Construction of artificial signal transducers on a lectin surface by post-photoaffinity-labeling modification for fluorescent saccharide biosensors

Article abstract of DOI:10.1002/chem.200304925

A new general method, post-photoaffinity-labeling modification (PPALM), for constructing fluorescent saccharide biosensors based on naturally occurring saccharide-binding proteins, lectins, is described in detail. An active-site-directed incorporation of

Full text of DOI:10.1002/chem.200304925

FULL PAPER
Construction of Artificial Signal Transducers on a Lectin Surface by Post-
Photoaffinity-Labeling Modification for Fluorescent Saccharide Biosensors
Tsuyoshi Nagase,[b] Eiji Nakata,[b] Seiji Shinkai,[b] and Itaru Hamachi*[a]
Abstract: A new general method, post-
photoaffinity-labeling modification
(PPALM), for constructing fluorescent
saccharide biosensors based on naturally
occurring saccharide-binding proteins,
lectins, is described in detail. An ac-
irradiation, the mannoside unit was
cleaved by reduction. The unique thiol
group thus produced was site-specifical-
ly modified with various fluorescent
groups (dansyl, coumarin, or dimethyla-
minobenzoate derivatives) to afford flu-
orescent Con As. The labeling site was
characterized by protease-catalyzed di-
gestion followed by HPLC, MALDI-
TOF MS, and tandem mass mass spec-
trometry; these methods indicated that
the photolabeling step is remarkably site
specific. Strong fluorescence was ob-
served in the engineered Con A with a
fluorophore, and the emission changed
sensitively upon saccharide complexa-
tion. The binding constants for various
saccharides were determined by fluores-
cence titration and demonstrated that
the binding selectivity and affinity of the
engineered Con As are comparable to
those of native Con A. The red shift of
the emission maximum, the decrease in
the fluorescence anisotropy of the dan-
syl unit, and the increase in the twisted
intramolecular charge transfer emission
caused by sugar binding to the engi-
neered Con A explicitly indicate that
the microenvironment of the appended
fluorophores changes from a restricted
and relatively hydrophobic environment
into a rather freely mobile and hydro-
philic environment.
tive-site-directed incorporation of
a
masked reactive site into a lectin was
conducted by using a photoaffinity la-
beling technique followed by demasking
and then chemical modification to yield
a fluorescent lectin. Two photoaffinity
labeling reagents were designed and
synthesized in this study. The labeling
reagent with a photoreactive site ap-
pended through a disulfide link to a
mannoside unit was bound to the sac-
charide-binding pocket of the lectin
concanavalin A (Con A). After light
Keywords: glycochemistry ¥ lectins
¥
photoaffinity labeling
¥ protein
engineering ¥ saccharide biosensors
artificial receptors that can selectively bind biologically
relevant molecules in aqueous solution. On the other hand,
a biosensor is an ideal system for precisely monitoring ions
and molecules of biological importance both in vivo and in
vitro,[3] because the protein framework can provide a sophis-
ticated molecular-recognition scaffold with high affinity and
specificity. A key problem in this case is how to carry out the
rational coupling of a signal-transduction device with molec-
ular-recognition events on a protein scaffold. Unlike the
situation with small, synthetic molecules, versatile methods
for site-specific chemical manipulation of biomacromolecules
are still poorly developed. Among the limited number of
examples, Tsien,[4] Imperiali,[5] Berg[6] and their co-workers
independently reported a fluorescence resonance energy
transfer (FRET) type of biosensor for calcium or zinc cations
that involved guest-induced peptide/protein folding. Hellinga
and co-workers developed a fluorescent saccharide sensor
equipped with an allosteric signal transducer.[7] The guest-
induced change in the fluorescence depolarization was
recently utilized by Thompson and co-workers.[8] Most of
these successful strategies are based on genetic protein
engineering or total/partial synthesis of proteins.
Introduction
In recent years, chemosensors based on organic scaffolds have
been actively investigated for monitoring molecules of bio-
logical importance.[1] In contrast to the great progress in the
detection of simple, small molecules such as ions, successful
examples of chemosensors for complex biomolecules such as
hormones, secondary messengers, and mono- or oligosacchar-
ides are still limited.[2] This is mainly due to the lack of
[a] Prof. I. Hamachi
PRESTO (Organization and Function, JST)
Institute for Materials Chemistry and Engineering (IMCE)
Department of Chemistry and Biochemistry
Graduate School of Engineering
Kyushu University, Fukuoka 812-8581 (Japan)
Fax : (‡81)92-642-2715
E-mail: itarutcm@mbox.nc.kyushu-u.ac.jp
[b] Dr. T. Nagase, E. Nakata, Prof. S. Shinkai
Department of Chemistry and Biochemistry
Graduate School of Engineering Kyushu University
Fukuoka 812-8581 (Japan)
Supporting information for this article is available on the WWW under
http://www.wiley-vch.de/home/chemistry/ or from the author.
3660
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200304925
Chem. Eur. J. 2003, 9, 3660 3669

3660 3669
Saccharides are important targets to be monitored quanti-
tatively, not only for clinical usage but also for the further
development of glycochemistry, because great advances in
glycoscience and -technology have rapidly clarified many of
the crucial roles of saccharide derivatives in biological
phenomena such as cell signaling, organogenesis, fertilization,
and inflammation.[9] Although an electrochemistry-based
amperometric glucose sensor that uses glucose oxidase has
now been commercialized for monitoring diabetes,[10] an
optical sensing system for various saccharides is expected to
be more advantageous and flexible.
To construct a fluorescent biosensor for saccharides, we
recruited a naturally occurring saccharide-binding protein,
lectin, as a protein scaffold.[11] Concanavalin A (Con A) from
Canavalia ensiformis, the most extensively explored lectin,
was utilized as a suitable model. Con A exists as a dimer at
low pH values (<5.5) and as a tetramer at high pH values
(>7).[12] Each monomer (MW ꢀ 26000 Da) contains 237 amino
acid residues with one sugar-binding site, as well as binding
sites for the ions Mn2‡ and Ca2‡.[13] Con A binds specifically to
the a anomers of d-mannopyranoside and d-glucopyrano-
side.[14]
Figure 1. a) Molecular design and structure of PPALM reagents 1a and 1b.
b) Representation of the PPALM scheme for semisynthesis of fluorescent
biosensors.
We recently reported a general semisynthetic method for
constructing a fluorescent saccharide biosensor based on
Con A.[15] Photoaffinity labeling of the sugar-binding pocket
of Con A, cleavage of the ligand, and chemical modification
with a fluorophore (post-photoaffinity-labeling modification,
PPALM) affords a fluorescent Con A that can act as a
saccharide biosensor. Here we describe several features of our
PPALM technique for engineering fluorescent Con A. De-
tailed investigation of the specificity of the photoaffinity
labeling site, the sugar-sensing mechanism of the fluorophore-
appended Con A, and the introduction of other read-out
modes, such as twisted intramolecular charge transfer emis-
sion, reveals that the present methodology is generally
valuable for the rational design of fluorescent biosensors
based on naturally occurring receptor proteins.
photoaffinity labeling reagents (1a, b) directed towards
[18] [19]
Con A.[17]
,
,
The synthetic routes to these reagents are
outlined in Figure S1 in the Supporting Information. Thio-
glycosylation of peracetyl-d-mannose (4) with thioacetic acid
in the presence of BF3 ¥ OEt2 selectively gave peracetyl-a-d-1-
thiomannoside (5). Deacetylation of 5 with LiOH followed by
treatment with 2,2'-dithiodipyridine gave pyridyl-a-d-dithio-
mannoside 2. The disulfide exchange reaction of 2 with the
thiol derivatives of the corresponding labeling units (3a, b)
afforded the photoaffinity labeling reagents 1a and 1b,
respectively.
Post-photoaffinity-labeling modification of Con A: The
PPALM method is briefly illustrated in Figure 1b. When the
affinity ligand 1 was bound to the binding pocket of Con A,
photolabeling was carried out by UV-light irradiation
(330 nm < l < 400 nm for trifluoromethylphenyldiazirine).
The labeled Con A was purified by gel chromatography
(Biogel P-30) and affinity column chromatography (Sepha-
dex G-100).[20] As shown in Figure 2, three peaks were
observed in the affinity chromatography results. Each of
these was characterized by MALDI-TOF MS, the results of
which allowed the first peak to be assigned to a homodimer of
Con A labeled with one molecule of 1 and the second peak to
be assigned to a heterodimer consisting of a monolabeled
Con A and a nonlabeled Con A. The third peak was assigned
to a homodimer of the nonlabeled Con A (that is, native
Con A). For the structural and functional studies mentioned
later, the heterodimer was mainly used because of its higher
yield.[21] The yield of labeled Con A is greatly dependent on
the molecular structure of the PPALM reagents: the yield is
about 10% for 1a and less than 1% for 1b.
Results
Molecular design of the photoaffinity labeling reagents for
Con A: We designed two PPALM reagents, 1a and 1b, which
are shown in Figure 1a. The reagents needed to include three
functionalities in the molecular structure, that is, 1) a high-
affinity ligand (a-d-mannoside) to bind to Con A selectively,
2) a photoreactive group (trifluoromethylphenyl diazirine) to
label Con A by photoirradiation, and 3) a suitable cleavage
site (disulfide) to remove the ligand after photolabeling and to
generate a chemoselective modification site[15a] (thiol group).
It was previously reported that the strongest monosaccharide
ligand for Con A is methyl-a-d-mannopyranoside.[16] Thus, a
mannoside group attached through a disulfide link was
employed as a high-affinity ligand. The disulfide bond could
be reductively cleaved by (Æ)-dithiothreitol (DTT) or tris(2-
carboxyethyl)phosphine (TCEP) treatment so that a reactive
benzylthiol moiety is produced. As a photoreactive unit,
trifluoromethylphenyldiazirine, which is conventionally used
in these experiments, was selected for the design of the
Subsequent treatment of the labeled Con A with DTT or
TCEP gave a unique benzylthiol site (SH-Con A), which was
chemoselectively modified with a haloacetylated fluorophore
(Fluoro-Con A). These two steps proceeded almost quantita-
tively. The modified Con As (from labeling with 1a) at all
Chem. Eur. J. 2003, 9, 3660 3669
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3661

I. Hamachi et al.
FULL PAPER
Figure 2. Fraction curves in the affinity purification of a) 1a-labeled
Con A; b) 1b-labeled Con A.
stages were characterized by MALDI-TOF MS (Figure 3).
Mass peaks (m/z) were observed at 25981 Æ 10 for 1a-labeled
Con A, at 25796 Æ 11 for SH-Con A, and at 26097 Æ 6 for
IAEDANS-Con A (see Scheme 1b for structure of IAE-
DANS); these data agree well with the sum of the MW of
native Con A (25584 Da) and the corresponding residues.[22]
The secondary structure of the modified Con As was moni-
tored by CD spectroscopy. For the modified Con A at each
stage (the 1a-labeled, SH-, and IAEDANS-Con A), a neg-
ative Cotton peak at 218 nm, characteristic of a typical b
sheet, was observed; this peak is comparable to that of native
Con A,[23] a fact suggesting that the secondary structure of
Con A is not significantly perturbed by the chemical mod-
ification.
Figure 3. MALDI-TOF mass spectra of a) native Con A; b) 1a-labeled
Con A; c) SH-Con A; d) IAEDANS-Con A.
by comparison with the HPLC chart of native Con A, it could
be seen that one original peak (Rt ˆ 74.5 min, assigned to
fragment A4 by MALDI-TOF MS in the case of native
Con A) disappeared and a new peak (Rt ˆ 81.4 min) appeared
instead. Consistently, MALDI-TOF MS analysis of the new
peak showed that it can be ascribed to the peptide fragment
A4 (Arg60 Lys101)) plus one molecule of the denitrogen-
ated ligand 1a. Subsequently, the modified fragment A4 was
collected by HPLC and subjected to trypsin-catalyzed diges-
tion; the resultant mixture was analyzed by mass spectrom-
etry. Fragment A4 involves two arginine residues so it was
divided into one arginine and two peptides by trypsin. The
longer peptide (Leu61 Arg90) had a molecular weight
identical to the native one, while the shorter peptide
(Val91 Lys101) showed the sum of the corresponding pep-
tide and the denitrogenated ligand 1a in its MW (747.32 Da as
a dication form). Precise sequencing of the shorter fragment
was next conducted by the tandem mass mass method
(Figure 4b). Each difference in the mass peaks showed good
agreement with the mass unit of the corresponding natural
amino acid with the exception of the Tyr100. The difference
mass of the Tyr100 site was determined to be 561.11 Da, a
Determination of the labeled site of the modified Con A: To
determine the labeled site, the 1a-labeled Con A was digested
by lysyl endopeptidase (LEP), which can cleave peptide
bonds on the C-terminal side of lysine residues.[24] Figure 4a
shows the reverse-phase HPLC trace of the digested peptides
for the labeled and native Con As. There are 12 lysine
residues in Con A (the full amino acid sequence is given in
the Supporting information) so 13 fragments were generated
by the LEP digestion. In native Con A, eight peaks appeared
in the HPLC trace, all of which were assigned by MALDI-
TOF MS. Five other fragments (Ser31 Lys35, Lys36, Thr37
Lys39, Leu115 Lys116, Asp136 Lys138) were too short to
be detected under the HPLC conditions used. In the case of
the labeled Con A, eight peaks were also observed. However,
3662
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3660 3669

Fluorescent Saccharide Biosensors
3660 3669
Figure 4. a) Reverse-phase HPLC trace of native and 1a-labeled Con As
after lysyl-endopeptidase digestion. b) Tandem mass mass analysis of the
1a-labeled fragment A4.
value that is completely coincident with the sum of the MW of
tyrosine and the denitrogenated labeling ligand 1a (a
calculated value of 561.1103 Da). Thus, the labeling site was
assigned to Tyr100. It is clear that the photoaffinity labeling
process of the present PPALM method is remarkably site
selective on a protein surface.
Fluorescent sensing of saccharides by using fluoro-Con As:
IAEDANS-Con A (from labeling with 1a) prepared by the
PPALM method with the iodoacetylated dansyl derivative
showed a strong emission due to the dansyl fluorophore, an
environmentally sensitive probe.[25] A spectral change in the
emission of IAEDANS-Con A was observed on addition of
methyl-a-d-mannoside (Me-a-Man), one of the strongest
ligands for native Con A (Figure 5a). With increasing con-
centrations of Me-a-Man, the emission intensity at 484 nm
decreased and the emission peak maximum was slightly red
shifted (488 nm). This implies that the dansyl unit moves into
a more polar microenvironment than the original position
upon complexation of IAEDANS-Con A with Me-a-Man. It
is noteworthy that the binding process of IAEDANS-Con A
with Me-a-Man can be directly monitored by the fluorescence
signal change. Fluorescence titration gave the typical satu-
ration curve, which yielded a binding constant of 7.8 Â 103 mÀ1
for Me-a-Man by the Benesi Hildebrand plot analysis[26]
(Figure 5b). This value is almost comparable to that of native
Con A (1.1 Â 104 mÀ1) as determined by isothermal titration
calorimetry.[12a]
Figure 5. a) Spectral changes in the fluorescence of IAEDANS-Con A
upon addition of Me-a-Man (0 ! 20 mm). [IAEDANS-Con A] ˆ 5 mm, in
50 mm HEPES buffer (pH 7.0) containing 1 mm CaCl2, 1 mm MnCl2, and
0.1m NaCl. Temperature ˆ 15 Æ 18C, lex ˆ 340 nm. b) The binding curve for
IAEDNAS-Con A. Inset: Benesi Hildebrand plot analysis. HEPES ˆ 2-
[4-(2-hydroxyethyl)-1-piperazinyl]ethanesulfonic acid.
than glucose derivatives (Me-a-Man > Me-a-Glc, Man >
Glc); 2) the affinities of galactose, ribose, and arabinose were
negligible; 3) the a configuration in the glycoside bond was
preferable to the b configuration (Me-a-Glc > Me-b-Glc);
and 4) l-glucose, a mirror image of native d-glucose, was
practically not bound. The observed binding selectivity of
IAEDANS-Con A was identical to that of native Con A,[16]
a
fact suggesting that the binding pocket of IAEDANS-Con A
for the monosaccharide is not disturbed by the PPALM.
For a series of the oligosaccharides, the selectivity of
IAEDANS-Con A observed by fluorescence titration was
also in good agreement with that of native Con A as
determined by isothermal titration calorimetry, that is, Man-
tri > Man-bi > Isomal ꢁ Cel, Lac (Figure 6b). Man and Glc at
the nonreducing saccharide terminal were preferable to Gal,
and there was also a preference for the a-glycoside anomer to
the b-glycoside. It is noteworthy that the binding constant of
IAEDANS-Con A for Man-tri was found to be identical with
that of native Con A.[12a,b] Man-tri is known to be one of the
main core structures in branched saccharides that attach to
We next conducted similar titration experiments for other
saccharides (the structures are summarized in Scheme 1a and
the Supporting Information) with IAEDANS-Con A. The
saturation curves shown in Figure 6 gave the binding con-
stants for various saccharides that are displayed in Table 1.
Among the monosaccharides (Figure 6a) these show signifi-
cant features: 1) Mannose derivatives showed a higher affinity
Chem. Eur. J. 2003, 9, 3660 3669
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3663

I. Hamachi et al.
FULL PAPER
Figure 6. Fluorescence titration profile of the relative intensity (I/I0)
versus the saccharide concentration (log [saccharide]) for: a) IAEDANS-
*
^
Con A (5 mm) with the monosaccharides Me-a-Man ( ), Me-a-Glc ( ),
Scheme 1. a) Representative saccharide structures used in this study. The
abbreviation is given below the saccharide name. b) Molecular structures of
the fluorescent probes attached to SH-Con A. The wavelengths given in the
parentheses are the excitation and emission wavelengths for each
fluorophore.
&
~
!
~
&
Man ( ), Glc ( ), Me-a-Glc ( ), l-Glc ( ), Ara ( ), Me-a-Gal (‡), Gal
*
(
), and Rib (Â); b) IAEDANS-Con A (0.2 mm) with the oligosaccharides
~
&
^ * ^ *
Man-tri ( ), Man-bi ( ), Pal ( ), Isomal ( ), Isomalti ( ), Malti ( ), Cel
!
(Â), and Lac ( ).
dansyl group is crucial to the above-mentioned fluorescent
response of IAEDANS-Con A.
natural glycoproteins. Therefore, the ability to detect Man-tri
is useful for quantitative estimation of such a glycoprotein
family. This is the first example of fluorescence detection
of Man-tri at mm concentrations, to the best of our knowledge.
It was reported that the Con A surface in the proximity of
the binding pocket is involved in Man-tri binding.[12c] The
present data implied that the protein surface responsible for
the sugar binding was structurally conserved even with the
modification, in addition to the binding pocket. Thus, not
only binding selectivity but also the affinity of native
Con A for various saccharides is quantitatively retained in
IAEDANS-Con A.
As a control sample, we prepared a Con A randomly
modified with a dansyl group. Some lysine residues may have
be modified because dansyl chloride was used as the reagent.
After purification of the Con A modified randomly with an
average of 1 molecule of dansyl unit per protein, a fluores-
cence titration with Me-a-Man was conducted. No significant
changes in either its emission intensity or wavelength were
observed (see Supporting Information). This control ex-
periment indicates that the site-selective attachment of a
The fluorescence depolarization measurements[8a] strongly
supported the molecular mechanism of the fluorescence
response of IAEDANS-Con A to saccharide binding. The
fluorescence anisotropy ratio (r) was determined to be 0.40
without sugars, which is a typical value for the probe bound to
proteins. Interestingly, the anisotropy ratio decreased to 0.35
on addition of Me-a-Man (see Supporting Information). The
saturation curve showed good agreement with the curve
obtained by the steady-state fluorescence titration and gave a
binding constant of 7.6 Â 103 mÀ1, which is almost identical to
the value from the fluorescence titration. A decrease in the
anisotropy is usually due to an increase in the freedom of the
fluorophore in motion. Thus, the results clearly imply that the
dansyl moiety is restricted in the binding pocket of Con A
without sugars but the environment changes into one of rather
higher mobility upon complexation with Me-a-Man. This is
nicely consistent with our tentative explanation of the
saccharide-induced fluorescence change in IAEDANS-
Con A.
3664
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3660 3669

Fluorescent Saccharide Biosensors
3660 3669
Table 1. Comparison of the association constants of Fluro-Con As with those of native Con A.
Saccharide
Ka [mÀ1
]
native Con A
IAEDANS-Con A
IAANS-Con A
DACIA-Con A
DCIA-Con A
Me-a-Man
Me-a-Glc
Me-b-Glc
Me-a-Gal
Man
Glc
l-Glc
Gal
Rib
1.1 Â 104[a]
7.8 Â 103[b]
1.6 Â 103[b]
2.6 Â 103[b]
3.7 Â 103[b]
1.9 Â 103[b]
3.0 Â 103[a]
1.4 Â 103[b]
1.2 Â 103[b]
8.3 Â 102[b]
[c]
[d]
[d]
[d]
< 1.0 Â 102[b]
[c]
[c]
[d]
[d]
[d]
[d]
[c]
[d]
[d]
[d]
[d]
[d]
[d]
[c]
[d]
[d]
[d]
[d]
[d]
[d]
[c]
[d]
[d]
2.2 Â 103[a]
1.3 Â 103[b]
8.0 Â 102[a]
1.0 Â 103[b]
[c]
[c]
[c]
[e]
[e]
[c]
[c]
[c]
Ara
Man-tri
Man-bi
Malti
2.5 Â 105[a]
2.5 Â 105[f]
5.8 Â 103[b]
1.0 Â 103[b]
1.6 Â 103[b]
1.3 Â 103[b]
6.5 Â 105[f]
5.4 Â 105[f]
1.2 Â 105[f]
3.0 Â 104[a]
1.9 Â 104[b]
6.9 Â 103[b]
5.2 Â 103[b]
[e]
[d]
[d]
[d]
[e]
[d]
[d]
[d]
[c]
[d]
[d]
[d]
[d]
[c]
[d]
[d]
[d]
[d]
[c]
[d]
Pal
Isomal
Isomalti
Cel
1.7 Â 103[a]
[e]
8.8 Â 102[b]
[c]
[e]
[c]
[c]
Lac
[a] Reported values detrermined by isothermal titration calorimetry.[12] [b] Determined by Benesi Hildebrand plot analysis. [c] Cannot be determined
because of the low affinity. [d] Titration experiments were not conducted. [e] Not reported previously. [f] Determined by nonlinear curve fitting analysis.
Other modified Con As with different fluorophores (cou-
marin derivatives DCIA and DACIA, or aminonaphthalene-
sulfonate derivative IAANS; shown in Scheme 1b) were
prepared by the same procedure. These Fluoro-Con As (from
labeling with 1a) showed good fluorescent sensing affinity
and selectivity in a similar way to IAEDANS-Con A (from
labeling with 1a), that is, they show sensitive fluorescent
change with Me-a-Glc, Me-a-Man, Man-bi, and Man-tri, but
do not respond to Gal and Cel (see Table 1). In addition to
varying the monitoring wavelength (450 490 nm), it was an
advantage to shift the excitation wavelength into the visible
light region to avoid the undesirable interference by other
impurities derived from proteins or DNA in the analyte
solutions. In the present case, we could shift the excitation
from 330 to 420 nm by changing the fluorophore without
decreasing the sensing capability.
Twisted intramolecular charge transfer (TICT) mode for
reading out the saccharide by DMAB-Con A: In addition to
the environmentally sensitive probes (see above), a TICT
emission probe can be engineered onto a Con A surface by
using the PPALM method. Dimethylaminobenzoate
(DMAB), a typical fluorophore for TICT, was converted into
a bromoacetyl amide derivative, DMAB-Br (Figure 7a).
Figure 7. a) Structure of DMAB-Br. b) Differential spectral changes in the
DMAB-Br was appended to the benzylthiol site of SH-Con
A (DMAB-Con A). Figure 7b shows difference fluorescence
fluorescence of DMAB-Con A upon addition of Me-a-Man (0 ! 10 mm).
[DMAB-Con A] ˆ 1 mm in 50 mm HEPES buffer (pH 7.0) containing 1 mm
spectra of DMAB-Con A due to the addition of Me-a-Man.
In contrast to the monotonous decrease in the fluorescence
intensity for IAEDANS-Con A, an increase in the fluores-
cence at 435 nm was observed with a concurrent decrease in
the fluorescence at 360 nm with an isosbestic emission point at
395 nm in the case of DMAB-Con A. Such a fluorescence
change at two distinct wavelengths is, in principle, applicable
to the ratiometric measurement, which enables the more
precise (quantitative in some cases) estimation of analytes by
CaCl2, 1 mm MnCl2, and 0.1m NaCl. Temperature ˆ 15 Æ 18C, lex
ˆ
310 nm. Inset: fluorescence spectral changes.
reducing the influence of photobleaching, the effect of the
probe concentration, and other environmental side effects.
According to the literature,[27] the peak at 435 nm can be
ascribed to the emission from the TICT excited state, and the
peak at 360 nm is due to the emission from the normal planar
Chem. Eur. J. 2003, 9, 3660 3669
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3665

I. Hamachi et al.
FULL PAPER
(NP) excited state. It is generally considered that the TICT
emission increases in a polar solvent rather than a less polar
solvent because the TICT excited state is more charge-
separated, relative to the NP excited state. Therefore, the
increase in the TICT emission and the decrease in the NP
emission induced by Me-a-Man indicate that the DMAB
moiety shifts from a rather hydrophobic sugar-binding pocket
to a bulk aqueous solution. This is absolutely consistent with
the mechanism of the saccharide-induced fluorescence
change in IAEDANS-Con A. Both emissions (due to TICT
and NP) change with typical saturation behaviors, which give
us the apparent binding constants as follows: 2.3 Â 103 mÀ1 for
Me-a-Man, 1.3 Â 103 mÀ1 for Me-a-Glc, and a negligible value
for Gal. The saccharide selectivity is again coincident with
that of native Con A, similarly to the case of IAEDANS-
Con A.
as Tyr12, Thr15, Tyr100, and His205, seem to be intrinsic
candidates for a photolabeling reaction by the electrophilic
trifluoromethyl carbene moiety. Among them, only Tyr100
was modified as shown in our peptide-mapping experiments
in the case of 1a. Importantly, it was reported that the hydroxy
group of Tyr100 is not responsible for the sugar binding in
native Con A. Tyr12 and Thr15 on the Con A surface, which
can interact with the branched mannose unit of Man-tri, are
considered to be on the opposite side to the hydrophobic
fence consisting of Tyr100.[12c] Thus, the modification at
Tyr100 does not strongly disturb the structure of the binding
surface of Con A. These factors should contribute to the
satisfactory retainment of the original selectivity and affinity
for various saccharides in IAEDANS-Con A.
In contrast, another photolabeling reagent 1b did not yield
any substantial amount of labeled Con A. This is probably due
to the fact that the para substitution is oriented to the outside
of Con A and is thus far from the residues surrounding the
binding pocket. On the other hand, we previously reported
that p-azidophenyl-a-d-mannoside (p-APM) reacted with
Con A in 20% yield and that the modification reaction is
distributed to two sites, such as Tyr12 or Tyr100, both of which
are positioned in the proximity of the sugar-binding pock-
et.[15b] Compared to 1a, the linkage of p-APM between the
mannose group and the phenylazide group is shorter by two
bond lengths. Thus, the azide group can reach Tyr12, which is
located on the more buried side, as well as Tyr100, so that
photolabeling occurs at both positions. In contrast to p-APM,
the photoreactive species of 1a and 1b is expected to be
generated on the external protein surface, and thus it is
reasonable that the photoaddition predominantly takes place
Discussion
The present results reveal that elaborate organic chemistry on
a protein surface is a powerful tool for converting a lectin into
a fluorescent saccharide sensor. As a pioneering example,
Schultz and co-workers proposed a unique approach to
conducting organic chemistry on an antibody cavity by a
combination of affinity labeling with subsequent modification
to produce a fluorescent antibody.[28] The reactive species of
photolabeling reagents such as carbene or nitrene are so
reactive that PPALM may be applied to a wide range of
proteins which do not have such highly reactive amino acids
around the ligand binding site. The photoaffinity labeling
technique has been widely uti-
lized so far in structural and
functional biology to deter-
mine unknown enzyme active
sites or protein/protein inter-
19]
action surfaces.[17
The la-
beled proteins are usually
fragmented into many pepti-
des for the structural analysis.
In contrast, the present
PPALM method provides nov-
el functionality to the target
protein without fragmenta-
tion. Thus, the photoaffinity
labeling method can be ex-
tended into protein engineer-
ing.
It is remarkable that the
photoaffinity labeling reaction
with 1a is highly site-selective.
As shown in Figure 8a, the
crystal structure of Con A
complexed with p-nitrophen-
yl-a-d-mannoside
suggests
that there are several nucleo-
philic amino acid residues
around the saccharide-binding
pocket.[29] These residues, such
Figure 8. The molecular structures of the residues in the proximity of the saccharide-binding site of Con A: a) the
crystal structure of the p-nitrophenyl-a-d-mannopyranoside complex (PDB file code: 1VAM); b) molecularly
modeled structure of the complex with 1a. c) Schematic illustration of the present sensing mode of IAEDANS-
Con A.
3666
¹ 2003 Wiley-VCH
Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3660 3669

Fluorescent Saccharide Biosensors
3660 3669
at the more solvent-exposed Tyr100. A molecular modeling
study of the 1-labeled Con A (Figure 8b) gave us the roughly
estimated distance between the reactive center of 1 and
Tyr100. In the case of meta-substituted 1a, the average
distance is in the range of 3.5 ä, which is close enough to be
photolinked with Con A, whereas the distance is more than
5 ä in the complex with the linear para substituent 1b. These
results clearly demonstrate that fine tuning of the structure of
the PPALM reagents will allow regulation of the modification
position, which is crucial for the introduced novel func-
tion.
Conclusion
We have successfully proposed a simple method called
PPALM (post-photoaffinity-labeling modification) for con-
verting a lectin protein into a fluorescent saccharide sensor.
The resultant engineered Con A showed a fluorescent re-
sponse to a series of saccharides with superior selectivity and
affinity identical to that of native Con A. The high site
specificity of the attachment of the probe and the chemical
flexibility for selection of a suitable fluorophore are advanta-
geous in the PPALM method. It was reported in the literature
In addition, the 3D structure suggests that the saccharide-
binding pocket is partially hydrophobic because it is sur-
rounded by the hydrophobic fence consisting of Tyr12, Tyr100,
and Leu99 (Figure 8a). Thus, it is reasonable that the
hydrophobic fluorophore appended to the proximity of the
binding pocket is preferably located in the rather hydrophobic
position without saccharides. The fluorophore moves to the
more hydrophilic microenvironment upon sugar binding, as
suggested by the results in which the emission is lessened in its
intensity and red-shifted upon sugar binding. The decrease in
the fluorescence anisotropy upon saccharide binding strongly
supported the tentative explanation mentioned above. That is,
the mobility of the fluorophore is restricted in the saccharide-
binding pocket, and the sugar-induced relocation causes a
partial recovery of the freedom in motion. This is also nicely
coincident with the increase in the TICT emission induced by
Me-a-Man complexation to DMAB-Con A, along with the
concurrent decrease in the NP emission. In the control
experiment with randomly modified Con A, no significant
fluorescence changes occurred. Therefore, the signal-trans-
duction modes displayed in the present examples are based on
the response caused by the microenvironmental change in the
fluorescent probes induced by sugar binding (Figure 8c). It is
concluded that the active-site-directed incorporation of the
fluorophores effectively coupled the sugar-binding event to
the fluorescent signal change.
Various fluorophore-labeled Con As can be prepared from
SH-Con A by simple mixing with the haloalkylated species.
The binding constants for Me-a-Man evaluated by the
corresponding Fluoro-Con A (from labeling with 1a) are
varied: 7.8 Â 103 (IAEDANS), 3.7 Â 103 (DACIA), 2.6 Â 103
(IAANS), and 1.9 Â 103 mÀ1 (DCIA). Apparently, the more
hydrophobic and bulky fluorophores suppressed more strong-
ly the binding affinity to Me-a-Man. This might be ascribed to
the partial blocking of the binding pocket of Con A by the
bulky fluorophore or to the fluorophore-induced disorder of
the binding pocket. However, the considerable suppression of
the affinity by the attached fluorophore was not observed in
the case of oligosaccharide sensing, which may be due to the
fact that the protein surface, as well as the binding pocket, is
greatly involved in the oligosaccharide binding. The present
PPALM strategy has the advantage that one can conveniently
and rapidly attach various kinds of fluorophores to Con A.
This facilitates the screening for satisfactory sensitivity, the
modulation of the excitation, the monitoring of the wave-
length, and the selection of an appropriate transduction mode,
such as environmentally sensitive probes and TICT emission
probes, to meet each objective in biosensing.
that the binding pocket of naturally occurring lectins is rather
[30]
hydrophobic in many cases.[11a]
,
Therefore, this strategy is
anticipated to be widely applicable to other saccharide-
binding proteins, simply by exchanging the sugar (that is, the
ligand) part of the PPALM reagent 1, to produce other
fluorescent biosensors with distinct saccharide specificities.
These will advance the quantitative estimation of various
saccharides and facilitate glycoscience and -technology.
Furthermore, many other naturally occurring receptor pro-
teins might become intriguing targets for PPALM, which will
extend the research field of biosensors.
Experimental Section
General methods: CH3CN and MeOH were distilled under a nitrogen
atmosphere prior to use. CH3CN was dried over calcium hydride. MeOH
was dried over Mg and I2. Unless otherwise stated, all of the air- or
moisture-sensitive reactions were performed under a nitrogen atmosphere.
1H NMR spectra were recorded either on Bruker AC250-P (250 MHz) or
Bruker DRX-600 (600MHz) spectrometer. ESI-TOF MS was measured
with a PerSeptive Mariner apparatus. MALDI-TOF MS were recorded on
a Perkin Elmer Voyager DE-RP apparatus, with 2,5-dihydroxybenzoic
acid, a-cyano-4-hydroxycinnamic acid (CHCA), or sinapinic acid (SA) as
the matrix. UV-Vis spectra were obtained on a Hitachi U-3000 spectrom-
eter. CD spectra were recorded on a Jasco J-720 spectrometer. Fluores-
cence spectra were recorded on a Hitachi F-4500 spectrometer. Synthetic
methods and characterization data for 1a, 1b, and DMAB-Br are described
in the Supporting Information. Con A was purchased from Funakoshi and
used without further purification. All of the fluorescent probes were
purchased from Molecular Probes. The other chemical reagents were
purchased from Aldrich or Tokyo Chemical Industry.
Photoaffinity labeling of Con A with the labeling reagent 1a[31]: Con A
(30 mg, 1.2 mmol/monomer-based) was dissolved in 10 mm acetate buffer
(2 mL; pH 5.0) containing 1 mm CaCl2, 1 mm MnCl2, and 0.2 m NaCl. The
labeling reagent 1a (2.0 mg, 4 equiv to 1 sugar-binding site of Con A) in
DMF (80 mL) was slowly added to the Con A solution at 08C. This solution
was stored for 30 min at 08C in the dark and then gently bubbled with
nitrogen gas to remove dioxygen; this was followed by centrifugation
(10000 rpm, 5 min) at 48C. The supernatant was transferred into a UV cell
(1 cm) and photoirradiated for 40 min at 108C by a 500 W high-pressure Hg
lamp (USHIO Optical ModuleX) with a glass filter that cuts out light of
wavelengths below 320 and above 380 nm. After removal of a trace amount
of precipitate by centrifugation (10000 rpm, 5 min), the supernatant was
subjected to gel filtration chromatography (Biogel P-30, 2 Â 20 cm), eluted
with the buffer A (10 mm acetate buffer (pH 5.0), 1 mm CaCl2, 1 mm
MnCl2) at 48C. The protein fractions were collected and concentrated by
ultrafiltration (YM10 filters, Amicon). The concentrated protein solution
was subsequently purified by affinity chromatography (Sephadex G100,
2 Â 30 cm), eluted with a linear gradient of buffer A with 0 ! 40 mm d-
glucose. Three peaks were observed in the trace. Each fraction was
characterized by MALDI-TOF MS. The first peak (m/z 25981) was
assigned to a homodimer of Con A labeled with 1 molecule of 1a, the
second peak was assigned to a heterodimer consisting of monolabeled
Con A and nonlabeled Con A. The third peak (m/z 25583) was assigned to
Chem. Eur. J. 2003, 9, 3660 3669
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3667

I. Hamachi et al.
FULL PAPER
a homodimer of nonlabeled Con A (that is, native Con A). (It has been
reported that Con A exists as a dimer at pH 5.0.) The fractions which made
up the second peak was collected and concentrated.
emission wavelengths used for each Fluoro-Con A are described in the
corresponding figures. The fluorescence titration curves were analyzed with
a Benesi Hildebrand plot or nonlinear curve-fitting analysis to yield the
association constants (Ka).
Fluorescence depolarization measurement[34]: The fluorescence anisotropy
ratio (r) of IAEDANS-Con A was evaluated from Equation (1) (Weber
equation), where Ik and I? are the intensities of fluorescence observed
through polarizers parallel and perpendicular to the polarization of the
exciting light, respectively.
Preparation of SH-Con A: DTT (final concentration: 5 mm) was added to a
solution of 1a-labeled Con A (from the fractions that made up the second
peak in the affinity column; 20 mm, 10 mL) in 100 mm phosphate buffer
(pH 8.0), and the reaction mixture was incubated for 20 h at 48C under a
nitrogen atmosphere. The reaction solution was purified by gel filtration
chromatography (Biogel P-30, 2 Â 20 cm) eluted with 100 mm phosphate
buffer (pH 7.2) to give SH-Con A (12 mm, 16 mL) in quantitative yield. The
thiol content in the SH-Con A was determined by the Ellman method,[32]
which showed that one Con A dimer has one SH group. MALDI-TOF MS
(SA) demonstrated the cleavage of the thiomannose unit: m/z calcd:
25787; found: 25796. The SH-Con A was quickly subjected to the next
modification.
r ˆ (Ik À I?)/(Ik ‡ 2I?)
(1)
The fluorescence intensity was measured with a fluorescence spectropho-
tometer (Hitachi F-4500) equipped with polarizer.
Molecular modeling: Model building and molecular dynamics calculations
were performed with the Insight II/Discover program packaged in the
context of Molecular Simulations Inc. All calculations were performed with
the consistent valence force field. The structure shown in Figure 8 is a
minimized low-energy conformation selected from 50000 dynamics
interactions at 450 K. The minimization was performed until a maximum
derivative of 0.001 was achieved. During the dynamics run and the
minimization, the Con A and the mannopyranoside ring were restrained
according to the crystallographic structure of the Con A p-nitrophenyl-a-
d-mannoside complex published by Kanellopoulos and co-workers.[29]
Preparation of Fluoro-Con A: Iodoacetyl-fluorophore (in the case of
IAEDANS: 0.22 mg, 5 Â 10À7 mol, 10 equiv to 1 SH group) in DMF
(100 mL) was slowly added to a solution of SH-Con A (20 mm, 5 mL) in
100 mm phosphate buffer (pH 7.2) at 08C, and the reaction mixture was
incubated for 12 h at 48C in the dark. After incubation, the solution was
submitted to gel filtration chromatography (Biogel P-30, 1 Â 15 cm), eluted
with the buffer B (50 mm HEPES buffer (pH 7.0), 1 mm CaCl2, 1 mm
MnCl2, and 0.1m NaCl). The protein fraction was collected, concentrated
by ultrafiltration, and dialyzed against buffer B to afford Fluoro-Con A in
quantitative yield (in the case of IAEDANS-Con A: 12 mm, 8 mL). For
IAEDANS-Con A, the labeling efficiency was estimated from the ratio of
[IAEDANS]/([Con A]/2) to be 0.75 0.90. IAEDANS-Con A was identi-
fied by MALDI-TOF MS (SA): m/z calcd: 26093; found: 26097. The same
procedure was employed for modification with the other fluorophores and
the overall labeling efficiency of SH group was shown to be 0.7 0.9.
IAEDANS: e336 ˆ 5700 mÀ1 cmÀ1; IAANS: e327 ˆ 26000 mÀ1 cmÀ1; DCIA:
Acknowledgement
We gratefully thank Dr. Masato Ikeda for valuable comments on the
molecular modeling study. This research was partially supported by
PRESTO (JST), a specially promoted area (Dynamic Control of Strongly
Correlated Soft Materials, No. 413), and a Grant-in-Aid for 21st COE
Research (™Functional Innovation of Molecular Informatics∫) from the
Ministry of Education, Science, Sports, and Culture of Japan.
e384 ˆ 33000 mÀ1 cmÀ1
;
DACIA: e384 ˆ 31000 mÀ1 cmÀ1
;
DMAB: e300
ˆ
22500 mÀ1 cmÀ1
.
Preparation of the randomly modified Con A: A solution of dansyl
chloride (0.24 mg, 8.9 Â 10À7 mol) in acetone (0.1 mL) was added to a
solution of Con A (15 mg, 5.9 Â 10À7 mol) in 0.1m NaHCO3 buffer (pH 8.4;
15 mL). The mixed solution was stored for 24 h at 48C in the dark. The
solution was dialyzed against 10 mm acetate buffer (pH 5.0) and then
50 mm HEPES buffer (pH 7.0) containing 1 mm MnCl2, 1 mm CaCl2, and
0.1m NaCl to yield the randomly dansyl-modified Con A. The UV/Vis
spectrum showed that one Con A monomer had one dansyl group.
[1] a) A. W. Czarnik, Acc. Chem. Res. 1994, 27, 302 308; b) A. P. de Silva,
H. Q. N. Gunaratne, T. Gunnlaugsson, A. J. M. Huxley, C. P. McCoy,
J. T. Rademacher, T. E. Rice, Chem. Rev. 1997, 97, 1515 1566.
[2] a) T. D. James, K. R. A. S. Sandanayake, S. Shinkai, Angew. Chem.
1996, 108, 2038 2050; Angew. Chem. Int. Ed. Engl. 1996, 35, 1910
1922; b) C. T. Chen, H. Wagner, W. C. Still, Science 1998, 279, 851
853; c) A. Sugasaki, K. Sugiyasu, M. Ikeda, M. Takeuchi, S. Shinkai, J.
Am. Chem. Soc. 2001, 123, 10239 10244.
[3] a) H. W. Hellinga, J. S. Marvin, Trends Biotechnol. 1998, 16, 183 189;
b) D. A. Zachrias, G. S. Baird, R. Y. Tsien, Curr. Opin. Neurobiol.
2000, 10, 416 421; c) L. Tolosa, I. Gryczynski, L. R. Eichhorn, J. D.
Dattelbaum, F. N. Castellano, G. Rao, J. R. Lakowicz, Anal. Biochem.
1999, 267, 114 120; d) M. Hirshberg, K. Henrick, L. L. Haire, N.
Vasisht, M. Brune, J. E. T. Corrie, M. R. Webb, Biochemistry 1998, 37,
10381 10385; e) J. D. Brennan, K. K. Flora, G. N. Bendiak, G. A.
Baker, M. A. Kane, S. Pandey, F. V. Bright, J. Phys. Chem. B 2000, 104,
10100 10110; f) G. J. Cotton, B. Ayers, R. Xu, T. W. Muir, J. Am.
Chem. Soc. 1999, 121, 1100 1101; g) I. Hamachi, R. Eboshi, J.
Watanabe, S. Shinkai, J. Am. Chem. Soc. 2000, 122, 4530 4531; h) T.
Morii, K. Sugimoto, K. Makino, M. Otsuka, K. Imoto, Y. Mori, J. Am.
Chem. Soc. 2002, 124, 1138 1139.
[4] a) A. Miyawaki, J. Llopis, R. Heim, J. M. McCaffery, J. A. Adams, M.
Ikura, R. Y. Tsien, Nature 1997, 388, 882 887; b) M. Zaccolo, F.
De Giorgi, C. Y. Cho, L. Feng, T. Knapp, P. A. Negulescu, S. S. Taylor,
R. Y. Tsien, T. Pozzan, Nat. Cell. Biol. 2000, 2, 25 29.
[5] G. K. Walkup, B. Imperiali, J. Am. Chem. Soc. 1997, 119, 3443 3450.
[6] H. A. Godwin, J. M. Berg, J. Am. Chem. Soc. 1996, 118, 6514 6515.
[7] a) D. E. Benson, D. W. Conrad, R. M. de Lorimier, S. A. Trammell,
H. W. Hellinga, Science 2001, 293, 1641 1644; b) J. S. Marvin, H. W.
Hellinga, J. Am. Chem. Soc. 1998, 120, 7 11; c) G. Gilardi, L. Q.
Zhou, L. Hibbert, A. E. G. Cass, Anal. Chem. 1994, 66, 3840 3847.
[8] a) R. B. Thompson, B. P. Maliwal, V. L. Feliccia, C. A. Fierke, K.
McCall, Anal. Chem. 1998, 70, 4717 4723; b) D. Elbaum, S. K. Nair,
M. W. Patchan, R. B. Thompson, D. W. Christianson, J. Am. Chem.
Lysyl-endopeptidase digestion: A solution of the 1a-labeled Con A (from
the fractions making up the first peak in the affinity purification; 100 mm,
500 mL) in 50 mm tris(hydroxymethyl)aminomethane (Tris)/HCl buffer
(pH 9.0) containing 3 m urea was treated with lysyl-endopeptidase at 378C
for 15 h at the enzyme/substrate ratio of 1:50 (w/w). The digestion reaction
was stopped by addition of trifluoroacetic acid (final concentration: 0.1%
(v/v)). The digested peptides were separated by HPLC and the collected
fractions were analyzed by MALDI-TOF MS (SA or CHCA).
Peptide sequencing by tandem mass mass analysis: The sample including
fragment A4 was treated with trypsin (in Tris/HCl buffer, pH 8.0) at 358C
for 20 h. The resultant mixture was purified through a ZipTip C18
(Millipore), eluted with 70% aqueous acetonitrile containing 1% formic
acid). The used eluent was subjected to ESI-TOF MS (Q-Tof 2, Micromass,
UK). The target peak (747.32/2 ‡ ) was sequenced by the tandem mass
mass method and analyzed by BioLynx software (manual mode).
UV/Vis, fluorescence, and CD spectroscopies: UV/Vis, fluorescence, and
CD spectra of the modified Con As were obtained by using 5 mm protein
concentration as determined by a Bradford assay in 100 mm phosphate
buffer (pH 7.0) at 258C. For the measurement of the fluorescence spectra
the slit widths of the excitation and emission were set to 10 nm, and an
excitation wavelength of lex ˆ 280 nm was used for characterization of
labeled- and SH-Con A. In the measurement of the CD spectra the
following conditions were used: sensitivity: 200 mdeg; resolution: 0.5 nm;
bandwidth: 1.0 nm; response: 0.5 s; scan speed: 100 nm/min.
Fluorescence titration of saccharide[33]: Saccharide stock solution was
added dropwise to 0.2 5 mm Fluoro-Con A in 50 mm HEPES buffer
(pH 7.0) containing 1 mm CaCl2, 1 mm MnCl2, and 0.1m NaCl at 15 Æ 18C.
The fluorescence spectra were measured. The slit widths for the excitation
and emission were set to 10 nm and 5 nm, respectively. The excitation and
3668
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3660 3669

Fluorescent Saccharide Biosensors
3660 3669
Soc. 1996, 118, 8381 8387; c) R. B. Thompson, B. P. Maliwal, C. A.
Fierke, Anal. Biochem. 1999, 267, 185 195.
[9] a) C. R. Bertozzi, L. L. Kiessling, Science 2001, 291, 2357 2364;
b) P. M. Rudd, T. Elliott, P. Cresswell, I. A. Wilson, R. A. Dwek,
Science 2001, 291, 2370 2376.
[10] A. Riklin, E. Katz, I. Willner, A. Stocker, A. F. Buckmann, Nature
1995, 376, 672 675.
[11] a) H. Lis, N. Sharon, Chem. Rev. 1998, 98, 637 674, and references
therein; b) H. Bittiger, H. P. Schnebli, Concanavalin A as a Tool, John
Wiley and Sons, New York, 1976.
1994, 59, 383 387; c) J. Brunner, H. Senn, F. M. Richards, J. Biol.
Chem. 1980, 255, 3313 3318; d) Y. Hatanaka, M. Hashimoto, H.
Nakayama, Y. Kanaoka, Chem. Pharm. Bull. 1994, 42, 826 831; e) M.
Nassal, J. Am. Chem. Soc. 1984, 106, 7540 7545.
[20] M. Beppu, T. Terao, T.Osawa, J. Biochem. 1975, 78, 1013 1019.
[21] We confirmed that the sensing capability of the IAEDANS-Con A
prepared from fractions making up the second peak is comparable to
that of IAEDANS-Con A prepared from fractions making up the first
peak.
[22] a) J. L. Wang, B. A. Cunningham, G. M. Edelman, Proc. Nat. Acad.
Sci. USA 1971, 68, 1130 1134; b) J. L. Wang, B. A. Cunningham, M. J.
Waxdal, G. M. Edelman, J. Biol. Chem. 1972, 250, 1490 1502.
[23] M. S. Herrmann, L. M. Richardson, L. M. Setzler, W. D. Behnke,
Biopolymers 1978, 17, 2107 2120.
[12] a) D. K. Mandal, N. Kishore, C. F. Brewer, Biochemistry 1994, 33,
1149 1156; b) J. H. Naismith, R. A. Field, J. Biol. Chem. 1996, 271,
972 976; c) R. Loris, D. Maes, F. Poortmans, L. Wyns, J. Bouckaert, J.
Biol. Chem. 1996, 271, 30614 30618.
[13] C. Emmerich, J. R. Helliwell, M. Redshaw, J. H. Naismith, S. J.
Harrop, J. Raftery, A. J. Kalb, J. Yariv, Z. Dauter, K. S. Wilson, Acta
Crystallogr. Sect. D 1994, 50, 749 756.
[14] a) I. J. Goldstein, C. E. Hollerman, E. E. Smith, Biochemistry 1965, 4,
876 883; b) Z. Derevenda, J. Yariv, J. R. Helliwell, A. J. Kalb, E. J.
Dodson, M. Z. Papiz, T. Wan, J. Campbell, EMBO J. 1989, 8, 2189
2193; c) J. H. Naismith, C. Emmerich, J. Habash, S. J. Harop, J. R.
Helliwell, W. N. Hunter, J. Raftery, A. J. Kalb (Gilboa), J. Yariv, Acta
Crystallogr. 1994, D50, 847 858; d) T. K. Dam, C. F. Brewer, Chem.
Rev. 2002, 102, 387 429.
[24] Y. Shikata, M. Kuwada, Y. Hayashi, A. Hashimoto, A. Koide, N.
Asakawa, Anal. Chim. Acta 1998, 365, 241 247.
[25] a) E. N. Hudson, G. Weber, Biochemistry 1973, 12, 4154 4161;
b) R. P. Haugland, Handbook of Fluorescent Probes and Research
Chemicals, 6th ed., Molecular Probes, Inc., Eugene, OR, 1996.
[26] H. Benesi, J. H. Hildebrand, J. Am. Chem. Soc. 1949, 71, 2703 2707.
[27] K. Hamasaki, H. Ikeda, A. Nakamura, A. Ueno, F. Toda, I. Suzuki, T.
Osa, J. Am. Chem. Soc. 1993, 115, 5035 5040.
[28] S. J. Pollack, G. R. Nakayama, P. G. Schultz, Science 1988, 242, 1038
1040.
[15] a) I. Hamachi, T. Nagase, S. Shinkai, J. Am. Chem. Soc. 2000, 122,
12065 12066; b) T. Nagase, S. Shinkai, I. Hamachi, Chem. Commun.
2001, 229 230.
[29] P. N. Kanellopoulos, K. Pavlou, A. Perrakis, B. Agianian, C. E.
Vorgias, C. Mavrommatis, M. Soufi, P. A. Tucker, S. J. Hamodrakas,
J. Struct. Biol. 1996, 116, 345 355.
[16] F. P. Schwarz, K. D. Puri, R. G. Bhat, A. Surolia, J. Biol. Chem. 1993,
268, 7668 7677.
[30] H. Lis, N. Sharon, Annu. Rev. Biochem. 1986, 55, 35 67.
[31] J. L. Wang, G. M. Edelman, J. Biol. Chem. 1978, 253, 3000 3007.
[32] G. L. Ellman, Arch. Biochem. Biophys. 1959, 82, 70 77.
[33] Since Mn2‡ and Ca2‡ ions are essential for the selective and tight
binding of sugars to Con A, we added these metal cations to the
solution both in the PPALM step and the fluorescence titration
process. Without them, we observed a decrease of the binding affinity.
[34] G. Weber, Adv. Protein Chem. 1953, 8, 415 459.
[17] For review articles, see: a) G. Dorman, G. D. Prestwich, Trends
Biotechnol. 2000, 18, 64 77; b) S. A. Fleming, Tetrahedron 1995, 51,
12479 12520; c) J. Brunner, Annu. Rev. Biochem. 1993, 62, 483 514;
d) D. G. Knorre, T. S. Godovikova, FEBS Lett. 1998, 433, 9 14.
[18] a) Y. Hatanaka, M. Hashimoto, Y. Kanaoka, J. Am. Chem. Soc. 1998,
120, 453 454; b) K. Fang, M. Hashimoto, S. Jockusch, N. J. Turro, K.
Nakanishi, J. Am. Chem. Soc. 1998, 120, 8543 8544; c) H. Li, Y. Liu,
K. Fang, K. Nakanishi, Chem. Commun. 1999, 365 366.
[19] a) M. Nassal, Liebigs Ann. Chem. 1983, 1510 1523; b) Y. Hatanaka,
M. Hashimoto, H. Kurihara, H. Nakayama, Y. Kanaoka, J. Org. Chem.
Received: March 7, 2003 [F4925]
Chem. Eur. J. 2003, 9, 3660 3669
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3669