Enantioselective ring opening of epoxides with cyanide catalysed by halohydrin dehalogenases: A new approach to non-racemic β-hydroxy nitriles

Article abstract of DOI:10.1002/adsc.200505333

Halohydrin dehalogenases (HheA, HheB and HheC) were found to efficiently catalyse a carbon-carbon bond forming reaction between terminal aliphatic epoxides and cyanide, yielding β-hydroxy nitriles. With all three enzymes nucleophilic ring opening of epoxi

Full text of DOI:10.1002/adsc.200505333

FULL PAPERS
DOI: 10.1002/adsc.200505333
Enantioselective Ring Opening of Epoxides with Cyanide
Catalysed by Halohydrin Dehalogenases: A New Approach to
Non-Racemic b-Hydroxy Nitriles
a
b
a,
´
Maja Majeric Elenkov, Bernhard Hauer, Dick B. Janssen *
a
Biochemical Laboratory, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen,
Nijenborgh 4, 9747 AG, Groningen, The Netherlands
Fax: (þ31)-50-363-4165, e-mail: d.b.janssen@rug.nl
b
BASF AG, Fine Chemicals and Biocatalysis Research, 67056 Ludwigshafen, Germany
Received: August 31, 2005; Accepted: December 28, 2005
Abstract: Halohydrin dehalogenases (HheA, HheB varies from moderate to high (E¼5–106), while res-
and HheC) were found to efficiently catalyse a car- olution of 2,2-disubstituted epoxides proceeds with
bon-carbon bond formingreaction between terminal very high enantioselectivity (E¼141 and 200). The re-
aliphatic epoxides and cyanide, yielding b-hydroxy ni- sults show that halohydrin dehalogenases may be-
triles. With all three enzymes nucleophilic ringopen- come attractive catalysts for the facile preparation
ing of epoxides proceeds with high regioselectivity to of enantiopure b-hydroxy nitriles from racemic epox-
the b-carbon atom. Activity, enantioselectivity and ides.
enantiopreference depend on the type of enzyme
and the substrate structure. HheC was found to be
the most selective amongthe tested enzymes. The
Keywords: cyanide; epoxides; halohydrin dehaloge-
enantioselectivity toward monosubstituted epoxides nase; kinetic resolution
Introduction
alcohols,[5] b-nitro alcohols[6] and b-hydroxy nitriles[7,8]
were prepared.
Halohydrin dehalogenases are enzymes involved in the
Amongthese nucleophiles cyanide is particularly in-
biodegradation of some halogenated organic com- terestingbecause of its synthetic versatility. Since a ni-
pounds. They catalyse the reversible ringclosure of vici-
nal halo alcohols, yieldingepoxides and halide ions
(Scheme 1).[1]
trile group can be transformed into an amino, amide, car-
boxy, or carbonyl group, enantiopure b-hydroxy nitriles
can serve as versatile buildingblocks for the synthesis of
Several microorganisms are known to possess halohy- biologically active compounds. There are several bioca-
drin dehalogenases.[2] Although not yet commercially talytic approaches for their preparation, such as stereo-
available, these enzymes are highly promising biocata- selective microbial reduction of b-keto nitriles,[9,10] li-
lytic tools for the synthesis of optically pure epoxides, pase-catalysed resolution of b-hydroxy nitrile acetates
halohydrins and other b-substituted alcohols due to or alcohols[11] and resolution of b-hydroxy nitriles by ni-
their enantioselectivity[1,3] and capability to accept
non-natural nucleophiles.[4] It was found that besides
the halide, other small negatively charged ions like
azide, cyanide and nitrite are also accepted in the ring
openingreaction (Figure 1). [4] In this way some b-azido
Figure 1. Ringopeningreactions catalysed by halohydrin de-
halogenase.
Scheme 1.
Adv. Synth. Catal. 2006, 348, 579 – 585
ꢀ 2006 Wiley-VCH VerlagGmbH & Co. KGaA, Weinheim
579

´
Maja Majeric Elenkov et al.
FULL PAPERS
trilases.[12] A valuable alternative to these methods
would be the use of halohydrin dehalogenases in a kine-
ꢀ
tic resolution of epoxides with the formation of a C C
bond at the same time. Although halohydrin dehaloge-
nases have been known for several years, their potential
for the synthesis of b-hydroxy nitriles has not been well
explored. Until now, there are only two reports describ-
ingthis conversion. [7,8] A halohydrin dehalogenase iso-
lated from a Corynebacterium sp. strain N-1074, called
H-lyase A (closely related to HheA, see below), cata-
lysed the ringopeningof 1,2-epoxybutane resultingin
racemic 3-hydroxypentanenitrile.[7] H-Lyase B (closely
related to HheB), isolated from the same microorgan-
ism, showed enantioselectivity towards epichlorohy-
drin, yieldingenantiomerically enriched ( R)-g-chloro-
b-hydroxybutyronitrile upon reaction with cyanide.[8]
In order to explore the catalytic scope of the halohydrin
dehalogenase-catalysed synthesis of b-hydroxy nitriles,
more thorough studies are needed.
Here we describe the substrate specificity and enan-
tioselectivity in cyanide-mediated epoxide ringopening
of three different halohydrin dehalogenases. The en-
zymes are obtained from Arthrobacter sp. AD2
(HheA), Mycobacterium sp. GP1 (HheB) and Agrobac-
terium radiobacter AD1 (HheC). Based on sequence
analysis and comparison with other genes, they can be
divided into three distinct groups that share 24 to 32%
Scheme 2.
sequence identity, which are referred as groups A, B and disubstituted terminal epoxides tested were con-
and C.[1] The best studied amongthese enzymes is
verted and in each case GC analysis indicated formation
HheC. It is a homotetrameric protein with 28 kDa sub- of only one product. The structures of the products were
units. The crystal structure and catalytic mechanism of confirmed usingchemically synthesized racemic stand-
HheC have been determined recently.[13]
ards 1–9b that were co-injected on GC columns. With
non-terminal epoxides 10–14a, no product formation
was observed within 24 h under the described experi-
mental conditions, indicatingthat the enzymes have no
significant activity with these compounds. The enzyma-
Results and Discussion
Since the three halohydrin dehalogenases isolated in our tic reactions with terminal epoxides proceeded with
laboratory have not been studied as biocatalysts in the preferential attack of CNꢀ at the less substituted carbon
epoxide ringopeningreaction by cyanide, we decided
atom, in all cases with high regioselectivity. Only epox-
to start a systematic evaluation of this reaction. Our in- ide 8a was found to be unstable in water, due to the pres-
vestigations are focused on the enantioselectivity of the ence of a vinyl group at the a-carbon atom of the oxirane
wild-type dehalogenases (HheA, HheB and HheC) and ring. This substrate undergoes spontaneous hydrolysis
the dependence of the activity and selectivity on the sub- (ca. 5% per h) and chemical formation of the diol was
strate structure. Several structurally different aliphatic competitive with enzymatic conversion. Replacingthe
epoxides (1–14a) were used as substrates to cover varia- vinyl group on the 2,2-disubstituted oxirane ring with
tions in the number and size of substituents on the oxir- an ethyl group resulted in substrate 9a, which was resis-
ane ring(Scheme 2).
All enzymatic reactions were carried out in Tris-SO4 were repeated in the absence of enzyme to determine
buffer (pH 7.5) containing0.5% DMSO to increase the rate of the chemical ringopeningreaction. The
tant to hydrolysis. For each substrate, the experiments
the solubility of the substrate, NaCN (15 mM) and epox- non-enzyme-catalysed formation of b-hydroxy nitriles
ide (5 mM), except with 3a and 14a for which, because of was found to be absent or very slow. Chemical conver-
low aqueous solubility, a 2 mM solution of epoxide was sion was negligible at epoxide and cyanide concentra-
used. Reactions were performed with purified enzyme tions of 5 and 15 mM, respectively, and only became sig-
at room temperature and monitored by periodically tak- nificant at higher concentrations and with prolonged re-
ingsamples, followed by a GC analysis.
action time. With 50 mM 1a and 100 mM NaCN 4% of
All three enzymes showed a broad substrate range racemic product 1b was observed after 2 h at room tem-
with the terminal epoxides 1–9a (Table 1). All mono- perature.
580
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH VerlagGmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 579 – 585

Enantioselective RingOpeningof Epoxides with Cyanide
FULL PAPERS
Table 1. Kinetic resolution of rac 1–9a.
Enzyme Substrate Initial activity[a]
Conversion[b] [%] Epoxide
ee [%] Config.[d, f] ee [%] Config.[d]
b-Hydroxy nitrile
E value[c]
[mmol minꢀ1 mgꢀ1
]
1
2
3
4
5
6
7
8
9
HheA
HheA
HheA
HheA
HheA
HheA
HheA
HheA
HheA
1a
2a
3a
4a
5a
6a
7a
8a
9a
1a
2a
3a
4a
5a
6a
7a
8a
9a
1a
2a
3a
4a
5a
6a
7a
8a
9a
1.4
1.8
0.8
0.6
0.3
0.6
0.7
0.6
1.2
0.4
1.2
0.2
0.2
0.1
0.3
0.2
0.8
2.2
1.0
0.2
0.1
0.8
0.1
1.1
0.7
2.1
0.3
48
46
54
33
34
27
40
30
54
43
36
31
24
18
21
29
42
85
58
32
41
57
20
49
51
53
34
43
22
nd
35
51
28
44
18
15
6
40
nd
20
18
10
19
46
69
99
32
nd
71
22
92
82
100
51
(R)
(R)
(R)
(R)
(R)
(S)
(R)
(S)
(S)
(R)
(R)
(R)
(R)
(R)
(R)
(S)
(S)
(R)
(S)
(S)
(S)
(S)
(R)
(S)
(S)
(S)
(S)
46
26
27
71
97
75
66
42
13
8
71
75
62
80
37
48
63
71
71
67
54
54
86
94
79
90
99
(S)
(S)
(S)
4
2
2
8
109
9
7
3
1
1
9
10
5
10
2
3
7
2
29
7
5
(R)[e]
(R)[e]
(S)[e]
(R)[e]
(S)[e]
(R)
10 HheB
11 HheB
12 HheB
13 HheB
14 HheB
15 HheB
16 HheB
17 HheB
18 HheB
19 HheC
20 HheC
21 HheC
22 HheC
23 HheC
24 HheC
25 HheC
26 HheC
27 HheC
(S)
(S)
(S)
(R)[e]
(R)[e]
(R)[e]
(S)[e]
(S)[e]
(R)
(R)
(R)
(R)
(S)[e]
(R)[e]
(S)[e]
(S)[e]
(S)[e]
(R)
7
16
106
21
141
>200
[a]
Expressed in mmol of product formed per min per mgenzyme.
[b]
Data given for conversion are calculated according to formula c¼ees/(ees þeep), and are in good correlation with conver-
sions determined experimentally usingGC. Large deviations were observed only with substrate 8a where, due to the che-
mical instability of the epoxide, a significant amount of diol was formed as well.
[c]
E values were calculated from eep and ees accordingto formula
E¼ln[(1 – ees)/(1þees/eep)]/ln[(1þees)/(1þees/eep)], except for 3a, where E was calculated from c and eep.[14] E values
above 100 should be considered as approximate values.
[d]
[e]
[f]
Absolute configurations were determined by GC analysis on a chiral stationary phase by comparison of retention times with
reference materials.
Apparent inversion of configuration of epoxides (S)-4–8a to alcohols (R)-4–8b is due to a different substituent priority
ꢀ
accordingto the Cahn–Ingold Prelogconvention.
Configuration of the remaining enantiomer.
The three enzymes converted the terminal epoxides strate 5a, which has the bulkiest substituent. Most of
1–9a with different rates and enantioselectivities. In the tested substrates were converted with low enantiose-
general, HheA and HheB catalysed their reactions lectivity by HheB. The highest E values (E¼10) were
with lower enantioselectivity, but for both enzymes it obtained with substrates havinga longer chain (entries
was observed that enantioselectivity rises with increas- 11 and 12), or a bulky substituent (entry 14).
ingbulkiness of the substituent, although the activity
The results show that HheC is the most selective
drops (Table 1, entries 1, 4, 5 and 10, 13, 14). The influ- amongthe tested enzymes. Previously this enzyme
ence of bulkiness is more evident for HheA, where an in- showed a high selectivity in azide- and nitrite-mediated
crease of enantioselectivity was observed from E¼4 for ringopeningreactions of aliphatic and aromatic epox-
1a, to E¼8–9 for 4a and 6a, reaching E¼109 for sub- ides.[4–6] Whereas its selectivity towards monosubstitut-
Adv. Synth. Catal. 2006, 348, 579 – 585
ꢀ 2006 Wiley-VCH VerlagGmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
581

´
Maja Majeric Elenkov et al.
FULL PAPERS
ed aliphatic epoxides varied from moderate to high (en- groups is prevented, which is important because the cat-
tries 19–25), the enzyme showed a enhanced enantiose- alytic activity is not very high (kcat up to about 1 sꢀ1).
lectivity with 2,2-disubstituted epoxides (entries 26 and High substrate concentrations are tolerated (e.g., up to
27). For monosubstituted epoxides its initial activity sig- 300 mM for epoxybutane), allowingpreparative scale
nificantly dropped with increasinglength of alkyl sub-
stituent (initial activity for R¼Et > n-Bu > n-Hex, en-
tries 19–21), as well aswith branching(initial activity for
applications of these novel biocatalysts.
R¼Et > i-Pr > t-Bu, entries 19, 22, 23). Branchingof Conclusion
the alkyl substituent did not increase enantioselectivity,
except for substrate 6a, which bears a cyclohexyl sub- Wild-type halohydrin dehalogenases were able to cata-
stituent. The enantioselectivity of HheC towards 6a is lyse the cyanide-mediated ringopeningreaction of ep-
the highest among the monosubstituted epoxides (E¼ oxides yieldingenantiomerically enriched b-hydroxy ni-
106), with no loss of activity (entry 24).
triles. HheA, HheB and HheC showed different catalyt-
Introduction of a methyl group as a second substituent ic properties concerningactivity, enantioselectivity and
at the chiral centre led to a huge increase in enantiose- enantiopreference. Although HheB had a low enantio-
lectivity with E values above 100 for 8a (compare entries selectivity towards aliphatic epoxides, this can be con-
25 and 26) and 9a (compare entries 19 and 27). A similar sidered as an advantage when only regioselectivity is
enhanced selectivity with disubstituted epoxides was re- an issue. Chemical methods[16] for epoxide ringopening
ported for some bacterial epoxide hydrolases.[15] The usually proceed with low regioselectivity, require a
2,2-disubstituted epoxides turned out to be the best sub- Lewis acid catalyst, high temperatures, a prolonged re-
strates for HheC with regard to enantioselectivity and action time, and the use of organic solvents. In contrast,
these conversions yielded highly optically enriched the biocatalytic ringopeningwith cyanide in water pro-
product (ee 99%). The high enantioselectivity towards ceeds under mild conditions with high regioselectivity.
2,2-disubstituted oxiranes (entries 26 and 27) and the Therefore, when racemic 3-substituted 3-hydroxy ni-
lower activity of HheC toward substrates with larger triles are the desired products, HheB is the catalyst of
substituents (entries 20, 21, and 23) disclose HheC as a choice because of its high regioselectivity, low enantio-
biocatalyst suitable for the resolution of epoxides bear- selectivity, and broad substrate range. HheA shows po-
inga fully substituted chiral centre with small substitu-
tential as a biocatalyst for the resolution of epoxides
ents. In this way, enantiomerically pure tertiary alcohols bearingbulky substituents. In spite of its low enantiose-
containinga synthetically useful nitrile group can be ob-
tained.
lectivity towards the majority of the tested substrates, it
showed promisingresults with 5a, which is an indication
Absolute configurations of the nitriles produced by that more suitable substrates can be found.
halohydrin dehalogenases were determined using The catalytic properties of HheC strongly differ from
chemically prepared standard compounds of known those of the other two halohydrin dehalogenases. The
configuration. We found that the stereochemical course enantiopreference of HheC is opposite to that of
of the enzymatic reactions mostly depends on the bioca- HheA and HheB. It is an (R)-selective enzyme and it
talyst and less on the substrate structure. The (S)-enan- shows the highest enantioselectivity among the tested
tiomers were transformed more rapidly in reactions cat- enzymes. The large enhancement of enantioselectivity
alysed by HheA and HheB. However, for HheB a vinyl that is obtained when a second substituent is present
substituent (7a, 8a) turns the enantiopreference in the on the chiral centre of the epoxide ringis remarkable.
opposite direction (entries 16 and 17), while for HheA The high enantioselectivity toward 2,2-disubstituted ep-
there was an inversion of selectivity for 2,2-disubstituted oxides makes HheC a promisingbiocatalyst for the syn-
epoxides (8a, 9a) as well as for 6a (entries 6, 8, and 9). thesis of enantiomerically pure tertiary alcohols.
The stereopreference of HheC was found to be opposite
to that of the A- and B-type enzymes. For all tested sub- b-hydroxy nitriles is a C C bond formingreaction. In
strates, except 5a, the (R)-enantiomer was converted general C C bond formingenzymes are useful tools
The halohydrin dehalogenase-catalysed production of
ꢀ
ꢀ
first. This (S)-enantiopreference with 5a is the only ex- for the synthesis of important groups of compounds in
ample of reversed enantioselectivity for HheC. With a chemo-, regio- and enantioselective manner. In this
all other substrates tested in this study and in all our pre- way, enantioselective syntheses of cyanohydrins, b-hy-
vious work with other nucleophiles[4–6] HheC behaved droxy ketones, and a-hydroxy carbonyl compounds
as an (R)-selective enzyme. The steric properties of sub- were achieved.[17] The range of C C bond formingen-
ꢀ
strate 5a are probably responsible for this inversion of zymes is now expanded with three halohydrin dehaloge-
stereopreference.
nases. These halohydrin dehalogenases also comple-
Halohydrin dehalogenases are cofactor-independent ment the biocatalytic scope of hydroxynitrile lyases,
enzymes that are easy to prepare from recombinant E. since both catalyse formation of products that contain
coli cells and they show high stability under the incuba- a hydroxy and a nitrile group and that are, therefore,
tion conditions used when oxidation of sulfhydryl of the same bifunctional nature, but halohydrin dehalo-
582
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH VerlagGmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 579 – 585

Enantioselective RingOpeningof Epoxides with Cyanide
FULL PAPERS
then a solution of CH3CN (0.78 mL, 15 mmol) in THF
(7.5 mL) was added. After 1 h stirringat ꢀ788C, a solution
of propionaldehyde (0.9 mL, 12.5 mmol) in THF (7.5 mL)
was added. After 3 h of stirringat ꢀ788C, the reaction was
quenched with a saturated aqueous solution of NH4Cl and
treated with aqueous 1 M HCl, and the pH was adjusted to neu-
tral. The reaction mixture was extractedwith Et2O. The organic
extracts were dried over Na2SO4, filtered, and solvent was
evaporated. Crude product was purified by column chroma-
tography on silica gel using hexane/ethyl acetate (7:3) as elu-
ent.
genases yield b-cyanohydrins whereas hydroxynitrile
lyases yield a-cyanohydrins.[18]
Experimental Section
General
1H and 13C NMR spectra were recorded on Varian 300
(75.48 MHz) or Varian 400 (100.00 MHz) spectrometer in
CDCl3. Chemical shift values are denoted in d values (ppm)
Usingthe same procedure, b-hydroxy nitriles 2–9b were
also prepared from the correspondingcarbonyl compounds.
1
relative to residual solvent peak (CHCl3, H d¼7.26, 13C d¼
77.16). Optical rotations were measured on a Polartronic
MH8 (SchmidtþHaensch) polarimeter. Gas chromatography
(GC) was performed on Shimadzu GC-17A and Hewlett-Pack-
ard 6890 gas chromatographs equipped with FID detectors us-
ingHe as a carrier gas. Enzymatic reactions were monitored us-
inga HP-1 column (60 m ꢁ0.25 mmꢁ0.25 mm, Agilent). Opti-
cal purities were determined usinga Chiraldex G-TA column
(30 mꢁ0.25 mmꢁ0.25 mm, Astec). Concentrations of epox-
ides and hydroxynitriles were derived from calibration curves.
The commercially available substrates 1,2-epoxybutane
(1a), 1,2-epoxyoctane (3a), cyclopentene oxide (13a), cyclo-
hexene oxide (14a) were supplied by Aldrich Chemie, as
were the enantiomerically pure (R)-1a, (R)-2a and (R)-3a.
1,2-Epoxyhexane (2a), 2-vinyloxirane (7a), (R)-7a and trans-
2,3-epoxybutane (10a) were acquired from Fluka. 1,2-Epoxy-
3-methylbutane (4a), 3,3-dimethyl-1,2-epoxybutane (5a)
were purchased from Lancaster and 1,2-epoxy-2-methylbu-
tane (9a), cis-2,3-epoxybutane (11a), 2,3-epoxy-2-methylbu-
tane (12a) from Acros Organics.
1
3-Hydroxypentanenitrile (1b): H NMR (CDCl3): d¼1.00
(3H, t, J¼7.0 Hz), 1.78 (2H, dq, J1 ¼8.0 Hz, J2 ¼8.0 Hz), 1.94
(1H, bs), 2.48 (1H, dd, J1 ¼17.0 Hz, J2 ¼6.0 Hz), 2.56 (1H, dd,
J1 ¼17.0 Hz, J2 ¼5.0 Hz), 3.85–3.90 (1H, m); 13C NMR
(CDCl3): d¼9.9, 25.9, 29.7, 69.8, 118.25.
1
3-Hydroxyheptanenitrile (2b): H NMR (CDCl3): d¼0.92
(3H, t, J¼7.0 Hz), 1.34–1.48 (4H, m), 1.56–1.63 (2H, m),
2.05 (1H, bs), 2.48 (1H, dd, J1 ¼17.0 Hz, J2 ¼6.0 Hz), 2.56
(1H, dd, J1 ¼17.0 Hz, J2 ¼5.0 Hz), 3.92–3.96 (1H, m); 13C
NMR (CDCl3): d¼14.2, 22.7, 26.3, 26.9, 66.0, 118.1.
1
3-Hydroxynonanenitrile (3b): H NMR (CDCl3): d¼0.88
(3H, t, J¼7.0 Hz), 1.29–1.46 (8H, m), 1.56–1.61 (2H, m),
1.92 (1H, bs), 2.48 (1H, dd, J1 ¼17.0 Hz, J2 ¼6.0 Hz), 2.56
(1H, dd, J1 ¼17.0 Hz, J2 ¼5.0 Hz), 3.90–3.98 (1H, m); 13C
NMR (CDCl3): d¼14.3, 22.8, 24.6, 26.3, 29.3, 31.9, 36.8, 68.9,
118.2.
3-Hydroxy-4-methylpentanenitrile (4a): 1H NMR (CDCl3):
d¼0.96 (3H, d, J¼7.0 Hz), 0.98 (1H, d, J¼7.0 Hz), 1.81 (1H,
dsept, J1 ¼6.5 Hz, J2 ¼6.5 Hz), 2.06 (1H, bs), 2.50 (1H, dd,
J1 ¼17.0 Hz, J2 ¼7.0 Hz), 2.55 (1H, dd, J1 ¼17.0 Hz, J2 ¼
5.0 Hz), 3.67–3.73 (1H, m); 13C NMR (CDCl3): d¼17.6, 18.7,
23.9, 33.5, 72.8, 118.6.
Enzymes were prepared from recombinant E. coli cells ex-
pressingthe hheA, hheB, or hheC genes according to a previ-
ously described protocol.[1]
3-Hydroxy-4,4-dimethylpentanenitrile (5a): 1H NMR
(CDCl3): d¼1.65 (9H, s), 2.27 (1H, d, J¼5.0 Hz), 2.41 (1H,
dd, J1 ¼16.5 Hz, J2 ¼9.5 Hz), 2.54 (1H, dd, J1 ¼6.5 Hz, J2 ¼
3.0 Hz), 3.65 (1H, ddd, J1 ¼9.5 Hz, J2 ¼5.0 Hz, J3 ¼3.0 Hz);
13C NMR (CDCl3): d¼18.7, 21.8, 25.5, 35.2, 75.7, 120.0.
3-Cyclohexyl-3-hydroxypropionitrile (6a): 1H NMR
(CDCl3): d¼0.98–1.31 (5H, m), 1.43–1.54 (1H, m), 1.61–
1.90 (5H, m), 2.05 (1H, d, J¼5.0 Hz), 2.50 (1H, dd, J1 ¼
16.5 Hz, J2 ¼6.5 Hz), 2.58 (1H, dd, J1 ¼16.5 Hz, J2 ¼5.0 Hz),
3.64–3.73 (1H, m); 13C NMR (CDCl3): d¼23.8, 25.9, 26.1,
26.4, 28.1, 29.2, 43.3, 72.2, 118.5.
Synthesis of Substrate 6a
To a solution of cyclohexanecarboxaldehyde (2.24 g, 20 mmol)
in CH2Cl2 (20 mL) trimethylsulphonium methyl sulphate was
added (4.4 g, 23.3 mmol). Aqueous NaOH (50%, 10 mL) was
added and the reaction mixture was stirred overnight. Water
was added and the organic phase was separated. The water
phase was extracted twice with CH2Cl2. The combined organic
phase was washed twice with water, shaken for 20 min with a
saturated solution of sodium metabisulphite to remove un-
reacted aldehyde, and washed twice with water. The organic
phase was dried and evaporated. Bulb-to-bulb distillation un-
der reduced pressure (10 mm Hg, 60–708C) afforded 6a as a
colourless liquid; yield: 1.3 g(43%). 1H NMR (CDCl3): d¼
1.05–1.29 (6H, m), 1.65–1.74 (4H, m), 1.85–1.88 (1H, m),
2.51 (1H, dt, J1 ¼4.5 Hz, J2 ¼1.0 Hz), 2.70 (2H, m); 13C NMR
(CDCl3): d¼25.8, 25.9, 26.6, 44.3, 29.0, 40.6, 46.3, 57.0 ppm.
3-Hydroxy-4-pentenenitrile (7a): 1H NMR (CDCl3): d¼
2.20 (1H, d, J¼4.5 Hz), 2.58 (1H, dd, J1 ¼17.0 Hz, J2 ¼
6.0 Hz), 2.64 (1H, dd, J1 ¼17.0 Hz, J2 ¼5.0 Hz), 4.45–4.48
(1H, m), 5.31 (1H, d, J¼10.0 Hz), 5.41 (1H, d, J¼17 Hz),
5.93 (1H, ddd, J1 ¼17.0 Hz, J2 ¼10.0 Hz, J3 ¼6.0 Hz); 13C
NMR (CDCl3): d¼26.3, 68.8, 117.5, 117.9, 135.6.
3-Hydroxy-3-methyl-4-pentenenitrile (8a): 1H NMR
(CDCl3): d¼0.97 (3H, t, J¼7.5 Hz), 1.35 (3H, s), 1.63–1.68
(3H, m), 2.50 (1H, s), 2.51 (1H, s); 13C NMR (CDCl3): d¼8.3,
26.3, 31.1, 34.4, 71.6, 118.0.
Synthesis of Racemic b-Hydroxy Nitriles 1–9b;
General Procedure[11]
3-Hydroxy-3-methylpentanenitrile (9a): 1H NMR (CDCl3):
d¼0.97 (3H, t, J¼7.5 Hz), 1.35 (3H, s), 1.63–1.68 (2H, m), 2.51
(2H, d, J¼1.0 Hz); 13C NMR (CDCl3): d¼8.3, 26.3, 31.0, 34.4,
71.5, 118.0.
To a solution of (i-Pr)2NH (2.01 mL, 15 mmol) in dry THF
(15 mL), a 1.6 M solution of n-BuLi in hexanes (8.75 mL,
14 mmol) was added under an inert atmosphere of N2 at 08C.
The mixture was stirred for 15 min, cooled to ꢀ788C, and
Adv. Synth. Catal. 2006, 348, 579 – 585
ꢀ 2006 Wiley-VCH VerlagGmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
583

´
Maja Majeric Elenkov et al.
FULL PAPERS
Table 2. Chiral GC analysis of epoxides and b-hydroxy nitriles.
Compound
Conditions
Retention time [min]
Epoxide (a)
b-Hydroxy nitrile (b)
1
2
3
4
5
6
7
8
9
408C 4 min, 108C/min to 1208C, 15 min at 120 8C
608C 7 min, 158C/min to 1508C, 5 min at 1508C
408C 50 min, 158C/min to 1508C, 5 min at 1508C
408C 4 min, 108C/min to 1508C, 5 min at 1508C
408C 4 min, 108C/min to 1208C, 15 min at 120 8C
608C 7 min, 158C/min to 1708C, 10 min at 170 8C
408C 5 min, 108C/min to 1508C, 3 min at 1508C
408C 5 min, 158C/min to 1358C, 5 min at 1358C
408C 5 min, 158C/min to 1358C, 5 min at 1358C
3.9 (R)/4.1 (S)
4.1 (S)/4.3 (R)
52.7 (S)/52.8 (R)
4.7 (R)/4.9 (S)
6.2 (R)/6.5 (S)
11.8 (R)/11.9 (S)
3.4 (S)/3.5 (R)
4.2 (R)/4.6 (S)
4.3 (R)/4.6 (S)
22.0 (R)/23.8 (S)
15.1 (R)/15.4 (S)
66.7 (R)/67.1 (S)
16.8 (S)/17.1 (R)
26.0 (S)/26.5 (R)
21.4 (S)/21.7 (R)
16.3 (S)/16.7 (R)
15.2 (S)/15.6 (R)
15.1 (R)/15.3 (S)
ꢀ23.4 (c 0.76 CHCl3); (S)-2-methyl-butane-1,2-diol [a]D:
ꢀ3.9 (c 2.89 CHCl3), Lit.[22] (R) [a]D: þ8.7 (c 1.35 CHCl3).
From isoprene, a mixture of two diols was obtained, they
were not separated but used as a mixture. The absolute config-
uration of (S)-2-methylbut-3-ene-1,2-diol was assigned by
analogy to the stereochemistry of dihydroxylation of 3,3-di-
methyl-1-butene and 2-methyl-1-butene. Diols were cyclised
to (S)-5a (ee 63%), (S)-8a (ee 8%) and (S)-9a (ee 50%) accord-
ingto a literature procedure. [20] Since these epoxides are highly
volatile, they were not isolated but used as ethereal solutions.
Kinetic Resolution of Epoxides
To 20 mL of Tris-SO4 buffer (0.2 M, pH 7.5) containingepoxide
(0.1 mmol, 5 mM) and DMSO (100 mL, 0.5%) purified enzyme
was added (0.5 mg, final concentration ca. 0.8 mM). Because of
the low solubility of 3a and 14a, they were used as 2 mM solu-
tions. Enzymatic reactions were performed at ambient temper-
ature (228C). The progress of the reaction was followed by pe-
riodically takingsamples (0.5 mL) from the reaction mixture.
Samples were extracted with diethyl ether (1 mL) containing
an internal standard (1-chlorohexane or mesitylene), dried
over Na2SO4 and analysed by GC. The conversion of epoxides
was monitored on an achiral column. Non-enzymatic ring
openingwas determined by monitoring b-hydroxy nitrile for-
mation in the absence of enzyme. The enantiomeric excess
(ee) of the remainingepoxides was determined by chiral GC
under conditions described in Table 2. All separations were
carried out on a Chiraldex G-TA column.
Synthesis of (S)-4b and (S)-6b; General Procedure
Racemic alcohols 4b or 6b (200 mg) were dissolved in vinyl
acetate (20 mL) and hexane (100 mL). The reaction was initi-
ated by the addition of Pseudomonas cepacia lipase
(200 mg). The progress of the reaction was followed by GC
and reactions were terminated when ca. 50% conversion to
acetate was reached. Enzyme was filtered off, solvent was
evaporated, and crude product was purified on silica gel using
hexane/ethyl acetate (9:1) as eluent. The enantiomerically en-
riched (R)-acetates and (S)-alcohols (S)-4b (ee 61%) and (S)-
6b (ee 90%) were isolated. Specific optical rotation values
are: (R)-1-(cyanomethyl)-2-methylpropyl acetate [a]D: ꢀ60
(c 0.75 CHCl3), Lit.[11] (R) [a]D: ꢀ52.2 (c 0.76 CHCl3); (R)-2-cy-
ano-1-cyclohexylethyl acetate [a]D: ꢀ32 (c 0.85 CHCl3), Lit.[11]
(R) [a]D: ꢀ31.4 (c 0.52 CHCl3).
Absolute Configurations
Absolute configurations were assigned by chiral GC analysis
usingreference compounds. Commercially available were op-
tically pure (R)-1a, (R)-2a, (R)-3a and (R)-7a. Enantiomerical-
ly enriched epoxides (S)-5a, (S)-8a and (S)-9a were prepared
by Sharpless asymmetric dihydroxylation[19] of 3,3-dimethyl-
1-butene, isoprene, or 2-methyl-1-butene followed by ringclo-
sure (TsCl, NaH, THF).[20] Their absolute configurations (S)
were assigned by comparison of the optical rotations of the
formed diols with literature data.[21,22] Enantiomerically en-
riched b-hydroxy nitriles (S)-4b and (S)-6b were obtained by li-
pase-catalysed resolution of racemic alcohols using Pseudomo-
nas cepacia lipase (Amano). Their (S) absolute configuration
was assigned by comparison of the optical rotations of the
formed acetates with literature data.[11]
References
[1] J. E. T. van Hylckama Vlieg, L. Tang, J. H. Lutje Spel-
berg, T. Smilda, G. J. Poelarends, T. Bosma, A. E. J. van -
Merode, M. W. Fraaije, D. B. Janssen, J. Bacteriol. 2001,
183, 5058–5066.
[2] a) T. Nakamura, T. Nagasawa, F. Yu, I. Watanabe, H. Ya-
mada, Appl. Environ. Microbiol. 1994, 60, 1297–1301;
b) C. E. Castro, E. W. Bartnicki, Biochemistry 1968, 7,
3213–3218; c) H. M. S. Assis, P. J. Sallis, A. T. Bull, D. J.
Hardman, Enzyme Microb. Technol. 1998, 22, 568–574;
d) A. J. van den Wijngaard, P. T. W. Reuvekamp, D. B.
Janssen, J. Bacteriol. 1991, 173, 124–129; e) G. J. Poelar-
ends, J. E. T. van Hylckama Vlieg, J. R. Marchesi, L. M.
Freitas Dos Santos, D. B. Janssen, J. Bacteriol. 1999,
181, 2050–2058.
Synthesis of (S)-5a, (S)-8a and (S)-9a; General
Procedure
Accordingto Sharpless et al. [19] enantiomerically enriched di-
ols were prepared usingAD-mix- a, startingfrom 3,3-dimeth-
yl-1-butene, isoprene, or 2-methyl-1-butene (Fluka). Their ab-
solute configurations were assigned on the basis of optical ro-
tations. Specific optical rotation values are: (S)-3,3-dimethyl-
butane-1,2-diol [a]D: þ13.9 (c 1.25 CHCl3), Lit.[21] (R) [a]D:
584
asc.wiley-vch.de
ꢀ 2006 Wiley-VCH VerlagGmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 579 – 585

Enantioselective RingOpeningof Epoxides with Cyanide
FULL PAPERS
[3] J. H. Lutje Spelberg, J. E. T. van Hylckama Vlieg, T. Bos-
ma, R. M. Kellogg, D. B. Janssen, Tetrahedron: Asymme-
try 1999, 10, 2863–2870.
[4] J. H. Lutje Spelberg, L. Tang, M. van Gelder, R. M. Kel-
logg, D. B. Janssen, Tetrahedron: Asymmetry 2002, 13,
1083–1089.
[5] J. H. Lutje Spelberg, J. E. T. van Hylckama Vlieg, L.
Tang, D. B. Janssen, R. M. Kellogg, Org. Lett. 2000, 3,
41–43.
C. L. Rand, D. P. Weiner, D. E. Robertson, M. J. Burk,
J. Am. Chem. Soc. 2002, 124, 9024–9025.
[13] R. M. de Jong, J. J. W. Tiesinga, H. J. Rozeboom, K. H.
Kalk, L. Tang, D. B. Janssen, B. W. Dijkstra, EMBO J.
2003, 22, 4933–4944.
[14] C.-S. Chen, Y. Fujimoto, G. Girdaukas, C. J. Sih, J. Am.
Chem. Soc. 1982, 104, 7294–7299.
[15] M. Mischitz, W. Kroutil, U. Wandel, K. Faber, Tetrahe-
dron: Asymmetry 1995, 6, 1261–1272.
[6] G. Hasnaoui, J. H. Lutje Spelberg, E. de Vries, L. Tang,
B. Hauer, D. B. Janssen, Tetrahedron: Asymmetry 2005,
16, 1685–1692.
[7] T. Nakamura, T. Nagasawa, F. Yu, I. Watanabe, H. Ya-
mada, Biochem. Biophys. Res. Commun. 1991, 180,
124–130.
[16] a) M. Chini, P. Crotti, L. Favero, F. Macchia, Tetrahedron
Lett. 1991, 32, 4775–4778; b) J. A. Ciaccio, M. Smrtka,
W. A. Maio, D. Rucando, Tetrahedron Lett. 2004, 45,
7201–7204.
[17] J. Sukumaran, U. Hanefeld, Chem. Soc. Rev. 2005, 34,
530–542.
[8] T. Nakamura, T. Nagasawa, F. Yu, I. Watanabe, H. Ya-
mada, Tetrahedron 1994, 50, 11821–11826.
[9] J. R. Dehli, V. Gotor, Tetrahedron: Asymmetry 2000, 11,
3693–3700.
[10] A. J. Smallridge, A. Ten, M. A. Trewhella, Tetrahedron
Lett. 1998, 39, 5121–5124.
[11] T. Itoh, K. Mitsukura, W. Kanphai, Y. Takagi, H. Kihara,
H. Tsukube, J. Org. Chem. 1997, 62, 9165–9172.
[12] G. DeSantis, Z. Zhu, W. A. Greenberg, K. Wong, J.
Chaplin, S. R. Hanson, B. Farwell, L. W. Nicholson,
[18] M. North, Tetrahedron: Asymmetry 2003, 14, 147–176.
[19] K. B. Sharpless, W. Amberg, Y. L. Bennani, G. A. Crisp-
ino, J. Hartung, K.-S. Jeong, H.-L. Kwong, K. Morikawa,
Z.-M. Wang, D. Xu, X.-L. Zhang, J. Org. Chem. 1992, 57,
2768–2771.
[20] P. Moussou, A. Archelas, R. Furstoss, J. Mol. Catal. B
Enzym. 1998, 5, 447–458.
[21] J. P. Guette, N. Spassky, Bull. Soc. Chim. Fr. 1972, 4217–
4224.
[22] D. Seebach, R. Naef, G. Calderari, Tetrahedron 1984, 40,
1313–1324.
Adv. Synth. Catal. 2006, 348, 579 – 585
ꢀ 2006 Wiley-VCH VerlagGmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
585