Single-Molecule Determination of the Isomers of d-Glucose and d-Fructose that Bind to Boronic Acids

Article abstract of DOI:10.1002/anie.201712740

Monosaccharides, such as d-glucose and d-fructose, exist in aqueous solution as an equilibrium mixture of cyclic isomers and can be detected with boronic acids by the reversible formation of boronate esters. The engineering of accurate, discriminating and continuous monitoring devices relies on knowledge of which cyclic isomer of a sugar binds to a boronic acid receptor. Herein, by monitoring fluctuations in ionic current, we show that an engineered α-hemolysin (αHL) nanopore modified with a boronic acid reacts reversibly with d-glucose as the pyranose isomer (α-d-glucopyranose) and d-fructose as either the furanose (β-d-fructofuranose) or the pyranose (β-d-fructopyranose). Both of these binding modes contradict current binding models. With this knowledge, we distinguished the individual sugars in a mixture of d-maltose, d-glucose, and d-fructose.

Full text of DOI:10.1002/anie.201712740

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Single-molecule determination of the isomers of D-glucose and
D-fructose that bind to boronic acids
Authors: William Jamieson Ramsay and Hagan Bayley
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.201712740
Angew. Chem. 10.1002/ange.201712740
Link to VoR: http://dx.doi.org/10.1002/anie.201712740
http://dx.doi.org/10.1002/ange.201712740

10.1002/anie.201712740
Angewandte Chemie International Edition
COMMUNICATION
Single-molecule determination of the isomers of D-glucose and
D-fructose that bind to boronic acids
William J. Ramsay and Hagan Bayley*[a]
Abstract: Monosaccharides, such as D-glucose and D-fructose,
exist in aqueous solution as an equilibrium mixture of cyclic isomers
and can be detected with boronic acids by the reversible formation of
boronate esters. The engineering of accurate, discriminating and
continuous monitoring devices relies on knowledge of which cyclic
isomer of a sugar binds to a boronic acid receptor. Here, by
monitoring fluctuations in ionic current, we show that an engineered
α-hemolysin (αHL) nanopore modified with a boronic acid reacts
reversibly with D-glucose as the pyranose isomer (α-D-
glucopyranose) and D-fructose as either the furanose (β-D-
fructofuranose) or the pyranose (β-D fructopyranose). Both of these
binding modes contradict current binding models. With this
knowledge, we distinguished the individual sugars in a mixture of D-
maltose, D-glucose and D-fructose.
protein pore. Individual bond-making and bond-breaking events
are tracked by monitoring the change in ionic current flowing
through the nanoreactor, often an α-hemolysin (αHL) protein
pore.[9] By this means, a variety of aqueous, reversible covalent
reactions have been resolved with single-molecule resolution,
including the formation and cleavage of disulfide bonds,[10] the
creation and inversion of chiral centers,[11] in a multicomponent
network,[12] and the continuous, stochastic motion of a small-
molecule walker.[13] Here, we show that the nanoreactor
approach can be used to observe the reversible reaction
between a boronic acid and diol. This has enabled us to
determine the isomeric structures of sugars that bind to a
boronic acid from equilibrium mixtures.
The nanoreactor used here was an αHL heteroheptamer
consisting of one copy of the single-cysteine mutant T117C and
six copies of the wild-type (WT) protein (WT)6(T117C)1 (referred
to as PC, Figure 1a) (Section 1.2 in the Supporting Information).
The presence of only one cysteine residue in the β barrel of the
PC pore allowed a single boronic acid group to be introduced into
the nanoreactor for single-molecule diol detection. Maleimide 1
(3-(maleimide)phenylboronic acid) (Figure S2 in the Supporting
Information) was reacted with the Cys side-chain at position 117
to form a thioether adduct (Figure 1b). When 1 entered the 
barrel of PC from the trans compartment, at pH 7.0 in 2 M KCl
under an applied potential of −50 mV, a step decrease in the
negative current was observed, after a delay, which was
attributed to the formation of the adduct, PB: ΔI = +4.7 ± 0.3 pA
(number of experiments n = 20), over the unreacted pore current
of −77 ± 2 pA (n = 20) (Figure S3). 1 did not react with WT7
pores (Figure S4).
The ability of the PB pore to sense individual diol molecules
was first examined with 1,2-dihydroxybenzene (catechol) (Figure
1b). The addition of catechol to the trans compartment resulted
in the fluctuation of the ionic current between two discrete levels
separated by ΔI = +3.1 ± 0.1 pA (Figure 1c), where ΔI is the
difference between the current carried by the pore with bound
catechol (level 2) and that of the unoccupied pore (level 1). The
events associated with catechol were not seen with unmodified
(WT)6(T117C)1 pores (Figure S5). We also did not observe
binding events when phenol was added to PB from the trans
compartment (Figure S6), suggesting two hydroxyl groups are
necessary to form an adduct. In a kinetic analysis of the binding
events of catechol to PB, we assigned level 2 to the boronate
ester complex and assumed that the catechol concentration
inside the pore equals that in solution;[9] nanopore
confinement[14] in the αHL pore has little effect (<10-fold) on the
rates of reactions of uncharged small molecules.[9] The
reciprocal of the mean inter-event interval (on) is proportional to
The reversible, covalent interactions of boronic acids with diols
under aqueous conditions[1] have become a lynchpin in the
development of receptors for sensing biologically important 1,2-
diols,[2] including saccharides.[3] For example, boronic acid-
containing hydrogels have been used recently for biocompatible
continuous glucose monitoring.[4] Despite the growing number of
sophisticated boronic acid-based chemosensors that can detect
and discriminate sugars,[5] knowledge of the structural
interactions between saccharides and boronic acids remains
limited.
In aqueous solution, monosaccharides such as D-fructose
and D-glucose exist as mixtures of pyranose and furanose
isomers, 6- and 5-membered rings, each with an α- and β-
anomer. Due to the syn-periplanar pair of hydroxyl groups
present in the glucose isomer α-D-glucofuranose, which is in an
optimal orientation for boronic acid binding, it is generally
accepted that this isomer binds to the boronic acid, despite
making up <1% of the total composition in solution at
equilibrium.[3a, 6] Similarly, the fructose isomer β-D-fructofuranose,
which has ideal syn hydroxyl groups and is present as ~25% of
the isomer population in solution,[7] is the proposed binding
isomer. The stronger binding affinity of D-fructose to boronic
acids over D-glucose has been attributed to the higher fraction
of the optimal binding isomer in solution.[3a] However, ensemble
NMR measurements[8] that attempt to characterize sugar-boronic
acid interactions are hampered by the complexity of the mixture
of isomers that form.
Single-molecule approaches to chemical reaction analysis
provide information that is hidden or not accessible when
ensembles are studied. In the nanoreactor approach, small
molecule analytes form covalent bonds with a single sensing
group that has been engineered onto the wall of a modified
the catechol concentration, which is consistent with
a
bimolecular interaction for which 1/on = kon[catechol]. In contrast,
a plot of the reciprocal of the mean lifetime of the complex (off)
versus the catechol concentration has a near zero slope, which
is consistent with a unimolecular dissociation mechanism (1/off
= koff) (Figure 1d). The values of on and off were determined by
fitting dwell-time histograms for each catechol concentration to
[a]
Dr. W. J. Ramsay, Prof. H. Bayley
Department of Chemistry
University of Oxford
Oxford OX1 3TA
E-mail: hagan.bayley@chem.ox.ac.uk
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/anie.201712740
Angewandte Chemie International Edition
COMMUNICATION
single exponential functions. The forward and reverse rate
a bimolecular interaction between D-glucose and the boronic
acid; the binding affinity (Ka) of 1.5 ± 0.1 × 10-1 M-1 was weak
(Table 1). Although D-glucose exists in solution as an
equilibrium mixture between pyranose and furanose cyclic
isomers, both with α- and β-anomers (Figure 2g), single-
molecule analysis of the binding events of D-glucose suggested
that one isomer was reacting with the boronic acid sensor as
only one dissociation rate constant was observed. To identify
this isomer, we compared the single-molecule binding data of D-
glucose to other sugars and diols with known conformations in
solution.
constants derived from the
 values (Table 1) yield an
equilibrium association constant of catechol to the boronic acid
of Ka = 2.4 ± 0.2 × 103 M-1 at 21 ± 1 °C (n = 3), consistent with
literature values.[5] Short, spike events of varying amplitudes
observed in the baseline of single-channel recordings with a PB
pore (Figure S6), which can be seen in addition to the dominant
binding events of catechol in Figure 1c, were attributed to
intrinsic fluctuations within the pore and were neglected during
analysis.
Figure 2. Single-channel recordings for various diols, added to the trans
compartment (a – f). All traces are displayed on the same scale. Concentration
of diols: (a – c, e, f) 12 mM; (d) 8 mM. Conditions: applied potential −50 mV;
buffer: 2 M KCl, 100 mM MOPS (pH 7.0) in both chambers. g) Scheme
showing the binding of the α-D-glucopyranose isomer (cyan) to the PB pore
from an equilibrium mixture. The percentages of each isomer in the population
are shown.[6]
Figure 1. a) Ribbon representation of the PC pore in a lipid bilayer. PC contains
a single Cys residue (blue) at position 117 in one of the seven subunits
(green). b) Reaction of the thiol in PC (Cys-117) with 1 to irreversibly form a
thioether adduct by a Michael addition. The modified pore (PB) reversibly
forms boronate esters with catechol. c) Single-channel recordings at −50 mV
with 2 M KCl, 10 mM MOPS (pH 7.0) in both compartments, and 0 to 0.6 mM
catechol in the trans compartment. The two current levels correspond to the
unoccupied pore (1) and to the state with catechol attached to the boronic acid
(2). d) Plots of the reciprocals of the mean inter-event intervals (on) and the
dwell times of catechol (off) versus the catechol concentration.
The addition of D-maltose, a disaccharide of two glucose
units, to the PB pore resulted in reversible current blockades
(15.6 ± 0.7 pA) with twice the amplitude of the blockades
recorded with D-glucose (Figure 2b, S10). The mean life-time of
the events (5.2 ± 0.6 ms) and the affinity of D-maltose (Ka
=
1.3 ± 0.2 × 10-1 M-1) were nearly identical to the values for D-
glucose (Table 1). As the two glucose units in D-maltose are in
the pyranose form,[15] and koff for the dissociation of D-maltose
and D-glucose are the same, it is likely that a pyranose isomer
of D-glucose binds to the boronic acid. Further support for this
argument was provided by comparing the binding kinetics of D-
glucose to cis-1,2,-cyclohexanediol (Figure 2c, S12) and cis-1,2-
cyclopentanediol (Figure 2d, S14), which can be regarded as
model compounds for pyranose and furanose binding,
respectively. The rigid cyclopentane ring fixes the adjacent
hydroxyl groups on one side of the ring, whereas the
We also observed binding of D-glucose to the PB pore
(Figure 2a). Similarly to catechol, the addition of D-glucose to
the trans compartment resulted in reversible current blockades
(ΔI = +8.1 ± 0.6 pA) with a short mean life-time (5.2 ± 0.3 ms);
again, PC pores without the boronic acid sensor did not bind D-
glucose (Figure S7). The association rates show a first-order
dependence on the concentration of D-glucose (Figure S8),
while dissociation is independent of the concentration, indicating
This article is protected by copyright. All rights reserved.

10.1002/anie.201712740
Angewandte Chemie International Edition
COMMUNICATION
conformational flexibility of cyclohexane allows limited
movement of the cis hydroxyls. We found that the 6-membered
ring diol bound to the boronic acid with koff (118 ± 9 s-1) and Ka
(2.9 ± 0.4 × 10-1 M-1) values similar to those of D-glucose and D-
maltose (Table 1), consistent with our assignment of the binding
forms of these sugars as the pyranose isomers. Further, the koff
of the 5-membered ring diol (2.40 ± 0.09 s-1) was nearly two
orders of magnitude lower than that of D-glucose and this diol
bound with a much stronger affinity (Ka = 2.7 ± 0.1 × 101 M-1)
(Table 1); we suggest that the 5-membered ring diol forms a
more stable boronate ester than the 6-membered ring diol, thus
giving rise to the longer life-time. Therefore, our results indicate
that D-glucose binds to a boronic acid as the pyranose isomer,
and not as the furanose isomer as currently proposed.[3a]
We next examined the binding of D-fructose to the PB pore
(Figure 3, S19). Similar to the binding of D-glucose, the addition
of D-fructose to the trans compartment resulted in reversible
current blockades (ΔI = +7.5 ± 0.3 pA) (Figure 3a). However, the
dwell-time histogram of the current blockades, which was
obtained from idealized current traces by using QuB software
(QUB 2.0, University at Buffalo), was well fitted by a two-
component probability density function that yielded mean dwell-
times of 143 ± 8 ms and 4762 ± 454 ms[16] (Figure 3b). This
indicated that two different isomers in the equilibrium mixture of
D-fructose were stochastically binding to the boronic acid. The
nearly two orders of magnitude difference in the mean dwell-
times was similar to the difference observed between the
pyranose (cis-1,2-cyclohexanediol) and furanose (cis-1,2-
cyclopentanediol) model compounds, and therefore we initially
assigned the shorter dwell-time to the binding of β-D-
fructopyranose (68% of the total isomer population in solution),
which has been characterized with X-ray crystallography,[17] and
the longer dwell-time to the binding of β-D-fructofuranose (22%)
(Figure 3c). The ratio of the calculated uncorrected kon values for
We also found that the orientation of the hydroxyl groups in
the 1,2-cyclohexane and 1,2-cyclopentane diols was critical for
binding to the boronic acid. The addition of (1S,2S)-trans-1,2-
cyclohexanediol
(Figure
2e,
S15)
or
(±)-trans-1,2-
cyclopentanediol (Figure S16), in which the hydroxyl groups
point in opposite directions with respect to the plane of the
cycloalkane, did not result in fluctuations of the ionic current
during electrical recordings. These results, in combination with
the binding events observed with cis-1,2-cyclohexanediol,
suggested that only 1,2-hydroxyl groups positioned on the same
side of a cyclohexane ring (cis) bind to a boronic acid. Because
the D-glucose pyranose anomer α-D-glucopyranose has 1,2-
hydroxyl groups positioned cis at C1 and C2, and β-D-
glucopyranose has hydroxyl groups at C1 through C4 positioned
trans to each other (Figure 2g), our results suggest that α-D-
glucopyranose binds to a boronic acid through the cis-hydroxyl
groups at C1 and C2. This proposition was further supported by
the lack of interaction of α-D-glucose 1-phosphate with the PB
pore (Figure 2f, S17), consistent with previous reports that did
not report binding of this phosphorylated sugar to a fluorinated
boronic acid-appended pyridinium salt receptor.[5]
β-D-fructopyranose (43 ±
2
M-1s-1) and β-D-fructofuranose
(18.2 ± 0.4 M-1s-1) reflected the proposed isomer fractions in
solution[7] (Figure S20). Although α-D-fructopyranose has syn
hydroxyl groups at C4 and C5 ideal for binding to the boronic acid
(Figure 2c), our single-molecule analysis is not able to detect an
isomer with a low population (here 3%), especially if, as is likely,
its binding events are similar in amplitude and duration to those
of the dominant β-D-fructopyranose (68%).
Table 1. Kinetic constants for the formation of complexes between the PB pore and diol.
[b]
Diol
Amplitude
Block (pA)[a]
kon (M-1s-1
)
koff (s-1)
Ka (M-1
)
catechol[c]
3.1 ± 0.1
8.1 ± 0.6
15.6 ± 0.7
5.3 ± 0.3
1460 ± 20
29.8 ± 0.5
24 ± 1
0.61 ± 0.05
193 ± 13
191 ± 22
118 ± 9
2.4 ± 0.2 × 103
1.5 ± 0.1 × 10-1
1.3 ± 0.2 × 10-1
2.9 ± 0.4 × 10-1
D-glucose[d]
D-maltose[d]
cis-1,2-
34 ± 3
cyclohexane-
diol[d]
cis-1,2-
3.6 ± 0.1
65 ± 2
2.40 ± 0.09
2.7 ± 0.1 × 101
cyclopentane
-diol[d]
[a] The errors were obtained from the fitting of a dwell-time histogram to the binding events.
[b] Calculated by using Ka = kon/koff. [c] 2 M KCl, 10 mM MOPS, Chelex, pH 7.0, recorded at
−50 mV and 21 ± 1 °C. [d] 2 M KCl, 100 mM MOPS, Chelex, pH 7.0, recorded at −50 mV
and 21 ± 1 °C.
Figure 3. a) Representative current recording from a PB pore at −50 mV with 8
mM D-fructose in the trans compartment (conditions: 2 M KCl, 100 mM MOPS,
pH 7.0, 21 ± 1 °C). b) Dwell-time histogram of the current blockades (level 4,
This article is protected by copyright. All rights reserved.

10.1002/anie.201712740
Angewandte Chemie International Edition
COMMUNICATION
blue) fitted with a two-component probability density function. c) Proposed
scheme describing the binding of the β-D-fructopyranose and β-D-
fructofuranose isomers from an equilibrium mixture to the PB pore (the
percentages of each isomer in the mixture are shown).[7] d) Further analysis of
the current recordings during the reaction identified additional events
accessed from level 4, which were assigned as level 3 (grey) and level 2
(cyan).
of the sugars were the same. A mixture of D-fructose (15 mM),
D-glucose (30 mM) and D-maltose (30 mM) was also analyzed
(Figure 4c). The binding events of the disaccharide (D-maltose)
were easily identified and separated from the binding events of
the monosaccharides (D-fructose and D-glucose) by analyzing
the blockade amplitudes (Figure 4d). Although the two
monosaccharides gave indistinguishable blockade levels, the
identity of D-glucose and D-fructose could be distinguished by
dwell-time analysis (Figure 4e). The short dwell time of α-D-
glucopyranose bound to the boronic acid was separated from
the two longer dwell-times of β-D-fructopyranose and β-D-
fructofuranose.
Further analysis of the blockade level (assigned as level 4,
Figure 3d) with the longer life-time (4762 ± 454 ms), which is
attributed to the binding of β-D-fructofuranose, showed that two
additional levels (ΔI = +5.4 ± 0.4 pA and +3.4 ± 0.5 pA, assigned
as levels 3 and 2, respectively) were derived from level 4. Levels
3 and 2 were accessed only from level 4, and always returned to
level 4; transitions between levels 3 and 2 were not observed.
We constructed a four-state model to describe the observed
transitions (Figure S20). The associated rate constants were
determined by analyzing idealized current traces with Hidden
Markov modelling (QUB 2.0, University at Buffalo). The
association rates for D-fructose (describing the transition from
level 1 to level 4), comprising the rates k14(β-D-fructopyranose)
and k14(β-D-fructofuranose), both showed
a
first-order
dependence on the concentration of D-fructose, while
dissociation (rates k41(β-D-fructopyranose) and k41(β-D-
fructofuranose)) were independent of the concentration,
consistent with bimolecular interactions between β-D-
fructopyranose and PB and between β-D-fructofuranose and PB.
The transitions between levels 4 (β-D-fructofuranose) and 3 (k43
and k34) and between levels 4 (β-D-fructofuranose) and 2 (k42
and k24) were concentration independent, suggesting that level 2
and level 3 arose from rearrangement of the sugar complex
bound to the boronic acid. Because we assigned level 4 to the
adduct between PB and β-D-fructofuranose (bound via the cis-
diol pair at C2 and C3), we postulated that level 3 and level 2
arose from additional coordination of the hydroxyl groups at C1
and C6, respectively, to from an anionic tetrahedral boron
complex[1b] (Figure 3c). In support of this argument, we observed
that D-fructose 1,6-bisphosphate, in which the hydroxyl groups
at C1 and C6 are replaced with non-coordinating phosphate
groups, bound to PB with only one blockade level (Figure S21).
In addition, we also observed only two different current
blockades when D-fructose 6-phosphate bound to PB (Figure
S22), consistent with the initial binding of the cis hydroxyl groups
at C2 and C3 to the boronic acid, followed by the intramolecular
coordination of the hydroxyl group at C1 (but not C6, which is
phosphorylated). The assignment of level 3 to the four-
coordinate binding species with β-D-fructofuranose (Figure 3c)
was based on the similar blockade level observed with D-
fructose 6-phosphate.
Figure 4. a) Single-channel recording at −50 mV with D-glucose (30 mM)
(level 2, cyan) and D-maltose (30 mM) (level 3, grey) in the trans compartment.
b) 2D event-distribution plot displaying the extent of current block versus
log(event duration). c) Single-channel recording with a mixture of D-fructose
(15 mM), D-glucose (30 mM) and D-maltose (30 mM) in the trans
compartment. d) 2D event-distribution plot displaying current block versus
log(event duration) for a mixture of three sugars. D-maltose (level 3) was first
discriminated from the monosaccharides (level 2) by analyzing the extents of
current block. e) D-glucose and D-fructose were identified by analyzing the
dwell times. Histograms were fitted with one- or three component probability
density functions. All measurements were performed in 2 M KCl, 100 mM
MOPS, pH 7.0, 21 ± 1 °C.
In this study, we have shown that single-molecule analysis of
reversible, covalent interactions with an engineered protein
nanopore identifies which cyclic isomers of D-glucose and D-
fructose bind to a boronic acid in solution. The approach can
also identify and discriminate the components of sugar mixtures.
We anticipate that this new binding information will impact the
design of more accurate continuous glucose monitoring devices
for medical diagnostics.[18]
To demonstrate the potential of PB as a saccharide sensor,
the discrimination of the components of solutions of different
sugars was accomplished by analyzing the current blockade
amplitudes and mean dwell-times (Figure 4, S23). The
components in a mixture of D-maltose (30 mM) and D-glucose
(30 mM) were readily distinguished by visual inspection of a
current recording (Figure 4a). In a 2D event-distribution plot of
current block versus log(dwell-times), two clearly separated
populations were apparent based on differences in the
amplitudes of current block (Figure 4b), although the dwell-times
Acknowledgements
This article is protected by copyright. All rights reserved.

10.1002/anie.201712740
Angewandte Chemie International Edition
COMMUNICATION
[8]
[9]
a) J. C. Norrild, H. Eggert, J. Am. Chem. Soc. 1995, 117, 1479-1484; b)
J. C. Norrild, H. Eggert, J. Chem. Soc., Perkin Trans. 2 1996, 2583-
2588.
This work was supported by a European Research Council
Advanced Grant.
H. Bayley, T. Luchian, S. H. Shin, M. B. Steffensen in Single Molecules
and Nanotechnology (Eds.: R. Rigler, H. Vogel), Springer, 2008, pp.
251-277.
Keywords: single-molecule • α-hemolysin • nanoreactor •
boronic acid • sugar detection
[10] T. Luchian, S.-H. Shin, H. Bayley, Angew. Chem. Int. Ed. 2003, 42,
3766-3771.
[1]
[2]
[3]
[4]
a) Z. Guo, I. Shin, J. Yoon, Chem. Comm. 2012, 48, 5956-5967; b) S. D.
Bull, M. G. Davidson, J. M. H. van den Elsen, J. S. Fossey, A. T. A.
Jenkins, Y.-B. Jiang, Y. Kubo, F. Marken, K. Sakurai, J. Zhao, T. D.
James, Acc. Chem. Res. 2013, 46, 312-326.
[11] S.-H. Shin, M. B. Steffensen, T. D. W. Claridge, H. Bayley, Angew.
Chem. Int. Ed. 2007, 46, 7412-7416.
[12] M. B. Steffensen, D. Rotem, H. Bayley, Nat. Chem. 2014, 6, 603.
[13] G. S. Pulcu, E. Mikhailova, L.-S. Choi, H. Bayley, Nat. Nanotech. 2014,
10, 76.
a) J. Yoon, A. W. Czarnik, Bioorg. & Med. Chem. 1993, 1, 267-271; b)
K. E. Secor, T. E. Glass, Org. Lett. 2004, 6, 3727-3730; c) A. E.
Hargrove, R. N. Reyes, I. Riddington, E. V. Anslyn, J. L. Sessler, Org.
Lett. 2010, 12, 4804-4807.
[14] Y. Lin, Y.-L. Ying, Y.-T. Long, Current Opinion in Electrochemistry 2018,
7, 172-178.
[15] D. E. Levy, P. Fügedi in The Organic Chemistry of Sugars, CRC Press,
2005.
a) X. Wu, Z. Li, X.-X. Chen, J. S. Fossey, T. D. James, Y.-B. Jiang,
Chem. Soc. Rev. 2013, 42, 8032-8048; b) B. M. Chapin, P. Metola, S. L.
Vankayala, H. L. Woodcock, T. J. Mooibroek, V. M. Lynch, J. D. Larkin,
E. V. Anslyn, J. Am. Chem. Soc. 2017, 139, 5568-5578.
[16] N. Ni, S. Laughlin, Y. Wang, Y. Feng, Y. Zheng, B. Wang, Bioorg. &
Med. Chem. 2012, 20, 2957-2961.
[17] S. P. Draffin, P. J. Duggan, G. D. Fallon, Acta Cryst. E 2004, 60,
o1520-o1522.
a) Q. Dou, D. Hu, H. Gao, Y. Zhang, A. K. Yetisen, H. Butt, J. Wang, G.
Nie, Q. Dai, RSC Adv. 2017, 7, 41384-41390; b) C. Zhang, M. D.
Losego, P. V. Braun, Chem. Mat. 2013, 25, 3239-3250; c) S. Gamsey,
J. T. Suri, R. A. Wessling, B. Singaram, Langmuir 2006, 22, 9067-9074.
J. Axthelm, S. H. C. Askes, M. Elstner, U. R. G, H. Görls, P. Bellstedt, A.
Schiller, J. Am. Chem. Soc. 2017, 139, 11413-11420.
[18] a) I. R. Scorei in Breast Cancer - Current and Alternative Therapeutic
Modalities (Ed.: E. Gunduz), InTech, 2011, pp. 91-114; b) S. Jin, Y.
Cheng, S. Reid, M. Li, B. Wang, Med. Res. Rev. 2010, 30, 171-257; c)
X. Sun, T. D. James, Chem. Rev. 2015, 115, 8001-8037.
[5]
[6]
[7]
R. Polacek, J. Stenger, U. Kaatze, J. Chem. Phys. 2002, 116, 2973-
2982.
T. Barclay, M. Ginic-Markovic, M. R. Johnston, P. Cooper, N. Petrovsky,
Carbohydr. Res. 2012, 347, 136-141.
This article is protected by copyright. All rights reserved.

10.1002/anie.201712740
Angewandte Chemie International Edition
COMMUNICATION
COMMUNICATION
Sweet reaction: The aqueous,
reversible covalent chemistry of
boronic acids and diols can be
monitored inside a protein
William J. Ramsay and
Hagan Bayley*
Page No. – Page No.
nanoreactor. The approach is used to
identify which cyclic isomers of D-
glucose and D-fructose bind to a
boronic acid in aqueous solution.
Both of these binding modes
Single-molecule
determination of the
isomers of D-glucose
and D-fructose that
bind to boronic acids
contradict current binding models.
This article is protected by copyright. All rights reserved.