Macrocyclic mechanism-based inhibitor for neuraminidases

Article abstract of DOI:10.1002/chem.201200859

A macrocyclic mechanism-based inhibitor for neuraminidases (NAs) bearing a 2-difluoromethylphenyl aglycone and a linker between the aglycone and C-9 positions of sialic acid was synthesized and evaluated. The macrocyclic structure was designed to keep the aglycone moiety in the active site of the neuraminidase after cleavage of the glycoside bond. When Vibrio chorelae neuraminidase (VCNA) was treated with a similar acyclic derivative in the presence of detergent, the irreversible inhibition property was disabled. In contrast, this macrocyclic compound acted as an irreversible inhibitor for VCNA in the presence of detergent. Inhibition assay for various NAs using this macrocyclic compound revealed that the irreversible inhibition property depends on the k_cat of the neuraminidase treated. NAs having small k _cat values, such as Influenza viruses, Clostridium, Trypanosoma cruzi, and Human, were also inhibited irreversibly. However, Salmonella typhimurium NA, which has an extremely high k_cat, was not affected irreversibly by the inhibitor. Interestingly, in contrast to common k _cat inhibitors, the irreversibility of inhibition by this macrocyclic compound is inversely proportional to the k_cat of the target neuraminidase. Copyright

Full text of DOI:10.1002/chem.201200859

DOI: 10.1002/chem.201200859
Macrocyclic Mechanism-Based Inhibitor for Neuraminidases
Hirokazu Kai, Hiroshi Hinou,* Kentaro Naruchi, Takahiko Matsushita, and
Shin-Ichiro Nishimura*[a]
Abstract: A macrocyclic mechanism-
based inhibitor for neuraminidases
(NAs) bearing a 2-difluoromethylphen-
yl aglycone and a linker between the
aglycone and C-9 positions of sialic
acid was synthesized and evaluated.
The macrocyclic structure was designed
to keep the aglycone moiety in the
active site of the neuraminidase after
cleavage of the glycoside bond. When
tergent, the irreversible inhibition
property was disabled. In contrast, this
macrocyclic compound acted as an ir-
on the kcat of the neuraminidase treat-
ed. NAs having small kcat values, such
as Influenza viruses, Clostridium, Try-
panosoma cruzi, and Human, were also
inhibited irreversibly. However, Salmo-
nella typhimurium NA, which has an
extremely high kcat, was not affected ir-
AHCTUNGTRENGNrUN eversible inhibitor for VCNA in the
presence of detergent. Inhibition assay
for various NAs using this macrocyclic
compound revealed that the ir-
reversible inhibition property depends
ACHTUNGTRENUNrNG eversibly by the inhibitor. Interesting-
ly, in contrast to common kcat inhibi-
tors, the irreversibility of inhibition by
this macrocyclic compound is inversely
proportional to the kcat of the target
neuraminidase.
Keywords: drug design · enzyme
catalysis · inhibitors · macrocycles ·
metathesis
Vibrio
chorelae
neuraminidase
(VCNA) was treated with a similar
acyclic derivative in the presence of de-
Introduction
types of inhibitors depend on the kcat of their target en-
zymes; these inhibitors are called kcat inhibitors.[3] In the
Neuraminidases (NAs, EC 3.2.1.18), a family of glycoside
hydrolases that cleave the glycosidic linkages of sialic acid
(Neu5Ac, N-acetylneuraminic acid) from sialylglycoconju-
gates, regulate the interactions between a cell and the ex-
tracellular world during infection, development, inflamma-
context of glyACTHNGUTRENNUcG oACTHUNGTRENsNUNG idases, 2- or 4-difluoromethylphenyl glyco-
sides, which were first reported in 1990 by Danzin et al.,[4]
are some of the most utilized mechanism-based inhibitors
(Figure 1a). These aglycones are hydrolytically released by
their target enzyme (Figure 1b, E·I), and they rapidly re-
lease HF to be transformed into their reactive fluorinated
quinone methide (Figure 1b, E·I*). These intermediates are
highly reactive electrophilic species that can form covalent
bonds with their target enzyme (Figure 1b, E–I) and ir-
tion, antigenicity, and cell–cell adhesion.[1] In higher or
nisms, a NA gene family (Neu1-4) are expressed in cells
ACHTUNGERTNgNUNG a-
ACHTUNGTRENNUNG
with particular cell distribution and specificity for homeo-
static control of the sialylglycoconjugates.[1c] In contrast, a
large number of virus and microbes display or secrete NAs
for propagation in a host cell or organism. Thus, potent and
selective inhibitors of NAs could be a potential field of
study, not only as tools to assess the biological functions of
the NAs but also to develop therapeutic agents that could
be used to treat infectious diseases, such as influenza.[2]
Mechanism-based inhibitors — a class of irreversible inhibi-
tors — are substrates that have cryptic reactive moieties
that are specifically activated by the actions of their target
enzymes. Commonly, the activation efficiencies of these
AHCTUNGTREGrNNUN eversibly inhibit the activity of the enzyme; the latter result
occurs when the nucleophile is an amino acid residue of the
enzyme. This mechanism has been applied to multiple glyco-
sidases, including galactosidases[5] and N-acetyl glucosamini-
dase,[6] and much attention has been paid to the NAs.[7] For
example, we reported a mechanism-based labeling study of
Vibrio cholerae neuraminidase (VCNA) using 2-difluoro-
AHCTUNGTREGmNNUN ethylphenyl sialoside 1a, which has a dansyl group as a
fluorescence reporter moiety (Figure 1a), and, following a
proteomic study, MS/MS analysis identified that the labeled
residues were located in the peripheral hydrophobic loop in
the catalytic domain.[7c] Using this labeling information, we
also designed a novel inhibitor skeleton and created a highly
potent and selective inhibitor for VCNA.[8] These results
showed that a mechanism-based inhibitor for NAs can, in
addition to being a drug candidate, also be a versatile tool
with which to elucidate the catalytic mechanism of a target
enzyme and develop a potent and selective inhibitor using
the labeling information. However, some recent studies indi-
cate that the labeling efficiencies of these types of mecha-
nism-based inhibitors are not high because the activations of
[a] Dr. H. Kai, Dr. H. Hinou, Dr. K. Naruchi, Dr. T. Matsushita,
Prof. S.-I. Nishimura
Graduate School of Life Science
and Faculty of Advanced Life Science
Hokkaido University, N21, W11
Kita-ku, Sapporo 001-0021 (Japan)
Fax : (+81)11-707-9042
E-mail: hinou@sci.hokudai.ac.jp
shin@sci.hokudai.ac.jp
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201200859.
1364
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1364 – 1372

FULL PAPER
characterized NA among mammalian NA family, found
from the focused library, suggested that this type of design
affected the binding property of the sialoside (k1, k2) but not
the aglycone alone (k4). In another study, a hydrophobic
loop in the active site domain of VCNA was labeled by a
pair of aglycone moieties of 1a.[7c] A detergent-based
assay[11] of compound 1a against VCNA in the presence or
absence of 0.01% Triton X-100 was used, and the time-de-
pendent inhibition by compound 1a almost disappeared
upon addition of 0.01% Triton X-100 (see Figure S1 in the
Supporting Information). This result suggested that the hy-
drolyzed aglycone moiety of compound 1a interacted with
the catalytic domain of VCNA mostly through hydrophobic
interactions. Whereas the hydrophobic interaction is impor-
tant for the recognition of a sugar residue by the gly
dases and lectins,[12] the interaction, which is mainly hydro-
phobic, is weak and of low selectivity, and could cause non-
specific inhibition to occur.[11] To overcome this diffusion
ACHTUNGTRENNUNGcoACHTUNGTRENNUNGsi-
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
property of the difluoromethylphenyl-type glycosides, we
planned to add a covalent bond between the sialic acid and
aglycone moiety to form a macrocyclic structure based on a
catalytic feature of the NAs (Figure 1c and Figure 1d). By
using 3-fluorosialic acid derivatives and Tripanosoma cruzi
trans-sialidase (TcTS), a conserved tyrosine residue of the
NAs was identified to act as the catalytic nucleophile that
forms a b-linked sialic acid-NA intermediate during the hy-
drolysis.[13] Before the covalently bonded sialic acid moiety
hydrolyzed, the aglycone moiety of the substrate dissociates
from the active site of the NA and is replaced by a water
molecule involved in the hydrolysis to release a-sialic acid
as the hydrolyzed product (Figure 1c).[9b] Although forma-
tion of the covalent sialic acid–NA bond might be not be
the rate determining step in the hydrolytic mechanism for
most NAs,[14] the ordered mechanism is conserved in all nat-
ural NAs having a retention mechanism.[14,15] Therefore, the
additional sialic acid–aglycone linkage at the nonanomeric
position could delay the dissociation of the aglycone moiety
from the active site of NA during the hydrolysis process of
the b-linked sialic acid–NA complex (Figure 1c, k5 ꢀk7; Fig-
ure 1d), and if the aglycone had a difluoromethylphenyl
type structure, the efficiency of the mechanism-based label-
ing could be improved. In addition, the remaining aglycone
moiety could disturb the binding step of a water molecule
(k5).[10] In this paper, we designed and synthesized a novel
mechanism-based NA inhibitor with a macrocyclic structure.
The difluoromethylphenyl-type aglycone moiety of this in-
hibitor stays within the active site of the NA after enzymatic
cleavage of the sialoside bond, and forms a covalent bond
with a nucleophilic amino acid side chain. To evaluate the
effect of the macrocyclic structure separate from the hydro-
phobic interaction, the inhibitor was assayed in the presence
of Triton X-100 (0.01%) against three bacterial NAs:
VCNA, Salmonella typhimurium neuraminidase (STNA),
and Clostridium perfringes neuraminidase (CPNA). Further-
more, following the bacterial assays, protozoan, viral, and
mammalian NAs: TcTS, Influenza A neuraminidase (FluA
NA, A/Puerto Rico/8/34), Influenza neuraminidase (FluB
Figure 1. Structures and mechanisms of the mechanism-based inhibitors
and NAs. a) Structures of the 2- and 4-difluoromethylphenyl glycosides
and sialic acid derivatives 1a and 1b. b) Plausible reaction mechanism
for the enzymatic activation and covalent bond formation of the difluoro-
methylphenyl moiety in the presence of glycosidase. c) Catalytic mecha-
nism of NAs. d) Macrocyclization of sialoside and its hydrolyzed product
by NAs.
their difluoroACHTUNGTRENNUNGmethylphenol groups to form their respective
reactive quinone methides and subsequent covalent bond
formation are relatively slow processes compared with diffu-
sion of the difluoromethylphenol groups from the cavities of
the active sites of the hydrolases.[9] As shown in Figure 1c,
hydrolysis of sialosides by NAs proceeds in a manner that
retains the anomeric configuration through a ping-pong
mechanism.[9b] In this mechanism, the hydrolyzed aglycone
moiety rapidly diffuses from the active site (k4) and is re-
placed by a water molecule (k5), and the difluoromethylphe-
nol group must activate and form a covalent bond with NAs
faster than this diffusion. Recently, Withers et al. applied
this diffusion property to test multiple fluorescent aglycone
labels in a specific cell in which each corresponding enzyme
was expressed.[10] To improve the labeling efficiency and se-
lectivity of the 2-difluoromethylphenyl-type sialoside for
NAs, we recently reported an aglycone focused library of
the sialoside that was constructed by using the click reaction
between an azide terminated sialoside 1b and an alkyne-ter-
minated compound library.[7f] Analysis of the kinetics of the
reaction with the best inhibitor for VCNA or human neur-
ACHTUNGTRENNUNGaminidase 2 (hNeu2), cytosolic NA and the only structurally
Chem. Eur. J. 2013, 19, 1364 – 1372
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1365

H. Hinou, S.-I. Nishimura et al.
NA, B/Hong Kong/5/72), and hNeu2 were also evaluated to
(5),[18] which can be made from the 9-O-selective allylation
elucidate the universal properties of this inhibitor.
of the sialic acid residue (Figure 2c).
Synthesis of the macrocyclic mechanism-based inhibitor 2:
3-Allyl-5-nitrosalicylaldehyde (5) was prepared from 5-nitro-
salicylaldehyde through O-allylation and Claisen rearrange-
ment (Scheme 1). Compound 5 was coupled with crude gly-
cosyl chloride 4, which was prepared from the methyl ester
of sialic acid 6 in one step,[18] under basic conditions, to
afford 7 in 62% yield. During this step, pretreatment of 5
with sodium methoxide was required before coupling with 4
in N,N-dimethylformamide (DMF). One of the common
side reactions that occur during the coupling with diiso-
Results and Discussion
Design of the macrocyclic mechanism-based inhibitor 2: As
shown in Figure 2, the 9-OH position of the sialic acid
moiety was chosen as the modification point for cyclization
based on the following three considerations: 1) the co-crys-
tal structure of VCNA with 2,3-dehydro-2-deoxy-N-acetyl-
neuraminic acid (DANA) (PDB: 1W0O, Figure 2a and Fig-
ure 2b)[15a] indicate that the C-3 to C-5 positions, the
N-acetyl group, and the b-face of sialic acid are within the
cavity of VCNA, whereas the ring oxygen atom and glycerol
moiety (C-7 to C-9 positions) lie outside of the binding
cavity; 2) among the three hydroxyl groups of the glycerol
moiety, we have reported that the C-9 modified sialic acid
derivative can bind to the catalytic domain of VCNA; and[8]
3) among the eukaryotic, prokaryotic, and viral NAs, all of
the reported co-crystal structures of the NAs with DANA
have clefts between the glycerol moiety and the a-face of
the anomeric carbon (C-2) over the ring oxygen of the sialic
acid.[15] To construct the cyclic structure, transition-metal-
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
due to the low reactivity of the phenol derivative 4, owing
to the steric hindrance caused by the substituent groups
ortho to the phenolic hydroxyl group. The formyl group of 7
was converted into a difluoromethyl group by treatment
with diethylamino sulfur trifluoride (DAST) to give 8 in
1
63% yield.[4] The H NMR chemical shifts of 8 indicate that
this compound is composed of two structures, which may
arise from rotational isomers that can form from the ortho-
disubstituted phenolic aglycone of 8. The nitro group of 8
was reduced with zinc powder in glacial acetic acid, and
then treated with tbutyl nitrite (tBuONO) and azidotrime-
thylsilane (TMSN3)[20] to give compound 9 in 55% yield.
De-O-acetylation in NaOMe/MeOH and subsequent protec-
tion of the hydroxyl groups with acrolein diethylacetal in
DMF and a catalytic amount of camphorsulfonic acid at
room temperature[21] afforded a (1’R) and (1’S) diastereo-
meric mixture of 8,9-O-(prop-2-enylidene) acetal, 10a and
catalyzed olefin metathesis, especially ring-closing me
ACHTUNGTNERtNUNG ath-
esis (RCM)[16] using a ruthenium catalyst, was adopted.
Therefore, the 14-membered cyclic mechanism-based inhibi-
tor 2 for the NAs was designed starting from a di-allyl inter-
mediate 3 by RCM reaction. The di-allyl intermediate 3
could, in turn, be synthesized from methyl 5-acetomido-
4,7,8,9-tetra-O-acetyl-2-chloro-2,3,5-trideoxy-b-d-galacto-2-
nonuropyranosonate (4)[17] and 3-allyl-5-nitrosalicylaldehyde
Figure 2. Structure-based analysis of the VCNA (PDB ID: 1W0O) active site, and design of the macrocyclic mechanism-based inhibitor. a) Recognition
mechanism of DANA by the catalytic domain of VCNA. DANA and the conserved amino acid residues (W311, D250, E619, Y740, E756, R224, R635,
and R712) and other residues (R245, N318, D292, and D637) of the active site that interact with DANA are represented by stick models. b) Surface
model of the VCNA active site in which DANA is placed in the deep pocket of the active site. c) Design of the novel cyclic mechanism-based inhibitor 2
and its retrosynthetic analysis.
1366
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1364 – 1372

Neuraminidase Inhibitors
FULL PAPER
each diastereomer on an analytical scale. The diastereomeric
mixture of 10a and 10b was acetylated by Ac2O/pyridine to
give the corresponding mixture of 11a and 11b, which could
not be separated.
To identify the structure and stereochemistry of the allyli-
dene acetal groups, a set of standard NMR experiments (1H,
DQF-COSY, TOCSY, HSQC, HMBC, and NOESY) were
performed on the isolated analytical samples of 10a, 10b,
1
11a, and 11b. According to the H NMR chemical shifts pre-
sented in Table 1 for the neuraminic acid region, the H-7
chemical shifts for compounds 7–9, 11a, and 11b were in the
range 5.12–5.53 ppm, whereas those for compounds 10a and
10b were 3.43 and 3.53 ppm, respectively. Additionally, the
H-8 chemical shifts of compounds 7–9 were between 4.99–
5.30 ppm, whereas those of compounds 10a, 10b, 11a, and
11b were between 4.28–4.37 ppm. These results clearly indi-
cate that the 8-OH and 9-OH groups were protected by the
allylidene acetal. In addition, HMBC experiments demon-
strated long-range H-acetal/H-8 correlations (see Figure S2
in the Supporting Information). The results of the NOESY
experiments and the stereochemistries of compounds 11a
and 11b are shown in Figure 3a and Figure 3b, respectively.
For 11a, NOE cross-peaks between H-acetal/H-8, H-acetal/
H-9b, and H-8/H-9b were observed. However, in 11b, NOE
cross-peaks between H-acetal/H-9b and H-9a/H-8 were only
observed, and the cross-peaks between H-acetal/H-8 were
absent. Therefore, from these results, it is confirmed that
11a has (R) configuration and 11b has (S) configuration at
the acetal carbon.
Our first regioselective, reductive ring opening reactions
of the allylidene acetal mixture of 11a and 11b were at-
tempted by using sodium cyanoborohydride/hydrogencloride
(NaBH3CN/HCl)[22] and triethylsilane/trifluoromethanesul-
fonic acid (Et3SiH/TfOH)[23] conditions, which afforded
complex mixtures containing the desired product 12 (detect-
ed by TLC and ESI-MS analyses). After several attempts,
regioselective cleavage of the allylidene mixture was ach-
ieved by using borane-triethylamine/aluminum chloride
(Et3N·BH3/AlCl3),[24] which delivered the desired compound
12 in high yield (83%); compound 12 was then acetylated to
give 3 (95%). The difference between the H-8 chemical
shifts for 12 and 3 indicated the regioselectivity of the reduc-
tive ring-opening reaction under these conditions (Table 1).
Scheme 1. Synthesis of the macrocyclic mechanism-based inhibitor 2. Re-
agents and conditions: a) AcCl; b) i. allyl bromide, Bu4NI, K2CO3,
CH3CN, reflux; ii. 2008C, 2 steps 37%; c) NaOMe, DMF, 62% from 6;
d) DAST, CH2Cl2, 63%; e) i. Zn, AcOH, RT; ii. tBuONO, TMSN3,
CH3CN, 08C, 2 steps 55%; f) i. NaOMe, MeOH; ii. acrolein diethylacetal,
camphorsulfonic acid, DMF,
2 steps 52%; g) Ac2O, pyridine, 98%;
h) borane–triethylamine complex, AlCl3, 4 ꢁ MS, THF, 83%; i) Ac2O,
pyridine, 95%; j) Hoveyda-Grubbs 2nd generation catalyst, CH2Cl2, 3 d,
19%; k) NaOMe, MeOH; l) 1 N NaOH, H2O, 6 h, 2 steps 73%. DAST=
N,N-diethylaminosulfur trifluoride, DMF=N,N-dimethylformamide,
TMS=trimethylsilyl.
10b, in moderate yield (52% yield, 1:1 mixture). Although
the mixture could not be separated by flash silica gel chro-
matography, normal-phase HPLC allowed the isolation of
Treatment of di-allyl compound 3 under ring-closing meACHTUNGTRENNUNGtath-
Table 1. 1H NMR chemical shifts of compounds 2, 3, and 7—14.
Compound[a]
7
8
8’
9
10a
10b
11a
11b
12
3
13
14[b]
2[c]
H-3 (ax)
H-3 (eq)
H-4
H-5
NH
H-6
H-7
H-8
H-9a
H-9b
1.97
2.96
5.11
3.84
5.33
3.96
5.24
4.99
4.17
4.01
1.83
2.76
5.13
3.88
5.33
4.04
5.30
5.18
4.28
4.09
2.08
3.06
5.51
4.33
5.56
4.63
5.53
5.30
4.87
4.18
1.76
2.63
5.11
3.87
5.33
3.98
5.31
5.30
4.35
4.17
1.69
2.79
3.89
3.81
5.82
3.34
3.43
4.28
4.13
4.04
1.67
2.78
3.88
3.81
5.81
3.30
3.53
4.31
4.19
3.96
1.75
2.64
5.20
3.94
5.26
3.82
5.12
4.35
3.99
3.93
1.75
2.65
5.18
3.96
5.30
3.80
5.16
4.37
4.11
3.84
1.83
2.71
5.04
4.05
5.33
3.93
5.09
4.07
3.51
3.42
1.74
2.62
5.13
3.82
5.37
4.01
5.36
5.25
3.72
3.51
2.20
2.92
4.92
4.22
5.44
4.00
5.23
5.37
3.98
3.52
1.75
2.79
3.71
3.86
–
3.57
4.03
3.78
3.96
3.91
1.73
2.81
3.64
3.81
–
3.60
4.23
3.69
3.91
3.66
[a] All of the compounds except 14 and 2 were dissolved in CDCl3. [b] Dissolved in CD3OD. [c] Dissolved in D2O.
Chem. Eur. J. 2013, 19, 1364 – 1372
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1367

H. Hinou, S.-I. Nishimura et al.
Figure 3. NMR spectroscopic experiments undertaken to establish the stereochemistries of compounds 11a, 11b, and 13: a) NOESY spectrum of com-
pound 11a (R-isomer); b) NOESY spectrum of compound 11b (S-isomer); c) overlay of the HSQC and HMBC spectra of compound 13. Green repre-
sents the cross peaks of the HMBC experiment, and blue and red represent the positive and negative cross peaks of the HSQC experiment, respectively.
A
macrocyclic structure of 13 and degradation of the azide
group under these reaction conditions.
The cyclic structure of compound 13 was confirmed by
NMR analysis. As shown in Figure 3c, HMBC cross-peaks
around the olefin bond indicate that compound 13 has the
eration catalyst[16b] (15 mol% catalyst, 0.005m, CH2Cl2, RT,
3 days) provided the desired 14-membered cycloalkene 13
in 19% yield. The low yield and slow rate of the RCM re-
ACHTUNGTRENNUNGaction might be due to the conformational constraint of the
1368
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1364 – 1372

Neuraminidase Inhibitors
FULL PAPER
desired 14-membered cycloalkene structure, without migra-
tion of the alkene bond. The E/Z-stereochemistry of the
olefin moiety was identified by NOESY or high-resolution
DQF-COSY (TD: 512ꢂ64k, see Figure S3 in the Supporting
Information) spectra of the de-O-acetylated derivative 14
because the two olefin protons of compound 13 overlapped.
The coupling constant between the two olefin protons of 14
was observed to be 10.9 Hz, which is consistent with the
standard coupling constants of cis alkenes.[25] This cis selec-
tivity of the RCM reaction for di-allyl compound 3 is sug-
gested to be due to the energy barrier constriction of the
ruthenium-bridged RCM intermediate formed. Finally, two-
step deprotection of 13 via 14 afforded the macrocyclic
mechanism-based inhibitor 2 in 73% yield. Although it is
reported that saponification of the carboxylate moiety of
sialic acid destabilizes the sialoside bond towards hydroly-
sis,[26] an NMR monitoring study[26b,c] of compound 2 under
near neutral conditions in deuterium oxide indicated that
the sialoside bond of 2 is stable for hours, and spontaneous
hydrolysis was observed only after a few days (see Figure S4
in the Supporting Information).
Assay of the macrocyclic mechanism-based inhibitor 2 for
neuraminidases: For this study, the inhibitory properties of
macrocyclic sialoside 2 were first evaluated against three
commercially available viral neuraminidases (VCNA,
STNA, and CPNA). All of the assays were performed at
least in duplicate, unless otherwise noted, and several con-
centrations of the cyclic mechanism-based inhibitor 2 (0.5–
7.5 mm) were preincubated with each NA in the presence of
Triton X-100 (0.01%).[11] From the preincubated enzyme
solutions, aliquots (20 mL) were transferred into the assay
medium (180 mL) in 10 min intervals (0–50 min), and the re-
sidual NA activities for 4-MU-NANA (final concentration
of 400 mm), as the substrate, were monitored by using a fluo-
rometric method (Figure 4).[27] Without preincubation
(0 min), 2 showed moderate inhibition activity, depending
upon the concentration of 2 in all of the three NAs (Fig-
ure 4a, c, and e, orange). This result suggests that 2 was rec-
ognized by all of the NAs and competed with the substrate,
4-MU-NANA, and that 750 mm (7.5 mm for preincubation)
of compound 2 can decrease the NA activity almost half for
VCNA and CPNA and two-thirds for STNA in the presence
Figure 4. Inhibitory patterns of compound 2 against three bacterial NAs in the presence of 0.01% Triton X-100: a), c), and e) residual NA activities after
preincubation with 2 for 0 or 50 min against VCNA, CPNA, and STNA, respectively; b), d), and f) relative NA activities after 0–50 min of preincubation
against VCNA, CPNA, and STNA, respectively.
Chem. Eur. J. 2013, 19, 1364 – 1372
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1369

H. Hinou, S.-I. Nishimura et al.
of 400 mm of 4-MU-NANA. The residual NA reactivities for
VCNA and CPNA decreased in a time-dependent manner
after preincubation (10–50 min), but the time-dependent in-
hibition rates were independent of the concentrations of 2
in the 0.5–7.5 mm range (Figure 4a–d). These results indicate
that the macrocyclic sialoside 2 was recognized and that the
glycoside bond was cleaved by these NAs as the substrate
formed a complex between the catalytic Tyr residue in the
active site of NA and anomeric position of the sialic acid
moiety of 2 (k3 in Figure 1c). During the hydrolytic dissocia-
tion between the Tyr residue and 2 (k5 ꢀk7), the aglycone
moiety was activated and acted as a mechanism-based inhib-
itor, as expected, even in the presence of 0.01% Triton X-
100, which could potentially disturb the hydrophobic inter-
action of the aglycone moiety with the active site. In con-
trast to VCNA and CPNA, STNA showed slight activation
rather than deactivation by 2 in a time-dependent manner
(Figure 4e and f). This result indicates that STNA must rec-
ognize 2 as substrate to compete with 4-MU-NANA, but
when hydrolyzed, 2 did not stay in the active site of STNA.
To rationalize this observation, it is possible to envisage a
plausible pathway in which the rate of a hydrolytic dissocia-
tion step of the NA-sialic acid complex (k5 ꢀk7) is much
higher than the rate of NA-aglycone bond formation. To
elucidate the mechanistic difference between STNA and the
other two bacterial NAs, the kinetic constant of each NA
was measured and compared (see Figure S5 in the Support-
ing Information). As shown in Table 2, STNA has an excep-
against all of the four NAs. Quash et al. reported that non-
cyclic difluoromethylphenyl sialosides exhibit only competi-
tive inhibition of FluA NA (A/HK/1/68), but both competi-
tive and mechanism-based irreversible inhibition modes
were observed for CPNA.[7b] The difference between the re-
sults obtained for FluA NA and our results may be due to
the macrocyclic structure staying attached to the aglycone
moiety in the active site after cleavage of the glycoside
bond. The same inhibition patterns recorded for 2 were ob-
served for all of the four NAs, which indicate that the mech-
anisms of reactivity were consistent and that all of the con-
formations for the NAs were retained.[30] Among all of the
seven NAs investigated in this study, FluB NA was the most
inhibited by 2, both competitively and irreversibly, and
FluA NA was most insensitive to treatment with 2, competi-
tively (Figure 4 and Figure 5). In contrast to the 2,3-dehydro
neuraminic acid derivatives, such as Tamiflu and Relenza,
which inhibit both NAs of FluA and FluB more selectively
than other neuraminidases,[2] the macrocyclic structure of 2
seems to recognize the structural differences between the
FluA and FluB NAs.
Conclusion
A novel macrocyclic mechanism-based inhibitor 2 for neur-
AHCTUNGTREGaNNUN minidases was prepared by using an RCM reaction. This
inhibitor was designed to stay and activate the difluorome-
thylphenyl moiety in the active site of NA after cleavage of
the sialoside bond of 2 based on the retention-type mecha-
nism of NAs. Compound 2 was recognized by all seven of
the NAs tested (viral, bacterial, and eukaryotic), found to
inhibit all of the NAs competitively, and irreversibly inhibit-
ed all of the six NAs that have normal kcat values for NAs.
Only STNA, which has an especially large kcat for an NA,
was not inhibited irreversibly upon treatment with 2. These
results suggest that 2 forms an enzyme–2 complex to stay in
the active site during activation and that the formed
enzyme–aglycone covalent bonds inhibit the NA activities ir-
Table 2. Kinetic constants of the microbial neuraminidases.
Enzymes
VCNA
CPNA
STNA
Km [mM][a]
50
233
36
37
135
55
378 (250)[b]
714
Vmax [RFU[c] minÀ1
]
kcat [sÀ1
]
1.8ꢂ104 (2700)[b]
tionally high kcat (1.8ꢂ104 sÀ1) compared with VCNA and
CPNA (36 and 55 sÀ1, respectively). This difference is con-
A
N
ACHTUNGTRENrNUNG eversibly, as expected. Overall, our macrocyclic inhibitor
much higher k5 ꢀk7 may contribute particularly to the ex-
ceptionally high kcat of STNA. In addition, the rate of k5 ꢀk7
might be the rate-limiting step for the kcat values of VCNA
and CPNA. It is interesting that VCNA and STNA belong
to same nanH gene family; both can be spread by a phage-
mediated horizontal gene transfer and the alignment of
their catalytic amino acid residues in their crystal structures
are highly conserved.[29] Vimr and Taylor et al. reported that
differences in binding to O-4 and to the glycerol side-chain
of sialic acid may reflect the different kinetics of STNA
compared to other NAs.[29a]
For the next part of the study, the inhibitory properties of
macrocyclic mechanism-based inhibitor 2 were further eval-
uated against a protozoa, a mammalian, and two viral NAs:
TcTS, hNeu2, FluA (A/PR/8/34, H1N1) NA, and FluB (B/
Hong Kong/5/72) NA, respectively. As shown in Figure 5, 2
showed both competitive and irreversible inhibition modes
can be a potential tool that can be used to elucidate the hy-
drolytic dissociation mechanism of each NA–sialic acid com-
plex by a simple photometric assay. Although the inhibition
efficiency of 2 for the tested NAs is low, these results afford
a novel strategy through which to inhibit neuraminidases
based on their mechanisms and to elucidate detailed mecha-
nism of NAs by capturing of amino acid residue involved in
the mechanism.[7–9,12,13] In the context of medical applica-
tions, universal blocking reagents for NAs are definitely un-
AHCTUNGTREGdNNUN esirable because mammalian NAs are highly relevant to
cell physiology.[1,31] A more potent and selective inhibitor
with a macrocyclic structure could also be achieved by com-
bining aglycone-focused library strategies[7f] and structural
information gained from noncompetitive inhibitors of
NAs.[32]
1370
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1364 – 1372

Neuraminidase Inhibitors
FULL PAPER
Figure 5. Inhibitory patterns of compound 2, in the presence of 0.01% Triton X-100, against a protozoa, a mammalian, and two viral NAs: TcTS, hNeu2,
FluA (A/PR/8/34, H1N1) NA, and FluB (B/Hong Kong/5/72) NA, respectively. a), c), e), and g) Residual NA activities after preincubation with 2 for 0
or 50 min against TcTS, hNeu2, FluA NA, and FluB NA, respectively. b), d), f), and h) Relative NA activities between 0 min of preincubation and 0–
50 min of preincubation against TcTS, hNeu2, FluA NA, and FluB NA, respectively.
thank Dr. S. Arioka, Dr. T. Noshi, and Dr. T. Yoshinaga at Discovery Re-
search Laboratories, Shionogi & Co., Ltd. for donating the TcTS, FluA,
and FluB NA utilized in this study.
Acknowledgements
This work was supported by funding from the Ministry of Education,
Culture, Science, Sports and Technology of Japan through the “Grant-in-
[1] a) A. Varki, Nature 2007, 446, 1023–1029; b) T. Angata, A. Varki,
Aid for Young Scientists (A) and for Scientific Research (B)” and the
Chem. Rev. 2002, 102, 439–469; c) E. Monti, E. Bonten, A. D’Azzo,
R. Bresciani, B. Venerando, G. Borsani, R. Schauer, G. Tettamanti,
Adv. Carbohydr. Chem. Biochem. 2010, 64, 403–479.
“Innovation COE Program for Future Drug Discovery and Medical
Care” programs. This work was also supported in part by funding from
Hokkaido University through the “Special Incentive for Young Scien-
[2] a) C. U. Kim, W. Lew, M. A. Williams, H. T. Liu, L. J. Zhang, S. Swa-
tists”. We thank Professor J. O. Previato and L. Mendonca-Previato for
their donation of the plasmid used for the expression of TcTS. We thank
R. Miyoshi and Dr. X.-D. Gao for donating the hNeu2 used. We also
minathan, 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;
b) M. von Itzstein, W. Y. Wu, G. B. Kok, M. S. Pegg, J. C. Dyason, B.
Chem. Eur. J. 2013, 19, 1364 – 1372
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1371

H. Hinou, S.-I. Nishimura et al.
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.
[15] a) I. Moustafa, H. Connaris, M. Taylor, V. Zaitsev, J. C. Wilson, M. J.
Kiefel, M. von Itzstein, G. Taylor, J. Biol. Chem. 2004, 279, 40819–
40826; b) L. M. G. Chavas, C. Tringali, P. Fusi, B. Venerando, G. Tet-
tamanti, R. Karo, E. Monti, S. Wakatsuki, J. Biol. Chem. 2005, 280,
469–475; c) S. L. Newstead, J. A. Potter, J. C. Wilson, G. Xu, C. H.
Chien, A. G. Watts, S. G. Withers, G. Taylor, J. Biol. Chem. 2008,
283, 9080–9088; d) A. Buschiazzo, P. M. Alzari, Curr. Opin. Chem.
Biol. 2008, 12, 565–572.
[16] a) R. G. Grubbs, Tetrahedron 2004, 60, 7117–7140; b) S. B. Garber,
J. S. Kingsbury, B. L. Gray, A. H. Hoveyda, J. Am. Chem. Soc. 2000,
122, 8168–8179.
[17] K. Furuhata, K. Anazawa, M. Itoh, Y. Shitori, H. Ogura, Chem.
Pharm. Bull. 1986, 34, 2725–2731.
[18] E. Berman, R. E. Fox, F. D. Thomson, J. Am. Chem. Soc. 1959, 81,
5605–5608.
[19] A. Kuboki, T. Sekiguchi, T. Sugai, H. Ohta, Synlett 1998, 479–482.
[20] a) K. Barral, A. D. Moorhouse, J. E. Moses, Org. Lett. 2007, 9,
1809–1811; b) J. Das, S. N. Patili, R. Awasthi, C. P. Narasimhulu, S.
Trehan, Synthesis 2005, 1801–1806.
[21] N. M. Spijker, C. A. Keuning, M. Hooglugt, G. H. Veeneman,
C. A. A. van Boeckel, Tetrahedron 1996, 52, 5945–5960.
[22] a) P. J. Garegg, H. Hultberg, Carbohydr. Res. 1981, 93, C10–C11;
b) P. J. Garegg, H. Hultberg, S. Oscarson, J. Chem. Soc. Perkin
Trans. 1 1982, 2395–2397.
[23] M. Sakagami, H. Hamana, Tetrahedron Lett. 2000, 41, 5547–5551.
[24] M. Ek, P. J. Garegg, H. Hultberg, S. Oscarson, J. Carbohydr. Chem.
1983, 2, 305–311.
[25] H. Gꢃnther, G. Jikeli, Chem. Rev. 1977, 77, 599–637.
[26] a) H. Hinou, X.-L. Sun, Y. Ito, Tetrahedron Lett. 2002, 43, 9147–
9150; b) H. Hinou, X.-L. Sun, Y. Ito, J. Org. Chem. 2003, 68, 5602–
5613; c) Y. Kajihara, S. Nishigaki, D. Hanzawa, G. Nakanishi, R.
Okamoto, N. Yamamoto, Chem. Eur. J. 2011, 17, 7645–7655.
[27] R. W. Myers, R. T. Lee, Y. C. Lee, G. H. Thomas, L. W. Reynolds, Y.
Uchida, Anal. Biochem. 1980, 101, 166–174.
[28] L. L. Hoyer, P. Roggentin, R. Schauer, E. R. Vimer, J. Biochem.
1991, 110, 462–467.
[29] a) S. J. Crennell, E. F. Garman, C. Philippon, A. Vasella, W. G.
Laver, E. R. Vimr, G. L. Taylor, J. Mol. Biol. 1996, 259, 264–280;
b) S. Crennell, E. Garman, G. Laver, E. Vimr, G. Taylor, Structure
1994, 2, 535–544.
[3] R. R. Rando, Science 1974, 185, 320–324.
[4] S. Halazy, V. Berges, A. Ehrhard, C. Danzin, Bioorg. Chem. 1990,
18, 330–344.
[5] a) K. D. Janda, L.-C. Lo, C.-H. L. Lo, M.-M. Sim, R. Wang, C.-H.
Wong, Science 1997, 275, 945–948; b) M. Kurogochi, S.-I. Nishi-
mura, Y. C. Lee, J. Biol. Chem. 2004, 279, 44704–44712.
[6] M. Ichikawa, Y. Ichikawa, Bioorg. Med. Chem. Lett. 2001, 11, 1769–
1773.
[7] a) P.-A. Driquez, B. Barrere, B. Chantegral, C. Deshayes, A. Dou-
theau, G. Quash, Bioorg. Med. Chem. Lett. 1992, 2, 1361–1366;
b) B. Barrere, P.-A. Driguez, J. Maudrin, A. Doutheau, M. Aymard,
G. Quash, Arch. Virol. 1997, 142, 1365–1380; c) H. Hinou, M. Kuro-
gochi, H. Shimizu, S.-I. Nishimura, Biochemistry 2005, 44, 11669–
11675; d) C.-P. Lu, C.-T. Ren, Y.-N. Lai, S.-H. Wu, W.-M. Wang, J.-
Y. Chen, L.-C. Lo, Angew. Chem. 2005, 117, 7048–7052; Angew.
Chem. Int. Ed. 2005, 44, 6888–6892; e) S. T. Carvalho, M. Sola-
Penna, I. A. Oliveira, S. Pita, A. S. GonÅalves, B. C. Neves, F. R.
Sousa, L. Freire-de-Lima, M. Kurogochi, H. Hinou, S.-I. Nishimura,
L. MendonÅa-Previato, J. O. Previato, A. R. Todeschini, Glycobiolo-
gy 2010, 20, 1034–1045; f) H. Kai, H. Hinou, S.-I. Nishimura,
Bioorg. Med. Chem. 2012, 20, 2739–2746.
[8] H. Hinou, R. Miyoshi, Y. Takasu, H. Kai, M. Kurogochi, S. Arioka,
X.-D. Gao, N. Miura, N. Fujitani, S. Omoto, T. Yoshinaga, T. Fuji-
wara, T. Noshi, H. Togame, H. Takemoto, S.-I. Nishimura, Chem.
Asian J. 2011, 6, 1048–1056.
[9] a) V. Ahmed, Y. Liu, S. D. Taylor, ChemBioChem 2009, 10, 1457–
1461; b) B. P. Rempel, S. G. Withers, Glycobiology 2008, 18, 570–
586.
[10] D. H. Kwan, H.-M. Chen, K. Ratananikom, S. M. Hancock, Y. Wata-
nabe, P. T. Kongsaeree, A. L. Samuels, S. G. Withers, Angew. Chem.
2011, 123, 314–317; Angew. Chem. Int. Ed. 2011, 50, 300–303.
[11] a) B. Y. Feng, A. Shelat, T. N. Doman, R. K. Guy, B. K. Shoichet,
Nat. Chem. Biol. 2005, 1, 146–148; b) B. Y. Feng, B. K. Shoichet,
Nat. Protoc. 2006, 1, 550–553.
[12] a) W. I. Weis, K. Drickamer, Annu. Rev. Biochem. 1996, 65, 441–
473; b) C.-H. Wong, Acc. Chem. Res. 1999, 32, 376–385; c) E. C.
Stanca-Kaposta, D. P. Gamblin, J. Screen, B. Liu, L. C. Snoek, B. G.
Davis, J. P. Simons, Phys. Chem. Chem. Phys. 2007, 9, 4444–4451;
d) H.-J. Gabius, S. Andre, J. Jimenez-Barbero, A. Romero, D. Solis,
Trends Biochem. Sci. 2011, 36, 298–313; e) H.-J. Gabius, Biochem.
Soc. Trans. 2011, 39, 399–405.
[30] a) A. R. Todeschini, L. MendonÅa-Previato, J. O. Previato, A. Varki,
H. van Halbeek, Glycobiology 2000, 10, 213–221; b) S. Albouz-Abo,
R. Turton, J. C. Wilson, M. von Itzstein, FEBS Lett. 2005, 579, 1034–
1038; c) A. K. J. Chong, M. S. Pegg, N. R. Taylor, M. von Itzstein,
FEBS Lett. 1992, 207, 335–343.
[13] a) A. G. Watts, I. Damager, M. L. Amaya, A. Buschiazzo, P. Alzari,
A. C. Frasch, S. G. Withers, J. Am. Chem. Soc. 2003, 125, 7532–
7533; b) G. Pierdominici-Sottile, N. A. Horenstein, A. E. Roitberg,
Biochemistry 2011, 50, 10150–10158.
[14] a) J. Chan, A. R. Lewis, M. Gilbert, M.-F. Karwaski, A. J. Bennet,
Nat. Chem. Biol. 2010, 6, 405–407; b) J. Chan, A. R. Lewis, D. In-
durugalla, M. Schur, W. Wakarchuk, A. J. Bennet, J. Am. Chem.
Soc. 2012, 134, 3748–3757; c) J. Chan, J. N. Watson, A. Lu, V. C.
Cerda, T. J. Borgford, A. J. Bennet, Biochemistry 2012, 51, 433–441.
[31] A. Mozzi, P. Mazzacuva, G. Zampella, M. E. Forcella, P. A. Fusi, E.
Monti, Proteins Struct. Funct. Bioinf. 2012, 80, 1123–1132.
[32] 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: March 14, 2012
Revised: October 24, 2012
Published online: December 11, 2012
1372
www.chemeurj.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 1364 – 1372