Molecular design of novel transition state analogues for molecular imprinting

Article abstract of DOI:10.1039/b102500g

The synthesis and characterisation of novel polymerisable heterocycles containing boron is described. The compounds were designed to serve as transition state analogues of the enantioselective borane reduction of prochiral ketones with oxazaborolidine cat

Full text of DOI:10.1039/b102500g

View Article Online / Journal Homepage / Table of Contents for this issue
Molecular design of novel transition state analogues for molecular
imprinting
Andrea Biffis*y and Gunter Wulff
Institut fur Organische Chemie und Makromolekulare Chemie, Heinrich-Heine-Universitat
¨ ¨
Dusseldorf, Universitatstr. 1, D-40225 Dusseldorf, Germany
¨
¨
¨
Received (in Strasbourg, France) 16th March 2001, Accepted 15th September 2001
First published as an Advance Article on the web 21st November 2001
The synthesis and characterisation of novel polymerisable heterocycles containing boron is described. The
compounds were designed to serve as transition state analogues of the enantioselective borane reduction of
prochiral ketones with oxazaborolidine catalysts (CBS reduction). Preliminary results show that the
heterocycles can be conveniently used as template monomers in the synthesis of molecularly imprinted
polymers.
Reproducing the features of natural enzymes in artificial sys-
tems represents one of the most challenging tasks for today’s
chemists. The successful development of such systems con-
tributes to the understanding of the principles and strategies of
enzyme catalysis, enabling the discovery of novel non-biolo-
gical catalysts as well as the targeted use of modern protein
engineering techniques for the preparation of tailor-made
biocatalysts with novel properties.
Since the publication of the first crystal structures of enzymes
more than 40 years ago, a continuously increasing number of
researchers have been working on this topic. Early attempts
making use of host–guest chemistry1 or functionalised linear
polymers2 have been paralleled by a variety of other approaches
based on supramolecular complexes,1 functionalised micelles
and vescicles,1b cyclodextrins,1,3 artificial polypeptides,1c,4 cat-
alytic antibodies,5 and molecularly imprinted materials.6
Enzymes are very efficient catalysts that usually work with
high activity and selectivity. High chemo-, regio- and especially
stereoelectivity is induced by specific interactions (steric effects,
hydrophobic and electrostatic interactions) between the sub-
strates and the active site of the enzyme. Co-ordination to the
active site brings the reagents into particular conformations
and relative positions, which allow for just one of the possible
reaction patterns.
After removal of the template molecules, a solid is left pos-
sessing cavities that are complementary in size, shape and
functionalisation to the template being employed; such cavities
are able to recognise and selectively rebind the template
molecule.
In order to prepare enzyme-like polymers by molecular
imprinting, molecules that mimick the transition state of a
given reaction are used as templates; furthermore, the system
must be designed in such a way that a reactive moiety or a
convenient precursor thereof is placed in the cavity in the
course of the imprinting procedure (Scheme 1). The resulting
functionalised cavities are then potentially able to accelerate
the rate of a reaction taking place inside them by stabilising the
transition state, and=or to impart higher selectivity by
favouring the formation of one of the possible transition states.
Molecular imprinting has already proven useful in the
preparation of polymer-bound reagents with unusual regio-
and stereoselectivities,8 as well as in the development of
catalysts showing a certain degree of substrate selectivity.1c,9
The application of a similar strategy to the catalysis of
abiotic reactions of synthetic interest appears very intriguing.
To achieve this, a catalytically active moiety should be posi-
tioned into artificial clefts that are able, through their size,
shape and three-dimensional arrangement of functional
groups, to direct the reaction towards one of the possible
products. A rough but illustrative application of this principle
can be seen in the well-known regioselective isomerisation and
disproportionation of hydrocarbons within zeolite cavities.7
More difficult, however, appears to be the development of
general strategies for the preparation of stereoselective cata-
lysts of this kind.
We, together with other research groups, have decided to
tackle this problem by making use of molecular imprinting.6
This technique involves the preparation of highly crosslinked
polymers in the presence of molecules acting as templates.
Scheme 1 Preparation of enzyme-like reactive polymers by mole-
cular imprinting. (i) Copolymerisation of a transition state analogue
(TSA) containing a reactive group (R) with an excess of crosslinker; (ii)
removal of the TSA leaving R exactly placed in the cavity.
y Permanent address: Dipartimento di Chimica Inorganica, Metallor-
ganica ed Analitica, Universita di Padova, Via Marzolo 1, I-35131
Padova, Italy; E-mail andrea.biffis@unipd.it; Fax þ 39 049 827 5223.
DOI: 10.1039/b102500g
New J. Chem., 2001, 25, 1537–1542
1537
This journal is # The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2001

View Article Online
Very recently, initial attempts to improve the performance of
enantioselective transition metal catalysts by molecular
imprinting have also been reported.10 Higher activities and
enantioselectivities in comparison with non-imprinted polymer
supports were reached with these catalysts, but their perfor-
mance was never found to be superior to that of the corre-
sponding metal complexes in homogeneous solution.
As a test reaction we have chosen the enantioselective borane
reduction of prochiral ketones with a 1,3,2-oxazaborolidine
as the catalyst (CBS reduction shown in Scheme 2).11,12
Results and discussion
The first problem to be addressed was the choice of a transition
state analogue of the CBS reduction suitable for molecular
imprinting. The target molecule should be sterically and elec-
tronically similar to the transition state, and be polymerisable.
Furthermore, after polymerisation it should be easily removed
from the imprinted polymer, leaving functional groups in the
cavities that allow the exact placement of the oxazaborolidine
catalyst. These conditions can be fulfilled by using molecules of
the type shown in Scheme 4. In this model compound for the
exo conformation of the transition state, the bridging borane is
substituted by two methylene groups, the coordinated ketone is
simulated by an alkoxide functionality, and the B–N covalent
bond is replaced by a co-ordinative intramolecular interaction.
Scheme 2 The CBS reduction.
This reaction has aroused considerable interest in the last ten
years due to its versatility and high enantioselectivity.13
Although the detailed kinetics of this complex reaction are still
not completely understood, there is an experimentally well-
founded hypothesis about the reaction mechanism.14 In the first
step, the reagents co-ordinate to the catalyst forming a ternary
complex (Scheme 3). The borane interacts with the nitrogen
atom and the ketone with the boron atom; both reagents bind
anti to the oxazaborolidine substituent R00 at the asymmetrical
carbon atom, which shields one face of the ring. The co-
ordination of the ketone with the larger substituent R2 anti to
the oxazaborolidine ring is favoured for steric reasons. How-
ever, the complex can still assume two different conformations,
namely exo and endo.15
The key step, that is, the hydride transfer from borane to
ketone, takes place intramolecularly in the complex through a
cyclic six-membered transition state. Since exo and endo
complexes afford products with opposite configuration, the
stereochemical output of the reaction is critically dependent on
the energetic differences between the two conformations.
According to ab initio calculations, the endo conformation in
the ternary complex is slightly more stable;14a,15 in spite of this,
the product from the exo conformation usually predominates,
since the activation energy for the hydride transfer in the exo
conformation is lower.15
Scheme 4 Transition state (left) and the proposed transition state
analogue (right) of the CBS reduction.
The molecule can be made polymerisable through the choice
of an appropriate substituent R0. After polymerisation, the
dialcoholamine part of the molecule can be split off by simple
hydrolysis, leaving a polymer-bound boronic acid group that
can be easily converted into an oxazaborolidine catalyst by
condensation with an appropriate amino alcohol. Interest-
ingly, chiral as well as achiral oxazaborolidines, depending on
the amino alcohol employed, can be bound to the polymer
with this approach, making it possible on the one hand to
study the cavity effect on the intrinsic enantioselectivity of the
reaction and on the other to evaluate the enantioselective
induction of the cavity alone.
Based on the structure shown in Scheme 4, we prepared the
transition state analogues 1–4. These molecules are models for
the CBS reduction of acetophenone or related compounds, a
standard substrate class for this reaction, with oxazaboro-
lidines derived from the commercially available chiral amino
alcohols norephedrine and diphenylalaninol.
The selective stabilisation of one particular complex con-
formation should in principle have a profound influence on the
enantioselectivity of the reaction. This can actually be achieved
by molecular imprinting. Using the synthetic strategy outlined
above, it is possible to prepare imprinted polymers that possess
cavities complementary in size and shape to a predetermined
transition state analogue. Following removal of the template
molecule, oxazaborolidines can be placed within these cavities,
resembling one of the possible transition states. By testing
these catalysts in the CBS reduction, the cavity effect on the
stereochemical outcome of the reaction can be investigated. In
this paper we wish to report on the synthetic methodology for
the preparation of these novel materials.
It should be noted that compounds 1, 3 and 4 represent true
transition states of the CBS reduction, whereas 2 is a model of
a less-favoured transition state (diastereomeric to 1), in which
the ketone co-ordinates with the larger substituent syn to the
oxazaborolidine. Imprinting with this template affords cavities
with the ‘‘wrong’’ shape, whose behaviour in the CBS reduc-
tion should disclose additional information on the cavity
effect. In the case of compound 4 the methyl group in the
ketone-mimicking part of the molecule was substituted by a
hydrogen atom to maximise the difference in the steric demand
Scheme 3 Representation of the two possible conformations of the
intermediate complex in the CBS reduction
1538
New J. Chem., 2001, 25, 1537–1542

View Article Online
of the two ketone substituents. As a conveniently poly-
merisable substituent at the boron atom the p-vinylphenyl
group was employed.
The synthesis of compounds 1 and 2 (Scheme 5) was
accomplished through a Michael-type addition of phenyl vinyl
interaction and also to quantify its strength, that is, the degree
of quaternisation at the boron atom.17 Tricoordinate, cyclic
boronic acid diesters show chemical shifts in the range 26–33
ppm.17c Compounds 1 and 2 exhibit instead considerably
smaller values (6.15 and 6.63 ppm, respectively). These figures
are in good agreement with the literature value (6.62 ppm) for
the complex between N-(2-hydroxyethyl)-3-aminopropanol
and phenylboronic acid, a non-substituted analogue of the
above listed compounds with a fully quaternised boron
atom.17a
ketone to diphenylalaninol, affording 5, followed by
a
Grignard reaction with methylmagnesium iodide, yielding the
two diastereomers 6 and 7 in comparable amounts. The dia-
stereomers can be conveniently separated by column chro-
matography. Condensation of 6 or 7 with tris( p-vinylphenyl)-
boroxine 10 yields the desired products. Use of methyl vinyl
ketone and phenylmagnesium bromide results in an excess of 6
(diastereomeric ratio 7 : 3) but also in lower reaction yields.
To unequivocally assign the structures 1 and 2 to the prepared
compounds, the configuration at the newly formed chiral cen-
tres has to be determined. Furthermore, experimental evidence
for the intramolecular B–N interaction has to be obtained.
The isolation of a definite condensation product between 6
or 7 and the anhydride 10 is already an indication of the
existence of this interaction; simple nine-membered cyclic
boronic acid esters are in fact usually unstable and oligomeric
products are preferentially formed under the employed reac-
tion conditions.16a The fragmentation pattern of the com-
pounds in mass spectrometry is also in accordance with the
postulated structures. The B–N interaction greatly stabilises
the radical ion, which results from the splitting off of the
p-vinylphenyl substituent; the resulting signal appears as the
strongest fragment peak.16 Further evidence is provided by
11B-NMR. By considering the chemical shift of the boron
signal, it is possible to demonstrate the existence of the B–N
Having proved the existence of the B–N interaction, the
three-dimensional structures of the molecules 1 and 2 were
carefully investigated by molecular modelling. The B–N bond
was treated as a covalent bond with full charge separation. The
energy-minimised structures of the compounds are reported in
Fig. 1. The six-membered ring orients itself in both cases exo to
the oxazaborolidine ring, in accordance with the preferred
transition state of the CBS reduction. It takes, however, a
twisted boat conformation instead of the predicted chair-like
conformation. In any case, it must be remarked that the slight
preference of the reaction transition state for a chair-like
conformation has been theoretically proved only in a limited
number of cases,14a,15 and that the change from chair to
twisted boat does not significantly vary the spatial orientation
of the ring substituents, which is the most important factor for
the molecular imprinting procedure.
The assignment of the structures 1 and 2 to the prepared
compounds was accomplished through NOE investigations.
Irradiation of the 9-methyl group signal in the 1H-NMR
spectra of the two compounds yields, as expected in only one
case, an effect on the ortho-protons of the p-vinylphenyl ring.
The substitution pattern at this ring shifts the signal of
the ortho-protons to lower fields and makes them easily
Fig. 1 Modelled structures of the transition state analogues 1 (above)
and 2 (below).
Scheme 5 Synthesis of the TSAs 1–4.
New J. Chem., 2001, 25, 1537–1542
1539

View Article Online
Buchi 510 melting point apparatus. Optical rotations were
distinguishable from the other phenyl protons. The NOE effect
amounts to 4.8% and is reciprocal. The corresponding com-
pound was therefore assigned the structure 1.
measured on a Perkin Elmer 241 MC polarimeter. NMR
spectra were recorded on Varian VXR-300 (1H, 13C) and
Brucker DRX 500 (11B) spectrometers; TMS and BF3 Á Et2O
were used as standards. IR spectra were recorded on a Perkin
Elmer 1420 spectrophotometer. Mass spectra were obtained at
the Institute of Pharmaceutical Biology, University of Dus-
seldorf, using a Varian MAT CH7A mass spectrometer. Ele-
mental analyses were performed at the Institute of
Pharmaceutical Chemistry, University of Dusseldorf. Silica gel
60 (Merck) was used as adsorbent for column chromato-
graphy. Diphenylalaninol,18p-divinylbenzene,19 phenyl vinyl
ketone,20 and tris( p-vinylphenyl)boroxine 1021 were prepared
according to literature procedures.
Compound 3 was prepared in a similar reaction sequence
starting from norephedrine. Methyl vinyl ketone and phenyl-
magnesium bromide were employed. One of the two diaster-
eomeric dialcoholamine products (structure 8) was isolated
from the reaction mixture by simple crystallisation. NOE
analysis of the condensation product of 8 with 10 indicated the
same configuration for the new chiral carbon atom as in 1.
The dialcoholamine 9, the intermediate product for the
synthesis of compound 4, was synthetised by diastereoselective
reduction of 5 with borane–THF. To maximise the asymmetric
induction through the amino alcohol moiety, two equivalents
of the borane reagent were added to the substrate solution at
low temperature. In this way, the first borane equivalent builds
an oxazaborolidine with the amino alcohol moiety, which
subsequently catalyses the diastereoselective reaction of the
keto group with the second borane equivalent through an
intramolecular or intermolecular process. A diastereomeric
ratio of 8 : 2 (1H-NMR) was recorded. The configuration of
the diastereomer produced in excess was investigated by NOE
after condensation with 10. In this case, the signal of the 9-
hydrogen atom was irradiated, yielding once more a significant
effect on the ortho-protons of the p-vinylphenyl ring (7%).
The structural similarity of the synthetised model com-
pounds with the transition state of the CBS reduction is
apparent. Their application to copolymerisation, as well as
their ease of conversion to oxazaborolidines after poly-
merization, were preliminarily investigated by the synthesis of
a test polymer. Monomer 3 was copolymerised with an excess
of p-divinylbenzene as the crosslinker and THF as the porogen
by radical batch polymerisation. The imprinted polymer was
obtained as a stiff rod, which was subsequently ground to a
particle size between 125 and 45 mm, packed in glass columns,
and treated with an appropriate medium to wash out the
dialcoholamine part of the template molecule. Best results
were obtained by preswelling the polymer with pure THF and
then washing it with 3 : 1 THF–0.01 M HCl. Under these
conditions, the dissociation yield turned out to be satisfactory
(86%). The polymer-bound boronic acid groups could then be
converted to polymer-bound oxazaborolidines by condensa-
tion with norephedrine. Due to the heterogeneous nature of
the polymeric reagent, the reaction is rather slow, but it affords
the polymer-bound heterocycle in quantitative yield. A first
evaluation of the catalytic performance of the resulting
oxazaborolidine-containing, molecularly imprinted catalyst in
the reduction of acetophenone with borane–dimethylsulfide
adduct (THF, RT, 0.3 equiv. catalyst) yielded the alcohol
product in quantitative yield and 77% ee, very close to the
value for the corresponding free oxazaborolidine [(4S,5R)-2,5-
diphenyl-4-methyl-1,3,2-oxazaborolidine] under identical con-
ditions (82% ee).
Synthesis of the transition state analogues
(S)-2-(3-Oxo-3-phenylpropyl)amino-1,1-diphenylpropanol 5. To
a solution of vinyl phenyl ketone (7.80 g, 59 mmol) in diethyl
ether (100 ml) was added a solution of diphenylalaninol
(13.51 g, 59 mmol) in diethyl ether (100 ml), and the reaction
mixture was stirred overnight. The product slowly separated
as a white solid. The precipitation was completed by
concentrating the reaction mixture to 50 ml. The product was
filtered off and recrystallised from 1 : 1 CHCl3–n-hexane. Yield
18.7 g (88%). M.p.: 103 ꢀC. [a]20 (c ¼ 0.5 in CHCl3): 38.1ꢀ.
D
1H-NMR (CDCl3 , ppm): 1.00 (d, J¼ 6.35 Hz, 3H), 1.46 (br s,
1H), 2.81 (m, 1H), 3.01 (m, 3H), 3.80 (q, J ¼ 6.35 Hz, 1H), 4.46
(br s, 1H), 7.11–7.59 (m, 13H), 7.87 (m, 2H).13C-NMR (CDCl3 ,
ppm): 14.81 (CH3), 38.70, 42.14 (CH2), 58.51 (CH), 78.75 (C),
125.81–128.57, 133.17 (CH), 136.70, 144.80, 146.46, 199.20 (C).
IR (KBr, cm 1): 3360 (OH), 3280 (NH), 3070, 3040, 2960,
2850, 2810 (CH), 1670 (C¼O), 1480 (C¼C), 1440, 1370 (CH).
MS (DCI-NH3): m=z 360 (MHþ ). Anal. calc. for C24H25NO2 :
C 80.19, H 7.01, N 3.90; found C 80.18, H 7.16, N 3.63%.
(2S)-2-[(3S)-3-Hydroxy-3-phenylbutyl]amino-1,1-diphenyl-
propanol 6 and (2S)-2-[(3R)-3-Hydroxy-3-phenylbutyl]amino-
1,1-diphenylpropanol 7. A solution of methylmagnesium
iodide was prepared by standard methods from magnesium
turnings (7.44 g, 0.31 mol), methyl iodide (43.74 g, 0.28 mol)
and dry diethyl ether (400 ml). A solution of 5 (33.5 g, 0.093
mol) in diethyl ether (200 ml) was added dropwise at room
temperature. The reaction mixture was stirred for 3 h and
hydrolysed with saturated NH4Cl and water. Standard
workup yielded a crude oil that was purified by column
chromatography over silica (eluent: n–hexane–ethyl acetate–
triethylamine 20 : 10 : 1).
The fractions with 6 (Rf ¼ 0.35) were evaporated to dryness
and the residue taken up in hot n-hexane. The product forms
colourless crystals upon slow cooling to room temperature.
The feasibility of the proposed synthetic strategy for the
preparation of molecularly imprinted catalysts for the CBS
reduction has been demonstrated. The performance of such
catalysts in enantioselective reductions upon variation of
parameters such as the nature of the template molecule, the
nature of the co-monomers and porogen, the crosslinking
degree and the polymerisation conditions are currently under
investigation.
M.p.: 105 ꢀC. [a]20 (c ¼ 0.5 in CHCl3):
67.5 (589),
70.1
(578),
79.3 (546),
147.1 (435),
255.6ꢀ(365 nm). 1H-
NMR (CDCl3 , ppm): 0.95 (d, J ¼ 6.32 Hz, 3H), 1.38 (s, 3H)
1.72 (m, 1H), 1.93 (m, 1H), 2.26 (m, 1H), 2.85 (m, 1H), 3.49 (q,
J ¼ 6.32 Hz, 1H), 3.50 (br s, 1H), 7.12–7.49 (m, 15H).13C-
NMR (CDCl3 , ppm): 13.98 (CH3), 31.54 (CH3), 42.13, 44.22
(CH2), 59.06 (CH), 75.14, 80.07 (C), 125.00–128.41 (CH),
145.28, 145.80, 148.12 (C). IR (KBr, cm 1): 3460, 3290 (OH),
3265 (NH), 3070, 3040, 3010, 2960, 2830 (CH), 1490, 1470
(C¼C), 1440, 1375 (CH). MS (DCI-NH3): m=z 376 (MHþ ).
Anal. calc. for C25H29NO2 : C 79.96, H 7.78, N 3.73; found C
79.71, H 7.81, N 3.77%.
Experimental
The fractions with 7 (Rf ¼ 0.45) were concentrated and the
residue chromatographed again with the same eluent. The
fractions were evaporated to dryness and the residue treated
with hot n-hexane. As for its diastereomer, 7 crystallised upon
cooling to room temperature. M.p.: 90 ꢀC. [a]20 (c ¼ 0.5 in
General
Molecular modelling calculations were performed with the
program Cerius2, (Molecular Simulations Inc.) using the force
field Dreiding 2.2.1. Melting points were determined on a
1540
New J. Chem., 2001, 25, 1537–1542

View Article Online
CHCl3): þ 8.8 (589), þ 9.0 (578), þ 9.8 (546), þ 12.8 (435),
temperature. The solid was washed with n-hexane and dried.
1
þ 11.4ꢀ (365 nm). H-NMR (CDCl3 , ppm): 0.82 (d, J ¼ 6.32
Yield: 1.9 g (53%). M.p.: 76–77 ꢀC. [a]20 (c ¼ 0.5 in CHCl3):
Hz, 3H), 1.41 (s, 3H) 1.78 (m, 1H), 1.94 (m, 1H), 2.53 (m, 2H),
3.22 (br s, 1H), 3.52 (q, J ¼ 6.32 Hz, 1H), 7.11–7.53 (m,
15H).13C-NMR (CDCl3 , ppm): 14.18, 31.31 (CH3), 42.23,
44.02 (CH2), 59.22 (CH), 75.34, 79.73 (C), 124.96–128.58
(CH), 144.53, 145.53, 148.02 (C). IR (KBr, cm 1): 3500, 3310
(OH), 3260 (NH), 3070, 3040, 3010, 2950, 2910, 2830 (CH),
1492 (C¼C), 1445, 1370 (CH). MS (DCI-NH3): m=z 376
(MHþ ). Anal. calc. for C25H29NO2 : C 79.96, H 7.78, N 3.73;
found C 79.63, H 7.79, N 3.78%.
16.2 (589), 16.7 (578), 19.8 (546), 38.1 (435), 70.8ꢀ
1
(365 nm). H-NMR (CDCl3 , ppm): 0.98 (d, J ¼ 6.32 Hz, 3H),
1.74 (m, 2H), 2.57 (m, 1H), 2.83 (m, 1H), 3.69 (q, J ¼ 6.32 Hz,
1H), 3.87 (br s, 1H), 4.66 (m, 1H), 7.13–7.56 (m, 15H). 13C-
NMR (CDCl3 , ppm): 14.50 (CH3), 38.58, 44.96 (CH2), 59.02
(CH), 73.85, 79.46 (C), 125.63–128.40, 144.54 (CH), 144.82,
146.02 (C). IR (KBr, cm 1): 3414 (OH), 3060, 3020, 2960,
2935, 2810 (CH), 1490 (C¼C), 1450, 1380 (CH). MS (DCI-
NH3): m=z 362 (MHþ ). Anal. calc. for C24H27NO2 : C 79.74,
H 7.53, N 3.87; found C 79.63, H 7.24, N 3.81%.
(1R,2S)-N-(3-Oxobutyl)norephedrine. To
a
solution of
(5S,9S)-Perhydro-5,9-dimethyl-4,4,9-triphenyl-2-(p-vinyl-
phenyl)-1,3,6,2-dioxazaboronine 1. A solution of 6 (5.00 g, 13.3
mmol) and 10 (1.73 g, 4.4 mmol) in benzene (200 ml) was
heated in a round-bottomed flask fitted with a Dean–Stark
trap and a condenser. The solution was refluxed for 4 h and
then cooled to room temperature. The solvent was
evaporated and the residue was purified by column
norephedrine (9.6 g, 63 mmol) in CHCl3 (100 ml) cooled to
0 ꢀC was added dropwise a solution of freshly distilled
methyl vinyl ketone (4.45 g, 63 mmol) in CHCl3 (50 ml). The
solution was stirred overnight and allowed to reach room
temperature. The solvent was evaporated and the resulting
solid was used without further purification. The yield is
practically quantitative. M.p. 72 ꢀC. [a]20
(c ¼ 0.5 in
D
3.3ꢀ. 1H-NMR (CDCl3 , ppm): 0.81 (d, J ¼ 6.53
chromatography (eluent: n-hexane–ethyl acetate
2 : 1;
CHCl3):
Rf ¼ 0.55). The product (4.41 g, 68%) was obtained as a
Hz, 3H), 2.15 (s, 3H), 2.63 (t, J ¼ 6.21 Hz, 2H), 2.88 (m,
2H), 2.97 (dq, J ¼ 6.53 Hz and 3.78 Hz, 1H), 4.74 (d,
J ¼ 3.78 Hz, 1H), 7.30 (m, 5H).13C-NMR (CDCl3 , ppm):
14.50, 30.20 (CH3), 41.52, 43.87 (CH2), 58.48, 73.05, 126.03,
126.99, 128.05 (CH), 141.46, 208.34 (C). IR (KBr, cm 1):
3280 (NH), 3110 (OH), 3060, 3040, 3010, 2970, 2890 (CH),
1705 (C¼O), 1485 (C¼C), 1425, 1365 (CH). MS (DCI-NH3):
m=z 222 (MHþ ). Anal. calc. for C13H19NO2 : C 70.66, H
8.65, N 6.33; found C 70.82, H 8.85, N 6.11%.
1
colorless foam. [a]20 (c ¼ 0.5 in CHCl3): 112.2ꢀ. H-NMR
D
(CDCl3 , ppm): 1.04 (d, J ¼ 6.72 Hz, 3H), 1.23 (s, 3H), 1.93
(m, 2H), 3.35 (m, 1H), 3.98 (m, 1H), 4.33 (dq, J ¼ 6.72 Hz
and 10.22 Hz, 1H), 5.21 (dd, J ¼ 10.87 Hz and 1.04 Hz, 1H),
5.78 (dd, J ¼ 17.62 Hz and 1.04 Hz, 1H), 6.77 (dd, J ¼ 10.87
Hz and 17.62 Hz, 1H), 7.08–7.52 (m, 17H), 7.60 (d, J ¼ 7.93
Hz, 2H).13C-NMR (CDCl3 , ppm): 16.44, 30.33 (CH3), 34.43,
42.86 (CH2), 62.95 (CH), 73.33, 85.07 (C), 112.44 (CH2),
124.91–128.09, 131.32 (CH), 135.51 (C), 137.40 (CH), 144.39,
147.01, 150.09 (C). 11B-NMR (CDCl3 , ppm): 6.15. IR (KBr,
cm 1): 3040, 3010, 2950 (CH), 1625 (vinyl-C¼C), 1595, 1490
(C¼C), 1440, 1390, 1365 (CH). MS (DCI-NH3): m=z 505
(1R,2S)-2-[(3S)-3-Hydroxy-3-phenylbutyl]amino-1-phenyl-
propanol 8. A solution of phenylmagnesium bromide was
prepared by standard methods from magnesium turnings
(5.36 g, 0.22 mol), bromobenzene (31.49 g, 0.20 mol) and dry
(Mþ þ NH3), 488 (MHþ ), 384 (MHþ
vinylphenyl). Anal.
calc. for C33H34BNO2 : C 81.31, H 7.03, N 2.87; found C
80.93, H 6.95, N 2.78%.
diethyl ether (200 ml).
A
solution of 1(R,2S)-N-(3-
oxobutyl)norephedrine (14.8 g, 0.067 mol) in diethyl ether
(200 ml) was added dropwise at room temperature. The
reaction mixture was stirred for 3 h and hydrolysed with
saturated NH4Cl and water. After standard workup the crude
product was crystallised twice from 1 : 1 CHCl3–n-hexane.
Yield: 3.84 g (19%). M.p.: 140 ꢀC. [a]20 (c ¼ 0.5 in CHCl3):
225.7ꢀ (365 nm). 1H-NMR (CDCl3 , ppm): 0.85 (d,
J ¼ 6.53 Hz, 3H), 1.48 (s, 3H), 1.93 (m, 2H), 2.34 (m, 1H),
2.66 (dq, J ¼ 6.53 Hz and 3.84 Hz, 1H), 2.91 (m, 1H), 4.66
(d, J ¼ 3.84 Hz, 1H), 7.17–7.38 (m, 10H).13C-NMR CDCl3 ,
ppm): 13.66, 31.52 (CH3), 41.50, 43.90 (CH2), 59.00, 75.20
(CH), 75.68 (C), 124.91–128.21 (CH), 142.03, 148.36 (C). IR
(KBr, cm 1): 3310 (OH), 3250 (NH), 3040, 2960, 2900, 2840
(CH), 1490, 1470 (C¼C), 1440, 1365 (CH). MS (DCI-NH3):
m=z 300 (MHþ ). Anal. calc. for C19H25NO2 : C 76.22, H
8.42, N 4.68; found C 76.42, H 8.45, N 4.95%.
General procedure for the synthesis of the template monomers
2–4. A solution of one of the dialcoholamines 7–9 (10–20
mmol) and 10 (0.33 equiv.) in benzene (200 ml) was heated
in a round-bottomed flask fitted with a Dean–Stark trap and
a condenser. The solution was refluxed for 1 h. During this
time, the Dean–Stark trap was periodically emptied so that
about 150 ml of the solvent were distilled off. The product
(about 80% yield) separated from the solution as a white
amorphous solid upon cooling.
66.9 (589),
69.7 (578),
79.2 (546),
138.5 (435),
(5S,9R)-Perhydro-5,9-dimethyl-4,4,9-triphenyl-2-(p-vinyl-
phenyl)-1,3,6,2-dioxazaboronine 2. M.p.: 205 ꢀC. [a]20D (c ¼ 0.5
in CHCl3): 38.0ꢀ. 1H-NMR (CDCl3 , ppm): 0.95 (s, 3H) 0.97
(d, J ¼ 6.92 Hz, 3H), 2.07 (m, 2H), 3.08 (m, 1H), 3.27 (m, 1H),
3.53 (m, 1H), 4.33 (dq, J ¼ 6.92 Hz and 3.70 Hz, 1H), 5.11 (dd,
J ¼ 10.86 Hz and 1.16 Hz, 1H), 5.65 (dd, J ¼ 17.66 Hz and 1.16
Hz, 1H), 6.53 (dd, J ¼ 10.86 Hz and 17.66 Hz, 1H), 7.02–7.53
(m, 17H), 7.82 (d, J ¼ 7.12 Hz, 2H).13C-NMR (CDCl3 , ppm):
18.60, 32.77 (CH3), 34.23, 44.61 (CH2), 63.75 (CH), 72.25,
85.51 (C), 111.95 (CH2), 124.90–128.31, 131.96 (CH), 134.91
(C), 137.50 (CH), 145.53, 147.27, 149.72 (C). 11B-NMR
(CDCl3 , ppm): 6.63. IR (KBr, cm–1): 3280 (NH), 3050, 3020,
2970 (CH), 1625 (vinyl-C¼C), 1595, 1490 (C¼C), 1440,
1390, 1365 (CH). MS (DCI-NH3): m=z 505 (Mþ þ NH3), 488
(MHþ ), 384 (MHþ vinylphenyl). Anal. calc. for
C33H34BNO2 : C 81.31, H 7.03, N 2.87; found C 80.98, H
7.14, N 2.78%.
(2S)-2-[(3S)-3-Hydroxy-3-phenylpropyl]amino-1,1-diphenyl-
propanol 9. A solution of 5 (3.6 g, 10 mmol) in 100 ml dry THF
was cooled to
60 ꢀC under an inert atmosphere. A 1 M
solution of borane in THF (20 ml) was added with stirring
during 30 min with a syringe pump. After the addition, the
solution was stirred overnight and allowed to slowly reach
room temperature. The solution was hydrolysed by cautious
dropwise addition of methanol and evaporated to dryness.
The residue was taken up in diethyl ether (100 ml) and washed
two times with 0.2 M NaOH (50 ml) and two times with water
(50 ml). The solution was dried over Na2SO4 and evapor-
ated to dryness. The residue was purified by column
chromatography (eluent: n-hexane–ethyl acetate–triethylamine
20 : 10 : 1; Rf ¼ 0.23). The fractions were evaporated to dryness
and the oily residue crystallised upon standing at room
(4R,5S,9S)-Perhydro-5,9-dimethyl-4,9-diphenyl-2-(p-vinyl-
phenyl)-1,3,6,2-dioxazaboronine 3. M.p.: 202 ꢀC. [a]20D (c ¼ 0.5
in CHCl3): 69.1ꢀ. 1H-NMR (CDCl3 , ppm): 0.76 (d, J ¼ 6.92
New J. Chem., 2001, 25, 1537–1542
1541

View Article Online
33.5 (578), 37.6 (546), 60.4 (435), 86.4ꢀ (365 nm).
Hz, 3H) 1.45 (s, 3H), 1.89–2.19 (m, 2H), 2.87 (m, 1H), 3.18–
3.32 (m, 2H), 3.72 (m, 1H), 5.19 (dd, J ¼ 10.85 Hz and 1.07
Hz, 1H), 5.40 (d, J ¼ 4.95 Hz, 1H), 5.75 (dd, J ¼ 17.61 Hz
and 1.07 Hz, 1H), 6.74 (dd, J ¼ 10.85 Hz and 17.61 Hz, 1H),
7.16–7.41 (m, 12H), 7.69 (d, J ¼ 7.97 Hz, 2H).13C-NMR
(CDCl3 , ppm): 15.68, 32.29 (CH3), 36.14, 44.63 (CH2), 62.93
(CH), 72.84 (C), 76.60 (CH), 112.56 (CH2), 125.34–128.06,
132.17 (CH), 135.90 (C), 137.36 (CH), 139.97, 149.87 (C).
11B-NMR (CDCl3 , ppm): 6.67. IR (KBr, cm 1): 3260 (NH),
3040, 2950, 2900, 2830 (CH), 1625 (vinyl-C¼C), 1595, 1490
(C¼C), 1440, 1390, 1365 (CH). MS (DCI-NH3): m=z 429
(Mþ þ NH3), 412 (MHþ ), 308 (MHþ vinyl- phenyl).
Anal. calc. for C27H30BNO2 : C 78.84, H 7.35, N 3.41; found
C 79.05, H 7.33, N 3.21%.
measured polarimetrically. [a]20 (c ¼ 1 in CHCl3): –32.2 (589),
Acknowledgements
We wish to thank Dr U. Matthiesen, University of Dusseldorf,
for the mass spectrometric characterisation of the compounds.
Financial support through the BMBF (reference number
03D0029C6) and ‘‘Fonds der Chemischen Industrie’’ is
gratefully acknowledged.
References
1
(a) A. J. Kirby, Angew. Chem., Int. Ed. Engl., 1996, 35, 707; (b)
Y. Murakami, J. Kikuchi, Y. Hisaeda and O. Hayashida, Chem.
Rev., 1996, 96, 721; (c) W. B. Motherwell, M. J. Bingham and Y.
Six, Tetrahedron, 2001, 57, 4663.
(5S,9S)-Perhydro-5-methyl-4,4,9-triphenyl-2-(p-vinylphenyl)-
1,3,6,2-dioxazaboronine 4. M.p.: 195 ꢀC. [a]20 (c ¼ 0.5 in
D
34.1ꢀ. 1H-NMR (CDCl3–d-DMSO 7:1, ppm):
2
3
4
W. K. Fife, Trends Polym. Sci., 1995, 3, 214.
CHCl3):
R. Breslow and S. D. Dong, Chem. Rev., 1998, 98, 1997.
K. Johnsson, R. K. Allemann, H. Widmer and S. A. Benner,
Nature (London), 1993, 365, 530.
G. M. Blackburn and A. Garcon, in Biotechnology, ed. D. R. Kelly,
Wiley–VCH, Weinheim, Germany, 2nd edn., 2000, vol. 8b, p. 403.
Molecularly Imprinted Polymers, ed. B. Sellergren, Elsevier, Am-
sterdam, 2001.
See, for example: B. C. Gates, Catalytic Chemistry, Wiley, New
York, 1992, p. 285.
(a) S. E. Bystrom, A. Borje and B. Akermark, J. Am. Chem. Soc.,
1993, 115, 2081; (b) G. Wulff and J. Vietmeier, Makromol. Chem.,
1989, 190, 1727.
0.99 (d, J ¼ 7.08 Hz, 3H) 1.55–1.70 (m, 2H), 2.75 (m, 1H),
3.27 (m, 1H), 4.20 (q, J ¼ 7.08 Hz, 1H), 4.46 (dd, J ¼ 10.25
Hz and 2.65 Hz, 1H), 5.13 (dd, J ¼ 10.86 Hz and 1.07 Hz,
1H), 5.69 (dd, J ¼ 17.58 Hz and 1.07 Hz, 1H), 5.95 (m, 1H),
6.68 (dd, J ¼ 10.86 Hz and ¼ 17.58 Hz, 1H), 7.09–7.78 (m,
17H), 7.88 (d, J ¼ 7.26 Hz, 2H).13C-NMR (CDCl3–d-DMSO
7:1, ppm): 18.11 (CH3), 33.94, 48.08 (CH2), 65.32, 70.68
(CH), 84.74 (C), 111.84 (CH2), 124.62–128.19, 133.26 (CH),
134.92 (C), 137.50 (CH), 146.14, 146.74, 149.84 (C). 11B-
NMR (CDCl3–d-DMSO 7:1, ppm): 7.09. IR (KBr, cm 1):
3260 (NH), 3060, 3030, 2960, 2810 (CH), 1627 (vinyl-C¼C),
1600, 1491 (C¼C), 1449, 1391 (CH). MS (DCI-NH3): m=z
491 (Mþ þ NH3), 474 (MHþ ), 370 (MHþ vinylphenyl).
Anal. calc. for C32H32BNO2 : C 81.19, H 6.81, N 2.96; found
C 80.87, H 6.72, N 3.04%.
5
6
7
8
9
M. J. Whitcombe, C. Alexander and E. N. Vulfson, Synlett., 2000,
6, 911 and references cited therein
10 (a) K. Severin, Curr. Opin. Chem. Biol., 2000, 4, 710; (b) F. Loca-
telli, P. Gamez and M. Lemaire, J. Mol. Catal. A., 1998, 135, 89.
11 (a) A. Hirao, S. Itsuno, S. Nakahama and N. Yamazaki, J. Chem.
Soc., Chem. Commun., 1981, 315; (b) S. Itsuno, A. Hirao, S. Na-
kahama and N. Yamazaki, J. Chem. Soc., Perkin Trans. 1, 1983,
1673; (c) S. Itsuno, K. Ito, A. Hirao and S. Nakahama, J. Org.
Chem., 1984, 49, 555; (c) S. Itsuno, M. Nakano, K. Miyazaki, H.
Masuda, K. Ito, A. Hirao and S. Nakahama, J. Chem. Soc., Perkin
Trans. 1, 1985, 2039; (e) S. Itsuno, Y. Sakurai, K. Ito, A. Hirao
and S. Nakahama, Bull. Chem. Soc. Jpn., 1987, 60, 395.
12 (a) E. J. Corey, R. K. Bakshi and S. Shibata, J. Am. Chem. Soc.,
1987, 109, 5551; (b) E. J. Corey, R. K. Bakshi, S. Shibata, C.-P.
Chen and V. K. Singh, J. Am. Chem. Soc., 1987, 109, 7925.
13 (a) E. J. Corey and C. J. Helal, Angew. Chem., Int. Ed. Engl., 1998, 37,
1986; (b) L. Deloux and M. Screbnik, Chem. Rev. 1993, 93, 763; (c) S.
Wallbaum and J. Martens, Tetrahedron: Asymmetry, 1992, 3, 1475.
14 (a) D. K. Jones, D. C. Liotta, I. Shinkai and D. J. Mathre, J. Org.
Chem., 1993, 58, 799; (b) L. P. Linney, C. R. Self and I. H. Wil-
liams, J. Chem. Soc., Chem. Commun, 1994, 1651; (c) A. W. Dou-
glas, D. M. Tschaen, R. A. Reamer and Y.-J. Shi, Tetrahedron:
Asymmetry, 1996, 7, 1303.
15 G. J. Quallich, J. F. Blake and T. M. Woodall, J. Am. Chem. Soc.,
1994, 116, 8516.
16 (a) U. W. Gerwarth, Z. Naturforsch., B, 1977, 32, 1408; (b) I. B.
Mazheika, A. P. Gaukhman, I. P. Urtane, G. I. Zelchan and
E. Lukevits, Zh. Obshch. Khim., 1979, 49, 2528.
17 (a) R. Csuk, H. Honig and C. Romanin, Monatsh. Chem., 1982,
113, 1025; (b) H. Noth and B. Wrackmeyer, Nuclear Magnetic
Resonance Spectroscopy of Boron Compounds, Springer, Berlin,
1978; (c) W. V. Dahlhoff and R. Koster, Liebigs Ann. Chem., 1975,
1625.
Preparation of an imprinted polymer catalyst
A solution of 3 (4.32 g), p-divinylbenzene (10.10 g) and AIBN
(0.15 g) in dry THF (14.40 g) was placed in a polymerisation vial.
The solution was degassed three times through freeze–thaw
cycles, the vial was sealed and the mixture polymerised at 65 ꢀC
for 70 h. The resulting polymer rod was crushed in a mortar,
ground in a laboratory mill and sieved. The fraction between 125
and 45 mm was dried, weighed and packed in a small chroma-
tographic column. The polymer was swollen overnight with
THF, washed with 3 : 1 THF–0.01 M HCl (100 ml per g poly-
mer), rinsed with 3 : 1 THF–water and vacuum-dried. The
splitting percentage was determined on the wash solution as
follows: the THF was evaporated, and the water solution neu-
tralised with 0.1 M NaOH. The resulting suspension was eva-
porated to dryness and taken up in 100 ml CHCl3 . The
concentration of 8 in the CHCl3 solution was measured
polarimetrically; the specifical optical rotations are given above.
To synthesise the polymer-bound oxazaborolidine catalysts,
the washed polymer (5.04 g), norephedrine (2.00 g, 3 equiv.) and
benzene (200 ml) were placed in a round-bottomed flask fitted
with a Soxhlet extractor filled with 4 A molecular sieves. The
+
mixture was refluxed under nitrogen for 90 h. The suspension
was cooled to room temperature and the polymer was filtered
under nitrogen. The polymer was vacuum-dried, poured into a
Soxhlet extractor, and extracted with dry benzene for 24 h under
nitrogen to wash out excess amino alcohol. The solutions were
combined and evaporated to dryness. The residue was taken up
in 100 ml CHCl3 and the concentration of norephedrine was
18 F. Dammast and H.-U. Reißig, Chem. Ber., 1993, 126, 2449.
19 (a) T. W. Campbell and R. N. McDonald, J. Org. Chem., 1959, 24,
1246; (b) G. Wulff and A. Akelah, Makromol. Chem., 1979, 179,
2647.
20 A. F. Popov, Z. Piskunova and V.N. Matvienko, J. Org. Chem.
USSR, 1986, 22, 1299.
21 (a) W. J. Dale and J. E. Rush, J. Org. Chem., 1962, 27, 2598; (b) W.
Vesper, Ph.D. Dissertation, Universitat Bonn, Germany, 1978.
1542
New J. Chem., 2001, 25, 1537–1542