Synthesis of aliphatic (S)-α-hydroxycarboxylic amides using a one-pot bienzymatic cascade of immobilised oxynitrilase and nitrile hydratase

Article abstract of DOI:10.1002/adsc.200800625

A one-pot bienzymatic cascade combining a hydroxynitrile lyase (Manihot esculenta, E.C. 4.1.2.10) and a nitrile hydratase (Nitriliruptor alkaliphilus, E.C. 4.2.1.84) for the synthesis of enantiopure aliphatic α- hydroxycarboxylic amides from aldehydes is described. Both enzymes were immobilised as cross-linked enzyme aggregates (CLEAs). Stability tests show that the nitrile hydratase CLEAs are sensitive to water-immiscible organic solvents as well as to aldehydes and hydrogen cyanide (HCN), but are remarkably stable and show useful activity in acidic aqueous environments of pH 4-5. The cascade reactions are consequently carried out by using a portionwise feed of HCN and moderate concentrations of aldehyde in acidic aqueous buffer to suppress the uncatalysed hydrocyanation background reaction. After optimisation, this method was used to synthesise five different kinds of aliphatic α- hydroxycarboxylic amides from the corresponding aldehydes with good yields and with enantiomeric purities comparable to those obtained for the α-hydroxynitriles in the microaqueous hydrocyanation using hydroxynitrile lyase and an excess of HCN.

Full text of DOI:10.1002/adsc.200800625

FULL PAPERS
DOI: 10.1002/adsc.200800625
Synthesis of Aliphatic (S)-a-Hydroxycarboxylic Amides using a
One-Pot Bienzymatic Cascade of Immobilised Oxynitrilase and
Nitrile Hydratase
Sander van Pelt,a Fred van Rantwijk,a and Roger A. Sheldona,*
a
Laboratory of Biocatalysis & Organic Chemistry, Department of Biotechnology, Delft University of Technology,
Julianalaan 136, 2628 BL Delft, The Netherlands
Fax : (+31)-15-278-1415; phone: (+31)-15-278-2683; e-mail: r.a.sheldon@tudelft.nl
Received: October 13, 2008; Revised: December 26, 2008; Published online: January 26, 2009
Abstract: A one-pot bienzymatic cascade combining of aldehyde in acidic aqueous buffer to suppress the
a hydroxynitrile lyase (Manihot esculenta, E.C. uncatalysed hydrocyanation background reaction.
4.1.2.10) and a nitrile hydratase (Nitriliruptor alkali- After optimisation, this method was used to synthe-
philus, E.C. 4.2.1.84) for the synthesis of enantiopure sise five different kinds of aliphatic a-hydroxycar-
aliphatic a-hydroxycarboxylic amides from aldehydes boxylic amides from the corresponding aldehydes
is described. Both enzymes were immobilised as with good yields and with enantiomeric purities com-
cross-linked enzyme aggregates (CLEAs). Stability parable to those obtained for the a-hydroxynitriles
tests show that the nitrile hydratase CLEAs are sen- in the microaqueous hydrocyanation using hydroxy-
sitive to water-immiscible organic solvents as well as nitrile lyase and an excess of HCN.
to aldehydes and hydrogen cyanide (HCN), but are
remarkably stable and show useful activity in acidic
aqueous environments of pH 4–5. The cascade reac- Keywords: amides; cascade reaction; cross-linked
tions are consequently carried out by using a por- enzyme aggregates; hydrocyanation; hydroxynitrile
tionwise feed of HCN and moderate concentrations lyase; nitrile hydratase
Introduction
matic hydrolysis of the cyano group using a stereose-
lective nitrilase (NLase, E.C. 3.5.5.1).[4] The remaining
Chiral a-hydroxycarboxylic amides, like the corre- undesired cyanohydrin enantiomer can easily be
sponding acids, are versatile building blocks in organic racemised in situ under basic reaction conditions, re-
synthesis because of the ease of transformation of the sulting in a dynamic kinetic resolution (DKR) of the
functional groups in these molecules. In the last two nitrile. The main limitation of this method is that
decades, research was mainly focused on the develop- there are only a few nitrilases available that are truly
ment of biocatalytic synthesis routes to enantiopure enantioselective and that this enantioselectivity is usu-
a-hydroxycarboxylic acids.[1] However, a recent brain- ally limited to mandelonitrile and its derivatives.
storm session and cross company debate of the Phar-
Besides the aforementioned chemoenzymatic
maceutical Roundtable identified a definite need for routes, a novel method employing a bienzymatic one-
greener and more atom efficient methods of amide pot cascade for the synthesis of (S)-a-hydroxycarbox-
formation in the pharmaceutical industry.[2]
ylic acids was recently published. In this method an
Enantiomerically pure (R)- and (S)-a-hydroxyni- (S)-selective HnL from Manihot esculenta is com-
triles can be synthesised enzymatically from the corre- bined with a non-selective NLase from P. fluorescens
sponding aldehydes using an oxynitrilase (hydroxyni- in a combi-CLEA or in the acidotolerant fungal host
trile lyase, HnL, E.C. 4.1.2.10).[3] Chemical hydration strain Pichia pastoris.[5] This fully enzymatic route to
or hydrolysis of these enantiopure nitriles using a (S)-a-hydroxy acids is environmentally benign and
strong acid will then give rise to the corresponding compatible with a wide range of hydrolytically sensi-
enantiopure carboxylic amides or acids. However, tive groups because of the mild reaction conditions.
chemical hydration or hydrolysis is usually not com-
An analogous bienzymatic cascade can be envis-
patible with sensitive functional groups and generates aged for the one-pot synthesis of a-hydroxycarboxylic
large amounts of salt waste. A greener route to a-hy- amides by using a nitrile hydratase (NHase, E.C.
droxycarboxylic acids comprises the asymmetric enzy- 4.2.1.84) instead of a NLase in combination with a
Adv. Synth. Catal. 2009, 351, 397 – 404
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
397

Sander van Pelt et al.
FULL PAPERS
Scheme 1. Access to (R)- and (S)-a-hydroxycarboxylic amides and acids using a one-pot bienzymatic cascade of HnL and
NLase or HnL and NHase. The stereogenic centre is set in the molecule by an (R)- or (S)-selective HnL. Subsequent hydra-
tion/hydrolysis is catalysed by an aselective NHase/NLase.
(R)- or (S)-HnL. Like NLases, NHases are industrial- tions quite well, especially when diisopropyl ether
ly relevant enzymes.[6] Their main drawbacks are low (DIPE) is used as the water-immiscible organic sol-
stability in cell-free preparations and low to nonexis- vent.[11]
tent enantioselectivity on most substrates tested,
For NHases, in contrast, this environment is far
which makes the use of a DKR on a-hydroxynitriles from ideal. In order to find out whether NHase
practically impossible. Recently, a new NHase-con- CLEAs can still function under HnL hydrocyanation
taining organism from soda lakes was isolated by our conditions, an assessment was made of their stability
group[7] and subsequently the cell-free NHase from in DIPE/buffer mixtures of different composition.
this organism was stabilised by the formation of MeHnL CLEAs tolerate lyophilisation very well and
CLEAs.[8] When a bienzymatic cascade of HnL and can form a homogeneous suspension in DIPE, which
NHase is used for the production of a-hydroxycarbox- can be kept for weeks in the refrigerator without loss
ylic amides, the low selectivity of the NHase is no of activity. In contrast, NHase CLEAs suffer from a
longer a drawback since either an (R)- or (S)-selective high loss of activity during freeze drying and are most
HnL will impart its enantioselectivity on the reaction stable when stored in an aqueous environment.[8] In-
(Scheme 1).
cubation of the NHase CLEAs in different DIPE/
We initially studied the stability of NHase CLEAs buffer mixtures resulted in an average activity loss of
under the conditions used for HnL-catalysed hydro- 40% in one hour. In addition to the low stability, it is
cyanation. After establishing the stability, the condi- cumbersome to suspend the NHase CLEAs in these
tions of the one-pot enzymatic cascade were opti- mixtures, since the particles coagulate easily and tend
mised and consequently a-hydroxycarboxylic amides to stick to the walls of the reaction vessel.
were produced from several different aliphatic alde-
hydes.
Since the use of biphasic as well as micro-aqueous
systems was not feasible, the stability of the NHase
CLEAs in low pH aqueous buffer systems was deter-
mined. Although the optimum pH of the NHase from
Nitriliruptor alkaliphilus is located around 8 (Figure 1
A), it proved to be surprisingly stable at pH values 2–
4 units below the optimum (Figure 1 B) and still had
a useful activity under these conditions.
Results and Discussion
Stability of NHase CLEAs under HnL
Hydrocyanation Conditions
It is well known that NHases are highly sensitive to
HnL-catalysed hydrocyanations are preferably carried low to moderate cyanide concentrations.[12] Presuma-
out under acidic conditions (pH <5) to suppress the bly, the deactivation and/or inhibition of the NHase
competing unselective uncatalysed hydrocyanation,[9] by cyanide is caused by the strong interaction of cya-
which will cause erosion of the enantiomeric excess nide with the cobalt or iron metal centre of this
(ee) of the a-hydroxynitrile product. Biphasic aque- enzyme, which is assumed to be essential for cataly-
ous-organic as well as micro-aqueous systems can sis.[13] NHases that can tolerate moderate to high cya-
help to further suppress the uncatalysed hydrocyana- nide concentrations may have interesting applications
tion.[10] The MeHnL CLEA can tolerate these condi- in the conversion of nitrile compounds that readily re-
398
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 397 – 404

Synthesis of Aliphatic (S)-a-Hydroxycarboxylic Amides
FULL PAPERS
Figure 2. A Deactivation of the NHase CLEAs caused by
incubation in the presence of different concentrations of
HCN at 218C and pH 5.5, incubation of 0 min (&), 30 min
(&), 90 min (&). B Deactivation of NHase CLEAs caused
by incubation in presence of different concentrations of
acrolein at 218C and pH 5.5, incubation of 0 h (&), 2 h (&),
3.5 h (&), 18.5 h (&).
Figure 1. A NHase CLEA pH activity profile.[8] B Deactiva-
tion of the NHase CLEAs caused by incubation in acidic
buffer solutions at 218C (pH 4–5.5, 0.01M citrate buffer;
pH 6, MilliQ), pH 4 (^), pH 4.5 (&), pH 5 (~), pH 5.5
(*), pH 6 ( ).
*
lease cyanide because of decomposition, like a-hy- Schiffꢁs base formation with the e-amino groups of
droxy- or a-aminonitriles.[14] In HnL-catalysed hydro- free lysine residues but they can also react with thiols,
cyanation reactions, HCN is usually added in excess phenols and imidazoles.[15] These interactions with the
to drive the equilibrium to complete conversion. The enzyme can lead to a change in enzyme conformation
stability of the NHase CLEA activity during incuba- and subsequent deactivation. In order to investigate
tion in the presence of different HCN concentrations the effects of aldehydes on the NHase CLEAs, these
is therefore of importance and was investigated were incubated in the presence of different concentra-
(Figure 2 A). The activity was reasonably preserved tions of acrolein, which was one of the intended sub-
during incubation in the presence of HCN concentra- strates of the cascade reaction (Figure 2 B). Deactiva-
tions ꢀ11 mM. HCN concentrations ꢁ21 mM result- tion of the NHase CLEAs in the presence of moder-
ed in severe loss of activity after an incubation of ate acrolein concentrations was quite severe. When
30 min. These results compare favourably with previ- the concentration of acrolein was monitored during
ously published results.[12] The activity of the CLEAs experiments without HCN it is clear that the presence
remained the same after washing them several times of NHase CLEAs is connected to a disappearance of
with buffer and storage for a week in an open atmos- acrolein (Figure 3). No alien peaks were detected
phere, indicating that the loss of activity is permanent. during HPLC measurements, which might indicate
Nonetheless, the relatively high HCN tolerance of the that a part of the acrolein binds to the cross-linked
NHase from Nitriliruptor alkaliphilus made this enzyme aggregates. Apparently the acrolein is able to
enzyme a good candidate for use in the cascade reac- react either with amine groups that were not occupied
tion system.
during cross-linking of the NHase aggregates or with
In general, aldehydes have the characteristic to other essential functional groups. In contrast, propio-
bind to enzymes mostly through non-conjugated naldehyde did not disappear during incubation with
Adv. Synth. Catal. 2009, 351, 397 – 404
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
399

Sander van Pelt et al.
FULL PAPERS
the NHase CLEAs, which indicates that the enzyme
reacts with the C=C double bond of acrolein in a Mi-
chael-type addition. Nevertheless, the enzyme was de-
activated by moderate concentrations of propionalde-
hyde.
During the preparation of NHase CLEAs with glu-
taraldehyde no reduction step was carried out since
no leakage of free enzyme from the CLEA matrix
was detected. No noticeable effect on CLEA stability
in organic solvent or in presence of acrolein was ob-
served after reduction of the CLEAs using NaBH4.
The Schiffꢁs bases formed by glutaraldehyde with the
free amino groups of the enzyme are apparently very
stable or cross-linking of the NHase CLEAs also
takes place through the formation of other (stable)
connections.[15]
Figure 3. Disappearance of acrolein during incubation in
buffer (0.01M citrate buffer, pH 5.5) with and without the
presence of NHase CLEAs, acrolein in buffer (^), acrolein
in buffer+3.7 mg NHase CLEA (&).
In view of the results discussed above, carrying out
a one-pot bienzymatic cascade using HnL and NHase
to synthesise a-hydroxyamides from aldehydes and
HCN is affected by several severe limitations. Be-
cause of the bad performance of the NHase CLEAs
in DIPE mixtures, the reactions have to be carried
out in an aqueous environment and, hence, the unca-
talysed hydrocyanation has to be suppressed by lower-
ing the pH of the reaction as much as possible. Be-
cause of the low HCN tolerance of the metalloen-
zyme, the HCN concentration in the reaction should
be kept low. A low HCN concentration will also have
the benefit of reducing the uncatalysed hydrocyana-
tion reaction rate and since the NHase will convert
the formed a-hydroxynitrile to the stable amide, the
equilibrium will shift to the right without the need of
an HCN excess. The preferred mode of carrying out
Figure 4. Rate of the uncatalysed hydrocyanation of acrolein
at different pH under cascade reaction conditions, pH 4
(^), pH 4.5 (&), pH 5 (~), pH 5.5 (&).
the cascade reactions was therefore the fed-batch ad- pH values (Figure 4). As expected, lowering the pH
dition of portions of HCN to modest aldehyde con- of the cascade reaction decreased the rate of the un-
centrations in acidic aqueous environment.
catalysed hydrocyanation.
The relative contribution of the background reac-
tion is also dependent on the rate of enzymatic hydro-
cyanation. An increase of HnL activity in the cascade
reaction would decrease the effect of the background
HnL-NHase One-Pot Bienzymatic Cascade
Because of the unfavourable reaction equilibrium and reaction on the ee of the intermediate nitrile product.
the low ees obtained in the MeHnL-catalysed hydro- However, there is a limit to the amount of HnL that
cyanation of acrolein,[11] this challenging aldehyde was can be added to the cascade reaction, since its cascade
chosen as the substrate for optimising the cascade re- partner should still be able to keep the intermediate
action conditions, since changes in reaction conditions nitrile concentration low to prevent chemical and/or
together with enzyme ratio were expected to have a enzymatic racemisation. The activity ratio of the two
more profound impact on the ee of the amide prod- enzymes in this cascade is therefore an important var-
uct.
As was previously mentioned, the uncatalysed
iable.
During the cascade reaction studies (Scheme 2), the
background hydrocyanation reaction should be sup- HnL activity added to the reaction mixture was fixed
pressed as much as possible, because it is unselective and the NHase activity was varied. The activity of the
and will deteriorate the ee of the a-hydroxynitrile in- HnL CLEA for acrolein under cascade reaction con-
termediate and consequently of the final amide prod- ditions at pH 4 was 0.98 mmol minÀ1 mgÀ1. During the
uct. In order to investigate the effect of the pH on the cascade reactions a fixed amount of 1.8–1.9 mgmLÀ1
background reaction in more detail, the latter was of HnL CLEA was used, which enabled reasonably
monitored for acrolein under cascade reaction condi- fast reactions without building up high concentrations
tions (5 mM HCN, 45 mM acrolein, 218C) at different of intermediate nitrile. Under these conditions the
400
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 397 – 404

Synthesis of Aliphatic (S)-a-Hydroxycarboxylic Amides
FULL PAPERS
Scheme 2. Synthesis of aliphatic (S)-a-hydroxyamides using
a bienzymatic one-pot cascade of the (S)-HnL from Manihot
esculenta and the NHase from Nitriliruptor alkaliphilus.
Figure 5. HnL-NHase one-pot bienzymatic cascade conver-
sion of acrolein at pH 4, 1a (^), 2a (&), 3a (~), HCN
spike (ꢂ).
Table 1. HnL-NHase one-pot bienzymatic cascade reactions
with 1a (45–46 mM) at different pH, 218C, and a maximum
HCN concentration of 5 mM.
pH Background
[%][a]
Factor[b] 3a produced
ee [%]
(S)
intermediate by the NHase CLEA is fast enough to
prevent enzymatic racemisation of 2a at this pH.
Complete conversion of 1a at pH 4.5 did not result
in a lower ee of 3a compared to the reaction where 1a
was not fully converted. This indicates that stopping
the reaction prematurely to maintain a favourable
equilibrium and to prevent enzymatic racemisation
was not necessary. The yield on 1a was between 77
and 89%, depending on the conditions and time
course of the reaction (Figure 5). The loss of 1a in the
reaction is probably caused by attachment of the sub-
strate to the enzyme, as was demonstrated before
(Figure 3). In these reactions, the formation of a-hy-
droxy-3-butenoic acid was not observed by HPLC,
GC, and NMR.
[mM]
5.5[c] 3.1
12
24
24
24
24
24
48
24.4
26.6
24.8
20.0
35.5
28.1
40.0
35
35
39
43
43
39
48
5.5[c] 3.1
5[c]
1.6
4.5[d] 0.5
4.5[e] 0.5
4[c]
0.1
4[e] 0.1
[a]
Relative background reaction when using 1.8–1.9 mg
HnL CLEA (T=218C).
Ratio of NHase activity to HnL activity. NHase activity
was measured for the hydration of 2a at pH 6 and HnL
activity was measured for the production of 2a under cas-
cade reaction conditions at pH 4.
Reaction was stopped after spiking 5 times with 5 mM
HCN.
Reaction was stopped after spiking 4 times with 5 mM
HCN.
[b]
[c]
[d]
[e]
The Manihot esculenta HnL is known to have a
rather low selectivity in converting acrolein (47–
56%).[11,16] In order to attempt the production of a-hy-
droxycarboxylic amides with higher enantiomeric
purity, the same cascade reaction was used for the
conversion of aliphatic aldehydes for which the HnL
Reaction was continued to the point were 1a was com-
pletely converted.
relative background reaction during the cascade reac-
tion of acrolein (1a) was 3.1% or less (Table 1).
Table 2. HnL-NHase one-pot bienzymatic cascade reactions
of 1b–1e (35–45 mM) at pH 4.5, 218C, and a maximum
HCN concentration of 5 mM. Unless stated otherwise, reac-
tions were stopped after full conversion of the aldehyde.
The pH of the cascade reaction clearly had an
effect on the final ee of 3a. This is to be expected,
since the competing unselective background hydro-
cyanation rate increases at a higher pH. The drop in
ee between pH 4.5 and the first experiment at pH 4 is
most probably the result of enzymatic racemisation of
2a, since moderate concentrations of 2a were detected
by HPLC at this pH (Figure 5).[11] Apparently, it was
not possible for the NHase to keep up with the HnL
in this situation. When the NHase activity at pH 4
was doubled, the nitrile intermediate was no longer
detected and the enantiomeric purity of the amide im-
proved. A two-fold increase of the relative NHase ac-
tivity in the cascade reaction at pH 5.5, in contrast to
pH 4, did not lead to an improvement in product ee,
indicating that the conversion of the a-hydroxynitrile
Compound 3 produced
Factor[a] ee [%] ee [%] (S)-
[mM]
(S)
2[16,17]
b
44
28
35
41
43
24
24
48
24
48
86
84
88
88
90
91
95
95
88
91
c[b]
c
d
e
[a]
Ratio of NHase activity to HnL activity. NHase activity
is measured for the hydration of 2a at pH 6 and HnL ac-
tivity is measured for the production of 2a under cascade
reaction conditions at pH 4.
Full conversion was not achieved with this amount of
NHase.
[b]
Adv. Synth. Catal. 2009, 351, 397 – 404
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
401

Sander van Pelt et al.
FULL PAPERS
from Manihot esculenta is known to have a higher se- Experimental Section
lectivity (Table 2).
The ee of the amide products significantly improved General Comments
when aldehydes were used for which the MeHnL is
Reactions and stability tests were carried out in Eppendorf
tubes using a ThermoTWISTER comfort shaker of QUAN-
TIFOIL Instruments. Freeze drying was performed by a
Christ Alpha 2–4 freeze dryer at a reduced pressure of
0.1 mbar and a temperature of À808C, after freezing the
samples with liquid N2. An Eppendorf 5415R centrifuge was
used for centrifugation at 13200 rpm. The temperature of
the centrifuge was adjusted according to the needs of the ex-
periment.
known to show a higher selectivity. In all cases,
except for d and e, the ee of the amide products is
lower than the ee of the nitriles as reported in litera-
ture for MeHnL catalysed hydrocyanations in micro-
ACHTUNGTRENNUNGaqueous systems using an excess of HCN. The higher
ees obtained for compounds 3d and 3e could be the
result of a low preference for the corresponding (S)-
a-hydroxy nitriles by the NHase. The enantioselective
properties of the NHase from Nitriliruptor alkaliphi-
lus are currently under investigation.
Materials
The yields of the a-hydroxy amide on the aldehyde
are >95%, which contrasts with the lower yields ob-
tained when acrolein is used as the substrate. Unlike
acrolein, the aliphatic aldehydes did not disappear
during the reaction because of possible attachment to
the enzyme. No traces of the corresponding a-hydroxy-
carboxylic acids were detected during these experi-
ments on HPLC and GC.
Semi-purified (S)-hydroxynitrile lyase from Manihot escu-
lenta was obtained from Jꢃlich Fine Chemicals (now Codex-
is). The NHase was produced by cultivation and induction
of the haloalkaliphilic actinobacterium Nitriliruptor alkali-
philus.[7] NHase and MeHnL CLEAs were prepared as de-
scribed previously.[8,11] NHase CLEAs were stored at +48C
in Tris-HCl buffer (0.01M, pH 8), while the MeHnL CLEAs
were stored at À208C in lyophilised form. Hexanenitrile
(98%, Aldrich), hexanamide (98%, Aldrich), sodium boro-
hydride (>96%, Fluka), diisopropyl ether (DIPE, >99%,
Fluka), acrolein (>95%, Fluka), propionaldehyde (99+%,
Janssen), isobutyraldehyde (99+%, Acros), butyraldehyde
(>99.5%, Aldrich), valeraldehyde (>97%, Fluka), a-hy-
droxybutyronitrile (Fluka, purum), a-hydroxybutyric acid
(Fluka, >97%), and a-hydroxycaproic acid (Aldrich, 98%)
were used in the experiments as received without additional
purification. A 2M solution of hydrogen cyanide in DIPE
was prepared from sodium cyanide (98+%, Acros) as de-
scribed previously.[18]
Conclusions
A bienzymatic one-pot cascade reaction to synthesise
aliphatic a-hydroxycarboxylic amides using MeHnL
and NHase CLEAs was made possible by using aque-
ous acidic citrate buffer of pH 4–pH 5.5 as reaction
medium containing moderate concentrations of alde-
hyde and the subsequent portionwise feed of HCN.
After optimisation of the pH together with the
HnL:NHase activity ratio in the reaction, five aliphat-
ic aldehydes were successfully converted into the a-
hydroxycarboxylic amides. Except for acrolein, the al-
dehydes were converted into the stable a-hydroxycar-
boxylic amides with good yield and high ee without
the need for organic solvent and an excess of HCN.
The ee of the a-hydroxycarboxylic amides as well
as the stability of the immobilised nitrile hydrating
biocatalyst can probably be further improved by the
controlled continuous addition of low concentrations
of aldehyde and pure HCN and by co-immobilisation
of the two biocatalysts in the form of a combi-CLEA.
An NHase with an inherent or designed increased sta-
bility in microaqueous systems could further increase
the performance of this system and the substrate
scope of the cascade reaction system could possibly
be broadened to aromatic aldehydes by using an
NHase with a higher activity for aromatic nitriles.
Warning: both sodium cyanide and HCN are highly poi-
sonous. They should always be handled in a fume cupboard
with a good draught. It is strongly advised to use a well-cali-
brated HCN detector during HCN work.
Analytical Procedures
HPLC analysis: The progress of the reactions was moni-
tored by HPLC using one or three 4.6ꢂ50 Merck Chromo-
lith SpeedROD RP-18e columns, depending on the polarity
of the compounds analysed. The eluent (MilliQ with 0.1 v%
TFA) was pumped through the column(s) at a flow rate of
1 mLminÀ1 using a Waters 590 HPLC pump. The HPLC
analyses were carried out at room temperature. All com-
pounds were detected using a Shimadzu RID 10A refractive
Table 3. Retention times on HPLC of the different com-
pounds in the cascade reaction.
Retention time [min]
Compound
Column #
1
2
3
a
b
c
d
e
3
3
3
3
1
5.2
4.8
9.4
8.9
7.6
4.4
4.6
12.2
10.9
11.6
3.2
3.5
5.3
5
4
402
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 397 – 404

Synthesis of Aliphatic (S)-a-Hydroxycarboxylic Amides
FULL PAPERS
Table 4. Retention times and temperature programs used for
separation of the (S)- and (R)-enantiomers of the formed
aliphatic a-hydroxycarboxylic amides.
3 times with Tris-HCl buffer (0.01M, pH 8) to remove all
traces of DIPE. The residual activities were determined by
the standard activity assay.
Acidic pH: NHase CLEAs were incubated in buffers with
different acidic pH values (pH 4–5.5: 0.01M citrate buffer)
and in MilliQ (pH ~6) at 218C. The residual activities after
different incubation times were determined by the standard
activity assay. Changing buffers for incubation was carried
out by washing the CLEAs 3 times with the corresponding
buffer.
Compound T program [oC] Rt (R) [min][a] Rt (S) [min][a]
3a
3b
3c
3d
3e
100
100
105
105
115
34
37
36
39
43
46
21[b]
39[c]
25[b]
36[c]
HCN: NHase CLEAs were washed 3 times with citrate
[a]
buffer (0.01M, pH 5.5). To
a mixture of this buffer
Since the MeHnL is (S)-selective, the peak with the larg-
est area was designated as the (S)-enantiomer.
This a-hydroxycarboxylic amide was first derivatised by
a one hour incubation in a mixture of acetic anhydride
(0.8 mL) and pyridine (0.8 mL) in dichloromethane
(10 mL).
(1000 mL) and the CLEAs, HCN was added (2.5, 5, 11, 21,
50, and 98 mM). These mixtures were then shaken at
900 rpm and 218C. After 30 and/or 90 min of incubation the
incubation mixture was centrifuged and the CLEAs washed
3 times with citrate buffer. The residual activity was deter-
mined by the standard activity test at pH 5.5. After the ac-
tivity test the CLEAs were washed 3 times with Tris-HCl
buffer (0.01M, pH 8) and stored at +48C in an open atmos-
phere. Activity of the CLEAs was reassessed after one week
of storage.
Acrolein and propionaldehyde: NHase CLEAs were
washed 3 times with citrate buffer (0.01M, pH 5.5). To a
mixture of this buffer (1000 mL) and the CLEAs, acrolein or
propionaldehyde were added (46, 131, and 269 mM). After
2, 3.5 and 18.5 h of incubation (900 rpm, 218C), the incuba-
tion mixture was centrifuged and the CLEAs washed 3
times with citrate buffer. The residual activities were deter-
mined by the standard activity test at pH 5.5.
[b]
[c]
The elution of the (S)- and (R)-enantiomers on this
column are in reversed order.
index detector and/or a Shimadzu SPD-10A VP UV-VIS de-
tector at a wavelength of 210 nm (Table 3).
GC analysis: Enantiomeric excess (ee) of the formed ali-
phatic a-hydroxycarboxylic amides 3a–d was determined
using a 25 m Chirasil dex CB column with i.d. 0.32 mm and
df 0.25 mm on a Shimadzu GC-17A gas chromatograph
equipped with a FID detector (Tinjector =2008C, Tdetector
=
2008C). The carrier gas used was N2 at a column flow rate
of 5.92 mLminÀ1. The (S)- and (R)-enantiomers were sepa-
rated using the temperature programmes in Table 4. The en-
antiomers of compound 3e were separated using a 50 m
Chiradex GTA column with i.d. 0.25 mm and df 0.12 mm on
a Shimadzu GC-2010 gas chromatograph equipped with an
FID detector (Tinjector =2008C, Tdetector =2008C). The carrier
gas used was N2 at a column flow rate of 5.00 mLminÀ1.
Cyanide concentration: Cyanide concentrations in the re-
actions and stability tests were determined by using the
Merck Spectroquantꢄ cyanide test kit for the determination
of free and readily liberated cyanide in water according to
the manual of the manufacturer. Cyanide stock solutions for
calibration were prepared by dissolving potassium cyanide
in a KOH solution (35 mM).
Preparation of the Racemic Amides
Racemic 2a was prepared according to a literature proce-
dure.[19] Racemic 2b (purum) was purchased at Fluka. Race-
mic 3a and 3b were prepared by adding racemic 2a and 2b
in fed-batch fashion to a suspension of NHase CLEAs in
MilliQ (218C, 900 rpm). After the hydration reaction was
finished the CLEA particles were spun down by centrifuga-
tion and the supernatant was removed. Freeze drying of the
obtained supernatant resulted in fluffy white amide crystals
of high purity. Racemic 2c–2e were produced by adding
equimolar amounts of 1c–1e and HCN to a suspension of
HnL CLEAs in citrate buffer (pH 4.5, 0.01M), after which
the reaction was allowed to proceed overnight. Racemic 3c–
3e were then prepared by the removal of the HnL CLEAs
by centrifugation, after which NHase CLEAs were added to
the supernatant of the enzymatic hydrocyanation reaction.
After the hydration reaction was completed, the NHase
CLEAs were removed by centrifugation and the superna-
tant was subsequently freeze dried to obtain the racemic
amide crystals.
Standard Activity Assay
NHase CLEAs were suspended in buffer (1000 mL, 0.01M
Tris-HCl of pH 8, unless stated otherwise) in an Eppendorf
tube and subsequently hexanenitrile was added (15 mL,
~120 mM). The reaction was allowed to proceed for 5 min
after which the CLEAs were centrifuged off in 10–20 sec-
onds. A sample (200 mL) was withdrawn from the superna-
tant and mixed with the same amount of HCl (1M) to make
sure that any accidentally transferred CLEA was deactivat-
ed. After centrifugation the supernatant was directly inject-
ed on HPLC.
HnL-NHase Cascade
In order to obtain better suspension behaviour, HnL
CLEAs were incubated in citrate buffer (0.01M) under re-
duced pressure to remove air from the very fluffy lyophi-
lised particles before the reaction. To a suspension of
NHase and HnL CLEAs with a certain activity ratio in cit-
rate buffer (0.01M, pH 4–6), aldehydes were added (30–
50 mM). At these concentrations the aldehydes were com-
pletely soluble in citrate buffer. After sampling for determi-
NHase CLEA Stability Tests
Two-phase DIPE-buffer system: NHase CLEAs were incu-
bated in mixtures of DIPE (50 v%, 90 v% and >99 v%)
and Tris-HCl buffer (0.01M, pH 8). After an incubation
time of one hour (900 rpm, 218C), the CLEAs were washed
Adv. Synth. Catal. 2009, 351, 397 – 404
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
403

Sander van Pelt et al.
FULL PAPERS
nation of the exact aldehyde starting concentration, the cas-
cade reaction was initiated by the addition of HCN (5 mM).
The reaction was spiked with portions of HCN (5 mM) sev-
eral times. The time intervals between the HCN spikes de-
pended on the speed of hydration of the a-hydroxynitrile in-
termediate. Sampling was carried out by stopping the reac-
tion with a 30 second centrifugation after which a sample
(50 mL) was withdrawn. The sample was directly injected on
the HPLC and the excess sample liquid was returned to the
reaction. At the end of the reaction the NHase and HnL
CLEAs were spun down by a 10 min centrifugation and sub-
sequently all supernatant was removed. After removing
excess HCN under vacuum, the supernatant was freeze
dried, after which white fluffy crystals were obtained. Cit-
rate, which is an impurity in the crystalline product, can be
removed by the addition of methanol in which the buffer
does not dissolve. After removing insoluble citrate by filtra-
tion, the enantiomeric excess of the final product was deter-
mined by chiral GC as discussed previously.
17, 320–323; b) S. Rustler, H. Motejadded, J. Alten-
buchner, A. Stolz, Appl. Microbiol. Biotechnol. 2008,
80, 87–97.
[6] R. DiCosimo, in: Biocatalysis in the Pharmaceutical
and Biotechnology Industries, (Ed.: R. N. Patel), CRC
Press, New York, 2007, pp 1–26.
[7] a) D. Y. Sorokin, S. van Pelt, T. P. Tourova, G. Muyzer,
Appl. Environ. Microbiol. 2007, 73, 5574–5579;
b) D. Y. Sorokin, S. van Pelt, T. P. Tourova, L. I. Ev-
tushenko, Int. J. Syst. Evol. Microbiol. 2008, in press.
ˇ
[8] S. van Pelt, S. Quignard, D. Kubac, D. Yu. Sorokin, F.
van Rantwijk, R. A. Sheldon, Green Chem. 2008, 10,
395–400.
[9] M. Schmidt, H. Griengl, Top. Curr. Chem. 1999, 200,
193–226.
[10] a) W. T. Loos, H. W. Geluk, M. M. A. Ruijken, C. G.
Kruse, J. Brussee, A. van der Gen, Biocatal. Biotrans-
form. 1995, 12, 255–266; b) S. Q. Han, G. Q. Lin, Z. Y.
Li, Tetrahedron: Asymmetry 1998, 9, 1835–1838.
[11] A. Chmura, G. M. van der Kraan, F. Kielar, L. M. van
Langen, F. van Rantwijk, R. A. Sheldon, Adv. Synth.
Catal. 2006, 348, 1655–1661.
[12] a) T. Nagasawa, A. Matsuyama, (Daicel), EU Patent
1266 962A2, 2002; Chem. Abstr. 2002, 138, 23774; b) S.
Osswald, C. Weckbecker, K. Huthmacher, T. Gerasi-
mova, A. Novikov, L. Ryabchenko, A. Yanenko, K.
Egorova, (Degussa), WO Patent 2005/090394A2, 2005;
Chem. Abstr. 2005, 143, 340641.
Acknowledgements
The authors wish to thank Andrzej Chmura (TU Delft) for
the preparation of the MeHnL CLEAs and Marion Livecchi
for carrying out the experiments on NHase CLEA reduction.
Financial support of the Delft Research Center of Life Scien-
ces and Technology (DRC-LST) is gratefully acknowledged.
[13] S. Mitra, R. C. Holz, J. Biol. Chem. 2007, 282, 7397–
7404.
[14] T. Gerasimova, A. Novikov, S. Osswald, A. Yanenko,
Eng. Life Sci. 2004, 4, 543–546.
References
[15] Y. Wine, N. Cohen-Hadar, A. Freeman, F. Frolow, Bio-
[1] H. Groger, Adv. Synth. Catal. 2001, 343, 547–558.
[2] D. J. C. Constable, P. J. Dunn, J. D. Hayler, G. R. Hum-
phrey, J. L. Leazer, R. J. Linderman, K. Lorenz, J.
Manley, B. A. Pearlman, A. Wells, A. Zaks, T. Y.
Zhang, Green Chem. 2007, 9, 411–420.
[3] R. J. H. Gregory, Chem. Rev. 1999, 99, 3649–3682.
[4] M. Breuer, K. Ditrich, T. Habicher, B. Hauer, M. Kess-
eler, R. Sturmer, T. Zelinski, Angew. Chem. 2004, 116,
806–843; Angew. Chem. Int. Ed. 2004, 43, 788–824.
[5] a) C. Mateo, A. Chmura, S. Rustler, F. van Rantwijk,
A. Stolz, R. A. Sheldon, Tetrahedron: Asymmetry 2006,
technol. Bioeng. 2007, 98, 711–718.
[16] S. Forster, J. Roos, F. Effenberger, H. Wajant, A. Spra-
uer, Angew. Chem. 1996, 108, 493–494; Angew. Chem.
Int. Ed. Engl. 1996, 35, 437–439.
[17] H. Wajant, S. Forster, H. Bottinger, F. Effenberger, K.
Pfizenmaier, Plant Sci. 1995, 108, 1–11.
[18] L. M. van Langen, F. van Rantwijk, R. A. Sheldon,
Org. Process Res. Dev. 2003, 7, 828–831.
[19] P. G. Gassman, J. J. Talley, Tetrahedron Lett. 1978,
3773–3776.
404
asc.wiley-vch.de
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 397 – 404