Enzymatic transformations; 61. Preparation of enantiopure trifluoromethyl-substituted aromatic epoxides and vicinal diols using the Aspergillus niger epoxide hydrolase-catalysed resolution

Article abstract of DOI:10.1002/adsc.200606061

The resolution of eight differently substituted trifluoromethylstyrene oxide derivatives was explored using the Aspergillus niger epoxide hydrolase. The obtained results indicate that all (but one) epoxides were efficiently processed by this enzyme, under

Full text of DOI:10.1002/adsc.200606061

COMMUNICATIONS
DOI: 10.1002/adsc.200606061
Enzymatic Transformations; 61. Preparation of Enantiopure
Trifluoromethyl-Substituted Aromatic Epoxides and Vicinal Diols
using the Aspergillus niger Epoxide Hydrolase-Catalysed
Resolution
Justine Deregnaucourt,a Alain Archelas,a Fabien Barbirato,b Jean-Marc Paris,b
and Roland Furstossa,*
a
Groupe Biocatalyse et Chimie Fine, FRECNRS 2712, UniversitØ de la MØditerranØe, FacultØ des Sciences de Luminy,
Case 901, 163 avenue de Luminy, 13288 Marseille Cedex 9, France
Fax : (+33)-4-91-82-91-45; e-mail: furstoss@luminy.univ-mrs.fr
StØ RHODIA, Centre de Recherches de Lyon, 85 avenue des Fr›res Perret, 69192 Saint-Fons Cedex, France
b
Received: February 17, 2006; Accepted: May 19, 2006
Abstract: The resolution of eight differently substi-
tuted trifluoromethylstyrene oxide derivatives was
explored using the Aspergillus niger epoxide hydro-
lase. The obtained results indicate that all (but one)
epoxides were efficiently processed by this enzyme,
under very mild experimental conditions. The spe-
cific activity and the enantioselectivity of the
enzyme against each substrate were determined.
Circular dichroism analysis was also performed, al-
lowing us to establish the absolute configuration of
the products.
In this context, we were strongly interested in explor-
ing the possibility to use the overexpressed (and com-
mercially available)[2] Aspergillus niger (A. niger) ep-
oxide hydrolase to perform the biocatalysed hydrolyt-
ic kinetic resolution of some selected trifluoromethyl-
substituted styrene oxide derivatives. We here de-
scribe our results in this context.
Over the last decades we and others have demon-
strated that preparation of various enantiomerically
pure epoxides and vicinal diols, including aliphatic as
well as aromatic substrates, can be very efficiently
achieved by carrying out a so-called biocatalysed hy-
drolytic kinetic resolution (BHKR) of the corre-
sponding racemate.[3] These “green chemistry” resolu-
tions were performed using various epoxide hydrolas-
es from different origins, thus offering an easy, cheap
and environmentally gentle alternative to substrate
limited and potentially toxic transition metal-based
methodologies. Our studies using the A. niger enzyme
Keywords: A. niger epoxide hydrolase; circular di-
chroism; enantiopure epoxides; enzyme catalysis;
fluorinated derivatives; kinetic resolution
Asymmetric synthesis of selectively fluorinated com- demonstrated that this enzyme was able to operate in
pounds, driven initially by the necessity of preparing plain water at room temperature and in two-phase re-
biologically relevant compounds in an enantiomeri- actors allowing us to run the reaction within a few
cally pure form, has gained tremendous impetus over hours, at substrate concentrations as high as 2.5M
the recent years.[1] As a consequence, the develop- (i.e., 500 gLÀ1).[4,5] We have now raised the question
ment of practical methods for preparing selectively whether this type of process could be applied to some
fluorinated yet stereochemically defined compounds fluorine-bearing substrates. The trifluoromethyl-sub-
is definitely critical to the further advances of fluorine stituted styrene oxide derivatives rac-1–8 were select-
chemistry at the biomedical interface. Fluorinated or- ed to explore this possibility and, more precisely, to
ganic compounds are, however, well known for their determine the relative influence of (a) their ortho-,
unexpected and generally unusual reactivity. Thus, meta- and para- substitution, (b) the intrinsic nature
due to the specific fluorine factors such as steric, elec- of different para substituents, (c) the presence of an
trostatic and electronic features, as well as to the abil- additional gem-methyl group on the (thus trisubstitut-
ity of fluorine in coordinating positively charged and ed) epoxide moiety (Scheme 1).
electron-deficient species, many of the methods estab-
The racemic epoxides 1–8 used in this study were
lished for the asymmetric synthesis of fluorine-free synthesised with moderate to excellent (46–99%)
compounds do not work or give impractical outcomes yields by reaction of the corresponding commercially
when applied to prepare fluorine-containing targets. available aldehyde or ketone with either trimethylsul-
Adv. Synth. Catal. 2006, 348, 1165 – 1169
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1165

Justine Deregnaucourt et al.
COMMUNICATIONS
whereas some other ones (for instance, 7) were very
stable. Initial velocity of the EH-catalysed resolution
was also determined and allowed us to quantify the
specific activity of the enzyme against each epoxide.
As usual, the activity was dependent on the structure
of each substrate. Thus, ortho-substituted epoxides 1
and 6 showed much lower reactivity as compared to
styrene oxide itself (11 U·mgÀ1),[8] whereas the other
(meta- and para-substituted) substrates exhibited
comparable – if not better – reactivity. Surprisingly
enough, epoxides 7 and 8, bearing an additional gem-
methyl group showed the highest activity. This is a
very interesting and noteworthy observation since
even the Jacobsenꢁs salen (Co) catalysts, claimed to
Scheme 1. Structure of the racemic epoxides used in this
study.
fonium or trimethyloxosulfonium iodide, following be the best transition metal-based catalysts for achiev-
the procedure previously described by Corey.[6] Each ing the hydrolytic kinetic resolution of epoxides, are
of these epoxides was also quantitatively hydrolysed inoperative on such trisubstituted substrates.[9]
in H2O/THF solution in the presence of catalytic
Analytical scale resolutions: Each one of the epox-
amounts of sulphuric acid, thus affording racemic ides rac-1–8 was submitted to small-scale experiments
probes of the corresponding vicinal diols rac-1d–8d. in order to (a) establish the kinetic profile of the reso-
The solubility of each of the epoxides in plain water lution, (b) follow the evolution of the conversion
and room temperature was estimated as being in the ratio c in correlation with the ee of the less reactive
order of 1 mM. However, using an aqueous solution (recovered) epoxide enantiomer, and (c) obtain a
containing 20 to 30% DMSO allowed enhancement small amount of the corresponding formed diol, in
of this solubility to about 2.5 mM. Therefore, the ex- order to determine the absolute configuration (see
periments described in this exploratory study were below). The (weight to weight) substrate over enzy-
achieved, at 278C, using this solvent mixture, owing matic powder R ratio used is indicated in Table 1. It is
to our previous observation indicating that the A. to be stressed that these R ratios can be valuably
niger epoxide hydrolase was not noticeably affected compared since all the experiments were conducted
in these conditions over a few hours period.[7]
using the same enzymatic preparation (showing an ac-
The precise experimental conditions used to run tivity of 11 U·mgÀ1 as measured by UV spectroscopy
each one of the experiments, including the concentra- against styrene oxide[8]). Thus, by using R ratios of 58
tion of substrate, the percentage of DMSO, the spon- or 120 (i.e.,1 g of enzyme powder for 58 or 120 g of
taneous consumption and the specific activity of the racemic epoxide) the resolution was complete within
enzyme against each epoxide are indicated in Table 1. 20 to 60 min, depending on the substrate.
Blank experiments allowing us to accurately measure
Our results indicate that all – but one (rac-6) – ep-
the spontaneous consumption of each specific sub- oxides were indeed efficiently processed by the
strate were performed under these experimental con- enzyme, following a classical one-phase resolution
ditions, indicating that some epoxides (for instance, 1) profile.[10] As an example, the kinetic profile of rac-4
were rather sensitive to spontaneous hydrolysis is illustrated in Figure 1. The apparent E value (Eapp
)
Table 1. Experimental results obtained upon hydrolytic kinetic resolution of rac-1–8 catalysed by the A. niger EH.
Substrate
1
2
3
4
5
6
7
8
Epoxide concentration [mM]
% DMSO
2.5
25
44.4
4.9
58
2.66
20
0
8.9
67
50
2.5
20
11.3
16.0
58
2.5
30
8.8
10.8
116
60
2
30
2
12.1
116
60
2.5
20
4
2.3
58
-
2.5
30
0
29.9
58
1.8
30
18.9
16.9
58
% Spontaneous consumption [hÀ1]
À1
Specific activity [U·mgpowder
]
R
Reaction time[a] [min]
120
30
20
40
Absolute configuration (epoxide)
Absolute configuration (diol)
Eapp
(S)(S)(S)(S)(S)
-
-
1
(S)(S)
(R)(R)
25
(R)(R)(R)(R)(R)
[b]
5[c]
10[c]
50
30
160
30
[a]
Reaction time to reach an epoxide of ee>99%.
[b] The ee was determined by chiral GC on a Chirasil-Dex CB Chrompack column.
[c]
The ee determined by chiral GC on a Lipodex G column.
1166
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1165 – 1169

Enantiopure Trifluoromethyl-Substituted Aromatic Epoxides and Vicinal Diols
COMMUNICATIONS
sign of the Cotton effect of the CD band around
305 nm (“band IV”) observed for a complex formed
between a given vicinal diol and dimolybdenum tet-
raacetate. According to their study, a negative sign of
the “band IV” is associated with the (R) absolute con-
figuration of the diol and, although being an empirical
approach, no exception to this rule is known up to
now. Thus, each one of our obtained diols 1d–8d was
submitted to CD spectroscopy measurements follow-
ing this methodology. The results obtained are de-
scribed in Table 2 and Figure 2. These measurements
indicate that all our diols led to a negative sign for
the band IV, leading to the conclusion that they all
were of (R) absolute configuration. This is in perfect
agreement with our previous observations that, as a
general feature, the A. niger epoxide hydrolase (a)
preferably hydrolyses the (R) epoxide of styrene
oxide derivatives by attack at the terminal – less sub-
stituted – carbon atom of the oxirane ring (i.e., with
retention of configuration at the benzylic carbon
atom), thus leading to the (R) diol and (b) that the
Figure 1. . Kinetic profile of the resolution of rac-4 catalysed
by the A. niger EH. Epoxide enantiomeric excess, eeepox (tri-
angles); conversion ratio c (circles); chemical hydrolysis=
blank experiment (diamonds). Experimental conditions:
278C; H2O/DMSO 30% (v/v); R=116.
of each resolution was calculated by using both the c para substitution allows the most efficient resolution
and eeepox values on the basis of the appropriate Sihꢁs
equation, following a curve-fitting methodology.[11]
Interestingly, it can be observed that this Eapp value
appeared to be very low for the ortho-substituted
compounds 1 and 6, moderate for the meta-substitut-
ed substrates 2 and 7 and good to even excellent for
the para-substituted epoxides 3, 4, 5 and 8.
Determination of the absolute configuration of the
obtained epoxides and diols: As far as the absolute
configuration of the obtained enantiomerically en-
riched epoxides and diols is concerned, examination
of the literature indicated that, except for epoxide
(S)-2[12] and diol (R)-3d,[13] their absolute configura-
tion was unknown. We therefore had to achieve this
determination. This was performed by using circular
dichroism spectroscopy, following a methodology de-
veloped by Schnatzke and co-workers[14] and recently
revisited by Di Bari and co-workers.[15] According to
these authors, the absolute configuration of various
vicinal diols can be determined on the basis of the 1d–8d in the presence of Mo2
Figure 2. Circular dichroism spectra of enantioenriched diol
(AcO4) (DMSO solution).
AHCTREUNG
Table 2. Circular dichroism studies of enantioenriched diols 1d–8d in the presence of Mo2
ACHTREUNG
Diol
ee[a]
G
Ratio diol/Mo2
G
)
V
IV
III
II
1d
2d
3d
4d
5d
7d
8d
77[b]
13.2
84.3
94.5
85
0.375
0.75
0.375
0.375
0.375
0.75
1
1
1
1
1
1
1
280 (0.12)
279 (0.04)
270 (0.07)
272 (0.26)
271 (0.22)
277 (0.15)
273 (0.15)
305 (À0.66)
301 (À0.14)
308 (À0.60)
310 (À1.22)
308 (À1.12)
316 (À0.22)
308 (À0.18)
348 (0.06)
385 (À0.18)
375 (À0.05)
381 (À0.20)
375 (À0.33)
379 (0.30)
352 (À0.01)
350 (À0.12)
352 (À0.22)
352 (À0.18)
342 (À0.01)
350 (0.02)
59
88.3
385 (À0.05)
0.375
379 (À0.06)
[a]
Determined by chiral GC on a Lipodex column after derivatisation into the corresponding acetonide.
[b] Determined by chiral GC on a Chirasil-Dex CB Chromback column after dimethylation.
Adv. Synth. Catal. 2006, 348, 1165 – 1169
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1167

Justine Deregnaucourt et al.
COMMUNICATIONS
(unpublished results). Each one of the obtained diols enzyme powder ratio of about 60 to 120, these resolu-
was further cyclised back, without loss of stereochem- tions could be performed within one hour (or less).
ical integrity, into the corresponding epoxide. Compa- The reactivity as well as the enantioselectivity of the
rative chiral GC analysis confirmed that the unreact- selected substrates was in perfect agreement with the
ed (recovered) epoxide was, as expected, of (S) abso- previous tendency we have observed using this
lute configuration.
enzyme. Thus, the para-substituted derivatives exhib-
As an overall result, it thus appears that this meth- ited the highest E value (up the 160) and, interesting-
odology allowed us to prepare both the epoxides of ly, the 1,1-gem-disubstituted epoxides were also proc-
(S) configuration and the diols of (R) configuration. essed. The HKR of such disubstituted epoxides has
To the best of our knowledge, most – if not all – of never been described using Jacobsenꢁs salen(Co)OAc
these products have never been prepared previously catalysts, and control experiments we have attempted
in enantiopure form. Moreover since, in principle, the following this methodology on our substrates were
enantiopure (S) epoxides can be chemically hydro- unsuccessful. The described methodology thus al-
lysed to the corresponding (S) diol, and the enan- lowed us to prepare all (but two) of the fluorinated
tioenriched (R) diols cyclised back to the (R) epoxide, epoxides or diols in enantioenriched form, most (but
this methodology opens the way to all the possible ep- two) of them for the first time. Preliminary experi-
oxide and diol enantiomers.[5a,16] Indeed, a second ments aimed at increasing the substrate concentration
resolution cycle applied to the thus obtained enan- by using a two-phase reactor approach, however, did
tioenriched (R) epoxides would afford the corre- not lead to satisfactory results. Further work toward
sponding diols in enantiomerically pure form.
this goal is ongoing in our laboratory and will be de-
Further experimental improvements: We have previ- scribed later on.
ously described several examples illustrating the fact
that, by using a two-phase reactor, such a hydrolytic
kinetic resolution could be performed on different Experimental Section
styrene oxide derivatives at a substrate concentration
as high as several hundred grams per litre.[4] This very Analytical Scale BHKR Experiments
interesting achievement, paving the way to possible
In a typical experiment, 5.1 mg (2.5 mm) of 4-(trifluorome-
industrial-scale application, was in fact based on the
possibility to assure a sufficient phase-transfer rate
between the aqueous and organic (the substrate
itself) phases by creating an emulsion under vigorous
stirring. Unfortunately, our preliminary experiments
aimed at applying this methodology to either sub-
strate 1–8 did not lead to satisfactory results. This is
obviously due to the presence of fluorine atoms, as it
is well known that fluorinated derivatives do exhibit
rather surprising (and specific) physico-chemical
properties.[17] As a consequence, this interestingly il-
lustrates the fact that, although chemical reactivities
of fluorine-bearing epoxides toward the A. niger ep-
oxide hydrolase are not altered, practical limitations
to scale-up may be encountered due to such specific
properties. Obviously, more elaborated methodologies
will have to be set up in order to allow efficient reso-
lution to be performed.
thoxyphenyl)-oxirane 4 were dissolved in 3 mL of DMSO.
Distilled water (7.9 mL) as well as 3 mL of 3-(trifluoro-
methyl)-acetophenone (as an internal standard) were added.
A solution of A. niger epoxide hydrolase was prepared sepa-
rately by dissolving 2 mg of recombinant enzymatic powder
(about 25% purity) in 4.55 mL of distilled water. The two
solutions were placed at 278C for half an hour in order to
equilibrate the temperature. The reaction was started by ad-
dition of 100 mL of the enzymatic solution to the epoxide so-
lution. Aliquots (400 mL) were withdrawn at definite time
intervals and mixed with 200 mL of acetonitrile to stop the
reaction, under vortex agitation. This sample was further ex-
tracted with 400 mL of isooctane and the organic phase ana-
lysed by chiral GC for conversion ratio and ee of the epox-
ide.
Circular Dichroism Measurements
The aim of this study was to explore the possibility
to perform the hydrolytic kinetic resolution of various
trifluoromethyl-substituted styrene oxide derivatives
using the A. niger epoxide hydrolase. Our results de-
scribed in this paper indicate that, in spite of the pres-
ence of fluorine atoms, which very often disturb
chemical reactivity, this is indeed possible. Thus, we
have shown that, for all (but one) of the epoxides we
have studied, such a resolution can be performed very
efficiently under gentle experimental conditions, i.e.,
at room temperature and in plain water/DMSO solu-
tion. Moreover, by using a (w/w) substrate over
A DMSO solution of Mo2(OAc)4 (2 mL) is placed in a 3 mL
A
UV cuvette and a DMSO solution (200 mL) of diol is added.
The final concentrations of reactant and diol are between
0.375 and 0.750 mm depending on the specific diol. The cuv-
ette is shaken and the ICD spectra observed immediately
(208C; 50 nmminÀ1; time constant 1 s; bandwith 2 nm). A
new spectrum was recorded each 10 min until stabilisation
of the signal.
1168
www.asc.wiley-vch.de
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2006, 348, 1165 – 1169

Enantiopure Trifluoromethyl-Substituted Aromatic Epoxides and Vicinal Diols
COMMUNICATIONS
[7] C. Morisseau, A. Archelas, C. Guitton, D. Faucher, R.
Furstoss, J. Baratti, Eur. J. Biochem. 1999, 263, 386–
396.
Acknowledgements
This work is part of the PhD thesis of J. Deregnaucourt. The
CNRS and the Rhodia Company (division Organique fine)
are greatly acknowledged for their financial support.
[8] C. Mateo, A. Archelas, R. Furstoss, Anal. Biochem.
2003, 314, 135–141.
[9] D. E. J. E. Robinson, S. D. Bull, Tetrahedron: Asymme-
try 2003, 14, 1407–1446.
[10] In some rare cases, the kinetic profile of such a resolu-
tion can exhibit a two-phase profile: R. Rink, D. B.
Janssen, Biochemistry 1998, 37, 18119–18127.
[11] C.-S. Chen, Y. Fujimoto, G. Girdaukas, C. J. Sih, J. Am.
Chem. Soc. 1982, 104, 7294–7299.
References
[1] Enantiocontrolled synthesis of fluoro-organic com-
pounds: Stereochemical Challenges and biomedical tar-
gets, (Ed.: V. A. Soloshonok), Wiley, Chichester, 1999.
[2] Fluka Catalogue, ref. 71832.
[12] M. J. Ferris, European Patent 40,000; Chem. Abstr.
[3] a) W. C. Choi, C. Y. Choi, Biotechnol. Bioprocess Eng.
2005, 10, 167–179; b) E. J. de Vries, D. B. Janssen,
Curr. Opin. Biotechnol. 2003, 14, 414–420; c) A. Ar-
chelas, R. Furstoss, Curr. Opin. Chem. Biol. 2001, 5,
112–119.
[4] A. Archelas, M. Arand, J. Baratti, R. Furstoss, French
Patent Application 9905711, 1999; International Patent
Application PCT/FR00/01217, 2000.
1981, 96, 199276.
[13] K. Hirose, K. Ogasahara, K. Nishioka, Y. Tobe, K. Nae-
mura, J. Chem. Soc., Perkin Trans. 2 2000, 1984–1993.
[14] a) J. Frelek, M. Geiger, W. Voelter, Curr. Org. Chem.
1999, 3, 117–146; b) G. Snatzke, U. Wagner, H. P.
Wolff, Tetrahedron 1981, 37, 349–361.
[15] L. Di Bari, G. Pescitelli, C. Pratelli, D. Pini, P. Salva-
dori, J. Org. Chem. 2001, 66, 4819–4825.
[5] a) N. Monfort, A. Archelas, R. Furstoss, Tetrahedron
2004, 60, 601–604; b) K. M. Manoj, A. Archelas, J. Bar-
atti, R. Furstoss, Tetrahedron 2001, 57, 695–701; c) M.
Cleij, A. Archelas, R. Furstoss, J. Org. Chem. 1999, 64,
5029–5035.
[16] a) H. C. Kolb, K. B. Sharpless, Tetrahedron 1992, 48,
10515–10530; b) B. T. Golding, D. R. Hall, S. Sakrikar,
J. Chem. Soc., Perkin Trans. 1 1973, 1214–1220.
[17] a) B. E. Smart, J. Fluorine Chem. 2001, 109, 3–11;
b) J. D. Dunitz, ChemBioChem. 2004, 5, 614–621.
[6] E. J. Corey, M. Chaykovsky, J. Am. Chem. Soc. 1965,
87, 1353–1364.
Adv. Synth. Catal. 2006, 348, 1165 – 1169
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.asc.wiley-vch.de
1169