Post on 06-Jun-2016
Study on the Inhibitory Mechanism and Binding Mode of theHydroxycoumarin Compound NSC158393 to HIV-1 Integraseby Molecular Modeling
Ming Liu, Xiao Jing Cong, Ping Li, Jian Jun Tan, Wei Zu Chen, Cun Xin WangCollege of Life Science and Bioengineering, Beijing University of Technology, Beijing 100124, China
Received 11 March 2009; revised 13 April 2009; accepted 13 April 2009
Published online 20 April 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/bip.21211
This article was originally published online as an accepted
preprint. The ‘‘Published Online’’ date corresponds to the
preprint version. You can request a copy of the preprint by
emailing the Biopolymers editorial office at biopolymers@wiley.
com
INTRODUCTION
The pol gene of human immunodeficiency virus type 1
(HIV-1) encodes three enzymes that are essential for
the virus: protease (PR), reverse transcriptase (RT),
and integrase (IN). Inhibitors targeting at PR and RT
have become successful drugs in the fight against
AIDS; however, efforts toward development of effective IN
inhibitors have been hampered by the absence of crystal struc-
tures of the full-length enzyme and by uncertainties in the
understanding of the biochemical mechanism of integration.
Study on the Inhibitory Mechanism and Binding Mode of theHydroxycoumarin Compound NSC158393 to HIV-1 Integraseby Molecular Modeling
Additional Supporting Information may be found in the online version of this
article.Correspondence to: Cun Xin Wang; e-mail: cxwang@bjut.edu.cn
ABSTRACT:
Human immunodeficiency virus type 1 integrase (IN) is
an essential enzyme in the life cycle of this virus and also
an important target for the study of anti-HIV drugs. In
this work, the binding modes of the wild type IN core
domain and the two mutants, that is, W132G and
C130S, with the 4-hydroxycoumarin compound
NSC158393 were evaluated by using the ‘‘relaxed
complex’’ molecular docking approach combined with
molecular dynamics (MD) simulations. Based on the
monomer MD simulations, both of the two substitutions
affect not only the stability of the 128–136 peptides, but
also the flexibility of the functional 140s loop. In
principle, NSC158393 binds the 128–136 peptides of IN;
however, the specific binding modes for the three systems
are various. According to the binding mode of
NSC158393 with WT, NSC158393 can effectively
interfere with the stability of the IN dimer by causing a
steric hindrance around the monomer interface.
Additionally, through the comparative analysis of the MD
trajectories of the wild type IN and the IN-NSC158393
complex, we found that NSC15893 may also exert its
inhibitory function by diminishing the mobility of the
function loop of IN. Three key binding residues, that is,
W131, K136, and G134, were discovered by energy
decomposition calculated with the Molecular Mechanics
Generalized Born Surface Area method. Characterized by
the largest binding affinity, W131 is likely to be
indispensable for the ligand binding. All the above results
are consistent with experiment data, providing us some
helpful information for understanding the mechanism of
the coumarin-based inhibitors. # 2009 Wiley Periodicals,
Inc. Biopolymers 91: 700–709, 2009.
Keywords: HIV-1; coumarins; molecular docking;
molecular dynamics simulation; MM-GBSA
Contract grant sponsor: National Natural Science Foundation of China
Contract grant number: 30670497
Contract grant sponsor: Beijing Natural Science Foundation
Contract grant numbers: 5072002, 7082006
Contract grant sponsor: National Basic Research Program of China
Contract grant number: 2009CB930203VVC 2009 Wiley Periodicals, Inc.
700 Biopolymers Volume 91 / Number 9
So far, Raltegravir is the only FDA-approved HIV IN inhibi-
tor. HIV-1 IN is composed of 288 residues (32 kDa), and can
be divided into three distinct functional domains: the N-
terminal domain (residues 1–49), the catalytic core domain
(residues 50–212), and the C-terminal domain (residues
213–288). The N-terminal domain contains a conserved
‘‘HHCC’’ motif that binds with a Zn21 ion. This domain has
been proved that it can facilitate the multimerization of HIV-
1 IN in vitro.1,2 The catalytic core domain contains three
highly conserved residues, namely D64, D116, and E152,
which coordinate divalent cations such as Mg21 or Mn21.3–5
This domain mainly serves as endonuclease and polynucleo-
tidyl transferase.3 The less conserved C-terminal domain is
responsible for the nonspecific recognition of DNA.6,7 HIV-1
IN mediates the insertion of viral DNA into the host genome.
This process occurs through two separate steps, that is, the
30-processing and the strand transfer reactions, both of which
are catalyzed by IN. In the 30-processing reaction, IN cleaves
a dinucleotide adjacent to a conserved CA on each terminus
of the reverse-transcribed viral DNA. This cleavage results in
two 30 hydroxyl groups that are used for a subsequent nucle-
ophilic attack. IN then inserts this 30-processed virus DNA
into the host genome in the subsequent reaction, namely
strand transfer.8
Recent studies have reported a variety of compounds with
HIV IN inhibitory potency. In many cases, multiple aromatic
rings and aryl orthohydroxylation are required for good in-
hibitory potency. Examples of these compounds include fla-
vones, such as quercetin, caffeic acid phenethyl ester (CAPE),
and analogs as well as some certain ‘‘tyrphostins’’9–11 (see
Figure 1). These inhibitors can be described in general as
consisting of two aryl units, at least one of which contains
the 1,2-dihydroxy pattern, separated by an appropriate linker
segment. The practical utility of these catechol-containing
inhibitors is significantly diminished by cytotoxicity, presum-
ably attributable in part to the in situ oxidation of the cate-
chol moiety to reactive quinone species.12 This is exemplified
by the finding that CAPE-like compounds can cross-link cel-
lular protein at micromolar concentrations, most likely via
Michael-type addition of protein nucleophiles to such qui-
none intermediates.13 Coumarins are affiliated to a class of
potent IN inhibitors that do not contain a catechol group,
however maintaining the inhibitory activity.11 The com-
pound NSC158393 (see Figure 1), which contains four iden-
tical 4-hydroxycoumarin side-chains, inhibited IN at an IC50
5 1.5 lM and inhibited disintegration activity of the catalytic
core domain at a concentration of 15 lM.14 This experiment
provided evidence that the coumarins bind to the catalytic
core domain, and binding this region alone is responsible for
inhibition activity. According to a recent experiment, coumar-
ins do not interact with the highly conserved DDE motif or
the metal cofactors like other inhibitors,15 such as Raltegravir.
FIGURE 1 Structures of three catechol-containing IN inhibitors and two coumarin-containing IN inhibitors.
Hydroxycoumarin Compound NSC158393 701
Biopolymers
In that study, a series of coumarin-containing compounds
(see Figure 1) with a photo-activatable benzophenone moiety
attached were tested. It was found that the benzophenones
of these coumarins were able to covalently bind the
128AACWWAGIK136 peptide in the IN catalytic core, and the
nonconservative substitution of W132 and C130 exhibited
clear resistance to these photo-activatable coumarins.15
Although the binding sites of these photo-activatable cou-
marins were given, the specific binding modes were not
clearly discussed. Besides, the most potent coumarin-con-
taining IN inhibitor, that is, the NSC158393 was not involved
in that study. Therefore, the inhibitory mechanism and bind-
ing mode of this inhibitor remain elusive. Based on these
backgrounds, we intend to investigate the binding modes
and inhibitory mechanism of the hydroxycoumarin com-
pound NSC158393. Understanding of the binding modes
and inhibitory mechanism of NSC158393 is helpful to our
future study on drug design and modification.
In this study, the binding modes of NSC158393 with the
wild type IN and two mutants, that is, W132G and C130S,
were evaluated with the relaxed complex molecular docking
method,16,17 and then the obtained IN-NSC158393 complex
were subject to molecular dynamics (MD) simulations. The
key residues of binding were analyzed through the Molecular
Mechanics Generalized Born Surface Area (MM-GBSA)
approach.18,19 According to the results of molecular docking
and MD simulation, the inhibitory mechanism of
NSC158393 against IN was explained.
MATERIALS AND METHODS
System PreparationThe core domain structure of IN was obtained form the Protein
Data Bank20 with the PDB entry of 1BL3.21 This PDB contains three
IN cores, termed chain A, B, and C, of which chain A and B consti-
tute an IN core dimer. In the current study, the dimer was taken
into the subsequent MD simulation to make sure the simulations
were performed under a relatively reasonable condition. All analyses
are focused on monomer A, if not otherwise specified. The missing
functional 140s loop on each monomer (residues 141–150 in chain
A; 140–149 in chain B) was built by using the Biopolymer module
of Sybyl 7.0. The modeled 140s loops were then minimized by 1000
steps steepest descent followed by 1000 steps of the conjugate gradi-
ent energy minimization. Finally, the structure of the modeled IN
dimer was then checked by Ramachandran map. Most of the resi-
dues were located at the most allowed regions. All hydrogen atoms
were added with xLeap program according to the amber ff03 force
field.22 The processed core domain was then merged into a 70 3 66
3 60 A3 TIP3P23 water box, and four Cl2 ions were added to neu-
tralize the system. The wild type IN core was denoted as WT. After
finishing the preparation of the wild type IN core structure, two
mutated systems were prepared by respectively introducing the
W132G and the C130S substitutions with the Mutator plug-in of
VMD 1.8.6.24
Molecular Dynamics SimulationsAll MD simulations were performed with the NAMD 2.623 in a 28-
nodes HP cluster, and the AMBER all-atom force field22,24 was used.
During the simulations, all bond lengths containing hydrogen atoms
were constrained employing the SHAKE algorithm,25 and the inte-
gration time step was set to 2 fs. First, the three systems were respec-
tively minimized by 20,000 steps with solutes constrained, followed
by 20,000 steps minimizations without any constraint. Then, the
minimized systems were slowly heated form 0 up to 310 K within
120 ps with all Ca atoms and the two Mg21 ions constrained.
Finally, the nonconstrained simulations were performed for 5 ns at
a constant temperature of 310 K and a constant pressure of 1 atm
through the Langevin piston method.26 The subsequent IN-
NSC158393 complex MD simulations were performed as the same
procedures mentioned above as well. The simulation time for the
three complex run was prolonged to 6 ns. Altogether, six independ-
ent MD simulations were performed for the WT and the two
mutant systems.
Molecular Docking with the Relaxed
Complex MethodPresently, the flexibility of receptor is still not taken into considera-
tion in most molecular docking methods. Additionally, the flexibility
of ligand is only partly concerned owing to the costly computation.
As a matter of fact, the conformation of receptor, particularly the
active site, would change to some extent to accommodate the insert
of ligand. Consequently, the docking result could better describe the
actual binding mode if the receptor flexibility can be partly taken in
to account. To accommodate the flexibility of receptor, the relaxed
complex molecular docking approach16,17 was employed in this
study. First, a 5-ns MD simulation was performed for each system
(see the section above for details), and then 12 snapshots were
extracted for each system to dock with the NSC158393. Among the
12 snapshots, the first and the last one are the first and the last
recorded structures of the 5-ns simulation, respectively. The other
10 snapshots are extracted from the beginning to the end of the sim-
ulation according to their Ca RMSD values versus the starting struc-
ture. Basically, we preferred to choose the snapshots with different
conformations in the coumarin binding site. The structure of
NSC158393 was generated and minimized with the Sybyl 7.0
program package. The multiple conformation molecular dockings
calculations were performed with the AutoDock 4.0 package.27,28
To properly employ the force field of AutoDock 4.0, the Gasteiger
partial charges29 were assigned to the ligand and the receptor with
Autodocktools 1.52,30 as suggested by the user guide. All single
bonds of the ligand were treated as rotatable during the docking cal-
culation, and altogether 10 flexible torsions were defined. The maxi-
mum number of energy evaluations was increased to 2.5 3 107 to
explore the conformational space sufficiently. The center of the grid
box was set to the centroid of the 128–136 region of monomer A.
The box size was set to 60 3 60 3 60 A3 with grid spacing 0.375 A
in each dimension, which is large enough for the free rotation of the
ligand. Each docking calculation generated 100 complex structures.
All other docking parameters were set to default. The final solutions
702 Liu et al.
Biopolymers
were selected according to size of clusters and the estimated free
energy of binding (FEB). Basically, the best solution of the largest
cluster was selected as the best mode for each docking test. After
carefully checking the docking results, we found that the binding
modes of the 12 structures are similar (see Supporting Information
Figures 1A–1C for details). Therefore, the mode with the lowest FEB
value was taken as the final bind mode and starting structure for the
following complex MD simulation. The final selected complex of
the wild type IN core and NSC158393 was denoted as WT_NSC,
while that of the two mutants and the NSC158393 were denoted as
W132G_NSC and C130S_NSC, respectively. The three complex sys-
tems were then subjected to minimizations and MD simulations.
The procedures have been addressed above in detail.
MM-GBSAThe MM-GBSA method18,19 was used to perform energy decompo-
sition. The essential idea of the MM-GBSA approach is to divide the
FEB into three parts, namely the gas phase energies, the solvation
free energies, and the nonpolar contributions, and then decompose
these energies into individual contributions for all residues of the
receptor. The gas phase energies are computed with the molecular
mechanics method; the electrostatic contribution to the solvation
free energies are calculated with the Generalized Born (GB) approxi-
mation model,31–34 and the nonpolar contributions to the solvation
free energy are determined with the LCPO model.35,36 All kinds of
energies were decomposed into backbone and side-chain atoms.
Through energy decomposition, we can analyze the contributions of
the key residues to the binding.
RESULTS AND DISCUSSION
Comparative Analysis of the MD Trajectories of the
IN Core Domains
The root mean square deviation (RMSD) for Ca atoms of the
WT, the W132G, and the C130S systems with regard to their
respective starting structures are shown in Figure 2A. Illus-
trated by this figure, the RMSD curves become flat after 2 ns,
indicating the conformations of the protein reach equilib-
rium. The average Ca RMSD values (from 2 ns to the end)
for the three systems are 2.2 6 0.1 A, 1.8 6 0.2 A, and 1.6 6
0.2 A, respectively. The Ca RMSD of the 128–136 peptides
for each system is plotted in Figure 2B. Although the overall
RMSD of the three systems are similar, the RMSD of the
128–136 peptides for the W132G mutant system is slightly
higher than that for the WT, indicating that the 128–136
peptides of this mutant is less stable than the corresponding
region of the WT. This result can be validated by calculating
the Ca root mean square fluctuation (RMSF) for the three
FIGURE 2 Comparative analyses of the MD trajectories of the
WT and the two mutant systems. (A) RMSD of the Ca atoms of the
three systems; (B) RMSD of the Ca atoms of the 128–136 regions of
the three systems. For the two sets RMSD plots, the corresponding
histograms are also given as insets. The plots of Ca RMSD for the
128–136 peptides for the three systems are smoothed out by the ad-
jacent-averaging method (every 20 points) for clarity. (C) RMSF of
Ca atoms of the three systems. For the RMSF plot for all Ca atoms,
the 128–136 regions and the 140s loops of the three systems are
highlighted by green and gray shade bars, respectively. The RMSF
plot for 128–136 regions of the three systems is also given as zoom-
in inset for clarity. In this article, we apply a consistent coloring
scheme to all figures: black for WT, red for W132G, and blue for
C130S. All curves and graphs are represented by corresponding col-
ors, if not otherwise addressed.
Hydroxycoumarin Compound NSC158393 703
Biopolymers
systems (see Figure 2C). As illustrated by Figure 2C, the 128–
136 peptides of the W132G mutant system fluctuate much
more than that of the wild type system. This is mainly owing
to the nonconservative W132G substitution. In the wild type
IN core dimer, the aromatic side-chain of W132 in one
monomer makes p–p stacking interaction with that of the
F181 in the other.15 The substitution of glycine for trypto-
phan breaks this interaction, because glycine lacks the pres-
ence of any side-chain. Since the 128–136 peptides is adjacent
to the interface between the two monomers, the reduction in
the stability of this very region may interfere with the dimeri-
zation process between monomers, thus seriously affecting
the catalytic activity of IN. This speculation is basically in
line with the fact that the W132G and the C130S single sub-
stitution seriously abolish the catalytic activity of strand-
transfer.15 As to the C130S mutant, the flexibility of the
128–136 peptides is just slightly higher than that of the WT,
suggesting that the reduction of the catalytic activity of this
mutant is probably not caused by the reduction of the stabil-
ity of the 128–136 region. In addition to these findings, we
also noted the differences among the flexibilities of the 140s
loops of the three systems. Greenwald et al. indicated that the
mobility of the 140s loop is correlated with the catalytic ac-
tivity of IN. The G140A/G149A double mutations directly
constrained the mobility of this loop.37 Besides, some com-
putional works also implied that the flexibility of this loop
may be a requirement for efficient biological activities.38–40
According to our MD simulations, it was found that the
functional 140s loops of the two mutant systems, especially
the C130S, are not flexible as that of the WT system (see Fig-
ure 2C). This may be another important reason why the cata-
lytic activities of such two mutants are much weaker than
that of the wild type IN.
The Prediction of Binding Modes
The predicted binding modes of NSC158393 with the WT
and the two mutants are illustrated in Figures 3A–5A, and
the detailed information about interactions between the
ligand and the protein are exhibited by Figures 3B–5B. To
simplify the following analyses, we denoted the four side-
chains of the ligand as S1 (abbreviation of side-chain one),
S2, S3, and S4 according to their orientations about the
screen, respectively. For example, S1 represents the side-chain
in the top left of the screen, S2 is used to represent the side-
chain in the bottom left, and so on. The exact order is mean-
ingless, since the four side-chains of NSC158393 are identical
in chemical property.
As is illustrated in Figure 3A, NSC158393 binds the back-
bone of the peptide in the 128–136 region of chain A, which
is in good agreement with the recent photocrosslinking
experiment.15 From Figure 3B, it is found that the
NSC158393 makes hydrophobic interactions with the C130,
W131, W132, G134, I135, K136, Q137, and F139 of mono-
mer A. Actually, only three of the hydroxycoumarin side-
chains of NSC158393 are responsible for the binding, while
the forth side-chain, that is, the S4, does not make any inter-
action with the protein. Two of the three binding side-chains
(S1 and S2) dip into the captivity constituted by the C130,
W131, I135, K136, Q137, and F139 of chain A; the third
binding side-chain (S3) binds the right side of this peptide,
and inserts into the deep trench formed at the interface of
two monomers, thus possibly resulting in the destabilization
of the IN dimer by causing a steric hindrance around the
FIGURE 3 The predicted binding mode of the wild type IN with
the NSC158393 calculated with molecular docking (A), and the
contact residues around the binding site (B). The contact residues
are calculated with a cutoff of 4 A with the LIGPLOT program pack-
age.41 The IN dimer are represented by its solvent-accessible-surface
area (SASA) model, and the contact residues are labeled by corre-
sponding texts. NSC158393 are represented by stick model. The
same representation scheme is also used for Figures 4 and 5.
704 Liu et al.
Biopolymers
monomer interface. This finding is in good agreement with
the previous results.15 The overall binding mode of
NSC158393 here resembles a talon. S1, S2, and S3 tightly
hold the a-helix consisting of the 128–136 peptides (see the
Supporting Information Figure 2), whereas S4 extends to the
opposite direction of the binding site. Although the forth
side-chain (S4) of NSC158393 does not contribute to the
binding with IN, it may interfere with the formation of mul-
timer of IN by causing a steric hindrance at the interface
between dimers; however, this speculation needs to be vali-
dated further.
As revealed by a previous experiment, some benzophe-
none-containing coumarins can also bind the 128–136 pep-
tides of the W132G and the C130S mutants.15 According to
Figure 4A, the binding mode of NSC158393 with the W132G
mutant is different from that with the wild type IN. In gen-
eral, NSC158393 binds the left side of the a-helix constitutedby residues from 128 to 136, and only the S3 and S4 side-
chains of the ligand are involved in these interactions. As
addressed above, previous study has indicated that the p–pstacking interaction between the aromatic side-chains of
W132 in one monomer and F181 in the other monomer may
be important for the dimerization of two IN monomers.15
Since the nonconservative W132G substitution eliminates
the p–p stacking interaction, the previous binding pocket is
slightly distorted in the mutated system. This may partly
account for the shift of the NSC158393 binding site in the
W132G mutant. Moreover, it is worth noting that
NSC158393 makes hydrophobic interactions with most of
residues that are involved in the WT case (see Figure 4B).
This finding confirms that NSC158393 do interact with the
128–136 peptides, which is in accordance with previous
assumptions.15 Although the NSC158393 still binds the 128–
136 peptides of the W132G mutant, this compound is prob-
ably not capable of interfering with the stability of the IN
dimer as effectively as it does to the wild type IN, because it
is far away from the dimer interface. As for the C130S mu-
tant, the NSC158393 binds the left side of 128–136 peptides
FIGURE 4 The predicted binding mode of the W132G mutant
with the NSC158393 calculated with molecular docking (A), and
the contact residues around the binding site (B).
FIGURE 5 The predicted binding mode of the C130S mutant
with the NSC158393 calculated with molecular docking (A), and
the contact residues around the binding site (B).
Hydroxycoumarin Compound NSC158393 705
Biopolymers
with three side-chains, namely S1, S3, and S4, and the other
side-chain (S2) interacts with the adjacent b-sheet (Figure
5A). In this binding mode, the aromatic groups of S1 and S4
make p–p stacking interactions with the side-chain of W131
(Figures 5A and 5B). Like the binding mode with W132G,
the NSC158933 seems can not effectively interfere with the
stability of the IN dimer either, since it is far away from the
dimer interface too.
Comparative Analysis of the MD Trajectories
of the Complexes
According to Figure 6A, the Ca RMSD for the WT_NSC, the
W132G_NSC and the C130S_NSC systems with regard to
their respective starting structures reach equilibrium within
the first 1.5 ns, and the average Ca RMSD (from 1.5 ns to the
end) for the three systems are 1.3 6 0.1, 2.6 6 0.3, and 1.9 6
0.1 A, respectively. It is worth noting that the Ca RMSD for
the WT_NSC is much smaller than those for the two mutant
complex systems. This is particularly true for the
W132G_NSC system, whose Ca RMSD plot fluctuates much
more than the other two. Based on these results, it is sug-
gested that the binding between NSC158393 and wild type
IN is more stable than the two mutants, especially the
W132G substitution.
For the sake of comparing the differences in the flexibility
of IN before and after inhibitor binding, the Ca RMSF of
WT, W132G, and C130S versus their respective complex sys-
tems are given by Figures 6B–6D. Based on Figure 6B, it is
FIGURE 6 Comparative analyses of the MD trajectories of the WT_NSC, W132G_NSC and
C130S_NSC systems. (A) RMSD of the Ca atoms of the three systems versus simulation time and
the corresponding histogram of the RMSD values; (B) RMSF of Ca atoms for the WT and the
WT_NSC; (C) RMSF of Ca atoms for the W132G and its corresponding complex system; (D)
RMSF of Ca atoms for the WT and its corresponding complex system. For all RMSF plots men-
tioned here, the 128–136 regions and the 140s loops of the three systems are highlighted by green
and gray shade bars, respectively. The RMSF plots for 128–136 regions are also given as zoom-in
insets for clarity.
706 Liu et al.
Biopolymers
found that the mobility of the flexible functional 140s loop
in the WT system is much higher than that in the complex
run. This indicates that the binding of NSC158393 may con-
strain the mobility of the functional loop, thus interfering
with the 30-processing as well as the subsequent strand trans-
fer reactions. Although NSC158393 binds the 128–136
peptides, it seems that it does not notably affect the mobility
of this region.
As to the Ca RMSF plots of the two mutant systems and
their corresponding complex systems (see Figures 4C and
4D), differences in flexibilities are still observed in the func-
tional 140s loop regions. However, the differences are much
smaller than those between the WT and the WT_NSC runs,
indicating that the binding of NSC158393 does not largely
affect the flexibility of the functional loop. This may partially
account for the drug resistance to NSC158393 induced by
the W132G and the C130S substitutions.
Energy Decomposition
Figures 7A–7C show the energy contribution of each residue
in the WT, W132G, and C130S systems to the binding with
NSC158393, respectively. Here, we nominate the residues
that notably contribute (no matter positive or negative) to
the binding of NSC158393 as binding residues, and define
the binding range is from the binding residue with the small-
est residue index to the one with the largest.
It can be seen from Figure 7A that eight residues are
mainly responsible for the ligand binding, namely K127,
C130, W131, W132, G134, K136, I135, and F139. In addi-
tion, nearly no residue makes unfavorable contribution to
the binding. The binding range of the wild type IN is from
K127 to F139. All of these binding residues were discovered
by the previous docking calculations except F139 (see Figure
3B). Therefore, it is necessary to refine the result obtained
form the multiple conformation molecular docking with MD
simulation. Among these binding residues, W131, K136, and
G134 make the first three strongest interactions with the in-
hibitor. Characterized by the high affinity to NSC158393, it
is speculated that the three residues are indispensable for the
binding of this inhibitor. Checking Figures 7B and 7C, two
main results can be found. One is that the binding ranges of
the two mutants are wider than that of the wide type IN, and
the contribution of each residue to ligand binding becomes
moderate when comparing with the large peaks in Figure 7A.
This finding represents that the NSC158393 binds more resi-
dues of the mutant IN than it does in the WT case. The other
result is that the first three strongest binding residues in WT,
that is, the W131, the K136, and the G134, are still very
strong in the two mutants. This finding confirms that W131,
K136, and G134, particularly the W131, are important for
the binding of NSC158393. According to these results, it is
suggested that the binding modes seem well-defined and may
vary amongst the three systems.
FIGURE 7 Binding energy contribution of each residue of WT
(A), W132G (B), and C130S (C) to the NSC158393. The plots are
focused on a small range of residues (residue index: 125 to 150) to
provide a clear representation. The binding energy contributions ver-
sus full range of residues for the three systems are given as insets. The
important residues for binding are marked by corresponding texts.
Hydroxycoumarin Compound NSC158393 707
Biopolymers
Additionally, based on the comparison between Figures
7A and 7B, it is noted that the contribution of I135 for bind-
ing energy in WT is 21.00 kcal/mol; however, the contribu-
tion becomes 0.68 kcal/mol in the W132G system. The repul-
sion between NSC158393 and I135 is partly responsible for
the resistance of W132G mutant to NSC158393. According
to Figures 7A and 7C, the affinity of NSC158393 for the S130
residue of the C130S mutant (21.43 kcal/mol) is nearly seven
times stronger than that for the C130 residue of the wild type
IN (merely 20.22 kcal/mol). This is mainly owing to the dif-
ference in the polarities of serine and cysteine residue.
CONCLUSIONIn this study, MD simulation and molecular docking were per-
formed to evaluate the binding modes of the NSC158393 with
the wild type IN and two mutants, that is, W132G and C130S.
According to the results of the monomer MD simulations, it is
found that both of the two substitutions affect the stability of
the 128–136 peptides and the flexibility of the functional loop.
These findings may partly account for the drug resistance of IN
to the NSC158393. In principle, NSC158393 binds the 128–136
peptides of IN; however, the specific binding modes for the
three systems are various. Based on the binding mode of
NSC158393 with WT, it is suggested that NSC158393 can effec-
tively interfere with the stability of the IN dimer by causing a
steric hindrance around the monomer interface. This binding
mode fits well with the previous experiment.15 Moreover, this
inhibitor may also exert its function by disturbing the forma-
tion of multimer. As for the two mutants, it seems that
NSC158393 can not effectively interfere with the stability of the
IN dimer since it is not close enough to the dimer interface. On
the ground of the complex MD simulations, it is noticed that
the binding of NSC158933 with WT largely affects the flexibil-
ity of the functional loop, partly accounting for the inhibitory
potency of this inhibitor. In contrast, flexibility of the func-
tional loop in the two mutant complexes is not apparently
hampered by the ligand binding. Finally, three key binding resi-
dues, that is, W131, K136, and G134, were discovered by the
results of energy decomposition. Characterized by the largest
binding affinity, W131 seems to be indispensable for the ligand
binding. In conclusion, the mutual recognition process between
IN and NSC158393 was explained in this work. Understanding
of the binding modes and inhibitory mechanism of
NSC158393 is helpful to subsequent IN inhibitor design and
the modification of coumarin-containing compounds.
REFERENCES1. Zheng, R.; Jenkins, T. M.; Craigie, R. Proc Natl Acad Sci USA
1996, 93, 13659–13664.
2. Lee, S. P.; Xiao, J.; Knutson, J. R.; Lewis, M. S.; Han, M. K. Bio-
chemistry 1997, 36, 173–180.
3. Engelman, A.; Craigie, R. J Virol 1992, 66, 6361–6369.
4. Kulkosky, J.; Jones, K. S.; Katz, R. A.; Mack, J. P. G. Mol Cell
Biol 1992, 12, 2331–2338.
5. Polard, P.; Chandler, M. Mol Microbiol 1995, 15, 13–23.
6. Engelman, A.; Hickman, A. B.; Craigie, R. J Virol 1994, 68,
5911–5917.
7. Vink, C.; Groeneger, A. A. M. O.; Plasterk, R. H. A. Nucleic
Acids Res 1993, 21, 1419–1425.
8. Asante-Appiah, E.; Skalka, A. M. Adv Virus Res 1999, 52, 351–
369.
9. Fesen, M. R.; Pommier, Y.; Leteurtrea, F.; Hiroguchia, S.; Yunga,
J.; Kohna, K. W. Biochem Pharmacol 1994, 48, 595–608.
10. Burke, T. R., Jr.; Fesen, M.; Mazumder, A.; Yung, J.; Wang, J.;
Carothers, A. M.; Grunberger, D.; Driscoll, J.; Pommier, Y.;
Kohn, K. J Med Chem 1995, 38, 4171–4178.
11. Mazumder, A.; Gazit, A.; Levitzki, A.; Nicklaus, M.; Yung, J.;
Kohlhagen, G.; Pommier, Y. Biochemistry 1995, 34, 15111–
15122.
12. Zhao, H.; Neamati, N.; Hong, H.; Mazumder, A.; Wang, S.; Sun-
der, S.; Milne, G. W. A.; Pommier, Y.; Burke, T. R. J Med Chem
1997, 40, 242–249.
13. Stanwell, C.; Ye, B.; Yuspa, S. H.; Burke, T. R. Biochem Pharma-
col 1996, 52, 475–480.
14. Mazumder, A.; Wang, S.; Neamati, N.; Nicklaus, M.; Sunder, S.;
Chen, J.; Milne, G. W. A.; Rice, W. G.; Burke, T. R.; Pommier, Y.
J Med Chem 1996, 39, 2472–2481.
15. Al-Mawsawi, L. Q.; Fikkert, V.; Dayam, R.; Witvrouw, M.;
Burke, T. R.; Borchers, C. H.; Neamati, N. Proc Natl Acad Sci
USA 2006, 103, 10080–10085.
16. Lin, J.-H.; Perryman, A. L.; Schames, J. R.; McCammon, J. A.
J Am Chem Soc 2002, 124, 5632–5633.
17. Amaro, R.; Baron, R.; McCammon, J. J Comput-Aided Mol Des
2008, 22, 693–705.
18. Kollman, P. A.; Massova, I.; Reyes, C.; Kuhn, B.; Huo, S.; Chong,
L.; Lee, M.; Lee, T.; Duan, Y.; Wang, W.; Donini, O.; Cieplak, P.;
Srinivasan, J.; Case, D. A.; Cheatham, T. E. Acc Chem Res 2000,
33, 889–897.
19. Wang, W.; Donini, O.; Reyes, C. M.; Kollman, P. A. Annu Rev
Biophys Biomol Struct 2001, 30, 211–243.
20. Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T.
N.; Weissig, H.; Shindyalov, I. N.; Bourne, P. E. Nucleic Acids
Res 2000, 28, 235–242.
21. Maignan, S.; Guilloteau, J. P.; Zhou-Liu, Q.; Clement-Mella, C.;
Mikol, V. J Mol Biol 1998, 282, 359–368.
22. Duan, Y.; Wu, C.; Chowdhury, S.; Lee, M. C.; Xiong, G.; Zhang,
W.; Yang, R.; Cieplak, P.; Luo, R.; Lee, T.; Caldwell, J.; Wang, J.;
Kollman, P. J Comput Chem 2003, 24, 1999–2012.
23. Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R.
W.; Klein, M. L. J Chem Phys 1983, 79, 926–935.
24. Humphrey, W.; Dalke, A.; Schulten, K. J Mol Graph 1996, 14,
33–38.
25. Kale, L.; Skeel, R.; Bhandarkar, M.; Brunner, R.; Gursoy, A.; Kra-
wetz, N.; Phillips, J.; Shinozaki, A.; Varadarajan, K.; Schulten, K.
J Comput Phys 1999, 151, 283–312.
26. Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K.
M.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.;
Kollman, P. A. J Am Chem Soc 1995, 117, 5179–5197.
708 Liu et al.
Biopolymers
27. Ryckaert, J. P.; Ciccotti, G.; Berendsen, H. J. C. J Comput Phys
1977, 23, 327–341.
28. Feller, S. E.; Zhang, Y.; Pastor, R. W.; Brooks, B. R. J Chem Phys
1995, 103, 4613–4621.
29. Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart,
W. E.; Belew, R. K.; Olson, A. J. J Comput Chem 1998, 19,
1639–1662.
30. Huey, R.; Morris, G. M.; Olson, A. J.; Goodsell, D. S. J Comput
Chem 2007, 28, 1145–1152.
31. Gasteiger, J.; Marsili, M. Tetrahedron 1980, 36, 3219–3228.
32. Sanner, M. F. J Mol Graph Model 1999, 17, 57–61.
33. Tsui, V.; Case, D. A. Biopolymers 2000, 56, 275–291.
34. Simonson, T. Curr Opin Struct Biol 2001, 11, 243–252.
35. MacKerell, A. D.; Bashford, D.; Bellott, M.; Dunbrack, R. L.;
Evanseck, J. D.; Field, M. J.; Fischer, S.; Gao, J.; Guo, H.; Ha, S.;
Joseph-McCarthy, D.; Kuchnir, L.; Kuczera, K.; Lau, F. T. K.;
Mattos, C.; Michnick, S.; Ngo, T.; Nguyen, D. T.; Prodhom, B.;
Reiher, W. E.; Roux, B.; Schlenkrich, M.; Smith, J. C.; Stote, R.;
Straub, J.; Watanabe, M.; Wiorkiewicz-Kuczera, J.; Yin, D.; Kar-
plus, M. J Phys Chem B 1998, 102, 3586–3616.
36. Bashford, D.; Case, D. A. Annu Rev Phys Chem 2000, 51, 129–
152.
37. Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T. J Am
Chem Soc 1990, 112, 6127–6129.
38. Weiser, J.; Shenkin, P. S.; Still, W. C. J Comput Chem 1999, 20,
217–230.
39. Greenwald, J.; Le, V.; Butler, S. L.; Bushman, F. D.; Choe, S. Bio-
chemistry 1999, 38, 8892–8898.
40. Lins, R. D.; Briggs, J. M.; Straatsma, T. P.; Carlson, H. A.; Green-
wald, J.; Choe, S.; Andrew McCammon, J. Biophys J 1999, 76,
2999–3011.
41. Barreca, M. L.; Lee, K. W.; Chimirri, A.; Briggs, J. M. Biophys J
2003, 84, 1450–1463.
42. Lee, M. C.; Deng, J.; Briggs, J. M.; Duan, Y. Biophys J 2005, 88,
3133–3146.
43. Wallace, A. C.; Laskowski, R. A.; Thornton, J. M. Protein Eng
1995, 8, 127–134.
Reviewing Editor: J. McCammon
Hydroxycoumarin Compound NSC158393 709
Biopolymers