Re-investigation of optical sensing properties of boronic-acid-appended ReI complexes for saccharides

Article abstract of DOI:10.1039/a906708f

A number of unanswered questions occurred to us upon reading a communication by Yam and Kai (ref. 16) which had reported optical sensing properties of a boronic-acid-appended Re^I complex for saccharides. Careful re-examination has disclosed that the pK_a-value proposed by them (5.9) is wrong and that the saccharide-binding mode at pH above the pK_a is totally different from that at pH below the pK_a. The absorption spectral change, which reflects an sp²-to-sp³ boron hybridisation change induced by the saccharide complexation, was observed only at pH below the pK_a, and the CD band, which reflects the formation of 1:1 cyclic complexes, appeared only at pH above the pK_a. The results imply that the optimum pH should be carefully selected for the precise optical sensing of saccharides.

Full text of DOI:10.1039/a906708f

View Article Online / Journal Homepage / Table of Contents for this issue
Re-investigation of optical sensing properties of boronic-acid-
appended ReI complexes for saccharides
Toshihisa Mizuno, Takayuki Fukumatsu, Masayuki Takeuchi and Seiji Shinkai*
1
Department of Chemistry & Biochemistry, Graduate School of Engineering, Kyushu University,
Fukuoka 812-8581, Japan
Received (in Cambridge, UK) 18th August 1999, Accepted 1st December 1999
Published on the Web 14th January 2000
A number of unanswered questions occurred to us upon reading a communication by Yam and Kai (ref. 16)
which had reported optical sensing properties of a boronic-acid-appended ReI complex for saccharides. Careful
re-examination has disclosed that the pKa-value proposed by them (5.9) is wrong and that the saccharide-binding
mode at pH above the pKa is totally different from that at pH below the pKa. The absorption spectral change, which
reflects an sp2-to-sp3 boron hybridisation change induced by the saccharide complexation, was observed only at pH
below the pKa, and the CD band, which reflects the formation of 1:1 cyclic complexes, appeared only at pH above
the pKa. The results imply that the optimum pH should be carefully selected for the precise optical sensing of
saccharides.
ides.2,7,9,10 Thus, the spectral change observed for the saccharide
complexation is mainly due to a change in neutral boronic
Introduction
Recognition of neutral organic species by synthetic molecular
receptors has been of great interest to many chemists, but some
new methodology which is different from recognition of ionic
species has to be exploited. Most of the known synthetic
acids [R-B(OH)2] to anionic boronate–saccharide complexes
[R-(HO)BϪ〈OO〉(saccharide)].2,6,7 If the pKa for 1 is really 5.9 as
reported by Yam and Kai,16 it follows that the boronic acid
groups should exist as anionic boronate groups [R-BϪ(OH)3]
molecular receptors utilise hydrogen-bonding interactions in
under the measurement conditions (pH 8.3) and the significant
order to recognise and bind with neutral guest molecules.1
spectral change should not take place upon complexation with
However, these interactions are less effective in aqueous media
where guest species are water-soluble and well hydrated. On the
other hand, covalent interactions found in the binding between
saccharides.6,7 In fact, however, they observed large spectral
changes for 1–saccharide complexation. Thirdly, it is not clear
how the association constants (Kass) were determined. They
boronic acids and saccharides are stronger than such hydrogen-
suggested the formation of cyclic 1:1, noncyclic 1:1 and non-
bonding interactions in aqueous media and, therefore, should
cyclic 1:2 complexes on the basis of mass spectrometric data.
be more effective. The usefulness of the boronic acid function
as a saccharide receptor has been demonstrated in saccharide
Nevertheless, the complexation processes were expressed by a
single Kass-value. Fourthly, the interaction of saccharides and
the deprotonated amide anions was proposed to occur at pH
recognition in rigid matrices, CD detection, fluorescence detec-
tion, molecular assemblies, membrane transport, etc.2–9
A
12.1 on the basis of comparison with an N-methylated refer-
ence compound. To the best of our knowledge, however, it is
quite rare for a significant contribution of hydrogen-bonding
interactions to molecular recognition to be found in an aqueous
host–guest system and, in fact, they did not present any con-
structive evidence for this interaction. Fifthly, they used glycine
as a buffer, which is known to interact with boronic acids.18
Because of these queries, we considered that it was worthwhile
to carefully re-examine the content of the paper reported by
Yam and Kai.16 We thus synthesised compounds 1 and 2 and
extensively studied their saccharide-binding properties. The
results showed that they have pKa-values of 8.4–8.9 and the
association mode is totally different depending on the
medium’s pH.
potential breakthrough found through these studies is the
finding that the acidity of boronic acids is intensified when
they form cyclic boronate esters with diols.2,7,9,10 Thus, the
sp2-hybridised boronic acid group can be converted to the
sp3-hybridised boronate anion by the saccharide complex-
ation.2,7,9,10 In addition, the tertiary amine bearing an intra-
molecular boronic acid group changes its basicity upon
saccharide-binding through the boronic acid–nitrogen (B–N)
interaction. Therefore, the spectroscopic properties of amine-
containing chromophores are also changed by the saccharide
binding.9 These findings have enabled us to detect the
saccharide-binding event with absorption and fluorescence
spectroscopic methods.6,7,9,11–14
From a practical viewpoint for optical saccharide sensing, it
is desirable to use a longer wavelength region in both absorp-
tion and fluorescence spectroscopic methods. One potential
issue of this requirement is to utilise metal complexes as a
chromophoric site.8,15 More recently, Yam and Kai16 reported a
ReI complex bearing two boronic acids (compound 1) which
changed electronic absorption characteristics in response to
saccharide binding. Although this system is useful for the detec-
tion of saccharides in the visible wavelength region, some of
their results are curious to our eyes and quite difficult to under-
stand. First, the pKa of the boronic acid groups was estimated
to be 5.9, which is much lower than those for conventional
boronic acids (≈9).2,7,17 Secondly, it is known that the pKa is
lowered by ≈ 2 pK units upon complexation with sacchar-
Chart 1
J. Chem. Soc., Perkin Trans. 1, 2000, 407–413
This journal is © The Royal Society of Chemistry 2000
407

View Article Online
Fig. 1 pH Dependence of the absorption spectra for 1: buffered
water–Me2SO = 1:1 v/v, [1] = 5.00 × 10Ϫ5 mol dmϪ3, 25 ЊC.
Results and discussion
Determination of pKa-values
Fig. 2 Plots of A370 vs. pH for (A) 1 and (B) 2 in the absence and the
presence of -fructose (100 mmol dmϪ3): buffered water–Me2SO = 1:1
v/v, [1] or [2] = 5.00 × 10Ϫ5 mol dmϪ3, 25 ЊC. -᭜-, In the presence of
-fructose, -᭿-, in the absence of -fructose.
The absorption spectral measurements were carried out in a
water–Me2SO (1:1 v/v) mixed solvent. The medium pH was
adjusted with 50 vol% water using AcOH–AcONa for pH 4.5–
5.0, KH2PO4–Na2HPO4 for pH 5.5–8.5 and NaHCO3–Na2CO3
for pH 9.0–11.0 (50 mmol dmϪ3 for each), buffer constituents
which do not or scarcely interact with the boronic acid groups.
To confirm that 1 and 2 are completely dissolved into these
media we measured the absorption spectra as a function of [1]
or [2]. Plots of A370 vs. [1] or [2] showed a good linear relation-
ship over 0–1.50 mmol dmϪ3 (data not shown). The result sup-
ports the view that this is a “homogeneous” system satisfying
the Lambert–Beer law.
Next, the pKa-values were estimated by photometric titration
in the absence and the presence of saccharide (D-fructose
which shows the highest affinity with monoboronic acids2,7,10
was chosen). The spectral change observed for 1 in the absence
of saccharide is shown in Fig. 1. It is seen from Fig. 1 that there
is a large difference in absorbance between pH 8.5 and 9.0.
Plots of A370 vs. pH (Fig. 2A) establish that the titration curves
are satisfactorily simulated with a single pKa of 8.9 in the
absence, and 5.3 in the presence, of -fructose. Hence, one can
speculate that the two boronic acid groups form anionic OHϪ
adducts simultaneously or nearly so in the same pH region.
Since the affinity of -fructose with boronic acids is very
high,2,7,10 one can assume that the boronic acid groups in 1 are
totally converted to the -fructose complexes in the presence of
100 mmol dmϪ3 -fructose. This implies that the acidity of the
boronic acid groups is intensified by 3.6 pK units by -fructose
complexation. From the similar A370 vs. pH plots the pKa-values
for 2 were estimated to be 8.4 in the absence, and 5.5 in the
presence, of -fructose (100 mmol dmϪ3: Fig. 2B).
Fig. 3 CD spectral change in 1 (5.00 × 10Ϫ5 mol dmϪ3) in the presence
of -fructose (0–5.00 × 10Ϫ4 mol dmϪ3): water (pH 10.0 with 50 mmol
dmϪ3 carbonate)–Me2SO = 1:1 v/v, 25 ЊC.
The foregoing results consistently indicate that the ‘true’ pKa-
values for 1 and 2 are much higher than those reported by Yam
and Kai (i.e., 5.9)16 and rather closer to those of conventional
boronic acids (i.e., ≈9).2,7,17
Fig. 4 Continuous-variation plot monitored at CDmax (370 nm):
[1] ϩ [-fructose] = 1.00 × 10Ϫ4 mol dmϪ3 (constant): water (pH 10.0
with 50 mmol dmϪ3 carbonate)–Me2SO = 1:1 v/v, 25 ЊC.
Saccharide-binding properties at pH 10.0
First, the saccharide-binding properties were estimated at pH
10.0 (50 mmol dmϪ3 NaHCO3–Na2CO3) and 25 ЊC where the
boronic acid groups are converted to the OHϪadducts and the
Kass-values are apparently enhanced owing to the high OHϪ
concentration. In this medium, the absorption spectra were
scarcely changed by the saccharide addition. This implies that
a change in the boronic acid group from R-BϪ(OH)3 to R-
(HO)BϪ〈OO〉(saccharide) scarcely affects the absorbance of the
chromophoric ReIؒbipyridine moiety. However, the saccharide
binding could be conveniently monitored by a change in CD
spectroscopy.
which is the same wavelength as the λmax (shoulder) in absorp-
tion spectroscopy. It is known that the appearance of CD bands
in diboronic acid receptors is due to the rigidification effect
arising from the formation of 1:1 cyclic complexes.6,7,9,13,14,19,20
In fact, a continuous-variation plot (Job’s plot)21 at [1] ϩ
[-fructose] = 1.00 × 10Ϫ4 mol dmϪ3 (constant) gave a maxi-
mum at [1]/([1] ϩ [-fructose]) = 0.5 (Fig. 4), indicating that the
CD-active species is the 1:1 cyclic complex.
It is seen from Fig. 3 that the CD intensity first increases
up to 2 equivalents of [-fructose]/[1] and then decreases with
further increase in the -fructose concentration. This biphasic
As shown in Fig. 3, a negative CD band appeared at 370 nm
408
J. Chem. Soc., Perkin Trans. 1, 2000, 407–413

View Article Online
Table 1 Kass (1:1) and Kass (1:2) for 1 ϩ saccharide systems
involved in this classification. Presumably, complexation of the
first boronic acid group with the 1,2-cis-diol group governs the
twisting direction of the ReIؒbipyridine moiety.
Kass (1:1)
/
Kass (1:2)/
Saccharide
dm3 molϪ1
dm3 molϪ1
Saccharide-binding properties at pH 6.5
-Fructose
-Galactose
-Arabinose
-Mannose
-Glucose
2.0 × 105
1.5 × 105
1.5 × 105
1.0 × 105
5.0 × 104
1.5 × 103
7.0 × 102
6.0 × 102
2.5 × 102
2.0 × 102
At pH 10.0, above the pKa, the boronic acid groups show high
affinity with saccharides and exist as sp3-hybridised boronate
anions both before and after the saccharide complexation. As a
result, 1 could form a 1:1 cyclic complex (as evidenced by CD
spectroscopy) whereas the absorption spectra scarcely changed.
Next, we carried out spectral measurements at pH 6.5, which is
higher than the pKa in the absence of saccharides but lower
than that in the presence of saccharides (e.g., 8.9 in the presence
of -fructose: see Fig. 2A). Hence, the absorption spectra
would be changed by the saccharide complexation but the
formation of a 1:1 cyclic complex would become more difficult
because of the low affinity with saccharides.
When saccharides were added to a solution of 1 adjusted to
pH 6.5, the absorption spectral change was really induced. The
spectral change was very similar to that induced by the pH
change (Fig. 1), indicating that this change is attributable
to a conversion of sp2-hybridised R-B(OH)2 to sp3-hybridised
R-(OH)BϪ〈O〉(saccharide). Plots of A370 vs. saccharide concen-
trations gavOe apparent saturation curves (Fig. 7). It is supposed
that Yam and Kai16 calculated the Kass-values from an analysis
of these plots, assuming the formation of 1:1 cyclic com-
plexes.16 However, this system is not so simple at pH 6.5. For
example, it is not clear yet whether 1 forms only 1:1
cyclic complexes. The validity of this assumption may be con-
firmed by a continuous-variation plot. As shown in Fig. 8, the
plot for -fructose apparently gave a maximum at [1]/([1] ϩ
[-fructose]) = 0.5. This implies that 1 tends to form a 1:1 com-
plex with -fructose, at least, at this concentration region (i.e.,
[1] ϩ [-fructose] = 1.00 × 10Ϫ3 mol dmϪ3). However, the solu-
tion was totally CD-silent even at [1]/([1] ϩ [-fructose]) = 0.5
where the macrocycle formation becomes most advantageous.
The result supports the view that this 1:1 complex is not cyclic
and -fructose is appended only to one boronic acid group.
This is due to the low saccharide affinity at pH 6.5. As described
above, the phototitration data (Fig. 1) showed that the two
boronic acid groups act independently, forming the OHϪ
adduct simultaneously or nearly so in the same pH region. This
suggests that the saccharide complexation, which is similar to
the OHϪ adduct formation, should also occur independently.
This implies that, in Fig. 7, not only 1:1 noncyclic complexes
but also 1:2 noncyclic complexes can be formed (particularly at
[saccharide] > [1]). In fact, the absorbance measured at pH
6.5 and [-fructose] = 3.00 × 10Ϫ2 mol dmϪ3 was 0.47, which
corresponds to the absorbance of two boronate anionic groups.
The results imply, therefore, that at [-fructose] = 1.00 × 10Ϫ3
mol dmϪ3 (Job’s plot: Fig. 8) 1 tends to form the 1:1 complex
whereas at [-fructose] = 3.00 × 10Ϫ2 mol dmϪ3 (Fig. 7) 1 tends
to form the 1:2 complex. Therefore, calculation of Kass from the
plots in Fig. 7, assuming the formation of 1:1 complexes, is
meaningless.
Fig.
5
Plots of [θ]370 vs. (A) [-fructose]/[1] and (B) [other
saccharides]/[1]: water (pH 10.0 with 50 mmol dmϪ3 carbonate)–
Me2SO = 1:1 v/v, [1] = 5.00 × 10Ϫ5 mol dmϪ3, 25 ЊC.
binding manner (Fig. 5) is frequently observed for diboronic
acid receptors and can be explained by the primary step to
form CD-active 1:1 cyclic complexes at low saccharide
concentration, followed by the secondary step to form CD-
silent 1:2 receptor/saccharide noncyclic complexes at high
saccharide concentration.6,7,9,13,14,19,20 Fig. 5A shows that in the
1 ϩ -fructose system the maximum concentration for the 1:1
cyclic complex appears at [-fructose]/[1] = 2.0. The biphasic
dependence was analysed using a nonlinear least-squares
method22 to estimate the association constants for the 1:1
cyclic complex [Kass (1:1)] and the 1:2 receptor/saccharide non-
cyclic complexes [Kass (1:2)]. The similar biphasic dependence
was also observed for other saccharides such as -arabinose,
-galactose, -mannose and -glucose (Fig. 5B). The Kass (1:1)
-
and Kass( 1:2)-values for these saccharides were estimated by
the same method and are summarised, together with those for
-fructose, in Table 1.
Fig. 6 shows the partial CD spectra in the presence of nine
saccharides. The strongest CD spectra which are obtained at
the maximum 1:1 cyclic complex concentration are recorded.
Since the CD spectrum in the presence of -glucose was so
weak, the concentration was enhanced to 5.00 × 10Ϫ4 mol dmϪ3
(10-times higher than the standard concentration). Careful
examination of the CD sign at around 370 nm reveals that there
exists a close correlation between the CD sign and the mono-
saccharide configuration. When the 2-OH group is ‘down’, the
CD sign is always positive (Group A). -Xylose, -glucose,
-galactose, -ribose and -allose are classified in this category.
When the 2-OH (in -fructose this OH group is counted as
3-OH) group is ‘up’, the CD sign is always negative (Group B).
-Mannose, -talose and -arabinose are classified in this
category. Although the structure of -fructose (ketose) is
somewhat different from others, the 2-OH is ‘up’ and the CD
sign is negative (Fig. 3). Hence, this saccharide can be also
Saccharide-binding properties of an N-ethylated reference
compound (2)
The spectral properties of 2 for saccharide-binding at pH 6.5
were very similar to those of 1. For example, plots of ∆A370 vs.
saccharide concentration gave saturation curves as shown in
Fig. 9. A continuous-variation plot for -fructose resulted in a
maximum at [2]/([2] ϩ [-fructose]) = 0.5, indicating that 2 also
forms a 1:1 complex (Fig. 10). On the other hand, the CD band
did not appear at any 2/-fructose ratio. The CD spectra were
also measured for other saccharides but none of the 2 ϩ sacchar-
ide systems became CD-active at pH 6.5. The foregoing results
support the view that 2 also forms 1:1 noncyclic complexes
with saccharides at pH 6.5.
J. Chem. Soc., Perkin Trans. 1, 2000, 407–413
409

View Article Online
Fig. 6 Partial CD spectra of 1 (5.00 × 10Ϫ5 mol dmϪ3; 5.00 × 10Ϫ4 mol dmϪ3 for -glucose) obtained in the presence of various saccharides. The CD
spectra are those around the maximum 1:1 cyclic complex concentration (i.e., [saccharide]/[1] = 2.0 for -galactose, 5.0 for -arabinose and
-mannose, and 1.5 for -glucose). Saccharide structures are illustrated with their furanose forms.
The amide protons in 1 might participate in the binding of
saccharides through a hydrogen-bonding interaction even in
aqueous solution. It is seen, however, that the curves for 2 in
Fig. 9 are very similar to those for 1 in Fig. 7. This implies that
the association properties for 2 to form 1:1 noncyclic com-
plexes are not much different from those for 1 (although it is not
correct to estimate the Kass directly from these curves). Hence, it
seems to be unnecessary to take the additional participation of
the hydrogen-bonding interaction into consideration.
In high-pH region, Yam and Kai considered that the amide
groups are dissociated into their anionic groups, which partici-
pate in saccharide binding through hydrogen-bonding inter-
actions.16 We observed that with increasing medium pH up to
12.5 the absorption spectrum of 1 changes whereas that of 2
does not change, supporting the fact that deprotonation occurs
from 1. However, this significant difference was not seen
between 1 and 2 when -fructose (1.00 × 10Ϫ3 mol dmϪ3) was
added. We consider, therefore, that there exists no strong
evidence for the interaction between the amide anions and the
saccharide guest.
As shown in Fig. 11, a 2 ϩ -fructose system gave only a very
weak CD band.23 Molecular modelling suggests that when the
distance between the two boronic acid groups becomes the
shortest possible (that is, the conformation illustrated in Chart
1), the N-ethyl groups must be directed toward the inward-
position. This conformation is seriously hampered by steric
crowding between the ethyl groups. This should cause some
rearrangement in the N-ethyl groups and lessen the Kass. Assum-
410
J. Chem. Soc., Perkin Trans. 1, 2000, 407–413

View Article Online
Fig. 7 Plots of absorbance difference at 370 nm (∆A370) vs. saccharide
concentration at pH 6.5: water (pH 6.5 with 50 mmol dmϪ3 phosphate)–
Me2SO = 1:1 v/v, [1] = 5.00 × 10Ϫ5 mol dmϪ3, 25 ЊC.
Fig. 10 Continuous-variation plot for a 2 ϩ -fructose system at pH
6.5: water (pH 6.5 with 50 mmol dmϪ3 phosphate)–Me2SO = 1:1 v/v,
[2] ϩ [-fructose] = 1.00 × 10Ϫ3 mol dmϪ3 (constant), 25 ЊC.
Fig. 8 Continuous-variation plot for a 1 ϩ -fructose system at pH
6.5: water (pH 6.5 with 50 mmol dmϪ3 phosphate)–Me2SO = 1:1 v/v,
[1] ϩ [-fructose] = 1.00 × 10Ϫ3 mol dmϪ3 (constant), 25 ЊC.
Fig. 11 CD spectrum of 2 in the presence of -fructose at pH 10.0:
water (pH 10.0 with 50 mmol dmϪ3 carbonate)–Me2SO = 1:1 v/v,
[2] = 5.00 × 10Ϫ4 mol dmϪ3, [-fructose] = 0–1.00 × 10Ϫ3 mol dmϪ3
25 ЊC.
,
properties are not so simple as proposed by Yam and Kai, the
molecular design of 1 is still meaningful, particularly for sac-
charide sensing using longer-wavelength radiation. We believe
that further modification of 1 based on skillful molecular
design would enhance both the affinity and the selectivity for
saccharides and enable us to read the saccharide-binding
process directly in visible light region.
Experimental
Materials
2,2Ј-Bipyridine-4,4Ј-dicarboxylic acid was synthesised from
4,4Ј-dimethyl-2,2Ј-bipyridine according to the literature
method.24
Fig. 9 Plots of absorbance difference at 370 nm (∆A370) vs. saccharide
concentration at pH 6.5: water (pH 6.5 with 50 mmol dmϪ3 phosphate)–
Me2SO = 1:1 v/v, [2] = 5.00 × 10Ϫ5 mol dmϪ3, [saccharide] = 0–50 mmol
dmϪ3, 25 ЊC.
3-Acetamidophenylboronic acid 3. To a THF (50 ml) solution
containing 3-aminophenylboronic acid (2.0 g, 12.9 mmol) and
triethylamine (8.96 ml, 64.5 mmol) was gradually added acetyl
bromide (1.9 ml, 25.8 mmol). The mixture was stirred for 14 h
at room temperature. The precipitate was removed by filtration,
the filtrate being washed with saturated aq. NaCl and dried
over MgSO4. Evaporation of this solution yielded a yellow
powder, the purity of which was sufficient for its use in the next
reaction without further purification: (2.1 g), mp 138–140 ЊC;
δH (250 MHz; DMSO-d6) 3.58 (3H, s, CH3), 6.50 (2H, m, 4-,
5-ArH), 6.68 (1H, s, NH), 6.78 (1H, d, 6-ArH), 6.95 (1H, s,
2-ArH).
ing that the major species is still a 1:1 cyclic complex, the
Kass (1:1) was estimated to be roughly 4 × 103 dm3 molϪ1 from the
CD spectra.
Concluding remarks
From the foregoing findings based on our careful re-
examination, we can now derive the following conclusions
which are not necessarily compatible with those proposed by
Yam and Kai.16 First, the pKa is not so low as proposed by
them16 but, rather, is close to the conventional value for boronic
acids. Secondly, 1 does not form 1:1 cyclic complexes at pH-
values below the pKa. Thirdly, the Kass-values should not be
estimated directly from the plot of ∆A370 vs. [saccharide].
Fourthly, neither the CONH groups nor the CONϪ groups
significantly contribute to the saccharide binding in aqueous
solution. We consider that although the saccharide-binding
3-(N-Ethylamino)phenylboronic acid 4. Compound 3 (1.5 g,
7.6 mmol) was dissolved in dehydrated THF (100 ml) in a N2-
substituted flask. The solution was cooled in an ice-bath and
BH3ؒMe2S (7.2 ml, 76 mmol) was gradually added. The mixture
was heated at 60 ЊC for 6 h and then concentrated to dryness.
The residue was mixed with 6 mol dmϪ3 aq. HCl (70 ml) and the
J. Chem. Soc., Perkin Trans. 1, 2000, 407–413
411

View Article Online
mixture was refluxed for 2 h. It was neutralised by Na2CO3 and
extracted with diethyl ether. The organic solution was dried
over MgSO4 and evaporated to dryness to give a slightly yellow
powder, yield from 3-aminophenylboronic acid 71.6%, mp 48–
50 ЊC; δH (250 MHz; DMSO-d6) 1.15 (3H, t, CH3), 3.03 (2H, q,
CH2), 5.64 (1H, br, NH), 6.60 (1H, s, 4-ArH), 6.98–7.07 (3H, m,
2-, 5-, 6-ArH), 7.81 (2H, s, BOH).
(52 mg, 50%); mp > 320 ЊC (decomp.) (Found: C, 40.70; H,
2.66; N, 7.03. C27H20N4B2ClO9Re ϩ 0.5 H2O requires C, 40.83;
H, 2.49; N, 6.91%); IR (KBr)/cmϪ1 3420 (O–H), 2027, 1948,
1909 (M–C᎐O), 1647 (C᎐O), 1339 (B–O); δ (250 MHz; CDCl )
᎐
᎐
H
3
7.40 (2H, t, 5-ArH), 7.63 (2H, d, 6-ArH), 7.92 (2H, d, 4-ArH),
8.07 (2H, s, 2-ArH), 8.24 (2H, s, 5-PyH), 9.24–9.28 (4H, m, 3-,
6-PyH), 10.83 (2H, s, BOH).
N-Ethyl-3-(1,3,2-dioxaborinan-2-yl)aniline 5. To protect the
boronic acid group, compound 4 (1.2 g, 6.56 mmol) was treated
with propane-1,3-diol (0.95 ml, 13.1 mmol) in toluene (100 ml)
using a Dean–Stark apparatus. After treatment at reflux for
14 h, the solution was evaporated to dryness. The product was
purified by column chromatography (silica gel–ethyl acetate) to
give a slightly yellow oil (1.5 g, quant.); δH (250 MHz; CDCl3)
1.20 (3H, t, CH3), 2.00 (2H, qv, CH2CH2CH2), 3.14 (2H, q,
CH3CH2), 4.21 (4H, t, OCH2), 6.64 (1H, d, 6-ArH), 7.06–7.21
(3H, m, 2-, 4-, 5-ArH).
Miscellaneous
For the phototitration of compounds 1 and 2, the following
buffer solutions were used: 50 mmol dmϪ3 acetate for pH 4.5–
5.0, 50 mmol dmϪ3 phosphate for pH 6.5–9.0, 50 mmol dmϪ3
carbonate for pH 9.5–11.0. Absorption spectra, CD spectra
1
and H NMR spectra were measured with Shimadzu 2500PC,
JASCO J-720 and Bruker AC-205P instruments, respectively.
References and notes
4,4Ј-Bis{N-[3-(1,3,2-dioxaborinan-2-yl)phenyl]-N-ethylcarb-
amoyl}-2,2Ј-bipyridine 6. 2,2Ј-Bipyridine-4,4Ј-dicarboxylic acid
(2.0 g, 8.2 mmol) was converted to 4,4Ј-bis(chlorocarbonyl)-
2,2Ј-bipyridine by treatment with SOCl2. To a THF solution
(50 ml) containing compound 5 (3.0 g, 18.1 mmol) and triethyl-
amine (12.5 ml, excess) was added dropwise a THF solution
containing 4,4Ј-bis(chlorocarbonyl)-2,2Ј-bipyridine prepared
above. The mixture was stirred for 15 h at room temperature.
The precipitate was removed by filtration, the filtrate being
concentrated to dryness. The product 6 was isolated by GPC
(Japan Industry Co., LC-908; JAIGEL-2H, chloroform). The
title product was obtained as a white powder (130 mg, 6.8%),
mp 194–195 ЊC; IR (KBr)/cmϪ1 1647 (C᎐O), 1308 (B–O); δ
1 For comprehensive reviews on hydrogen-bond-based synthetic
receptors, see: J. Rebek, Jr., Angew. Chem., Int. Ed. Engl., 1990, 29,
245; A. D. Hamilton, Bioorg. Chem. Front., 1991, 2, 115; J.-H.
Fuhrhop and J. Köning, Membranes and Molecular Assemblies: The
Synkinetic Approach, ed. J. F. Stoddart, Royal Society of Chemistry,
Cambridge, 1994; G. M. Whitesides, E. E. Simanek, J. P. Mathias,
T. S. Christopher, N. C. Donovan, M. Mathai and M. G. Dana, Acc.
Chem. Res., 1995, 28, 37.
2 J. Yoon and A. W. Czarnik, J. Am. Chem. Soc., 1992, 114, 5874;
L. K. Mohler and A. W. Czarnik, J. Am. Chem. Soc., 1993, 115,
2998.
3 P. R. Westmark and B. D. Smith, J. Am. Chem. Soc., 1994, 116, 9343
and reference cited therein.
4 Y. Nagai, K. Kobayashi, H. Toi and Y. Aoyama, Bull. Chem. Soc.
Jpn., 1993, 66, 2965.
5 G. Wulff, S. Krieger, B. Kubneweg and A. Steigel, J. Am. Chem.
Soc., 1994, 116, 409 and references cited therein; G. Wulff, Pure
Appl. Chem., 1982, 54, 2093.
6 K. Tsukagoshi and S. Shinkai, J. Org. Chem., 1991, 56, 4089;
Y. Shiomi, M. Saisho, K. Tsukagoshi and S. Shinkai, J. Chem. Soc.,
Perkin Trans. 1, 1993, 2111; J. C. Norrild and H. Eggert, J. Am.
Chem. Soc., 1995, 117, 1479.
7 H. Suenaga, M. Mikami, K. R. A. S. Sandanayake and S. Shinkai,
Tetrahedron Lett., 1995, 36, 4825; H. Murakami, T. Nagasaki,
I. Hamachi and S. Shinkai, J. Chem. Soc., Perkin Trans. 2, 1994, 975;
T. Kimura, S. Arimori, M. Takeuchi, T. Nagasaki and S. Shinkai,
J. Chem. Soc., Perkin Trans. 2, 1995, 1889; H. Shinmori, M.
Takeuchi and S. Shinkai, J. Chem. Soc., Perkin Trans. 2, 1996, 1.
8 For boronic-acid-appended porphyrins, see: H. Toi, Y. Nagai,
H. Kawabe, K. Aizawa and H. Ogoshi, Chem. Lett., 1993, 1043;
T. Imada, H. Murakami and S. Shinkai, J. Chem. Soc., Chem.
Commun., 1994, 1557; T. Imada, H. Kijima, M. Takeuchi and
S. Shinkai, Tetrahedron, 1996, 52, 2817; H. Murakami, T. Nagasaki,
I. Hamachi and S. Shinkai, J. Am. Chem. Soc., 1996, 118, 245.
9 For comprehensive reviews on boronic-acid-based saccharide
receptors, see: T. D. James, K. R. A. S. Sandanayake and S. Shinkai,
Supramol. Chem., 1995, 6, 141; T. D. James, P. Linnane and
S. Shinkai, Chem. Commun., 1996, 281; T. D. James, K. R. A. S.
Sandanayake and S. Shinkai, Angew. Chem., Int. Ed. Engl., 1996, 35,
1911; S. Shinkai and M. Takeuchi, Trends Anal. Chem., 1996, 15,
418.
᎐
H
(250 MHz; CDCl3) 1.21 (6H, t, CH3), 2.02 (4H, qv, CH2-
CH2CH2), 3.97 (4H, q, CH3CH2), 4.11 (8H, t, OCH2), 6.97–7.05
(4H, m, 4-, 6-ArH), 7.16 (2H, t, 5-ArH), 7.50–7.58 (4H, m,
2-ArH, 5-PyH), 8.24 (2H, s, 3-PyH), 8.34 (2H, d, 6-PyH).
4,4Ј-Bis{[3-(dihydroxyboryl)phenyl]carbamoyl}-2,2Ј-bipyrid-
ine 7. Compound 7 was synthesised from 3-(1,3,2-dioxaborinan-
2-yl)aniline and 4,4Ј-bis(chlorocarbonyl)-2,2Ј-bipyridine. The
precipitate was recovered by filtration and washed with water
to remove triethylamine hydrochloride and unchanged com-
pounds. The product was obtained as a white powder: 85%
yield, mp 328–330 ЊC; IR (KBr)/cmϪ1 3288 (O–H), 1643 (C᎐O),
᎐
1304 (B–O); δH (250 MHz; D2O ϩ NaOD) 7.06 (2H, d, 4-ArH),
7.30–7.33 (6H, m, 2-, 5-, 6-ArH), 7.94 (2H, d, 5-PyH), 8.45 (2H,
s, 3-PyH), 8.80 (2H, d, 6-PyH).
Tricarbonylchloro(4,4Ј-bis{N-[3-(1,3,2-dioxaborinan-2-yl)-
phenyl]-N-ethylcarbamoyl}-2,2Ј-bipyridine)rhenium 2Ј. Com-
pound 6 (20 mg, 4.2 × 10Ϫ5 mol) and Re(CO)5Cl (15 mg,
4.2 × 10Ϫ5 mol) were dissolved in dry ethanol in a N2-
substituted flask. The solution was refluxed for 14 h. Evapor-
ation resulted in a crude solid product, which was recrystallised
from chloroform to give compound 2Ј as an orange powder (52
mg, 50%), mp 132–134 ЊC; (Found: C, 45.79; H, 3.74; N, 5.70.
C37H36N4B2ClO9Re ϩ 0.5 CHCl3 requires C, 45.88; H, 3.97;
10 J. P. Lorand and J. D. Edwards, J. Org. Chem., 1959, 24, 769.
11 K. R. A. S. Sandanayake and S. Shinkai, J. Chem. Soc., Chem.
Commun., 1994, 1083.
N, 5.50%); IR (KBr)/cmϪ1 2021, 1917, 1890 (M–C᎐O), 1636
12 K. Koumoto, M. Takeuchi and S. Shinkai, Supramol. Chem., 1998, 9,
᎐
203.
(C᎐O), 1312 (B–O); δ (250 MHz; DMSO-d ) 1.28 (6H, t, CH ),
᎐
H
6
3
13 T. D. James, K. R. A. S. Sandanayake, R. Iguchi and S. Shinkai,
J. Am. Chem. Soc., 1995, 117, 8982; T. D. James, K. R. A. S.
Sandanayake and S. Shinkai, Nature, 1995, 374, 345.
14 K. R. A. S. Sandanayake, T. D. James and S. Shinkai, Chem. Lett.,
1995, 503.
15 H. Matsumoto, A. Ori, F. Inokuchi and S. Shinkai, Chem. Lett.,
1996, 301.
16 V. W.-W. Yam and A. S.-F. Kai, Chem. Commun., 1998, 109.
17 H. Nakatani, T. Morita and K. Hiromi, Biochim. Biophys. Acta,
1978, 525, 423; J. Juillard and N. C. R. Geugue, C. R. Hebd. Seances
Acad. Sci., Ser. C, 1967, 264, 259.
18 L. K. Mohler and A. W. Czarnik, J. Am. Chem. Soc., 1993, 115,
7037.
2.06 (4H, qv, CH2CH2CH2), 4.05 (4H, q, CH3CH2), 4.13 (8H, t,
OCH2), 7.05–7.10 (4H, m, 4-, 6-ArH), 7.33 (2H, t, 5-ArH), 7.57
(2H, t, 2-ArH), 7.71 (2H, t, 5-PyH), 8.15 (2H, s, 3-PyH), 8.65
(2H, d, 6-PyH).
Compound 2Ј was directly used for spectral measurements.
The results were basically identical with those of 2 (a small
amount was obtained by deprotection).
Tricarbonylchloro(4,4Ј-bis{[3-(dihydroxyboryl)phenyl]carb-
amoyl}-2,2Ј-bipyridine)rhenium 1. Compound 1 was synthesised
from 7 and Re(CO)5Cl in a manner similar to that used for 2:
412
J. Chem. Soc., Perkin Trans. 1, 2000, 407–413

View Article Online
19 K. Kondo, Y. Shiomi, M. Saisho, T. Harada and S. Shinkai,
Tetrahedron, 1992, 48, 8239.
20 T. D. James and S. Shinkai, J. Chem. Soc., Chem. Commun., 1995,
1483.
11237 and references cited therein. If this is the case in 2, two
interconvertible enantiomers can exist and the resultant -fructose
complexes become diastereomeric. The 1H NMR spectrum (250
MHz) shows only one set of sharp triplet–quartet peaks. This
1
21 H. A. Benesi and J. H. Hildebrand, J. Am. Chem. Soc., 1949, 71,
indicates that cis–trans isomerism occurs faster than the H NMR
2703.
time-scale. This may be the reason why the CD intensity and the
Kass (1:1) are smaller than those for 1.
22 H. Shinmori, M. Takeushi and S. Shinkai, J. Chem. Soc., Perkin
Trans. 2, 1998, 847.
23 One referee raised an interesting comment on the known
conformational preference of N-alkyl anilides to adopt a cis amide
conformation: I. Azumaya, I. Okamoto, S. Nakayama, A. Tanatani,
K. Yamaguchi, K. Shudo and H. Kagechika, Tetrahedron, 1999, 55,
24 G. Sprintschnik, H. W. Sprintschnik, P. P. Kirsch and D. G. Whitten,
J. Am. Chem. Soc., 1977, 99, 4947.
Paper a906708f
J. Chem. Soc., Perkin Trans. 1, 2000, 407–413
413