Novel HCV NS5B polymerase inhibitors derived from 4-(1′,1′-dioxo-1′,4′-dihydro-1′λ6-benzo[1′,2′,4′]thiadiazin-3′-yl)-5-hydroxy-2H-pyridazin-3-ones. Part 2: Variation of the 2- and 6-pyridazinone substituents

Article abstract of DOI:10.1016/j.bmcl.2008.01.005

5-Hydroxy-3(2H)-pyridazinone derivatives were investigated as inhibitors of genotype 1 HCV NS5B polymerase. The structure-activity relationship (SAR) associated with variation of the pyridazinone 2- and 6-substituents is discussed. The synthesis and metab

Full text of DOI:10.1016/j.bmcl.2008.01.005

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1419–1424
Novel HCV NS5B polymerase inhibitors derived from
4-(10,10-dioxo-10,40-dihydro-10k6-benzo[10,20,40]thiadiazin-30-yl)-5-
hydroxy-2H-pyridazin-3-ones. Part 2: Variation of the 2- and
6-pyridazinone substituents
Yuefen Zhou,* Lian-Sheng Li, Peter S. Dragovich, Douglas E. Murphy, Chinh V. Tran,
Frank Ruebsam, Stephen E. Webber, Amit M. Shah, Mei Tsan, April Averill,à
Richard E. Showalter, Rupal Patel, Qing Han,§ Qiang Zhao, Thomas Hermann,–
Charles R. Kissinger, Laurie LeBrun and Maria V. Sergeeva
Anadys Pharmaceuticals, Inc., 3115 Merryfield Row, San Diego, CA 92121, USA
Received 9 November 2007; revised 23 December 2007; accepted 2 January 2008
Available online 8 January 2008
Abstract—5-Hydroxy-3(2H)-pyridazinone derivatives were investigated as inhibitors of genotype 1 HCV NS5B polymerase. The
structure–activity relationship (SAR) associated with variation of the pyridazinone 2- and 6-substituents is discussed. The synthesis
and metabolic stability of this new class of compounds are also described.
Ó 2008 Elsevier Ltd. All rights reserved.
Hepatitis C virus (HCV) infects approximately 170 mil-
lion individuals worldwide and is a leading cause of
chronic liver disease. About 70% of infected people will
develop chronic histological changes in the liver (chronic
hepatitis) with a high risk of advancing to cirrhosis or
hepatocellular carcinoma.1 The current standard of care
for HCV is a combination therapy of pegylated inter-
feron and ribavirin. Unfortunately, this treatment pro-
vides benefit for less than 50% of genotype 1 infected
patients and is associated with adverse events.2 There-
fore there is an urgent need for the development of more
effective HCV therapies especially for genotype 1
patients.
As part of our efforts to develop small molecule non-
nucleoside inhibitors of HCV NS5B, a RNA-dependent
RNA polymerase (RdRp),3 we discovered a series of
5-hydroxy-3(2H)-pyridazinone derivatives (1, Fig. 1)
using a structure-based design approach. In a separate
communication,4 we discussed the structure–activity
relationships (SAR) associated with variation of
the R3 substituents in this series and identified the
–OCH2CONH2 moiety as a suitable fragment for
obtaining potent NS5B inhibition properties. Here we
Keywords: Pyridazinones; 5-Hydroxy-3(2H)-pyridazinone derivatives;
Hepatitis C virus (HCV); Structure-based design; RNA-dependent
RNA polymerase (RdRp); Small molecule; Non-nucleoside NS5B
inhibitors.
*
Corresponding author. Tel.: +1 858 530 3693; fax: +1 858 530
3644; e-mail: yzhou@anadyspharma.com
O
O
O
O
O
Present address: Illumina, 9885 Towne Centre Drive, San Diego, CA
92121-1975, USA.
S
R3
7'
S
O
OH
N
OH
N
NH2
à
§
R1
R1
Present address: University of Vermont, Microbiology and Molec-
ular Genetics, 95 Carrigan Drive, Burlington, VT 05401, USA.
Present address: Structure Based Design, Inc., 5560 Lusk Blvd. B104,
San Diego, CA 92121, USA.
6
N
H
N
H
4
2
N
N
N
O
N
O
R2
1
R2
2
–
Present address: Department of Chemistry and Biochemistry,
University of California, San Diego, 9500 Gilman Drive, La Jolla,
CA 92093-0358, USA.
Figure 1. 5-Hydroxy-3(2H)-pyridazinone derivatives as HCV NS5B
polymerase inhibitors.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.01.005

1420
Y. Zhou et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1419–1424
describe biological properties associated with the sys-
tematic alteration of the 2- and 6-pyridazinone substitu-
ents in a series of molecules (2) which contain this
optimal R3 substituent.
As previously described in a separate communication,4
compound 2a (Table 1) exhibited good inhibitory po-
tency against the genotype 1b enzyme. However, it
was 5-fold less potent in the genotype 1a enzymatic as-
say. Compounds bearing R2 groups that were compara-
ble in linear length to the isoamyl moiety present in 2a
exhibited slightly improved activity in the biochemical
assay (2b and 2c). However, they were significantly less
potent in the genotype 1b replicon assay compared to
compound 2a. Adding an F-atom to the terminus of
the isoamyl group (2d) or introducing a 2,2-dimethylbu-
tyl group (2e) led to a significant loss in activity in the
biochemical assay when compared with 2a. Interest-
ingly, compound 2e exhibited equipotent antiviral prop-
erties in the replicon assay as compared to 2a. In
addition, introduction of R2 alkyl groups that were
longer or shorter than the isoamyl moiety led to com-
pounds with NS5B inhibition activities comparable to
2a against the genotype 1b enzyme (2f and 2g). How-
ever, these molecules were less potent than 2a in both
1b replicon and 1a enzyme assays. Furthermore, com-
pounds bearing a methylene spacer between the pyridaz-
inone core and a cyclic moiety displayed decreasing
NS5B inhibition activities with increasing ring size
(2h–2j). When compared to 2j, compound 2k having a
benzyl R2 substituent displayed weaker antiviral po-
tency in the 1b replicon assay although the inhibitory
activity in the enzyme assay was improved. The different
EC50/IC50 ratios observed likely resulted from the differ-
ences in cell permeability and protein binding.10 In gen-
eral, this series of 5-hydroxy-3(2H)-pyridazinone
derivatives under study was 3- to 14-fold less potent
against genotype 1a NS5B relative to the genotype 1b
enzyme. We believe that this potency shift was likely
due to the amino acid residue difference of Y415F
(phenylalanine in genotype 1a vs tyrosine in genotype
1b) located near the R2 sub-pocket at the palm binding
site.11
As depicted in Scheme 1, three methods (A, B, and C)
were developed to synthesize pyridazinone benzothi-
adiazinedioxide derivatives (5) bearing a methoxy group
at the 70-position. Method A utilized the key 3(2H)-
pyridazinone ester intermediates (3), which were
prepared from a-keto-esters via an efficient three-step
synthesis.5 Condensation of 3 with 2-amino-5-meth-
oxy-benzenesulfonamide (4)6 in refluxing pyridine in
the presence of DBU afforded the desired products 5.
In method B, hydrazones 6 were prepared by condensa-
tion of the corresponding a-keto-esters with alkyl hydra-
zines (or their oxalic acid salts).5 Compounds 6
underwent DCC-mediated coupling with (7-methoxy-
1,1-dioxo-1,4-dihydro-1k6-benzo[1,2,4]-thiadiazin-3-yl)ace-
tic acid (7)6 and the resulting hydrazide intermediates
were subsequently cyclized to 5 by treatment with a base
such as triethylamine or NaOEt. Alternatively, hydra-
zones 6 could be converted to 5 directly in one step by
heating with (7-methoxy-1,1-dioxo-1,4-dihydro-1k6-
benzo[1,2,4]thiadiazin-3-yl)acetic acid ethyl ester (8)6 at
160 °C (method C). In general, methods A and B gave
higher yields than method C. Compounds 5 were then
transformed to the target molecules 2 by treatment with
BBr3 to form the corresponding 70-phenol intermediates
and subsequent O-alkylation with 2-bromoacetamide in
the presence of K2CO3 at 80 °C.
A variety of pyridazinone compounds 2 were synthe-
sized using the methods described above.7 The inhibi-
tory activities of these compounds in genotype 1a and
1b NS5B enzyme assays as well as in a 1b replicon assay
were evaluated. The results of these biological assess-
ments are summarized in Tables 1 and 2.
Method A:
OH
O
O
O
O
R1
S
OMe
OEt
H2N
H2N
+
N
N
R2
Pyridine, DBU
3
4
120
ºC
Method B:
O
O
O
O
O
S
OMe
1. BBr3, CH2Cl2
S
OMe
1. DCC
CH2Cl2 / DMF
OH
N
R1
N
O
OEt
R1
+
N
H
2
N
H
N
2. BrCH2CONH2
K2CO3 / DMF
NH
N
N
O
2. base
R2
HO
5
R2
80 ºC
7
6
6
O
O
160 ºC, neat
S
OMe
Method C:
N
N
H
EtO
O
8
Scheme 1. General methods for the synthesis of pyridazinone derivatives (2).7

Y. Zhou et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1419–1424
Table 1. R2 optimization with 2-thiophene as R1 substituent
1421
O
O
O
S
O
OH
N
NH2
S
N
H
N
N
O
2a - 2k
R2
b
HLM t1/2 (min)
Compounda
R2
IC50 (1b)b (lM)
IC50 (1a)c (lM)
EC50 (1b)b (lM)
CC50 (GAPDH)b (lM)
2a
0.07
0.34
0.4
78
23
2b
2c
2d
2e
2fe
2g
2h
2i
0.035
0.046
0.38
0.12
0.24
NDd
0.99
1.0
2.1
>33
>33
>33
>33
>33
>33
>33
>17
>33
>33
12
1.9
21
F
2.3
39
0.18
0.44
1.09
0.81
1.6
>60
25
0.07
0.088
0.055
0.28
0.84
0.54
ND
ND
0.54
27
22
2.1
26
2j
0.44
1.1
>60
23
2k
0.15
5.15
a See Ref. 8.
b See Ref. 4 for assay method and experimental error.
c See Ref. 9 for assay method and experimental error.
d ND, not determined.
e Racemic.
The above efforts led to the identification of the R2
substituents present in compounds 2a–2c and 2h as
optimal moieties for obtaining potent NS5B inhibitory
activity. We then explored combinations of these R2
fragments with various R1 substituents in an effort to
further improve potency. The impact of R1 substitution
on the inhibitory activity against genotype 1a and 1b
enzymes as well as the 1b replicon is summarized in
Table 2.
of the 2-thiophene ring resulted in a significant loss of
activities in both enzyme (30-fold) and replicon assays
(4-fold) (compare 2m with 2b). This trend was ob-
served with several other inhibitors (2n and 2o) bear-
ing different R2 substituents (compare with 2c and
2h, respectively). The co-crystal structure of 2m bound
to NS5B12 revealed that this methyl group was unable
to access the small sub-pocket in the R1 region as
originally intended. Instead, the methyl fragment
rested on the NS5B surface and generated unfavorable
interactions with nearby polar protein residues. These
interactions were the likely cause of the observed
decrease in the inhibition activity of 2m. A similar
reduction in potency was observed by introducing a
As shown in Table 2, the 3-thiophene analog (2l) was
9-fold more potent in the 1b replicon assay than the
corresponding 2-thiophene-containing compound (2b).
However, introducing a methyl group at the 5-position

1422
Y. Zhou et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1419–1424
Table 2. R1 Variation with optimal R2 substituents
O
O
O
S
O
OH
N
NH2
R1
N
H
N
N
O
R2
2l - 2y
Compounda R1
R2
IC50 (1b)b (lM) IC50 (1a)c (lM) EC50 (1b)b (lM) CC50 (GAPDH)b (lM) HLM t1/2 (min)
b
2b
0.035
0.088
1.1
0.12
0.14
NDd
ND
ND
ND
0.15
0.14
ND
ND
ND
ND
ND
ND
5.1
2.1
>33
>33
>33
>33
>33
>33
>33
>33
>33
>33
>10
>100
>33
>33
>33
12
S
S
2l
0.23
7.7
16
2m
2n
2o
42
S
S
S
1.4
23
45
1.2
17
25
58
2p
0.43
0.039
0.077
2.4
ND
60
S
N
2q
1.1
S
N
2r
1.8
5
>60
>60
>60
8
S
N
N
2s
S
2t
1.1
8.7
6.6
S
2u
2v
0.79
2.2
11
ND
ND
10
2w
2x
0.4
4.5
3
0.43
0.74
2y
9.7
ND
a See Ref. 8.
b See Ref. 4 for assay method and error.
c See Ref. 9 for assay method and error.
d ND, not determined.
methyl group at the 3-position of the thiophene ring
(compare 2p with 2a). The co-crystal structure of 2p
complexed with the NS5B protein12 showed that the
addition of the methyl group caused a change in ori-
entation of the thiophene ring relative to the unsubsti-
tuted ring. This change placed the methyl group in

Y. Zhou et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1419–1424
1423
close proximity with polar protein atoms and this may
explain the reduction in potency observed for 2p.
number of direct or water-mediated H-bonds were
formed between the sulfonamide O-atoms and residues
Ser556, Ser288, and Gly449 as well as between the acet-
amide R3 moiety and residues Asn291 and Asp318.
These favorable structure features, which were almost
identical to those observed in the X-ray co-crystal struc-
ture of 2a-NS5B,4 may explain the high potency of com-
pound 2q. A schematic diagram shown in Figure 3
summarizes the observed interactions in the co-crystal
structure of compound 2q bound to the NS5B protein.
In addition, replacement of the 2-thiophene ring with a
5-thiazole moiety led to very potent compounds (2q and
2r) possessing activities comparable to 2a and 2b in both
enzyme (1b and 1a) and replicon assays. This result was
consistent with the analysis of the co-crystal structure of
2q bound to the NS5B protein (Fig. 2). The pyridazi-
none and benzothiadiazine rings of 2q adopted a nearly
co-planar geometry which was highly favorable for
binding to the NS5B palm site. The 5-thiazole ring, ro-
tated approximately 16° out of plane relative to the
pyridazinone ring, was able to effectively access the
sub-pocket in the R1 region. This rotation is similar to
the rotation of the 2-thiophene moiety in 2a-NS5B X-
ray co-crystal structure.12 These rotations were smaller
than the rotation of a phenyl R1 group (R2 = isoamyl,
R3 = OMe) as we discussed previously.4 Due to its larger
size, a phenyl ring needs to rotate more to avoid a close
interaction with Gly410 and Asn411. This could explain
some of the differences between the IC50 values of 2q or
2a and the phenyl substituted analogs. Additionally, a
Surprisingly, the replacement of the 2-thiophene ring
with a 2-thiazole moiety (2s and 2t) caused a significant
loss in inhibitory potency in both enzyme and replicon
assays when compared with 2a and 2h, respectively.
The reasons for this loss in biological activity are not
yet well understood. Furthermore, introduction of alkyl
R1 substituents into the inhibitor design (2u–2y) did not
afford any improvements in NS5B inhibitory potency as
compared with compound 2a.
In order to better predict the in vivo metabolic stability
of this new class of pyridazinone derivatives, the effect
of R1 and R2 substituents on stability toward human
liver microsomes (HLM) was evaluated. As shown in
Table 1, the measured HLM half-life for compound
2a was relatively short (23 min). We reasoned that, in
addition to the unsubstituted thiophene ring present
in 2a, the secondary C-atom of the terminal isoamyl
group might also be a potential site for cytochrome
P450 (CYP) catalyzed metabolism. Accordingly, an
F-atom was introduced at the secondary C-atom of
the isoamyl fragment and this resulted in a more stable
compound (2d) (t1/2 = 39 min). However, the addition
of a methyl group at this position did not improve
the stability of the resulting molecule (2c) (t1/2
=
21 min). Interestingly, inclusion of a more branched
R2 group of the same linear length as the isoamyl moi-
ety or a large cyclohexyl R2 fragment resulted in very
stable compounds (2e and 2j, respectively; t1/
2 > 60 min). In contrast, compound 2k, bearing a ben-
zyl R2 moiety, had a short half-life (23 min) possibly
due to CYP-mediated modification of the benzene ring
(compare 2k with 2j). The measured HLM half-lives
for the rest of compounds (2b, 2f–2i) in Table 1 were
relatively short (<30 min).
Figure 2. Co-crystal X-ray structure of compound 2q bound to the
13
˚
NS5B protein (2.3 A resolution).
HN
Ser 556
O
HN
O
HN
Gly 449
NH
Alkyl R1 substituents shown in Table 2 led to the least
stable compounds (2u, 2x) with very short half-lives
(610 min). Similar to compound 2b bearing a 2-thio-
phene moiety, the 3-thiophene group also led to a
compound (2l) with a short half-life (16 min). As the
5-position of the 2-thiophene ring is a potential site for
oxidation catalyzed by CYPs,15 we introduced a methyl
group at this position to block such metabolism. This
change resulted in a significant improvement in the sta-
bility of the resulting compounds 2m, 2n, and 2o when
compared with 2b, 2c, and 2h, respectively. Both 2-thia-
zole and 5-thiazole moieties led to very stable com-
pounds (2q–2t) with t1/2 P 60 min. The above results
indicated that both R1 and R2 substituents significantly
affected the metabolic stability of the resulting
molecules.
Ser 288
OH
OH
H2O
O
H2O
OH
H2N
O
O
O
Asn 291
S
O
N
S
N
NH2
Gly 410
N
H
Met 414
Gln 446
O
N
N
O
HO
Asp 318
Tyr 448
Met 414
Pro 197
Arg 200
Leu 384
Figure 3. Schematic diagram of compound 2q bound in the NS5B
palm site. Hydrogen bonds are represented as dashed lines, and the
residues which make up the enzyme binding subsites are depicted.14

1424
Y. Zhou et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1419–1424
In summary, we synthesized a new class of 4-(10,10-di-
oxo-10,40-dihydro-10k6-benzo[10,20,40]thiadiazin-30-yl)-5-
hydroxy-2H-pyridazin-3-one derivatives (2) as potent
inhibitors of genotype 1 HCV RNA-dependent RNA
polymerase (NS5B). The optimization of the 2- and
6-substituents of the pyridazinone derivatives resulted
in significantly improved potencies in both biochemical
(1b and 1a) and replicon (1b) assays. In addition, our
SAR studies revealed that both R1 and R2 substituents
significantly affected the HLM stability of the molecules.
Further optimization of this new class of pyridazinone
compounds will be discussed in future communications.
8. All structures of compounds 2 under study were consistent
with 1H NMR and LC–MS analysis (P95% HPLC
purity). They are arbitrarily drawn in one of the possible
tautomer forms.
9. NS5B polymerase inhibition assay (1a IC50, lM): assays
were performed in a 96-well streptavidin-coated Flash-
Plate using 50 nM enzyme (NS5B H77 1a), 0.5 lCi of [a-
33P]GTP, 0.6 mM GTP, and 250 nM 50biotinylated oligo
(rG13)/poly rC in 20 mM Tris–HCl, pH 7.5, 5 mM MgCl2,
5 mM dithiothreitol, 0.1 g/L bovine serum albumin, and
100 U/mL RNAse inhibitor. The reaction was stopped by
aspiration after 75 min at 28 °C and the plate was washed
several times with 150 mM NaCl aqueous solution. After
washing and drying the plate, incorporated radioactivity
was counted using a Microbeta scintillation counter. IC50
values were calculated relative to the uninhibited control
and inhibition data were fitted to a 4-parameter IC50
equation. The estimated average standard deviation for 1a
IC50 data is 20% from the mean value.
10. The amounts of protein in replicon assay were much
higher than that in the NS5B enzyme assay. In the 1b
NS5B enzyme assay, there was 0.1 g/L bovine serum
albumin, while in the replicon assay there was about 4 g/L
of fetal bovine serum (FBS).
11. Pauwels, F.; Mostmans, W.; Quirynen, L. M. M.; Helm,
L. V. D.; Boutton, C. W.; Rueff, A.-S.; Cleiren, E.;
Raboisson, P.; Surleraux, D.; Nyanguile, O.; Simmen, K.
A. J. Virol. 2007, 81, 6909.
Acknowledgments
The authors thank Drs. Devron Averett and Steve Wor-
land for their support and helpful discussions during the
course of this work.
References and notes
1. Wong, T.; Lee, S. S. CMAJ 2006, 174, 649.
2. Hoofnagle, J. H.; Seeff, L. B. N. Engl. J. Med. 2007, 355,
2444.
3. Koch, U.; Narjes, F. Curr. Top. Med. Chem. 2007, 7, 1302.
4. Zhou, Y.; Webber, S. E.; Murphy, D. E.; Li, L.-S.;
Dragovich, P. S.; Tran, C. V.; Sun, Z.; Ruebsam, F.; Shah,
A. M.; Tsan, M.; Showalter, R. E.; Patel, R.; Li, B.; Zhao,
Q.; Han, Q.; Hermann, T.; Kissinger, C. R.; LeBrun, L.;
Sergeeva, M. V.; Kirkovsky, L. Bioorg. Med. Chem. Lett.
2008, 18, 1413.
5. (a) Li, L.-S.; Zhou, Y.; Zhao, J.; Dragovich, P. S.;
Stankovic, N.; Bertolini, T. M.; Murphy, D. E.; Sun, Z.;
Tran, C. V.; Ayida, B. K.; Ruebsam, F.; Webber, S. E.
Synthesis 2007, 3301; (b) Murphy, D. E.; Dragovich, P. S.;
Ayida, B. K.; Bertolini, T. M.; Li, L.-S.; Ruebsam, F.;
Stankovic, N. S.; Sun, Z.; Zhou, Y. Tetrahedron Lett.
2008, 49, 811.
12. The co-crystal structures of 2a, 2m, and 2p bound to the
NS5B protein are not shown.
13. Crystals of HCV NS5B polymerase (genotype 1b, strain
BK, D21) were grown by the hanging drop method at
room temperature using a well buffer of 20% PEG 4K,
50 mM ammonium sulfate, 100 mM sodium acetate, pH
4.7, with 5 mM DTT. The crystals formed in space group
˚
P212121 with approximate cell dimensions, a = 85 A,
˚
˚
b = 106 A, c = 127 A and two protein molecules in the
asymmetric unit. Protein-inhibitor complexes were pre-
pared by soaking these NS5B crystals for 24 h in solutions
containing 10% DMSO, 20% glycerol, 20% PEG 4K,
0.1 M Hepes, 10 mM MgCl2 at pH 7.6, and 2q at a
concentration of 3 mM. Diffraction data were collected to
˚
a resolution of 2.3 A on beamline 14IDB at the Advanced
6. 2-Amino-5-methoxy-benzenesulfonamide 4, benzothiadi-
azine acid 7 and its ethyl ester 8 were prepared using
modified literature methods as described in (a) Fitch, M.
D.; Evans, K. A.; Chai, D.; Duffy, K. J. Org. Lett. 2005, 7,
5521; (b) Kovalenko, S. N.; Chernykh, V. P.; Shkarlat, A.
E.; Ukrainets, I. V.; Gridasov, V. I.; Rudnev, S. A. Chem.
Heterocycl. Comp. (Engl. Transl.) 1999, 34, 791.
7. Detailed procedures for the synthesis of compounds 1 and
2 as well as their analytical data are described in Zhou, Y.;
Li, L.-S.; Webber, S. E. WO06066079A2, 2006.
Photon Source at Argonne National Laboratory. This
crystal structure has been deposited in the Protein
Databank (www.rcsb.org) with entry code 3BSA. Full
details of structure determination are given in the PDB
entry.
14. The O-atom at the 5-position of the pyridazin-3-one in
Figure 3 was arbitrarily drawn in the protonated form.
The precise extent of the protonation was not determined.
15. O’Donnell, J. P.; Dalvie, D. K.; Kalgutkar, A. S.; Obach,
R. S. Drug Metab. Dispos. 2003, 31, 1369.