Cross-linked aggregates of the hydroxynitrile lyase from Manihot esculenta: Highly active and robust biocatalysts

Article abstract of DOI:10.1002/adsc.200606140

The precipitation and cross-linking into CLEAs of the hydroxynitrile lyase (E.G. 4.1.2.10) from Manihot esculenta was investigated and an optimized procedure, which involved precipitation with (NH₄)₂SO ₄, was developed. It

Full text of DOI:10.1002/adsc.200606140

FULL PAPERS
DOI: 10.1002/adsc.200606140
Cross-Linked Aggregates of the Hydroxynitrile Lyase from
Manihot esculenta: Highly Active and Robust Biocatalysts
Andrzej Chmura,a Geert M. van der Kraan,a Filip Kielar,a Luuk M. van Langen,b
Fred van Rantwijk,a and Roger A. Sheldona,*
a
Laboratory of Biocatalysis and Organic Chemistry, Delft University of Technology, Julianalaan 136, 2628 BL Delft, The
Netherlands
Fax : (+31)-15-278-1415; e-mail: r.a.sheldon@tudelft.nl
CLEA Technologies, Julianalaan 136, 2628 BL Delft, The Netherlands
b
Received: March 24, 2006; Accepted: May 29, 2006
Abstract: The precipitation and cross-linking into CLEA was applied in the hydrocyanation of various
CLEAs of the hydroxynitrile lyase (E.C. 4.1.2.10) aldehydes and ketones in microaqueous medium,
from Manihot esculenta was investigated and an opti- conditions that rendered the non-enzymatic back-
mized procedure, which involved precipitation with ground reaction insignificant, even with unreactive
(NH4)2SO4, was developed. It was found that a fast substrates.
(photometric) assay severely underestimated the ac-
tivity recovery in the CLEA due to rate-limiting dif- Keywords: cross-linked enzyme aggregate; cyanohy-
fusion, whereas in a synthetic assay the activity re- drin synthesis; hydrocyanation; hydroxynitrile lyase;
covery was up to 93% of the starting activity. The Manihot esculenta; microaqueous medium
Introduction
added buffer,[5] conditions that are difficult to recon-
cile with the use of free enzyme. A particular advant-
The (R)- and (S)-selective oxynitrilases (hydroxyni- age of CLEAs, in comparison with carrier-fixed prep-
trile lyases, E.C. 4.1.2.10), which catalyze the addition arations,[6] is that a CLEA consists of nearly pure pro-
of HCN to aldehydes, are undergoing a rapid transi- tein and requires only a minute amount of water to
tion from laboratory curiosities into industrial biocat- maintain full hydration.[5]
alysts, in particular since the (S)-selective oxynitrilas-
es from Manihot esculenta[1] and Hevea brasiliensis[2] ry to recent results from our laboratory,[7] of the prep-
have become available via recombinant expression. aration of CLEAs of the oxynitrilase from M. esculen-
We now report an optimization study, complimenta-
The enantiomeric purity of the product is critically ta (MeHnL) and the use of the latter as a hydrocyana-
dependent on the competition of the enantioselective tion catalyst.
enzymatic hydrocyanation and the non-selective
chemical reaction. It is common practice nowadays to
suppress the chemical reaction by maintaining an Results and Discussion
acidic environment.
A biphasic, aqueous-organic
medium, with an aqueous working phase and an or- Precipitation Study of MeHnL
ganic extractive phase further reduces the contribu-
tion of chemical hydrocyanation.
The preparation of a CLEA involves the precipita-
The non-enzymatic reaction is reduced even further tion/aggregation of the protein followed by cross-link-
in a micro-aqueous reaction medium,[3] in which the ing.[8] We studied the precipitation step in isolation by
amount of aqueous phase is the minimum consistent treating a MeHnL solution with various amounts of
with maintaining enzymatic activity. We have shown organic solvents and salts. The enzymatic activity in
that this approach is particularly effective with a the precipitate – upon redissolution – and in the su-
slow-reacting aldehyde, such as 2-chlorobenzalde- pernatant were measured using a standard UV assay.
hyde.[4] Quite recently, we have shown that a cross- A precipitant threshold concentration was often ob-
linked enzyme aggregate (CLEA) of the (R)-specific served (see Figure 1, panel A for an example). The
oxynitrilase from Prunus amygdalus efficiently cata- activity recoveries at the optimum precipitant concen-
lyzed hydrocyanation in the presence of only 2% of tration, which varied widely, are listed in Table 1. Pre-
Adv. Synth. Catal. 2006, 348, 1655 – 1661
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1655

Andrzej Chmura et al.
FULL PAPERS
Preparation of MeHnL CLEAs
Cross-linking of the precipitated MeHnL with gluta-
raldehyde resulted in the formation of insoluble
CLEAs. These were rinsed and subsequently subject-
ed to a dissolution test. The amount of soluble
enzyme that could be recovered from the CLEAs was
low, typically 0.3–0.5% of the starting activity. The re-
covered activity of the CLEAs, as measured in the
UV test, was low and showed no obvious correlation
with the results of the precipitation-redissolution test
(see Table 1). The activity recoveries in a subsequent
synthetic assay, which were much higher, also did not
reveal a relationship with the precipitation-redissolu-
tion test. This latter observation indicates that struc-
tural changes of the protein upon exposure to the pre-
cipitant are fixed by subsequent cross-linking, where-
as the precipitation-redissolution test is only an indi-
cator of irreversible structural changes. The precipi-
tants of choice for the preparation of MeHnL CLEAs
are 1,2-dimethoxyethane and ammonium sulfate with
activity recoveries of 87 and 93%, respectively. The
morphology of the CLEA particles was investigated
by scanning electron microscopy (Figure 2), which re-
vealed the existence of micropores of 50–100 nm size.
The discrepancy between the results of the UV and
synthetic assays reasserts the importance of diffusion
effects on immobilized enzymes in a fast assay.
Hence, a meaningful assay should be based on a reac-
tion that is not diffusion limited. A hydrocyanation
reaction, such as used above, is quite suited although
the necessity to monitor its progress by HPLC makes
Figure 1. Precipitation of MeHnL with 1,2-dimethoxyethane
(A) and poly(ethylene glycol) (B); activities recovered from
~
^
the precipitate ( ) and the supernatant ( ).
cipitation with poly(ethylene glycol), which was very it somewhat less convenient than a spectrophotomet-
efficient with penicillin acylase,[8] gave an erratic ric assay. In the following, synthetic, part of the paper
result with MeHnL (see Figure 1, panel B) and was we have quantified amounts of MeHnL in mandeloni-
not investigated further.
trile synthesis units (MSU).
Table 1. Effects of the precipitant on the activity recovery.
Precipitant
Conc. [%, v/v][a] Activity in
precipitate [%][b,c]
Glutaraldehyde [%, v/v] CLEA
CLEA synthetic
UV activity [%][b] activity [%][b,d]
Acetone
2-Propanol
1,2-Dimethoxyethane 80
1-Propanol
Ammonium sulfate
tert-Butyl alcohol
Acetonitrile
Ethanol
Poly(ethylene glycol) 50
50
50
100
1.0
1.0
0.5
0.5
0.5
1.0
2.0
0.5
n.d.
11
22
11
14
23
35
15
14
n.d.
6
100
96
94
84
80
74
69
61
38
87
32
93
38
41
49
n.d.
67
80
80
50
80
[a]
Optimized for recovery from the precipitate.
[b] Activity in% recovery of the starting activity.
[c]
UV assay of the redissolved precipitate
[d] Determined in the hydrocyanation of cinnamic aldehyde.
1656
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1655 – 1661

Cross-Linked Aggregates of the Hydroxynitrile Lyase from M. esculenta
FULL PAPERS
Figure 3. Enzymatic hydrocyanation of aldehydes and ke-
tones.
Figure 2. SEM photographs of an MeHnL CLEA at 5·103
(top) and 20·103 (bottom) magnification.
Figure 4. Hydrocyanation of 1a (100 mM) in DIPE-citrate
buffer pH 5.5; MeHnL CLEA (22 MSU mLÀ1), 10% buffer,
Synthetic Application in Microaqeous Medium
~
~
500Àm1 M HCN, progress ( ), ee ( ); MeHnL CLEA (7 MSU
^
^
The MeHnL CLEA prepared by precipitation with mL ), 2% buffer, 700 mM HCN, progress ( ), ee ( ).
ammonium sulfate was selected for further study in
cyanohydrin synthesis. A major advantage of the
CLEA methodology is the possibility to perform reac- zymatic cycling of reactant and product may also play
tions in microaqueous (<2% buffer) diisopropyl a role in the loss of ee.
ether (DIPE), to restrict the uncatalyzed hydrocyana-
tion.
Enantiomerically pure 2a was formed in 96% yield
when the amount of buffer was reduced to 2%. Pre-
The hydrocyanation of benzaldehyde (1a, see sumably because of more efficient interphase mass-
Figure 3), which is a common test-bed for enantiose- transfer – the reaction mixture became monophasic as
lective HCN addition, was investigated first. The en- all of the buffer was absorbed by the biocatalyst – the
antiomeric purity of the produced (S)-mandelonitrile amount of biocatalyst could be reduced three-fold
(2a) was only 96% initially when the reaction was without sacrificing reaction rate (see Figure 4); the
performed in the presence of 10% (v/v) buffer solu- excess of HCN was increased from three- to seven-
tion. This modest enantioselectivity, which is lower fold to delay the enzymatic dehydrocyanation of (S)-
than previously reported, is attributed to a competing 2a. The product ee stayed high in the course of this
uncatalyzed reaction in the pH 5.5 buffer fraction. A latter reaction, presumably because, due to the sup-
slight erosion of the ee became manifest at the ap- pression of the non-enzymatic reaction, there was
proach of equilibrium (see Figure 4). We ascribe this much less (R)-2a formed.
latter phenomenon to equilibrium-related cycling of
The procedure was subsequently scaled up, with
reactant and product, which is bound to result in the minimum change, to a 25 g scale and afforded (S)-2a
accumulation of (R)-2a because the dehydrocyanation in near-quantitative yield and ee. Reducing the
is more (S)-selective than the HCN addition. Non-en- amount of buffer to 0.2% did not result in further im-
Adv. Synth. Catal. 2006, 348, 1655 – 1661
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1657

Andrzej Chmura et al.
FULL PAPERS
Table 2. Hydrocyanation of aldehydes and ketones in the presence of MeHnL CLEA.
Reactant
Enzyme [MSU mLÀ1]
Buffer [%]
HCN [equivs.]
Time [h]
Conversion [%]
ee [%]
1a
22
7
7
3
9
12
400
400
32
32
32
17
10
5
7
7
6
5
12
6
12
12
12
12
12
4.0
3.0
5.5
1.0
1.0
2.0
6.2
6.2
3.8
23
94
96
95
96
96
99
86
90
7
94
99
99
49
42
55
87
92
2.0
0.2
0.1
0.0
0.1
2.5
2.5
0.2
0.2
0.2
0.5
1b
08C
1c
3a
>99
91
98
12
11
90
08C
3b
23
2.5
>96
provement (see Table 2). This latter result attests, ic background reaction, because neither a decrease of
however, to the extreme stability of the MeHnL the pH to 4 nor reducing the amount of buffer had
CLEA, as is also noted in an accompanying study;[7] any effect (data not shown). The MeHNL CLEA
a CLEA of the (R)-specific oxynitrilase from P. amyg- even maintained its activity in completely dry DIPE
dalus, in contrast, required 2% of water.[5]
(Table 2), but no beneficial effect on the product ee
The reduction of the buffer phase that the CLEA became apparent. This latter experiment attests to the
methodology makes possible is particularly attractive extreme robustness of the CLEA.[7]
with a reactant such as 2-propenal (1b), which readily
An experiment with a reduced enzyme load
dissolves in water. It is also noteworthy that the enan- (Figure 5), in which equilibrium was not reached,
tioselectivities that have been obtained in the hydro- demonstrated that MeHnL effects the hydrocyanation
cyanation of 1b in the presence of two very similar of 1b with an inherent ee of 56–57%. We conclude
(S)-oxynitrilases (MeHnL: 56%;[1] HbHnL: 98%[9]) that the erosion of the ee of 2b noted above is caused
do rather diverge and the question arises whether the by (S)-selective enzymatic dehydrocyanation upon the
discrepancy is inherent or caused by an uncontrolled approach of equilibrium.
background reaction. We found that 1b reacted rapid-
A much improved result was obtained when the hy-
ly in microaqueous DIPE although, in contrast with a drocyanation of 1b was performed at 08C. The
previous report,[10] at least a three-fold excess of enzyme load was doubled to compensate for the loss
HCN was required to obtain>95% conversion. The in activity and the excess of HCN was increased to
enantiomeric purity of the product (2b) was a modest 12-fold to retard the enzymatic dehydrocyanation.
55–56% initially and decreased when the reaction Full conversion (>99%) into 2b with 55% ee was
proceeded towards equilibrium conversion (see reached in 2 h under these conditions.
Figure 5). This low ee is not caused by a non-enzymat-
The hydrocyanation of cinnamic aldehyde (1c) is
notoriously slow and suffers from an unfavorable
equilibrium.[11] We accordingly increased the amount
of enzyme (see Table 2), but obtained only 84% con-
version with 87% ee. Eventually, upon the application
of a very large (12-fold) excess of HCN, 2c was ob-
tained in 90% yield and 93% ee, which compares fa-
vorably with a reported result (80% yield, 95%
ee).[12] The question arises whether the somewhat
modest product ee is caused by a non-enzymatic back-
ground reaction, or is inherent to the MeHnL. An ex-
periment with a reduced amount of enzyme, to pro-
long the reaction over time (see Figure 6) showed the
absence of a background reaction. It also became
clear that the enantiopreference of the enzyme is high
but that erosion of the product ee already starts at<
50% conversion, presumably because enzymatic de-
hydrocyanation already makes itself felt.
Figure 5. Hydrocyanation of 1b (100 mM) in DIPE-citrate
buffer pH 5.5; MeHnL CLEA (3 MSU mLÀ1), 0.1% buffer,
~
~
The conclusions from the hydrocyanation of 1a–c in
the presence of the CLEA biocatalyst are two-fold.
600 mM HCN, progress ( ), ee ( ); MeHnL CLEA (1.8
À1
^
^
MSU mL ), 500 mM HCN, progress ( ), ee ( ).
1658
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1655 – 1661

Cross-Linked Aggregates of the Hydroxynitrile Lyase from M. esculenta
FULL PAPERS
Conclusions
An optimized procedure affords
a cross-linked
enzyme aggregate (CLEA) of the hydroxynitrile lyase
from Manihot esculenta with up to 93% recovery of
the synthetic activity. A fast (photometric) assay se-
verely underestimated the activity recovery in the
CLEA due to rate-limiting diffusion.
The CLEA is a very stable biocatalyst that is very
suitable for hydrocyanation reactions in microaqueous
and even anhydrous media. Such reaction conditions
render the non-enzymatic hydrocyanation insignifi-
cant. The synthetic potential of the CLEA has been
demonstrated in the enantioselective hydrocyanation
of a number of aldehydes and ketones. A decay of the
product ee was observed upon the approach of equi-
librium conversion and was ascribed to enzymatic cy-
cling of reactant and product.
Figure 6. Hydrocyanation of 1c (100 mM) in DIPE-citrate
buffer pH 5.5, 1.2M HCN; MeHnL CLEA (400 MSU
À1
~
~
mL ), 0.25%Àb1 uffer, progress ( ), ee ( ); MeHnL CLEA
^
^
(200 MSU mL ), progress ( ), ee ( ).
First, the severe reduction of the buffer phase that Experimental Section
the CLEA tolerates renders the non-enzymatic back-
ground reaction insignificant. Second, the erosion of Materials
the product ee upon the approach of equilibrium,
Semi-purified (S)-hydroxynitrile lyase from Manihot escu-
which should be expected on theoretical grounds but
previously was obscured by the non-enzymatic hydro-
cyanation, is now a major limiting factor whenever
the intrinsic enantiomer preference of the biocatalyst
is short of absolute.
The high enzyme concentrations that become possi-
ble upon adopting the CLEA methodology are partic-
ularly advantageous in the hydrocyanation of recalci-
trant reactants, such as ketones. Thus, the hydrocyana-
tion of acetophenone (3a, see Figure 6) is known to
be incomplete and the ee of the adduct (4a) is modest
(up to 87%)[1,12] when the reaction is performed in
the presence of MeHnL in biphasic medium.
lenta (3000 U mLÀ1) was obtained from Jülich Fine Chemi-
cals (Jülich, Germany). A CLEA of MeHnL [batch number
CLEAMEHNL-S03–150–03512, 78 mandelonitrile synthesis
units (MSU, see below) mgÀ1] from CLEA Technologies
(Delft, The Netherlands) was used in the synthetic hydro-
cyanation experiments.
Acetophenone, acrolein, benzaldehyde, cinnamic alde-
hyde, diisopropyl ether (DIPE), 1,3-dimethoxybenzene and
1-phenyl-2-propanone were obtained from Acros, Aldrich or
Fluka and were used as received. Racemic mandelonitrile
was obtained from Fluka. Racemic 2-hydroxy-3-buteneni-
trile,[13] 2-hydroxy-2-methyl-3-phenylpropionitrile[11b] and 2-
hydroxy-4-phenyl-trans-3-butenenitrile[11b] were prepared
according to literature procedures.
We performed the hydrocyanation of 3a at 0.2%
buffer concentration and found that the reaction did
not proceed beyond 8–11% conversion, depending on
the reactant concentrations. Because adding extra
enzyme had no effect – which attests that the stagna-
tion is not caused by biocatalyst deactivation – we
conclude that equilibrium is already reached at these
conversions. The non-catalytic hydrocyanation was
negligible (approx. 1% over 24 h) at 0.2% of buffer.
A 2M solution of hydrogen cyanide in DIPE was pre-
pared from sodium cyanide as described.[4] Warning: sodium
cyanide and HCN are highly poisonous. They should be han-
dled in a fume cupboard with a good draught. It is strongly
advised to keep an HCN alarm switched on.
Analytical Procedures
The product ee was initially>99% but slowly de- The progress of the reactions was measured by HPLC; en-
antiomeric purities were measured by chiral HPLC or GC
as described. HPLC analyses were carried out using a
Waters 510 pump and a Waters 468 variable wavelength UV
detector at 215 nm. GC analyses were performed on a Shi-
madzu GC-17 instrument equipped with a FID detector and
a Varian Inc. 25 m ” 0.32 mm Chirasil-Dex CB column; the
carrier gas was N2. Samples (10 mL) were treated with a de-
rivatization reagent (300 mL) that contained acetic anhy-
dride (0.8 mL) and pyridine (0.8 mL) in dichloromethane
(10 mL) and analyzed after standing for 1 h. Further details
are given below.
creased at the approach of equilibrium (see Table 2),
as before. The enantiomeric purity of the product was
improved to 98% (at 11% conversion, see Table 2)
by performing the reaction at 08C.
1-Phenylpropanone (3b) reacted much faster than
3a and the equilibrium conversion was much higher
as well. Its HCN adduct (4b) was obtained in 90%
yield and>96% ee.
Adv. Synth. Catal. 2006, 348, 1655 – 1661
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1659

Andrzej Chmura et al.
FULL PAPERS
citrate buffer) were assayed photometrically. The optimum
amounts of glutaraldehyde, as regards activity recovery in
the CLEA, are listed in Table 1. The residual activity in the
supernatant was<0.5% of the starting activity in all experi-
ments.
A leakage test was performed by keeping CLEAs for 3 d
in 20 mM citrate buffer pH 5.5. The activity recovery in the
supernatant was<0.5%.
The MeHnL CLEA for use in the hydrocyanation experi-
ments was prepared by precipitation with ammonium sulfate
and cross-linking with 0.5% (v/v) of glutaraldehyde solution,
washed with citrate buffer as described above and lyophi-
lized.
Photometric Assay
To citrate buffer pH 5.5 (3 mL, 20 mM) in a quartz cuvette
was added a solution of mandelonitrile in 1,2-dimethoxy-
ethane (0.4%, v/v, 20 mL) and the uncatalyzed dehydrocya-
nation was monitored spectrometrically (Cary 3 bio UV
spectrophotometer) at 250 nm and 258C. Then an appropri-
ate amount of MeHnL was added and the enzymatic dehy-
drocyanation was monitored as above.
The activity recovery from precipitation and cross-linking
experiments was measured with a 20 mL sample from the 10
times diluted pellet or supernatant solutions. One unit (U)
of MeHnL will liberate 1 mmol of benzaldehyde per min
under these conditions.
Kinetic Assay
Enzymatic Hydrocyanation; General Procedure
2-Hydroxy-4-phenyl-trans-3-butenenitrile: To citrate buffer
pH 5.5 (2 mL, 20 mM) were added MeHnL solution
(250 mL) or an equivalent amount of CLEA, DIPE
(0.4 mL), cinnamic aldehyde solution in DIPE (2M, 0.4 mL)
and HCN solution in DIPE (2M, 1.2 mL). The mixture was
stirred and a temperature of 08C was maintained. The reac-
tion vessel was kept closed to prevent escape of HCN.
A 20 mL sample was taken every h from the organic phase
for up to 3 h and dissolved into hexane containing 2% TFA
(1 mL). The remaining water was removed by adding some
Na2SO4. The sample was centrifuged and analyzed by
HPLC.
Mandelonitrile: To citrate buffer pH 5.5 (2 mL) in a
10 mL screw-cap vessel were added an appropriate amount
of enzyme preparation, HCN solution in DIPE (2M,
1.2 mL), and DIPE (700 mL). The mixture was stirred mag-
netically, chilled to 08C in ice/water and benzaldehyde
(100 mL) was added. After 10 min (<30% conversion) a
sample (10 mL) was taken from the organic phase, diluted
with eluant (990 mL), dried over Na2SO4 and analyzed by
HPLC. One mandelonitrile synthesis unit (MSU) will syn-
thesize one mmol of mandelonitrile per min under these con-
ditions.
The reagents and internal standard (1,3-dimethoxybenzene)
were taken from 2M stock solutions in DIPE. DIPE was
kept saturated with 20 mM citrate buffer pH 5.5 (unless
noted otherwise). The MeHnL CLEA biocatalyst (from
CLEA Technologies, equivalent to the one described above)
was kept as a suspension in DIPE and appropriate amounts
were pipetted into the reaction mixtures.
The hydrocyanation experiments were carried out in 2 mL
glass reactors equipped with a PTFE-lined screw cap. The
appropriate amounts of buffer, DIPE, MeHnL CLEA sus-
pension, HCN solution and internal standard were intro-
duced into the reactor, which was magnetically stirred and
thermostatted using a water bath (258C) or a cryostat
(08C). The reaction was started by adding the aldehyde or
ketone (100 mM with respect to the total volume).
Samples (10 mL) for HPLC analysis were taken periodi-
cally, diluted with eluant (990 mL), dried over sodium sulfate
and centrifuged prior to injection.
Benzaldehyde (1a): The stock solution of 1a was kept
over saturated NaHCO3 to prevent accumulation of benzoic
acid. The hydrocyanation reactions of 1a were carried out as
described in the general procedure. Progress and ee were
measured by HPLC [4.6”250 mm 5 m Chiralcel OB-H
column (Daicel), eluant hexane/2-propanol (90:10, v/v) at
0.5 mLminÀ1 at 358C]. Retention times of mandelonitrile:
17.3 min (R), 18.3 min (S).
Precipitation Experiments
A solution of MeHnL (250 mL, 3000 U mLÀ1) in an Eppen-
dorf tube was mixed with phosphate buffer pH 8 (50 mL,
0.1M), an appropriate amount of precipitant was added and
the mixture was mechanically shaken at 08C for 4 min, then
centrifuged for 3 min. To the supernatant citrate buffer
pH 5.5 (20 mM) was added to a total volume of 1 mL; the
pellet was dissolved in citrate buffer pH 5.5 (1 mL, 20 mM).
The enzymatic activity in both fractions was measured pho-
tometrically as described above.
Acrolein (1b): The hydrocyanation reactions of 1b were
carried out as described in the general procedure. The reac-
tion progress was measured by HPLC (4.6”250 mm 5 m
Chiralcel OB-H column (Daicel), eluant hexane/2-propanol
(90:10, v/v) at 0.5 mLminÀ1). Chiral analysis was carried out
by GC as described above; temperature program: 10 min at
1008C, then a gradient of 138C minÀ1 to 1808C. Retention
times of 2-hydroxy-3-butenenitrile: 2.2 min (R), 2.8 min (S).
Cinnamic aldehyde (1c): The hydrocyanation experiments
of 1c were carried out as described in the general procedure.
The reaction progress was measured by HPLC [4.6”250 mm
5 m Chiralcel OB-H column (Daicel), eluant hexane/2-prop-
anol (90:10, v/v) at 0.5 mLminÀ1]. Chiral analysis was car-
ried out by GC as described above at 1608C column temper-
ature. Retention times of 2-hydroxy-4-phenyl-trans-3-bute-
nenitrile: 7.3 min (R), 8.1 min (S).
CLEA Preparation
Precipitates of MeHnL were prepared using the solvents
and concentrations listed in Table 1. Glutaraldehyde (25%
in water) was added in amounts of 0.5, 1.0 and 2.0% (v/v)
with respect to the total volume and the resulting suspen-
sions were mechanically shaken at 08C for 5 min. The
CLEA was separated by centrifugation, resuspended in
20 mM citrate buffer pH 5.5 and centrifuged. The enzymatic
activity in the supernatant and the pellet (resuspended in
Acetophenone (3a): The hydrocyanation reactions of 3a
were carried out as described in the general procedure.
Progress and ee were measured by HPLC [4.6”250 mm 5 m
Chiralcel OB-H column (Daicel), eluant hexane-2-propanol
1660
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1655 – 1661

Cross-Linked Aggregates of the Hydroxynitrile Lyase from M. esculenta
FULL PAPERS
(90:10, v/v) at 0.5 mLminÀ1 at 358C]. Retention times of 2-
hydroxy-2-phenylpropionitrile: 14.0 (R), 14.6 (S).
Council NWO under the CERC3 programmed is gratefully
acknowledged.
1-Phenylpropanone (3b): The hydrocyanation reactions of
3b were carried out as described in the general procedure
but at 6 mL scale in a 10 mL reaction vessel, to accommo-
date the increased sample volume. The reaction progress
was monitored by HPLC {4.6”250 mm 5 m Chiralcel OD-H
column (Daicel), eluant hexane/2-propanol (90:10, v/v) at
0.5 mLminÀ1].
Samples for chiral analysis (200 mL) were mixed with a
derivatization reagent composed of butyric anhydride
(40 mL), N,N-dimethylaminopyridine (40 mg) in dioxane
(400 mL), stirred for 2 h and washed with dilute HCl. A
sample of the organic phase was taken, diluted with eluant
and analyzed by HPLC [4.6”250 mm 5 m Chiralcel OB-H
column (Daicel), eluant hexane/2-propanol (90:10, v/v) at
0.5 mLminÀ1 at 358C]. Retention times of 2-hydroxy-2-
methyl-3-phenylpropionitrile: 22.1 min (R), 33.0 min (S).
References
[1] S. Fçrster, J. Roos, F. Effenberger, H. Wajant, A. Spra-
uer, Angew. Chem. Int. Ed. Engl. 1996, 35, 437–439.
[2] M. Hasslacher, M. Schall, M. Hayn, H. Griengl, S. D.
Kohlwein, H. J. Schwab, J. Biol. Chem. 1996, 271,
5884–5891.
[3] a) S. Han, G. Lin, Z. Li, Tetrahedron: Asymmetry 1998,
9, 1835–1838; b) S. Han, P. Chen, G. Lin, H. Huang, Z.
Li, Tetrahedron: Asymmetry 2001, 12, 843–846.
[4] L. M. van Langen, F. van Rantwijk, R. A. Sheldon,
Org. Proc. Res. Dev. 2003, 7, 828–831.
[5] L. M. van Langen, R. P. Selassa, F. van Rantwijk, R. A.
Sheldon, Org. Lett. 2005, 7, 327–329.
Large-Scale Synthesis of (S)-mandelonitrile (2a)
[6] a) F. Effenberger, T. Ziegler, S. Fçrster, Angew. Chem.
Int. Ed. Engl. 1987, 26, 458–460; b) E. Wehtje, P.
Adlercreutz, B. Mattiasson, Biotechnol. Bioeng. 1993,
41, 171–178; c) H. Grçger, E. Capan, A. Barthuber,
K.-D. Vorlop, Org. Lett. 2001, 3, 1969–1972; d) L.
Veum, U. Hanefeld, A. Pierre, Tetrahedron 2004, 60,
10419–10425.
[7] F. L. Cabirol, U. Hanefeld, R. A. Sheldon, Adv. Synth.
Catal. 2006, 348, (F606139).
[8] L. Cao, F. van Rantwijk, R. A. Sheldon, Org. Lett.
To DIPE (0.25 L) in a 1 L round-bottomed flask were added
MeMnL CLEA (0.25 g, 19.6 kMSU) and 20 mM citrate
buffer pH 5.5 (2 mL). The mixture was stirred vigorously in
an ice-water batch until a fine suspension was obtained.
HCN was added as a 2M solution in DIPE (200 mL), fol-
lowed by a solution of benzaldehyde (25 g, 0.24 mol) in
DIPE (250 mL). The mixture was stirred at room tempera-
ture.
The progress and product ee were monitored as described
above. Complete conversion was obtained in 2 h. The reac-
tion mixture was filtered to remove the biocatalyst, dried
over Na2SO4 and concentrated under vacuum at 508C until
the last traces of HCN had been removed. Compound 2a
was obtained as a colorless oil in 99% yield and 99.6% ee.
2000, 2, 1361–1364.
[9] H. Griengl, N. Klempier, P. Pçchlauer, M. Schmidt, N.
Shi, A. A. Zabelinskaja-Mackova, Tetrahedron 1998,
54, 14477–14486.
[10] N. Klempier, U. Pichler, H. Griengl, Tetrahedron:
Asymmetry 1995, 6, 845–848.
[11] a) E. G. J. C. Warmerdam, A. M. C. H. Van den Nieu-
wendijk, C. G. Kruse, J. Brussee, A. van der Gen, Recl.
Trav. Chim. Pays-Bas 1996, 115, 20–24; b) P.-J. Gerrits,
W. F. Willeman, A. J. J. Straathof, J. J. Heijnen b, J.
Brussee, A. van der Gen, J. Mol. Catal. B: Enzym.
2001, 15, 111–121.
Acknowledgements
Experimental assistance by Mr.Adem Celik, Ms. Jessica
Cuevas and Ms. Elena Latorre is gratefully acknowledged.
The authors thank Dr. Rob Schoevaart (CLEA Technologies,
Delft, The Netherlands) for the SEM photos. Thanks are due
to Mr. Fabien Cabirol (TU Delft and ICES, Singapore) for
discussions. Financial support by the Netherlands Research
[12] H. Bühler, F. Effenberger, S. Fçrster, J. Roos, H.
Wajant, ChemBioChem. 2003, 4, 211–216.
[13] P. G. Gassman, J. J. Talley, Tetrahedron Lett. 1978, 19,
3773–3776.
Adv. Synth. Catal. 2006, 348, 1655 – 1661
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1661