Selectivity enhancement of enantio- and stereo-complementary epoxide hydrolases and chemo-enzymatic deracemization of (±)-2-methylglycidyl benzyl ether

Article abstract of DOI:10.1016/j.tetasy.2005.12.018

The kinetic resolution of (±)-2-methylglycidyl benzyl ether was achieved via enantioselective biohydrolysis using microbial and plant epoxide hydrolases. Depending on the type of enzyme, opposite enantiopreference and stereo-complementary mode of action (

Full text of DOI:10.1016/j.tetasy.2005.12.018

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 17 (2006) 402–409
Selectivity enhancement of enantio- and
stereo-complementary epoxide hydrolases and chemo-enzymatic
deracemization of ( )-2-methylglycidyl benzyl ether
*
´
Yolanda Simeo and Kurt Faber
Department of Chemistry, Organic and Bioorganic Chemistry, University of Graz, Heinrichstrasse 28, Graz A-8010, Austria
Received 12 December 2005; accepted 19 December 2005
Available online 23 January 2006
Abstract—The kinetic resolution of ( )-2-methylglycidyl benzyl ether was achieved via enantioselective biohydrolysis using micro-
bial and plant epoxide hydrolases. Depending on the type of enzyme, opposite enantiopreference and stereo-complementary mode
of action (i.e., retention vs inversion of configuration) led to hetero- and homochiral product mixtures. Optimization of the reaction
conditions for Rhodococcus sp. R312 led to significantly enhanced enantioselectivity (E >200), which enabled the deracemization of
( )-2-methylglycidyl benzyl ether via biohydrolysis (proceeding with retention of configuration) followed by inverting acid-catalyzed
hydrolysis to furnish (R)-1-benzyloxy-2-methylpropane-2,3-diol in >97% ee and 78% yield from the racemate.
ꢀ 2006 Elsevier Ltd. All rights reserved.
1. Introduction
Alternatively, modification of an E- or Z-carbon–car-
bon double or triple bond was investigated,4 as well as
the adjustment of a nitro group regarding its relative
position on an aromatic system.5
Enantiopure epoxides and their corresponding vicinal
diols are important building blocks for the synthesis of
numerous bioactive molecules. In particular, 2,2-disub-
stituted oxiranes and the corresponding vic-diols pro-
vide easy access to sterically demanding synthons
containing a fully substituted carbon atom resembling
a chiral tert-alcohol moiety, which is difficult to obtain
in its non-racemic form by conventional means.1 In this
context, enantioselective biocatalytic hydrolysis of rac-
oxiranes using epoxide hydrolases of microbial origin
has proven to be a powerful tool for obtaining these
compounds.2
On the other hand, a general limitation in the biohydroly-
sis of epoxides is the poor solubility of the lipophilic
oxiranes in aqueous buffer systems, causing a (apparent)
low substrate concentration. One option to overcome
this problem is the use of water-miscible organic co-sol-
vents, as demonstrated by Furstoss et al. for the resolu-
tion of p-nitrostyrene oxide used for the synthesis of the
6
´
b-blocker (R)-Nifenalol. Subsequent studies revealed a
strong influence of co-solvent concentration on enzyme
activity using water-miscible organic solvents.7,8 Like-
wise, the effects of two different water-soluble co-
solvents on the reaction rate and selectivity of 1,2-
epoxyoctane hydrolysis have been analyzed.9 However,
this procedure is not applicable to bacterial epoxide
hydrolases, as the addition of even small amounts of
water-soluble organic co-solvents led to rapid enzyme
deactivation.10
Since the presence of multiple (competing) enzymes in
whole-cell biocatalysts is a common phenomenon,
enantioselectivities in microbial epoxide hydrolysis are
often less than perfect and selectivity enhancement is
required. In this context, several procedures have been
developed: substrate modification by variation of pro-
tective groups has shown to be a highly valuable tool
for modulating enzyme selectivity on a range of 2,2-
disubstituted oxiranes bearing an ether-functionality.3
However, the use of high substrate concentrations
(where the substrate makes up a second ‘organic’ phase)
has proven to increase reaction rates and enantioselec-
tivities.11,12 Instead of using large amounts of substrate,
water-immiscible organic co-solvents have been used as
*
Corresponding author. Tel.: +43 316 380 5332; fax: +43 316 380
9840; e-mail: Kurt.Faber@uni-graz.at
0957-4166/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2005.12.018

Y. Simeo´, K. Faber / Tetrahedron: Asymmetry 17 (2006) 402–409
403
the second liquid phase. For instance, the enantioselec-
tivity in the resolution of glycidyl phenyl ether was
improved by using a biphasic i-octane/aqueous system13
and the biohydrolysis of rac-epichlorohydrin has been
accomplished in cyclohexane supplemented with a small
amount of water (2.0%, v/v) in the absence of a discrete
second aqueous buffer phase.14
During our studies toward an enantio-convergent che-
mo-enzymatic asymmetric total synthesis of (all-R)-a-
tocopherol (vitamin E), we required a reliable and
scalable protocol for the preparation of enantiopure 2-
methylglycidyl benzyl ether. Initial attempts to apply
our previously developed procedure based on the kinetic
resolution of ( )-1 to a deracemization protocol failed
due to incomplete conversion (i.e., at or slightly beyond
50%), which is critical for obtaining an optimal eeP(final)
in an enantio-convergent process.6,25
However, the benefit of two-phase solvent systems can-
not be generalized, since the hydrolysis of rac-1-{20,30-
dihydrobenzo[b]furan-40-yl}-1,2-oxirane using 10% of
methyl tert-butyl ether led to an improvement with Rho-
dotorula glutinis, but worse results were obtained with
Aspergillus niger.15 Although the exact details on the
effects of the biphasic systems exert on epoxide hydrolase
activity and -selectivity are not fully understood, inter-
facial activation (in case of Agrobacterium radiobacter
epoxide hydrolase)16 and mass transfer limitations17
seem to play a major role. In an attempt to correlate
the physicochemical properties of organic solvents and
their biocompatibility with epoxide hydrolase activity
of whole yeast cells, 40 water-miscible and -immiscible
solvents have been tested with Rhodotorula sp. UOFS
Y-0448. Among the water-miscible solvents, mono-alco-
hols (in particular 1-octanol) were considered most bio-
compatible, but no clear conclusions could be drawn for
biphasic solvent systems.18 Most recently, the biohydro-
lysis of rac-epoxides has also been performed in ionic
liquids, with reaction rates and E-values being compar-
able to those observed in aqueous buffer systems.19
2. Results and discussion
From a screen for epoxide hydrolase activity including
34 bacteria, fungi, and yeasts using rac-methylglycidyl
benzyl ether as substrate (data not shown), three bacte-
ria and one red yeast were selected as hits for further
selectivity-enhancement studies (Table 1).
In line with previous studies, Rhodococcus spp. showed
an (S)-enantiopreference with promising enantioselectiv-
ities ranging from moderate to good (E up to 71). On
the other hand, the enantioselectivity of the red yeast
Rhodotorula mucilaginosa DSM 70404 was comparably
low (E = 8.6), although the strain showed a rare oppo-
site enantiopreference for (R)-1 and was, therefore,
included in further studies. All strains exerted a
retention of configuration by directing the (formal)
nucleophile [OHꢀ] onto the less hindered unsubstituted
oxirane carbon atom during biohydrolysis, thus render-
ing hetero-chiral products (R)-1 and (R)-2 or (S)-1 and
(S)-2, respectively; note the switch in CIP sequence
priority from epoxide 1 to diol 2.
A rather puzzling picture emerged with regard to the
effects of detergents on the reaction rate and enantioselec-
tivity in the biohydrolysis of epoxides using R. glutinis.
Overall, the influence of detergents seemed to be depen-
dent on the substrate type, since the action of a certain
detergent on aliphatic and aromatic substrates was
found to be opposite.20
All attempts to improve the modest enantioselectivity of
R. mucilaginosa DSM 70404 by variation of the reaction
conditions proved disappointing (Table 2): In contrast
to the trends generally observed with epoxide hydrolases
from bacteria, fungi, and red yeasts,10–12 elevated
substrate concentration (entry 2) and low temperature
(entry 3) had negative effects on the enantioselectivity.
Enhanced enantioselectivities were also achieved by
using enzyme mutants of A. radiobacter AD1 epoxide
hydrolase.21,22 Likewise, selectivity enhancement by
enzyme modification was reported for A. niger23 and
R. glutinis24 with varying success. Finally, rather simple
selectivity enhancement was achieved by performing the
biohydrolysis of an epoxide at 4 ꢁC.11
More encouraging results were obtained for the selectiv-
ity enhancement using whole bacterial cells. Alth-
ough Rhodococcus ruber DSM 44540 gave a promising
Table 1. Enantioselectivities from the biohydrolysis of ( )-1 using selected microorganismsa
Microorganism
Conversion (%)
Selectivity E
Preferred enantiomer
Rhodococcus sp. CBS 717.73
16
15
8
16
37
71
8.6
(S)
(S)
(S)
(R)
Rhodococcus ruber DSM 44540
Rhodococcus ruber DSM 44190
Rhodotorula mucilaginosa DSM 70404
22
a Reaction conditions: 30 ꢁC, substrate concentration 168 mM, catalyst/substrate ratio 2:1 (w/w).
Table 2. Enantioselectivities from the biohydrolysis of ( )-1 using Rhodotorula mucilaginosa DSM 70404
Entry
Temp (ꢁC)
Substrate concn (mM)
Catalyst/substrate (w/w)
Conversion (%)
eeS (%)
eeP (%)
Selectivity E
1
2
3
30
30
4
28
168
168
10:1
2:1
2:1
71
22
36
62
2
8
13
7
14
8.6
1.2
1.4

404
Y. Simeo´, K. Faber / Tetrahedron: Asymmetry 17 (2006) 402–409
Table 3. Enantioselectivities from the biohydrolysis of ( )-1 using Rhodococcus ruber DSM 44540
Entry
Temp (ꢁC)
Substrate concn (mM)
Catalyst/substrate (w/w)
Conversion (%)
eeS (%)
eeP (%)
Selectivity E
1
2
3
4
5
30
30
30
4
28
168
786
168
786
10:1
2:1
2:1
2:1
2:1
60
15
3
8
10
95
76
3
8
11
63
94
>97
87
15
37
67
15
73
4
>97
enantioselectivity of E = 37 during the screening, scale-
up experiments using an enhanced catalyst/substrate
ratio of 10:1 (w/w) gave a lower value (E = 15, Table 3).
However, when the substrate concentration was gradu-
ally enhanced, the selectivity climbed to an E-value of
67 (entries 2 and 3). In addition, lowering the temperature
to 4 ꢁC (entries 4 and 5) had an additional beneficial
effect, which led to a final improvement of E = 73.
although a value of 168 mM showed a clear maximum,
which dropped off again at 768 mM (entries 2 and 3). A
strong positive effect was obtained by lowering the tem-
perature to 4 ꢁC (E = 70, entry 4), which largely com-
pensated for the selectivity decrease at high substrate
concentrations (768 mM). However, this went in hand
with a strong decrease in the reaction rate and the con-
version of 7% thus achieved within a reasonable reaction
time of 24 h was by far insufficient for a deracemization
protocol.
Encouraged by these results, the effect of organic co-sol-
vents was studied using a substrate concentration of
168 mM as standard (Table 4). Whereas water-miscible
DMSO had a negative effect (E = 22, entry 2), lipophilic
water-immiscible co-solvents, such as toluene (entry 3)
and hydrocarbons (entries 4–7) were generally more
suitable. Most striking was the difference between vari-
ous hydrocarbons, where straight-chain n-alkanes (n-
hexane, n-octane, entries 6 and 7) proved to be more
suitable than branched or cyclic analogues (cyclohexane,
i-octane, entries 4 and 5), which is contrary to the trends
observed with lipases.26 Among ethers, i-Pr2O was far
superior than t-BuOMe (entries 8 and 9).
Table 6. Enantioselectivities from the biohydrolysis of ( )-1 using
Rhodococcus ruber DSM 44190 in the presence of organic co-solvents
Entry
Co-solventa
Conversion
(%)
eeS
(%)
eeP
(%)
Selectivity
E
1
2
3
4
5
6
7
8
9
None
DMSO
t-BuOMe
i-Pr2O
Toluene
Cyclohexane
i-Octane
n-Octane
n-Hexane
8
42
3
8
1
25
31
29
44
8
56
2
>97
77
69
81
>97
74
80
88
94
71
13
5.6
10
66
8.5
12
22
73
7
<1
25
36
36
75
Table 4. Enantioselectivities from the biohydrolysis of ( )-1 using
Rhodococcus ruber DSM 44540 in the presence of organic co-solvents
a Aqueous/organic solvent 5:1 (v/v), except for dimethyl sulfoxide,
which was used in a proportion of 20:1; standard conditions 168 mM
substrate, catalyst/substrate 2:1 (w/w), 30 ꢁC.
Entry
Co-solventa
Conversion
(%)
eeS
(%)
eeP
(%)
Selectivity
E
1
2
3
4
5
6
7
8
9
None
DMSO
15
41
3
11
24
26
27
1
76
60
3
11
30
33
36
1
94
85
>97
90
90
95
93
85
>97
37
22
67
21
36
53
69
12
71
Toluene
Cyclohexane
i-Octane
n-Hexane
n-Octane
t-BuOMe
i-Pr2O
Co-solvent effects on the epoxide hydrolase activity of
R. ruber DSM 44190 (Table 6) largely paralleled that
of R. ruber DSM 44540. Again water-miscible DMSO
lowered the E-value of 71 under standard conditions
dramatically (E = 13, entry 2), similar to ethers (entries
3 and 4). No negative effects were observed in the pres-
ence of toluene (E = 66, entry 5). Likewise, hydrocar-
bons exhibited a familiar trend, that branched or
cyclic solvents were less suitable than n-alkanes, such
as n-octane or n-hexane (entries 6–9). In the latter case,
the reaction rate was not markedly diminished.
8
8
a Aqueous/organic solvent 5:1 (v/v) except for dimethyl sulfoxide,
which was used in a proportion of 20:1; standard conditions 168 mM
substrate, catalyst/substrate 2:1 (w/w), 30 ꢁC.
Related dependencies were observed for R. ruber DSM
44190 (Tables 5 and 6). Initially modest selectivities at
an enhanced catalyst/substrate ratio (E = 8.6) could be
improved by enhancing the substrate concentration,
Overall, the epoxide hydrolase system of whole cells of
Rhodococcus sp. CBS 717.7327 proved to be most flexible
in terms of selectivity enhancement (Tables 7 and 8).
Table 5. Enantioselectivities from the biohydrolysis of ( )-1 using Rhodococcus ruber DSM 44190
Entry
Temp (ꢁC)
Substrate concn (mM)
Catalyst/substrate (w/w)
Conversion (%)
eeS (%)
eeP (%)
Selectivity E
1
2
3
4
30
30
30
4
28
168
786
786
10:1
2:1
2:1
2:1
71
8
20
7
97
8
21
7
40
>97
84
8.6
71
14
>97
70

Y. Simeo´, K. Faber / Tetrahedron: Asymmetry 17 (2006) 402–409
405
Table 7. Enantioselectivities from the biohydrolysis of ( )-1 using Rhodococcus sp. CBS 717.73
Entry
Temp (ꢁC)
Substrate concn (mM)
Catalyst/substrate (w/w)
Conversion (%)
eeS (%)
eeP (%)
Selectivity E
1
2
3
4
5
30
30
30
4
28
168
786
168
786
10:1
2:1
2:1
2:1
2:1
53
16
6
7
7
>97
16
6
6
7
85
87
>97
82
51
16
69
10
70
4
>97
Table 8. Enantioselectivities from the biohydrolysis of ( )-1 using Rhodococcus sp. CBS 717.73 in the presence of organic co-solvents
Entry
Catalyst/substrate (w/w)
Co-solventa
Conversion (%)
eeS (%)
eeP (%)
Selectivity E
1
2
3
4
5
6
7
8
2:1
2:1
2:1
2:1
2:1
2:1
2:1
2:1
2:1
10:1
10:1
10:1
10:1
10:1
None
DMSO
t-BuOMe
i-Pr2O
Toluene
n-Hexane
Cyclohexane
n-Octane
i-Octane
n-Hexane
Toluene
i-Pr2O
16
27
9
16
34
7
23
37
61
42
66
73
>97
49
83
>97
>97
87
90
67
>97
>97
96
16
26
5.4
82
94
91
19
28
39
31
40
43
52
34
46
51
50
92
97
36
130
144
79
106
171
100
>200
9
>97
90
>97
>97
92
10
11
12
13
14
n-Octane
i-Octane
>97
a Aqueous/organic solvent 5:1 (v/v) except for dimethyl sulfoxide, which was used in a proportion of 20:1; 30 ꢁC.
Although initial selectivities at a high catalyst loading
(10:1) and low substrate concentration (28 mM) were
rather good (entry 1, E = 51), reduced values resulted
when a more realistic catalyst/substrate ratio of 2:1
was used (E = 16, entry 2). However, elevated substrate
concentration and lower temperatures (entries 3–5) com-
pensated these effects significantly (Emax = 70). In any
case, a significant drop in reaction rate was observed
again (entries 2–4).
In order to increase reaction rates, higher catalyst load-
ings were investigated (entries 10–14). Again, biphasic
systems consisting of C8-hydrocarbons proved to be
best, where the critical conversion of 50% could be
reached while maintaining excellent enantioselectivities
(E >200, entry 14).
In order to avoid the use of whole-cell systems contain-
ing an unknown number of enzymes, the use of cloned
and overexpressed epoxide hydrolases from plant and
microbial origin was investigated (Table 9). Soybean
epoxide hydrolase gave a modest selectivity (E = 14) at
high catalyst loading (2:1, entry 1); when this value
was decreased to 1:10, no reaction was observed. In con-
trast, the potato-enzyme was active at both catalyst/sub-
strate ratios, albeit at low enantioselectivities (entries 2
and 3; E <10). With respect to the stereochemical out-
come, both plant enzymes showed an opposite enantio-
preference, that is, (R) for soybean epoxide hydrolase
and (S) for the enzyme from potato. Both plant
enzymes, however, seemed to act through the same
The effects of organic co-solvents somewhat differed
from those observed with the previous Rhodococci
(Table 8). First, a weak selectivity enhancement was
observed in the presence of DMSO (entry 2). The ethers
behaved rather differently, as t-BuOMe proved to be
unsuitable in contrast to i-Pr2O (entries 3 and 4). Again,
toluene and hydrocarbons were best. In this case, the
difference between cyclic/branched and linear solvents
was small, but were more profound regarding their
molecular weight, as n- and i-octane gave superior
results than hexanes.
Table 9. Enantioselectivities from the biohydrolysis of ( )-1 using cloned epoxide hydrolases
Entry
Epoxide hydrolase
Catalyst/substrate (w/w)
Conversion (%)
eeS (%)
eeP (%)
Selectivity E
Enantiopreference
1
2
3
4
5
Soybeana
2:1d
1:5
1:10
1:5
50
98
82
86
75
74
>97
>97
>97
60
73
2
22
16
20
14
(R)
(S)
(S)
(R)
(R)
Potatob
2.3
5.1
4.3
2.5
Potatob
Rhodococcus erythropolisc
Rhodococcus erythropolisc
1:10
a Glycine max, expressed in Saccharomyces cerevisae WA6.29
b Solanum tuberosum, expressed in Spodoptera frugiperda.30
Limonene 1,2-epoxide hydrolase from Rhodococcus erythropolis, expressed in E. coli.31
c
d No reaction was observed at a catalyst/substrate ratio of 1:10.

406
Y. Simeo´, K. Faber / Tetrahedron: Asymmetry 17 (2006) 402–409
mechanism, that is with retention of configuration. A
possible explanation for the low enantioselectivities of
plant epoxide hydrolases might be the fact that the
natural substrates of these enzymes are epoxy-fatty
acids,28 which have a rather unrelated structure to that
of substrate 1.
mechanism, which is commonly observed in acid-catal-
yzed hydrolysis of epoxides, which leads to incorpora-
tion of the nucleophile at the higher substituted
oxirane carbon atom (Scheme 2, bottom).38,39 In this
case, the latter bears a stereogenic center (as in the case
of 1), with inversion of configuration observed.
Surprising results were obtained with limonene 1,2-
epoxide hydrolase from Rhodococcus erythropolis31 (en-
tries 4 and 5). Although this enzyme exhibited only low
enantioselectivities with a preference for (R)-1, it formed
diol (R)-2, which suggests that it acts through an inver-
sion of configuration. These observations can be
explained by taking the molecular mechanism of epoxide
hydrolases into consideration (Scheme 2). The majority
of enzymes investigated so far [e.g., from A. radio-
bacter34 and A. niger35] act through clear SN2-type dis-
placement of the oxirane oxygen atom by (formal)
hydroxyl ion. The latter is realized by nucleophilic
attack of an Asp-anion within the active site of the
enzyme, leading to the formation of a covalent glycol-
hemiester intermediate,32 which in turn is hydrolyzed
by [OHꢀ] provided by an adjacent histidine-acidic-resi-
due pair through proton-abstraction from water. Tyro-
sines positioned at the opposite side promote the ring
opening of the oxirane by delivering a proton for the
epoxy-oxygen. Overall, this mechanism represents a
clear SN2-type mechanism (Scheme 2, top).33–35 Since
nucleophilic attack at the fully substituted oxirane car-
bon atom of 1 bearing the chiral center is impossible
for steric reasons, the [OHꢀ] is inserted at the unsubsti-
tuted carbon atom, which results in retention of
configuration.
In order to obtain a single stereoisomer in 100% theoreti-
cal yield, an enantio-convergent process was envisaged
as follows (Scheme 1): biohydrolysis of rac-1 using
lyophilized whole cells of Rhodococcus sp. CBS 717.73
under optimized conditions (catalyst/substrate 10:1,
buffer/i-octane, 30 ꢁC) proceeded with (S)-enantiopre-
ference acting through retention of configuration to ren-
der the corresponding (R)-diol 2 (note the switch in CIP
sequence priority) and remaining non-converted (R)-1.
Acid-catalyzed hydrolysis of the latter in dioxane/water
proceeded with inversion of configuration to furnish (R)-
2 as the sole product in >97% ee and 78% yield.39 Finally,
cyclization of (R)-2 via tosylation of the prim-hydroxy
moiety followed by treatment of the corresponding
mono-tosylate with base furnished (S)-1 as the sole prod-
uct in 92% yield.
3. Conclusion
Kinetic resolution of ( )-2-methylglycidyl benzyl ether
was accomplished via biohydrolysis using whole micro-
bial cells or cloned epoxide hydrolases. While the abso-
lute configuration was retained during biohydrolysis for
the majority of biocatalysts, an opposite enantioprefer-
ence was observed depending on the type of enzyme:
(S) for Rhodococcus spp. and potato epoxide hydrolase
versus (R) for Rhodotorula sp. and soybean epoxide
hydrolase. In contrast, limonene epoxide hydrolase
acted with an inversion of configuration, thus producing
a homochiral product mixture. Optimization of the reac-
tion conditions with respect to temperature, substrate
concentration, catalyst loading, and the use of aque-
ous/organic media led to significantly enhanced enantio-
selectivity for Rhodococcus sp. R312 (E >200), which
enabled the deracemization of ( )-2-methylglycidyl ben-
zyl ether via biohydrolysis (proceeding with retention of
In contrast, limonene 1,2-epoxide hydrolase from
R. erythropolis was shown to possess a different mode of
action.36,37 Since its natural substrate (limonene 1,2-epox-
ide) is sterically very demanding, SN2-type displacement
would be energetically extremely unfavorable. As a con-
sequence, this enzyme acts via a single push–pull step
through simultaneous O-protonation and nucleophilic
attack by [OHꢀ] without the formation of a covalent
enzyme–substrate intermediate. Overall, this mechanism
bears a strong resemblance to a borderline-SN2-type
O
Epoxide Hydrolase
from
Epoxide Hydrolase
O
O-Bn
O-Bn
from
H+ cat.
Dioxane/
H2O
Soybean
Potato
Rhodococcus sp.
(S)-1
+
(R)-1
O
Rhodotorula sp.
O
+
O-Bn
rac-1
O-Bn
HO
HO
aq. Buffer
aq. Buffer
(S)-1
sole product
HO
HO
O-Bn
O-Bn
(R)-Preference
Retention
(S)-Preference
Retention
TsCl/Et3N/
Me3NH+Cl- then
NaH/THF
(S)-2*
Hetero-Chiral*
(R)-2*
Limonene
Hetero-Chiral*
(R)-Preference
Inversion
Epoxide Hydrolase
aq. Buffer
Bn = CH2-Ph
HO
O
HO
O-Bn
O-Bn
+
* switch in CIP sequence priority
(S)-1
(R)-2*
Homo-Chiral*
Scheme 1. Stereo- and enantio-complementary kinetic resolution and chemo-enzymatic deracemization of ( )-2-methylglycidyl benzyl ether (1).

Y. Simeo´, K. Faber / Tetrahedron: Asymmetry 17 (2006) 402–409
407
SN2-type mechanism
Tyr
HO
HO
R
*
R
HO
O
*
Nucleophilic attack
(Retention)
Hydrolysis
Tyr
OH
O
O
O
R
H
HO
OH
*
O
H
O
Asp
O-
N
His
N
H
Asp
Asp
Glycol-hemiester
Enzyme-substrate intermediate
Borderline-SN2-type mechanism
Asp
Asp
O-
O
OH
O
HO
O
*
Nucleophilic attack
R
R
*
(Inversion)
NH2
H
O
NH2
O
O
HO
O
OH
HO
Tyr
H
HO
O
O-
Asn
Asn
Asp
Tyr
Asp
Scheme 2. Two mechanistic principles of enzymatic epoxide hydrolysis. *Center of chirality.
configuration) followed by inverting acid-catalyzed
hydrolysis to furnish (R)-1-benzyloxy-2-methylpro-
pane-2,3-diol in >97% ee and 78% yield from the
racemate.
Retention times were as follows: (R)-1, 16.52 min, (S)-
1, 17.89 min, (R)-2, 19.05 min, (S)-2, 20.35 min; reten-
tion times for 1 could vary to some extent depending
on column pretreatment.
4.2. Biocatalytic hydrolysis of ( )-2-methylglycidyl
benzyl ether 1
4. Experimental
4.1. General
4.2.1. Screening. Lyophilized cells (30 mg) were rehy-
drated in Tris/HCl buffer (500 lL, 0.05 M, pH 8) for
1 h at 30 ꢁC and 130 rpm. Then, 15 lL (168 mM) of
rac-1 was added and the mixture agitated on a thermo-
stated shaker. After 24–48 h, the reactions were stopped
by extraction with ethyl acetate (twice 500 lL). To facili-
tate phase separation, cells were removed by centrifuga-
tion. The combined organic layers were dried over
Na2SO4, evaporated, and chromatographed (petroleum
ether/ethyl acetate 2:1) to obtain the remaining non-con-
verted epoxide and the formed diol was eluted using
ethyl acetate.
Reactions were monitored by TLC (Merck silica gel 60
F254) and compounds visualized by spraying with vanil-
lin–concd H2SO4 (5 g Lꢀ1). Column chromatography
was performed using silica gel Merck 60 (230–400
mesh). Petroleum ether with a boiling range of 60–
90 ꢁC was used. Solvents were dried and freshly distilled
by standard techniques. For anhydrous reactions, flasks
were dried at 150 ꢁC and flushed with dried argon just
before use. For biotransformations, lyophilized bacte-
rial cells were used. Bacteria were obtained from culture
collections and grown as previously described.40–43 ( )-
2-Methylglycidyl benzyl ether 1 was synthesized as pre-
viously described.10 Spectroscopic data were identical to
those previously reported.44 HPLC analyses were car-
ried out on a JASCO system containing a PU-980 pump
equipped with a Daicel Chiralpak AD column con-
nected to a MD-910 multi-wavelength detector. The
enantiomers of epoxide 1 were separated with n-heptane
as eluent and a flow of 0.4 mL/min and those of diol 2
using n-heptane/2-propanol 90:10 at a flow of 0.5 mL/
min at 18 ꢁC. The absolute configurations of 1 and 2
were assigned according to the relative elution order of
enantiomers on HPLC as previously determined.45
4.2.2. Optimization. The general procedure described
above was modified by altering the following
parameters.
4.2.2.1. Substrate concentration. Lyophilized cells
(70 mg), Tris/HCl buffer (250 lL, 0.05 M, pH 8), sub-
strate rac-1 (35 lL).
4.2.2.2. Catalyst loading. Catalyst/substrate ratio
10:1: Lyophilized cells (50 mg), Tris/HCl buffer (1 mL,
0.05 M, pH 8), substrate rac-1 (5 lL). Catalyst/substrate
ratio 1:5: Lyophilized cells (1 mg), Tris/HCl buffer

408
Y. Simeo´, K. Faber / Tetrahedron: Asymmetry 17 (2006) 402–409
(1 mL, 0.05 M, pH 8), substrate rac-1 (5 lL). Catalyst/
substrate ratio 1:10: Lyophilized cells (1 mg), Tris/HCl
buffer (1 mL, 0.05 M, pH 8), substrate rac-1 (10 lL).
rated NaHCO3 solution, and distilled water. After dry-
ing over Na2SO4, solids were filtered and the solution
was evaporated to yield crude product as an oil, which
was dissolved in 20 mL anhyd THF and cooled to
0 ꢁC. NaH (132 mg, 5.5 mmol) was added and the mix-
ture was allowed to reach room temperature while stir-
ring for 1.5 h. It was then poured into a mixture of ice
and NH4Cl. Distilled water (10 mL) was added, the
phases were separated, and the aqueous layer was ex-
tracted with CH2Cl2 (2 · 20 mL). The combined organic
layers were dried (Na2SO4), filtered, and concentrated
under reduced pressure to render 300 mg (92%) of (S)-
4.2.2.3. Temperature. Lyophilized cells (30 mg) were
rehydrated in a Tris/HCl buffer (500 lL, 0.05 M, pH 8)
for 1 h at 30 ꢁC and 130 rpm and after which for 1 h at
4 ꢁC. Then, 15 lL of rac-1 was added and the mixture
agitated at 4 ꢁC and 130 rpm for 24–48 h.
4.2.2.4. Aqueous/organic solvent systems. To a solu-
tion of rehydrated lyophilized cells (30 mg) in 500 lL of
Tris/HCl buffer (0.05 M, pH 8), 100 lL of the organic
co-solvent (25 lL in case of dimethyl sulfoxide) was
added and the mixture was agitated for 1 h at 30 ꢁC
and 130 rpm. Then, 15 lL of rac-1 was added and the
reaction was agitated for 24 h under the same condi-
tions. A catalyst/substrate ratio 10:1 in the presence of
organic co-solvents: Lyophilized cells (50 mg), Tris/
HCl buffer (1 mL, 0.05 M, pH 8), organic solvent
(200 lL), rac-1 (5 lL).
20
25
1. ½aꢁD ¼ þ9:2 (c 1.0, CHCl3); lit.:47 ½aꢁD ¼ þ10:9 (c
1.20, MeOH).
Acknowledgments
´
´
Financial support from Fundacion Ramon Areces is
gratefully acknowledged. Soybean epoxide hydrolase
´
was kindly provided by Elizabeth Blee (University of
Strasbourg) and potato- and limonene 1,2-epoxide
hydrolase by Michael Arand and Anette Cronin (Uni-
4.3. Chemo-enzymatic deracemization of ( )-1
versity of Zurich).
¨
Lyophilized cells (1 g) of Rhodococcus sp. CBS 717.73
were rehydrated in a mixture of Tris/HCl buffer
(20 mL, 0.05 M, pH 8) and i-octane (4 mL) by shaking
at 30 ꢁC for 1 h. Then, 100 lL of substrate rac-1 were
added. The mixture was agitated on a thermostated sha-
ker at 30 ꢁC and 130 rpm for 24 h. The products were
then extracted with ethyl acetate (four times, phase sepa-
ration was facilitated by centrifugation), and the
combined organic layers dried over Na2SO4 and evapo-
rated. The resulting bright yellow oil consisting of (R)-1
and (R)-2 was subjected to acid-catalyzed epoxide open-
ing under inversion of configuration without further
purification. Thus, the mixture was dissolved in dioxane
(20 mL) and cooled to 0 ꢁC. Then, 220 lL of 93% aq
H2SO4 was added to the solution, which was allowed
to reach rt and stirred for 20 min. The reaction was neu-
tralized with aq satd NaHCO3, EtOAc was added, and
the resulting biphasic mixture was stirred vigorously
for 30 min. Finally, the organic layer was separated,
the aqueous phase was extracted twice with ethyl
acetate, and the combined organic layers were dried,
filtered, and evaporated. Column chromatography ren-
References
1. Tietze, L. F.; Go¨rlitzer, J. Synthesis 1998, 873–878.
2. De Vries, E. J.; Janssen, D. B. Curr. Opin. Biotechnol.
2003, 14, 414–420; Steinreiber, A.; Faber, K. Curr. Opin.
Biotechnol. 2001, 12, 552–558; Archelas, A.; Furstoss, R.
Curr. Opin. Chem. Biol. 2001, 5, 112–119; Weijers, C. A.
G. M.; De Bont, J. A. M. J. Mol. Catal. B: Enzym. 1999,
6, 199–214; Ueberbacher, B. J.; Osprian, I.; Mayer, S. F.;
Faber, K. Eur. J. Org. Chem. 2005, 1266–1270.
3. Steinreiber, A.; Osprian, I.; Mayer, S. F.; Orru, R. V. A.;
Faber, K. Eur. J. Org. Chem. 2000, 3703–3711.
4. Osprian, I.; Stampfer, W.; Faber, K. J. Chem. Soc., Perkin
Trans. 1 2000, 3779–3785.
5. Xu, Y.; Xu, J.-H.; Pan, J.; Zhao, L.; Zhang, S.-L. J. Mol.
Catal. B: Enzym. 2004, 27, 155–159.
6. Pedragosa-Moreau, S.; Morisseau, C.; Baratti, J.; Zylber,
J.; Archelas, A.; Furstoss, R. Tetrahedron 1997, 53, 9707–
9714.
7. Nellaiah, H.; Morisseau, C.; Archelas, A.; Furstoss, R.;
Baratti, J. C. Biotechnol. Bioeng. 1996, 49, 70–77.
8. Li, C.; Liu, Q.; Song, X.; Ding, D.; Ji, A.; Qu, Y.
Biotechnol. Lett. 2003, 25, 2113–2116.
9. Krieg, H. M.; Botes, A. L.; Smit, M. S.; Breytenbach, J.
C.; Keizer, K. J. Mol. Catal. B: Enzym. 2001, 13, 37–
47.
dered (R)-2 (86 mg, 78% from rac-1, >97% ee) as the
20
sole product. ½aꢁD ¼ ꢀ6:6 (c 1.33, CH2Cl2); lit.:46 [a]D =
ꢀ6.3 (c 0.87, CH2Cl2). Spectroscopic data were in agree-
ment with those previously reported.46
10. Hellstro¨m, H.; Steinreiber, A.; Mayer, S. F.; Faber, K.
Biotechnol. Lett. 2001, 23, 169–173.
4.4. Cyclization of diol (R)-2 to epoxide (S)-1
11. Cleij, M.; Archelas, A.; Furstoss, R. Tetrahedron: Asym-
metry 1998, 9, 1839–1842.
12. Botes, A. L.; Litthauer, D.; Van Tonder, A.; Van Dyk, M.
S. Biotechnol. Lett. 1999, 21, 1137–1144.
13. Gong, P.-F.; Xu, J.-H. Enzyme Microb. Technol. 2005, 36,
252–257.
14. Choi, W. J.; Lee, E. Y.; Yoon, S. J.; Yang, S.-T.; Choi, C.
Y. J. Biosci. Bioeng. 1999, 88, 339–341.
15. Goswami, A.; Totleben, M. J.; Singh, A. K.; Patel, R. N.
Tetrahedron: Asymmetry 1999, 10, 3167–3175.
16. Baldascini, H.; Janssen, D. B. Enzyme Microb. Technol.
2005, 36, 285–293.
Diol (R)-2 (357 mg, 1.8 mmol) was dissolved in 25 mL
of CH3CN, after which Et3N (0.76 mL, 5.5 mmol) and
Me3NH+Clꢀ (35 mg, 0.4 mmol) were added. The solu-
tion was cooled to 0 ꢁC and p-toluenesulfonyl chloride
(1 g, 5.5 mmol) was added. The mixture was stirred at
0 ꢁC for 2 h and the reaction was quenched by addition
of satd aq NH4Cl solution (30 mL). The phases were
separated, the aqueous layer was extracted with AcOEt
(3 · 30 mL), and the combined organic layers were
washed with saturated NH4Cl solution (15 mL), satu-

Y. Simeo´, K. Faber / Tetrahedron: Asymmetry 17 (2006) 402–409
409
17. Baldascini, H.; Ganzeveld, K. J.; Janssen, D. B.; Been-
ackers, A. A. C. M. Biotechnol. Bioeng. 2001, 73, 44–54.
18. Lotter, J.; Botes, A. L.; Van Dyk, M. S.; Breytenbach, J.
C. Biotechnol. Lett. 2004, 26, 1191–1195.
19. Chiappe, C.; Leandri, E.; Lucchesi, S.; Pieraccini, D.;
Hammock, B. D.; Morisseau, C. J. Mol. Catal. B: Enzym.
2004, 27, 243–248.
32. Lacourciere, G. M.; Armstrong, R. N. J. Am. Chem. Soc.
1993, 115, 10466–10467.
33. Argiriadi, M. A.; Morisseau, C.; Hammock, B. D.;
Christianson, D. W. Proc. Natl. Acad. Sci. U.S.A. 1999,
96, 10637–10642, PDB-code 1ehy.
34. Nardini, M.; Rink, R.; Janssen, D. B.; Dijkstra, B. W. J.
Mol. Catal. B: Enzym. 2001, 11, 1035–1042; Nardini, M.;
Ridder, I. S.; Rozeboom, H. J.; Kalk, K. H.; Rink, R.;
Janssen, D. B.; Dijkstra, B. W. J. Biol. Chem. 1999, 274,
14579–14586, PDB-code 1qo7.
35. Zou, J.; Hallberg, B. M.; Bergfors, T.; Oesch, F.; Arand,
M.; Mowbray, S. L.; Jones, T. A. Structure 2000, 8, 111–
122.
36. Arand, M.; Hallberg, B. M.; Zou, J.; Bergfors, T.; Oesch,
F.; van der Werf, M.; de Bont, J. A. M.; Jones, T. A.;
Mowbray, S. L. EMBO J. 2003, 22, 2583–2592, PDB-
codes 1nu3, 1nww.
37. Hopmann, K. H.; Hallberg, B. M.; Himo, F. J. Am. Chem.
Soc. 2005, 127, 14339–14347.
38. Pocker, Y.; Ronald, B. P. J. Am. Chem. Soc. 1978, 100,
3122–3127; Pocker, Y.; Ronald, B. P. J. Am. Chem. Soc.
1980, 102, 5311–5316; Pocker, Y.; Ronald, B. P.; Ferrin,
L. J. Am. Chem. Soc. 1980, 102, 7725–7732.
39. Orru, R. V. A.; Mayer, S. F.; Kroutil, W.; Faber, K.
Tetrahedron 1998, 54, 859–874.
20. Kronenburg, N. A. E.; De Bont, J. A. M. Enzyme Microb.
Technol 2001, 28, 210–217.
21. Spelberg, J. H. L.; Rink, R.; Archelas, A.; Furstoss, R.;
Janssen, D. B. Adv. Synth. Catal. 2002, 344, 980–985.
22. Rink, R.; Spelberg, J. H. L.; Pieters, R. J.; Kingma, J.;
Nardini, M.; Kellog, R. M.; Dijkstra, B. W.; Janssen, D.
B. J. Am. Chem. Soc. 1999, 121, 7417–7418.
23. Reetz, M. T.; Torre, C.; Eipper, A.; Lohmer, R.; Hermes,
M.; Brunner, B.; Maichele, A.; Bocola, M.; Arand, M.;
Cronin, A.; Genzel, Y.; Archelas, A.; Furstoss, R. Org.
Lett. 2004, 6, 177–180.
24. Kronenburg, N. A. E.; De Bont, J. A. M.; Fischer, L. J.
Mol. Catal. B: Enzym. 2001, 16, 121–129.
25. Va¨nttinen, E.; Kanerva, L. T. Tetrahedron: Asymmetry
1995, 6, 1779–1786.
26. Nakamura, K.; Takebe, Y.; Kitayama, T.; Ohno, A.
Tetrahedron Lett. 1991, 32, 4941–4944.
27. Rhodococcus ruber CBS 717.73 was formerly classified
as Brevibacterium sp., see: Briand, D.; Dubreucq, E.;
Perrier, V.; Grimaud, J.; Galzy, P. Microbios 1994, 78,
205–214.
28. Morisseau, C.; Beetham, J. K.; Pinot, F.; Debernard, S.;
Newman, J. W.; Hammock, B. D. Arch. Biochem. Biophys.
2000, 378, 321–332.
40. Krenn, W.; Osprian, I.; Kroutil, W.; Braunegg, G.; Faber,
K. Biotechnol. Lett. 1999, 21, 687–690.
41. Kroutil, W.; Osprian, I.; Mischitz, M.; Faber, K. Synthesis
1997, 156–158.
42. Osprian, I.; Kroutil, W.; Mischitz, M.; Faber, K. Tetra-
hedron: Asymmetry 1997, 8, 65–71.
29. Summerer, S.; Hanano, A.; Utsumi, S.; Arand, M.;
Schuber, F.; Blee, E. Biochem. J. 2002, 366, 471–480.
30. Stapleton, A.; Beetham, J. K.; Pinot, F.; Garbarino, J. E.;
Rockhold, D. R.; Friedman, M.; Hammock, B. D.;
Belknap, W. R. Plant J. 1994, 6, 251–258.
43. Mischitz, M.; Kroutil, W.; Wandel, U.; Faber, K. Tetra-
hedron: Asymmetry 1995, 6, 1261–1272.
44. Ihara, M.; Tanaka, Y.; Takahashi, N.; Tokunaga, Y.;
Fukumoto, K. J. Chem. Soc., Perkin Trans. 1 1997, 3043–
3052.
31. van der Werf, M. J.; Overkamp, K. M.; de Bont, J. A. M.
J. Bacteriol. 1998, 180, 5052–5057; van der Werf, M. J.;
Orru, R. V. A.; Overkamp, K. M.; Swarts, H. J.; Osprian,
I.; Steinreiber, A.; de Bont, J. A. M.; Faber, K. Appl.
Microbiol. Biotechnol. 1999, 52, 380–385, PDB-code
1nww.
45. Steinreiber, A.; Strauss, U.; Mayer, S. F.; Uray, G. J.
Chromatogr., A 2001, 911, 295–298.
46. Tanner, D.; Somfai, P. Tetrahedron 1986, 42, 5985–5990.
47. Avenoza, A.; Cativiela, C.; Peregrina, J. M.; Sucunza, D.;
Zurbano, M. M. Tetrahedron: Asymmetry 2001, 12, 1383–
1388.