Synthesis, biological evaluation and molecular modeling studies of quinolonyl diketo acid derivatives: New structural insight into the HIV-1 integrase inhibition

Article abstract of DOI:10.1016/j.ejmech.2011.02.028

New quinolonyl diketo acid compounds bearing various substituents at position 6 of the quinolone scaffold were designed and synthesized as potential HIV-1 integrase inhibitors. These new compounds were evaluated for their antiviral and anti-integrase acti

Full text of DOI:10.1016/j.ejmech.2011.02.028

European Journal of Medicinal Chemistry 46 (2011) 1749e1756
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Synthesis, biological evaluation and molecular modeling studies of quinolonyl
diketo acid derivatives: New structural insight into the HIV-1 integrase inhibition
, ,
,
Pierre Vandurm a 1, Allan Guiguen b 1, Christine Cauvin a, Benoît Georges a, Kiet Le Van a,
*
Catherine Michaux a, Christelle Cardona b, Gladys Mbemba c, Jean-François Mouscadet c, László Hevesi a
**
,
,
1
,
,
,
,1
***
Carine Van Lint b 1, Johan Wouters a
a Département de Chimie, Facultés Universitaires Notre-Dame de la Paix (FUNDP), B-Namur 5000, Belgium
b Laboratoire de Virologie Moléculaire, Institut de Biologie et de Médecine Moléculaires, Université Libre de Bruxelles (ULB), Gosselies B-6041, Belgium
c LBPA, CNRS, Ecole Normale Supérieure de Cachan, Cachan F-94235, France
a r t i c l e i n f o
a b s t r a c t
Article history:
New quinolonyl diketo acid compounds bearing various substituents at position 6 of the quinolone
scaffold were designed and synthesized as potential HIV-1 integrase inhibitors. These new compounds
were evaluated for their antiviral and anti-integrase activity and showed inhibitory potency similar to
that of 6-bromide analog 2. Molecular modeling and docking studies were performed to rationalize these
data and to provide a detailed understanding of the mechanism of inhibition for this class of compounds.
Ó 2011 Elsevier Masson SAS. All rights reserved.
Received 25 November 2010
Received in revised form
10 February 2011
Accepted 13 February 2011
Available online 22 February 2011
Keywords:
Quinolonyl diketo acid
Antiviral activity
HIV-1 IN
Anti-integrase activity
Molecular modeling
Docking
1. Introduction
structures of the isolated CCDs available to date [8,9] show a single
Mg2þ ion coordinated by the two aspartate residues (D64 and
Integration of viral DNA into the host-cell chromosome is an
obligatory step in the human immunodeficiency virus (HIV) repli-
cation cycle [1]. This process is catalyzed by the virally encoded
integrase (IN) through two sequential reactions [2,3]. The first one,
named 30-processing (30-P), corresponds to an endonucleolytic
cleavage at each 30 ends of double-stranded viral DNA extremities.
The second step, strand transfer (ST), inserts the viral DNA into the
host-cell DNA through a transesterification mechanism. HIV-1 IN
belongs to a large family of polynucleotidyl transferase enzymes
that include transposases, polymerases and other retroviral inte-
grases [4e6]. The catalytic core domain (CCD) contains a conserved
DDE motif which consists of two aspartates (D64 and D116) and one
glutamate (E152), essential for enzymatic activity [7]. Crystal
D116). Nevertheless, similar to other polynucleotidyl transferases,
a second metal can be most likely coordinated by E152 (with D64)
once HIV-1 IN binds to its DNA substrate [6]. Inhibitors of IN appear
to be particularly promising since, unlike protease and reverse
transcriptase, this enzyme does not have direct human homologs
[10,11]. Among all reported IN inhibitors [12e14], the b-diketo acid
(DKA) class of compounds has emerged as the most potent and the
most promising. Raltegravir is the first approved IN inhibitor
whereas Elvitegravir and GSK364735 have reached clinical devel-
opment (Fig. 1) [15e19]. Like other well-known DKA inhibitors,
these one share two common structural chemotypes essential for
the anti-IN activity: a diketo acid chain able to interact with Mg2þ
metal ions (marked in bold) and a properly oriented hydrophobic
benzyl moiety (marked in dashed box). They selectively inhibit ST
reaction, suggesting that they bind at the IN/DNA interface, acting
as “interfacial inhibitors” [20,21]. Despite the lack of detailed
structural information about HIV-1 IN/DNA interactions, this
speculative mechanism of action tends to be validated by the recent
X-ray crystal structure of IN from the prototype foamy virus (PFV-1
IN) in complex with its cognate viral DNA and ST inhibitors [22,23].
* Corresponding author. Tel.: þ32 81 72 45 69.
** Corresponding author. Tel.: þ32 2 650 98 07; fax: þ32 2 650 98 00.
*** Corresponding author. Tel.: þ32 81 72 45 50; fax: þ32 81 72 54 40.
E-mail addresses: pierre.vandurm@fundp.ac.be (P. Vandurm), cvlint@ulb.ac.be
(C. Van Lint), johan.wouters@fundp.ac.be (J. Wouters).
1
These authors equally contributed to this work.
0223-5234/$ e see front matter Ó 2011 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2011.02.028

1750
P. Vandurm et al. / European Journal of Medicinal Chemistry 46 (2011) 1749e1756
Schemes 1 and 2. The key intermediate 3-acetyl-6-hydroxy-4(1H)-
O
OH
F
O
O
quinolinone 8 was obtained by reaction of commercially available
4-hydroxyaniline with ethylethoxymethyleneacetoacetate and
subsequent thermal condensation in diphenyl ether. 8 was then Ne
and Oealkylated with 4-fluorobenzyl bromide in alkaline medium
(K2CO3) to give compound 9. The formation of N and O-benzyl
derivative 9 was indicated by the presence of a singlet at 5,16 and
5,62 ppm for the NeCH2 and OeCH2. (E)-ethyl [3-acetyl-1-(4-flu-
orophenyl)methyl-4(1H)-quinolinon-6-yl]acrylate 11 was obtained
by palladium-catalyzed Heck reaction of 10 (previously synthesized
and described [24]) with acrylic acid ethyl ester. The presence of the
E isomeric form was confirmed by 1H NMR spectra, showing the
appearance of two doublets at 6,51 and 7,73 ppm characteristic of
O
N
H
Cl
OH
N
N
F
O
N
H
N
H
N
N
O
OH
Raltegravir
Elvitegravir
OH O
F
N
OH
N
H
a,b
-unsaturated olefinic protons, with coupling constant 16 Hz. The
diketo ester derivatives and were obtained by Claisen
3
5
N
O
condensation of compounds 9 and 11, respectively, with diethyl
oxalate in the presence of sodium ethoxide. These compounds were
in turn hydrolyzed by saponification reaction to give DKA deriva-
tives 4 and 6. Diketo ester/acid compounds were characterized by
IR and 1H NMR as well as by mass spectrometry. The presence of the
keto-enolic ester/acid chain in synthesized compounds 3e6 was
detected by the appearance of a strong singlet around 8 ppm
characteristic of an enolic proton and three characteristic intense
bands in the region around 1600e1750 cmꢀ1 due to nC¼O carboxylic
GSK364735
Fig. 1. Representative integrase strand transfer inhibitors (INSTIs). The two common
structural chemotypes include a diketo acid moiety in bold and a hydrophobic benzyl
moiety in dashed box.
Based on Di Santo’s quinolone series, we have reported in
a previous work [24] two compounds (compounds 1 and 2, Fig. 2) as
potential HIV-1 integrase inhibitors. Docking studies based on the
HIV-1 IN CCD crystal structure [9] allowed to propose two binding
modes for compound 2 pointing out the possibility to substitute the
bromine atom at C-6. In order to gain new insights into the structural
requirements for anti-IN activity as well as to elucidate the binding
mode of this series of compounds, we here explored the effect of
chemical modifications at position 6 of the quinolone scaffold. In this
context, we synthesizednew quinolonederivatives 3e6(Fig. 2)bythe
replacement of bromine atom of 2 with bulky hydrophilic or hydro-
phobic substituents. All these compounds were evaluated for their
anti-integrase and antiviral activity and structureeactivity relation-
ships were established. In addition, a new HIV-1 IN/DNA model was
built considering two Mg2þ ions octahedrally coordinated. Docking
studies were then performed to predict the binding mode of quino-
lone compounds and to rationalize the observed structureeactivity
relationships. Based on these results, a detailed mechanism of inhi-
bition for quinolonyl DKA compounds is proposed.
acid, nC¼O quinolone ring and coupled vibration band (nC¼O
þ nC¼C)
of keto-enol group.
2.2. Biological evaluation
Diketo ester and acid derivatives 3e6 were tested in vitro for
their ability to inhibit the 30-P and the ST reaction in the presence of
Mg2þ using a gel-based assay [27]. IC50 values were generated from
duplicate experiments and calculated using dose response curves
(Table 1). As already observed for quinolonyl DKA series [24e26],
acidic derivatives were more potent than the corresponding esters,
with a high selectivity against ST. The replacement of bromine atom
at C-6 by an anionic (acrylate) or a hydrophobic (4-fluoro-benzy-
loxy) group does not lead to a significant improvement of HIV-1 IN
inhibition. Compounds 4 and 6 showed IC50 values against ST (ST
IC50 ¼ 0.03e0.05
mM) similar to the one of 6-bromide analog 2 (ST
IC50 ¼ 0.03 M) with a moderate increase in the selectivity for
m
2. Results and discussion
ST versus 30-P. These results suggest that the size of substituents at
C-6 on quinolone scaffold affect the ability of the inhibitor to
discriminate between the active site of the DNA-bound or
-unbound form of HIV-1 IN. Moreover, once the inhibitor is bound
to the IN/DNA complex, substituents at C-6 do not make significant
interaction.
2.1. Chemistry
The title compounds 3e6 were synthesized according to Di
Santo’s procedure [25,26] with modifications as presented in
O
O
OH
O
O
OH
Br
OX
R6
OX
O
O
F
F
3: X= Et; R6 = O-CH2(4-F-Ph)
4: X= H; R6 = O-CH2(4-F-Ph)
5: X= Et; R6 = HC=CHCOOEt
6: X=H; R6 = HC=CHCOOH
1: X = Et
2: X = H
Fig. 2. Structure of 6-substituted quinolone compounds. Bromine atom of 2 was substituted by acrylic ester/acid and 4-fluoro-benzyloxy to generate new inhibitors 3e6.

P. Vandurm et al. / European Journal of Medicinal Chemistry 46 (2011) 1749e1756
1751
F
O
N
O
HO
O
O
O
HO
HO
a
b
c
N
H
O
NH2
N
H
O
O
F
9
7
8
d
F
F
O
O
OH
O
N
O
OH
O
O
O
OH
e,f
O
O
N
F
F
4
3
Scheme 1. Reagents and conditions: (a) ethylethoxymethyleneacetoacetate, 120 ꢁC, 10 min; (b) Ph2O, reflux 1.30 h; (c) 4-Fluorobenzyl bromide, K2CO3, DMF, 100 ꢁC, 5 h; (d) diethyl
oxalate, C2H5ONa, THF, 20 h, HCl, rt; (e) 1 N NaOH, THF/MeOH (1:1), 3 h; (f) HCl 1 N, rt.
The activity of the title quinolone compounds against wild-type
HIV-1 was also established by determining their ability to inhibit
viral infection of TZM-bl cells, as described elsewhere [28]. Cyto-
toxicity (CC50) and antiviral activities (EC50) of compounds 3e6 are
presented in Table 1. All tested compounds showed antiviral
activity at non-cytotoxic concentrations. All derivatives were
characterized by low cytotoxicity with CC50 values up to 1 mM. The
antiviral activities are correlated with the anti-integrase activities.
Indeed, acidic derivatives 4 and 6 were more potent than the cor-
responding esters 3,5 and showed EC50 values similar to that of
compound 2.
IN/DNA complex in its active and INSTI-inhibited forms have been
reported [29,30]. Nevertheless, they did not consider the octahedral
inner sphere coordination environments of the two active-site Mg2þ
ions, in particular explicit water molecules which could be involved
in the catalytic reaction mechanism. Interestingly, by superimposing
the Ca atoms of the Tn5 transposase (PDB code 1MUS) [31], PFV IN
(PDB code 3L2S) [22] and HIV-1 IN CCD (PDB code 1BL3C) [9] DDE
active-site residues, water molecules present in both Tn5 trans-
posase and HIV-1 IN crystal structures occupy similar positions and
complete the octahedral geometry of the two divalent metal ions
from PFV crystal structure (Fig. 3a). The 50 phosphate group of Tn5
transposon superimposes with both bridging (w4) and apical (w1)
water molecules from HIV-1 IN CCD. This superimposition reveals
a plausible conformation of a scissile phosphodiester group from the
target DNA interacting with the two divalent ions by displacing two
water molecules. In this way, the scissile phosphodiester group
might be properly positioned for the attack of the 30-OH nucleophilic
group of the viral DNA. These observations prompted us to build
a model of the HIV-1 IN/DNA complex in its INSTI-inhibited form by
2.3. Molecular modeling studies
2.3.1. HIV-1 IN/DNA complex building
Due to the lack of HIV-1 IN/DNA experimental structures, the
building of a reliable model is essential for predicting docking of
INSTIs. Based on the recent crystal structure of full-length foamy
virus IN in complex with its cognate DNA and INSTIs, models of HIV-1
O
O
O
N
O
O
N
O
O
N
O
OH
O
O
Br
O
a
b
O
F
F
F
10
11
5
c, d
O
O
O
OH
HO
OH
O
N
F
6
Scheme 2. Reagents and conditions: (a) Pd(OAc)2 cat., PPh3, NEt3, acrylic acid ethyl ester, DMF, 80 ꢁC, 48 h; (b) diethyl oxalate, C2H5ONa, THF, 20 h, HCl, rt; (c) 1 N NaOH, THF/MeOH
(1:1), 3 h; (d) HCl 1 N, rt.

1752
P. Vandurm et al. / European Journal of Medicinal Chemistry 46 (2011) 1749e1756
Table 1
Cytotoxicity, antiviral and anti-integrase activities of derivatives 3e6.
Compd
R
X
anti-IN
activity, IC50
antiviral activity
a
SId
b
c
30-P
ST
EC50
CC50
3
4
5
OeCH2ꢀ4ꢀFꢀPh
OeCH2ꢀ4ꢀFꢀPh
HC ¼ CHCOOEt
HC ¼ CHCOOH
Br
Et
H
Et
H
>300
3
>300
8
>300
0.05
>300
0.03
13
1
>140
5
4
>1000
>1000
>1000
>1000
>1000
77
1000
7
200
250
6
2e
H
1.6
0.03
a
Inhibitory concentration 50% (
m
M) determined from dose response curves.
M) (TZM-bl cell line [26]).
Cytotoxic concentration 50% ( M) (TZM-bl cells).
b
c
Ex vivo anti-HIV activity (
m
m
d
e
Selectivity index ¼ CC50/EC50
.
Literature data: see ref [23].
superimposing the Ca atoms of HIV-1 IN CCD (PDB code 1BL3C) onto
those of the Elvitegravir-bound PFV crystal structure (PDB code
3L2U) [22] and by considering explicit water molecules (see exper-
imental section for details). This new model was subsequently
optimized by a minimization procedure and used as template to
perform docking studies.
2.3.2. Docking studies
Docking of the best DKA quinolone compounds (2, 4 and 6) in
the newly built model was performed with the Gold Software [32].
All these compounds showed a similar interfacial-binding mode in
which the DKA moiety chelates the two Mg2þ ions (see docking
result of 2, Fig. 3b). In this orientation, the w1 (bridging) and w4
(apical) coordinated water molecules were totally displaced by
carboxylate and enolate oxygen atoms, respectively, whereas the
keto oxygen atom replaces the 30-OH extremity of viral DNA. As
a result, the displaced 30-adenosine terminal base (A17) was
involved in a p-stacking interaction with the quinolone ring in such
a way that bromine atom at C-6 pointed out to the solvent-exposed
surface. The p-F-benzyl group fitted within a tight pocket formed
by cytosine 16 (C16), guanine 4 (G4) and two catalytic loop residues
P145 and Q146. Fluorobenzyl group made
p-stacking interaction
with C16 taking the space originally occupied by A17. These docking
results were in agreement with the structureeactivity relationships
that we observed above (see Section 2.2). Indeed, the chelation of
the carboxylate group to Mg2þ ion could explain the stronger
inhibitory potency of acidic compounds compared to ester
compounds. Moreover, the orientation of the substituent at posi-
tion 6 of the quinolone scaffold towards the solvent-exposed
surface could also explain the unmodified potency of 6-substituted
quinolone derivatives.
These molecular modeling studies allow us to propose a mech-
anism of inhibition by DKA compounds (Fig. 4). After the processing
of viral DNA, the active site of HIV-1 IN adopts an active confor-
mation in which two Mg2þ ions are octahedrally coordinated by the
DDE motif, the 30-OH extremity and water molecules. This active
conformation facilitates the binding of a scissile phosphodiester
group from target DNA which displaces bridging (w1) and apical
(w4) water molecules from the active site, forming the ST complex.
Once the inhibitor binds in the active site, the chelating moiety
(DKA chain) interacts with the two Mg2þ ions by displacing the
30-OH extremity and the two water molecules originally occupied
by the phosphodiester group. The newly formed octahedral
complex is then stabilized by the properly oriented hydrophobic
benzyl moiety by replacing the 30 A17 displaced from the active site
to form a tighter complex. In this way, the inhibitor prevents the
formation of the ST complex rendering it catalytically inactive. This
mechanism of action is in agreement with the two-step mechanism
of binding of INSTIs proposed by Garvey et al. [33].
Fig. 3. a) Superimposition of the active sites of HIV-1 IN, Tn5 transposase and PFV IN
based on the positions of C- atoms of the active-site DDE residues. Side chains of DDE
a
motif and terminal nucleotides are shown as sticks. Divalent metal ions (labeled A and
B) and waters molecules (labeled w) are shown as spheres. The arrow indicates the 30-
OH end attacking the scissile phosphate of the target DNA mimicked by the 50-phos-
phate end of Tn5 transposon. b) Docking result of compound 2 (yellow) into the HIV-1
IN/DNA model. The two Mg2þ ions and coordinated water molecules (w) are shown as
green and red spheres, respectively. The nontransferred (orange) and reactive
(magenta) DNA strand are shown as cartoon. Catalytic core domain is shown as grey
cartoon with main amino acids as sticks. Note that the two Mg2þ ions are octahedrally
coordinated to DDE motif, water molecules and the diketo acid chain of the quinolone
inhibitor. (For interpretation of the references to colour in this figure legend, the reader
is referred to the web version of this article.)
3. Conclusion
In order to elucidate the binding mode of DKA quinolonyl series
of compounds, new derivatives were synthesized by replacing the
6-bromo atom of compound 2 by an acrylate and a p-fluoro-ben-
zyloxy group. These compounds were evaluated for their enzymatic

P. Vandurm et al. / European Journal of Medicinal Chemistry 46 (2011) 1749e1756
1753
Fig. 4. Mechanism of inhibition for quinolonyl DKA compounds.
and antiviral activity. The introduction of both hydrophilic and
4.1.1. Synthesis of 2-[[(4-phenol)amino]methylene]-3-oxobutanoic
acid ethyl ester (7)
hydrophobic substituents at C-6 did not have any effect on the
activity in comparison with the activity observed with compound 2.
Structural analysis of the catalytic core domain of HIV-1 IN, Tn5
transposase and PFV IN led us to build a new HIV-1 IN/DNA model
taking into account explicit water molecules to complete the
octahedral geometry of the two active-site magnesium ions. Our
further docking studies including water molecules predicted
a binding mode in agreement with the biological activities we
observed. Molecular modeling studies allowed to propose
a detailed mechanism of inhibition for quinolonyl DKA compounds.
We provided new structural insight into the mechanism of action of
INSTIs which could be useful for the design of new HIV-1 IN
inhibitors.
An equimolar mixture of ethylethoxymethyleneacetoacetate
(4.6 g, 23 mmol) and 4-hydroxyaniline (2.5 g, 23 mmol) was stirred
at 120 ꢁC for 10 min. After cooling, the formed precipitate was
filtered and washed with n-hexane to obtain pure compound 7
used in the next reaction without further purification. Yellow solid,
yield: 55%, mp: 113 ꢁC. 1H NMR (DMSO-d6, 400 MHz):
d 1.22 (t, 3H,
CH2CH3), 2.36 (s, 3H, CH3), 2.54 (s, 3H, CH3), 4.10 (q, 2H, CH2CH3),
6.76 (d, 2H, J ¼ 8.8 Hz, benzene H), 7.18 (d, 2H, J ¼ 8.8 Hz, benzene
H), 8.29 (d, 1H, J ¼ 13.7 Hz, C]CH), 9.54 (s, 1H, OH), 12.54 (d, 1H,
J ¼ 13.7 Hz, NH).
4.1.2. Synthesis of 3-acetyl-6-hydroxy-4(1H)-quinolinone (8)
4.2 g (17 mmol) was added portion wise to 200 ml of boiling
diphenyl ether previously heated at 250 ꢁC. The solution was kept
at 250 ꢁC for 1 h and 30 min. After cooling, the formed precipitate
was collected. The crude product was purified by chromatography
on silica gel column (EtOH/CH2Cl2 as eluant) to provide pure
derivative 8. Brown solid, yield: 46%, mp: >250 ꢁC. 1H NMR (DMSO-
4. Experimental
4.1. Chemistry
d6, 400 MHz):
d
2.56 (s, 1H, CH3), 7.15 (dd, 1H, J7,5 ¼ 2.6 Hz,
All the reagents and solvents employed were used without
further purification. 1H NMR and and 13C NMR spectra were
J7,8 ¼ 8.7 Hz, C7-H quinolinone), 7.46 (d, 1H, J8,7 ¼ 8.7 Hz, C8-H
quinolinone), 7.51 (d, 1H, J5,7 ¼ 2.6 Hz, C5-H quinolinone), 8.37 (s,
1H, C2-H quinolinone), 9.91 (bs, 1H, OH).
recorded on a Jeol JNM EX-400 spectrometer. Chemical shifts (
d
scale) are reported in parts per million (ppm) downfield from tet-
ramethylsilane (TMS) used as an internal standard. Infrared (IR)
spectra were recorded to the solid state on a PerkineElmer Spec-
trum 65. Melting points were determined with a TOTOLLI-BUCHI
Melting Point B-545 apparatus. Analytical thin-layer chromatog-
raphy (TLC) was done on Merck silica gel 60 F254 plates. Mass
spectra were recorded on an 1100 series MSD Trap spectrometer
equipped with an electron spray ionization (ESI) source.
4.1.3. Synthesis of 3-acetyl-1-((4-fluorophenyl)methyl)-6-(4-
fluorobenzyloxy)-4(1H) quinolinone (9)
4-Fluorophenylmethyl bromide (1.7 g, 9 mmol) and anhydrous
K2CO3 (1.3 g, 9 mmol) was added to a mixture of 8 in dry DMF
(30 ml). The resulting suspension was stirred at 100 ꢁC for 5 h. After
the mixture was cooled, water (50 ml) was added and the

1754
P. Vandurm et al. / European Journal of Medicinal Chemistry 46 (2011) 1749e1756
precipitate was filtered, washed with petroleum ether and purified
by chromatography on silica gel column (AcOEt/petroleum ether
4/6 and then AcOEt as eluant). Yellow solid, yield: 73%, mp: 219 ꢁC.
CH2CH3), 4.40 (q, 2H, CH2CH3), 5.42 (s, 2H, CH2), 6.52 (d, 1H,
Jtrans ¼ 16 Hz, CH]CH), 7.04e7.20 (m, 4H, AreH), 7.36 (d, 1H,
J8,7 ¼ 8.7 Hz, C8-H quinolinone), 7.69e7.76 (m, 2H, CH]CH and C7-
H quinolinone), 8.11 (s, 1H, C3-H butenoate), 8.67 (d, 1H, J5,7 ¼ 2 Hz,
C5-H quinolinone), 8.72 (s, 1H, C2-H quinolinone). 13C NMR (CDCl3,
400 MHz): 14.24 (CH3), 14.37 (CH3), 57.46 (NeCH2), 60.87 (CH2),
62.51 (CH2), 102.51 (CH), 115.80 (C), 116.78 (d, J ¼ 22.0 Hz, 2CH),
117.54 (CH), 120.37 (CH), 127.93 (CH), 128.10 (d, J ¼ 8.6 Hz, 2CH),
129.29 (C), 129.39 (d, J ¼ 2.8 Hz, C), 131.86 (C), 132.23 (CH), 139.62
(C), 142.36 (CH), 148.68 (CH), 162.45 (C), 162.90 (d, J ¼ 249.20 Hz, C),
166.58 (C), 169.58 (C), 174.96 (C), 187.55 (C). LC-MS: (M þ H)þ 494.0.
1H NMR (DMSO-d6, 400 MHz):
d 2.61 (s, 3H, CH3), 5.16 (s, 2H, CH2),
5.62 (s, 2H, CH2), 7.1e7.3 (m, 6H, AreH), 7.35 (dd, 1H, J7,5 ¼ 3 Hz,
J7,8 ¼ 9.4 Hz, C7-H quinolinone), 7.44e7.52 (m, 2H, AreH), 7.61 (d,
1H, J8,7 ¼ 9.4 Hz, C8-H quinolinone), 8.65 (d, 1H, J5,7 ¼ 3 Hz, C5-H
quinolinone), 8.78 (s, 1H, C2-H quinolinone).
4.1.4. Synthesis of (E)-ethyl [3-acetyl-1-(4-fluorophenyl)methyl-4
(1H)-quinolinon-6-yl] acrylate (11)
3-acetyl-6-bromo-1-((4-fluorophenyl)methyl)-4(1H)-quinoli-
none (10) (0.450 g, 1.2 mmol) was added to a mixture of acrylic
acid ethyl ester (0.202 g, 2.4 mmol), Pd(OAc)2 (0.013 g,
0.06 mmol), PPh3 (0.061 g, 0.24 mmol) and NEt3 (0.134 g,
1.2 mmol) in 3 ml of DMF. The reaction mixture was stirred under
argon for 48 h at 80 ꢁC. The reaction was cooled to room
temperature and extracted with ethyl acetate. The organic phases
were washed four times with brine (10 ml) and dried with MgSO4.
Upon removal of the solvent, the residue obtained was purified by
chromatography on silica gel column (ACOEt/petroleum ether 3/7
and then 1/1 as eluant) to afford pure derivative 11. Yellow solid,
4.1.6. General synthesis procedure for diketo acids (4 and 6)
A mixture of the appropriate ester 3 or 5 (1.3 mmol) and 1N
NaOH (12 ml) in THF/methanol (1:1, 15 ml) was stirred at room
temperature for 3 h under nitrogen atmosphere, poured onto
crushed ice and treated with 1N HCl to reach pH 2. The formed
precipitate was filtered off and successively washed with water,
ethanol and petroleum ether to give pure acids 4 and 6.
4.1.6.1. 4-[6-(4-fluorobenzyloxy)-1-(4-fluorophenyl)methyl-4(1H)-
quinolinon-3-yl]-2-hydroxy-4-oxo-2-butenoic acid (4). Yellow solid,
yield: 79%, mp: 194 ꢁC. IR ymax (KBr/cmꢀ1): 1731 (C]O carboxylic
acid), 1718 (C]O quinolinone), 1612 (C]O þ C]C keto-enol). 1H
yield: 62%, mp: 220 ꢁC. 1H NMR (CDCl3, 400 MHz):
d 1.33 (t, 3H,
CH2CH3), 2.82 (s, 3H, CH3), 4.26 (q, 2H, CH3CH2), 5.38 (s, 2H, CH2),
6.51 (d, 1H, Jtrans ¼ 16 Hz, CH]CH), 7.02e7.18 (m, 4H, AreH), 7.32
(d, 1H, J8,7 ¼ 8.9 Hz, C8-H quinolinone), 7.71 (dd, 1H, J7,5 ¼ 1 Hz,
J7,8 ¼ 8.9 Hz, C7-H quinolinone), 7.73 (d, 1H, Jtrans ¼ 16 Hz, CH]
CH), 8.59 (s, 1H, C2-H quinolinone), 8.65 (d, 1H, J5,7 ¼ 1 Hz, C5-H
quinolinone).
NMR (DMSO-d6, 400 MHz):
d 5.17 (s, 2H, CH2), 5.71 (s, 2H, CH2),
7.09e7.34 (m, 6H, AreH), 7.39 (m, 1H, C7-H quinolinone), 7.43e7.53
(m, 2H, AreH), 7.68 (d, 1H, J8,7 ¼ 9.4 Hz, C8-H quinolinone), 7.77 (d,
1H, J5,7 ¼ 3 Hz, C5-H quinolinone), 7.99 (s, 1H, C3-H butenoate), 9.03
(s, 1H, C2-H quinolinone). 13C NMR (DMSO-d6, 400 MHz): 56.13
(NeCH2), 69.47 (OeCH2),102.10 (CH),108.93 (CH),113.35 (C), 115.86
(d, J ¼ 22.0 Hz, 2CH), 116.32 (d, J ¼ 21.0 Hz, 2CH), 120.64 (CH), 123.11
(CH), 129.39 (d, J ¼ 7.6 Hz, 2CH), 130.31 (C), 130.57 (d, J ¼ 8.6 Hz,
2CH), 132.35 (C), 133.30 (d, J ¼ 1.9 Hz, C), 133.56 (C), 148.96 (CH),
156.61 (C), 162.22 (d, J ¼ 244.41 Hz, C), 162.36 (d, J ¼ 243.45 Hz, C),
164.22 (C), 174.33 (C). LC-MS: (M þ H)þ 492.0.
4.1.5. General synthesis procedure for diketo esters (3 and 5)
Freshly prepared sodium ethoxide (0.272 g, 2 mmol) was added
to a well stirred mixture of diethyl oxalate (0.584 g, 4 mmol) and
appropriate acetyl derivative 9 or 11 (2 mmol) in anhydrous THF
(2 ml) under nitrogen atmosphere. The reaction mixture was stir-
red at room temperature for 20 h. n-hexane (70 ml) was added
and the formed precipitate was filtered, washed with n-hexane and
stirred for 30 min in 1N HCl (70 ml). The solid was filtered off and
washed with n-hexane. Residual water was removed by azeotropic
distillation using acetonitrile to provide pure diketo esters 3 and 5.
4.1.6.2. 4-[6-((E)-3-hydroxy-3-oxoprop-1-enyl)-1-(4-fluorophenyl)
methyl-4(1H)-quinolinon-3-yl]-2-hydroxy-4-oxo-2-butenoic
acid
(6). Yellow solid, yield: 70%, mp: 163 ꢁC. IR ymax (KBr/cmꢀ1): 1703
(C]O carboxylic acid), 1698 (C]O acrylic acid), 1636 (C]O qui-
nolinone), 1599 (C]O þ C]C keto-enol). 1H NMR (DMSO-d6,
4.1.5.1. 4-[6-(4-fluorobenzyloxy)-1-(4-fluorophenyl)methyl-4(1H)-qui-
nolinon-3-yl]-2-hydroxy-4-oxo-2-butenoic acid ethyl ester (3). Yellow
solid, yield: 82%, mp: 160 ꢁC. IR ymax (KBr/cmꢀ1): 1743 (C]O
carboxylic acid), 1634 (C]O quinolinone), 1605 (C]O þ C]C keto-
400 MHz):
d
5.75 (s, 2H, CH2), 6.58 (d, 1H, Jtrans ¼ 16 Hz, CH]CH),
7.05e7.42 (m, 4H, AreH), 7.61e7.78 (m, 2H, CH]CH and C8-H
quinolinone), 7.96 (s, 1H, C3-H butenoate), 8.05 (m, 1H, C7-H qui-
nolinone), 8.43 (d, 1H, J5,7 ¼ 2 Hz, C5-H quinolinone), 9.08 (s, 1H, C2-
H quinolinone). 13C NMR (DMSO-d6, 400 MHz): 56.01 (NeCH2),
102.17 (CH), 116.35 (d, J ¼ 22.0 Hz, 2CH), 119.46 (CH), 121.19 (CH),
127.61 (CH), 129.41 (d, J ¼ 8.6 Hz, 2CH), 132.06 (C), 132.28 (d,
J ¼ 1.9 Hz, C), 140.07 (C), 142.85 (CH), 150.01 (C), 150.42 (CH), 162.32
(d, J ¼ 245.37 Hz, C), 162.85 (C), 164.19 (C), 167.87 (C), 174.65 (C). LC-
MS: (M þ H)þ 438.0.
enol). 1H NMR (DMSO-d6, 400 MHz):
d 1.28 (t, 3H, CH2CH3), 4.27 (q,
2H, CH2CH3), 5.17 (s, 2H, CH2), 5.72 (s, 2H, CH2), 7.09e7.34 (m, 6H,
AreH), 7.39 (dd, 1H, J7,5 ¼ 3 Hz, J7,8 ¼ 9.4 Hz, C7-H quinolinone),
7.43e7.53 (m, 2H, AreH), 7.66 (d, 1H, J8,7 ¼ 9.4 Hz, C8-H quinoli-
none), 7.79 (d, 1H, J5,7 ¼ 3 Hz, C5-H quinolinone), 8.00 (s, 1H, C3-H
butenoate), 9.04 (s, 1H, C2-H quinolinone). 13C NMR (CDCl3,
400 MHz): 14.26 (CH3), 57.59 (NeCH2), 62.42 (CH2), 69.91 (OeCH2),
102.47 (CH), 108.58 (CH), 114.27 (C), 115.72 (d, J ¼ 22.0 Hz, 2CH),
116.65 (d, J ¼ 22.0 Hz, 2CH), 118.71 (CH), 123.68 (CH), 128.12
(d, J ¼ 8.6 Hz, 2CH), 129.66 (d, J ¼ 8.6 Hz, 2CH), 130.55 (C), 131.90
(d, J ¼ 2.8 Hz, C), 133.28 (C), 147.50 (CH), 156.77 (C), 162.62 (C),
162.73 (d, J ¼ 247.28 Hz, C), 162.81 (d, J ¼ 248.24 Hz, C), 169.21 (C),
174.83 (C), 188.09 (C). LC-MS: (M þ H)þ 520.0.
4.2. Biological assays
4.2.1. Cells and viruses
The following reagents were obtained through the AIDS
Research and Reference Reagent Program, Division of AIDS, NIAID,
NIH: TZM-bl from Dr. John C. Kappes, Dr. Xiaoyun Wu and Tran-
zyme Inc. TZM-bl cells were grown at 37 ꢁC, 5% CO2 in DMEM
medium (Invitrogen) supplemented with 5% (v/v) heat-inactivated
4.1.5.2. 4-[6-((E)-3-ethoxy-3-oxoprop-1-enyl)-1-(4-fluorophenyl)met-
hyl-4(1H)-quinolinon-3-yl]-2-hydroxy-4-oxo-2-butenoic acid ethyl ester
(5). Yellow solid, yield: 79%, mp: 146 ꢁC. IR ymax (KBr/cmꢀ1): 1733
(C]O carboxylic acid), 1702 (C]O acrylic ester), 1644 (C]O qui-
bovine serum, 50 U of penicilin/ml and 50 mg of streptomycin/ml
(Invitrogen). Exponentially growing cells were trypsinised, centri-
fuged and split twice weekly. The HIV strain used in this study is
HIV-1 NL4-3 derived from the infectious molecular clone pNL4-3.
HIV-1 stocks were stored as cell-free supernatants at ꢀ80 ꢁC.
nolinone), 1600 (C]O
þ
C]C keto-enol). 1H NMR (CDCl3,
400 MHz): 1.33 (t, 3H, CH2CH3), 1.41 (t, 3H, CH2CH3), 4.26 (q, 2H,
d

P. Vandurm et al. / European Journal of Medicinal Chemistry 46 (2011) 1749e1756
1755
4.2.2. In vitro assays
Tn5 transposome crystal structure were then merged into the HIV-1
model to complete the octahedral geometry of the two Mg2þ ions.
The slightly resolved active-site loop (residues 141e148) was
removed and replaced by the equivalent PFV IN loop (residues
210e217), after which residues were mutated to correspond to the
HIV-1 IN sequence. To remove bad intramolecular contacts without
altering the architecture of the binding site, the resulting model was
energy-minimized by 500 steps of steepest descent using CHARMm
force field and Generalized Born (GB) model as solvation treatment
implemented in Discovery Studio 2.5 [34]. Coordinates of model are
available upon request to the authors.
Compounds 3e6 were tested in vitro for 30-P and ST inhibition in
the presence of magnesium (Mg2þ) using a gel-based assay [27]. In
this assay, the extent of 30-P and ST was analyzed by using as
substrate a 50-32P end-labeled 21-mer double-stranded DNA oligo-
nucleotide corresponding to the last 21 bases of the U5 viral LTR.
IN1ꢀ288 (400 nM) was incubated with 12.5 nM of the 50-end 32P-
labeled oligonucleotide substrate in a reaction buffer (20 mM HEPES
pH7.5,1 mM DTT,10 mM MgCl2,10% dimethyl sulfoxyde) 1 h at 37 ꢁC.
Oligonucleotides were precipitated in a precipitation buffer (3M
sodium acetate pH7, 10 mg/ml glycogen, Tris-EDTA pH7.5), extracted
with phenol/chloroform and washed with ethanol. The oligonucle-
otides were resuspended in 10
ml of 10 mM formamide/EDTA,
4.3.2. Docking
denatured and separated by electrophoresis on a polyacrylamide/
urea gel (16% acrylamide, 7 M urea). Gels were dried, exposed in
a PhosphorImager cassette, analyzed using a Storm mode imager
(Amersham Biosciences) and quantified using ImageQuant. The IC50
values were determined from a curve (drug concentration versus
percent inhibition) fitted on experimental data (Prism software) to
obtain the concentration that produced 50% inhibition.
Docking of quinolone compounds into the HIV-1 IN/DNA model
was carried out using GOLD (version 1.3.1) Software [30]. All
compounds were modeled by considering the stable conformation
previously identified by experimental and theoretical studies [35].
The binding region was defined as a 15 Å radius sphere centered on
Mg2þ ions chelated by E152. An octahedralgeometry was imposed on
the two Mg2þ ions whereas explicit water molecules were allowed to
switch on and off and to rotate around their three principal axes. For
the 30 genetic algorithm (GA) runs performed, a total of 100 000
genetic operations were carried out on five islands, each containing
100 individuals. The niche size was set to 2, and the value for the
selection pressure was set to 1.1. Genetic operator weights for cross-
over, mutation, and migration were set to 95, 95 and 10, respectively.
ThescoringfunctionusedtorankthedockingwasGoldscore.IN/DNA/
ligand complex obtained by docking studies were minimized using
steepest descent method (10 kcal molꢀ1 Åꢀ1 convergence) followed
by conjugate gradient minimization (0.0001 kcal molꢀ1 Åꢀ1 conver-
gence) imposing the coplanarity of the diketo acid moiety. All
complexes were minimized using CHARMm force field and GB model
as solvation treatment implemented in Discovery Studio 2.5 [34].
4.2.3. Antiviral activity
Antiviral activity of the compounds was evaluated using the
TZM-bl [28] cell line. This TZM-bl cell line uses a Tat protein-induced
transactivation of the HIV-1 LTR promoter driving the expression of
the b-galactosidase gene. In brief, exponentially growing cells were
plated at 5000 cells/well in flat-bottom 96-well microtiter plates.
After a 48 h incubation, the medium was removed and replaced by
100
ml of medium containing the quinolinone derivatives at
concentrations ranging from 0.00128 to 4
mM. Four hours after drug
treatment, cells were infected with equal amount of cell-free virus
(NL4-3 strain), corresponding to 100 ng of HIV p24 antigen. Forty-
eight hours post-infection, the medium was removed and cells were
lysed in PBS containing 0.1% NP-40 and 5 mM MgCl2.
b-galactosidase
activity was measured using chlorophenol red -galactopyrano-
side (Roche) assays. The 50% effective concentration (EC50) was
calculated from each dose response curve using Prism software.
b-D
Acknowledgements
This research was supported in part by grants from the DGTRE
(Direction générale des Technologies, de la Recherche et de
l’Energie, Région Wallonne) (Programs WALEO 06/1/6295 and 05/
1/5995). CVL is Directeur de Recherches of the Belgian “Fonds de la
Recherche Scientifique-FNRS” (FRS-FNRS). P. Vandurm is a fellow of
the FRIA/FNRS. A. Guiguen is a Postdoctoral fellow of the FRS-FNRS.
C. Michaux is a Research Associate of the FRS-FNRS.
4.2.4. Cytotoxicity
Flat-bottom 96-well plates were filled with 50
medium containing 5000 TZM-bl cells. Two hours later, 50
medium containing an appropriate dilution of the test compounds
(from 100 to 0.032 M) was added. Cells and compounds were
m
l of complete
ml of
m
incubated at 37 ꢁC in growth medium for 3 days. Cell viability was
determined by a colorimetric assay based on the cleavage of the
tetrazolium salt WST-1 (Roche) by mitochondrial dehydrogenases in
viable cells. Absorbance of converted dye was measured in an ELISA
plate reader at the wavelength of 570 nm. The cytotoxicity of the
compounds was calculated as the reduction percentage of viable cells
observed in the drug-free control cell culture. The drug concentration
References
[1] E. Asante-Appiah, A.M. Skalka, Antivir. Res. 36 (1997) 139e156.
[2] T.K. Chiu, D.R. Davies, Curr. Top. Med. Chem. 4 (2004) 965e977.
[3] O. Delelis, K. Carayon, A. Sa, E. Deprez, J.-F. Mouscadet, Retrovirology 5 (2008)
114.
[4] P. Rice, R. Craigie, D.R. Davies, Curr. Opin. Struct. Biol. 6 (1996) 76e83.
[5] P.A. Rice, T.A. Baker, Nat. Struct. Biol. 8 (2001) 302e307.
[6] M. Nowotny, EMBO Reports 10 (2009) 144e151.
required to reduce cell viability by 50% was named CC50
.
[7] D. Esposito, R. Craigie, Adv. Virus Res. 52 (1999) 319e333.
[8] Y. Goldgur, F. Dyda, A.B. Hickman, T.M. Jenkins, R. Craigie, D.R. Davies, Proc.
Natl. Acad. Sci. USA 95 (1998) 9150e9154.
[9] S. Maignan, J.P. Guilloteau, Q. Zhou-Liu, C. Clement-Mella, V. Mikol, J. Mol. Biol.
282 (1998) 359e368.
[10] R.L. LaFemina, C.L. Schneider, H.L. Robbins, P.L. Callahan, K. LeGrow, E. Roth,
W.A. Schleif, E.A. Emini, J. Virol. 66 (1992) 7414e7419.
[11] Z. Debyser, P. Cherepanov, B. Van Maele, E. De Clercq, M. Witvrouw, Antivir.
Chem. Chemother. 13 (2002) 1e15.
4.3. Molecular modeling
4.3.1. Model building
The HIV-1 IN/DNA model in its INSTI-inhibited form was initiated
by overlaying the structures of HIV-1 IN CCD (1BL3C), Tn5 trans-
posome (1MUS) and elvitegravir-bound PFV structure (3L2U) with
their Ca positions of active-site residues Asp64, Asp116, and Glu152
[12] P. Cotelle, Recent Pat. Anti-Infect. Drug Discov. 1 (2006) 1e15.
[13] E. Serrao, S. Odde, K. Ramkumar, N. Neamati, Retrovirology 6 (2009) 25.
[14] D.J. McColl, X. Chen, Antivir. Res 85 (2010) 101e118.
[15] V. Summa, A. Petrocchi, F. Bonelli, B. Crescenzi, M. Donghi, M. Ferrara, F. Fiore,
C. Gardelli, O. Gonzalez Paz, D.J. Hazuda, P. Jones, O. Kinzel, R. Laufer,
E. Monteagudo, E. Muraglia, E. Nizi, F. Orvieto, P. Pace, G. Pescatore,
R. Scarpelli, K. Stillmock, M.V. Witmer, M. Rowley, J. Med. Chem. 51 (2008)
5843e5855.
(DDE motif). Bases of the 19-mer PFV DNA properly oriented into the
HIV-1 IN CCD were mutated in order to correspond to the HIV-1 DNA
sequence. Two Mg2þ ions were then modeled into the HIV-1 active
site based on the two Mn2þ ions coordinates from PFV structure and
glutamate 152 residue (E152) was reoriented to interact with Mg2þ
ions. Coordinated water molecules present in both HIV-1 IN CCD and

1756
P. Vandurm et al. / European Journal of Medicinal Chemistry 46 (2011) 1749e1756
[16] M. Sato, T. Motomura, H. Aramaki, T. Matsuda, M. Yamashita, Y. Ito,
H. Kawakami, Y. Matsuzaki, W. Watanabe, K. Yamataka, S. Ikeda, E. Kodama,
M. Matsuoka, H. Shinkai, J. Med. Chem. 49 (2006) 1506e1508.
[17] M. Sato, H. Kawakami, T. Motomura, H. Aramaki, T. Matsuda, M. Yamashita,
Y. Ito, Y. Matsuzaki, K. Yamataka, S. Ikeda, H. Shinkai, J. Med. Chem. 52 (2009)
4869e4882.
[25] R. Di Santo, R. Costi, A. Roux, M. Artico, A. Lavecchia, L. Marinelli, E. Novellino,
L. Palmisano, M. Andreotti, R. Amici, C.M. Galluzzo, L. Nencioni, A.T. Palamara,
Y. Pommier, C. Marchand, J. Med. Chem. 49 (2006) 1939e1945.
[26] R. Di Santo, R. Costi, A. Roux, G. Miele, G.C. Crucitti, A. Iacovo, F. Rosi,
A. Lavecchia, L. Marinelli, C. Di Giovanni, E. Novellino, L. Palmisano,
M. Andreotti, R. Amici, C.M. Galluzo, L. Nencioni, A.T. Palamara, Y. Pommier,
C. Marchand, J. Med. Chem. 51 (2008) 4744e4750.
[27] H. Leh, P. Brodin, J. Bischerour, E. Deprez, P. Tauc, J-C. Brochon, E. Le Cam,
D. Coulaud, C. Auclair, J.-F. Mouscadet, Biochemistry 39 (2000)
9285e9294.
[28] X. Wei, J.M. Decker, H. Liu, Z. Zhang, R.B. Arani, J.M. Kilby, M.S. Saag, X. Wu,
G.M. Shaw, J.C. Kappes, Antimicrob. Agents Chemother. 46 (2002) 1896e1905.
[29] L. Krishnana, X. Lia, H.L. Naraharisettya, S. Hareb, P. Cherepanov, A. Engelman,
Proc. Natl. Acad. Sci. U.S.A. 107 (2010) 15910e15915.
[30] J. Tang, K. Maddali, Y. Pommier, Y.Y. Sham, Z. Wang, Bioorg. Med. Chem. Lett.
20 (2010) 3275e3279.
[18] D. Hazuda, M. Iwamoto, L. Wenning, Annu. Rev. Pharmacol. Toxicol. 49 (2009)
377e394.
[19] E.P. Garvey, B.A. Johns, M.J. Gartland, S.A. Foster, W.H. Miller, R.G. Ferris,
R.J. Hazen, M.R. Underwood, E.E. Boros, J.B. Thompson, J.G. Weatherhead,
C.S. Koble, S.H. Allen, L.T. Schaller, R.G. Sherrill, T. Yoshinaga, M. Kobayoshi,
C. Wakasa-Morimoto, S. Miki, K. Nakahara, T. Noshi, A. Sato, T. Fujiwara,
Antimicrob. Agents Chemother. 52 (2008) 901e908.
[20] A.S. Espeseth, P. Felock, A. Wolfe, M. Witmer, J. Grobler, N. Anthony,
M. Egbertson, J.Y. Melamed, S. Young, T. Hamill, J.L. Cole, D.J. Hazuda, Proc.
Natl. Acad. Sci. U.S.A. 97 (2000) 11244e11249.
[21] A. Bacchi, M. Biemmi, M. Carcelli, F. Carta, C. Compari, E. Fisicaro, D. Rogolino,
M. Sechi, M. Sippel, C.A. Sotriffer, T.W. Sanchez, N. Neamati, J. Med. Chem. 51
(2008) 7253e7264.
[22] S. Hare, S.S. Gupta, E. Valkov, A. Engelman, P. Cherepanov, Nature 464 (2010)
232e236.
[23] S. Hare, A.M. Vos, R.F. Clayton, J.W. Thuring, M.D. Cummings, P. Cherepanov,
Nature 107 (2010) 20057e20062.
[24] P. Vandurm, C. Cauvin, A. Guiguen, B. Georges, K. Le Van, V. Martinelli,
C. Cardona, G. Mbemba, J.-F. Mouscadet, L. Hevesi, C. Van Lint, J. Wouters,
Bioorg. Med. Chem. Lett. 19 (2009) 4806e4809.
[31] M. Steiniger-White, I. Rayment, W.S. Reznikoff, Curr.Opin.Struct.Biol. 14
(2004) 50e57.
[32] G. Jones, P. Willett, R.C. Glen, A.R. Leach, R. Taylor, J. Mol. Biol. 267 (1997)
727e748.
[33] E.P. Garvey, B. Schwartz, M.J. Gartland, S. Lang, W. Halsey, G. Sathe,
H.L. Carter III, K.L. Weaver, Biochemistry 48 (2009) 1644e1653.
[34] Accelrys Software Inc, Discovery Studio Modeling Environment, Release 2.5.
Accelrys Software Inc., San Diego, 2007.
[35] P. Vandurm, C. Cauvin, J. Wouters, E.A. Perpète, D. Jacquemin, Chem. Phys.
Lett. 478 (2009) 243e248.