Synthesis and biological evaluation of α-ketoamides as inhibitors of the Dengue virus protease with antiviral activity in cell-culture

Article abstract of DOI:10.1016/j.bmc.2011.05.015

The development of small molecule inhibitors of the viral protease is of considerable interest for the treatment of emergent flaviviral diseases such as Dengue or West Nile fever. Until today little progress has been made in finding drug-like compounds th

Full text of DOI:10.1016/j.bmc.2011.05.015

Bioorganic & Medicinal Chemistry 19 (2011) 4067–4074
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis and biological evaluation of
a-ketoamides as inhibitors of the Dengue
virus protease with antiviral activity in cell-culture
Christian Steuer a, Christian Gege a, Wolfgang Fischl b, Karl H. Heinonen a, Ralf Bartenschlager b,
Christian D. Klein a,
⇑
a Medizinische Chemie, Universität Heidelberg, Im Neuenheimer Feld 364, 69120 Heidelberg, Germany
b Molekulare Virologie, Department of Infectious Diseases, Molecular Virology, University Heidelberg, Im Neuenheimer Feld 345, 69120 Heidelberg, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
The development of small molecule inhibitors of the viral protease is of considerable interest for the
treatment of emergent flaviviral diseases such as Dengue or West Nile fever. Until today little progress
has been made in finding drug-like compounds that inhibit the protease and provide a starting point
for lead optimization. We describe here the initial steps of a drug discovery effort that focused on the sty-
ryl pharmacophore, combined with a ketoamide function to serve as electrophilic trap for the catalytic
serine. This resulted in a fragment-like lead compound with reasonable target affinity and good ligand
efficiency, which was extensively modified to explore structure–activity relationships. Selected com-
pounds were cross-tested against the West Nile virus protease and thrombin, indicating that selectivity
for one or more flaviviral proteases can be achieved. Finally, the antiviral activity of several protease
inhibitors was confirmed in a cell-culture model of Dengue virus replication. The SAR presented here
may serve as starting point for further drug discovery efforts with the aim of targeting flaviviral
proteases.
Received 8 April 2011
Revised 10 May 2011
Accepted 10 May 2011
Available online 18 May 2011
Keywords:
Ketoamides
NS2B–NS3 protease
Protease inhibitors
Dengue virus
Flavivirus
Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction
NS5). Host proteases (furin, signalase) as well as the two-compo-
nent viral protease NS2B–NS3 (the ‘DenV protease’) process the
The Dengue virus (DenV) belongs to the family Flaviviridae and
is closely related to the West Nile virus (WNV), the Yellow Fever
virus (YFV) and the Hepatitis C virus (HCV). Dengue viruses are
transmitted by mosquitos of the species Aedes aegypti and cause
clinical symptoms ranging from mild fever to Dengue hemorrhagic
fever (DHS) and Dengue shock syndrome (DSS). DenV can be sub-
grouped into four highly homologous but antigenically distinct ser-
otypes (DenV 1–4). Immunity against one serotype does not
protect against an infection by another but may enhance disease
severity if re-infection with another serotype occurs.1 At the mo-
ment an estimated 2.5 billion people live in regions at high risk
for epidemic transmission. Because of global warming and interna-
tional travel, the virus currently spreads into the temperate zones.
Accordingly, there is a considerable interest in developing effective
vaccines or chemotherapeutics against DenV and other closely re-
lated flaviviruses like WNV. Until today neither vaccination nor a
targeted drug therapy is available for treatment of Dengue infec-
tions. DenV consists of a single stranded and positive sensed RNA
genome of approximately 11 kb. The genome encodes for a poly-
protein with three structural proteins (C, prM, E) and seven non-
structural (NS) proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and
polyprotein.2 The correct processing of the precursor is essential
for the viral life-cycle.3,4 Therefore, the viral protease is an interest-
ing target for the development of DenV therapeutics.
NS3 is a multifunctional protein in which the N-terminal do-
main (NS3pro, 184 residues) represents the protease. The C-termi-
nal part comprises the enzymatic functions of a helicase, a
nucleoside triphosphatase and RNA 50-triphosphatase.3,5 The pro-
tease is a serine endoprotease with the catalytic triad His51-
Asp75-Ser135. In common with mammalian thrombin, flaviviral
NS2B–NS3 requires basic residues at the P1 position of the sub-
strate. Basic residues in the P2 position greatly increase sub-
strate-binding. The P10 position is preferably glycine or serine,
whereas the other C-terminal positions of the substrate are not rel-
evant for substrate recognition.6 The protease displays a chymo-
trypsin fold consisting of two b-barrels each formed by six b-
strands.7 For optimal catalytic activity the protease requires the
hydrophilic core domain of the NS2B protein as a cofactor.8 The
cofactor provides a single b-strand to the N-terminal b-barrel and
stabilizes this domain by covering hydrophobic residues of
NS3pro.9
Targeting the flaviviral proteases of DenV, WNV or YFV appears
to be similarly difficult for medicinal chemistry as the drug discov-
ery of inhibitors against the HCV NS3–NS4A protease. We were un-
able to identify any reports of small-molecular, drug-like inhibitors
of DenV protease that entered the clinical development phases.
⇑
Corresponding author. Tel.: +49 6221 544875; fax: +49 6221 546430.
E-mail address: c.klein@uni-heidelberg.de (C.D. Klein).
0968-0896/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2011.05.015

4068
C. Steuer et al. / Bioorg. Med. Chem. 19 (2011) 4067–4074
In the other published efforts aimed at flaviviral proteases, the
substrate-binding region is usually addressed by a tetrameric (or
larger) peptide chain that resembles the non-prime residues of
the natural substrate, combined with various electrophilic traps
such as aldehydes and fluoroketones, to address the catalytic ser-
ine residue.10,11 Due to the basic nature of the non-prime amino
acids of the substrate (Arg, Lys), and the considerable size of the
resulting peptide, these inhibitors are not drug-like and probably
not cell-permeant. A noteworthy exception is the work of Bodenre-
ider et al.,12 who described and characterized a set of drug-like
inhibitors with molecular masses in the 500–600 g/mol range
and KD values in the lower micromolar range that were probably
identified by high-throughput screening. Further development or
cell-culture studies with these compounds, however, have not
been reported.
The drug discovery effort described here was initiated with the
screening of small-molecular aldehydes. The rationale behind this
was that we required a reference inhibitor (without the require-
ment of drug-likeness) to assess the optimized screening proce-
dure described previously.13 Aldehydes can interact with the
nucleophilic serine of serine proteases in a covalent–reversible
manner by forming hemiketals. In order to generate compounds
with increased drug-likeness, we decided to replace the aldehyde
Figure 1. Reagents and conditions: (i) pyruvic acid, KOH (1.5 equiv), MeOH; (ii)
HNR1R2, HATU, HOAt, NEt3, DMF, 0 °C–rt; (iii) Cu(OTf)2, CH(OMe)3, CH2Cl2,
microwave, 100 °C; (iv) DBU, CH2Cl2, rt; (v) SiCl4, CH2Cl2, ꢀ78 °C, 4 h, then NaHCO3;
(vi) Dess–Martin periodinane, CH2Cl2, rt.
hydes. Aldol condensation of pyruvic amide under basic conditions
using 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) as catalyst re-
sulted in no improvements. Passerini reaction (Route C) was per-
formed as described by Denmark and Fan32 and gave moderate
yields. The hydroxyamides were subjected to Dess–Martin oxida-
tion as described by Speicher et al.33 This two-step method was
group with a more selective electrophilic trap—the
function—and thus to create a SAR dataset of -ketoamides and
-ketoamide derivatives. For the serine protease from HCV, which
is related to DenV protease, there are currently several inhibitors in
clinical trials that employ the -ketoamide group as electrophilic
a-ketoamide
chosen for the a-ketoamides that could not be obtained by Route
A and B (7–9 in Table 1).
a
a
With the goal of exploring the aryl moiety, we kept the b,c-
unsaturated
a-keto tert-butyl amide moiety constant. Inhibitors
a
listed in Table 2 were synthesized as depicted in Figure 1 with
the following exceptions: phenol 20 was synthesized from com-
pound 15 via O-demethylation using boron tribromide. Compound
21 was synthesized by reducing nitro derivative 26 with excess tin
chloride in refluxing ethanol. Derivative 34 was obtained by acet-
ylation of amine 21 with Ac2O in pyridine. Finally, nitrile 35 was
prepared from bromide 25 using zinc(II) cyanide and Pd(PPh3)4
as catalyst. For details see the data in the Supplementary data.
For the exploration of the substituent on the amide nitrogen,
the phenyl moiety was kept constant. The inhibitors listed in Ta-
ble 3 were synthesized as depicted in Figure 1 with the exception
of compound 46, which was synthesized from alcohol 44 by acet-
ylation with Ac2O in pyridine.
trap (Telaprevir, Boceprevir).14,15 This moiety is common in phar-
maceuticals like serine or cysteine protease inhibitors16–20 and
can be found in several bioactive natural products such as the
immunosuppressant drugs FK-506 and rapamycin.21 The binding
towards serine proteases is based on the formation of a hemiketal
or hemithioketal with the –OH or –SH group of serine or cysteine
protease, respectively.22
In contrast, b,c-unsaturated a-ketoamides are found only in a
few number of pharmaceuticals with biological activities such as
antibacterials23,24 and in one natural product.25 They are also used
as chemical intermediates for the synthesis of 5-aryl-3-carboxam-
ide-4,5-dihydro-(1H)-pyrazoles.26 Therefore, the work presented
here, which also includes a new synthetic route towards this com-
pound class, is also relevant in the light of synthetic organic
chemistry.
Table 1
Influence of modifications at the core structure of lead compound 1 on the inhibition
of DenV NS2B–NS3 serine protease. Inhibitor concentration: 50 lM
2. Results and discussion
2.1. Chemistry
#
Structure
% inhib.
(sd%)
#
Structure
% inhib.
(sd%)
26.3
(3.2)
20.4
(1.6)
For the synthesis of the targeted 4-substituted 2-oxo-but-3-
enoic acid amides, several approaches were pursued as outlined
in Figure 1. Route A started with an aldol condensation of aromatic
aldehydes with pyruvic acid in the presence of potassium hydrox-
ide (1.5 equiv),27 followed by peptidic coupling of the intermediate
with various amines and O-(7-azabenzotriazol-1-yl)-N,N,N0,N0-
tetramethyluronium-hexafluorophosphate (HATU) and 1-hydro-
xy-7-azabenzotriazole (HOAt) as powerful coupling reagent.28
The key compound 1 was crystallized and showed trans geometry
in X-ray diffraction structure analysis (data not shown). Since with
some aldehydes (e.g., with 2-pyridyl aldehyde) only a sluggish al-
dol condensation was observed, we investigated the direct cou-
pling of an aldehyde with N-tert-butyl pyruvic amide29 or N-
benzyl pyruvic amide30 using copper triflate as catalyst as recently
described for the coupling of pyruvic esters (Route B).31 The reac-
tion times could be reduced by the use of microwave irradiation.
Particularly good yields could be obtained from electron-rich alde-
1
2
3
4
5
6
7
0
(4.3)
6.9
(0.7)
8
7.4
(4.7)
21.5
(1.6)
9
19.6
3.1
8.2
(3.2)
10
11
25.6
(3.4)
12.6
(1.4)
15.2
(3.0)

C. Steuer et al. / Bioorg. Med. Chem. 19 (2011) 4067–4074
4069
Table 2
Table 3
Inhibitory activities of unsaturated
a
-ketoamides against DenV NS2B–NS3 serine
Inhibitory activities of unsaturated
a
-ketoamides against DenV NS2B–NS3 serine
protease: Variation of the
c
-aryl substituent. Inhibitor concentration: 50
l
M
protease: Variation of the substituent at the amide nitrogen. Inhibitor concentration:
50
l
M
% inhib.
#
% inhib.
(sd%)
#
R
% inhib.
(sd%)
#
R
% inhib.
(sd%)
#
(sd%)
26.3
(3.2)
25.9
(1.4)
26.3
(3.2)
36.1
(6.8)
1
43
1
18
19
23.8
(6.3)
26.1
(1.3)
27.6
(5.2)
27.8
(0.4)
39
40
41
44
45
46
12
26.9
(0.9)
28.9
(2.6)
25.5
(1.2)
37.6
(2.4)
31.2
(3.6)
35.8
(6.1)
13
20
14.5
(4.4)
14.6
(2.1)
8.9
(2.3)
40.6
(0.9)
42
48
47
51
14
15
16
21
22
23
16.7
(2.8)
23.9
(3.5)
0
(4.9)
35.1
(3.0)
17.8
(2.1)
16.8
(7.5)
29.9
(2.7)
29.0
(1.8)
49
50
52
53
23.3
(2.8)
15.9
(5.2)
29.4
(4.2)
30.8
(1.5)
17
25
24
32
15.1
(1.3)
39.1
(0.7)
study this compound class in more detail, in order to evaluate
the potential for further development.
12.7
(2.2)
13.8
(1.2)
26
27
28
33
34
35
The occurrence of unspecific-/promiscuous inhibitors is a wide-
spread phenomenon in drug discovery. Therefore many tools have
been developed to eliminate such candidates in early stages of this
process.12 We were well aware of the dangers of promiscuous
binding and took various measures to exclude this risk before
advancing the project. These measures included: Analysis of com-
pound solutions by dynamic light scattering; testing against
numerous other enzymes including metalloproteases, ligases and
other serine proteases; Analysis of the Hill slope of dose–response
curves; application of orthogonal assay procedures such as a HPLC-
based assay and the tryptophane-fluorimetric assay described by
Bodenreider et al.12 None of these procedures gave any hint what-
soever on a promiscuous binding mode of the compounds. Further-
more, kinetic analyzes of selected compounds demonstrated a
competitive binding mode.
9.6
(1.1)
32.9
(2.6)
12.9
(1.7)
13.5
(3.4)
18.4
(3.0)
31.3
(2.7)
29
36
23.3
(2.0)
29.8
(3.7)
30
31
37
38
27.7
(4.0)
4.4
(6.3)
The ligand efficiency (LE)36,37 as measure of affinity per atom (or
per molecular weight unit) of compd 1 is 16.6 Perola37 determined
ligand efficiencies of typical lead compounds to be generally above
12.4. Ligand efficiencies of the optimized compounds shown here,
such as 18 and 32, are in the range of 14–15, which is a typical va-
lue for clinical candidates and established drugs.
We therefore conclude that the inhibitors presented here ex-
ploit the potential interaction sites near the catalytic serine of Den-
gue protease with high effectivity. These analogs represent
valuable ‘serine anchors’ which may be expanded to yield high-
affinity inhibitors. Finally, the cell-culture data shown below
clearly indicate that these chemotypes have the potential to be
used in further preclinical studies on Dengue therapy.
To investigate the relevance of the central b,
c
-unsaturated a-
keto moiety, the analogs shown in Table 1 were synthesized. The
introduction of the cyclopropyl moiety in compound 1 was accom-
plished by the method of Corey and Chaykovsky34 using trimethyl-
sulfoxonium iodide and sodium hydride as base to give derivative
3. By reaction of compound 1 with trimethylsulfonium iodide and
potassium tert-butylate as base,34 oxirane 11 was obtained. Com-
pound 6 can be prepared as previously described.35 The other
derivatives were prepared by applying the Passerini reaction
(Fig. 1, Route C) to the corresponding aldehyde and, with the
exception of alcohol 10, with subsequent oxidation using Dess–
Martin periodinane.
2.3. Structure–activity relationships
2.2. General considerations: lead-likeness, absence of
promiscuous inhibition, ligand efficiency
In analogy to the
ases, we expected a nucleophilic attack by the catalytic serine at
the ketoamide -carbonyl group (1,2-addition), resulting in a cova-
lent-reversible binding mode. A Michael (1,4) addition of the serine
at the -position appears unlikely, although 1,4-addition to the
a-ketoamide inhibitors of other serine prote-
It has not escaped our attention that, in comparison to other
drug development projects, the compounds described here do
not have a high affinity towards the target enzyme. In the absence
of any other pursuable, small-molecular leads—either from litera-
ture or our own screening process—we nevertheless decided to
a
c
nucleophilic hydroxyl group of threonine proteases (e.g., in the
proteasome) has been described.38 To elucidate the binding mode

4070
C. Steuer et al. / Bioorg. Med. Chem. 19 (2011) 4067–4074
and to explore possible alternatives for the potentially problematic
acrylic moiety, a number of variants to the unsaturated -ketoa-
mide core structure were generated. These variants are presented
pocket. Many substituents are accepted in this pocket—bulky
tert-butyl residues as well as smaller alkyl groups, for example,
methyl, propyl or cyclopropyl, which can additionally be substi-
tuted with heteroatoms. While the size could be extended to long-
er alkyl chains, for example, the pentyl analog 43, ‘branched’ bulk
like cyclohexyl (compd 42) is less tolerated. In addition, hydropho-
bic aromatic systems like benzyl or 1-methylbenzyl should be
avoided, whereas more hydrophilic residues like pyridyl (compds
49 and 50) are able to compensate that effect. Based on the
a
in Table 1.
The saturated analog 2 has no effect on the DenV protease.
Reduction of the linker length between the electrophilic trap and
the phenyl-/cyclohexane moiety results in a decreased activity of
the compounds (compds. 7–9). Interestingly, reduction of the dis-
tance between the aryl moiety and the a-keto functionality is tol-
erated as depicted with compound 7. However, a further reduction
of the distance is not tolerated (compds 8 and 9). Replacement of
the double bond with a cyclopropyl group—to explore the rele-
vance of steric and conformational effects—resulted in marginal
assumption that the substituents on the a-ketoamide nitrogen oc-
cupy the S10-pocket, we decided to explore the potential of moie-
ties that resemble serine, which is a preferred P10 residue in
DenV protease substrates. In fact, remarkable activity was ob-
served with the serine analog 51, but also with the similar com-
pounds 44 and 46. Alkylation of the amide nitrogen does not
have significant influence on the activity (cf. 40 vs 52 and 53).
activity. The necessity of the
a-keto functionality is also exempli-
fied by the -hydroxy derivative 10 and the
a
a
-epoxy derivative
11, both showing reduced inhibitory activity. Methylation is com-
pletely tolerated at the b-position (compd 5): Further work may
aim at substituents in this position to activate the electrophilic trap
and to reach the S2 subsite of the DenV protease. Methylation at
2.4. Selectivity against other serine proteases
the
4). This indicates that the binding does not depend on the ‘Michael’
reactivity in the -position but rather on the reversible formation
c
-position leads to a marginally less active compound (compd
Selectivity data for a subgroup of compounds against the Den-
gue and West Nile virus proteases and thrombin—as a mammalian
serine protease with similar substrate recognition requirements—
are presented in Table 4. The lead compound 1 has limited selectiv-
ity under these conditions. Compound 5 is remarkable because a
very minor structural variation leads to selectivity against DenV
c
of a hemiketal at the ketoamide carbon. The acrylamide 6 was only
marginally active. Compound 11 was synthesized to explore the
potential of an epoxide electrophile, but appears less promising.
To summarize, work on the central moiety confirmed that the
unsaturated ketoamide structure is the essential pharmacophore.
With respect to the shallow binding area,39 the unsaturated
moiety and the resulting planarity and loss of conformational free-
dom is apparently essential for the molecular recognition of this
compound class. The SAR data also indicates that the b-position
may be well suited for the extension of the compounds towards
the S2 and further binding regions.
Variation of the c-aryl moiety (Table 2) revealed the following
picture: While electron donating residues (e.g., methoxy, pyridyl)
promote inhibitory activity, electron withdrawing residues (e.g.,
nitro, cyano, fluoro) diminish activity in most cases.
As shown with derivative 14, incorporation of hydrophobic bulk
is not tolerated. However, the branched isopropyl group 13 is still
accepted by the target. Substitution of the aryl moiety with less
hydrophobic hydrogen-bond-acceptor functions are tolerated to
some extent, as shown in methoxy derivatives 15–17. No clear
preference upon the site of substitution is detectable, also in con-
sideration of the three nitro isomers 26 to 28 or the pyridyl deriv-
atives 29 to 31. In the dimethoxy derivative 18, the increased
electron donating potential results in high inhibition, but this posi-
tive trend can not be maintained in the trimethoxy derivative 19,
probably due to steric constraints. The substituent in 4-position
is preferably a hydrogen bond donor: Phenol 20 has a significantly
higher activity than the methoxy analog 15. The relevance of a
hydrogen bond donor substituent on the aromatic moiety also be-
comes evident in the comparison of the 4-amino and the 4-dimeth-
ylamino analogs 21 and 33. In these cases the activity is abolished
Figure 2. Compounds 1 (magenta) and 32 (green) were covalently docked to the
oxygen of the Ser135 Important residues of the S1 pocket as well as residues of the
catalytic triad (Ser135, His51) are shown as sticks. The S2 pocket is represented as
surface (magenta). The picture was generated with CHIMERA.41
by methylation, while the electronic effects on the
a-ketoamide
moiety should be negligible. A similar interpretation is possible
for the detrimental effect of methylation (36) and acetylation
(37) on the activity of the 3-indole 32, which is among the most
potent compounds of the dataset. The binding of the 3-indole moi-
ety into the S1 subsite is in remarkable agreement with the results
of Yin et al., whose second most active inhibitor was a tetrapeptide
aldehyde with tryptophane in the S1 position that achieved a Ki of
7.5
l
M.40
Table 3 shows the effects of various substituents on the ketoa-
mide nitrogen. We reasoned that the -aryl moiety is probably lo-
cated in or near the S1 pocket of the DenV protease, while the
ketoamide nitrogen substituents are oriented towards the S10
Figure 3. Compounds 18 and 32 inhibit DenV replication in cell-culture. Error bars
c
represent standard deviation from the mean of three independent experiments.
àIndicates cytotoxic activity of compd 18 at 50 and 100
l
M. No cytotoxicity was
a
-
observed for compd 32 at any of the tested concentrations.

C. Steuer et al. / Bioorg. Med. Chem. 19 (2011) 4067–4074
4071
Table 4
Activity of selected compounds against DenV NS2B–NS3, WNV NS2B–NS3 and thrombin. Inhibitor concentration: 50
l
M for DenV and WNV protease, 25 lM for thrombin.
#
Structure
DenV % inhib.
(sd%)
WNV % inhib.
(sd%)
Thrombin % inhib.
(sd%)
26.3
(3.2)
20.0
(6.0)
19.6
(3.4)
1
25.6
(3.4)
0
(2.9)
5.7
(2.4)
5
27.6
(5.2)
20.3
(1.7)
12.1
(3.7)
11
15
16.7
(2.8)
27.3
(5.1)
14.2
(5.3)
36.1
(6.8)
31.0
(0.5)
22.5
(2.1)
18
20
21
24
37.6
(2.4)
31.4
(4.6)
17.3
(3.7)
40.6
(0.9)
39.5
(4.5)
21.1
(2.2)
15.9
(5.2)
6.3
(5.6)
6.7
(4.4)
39.1
(0.7)
40.6
(0.7)
16.9
(1.9)
32
33
13.8
(1.2)
33.9
(1.1)
28.9
(3.7)
31.2
(3.6)
20.0
(6.7)
17.1
(5.2)
41
46
48
35.8
(6.1)
29.8
(2.5)
16.3
(1.8)
0
(4.9)
12.2
(3.2)
5.1
(1.0)
35.1
(3.0)
27.5
(3.2)
12.6
(2.2)
51
protease. Selectivity is slightly improved in compounds 32, 46, and
51, which show higher activity against the flaviviral proteases than
against thrombin. Hydrogen-bond-donors in the aromatic moiety
(20, 21, 32) improve the activity and selectivity for flaviviral prote-
ases in comparison to thrombin. Variations at the amide nitrogen,
in particular those that mimic the preferred P10 residues glycine
and serine (46, but especially 51), lead to improved selectivity pro-
files. In summary, selectivity against flaviviral proteases can be
achieved by several, minor structural variations which may be
combined to yield an optimized structural scaffold.
the compounds, shown here for 1 and 32, occupy the S1 binding re-
gion. Hydrophilic substituents on the aromatic moieties, which have
been shown to increase activity, probably interact with the Tyr150
and Asp129 sidechains at the bottom of the S1 pocket. With respect
to further modifications targeting the S2 pocket, it appears most
promising to extend the analogs either at the cposition of the ketoa-
mide sidechain or at the ortho position on the aromatic ring.
2.6. Antiviral activity in cell-culture
Compounds 2, 18 and 32 were initially screened for antiviral
activity in a reporter-gen cell-culture assay of Dengue replication
(cf. Fig. 3), with 18 and 32 showing activity in that assay. As a neg-
ative control, compound 2, which is inactive against the protease,
was found to be inactive in cell-culture (data not shown). There-
fore, antiviral activity of this class of compounds appears to be
linked to protease inhibition. To further characterize the active
compounds, we treated DenV-2 (wild-type) infected Huh-7 cells
2.5. Molecular modeling
To rationalize the SAR and guide further syntheses, we performed
a covalent docking of selected compounds into the active site of
DenV protease. The search space is limited because of the defined,
covalent bond between the ketoamide moiety and the serine side-
chain. The results are shown in Figure 2. The aromatic moieties of

4072
C. Steuer et al. / Bioorg. Med. Chem. 19 (2011) 4067–4074
with different concentrations of compounds 18 and 32 immedi-
ately after infection and determined infectivity titers 24 h later
(cf. Fig. 3). Furthermore, cytotoxicity of all compounds was deter-
mined in Huh-7-Fluc cells. Compound 32 was not cytotoxic in all
added under ice cooling a solution of potassium hydroxide (1.6 g,
29 mmol) in methanol (20 mL). Precipitation occured during addi-
tion of the base and the mixture was allowed to reach room tem-
perature for 1 h. The mixture was allowed to stand at 8 °C for three
days and then the precipitate was filtered and washed with few
cold MeOH and then diethyl ether. The faint yellow crystals
(2.81 g, 68%) were used without further purification. To a suspen-
sion of the intermediate above (254 mg, 1.18 mmol), HATU
(450 mg, 1.18 mmol), HOAt (160 mg, 1.18 mmol) and tert-butyl
amine hydrochloride (137 mg, 1.25 mmol) in dry DMF (5 mL)
tested concentrations (up to 100 lM), as determined by visual
inspection of the cells as well as by measuring luciferase expres-
sion as a marker of cell viability. Importantly, for compound 32
we determined an inhibitory effect on DenV replication in a
dose-dependent manner, achieving a more than 1000-fold reduc-
tion of virus titers at non-cytotoxic concentrations (cf. Fig. 3).
was added triethyl amine (200 lL, 1.4 mmol) under ice cooling
and the suspension was temporarily sonicated. The mixture was
stirred overnight at room temperature and then concentrated.
The remaining residue was suspended in 5% aqueous citric acid
and extracted with CHCl3 and the solution was dried and purified
by flash chromatography using cyclohexane–ethyl acetate 1:1 to
0:1 as eluent to afford the title compound (251 mg, 92%) as yellow
crystals. NMR 1H d (300 MHz, acetone, 293 K): 8.93 (d, 1H,
J = 2.4 Hz), 8.65 (dd, 1H, J = 4.8, J = 1.7 Hz), 8.25–8.21 (m, 1H),
7.48 (dd, 1H, J = 4.8, J = 7.8 Hz), 7.36 (br s, 1H), 1.43 (s, 9H). NMR
13C d (75 MHz, acetone, 293 K): 188.0 (CO), 162.7 (CO), 153.5
(CH), 152.6 (CH), 144.8 (CH), 136.5 (CH), 132.2 (C), 125.8 (CH),
122.7 (CH), 52.8 (C), 29.5 (3 ꢁ CH3). MS (EI): C13H16N2O2; Mw:
232.1; MS 232.1 M+.
3. Conclusion
In summary, novel inhibitors for DenV protease were identified
and the SAR around the b,c-unsaturated a-ketoamide pharmaco-
phore was investigated. With minor modification, selectivity to-
wards related proteases such as West Nile virus protease and
thrombin could be achieved. Although the derivatives showed only
moderate affinity towards DenV protease in the enzymatic assay,
compound 32 inhibits DenV replication in a cell-culture assay in
a dose-dependent manner, achieving a more than 1000-fold reduc-
tion of virus titers at non-cytotoxic concentrations. In view of the
minimal size of the compounds described here and their remark-
able activity against Dengue virus replication in cell-culture they
appear as valuable starting points for the exploration of further
structural variations to discover and develop novel inhibitors of
flaviviral proteases. Future work should, under the perspectives
of potency and selectivity, aim at modifications on the b-position
and on the amide nitrogen, while retaining the indole moiety. Gi-
ven the antiviral activity and lack of cytotoxicity in cell-culture,
it is also tempting to pursue a medium-scale screening effort in a
cell-culture based assay of virus replication.
4.1.2. (E)-N-(tert-Butyl)-2-oxo-4-(3,4,5-trimethoxyphenyl)but-
3-enamide (19)
A
solution of 3,4,5-trimethoxy benzaldehyde (128 mg,
mol), Cu(OTf)2 (24 mg, 66 mol), trimethyl orthoformate
L, 71
mol) and N-tert-butyl pyruvic amide29 (113 mg,
lmol) in dry CH2Cl2 (2 mL) was degassed under Argon and
652
l
l
(78
l
l
778
heated under microwave irradiation to 100 °C for 10 min. The
brown solution was absorbed on silica and purified by flash chro-
matography using cyclohexane–ethyl acetate 1:0 to 3:1 as eluent
to afford the product (77 mg, 37%) as yellow crystals. NMR 1H d
(300 MHz, acetone, 293 K): 7.79 (d, 1H, J = 15.9 Hz), 7.61 (d, 1H,
J = 15.9 Hz), 7.31 (br s, 1H), 7.14 (s, 1H), 3.92 (s, 6H), 3.79 (s, 3H),
1.43 (s, 9H). NMR 13C d (75 MHz, acetone, 293 K): 188.1 (CO),
163.0 (CO), 155.7 (C), 148.9 (CH), 143.2 (C), 143.2 (C), 131.9, (C),
119.9 (CH), 108.4 (CH), 61.7 (CH3), 57.6 (CH3), 52.7 (C), 29.5
(CH3). MS (EI): C17H23NO5; Mw: 321.2; MS 321.2 M+.
4. Experimental section
4.1. Chemical methods
All chemicals were obtained from Sigma–Aldrich (Germany),
Bachem, (Germany) and Acros (Belgium) and were of analytical
grade. No further purification steps were performed unless indi-
cated. Passerini and Dess–Martin reactions were performed in
dried glassware and under a nitrogen atmosphere. All reactions
were performed with dried solvents (molecular sieve). 1H NMR
spectra were recorded on a Varian (300 MHz) instrument at
300 K in CDCl3, or acetone-d6. Chemical shifts are reported in d val-
ues (ppm) and the residuals of non-deuterated solvents were used
as internal standard (CDCl3: d 7.26 ppm and acetone: d 2.05 ppm).
Signals are described as s, d, t, dd, m, and br for singlet, doublet,
triplet, double-doublet, multiplet, and broad, respectively. Mass
spectra (EI) were measured on a Finnigan MAT 8200 instrument.
Chemical names follow IUPAC nomenclature. Column chromatog-
raphy was performed using silica gel cartridges (KP-Sil) and UV
monitoring, and the reaction progress was determined by thin
layer chromatography analyzes on Merck silica gel plastic plates
60F254. Compounds were visualized by UV and/or treatment with
4.1.3. (E)-N-(tert-Butyl)-4-(4-chlorophenyl)-2-oxobut-3-
enamide (24)
A mixture of (E)-3-(4-chlorophenyl)acrylaldehyde (1 mmol),
tert-butyl isocyanide (1.2 equiv) and SiCl4 (1.1 equiv) in CH2Cl2
(1 mL) was stirred for 4 h at ꢀ78 °C. The reaction was stopped by
the addition of cold aq NaHCO3. After stirring for 2 h, the mixture
was filtered through CeliteÒ. The filtrate was extracted with CH2Cl2
(4 ꢁ 20 mL). The combined organic layers were dried over MgSO4
and concentrated under reduced pressure. The product was used
without any further purification.
A mixture of a-hydroxyamide (1 mmol) and Dess–Martin peri-
odinane (1.15 equiv) in CH2Cl2 (3 mL) was stirred for 2 h at room
temperature. The reaction was stopped by addition of aq NaHCO3.
The organic phase was separated and the aqueous phase was ex-
tracted with diethyl ether (3 ꢁ 20 mL). The combined organic lay-
ers were dried over MgSO4 and concentrated under reduced
pressure. The residue was purified by flash chromatography elut-
ing with cyclohexane–ethyl acetate to give 199 mg of the desired
compound as a yellow solid, yield 75%. NMR 1H d (300 MHz, ace-
tone, 293 K): 7.82 (d, 1H, J = 16.2 Hz), 7.69 (d, 1H, J = 16.3 Hz),
7.83–7.80 (d, 2H), 7.53–7.48 (d, 2H), 7.33 (br s, 1H), 1.42 (s, 9H).
MS (EI): C14H16NO2Cl: Mw: 265.1; MS: 265.1 M+.
a
solution of KMnO4 (2.5 g) and Na2CO3 (12.5 g) in water
(250 mL) and (optionally) heating.
Synthetic and analytical details for the target compounds are gi-
ven in the Supplementary data. Three representative chemical syn-
theses are described below:
4.1.1. (E)-N-(tert-Butyl)-2-oxo-4-(pyridin-3-yl)but-3-enamide
(30)
To a solution of 3-pyridine carboxaldehyde (2.05 g, 19.1 mmol)
and pyruvic acid (1.70 g, 19.3 mmol) in methanol (10 mL) was
The internally quenched DenV NS2B–NS3 protease substrate
Abz-NleKRRS-3-(NO2)Y was synthesized by solid-phase synthesis

C. Steuer et al. / Bioorg. Med. Chem. 19 (2011) 4067–4074
4073
on Rink amide resin according to the Fmoc-protocol. It was purified
by preparative HPLC using an ÄKTA Purifier, GE Germany, with a
RP-18 chromatography column (LobarÒ, Merck, Germany). The
mobile phase consisted of MeOH/0.1% TFA and H2O/0.1% TFA fol-
lowing a gradient of 20–100% MeOH in water with a flow rate of
1.5 mL/min. The internally quenched substrate for the WNV assay
(Abz-GLKRGG-3-(NO2)Y) was synthesized and purified as de-
scribed above. The purity of both substrates was assessed by HPLC
and found to be higher than 95%. The identity was confirmed by
MALDI-ToF-MS.
20.42 The activity of the enzyme was determined as the initial slope
per second and monitored for 10 min.
4.2.4. Tryptophane quenching assay
This was performed as described by Bodenreider et al.12
Dynamic light scattering: Inhibitors were typically diluted from
10 mM stocks in DMSO with filtered 50 mM KPi buffer, pH 7, to a
final concentration of 100 lM. All measurements were performed
at room temperature using a Zetasizer 3000 HS (Malvern Instru-
ments GmbH, Herrenberg, Germany). The laser power and integra-
tion times were comparable for all experiments. The scattering
angle was 173°.
4.2. Biological methods
Expression and purification of DenV protease: The Dengue prote-
ase NS2B–NS3 gene, with the hydrophilic cofactor NS2B connected
to the protease domain NS3pro via a flexible glycine linker, was syn-
thesizedbyGeneart. TheexpressionplasmidpET28a(Novagen)with
the insertedgene was used totransform Escherichia coli BL 21 k (DE3)
cells. Overnight cultures of the transformed cells were grown at
4.2.5. Docking
All calculations were performed on an Intel(R) Core(TM)2 Quad
CPU Q9450 @ 2.66 GHz running open SuSE 11.0 while using GOLD
v5.0 and its graphical interface Hermes 1.4.43 As there is no cata-
lytic active crystal structure of the Dengue NS2B–NS3 protease a
homology model was applied.44 Prior to the docking the embedded
ligand was extracted and hydrogens were added except to the oxy-
gen of the Ser135 where the covalent bond is formed later using
the in Hermes implemented functions. A covalent docking to the
oxygen of Ser135 was performed using GOLD v5.0 with a search ra-
dius of 8 Å and a given substructure common to all inhibitors.
Automatic cavity detection was switched on. Furthermore, the
GoldScore fitness function and default parameters for the genetic
algorithm were applied. During the creation of the covalent bond
between the ligand and the Ser135 a chiral hemiacetal was formed.
For each compound both chiral conformers were use and 10 differ-
ent solutions were calculated and ranked according to their Gold-
Score. The conformer with the better score was used for
visualization.
37 °C in standard LB-medium containing50 lg/mLkanamycin. After
the OD of 0.6–0.8 at 600 nm was reached, the expression was in-
duced by addition of IPTG to a final concentration of 1 mM. The cells
were grown at 30 °C for a further 4 h and then collected by centrifu-
gation at 4500 g. The pellet was resuspended in buffer A (Lysis buf-
fer: 50 mM Tris–HCl pH 7.9, 100 mM NaCl, 5% glycerol and 5 mM
imidazol) and passed through a cell disruptor (One Shot, Constant
Systems). Afterward the solution was centrifuged at 18,500 g and
4 °C for 40 min. The supernatant was then purified by Ni2+-affinity
chromatography. The protein was eluted by increasing the imidazol
concentration from 5 to 250 mM. Stocks of purified protein were
stored at ꢀ70 °C in 100 mM Tris–HCl pH 7.9, 50 mM NaCl and 50%
glycerol. The expression and purification protocol for the WNV pro-
tease was similar to the one given for DenV (cf. Supplementary data
for details).
Antiviral activity in cell-culture: Huh-7 cells were infected with
DenV 2 at an MOI of 1 TCID50 per cell. Four hours after infection,
cell-culture supernatants were supplemented with the test com-
pounds at the indicated concentrations. Control wells were treated
with DMSO alone at a final concentration of 1% (diluent control).
Infectious titers were determined 24 h after infection by a limiting
dilution assay.45
4.2.1. Flourimetric DenV protease assay
The DenV protease assay was performed as described previ-
ously.13 In short, continuous enzymatic assays were performed
on a BMG Labtech Fluostar OPTIMA microtiter fluorescence plate
reader using black 96 well V-bottom plates from Greiner. The exci-
tation wavelength was 320 nm and the emission was monitored at
405 nm. The inhibitor concentration was 50 lM. The inhibitors
were preincubated for 15 min with the enzyme. Afterward, the
reaction was initiated by the addition of the substrate to a final
concentration of 50 lM. The activity of the enzyme was deter-
4.2.6. Cytotoxicity
Huh-7-Fluc cells were treated with the compounds at the indi-
cated concentrations and incubated for 48 h at 37 °C, 5% CO2. Sub-
sequently, cells were lysed and assayed for luciferase activity as a
surrogate for cell viability.
mined as the initial slope per second and monitored for 15 min.
Experiments were performed in triplicate (n = 3) and the experi-
mental values were averaged.
Acknowledgments
C.S. appreciates financial support by the Konrad Adenauer
Foundation. R.B. was supported by the Deutsche Forschungsgeme-
inschaft (SFB544, Teilprojekt B19). Substrate for the DenV-/WNV
protease assay was synthesized by Frederic Zoller. Compound 4
was kindly synthesized by Mrs. Lina Irsheid in the course of a
medicinal chemistry internship. The authors thank Prof. G. Fricker,
Pharmaceutical Technology/Heidelberg University for use of the
DLS equipment.
4.2.2. Flourimetric West Nile virus protease assay
The WNV protease assay was performed in analogy to the DenV
protease assay. Final concentration of the enzyme was 150 nM. The
reaction was initiated by addition of the substrate to a final con-
centration of 50 lM.
4.2.3. Thrombin assay
The thrombin assay was performed as a continuous fluorimetric
assay on a BMG Labtech Fluostar OPTIMA microtiter fluorescence
plate reader. The excitation wavelength was 355 nm and the emis-
sion wavelength was 460 nm. The protease was assayed against
the substrate Boc-Val-Pro-Arg-AMC, (Bachem, Germany). The final
concentrations of the enzyme and substrate were 10 nM and
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2011.05.015.
50
l
M, respectively. The inhibitors were preincubated with the en-
References and notes
zyme for 15 min at a concentration of 25
lM. The cleavage reaction
was initiated by addition of the substrate. The assay buffer con-
1. Leitmeyer, K. C.; Vaughn, D. W.; Watts, D. M.; Salas, R.; Villalobos de Chacon, I.;
Ramos, C.; Rico-Hesse, R. J. Virol. 1999, 73, 4738.
sisted of 50 mM Tris–HCL pH 7.5, 150 mM NaCl and 0.05% Tween

4074
C. Steuer et al. / Bioorg. Med. Chem. 19 (2011) 4067–4074
2. Melino, S.; Paci, M. FEBS J. 2007, 274, 2986.
3. Arakaki, T. L.; Fang, N. X.; Fairlie, D. P.; Young, P. R.; Martin, J. L. Protein
Expression Purif. 2002, 25, 241.
4. Gouvea, I. E.; Izidoro, M. A.; Judice, W. A.; Cezari, M. H.; Caliendo, G.; Santagada,
V.; dos Santos, C. N.; Queiroz, M. H.; Juliano, M. A.; Young, P. R.; Fairlie, D. P.;
Juliano, L. Arch. Biochem. Biophys. 2007, 457, 187.
5. Niyomrattanakit, P.; Yahorava, S.; Mutule, I.; Mutulis, F.; Petrovska, R.; Prusis,
P.; Katzenmeier, G.; Wikberg, J. E. Biochem. J. 2006, 397, 203.
6. Yusof, R.; Clum, S.; Wetzel, M.; Murthy, H. M.; Padmanabhan, R. J. Biol. Chem.
2000, 275, 9963.
21. Bierer, B. E.; Somers, P. K.; Wandless, T. J.; Burakoff, S. J.; Schreiber, S. L. Science
1990, 250, 556.
22. Maryanoff, B. E.; Qiu, X.; Padmanabhan, K. P.; Tulinsky, A.; Almond, H. R., Jr.;
Andrade-Gordon, P.; Greco, M. N.; Kauffman, J. A.; Nicolaou, K. C.; Liu, A. Proc.
Natl. Acad. Sci. U.S.A. 1993, 90, 8048.
23. Varshney, V.; Mishra, N. N.; Shukla, P. K.; Sahu, D. P. Bioorg. Med. Chem. Lett.
2009, 19, 6810.
24. Fawzy, N. M.; Swelam, S. A.; Farid, M. A. Biotechnology 2008, 7, 43.
25. Hänsel, R.; Leuschke, A.; Bohlmann, F. Planta Med. 1980, 40, 161.
26. Lange, J. H. M.; Zilaout, H.; Van Vliet, B. J. WO2009037244, 2009.
27. Srivastava, B. K.; Joharapurkar, A.; Raval, S.; Patel, J. Z.; Soni, R.; Raval, P.; Gite,
A.; Goswami, A.; Sadhwani, N.; Gandhi, N.; Patel, H.; Mishra, B.; Solanki, M.;
Pandey, B.; Jain, M. R.; Patel, P. R. J. Med. Chem. 2007, 50, 5951.
28. Allais, C.; Constantieux, T.; Rodriguez, J. Synthesis 2009, 2523.
29. Solladiecavallo, A.; Bencheqroun, M. J. Org. Chem. 1992, 57, 5831.
30. Weinstein, D. S.; Yang, B. V.; Kim, S.-H.; Vaccaro, W.; Sheppeck, J.; Gilmore, J.
WO2005072132, 2005.
31. Dujardin, G.; Leconte, S.; Benard, A.; Brown, E. Synlett 2001, 147.
32. Denmark, S. E.; Fan, Y. J. Am. Chem. Soc. 2003, 125, 7825.
33. Speicher, A.; Bomm, V.; Eicher, T. J. Prakt. Chem. 1996, 338, 588.
34. Corey, E. J.; Chaykovsky, M. J. Am. Chem. Soc. 1965, 87, 1353.
35. Schauer, D. J.; Helquist, P. Synthesis 2006, 3654.
7. Erbel, P.; Schiering, N.; D’Arcy, A.; Renatus, M.; Kroemer, M.; Lim, S. P.; Yin, Z.;
Keller, T. H.; Vasudevan, S. G.; Hommel, U. Nat. Struct. Mol. Biol. 2006, 13, 372.
8. Niyomrattanakit,
P.;
Winoyanuwattikun,
P.;
Chanprapaph,
S.;
Angsuthanasombat, C.; Panyim, S.; Katzenmeier, G. J. Virol. 2004, 78, 13708.
9. Lescar, J.; Luo, D.; Xu, T.; Sampath, A.; Lim, S. P.; Canard, B.; Vasudevan, S. G.
Antiviral Res. 2008, 80, 94.
10. Yin, Z.; Patel, S. J.; Wang, W. L.; Wang, G.; Chan, W. L.; Rao, K. R.; Alam, J.;
Jeyaraj, D. A.; Ngew, X.; Patel, V.; Beer, D.; Lim, S. P.; Vasudevan, S. G.; Keller, T.
H. Bioorg. Med. Chem. Lett. 2006, 16, 36.
11. Stoermer, M. J.; Chappell, K. J.; Liebscher, S.; Jensen, C. M.; Gan, C. H.; Gupta, P.
K.; Xu, W. J.; Young, P. R.; Fairlie, D. P. J. Med. Chem. 2008, 51, 5714.
12. Bodenreider, C.; Beer, D.; Keller, T. H.; Sonntag, S.; Wen, D.; Yap, L.; Yau, Y. H.;
Shochat, S. G.; Huang, D.; Zhou, T.; Caflisch, A.; Su, X. C.; Ozawa, K.; Otting, G.;
Vasudevan, S. G.; Lescar, J.; Lim, S. P. Anal. Biochem. 2009, 395, 195.
13. Steuer, C.; Heinonen, K. H.; Kattner, L.; Klein, C. D. J. Biomol. Screening 2009, 14,
1102.
36. Hopkins, A. L.; Groom, C. R.; Alex, A. Drug Discovery Today 2004, 9, 430.
37. Perola, E. J. Med. Chem. 2010, 53, 2986.
38. Groll, M.; Schellenberg, B.; Bachmann, A. S.; Archer, C. R.; Huber, R.; Powell, T.
K.; Lindow, S.; Kaiser, M.; Dudler, R. Nature 2008, 452, 755.
39. Kwong, A. HCV Protease Inhibitors: Activity and Resistance. In Conference on
Retroviruses and Opportunistic Infections (CROI), Denver, 2006.
40. Yin, Z.; Patel, S. J.; Wang, W. L.; Chan, W. L.; Ranga Rao, K. R.; Wang, G.; Ngew,
X.; Patel, V.; Beer, D.; Knox, J. E.; Ma, N. L.; Ehrhardt, C.; Lim, S. P.; Vasudevan, S.
G.; Keller, T. H. Bioorg. Med. Chem. Lett. 2006, 16, 40.
41. Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.;
Meng, E. C.; Ferrin, T. E. J. Comput. Chem. 2004, 25, 1605.
42. Diamond, S. L.; Thrombin 1536 HTS, PubChem Bio-Assay ID 1046. Available
from: <http://pubchem.ncbi.nlm.nih.gov/a ssay/assay.cgi?aid=1046&loc=ea_
ras>.
43. Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R. J. Mol. Biol. 1997, 267,
727.
44. Wichapong, K.; Pianwanit, S.; Sippl, W.; Kokpol, S. J. Mol. Recognit. 2010, 23,
283.
45. Lindenbach, B. D.; Evans, M. J.; Syder, A. J.; Wolk, B.; Tellinghuisen, T. L.; Liu, C.
C.; Maruyama, T.; Hynes, R. O.; Burton, D. R.; McKeating, J. A.; Rice, C. M. Science
2005, 309, 623.
14. Peese, K. Drug Discovery Today 2010, 15, 406.
15. Nelson, D. R. Hepatology 2009, 50, 997.
16. Han, W.; Hu, Z.; Jiang, X.; Decicco, C. P. Bioorg. Med. Chem. Lett. 2000, 10,
711.
17. Bennett, J. M.; Campbell, A. D.; Campbell, A. J.; Carr, M. G.; Dunsdon, R. M.;
Greening, J. R.; Hurst, D. N.; Jennings, N. S.; Jones, P. S.; Jordan, S.; Kay, P. B.;
O’Brien, M. A.; King-Underwood, J.; Raynham, T. M.; Wilkinson, C. S.;
Wilkinson, T. C.; Wilson, F. X. Bioorg. Med. Chem. Lett. 2001, 11, 355.
18. Arasappan, A.; Njoroge, F. G.; Parekh, T. N.; Yang, X.; Pichardo, J.; Butkiewicz,
N.; Prongay, A.; Yao, N.; Girijavallabhan, V. Bioorg. Med. Chem. Lett. 2004, 14,
5751.
19. Bogen, S. L.; Pan, W.; Ruan, S.; Nair, L. G.; Arasappan, A.; Bennett, F.; Chen, K. X.;
Jao, E.; Venkatraman, S.; Vibulbhan, B.; Liu, R.; Cheng, K. C.; Guo, Z.; Tong, X.;
Saksena, A. K.; Girijavallabhan, V.; Njoroge, F. G. J. Med. Chem. 2009, 52, 3679.
20. Chen, J. C.; Uang, B. J.; Lyu, P. C.; Chang, J. Y.; Liu, K. J.; Kuo, C. C.; Hsieh, H. P.;
Wang, H. C.; Cheng, C. S.; Chang, Y. H.; Chang, M. D.; Chang, W. S.; Lin, C. C. J.
Med. Chem. 2010, 53, 4545.