A nanomolar multivalent ligand as entry inhibitor of the hemagglutinin of avian influenza

Article abstract of DOI:10.1021/ja410918a

Influenza virus attaches itself to sialic acids on the surface of epithelial cells of the upper respiratory tract of the host using its own protein hemagglutinin. Species specificity of influenza virus is determined by the linkages of the sialic acids. Birds and humans have α2-3 and α2-6 linked sialic acids, respectively. Viral hemagglutinin is a homotrimeric receptor, and thus, tri- or oligovalent ligands should have a high binding affinity. We describe the in silico design, chemical synthesis and binding analysis of a trivalent glycopeptide mimetic. This compound binds to hemagglutinin H5 of avian influenza with a dissociation constant of K _D = 446 nM and an inhibitory constant of K_I = 15 μM. In silico modeling shows that the ligand should also bind to hemagglutinin H7 of the virus that causes the current influenza outbreak in China. The trivalent glycopeptide mimetic and analogues have the potential to block many different influenza viruses.

Full text of DOI:10.1021/ja410918a

Article
pubs.acs.org/JACS
A Nanomolar Multivalent Ligand as Entry Inhibitor of the
Hemagglutinin of Avian Influenza
Moritz Waldmann,† Raffael Jirmann,† Ken Hoelscher,† Martin Wienke,† Felix C. Niemeyer,†
Dirk Rehders,‡ and Bernd Meyer*,†
†Department of Chemistry, University of Hamburg, 20146 Hamburg, Germany
‡Joint Laboratory for Structural Biology of Infection and Inflammation, Institute of Biochemistry and Molecular Biology, University of
Hamburg and Institute of Biochemistry, University of Luebeck, c/o Deutsches Elektronen-Synchrotron (DESY), Notkestrasse 85,
22607 Hamburg, Germany
S
* Supporting Information
ABSTRACT: Influenza virus attaches itself to sialic acids on the
surface of epithelial cells of the upper respiratory tract of the host
using its own protein hemagglutinin. Species specificity of influenza
virus is determined by the linkages of the sialic acids. Birds and
humans have α2−3 and α2−6 linked sialic acids, respectively. Viral
hemagglutinin is a homotrimeric receptor, and thus, tri- or
oligovalent ligands should have a high binding affinity. We describe
the in silico design, chemical synthesis and binding analysis of a
trivalent glycopeptide mimetic. This compound binds to hemag-
glutinin H5 of avian influenza with a dissociation constant of KD =
446 nM and an inhibitory constant of KI = 15 μM. In silico modeling shows that the ligand should also bind to hemagglutinin H7
of the virus that causes the current influenza outbreak in China. The trivalent glycopeptide mimetic and analogues have the
potential to block many different influenza viruses.
published by Wiley et al. in 1981. It triggered an extensive
search for small molecules with high affinity for hemagglutinin.7
In fact, the search proved to be difficult, and despite the more
than 30 years of search, no competitive inhibitor with
therapeutic potential was found. This is due to the weak
interaction of hemagglutinin to its natural ligand sialic acid.8
The strong interaction between virus and host cell is caused by
multivalent binding of the homotrimeric hemagglutinin
molecules to the numerous carbohydrate moieties on the
surface of the host cell.9 Polymeric multivalent ligands were
shown to exert tight binding to hemagglutinin.10
INTRODUCTION
■
Avian influenza has become a serious threat to human health.
The infection is species specific and depends on the glycosidic
linkage of the sialic acids in the upper respiratory tract. In the
upper respiratory tract, humans have predominantly α2−6
linked sialic acid carbohydrate moieties, while birds carry α2−3
linked sialic acids. The surface protein hemagglutinin of
influenza is either specific for α2−3 linkages, e.g., H5 and
H7, or α2−6 linkages, e.g., H1, H2, H3.1 The first outbreak of
avian influenza was caused by a viral influenza strain that has
the H5 hemagglutinin on its surface.2 Infections of humans
with avian flu have very recently been shown to also occur with
viruses carrying the H7 hemagglutinin. This new virus found in
China seems to be much more infectious to humans than the
previous H5N1 virus. Furthermore, a highly pathogenic strain
like H5N1 with a high mortality rate can undergo genetic
reassortment with a strain that has an effective human-to-
human transmission like the swine influenza H1N1. Those
viruses would pose a serious pandemic threat.3
Antiviral therapeutics are available targeting two of the three
major influenza virus surface proteins: the neuraminidase and
the M2 ion channel.4 No compound targeting the third protein,
hemagglutinin, at the binding pocket has been approved as
drug. However, small molecules inhibiting the membrane
fusion by binding to the interface of the HA subunits have been
developed.5 An increasing number of resistant influenza strains
makes the search for new therapeutics indispensable.4c,6 The
crystal structure of an influenza hemagglutinin was first
RESULTS AND DISCUSSION
■
Here we describe the in silico design, synthesis and in vitro
analysis of a trivalent ligand that binds influenza hemagglutinin
H5 of avian influenza in the high nanomolar range. As a species
highly pathogenic to humans avian influenza is characterized by
an unusual high mortality rate (∼60%) compared to that of
seasonal influenza (0.01−0.001%), we chose H5 of avian
influenza as model system.11 On the basis of the crystal
structures of avian influenza H5 published in 2001 and 2006,
we designed in silico ligands and calculated their binding
energies.12 The position of the binding pocket at the top of the
globular domain distal to the virus membrane allowed the
development of a trimeric structure with a centrally positioned
Received: November 7, 2013
Published: December 30, 2013
© 2013 American Chemical Society
783
dx.doi.org/10.1021/ja410918a | J. Am. Chem. Soc. 2014, 136, 783−788

Journal of the American Chemical Society
Article
Figure 1. Crystal structures of H5 and H7 trimers shown as surface plots with ligands. (a) Sialic acids (yellow) in the binding pockets as determined
in pdb 1JSO of hemagglutinin H5. The top of the globular domain, which is distal to the virus membrane, is shown. The hemagglutinin monomers
are colored black, brown, and white, respectively. The exposed position of the binding site is predestinated for the development of a trimeric ligand
with a centrally positioned core structure with radial topology. Top: view of the top of hemagglutinin shown at an angle. Bottom: view along the axis
of the hemagglutinin H5 trimer with the distances between the binding sites highlighted. (b) View on top of hemagglutinin H5 with ligand 1 bound
to all three binding sites at two different times of an MD simulation. The ligand is shown in atom colors. Flexibility during the dynamics simulation is
predominantly found at the core while all three sialic acid residues remain in their binding pocket. (c) View on top of the hemagglutinin H7 (from
pdb 3M5G) with ligand 1 bound to all three binding sites at two different times of an MD simulation. Motion is predominantly found at the core
structure while all three sialic acid residues remain in their binding site during the simulation indicating that 1 is also a good ligand for H7.
core which is connected to the three binding pockets (Figure
1a, top). The distance between the binding pocket and the
symmetry axis of the trimer is approximately 30 Å (Figure 1a,
bottom). The radial topology is optimal for hemagglutinin
inhibitors because of the exposed position of the binding
pockets and also thermodynamically favored over all other
topologies as shown by Bundle et al.13 Different types of
scaffolds and the influence of structure and size regarding the
same target were reviewed by Kiessling et al.14
Structural as well as thermodynamic and entropic consid-
erations led to the development of ligands with a core structure
mimicking the C3 symmetry of the protein. In silico, potential
ligands with different core structures were derived from 1,3,5-
alkoxycyclohexane, 1,3,5-alkoxybenzene or benzene-1,3,5-tri-
carboxylic acid (trimesic acid). The value of trimesic acid as
scaffold for multivalent ligands was demonstrated by White-
sides and his group in his vancomycin studies.15 In order to
achieve water solubility, peptidic fragments as well as alkyl
groups were integrated. The flexible alkyl chains act as spacers
between the core and the more rigid peptide fragments and
between the sialic acid and the peptide. We chose the flexible
alkyl chains in two positions of the trimeric ligand to
compensate a nonperfect steric match between ligand and
protein. Compared to often used ethylene glycol oligomers, a
structurally well-defined peptide segment was introduced to
reduce overall flexibility, which in turn reduces the entropic loss
from conformational flexibility during binding. An alternating
sequence of glycine and serine was chosen, which does not
form a secondary structure as verified by CD-spectroscopy of
the peptidic fragment 7 as well as of ligand 1. This sequence is
often being used as biocompatible linker in the construction of
single chain antibodies and it is not intended to assist binding.16
Alanine was introduced for practical reasons to facilitate NMR
analysis by breaking the symmetry. The length of the peptide
and of the alkyl linkers was varied and their binding energy was
assessed by docking the corresponding structures into the
binding sites and by calculating their binding energies. The
optimized ligand complexes were used to identify structural
requirements. A nine amino acid long peptide fragment gave
the best energy. Shortening the peptide chain by one amino
acid led to a significant loss of binding affinity. We could also
obtain the optimal length of the alkyl chain proximal to the core
structure. Six methylene groups yielded a flexible spacer
between the aromatic core, i.e., trimesic acid, and the peptide
linker. The sialic acid was also connected to the peptide linker
by an alkyl chain with five carbon atoms. The docking results
784
dx.doi.org/10.1021/ja410918a | J. Am. Chem. Soc. 2014, 136, 783−788

Journal of the American Chemical Society
Article
for ligands with 1,3,5-trihydroxybenzene and trimesic acid
derived cores are shown in Table 1 (for an extended version
including also cis,cis-1,3,5-trihydroxy-cyclohexane see Table S1).
To compensate the reduced flexibility of the amide bond of the
trimesic acid core, we elongated the adjacent alkyl chain by one
methylene group for this ligand for direct comparison. Some
cyclohexyl containing ligands yielded similar binding energies,
i.e., +0.2 and 1.1 kcal/mol, respectively. Due to the complexity
of modeling the interaction of a trivalent ligand with trimeric
protein, the calculated energies are considered to be estimates
of the respective real binding energy. For obvious reasons, we
decided to synthesize the ligand with the best predicted energy.
We also wanted to synthesize the 1,3,5-trihydroxy-benzene
derived ligand with the best predicted energy (Table 1).
However, because of low yields in derivatizing the 1,3,5-
trihydroxy-benzene as core, we focused on the trimesic acid
core.17
a
Table 1. Summary of the in Silico Results
After orienting chemistry, we synthesized ligand 1 with the
trimesic acid derived core structure, a hexyl side chain, a nine
amino acid long peptide sequence, another pentyl alkyl chain
and a sialic acid glycoside (cf. structure in Table 1, bottom).
The aromatic core and the anomeric center of the sialic acid are
spaced by 42 bonds to bridge the distance of 30 Å and enable
simultaneously binding of all three sialic acids. This was verified
by a 10 ns molecular dynamic (MD) simulation of the protein/
ligand complex in a water box using Desmond as part of the
Schrodinger suite. During the MD a small amount of
movement was observed for the core while all three sialic
acids remained in their respective binding pocket over the full
simulation (Figure 1b; also watch movie 1 in Supporting
Information). The complex of ligand 1 with H5 was analyzed
showing dominantly interactions between the sialic acid and
H5, which are similar to the binding mode of sialic acid to H5
in the crystal.12a Thus, the pose of the three sialic acids in ligand
1 when docked to H5 is almost identical to that of the sialic
acids in the X-ray structure with the NHAc function and the
glycerol side chain in close proximity to the protein (Figure 1b;
for a ligand-interaction plot see Figure S1 in Supporting
Information). We also find few interactions of the pentyl chain
and the two N-terminal amino acids with H5.
Ligand 1 was synthesized by a convergent approach. The
sialic acid chloride acting as glycosyl donor for the synthesis of
building block 5 was prepared by standard procedure18 and
glycosylated with 5-hexene-1-ol, which in turn was oxidized
under cleavage of the terminal double bond by ruthenium(III)
chloride yielding building block 5, which was also used for
synthesis of 6 (Figure 2a). Synthesis of the peptide was carried
out by standard protocol for microwave assisted automated
SPPS. Sialic acid building block 5 was coupled to the nascent
peptide on solid support. Glycoconjugate 8 was cleaved from
the resin with simultaneous removal of all protecting groups of
the peptide side chain. This was essential to ensure solubility
and allowing purification of the compounds. Compound 3 was
reacted with 8 to give the trivalent glycoconjugate 10. The
latter two reactions are very sensitive to the excess and
concentration of 5 and 8, respectively. Final deprotection led to
1 which was purified by RP-HPLC (Figure 2c). We synthesized
fragments of 1 to test the binding affinity and the effect of
multivalency. Compound 9 was prepared as monovalent
fragment of 1 from compound 8, which was also used for the
synthesis of 1. Additionally, compounds 6 and 7 were prepared
as fragments of 1 (Figure 2b,c). All compounds were
characterized for their binding affinity.
a
Different peptidic linker length (m) and alkyl chain length (n and o)
were docked and binding energies were calculated. The energies are
expressed relative to the ligand with the best binding energy. The
structure of ligand 1 is shown at the bottom of the table. The ligand
consists of a trimesic acid derived core structure which is bound by
flexible hexyl chains to the nonapeptide linkers. The sialic acid epitope
is bound by a flexible alkyl chain to the N-terminus of the peptidic
linker.
For the experimental binding studies, trimeric H5 (A/
Vietnam/1203/2004) was used. The thermodynamic dissocia-
tion constant KD was determined by surface plasmon resonance
(SPR) on a Biacore T100 immobilizing hemagglutinin to the
sensor chip. Evaluation of the SPR data using the 1:1 binding
model yielded KD = 450 nM (Figure 3, top). Compared to
785
dx.doi.org/10.1021/ja410918a | J. Am. Chem. Soc. 2014, 136, 783−788

Journal of the American Chemical Society
Article
Figure 2. Synthesis of ligand 1: (a) synthesis of building block 519 that was also used for synthesis of fragment 6; (b) synthesis of the peptide
fragment by standard protocol for microwave assisted solid phase peptide synthesis; (c) synthesis of ligand 1 using building blocks 3 and 5. For
synthesis of compound 3, see Supporting Information.
Neu5Acα2Me, the affinity was increased by a factor of 4000.
For a bivalent interaction, one would expect an increase of a
factor of about 500 and for a trivalent interaction a factor of
∼300 000. Thus, we reached an interaction strength that is
between bi- and trivalency. Reported avidities of multivalent
ligands for influenza hemagglutinin show enhanced binding
relative to the methyl sialoside by a factor of 100 for a bivalent
molecule20 up to a factor of 1 000 00010 for polymeric
substrates. One of the first synthetic bivalent ligands was
reported by Knowles et al. and had an 100-fold increased
affinity against whole viruses in a hemagglutination assay when
compared to Neu5Acα2Me. Interestingly, the bivalent ligand
did not bind to bromelain-released hemagglutinin because of
unknown reasons as a cross-linking of viruses was due to the
low viral concentration excluded.20 Roy and co-workers were
one of the first to report higher multivalent ligands derived
from synthetic sialylated glycoconjugates, which reduced viral
infectivity by a factor of 130.21 The tightest binding of a
sialylated polymer was reported by Whitesides et al. with an
increased binding of up to 1 000 000 compared to Neu5-
Acα2Me as assessed from IC50 values determined in an ELISA
assay.10 Unfortunately, much of this exciting research was not
continued, mainly due to an insufficient binding or the high
molecular weight and most importantly polymeric derivatives of
sialic acid will likely bind to all receptors for sialic acids and thus
have a low specificity. Here, we demonstrate the power of
custom tailored ligands for influenza hemagglutinin.
We also used STD-NMR studies with α-methyl sialoside
(Neu5Acα2Me) as competitor to determine inhibition
constants.22 For competitive studies, a 0.55 μM solution of
H5 trimer containing 20 mM Neu5Acα2Me was used. Ligand 1
was added stepwise to this solution to a final concentration of
150 μM (micelle formation was observed at higher
concentrations as verified by dynamic light scattering measure-
ments). The absolute STD% for each competitor proton signal
was plotted against the logarithm of the ligand concentration
786
dx.doi.org/10.1021/ja410918a | J. Am. Chem. Soc. 2014, 136, 783−788

Journal of the American Chemical Society
Article
made shorter without losing significant amounts of the binding
affinity (cf. Supporting Information).
Ligand 1 utilizes only the binding sites of the sialic acids on
the trimeric hemagglutinin. Therefore, we hypothesized that
the trimeric compound 1 and possibly analogs of this
compound are generically able to interact with most influenza
hemagglutinins, which are all trimeric. In the current outbreak
of avian flu in China, the virus utilizes another hemagglutinin,
H7. Therefore, we tested in silico whether compound 1 would
be able to bind to H7. In a docking experiment and a molecular
dynamics simulation with trimeric H7 (pdb 3M5G),24 we
found that compound 1 is well capable to also block
hemagglutinin H7. During the 10 ns MD-simulation, all three
sialic acids of ligand 1 remained in the binding pockets of
hemagglutinin while flexibility was found among the alkyl
chains and the peptidic fragment (Figure 1c; also see movie 2 in
Supporting Information). This is an indicator that ligand 1
might have a very broad activity to various hemagglutinins.
For further development toward a drug for use in a pandemic
situation, more questions need to be addressed. One important
question is the bioavailability. As the target location, i.e., the
upper respiratory tract, is easily accessible, the molecule can be
applied as an aerosol as a delivery method. Thus, commonly
asked pharmacokinetic delivery issues are not relevant for this
molecule.
A decay of the molecule by the viral neuraminidase is, for
several reasons, unlikely. First, the high specific avidity for
hemagglutinin reduces the local concentration of the molecule
in close proximity to the neuraminidase drastically. Second, the
relatively low efficiency (kcat = 9 s−1) of neuraminidase makes a
cleavage of sialic acids from 1 less likely. Third, only about 5−
10% of the viral surface proteins are neuraminidases.25
Furthermore, there is an ongoing discussion about the function
of the influenza neuraminidase in primarily infection and the
activation of the hydrolase activity.26 Fourth, Knowles and co-
workers did not observe a decay of their sialic acid containing
ligands in a hemagglutination assay with whole viruses.20
Current research in our group addresses questions of improving
avidity and ease of synthesis before the present concept will be
tested in in vivo studies.
Figure 3. Top: SPR sensorgram (left) and analysis (right) of ligand 1.
H5 was immobilized on sensor chip and concentrations of 1 from
0.019 to 6 μM were measured. Thermodynamic analysis yielded a KD
= 446 nM. Bottom: results from competitive STD-NMR experiments.
To a solution of 20 mM Neu5Acα2Me and 0.55 μM H5 in PBS, ligand
1 was titrated to a final concentration of 150 μM. The STD-NMR
signals of the competitor were integrated and plotted against the
logarithm of the ligand concentration to yield IC50 = 162 μM and KI =
15 μM using the signal of H-3eq of sialic acid (left) and IC50 = 332 μM
and KI = 30 μM using the signal of H-7 (right).
and the data points were fitted to the one site competition
model yielding the IC50 value (Figure 3, bottom). Ligand 1 was
able to displace Neu5Acα2Me from the primary carbohydrate
recognition domain. The IC50 values for the competitor were
determined and converted into the inhibition constant KI using
the Cheng-Prusoff equation.23 KI values in the low micromolar
range with a good homogeneity were obtained with KI = 15 μM
using H-3eq of Neu5Acα2Me. The multivalent effect was
proven by comparing the inhibition constant of 1 with that of
the monovalent fragment 9, which showed no competitive
behavior in the STD measurements (Figure S9). It is important
to prove that the interaction is of a multivalent nature to further
development of this concept toward medical applications.
To analyze the structural influences of the multivalent effect,
a small substance library was synthesized. Analogs of 1 with 9,
10, and 11 amino acids (ligands 11−14, Figure S16) and also
with a shortened spacer to the aromatic core using a C4
diamine were prepared. None of these compounds performed
better than compound 1 (Figure S5−S8 and S10−S13). The
thermodynamic binding constants of these derivatives of 1 were
in the low millimolar range. This shows clearly that no
multivalent effect is operative for the molecules 11-14. More
importantly, in competition assays, these ligands have no
inhibitory activity for the primary binding site of hemagglutinin.
This is further demonstrated by analysis of the monovalent
fragment 9 which has almost identical binding properties as the
trimeric ligands 11−14. The structural properties of ligand 1
seem to match the geometry of hemagglutinin extremely well,
such that even slightly modified spacers show no multivalency:
We observed drastically worse interaction in constructs that
have shortened spacers. Molecules elongated in the peptide
chain are also worse binders than 1. Even with nine amino acids
in the peptide chain we find that the alkyl chains cannot be
CONCLUSION
■
Ligand 1 has a KD in the high nanomolar range and is therefore
a compound with one of the best reported affinities for small
molecules exhibiting a clear multivalent effect.27 This concept
of a trimeric entry inhibitor can easily be modified and adapted
to other viruses as new species arise as all influenza viruses have
a trimeric hemagglutinin. The ligand presented here can
interact with H5 and in silico also with H7 and, thus, has the
potential to block the virus present in the current infection
outbreak in China. In summary, we believe that the multimeric
concept is a powerful tool especially for easily accessible body
regions like the upper respiratory tract, where the application of
an aerosol is possible and pharmacokinetic effects are less
important. Because the binding epitope for all hemagglutinins,
i.e., sialic acid, is determined by the host and because of the
global nature of the multivalent interaction, these ligands are
less prone to a viral escape by genetic drift and can also be
easily structurally adjusted by simple modification of single
building blocks and represent, therefore, a precious tool in
order to develop powerful and long lasting drugs.
787
dx.doi.org/10.1021/ja410918a | J. Am. Chem. Soc. 2014, 136, 783−788

Journal of the American Chemical Society
Article
Armstrong, G. D.; Ling, H.; Pannu, N. S.; Read, R. J.; Bundle, D. R.
Nature 2000, 403 (6770), 669−672.
(14) Gestwicki, J. E.; Cairo, C. W.; Strong, L. E.; Oetjen, K. A.;
Kiessling, L. L. J. Am. Chem. Soc. 2002, 124 (50), 14922−14933.
(15) Rao, J.; Lahiri, J.; Isaacs, L.; Weis, R. M.; Whitesides, G. M.
Science 1998, 280 (5364), 708−711.
(16) Shen, Z.; Yan, H.; Zhang, Y.; Mernaugh, R. L.; Zeng, X. Anal.
Chem. 2008, 80 (6), 1910−1917.
(17) Boland, Y.; Hertsens, P.; Marchand-Brynaert, J.; Garcia, Y.
Synthesis 2006, 9, 1504−1512.
ASSOCIATED CONTENT
■
S
* Supporting Information
Details about the in silico data, a ligand interaction plot between
H5 and 1, SPR and STD-NMR data of all ligands and
experimental procedures for the synthesis of the building blocks
and the ligands; two 2 ns movies of the 10 ns molecular
dynamic simulations of ligand 1 with H5 (movie1.qt) and with
H7 (movie2.qt). This material is available free of charge via the
Internet at http://pubs.acs.org.
(18) (a) Kuhn, R.; Lutz, P.; Macdonald, D. L. Chem. Ber. 1966, 99
(2), 611−617. (b) Furuhata, K.; Anazawa, K.; Itoh, M.; Shitori, Y.;
Ogura, H. Chem. Pharm. Bull. 1986, 34 (7), 2725−2731.
(19) Toogood, P. L.; Galliker, P. K.; Glick, G. D.; Knowles, J. R. J.
Med. Chem. 1991, 34 (10), 3138−3140.
AUTHOR INFORMATION
■
Corresponding Author
bernd.meyer@chemie.uni-hamburg.de
(20) Glick, G. D.; Knowles, J. R. J. Am. Chem. Soc. 1991, 113 (12),
4701−4703.
Notes
The authors declare no competing financial interest.
(21) Gamian, A.; Chomik, M.; Laferrier
Microbiol. 1991, 37 (3), 233−237.
̀
e, C. A.; Roy, R. Can. J.
(22) (a) Mayer, M.; Meyer, B. Angew. Chem., Int. Ed. 1999, 38 (12),
1784−1788. (b) Meyer, B.; Peters, T. Angew. Chem., Int. Ed. 2003, 42
(8), 864−890.
ACKNOWLEDGMENTS
■
This work was supported by an instrumentation grant (DFG)
Me-1830/1-1 for the 700 MHz NMR spectrometer. We also
(23) Cheng, Y.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22 (23),
3099−3108.
like to thank I. Gruneberg for technical assistance and Dr. E.
̈
Magbanua for preparing the table of contents artwork.
(24) Yang, H.; Chen, L.-M.; Carney, P. J.; Donis, R. O.; Stevens, J.
PLoS Pathog. 2010, 6 (9), e1001081.
(25) Janakiraman, M. N.; White, C. L.; Laver, W. G.; Air, G. M.; Luo,
M. Biochemistry 1994, 33 (27), 8172−8179.
(26) (a) Matrosovich, M. N.; Matrosovich, T. Y.; Gray, T.; Roberts,
N. A.; Klenk, H. D. J. Virol. 2004, 78 (22), 12665−12667. (b) Ohuchi,
M.; Asaoka, N.; Sakai, T.; Ohuchi, R. Microbes Infect. 2006, 8 (5),
1287−1293.
(27) (a) Song, G.; Yang, S.; Zhang, W.; Cao, Y.; Wang, P.; Ding, N.;
Zhang, Z.; Guo, Y.; Li, Y. J. Med. Chem. 2009, 52 (23), 7368−7371.
(b) Weinhold, E. G.; Knowles, J. R. J. Am. Chem. Soc. 1992, 114 (24),
9270−9275.
REFERENCES
■
(1) Stevens, J.; Blixt, O.; Tumpey, T. M.; Taubenberger, J. K.;
Paulson, J. C.; Wilson, I. A. Science 2006, 312 (5772), 404−410.
(2) Salomon, R.; Webster, R. G. Cell 2009, 136 (3), 402−410.
(3) Webby, R. J.; Webster, R. G. Science 2003, 302 (5650), 1519−
1522.
(4) (a) von Itzstein, M. Nat. Rev. Drug Discovery 2007, 6 (12), 967−
974. (b) Dyason, J. C.; von Itzstein, M. Microbial Glycobiology;
Academic Press: Boston, MA, 2010; pp 269−283; (c) Lee, S. M.-Y.;
Yen, H.-L. Antivir. Res. 2012, 96 (3), 391−404.
(5) (a) Bodian, D. L.; Yamasaki, R. B.; Buswell, R. L.; Stearns, J. F.;
White, J. M.; Kuntz, I. D. Biochemistry 1993, 32 (12), 2967−2978.
(b) Russell, R. J.; Kerry, P. S.; Stevens, D. J.; Steinhauer, D. A.; Martin,
S. R.; Gamblin, S. J.; Skehel, J. J. Proc. Natl. Acad. Sci. U.S.A. 2008, 105
(46), 17736−17741.
(6) (a) Le, Q. M.; Kiso, M.; Someya, K.; Sakai, Y. T.; Nguyen, T. H.;
Nguyen, K. H.; Pham, N. D.; Ngyen, H. H.; Yamada, S.; Muramoto,
Y.; Horimoto, T.; Takada, A.; Goto, H.; Suzuki, T.; Suzuki, Y.;
Kawaoka, Y. Nature 2005, 437 (7062), 1108. (b) van der Vries, E.;
Stelma, F. F.; Boucher, C. A. B. New Engl. J. Med. 2010, 363 (14),
1381−1382. (c) Yang, J.; Li, M.; Shen, X.; Liu, S. Viruses 2013, 5 (1),
352−373.
(7) Wilson, I. A.; Skehel, J. J.; Wiley, D. C. Nature 1981, 289 (5796),
366−373.
(8) Sauter, N. K.; Bednarski, M. D.; Wurzburg, B. A.; Hanson, J. E.;
Whitesides, G. M.; Skehel, J. J.; Wiley, D. C. Biochemistry 1989, 28
(21), 8388−8396.
(9) Mammen, M.; Choi, S.-K.; Whitesides, G. M. Angew. Chem., Int.
Ed. 1998, 37 (20), 2754−2794.
(10) Sigal, G. B.; Mammen, M.; Dahmann, G.; Whitesides, G. M. J.
Am. Chem. Soc. 1996, 118 (16), 3789−3800.
(11) Thompson, M. G.; Shay, D. K.; Zhou, H.; Bridges, C. B.; Cheng,
P. Y.; Burns, E.; Bresee, J. S.; Cox, N. J. J. Am. Med. Assoc. 2010, 304
(16), 1778−1780.
(12) (a) Ha, Y.; Stevens, D. J.; Skehel, J. J.; Wiley, D. C. Proc. Natl.
Acad. Sci. U.S.A. 2001, 98 (20), 11181−11186. (b) Yamada, S.; Suzuki,
Y.; Suzuki, T.; Le, M. Q.; Nidom, C. A.; Sakai-Tagawa, Y.; Muramoto,
Y.; Ito, M.; Kiso, M.; Horimoto, T.; Shinya, K.; Sawada, T.; Usui, T.;
Murata, T.; Lin, Y.; Hay, A.; Haire, L. F.; Stevens, D. J.; Russell, R. J.;
Gamblin, S. J.; Skehel, J. J.; Kawaoka, Y. Nature 2006, 444 (7117),
378−382.
(13) (a) Kitov, P. I.; Bundle, D. R. J. Am. Chem. Soc. 2003, 125 (52),
16271−16284. (b) Kitov, P. I.; Sadowska, J. M.; Mulvey, G.;
788
dx.doi.org/10.1021/ja410918a | J. Am. Chem. Soc. 2014, 136, 783−788