–Alder adduct 21 was elaborated to the desired
benzo[d]isothiazole-1,1-dioxide 27 by first switching the t-butox-
ycarbonyl group to the desired methyl sulfonamide (Scheme 3).
Amide deprotection gave 27, which was then activated to provide
imidate 28. Final coupling of this imidate with the sodium salt de-
rived from tetramic acid 29 led to the desired piperidine fused
benzo[d]isothiazole-1,1-dioxide 16.
Compound 16 displayed NS5B inhibition activity of 0.005 lM.
We speculate that less than ideal interactions with D318 may be
partially offsetting the anticipated benefit of restricting the sulfon-
amide side chain conformation. Regardless, for this series fused
analogue 16 exhibited the most potent activity (0.026
replicon assay. However, this compound showed a high serum shift
(4.938 M). In order to confirm our initial hypothesis, compound
16 was soaked into C
21 NS5B for X-ray analysis.15 The observed
lM) in our
l
D
binding conformation of cyclic sulfonamide 16 was close to the
predicted ground state conformation and the sulfonamide group
position was almost identical to the corresponding acyclic ana-
logue (Fig. 4).
In conclusion, we have reported on the structure activity evalu-
ation of a novel benzo[d]isothiazole-1,1-dioxides series of non-
nucleoside inhibitors of HCV NS5B polymerase. The incorporation
of the extended methyl sulfonamide group led to potent inhibitors
in our standard and serum shift replicon assays. The most relevant
compounds in this series contained a 3-methyl-4-fluorobenzyl
Figure 4. X-ray crystal structure of compound 16 bound to the palm site of CD21
NS5B polymerase.
O
S
O
S
O
O
O
O
group. Pharmacokinetic data obtained from selected compounds
indicated only moderate levels of exposure. Structure-based design
and molecular modeling were employed to guide the discovery
and optimization of the methyl sulfonamide group. As a result,
the novel cyclic sulfonamide analogue 16 was shown to be the
most potent analogue in our replicon assay. The novel synthesis
of compound 16 also highlights an efficient DA approach for the
synthesis of complex polyheterocyclic systems. Although com-
pound 16 exhibited good potency in both the biochemical NS5B
and the replicon cell-based assays, the reduced potency in pres-
ence of 40% HS limited the development of this compound as drug
candidate.
O
O
S
S
HN
N
d-e
N
a-c
N
21
O
EtO
HO
27
28
f
N
O
29
F
O
S
O
O
O
S
N
HO
N
N
O
Acknowledgments
16
F
The authors thank Drs. Eric Sjogren, Hans Maag, and John Josey
for their support and helpful discussion during the course of this
work.
Scheme 3. Reagents and conditions: (a) TFA, DCM, 25 °C, 100%; (b) MsCl, Py, DCM,
25 °C; (c) BCl3, DCM, 25 °C; (d) SOCl2, DMF; (e) EtOH, 81% (four steps); (f) NaH, THF,
45%.

5656
J. de Vicente et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5652–5656
Klumpp, K.; Symons, J.; Browner, M. F.; Cammack, N.; Najera, I. J. Virol. 2006,
References and notes
80, 6146.
9. For a detailed description of the enzyme assay see: Ma, H.; Jiang, W.-R.;
Robledo, N.; Leveque, V.; Ali, S.; Lara-Jaime, T.; Masjedizadeh, M.; Smith, D. B.;
Cammack, N.; Klumpp, K.; Symons, J. J. Biol. Chem. 2007, 282, 29812.
10. For a detailed description of the replicon assay see: Klumpp, K.; Leveque, V.; Le
Pogam, S.; Ma, H.; Jiang, W.-R.; Kang, H.; Granycome, C.; Singer, M.; Laxton, C.; Hang, J.
Q.; Sarma, K.; Smith, D. B.; Heindl, D.; Hobbs, C. J.; Merrett, J. H.; Symons, J.; Cammack,
N.; Martin, J. A.; Devos, R.; Najera, I. J. Biol. Chem. 2006, 281, 3793.
11. A representative number of benzo[d]isothiazole-1,1-dioxides analogues (5, 6,
11, 14) were also evaluated in a GT-1a replicon assay. This class of compounds
had a minimal loss in inhibitory activity of two to eightfold relative to our
standard GT-1b replicon assay.
12. (a) Kneteman, N. M.; Howe, A. Y. M.; Gao, T.; Lewis, J.; Pevear, D.; Lund, G.;
Douglas, D.; Mercer, D. F.; Tyrrell, D. L. J.; Immermann, F.; Chaudhary, I.; Speth,
J.; Villano, S. A.; O’Connell, J.; Collett, M. Hepatology 2009, 49, 745; (b) Chandra,
P.; Raible, D.; Harper, D.; Speth, J.; Villano, S.; Bichier, G.; Presented at the
Digestive Disease Week, Los Angeles, CA, May 21–24, 2006.
13. GAMESS Version 24 March 2007 (R6). Schmidt, M. W.; Baldridge, K. K.; Boatz, J.
A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. J.; Koseki, S.; Matsunaga, N.; Nguyen, K.
A.; Su, S.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput. Chem. 1993,
14, 1347. All geometries and energetics were computed at the 6-311++G(d,p)
level of theory. Aqueous solvent corrections were applied via AMSOL5.0 with
the SM2.1 model. Cramer, C. J.; Hawkins, G. D.; Lynch, G. C.; Giesen, D. J.; Rossi,
I.; Storer, J. W.; Truhlar, D. G.; Liotard, D. A. (1995) AMSOL version 5.0, QCPE
Program 606.
1. Choo, Q. L.; Weiner, A. J.; Overby, L. R.; Bradley, D. W.; Houghton, M. Science
1989, 244, 359.
2. For a review of HCV epidemiology, pathogenesis, and clinical treatment, see:
(a) Lauer, G. M.; Walker, B. D. N. Engl. J. Med. 2001, 345, 41; (b) Di Bisceglie, A.
M. Lancet 1998, 351; (c) Alter, M. J.; Kruszon-Moran, D.; Nainan, O. V.;
McQuillan, G. M.; Gao, F.; Moyer, L. A.; Kaslow, R. A.; Margolis, H. S. N. Engl. J.
Med. 1999, 341, 556.
3. Manns, M. P.; McHutchison, J. G.; Gordon, S. C.; Rustgi, V. K.; Shiffman, M.;
Reindollar, R.; Goodman, Z. D.; Koury, K.; Ling, M.-H.; Albrecht, J. K. Lancet
2002, 347, 975.
4. For progress reports on non-nucleoside HCV polymerase inhibitors see: Koch,
U.; Narjes, F. Curr. Top. Med. Chem. 2007, 7, 1302; Rönn, R.; Sandström, A. Curr.
Top. Med. Chem. 2008, 8, 533; Zapf, C. W.; Bloom, J. D.; Levin, J. I. Ann. Rep. Med.
Chem. 2007, 281. see also references therein.
5. Blake, J. F.; Fell, J. B.; Fischer, J. P.; Hendricks, R. T.; Spencer, S. R.; Stengel, P. J.
U.S. Patent Application 2006/0252785, 2006.
6. During the course of this study, similar HCV NS5B inhibitors were described in
the literature. See: Kim, S. S.; Tran, M. T.; Ruebsama, F.; Xianga, A. X.; Ayidaa, B.;
McGuire, H.; Ellis, D.; Blazel, J.; Tran, C. V.; Murphy, D. E.; Webber, S. E.; Zhou,
Y.; Shah, A. M.; Tsan, M.; Showalter, R. E.; Patel, R.; Gobbi, A.; LeBrun, A. L.;
Bartkowski, D. M.; Nolan, T. G.; Norris, D. A.; Sergeeva, M. V.; Kirkovsky, L.;
Zhao, Q.; Han, Q.; Kissinger, C. R. Bioorg. Med. Chem. Lett. 2008, 18, 4181.
7. For a detailed description of the synthesis of these benzo[d]isothiazole-1,1-
dioxides analogues see Refs..5,6
14. Burri, K. F. Helv. Chim. Acta 1989, 72, 1416.
15. Atomic coordinates of the structure 16 bound to CD21 NS5B polymerase were
8. Atomic coordinates of the structure 5 bound to CD21 NS5B polymerase were
deposited with the RCSB Protein Data Bank (PDB) under the accession code
3H5U. The crystal structure was determined following previously described
procedures: LePogam, S.; Kang, H.; Harris, S. F.; Leveque, V.; Giannetti, A. M.;
Ali, S.; Jiang, W.; Rajyaguru, S.; Tavares, G.; Oshiro, C.; Hendricks, T.;
deposited with the RCSB Protein Data Bank (PDB) under the accession code
3H5S. The crystal structure was also determined following previously
described procedures as noted in Ref.8