An enantioselective fluorescent sensor for sugar acids

Article abstract of DOI:10.1021/ja046289s

Chiral fluorescent boronic acid 1 was found to be a highly enantioselective, chemoselective, and sensitive sensor for sugar acids, such as tartaric acid. Enantioselectivities (K_R/K_S) of up to 550:1, chemoselectivity up to 11 000:1, a

Full text of DOI:10.1021/ja046289s

Published on Web 11/16/2004
An Enantioselective Fluorescent Sensor for Sugar Acids
Jianzhang Zhao,† Matthew G. Davidson,† Mary F. Mahon,‡
Gabriele Kociok-Ko¨hn,‡ and Tony D. James*,†
Contribution from the Department of Chemistry and Bath Chemical Crystallography,
UniVersity of Bath, Bath BA2 7AY, U.K.
Received June 23, 2004; E-mail: t.d.james@bath.ac.uk
Abstract: Chiral fluorescent boronic acid 1 was found to be a highly enantioselective, chemoselective,
and sensitive sensor for sugar acids, such as tartaric acid. Enantioselectivities (KR/KS) of up to 550:1,
chemoselectivity up to 11 000:1, and sensitivities in the micromolar range with sensor 1 were observed.
Single-crystal X-ray analysis was used to confirm the structure of the fluorescent species.
Introduction
as has a boronic acid-based colorimetric indicator-displacement
assay for the determination of enantiomeric excess of R-hydroxy
Much recent attention has been paid to the development of
synthetic molecular receptors with the ability to recognize
selectively small molecules involved in biological pathways.
Fluorescent sensors are preferred because they are well suited
to meet the need for in vivo probes, such as mapping the spatial
and temporal distribution of the biological analytes.1,2 Boronic
acid molecular receptors for saccharides have attracted consider-
able interest due to their ability to bind saccharides in aqueous
media.3-6 In this case, the most common interaction is with
cis-1,2- or 1,3-saccharide diols to form five- or six-membered
rings with the boronic acid moiety. Over the past few years we
have been interested in developing new selective sensors for
saccharides employing a modular approach.7-12 Our idea was
to disconnect a sensor into three tuneable components: receptor,
linker, and “read-out” units. This approach can be illustrated
by the D-glucose-selective fluorescent sensor 3 (Scheme 1)
which contains two boronic acid units (binding units) and an
anthracene unit (acting as both a D-glucose selective linker and
the fluorophore read-out unit).13,14
acids.20 Hydrogen-bonding receptors for the binding of tartaric
acid,21 chiral discrimination of hydroxylcarboxylates,22 and
tartaric acid23,24 are also known.
Herein we report the tight and selective binding of the sugar
acids, D- and L-tartaric acid, D-glucaric acid, and D-gluconic
acid with the rationally designed chiral fluorescent sensors 1
and 2.
Most naturally occurring chemical species, including sugars
and sugar acids, are chiral compounds; therefore, a chiral sensor
is inherently better than an achiral one, since the chirality of
the analytes can be utilized as an extra differential factor to
enhance discrimination. In an attempt to address this need for
enantioselectivity we recently employed BINOL. The rigid axial
chirality of BINOL was key in the selective chiral sensor’s
construction (functioning as chirogenic center and fluoro-
phore).25 Unfortunately, in this case, the chiral center is not in
close proximity to the receptor’s binding site, and BINOL has
poor fluorescence properties. Therefore, we designed fluorescent
chiral sensor 1 which has two chiral centers in close proximity
to the binding site of the receptor and used anthracene, a good
fluorophore, as a rigid linker.
Fluorescent boronic acid-based sensors for tartaric acid,15,16
D-glucuronic acid,17,18 and D-glucaric acid19 have been reported,
† Department of Chemistry.
‡ Bath Chemical Crystallography.
Results and Discussion
(1) de Silva, A. P.; Gunaratne, H. Q. N.; Gunnlaugsson, T.; Huxley, A. J. M.;
McCoy, C. P.; Rademacher, J. T.; Rice, T. E. Chem. ReV. 1997, 97, 1515.
(2) Pu, L. Chem. ReV. 2004, 104, 1687.
The fluorescence response of the boronic acid sensor as well
as the binding of the carboxylic moiety, is pH-dependent;
therefore, the fluorescence-pH profiles of sensor 1 in the
(3) James, T. D.; Shinkai, S. Top. Curr. Chem. 2002, 218, 159.
(4) Wang, W.; Gao, X.; Wang, B. Curr. Org. Chem 2002, 6, 1285.
(5) Striegler, S. Curr. Org. Chem 2003, 7, 81.
(6) James, T. D.; Sandanayake, K. R. A. S.; Shinkai, S. Nature 1995, 374,
345.
(15) Lavigne, J. J.; Anslyn, E. V. Angew. Chem., Int. Ed. 1999, 38, 3666.
(16) Gray, C. W.; Houston, T. A. J. Org. Chem. 2002, 67, 5426.
(17) Takeuchi, M.; Yamamoto, M.; Shinkai, S. Chem. Commun. 1997, 1731.
(18) Yamamoto, M.; Takeuchi, M.; Shinkai, S. Tetrahedron 1998, 54, 3125.
(19) Yang, W.; Yan, J.; Fang, H.; Wang, B. Chem. Commun. 2003, 792.
(20) Zhu, J.; Anslyn, E. V. J. Am. Chem. Soc. 2004, 126, 3676.
(21) Goswami, S.; Ghosh, K.; Mukherjee, R. Tetrahedron 2001, 57, 4987.
(22) Tejeda, A.; Oliva, A. I.; Simon, L.; Grande, M.; Caballero, M.; Moran, J.
R. Tetrahedron Lett. 2000, 41, 4563.
(23) Hernandez, J. V.; Almaraz, M.; Raposo, C.; Martin, M.; A., L.; Crego,
M.; Cabellero, C.; Moran, J. R. Tetrahedron Lett. 1998, 39, 7401.
(24) Garcia-Tellado, F.; Albert, J.; Hamilton, A. D. J. Chem. Soc., Chem.
Commun. 1991, 1761.
(25) Zhao, J.; Fyles, T. M.; James, T. D. Angew. Chem., Int. Ed. 2004, 43,
3461.
(7) Arimori, S.; Bell, M. L.; Oh, C. S.; Frimat, K. A.; James, T. D. Chem.
Commun. 2001, 1836.
(8) Arimori, S.; Bell, M. L.; Oh, C. S.; Frimat, K. A.; James, T. D. J. Chem.
Soc., Perkin Trans. 1 2002, 803.
(9) Arimori, S.; Bell, M. L.; Oh, C. S.; James, T. D. Org. Lett. 2002, 4, 4249.
(10) Arimori, S.; Consiglio, G. A.; Phillips, M. D.; James, T. D. Tetrahedron
Lett. 2003, 44, 4789.
(11) Arimori, S.; Phillips, M. D.; James, T. D. Tetrahedron Lett. 2004, 45, 1539.
(12) Arimori, S.; Ushiroda, S.; Peter, L. M.; Jenkins, A. T. A.; James, T. D.
Chem. Commun. 2002, 2368.
(13) James, T. D.; Sandanayake, K.; Shinkai, S. Angew. Chem., Int. Ed. Engl.
1994, 33, 2207.
(14) James, T. D.; Sandanayake, K. R. A. S.; Iguchi, R.; Shinkai, S. J. Am.
Chem. Soc. 1995, 117, 8982.
9
10.1021/ja046289s CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 16179-16186
16179

A R T I C L E S
Zhao et al.
Scheme 1. Fluorescent Sensors for Monosaccharides or Sugar Acidsa
a Reagents and conditions: i, EtOH/THF (7:2), reflux, 6 h, 76.8% (R,R-4), 76.8% (S,S-4), 85.2% (R-6), 90.0% (S-6); ii, THF, NaBH4, 60.1% (R,R-5),
78.9% (S,S-5), 85.2% (R-7), 81.1% (S-7); iii, CH3CN, 2-(2-bromobenzyl)-1,3-dioxaborinane, reflux, 10 h, 66.4% (R,R-1)0.78.5% (S,S-1), 14.0% (R-2), 5.2%
(S-2).
Scheme 2. Analytes Used in the Study
presence of various analytes (Scheme 2) were determined since
this allows a rapid preview of the optimal pH region of the
sensor. That is, the appropriate pH value that will result in large
fluorescence changes. The results observed for sensor 1 with
D- and L-tartaric acid are shown in Figure 1a. In the presence
of D- and L-tartaric acid, sensor R,R-1 (or S,S-1) gave different
responses (Figure 1a). The apparent pKa value of S,S-1 is 4.82
( 0.04 and in the presence of D- or L-tartaric acid, this shifts to
7.62 ( 0.04 or 8.73 ( 0.04, respectively. While for R,R-1, the
pKa value of 4.81 ( 0.05 changes to 8.66 ( 0.05 or 7.61 (
0.05 respectively, in the presence of D- or L-tartaric acid (Figure
1a). The complementarities of the pKa values between enanti-
omers of sensor 1 and tartaric acid clearly demonstrate enan-
tioselective recognition by the boronic acid sensors for tartaric
acid. With the monoboronic acid sensor 2, however, no
enantioselectivity was observed (Figure 1b). The apparent pKa
value of R-2 is 5.91 ( 0.02, and in the presence of D- and
L-tartaric acid this shifts to 8.04 ( 0.02 and 8.14 ( 0.01,
9
16180 J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004

Enantioselective Fluorescent Sensor for Sugar Acids
A R T I C L E S
Figure 1. Fluorescence intensity-pH profile of sensors 1 and 2. (a) Bisboronic acid 1 with D- and L-tartaric acid, λex at 365 nm, λem at 429 nm. (b)
Monoboronic acid 2 with D- and L-tartaric acid, λex at 373 nm, λem at 421 nm. 3.0 × 10-6 mol dm-3 of sensors in 5.0 × 10-2 mol dm-3 NaCl ionic buffer
(52.1% methanol in water), [L- and D-tartaric acid] ) 5.0 × 10-2 mol dm-3, 22 °C.
Figure 2. Relative fluorescence intensity of sensors 1 and 2 vs concentration of D- or L-tartaric acid. (a) Bisboronic acid 1 with D- and L-tartaric acid, λex
at 365 nm, λem at 429 nm, pH 8.3. (b) Monoboronic acid 2 with D- and L-tartaric acid, λex at 373 nm, λem at 421 nm, pH 7.0. 3.0 × 10-6 mol dm-3 of sensors
in 5.0 × 10-2 mol dm-3 NaCl ionic buffer (52.1% methanol in water), 22 °C.
respectively. The pKa of S-2 is 5.89 ( 0.05, and in the presence
of D- and L-tartaric acid this shifts to 8.17 ( 0.06 and 8.23 (
0.06, respectively.
for this system, the enantioselectivity (KR:KS) is 490:1 for
D-tartaric acid and 1:550 for L-tartaric acid.
At pH 5.6 the enantioselectivity and relative fluorescence
enhancement is diminished. However, the binding of tartaric
acid with sensor 1 at pH 5.6 is 1000 times stronger than the
values obtained at pH 8.3. The binding constants (log K) of
D-tartaric acid with R,R-1 is 5.92 ( 0.17, compared to the value
of 2.79 ( 0.12 at pH 8.3. The improvement in the binding
constants is significant and desirable since the detection limit
can be improved from the millimolar to the micromolar range.
As the stability constants and fluorescence enhancement of
sensor 1 with D- or L-tartaric acid are highly enantioselective,
the ee of tartaric acid mixtures can be determined (Figure 4).
Also, since the dynamic range of the fluorescence response is
large the system is very sensitive. The range of ee chosen was
from 90 to 100 since this is the range of interest in monitoring
enantioselective reactions. The large fluorescence response with
sensor 1 makes it possible to detect ee differences of only 1%
(Figure 4).
Titrations of the sensors with tartaric acid were carried out
at pH 8.3, 7.0, and 5.6. At pH 8.3 (Figures 2 and 3) a significant
fluorescence enhancement was observed with S,S-1 (I/I0 ) 8.24)
in the presence of L-tartaric acid (Figures 2a and 3a), whereas
in the presence of D-tartaric acid (Figures 2a and 3b), only a
very small fluorescence enhancement was observed (I/I0 ) 1.5).
When R,R-1 was used, the fluorescence response mirrored that
observed with S,S-1, in that a large fluorescence enhancement
(I/I0 ) 8.87) was observed with D-tartaric acid (Figure 2a),
whereas in the presence of L-tartaric acid (Figure 2a) only a
small fluorescence response was observed (I/I0 ) 1.5). The
binding constants of the sensor with the D- or L-tartaric acid
are also significantly different (Table 1). For example, the
binding constants of R,R-1 with the D-tartaric acid is log KD
2.79 ( 0.12, while its binding constant with L-tartaric acid was
not determined due to the small changes in fluorescence.
The binding of the sensor 1 with other sugar acids, such as
glucaric acid, gluconic acid, glucuronic acid and galacturonic
acid as well as sorbitol and glucose were also investigated and
the binding constants, fluorescent enhancements and enantio-
At pH 7.0 the binding constants of R,R-1 with the D- and
L-tartaric acid were log KD 4.79 ( 0.07 and log KL 2.07 (
0.06, and the binding constants of S,S-1 with the D- and L-tartaric
acid were log KD 2.09 ( 0.06 and log KL 4.81 ( 0.06. Hence,
9
J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004 16181

A R T I C L E S
Zhao et al.
Figure 3. Normalized emission spectra of sensors S,S-1 and S-2 in the presence of tartaric acid. (a) S,S-1 with L-tartaric acid; (b) S,S-1 with D-tartaric acid,
λ
ex at 365 nm, pH 8.3; (c) S,S-2 with L-tartaric acid; (d) S-2 with D-tartaric acid, λex at 373 nm, pH 7.0. 3.0 × 10-6 mol dm-3 of sensors in 5.0 × 10-2 mol
dm-3 NaCl ionic buffer (52.1% methanol in water), 22 °C.
Table 1. Logarithm of 1:1 Stability Constants, Fluorescence Enhancement F on Binding, and Enantioselectivity (KR:KS) of Sensors R,R-1,
S,S-1, R-2, and S-2a
log K
Fb
analytes
pH
R,R-1
S,S-1
R,R-1
S,S-1
KR:KS
D-tartaric acid
5.6
7.0
8.3
5.6
7.0
8.3
7.0
5.6
8.3
5.6
8.3
5.6
8.3
5.6
8.3
5.6
8.3
5.92 ( 0.17
4.79 ( 0.07
2.79 ( 0.12
3.93 ( 0.13
4.00 ( 0.11
4.80 ( 0.12
9.14 ( 0.15
8.87 ( 0.22
3.71 ( 0.06
8.03 ( 0.12
1.5d
8.61 ( 0.12
4.80 ( 0.06
2.96 ( 0.15
3.77 ( 0.04
3.51 ( 0.12
4.86 ( 0.05
3.26 ( 0.12
4.97 ( 0.15
14.2 ( 0.76
5.07 ( 0.08
11.1 ( 0.13
3.68 ( 0.06
8.02 ( 0.14
1.5d
83:1
490:1
-
1:67
1:550
-
1:1.1
1:9.2
1:2.3
1:4.4
1:3.3
1:2.0
1:1.4
1.2:1
1.2:1
1:27
1:24
2.09 (0.06
c
-
L-tartaric acid
5.76 ( 0.16
4.81 ( 0.06
2.74 ( 0.10
4.71 ( 0.08
5.73 ( 0.13
2.68 ( 0.07
4.73 ( 0.04
2.93 ( 0.05
3.81 ( 0.04
3.01 ( 0.03
2.93 ( 0.04
1.38 ( 0.08
3.31 ( 0.09
3.30 ( 0.15
4.36 ( 0.14
9.42 ( 0.10
8.24 ( 0.21
8.33 ( 0.11
5.52 ( 0.11
9.24 ( 0.19
4.08 ( 0.03
7.62 ( 0.13
5.26 ( 0.19
14.5 ( 0.89
3.76 ( 0.04
15.8 ( 0.58
4.06 ( 0.06
9.51 ( 0.28
2.07 ( 0.06
c
-
DL-tartaric acid
D-glucaric acide
4.68 ( 0.08
4.76 ( 0.40
2.31 ( 0.20
4.09 ( 0.07
2.41 ( 0.15
3.52 ( 0.05
2.86 ( 0.19
3.02 ( 0.12
1.45 ( 0.12
1.87 ( 0.06
1.93 ( 0.04
D-gluconic acidf
D-glucuronic acid
D-galacturonic acid
D-glucose
D-sorbitol
R-2
S-2
R-2
S-2
D-tartaric acid
L-tartaric acid
5.6
7.0
5.6
7.0
3.47 ( 0.29
2.62 ( 0.03
3.47 ( 0.08
2.57 ( 0.04
3.48 ( 0.04
2.61 ( 0.04
3.54 ( 0.08
2.57 ( 0.03
1.70 ( 0.00
4.33 ( 0.03
1.75 ( 0.01
4.20 ( 0.04
1.80 ( 0.01
4.40 ( 0.04
1.74 ( 0.01
4.39 ( 0.02
1:1.0
1.0:1
1:1.2
1.0:1
a 5.0 × 10-6 mol dm-3 R,R-1 or S,S-1 in 0.05 mol dm-3 NaCl ionic buffer (52.1% methanol in water), λex ) 365 nm, λem ) 429 nm. 22 ( 1 °C.
Constants determined by fitting a 1:1 binding model to I/I0. Errors reported are two standard deviations (95% confidence limit); r2 ) 0.99 in most cases.
b Determined by 1:1 fitting. F values agree well with the experimental results. c Due to small changes in fluorescence accurate values could not be determined.
d Maximum fluorescence enhancement observed. e The potassium glucarate salt of glucaric acid was used in the titration. f The sodium gluconate salt of
gluconic acid was used in the titration.
selectivities (KR:KS) are listed in Table 1. The data in Table 1
indicates that this system is highly chemoselective for the sugar
acids. For example, the chemoselectivity of R,R-1 for D-glucaric
acid/D-gluconic acid is 9.9:1 (pH 5.6) and for D-glucaric acid/
9
16182 J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004

Enantioselective Fluorescent Sensor for Sugar Acids
A R T I C L E S
The best selectivity improvement is observed for R,R-1 and
D-tartaric acid/D-sorbitol, 7.2:1 (pH 8.3) and 11 000:1 (pH 5.6).
Such selectivity enhancements and in particular a switch in
selectivity with pH is highly desirable since in principle it will
allow the concentration of sugar acids in mixtures to be
determined.
From the data in Table 1, Figures 1b, 2b, 3c, and 3d, no
enantioselective discrimination between D- and L-tartaric acid
was observed when the monoboronic acid sensors R-2 and S-2
were used instead of the diboronic acid sensors R,R-1 and S,S-
1. This result indicates that for enantioselective discrimination
1:1 cyclic complexes must be formed between R,R-1 and S,S-1
and the sugar acids. The data in Table 1 also indicate that the
binding constants with the sugar acids are strongly pH-dependent
whereas those of the sugar alcohols are not.
Figure 4. Fluorescence intensity changes of S,S-1 vs enantiomeric
composition of D-tartaric acid. 3.0 × 10-6 mol dm-3 of S,S-1 in 5.0 ×
10-2 mol dm-3 M NaCl ionic buffer (52.1% methanol in water), λex 365
nm, λem 429 nm, pH 8.3, 22 °C. [D-tartaric acid] + [L-tartaric acid] ) 6.0
× 10-2 mol dm-3. The variation of the ee value of the D-tartaric acid was
achieved with mixing the D-tartaric acid and the L-tartaric acid.
Another advantage of these sensors is in utilizing their
chirality to provide chemoselectivity between the single enan-
tiomers of two different biological molecules. For example, at
pH 5.6 the chemoselectivity of S,S-1 for D-glucaric acid/D-
sorbitol is 260:1; however, by using R,R-1, the D-glucaric acid/
D-sorbitol chemoselectivity can be increased to 790:1. With
D-tartaric acid/D-sorbitol at pH 5.6 the chemoselectivity of S,S-1
is 4.9:1; however, by using R,R-1, the chemoselectivity is
increased to 11 000:1. With simple achiral sensors, such
optimization of sensitivity and selectivity is impossible.
D-sorbitol is 790:1. Moreover, the chemoselectivity of sensor 1
can be optimized by changing the pH of the solution. For
example, at pH 8.3, the chemoselectivity of R,R-1 for D-glucaric
acid/D-glucuronic acid is 1:3.5 but at pH 5.6 the selectivity
changes to 18:1. With S,S-1 the D-glucaric acid/D-glucuronic
acid selectivity changes from 1:2.2 (pH 8.3) to 82:1 (pH 5.6).
Figure 5. (a) Structure of R,R-1. (b) Structure of S,S-1 complex with L-tartaric acid. (c) Overlay for comparison of structure (a) (mirror image) and structure
(b).
9
J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004 16183

A R T I C L E S
Zhao et al.
controlled to within less than (0.03 unit from the desired pH. Titration
plots were generated using Origin 5.0 (Microcal software). The binding
constants were calculated using custom-written nonlinear least-squares
curve-fitting programs implemented within SigmaPlot 2000 (SPSS Inc.).
Single-Crystal Preparation and Crystallographic Methods. The
R,R-1 single crystal was obtained by dissolving the compound in
dichloromethane/methanol mixture (5:1) to give a concentrated solution.
Good quality crystals were obtained within about 2 weeks. The S,S-1
with L-tartaric acid single crystal was obtained by the two layer diffusion
method (sensor S,S-1 was dissolved in dichloromethane and the
L-tartaric acid was dissolved in methanol).
Crystallographic data for R,R-1 and S,S-1 with L-tartaric acid are
available as Supporting Information. Data were collected on a Nonius
KappaCCD diffractometer. Structure solution was performed using
SHELXS-86 [G. M. Sheldrick, SHELXS-86, Computer Program for
Crystal Structure Determination, University of Go¨ttingen, Germany,
1986], and refinement was by full-matrix least squares on F2 using
SHELXL-97 [G. M. Sheldrick, SHELXL-97, Computer Program for
Crystal Structure Refinement, University of Go¨ttingen, Germany, 1997]
software.
In R,R-1 the asymmetric unit consists of two boronic acid molecules
plus half a molecule of dichloromethane positionally disordered over
two sites in a 35:15 ratio. The boronic acid molecules are associated
via hydrogen bonding between the acidic hydrogen atoms (H1-H8)
of the boronic acid functionalities and the lone pairs on the tertiary
amine nitrogen atoms. H1-H8 were included at calculated positions
in the refinement, but their isotropic atomic displacement parameters
were refined without restraint.
In the S,S-1 with L-tartaric acid complex the asymmetric unit contains
one molecule of the boronate, plus two fragments of methanol, each
with 50% occupancy. Despite collecting four independent data sets for
this structure, on substantially sized, good quality crystals, a relatively
premature fall off in diffraction intensity was unavoidable. On refining
the structural model to the stage where the disordered methanol pre-
sented herein was identified, it became obvious that there was additional
partial occupancy (range 20-40%) solvent fragments located in the
lattice. It was also evident from early in the refinement procedure that
there is apparent void space in the lattice (confirmed by PLATON).
Thus, a data squeeze was implemented, and the refinement presented
here takes account of this. (R1 prior to this adjustment had converged
to 11.42%).
Although a large number of boronic acid PET sensors have
been studied, to our knowledge, no single-crystal X-ray data
for a fluorescent diboronic acid sensor bound with a guest have
been reported until now. With sensor 1, the structure of the 1:1
binding complex with tartaric acid was determined, as was the
structure of the unbound receptor (Figure 5). For the unbound
receptor, the hydrogen atoms of the boronic acid were located
and participate in O-H‚‚‚N hydrogen bonds with the nitrogen
atoms of the receptor framework (Figure 5a: N(1)‚‚‚O(1), 2.764
Å; N(2)‚‚‚O(3), 2.759 Å). For the tartrate complex (Figure 5b)
a methanol (CH3OH) is bound through its oxygen atom to the
boron center and also hydrogen bonds to the nitrogen atom of
the receptor framework (N(1)‚‚‚O(1), 2.655 Å; N(2)‚‚‚O(2),
2.693 Å). In this latter case the hydrogen atoms of the methanol
could not be located and refined, but their position is inferred
by the similarity in geometry between the unbound and bound
receptors (Figure 5c). These structural data clearly demonstrate
that binding of the sugar acid to the boronic acid does not result
in the formation of a strong B-N bond (the molecular property
believed to account for the fluorescence enhancement in this
type of sensor). These structural observations agree with a recent
computational study which indicated that the B-N bond was
weak (13 kJ mol-1 in the absence of solvent).26
Experimental Section
General. All the solvents and chemicals were supplied by Frontier
Scientific Ltd., Aldrich Chemical Co. Ltd., Lancaster Synthesis Ltd.,
and Fisher Scientific Ltd. Dry solvents for synthesis were freshly
1
distilled over suitable drying agents prior to use. H and 13C NMR
spectra were recorded on a Bruker AVANCE 300 (300.13 and 75.47
MHz respectively) spectrometer. All the chemical shifts (δ) are reported
in ppm using the deuterated solvent and or tetramethylsilane as the
internal reference. The mass spectra were recorded on a Waters
Micromass Autospec Spectrometer using fast atom bombardment (liquid
secondary ion mass spectrometry) using 3-nitrobenzyl alcohol (NOBA)
as the matrix liquid. Elemental Analyses were performed on an Exeter
Analytical CE 440. Thin-layer chromatography (TLC) was performed
on precoated aluminum-backed silica or alumina plates supplied by
Fluka Chemie. Visualization was achieved by UV light (254 nm). The
specific rotation of the chiral sensors ([R]22D) was measured with a
AA-10 automatic polarimeter (Optical Activity Ltd., England) using a
10 cm cell.
Fluorescence spectra were measured on a Perkin-Elmer LS 50B
luminescence spectrometer. Data were collected and exported via the
Pekin Elmer FL WinLab 4.00 software package. The fluorescence
intensity changes of bisboronic acid 1 was monitored at 429 nm (λex
365 nm). For monoboronic acid 2, the fluorescence was monitored at
421 nm (λex 373 nm). A 0.05 mol dm-3 NaCl (52.1% methanol, w/w)
ionic buffer was used in all the experiments. The final concentrations
of the sensors were fixed at 5.0 ×10-6 mol dm-3 (by dilution of the
stock solution into the buffer). All pH measurements were recorded
on a Hanna Instruments HI 9321 Microprocessor pH meter which was
routinely calibrated using Fisher Chemicals standard buffer solutions
(pH 4.0 for phthalate, 7.0 for phosphate, and 10.0 for borate).
pH Titration. The fluorescence emission spectra of the sensors with
or without the analytes were recorded as the pH was changed from pH
2 to 12 in approximate intervals of 0.5 pH unit. The pH of the solution
was controlled using minimum amount of sodium hydroxide and
hydrochloric acid solutions. For every concentration or pH data point
the fluorescence was recorded three times and the mean values were
used for the final curves.
Synthesis. S,S-(+)-Bisboronic Acid (1). 9,10-Anthracenedicarbox-
aldehyde (2.00 g, 6.80 mmol) and (S)-(-)-R-methylbenzylamine (3.00
g, 24.8 mmol) were dissolved in 90 mL EtOH/THF (7:2) and a few
drops of acetic acid were added as catalyst. The mixture was heated
with stirring at reflux for 6 h under an N2 atmosphere. The mixture
was then cooled to room temperature and filtered under vacuum to
1
give S,S-4 (2.30 g, 76.8%) as yellow needles. H NMR (300 MHz,
CDCl3) δ 9.44 (s, 2H), 8.40 (m, 4H), 7.63-7.33 (m, 14H), 4.90 (q,
2H, J ) 6.0 Hz), 1.87 (d, 6H, J ) 6.0 Hz); 13C NMR (75 MHz, CDCl3)
δ 159.4, 145.2, 131.2, 129.8, 129.1, 127.6, 127.3, 126.7, 125.7, 72.2,
25.7. S,S-4 was used without further purification.
S,S-4 (1.13 g, 2.56 mmol) was dissolved in dry THF, and then NaBH4
(0.96 g, 25.4 mmol) was added and the mixture stirred at room
temperature until the reduction was finished (the reaction was monitored
by TLC using alumina plates with dichloromethane as eluent). The
solvent was removed and the residue dissolved in dichloromethane then
washed with brine. The organic phase was dried over anhydrous MgSO4
and the solvent removed under vacuum to give S,S-5 (0.900 g, 78.9%)
1
as a light-yellow solid. H NMR (300 MHz, CDCl3) δ 8.16 (m, 4H),
7.55-7.27 (m, 14H), 4.44 (d, 2H, J ) 12.0 Hz), 4.40 (d, 2H, J ) 12.0
Hz), 4.08 (q, 2H, J ) 6.0 Hz), 1.41 (d, 6H, J ) 6.0 Hz); 13C NMR (75
MHz, CDCl3) δ 146.1, 132.5, 130.5, 129.0, 127.6, 127.4, 125.9, 125.3,
59.7. 44.6, 25.1. S,S-5 was used without further purification.
Titration with Analytes. The fluorescence spectra of the sensors
in the presence of analytes were recorded as increasing amounts of
analytes were added to the solution. For all titrations the final pH was
(26) Franzen, S.; Ni, W.; Wang, B. J. Phys. Chem. B 2003, 107, 12942.
9
16184 J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004

Enantioselective Fluorescent Sensor for Sugar Acids
A R T I C L E S
S,S-5 (0.95 g, 2.14 mmol) and 2-(2-bromomethyl)-1,3,2-dioxabori-
nane (1.75 g, 6.86 mmol) were dissolved in 30 mL of dry acetonitrile,
and then K2CO3 (1.18 g, 8.54 mmol) was added and the mixture heated
under reflux with stirring for 10 h. The solvent was evaporated and
water added to the residue, the resultant heterogeneous mixture was
then extracted with dichloromethane. The organic phases were recom-
bined and dried over MgSO4 and the solvent removed to give a dark-
yellow solid. The solid was dissolved in acetone and water was added
dropwise with stirring until a light-yellow precipitate formed. The solid
was vacuum filtered and washed with an acetone/water mixture to give
9.0 Hz), 1.76 (d, 3H, J ) 9.0 Hz); 13C NMR (75 M Hz, CDCl3) δ
159.3, 145.3, 131.7, 130.4, 129.6, 129.2, 129.1, 129.0, 128.7, 127.5,
127.2, 127.0, 126.3, 125.6, 125.2, 72.2, 25.8. S-6 was used without
further purification.
S-6 (1.90 g, 6.14 mmol) was dissolved in 50 mL of dry THF, and
then NaBH4 (1.13 g, 30.0 mmol) was added and the mixture was stirred
at room temperature until the reduction was finished. The solvent was
removed and the residue dissolved in dichloromethane and washed with
brine. The organic phase was dried over anhydrous MgSO4 and the
solvent removed under vacuum to give S-7 (1.55 g, 81.1%) as a light
yellow oil which slowly solidified at room temperature. 1H NMR (300
MHz, CDCl3) δ 8.22 (s, 1H), δ 8.00 (m, 2H), δ 7.84 (m, 2H), 7.43-
7.19 (m, 9H), 4.40 (d, 1H, J ) 12.0 Hz), 4.32 (d, 1H, J ) 12.0 Hz),
3.94 (q, 1H, J ) 6.0 Hz), 1.41 (d, 3H, J ) 6.0 Hz); 13C NMR (75
MHz, CDCl3) δ 146.2, 132.4, 132.1, 130.7, 129.6, 129.1, 127.7, 127.6,
127.5, 126.4, 125.1, 124.7, 59.8, 44.6, 25.1. S-7 was used without further
purification.
1
S,S-1 (1.20 g, 78.5%) as a light-yellow powder. H NMR (300 MHz,
CDCl3, and a few drops of CD3OD) δ 7.75-7.09 (m, 26H), 4.56 (d,
2H, J ) 15.0 Hz), 4.41 (d, 2H, J ) 12.0 Hz), 3.95 (q, 2H, J ) 6.0
Hz), 3.67 (s, 4H), 1.59 (d, 6H, J ) 6.0 Hz); m/z (FAB) 1252.0 ([M +
4NOBA - 4H2O]+, 7%). Anal. Calcd for C46H46B2N2O4‚0.5H2O: C,
22
76.6; H, 6.57; N, 3.88. Found: C, 76.5, H, 6.49; N, 3.78. [R]D
+28 ( 1° (c ) 1.0, CH3OH).
)
R,R-(-)-Bisboronic Acid (1). 9,10-Anthracenedicarboxaldehyde
(2.00 g, 6.80 mmol) and (R)-(+)-R-methylbenzylamine (3.00 g, 24.8
mmol) were dissolved in 90 mL of EtOH/THF (7:2) and a few drops
of acetic acid were added as catalyst. The mixture was heated under
reflux with stirring for 6 h under an N2 atmosphere. The mixture was
then cooled to room temperature and filtered under vacuum to give
R,R-4 (2.30 g, 76.8%) as yellow needles. 1H NMR (300 MHz, CDCl3)
δ 9.37 (s, 2H), 8.26 (m, 4H), 7.52-7.14 (m, 14H), 4.79 (q, 2H, J )
6.0 Hz), 1.74 (d, 6H, J ) 6.0 Hz); 13C NMR (75 MHz, CDCl3) δ 159.5,
145.1, 131.2, 129.8, 129.1, 127.6, 127.2, 126.7, 125.7, 72.2, 25.6. R,R-4
was used without further purification.
S-7 (1.30 g, 4.17 mmol) and 2-(2-bromomethyl)-1,3,2-dioxaborinane
(1.17 g, 4.59 mmol) were dissolved in 50 mL of dry acetonitrile, and
then K2CO3 (0.63 g, 4.56 mmol) was added and the mixture was stirred
under reflux overnight. The solvent was evaporated and water added
to the residue, the resultant heterogeneous mixture was then extracted
with dichloromethane. The organic phases were recombined and dried
over anhydrous MgSO4 and the solvent was removed. The solid was
then dissolved in acetone and water was added dropwise with stirring
until a light-yellow precipitate formed. The solid was vacuum filtered
and washed with an acetone/water mixture to give S-1 (96 mg, 5.2%)
as a light yellow powder. 1H NMR (300 MHz, CDCl3) δ 8.31 (s, 1H),
7.85 (d, 2H, J ) 9.0 Hz), 7.72 (d, 2H, J ) 6.0 Hz), 7.39-7.17 (m,
14H), 4.55 (d, 1H, J ) 12.0 Hz), 4.49 (d, 1H, J ) 18.0 Hz), 3.94 (q,
1H, J ) 6.0 Hz), 3.87 (d, 1H, J ) 15.0 Hz), 3.94 (d, 1H, J ) 12.0
Hz), 1.64 (d, 3H, J ) 6.0 Hz); 13C NMR (75 MHz, CDCl3) δ 142.0,
141.2, 136.9, 132.2, 131.7, 131.5, 130.7, 130.0, 129.6, 128.9, 128.3,
127.7, 126.3, 125.1, 124.7, 59.4, 57.3, 44.4, 31.3; m/z (FAB) 716.2
([M + 2NOBA - 2H2O]+, 30%). Anal. Calcd for C30H28BNO2: C,
R,R-4 (1.57 g, 3.56 mmol) was dissolved in 40 mL of dry THF, and
then NaBH4 (1.70 g, 44.9 mmol) was added and the mixture stirred at
room temperature until the reduction was finished (the reaction was
monitored using TLC alumina plates with dichloromethane as eluent).
The solvent was removed and the residue dissolved in dichloromethane
and washed with brine. The organic phase was dried over anhydrous
MgSO4 and the solvent removed under vacuum, to give R,R-5 (0.950
g, 60.1%) as a light-yellow solid. 1H NMR (300 MHz, CDCl3) δ 8.07
(m, 4H), 7.47-7.27 (m, 14H), 4.44 (d, 2H, J ) 12.0 Hz), 4.35 (d, 2H,
J ) 12.0 Hz), 3.99 (q, 2H, J ) 6.0 Hz), 1.68 (b, 2H); 1.32 (d, 6H, J
) 6.0 Hz); 13C NMR (75 MHz, CDCl3) δ 146.1, 132.5, 130.5, 128.9,
127.7, 127.4, 125.9, 125.3, 59.7, 44.6, 25.1. R,R-5 was used without
further purification.
80.9; H, 6.34; N, 3.15. Found: C, 80.5, H, 6.31; N, 3.12. [R]22
+56 ( 1° (c ) 1.0, CH3OH).
)
D
R-(-)-Monoboronic Acid (2). 9-Anthracenecarboxaldehyde (2.00
g, 9.70 mmol) and (S)-(-)-R-methylbenzylamine (1.76 g, 14.5 mmol)
were dissolved in 50 mL of EtOH/THF (7:2) and a few drops of acetic
acid were added as catalyst. The mixture was heated under reflux with
stirring for 6 h under an N2 atmosphere. The mixture was then cooled
to room temperature and filtered under vacuum to give R-6 (2.00 g,
66.6%) as yellow needles. 1H NMR (300 MHz, CDCl3) δ 9.35 (s, 1H),
8.31 (m, 3H), 7.94 (m, 2H), 7.17-7.49 (m, 9H), 4.72 (q, 1H, J ) 6.0
Hz), 1.69 (d, 3H, J ) 6.0 Hz); 13C NMR (75 MHz, CDCl3) δ 159.3,
145.4, 131.7, 130.5, 129.2, 129.1, 127.5, 127.3, 127.1, 125.7, 125.3,
72.3, 25.8. R-6 was used without further purification.
R-6 (1.90 g, 6.14 mmol) was dissolved in 50 mL of dry THF, and
then NaBH4 (1.13 g, 30.0 mmol) was added and the mixture stirred at
room temperature until the reduction was finished. The solvent was
removed and the residue dissolved in dichloromethane and washed with
brine. The organic phase was dried over anhydrous MgSO4 and the
solvent removed under vacuum to give R-7 (1.63 g, 85.2%) as a light
yellow oil which slowly solidified at room temperature. 1H NMR (300
MHz, CDCl3) δ 8.23 (s, 1H), δ 8.01 (m, 2H), 7.83 (m, 2H), 7.43-
7.21 (m, 9H), 4.40 (d, 1H, J ) 12.0 Hz), 4.32 (d, 1H, J ) 12.0 Hz),
3.94 (q, 1H, J ) 6.0 Hz), 1.29 (d, 3H, J ) 6.0 Hz); 13C NMR (75
MHz, CDCl3) δ 146.2, 132.4, 132.1, 130.7, 129.6, 129.1, 127.7, 127.6,
127.4, 126.4, 125.4, 124.7, 59.8, 44.6, 25.1. R-7 was used without
further purification.
R,R-5 (0.95 g, 2.14 mmol) and 2-(2-bromomethyl)-1,3,2-dioxabori-
nane (1.75 g, 6.86 mmol) were dissolved in 30 mL of dry acetonitrile,
and then K2CO3 (1.18 g, 8.54 mmol) was added and the mixture heated
under reflux with stirring for 10 h. The solvent was evaporated and
water added to the residue, the resultant heterogeneous mixture was
then extracted with dichloromethane. The organic phases were recom-
bined and dried over MgSO4 and the solvent was removed. The solid
was then dissolved in acetone and water was added dropwise with
stirring until a light-yellow precipitate formed. The solid was vacuum
filtered and washed with an acetone/water mixture to give R,R-1 (1.01
1
g, 66.4%) as a light-yellow powder. H NMR (300 MHz, CDCl3 and
a few drops of CD3OD) δ 7.72-7.06 (m, 26H), 4.55 (d, 2H, J ) 12.0
Hz), 4.39 (d, 2H, J ) 15.0 Hz), 3.94 (q, 2H, J ) 6.0 Hz), 3.78 (s, 4H),
1.58 (d, 6H, J ) 6.0 Hz); m/z (FAB) 1252.1 ([M + 4NOBA - 4H2O]+,
7%). Anal. Calcd for C46H46B2N2O4‚H2O: C, 75.6; H, 6.62; N, 3.83.
Found: C, 75.3, H, 6.54; N, 3.74. [R]22 ) -27 ( 1° (c ) 1.0,
D
CH3OH).
S-(+)-Monoboronic Acid (2). 9-Anthracenecarboxaldehyde (2.00
g, 9.70 mmol) and (S)-(-)-R-methylbenzylamine (1.76 g, 14.5 mmol)
were dissolved in 50 mL of EtOH/THF (7:2) and few drops of acetic
acid were added as catalyst. The mixture was heated under reflux with
stirring for 6 h under an N2 atmosphere. The mixture was then cooled
to room temperature and filtered under vacuum to give S-6 (2.70 g,
R-7 (1.30 g, 4.17 mmol) and 2-(2-bromomethyl)-1,3,2-dioxaborinane
(1.17 g, 4.59 mmol) were dissolved in 50 mL of dry acetonitrile, and
then K2CO3 (0.63 g, 4.56 mmol) was added and the mixture heated
under reflux with stirring overnight. The solvent was evaporated and
water added to the residue, the resultant heterogeneous mixture was
1
90.0%) as yellow needles. H NMR (300 M Hz, CDCl3) δ 9.44 (s,
1H), 8.40 (m, 3H), 7.94 (m, 2H), 7.17-7.53 (m, 9H), 4.79 (q, 1H, J )
9
J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004 16185

A R T I C L E S
Zhao et al.
then extracted with dichloromethane. The organic phases were com-
bined, dried over anhydrous MgSO4 and the solvent removed. The solid
was then dissolved in acetone and water was added dropwise with
stirring until a light-yellow precipitate formed. The solid was vacuum
filtered and washed with an acetone/water mixture to give R-1 (0.260
sugar acids, such as tartaric acid, glucaric acid, and gluconic
acid. The enantioselectivity is as high as 550:1 and the detection
limit for some sugar acids is in the micromolar range. Sensor 1
with high sensitivity, chemoselectivity, and enantioselectivity
has clearly demonstrated the value of the modular approach to
the construction of chiral sensors. These results, together with
the unambiguous structural data we report, should aid further
development and application of enantioselective chemical
sensors.
1
g, 14.0%) as a light yellow powder. H NMR (300 MHz, CDCl3) δ
8.31 (s, 1H), 7.86 (d, 2H, J ) 6.0 Hz), 7.72 (d, 2H, J ) 6.0 Hz),
7.40-7.17 (m, 14H), 4.55 (d, 1H, J ) 15.0 Hz), 4.49 (d, 1H, J ) 15.0
Hz), 4.04 (q, 1H, J ) 6.0 Hz), 3.88 (d, 1H, J ) 12.0 Hz), 3.77 (d, 1H,
J ) 12.0 Hz), 1.64 (d, 3H, J ) 9.0 Hz) ppm; 13C NMR (75 MHz,
CDCl3) δ 141.9, 141.2, 136.9, 132.2, 131.7, 131.5, 130.7, 130.0, 129.6,
128.9, 128.3, 127.7, 126.3, 125.2, 124.7, 59.4, 57.3, 44.4, 31.3; m/z
(FAB) 715.9 ([M + 2NOBA - 2H2O]+, 30%). Anal. Calcd for C30H28-
BNO2: C, 80.9; H, 6.34; N, 3.15. Found: C, 80.6, H, 6.23; N, 3.18.
Acknowledgment. T.D.J. and J.Z. wish to acknowledge the
University of Bath and Beckman-Coulter for support.
Supporting Information Available: Original titration curves
for the entries in Table 1 and X-ray crystallographic data (PDF,
CIF). This material is available free of charge via the Internet
at http://pubs.acs.org.
[R]22 ) -57 ( 1° (c ) 1.0, CH3OH).
D
Conclusions
In conclusion the chiral fluorescent sensor 1 was found to
be highly sensitive, chemoselective, and enantioselective to
JA046289S
9
16186 J. AM. CHEM. SOC. VOL. 126, NO. 49, 2004