Preferential binding of E. coli with type 1 fimbria to D-mannobiose with the Manα1→2Man structure

Article abstract of DOI:10.1080/09168451.2015.1075863

Manα1→2Man, Manα1→3Man, Manα1→4Man, and Manα1→6Man were converted to the glycosylamine derivatives. Then, they were mixed with monobenzyl succinic acid to obtain their amide derivatives. After removing the benzyl group by hydrogenation, the succinylamide derivatives were coupled with the hydrazino groups on BlotGlyco beads in the presence of water-soluble carbodiimide. D-Mannobiose-linked beads were incubated with fluorescence-labeled Escherichia coli with type 1 fimbria, and the number of the fluorescent dots associated with the beads was counted in order to determine the binding preference among D-mannobiose isomers. The results showed that the bacteria bind strongly to Manα1→2Man1→beads, Man- α1→3Man1→beads, Manα1→4Man1→beads, and Manα1→6Man1→beads, in order. In the presence of 0.1 M methyl α-D-mannopyranoside, most of the bacteria failed to bind to these beads. These results indicate that E. coli with type 1 fimbria binds to all types of D-mannobiose isomers but preferentially to Manα1→2Man disaccharide.

Full text of DOI:10.1080/09168451.2015.1075863

Bioscience, Biotechnology, and Biochemistry
ISSN: 0916-8451 (Print) 1347-6947 (Online) Journal homepage: http://www.tandfonline.com/loi/tbbb20
Preferential binding of E. coli with type 1 fimbria to
d-mannobiose with the Manα1→2Man structure
Katsumi Ajisaka, Kaoru Yuki, Kaori Sato, Nozomi Ishii, Ichiro Matsuo, Ryo
Kuji, Tatsuo Miyazaki & Kiyoshi Furukawa
To cite this article: Katsumi Ajisaka, Kaoru Yuki, Kaori Sato, Nozomi Ishii, Ichiro Matsuo, Ryo
Kuji, Tatsuo Miyazaki & Kiyoshi Furukawa (2016) Preferential binding of E. coli with type 1
fimbria to d-mannobiose with the Manα1→2Man structure, Bioscience, Biotechnology, and
Biochemistry, 80:1, 128-134, DOI: 10.1080/09168451.2015.1075863
To link to this article: http://dx.doi.org/10.1080/09168451.2015.1075863
Published online: 19 Aug 2015.
Submit your article to this journal
Article views: 21
View related articles
View Crossmark data
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=tbbb20
Download by: [University of Nebraska, Lincoln]
Date: 04 December 2015, At: 15:08

Bioscience, Biotechnology, and Biochemistry, 2016
Vol. 80, No. 1, 128–134
Preferential binding of E. coli with type 1 fimbria to D-mannobiose with the
Manα1→2Man structure
Katsumi Ajisaka1, Kaoru Yuki2, Kaori Sato2, Nozomi Ishii3, Ichiro Matsuo3, Ryo Kuji2,
Tatsuo Miyazaki1 and Kiyoshi Furukawa2,*
1Department of Food Science, Niigata University of Pharmacy and Applied Life Sciences, Niigata, Japan;
3
2Department of Bioengineering, Nagaoka University of Technology, Nagaoka, Japan; Division of Molecular
Sciences, Faculty of Science and Technology, Gunma University, Kiryuu, Japan
Received May 14, 2015; accepted July 15, 2015
http://dx.doi.org/10.1080/09168451.2015.1075863
Manα1→2Man, Manα1→3Man, Manα1→4Man,
and Manα1→6Man were converted to the glycosy-
lamine derivatives. Then, they were mixed with
monobenzyl succinic acid to obtain their amide
derivatives. After removing the benzyl group by
hydrogenation, the succinylamide derivatives were
coupled with the hydrazino groups on BlotGlyco™
beads in the presence of water-soluble carbodiimide.
D-Mannobiose-linked beads were incubated with
fluorescence-labeled Escherichia coli with type 1 fim-
bria, and the number of the fluorescent dots associ-
ated with the beads was counted in order to
determine the binding preference among D-manno-
biose isomers. The results showed that the bacteria
bind strongly to Manα1→2Man1→beads, Man-
α1→3Man1→beads, Manα1→4Man1→beads, and
Manα1→6Man1→beads, in order. In the presence
of 0.1 M methyl α-D-mannopyranoside, most of the
bacteria failed to bind to these beads. These results
indicate that E. coli with type 1 fimbria binds to all
types of D-mannobiose isomers but preferentially to
Manα1→2Man disaccharide.
Furthermore, Escherichia. coli binds high mannose-
type glycans rich in intestinal mucosa and glycans with
the Galα1→4Gal group in urinary tract depending on
the strains, and the pathogenic strains often cause
enterohemorrhagy with verotoxin produced, hemolytic
uremic syndrome, acute encephalopathy, pyelonephritis,
cystitis, and urethritis for humans.4–8) E. coli is a gram-
negative bacterium and covered with fimbria. At the
top of the fimbria, there are glycan-binding proteins
called adhesins. The most well characterized bacterial
adhesin is the type 1 fimbrial FimH adhesin which
recognizes α-D-mannose.9) FimH is displayed on the
bacterial surface as a component of the type 1 fimbrial
organelle.
Using beads immobilized with α-D-mannose-contain-
ing glycans, we found E. coli binds strongly to high
mannose-type glycans and wondered how many α-D-
mannose residues are required for the recognition by
the FimH adhesin. Then, we prepared beads immobi-
lized with D-mannose or D-mannobiose with
Manα1→2-Man structure, examined the binding activ-
ity of E. coli (strain W3110) to these beads, and found
that the bacteria bind stronger to D-mannobiose-linked
beads than α-D-mannose-linked beads. However, in the
case of D-mannobiose, there are four isomers with dif-
ferent glycosidic linkages. In this study, we examined
the binding preference of E. coli among D-mannobiose
Key words: E. coli; carbohydrates; D-mannobiose;
binding preference
isomers
with
Manα1→2Man,
Manα1→3Man,
Many microorganisms recognize cell surface glycans
as receptors for cell attachment and tissue colonization.
This is because viruses, bacteria, and protozoa express
a large number of glycan-binding proteins so-called
Manα1→4Man, and Manα1→6Man structures.
1
Materials and methods
lectins (reviewed in Ref. ). For example, human influ-
enza viruses recognize the Siaα2→6Gal structure by
Reagents. All the chemical reagents used for the
preparation of sugar-linked beads were from Wako Pure
Chemicals Co. (Osaka). D-Mannobiose isomers,
Manα1→2Man, Manα1→3Man, and Manα1→6Man
were synthesized by HCl-catalyzed condensation reac-
tion according to the method described previously.10)
Manα1→4Man was obtained from Carbosynth Ltd.
their hemagglutinins expressed on virus surface, and
Helicobacter
pylori
binds
Fucα1→2Galβ1→3
(Fucα1→4) GlcNAc1β→R structure (Leb antigen) by
its adhesin BabA expressed on flagella.2,3) In fact, such
glycans are expressed on cell surfaces of human
respiratory tract and stomach tissues, respectively.
*Corresponding author. Email: furukawa@vos.nagaokaut.ac.jp
Abbreviations: E. coli, Escherichia coli; HPLC, high-performance liquid chromatography; HRMS, high-resolution mass spectrometry; PBS, 10 mM
phosphate-buffered saline (pH 7.4); WSC, water-soluble carbodiimide.
© 2015 Japan Society for Bioscience, Biotechnology, and Agrochemistry

Binding of E. coli to D-mannobiose
129
(Berkshire, UK). A PKH67 green fluorescent cell linker
kit for general cell membrane labeling was purchased
from Sigma-Aldrich Corp. (St. Louis, MO).
GP™ column (1 × 25 cm) (Kanto Chemical Co., Inc.,
Tokyo). Fractions eluted at the retention time from 12 to
14 min were collected (the retention times were different
among D-mannobiose isomers), and the collected frac-
tions were concentrated to obtain a pure D-mannobiosyl
benzylsuccinylamide derivative (3, 9.8 mg, 18 μmol).
The compound 3 (7.1 mg, 13 μmol) was dissolved in
distilled water (2 mL), and PD-C (10 mg) was added to
the solution. It was then placed into a flask attached
with a H2 balloon and stirred at room temperature for
3 h. After completing the reaction, the catalyst was
separated by centrifugation at 2500 × g for 3 min. By
concentrating the supernatant, D-mannobiosyl succiny-
lamide derivative (4) was obtained quantitatively. The
1H-NHR and 13C-NMR spectra and high-resolution
mass spectrometry (HRMS) spectra of D-mannobiosyl
succinylamide derivatives were obtained with a Jeol
JNM-EC600 spectrometer (JEOL Ltd., Tokyo) and a
micrOTOF II spectrometer (Bruker Daltonics Inc.,
Billerica, MA), respectively.
Synthesis of D-mannobiosyl succinylamide deriva-
tives. Since the method for preparing D-mannobiose-
linked beads is the same among the isomers, the
preparation of Manα1→2-Man1→beads was described
here as a representative. The amination of the D-manno-
biose was performed in a saturated (NH4)2CO3 solution
by the method with
a
slight modification11) as
described in Fig. 1. A disaccharide, Manα1→2Man, (1,
51.5 mg, 0.15 mmol) was dissolved in a saturated
(NH4)2CO3 solution (5 mL), and the mixture was stirred
at 40 °C for 4 h. After confirming the disappearance of a
starting material spot (Rf: 0.43) by thin layer chro-
matography (CH3CN/H2O = 3:1 (v/v) containing 0.1%
triethylamine), the solution was subjected to evaporation
initially at 45 mm Hg and then at 12 mm Hg.
Monobenzyl succinic acid was synthesized by mix-
ing benzyl alcohol (1.0 mL, 9.6 mmol) and succinic
anhydride (1.2 g, 1.2 mmol) in pyridine (2.0 mL) at
room temperature for 1 h. The concentrated reaction
mixture was dissolved in ethyl acetate, washed with a
1 M HCl solution, brine (a saturated aqueous NaCl
solution), a saturated aqueous NaHCO3 solution and
brine, dried in the presence of Na2SO4, and then con-
centrated in vacuo to obtain monobenzyl succinic acid
as a pure compound. All the concentrated D-mannobio-
syl amine syrup (1) was dissolved in dimethyl for-
mamide (5 mL), to which monobenzyl succinic acid
(46.8 mg, 0.225 mmol) and 4-(4,6-dimethoxy-1,3,5-tri-
azin-2-yl)-4-methylmorpholinium chloride n-hydrate12)
(61.2 mg, 0.225 mmol) were added. The mixture was
stirred at 40 °C for 22 h and then concentrated in
vacuo to obtain syrup (34.5 mg). The syrup was puri-
fied by chromatography using a C-18 Sep-Pak™ col-
umn (Nihon Waters KK, Japan) and a Mightysil RP-18
Analytical data of Manα1→2Man succinylamide
derivative.
Rf: 0.28 (CH3CN/H2O = 3:1). 1H-NMR
(600 MHz, D2O): δH 5.57 (0.18H, d, J = 1.91 Hz, H-
1α), 5.15 (0.82H, bs, H-1β), 5.13 (0.16H, s, H-1α′),
4.97 (0.84H, d, J = 1.81 Hz, H-1β′), 4.00 (1H, dd, J =
1.92, 3.05 Hz, H-2), 3.93 (1H, bd, J = 3.04 Hz, H-2′),
3.81 (1H, dd, J = 1.21, 11.98 Hz, H-6), 3.78 (1H, dd, J
= 3.08, 9.51 Hz, H-3′), 3.89 (1H, bd, J = 2.81,
11.83 Hz, H-6), 3.70 (1H, dd, J = 3.08, 9.66 Hz, H-3),
3.62 (1H, dd, J = 6.81, 12.21, H-6′), 3.59 (1H, dd, J =
5.85, 12.01 Hz, H-6′) 3.53 (1H, t, J = 9.461 Hz, H-4),
3.51 (1H, t, J = 9.59 Hz, H-4), 3.34 (1H, ddd, J = 2.02,
5.97, 9.67 Hz, H-5), 2.57–2.49 (4H, m, –CH2–CH2–),
13C-NMR (150 MHz, D2O): δC 177.53, 175.49 (C=O),
103.10 (C-1′), 81.25, 78.91 (C-1α), 78.58 (C-1β), 78.33
(C-3), 77.75, 74.44, 73.89 (C-5), 70.62 (C-2), 67,39,
67.02, 61.45 (C-6), 61.40, 30.43, 29.51 (–CH2–CH2–).
Fig. 1. Synthetic pathways of BlotGlyco™ beads linked with D-mannobiose derivative (one of R1–R4 is D-mannose, and the others are H’s).

130
K. Ajisaka et al.
HRMS m/z calcd. for C16H26NO13 [M-H+] 440.1399,
found 440.1390.
lization was conducted by rotation of the mixture at
40 °C for 48 h, and then, the suspension was cen-
trifuged at 2500 × g for 30 s. The beads were washed
with water (300 µL × 3 times) and methanol
(500 µL × 3 times) and finally dried in vacuo to obtain
D-mannobiose-linked BlotGlyco™ beads (5) with keep-
ing its cyclic conformation of D-mannose residue at the
reducing end (Fig. 1).
Analytical data of Manα1→3Man succinylamide
derivative.
Rf: 0.27 (CH3CN/H2O = 3:1). 1H-NMR
(600 MHz, D2O): δH 5.28 (0.09H, d, J = 1.95 Hz, H-
1α), 5.13 (0.91H, s, H-1β), 5.10 (0.09H, s, H-1α′), 4.97
(0.91H, d, J = 1.38 Hz, H-1β′), 3.95 (1H bd, J =
3.15 Hz, H-2′), 3.93 (1H, dd, J = 1.48, 3.06 Hz, H-2),
3.77–3.72 (3H, m, H-6, H-6, H-3), 3.70 (1H, dd, J =
3.32, 9.47 Hz, H-3), 3.67–3.56 (4H, m, H-6′, H-6′, H-
5′, H-4), 3.51 (1H, t, J = 9.66 Hz, H4), 3.36 (1H, ddd,
J = 2.00, 5.53, 9.56 Hz, H-5), 2.48 (4H, bs, –CH2–
CH2–), 13C-NMR (150 MHz, D2O): δC 178.78, 176.14
(C=O), 102.56 (C-1′), 80.86, 78.19, 78.23, 78.23,
74.44, 73.92, 70.91, 70.57, 70.36, 67,36, 66.52, 61.55
(C-6), 61.30 (C-6′), 31.12, 30.43 (–CH2–CH2–). HRMS
m/z calcd. for C16H26NO13 [M-H+] 440.1399, found
440.1393.
Determination of amounts of D-mannose linked to
beads.
The amount of D-mannobiose linked to
BlotGlyco™ beads was determined as follows. The
disaccharide-linked beads (5, 0.9–2.3 mg) were placed
into 4 M CF3COOH (300 µL) in an Eppendorf tube.
After incubating the tube at 100 oC for 3 h, N2 gas was
flashed into the tube for evaporation of CF3COOH and
water. D-Mannose residues released were dissolved in
distilled water (200 µL), and the sample was filtered
through a Millex filter (0.45 μm) (Merk Millipore
Corp., Billerica, MA) at 700 × g for 30 s. The amount
of D-mannose was determined by high-performance
liquid chromatography (HPLC) equipped with an evap-
orative light scattering detector using an Asahi-Pak
NH2-P50™ column (Showa Denko K.K., Tokyo).
Analytical data of Manα1→4Man succinylamide
derivative.
Rf: 0.25 (CH3CN/H2O = 3:1). 1H-NMR
(600 MHz, D2O): δH 5.31 (0.12H, d, J = 1.70 Hz,
H-1α), 5.14 (0.88H, d, J = 1.46 Hz, H-1β), 5.10 (1H, d,
J = 0.85 Hz, H-1β′), 3.94 (1H, dd, J = 1.70, 2.97 Hz,
H-2), 3.79 (1H, dd, J = 0.81, 3.09 Hz, H-2′), 3.78–3.74
(2H, m, H-6, H-6), 3.71–3.60 (4H, m), 3.59–3.52 (3H,
m), 3.43 (1H, ddd, J = 1.95, 5.10, 9.40 Hz, H-5), 2.59–
2.50 (4H, m, CH2–CH2–), 13C-NMR (150 MHz, D2O)
δC 177.17, 175.42 (C=O), 101.57 (C-1′), 80.86, 78.37,
76.83, 74.41, 74.27, 71.23, 70.87, 67,12, 61.54 (C-6),
61.48 (C-6′), 30.04, 28.91 (–CH2–CH2–). HRMS m/z
calcd. for C16H26NO13 [M-H+] 440.1399, found
440.1400.
Culture and labeling with a fluorescent dye of
bacteria.
E. coli (strain W3110)14) donated by Dr.
M. Fukuda of Nagaoka University of Technology was
grown in a Luria–Bertani medium at 37 °C for 14–20 h
by shaking. After centrifugation, the bacteria were sus-
pended in Diluent C (1 mL) from the cell membrane
labeling kit, mixed with 10 µL of PKH67 solution
(4 µL of PKH67 mixed with 996 µL of Diluent C),
and incubated at 37 °C for 10 min to label the bacteria
with the fluorescent dye according to the manufac-
turer’s protocol. In this study, more than 90% of
E. coli was labeled with the dye and alive during the
period of experiments, which was confirmed by deter-
mining the growth rate of the fluorescence-labeled E
coli in culture. Higher percentages of cells labeled with
this dye together with a longer survival time of the
dye-labeled cells were reported.15)
Analytical data of Manα1→6Man succinylamide
derivative.
Rf: 0.25 (CH3CN/H2O = 3:1). 1H-NMR
(600 MHz, D2O): δH 5.31 (0.11H, d, J = 1.70 Hz, H-
1α), 5.11 (0.89H, d, J = 0.97 Hz, H-1β), 4.77 (0.88H,
d, J = 1.59 Hz, H-1β′), 4.75 (012H, d, J = 1.64 Hz, H-
1α′), 3.86 (1H, dd, J = 1.62, 3.98 Hz, H-2), 3.83 (1H,
dd, J = 1.02, 2.98 Hz, H-2′), 3.81 (1H, dd, J = 5.13,
11.46 Hz, H-6), 3.75 (1H, dd, J = 2.04, J = 12.17 Hz,
H-6), 3.71 (1H, dd, J = 3.18, 8.93 Hz, H-3′), 3.63 (1H,
dd, J = 5.56, 12.04, H-6′), 3.59 (1H, dd, J = 5.47,
11.64, H-6′), 3.58 (1H, dd, J = 3.92, 9.58, H-3), 3.58–
3.51 (3H, m, H-4, H-4′, H-5′), 3.47 (1H, ddd, J = 1.95,
5.10, 9.40 Hz, H-5), 2.59–2.52 (4H, m, –CH2–
CH2–),13C-NMR (150 MHz, D2O) δC 177.66, 175.79
(C=O), 100.10 (C-1′), 78.27, 76.34, 74.01, 73.27,
71.04, 74.01, 70.04, 67,23, 66.89, 66.28 (C-6), 61.44
(C-6′), 30.55, 29.40 (–CH2–CH2–). HRMS m/z calcd.
for C16H26NO13 [M-H+] 440.1399, found 440.1394.
Assay for binding of bacteria to D-mannobiose-linked
beads.
Fluorescence-labeled E. coli suspended in a
Luria-Bertani medium (OD600 = 0.1, 200 µL) was
mixed with 1 mg of D-mannobiose-linked beads sus-
pended in 40 µL of 10 mM phosphate-buffered saline
(pH 7.4) (PBS) in a glass-bottom dish and incubated at
37 °C for 1–2 h. Then, beads were washed with PBS
(200 µL) three times and suspended in PBS (200 µL)
for observation under a confocal laser microscope. In
some experiments, the bacteria and the beads were
incubated in the presence of 0.1 M methyl
α-D-mannopyranoside.
Immobilization of D-mannobiose isomers on beads.
To D-mannobiosyl succinylamide derivative (4, 2 mg,
4.5 µmol) dissolved in distilled water (300 μL), 8 mg
of BlotGlyco™ beads (Sumitomo Bakelite Co. Ltd.,
Tokyo)13) and water-soluble carbodiimide (WSC)
(86.3 mg, 450 µmol) was added (Fig. 1). The immobi-
Results and discussion
Preparation of D-mannobiose isomer-linked beads
In order to determine the binding preference of
E. coli with type 1 fimbria among D-mannobiose

Binding of E. coli to D-mannobiose
131
isomers,
Manα1→2Man,
Manα1→3Man,
and
proton signal was assigned to be axial, which indicates
that the succinylamide group was conjugated in a
β-linkage to the D-mannose residue. Contrary to the
empirical rule in the glycosylation of D-mannose, the
amount of the β-linked isomer was more than that of
the α-linked isomer, and furthermore, the coupling con-
stant between H-1α and H-2 was larger than that
between H-1β and H-2 in all the derivatives. For all
the succinylamide derivatives of D-mannobiose isomers,
the percentages of the β-isomers were 82–93% by
calculating the peak areas of the proton signals at indi-
vidual reducing ends. The α-/β-isomer ratios obtained
Manα1→6-Man were synthesized by HCl-catalyzed
condensation reaction,10) and Manα1→4-Man were
commercially obtained. Individual D-mannobiose iso-
mers were mixed with monobenzyl succinic acid in the
presence of 4-(4,6-dimethoxy-1,3,5-triazin–2-yl)–4-
methylmorpholinium chloride n-hydrate, and then, the
benzyl group was removed by hydrogenolysis to obtain
D-mannobiosyl
succinylamide
derivatives.
The
1H-NMR spectra of D-mannobiosyl succinylamide
derivatives (Fig. 2(a–d)) showed two sets of signals
derived from α-anomer (1α) and β-anomer (1β) at the
reducing end and at the non-reducing end, the latter of
which is indicated with an asterisk shown in Fig. 2.
The assignment of H-1 signals was performed by
nuclear Overhauser effect techniques in NMR spec-
troscopy. The H-1 signal giving a cross-peak with H-5
1
by H-NMR and HPLC analyses are shown in Table 1.
D-Mannobiosyl succinylamide derivatives were
mixed with BlotGlyco™ beads. The immobilization of
them on the beads was conducted according to the
reaction described in Fig. 1. Although the protocol of
Fig. 2. 600 MHz 1H-NMR spectra of D-mannobiosyl succinyl-amide derivatives. Panels (a), (b), (c), and (d) indicate Manα1→2Man1-,
Man-α1→3Man1-, Manα1→4-Man-1-, and Manα1→6Man1-succinyl-amide derivatives, respectively. The 1α and 1β indicate H-1α and H-1β
proton signals at the reducing end, respectively. *, a signal from the H-1 proton at the non-reducing end.

132
K. Ajisaka et al.
Table 1. The α/β percentages of D-mannobiosyl succinylamide
derivative 4’s.
When the numbers of the fluorescent dots associated
with other types of the beads were counted, the average
numbers of them were 26 dots/bead in the case of
Manα1→3-Man-1→bead (Fig. 3(b)), 16 dots/bead in
the case of Manα1→4-Man1→bead (Fig. 3(c)), and 10
dots/bead in the case of Manα1→6Man1→bead
(Fig. 3(d)), respectively. The binding assay was con-
ducted three times, and similar binding preference to
those shown in Fig. 3 was obtained. When the bacteria
were incubated with Manα1→2Man-1→beads in the
presence of 0.1 M methyl α-D-mannopyranoside, most
of them failed to bind to the beads (Fig. 3(e)). The
same results as that shown in Fig. 3(e) were obtained
for E. coli using Manα1→3Man1→beads, Manα1→4-
Man-1→beads, and Manα1→6Man1→beads, respec-
tively, in the presence of the haptenic sugar (data not
shown). These results indicate that E. coli with type 1
fimbria binds to D-mannobiose-linked beads through its
type 1 adhesion and that the bacteria have a preference
to bind to Manα1→2Man disaccharide among the iso-
mers. There are many fluorescent dots not bound to the
beads in Fig. 3. They can be removed from the reaction
mixtures by harsh washings with PBS. However, such
washings also take the beads out and reduce the num-
ber of the beads to be observed. Another approach to
reduce the background fluorescent dots is to decrease
the number of E. coli mixed with the beads. However,
in this case, it also decreases the number of the fluores-
cent dots associated with the beads. Therefore, the pre-
sent assay condition is suitable for determining the
binding preference of E. coli among D-mannobiose iso-
mers. Furthermore, once the bacteria bind to these
beads, they are not released from the beads under the
present condition unless the haptenic sugar is included
in the reaction mixture. Therefore, the fluorescent dots
in the background are not from those released from the
beads.
High mannose-type glycans attached to proteins con-
tain Manα1→2Man disaccharides at the non-reducing
termini,16) and E. coli may bind to these high man-
nose-type glycans expressed in the large intestine
where they reside. E. coli is quite often contaminated
in water for domestic use to cause us serious enterohe-
morrhagic diseases.8) If some filtration systems
immobilized with the D-mannobiose or high mannose-
type glycans are developed and applied to a water sup-
ply system to clean, they can trap specifically E. coli
with type 1 fimbria in water and provide us with safe
water. In some underdeveloped countries, local resi-
dents are still facing to serious infectious diseases with
E. coli contaminated in water for domestic use. It has
been reported that around 100 bacteria are contami-
nated in 1 mL of water.17) Our preliminary study
NMRa
HPLCb
Sugar part of compound 4
α
β
α
β
Manα1→2Man
Manα1→3Man
Manα1→4Man
Manα1→6Man
18
9
12
11
82
91
88
89
11
10
13
10
89
90
87
90
aDetermined from integral intensity of each H-1 signal in NMR spectra.
bDetermined from height of each peak in HPLC chart.
BlotGlyco™ beads shows the coupling between a
hydrazino group on the bead and an aldehyde group of
the reducing end of glycan,13) the present method was
to couple the carboxyl group of D-mannobiosyl suc-
cinylamide derivatives to a hydrazino group on the
bead with an aid of WSC to establish a hydrazide link-
age. As the reaction was to hydrazino groups on the
bead, a 100-fold higher amount of WSC than that of
bead was used. The amount of D-mannobiose isomers
linked to BlotGlyco™ beads was determined, which is
described in detail in the materials and methods sec-
tion, and 0.11–0.16 µmol of D-mannose was linked to
1 mg of the beads (Table 2). As the amount of
hydrazino group expressed on the bead is shown to be
1–2 μmol/mg bead according to the manufacturer’s
protocol,13) the obtained values in Table 2 appeared
reasonable.
Binding of E. coli to D-mannobiose isomer-linked
beads
E. coli with type 1 fimbria is shown to bind to α-D-
mannose-containing glycans.4,5,9) Our preliminary study
showed that the bacteria bind to D-mannobiose-linked
beads stronger than D-mannose-linked beads. In order
to examine whether E. coli shows binding preference
to a D-mannobiose isomer with a certain structure, the
fluorescence-labeled bacteria were incubated with
Manα1→2Man-1→beads, Manα1→3-Man1→beads,
Manα1→4Man1→beads, or Manα1→6Man-1→beads,
and the reaction mixtures were observed under a confo-
cal laser microscope for counting the numbers of the
dots associated with the circumferences of the beads.
The bacteria bound to all types of D-mannobiose
isomer-linked beads (Fig. 3). In the case of Manα-
1→2-Man1→beads, the average number of the dots
associated with the beads was 52 dots/bead, and more
dots whose numbers were unable to be counted were
associated with the two beads shown in Fig. 3(a).
Table 2. Amounts of D-mannose liberated from sugar-linked BlotGlyco™ beads.
D-mannose on the beads
Beadsa (mg)
D-mannoseb (µg/200 µL)
(μg/mg beads)
(μmol/mg beads)
Manα1→2Man
Manα1→3Man
Manα1→4Man
Manα1→6Man
0.577
0.576
0.569
0.570
32.2
20.2
44.0
28.6
56
37
42
50
0.16
0.11
0.12
0.15
aWeights of sugar-linked BlotGlyco™ beads used for the analysis.
bAmounts of D-mannose from BlotGlyco™ beads determined by HPLC analysis.

Binding of E. coli to D-mannobiose
133
Fig. 3. Incubation of fluorescence-labeled E. coli with sugar-linked beads. Panels (a), (b), (c), and (d) indicate the bacteria with Manα1→2-
Man1→beads, Manα1→3Man-1→beads, Man-α1→4Man-1→beads, and Manα1→6Man1→beads, respectively, and panel (e) indicates the bacte-
ria with Manα1→2Man1→beads in the presence of 0.1 M methyl α-D-mannopyranoside.
showed that 1 mg of beads immobilized with high
mannose-type glycans can trap at least a million of
Disclosure statement
E. coli. If such a tool immobilized with Manα1→2Man
No potential conflict of interest was reported by the authors.
disaccharide or high mannose-type glycans is used in
these countries, it can trap the bacteria in water for
domestic use and provide the local residents with safe
water.
Funding
This work was supported by the MEXT of Japan under a Grant-
in-Aid for Scientific Research [grant number 15K12314] to KF; Grant
of Strategic Research Foundation Grant-Aided Project for Private
Universities [grant number S1001030] to KA and TM, and by the
Japan Science and Technology Agency under a Research Promotion
Fund to KF.
Author contributions
KA and KF designed the research and wrote the
paper. KY, KS, NI, IM, RK, and TM performed the
experiments.

134
K. Ajisaka et al.
[9] Klemm P, Schembri MA. Bacterial adhesins: function and
References
structure. Int. J. Med. Microbiol. 2000;290:27–35.
[10] Ajisaka K, Yagura M, Miyazaki T. A novel two-step synthesis
of α-linked mannobioses based on an acid-assisted reverse
hydrolysis reaction. Carbohydr. Res. 2012;347:147–150.
[11] Likhosherstov LM, Novikova O, Derevitskaja VA, Kochetkov
NK. A new simple synthesis of amino sugar β-D-glycosylami-
nes. Carbohydr. Res. 1986;146:C1–C5.
[12] Kunishima M, Kawachi C, Hioki K, Terao K, Tani S. Formation
of carboxamides by direct condensation of carboxylic acids and
amines in alcohols using a new alcohol- and water-soluble
condensing agent: DMT-MM. Tetrahedron. 2001;57:1551–1558.
[13] Protocol for on “BlotGlycoTM Glycan purification and labeling
kit”. Sumitomo Bakelite Co., Ltd. (2012).
[1] Varki A, Cummings RD, Esko JD, et al. Essentials of glycobiol-
ogy. 2nd ed. New York (NY): Cold Spring Harbor Lab Press;
2008.
[2] Couceiro JN, Paulson JC, Baum LG. Influenza virus strains
selectively recognize sialyloligosaccharides on human respiratory
epithelium; the role of the host cell in selection of hemagglutinin
receptor specificity. Virus Res. 1993;29:155–165.
[3] Ilver D, Amrnqvist A, Ogren J, et al. Helicobacter pylori
adhesin binding fucosylated histo-blood group antigens revealed
by retagging. Science. 1998;279:373–377.
[4] Ofek I, Beachey EH. Mannose binding and epithelial cell adher-
ence of Escherichia coli. Infect. Immun. 1978;22:247–254.
[5] Firon N, Ashkenazi S, Mirelman D, Ofek I, Sharon N. Aromatic
alpha-glycosides of mannose are powerful inhibitors of the
adherence of type 1 fimbriated Escherichia coli to yeast and
intestinal epithelial cells. Infect. Immun. 1987;55:472–476.
[6] Gouin SG, Wellens A, Bouckaert J, Kovensky J. Synthetic
multimeric heptyl mannosides as potent antiadhesives of uro-
pathogenic Escherichia coli. ChemMedChem. 2009;4:749–755.
[7] Bock K, Breimer ME, Bringnole A, et al. Specificity of binding
of a strain of uropathogenic Escherichia coli to Galα1-4Gal-con-
taining glycosphingo-lipids. J. Biol. Chem. 1985;260:8545–8551.
[8] Konowalchuk J, Speirs J, Stavric S. Vero response to a cytotoxin
of Escherichia coli. Infect. Immun. 1977;18:775–779.
[14] Bachmann BJ. Pedigrees of some mutant strains of Escherichia
coli K-12. Bacteriol. Rev. 1972;36:525–557.
[15] Poon RYM, Ohlsson-Wilhelm BM, Bagwell CB, Muirhead KA.
Use of PKH membrane intercalating dyes to monitor cell traffick-
ing and function. In: Diamond RA, DeMaggio S, editors. In living
color. New York (NY): Springer-Verlag; 2000. p. 302–352.
[16] Liang C-J, Yamashita K, Kobata K. Structural study of the
carbohydrate moiety of bovine pancreatic ribonuclease B. J. Bio-
chem. 1980;88:51–58.
[17] Lee J-w. World Health Organization. Water, Sanitation, and
Hygiene Links to Health, Facts and Figures-updated. November
2004.