Microbiological Transformations. 38. Clues to the Involvement of a General Acid Activation during Hydrolysis of Para-Substituted Styrene Oxides by a Soluble Epoxide Hydrolase from Syncephalastrum racemosum

Article abstract of DOI:10.1021/jo9714371

In the course of this work, we have determined the regioselectivity as well as the rate of biohydrolysis of various para-substituted styrene oxide derivatives catalyzed by a new epoxide hydrolase activity found in the soluble cell extract of the fungus Syncephalastrum racemosum. We have observed that this regioselectivity switched progressively from the benzylic C_α carbon atom to the terminal C_β carbon atom depending upon the electronic character of the para substituent. Hammett plotting of the ratio of water incorporation at both the benzylic and terminal carbon atoms, i.e., log α/β versus σ, gave linear relationships for the two (R)- and (S)-epoxide enantiomers with slopes pα/β = -2.07 and -1.35, respectively. Apparent kinetic constants K_m and V_max were determined for the biohydrolysis of the enantiomers of R absolute configuration, which were the better substrates. Hammett correlation was investigated for V_max/K_m for the reaction on both the C_α and C_β carbon atoms. Log αV_max/K_m vs σ gave a linear relationship with a slope pαv_max/K_m = -1.8, suggesting that, in the case of these enzyme/substrate couples, the rate-determining step is the oxirane ring cleavage. These results give, for the first time, interesting clues to the fact that a general acid activation of the epoxide is very probably involved in a concerted process together with its nucleophilic attack.

Full text of DOI:10.1021/jo9714371

3
532
J . Org. Chem. 1998, 63, 3532-3537
Articles
Micr obiologica l Tr a n sfor m a tion s. 38. Clu es to th e In volvem en t of
a Gen er a l Acid Activa tion d u r in g Hyd r olysis of P a r a -Su bstitu ted
Styr en e Oxid es by a Solu ble Ep oxid e Hyd r ola se fr om
Syn ceph a la str u m r a cem osu m
P. Moussou, A. Archelas, J . Baratti, and R. Furstoss*
Groupe Biocatalyse et Chimie Fine, ERS 157 associ e´ e au CNRS, Facult e´ des Sciences de Luminy,
Case 901, 163 avenue de Luminy, 13288 Marseille Cedex 9, France
Received August 4, 1997
In the course of this work, we have determined the regioselectivity as well as the rate of biohydrolysis
of various para-substituted styrene oxide derivatives catalyzed by a new epoxide hydrolase activity
found in the soluble cell extract of the fungus Syncephalastrum racemosum. We have observed
that this regioselectivity switched progressively from the benzylic C
carbon atom depending upon the electronic character of the para substituent. Hammett plotting
of the ratio of water incorporation at both the benzylic and terminal carbon atoms, i.e., log R/â
versus σ, gave linear relationships for the two (R)- and (S)-epoxide enantiomers with slopes FR/â
2.07 and -1.35, respectively. Apparent kinetic constants K and Vmax were determined for the
biohydrolysis of the enantiomers of R absolute configuration, which were the better substrates.
Hammett correlation was investigated for Vmax/K for the reaction on both the C and C carbon
atoms. Log RVmax/K vs σ gave a linear relationship with a slope FRVmax/Km ) -1.8, suggesting that,
R
carbon atom to the terminal
C
â
)
-
m
m
R
â
m
in the case of these enzyme/substrate couples, the rate-determining step is the oxirane ring cleavage.
These results give, for the first time, interesting clues to the fact that a general acid activation of
the epoxide is very probably involved in a concerted process together with its nucleophilic attack.
Epoxide hydrolases (EHs) catalyze the addition of
efficient biocatalysts for achieving the resolution of
racemic epoxides and were shown to allow the prepara-
tion of several epoxides and diols thus obtained with good
to excellent enantiomeric purity.7 Therefore, these co-
factor-independent enzymes can be regarded as being,
as were lipases about 20 years ago, very promising new
tools in the organic chemist’s tool box, in particular for
performing asymmetric synthesis. Thus, the study of
their properties, i.e., of their catalytic behavior, molecular
mechanism, and three-dimensional structure, is becom-
ing a topic of utmost importance.
The mechanism implied in these enzymes has only
been recently elicited, at least for murine mEH and sEH.
It was shown that the reaction occurs via a two step
process involving a trans-specific nucleophilic attack of
the oxirane ring by an aspartic residue, which forms a
covalent enzyme-substrate ester intermediate in a first
step8 (Scheme 1). This ester intermediate is subse-
quently hydrolyzed, in a second step, by a water molecule
activated by an histidine-aspartate pair. The accurate
understanding of the results obtained by performing such
reactions is, however, complicated by the existence of two
water onto an epoxide ring, thus affording the corre-
sponding vicinal diol. Due to their essential involvement
in human detoxification processes, these enzymes have
been mostly studied in mammals up to the recent years.1
They were shown to exist predominantly in two different
forms, i.e., microsomal (mEH) or cytosolic (soluble) (sEH)
forms. However, similar EH activities have been identi-
2
3
fied recently in organisms as diverse as bacteria, yeasts,
4
5
6
fungi, plants, and insects, and these enzymes may thus
be regarded as being ubiquitous in nature. Interestingly,
they have been described in recent years to be very
*
To whom correspondence should be addressed. Tel.: +33 04 91
2 91 55. Fax: +33 04 91 82 91 45. E-mail: furstoss@lac.gulliver.fr.
1) (a) Hammock, B. D.; Grant, D. F.; Storms, D. H. In Comprehen-
8
(
sive Toxicology; Sipes, I., McQueen, C., Gandolfi, A., Eds.; Pergamon:
Oxford, 1996; Chapter 18.3. (b) Knehr, M.; Thomas, H.; Arand, M.;
Gebel, T.; Zeller, H.-D.; Oesch, F. J . Biol. Chem. 1993, 268, 17623. (c)
Bellucci, G.; Chiappe, C.; Cordoni, A.; Ingrosso, G. Tetrahedron Lett.
1
996, 37, 9089.
2) Osprian, I.; Kroutil, W.; Mischitz, M.; Faber, K. Tetrahedron:
Asymmetry 1997, 8, 65.
(
(
(
3) Weijers, C. A. G. M. Tetrahedron: Asymmetry 1997, 8, 639.
4) (a) Chen, X.-J .; Archelas, A.; Furstoss, R. J . Org. Chem. 1993,
5
8, 5528. (b) Pedragosa-Moreau, S.; Archelas, A.; Furstoss, R. J . Org.
Chem. 1993, 58, 5533. (c) Zhang, J .; Reddy, J .; Roberge, C.; Senanay-
ake, C.; Greasham, R.; Chartrain, M. J . Ferment. Bioeng. 1995, 80,
(6) (a) Linderman, R. J .; Walker; E. A.; Haney, C.; Roe, R. M.
Tetrahedron 1995, 51, 10845. (b) Graham, S. M.; Prestwich, G. D. J .
Org. Chem. 1994, 59, 2956. (c) Mullin, C. A. J . Chem. Ecol. 1988, 14,
1867.
(7) (a) Archelas, A.; Furstoss, R. Annu. Rev. Microbiol. 1997, 51,
491. (b) Faber, K.; Mischitz, M.; Kroutil W. Acta Chem. Scand. 1996,
50, 249. (c) Archer, I. V. J . Tetrahedron 1997, 53, 15617-15662. (d)
Archelas, A.; Furstoss, R. Trends Biotechnol. 1998, in press.
2
44.
5) (a) Sch o¨ ttler, M.; Boland W. Helv. Chem. Acta 1996, 79, 1488.
b) Bl e´ e, E.; Schuber, F. Eur. J . Biochem. 1995, 230, 229. (c) 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,
2
51.
S0022-3263(97)01437-0 CCC: $15.00 © 1998 American Chemical Society
Published on Web 05/08/1998

Microbiological Transformations
J . Org. Chem., Vol. 63, No. 11, 1998 3533
Sch em e 1
Sch em e 2
possible reaction centers at the oxirane ring. Thus, the
aspartic residue may attack following two different
regioselectivities, thus leading to (formal) incorporation
of water at onesor at the otherscarbon atom. Further-
more, this regioselectivity may be shared in different
proportions on the two carbon atoms and may also be
different for each of the two enantiomers.
On the basis of the absolute configurations obtained
for the formed diol, as well as from 18O-labeling experi-
ments, it is now well established that, for mammalian
microsomal EH, the aspartic acid attack occurs highly
preferentially at the less hindered carbon atom for both
enantiomers.9 Furthermore, series of styrene oxides and
of cis-stilbene oxide derivatives, bearing aryl substituents
of increasing electronic-withdrawing power, have been
studied with the (partially purified) rat mEH.10 From
the observed variation of reaction rates, it has been con-
cluded, via Hammett coefficient plotting, that (a) the rate-
determining step was a nucleophilic attack and (b) a free
carbocation ion was not involved in the mechanism.
Interestingly, we have ourselves recently found results
leading to the same conclusion with the sEH (crude solu-
ble enzymatic extract) from the fungus Aspergillus niger,
which catalyzes also the water incorporation at the less
hindered carbon atom.11 On the other hand, Bellucci et al.
have observed that the hydrolysis of trans-3-bromo-1,2-
epoxycyclohexane, catalyzed by rabbit mEH (crude micro-
the above-cited examples, the oxirane ring cleavage
would involve an electrophilic activation (via either
protonation or hydrogen bonding) of the epoxide in the
transition state.1
3,15
However, direct experimental proofs
that confirm this hypothesis, or kinetic studies similar
to those carried out with mammalian mEH or A. niger
EH, are not presently available. In the course of our
present work, we have focused on a new fungal EH
activity from the fungus Syncephalastrum racemosum
MUCL 2876616 (crude soluble protein extract), which, in
contrast to what was observed with the fungus A. niger,
led to a highly preferential attack at the more hindered
carbon atom for styrene oxide itself. We report here our
results about the regioselectivity of the water molecule
incorporation, as well as the kinetic study of the hydroly-
sis of a series of differently para-substituted styrene
oxides by this new biocatalyst.
Resu lts a n d Discu ssion
The various racemic para-substituted styrene oxides
somes), led to trans-2,3-cyclohexanol, presumably via in-
1
1
tramolecular trapping of the transient oxyanion.12 This
1-7 (Scheme 2), prepared as previously described, were
submitted to biohydrolysis by a crude soluble enzymatic
extract obtained from the fungus Syncephalastrum race-
mosum MUCL 28766. This extract was prepared from
the soluble fraction of the cell extracts, similar to the one
obtained from A. niger.17
was interpreted by the authors as evidence that the nu-
cleophilic attack occurred at the oxirane ring without ap-
parent electrophilic activation/protonation of the epoxide.
Unexpectedly, an opposite regioselectivity, i.e., indicat-
ing that the water molecule is partially (or even totally)
incorporated at the more hindered (benzylic) center, has
been observed for the opening of aryl epoxides by rabbit-
The enantiomeric ratios (E values) for the reactions
with X ) CH , H, Cl, and NO have been determined
3 2
according to Sih et al.18 Unfortunately, since these E
values were very low (respectively, 1.1, 8, 5, and 1.6), this
soluble enzymatic extract appeared not to be a good
catalyst for the kinetic resolution of these racemic para-
substituted styrene oxides. Interestingly, however, for
all the epoxides studied, the residual enantiomer was
always of S absolute configuration, which means that the
preferred antipode was always of R configuration, whereas
the absolute configuration of the diol formed in excess
1
3
14
and murine-soluble EH, as well as by whole cells of
the fungus Beauveria bassiana (formerly B. sulfure-
scens).15 On the basis of these regioselectivities, it has
been suggested that, in these cases and in contrast to
(
8) (a) Lacourci e` re, G. M.; Armstrong, R. N. J . Am. Chem. Soc. 1993,
15, 10466. (b) Pinot, F.; Grant, D. F.; Beetham, J . K.; Parker, A. G.;
Borhan, B.; Landt, S.; J ones, A. D.; Hammock, B. D. J . Biol. Chem.
1
1
1
995, 270, 7968. (c) Arand, M.; Wagner, H.; Oesch, F. J . Biol. Chem.
996, 271, 4223. (d) M u¨ ller, F.; Arand, M.; Seidel, A.; Hinz, W.;
Winkler, L.; H a¨ nel, K.; Bl e´ e, E.; Beetham, J . K.; Hammock, B. D.;
Oesch, F. Eur. J . Biochem. 1997, 245, 490.
switched from S to R, respectively, for the CH
3
, H, F, Cl,
and Br, CN, NO groups. This in fact indicates that the
2
(
9) (a) Barili, P. L.; Berti, G.; Mastrorilli, E. Tetrahedron 1993, 49,
nature of the substituent born by the aromatic ring did
strongly affect the regioselectivity of the hydrolysis.
Thus, the water molecule was preferentially incorporated
at the more hindered carbon atom for the substrate
bearing weak electron-donating (CH ) or -withdrawing
3
groups, whereas incorporation at the less substituted
6
263. (b) Hanzlik, R. P.; Edelman, M.; Michaely, W. J .; Scott, G. J .
Am. Chem. Soc. 1976, 98, 1952. (c) J erina, D. M.; Ziffer, H.; Daly, J .
W. J . Am. Chem. Soc. 1970, 92, 1056.
(
10) Dansette, P. M.; Makedonska, V. B.; J erina, D. M. Arch.
Biochem. Biophys. 1978, 187, 290.
11) Pedragosa-Moreau, S.; Morisseau, C.; Zylber, J .; Archelas, A.;
Baratti, J . C.; Furstoss, R. J . Org. Chem. 1996, 61, 7402.
12) Bellucci, G.; Berti, G.; Ferretti, M.; Marioni, F.; Re, F. Biochem.
Biophys. Res. Commun. 1981, 102, 838.
13) Bellucci, G.; Chiappe, C.; Cordoni, A.; Marioni, F. Tetrahedron
Lett. 1994, 35, 4219.
14) Dietze, E. C.; Kuwano, E.; Casas, J .; Hammock, B. D. Biochem.
Pharmacol. 1991, 42, 1163.
15) Pedragosa-Moreau, S.; Archelas, A.; Furstoss, R. Bioorg. Med.
Chem. 1994, 2, 609.
(
(
(
(16) This fungus was selected, because of its peculiar regioselectivity,
from a screening for new fungal EH activity. Unpublished results.
(17) Morisseau, C.; Nellaiah, H.; Archelas, A.; Furstoss, R.; Baratti,
J . C. Enzymol. Microb. Technol. 1997, 20, 446.
(18) Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J . J . Am. Chem.
Soc. 1982, 104, 7294.
(
(

3
534 J . Org. Chem., Vol. 63, No. 11, 1998
Moussou et al.
Ta ble 1. Regioselectivity of th e Bioh yd r olysis of th e R
a n d S En a n tiom er s of P a r a -Su bstitu ted Styr en e Oxid es
Usin g a Solu ble P r otein Extr a ct of S. r a cem osu m
substituent
CH3
H
0
F
CL
Br
CN
NO2
0.81
σpa
-0.14
0.15
0.24
0.26
0.70
substrate
R)-1 (R)-2 (R)-3 (R)-4 (R)-5 (R)-6 (R)-7
(
b
R (% diol (S) )
93
7
90
10
87
13
67
33
56
44
29
71
11
89
b
â (% diol (R) )
substrate
(S)-1 (S)-2 (S)-3 (S)-4 (S)-5 (S)-6 (S)-7
b
R (% diol (R) )
93
7
53
47
43
57
73
27
69
31
53
47
34
66
b
â (% diol (S) )
a
σp values are from: March, J . Advanced Organic Chemistry,
b
3
rd ed.; Wiley: New York, 1985. The absolute configurations
have been determined previously. See ref 10.
carbon atom was preferred for more powerful electron-
withdrawing substituents.
Regioselectivity of th e Hyd r olysis. To get a more
accurate explanation of the reasons governing these
differences, and to determine the regioselectivity involved
for each one of the two antipodes, we performed sepa-
rately, for all the substrates, the biohydrolysis of both
enantiomers (Table 1). This determination could be
achieved in a straightforward way, since the percentage
of diol of opposite absolute configuration formed (as
compared to the one of the starting substrate) corre-
sponds to the percentage of attack at the stereogenic
benzylic carbon atom (R attack), whereas the percentage
of diol showing the same absolute configuration as the
starting epoxide reflects the percentage of attack at the
terminal carbon atom (â attack). The data obtained show
a remarkable effect of the nature of the para-substituent
on the regioselectivity, and this effect seems to depend
essentially upon the electronic power of the substituent.
Indeed, for hydrolysis of both (R)-1 and (S)-1 (bearing
the electron-donating methyl group), the attack was
p
F igu r e 1. Hammett plot of log(R/â) versus σ for the hydroly-
sis of (R)-p-substituted styrene oxide (b) and (S)-p-substituted
styrene oxide (O) catalyzed by a soluble enzymatic extract of
S. racemosum.
be that a lack of steric hindrance at the para position
could allow the substrate to adopt a different positioning
into the enzymatic active site. In this context, it is
interesting to emphasize that the fact that these correla-
tions appear to be rather linear is a good indication that
only one EH should be involved in this crude enzyme
preparation.
Kin etic Stu d y. This surprising gradual switch in
regioselectivity related to para substitution lies in sharp
contrast with the results we have previously obtained
with the A. niger EH, and led us to search for more
information about the rates of these hydrolyses. To avoid
the problems due to competition between the two enan-
tiomers of a racemic substrate, and to the difference of
regioselectivity observed in some cases for each one of
these enantiomers, we performed this kinetic study using
the faster hydrolyzed R enantiomer of epoxides 1-7 in
enantiopure form. Also, due to the low solubility of these
epoxides in water, the experiments were carried out using
10% of DMSO as a cosolvent.17 The initial hydrolysis
rates were measured as a function of the substrate
concentration, up to their limit of solubility (i.e., 30, 37,
31, 23, 7, 36, and 5 mM, respectively, for epoxides 1-7).
R
highly preferential (93%) at the benzylic carbon atom C ,
whereas this percentage decreased when the electron-
withdrawing character of the substituent group in-
creases. Thus, for the powerful nitro electron-withdraw-
ing group, the attack became highly preferential at the
â
terminal C carbon atom, i.e., 89% for (R)-7 and 66% for
(
S)-7. Plotting of the R/â ratio log versus the Hammett
p
constant σ for each substituent is given in Figure 1. It
can be seen that, for hydrolysis of the (R)-epoxides, a
rather good linear relationship exists (correlation coef-
2
ficient r ) 0.96) with a negative slope of FR/â ) -2.07.
This correlation indicates that, for this series of sub-
strates, the variation of regioselectivity essentially de-
pends on the electronic properties of the substituent.
Similarly, for the hydrolysis of almost all the (S)-
epoxides, a rather good linear relationship (correlation
m
The apparent kinetic parameters (K and Vmax) were
calculated via a nonlinear regression of the Michaelis-
Menten equation and are reported in Table 2.
It appears as a general trend that, when the electron-
withdrawing character of the substituent group in-
creases, the maximal enzymatic rate (Vmax) and the
2
coefficient r ) 0.94) with a different, but still negative,
slope of FR/â ) -1.35 was also obtained. As previously,
the variation of regioselectivity seems to depend, for the
major part, upon the electronic effect of the substituent.
However, for the S enantiomer of styrene oxide itself, as
well as for its p-fluoro-substituted derivative (S)-3, the
data point repeatedly lied outside this correlation line.
At the present time, we do not have any satisfactory
explanation for this observation. One suggestion could
m
specificity constant (Vmax/K ) decrease (respectively, by
a factor of 46 and 10 going from the electron-donating
methyl group to the powerful electron-withdrawing nitro
group). So, the gradual switch in regioselectivity related
to para-substitution is combined with the Vmax and the
V
max/K
m
decrease (Scheme 3).
It is well documented that the interpretation of Ham-
mett-type studies in enzymatic reactions may be com-

Microbiological Transformations
J . Org. Chem., Vol. 63, No. 11, 1998 3535
Ta ble 2. Su bstitu en t Effects on th e Ap p a r en t Kin etic P a r a m eter s of th e Hyd r olysis of P a r a -Su bstitu ted Styr en e Oxid es
Ca ta lyzed by a Solu ble P r otein Extr a ct of S. r a cem osu m e
substrate
(R)-1
(R)-2 l
(R)-3
(R)-4
(R)-5
(R)-6
(R)-7
Kma
Vmax
Vmax/Km
Vmax/Km
57 ( 6
100 ( 19
57 ( 8
25 ( 5
17 ( 2
20 ( 9
29.5 ( 1.5
9.2 ( 0.3
12.0 ( 1.2
2.3 ( 0.2
a
106 ( 8
30 ( 4
17 ( 1
15 ( 5
b
c
1.84( 0.06
0.57( 0.02
1.20( 0.13
0.99( 0.06
0.75( 0.08
0.31( 0.01
0.19( 0.005
2
2
NDd
2
2
2
2
r
> 0.99
r
> 0.99
r
> 0.99
r
> 0.99
r
> 0.99
r
) 0.97
a
Km are expressed in mm, Vmax in µmol‚min-1‚g-1 of protein. Obtained from the substrate dependence of initial velocities by fitting
b
the equation vi ) a([S]/([S] + b), with a ) Vmax and b ) Km.c Obtained from the substrate dependence of initial velocities by fitting the
equation vi ) a(b[S]/([S] + b), with a ) Vmax/Km and b ) Km. Obtained from the substrate dependence of initial velocities for [S] < Km/10
d
e
by fitting the equation vi ) a[S], with a ) Vmax/Km. Not determined. The obtained Km values, which were in the range of 12-100 mM,
were of the same magnitude (or higher) than the solubility limit of the epoxides 1-7. Consequently, the Km values have been obtained
by fitting the Michaelis-Menten equation only with relatively high standard errors. However, the standard errors for the Vmax and
Vmax/Km values thus obtained were better. Furthermore, the Vmax/Km values calculated as the slopes of the linear correlations between
the initial velocities and [S] for [S] < Km/10 (for three different values of [S]), i.e., at low substrate concentrations when the enzyme is
largely unbound, gave values similar to those previously obtained (differences between 4 and 24%). We therefore consider that the accuracy
of these values was reasonably good.
Sch em e 3
plicated by theoretical difficulties.19 Indeed, the observed
Vmax may represent different chemical events or the
combination of different mechanistic steps and may,
furthermore, be altered by nonproductive binding, strain,
or induced fit. However, all these artifacts are canceled
m
out for the Vmax/K value, and this parameter may thus
be considered as being relevant to the rate-limiting
transition state, even if it still contains binding energy
terms. We therefore examined the substituent effect on
the Vmax/K
m
values as well as, because of the possible
addition (of water) to either C and C carbon atoms, the
substituent effect for each of these two reaction sites.
Hammett plotting of log RVmax/K as well as of log âVmax
were performed as functions of the σ value of each
substituent, as shown in Figure 2.
Our results indicate that, for the benzylic carbon atom
), a good linear relationship exists for the RVmax/K
values, showing substantial negative slopes, i.e., FRVmax/Km
R
â
m
/
K
m
p
F igu r e 2. Hammett plot of log(RVmax) (b), log(âVmax) (O),
log(RV /K ) (9), and log(âV /K ) (0) versus σ for the
max
m
max
m
p
hydrolysis of (R)-p-substituted styrene oxide catalyzed by a
soluble enzymatic extract of S. racemosum. Vmax/K
have been obtained from the substrate dependence of initial
velocities by fitting the equation v ) a(b[S]/([S] + b), with a
and b ) K
m
values
(C
R
m
2
i
)
-1.8 (correlation coefficient r ) 0.88). On the basis
)
Vmax/K
m
m
.
of these results, one can deduce some information on the
rate-limiting step as well as on the intimate mechanism
of the reaction (see below). Also, despite the use of a
crude protein extract, this again strongly suggests that
only one EH was operating. On the other hand, the
Ra te-Lim itin g Step . As mentioned above, it has been
shown that for mammalian EHs the reaction occurs via
a two-step process involving a covalent enzyme-sub-
strate ester intermediate. Furthermore, using presteady
and steady-state kinetic experiments, Tzeng et al. have
concluded recently that, in the rat mEH-catalyzed hydra-
tion of glycidyl 4-nitrobenzoates, the rate-limiting factor
was the second step, i.e., hydrolysis of the enzyme-
substrate ester intermediate.20 Surprisingly, the nega-
tive sign of the FRVmax value we have obtained for the R
magnitude of the substituent effect on the âVmax/K
values was much smaller, and due to the insufficient
accuracy of the Vmax/K value determination, no conclu-
sion could be drawn from these results.
m
m
(19) (a) Kirsch, J . F. In Advances in Linear Free Energy Relation-
ships; Chapman, N. B., Shorter, J ., Eds.; Plenum Press: London, New
York, 1972; Chapter 8, pp 369-400. (b) Fersht, A. In Enzyme Stucture
and Mechanism, 2nd ed.; Freeman, W. H., Ed.; New York, 1985; pp
(20) Tzeng, H.-F.; Laughlin, L. T.; Lin, S.; Amstrong, R. N. J . Am.
Chem. Soc. 1996, 118, 9436.
4
7-120 and 311-346.

3
536 J . Org. Chem., Vol. 63, No. 11, 1998
Moussou et al.
enantiomers of 1-7, together with its important magni-
tude of -2.4 (correlation coefficient 0.94), indicate that
in our case the rate-limiting step cannot be hydrolysis of
an enzyme-substrate ester intermediate, whatever mech-
anism is implied (acid or base catalysis). Indeed, these
values are not compatible with those obtained for such
than the one obtained for the acid hydrolysis of para-
substituted styrene oxides (F
R
) -4.2).23
Con clu sion
We have studied in this work the biohydrolysis of
various racemic or enantiopure para-substituted styrene
oxide derivatives using a crude enzymatic extract pre-
pared from the fungus S. racemosum. This led us to
observe (a) that, in contrast to previously described
examples with several mammalian mEH as well as of a
fungal (A. niger) EH, the regioselectivity of these reac-
tions is directed toward the more substituted carbon atom
for substrates bearing weak electron-donating or -with-
drawing groups, (b) that a striking progressive switch of
regioselectivity was observed when the electron-with-
drawing power increased, and (c) that the values of Vmax
(chemical) base and acid-catalyzed hydrolyses of para-
substituted benzyl acetate derivatives (which can be
considered as being analogues of the hypothetical enzyme-
substrate ester intermediate formed when the nucleo-
philic attack of the enzyme occurs at the benzylic carbon
R
atom C ). Indeed, these have been plotted versus their
σ
-
p
Hammett coefficient, leading to F values of +0.743 and
0.054, respectively.21 On the other hand, the FR/â and
F
RVmax values we have obtained (-2.07 and -2.4, respec-
tively) are consistent with the Hammett correlation
and Vmax/K
m
decreased when the electron-withdrawing
2
2-24
slopes calculated for the regioselectivity
for the rates2
as well as
power increased.
3-25
of nucleophilic oxirane opening reactions
Analysis of these results via Hammett relationships
yields a consistent indication to the existence, in this
case, of a positive charge in the oxirane cleavage rate-
determining transition state. Thus, this enzymatic mech-
anism seems to be best described by a concerted process
implying a general acid activation of the oxirane ring,
together with a nucleophilic attack. Interestingly, this
mechanism thus appears to be very probably different
from the one presumed to be operating for such microso-
mal enzymes. Interestingly, this difference of mechanism
is reminiscent of the proteases, for which there are the
divergent general baselike (e.g., chymotrypsin) and gen-
eral acid-like (e.g., pepsin) mechanisms. Although this
conclusion must still be taken with caution, because of
the use of a crude enzyme and of the lack of any absolute
proof, we consider this explanation to be the most
reasonable in view of the overall results we have ob-
tained. To the best of our knowledge, this is the first
time that such experimental outcomes point to the highly
probable occurrence of such a mechanism for an epoxide
hydrolase. However, since other explanations, such as
variable substrate positioning into the active site, can of
course not be totally ruled out for the moment, further
work is in progress in our laboratory for gaining ad-
ditional experimental clues to support this hypothesis.
cited in the literature. For example, the hydroxide-
catalyzed hydrolysis of para-substituted styrene oxides
derivatives was described to afford FR/â and F
respectively, -1.1 and -0.9,23 and the nucleophilic attack
of benzylamine in ethanol led to a FR/â ) -2.02 and a F
-1.15.24 Thus, our kinetic values, together with the
R
values of,
R
)
observed regioselectivities, strongly suggest that, to the
contrary of the result obtained by Tzeng et al. with rat
mEH,20 the rate-limiting reaction very probably is a
nucleophilic attack at the oxirane ring in the case of the
epoxide hydrolase studied here.
Mech a n ism of th e Oxir a n e Rin g-Op en in g Step .
In theory, two limiting types of mechanism may operate
for the oxirane cleavage: (a) a base-catalyzed nucleophilic
event, without apparent activation of the oxirane ring,
and (b) an acid-catalyzed event, implying an electrophilic
activation of the oxirane ring (by hydrogen-bonding or
protonation for example) concerted with the nucleophilic
attack.
The negative sign, as well as the significant magnitude
of the FRVmax/Km value (-1.8) we have obtained using the
enzymatic extract from S. racemosum, strongly suggest
the occurrence of an electron demand (i.e., the develop-
ment of a positive charge) at the benzylic carbon atom.
In other words, this suggests that, in the course of this
process, the C-O bond breaking is more advanced than
the ester-enzyme bond formation in the transition state
and, therefore, that an electrophilic activation of the
epoxide must be occurring. Thus, a “push-pull” mecha-
nism, i.e., a general acid activation concerted with a
nucleophilic attack of the oxirane ring, in which the “pull”
step would be predominant, seems to be operating in this
case. Consequently, the major factor influencing the
regioselectivity of the oxirane ring nucleophilic attack
should be the electronic stabilization of this incipient
positively charged carbon atom implied in the transition
state. Nevertheless, protonation of the epoxide oxygen
atom by a free hydronium ion may be eliminated because
the magnitude of FRVmax/Km (-1.8) is significantly lower
Exp er im en ta l Section
Gen er a l Meth od s. The strain of S. racemosum used in
this work was purchased at the MUCL collection (no. 28766).
Corn steep liquor (CSL) is from Roquette SA. 1H NMR spectra
were recorded in CDCl solution on a Bruker AC 250. Optical
3
rotation values were measured on a Perkin-Elmer 241C
polarimeter at 589 nm. The absolute configurations of ep-
oxides and diols were assigned via GC analysis on a chiral
stationary phase by comparison with previously described
data.11 Vapor-phase chromatography analyses were performed
using a chiral 25 m capillary column [heptakis(6-O-methyl-
2,3-di-O-pentyl)-â-cyclodextrin]. Determination of the ee of the
diols were performed after derivatization into their acetonide
as described previously.11 HPLC analyses were carried out
using an UV detector and a C18 reversed-phase column (250
×
4.6 mm, 5 µm) using water-acetonitrile as eluent.
Syn th esis of Ep oxid es 1-7 a n d Diols 8-14. Racemic
epoxides 1-7 and diols 8-14 have been previously synthesized
in our laboratory11 (except 2 and 9 purchased from J anssen).
Optically pure epoxides (S)-2 and (R)-2 were purchased from
Fluka. (S)-Epoxides 3-7 have been previously obtained with
ee > 98% in our laboratory by bioconversion of the racemates
using the whole cells of A. niger.11 (R)-7 (ee > 99%) has been
previously synthesized in our laboratory by cyclization of the
optically pure diol (R)-14 obtained by biohydrolysis of (()-7,
(
21) J aff e´ , H. H. Chem. Rev. 1953, 53, 191.
(22) (a) Tamura, N.; Takahashi, K.; Shirai, N.; Kawazoe, Y. Chem.
Pharm. Bull. 1982, 30, 1393. (b) Sugiura, K.; Kimura, T.; Goto, M.
Mutation Res. 1978, 58, 159.
(
D. L. J . Org. Chem. 1993, 58, 924.
(
23) Blumenstein, J . J .; Ukachukwu, V. C.; Moham, R. S.; Whalen,
24) Laird, R. M.; Parker, R. E. J . Am. Chem. Soc. 1961, 83, 4277.
25) Parker, R. E.; Isaacs, N. S. Chem. Rev. 1959, 737.
(

Microbiological Transformations
J . Org. Chem., Vol. 63, No. 11, 1998 3537
using an enzymatic extract from A. niger, followed by recrys-
0.1 mM cysteine, 0.1 mM EDTA, and desalted by passing on
a Sephadex G25 column (6 × 24 cm) equilibrated with the
same buffer. The protein solution obtained was lyophilized
and stored at 4 °C.
2
6
tallization.
The epoxides (S)-1, (R)-1, and (R)-3-6 were
obtained following a two-step synthesis: (i) catalytic asym-
metric dihydroxylation of the corresponding olefin using
commercial AD-mix-R or AD-mix-â (Aldrich) according to
Bioh yd r olysis of Ra cem ic P a r a -Su bstitu ted Styr en e
Oxid es 1-7 Usin g th e En zym a tic Extr a ct of S. r a cem o-
su m . All bioconversions were carried out at 27 °C in a stirred
(750 rpm) minivial (3 mL) containing 0.9 mL of a pH 8.0
sodium phosphate buffer (0.1 M) solution and 14.6 mg of
protein. Reactions were initiated by addition of 0.1 mL of the
racemic epoxide in DMSO (final concentration in the reaction
medium of 3.9-4.9 mM). After 3-7.5 h of reaction, 200 µL of
the medium was extracted with pentane (100 µL) and sub-
jected to direct chiral GC analysis, which allowed us to
determine the ee and the absolute configuration of the residual
epoxide. After addition of 600 µL of MeOH to a second 600
µL sample to stop the reaction and extraction of the residual
epoxide with 600 µL of pentane, the produced diol was
quantified by HPLC analysis using an external calibration to
determine the conversion rate c. The enantiomeric ratio E was
then calculated using the Sih equation E ) ln[(1 - c)(1 - ees)]/
2
7
Sharpless, and (ii) cyclization of the diol according to Kolb
2
8
and Sharpless. The procedure for the cyclization of activated
diols was used for the diols 8 and 10, substituted by the CH
3
and F groups. The procedure used for the inactivated diols
was followed for diols 11-13. The diols and the epoxides were
obtained, respectively, with yields of 73-94% and 73-86%
after purification by bulb-to-bulb distillation. The 1H NMR
spectra of all these compounds were identical to those previ-
1
1
ously described. Optical rotation and chiral GC analysis of
the isolated products are as follows: (R)-1, +27 (c 0.98, PhH,
2
2
11
20
ee 95%); (S)-1, [R]
D
-27 (c 1.01, PhH, ee 95%) [lit. [R]
D
2
0
+
25.5 (c 1.3, PhH, ee > 98%) for (R)-1]; (R)-3, [R]
D
-17 (c
1
1
20
1
9
+
1
9
.03, CHCl
6%)]; (R)-4, [R]
3
, ee 97%) [lit. [R]
D
-15.1 (c 0.45, CHCl , ee
3
2
0
11
20
D
-24 (c 1.08, CHCl
3
3
, ee 97%) [lit. [R]
D
2
0
19.3 (c 1.16, CHCl
, ee > 98%) for (S)-4]; (R)-5, [R]
D
-22 (c
1
1
20
.00, CHCl
3
, ee 98%) [lit. [R]
D
-12.9 (c 1.04, CHCl , ee >
3
2
0
11
20
18
8%)]; (R)-6, [R]
D
-6 (c 1.26, EtOH, ee 96%) [lit. [R]
D
+7.3
ln[(1 - c)(1 + ees)]. The ee and the absolute configuration
2
3
(
c 1.09, EtOH, ee > 98%) for (S)-6]; (R)-8, [R]
D
-67 (c 0.90,
of the diol were further determined by chiral GC analysis after
saturation with NaCl, ether extraction of diol, purification
through silica gel, and derivatization.
1
1
20
CHCl
3
, ee > 97%) [lit. [R]
D
3
-50 (c 0.87, CHCl , ee 76%)];
2
3
11
20
(
(
R)-10, [R]
D
-63 (c 1.06, CHCl
3
, ee > 98%) [lit. [R]
D
-49
, ee
3
, ee 79%)]; (R)-12,
2
3
c 1.07, CHCl
3
, ee 81%)]; (R)-11, [R]
D
-60 (c 1.00, CHCl
-52.1 (c 1.03, CHCl
3
Bioh yd r olysis of Op tica lly P u r e P a r a -Su bstitu ted
Styr en e Oxid es 1-7 Usin g th e En zym a tic Extr a ct of S.
r a cem osu m . The biohydrolyses were carried out as described
above in the presence of 10% DMSO for enzymatic extract
concentrations of 20-50 mg/mL and for substrate concentra-
tions of 3.5-4.5 mM. For each biohydrolysis, at two different
conversion rates, between 15 min and 2 h of the reaction, the
ee of the formed diols were analyzed as described above.
Under these conditions, the nonenzymatic hydrolysis was
insignificant.
1
1
20
>
98%) [lit. [R]
-47 (c 1.02, CHCl
, ee 79%)]; (R)-13 [R]
D
11
20
[
R]23
D
3
, ee > 98%) [lit. [R]
D
-37.2 (c 1.03,
-23 (c 0.80, EtOH, ee > 98%)
-17.1 (c 0.53, EtOH, ee 76%)].
En zym e P r ep a r a tion . S. racemosum was grown in a 2 L
2
3
CHCl
3
D
1
1
20
[lit. [R]
D
fermentor (SETRIC) containing 1.4 L of culture medium (20
g of CSL, 10 g of glucose in 1 L of tap water). Before
sterilization (20 min at 115 °C), 0.28 g of Pluronic PE 8100
(BASF) and 70 µL of antifoam silicone 426R (PROLABO) were
added to prevent overflowing during the growth. The medium
was maintained at 27 °C, stirred at 700 rpm, and aerated at
Deter m in a tion of th e Kin etic P a r a m eter s. All reactions
were carried out at 27 °C in a stirred (750 rpm) Eppendorf
tube (1.5 mL) containing 1-4 mg of enzymatic extract dis-
solved in 360 µL of a pH 8.0 sodium phosphate buffer (0.1 M)
solution. Reactions were initiated by addition of 40 µL of the
(R)-epoxide in DMSO. The final concentration ranges of the
epoxides were between 0.6 mM and their maximal solubility
in the reaction medium, i.e., 30, 37, 31, 23, 7, 36, 5 mM,
respectively, for the epoxides 1-7. After 10 min, the reactions
2
0 ( 5 L/h. The broth was inoculated with 5 mL of a
suspension of mycelium and spores in physiologic water (NaCl
.9%) from an agar slant of S. racemosum. After incubation
0
for 70 h, the mycelium (dry mass 9.8 g) was filtered off, washed
with water, and suspended in 700 mL of 20 mM ammonium
acetate buffer pH 7.0 containing 1 mM cysteine and 1 mM
EDTA. After addition of 1.75 mL of a 200 mM solution of
phenylmethylsulfonyl fluoride in acetone, the cells were broken
by two passages through a Cell Disrupt (CELL-D) at a
pressure of 1800 bar. Unbroken cells and cells debris were
separated by centrifugation (9600g for 30 min at 4 °C) and
discarded. After addition of solid ammonium sulfate at 25%
of the saturation to the supernatant, the solution was stirred
for 30 min and centrifuged at 9600g for 30 min. After having
discarded insoluble particles on the surface of the supernatant,
ammonium sulfate at 80% of the saturation was added to this
one, and this solution was again stirred for 30 min and
centrifuged at 9600g for 30 min. The pellet was dissolved in
3
were stopped by addition of 400 µL of CH CN. After centrifu-
gation (3 min, 10000 rpm), the produced diols were quantified
by HPLC analysis of the medium using an external calibration
for each diol. For all these reactions, the conversion was below
5%. Initial rates of hydration were calculated at each initial
substrate concentration from the amounts of diol formed. The
m m
apparent kinetic parameters (K , Vmax and Vmax/K ) were
calculated by nonlinear regression of the Michaelis-Menten
equation (see Table 2), using the commercial J andel software
Sigmaplot (Marquardt-Levenberg algorithm).
2
5 mL of 10 mM ammonium acetate buffer pH 7.0 containing
Ack n ow led gm en t. This work is part of the Ph.D.
thesis presented by P. Moussou. Financial support by
the Soci e´ t e´ Rh oˆ ne-Poulenc (BIOAVENIR programs
including a Ph.D. stipend to P.M.) and by the European
Community (BIOC4-CT95-0005) is greatly acknowledged.
(
26) Pedragosa-Moreau, S.; Morisseau, C.; Baratti, J . C.; Zylber, J .;
Archelas, A.; Furstoss, R. Tetrahedron 1997, 53, 9707.
27) Sharpless, K. B.; Amberg, W.; Bennani, Y. L.; Crispino, G. A.;
Hartung, J .; J eong, K.-S.; Kwong, H.-L.; Morikawa, K.; Wang, Z.-M.;
Xu, D.; Zhang, X.-L. J . Org. Chem. 1992, 57, 2768.
(
(28) Kolb, H. C.; Sharpless, K. B. Tetrahedron 1992, 48, 10515.
J O9714371