The first encapsulation of hydroxynitrile lyase from Hevea brasiliensis in a sol-gel matrix

Article abstract of DOI:10.1016/j.tet.2004.06.135

A straightforward process for the encapsulation of HbHNL under low methanol conditions has been developed. By adding a sol, prepared by hydrolysis of TMOS/MTMS at pH 2.8 with continuous removal of methanol, to a stirred solution of the enzyme in a buffer at pH 6.5, at least 65% of the activity of the free enzyme could be recovered after the encapsulation. The aquagels were successfully used in the synthesis of (S)-cyanohydrins. Graphical Abstract.

Full text of DOI:10.1016/j.tet.2004.06.135

Tetrahedron 60 (2004) 10419–10425
The first encapsulation of hydroxynitrile lyase from
Hevea brasiliensis in a sol–gel matrix
a,b
Lars Veum, Ulf Hanefeld and Alain Pierre
a,*
b
a
Gebouw voor Scheikunde, Technische Universiteit Delft, Julianalaan 136, 2628 BL Delft, The Netherlands
Institut de Recherches sur la Catalyse, Universit e´ Claude-Bernard-Lyon1, UPR-CNRS 5401, 2 Avenue Albert Einstein,
9626 Villeurbanne cedex, France
b
6
Received 23 April 2004; revised 14 June 2004; accepted 25 June 2004
Available online 21 August 2004
Abstract—A straightforward process for the encapsulation of HbHNL under low methanol conditions has been developed. By adding a sol,
prepared by hydrolysis of TMOS/MTMS at pH 2.8 with continuous removal of methanol, to a stirred solution of the enzyme in a buffer at pH
6.5, at least 65% of the activity of the free enzyme could be recovered after the encapsulation. The aquagels were successfully used in the
synthesis of (S)-cyanohydrins.
q 2004 Elsevier Ltd. All rights reserved.
1
. Introduction
belong to different classes of enzymes and are therefore not
always comparable. For example, HNL from Hevea
brasiliensis is closely related to the a/b hydrolases, which
also include lipases, while Prunus amygdalus is closely
Hydroxynitrile lyases (HNL’s) are a class of enzymes that
can be found in a wide range of plants, such as millet and the
1
apple, almond, rubber and plum trees. When the plant
16,17
related to FAD dependant oxidoreductases.
material is damaged, for instance by a herbivore, HNL
catalyses the breakdown of a cyanohydrin into an aldehyde/
ketone and toxic HCN. More interestingly, the enzyme may
also catalyse the reverse reaction, enabling the synthesis of
enantiopure cyanohydrins from aldehydes/ketones and
The (S)-hydroxynitrile lyase from Hevea brasiliensis
(HbHNL) has been used for the addition of HCN to a
1
8–23
wide range of aldehydes and ketones.
In spite of its
versatility, the only immobilisation reported is the adsorp-
2
–8
11
tion of HbHNL onto celite. Here it was found that the
HCN with excellent yields and enantioselectivities.
These enantiopure cyanohydrins can in turn readily be
converted into a wide range of compounds that are versatile
building blocks for the synthesis of fine chemicals,
maximum activity of the enzyme was only obtained when
there was a discrete water layer surrounding the enzyme.
This indicates that the enzyme is only active in an aqueous
environment. To assure this whilst avoiding a separate
aqueous phase in the reaction, the HbHNL can be
encapsulated in a sol–gel matrix. The pores of such a
matrix can be filled with the aqueous buffer of choice or any
organic phase. In this manner, the versatile HbHNL will
become available for an even wider range of reaction
conditions.
2
,9,10
pharmaceuticals and agrochemicals.
Recently, the immobilisation of enzymes has attracted much
attention, increased stability and recyclability being the
main research objectives. Cross-linked crystals of HNL’s
and HNL’s adsorbed on solid supports or encapsulated in
sol–gels and PVA-gels have been used as catalysts in the
1
1–15
synthesis of optically active cyanohydrins.
However,
The sol–gel technique allows the synthesis of chemically
inert glasses, which can be formed into any desired shape.
They have high porosity (up to 98% pore volume) and high
mechanical and thermal resistance. In addition they can be
produced under conditions that are relatively benign to
enzymes. This technique has successfully been applied for
the encapsulation of lipases into various sol–gel
materials, and as the HbHNL is structurally related to
lipases, we assumed that the encapsulation of HbHNL
should proceed in a similar manner.
only the PVA-entrapped HNL from Prunus amygdalus and
the crosslinked enzyme crystals from Manihot Esculenta
proved to be stable upon recycling. In this context it is
important to notice, that HNL’s only have in common that
they all catalyse the cyanogenesis. Structurally they can
Keywords: Asymmetric catalysis; Hydroxynitrile lyase; Oxynitrilase; Sol–
2
4–28
gel process; Immobilisation; Cyanohydrins.
*
Corresponding author. Tel.: C31-15-278-9304; fax: C31-15-278-4289;
e-mail: u.hanefeld@tnw.tudelft.nl
0040–4020/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2004.06.135

10420
L. Veum et al. / Tetrahedron 60 (2004) 10419–10425
As sol–gel encapsulated CAL-B (33 kDa) which belongs to
the a/b hydrolase family just like the HbHNL, has been
Varying the MTMS (methyltrimethoxysilane)/TMOS
(tetramethoxysilane) ratio is also known to change the
activity of the sol–gel entrapped lipases. Aquagels with a
2
4
40
recycled up to eight times without any loss of activity, it is
reasonable to assume that the dimer of HbHNL (58.4 kDa)
is equally well encapsulated in a sol–gel matrix (HbHNL is
concentration of MTMS in TMOS varying from 0 to
4
1
50 vol% were prepared and used in the addition of HCN to
benzaldehyde, but, no difference in activity could be
detected. No higher MTMS concentrations were examined
since the gelation process then took several hours and the
gels had a paste like aspect. A mixture of 20 vol% MTMS in
TMOS was chosen for further studies since this mixture
gave the most convenient gelation time.
2
9
a dimer in aqueous solutions ). In spite of this structural
similarity, the HbHNL is much more susceptible towards
3
0–34
deviations from its optimum conditions.
Due to this,
newly developed methodology rather than the standard
immobilisation procedures, is applied.
Here we present the first successful encapsulation of the
(S)-selective hydroxynitrile lyase from Hevea brasiliensis in
a sol–gel matrix.
Attempts to dry the aquagels as aerogels and xerogels
resulted in a total loss of activity. In the case of the xerogel
there are two possible reasons for deactivation. When the
gel is dried by evaporation of the water phase in open air
(xerogels), capillary stress will cause partial collapse of the
gel structure as the liquid–gas interface moves in through
the gel. As the gel shrinks, some of the enzymes will also be
crushed. Secondly, it has been suggested that the exposure
of HbHNL to a gas–liquid interface drastically reduces the
2
. Results and discussion
In preliminary experiments, we encapsulated the HbHNL
into a sol–gel matrix following standard procedures.
2
4,26
In
these procedures, the methanol released during the for-
mation of the sol was not eliminated, which caused a
complete deactivation of the HbHNL during the gelation
process. This is in line with the previously described
3
4
half-life of the enzyme. In the preparation of the aerogels,
the water in the aquagel is replaced with acetone, which
again is replaced with CO
supercritical conditions for CO
in an autoclave. When the
are reached, by increasing
2
3
5
methanol sensitivity of HbHNL. In another procedure, the
alkoxy silanes are partially hydrolysed and then transesteri-
fied with glycerol. The methanol is removed from the sol,
and then the sol is mixed with water containing the bio-
2
the temperature of the autoclave, the autoclave is slowly
evacuated. The acetone is most likely causing the deactiva-
tion in this drying procedure. To verify whether this was the
case, the buffer filling the pores of the aquagel was
exchanged with acetone and then back to the buffer by
dialysis. The resulting gel showed no activity, indicating
that indeed acetone or the acetone water mixture did
deactivate the HbHNL. It has been shown that stirring
HbHNL in acetone containing 0.25% water over 15 h at
1
4
molecules. This method is so far the best method reported
for the encapsulation of methanol sensitive enzymes,
however, applied to HbHNL it gave unsatisfactory results.
Therefore, a new procedure was developed, were the alkoxy
silanes were almost 100% hydrolysed by acid mediated
hydrolysis, and the released methanol was removed by
evaporation. The enzyme, dissolved in a buffer with pH 6.5,
was then added to this precursor sol. At this pH, the gelation
was catalysed and at the same time the enzyme was
3
5
room temperature gives only 15% loss of activity. From
this it can be concluded that it is the acetone–water mixture,
rather than the acetone itself, that deactivates the enzyme.
However, as it is known that HbHNL is inactive at low
water concentrations, it is not desirable to dry the gels, but
rather to use them directly as aquagels. In this manner the
enzyme will be completely hydrated with the buffer of
choice. The buffer remains inside the pores of the aquagel
and no separate macroscopic water phase is formed in the
reaction mixture.
3
4,36
stabilized.
As soon as the gel was formed, it was
submerged in the same buffer, pH 6.5, to remove any
remaining methanol, possibly formed from hydrolysis of
residual methoxy groups, by dialysis.
3
7
With this procedure, the aquagels showed an activity of at
least 65% relative to the free enzyme in the standard
aqueous activity test. The apparent decrease of activity is
probably due to deactivation by residual methanol and
diffusion limitations. Initial rate studies showed that the
The encapsulated HbHNL was applied in the enantioselec-
tive addition of HCN to benzaldehyde 3a, furaldehyde 3b,
hexanal 3c, m-phenoxybenzaldehyde 3d and methyl iso-
propyl ketone 3e to give their (S)-cyanohydrins 4a–e. For
safety reasons acetone cyanohydrin was used as the cyanide
source, even though the liberated acetone has a negative
effect on the enzyme (Scheme 1).
3
8
system is indeed limited by diffusion, which indicates that
the actual loss of activity during the encapsulation
procedure is lower than 35%. Due to changes of the specific
particle size of the ground aquagels under the reaction
conditions, quantification of the diffusion limitations was
not pursued.
As both, the generation of HCN and the addition of the HCN
to the carbonyl group, are reversible reactions, the
maximum conversion will be determined by the equilibrium
constants of the two reactions, and, can therefore not be
compared to the isolated yields (typically 95–99%) that are
obtained in the standard method where a five fold excess of
The use of poly vinyl alcohol (PVA) as an additive in the
sol–gels is known to increase the activity of lipases in
2
5
hydrophobic sol–gel materials. Since the structure of
HbHNL is comparable with that of lipases, this possibility
was also investigated, but, no effect of the PVA on the
2
1
free HCN is used. When the amount of enzyme used in the
literature procedure (2.5 times more) is taken into account,
the reaction time that we observe, approximately 40–50 min
(Fig. 1), is comparable with that found in the literature
3
9
enzyme activity could be observed. This indicates that the
presence of PVA might only have an effect on the lipase
activity after or during the drying of the gel.

L. Veum et al. / Tetrahedron 60 (2004) 10419–10425
10421
Scheme 1. The synthesis of optically active cyanohydrins.
2
15 min). The overall equilibrium in Scheme 1 was
1
(
established rapidly in all cases except for 3d, a substrate
Figure 2. Recycling of the sol–gel encapsulated HbHNL using benzal-
dehyde as the substrate. In (A) the gel was washed with diisopropyl ether
between each cycle and in (B) the gel was washed with a 50 mM phosphate
buffer pH 5.0. Cycle 1 (B), cycle 2 (!), cycle 3 (6) and cycle 4 (,).
3
known to be “difficult” for HNL’s. A satisfactory result for
this substrate has only been described in a highly enzyme
loaded emulsion, using free HCN.
2
1
The enantiomeric excess of the products is in line with those
reported in the literature (O98% for 4a–d, and 75% for
4
each cycle. Then, the initial rate only dropped by around
50% for each cycle. In all cases the enantiomeric excess of
the product was higher than 98% (Fig. 2B).
2
e) for the same reactions. Small differences are due to the
1
chemical background reactions (which form the racemic
cyanohydrin) under the specific reaction conditions. To
establish the stability of the sol–gel encapsulated HbHNL, it
was recycled 3 times in the reaction of HCN with
benzaldehyde, using acetone cyanohydrin as the cyanide
source. When the gels were washed with diisopropyl ether
between each cycle, there was a rapid loss of activity and the
gels were completely deactivated after the second recycle
A gel identical to the one used in the experiment described
in Figure 2B was stirred with diisopropyl ether for four
hours (which is the same time as it took to perform the first
three cycles of the experiment described in Fig. 2B) before
benzaldehyde and acetone cyanohydrin were added. The
conversion observed for this gel was comparable with the
conversion in cycle one of Figure 2B, indicating that there is
no deactivation due to the buffer inside the gel or the
solvent. Instead, the loss of activity must be caused by HCN,
(Fig. 2A). This result was improved when the gels were
washed with a 50 mM phosphate buffer of pH 5.0 between
Figure 1. The conversion (B) of 3a (A), 3b (B), 3c (C), 3d (D) 3e (E) into their respective (S)-cyanohydrins 4a–e. 0.98 mM 3a–e, diisopropyl ether, 50 mM
phosphate buffer pH 5.0, 3 equiv acetone cyanohydrin and 1.8 KU HbHNL. The enantiomeric excess (!) was determined by chiral GC (see materials and
methods).

10422
L. Veum et al. / Tetrahedron 60 (2004) 10419–10425
acetone, acetone cyanohydrin, the substrate, the product or
leakage of the enzyme.
washed powder showed a decrease of the initial reaction rate
by 48% compared to the unwashed sol–gel powder. Even
though the gels were shaken and not magnetically stirred in
the recycling experiment, it is probable that some of the gels
got further crushed leading to loss of activity in the
subsequent washing.
In the case of washing with diisopropyl ether, the
deactivation is probably due to the acetone, which is formed
during the reaction. Firstly, acetone is harmful for the
enzyme, secondly, it changes the solubility of the solvent in
the aqueous phase and vice versa. When the solvent is used
for washing some of the water will be washed away from the
gel, lowering the enzyme activity.
In order to characterise the gels, they were dried under
supercritical conditions with CO to give aerogels. BET-
2
2
analysis gave a surface area of 1000 m /g and a pore volume
3
of 1.74 cm /g. The pore size distribution (Fig. 5) is
Whenthegelsarewashedwiththebuffer, the initial conditions
for the enzyme are re-established and the activity is relatively
higher than when washing with a solvent. The loss of activity
willin this case, too, be due to the acetone formed, but alsodue
to leakage of enzyme during the washing procedure.
relatively wide, where the maximum pore radius is 5 nm.
In comparison the dimensions of the HbHNL monomer is
4
3
approximately 3.0!3.8!4.8 nm. After the formation of
the gel, the aquagel is usually aged for 12–72 h in order to
complete the hydrolysis and condensation. To investigate if
there is an optimal ageing time for the HbHNL, we tested
the activity of gels aged up to 16 days. The result (Fig. 6)
shows that there is a significant drop in activity during the
first four to five days; then the activity stabilizes. The initial
drop of activity is probably due to the formation of methanol
during the beginning of the aging process and structural
changes within the gel. The structural changes during the
first days then level off.
In order to rule out any possible leakage of activity into the
organic phase, the suspension of one reaction was filtered
(
taking care that no HCN could escape during the filtration)
and the filtrate was monitored for activity. As expected the
filtrate showed no activity (Fig. 3). However, when the
reaction was performed in water, it was found that up to
1
leached into the aqueous phase.
7% of the observed activity derives from HbHNL that has
4
2
Figure 5. Pore size distribution of the sol–gel; the vertical axis is the
differential dV/dR, where V is the adsorbed nitrogen gas volume, at standard
3
conditions, per gram of gel (cm /g) and R is the pore radius (nm).
Figure 3. Test for leakage of activity from the gel into the organic phase.
Unfiltered reaction (B) and filtered reaction (!).
Encapsulated enzymes are in principle not bound to the sol
gel surface. This means that if the gel breaks and the
‘
capsule’ around an enzyme opens, the enzyme will be
washed away. Indeed, when the activity of an aquagel,
which had been crushed into a fine powder was compared to
an aquagel that had been crushed into a fine powder and then
washed, a significant difference was found (Fig. 4). The
Figure 6. Aging of the gels.
3. Conclusion
HbHNL, an enzyme very sensitive to organic solvents and
requiring a near neutral pH, was successfully immobilized
in a sol–gel. This low methanol immobilisation and first
application of an aquagel holds great potential for the sol–
gel encapsulation of many sensitive enzymes.
Figure 4. Test for leakage of activity during the washing procedure.
Ground gel (B) and ground and washed gel (!).

L. Veum et al. / Tetrahedron 60 (2004) 10419–10425
10423
4
. Experimental
procedure B. Encapsulation of HbHNL in a sol–gel matrix
for standard activity test.
4
.1. General
4.3. General procedure B. Encapsulation of HbHNL in a
sol–gel matrix for standard activity test
HbHNL was made available in a 25 mM potassium
phosphate buffer pH 6.5 with 0.09% sodium azide
(
The activity of the homogenous enzyme was determined as
3600 IU/ml) by Roche Diagnostics (Penzberg, Germany).
The stock solution of HbHNL (100 mg, 3.6 KU/ml) was
diluted to 6.0 g with buffer A. This solution (40 ml) was
added to a mixture of the precursor (500 ml; prepared as
described in general procedure A), and buffer A (460 ml)
and stirred magnetically for 20 s. The stirring bar was
removed and when the mixture gelled (4–5 min), the gel
was submerged in buffer A and aged at 4 8C for 24 h. Buffer
A was then replaced with distilled water and the gels were
aged for further 20 h at 4 8C. The aquagel was ground into a
fine powder and tested for catalytic activity.
3
0
described in the literature. The reactants used in this study
were poly vinyl alcohol with an average molar mass of MZ
1
5,000 (Fluka), methyltrimethoxysilane (MTMS, 98%,
Aldrich) and tetramethoxysilane (TMOS, 98%, Aldrich).
Benzaldehyde, m-phenoxybenzaldehyde, furaldehyde, hex-
anal and methyl isopropyl ketone were all of analytical
grade and distilled under a nitrogen atmosphere less than
two hours before use. Mandelonitrile was purified by
column chromatography at most 24 h prior to use and
stored under nitrogen at K4 8C. Acetone cyanohydrin was
distilled and stored under nitrogen at 4 8C. Diisopropyl
ether was of analytical grade and used without further
purification. The derivatised samples were analysed on a
b-cyclodextrin column (CP-Chirasil-Dex CB 25 m!
4.4. General procedure C. Encapsulation of HbHNL in a
sol–gel matrix for synthetic reaction
A mixture of the stock solution of HbHNL (0.5 ml,
3.6 KU/ml) and the precursor mixture (500 ml; prepared
0
.25 mm) using a Shimadzu Gas Chromatograph GC-14B
equipped with a FID detector and a Shimadzu Auto-injector
according to standard procedure A) was stirred magnetically
for 20 s. The stirring bar was removed and when the mixture
gelled (4–5 min), the gel was submerged in buffer A and
aged at 4 8C for 24 h. Then the buffer in the pores was
exchanged against buffer B by dialysis over 1 h. The
aquagel was ground into a fine powder and used for the
synthesis of enantiopure cyanohydrins.
AOC-20i, using N as the carrier gas. The conversion and
2
the enantiomeric excess were calculated from the peak
areas. The temperature programs and retention times are
given in Table 1. UV measurements were performed on a
UNICAM UV/Vis spectrometer. A Brunauer, Emmett and
Teller (BET) analysis of a gel dried as an aerogel, desorbed
at 200 8C, gave the specific surface area and the pore size
distribution. Buffer A is a 25 mM potassium phosphate
buffer pH 6.5 with 0.09% sodium azide, buffer B is a 50 mM
citrate/potassium phosphate buffer pH 5.0. Racemic refer-
ence compounds were prepared according to standard
4.5. General procedure D. Standard aqueous activity test
for the encapsulated enzyme in aqueous media
Mandelonitrile (80 ml) was dissolved in 10 ml of a citric
acid/potassium phosphate buffer (3 mM, pH 3.5). This
solution (1.4 ml), and all of the sol–gel encapsulated
HbHNL, prepared as described in general procedure B,
were added to buffer B (4.9 ml) at 25 8C. The mixture was
stirred magnetically for another 6 min, before the reaction
was stopped by addition of concentrated HCl (2 drops). A
sample was filtered through cotton and the UV-absorption
of the supernatant was measured at 280 nm against a blank
reaction. The activity was calculated according to the
equation below.
4
4
procedures and their NMR-spectra were in accordance
4
with literature.
4–47
4.2. General procedure A. Preparation of the sol–gel
precursor
Acidic water (1.38 ml, pH adjusted to 2.85 by addition of
HCl) was added to a mixture of MTMS (2.10 g, 15.4 mmol),
TMOS (9.08 g, 58.5 mmol) and distilled water (10.4 ml)
and stirred in a 100 ml round bottom flask until a
homogenous mixture was obtained. The formed methanol
was continuously removed on a rotary evaporator until the
characteristic odours of MTMS, TMOS and MeOH were not
detectable any more. The mixture was then cooled to 0 8C
and water was added until the total volume corresponded to
the initial MTMS/TMOS—volume. The precursor (sol) was
used immediately for the encapsulation of HbHNLGeneral
V
Activity Z
DAbs=min
3280
lST
Activity. The activity of the sample (U/ml); V: total reaction
K1
K1
volume (ml); 3280: 1.376 (l!mmol !cm ); l: path
length in the UV-cell (cm); S: volume of enzyme solution
added in the preparation of the gel (ml); T: reaction time
(min).
Table 1. Temperature programs and retention times for 3a–e, (R)-4a–e and (S)-4a–e
a
Substrate
Temperature program
R
t
(min) 3
R
t
(min) (R)-4
t
R (min) (S)-4
a
b
c
d
e
125 8C (3 min)–20 8C/min–200 8C (0 min)
100 8C (3 min)–20 8C/min–200 8C (0 min)
75 8C (5 min)–30 8C/min–200 8C (1 min)
125 8C (3 min)–20 8C/min–200 8C (13 min)
60 8C (5 min)–2 8C/min–98 8C (0 min)
3.11
3.17
5.49
10.16
2.80
6.35
6.50
9.03
19.19
22.85
6.64
6.80
9.27
19.58
23.07
a
Initial temperature (holding time)–temperature gradient–final temperature (holding time).

10424
L. Veum et al. / Tetrahedron 60 (2004) 10419–10425
4.6. General procedure E. Synthesis of optically active
cyanohydrins
according to general procedure D. The results are given in
Figure 6.
At 25 8C, all of the sol–gel encapsulated HbHNL, prepared
according to general procedure C (1.8 kU HbHNL), was
added to a magnetically stirred solution of the freshly
distilled aldehyde/ketone (4.92 mmol) in diisopropyl ether
Acknowledgements
The authors gratefully acknowledge Dr. Sylvie Maury for
fruitful discussions and ideas, also Roche Diagnostics
Penzberg (W. Tischer) for the generous gift of the enzyme
(
5 ml) which was saturated with buffer B. The reaction was
started by the addition of acetone cyanohydrin (1.35 ml,
4.76 mmol). Samples of 10 ml were taken through the
1
(
HbHNL). L. Veum thanks the European Union for a Marie
septum at different stages of conversion. The samples were
added to a mixture of dichloromethane (0.5 ml), acetic
anhydride (40 ml) and pyridine (40 ml). After at least 12 h at
room temperature the samples were analysed by chiral GC.
The conversion and the enantiomeric excess were calculated
from the relative peak areas of the aldehyde and the
cyanohydrin derivative. The results are given in Figure 1.
Curie fellowship and U. Hanefeld thanks the Royal
Nederlands Academy of Arts and Sciences (KNAW) for a
fellowship.
4.7. Recycling, washing with diisopropyl ether saturated
with a 50 mM citrate/phosphate buffer pH 5.0
References and notes
1
. Kiljunen, E.; Kanerva, L. T. Tetrahedron: Asymmetry 1997, 8,
1551–1557.
The gels were prepared according to general procedure C
and used in the addition of HCN to benzaldehyde according
to general procedure E. The synthetic reaction was stopped
after 1h and the gel was washed with diisopropyl ether
saturated with buffer B before it was reused in a new cycle.
The results of three cycles are given in Figure 2A.
2. North, M. Tetrahedron: Asymmetry 2003, 14, 147–176.
3. Gregory, R. J. H. Chem. Rev. 1999, 99, 3649–3682.
4. Griengl, H.; Hickel, A.; Johnson, D. V.; Kratky, C.; Schmidt,
M.; Schwab, H. Chem. Commun. 1997, 1933–1940.
5. Effenberger, F. Chimia 1999, 53, 3–10.
6
. Schmidt, M.; Griengl, H. Top. Curr. Chem. 1999, 200,
193–226.
4.8. Recycling, washing with a 50 mM citrate/phosphate
buffer pH 5.0
7. Seoane, G. Curr. Org. Chem. 2000, 4, 283–304.
. Effenberger, F.; F o¨ rster, S.; Wajant, H. Curr. Opin. Biotech-
nol. 2000, 11, 532–539.
8
The gels were prepared according to general procedure C
and used in the addition of HCN to benzaldehyde according
to general procedure E, with the following exception: The
synthetic reaction was shaken orbitally and after one hour
the gel was washed on a filter with a 50 mM phosphate
buffer pH 5.0 (40 ml) and then reused in a new cycle. The
results of four cycles are given in Figure 2B.
9. Griengl, H.; Schwab, H.; Fechter, M. Trends Biotechnol. 2000,
18, 252–256.
10. Brussee, J.; van der Gen, A. Stereoselective Biocatalysis,
Ramesh, P. N., Ed.; Marcel Dekker: New York, 2000, pp 289–
320.
11. Costes, D.; Wehtje, E.; Adlercreutz, P. J. Mol. Catal. B:
Enzym. 2001, 11, 607–612.
4
.9. Test for leakage of activity to the organic solvent
12. Costes, D.; Wehtje, E.; Adlercreutz, P. Enzyme Microb.
Technol. 1999, 25, 384–391.
Two gels were prepared according to general procedure C
and used in the addition of HCN to benzaldehyde according
to general procedure E, in two separate reactions A and B.
After 10 min reaction B was filtered in a closed system to
avoid any leakage of HCN. No more conversion could be
detected in the supernatant. The conversion curves of the
two reactions are given in Figure 3.
13. Effenberger, F.; Eichhorn, J.; Roos, J. Tetrahedron: Asym-
metry 1995, 6, 271–282.
14. Gill, I.; Ballesteros, A. J. Am. Chem. Soc. 1998, 120,
8587–8598.
15. Gr o¨ ger, H.; Capan, E.; Barthuber, A.; Vorlop, K. Org. Lett.
2001, 3, 1969–1972.
16. Dreveny, I.; Gruber, K.; Glieder, A.; Thompson, A.; Kratky, C.
Structure 2001, 9, 803–815.
17. Gruber, K.; Kratky, C. J. Polym. Sci. Part A: Polym. Chem.
4
.10. Test for leakage of activity by washing
2
004, 42, 479–486.
Two gels were prepared according to general procedure C
and used in the addition of HCN to benzaldehyde according
to general procedure E, in two separate reactions A and B.
The crushed gel used in reaction B was washed with buffer
B prior to its use in the synthetic reaction. The conversion
curves for the two reactions are given in Figure 4.
18. Klempier, N.; Griengl, H.; Hayn, M. Tetrahedron Lett. 1993,
34, 4769–4772.
19. Klempier, N.; Pichler, U.; Griengl, H. Tetrahedron: Asym-
metry 1995, 6, 845–848.
20. Schmidt, M.; Herv e´ , S.; Klempier, N.; Griengl, H. Tetrahedron
1996, 52, 7833–7840.
2
1. Griengl, H.; Klempier, N.; P o¨ chlauer, P.; Schmidt, M.; Shi, N.;
Zabelinskaja-Mackova, A. A. Tetrahedron 1998, 54,
4
.11. The effect of ageing on the gels
1
4477–14486.
Seven gels were prepared according to general procedure B
and stored for 1, 2, 3, 5, 7, 13 and 16 days respectively. The
activity relative to the free enzyme was then measured
22. Bianchi, P.; Roda, G.; Riva, S.; Danieli, B.; Zabelinskaja-
Mackova, A.; Griengl, H. Tetrahedron 2001, 57, 2213–2220.
23. Roda, G.; Riva, S.; Danieli, B.; Griengl, H.; Rinner, U.;

L. Veum et al. / Tetrahedron 60 (2004) 10419–10425
10425
Schmidt, M.; Zabelinskaja Mackova, A. Tetrahedron 2002,
8, 2979–2983.
4. Reetz, M. T.; Wenkel, R.; Anvir, D. Synthesis 2000, 6,
81–783.
5. Reetz, M. T.; Zonta, A.; Simpelkamp, J. Biotechnol. Bioeng.
996, 49, 527–534.
38. Gels prepared according to standard procedure C were cut or
ground into a particle size of 3–5 mm, 1–2 mm and !0.1 and
used in the addition of HCN to benzaldehyde according to
standard procedure E. The gels showed relative initial rates of
0.2, 0.3 and 1, respectively.
5
2
2
2
2
2
2
7
1
3
9. Three gels were prepared according to procedure B where the
5 mM phosphate buffer pH 6.5 contained 0.0, 1.0 and 2.0%
6. Buisson, P.; Hernandez, C.; Pierre, M.; Pierre, A. C. J. Non-
Cryst. Solids 2001, 285, 295–302.
7. Maury, S.; Buisson, P.; Pierre, A. C. J. Mol. Catal. B: Enzym.
2
PVA respectively. The gels were tested for activity according
to procedure D.
2
002, 19-20, 269–278.
4
0. Reetz, M. T.; Zonta, A.; Simpelkamp, J. Angew. Chem., Int.
Ed. Engl. 1995, 34, 301–303.
8. Reetz, M. T.; Tielmann, P.; Wiesenh o¨ fer, W.; K o¨ nen, W.;
Zonta, A. Adv. Synth. Catal. 2003, 345, 717–728.
4
1. The gels were prepared according to standard procedure C
with different MTMS/TMOS ratios and used in the addition of
HCN to benzaldehyde according to standard procedure E.
2. Gels prepared according to standard procedure B and tested for
activity according to standard procedure D. After 10 min the
reaction mixtures were filtered and the activity of the filtrates
were monitored.
9. Hanefeld, U.; Stranzl, G.; Straathof, A. J. J.; Heijnen, J. J.;
Bergmann, A.; Mittelbach, R.; Glatter, O.; Kratky, C. Biochim.
Biophys. Acta 2001, 1544, 133–142.
4
3
3
3
3
3
3
0. Hanefeld, U.; Straathof, A. J. J.; Heijnen, J. J. Biochim.
Biophys. Acta 1999, 1432, 185–193.
1. Bauer, M.; Geyer, R.; Boy, M.; Griengl, H.; Steiner, W. J. Mol.
Catal. B: Enzym. 1998, 5, 343–347.
4
4
4
4
4
3. Wagner, U. G.; Hasslacher, M.; Griengl, H.; Schwab, H.;
Kratky, C. Structure 1996, 4, 811–822.
2. Persson, M.; Costes, D.; Wehtje, E.; Adlercreutz, P. Enzyme
Microb. Technol. 2002, 30, 916–923.
4. Veum, L.; Kuster, M.; Telalovic, S.; Hanefeld, U.; Maschmeyer,
T. Eur. J. Org. Chem. 2002, 1516–1522.
3. Hickel, A.; Graupner, M.; Lehner, D.; Hermetter, A.; Glatter,
O.; Griengl, H. Enzyme Microb. Technol. 1997, 21, 361–366.
4. Bauer, M.; Griengl, H.; Steiner, W. Enzyme Microb. Technol.
5. Ohta, H.; Kimura, Y.; Sugano, Y.; Sugai, T. Tetrahedron 1989,
4
5, 5469–5476.
1
999, 24, 514–522.
5. Pogorevc, M.; Stecher, H.; Faber, K. Biotechnol. Lett. 2002,
4, 857–860.
6. Ooi, T.; Miura, T.; Takaya, K.; Ichikawa, H.; Maruoka, K.
Tetrahedron 2001, 57, 867–873.
2
7. Inagaki, M.; Hatanaka, A.; Mimura, M.; Hiratake, J.; Nishioka,
T.; Oda, J. Bull. Chem. Soc. Jpn 1992, 65, 111–120.
3
3
6. Gill, I. Chem. Mater. 2001, 13, 3404–3421.
7. In an aquagel, the pores of the gel are filled with water.