Optimisation of the enantioselective synthesis of cyanohydrin esters

Article abstract of DOI:10.1002/adsc.200505031

The base- and lipase-catalysed enantioselective synthesis of cyanohydrin esters was investigated, and the problem of previously reported low yields due to residual water in the reaction mixture was addressed. When the lipase was immobilised on Celite R-633 as a carrier, both the enantioselectivity and the reaction times for this dynamic kinetic resolution were improved, thus enabling a highly enantioselective synthesis of aromatic and heteroaromatic cyanohydrin acetates.

Full text of DOI:10.1002/adsc.200505031

FULL PAPERS
Optimisation of the Enantioselective Synthesis of Cyanohydrin
Esters
Lars Veum,a Liisa T. Kanerva,b Peter J. Halling,c Thomas Maschmeyer,d
Ulf Hanefelda,*
a
Gebouw voor Scheikunde, Technische Universiteit Delft, Julianalaan 136, 2628 BL Delft, The Netherlands
Fax: (þ31)-15-278-4289, e-mail: u.hanefeld@tnw.tudelft.nl
b
Laboratory of Synthetic Drug Chemistry, University of Turku, Lemminkꢀisenkatu 5 C, FIN-20520 Turku, Finland
Fax: (þ358)-2-333-7955, e-mail: lkanerva@utu.fi
c
Department of Pure and Applied Chemistry, University of Strathclyde, Glasgow G1 1XL, United Kingdom
Fax: (þ44)-141-548-4822, e-mail: p.j.halling@strath.ac.uk
d
Laboratory of Advanced Catalysis for Sustainability, School of Chemistry, Building F11, The University of Sydney, NSW,
2006, Australia
Fax: (þ61)-2-9351-3329, e-mail: th.maschmeyer@chem.usyd.edu.au
Received: January 21, 2005; Accepted: March 25, 2005
Abstract: The base- and lipase-catalysed enantioselec- were improved, thus enabling a highly enantioselec-
tive synthesis of cyanohydrin esters was investigated, tive synthesis of aromatic and heteroaromatic cyano-
and the problem of previously reported low yields hydrin acetates.
due to residual water in the reaction mixture was ad-
dressed. When the lipase was immobilised on Celite Keywords: asymmetric synthesis; Celite; dynamic ki-
R-633 as a carrier, both the enantioselectivity and netic resolution; enzyme catalysis; lipase B from Can-
the reaction times for this dynamic kinetic resolution dida antarctica
Introduction
dynamic kinetic resolution (DKR) of cyanohydrins
(Scheme 1).[1,10] While the kinetic resolution has a max-
Chiral cyanohydrins can be converted into a wide range imum yield of 50% only, and involves at least two sepa-
of compounds, which are versatile building blocks for rate steps to reach higher yields, the DKR starting from
the synthesis of fine chemicals, pharmaceuticals and a prochiral aldehyde yields theoretically 100% in just
agrochemicals, and they are therefore an important sub- one step.[11]
ject of research.[1–3] There are four main approaches for
The DKR combines a base-catalysed equilibrium be-
the synthesis of enantiopure cyanohydrins: either chem- tween an aldehyde, acetone cyanohydrin and the result-
ically, using cyclic dipeptides or transition metal com- ing cyanohydrin of the aldehyde, and a lipase-catalysed
plexes, or enzymatically, using hydroxynitrile lyases acylation of one enantiomer of this cyanohydrin. As the
(HNLs) or lipases.[1]
remaining enantiomer of the cyanohydrin is racemised
The enzyme-catalysed methods are of great interest by the base, one can theoretically obtain a 100% yield
due to their low costs and environmentally friendly as- of the acylated cyanohydrin. This reaction was already
pects. In the HNL-catalysed additions of HCN to alde- published in 1991,[12–14] but, in spite of its elegance, it
hydes and ketones, both excellent yields and enantio- has only been used successfully in a limited number of
meric purities were achieved within short reaction times cases.[15–19] This might be due to the fact that hydroxyni-
(<1 h).[4–9] However, the reaction is mostly performed trile lyase-catalysed cyanohydrin synthesis tends to be
in a two-phase system, so water-sensitive and water-in- faster and give higher yields and enantiomeric purities
soluble substrates can only be used with difficulties. for most of the substrates that have been studied.[4]
Since the reaction is an equilibrium reaction, a large ex- But this could also be caused by the unsuccessful exam-
cess of HCN or of the less toxic acetone cyanohydrin is ples of the DKR that have been reported.[20–22]
needed to obtain high conversions. Furthermore, not
It has previously been shown that one of the most
all substrates are accepted by the HNLs and some of widely applied lipases in synthesis, Candida antarctica li-
the resulting cyanohydrins racemise very easily. As pase B immobilised on a methacrylate polymer (Novo-
an alternative to the HNLs, one can use lipases. zyme 435, or Chirazyme L-2, cf. C2, Lyo; both abbrevia-
They have been applied both in the kinetic and in the tions stand for the same enzyme preparation), is par-
Adv. Synth. Catal. 2005, 347, 1015–1021
DOI: 10.1002/adsc.200505031
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1015

Lars Veum et al.
FULL PAPERS
ticularly enantioselective in the kinetic resolution of cy- ious amounts of base and Novozyme 435 in 4 mL of tol-
anohydrins.[22–24] In order to increase the utility of the uene at 40 0C. In our previous work we could show that
DKR, we recently investigated the reaction. Starting 380 U of Novozyme 435 were sufficient for the kinetic
from 1a, using Novozyme 435, basic Amberlite was resolution of 2a to proceed to 50% conversion within
used as the racemisation catalyst.[22] The racemisation 4 h and with an excellent E (>100). Therefore 380 U/
of mandelonitrile (2a), the transcyanation between the mmol substrate were employed for the reaction[22]
acetone cyanohydrin and the aldehyde, as well as the ki- (Scheme 1).
netic resolution of mandelonitrile via acylation, all
Solid buffers have previously been used to adjust the
worked well separately. However, when the three reac- ionisation state of enzymes in organic media, and to
tions were combined, the reaction hardly even went to maintain it.[25] It was therefore investigated whether
16% conversion in the DKR. It was concluded that the these solid buffers could replace the traditionally used
water present in the system hydrolysed the acylating Amberlite (Amberlite IRA-904 in OHꢀ form). Since
agent, yielding an acid, which in turn neutralised the fast racemisation of the non-acylated enantiomer of 2
base that catalyses the dynamic equilibrium. Eventually, is essential for the DKR (to reach 100% conversion
the acid might even render the enzyme inactive. HCN with a high enantioselectivity), we tested various solid
formed during the reaction could have an equivalent ef- buffers (MOPS, HEPES, HEPPSO, AMPSO and
fect. If simply more base was added, another problem CAPS) as catalysts for the dynamic equilibrium
arose, namely the base-catalysed polymerisation of (Scheme 2, Table 1). Only CAPS gave results that
HCN. These polymers inhibited the enzyme and led to were similar to Amberlite, but CAPSO too catalysed
a full stop in the reaction.
this reaction within a reasonable time frame. Based on
these results, CAPS, CAPSO and Amberlite were used
for our further studies. The active catalyst in the buffer
pair is probably the solid Na salt, which has a free basic
amino group.
Results and Discussion
In this work we now look at five different possibilities to
In order to find the optimal quantity of base to use in
solve this problem. Solid buffers have previously proven the DKR starting from 1a and 1b, various amounts of
to be useful in establishing and maintaining optimum CAPSO, CAPS (40, 60, 80, 100, 120 mg of each salt of
acid-base conditions for enzymes in organic media.[25] the buffer pair), Amberlite (20, 40, 100, 200, 300 mg,
We use the shorthand “pH” to refer to this, although 1 mmol OHꢀ/g Amberlite) and 380 U of Novozyme
in the absence of a liquid water phase, simple pH is
not the appropriate parameter. If solid buffers are
used in the DKR, they should help in ensuring a favour-
able “pH”, buffering any acetic acid formed. Another
option is to use cyanide salts as bases. They should also
neutralise any acetic acid formed in the reaction, yield-
ing HCN that in turn should add to the aldehyde. In
the first description of the DKR, molecular sieves
were used to dry the reaction.[12–14] We look at this pos-
sibility, too, even though molecular sieves also work as
ion exchangers, and therefore may alter the “pH” of
the reaction in an unfavourable direction.[26] The carrier
of the enzyme may also be of importance. As Celite has
been used as the carrier in all the successful applications
Scheme 1. Enantioselective synthesis of cyanohydrin esters
of the DKR, it is also of interest to compare Novozyme via a dynamic kinetic resolution.
435 with CAL-B immobilised on Celite. In addition it is
known that Celite R-640 can be used to control water ac-
tivities in organic media.[27] It can therefore be expected
that the Celite used for the immobilisation has this effect
on the reaction, too. We chose Celite R-633 in this study
since we have applied it successfully in the immobilisa-
tion of HbHNL before.[28] The acylating agent is known
to have an effect on both the enantioselectivity and the
activity of the enzyme and the possibility of using other
acylating agents was explored.[10]
Based on the first DKR reported,[12–14] 1 mmol of al-
dehyde, 2 mmol of acetone cyanohydrin, and 3 mmol
of acetylating agent were used in combination with var- nohydrin, HCN and acetone and between 1a, HCN and 2a.
Scheme 2. Base-catalysed equilibrium between acetone cya-
1016
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1015–1021

Enantioselective Synthesis of Cyanohydrin Esters
FULL PAPERS
Table 1. pKa and time to reach the base-catalysed equili-
brium between 1a, acetone cyanohydrin, 2a and acetone.
Even though the two gave comparable results, the
OHꢀ form was chosen because it gave a slightly higher
ee. Instead of using CNꢀ-conditioned Amberlite, we ex-
plored the possibility of employing different cyanide
salts. The idea is that any acetic acid formed as a side
product in the reaction will be neutralised by the cyanide
salt. The HCN formed will then add to the aldehyde
while the metal salt of the acetate precipitates. In order
to pair the soft cyanide anion with a hard, intermediate
and soft acid, we chose for the salts NaCN, Zn(CN)2 and
CuCN, respectively, and used these as bases/salts in par-
allel with Amberlite OHꢀ and NaOAc in the DKR start-
ing from 1a and 1b (Table 2). The composition of re-
agents remained unchanged; i.e., 2 equivalents of ace-
tone cyanohydrin were added.
[a]
Solid buffer
pKa
Teq
MOPS
7.2
7.5
7.8
no reaction
>30 h
no reaction
24 h
HEPES
HEPPSO
TAPS
8.4
AMPSO
CAPSO
CAPS
9.0
9.6
10.4
30 h
5 h
10 min
20 min
Amberlite
[a]
Time to reach equilibrium with 100 mg of buffer.
435 per mmol substrate were used. The reactions were
analysed after 7 days.
As expected there is a correlation between the hard-
ness of the acid and the enantioselectivity. According
In the case of CAPSO, an optimal conversion is to the hard-soft acid-base principle, the bond in CuCN
reached for both 1a (conversion 100%, ee 91%) and 1b has a covalent character, NaCN is completely ionic
(conversion 66%, ee 74%) when 100 mg of each salt of while Zn(CN)2 is an intermediate between the two. Con-
the pair are used. With CAPS the maximum conversions sequently, the CNꢀ is more “available” as a base in
were already achieved when only 60 mg of each salt of NaCN than in CuCN. As predicted this trend is evident
the pair were used (1a: conversion 98%, ee 91% and in the results. NaCN catalyses the reaction, but in the
1b: conversion 57%, ee 26%). However, for 1b the con- case of the more reactive substrates the ee is low due
versions and ees were slightly lower than with CAPSO. to the base-catalysed acylation of the cyanohydrin or
Amberlite is a strong base and, in the case of 1a and 1b, the base-catalysed racemisation of the product. CuCN
the conversion drops slightly when more base is added. is a poor catalyst, and it only works in the case of
This is probably due to polymerisation of HCN and a 1b. However, as 1b is a base in itself, the observed con-
consequential deactivation of the enzyme.
version might also be autocatalysed. Zn(CN)2 gives by
For 1a, the ee is generally only slightly reduced if more far the best ees but only a moderate conversion. Both
base is used in the reactions. But in the case where the NaOAc and Amberlite gave higher conversions than
amount of Amberlite OHꢀ utilised in combination Zn(CN)2, but, the ees were lower.
with 1b increases from 40 mg to 200 mg, the ee for 3b de-
These experiments were repeated in the presence of
creases from 70% to 20%. This might be due to the fact molecular sieves as drying agents. As expected there
that 1b is also a base in itself, aiding the chemical back- was no significant difference between the two sets of ex-
ground reaction or the base-catalysed racemisation of periments, whether molecular sieves are used or not.
the product.
Any acid is neutralised by the cyanide salt to form
All the experiments with the solid buffers were re- HCN and the corresponding salt of the acid. As long
peated using molecular sieves to dry the reaction mix- as there is cyanide salt present, the “pH” of the reaction
ture. However, independent of how much buffer or Am- should be constant and the molecular sieves should not
berlite was used, the molecular sieves did not influence act as ion exchangers.
the reaction significantly: neither the conversion nor the
ee obtained showed much deviation from the results ob- cyanohydrins, the lipases were immobilised on Celite.
tained without molecular sieves. Celites are natural silicates, and some of them can ad-
In all the successful, albeit sometimes slow, DKR of
In the original paper describing the DKR, both Am- sorb large amounts of water. They bind this water tightly
berlite conditioned with NaOH and NaCN were used. via hydrogen bridges; indeed Celite R-640 can be used to
Table 2. The use of salts as bases in the DKR starting from 0.98 mmol 1a and 1b, using 380 U/mmol substrate of Novozyme
435, after 6 days.
[a, b]
Compound
NaCN[a, b]
Zn(CN)2
CuCN[a, b]
NaOAc[a.b]
Amberlite[a, c]
1a
1b
90 (36)
96 (0)
24 (94)
25 (85)
0 (ꢀ)
41 (89)
32 (35)
55 (90)
63 (57)
3 (91)
[a]
Conversion (ee) [%].
1 equivalent.
0.3 equivalents.
[b]
[c]
Adv. Synth. Catal. 2005, 347, 1015–1021
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1017

Lars Veum et al.
FULL PAPERS
efficiently control low water activities in organic sol- with various amounts of base (CAPSO, CAPS) (Ta-
vents, since it can adsorb more than 90% of the Celiteꢂs ble 4).
weight.[27] In comparison, Novozyme 435 is adsorbed on
Although CAPSO gave fairly high ees, a substantial
a divinylbenzene-cross-linked, hydrophobic macropo- amount of buffer is necessary to obtain the same conver-
rous polymer based on methyl and butyl methacrylic es- sion as with CAPS. The conversion that is obtained with
ters.[29] This lipophilic material will readily release any CAPS is not influenced by how much base is used, how-
water that is attached to it into the dry reaction mixture, ever, the ee drops as a function of the amount of base
thereby enabling the hydrolysis of both the product and added. These differences can be rationalised by the
the acyl donor. Not all types of Celite are suitable for en- base strength of the solid buffers. More CAPSO is need-
zyme immobilisations and care has been taken to choose ed for the racemisation since it is the weaker base, while
a suitable one. We immobilised CAL-B on Celite Bio- larger amounts of the stronger base CAPS lead to prod-
catalyst carrier R-633 according to standard proce- uct racemisation. When performing the same reaction
dures[30] and tested it with substrates 1a, 1c, 1d and the using 30 mg of Amberlite, 83% conversion and an ee
bases NaCN, Zn(CN)2, NaOAc and Amberlite. The re- of 97% were obtained. This is better than what was ob-
sults obtained after 6 days are shown in Table 3.
tained with both CAPSO and CAPS. These results can-
The observed trend is the same as when Novozyme not be directly compared with those obtained with No-
435 is used. However, if one considers both conversion vozyme 435 since less units of the enzyme were em-
and enantiomeric excess, Amberlite is clearly the best ployed in those experiments.
base to use. The result we obtained for 1c is a significant
However, it is clearthat Amberliteincombinationwith
improvement over what has earlier been reported. In CAL-B on Celite R-633 not only gives good yields but
the original work on the DKR, a conversion of 73% of also an excellent enantioselectivity. In order to directly
1c to its corresponding cyanohydrin acetate 3c was ach- compare the two different lipase immobilisations, we
ieved with an ee of 47% in 6 days using Amberlite as the tested various amounts of both the CAL-B on Celite R-
base. The results that are reported here for 1a are also an 633 and Novozyme 435 in the synthesis of 3a (Figure 1).
improvement compared to those obtained earlier with
Novozyme 435, both in terms of conversion and enantio- amount of activity is used, the CAL-B adsorbed on Cel-
purity. iteR-633performs significantly better, both in respect to
The results clearly demonstrate that even if the same
In order to see how CAL-B on Celite R-633 (910 KU/ yield and enantioselectivity. In the case of CAL-B on
mmol substrate) performed in combination with the sol- Celite R-633 the conversion increases with an increasing
id buffers, it was used for the synthesis of 3a, together amount of enzyme, until a plateau is reached, indicating
that the enzymatic reaction is no longer the rate-limiting
step. The slight drop in ee for 3a is consistent with this,
indicating that the dynamic equilibrium now is the
rate-limiting step.[31] It also shows that the 380 U, which
were enough for the kinetic resolution of 2a to proceed
rapidly, are not enough for the DKR to proceed smooth-
ly: at least 1.3 KU/mmol substrate have to be used. To
ensure that it is really the effect of the carrier that causes
the difference described in Figure 1, and not the immo-
Table 3. The use of salts as bases in the DKR starting from
1a, 1c and 1d using CAL-B on Celite R-633 (380 U), after
6 days.
[a, b]
Compound NaCN[a, b] Zn(CN)2
Amberlite[a, c] NaOAc[a, b]
1a
1c
1d
–
18(92)
23(75)
10(81)
87(95)
97(86)
71(79)
56(91)
71(84)
28(74)
69(4)
76(11)
[a]
[b]
[c]
Conversion (ee) [%].
1 equivalent.
0.3 equivalents.
Table 4. The use of various amounts of CAPSO and CAPS as
bases in the DKR starting from 1a using CAL-B on Celite R-
633 (910 U/mmol substrate) with a reaction time of 48 hours.
Mass acid/base [mg] CAPSO[a] CAPS[a] Amberlite[a, b]
-/30
–
–
83 (97)
40/40
60/60
80/80
100/100
48 (95)
52 (95)
78 (91)
86 (92)
87 (91)
85 (82)
85 (79)
86 (77)
–
–
–
–
Figure 1. The yield of 3a after 3 days using Amberlite OHꢀ as
the base, isopropenyl acetate as the acylating agent, and var-
[a]
&
&
Conversion (ee) [%].
ious amounts of Novozyme 435 (conversion , ee ) were
[b]
*
*
0.3 equivalents.
tested vs. CAL-B on Celite R-633 (conversion , ee ).
1018
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1015–1021

Enantioselective Synthesis of Cyanohydrin Esters
FULL PAPERS
bilisation procedure, we stirred the Novozyme 435 in the conversions and long reaction times due to hydrolysis
same buffer-sucrose solution that was used for the im- of the acylating agents, have been addressed and solved.
mobilisation of CAL-B on Celite R-633 and dried it in
Varying the nature and the amount of base used in the
the same way. The treated Novozyme 435 showed no dif- DKR of cyanohydrins already gave improvements com-
ference compared to the untreated Novozyme 435 prep- pared to previous work using Novozyme 435, although a
aration.
straightforward change in the carrier of the enzyme had
Even though vinyl acetate and vinyl butanoate give the largest effect on the reaction. Whereas Novozyme
acetaldehyde as a side product, which can in general 435 probably releases water from the carrier into the re-
be harmful for enzymes,[10] and can act as a substrate action media, causing hydrolysis which subsequently
for the DKR, they have still been the acylating agents stopped the reaction, Celite R-633 suppressed the nega-
of choice in several of the reported successful applica- tive side effects of water most likely by binding it. With
tions of the DKR. We tested isopropenyl acetate, vinyl lipase B from Candida antarctica adsorbed on Celite R-
acetate and vinyl butanoate with CAL-B adsorbed on 633, both the conversion and the ee was improved signif-
Celite R-633, and, after 16 hours the conversions and icantly when compared to earlier results.
ees were 35 (98), 13 (83) and 14 (92) respectively. Since
CAL-B in combination with isopropenyl acetate gives
both the best conversion and enantiomeric excess, all
further studies were performed with this acylatingagent.
The reaction was performed with 1350 U/mmol CAL-
B on Celite R-633 on a preparative scale, starting from
Experimental Section
1a and 1c; using Amberlite as the base and isopropenyl
acetate as the acylating agent (Table 5).
General Remarks
1H and 13C NMR spectra were recorded on a Varian Unity In-
ova 300 (300 MHz and 75, MHz, respectively), instrument.
Chemical shifts are expressed in parts per million (d) relative
to tetramethylsilane. Abbreviations are as follows: s (singlet),
d (doublet), t (triplet), q (quartet) and m (multiplet). Optical
rotations were obtained using a PerkinElmer 241 polarimeter.
Column chromatography was carried out with silica gel 0.060–
0.200 mm, pore diameter ca. 6 nm. TLC was performed on
0.20 mm silica gel and developed in a vanillin bath [vanillin
(15 g) in ethanol (250 ml)þconcentrated H2SO4 (2.5 mL)]
where all products containing a CN group gave orange spots.
Dry toluene was purchased from Aldrich. Lyophilised lipase
B from Candida antarctica was purchased from Roche. Immo-
bilised lipase B from Candida antarctica (CALB, Novozyme)
was a generous gift from Novo Nordisk (Dr. Deussen). Lyophi-
lised CAL-B was adsorbed on Celite Bio-Catalyst Carrier R-
633 from World Minerals.[30] In the original procedure Celiteꢃ
(Filter agent) from Aldrich was used. The activity of the en-
zymes used was determined as described earlier.[24] Amberlite
IRA-904 was conditioned with NaOH,[13] and the basicity
of the conditioned ion exchanger was found to be 1 mmol
OHꢀ/g. All aldehydes, acetone cyanohydrin and isopropenyl
acetate were distilled prior to use and stored under nitrogen.
The racemic cyanohydrin acetates 3a–f were prepared accord-
ing to a standard procedure[33] and their spectroscopic data cor-
respond to those in the literature.[14,24,34–36] All reactions were
performed in a 10-mL glass vial equipped with a silicone sup-
ported teflon septum and a screw cap. 1,3,5-Triisopropylben-
zene was used as an internal standard in the reactions.
Table 5. Results of the DKR starting from 1a and 1c using
optimised reaction conditions.
Substrate
Time [days]
Yield [%]
ee [%]
1a
1c
4.0
4.5
97
92
98
89
For both 1a and 1c the preparative reaction proceeded
within four days, which is faster than described earlier (4
and 4.5 days rather than 6 days for both 3a and 3c, re-
spectively). More importantly, the enantioselectivity
for both compounds was significantly improved. For 1c
the yield is increased from 57% to 92%, and the enantio-
meric excess is increased from 47% to 89%. For 1a, it is
now an almost enantiospecific reaction (ee¼98%), in-
stead of an enantioselective reaction (ee¼84%).[13]
When comparing it to the vanadium-salen-catalysed for-
mation of 3a from 1a, KCN and acetic anhydride it also is
a significant step forward. Although the vanadium-cata-
lysed reaction proceeds faster (10 h), the yield (88%)
and the enantioselectivity (ee¼90%) are lower.[32]
Work on the DKR starting from aliphatic aldehydes is
in progress. Preliminary results indicate that the system
that is described in this paper is not the best solution.
The optimisation for these substrates will be reported
elsewhere.
HPLC Analysis
Conclusion
The HPLC analysis were performed using a 4.6ꢁ250 mm Chir-
acel OB-H column with a Waters 510 pump, and a Waters 486
UV detector. The eluent was a 90:10 mixture of hexane and 2-
propanol containing 0.1% acetic acid. The flow was
This work describes the synthesis of cyanohydrins via a
DKR using the readily available lipase B from Candida
antarctica. The problems experienced earlier, of low 0.8 mL minꢀ1
.
Adv. Synth. Catal. 2005, 347, 1015–1021
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1019

Lars Veum et al.
FULL PAPERS
Table 6. Temperature programs and retention times for 1a–d, (R)-3a–d and (S)-3a–d.
Substrate
Temp. [8C][a]
Rt [min] 1
Rt [min] (R)-3
Rt [min] (S)-3
a
b
c
145
145
120
120
1.00
0.99
1.00
0.99
2.62
3.89
3.04
2.96
3.12
5.05
3.83
3.91
d
[a]
All analyses were performed isothermally.
as a clear oil; yield: 798 mg (97%); 98% ee; [a]2D0: ꢀ6.2 (c 1,
GC Analysis
1
CHCl3); H NMR (300 MHz, CDCl3): d¼2.14 (s, 3H, CH3
The conversion and enantiomeric purity were determined by
chiral GC using a b-cyclodextrin column (CP-Chirasil-Dex
CB 25 mꢁ0.25 mm) using a Shimadzu Gas Chromatograph
GC-17A equipped with a FID detector and a Shimadzu
Auto-injector AOC-20i, using He with a linear gas velocity of
75 cm/s as the carrier gas. The temperature programs and re-
tention times are given in Table 6.
CO), 6.40 (s, 1H, CH-O), 7.20–7.58 (m, 5H, aromatic);
13C NMR (75 MHz, CDCl3): d¼20.5 (CH3CO), 62.9 (CH-O),
116.2 (CN), 127.9 (C-2,6), 129.3 (C-3,5), 130.4 (C4), 131.8
¼
(C-1), 169.9 (C O).
(S)-(ꢀ)-2-Acetoxy-2-(2-furyl)ethanenitrile [(S)-3c]
The title compound was prepared from furfural (452 mg,
4.7 mmol), following general procedure B. (S)-3d was isolated
as a clear oil; yield: 718 mg (92%); 89%; [a]2D5: þ22.4 (c 1.0
Base-Catalysed Transcyanation from Acetone
Cyanohydrin to Benzaldehyde
1
CHCl3); H NMR (300MmHz, CDCl3): d¼2.14 (s, 3H, CH3
¼
Acetone cyanohydrin (0.17 ml, 1.84 mmol) was added to a stir-
red mixture of benzaldehyde (98 mg, 0.92 mmol) and either the
solid buffer (50 mg of both the acid and its corresponding so-
dium salt) or Amberlite (30 mg), in toluene (4 mL) at 40 8C.
Samples (10 mL) were diluted in hexane and analysed by
HPLC.
C O), 6.47 (m, 2H, CH-CN and C3-H), 6.69 (m, 1H, C2-H),
7.51 (m, C4-H); 13C NMR (75 MHz, CDCl3): d¼20.2 (CH3),
55.8 (CH-O), 111.2 and 112.7 (C-2 or C-3), 114.3 (CN), 144.2
¼
(C-1), 145.1 (C-4), 168.8 (C O).
Acknowledgements
General Procedure A: The DKR of Cyanohydrins,
Analytical Scale
U. H. thanks the Royal Netherlands Academy of Arts and Scien-
ces (KNAW) for a Fellowship. The authors gratefully acknowl-
edge COST D25 for enabling their collaboration.
Acetone cyanohydrin (0.17 mL, 1.84 mmol) was added to a
stirred mixture of the aldehyde (0.98 mmol), the acylating
agent (2.76 mmol), the immobilised enzyme (as stated in the
text) and the base (as stated in the text) in dry toluene
(4 mL). The reaction was stirred magnetically at 40 8C. Sam-
ples (10 mL) were diluted in acetone and centrifuged. The su-
pernatant was analysed by chiral GC.
References and Notes
[1] M. North, Tetrahedron: Asymmetry 2003, 14, 147–176.
[2] H. Griengl, H. Schwab, M. Fechter, Trends Biotechnol.
2000, 18, 252–256.
[3] J. Brussee, A. van der Gen, in: Stereoselective Biocataly-
sis, (Ed.: P. N. Ramesh), Marcel Dekker, Inc., New York,
2000, pp. 289–320.
[4] R. J. H. Gregory, Chem. Rev. 1999, 99, 3649–3682.
[5] H. Griengl, A. Hickel, D. V. Johnson, C. Kratky, M.
Schmidt, H. Schwab, Chem. Commun. 1997, 1933–1949.
[6] F. Effenberger, Chimia 1999, 53, 3–10.
[7] M. Schmidt, H. Griengl, Top. Curr. Chem. 1999, 200,
193–226.
[8] G. Seoane, Curr. Org. Chem. 2000, 4, 283–304.
[9] F. Effenberger, S. Fçrster, H. Wajant, Curr. Opin. Bio-
technol. 2000, 11, 532–539.
General Procedure B: The DKR of Cyanohydrins,
Preparative Scale
Acetone cyanohydrin (0.86 mL, 9.4 mmol) was added to a stir-
red mixture of the aldehyde (4.7 mmol), isopropenyl acetate
(1.62 ml, 14.9 mmol), CAL-B on Celite R-633 (6.83 KU) and
Amberlite IRA-904 in OHꢀ form (150 mg) in dry toluene
(20 mL). When the reaction was completed, the enzyme and
the resin was filtered off and washed with toluene (2ꢁ
10 mL). The solvents were removed under vacuum and the res-
idue was purified by column chromatography on silica gel (PE/
EtOAc, 90:10).
[10] U. Hanefeld, Org. Biomol. Chem. 2003, 1, 2405–2415.
[11] J. Sukumaran, U. Hanefeld, Chem. Soc. Rev. 2005, 34,
DOI: 10.1039/b412490a, available on the web.
[12] M. Inagaki, J. Hiratake, T. Nishioka, J. Oda, J. Am.
Chem. Soc. 1991, 113, 9360–9361.
(S)-(ꢀ)-2-Acetoxy-2-phenylethanenitrile [(S)-3a]
The title compound was prepared from benzaldehyde (499 mg,
4.7 mmol), following general procedure B. (S)-3a was isolated
1020
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 1015–1021

Enantioselective Synthesis of Cyanohydrin Esters
FULL PAPERS
[13] M. Inagaki, J. Hiratake, T. Nishioka, J. Oda, J. Org.
Chem. 1992, 57, 5643–5649.
[25] E. Zacharis, B. D. Moore, P. J. Halling, J. Am. Chem. Soc.
1997, 119, 12396–12397.
[14] M. Inagaki, A. Hatanaka, M. Mimura, J. Hiratake, T.
Nishioka, J. Oda, Bull. Chem. Soc. Jpn. 1992, 65, 111–
120.
[15] L. T. Kanerva, K. Rahiala, O. Sundholm, Biocatalysis
1994, 10, 169–180.
[16] C. Paizs, M. Tosa, C. Majdik, P. Tꢀhtinen, F. D. Irimie,
L. T. Kanerva, Tetrahedron: Asymmetry 2003, 14, 619–
627.
[17] C. Paizs, P. Tꢀhtinen, K. Lundell, L. Poppe, F. D. Irimie,
L. T. Kanerva, Tetrahedron: Asymmetry 2003, 14, 1895–
1904.
[26] N. Fontes, J. Partridge, P. J. Halling, S. Barreiros, Biotech-
nol. Bioeng. 2002, 77, 296–305.
[27] A. Basso, L. De Martin, C. Ebert, L. Gardossi, P. Linda,
J. Mol. Catal. B Enzym. 2000, 8, 245–253.
[28] U. Hanefeld, A. J. J. Straathof, J. Heijnen, J. Mol. Cat-
al. B. Enzym. 2001, 11, 213–218.
¨
[29] N. W. J. T. Heinsman, C. G. P. H. Schroen, A. van der
Padt, M. C. R. Franssen, R. M. Boom, K. vanꢂt Riet, Tet-
rahedron: Asymmetry 2003, 14, 2699–2704.
[30] L. T. Kanerva, O. Sundholm, J. Chem. Soc. Perkin Trans.
1 1993, 2407–2410.
[18] C. Paizs, P. Tꢀhtinen, M. Tos¸a, C. Majdik, F.-D. Irime,
L. T. Kanerva, Tetrahedron 2004, 60, 10533–10540.
[19] L. Veum, U. Hanefeld, Tetrahedron: Asymmetry 2004, 15,
3707–3709.
[20] Y. Zhu, L.-R. Yang, Z.-Q. Yao, S. Zhu, P. Cen, Ann. New
York Acad. Sci. 1998, 864, 646–648.
[31] M. Kitamura, M. Tokunaga, R. Noyori, Tetrahedron
1993, 49, 1853–1860.
[32] Y. N. Belekon, P. Carta, A. V. Gutnov, V. Maleev, M. A.
Moskalenko, L. V. Yashkina, N. S. Ikonnikov, N. V. Vos-
koboev, V. N. Khrustalev, M. North, Helv. Chim. Acta
2002, 85, 3301–3312.
[21] T. Zhang, L. Yang, Z. Zhu, J. Wu, J. Mol. Catal. B En-
zym. 2002, 18, 315–323.
[33] A. Fishman, M. Zviely, Tetrahedron: Asymmetry 1998, 9,
107–118.
[22] Y.-X. Li, A. J. J. Straathof, U. Hanefeld, Tetrahedron:
Asymmetry 2002, 13, 739–743.
[23] U. Hanefeld, Y. Li, R. A. Sheldon, T. Maschmeyer, Syn-
lett 2000, 12, 1775–1776.
[34] H. M. R. Hoffmann, Z. M. Ismail, R. Hollweg, A. R.
Zein, Bull. Chem. Soc. Jpn. 1990, 63, 1807–1810.
[35] M. Schmidt, S. Herve, N. Klempier, H. Griengl, Tetrahe-
dron 1996, 52, 7833–7840.
[24] L. Veum, M. Kuster, S. Telalovic, U. Hanefeld, T. Masch-
meyer, Eur. J. Org. Chem. 2002, 1516–1522.
[36] M. Scholl, C. K. Lim, G. C. Fu, J. Org. Chem. 1995, 60,
6229–623.
Adv. Synth. Catal. 2005, 347, 1015–1021
asc.wiley-vch.de
ꢁ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1021