Discovery, Synthesis, and invitro Evaluation of West Nile Virus Protease Inhibitors Based on the 9,10-Dihydro-3H,4aH-1,3,9,10a-tetraazaphenanthren-4-one Scaffold

Article abstract of DOI:10.1002/cmdc.201200136

West Nile virus (WNV), a member of the Flaviviridae family, is a mosquito-borne pathogen that causes a great number of human infections each year. Neither vaccines nor antiviral therapies are currently available for human use. In this study, a WNV NS2B-NS3 protease inhibitor with a 9,10-dihydro-3H,4aH-1,3,9,10a-tetraazaphenanthren-4-one scaffold was identified by screening a small library of non-peptidic compounds. This initial hit was optimized by solution-phase synthesis and screening of a focused library of compounds bearing this scaffold. This led to the identification of a novel, uncompetitive inhibitor (1a40, IC₅₀=5.41±0.45μM) of WNV NS2B-NS3 protease. Molecular docking of this chiral compound onto the WNV protease indicates that the Senantiomer of 1a40 appears to interfere with the productive interactions between the NS2B cofactor and the NS3 protease domain; (S)-1a40 is a preferred isomer for inhibition of WNV NS3 protease.

Full text of DOI:10.1002/cmdc.201200136

MED
DOI: 10.1002/cmdc.201200136
Discovery, Synthesis, and in vitro Evaluation of West Nile
Virus Protease Inhibitors Based on the 9,10-Dihydro-
3H,4aH-1,3,9,10a-tetraazaphenanthren-4-one Scaffold
Sanjay Samanta,[a] Taian Cui,[b] and Yulin Lam*[a]
West Nile virus (WNV), a member of the Flaviviridae family, is
a mosquito-borne pathogen that causes a great number of
human infections each year. Neither vaccines nor antiviral
therapies are currently available for human use. In this study,
a WNV NS2B–NS3 protease inhibitor with a 9,10-dihydro-
3H,4aH-1,3,9,10a-tetraazaphenanthren-4-one scaffold was iden-
tified by screening a small library of non-peptidic compounds.
This initial hit was optimized by solution-phase synthesis and
screening of a focused library of compounds bearing this scaf-
fold. This led to the identification of a novel, uncompetitive in-
hibitor (1a40, IC50 =5.41ꢀ0.45 mm) of WNV NS2B–NS3 pro-
tease. Molecular docking of this chiral compound onto the
WNV protease indicates that the S enantiomer of 1a40 appears
to interfere with the productive interactions between the NS2B
cofactor and the NS3 protease domain; (S)-1a40 is a preferred
isomer for inhibition of WNV NS3 protease.
Introduction
The West Nile virus (WNV), a neurotropic flavivirus, is a mosqui-
to-borne reemerging pathogen causing outbreaks worldwide.[1]
While a larger proportion of infected people suffer from milder
symptoms such as West Nile fever, there has been an observed
increase in cases of severe neuroinvasive diseases such as
meningitis, flaccid paralysis, and encephalitis over the past
decade.[2] Furthermore, close follow-ups with WNV-infected
people have shown that patients who manifest symptomatic
infections often suffer lifetime consequences such as long-
term cognitive and neurological impairment.[3] Despite the
growing public health concern associated with WNV infection,
to date there are no successful antiviral therapies or vaccines
for human use against WNV.
coupled with unique substrate preference make it an attractive
target for selective viral inhibition.
Efforts toward the development of inhibitors against the
WNV NS2B–NS3 protease have thus far focused mainly on nat-
ural or modified peptidic compounds,[10] and only a few non-
peptidic compounds have been reported.[11] To date, various
peptides containing highly charged moieties have been shown
to have good inhibitory activity against the WNV NS2B–NS3
protease in vitro,[10a,b] underscoring the role of electrostatic in-
teractions. However, a major disadvantage of peptidic com-
pounds is their short physiological half-life, as peptides are
easily cleaved and inactivated in the body. Furthermore, high
drug clearance rates in vivo lead to diminished therapeutic ef-
fects. For non-peptidic inhibitors, scaffolds reported earlier in-
clude compounds with isothiourea (IC50 =183 mm)[11a] or guani-
dino[11b] moieties (Ki =13ꢀ1 mm), sulfonylpyrazolyl (IC50 =
0.105 mm),[11c] 8-hydroxyquinoline (Ki =3.2ꢀ0.3 mm),[11d] and iso-
quinoline (IC50 =30 mm)[11e] derivatives. The most potent non-
peptidic inhibitors known are several sulfonylpyrazolyl deriva-
tives, which were identified by high-throughput screening of
more than 65000 compounds at the National Institutes of
Health.[11c] However, these compounds were unstable under
physiological conditions and degraded within one hour in
aqueous buffer. Attempts to improve their stability without
compromising biological activity proved unsatisfactory.[12]
The West Nile virion contains a 10.8-kb single-stranded posi-
tive-sense RNA genome 11000–12000 nucleotides long.[4] The
RNA genome encodes three structural proteins (C, capsid; M,
membrane; E, envelope) and seven non-structural proteins
(NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5),[5] which form
the machinery involved in viral synthesis and replication.[4,6]
Site-directed mutagenesis studies have shown that mutation
of essential catalytic residues at NS3 protease cleavage sites
lead to a halt in viral infectivity, indicating its importance in
the WNV life cycle.[7] In addition, the NS2B protein has been
found to be an essential cofactor that increases NS3 protease
activity significantly.[8] The WNV NS2B–NS3 protease cleaves
peptide substrates at their C termini; substrate recognition is
brought about by a pair of basic amino acids (R and K) at the
P1 and P2 positions. This mode of substrate recognition is con-
served among other flaviviral proteases, but not in host cellular
enzymes,[9] indicating the possibility of designing inhibitors
that are selective for flaviviral proteases without adversely af-
fecting normal cellular functions. Thus, the biological impor-
tance of the NS2B–NS3 complex in activation of viral proteins
[a] S. Samanta, Prof. Y. Lam
Department of Chemistry
National University of Singapore, 3 Science Drive 3, 117543 (Singapore)
E-mail: chmlamyl@nus.edu.sg
[b] Dr. T. Cui
School of Applied Science
Tampines Polytechnic, 21 Tampines Avenue 1, 529757 (Singapore)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201200136.
ChemMedChem 0000, 00, 1 – 8
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

Y. Lam et al.
MED
Hence, there is a need to examine other compounds for their
activities and suitability as WNV protease inhibitors. Herein we
describe the identification of a new scaffold with promising in-
hibitory activity against WNV NS2B–NS3 protease, the library
synthesis and biological evaluation of 72 analogues of the lead
compound, and the structure–activity relationship (SAR) data
obtained.
Results and Discussion
Identification of the initial hit and synthesis of its analogues
In the preliminary WNV NS2B–NS3 protease inhibition assay,
a collection of 110 compounds with various scaffolds were
screened at 100 mm to filter for potential inhibitors. This initial
screen provided a rapid identification of compounds that
show the potential to inhibit WNV NS2B–NS3 protease. One of
the hit compounds identified, compound 1a1 (Figure 1), was
Scheme 1. Synthesis of compounds 1a and 2a. Reagents and conditions:
a) Cl3CCH(OH)2, Na2SO4, HCl, H2O, NH2OH·HCl, 908C, 2.5 h (80–95% yield);
b) conc H2SO4, 60!808C, 45 min; c) NH2NHCSNH2, 10% aq KOH, 1158C, 1 h,
then !pH 5 (52–70% yield); d) ArCH2Br, EtOH/10% aq KOH mixture, RT, 4 h
(80–90% yield); e) R2CHO, glacial acetic acid, EtOH, reflux, 10–15 min (50–
90% yield); f) DDQ, CH2Cl2, 08C!RT, 4 h (70–85% yield).
bazide in 10% aqueous potassium hydroxide[13a] to yield inter-
mediate 6, which was precipitated from the reaction mixture
at pH 5 as a yellow solid. Intermediate 6 was treated with the
respective benzyl bromide derivatives and then acidified with
glacial acetic acid to pH 5–6 to yield compound 7 as a pure
yellow solid. Reaction of compound 7 with various aldehydes
in the presence of catalytic amounts of glacial acetic acid pro-
vided racemic 1a, which could be further oxidized with 2,3-di-
chloro-5,6-dicyano-1,4-benzoquinone (DDQ) at room tempera-
ture[16] to the corresponding compound 2a.
Figure 1. Structure and IC50 value of initial hit compound against WNV
NS2B–NS3 protease.
of particular interest, because to our knowledge, this is the
first zwitterion-type inhibitor identified, and we reasoned that
such an entity, together with its surrounding amino groups,
could function like highly charged peptidyl-recognition com-
ponents to provide favorable binding interactions. To provide
access to structurally diverse analogues of compound 1a1 for
biological evaluation of their inhibitory activities, a focused li-
brary of 1a1 analogues was synthesized by modifying a pub-
lished procedure.[13]
SAR studies
A representative set of 69 analogues of compound 1a and
three analogues of compound 2a were synthesized and tested
in a WNV assay at 100 mm to identify potential inhibitors. Inhib-
itor candidates were determined by their ability to block
NS2B–NS3 protease activity by at least 50%. The effects of
these compounds were then evaluated at various concentra-
tions, and the IC50 values of each compound were calculated
with GraphPad Prism software (Table 1). The KM value for the
fluorogenic peptide substrate Pyr-RTKR-AMC for WNV NS2B–
NS3 protease was found to be 3.45ꢀ0.41 mm by using the pro-
tocol from the SensoLyte 440 West Nile Virus Protease Assay
Kit.
In the SAR studies, R2 was first changed from 2-(3-chlorophe-
nyl)furan-5-yl to various alkyl, aromatic, or heteroaromatic
groups whilst keeping R1 and Ar unchanged (Table 1, com-
pounds 1a2–1a32). Replacing the 2-(3-chlorophenyl)furan-5-yl
group in 1a1 with a 2-(4-nitrophenyl)furan-5-yl, 2-(4-bromo-
phenyl)furan-5-yl, or 2-(4-chlorophenyl)furan-5-yl moiety gave
comparatively better inhibition (Table 1, compounds 1a3–1a5),
indicating that a 4-substituted phenyl group plays an impor-
tant role in the compound’s inhibitory activity. In addition,
The general strategy for the synthesis of compound 1a is
shown in Scheme 1. A warm, aqueous solution of chloral hy-
drate and sodium sulfate was first reacted with an acidic solu-
tion of the respective aniline 3 followed by aqueous hydroxyl-
amine at 908C to give isonitrosoacetanilide[14] 4, which precipi-
tated upon cooling from the reaction mixture in 80–95% yield.
Polychronopoulos et al.[15] reported the cyclization of 4 with
~68 equiv of concentrated sulfuric acid. However, in our
hands, this large amount of concentrated acid hindered the
precipitation for isatin 5. We explored the reaction by using
a lower amount of concentrated sulfuric acid and found that
at 46 equiv H2SO4, the cyclization reaction proceeded equally
readily; the crude 5, which precipitated easily, could be used in
the next step of the reaction without purification. Compound
5 was subsequently allowed to react at reflux with thiosemicar-
&2
&
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 8
ÝÝ
These are not the final page numbers!

WNV Protease Inhibitors
MED
Table 1. Inhibition of WNV NS2B–NS3 protease by compounds 1a and
2a.
Table 1. (Continued)
Compd R1
1a67
1a68 Me 4-chlorophenyl
1a69 Me 4-(methylthio)phenyl
R2
Ar
IC50 [mm][a]
>100
Compd R1
R2
Ar
IC50 [mm][a]
H
4-(methylthio)phenyl
4-CNC6H4
4-MeOC6H5
4-MeOC6H5
C6H5
C6H5
C6H5
14.56ꢀ1.03
1a2
1a3
1a4
1a5
1a6
1a7
1a8
1a9
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
F
2-(2-chlorophenyl)furan-5-yl C6H5
2-(4-chlorophenyl)furan-5-yl C6H5
2-(4-bromophenyl)furan-5-yl C6H5
72.84ꢀ6.00
13.56ꢀ0.53
29.41ꢀ3.97
10.05ꢀ1.34
NA
11.92ꢀ1.32
2a1
2a2
2a3
Me 4-chlorophenyl
Me 4-(methylthio)phenyl
Et 4-(methylthio)phenyl
NA
NA
NA
2-(4-nitrophenyl)furan-5-yl
ethyl
thiophene-2-yl
5-methylfuran-2-yl
phenyl
quinoline-3-yl
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
45.06ꢀ2.85
49.17ꢀ5.51
27.33ꢀ5.81
11.02ꢀ0.78
[a] Data represent concentrations required to inhibit WNV NS2B–NS3 pro-
tease activity by 50% and are the mean ꢀSE of triplicate experiments;
compounds used are ꢁ98% pure. NA: no activity.
1a10
1a11
1a12
1a13
1a14
1a15
1a16
1a17
1a18
1a19
1a20
1a21
1a22
1a23
1a24
1a25
1a26
1a27
1a28
1a29
1a30
1a31
1a32
1a33
1a34
1a35
1a36
1a37
1a38
1a39
pyridine-3-yl
pyridine-2-yl
>100
>100
20.46ꢀ1.47
29.99ꢀ2.52
>100
2-methylphenyl
2-chlorophenyl
2-(methylthio)phenyl
3-phenoxyphenyl
3-carboxyphenyl
3-chlorophenyl
3-nitrophenyl
4-fluorophenyl
4-chlorophenyl
4-bromophenyl
with R2 as a phenyl group (Table 1, compound 1a9), the inhibi-
tory activity was also much better than that of compound 1a1.
As having a phenyl group also made the compound more
drug-like,[17] we carried out further SAR studies of this novel
scaffold type. Aromatic rings carrying an electron-donating or
electron-withdrawing group at different positions of the
phenyl ring were synthesized (Table 1, compounds 1a13–
1a32), and the screening results showed that the presence of
a p-chloro, p-bromo, or p-thiomethyl substituent on the phenyl
ring (Table 1, compounds 1a21–1a23) significantly amplified
the compound’s inhibitory activity.
We next proceeded to tune the R1 group of the two stron-
gest lead compounds—1a21 and 1a23—by first replacing H
(=R1) with an electron-withdrawing group, such as a halide.
This did not provide any improvement in the inhibitory activi-
ties of the two compounds (Table 1, compounds 1a33–1a39).
However, if R1 was changed to an electron-donating methyl
group (Table 1, compounds 1a40–1a47), a slight enhancement
in the inhibition of WNV NS2B–NS3 protease was observed in
one of the two compounds (Table 1, compound 1a40). Howev-
er, increasing the size of the alkyl group from methyl to ethyl
resulted in a complete loss of inhibitory activity for the two
compounds (Table 1, compounds 1a48–1a49).
NA
>100
NA
11.39ꢀ0.31
45.22ꢀ3.28
7.66ꢀ0.91
9.39ꢀ1.00
6.28ꢀ1.39
NA
4-(methylthio)phenyl
4-methoxyphenyl
4-nitrophenyl
NA
>100
3-methyl-4-hydroxyphenyl
2-chloro-4-hydroxyphenyl
3,4-dichlorophenyl
4-hydroxy-3-nitrophenyl
4-fluoro-3-nitrophenyl
4-bromo-3-nitrophenyl
3,5-dibromophenyl
4-bromophenyl
>100
>100
>100
35.72ꢀ3.77
18.73ꢀ0.99
>100
>100
>100
F
F
F
4-chlorophenyl
4-(methylthio)phenyl
3-nitrophenyl
23.91ꢀ7.37
>100
>100
Cl phenyl
Cl 4-chlorophenyl
Cl 4-(methylthio)phenyl
16.28ꢀ2.56
>100
1a40 Me 4-chlorophenyl
1a41 Me 4-(methylthio)phenyl
1a42 Me 2-(4-chlorophenyl)furan-5-yl C6H5
1a43 Me phenyl
5.41ꢀ0.45
10.70ꢀ0.72
8.21ꢀ0.28
C6H5
>100
To explore the effect of the Ar group on the biological activi-
ty of the compound, we replaced the phenyl ring with
a phenyl group containing either an electron-donating or elec-
tron-withdrawing substituent (Table 1, compounds 1a50–
1a67). The best inhibition was observed with compound 1a53,
which contains a 4-methoxyphenyl group at the Ar position
and a 4-(methylthio)phenyl moiety at R2. In addition, the inhibi-
tory activity of compound 1a52 is similar to that of 1a21. Final-
ly, we synthesized compounds 1a68 and 1a69 to analyze the
combined effects of the preferred substituents at R1, R2, and
Ar. However, the inhibitory activities of compounds 1a68 and
1a69 were weaker than those of compounds 1a23, 1a40, or
1a53, the three strongest candidates identified from our earlier
SAR studies.
1a44 Me 2-(4-nitrophenyl)furan-5-yl
1a45 Me 4-bromophenyl
1a46 Me 3-nitrophenyl
C6H5
C6H5
C6H5
C6H5
C6H5
C6H5
>100
>100
>100
>100
NA
NA
NA
NA
8.47ꢀ2.54
5.56ꢀ0.10
>100
>100
>100
1a47 Me 4-bromo-3-nitrophenyl
1a48
1a49
1a50
1a51
1a52
1a53
1a54
1a55
1a56
1a57
1a58
1a59
1a60
1a61
1a62
1a63
1a64
1a65
1a66
Et 4-chlorophenyl
Et 4-(methylthio)phenyl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
phenyl
3-nitrophenyl
4-chlorophenyl
4-(methylthio)phenyl
4-chlorophenyl
4-(methylthio)phenyl
4-chlorophenyl
4-(methylthio)phenyl
4-chlorophenyl
4-(methylthio)phenyl
4-chlorophenyl
4-(methylthio)phenyl
4-chlorophenyl
4-MeOC6H4
4-MeOC6H4
4-MeOC6H4
4-MeOC6H4
4-BrC6H4
4-BrC6H4
3-BrC6H4
3-BrC6H4
4-ClC6H4
4-ClC6H4
3-ClC6H4
3-ClC6H4
4-FC6H4
4-FC6H4
3-FC6H4
3-FC6H4
4-CNC6H4
11.15ꢀ1.32
>100
6.20ꢀ0.15
23.35ꢀ5.05
11.14ꢀ2.42
Considering that the compounds tested thus far had been
racemates, we proceeded to investigate the influence of ste-
reochemistry on the compounds’ biological activities. However,
attempts to separate the R and S enantiomers of 1a23, 1a40,
and 1a53 proved futile, as a single peak was constantly ob-
served in chiral HPLC analysis. In addition, removal of the chiral
>100
4-(methylthio)phenyl
4-chlorophenyl
4-(methylthio)phenyl
4-chlorophenyl
13.29ꢀ1.55
>100
13.80ꢀ2.69
>100
ChemMedChem 0000, 00, 1 – 8
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
&
3
&
ÞÞ
These are not the final page numbers!

Y. Lam et al.
MED
center by DDQ oxidation gave a new scaffold 2a, the inhibitors
from which were completely inactive toward WNV NS2B–NS3
protease. (Table 1, compounds 2a1–2a3). Thus compounds
1a23, 1a40, and 1a53 were used for subsequent in vitro evalu-
ations.
The stabilities of compounds 1a23, 1a40, and 1a53 in the
assay buffer (pH 8) were determined by analyzing the amount
of compound remaining over time by LC–MS-IT-TOF detection.
Quantification of the compounds was determined by LCsolu-
tion software (Shimadzu). Compounds 1a23, 1a40, and 1a53
showed excellent stability in pH 8 buffer.
We next evaluated the cytotoxicities of compounds 1a23,
1a40, and 1a53 against baby hamster kidney fibroblasts
(BHK21 cells) for incubation times of 48 and 72 h. Dimethyl
sulfoxide (DMSO) was also included to serve as vehicle control,
as the test compounds were dissolved in DMSO. The amount
of DMSO used in the assay was kept constant at 0.2% of the
total volume; this is to ensure that DMSO does not result in cy-
totoxicity toward the cells. From the results obtained
(Figure 2), compounds 1a23, 1a40, and 1a53 were shown to
Figure 3. Uncompetitive inhibition of 1a40 with WNV NS2B–NS3 protease:
~
WNV protease assays were performed at 1a40 concentrations of 0 ( ), 5.0
&
^
( ), and 25 mm ( ), and substrate concentrations of 3.3, 8.3, 16.7, 33.3, and
83.3 mm. Double-reciprocal plots of 1/V versus 1/[S] for 1a40 at each inhibi-
tor concentration were generated with OriginPro 8.
tease, we focused our analysis on regions beyond the sub-
strate binding sites. Earlier studies have shown that integration
of residues 78–87 of NS2B into the protease–cofactor complex
affected the formation of the catalytic active sites.[18,19] Deletion
or mutagenesis of these residues produced an inert enzyme.
Therefore, the region on NS3 protease where key interactions
with the NS2B cofactor occur (residues 78–87) is important for
NS2B–NS3 protease activity, and particular attention was paid
to this region during our docking studies.
Because compound 1a40 contains a chiral center, the R and
S enantiomers were separately analyzed for their interactions
with the WNV NS2B–NS3 protease. Our docking analysis pre-
dicted that the binding energies for both enantiomers with
the protease are similar [(S)-1a40: ꢂ8.10 kcalmolꢂ1; (R)-1a40:
ꢂ8.22 kcalmolꢂ1]. Amino acid residues that are in close prox-
imity (<5 ꢁ) to the atoms of the S enantiomer include Arg78,
Leu79, and Phe85 of NS2B and Ser71, Lys73, Glu74, Arg76,
Lys88, Pro119, Glu120, Ile123, Ala164, Ile165, and Gln167 of
NS3 protease. Residues in close proximity (<5 ꢁ) to the atoms
of the R enantiomer are Val77, Arg78, and Leu79 of NS2B and
Lys88, Thr118, Glu120, Ile123, and Gln167 of NS3 protease. The
guanidine group in Arg78 of the NS2B cofactor binds strongly
with the Ar group of the S enantiomer via p–cation-type inter-
actions. A strong p–cation-type interaction was also observed
between the 4-chlorophenyl moiety of (S)-1a40 and Lys73 on
the NS3 protease (Figure 4a). On the other hand, (R)-1a40
could potentially form a hydrogen bond with Lys73 on the
NS3 protease as well as a p–cation interaction with the same
residue. However, we did not observe any strong interaction
between the R enantiomer and key residues from the NS2B
binding cavity (Figure 4b). This indicates that the S enantiomer
of 1a40 interferes more strongly with the productive interac-
tions between the NS2B cofactor and the NS3 protease
domain; it is therefore the preferred isomer for inhibition of
WNV NS2B–NS3 protease.
Figure 2. MTS assay results obtained at incubation times of 48 (&) and 72 h
(&) with the BHK21 cell line. Compounds were added at a final concentra-
tion of 50 mm.
be non-cytotoxic to BHK21 cells. To determine the mode of in-
hibition of compound 1a40, we followed the kinetics of inhibi-
tion using Pyr-RTKR-AMC as substrate. This provided a sub-
strate-concentration-dependent inhibition of NS2B–NS3 pro-
tease activity (Figure 3). Lineweaver–Burk analysis of the initial
velocities suggested that inhibition by 1a40 is uncompetitive
with respect to substrate, with a Ki value of 4.47ꢀ0.37 mm.
Docking analysis
The crystal structure of the WNV NS2B–NS3 protease in com-
plex with peptide Bz-Nle-Lys-Arg-Arg-H (PDB ID: 2FP7) was re-
ported earlier.[18] To identify the potential binding site of com-
pound 1a40 in NS2B–NS3 protease, in silico docking studies
were performed by using these crystal structure coordinates.
As 1a40 is an uncompetitive inhibitor of WNV NS2B–NS3 pro-
&4
&
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 8
ÝÝ
These are not the final page numbers!

WNV Protease Inhibitors
MED
system with a Phenomenex C18 column (50 mmꢂ3.0 mm, 5 mm).
Purity of the compounds was determined at l 254 nm, and inte-
1
gration was carried out with Shimadzu LCsolution software. H and
13C NMR spectra were recorded using CDCl3, [D6]acetone, or
[D6]DMSO as solvents and tetramethylsilane (TMS) as internal refer-
ence on a Bruker AMX500 or a Bruker ACF 300 Fourier-transform
spectrometer. The following abbreviations are used to denote mul-
tiplicities: s (singlet), d (doublet), t (triplet), dd (doublet of dou-
blets), q (quartet), and m (multiplet). The number of protons (n) for
a given resonance is indicated as nH. Mass spectrometry was per-
formed on a Finnigan/MAT LCQ mass spectrometer using electron
spray (ESI) ionization.
General procedure for the synthesis of 4: Cl3CCH(OH)2 (3.84 g,
23.2 mmol) and Na2SO4 (60 g, 422 mmol) were dissolved in H2O
(20 mL) in a 500 mL beaker and warmed to 358C. A warm solution
of the commercially available aniline or its derivatives
3
(21.5 mmol) in H2O (6 mL) and an aqueous solution of concentrat-
ed HCl (1.83 mL, 22 mmol) were added (a white precipitate of the
aniline sulfate was formed) followed by a warm solution of
NH2OH·HCl (4.72 g, 67.94 mmol) in H2O (8.25 mL). The mixture was
heated at 908C for 2.5 h and then allowed to cool to 508C before
filtering. The pale-cream solid obtained was washed by stirring
with H2O (100 mL), filtered, and dried overnight in a vacuum oven
at 508C to give the corresponding isonitrosoacetanilide 4 in 80–
95% yield.
General procedure for the synthesis of 5: Concentrated H2SO4
(18 mL, 337.7 mmol) was heated in a 250 mL beaker at 608C. The
respective dried compound 4 (7.3 mmol) was added portionwise
with stirring over 30 min so that the temperature did not exceed
658C. The mixture was then heated at 808C for 15 min, allowed to
cool to 708C, and cooled on ice-cold H2O. The solution was then
poured very slowly into ice and allowed to stand until a reddish
precipitate of isatin 5 was formed. The crude solid 5 was filtered
and used in the next step without further purification.
Figure 4. Molecular docking of compound 1a40 onto WNV NS2B–NS3 pro-
tease: The WNV protease crystal structure coordinates (PDB ID: 2FP7) were
used for docking. The protease (ribbon backbone and wireframe side chain
residues) and docked compounds a) (S)-1a40 and b) (R)-1a40 (stick models)
were generated with Discovery Studio 3.1 client.
General procedure for the synthesis of 6: The respective isatin 5
(10.2 mmol) was dissolved in 10% aqueous KOH (10 mL,
45.5 mmol) and then treated with thiosemicarbazide (985 mg,
10.7 mmol). The reaction mixture was heated at 1158C for 1 h,
cooled to room temperature, and then poured over ice. Acidifica-
tion of the reaction mixture to ~pH 5 with glacial acetic acid pro-
vided a yellow fluffy precipitate, which was filtered, washed copi-
ously with H2O, dried in air and then dried in a vacuum oven (at
508C) to afford a yellow solid 6 in 60–78% yield. Purification of
compound 6 by column chromatography gave pure product at
52–70%.
Conclusions
This study reports an in vitro assay for the WNV NS2B–NS3 pro-
tease that yielded a lead inhibitor with a 9,10-dihydro-3H,4aH-
1,3,9,10a-tetraazaphenanthren-4-one scaffold. Refinement of
this initial hit by synthesizing and evaluating a focused library
of
9,10-dihydro-3H,4aH-1,3,9,10a-tetraazaphenanthren-4-one
General procedure for the synthesis of 7: The respective crude
compound 6 (6 mmol) was dissolved in a mixture of EtOH (12 mL)
and 10% aqueous KOH (18 mL, 27.3 mmol), then treated with a so-
lution of the respective benzyl bromide (6 mmol) and allowed to
stir overnight. Thereafter, the reaction mixture was diluted with
Et2O (100 mL) and H2O (70 mL). The ether layer was separated, and
the aqueous layer was washed with Et2O (3ꢂ100 mL) and then
poured over ice. Upon careful dropwise addition of glacial acetic
acid with vigorous shaking at 0–48C, a yellow precipitate formed.
This precipitate was filtered, washed with H2O (3ꢂ30 mL), and
dried in a vacuum oven (at 508C) to provide compound 7 in 80–
90% yield.
compounds provided analogue 1a40, which shows low-micro-
molar inhibition of the WNV NS2B–NS3 protease. Compound
1a40 blocks the interaction between the NS2B cofactor and
NS3 protease in an uncompetitive manner.
Experimental Section
Chemistry
All reagents and anhydrous solvents were obtained from Aldrich,
Merck, Lancaster, or Fluka and were used without further purifica-
tion. Flash chromatography was performed with silica (Merck, 70–
230 mesh), whilst TLC was carried out on pre-coated plates (Merck
silica gel 60, F254) and visualized with UV light. Purity of final com-
pounds was determined by HPLC using a Shimadzu LC–MS-IT-TOF
General procedure for the synthesis of 1a: To a suspension of the
respective compound 7 (0.32 mmol) in anhydrous EtOH (6 mL) was
added glacial acetic acid (25 mL, 0.42 mmol) followed by the re-
spective aldehyde (0.42 mmol, 1.3 equiv). The reaction mixture was
ChemMedChem 0000, 00, 1 – 8
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
&
5
&
ÞÞ
These are not the final page numbers!

Y. Lam et al.
MED
held at reflux for 10–15 min, resulting in a dark-red solution. Upon
cooling to room temperature, the orange–yellow solid 1a precipi-
tated from solution. Sometimes the hot reaction mixture contained
insoluble particles, which were removed by filtering the reaction
mixture under hot conditions. The filtrate was then cooled to room
temperature to provide the crude solid compound 1a. The crude
product was recrystallized with EtOH, filtered, and washed with
cold EtOH (2ꢂ3 mL). In a few cases, the product was purified by
column chromatography and then dried in a vacuum oven at
508C. Compound 1a was obtained in 50–90% yield.
Determination of inhibition mechanism: Two different inhibitor con-
centrations and a no-inhibitor control (0–25 mm) were each assayed
at five substrate concentrations ranging from 3.3 to 83.3 mm. In
each assay, the enzyme and inhibitor were incubated at 378C for
15 min, followed by the addition of the substrate to start the kinet-
ics measurement. The rate of substrate cleavage (v) was monitored
using the F200 microplate reader. To illustrate the inhibition mech-
anism, 1/V versus 1/[S] was plotted for each inhibitor concentration
using OriginPro 8.
General procedure for the synthesis of 2a: The respective com-
pound 1a (0.16 mmol) was dissolved in CH2Cl2 (10 mL), and the re-
action mixture was kept at 08C followed by slow addition of DDQ
(0.2 mmol). After the addition of DDQ, the temperature of the reac-
tion mixture was allowed to rise to room temperature and was
kept at this temperature for 4 h. During this time, the reaction was
monitored by TLC. Upon completion, the reaction mixture was ex-
tracted with a mixture of CH2Cl2/H2O (20:30). The organic layer was
concentrated, purified by column chromatography (EtOAc/hexane
1:4!1:1) and dried under vacuum at 508C to yield compound 2a
in 70–85% yield.
In silico studies
Molecular docking simulations were performed with the Auto-
Dock4 program. Both enantiomers of compound 1a40 were con-
structed in ChemBio 3D Ultra 11.0 by energy minimization. The
active site of WNV protease was prepared with the crystal structure
of WNV NS2B–NS3 protease in complex with peptide Bz-Nle-Lys-
Arg-Arg-H (PDB ID: 2FP7).[18] The structure was stripped of all water
molecules and bound ligand. Hydrogen atoms were added to all
polar atoms of the protein, followed by addition of Gasteiger–Mar-
sili charges. AutoDock simulations were performed using the La-
marckian Genetic Algorithms (GA) subroutine at default settings
for GA population size, crossover rate, mutation rate, and starting
with fully randomized ligand position, orientation, and conforma-
tion; 50 GA runs were performed for the inhibitor–enzyme pair,
and results were analyzed using Discovery Studio 3.1 client.
Biology
All compounds used in biological assays were ꢁ98% pure. Com-
pounds were tested against WNV NS2B–NS3 protease using the
SensoLyte 440 WNV Protease Assay Kit (Cat. #72079) and active re-
combinant WNV protease (Cat. #72081), which were purchased
from Anaspec (USA). This protease assay kit uses the fluorogenic
peptide Pyr-RTKR-AMC as substrate. Eight different substrate con-
centrations ranging from 1 to 80 mm were incubated in 96-well
plates with 0.3 mgmLꢂ1 recombinant WNV protease at 378C in the
buffer provided. The increase in fluorescence intensity was moni-
tored with an Infinite F200 microplate reader (Tecan, Switzerland)
at an excitation wavelength of 354ꢀ10 nm and an emission wave-
length of 442ꢀ15 nm. The initial velocity was determined from
the linear portion of the progress curve, and the value of KM
(3.45ꢀ0.41 mm) was determined by the Michaelis–Menten equa-
tion: v=Vmax [S]/([S]+KM). Triplicate measurements were taken at
each data point, and the data are reported as mean ꢀSE. In the
preliminary WNV NS2B–NS3 protease inhibition assay, compounds
1a and 2a were screened at a fixed concentration of 100 mm to
highlight potential inhibitors. Compounds effecting >50% inhibi-
tion were further investigated for their IC50 determination.
Acknowledgements
We thank the National University of Singapore for financial sup-
port of this work (ARF: R-143-000-308-112).
Keywords: antiviral agents · bioorganic chemistry · docking
studies · inhibitors · proteases
[1] B. E. E. Martina, P. Koraka, A. D. M. E. Osterhaus, Curr. Opin. Investig.
Drugs 2010, 11, 139–146.
[2] a) K. O. Murray, S. Baraniuk, M. Resnick, R. Arafat, C. Kilborn, R. Shallen-
berger, T. L. York, D. Martinez, M. Malkoff, N. Elgawley, W. McNeely, S. A.
Khuwaja, Vector Borne Zoonotic Dis. 2008, 8, 167–174; b) M. Y. Chowers,
R. Lang, F. Nassar, D. Ben-David, M. Giladi, E. Rubinshtein, A. Itzhaki, J.
Mishal, Y. Siegman-Igra, R. Kitzes, N. Pick, Z. Landau, D. Wolf, H. Bin, E.
Mendelson, S. D. Pitlik, M. Weinberger, Emerging Infect. Dis. 2001, 7,
675–678.
[3] a) G. Bjorn, Nat. Med. 2008, 14, 700; b) J. J. Sejvar, Clin. Infect. Dis. 2007,
44, 1617–1624.
Determination of IC50: For IC50 calculations, recombinant WNV pro-
tease at a concentration of 0.3 mgmLꢂ1 and seven different concen-
trations of the inhibitor ranging from 10 nm to 100 mm were used.
For each experiment, the protease was pre-incubated with the in-
hibitor at 378C for 15 min in separate wells, and enzyme reactions
were initiated by adding the Pyr-RTKR-AMC substrate at a final
concentration of 16.7 mm. The increase in fluorescence intensity
was monitored continuously with an Infinite F200 microplate
reader at an excitation wavelength of 354 nm and emission wave-
length of 442 nm. Fluorescence values obtained from the positive
control (no inhibitor) were considered as 100% complex formation,
and those values obtained in the presence of inhibitors were calcu-
lated as the percentage of inhibition relative to control. Triplicate
measurements were obtained at each data point. The IC50 values
were calculated using a sigmoidal dose–response curve with the
GraphPad Prism 3.0 software package (San Diego, CA, USA). Tripli-
cate measurements were recorded as the mean ꢀSE.
[4] S. Mukhopadhyay, R. J. Kuhn, M. G. Rossman, Nat. Rev. Microbiol. 2005,
3, 13–22.
[5] M. A. Brinton, Annu. Rev. Microbiol. 2002, 56, 371–402.
[6] S. J. Wong, R. H. Boyle, V. L. Demarest, A. N. Woodmansee, L. D. Kramer,
H. Li, M. Drebot, R. A. Koski, E. Fikrig, D. A. Martin, P.-Y. Shi, J. Clin. Micro-
biol. 2003, 41, 4217–4223.
[7] K. J. Chappell, M. J. Stoermer, D. P. Fairlie, P. R. Young, J. Biol. Chem.
2006, 281, 38448–38458.
[8] R. Yusof, S. Clum, M. Wetzel, H. M. Murthy, R. Padmanabhan, J. Biol.
Chem. 2000, 275, 9963–9969.
[9] T. A. Nall, K. J. Chappell, M. J. Stoermer, N.-X. Fang, J. D. A. Tyndall, P. R.
Young, D. P. Fairlie, J. Biol. Chem. 2004, 279, 48535–48542.
[10] a) S. A. Shiryaev, B. I. Ratnikov, A. V. Chekanov, S. Sikora, D. V. Rozanov, A.
Godzik, J. Wang, J. W. Smith, Z. Huang, I. Lindberg, M. A. Samuel, M. S.
Diamond, A. Y. Strongin, Biochem. J. 2006, 393, 503–511; b) J. E. Knox,
N. L. Ma, Z. Yin, S. J. Patel, W.-L. Wang, W.-L. Chan, K. R. Ranga Rao, G.
Wang, X. Ngew, V. Patel, D. Beer, S. P. Lim, S. G. Vasudevan, T. H. Keller, J.
Med. Chem. 2006, 49, 6585–6590; c) M. J. Stoermer, K. J. Chappell, S.
&6
&
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 8
ÝÝ
These are not the final page numbers!

WNV Protease Inhibitors
MED
Liebscher, C. M. Jensen, C. H. Gan, P. K. Gupta, W.-J. Xu, P. R. Young, D. P.
Fairlie, J. Med. Chem. 2008, 51, 5714–5721; d) S. M. Tomlinson, S. J. Wa-
towich, Biochemistry 2008, 47, 11763–11770.
[14] Y. Lai, L. Ma, W. Huang, X. Yu, Y. Zhang, H. Ji, J. Tian, Bioorg. Med. Chem.
Lett. 2010, 20, 7349–7353.
[15] P. Polychronopoulos, P. Magiatis, A.-L. Skaltsounis, V. Myrianthopoulos,
E. Mikros, A. Tarricone, A. Musacchio, S. M. Roe, L. Pearl, M. Leost, P.
Greengard, L. Meijer, J. Med. Chem. 2004, 47, 935–946.
[16] L.-J. Ma, X.-X. Li, T. Kusuyama, I. E.-T. El-Sayed, T. Inokuchi, J. Org. Chem.
2009, 74, 9218–9221.
[17] C. A. Lipinski, Drug Discovery Today Technol. 2004, 1, 337–341.
[18] P. Erbel, N. Schiering, A. D’Arcy, M. Renatus, M. Kroemer, S. P. Lim, Z. Yin,
T. H. Keller, S. G. Vasudevan, U. Hommel, Nat. Struct. Mol. Biol. 2006, 13,
372–373.
[19] a) A. E. Aleshin, S. A. Shiryaev, A. Y. Strongin, R. C. Liddington, Protein Sci.
2007, 16, 795–806; b) I. Radichev, S. A. Shiryaev, A. E. Aleshin, B. I. Ratni-
kov, J. W. Smith, R. c. Liddington, A. Y. Strongin, J. Gen. Virol. 2008, 89,
636–641.
[11] a) D. Ekonomiuk, X.-C. Su, K. Ozawa, C. Bodenreider, S. P. Lim, Z. Yin,
T. H. Keller, D. Beer, V. Patel, G. Otting, A. Caflisch, D. Huang, PLoS Negl.
Trop. Dis. 2009, 3, e356; b) V. K. Ganesh, N. Muller, K. Judge, C.-H. Luan,
R. Padmanabhan, K. H. M. Murthy, Bioorg. Med. Chem. 2005, 13, 257–
264; c) P. A. Johnston, J. Phillips, T. Y. Shun, S. Shinde, J. S. Lazo, D. M.
Huryn, M. C. Myers, B. I. Ratnikov, J. W. Smith, Y. Su, R. Dahl, N. D. P. Cos-
ford, S. A. Shiryaev, A. Y. Strongin, Assay Drug Dev. Technol. 2007, 5,
737–750; d) N. H. Mueller, N. Pattabiraman, C. Ansarah-Sobrinho, P. Vis-
wanathan, T. C. Pierson, R. Padmanabhan, Antimicrob. Agents Chemother.
2008, 52, 3385–3393; e) D. Dou, P. Viwanathan, Y. Li, G. He, K. R. Allis-
ton, G. H. Lushington, J. D. Brown-Clay, R. Padmanabhan, W. C. Groutas,
J. Comb. Chem. 2010, 12, 836–843.
[12] S. Sidique, S. A. Shiryaev, B. I. Ratnikov, A. Herath, Y. Su, A. Y. Strongin,
N. D. P. Cosford, Bioorg. Med. Chem. Lett. 2009, 19, 5773–5777.
[13] a) G. Doleschall, K. Lempert, Tetrahedron 1973, 29, 639–649; b) G. Dole-
schall, K. Lempert, Tetrahedron 1974, 30, 3997–4012.
Received: March 13, 2012
Revised: April 5, 2012
Published online on && &&, 0000
ChemMedChem 0000, 00, 1 – 8
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
&
7
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
S. Samanta, T. Cui, Y. Lam*
Running cofactor interference: In vitro
assays with West Nile virus (WNV)
&& – &&
NS2B–NS3 protease resulted in the dis-
covery of 9,10-dihydro-3H,4aH-1,3,9,10a-
tetraazaphenanthren-4-ones as a new
class of inhibitors of this enzyme. Opti-
mization of the lead compound led to
an uncompetitive WNV NS2B–NS3 inhib-
itor with an IC50 value of 5.41ꢀ0.45 mm.
Discovery, Synthesis, and in vitro
Evaluation of West Nile Virus Protease
Inhibitors Based on the 9,10-Dihydro-
3H,4aH-1,3,9,10a-
tetraazaphenanthren-4-one Scaffold
&8
&
www.chemmedchem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 0000, 00, 1 – 8
ÝÝ
These are not the final page numbers!