Towards selective recognition of sialic acid through simultaneous binding to its cis-diol and carboxylate functions

Article abstract of DOI:10.1002/ejoc.201000186

A series of receptors containing phenylboronic acid and urea or thiourea units have been designed for simultaneous recognition of the cis-diol and carboxylate functions of sialic acids, which are known to be overexpressed on the surfaces of tumor cells. The interaction of the receptors with 5-acetylneuraminic acid (Neu5Ac) and 2-α-O-methyl Neu5Ac (MeNeu.5Ac) in DMSO solution has been investigated, by means of spectrophotometric titrations and ¹H, ¹³C, and ¹¹B NMR spectroscopy. Additionally, we have also investigated the binding of these receptors with competing monosaccharides such as D-(+)-glucose, D-fructose, methyl α-D-galactoside, and methyl α-D-ma:rmosid.e. Our results show that 2([3-(4-nitrophenyl)thioureido]methyl}phenylboronic acid (3a) recognizes both Neu5Ac and MeNeu5Ac with good selectivity with regard to the remaining monosaccharides investi-gated. DFT calculations performed at the B3LYP/6-31G(d) level show that this selectivity is due to a cooperative twosite binding of Neu5A.c through 1) ester formation by interaction at the phenylboronic acid function of the receptor and 2) hydrogen-bond interaction between the thiourea moiety and the carboxylate group of Neu5Ac. Compound 3a can therefore be considered a promising synthon for the design of contrast agents for magnetic resonance imaging of tumors. In contrast, the analogue of 3a containing a urea moiety compound 3b - displays strong binding to all monosaccharides investigated, due to two-site binding through interaction on the phenylboronic acid function of the receptor and a hydrogen-bond interaction between the urea moiety and the sugar hydroxy groups.

Full text of DOI:10.1002/ejoc.201000186

FULL PAPER
DOI: 10.1002/ejoc.201000186
Towards Selective Recognition of Sialic Acid Through Simultaneous Binding to
Its cis-Diol and Carboxylate Functions
Martín Regueiro-Figueroa,[a] Kristina Djanashvili,[b] David Esteban-Gómez,[a]
Andrés de Blas,[a] Carlos Platas-Iglesias,*[a] and Teresa Rodríguez-Blas*[a]
Keywords: Carbohydrates / Supramolecular chemistry / Sialic acids / Hydrogen bonds / Molecular recognition
A series of receptors containing phenylboronic acid and urea
or thiourea units have been designed for simultaneous re-
cognition of the cis-diol and carboxylate functions of sialic
acids, which are known to be overexpressed on the surfaces
gated. DFT calculations performed at the B3LYP/6-31G(d)
level show that this selectivity is due to a cooperative two-
site binding of Neu5Ac through 1) ester formation by interac-
tion at the phenylboronic acid function of the receptor and
of tumor cells. The interaction of the receptors with 5-acetyl- 2) hydrogen-bond interaction between the thiourea moiety
neuraminic acid (Neu5Ac) and 2-α-O-methyl Neu5Ac
(MeNeu5Ac) in DMSO solution has been investigated by
means of spectrophotometric titrations and 1H, 13C, and 11B
NMR spectroscopy. Additionally, we have also investigated
the binding of these receptors with competing monosac-
and the carboxylate group of Neu5Ac. Compound 3a can
therefore be considered a promising synthon for the design
of contrast agents for magnetic resonance imaging of tumors.
In contrast, the analogue of 3a containing a urea moiety –
compound 3b – displays strong binding to all monosac-
charides investigated, due to two-site binding through inter-
action on the phenylboronic acid function of the receptor and
a hydrogen-bond interaction between the urea moiety and
the sugar hydroxy groups.
charides such as
D-(+)-glucose,
D-fructose, methyl α-D-galac-
toside, and methyl α-
D-mannoside. Our results show that 2-
{[3-(4-nitrophenyl)thioureido]methyl}phenylboronic acid (3a)
recognizes both Neu5Ac and MeNeu5Ac with good selectiv-
ity with regard to the remaining monosaccharides investi-
a tumor marker because it is known to be overexpressed on
the surfaces of tumor cells.[4]
Introduction
Most of the currently available contrast agents for mag-
The most common member of the sialic acid family is
netic resonance imaging (MRI) are small GdIII complexes the nine-carbon amino sugar 5-acetylneuraminic acid
with polyaminocarboxylate ligands such as DTPA and (Neu5Ac, Scheme 1).[5] Neu5Ac and its derivatives gen-
DOTA (DTPA = diethylenetriamine-N,N,NЈ,NЈЈ,NЈЈ-penta- erally occupy the terminal positions of carbohydrate chains
acetate; DOTA = 1,4,7,10-tetraazacyclododecane-1,4,7,10- of glycoproteins and glycolipids in biological membranes,
tetraacetate).[1] However, these MRI contrast agents distrib- and they appear to play important roles in cellular recogni-
ute over the extracellular space in a non-specific way. There tion processes.[6] Neu5Ac exists in solution in equilibrium
is therefore currently great interest in the development of between its α-pyranose (5–8%) and β-pyranose forms (92–
more specific contrast agents that would be able either to 95%), but the furanose type of Neu5Ac is absent because
accumulate in the target tissue[2] or to report on their bio- of the acetamide moiety present at the C5 position.[7] It
logical environments through molecular recognition mecha- must be pointed out, however, that all known glycosides of
nisms.[3] A specific MRI contrast agent could take advan- Neu5Ac are α-linked, often to galactose or N-galactos-
tage of these molecular recognition processes to respond to amine residues through α(2Ǟ3) or α(2Ǟ6) linkages.[8]
functional groups occurring at higher concentrations in the
diseased tissue. Sialic acid, for instance, is considered to be
[a] Departamento de Química Fundamental, Facultade de
Ciencias, Universidade da Coruña,
Campus da Zapateira s/n, 15071 A Coruña, Spain
Fax: +34-981-167065
E-mail: cplatas@udc.es
mayter@udc.es
[b] Biocatalysis and Organic Chemistry, Department of
Biotechnology, Delft University of Technology,
Julianalaan 136, 2628 BL Delft, The Netherlands
Scheme 1. α-5-Acetylneuraminic acid (Neu5Ac) and 2-α-O-methyl-
5-acetylneuraminic acid (MeNeu5Ac).
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201000186.
Eur. J. Org. Chem. 2010, 3237–3248
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3237

C. Platas-Iglesias, T. Rodríguez-Blas et al.
FULL PAPER
Boronic acids are a promising class of recognition moie- of carbohydrate chains we selected 2-α-O-methyl Neu5Ac
ties for the synthesis of artificial receptors and sensors for (MeNeu5Ac). The X-ray crystal structures of four of the
saccharides.[9] The use of boronic acids for this purpose is eight receptors are reported. Finally, the binding trends of
based on their ability to form reversible covalent complexes the different receptors were interpreted with the aid of DFT
with 1,2- and 1,3-diol units in saccharides. Detailed investi- calculations at the B3LYP/6-31G(d) level.
gations of the interactions of Neu5Ac with phenylboronic
acid[10] and phenylboronic acid derivatives[11] have been re-
ported in the literature. Moreover, a few conjugates of
phenylboronic acid and lanthanide(III) complexes of
DTPA,[12] DTPA-bisamides,[13] and DOTA-tetraamides[14]
have also been reported. It has been suggested that a more
selective artificial receptor for sialic acid residues in glyco-
proteins should contain both a phenylboronic acid function
and a group capable of recognizing the negatively charged
COO– groups of sialic acids (Scheme 1).[13] In particular, a
specific MRI contrast agent for sialic acid recognition
should bind selectively with sialic acids in preference both
to other sugar residues present in glycan chains and to sac-
charides such as glucose and fructose, which occur in rela-
tively high concentrations in the blood.[15] Positively
charged benzimidazolium and guanidinium-based receptors
have recently been used for molecular recognition of
Neu5Ac.[16] It has also been shown that bis-boronates can
be designed to bind both the glycerol tail and the hydroxy
acid at the anomeric center.[17]
Urea- and thiourea-based receptors are known to be suit-
able for anion recognition and, in particular, can establish
complementary hydrogen-bond interactions with Y-shaped
anions such as carboxylates.[18] In view of this, we envisaged
Scheme 2. Receptors reported in this work.
that synthetic receptors containing urea or thiourea moie-
ties might be also useful for recognition of the COO– group
of sialic acid. A suitable receptor for sialic acid recognition
might therefore be based on urea or thiourea units contain-
ing phenylboronic acid functions for the recognition of the
1,2- and 1,3-diol groups of sialic acids. However, the design
of new targeting contrast agents selective for sialic acids
requires insight into the optimal arrangement of the dif-
ferent recognition moieties in the receptor.
Results and Discussion
Synthesis and Characterization of the Receptors
Here we report the receptors 1a, 1b, 2b 3a, and 3b
Compounds 1a, 1b, 2b, 3a, and 3b were easily obtained
(Scheme 2), which contain thiourea (1a and 3a) or urea (1b each in one step by the standard method involving the reac-
and 3b) groups and phenylboronic acid functions. Further- tions between amines and phenyl iso(thio)cyanates. How-
more, we also report the benzodiazaborine receptor 2a. We ever, the reaction between (2-aminophenyl)boronic acid and
show that 2a maintains the cyclic benzodiazaborine form in 4-nitrophenyl isothiocyanate under analogous conditions
solution and is therefore not able to interact with 1,2- and resulted in the formation of the benzodiazaborine 2a
1,3-diol groups of sugars. Compound 2b, however, is cap- (Scheme 2).[19] All six receptors were fully characterized by
able of recognition of sialic acids, due to its open structure. conventional techniques. The 1H and 13C NMR spectro-
With the goal of evaluation of a possible synergetic effect scopic data for these receptors are summarized in Tables S1
1
of the (thio)urea and pheylboronic acid functions on the and S2 in the Supporting Information. The H NMR spec-
binding of sialic acids we also report the receptors 4a and tra of compounds 1a, 1b, 2b, 3a, and 3b (500 MHz, [D6]-
4b. The interactions of the eight receptors with Neu5Ac and DMSO) each show, for instance, two signals due to (thio)-
with other competing monosaccharides such as -(+)-glu- urea protons and one signal due to the –OH protons of the
cose or -fructose were also investigated in DMSO solution phenylboronic acid function. The 1H NMR spectrum of 2a
by means of spectrophotometric titrations and 1H, 13C, and confirms the formation of the benzodiazaborine unit, be-
11B NMR spectroscopy. Additionally, we have also investi- cause it shows only one resonance due to NH protons, to-
gated the interactions of these receptors with methyl α-- gether with one signal of the same intensity due to the
galactoside and methyl α--mannoside, which are models –OH protons. The structure proposed for compound 2a on
of units commonly present in glycan chains. As a model the basis of the spectroscopic data was confirmed by single-
for neuraminic residues present in the terminal positions crystal X-ray diffraction analysis (Figure 1). The B(1)–N(2)
3238
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 3237–3248

Selective Recognition of Sialic Acid
distance [1.465(2) Å] is considerably shorter than that ex-
pected for a B–N single bond [1.57–1.63 Å], which suggests
a partial double bond character of this bond in the benzodi-
azaborine 2a. The benzodiazaborine unit is essentially co-
planar (mean deviation from planarity 0.02343 Å) and
forms an angle of 74.6° with the plane described by the
nitrophenyl unit.
Figure 1. ORTEP diagram (30% probability level) of the crystal
structure of 2a.
Receptors 4a and 4b were prepared by a slight modifica-
tion of the reported procedure.[20]
Single crystals of 1b, 4a, and 4b were grown in DMSO/
H2O (1:1, v:v) mixtures. Compounds 1b and 4a crystallized
as DMSO solvates, whereas 4b crystallized in the unsol-
vated form. The X-ray crystal structure of a DMF solvate
of 4b has been reported previously.[20]
Compound 4b crystallizes in the monoclinic P21 space
group and the asymmetric unit contains one molecule. Self
assembly in 4b occurs through the characteristic urea α-
network of N–H···O hydrogen bonds (Figure 2).[21] The
urea groups are aligned along the a-axis, and the phenyl
and nitrophenyl groups are twisted out of the urea plane by
43.1 and 43.7°, respectively, to accommodate the hydrogen
bonded α-network. Weak C–H···O interactions appear to
provide additional support to the crystal packing.
Compound 1b·DMSO crystallizes in the monoclinic P21/
c space group and the asymmetric unit contains two urea
and two DMSO molecules. The urea molecule adopts a flat
planar conformation stabilized by intramolecular C–H···O
hydrogen bonds involving the O atom of the urea group,
further stabilized by resonance in the extended aromatic
molecule (Figure 2).[22] The presence of the phenylboronic
acid function disrupts the urea α-network through intermo-
lecular N–H···OBC6H4 hydrogen bonds. The DMSO mole-
cules stabilize the crystal packing through the formation of
DMSO···HOBC6H4 hydrogen bonds.
Figure 2. ORTEP diagrams (30% probability) of the crystal struc-
tures of 1b (top), 4a (middle), and 4b (bottom), highlighting the H-
bond interactions involving the (thio)urea units.
Study of the Interactions of 1a/b–4a/b with CH3COO–
Receptors 1a/b–3a/b were designed for simultaneous
recognition of the cis-diol and carboxylate functions of
Neu5Ac. We therefore initially investigated the interactions
of these receptors with acetate as a model for the carboxyl-
Compound 4a·DMSO crystallizes in the monoclinic P21/ ate function of Neu5Ac (Scheme 1). The binding of recep-
n space group and the asymmetric unit contains one thio- tors 1a/b–3a/b with CH3COO– was followed by means of
urea and one DMSO molecule. The phenyl and nitrophenyl spectrophotometric titrations in DMSO solution: a solution
groups are twisted out of the thiourea plane by 39.8 and of the receptor (10–4 ) was titrated with a standard solu-
26.3°, respectively, as previously observed for different di- tion of the anion (as the tetrabutylammoniun salt) up to a
aryl thioureas.[23] The sulfoxide oxygen of DMSO acts as a five- to tenfold excess. The absorption spectra of the thio-
bifurcated acceptor to the thiourea NH donors (Figure 2).
ureas 1a and 3a each show a maximum at ca. 360 nm,
Eur. J. Org. Chem. 2010, 3237–3248
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3239

C. Platas-Iglesias, T. Rodríguez-Blas et al.
FULL PAPER
whereas those of the ureas 1b, 2b, and 3b each show a maxi-
mum at ca. 350 nm. These absorption bands are typical of
the nitrophenyl chromophore.[24] Nonlinear least-squares
fits of the UV/Vis titration profiles allowed us to determine
the binding constants listed in Table 1.
Table 1. Binding constants (logK values) determined for the inter-
action of receptors 1a/b–4a/b with acetate in DMSO solution at
25 °C.[a]
1a[b]
logK1
logK2
logK1
logK1
5.28(6)
4.45(5)
3.61(1)
5.39(3)
3a
logK1
5.07(1)
1b
2a
3b
logK1
logK1
logK2
logK1
3.59(1)
5.49(2)
5.05(5)
3.35(5)
4a[b]
2b
logK1
2.87(2)
4b
[a] The errors given correspond to one statistical deviation; the ex-
perimental errors are estimated to be Ϯ0.2 logK units. [b] The ti-
tration data were fitted according to the equilibria described by
Equations (1) (K1) and (2) (K2).
The experimental data can be interpreted in terms of the
equilibria shown in Equations (1) and (2).[25]
LH + CH3COO– p [CH3COO···HL]–
(1)
[CH3COO···HL]– + CH3COO– p L– + [(CH3COO)2H]–
(2)
All the investigated receptors undergo the first equilib-
rium [Equation (1)], characterized by the association con-
stant K1, giving a more or less stable hydrogen-bond com-
plex. The second process [Equation (2)], characterized by
an association constant K2, involves the deprotonation of
the receptor by a second CH3COO– anion, and so it is ex-
pected to occur in those cases in which the receptor is par-
ticularly acidic. This is indeed the case for thiourea 1a, for
which the titration data are best fitted on the assumption
Figure 3. Spectra taken during the course of a titration of a DMSO
solution of 1a (top) or 3a (bottom) with a standard solution of
[Bu4N]CH3COO. Insets: changes in the molar absorbance at se-
lected wavelengths upon addition of CH3COO–.
that the two consecutive processes described by Equa- thiourea moiety increases the electron-donor properties of
tions (1) and (2) are taking place. Upon addition of the donor, due to the negative charge of the anion. Anion
CH3COO– to a solution of 1a in DMSO, the intensity of binding provides stabilization of the excited state, resulting
the band at 362 nm progressively decreases, while three new in a bathochromic shift. The logK1 value obtained for the
maxima at 290, 398, and 470 nm develop (Figure 3). The interaction between 3a and CH3COO– is only slightly lower
new band at 470 nm is characteristic for deprotonation of than those obtained for 1a and 4a, indicating similar capa-
the receptor.[18c,26] A comparison of the equilibrium con- bilities of the three receptors to enter into H-bond interac-
stants determined for 1a and 4a (Table 1) indicates that the tions with acetate.
introduction of the boronic acid function of 1a does not
have an important effect on the K1 and K2 values.
The data reported in Table 1 indicate that the thiourea-
containing receptors enter into stronger H-bond interac-
Unlike 1a and 4a, thiourea 3a does not undergo depro- tions with CH3COO– than their urea-containing counter-
tonation upon addition of CH3COO–, probably because 3a parts. This would be expected, however, because thiourea
is a weaker acid than the former receptors. Indeed, upon is a much stronger acid than urea (pKa = 21.1 and 26.9,
addition of CH3COO– to a solution of 3a in DMSO the respectively in DMSO).[29] This is also the reason why thio-
intensity of the band at 360 nm progressively decreases, ureas 1a and 4a undergo deprotonation of the N–H groups
while the formation of a new band at 389 nm is observed in the presence of CH3COO–, whereas for the urea ana-
(Figure 3). However, no band characteristic of the depro- logues 1b, 2b, 3b, and 4b this situation is not observed. The
tonation of the receptor at ca. 470 nm is observed. The ba- spectral variations observed in the ureas 1b, 2b, 3b, and 4b
thochromic shift of the absorption maximum at 360 nm in- are also less pronounced than those observed for the thio-
dicates anion recognition through hydrogen bonding.[27] In- urea analogues. Figure 4 shows the representative example
deed, the electronic excitation in (4-nitrophenyl)thioureido of 1b. It can be seen that the band of the free receptor at
receptors is accompanied by charge transfer from the donor 353 nm decreases upon anion addition, while a new maxi-
nitrogen of the thiourea to the acceptor substituent of the mum is formed at ca. 370 nm. As mentioned above, the ba-
chromophore (–NO2).[28] The coordination of acetate to the thochromic shift of the absorption maximum indicates
3240
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 3237–3248

Selective Recognition of Sialic Acid
anion recognition. The binding constants reported in
Table 1 show that receptor 2b enters into weaker H-bond
interactions with CH3COO– than 1b, 3b, and 4b. This is
attributed to the steric hindrance in the vicinity of the urea
moiety caused by the presence of the –B(OH)2 group in the
ortho position.
1
Figure 5. H NMR spectra taken over the course of a titration of
a DMSO solution of 1b (10 m) with [Bu4N]CH3COO.
upfield shift of the resonance due to the –NH proton. Upon
addition of 3 equiv. of the anion this signal is observed at δ
Figure 4. Spectra taken during the course of a titration of a DMSO
solution of 1b with a standard solution of [Bu4N]CH3COO. Inset: = 9.69 ppm, which clearly confirms acetate recognition by
changes in the molar absorbance at 390 nm upon addition of
the receptor. It has long been established that triarylboranes
interact with anions such as fluoride or cyanide to form the
CH3COO–.
corresponding borate complexes.[31] The results reported
1H NMR spectroscopy confirms the binding of acetate
here for 2a demonstrate that benzodiazaborines, which also
to the receptors 1b, 2b, 3b, and 4b. Indeed, addition of
contain a trigonal planar boron center, can be also used for
CH3COO– to solutions of the receptors in DMSO causes
anion recognition purposes.
important downfield shifts of the two N–H proton signals,
which reflects the establishment of H-bond interactions
with the anion (Figure 5).[18a,30] We also note that in the
cases of 1b, 2b, and 3b the –OH protons of the phenylbor-
onic acid groups undergo significant downfield shifts upon
addition of acetate. In contrast, the positions of the signals
due to aromatic protons are less affected by the interaction
with acetate. Even so, the protons of the nitrophenyl unit
closer to the urea moiety experience a relatively significant
downfield shift (ca. 0.18 ppm), whereas those positioned
further away from the urea unit are shifted upfield by ca.
0.07 ppm (Figure 5). These spectral variations also indicate
H-bond interaction with the anion.[18a] The ∆δ values expe-
rienced by the –NH and –OH protons show saturation pro-
files that confirm the 1:1 stoichiometries of the supramolec-
ular adducts (Figure 5).
Figure 6. Spectra taken during the course of a titration of a DMSO
The benzodiazaborine compound 2a also shows interest-
solution of 2a with a standard solution of [Bu4N]CH3COO. Inset:
ing properties in relation to the recognition of CH3COO–.
changes in the molar absorbance at 375 nm upon addition of
The absorption spectrum of 2a has two maxima, at 290 and
350 nm (Figure 6). Upon addition of acetate the intensities
of these maxima decrease, while two new bands at 280 and
390 nm are formed. The spectrum of the free receptor is
very different to that recorded for, for instance, receptor
1a, which indicates that the benzodiazaborine unit remains
intact in DMSO solution (see Figure 3 and Figure 6). This
CH3COO–.
Study of the Interactions between 1a/b-4a/b and
Carbohydrates
The interactions between the receptors 1a/b–3a/b and
is also confirmed by the 1H NMR spectra (see Figure S1 in Neu5Ac and MeNeu5Ac were followed by means of spec-
1
the Supporting Information). The H NMR spectrum of trophotometric titrations on DMSO solutions of the recep-
the free receptor shows two signals of equal intensity at tors (10–4 ). These experiments were performed in the
12.14 and 9.67 ppm, which are attributed to –NH and –OH presence of equimolar amounts of triethylamine to ensure
protons, respectively. Anion addition results in a dramatic the deprotonation of both Neu5Ac and MeNeu5Ac. To
Eur. J. Org. Chem. 2010, 3237–3248
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3241

C. Platas-Iglesias, T. Rodríguez-Blas et al.
FULL PAPER
confirm that the functioning of the phenylboronic moieties
of the receptors were not affected by triethylamine, 11B
NMR shift measurements were performed on phenylbor-
onic acid (PBA) upon addition of the base in DMSO (Fig-
ure S2 in the Supporting Information). The chemical shift
of 9 ppm indicated that the boron was remaining in the tri-
gonal planar configuration.[10] In addition to Neu5Ac and
MeNeu5Ac, we also investigated the affinities of these
receptors for competing monosaccharides such as -(+)-
glucose, -fructose, methyl α--galactoside, and methyl
α--mannoside. Methyl α--galactoside and methyl α--
mannoside are models of units commonly present in glycan
chains, whereas glucose and fructose occur at relatively high
concentrations in the blood. Aiming to understand whether
or not binding to the (thio)urea moiety plays an important
role in the recognition of Neu5Ac, we also investigated the
interactions of these sugars with receptors 4a and 4b, which
each contain a (thio)urea unit, but lack a phenylboronic
acid moiety (Scheme 2). Furthermore, spectrophotometric
titrations to investigate the interactions of these sugars with
PBA in DMSO solution were also carried out.
The spectrophotometric titration of 1b with Neu5Ac
shows two inflection points at Neu5Ac/1b molar ratios
close to 1 and 2, indicating the stepwise interaction of two
1
Neu5Ac molecules with the receptor (Figure 7). H NMR
spectroscopy confirms the stepwise binding of two Neu5Ac
units to this receptor. Indeed, 1b shows two signals at δ =
9.39 and 8.82 ppm due to the –NH groups of the urea moi-
ety. Addition of Neu5Ac to a solution of 1b in DMSO
(30 m) results in the formation of two new signals for the
–NH protons at ca. 9.28 and 8.60 ppm. Moreover, the signals
due to the protons of the phenylboronic acid function at δ
= 7.28, 7.46, and 7.62 ppm undergo sgnificant shifts upon
Figure 7. Top: UV/Vis spectra taken during the course of a titration
of a DMSO solution of 1b (10–4 ) with a standard solution of
Neu5Ac and Et3N. Inset: changes in the molar absorbance at
390 nm upon addition of Neu5Ac. Bottom: 1H NMR spectra taken
addition of Neu5Ac. In contrast, the signals of the proton over the course of a titration of a DMSO solution of 1b (30 m)
with Neu5Ac and equimolar Et3N.
nuclei of the nitrophenyl unit at δ = 8.19 and 7.69 ppm are
almost unaffected (Figure 7). This indicates that the interac-
tion between 1b and Neu5Ac occurs through the phenyl-
boronic acid function of the receptor. 11B NMR spectra
13C NMR spectra taken after addition of different
additionally demonstrate the change of boron hybridization amounts of Neu5Ac provide a better insight into the bind-
from trigonal to tetrahedral (chemical shift around ing mode. Notably, no peaks corresponding to the free Ne-
–15 ppm), which is a result of covalent binding to Neu5Ac u5Ac were found even in the presence of 2 equiv. This indi-
(Figure S3 in the Supporting Information). The binding cates that after the binding at the phenylboronic acid group
would be expected to involve both the glycerol tail and the of the receptor has occurred, an additional interaction at
α-hydroxycarboxylic acid group of sialic acid. These two the thio(urea) moiety takes place. It also explains the strong
binding profiles were shown to be feasible by measurement chemical shift of the thiourea carbon (from 179 to
of 11B NMR spectra of 1b with equimolar amounts of fruc- 169 ppm), not observed when the equimolar amounts of
tose and lactic acid in DMSO. In both cases the tetrahedral Neu5Ac and the receptor are used. These data support a
configuration of boron observed by NMR indicated bind- 1:2 (receptor:Neu5Ac) binding ratio at high concentrations
ing through boronate ester formation (Figure S4 in the Sup- of sugar. We thus conclude that receptors 1a and 1b are not
porting Information). After addition of two equivalents of able to recognize Neu5Ac through its simultaneous binding
Neu5Ac the signals corresponding to the free receptor in to the phenylboronic acid and (thio)urea functions of the
the proton spectrum have nearly disappeared. Addition of receptors. Instead, Neu5Ac binds preferentially to the phen-
a larger excess of Neu5Ac causes significant downfield ylboronic acid unit, with use of an excess of Neu5Ac also
shifts of the two N–H proton signals, which reflects the es- resulting in binding to the (thio)urea moiety. This picture is
tablishment of a H–bond interaction between the sugar and also in line with the UV/Vis spectral changes observed upon
the thiourea moiety of the receptor.[18a] A similar situation addition of Neu5Ac (Figure 7). Indeed, at the first stages
1
is observed in the H NMR spectrum of 1a upon addition of the titration addition of Neu5Ac causes only a slight
of Neu5Ac (Figure S5 in the Supporting Information).
decrease in the absorption band of 1b at 353 nm, whereas
3242
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 3237–3248

Selective Recognition of Sialic Acid
addition of an excess of Neu5Ac induces the characteristic
bathochromic shift that signals carboxylate recognition.
The UV/Vis spectra of receptors 3a and 3b show maxima
at 360 and 350 nm, respectively (Figure 8; see also Fig-
As expected, the benzodiazaborine compound 2a binds ure S8 in the Supporting Information). Addition of
Neu5Ac, but no interaction between this receptor and the Neu5Ac results in a red shifting of the absorption maxima,
other monosaccharides studied in this work is observed. while their intensities remain almost unaffected. These spec-
The spectral changes observed in the UV/Vis spectrum of tral variations are indicative of the recognition of the car-
2a upon addition of Neu5Ac are very similar to those ob- boxylate function of Neu5Ac by the (thio)urea moiety (see
served upon interaction with acetate (see Figure 6 and Fig- above). At the same time, 11B NMR shift measurements
ure S6 in the Supporting Information). These results indi- reported binding at the phenylboronic group of the recep-
cate that Neu5Ac binds to 2a through the carboxylate func- tor, through the presence of boron in its tetrahedral config-
tion. 11B NMR spectra additionally demonstrate the uration, upon addition of different amounts of Neu5Ac
change in the boron hybridization from trigonal to tetrahe- (Figure S8 in the Supporting Information). 13C NMR spec-
dral (chemical shift around –15 ppm) resulting from coval- tra taken at different Neu5Ac/3a ratios indicate simulta-
ent binding to Neu5Ac (Figure S6 in the Supporting Infor- neous binding of the receptor to the glycerol tail and hydro-
mation). However, the binding constant obtained for the gen-bond formation with the thiourea unit. At Neu5Ac/3a
interaction with Neu5Ac (logK1 = 4.71, Table 2) is slightly molar ratios of 1 and 2, no peaks corresponding to the free
lower than that obtained for the interaction with CH3COO– receptor were detected. After addition of an excess of the
(logK1 = 5.39). This can be attributed to the larger steric receptor the peaks of free 3a appeared, but all resonances
demand of Neu5Ac.
corresponding to the bound species remained unchanged in
the spectrum. This indicates that the simultaneous binding
through the phenylboronic group and the thio(urea) moiety
takes place regardless of the amount of Neu5Ac used for
the interaction. The overall complexity of the 13C NMR
spectra of the receptors upon binding with Neu5Ac can be
explained in terms of the variety of the bound species due
to the diastereomeric character of the boronate esters that
are formed. Analysis of the spectrophotometric titrations
provides especially high binding constants for the interac-
tions of 3a and 3b with Neu5Ac: logK1 = 4.78(1) and
4.65(1), respectively. These binding constants are substan-
tially higher than those obtained for the interaction of 4a
or 4b with Neu5Ac. Moreover, they are also ca. 1.6 logK
units higher than those obtained for the interaction of
Neu5Ac with phenylboronic acid. It can thus be concluded
that receptors 3a and 3b bind to Neu5Ac both through their
(thio)urea and through their phenylboronic acid functions,
Table 2. Binding constants (logK values) determined for receptors
2a/b–4a/b in DMSO solution at 25 °C.[a]
Neu5Ac MeNeu5Ac Fructose Glucose MeGal MeMann
[c]
[b]
[b]
[b]
[b]
2a
2b
3a
3b
4a
4b
4.71(1)
3.41(1)
4.78(1)
4.65(1)
2.91(2)
2.63(3)
3.19(4)
4.38(4)
2.19(1)
3.49(4)
2.32(1) 2.06(1) 2.77(2)
3.24(4) 2.75(2) 2.95(4)
4.69(2) 5.02(1) 4.58(3)
3.14(6)
5.51(5)
[c]
[b]
[b]
[b]
[b]
[b]
[b]
[b]
[c]
[c]
[b]
PBA 3.10(1)
3.28(7)
3.21(4) 2.29(8) 2.28(8)
[a] The errors given correspond to one statistical deviation; the ex-
perimental errors are estimated to be Ϯ0.2 logK units. [b] No ap-
parent binding, as reflected in the lack of changes observed in the
UV/Vis spectral titrations upon addition of saccharide. [c] Not de-
termined.
The UV/Vis spectrum of 2b shows a maximum at such cooperative binding being responsible for the espe-
290 nm, the intensity of which quickly diminishes upon ad- cially high binding constants observed. Of these two recep-
dition of Neu5Ac; addition of Neu5Ac also results in the tors, 3a displays an important selectivity for Neu5Ac with
formation of a new band at ca. 350 nm (Figure S7 in the
Supporting Information). Spectrophotometric titrations of
2b with -(+)-glucose, -fructose, methyl α--galactoside,
or methyl α--mannoside result in similar spectral varia-
tions. The binding constants obtained from the non-least-
squares fit of the UV/Vis titration data are shown in
Table 2. The logK1 values obtained for -(+)-glucose, -
fructose, and methyl α--galactoside are slightly lower than
those obtained for the interactions of these sugars with
PBA. However, the logK1 value obtained for the interaction
between 2b and Neu5Ac is slightly higher than that ob-
tained for phenylboronic acid (ca. 0.3 logK units), and also
slightly higher than the binding constant of 2b with acetate
(ca. 0.5 logK units). Furthermore, 2b binds more strongly
to Neu5Ac than 4b. This can be attributed to simultaneous
recognition of the carboxylate and cis-diol functions of
Figure 8. UV/Vis spectra taken during the course of a titration of
Neu5Ac by this receptor, which results in a certain degree
a DMSO solution of 3b with a standard solution of Neu5Ac and
Et3N. Inset: changes in the molar absorbance at 380 nm upon ad-
dition of Neu5Ac.
of selectivity for Neu5Ac with respect to the remaining sac-
charides studied in this work.
Eur. J. Org. Chem. 2010, 3237–3248
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3243

C. Platas-Iglesias, T. Rodríguez-Blas et al.
FULL PAPER
respect to potentially competing monosaccharides such as by interaction with the phenylboronic acid function of the
-(+)-glucose, -fructose, methyl α--galactoside, and receptor, and 2) a hydrogen-bond interaction between the
methyl α--mannoside. This, however, is not the case with thiourea moiety and the carboxylate group of Neu5Ac. In-
receptor 3b, because the binding constants determined for deed, one of the oxygen atoms of the carboxylate group of
the interaction with these monosaccharides are higher than Neu5Ac is involved in hydrogen-bond interaction with a
that with Neu5Ac. This receptor has a particularly high af- –NH group of the thiourea unit [O···N 2.79 Å, O···HN
finity for -fructose [logK = 5.51(5), Table 2]. The binding 1.77 Å, O···H–N 166.0°]. Furthermore, a weak C–H···O in-
constants reported in Table 2 show that the substitution of teraction appears to provide additional support for the in-
the thiourea moiety of 3a by a urea group does not substan- teraction of the carboxylate group and the receptor [O···C
tially affect the affinity of the receptor for Neu5Ac, but in- 3.33 Å, O···HN 2.25 Å, O···H–N 169.0°]. For the β-pyr-
creases the binding constants towards the remaining sac- anose form, however, our calculations predict that the ester
charides by 1.45–2.27 logK units.
formed at C8 and C9 should be more stable than that
Neu5Ac exists in solution in equilibrium between its α- formed at C7 and C9 by 30.8 kJmol–1. One of the oxygen
pyranose (5–8%) and its β-pyranose forms (92–95%).[7] atoms of the carboxylate group of Neu5Ac is involved in
Thus, in solution the major form of free Neu5Ac is hydrogen-bond interaction with the two NH groups of the
expected to be the β-pyranose form, whereas all known thiourea unit [Figure 9; O···N(1) 2.73 Å, O···HN(1) 1.71 Å,
glycosides of Neu5Ac are α-linked.[8] We therefore investi- O···H–N(1) 164.6°; O···N(2) 2.95 Å, O···HN(2) 2.02 Å,
gated the interactions of the receptors 2b, 3a, and 3b with O···H–N(2) 148.8°].
2-α-O-methyl Neu5Ac (MeNeu5Ac) as a model for neur-
aminic residues present in the terminal positions of carbo-
hydrate chains. The changes observed in the absorption
spectra of the receptors are similar to those observed upon
addition of Neu5Ac (Figure S9 in the Supporting Infor-
mation). The binding constants obtained are slightly lower
than those obtained for Neu5Ac, which suggests that the β-
pyranose form provides stronger binding to these receptors
than the α-pyranose one. Even so, 3a shows an important
selectivity for MeNeu5Ac over other possible competing
saccharides such as glucose, as well as over methyl α--ga-
lactoside and methyl α--mannoside, which are models of
units commonly present in glycan chains. The receptor 2b
also shows a certain selectivity for MeNeu5Ac over these
monosaccharides, whereas for 3b this is not the case
(Table 2).
To explain the selectivity of 3a for Neu5Ac, we charac-
terized the 3a·Neu5Ac– system with the aid of DFT calcula-
tions [B3LYP/6-31G(d)]. This theoretical approach has pre-
viously been used to investigate hydrogen-bond interactions
established between (thio)urea receptors and different sub-
strates,[32] including carboxylates.[33] Phenylboronic acid is
known to bind covalently and reversibly to 1,2- and 1,3-
diols to give five- and six-membered cyclic esters.[10] It has
previously been shown that the formation of a five-mem-
bered ester at C7 and C8 is limited by the unfavorable
erythro configuration of the glycerol tail. In our calculations
we therefore only considered the five-membered cyclic ester
formed at C8 and C9 and the six-membered cyclic ester
formed at C7 and C9 (Scheme 1). 11B NMR studies demon-
Figure 9. Structures of (top) the 3a·α-Neu5Ac– and (bottom) the
3a·β-Neu5Ac– systems obtained from DFT calculations at the
B3LYP/6-31G(d) level.
According to our calculations the 3a·β-Neu5Ac– system
strate that the boron atom of the receptor is in tetrahedral is more stable than the 3a·α-Neu5Ac– one by 27.3 kJmol–1.
hybridization upon binding to cis-diol functions (see This result is consistent with the lower association con-
above). For the reasons given above, our calculations were stants determined for the interaction of 3a with Me-
performed both for the α-pyranose and for the β-pyranose Neu5Ac, which exists in its α-pyranose form, than with
forms of Neu5Ac. Neu5Ac, the major form of which adopts the β-pyranose
For the α-pyranose form of Neu5Ac our results show form in solution.
that the adduct formed at C8 and C9 is less stable than
DFT calculations at the B3LYP/6–31G(d) level were also
that formed at C7 and C9 by 15.9 kJmol–1. The optimized performed on the 3b·methyl α--galactoside system in
geometry of the adduct (Figure 9) indeed shows a coopera- an attempt to explain the reasons for the high affinities of
tive two-site binding of Neu5Ac through 1) ester formation this receptor for -(+)-glucose, -fructose, methyl α--
3244
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 3237–3248

Selective Recognition of Sialic Acid
without further purification. Solvents were of reagent grade and
purified by the usual methods.
galactoside, and methyl α--mannoside. Methyl α--
galactoside can potentially form five-membered esters at C2
and C3 or at C3 and C4 (Figure S10 in the Supporting In-
formation). Additionally, a six-membered ester at position
C4 and C6 is also possible. According to our calculations
the formation of the six-membered ester at positions C4
and C6 is more favorable than the formation of five-mem-
bered esters at C2–C3 and C3–C4. This is nicely in agree-
ment with experimental data reported in the literature.[34]
The minimum-energy conformation obtained for the
3b·methyl α--galactoside system (see Figure S10 in the
Supporting Information) shows the hydroxy group at the
C3-position involved in a hydrogen-bond interaction with
the oxygen atom of the urea group (O···O 2.98 Å, O···HO
2.12 Å, O···H–N 146.2°). The high binding constants ob-
tained for the interactions between 3b and -(+)-glucose, -
fructose, methyl α--galactoside, and methyl α--mannos-
ide can thus be attributed to two-site binding through inter-
action at the phenylboronic acid function of the receptor
and a hydrogen-bond interaction between the urea moiety
and the hydroxy groups of the saccharides. The sulfur atom
of thiourea is a poorer hydrogen-bond acceptor than the
oxygen atom of urea,[35] so this cooperative effect is not
observed for 3a.
Elemental analyses were carried out with
a ThermoQuest
Flash 1112 elemental analyzer. ESI-TOF mass spectra were re-
corded with a LC-Q-q-TOF Applied Biosystems QSTAR Elite
spectrometer in the negative mode. IR spectra were recorded with
a Bruker Vector 22 spectrophotometer with a Golden Gate Attenu-
ated Total Reflectance (ATR) accessory (Specac). 1H and 13C
NMR spectra were run with Bruker Avance 300, Bruker Av-
ance 500, or Varian Inova 300 spectrometers. Chemical shifts are
reported in δ values. Spectral assignments were based in part on
two-dimensional COSY, HMQC, and HMBC experiments. 11B
NMR was measured with a Bruker Avance 400 spectrometer with
use of a solid-state probe and 4 mm zirconia rotors, spinning at the
magic angle with a rate of 20 Hz. Peak positions were measured
with respect to the resonance of a solution of boric acid (0.1 , δ
= 0 ppm). Electronic spectra in the UV/Vis range were recorded at
25 °C with a Perkin–Elmer Lambda 900 UV/Vis spectrophotome-
ter in 1.0 cm quartz cells. Acetate- and sugar-binding studies were
performed with DMSO solutions of the receptors (10–4 , 10 mL).
Typically, aliquots of a fresh standard solution of the envisaged
substrate (0.1–0.01 ) in the same solvent were added and the UV/
Vis spectra of the samples were recorded. For titrations with Ne-
u5Ac and MeNeu5Ac one equivalent of triethylamine was used to
ensure deprotonation of the carboxylate group of the sialic acids.
All spectrophotometric titration curves were fitted with the aid of
the HYPERQUAD program.[37] Binding constants were obtained
from a simultaneous fit of the UV/Vis absorption spectral changes
at six to eight selected wavelengths. A minimum of 25 absorbance
data points at each of these wavelengths was used.
Conclusions
We have designed a series of receptors containing (thio)-
urea and phenylboronic acid functions for the recognition
of sialic acids. Our results show that receptors 1a and 1b
are able to recognize Neu5Ac preferentially through their
phenylboronic acid units. Additional binding to the (thio)-
urea moiety can be observed when an excess of Neu5Ac is
applied. No simultaneous binding of Neu5Ac to the func-
tional groups was observed. The receptors 2b, 3a, and 3b,
however, bind to Neu5Ac through cooperative two-site
binding involving 1) ester formation by interaction at the
phenylboronic acid function of the receptor, and 2) a hydro-
gen-bond interaction between their (thio)urea moieties and
the carboxylate group of Neu5Ac. Of these receptors, 3a
displays the highest binding affinity for Neu5Ac, combined
with the best selectivity with respect to other competing
saccharides. Furthermore, this receptor also shows good
selectivity for MeNeu5Ac over methyl α--galactoside and
methyl α--mannoside, which are models of units com-
monly present in glycan chains, as well as over glucose,
which occurs in relatively high concentrations in blood. The
results presented here may be useful for the design of spe-
cific MRI contrast agents for the recognition of sialic acid
residues on the surfaces of tumor cells.
Computational Methods: All calculations were performed with the
aid of hybrid DFT, the B3LYP exchange-correlation func-
tional,[38,39] and the Gaussian 03 package (revision C.01).[40] Full
geometry optimizations of the 2b·Neu5Ac–, 3a·Neu5Ac–, and
3b·methyl α--galactoside systems were performed in vacuo with
use of the standard 6–31G(d) basis set. The stationary points found
on the potential energy surfaces as a result of the geometry optimi-
zations were tested to ensure they represented energy minima rather
than saddle points by frequency analysis. The relative free energies
of the different conformations calculated for each system include
non-potential-energy contributions (that is, zero point energy and
thermal terms) obtained by frequency analysis.
3-[3-(4-Nitrophenyl)thioureido]phenylboronic Acid (1a): 3-Amino-
phenylboronic acid (0.130 g, 0.842 mmol) and 4-nitrophenyl iso-
thiocyanate (0.152 g, 0.842 mmol) were dissolved in a diethyl ether/
dioxane mixture (5:1, v:v, 5 mL). The resulting solution was stirred
at room temperature for 14 h. The precipitate formed was collected
by filtration and washed with diethyl ether to give 1a (0.270 g, 89%
yield) as a white solid; m.p. 124 °C. 1H NMR (500 MHz, [D6]-
DMSO, 25 °C, Me4Si): δ = 10.30 (s, 1 H, NH), 10.20 (s, 1 H, NH),
3
8.20 (d, JH,H = 9.2 Hz, 2 H, C6H4NO2), 8.07 (s, 2 H, OH), 7.84
3
(d, JH,H = 9.2 Hz, 2 H, C6H4NO2), 7.76 [s, 1 H, C6H4B(OH)2],
7.61 [d, 3JH,H = 7.3 Hz, 1 H, C6H4B(OH)2], 7.55 [d, 3JH,H = 9.0 Hz,
1 H, C6H4B(OH)2], 7.34 [m, 1 H, C6H4B(OH)2] ppm. 13C NMR
(125.8 MHz, [D6]DMSO, 25 °C, Me4Si): δ = 179.4, 146.3, 142.2,
138.1, 135.0, 130.9, 129.7, 127.6, 126.0, 124.2, 121.5 ppm. IR
(ATR): ν = 3471 (OH), 3374 (NH), 1490 (NO ), 1327 (NO ) cm–1.
˜
2
2
MS (ESI–): m/z = 316 [M – H]–. C13H12BN3O4S·0.5C4H8O2
(361.18): calcd. C 49.88, H 4.47, N 11.63; found C 49.81, H 4.42,
N 11.60.
Experimental Section
General: 1-(4-Nitrophenyl)-3-phenylthiourea (4a)[36] and 1-(4-nitro-
phenyl)-3-phenylurea (4b)[20] were prepared by a slight modification
of the literature methods with use of diethyl ether as solvent. All
other chemicals were purchased from commercial sources and used
3-[3-(4-Nitrophenyl)ureido]phenylboronic Acid (1b): 3-Aminophen-
ylboronic acid (0.132 g, 0.851 mmol) and 4-nitrophenyl isocyanate
Eur. J. Org. Chem. 2010, 3237–3248
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3245

C. Platas-Iglesias, T. Rodríguez-Blas et al.
FULL PAPER
(0.140 g, 0.851 mmol) were dissolved in diethyl ether and the mix-
ture was stirred at room temperature for 14 h. The precipitate
formed was collected by filtration and washed with diethyl ether to
give 1b (0.202 g, 79% yield) as a yellow solid; m.p. 267 °C. 1H
1
(0.067 g, 27% yield) as a pale yellow solid; m.p. 124 °C. H NMR
(500 MHz, [D6]DMSO, 25 °C, Me4Si): δ = 10.30 (s, 1 H, NH), 8.38
3
(s, 1 H, NH), 8.23 (s, 2 H, OH), 8.17 (d, JH,H = 9.2 Hz, 2 H,
C6H4NO2), 7.84 (d, 3JH,H = 9.2 Hz, 2 H, C6H4NO2), 7.59 [d, 3JH,H
NMR (500 MHz, [D6]DMSO, 25 °C, Me4Si): δ = 9.39 (s, 1 H, NH),
= 7.2 Hz, 1 H, C6H4B(OH)2], 7.35 [m, 2 H, C6H4B(OH)2], 7.24 [m,
3
3
8.82 (s, 1 H, NH), 8.19 (d, JH,H = 9.3 Hz, 2 H, C6H4NO2), 8.03 1 H, C6H4B(OH)2], 4.88 (d, JH,H = 5.2 Hz, 2 H, -CH2-) ppm. 13C
3
(s, 2 H, OH), 7.72 [s, 1 H, C6H4B(OH)2], 7.70 (d, JH,H = 9.3 Hz,
NMR (125.8 MHz, [D6]DMSO, 25 °C, Me4Si): δ = 180.0, 146.5,
3
2 H, C6H4NO2), 7.62 [d, JH,H = 8.1 Hz, 1 H, C6H4B(OH)2], 7.46 142.2, 141.9, 134.3, 129.6, 128.2, 126.5, 124.6, 120.8, 47.8 ppm. IR
3
[d, JH,H = 7.4 Hz, 1 H, C6H4B(OH)2], 7.28 [m, 1 H, C6H4B- (ATR): ν = 3519 (OH), 3395 (NH), 1469 (NO ), 1320 (NO ) cm–1.
˜
2
2
(OH)2] ppm. 13C NMR (125.8 MHz, [D6]DMSO, 25 °C, Me4Si): δ MS (ESI–): m/z = 313 [M – H]–. C14H14BN3O4S·0.5C4H8O2
= 151.9, 146.4, 140.8, 138.0, 134.9, 128.4, 127.8, 125.1, 124.5, 120.5,
(375.21): calcd. C 51.22, H 4.84, N 11.20; found C 51.60, H 4.86,
N 11.49.
117.3 ppm. IR (ATR): ν = 3403 (OH), 3327 (NH), 3308 (NH), 1673
˜
(C=O), 1487 (NO2), 1323 (NO2) cm–1. MS (ESI–): m/z = 300 [M –
H]–. C13H12BN3O5 (301.06): calcd. C 51.86, H 4.02, N 13.96; found
C 51.92, H 4.12, N 13.90.
2-{[3-(4-Nitrophenyl)ureido]methyl}phenylboronic Acid (3b): The re-
ceptor was prepared as described for 3a from 2-(aminomethyl)-
phenylboronic acid (0.102 g, 0.678 mmol) and 4-nitrophenyl isocy-
anate (0.111 g, 0.678 mmol) to give 3b (0.058 g, 24% yield) as a
pale yellow solid; m.p. 217 °C. 1H NMR (500 MHz, [D6]DMSO,
25 °C, Me4Si): δ = 9.51 (s, 1 H, NH), 8.32 (s, 2 H, OH), 8.12 (d,
Single crystals of 1b·DMSO suitable for X-ray diffraction analysis
were grown from an oversaturated solution of the isolated solid in
DMSO/H2O (1:1, v:v), which was heated until complete dissolution
of the solid and then allowed to cool down to room temperature.
3
3JH,H = 9.3 Hz, 2 H, C6H4NO2), 7.59 (d, JH,H = 9.3 Hz, 2 H,
3
C6H4NO2), 7.55 [d, JH,H = 7.5 Hz, 1 H, C6H4B(OH)2], 7.32 [m, 2
1-Hydroxy-2-(4-nitrophenyl)-1,2-dihydrobenzo[c][1,5,2]diaza-
borinine-3(4H)-thione (2a): The receptor was prepared as described
for 1a from 2-aminophenylboronic acid (0.132 g, 0.962 mmol) and
4-nitrophenyl isothiocyanate (0.173 g, 0.962 mmol) to give 2a
(0.281 g, 85% yield) as white solid; m.p. 224 °C. 1H NMR
(500 MHz, [D6]DMSO, 25 °C, Me4Si): δ = 12.14 (s, 1 H, NH), 9.67
H, C6H4B(OH)2], 7.21 [m, 1 H, C6H4B(OH)2], 6.85 (s, 1 H, NH),
3
4.42 (d, JH,H = 5.8 Hz, 2 H, -CH2-) ppm. 13C NMR (125.8 MHz,
[D6]DMSO, 25 °C, Me4Si): δ = 154.7, 147.2, 143.7, 140.6, 134.1,
129.5, 128.0, 126.2, 125.3, 117.0, 43.4 ppm. IR (ATR): ν = 3521
˜
(OH), 3317 (NH), 1672 (C=O), 1495 (NO2), 1322 (NO2) cm–1. MS
(ESI–): m/z = 296 [M – H]–. C14H14BN3O5·0.5C4H8O2 (359.14):
calcd. C 53.51, H 5.05, N 11.70; found C 53.72, H 4.90, N 12.09.
3
(s, 1 H, OH), 8.27 (d, JH,H = 8.9 Hz, 2 H, C6H4NO2), 8.06 [d,
3JH,H = 6.6 Hz, 1 H, C6H4B(OH)], 7.60 [m, 1 H, C6H4B(OH)], 7.44
[m, 3 H, C6H4NO2 + C6H4B(OH)], 7.21 [m, 1 H, C6H4B-
(OH)] ppm. 13C NMR (125.8 MHz, [D6]DMSO, 25 °C, Me4Si): δ
= 179.1, 148.8, 146.0, 144.8, 133.2, 132.5, 130.6, 123.8, 122.8,
X-ray Crystal Structures: Three-dimensional X-ray data were col-
lected with a Bruker SMART 1000 CCD for 2a and with a
Bruker X8 APEXII CCD for 1b·DMSO, 4a·DMSO, and 4b. Data
were corrected for Lorentz and polarization effects and for absorp-
tion by semiempirical methods[41] based on symmetry-equivalent
reflections. Complex scattering factors were taken from the pro-
gram SHELX97[42] running under the WinGX program system[43]
implemented on a computer with an Intel Pentium® processor. All
115.4 ppm. IR (ATR): ν = 3498 (OH), 3398 (NH), 1482 (NO ),
˜
2
1343 (NO2) cm–1. MS (ESI–): m/z
=
298 [M
–
H]–.
C13H10BN3O3S·0.5C4H8O2 (343.17): calcd. C 52.50, H 4.41, N
12.24; found C 52.43, H 3.85, N 12.42.
Single crystals of 2a suitable for X-ray diffraction analysis were
grown from an oversaturated solution of the isolated solid in
DMSO/H2O (1:1, v:v), which was heated until complete dissolution
of the solid and then allowed to cool down to room temperature.
Table 3. Crystal data and refinement details for 1b, 2a, 4a, and 4b.
1b·DMSO
2a
4a·DMSO
4b
2-[3-(4-Nitrophenyl)ureido]phenylboronic Acid (2b): The receptor
was prepared as described for 1a from 2-aminophenylboronic acid
(0.131 g, 0.956 mmol) and 4-nitrophenyl isocyanate (0.157 g,
0.957 mmol). The precipitate was washed with dioxane and dried
under vacuum to give 2b (0.140 g, 45% yield) as a pale yellow solid;
Emp. formula
Formula weight
Crystal system
Space group
Temp. [K]
a [Å]
b [Å]
c [Å]
α [°]
β [°]
C15H18BN3O6S C13H10BN3O3S C15H17N3O3S2 C13H11N3O3
379.19
monoclinic
P21/c
100.0(2)
13.6530(6)
12.8836(6)
19.965(1)
90
105.750(1)
90
3380.0(3)
1584
299.11
triclinic
¯
P1
351.44
monoclinic
P21/n
100.0(2)
5.4120(6)
9.3484(9)
32.776(3)
90
91.154(3)
90
1657.9(3)
736
257.25
monoclinic
P21
100.0(2)
4.5871(3)
8.3370(5)
15.3036(9)
90
97.009(4)
90
580.88(6)
268
1
100.0(2)
5.6677(5)
8.5250(7)
13.681(1)
87.589(2)
79.481(2)
85.447(2)
647.63(9)
308
m.p. 206 °C. H NMR (500 MHz, [D6]DMSO, 25 °C, Me4Si): δ =
3
10.54 (s, 1 H, NH), 9.34 (s, 1 H, NH), 8.26 (d, JH,H = 9.1 Hz, 2
3
H, C6H4NO2), 8.01 [d, JH,H = 7.4 Hz, 1 H, C6H4B(OH)2], 7.50
3
[m, 3 H, C6H4NO2, C6H4B(OH)2], 7.09 [d, JH,H = 8.1 Hz, 1 H,
C6H4B(OH)2], 7.07 [m, 1 H, C6H4B(OH)2], 6.52 (s, 2 H, OH) ppm.
13C NMR (125.8 MHz, [D6]DMSO, 25 °C, Me4Si): δ = 153.4,
145.7, 145.6, 145.3, 132.7, 132.6, 130.2, 123.5, 120.9, 114.3 ppm.
γ [°]
V [Å3]
F(000)
Z
IR (ATR): ν = 3611 (OH), 3536 (OH), 3208 (NH), 1672 (C=O),
˜
1496 (NO2), 1338 (NO2) cm–1. MS (ESI–): m/z = 282 [M – H]–.
C13H10BN3O4·0.5C4H8O2 (327.10): calcd. C 55.08, H 4.31, N
12.85; found C 54.80, H 3.95, N 12.51.
8
2
4
2
λ [Å]
0.71073
1.490
0.231
1.55–28.29
0.0441
8380
5638
1.047
0.0407
0.1388
0.71073
1.534
0.263
1.51–28.29
0.0275
3211
2662
1.035
0.0338
0.0873
0.71073
1.408
0.339
1.24–28.28
0.0561
4079
2947
1.120
0.0438
0.1419
0.71073
1.471
0.108
3.63–28.45
0.0565
1561
1145
1.031
0.048
0.1233
ρcalcd. [gcm–3]
µ [mm]–1
θ range [°]
Rint
2-{[3-(4-Nitrophenyl)thioureido]methyl}phenylboronic Acid (3a): A
mixture of 2-(aminomethyl)phenylboronic acid (0.101 g,
0.668 mmol)
and
4-nitrophenyl
isothiocyanate
(0.120 g,
Measured reflns.
Observed reflns.
GOF on F2
R1 [IϾ2σ(I)][a]
wR2 (all data)[b]
0.668 mmol) in dioxane (10 mL) was heated at reflux under Ar over
a period of 24 h. The precipitate formed was filtered off and
washed with dioxane. The filtrate was concentrated to dryness to
give an orange oil that was purified by column chromatography on
SiO2 with a CH2Cl2/MeOH (5%) mixture as eluent to give 3a
[a] R1 = Σ||Fo| – |Fc||/Σ|Fo|. [b] wR2 = {Σ[w(||Fo|2 – |Fc|2)2]/Σ[w(Fo )]}1/2
.
4
3246
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 3237–3248

Selective Recognition of Sialic Acid
Chem. Eur. J. 2004, 10, 5205–5217; b) K. Djanashvili, G. A.
Koning, A. J. G. M. van der Meer, H. T. Wolterbeek, J. A. Pe-
ters, Contrast Media Mol. Imaging 2007, 2, 35–41.
a) R. Trokowski, S. Zhang, A. D. Sherry, Bioconjugate Chem.
2004, 15, 1431–1440; b) S. Zhang, R. Trokowski, A. D. Sherry,
J. Am. Chem. Soc. 2003, 125, 15288–15289; c) J. Ren, R. Tro-
kowski, S. Zhang, C. R. Malloy, A. D. Sherry, Magn. Reson.
Med. 2008, 60, 1047–1055.
T. Kawasaki, H. Akanuma, T. Yamanouchi, Diabetes Care
2002, 25, 353–357.
M. Mazik, H. Cavga, J. Org. Chem. 2007, 72, 831–838.
S. M. Levonis, M. J. Kiefel, T. A. Houston, Chem. Commun.
2009, 2278–2280.
a) P. J. Smith, M. V. Reddington, C. S. Wilcox, Tetrahedron
Lett. 1992, 41, 6085–6088; b) E. Fan, S. A. van Arman, S. Ki-
neaid, A. D. Hamilton, J. Am. Chem. Soc. 1993, 115, 369–370;
c) M. Boiocchi, L. Del Boca, D. E. Gómez, L. Fabbrizzi, M.
Licchelli, E. Monzani, J. Am. Chem. Soc. 2004, 126, 16507–
16514; d) D. Esteban-Gómez, L. Fabbrizzi, M. Liccheli, D.
Sacchi, J. Mater. Chem. 2005, 15, 2670–2675; e) T. Gunnlaugs-
son, M. Glynn, G. M. Tocci, P. E. Kruger, F. M. Pfeffer, Coord.
Chem. Rev. 2006, 250, 3094–3117; f) Y.-P. Yen, K.-W. Ho, Tet-
rahedron Lett. 2006, 47, 1193–1196; g) W.-X. Liu, R. Yang, A.-
F. Li, Z. Li, Y.-F. Gao, X.-X. Luo, Y.-B. Ruan, Y.-B. Jiang,
Org. Biomol. Chem. 2009, 7, 4021–4028; h) J. R. Hiscock, C.
Caltagirone, M. E. Light, M. B. Hursthouse, P. A. Gale, Org.
Biomol. Chem. 2009, 7, 1781–1783; i) C. Caltagirone, P. A.
Gale, Chem. Soc. Rev. 2009, 38, 520–563.
a) M. C. Davis, S. G. Franzblau, A. R. Martin, Bioorg. Med.
Chem. Lett. 1998, 8, 843–846; b) J.-C. Zhuo, A. H. Soloway,
J. C. Beeson, W. Ji, B. A. Barnum, F.-G. Rong, W. Tjarks, G. T.
Jordan IV, J. Liu, S. G. Shore, J. Org. Chem. 1998, 64, 9566–
9574; c) M. P. Groziak, A. D. Ganguly, P. D. Robinson, J. Am.
Chem. Soc. 1994, 116, 7597–7605.
L. S. Reddy, S. K. Chandran, S. George, N. J. Babu, A. Nangia,
Cryst. Growth Des. 2007, 7, 2675–2690.
a) S. George, A. Nangia, C.-K. Lam, T. C. W. Mak, J.-F. Nic-
oud, Chem. Commun. 2004, 1202–1203; b) J. P. Clare, A. Statni-
kov, V. Lynch, A. L. Sargent, J. W. Sibert, J. Org. Chem. 2009,
74, 6637–6646.
L. S. Reddy, S. Basavoju, V. R. Vangala, A. Nangia, Cryst.
Growth Des. 2006, 6, 161–173.
a) M. Soriano-García, G. T. Chávez, A. E. D. Pérez, G. A.
Hernández, Anal. Sci. 2001, 17, 907–908; b) M. S. Fonari, Y. A.
Simonov, G. Bocelli, M. M. Botoshansky, E. V. Ganin, J. Mol.
Struct. 2005, 738, 85–89.
R. Martínez-Máñez, F. Sacenon, Chem. Rev. 2003, 103, 4419–
4476.
D. E. Gómez, L. Fabbrizzi, M. Licchelli, E. Monzani, Org. Bi-
omol. Chem. 2005, 3, 1495–1500.
M. Bonizzoni, L. Fabbrizzi, A. Taglietti, F. Tiengo, Eur. J. Org.
Chem. 2006, 3567–3574.
a) S. Nishizawa, R. Kato, T. Hayashita, N. Teramae, Anal. Sci.
1998, 14, 595–597; b) L. Nie, Z. Li, J. Han, X. Zhang, R. Yang,
W.-X. Liu, F.-Y. Wu, J.-W. Xie, Y.-F. Zhao, Y.-B. Jiang, J. Org.
Chem. 2004, 69, 6449–6454.
the structures were solved by direct methods (SHELXS97 for
1b·DMSO, 2a, and 4a·DMSO; SUPERFLIP[44] for 4b) and re-
fined[41] by full-matrix, least-squares on F2. Hydrogen atoms were
included in calculated positions and refined in riding mode. Refine-
ment converged with anisotropic displacement parameters for all
non-hydrogen atoms. Crystal data and details on data collection
and refinement are summarized in Table 3.
[14]
[15]
CCDC-740810 (for 1b·DMSO), -740808 (for 2a), -740809 (for
4a·DMSO) and -740810 (for 4b) contain the supplementary crystal-
lographic data for this paper. These data can be obtained free of
charge from the Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[16]
[17]
[18]
Supporting Information (see also the footnote on the first page of
this article): Tables S1 and S2, giving NMR spectroscopic data for
receptors 1a/b–3a/b, Figures S1–S9 showing 1H, 11B, and UV/Vis
titrations, Figures S10 and S11, showing the optimized geometries
of the 3b·methyl α--galactoside and 2b·Neu5Ac– systems and opti-
mized Cartesian coordinates obtained from DFT/6-31G(d) calcula-
tions.
Acknowledgments
The authors thank the Spanish Ministerio de Educación y Ciencia
(MEC) and Fondo Europeo de Desarrollo Regional (FEDER)
(CTQ2006-07875/PPQ)
and
the
Xunta
de
Galicia
[19]
(INCITE08ENA103005ES) for financial support. This research
was performed in the framework of the EU COST Action D38
(Metal-Based Systems for Molecular Imaging Applications). The
authors are indebted to the Centro de Supercomputación de Gal-
icia (CESGA) for providing the computer facilities. M. R.-F.
thanks the MEC (FPU program) for a predoctoral fellowship. The
authors gratefully acknowledge Dr. Joop A. Peters for fruitful sci-
entific discussions.
[20]
[21]
[1] The Chemistry of Contrast Agents in Medical Magnetic Reso-
nance Imaging (Eds.: A. E. Merbach, É. Tóth), Wiley, New
York, 2001.
[2] V. Jacques, J.-F. Desreux, Top. Curr. Chem. 2002, 221, 123–164.
[3] a) W.-H. Li, S. E. Fraser, T. J. Meade, J. Am. Chem. Soc. 1999,
121, 1413–1414; b) G. Angelovski, P. Fouskova, I. Mamedov,
S. Canals, E. Toth, N. K. Logothetis, K. Nikos, ChemBioChem
2008, 9, 1729–1734; c) T. Chauvin, P. Durand, M. Bernier, H.
Meudal, B.-T. Doan, F. Noury, B. Badet, J.-C. Beloeil, E. Toth,
Angew. Chem. Int. Ed. 2008, 47, 4370–4372.
[22]
[23]
[24]
[25]
[26]
[27]
[4] R. Schauer, Glycoconjugate J. 2000, 17, 485–499.
[5] R. Schauer, Zoology 2004, 107, 49–64.
[6] A. Varki, Trends Mol. Med. 2008, 14, 351–360.
[7] H. Friebolin, P. Kunzelmann, M. Supp, R. Brossmer, G. Keil-
ich, D. Ziegler, Tetrahedron Lett. 1981, 22, 1383–1386.
[8] a) A. Varki, Nature 2007, 446, 1023–1029; b) H. E. Murrey, C.
Hsieh-Wilson, Chem. Rev. 2008, 108, 1708–1731; c) T. Angata,
A. Varki, Chem. Rev. 2002, 102, 439–469.
[9] a) H. S. Mader, O. S. Wolfbeis, Microchim. Acta 2008, 162, 1–
34; b) T. D. James, Top. Curr. Chem. 2007, 277, 107–152; c)
T. D. James, S. Shinkai, Top. Curr. Chem. 2002, 218, 159–200;
d) T. D. James, D. Tony, K. R. A. Sandanayake Samankumara,
S. Shinkai, Angew. Chem. Int. Ed. Engl. 1996, 35, 1911–1922.
[10] K. Djanashvili, L. Frullano, J. A. Peters, Chem. Eur. J. 2005,
11, 4010–4018.
[11] H. Otsuka, E. Uchimura, H. Koshino, T. Okano, K. Kataoka,
J. Am. Chem. Soc. 2003, 125, 3493–3502.
[12] E. Battistini, A. Mortillaro, S. Aime, J. A. Peters, Contrast Me-
dia Mol. Imaging 2007, 2, 163–171.
[13] a) L. Frullano, J. Rohovec, S. Aime, T. Maschmeyer, M. I.
Prata, J. J. Pedroso de Lima, C. F. G. C. Geraldes, J. A. Peters,
[28]
[29]
[30]
H.-G. Löhr, F. Vögtle, Acc. Chem. Res. 1985, 18, 65–72.
F. G. Bordwell, Acc. Chem. Res. 1988, 21, 456–463.
a) V. Amendola, M. Boiocchi, D. Esteban-Gómez, L. Fab-
brizzi, E. Monzani, Org. Biomol. Chem. 2005, 3, 2632–2639; b)
C. M. G. dos Santos, T. McCabe, G. W. Watson, P. E. Kruger,
T. Gunnlaugsson, J. Org. Chem. 2008, 73, 9235–9244.
a) T. W. Hudnall, F. P. Gabba, J. Am. Chem. Soc. 2007, 129,
11978–11986; b) Z.-Q. Liu, M. Shi, F.-Y. Li, Q. Fang, Z.-H.
Chen, T. Yi, C.-H. Huang, Org. Lett. 2005, 7, 5481–5484; c) Y.
Kubo, M. Yamamoto, M. Ikeda, M. Takeichi, S. Shinkai, S.
Yamaguchi, K. Tamao, Angew. Chem. Int. Ed. 2003, 42, 2036–
2040; d) Z. Zhou, F. Li, T. Yi, C. Huang, Tetrahedron Lett.
2007, 48, 6633–6636.
[31]
Eur. J. Org. Chem. 2010, 3237–3248
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3247

C. Platas-Iglesias, T. Rodríguez-Blas et al.
FULL PAPER
[32] a) R. Custelcean, M. G. Gorbunova, P. V. Bonnesen, Chem.
Eur. J. 2005, 11, 1459–1466; b) W.-R. Zheng, Y. Fu, K. Shen,
L. Liu, Q.-X. Guo, J. Mol. Struct. 2007, 822, 103–110.
[33] V. Ruangpornvisuti, J. Mol. Struct. 2004, 686, 47–55.
[34] a) G. T. Morin, M.-F. Paugam, M. P. Hughes, B. D. Smith, J.
Org. Chem. 1994, 59, 2724–2728; b) T. D. James, T. Harada, S.
Shinkai, J. Chem. Soc., Chem. Commun. 1993, 857–860.
[35] Z. Li, F.-Y. Wu, L. Guo, A.-F. Li, Y.-B. Jiang, J. Phys. Chem.
B 2008, 112, 7071–7079.
[36] a) E. Dyer, T. B. Johnson, J. Am. Chem. Soc. 1932, 54, 777–
787; b) S. Kubota, K. Horie, H. K. Misra, K. Toyooka, M.
Uda, M. Shibuka, H. Terada, Chem. Pharm. Bull. 1985, 33,
662–666.
[37] P. Gans, A. Sabatini, A. Vacca, Talanta 1996, 43, 1739–1753.
[38] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[39] C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789.
[40] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery Jr., T.
Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P.
Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R.
Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R.
Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morok-
uma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzew-
ski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K.
Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V.
Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L.
Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A.
Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian 03, Re-
vision C.01, Gaussian, Inc., Wallingford CT, 2004.
[41]
[42]
G. M. Sheldrick, SADABS, version 2.10, University of
Göttingen, Germany, 2004.
SHELX, G. M. Sheldrick, Acta Crystallogr., Sect. A 2008, 64,
112–122.
[43] WinGX, Microsoft Windows® system of programs for solving,
refining, and analyzing single-crystal X-ray diffraction data for
small molecules: L. J. Farrugia, J. Appl. Crystallogr. 1999, 32,
837–838.
[44] SUPERFLIP: L. Palatinus, G. Chapuis, J. Appl. Crystallogr.
2007, 40, 786–790.
Received: February 10, 2010
Published Online: April 30, 2010
3248
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 3237–3248