Mechanism of drug resistance of hemagglutinin of influenza virus and potent scaffolds inhibiting its function

Article abstract of DOI:10.1021/cb200332k

Highly pathogenic influenza viruses have become a global threat to humans. It is important to select an effective therapeutic option suitable for the subtypes in an epidemic or pandemic. To increase the options, the development of novel antiviral agents acting on targets different from those of the currently approved drugs is required. In this study, we performed molecular dynamics simulations on a spike protein on the viral envelop, hemagglutinin, for the wild-type and three kinds of mutants using a model system consisting of a trimeric hemagglutinin complex, viral lipid membrane, solvation waters, and ions. A natural product, stachyflin, which shows a high level of antiviral activity specific to some subtypes of influenza viruses, was examined on binding to the wild-type hemagglutinin by docking simulation. The compound potency of stachyflin is, however, easily lost due to resistant mutations. From a comparison of simulation results between the wild-type and the resistant mutants, the reason for the drug resistance of hemagglutinin was clarified. Next, 8 compounds were selected from a chemical database by in silico screening, considering the findings from the simulations. Inhibitory activities to suppress the proliferation of influenza virus were measured by cell-based antiviral assays, and two chemical scaffolds were found to be potent for an inhibitor. More than 30 derivatives bearing either of these two chemical scaffolds were synthesized, and cell culture assays were carried out to evaluate the compound potency. Several derivatives displayed a high compound potency, and 50% effective concentrations of two synthesized compounds were below 1 μM.

Full text of DOI:10.1021/cb200332k

Articles
pubs.acs.org/acschemicalbiology
Mechanism of Drug Resistance of Hemagglutinin of Influenza Virus
and Potent Scaffolds Inhibiting Its Function
Hiroshi Yanagita,†,# Norio Yamamoto,‡,§,#,* Hideyoshi Fuji,† Xinli Liu,† Masakazu Ogata,†
Mizuho Yokota,† Hiroshi Takaku,∥ Hideki Hasegawa,⊥ Takato Odagiri,‡ Masato Tashiro,‡
and Tyuji Hoshino†,*
†Graduate School of Pharmaceutical Sciences, Chiba University, Inohana 1-8-1, Chuo-ku, Chiba 260-8675, Japan
⊥
‡Influenza Virus Research Center and Department of Pathology, National Institute of Infectious Diseases, 4-7-1 Gakuen,
Musashimurayama, Tokyo 208-0011, Japan
§Department of General Medicine, Juntendo University School of Medicine, 2-1-1 Hongo, Bunkyo-ku, Tokyo 113-8421, Japan
∥Department of Life and Environmental Science, Chiba Institute of Technology, 2-17-1 Tsudanuma, Narashino-shi, Chiba 275-0016,
Japan
S
* Supporting Information
ABSTRACT: Highly pathogenic influenza viruses have
become a global threat to humans. It is important to select
an effective therapeutic option suitable for the subtypes in an
epidemic or pandemic. To increase the options, the develop-
ment of novel antiviral agents acting on targets different from
those of the currently approved drugs is required. In this study,
we performed molecular dynamics simulations on a spike
protein on the viral envelop, hemagglutinin, for the wild-type
and three kinds of mutants using a model system consisting of
a trimeric hemagglutinin complex, viral lipid membrane,
solvation waters, and ions. A natural product, stachyflin,
which shows a high level of antiviral activity specific to some
subtypes of influenza viruses, was examined on binding to the
wild-type hemagglutinin by docking simulation. The compound potency of stachyflin is, however, easily lost due to resistant
mutations. From a comparison of simulation results between the wild-type and the resistant mutants, the reason for the drug
resistance of hemagglutinin was clarified. Next, 8 compounds were selected from a chemical database by in silico screening,
considering the findings from the simulations. Inhibitory activities to suppress the proliferation of influenza virus were measured
by cell-based antiviral assays, and two chemical scaffolds were found to be potent for an inhibitor. More than 30 derivatives
bearing either of these two chemical scaffolds were synthesized, and cell culture assays were carried out to evaluate the compound
potency. Several derivatives displayed a high compound potency, and 50% effective concentrations of two synthesized
compounds were below 1 μM.
The currently available anti-influenza drugs target one of two
viral proteins: M2 protein and neuraminidase. M2 protein is
embedded in the lipid membrane of the viral envelope and
functions as an ion channel to pump protons into the viral
particles. Amantadine and Rimantadine block the function of
M2 protein by combining at the center of the channel or the
side domain of this enzyme.3,4 Neuraminidase is a kind of spike
protein sticking out on the viral particle surface. Neuraminidase
causes the hydrolysis of neuraminic acid of the glycan of the
host cell. Zanamivir, Oseltamivir, Peramivir, and Laninamivir
have been used as neuraminidase inhibitors.5,6 Emergence of
drug-resistant viruses has been reported for the above approved
nfluenza viruses cause acute respiratory infection in humans
Ithat occasionally progresses to a severe pulmonary
condition. Even seasonal epidemics account for 300,000 or
more deaths per a year all over the world. Recently, the
emergence of highly pathogenic avian and swine influenza
viruses has become a global threat to humans. Avian influenza
H5N1 virus infections have been reported since 2003,1 and a
pandemic of transmissible H5N1 virus is a serious concern for
public health.2 The recent outbreak of swine influenza subtype
H1N1 resulted in considerable mortalities for infants and the
elderly. While several anti-influenza drugs are currently
approved, their effectiveness for pandemic viruses may be
limited due to drug resistance. Therefore, the development of
additional antiviral agents against influenza virus infection is
needed.
Received: September 1, 2011
Accepted: January 4, 2012
Published: January 4, 2012
© 2012 American Chemical Society
552
dx.doi.org/10.1021/cb200332k | ACS Chem. Biol. 2012, 7, 552−562

ACS Chemical Biology
Articles
drugs. For example, resistance against Amantadine and
Rimantadine was shown in the H3N2 and H1N1 viruses, and
resistance against Oseltamivir was shown for the H1N1
virus.7−9 Since an RNA virus easily acquires amino acid
mutations, the emergence rate of drug-resistant viruses is high.
A drug-resistant virus is a serious issue in infectious diseases
because a chemotherapeutic approach is restricted. In order to
combat drug-resistant viruses, it is important to prepare many
chemotherapeutic options and select an effective option
suitable for subtypes in an epidemic. Accordingly, development
of novel antiviral drugs that act on a target different from those
of the currently approved drugs is needed.
A species of moss, Stachybotrys s. RF-7260, generates a
unique natural product named stachyflin. Stachyflin was found
to have strong antiviral activity against some subtypes of
influenza viruses.10−12 The inhibitory mechanism of stachyflin
is different from the inhibitory mechanism of the currently
approved anti-influenza drugs. Stachyflin binds with hemag-
glutinin (HA) on the viral envelope and blocks conformation
change of HA to prevent the lipid membrane of the viral
envelope from merging with that of host cell. HA is one of the
attractive targets of antiviral agents for the following reasons.
First, HA is a key component in the viral infusion process that
has no cellular counterparts and therefore has a potential
advantage in selectivity and toxicity. Second, HA inhibitors will
complement other currently approved drugs since they act on a
different molecular target in the virus life cycle.
Stachyflin is highly effective for influenza virus of A/WSN/33
H1N1 subtype, and the 50% inhibitory concentration (IC50)
value for the WSN strain was reported to be 3 nM.12 However,
its inhibitory activity for other strains including other H1N1
virus strains is not so high, and its inhibitory activity is easily
decreased by amino acid mutations of the virus. Since several
amino acid mutations involved in drug resistance are not
localized in one domain of HA, it is difficult to understand the
mechanism of drug resistance caused by mutations straightfor-
wardly. The chemical structure of stachyflin is complicated.
Five rings merge to form a structure called the 3H-naphtho-
pyrano-isoindol-3-one scaffold, and stachyflin also contains five
chiral centers (Figure 1a). This complexity in its chemical
structure is another reason for the difficulty in improving
compound potency for stachyflin derivatives.
Yoshimoto and co-workers performed an experiment on
resistance induction with stachyflin13 and demonstrated that
K51R, K121E, S206L in the HA2 subunit and V176I in the
HA1 subunit appeared in resistant viruses (Figure 1b). K51R
and K121E mutations were suggested to be essential for drug
resistance. They suggested from a docking simulation that
stachyflin was bound to a position close to Lys51 or Phe110 of
HA. This simulation, however, provided no clear explanation
for the mechanism of drug resistance due to the above amino
acid mutations. The mechanism of the drug resistance of HA
should be clarified for producing promising inhibitory
compounds.
In this study, we performed computational analysis, screening
to find candidate compounds, a cell-based antiviral assay, and
synthesis of analogue compounds in the following manner: (1)
Molecular dynamics (MD) simulations were carried out for the
wild-type and three kinds of variants containing amino acid
mutations responsible for drug resistance in order to clarify the
mechanism. (2) A point for designing a potential inhibitor was
deduced from the concept of minimizing the influence of the
drug-resistance-related conformational change of HA. (3) An in
Figure 1. (a) Chemical structure of stachyflin. Stachyflin is composed
of five complex rings and contains five chiral centers. The rings are
labeled A−E, and chiral carbon atoms are marked by asterisks. (b)
Structure of hemagglutinin in a trimer conformation. Spheres denote
the residues introducing amino mutations in the respective mutants.
The center residue targeted in the ligand docking simulation is
indicated by a circle. (c) Calculation model of a complex of
hemagglutinin trimer and lipid membrane. Hemagglutinin subunits
HA1 and HA2 are colored blue and green, respectively. The
membrane consists of 6 different kinds of lipid molecules, and its
composition is presented in Supplementary Table S1. No water
molecules or ions are shown for visual clarity.
silico screening was performed to find low-molecular-weight
compounds showing inhibitory activity against HA, considering
the above point and using the pharmacophore of stachyflin. (4)
A cell-based assay was carried out to evaluate the inhibitory
potencies of the compounds collected by the in silico screening.
(5) Derivatives of the hit compounds found from the screening
were synthesized, and their inhibitory activities were measured
to elucidate the antiviral potency of the scaffold proposed in
this work.
RESULTS AND DISCUSSION
■
Structural Difference in HA between the Wild-Type
and Mutants. MD simulation was carried out for 30 ns to
obtain the probable protein structure of HA in a trimer form for
the wild-type and three mutants, using the model system
containing a trimeric hemagglutinin and viral lipid membrane as
shown in Figure 1c. The respective mutants contain the
resistant mutations K51R and K121E in HA2 for mutant 1,
V176I in HA1 and K51R, K121E in HA2 for mutant 2, and
V176I in HA1 and K51R in HA2 for mutant 3. Root mean
square deviation (rmsd) relative to the structure after heating is
shown in Supplementary Figure S1. The rmsd value for the
wild-type (Figure S1a) was scarcely changed during 30 ns. Each
of the rmsd curves for the mutants (Figures S1b−d) shows a
553
dx.doi.org/10.1021/cb200332k | ACS Chem. Biol. 2012, 7, 552−562

ACS Chemical Biology
Articles
gradual increase up to 15 ns and seems to be almost constant
after 20 ns. Plots of rmsd in Figure S1 were obtained from the
coordinates of main chain atoms of the whole HA. The N-
terminal domain of the HA1 subunit is so flexible that rmsd
values are considerably large. Hence, rmsd values were
calculated again with respect to the main chain atoms of only
the HA2 subunit with excluding the C-terminal region, aa 176−
222. The rmsd curves only for HA2 in Supplementary Figure
S2 also became constant after 20 ns for every model.
Accordingly, protein conformations for the respective models
were judged to be equilibrated.
Principal component analysis (PCA) in Supplementary
Figure S3 indicates that the trajectory structures for the last 5
ns are in a single conformation for every model. The
equilibration of the simulation is also confirmed from these
PCA plots. In order to extract the plausible protein structure,
the averaged structure was obtained using 500 trajectory
structures from the last 5 ns of MD simulation. The rmsd
between each trajectory structure and the average structure was
calculated, and then one trajectory structure with the smallest
rmsd value was determined to be the plausible protein
structure. At a glance, there is no prominent difference
among the 4 models in terms of shape of the trimer,
conformation of the HA1 and HA2 subunits, or position of
helices. Although no significant apparent change is seen in the
backbone of HA, there appears a notable difference in the
location of side chains. The differences in the side chain will be
responsible for the change in binding affinity and inhibitory
activity of inhibitors.
Binding of an Inhibitor to HA. By means of docking
simulation, an inhibitor, stachyflin, was bound to the HAs,
using the respective plausible protein structures obtained by
MD calculations. In the wild-type HA, stachyflin was bound
near Asp109 of the HA2 subunit (Figure 2). Hydrophobic
interactions were observed between the B ring of stachyflin and
Phe37 of HA1, between the C ring of stachyflin and Phe110 of
HA2, and between the D ring and Leu113. Hydrogen bonds
were formed between the O atom on the D ring of stachyflin
and the amino group of Asn114 in HA2 and between the O
atom on the E ring and the amino group of Asn117.
In mutant 2, stachyflin was bound to a location similar to that
of the wild-type (Supplementary Figure S4c), while the docking
simulation showed binding of stachyflin at the central space
among three helices of the HA2 trimer in mutants 1 and 3
(Figures S4b and d). Judging from the binding affinity
evaluated by ASP score function (Supplementary Table S2),
the binding of stachyflin to the wild-type HA is the most stable.
All of the mutants showed notable decrease in binding affinity
compared to that of the wild-type.
An amino acid mutation of K51R in HA2 commonly
appeared in the three mutants, suggesting that K51R was the
primary mutation for drug resistance. Other mutations, K121E
and V176I, will enhance the resistance. All of these amino
mutations are, however, distant from the stachyflin binding site.
Our MD simulation clearly indicated that inner helices of HA2
subunit were rotated (Figure 3a). One of the three inner
helices, chain D in nomenclature in PDB 1RD8, was rotated by
10.8° in mutant 1 and by 15.0° in mutant 2 compared to the
wild-type (Figure 3b). The amino group on the side chain of
Lys51 makes a strong hydrogen bond with the hydroxy group
of Thr107 of HA2 (Figure 3c). Arg is also a positively charged
amino acid residue, but the length of the side chain is longer
than that of Lys. When Lys is converted into Arg, the side chain
Figure 2. (a) Complex structure of stachyflin and hemagglutinin
obtained by the ligand docking simulation. (b) Binding site of
stachyflin viewed in a plane perpendicular to the helices of the HA2
subunit. (c) Binding mode of stachyflin, shown in a magnified view of
the area indicated by a red frame in panel a. Stachyflin is bound to the
space between two helices of HA2 subunit, making strong interaction
with side chains of Asp109, Phe110, and Leu113. The interaction
distances are in Å.
of the residue at codon 51 expands and pushes T107. The side
chain of Thr107 serves as a lever to rotate the helix. The
position of the side chain of Asp109 is largely deviated from
that of the wild-type because of the closeness to Thr107.
Lys121 has a strong interaction with the carboxy group of
Asp18 of the HA1 subunit. When Lys is converted into Glu in
the K121E mutation, the side chain of Glu121 and Asp18 of
HA1 causes repulsion to increase the distance between them.
This repulsion assists the helix rotation, and the deviation of
Leu113 and Asn114 from the wild-type will be enhanced
because of the closeness to Glu121. To monitor the helix
rotation, the angle between the line connecting the Cα and Cβ
atoms of Asp109 on the inner helix chain D and the line
connecting the Asp109 Cα atom on chain D and the Phe110
Cα atom on another inner helix chain B was measured through
the simulation as shown in Supplementary Figure S5. A
significant angle change was observed after 10 ns for every
model. The distances between the Nζ atom of Lys51 (or Cζ in
K51R) of HA2 and the Oγ atom of Thr107 of HA2 and also
between the Nζ atom of Lys121 (or Cδ in K121E) of HA2 and
Cγ of Asp18 of HA1 were monitored as shown in
Supplementary Figure S6. Because HA is a trimer and there
exist three HA1 and three HA2 subunits in the calculation
models, three combinations of those interatomic distances were
measured through the simulation. The distance plots in Figure
S6 indicate that the interaction between the residues at codon
51 and codon 107 is quite stable in the wild-type HA. In
contrast, some of these distances occasionally increased in the
mutants. The distance between the residues at codon 121 of
HA2 and at codon 18 of HA1 ceaselessly fluctuated in all
554
dx.doi.org/10.1021/cb200332k | ACS Chem. Biol. 2012, 7, 552−562

ACS Chemical Biology
Articles
adrenergic G-protein-coupled receptor14 suggested that a
rotational motion was observed for one of the transmembrane
helices owing to the binding of a ligand for the receptor. An
electron paramagnetic resonance measurement15 and a
computational analysis16 indicated that light adsorption
induced a conformational change of retinal chromophore in
sensory rhodopsin and this conformational change caused the
rotation of transmembrane helix TM1 to lead signal trans-
duction of the sensory protein.
Search for Active Compounds. The natural product
stachyflin possesses a unique pentacyclo structure in which each
ring is labeled as A, B, C, D, and E, respectively. Rings AB are
composed of a naphthol skeleton, and rings DE are composed
of an indol frame. Ring C bears a pyran structure and connects
rings AB and rings DE. Rings ABC are fused in cis-form, and
the oxygen of pyran is bonded to the carbon at the junction of
rings AB. Furthermore, stachyflin contains 5 chiral centers, in
which all of the chiral carbons are located on rings AB.
Therefore, the naphtol moiety and its connection to pyran
make the chemical structure of stachyflin highly complicated. A
hydroxy group and carbonyl oxygen are bound to the indol
ring, which characterizes the electrostatic property of stachyflin.
The hydroxy group of naphthol is another factor to characterize
the polar feature of this natural product.
Eight chemical compounds were selected by an in silico
screening using the pharmacophore of stachyflin (Table 1). All
of the selected compounds were produced by organic synthesis
and are available by purchase. The molecular weights of these
compounds range from 304 to 341. Compound 1 bears two
ester bonds and a benzofuran moiety corresponding to rings C
and D of stachyflin in superimposition. Compound 2 bears a
thieno-pyrimidine, which corresponds to rings D and E of
stachyflin and a phenyl-cyclopentane corresponding to rings A
and B. Two benzene rings are connected via a dichloromethyl-
carbonyl group in compound 3. Methylbenzoic acid methyl-
ester in compound 4 corresponds to rings D and E, and
dimethyl-diazolane is connected by a sulfonyloxy group.
Thieno-pyrimidine in compound 5 corresponds to rings D
and E, and another thiophene corresponds to ring A. The
molecular weight of compound 6 is the largest among the
selected compounds, and methylester-benzene corresponds to
rings D and E and pyridyl-triazole corresponds to rings A and
B. Compound 7 bears a seven-membered ring containing two
nitrogen atoms. Hydroxybenzene in compound 8 corresponds
to ring B of stachyflin.
Assay for Antiviral Activity. The 8 selected compounds
were tested in an influenza virus cell culture assay (Table 1).
Compounds 4 and 5 were found to have significant antiviral
activity (EC50 < 5 μM). Although compound 2 was the most
highly potent with an EC50 of 3 μM, compound 2 exhibited
significant cytotoxicity at a concentration of 6 μM, measured by
an MTT assay with MDCK cells. Since compounds 4 and 5
showed no noticeable cytotoxicity, these two compounds were
chosen as the structural core in the next step for organic
synthesis.
Figure 3. (a) Superposition of the hemagglutinin structures of mutants
on that of the wild-type. The wild type is colored green and mutants 1,
2, and 3 are colored cyan, magenta, yellow, respectively. The binding
area of stachyflin to the wild-type HA is shown in a mesh
representation. There is no significant change in positions of helices.
However, the amino acid mutations cause rotation of helices. The
rotation angle is estimated from the position of Cβ atom of Asp109, as
shown in panel b. Thr107 in HA2 subunit and Asp18 in HA1 subunit
are deeply involved in the helix rotation induced by K51R and K121E
mutations, respectively, as shown in panel c. Two illustrations on the
left side are depicted around the planes perpendicular to the helices
and containing Thr107 or Lys121, respectively. The planes are
indicated by blue and red frames in the right side illustration in panel c.
models. A strong interaction was, however, established at least
for one combination of Lys121 and Asp18 in the wild-type.
The helix rotation pointed out above is the reason for the
decrease in inhibitory activity of stachyflin to the resistant
mutants. In the mutants, the side chains of Asp109 and Leu113
are displaced and occupy the space that stachyflin was bound to
in the wild-type HA. The side chains of Phe110 and Asn114 are
also displaced and move away from the stachyflin binding site
as shown in Supplementary Figure S7. Therefore, stachyflin
would not be bound to HA stably any more. The helix rotation
in mutant 3 is small. This is naturally understood because the
influence of V176I mutation in the HA1 subunit is slight. This
finding suggests that some degree of flexibility is favorable to
HA inhibitors for releasing the strain due to the helix rotation.
Therefore, compounds bearing acomplicated heteroring
structure are disadvantageous. Instead, a single bond con-
nection of separated ring domains can be a good chemical
frame to maintain structural flexibility.
The binding modes of compounds 4 and 5 to the wild-type
HA were predicted by performing docking simulation. The
most probable docking structures are shown in Figure 4a and b,
which were determined from the score ranking for 50 docking
poses. The two compounds are bound to almost the same
position as stachyflin is. That is, these compounds are located
not in the central space of HA2 trimer but at a position in the
middle of two helices of the HA2 subunit. The calculated
Conformational change accompanying helix rotation has
been reported for other kinds of transmembrane proteins. For
example, an X-ray crystallographic analysis of the human β2
555
dx.doi.org/10.1021/cb200332k | ACS Chem. Biol. 2012, 7, 552−562

ACS Chemical Biology
Articles
Table 1. Structures and Inhibitory Activities of the Chemical Compounds Obtained through an in Silico Screening
a
Stachyflin is depicted in green stick representation, while compounds are blue.
binding affinities of these two compounds in ASP score are
lower than that of stachyflin (TSupplementary able S2) but
higher than that in the cases of stachyflin bound to the three
kinds of mutants.
the ester. Substitution of the sulfonyl group by an alkyl chain
(16−20) resulted in loss of inhibitory activity. Only compound
18 in which the sulfonyl group was substituted by benzyl
showed slight inhibitory activity. Then the effect of addition of
functional groups to the benzyl (21−28) was surveyed.
Compound 22 containing an oxybenzyl group at the para-
position of the benzyl exhibited high compound potency. The
inhibitory activity was maintained with the addition of methoxy
or trifluoromethyl at the para-position, while the incorporation
of other kinds of functional groups (23−25) or the addition of
trifluoromethyl at the ortho- or meta-position (27, 28) resulted
in loss of inhibitory activity. Conversion of the sulfonyloxy
group into an amino group (29, 30) was tested. None of the
derivatives showed noticeable increase in compound potency.
Compound 5 bears a heteroring core of thieno-pyrimidine,
and a dimethyl-aminyl-thienyl group is connected to the
heteroring via amine. Keeping the heteroring core, 9 derivatives
from compound 5 were synthesized as shown in Table 3. The
Synthesis of Analogue Compounds. Compound 4 bears
a structural core of vanillic acid. Based on the vanillic acid
methylester core, 22 derivatives from compound 4 were
synthesized as shown in Table 2, where the functional group
at the fourth position of the benzoic acid was modulated. A
highly potent compound (9) was found in the analogues in
which methylsulfonyl was connected to the fourth position.
The incorporation of benzyl or methylphenyl (10, 12) showed
no inhibitory activity, while moderate activity was observed in
case of phenyl only (11). This means that a small chemical
group is favorable for the substitute connecting via the
sulfonyloxy group. Conversion of the sulfonyl group into an
ester bond (13−15) resulted in complete loss of compound
potency, regardless of the size of the substitutes connected to
556
dx.doi.org/10.1021/cb200332k | ACS Chem. Biol. 2012, 7, 552−562

ACS Chemical Biology
Articles
containing piperadine connected to trifluoro-methylbenzoyl
also showed high inhibitory activity for influenza virus
replication.20 Ligand docking calculation suggested that this
compound was bound to a site near Phe110 of the HA2
subunit, which is almost identical to the binding site of
stachyflin obtained in this work as shown in Figure 2. A
compound analogue to podocarpic acid was identified as an
inhibitor of type A viruses21 and showed a high inhibitory
activity especially for Kawasaki strain but was not so sensitive to
WSN strain. In contrast, some recently identified blockers are
targeted at another domain of HA.23,24 Some peptides
mimicking sialic acid were shown to inhibit the entry of viruses
into host cells, attached to the receptor-binding site of HA.23
Synthesized macromolecules containing three sialyllactoses
linked with trisphenol or trisaniline were reported to inhibit
viral replication, combined with the receptor-binding site of
HA.24 Recently, an HA inhibitor bearing a benzenesulfonamide
core was identified through structural modifications of a
salicylamide derivative.25 It is interesting to note that the
chemical structure of this agent resembles that of compound 9
in Table 2 to some extent.
In most of the active compounds, susceptibility varies among
strains of influenza viruses and the inhibitory activity is
drastically decreased due to the acquirement of resistant
mutation. Our MD simulation demonstrated that the rotation
of helices is the reason for the reduction of compound potency.
Due to the increase in performance of computers and the
development of calculation methodology, several computa-
tional analyses have recently been carried out.26−32 Quantum
mechanical calculations using the fragment molecular orbital
method were used to investigate the role of key amino acid
residues in recognizing sialoglycoproteins on the host cell
surface26 or in combining with neutralizing monoclonal
antibodies.27 Huge MD simulations were performed to examine
the interaction between HA and sialoglycans28 and clarify the
reason for mutations at the receptor-recognizing site.29 MD
simulations were also employed to evaluate the binding free
energy30 and to interpret the difference in receptor specificity.31
A computational approach was further demonstrated to be
quite helpful for designing protein peptides that strongly inhibit
the function of HA.32
Pharmaceutical Properties of Hit Chemicals. Since
stachyflin exhibits a highly potent anti-influenza activity for the
wild-type WSN strain and is a challenging synthetic target due
to its unique alkaloid structure, two research groups have so far
attempted total synthesis of stachyflin. The first total synthesis
of racemic ( )-stachyflin was reported by Taishi et al.,11 in
which the characteristic structure of 5 heterorings was built one
by one in 29 steps. The first enantioselective total synthesis of
(+)-stachyflin was recently achieved by Watanabe et al.,33
utilizing an acid-induced domino epoxide-opening, rearrange-
ment, cyclization reaction.34 These synthetic studies suggest the
possibility for producing stachyflin analogues and encourage
the development of stachyflin-based antiviral agents. The
complexity in synthesis is, however, disadvantageous from the
viewpoint of productivity in manufacturing. Carey and co-
workers analyzed the reactions used for the preparation of drug
candidate molecules,35 surveying 128 compounds produced in
the departments of process chemistry of three major
pharmaceutical companies. According to their analysis, the
average number of chemical steps for synthesizing one
candidate is 8.1. Compounds containing a chiral center account
for about half of the 128 molecules. Therefore, the structural
Figure 4. Hit chemicals found through in silico screening: (a)
compound 4 and (b) compound 5. Top: chemical structure. Middle:
binding site of the compound predicted by the docking simulation.
Bottom: binding position of the compound viewed from the direction
of subunit HA1.
introduction of benzyl via amine group (31) resulted in an
increase in inhibitory activity. In contrast, conversion into
dimethylamine (32, 33) resulted in loss of compound potency.
Interestingly, while conversion into methyl-piperazine (34)
resulted in a decrease in inhibitory activity, its hydrochloride
salt (35) exhibited a high compound potency. Phenyl-
piperazine (36) also exhibited a considerably high inhibitory
activity. The conversion of thiophene of compound 5 into
benzene (37) exhibited a high compound potency, while its
chiral analogue (38) showed no inhibitory activity. This chiral-
selective compound potency was confirmed by substitution of a
Boc protection group (39) for the dimethylamine of compound
37.
Actions of Active Compounds for Blocking HA. Fusion
of the membrane is an essential process in the entry of
influenza virus into the host cell. HA mediates this process
through two functions in the early stage of the viral life cycle.17
One is anchoring at sialylated glycoprotein receptors on the cell
membrane surface. The other is the low-pH induced conforma-
tional change that initiates the exposure of a fusion peptide, a
hydrophobic N-terminal segment buried in the HA trimer
interface, to be inserted into the endosomal membrane.
Stachyflin is assumed to block this fusogenic process of
HA.12,13 Indeed, in the predicted binding structure shown in
Figure 2, stachyflin is combined with HA with the formation of
two strong hydrogen bonds and three strong aromatic ring-
involved hydrophobic interactions.
Several other compounds blocking the fusogenic activity of
HA have been identified.18−22 An active compound18 bears a
naphthoquinone skeleton, which is commonly seen in
stachyflin as rings AB. Another compound19 contains a
quinolizin moiety connected to a benzamide skeleton, and a
one-step virus growth experiment indicated that this compound
mainly inhibited virus proliferation at the early stage of the
replication cycle. Therefore, a quinoline skeleton is one of the
effective chemical cores for binding to HA. A compound
557
dx.doi.org/10.1021/cb200332k | ACS Chem. Biol. 2012, 7, 552−562

ACS Chemical Biology
Articles
a
Table 2. Structures and Inhibitory Activities of the Synthesized Analogues to Compound 4, Which Bears a Vanillic Acid Core
a
Condition: (a) K2CO3, R-Cl, CH3CN, temp, time, 21−93%.
core of stachyflin would not be suitable for scale-up synthesis in
terms of the number of synthetic steps and the number of chiral
centers. In this study, we provided two scaffolds exhibiting
antiviral activity. Most of the compounds shown in Tables 2
and 3 were synthesized within 3 steps, except for compounds
37−39. Compounds 37−39 include chiral centers, and the
synthesis of these compounds was achieved at most within 8
steps. Accordingly, the proposed scaffolds are feasible for
diverse modulation of functional groups and then ones of the
chemical bases for the development of HA inhibitors.
A requirement of antiviral drugs in clinical use is inhibitory
activity for a broad range of viral strains. We tested the potency
of several compounds synthesized in this study using A/
Vietnam/1194/2004 (H5N1) strain,36 which causes severe
pathological conditions for humans and is one of the viruses
attracting keen concern for the threat of a pandemic. A cell-
based antiviral assay indicated that compounds 31 and 39 were
effective for this H5N1 type virus with IC50 values of 0.2 and
5.2 μM, respectively. This suggests that the scaffolds found in
this study will maintain compound potency over different types
of influenza viruses.
three features of aromatic ring, hydrophobic region, and
hydrogen-bond acceptor. Hit chemical compound 5 is
compatible in all four features. All of the 8 selected compounds
are compatible with stachyflin at least in two features, aromatic
ring and hydrogen-bond acceptor. That is, these two features
are commonly observed in pharmacophores of all selected
compounds. This suggests that aromatic ring and hydrogen-
bond acceptor are indispensable for inhibitors analogous to
stachyflin. From the viewpoint of compatibility in the
pharmacophore, compound 5 is the most advantageous for
an HA inhibitor. In the binding modes of compounds 4 and 5
to the wild-type HA shown in Figure 3, compound 4 generated
three hydrogen bonds with HA. In contrast, one hydrogen
bond was observed between compound 5 and HA. The binding
position of compound 5 is considerably close to that of
stachyflin. These findings in binding mode suggest that the
shape complementarity between compound 5 and the binding
site in HA is high, while electrostatic complementarity is more
significant in the binding of compound 4.
Methylester of compound 4 corresponds to carbonyl oxygen
on ring E of stachyflin, and methoxy corresponds to the
hydroxy group bound to ring D. It should be noted that rings
DE of stachyflin compose a large flat region and that the vanillic
acid core of compound 4 mimics the flat region and the
distribution of polar atoms. Diazolane of compound 4 is
compatible with ring A of stachyflin, and amine on the
Plan of the Design of Potent Agents. In the in silico
screening carried out in this study, 4 features, i.e., hydrophobic
region, hydrogen-bond donor, hydrogen-bond acceptor, and
aromatic ring, were monitored for pharmacophore. Hit
chemical compound 4 is compatible with stachyflin in the
558
dx.doi.org/10.1021/cb200332k | ACS Chem. Biol. 2012, 7, 552−562

ACS Chemical Biology
Articles
region. Another thiophene and dimethylamine are compatible
with rings A and B of stachyflin, respectively. The high
inhibitory activity of compound 35 in Table 3 suggests that
thiophene is not necessarily required to maintain compound
potency. This finding provides a sound explanation for the high
inhibitory activity of compound 9. Accordingly, the region of
ring A of stachyflin is not so important for inhibitory activity.
Instead, the flat region at rings DE and the charge distribution
on the flat region are essential for compound potency.
Table 3. Structures and Inhibitory Activities of the
Synthesized Analogues to Compound 5, Which Bears a
Thieno-pyrimidine Core
a
Compound 9 and 35 were docked to the wild-type HA and
the three mutants in the same manner as stachyflin was. The
predicted binding structures in Supplementary Figure S8
indicate that the binding site of compound 9 changed among
the models. In contrast, compound 35 was bound to almost the
same location among all models. The thieno-pyrimidine moiety
is positioned near D109. This position is, however, different
from that of stachyflin. The sizes of compounds 9 and 35 are
small compared to stachyflin and compounds 4 and 5. Hence
this smallness may be a reason for the incompatibility of the
binding mode. To produce potent compounds appropriately
fitted to the domain between inner helices of HA, compounds
9 and 35 should be converted with retention of the vanillic acid
core and/or the thieno-pyrimidine moiety and with attachment
of some chemical group containing a chiral center at the
opposite side to cling to the helix round surface.
METHODS
■
Molecular Dynamics Simulation. Information on the structural
difference between the wild-type HA and the mutants is essential for
clarifying the reason why some amino acid mutations in HA diminish
the inhibitory activity of stachyflin. The initial structure of the wild-
type HA of WSN strain was constructed by homology modeling using
Modeler ver. 9.4.37 The multiple-alignment technique was employed,
where the X-ray crystal structures of A/Puerto Rico/8/1934 H1N1
with PDB codes 1RVZ and 1RU738 and A/Brevig Mission/1/1918
H1N1 with code 1RD839 were selected for references in modeling.
Homology modeling was also performed to build the initial structures
for variants; K51R and K121E mutations were introduced in the HA2
subunit in mutant 1, V176I in HA1 and K51R, K121E in HA2 in
mutant 2, and V176I in HA1 and K51R in HA2 in mutant 3. A lipid
bilayer of 100 Å × 100 Å was generated by using an in-house software
GLYMM implemented in VMD ver. 1.840 for the purpose of
embedding the C-terminal side of HA2 subunit into the lipid
membrane mimicking the viral envelope. The composition of lipid
molecules in the membrane was set to be as compatible as possible
with the composition of lipid molecules in the influenza viral envelope,
as shown in Supplementary Table 1. Judging from the prediction
results with UniProtKB,41 amino residues 186−206 were assumed to
be the transmembrane region embedded in the viral envelope. It is
natural to consider that an α helix structure is formed in the
transmembrane region. However, there is no experimental ground for
the formation of an α helix in this region. Accordingly, this region was
set to have no secondary structure, that is, to be in a strand form in the
initial structure of the present simulations. In our preliminary
calculation without embedding the C-terminal side of HA2 subunit
into the lipid membrane, each helix in HA2 gradually changed its
conformation to separate itself from other helices. HA trimer was
converted from a closed shape to an open one with the progress of
simulation. The motion of the C-terminal region of HA2 was large in
the preliminary calculation, which seemed to be a cause of structural
instability and became a trigger for the drastic conformational change.
That is, the helices of the HA trimer cannot maintain the closed form
unless the C-terminal side of HA2 is embedded in the lipid membrane.
Hence, calculation models in this work included the membrane
mimicking the viral envelope and the transmembrane region of HA.
TIP3P water molecules and ions to neutralize the calculation cell were
generated to solvate the complex of HA and lipid membrane, making a
a
Condition: (b) amine, NaOH, THF, reflux, 6 h.
diazolane corresponds to the hydroxy group bound to ring A.
Compound 9 in Table 2, in which diazolane is replaced by
methoxy, shows a considerably high inhibitory activity.
Therefore, a polar group at the position of the diazolane is
not necessarily required to maintain compound potency.
Sulfur and nitrogen atoms of thieno-pyrimidine moiety of
compound 5 mimic the charge distribution of rings DE. It is
notable that thieno-pyrimidine also composes a large flat
559
dx.doi.org/10.1021/cb200332k | ACS Chem. Biol. 2012, 7, 552−562

ACS Chemical Biology
Articles
periodic boundary box of 100 Å × 100 Å × 250 Å in which the top and
bottom parts of HA were set apart from the boundary by more than 10
Å (Figure 1c). Consequently, the total number of atoms was about
259,300 in each model.
the combined organic layer was washed with brine, dried over MgSO4,
and concentrated in vacuo. The product was purified by thin-layer
chromatography with hexane/EtOAc in a ratio of 3:1. The solid
obtained was resuspended with a solution of hexane and EtOAc, and
recrystallization produced the final compound as a white solid (1.01 g,
yield 71%).
MD simulations were carried out for every model using NAMD ver.
2.6.42 Initially, energy minimization was executed for 1,000,000 steps
with the conjugate gradient method. Next, the temperature of the
model system was elevated up to 310 K. Then 30 ns equilibrating
simulation was performed in the NTP ensemble condition to obtain
the equilibrated structures of the wild-type HA and the three kinds of
mutants. Nonbonded interaction terms were computed with a cutoff
distance of 12 Å, where a switching distance of 10 Å was applied to
make the nonbonded interaction zero at the cutoff distance smoothly.
A periodic boundary condition was applied to all directions of the
calculation cell, and the particle mesh Ewald method was employed to
compute the long-distance nonbonded interaction. CHARMM27 force
field43 was adopted for all atoms.
Docking Simulation. The binding modes of stachyflin to the wild-
type HA and its mutants were predicted by docking simulation using
GOLD ver. 4.44,45 The equilibrated structure obtained from the MD
simulation was utilized as a plausible protein structure of HA in the
wild-type and three mutants. Binding score was also calculated to
evaluate the difference in binding affinity due to the mutations. A
preliminary docking calculation was executed to search for the docking
area within 30 Å from Phe110 of the HA2 subunit. Since an adequate
docking space was found inside the area and stachyflin was positioned
near Asp109 in the binding mode ranking first, recalculation of
stachyflin docking was performed with the search area set within 10 Å
from Asp109 of HA2 subunit. Fifty binding poses were generated and
the binding affinities of those binding poses were estimated by GOLD
score function. On the basis of the ranking in the estimated GOLD
score, the most probable binding pose was selected. The binding
affinity for the selected binding pose was re-estimated using ASP score
function. This two-step approach, i.e., determination of the binding
pose with GOLD score and subsequent estimation of binding affinity
with ASP score, was reported to successfully provide reliable
prediction in docking of low-molecular-weight ligands to an enzyme
or receptor.46,47
In Silico Screening. A search for compounds bearing chemical
features similar to those of stachyflin was made by an in silico
pharmacophore screening. A chemical database was provided by
Namiki Co. Ltd., in which about 3 million synthesized compounds are
listed and all of the compounds are available by purchase. First,
conformations of every compound were generated by using OMEGA
module of OpenEye software.48 Totally, more than 200 million
chemical conformations were generated. Second, the pharmacophore
of stachyflin was extracted for setting queries, in which hydrogen-bond
donor and acceptor, aromatic ring, and hydrophobic region appeared
to be key features. Third, chemical screening was carried out from the
viewpoint of structural similarity to stachyflin using ROCS module of
OpenEye.49 A total of 5094 compounds were extracted from the
Namiki database under the condition of the Tanimoto coefficient
being more than 0.75. Fourth, more condensed selection of chemicals
from the 5094 compounds was performed using EON module49 from
the viewpoint of similarity in charge distribution. The compounds
without structural flexibility were excluded. Consequently, 8 chemical
compounds were selected as candidates for purchase.
Synthesis of Analogue Compounds. Two series of derivatives
were synthesized in this work. One is an analogue containing a vanillic
acid skeleton, and the other is one bearing thieno-pyrimidine.
Compound 9, 3-methoxy-4-[(methyl-sulfonyl)oxy]-benzoic acid me-
thoxy-ester, is a typical derivative of the former series. A mixture of
vanillic acid methyl (1.0 g, 5.49 mmol), KCO3 (1.13 g, 8.24 mmol),
and acetonitrile (30 mL) was cooled in an ice bath under Ar
atmosphere. Methane-sulfonyl chloride (0.51 mL, 6.59 mmol) was
slowly added to the mixture, and then the solution was mechanically
stirred for 3 h at RT. The reaction mixture was filtered with Celite, and
the solvent was removed in vacuo. The resulting product was extracted
with a solution of EtOAc (30 mL) and 1 M aqueous HCl (30 mL).
The aqueous layer was treated with EtOAc (20 mL) two times, and
A typical derivative of the latter series is compound 35, 4-(4-methyl-
1-pyperazinyl)-thieno[2,3-d]pyrimidine hydrochloride salt. 4-Chloro-
thieno[2,3-d]pyrimidine (200 mg, 1.17 mmol) was solvated with THF
(15 mL). After addition of 1-methyl-pyperazine (0.26 mL, 2.34 mmol)
and NaOH (0.94 g, 2.34 mmol), the mixture was heated under reflux
for 6 h. The solvent was removed in vacuo, and the reaction mixture
was treated with a solution of EtOAc (20 mL) and distilled H2O (15
mL). The product was extracted with EtOAc two times, and the
organic layer was washed with brine and dried over Na2SO4. The
solvent was evaporated in vacuo, and the resulting product was purified
by two-dimensional thin-layer chromatography with hexane/EtOAc in
ratios of 3:1 and 1:1. A brown solid of (phenyl pyperazinyl)-theino-
pyrimidine was obtained in a yield of 54% (148 mg). The solid
obtained was resuspended in a solution of toluene (20 mL) and 1 M
aqueous HCl (0.6 mL) and heated under reflux for 1 h. The reaction
solution was cooled to RT, and filtration gave the final compound as a
brownish solid (127 mg, yield 45%).
Antiviral Assay. Compound potency was tested by an influenza
virus cell culture assay with measurement of the quantity of viral RNA
using real-time polymerase chain reaction (RT-PCR). Test com-
pounds were mixed with minimum essential medium (MEM)
containing bovine serum albumin (BSA). To prepare a virus-
containing compound-mixed medium, 100 units of 50% tissue culture
infective dose (TCID50) of influenza virus A/Perto Rico/8/34 strain
(PR8) was suspended in 100 μL of the MEM-BSA containing test
compounds and 10 μg/mL of acetylated trypsin. Madin-Darby canine
kidney (MDCK) cells were loaded in a 96-well plate. The cells were
washed with phosphate-buffered saline (PBS), followed by 0.5−1 h
incubation with compound-mixed medium. The virus-containing
compound-mixed medium was also incubated for 0.5−1 h. After
incubation, MDCK cells in the compound-mixed medium without
viruses were transferred to the compound-mixed medium with viruses.
Then the cells were incubated for 1 h at 37 °C in 100 μL of the virus-
containing compound-mixed medium (100 TCID50 influenza virus,
10 μg/mL acetylated trypsin, and test compound at several
concentrations). MDCK cells were washed with compound-mixed
medium without viruses and incubated in the compound-mixed,
acetylated trypsin-containing medium without influenza virus for 24 h.
Culture supernatants of MDCK cells were collected after the
incubation, and RNA was extracted from the supernatants. The
amount of viral RNA was measured by the RT-PCR method. The
measurement was compared to that of the control that was performed
in a similar manner without any test compound. The compound
concentration to suppress viral proliferation to 50% (EC50) was
estimated from the comparison.
RNA Extraction and RT-PCR. To monitor the efficiency of RNA
purification, uninfected VeroE6 cells were mixed in the ISOGEN
reagent as a source of 18S rRNA for normalization. Supernatants from
MDCK cell culture medium were mixed with ISOGEN reagent and
RNA was purified according to the manufacturer’s protocol. For
quantification of PR8 HA RNA, real-time RT-PCR was performed
using the primers and the probe with the sequences of PR8-HA-F: 5′-
GGCAAATGGAAATCTAATAGCACC-3′, PR8-HA-R: 5′-
TGATGCTTTTGAGGTGATGA-3′, and PR8-HA-probe: 5′-FAM-
TCGCACTGAGTAGAGGCTTTGGGTCC-TAMRA-3′. For moni-
toring efficiency of RNA purification, 18S rRNA was quantified using
the primers and the probe with the sequences of 18S-F: 5′-
GTAACCCGTTGAACCCCATT-3′, 18S-R: 5′-CCATCCAATCGG-
TAGTAGCG-3′, and 18S-probe: 5′-FAM-TGCGTTGAT-
TAAGTCCCTGCCCTTTGTA-TAMRA-3′. The intensity of fluo-
rescence emitted from the probe was detected by the ABI-7700
sequence detector system (Applied Biosystems).
560
dx.doi.org/10.1021/cb200332k | ACS Chem. Biol. 2012, 7, 552−562

ACS Chemical Biology
Articles
2007−2008 season in Nara Prefecture, Japan. Jpn. J. Infect. Dis 61,
253−254.
ASSOCIATED CONTENT
* Supporting Information
■
S
(10) Kamigauchi, T.; Fujiwara, T.; Tani, H.; Kawamura, Y.; Horibe, I.
Shionogi & Co, Ltd. Sesquiterpene derivatives having antiviral activity,
Patent WO/1997/011947, 1997.
Composition of lipid membrane models, predicted binding
affinities, RMSDs during MD simulation, predicted binding
structures of stachyflin, superimposition of HA structures,
chemical properties of synthesized compounds, and their NMR
spectra. This material is available free of charge via the Internet
at http://pubs.acs.org.
(11) Taishi, T., Takechi, S., and Mori, S. (1998) First total synthesis
of ( )-stachyflin. Tetrahedron Lett. 39, 4347−4350.
(12) Yoshimoto, J., Kakui, M., Iwasaki, H., Fujiwara, T., Sugimoto,
H., and Hattori, N. (1999) Identification of a novel HA conforma-
tional change inhibitor of human influenza virus. Arch. Virol. 144, 865−
878.
AUTHOR INFORMATION
Corresponding Author
*E-mail: (N.Y.) n-yama-5@nih.go.jp; (T.H.) hoshino@chiba-u.
■
(13) Yoshimoto, J., Kakui, M., Iwasaki, H., Sugimoto, H., Fujiwara,
T., and Hattori, N. (2000) Identification of amino acids of influenza
virus HA responsible for resistance to a fusion inhibitor, Stachyflin.
Microbiol. Immunol. 44, 677−685.
(14) Rosenbaum, D. M., Cherezov, V., Hanson, M. A., Rasmussen, S.
G. F., Thian, F. S., Kobilka, T. S., Choi, H. J., Yao, H. J., Weis, W. I.,
Stevens, R. C., and Kobilka, B. K. (2007) GPRC engineering yields
high-resolution structural insights into β2-adrenergic receptor
function. Science 318, 1266−1273.
(15) Wegener, A. A., Klare, J. P., Engelhard, M., and Steinhoff, H. J.
(2001) Structural insights into the early steps of receptor-transducer
signal transfer in archaeal pototaxis. EMBO J. 20, 5312−5319.
(16) Sato, Y., Hata, M., Neya, S., and Hoshino, T. (2005)
Computational analysis of the transient movement of helices in
sensory rhodopsin II. Protein Sci. 14, 183−192.
(17) Cross, K. J., Burleigh, L. M., and Steinhauer, D. A. (2001)
Mechanisms of cell entry by influenza virus. Exp. Rev. Mol. Med. 3, 1−
15.
(18) Bodian, D. L., Yamasaki, R. B., Buswell, R. L., Stearns, J. F.,
White, J. M., and Kuntz, I. D. (1993) Inhibition of the fusion-inducing
conformational change of influenza hemagglutinin by benzoquinones
and hydroquinones. Biochemistry 32, 2967−2978.
(19) Luo, G., Colonno, R., and Krystal, M. (1996) Characterization
of hemagglutinin-specific inhibitor of influenza A virus. Virology 226,
66−76.
(20) Plotch, S. J., O’Hara, B., Morin, J., Palant, O., LaRocque, J.,
Bloom, J. D., Lang, S. A. Jr., DiGrandi, M. J., Bradley, M., Nilakantan,
R., and Gluzman, Y. (1999) Inhibition of influenza A virus replication
by compounds interfering with the fusogenic function of the viral
hemagglutinin. J. Virol. 73, 140−151.
(21) Staschke, K. A., Hatch, S. D., Tang, J. C., Hornback, W. J.,
Munroe, J. E., Colacino, J. M., and Muesing, M. A. (1998) Inhibition
of influenza virus hemagglutinin-mediated membrane fusion by a
compound related to podocarpic acid. Virology 248, 264−274.
(22) Hoffman, L. R., Kuntz, I. D., and White, J. M. (1997) Structure-
based identification of an inducer of the low-pH conformational
change in the influenza virus hemagglutinin: Irreversible inhibition of
infectivity. J. Virol. 71, 8808−8820.
(23) Matsubara, T., Onishi, A., Saito, T., Shimada, A., Inoue, H., Taki,
T., Nagata, K., Okahata, Y., and Sato, T. (2010) Sialic acid-mimic
peptides as hemagglutinin inhibitors for anti-influenza therapy. J. Med.
Chem. 53, 4441−4449.
(24) Feng, F., Miura, N., Isoda, N., Sakoda, Y., Okamatsu, M., Kida,
H., and Nishimura, S. (2010) Novel trivalent anti-influenza reagent.
Bioorg. Med. Chem. Lett. 20, 3772−3776.
(25) Tang, G., Lin, X., Qiu, Z., Li, W., Zhu, L., Wang, L., Li, S., Li, H.,
Lin, W., Yang, M., Guo, T., Chen, L., Lee, D., Wu, J. Z., and Yang, W.
(2011) Design and synthesis of benzenesulfonamide derivatives as
potent anti-influenza hemagglutinin inhibitors. ACS Med. Chem. Lett. 2,
603−607.
(26) Sawada, T., Fedorov, D. G., and Kitaura, K. (2010) Role of the
key mutation in the selective binding of avian and human influenza
hemagglutinin to sialosides revealed by quantum-mechanical calcu-
lations. J. Am. Chem. Soc. 132, 16862−16872.
(27) Takematsu, K., Fukuzawa, K., Omagari, K., Nakajima, S.,
Nakajima, K, Mochizuki, Y., Nakano, T., Watanabe, H., and Tanaka, S.
(2009) Possibility of mutation prediction of influenza hemagglutinin
jp.
Author Contributions
#These authors contributed equally to this work.
ACKNOWLEDGMENTS
■
We thank L.Q. Mai (National Institute of Hygiene and
Epidemiology, Hanoi) for providing the A/Vietnam/1194/04
virus strain. This work was partly supported by a Health and
Labor Science Research Grant for Research from the Ministry
of Health and Labor of Japan. This work was also supported by
a Grant-in-Aid for Scientific Research (C) from Japan Society
for the Promotion of Science (JSPS). This study was also
supported in part by a grant-in-aid (S0991013) from the
Ministry of Education, Culture, Sport, Science, and Technol-
ogy, Japan (MEXT) for the Foundation of Strategic Research
Projects in Private Universities. Theoretical calculations were
performed at the Research Center for Computational Science,
Okazaki, Japan and at the Information Technology Center of
the University of Tokyo and also by the high-performance
computer system at Institute for Media Information Technol-
ogy in Chiba University.
REFERENCES
■
(1) Gambotto, A., Barratt-Boyes, S. M., de Jong, M. D., Neumann, G.,
and Kawaoka, Y. (2008) Human infection with highly pathogenic
H5N1 influenza virus. Lancet 371, 1464−1475.
(2) Beigel, J. H., Farrar, J., Han, A. M., Hayden, F. G., Hyer, R., de
Jong, M. D., Lochindarat, S., Nguyen, T. K. T., Nguyen, T. H., Tran, T.
H., Nicoll, A., Touch, S., Yuen, K. Y., and Writing Committee of the
World Health Organization Consultation on Human Influenza A/H5.
(2005) Avian influenza A (H5N1) infection in humans. N. Engl. J.
Med. 353, 1374−1385.
(3) Schnell, J. R., and Chou, J. J. (2008) Structure and mechanism of
the M2 proton channel of influenza A virus. Nature 451, 591−595.
(4) Stouffer, A. L., Acharya, R., Salom, D., Levine, A. S., Di Costanzo,
L., Soto, C. S., Tereshko, V., Nanda, V., Stayrook, S., and DeGrado, W.
F. (2008) Structural basis for the function and inhibition of an
influenza virus proton channel. Nature 451, 596−599.
(5) McNicholl, I. R., and McNicholl, J. J. (2001) Neuraminidase
inhibitors: zanamivir and oseltamivir. Ann. Pharmacother. 35, 57−70.
(6) Moscona, A. (2005) Neuraminidase inhibitors for influenza. N.
Engl. J. Med. 353, 1363−1373.
(7) Bright, R. A., Medina, M.-j., Xu, X., Perez-Oronoz, G., Wallis, T.
R., Davis, X. M., Povinelli, L., Cox, N. J., and Klimov, A. I. (2005)
Incidence of adamantane resistance among influenza A (H3N2)
viruses isolated worldwide from 1994 to 2005: a cause for concern.
Lancet 366, 1175−1181.
(8) Deyde, V. M., Xu, X., Bright, R. A., Shaw, M., Smith, C. B., Zhang,
Y., Shu, Y., Gubareva, L. V., Cox, N. J., and Klimov, A. I. (2007)
Surveillance of resistance to adamantanes among influenza A(H3N2)
and A(H1N1) viruses isolated worldwide. J. Infect. Dis. 196, 249−257.
(9) Kitahori, Y., Imanishi, Y., and Inoue, Y. (2008) High incidence of
amantadine-resistant influenza H1N1 viruses isolated during the
561
dx.doi.org/10.1021/cb200332k | ACS Chem. Biol. 2012, 7, 552−562

ACS Chemical Biology
Articles
by combination of hemadsorption experiment and quantum chemical
calculation for antibody binding. J. Phys. Chem. B. 113, 4991−4994.
(28) Newhouse, E. I., Xu, D., Markwick, P. R., Amaro, R. E., Pao, H.
C., Wu, K. J., Alam, M., McCammon, J. A., and Li, W. W. (2009)
Mechanism of glycan receptor recognition and specificity switch for
avian, swine, and human adapted influenza virus hemagglutinins: a
molecular dynamics perspective. J. Am. Chem. Soc. 131, 17430−17442.
(29) Kasson, P. M., Ensign, D. L., and Pande, V. S. (2009)
Combining molecular dynamics with bayesian analysis to predict and
evaluate ligand-binding mutations in influenza hemagglutinin. J. Am.
Chem. Soc. 131, 11338−11340.
(45) Jones, G., Willett, P., and Glen, R. C. (1995) Molecular
recognition of receptor sites using a genetic algorithm with a
description of desolvation. J. Mol. Biol. 245, 43−53.
(46) Verdonk, M. L., Cole, J. C., Hartshorn, M. J., Murray, C. W., and
Taylor, R. D. (2003) Improved protein-ligand docking using GOLD.
Proteins: Struct., Funct., Genet. 52, 609−623.
(47) Mooij, W. T. M., and Verdonk, M. L. (2005) General and
targeted statistical potentials for protein-ligand interactions. Proteins:
Struct., Funct., Genet. 61, 272−281.
(48) Bostrom, J., Greenwood, J. R., and Gottfries, J. (2003) Assessing
the performance of OMEGA with respect to retrieving bioactive
conformations. J. Mol. Graphics Modell. 21, 449−462.
(49) OpenEye Scientific Software, Inc., Santa Fe, NM, USA, www.
eyesopen.com.
(30) Das, P., Li, J., Royyuru, A. K., and Zhou, R. (2009) Free energy
simulations reveal a double mutant avian H5N1 virus hemagglutinin
with altered receptor binding specificity. J. Comput. Chem. 30, 1654−
1663.
(31) Xu, D., Newhouse, E. I., Amaro, R. E., Pao, H. C., Cheng, L. S.,
Markwick, P. R., McCammon, J. A., Li, W. W., and Arzberger, P. W.
(2009) Distinct glycan topology for avian and human sialopenta-
saccharide receptor analogues upon binding different hemagglutinins:
a molecular dynamics perspective. J. Mol. Biol. 387, 465−491.
(32) Fleishman, S. J., Whitehead, T. A., Ekiert, D. C., Dreyfus, C.,
Corn, J. E., Strauch, E. M., Wilson, I. A., and Baker, D. (2011)
Computational design of proteins targeting the conserved stem region
of influenza hemagglutinin. Science 332, 816−821.
(33) Watanabe, K., Sakurai, J., Abe, H., and Katoh, T. (2010) Total
synthesis of (+)-stachyflin: a potential anti-influenza A virus agent.
Chem. Commun. 46, 4055−4057.
(34) Nakatani, M., Nakamura, M., Suzuki, A., Inoue, M., and Katoh,
T. (2002) A new strategy toward the total synthesis of stachyflin, a
potent anti-influenza A virus agent: concise route to the tetracyclic
core structure. Org. Lett. 4, 4483−4486.
(35) Carey, J. S., Laffan, D., Thomson, C., and Williams, M. T.
(2006) Analysis of the reactions use for the preparation of drug
candidate molecules. Org. Biomol. Chem. 4, 2337−2347.
(36) Mizukami, T., Imai, J., Hamaguchi, I., Kawamura, M., Momose,
H., Naito, S., Maeyama, J., Masumi, A., Kuramitsu, M., Takizawa, K.,
Nomura, N., Watanabe, S., and Yamaguchi, K. (2008) Application of
DNA microarray technology to influenza A/Vietnam/1194/2004
(H5N1) vaccine safety evaluation. Vaccine 26, 2270−2283.
(37) Sali, A., and Blundell, T. L. (1993) Comparative protein
modelling by satisfaction of spatial restraints. J. Mol. Biol. 234, 779−
815.
(38) Gamblin, S. J., Haire, L. F., Russell, R. J., Stevens, D. J., Xiao, B.,
Ha, Y., Vasisht, N., Steinhauer, D. A., Daniels, R. S., Elliot, A., Wiley,
D. C., and Skehel, J. J. (2004) The structure and receptor binding
properties of the 1918 influenza hemagglutinin. Science 303, 1838−
1842.
(39) Stevens, J., Corper, A. L., Basler, C. F., Taubenberger, J. K.,
Palese, P., and Wilson, I. A. (2004) Structure of the uncleaved human
H1 hemagglutinin from the extinct 1918 influenza virus. Science 303,
1866−1870.
(40) Humphrey, W., Dalke, A., and Schulten, K. (1996) VMD: visual
molecular dynamics. J. Mol. Graph. 14, 33−38.
(41) The UniProt Consortium. (2011) Ongoing and future
developments at the Universal Protein Resource. Nucleic Acids Res.
39, D214−D219.
(42) Phillips, J. C., Braun, R., Wang, W., Gumbart, J., Tajkhorshid, E.,
́
Villa, E., Chipot, C., Skeel, R. D., Kale, L., and Schulten, K. (2005)
Scalable molecular dynamics with NAMD. J. Comput. Chem. 26, 1781−
1802.
(43) Feller, S. E., Gawrisch, K., and MacKerell, A. D. (2002)
Polyunsaturated fatty acids in lipid bilayers: intrinsic and environ-
mental contributions to their unique physical properties. J. Am. Chem.
Soc. 124, 318−326.
(44) Jones, G., Willett, P., Glen, R. C., Leach, A. R., and Taylor, R.
(1997) Development and validation of a genetic algorithm for flexible
docking. J. Mol. Biol. 267, 727−748.
562
dx.doi.org/10.1021/cb200332k | ACS Chem. Biol. 2012, 7, 552−562