Novel trivalent anti-influenza reagent

Article abstract of DOI:10.1016/j.bmcl.2010.04.060

We designed and synthesized novel trivalent anti-influenza reagents. Sialyllactose was located at the terminal of each valence which aimed to block each receptor-binding site of the hemagglutinin (HA) trimer on the surface of the virus. Structural analyse

Full text of DOI:10.1016/j.bmcl.2010.04.060

Bioorganic & Medicinal Chemistry Letters 20 (2010) 3772–3776
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel trivalent anti-influenza reagent
Fei Feng a, , Nobuaki Miura a, , Norikazu Isoda b, Yoshihiro Sakoda b, Masatoshi Okamatsu b,
Hiroshi Kida b, ,à, Shin-Ichiro Nishimura a, ,§
*
*
a Department of Advanced Transdisciplinary Science, Faculty of Advanced Life Science, and Frontier Research Center for Post-Genome Science and Technology,
Hokkaido University, Sapporo, Japan
b Laboratory of Microbiology, Graduate School of Veterinary Medicine, Hokkaido University, Sapporo, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
We designed and synthesized novel trivalent anti-influenza reagents. Sialyllactose was located at the ter-
minal of each valence which aimed to block each receptor-binding site of the hemagglutinin (HA) trimer
on the surface of the virus. Structural analyses were carried out with a model which was constructed with
a computer simulation. A previously reported cyclic glycopeptide blocker [Ohta, T.; Miura, N.; Fujitani, N.;
Nakajima, F.; Niikura, K.; Sadamoto, R.; Guo, C.-T.; Suzuki, T.; Suzuki, Y.; Monde, K.; Nishimura, S.-I.
Angew. Chem. Int. Ed., 2003, 42, 5186] bound to the HA in the model. The analyses suggest that the glu-
Received 1 December 2009
Revised 28 March 2010
Accepted 15 April 2010
Available online 18 April 2010
Keywords:
tamine residue in the cyclic peptide bearing Neu5Aca2,3Galb1,4Glc trisaccharide via a linker interacts
Computer-based drug design
Anti-influenza virus infection
Trivalent glycoside
Scaffold
with the Gln189 in HA through hydrogen bonding. The present anti-influenza reagents likely interact
with a glutamine residue included in the vicinity of Gln189. A plague reduction assay of the influenza
virus, A/PR/8/1934 (H1N1), was performed in MDCK cells to evaluate for the synthesized compounds
to inhibit viral replication. One of the compounds showed approximately 85% inhibition at the concen-
tration of 400 lM at 4 °C.
Ó 2010 Elsevier Ltd. All rights reserved.
An influenza outbreak poses a significant threat to public the
health worldwide as highlighted by the novel swine origin influ-
enza A (H1N1). Since the first detection in April 2009 until Novem-
ber 2009, the human case of infection with the virus has been
confirmed over 6770 deaths in 206 countries by the WHO.1 The
pandemic of novel H1N1 virus hasten the development of new
anti-virus reagents that are effective for various subtypes of influ-
enza A virus. Oseltamivir and zanamivir as neuraminidase (NA)
inhibitors are now widely used as effective anti-influenza drugs.
However, strains resistant to these inhibitors have been selected
mutations of NA result in inhibitor resistance.2,3 Therefore, novel
inhibitors of influenza virus are strongly required.
Influenza virus expresses the other glycoprotein, hemagglutinin
(HA), which plays a fundamental role in the initial step of the infec-
tion process.4 HA binds to the receptor carbohydrate chain termi-
nated with neuraminic acid on the host cell-surface. Thus, HA is
a potential target of anti-influenza virus reagents. Inhibition of
the binding between HA and the receptor on the cell-surface
should prevent humans from being infected by virus. HA blocker
is expected to act as a prophylactic reagent. The structure of the
HA has been analyzed in detail by extensive biochemical and crys-
tallographic studies.5 These studies indicated that the HA has three
identical receptor-binding sites on the top of the trimer that pro-
truded from the viral envelope.6 In addition, sialyllactose binds
to the HA of influenza virus more strongly than one neuraminic
acid unit.7 It is important to efficiently occupy the binding sites
of the HA. We have reported that a trivalent glycopeptide inhibits
efficiently hemagglutination mediated by viral hemagglutinin.8
Several groups also demonstrated that synthetic polymers bearing
multiple neuraminic acids showed greatly enhanced affinity to
influenza HA.9 In recent years, multivalent blockers designed with
polyamino acid have also been developed.10 Since multivalent gly-
coconjugate blockers have densely displayed neuraminic acids at
the terminal of the sugar chain, it is thought to amplify the affinity
between the HA and the blockers. Since syntheses of glycoproteins
and glycopeptides are quite expensive, complicated and difficult to
complete in short period of time with the large quantity, it is defec-
tive for HA blockers as preventive medicines. In this Letter, we re-
port the synthesis and evaluation of the inhibitory effect of novel
trivalent blockers against influenza virus A/PR/8/1934 (H1N1).
We have developed much more simple and practical trivalent
blockers with siallyllactose at each terminal of valence (Fig. 1). In
our previous Letter, we reported that the monovalent glycopep-
tide showed no inhibition of hemagglutination. Therefore, in this
study, we only designed trivalent blockers. These blockers were
* Corresponding authors.
E-mail addresses: kida@vetmed.hokudai.ac.jp (H. Kida), shin@glyco.sci.
hokudai.ac.jp (S.-I. Nishimura).
These authors contributed equally to this work.
Tel.: +81 11 706 5207; fax: +81 11 706 5273.
Tel.: +81 11 706 9043; fax: +81 11 706 9042.
à
§
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.04.060

F. Feng et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3772–3776
3773
Figure 1. Chemical structures of novel trivalent anti-influenza reagetns.
designed with structure-based methods shown in Figure 2a and b.
Geometrical observation was performed with the model struc-
tures which were constructed with the HA (PDBID is 1HGG)
and our previously developed glycopeptides. The major character
of the binding between HA and the blocker were kept in the
model structure. Then novel alternative core component were
Figure 2. Model structures of binding between trivalent trivalent anti-influenza reagents and HA. (a) trisphenol type, (b) trisaniline type, and (c) previously synthesized HA
blocker (Ref. 8).

3774
F. Feng et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3772–3776
examined. The length of the linker was determined and the triva-
lent blockers were designed and synthesized. In order to induce
an interaction with the Q189 residue as shown in Figure 2c, glu-
tamine residue was introduced in the linker part as connectors
between aglycon core and trisaccharides in addition to the inter-
action between sialyllactose and the HA receptor site.
Our designed compounds bound to the HA (PDB ID:1HGG) on
model structures with computer simulations. We carried out short
molecular dynamics simulations (200 ps) at 1000 K with the time
step of 1 fs by using MMFF94S force fields. During the simulation,
the geometrical position of sialylllactoses and the HA were fixed.
The conformation with the lowest potential energy as shown in
Figure 2 was selected and observed. This observation suggests that
the lengths of the linkers and the position of the glutamine residue
were appropriate and the designed compounds could bind effi-
ciently with HA.
valences can spread out evenly. The flexible hydrophilic linker
between trisaccharide and the core was built from oligoethylene-
glycol structure derived from 2-[2-(2-chloroethoxy)ethoxy]etha-
nol and 3-{2-[2-(9H-fluoren-9-ylmethoxycarbonylamino)-ethoxy]
-ethoxy}-propionic acid. First, 3 and 7 were synthesized according
to the above design. Then the tosyl group was introduced into 3 as
a leaving group in order to convert it to the azide compound 5
which was transformed to the trisamine 6 by a single reduction
reaction. Another core aglycon 8 was built briefly in two steps by
normal deprotection of Fmoc in 7.
N-
a
-Fluorenylmethoxycarbonyl-O-t-butyl
L-glutamic acid
(Fmoc-Glu-OtBu) was activated into its N-succinimidyl ester inter-
mediate 9 and coupled with the full-protected lactose derivative
1011 to give the key intermediate 11 which was converted effec-
tively into the corresponding acid form 12 by TFA (Scheme 2).
As depicted in Schemes 3 and 4, the key condensation reac-
tions between trisamine (6 or 8) and acid 12 were carried out
by the use of condensing agents EDC and DPPA, respectively. After
the Fmoc protecting group at amino group and the acetyl group
on sugar were removed smoothly, sialic acid was introduced by
Chemical synthesis of the core parts of the HA blockers is out-
lined in Scheme 1. Trisphenol and trisaniline skeleton were
adopted as the starting materials because both have rigid confor-
mations because of the existence of a center sp3 carbon, so three
Scheme 1. Reagents and conditions: (i) 2-[2-(2-chloroethoxy)ethoxy]ethanol, K2CO3, KI, DMF, 100 °C, 46 h, 65%; (ii) TsCl, TEA, CH2Cl2, rt, 3 h, 67%; (iii) NaN3, DMF, 60 °C, 9 h,
94%; (iv) H2 gas, 10% Pd–C, MeOH/EtOAc = 2:1, rt, 2 h, 91%; (v) 3-{2-[2-(9H-fluoren-9-ylmethoxycarbonylamino)-ethoxy]- ethoxy}-propionic acid, DCC, DMAP, CH2Cl2, rt, 6 h,
83%; (vi) piperidine, CH3CN–CHCl3, rt, 1.5 h, 60%.

F. Feng et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3772–3776
3775
Scheme 2. Reagents and conditions: (i) TEA, CHCl3, rt., 24 h, 56%); (ii) TFA, CH2Cl2, rt, 3 h, 94%.
Scheme 3. Reagents and conditions: (i) EDC, DMAP, CH2Cl2, rt, 22 h, 59%; (ii) (a)
piperidine, CH3CN, rt, 1.5 h; (b) NaOMe–MeOH, rt, 1 h, 87% (two steps); (iii)
a2,3-
Scheme 4. Reagents and conditions: (i) DPPA, TEA, DMF, rt, 18 h, 32%; (ii) (a)
(N)-sialyltransferase, CMP-NANA, cacolylate buffer, MnCl2, 37 °C, 28 h, 45%.
piperidine, CH3CN, rt, 1.5 h; (b) NaOMe–MeOH, rt, 1 h, 80% (two steps); (iii) a2,3-
(N)-sialyltransferase, CMP-NANA, cacolylate buffer, MnCl2, 37 °C, 64 h, 46%.
an enzymatic method with
condition to give the target trisphenol–sialyllactose 1 and trisan-
iline–sialyllactose 2. Since 2,3-(N)-sialyltransferase was easy to
purchase, we used the enzyme and synthesized the compounds
with the Neu5Ac- 2,3Gal terminal. The final target compounds
were purified by reverse phase high-performance liquid chroma-
tography, followed by lyophilization. All of the structures of the
synthesized compounds were confirmed by 1H, 13C NMR and
MS measurements.12
a2,3-sialyltransferase under a mild
on MDCK cells.13 The employed virus strain, A/PR/8/1934 (H1N1),
recognized the receptor with saccharide terminated in Neu5Ac-
a
sized compounds 1 and 2 is quite small, the inhibition activity
against the concentration of the compounds could not be observed.
Thus, we could not evaluate the IC50 of the inhibition. We evalu-
ated the inhibition effect with MDCK cells at a certain concentra-
a
2,3Gal and Neu5Ac-a
2,6Gal.14 Since the amount of the synthe-
a
tion of 400 lM. As the results are shown in Table 1, 1 showed a
Four synthesized substances, 1, 2, 14 and 16 showed higher
water solubility, thus the test was facilitated in an aqueous system.
The inhibition effects against human influenza A virus infection
were investigated by means of inhibition of the cytopathic effect
significant inhibitory effect, particularly at lower temperature
4 °C, as the number of plaques was reduced and approximately
85% inhibition of plaque formation while 2 gave a weaker effect
than 1. Although 14 and 16 possess the same trivalent skeleton

3776
F. Feng et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3772–3776
Table 1
the influenza virus removes the terminal neuraminic acid in each
valence. Therefore, unnatural glycosidic bonds have to be exam-
ined to prevent the neuraminic acid being removed by NA. On
the basis of the results from this study, we are examining the more
effective anti-influenza reagents.
Blocking effects of novel trivalent blockers on virus replication
Numbers of plagues at different adsorption
temperature (percentage of plaque [%])
4 °C
35 °C
Trisphenol–sialyllactose (1) 9 Â 10 (15.5)
Trisaniline–sialyllactose (2) 39 Â 10 (67.2)
38 Â 10 (45.6)
47 Â 10 (56.6)
64 Â 10 (77.1)
68 Â 10 (81.9)
83 Â 10 (100.0)
0 (0.0)
Acknowledgments
Trisphenol–lactose (14)
Trisaniline–lactose (16)
Virus only
42 Â 10 (72.4)
45 Â 10 (77.6)
58 Â 10 (100.0)
0 (0.0)
This work was partly funded by a Program of Founding Re-
search Centers for Emerging and Reemerging Infectious Diseases
grant. We thank Ms. Hitomi Shibuya for her experimental support.
The authors also wish to thank Ms. S. Oka at the Center for Instru-
mental Analysis of Hokkaido University for mass measurement.
N.M. special thanks to Ms. Kana Tosho who contributed her sup-
port in preparation of the Letter.
No virus (inoculation)
as 1 and 2 with no neuraminic acid they indicated the small block-
ing effect. It suggests that there are no-specific interaction between
14 or 16 and virus. 1 showed a higher activity at a low temperature
(4 °C) than it was at 35 °C. This suggests that neuraminidase activ-
ity was suppressed at this temperature. Terminal neuraminic acid
might be partly cleaved by neuraminidase at 35 °C.
Supplementary data
Unexpectedly, 1 and 2 exhibited no hemagglutination-inhibi-
tion assay (HI) activity (data not shown). In the molecular simula-
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2010.04.060.
tions, we investigated
a geometrical aspect of the complex
between the compounds and hemagglutinin. In this study, we
challenged to simplify the complex HA blocker which previously
designed. The compounds 1 and 2 were quite simple compared
with the previously designed glycopeptides type HA blocker. Our
compounds can suppress the replication of the influenza virus. It
is necessary to elucidate the interaction between virus and the
anti-influenza reagents. This study was the first step of such a com-
puter assisted drug design. We could not perform the experiment
to evaluate the interaction between virus and our synthesized
compounds, because the amount of obtained compounds was quite
small. We also could not determine the reason why 2 had less virus
replication activity than 1. The two compounds carried three
equivalent neuraminic acid sugar units. They only differed in core
structure. Exceedingly, the freedom of the sugar moieties in 2 was
different from that of 1 because of the stiff amido bond, hence the
trisaccharide in 2 was unable to approach the viral HA binding sites
efficiently. Perhaps some other steric factors were limited the con-
formational flexibility of 2.
References and notes
1. World Health Organization. Pandemic (H1N1) 2009—Update 75. (http://
www.who.int/csr/don/2009_11_20a/en/print.html).
2. Ferraris, O.; Lina, B. J. Clin. Virol. 2008, 41, 13.
3. Collins, P. J.; Haire, L. F.; Lin, Y. P.; Liu, J.; Russell, R. J.; Walker, P. A.; Skehel, J. J.;
Martin, S. R.; Hay, A. J.; Gamblin, S. J. Nature 2008, 453, 1258.
4. (a) White, J. M. Annu. Rev. Physiol. 1990, 52, 675; (b) Suzuki, Y.; Nagao, Y.; Kato,
H.; Matsumoto, M.; Nerome, K.; Nakajima, K.; Nobusawa, E. J. Biol. Chem. 1986,
261, 17057.
5. (a) Wilson, I. A.; Skehel, J. J.; Wiley, D. C. Nature 1981, 289, 366; (b) Carr, C. M.;
Kim, P. T. Cell 1993, 73, 823; (c) Matrosovich, M. N.; Gambaryan, A. S.; Tuzikov,
A. B.; Byramova, N. E.; Mochalova, L. V.; Golbraikh, A. A.; Shenderovich, M. D.;
Finne, J.; Bovin, N. V. Virology 1993, 196, 111.
6. Lees, W. J.; Spaltenstein, A.; Kingery-Wood, J. E.; Whitesides, G. M. J. Med. Chem.
1994, 37, 3419.
7. Suzuki, Y. Prog. Lipid Res. 1994, 33, 429.
8. Ohta, T.; Miura, N.; Fujitani, N.; Nakajima, F.; Niikura, K.; Sadamoto, R.; Guo, C.-
T.; Suzuki, T.; Suzuki, Y.; Monde, K.; Nishimura, S.-I. Angew. Chem., Int. Ed. 2003,
42, 5186.
9. (a) Gamian, A.; Chomik, M.; Laferrière, C. A.; Roy, R. Can. J. Microbiol. 1991, 37,
233; (b) Kamitakahara, H.; Suzuki, T.; Nishigori, N.; Suzuki, Y.; Kanie, O.; Wong,
C.-H. Angew. Chem., Int. Ed. 1998, 37, 1524; (c) Reuter, J. D.; Myc, A.; Hayes, M.
M.; Gan, Z.; Roy, R.; Qin, D.; Yin, R.; Piehler, L. T.; Esfand, R.; Tomalia, D. A.;
Baker, J. R., Jr. Bioconjugate Chem. 1999, 10, 271.
10. (a) Ogata, M.; Murata, T.; Murakami, K.; Suzuki, T.; Hidari, I. P. J. K.; Suzuki, Y.;
Usui, T. Bioorg. Med. Chem. 2007, 15, 1383; (b) Ogata, M.; Hidari, I. P. J. K.;
Kozaki, W.; Murata, T.; Hiratake, J.; Park, Y. E.; Suzuki, T.; Usui, T.
Biomacromolecules 2009, 10, 1894.
11. Yamada, K.; Fujita, E.; Nishimura, S.-I. Carbohydr. Res. 1998, 305, 443.
12. See Supplementary data.
In conclusion, novel trivalent anti-influenza reagents were de-
signed and constructed effectively. The inhibition effects were
examined in the virus replication of the A/PR/8/34 (H1N1) virus
strain in MDCK cells. Based on the significant activity of trisphe-
nol–sialyllactose 1 at 400
lM, this may be a potential candidate
for anti-influenza drug. It is impressed that our compounds has
inhibition of virus replication without HAI activity. It suggests that
the further investigation to understand the detail of interaction be-
tween the anti-influenza blocker and virus were indispensable to
better design of novel anti-influenza reagents. The temperature
dependence of the MDCK assay indicated that the NA activity of
13. (a) Watanabe, W.; Sudo, K.; Asawa, S.; Konno, K.; Yokota, T.; Shigeta, S. J. Viral
Methods 1995, 51, 185; (b) Mochalova, L. V.; Tuzikov, A. B.; Marinina, V. P.;
Gambaryan, A. S.; Byramova, N. E.; Bovin, N. V.; Matrosovich, M. N. Antiviral
Res. 1994, 23, 179.
14. Rogers, G. N.; Paulson, J. C. Virology 1983, 127, 361.