Non-nucleoside inhibitors of HCV NS5B polymerase. Part 1: Synthetic and computational exploration of the binding modes of benzothiadiazine and 1,4-benzothiazine HCV NS5b polymerase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2009.04.119

The importance of internal hydrogen bonding in a series of benzothiadiazine and 1,4-benzothiazine NS5b inhibitors has been explored. Computational analysis has been used to compare the protonated vs. anionic forms of each series and we demonstrate that activity against HCV NS5b polymerase is best explained using the anionic forms. The syntheses and structure-activity relationships for a variety of new analogs are also discussed.

Full text of DOI:10.1016/j.bmcl.2009.04.119

Bioorganic & Medicinal Chemistry Letters 19 (2009) 3637–3641
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Non-nucleoside inhibitors of HCV NS5B polymerase. Part 1: Synthetic
and computational exploration of the binding modes of benzothiadiazine
and 1,4-benzothiazine HCV NS5b polymerase inhibitors
a
a
a
a
a
Robert T. Hendricks b, , Jay B. Fell , James F. Blake , John P. Fischer , John E. Robinson , Stacey R. Spencer ,
Peter J. Stengel a, April L. Bernacki a, Vincent J. P. Leveque b, Sophie Le Pogam b, Sonal Rajyaguru b,
Isabel Najera b, John A. Josey a, Jason R. Harris a, Steven Swallow b,
*
a Array BioPharma, Inc., 3200 Walnut Street, Boulder, CO 80301, USA
b Roche Palo Alto, LLC, 3431 Hillview Avenue, Palo Alto, CA 94304, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The importance of internal hydrogen bonding in a series of benzothiadiazine and 1,4-benzothiazine NS5b
inhibitors has been explored. Computational analysis has been used to compare the protonated vs. anio-
nic forms of each series and we demonstrate that activity against HCV NS5b polymerase is best explained
using the anionic forms. The syntheses and structure–activity relationships for a variety of new analogs
are also discussed.
Received 25 February 2009
Revised 23 April 2009
Accepted 24 April 2009
Available online 3 May 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
HCV
NS5b
Polymerase
Inhibitor
Non-nucleoside
Replicon
1,4-Benzothiazine
Computational
GAMESS
Conformation
Anion
Approximately 200 million people worldwide are infected with
Hepatitis C virus (HCV), the leading cause of cirrhosis of the liver.1
Infection with HCV can lead to hepatocellular carcinoma, and ulti-
mately, liver failure requiring transplantation surgery.2 The current
best treatment option consists of a combination of pegylated inter-
hydrogen bonds, as shown in Figure 1. As described, these hydro-
gen bonds are thought to stabilize the bound conformation, thus
contributing to the potency observed for this class of NS5b
inhibitors.4b
feron-a
-2a and ribavirin, a broad-spectrum anti-viral agent.3 The
limited efficacy and potential for severe side-effects of this
approach continues to drive the search for more efficacious and
targeted therapies.
O
O
O
O
S
N
S
OH
N
N
OH
In 2002 GSK reported that compound 1 inhibited the HCV NS5b
X
N
H
polymerase with an IC50 of 0.03
later revealed as a more potent analog with an enzyme IC50 of
0.01 M and a subgenomic HCV replicon IC50 of 0.04 M. A key fea-
l
M (Fig. 1).4a Compound 2 was
O
N
l
l
1: R = i-propyl; X = H
2: R = c-propyl; X = F
ture of these compounds is their ability to form two internal
R
F
3
* Corresponding author. Tel.: +1 650 855 6486.
E-mail address: Than.Hendricks@Roche.com (R.T. Hendricks).
Present address: AstraZeneca R&D, Mereside, Alderley Park, Macclesfield, Chesh-
ire SK10 4TG, UK.
Internal Hydrogen Bonds
Figure 1. Internal hydrogen bond networks for 4-hydroxyquinolone-benzothiadi-
azines 1 and 2 and isoquinoline 3.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.04.119

3638
R. T. Hendricks et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3637–3641
O
O
O
O
N
O
O
O
O
S
S
COR
ONa
N
S
O
O
N
HN
a
CO2Et
+
O
N
H
N
N
N
H
O
4
18: R = OEt
19: R = OH
N
17
b
N
O
16
c
H2NSO2
O
O
S
O
6
5
O
N
cpKa: 9.1
cpKa: 11.4
cpKa: 7.6
O
N
d
O
O
N
O
S
O
N
S
HO
OH
N
R'
N
R
N
H
6
20
N
5
X
N
R
O
Scheme 2. Reagents and conditions: (a) DMF, 110 °C; (b) THF, NaOH; (c) (1) oxalyl
chloride, toluene; (2) 2-aminobenzenesulfonamide, TEA, toluene; (d) Cs2CO3, EtOH,
reflux.
cpKa: 7.6
cpKa: 6.5
8: R = H
7a: X = CH; R' = H; R = isoamyl
cpKa: 4.6
F
9: R = Me
Figure 2. Set of analogs with the capacity to form only a single internal hydrogen
bond.
HS
S
a
c, d
EtO2C
H2N
N
R
We recently described a new series of 3-hydroxyisoquinoline
22: R = H
O
S
O
21
b
derived NS5b inhibitors (Fig. 1, compound 3), which demonstrated
23: R = Boc
reasonable NS5b inhibitory potency (IC50: 2.2 l
M).5 While 3 can
O
O
S
also exist in an internal hydrogen-bond stabilized conformation
similar to that of 2, we hypothesized that a delocalized enolic
anion (pKa of 3 = 7.3) is the active form that interacts with the
NS5B polymerase enzyme. To test this hypothesis further, we
designed a set of compounds related to 1 and 3 (Fig. 2) that would
probe both the importance of this key hydrogen bond and the
anionic binding model. Compounds 4 and 54c can only form one
internal hydrogen bond to the thiadiazine ring system, and neither
compound can form a delocalized anion in the quinolone ring. The
combined 2-quinolone and 1,4-thiazine ring system in 6 lacks both
hydrogen bond elements along the upper portion of the ring sys-
tem. Compounds 7, 8 and 9 also possess the 1,4-thiazine system,
but retain a hydrogen bond donating OH group on the left-hand
ring system and have the ability to form delocalized enolic anions.
Both internal hydrogen bonds have been blocked entirely in com-
pound 9. In general, measured pKa’s for compounds across these
series have tracked very closely with calculated pKa (cpKa) values.6
Our synthesis of compound 4 is shown in Scheme 1. 2-Amino-
methylbenzoate was acylated using 3-methylbutanoyl chloride to
give 10, which was reduced with lithium aluminum hydride to give
the anilino alcohol 11. Oxidation to the aldehyde 12 was accom-
plished using MnO2. Treatment with Meldrum’s acid followed by
chlorination gave quinolone acyl chloride 14 in modest yield. Reac-
tion with 2-amino-benzenesulfonamide and subsequent ring clo-
sure of 15 using Cs2CO3 in ethanol at reflux provided 4.
N
e
EtO2C
N
O
N
H
24
6
Scheme 3. Reagents and conditions: (a) ethyl 4-chloro-3-oxobutanoate, MeOH; (b)
BOC2O, THF, DMAP; (c) MCPBA, THF; (d) TFA, CH2Cl2; (e) compound 12, THF, 130 °C;
then aq NaHCO3 workup.
The synthesis of the 1H-quinolin-4-one 5 was carried out by
reaction of the isatoic anhydride 167 and the sodium enolate 178
to give the quinolone ester 18 (Scheme 2). Saponification to the
acid was achieved in quantitative yield using NaOH in THF. Activa-
tion with oxalyl chloride and reaction with 2-aminobenzene-sul-
fonamide gave intermediate 20. Using conditions similar to that
found in Scheme 1, the product was cyclized with CsCO3 in ethanol.
The 1,4-thiazine 6 was synthesized using the route shown in
Scheme 3. Alkylation of 2-aminothiophenol 21 with ethyl 4-
chloro-3-oxobutanoate resulted in a spontaneous ring closure to
give 1,4-benzothiazine 22. Direct oxidation of 22 proved problem-
atic. However, following acylation with Boc2O, intermediate 23
underwent efficient oxidation and spontaneous olefin migration
using MCPBA. Condensation of the deprotected 1,4-thiazine diox-
ide 24 (as the TFA salt) with aldehyde 12 in a sealed tube at
130 °C gave the desired cyclized product 6 in low (15%) yield.
The syntheses of a variety of key isatoic anhydride intermedi-
ates are outlined in Scheme 4. An SnAr reaction on 2-bromo-5-flu-
oro-benzoic acid using 2-cyclopropylethylamine gave aniline 26.
Treatment with triphosgene in the presence of K2CO3 provided isa-
toic anhydride 27.4d Desulfurization of 289 followed by amine dis-
placement and saponification gave the key aza intermediate 29,
which was converted to anhydride 30 as previously described.
Additional variously substituted anhydrides were prepared in a
similar fashion using readily available materials. Keto-ester 31
was first converted to cyclic anhydride 32 using urethane and
POCl3. Alkylation of the sodium salt of 32 with isoamylbromide
gave the monocyclic precursor 33 in good (73%) yield.
CO2Me
NH
R
CO2Me
NH2
d
a
b
NH
O
N
11: R = CH2OH
12: R = CHO
10
c
H2NSO2
O
O
O
S
COR'
N
N
H
N
N
O
f
g
O
N
O
Following Scheme 5, the 1,4-thiazines 7a and 34 were synthesized
using the key 1,4-thiazine dioxide intermediate 24 from Scheme 3.
Reaction of 24 with the appropriately substituted isatoic anhydrides
gives the desired products in modest yield. In the case of 34, use of
DBU gave poor conversion, however, switching to the sodium salt of
24 led to a modest (41%) yield of the desired 1,4-thiazine.
13: R' = OH
14: R' = Cl
15
e
4
Scheme 1. Reagents: (a) 3-methylbutanoyl chloride, NaHCO3, THF; (b) LAH, THF;
(c) MnO2, DCM; (d) Meldrum’s acid, ethylenediamine, MeOH, AcOH; (e) SOCl2; (f) 2-
aminobenzenesulfonamide, toluene; (g) Cs2CO3, EtOH.

R. T. Hendricks et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3637–3641
3639
Table 1
O
N
NS5b and replicon inhibitory potency of compounds 1–9 and 34
CO2H
NH
CO2H
Br
F
F
F
O
a
b
Compound #
NS5b Enzyme IC50
(
lM)
GT-1b Replicon EC50 (lM)
O
1
2
3
4
5
6
7a
8
9
1.1
0.1
2.2
682
>1000
868
2.3
6.9
19.5
87
0.87
0.26
0.44
nd
nd
nd
2.5
7.6
nd
25
27
O
26
O
CO2H
NH
CO2Et
Cl
F
F
F
b
c
N
MeS
N
N
N
O
28
29
30
34
nd
O
N
O
CO2Et
O
O
N-methyl aniline precursor in the acylation step gave 37 as the sole
product. Treatment with BBr3 led directly to the desired N-methyl
1,4-thiazine 9.
e
d
O
O
N
H
O
31
32
Heterocycles 1–9 and 34 were tested against the isolated HCV
NS5b polymerase (GT-1b) and in a replicon assay system.10 Table
1 summarizes the potency data for this set of compounds. Analogs
4 and 5, which retain the thiadiazine moiety found in compounds 1
and 2 but lack an ionizable 4-hydroxy group, showed greater than
two orders of magnitude loss in potency against the enzyme.
Compound 6, which differs from 4 by a single N-for-C exchange,
exhibited a similar ca. 800-fold drop in activity compared to 1.
The 1,4-thiazine 7a, now with an ionizable 4-hydroxy quinolone
33
Scheme 4. Reagents and conditions: (a) 2-cyclopropylethylamine HCl, K2CO3,
CuBr2, THF; (b) triphosgene, K2CO3, EtOAc; (c) (1) Ra-Ni, EtOH; (2) isoamylamine,
DIEA, EtOH, 150 °C (sealed tube); (3) aq NaOH, dioxane (d) urethane, POCl3, 90 °C;
(e) isoamylbromide, NaH, DMA, 80 °C (sealed tube).
O
O
S
N
OH
group, inhibited the NS5b enzyme with an IC50 of 2.3
lM and
a
b
showed promising replicon activity (2.5 M). This finding sug-
l
O
O
gested the conformational stabilization provided by two internal
hydrogen bonds was not an absolute requirement for NS5B activity
within this general class of inhibitors. The expanded 6-membered
1,4-thiazine ring of 8 is less well tolerated in the isoquinoline ser-
7a
N
O
S
O
O
EtO2C
S
N
H
HO
ies (6.9 lM vs 2.2 lM for 3), and as expected, methylation of the
N
24
nitrogen (i.e., 9) further reduced activity. The mono-cyclic 1,4-thi-
azine analog 34 is notably less active against the NS5b enzyme,
regardless of its greater ionization potential.
N
O
34 cpKa: 3.6
Publication of the X-ray co-crystal structure of 1 complexed
with the NS5b polymerase demonstrated that this class of inhibi-
tors adopts a shallow U-shaped conformation in the active site,
with the internal H-bonding partners aligned as depicted in Figure
1.4b,11 Given the general similarity of compounds 1 and 7, it is rea-
sonable to assume that they may inhibit the NS5b polymerase in a
similar fashion, while maintaining key contacts with the polymer-
ase. However, it would seem unlikely that 1,4-thiazines 7 can
adopt a similar shallow U-shaped conformation in the NS5b active
site, without a significant energetic cost, due to the lack of compli-
mentary H-bond partners. To investigate this more fully, we
Scheme 5. Reagents and conditions: (a) compound 16, DBU, THF/EtOAc, THF; (b)
compound 33, NaH, THF, reflux.
The preparation of isoquinoline-1,4-thiazines 8 and 9 is out-
lined in Scheme 6. Acylation of the dianion of 2-methanesulfonyl-
phenylamine with 355 led to the spontaneously cyclized
intermediate imine 36. Deprotection of this imine with BBr3 gave
the tautomeric 1,4-benzothiazine dioxide 8 as a single product.
Direct methylation of 8 with MeI/K2CO3 proved unsuccessful,
giving primarily the C-alkylated imine product. Switching to the
O
O
S
N
MeO
N
O
O
MeO
O
S
HO
36
OMe
a or b
N
R
N
c
F
N
NHR
MeO
O
SO2
35
8: R = H
F
9: R = Me
F
N
37: R = Me
F
Scheme 6. Reagents and conditions: (a) 2-methanesulfonyl-phenylamine; 2.2 equiv n-BuLi, THF, ꢀ78 °C; (b) (2-methanesulfonylphenyl)methylamine, 2.2 equiv n-BuLi, THF,
ꢀ78 °C; (c) BBr3, DCM.

3640
R. T. Hendricks et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3637–3641
computed the energy of the ‘unbound’ (i.e., lowest energy) struc-
tures of N-methyl enolic prototypes of 1 and 7 via ab initio molec-
ular orbital calculations at the 6-311++G(d,p)//6-311++G(d,p) level
of theory with the GAMESS program.12 We then compared these min-
imum energy conformations with the published thiadiazine
structure.
For the thiadiazine enolic prototype, we found the lowest
energy conformation corresponds to a doubly hydrogen bonded
structure, illustrated in Figure 3 (top row). Interestingly, our com-
puted lowest energy structure is not fully co-planar, but rather
adopts a U-shaped conformation quite similar to that in the
published thiadiazine/NS5b X-ray co-crystal structure. The ‘bend’
in our computed structure is caused by the boat-like geometry of
the thiadiazine ring.13
The minimum energy structure of the 1,4-thiazine enolic proto-
type was found to possess a substantially different orientation of
the quinolone and 1,4-thiazine rings. The computed torsion for
the C(OH)@C–C@C(SO2) angle is 51.5°, compared to 2.6° for the
equivalent C(OH) @C–C@N(SO2) angle in the quinolone thiadiazine
prototype (Fig. 3; top row). Based on the published X-ray co-crystal
of 1, it is unlikely that this lowest energy structure of the 1,4-thi-
azine prototype corresponds to an ‘active’ conformation.
Constraining the torsion angle connecting the 1,4-thiazine and
quinolone rings to match that of the thiadiazine series gave a struc-
ture ca. 12 kcal/mol higher in energy. Given the relative inhibitory
Figure 3. Lowest energy conformations of the enolic (top row) and anionic (middle
row) thiadiazine and 1,4-thiazine series prototypes. Bottom row: computed NS5B-
bound conformations for each anionic form (i.e., ca. 180° shifted about the central
C–C bond).
Table 2
NS5b and replicon inhibitory potency of compounds 7a–o
Compound #
X
R0
R
NS5b IC50
M)
Replicon EC50
M)
(
l
(l
potencies of the 1,4-thiazine (2.3 lM) and thiadiazine (1.1 lM)
series of compounds, it seems unlikely that a 12 kcal/mol confor-
mational penalty must be overcome in order for the 1,4-thiazines
to bind to the NS5b polymerase.
7a
7b
7c
7d
7e
7f
CH
CH
CH
CH
CH
N
H
2.3
1.1
2.4
1.3
3.8
1.4
0.83
2.7
17.5
2.5
1.2
1.6
0.64
7.1
nd
We sought to develop an alternate binding model for these com-
pounds through consideration of the acidic nature of the hydroxy
quinolone ring system; we hypothesized that the anionic species,
resulting from loss of the 2-quinolone hydroxyl proton, may be the
active form of the inhibitors.5,6 Upon computing the lowest energy
and ‘NS5b-bound’ conformations for the anionic form of each of
the prototypes, we found a shift in the conformational preference
for both series. Interestingly, both anionic forms now prefer
‘inverted’ conformations corresponding to a ca. 180° shift in the rel-
ative orientation of the quinolone and either of the thiadiazine or
1,4-thiazine rings (Fig. 3; middle row). The lowest energy conforma-
tion for the anionic thiadiazine prototype is ca. 1.5 kcal/mol lower in
energy than the bound conformation, while the lowest energy con-
formation of the 1,4-thiazine is preferred by only ca. 1.1 kcal/mol
(Fig. 3; middle vs bottom row). Both anionic prototypes display the
characteristic shallow U-shaped conformation due to a boat-like thi-
adiazine/1,4-thiazine ring. Since the anionic prototypes for both ser-
ies require only a modest amount of energy to adopt the predicted
NS5b binding mode, these computational results are in excellent
accord with the data from the biochemical assays in Table 1, and fur-
ther support our anionic binding hypothesis.14
A more detailed analysis of the structure–activity relationships
within the 1,4-thiazine series is outlined in Table 2. Based on our
own determination of structure–activity limitations within the
original thiadiazine series, we chose to focus only on small substi-
tutions at C-6, combined with a limited set of lipophilic N-substit-
uents. Compounds 7b–o were prepared as outlined in Scheme 5,
using the appropriately substituted isotoic anhydrides (Scheme
4). Small C-6 substituents were tolerated when combined with
either the cyclopropylethyl (7b–e) or the 4-fluorobenzyl (7j–l)
lipophilic sidechain. The 1H-naphthyridone analogs (7f–h) also
showed similar NS5b potencies, but were significantly less active
compared with a similar thiadiazine series15 and 7g showed a fur-
F
Cl
Me
MeO
H
7g
7h
7i
N
F
8.6
nd
N
H
CH
H
46
F
7j
CH
CH
CH
CH
CH
CH
H
F
0.72
0.7
1.5
1.2
0.4
37
0.79
0.7
1.5
1.5
0.8
nd
F
F
7k
7l
Me
F
F
F
7m
7n
7o
F
F
F
ther 10-fold loss in the replicon assay (8.6 lM). In general, the
CN
NS5b and replicon potencies across this series show only minor
variations. Two key exceptions showing notable losses in potency
are the 2-fluorobenzyl (7i) and 3-cyanobenzyl (7o) analogs. Com-
F

R. T. Hendricks et al. / Bioorg. Med. Chem. Lett. 19 (2009) 3637–3641
3641
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; (c) During the course of our investigation, a related hydrogen bond
analysis was published describing compounds 4 and 5. See Ref. 4b Herein, we
describe an alternate synthesis of compound 4.; (d) For an alternate
preparation of 27, see Ref. 4b.
pared with 7o, the smaller, less-polar 4-fluoro-3-methylbenzyl
analog 7n is two orders of magnitude more potent (0.4 M vs
37 M). Somewhat surprisingly, replicon potency generally
tracked closely with NS5b potency. Cellular permeabilities based
on a 7-day caco cell assay are quite reasonable (7b, 7d, 7j;
Papp = 6.4, 9.4, 4.4 ꢁ 10ꢀ6 cm/s, respectively), however, a 21-day
caco cell assay suggests significant efflux occurs across the series
(7b, 7k, 7n; Papp AB/BA = 0.6/7.1, 0.6/4.9, 0.5/5.3 ꢁ 10ꢀ6 cm/s,
respectively).
A novel class of hydroxy quinolone thiadiazine inhibitors of
HCV NS5b polymerase was recently described.4a,4b A key feature
of this class of inhibitors is that they can form two internal hydro-
gen bonds which help to stabilize an NS5b-bound-like conforma-
tion, thus contributing their inhibitory potency. We have further
tested this hypothesis via synthesis of a series of closely related
molecules that contain varying degrees of hydrogen bonding func-
tionality. Our results suggest an alternate key feature of these
hydroxy quinolone based inhibitors is the acidic nature of this
functionality, leading to their interaction with the enzyme as
anions. From our effort, a novel series of hydroxy quinolone
1,4-thiazine derived inhibitors has emerged with sub-micromolar
potencies against the HCV NS5b polymerase enzyme.
l
l
5. Hendricks, R. T.; Spencer, S. R.; Blake, J. F.; Fell, J. B.; Fischer, J. P.; Stengel, P. J.;
Leveque, V. J. P.; Le Pogam, S.; Rajyaguru, S.; Najera, I.; Josey, J. A.; Swallow, S.
Bioog. Med. Chem. Lett. 2009, 19, 410. doi: 10.1016/j.bmcl.2008.11.060.
6. Milletti, F.; Storchi, L.; Sforna, G.; Cruciani, G. J. Chem. Inf. Model. 2007, 47,
2172–2181. doi: 10.1021/ci700018y. Measured versus calculated pKa for
compounds 2: 5.7 versus 4.4; 3: 7.3 versus 7.4; 7b: 3.9 versus 4.6; 7j: 4.0
versus 4.4.
7. Hardtmann, G. E.; Koletar, G.; Pfister, O. R. J. Heterocycl. Chem. 1975, 12, 565.
8. Mylari, B. L.; Oates, P. J.; Beebe, D. A.; Brackett, N. S.; Coutcher, J. B.; Dina, M. S.;
Zembrowski, W. J. J. Med. Chem. 2001, 44, 2595.
9. Winn, M.; De, B.; Zydowsky, T. M.; Altenbach, R. J.; Basha, F. Z.; Boyd, S. A.;
Brune, M. E.; Buckner, S. A.; Crowell, D.; Drizin, I.; Hancock, A. A.; Jae, H.-S.;
Kester, J. A.; Lee, J. Y.; Mantei, R. A.; Marsh, K. C.; Novosad, E. I.; Oheim, K. W.;
Rosenberg, S. H.; Shiosaki, K.; Sorensen, B. K.; Spina, K.; Sullivan, G. M.; Tasker,
A. S.; von Geldern, T. W.; Warner, R. B.; Obgenorth, T. J.; Kerkman, D. J.;
DeBernardis, J. F. J. Med. Chem. 1993, 36, 2676.
10. 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; Ma, H.; Leveque, V.; De Witte,
A.; Li, W.; Hendricks, T.; Clausen, S. M.; Cammack, N.; Klumpp, K. Virology 2005,
332, 8. Likely due to differences in our assay conditions (100 nM NS5b; cIRES
RNA template) compared with those reported by GSK (50 nM NS5b; poly-C
RNA template), our NS5b IC50’s for 1 and 2 are ca. 10–30-fold higher, however,
their relative potencies are similar and our replicon EC50 values are within ca.
2–6-fold of previously published values.
Acknowledgments
The authors wish to thank Drs. Eric Sjogren, Hans Maag, and
Keith Walker for their support and helpful discussions during the
course of this work.
11. PDB code: 2fvc.
12. GAMESS Version 27 June 2005 (R2): 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 stationary points were verified as true minima via computation of the
corresponding analytical frequencies.
References and notes
13. For comparison, the fully co-planar benzothiadiazine structure was found to be
ca. 0.80 kcal/mol higher in energy than the boat form, and shown to be a
transition state via computation of the analytical frequencies. A search of the
1. (a) Alter, M. J.; Moran, D. K.; Nainan, O. V.; McQuillan, G. M.; Gao, F.; Moyer, L.
A.; Kaslow, R. A.; Margolis, H. S. N. Eng. J. Med. 1999, 341, 556; (b) Tan, S.-L.;
Pause, A.; Shi, Y.; Sonenberg, N. Nat. Rev. Drug Disc. 2002, 1, 867; (c) Brass, V.;
Blum, H. E.; Moradpour, D. Exp. Opin. Ther. Targets 2004, 8, 295; (d) Beaulieu, P.
L.; Tsantrizos, Y. S. Curr. Opin. Invest. Drugs 2004, 5, 838.
2. World Health Organization, Hepatitis C Weekley Epidemiological Record, 1997,
72, 65.
3. Manns, M. P.; McHutchison, J. G.; Gordon, S. C.; Rustgi, V. K.; Shiffman, M.;
Reindollar, R.; Goodman, Z. D.; Koury, K.; Ling, M.; Albrecht, T. K. Lancet 2001,
358, 958.
4. (a) Dhanak, D.; Duffy, K. J.; Johnston, V. K.; Lin-Goerke, J.; Darcy, M.; Shaw, A.
N.; Gu, B.; Silverman, C.; Gates, A.; Nonnemacher, M. R.; Earnshaw, D. L.;
Casper, D. J.; Kaura, A.; Baker, A.; Greenwood, C.; Gutshall, L. L.; Maley, D.;
DelVecchio, A.; Nacarron, R.; Hoffmann, G. A.; Alnoah, D.; Cheng, H. Y.; Chan, G.;
Khandekar, S.; Keenan, R. M.; Sarisky, R. T. J. Biol. Chem. 2002, 277, 38322; (b)
CSD for thiadiazine containing molecules yields
a number of relevant
examples. In the majority of small molecule X-ray structures the thiadiazine
fragment is largely planar, though subtle puckering is noted in a few examples.
In light of the ab initio results, the planar conformation of the thiadiazine
group is likely due to crystal packing effects. CSD, Version 5.13, April 1997
release. Entries CACTAZ0001, DIAZOX0001, JIFYUS0001, JIFZAZ0001,
SEFMEV0001, SEFMIZ0001.
14. These findings do not exclude the possibility that the thiadiazines (i.e., 1)
might still bind in the enolic form. The available X-ray structure of 1 does not
allow for differentiation between the enolic and anionic forms.
15. 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. Bioog. Med. Chem. Lett. 2006, 16, 3833.