Resolution of limonene 1,2-epoxide diastereomers by mercury(II) ions

Article abstract of DOI:10.1016/S0040-4039(01)01037-1

When HgCl₂ was added to a diastereomeric mixture of cis- and trans-(4S)-limonene 1,2-epoxide, the Hg(II) ions stereoselectively complexed to the cis epoxide, enabling ring opening by water. The resulting mercuric salt could be demetalated by treatment with NaBH₄, giving a mixture of diastereomeric (1S,2S,4S)- and (1R,2R,4S)-diols. The remaining trans-(4S)-epoxide was obtained in >98% d.e. and 40% yield. For reactions on a larger scale, the most convenient reaction system was Hg(OAc)₂ in 50% acetone/tris-HCl buffer pH 7.0. The reaction rate was affected by the pH, with pH 6-8 as optimum.

Full text of DOI:10.1016/S0040-4039(01)01037-1

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 5521–5524
Resolution of limonene 1,2-epoxide diastereomers by
mercury(II) ions
Marie¨t J. van der Werf,a Hugo Jongejanb and Maurice C. R. Franssenb,*
aDepartment of Applied Microbiology and Gene Technology, TNO Nutrition and Food Research, PO Box 360,
3700 AJ Zeist, The Netherlands
bLaboratory of Organic Chemistry, Wageningen University, Dreijenplein 8, 6703 HB Wageningen, The Netherlands
Received 17 April 2001; accepted 8 June 2001
Abstract—When HgCl2 was added to a diastereomeric mixture of cis- and trans-(4S)-limonene 1,2-epoxide, the Hg(II) ions
stereoselectively complexed to the cis epoxide, enabling ring opening by water. The resulting mercuric salt could be demetalated
by treatment with NaBH4, giving a mixture of diastereomeric (1S,2S,4S)- and (1R,2R,4S)-diols. The remaining trans-(4S)-epoxide
was obtained in >98% d.e. and 40% yield. For reactions on a larger scale, the most convenient reaction system was Hg(OAc)2 in
50% acetone/tris-HCl buffer pH 7.0. The reaction rate was affected by the pH, with pH 6–8 as optimum. © 2001 Elsevier Science
Ltd. All rights reserved.
Optically pure epoxides are versatile building blocks for
the synthesis of optically pure bioactive compounds.
Both biological1–3 and chemical3,4 methods for the pro-
duction of optically pure epoxides have been described.
Kinetic resolution is an attractive method because
racemic epoxides are easily accessible.5 In recent years,
many reports have appeared in which eukaryotic and
bacterial epoxide hydrolases (EH) have been employed
(Refs. 6, 7 and 8 and papers cited therein). Only
recently an efficient chemical kinetic resolution method
was developed, using chiral Co(salen) complexes as the
catalysts, for the production of enantiopure epoxides.9
Limonene is a chiral natural product, of which the
enantiomers are readily available in optically pure
form. Their 1,2-epoxides are also commercially avail-
able, but the pure isomers of the (+)-(4R)-epoxides are
expensive and the (−)-(4S)-epoxides can only be
obtained as a mixture of diastereomers. Searching for
inhibitors of limonene 1,2-epoxide hydrolase from
NaBH4
OH
OH
OH
Hg2+
OHg+
fast
fast
O
O
Hg2+
fast
3
5 (1S,2S,4S)
70% OH2
30%
(1R,2S,4S)-cis-1
2
OHg+
OH
NaBH4
OH
OH
Hg2+
very slow reaction
4
6 (1R,2R,4S)
O
(1S,2R,4S)-trans-1
Scheme 1.
Keywords: limonene epoxide; mercury; complexation; diastereomer resolution.
* Corresponding author.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)01037-1

5522
M. J. 6an der Werf et al. / Tetrahedron Letters 42 (2001) 5521–5524
Rhodococcus erythropolis,10 an unexpected phenomenon
was observed when HgCl2 was tested. Even in the
absence of enzyme, the diastereoselective conversion of
(4S)-limonene 1,2-epoxide [(4S)-1] was observed. In
this report we describe this reaction in more detail,
since it can be used for the preparative separation of
limonene 1,2-epoxide diastereomers.
tion mixture was treated with NaBH4, two 1,2-diols
were formed as a 7:3 diastereomeric mixture of the
(1S,2S,4S) diequatorial diol 5 and the (1R,2R,4S) diax-
ial diol 6.12
The following mechanism is proposed for this reaction.
Just like in the well-known oxymercuration reaction,13–15
the first step is the formation of a complex of Hg(II)
with the p-orbital of the olefin.16,17 In the oxymercura-
tion reaction, this p-complex is attacked by a nucleo-
phile from the back side, leading to a covalent
organomercuric species. In our case, ligation of the
oxygen atom of the epoxide to the mercuric ion is
faster, giving a bridged complex in which the cyclohex-
ane ring is presumably in the boat formation (structure
2, Scheme 1). This bridged p-complex is more readily
formed with (4S)-cis-1 than with (4S)-trans-1, since
these groups are quite close to each other in the cis
isomer whereas they are remote in the trans isomer.
Similar steric restrains have been reported for the bromo-
etherification of 118 and the iodoetherification of
carveol.19 The Lewis acid Hg2+ polarizes the car-
bonꢀoxygen bonds of the epoxide, enabling rapid ring
opening by water which leads to the mercuric salts 3
and 4. Reduction of the Hg2+ ion with NaBH4 breaks
up 3 and 4 and releases the diastereomeric diols 5 and
6, respectively.
Commercially available (4S)-1 is a diastereomeric mix-
ture of 57% (1R,2S,4S)-cis-1 and 43% (1S,2R,4S)-
trans-1, corresponding to an d.e. of 14%. When a
solution of (4S)-1 in a phosphate buffer (pH 7.0) was
incubated with HgCl2, cis-(1R,2S,4S)-1 was converted
preferentially, leaving unreacted trans-(1S,2R,4S)-1
(Scheme 1).11
The yield and diastereomeric excess of the trans-(4S)-1
remaining depended on the amount of HgCl2 added. If
the amount of HgCl2 added was equimolar to the
amount of cis-(4S)-1 present, trans-(4S)-1 remained in
>98% d.e. and 40% yield (=93% of the maximal theore-
tical yield).
The stereoselectivity of this reaction was due to the
difference in reaction rates of the two diastereomers of
(4S)-1 with the Hg(II) ions. If <57 mol% HgCl2 was
added, the reaction occurred instantaneous; samples
taken after 1 minute showed the same yield as those of
samples incubated for 2 h at 30°C. However, if more
than 57 mol% HgCl2 was added, the trans-epoxide
reacts too, although it takes 2 h of incubation to reach
maximal conversion (see Table 1).
The reaction of Hg2+ with (4S)-1 was observed only at
neutral pH (6–9) (Table 1). The type of buffer used did
not affect the stereoselectivity of the reaction. At pH-
values >9 the conversion rate decreased dramatically
(Table 1), probably due the formation of insoluble HgO
hydrates.20
No product was detected when the reaction mixture
was extracted with ethyl acetate and analyzed by GC.
Also when the reaction mixture was treated with an
equimolar amount of EDTA before extraction, no
product was detected by GC analysis of the ethyl
acetate phase. This indicates that Hg2+ ion forms a
tight, water-soluble adduct with (4S)-1. When the reac-
The reaction system investigated is rather dilute and
therefore inconvenient for scale-up. However, simply
increasing the concentration of (4S)-1 and Hg(II) led to
decreased selectivity, probably due to the two-phase
system formed. The addition of cosolvents increases the
Table 1. Effect of pH, buffer, Hg(II), and incubation time on the conversion of (4S)-1
Buffer
pH
10 min 30°C
+57 mol% Hg(II)
2 h 30°C
+57 mol% Hg(II)
no Hg(II)
no Hg(II)
Yield
d.e.
(4S)-1
(%)
Yield
(4S)-1
(%)
d.e.
(4S)-1
(%)
Yield
(4S)-1
(%)
d.e.
(4S)-1
(%)
Yield
(4S)-1
(%)
d.e.
(4S)-1
(%)
(4S)-1
(%)
None
Citrate
–
\98
\98
\98
\98
\98
\98
\98
79
53
62
−9
−11
0
1
18
\98
28
15
14
15
15
15
15
15
15
15
15
100
49
86
97
98
100
100
100
100
100
100
100
100
–
–
–
0
0
0
44
\98
\98
27
27
20
19
20
18
18
82
13
51
89
89
98
99
98
100
100
100
100
100
4.0
5.0
6.0
6.0
7.0
8.0
7.0
8.0
9.0
24
35
36
37
38
40
47
55
52
94
94
\98
\98
\98
\98
\98
\98
\98
\98
21
22
20
34
36
20
39
40
38
66
83
KPi
Tris
Glycine
9.0
10.0
11.0
18
18
18
−4

M. J. 6an der Werf et al. / Tetrahedron Letters 42 (2001) 5521–5524
5523
References
solubility of the limonene epoxide but decreases the
solubility of the Hg(II) salt. After much experimenta-
tion, reproducible results on a 3 g scale were obtained
using Hg(OAc)2 in 50% acetone/tris-buffer pH 7.0
during a very short time.21
1. Archelas, A.; Furstoss, R. Ann. Rev. Microbiol. 1997, 51,
491–525.
2. Swaving, J.; de Bont, J. A. M. Enzyme Microb. Technol.
1998, 22, 19–26.
3. Besse, P.; Veschambre, H. Tetrahedron 1994, 50, 8885–
8927.
4. Schurig, V.; Betschinger, F. Chem. Rev. 1992, 92, 873–
888.
5. Imuta, M.; Ziffer, H. J. Org. Chem. 1979, 44, 1351–1352.
6. Archer, I. V. J. Tetrahedron 1997, 53, 15617–15662.
7. Orru¯, R. V. A.; Archelas, A.; Furstoss, R.; Faber, K.
Adv. Biochem. Eng. Biotechnol. 1999, 63, 145–167.
8. Weijers, C. A. G. M.; de Bont, J. A. M. J. Mol. Catal. B.
Enzyme 1999, 6, 199–214.
At pH-values lower than 6, the acid-catalyzed hydroly-
sis of (4S)-1 competes with the formation of the
Hg(II)-(4S)-1 complex, and as a consequence yields
were lower. However, under acidic conditions, the
diaxial (1R,2R,4S)-diol 6 is formed as the major
(>85%) reaction product and the (1R,2S,4S)-cis-
rather than the (1S,2R,4S)-trans-isomer of (4S)-1
(>98% e.e.) remained in 49% yield (=86% of the maxi-
mal theoretical yield). (Table 1). This reaction has
been reported previously.22,23
9. Tokunaga, M.; Larrow, J. F.; Kakiuchi, F.; Jacobsen, E.
N. Science 1997, 277, 936–938.
10. van der Werf, M. J.; Overkamp, K. M.; de Bont, J. A. M.
J. Bacteriol. 1998, 180, 5052–5057.
Also (4R)-1 was stereoselectively converted by both
Hg(II) and low pH. The reaction of Hg(II) with
(4R)-1 resulted in >99% pure (1R,2S,4R)-trans-1,
while the incubation of (4R)-1 in citrate buffer pH 4
resulted in >99% pure (1S,2R,4R)-cis-1.
11. To 2 ml of a freshly prepared 5 mM solution of
diastereomeric (4S)-1 in 50 mM potassium phosphate
buffer (pH 7.0) was added 225 ml of a 25 mM HgCl2
solution in water and the reaction mixture was incubated
for 10 min at 30°C. One ml of EtOAc was added and the
vials were vigorously shaken to accomplish quantitative
extraction of the epoxide trans-(1S,2R,4S)-1 (40% yield).
tR=10.84 min (a-DEX 120, 100°C). Trans-(1R,2S,4R)-1
was prepared in the same way from a diastereomeric
mixture of (4R)-1 (38% yield). tR=10.62 min (a-DEX
120, 100°C). These retention times and the MS-spectra
were the same as those of authentic samples of the
isomers of trans-1.
The stereoselective conversion of 1 was previously
described using different catalysts. The biological
kinetic resolution of 1 has been described using whole
cells of Rhodotorula glutinis24 and limonene 1,2-epox-
ide hydrolyze from Rhodococcus erythropolis.25 With
both epoxide hydrolase activities, the (1S,2R)-isomers
of (4S)-1 and (4R)-1 remained. Davies et al.18 have
reported on the bromine-induced stereoselective
cyclization of (4R)-1, leaving optically pure
(1R,2S,4R)-1 after the addition of 50 mol%
bromine.26 Our method avoids the use of bromine
and carbon tetrachloride and has superior yields. We
have not found any previous reports on the stereose-
lective conversion of epoxides by Hg(II), although the
ring opening of cyclopropanes by Hg(II) is a known
reaction.27
12. To a solution of diastereomeric (4S)-1 (288 mg, 1.83
mmol) in 50 mM tris-HCl buffer (pH 7.0, 250 ml) was
added HgCl2 (41.4 ml of a 25 mM solution in water, 281
mg, 1.04 mmol). After 10 min at 30°C, 50 ml EtOAc was
added and unreacted (1S,2R,4S)-trans-1 was extracted
from the water phase. The water phase was separated and
24 mg of sodium borohydride was added. The solution
was stirred for 21 h, after which the solution was
extracted with 3 portions of 50 ml of CH2Cl2 and there-
after with 3 portions of 50 ml of EtOAc. The collected
organic phases were dried over MgSO4, filtered and evap-
orated till dryness, yielding 180 mg of a solid. The solid
consisted of the (1S,2S,4S)-diol 5 and the (1R,2R,4S)-
diol 6 as a 7:3 mixture. (1S,2S,4S)-diol 5: tR=15.04 min
In conclusion, the method described in this paper in
combination with the published one22,23 provides a
feasible access to all enantiopure stereoisomers of
limonene 1,2-epoxide. The Hg(II) system preferentially
converts the cis-isomers of 1, leaving enantiopure
(1S,2R,4S)-trans-1 and (1R,2S,4R)-trans-1 from (4S)-
1 and (4R)-1, respectively. The complementary, enan-
tiopure (1R,2S,4S)-1 and (1S,2R,4R)-1 (i.e. (4S)-cis-1
and (4R)-cis-1), can be obtained by incubating (4S)-1
or (4R)-1, respectively, at pH values <5.
1
(a-DEX 120, 140°C). H NMR: l 1.20 (s, 3H); 1.72 (br s,
3H); 1.21–2.30 (m, 9H); 3.54 (dd Ja,a=11.8 Hz, Ja,e=4.5
Hz, 1H); 4.69–4.73 (m, 2H). (1R,2R,4S)-diol 6: tR=16.12
min (a-DEX 120, 140°C) and its MS- and 1H NMR
spectra were identical to that of enzymatically prepared
optically pure (1R,2R,4S)-6.25
13. Kitching, W. Organomet. React. 1972, 3, 319–398.
14. Henry, P. M.; Lange, G. L. In Supplement A: The
chemistry of double-bonded functional groups, part 2;
Patai, S., Ed. Oxidation of CꢁC and CꢁN groups; John
Wiley: London, 1977; pp. 965–1083.
15. Houghton, R. P. Metal complexes in organic chemistry;
Cambridge University Press: Cambridge, UK, 1979; pp.
195–201.
Acknowledgements
This work was supported by a grant from the
European Community (BIO4-CT95-0049). We thank
Jan A. M. de Bont and Carel A. G. M. Weijers for
fruitful discussions and Martin de Wit and Henk J.
Swarts for technical assistance.
16. Hartley, F. R. Chem. Rev. 1973, 73, 163–190.

5524
M. J. 6an der Werf et al. / Tetrahedron Letters 42 (2001) 5521–5524
17. Negishi, E.-I. Organometallics in organic synthesis; John
Wiley: New York, 1980; Vol. 1, pp. 455–496.
18. Davies, S. G.; Polywka, M. E. C.; Thomas, S. E. J.
theoretical yield) of unreacted trans-(1S,2R,4S)-1 (d.e.>
98%, [h]2D0=−38.0° (c=3.5, CHCl3); lit.29 −30.4° (c=0.56,
MeOH).
Chem. Soc., Perkin Trans. 1 1986, 1277–1282.
19. Yasui, K; Fugami, K.; Tanaka, S.; Tamaru, Y. J. Org.
Chem. 1995, 60, 1365–1380.
22. Leffingwell, J. C.; Shackelford, R. E. Fr. Demande 1969,
2, 595.
23. Royals, E. E.; Leffingwell, J. C. J. Org. Chem. 1966, 31,
1937–1944.
24. Weijers, C. A. G. M. Tetrahedron: Asymmetry 1997, 8,
639–647.
25. van der Werf, M. J.; Orru¯, R. V. A.; Overkamp, K. M.;
Swarts, H. J.; Osprian, I.; Steinreiber, A.; de Bont, J. A.
M. Appl. Microbiol. Biotechnol. 1999, 52, 380–385.
26. In this paper, the stereochemistry in the drawings is not
in accordance with the text. We have assumed that
(+)-limonene epoxide (i.e. (4R)-1) has been used by the
authors, i.e. the wrong enantiomer has been drawn in
their schemes.
27. Barrett, A. G. M.; Tam, W. J. Org. Chem. 1997, 62,
4653–4664 and references cited therein.
28. van Beek, T. A.; Subrtova, D. Phytochem. Anal. 1995, 6,
1–19.
20. When other salts were tested, the stereoselective conver-
sion of (4S)-1 was only observed when HgO or Hg(OAc)2
were used as reagents. Hg2Cl2, ZnCl2, and CdCl2 were
neither able to accomplish kinetic resolution, nor did they
convert (4S)-1. Also AgNO3, as HgCl2 a salt containing
a metal with a filled outer d orbitals which easily forms
p-complexes with alkenes,16,28 showed no reaction with
(4S)-1.
21. To a solution of 3 g (20 mmol) diastereomeric (4S)-1 in
70 ml acetone was added 70 ml 200 mM Hg(OAc)2 in
tris-HCl buffer pH 7.0. After stirring the reaction mixture
for 20 seconds, 100 ml petroleum ether (40–60) was
added and the mixture was stirred for 1 min. The organic
phase was separated and the water phase was extracted
with 100 ml of petroleum ether. The collected organic
fractions were dried over MgSO4, filtered and evaporated
till dryness, yielding 1.28 g (43%=99% of the maximal
29. Baptistella, L. H. B.; Aleixo, A. M. Liebigs Ann. Chem.
1994, 785–789.
.