Deracemization of (±)-2,3-disubstituted oxiranes via biocatalytic hydrolysis using bacterial epoxide hydrolases: Kinetics of an enantioconvergent process

Article abstract of DOI:10.1039/a704812b

Asymmetric biocatalytic hydrolysis of (±)-2,3-disubstituted oxiranes leading to the formation of vicinal diols in up to 97% ee at 100% conversion was accomplished by using the epoxide hydrolase activity of various bacterial strains. The mechanism of this deracemization was elucidated by ¹⁸OH₂-labelling experiments using a partially purified epoxide hydrolase from Nocardia EH1. The reaction was shown to proceed in an enantioconvergent fashion by attack of OH^- at the (S)-configured oxirane carbon atom with concomitant inversion of configuration. A mathematical model developed for the description of the kinetics was verified by the determination of the four relative rate constants governing the regio- and enantio-selectivity of the process.

Full text of DOI:10.1039/a704812b

View Article Online / Journal Homepage / Table of Contents for this issue
Deracemization of ( )-2,3-disubstituted oxiranes via biocatalytic
hydrolysis using bacterial epoxide hydrolases: kinetics of an
enantioconvergent process
Wolfgang Kroutil, Martin Mischitz and Kurt Faber*
Institute of Organic Chemistry, Graz University of Technology, Stremayrgasse 16,
A-8010 Graz, Austria
Asymmetric biocatalytic hydrolysis of ( )-2,3-disubstituted oxiranes leading to the formation of vicinal
diols in up to 97% ee at 100% conversion was accomplished by using the epoxide hydrolase activity of
various bacterial strains. The mechanism of this deracemization was elucidated by 18OH2-labelling
experiments using a partially purified epoxide hydrolase from Nocardia EH1. The reaction was shown to
proceed in an enantioconvergent fashion by attack of OH؊ at the (S)-configured oxirane carbon atom
with concomitant inversion of configuration. A mathematical model developed for the description of
the kinetics was verified by the determination of the four relative rate constants governing the regio- and
enantio-selectivity of the process.
value).11 In contrast, incomplete regioselectivity was observed
Introduction
for styrene oxide-type substrates—presumably due to the pres-
ence of a benzylic carbon atom—and for 2,3-disubstituted
analogues.10 As a consequence, E-values were found to be
inapplicable due to the non-classical nature of the kinetics
involved.
In order to broaden the applicability of bacterial epoxide
hydrolases, we extended our investigation in two directions, i.e.
by (i) using substrates bearing two stereogenic centres, and (ii)
the elucidation of the kinetics underlying the reaction.12
Chiral epoxides and vicinal diols—the latter employed as their
corresponding cyclic sulfite or sulfate esters as reactive
intermediates1—are extensively used intermediates for the syn-
thesis of chiral target compounds due to their ability to react
with a broad variety of nucleophiles in a highly stereoselective
fashion. In recent years, intense research has been devoted to
the development of catalytic methods for their production.1–4
Besides chemocatalytic methods,3 biocatalysis4 has been
shown to provide a useful arsenal of methods as alternatives
to the asymmetric epoxidation or dihydroxylation of alkenes.
Among several types of enzymes employed, epoxide hydro-
lases [EC 3.3.2.X] are particularly attractive due to their
cofactor-independence. Although enzymes from mammalian
sources, such as liver tissue, have been investigated in great
detail, aiming at the elucidation of detoxification mechan-
isms,5 biotransformations on a preparative scale were severely
impeded by the limited supply of enzyme, which is generally
isolated from (rat or rabbit) liver tissue derived from
phenobarbital-treated animals. As a consequence, the scale of
these biotransformations did not surpass the millimolar range.
It was only recently that several microbial strains from fun-
gal6,7 and bacterial origin8 were identified as convenient and
reliable sources of highly selective epoxide hydrolases.9 This
allows the production of biocatalysts in (almost) unlimited
amounts and opens the way to the transformation of sub-
strates on a multigram scale.7
Results and discussion
Biocatalytic hydrolyses
Guided by our experience that bacterial epoxide hydrolases in
general do not accept substrates bearing polar functional
groups, cis- and trans-configurated 2,3-dialkyloxiranes bearing
alkyl groups of varying chain length were chosen as substrates
(Scheme 1). The most promising bacterial strains showing high
Lyophilized
HO OH
O
R2
O
R2
bacterial cells
Buffer, pH 7.5
+
R1
R1
R1
R2
rac-1 R1 = CH3, R2 = n-C4H9
rac-2 R1 = C2H5, R2 = n-C3H7
rac-3 R1 = CH3, R2 = n-C5H11
rac-4 R1 = CH3, R2 = n-C6H13
(2R,3R )-1
(3R,4R )-2
(2R,3R )-3
(2R,3R )-4
(2R,3S )-1a
(3R,4S )-2a
(2R,3S )-3a
(2R,3S )-4a
During our recent studies on the substrate-structure selectiv-
ity relationship of several bacterial epoxide hydrolases, several
structurally different types of substrates were investigated with
varying success.
Lyophilized
HO OH
O
O
bacterial cells
+
Buffer, pH 7.5
R1
R2
R1
R2
R1
R2
Whereas monosubstituted oxiranes were resolved with
insufficient selectivity by bacterial enzymes—fungal epoxide
rac-5 R1 = CH3, R2 = n-C4H9
rac-6 R1 = C2H5, R2 = n-C3H7
(2R,3S )-5
(3S,4R )-6
(2R,3R )-5a
(3R,4R )-6a
hydrolases seem to be more suitable for these substrates6,7
—
2,2-disubstituted oxiranes were shown to be readily accepted
by bacterial enzymes, and kinetic resolutions proceeded with
virtually absolute enantiospecificity (E-values >200).8a For both
of these types of substrates, the regioselectivity of the oxirane
opening was found to be absolute, i.e. attack always occurred
at the less substituted carbon atom and the configuration of
the stereogenic centre was retained.10 Thus, the reaction fol-
lows a kinetic resolution pattern and the enantioselectivity can
accurately be described by using the ‘Enantiomeric Ratio’ (E-
Scheme 1 Biocatalytic hydrolysis of epoxides ( )-1–6
epoxide hydrolase activity were selected from our previous stud-
ies: Rodococcus sp. NCIMB 11216,8a Rhodococcus equi IFO
3730,13 Rhodococcus ruber DSM 43338,8b Mycobacterium par-
affinicum NCIMB 10420,13 Arthrobacter sp. DSM 312 and three
Nocardia strains (H8, TB1 and EH1).8b Substrates were incu-
bated with rehydrated lyophilized whole cells in Tris-buffer
(pH = 7.5). After a certain degree of conversion had been
J. Chem. Soc., Perkin Trans. 1, 1997
3629

View Article Online
Table 1 Asymmetric hydrolysis of trans-epoxides ( )-1–4
Entry
Biocatalyst
Substrate
Epoxide
ee (%)
~2
Diol
ee (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Rhodococcus sp. NCIMB 11216
Rhodococcus equi IFO 3730
Rhodococcus ruber DSM 43338
Mycobacterium paraffinicum NCIMB 10420
Arthrobacter sp. DSM 312
Nocardia H8
( )-1
( )-1
( )-1
( )-1
( )-1
( )-1
( )-1
( )-1
( )-2
( )-2
( )-2
( )-2
( )-2
( )-2
( )-2
( )-2
( )-3
( )-4
(2R,3R)-1
(2R,3R)-1
(2R,3R)-1
(2R,3R)-1
(2R,3R)-1
(2R,3R)-1
(2R,3R)-1
(2R,3R)-1
(3R,4R)-2
(3R,4R)-2
(3R,4R)-2
(3R,4R)-2
(3R,4R)-2
(3R,4R)-2
(3R,4R)-2
(3R,4R)-2
(2R,3R)-3
(2R,3R)-4
(2R,3S)-1a
(2R,3S)-1a
(2R,3S)-1a
(2R,3S)-1a
(2R,3S)-1a
(2R,3S)-1a
(2R,3S)-1a
(2R,3S)-1a
(3R,4S)-2a
(3R,4S)-2a
(3R,4S)-2a
(3R,4S)-2a
(3R,4S)-2a
(3R,4S)-2a
(3R,4S)-2a
(3R,4S)-2a
(2R,3S)-3a
(2R,3S)-4a
69
65
86
18
90
83
84
90
36
30
40
~4
63
47
46
38
77
78
~4
14
20
~5
13
28
35
~6
17
10
~4
~4
12
21
16
~4
~2
Nocardia TB1
Nocardia EH1
Rhodococcus sp. NCIMB 11216
Rhodococcus equi IFO 3730
Rhodococcus ruber DSM 43338
Mycobacterium paraffinicum NCIMB 10420
Arthrobacter sp. DSM 312
Nocardia H8
Nocardia TB1
Nocardia EH1
Rhodococcus sp. NCIMB 11216
Rhodococcus sp. NCIMB 11216
Table 2 Asymmetric hydrolysis of cis-epoxides ( )-5,6
Entry
Biocatalyst
Substrate
Epoxide
ee (%)
~2
~4
12
~5
Diol
ee (%)
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Rhodococcus sp. NCIMB 11216
Rhodococcus equi IFO 3730
Rhodococcus ruber DSM 43338
Mycobacterium paraffinicum NCIMB 10420
Arthrobacter sp. DSM 312
Nocardia H8
( )-5
( )-5
( )-5
( )-5
( )-5
( )-5
( )-5
( )-5
( )-6
( )-6
( )-6
( )-6
( )-6
( )-6
( )-6
( )-6
(2R,3S)-5
(2R,3S)-5
(2R,3S)-5
(2R,3S)-5
(2R,3R)-5a
(2R,3R)-5a
(2R,3R)-5a
(2R,3R)-5a
n.c.a
(2R,3R)-5a
(2R,3R)-5a
(2R,3R)-5a
(3R,4R)-6a
(3R,4R)-6a
(3R,4R)-6a
(3R,4R)-6a
n.c.a
71
78
95
66
(2R,3S)-5
(2R,3S)-5
(2R,3S)-5
(3S,4R)-6
(3S,4R)-6
(3S,4R)-6
(3S,4R)-6
~5
16
24
~3
~1
~8
~8
78
88
97
58
36
42
71
Nocardia TB1
Nocardia EH1
Rhodococcus sp. NCIMB 11216
Rhodococcus equi IFO 3730
Rhodococcus ruber DSM 43338
Mycobacterium paraffinicum NCIMB 10420
Arthrobacter sp. DSM 312
Nocardia H8
(3S,4R)-6
(3S,4R)-6
(3S,4R)-6
~1
~3
~1
(3R,4R)-6a
(3R,4R)-6a
(3R,4R)-6a
66
77
66
Nocardia TB1
Nocardia EH1
a n.c. = no conversion.
reached within 10–200 h, the diol formed and the remaining
epoxide were isolated and analysed for enantiomeric purity and
absolute configuration.
The results depicted in Tables 1 and 2 show the following
trends:
strates ( )-1 and ( )-2 or ( )-5 and ( )-6]. This fact is under-
standable since one moves towards quasi-symmetrical sub-
strates, which fits well with the observation that meso-oxiranes
were found to be unsuitable substrates.8a,9a
(v) Despite the fact that a range of eight different bacterial
strains were employed, the stereochemical course of the reac-
tion showed a remarkable homogeneous picture: In each case,
trans-epoxides led to the exclusive formation of threo-diols
whereas cis-configurated substrates gave the corresponding
erythro-counterparts. No trace of products from non-matching
stereoselectivity could be detected. As a consequence, the reac-
tion can be formulated as proceeding via SN2-type hydrolytic
opening of the epoxide, leading to inversion of configuration at
the oxirane carbon atom being attacked. This fact is in line with
the generally accepted proposal for the mechanism of epoxide
hydrolases.14
(vi) Regardless of the stereochemistry of the substrate, all of
the strains showed a marked preference for attack at (S)-
configurated oxirane carbon atoms (Scheme 2). This strict pref-
erence for the configuration of the oxirane carbon atom being
attacked was coupled by a trend in the regio-selectivity in
favour of the less sterically hindered position. Thus, with sub-
strates from the trans-series [( )-1 through ( )-4], inversion of
the more accessible (S)-configurated carbon atom led to
(2R,3S)- and (3R,4S)-diols, respectively, while (R,R)-epoxides
were hydrolysed at a slower rate. This was also true for cis-2,3-
epoxyheptane, which led to (2R,3R)-heptanediol and remaining
(2R,3S)-epoxide. The only exception to this rule was substrate
( )-6, where the less accessible (S)-configurated oxirane C-atom
was preferred.
(i) The cis- or trans-configuration of the epoxide has only a
minor impact on the acceptance of substrates, as both series
were shown to be converted by the majority of the strains.
However, in general, the reaction rate was about twice as high
for trans-epoxides as compared with their cis-analogues. The
only notable exception was Arthrobacter sp. DSM 312, which
was unable to hydrolyse cis-epoxides ( )-5 and ( )-6 (entries 23
and 31).
(ii) As may be deduced from the enantiomeric purity of diols
formed and the remaining non-hydrolysed epoxides, the select-
ivity of the reaction covered a wide range, from low to excellent.
In general, the selectivities obtained from cis-epoxides ( )-5
and ( )-6 were slightly higher than those for trans-oxiranes ( )-
1 and ( )-2, although this effect was not pronounced. This can
also be verified by comparison of the kinetic data (see below).
The only notable exception to this trend was found with sub-
strates ( )-1 and ( )-5 with Nocardia H8 (entries 6 and 24).
(iii) By comparing the results obtained with Rhodococcus sp.
NCIMB 11216 using an homologous series of trans-configur-
ated substrates [( )-1, ( )-3, ( )-4] the length of the alkyl chain
was found to have a negligible impact on the selectivity. This
is in sharp contrast to the findings obtained when 2,2-dialkyl-
oxiranes were used as substrates.8a
(iv) The selectivity dropped when the epoxy moiety was
moved towards the centre of the n-alkyl chain [compare sub-
3630
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
O
R
R
O
O
R
R
O
R
S
S
R
R
S
R
S
slow
fast
rac-1 rac-3 rac-4
R = n-C3H7, n-C4H9, n-C6H13
slow
fast
rac-2
R = n-C3H7
O
O
O
O
R
R
S
S
S
S
R
R
R
R
R
R
slow
fast
fast
slow
rac-5
R = n-C4H9
rac-6
R = n-C3H7
preferred enzyme attack
Fig. 1 Time-course for hydrolysis of compound ( )-5
Scheme 2 Stereochemical course of enzymatic hydrolysis
Kinetics
Determination of regio- and enantio-selectivity
The experiment showing the highest apparent selectivity—i.e.
hydrolysis of ( )-5 with Nocardia EH1 (entry 26)—was selected
for the investigation of the kinetics of the system. When the
time-course of the hydrolysis of compound ( )-5 was followed
with respect to the enantiomeric purity of formed diol (2R,3R)-
5a and remaining non-hydrolysed epoxide (2R,3S)-5, an
unexpected phenomenon was observed (Fig. 1): the enantio-
meric excess (ee) of the diol did not decline beyond a conversion
of ~50%—which would be expected from classical kinetic
resolution—but rather remained at a constant level of у90% ee
until the starting material was completely consumed. From
these data, it was concluded that the overall process comprises
four different individual reactions, characterized by the relative
rate constants k1 through k4 (Scheme 4).16 First, two substrate
The absolute configuration of epoxides was determined as fol-
lows (Scheme 3): LiAlH4 reduction of 2,3-epoxyalkenes 1, 3, 4
OH
R1
OH
R2
O
LiAlH4
Et2O
+
R1
R2
R2
R1
cis
trans
R
R
S
R
R1 = Me, Et
R2 = n-C3H7, n-C4H9, n-C5H11, n-C6H13
O
O
R2MgBr
THF
LiAlH4
Et2O
(2R.3RS )-1a, 3a–5a
(3R.4RS )-2a, 6a
R1
R
R1
OH
OEt
R2
k1
P2
OH
OH
O18
H
(2R )-7a R1 = Me, R2 = n-C4H9
(2R )-7b R1 = Me, R2 = n-C5H11
(2R )-7c R1 = Me, R2 = n-C6H13
(3R )-7d R1 = Et, R2 = n-C3H7
O
P
O18
H
H
(2S,3S )-5a
A
k3
P3
(2R,3S )-5
Scheme 3 Determination of absolute configuration
OH
and 5 led to a mixture of corresponding alkan-2-ols and -3-ols,
respectively. The configuration of the alkan-2-ols thus obtained
was determined through GLC analysis of the corresponding
trifluoroacetate esters on a chiral stationary phase via co-
injection with samples obtained from commercially available
reference compounds. Once the configuration of the centre at
C-2 was determined, the remaining configuration at C-3 could
be elucidated from the cis- and trans-configuration of the start-
ing epoxides, which were obtained through MCPBA-oxidation
of the corresponding Z- and E-alkene. An analogous approach
was chosen for 3,4-epoxyalkanes 2 and 6. Thus, LiAlH4 reduc-
tion led to a mixture of heptan-3-ol and (achiral) heptan-4-ol.
The configuration of the former was elucidated via GLC anal-
ysis of the trifluoroacetic acid (TFA) derivative on a chiral sta-
tionary phase using reference material obtained as described
above.
Enantiopure reference material for diols obtained from bio-
catalytic hydrolysis was synthesized as follows (Scheme 3): add-
ition of the corresponding Grignard reagent to commercially
available ethyl (R)-lactate or (R)-2-hydroxybutanoate, respect-
ively, gave rise to (R)-α-hydroxyalkanones 7a–d. The latter were
reduced with LiAlH4 to yield mixtures of the corresponding
diastereomeric diols (2R,3RS)-1a, -3a, -5a and (3R,4RS)-2a,
-6a. Comparison with independently synthesized rac-threo- and
rac-erythro-diols15 revealed their relative configuration. The
configuration of all of the diols obtained from biocatalytic
hydrolyses was elucidated via co-injection on GLC using a
chiral stationary phase.
O18
k2
Q3
OH
O
Q
(2R,3R )-5a
OH
B
k4
Q2
(2S,3R )-5
O18
H
Scheme 4 Kinetics for hydrolysis of epoxide ( )-5
enantiomers A and B may react at different reaction rates (cor-
responding to the sums of k1 ϩ k2 and k3 ϩ k4, respectively),
leading to product enantiomers P and Q. Second, each enan-
tiomer may be hydrolysed through two pathways of different
regioselectivity (k1 or k2, and k3 or k4, respectively), involving
attack at carbon 2 or 3 (Scheme 4).10 In order to determine the
individual relative rate constants, the reaction was performed in
18OH2 using a partially purified epoxide hydrolase prepar-
ation.17 The labelled diol products thus obtained were analysed
by GLC/MS using a chiral stationary phase. The data show that
(i) only one oxygen atom was introduced into the substrate,
which allowed (ii) the position of the 18O-label to be determined
by analysis of the diol fragmentation pattern during glycol
cleavage upon electron-impact ionization. Additionally, the
enantiomeric composition of epoxide (A/B) and diol (corre-
sponding to the sums of Q2 ϩ Q3/P2 ϩ P3)16 was analysed by
GLC on a chiral stationary phase (Table 3).
J. Chem. Soc., Perkin Trans. 1, 1997
3631

View Article Online
For the calculation of the individual relative rate constants
from experimentally obtained data, a mathematical model was
developed.18 The derivations are based upon the following
assumptions:
(i) The specific activity of the enzyme remains constant
throughout the reaction, implying that enzyme deactivation
caused by mechanical or chemical stress is negligible. (ii) Nei-
ther inhibition, (iii) nor spontaneous—and thus non-specific—
hydrolysis19 is impeding the reaction, (iv) the reactions are
irreversible and (v) at the time of measurement, the starting
material is still in excess.
k3B0
3
4
(8)
P3 =
[1 Ϫ eϪ(k ϩk )t
]
k3 ϩ k4
and (8), with A0 and B0 being the initial concentrations of sub-
strates A and B at t = 0, respectively.21 Equations (9) and (10)
are derived in an analogous fashion:
k4B0
3
4
(9)
Q2 =
Q3 =
[1 Ϫ eϪ(k ϩk )t
]
k3 ϩ k4
k2A0
[1 Ϫ eϪ(k ϩk )t
]
1
2
The system described in Scheme 4 can be characterized by the
following differential equations, with ki being the first-order
relative rate constants:
(10)
k1 ϩ k2
After division of equations (1) and (2), followed by integra-
tion, equation (11) is obtained:
dA
(1)
(2)
(3)
(4)
= Ϫ(k1 ϩ k2)A
= Ϫ(k3 ϩ k4)B
dt
eϪ(k ϩk )t
1
2
A
B
(11)
=
eϪ(k ϩk )t
3
4
dB
dt
The quotient (k1 ϩ k2)/(k3 ϩ k4) is calculated from the
experimentally determined enantiomeric composition of the
diol, which corresponds to P2 ϩ P3/Q2 ϩ Q3. If—for reasons
dP
2 = k1A
dt
of clarity—eϪ(k ϩk )t is substituted for x, eϪ(k ϩk )t can be
expressed as Bx/A through equation (11). Now, the relative rate
constants can be related to the enantiomeric composition of the
product P and Q:
1
2
3
4
dQ
3 = k2A
dt
The corresponding equations for P3 and Q2 are obtained in
an analogous fashion. After dividing (3) by (4) and subsequent
integration, the quotients of relative rate constants are obtained
as (5) and (6):
k1
k3
B
A
B
A
(1 Ϫ x) ϩ
(1 Ϫ x) ϩ
ͩ
ͩ
1 Ϫ
1 Ϫ
x
x
ͪ
ͪ
P2 ϩ P3 k1 ϩ k2
k3 ϩ k4
k4
=
(12)
Q2 ϩ Q3
k2
k1 ϩ k2
k3 ϩ k4
k1 P2
(5)
(6)
=
k2 Q3
If x is expressed explicitly from equation (11), one obtains
equation (13).
k3 P3
=
1
1
P2ϩ3
1
1
k4 Q2






ϩ
Ϫ
ؒ
ؒ
ϩ
ϩ
k2
k1
k4 Q2ϩ3
k1
k2
k
3

1 ϩ
1 ϩ
The ratios of P2/Q3 and P3/Q2 were experimentally deter-
mined by GLC/MS via analysis of the peaks having an m/z
of 87 (products P2 and Q2) and 89 (corresponding to P3 and
Q3), which result from 18O-incorporation at C2 and C3,
respectively.20
To obtain the values of k1/k3 and k2/k4, (k1 ϩ k2)/(k3 ϩ k4)
has to be calculated first: thus, integration of equations (3)
and (4)—with P2 and P3 at t = 0 being nil—gives equations (7)
1 ϩ
1 ϩ
k3
k4
x =
(13)
1
B
1
P2ϩ3
Ϫ
k4 Q2ϩ3
1
B
ؒ
1





ϩ
ؒ
k2
k1
A
k1
k2
A
k
3 

1 ϩ
1 ϩ
1 ϩ
1 ϩ

k3
k4
Now, the desired quotient can be expressed as in equation
(14):
k1A0
k1 ϩ k2
k3 ϩ k4
ln x
[1 Ϫ eϪ(k ϩk )t
]
1
2
(7)
P2 =
(14)
=
k1 ϩ k2
B
ln
ͩ
x
ͪ
A
Table 3 Enantiomeric purities and label-distribution for substrate
( )-5 using Nocardia EH1
Other quotients of relative rate constants, such as k1/k3, k2/k4,
etc., can be obtained from equations (14), (5) and (6). Finally—
if the lowest k-value of a given set of four is arbitrarily set to
one—the following kis are obtained (Table 4).
ee (%)
Intensity (counts)
Time (t/h)
(2R,3S)-5
(2R,3R)-5a
P2
P3
Q2
Q3
The accuracy of the mathematical model and the validity of
the basic assumptions discussed above were verified by hydro-
lysing substrate ( )-5 with Nocardia EH1 epoxide hydrolase
22
110
3.3
15.7
95.9
97.3
62
120
10
31
2701
5309
144
308
Table 4 Relative rate constants for substrates ( )-5 and ( )-1
Entry
Substrate
Biocatalyst
Time (t/h)
k1
k2
k3
k4
1
2
3
4
5
6
( )-5
( )-5
( )-5
( )-5
( )-1
( )-1
Nocardia EH1
Nocardia EH1
Rhodococcus ruber DSM 43338
Rhodococcus equi IFO 3730
Nocardia EH1
22
110
22
22
23
6
3.5
14
4
1
17
17
63
6
7
1.4
1
1
1
1
1.1
1
326
343
498
39
36
34
Rhodococcus ruber DSM 43338
23
1.2
3632
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
using 18OH2 as solvent (label content 97%). The reaction was
stopped at two different intervals, after 22 and 110 h, respect-
ively. The enantiomeric purity of the diol (2R,3R)-5a formed
and of the remaining non-hydrolysed epoxide (2R,3S)-5 were
determined by GLC analysis and the label-distribution was
analysed by GLC/MS on a chiral stationary phase (Table 3).
These data were used for the calculation of the relative kis. As
may be seen from Table 4 (entries 1 and 2), the values for the
relative rate constants were within a very narrow margin of
experimental error, indicating the applicability of the math-
ematical model and the validity of the basic assumptions for
the time period investigated. Analysis of the relative rate con-
stants for substrate ( )-5 reveals that the (2S,3R)-epoxide (B,
Scheme 4) is the fast reacting enantiomer, which is hydrolysed
with excellent regioselectivity at the more accessible C-2 atom
(k4/k3 >300:1). On the other hand, the (2R,3S)-enantiomer
(A) is transformed about 14 times slower (k3 ϩ k4/k1 ϩ k2 =
327:23) and low regioselectivity in favour of the sterically
more hindered C-3 atom is observed (k2/k1 = 17:5). Although
the selectivity for the enantiodiscrimination per se is relatively
low (E ~14–17),22 the fact that both enantiomers are hydro-
lysed with opposite regioselectivity yields (2R,3R)-5a in >95%
enantiomeric excess. Since both epoxide substrate enantiomers
are hydrolysed to give the same (2R,3R)-diol product, the
reaction constitutes an enantioconvergent process. Thus it can
be run to completion and 100% theoretical yield can be
accrued.
Comparable results—i.e. low to moderate enantioselectivity in
favour of the (2S,3R)-enantiomer of 5 (E ~6)22 with matching
(opposite) regioselectivity—were obtained for Rhodococcus
ruber DSM 43338 (Table 4, entry 3). On the other hand, signifi-
cantly lower regio- and enantio-selectivities were observed with
Rhodococcus equi IFO 3730 (entry 4).
A similar situation was found for the trans-configurated sub-
strate ( )-1. Both strains investigated showed reduced enantio-
selectivities, mainly due to the drop of the k4-values by about
one order of magnitude, while the regioselectivities remained
within the same range. Thus, the enantiomeric purity of diol
(2R,3S)-1a did not exceed a value of 90%.
ative chromatography was performed on silica gel Merck 60
(40–63 µm). The ee of epoxides and diols was determined by
GLC (Shimadzu GC-14A, equipped with FID) on a CP-
Chirasil-DEX CB column (25 m; 0.32 mm; 0.25 µm film; H2).
GLC/MS data were recorded on an HP 6890 GC/MS spec-
trometer. The EI quadrupole detector was set to 70 eV, 130 ЊC,
auxiliary temp. 180 ЊC. 18O-Labelled water (95–98% 18O-
content) was purchased from Promochem. Optical rotation
values were measured on a Perkin-Elmer polarimeter 341 at
589 nm (Na-line) in a 1 dm cuvette, and are given in units of
1
10Ϫ1 deg cm2 gϪ1. H and 13C NMR spectra were recorded in
CDCl3 solution on a Bruker MSL 300 (300 MHz and 75.47
MHz, respectively). Chemical shifts are reported in δ from
SiMe4 as internal standard; coupling constants are given in Hz.
Light petroleum refers to the fraction with distillation range
60–90 ЊC.
Synthesis of substrates
Epoxides ( )-1–6 were obtained via epoxidation of the corres-
ponding alkene (17 mmol) in CH2Cl2 (50 cm3) using m-
chloroperbenzoic acid (70%; 1.2 mol equiv.) containing finely
powdered K2CO3 (2 mol equiv.) at 0 ЊC (14 h). After extractive
work-up (10% aq. Na2S2O5, then saturated aq. NaHCO3), the
organic layer was dried (Na2SO4) and evaporated, and the
products were distilled through a short Vigreux column; yields
ranged from 58 to 96%.
trans-2,3-Epoxyheptane 1. δH 0.88 (t, J 7.0, 3 H, 7-H3), 1.12
(d, J 6.0, 3 H, 1-H3), 1.21–1.44 (m, 6 H, 4-, 5- and 6-H2), 2.41
(dt, Jt 5.3, Jd 2.2, 1 H, 3-H) and 2.55 (dq, Jq 5.2, Jd 2.2, 1 H,
2-H); δC 13.73 (C-7), 17.43 (C-1), 22.40 (C-6), 28.03 (C-5),
31.62 (C-4), 54.23 (C-2) and 59.46 (C-3).
trans-3,4-Epoxyheptane 2. δH 0.82–0.93 (m, 6 H, 1- and 7-H3),
1.43–1.55 (m, 6 H, 2-, 5- and 6-H2) and 2.52–2.63 (m, 2 H, 3-
and 4-H); δC 9.73, 13.75, 19.25, 25.08, 34.06, 58.18 and 59.65.
trans-2,3-Epoxyoctane 3. δH 0.80 (t, J 7.0, 3 H, 8-H3), 1.16–
1.43 (m, 11 H, 1-H3, 4–7-H2), 2.52 (dt, Jt 5, Jd 2.0, 1 H, 3-H) and
2.64 (dq, Jq 6.0, Jd 2.0, 1 H, 2-H); δC 13.90, 17.61, 22.54 and
25.65 (C-5), 31.61, 31.97 (C-1 and -4), 54.45 (C-3) and 59.72
(C-2).
Finally, the applicability of the enantioconvergent biocat-
alytic hydrolysis was demonstrated with a preparative-scale
reaction. Thus, when substrate ( )-5 was treated with lyophil-
ized cells of Nocardia EH1 until the starting material had been
completely consumed, diol (R,R)-5a was isolated in 79% chem-
ical yield and 91% enantiomeric purity as the sole product.
Cases for a non-classical deracemization of epoxides, which
leads to the highly desired formation of a single product enan-
tiomer in 100% theoretical yield, are rare. For instance, the
hydrolysis of ( )-3,4-epoxytetrahydropyran23 and several cis-β-
alkyl-substituted styrene oxides24 by hepatic microsomal
epoxide hydrolase proceeded in an enantioconvergent manner
leading to the corresponding (R,R)-diols as the sole product.
Similarly, soybean epoxide hydrolase converted ( )-cis-9,10-
epoxyoctadec-12(Z)-enoic and ( )-cis-12,13-epoxydec-9(Z)-
enoic acid into the corresponding (R,R)-dihydroxy acids.25 All
of the above-mentioned transformations, however, were aimed
at the elucidation of enzyme mechanisms and were performed
only on an analytical (<1 mmol) scale. It was only recently that
preparative-scale hydrolysis of ( )-cis-β-methylstyrene oxide by
the fungus Beauveria bassiana9 was reported to afford (1R,2R)-
1-phenylpropane-1,2-diol in 85% yield and 98% ee. However,
in none of these studies were the kinetics of these systems
investigated.
trans-2,3-Epoxynonane 4. δH 0.80 (t, J 7.0, 3 H, 9-H3), 1.16–
1.48 (m, 13 H, 1-H3, 4–8-H2), 2.53 (dt, Jd 2.0, Jt 8.0, 1 H, 3-H)
and 2.64 (dq, Jd 2.0, Jq 5.0, 1 H, 2-H); δC 13.98 (C-9), 17.61,
22.53, 25.94, 29.09, 31.74, 32.02, 54.45 (C-2) and 59.72 (C-3).
cis-2,3-Epoxyheptane 5. δH 0.87 (t, J 7.0, 3 H, 7-H3), 1.21 (d,
J 6.0, 3 H, 1-H3), 1.25–1.52 (m, 6 H, 4–6-H2), 2.85 (dq, Jq 6.2,
Jd 2.0, 1 H, 2-H) and 2.98 (dt, Jt 5.5, Jd 2.0, 1 H, 3-H); δC 13.15
(C-7), 13.98 (C-1), 22.57 (C-6), 27.21 (C-5), 28.58 (C-4), 52.60
(C-2) and 57.1 (C-3).
cis-3,4-Epoxyheptane 6. δH 0.83–0.95 (m, 6 H, 1- and 7-H3),
1.32–1.52 (m, 6 H, 2-, 5- and 6-H2) and 2.71–2.86 (m, 2 H,
3- and 4-H); δC 10.47, 13.52, 19.86, 21.06, 29.67, 56.98 and
58.13.
Synthesis of reference material
Diols ( )-1a–6a were obtained via OsO4-catalysed dihydroxyl-
ation of the corresponding Z- or E-alkene (5 mmol) in acetone
(5 cm3) using N-methylmorpholine N-oxide (6 mmol) as reoxi-
dant (0 ЊC, 12 h). After extractive work-up (10% aq. Na2S2O5,
semisaturated aq. NH4Cl) the products were purified via silica
gel chromatography. Yields ranged from 54 to 80%.
erythro-Heptane-2,3-diol 1a. δH 0.85 (t, J 7.0, 3 H, 7-H3), 1.05
(d, J 7.0, 3 H, 1-H3), 1.22–1.50 (m, 6 H, 4–6-H2), 2.20–2.45 (s,
br, 2 H, OH) and 3.55–3.65 and 3.72–3.83 (m, 2 × 1 H, 2- and 3-
H); δC 13.99 (C-7), 16.25 (C-1), 22.73 (C-6), 28.32 (C-5), 31.51
(C-4), 70.44 (C-2) and 74.91 (C-3).
erythro-Heptane-3,4-diol 2a. δH 0.92 (t, J 7.0, 3 H, 7-H3), 0.96
(t, J 7.0, 3 H, 1-H3), 1.05–1.60 (m, 6 H, 2-, 5- and 6-H2) and
3.44–3.68 (m, 2 H, 3- and 4-H); δC 10.01 (C-7), 14.10 (C-1),
18.85 (C-6), 26.32 (C-2), 35.61 (C-5), 73.94 and 76.03 (C-3,
-4).
Experimental
Reactions were monitored by TLC (silica gel Merck 60 F254),
compounds were visualized by spraying with vanillin–conc.
H2SO4 (5 g dmϪ3) or Mo-reagent [(NH4)6Mo7O24ؒ4H2O (1.1 g
dmϪ3), Ce(SO4)2ؒ4H2O (4 g dmϪ3) in H2SO4 (10%)]. Prepar-
J. Chem. Soc., Perkin Trans. 1, 1997
3633

View Article Online
erythro-Octane-2,3-diol 3a. δH 0.80 (t, J 7.0, 3 H, 8-H3),
1.18 (d, J 7.0, 3 H, 1-H3), 1.05–1.50 (m, 8 H, 4–7-H2), 2.95 (s,
br, 2 H, OH), 3.64 (dt, Jt 4.0, Jd 2.0, 1 H, 3-H) and 3.80 (dq,
Jq 6.5, Jd 3.0, 1 H, 2-H); δC 14.10 (C-8), 16.61 (C-7), 22.64 (C-
1), 25.74 (C-4), 29.8 (C-5), 31.95 (C-6), 70.51 (C-3) and 75.02
(C-2).
erythro-Nonane-2,3-diol 4a. δH 0.84 (t, J 7.0, 3 H, 9-H3), 1.18
(t, J 7.0, 3 H, 1-H3), 1.02–1.65 (m, 10 H, 4–8-H2), 2.95 (s, br, 2 H,
OH), 3.64 (dt, Jt 4.0, Jd 2.0, 1 H, 3-H) and 3.80 (dq, Jq 6.5, Jd
3.0, 1 H, 2-H); δC 14.03 (C-9), 16.64 (C-8), 22.65 (C-1), 23.40
(C-7), 25.73 (C-4), 31.82 (C-5), 31.94 (C-6), 70.53 (C-3) and
75.01 (C-2).
threo-Heptane-2,3-diol 5a. δH 0.85 (t, J 7.2, 3 H, 7-H3), 1.22
(d, J 6.0, 3 H, 1-H3), 1.22–1.50 (m, 6 H, 4–6-H2), 2.85 (s, br, 1 H,
OH), 3.25 (s, br, 1 H, OH) and 3.20–3.35 and 3.47–3.61 (m,
2 × 1 H, 2- and 3-H); δC 14.04 (C-7), 19.45 (C-1), 22.75 (C-6),
27.79 (C-5), 32.99 (C-4), 70.93 (C-2) and 76.77 (C-3).
threo-Heptane-3,4-diol 6a. δH 0.91 (t, J 7.1, 3 H, 7-H3), 0.97 (t,
J 7.0, 3 H, 1-H3), 1.20–1.63 (m, 6 H, 2-, 5- and 6-H2) and 3.24–
3.51 (m, 2 H, 3- and 4-H); δC 10.50, 14.12, 19.21, 24.15, 33.24,
74.35 and 76.42.
removed, while the epoxide hydrolase remained still bound.
Finally, the epoxide hydrolase was desorbed by using further
volumes of deionized water (flow rate 10 cm3 minϪ1). The active
fractions were pooled and adjusted with Tris-HCl buffer (0.5 ,
pH 8.0) to yield a final buffer concentration of 0.05 . This
solution was then lyophilized to give a stable, partially purified
epoxide hydrolase preparation which was used for 18OH2-
labelling experiments.
General procedure for the asymmetric hydrolysis of epoxides
Lyophilized bacterial cells (90 mg) were rehydrated in Tris-HCl
buffer (0.05 , pH 8.0; 5 cm3) for 30 min on a rotary shaker (180
rpm) at ambient temperature. Substrate was then added (100
mm3) in one portion and the mixture was agitated at room
temp., while the reaction was monitored by TLC or GLC. The
specific activity of the reactions was in the range ~1000–2000
µmol mgϪ1 hϪ1 (with respect to pure protein). After an
appropriate degree of conversion was reached, acetone was
added (1 cm3) and the cells were removed by centrifugation
(12 000 g; 10 min). The organic material was extracted from
the supernatant with ethyl acetate, and the organic phase was
dried (Na2SO4) and carefully evaporated under partially
reduced pressure. The ee of epoxides and diols was determined
by GLC on a chiral stationary phase without further
purification. For optical rotation values of products having
an ee >50% see Table 6. Retention times for GLC on a chiral
stationary phase are given in Table 5.
Growth of microorganisms
Bacteria were obtained from culture collections except for the
Nocardia spp., which were a kind gift of J. de Bont (Wagenin-
gen, NL) and C. Syldatk (Stuttgart). The strains were grown in
shake-flask cultures at 30 ЊC on the following medium: yeast
extract (10 g dmϪ3), peptone (10 g dmϪ3), glucose (10 g dmϪ3),
NaCl (2 g dmϪ3), MgSO4ؒ7H2O (0.147 g dmϪ3), NaH2PO4 (1.3
g dmϪ3) and K2HPO4 (4.4 g dmϪ3). At the late exponential
growth phase (~25–40 h) the cells were harvested by centrifug-
ation (5000 g; 20 min; 25–30 g dmϪ3 wet cells), resuspended in
Tris-HCl buffer (0.05 , pH 8.0), centrifuged again and lyophil-
ized. Typical yields of dry cells ranged from ~3 to 5 g dmϪ3. The
cells could be stored over several months at ϩ5 ЊC without sig-
nificant loss of activity.
Preparative-scale reaction
Lyophilized whole cells of Nocardia EH1 (4 g) were rehydrated
in Tris-buffer (50 cm3; 50 m, pH 8.0) for 1 h before addition of
epoxide ( )-5 (0.82 g, 7.1 mmol). The mixture was agitated on
a rotary shaker at room temp. until the starting material was
completely consumed (~64 h). After addition of acetone (30
cm3), the mixture was centrifuged and the cells and the super-
natant were separately extracted with ethyl acetate (3×). The
combined organic layer was dried (Na2SO4), the volatiles were
evaporated off, and the residue was purified by silica gel chro-
matography to remove minor impurities and dyes which arose
from the cells, to furnish diol (R,R)-5a (0.75 g, 79%, ee 91%),
[α]D20 ϩ20.45 (c 1, EtOH).
Partially purified Nocardia EH1 epoxide hydrolase preparation
All steps of the protein purification were performed at ϩ4 ЊC.
Freshly grown wet cells were resuspended in Tris-HCl buffer
(0.05 , pH 8.0) and disintegrated using a glass bead mill (Vi-
brogen Zellmühle, shaking frequency 75 Hz; glass bead diam-
eter 0.35 mm; 8 min). After removal of the cell debris by centri-
fugation (12 000 g; 40 min), the supernatant was subjected to
hydrophobic interaction chromatography (Phenyl Sepharose
CL-4B, Sigma) using a 9.5 × 2.5 cm column. The enzyme
appeared very lipophilic since it was readily adsorbed onto the
stationary phase from 50 m Tris-buffer without addition of
(NH4)2SO4. By using a linear gradient of adsorption buffer
against deionized water, most of the adsorbed proteins could be
18O-Labelling studies
A sample of partially purified lyophilized epoxide hydrolase (10
mg) was rehydrated in 18OH2 (label content 95–97%; 120 mm3)
at 30 ЊC for 1 h. Then substrate (4 mm3) was added and shaking
was continued. Samples were extracted with ethyl acetate (120
mm3), the organic layer was dried (Na2SO4) and analysed by
GLC/MS.
Table 5 GLC data
Compound
Gas pressure (bar)
Temperature (T/ЊC)
Retention time (tR/min)
1
0.75
1.0
0.5
1.0
1.0
1.0
0.5
1.0
0.5
40
100
45
100
115
120
60
100
60
100
50
50
50
60
135
120
6.5 (2S,3S)/6.7 (2R,3R)
7.8 (2R,3S)/8.1 (2S,3R)
6.8 (3S,4S)/7.2 (3R,4R)
7.8 (3R,4S)/8.0 (3S,4R)
5.9 (2R,3S)/6.3 (2S,3R)
7.5 (2R,3S)/8.2 (2S,3R)
5.5 (2S,3R)/5.6 (2R,3S)
6.7 (2S,3S)/7.3 (2R,3R)
5.1 (3S,4R)/5.6 (3R,4S)
6.1 (3S,4S)/6.8 (3R,4R)
4.3 (S)/4.5 (R)
3.4 (S)/3.6 (R)
7.1 (S)/7.4 (R)
9.1 (S)/9.6 (R)
2.67 (2R,3S)/2.74 (2S,3R)
3.95 (2S,3S)/4.13 (2R,3R)
1a
2
2a
3a
4a
5
5a
6
6a
1.0
Heptan-2-yl trifluoroacetate
Heptan-3-yl trifluoroacetate
Octan-2-yl trifluoroacetate
Nonan-2-yl trifluoroacetate
1a
0.75
0.75
1.0
1.0
1.9b
2.4b
5a
a GLC/MS analysis. b Flow rate (cm3 minϪ1).
3634
J. Chem. Soc., Perkin Trans. 1, 1997

View Article Online
Table 6 Optical rotation valuesa
References and notes
1 H. C. Kolb, M. S. VanNieuwenhze and K. B. Sharpless, Chem.
Rev., 1994, 94, 2483.
Specific
Compound
rotation [α]D20
ee (%)
(c, Solvent)
2 V. Schurig and F. Betschinger, Chem. Rev., 1992, 92, 873.
3 R. A. Johnson and K. B. Sharpless, in Catalytic Asymmetric
Synthesis, ed. I. Ojima, Verlag Chemie, New York, 1993, p. 103;
E. N. Jacobsen, W. Zhang, A. R. Muci, J. R. Ecker and L. Deng,
J. Am. Chem. Soc., 1991, 113, 7063; K. Konishi, K. Oda, K. Nishida,
T. Aida and S. Inoue, J. Am. Chem. Soc., 1992, 114, 1313.
4 J. A. M. de Bont, Tetrahedron: Asymmetry, 1993, 4, 1331; D. J. Leak,
P. J. Aikens and M. Seyed-Mahmoudian, Trends Biotechnol.,
1992, 10, 256; A. N. Onumonu, A. Colocoussi, C. Matthews,
M. P. Woodland and D. J. Leak, Biocatalysis, 1994, 10, 211; P. Besse
and H. Veschambre, Tetrahedron, 1994, 50, 8885; S. Pedragosa-
Moreau, A. Archelas and R. Furstoss, Bull. Soc. Chim. Fr., 1995,
132, 769.
(2R,3S)-1a
(3R,4S)-2a
(2R,3S)-5
(2R,3R)-5a
(3R,4R)-6a
Ϫ23.9
Ϫ8.8
ϩ0.7
ϩ21.8
ϩ8.5
82
52
83
97
72
0.67, EtOH
0.17, EtOH
0.3, Et2O
0.66, EtOH
0.8, EtOH
a Compounds having an ee of <50% were not measured.
Determination of absolute configuration of epoxides
The remaining non-hydrolysed epoxide from the biotransform-
ation was separated from the formed diol by column chrom-
atography (light petroleum–ethyl acetate 1:1). The epoxide was
reduced with LiAlH4 (0.3 g, 8 mmol) in diethyl ether (4 cm3) at
ambient temp. After the suspension had been stirred at room
temp. for 16 h, the reaction was quenched by addition of satur-
ated aq. NH4Cl. The secondary alcohols thus obtained were
extracted with ethyl acetate and dried (Na2SO4). For GLC
analysis, the samples were transformed into the corresponding
trifluoroacetates (CH2Cl2, pyridine 1.25 mol equiv., trifluoro-
acetic anhydride, 1.25 mol equiv., room temp., 16 h). The organic
layer was washed with 2  HCl, dried (Na2SO4), and the abso-
lute configuration of trifluoroacetates was determined via co-
injection on GLC using independently synthesized reference
material obtained from commercially available (R)-heptan-2-ol.
Enantiomerically enriched reference material for heptan-3-ol
was synthesized via reduction epoxide (2R,3S)-5, obtained
from the biocatalytic hydrolysis of epoxide ( )-5 using Nocar-
dia EH1 (Table 2, entry 26).
5 F. Oesch, Xenobiotica, 1972, 3, 305.
6 S. Pedragosa-Moreau, A. Archelas and R. Furstoss, Tetrahedron,
1996, 52, 4593.
7 S. Pedragosa-Moreau, A. Archelas and R. Furstoss, J. Org. Chem.,
1993, 58, 5533.
8 (a) M. Mischitz, W. Kroutil, U. Wandel and K. Faber, Tetrahedron:
Asymmetry, 1995, 6, 1261; (b) I. Osprian, W. Kroutil, M. Mischitz
and K. Faber, Tetrahedron: Asymmetry, 1997, 8, 65.
9 (a) For a review see: K. Faber, M. Mischitz and W. Kroutil, Acta
Chem. Scand., 1996, 50, 249; (b) For the isolation and character-
ization of a highly selective bacterial epoxide hydrolase see:
M. Mischitz, K. Faber and A. Willetts, Biotechnol. Lett., 1995, 17,
893.
10 M. Mischitz, C. Mirtl, R. Saf and K. Faber, Tetrahedron:
Asymmetry, 1996, 7, 2041.
11 C.-S. Chen, Y. Fujimoto, G. Girdaukas and C. J. Sih, J. Am. Chem.
Soc., 1982, 104, 7294.
12 For a preliminary communication see: W. Kroutil, M. Mischitz,
P. Plachota and K. Faber, Tetrahedron Lett., 1996, 37, 8379.
13 W. Kroutil, I. Osprian, M. Mischitz and K. Faber, Synthesis, 1997,
156.
14 G. M. Lacourciere and R. N. Armstrong, J. Am. Chem. Soc., 1993,
115, 10 466; B. D. Hammock, F. Pinot, J. K. Beetham, D. F. Grant,
M. E. Arand and F. Oesch, Biochem. Biophys. Res. Commun., 1994,
198, 850; M. Arand, H. Wagner and F. Oesch, J. Biol. Chem., 1996,
271, 4223.
15 Obtained via OsO4-dihydroxylation of the corresponding Z- and E-
alkene.
16 For reasons of clarity, the following nomenclature is used
throughout this paper: epoxide and diol enantiomers are denoted as
A/B and P/Q, respectively; the position of the oxygen incorporation
in diols P and Q (as determined by 18O-labelling experiments) is
indicated by subscript arabic numerals.
17 Full details of the isolation and characterization of the epoxide
hydrolase from Nocardia EH1 will be reported elsewhere.
Preliminary data indicate that the enzyme is related to the epoxide
hydrolase from Rhodococcus sp. NCIMB 11216, i.e. a constitutive,
soluble protein having a relative molecular mass of ~35 kDa. See ref.
9b.
18 A related approach was elaborated by Kagan and co-workers, which
allows the determination of the quotients k1/k2 and k3/k4, whereas
other quotients of kis can only be determined for t approaching zero.
As a consequence, preparatively useful applications cannot be
sufficiently treated. By using our approach, all quotients of kis can
be determined at any ti during the whole course of the reaction. See:
S. El-Baba, J.-C. Poulin and H. B. Kagan, Tetrahedron, 1984, 40,
4275.
19 The absence of spontaneous hydrolysis under the reaction
conditions employed was verified using blank-experiments in the
absence of biocatalyst. No trace of hydrolysis was observed within
~600 h.
Determination of absolute configuration of diols
To a Grignard reagent, prepared from the corresponding n-
alkyl bromide (1.16 g, 8.5 mmol) and magnesium (0.205 g, 8.5
mmol) in tetrahydrofuran (20 cm3), was added ethyl (R)-lactate
or ethyl (R)-2-hydroxybutanoate26 (4.2 mmol), dropwise (room
temp.; N2). After the solution had been stirred for 5 h at room
temp., the reaction was quenched by addition of saturated aq.
NH4Cl. The phases were separated, the aqueous solution was
extracted with diethyl ether (2×), the combined organic phases
were dried (Na2SO4), and the volatiles were evaporated off to
give ~90% of crude α-hydroxy ketone 7a–d. The latter were
reduced without further purification, using LiAlH4 (70 mg, 1.8
mmol) in diethyl ether (10 cm3) at ambient temp. The suspen-
sion was stirred at room temp. for 18 h and the reaction was
quenched by addition of saturated aq. NH4Cl. The diastereo-
meric mixtures of diols (2R,3RS)-1a, -3a, -4a, -5a and (3R,
4RS)-2a, -6a were extracted with ethyl acetate and dried
(Na2SO4). The relative configuration of diols was determined
via co-injection with racemic threo- and erythro-diols (obtained
via OsO4-catalysed dihydroxylation of the corresponding Z-
and E-alkene). The absolute configuration was elucidated from
the (R)-lactate or (R)-2-hydroxybutanoate, respectively, used
for the synthesis.
20 The GLC/MS data from the 18O-experiments had to be subjected to
background correction for the following reason: the fragmentation
pattern of (unlabelled) heptane-2,3-diol shows not only a major peak
with m/z 87, but also a minor signal with m/z 85 which corresponds
to [(m/z) Ϫ 2 H]. Since the major m/z-peak from the 16O-fragment
(87) coincides with the minor [(m/z) Ϫ 2 H] signal from the 18O-
fragment (87), this signal was subjected to background correction by
comparison with the spectra obtained from unlabelled material.
The corrected values thus obtained accurately describe the
regioselectivity of the 18O-incorporation. Since the MS data
obtained correspond to the average of scans, but not absolute values,
the peak integration values were corrected with an additional factor
( f ) via calibration with the (absolute) values from the enantiomeric
composition obtained via GLC. PGC, QGC correspond to amounts
Acknowledgements
We express our cordial thanks to I. Osprian, A. Steinreiber and
P. Plachota (Graz) for their valuable technical assistance and
J. de Bont (Wageningen, NL) and C. Syldatk (Stuttgart) for the
kind donation of Nocardia strains. This study was performed
within the Spezialforschungsbereich ‘Biokatalyse’ and was
financed by the Fonds zur Förderung der wissenschaftlichen
Forschung (project no. F 104), the Austrian Federal Ministry
of Science (Vienna) and the European Commission (project
BIO4-CT95-0005).
J. Chem. Soc., Perkin Trans. 1, 1997
3635

View Article Online
measured by areas, whereas PMS, QMS stand for amounts via average
counts. f = (QGCؒPMS/PGCؒQMS).
21 Since the starting material is racemic, A0 equals B0.
22 E corresponds to the quotient of the reaction rates of the two
enantiomers (k3 ϩ k4/k1 ϩ k2).
25 E. Blée and F. Schuber, Eur. J. Biochem., 1995, 230, 229.
26 Obtained via transesterification of commercially available tert-butyl
(R)-2-hydroxybutanoate using catalytic NaOEt in ethanol (room
temp.; 24 h).
23 G. Bellucci, G. Berti, G. Catelani and E. Mastrorilli, J. Org. Chem.,
1981, 46, 5148.
24 G. Bellucci, C. Chiappe and A. Cordoni, Tetrahedron: Asymmetry,
1996, 7, 197.
Paper 7/04812B
Received 7th July 1997
Accepted 8th August 1997
3636
J. Chem. Soc., Perkin Trans. 1, 1997