Microbiological transformations 43. Epoxide hydrolases as tools for the synthesis of enantiopure α-methylstyrene oxides: A new and efficient synthesis of (S)-ibuprofen

Article abstract of DOI:10.1021/jo982101+

Biohydrolysis of various α-methylstyrene oxide derivatives, differently substituted at the aromatic ring, was investigated using 10 epoxide hydrolases from different origins. Our results indicate that the enantioselectivity of these biohydrolyses strongly depends on the nature of the enzyme and of the substituent. Using some of these enzymes, this approach allows to prepare these epoxides in high optical purity. The potentiality to perform efficient preparative-scale resolution using such a biocatalyst was illustrated by the four-step synthesis of (S)-ibuprofen, a nonsteroidal antiinflammatory drug and household pain killer, one of the top-ten drugs sold worldwide. Using a combined chemoenzymatic strategy, we were thus able to set up a four-step enantioconvergent procedure allowing for the synthesis of this compound in optically pure form and with a 47% overall yield, including the resolution process, due to a possible recycling of the formed diol via chemical racemisation.

Full text of DOI:10.1021/jo982101+

J . Org. Chem. 1999, 64, 5029-5035
5029
Micr obiologica l Tr a n sfor m a tion s 43. Ep oxid e Hyd r ola ses a s Tools
for th e Syn th esis of En a n tiop u r e r-Meth ylstyr en e Oxid es: A New
a n d Efficien t Syn th esis of (S)-Ibu p r ofen
M. Cleij, A. Archelas, and R. Furstoss*
Groupe Biocatalyse et Chimie Fine, ESA 6111 associe´e au CNRS, Faculte´ des Sciences de Luminy,
Universite´ de la Me´diterrane´e, Case 901, 163 avenue de Luminy, 13288 Marseille Cedex 9, France
Received October 19, 1998
Biohydrolysis of various R-methylstyrene oxide derivatives, differently substituted at the aromatic
ring, was investigated using 10 epoxide hydrolases from different origins. Our results indicate that
the enantioselectivity of these biohydrolyses strongly depends on the nature of the enzyme and of
the substituent. Using some of these enzymes, this approach allows to prepare these epoxides in
high optical purity. The potentiality to perform efficient preparative-scale resolution using such a
biocatalyst was illustrated by the four-step synthesis of (S)-ibuprofen, a nonsteroidal antiinflam-
matory drug and household pain killer, one of the top-ten drugs sold worldwide. Using a combined
chemoenzymatic strategy, we were thus able to set up a four-step enantioconvergent procedure
allowing for the synthesis of this compound in optically pure form and with a 47% overall yield,
including the resolution process, due to a possible recycling of the formed diol via chemical
racemisation.
In tr od u ction
likely that a suitable enzyme can be found for achieving
the enantioselective hydrolysis of either type of epoxide,
and the screening of various EH for synthetically inter-
esting chiral epoxides is therefore of paramount impor-
tance. Furthermore, it appears that, whereas the large
preparative scale use of some of these enzymes, for
instance, the once from mammalian origin, is severely
hampered due to their low availability, several such
biocatalysts are now easily available in large scale from
microorganisms.2
In our laboratory, we have previously detected different
EHs from fungal origin, and we have developed various
studies aimed at exploring the potentialities of these
enzymes on different types of substrates.6 In this paper,
we report the enantioselective hydrolysis of seven dif-
ferently substituted R-methylstyrene oxide derivatives
with 10 different soluble EHs. The choice of these
substrates stemmed from the fact that they constitute
interesting chiral building blocks allowing for the syn-
thesis of various R-arylpropionic acid derivatives, a major
class of nonsteroidal antiinflammatory drugs and house-
hold painkillers. Among these, ibuprofen (2-[4-(2-meth-
ylpropyl)phenyl]propanoic acid) and naproxen (2-(4-
methoxynaphthyl)propanoic acid) are probably the best
known and are sold worldwide in large quantities. These
molecules are chiral, and it was shown that it is the S
enantiomer that is responsible for the desired therapeutic
effect.7 Despite this fact, ibuprofen is currently admin-
istered as the racemate, although it has been demon-
strated that its R enantiomer accumulates in fatty tissue
as a glycerol ester, whose long-term effects are not
A hot topic in synthetic organic chemistry is the
preparation of optically pure epoxides as well as of their
corresponding vicinal diols.1 These compounds are highly
versatile chiral synthons that are often used to achieve
the synthesis of biologically active molecules. One of the
emerging approaches to obtain such optically pure build-
ing blocks is the enantioselective hydrolysis of a racemic
epoxide catalyzed by a “new” type of hydrolytic enzymes,
i.e., epoxide hydrolases (EHs). This reaction will afford
the corresponding vicinal diol together with the unre-
acted epoxide enantiomer, both of these products being,
in certain cases, obtained in good to excellent enantio-
meric purity.2 Only recently, it became clear that these
biocatalysts are ubiquitous in nature, since they have
been detected in organisms as diverse as mammals,3
plants,4 or microorganisms.5 As a consequence, it is now
* To whom correspondence should be addressed. Phone: +33 04 91
82 91 55. Fax: +33 04 91 82 91 45. E-mail: furstoss@luminy.univ-
mrs.fr.
(1) (a) Brandes, B. D.; J acobsen, E. N. Tetrahedron: Asymmetry
1997, 8, 3927. (b) Katsuki, T. Coord. Chem. Rev. 1995, 140, 189. (c)
J acobsen, E. S.; Kakiuchi, F.; Fonsler, R. G.; Larrow, J . F.; Tokunaga,
M. Tetrahedron Lett. 1997, 38, 773. (d) Kolb, H. C.; van Nieuwenhuize,
M. S.; Sharpless, K. B. Chem. Rev. 1994, 94, 2483.
(2) (a) Faber, K.; Mischitz, M.; Kroutil W. Acta Chem. Scand. 1996,
50, 249. (b) Archelas, A.; Furstoss, R. Annu. Rev. Microbiol. 1997, 51,
491. (c) Archer, I. V. J . Tetrahedron 1997, 53, 15617. (d) Archelas, A.;
Furstoss, R. Trends Biotechnol. 1998, 16, 108.
(3) (a) Hammock, B. D.; Grant, D. F.; Storms, D. H. In Comprehen-
sive Toxicology; Sipes, I., McQueen, C., Gandolfi, A., Eds.; Pergamon:
Oxford, 1996; Chapter 18.3. (b) Arand, M.; Mu¨ller, F.; Mecky, A.; Hinz,
W.; Urban, P.; Pompon, D.; Kellner, R.; Oesch, F. Biochem. J . 1999,
337, 37-43.
(4) Ble´e, E.; Schuber, F. Eur. J . Biochem. 1995, 230, 229.
(5) (a) Pedragosa-Moreau, S.; Archelas, A.; Furstoss, R. J . Org.
Chem. 1993, 58, 5533. (b) Osprian, I.; Kroutil, W.; Mischitz, M.; Faber,
K. Tetrahedron: Asymmetry 1997, 8, 65. (c) Weijers, C. A. G. M.; Botes,
A. L.; van Dyk, M. S.; de Bont, J . A. M. Tetrahedron: Asymmetry 1998,
9, 467. (d) Rink, R.; J anssen, D. B. Biochemistry 1998, 37, 18119-
18127. (e) Barbirato, F.; Verdoes, J . C.; de Bont, J . A. M.; van der Werf,
M. J . FEBS Lett. 1998, 438, 293-296. (f) Misawa, E.; Chan Kwo Chion,
K. C.; Archer, I. V.; Woodland, M. P.; Zhou, N. Y.; Carter, S. F.;
Widdowson, D. A.; Leak, D. J . Eur. J . Biochem. 1998, 253, 173-183.
(6) (a) Chen, X. J .; Archelas, A.; Furstoss, R. J . Org. Chem. 1993,
58, 5528. (b) Pedragosa-Moreau, S.; Archelas, A.; Furstoss, R. Tetra-
hedron 1996, 52, 4593. (c) Pedragosa-Moreau, S.; Archelas, A.; Moris-
seau, C.; Zylber, J .; Baratti, J .; Furstoss, R. J . Org. Chem. 1996, 61,
7402. (d) Moussou, P.; Archelas, A.; Furstoss, R. Tetrahedron 1998,
54, 1563.
(7) (a) Wetchter, W. J . J . Clin. Pharmacol. 1994, 34, 1036. (b)
Wetcher, W. J .; Bigornia, A. E.; Murray, E. D., J r.; Levine, B. H.;
Young, J . W. Chirality 1993, 5, 492.
10.1021/jo982101+ CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/24/1999

5030 J . Org. Chem., Vol. 64, No. 14, 1999
Cleij et al.
and rat12 origin and were obtained as overexpressed
proteins from baculovirus infected insect cells. They were
obtained from these cells as freeze-dried extracts with a
relatively high enzyme content (approximately 5%).
In the context of this work, our first aim was to develop
a “matrix-type” approach allowing us to explore the
various possibilities offered by the combination of each
of the substrates with each of the specific enzymes. Our
preliminary results indicated that all the substrates were
hydrolyzed by the various enzymes, although with dif-
ferent rates. For each extract, the enzymatic activity was
determined at three different pH values, i.e., pH 7.2, 8.0,
and 9.0, relative to p-chlorostyrene oxide chosen as a
reference substrate. The results are presented in Table
1. Generally, the enzymatic activity was strongly pH
dependent. The highest values were always observed at
pH 7.2, and increasing to pH 9 often led to a noticeable
decrease of activity. The only exception was for rat-EH,
whose activity stayed almost constant whatever the pH.
Table 1 also indicates the rates of spontaneous chemical
hydrolysis as a function of the pH. At pH 7.2, it was
relatively important but slowed going to pH 9. Thus, the
best compromise to achieve our experiments appeared
to be at pH 8.0. Indeed, at this pH, the loss of enzymatic
activity was still acceptable and the spontaneous hy-
drolysis stayed moderate. Owing to these results, more
accurate studies using the extracts from C. globosum, M.
isabellina, C. echinulata, and B. bassiana were not
pursued due to low enzymatic activity.
The other extracts were, however, more promising and
were therefore studied in more details, as indicated in
Table 2. For each substrate/enzyme extract combination,
five parameters were determined, i.e., (a) the specific
activity; (b) the value of the observed enantioselectivity
ratio E13 (it is to be stressed that, because of the
occurrence of chemical hydrolysis, the “real” E value of
each enzyme must be somewhat higher); (c) the config-
uration of the recovered (slow reacting) epoxide enanti-
omer; (d) the enzymatic activity observed with this
specific substrate at a 8 mM concentration (except for 7,
where it was measured at 4 mM because of its low
solubility) (these activity values have been corrected for
nonenantioselective spontaneous chemical hydrolysis,
which occurred at a different extend for each substrate,
as indicated in Table 2); and (e) the amount of enzymatic
activity (expressed in units and calculated using 3 as
reference substrate) used to perform this specific biohy-
drolysis. In certain cases, chemical hydrolysis being sign-
ificant relative to the enzymatic hydrolysis, the amount
of biocatalyst used to perform the reaction was enhanced
in order to obtain a reasonable enzymatic/chemical
hydrolysis ratio. This ratio is indicated in footnote f.
As a first result, it appears that in all cases, the diol
with an absolute configuration opposite to the one of the
residual epoxide was obtained. This indicates that, as
expected, the nucleophilic attack always preferentially
(if not exclusively) occurred at the C-â carbon atom of
the oxirane ring. The structures of the formed diols have
been ascertained either by comparison with literature
Sch em e 1
known.8 Therefore, it is desirable to develop synthetic
methods allowing for the preparation of these drugs in
an enantiopure form. We illustrate in this work the
synthetic potentialities of EH enzymes by describing a
four-step synthesis of ibuprofen, using a combined
chemoenzymatic strategy. This allowed us to prepare this
drug in enantiopure form from racemic 4-isobutyl-R-
methylstyrene oxide 5 with an overall yield as high as
47%, including the resolution process.
Resu lts a n d Discu ssion
The various para-substituted R-methylstyrene oxide
derivatives that we have studied are presented in Scheme
1. Their structures have been ascertained on the base of
previously described syntheses described in the literature
(see the Experimental Section). Theoretically, a nucleo-
phile (or an enzyme) can attack an epoxide at either of
the two carbon atoms of the oxirane ring, and it is obvious
that this regioselectivity depends on the nature of the
nucleophile, the structure of the epoxide, the reaction
conditions, and for biocatalytic reactions, the nature of
the enzyme. In this last case, it is to be stressed that both
the regioselectivity and the kinetic parameters can be
different from one enantiomer to the other, thus leading
to quite puzzling results.9 For instance, owing to a
difference of regioselectivity for each enantiomer, hy-
drolysis of a racemic substrate can lead to one single, i.e.
enantiopure, enantiomer of the diol product, following a
so-called “enantioconvergent” process. However, in the
case of the substrates studied in this work, the R-carbon
atom is a severely sterically hindered tertiary carbon
atom, and consequently, most nucleophiles should ex-
clusively attack at the â-carbon atom.
In the course of this study, we used 10 different
enzymatic extracts known to possess epoxide hydrolase
activity, i.e., seven fungal strains and three overex-
pressed enzymes from various origins. The seven fungal
EHs were obtained as a freeze-dried crude enzymatic
extract from, respectively, the following: Aspergillus
niger LCP 521; Aspergillus terreus CBS 116-46; Chaeto-
mium globosum LCP 679; Syncephalastrum racemosum
MUCL 28766; Mortierella isabellina ATCC 42613; Cun-
ninghamella elegans LCP 1543 and Beauveria bassiana
ATCC 7159 (previously named Beauveria sulfurescens).
The other three soluble EHs are from human,10 potato,11
(8) (a) Williams, K.; Day, R.; Romualda, K.; Duffield, A. Biochem.
Pharmacol. 1986, 35, 3403. (b) Williams, K. M. Problems and Wonders
of Chiral Molecules; Simonyi, M., Ed; Akademiai: Kiado, Budapest,
1990; p 181.
(9) Moussou, P.; Archelas, A.; Baratti, J .; Furstoss, R. Tetrahedron:
Asymmetry 1998, 9, 1539.
(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.
(12) Lacourciere, G. M.; Vakharia, V. N.; Tan, C. P.; Morris, D. I.;
Edwards, G. H.; Moos, M.; Armstrong, R. N. Biochemistry 1993, 32,
2610.
(10) Beetham, J . K.; Tian, T.; Hammock, B. D. Arch. Biochem.
Biophys. 1993, 305, 197.
(13) Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J . J . Am. Chem.
Soc. 1982, 104, 7294.

Synthesis of Enantiopure R-Methylstyrene Oxides
J . Org. Chem., Vol. 64, No. 14, 1999 5031
Ta ble 1. En zym e Extr a ct Activitya ,b a n d Sp on ta n eou s Hyd r olysisc of r a c-3 a s a F u n ction of p H
pH
A. n.
A. t.
C. g.
S. r.
M. i.
C. e.
B. b.
h-EH
p-EH
r-EH
chem hyd
7.2
8.0
9.0
23.3
19.2
7.9
5.8
5.9
4.4
0.9
0.7
0.4
2.9
2.8
1.4
1.95
1.8
1.2
0.4
0.2
0.06
0.2
0.1
0.1
19.8
15.8
9.3
3.15
3.2
2.6
4.9
4.8
4.9
3.7
1.2
0.6
a
b
Experimental conditions: cextract 5-20 mg/mL; csubstrate 8.0 mM; 0.2 M Tris/HCl; t0 27 °C; 1% DMF. Enzymatic activity in nmol/
min/mg per mL. c Chemical hydrolysis in nmol/min per mL.
Ta ble 2. Bioca ta lyzed Hyd r olysis of Ep oxid es 1-7 Usin g Va r iou s Ep oxid e Hyd r ola ses
enzymatic
extract
specific
activitya
H 1
F 2
Cl 3
Br 4
iBu 5
CN 6
NO2 7
A. n.
A. t.
19.2
5.9
3b (S)c
9.6d
10 (S)
19.1
576 (39)
3 (S)
15 (S)
19.2
96 (80)
1.3 (R)
5.9
20 (S)
14.2
192 (183)
3 (R)
20 (S)
0.6
384 (25)
8 (R)
0.7
26 (S)
19.7
96 (462)
1.4 (R)
7.4
39 (S)
24.8
96 (1859)
1.1 (S)
11.7
192e (9)f
2 (S)
7.5
12.7
4.1
118 (14)
3 (R)
0.4
177 (26)
3 (R)
0.6
119 (99)
3 (R)
2.8
118 (106)
2 (R)
2.4
177 (42)
2 (S)
0.12
118 (693)
2 (S)
1.3
118 (3499)
2 (S)
1
S. r.
2.8
70 (1)
3 (S)
1.1
54 (3)
8 (R)
22.4
158 (21)
3 (S)
2.7
64 (5)
6 (R)
10.1
70 (1)
2 (S)
1.5
54 (3)
5 (R)
32.3
158 (22)
2 (S)
56 (46)
3 (S)
1.8
36 (30)
1.5 (S)
15.8
158 (132)
2 (S)
42 (46)
3.4(S)
1.6
36 (42)
2.5 (S)
9.9
158 (128)
2 (S)
1.7
84 (8)
3 (S)
0.3
54 (22)
1.1 (R)
1.5
474 (98)
3 (R)
0.3
56 (124)
5 (R)
1.4
54 (199)
4 (R)
4.9
158 (230)
7 (R)
1.5
32 (68)
10 (R)
4.2
56 (299)
4 (R)
2.0
36 (599)
2 (R)
4.8
79 (359)
3 (R)
1.1
32 (169)
4 (R)
2.8
M. i.
h-EH
p-EH
r-EH
1.8
15.8
3.2
2.9
3.2
64 (4)
14 (R)
25.7
64 (53)
1.4 (R)
4.8
64 (48)
1.4 (S)
3.2
96 (17)
2 (R)
0.8
4.8
96 (19)
10.7
96 (35)
14.7
48 (40)
1.2
48 (42)
0.77
144 (51)
0.47
48 (199)
0.21
48 (419)
0.13
spont hydr (nmol/min)
a
Specific enzymatic activity in nmol/min/mg, calculated using 3 as reference substrate. Experimental conditions: cextract (5-60 mg/
mL); csubstrate ) 8 mM (4 mM for 7); pH 8.0 (pH 9 for 5) (0.2 M Tris/HCl); T ) 27 °C; 1% DMF. Calculated E value. c Absolute configuration
b
of the residual epoxide. Enzymatic activity in nmol/min/mg. e Amount (in units) of enzyme used to perform this biohydrolysis. f Ratio of
d
enzymatic hydrolysis versus chemical hydrolysis.
data when available or, for the new diols 2d , 6d , and 7d ,
by elemental analysis, accurate NMR spectroscopy, and
exact mass determination as indicated in the Experi-
mental Section.
oxirane opening switched progressively from the benzylic
CR carbon atom to the terminal carbon atom depending
upon the electronic character of the para substituent. The
results obtained in the present work show that (a) the
presence of a methyl group at CR does not improve the
enantioselectivity of the reaction (E ) 2-3), and (b) to
the contrary of our previous observation, no regioselec-
tivity switch was observed, the attack occurring always
preferentially at the terminal carbon atom, whatever the
nature of the para substituent. This lack of regioselec-
tivity switch is best explained by the steric hindrance at
CR introduced by the additional methyl group. Interest-
ingly in this case, the enantioselectivity, i.e., the choice
for the fastest hydrolyzed substrate, switched from the
S (epoxides 1-4) to the R (epoxides 5-7) absolute
configuration depending on the nature of the aromatic
substituent, thus leading to the formation of either the
(S)-diol (for 1d -4d ) or the (R)-diol (for 5d -7d ) respec-
tively.
Second, it can be observed that, for all substrates, the
best results were obtained with the A. niger enzymatic
extract, which in general afforded good enantioselectivity
as well as high enzymatic activity. This extract showed
a clear preference for hydrolyzing the R enantiomer, thus
affording the recovered epoxide of S absolute configura-
tion. Interestingly, the various E values clearly depend
on the nature of the para substituent. As a rule of thumb,
it seems that increasing the size of the substituent
resulted in an appreciable increase of enantioselectivity,
e.g., from E ) 3 for 1 (R ) H) to E ) 39 for 7 (R ) NO2).
Good activities were also observed for the enzyme
extract of A. terreus. However, the enantioselectivities
were relatively low, the best E value (E ) 8) being
observed for 5. Interestingly, this biohydrolysis was
enantiocomplementary to the one performed with the A.
niger extract for some of the substrates; i.e., the residual
epoxide was of R configuration for 3-6. In the case of
the enzymatic extract obtained from S. racemosum, it is
interesting to compare the results obtained in this study
with those we have previously described using various
para-substituted styrene oxide derivatives as substrates.14
Indeed, we had observed that (a) although the enantio-
selectivity was rather low, hydrolysis led always to the
residual (S) epoxide and (b) the regioselectivity of the
Finally, the enantioselectivities observed with the
three overexpressed enzymes (human-EH, potato-EH,
and rat-EH) did vary from low to moderate. However,
for the substrates showing reasonable enantioselectivities
(E > 5), the residual enantiomer was always of the R
configuration, and in this respect, these enzymes again
appear to be enantiocomplementary to the A. niger
extract. The best enantioselectivities (E ) 10 for 6 and
E ) 14 for 2) were observed with the rat liver EH, and
for the latter substrate, it was better than for all the other
extracts. Generally, the activities with the potato EH
were moderate, but the activities with human EH and
(14) Moussou, P.; Archelas, A.; Baratti, J .; Furstoss, R. J . Org. Chem.
1998, 63, 3532.

5032 J . Org. Chem., Vol. 64, No. 14, 1999
Cleij et al.
Sch em e 2
rat liver EH were rather good. However, it should be kept
in mind that these enzymatic extracts were obtained from
genetically engineered organisms (thus affording over-
expression) and therefore contain relatively large quanti-
ties of EH (up to 5% enzyme), whereas this is not the
case for the enzyme prepared from the wild-type strain,
which has been shown to only contain about 0.1% enzyme
for A. niger, for example.15 This indicates that the specific
activity of the overexpressed enzymes toward these
substrates is quite low as compared to the A. niger EH.
Assign m en t of th e Absolu te Con figu r a tion of th e
Ep oxid es a n d Diols. The absolute configuration of
epoxide 16b and of diol 1d 16 has been determined previ-
ously. The chloro- and bromo-substituted diols 3d and
4d were dehalogenated by treatment with sodium in
ethanol,17 thus affording the corresponding diol 1d of
known absolute configuration. Each of these diols could
also be cyclized into their corresponding epoxide 3 or 4
(with retention of configuration), thus attributing to their
absolute configuration. The absolute configuration of 5
(and therefore of 5d ) was proven by its transformation
into known (S)-ibuprofene (see below). Until now, we
have not been able to unequivocally prove the absolute
configuration of the epoxides and diols 2, 6, and 7.
Therefore, these were tentatively assigned on the basis
of three convergent features: (a) Their elution order on
our chiral GC column (column I). Indeed, we observed
previously that the (R)-epoxides from various para-
substituted styrene oxide derivatives (of known absolute
configuration) always exhibited the shortest retention
time on this column.6c (b) Very probably, the A. niger
enzyme should preferentially hydrolyze the R enanti-
omer. Indeed, in our hands, this enzyme always showed
a high R enantiomer preference for 1-monosubstituted
epoxides. (c) Both of these features appear to be con-
firmed for epoxides 1, 3, and 4 of known configuration
(see above). Thus, a similar behavior of the other three
epoxides seems reasonable and interestingly fits indeed
with the GC elution order.
Syn th esis of (S)-Ibu p r ofen 9. As pointed out previ-
ously, ibuprofen is one of the top-ten drugs sold world-
wide, and it has been shown that only its S enantiomer
is endowed with the desired biological activity. Despite
this fact, it is presently sold as a racemate, and we
therefore were interested in setting up a procedure
allowing for the specific preparation of the biologically
active enantiomer. One of the possible strategies was to
achieve the enantioselective hydrolysis of racemic 4-isobu-
tyl-R-methylstyrene oxide 5, using an enzyme that would
specifically hydrolyze the undesired (R)-5, and to further
transform the thus-obtained (S)-5 enantiopure epoxide
into (S)-ibuprofen 9, using classical chemical synthesis.
Examination of Table 2 indicates that the A. niger
extract gives the best result, since the E value is quite
reasonable (E ) 20) and the residual epoxide has the
desired S configuration. To set up the most efficient
experimental conditions, small-scale trials were further
performed, using this fungal extract, by varying different
parameters. This led us to perform this biohydrolysis at
4 °C, which allowed us to both enhance enzyme stability
and decrease spontaneous hydrolysis of the substrate. At
this temperature, a 2- to 3-fold larger amount of epoxide
could be resolved as compared to 27 °C. Thus, a prepara-
tive-scale experiment was carried out using 1.5 g (7.9
mmol) of racemic 5 in 30 mL buffer solution. This
corresponds to a “concentration” of about 50 g/L (285 mM)
of substrate, which in fact formed a two-phase system
still allowing us to efficiently achieve this biohydrolysis,
as we have previously described.18 The ee of the unre-
acted epoxide was followed by chiral GC analysis, and
when it reached a value higher than 95%, the reaction
was stopped by extracting the remaining epoxide with
pentane. The formed diol was further extracted from the
reaction medium with ether and purified by column
chromatography, leading to a 55% yield of (R)-5d (0.91
g; 4.38 mmol) showing a 70% ee (Scheme 2).
Crude (S)-5 was directly converted into the corre-
sponding alcohol (S)-8 by catalytic hydrogenation. Pre-
liminary experiments aimed at exploring this reduction
indicated that the ee of the product was strongly depend-
ent upon the reaction conditions used (temperature,
solvent, and nature of the Pd catalyst). We obtained
excellent results (less than 1% loss of optical purity) when
performing this reduction at 0 °C, using N-methylfor-
mamide as a solvent and pure palladium (powder) as
catalyst (submicron/Aldrich). Using these optimized ex-
perimental conditions, the reduction was shown to occur
without any loss of stereochemical integrity, leading to
the corresponding alcohol (S)-8. This could be easily
purified by flash chromatography, affording 0.53 g of
product (2.76 mmol, 35% overall yield for both the chiral
resolution and the catalytic hydrogenation). Subse-
quently, (S)-8 was oxidized with KMnO4/H2SO4, leading
to (S)-ibuprofen 9 in 76% yield (0.43 g, 2.09 mmol). Chiral
GC analysis of the methyl ester of 9 showed an ee of 95%
indicating that the optical purity of the product stayed
unaffected all over the reaction scheme. It is to be
stressed that, therefore, it is possible to prepare enan-
tiopure (S)-ibuprofen simply by slightly increasing the
biohydrolysis time. Following this strategy, the overall
yield of 9, starting from racemic 5, is about 27%.
Recyclin g of Diol 5d . Obviously, one interesting
improvement of this synthetic scheme would be to avoid
the intrinsic 50% loss of material due to the resolution
(15) Morisseau, C.; Archelas, A.; Furstoss, R.; Baratti, J . Eur. J .
Biochem. in press.
(16) Nakajima, M.; Tomioka, K.; Iataka, Y.; Koga, K. Tetrahedron
1993, 49, 10793.
(17) Hudlicky, T.; Boros, E. E.; Boros, C. H. Tetrahedron: Asymmetry
1993, 4, 1365.
(18) Cleij, M.; Archelas, A.; Furstoss R. Tetrahedron: Asymmetry
1998, 9, 1839.

Synthesis of Enantiopure R-Methylstyrene Oxides
J . Org. Chem., Vol. 64, No. 14, 1999 5033
3.7, 3.1, and 4.1 g of dry enzymatic extract. It was similar
to method A except that centrifugation at 135000g was
omitted.
process. We have already described such strategies, based
on using either an enzyme mixture5a or a chemoenzy-
matic approach.19 In the present case, it was necessary
to reuse the formed diol and to find a way to cyclize 5d
into epoxide 5 using experimental conditions that would
lead, in the same time, to racemization. We found that
this could be achieved by treatment of 5d with HBr/AcOH
and subsequent cyclization of the bromhydrin intermedi-
ate under basic conditions.20 This indeed afforded racemic
epoxide 5 in 80% overall yield, which could thus be
resubmitted to enzymatic resolution. In other words, this
synthetic combination makes the entire process enan-
tioconvergent, allowing us to achieve the synthesis of
enantiopure (S)-ibuprofen 9 following a four-step strategy
that, starting from racemic epoxide 5, can afford a 47%
overall yield, including the resolution step.
Meth od C. This method was used in the case of C. globosum
and S. racemosum, which led, respectively, to 0.706 and 1.16
g of dry enzymatic extract. This method was similar to method
A except that, after disruption of the cells, the suspension was
directly subjected to ultrafiltration using a 0.1 µm filter. The
suspension was concentrated to 200 mL, and subsequently,
1.0 L of a pH 7 buffer solution (10 mM ammonium acetate;
0.1 mM EDTA; 0.1 mM cysteine) was added. This suspension
was concentrated again to 200 mL. The latter procedure was
repeated one more time. The collected clear solution (that has
passed through this filter) was concentrated to 75 mL by
means of ultrafiltration using a 10 kD filter. The resulting
solution was again diluted to 500 mL with a pH 7 buffer
solution (10 mM ammonium acetate; 0.1 mM EDTA; 0.1 mM
cysteine) and further concentrated to 75 mL. The latter
procedure was repeated two more times. The final concen-
trated solution was freeze-dried, and the thus-obtained enzyme
extract was stored at 4 °C.
Exp er im en ta l Section
Gen er a l Meth od s. The strains used in this work were
purchased from the ATCC, CBS, LCP, or MUCL collections.
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. Determination of the enantiomeric
excesses were performed using two different chiral columns
(0.22 mm, 25 m, Macherey-Nagel), i.e.: heptakis(6-O-methyl-
2,3-di-O-pentyl)-â-cyclodextrin (column I) and octakis(6-O-
methyl-2,3-di-O-pentyl)-γ-cyclodextrin in OV 1701 (column II).
HPLC analyses were performed using a Spheri-5 silica column
(220 × 4.6 mm, 5 µm).
P r ep a r a tion of th e F r eeze-Dr ied Cells fr om Ba cu lovi-
r u s-In fected In sect Cells.21 The human, potato, and rat EH
extracts were obtained from baculovirus infected insect cells
in which these enzymes were overexpressed.10-12 Two simple
spinner-cultures of 500 mL volume, achieved in flat-bottomed
flasks equipped with a magnetic stirrer, and containing 100
mL of rich medium (EX-CELL 401 medium supplemented with
fetal calf serum (3%), and penicillin-streptomycin (1%)) were
inoculated with SF21 insect cells (Spodoptera frugiperda) in
order to obtain a cell density of 105 cell/mL.
After 96 h incubation at 28 °C under gentle stirring (50-
70 rpm), the cell density reached 2-2.5 × 106 cell/mL, and
they were infected with the required recombinant virus at a
MOI of 0.1 pfu/mL, from a high-titer working stock of bacu-
lovirus, by simply adding the adequate inoculum to the insect
cell culture. The culture mediums of the two flasks were again
stirred 1 h at 28 °C before being poured into a 1.8 L Fernbach
flask containing 200 mL of rich medium previously warmed
at 28 °C. The medium was maintained at 28 °C during 72 h
under gentle shaking (50 rpm) on an circular shaker and then
centrifuged at 3000 tr/min during 5 min. The supernatant was
P r ep a r a tion of th e F u n ga l En zym a tic Extr a cts. The
fungal enzymatic extracts were prepared from each particular
fungus. This was achieved in two steps (1) culture of the fungi
and (2) breakage of the cells and extraction of the crude
enzymatic epoxide hydrolase activity followed by lyophiliza-
tion. The seven fungi were cultivated in a 5 L fermentor at 27
°C as previously described.6d The fungal cake was collected
by filtration and washed with water. The following procedures
were all carried out at 4 °C. The extracts were obtained in
three different ways (methods A-C, see below).
Meth od A. This method was used in the case of A. niger
(4.8 g of dry enzymatic extract) and with M. isabellina (2.1 g
of dry enzymatic extract). The fungal cake was suspended in
0.8-1.5 L buffer solution (pH 7.0: 20 mM ammonium acetate;
1.0 mM EDTA; 1.0 mM cysteine) with the use of a mixer.
Subsequently, a 200 mM solution of PMSF (protease inhibitor)
in acetone was added such that the final concentration
amounted to 0.5 mM. To break the cells, this suspension was
passed through a cell disruptor (Constant Systems from CELL-
D) at 27 kpsi and collected in an ice-cooled flask containing
ice made from distilled water. The thus-obtained suspension
was subjected to a centrifugation at 10000g for 20 min to
remove the cell walls and unbroken cells. The supernatant was
concentrated to 75 mL by means of ultrafiltration using a 10
kD filter. To this concentrated solution was added 500 mL of
a pH 7.0 buffer solution (10 mM ammonium acetate; 0.1 mM
EDTA; 0.1 mM cysteine), and the solution was again concen-
trated to 75 mL. The latter procedure was repeated two more
times. The final concentrated solution was subjected to a 1.5
h ultracentrifugation at 135000g to remove the membranes.
Sometimes, some lipidic material stayed floating on top of the
solution. This could be removed by collecting the solution with
a pipet and/or filtration of the solution through glass wool. The
thus-obtained clear solution was freeze-dried, and the obtained
powder was stored at 4 °C.
discarded and the pellet was lyophilized in
a carbonate
ammonium buffer (0.1 M), thus leading to about 1 g of freeze-
dried cells containing about 5% of overexpressed epoxide
hydrolase.
Syn th esis of Ep oxid es 1 to 7. Substrates 1-7 were
synthesized from the corresponding acetophenone precursor
by reaction with either (CH3)3S+ I- (method I) or (CH3)3SO+
I- (method II) in the presence of NaH. Better yields were
obtained with method I for the substrates 1-5. For 6 and 7
(strong electron-attracting substituents), better results were
obtained using method II. In these cases, an excess of NaH
relative to (CH3)3SO+ I- should be avoided.
Gen er a l P r oced u r e. (CH3)3S+ I- (or (CH3)3SO+ I)- (0.12
mol) was dissolved in 200 mL of DMSO. To this solution was
added 0.12 mol NaH under nitrogen at room temperature.
After the solution was stirred for 20 min, 0.1 mol of the
corresponding acetophenone in 40 mL of DMSO was added
dropwise within 20 min. After being stirred for 10-15 h at
room temperature, the reaction mixture was poured into 1 L
water, and the epoxide was extracted with 2 × 150 mL of ether.
The collected ether fractions were washed two times with 300
mL water, dried over MgSO4, and concentrated into vacuo. The
crude epoxide was purified by distillation under reduced
pressure or by flash chromatography.
Meth od B. This method was used in the case of A. terreus,
C. echinulata, and B. bassiana, which led respectively to
r-Meth ylstyr en e Oxid e 1. This epoxide was obtained as
a colorless liquid in 88% yield. Bp: 55 °C (0.3 mm). H NMR:
1
δ 1.72 (s, 3H); 2.80 and 2.98 (2 × d, 2 × 1H); 7.2-7.4 (m, 5H).
These data were consistent with the literature.22 The two
enantiomers can be separated by GC using column I (T ) 60
°C: R ) 12.6 min; S ) 14 min).
(19) Pedragosa-Moreau, S.; Morisseau, C.; Baratti, J .; Zylber, J .;
Archelas, A.; Furstoss, R. Tetrahedron 1997, 53, 9707.
(20) Golding, B. T.; Hall, D. R.; Sakrikar, S. J . Chem. Soc., Perkin
Trans. 1 1973, 1214.

5034 J . Org. Chem., Vol. 64, No. 14, 1999
Cleij et al.
p-F lu or o-r-m eth ylstyr en e Oxid e 2.23 This epoxide was
obtained as a colorless oil in 79% yield. Bp: 54 °C (4.5 mm).
1H NMR δ: 1.70 (s, 3H); 2.77 and 2.97 (2 × d, 2 × 1H); 7.0
and 7.33 (2 × t, 2 × 2H). 13C NMR δ: 22.07; 57.18; 115.35 (d,
J ) 21 Hz), 127.24 (d, J ) 8 Hz). The two enantiomers can be
separated with column I (GC) at T ) 70 °C: R ) 10.0 min; S
) 11.5 min.
p-Ch lor o-r-m eth ylstyr en e Oxid e 3. This epoxide was
obtained as a colorless oil in 82% yield. Bp: 80 °C (0.2 mm).
1H NMR δ: 1.70 (s, 3H); 2.75 and 2.97 (2 × d, 2 × 1H); 7.30
(s, 4H). These data were consistent with literature.22 The two
enantiomers can be separated by chiral GC (column I) (T )
90 °C: R ) 14.6 min; S ) 15.9 min).
p-Br om o-r-m eth ylstyr en e Oxid e 4. This epoxide was
obtained as a colorless oil in 80% yield. Bp: ) 92 °C (2.5 mm).
1H NMR δ: 1.68 (s, 3H); 2.73 and 2.96 (2 × d, 2 × 1H); 7.1-
7.5 (2 × d, 4H). 13C NMR δ: 21.59; 56.36; 57.00; 121.43; 127.11;
131.44; 140.31. These 1H and 13C NMR data were consistent
with the literature.22 The two enantiomers can be separated
by chiral GC (column I) (T ) 100 °C: R ) 20 min; S ) 21
min).
26.16; 70.96; 74.75; 126.75; 128.47; 133.18; 143.69. The 1H and
13C NMR were consistent with the literature.22 The enanti-
omers of the acetonide derivative of this diol can be separated
by chiral GC with column I at T ) 65 °C: S ) 185 min; R )
201 min.
2-(p -Br om op h en yl)-1,2-p r op a n ed iol 4d . White solid.
Mp: ) 59 °C.1H NMR δ: 1.50 (s, 3H); 2.05 (t, 1H); 2.72 (s,
1H); 3.5-3.8 (o, 2H); 7.2-7.5 (2 × d, 4H). 13C NMR δ: 25.98;
70.83; 74.60; 121.20; 127.01; 131.47; 144.05. The 1H and 13C
NMR were consistent with the literature.25 The enantiomers
of the acetonide derivative of this diol can be separated by GC
(column I) (T ) 75 °C: S ) 208 min; R ) 222 min).
2-(p-Isobu tylp h en yl)-1,2-p r op a n ed iol 5d . This diol was
1
obtained as a white solid. Mp: ) 92 °C. H NMR δ: 0.88 (d,
6H); 1.54 (s, 3H); 1.4-2.0 (s (br), 2H); 1.83 (m, 1H); 2.45 (d,
2H); 3.45-3.85 (2 × d, 2 × 1H); 7.1-7.4 (2 × d, 4H).13C NMR
δ: 22.62; 26.23; 30.40; 45.17; 71.39; 75.00; 125.04; 129.41;
1
140.90; 144.71. The H and 13C NMR were consistent with the
literature.22 The acetonide derivative of this diol could be
analyzed by chiral GC analysis (column I) (T ) 95 °C: S )
82.9 min; R ) 84.6 min).
p-Isobu tyl-r-m eth ylstyr en e Oxid e 5. This epoxide was
2-(p-Cya n op h en yl)-1,2-p r op a n ed iol 6d . This diol was
obtained as a viscous oil. 1H NMR δ: 1.53 (s, 3H); 2.05 (t, 1H);
2.85 (s, 1H); (o, 2H); 7.48-7.7 (2 × d, 4H). 13C NMR δ: 26.13;
70.66; 74.97; 110.90; 118.97; 126.34; 132.34; 151.04. Anal.
Calcd for C10H11NO2 (177.20): C, 67.8; H, 6.3; N, 7.9; O, 18.1.
obtained as a colorless oil in 87% yield following method I.
1
Bp: ) 95 °C (4 mm). H NMR δ: 0.9 (d, 6H); 1.7 (s, 3H); 1.8
(m, 1H); 2.45 (d, 2H); 2.80 and 2.95 (2 × d, 2 × 1H); 7.0-7.3
(2 × d, 4H). These data and bp were consistent with the
literature.22 The two enantiomers can be separated by chiral
GC (column I) (T ) 90 °C: R ) 31.5 min; S ) 32.4 min).
p-Cya n o-r-m eth ylstyr en e Oxid e 6.24 This epoxide is
obtained as a low-melting solid in 77% following method II.
Found: C, 67.46; H, 6.38; N, 7.64; O, 18.78. Calcd for C10H11
-
NO2 177.078 978 7; found 177.078 300 0. The acetonide deriva-
tive of this diol could not be separated efficiently enough using
the chiral GC columns available. Therefore, this diol was
converted back into the epoxide by reaction with p-toluene-
sulfonyl chloride and NaH in ether, and the thus-obtained
epoxide was analyzed as described above.
2-(p-Nitr op h en yl)-1,2-p r op a n ed iol 7d . This diol is a very
viscous slightly yellow oil. 1H NMR δ: 1.53 (s, 3H); 2.50 (t,
1H); 3.13 (s, 1H); 3.6-3.9 (o, 2H); 7.45-8.25 (2 × d, 4H). 13C
NMR δ: 26.21; 70.72; 75.04; 123.69; 126.46; 152.88. Anal.
Calcd for C9H11NO4 (197.19): C, 54.8; H, 5.6; N, 7.1; O, 32.5.
1
Mp: ) 41 °C. H NMR δ: 1.76 (s, 3H); 2.75 and 3.06 (2 × d,
2 × 1H); 7.4-7.65 (2 × d, 4H). 13C NMR δ: 21.30; 56.36; 57.29;
111.42; 118.81; 126.23; 132.33; 146.78. The 1H NMR data were
consistent with literature.24 The two enantiomers can be
separated by chiral GC (column I) (T ) 110 °C: R ) 21 min;
S ) 22 min).
p-Nitr o-r-m eth ylstyr en e Oxid e 7.24 This epoxide was
obtained as a low-melting, pale yellow, solid in 35% yield
following method II. Mp: ) 33-34 °C. It was purified by flash
chromatography (silica gel: pentane/CH2Cl2 (9/1)). 1H NMR
δ: 1.77 (s, 1H); 2.78 and 3.05 (2 × d, 2 × 1H); 7.5-8.25 (2 ×
Found: C, 54.15; H, 5.72; N, 7.14; O, 32.85.. Calcd for C9H11
-
NO4 197.068 808 0; found 197.069 000 0. As for 6d , the ac-
etonide derivative could not be separated efficiently enough
on the chiral GC columns available. Thus, the diol was
transformed back into the epoxide using p-toluenesulfonyl
chloride and NaH in ether and was analyzed as described
above.
1
d, 4H). 13C NMR δ: 21.27; 57.20; 123.64; 126.25. The H NMR
data were consistent with literature.24 The two enantiomers
can be separated by chiral GC (column I) (T ) 120 °C: R )
20.0 min; S ) 21.3 min).
Syn th esis of Diols 1d to 7d . The racemic diols 1d to 7d
were synthesized by acid hydrolysis of the corresponding
epoxides. This was achieved by adding a drop of concentrated
sulfuric acid to a mixture of 200 mg of the racemic epoxide in
a 10 mL water/2-methyl-2-propanol mixture (3/1). The reaction
was followed by TLC, and the crude product was purified by
means of flash chromatography (silica gel: dichloromethane/
ether) (9/1 v/v).
2-P h en yl-1,2-p r op a n ed iol 1d . This diol has been previ-
ously synthesized in our laboratory.6b The enantiomers of the
acetonide of this diol can be separated by chiral GC (column
I) (T ) 60 °C: S ) 46.3 min; R ) 47.8 min) (column II) (T )
70 °C: R ) 45.4 min; S ) 47.0 min).
Kin etic Mea su r em en ts. Lyophilized enzymatic extract
(5-60 mg) was dissolved in 1.0 mL of buffer (0.4 M Tris/HCl)
of the desired pH. The substrate was injected as a 0.8 M
solution in DMF (10 µL) into the buffer solution and stirred
at 27 °C. The reaction was followed by periodically withdraw-
ing 0.1 mL samples. These were saturated with NaCl and
extracted with 0.3 mL of ethyl acetate. The amount of diol
present in the organic phase (and therefore the degree of
conversion) was determined by HPLC analysis (silica gel:
hexane/ethanol 95/5 (v/v)). The ee of the epoxide was deter-
mined by GC, and the E value was calculated using the ee of
the epoxide and the calculated conversion ratio. The initial
rate of the reaction was determined by plotting the amount of
formed diol against the time (t) and extrapolation to t ) 0.
Syn th esis of Ibu p r ofen . (R/S)-4-Isobu tyl-r-m eth ylsty-
r en e Oxid e 5. To a solution of 20.8 g (102 mmol) of (CH3)3S+I-
in 200 mL of DMSO was carefully added 4.1 g of NaH
(approximately 60% ) 103 mmol) in an N2 atmosphere. After
the solution was stirred for 15 min, 15 g (85 mmol) of
4-isobutylacetophenone in 50 mL of DMSO was added drop-
wise within a 20 min period at room temperature. The solution
was stirred overnight. Subsequently, the reaction mixture was
poured into 1 L of tap water, and the epoxide was extracted
with 2 × 200 mL pentane. The collected pentane fractions were
2-(p-F lu or op h en yl)-1,2-p r op a n ed iol 2d . This diol was
obtained as a viscous oil. 1H NMR δ: 1.47 (s, 3H); 2.42 (2 × d,
1H); 2.96 (s, 1H); 3.5-3.8 (o, 2H); 6.9-7.5 (m, 4H). 13C NMR
δ: 26.16; 71.09; 115.17 (d, J ) 21 Hz); 126.85 (d, J ) 8 Hz).
Anal. Calcd for C9H11O2F (170.18): C, 63.5; H, 6.5; F, 11.2.
Found: C, 63.32; H, 6.35; F, 11.2. Exact mass calcd for
C9H11O2F 170.074 307 9; found 170.074 600 0. The acetonide
of this diol can be separated by chiral GC with column I at T
) 50 °C: S ) 116 min; R ) 126 min.
2-(p-Ch lor op h en yl)-1,2-p r op a n ed iol 3d . This diol is a
1
very viscous oil. H NMR δ: 1.47 (s, 3H); 2.43 (t, 1H); 2.97 (s,
1H, OH); 3.45-3.8 (o, 2H); 7.25-7.45 (2 × d, 4H). 13C NMR δ:
(23) Niffeler, R.; Zondler, H.; Strum, E. Eur. Pat. Appl. EP12 6430
A2, 1984.
(24) Rosman, L. B.; Beylin, V. G.; Gaddamidi, V.; Hooberman, B.
H. Mutat. Res. 1986, 171, 63.
(21) O’Reilly, D. R.; Miller, L. K.; Luckow V. A. In Baculovirus
expression vectors. A laboratory manual; W. H. Freeman Co.: New
York, 1992.
(22) Basavaiah, D.; Raju, S. B. Synth. Commun. 1995, 25, 3293.
(25) Suprun, W. J . Pratkt. Chem./ Chem.-Ztg 1996, 3, 231.

Synthesis of Enantiopure R-Methylstyrene Oxides
J . Org. Chem., Vol. 64, No. 14, 1999 5035
washed with 2 × 300 mL of water, dried over MgSO4, and
concentrated in vacuo. The oily residue was subjected to
distillation, which gave pure (R/ S)-5 as a colorless oil (bp )
95 °C (4 mm)). This afforded 14.1 g (74 mmol; 87% yield).
P r ep a r a tive-Sca le En zym a tic Resolu tion of (R/S)-5. A.
niger enzymatic extract (2.6 g) was dissolved in 30 mL of buffer
(pH 8.0; 0.4 M Tris/HCl) at 4 °C. To this solution was added
1.5 g (7.9 mmol) of 5. The solution was shaken for 24 days at
4 °C (in a cold room), and the ee of the remaining substrate 5
was followed by chiral GC analysis. When it reached 96%, the
reaction was stopped by extraction of the epoxide with pentane
(3 × 10 mL). Subsequently, the diol was extracted from the
remaining aqueous phase with diethyl ether (3 × 15 mL). The
formed diol 5d was purified by column chromatography (silica
gel, dichloromethane). The collected pentane fractions not only
contained (S)-5 but also some diol 5d (and some uncharacter-
ized compounds that were presumably present in the enzy-
matic extract). They were concentrated in vacuo and used
without further purification since, on a silica gel column, 5
rapidly rearranged into the corresponding (racemic) aldehyde.
Distillation also led to formation of this aldehyde, which we
were unable to separate from the epoxide. The total isolated
amount of (R)-5d was 0.91 g (4.38 mmol) corresponding to a
(S)-Ibu p r ofen 9. A solution of 2.76 mmol (0.53 g) of (S)-8
(ee 95%) in 12 mL acetone/H2SO4 (3 N) was cooled in an ice
bath under stirring. To this solution was added 5.0 mmol of
KMnO4 portionwise at 0 °C. After the solution was stirred for
1 h (at 0 °C), the ice bath was removed and the mixture was
stirred for 2 more hours at room temperature. Subsequently,
solid NaHSO3 was added until the solution became colorless.
The mixture was dissolved in 50 mL of water and the aqueous
phase was extracted two times with 25 mL of ether/pentane
(1/2, v/v). The collected organic phase was extracted with a
2% NaOH solution (2 × 50 mL), and the thus-obtained
collected aqueous phase was acidified to pH 1 with HCl (35%).
The turbid aqueous phase was extracted with CHCl3 (2 × 50
mL), and the collected organic fractions were dried over MgSO4
and concentrated in vacuo. The residue was subjected to flash
chromatography (silica gel: dichloromethane/ether, 4/1, v/v),
which gave 0.43 g of 9 (2.09 mmol, 76% yield) as a sticky low-
melting solid. Mp: 48 °C. [R]20
: +55 (c 2, EtOH). The
D
commercially available (S)-ibuprofen from Aldrich (ee 98%)
shows an mp of 51-53 °C and an optical rotation [R]20 +59
D
1
(c 2, EtOH). H NMR δ: 0.88 (d, 6H); 1.50 (d); 1.81 (m, 1H);
2.45 (d, 2H); 3.69 (q, 1H); 7.0-7.3 (2 × d, 4H); 9.8 (s (very
broad), 1H).
55% yield (ee 70%). [R]20D: -11 (c 1.17; CHCl3) (lit.22 [R]20
A small amount of 9 was converted into its methyl ester
with SOCl2 in methanol. This could be analyzed by chiral GC
(column I) (80 °C: S ) 118 min and R ) 121 min) and showed
an ee of 95%.
D
-4.5 (c 1.55; CHCl3)). The corresponding acetonide could be
analyzed by chiral GC (column I) (T ) 95 °C: S ) 82.9 min
and R ) 84.6 min).
(S)-2-(4-Isobu tylp h en yl)p r op a n ol 8. A suspension of 0.2
g of Pd powder (submicron/Aldrich) in 20 mL of N-methylfor-
mamide was stirred for 20 min in a H2 atmosphere at room
temperature. The mixture was cooled to 0 °C, and the crude
(S)-5 epoxide obtained above was added. The reaction mixture
was stirred overnight at 0 °C under H2. Subsequently, the
reaction mixture was poured into 150 mL of water. The
aqueous phase was extracted with pentane (2 × 35 mL). The
collected pentane fractions were dried over MgSO4 and con-
centrated in vacuo. The residue was subjected to flash chro-
matography (silica gel: pentane/dichloromethane 3/1 (v/v)).
The formed alcohol (S)-8 (0.53 g; 2.76 mmol) was obtained as
a colorless oil. A fraction of 5d , present in the crude epoxide,
was eluted from the chromatography column with dichlo-
romethane and added to the previously obtained diol. The
overall yield of 8, starting from racemic 5, amounted to
Syn th esis of Ra cem ic Ep oxid e 5 fr om Diol (R)-5d . The
method we have used was described by Golding et al.20 To 8.0
mL of acetic acid saturated with HBr was added 0.91 g (4.38
mmol) of (R)-5d (ee ) 70%) at room temperature. After being
stirred for 30 min, the reaction mixture was poured into 40
mL of water and neutralized with solid Na2CO3. The solution
was extracted with ether (2 × 20 mL), and the combined ether
fractions were dried over MgSO4 and concentrated in vacuo.
The residue was dissolved in 3 mL of methanol, and NaOH
(4.4 mmol) in 4 mL of methanol, was added for 15 min at room
temperature. After being stirred for 30 min, the reaction
mixture was poured into 25 mL water and was extracted with
pentane (2 × 15 mL). The combined pentane fractions were
dried over MgSO4 and concentrated in vacuo. The residue was
distilled (bulb-to-bulb distillation), which gave 5 as a colorless
oil. The yield amounted to 80% (0.66 g, 3.50 mmol), and the
ee was shown to be about 0%, showing that complete racem-
ization occurred during this process.
35%. [R]20
:
-16 (c 1.0, CHCl3) (lit.26 [R]20 -14.8 (c 1.6;
D
D
CHCl3)). Alcohol 8 was analyzed by chiral GC (column III) (90
°C: R ) 133 min and S ) 139 min); the measured ee was
95%. 1H NMR δ: 0.9 (d, 6H); 1.25 (d, 3H); 1.34 (t, 1H); 1.82
(m, 1H); 2.42 (d, 2H); 2.8 (m, 1H); 3.67 (2 × d, 2H); 7.0-7.2 (2
× d, 4H).
Ack n ow led gm en t. This work has been achieved in
the context of the EC BIOC4-CT95-0005. Financial
support to one us (M.C.) in the frame of this contract is
greatly acknowledged.
(26) Chiarino, D.; Ferrario, F.; Pellacini, F.; Sala, A. J . Heterocycl.
Chem. 1989, 26, 589.
J O982101+