Identification of a novel glycan processing enzyme with exo-acting β-allosidase activity in the Golgi apparatus using a new platform for the synthesis of fluorescent substrates

Article abstract of DOI:10.1016/j.bmc.2014.11.023

The majority of eukaryotic proteins undergo post-translational modifications (PTMs) involving the attachment of complex glycans, predominantly through N-glycosylation and O-glycosylation. PTMs play important roles in virtually all cellular processes, and aberrant regulation of protein glycosylation and glycan processing has been implicated in various diseases. However, glycan processing on proteins in various cellular contexts has not been visualized. We had previously developed a quinone methide cleavage (QMC) platform for enhanced substrate design. This platform was applied here to screen for novel glycan-processing enzymes. We designed and synthesized fluorescent substrates with β-allopyranoside residues using the QMC platform. When applied in cell-based assays, the fluorescent substrates allowed rapid and clear visualization of β-allosidase activity in the Golgi apparatus of human cultured cells. The QMC platform will likely find broad applications in visualizing the activities of glycan processing enzymes in living cells and in studying PTMs.

Full text of DOI:10.1016/j.bmc.2014.11.023

Bioorganic & Medicinal Chemistry 23 (2015) 73–79
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Identification of a novel glycan processing enzyme with exo-acting
b-allosidase activity in the Golgi apparatus using a new platform
for the synthesis of fluorescent substrates
⇑
Wataru Hakamata , Kazuki Miura, Takako Hirano, Toshiyuki Nishio
Department of Chemistry and Life Science, College of Bioresource Sciences, Nihon University, 1866 Kameino, Fujisawa-shi, Kanagawa 252-0880, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
The majority of eukaryotic proteins undergo post-translational modifications (PTMs) involving the
attachment of complex glycans, predominantly through N-glycosylation and O-glycosylation. PTMs play
important roles in virtually all cellular processes, and aberrant regulation of protein glycosylation and
glycan processing has been implicated in various diseases. However, glycan processing on proteins in
various cellular contexts has not been visualized. We had previously developed a quinone methide cleav-
age (QMC) platform for enhanced substrate design. This platform was applied here to screen for novel
glycan-processing enzymes. We designed and synthesized fluorescent substrates with b-allopyranoside
residues using the QMC platform. When applied in cell-based assays, the fluorescent substrates allowed
rapid and clear visualization of b-allosidase activity in the Golgi apparatus of human cultured cells. The
QMC platform will likely find broad applications in visualizing the activities of glycan processing
enzymes in living cells and in studying PTMs.
Received 22 September 2014
Revised 13 November 2014
Accepted 14 November 2014
Available online 22 November 2014
Keywords:
b-Allosidase
Golgi apparatus
Fluorescent substrate
Protein glycosylation
Glycan processing
Post-translational modification
Quinone methide cleavage
Ó 2014 Elsevier Ltd. All rights reserved.
1
. Introduction
co-translationally appended with asparagine-linked oligosaccha-
rides (N-glycans). N-glycans and N-glycan processing enzymes
Although more than 300 post-translational modifications
PTMs) have been reported, many of these have only been
are thought to play important roles in protein quality control
during protein synthesis and folding, functions that are essential
for cellular health. In contrast, the hydroxyl groups of serine and
(
1
5
described in eukaryotes. PTMs comprise an array of reversible
modifications that promote signaling diversity and the rapid
threonine in various proteins are labeled with O-linked oligosac-
charides (O-glycans). O-Glycans and O-glycan processing enzymes
are thought to play important roles not only in protection of
mucosa but also in abnormal behavior at the cellular level and
the avoidance of immunological defenses by cancer cells. These
2
conveyance of cellular messages. Some irreversible PTMs, includ-
ing glycosylation, lipidation, and disulfide bond formation, are
stable and are important for the maturation and proper folding
3
of nascent proteins. Protein glycosylation involves the attachment
6
of monosaccharides and glycans to proteins and provides greater
proteomic diversity than do other PTMs.4 Protein glycosylation is
mediated by various glycosyltransferases such as oligosaccharyl-
transferase (OST, EC 2.4.1.119) and is processed by an exo-acting
glycan-processing enzyme such as endoplasmic reticulum (ER)
glucosidase I and II, resulting in the attachment of various N- and
O-linked glycans. The majority of secretory proteins are
functions are essential for the proliferation of cancer cells. Despite
their importance in maintaining cellular health and promoting
cancerous growths, the development of technology for visualizing
most glycan-processing enzyme activities remains an important
challenge. Several fluorescent substrates are commercially
available for visualizing and monitoring
a-mannosidase and b-
galactosidase as exo glycan intracellular hydrolase activity in cul-
tured cells (Fig. 1). In the future, these substrates may facilitate
functional imaging analysis of protein glycosylation. Previous
molecular designs of glycan-processing enzyme substrates have
not targeted active sites using 3-dimensional enzyme structures.
In general, the active sites of exo-acting glycan-processing enzymes
are composed of subsite À1, which recognizes a sugar moiety at
the non-reducing end of the substrate, and subsite +1, which
Abbreviations: 2MeTG, 2-methyl TokyoGreen; BFA, brefeldin A; EM, emission
wavelength; ER, endoplasmic reticulum; EX, excitation wavelength; Man, mannose;
MS, mass spectrometry; NMR, nuclear magnetic resonance; PBS, phosphate-
buffered saline; PTM, post-translational modifications; QMC, quinone methide
cleavage; TBP, tributyl phosphine; TFMU, 4-Trifluoromethylumbelliferone.
⇑
Corresponding author.
E-mail address: hakamata.wataru@nihon-u.ac.jp (W. Hakamata).
http://dx.doi.org/10.1016/j.bmc.2014.11.023
968-0896/Ó 2014 Elsevier Ltd. All rights reserved.
0

7
4
W. Hakamata et al. / Bioorg. Med. Chem. 23 (2015) 73–79
Figure 1. Structure of commercially available fluorescent substrates for
a
-mannosidase and b-galactosidase in cultured cells.
recognizes the penultimate sugar moiety at the non-reducing end.
our QMC platform in discovering novel glycan-processing
enzymes, we sought to identify exo-acting b-allosidase activity as
an undiscovered activity in human cells. Allose, which is an epimer
of glucose at the C-3 hydroxyl group, is a rare sugar. Yanagita et al.
7–9
Hydrolysis occurs between subsites À1 and +1.
Using small
synthetic molecules to design the substrate, we previously
reported the molecular recognition of an active site in ER glucosi-
dases functioning as N-glycan processing enzymes.1
0,11
To develop
reported the anti-proliferative activity of 6-O-acyl-D-allose against
1
4
practical substrates that are operational in cells, it is essential to
base the precise molecular design of substrates on known struc-
tural information for glycan-processing enzymes.
the human leukemia MOLT-4F cell line. If b-allosidase activity is
observed in human cells, we can postulate the existence of b-allos-
idase as an important player of an undiscovered physiological
function. Moreover, b-allosidases potentially regulate novel path-
ways of N- and O-glycan processing and may serve as novel targets
for drug discovery. We report herein the design and synthesis of a
QMC platform based on novel fluorescent substrates for measuring
b-allosidase activity in human cultured cells. Using these sub-
strates, we also localized the b-allosidase activity to the Golgi
apparatus in human cultured cells with bright fluorescence.
Commercially available fluorescent substrates targeting the
modeled enzyme subsites bind subsite +1 poorly, which may be
due to steric hindrance and a flat, non-sugar-like structure on the
part of the fluorophore. Reducing the size of fluorophores is
impractical because small-sized fluorescent moieties have low
quantum yields, decreasing their sensitivity. Therefore, a different
substrate-design platform is required to design fluorescent sub-
strates for exo glycan-processing enzymes. Our group reported
the development of a substrate-design platform using quinone
methide cleavage (QMC) reactions to study the activity of an ER-
localized endogenous carboxyl esterase as a pro-drug activation
enzyme. The multicolor fluorescence probes detected carboxyl
esterases in the ER via two-step reactions: ester hydrolysis and
spontaneous QMC reaction at the cellular level. These probes pen-
etrate into the ERs, after hydrolysis, the release and accumulation
of water insoluble fluorophores from these probes specifically
2
2
. Material and methods
.1. Substrate synthesis
Structures of all designed substrates 1–6 are shown in Figure 5.
The fluorescent substrates 1–3 and 6 were synthesized to evaluate
b-allosidase activities in living cells. Details of the synthetic
scheme of 1–3 and 6 are outlined in Figure S1. The starting
1
2,13
stained ERs with bright fluorescence in 3 human cell lines.
To
material (4-formylphenyl b-
D
-allopyranoside) of 1–3 and 6 was
acetylated, and an aldehyde of obtained acetyl derivative was
reduced to a benzyl alcohol derivative using NaBH . In the next
were used to react the
benzyl alcohol derivative with the fluorophores resorufin, TFMU,
1
5
develop probes with improved binding to subsite +1 of the enzyme
active site, we applied the QMC platform for substrate design of an
exo glycan-processing enzyme involved in N- and O-glycan pro-
cessing (Fig. 2). The sugar and benzyl moieties of the substrates
occupy subsite À1 and subsite +1, respectively. The substrate fluo-
rophore is located outside of the enzyme active site. A schematic
representation of our designed substrates and commercially avail-
able substrates bound to enzyme subsites is shown in Figure 3A
and 3B, respectively.
4
1
2,13
step, Mitsunobu reaction conditions
1
6,17
or 2MeTG,
yielding substrates 1–3, respectively. Substrate 4
was synthesized by deacetylation of compound 3. The materials
used for substrate synthesis, the synthesis details, and the
instrumental analysis of substrates 1–3 and 6 are described in
the Supplementary data.
The present study describes a new imaging technology for
identifying glycan-processing enzymes using fluorescently labeled
substrates, based on the QMC platform. These substrates can also
be used to conduct mechanistic experiments in living cells by con-
focal microscopy. Thus, our platform may also reveal novel glycan-
processing enzymes in living cells. To demonstrate the potential of
2
.2. Cell-based assays
Cell-base assays for measuring b-allosidase activity were
performed using substrates 1–3, substrate 6, and Golgi-ID Green
with HT-1080, HeLa, and SK-N-SH cells. Co-staining assays with
Figure 2. Rational design of substrates for an exo glycan-processing enzyme, based on the QMC platform.

W. Hakamata et al. / Bioorg. Med. Chem. 23 (2015) 73–79
75
Figure 3. Schematic diagrams of subsite structures of an exo glycan-processing enzyme. (A) QMC-based designed substrate bound in the active site. (B) Poor fitting of a
commercially available substrate in subsite +1.
substrate 1 and Golgi-ID Green were also performed with each of
the 3 cell lines. We further determined the identity of the stained
organelle in HT-1080 cells using by substrate 6 and BFA. All fluo-
rescence signals in cells were recorded using a confocal laser-scan-
ning microscope. The materials and methods used for performing
cell-based assays are described in the Supporting information.
surface, as shown in Figure 4A. The a-mannopyranoside and ben-
zyl moieties of the QMC-based substrate occupied subsites À1
and +1, respectively. The 2MeTG fluorophore of the QMC-based
substrate was located outside of the enzyme active site, as shown
in Figure 4B. The commercially available substrate was not pre-
dicted to occupy subsite À1. The
the commercially available substrate occupied subsite +1 and the
MeTG fluorophore of the substrate was located outside of enzyme
a-mannopyranoside moiety of
2
3
3
. Results and discussion
active site, as shown in Figure 4C. Given these results, QMC plat-
form-based substrates for b-allosidase are predicted to have
greater compatibility with subsite +1 of the enzyme than are com-
mercially available substrates and some reported substrates.1
We designed 6 fluorescent b-allosidase substrates based on the
.1. Design and synthesis of fluorescent b-allosidase substrates
based on the QMC platform
9,20
To confirm the efficacy of adapting our QMC platform-based
substrates for targeting enzyme active sites, we performed a
molecular docking study to analyze the interactions of 3 ligands
QMC platform and docking study of human class I a-1,2-mannos-
idase as it provides suitable models for enzyme–ligand complex, as
shown in Figure 5. Cleavage of these substrates by b-allosidase in
cultured cells can occur through 2 reactions, namely b-allopyrano-
side linkage hydrolysis and QMC reactions. Hydrolysis releases the
fluorophore from the substrate, staining the organelle that the b-
allosidase enzyme resides in. b-Allosidase substrates 1–6 showed
the following EX and EM maxima: resorufin (EX, 571 nm; EM,
with human class I
ing, N-linked glycan-processing enzyme. For this computational
study, we analyzed 3 substrates: a natural substrate (Man -1,2
Man -1,2 Man), a QMC platform-based mannosidase substrate
a-1,2-mannosidase (PDB ID: 1X9D), an exo-act-
18
a
a
with 2MeTG, and a commercially available mannosidase substrate.
Our docking models demonstrated potential interactions between
each substrate and the enzyme subsites. The Man a-1,2 Man moi-
ety of the non-reducing end of the natural substrate was predicted
to occupy subsites À1 and +1 of the enzyme active site. The Man
residue of reducing end of the substrate was exposed to the solvent
5
4
85 nm), TFMU (EX, 385 nm; EM, 502 nm), and 2MeTG (EX,
1
2,13
91 nm; EM, 510 nm),
producing red, blue, and green fluores-
cence, respectively. These fluorophores also show distinct EX and
EM, excellent fluorescence quantum yields, and low pH-dependent
Figure 4. Molecular modeling of the lowest-energy complex of human Golgi mannosidase bound with a mannopyranoside derivative. (A) Magnified view of the enzyme-
natural substrate complex. (B) Magnified view of the enzyme-QMC platform-based substrate complex. (C) Magnified view of the enzyme-commercially available substrate
complex.

7
6
W. Hakamata et al. / Bioorg. Med. Chem. 23 (2015) 73–79
Figure 5. Structure of QMC platform-based substrates for b-allosidase.
Figure 6. Fluorescence images of three different cell lines showing the location of substrate hydrolysis. Substrates 1–3 and 6, each tagged with a specific color fluorophore,
were introduced to three cell types, HT1080, HeLa, and Sk-N-SH. Substrate 1 was tagged with resorufin (red), while substrate 2 was tagged with TFMU (blue). Substrates 3 and
6
were both tagged with 2MeTG (green).
changes in fluorescence, which make them suitable for multicolor
imaging of enzyme activity in living cells. Thus, these fluorophores
and related derivatives are used widely in biological research.
Moreover, these designed substrates exhibit membrane permeabil-
ity, making them compatible for use as fluorescent bioimaging
tools. Results from a previous study with Caco-2 cells showed that
a logD value of >4.5 is required for compounds with molecular
2
1
weights >500 to have high membrane permeability. The clogD
2
2
values of substrates 1–6 were calculated, and the resulting values
were evaluated based on above index. Substrates 4 and 5 lack

W. Hakamata et al. / Bioorg. Med. Chem. 23 (2015) 73–79
77
Figure 7. Fluorescence images of substrate 1 (resorufin; red) and Golgi-ID Green (green) expression, as well as the merged fluorescence images, indicating the co-localization
of the substrate and the Golgi apparatus-specific dye in three cell lines.
acetyl groups and showed low clogD values (data not shown). In
contrast, substrates 1–3 and 6, which have molecular weights of
3.2. Fluorescence of QMC platform-based substrates in cultured
human cells
6
4
50, 612, 738, and 571 had cLogD values of 4.20, 3.36, 7.43, and
.73, respectively, at physiological pH in cells, indicative of ade-
To screen for b-allosidase activity in cultured cells, we
incubated a human fibrosarcoma cell line (HT-1080), a human cer-
vical cancer cell line (HeLa), and a human neuroblastoma cell line
(SK-N-SH) with substrates 1–3 and 6, at final concentrations of
quate membrane permeability. An acetyl modification of substrate
hydroxyl groups to improve membrane permeability may raise
concerns regarding deacetylation in cells living. However, numer-
ous commercially available bioimaging probes have acetyl groups
10 lM. The resulting fluorescence signals were imaged using a
2
3
that are hydrolyzed and activated by cytosolic esterases.
confocal laser-scanning microscope. Fluorescence signaling from
substrates 1–3 (with acetyl groups) and substrate 6 (without an
acetyl group) revealed an intracellular accumulation of each
fluorophore in the nuclear peripheries (Fig. 6), suggesting that they
were cleaved both by b-allosidase and esterase enzymes. The
fluorescence intensities of substrates 1–3 and 6 generated by b-
allosidase activity in each cell line was independent of fluorophore
structures and acetyl modification. The substrates bound to subsite
À1 and subsite +1 of b-allosidase, and the enzyme has no recogni-
tion site for the fluorophore based on molecular modeling data
with a mannosidase. Moreover, substrates 1–3 and 6 stained the
Considering these results, substrates 1–3 and 6 were synthe-
sized in preparation for testing in cell-based assays. To be useful
as b-allosidase substrates and multicolor bioimaging tool in cells,
the fluorescence from substrates 1–3 and 6 should show minimal
emission before hydrolyzation, as was observed with substrates
1–3 and 6 in PBS. Therefore, these substrates will be in a quenched
1
state in cells before hydrolyzation. For synthesis details, H NMR,
1
3
C NMR, MS, elemental analysis (C, H, N, and F), and fluorescence
spectra of substrates 1–3 and 6, refer to the METHODS and
Supporting information sections.

7
8
W. Hakamata et al. / Bioorg. Med. Chem. 23 (2015) 73–79
Figure 8. Fluorescence images of HT-1080 cells stained with Golgi-ID Green and Substrate 6 (both green) following dispersion of the Golgi apparatus with BFA along with
differential interference contrast microscopy (DIC) images highlighting the altered morphology of the cells.
same cellular compartment. A dynamic equilibrium may occur as
deacetylated substrates are transferred into the lumens of organ-
elles and are then trapped by hydrolysis at the b-allopyranoside
linkage. The pattern of bright fluorescence observed around the
nuclear peripheries suggested that the fluorescent substrates were
localized to the Golgi apparatus, as a reticulated staining pattern
characteristic of the ER was not observed.
6 can be used to investigate the role of b-allosidase in glycan
processing and to determine aberrant activities in diseases. The
preliminary conclusions was reached that the Golgi apparatus of
human cells contains active b-allosidase. Further consideration will
be needed to yield any findings about an isolation of b-allosidase
and a determination of enzymatic properties. Substrates 1–3 and
6 also provide a starting point for developing fluorescent sub-
strates to explore novel glycan-processing enzymes mediating
PTMs in living cells. The results of this study provide a good exam-
ple of the utility of the QMC platform for designing substrates of
glycan-processing enzymes; thus, this platform should be useful
for analyzing both static and dynamic events in biological applica-
tions. Using this strategy, we are currently building a library of
fluorescent substrates for exo glycan-processing hydrolases
involved in PTMs, such as glucosidase in the ER, mannosidase in
the ER and Golgi apparatus, and fucosidase in the lysosome, which
will greatly impact the bioimaging field of glycobiology. Glycosyl-
ation studies deeply affected our understanding of PTMs in the last
century. Similarly, the bioimaging of processing enzyme activities
in living cells should advance our understanding of protein
glycosylation and provide breakthroughs needed to monitor gly-
can-processing pathways in live cells.
3
.3. b-Allosidase activity in the Golgi apparatus
The subcellular localization of the observed b-allosidase activity
was investigated using a commercially available organelle-specific
dye. A well-known fluorescent reagent for the Golgi apparatus,
Golgi-ID Green (EX, 473 nm; EM, 543 nm), was used to co-stain
HT-1080, HeLa, and SK-N-SH cells with substrate 1. Substrate 1
contains resorufin as the fluorophore, which provides a fluores-
cence window that is distinct from Golgi-ID Green. The capacity
for generating alternative fluorescence wavelengths by the QMC
platform allows flexibility in selecting the most suitable substrates
for biological observations. Fluorescence signaling in cells co-
stained with substrate 1 and Golgi-ID Green was evident in all 3
cell lines. Moreover, substrate 1 and Golgi-ID Green fluorescence
colocalized to the Golgi apparatus, as shown in Figure 7. It is evi-
dent from these observations that substrates 1–3 and 6 were spe-
cifically hydrolyzed in the Golgi apparatus. In parallel experiments,
we analyzed the dispersion of the stained Golgi apparatus caused
by BFA. BFA inhibits the transport of proteins from the ER to the
Golgi apparatus; consequently, the Golgi apparatus is dispersed
Acknowledgments
The authors wish to express their sincere thanks to Mrs. T. Kiu-
chi of the General Research Institute of the College of Bioresource
Sciences of Nihon University for performing FABMS measurements
and Mrs. A. Sato of A-Rabbit-Science Japan Co., Ltd for performing
the elemental analyses. This work was partly supported by JSPS
KAKENHI Grant Number 26460157.
2
4
in living cells. Treatment of HT-1080 cells with BFA and either
Golgi-ID Green or substrate 6 revealed similar morphological and
staining patterns in the Golgi apparatus, as shown in Figure 8. In
these cell-based assays, traces of released fluorophore leaking out
of the Golgi apparatus were observed. After carefully considering
the concern, it is evident from our data that the Golgi apparatus
of human cells possesses b-allosidase.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2014.11.023.
4
. Conclusion
References and notes
We successfully developed a QMC platform for synthesizing
fluorescent substrates of exo glycan-processing enzymes, based
on the known subsite structures. Fluorescent substrates 1–3 and
1. Witze, E. S.; Old, W. M.; Resing, K. A.; Ahn, N. G. Nat. Methods 2007, 4, 798.
2
3
.
.
Deribe, Y. L.; Pawson, T.; Dikic, I. Nat. Struct. Mol. Biol. 2010, 17, 666.
Siman, P.; Brik, A. Org. Biomol. Chem. 2012, 10, 5684.

W. Hakamata et al. / Bioorg. Med. Chem. 23 (2015) 73–79
79
4
.
Pan, S.; Chen, R.; Aebersold, R.; Brentnall, T. A. Mol. Cell Proteomics 2011, 10.
R110.003251.
Needham, P. G.; Brodsky, J. L. Biochim. Biophys. Acta 2013, 1833, 2447.
Brooks, S. A.; Carter, T. M.; Royle, L.; Harvey, D. J.; Fry, S. A.; Kinch, C.; Dwek, R.
A.; Rudd, P. M. Anticancer Agents Med. Chem. 2008, 8, 2.
15. Ye, D.; Zhang, K.; Chen, H.-F.; Yin, S.-F.; Li, Y. Struct. Rep. Online 2009, 65, o1338.
16. Urano, Y.; Kamiya, M.; Kanda, K.; Ueno, T.; Hirose, K.; Nagano, T. J. Am. Chem.
Soc. 2005, 127, 4888.
17. Mottram, L. F.; Boonyarattanakalin, S.; Kovel, R. E.; Peterson, B. R. Org. Lett.
2006, 8, 581.
5
6
.
.
7
8
9
.
.
.
Kurakata, Y.; Uechi, A.; Yoshida, H.; Kamitori, S.; Sakano, Y.; Nishikawa, A.;
Tonozuka, T. J. Mol. Biol. 2008, 381, 116.
Karaveg, K.; Siriwardena, A.; Tempel, W.; Liu, Z. J.; Glushka, J.; Wang, B. C.;
Moremen, K. W. J. Biol. Chem. 2005, 280, 16197.
18. CLC bio, a QIAGEN Company Inc, Molegro Virtual Docker Ver. 6.0.
19. Kamiya, M.; Asanuma, D.; Kuranaga, E.; Takeishi, A.; Sakabe, M.; Miura, M.;
Nagano, T.; Urano, Y. J. Am. Chem. Soc. 2011, 133, 12960.
20. Ichikawa, Y.; Kamiya, M.; Obata, F.; Miura, M.; Terai, T.; Komatsu, T.; Ueno, T.;
Hanaoka, K.; Nagano, T.; Urano, Y. Angew. Chem., Int. Ed. 2014, 53, 6772.
21. Waring, M. J. Expert Opin. Drug Discovery 2010, 5, 235.
22. Advanced Chemistry Development Inc., Structure Design Suite Ver. 12.
23. Products list: Carboxy-H2DFFDA, C-13293, H2DCFDA, D-399, Oregon green
carboxylic acid diacetate, O6151 (Molecular Probes); FluoForte Calcium Assay
Kit, ENZ-51017, eFluxx-ID Green Multidrug Resistance Assay Kit, ENZ-51029
(ENZO Life Sciences); HaloTag diAcFAM Ligand, G8273, (Promega).
24. Fujiwara, T.; Oda, K.; Yokota, S.; Takatsuki, A.; Ikehara, Y. J. Biol. Chem. 1988,
263, 18545.
Shah, N.; Kuntz, D. A.; Rose, D. R. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 9570.
1
1
1
1
1
0. Hakamata, W.; Muroi, M.; Nishio, T.; Oku, T.; Takatsuki, A. J. Carbohydr. Chem.
004, 23, 27.
1. Hakamata, W.; Kurihara, M.; Okuda, H.; Nishio, T.; Oku, T. Curr. Top. Med. Chem.
009, 9, 3.
2. Hakamata, W.; Machida, A.; Oku, T.; Nishio, T. Bioorg. Med. Chem. Lett. 2011, 21,
206.
3. Hakamata, W.; Tamura, S.; Hirano, T.; Nishio, T. ACS Med. Chem. Lett. 2014, 5,
21.
2
2
3
3
4. Yanagita, R. C.; Kobashi, K.; Ogawa, C.; Ashida, Y.; Yamaashi, H.; Kawanami, Y.
Biosci. Biotech. Biochem. 2014, 78, 190.