Microbiological transformations. 47. A step toward a green chemistry preparation of enantiopure (S)-2-, -3-, and -4-pyridyloxirane via an epoxide hydrolase catalyzed kinetic resolution

Article abstract of DOI:10.1021/jo001406x

The biocatalyzed hydrolytic kinetic resolution of 2-, 3-, and 4-pyridyloxirane by the Aspergillus niger epoxide hydrolase (EH) has been explored. This was used to perform a gram scale preparation of these epoxides of (S) absolute configuration using a process performed at a concentration as high as 10 g/L (82 mM). All three epoxides have been obtained in a nearly enantiopure form (ee > 98%). Interestingly, it was shown that this biotransformation could be achieved using plain water instead of buffer solution, an important improvement as far as downstream processing of an eventual industrial process is concerned. Neither of these substrates could be obtained in reasonable enantiomeric purity and yield using the nowadays most efficient metal-based catalysts.

Full text of DOI:10.1021/jo001406x

538
J . Org. Chem. 2001, 66, 538-543
Micr obiologica l Tr a n sfor m a tion s. 47. A Step tow a r d a Gr een
Ch em istr y P r ep a r a tion of En a n tiop u r e (S)-2-, -3-, a n d
-4-P yr id yloxir a n e via a n Ep oxid e Hyd r ola se Ca ta lyzed Kin etic
Resolu tion
Yvonne Genzel,† Alain Archelas,*,† Q. B. Broxterman,‡ Birgit Schulze,‡ and Roland Furstoss†
Groupe Biocatalyse et Chimie Fine, ESA 6111 associe´e au CNRS, Universite´ de la Me´diterrane´e,
Faculte´ des Sciences de Luminy, Case 901, 163 avenue de Luminy, 13288 Marseille Cedex 9, France, and
DSM Research, Life Science Chemistry and Catalysis, P.O. Box 18, 6160 MD Geleen, The Netherlands
archelas@luminy.univ-mrs.fr
Received September 22, 2000
The biocatalyzed hydrolytic kinetic resolution of 2-, 3-, and 4-pyridyloxirane by the Aspergillus
niger epoxide hydrolase (EH) has been explored. This was used to perform a gram scale preparation
of these epoxides of (S) absolute configuration using a process performed at a concentration as
high as 10 g/L (82 mM). All three epoxides have been obtained in a nearly enantiopure form (ee >
98%). Interestingly, it was shown that this biotransformation could be achieved using plain water
instead of buffer solution, an important improvement as far as downstream processing of an eventual
industrial process is concerned. Neither of these substrates could be obtained in reasonable
enantiomeric purity and yield using the nowadays most efficient metal-based catalysts.
In tr od u ction
various examples of EHs have been recently shown to
exhibit complementary substrate- as well as enantio- and/
or regioselectivity. Moreover, some of them were de-
scribed to even display an opposite regioselectivity for
each enantiomer of one particular substrate. Combination
of some of these aspects led, in certain cases, to the
possibility to elaborate highly valuable enantioconvergent
processes allowing one to overcome the 50% yield limita-
tion intrinsic to a resolution process, as illustrated for
the synthesis of some enantiomerically pure bioactive
compounds.3,5 Of course, such methods may be even more
attractive if they allow the preparation of epoxides that
cannot be obtained in enantiopure form by using the
chemical approaches mentioned above.6,7
Enantiopure epoxides are nowadays recognized as
highly valuable chiral synthons. Therefore, some elegant
conventional chemistry methods, mainly based on the use
of transition-metal-based catalysts, have been developed
lately in order to synthesize these important building
blocks in enantiopure form.1 As a result of, in particular,
increased environmental as well as regulatory pressure
on chemical processes, other approaches using, for in-
stance, biological catalysts may constitute interesting
alternative approaches to these techniques.
Recently novel methods have been described that allow
the preparation of enantiopure epoxides using a biocata-
lyzed hydrolytic kinetic resolution approach. This strat-
egy is based on the use of “new enzymes”, i.e., microbial
epoxide hydrolases (EHs) (EC 3.3.2.3).2 In the recent
years, such biocatalysts have been shown to be ubiquitous
in nature, to be cofactor-independent, and to catalyze the
enantioselective addition of water to an epoxide, thus
leading to the corresponding diol and to the recovery of
the less reactive substrate enantiomer. Both products can
thus be, ideally, obtained in enantiopure form. Although
not yet available on preparative scale and therefore not
commercially available, such enzymes are highly promis-
ing for synthetic, i.e., industrial application. Indeed,
Other general advantages of the use of an epoxide
hydrolase based process are the facts that it can in some
cases be performed at high substrate concentration,4a,5a,8
uses water as the only solvent and reactant, and only
implies an enzyme, i.e., a natural and biodegradable
(3) Pedragosa-Moreau, S.; Archelas, A.; Furstoss, R. J . Org. Chem.
1993, 58, 5533.
(4) See for instance: (a) Cleij, M.; Archelas, A.; Furstoss, R. J . Org.
Chem. 1999, 64, 5029. (b) Pedragosa-Moreau, S.; Morisseau, C.; Zylber,
J .; Archelas, A.; Baratti, J .; Furstoss, R. J . Org. Chem. 1996, 61, 7402.
(c) Krenn, W.; Ospirian, I.; Kroutil, W.; Braunegg, G.; Faber, K.
Biotechnol. Lett. 1999, 21, 687.
(5) (a) Pedragosa-Moreau, S.; Morisseau, C.; Baratti, J .; Zylber, J .;
Archelas, A.; Furstoss, R. Tetrahedron 1997, 53, 9707. (b) Orru, R. V.
A.; Kroutil, W.; Faber, K. Tetrahedron Lett. 1997, 38, 1753.
(6) As can be seen in the above cited reference, numerous examples
of enantioselective epoxidation of prochiral substrates have been
described, most of them using heavy-metal-based catalysts. However,
the obtained product rarely show excellent ee’s (>98%). Only few
applications of the new chemical HKR strategy have been described
up to now. For a direct comparison of such strategies on one particular
compound, see for instance: Brandes, B. D.; J acobsen, N. E. Tetrahe-
dron: Asymmetry 1997, 23, 3927-3993.
(7) Recently, the biocatalyzed direct epoxidation of 1 and 3 with a
styrene monooxygenase overexpressing bacteria was also described,
albeit in very low yield: Di Gennaro, P.; Colmegna, A.; Galli, E.; Sello,
G.; Pelizzoni, F.; Bestetti, G. Appl. Environ. Microbiol. 1999, 65, 2794.
(8) Manoj, K. M.; Archelas, A.; Baratti, J .; Furstoss, R. Tetrahedron
2000, in press.
* Ph: +33 04 91 82 91 58. Fax: +33 04 91 82 91 45.
† Universite´ de la Me´diterrane´e.
‡ DSM Research.
(1) (a) Ito, Y. N.; Katsuki, T. Bull. Chem. Soc. J pn. 1999, 71, 603.
(b) Savle, P. S.; Lamoreaux, M. J .; Berry, J . F.; Gandour, R. D.
Tetrahedron: Asymmetry 1998, 9, 1843. (c) Kolb, H. C.; VanNieuenhze,
M. S.; Sharpless, K. B. Chem. Rev. 1994, 94, 2483. (d) Schurig, V.;
Betschinger, F. Chem. Rev. 1992, 92, 873.
(2) (a) Archelas, A.; Furstoss, R. In BiocatalysissFrom Discovery
to Application; Fessner, W. D., Ed.; Springer-Verlag: Berlin, Heidel-
berg, 1999; pp 159-192. (b) Archelas, A.; Furstoss, R. Trends Biotech-
nol. 1998, 16, 108. (c) Svaving, J .; de Bont, J . A. M. Enzymol. Microb.
Technol. 1998, 22, 19. (d) Orru, R. V. A.; Archelas, A.; Furstoss, R.;
Faber, K. Adv. Biochem. Eng. Biotechnol. 1998, 63, 145. (e) Archer, I.
V. J . Tetrahedron 1997, 53, 15617.
10.1021/jo001406x CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/23/2000

Microbiological Transformations
J . Org. Chem., Vol. 66, No. 2, 2001 539
Ta ble 1. Com p a r ison of Ch em ica l Meth od s for th e
Asym m etr ic Syn th esis of (R)-1, -2, a n d -3 (An a lytica l
Sca le)
Sch em e 1
J acobsen
J acobsen
HKRc
Sharpless
epoxidationa
dihydroxylationd
ee temp time ee temp time ee temp yield
product (%) (°C)
(h) (%) (°C)
(h) (%) (°C)
(%)
1
9
4
18
1
2.5
18
4
20
70
79
0
8
1b
2
23 -78
34
31
4
4
5
0
20
20
70
90
72
63
0
0
nd
10
protein, as catalyst. This can theoretically be produced
in unlimited amount. This strategy thus constitutes an
attractive “green chemistry” alternative to chemical
methodologies. Moreover, significant improvement of
such a native enzyme can be achieved using new (and
presently fast developing) so-called directed evolution
techniques.
We describe here our results focused on the prepara-
tion of enantiopure pyridine-type oxiranes, i.e., 2-, 3-, and
4-pyridyloxirane 1-3, using the epoxide hydrolase from
the filamentous fungus Aspergillus niger. These com-
pounds are key-step building blocks for the synthesis of
several biologically active compounds, such as â-adren-
ergic receptor agonists or antiobesity drugs.9
3
a
(S,S)-salen(Mn) (0.019 mmol)/olefin (0.14 mmol)/NaOCl 14%
(1.5 mL)/CH2Cl2 (5 mL).10 (S,S)-salen(Mn) (0.03 mmol)/olefin
(0.61 mmol)/N-methylmorpholine N-oxide (2.98 mmol)/CH2Cl2 (5
mL).11 c (R,R)-salen(CoIII)OAc was prepared from (R,R)-salen-
(CoII) (0.085 mmol) using acetic acid (0.17 mmol) in toluene (0.5
mL). HKR: (R,R)-salen(CoIII)OAc (0.01 mmol)/epoxides (0.42
b
mmol)/H2O (0.22 mmol).6 AD-mix â (12 mg)/olefin (1.2 mL)/t-
d
BuOH-H2O (41 mL:41 mL)/23 H.12
Scr een in g for Ap p r op r ia t e E p oxid e H yd r ola se
Activity. Owing to the above-described difficulty, a
biocatalytic equivalent of J acobsen’s HKR method was
obviously of great interest. To determine the most suit-
able biocatalyst to achieve the resolution of 1-3, we
screened 14 enzyme extracts from our present collection
of EH-containing enzyme preparations, i.e., nine origi-
nating from fungal strains and five obtained from various
origins (overexpressed using a baculovirus/insect cells
procedure). Analytical scale experiments were pursued
to determine the overall reaction profile of these enzymes,
including their enantio- and regioselectivity. The results
obtained throughout this study showed that the EH from
A. niger (AnEH) was the best choice to perform the
preparative scale resolution of substrates 1-3.14 We thus
performed a more accurate determination of the various
parameters of these BHKR by using this EH prepared
from an overexpressing fungal strain we have previously
elaborated.15,16
Resu lts a n d Discu ssion
Independently of their potential interest as chirons for
the preparation of bioactive compounds, pyridyloxiranes
1-3 (Scheme 1) were chosen as a target because the
corresponding olefins should a priori be poor substrates
for heavy-metal-based catalysts, which are essentially
electrophilic reactants and therefore react more readily
with electron-rich olefins. Moreover, it is not to be
excluded that the nitrogen atom of the pyridine moiety
would either react with the reagents used for achieving
such reactions (commercial bleach or peracids, for ex-
ample) and/or be complexed with the metal-based cata-
lyst.
Deter m in a tion of th e En a n tio- a n d Regioselec-
tivity. Interestingly, the AnEH exhibited a rather high
enantioselectivity toward all three substrates, hydrolyz-
ing preferentially the (R) enantiomer of 1-3 and thus
leading to the recovery of the slow reacting (S) epoxide.
The E values (calculated on the basis of the substrate ee
and the percentage of conversion at about 40-50%) were
shown to be respectively 96, 27, and 47 at a 5 mM
substrate concentration.
Ch em ica l Asym m etr ic Oxid a tion . The above hy-
pothesis was confirmed by the results we obtained by
exploring the various methodologies presently known for
the direct preparation of enantiomerically enriched ep-
oxides or diols. Thus, using either the commercially
available (salen)Mn catalyst for direct epoxidation of
these olefins10,11 or the AD-mix â-catalyzed direct dihy-
droxylation12 of these olefins, only poor to moderate ee’s
for the resulting product were observed. Similarly, the
resolution of racemic 1-3 using the recently discovered
so-called “HKR” strategy catalyzed by (salen)Co com-
plexes13 proved to be ineffective. (Table 1). Moreover, the
low yield obtained for the dihydroxylation procedure
precluded any further use of the thus obtained diol.
The regioselectivity of the enzymatic attack on the
oxirane ring is also an important factor in such types of
reactions. Indeed, it has been previously documented that
(14) These detailed results will be described in a separate paper.
(15) This recombinant strain has been constructed by Professor J .
Visser (Wageningen Agricultural University, Section Molecular Genet-
ics of Industrial Microorganisms, Wageningen, 6703H, The Nether-
lands) by cloning the epoxide hydrolase gene from the wild strain
Aspergillus niger LCP 521 into an appropriate Aspergillus host
(unpublished results, for the corresponding patent application, see ref
16). The enzyme has been previously purified in our laboratory (see
Morisseau, C.; Archelas, A.; Guitton, C.; Faucher, D.; Furstoss, R.;
Baratti, J . C. Eur. J . Biochem. 1999, 263, 386) and further on
sequenced and cloned in an E. coli host (see Arand, M.; Hemmer, H.;
Durk, H.; Baratti, J .; Archelas, A.; Furstoss, R.; Oesch, F. Biochem. J .
1999, 344, 273). The X-ray crystal structure of this enzyme has also
been published recently (Zou, J .-Y.; Hallberg, M.; Bergfors, T.; Oesch,
F.; Arand, M.; Mowbray, S. L.; J ones, T. A. Structure (London) 2000,
8, 111). This enzyme should become available in the near future from
the Fluka catalog (year 2001-2002).
(9) (a) Mathvink, R. J .; Barritta, A. M.; Candelore, M. R.; Cascieri,
M. A.; Deng, L.; Tota, L.; Strader, C. D.; Wyvratt, M. J .; Fisher, M.
H.; Weber, A. E. Bioorg. Med. Chem. Lett. 1999, 9, 1869. (b) Devries,
K. M.; Dow, R. L.; Wright, S. W. Patent WO9821184A1, 1998. (c)
Fisher, M. H.; Naylor, E. M.; Ok, D.; Weber, A. E.; Shih, T.; Ok, H.
Patent US5561142, 1996. (d) Liang, C. D.; Walsh, G. M. Patent
US4600710, 1986.
(10) Hentemann, M. F.; Fuchs, P. L. Tetrahedron Lett. 1997, 38,
5615.
(11) Paluki, M.; McCormick, G. J .; J acobsen, E. N. Tetrahedron Lett.
1995, 36, 5457.
(12) Sharpless, B. K.; Amberg, W.; Bennani, Y. L.; Crispino, G. A.;
Hartung, J .; J eong, K.; Kwong, H.; Morikawa, K.; Wang, Z.; Xu, D.;
Zhang, X. J . Org. Chem. 1992, 57, 2768.
(13) (a) Tokunaga, M.; Larrow, J . F.; Kakiuchi, F.; J acobsen, E. N.
Science 1997, 277, 936. (b) Furrow, M. E.; Schaus, S. E.; J acobsen, E.
N. J . Org. Chem. 1998, 63, 6776.
(16) Archelas, A.; Arand, M.; Baratti, J .; Furstoss, R. French Patent
Application 9905711, 1999, and International Patent Application PCT/
FR00/01217, 2000.

540 J . Org. Chem., Vol. 66, No. 2, 2001
Genzel et al.
Ta ble 2. Regioselectivity for th e Con ver sion of 1, 2, a n d
3 w ith th e An EH
1
2
3
R(S)/â(S)a
R(R)/â(R)b
eefc
3/97
3/97
0
7/93
7/93
0
11/89
11/89
0
a
R(S)/â(S): regioselectivity coefficient respectively for the attack
at the R-carbon atom/â-carbon atom of the (S)-enantiomer given
as percentage. R(R)/â(R): regioselectivity coefficient respectively
b
for the attack at the R-carbon atom/â-carbon atom of the (R)-
enantiomer given as percentage. c eef: enantiomeric excess of the
diol at total conversion of the racemic epoxide. These values were
calculated from the equation R(R) ) R(S) - eef that we have
established previously.17
F igu r e 2. Remaining percentage of each enantiomer through-
out the conversion of rac-1 with an enzyme extract of A. niger
(strain GBCF79) at different substrate concentrations. Filled
symbols represent the (R)-enantiomer, unfilled symbols rep-
resent the (S)-enantiomer: b, 5 mM rac-1; [, 10 mM rac-1;
9, 20 mM rac-1; 2, 40 mM rac-1.
or (c) a substrate inhibition. Therefore, further experi-
ments were performed to determine the origin of this
phenomenon. This led to the following results. (a) De-
termination of the enantioselectivity at three different
enzyme-substrate ratios versus two different substrate
concentrations (20 and 40 mM) showed no modification
of the E value, ruling out the first hypothesis. (b) A
reaction was carried out at 5 mM rac-1 after prior
addition of rac-1d at a concentration of 3 or 41 mM. A
decrease of the reaction rate could be observed at 41 mM
diol concentration (factor 1.8) but again no modification
of the E value was observed. (c) The remaining percent-
age of each enantiomer throughout conversion of rac-1
was determined at different substrate concentrations.
This is represented in Figure 2. It can be seen that,
whereas we observed a normal decrease of (R)-1 con-
sumption rate upon increase of substrate concentration,
the (S)-1 enantiomer consumption was, surprisingly,
nearly identical whatever the substrate concentration.
To get a better insight into this phenomenon, the initial
rates19 of the two enantiomers were determined sepa-
rately at different substrate concentrations either by
using the racemic compound as a substrate or by using
each enantiopure enantiomer separately.20 In theory,
when using the racemic substrate, the conversion velocity
of both enantiomers should be lowered (as compared with
the intrinsic values of the pure enantiomer) upon sub-
strate concentration increase as a result of competitive
inhibition between enantiomers. Surprisingly, it can be
seen in Figure 3 that the initial rates for (S)-1, calculated
from the racemic mixture experiment on one hand and
from the experiment using enantiopure (S)-1 on the other
hand, were nearly identical. Thus, the conversion rate
of the “slow reacting” (S)-1 enantiomer was only slightly
influenced by the presence of the (fast reacting) (R)-1
enantiomer. In contrast, the rate of (R)-1 transformation
was substantially lowered in the racemic substrate as
compared to the pure enantiomer, indicating that hy-
drolysis of (R)-1 was significantly inhibited by (S)-1. This
discrepancy of kinetic behavior of the two enantiomers
F igu r e 1. Calculated E value for the conversion of 1 with an
enzyme extract of A. niger (strain GBCF79) at various sub-
strate concentration.
epoxide hydrolases may exhibit the peculiar property to
attack an epoxide moiety at either the less or the more
substituted carbon atom. Moreover, it has been observed
that such changes of regioselectivity may occur from one
enantiomer to the other, even in a racemic mixture, thus
leading to a (theoretically) 100% yield formation of nearly
enantiopure diol.17 Therefore, to get accurate information,
it is important to establish the regioselectivity of the
attack for each one of the two enantiomers of a specific
substrate. We have determined the R(S)/â(S) and R(R)/
â(R) ratios corresponding to each one of the three
substrates 1-3 by studying separately the stereochemical
outcome of the enzymatic hydrolysis of their enantiopure
(S) enantiomer as well as of the racemic substrate. The
overall results are described in Table 2. They indicate
that attack at the â-carbon atom was highly predominant
for both enantiomers, resulting in retention of configu-
ration of the corresponding diol. This is consistent with
the fact that the ee of the diol decreased to about 0% at
total conversion of the racemic substrate.
High Con cen tr a tion Exp er im en ts. To achieve ef-
ficient preparative scale experiments, optimum condi-
tions had to be found, including determination of the
highest possible substrate concentration. Quite unexpect-
edly, we observed that, for all three substrates 1-3, the
E value decreased when substrate concentration in-
creased, a quite surprising result as compared to one of
our previous observations where, in the case of p-
bromostyrene oxide,18 the E value was shown to increase
upon substrate concentration increase (Figure 1).
A priori, this could be due to either (a) the modification
of the enzyme/substrate ratio, (b) a product inhibition,
(19) The initial rates were determined from the slope of the derived
second-order polynomal regression of the extent of conversion (0-30%).
This could be done for each enantiomer, even from the racemic starting
compound, as the extent of conversion was determined by using an
internal standard together with chiral GC analysis.
(17) Moussou, P.; Archelas, A.; Baratti, J .; Furstoss, R. Tetrahe-
dron: Asymmetry 1998, 9, 1539.
(18) Cleij, M.; Archelas, A.; Furstoss, R. Tetrahedron: Asymmetry
1998, 9, 1839.
(20) Enantiopure (R)-1 was obtained by using other epoxide hydro-
lases; these results will be published elsewhere.

Microbiological Transformations
J . Org. Chem., Vol. 66, No. 2, 2001 541
an ee of 56%. The calculated turn over number and space
time yield were similar to the one estimated for 1.
The absolute configuration of 1 and of the correspond-
ing diol 1d were assigned on the basis of their optical
rotation sign by comparison with previously described
data.22,23 Those of epoxides 2 and 3 were established by
chemical correlation with the corresponding pyridyl-1-
ethanols 2e and 3e, for which the optical rotation/
absolute configuration assignment was known.24,25 Thus,
reduction of (optically active) epoxides 2 and 3 with
LiAlH4 led to the corresponding pyridyl-1-ethanols. Since
there was retention of absolute configuration at the
benzylic carbon atom during this transformation, the
absolute configuration of the starting epoxide could be
deduced from the obtained stereochemistry of these
products. Similarly, the absolute configuration of 2d and
3d was established by cyclization of the diol back to the
corresponding epoxide.26 We have verified that no change
of stereochemical integrity occurred during this chemical
step.
F igu r e 3. Initial rates for the conversion of each enantiomer
of 1 with an enzyme extract of A. niger (strain GBCF79) at
different substrate concentrations: 0, (S)-enantiomer of rac-
1; 9, enantiopure (S)-1; O, (R)-enantiomer of rac-1; b, enan-
tiopure (R)-1.
upon substrate concentration increase clearly explains
the decrease in E value we have observed. Further
studies, i.e., determination of the corresponding kinetic
constants using highly purified enzyme, would be neces-
sary to understand the intrinsic reason of this behavior
(it must be noted that the velocity increase observed in
the substrate range we have used was nearly linear,
indicating a much higher KM value for both enantiomers).
This will be achieved in the near future and described
elsewhere.
Con clu sion
The biocatalyzed hydrolytic kinetic resolution of 2-, 3-,
and 4-pyridyloxiranes 1, 2, and 3 was performed using
an enzyme extract of the fungus Aspergillus niger GBCF
79. Up to now none of these products was directly
accessible using presently available conventional chemi-
cal procedures, including the most effective heavy-metal-
catalyzed approaches known to date. In the course of this
study, we have observed that, for either of these three
substrates, an increase of substrate concentration led to
an unexpected decrease of the corresponding E value.
This was shown to be due to the modification of the
kinetic profile of the fast reacting enantiomer upon
increase of substrate concentration. We have shown that
this was due to an inhibition by the slow reacting (S)-1
enantiomer, resulting in a (proportional) decrease of the
hydrolysis rate of the fast reacting (R)-1 enantiomer.
Nevertheless, the preparative scale synthesis of each of
these three targets could be performed at about 10 g/L
substrate concentration, thus allowing one to obtain these
compounds in nearly enantiopure form (ee > 98%).
Moreover, we have shown that this biocatalyzed hydro-
lytic kinetic resolution could be performed using plain
water instead of buffer solution, an important step in
favor of the “green chemistry” aspect. Thus, this very
simple procedure, which only uses water as a reactant
and solvent, appears to be nowadays the best and most
direct way to prepare these very valuable chiral synthons
in enantiopure form.
P r epar ative Scale Exper im en ts. Owing to the above-
described results, the best substrate concentration/E
value compromise was chosen to perform a gram scale
preparation of each of the three epoxides 1-3. Thus, as
a typical example, 2 g of rac-1 were submitted to
biohydrolysis using 0.1 g of a partially purified enzyme
extract (specific activity 23 U/mg protein).21 The reaction
was carried at a 10 g/L (82 mM) substrate concentration
and at 4 °C, which we have found previously to be the
best temperature choice for this enzyme.4a This afforded
a 27% (purified) yield of (S)-1, which showed an ee of 99%,
and a 55% (purified) yield of (R)-1d with an ee of 36%.
The calculated turnover number (TON) could thus be
estimated to be about 4500 mol substrate/mol enzyme/h
and the space time yield to be about 8 g/L/day. Interest-
ingly enough, we have recently observed that this biotrans-
formation could also be achieved using plain (deionized)
water instead of a buffer solution. This obviously consti-
tutes an interesting improvement of the procedure, as
far as “green chemistry” is concerned, thereby avoiding
addition of buffering salts and thus simplifying the
downstream processing for a possible industrial applica-
tion. The catalytic process thus involved (a) mixing the
substrate and the enzyme in plain water (b) stirring at
25 °C for 1.5 h, and (c) extraction and bulb to bulb
distillation of the unreacted (S) epoxide. The formed diol
could be obtained directly by lyophilization of the aqueous
solution and was purified by simple flash chromatogra-
phy. Similarly, biohydrolysis of rac-2 at a 9.7 g/L sub-
strate concentration led to (S)-2 (96% ee, 6% yield) and
(R)-2d (7% ee, 54% yield). For this substrate, the overall
yield of epoxide was unfortunately very low as a result
of the poor enantioselectivity of the epoxide hydrolase
toward this substrate (E ) 3). More interesting results
were, however, obtained with rac-3. Thus 2.2 g of this
substrate was resolved within about 7 h at a 10 g/L
substrate concentration and afforded (S)-3 (28% yield),
which showed an ee of 98%, and (R)-3d (34% yield) with
Exp er im en ta l Section
Gen er a l. NMR spectra (1H and 13C) were recorded in CDCl3
at 250 and 100 MHz, respectively. Chemical shifts are reported
in δ from TMS as internal standard. For gas chromatography
analysis, the chiral column Chiralsil Dex CB (Chrompack) was
used.
(22) Imuta, M.; Kawai, K.; Ziffer, H. J . Org. Chem. 1980, 45, 3352.
(23) Chelucci, G.; Cabras, M. A.; Saba, A. Tetrahedron: Asymmetry
1994, 5, 1973.
(24) (a) Imuta, M.; Ziffer, H. J . Org. Chem. 1978, 43, 3530. (b) Ziffer,
H.; Kawai, K.; Kasai, M.; Imuta, M.; Froussios, C. J . Org. Chem. 1983,
48, 3017. (c) Seemayer, R.; Schneider, M. P. Tetrahedron: Asymmetry
1992, 3, 827.
(25) It should be noted that due to the rule of priority of substituents
by Cahn, Ingold and Prelog rule the (S)-epoxides led to the correspond-
ing (R)-pyridyl-1-ethanols.
(26) Golding, B. T.; Hall, D. R.; Sarkrikar, S. J . Chem. Soc., Perkin
Trans. 1 1973, 1214.
(21) The enzymatic activity was calculated using p-chlorostyrene
oxide as a reference substrate. It is expressed in µmol/min/mg of
protein. We estimate the specific activity of pure AnEH to be about 90
U/mg pure enzyme.

542 J . Org. Chem., Vol. 66, No. 2, 2001
Genzel et al.
Syn th esis of 2-P yr id yloxir a n e 1. This was synthesized
following the procedure described by Hanzlik et al.27 From 10.3
mL of 2-vinylpyridine (Fluka), 4.9 g of 1 was obtained (41%
yield). Purification was by flash chromatography (hexane/ethyl
acetate 4/6), followed by removal of solvent by bulb-to-bulb
by addition of 500 µL of water and about 200 mg of solid
Na2CO3 and extracted twice with 300 µL of ethyl acetate. The
combined organic layers were evaporated and dried under
vacuum. To this, 50 µL of KOH in MeOH (1 N) was added
and after 30 min at room temperature 300 µL of water was
added. The epoxide was extracted with 300 µL of CHCl3 and
analyzed by GC.
1
distillation (40 °C; 4 × 10-2 mbar). H NMR: δ 2.78 (dd, 1H,
H2c, J gem ) 5 Hz, J 1-2c ) 2.5 Hz); 2.99 (dd, 1H, H2t, J gem ) 5
Hz, J 1-2t ) 3.5 Hz); 3.84 (dd, 1H, H1, J 1-2c ) 2.5 Hz, J 1-2t
)
Gen er a l P r oced u r e for th e Bioh yd r olysis of 1-3.
P r ep a r a tive Sca le Con ver sion . An amount of 2 g (16.5
mmol) of rac-epoxide 1-3, dissolved in 200 mL of 0.1 M Na
phosphate buffer at pH 8 (cooled at 4 °C), was added directly
to 100 mg of the enzyme extract of A. niger (4 °C, 250 rpm).
The reaction was stopped by adding methanol to the reaction
mixture (up to a 1/1 methanol/buffer proportion) when the ee
of the remaining substrate reached a value of about 96%. After
extraction with CHCl3 (3 × 300 mL) the aqueous phase was
saturated with NaCl and extracted twice with ethyl acetate
(800 mL). The aqueous phase was then further extracted by
continuous extraction with CH2Cl2 (500 mL) for 4 days. The
epoxide and the diol were purified by flash chromatography
(only in the first extraction with CHCl3 was the epoxide
present; the diol was present in all organic phases). Flash
chromatography of the epoxide was performed with hexane/
ethyl acetate 2/8 and for the diol with ethyl acetate/methanol
9/1. The fractions containing the epoxide and the diol, respec-
tively, were pooled and dried over Na2SO4 and the solvent was
removed by stripping. The epoxide was purified via bulb-to-
bulb distillation without heating (4 × 10-2 mbar).
3.75 Hz); 7.02-7.08 (m, 2H); 7.50 (t, 1H); 8.39 (d, 1H, J ) 5
Hz).28 13C NMR: δ 50.38; 52.81; 119.69; 123.12; 136.82; 149.39;
157.18. GC: 110 °C, 1 kg/cm2 helium; (R) ) 7.8 min; (S) ) 8.6
min.
Syn th esis of 3-P yr id yloxir a n e 2. As described by Gian-
nini et al.29 From 3.29 g of 3-pyridylaldehyde (Fluka), 2.1 g of
2 was obtained (56% yield). DMSO was removed by washing
the organic phase four times with distilled water. Purification
was by flash chromatography (hexane/ethyl acetate 4/6),
followed by removal of solvent by bulb-to-bulb distillation
without heating (4 × 10-2 mbar). 1H NMR: δ 2.83 (dd, 1H,
H2, J gem ) 5.3 Hz, J 1-2 ) 2.51 Hz); 3.20 (dd, 1H, H2, J gem
)
5.26 Hz, J 1-2 ) 4.1 Hz); 3.9 (dd, 1H, H1, J 1-2 ) 3.87 Hz, J 1-2
) 2.7 Hz); 7.25-7.31 (m, 1H); 7.52-7.57 (m, 1H); 8.55-8.59
(m, 2H).9c,28 13C NMR: δ 50.38; 51.1; 123.46; 132.73; 147.88;
149.62. GC: 110 °C, 1 kg/cm2 helium; R ) 12.2 min; S ) 12.8
min.
Syn th esis of 4-P yr id yloxir a n e 3. As described by Corey
and Chaykovsky.30 From 23.3 g of 4-pyridylaldehyde (Fluka),
6.7 g of 3 was obtained (25% yield). DMSO was removed by
washing the organic phase four times with distilled water.
Purification was by flash chromatography (hexane/ethyl ace-
tate 4/6), followed by removal of solvent by bulb-to-bulb
distillation without heating (4 × 10-2 mbar). 1H NMR: δ 2.77
(dd, 1H, H2, J gem ) 5.6 Hz, J 1-2 ) 2.5 Hz); 3.20 (dd, 1H, H2,
J gem ) 5.6 Hz, J 1-2 ) 4.2 Hz); 3.85 (dd, 1H, H1, J 1-2 ) 4 Hz,
J 1-2 ) 2.5 Hz); 7.21 (dd, 2H, J ) 4.5 Hz, J ) 1.5 Hz); 8.58 (d,
2H, J ) 5.7 Hz).28,31 13C NMR: δ 50.95; 51.33; 120.36; 146.96;
149.93. GC: 110 °C, 1 kg/cm2 helium; R ) 12.9 min; S ) 13.8
min.
Ra cem ic 2-P yr id yld iol 1d . As described by Hanzlik et al.27
A total of 1.6 g of 1 was hydrolyzed under acidic conditions
(H2SO4) for 64 h to give 1.1 g of 1d (yield 60%). 1H NMR
(acetone-d6): δ 3.64 (dd, 1H, H2, J gem ) 10.8 Hz, J 1-2 ) 5.6
Hz); 3.82 (dd, 1H, H2, J gem ) 11.3 Hz, J 1-2 ) 6.1 Hz); 3.97-
3.94 (m, 1H, HOH); 4.66 (d, 1H, HOH, J ) 4.95); 4.77-4.73 (m,
1H, H1); 7.28-7.24 (m, 1H); 7.56 (d, 1H, J ) 7.8 Hz); 7.82-
7.75 (m, 1H); 8.51 (d, 1H, J ) 4.5 Hz).28,31 13C NMR: δ 67.81;
75.06; 121.59; 123.07; 137.32; 149.12;162.57.
Gen er a l P r oced u r e for th e Bioh yd r olysis of 1-3.
An a lytica l Sca le Con ver sion . A substrate solution (5-80
mM) in a 0.1 M pH 8 Na phosphate buffer was used and
diethylendiethylglycolether (Aldrich) was added as an internal
standard. Before addition of the enzymatic extract, a sample
was taken for the 100%-epoxide value. The reaction was
started with addition of the substrate solution to the enzyme
extract. The kinetic resolution (28 °C, 750 rpm) was followed
by taking aliquots from the reaction medium (1 vol sample
was added to 1 vol of methanol and 2 vol of CHCl3). The organic
phase was analyzed by GC to calculate the degree of conversion
and the ee of the remaining epoxide (GC, 110 °C; internal
standard ) 4.6 min). The diol was extracted (after total
extraction of the epoxide with CHCl3) using ethyl acetate after
saturation of the aqueous phase with NaCl. To determine its
ee, the diol was cyclized back to the epoxide following the
procedure previously described by Golding et al.26 (without
change of its absolute configuration) as follows. To the dried
diol, 100 µL of HBr in glacial acetic acid (33%) were added.
After 1.25 h at room temperature, the mixture was neutralized
For the conversion of 1 a different workup was used. The
unreacted 1 was extracted with CH2Cl2 and the aqueous phase
was filtered (0.45 µm) to remove the enzyme extract. The
aqueous phase was then lyophilized, and 1d was directly
extracted from the resulting dry diol salt with ethyl acetate
and purified by flash chromatography.
Using the above-described protocol, 2 g of rac-1 afforded
after about 8 h 0.55 g of (S)-1 (27% purified yield, 99% ee)
and 1.26 g (55% purified yield, 36% ee) of (R)-1d . Similarly,
1.93 g of rac-2 at a 9.7 g/L substrate concentration led to 0.12
g of (S)-2 (6% yield, 96% ee) and 1.2 g of (R)-2d (54% yield,
7% ee). Also, 0.61 g of (S)-3 (28% yield, ee 98%) and 0.85 g of
(R)-3d (34% yield, ee 56%.) were obtained from rac-3 (2 g).
Bioh yd r olysis of r a c-1 Usin g P la in Wa ter In stea d of
a Bu ffer Solu tion . A semipreparative scale experiment was
conducted using a 5 mM concentration of racemic 1 (0.61 g/L).
Thus, 610 mg of 1 in 1 L of plain (deionized) water were treated
with 990 mg of a crude enzymatic extract of EH. After 1.5 h
at 25 °C, the medium was extracted with chloroform. Normal
workup of this organic phase led to 260 mg (43% yield) of (S)-
2, of which the ee was shown to be higher than 99%.
Lyophilization of the aqueous phase afforded directly 300 mg
of (R)-3 diol (43% yield) with 62% ee, purified using flash
chromatography.
Op tica l r ota tion s and NMR data of the isolated products
after biohydrolysis were as follows: (S)-1 [R]19 +14 (c 0.56,
D
CHCl3), ee 99%; lit.22 (R)-1 [R]25 -15 (c 0.41, CHCl3); (R)-1d
D
[R]19 -29 (c 1.64, EtOH), ee 36%; lit.32 (S)-1d [R]25 +80.6 (c
D
D
1.6, EtOH); (S)-2 [R]25 +18.3 (c 0.91, CHCl3), ee 96%; (R)-2d
D
[R]25 -3.5 (c 0.57, MeOH), ee 7.3%; (R)-2e [R]25 +44 (c 0.94,
D
D
CHCl3); lit.24c (R)-2e [R]25 +52.4 (c 1.4, CHCl3); (S)-3 [R]25
D
D
+28.5 (c 1.16, CHCl3), ee 98%; (R)-3d [R]20D -11 (c 0.53, EtOH),
ee 56%; (R)-3e [R]25 +38 (c 0.88, MeOH); lit.24c (R)-3e [R]25
D
D
+42.5 (c 1.04, MeOH).
The 1H NMR data for epoxides 1-3 and diol 1d were
identical to those obtained with the corresponding racemic
1
substrates previously prepared. 3-P yr id yld iol 2d : H NMR
(acetone-d6): δ 3.69-3.58 (m, 2H, H2); 4.10 (s, 1H, HOH); 4.66
(s, 1H, HOH); 4.8-4.78 (m, 1H, H1); 7.34 (dd, 1H, J ) 7.86 Hz,
J
) 4.75 Hz); 7.81-7.77(m, 1H); 8.60-8.44 (m, 2H). 13C
NMR: δ 68.45; 73.16; 123.91; 134.63; 139.00; 149.07;149.28.
1
(27) Hanzlik, R. P.; Edelman, M.; Michaely, W. J .; Scott, G. J . Am.
Chem. Soc. 1976, 98, 1952.
(28) Kloc, K.; Kubicz, E.; Mlochowski, J . Heterocycles 1984, 22, 2517.
(29) Giannini, M.; Bonacchi, G.; Fedi, M.; Sesto, F. German patent
DE 3425477 A1, 1985.
(30) Corey, E. J .; Chaykovsky, M. J . Am. Chem. Soc. 1965, 87, 1353.
(31) Carrol, F. I. Org. Prep. Proced. Int. 1971, 3, 121.
4-P yr id yld iol 3d : H NMR (250 MHz, acetone-d6): δ 3.67-
3.58 (m, 2H, H2); 4.03 (s, 1H, HOH); 4.66 (s, 1H, HOH); 4.74 (s,
1H, H1); 7.40-7.37 (m, 2H); 8.52-8.49 (m, 2H).33 13C NMR: δ
1
68.22; 73.96; 122.27; 150.13; 152.59. The H NMR data for the
corresponding pyridylethanol were identical to those described
previously in the literature.23

Microbiological Transformations
J . Org. Chem., Vol. 66, No. 2, 2001 543
Ack n ow led gm en t. We are extremely grateful to the
following colleagues, whose previous work helped us to
achieve this study: Mrs. C. Guitton and Mr. D. Faucher
(Rhoˆne-Poulenc Company, Lyon, France) for having
constructed the appropriate probe for the Aspergillus
niger epoxide hydrolase, allowing further molecular
biology work; Prof. M. Arand (Institute of Toxicology,
Mainz, Germany) for having achieved the cloning of the
enzyme; Prof. J . Visser (University of Wageningen,
Holland) for having performed the overexpression of the
wild type A. niger enzyme into an appropriate Asper-
gillus host. Financial support of this work by DSM, as
well as a fellowship to one of us (Y.G.) is greatly
acknowledged.
(32) Chelucci, G.; Cabras, M. A.; Saba, A. Tetrahedron: Asymmetry
1994, 5, 1973.
(33) Aaron, H. S.; Miller, J . I. Bull. Chem. Soc. J pn. 1974, 47(5),
1305.
J O001406X