Highly selective lipophilic diboronic acid that transports fructose as the trisdentate 2,3,6-β-D-fructofuranose ester

Article abstract of DOI:10.1016/j.tet.2003.09.068

The transport of fructose and glucose through supported liquid membranes promoted by a tetrahedral shaped lipophilic monoboronic acid and diboronic acid, both based on a pentaerythritol core, has been studied. The diboronic acid gave a high fructose/gluco

Full text of DOI:10.1016/j.tet.2003.09.068

Tetrahedron 59 (2003) 9075–9082
Highly selective lipophilic diboronic acid that transports fructose
as the trisdentate 2,3,6-b-D-fructofuranose ester
†
b
,
a
Scott P. Draffin,a Peter J. Duggan,a Sandhya A. M. Duggan and Jens Chr. Norrild
,
*
aSchool of Chemistry, Monash University, Clayton, Melbourne, Vic. 3800, Australia
bDepartment of Chemistry, University of Copenhagen, Universitetsparken 5, DK-2100 Copenhagen, Denmark
Received 22 August 2003; revised 12 September 2003; accepted 18 September 2003
Abstract—The transport of fructose and glucose through supported liquid membranes promoted by a tetrahedral shaped lipophilic
monoboronic acid and diboronic acid, both based on a pentaerythritol core, has been studied. The diboronic acid gave a high fructose/glucose
selectivity (7.7:1.0). The results of molecular modelling studies and 13C NMR experiments with uniformly labelled 13C-D-fructose suggest
that the enhanced selectivity results from the formation of a trisdentate 2,3,6-b-D-fructofuranose diester within the membrane. Experimental
evidence for a previously proposed macrocyclic D-fructopyranose diester formed with an o-phenylene linked diboronic acid has also been
found.
q 2003 Elsevier Ltd. All rights reserved.
1. Introduction
selectively transport carbohydrates across lipophilic mem-
branes have potential applications in drug delivery4,5 and in
environmentally benign industrial sugar production.5b,6
With the latter application in mind, we have been
investigating the inherent preference of aryl boronic acids
to transport fructose over other sugars, with the ultimate aim
of developing a new method for the production of high
fructose syrup.
It is almost fifty years since the esterification of polyols with
aryl boronic acids was first examined,1 and applications of
this reversible process in organic synthesis have been
investigated for almost as long.2 Shinbo and co-workers
were the first to demonstrate that an aryl boronic acid, in
combination with a lipophilic ammonium ion, could
facilitate the passage of monosaccharides through a
lipophilic liquid membrane.3 Boronic acids that are able to
Figure 1 shows how the transport of a sugar out of an
alkaline aqueous departure phase, through a lipophilic
membrane and into a slightly acidic aqueous receiving
phase can be promoted by a boronic acid. This transport
process is thought to be diffusion controlled,5 with the
formation of the boronate esters at the interface being rapid
and reversible. Since, in this mechanism, every sugar
molecule is co-transported with a hydroxide ion, sugar
transport can be driven uphill via the application of a pH
gradient across the membrane.6
Membrane stability and ease of preparation of carriers are of
prime importance if this process is to become industrially
feasible. In our early experiments with bulk liquid
membranes (BLM’s),6 we tried to address these issues
through the use of a boronic acid derived from cholesterol
(1, Fig. 2).7 This highly lipophilic boronic acid, which is
expected to be resistant to leaching, unfortunately possesses
very poor transport properties in a BLM, with virtually no
sugar transport being observed under standard conditions.
We suggested that steroidal boronic acids of this type
probably self assemble at the membrane interface, inhibi-
ting transport.6 This seems reasonable, given that 1
possesses liquid crystalline properties.7
Figure 1. Accepted ‘tetrahedral’ transport mechanism for the passage of
sugars through a lipophilic membrane promoted by an aryl boronic acid and
a quaternary ammonium salt (QþX2).
Keywords: boronic acids; membrane transport; monosaccharides.
*
Corresponding author. Tel.: þ61-3-9905-4571; fax: þ61-3-9905-4597;
e-mail: peter.duggan@sci.monash.edu.au
Present address: PreciSense A/S, Dr Neergaards Vej 3, DK-2970
†
Hørsholm, Denmark. Fax: þ45-45-57-09-10; jcn@precisense.com
0040–4020/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2003.09.068

9076
S. P. Draffin et al. / Tetrahedron 59 (2003) 9075–9082
2. Results and discussion
2.1. Carrier design and synthesis
In order to circumvent any possible unfavourable stacking
interactions at the aqueous-membrane interface, an overall
tetrahedral shape, derived from a pentaerythritol core, was
chosen for the new boronic acid carriers. High lipophilicity
was incorporated through the attachment of p-tert-octyl-
phenyl ethers, while the boronic acid functionality was
attached through an ester link to p-carboxyphenyl boronic
acid. Following this strategy, a mono- and diboronic acid (7
and 10) were synthesised, as outlined in Scheme 1. The use
of a pinacol protecting group for the boronic acid appears to
minimise side reactions during the alkylation of the
carboxylic acid (5) and makes the boron-containing
products (6 and 9) easier to isolate and purify. The
deprotection of these boronates with aq. HCl/acetone,
although slow, was found to be very a clean process.
Figure 2. Previously described monoboronic acids used in this study.
More recently, we have been studying sugar transport
through supported liquid membranes (SLM’s) because these
membranes are more likely to be industrially useful than
BLM’s, and experiments involving SLM’s require much
lower quantities of carrier.8 Here we report the transport
properties exhibited by 1 in an SLM, together with those of
a new class of carriers based on a pentaerythritol core.9 The
study of the sugar transport properties of these new carriers,
as well their interaction with 13C-labelled fructose has
allowed us to better understand the sugar transport process
promoted by boronic acid carriers, and make predictions on
how saccharide selectivity can be improved.
2.2. Sugar transport
In the production of high fructose syrup, one of the critical
steps is the enrichment of the fructose concentration in
mixtures that contain similar quantities of glucose and
fructose. Hence, in this study, we looked for the enhance-
ment of fructose flux over that of glucose. Competitive SLM
transport fluxes were determined using apparatus described
previously.8a The liquid membrane consisted of the carriers
Scheme 1. Synthesis of tetrahedral shaped lipophilic boronic acids.

S. P. Draffin et al. / Tetrahedron 59 (2003) 9075–9082
9077
Table 1. Sugar fluxes through a supported liquid membranea
promoted by 7 is approximately 30% lower than that of the
lighter monoboronic acid (2), which has a molecular weight
of less than half of that of 7 (415 g mol21 cf. 849 g mol21).
Entry
Boronic acid
Flux
(1028 mol m22 s21
ratio of fluxes
)
Fructose
Glucose
More remarkable are the transport properties of the
diboronic acid (10). This compound, while also showing
no evidence of leaching during the course of transport
experiments, exhibits unexpectedly high selectivity for
fructose over glucose (Table 1, entry 4). This selectivity
stems from an enhanced fructose flux, compared with that of
7, a flux that is very similar to that promoted by the much
lighter monoboronic acid (2; 415 g mol21 cf. 809 g mol21
for 10). In the following two sections, evidence is presented
that supports the notion that the enhanced fructose flux
promoted by 10 stems from its ability to simultaneously
carry 2 equiv. of fructose through the membrane in the form
of a stabilised fructofuranose diester.
1
2
3
4
5b
2
1
7
10
11
28.4
7.6
17.9
26.1
21
6.7
4.1
3.5
3.4
3.4
4.2
1.9
5.1
7.7
6.2
a
[Boronic acid] in membrane¼50 mM, fluxes shown are averages of 2–3
runs, T¼298 K, flux uncertainty^10%.
Data obtained in a previous study under similar conditions.8a
b
dissolved in 2-nitrophenyl octyl ether supported by a thin,
flat sheet of microporous polypropylene (Accurelw). The
membrane separated two aqueous phases, with the source
phase containing 0.3 M fructose and glucose buffered at pH
11.3 with Na2CO3 and the receiving phase buffered at pH
6.0 with sodium phosphate. The appearance of fructose and
glucose in the receiving phase was monitored by enzymatic
methods.
2.3. Molecular modelling
It has previously been reported that the o-phenylene linked
diboronic acid (11, Fig. 3) displays enhanced fructose
selectivity relative to 2 (Table 1, entry 5),8a and based on the
results of molecular modelling experiments, it was
suggested that this enhanced selectivity is due to the ability
of 11 to form a 1:1 macrocyclic diester with fructopyranose,
shown in Figure 4(A). The formation of such an ester should
minimize the hydrophilic surface area of fructose, thus
enhancing extraction into, and promoting smooth passage
through, the hydrophobic liquid membrane. Similar model-
ling experiments11 with the fructopyranose ester of 10,
however, showed that the spatial arrangement of the boronic
acid functionalities of 10 are inappropriate to allow it to
form a similarly stable cyclic 1:1 diester. Other possibilities
were sought, and after recognizing that in aqueous alkaline
solution, the major boronate ester formed between p-tolyl-
boronic acid and fructose is the trisdentate 2,3,6-b-D-
fructofuranose ester (12, Fig. 3),12 a 2:1 diester of 10 was
identified (Fig. 4(B)). These molecular modelling experi-
ments further suggested that this diester could form two
intramolecular hydrogen bonds with interatomic distances
The mono-saccharide transport promoted by a reference
monoboronic acid (2)10 is shown in Table 1 (entry 1),
compared with that induced by 1 (entry 2), the penta-
erythrityl boronic acids (entries 3 and 4) and a previously
reported o-phenylene linked diboronic acid (11, entry 5). In
analogy to the results obtained in BLM transport experi-
ments,6 1 showed poor SLM transport characteristics,
inducing lower sugar fluxes than the reference monoboronic
(2). The very low fructose/glucose selectivity induced by 1
is a curious result and suggests that the properties of this
compound retard fructose transport more than glucose
transport.
In contrast to the transport properties of 1, the highly
lipophilic, tetrahedral shaped carrier (7) proved to be an
excellent sugar transporter. No leaching of this carrier was
observed during the course of the transport experiments, and
the fructose flux produced by 7 was found to be significantly
higher than that of 1, despite having a substantially higher
molecular weight (849 g mol21 cf. 550 g mol21). The
fructose selectivity of 7 was also superior to that of 1,
being more like to that of 2. The choice of a tetrahedral
shape for lipophilic carriers thus appears to be validated.
However, the higher molecular weight of 7, which would be
expected to lower the diffusion constant of its sugar
boronates within the membrane, does in fact appear to
have a detrimental impact on sugars fluxes. The flux
of less than 1.2 A.13 This structure is an attractive possibility
˚
as the additional hydrogen bonding interactions should
further stabilize the fructofuranose esters, and association
between the fructose portions of the esters would reduce the
amount of hydrophilic surface exposed to the hydrophobic
medium, thus improving fructose extraction and diffusion
through the membrane. The ability of each molecule of 10 to
readily carry two fructose molecules across the membrane at
once, in the form of a frutofuranose diester, similar to that
Figure 3. Previously studied o-phenylene linked diboronic acid (11)8a and previously characterised p-tolylboronates of fructose (12 and 13).12

9078
S. P. Draffin et al. / Tetrahedron 59 (2003) 9075–9082
Figure 4. Fructose esters of diboronic acids identified through molecular modelling experiments. (A) Macrocyclic D-fructopyranose diester of 11.8a
(B) Trisdentate 2,3,6-b-D-fructofuranose diester of 10.
shown in Figure 4, may also help to account for the
improved fructose flux promoted by this diboronic acid.
eliminates interference by signals from the carriers and
ammonium ions, and a considerable amount of reference
data on the various boronate esters that can be formed with
fructose is now available.12,14
2.4. Solution structures of fructose boronates
To test the above predictions, experimental evidence for the
structures shown in Figure 4 was sought. Tetrahedral
boronate esters are notoriously difficult to characterise,
and attempts to identify the fructose esters of 10 and 11,
formed by extraction of fructose from aqueous alkaline
solution into organic solutions containing the boronic acid
and aliquat 336, with a range of mass spectrometry
techniques, including electron impact ionisation (EI),
atmospheric pressure chemical ionisation (APCI), electro-
spray (ES) and, in the case of 11, fast atom bombardment
(FAB), proved fruitless. Attempts to form crystals of these
esters suitable for X-ray crystal structure determination also
proved unsuccessful.‡ However, the aqueous solution
structures of several aryl boronates formed with 13C
enriched fructose have previously been identified by 13C
NMR spectroscopy through the use of chemical shift
The initial experiment involved the extraction of fructose,
uniformally labelled with 13C with an enrichment of 99%,
out of D2O adjusted to ,pD11.0 with Na2CO3, into CDCl3
containing Aliquat 336w and 2. These conditions closely
model the initial stages of the transport process facilitated
by a typical monoboronic acid, in which fructose is
sequestered from the source phase and drawn into the
membrane. Thus, the boronate esters formed in the
membrane, and in CDCl3 in the extraction experiment, are
likely to be very similar. A portion of the resulting 13C NMR
spectrum of the CDCl3 solution thus obtained is shown in
1
Figure 5(A), with the major peaks and JCC coupling
constants listed at entry 3 in Tables 2 and 3, respectively.
For comparison, data previously obtained with the trisden-
tate 2,3,6-b-D-fructofuranose p-tolylorthoboronate ester
(12) and the four possible diastereomeric forms of b-D-
fructopyranose p-tolylhydroxyboronate esters (13) in D2O12
are shown at entries 1 and 2 in these tables.
1
correlations and JCC coupling constants,12 and modifi-
cation of that method allowed the identification of the likely
structures of the major boronate esters extracted into the
membrane during fructose transport promoted by 2 and the
diboronic acids (10 and 11). This method is especially
suitable here as the use of 13C enriched fructose virtually
On first inspection of Figure 5(A) it is immediately clear
that only one major fructose ester is formed with 2 in
CDCl3, and by comparison of the NMR chemical shifts and
1JCC coupling constants of this ester (Tables 2 and 3) with
the data previously obtained for 12 and 13, it is apparent that
this ester is the trisdentate 2,3,6-b-D-fructofuranose ester.
In particular, signals at .d 110 ppm for C-2 and signals for
‡
Even if such experiments had been successful, the relevance to transport
experiments of structures identified in the gas phase or solid state is
somewhat questionable.

S. P. Draffin et al. / Tetrahedron 59 (2003) 9075–9082
9079
Table 3. 1JCC Coupling constants (in Hz) for fructose portion of tetrahedral
boronate esters
Entry
Boronate
J1,2
J2,3
J3,4
J4,5
J5,6
1
2
3
4
12a
54.1
51.3
53.8
53.5
37.4
38.3
36.6
37.1
40.3
44.3
38.4
38.4
36.9
34.1
36.2
36.8
36.3
38.9
36.4
37.0
13a,b
2·boronate
10·diboronate
a
Data obtained in a previous study,12 pD¼11–12 in D2O.
b
These coupling constants are approximate and are within ^0.5 Hz of
those found for the four diastereoisomers of 13 (where measured).
similar position to that observed for C-2 in 13, suggesting
that, although the trisdentate 2,3,6-b-D-fructofuranose ester
of 2 is dominant, some pyranose esters are still present in the
CDCl3 solution.
Analogous experiments were then performed with the
diboronic acids (10 and 11). In the case of the penta-
erythrityl diboronic acid (10), a very similar spectrum to
that obtained with 2 was produced, the major fructose
derivative observed again being the trisdentate 2,3,6-b-D-
fructofuranose ester (see correlations in Tables 2 and 3). In
fact, notwithstanding that the signal-to-noise in Figure 5(B)
is inferior to that in Figure 5(A), there is no evidence for
fructopyranose esters in this spectrum.
Remarkably, the spectrum obtained with the o-phenylene
linked diboronic acid (11) is distinctly different from the
previous two (Fig. 5(C)). Although there is evidence for the
presence of the trisdentate 2,3,6-b-D-fructofuranose ester,
especially in the region just above d 110 ppm, very strong
signals in the region d 100–103 ppm and in the range d 63–
76 ppm suggest the presence of significant quantities of
b-D-fructopyranose esters, thus supporting the previous
molecular modelling results.8a In addition, significant
amounts of polymeric material, both furanose and pyranose
in nature, appear to be present, as shown by the broad
signals at ,d 112 ppm, in the region d 70–77 ppm, as well
as those overlapping with many of the sharper peaks
throughout the spectrum. Similar results have recently been
reported for glucose adducts of ferrocene containing
diboronic acids.15
Figure 5. 13C NMR spectra of UL-13C-D-fructose extracted from D2O
(pD,11) into CDCl3 containing 2, 10 and 11 and aliquat 336.
(A) Extraction with 2 (75 MHz). (B) Extraction with 10 (125 MHz).
(C) Extraction with 11 (100 MHz).
3. Conclusions
A new class of tetrahedral shaped lipophilic boronic acids
based on a pentaerythritol core (7 and 10) have been
prepared and tested for their ability to transport fructose and
glucose through a lipophilic SLM. These compounds have
shown far superior transport properties to those of a
previously described cholesterol derivative (1), despite
C-3 to C-5 at .80 ppm are characteristic of such structures.
1
In addition, the JCC coupling constants for the fructose
ester of 2 are more in line with those of 12 than 13. The
weak, broad signal at d 101–103 ppm in Figure 5(A) is in a
Table 2. 13C Chemical shifts (ppm) for fructose portion of tetrahedral boronate esters
Entry
Boronate
C-1
C-2
C-3
C-4
82.1
73.3^1.0
81.0
81.1
C-5
87.6
71.7^0.3
85.7
85.8
C-6
67.9
66.7^0.4
65.5
65.4
1
2
3
4
12a
66.1
113.4
84.7
13a,b
69.4^0.2
64.8
64.9
103.8^0.8
111.4
111.4
76.4^0.5
83.1
83.2
2·boronate
10·diboronate
a
Data obtained in a previous study,12 pD¼11–12 in D2O.
b
Uncertainties indicate the range over which signals occur for the four diastereoisomers of 13.

9080
S. P. Draffin et al. / Tetrahedron 59 (2003) 9075–9082
having significantly higher molecular weight. The overall
tetrahedral shape of 7 and 10 appears to prevent unfavour-
able interfacial stacking interactions that are likely to occur
with more planar boronic acids such as 1.
4.2. Synthesis
4.2.1. 1,10-[[2-(Bromomethyl)-2-[[4-(1,1,3,3-tetramethyl-
butyl)phenoxy]methyl]-1,3-propanediyl]bis(oxy)]bis-
[4-(1,1,3,3-tetramethylbutyl)-benzene (4). Pentaerythrityl
tetrabromide (2.32 g, 6 mmol), p-t-octyl phenol (3.70 g,
18 mmol) potassium carbonate (4.14 g, 30 mmol) and DMF
(30 mL) were combined and stirred at 1008C for 24 h under
argon. The reaction mixture was allowed to cool, then
filtered. The solvent was removed in vacuo and the residue
taken up in dichloromethane and washed with 5% HCl
(20 mL), water (2£20 mL), brine (30 mL) and dried over
MgSO4. The solvent was then evaporated to give a yellow
oil (2.40 g). This oil was submitted to flash chroma-
tography (hexane 100%) to give pure monobromide (4)
(1.32 g, 28%) and dibromide (8) (1.02 g, 22%). Charac-
terisation data for 8 are shown below. 4, mp 98.58C. IR
(KBr) n: 1610 w, 1511 m, 1463 s, 1377 m, 1241 m, 1184 m,
The diboronic acid (10) shows unusually high fructose/
glucose transport selectivity, which appears to result from
the formation within the membrane of a trisdentate 2,3,6-b-
D-fructofuranose diester (Fig. 4(B)), which can be stabilized
by intramolecular hydrogen bonds. This contrasts markedly
with results obtained with a previously described o-phenyl-
ene linked diboronic acid (11), which appears to gain
advantage from the formation of a macrocyclic D-fructo-
pyranose diester.
In general, the minimisation of the hydrophilic surface area
of the transported sugar should improve flux by enhancing
both extraction into the membrane and diffusion through the
lipophilic medium. While the formation of macrocyclic
Dfructopyranose diesters with diboronic acids appears to be
beneficial in this regard, greater advantage can apparently
be obtained through the formation of adjacent trisdentate
2,3,6-b-D-fructofuranose diesters. Results described here
suggest that these esters gain stability by intramolecular
association within the membrane, which also leads to
reduced exposure of the hydrophilic surface of the fructose
portion of the esters. The implication is that the projection of
multiple boronic acids in a parallel fashion from a rigid,
hydrophobic scaffold should lead to improved fructose flux
and selectivity. Studies that test this design hypothesis have
been initiated and the preliminary results are very
encouraging.8b
1
cm21. H NMR: d 7.24 (d, J¼7.8 Hz, 6H), 6.82 (d, J¼
7.8 Hz, 6H), 4.23 (s, 6H), 3.90 (s, 2H), 1.69 (s, 6H), 1.33
(s, 18H), 0.72 (s, 27H) ppm. 13C NMR: d 156.3, 142.8,
127.1, 114.0, 67.3, 57.1, 44.5, 38.2, 34.8, 32.5, 32.1,
31.9 ppm. HRMS calcd for C47H71BrO3Na (MNaþ):
785.4484. Found: 785.4477. Anal. calcd for C47H71BrO3:
C, 73.89; H, 9.37; Br, 10.21. Found: C, 74.35; H, 9.37; Br,
10.46.
4.2.2. Benzoic acid, 4-(4,4,5,5-tetramethyl-1,3,2-dioxa-
borolan-2-yl)-, 3-[4-(1,1,3,3-tetramethylbutyl)phenoxy]-
2,2-bis[[4-(1,1,3,3-tetramethylbutyl)phenoxy]methyl]-
propyl ester (6). The monobromide (4), (0.20 g,
0.26 mmol, 5 (0.06 g, 0.03 mmol), cesium carbonate
(0.23 g, 0.78 mmol), potassium iodide (0.55 g, 3.30 mmol)
and dimethyl acetamide (30 mL) were combined and heated
at 1008C for 24 h under argon. Upon cooling the solvent was
removed in vacuo and the residue adsorbed onto a small
plug of silica. The plug was washed with hexane then the
product was desorbed by washing with dichloromethane.
Removal of the dichloromethane under vacuum gave an oil
which was submitted to flash chromatography (dichloro-
methane/hexane 8:2) to give white crystalline product
(0.21 g, 75%), mp 114–1168C. IR (Nujol) n: 1716 m,
4. Experimental
4.1. General
Melting points of solid compounds were measured on a
Reichert Hot stage melting point apparatus. All infrared
spectra were recorded on a Perkin–Elmer 1600 Series
Fourier Transform spectrophotometer, either as Nujol
(paraffin) mulls mounted on sodium chloride plates, or as
potassium bromide disks. All NMR spectra were recorded
of solutions of compounds dissolved in CDCl3 (unless stated
otherwise), referenced to tetramethylsilane (d 0.00 ppm) at
spectrometer frequencies of 300 MHz (1H NMR) or
75 MHz (13C NMR), with the exception of the 13C NMR
spectrum of the fructose ester of 10, which was recorded at
125 MHz, and of 11, which was recorded at 100 MHz. Mass
spectra were recorded on a Bruker–Bioapex 47e Fourier
Transform mass spectrometer. Unless otherwise stated,
compounds were dissolved in methanol and ionised using an
electrospray ionisation source and recorded in positive ion
mode. Microanalyses were performed by Campbell Micro-
analytical Laboratory, University of Otago, Dunedin, New
Zealand. All solvents were AR grade and used as supplied,
1609 w, 1510 m, 1242 m, 1122 m, 1099 m, 1025 m, cm21
.
1H NMR: d 7.95 (d, J¼7.8 Hz, 2H), 7.82 (d, J¼7.8 Hz, 2H),
7.24 (d, J¼7.8 Hz, 6H), 6.82 (d, J¼7.8 Hz, 6H), 4.75 (s,
2H), 4.31 (s, 6H), 1.69 (s, 6H), 1.34 (s, 12H), 1.32 (s, 18H)
0.71 (s, 27H) ppm. 13C NMR: d 166.3, 156.4, 142.7, 134.7,
132.3, 128.6, 127.1, 113.9, 84.3, 67.1, 64.5, 57.1, 44.5, 38.2,
32.5, 32.0, 31.9, 25.1 ppm. (Carbon attached to boron
not observed due to broadening) MS (APCI) m/z 931.4
(Mþ1þ, 100%).
4.2.3. Benzoic acid, 4-borono-, 1-[3-[4-(1,1,3,3-tetra-
methylbutyl)phenoxy]-2,2-bis[[4-(1,1,3,3-tetramethyl-
butyl)phenoxy]methyl]propyl] ester (7). The mono
pinacol ester (6) (0.10 g), hydrochloric acid (3 mL, 1 M)
and acetone (30 mL) were combined and stirred for 24 h.
The solvent was removed in vacuo and the residue taken up
in dichloromethane (30 mL). The organics were washed
with water (3£50 mL) and dried with MgSO4. Filtration and
removal of the solvent in vacuo gave glass like crystals
(0.06 g, 66%), mp 126–1288C. IR (Nujol) n: 3385 b, 3268 s,
1721 m, 1655 s, 1540 m, 1243 m, 1028 m, cm21. 1H NMR
˚
except dimethyl formamide, which was dried over 4 A
molecular sieves. Column chromatography was performed
according to the method reported by Still et al.16 using the
eluents stated, expressed as volume ratios. Compound 5 is
commercially available from Boron Molecular Ltd, Noble
Pk, Victoria, Australia (http://www.boronmolecular.com).

S. P. Draffin et al. / Tetrahedron 59 (2003) 9075–9082
9081
(d6-acetone/D2O, 9:1): d 8.02 (d, J¼7.8 Hz, 2H), 7.96 (d,
J¼7.8 Hz, 2H), 7.28 (d, J¼7.8 Hz, 6H), 6.92 (d, J¼7.8 Hz,
6H), 4.75 (s, 2H), 4.41 (s, 6H), 1.71 (s, 6H), 1.31 (s, 18H)
0.69 (s, 27H) ppm. 13C NMR: d 166.3, 156.4, 142.7, 133.7,
132.2, 128.9, 127.1, 113.9, 67.0, 64.5, 57.1, 44.5, 38.2, 32.5,
32.0, 31.9 ppm. (Carbon attached to boron not observed due
to broadening).
32.0, 31.9 ppm. (Carbon attached to boron not observed due
to broadening).
4.3. Transport experiments
The transport cell consisted of two identical water jacketed
cylindrical halves (T¼298 K), with a half cell volume of
34 mL and a cell membrane surface area of 12.6 cm2. Each
half cell was stirred by a magnetic stirrer at a rate of
250 rpm. The SLM consisted of a polypropylene support
(Accurelw type 1E flat sheet (thickness 0.1 mm pore size
0.1 mm)) obtained from Membrana GmbH, Wuppertal,
Germany, coated in a solution of carriers dissolved in
2-nitrophenyl octyl ether. Liquid membranes were prepared
in duplicate by dissolving 25 mmol carrier and 25 mmol
Aliquat 336w in a suitable solvent, (usually chloroform)
dividing the solution into two equal parts, then adding
0.25 g of nitrophenyl octyl ether to each. The solvent
was then removed in vacuo to leave oils. Two polymer
supports were then coated with these oils and kept under
vacuum (,1 Torr) for 24 h. The departure phase consisted
of a solution of 0.3 mol dm23 fructose, 0.3 mol dm23
glucose dissolved in a buffer solution of 0.5 mol dm23
Na2CO3 at pH 11.3. The sugar solutions were freshly
prepared before each transport experiment. The receiving
phase consisted of a 0.1 mol dm23 solution of sodium
phosphate at pH 6.0.
4.2.4. 1,10-[[2,2-Bis(bromomethyl)-1,3-propanediyl]bis-
(oxy)]bis[4-(1,1,3,3-tetramethylbutyl)-benzene (8). Penta-
erythrityl tetrabromide (2.00 g, 5.15 mmol) and p-t-octyl
phenol (2.13 g, 10.30 mmol) were treated with potassium
carbonate (5.38 g, 16.50 mmol) in DMF (30 mL) in the
same manner as that described for the preparation of
4. Flash chromatography was used to obtain pure
dibromide (8) (1.58 g, 48%) and monobromide (4)
(0.42 g 12%). 8, mp 109–1108C. IR (KBr) n: 1511
1
m, 1460 m, 1376 m cm21. H NMR: d 7.26 (d, J¼7.8 Hz,
4H), 6.82 (d, J¼7.8 Hz, 4H), 4.13 (s, 4H), 3.79 (s, 4H), 1.69
(s, 4H), 1.33 (s, 12H), 0.72 (s, 18H) ppm. 13C NMR: d
156.2, 143.3, 127.3, 114.2, 67.6, 57.2, 44.3, 38.2, 35.2, 32.7,
32.2, 32.0 ppm. MS (DCM/MeOH 1:2) m/z: 559.3
(M2Br, 3%), 637.4 (Mþ(79Br2)þ1, 3%). Anal. calcd for
C33H50Br2O2: C, 62.07; H, 7.89; Br, 25.03. Found: C, 61.98;
H, 8.01; Br, 24.99.
4.2.5. Benzoic acid, 4-(4,4,5,5-tetramethyl-1,3,2-dioxa-
borolan-2-yl)-, 2,2-bis[[4-(1,1,3,3-tetramethylbutyl)phen-
oxy]methyl]-1,3-propanediyl ester (9). The dibromide
(8) (1.00 g, 1.5 mmol) was treated with 5 (0.86 g,
3.47 mmol), cesium carbonate (2.26 g, 6.9 mmol) and
potassium iodide (2.62 g, 15.8 mmol) in dimethyl aceta-
mide (30 mL) in the same way as that described for
the preparation of 6. Flash chromatography (eluent 5%
ethyl acetate/hexane!100% ethyl acetate) was used to
obtain 9 as a white crystalline solid (0.40 g, 28%), mp 206–
2088C. IR (Nujol) n: 1718 m, 1243 m, 1114 m, 1018 m,
Aliquots were removed from the receiving phase at hourly
intervals and analysed in triplicate for glucose with a
coupled hexokinase/glucose-6-phosphate dehydrogenase
assay, and fructose with the addition of phosphoglucose
isomerase, by modification of a previously described
method.17 The volume of aliquots removed from the
receiving phase was adjusted so that the final absorbance
change was in the range 0.05–0.50 absorbance units. These
aliquots were added to 1 mL plastic disposable cuvettes, and
enough NaH2PO4 (0.1 mol dm23, pH 7.5) was added to give
a combined volume of 890 mL. ATP (5 mL of a
0.2 mol dm23 soln) and NADP (5 mL of a 0.1 mol dm23
soln) both dissolved in NaH2PO4 (0.07 mol dm23, pH 7.5)
were added to each cuvette, and 2.5 units of hexokinase/
glucose-6-phosphate dehydrogenase dissolved in 25 mL
1
710 m, ppm. H NMR: d 7.96 (d, J¼7.8 Hz, 4H), 7.82
(d, J¼7.8 Hz, 4H), 7.23 (d, J¼7.7 Hz, 4H), 6.82 (d, J¼
7.8 Hz, 4H), 4.73 (s, 4H), 4.28 (s, 4H), 1.68 (s, 4H), 1.35
(s, 24H), 1.33 (s, 12H), 0.72 (s, 18H) ppm. 13C NMR: d
166.1, 156.1, 142.7, 134.6, 131.9, 128.5, 127.0, 113.7,
84.1, 67.0, 64.1, 57.0, 43.9, 38.0, 32.4, 31.9, 31.7,
25.0 ppm. (Carbon attached to boron not observed due to
broadening) MS (APCI) m/z 973.3 (Mþ1þ, 33%), 995.2
(MþNaþ, 9%).
of NaH2PO4 (0.07 mol dm23
, pH 7.5 containing
0.004 mol dm23 MgCl2) was added. The absorbance change
at 340 nm could be used to determine [glucose]. Once
glucose determination was complete, 14 units of phospho-
glucose isomerase dissolved in 25 mL of NaH2PO4
(0.07 mol dm23, pH 7.5 containing 0.004 mol dm23
MgCl2) was added, and the change in absorbance at
340 nm could be used to determine [fructose]. Fluxes
were determined from slopes of plots of [Total NADPH Abs
(340 nm)] vs time over periods of at least 5 h, according to
Eq. (1), and the results averaged over two runs. All transport
experiments were performed at least in duplicate and all
assays were done in triplicate. Flux uncertainty ^10%.
4.2.6. Benzoic acid, 4-borono-, 1,10-[2,2-bis[[4-(1,1,3,3-
tetramethylbutyl)phenoxy]methyl]-1,3-propanediyl]
ester (10). The dipinacol ester (9) (0.20 g), 1 M hydro-
chloric acid (10 mL) and an acetone/water mix (4:1,
150 mL) were combined and stirred at room temperature
for 48 h. The solvent was removed in vacuo and a white
crystalline precipitate was collected. This was dissolved in
chloroform and washed with water (20 mL£3) and the
solvent removed to leave white crystals (0.12 g, 73%), mp
110–1118C. IR (Nujol) n: 3400 b, 3330 s, 1779 m, 1698 s,
1543 m, 1264 m, 1181 m, 1014 m, cm21 1H NMR
.
Fluxðmol m22s21Þ ¼
(d6-acetone/D2O, 9:1): d 7.97 (d, J¼7.8 Hz, 4H), 7.90 (d,
J¼7.8 Hz, 4H), 7.26 (d, J¼7.7 Hz, 4H), 6.90 (d, J¼7.8 Hz,
4H), 4.75 (s, 4H), 4.41 (s, 4H), 1.67 (s, 4H), 1.27 (s, 12H),
0.65 (s, 18H) ppm. 13C NMR: d 166.3, 156.2, 142.9, 134.8,
132.0, 128.6, 127.1, 113.9, 67.1, 64.2, 57.1, 44.0, 38.2, 32.5,
slopeðmin21Þ £ 0:034ðLÞ
6220ðL mol21 cm21Þ £ 1ðcmÞ £ 0:00126ðm2Þ £ 60ðs min21
Þ
ð1Þ

9082
S. P. Draffin et al. / Tetrahedron 59 (2003) 9075–9082
4.4. NMR Studies of 13C-fructose boronate esters
6. Karpa, M. J.; Duggan, P. J.; Griffin, G. J.; Freudigmann, S. J.
Tetrahedron 1997, 53, 3669–3678.
7. James, T. D.; Harada, T.; Shinkai, S. J. Chem. Soc., Chem.
Commun. 1993, 857–858.
Each boronic acid (10 mg) was briefly shaken in an
eppendorf tube with Aliquat 336w (1 equiv. for each
boronic acid), uniformly labelled 13C-fructose (99%
enrichment, 6 equiv.) and 0.5 mol dm23 Na2CO3 in D2O
(1.5 mL pD¼11.0). CDCl3 (0.7 mL) was then added and the
mixture shaken for 1 h. The organic layer was removed,
dried (MgSO4) and placed in an NMR tube.
8. (a) Gardiner, S. J.; Smith, B. D.; Duggan, P. J.; Karpa, M. J.;
Griffin, G. J. Tetrahedron 1999, 55, 2857–2864. (b) Altamore,
T. M.; Barrett, E. S.; Duggan, P. J.; Sherburn, M. S.; Szydzik,
M. L. Org. Lett. 2002, 4, 3489–3491. (c) Duggan, P. J.;
Szydzik, M. L. Aust. J. Chem. 2003, 56, 17–21.
9. A preliminary report of part of the work described here has
already appeared; Draffin, S. P.; Duggan, P. J.; Duggan, S. A.
M. Org. Lett. 2001, 3, 917–920.
Acknowledgements
10. (a) Paugam, M.-F.; Riggs, J. A.; Smith, B. D. J. Chem. Soc.,
Chem. Commun. 1996, 2539–2540. (b) Paugam, M.-F.; Bien,
J. T.; Smith, B. D.; Chrisstoffels, L. A. J.; de Jong, F.;
Reinhoudt, D. N. J. Am. Chem. Soc. 1996, 118, 9820–9825.
11. Molecular Mechanics simulations were performed with MSI’s
Discover Minimisation Module within the Insight II suite of
programs. Simulations were performed in the gas phase using
the ESFF force field and the ‘Structural’ analysis option.
Further details of the computational procedures used can be
found at http://www.accelrys.com/insight/discover.html.
12. Norrild, J. C.; Eggert, H. J. Chem. Soc., Perkin Trans. 2 1996,
2583–2588.
We thank Tim Altamore for preparing 3 and Dr Peter
Nichols for assistance with molecular modelling. We
gratefully acknowledge financial support from the
Australian Research Council and Monash University.
References
1. (a) Kuivila, H. G.; Keough, A. H.; Soboczenski, E. J. J. Org.
Chem. 1954, 19, 780–783. (b) Wolfrom, M. L.; Solms, J.
J. Org. Chem. 1956, 21, 815–816. (c) Sugihara, J. M.;
Bowman, C. M. J. Am. Chem. Soc. 1958, 80, 2443–2446.
(d) Lorand, J. P.; Edwards, J. O. J. Org. Chem. 1959, 24,
769–774.
13. Shinkai and co-workers have suggested that a similar type of
H-bonding might occur in difructose boronate diesters
appended to poly(L-lysine); Kobayashi, H.; Nakashima, K.;
Ohshima, E.; Hisaeda, Y.; Hamachi, I.; Shinkai, S. J. Chem.
Soc., Perkin Trans. 2 2000, 997–1002.
2. (a) Ferrier, R. J. Adv. Carbohydr. Chem. Biochem. 1978, 35,
31–80. (b) Duggan, P. J.; Tyndall, E. M. J. Chem. Soc., Perkin
Trans. 1 2002, 1325–1339, and references sited therein.
3. Shinbo, T.; Nishimura, K.; Yamaguchi, T.; Sugiura, M.
J. Chem. Soc., Chem. Commun. 1986, 349–351.
14. (a) Shull, B. K.; Spielvogel, D. E.; Head, G.; Gopalaswamy,
R.; Sankar, S.; Devito, K. J. Pharm. Sci. 2000, 89, 215–222.
(b) He, M.; Johnson, R. J.; Escobedo, J. O.; Beck, P. A.; Kim,
K. K.; St Luce, N. N.; Davis, C. J.; Lewis, P. T.; Fronczek,
F. R.; Melancon, B. J.; Mrse, A. A.; Treleaven, W. D.;
Strongin, R. M. J. Am. Chem. Soc. 2002, 124, 5000–5009.
15. Norrild, J. C.; Søtofte, I. J. Chem. Soc., Perkin Trans. 2 2002,
303–311.
4. (a) Riggs, J. A.; Hossler, K. A.; Smith, B. D.; Karpa, M. J.;
Griffin, G.; Duggan, P. J. Tetrahedron Lett. 1996, 37,
6303–6306. (b) Westmark, P. R.; Smith, B. D. J. Pharm.
Sci. 1996, 85, 266–269.
5. (a) Smith, B. D. Supramol. Chem. 1996, 7, 55–60. (b) Smith,
B. D.; Gardiner, S. J. Adv. Supramol. Chem. 1999, 5,
157–202, and references cited therein.
16. Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43,
2923–2925.
17. Kinksy, S. C. Methods Enzymol. 1974, 32, 501–504.