An enantioselective fluorescence sensor for glucose based on a cyclic tetrapeptide containing two boronic acid binding sites

Article abstract of DOI:10.1002/ejoc.200600245

The synthesis and binding properties of carbohydrate receptor 1d, which contains two boronic acid binding sites facing each other at opposing sides of a cyclotetrapeptide cavity, are presented. According to electrospray mass spectrometry, ¹H NM

Full text of DOI:10.1002/ejoc.200600245

FULL PAPER
DOI: 10.1002/ejoc.200600245
An Enantioselective Fluorescence Sensor for Glucose Based on a Cyclic
Tetrapeptide Containing Two Boronic Acid Binding Sites
Guido Heinrichs,[a] Marc Schellenträger,[b] and Stefan Kubik*[a]
Keywords: Fluorescence / Molecular recognition / Peptides / Receptors / Sensors
The synthesis and binding properties of carbohydrate recep-
tor 1d, which contains two boronic acid binding sites facing
each other at opposing sides of a cyclotetrapeptide cavity,
are presented. According to electrospray mass spectrometry,
1H NMR spectroscopy, and fluorescence spectroscopy, this
while D-fructose, D-ribose, and D-xylose cannot be bound by
cooperative action of both boronic acid binding sites. Thus,
1d possesses high affinity and selectivity in glucose recogni-
tion combined with good enantioselectivity. Moreover, fluo-
rescence of 1d is quenched upon substrate binding, which
allows the use of this receptor as an optical sensor for glucose
in aqueous solution.
receptor forms stable 1:1 complexes with
24800 1200 -glucose (Ka = 11900 1600
–1) and
water/methanol (1:1) at pH = 11.7. Complexes with
tose, -mannose, and -allose are significantly less stable,
D
-glucose (Ka
=
M
L
M
–1) in
D-galac- (© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
D
D
Germany, 2006)
of the systems developed so far exhibit appreciable activity
only in non-competitive organic media.[4] Alternative ap-
proaches towards the development of artificial carbo-
hydrate receptors with a high long-term stability and, if nec-
essary, a specificity for substrates that cannot be targeted
with natural systems are therefore actively being pursued.
The most promising strategy relies on the reversible forma-
tion of cyclic esters between boronic acids and the hydroxy
groups of sugar molecules.
Boronic acids are known to form five- or six-membered
cyclic esters with 1,2-diols and 1,3-diols, respectively, in ba-
sic aqueous solution.[10–12] The underlying equilibrium is
depicted in Scheme 1 (a).
Introduction
The specific recognition of cell-surface carbohydrates by
proteins initiates important biochemical processes such as
cell–cell interactions or the immune response. Carbo-
hydrate–protein interactions are also important for carbo-
hydrate metabolism and for the transport of carbohydrates
across cell membranes. Important insights into the prin-
ciples of the underlying non-covalent interactions have
come from crystal structures of carbohydrate binding pro-
teins with the substrate or an inhibitor bound inside the
active centre.[1–3] In addition, synthetic carbohydrate recep-
tors have contributed significantly to our current under-
standing of the factors that control affinity as well as selec-
tivity in carbohydrate recognition.[4] In spite of the signifi-
cant progress that has undoubtedly been made in the devel-
opment of artificial carbohydrate receptors, including the
first systems active in aqueous solution,[5–9] practical appli-
cations for such receptors, for example as chemosensors to
detect the presence and concentration of biologically im-
portant sugars in aqueous solution, are not yet in sight.
These applications, examples of which are monitoring the
glucose level in blood or the sugar concentration during
fermentation processes, require a receptor to bind to a given
substrate and signal complex formation in water, yet most
Scheme 1.
Equilibrium constants determined for the reactions be-
tween phenylboronic acid and several monosaccharides
show that the stability of the corresponding esters decreases
in the order fructose Ͼ galactose Ͼ mannose Ͼ glucose (at
25 °C).[13] This order seems to be retained by all mono-
boronic acids and therefore reflects the intrinsic carbo-
hydrate selectivity of a receptor containing one boronic acid
[a] Technische Universität Kaiserslautern, Fachbereich Chemie –
Organische Chemie,
Erwin-Schrödinger-Strasse, 67663 Kaiserslautern, Germany
Fax: +49-631-205-3921
E-mail: kubik@chemie.uni-kl.de
[b] Bergische Universität Wuppertal, Fachbereich C,
Gaußstrasse 20, 42097 Wuppertal, Germany
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2006, 4177–4186
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4177

G. Heinrichs, M. Schellenträger, S. Kubik
FULL PAPER
binding site. The selectivity order changes when a receptor high pH. The binding affinity towards -glucose is signifi-
contains two or more boronic acid groups that can interact cantly higher than towards -glucose and a change in the
with the substrate in a cooperative fashion. In this case, optical properties of 1d upon binding allows complex for-
selectivity depends mainly on the mutual arrangement of mation to be monitored by means of fluorescence spec-
the binding sites, which, in turn, is controlled by the struc- troscopy. Interestingly, electrospray ionisation mass spec-
ture of the receptor. Thus, attaching two boronic acids to a trometry and fluorescence spectroscopy indicate that only
suitable molecular scaffold represents an efficient modular glucose among the series of monosaccharides tested forms
strategy for the development of selective receptors[10–12] not 1:1 complexes with 1d of appreciable stability. Tetrapeptide
only for monosaccharides but also for sugar acids[14,15] and 1d therefore represents a new type of selective fluorescence
sugar alcohols.[16] Enantiodifferentiation in carbohydrate sensor for the detection of glucose in aqueous solution.
recognition can be achieved by using chiral scaffolds,[14–17]
while scaffolds containing chromophores yield fluorescence
or colour sensors that signal the binding event by a change
in their optical properties.[18–21]
Results
The structural complexity of most boronic acid based
carbohydrate receptors developed so far is relatively low,
which is a clear advantage in terms of synthesis.[10–12] One
can speculate, however, that the use of structurally more
Synthesis
We initially chose 1c as our first target because we ex-
elaborate scaffolds based on, for example, (chiral) macro- pected that this compound would be easily accessible from
cyclic compounds could have a positive effect on binding the previously described tetrapeptide 1b, which contains
selectivity. The tetraboronic acid resorcinarene system de- free amino groups in the 5-position of the aromatic rings.[23]
scribed by Strongin represents a step in this direction.[22] Compound 1c could also allow for N–B interactions be-
Another potentially useful scaffold is the cyclic tetrapeptide tween the boron centres and the neighbouring amino
1a, which contains alternating -proline and 3-aminoben- groups. Such interactions have been shown to induce the
zoic acid residues and has been developed in our group.[23] formation of tetrahedral boron centres also at neutral pH,
A conformational analysis revealed that this peptide can thus shifting the working pH of a boronic acid based carbo-
adopt conformations in solution with converging aromatic hydrate receptor from basic solution to an almost physio-
subunits; the angle between these subunits is variable, how- logical environment [Scheme 1 (b)].[10–12,24] It should be
ever. The attachment of binding sites to the aromatic resi- noted that a detailed analysis has recently shown that the
dues should therefore yield receptors in which the peptide N–B dative bond depicted in Scheme 1 (b) is usually only
ring functions as a molecular hinge that allows the distance present in aprotic solvents; in protic media a solvent inser-
and the orientation of the binding sites to adapt to the size tion pathway dominates that leads to the formation of a
of a bound substrate. Molecular modelling indicated that hydrogen-bonded zwitterionic species with little or no N–B
the angle between the aromatic rings in certain tetrapeptide interaction.[25]
conformations is well suited for the inclusion of a monosac-
Unfortunately, our attempts to synthesise 1c by reductive
charide and we therefore expected a derivative of 1a con- amination of 1b with commercial 2-formylphenylboronic
taining two boronic acid groups to exhibit the characteristic acid in the presence of sodium cyanoborohydride only
binding properties of a bis(boronic acid) based carbo- yielded inseparable product mixtures consisting mainly of
hydrate receptor.[10–12] In addition, the restricted conforma- mono- and disubstituted products. We therefore decided to
tional flexibility and the chirality of 1a could have the ad- assemble the bis(boronic acid) containing cyclic tetrapep-
vantage of imposing a high degree of selectivity in complex tide stepwise from appropriately protected preformed sub-
formation.
units. The corresponding synthesis started with the re-
ductive amination of protected 3,5-diaminobenzoic acid de-
rivative 3 (Scheme 2). The product formed was converted
directly into the corresponding pinacol ester 4. Because the
secondary amino group in 4 can potentially interfere during
the following peptide coupling reactions, it was transformed
into a tertiary amine by another reductive amination prior
to peptide synthesis. After acidic cleavage of the Boc group,
5 was coupled with Boc-protected -proline. The resulting
dipeptide 6 was chain-elongated to the corresponding tetra-
peptide 7 by conventional deprotection and coupling meth-
ods. Compound 7 was then deprotected at the N- and the
C-terminus, and the resulting product was cyclised under
pseudo-high-dilution conditions. Chromatographic workup
In this article, we describe the synthesis and binding and acidic cleavage of the pinacol esters in 8 finally afforded
properties of tetrapeptide 1d, and show that 1d binds analytically pure receptor 1d in amounts sufficient for the
strongly to glucose in water/methanol (1:1), albeit only at following binding studies.
4178
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4177–4186

An Enantioselective Fluorescence Sensor for Glucose
FULL PAPER
Scheme 2.
ESI Mass Spectrometry
substrate excess was chosen to estimate the relative stability
of a potential 1:1 complex between the receptor and a
monosaccharide with respect to the 1:2 complex in which
the two boronic acids act as individual binding sites. Only
if the 1:1 complex is sufficiently stable should significant
amounts of this complex be detectable in solution in the
presence of an excess of substrate. As substrates we used
-glucose, -glucose, -galactose, -mannose, -allose, -
fructose, -ribose, and -xylose. The ESI mass spectrum of
the 1d/-glucose mixture is depicted in Figure 1.
To obtain information about the nature and the number
of products formed during the reaction of 1d with carbo-
hydrates, and to see whether the two boronic acids in the
receptor are indeed able to cooperatively bind to a given
substrate to form 1:1 complexes, we investigated solutions
of 1d containing various monosaccharides by means of elec-
trospray ionisation mass spectrometry (ESI-MS). The sam-
ples were prepared by adding 40 equiv. of a substrate to a
0.2 m solution of 1d in water/methanol (1:1) adjusted to
pH = 11.7 with NaOH. After 2 h at room temperature, the
Six major peaks are visible in the spectrum, the most
ESI mass spectrum of each mixture was recorded. A large prominent of which at m/z = 865.61 can be assigned to the
Figure 1. ESI mass spectrum (negative mode) of a solution of 1d (0.2 m) and -glucose (8 m) in water/methanol (1:1; pH = 11.7).
4179
Eur. J. Org. Chem. 2006, 4177–4186
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org

G. Heinrichs, M. Schellenträger, S. Kubik
FULL PAPER
deprotonated 1:1 complex between 1d and -glucose [1d·- Figure 2 shows the 1H NMR spectrum of free receptor
glucose – 4H2O – H+]. The peak at m/z = 883.64 also repre- (1 m) in D2O/CD3OD (1:1) at pD = 11.7 and the effect
sents a 1:1 complex in which only three ester bonds are of the addition of increasing amounts (0.5–25 equiv.) of -
formed [1d·-glucose – 3H2O – H+]. Peaks of the 1:2 com- glucose.
plex are visible at m/z = 1045.75 [1d·2 -glucose – 4H2O –
The free receptor exhibits a 1H NMR spectrum with
H+] and 522.39 [1d·2 -glucose – 4H2O – 2H+]. The peak broad peaks, probably caused by slow conformational in-
at m/z = 767.57 can be assigned to a dehydrated dimethyl terconversion and/or intra- or intermolecular interactions
ester of the free receptor [1d + 2CH3OH – 3H2O – H+] and between the boronic acid groups. Upon addition of -glu-
1
the one at m/z = 873.62 to a compound in which one bo- cose, a new set of much sharper signals appears in the H
ronic acid group of 1d is esterified with glucose and the NMR spectrum, which indicates the formation of a struc-
other is replaced by a hydroxy group [{1d – B(OH)2
+
turally defined complex which, on the NMR timescale, is in
slow equilibrium with free 1d. Complex formation is almost
OH}·-glucose – 2H2O – H+].
The results obtained for the 1d/-glucose mixture were complete in the presence of 1 equiv. of -glucose since a
very similar (see Supporting Information). Interestingly, further increase in substrate concentration has practically
significant amounts of the 1:1 complexes could only be ob- no effect on the spectrum. Only when the substrate is pres-
served for these two monosaccharides. The other substrates ent in large excess are minor signals visible in the 1H NMR
tested can be divided into two groups. Only minor amounts spectrum; these may be caused by the presence of ad-
of 1:1 complexes were detected in the spectra of samples ditional species in solution.
containing -galactose, -mannose, or -allose; the more
prominent peaks in these spectra represent 1:2 complexes,
free receptor, or decomposition products of 1d (see Sup-
porting Information). The second group of substrates con-
sists of -fructose, -ribose, and -xylose. Here, no 1:1
complex formation could be observed and the peaks in the
spectra were those of 1:2 complexes and decomposition
products in which either one or both boronic acid groups
of 1d are substituted by hydroxy groups (see Supporting
Information).
Fluorescence Spectroscopy
The fact that 1d is strongly fluorescent and that fluores-
cence is quenched upon complex formation with glucose
allowed us to gain insight into the rate of complex forma-
tion and to quantitatively determine complex stability.
Upon excitation of a solution of 1d in water/methanol
(1:1; pH = 11.7) at 285 nm a strong emission band is ob-
served in the fluorescence spectrum of 1d with a maximum
at 480 nm. Addition of an excess of -glucose to the solu-
tion causes the fluorescence intensity of this band to de-
NMR Spectroscopy
Additional information on the binding equilibria be- crease. This effect, which demonstrates the usefulness of 1d
1
tween 1d and -glucose came from H NMR spectroscopy. as a chemosensor for glucose, is shown in Figure 3.
Figure 2. Section of the 1H NMR spectrum of 1d (1 m) in D2O/CD3OD (1:1) at pD = 11.7 (a) and the effect of the addition of 0.5 (b),
1 (c), 2 (d), 5 (e), 10 (f), and 25 (g) equiv. of -glucose.
4180
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4177–4186

An Enantioselective Fluorescence Sensor for Glucose
FULL PAPER
480 nm against the glucose/receptor ratio yields the satura-
tion curve depicted in Figure 5 (circles), from which the sta-
bility of the complex between 1d and -glucose can be cal-
culated by nonlinear regression using the mathematical
treatment for 1:1 binding equilibria.[26,27] The excellent
agreement between experimental and calculated results also
shows that the assumption of 1:1 binding is valid.[28]
Figure 3. Time-dependent decrease of the fluorescence intensity of
a solution of 1d (10 µ) in water/methanol (1:1; pH = 11.7) after
the addition of 10 equiv. of -glucose.
Figure 3 shows that under the chosen conditions, equilib-
rium is reached after about 90 min. We therefore prepared
each solution exactly 2 h prior to a spectroscopic or mass
spectrometric analysis to make sure that all samples had
reached thermodynamic equilibrium before the measure-
ment. No decrease in fluorescence intensity was observed
upon addition of -glucose to a solution of 1d in water/
methanol (1:1) at neutral pH, thus indicating that forma-
tion of the tetrahedral boronate shown in Scheme 1 (a) is a
prerequisite for complex formation.
Figure 4 shows a series of fluorescence spectra of solu-
tions of 1d (5 µ) in water/methanol (1:1; pH = 11.7) con-
taining 0–40 equiv. of -glucose. The arrow indicates the
decrease in fluorescence intensity upon increasing the glu-
cose concentration. Plotting the fluorescence intensity at
Figure 5. Saturation curves obtained by plotting the fluorescence
intensity at 480 nm of solutions of 1d (5 µ) in water/methanol
(1:1; pH = 11.7) containing increasing amounts of -glucose
(circles) and -glucose (squares) against the glucose/receptor ratio.
The lines represent the best fit of the nonlinear regression to the
experimental data.
The stability constant, Ka, determined from three indi-
vidual measurements for the complex between 1d and -
glucose amounts to 24800 1200 –1. A similar fluores-
cence titration using -glucose as substrate yielded a Ka
value of 11900 1600 –1, which shows that 1d binds -glu-
cose ca. two times better than -glucose. In other words,
the enantioselectivity of 1d in glucose recognition, as ex-
pressed by the ratio of the stability constants Ka()/Ka(),
is 2.1. A large scattering of the data or a straight line were
obtained during attempts to perform similar fluorescence
titrations for complexes of 1d with other monosaccharides,
which is further evidence for the significantly lower stability
of these complexes with respect to those of glucose.[29]
Discussion
The mass spectrometric analyses of the products formed
during reactions between 1d and various monosaccharides
clearly demonstrate that the two boronic acids in the recep-
tor are able to cooperatively bind to some of the investi-
gated substrates. According to the ESI mass spectra, the
monosaccharides investigated can be divided into three
groups. Significant amounts of 1:1 complexes with respect
to 1:2 complexes were only detected in mixtures of 1d con-
taining - or -glucose. Considering the large substrate ex-
cess used for the measurements, this result is strong evi-
Figure 4. Fluorescence spectra of solutions of 1d (5 µ) in water/
methanol (1:1; pH = 11.7) containing 0–40 equiv. of -glucose (ex-
citation wavelength 285 nm). The arrow indicates the decrease in
fluorescence intensity upon increasing the glucose concentration.
Eur. J. Org. Chem. 2006, 4177–4186
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4181

G. Heinrichs, M. Schellenträger, S. Kubik
FULL PAPER
dence that both monosaccharides form rather stable 1:1 tion. A variety of boronic acid based fluorescence sensors
complexes with the receptor. Significantly smaller amounts for carbohydrates have been described previously, of which
of 1:1 complexes were detected in the mass spectra of sam- many have in common that fluorescence is enhanced upon
ples containing -galactose, -mannose, or -allose, thus complex formation.[18] In some receptors, a photoinduced
indicating that the complexes of these monosaccharides are electron transfer quenching process is inhibited upon carbo-
considerably less stable. Finally, -fructose, -ribose, and - hydrate binding, whereas other receptors are conformation-
xylose seem to be unable to form 1:1 complexes with the ally rigidified upon complex formation, which also causes
receptor.
an increase in fluorescence intensity.[10–12] To explain the
The observed selectivity order significantly deviates from opposite effect in the system described here, we assume that
that of monoboronic acid carbohydrate receptors,[13] which fluorescence quenching upon glucose binding is due to pho-
clearly shows that the structure and conformation of the toinduced electron transfer from the analyte to the sensor
tetrapeptide scaffold in 1d have a decisive effect on binding or to structural constraints in the complex that prevent the
selectivity (including enantioselectivity). Moreover, the fact efficient conjugation of the tertiary amino groups with the
that only two of the eight sugars bind strongly to 1d demon- aromatic subunits in the peptide ring.[30]
strates that, as predicted, the use of structurally well-de-
fined and not too flexible scaffolds can have a beneficial 1d with - and -glucose by fluorescence titration are quite
effect on binding selectivity. large, thereby indicating a good structural complementarity
The stability constants determined for the complexes of
It must also be noted that at the high pH used for the between the receptor and these substrates. A comparison of
binding studies, nucleophilic replacement of one or both the glucose affinity of 1d with that of known boronic acid
boronic acid groups by hydroxy groups causes the slow de- based carbohydrate receptors has to consider factors such
composition of the receptor. This is, of course, disadvan- as number of boronic acid binding sites, pH (complex sta-
tageous with respect to long-term stability, and improve- bility is pH-dependent and substantially increases with in-
ment of the properties of 1d must therefore also address creasing pH[31]), and solvent composition. Although this
strategies to increase stability or to shift the working pH of comparison is not straightforward, it should be noted that
this receptor to neutral media. Interestingly, complex for- bis(boronic acid) receptors with similar glucose affinity in
mation seems to have a protective effect against decomposi- protic media at high pH have been reported.[32,33] Also, the
tion since large quantities of substitution products were enantioselectivity of 1d in glucose recognition is comparable
only detected in solutions containing monosaccharides that to that of other chiral bis(boronic acid) receptors. The binol
interact only weakly with 1d.
receptor reported by Shinkai and James, for example, binds
Support for the mass spectrometric results came from -glucose about 1.6 times better than -glucose.[17] The
NMR spectroscopic investigations. The low resolution of enantioselectivity of the same receptor in fructose binding
the spectrum of the free receptor and the number of peaks is even larger [Ka(-fructose)/Ka(-fructose) = 3.2]. More-
indicate that, in the absence of suitable guests, several (pre- over, James and co-workers have recently described a chiral
sumably non-symmetrical) conformations of 1d, some of bis(boronic acid) receptor whose (R) enantiomer binds -
which are possibly stabilized by interactions between the mannitol more than 2000 times better than the (S) enanti-
boronic acid groups, slowly interconvert in solution. No omer.[16] Hence, enantioselectivity in glucose recognition of
1
break of symmetry could therefore be detected in the H 1d can be considered moderate while affinity, and in par-
NMR spectrum of 1d upon complex formation. Neverthe- ticular carbohydrate selectivity, are significantly better.
less, the pronounced changes in the spectrum caused by the
We were, of course, disappointed to observe complex for-
presence of -glucose clearly demonstrate that 1d forms a mation between 1d and glucose only at high pH even
structurally well-defined complex with this monosaccharide though the tertiary amines in the receptor should, in prin-
which, in the presence of sub-stoichiometric amounts of the ciple, be able to induce binding also in a neutral environ-
substrate, is in slow equilibrium with free receptor. More- ment by interacting with the boron centres.[10–12,24] One
over, the fact that the spectrum of the complex is almost reason for this result could be that the basicity of the aro-
unaffected by increasing the substrate concentration beyond matic amino groups of 1d is too low to allow for N–B inter-
1 equiv. is further evidence that the most stable complex actions, although colour sensors are known in which the
species in solution is the 1:1 complex. Unfortunately, the coordination of an aromatic amine to the boronic acid
complexity of the spectrum prevents a more detailed struc- binding site does permit complex formation at pH = 8.2.[21]
tural assignment. Also, quantitative information with re- Geometrical features in the glucose complex of 1d are there-
spect to complex stability could not be derived from the fore probably more important in preventing the formation
NMR spectroscopic investigations because of the low reso- of intramolecularly stabilized zwitterions[25] or coordinative
lution of the spectrum of the free receptor and the extensive N–B interactions. The results of molecular modelling stud-
signal overlap in the spectra containing free receptor and ies support this assumption. The corresponding calcula-
complex.
tions were performed at a semi-empirical AM1 level with
Complex stability could, however, be determined from the program MacSpartan (Spartan 04 for Macintosh, Wave-
fluorescence titrations. These measurements rely on the fact function, Inc., Irvine, CA) and were based on the crystal
that the fluorescence of 1d, which is due to the 3,5-diamino- structure of tetrapeptide 1a[23] assuming that the boronic
benzoic acid subunits, is quenched upon complex forma- acids in 1d preferentially form cyclic esters with the hydroxy
4182
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4177–4186

An Enantioselective Fluorescence Sensor for Glucose
FULL PAPER
groups in the 1,2- and in 4,6-positions of the α-anomer of
the substrate.[33] The optimised structure of the complex be-
tween 1d and -glucose is depicted in Figure 6.
Conclusions
We have demonstrated that introducing boronic acid
binding sites in the periphery of a chiral macrocyclic scaf-
fold is a promising approach towards new selective carbo-
hydrate sensors. Our first receptor, the cyclic tetrapeptide
1d, possesses high affinity and selectivity in glucose recogni-
tion combined with good enantioselectivity. Moreover, fluo-
rescence of 1d is quenched upon substrate binding, which
allows the use of this compound as an optical sensor for
glucose in aqueous solution. Molecular modelling indicates
that interactions between the boron centres and the neigh-
bouring tertiary amines, which are required to shift the
working pH of the receptor to neutral media, are prevented
by the convergent arrangement of the binding sites in the
1:1 complex. As a consequence, complex formation of 1d
requires high pH, which limits the use of this compound in
practical applications. Future work has to address this issue,
for example by the introduction of additional Lewis basic
substituents suitably arranged with respect to the receptor’s
boronic acid groups.
Figure 6. Calculated structure of the complex between 1d and α--
glucose. Calculations were performed at a semi-empirical AM1
level with the program MacSpartan. Hydrogen atoms at the C
atoms of the receptor have been omitted for clarity.
Experimental Section
General Details: Melting points: Büchi 510 apparatus. Optical rota-
tion: Perkin–Elmer 241 MC digital polarimeter (d = 10 cm). NMR:
Bruker DRX 500 equipped with an automatic sampler. FT-IR: Fin-
nigan MAT 8200. ESI-MS: Micromass Q-Tof UltimaTM API.
Fluorescence spectroscopy: JASCO FP-6200. Elemental analysis:
Pharmaceutical Institute of the Heinrich Heine University, Düssel-
dorf. RP chromatography: Merck LiChroprep RP-8 (40–63 µm)
prepacked column size B (310-25). Abbreviations: All: allyl; Boc:
tert-butoxycarbonyl; DIEA: N-ethyldiisopropylamine; PyCloP:
chlorotripyrrolidinophosphonium hexafluorophosphate; TBTU:
This figure shows that the structure of the tetrapeptide
in the complex does not deviate significantly from the ones
previously reported for other structurally related cyclic tet-
rapeptides.[23] Thus, complex formation most probably does
not induce much strain in the ring. The bound glucose
adopts the energetically favourable 4C1 conformation,
which also contributes to the high stability of the complex.
The calculated structure demonstrates, however, that in or-
der to be oriented in a convergent fashion to allow for si-
multaneous interactions with the substrate, the boron
centres of the cyclic esters must turn away from the neigh-
bouring aromatic amino groups. This orientation prevents
intramolecular N–H···O–B hydrogen bonds or N–B interac-
tions and explains why complex formation of 1d requires
the formation of boronates and, as a consequence, high pH.
It should be pointed out that we have currently no exper-
imental evidence whether the calculated structure in Fig-
ure 6 reflects the true structure of the -glucose complex of
1d in solution. In particular, we have no information about
which anomer of glucose is bound or which cyclic form. It
has been proposed that other bis(boronic acid) receptors
bind to the pyranose form of -glucose in a similar fash-
ion,[34,35] although subsequent NMR investigations showed
that the pyranose form is only bound initially and rapidly
rearranges to the furanose form in aqueous solution.[36] We
cannot rule out that a similar process also occurs in the
system described here. Yet, independent of the exact struc-
ture of the substrate, a prerequisite for the cooperative ac-
tion of both boronic acids in monosaccharide binding is a
convergent arrangement which, according to our calcula-
tions, cannot simultaneously permit interactions of the aro-
matic amino groups with the boronic acid binding sites.
O-(1H-benzotriazol-1-yl)-N,N,NЈ,NЈ-tetramethyluronium
tetra-
fluoroborate; Pro: -proline; AB: 3,5-diaminobenzoic acid; BMAB:
3-amino-5-[(2-boronobenzyl)(methyl)amino]benzoic acid; BMAB-
(pin): 3-amino-5-{methyl[2-(4,4,5,5-tetramethyl-1,3,2-dioxaboro-
lan-2-yl)benzyl]amino}benzoic acid.
Syntheses: All solvents were dried according to standard pro-
cedures prior to use. DMF p.A. was purchased from Merck and
used without further purification. PyCloP was prepared according
to a literature procedure.[37] TBTU and 2-formylphenylboronic acid
are commercially available.
General Procedure for the Cleavage of tert-Butoxycarbonyl Groups:
The starting material was suspended in 1,4-dioxane (30 mL). The
resulting suspension was cooled with an ice bath, and a 6  solu-
tion of HCl in 1,4-dioxane (60 mL) was added dropwise. The reac-
tion mixture was stirred at 0–5 °C for 2 h and then concentrated to
dryness in vacuo.
General Procedure for the Cleavage of Allyl Esters: The ester was
dissolved in THF (20 mL per mmol) under inert conditions.
[Pd(PPh3)4] (10 mg) and morpholine (2 equiv.) were added, and the
reaction mixture was stirred at room temperature for 30 min. Com-
pletion of the reaction was checked by TLC. The solvent was evap-
orated in vacuo, the residue was redissolved in ethyl acetate, and
the organic layer was extracted three times with aqueous 3%
KHSO4 and three times with water. After drying, the solvent was
removed in vacuo.
Eur. J. Org. Chem. 2006, 4177–4186
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4183

G. Heinrichs, M. Schellenträger, S. Kubik
FULL PAPER
Allyl 3,5-Bis(tert-butoxycarbonylamino)benzoate (2): A mixture of
3
1.45 (s, 9 H, Boc-CH3), 4.47 (d, JH,H = 6.0 Hz, 2 H, BM-CH2),
3
3
3,5-bis(tert-butoxycarbonylamino)benzoic
acid[38]
(17.6 g,
4.72 (dt, JH,H = 5.4 Hz, 2 H, All-CH2), 5.24 (dd, JH,H = 12.0,
3
2
50.0 mmol) and NaHCO3 (12.6 g, 150.0 mmol) was suspended in
DMF (150 mL). Allyl bromide (13.0 mL, 150.0 mmol) was added,
and the reaction mixture was stirred at room temperature for 72 h.
After addition of 10% aqueous Na2CO3 (450 mL), the resulting
solution was extracted three times with ethyl acetate. The combined
organic layers were washed with water, and dried. The solvent was
removed in vacuo, and the product was isolated from the residue by
column chromatography (hexane/ethyl acetate, 1:1). It was finally
recrystallised from hexane/ethyl acetate. Yield: 19.2 g (98%); m.p.
2JH,H = 1.6 Hz, 1 H, All-H3(Z)), 5.34 (dd, JH,H = 17.0, JH,H
=
1.6 Hz, 1 H, All-H3(E)), 5.94–6.04 (m, 1 H, All-H2), 6.40 (t, JH,H
= 5.7 Hz, 1 H, NH), 6.84 (t, JH,H = 1.9 Hz, 1 H, AB-H), 7.00 (s,
1 H, AB-H), 7.23 (td, JH,H = 7.3, JH,H = 1.0 Hz, 1 H, BM-H),
7.30–7.42 (m, 3 H, AB-H/BM-H), 7.68 (dd, JH,H = 7.6, JH,H
3
4
3
4
3
4
=
1.2 Hz, 1 H, BM-H), 9.27 (s, 1 H, AB-NH) ppm. C28H37BN2O6
(508.4): calcd. C 66.15, H 7.34, N 5.51; found C 65.89, H 7.41, N
5.35. MS (EI): m/z (%) = 508 (41) [M+].
Allyl 3-(tert-Butoxycarbonylamino)-5-{methyl[2-(4,4,5,5-tetrameth-
yl-1,3,2-dioxaborolan-2-yl)benzyl]amino}benzoate (5): Compound 4
(7.2 g, 14.0 mmol) was dissolved in ethanol (100 mL). Aqueous
formaldehyde (37%, 12 mL, 140.0 mmol) was added and the mix-
ture was stirred at room temperature. Sodium cyanoborohydride
(1.4 g, 22.4 mmol) was added in three portions followed by glacial
acetic acid (8 mL). After stirring overnight, the solvent was re-
moved in vacuo. The residue was dissolved in ethyl acetate, and the
organic layer was extracted once with water, three times with 10%
aqueous Na2CO3, and three times with water. After drying with
sodium sulfate, the solvent was evaporated, and the product was
isolated from the residue by column chromatography (hexane/ethyl
acetate, 6:1; Rf = 0.32). Yield: 7.2 g (98%); m.p. 148–150 °C. 1H
NMR (500 MHz, [D6]DMSO): δ = 1.30 (s, 12 H, Pin-CH3), 1.44
1
148–157 °C. H NMR (500 MHz, [D6]DMSO): δ = 1.48 (s, 18 H,
3
3
Boc-CH3), 4.78 (d, JH,H = 5.3 Hz, 2 H, All-CH2), 5.29 (dd, JH,H
2
3
= 10.6, JH,H = 1.6 Hz, 1 H, All-H3(Z)), 5.41 (dd, JH,H = 17.3,
2JH,H = 1.6 Hz, 1 H, All-H3(E)), 5.99–6.08 (m, 1 H, All-H2), 7.74
(s, 1 H, AB-H), 7.75 (s, 1 H, AB-H), 7.93 (m, 1 H, AB-H), 9.56 (s,
2 H, AB-NH) ppm. C20H28N2O6 (392.5): calcd. C 61.21, H 7.19,
N 7.14; found C 61.45, H 7.30, N 6.89.
Allyl 3-Amino-5-(tert-butoxycarbonylamino)benzoate (3): Allyl 3,5-
bis(tert-butoxycarbonylamino)benzoate (2; 17.7 g, 45.0 mmol) was
first deprotected at the amino groups according to the general pro-
cedure and isolated as the bis(hydrochloride) salt. Without further
purification, the product obtained was suspended in dichlorometh-
ane (250 mL). After the addition of DIEA (17.1 mL, 99.0 mmol),
the mixture was cooled in an ice bath. A solution of di-tert-butyl
dicarbonate (9.8 g, 45.0 mmol) in dichloromethane (100 mL) was
added dropwise whilst stirring over the course of 30 min. Stirring
was continued at about 5 °C for 2 h, and at room temperature over-
night. Afterwards, the solvent was removed in vacuo and the prod-
uct was isolated from the residue by chromatographic workup (hex-
ane/ethyl acetate, 1:1). Allyl 3,5-bis(tert-butoxycarbonylamino)ben-
zoate (2) eluted first and then the desired monoprotected product.
Fractions containing pure 3 were pooled and the solvents evapo-
rated to dryness. The residue was triturated with hexane, filtered
off and dried with P4O10 in vacuo. Yield: 8.2 g (62%); m.p. 104 °C.
1H NMR (500 MHz, [D6]DMSO): δ = 1.54 (s, 9 H, Boc-CH3), 4.81
3
(s, 9 H, Boc-CH3), 3.00 (s, 3 H, N–CH3), 4.72 (dt, JH,H = 5.3 Hz,
3
2 H, All-CH2), 4.77 (s, 2 H, BM-CH2), 5.19 (dd, JH,H = 10.5,
3
2
2JH,H = 1.7 Hz, 1 H, All-H3(Z)), 5.28 (dd, JH,H = 17.3, JH,H
=
1.7 Hz, 1 H, All-H3(E)), 5.88–6.10 (m, 1 H, All-H2), 6.88 (s, 1 H,
AB-H), 7.00 (d, 3JH,H = 7.3 Hz, 1 H, BM-H), 7.06 (s, 1 H, AB-H),
3
7.17–7.43 (m, 2 H, BM-H), 7.51 (s, 1 H, AB-H), 7.73 (dd, JH,H
=
4
7.3, JH,H = 1.2 Hz, 1 H, BM-H), 9.31 (s, 1 H, AB-NH) ppm.
C29H39BN2O6 (522.5): calcd. C 66.67, H 7.52, N 5.36; found C
66.40, H 7.42, N 5.32. MS (EI): m/z (%) = 522 (48) [M+].
Dipeptide Boc-Pro-BMAB(pin)-OAll (6): Before the coupling reac-
tion, compound 5 (6.3 g, 12.0 mmol) was deprotected at the ter-
minal amino group according to the general procedure. The prod-
uct, Boc--proline (3.1 g, 14.4 mmol), and PyCloP (6.1 g,
14.4 mmol) were dissolved in CH2Cl2 (240 mL). DIEA (9.1 mL,
52.8 mmol) was then added dropwise at room temperature and the
reaction mixture was stirred overnight. Afterwards, the solvent was
evaporated in vacuo, and the product was isolated chromatographi-
cally from the residue (hexane/ethyl acetate, 2:1; Rf = 0.42). Yield:
4.8 g (65%); m.p. 95 °C (dec.) (softening from 80 °C). [α]2D5 = –64.3
(c = 2, CHCl3). 1H NMR (500 MHz, [D6]DMSO, 100 °C): δ = 1.30
(s, 12 H, Pin-CH3), 1.32 (s, 9 H, Boc-CH3), 1.73–1.98 (m, 3 H, Pro-
Hβ/Pro-Hγ), 2.05–2.23 (m, 1 H, Pro-Hβ), 2.94 (s, 3 H, N–CH3),
3
3
2
(d, JH,H = 5.4 Hz, 2 H, All-CH2), 5.34 (dd, JH,H = 10.4, JH,H
=
1.6 Hz, 1 H, All-H3(Z)), 5,42 (s, 2 H, NH2), 5.46 (dd, JH,H = 17.3,
3
2JH,H = 1.6 Hz, 1 H, All-H3(E)), 6.05–6.13 (m, 1 H, All-H2), 6.92
(t, 4JH,H = 1.6 Hz, 1 H, AB-H), 7.08 (s, 1 H, AB-H), 7.35 (t, JH,H
4
= 1.6 Hz, 1 H, AB-H), 9.34 (s, 1 H, AB-NH) ppm. C15H20N2O4
(292.3): calcd. C 61.63, H 6.90, N 9.58; found C 61.77, H 7.05, N
9.45.
Allyl 3-(tert-Butoxycarbonylamino)-5-[2-(4,4,5,5-tetramethyl-1,3,2-
dioxaborolan-2-yl)benzylamino]benzoate (4): Compound 3 (7.3 g,
25.0 mmol) and 2-formylphenylboronic acid (5.2 g, 35.0 mmol)
were dissolved in ethanol (150 mL) and the mixture was stirred at
room temperature. Sodium cyanoborohydride (3.9 g, 62.5 mmol)
was added in three portions followed by glacial acetic acid
(9.5 mL). After stirring overnight, another 4 mL of glacial acetic
acid was added to the reaction mixture and stirring was continued
for 24 h. The solvent was evaporated in vacuo and the residue was
suspended in toluene (150 mL). After the addition of 2,3-dimeth-
ylbutane-2,3-diol (3.0 g, 25.0 mmol), the resulting mixture was
heated to reflux in a Dean–Stark trap until TLC indicated complete
conversion (hexane/ethyl acetate, 6:1; Rf = 0.26). The solvent was
removed in vacuo, and the residue was dissolved in ethyl acetate.
The organic layer was extracted once with water, three times with
10% aqueous Na2CO3, and three times with water. After drying
with sodium sulfate, the solvent was evaporated and the product
was isolated from the residue by column chromatography (hexane/
3
3
3.26–3.48 (m, 2 H, Pro-Hδ), 4.21 (dd, JHa,He = 4.0, JHa,Ha
=
8.0 Hz, 1 H, ProHα), 4.70–4.77 (m, 4 H, All-CH2/BM-CH2), 5.22
(dd, JH,H = 10.6, JH,H = 1.5 Hz, 1 H, All-H3(Z)), 5.33 (dd, JH,H
3
2
3
= 17.1, 2JH,H = 1.5 Hz, 1 H, All-H3(E)), 5.86–6.13 (m, 1 H, All-H2),
4
7.02–7.41 (m, 5 H, AB-H/BM-H), 7.57 (t, JH,H = 1.5 Hz, 1 H,
4
AB-H), 7.71 (d, JH,H = 1.5 Hz, 1 H, BM-H), 9.57 (s, 1 H, AB-
NH) ppm. C34H46BN3O7 (619.6): calcd. C 65.91, H 7.48, N 6.78;
found C 66.02, H 7.65, N 6.52. MS (EI): m/z (%) = 619 (52) [M+].
Tetrapeptide Boc-[Pro-BMAB(pin)]2-OAll (7): Prior to coupling,
equimolar amounts of 6 (2.2 g, 3.5 mmol) were deprotected at the
amino group and at the carboxyl group as described in the general
procedures for the deprotection of Boc groups and of allyl esters,
respectively. Both products were dissolved in CH2Cl2 (80 mL). Py-
CloP (1.8 g, 4.2 mmol) and DIEA (2.7 mL, 15.4 mmol) were added,
ethyl acetate, 6:1; Rf = 0.26). Yield: 7.6 g (60%); m.p. 118–120 °C. and the solution was stirred overnight. The solvent was evaporated
1H NMR (500 MHz, [D6]DMSO): δ = 1.29 (s, 12 H, Pin-CH3),
in vacuo, and the residue was purified chromatographically (hex-
4184
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4177–4186

An Enantioselective Fluorescence Sensor for Glucose
FULL PAPER
ane/ethyl acetate, 1:2; Rf = 0.36). Yield: 2.3 g (62%). MS (FAB): Fluorescence Titrations: Stock solutions of the guest (0.6 m) and
m/z (%) = 1103 (100) [M + Na+], 1003 (65) [M – Boc + Na+].
of receptor 1d (10 µ) were prepared in methanol/water (1:1) ad-
justed to pH = 11.7 with NaOH. In total, 11 samples were set up
by adding increasing amounts of the guest solution (0–1000 µL in
100-µL steps) to 1.5 mL of the host solution. All samples were
made up to a volume of 3 mL with the solvent mixture (methanol/
water, 1:1; pH = 11.7). The fluorescence emission intensity at
480 nm was measured for each sample after 2 h upon excitation at
285 nm. Ka was calculated from the resulting saturation curves by a
nonlinear least-squares fitting method and using the mathematical
treatment for 1:1 complex formation.[26,27]
Cyclopeptide Cyclo[Pro-BMAB(pin)]2 (8): The tetrapeptide 7 (1.1 g,
1.0 mmol) was first deprotected at the carboxyl group and sub-
sequently at the terminal amino group according to the general
procedures. For the cyclisation, the deprotected tetrapeptide was
dissolved in a mixture of degassed DMF (40 mL) and DIEA
(1.0 mL, 6 mmol). The resulting solution was added dropwise to a
solution of TBTU (1.60 g, 5 mmol) and DIEA (0.4 mL, 2.5 mmol)
in degassed DMF (200 mL) at 80 °C over the course of 4 h. If nec-
essary, the pH of the reaction mixture was adjusted afterwards to
about 9 by adding more DIEA, and stirring was continued at 80 °C
for 1 h. The solvent was then evaporated in vacuo, and the product
was isolated from the residue by two subsequent chromatographic
separations using first chloroform/acetone (1:1; Rf = 0.27) and then
2-propanol/ethyl acetate (1:5; Rf = 0.46) as eluents. Yield: 0.48 g
Supporting Information (see footnote on the first page of this arti-
cle): ESI mass spectra of mixtures of 1d with various monosaccha-
rides.
Acknowledgments
(52%); m.p. 210 °C (dec.). [α]2D5 = –193.0 (c = 1, DMF). H NMR
1
(500 MHz, [D6]DMSO): δ = 1.30 (s, 24 H, Pin-CH3), 1.74–1.97 (m,
S. K. thanks the Deutsche Forschungsgemeinschaft for generous
funding.
6 H, Pro-Hβ/Pro-Hγ), 2.01–2.17 (m, 2 H, Pro-Hγ), 2.96 (s, 6 H, N–
3
CH3), 3.46–3.62 (m, 4 H, Pro-Hδ), 3.85 (d, JH,H = 8.5 Hz, 2 H,
2
2
Pro-Hα); 4.63 (d, JH,H = 17.7 Hz, 2 H, BM-CH2), 4.79 (d, JH,H
= 17.7 Hz, 2 H, BM-CH2), 6.07 (s, 2 H, AB-H), 6.29 (s, 2 H, AB-
H), 7.03 (d, 3JH,H = 7.6 Hz, 2 H, BM-H), 7.24 (dd, 3JH,H = 7.3 Hz,
[1] F. A. Quiocho, Pure Appl. Chem. 1989, 61, 1293–1306.
[2] R. U. Lemieux, Chem. Soc. Rev. 1989, 18, 347–374.
[3] W. I. Weis, K. Drickamer, Annu. Rev. Biochem. 1996, 65, 441–
473.
[4] A. P. Davis, R. S. Wareham, Angew. Chem. 1999, 111, 3160–
3179; Angew. Chem. Int. Ed. 1999, 38, 2978–2996.
[5] R. D. Hubbard, S. R. Horner, B. L. Miller, J. Am. Chem. Soc.
2001, 123, 5810–5811.
[6] E. Klein, M. P. Crump, A. P. Davis, Angew. Chem. 2005, 117,
302–306; Angew. Chem. Int. Ed. 2005, 44, 298–302.
[7] V. Král, O. Rusin, F. P. Schmidtchen, Org. Lett. 2001, 3, 873–
876.
[8] M. Mazik, H. Cavga, J. Org. Chem. 2006, 71, 2957–2963.
[9] R. Yanagihara, Y. Aoyama, Tetrahedron Lett. 1994, 35, 9725–
9728.
3
2 H, BM-H), 7.28 (s, 2 H, AB-H), 7.40 (dd, JH,H = 7.6 Hz, 2 H,
3
BM-H), 7.72 (d, JH,H = 7.3 Hz, 2 H, BM-H), 9.49 (s, 2 H, AB-
NH) ppm. 13C NMR (125 MHz, [D6]DMSO): δ = 23.50 (Pro-Cγ),
25.60 (Pin-CH3), 32.13 (Pro-Cβ), 39.45 (N–CH3), 47.51 (Pro-Cδ),
56.22 (Pro-Cα), 62.69 (BM-CH2), 84.54 (Pin-C), 103.67 (AB-C),
104.89 (AB-C), 107.05 (AB-C), 126.34 (BM-C), 126.95 (BM-C),
132.25 (BM-C), 137.09 (BM-C), 139.81 (AB-C), 145.92 (AB-C),
150.59 (AB-C), 170.74 (AB-CO), 172.43 (Pro-CO) ppm.
C52H64B2N6O8·1.5H2O (949.8): calcd. C 65.76, H 7.11, N 8.85;
found C 65.88, H 6.99, N 8.88. MS (FAB): m/z (%) = 923 (100) [M
+ H+], 945 (19) [M + Na+].
[10] T. D. James, K. R. A. S. Sandanayake, S. Shinkai, Supramol.
Chem. 1995, 6, 141–157.
[11] 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.
[12] T. D. James, S. Shinkai, Top. Curr. Chem. 2002, 218, 159–200.
[13] J. P. Lorand, J. O. Edwards, J. Org. Chem. 1959, 24, 769–774.
[14] J. Zhao, M. G. Davidson, M. F. Mahon, G. Kociok-Köhn,
T. D. James, J. Am. Chem. Soc. 2004, 126, 16179–16186.
[15] J. Zhao, T. M. Fyles, T. D. James, Angew. Chem. 2004, 116,
3543–3546; Angew. Chem. Int. Ed. 2004, 43, 3461–3464.
[16] J. Zhao, T. D. James, J. Mater. Chem. 2005, 15, 2896–2901.
[17] T. D. James, K. R. A. S. Sandanayake, S. Shinkai, Nature 1995,
374, 345–347.
Cyclopeptide cyclo[Pro-BMAB]2 (1d): The cyclopeptide 8 (200 mg,
220 µmol) was dissolved in acetone/water (4:1; 220 mL), 1  HCl
was added (22 mL), and the solution was stirred at room tempera-
ture for 4 d. Afterwards, the solvent was evaporated in vacuo, and
the material recovered was purified on an RP-8 column. For this,
it was dissolved in a small amount of acetone/water (1:1) and ap-
plied to a column conditioned with acetone/water (1:10). The elu-
ent composition was gradually changed until the pure product
eluted (acetone/water, 2:3). Fractions containing pure product were
pooled and the solvents evaporated to dryness. Finally, the product
was dried with P4O10 in vacuo. Yield: 85 mg (51%); m.p. Ͼ250 °C
(softening from 175 °C). [α]2D5 = –103.1 (c = 1, DMF). 1H NMR
(500 MHz, [D6]DMSO): δ = 1.79–1.97 (m, 6 H, Pro-Hβ/Pro-Hγ),
2.03–2.16 (m, 2 H, Pro-Hβ), 2.89 (s, 6 H, N–CH3), 3.41–3.61 (m, 4
[18] T. D. James, P. Linnane, S. Shinkai, Chem. Commun. 1996,
281–288.
[19] C. J. Ward, P. Patel, T. D. James, Org. Lett. 2002, 4, 477–479.
[20] S. Arimori, M. L. Bell, C. S. Oh, K. A. Frimat, T. D. James, J.
Chem. Soc., Perkin Trans. 1 2002, 803–808.
[21] C. J. Ward, P. Patel, T. D. James, J. Chem. Soc., Perkin Trans.
1 2002, 462–470.
[22] C. J. Davis, P. T. Lewis, M. E. McCarroll, M. W. Read, R.
Cueto, R. M. Strongin, Org. Lett. 1999, 1, 331–334.
[23] S. Pohl, R. Goddard, S. Kubik, Tetrahedron Lett. 2001, 42,
7555–7558.
[24] G. Wulff, Pure Appl. Chem. 1982, 54, 2093–2102.
[25] L. Zhu, S. H. Shabbir, M. Gray, V. M. Lynch, S. Sorey, E. V.
Anslyn, J. Am. Chem. Soc. 2006, 128, 1222–1232.
[26] K. A. Connors, Binding Constants, Wiley, New York, 1987.
[27] R. S. Macomber, J. Chem. Educ. 1992, 69, 375–378.
[28] Attempts to corroborate the 1:1 complex stoichiometry using
Job plots failed because the fluorescence emission of solutions
3
2
H, Pro-Hδ), 3.88 (d, JH,H = 7.7 Hz, 2 H, Pro-Hα), 4.62 (d, JH,H
2
= 17.7 Hz, 2 H, BM-CH2), 4.65 (d, JH,H = 17.7 Hz, 2 H, BM-
3
CH2), 6.14 (s, 2 H, AB-H), 6.32 (s, 2 H, AB-H), 6.96 (d, JH,H
=
7.6 Hz, 2 H, BM-H), 7.20 (dd, 3JH,H = 7.6 Hz, 2 H, BM-H), 7.25–
7.33 (m, 4 H, AB-H/BM-H), 7.54 (d, 3JH,H = 7.0 Hz, 2 H, BM-H),
9.52 (s, 2 H, AB-NH) ppm. 13C NMR (125 MHz, [D6]DMSO): δ
= 23.59 (Pro-Cγ), 32.16 (Pro-Cβ), 39.32 (N–CH3), 47.57 (Pro-Cδ),
56.91 (Pro-Cα), 62.75 (BM-CH2), 104.16 (AB-C), 105.34 (AB-C),
107.26 (AB-C), 125.95 (BM-C), 126.67 (BM-C), 130.04 (BM-C),
134.89 (BM-C), 139.82 (AB-C), 143.14 (AB-C), 150.95 (AB-C),
170.85 (AB-CO), 172.53 (Pro-CO) ppm. C40H44B2N6O8·3.5H2O
(821.5): calcd. C 58.48, H 6.26, N 10.23; found C 58.47, H 6.05, N
10.08. MS (FAB): m/z (%) = 740 (2) [M+ – H2O]. MS (ESI): m/z
(%) = 721 (100) [M – 2 H2O – H+].
Eur. J. Org. Chem. 2006, 4177–4186
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4185

G. Heinrichs, M. Schellenträger, S. Kubik
FULL PAPER
containing a small molar fraction of 1d could not be deter-
mined with sufficiently high accuracy and reproducibility.
[29] The fact that proper binding isotherms were only observed in
the presence of sugars that form significant amounts of 1:1
complexes with 1d at pH = 11.7, thus protecting the receptor
from decomposing, and not in the presence of substrates where
receptor decomposition is much more pronounced, makes us
confident that fluorescence quenching is indeed caused by the
binding event.
[33] K. Tsukagoshi, S. Shinkai, J. Org. Chem. 1991, 56, 4089–4091.
[34] T. D. James, K. R. A. S. Sandanayake, R. Iguchi, S. Shinkai, J.
Am. Chem. Soc. 1995, 117, 8982–8987.
[35] T. D. James, K. R. A. S. Sandanayake, S. Shinkai, Angew.
Chem. 1994, 106, 2287–2289; Angew. Chem. Int. Ed. Engl. 1994,
33, 2207–2209.
[36] M. Bielecki, H. Eggert, J. C. Norrild, J. Chem. Soc., Perkin
Trans. 2 1999, 449–456.
[37] J. Coste, E. Frérot, P. Jouin, J. Org. Chem. 1994, 59, 2437–2446.
[38] A. Kumar, E. W. Meijer, Chem. Commun. 1998, 1629–1630.
Received: March 21, 2006
[30] L. Pu, Chem. Rev. 2004, 104, 1687–1716.
[31] L. I. Bosch, T. M. Fyles, T. D. James, Tetrahedron 2004, 60,
11175–11190.
Published Online: July 12, 2006
[32] Y. Shiomi, M. Saisho, K. Tsukagoshi, S. Shinkai, J. Chem.
Soc., Perkin Trans. 1 1993, 2111–2117.
4186
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 4177–4186