Glycosylated tris-bipyridine ferrous complexes for probing a mechanism behind carbohydrate-carbohydrate interactions: Spatial carbohydrate packing of glycoclusters changes on additions of salts in carbohydrate- and anion-dependent manners

Accepted Manuscript

Glycosylated tris-bipyridine ferrous complexes for probing a mechanism behind

carbohydrate-carbohydrate interactions: Spatial carbohydrate packing of glycoclusters

changes on additions of salts in carbohydrate- and anion-dependent manners

Naoto Chigira, Fumiko Dai, Yuki Nonaka, Koki Sato, Yoshitsugu Amano, Maki

Sekiguchi, Mayu Inokuchi, Masahito Hagio, Teruaki Hasegawa

PII:

S0040-4020(18)30970-0

10.1016/j.tet.2018.08.017

TET 29737

DOI:

Reference:

To appear in: Tetrahedron

Received Date: 14 June 2018

Revised Date: 24 July 2018

Accepted Date: 13 August 2018

Please cite this article as: Chigira N, Dai F, Nonaka Y, Sato K, Amano Y, Sekiguchi M, Inokuchi M,

Hagio M, Hasegawa T, Glycosylated tris-bipyridine ferrous complexes for probing a mechanism behind

carbohydrate-carbohydrate interactions: Spatial carbohydrate packing of glycoclusters changes on

additions of salts in carbohydrate- and anion-dependent manners, Tetrahedron (2018), doi: 10.1016/

j.tet.2018.08.017.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to

our customers we are providing this early version of the manuscript. The manuscript will undergo

copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please

note that during the production process errors may be discovered which could affect the content, and all

legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Graphical Abstract

Glycosylated tris-bipyridine ferrous

Leave this area blank for abstract info.

complexes for probing a mechanism behind

carbohydrate-carbohydrate interactions:

spatial carbohydrate packing of glycoclusters

changes on additions of salts in

carbohydrate- and anion-dependent

manners

Naoto Chigira ^{a §}, Fumiko Dai ^{a §}, Yuki Nonaka ^a, Koki Sato ^b, Yoshitsugu Amano ^a, Maki Sekiguchi ^b, Mayu

Inokuchi ^b, Masahito Hagio ^band Teruaki Hasegawa^{b, c,}

^aGraduate School of Life Sciences, Toyo University, 1-1-1 Izumino, Itakura-machi, Ora-gun, Gumma 374-0193,

Japan, ^bFaculty of Life Sciences, Toyo University, 1-1-1 Izumino, Itakura-machi, Ora-gun, Gumma 374-0193,

Japan, ^cBio-Nano Electronics Research Centre, Toyo University, 2100 Kujirai, Kawagoe, Saitama 350-8585,

Japan.

^§These two authors contributed equally to this work.

1

ACCEPTED MANUSCRIPT

Tetrahedron

journal homepage: www.elsevier.com

Glycosylated tris-bipyridine ferrous complexes for probing a mechanism behind

carbohydrate-carbohydrate interactions: spatial carbohydrate packing of glycoclusters

changes on additions of salts in carbohydrate- and anion-dependent manners

Naoto Chigira ^{a, §}, Fumiko Dai ^{a, §}, Yuki Nonaka ^a, Koki Sato ^b, Yoshitsugu Amano ^a, Maki Sekiguchi ^b,

Mayu Inokuchi ^b, Masahito Hagio ^band Teruaki Hasegawa ^{b, c,}

^aGraduate School of Life Sciences, Toyo University, 1-1-1 Izumino, Itakura-machi, Ora-gun 374-0193, Japan

^bFaculty of Life Sciences, Toyo University, 1-1-1 Izumino, Itakura-machi, Ora-gun 374-0193, Japan

^cBio-Nano Electronics Research Centre, Toyo University, 2100 Kujirai, Kawagoe 350-8585, Japan

^§These two authors contributed equally to this work

Corresponding author

E-mail address: t-hasegawa@toyo.jp

ARTICLE INFO

ABSTRACT

Article history:

Received

Received in revised form

Accepted

2,2'-Bipyridines containing two β-maltoside, β-lactoside, or β-isomaltoside appendages were

prepared and successively complexed with ferrous ion to afford hexavalent glycoclusters having

tris-bipyridine ferrous complex cores. Each of these metalloglycoclusters showed unique UV-vis

and CD spectral changes upon addition of chlorides, nitrates, and sulfate salts in carbohydrate-

and anion-dependent manners. The results indicate that spatial carbohydrate packing of the

metalloglycoclusters changes upon addition of these anions and that different anions stabilize

different carbohydrate packings. Furthermore, the sulfates specifically enhance the rheological

properties of aqueous solutions containing the metalloglycocluster containing β-lactoside

appendages.

Available online

Keywords:

carbohydrate-carbohydrate interactions

tris-bipyridine ferrous complex

glycocluster

dynamic combinatorial library

circular dichroism

.

spatial carbohydrate packing

1. Introduction

concerning the conformation (spatial packing of the carbohydrate

units) in glycoclusters.

Glycosphingolipids (GSLs) in cell membranes aggregate

laterally to form GSL-enriched microdomains presenting densely

packed carbohydrate clusters (glycoclusters) on cell surfaces

(Fig. 1-a).^1–3It has been reported that these cell surface

glycoclusters play essential roles in cell-cell adhesion via

carbohydrate-carbohydrate interactions (CCIs), wherein GSLs on

the surface of one cell recognize GSLs on the surface of an

adjacent cell in specific, multivalent, and, in some cases, Ca²⁺-

dependent manners.⁴For example, Le^X-Le^Xand G_M3-Gg3

interactions were reported to induce compaction of blastomeres

and migration of cancer cells, respectively.^5,6Since cell-cell

adhesion is a fundamental bioprocess in multicellular organisms,

investigation into CCIs is of quite high importance from an

academic viewpoint. It is also an attractive research topic for

industry, since novel knowledge will support the design of new

drugs for the prevention of various diseases triggered by

unfavorable cell-cell adhesions (metastases, inflammations, etc.).

In spite of the importance of CCIs, the fluidic nature of cell

membranes as well as dynamic fluctuations in their GSL levels

make it quite difficult to investigate CCIs in a detailed manner. In

fact, very little molecular-level information about CCIs has been

obtained thus far. In particular, no information is available

The weakness of CCIs further compounds the difficulty in

their investigation;

a single CCI between two isolated

carbohydrates is too weak to be detected. Well-designed artificial

systems are thus vital in this research field^7–22; such artificial

systems may contain synthetic polymers,⁸peptides,^9,10

foldamers,¹¹cyclodextrins,¹²and liposomes,^13–15all having

multiple carbohydrate appendages. Gold nanoparticles (GNPs)

coated with densely packed glycoclusters are also well-known

model systems for probing CCIs.^16–18For example, Le^X-Le^X

interactions were successfully probed through Ca²⁺-mediated

aggregation of GNPs carrying multiple Le^X-trisaccharides (Le^X;

3

Galβ1,4(Fucα1,3)GlcNAcβ).¹⁸Dendrimers having multiple

lactosides (Lac; Galβ1,4Glcβ) and lipid monolayers containing

G_M3 (NeuAcα2,3Galβ1,4Glcβ) were also useful for probing

G_M3-Lac interactions at an air-water interface.¹⁹Lac-Lac

interactions were also successfully probed using a polyfluorene

with multiple Lac appendages; Ca²⁺mediated association of the

Lac-appended polyfluorene more efficiently than did K⁺, Na⁺,

Mg²⁺, or Ba²⁺.²⁰As these previous researches have exemplified,

researchers in this field commonly believe that the ions function

as a molecular glue to join two carbohydrate units together, and

that formation of multiple carbohydrate-ion-carbohydrate

2

Tetrahedron

ACCEPTED MANUSCRIPT

2. Results and Discussion

complexes is a main driving force mediating the CCIs.

However, no attention has been paid to the spatial carbohydrate

packings within these glycoclusters.

2.1. Syntheses of glycosylated bipyridines (glycobpys)

Commercially available 5,5'-dimethyl-2,2'-bipyridine was

oxidized with KMnO₄in boiling water, and the resultant 2,2'-

bipyridine-5,5'-dicarboxylic acid (bpyCOOH) was then converted

into 2,2'-bipyridine-5,5'-dicarboxyl chloride (bpyCOCl) by

SOCl₂(Scheme 1). The coupling of bpyCOCl with per-O-acetyl-

We have also developed various model systems for probing

CCIs.^23-26Recently, we developed tris-bipyridine (tris-bpy)

ferrous complex having six Lac appendages ([Fe(bpy)Lac)₃]²⁺)

and the same having six maltoside (Mal; Glcα1,4Glcβ)

appendages

([Fe(bpyMal)₃]²⁺).²⁶

Each

of

these

p-aminophenyl-β-maltoside

aminophenyl-β-lactoside (pAPLac), and

(pAPMal),

per-O-acetyl-p-

metalloglycoclusters exists as a mixture of two diastereomers (∆

and Λ) with different carbohydrate packings (Fig. S1). Their ∆-

and Λ-[Fe(bpy)₃]²⁺cores have screw-propeller-like 3D structures,

in which three bpy units coordinate the Fe²⁺in right- and left-

handed manners, respectively. Since the N-Fe bonds are

reversible at ambient temperature, the two diastereomers are

under dynamic equilibrium, in which the ∆-Λ ratio changes

depending on their relative stabilities (Fig. S2). Molecular

aminophenyl-β-isomaltoside (pAPIma) was followed by

deacetylation to afford bpy having two lactoside (bpyLac),

maltoside (bpyMal), or isomaltoside (bpyIma) appendages,

respectively. Here, isomaltoside (Ima; Glcα1,6Glcβ) is a

structural isomer of Mal, and differs from Mal only in its

glycosidic linkages. The synthesis of bpyIma is reported for the

first time in this work and its detailed synthetic procedure and

characterization are reported, while those of bpyMal and bpyLac

were reported in our previous work.²⁶

dynamics

(MD)

simulations

revealed

that

these

metalloglycoclusters have unique 3D structures in which two

trivalent glycoclusters are positioned at the northern and southern

poles of the spherical [Fe(bpy)₃]²⁺cores (Fig. 1-b). Each of these

trivalent glycoclusters would mimic the natural glycoclusters on

cell surfaces. An advantage of the metalloglycoclusters created

by our group is that the [Fe(bpy)₃]²⁺cores, which act as

molecular probes, are located in close proximity to the trivalent

glycoclusters. Since the trivalent glycoclusters are energetically

coupled with the [Fe(bpy)₃]²⁺cores, ion-induced conformational

changes in the former can be readily monitored by the ∆-Λ ratio

of the latter. Using such model systems, we investigated

carbohydrate packings in the systems both with and without ions,

and found that various ions including Na⁺, K⁺, Mg²⁺, Ca²⁺, and

Cl^-bind to the glycoclusters and greatly change their

carbohydrate packing in both carbohydrate- and ion-dependent

manners. It should be noted that Lac is the carbohydrate unit of

lactosylceramide (LacCer), which is one of major GSLs in cell

membranes as well as a common structural motif of various

GSLs (e.g., GM3 and Gg3). This finding is interesting since it

suggests a mechanism behind CCI, which is that the ions make

the carbohydrate packings suitable for the CCI and that the

resultant glycoclusters smoothly interact with other glycoclusters

intercellularly.

Fig. 1 (a) Illustration of glycoclusters on a cell surface and (b) the most stable

conformation of ∆-[Fe(bpyLac)₃]²⁺during the MD simulations (See Fig. S3

for the colored version of this figure).

Scheme 1 Reaction conditions: i) KMnO₄, water, reflux, overnight, 74%, ii)

SOCl₂, reflux, 5 h, quant., iii) pAPMal, Et₃N, THF, rt, 20 min, 46%, vi)

aqueous ammonia, MeOH, rt, 3 days, quant., v) pAPLac, Et₃N, THF, rt, 5

min, 36%, vi) aqueous ammonia, MeOH, rt, overnight, quant., vii) pAPIma,

Et₃N, THF, rt, 20 min, 61%, viii) aqueous ammonia, MeOH, rt, overnight,

quant..

In our previous paper, attention was primarily given to the

effects of cations on the spatial carbohydrate packings, since it is

widely believed that the cations (Ca²⁺in some cases) play

important roles in CCIs. However, in our further studies it was

revealed that anions also have a great impact on spatial

carbohydrate packings. We herein report novel experimental data

concerning the ability of the anions to change spatial

carbohydrate packings, and preliminary data concerning

intermolecular CCIs between metalloglycoclusters.

2.2. Preparation of tris-bipyridine ferrous complexes having six

carbohydrate appendages

Usually, bpy derivatives are readily converted to the

corresponding tris-bipyridine ferrous complexes in quantitative

3

yields by mixing them with a one-third molar equivalent of Fe²⁺

bound to [Fe(bpyMal)₃]²⁺and induced its UV-vis spectral

response.

ACCEPTED MANUSCRIPT

in aqueous solution. However, water solubility of the synthesized

glycobpys was so low that their simple complexation with Fe²⁺

was not accomplished. In fact, repeated heating/sonication

processes with 3 molar equivalents of Fe²⁺were required for their

complexation. The resultant solutions were radish purple in color

without any precipitates, indicating successful and quantitative

conversion of bpyMal, bpyLac, and bpyIma into [Fe(bpyMal)₃]²⁺,

[Fe(bpyLac)₃]²⁺, and [Fe(bpyIma)₃]²⁺, respectively. Even in the

presence of excess amounts of Fe²⁺, the metalloglycoclusters

having [Fe(bpy)₃]²⁺cores were formed exclusively, as

determined from their strong metal-to-ligand charge transfer

(MLCT) UV-vis bands at ~575 nm (Fig. S4).

In contrast to the clear UV-vis spectral changes induced by the

chloride and sulfate salts, the nitrate salts induced only negligible

UV-vis spectral changes (Fig. 2, black symbols). However, it

should be noted that these negligible UV-vis spectral changes

does not signify that no affinity exists between the nitrates and

[Fe(bpyMal)₃]²⁺. On the contrary, their binding was clearly

demonstrated via circular dichroism (CD) spectroscopy

measurements, as discussed in section 2.4.

2.3.2. Salt-induced UV-vis spectral changes of

[Fe(bpyLac)₃] ^{2 +}

The salt-induced UV-vis spectral changes of [Fe(bpyLac)₃]²⁺

were clearly different from those of [Fe(bpyMal)₃]²⁺. In the case

of the chloride salts, [Fe(bpyLac)₃]²⁺showed negligible UV-vis

spectral changes, in clear contrast to the enhanced MLCT

intensity induced in [Fe(bpyMal)₃]²⁺(Fig. 3, gray symbols).

Different UV-vis spectral responses between [Fe(bpyMal)₃]²⁺and

[Fe(bpyLac)₃]²⁺were also observed for the sulfate salts (Fig. 3,

white symbols). In this case, the MLCT intensity of

[Fe(bpyLac)₃]²⁺decreased with increasing sulfate salt

concentration, falling nearly to zero when sulfate salt

concentration was increased to ~0.3 M. In contrast to what was

observed for [Fe(bpyLac)₃]²⁺, [Fe(bpyMal)₃]²⁺still retained

substantial MLCT intensity (I/I₀= 0.5–0.6) even under much

higher sulfate-concentrations (e.g., [Na₂SO₄] = 1.25 M, See

section 2.3.1).

Fig. 2 (a) MLCT intensity of [Fe(bpyMal)₃]²⁺in the presence of various

concentrations of salts and (b) an enlarged version. [bpyMal] = 2.44 mM,

[FeCl₂] = 7.33 mM, d = 0.2 mm, rt, in water.

2.3. Salt-induced UV-vis spectral changes of the

metalloglycoclusters

2.3.1. Salt-induced UV-vis spectral changes of

[Fe(bpyMal)₃]^{2 +}

The metalloglycoclusters showed unique UV-vis spectral

responses toward chloride (NaCl, KCl, and CaCl₂), nitrate

(NaNO₃, KNO₃, and Ca(NO₃)₂), and sulfate salts (Na₂SO₄and

K₂SO₄) (Fig. S4). For example, in the case of [Fe(bpyMal)₃]²⁺,

the intensity of its MLCT band clearly increased in the presence

of the chloride salts (Fig. 2, gray symbols). In contrast, the

sulfate salts exhibited different effects, drastically decreasing the

MLCT intensity upon their addition (Fig. 2, white symbols). Note

that in both groups (the chloride and the sulfate salts), little

difference was induced in the UV-vis spectral response by the

Fig. 3 (a) MLCT intensity of [Fe(bpyLac)₃]²⁺in the presence of various

concentrations of the salts and (b) an enlarged version. [bpyLac] = 2.44 mM,

[FeCl₂] = 7.33 mM, d = 0.2 mm, rt, in water.

2.3.3. Salt-induced UV-vis spectral changes of

[Fe(bpyIma)₃] ^{2 +}

We also observed unique salt-induced UV-vis spectral

responses of [Fe(bpyIma)₃]²⁺which differed from those of both

2-

different cations, indicating that their anions (Cl^-and SO₄)

4

Tetrahedron

[Fe(bpyMal)₃]²⁺and [Fe(bpyLac)₃]²⁺. The most interesting data

packing suitable for the ∆-[Fe(bpy)₃]²⁺core. This Cl^--induced

ACCEPTED MANUSCRIPT

were obtained for the sulfate salts, which induced negligible UV-

vis spectral changes in [Fe(bpyIma)₃]²⁺(Fig. 4, white symbols).

This is in clear contrast to the decreased MLCT intensity

observed for [Fe(bpyMal)₃]²⁺and [Fe(bpyLac)₃]²⁺in the presence

of the sulfate salts. The negligible UV-vis spectral changes

provide strong evidence of the stability of the [Fe(bpy)₃]²⁺cores

in the presence of these sulfate salts.

conformational change of [Fe(bpyMal)₃]²⁺was discussed in detail

in our preceding work, in which it was assumed that multiple

electrostatic interactions stemmed from the amide linkages (-

NHCO-) in [Fe(bpyMal)₃]²⁺are responsible for the Cl^--binding.²⁶

Fig. 4 (a) MLCT intensity of [Fe(bpyMal)₃]²⁺in the presence of various

concentrations of the salts and (b) an enlarged version. [bpyMal] = 2.44 mM,

[FeCl₂] = 7.33 mM, d = 0.2 mm, rt, in water.

2.4. Circular dichroism (CD) spectral changes of the

metalloglycoclusters

2.4.1. Salt-induced CD spectral changes of

[Fe(bpyMal)₃]^{2 +}

In the absence of salts, [Fe(bpyMal)₃]²⁺exhibited a negative-

to-positive CD pattern, as measured from longer to shorter

wavelengths, centered around 315 nm (Fig. 5-a, bold line). This

CD pattern resembles that reported for enantio-pure ∆-

[Fe(bpy)₃]²⁺, indicating that [Fe(bpyMal)₃]²⁺predominantly

exists in its ∆ form in pure water.²⁷However, the signal

intensities of [Fe(bpyMal)₃]²⁺were much weaker in comparison

to those of enantio-pure ∆-[Fe(bpy)₃]²⁺, indicating its low

diastereomeric excess. This data indicates that the glycoclusters

of [Fe(bpyMal)₃]²⁺have flexible packings and that both

diastereomers (∆ and Λ) have similar stabilities.

The addition of salts greatly changed the CD spectrum of

[Fe(bpyMal)₃]²⁺. For example, the addition of chloride salts

markedly enhanced the CD signal (Figs. 5-a and -b, gray

symbols). Again, negligible differences were observed between

the chloride salts of different cations (NaCl, KCl, and CaCl₂),

suggesting that Cl^-is responsible for the CD spectral responses.

The enhanced CD signal indicate that Cl^-changed the

carbohydrate packing of the Mal clusters into a more rigid

Fig. 5 (a) CD spectra of [Fe(bpyMal)₃]²⁺with and without various salts, (b)

plots of the intensity of the CD peak at the shorter wavelength (the left-most

peaks in (a)) against salt concentration, and (c) wavelength of the CD peak at

the shorter wavelength against salt concentration. [bpyMal] = 2.44 mM,

[FeCl₂] = 7.33 mM, d = 0.2 mm, rt, in water.

The negative-to-positive CD pattern of [Fe(bpyMal)₃]²⁺in

pure water was inverted, becoming positive-to-negative (when

scanned from longer to shorter wavelengths) upon addition of the

5

sulfate or nitrate salts (Figs. 5-a and -b, white and black symbols,

ACCEPTED MANUSCRIPT

respectively). This data indicates that Λ-[Fe(bpyMal)₃]²⁺became

predominant upon addition of these salts. This observation is

quite interesting since it means that different salts stabilize

different carbohydrate packings (where chlorides stabilize

different packings than sulfates and nitrates; Table 1). Again, the

anions are responsible for the salt-induced CD spectral changes.

Although both the sulfate and the nitrate salts stabilized Λ-

[Fe(bpyMal)₃]²⁺,

they showed

different

concentration

dependences. For example, in the case of the sulfate salts,

relatively low concentrations were sufficient to induce the CD

spectral inversion; the original negative-to-positive CD pattern

was changed to a positive-to-negative pattern when sulfate salt

concentrations were higher than 10–20 mM (Fig. 5-b, white

symbols). The nitrate salts also induced similar CD spectral

changes, but much higher concentrations (50–100 mM) were

required to induce the CD spectral inversion (Fig. 5-b, black

symbols), indicating that [Fe(bpyMal)₃]²⁺has a stronger affinity

-

to SO₄^2-than to NO₃.

Along with the unique CD spectral changes, peak shifts of the

CD signals were also induced upon addition of certain salts. For

example, upon the addition of sulfate salts, the positive peak at

~304 nm was clearly blue-shifted to ~295 nm (Fig. 5-c, white

symbols). Again, significant peak shifts were not observed

between the solutions containing different cations, indicating that

2-

SO₄was responsible for the blue shift. Interestingly, neither

nitrate nor chloride salts induced shifts of the CD signal peak

(Fig. 5-c, black and gray symbols, respectively).

Table 1 Diastereomeric excesses of the glycosylated tris-

bipyridine ferrous complexes induced by the different salts^a

No

salt

∆

Chloride

salts

∆∆

Sulfate

salts

ΛΛ

Nitrate

salts

[Fe(bpyMal)₃]²⁺

[Fe(bpyLac)₃]²⁺

[Fe(bpyIma)₃]²⁺

ΛΛ

∆ (∆∆)^b

∆

∆∆

∆

c

∆

ΛΛ

-

∆∆

^a∆ indicates that the ∆ isomer dominated in a low diastereomeric excess,

b

while ∆∆ indicates a amplified diastereomeric excess. ∆∆ only for NaNO₃

c

(1–300 mM). Stable diastereomers could not be clarified, since the resultant

CD spectra were entirely different from those of ∆- and Λ-[Fe(bpy)₃]²⁺.

2.4.2. Salt-induced CD spectral changes of

[Fe(bpyLac)₃] ^{2 +}

In the absence of salts, [Fe(bpyLac)₃]²⁺showed a negative-to-

positive CD pattern of weak intensity, as measured from longer

to shorter wavelengths, centered around 315 nm. Its CD spectrum

was quite similar to that of [Fe(bpyMal)₃]²⁺, indicating that the

glycoclusters of [Fe(bpyLac)₃]²⁺also have a flexible packing in

pure water.

Before beginning a discussion of the CD spectral responses of

[Fe(bpyLac)₃]²⁺towards the salts, the most conceivable

mechanism of its salt-binding behavior will first be explained. As

reported in our preceding work, [Fe(bpyLac)₃]²⁺showed unique

2-step binding profiles upon addition of chlorides (Fig. 6-b, gray

symbols). These experiments revealed that anion (Cl^-) and cation

(Na⁺, K⁺, Ca²⁺, and Mg²⁺) bindings are responsible for the 1st and

2nd CD spectral responses, respectively (see our preceding paper

for details).²⁶It was also confirmed that the Lac appendages are

responsible for the cation bindings.

Fig. 6 (a) CD spectra of [Fe(bpyLac)₃]²⁺with and without various salts, (b)

plots of the peak intensity of the CD peak at the shorter wavelength (the left-

most peaks in (a)) against salt concentration, and (c) wavelength of the CD

peak at the shorter wavelength against salt concentration. [bpyLac] = 2.44

mM, [FeCl₂] = 7.33 mM, d = 0.2 mm, rt, in water.

Interestingly, the response of [Fe(bpyLac)₃]²⁺to sulfate salts

was quite unique; aqueous solutions of [Fe(bpyLac)₃]²⁺became

turbid upon addition of small amounts of sulfate salts (10 mM),

with white precipitates observed. The white precipitates may

suggest the dissociation of [Fe(bpyLac)₃]²⁺to its precursor,

bpyLac. In contrast, no such precipitation was observed upon the

addition of chloride or nitrate salts. Furthermore, [Fe(bpyMal)₃]²⁺

and [Fe(bpyIma)₃]²⁺did not give rise to any precipitates even

6

Tetrahedron

when subjected to much higher sulfate concentrations. These

the CD spectrum having a broad and positive CD signal centered

ACCEPTED MANUSCRIPT

results clearly show that the dissociation of [Fe(bpyLac)₃]²⁺

at ~320 nm.

2-

arises from interactions between the Lac clusters and SO₄. Due

In the case of [Fe(bpyIma)₃]²⁺, most salts did not induce any

to the aggregation, the CD spectral responses of [Fe(bpyLac)₃]²⁺

were monitored only when sulfate salt concentrations were lower

than 0.1 M. The positive-to-negative CD signal was slightly

weakened upon addition of the sulfate salts (Fig. 6-b, white

symbols).

shift of its CD signal (Fig. 7-c). The one exception was Na₂SO₄,

which caused [Fe(bpyIma)₃]²⁺to exhibit a CD spectrum that

differed significantly from those of both ∆- and Λ-[Fe(bpy)₃]², at

a concentration of 1.25 mM (Fig. 7-a, line with white square

symbols).

The addition of nitrate salts also induced unique CD spectral

responses in [Fe(bpyLac)₃]²⁺. For example, additions of small

amounts of NaNO₃greatly enhanced the signal intensity until the

concentration reached 60 mM (Fig. 6-b, black square symbols).

Further addition of NaNO₃weakened the signal intensity until it

was almost identical to that obtained in pure water ([NaNO₃] =

300 mM). The enhanced CD signal observed in the presence of a

specific range of nitrate salt concentrations (from 1 to 300 mM)

was unique to NaNO₃; no such CD signal enhancement was

induced by either KNO₃nor Ca(NO₃)₂(Fig. 6-b, black circles and

triangles, respectively). These results indicate that simultaneous

-

binding of both Na⁺and NO₃to [Fe(bpyLac)₃]²⁺was achieved

under certain NaNO₃concentrations and such binding changed

the flexible Lac packing into a rigid one suitable for the ∆-

[Fe(bpy)₃]²⁺core.

The addition of chloride and nitrate salts induced blue shifts of

the CD signals. For example, the additions of small amounts of

chloride salts (~30 mM) induced blue shifts of the positive CD

signals from 302 to 298 nm, although further addition of chloride

salts shifted the CD signals back to their original wavelengths

obtained without any added salts (Fig. 6-c, gray symbols). The

nitrate salts also induced similar blue shifts, but in contrast this

blue shift was not reversed upon further addition of nitrate salts,

even to much higher concentrations (Fig. 6-c, black symbols).

2.4.3. Salt-induced CD spectral changes of

[Fe(bpyIma)₃] ^{2 +}

In the absence of salts, [Fe(bpyIma)₃]²⁺showed a CD signal

with a negative-to-positive pattern, as measured from longer to

shorter wavelengths, centered around 315 nm (Fig. 7-a, bold

line). This CD signal was similar to those of both

[Fe(bpyMal)₃]²⁺and [Fe(bpyLac)₃]²⁺, indicating its flexible

carbohydrate packings.

[Fe(bpyIma)₃]²⁺exhibited salt-induced CD spectral changes

that were entirely different from those of both [Fe(bpyMal)₃]²⁺

and [Fe(bpyLac)₃]²⁺. For example, in the case of chloride salts,

the CD signal of [Fe(bpyIma)₃]²⁺was almost unchanged until

chloride salt concentrations reached ~150 mM (Fig. 7-b, gray

symbols). However, when chloride salt concentrations were

increased beyond 150 mM, the intensities of the CD signals

weakened with increasing chloride salt concentrations. The CD

signals finally inverted to the positive-to-negative patterns when

chloride concentrations were above 300 mM (KCl and CaCl₂) or

600 mM (NaCl).

The nitrate and sulfate salts also induced CD spectral changes

in [Fe(bpyIma)₃]²⁺. For example, additions of small amounts of

nitrate salts clearly enhanced intensities of the CD signals, but

such enhancements reached plateaus when nitrate concentrations

reached ~100 mM (Fig. 7-b, black symbols). Again, the different

cations of the different nitrate salts did not induce different CD

-

spectral response of [Fe(bpyIma)₃]²⁺, indicating that NO₃is

responsible for the CD spectral response. In the case of the

sulfate salts (Na₂SO₄and K₂SO₄), the CD signal of

[Fe(bpyIma)₃]²⁺remained almost unchanged until the sulfate salts

concentrations reached ~600 mM. Further addition of Na₂SO₄

entirely changed the CD response of [Fe(bpyIma)₃]²⁺, resulting in

Fig. 7 (a) CD spectra of [Fe(bpyIma)₃]²⁺with and without various salts, (b)

plots of the peak intensity of the CD peak at the shorter wavelength (the left-

most peaks in (a)) against salt concentration, and (c) wavelength of the CD

peak at the shorter wavelength against salt concentration. [bpyIma] = 2.44

mM, [FeCl₂] = 7.33 mM, d = 0.2 mm, rt, in water.

7

2.5. Additional experiments to reveal the mechanism behind the

salt-induced CD spectral changes

to dry chemistries, including MD simulations, to reveal which

ACCEPTED MANUSCRIPT

parts of the metalloglycoclusters were responsible for the anion

binding. For example, a MD simulation of ∆-[Fe(bpyMal)₃]²⁺co-

existing with K⁺and NO₃(Discovery Studio 4.5, CHARMm, ε =

-

So far, insufficient information has been discussed to allow

clarification of the mechanism behind the salt-induced UV-vis

spectral changes. However, the fact that these spectral changes

80, 10 ps equilibrium, 90 ps dynamics) indicated that the most

stable conformation of ∆-[Fe(bpyMal)₃]²⁺captured NO₃via

-

mainly depended on the anions (Cl^-, NO₃, or SO₄) demonstrated

the substantial roles of the anions in the UV-vis spectral

responses. It was assumed that the binding of the anions to the

metalloglycoclusters would affect the conformation of the

[Fe(bpy)₃]²⁺cores, resulting in the UV-vis spectral response.

-

2-

hydrogen bonding with its amide linkages. On the contrary, K⁺

was positioned away from ∆-[Fe(bpyMal)₃]²⁺during the MD

simulation, showing their low affinity for one another. The data

obtained through the MD simulations suggest the amide linkages

as possible anion-binding sites.

On the other hand, the CD spectral changes of the

metalloglycoclusters clearly arose from the salt-induced

conformational changes of their glycoclusters. Again, the CD

spectral changes mainly depended on the anions, showing that

the anions greatly affected the spatial carbohydrate packings of

the glycoclusters. Only one exception was found in

[Fe(bpyLac)₃]²⁺; it exhibit CD spectral response to both cations

and anions.

2.6. Salt-induced changes in rheological properties of solutions

of the aqueous metalloglycoclusters

During the previously described experiments, it was noticed

that some metalloglycocluster solutions became viscous upon the

addition of certain salts. Unfortunately, the typical experiments

for assessing the rheological properties of solutions include

viscosity measurements, which usually require large sample

volumes (~10 mL) and thus this method was not appropriate for

our metalloglycocluster solutions (~100 µL). Instead, an

experiment to assess the rheological properties of the aqueous

metalloglycocluster solutions was performed as follows. Briefly,

the aqueous metalloglycocluster solutions were slowly expelled

from a syringe with a needle, and drops were weighed as they fell

from the point of the needle.

Several experiments were conducted to reveal which parts of

the metalloglycoclusters were responsible for the binding of the

anions. One such experiment using [Fe(bpy)₃]²⁺showed no UV-

vis spectral changes under high concentrations of chloride,

nitrate, or sulfate salts, indicating that the [Fe(bpy)₃]²⁺cores do

not have possible anion binding sites. Previously, we tried but

failed to perform an experiment using tris-bpy ferrous complexes

carrying glucoside (Glc) clusters or cellobioside (Cel) clusters to

clarify the roles of carbohydrate appendages in the anion binding.

Conversions of bpyGlc and bpyCel into the corresponding

ferrous complexes ([Fe(bpyGlc)₃]²⁺and [Fe(bpyCel)₃]²⁺,

respectively) failed due to the insolubility of bpyGlc and bpyCel

in water (see our preceding paper).²⁶We also synthesized other

bpy derivatives having amide linkages and phenyl spacers but no

carbohydrate units in order to reveal the anion binding sites

(Schemes S1 and S2). However, these bpy derivatives were also

insoluble in water and thus their successful conversion into the

corresponding [Fe(bpy)₃]²⁺derivatives was not accomplished.

We also tried to reveal the mechanism by which the anions

As shown in Fig. 9, changes in the weight of each drop were

negligible when salts were added to aqueous solutions of

[Fe(bpyMal)₃]²⁺or [Fe(bpyIma)₃]²⁺. On the other hand, the

weight per drop of the aqueous [Fe(bpyLac)₃]²⁺solution

increased, specifically upon addition of sulfate salts. In addition,

K₂SO₄and Na₂SO₄elicited different impacts on the weight per

drop; the former increased the weight of each drop more

significantly than the latter. This finding is quite interesting since

it indicates that not only SO₄, but also the cations (Na⁺and K⁺)

2-

affected the rheological properties of the metalloglycocluster

solutions, specifically for that having Lac appendages.

1

induce the UV-vis spectral changes through H NMR spectral

1

analyses. The H NMR spectra of the metalloglycoclusters were,

however, quite broad and thus no informative clues were

obtained.

Fig. 8 The most stable conformation of Λ-[Fe(bpyMal)₃]²⁺in co-existence

with K⁺and NO₃^-during the MD simulation, in which only the complex core,

K⁺and NO₃^-are presented in CPK models for clear presentation (See Fig. S5

for the colored version of this figure).

We were unable to obtain enough information to reveal the

mechanism behind the salt-induced CD spectral changes through

the aforementioned wet chemistries. Therefore, focus was shifted

Fig. 9 Plots of the weight per drop of the aqueous metalloglycocluster

solutions against the concentration of the co-existing salts. [bpyMal],

[bpyLac], or [bpyIma] = 2.44 mM, [FeCl₂] = 7.33 mM, n = 3, rt, in water.

8

Tetrahedron

ACCEPTED MANUSCRIPT

3. Conclusions

(spherical, neutral, particle size 63–210 µm) for column

chromatography was purchased from Kanto Chemical Co., Inc.,

Japan. Thin layer chromatography (TLC) was carried out with

Whatman TLC glass plates (Partisil^®K6F) pre-coated with silica

gel (60 Å) with a fluorescent indicator (λ = 254 nm). All other

chemicals were purchased from Wako Pure Chemicals Co., Ltd.,

Japan, Kishida Chemicals Co., Ltd., Japan, GODO Co., Japan, or

Kanto Chemical Co., Inc., Japan, and used without further

purification.

We developed [Fe(bpy)₃]²⁺derivatives having six

carbohydrate appendages as molecular mimics of cell surface

glycoclusters. Using these metalloglycoclusters, the spatial

carbohydrate packing within the glycoclusters was assessed

through UV-vis and CD spectral analyses, revealing that they

contained flexible carbohydrate packings in pure water.

In the presence of various salts, the spatial carbohydrate

packings changed, mainly in anion-dependent manners. The

anion-dependent

4.2. 2,2'-Bipyridine-5,5'-dicarboxyl chloride

conformational

changes

were

well-

demonstrated in [Fe(bpyMal)₃]²⁺and [Fe(bpyIma)₃]²⁺. This

anion-dependency was in clear contrast to the anion- and cation-

dependent conformational changes induced in [Fe(bpyLac)₃]²⁺. In

addition, it was found that the addition of sulfate salts changed

the rheological properties of an aqueous solution of

[Fe(bpyLac)₃]²⁺. These results may indicate that the salt-induced

conformational changes are linked to CCIs between the

metalloglycoclusters. However, this assumption requires further

discussion, since partial dissociation of [Fe(bpyLac)₃]²⁺was also

observed upon addition of these sulfate salts.

To 2,2-bipyridine-5,5'-dicarboxylic acid, thionyl chloride was

added and then, the resultant heterogeneous mixture was refluxed

with magnetic stirring until it became homogeneous (ca. 5 h).

The resultant reaction mixture was evaporated to dryness and

dried in vacuo to afford the title compound as a pale-yellow

powder. The product was then used for the synthesis of

bpyLacAc, bpyMalAc, bpyIma, bpyPhe, or bpyPhe(OC₂H₄)₂OH

without any purification and characterization. The amounts of

2,2-bipyridine-5,5'-dicarboxylic acid and thionyl chloride used to

synthesize each of the bpy derivatives were reported in its

synthetic procedure.

Altogether, the results indicate that the conformations of the

glycoclusters change upon addition of numerous salts, and that

carbohydrate packing may play an important role in inducing

CCIs. In studies of CCIs the ions (Ca²⁺in many cases) have been

thought of as molecular glues which connect two carbohydrate

units. The formation of multiple carbohydrate-ion-carbohydrate

complexes has been thus assumed, without clear evidence, as a

main driving force for the induction of CCIs. Based on these

proposed mechanisms, no attention has been paid to the ion-

induced conformational changes in the glycoclusters. However,

the findings in this paper suggest an alternative mechanism,

specifically that the ions change the carbohydrate packing in

glycoclusters into those more suitable for CCIs.

4.3. 1,2,3,4,2',3',4',6'-hexa-O-acetyl-isomaltose (AcIma, 1)

To isomaltose (0.61 g) in pyridine (100 ml), acetic anhydride

(50 ml) was added and the resultant mixture was stirred at

ambient temperature for overnight. The resultant mixture was

poured into ice and diluted with ethyl acetate. The organic layer

was then washed with 0.5 N HCl aq. and saturated NaHCO₃aq.

repeatedly. The organic layer was dried over anhydrous MgSO₄,

filtered, evaporated and dried in vacuo to give 1.21 of the titled

compound (AcIma) as white powder. The compound was used as

a

substrate in the following section without any

purification/identification process.

4.4. 2,3,4,2',3',4',6'-hexa-O-acetyl-isomaltosyl bromide

(AcImaBr, 2)

Although the synthesized metalloglycoclusters offered useful

insights into the spatial carbohydrate packing in the glycoclusters

and their salt-induced conformational changes, their chemical

structures differ from those of the natural GSLs on cell surfaces.

Specifically, the spacer units tethering the carbohydrate units to

the [Fe(bpy)₃]²⁺cores are phenyl rings in the synthesized

metalloglycoclusters, which are clearly different from the spacer

units tethering the carbohydrate units to the hydrophobic tails in

To AcIma in CH₂Cl₂(20 ml), acetic anhydride (6 ml) and

hydrogen bromide in acetic acid (30%, 20 ml) were added and

the resultant mixture was stirred at ambient temperature for 2.5 h.

The resultant mixture was poured into ice and diluted with ethyl

acetate. The organic layer was washed with saturated NaHCO₃

aq. repeatedly. The organic layer was then dried over anhydrous

MgSO₄, filtered, evaporated, dried in vacuo and used as a

natural GSLs. Currently, new metalloglycoclusters having

L-

serine as the spacer units are under development in our group to

better mimic the natural glycoclusters on cell surfaces and allow

evaluation of the effect of the ions on their spatial carbohydrate

packings.

substrate

of

the

following

reaction

without

any

purification/identification process.

4.5. 1-(p-Nitrophenyl)-2,3,4,2',3',4',6'-hexa-O-acetyl-

isomaltoside (pNPIma, 3)

4. Experimental Section

To AcImaBr and sodium p-nitrophenoxide (1.12 g) in dry

acetone (150 ml) was stirred at 75 ºC for overnight. The resultant

mixture was diluted with ethyl acetate and the resultant organic

layer was washed with NaCl aq. repeatedly. The organic layer

was then dried over anhydrous MgSO₄, filtered, evaporated and

dried in vacuo. The residue was then subjected to purification on

silica-gel (toluene/AcOEt = 5/1 to 1/2, v/v) to give the titled

4.1. General

¹H and ¹³C NMR spectra were acquired on a JEOL AL300

spectrometer (JEOL Resonance, Ltd., Japan) in CDCl₃, D₂O, or

DMSO-d₆. Chemical shifts were reported in ppm (δ) relative to

TMS or internal references such as HOD. Matrix-assisted laser

desorption ionization time-of-flight (MALDI-TOF) mass spectra

were recorded on a SHIMAZU AXIMA-CFR+ (SHIMAZU,

Ltd., Japan) or SHIMAZU AXIMA-Confidence (SHIMAZU,

Ltd., Japan). IR spectra were recorded on a JASCO FT/IR-4100

Fourier transform infrared spectrometer (JASCO Co., Ltd.,

Japan). UV-vis spectra were recorded on a V-630 UV-visible

spectrophotometer (JASCO Co., Ltd., Japan). Circular dichroism

1

compound (pNPIma) as white powder; H NMR (300 MHz,

CDCl₃) δ 8.33 (d, J = 9.3 Hz, 2H), 7.15 (d, J = 9.0 Hz, 2H), 5.49

(t, J = 9.9 Hz, 1H), 5.36 (t, J = 9.0 Hz, 1H), 5.31–5.20 (m, 2H),

5.09–5.00 (m, 3H), 4.88 (dd, J = 3.6, 10.2 Hz, 1H), 4.08 (dd, J =

4.5, 12.6 Hz, 1H), 4.04–3.79 (m, 4H), 3.53 (dd, J = 1.8, 9.6 Hz,

2H), 2.12 (s, 3H), 2.10 (s, 3H), 2.08 (s, 3H), 2.07 (s, 3H), 2.06 (s,

3H), 2.06 (s, 3H), 1.81 (s, 3H); ¹³C NMR (75 MHz, CDCl₃) δ

170.5, 170.2, 170.1, 170.1, 169.5, 169.4, 169.2, 161.3, 143.3,

126.2, 116.4, 98.1, 95.6, 73.3, 72.4, 70.9, 70.7, 69.5, 68.7, 68.3,

(CD) spectra were acquired on

spectropolarimeter (JASCO Co., Ltd., Japan). Silica gel 60 N

a

JASCO J-820

9

67.9, 67.6, 66.4, 61.5, 20.8, 20.7, 20.7, 20.6, 20.6, 20.1; FT-IR

(KBr, cm^–1): 1755, 1554, 1522, 1370, 1346, 1234, 1038.

MALDI-TOF-MS, [M+Na]⁺= 780.19 (calc. 780.21).

Before the titration experiments, we prepared the aqueous

ACCEPTED MANUSCRIPT

metalloglycocluster solutions and the aqueous salt solutions as

the stock solutions. In the case of the former, we mixed the

glycobpys and FeCl₂in water to prepare the aqueous

metalloglycocluster solutions ([bpyMal], [bpyLac], or [bpyIma]

= 24.4 mM, [FeCl₂] = 73.3 mM). We then mixed the

metalloglycocluster solutions (3 µl), aqueous salt solutions (0–27

µl), and pure water (27–0 µl) to prepare the sample solutions (30

µl) for the spectroscopy titration assays. After the incubation of

these sample solutions for 1-day, their UV-vis and CD spectra

were measured by using a V-630 UV-visible spectrophotometer

and a JASCO J-820 spectropolarimeter equipped with SAH-769

and MSD-462, respectively.

4.6. 1-(p-Aminophenyl)-2,3,4,2',3',4',6'-hexa-O-acetyl-

isomaltoside

To pNPIma in THF (70 ml), Pd/C (0.42 mg) was added and

the resultant mixture was vigorously stirred under H₂atmosphere

at ambient temperature for overnight. The resultant mixture was

filtrated through a celite pad and the filtrate was evaporated and

dried in vacuo. The residue was then subjected to purification on

silica-gel (CHCl₃/MeOH = 50/1, v/v) to give the titled compound

(pAPIma) as white powder. This compound was used in the

1

following reaction without H/¹³C NMR, MS and IR spectral

4.10. MD simulations

characterizations.

Before the MD simulation of the molecular assembly

4.7. 5,5’-Bis-(-N-(p-O-(2,3,6,2,3,4,6-hepta-O-acetyl-

isomaltosyl)phenyl)-aminocarbonyl)-2,2'-bipyridine

(bpyImaAc)

β-

-

consisting of Λ-[Fe(bpyMal)₃]²⁺, K⁺, and NO₃, we built a Λ-

[Fe(bpyMal)₃]²⁺with a stable conformation through a molecular

mechanics (MM) calculation. Briefly, we constructed a Λ-

[Febpy₃]²⁺and then, its stable conformation was obtained through

its MM calculation. After the construction of Λ-[Febpy₃]²⁺, the

spatial positions of all atoms in its structure were fixed. This

spatial fixing was essential to successfully achieve the MD

simulation on the three-component molecular assembly, since

CHARMm force field was not suitable for molecules having Fe-

N bonds. We then attached β-maltosyloxyphenylaminocarbonyl

(MalβPheNHCO-) units onto all C5 positions of the pyridine (py)

units of Λ-[Febpy₃]²⁺and constructed Λ-[Fe(bpyMal)₃]²⁺with E

and Z configurations about its C5_py-C_C=Obonds and amide bonds,

To pAPIma (0.37 g, 0.51 mmol) and Et₃N (2 ml) in THF (65

ml), bpyCOCl prepared from 28 mg (0.11 mmol) of bpyCOOH

in THF (5.0 ml) was added and then, the resultant mixture was

stirred at the ambient temperature for 20 min. The resultant

reaction mixture was poured into iced water and diluted with

ethyl acetate. The organic layer was then washed with 0.5 N HCl

aq. and NaHCO₃saturated aqueous solution repeatedly. The

organic layer was dried over anhydrous Na₂SO₄, filtered, and

evaporated to give pale yellow syrup. The resultant syrup was

then subjected to silica-gel column chromatographic purification

process (hexane only ~ hexane/AcOEt = 1:9, gradient) to give the

titled compound (116 mg) in 61% yield (from bpyCOOH) as a

-

respectively. We then added K⁺and NO₃ions nearby the

trivalent Mal clusters of Λ-[Fe(bpyMal)₃]²⁺to obtain the three-

component molecular assembly consisting of Λ-[Fe(bpyMal)₃]²⁺,

1

pale yellow powder.; H NMR (300 MHz, CDCl₃) δ 9.24 (d, J =

-

K⁺and NO₃. Before starting the MD simulation, we applied

1.8 Hz, 2H), 8.70 (s, 2H), 8.58 (d, J = 8.4 Hz, 2H), 8.37 (dd, J =

2.1, 8.4 Hz, 2H), 7.76 (d, J = 9.0 Hz, 4H), 7.06 (d, J = 8.7 Hz,

4H), 5.41 (t, J = 9.6 Hz, 2H), 5.34 (t, J = 9.0 Hz, 2H), 5.25 (t, J =

9.7 Hz, 2H), 5.12–4.97 (m, 8H), 4.90 (dd, J = 3.9, 9.9 Hz, 2H),

4.19–4.14 (m, 2H), 4.01–3.76 (m, 8H), 3.48 (d, J = 9.6 Hz, 2H),

1.84 (s, 6H). ¹³C NMR (75 MHz, CDCl₃) δ 170.398, 170.357,

170.160, 170.094, 169.815, 169.675, 169.165, 163.649, 157.476,

153.974, 148.005, 136.085, 132.953, 130.684, 122.11, 121.420,

117.761, 101.147, 99.142, 76.198, 72.868, 72.548, 71.479,

70.953, 70.764, 69.111, 66.612, 62.017, 60.817, 20.831, 20.724,

20.659, 20.527. FT-IR (KBr, cm^–1): 1754, 1509, 1370, 1224,

1038. MALDI-TOF-MS, [M+Na]⁺= 1685.37 (calc. 1685.49).

distance restrictions (Fe²⁺-K⁺and Fe²⁺-NO₃) to this three-

-

component molecular assembly so that the ions (K⁺and NO₃)

did not leave from Λ-[Fe(bpyMal)₃]²⁺. The subsequent MD

simulation (CHARMm, ε = 80, 10 ps equilibrium, 90 ps

dynamics) gave the most stable conformation of the three-

component molecular assembly (Fig. 8).

4.11. Dropping test of metalloglycocluster solutions

According to the procedure described in section 4.9, we

prepared the aqueous metalloglycocluster solutions containing

varying concentrations of the chloride, nitrate, or sulfate salts.

After the 1-day incubation at ambient temperature, these

solutions were slowly expelled from a plastic syringe (Tuberculin

syringe, SS-01T, TERUMO Co., Japan) with a needle whose

outer and inner diameters were 0.72 and 0.41 mm, respectively.

The drops were weighed as they fell from the point of the needle.

4.8. 5,5’-Bis-(-N-(p-O-(β-isomaltosyl)phenyl)-

aminocarbonyl)-2,2'-bipyridine (bpyIma)

To bpyImaAc (0.67 g, 0.40 mmol) in methanol/THF (1:1, 50

ml), aqueous ammonia was added dropwise until pH value of the

resultant mixture became larger than 10. The resultant reaction

mixture was stirred at ambient temperature for overnight. After

confirming that the pH value of the reaction mixture was still

larger than 10, the resultant mixture was concentrated to form

white precipitate. The precipitate was retrieved through filtration

and dried over filter paper overnight to give the title compound

as a white powder (0.37 g, >99 %); d_H(300 MHz, DMF-d₇+

D₂O): 9.11 (t, J 1.5 Hz, 2H), 8.23 (d, J 1.5 Hz, 2H), 7.61 (d, J 9.0

Hz, 4H), 6.94 (d, J 9.0 Hz, 4H), 4.66 (d, J 7.8 Hz, 2H), 4.58 (d, J

3.9 Hz, 2H), 3.64-3.01 (m, 24H); d_C(75 MHz, DMF-d₇+ D₂O):

162.9, 162.0, 156.2, 153.8, 148.2, 136.0, 132.6, 130.528, 121.2,

120.0, 116.2, 100.9, 98.2, 76.4, 74.5, 73.2, 73.1, 72.1, 71.8, 70.0,

79.8, 66.0, 60.7; HR-ESI-TOF-MS, [M+H]⁺= 1075.3550 (calc.

1075.3519); n_max(KBr) 3375, 1647, 1542, 1508, 1417, 1224,

1078 cm^-1.

Declaration of interest

Declarations of interest: none

Acknowledgements

This work was supported by Grant-in-Aid for Scientific

Research (No. 25810105) from the Ministry of Education,

Culture, Sports, Science and Technology, Japan and Special

Research Fund of Toyo University.

References and notes

1. Hirohashi N, Kamei N, Kubo H, Sawada H, Matsumoto M, Hoshi

M. Dev. Growth Differ. 2008;50:S221–S238.

2. Schengrund CL. Trends Biochem. Sci. 2015;40:397–406.

4.9. UV-vis and CD spectroscopy titration experiments

10

Tetrahedron

3. Nakayama H, Ogawa H, Takamori K, Iwabuchi K. Arch.

Immunol. Ther. Exp. 2013;61;217–228.

21. Lai CH, Hütter J, Hsu CW, Tanaka H, Varela-Aramburu A, Cola

ACCEPTED MANUSCRIPT

LD, Lepenies B, Seeberger PH. Nano Lett. 2016;16:807–811.

22. Vilanova E, Santos GRC, Aquino RS, Valle-Delgado JJ,

Anselmetti D, Fernàndez-Busquets X, and Mourão PAS. J. Biol.

Chem. 2016;291:9425-9437.

23. Otsuka A, Sakurai K, Hasegawa T. Chem. Commun.

2009;36:5442–5444.

4. Fenderson BA, Eddy EM, Hakomori S. BioEssays 1990;12:173–

179.

5. Hakomori S. Pure Appl. Chem. 1991;63:473–482.

6. Iwabuchi K, Yamamura S, Prinetti A, Handa K, Hakomori S. J.

Biol. Chem. 1998;273:9130–9138.

7. Delbianco M, Bharate P, Varela-Aramburu S, Seeberger PH.

Chem. Rev. 2016;116:1693–1752.

24. Iwamura M, Koyama R, Nonaka T, Dai F, Matsuoka R, Nakamura

M, Iwabuchi H, Okada T, Hasegawa T. Bull. Chem. Soc. Jpn.

2016;89:617–625.

25. Nakamura M, Tsutsumi M, Ishikawa Y, Umemiya H, Hasegawa

T, Izawa K, Abe H, Togashi Y, Kinone T, Sekiguchi S, Igumi M,

Ide K, Hasegawa T. Tetrahedron 2013;69:3019–3026.

26. Nonaka Y, Uruno R, Dai F, Matsuoka R, Nakamura M, Iwamura

M, Iwabuchi H, Okada T, Chigira N, Amano Y, Hasegawa T.

Tetrahedron 2016;72:5456–5464.

8. Matsuura K, Kobayashi K. Glycoconjugate J. 2004;21:139–148.

9. Hasegawa T, Sasaki T. Chem. Commun. 2003;8:978–979.

10. Altamore TM, Fernández-García C, Gordon AH, Hübscher T,

Promsawan N, Ryadnov MG, Doig AJ, Woolfson DN, Gallagher

T. Angew. Chem. Int. Ed. 2011;50:11167–11171.

11. Simpson GL, Gordon AH, Lindsay DM, Promsawan N, Crump

MP, Mulholland K, Hayter BR, Gallagher T. J. Am. Chem. Soc.

2006;128:10638–10639.

27. Milder SJ, Gold JS, Kliger DS. J. Am. Chem. Soc.

1986;108:8295–8296.

12. Bavireddi H, Bharate P, Kikkeri R. Chem. Commun.

2013;49:3988–3990.

13. Lorentz B, de Cienfuegos LÁ, Oelkers M, Krienmen E, Brand C,

Stephan M, Sunnick E, Yüksel D, Kalsani V, Kumar K, Werz DB,

Janshoff A. J. Am. Chem. Soc. 2012;134:3326–3329.

14. Kunze A, Bally M, Hoeoek F, Larson G. Sci. Rep. 2013;3:1452–

1459.

15. Jayaraman N, Maiti K, Naresh K. Chem. Soc. Rev. 2013;42:4640–

4656.

16. de la Fuente, JM, Eaton P, Barrientos AG, Menéndez M, Penadés

S. J. Am. Chem. Soc. 2005;127:6192–6197.

Supplementary Materials

Supplementary data related to this article containing length of

each side of the triangles whose corners are occupied by C4

atoms of the non-reducing carbohydrate terminals (Fig. S1),

schematic illustration of our metalloglycoclusters for

investigating stable carbohydrate packings (Fig. S2), a colored

version of Fig. 1 (Fig. S3), representative UV-vis spectral

changes of the metalloglycoclusters (Fig. S4), a colored version

of Fig. 8 (Fig. S5), synthetic routes to access the bpy derivatives

17. Reynolds AJ, Haines AH, Russell DA. Langmuir 2006;22:1156–

1163

.

18. Zhao J, Liu Y, Park HJ, Boggs JM, Basu A. Bioconjugate Chem.

2012;23:1166–1173.

19. Seah N, Santacrose PV, Basu A. Org. Lett. 2009;11:559–562.

20. Chen Q, Cui Y, Zhang TL, Cao J, Han BH. Biomacromolecules,

2010;11:13–19.

(Schemes S1 and S2), and H/¹³C NMR spectra of the newly

synthesized compounds in this paper are available online free of

charge.

1

ACCEPTED MANUSCRIPT

Article Doi

Glycosylated tris-bipyridine ferrous complexes for probing a mechanism behind carbohydrate-carbohydrate interactions: Spatial carbohydrate packing of glycoclusters changes on additions of salts in carbohydrate- and anion-dependent manners

DOI: 10.1016/j.tet.2018.08.017

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.tet.2018.08.017

Full text of DOI:10.1016/j.tet.2018.08.017

Products guided by the article

R&D Labs maybe for 152387-95-6

Relevant to this article

Hot Product