A strategy for neuraminidase inhibitors using mechanism-based labeling information

Article abstract of DOI:10.1002/asia.201000594

A potent inhibitor for Vibrio cholerae neuraminidase (VCNA) was developed by using a novel two-step strategy, a target amino acid validation using mechanism-based labeling information, and a potent inhibitor search using a focused library. The labeling in

Full text of DOI:10.1002/asia.201000594

FULL PAPERS
DOI: 10.1002/asia.201000594
A Strategy for Neuraminidase Inhibitors Using Mechanism-Based Labeling
Information
Hiroshi Hinou,*[a] Risho Miyoshi,[a] Yasuaki Takasu,[a] Hirokazu Kai,[a]
Masaki Kurogochi,[a] Shingo Arioka,[b] Xiao-Dong Gao,[a] Nobuaki Miura,[a]
Naoki Fujitani,[a] Shinya Omoto,[c] Tomokazu Yoshinaga,[c] Tamio Fujiwara,[c]
Takeshi Noshi,[b] Hiroko Togame,[b] Hiroshi Takemoto,[b] and Shin-Ichiro Nishimura*[a]
Abstract: A potent inhibitor for Vibrio
cholerae neuraminidase (VCNA) was
developed by using a novel two-step
strategy, a target amino acid validation
using mechanism-based labeling infor-
mation, and a potent inhibitor search
using a focused library. The labeling in-
formation suggested the hidden dynam-
ics of a loop structure of VCNA, which
can be a potential target of the novel
inhibitor. A focused library composed
of 187 compounds was prepared from a
9-azide derivative of 2,3-dehydro-N-
acetylneuraminic acid (DANA) to in-
terrupt the function of the loop of the
labeled residues. Inhibitor 3c showed
potent inhibition properties and was
the strongest inhibitor with FANA, a
N-trifluoroacetyl derivative of DANA.
Validation studies of the inhibitor with
a detergent and a Lineweaver–Burk
plot suggested that the 9-substitution
group would interact hydrophobically
with the target loop moiety, adding a
noncompetitive inhibition property to
the DANA skeleton. This information
enabled us to design compound
4
having the combined structure of 3c
and FANA. Compound 4 showed the
most potent inhibition (Ki =73 nm,
mixed inhibition) of VCNA with high
selectivity among the tested viral, bac-
terial, and mammal neuraminidases.
Keywords: drug design
· hydro-
lases · inhibitors · sialic acid · struc-
ture–activity relationship
Introduction
of glycoproteins and the cells of vertebrates—regulate inter-
actions with the extracellular world including infection, im-
munity, and cell–cell adhesion.[1] In particular, many patho-
gens, such as viruses, bacteria, and protozoa, utilize their
own NAs for invasion, nutrition, detachment, and immuno-
logic escape.[2] Thus, NA inhibitors are an attractive target
for drug discovery. Vibrio cholerae neuraminidase (VCNA),
the oldest and best studied NA,[3] is involved in cholera
pathogenesis by degrading the mucin layer of the gastroin-
testinal tract to enhance the attachment of the bacteria by
trimming higher-order gangliosides to GM1, a putative re-
Neuraminidases (NAs; EC 3.2.1.18)—a family of glycoside
hydrolase enzymes that cleave the glycosidic linkages of
sialic acid (Neu5Ac; neuraminic acid) located on the surface
[a] Dr. H. Hinou, R. Miyoshi, Y. Takasu, H. Kai, Dr. M. Kurogochi,
Dr. X.-D. Gao, Dr. N. Miura, Dr. N. Fujitani, Prof. S.-I. Nishimura
Graduate School of Life Science and Frontier Research
Center for the Post-Genome Science and Technology
Hokkaido University
N21, W11, Kita-ku, Sapporo 001-0021 (Japan)
Fax : (+81)11-707-9042
ceptor of the cholera toxin,[4] and catabolizing Neu5Ac.[5]
A
structural study of VCNA revealed that this enzyme has two
lectin-like domains other than the central catalytic domain,
and alignment of the active site is very similar to that ob-
served in influenza (Flu) NA and Salmonella NA.[6] Despite
the success of the “structure-based inhibitor design”[7] for
production of anti-influenza drugs such as zanamivir (Relen-
za)[8] and oseltamivir (Tamiflu; ethyl ester of GS4071),[9] an
inhibitor design for VCNA lacks the ability to surpass the
nonselective NA inhibitor 2,3-dehydro-2-deoxy-N-acetyl-
neuraminic acid (DANA) and its N-trifluoromethyl deriva-
E-mail: hinou@glyco.sci.hokudai.ac.jp
[b] Dr. S. Arioka, Dr. T. Noshi, Dr. H. Togame, Prof. H. Takemoto
Shionogi Innovation Center for Drug Discovery
Shionogi & Co., Ltd.
N21, W11, Kita-ku, Sapporo 001-0021 (Japan)
[c] Dr. S. Omoto, Dr. T. Yoshinaga, Dr. T. Fujiwara
Discovery Research Laboratories
Shionogi & Co., Ltd., 5-1, Mishima 2-Chome
Settsu, Osaka 566-0022 (Japan)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/asia.201000594.
1048
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 1048 – 1056

actual affinity information and
design a new inhibitor of the
loop moiety.
Previously, we reported
mechanism-based labeling
study of VCNA by using 2-di-
a
fluoromethylaryl[16]
sialo-
side 1[17] (Scheme 1), and fol-
lowing protease digestion, sepa-
ration of the labeled peptide
fragment using an anti-dansyl
column.
MALDI-TOF/TOF
MS analysis revealed that the
Asp576 and Arg577 residues lo-
Scheme 1. Reported neuraminidase inhibitors.
cated in a peripheral hydropho-
bic loop (575LDRFFL580) in the
tives (FANA, Scheme 1),[10] which were reported in 1974.[11]
Conformational flexibility of the loop moieties in the cata-
lytic site of the NA[12,13] might be a novel target for the
design of potent and selective VCNA inhibitors because the
active site of the NAs is mainly composed of amino acid res-
idues on a loop moiety in the b-propeller-type catalytic
domain,[6] and the conformational change in the loop moiety
is involved in the catalytic mechanism (Figure 1a). Chavas
et al.[12] demonstrated in crystallographic studies that the
human cytosolic NA (Neu2) has two flexible loops contain-
ing Glu111 and catalytic Asp46 residues, which are disor-
dered in the apo form but ordered with DANA to adapt
two short a helices to cover the inhibitor and locate Asp46
for catalysis. Russell et al.[13] demonstrated that the influenza
virus NAs (N1–N9) could be classified into two structurally
distinct groups by conformational changes in the so-called
“150-loop”, which contains the catalytic Asp151 residue.
Group 1 is composed of the N1, N4, N5, and N8 subtypes
with an additional cavity adjacent to their active sites rela-
tive to group 2 (N2, N3, N6, N7, and N9). Despite this infor-
mation, only one FluNA inhibitor recognized the structural
change in the 150-loop[14] by sacrifice of the affinity of the
zanamivir skeleton because such a dynamic target is not
suitable for design based only on static structural informa-
tion. We thought that mechanism-based labeling of these
flexible loops[15] could predict the dynamic nature with
catalytic domain were labeled by the aglycone of 1 (Fig-
ure 1a).[17] The two labeled sites indicate that this enzyme is
active after the first aglycone 1 attack but is inactive after
the second attack. Interestingly, crystallographic data (PDB:
1W0O)[18] of VCNA indicated that the two amino acid resi-
dues are about 20 ꢁ away from the anomeric position of
NeuAc in the active-site pocket covered with a loop
(
246VGGGDPGALSN256), which includes a putative catalytic
Asp251 residue; the Asp576 residue is buried within the cat-
alytic domain, and the consecutive phenylalanine residues
next to the labeled sites are exposed to the composed hy-
drophobic surface on VCNA (Figure 1b). This suggested
that the labeled residues are located on the hydrophobic
flexible loop, and that the Asp576 and Arg577 residues
appear to interact with the aglycone of 1 after hydrophobic
recognition corresponding to the catalytic action of VCNA.
Although the crystallographic data of VCNA with DANA
(PDB: 1W0O) and the apo form (PDB: 1W0P) did not
show the dynamism (Figure 1c)[18] that occurs with human
neuraminidase 2 (hNeu2)[12] and FluNA,[13] the hydrophobic
loop 575–580 and an intermediate large loop
(
312PTDAAQNGDRIKPWMP327) lying between the hydro-
phobic loop and DANA were not conformationally
stabilized[6,18](Figure 1a). Thus, we designed potent and spe-
cific inhibitors based on the labeling information, hydropho-
bic affinity, and putative dynamism of the loop structures. In
this paper, we describe a novel strategy for neuraminidase
inhibitors by using the information of mechanism-based la-
beling and a focused library designed by validation of the la-
beling mechanism.
Abstract in Japanese:
Results and Discussion
Design and Synthesis of a Focused Library for the
Inhibition of VCNA
To design the novel inhibitor skeleton for VCNA, the fol-
lowing four points were considered: 1) the ring skeleton of
DANA interacts with the strictly conserved amino acid resi-
dues among the NAs; 2) the 9-position of DANA is directed
to the labeled loop; 3) the steric barrier of an intermediate
Chem. Asian J. 2011, 6, 1048 – 1056
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1049

H. Hinou, S.-I. Nishimura et al.
FULL PAPERS
than 70% for product 3 versus
2 were used for the screening of
the VCNA inhibition activity
(see the Supporting Informa-
tion). The triazole derivatives 3
(100 mm; calculated as the click
reactions 100% yield) were in-
cubated with 2-a-(4’-methylum-
belliferyl)-N-acetylneuraminic
acid (4-MU-NANA; 2 mm) and
VCNA (0.15 mU; Figure 2a).
The reagents used for the click
reaction (copper ion, etc.) did
not inhibit the VCNA activity
(Figure 2b). We classified each
compound into five groups
based on the relative remaining
activity in the screening result,
and eight compounds (3a–h;
each structure is shown in the
Supporting Information) that
belonged to the most potent
group (0–0.2 relative activity)
were selected for the second
screening (Figure 2a). The eight
compounds
were
prepared
again as purified compounds,
and 50 mm of each compound
was incubated again with 4-
MU-NANA (1 mm) and VCNA
(0.45 mU) for a second screen-
ing. Three compounds (3a, 3c,
and 3d) showed the significant-
ly potent inhibition of VCNA
(78, 95, and 88% inhibition ac-
tivity, respectively), whereas
DANA showed only a 12% in-
hibition (Figure 2c).
Figure 1. Structural analysis of VCNA (PDB ID 1W0O) active site: a) the catalytic domain of VCNA [DANA
(pink); conserved residues (W311, D250, E619, Y740, E756, R224, R635, R712) and catalytic loop
246VGGGDPGALSN256 (gray), intermediate large loop 312PTDAAQNGDRIKPWMP327 (light blue), labeled
residues by 1 (D576 and R577), and hydrophobic loop 575LDRFFL580 (yellow)]; b) surface model of VCNA
active site in which the ligand was placed in the deep pocket of the active site and the catalytic loop and the
intermediate loop must move away to open the pocket for the ligand recognition; c) superposition of the
VCNA–DANA complex (green, PDB ID 1W0O) and an apo VCNA (yellow, PDB ID 1W0P). Clear structural
changes cannot be found from these structures; d) inhibitor design based on the mechanism-based labeling in-
formation.
large loop from Pro312 to Pro327 (Figure 1a) can be ignored
because this loop moves away from DANA during the rec-
Validation of the Inhibition Mode of 3a, 3c, and 3d
ognition process; and 4) the hydrophobic surface of the la-
beled loop moves toward DANA and the distance is varia-
ble. Based on these four considerations, a focused library 3
with a modification point at the C9 position of the DANA
skeleton was designed. The random structure at the C9 posi-
tion of library 3 was constructed by a [2+3] cyclization
(click reaction)[19] between the azide group of 9-azide-9-
deoxy-DANA 2[20] with the terminal alkyne group of a com-
mercially available compound library (Figure 1d).
The order of the inhibition strength in the second screening
changed from the order in the first screening (3 f>3c>3e>
3g>3h>3affi3d>3b) and each selected compound (3a,
3c, and 3d) with an amphiphilic structure was suspected to
form an aggregate and act as a promiscuous inhibitor.[21] To
validate the inhibition mode of compounds 3a, 3c, and 3d,
the influence of the detergent Triton-X100 on inhibition
strength and concentrations of the inhibitors[22] was deter-
mined (Figure 3). The reserved structure of the sigmoid
curve after addition of detergent (0, 0.01, and 0.1%) clearly
showed that all three compounds interacted with VCNA in
a 1:1 ratio. Among these three compounds, 3c and 3d had a
slight shift in the sigmoid curve, which suggested weakening
of the inhibition property by the addition of the detergent.
This suggested that an independent hydrophobic interaction
was involved in the binding of VCNA and 3c (and 3d) as a
Screening From the Focused Library
For the first screening, the alkyne compounds (266 com-
pounds) were coupled with 2 in a uniform click reaction.[19]
The efficiency of each [2+3] cyclization reaction for 3 was
roughly estimated by ESI-TOF MS without any purification,
and 187 reaction mixtures with a ratio of intensity greater
1050
www.chemasianj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 1048 – 1056

Neuraminidase Inhibitors
Figure 2. Screening results of the focused library: a) result of the first screening for VCNA (0.15 mU) with substrate (2 mm) and click-reaction solutions
from 100 mm of 2 and alkyne compounds; b) effect of the reagents for click reactions towards VCNA activity: I) DMSO, II) tBuOH/H2O, III) Cu2+
,
IV) ascorbate, V) TBTA, VI) mixture of I–V, VII) blank; c) second screening for VCNA (0.15 mU) with substrate (1mm) and purified compounds 3a–h
(50 mm), and DANA. NI=no inhibitor.
Table 1. Km [mm] values of 4-MU-NANA and Ki [mm] values of DANA,
FANA, 3a, 3c, and 3d for viral, microbial, protozoal, and mammal neu-
source of inhibition potency. The dissociation constant for
inhibitor binding (Ki) values of DANA, FANA, 3a, 3c, and
3d were determined from the Lineweaver–Burk (LB) plots
as 1.3, 0.15, 1.4, 0.22, and 0.52 mm, respectively. The LB plot
showed that 3c and 3d exhibited a mixed inhibition mode
compared to that of DANA and FANA, which showed a
clear competitive inhibition mode (Figure 3). The detergent-
based assay and LB plot suggested that 3c and 3d bind the
catalytic site of VCNA with the DANA skeleton to provide
competitive inhibition, and the hydrophobic biphenyl
moiety presumably interacts with the hydrophobic loop
575LDRFFL580 and is attributed with providing a noncompe-
titive inhibition mode.
raminidases.
Compounds
VCNA
FluB NA
hNeu2
4-MU-NANA
DANA
9-N3 DANA (2)
3a
3c
3d
FANA
41
10
540
1.3 [32][a]
2.7 [15][a]
1.4 [29][a]
0.22 [186][a]
0.52 [79][a]
0.15 [273][a]
0.93 [11][a]
114 [4.7][a]
–
–
–
–
>1000 [<0.01][a]
–
501 [1.1][a]
–
0.33 [30][a]
43 [13][a]
[a] Km/Ki ratios are shown in brackets.
FluB NA (1= and 5 times the Ki values, respectively). This
3
The specificity of 3c for VCNA among the various neura-
minidases is also of great interest to validate our strategy.
DANA, FANA, and 3c were assayed for influenza B virus
NA (FluB NA) and human neuraminidase 2 (hNeu2) as
viral and mammal neuraminidases, respectively. As shown in
Table 1, 3c showed a high specificity for VCNA among the
three tested neuraminidases. The Km/Ki ratios (Km =Michae-
lis–Menten constant) of 3c and FANA for VCNA were
almost the same, but the ratios for FluB NA and hNeu2
were completely different. In the case of FluB NA, the N-
substitution from DANA to FANA affords half Ki values,
and the 9-substituent group from DANA to 3c affords more
than 1000 times greater Ki values. In the case of hNeu2,
each substitution resulted in a similar shift of Ki values as
suggests that the N-trifluoroacetyl group interacts with hy-
drophobic amino acid residues that are structurally stabi-
lized by b-sheet formation among the NAs (W311, W101,
and I103 residues in VCNA, FluB NA, and hNeu2, respec-
tively) to nonselectively enhance the affinity. On the other
hand, the 9-substituent group of 3c specifically interacts
with the “movable” hydrophobic loop of VCNA to enhance
affinity but repulses the corresponding position in FluB NA
and hNeu2.
Design, Preparation, and Evaluation of 4
The independent interaction of the 9-substitution group of
3c and the N-trifluoroacetyl group of FANA suggests that a
Chem. Asian J. 2011, 6, 1048 – 1056
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1051

H. Hinou, S.-I. Nishimura et al.
FULL PAPERS
Figure 3. Structure, sigmoid curve profile of detergent-dependent identification, and LB plot of a) 3a, b) 3c, and c) 3d, and LB plot of d) DANA and
e) FANA.
compound with both substituted groups, 4, can be a more
potent inhibitor for VCNA along with a high selectivity like
3c. Compound 4 was prepared from the methyl ester deriva-
tive of FANA (5)[10] in four steps (Scheme 2). Monotosyla-
tion of 5 gave 6, and this was followed by SN2 replacement
to give the 9-N3 derivative 7. Saponification of the methyl
ester group of 7 gave the 9-N3 analogue of FANA 8 with a
15% overall yield over three steps. Kinetic analysis of 8 by
using the LB plot showed that 8 competitively inhibited
VCNA with low Ki values (0.37 mm); the 9-N3 substitution of
8 did not affect the inhibition property of FANA. The same
tendency was observed with the introduction of an azide
group at the C9 atom of DANA. The click reaction of 8
with alkyne 9 gave 4 in 70% yield. An inhibition assay
showed that 4 has the lowest Ki value for VCNA (73 nm)
among the reported compounds[10,11] with a high selectivity
(Figure 4). The mixed inhibition mode and high selectivity
among the three NAs means that 4 maintains the inhibition
character of 3c. Although, the substitution effect decreased
relative to the change from DANA to FANA and 2 to 8, the
trifluoroacetyl group afforded additional interaction potency
to 3c as expected.
Conclusion
Protein dynamism, which is difficult to detect by means of
current methodologies, can be a potential target for the de-
1052
www.chemasianj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 1048 – 1056

Neuraminidase Inhibitors
Scheme 2. Preparation of 4. Reagents: a) TsCl, pyridine, 08C, 24 h, 40%;
b) NaN3, DMF, 708C, 6 h, 68%; c) NaOH, H2O, RT, 20 min, 66%;
d) tris(benzyltriazolylmethyl)amine (TBTA), ascorbate, CuSO4, H2O,
tBuOH, DMSO, 1 d, 70%.
velopment of a potent and selective inhibitor. Mechanism-
based labeling can be a versatile complement to current
drug designs based on the X-ray crystallographic structure
of the target proteins. In the case of neuraminidases, it was
anticipated that the structural change in the peripheral loop
structure of the catalytic domain can be a potential target of
novel drug design; however, a clear and practical approach
has not yet been reported. In this study, the unexpected
modification of amino acid residues observed in the mecha-
nism-based labeling study revealed the presence of an addi-
tional potential target loop and enabled us to design a fo-
cused library. A novel lead compound 3c with a potent and
selective inhibition property toward VCNA was found from
the focused library. The comparative kinetic study of 3c and
the most potent inhibitor, FANA, revealed the orthogonal
effect of modification of each compound from the common
DANA skeleton to give the most potent inhibitor of VCNA,
4. Development of a more potent VCNA inhibitor from this
lead compound is ongoing. This two-step approach could
afford a new strategy for the drug design of structurally
novel anti-influenza drugs and for targeting various infective
diseases and could complement conventional structure-
based drug design strategies as well as recent fragment-
based drug design.[23] Recently, we found that a derivative of
the suicide substrate 1 inhibits Trypanosoma cruzi trans-sia-
lidase (TcTS) to decrease the infection ratio of the protozoa
to the host cell, and the labeling moiety was revealed to be
another pair of Asp and Arg residues using 1.[24] We also
identified flavonoid and anthraquinone derivatives as poten-
tial and noncompetitive TcTS inhibitors from a natural
product library.[25] The development of more potent inhibi-
tors of TcTS using this information is currently in progress.
Figure 4. Inhibition property of 4. LB-plot and Ki value for a) VCNA and
b) hNeu2, and inhibition pattern for c) FluB NA.
Experimental Section
General
Molecular graphics in Figure 1 were generated and rendered with
PyMOL 0.99 software (http://www.pymol.org). 2-Deoxy-2,3-dehydro-N-
acetylneuraminic acid (DANA) and 5-acetamido-2,6-anhydro-9-azido-
3,5,9-tri-deoxy-d-glycero-d-galacto-non-2-enoic acid (2), methyl ester of
FANA (5), and 2-a-(4’-methylumbelliferyl)-N-acetylneuraminic acid (4-
MU-NANA) were synthesized by following published methods.[10,26]
Alkyne compounds were purchased from Thermo Fisher Scientific Inc.
(230 compounds) and Tokyo Chemical Industry Co., Ltd (36 com-
pounds). Vibrio cholerae neuraminidase (lot No. 088K4020) was pur-
chased from Sigma–Aldrich (Japan). Influenza B neuraminidase was do-
nated from SHIONOGI & Co., Ltd. Human neuraminidase 2 was pre-
pared as in a reported procedure.[12] The purity (>95%) of each com-
pound was also checked by 1H NMR spectroscopy (see the Supporting
Information). The activity of neuraminidase was measured by fluores-
cence spectrometry in the presence of 4-MU-NANA as a substrate using
a microplate reader (SpectraMax M5, Molecular Devices Co., Sunnyvale,
CA). Analyses of assay results were performed by using PRISM software
5.01 (GraphPad Software Inc., San Diego, CA).
Chem. Asian J. 2011, 6, 1048 – 1056
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1053

H. Hinou, S.-I. Nishimura et al.
FULL PAPERS
Construction of Compound Libraries for First Screening
tonon-2-enonate 5 (184 mg, 0.51 mmol) in pyridine (8 mL) and the mix-
ture was stirred under a nitrogen atmosphere at 08C for 24 h. Next,
MeOH (2 mL) was added to the mixture, which was then concentrated
under vacuum. Purification of the residue on a column of silica gel (4:3
A
solution of 2 (50 mm), alkyne compounds (50 mm), CuSO4·5H2O
(2.5 mm), and sodium ascorbate (25 mm) in tert-butyl alcohol/H2O (1:1),
and a solution of tris(benzyltriazolylmethyl)amine (TBTA; 2.5 mm) in di-
methyl sulfoxide (DMSO) were prepared, respectively. Then aliquots of
these solutions (10 mL) were mixed at room temperature and the mixture
was left for 8 h. The yields of triazole products 3 were roughly estimated
by comparison of the signal intensity of 2 and each triazole product by
ESIMS. The solutions that contained more than 70% yield of the triazole
product were used for the following first screening assays of VCNA as
100% yield solutions.
hexane/acetone) afforded compound
6
(105 mg, 40%). 1H NMR
(500 MHz, CD3OD): d=7.84 (d, J=8.3 Hz, 2H; aromatic), 7.47 (d, J=
8.3 Hz, 2H; aromatic), 5.97 (d, J=2.4 Hz, 1H; H-3), 4.52 (dd, J=2.4,
8.8 Hz, 1H; H-4), 4.37 (m, 2H; H-6, 9a), 4.17 (dd, J=9.0, 10.7 Hz, 1H;
H-5) , 4.13 (dd, J=6.1, 10.0 Hz, 1H; H-9b), 4.07 (ddd, J=2.0, 6.1, 8.9 Hz,
1H; H-8), 3.82 (s, 3H; COOCH3), 3.50 (d, J=8.6 Hz, 1H; H-7),
2.48 ppm (s, 3H; Ar-CH3); 13C NMR (125 MHz, CD3OD): d=162.82,
145.04, 143.48, 132.90, 129.62, 129.49, 127.76, 127.69, 112.35, 75.55, 72.25,
68.28, 67.76, 66.31, 51.50, 50.57, 20.18 ppm; HRMS (FAB): m/z calcd for
C19H21F3NO10S [MꢀH]ꢀ: 512.0844; found: 512.0844.
Preparation of 3a, 3c, and 3d
Aliquots (75 mL) of solutions of 2 (200 mm), CuSO4·5H2O (10 mm),
sodium ascorbate (100 mm), and alkyne compound (200 mm) in tert-butyl
alcohol/H2O (1:1), and of TBTA (10 mm) in DMSO were mixed at room
temperature and the mixtures were shaken on a vortex mixer overnight.
After checking the completion of the reactions by ESIMS, each reaction
mixture was purified by reversed phase (RP) HPLC to give 3a (31%),
3c (40%), and 3d (48%), respectively.
Methyl 2,6-anhydro-9-azido-3,5,9-trideoxy-5-trifluoroacetamido-d-glycero-
d-galactonon-2-enonate (7)
Sodium azide (78 mg, 1.2 mmol) was added to a solution of 6 (100 mg,
0.19 mmol) in N,N-dimethylformamide (DMF; 1.5 mL) and the mixture
was stirred at 708C. After 6 h, the reaction mixture was allowed to cool
to room temperature. Purification of the mixture on a column of silica
gel (2:1 hexane/acetone) afforded compound 7 (50 mg, 68%). 1H NMR
(500 MHz, CD3OD): d=5.95 (d, J=2.4 Hz, 1H; H-3), 4.49 (dd, J=2.3,
8.8 Hz, 1H; H-4), 4.40 (d, J=10.9 Hz, 1H; H-6), 4.19 (dd, J=1.4,
10.6 Hz, 1H; H-5), 4.03 (ddd, J=2.5, 6.5, 9.3 Hz, 1H; H-8), 3.79 (s, 3H;
COOCH3), 3.56 (dd, J=2.5, 10.8 Hz, 1H; H-9a), 3.49 (d, J=9.3 Hz, 1H;
H-7), 3.39 ppm (dd, J=6.3, 12.8 Hz, 1H; H-9b); 13C NMR (125 MHz,
CD3OD): d=162.90, 143.57, 112.37, 75.69, 69.20, 69.12, 66.46, 54.20,
51.48, 50.60 ppm; HRMS (FAB): m/z calcd for C12H14F3N4O7 [MꢀH]ꢀ:
383.0820; found: 383.0812.
5-Acetamido-2,6-anhydro-9-{4-[(3,5-dichlorobenzamido)cyclohexyl]-1H-
1,2,3-triazol-1-yl}-3,5,9-trideoxy-d-glycero-d-galactonon-2-enoic acid (3a)
1H NMR (600 MHz, CD3OD): d=7.92–7.62 (m, 4H; aromatic), 5.97 (d,
J=1.8 Hz, 1H; H-3), 4.84 (m, 1H; H-9a), 4.43 (m, 1H; H-4, 9b), 4.27
(ddd, J=2.4, 6.6, 9.0 Hz, 1H; H-8), 4.18 (d, J=10.8 Hz, 1H; H-6), 4.01
(dd, J=9.0, 10.8 Hz, 1H; H-5), 3.46 (d, J=9.0 Hz, 1H; H-7), 2.55 (m,
2H), 2.10 (m, 2H), 2.04 (s, 3H; Ac), 1.69–1.44 ppm (m, 6H); 13C NMR
(125 MHz, CD3OD): d=173.62, 138.59, 134.88, 130.57, 125.91, 112.13,
76.42, 70.00, 68.46, 66.50, 53.88, 50.53, 35.10, 34.91, 25.14, 21.62,
21.27 ppm; HRMS (FAB-MS): m/z calcd for C26H32Cl2N5O8 [M+H]+:
612.1628; found: 612.1626.
5-Trifluoroacetamido-2,6-anhydro-9-azido-3,5,9-trideoxy-d-glycero-d-
galactonon-2-enoic acid (8)
NaOH (1m, 0.1 mL) was added to a solution of 7 (25 mg, 0.07 mmol) in
H2O (1.6 mL) and the mixture was stirred at room temperature for
20 min. The reaction mixture was neutralized by the addition of Dowex
50Wꢂ8 resin. The mixture was filtered and lyophilized to give compound
8 (16 mg, 66%) as a white solid. The product was used for the following
reaction without further purification. 1H NMR (500 MHz, D2O): d=5.63
(d, J=2.2 Hz, 1H; H-3), 4.47 (dd, J=2.2, 8.9 Hz, 1H; H-4), 4.29 (d, J=
11.0 Hz, 1H; H-6), 4.14 (dd, J=9.3, 10.6 Hz, 1H; H-5), 4.03 (ddd, J=2.7,
5.8, 9.1 Hz, 1H; H-8), 3.57 (dd, J=2.7, 13.2 Hz, 1H; H-9a), 3.50 (d, J=
9.4 Hz, 1H; H-7), 3.43 ppm (dd, J=5.8, 13.1 Hz, 1H; H-9b); 13C NMR
(125 MHz, D2O): d=169.32, 147.88, 107.73, 74.58, 68.63, 68.60, 67.25,
53.66, 50.57 ppm; HRMS (FAB): m/z calcd for C11H12F3N4O7 [MꢀH]ꢀ:
369.0664; found: 369.0652.
5-Acetamido-2,6-anhydro-9-{4-[(4’-ethylbiphen-4-yl)carboxamidoprop-2-
yl]-1H-1,2,3-triazol-1-yl}-3,5,9-trideoxy-d-glycero-d-galactonon-2-enoic
acid (3c)
1H NMR (600 MHz, CD3OD): d=7.92–7.30 (m, 9H; aromatic), 5.97 (d,
J=2.4 Hz, 1H; H-3), 4.85 (m, 1H; H-9a), 4.43 (m, 2H; H-4, 9b), 4.29
(ddd, J=2.4, 6.6, 9.0 Hz, 1H; H-8), 4.19 (dd, J=1.2, 10.8 Hz, 1H; H-6),
4.01 (dd, J=8.4, 10.8 Hz, 1H; H-5), 3.48 (dd, J=1.2, 9.0 Hz, 1H; H7),
2.70 (q, J=7.8 Hz, 2H; CH2), 2.05 (s, 3H; Ac), 1.84 (m, 6H; CH3 ꢂ2),
1.28 ppm (t, J=7.8 Hz, 1H; CH2CH3); 13C NMR (125 MHz, CD3OD):
d=173.78, 144,18, 137.26, 133.41, 128.40, 128.10, 127.64, 126.64, 126.26,
122.46, 112.11, 76.41, 69.97, 68.50, 66.52, 54.03, 50.53, 28.10, 27.05, 21.27,
14.67 ppm; HRMS (FAB-MS): m/z calcd for C31H36N5O8 [MꢀH]ꢀ:
606.2569; found: 606.2562.
2,6-Anhydro-9-{4-[(4’-ethylbiphen-4-yl)carboxamido-prop-2-yl]-1H-1,2,3-
triazol-1-yl}-3,5,9-trideoxy-5-trifluoroacetamido-d-glycero-d-galactonon-
2-enoic acid (4)
5-Acetamido-2,6-anhydro-9-{4-[(4’-ethylbiphen-4-yl)carboxamino-pent-3-
yl]-1H-1,2,3-triazol-1-yl}-3,5,9-trideoxy-d-glycero-d-galactonon-2-enoic
acid (3d)
Solutions of azido 8 (200 mm), CuSO4·5H2O (10 mm), and sodium ascor-
bate (100 mm) in tert-butyl alcohol/H2O (1:1) and solutions of alkyne 9
(300 mm) and TBTA (10 mm) in DMSO were prepared, respectively. Ali-
quots (135 mL) of each solution were mixed at room temperature and left
for 24 h. The reaction mixture was purified by HPLC to give 4 (13 mg,
1H NMR (600 MHz, CD3OD): d=7.96–7.31 (m, 9H; aromatic), 5.98 (d,
J=2.4 Hz, 1H; H-3), 4.88 (m, 1H; H-9a), 4.48 (dd, J=7.8, 14.4 Hz, 1H;
H-9b), 4.44 (dd, J=2.4, 9.0 Hz, 1H; H-4), 4.30 (ddd, J=2.4, 7.8, 9.0 Hz,
1H; H-8), 4.18 (d, J=10.8 Hz, 1H; H-6), 4.01 (dd, J=9.0, 10.8 Hz, 1H;
H-5), 3.47 (d, J=9.0 Hz, 1H; H-7), 2.71 (q, J=7.8 Hz, 2H; Ar-CH2CH3),
2.28 (m, 4H; CH2 ꢂ2), 2.04 (s, 3H; Ac), 1.28 (t, J=7.8 Hz, 3H; Ar-
CH2CH3), 0.84 ppm (m, 6H; CH3 ꢂ2); 13C NMR (125 MHz, CD3OD):
d=173.81, 168.08, 164.09, 150.65, 144,21, 143.98, 137.23, 135.39, 133.00,
128.10, 127.50, 126.66, 126.38, 123.31, 112.14, 76.43, 70.00, 68.42, 66.48,
54.09, 50.55, 28.52, 28.52, 28.11, 21.28, 14.68, 6.62, 6.62 ppm; HRMS
(FAB-MS): m/z calcd for C33H40N5O8 [MꢀH]ꢀ: 634.2882; found:
634.2859.
70%) as
a
white powder after lyophilization. 1H NMR (500 MHz,
(CD3)2CO): d=8.69 (d, 1H; CF3CONHR), 7.95 (s, 1H; N-CH=CRN),
7.91 (d, J=8.3 Hz, 2H; aromatic), 7.82 (s, 1H; Ar-CONHR), 7.71–7.61
(m, 4H; aromatic), 7.33 (d, J=8.2 Hz, 2H; aromatic), 5.98 (d, J=2.5 Hz,
1H; H-3), 4.91 (dd, J=2.3, 14.0 Hz, 1H; H-9a), 4.71 (dd, J=2.5, 8.6 Hz,
1H; H-4), 4.54 (d, J=10.6 Hz, 1H; H-6), 4.42 (dd, J=8.3, 14.0 Hz, 1H;
H-9b), 4.31–4.28 (m, 2H; H-5,8), 3.60 (dd, J=0.8, 8.9 Hz, 1H; H-7), 2.69
(q, J=7.6 Hz, 2H; CH2CH3), 1.85 (s, J=7.6 Hz, 6H; CH3 ꢂ2), 1.24 ppm
(t, J=7.6 Hz, 3H; CH2CH3); 13C NMR (125 MHz, (CD3)2CO): d=
165.93, 165.87, 152.97, 144.11, 143.82, 143.46, 137.35, 134.23, 128.42,
127.75, 126.92, 126.38, 122.26, 112.39, 75.65, 70.16, 69.26, 66.26, 53.80,
15.10 ppm; HRMS (FAB): m/z calcd for C31H33F3N5O8 [MꢀH]ꢀ:
660.2287; found: 660.2289.
Methyl 2,6-anhydro-3,5-dideoxy-9-O-tosyl-5-trifluoroacetamido-d-
glycero-d-galactonon-2-enonate (6)
At ꢀ108C, tosyl chloride (251 mg, 1.32 mmol) was added to a solution of
methyl 2,6-anhydro-3,5-dideoxy-5-trifluoroacetamido-d-glycero-d-galac-
1054
www.chemasianj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 1048 – 1056

Neuraminidase Inhibitors
VCNA: First and Second Screening
Acknowledgements
Each compound of the first and second screening library (100 and 50 mm,
respectively) was incubated with 4 mm CaCl2, VCNA, and 4-MU-NANA
in sodium acetate buffer (100 mm, pH 6.5) at 378C for 5 min. The relative
inhibition ratios of VCNA by the first and second screening library were
estimated by using a microplate reader at excitation and emission wave-
lengths of 365 and 450 nm, respectively.
This work was supported by funding from the Ministry of Education,
Culture, Science, Sports and Technology of Japan for “Grant-in-Aid for
Young Scientists (A)” and “Innovation COE Program for Future Drug
Discovery and Medical Care”. We thank Ms. M. Kikuchi, Ms. S. Oka,
and Mr. T. Hirose at the Center for Instrumental Analysis, Hokkaido
University, for ESIMS measurements.
Detergent-Based Assays
The experiments for detergent-based assays were performed by using
150 mm 4-MU-NANA concentrations and several concentrations of inhib-
itor. The assay mixture—containing 4 mm CaCl2, 0, 0.01, or 0.1% Triton
X-100, 100 mgmLꢀ1 bovine serum albumin (BSA), 0.076 mU VCNA, and
inhibitor in 2-(N-morpholino)ethanesulfonic acid (MES)/NaOH buffer
(33 mm, pH 6.5)—was preincubated for 30 min at 258C and the reaction
was initiated by addition of 4-MU-NANA in a final volume of 120 mL.
After incubation (10, 20, 30, and 40 min), aliquots (20 mL) of the reaction
mixture were taken and the enzymatic reaction was quenched by addi-
tion to a 100 mm glycine solution at pH 10.8 (80 mL) in a 96-well plate.
The fluorescence of the released product (4-MU) was measured at 258C
with excitation and emission wavelengths of 365 and 450 nm, respectively.
Each sigmoid curve was derived from a plot of percent activity versus in-
hibitor concentration.
[1] a) T. Angata, A. Varki, Chem. Rev. 2002, 102, 439–469; b) A. Varki,
Nature 2007, 446, 1023–1029.
[2] E. Severi, D. W. Hood, G. H. Thomas, Microbiology 2007, 153,
2817–2822.
[3] A. Gottschalk, Biochim. Biophys. Acta 1957, 23, 645–646.
[4] J. E. Galen, J. M. Ketley, A. Fasano, S. W. Richardson, S. S. Wasser-
man, J. B. Kaper, Infect. Immun. 1992, 60, 406–415.
[5] S. Almagro-Moreno, E. F. Boyd, Infect. Immun. 2009, 77, 3807–
3816.
[6] S. Crennell, E. Garman, G. Laver, E. Vimr, G. Taylor, Structure
1994, 2, 535–544.
[7] M. von Itzstein, Curr. Opin. Struct. Biol. 2008, 18, 558–566.
[8] M. von Itzstein, W. Y. Wu, G. B. Kok, M. S. Pegg, J. C. Dyason, B.
Jin, T. Van Phan, M. L. Smythe, H. F. White, S. W. Oliver, P. M.
Colman, J. N. Varghese, D. M. Ryan, J. M. Woods, R. C. Bethell,
V. J. Hotham, J. M. Cameron, C. R. Penn, Nature 1993, 363, 418–
423.
[9] C. U. Kim, W. Lew, M. A. Williams, H. T. Liu, L. J. Zhang, S. Swami-
nathan, N. Bischofberger, M. S. Chen, D. B. Mendel, C. Y. Tai, W. G.
Laver, R. C. Stevens, J. Am. Chem. Soc. 1997, 119, 681–690.
[10] D. Meindl, G. Bodo, P. Palese, J. Schulman, H. Tuppy, Virology
1974, 58, 457–463.
[11] a) P. Florio, R. J. Thomson, A. Alafaci, S. Abo, M. von Itzstein,
Bioorg. Med. Chem. Lett. 1999, 9, 2065–2068; b) J. C. Wilson, R. J.
Thomson, J. C. Dyason, P. Florio, K. J. Quelch, S. Abo, M. von Itz-
stein, Tetrahedron: Asymmetry 2000, 11, 53–73; c) H. Streicher, C.
Bohner, Tetrahedron 2002, 58, 7573–7581; d) M. C. Mann, R. J.
Thomson, M. von Itzstein, Bioorg. Med. Chem. Lett. 2004, 14, 5555–
5558; e) M. C. Mann, R. J. Thomson, J. C. Dyason, S. McAtamney,
M. von Itzstein, Bioorg. Med. Chem. 2006, 14, 1518–1537; f) H.
Streicher, H. Busse, Bioorg. Med. Chem. 2006, 14, 1047–1057.
[12] L. M. G. Chavas, C. Tringali, P. Fusi, B. Venerando, G. Tettamanti,
R. Kato, E. Monti, S. Wakatsuki, J. Biol. Chem. 2005, 280, 469–475.
[13] R. J. Russell, L. F. Haire, D. J. Stevens, P. J. Collins, Y. P. Lin, M.
Blackburn, A. J. Hay, S. J. Gamblin, J. J. Skehel, Nature 2006, 443,
45–49.
Determination of Ki values
The experiments for Ki determination were performed in 30 to 150 mm 4-
MU-NANA concentrations and several concentrations of inhibitor. The
assay mixture—containing 33 mm MES, 4 mm CaCl2, 100 mgmLꢀ1 BSA,
0.076 mU VCNA, and inhibitor in MES/NaOH buffer (33 mm, pH 6.5)—
was preincubated for 30 min at 258C and the reaction was initiated by
addition of 4-MU-NANA in a final volume of 120 mL. After incubation
(10, 20, 30, and 40 min), aliquots (20 mL) of the reaction mixture were
taken and the enzymatic reaction was quenched by addition to a 100 mm
glycine solution at pH 10.8 (80 mL) in a 96-well plate. The fluorescence of
the released product (4-MU) was measured at 258C with excitation and
emission wavelengths of 365 and 450 nm, respectively. The Ki value and
the competitive nature of the inhibition were determined by Linewea-
ver–Burk plot analysis.
FluB NA
The experiments for Ki determination were performed in 8 to 30 mm 4-
MU-NANA concentrations and several concentrations of inhibitor. The
assay mixture—containing 4 mm CaCl2, 0.067 mU Influenza B, and inhibi-
tor in MES/NaOH buffer (33 mm, pH 6.5)—was preincubated for 30 min
at 258C and reaction was initiated by addition of 4-MU-NANA in a final
volume of 120 mL at 378C. After incubation (10, 20, 30, and 40 min), ali-
quots (20 mL) of the reaction mixture were taken and the enzymatic reac-
tion was quenched by addition to a 100 mm glycine solution at pH 10.8
(80 mL) in a 96-well plate. The fluorescence of the released product (4-
MU) was measured at 258C with excitation and emission wavelengths of
365 and 450 nm, respectively. The Ki value and the competitive nature of
the inhibition were determined by Lineweaver–Burk plot analysis.
[14] W.-H. Wen, S.-Y. Wang, K.-C. Tsai, Y.-S. E. Cheng, A.-S. Yang, J.-M.
Fang, C.-H. Wong, Bioorg. Med. Chem. 2010, 18, 4074–4084.
[15] H. Hinou, S.-I. Nishimura, Curr. Top. Med. Chem. 2009, 9, 106–116.
[16] S. Halazy, V. Berges, A. Ehrhard, C. Danzin, Bioorg. Chem. 1990,
18, 330–344.
[17] H. Hinou, M. Kurogochi, H. Shimizu, S.-I. Nishimura, Biochemistry
2005, 44, 11669–11675.
[18] I. Moustafa, H. Connaris, M. Taylor, V. Zaitsev, J. C. Wilson, J. M.
Kiefel, M. von Itzstein, G. Taylor, J. Biol. Chem. 2004, 279, 40819–
40826.
[19] H. C. Kolb, M. G. Finn, K. B. Sharpless, Angew. Chem. 2001, 113,
2056–2075; Angew. Chem. Int. Ed. 2001, 40, 2004–2021.
[20] T. G. Warner, A. Louie, M. Potier, A. Ribeiro, Carbohydr. Res. 1991,
215, 315–321.
[21] B. Y. Feng, A. Shelat, T. N. Doman, R. K. Guy, B. K. Shoichet, Nat.
Chem. Biol. 2005, 1, 146–148.
[22] B. Y. Feng, B. K. Shoichet, Nat. Protoc. 2006, 1, 550–553.
[23] a) S. H. Rotstein, M. A. Murcko, J. Med. Chem. 1993, 36, 1700–
1710; b) P. J. Hajduk, J. Greer, Nat. Rev. Drug Discovery 2007, 6,
211–219.
Neu2
The experiments for Ki determination were performed in 200 to 750 mm
4-MU-NANA concentrations and several concentrations of inhibitor. The
assay mixture—containing 100 mm Na citrate/phosphate pH 6.5, buffer,
100 mgmLꢀ1 BSA, 0.30 mU Neu2, and inhibitor solution—was preincu-
bated for 30 min at 258C and the reaction was initiated by addition of 4-
MU-NANA in a final volume of 120 mL at 378C. After incubation (10,
20, 30, and 40 min), aliquots (20 mL) of the reaction mixture were taken
and the enzymatic reaction was quenched by addition to a 100 mm gly-
cine solution at pH 10.8 (80 mL) in a 96-well plate. The fluorescence of
the released product (4-MU) was measured at 258C with excitation and
emission wavelengths of 365 and 450 nm, respectively. The Ki value and
the competitive nature of the inhibition were determined by Linewea-
ver–Burk plot analysis.
[24] S. T. Carvalho, M. Sola-Penna, I. A. Oliveira, S. Pita, A. S. Gon-
Åalves, B. C. Neves, F. R. Sousa, L. Freire-de-Lima, M. Kurogochi,
Chem. Asian J. 2011, 6, 1048 – 1056
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
1055

H. Hinou, S.-I. Nishimura et al.
FULL PAPERS
H. Hinou, S.-I. Nishimura, L. MendonÅa-Previato, J. O. Previato,
A. R. Todeschini, Glycobiology 2010, 20, 1034–1045.
nolds, Y. Uchida, Anal. Biochem. 1980, 101, 166–174; c) T. G.
Warner, J. S. OꢃBrien, Biochemistry 1979, 18, 2783–2787.
[25] S. Arioka, M. Sakagami, R. Uematsu, H. Yamaguchi, H. Togame, H.
Takemoto, H. Hinou, S.-I. Nishimura, Bioorg. Med. Chem. 2010, 18,
1633–1640.
Received: August 20, 2010
Published online: February 8, 2011
[26] a) H. Kamei, Y. Kajihara, Y. Nishi, Carbohydr. Res. 1999, 315, 243–
250; b) R. W. Myers, R. T. Lee, Y. C. Lee, G. H. Thomas, L. W. Rey-
1056
www.chemasianj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2011, 6, 1048 – 1056