Preparation and functional analysis of gossypols having two carbohydrate appendages with enaminooxy linkages

Article abstract of DOI:10.1016/j.carres.2018.02.001

We developed new gossypol (Gos)-based glycoconjugates through dehydration condensation of native Gos and chemically modified glycosides having aminooxy groups. The resultant glycoconjugates (glycoGos) were resistant to hydrolysis, although they were light-sensitive and slowly decomposed even under indoor lighting. The glycoGos also exhibited improved water solubility compared with native Gos, but their saturated concentrations in water were still low (6.4–17 μM), due to their hydrophobic naphthyl rings. We also carried out WST-8 assays to assess the anticancer activity of the glycoGos on DLD-1 and HepG2 cells and found that the glycoGos having β-lactosides and having β-galactosides (specific ligands for asialoglycoprotein receptors) showed enhanced anticancer activity on HepG2 cells.

Full text of DOI:10.1016/j.carres.2018.02.001

Carbohydrate Research 458-459 (2018) 67e76
Contents lists available at ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Preparation and functional analysis of gossypols having two
carbohydrate appendages with enaminooxy linkages
a
a
b
a
Yoshitsugu Amano , Masaki Nakamura , Shinya Shiraishi , Naoto Chigira ,
a
b
c
b, d, *
Nobuya Shiozawa , Masahito Hagio , Tomohiro Yano , Teruaki Hasegawa
Graduate School of Life Sciences, Toyo University, 1-1-1 Izumino, Itakura-machi, Ora-gun, Gumma 374-0193, Japan
Faculty of Life Sciences, Toyo University, 1-1-1 Izumino, Itakura-machi, Ora-gun, Gumma 374-0193, Japan
Faculty of Food and Nutritional Sciences, 1-1-1 Izumino, Itakura-machi, Ora-gun, Gumma 374-0193, Japan
a
b
c
d
Bio-Nano Electronics Research Centre, Toyo University, 2100 Kujirai, Kawagoe, Saitama 350-8585, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 7 December 2017
Received in revised form
We developed new gossypol (Gos)-based glycoconjugates through dehydration condensation of native
Gos and chemically modified glycosides having aminooxy groups. The resultant glycoconjugates (gly-
coGos) were resistant to hydrolysis, although they were light-sensitive and slowly decomposed even
under indoor lighting. The glycoGos also exhibited improved water solubility compared with native Gos,
29 January 2018
Accepted 1 February 2018
Available online 14 February 2018
but their saturated concentrations in water were still low (6.4e17
naphthyl rings. We also carried out WST-8 assays to assess the anticancer activity of the glycoGos on
DLD-1 and HepG2 cells and found that the glycoGos having -lactosides and having -galactosides
specific ligands for asialoglycoprotein receptors) showed enhanced anticancer activity on HepG2 cells.
2018 Elsevier Ltd. All rights reserved.
mM), due to their hydrophobic
b
b
Keywords:
Gossypol
Glycocluster
(
©
Chemoselective coupling
Anticancer activity
Drug delivery system
1
. Introduction
(-C¼N-R) are usually less cytotoxic than native Gos [13e15]. The
advantage of dehydration condensation as a synthetic strategy re-
lies on the following. First, it proceeds smoothly in an almost
quantitative manner just by mixing the precursors (Gos and the
amines). Second, neither catalyst nor additional reagent is required
for the reaction. Third, no byproduct except water forms through
the reaction. These points are quite advantageous for obtaining
pure Gos derivatives without time/labor-consuming purification
processes. Considering these advantages, a variety of Gos de-
rivatives have been prepared. For example, Yang et al. reported
dehydration condensation between Gos and various dipeptides to
afford Gosepeptide conjugates and assessed their anticancer ac-
tivity [16].
The introduction of carbohydrate units onto a Gos scaffold is
also attractive for pharmaceutical applications, since the resultant
Gos-based glycoconjugates (glycoGos) would have improved water
solubility and cell specificity originating from their carbohydrate
appendages. However, to date, only a couple of glycoGos have been
reported in the literature [16,17].
Gossypol (Gos), a yellow pigment from cotton plants, exerts
antioxidative [1,2], antibacterial [3], and insecticidal activities [4]
and it is widely believed to play important roles in biological de-
fense systems in these plants (Chart 1). It has also been reported
that Gos has antiproliferative activity against cancer cells [5e8] and
antiviral activity against influenza virus [9e11]. All of these bio-
activities make Gos attractive as a research target in pharmaceu-
tical/medicinal chemistry. However, the high cytotoxicity and low
water solubility of Gos, arising from its two aldehydes and two
naphthyl rings, respectively, strongly hinder its clinical applica-
tions. Many research groups have attempted to develop chemically
modified Gos (Gos derivatives) with amplified bioactivity and
reduced cytotoxicity [12]. A series of these works revealed that
dehydration condensation between the aldehydes of Gos and
various amines (H
Gos derivatives, and such Gos derivatives having imino linkages
2
N-R) is the most promising strategy to access
One of these preceding works featured glycoGos having D-
glucosamine appendages [16]. In this preceding paper, Gos was
coupled with D-glucosamine through dehydration condensation, in
which aldehydes of the former and an amino group at the C2
*
Corresponding author. Faculty of Life Sciences, Toyo University, 1-1-1 Izumino,
Itakura-machi, Ora-gun, Gumma 374-0193, Japan.
E-mail address: t-hasegawa@toyo.jp (T. Hasegawa).
https://doi.org/10.1016/j.carres.2018.02.001
0
008-6215/© 2018 Elsevier Ltd. All rights reserved.

68
Y. Amano et al. / Carbohydrate Research 458-459 (2018) 67e76
position of the latter reacted with each other to give glycoGos
carrying two glucosamine appendages. In this glycoGos, two alde-
hydes in native Gos were converted into two imino linkages and the
resultant glycoGos had reduced cytotoxicity. Its enhanced anti-HIV-
routes reported in the literature (Scheme 1) [22]. In the case of
lactose (Lac), for example, commercially available Lac monohydrate
was acetylated with acetic anhydride and pyridine, and the resul-
tant per-acetylated lactose (AcLac) was then brominated with HBr
[23]. The resultant lactosyl bromide (AcLacBr) was glycosylated
with N-hydroxysuccinimide (NHS) in a biphasic solvent system
1
activity was also reported in this previous paper.
However, this dehydration condensation is still associated with
certain problems as a general synthetic strategy to access glycoGos
for pharmaceutical use, as follows. (1) Amino groups of carbohy-
drates are essential for the dehydration condensation and thus
carbohydrates without a free amino group (e.g., glucose, galactose,
mannose, N-acetyl-glucosamine) are not acceptable in this
approach. (2) The amino group at the C2 position of D-glucosamine
is usually a critical structural motif for binding with its specific
lectins; therefore, the glucosamine appendages having imino
groups at their C2 position should in most cases have lower affinity
for these lectins. (3) In addition, the imino linkages obtained
through the dehydration condensation are usually not resistant to
hydrolysis (the reverse reaction of the dehydration condensation).
The products would thus readily be hydrolyzed to give their pre-
cursors (Gos and the amines) under physiological conditions. The
resultant Gos liberated from the Gos derivatives would have toxic
effects on patients. In this respect, it is highly attractive to design
glycoGos with improved water resistance.
containing Na
(TBAHS) to afford per-acetylated lactoside having
aglycon (AcLac NHS) at a 56% yield (two steps from AcLac). The
-linked NHS-glycosides (AcMal NHS, AcGal NHS, and
NHS) were also obtained through similar synthetic proced-
2
CO
3
and tetrabutylammonium hydrogen sulfate
b
-linked NHS
b
other
AcGlc
b
b
b
b
ures at moderate yields (40%, 56%, and 35%, respectively).
On the other hand, the corresponding glycosides having
a
-
linked NHS aglycons were prepared through alternative routes. For
example, in the case of maltose (Mal), AcMal was subjected to
glycosylation with NHS in the presence of BF
acetylated Mal having -linked NHS aglycon (AcMal
reaction, the product was a mixture containing AcMal
product) and AcMal NHS (minor product). We carried out repeated
column purification processes using various solvent systems and
finally retrieved AcMal NHS in its pure form. This purification
process greatly deteriorated the yield of this glycosylation step. Per-
acetylated galactoside having -linked NHS aglycon (AcGal NHS)
3
OEt
2
to give per-
a
aNHS). In this
aNHS (major
b
a
a
a
Recently, we established an alternative synthetic approach to
access glycoGos, in which chemically modified glycosides having
was also obtained through a similar reaction scheme starting from
galactose.
aminooxy (-ONH
2
) groups at their C1 positions (1-O-aminoglyco-
With these six a/b-linked NHS-glycosides in hand, we carried
sides) were synthesized and then coupled with Gos to afford new
glycoGos, in which two glycosides were attached onto a Gos scaf-
fold with iminooxy (-C¼N-O-R) linkages. It is widely recognized by
organic chemists that aminooxy groups react with aldehydes in a
chemoselective manner to form iminooxy linkages that are robust
even in aqueous media [18e21]. Furthermore, these new glycoGos
have O-linked glycosides and thus the carbohydrate units should
retain their inherent affinities towards their specific lectins. We
herein report synthetic procedures to access our new glycoGos,
their solubility, and their anticancer activity.
out simultaneous removal of the O-acetyl and N,N-succinimidyl
groups from these NHS-glycosides by treating them with hydrazine
monohydrate to give the corresponding 1-O-aminoglycosides.
2.2. Coupling between Gos and 1-O-aminoglycosides
The dehydration condensation of Gos with the 1-O-amino-
glycosides was achieved by mixing the former (1 eq.) and excess
amounts of the latter (3 eq.) in dry MeOH and evaporation of the
resultant mixtures (Scheme 2). Removal of the unreacted 1-O-
aminoglycosides from the residue was achieved by washing them
with small amounts of cold water to obtain the glycoGos in their
pure forms. Successful syntheses of the desired glycoGos were
confirmed through electrospray ionization time-of-flight mass
2
. Results and discussion
2.1. Synthesis of 1-O-aminoglycosides
(ESI-TOF-MS) spectral measurements. For example, in the case of
The 1-O-aminoglycosides were obtained through synthetic
the dehydration condensation of Gos with 1-O-amino- -lactoside
b
2 2 2 2 2 2 2 3
Scheme 1. Synthesis of 1-O-aminoglycosides: i) Ac O, Py, rt, 17e43 h, ii) HBr, acetic acid, Ac O, CH Cl , rt, 1.5e2.0 h, iii) NHS, TBAHS, CH Cl , Na CO aq. rt, 8.5e13 h, iv) NHS,
BF OEt , CH Cl , under N , rt, 22e24 h, v) hydrazine monohydrate, dry MeOH, rt, 2e5 h (see Experimental for the detailed reaction time and yield of each carbohydrate derivative).
3
2
2
2
2

Y. Amano et al. / Carbohydrate Research 458-459 (2018) 67e76
69
Fig. 1. Partial 1H NMR spectra (15e8 ppm) of native Gos (a) and
bLac
Gos (b) in DMF-
2
7
d (300 MHz, rt). See supplementary information for their whole spectra.
were also reported in the literature, although the linker structures
were enamine linkages in these previous works.
Scheme 2. Dehydration condensation of Gos and the 1-O-aminoglycosides followed
by the structural rearrangement to afford glycoGos having two enaminooxy linkages
The glycoGos having enaminooxy linkages were also supported
by molecular dynamics (MD) calculations on
representative. In these calculation processes, we built three
Mal Gos models in which the linkages between the Gos scaffold
and the carbohydrate appendages were different from each other;
that is, one model had two iminooxy linkages ( Mal Gosii), another
had enaminooxy and iminooxy ones ( Mal Gosei), and the other
had two enaminooxy ones ( Mal Gosee) (Chart S1). We then car-
2
aMal Gos as a
(
See Scheme 1 for the detailed carbohydrate structures): i) the 1-O-aminoglycosides (3
eq. for Gos), in dry MeOH, rt, 4e5 h.
a
2
a
2
(
Lac
molecular ion peak (m/z) at 1197.4374, which was attributed to Gos
having two -lactoside appendages ( Lac
Gos, calc. ¼1197.4350 for
2
bONH ), the ESI-TOF-MS spectrum of the product showed a
a
2
a
2
b
b
2
ried out MD calculations (Discovery Studio 4.5, CHARMm, 300 K,
ε ¼ 80, 10 ns equilibration, 90 ns dynamics) on these three models,
revealing that their average potential energies during the dynamics
C
54
H N
72 2
O28þH), proving its successful synthesis. Other glycoGos
also showed their molecular ion peaks at m/z values expected from
their elemental compositions.
1
In addition to the ESI-TOF-MS spectra, H NMR spectra of the
were 221.7 ± 5.8
11.0 ± 6.1 kcal/mol (
cate that Mal
(
a
a
Mal
Mal
2
Gosii), 219.1 ± 7.2
Gosee) (Fig. S1). These data clearly indi-
Mal Gosee
2
(aMal Gosei), and
2
2
products also showed clear evidence of the quantitative conversion
from Gos to glycoGos. For example, native Gos showed two singlet
a
2
Gos having two enaminooxy linkages (
a
2
)
1
is the most stable isomer.
H NMR peaks at 10.68 and 10.66 ppm (Fig. 1-a). These two peaks
should stem from some of the aldehyde-related structures (cis/
trans-aldehydes, hemiacetals, or enols; see Scheme S1) of Gos. In
2.4. UV-vis spectra of glycoGos
contrast,
2
bLac Gos obtained through our synthetic procedure
showed no peak around 10.6 ppm, indicating that the aldehydes
were quantitatively converted into the iminooxy linkages through
the reaction (Fig. 1-b).
Native Gos showed a broad absorption peak at 378 nm in its UV-
vis spectrum (Fig. 2-a). On the other hand, each glycoGos showed a
peak at around 355 nm that was blue-shifted (ca. 23 nm) from that
of native Gos. In addition, their intensities were amplified by factors
of 1.05e1.25 compared with that of native Gos.
2.3. Structural details of glycoGos
Interestingly, the UV-vis spectra of the glycoGos changed in a
concentration-dependent manner. In the case of
2
bLac Gos, for
As we mentioned above, our glycoGos showed no aldehyde-
derived peak in their H NMR spectra, clearly showing that both
of the two aldehydes in native Gos quantitatively reacted with the
example, it showed a strong absorption peak at 356 nm with its
1
4
ꢁ1
ꢁ1
when
molar extinction coefficient (ε) of 1.8 ꢀ 10 M cm
Lac Gos] was 1000 M (in DMF, Fig. 2-b). In contrast, its ε value
[b
2
m
1
-O-aminoglycosides. In addition, we also found that a broad peak
became small upon dilution and that of a shoulder peak (430 nm)
was simultaneously amplified. We carried out the following ex-
periments to confirm that the hydrolysis of glycoGos is not
responsible for the concentration-dependent UV-vis spectral
change. (1) Even when we used carefully dried DMF and wet DMF,
similar concentration-dependent UV-vis spectral changes were
observed for both of these glycoGos solutions. This finding showed
that water molecules in the DMF solutions have no critical effect on
(
14.1 ppm) of native Gos attributable to its C7OH also disappeared
through the reaction and two singlet peaks derived from enami-
nooxy (C8¼CH-NHO-) linkages newly appeared at around 10.0 and
11.7 ppm. This spectral feature shows that structural rearrange-
ments occurred and the iminooxy linkages first formed through the
dehydration condensation were thereafter converted into enami-
nooxy ones (Scheme 2, bottom). Similar structural rearrangements

70
Y. Amano et al. / Carbohydrate Research 458-459 (2018) 67e76
have a clear isosbestic point at around 382 nm. Since the ε value of
this isosbestic point is not affected by the glycoGos concentrations,
the absorbance at this wavelength is quite useful to assess the
concentrations of glycoGos solutions in a quantitative manner (see
the next section for details).
2.5. Water solubility of glycoGos
The introduction of carbohydrate units is one of the most widely
accepted strategies to improve the water solubility of organic/
inorganic compounds. This was also the case for Gos; that is, gly-
coGos showed enhanced water solubility in comparison to native
Gos. To assess their water solubility, we put native Gos and the
glycoGos into limited amounts of water and incubated the resultant
suspensions at ambient temperature for ca. 10 min. In the case of
native Gos, the supernatant of the resultant suspension was
colorless, confirming that it was hardly water-soluble. However, the
supernatants obtained from the glycoGos were pale yellow, sug-
gesting their improved solubility. We then estimated the saturated
concentrations of the glycoGos in water via the following proced-
ure. Briefly, we mixed a small aliquot of the supernatant obtained
from each of the aqueous glycoGos suspensions with DMF (final
DMF/water ratio ¼ 90:10 v/v). We then estimated the glycoGos
concentration of the original supernatant based on absorbance at
the isosbestic point (382 nm) of these DMF-rich solutions (Table 1).
The saturated concentration of native Gos in water is lower than
0 mM, which was too low to be quantified. The sub-micromolar
.1
solubility of native Gos has also been reported in the literature
24]. In contrast, the saturated concentrations of our glycoGos were
in the range of 6.4e17 M.
[
m
2.6. Photo-stability of glycoGos in solution
We designed our glycoGos to have improved resistance to hy-
drolysis. In fact, all of our experimental data, especially the ESI-TOF-
MS spectra, showed that our glycoGos are resistant to hydrolysis
and they do not give rise to their precursors even if stored in
aqueous media. However, it should be noted that our glycoGos are
light-sensitive and gradually decompose into various impurities in
few days, even under indoor lighting. For example, in the first stage
of our study, we attempted to purify the glycoGos obtained from
the dehydration condensation through size-exclusion chromatog-
raphy (TOYOPEARL HW40S, TOSOH) using methanol as an eluent.
The glycoGos were almost pure before the column chromato-
graphic process, but the compounds eluted out from the column
were mixtures of various impurities. The TLC analysis showed that
the impurities were neither Gos nor the 1-O-aminoglycosides,
suggesting that hydrolysis of the glycoGos is not responsible for
their decomposition. Note that this chromatographic process was
carried out under indoor lighting and it took 2e3 days until the
compounds were eluted out. It should also be noted that the gly-
coGos were also decomposed when we stored their solutions in
flasks without the size-exclusion gel. We assume that the glycoGos
are light-sensitive and they were slowly decomposed even under
Fig. 2. UV-vis spectra of glycoGos and Gos ([glycoGos] or [Gos] ¼ 1 mM) (a) and
concentration-dependent UV-vis spectral changes of
rt. (See SI for colored version of Fig. 2-a).
2
bLac Gos (b) and Gos (c): in DMF,
the concentration-dependent UV-vis spectral change. (2) No peak
attributable to Gos or Lac ONH , which are hydrolysis products
liberated from Lac Gos, was observed through the ESI-TOF-MS
b
2
b
2
Table 1
spectral analysis. (3) Native Gos also showed concentration-
dependent UV-vis spectral change in its DMF solution (Fig. 2-c).
Taking these findings together, we rule out the possibility of the
hydrolysis of the enaminooxy linkages being the mechanism
behind these concentration-dependent UV-vis spectral changes.
Instead, we assume that molecular-level aggregation of the glyco-
Gos contributes to the concentration-dependent spectral changes.
Note that these concentration-dependent UV-vis spectral changes
Solubility of the glycoGos and native Gos in water.
Saturated concentration/mM
b
b
b
a
Lac
Gal
Mal
Mal
2
Gos
Gos
6.4
7.3
8.5
17
2
2
Gos
Gos
2
a
Gos
<0.1
a
Too low to be quantified.

Y. Amano et al. / Carbohydrate Research 458-459 (2018) 67e76
71
indoor lighting, although chemical structures of the products
resulting from the light-induced decomposition have not yet been
clarified.
The light-induced decomposition of the glycoGos was clearly
supported by their limited UV-vis spectral change without indoor
lighting. In detail, we prepared two DMF solutions containing the
glycoGos and kept each of them under indoor lighting or in the dark
to measure their UV-vis spectra every 1 h. Through this experiment,
we found that the absorption peak at 354 nm that is characteristic
of the glycoGos was gradually weakened under the lighting and
after 24 h it reached a level 83% of that of freshly prepared solution
(Fig. 3). In contrast, the other sample kept in the dark still had 91%
intensity of this absorption peak in comparison to the original one.
Note that the slight weakening of intensity (9%) of the latter sample
should have arisen from photo-induced decomposition via un-
avoidable indoor light irradiation during sample manipulation.
Oxidative decomposition with
2
O in the medium was also
conceivable as one of the mechanisms behind the UV-vis spectral
change.
2.7. Anticancer activity of glycoGos
Anticancer activity of glycoGos was assessed on colorectal
adenocarcinoma (DLD-1) and hepatocellular carcinoma (HepG2)
cells. Briefly, these cells were treated with increasing doses of the
glycoGos for 48 h and then viable cell numbers were assessed
through WST-8 assays, which revealed that they showed unique
anticancer activities that were clearly different from those of native
Gos. In detail, in the case of native Gos, the viable cell numbers
monotonically decreased in a dose-dependent manner, with IC50
values of 13 and 17
mM for DLD-1 and HepG2 cells, respectively
(Fig. 4). Such strong anticancer activity should in part stem from the
highly cytotoxic aldehyde groups of native Gos. Note that these data
on native Gos showed that it had lower anticancer activity on
HepG2 than on DLD-1.
On the other hand, some glycoGos showed unique anticancer
activities that were clearly different from those of native Gos. For
Fig. 4. Viable cell numbers of DLD-1 (a) and HepG2 (b) after the treatment with Gos or
glycoGos.
example, in the case of
DLD-1 and HepG2 increased in a dose-dependent manner until the
doses reached ca. 10e40 M; then, they conversely decreased in a
dose-dependent manner. Lac Gos and Gal Gos also showed
2
bGal Gos, the viable cell numbers of both
m
b
2
b
2
similar anticancer activities for DLD-1 and HepG2 cells, respec-
tively. Although we have no information to reasonably explain the
mechanisms of these unique anticancer activities, the fact that only
certain combinations of the glycoGos and the cells showed unique
Fig. 3. UV-vis spectral change of
and time-courses of the UV-vis spectral changes of
door lighting (open circle) and in the dark (closed circle); [
DMF, rt,
¼ 1 mm.
a
Mal
2
Gos solution under the indoor lighting (inset)
Mal Gos solution under the in-
Mal M, in
Gos] ¼ 270
a
2
a
2
m
l

72
Y. Amano et al. / Carbohydrate Research 458-459 (2018) 67e76
their water solubility was still limited (6.4e17
mM). The glycoGos
(b
Lac Gos and Gal Gos) carrying specific carbohydrate ligands for
2
b
2
AGPR showed enhanced anticancer activities on HepG2 cells,
although their cell specificities were still limited. These findings
clearly show that the introduction of two carbohydrate units is not
sufficient to develop glycoGos with excellent water solubility and
cell specificity. We are now focusing our synthetic efforts on
developing dendritic compounds having Gos cores and more than
eight carbohydrate appendages. We expect that such dendritic
compounds should have excellent properties as potential water-
soluble/cell-specific anticancer drugs.
Chart 1. Structure of Gos.
4
. Experimental
anticancer activities suggests that the carbohydrate appendages
play substantial roles in their anticancer activities.
4.1. General
We estimated the IC50 values of the glycoGos on these cells
based on their anticancer activities, revealing that most glycoGos
showed higher IC50 (lower anticancer activities) on these cells than
those obtained by native Gos (Table 2). Specifically, in the case of
HepG2, all glycoGos showed higher IC50 than that obtained for
1H and 13C NMR spectra were acquired on a JEOL AL300 (JEOL
Resonance, Ltd) in CDCl
shifts were reported in ppm (
3
, CD
3
OD, D
) relative to Me
2
O or DMF-d
7
. The chemical
d
4
Si or internal refer-
ences such as HOD. IR spectra were recorded on an FT/IR-4100
Fourier transform infrared spectrometer (JASCO Co., Ltd). Matrix
assisted laser desorption ionization time-of-flight (MALDI-TOF)
mass spectra were recorded on AXIMA-CFRþ (SHIMADZU, Ltd) or
AXIMA-Confidence (SHIMADZU, Ltd). High-resolution electrospray
ionization time-of-flight (HR-ESI-TOF) mass spectra were recorded
on SYNAPT G2 HDMS (Waters, Co.). UV-vis spectra were recorded
on a JASCO V-630 Spectrophotometer (JASCO Co., Ltd). Molecular
dynamics calculations were carried out on Discovery Studio 4.5
native Gos (IC50 ¼ 17
this case, although certain glycoGos (
showed IC50 values that were equal to or lower than that obtained
by native Gos (13 M), most glycoGos showed higher ones. We
assume that chemical conversion of the highly cytotoxic aldehydes
into enaminooxy linkages was responsible for the lower anticancer
activities of the glycoGos.
m
M). This was also largely true for DLD-1. In
a
Gal Gos and Mal Gos)
2
a
2
m
Most glycoGos (
showed a higher IC50 for HepG2 than for DLD-1. The exceptions
were Lac Gos and Gal Gos, which showed similar or lower IC50
for HepG2 than for DLD-1. These findings showed that the intro-
duction of Lac or Gal units onto the Gos scaffold enhances its
2 2 2 2
aGal Gos, bMal Gos, aMal Gos, and bGlc Gos)
(
Dassault systems). Observations of DLD-1 and HepG2 cells were
carried out using Primo Vert (Zeiss). Silica gel 60 N (spherical,
neutral, particle size 63e210 m) for column chromatography was
b
2
b
2
m
b
b
purchased from Kanto Chemical Co., INC. Thin layer chromatog-
raphy (TLC) was carried out with Whatman TLC glass plates (Par-
anticancer activity against HepG2 in a cell-specific manner. Asiallo-
glycoprotein receptors (AGPR) that are specifically expressed on
HepG2 cell surfaces should bind the carbohydrate appendages of
®
tisil K6F) pre-coated with silica gel 60 Å with fluorescent indicator
(
l
¼ 254 nm). Gossypol was purchased from Funakoshi. All other
bLac
2 2
Gos and bGal Gos to increase their cellular uptake.
chemicals were purchased from Wako Pure Chemicals Industries,
Ltd., Kishida Chemicals Co., Ltd., GODO Co. Ltd., or Kanto Chemical
Co., INC. and used without purification.
Finally, we observed the DLD-1 and HepG2 cells under an optical
microscope after their treatment with the glycoGos or native Gos.
We clearly observed apoptotic bodies for these cells, showing that
not only native Gos but also the glycoGos induce their apoptosis
and that such enhancement of apoptosis is responsible for their
anticancer activities (Fig. S3). The enhanced apoptosis induced by
native Gos and the Gos derivatives has also been reported in the
literature [25].
4.2. General synthetic protocol to access per-acetylated
carbohydrates
To each of the commercially available carbohydrates (glucose,
galactose, maltose, and lactose monohydrate), pyridine and acetic
anhydride were added and the resultant heterogeneous mixture
was stirred at the ambient temperature for several hours (see the
following sections for details). After the mixture turned into
transparent and TLC analyses on the mixture (toluene/AcOEt ¼ 3/1
v/v or 1/1 v/v) showed complete conversion of the substrate, the
resultant mixture was diluted with ethyl acetate. We then added ice
to the resultant organic solution and then, the organic layer was
3
. Conclusion
We synthesized bis-glycosylated Gos through the dehydration
condensation between native Gos and 1-O-aminoglycosides. The
resultant glycoGos were resistant to hydrolysis but readily
decomposed in few days under indoor lighting. Although the gly-
coGos had improved water solubility in comparison to native Gos,
washed with 0.5 N HCl aq and NaHCO
tions, repeatedly. The organic layer was then dried over anhydrous
MgSO , filtrated and evaporated to dryness to afford the corre-
sponding per-acetylated carbohydrate as syrup or white powder.
3
-saturated aqueous solu-
Table 2
4
IC50 estimated through the WST-8 assay.
IC50
/
m
M
Relative HepG2-specificitya
4
.2.1. 1.2,3,4,6-Penta-O-acetyl-glucose (AcGlc, 1)
Glucose: 3.18 g; pyridine: 100 ml; acetic anhydride: 50 ml; re-
DLD-1
HepG2
þ
b
b
a
b
a
b
Lac
2
Gos
Gos
Gos
46
77
13
33
9
47
59
>100
42
74
1.28
1.71
<0.17
1.03
0.16
0.84
1.00
action time: 17 h; yield: 87%; MALDI-TOF-MS: [MþK] ¼ 429.01
Gal
Gal
Mal
Mal
2
(calc. 429.12).
2
2
Gos
Gos
4.2.2. 1,2,3,4,6-Penta-O-acetyl-galactose (AcGal, 2)
2
Glc
2
Gos
50
13
78
17
Galactose: 4.52 g; pyridine: 100 ml; acetic anhydride: 50 ml;
Gos
þ
reaction time: 24 h; yield: 85%; MALDI-TOF-MS: [MþNa] ¼ 413.40
a
Relative IC50(HepG2)/IC50(DLD-1) in comparison to that of Gos.
(calc. 413.12).

Y. Amano et al. / Carbohydrate Research 458-459 (2018) 67e76
73
0
0
0
0
4
.2.3. 1,2,3,6,2 ,3 ,4 ,6 -Hexa-O-acetyl-maltose (AcMal, 3)
reaction time: 13 h; yield: 35% (2 steps); white powder; R
f
(hexane/
1
Maltose: 4.12 g; pyridine: 100 ml; acetic anhydride: 75 ml; re-
action time: 43 h; yield: 97%; MALDI-TOF-MS: [MþNa] ¼ 701.46
AcOEt ¼ 1/4) ¼ 0.19; H NMR (300 MHz, CDCl
3
): d 5.30e5.19 (m,
3H), 5.07 (d, J 6.6 Hz, 1H), 4.30 (dd, J 4.8, 12.6 Hz, 1H), 4.16 (dd, J 3.0,
12.6 Hz, 1H), 3.77e3.69 (m, 1H), 2.74 (s, 4H), 2.13 (s, 3H), 2.09 (s,
þ
(
calc. 701.20).
1
3
3
3
H), 2.03 (s, 6H); C NMR (75 MHz, CDCl ): d170.6, 170.1, 170.0,
0
0
0
0
4
.2.4. 1,2,3,6,2 ,3 ,4 ,6 -Hexa-O-acetyl-lactose (AcLac, 4)
169.4, 169.3, 103.8, 72.3, 72.2, 69.6, 68.1, 61.7, 25.4, 20.8, 20.7, 20.6,
þ
Lactose monohydrate: 10.02 g; pyridine: 200 ml; acetic anhy-
dride: 100 ml; reaction time: 18 h; MALDI-TOF-MS:
MþNa] ¼ 701.54 (calc. 701.20).
20.6; MALDI-TOF-MS: [MþNa] ¼ 468.44 (calc. 468.12); IR (KBr)
ꢁ1
2965, 1755, 1738, 1380, 1278, 1227, 1188, 1076, 1032 cm .
þ
[
4
.4.2. 1-[(2,3,4,6-Tetra-O-acetyl-
pyrrolidinedione (AcGal NHS, 10)
NHS: 5.32 g; TBAHS: 3.06 g; CH
reaction time: 10 h; yield: 56% (2 steps); white powder; R
b-D-galactosyl)oxy]-2,5-
4
.3. General synthetic protocol to access per-acetylated glycosyl
b
bromides
2
Cl
2
: 70 ml; Na
2
CO
3
aq.: 70 ml;
(hexane/
3
): d 5.44e5.37 (m,
f
1
To each of the per-acetylated carbohydrates (AcGlc, AcGal,
AcMal, or AcLac) in CH Cl , acetic anhydride and 30% HBr in acetic
AcOEt ¼ 1/2) ¼ 0.08; H NMR (300 MHz, CDCl
2
2
2H), 5.08 (dd, J 4.8, 10.5 Hz, 1H), 4.92 (d, J 8.1 Hz, 1H), 4.24 (dd, J 6.3,
10.8 Hz, 1H), 4.15 (t, J 7.5 Hz, 1H), 4.11 (dd, J 7.2, 9.0 Hz, 1H), 3.92 (td J
0.9, 6.9 Hz, 1H), 2.75 (s, 4H), 2.18 (s, 3H), 2.15 (s, 3H), 2.04 (s, 3H),
acid were added and the resultant mixture was stirred at the
ambient temperature for several hours. After TLC analysis of the
mixture (toluene/AcOEt ¼ 3/1 v/v or 1/1 v/v) showed complete
consumption of the substrate, the resultant mixture was diluted
with ethyl acetate. We then added ice to the resultant solution and
then, the organic layer was washed with NaHCO
aqueous solutions, repeatedly. The organic layer was dried over
anhydrous MgSO , filtrated and evaporated to dryness to afford the
13
3
2.01 (s, 3H); C NMR (75 MHz, CDCl ): d 170.1, 170.0, 169.7, 169.5,
105.0, 70.9, 70.1, 66.6, 66.0, 60.4, 25.1, 20.5, 20.4, 20.4, 20.3; MALDI-
þ
TOF-MS: [MþNa] ¼ 468.44 (calc. 468.12); IR (KBr) 2951, 1759,
ꢁ
1
3
-saturated
1726, 1371, 1247, 1226, 1084 cm .
0
0
0
0
4
4.4.3. 1-[(2,3,6,2 ,3 ,4 ,6 -Hepta-O-acetyl-
pyrrolidinedione (AcMal NHS, 11)
NHS: 2.75 g; TBAHS: 1.67 g; CH
reaction time: 12 h; yield: 40% (2 steps); white powder; R
b-maltosyl)oxy]-2,5-
corresponding bromide as syrup or white powder. The bromide
was subjected to the subsequent glycosylation without any column
purification processes.
b
2
Cl
2
: 50 ml; Na
2
CO
3
aq.: 50 ml;
(hexane/
3
): d 5.43 (d, J 3.9 Hz,
f
1
AcOEt ¼ 1/4) ¼ 0.18; H NMR (300 MHz, CDCl
4
.3.1. 1-[(2,3,4,6-Tetra-O-acetyl-
pyrrolidinedione (AcGlcBr, 5)
AcGlc: 4.17 g; CH Cl : 50 ml; acetic anhydride: 4.0 ml; 30% HBr
in acetic acid: 20 ml; reaction time: 1.5 h.
b
-D-glucosyl)oxy]-2,5-
1H), 5.37 (t, J 9.9 Hz, 1H), 5.22 (dd, J 6.9, 8.1 Hz, 1H), 5.17 (d, J 6.6 Hz,
1H), 5.09 (t, J 6.9 Hz, 1H), 5.05 (t, J 9.9 Hz,1H), 4.84 (dd, J 4.2, 10.2 Hz,
1H), 4.45 (dd, J 3.3, 12.0 Hz, 1H), 4.38e4.25 (m, 3H), 4.15e3.98 (m,
2H), 3.86e3.81 (m, 1H), 3.98 (s, 4H), 2.75 (s, 4H), 2.15 (s, 3H), 2.12 (s,
2
2
1
3
6
H), 2.04 (s, 6H), 2.03 (s, 3H), 2.01 (s, 3H); C NMR (75 MHz,
4
.3.2. 1-[(2,3,4,6-Tetra-O-acetyl-
pyrrolidinedione (AcGalBr, 6)
AcGal: 4.00 g; CH Cl : 50 ml; acetic anhydride: 4.0 ml; 30% HBr
in acetic acid: 18 ml; reaction time: 2.0 h.
b
-D-galactosyl)oxy]-2,5-
CDCl ): 170.4, 170.3, 170.2, 170.1, 169.8, 169.8, 169.2, 169.2, 102.2,
3
d
95.3, 74.7, 72.3, 72.2, 70.2, 69.9, 69.0, 68.3, 67.8, 62.5, 61.4, 25.2, 20.8,
þ
2
2
20.6, 20.6, 20.5, 20.4, 20.4; MALDI-TOF-MS: [MþNa] ¼ 756.55
ꢁ
1
(calc. 756.21); IR (KBr) 2961, 1748, 1433, 1371, 1234, 1038 cm
.
0
0
0
0
0
0
0
0
4
.3.3. 1-[(2,3,6,2 ,3 ,4 ,6 -Hepta-O-acetyl-
pyrrolidinedione (AcMalBr, 7)
AcMal: 3.48 g; CH Cl : 40 ml; acetic anhydride: 2.6 ml; 30% HBr
in acetic acid: 12 ml; reaction time: 2.0 h.
b
-maltosyl)oxy]-2,5-
4.4.4. 1-[(2,3,6,2 ,3 ,4 ,6 -Hepta-O-acetyl-
pyrrolidinedione (AcLac NHS, 12)
NHS: 3.32 g; TBAHS: 1.94 g; CH
reaction time: 8.5 h; yield: 56% (2 steps); white powder; R
b-lactosyl)oxy]-2,5-
b
2
2
2
Cl
2
: 50 ml; Na
2
CO
3
aq.: 50 ml;
(hexane/
3
): d 5.35 (d, J 2.1 Hz,
f
1
AcOEt ¼ 1/4) ¼ 0.19; H NMR (300 MHz, CDCl
0
0
0
0
4
.3.4. 1-[(2,3,6,2 ,3 ,4 ,6 -Hepta-O-acetyl-
pyrrolidinedione (AcLacBr, 8)
AcLac: 4.13 g; CH Cl : 40 ml; acetic anhydride: 3.2 ml; 30% HBr
in acetic acid: 15 ml; reaction time: 2.0 h.
b
-lactosyl)oxy]-2,5-
1H), 5.24e5.09 (m, 4H), 4.98 (dd, J 3.3, 10.8 Hz, 1H), 4.56 (d, J 8.1 Hz,
1H), 4.44 (dd, J 1.8, 11.7 Hz, 1H), 4.23e4.05 (m, 4H), 3.90 (dd, J 6.6,
7.2 Hz, 1H), 3.80e3.75 (m, 1H), 2.74 (s, 4H), 2.16 (s, 3H), 2.13 (s, 3H),
2
2
13
2.12 (s, 3H), 2.09 (s, 3H), 2.07 (s, 3H), 2.06 (s, 3H), 1.97 (s, 3H);
C
NMR (75 MHz, CDCl ): 170.3, 170.2, 170.1, 170.0, 169.5, 169.2, 169.1,
3
d
4
.4. General synthetic protocol to access per-acetylated
102.1, 101.1, 75.7, 72.8, 72.6, 70.8, 70.7, 69.9, 68.9, 66.6, 61.8, 60.8,
þ
carbohydrate derivatives having -linked NHS aglycon
b
25.3, 20.8, 20.7, 20.6, 20.4; MALDI-TOF-MS: [MþNa] ¼ 756.49
(
calc. 756.21); IR (KBr) 2948, 1750, 1637, 1433, 1372, 1230,
ꢁ
1
To each of the bromides (AcGlcBr, AcGalBr, AcMalBr, or AcLacBr),
NHS, TBAHS, CH Cl and 1 M Na CO aqueous solution were added,
1070 cm
.
2
2
2
3
and the resultant heterogeneous solution was vigorously stirred at
the ambient temperature. After TLC analysis (hexane/AcOEt ¼ 1/4
v/v) of the mixture showed complete consumption of the substrate,
the resultant mixture was diluted with ethyl acetate. We then
washed the resultant organic layer with NaCl-saturated aqueous
solutions, repeatedly. The organic layer was dried over anhydrous
4.5. General synthetic protocol to access carbohydrate derivatives
having O- -linked NHS aglycon
b
To each of the per-acetylated carbohydrates having NHS aglycon
(AcGlc NHS, AcGal NHS, AcMal NHS, or AcLac NHS), dry MeOH
b
b
b
b
and hydrazine monohydrate were added and the resultant mixture
in tightly sealed flask was stirred for 3 min at the ambient tem-
perature. We then took off the seal and kept the resultant mixture
to be stirred for several hours under atmospheric conditions. White
precipitate was gradually formed as the reaction proceeded. The
resultant solution containing the white precipitate was again
MgSO
jected to silica-gel column purification (hexane/AcOEt ¼ 10/0 to 0/
0, gradient) to afford the titled compound.
4
, filtrated and evaporated to dryness. The residue was sub-
1
4
.4.1. 1-[(2,3,4,6-Tetra-O-acetyl-
pyrrolidinedione (AcGlc NHS, 9)
NHS: 5.30 g; TBAHS: 3.06 g; CH
b-D-glucosyl)oxy]-2,5-
ꢂ
b
tightly sealed and kept at ꢁ69 C overnight. The white precipitate
2
Cl
2
: 70 ml; Na
2
CO
3
aq.: 70 ml;
was then filtered off and the filtrate was evaporated and dried in

74
Y. Amano et al. / Carbohydrate Research 458-459 (2018) 67e76
vacuo. The resultant white powder was then re-dissolved into hot
AcMal
only through the silica-gel column purification process and the
yield of the obtained AcMal NHS was quite limited (7%). We
therefore combined the fractions containing mixtures of
AcMal NHS and AcMal NHS and the resultant solution was evap-
orated to dryness and then re-dissolved in AcOEt. Keeping the
resultant AcMal NHS/AcMal NHS solution for several days in the
NHS as white precipitates (19%).
bNHS, it is quite difficult to retrieve AcMalbNHS in pure form
ꢂ
MeOH (ca. 60 C), to which excess amount of CH
2
Cl
2
was added. The
resultant white precipitate was retrieved from the mixture through
filtration and then, purified on silica-gel (CH Cl
/MeOH ¼ 10/0 to 7/
v/v, gradient). Note that the silica-gel was carefully washed with
b
2
2
3
b
a
MeOH and dried before it was used in this column purification
process. The titled compound was obtained as white powder.
b
a
ambient temperature gave AcMal
Total yield was thus 26%.
b
4
.5.1. 1-O-Amino-
AcGlc NHS: 893 mg; dry MeOH: 20 ml; hydrazine mono-
hydrate: 534 l; reaction time: 3.3 h; yield: 17%; white powder; R
2
b-D-glucopyranoside (GlcbONH , 13)
b
m
f
4
.6.1. 1-[(2,3,4,6-Tetra-O-acetyl-
pyrrolidinedione (AcGal NHS, 17)
AcGal: 5.53 g; NHS: 4.90 g; CH
action time: 22 h; yield: 20%; white powder; R
a-D-galactosyl)oxy]-2,5-
1
(
CHCl
3
/MeOH/H
2
O ¼ 4/5/1) ¼ 0.49; H NMR (300 MHz, D
2
O): d 4.42
a
(
d, 8.1 Hz, 1H), 3.78 (d, J 12.0 Hz, 1H), 3.58 (dd, J 6.0, 12.3 Hz, 1H),
2
Cl
2
: 38 ml; BF
(hexane/AcOEt ¼ 1/
5.58 (dd, J 1.2, 3.3 Hz, 1H),
.52 (d, J 3.9 Hz, 1H), 5.48 (dd, J 3.3, 11.4 Hz, 1H), 5.23 (dd, J 4.2,
3 2
OEt : 8.9 ml; re-
13
3
D
.39e3.30 (m, 2H), 3.23 3.26e3.12 (m, 2H); C NMR (75 MHz,
O): 105.6, 76.4, 76.4, 72.2, 70.1, 61.3; IR (KBr) 3421, 3294, 2862,
f
d
1
2
2
5
) ¼ 0.31; H NMR (300 MHz, CDCl
):
d
3
ꢁ
1
1607, 1072 cm
.
1
1.1 Hz, 1H), 5.06 (td, J 0.6, 6.3 Hz, 1H), 4.25 (dd, J 6.0, 11.4 Hz, 1H),
4
.5.2. 1-O-Amino-
AcGal NHS: 895 mg; dry MeOH: 20 ml; hydrazine mono-
hydrate: 534 l; reaction time: 2.5 h; yield: 22%; white powder; R
2
b-D-galactopyranoside (GalbONH , 14)
3.95 (dd, 6.3, 10.8 Hz, 1H), 2.74 (s, 4H), 2.19 (s, 3H), 2.15 (s, 3H), 2.07
b
13
3
(s, 3H), 2.02 (s, 3H); C NMR (75 MHz, CDCl ): d 170.7, 170.4, 170.4,
m
f
1
2
70.0, 169.7, 101.2, 68.5, 67.7, 66.7, 66.5, 61.2, 25.3, 20.7, 20.6, 20.6,
1
(
CHCl
3
/MeOH/H
2
O ¼ 4/5/1) ¼ 0.41; H NMR (300 MHz, D
2
O):
d
4.37
ꢁ1
0.6; IR (KBr) 2970, 1750, 1734, 1433, 1372, 1228, 1137, 1074 cm .
(
d, J 8.1 Hz, 1H), 3.77 (d, J 3.3 Hz, 1H), 3.70e3.56 (m, 3H), 3.52 (dd, J
13
3
d
.3, 9.9 Hz, 1H), 3.37 (dd, J 8.4, 9.9 Hz, 1H); C NMR (75 MHz, D
2
O):
0
0
0
0
4
.6.2. 1-[(2,3,6,2 ,3 ,4 ,6 -Hepta-O-acetyl-
NHS, 18)
AcMal: 10.63 g; NHS: 5.40 g; CH
reaction time: 24 h; yield: 26%; white powder; R
a-maltosyl)oxy]-2,5-
106.1, 75.7, 73.4, 69.9, 69.2, 61.6; IR (KBr) 3348, 1432, 1184, 1146,
pyrrolidinedione (AcMal
a
ꢁ
1
1075, 1033 cm
.
2
Cl
2
: 50 ml; BF
3
OEt
2
: 10.0 ml;
(toluene/
f
4
.5.3. 1-O-Amino-
AcMal NHS: 732 mg; dry MeOH: 10 ml; hydrazine mono-
hydrate: 468 l; reaction time: 5.0 h; yield: 67%; white powder; R
2
b-maltoside (MalbONH , 15)
1
AcOEt ¼ 1/1) ¼ 0.30; H NMR (300 MHz, CDCl
3
):
d 5.65 (dd, J 9.0,
b
1
9
4
0.2 Hz, 1H), 5.46 (d, J 3.6 Hz, 1H), 5.42e5.35 (m, 2H), 5.10 (t, J
.6 Hz, 1H), 4.90e4.81 (m, 3H), 4.60 (dd, J 2.4, 12.6 Hz, 1H),
.29e4.21 (m, 2H), 4.08e3.96 (m, 3H), 2.73 (s, 4H), 2.15 (s, 3H), 2.14
m
f
1
(
CHCl
3
/MeOH/H
2
O ¼ 4/5/1) ¼ 0.30; H NMR (300 MHz, D
2
O): d 5.24
(
d, J 4.2 Hz, 1H), 4.42 (d, J 8.1 Hz, 1H), 3.81e3.45 (m, 9H), 3.49 (dd, J
(s, 3H), 2.11 (s, 3H), 2.08 (s, 3H), 2.06 (s, 3H), 2.04 (s, 3H), 2.01 (s,
13
3
2
.6, 9.9 Hz,1H), 3.27e3.14 (m, 2H); C NMR (75 MHz, D O): d 105.6,
13
3
3H); C NMR (75 MHz, CDCl ): d 170.7, 170.6, 170.5, 170.3, 170.2,
100.3, 77.4, 77.0, 75.2, 73.6, 73.5, 72.4, 72.3, 70.1, 61.5, 61.3; IR (KBr)
1
6
69.8, 169.6, 169.5, 100.3, 95.6, 71.8, 71.0, 70.1, 70.0, 69.9, 69.3, 68.4,
7.8, 61.9, 61.2, 25.3, 20.9, 20.8, 20.7, 20.6, 20.6, 20.5; IR (KBr) 1748,
ꢁ
1
3
531, 3307, 1160, 1056, 1028 cm
.
ꢁ
1
1434, 1371, 1235, 1038 cm
.
4
.5.4. 1-O-Amino-
AcLac NHS: 733 mg; dry MeOH: 10 ml; hydrazine mono-
hydrate: 468 l; reaction time: 2.0 h; yield: 61%; white powder; R
2
b-lactoside (LacbONH , 16)
b
4.7. General synthetic protocol to access carbohydrate derivatives
m
f
1
having O-
a
-linked NHS aglycon
(
CHCl
3
/MeOH/H
2
O ¼ 4/5/1) ¼ 0.31; H NMR (300 MHz, D
2
O): d 4.43
(
(
d, J 8.1 Hz,1H), 4.27 (d, J 7.5 Hz,1H), 3.83 (dd, J 1.8,12.9 Hz,1H), 3.75
d, J 3.3 Hz, 1H), 3.68e3.45 (m, 8H), 3.37 (dd, J 7.5, 9.9 Hz, 1H),
To each of the per-acetylated carbohydrates having NHS aglycon
(AcGal NHS or AcMal NHS), dry MeOH and hydrazine mono-
.23e3.13 (m, 1H); 1 C NMR (75 MHz, D
3
a
a
3
2
O):
d 105.4, 103.5, 78.7,
hydrate were added and the resultant mixture in tightly sealed
flask was stirred for 3 min at the ambient temperature. We then
took off the seal and kept the resultant mixture to be stirred for 3 h
under atmospheric conditions. White precipitate was gradually
formed as the reaction proceeded. The resultant solution contain-
ing the white precipitate was again tightly sealed and kept
7
5.9, 75.3, 75.0, 73.0, 71.9, 71.5, 69.1 61.6, 60.6; IR (KBr) 3366, 2890,
ꢁ1
1622, 1077, 1056 cm
.
4.6. General synthetic protocol to access per-acetylated
carbohydrate derivatives having O-
a-linked NHS aglycon
ꢂ
at ꢁ57 C for 13 h. The white precipitate was then filtered off and
To each of the per-acetylated carbohydrates (AcGal or AcMal),
the filtrate was evaporated and dried in vacuo. The resultant white
NHS, CH
was stirred under N
consumption of the substrate was confirmed through TLC analysis
2
Cl
2
and BF
3
OEt
2
were added, and the resultant mixture
ꢂ
powder was then re-dissolved into hot MeOH (ca. 60 C), to which
3
at the ambient temperature. After complete
excess amount of CH
was retrieved from the mixture through filtration and then, puri-
fied on silica-gel (CH Cl
2 2
Cl was added. The resultant white precipitate
(
hexane/AcOEt ¼ 2/1 v/v), precipitate that formed during the re-
/MeOH ¼ 10/0 to 7/3 v/v, gradient). Note
action was removed by filtration using Celite and the filtrate was
2
2
that the silica-gel was carefully washed with MeOH and dried
before it was used in this column purification process. The titled
compound was obtained as white powder.
diluted with ethyl acetate. The organic layer was then washed with
NaHCO
organic layer was dried over anhydrous MgSO
3
-saturated aqueous solutions, repeatedly. The resultant
, filtrated and
4
evaporated to dryness. TLC analysis showed that the residue con-
tained the titled compound along with some impurities containing
4.7.1. 1-O-Amino-
AcGal NHS: 446 mg; hydrazine: 267
action time: 3 h; yield: 27%; white powder; R
a
-D-galactoside (Gal
a
ONH
l; dry MeOH: 10 ml; re-
(CH Cl
/MeOH ¼ 7/
4.86 (d, J 3.9 Hz, 1H),
3.82e3.75 (m, 2H), 3.69 (dd, J 3.9, 10.5 Hz, 1H), 3.64e3.54 (m, 3H);
2
, 19)
the corresponding
b-isomer (AcGal
b
NHS or AcMal
bNHS). The res-
a
m
idue was then subjected to silica-gel column purification (hexane/
AcOEt ¼ 10/0 to 1/2, gradient) to retrieve the titled compound in
f
2
2
1
3) ¼ 0.67; H NMR (300 MHz, D
2
O): d
pure form. Since R
sponding
process greatly lowered the final yield. Especially, in the case of
f
values of the titled compound and the corre-
13
b
-isomer were so close to each other, this purification
C NMR (75 MHz, D
2
O): d 102.4, 71.5, 69.8, 69.7, 68.4, 61.6; IR (KBr)
ꢁ
1
3431, 2538, 1602, 1457, 1417, 1353, 1227, 1133, 1074 cm
.

Y. Amano et al. / Carbohydrate Research 458-459 (2018) 67e76
75
4
.7.2. 1-O-Amino-
AcMal NHS: 367 mg; hydrazine: 234
action time: 3 h; yield: 50%; white powder; R
a
-maltoside (Mal
a
ONH
2
, 20)
l; dry MeOH: 5 ml; re-
(CHCl /MeOH/
5.21 (d, J 3.6 Hz,
H), 4.81 (d, J 4.2 Hz, 1H), 3.74e3.36 (m, 11H), 3.23 (t, J 9.6 Hz, 1H);
4.8.4.
b
Lac
ONH
2
Gos (24)
: 185 mg; Gos: 82 mg; dry MeOH: 5 ml; reaction time:
(CHCl /MeOH/H
O ¼ 4/5/
11.84 (br s, 2H), 10.15 (s,
a
m
Lac
b
2
f
3
5 h; yield: 31%; pale yellow powder; R
f
3
2
1
1
H
2
O ¼ 4/5/1) ¼ 0.30; H NMR (300 MHz, D
2
O):
d
1) ¼ 0.64; H NMR (300 MHz, DMF-d
2H), 8.93 (br s, 2H), 7.71 (s, 2H), 5.63e1.52 (m); C NMR (75 MHz,
DMF-d ): 155.4, 150.9, 147.4, 132.7, 128.9, 127.6, 117.5, 115.5, 105.2,
7
): d
13
1
1
3
C NMR (75 MHz, D
2
O):
d
101.8, 100.0, 77.0, 73.7, 73.3, 73.1, 72.1,
7
d
7
1.1, 70.7, 69.7, 60.9, 60.7.
104.1, 103.8, 80.3, 75.4, 74.9, 73.3, 71.7, 70.7, 68.2, 60.5, 19.5, 19.3;
þ
ESI-TOF-MS:
[MþH] ¼ 1197.4374
(calc.
1197.4350
for
ꢁ
1
54 73 2
C H N O28); IR (KBr) 3390, 2926, 1633, 1424, 1323, 1072 cm .
4.8. General synthetic protocol to access Gos having two
carbohydrate appendages
4
.8.5.
Gal
.0 h; yield: 27%; pale yellow powder; R
a
Gal
ONH
2
Gos (25)
: 104 mg; Gos: 50 mg; dry MeOH: 2 ml; reaction time:
(CHCl
/MeOH ¼ 2/
10.03 (br s, 2H), 7.69 (br s,
H), 5.56 (d, J 4.2 Hz, 2H), 4.00e3.58 (m), 3.31 (m), 2.04 (s),
a
2
Each of the 1-O-amino-carbohydrates (3 eq.) was added to Gos
3
1
2
f
3
(
1 eq.) in dry MeOH, and the resultant solution was stirred for
1
) ¼ 0.32; H NMR (300 MHz, CD
3
OD): d
several hours. The resultant solution was evaporated to dryness
followed by dissolution with dry MeOH. This evaporation-
dissolution process was repeated for several times. After TLC
13
1.55e1.51 (m); C NMR (75 MHz, CD
3
OD): d 158.2, 158.1, 152.1,
149.1, 145.3, 134.5, 133.6, 131.2, 129.6, 117.9, 117.9, 117.6, 117.0, 107.0,
analysis (CHCl
3
/MeOH/H
2
O ¼ 4/5/1, v/v/v) of the resultant solution
102.8, 72.9, 71.3, 70.7, 69.5, 62.3, 30.7, 28.5, 28.4, 20.9; ESI-TOF-MS:
showed complete conversion from Gos to the desired product
having two carbohydrate-appendages, limited amount of water
was added. The water-insoluble compound (the glycoGos) was
retrieved by filtration followed by drying in vacuo to give the titled
compound. Note that the glycoGos were slightly water-soluble so
that certain amounts of the glycoGos were dissolved into the fil-
trates. We assume that the partial dissolution of the glycoGos into
the filtrates should be responsible for the deteriorated yields in this
synthetic step. This assumption was paradoxically supported by the
þ
[
3
MþH] ¼ 873.3293 (calc. 873.3293 for C42
53 2
H N O18); IR (KBr)
ꢁ
1
398, 2931, 1635, 1427, 1319, 1080 cm .
4
.8.6.
Mal
.0 h; yield: 50%; pale yellow powder; R
a
Mal
ONH
2
Gos (26)
: 91 mg; Gos: 15 mg; dry MeOH: 2 ml; reaction time:
(CHCl /MeOH/H
O ¼ 4/5/
10.07e10.04 (m, 2H), 7.69
a
2
3
1
f
3
2
1
) ¼ 0.59; H NMR (300 MHz, CD
3
OD): d
(s, 2H), 5.55e5.54 (m, 2H), 5.19e6.16 (m, 2H), 3.99e3.20 (m),
13
2
3
.17e2.03 (m), 1.55e1.52 (m); C NMR (75 MHz, CD OD): d 158.3,
fact that
study was obtained in the highest yield (99%). In this case, the
hardly water-soluble nature of Glc Gos should minimize the lost
of Glc Gos into the filtrate to improve the recovery efficiency.
2
bGlc Gos that is the least water-soluble glycoGos in this
1
1
7
58.3, 152.2, 152.1, 149.1, 149.0, 145.2, 134.5, 131.3, 129.7, 118.0, 118.0,
17.6, 117.0, 107.0, 107.0, 102.8, 102.8, 102.4, 102.3, 81.2, 81.1, 75.0,
4.9, 74.8, 74.7, 74.7, 74.2, 74.1, 72.9, 72.9, 72.4, 71.4, 71.4, 62.7, 61.7,
b
2
b
2
þ
6
1.7, 58.3, 20.9, 20.8, 20.8, 18.4; ESI-TOF-MS: [MþH] ¼ 1197.5671
(calc. 1197.4350 for C54 28); IR (KBr) 3389, 2927, 1617, 1426,
73 2
H N O
4
.8.1.
Glc
h; yield: 99%; pale yellow powder; R
) ¼ 0.76; H NMR (300 MHz, CD
b
Glc
ONH
2
Gos (21)
: 28 mg; Gos: 32 mg; dry MeOH: 5 ml; reaction time:
(CHCl /MeOH/H
O ¼ 4/5/
10.03 (br s, 2H), 7.67 (br s,
H), 5.03 (d, J 6.3 Hz, 2H), 3.89 (dd, J 2.1, 9.3 Hz, 2H), 3.71 (dd, J 4.5,
ꢁ
1
1320, 1035 cm
.
b
2
4
1
2
f
3
2
4.9. WST-8 assays
1
3
OD): d
Cell viability upon the glycoGos administration was assessed
through a well-known protocol using a Cell Counting Kit-8
Dojondo Molecular Technologies, Inc.). Briefly, 100 l of DLD-1
13
12 Hz, 2H), 3.49e3.29 (m), 2.04 (s), 1.55e1.52 (m); C NMR
(
75 MHz, CD OD): 157.8, 152.2, 149.1, 145.3, 134.5, 131.3, 129.7,
3
d
(
m
1
18.0, 117.1, 107.0, 106.1, 78.3, 78.2, 73.6, 71.2, 62.5, 30.7, 20.9, 20.7;
and HepG2 cell suspensions (30000 cells/ml) were incubated in
9
þ
ESI-TOF-MS: [MþH] ¼ 873.4611 (calc. 873.3293 for C42
53 2
H N O18);
ꢂ
6-well plates at 37 C with RPMI 1640 culture media, or 10% fetal
ꢁ
1
IR (KBr) 3399, 2926, 1616, 1424, 1323, 1244, 1179, 1075 cm
.
bovine serum (FBS) containing 0.5% penicillin streptomycin, for
4 h in a CO incubator. After the incubation, the media was
removed from the wells using an aspirator and then, 100 l of RPMI
640 media containing 0.2% DMF and varying concentration of the
glycoGos were applied to the wells. After 48 h incubation, 10 l of
2
2
4
.8.2.
Gal
h; yield: 48%; pale yellow powder; R
) ¼ 0.70; H NMR (300 MHz, CD
b
Gal
ONH
2
Gos (22)
: 27 mg; Gos: 36 mg; dry MeOH: 5 ml; reaction time:
(CHCl /MeOH/H
O ¼ 4/5/
10.03 (s, 2H), 7.69 (s, 2H),
.01 (d, J 8.1 Hz, 2H), 3.95 (br s, 2H), 3.89 (d, J 4.5 Hz, 2H), 3.81e3.62
m
b
2
1
4
1
5
f
3
2
m
1
3
OD): d
Cell Counting Kit-8 solution were applied to the wells, and cells
were further incubated at 37 C for 30 min in the CO
Color development was monitored at 450 nm with a multi-well
plate reader (SUNRISE Rainbow RC-R, Tecan) in which absorbance
at 650 nm was used as a reference.
ꢂ
2
incubator.
(
m, 8H), 3.57 (dd, J 3.3, 9.6 Hz, 2H), 2.04 (s, 4H), 1.55e1.51 (m, 12H);
1
3
C NMR (75 MHz, CD
3
OD): d 157.7, 152.3, 149.0, 145.2, 134.5, 131.3,
129.7, 118.0, 117.6, 117.1, 107.0, 106.7, 77.0, 75.1, 71.0, 70.2, 62.4, 35.0,
3
3.1, 30.7, 30.4, 30.2, 28.6, 26.1, 23.7, 20.9, 20.7, 14.2; ESI-TOF-MS:
þ
[MþH] ¼ 873.4482 (calc. 873.3293 for C42
53 2
H N O18); IR (KBr)
4.10. Trypan blue exclusion test
ꢁ1
3
421, 2926, 1618, 1424, 1323, 1078 cm .
DLD-1 and HepG2 cell suspensions (3 ml, 30000 cells/ml) were
ꢂ
4
.8.3.
Mal
.5 h; yield: 42%; pale yellow powder; R
b
Mal
ONH
2
Gos (23)
incubated on culture dishes for 24 h at 37 C with RPMI 1640 cul-
ture media (10% FBS containing 0.5% penicillin streptomycin) for
24 h. After the incubation, the media were removed from the plates
using an aspirator and then, 3 ml of the glycoGos solutions (RPMI
1640 media containing 2% FBS, 0.2% DMF and varying concentration
of the glycoGos or Gos) were applied to the plates. After 48 h in-
b
2
: 55 mg; Gos: 40 mg; dry MeOH: 5 ml; reaction time:
4
1
2
f
(CHCl
3
/MeOH/H
2
O ¼ 4/5/
1
) ¼ 0.54; H NMR (300 MHz, DMF-d
7
):
d
11.85 (br s, 2H), 10.15 (s,
13
H), 8.94 (br s, 2H), 7.71 (s, 2H), 5.76e1.53 (m); C NMR (75 MHz,
): 155.4, 150.9, 147.4, 143.4, 139.2, 132.7, 128.9, 127.6, 117.6,
DMF-d
7
d
ꢂ
115.5, 105.2, 104.3, 100.9, 79.1, 76.2, 75.4, 73.3, 72.6, 71.6, 70.1, 61.0,
cubation of the cells at 37 C, the supernatants were retrieved from
þ
6
0.2, 19.5, 19.3; ESI-TOF-MS: [MþH] ¼ 1197.4374 (calc. 1197.4350
the culture dishes (supernatant-1). The dishes with DLD-1 or
HepG2 cells were then added with FBS and then, the supernatants
were again retrieved from the culture dishes (supernatant-2). The
for C54
73 2
H N O
28); IR (KBr) 3389, 2928, 1631, 1424, 1323, 1244, 1146,
.
ꢁ1
1071, 1047 cm

76
Y. Amano et al. / Carbohydrate Research 458-459 (2018) 67e76
DLD-1 and HepG2 cells on the culture dishes were washed with
PBS. Trypsin-EDTA solution was added to the dishes those were
then incubated for 5 min. The resultant media containing DLD-1 or
HepG2 cells were retrieved and then mixed with the supernatants
474e483.
[
[
[
[
5] S.Z.A. Badawy, A. Souid, V. Cuenca, N. Montalto, F. Shue, Asian J. Androl. 9
(2007) 388e393.
6] L. Huang, J. Hu, W. Tao, Y. Li, G. Li, P. Xie, X. Liu, J. Jiang, Eur. J. Pharmacol. 645
(2010) 9e13.
7] X. Wang, J. Wang, S.C.H. Wong, L.S.N. Chow, J.M. Nicholls, Y.C. Wong, Y. Liu,
D.L.W. Kwong, J.S.T. Sham, S.W. Tsao, Life Sci. 67 (2000) 2663e2671.
8] C.V. Poznak, A.D. Seidman, M.M. Reidenberg, M.M. Moasser, N. Sklarin,
K.V. Zee, P. Borgen, M. Gollub, D. Bacotti, T. Yao, R. Bloch, M. Ligueros,
M. Sonenberg, L. Norton, C. Hudis, Breast Canc. Res. Treat. 66 (2001) 239e248.
9] J. Yang, F. Zhang, J. Li, G. Chen, S. Wu, W. Ouyang, W. Pan, R. Yu, J. Yang, P. Tien,
Bioorg. Med. Chem. Lett 22 (2012) 1415e1420.
(
suparnatant-1 and -2) in 15 ml centrifuge tube. The DLD-1 or
HepG2 cells were collected through the centrifugation and re-
suspended with RPMI 1640 media. The resultant cell suspensions
were mixed with the equal volumes of trypan blue solution to be
observed by using an optical micrometer Primo Vert, Zeiss).
[
[
[
10] J. Yang, G. Chen, L.L. Li, W. Pan, F. Zhang, J. Yang, S. Wu, P. Tien, Bioorg. Med.
Chem. Lett 23 (2013) 2619e2623.
11] N.I. Baram, A.I. Ismailov, K.L. Ziyaev, K.Z. Rezhepov, Chem. Nat. Compd. 40
Acknowledgement
(
2004) 199e205.
This research was partially supported by Special Research Fund
of Toyo University.
[12] V. Dao, C. Gaspard, M. Mayer, G.H. Werner, S.N. Nguyen, R.J. Michelot, Eur. J.
Med. Chem. 35 (2000) 805e813.
13] L. Li, Z. Li, K. Wang, S. Zhao, J. Feng, J. Li, P. Yang, Y. Liu, L. Wang, Y. Li, H. Shang,
[
Q. Wang, Design, J. Agric. Food Chem. 62 (2014) 11080e11088.
[14] J. Yang, F. Zhang, J. Li, G. Chen, S. Wu, W. Ouyang, W. Pan, R. Yu, J. Yang, P. Tien,
Appendix A. Supplementary data
Bioorg. Med. Chem. Lett 22 (2012) 1415e1420.
[
[
[
[
15] V. Dao, M.K. Dowd, M.T. Martin, C. Gaspard, M. Mayer, R.J. Michelot, Eur. J.
Supplementary data associated with this article (structures of
the representative Gos isomers with interconversions, chemical
structures of aMal Gosii, aMal Gosei, and aMal Gosee for the MD
2 2 2
calculations, plots of the potential energies during the dynamics,
colored version of Fig. 2-(a), microscopic image of the HepG2 cells
Med. Chem. 39 (2004) 619e624.
16] J. Yang, J. Li, J. Yang, L. Li, W. Ouyang, S. Wu, F. Zhang, Chin. Chem. Lett. 25
(2014) 1052e1056.
17] J. Yin, L. Jin, F. Chen, X. Wang, A. Kitaygorodskiy, Y. Jiang, Carbohydr. Res. 346
(2011) 2070e2074.
18] J. Zhang, J. Lv, X. Wang, D. Li, Z. Wang, G. Li, Nano Res. 8 (2015) 3853e3863.
after the
b
Gal
2
Gos-treatment followed by trypan blue staining, and
[19] K.D. McReynolds, D. Dimas, H. Le, Tetrahedron Lett. 55 (2014) 2270e2273.
[20] P.K. Dhal, S.C. Polomoscanik, D.A. Gianolio, P.G. Starremans, M. Busch,
K. Alving, B. Chen, R.J. Miller, Bioconjugate Chem. 24 (2013) 865e877.
1
13
H/ C NMR spectra of the synthetic compounds) can be found at
https://doi.org/10.1016/j.carres.2018.02.001.
[
21] C. Pifferi, G.C. Daskhan, M. Fiore, T.C. Shiao, R. Roy, O. Renaudet, Chem. Rev.
17 (2017) 9839e9873.
22] T.C. Shiao, A. Papadopoulos, O. Renaudet, R. Roy, Preparation of O-
1
[
b
-D-gal-
References
actopyranosyl-hydroxylamine, in: P. Kov aꢀ ꢁc (Ed.), Carbohydrate Chemistry:
Proven Synthetic Methods, vol. 1, CRC Press, 2012, pp. 289e293.
23] Y. Nonaka, R. Uruno, F. Dai, R. Matsuoka, M. Nakamura, M. Iwamura,
[
1] M. Ionov, N.V. Gordiyenko, I. Zukowska, E. Tokhtaeva, O.A. Mareninova,
[
N. Baram, K. Ziyaev, K. Rezhepov, M. Zamaraeva, Int. J. Biol. Macromol. 51
H. Iwabuchi, T. Okada, N. Chigira, Y. Amano, T. Hasegawa, Tetrahedron 72
(
2012) 908e914.
(2016) 5456e5464.
[
[
[
2] K. Dodou, R.J. Anderson, W.J. Lough, D.A.P. Small, M.D. Shelley,
[
[
24] F. Yan, X. Cao, H. Jiang, X. Zhao, J. Wang, Y. Lin, Q. Liu, C. Zhang, B. Jiang, F. Guo,
J. Med. Chem. 53 (2010) 5502e5510.
25] K. Sadahira, M. Sagawa, T. Nakazato, H. Uchida, Y. Ikeda, S. Okamoto,
H. Nakajima, M. Kizaki, Int. J. Oncol. 45 (2014) 2278e2286.
P.W. Groundwater, Bioorg. Med. Chem. 13 (2005) 4228e4237.
3] P. Przybylski, K. Pyta, J. Stefa nꢀ ska, M. Ratajczak-Sitarz, A. Katrusiak,
A. Huczy nꢀ ski, B. Brzezinski, Eur. J. Med. Chem. 44 (2009) 4393e4403.
4] L. Li, Z. Li, K. Wang, Y. Liu, Y. Li, Q. Wang, Bioorg. Med. Chem. 24 (2016)