Enzymatic transformations. Part 58: Enantioconvergent biohydrolysis of styrene oxide derivatives catalysed by the Solanum tuberosum epoxide hydrolase

Article abstract of DOI:10.1016/j.tetasy.2004.06.032

The biohydrolysis of four racemic styrene oxide derivatives has been explored, using the (recombinant) Solanum tuberosum epoxide hydrolase. Interestingly, this enzyme showed a marked tendency to operate a so-called enantioconvergent process, thus affording the corresponding (R)-diol in a nearly quantitative yield and good to excellent ee. We have demonstrated that this is due to the fact that the (S)-enantiomer of these epoxides was preferably attacked at the (benzylic) more substituted carbon atom, whereas the (R)-epoxide was attacked at the (terminal) less substituted carbon atom. The thus obtained meta- and para-chlorostyrene diol derivatives are important building blocks in the synthesis of various biologically active molecules. A nine cycles repeated batch reactor was performed starting from racemic meta-chlorostyrene oxide and afforded a 100% analytical yield (88% preparative) of the corresponding diol, obtained with ees as high as 97%.

Full text of DOI:10.1016/j.tetasy.2004.06.032

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 15 (2004) 2801–2805
Enzymatic transformations. Part 58: Enantioconvergent
biohydrolysis of styrene oxide derivatives catalysed by
the Solanum tuberosum epoxide hydrolase
Maria I. Monterde,a Murielle Lombard,a Alain Archelas,a Annette Cronin,b
Michael Arandb and Roland Furstossa,*
a
´ ´ ´ ´ ´
Groupe Biocatalyse et Chimie Fine, FRE 2712 associee au CNRS, Universite de la Mediterranee, Faculte des Sciences de Luminy,
Case 901, 163 avenue de Luminy, 13288 Marseille Cedex 9, France
bInstitute of Pharmacology and Toxicology, University of Wu¨rzburg, Versbacher Strasse 9, D-97078 Wu¨rzburg, Germany
Received 20 May 2004; accepted 10 June 2004
Available online 3 August 2004
Abstract—The biohydrolysis of four racemic styrene oxide derivatives has been explored, using the (recombinant) Solanum tubero-
sum epoxide hydrolase. Interestingly, this enzyme showed a marked tendency to operate a so-called enantioconvergent process, thus
affording the corresponding (R)-diol in a nearly quantitative yield and good to excellent ee. We have demonstrated that this is due to
the fact that the (S)-enantiomer of these epoxides was preferably attacked at the (benzylic) more substituted carbon atom, whereas
the (R)-epoxide was attacked at the (terminal) less substituted carbon atom. The thus obtained meta- and para-chlorostyrene diol
derivatives are important building blocks in the synthesis of various biologically active molecules. A nine cycles repeated batch
reactor was performed starting from racemic meta-chlorostyrene oxide and afforded a 100% analytical yield (88% preparative) of
the corresponding diol, obtained with ees as high as 97%.
Ó 2004 Elsevier Ltd. All rights reserved.
1. Introduction
vicinal diols is a most actively studied topic. This is
due to the fact that these compounds are essential chiral
intermediates for the production of biologically active
molecules, such as SR 58611A or AJ-9677, two exam-
ples of various b3-adrenergic receptor agonists, which
are known to require a 3-chlorophenyl moiety2 or Eli-
prodil,3 an effective NMDA receptor antagonist known
to possess neuroprotective properties, whose structure
involves a 4-chlorophenyl entity (Scheme 1).
Due to the ongoing search for increased safety and effi-
ciency of drugs, and to the release of new FDAÕs mar-
keting guidelines, the involvement of contemporary
pharmaceutical chemistry in the elaboration and com-
mercialisation of enantiopure compounds is nowadays
beyond discussion.1 In this context, the synthesis of
enantiopure epoxides, as well as of their corresponding
OH
OH
(R)
CO2Et
H
OH
(R)
N
(R)
H
N
Cl
N
O
CH3
N
H
Cl
H
Cl
F
AJ-9677
SR 58611A
Eliprodil
Scheme 1.
*
Corresponding author. Tel.: +33-491-82-91-55; fax: +33-491-82-91-45; e-mail: furstoss@luminy.univ-mrs.fr
0957-4166/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2004.06.032

2802
M. I. Monterde et al. / Tetrahedron: Asymmetry 15 (2004) 2801–2805
Therefore, the elaboration of efficient and cost effective
processes allowing for the preparation of these chiral
synthons in enantiopure form are of utmost importance,
with several methodologies already explored.1 These in-
clude the hydrolytic kinetic resolution (HKR) of race-
mic epoxides, one of the most promising methods in
this context being a biocatalytic approach. This environ-
mentally safe methodology is particularly attractive
since it allows us to minimise the cost of resources and
the production of (potentially toxic) waste in industrial
applications. It is based on the use of dedicated enzymes,
that is, epoxide hydrolases (EH) (EC 3.3.2.3), which spe-
cifically catalyse the hydrolysis of an epoxide into the
corresponding vicinal diol.4–7 We, and others, have de-
scribed over the last decade that, in many cases, this
can be achieved in an enantioselective manner. Nowa-
days, an even more stringent industrial demand relates
to the possibility to elaborate enantioconvergent proc-
esses, thus allowing us to approach the ideal Ô100% yield,
100% eeÕ goal starting from a racemic mixture. This is
obviously of utmost importance for process cost effi-
ciency, since it allows us to avoid the loss of 50% of
the starting product, a feature intrinsic to any resolution
process. In this context we have demonstrated that in
some cases, it is possible to take advantage of the poten-
tial Ôstereochemical flexibilityÕ of the EHs approach, and
to set up a single process implying combination of com-
plementary enantio- and regioselectivities. This can be
obtained either by combining two different but enantio-
and regiocomplementary epoxide hydrolases8 or by
using one single enzyme.9,10 In this last case, the chosen
epoxide hydrolase has to display an opposite regioselec-
tivity of the oxirane ring opening for each enantiomer of
a racemic mixture. This aspect offered by EHs is an
intriguing and potentially highly useful property of such
enzymes. To the best of our knowledge, no chemical
counterpart of this type has yet been described.
behaviour of these four epoxides when submitted to bio-
hydrolysis by the above-cited StEH. This was in partic-
ular aimed at evaluating the influence of the chlorine
atom positioning on the enantioconvergency, a phenom-
enon we had previously observed in preliminary studies
conducted on substrate 4.8 Ideally, this would allow the
100% yield preparation of diols 1d–4d, in their enantio-
pure forms.
Our initial experiments were designed to find the most
suitable conditions for the biohydrolysis of rac-1–4. In
a first set of experiments, four different conditions, that
is, a combination of either deionised water or pH7 phos-
phate buffer with the use of either DMF or DMSO as
co-solvent (to increase the substrate solubility in the
aqueous phase), were explored. Thus, as a model exper-
iment, biohydrolysis of meta-chlorostyrene oxide 3 was
carried out for each one of these possible combinations
at 28°C, 4mM substrate concentration, and using a sub-
strate over enzyme ratio (w/w) of 4.6. Figure 1 shows the
evolution of the enantiomeric excess (ee) and conversion
ratio (c) versus time dependent biohydrolysis of rac-3,
using each one of these four media. The best results were
obtained with the deionised H2O/DMSO combination.
This was therefore used for the following analytical
studies. Increasing either the DMSO or DMF propor-
tion (upto 30%), as well as using the three other combi-
nations, led to noticeable decreases in the reaction rate.
100
90
80
70
60
50
40
30
20
10
0
0
200
400
600
800
1000
1200
1400
Herein we report on our recent work aimed at exploring
the Ôenantioconvergent potentialÕ of the Solanum tubero-
sum epoxide hydrolase (StEH), a monomeric recombi-
nant protein of plant origin,11,12 applied to four
styrene oxide derivatives 1–4 (Scheme 2).
time (min)
Figure 1. Analytical follow-upof the biohydrolysis of rac-3 at 28°C
using StEH, under different experimental conditions: (a) H2O, DMSO
(1%): j ee, h conversion ratio; (b) Buffer, DMSO (1%): m ee, n
conversion ratio; (c) H2O, DMF (1%): d ee, s conversion ratio; (d)
Buffer, DMF (1%): r ee, e conversion ratio.
O
O
O
O
2.2. Analytical scale biohydrolysis
Cl
Cl
Experiments conducted on substrates 1–4 using the S.
tuberosum EH led to the results reported in Table 1.
The initial rate measured for biohydrolysis of rac-1
(4mM aqueous solution, 20°C) indicated an enzymatic
activity of about 2U/mg and, after about 40min, the
conversion ratio reached 53%. The unreacted (remain-
ing) epoxide showed an ee as high as 94%. It was shown
to be an (R)-configuration by comparison with a com-
mercial sample. The E value, calculated on the basis of
Cl
3
1
2
4
Scheme 2. Structure of the racemic epoxides 1–4 studied.
2. Results and discussion
2.1. Preliminary experiments
Owing to the interest of enantiopure styrene oxide deriv-
atives 1–4 as well as of their corresponding vicinal diols
1d–4d, as key-stepbuilding blocks for the synthesis of
various pharmaceutical compounds, we explored the
The StEH enzymatic extract used herein was prepared by over-
expression in Escherichia coli BL21(DE3) (unpublished results). This
was used as a crude extract for the analytical work and in a purified
form for the reactor process.

M. I. Monterde et al. / Tetrahedron: Asymmetry 15 (2004) 2801–2805
Table 1. Parameters of the biohydrolysis of 1–4 (4mM, 28°C) using S. tuberosum EH
2803
Substrate
t (h)
c (%)
Eeepoxi
Eediol
Eediol
(total conv.)a
E
Activityb
a(S)
b(S)
a(R)
b(R)
Chem.
hydr %/h
rac-1c
rac-2
rac-3
rac-4
0.7
5.5
2
53
45
57
50
94
31
95
30
94
88
86
29
91
74
30
2
98
62
97
97
2
38
3
7
25
6
93
75
94
92
1.5
0.5
0.5
1.2
2.5
0.01
0.4
1.8
61
>99
6
70
1.4
3
8
a Ee of the formed diol at c=100%.
b U/mg enzymatic extract.
c Biocatalysis conducted at 20°C.
the remaining substrate ee and of the conversion ratio,
was about 30. After derivatisation into its acetonide
and GC analysis, the resulting vicinal diol 1d was shown
(by chiral GC analysis) to also be 95% ee. However, this
ee surprisingly only slightly decreased to about 86% at
total conversion (c=100%). Moreover, the formed diol
1d appeared to be also of (R)-configuration. This result
obviously highlighted the occurrence of an unexpected
enantioconvergent process.
ous explanation would be that, in each of these cases,
the enzyme displayed an opposite regioselectivity for
each enantiomer of the racemic mixture. Thus, the (S)-
enantiomer should be essentially attacked at the more
substituted (benzylic) carbon atom, thus leading to the
(R)-diol by inversion of configuration, whereas the
(R)-enantiomer should be essentially attacked at the less
substituted carbon atom, affording the diol of un-
changed (R)-configuration. A detailed analysis of the
specific regioselectivity [i.e., the determination of the
so-called regioselectivity coefficients a(R), b(R), a(S)
and b(S)] had therefore to be achieved for total under-
standing of the process.9 This was conducted using each
one of the two enantiomers of substrates 1–4. Our
results are reported in Table 1. Interestingly enough, it
was observed that the (S)-enantiomer of substrates 1,
3 and 4 were indeed very preferably attacked at the ben-
zylic position [a(S) > 97%], whereas the (R)-enantiomer
was mostly attacked at the terminal carbon atom
[b(R) > 92%]. This behaviour is illustrated for rac-3 on
Scheme 3. The same tendency stayed true, although to
a lesser degree, for epoxide 2 where the regioselectivity
was more balanced between the two a and b carbon
atoms. We also found that, an opposite, spontaneous
(chemical) hydrolysis essentially led to an attack at the
benzylic position in all cases. However, we also have
determined that this hydrolysis remained rather negligi-
ble (Table 1), thus precluding any noticeable interfer-
ence with the enzymatic hydrolysis regioselectivity
after a few hours.
Racemic ortho-chlorostyrene oxide 2 behaved quite dif-
ferently when compared to rac-1. Thus, the enzymatic
activity was estimated to only be 0.01U/mg. Therefore,
about 5.5h were necessary to reach a conversion ratio
of 45%. Both the ee of the residual epoxide 2 as well
as of the formed diol 2d were rather low (30%) at this
stage. A similar ee was found for 2d at completion of
the reaction. Interestingly however, both these products
were shown to be of an (R)-configuration as in the case
of 1. The E value was estimated to be only about 2.5.
Obviously, the introduction of a chlorine atom at the
ortho position severely disturbed the enzymatic hydroly-
sis.
Biohydrolysis of meta-chlorostyrene oxide rac-3 simi-
larly revealed a rather low E value (E ꢀ 6). However,
contrary to the previous case, a much better enzymatic
activity was observed (0.4U/mg). Thus, after 2 h reac-
tion, biohydrolysis of rac-3 led to a conversion ratio of
57%. The remaining epoxide exhibited a 61% ee while
the obtained diol exhibited an ee as high as 94%. Again,
both these products were of an (R)-configuration, as
determined by comparison with commercial samples.
It is noteworthy however, that at total conversion, the
formed diol still exhibited an ee of 91%.
OH
O
(S)
StEH
OH
(R)
H2O/1%DMSO
α(S)=97%
ee= 94%
OH
Cl
Cl
Finally, para-chloro styrene 4 behaved similarly to 1. An
enzymatic activity of about 1.8U/mg was estimated
from the initial rate measurements and a 50% conver-
sion ratio reached after 1.4h. The remaining epoxide
exhibited an ee higher than 99% whereas the obtained
diol exhibited an ee of 94%. Again, both these products
were shown to be of an (R)-configuration, as determined
by comparison with commercial samples. At total con-
version, the formed diol still exhibited an ee of 74%.
O
(R)
StEH
OH
(R)
H2O/1%DMSO
ee= 89%
β(R)=94%
3d
Cl
3
Cl
Scheme 3. Enantioconvergent biohydrolysis of rac-3 due to comple-
mentary regioselectivity.
2.4. Optimisation of substrate concentration
2.3. Regioselectivity coefficients
Owing to the high potential of the above-described re-
sults for setting upan enantioconvergent process, we
decided to explore the possibility of performing a pre-
parative scale biohydrolysis of one of these substrates.
Throughout this study, the most surprising observation
was that for each substrate, the StEH appeared to afford
a good to excellent enantioconvergency. The most obvi-

2804
M. I. Monterde et al. / Tetrahedron: Asymmetry 15 (2004) 2801–2805
The substrate of choice was rac-3, due to (a) the interest
of the corresponding (R)-diol 3d for the synthesis of var-
ious b3-adrenergic receptor agonists;2 (b) its low E value
(and therefore the moderate biohydrolysis rate differ-
ence between the two enantiomers); and (c) its reasona-
ble enzymatic activity. Obviously, one important
parameter for implementation of such a process is the
determination of the maximum substrate concentration
usable. We have shown previously, that when using
the A. niger EH, in certain cases, a substrate concentra-
tion as high as 500g/L could be used.13 Exploration of
this possibility was therefore conducted by increasing
the concentration of rac-3 upto 200mM (34.5g/L).
Our exploratory results indicated that contrary to our
previous results, a concentration limitation appeared
at a value higher than 75mM of substrate. Due to pre-
vious experiences, we felt that this limitation could be
due to product inhibition. We therefore tested this
hypothesis by conducting several experiments (at 2mM
substrate concentration), adding various concentrations
of the (formed) (R)-3 diol enantiomer at the start of the
biohydrolysis. The results shown in Figure 2 clearly indi-
cate that this concentration limitation was indeed due to
product inhibition. They also indicate that a 75mM
(11.6g/L) substrate concentration (which corresponds
to a concentration of about 13g/L of formed diol) was
a reasonable compromise for further elaboration of an
enzymatic reactor.
sponded to an overall concentration of about 10g/L of
substrate. The mixture was then vigorously stirred (mag-
netic stirrer). In order to further minimise any spontane-
ous hydrolysis of the substrate and favour enzyme
stability, the process was conducted at 20°C (instead
of 28°C as used for the analytic experiments). Due to
low solubility of 3 in water, this was done at the start
of a biphasic system. However, the solution became
homogeneous as the reaction proceeded, since the diol
formed is totally soluble in water at that concentration.
The reaction was followed by HPLC analysis, showing
that total conversion occurred after a 1.5h reaction time
for this first run. The aqueous (homogeneous) solution
was filtered through an ultrafiltration membrane, which
allowed us to recover the enzyme (as a concentrated
solution). This was re-used to perform the next biohy-
drolysis run on another 340mg substrate sample, using
the same reactor and experimental conditions. Nine
runs were thus performed successively, using the same
enzyme, which corresponds to a biohydrolysis of a total
of 3g (19.4mmol) of rac-3. The nine aqueous solutions
containing the product 3d were pooled and extracted
(ethyl acetate). This afforded 3.5g of 3d (100% analytical
yield), which were purified by bulb-to-bulb distillation.
A total amount of 3g of pure 3d (waxy solid) was
thus obtained, which corresponds to an 88% prepara-
tive yield. As expected, the ee of the thus obtained
22
D
(R)-3d was excellent (97%) {½aŠ ¼ À23:4 (c 1.55,
26
EtOH); lit.15 for (S)-3d ½aŠ ¼ þ21:1 (c 1.31, EtOH),
D
ee=91%}.
100.00
90.00
80.00
70.00
60.00
50.00
40.00
30.00
20.00
10.00
0.00
3. Conclusion
Herein we have reported the biohydrolysis of the substi-
tuted styrene oxide derivatives 1–4 catalysed by the S.
tuberosum epoxide hydrolase, overexpressed in an
appropriate E. coli strain. We observed that each one
of the two enantiomers of these substrates were attacked
by this enzyme with a different and complementary regio-
selectivity. Thus, the (S)-enantiomers were preferably
attacked at the benzylic position, whereas the (R)-
enantiomers were mostly attacked at the terminal
carbon atom. As a consequence, this led to a so-called
enantioconvergent process. Based on this very interest-
ing property, we were able to perform the preparative
scale biohydrolysis of 3g (19.4mmol) of rac-3 in a re-
peated batch reactor, which was run over nine cycles
using the same enzyme sample. The corresponding diol
(R)-3d was thus obtained with an excellent yield (100%
analytic, 88% preparative), in nearly enantiopure form
(ee 97%). This illustrates the highly interesting potential
offered by the enantio-dependent regioselectivity of this
enzyme, which allows to efficiently overcome the 50%
yield limitation theoretically linked to any resolution
process. It should be stressed that this biohydrolysis
was performed using plain water instead of a buffer solu-
tion, thus affording a so-called salt-free process. More-
over, it should also be emphasized that, the best of
our knowledge such an enantio- and regio-complemen-
tary behaviour has never been described for a metal-
based catalysed hydrolytic kinetic resolution of an epox-
ide. Work is currently in progress in our laboratory in
0
200 400 600 800 1000 1200 1400
time (min)
Figure 2. Influence of the initial (R)-3 diol concentration on the
biohydrolysis rate. [(R)-3 diol]: j=0mM; m=75mM (13g/L);
d=100mM (17.2g/L); r=200mM (34.5g/L).
2.5. Preparative scale biohydrolysis of rac-3 using a
repeated batch reactor
Owing to the above described results, we decided to run
a preparative scale repeated batch reactor, using the
most appropriate experimental conditions for the biohy-
drolysis of rac-3. Thus, 340mg of rac-3 dissolved in
300lL of DMSO were placed in a 100mL reactor, to-
gether with 22mL of plain water. The StEH was added
(960U)à,14 as an aqueous solution (10mL, containing
10% of glycerol for enzyme stabilisation), which corre-
à This enzymatic activity was measured against styrene oxide, using the
same methodology as the one we have described previously for the A.
niger EH.14

M. I. Monterde et al. / Tetrahedron: Asymmetry 15 (2004) 2801–2805
2805
4. Archelas, A.; Furstoss, R. Curr. Opin. Chem. Biol. 2001, 5,
112–119.
5. de Vries, E. J.; Janssen, D. B. Curr. Opin. Biotechnol. 2003,
14, 414–420.
order to further explore the potential of this new epox-
ide hydrolase.
6. Smit, M. S. Trends Biotechnol. 2004, 22, 123–129.
7. Steinreiber, A.; Faber, K. Curr. Opin. Chem. Biol. 2001,
12, 552–558.
8. Manoj, K. M.; Archelas, A.; Baratti, J.; Furstoss, R.
Tetrahedron 2001, 57, 695–701.
9. Moussou, P.; Archelas, A.; Baratti, J.; Furstoss, R.
Tetrahedron: Asymmetry 1998, 9, 1539–1547.
10. Steinreiber, A.; Mayer, S. F.; Faber, K. Synthesis 2001, 13,
2035–2039.
Acknowledgements
This work was partly supported by the European Com-
munity Contract ÔEngineering integrated biocatalyst for
the production of chiral epoxides and other pharmaceu-
tical intermediatesÕ (No QLK3-CT2000-00426). Two
of us (M.M. and M.L.) are very grateful to the EC for
financial support in this context.
11. 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.
12. Morisseau, C.; Beetham, J. K.; Pinot, F.; Debernard, S.;
Newman, J. W.; Hammock, B. D. Arch. Biochem. Biophys.
2000, 378, 321–332.
References
1. Breuer, M.; Ditrich, K.; Habicher, T.; Hauer, B.; Keßeler,
M.; Sturmer, R.; Zelinski, T. Angew. Chem., Int. Ed. Engl.
¨
2004, 43, 788–824.
13. Monfort, N.; Archelas, A.; Furstoss, R. Tetrahedron 2004,
60, 601–605.
2. Harada, H.; Hirokawa, Y.; Suzuki, K.; Hiyama, Y.; Oue,
M.; Kawashima, H.; Yoshida, N.; Furutani, Y.; Kato, S.
Bioorg. Med. Chem. Lett. 2003, 13, 1301–1305.
3. Pabel, J.; Ho¨fner, G.; Wanner, K. T. Bioorg. Med. Chem.
Lett. 2000, 10, 1377–1380.
14. Mateo, C.; Archelas, A.; Furstoss, R. Anal. Biochem.
2003, 314, 135–141.
15. Schaus, S. E.; Brandes, B. D.; Larrow, J. F.; Tokunaga,
M.; Hansen, K. B.; Gould, A. E.; Furrow, M. E.;
Jacobsen, E. N. J. Am. Chem. Soc. 2002, 124, 1307–1315.