Synthesis of enantiopure 9-oxabicyclononanediol derivatives by lipase-catalyzed transformations and determination of their absolute configuration

Article abstract of DOI:10.1002/ejoc.200300783

Mixtures of endo,endo-9-oxabicyclo[4.2.1]nonane-2,5-diol (meso-2) and endo,endo-9-oxabicyclo[3.3.1]nonane-2,6-diol [(±)-3] were prepared from cycloocta-1,5-diene (1) upon 09174874200 treatment with peracids by transannular O-heterocyclization and subsequent saponification of the formed diol monoesters such as (±)-4 and (±)-5. The corresponding diacetates, meso-6 and (±)-7, were formed by acetylation of either meso-2 and (±)-3 or (±)-4 and (±)-5 with acetic anhydride/pyridine. These diacetates were enantioselectively hydrolyzed by microbial enzymes such as the lipases from Candida antarctica (CAL) or Candida rugosa (CRL). The corresponding enantiomers were formed by lipase-catalyzed acetylation of the diols meso-2 and (±)-3 with vinyl acetate. The skeletal isomers can also be separated in this way because the enantiopure monoacetates 4 were formed from the meso-compounds 2 or 6, while one enantiomer of the racemic diacetate (±)-7 [or the diol (±)-3] was transformed into the enantiopure diol 3 (or the enantiopure diacetate 7, respectively) via the corresponding enantiomers of the monoacetate 5. The other enantiomer remained untouched in both cases. The lipases reacted enantioselectively to give the R isomer. Cycloocta-1,5-diene (1) was also used to synthesize 2-oxa-6-thiatricyclo[3.3.1. 13,7]decane-4,8-diol [(±)-11] in a four-step sequence. This racemic diol was also acetylated selectively (R isomer) with vinyl acetate and CRL. Reductive desulfuration of (±)-11 gave exo,exo-9-oxabicyclo[3.3.1]nonane-2,6-diol [(±)-12], which was acetylated selectively (S isomer) with CRL under the same conditions. The similarity in size and particularly in shape is responsible for the observed stereoselectivity of the lipases for the racemic endo,endo compounds (±)-3 and (±)-7 on the one hand and the exo,exo compound (±)-12 on the other hand. The absolute configuration and crystal packing of the products was determined by X-ray structural analysis. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2004.

Full text of DOI:10.1002/ejoc.200300783

FULL PAPER
Synthesis of Enantiopure 9-Oxabicyclononanediol Derivatives by Lipase-
Catalyzed Transformations and Determination of Their Absolute Configuration
Klaus Hegemann,[a] Roland Fröhlich,[a] and Günter Haufe*[a]
Dedicated to Professor Manfred Mühlstädt on the occasion of his 75th birthday
Keywords: Alcohols / Chiral resolution / Diols / Enantioselectivity / Lipases / Polycycles
Mixtures of endo,endo-9-oxabicyclo[4.2.1]nonane-2,5-diol
(meso-2) and endo,endo-9-oxabicyclo[3.3.1]nonane-2,6-diol
[( )-3] were prepared from cycloocta-1,5-diene (1) upon
treatment with peracids by transannular O-heterocyclization
and subsequent saponification of the formed diol monoesters
such as ( )-4 and ( )-5. The corresponding diacetates, meso-
6 and ( )-7, were formed by acetylation of either meso-2 and
( )-3 or ( )-4 and ( )-5 with acetic anhydride/pyridine. These
diacetates were enantioselectively hydrolyzed by microbial
enantiomers of the monoacetate 5. The other enantiomer re-
mained untouched in both cases. The lipases reacted enanti-
oselectively to give the R isomer. Cycloocta-1,5-diene (1) was
also used to synthesize 2-oxa-6-thiatricyclo[3.3.1.13,7]de-
cane-4,8-diol [( )-11] in a four-step sequence. This racemic
diol was also acetylated selectively (R isomer) with vinyl
acetate and CRL. Reductive desulfuration of ( )-11 gave
exo,exo-9-oxabicyclo[3.3.1]nonane-2,6-diol [( )-12], which
was acetylated selectively (S isomer) with CRL under the
enzymes such as the lipases from Candida antarctica (CAL) same conditions. The similarity in size and particularly in
or Candida rugosa (CRL). The corresponding enantiomers
were formed by lipase-catalyzed acetylation of the diols
meso-2 and ( )-3 with vinyl acetate. The skeletal isomers can
also be separated in this way because the enantiopure
monoacetates 4 were formed from the meso-compounds 2 or
6, while one enantiomer of the racemic diacetate ( )-7 [or the
diol ( )-3] was transformed into the enantiopure diol 3 (or the
enantiopure diacetate 7, respectively) via the corresponding
shape is responsible for the observed stereoselectivity of the
lipases for the racemic endo,endo compounds ( )-3 and ( )-
7 on the one hand and the exo,exo compound ( )-12 on the
other hand. The absolute configuration and crystal packing
of the products was determined by X-ray structural analysis.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2004)
Introduction
Thus, treatment of compound 1 with performic acid formed
in situ from formic acid and 30% hydrogen peroxide gave a
38:62 mixture of meso-2 and (Ϯ)-3 in 70% combined
yield,[11,14] whereas a 1:1 mixture of these products was ob-
tained from 1 by treatment with peracetic acid and sub-
sequent saponification, although with a rather low yield of
28%.[13] A third alternative pathway by acid-catalyzed ring
opening of the bis-epoxide of 1 and transannular O-hetero-
cyclization gave a mixture of meso-2 and (Ϯ)-3 in 55% over-
all yield based on 1.[15] Unfortunately, separation of the iso-
meric diols meso-2 and (Ϯ)-3 or the corresponding diacet-
ates meso-6 and (Ϯ)-7 by either crystallization or column
chromatography was very difficult on a preparative scale.
Consequently, we designed a lipase-catalyzed separation:
acetylation with hydrolytic enzymes as catalysts trans-
formed meso-2 into the enantiopure monoester 4. Similarly,
one enantiomer of (Ϯ)-3 should give the corresponding di-
ester 7, while the other should remain untouched. Starting
from the diacetates meso-6 and (Ϯ)-7 the other enantiomers
should be available in each case. Several reviews on lipase-
catalyzed desymmetrizations and kinetic resolutions have
been published.[16Ϫ19]
Small enantiopure molecules, due to their versatile poten-
tial as starting materials in the synthesis of natural prod-
ucts, are favored targets for enzymatic desymmetrization of
prochiral compounds or resolution of racemates.[1Ϫ5] More-
over, such molecules gain interest as auxiliaries or as a
source of ligands for catalytically active metal complexes
for asymmetric synthesis.[6Ϫ10] Consequently, new ways to
synthesize such compounds are desirable. In this context we
became interested in several enantiopure polycyclic diols,
which are available from cycloocta-1,5-diene (1).
There are several alternatives described in the literature
to produce the diols 2 and 3 from the diene 1 by treatment
with peracids and subsequent transformation of the pri-
mary products.[11Ϫ15] The ratio of the two skeletal isomers,
which are formed by transannular O-heterocyclization, can
be influenced slightly by the nature of the applied reagents.
[a]
Organisch-Chemisches Institut, Westfälische Wilhelms-
Universität Münster,
Corrensstr. 40, 48149 Münster, Germany
Fax: ϩ 49-(0)251-83-39772
E-mail: haufe@uni-muenster.de
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Recently, we communicated the application of Candida
rugosa lipase (CRL) for this purpose,[20] and we now report
Eur. J. Org. Chem. 2004, 2181Ϫ2192
DOI: 10.1002/ejoc.200300783
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2181

K. Hegemann, R. Fröhlich, G. Haufe
FULL PAPER
the results with alternative hydrolases, the assignment of the
Subsequently, the activity and selectivity of different li-
absolute configuration of the products and the crystal pack- pases and esterases were tested in the hydrolysis of mixtures
ing of the enantiopure compounds. In forthcoming papers, of diacetates meso-6 and (Ϯ)-7, employed in different ratios
these chiral compounds will be applied as building blocks.
depending on the mode of synthesis (Scheme 3).
Results and Discussion
It has already been mentioned that compounds meso-2
and (Ϯ)-3 can be synthesized by different methods. Accord-
ing to the protocol of Eaton and Millikan[14] we obtained
a 40:60 mixture of the diols meso-2 and (Ϯ)-3 in 65% yield,
while we also improved the yield of Ganter’s[13] procedure
of reaction of 1 with 32% peracetic acid. A 64:36 mixture
of the monoacetates (Ϯ)-4 and (Ϯ)-5 was obtained in 38%
yield (Scheme 1).
Scheme 3
The reaction with porcine pancreatic lipase (PPL)
showed only 6% conversion of the diacetates after two
weeks and this line of investigation was therefore stopped.
Porcine liver acetone powder (PLAP, Sigma, mostly porcine
liver esterase[21,22]) was much more active and showed 53%
conversion of the diacetates after 36 hours (Table 1). How-
ever, only the enantiomeric excess of (Ϫ)-4 (92% ee) was sat-
isfactory.
Scheme 1
Hydrolytic enzymes such as esterases and lipases have
traditionally been employed for ester hydrolysis in buffer
systems. Thus, mixtures of the diols meso-2 and (Ϯ)-3
(40:60) or mixtures of corresponding monoacetates (Ϯ)-4
and (Ϯ)-5 (64:36) were first transformed into the diacetates
meso-6 and (Ϯ)-7 by reaction with acetic anhydride/pyri-
dine (Scheme 2). The mixture of these diacetates, in con-
trast to the diols meso-2 and (Ϯ)-3 or the monoacetates
(Ϯ)-4 and (Ϯ)-5, could be partially separated by repeated
recrystallization[13] or extensive column chromatography.
Hydrolysis of a 64:36 mixture of meso-6 and (Ϯ)-7 with
Candida antarctica lipase B (CAL-B) after 48 hours showed
a 36:9:10:45 mixture of (Ϫ)-4 (80% ee), (ϩ)-5 (94% ee),
meso-6 and (Ϫ)-7 (88% ee). The enantiomeric excesses were
1
determined by H NMR shift experiments using Eu(hfc)3
(see Supporting Information). The minor amounts of diols
that were also formed could not be extracted from the
buffer system under the applied conditions.
The best result with slightly enhanced enantioselectivity
was obtained with Candida rugosa lipase (CRL): after 46
hours a conversion of 52% of the diacetates was reached
and a 42 (94% ee):10 (Ͼ98% ee):9:39 (92% ee) mixture of
the mentioned mono- and diesters was obtained. However,
the separation of the skeletal isomers was not optimal in
these transformations, although quite high enantiomeric ex-
cesses were obtained; the minor amounts of diols formed
could also not be isolated, even by continuous extraction.
This disadvantage could be overcome by the use of or-
ganic solvents (Scheme 4). Moreover, in several cases the
lipases exhibited better enantioselectivity in organic sol-
Scheme 2
Table 1. Results of enzyme-catalyzed hydrolyses of the diacetates meso-6 and (Ϯ)-7
Entry Enzyme
Ratio of starting
materials
Ratio of
products
ee
Combined
yield
Ratio of remaining
substrates
ee
Combined
yield
meso-(6)
(Ϯ)-7
(Ϫ)-4 (ϩ)-5 (Ϫ)-4 (ϩ)-5
(%)
meso-(6)
(Ϫ)-7
(Ϫ)-7
(%)
1
2
3
4
PPL
68
53
64
64
32
47
36
36
n.d.
84
80
n.d.
92
80
6
53
45
52
64
24
18
19
36
76
82
81
n.d.
n.d.
88
93
20
27
26
PLAP
CAL-B
CRL
16
20
20
28
94
Ͼ98
80
94
92
2182
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 2181Ϫ2192

Synthesis of Enantiopure 9-Oxabicyclononanediol Derivatives
FULL PAPER
vents than in aqueous systems.[23Ϫ27] Thus, we first investi- ee) and (ϩ)-5 (60% ee) was isolated with 56% yield; 12% of
gated the transformation of a 40:60 mixture of meso-2 and a 21:79 mixture of the meso-diacetate 6 and the diacetate
(Ϯ)-3 with CAL-B and CRL with different acyl donors in (ϩ)-7 (82% ee) was isolated. Thus, only the diol (Ϫ)-3 was
different solvents (see Supporting Information). The trans- prepared as a pure skeletal isomer in enantiopure form. The
formations were shown to be most efficient and selective in other compounds exhibited low enantiomeric excesses and
pure vinyl acetate. In addition to the two lipases of Candida low chemical purity. This seems to be due to the fact that
sp. the lipase of Pseudomonas cepacia (PCL) was also the monoacetates are acetylated to the corresponding di-
tested, because this lipase has been shown to transform acetates meso-6 or (ϩ)-7 very slowly. The esterification of
other substituted and even heterocyclic alcohols to the cor- the second OH group of 4 and 5 started only when about
responding esters with high enantioselectivity.[28Ϫ31]
60% of the diol mixture had already been monoacetylated.
We therefore did not continue the investigations with this
enzyme.
Next we used CAL-B, immobilized on a porous support
and sold under the trade name Novozym435 by Novo
Nordisk. With vinyl acetate as the acylation reagent and
solvent the acetylation stopped after 22% conversion of the
diols meso-2 and (Ϯ)-3. A 5:1 mixture of tert-butyl methyl
ether (MTBE) and vinyl acetate was found to be more use-
ful. The reaction of a 40:60 mixture of meso-2 and (Ϯ)-3
was stopped after six days. At this stage 57% conversion of
the diols had occurred and a 13:87 mixture of the diols
meso-2 and (Ϫ)-3 (Ͼ98% ee) was isolated from the product
mixture by crystallization. Column chromatography of the
mother liquor gave a 4:1 mixture of (ϩ)-4 (Ͼ98% ee) and
(ϩ)-5 (95% ee) in 38% isolated yield; 18% of (ϩ)-7 (Ͼ98%
ee, contaminated with 1% of meso-6) was also isolated. Al-
though this lipase was able to transform the diols meso-2
and (Ϯ)-3 into the monoacetates with very high enantiose-
lectivity (E value Ͼ 200), the reaction was very slow. Thus,
we decided to apply CRL, which has been used for the es-
terification of secondary alicyclic alcohols, even sterically
demanding ones.[34]
Scheme 4
In the case of the acetylation of a 40:60 mixture of meso-
2 and (Ϯ)-3 the optimal conversion was reached when the
meso-[4.2.1]isomer 2 was quantitatively transformed to the
monoacetate 4 and one enantiomer of the diol (Ϯ)-3 was
completely acetylated to one enantiomer of 5. However, the
Initially, different solvents, different acetylation reagents,
latter monoacetate 5 is only an intermediate, and was such as vinyl acetate, isopropenyl acetate and acetic anhy-
further acetylated to the corresponding diacetate 7. Unfor- dride, as well as different amounts of acetylation reagent
tunately, the monoacetates (Ϯ)-4 and (Ϯ)-5 could not be and lipase were evaluated for the transformation of a 40:60
separated by GC. Thus, it was difficult to follow up the mixture of meso-2 and (Ϯ)-3 (see Supporting Information).
conversion of the mixture of diols meso-2 and (Ϯ)-3; the With solvents such as toluene, n-heptane or cyclohexane the
value of 70% consumption of the diols 2 and 3 was used as reaction stopped after 10Ϫ25% conversion of the diols after
a criterion for the desired conversion. However, the reaction one week. In MTBE in the presence of two, three or four
with PCL was stopped at the point when 66% (GC) of the equivalents of vinyl acetate relative to the total amount of
diols had been consumed after three weeks of reaction the diols meso-2 and (Ϯ)-3, the reaction stopped after 32%,
(Table 2).[32]
53% or 63% conversion, respectively, after one week. Thus,
For the remaining diol (Ϫ)-3 (14% isolated yield, 3% im- we decided to use vinyl acetate as the acylation reagent and
purity of meso-2) an enantiomeric excess of more than 98% as the solvent. In contrast to the reaction with CAL-B, the
was determined by chiral GC (E value Ͼ 200[33]). Moreover, acetylation was very fast in the case of the reaction of a
a 48:52 mixture (1H NMR) of the monoacetates (ϩ)-4 (28% 40:60 mixture of meso-2 and (Ϯ)-3. After six hours, 67% of
Table 2. Results of enzyme-catalyzed acetylations of diols meso-2 and (Ϯ)-3 in vinyl acetate
Entry Enzyme
Ratio of
Conversion Ratio and ee (%) Comb. yield Ratio and ee (%) Comb. yield Ratio and ee (%) of Comb. yield
starting materials of 2 and 3
of products
of 4 and 5
of products
of 6 and 7 remaining substrates
of 2 and 3
meso-(2) (Ϯ)-3
(%)
(ϩ)-4
(ϩ)-5
(%)
meso-6 (ϩ)-7
(%)
meso-(2)
(Ϫ)-3
(%)
1
2
3
PCL
CAL-B
CRL
40
40
40
60
60
60
66
57
67
48 (28)
80 (Ͼ98) 20 (95)
68 (Ͼ98) 32 (80)
52 (60)
56
38
40
21
1
6
79 (82)
99 (Ͼ98)
94 (Ͼ98)
12
18
10
3
13
0
97 (Ͼ98)
87 (Ͼ98)
100 (Ͼ98)
14
23
14
Eur. J. Org. Chem. 2004, 2181Ϫ2192
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2183

K. Hegemann, R. Fröhlich, G. Haufe
FULL PAPER
the total content of the diols had been consumed. After
In order to determine the absolute configuration of the
work up, three fractions of products were isolated; 14% of monoacetate (ϩ)-4 and to separate impurities of (ϩ)-5, sev-
(Ϫ)-3 (Ͼ98% ee) was isolated by crystallization as a pure eral derivatives such as the mesylates, the N-phenylcarbam-
skeletal isomer (E value Ͼ200). This compound was used to ates, the (1S)-camphor-10-sulfonates and the carbamates
determine the absolute configuration (vide infra). A 68:32 with (R)-(ϩ)-phenylethyl isocyanate were synthesized (see
mixture of the monoesters (ϩ)-4 (Ͼ98% ee) and (ϩ)-5 (80% Supporting Information). However, these derivatives could
ee) was isolated in 38% yield by column chromatography. not be isolated as pure skeletal isomers, and therefore were
Additionally, the diacetate (ϩ)-7 (Ͼ98% ee, contaminated not suitable for the determination of the absolute configura-
with 6% of meso-6) was separated in 10% yield.
tion of (ϩ)-4. In a different context, the (1R,2R,5S,6S)-con-
In summary, three practically enantiopure compounds of figuration of (ϩ)-4 was indirectly determined by X-ray
both enantiomeric series were produced either by lipase-cat- structural analysis of a tricyclic 1,4-dithiepane derivative
alyzed hydrolysis of the mixture of diacetates meso-6 and produced from (ϩ)-4 by selective oxidation of the hydroxy
(Ϯ)-7 or by acetylation of the mixture of diols meso-2 and group and subsequent reaction of the formed ketone with
(Ϯ)-3. However, the absolute configuration of these com- propane-1,3-dithiol.[41]
pounds remained to be determined.
The X-ray analysis of (1S,2S,5S,6S)-(Ϫ)-3 allowed inter-
From the literature[35Ϫ37] it is known that CRL according esting insights into the crystal packing of this compound.
to Kazlauskas’ rule prefers the enantiomer shown in Fig- The elemental cell along the y-axis (Figure 3) shows a layer
˚
ure 1. This is the (R)-enantiomer in cases where the larger structure. The diameter of one layer is about 4.2 A. The
substituent has also the higher priority according to the layers are linked to each other with hydrogen bonds be-
CahnϪIngoldϪPrelog (CIP) rule.[38,39] This is true for the tween the OH groups. The OH···O contact is about 1.91 A,
˚
˚
acetylation of meso-2 and (Ϯ)-3, and was observed earlier the O···O distance is about 2.74 A and the OϪHϪO angle
also in hydrolysis of endo,endo-2,6-diacetoxybicyclo- is about 174°.
[3.3.1]nonane.[40]
Figure 1. Enantiomer preferably transferred by CRL according to
Kazlauskas’ rule[36,37] (adapted to the investigated diols)
In order to determine the absolute configuration of the
products the diol (Ϫ)-3, which remained unchanged in the
CRL-catalyzed acetylation of the mixture of the diols meso-
2 and (Ϯ)-3, was recrystallized and excellent crystals were
grown for X-ray structural analysis. Due to the very small
error in the enantiopole parameter, the absolute configura-
tion of (Ϫ)-3 could be determined directly to be
(1S,2S,5S,6S)-(Ϫ)-9-oxabicyclo[3.3.1]nonane-2,6-diol (Fig-
ure 2). Consequently, the absolute configuration of the
monoacetylation product must be (1R,2R,5R,6R)-(ϩ)-5
and that of the diacetate must be (1R,2R,5R,6R)-(ϩ)-7.
Since in all acetylation experiments with CRL, CAL-B or
PCL the same sign of optical rotation was found for the
corresponding compounds, all these lipases must be (R)-
selective with these substrates.
Figure 3. View of the elemental unit of (1S,2S,5S,6S)-(Ϫ)-9-oxa-
bicyclo[3.3.1]nonane-2,6-diol (Ϫ)-3 along the y-axis
The view in the direction of the z-axis (Figure 4) shows
a supramolecular structure with two types of channels. The
Figure 2. Stereoview of (1S,2S,5S,6S)-(Ϫ)-9-oxabicyclo[3.3.1]non-
ane-2,6-diol [(Ϫ)-3]
Figure 4. View of the elemental unit of (1S,2S,5S,6S)-(Ϫ)-9-oxa-
bicyclo[3.3.1]nonane-2,6-diol (Ϫ)-3 along the z-axis
2184
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 2181Ϫ2192

Synthesis of Enantiopure 9-Oxabicyclononanediol Derivatives
FULL PAPER
non-polar type (center) is coated with the equatorial pro- at 120 °C in an autoclave for 96 hours[59] gave 40% of the
tons at C-3 and C-7. In the second, more polar and bigger diene (Ϯ)-9. By modification of a protocol of Stetter and
(8.8 A in diameter) channel, OH groups are involved. It is Heckel,[60,61] we obtained the diene (Ϯ)-9 as a crystalline
˚
noteworthy that a 43 helical axis causes a permutation of compound in 88% yield by refluxing the diiodide (Ϯ)-8 with
one upon the other molecules.
dicyclohexylmethylamine.
A similar observation has also been made for racemic
exo,exo-2,6-dimethyl-9-oxabicyclo[3.3.1]nonane-2,6-
X-ray analysis of this diene (Figure 6) showed that the
double bonds can approach each other and therefore they
diol.[42Ϫ44] For this compound, which has an opposite are suitable for another transannular heterocyclization.
orientation of the OH groups compared to 3, additional
hydrogen bonds were observed.[43] Moreover, the channels
in the crystalline state were shown to be much bigger, and
hence inclusion compounds were formed with a variety of
small organic molecules.[44] Because of the smaller cavities,
the endo,endo-diol (Ϫ)-3 could not be co-crystallized with
chiral molecules and hence could not be used for racemate
resolution. In order to elucidate whether bigger cavities
could be obtained from the diastereomer, we next synthe-
sized exo,exo-9-oxabicyclo[3.3.1]nonane-2,6-diol (12) in
enantiopure form.
Figure 6. Stereoview of (Ϯ)-9-oxabicyclo[3.3.1]nona-2,6-diene
This compound cannot be synthesized by an SN2 reac- [(Ϯ)-9]
tion, for example from the tosylate or mesylate of (Ϫ)-3,
for steric reasons. However, for the racemic compound a
Such a transannular S-heterocyclization was reported al-
synthesis has been reported.[13] We modified and optimized
this synthetic sequence. First endo,endo-2,6-diiodo-9-oxa-
bicyclo-[3.3.1]nonane [(Ϯ)-8] was synthesized from diene 1,
most simultaneously by Stetter et al.[59] and by Ganter and
Wicker more than three decades ago.[62] According to the
published procedure,[59] treatment of the diene (Ϯ)-9 with
according to our procedure,[45] by transannular O-hetero-
dichlorosulfane gave, after chromatographic purification,
cyclization with N-iodosuccinimide (NIS) and methanol in
90% isolated yield as a pure skeletal isomer (Scheme 5).
Other protocols for transannular O-heterocyclization reac-
4,8-dichloro-2-oxa-6-thiaadamantane [(Ϯ)-10] as a single
diastereomer in 45% yield via the episulfonium ion I
(Scheme 6).
tions using different electrophiles have, in most cases, led to
mixtures of skeletal isomers.[46Ϫ58]
Scheme 6
Scheme 5
The X-ray structure of compound (Ϯ)-10 (see Supporting
Information) shows that nucleophilic substitution of chlor-
X-ray analysis (Figure 5) shows that any SN2 replacement
ine by an oxygen function is favored by anchimeric assist-
of iodine is impossible. From the drawing it becomes obvi-
ance of sulfur, again giving the episulfonium ion I, which,
ous as well that HI elimination, which must be the next step
in the synthesis, will need a conformational change because
in turn, can be opened by an oxygen nucleophile. Thus, re-
fluxing of compound (Ϯ)-10 with concentrated aqueous so-
the anti-periplanar arrangement of H and I necessary for
dium carbonate[62] gave the diol (Ϯ)-11 in 59% isolated
an E2 process is not possible in the most stable chair-chair
conformation of compound (Ϯ)-8.
yield (Scheme 7).
However, such an arrangement is possible in a boat con-
formation. Thus, treatment of (Ϯ)-8 with KOH in ethanol
Scheme 7
An alternative two-step procedure, involving refluxing of
the dichloride (Ϯ)-10 with potassium acetate in DMF and
subsequent saponification of the formed diacetate, did not
improve the yield of (Ϯ)-11. The relative configuration of
Figure 5. Stereoview of endo,endo-(Ϯ)-2,6-diiodo-9-oxabicyclo-
[3.3.1]nonane [(Ϯ)-8]
Eur. J. Org. Chem. 2004, 2181Ϫ2192
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2185

K. Hegemann, R. Fröhlich, G. Haufe
FULL PAPER
product (Ϯ)-11 was proven by X-ray structural analysis (see
Supporting Information).
The synthesis of compound (Ϯ)-12 was completed by re-
ductive desulfuration with Raney-nickel, which gave the de-
sired exo,exo-9-oxabicyclo[3.3.1]nonane-2,6-diol [(Ϯ)-12] in
63% yield. This diol was contaminated with 5% of the
endo,exo-diol showing that the nucleophilic substitution of
chlorine with oxygen was not completely diastereoselective
(see Supporting Information). The corresponding di-
hydroxyoxathiaadamantane could not be isolated from the
reaction mixture of the nucleophilic substitution step.
With the racemic diols (Ϯ)-11 and (Ϯ)-12 in hand, we
now investigated the lipase-catalyzed acetylation in order to
compare the enantioselectivity of the lipase of Candida ru-
gosa towards these compounds. Comparing the molecular
geometry of the corresponding enantiopure diols 3 and 11
(cf. Figure 10) their similarity in size becomes obvious.
Thus, we also expected (R)-selectivity of the CRL (see dis-
cussion in ref.[20]).
Figure 7. Stereoview of (1S,2S,4S,5S,7S,8S)-(ϩ)-2-oxa-6-thiaadam-
antane-4,8-diol (ϩ)-(11)
[2.2.2]alkanols bearing an OH group in the exo-position are
not acetylated with CRL. In our case, under the conditions
mentioned for the reactions of the diols (Ϯ)-3 and (Ϯ)-11,
one enantiomer of the exo,exo-diol (Ϯ)-12 was acetylated,
although with a lower reaction rate than the other above-
mentioned diols. After a reaction time of eight hours 50%
conversion of the racemate was detected and three com-
pounds were found in the reaction mixture (Scheme 9).
This lipase reacted very fast with the racemic tricyclic
diol (Ϯ)-11 to give the monoacetate (Ϫ)-13 and the diacet-
ate (Ϫ)-14: after four hours 56% of racemic (Ϯ)-11 had al-
ready been consumed (Scheme 8).
Scheme 9
This mixture was separated by column chromatography.
The diol (ϩ)-12 (94% ee, E value 115) eluted as the last
compound from the column and was isolated in 23% yield.
For the monoacetate (Ϫ)-15, isolated in 28% yield, a quite
low enantiopurity of only 30% ee was detected. The diacet-
ate (Ϫ)-16 (92% ee) was isolated in 17% yield.
The diol (ϩ)-12 was recrystallized from acetone and
single crystals were obtained. X-ray crystallography showed
a tetragonal space group P41212 (Figure 8). No big cavities
were found in the packing scheme of this compound (see
Supporting Information).
Scheme 8
The remaining diol (ϩ)-11 (31% isolated yield) had more
than 98% ee (E value 34). A second fraction isolated in 17%
yield by column chromatography contained the monoacet-
ate (Ϫ)-13 (34% ee). The third fraction, the diacetate (Ϫ)-
14 (92% ee), was isolated in 31% yield. Single crystals of
the diol (ϩ)-11 were grown and an X-ray crystal structural
analysis was performed. The presence of the sulfur atom
allowed us to determine the absolute configuration of this
compound to be (1S,2S,4S,5S,7S,8S)-(ϩ)-2-oxa-6-thiaada-
mantane-4,8-diol [(ϩ)-11] (Figure 7). Thus, (R)-selectivity
of the lipase was also observed for this tricyclic heterocycle.
Compound (1S,2S,4S,5S,7S,8S)-(ϩ)-(11) crystallized in
the orthorhombic space group P212121, with hydrogen
bonds between the hydroxyl groups attached to C-4 and C-
8 contributing to the high symmetry of the crystal system;
large channels were not found in the crystal packing.
Next we investigated the CRL-catalyzed acetylation of
the C2-symmetric exo,exo-diol (Ϯ)-12. According to the in-
Figure 8. Stereoview of (1S,2R,5S,6R)-(ϩ)-9-oxabicyclo[3.3.1]non-
vestigations of Griengl et al.[63] bicyclo[2.2.1]- and bicyclo- ane-2,6-diol [(ϩ)-12]
2186
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 2181Ϫ2192

Synthesis of Enantiopure 9-Oxabicyclononanediol Derivatives
The assignment of the absolute configuration was not
FULL PAPER
As mentioned above, the absolute configuration of the
possible in a direct way due to the absence of a heavy atom endo,endo-monoacetate (ϩ)-4 — the product of acetylation
and a too big error of the enantiopole parameter, but was of meso-2 by CRL — could not be determined directly. Ac-
successful indirectly. Reductive desulfuration of the enanti- cording to Kazlauskas’ rule, this compound with CRL/vinyl
opure (1S,2S,4S,5S,7S,8S)-(ϩ)-2-oxa-6-thia-adamantane- acetate should be acetylated at the (R)-center. This assump-
4,8-diol [(ϩ)-11] with Raney nickel gave enantiopure tion is supported by the following consideration: Compari-
(1S,2R,5S,6R)-(ϩ)-9-oxabicyclo[3.3.1]nonane-2,6-diol [(ϩ)- son of the projections of the (R)-center and the (S)-center
12] (Scheme 10).
of meso-2 with the (R)-center of the diol (R,R)-3 suggests
strongly that the environments of the (R)-centers of 2 and
3 are much more similar (Figure 10). By comparison of
these two projections the (R)-center of 2 should also be
transformed selectively. As mentioned above, this assump-
tion was proven by X-ray structural analysis of a derivative
obtained after Jones oxidation of (2R,5S)-(ϩ)-4 and sub-
sequent treatment of the formed ketone with 1,3-pro-
panethiol.[41]
Scheme 10
This product exhibited the same sign of optical rotation
as the non-acetylated product in the kinetic resolution of
(Ϯ)-12 with CRL. As expected, the value of the specific
rotation was slightly bigger ([α]2D0 ϭ ϩ25.4), due to the
higher enantiomeric excess (Ͼ98% ee) in relation to ϩ24.5
(94% ee) obtained for the diol (ϩ)-12 not acetylated in the
resolution reaction. Thus, CRL acetylates the exo,exo-diol
(Ϯ)-12 (S)-selectively.
Comparing the acetylation of the three C2-symmetric di-
ols 3, 10 and 12 it is noteworthy to mention that all com-
pounds were transformed with very high enantioselectivity;
endo,endo-2,6-diacetoxybicyclo[3.3.1]nonane can also be
hydrolyzed with high enantioselectivity.[40] In order to find
an explanation for the opposite stereoselectivity of the li-
pase in the transformation of exo,exo-diol (Ϯ)-12, in ad-
dition to the X-ray structures, the most stable confor-
mations of these compounds were calculated (AM1) to ex-
Figure 10. Comparison of the projections of the (5S)- and the (2R)-
centers of meso-2 with (2R,6R)-3
Conclusion
A very simple one-pot reaction of the bulk chemical
clude steric distortions caused by packing effects. The calcu- cycloocta-1,5-diene (1) with peracids gave mixtures of
lated conformations did not differ from the structures found endo,endo-9-oxabicyclo[4.2.1]nonane-2,5-diol (2) and
in the crystal (Figure 9). endo,endo-9-oxabicyclo[3.3.1]nonane-2,6-diol [(Ϯ)-3] by
While the (R,R)-enantiomers of the endo,endo-diol (Ϯ)-3 transannular O-heterocyclization and hydrolysis. Applying
and the tricyclic diol (Ϯ)-11 were acetylated, as expected (R)-selective lipases, simultaneous separation of the skeletal
from Kazlauskas’ rule,[36] the enzyme prefers the (S,S)-en- isomers, desymmetrization of meso-2 and kinetic resolution
antiomer of exo,exo-diol (Ϯ)-12. A comparison of the mo- of (Ϯ)-3, three practically enantiopure compounds, namely
lecular geometry of all the faster-acetylated enantiomers (1S,2S,5S,6S)-(Ϫ)-9-oxabicyclo[3.3.1]nonane-2,6-diol (Ϫ)-
indicates that the shape of these molecules is very similar 3, (1R,2R,5S,6S)-(ϩ)-2-acetoxy-9-oxabicyclo[4.2.1]nonan-
(Figure 9); the (R)-selectivity in the CRL-catalyzed hydroly- 5-ol (ϩ)-4 [contaminated with 20% of (ϩ)-5] and
sis of the diacetates of racemic bicyclo[3.3.1]nonane-2,6- (1R,2R,5R,6R)-(ϩ)-2,6-diacetoxy-9-oxabicyclo[3.3.1]-
diol[40] also fits into this picture.
nonane (ϩ)-7 were isolated. The enantiomers of these com-
Figure 9. View of the four C2-symmetric substrates preferentially transformed by the lipase of Candida rugosa (CRL): (a) (1R,2R,5R,6R)-
(ϩ)-9-oxabicyclo[3.3.1]nonane-2,6-diol [(ϩ)-3], (b) (1R,3R,4R,5R,7R,8R)-(Ϫ)-2-oxa-6-thiaadamantane-4,8-diol [(Ϫ)-11], (c)
(1R,2S,5R,6S)-(Ϫ)-9-oxabicyclo[3.3.1]nonane-2,6-diol [(Ϫ)-12], (d) (1S,2R,5S,6R)-(ϩ)-2,6-diacetoxybicyclo[3.3.1]nonane
Eur. J. Org. Chem. 2004, 2181Ϫ2192
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2187

K. Hegemann, R. Fröhlich, G. Haufe
FULL PAPER
Complete spectroscopic data of all prepared compounds are given
in the Supporting Information.
pounds were available using the same lipases to catalyze the
hydrolysis of the corresponding diacetates meso-6 and (Ϯ)-
7. Candida rugosa lipase (CRL) was shown to be the most
efficient for both types of reaction. Analogously, 2-oxa-6-
thiatricyclo[3.3.1.13,7]decane-4,8-diol [(Ϯ)-11] and exo,exo-
9-oxabicyclo[3.3.1]nonane-2,6-diol [(Ϯ)-12] were resolved
by CRL-catalyzed acetylation. While the diols meso-2, (Ϯ)-
3 and (Ϯ)-11 or the diacetates (Ϯ)-5 and meso-6 were trans-
formed (R)-selectively, the exo,exo-diol (Ϯ)-12 was acetyl-
ated with (S)-selectivity by the same lipase under similar
conditions. The similarity of the shape of the molecules is
responsible for the observed stereoselectivity. The absolute
configuration of the products and crystal packing were de-
termined by X-ray crystallography. Forthcoming papers will
describe synthetic applications of the thus-formed enanti-
opure building blocks.
Reaction of Cycloocta-1,5-diene (1) with Peracetic Acid: Similar to
a known procedure,[13] a 32% peracetic acid solution in acetic acid
(58.9 g, 242 mmol) was cooled to Ϫ5 °C and stirred vigorously. A
solution of cycloocta-1,5-diene (1; 11.5 g, 106 mmol) in acetic acid
(22 mL) was added dropwise over a period of 2 hours, and the
temperature was kept between Ϫ5 °C and 0 °C. This mixture was
stirred at this temperature for two more hours and then allowed to
warm up slowly, while stirring was continued for 18 hours. The
reaction mixture was then treated with 10% aqueous sodium hy-
droxide until pH 7 was reached. Subsequently, this mixture was
extracted continuously with ethyl acetate (200 mL) for 15 hours.
The organic layer was stirred with a solution of FeSO4 (50 g) in
brine (100 mL) and dried over MgSO4. After evaporation of the
solvent a 64:36 mixture of the monoacetates (Ϯ)-4 and (Ϯ)-5 was
isolated as a yellowish oil. Yield: 8.12 g (38%). The product ratio
varied between 64:36 and 55:45 depending on the reaction tempera-
ture and work-up conditions. The mixture could not be separated
by column chromatography. The spectroscopic data agree with
published values.[13]
Experimental Section
endo,endo-2,5-Diacetoxy-9-oxabicyclo[4.2.1]nonane (meso-6) and
endo,endo-2,6-Diacetoxy-9-oxabicyclo[3.3.1]nonane [(؎)-7]: The
General Remarks: 1H NMR (300.1 MHz) and 13C NMR
(75.5 MHz), if not otherwise stated, were recorded from ca. 20%
solutions in CDCl3 on a 300-MHz spectrometer. Chemical shifts
64:36
mixture
of
(Ϯ)-endo,endo-2-acetoxy-9-oxabicyclo-
[4.2.1]nonan-5-ol (meso-4) and (Ϯ)-endo,endo-2-acetoxy-9-oxabicy-
clo[3.3.1]nonan-6-ol [(Ϯ)-5] (4.51 g, 22.6 mmol), was stirred with
acetic anhydride (10.2 g, 100 mmol) and pyridine (9.5 g, 120 mmol)
at room temperature for 18 hours. The solvents were then evapo-
rated in vacuo to dryness. The brownish solid was filtered through
a silica gel column (cyclohexane/acetone, 2:1). After evaporation of
the solvent, a colorless solid was isolated. Yield: 4.98 g (92%). M.p.
65 °C. The spectroscopic data agree with published values.[13]
1
are reported as δ values (ppm) relative to TMS (for H as internal
standard; for 13C, CDCl3, δ ϭ 77.0 ppm was used as internal stand-
ard). The multiplicity of the 13C signals was determined by the
DEPT technique. Mass spectra (electron impact ionization, 70 eV)
were recorded by GC/MS coupling. The conversion of substrates
during enzymatic transformations was followed by GC on quartz
capillary columns (25 m ϫ 0.33 mm, 0.52 µm HP-5 and 30 m ϫ
0.32 mm, 0.25 µm SPB-1, temperature program, 50 °C Ǟ 180 °C
with 5 °C/min then from 180 °C Ǟ 280 °C with 30 °C/min heating
rates with N2 as the carrier gas). The ratio of compounds was deter-
mined by integration of the peak area and was not corrected. The
products of enzymatic transformations were separated by column
chromatography (silica gel, Merck 60, 70Ϫ230 mesh). The solvents
are mentioned in the corresponding procedures. The enantiomeric
excesses of the diols were determined by chiral GC on a β-cyclodex-
trin column (30 m ϫ 0.25 mm, 0.25 µm, Beta-DexTM 120, isotherm
170 °C for compounds 3 and 12, and 190 °C for compound 11; N2
as carrier gas). The enantiomeric excesses of the mono- and diacet-
Lipase-Catalyzed Hydrolysis of the Mixture of Diacetates meso-6
and (؎)-7: A suspension of a 64:36 mixture of the diacetates meso-
6 and (Ϯ)-7 (480 mg, 2 mmol) in 0.1  phosphate buffer (pH 7.1,
45 mL) was treated with 90 mg of CRL and stirred at room tem-
perature for 48 hours. During that time, the liberated acetic acid
was neutralized by addition of 0.1  NaOH. After termination of
the reaction by addition of solid NaCl until saturation, the reaction
mixture was continuously extracted with diethyl ether (50 mL) for
10 hours. The organic layer was separated, dried over MgSO4 and
the solvent was evaporated. The residue was subsequently sepa-
rated by column chromatography (cyclohexane/ethyl acetate, 2:1).
In this way a 19:81 mixture of the diacetates meso-6 and (Ϫ)-7
(127 mg, 26%, 92% ee, determined by chiral GC after saponifi-
cation) was isolated. Furthermore, an 80:20 mixture (207 mg, 52%)
of the monoacetates (Ϫ)-4 (94% ee, 1H shift NMR) and (Ϫ)-5
1
ates were determined by H NMR spectroscopy using Eu(hfc)3 as
a chiral shift-reagent. Optical rotations were measured with a
polarimeter (NaD line, λ ϭ 589 nm). Microanalyses were carried
out by the ‘‘Mikroanalytisches Laboratorium, Organische
Chemie’’, University of Münster using a CHN-O analyzer. X-ray
data sets were collected with CAD4 diffractometers. Programs
used: data collection EXPRESS (Nonius B.V., 1994), data re-
duction MolEN (K. Fair, EnrafϪNonius B.V., 1990), structure
solution SHELXS-97,[64] structure refinement SHELXL-97,[65]
graphics SCHAKAL.[66]
1
(Ͼ98% ee, H shift NMR) was isolated.
The hydrolyses with the other lipases were done analogously. The
results are collected in Table 1.
Lipase-Catalyzed Acetylation of a Mixture of Diols meso-2 and (؎)-
3: A 40:60 mixture of the diols meso-2 and (Ϯ)-3 (10.0 g,
63.3 mmol), prepared according to ref.[14] was dissolved in vinyl
acetate (250 mL), CRL (2.5 g) was added and the mixture was
stirred for 6 hours, after which time 67% of the diol mixture had
been consumed. The progress of acetylation was followed by GC.
The lipase was then filtered off by passing the reaction mixture
through a silica gel column, which was subsequently eluted with
acetone (300 mL). The organic layers were combined and the sol-
vent was evaporated in vacuo. The crude product was dissolved in
ethyl acetate (60 mL) and subsequently n-pentane was added whilst
Chemicals were purchased from Acros, Aldrich, Fluka, Sigma or
Merck. Cycloocta-1,5-diene was kindly donated by Degussa-Hüls,
Marl, Germany. The lipase from Pseudomonas cepacia (Amano PS)
was a kind gift of Amano Pharmaceutical Co., Ltd., Nagoya, Ja-
pan. The lipase Candida antarctica (Novozym435) was kindly do-
nated by Novo Nordisk A/S, Copenhagen, Denmark. The lipase of
Candida rugosa and porcine liver esterase (acetone powder, PLAP)
were purchased from Sigma. According to a given protocol,[14]
a
40:60 mixture of the diols 2 and 3 (31.0 g, 65%) was synthesized.
2188
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 2181Ϫ2192

Synthesis of Enantiopure 9-Oxabicyclononanediol Derivatives
FULL PAPER
stirring until all unchanged diol (Ϫ)-3 precipitated. The solid was
separated as an analytically pure white powder by filtration. A frac-
tion of the diol was recrystallized from acetone for X-ray analysis.
data agree with published values.[60] Single crystals for X-ray struc-
tural analysis were grown by careful recrystallization from n-pen-
tane.
4,8-Dichloro-2-oxa-6-thiatricyclo[3.3.1.13,7]decane [(؎)-10]: Synthe-
sized according to
(1S,2S,5S,6S)-(؊)-9-Oxabicyclo[3.3.1]nonane-2,6-diol [(؊)-3]: Yield
1.44 g (48% of the theoretically possible amount). M.p. 84 °C (M.p.
of the racemate 127Ϫ128 °C[13]). [α]2D0 ϭ Ϫ45.8 (c ϭ 1.0, ethanol);
Ͼ98% ee (chiral GC). 13C NMR (CDCl3/CD3OD): δ ϭ 21.8 (t),
28.1 (t), 68.0 (d), 69.9 (d) ppm. The 1H NMR (100 MHz) and mass
spectra of the racemate are known.[13]
The solvent was evaporated from the mother liquor and the residue
was separated by column chromatography (cyclohexane/ethyl acet-
ate, 1:1) to give the inseparable monoacetates (ϩ)-4 and (ϩ)-5 as a
light-yellowish oil (5.10 g, 68:32 mixture) and the diacetate (ϩ)-7
as a sweet-smelling, colorless solid.
a
known procedure[59] from 9-oxabicy-
clo[3.3.1]nona-2,6-diene [(Ϯ)-9] (12.5 g, 0.102 mol) and SCl2
(10.5 g, 0.102 mol) in dry CH2Cl2 (250 mL). Purified by column
chromatography (cyclohexane/ethyl acetate, 1:1). Yield 10.0 g
(43%). M.p. 181 °C (cyclohexane/ethyl acetate), (ref.[59] M.p.
184Ϫ185 °C, in a sealed capillary after sublimation). A 100 MHz
1H NMR spectrum of this compound has been published.[62] 13
C
NMR: δ ϭ 32.3 (t), 36.3 (d), 60.9 (d), 71.0 (d) ppm. MS (GC/MS):
m/z (%) ϭ 227/225/223 (16:94:94), 190/188 (48:100), 157 (16), 155
(49), 145 (10), 125 (9), 117 (14), 115 (30), 97 (10), 91 (26), 85 (16).
Single crystals were grown for X-ray structural analysis by recrys-
tallization from acetone.
(1R,2R,5S,6S)-(؉)-2-Acetoxy-9-oxabicyclo[4.2.1]nonan-5-ol [(؉)-
4]: Ͼ98% ee (1H shift NMR, 100 mol % Eu(hfc)3]. 13C NMR: δ ϭ
21.0 (q), 23.6 (t), 25.2 (t), 25.4 (t), 28.0 (t), 66.4 (d), 71.6 (d), 73.6
(d), 77.5 (d), 170.2 (s) ppm. The 1H NMR (100 MHz) and mass
spectra of the racemate are known.[13]
2-Oxa-6-thiatricyclo[3.3.1.13,7]decane-4,8-diol [(؎)-11]: Similar to
ref.[62] a mixture of 4,8-dichloro-2-oxa-6-thiatricyclo[3.3.1.13,7]de-
cane [(Ϯ)-10] (8.5 g, 36.5 mmol), Na2CO3 (15.0 g, 141.5 mmol) and
water (50 mL) was refluxed for 12 hours. The cold mixture was
then neutralized with 2  aqueous HCl and lyophilized. The crude
solid was extracted with ethyl acetate (100 mL), the extract was
dried over MgSO4 and concentrated to about 10 mL. This solution
was purified by column chromatography (ethyl acetate) to give the
diol (Ϯ)-11 as a yellowish solid. Yield 4.07 g (59%). Single crystals
were grown for X-ray structural analysis by recrystallization from
ethyl acetate. M.p. 312 °C (ethyl acetate, ref.[62] M.p. 316Ϫ317 °C).
1H NMR (CD3OD): δ ϭ 1.99 (dm, 2J ϭ 12.8 Hz, 2 H), 2.72Ϫ2.78
(m, 2 H), 2.88 (dm, 2J ϭ 12.8 Hz, 2 H), 3.79Ϫ3.88 (m, 2 H),
3.95Ϫ4.03 (m, 2 H) ppm. A 100 MHz 1H NMR spectrum in CDCl3
has been published,[62] but the assignment of the signals seems not
to be conclusive (cf. Supporting Information). 13C NMR
(CD3OD): δ ϭ 32.3 (t), 36.8 (d), 71.1 (d), 72.4 (d) ppm. MS (GC/
MS): m/z (%) ϭ 187 (100) [Mϩ Ϫ 1], 171 (38), 155 (23), 137 (46),
131 (18), 113 (34), 109 (22), 97 (93), 85 (46), 81 (51), 69 (38).
(1R,2R,5R,6R)-(؉)-2-Acetoxy-9-oxabicyclo[3.3.1]nonan-6-ol [(؉)-
5]: Ͼ80% ee [1H shift NMR, 100 mol % Eu(hfc)3]. 13C NMR: δ ϭ
21.0 (q), 21.5 (t), 22.8 (t), 28.2 (t), 30.7 (t), 68.1 (d), 69.6 (d), 70.6
(d), 81.5 (d), 170.2 (s) ppm. The 1H NMR (100 MHz) and mass
spectra of the racemate are known.[13]
(1R,2R,5R,6R)-(؉)-2,6-Diacetoxy-9-oxabicyclo[3.3.1]nonane [(؉)-
7]: Yield 1.53 g (31% of theoretically possible amount, contami-
nated with 6% of meso-6). M.p. 110 °C (cyclohexane/ethyl acetate;
m.p. of the racemate 112Ϫ113 °C[13]). [α]2D0 ϭ ϩ50.2 (c ϭ 1.0, ethyl
acetate); Ͼ98% ee (determined by chiral GC after saponification),
corrected value of specific rotation for skeletal pure (ϩ)-7, [α]2D0
ϭ
ϩ53.4. 13C NMR: δ ϭ 20.7 (q), 22.5 (t), 24.8 (t), 66.1 (d), 70.0 (d),
1
169.8 (s) ppm. The H NMR (100 MHz) and mass spectra of the
racemate are known.[13]
(1S,2S,5S,6S)-(؊)-2,6-Diacetoxy-9-oxabicyclo[3.3.1]nonane [(؊)-7]:
Enantiopure (1S,2S,5S,6S)-(Ϫ)-3 (0.31 g, 2 mmol) was acetylated
with acetic anhydride (1.02 g, 10 mmol) and pyridine (0.95 g,
12 mmol) for 18 hours at room temperature. After removal of the
volatile material in vacuo, the solid residue was purified by column
chromatography (silica gel; cyclohexane/ethyl acetate, 1:1). Yield
exo,exo-9-Oxabicyclo[3.3.1]nonane-2,6-diol [(؎)-12]: Similar to
ref.[13]
2-oxa-6-thiatricyclo-[3.3.1.13,7]decane-4,8-diol
[(Ϯ)-11]
(6.7 g, 35.6 mmol) in dry ethanol (250 mL) was treated with Raney-
nickel (25 g) and refluxed for 24 hours. After cooling to room tem-
perature, the solid was removed by filtration, the filter was washed
with ethanol (50 mL), the layers were combined and the solvent
was evaporated in vacuo. The remaining white solid was purified
by column chromatography (acetone) to give the diol (Ϯ)-12 as the
major product. Yield 3.51 g (63%). M.p. 91 °C (acetone, ref.[13]
M.p. 85Ϫ86 °C after sublimation). 13C NMR ([D6]DMSO): δ ϭ
21.7 (t), 25.1 (t), 67.3 (d), 72.2 (d) ppm. 60 MHz 1H NMR and
mass spectra of this compound are known.[13] Additionally 180 mg
(3.4%) of endo,exo-9-oxabicyclo[3.3.1]nonane-2,6-diol was also iso-
lated (for spectroscopic data see Supporting Information).
0.44 g (93%). M.p. 110 °C (cyclohexane/ethyl acetate). [α]2D0
ϭ
Ϫ53.7 (c ϭ 1.0, ethyl acetate); Ͼ98% ee (chiral GC, after saponifi-
cation).
endo,endo-2,6-Diiodo-9-oxabicyclo[3.3.1]nonane [(؎)-8]: According
to ref.[45], (Ϯ)-8 was prepared as white crystals from cycloocta-
1,5-diene (1) (10.80 g, 0.1 mol) and N-iodosuccinimide (35.8 g, 0.16
mol) in methanol after recrystallization from methanol. Yield
27.2 g (90%). M.p. 123Ϫ124 °C (ref.[46] M.p. 124 °C). The spectro-
scopic data agree with published values.[45]
Kinetic Resolution of 2-Oxa-6-thiatricyclo[3.3.1.13,7]decane-4,8-diol
[(؎)-11]: 2-Oxa-6-thiatricyclo[3.3.1.13,7]decane-4,8-diol [(Ϯ)-11]
(1.20 g, 6.4 mmol) was partially dissolved in vinyl acetate (45 mL)
and CRL (0.72 g) was added. The mixture was stirred at room tem-
perature for 4 hours. The lipase was then separated by filtration
through a short silica-gel column and the column was eluted with
acetone (200 mL) and ethanol (200 mL). GC analysis of the prod-
uct mixture showed 45% remaining diol (Ϯ)-11. After evaporation
of the solvent the residue was separated by column chromatography
(ethyl acetate).
9-Oxabicyclo[3.3.1]nona-2,6-diene [(؎)-9]: endo,endo-2,6-Diiodo-9-
oxabicyclo[3.3.1]nonane [(Ϯ)-8] (78.0 g, 0.21 mol) and dicyclo-
hexylmethylamine (88.5 g, 0.5 mol) were refluxed for 2 hours. After
cooling to room temperature the reaction cake was dissolved in a
mixture of 2  hydrochloric acid (350 mL) and Et2O (250 mL). The
layers were separated and the aqueous phases was extracted with
Et2O (2 ϫ 100 mL). The combined organic layer was neutralized
with 5% aqueous NaHCO3 (100 mL), washed with water (100 mL)
and dried over MgSO4. After evaporation of the solvent a light
yellowish waxy solid remained. Yield 22.3 g (88%, 96% purity, GC).
M.p. 37 °C (n-pentane, ref.[59] M.p. 34Ϫ36 °C). The spectroscopic
(1S,2S,4S,5S,7S,8S)-(؉)-2-Oxa-6-thiatricyclo[3.3.1.13,7]decane-4,8-
diol [(؉)-11]: Yield 0.37 g, (31%). [α]2D0 ϭ ϩ22.5 (c ϭ 1.0, ethanol),
Eur. J. Org. Chem. 2004, 2181Ϫ2192
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2189

K. Hegemann, R. Fröhlich, G. Haufe
FULL PAPER
Ͼ98% ee (chiral GC, 190 °C, isotherm). All spectroscopic data X-ray Crystallographic Studies
agree with those given for the racemic compound. Careful recrys-
(1S,2S,5S,6S)-(؊)-9-Oxabicyclo[3.3.1]nonane-2,6-diol
[(؊)-3]:
tallization from acetone gave single crystals suitable for X-ray
structural analysis.
C8H14O3, M ϭ 158.19, colorless crystal 0.40 ϫ 0.25 ϫ 0.15 mm,
3
˚
˚
a ϭ 9.572(1), b ϭ 9.572(1), c ϭ 16.928(1) A, V ϭ 1551.0(2) A ,
calcd. ϭ 1.355 g cmϪ3, µ ϭ 8.46 cmϪ1, empirical absorption correc-
(1R,2R,4R,5R,7R,8R)-(؊)-4-Acetoxy-8-hydroxy-2-oxa-6-thiatri-
cyclo[3.3.1.13,7]decane [(؊)-13]: Yield 0.26 g (17%). M.p. 144 °C, [α]
ρ
tion (0.728 Յ T Յ 0.884), Z ϭ 8, tetragonal, space group P43212
20 ϭ Ϫ17.6 (c ϭ 1.0, ethyl acetate), 34% ee (chiral GC after saponi-
˚
(no. 96), λ ϭ 1.54178 A, T ϭ 223 K, ω/2θ scans, 1804 reflections
D
1
collected (Ϫh, Ϫk, Ϫl), [(sinθ)/λ] ϭ 0.62 AϪ1, 1573 independent
˚
fication to the diol). H NMR: δ ϭ 2.06Ϫ2.21 (m, 2 H), 2.17 (s, 3
H), 2.81Ϫ2.96 (m, 4 H), 3.92Ϫ3.99 (m, 2 H), 4.01Ϫ4.07 (m, 1 H),
5.13 (dd, 3J ϭ 1.7 Hz, 1 H) ppm. 13C NMR: δ ϭ 21.1 (q), 31.3 (t),
31.7 (d), 32.6 (d), 68.5 (d), 69.7 (d), 71.0 (d), 72.0 (d), 170.2 (s)
ppm. MS (GC/MS): m/z (%) ϭ 230 (30), 188 (4), 170 (40), 137
(100), 123 (6), 113 (8), 111 (18), 97 (28), 94 (29), 85 (14), 81 (24),
73 (8), 69 (7). C10H14O4S (230.3): calcd. C 52.16, H, 6.13; found C
52.31, H 6.28.
(Rint ϭ 0.035) and 1478 observed reflections [I Ն 2σ(I)], 103 refined
parameters, R ϭ 0.031, wR2 ϭ 0.086, Flack parameter Ϫ0.03(19),
max. residual electron density 0.17 (Ϫ0.13) e·AϪ3; hydrogen atoms
˚
calculated and refined as riding atoms.
endo,endo-2,6-Diiodo-9-oxabicyclo[3.3.1]nonane [(؎)-8]: C8H12I2O,
M ϭ 377.98, colorless crystal 0.40 ϫ 0.30 ϫ 0.25 mm, a ϭ
˚
19.221(1), b ϭ 5.394(1), c ϭ 12.996(1) A, β ϭ 128.77(1)°, V ϭ
3
1050.5(2) A , ρcalcd. ϭ 2.390 g cmϪ3, µ ϭ 59.37 cmϪ1 empirical
˚
(1R,2R,4R,5R,7R,8R)-(؊)-4,8-Diacetoxy-2-oxa-6-thiatricyclo-
[3.3.1.13,7]decane [(؊)-14]: Yield 0.55 g (31%). [α]2D0 ϭ Ϫ55.9 (c ϭ
1.0, ethyl acetate); 92% ee (chiral GC after saponification to the
diol). All spectroscopic data agree with those given for the race-
mate.
absorption correction (0.142 Յ T Յ 0.226), Z ϭ 4, monoclinic,
˚
space group C2/c (No. 15), λ ϭ 0.71073 A, T ϭ 223 K, ω/2θ scans,
2141 reflections collected (Ϯh, Ϫk, Ϯl), [(sinθ)/λ] ϭ 0.62 AϪ1, 1074
˚
independent (Rint ϭ 0.029) and 975 observed reflections [I Ն 2σ(I)],
52 refined parameters, R ϭ 0.037, wR2 ϭ 0.097, max. residual elec-
Reductive Desulfuration of (1S,2S,4S,5S,7S,8S)-(؉)-2-Oxa-6-thia-
tricyclo[3.3.1.13,7]decane-4,8-diol [(؉)-11]: (1S,2S,4S,5S,7S,8S)-(ϩ)-
2-Oxa-6-thiatricyclo[3.3.1.13,7]decane-4,8-diol [(ϩ)-11] (180 mg,
0.95 mmol) was dissolved in ethanol (20 mL), Raney-nickel (3 g)
was added and the mixture refluxed for 36 hours. Work-up as de-
scribed above gave (1S,2R,5S,6R)-(ϩ)-9-oxabicyclo[3.3.1]nonane-
2,6-diol [(ϩ)-12]. Yield 70 mg (47%). [α]2D0 ϭ ϩ25.4 (c ϭ 1.0,
ethanol); Ͼ98% ee (chiral GC).
tron density 1.87 (Ϫ1.27) e·AϪ3; hydrogen atoms calculated and
˚
refined as riding atoms. An X-ray structure of this compound is
already known.[67]
9-Oxabicyclo[3.3.1]nona-2,6-diene [(؎)-9]: C8H10O, M ϭ 122.16,
colorless crystal 0.60 ϫ 0.55 ϫ 0.40 mm, a ϭ 5.429(1), b ϭ
3
˚
˚
11.656(2), c ϭ 10.316(1) A, β ϭ 100.77(1)°, V ϭ 641.3(2) A ,
ρ
calcd. ϭ 1.265 g cmϪ3, µ ϭ 6.42 cmϪ1, empirical absorption correc-
tion (0.699 Յ T Յ 0.783), Z ϭ 4, monoclinic, space group P21/c
Kinetic Resolution of exo,exo-9-Oxabicyclo[3.3.1]octane-2,6-diol
[(؎)-12]: exo,exo-9-Oxabicyclo[3.3.1]octane-2,6-diol [(Ϯ)-12] (0.8 g,
5.1 mmol) was dissolved in vinyl acetate (32 mL), treated with CRL
(0.6 g) and the mixture stirred at room temperature. The progress
of the reaction was followed by GC. After 8.5 hours 50% of the diol
was consumed and the lipase was separated by filtration through a
short silica-gel column with ethyl acetate (200 mL) as eluent. After
evaporation of the solvent the residue was separated by column
chromatography (ethyl acetate).
˚
(no. 15), λ ϭ 1.54178 A, T ϭ 223 K, ω/2θ scans, 1381 reflections
collected (Ϯh, Ϫk, Ϫl), [(sinθ)/λ] ϭ 0.62 AϪ1, 1310 independent
˚
(Rint ϭ 0.035) and 1269 observed reflections [I Ն 2σ(I)], 83 refined
parameters, R ϭ 0.037, wR2 ϭ 0.097, max. residual electron density
0.21 (Ϫ0.15) e·AϪ3; hydrogen atoms calculated and refined as rid-
˚
ing atoms.
4,8-Dichloro-2-oxa-6-thiatricyclo[3.3.1.13,7]decane
[(؎)-10]:
C8H10Cl2OS, M ϭ 225.12, colorless crystal 0.50 ϫ 0.40 ϫ 0.10 mm,
˚
a ϭ 12.488(2), b ϭ 9.816(2), c ϭ 7.432(1) A, β ϭ 96.65(1)°, V ϭ
(1S,2R,5S,6R)-(؉)-9-Oxabicyclo[3.3.1]nonane-2,6-diol
[(؉)-12]:
3
904.9(3) A , ρcalcd. ϭ 1.652 g cmϪ3, µ ϭ 81.69 cmϪ1, empirical ab-
˚
Yield 0.18 g (23%). [α]2D0 ϭ ϩ22.5 (c ϭ 1.0, ethanol); 94% ee (chiral
GC). The spectroscopic data agree with those given above for the
racemic compound.
sorption correction (0.106 Յ T Յ 0.496), Z ϭ 4, monoclinic, space
˚
group C2/c (no. 15), λ ϭ 1.54178 A, T ϭ 223 K, ω/2θ scans, 920
reflections collected (ϩh, Ϫk, Ϯl), [(sinθ)/λ] ϭ 0.62 AϪ1, 882 inde-
˚
(1R,2S,5R,6S)-(؊)-2-Acetoxy-9-oxabicyclo[3.3.1]nonan-6-ol [(؊)-
pendent (Rint ϭ 0.089) and 874 observed reflections [I Ն 2σ(I)], 57
15]: Yield 0.30 g (28%). M.p. 70Ϫ71 °C (ethyl acetate). [α]2D0
ϭ
refined parameters, R ϭ 0.045, wR2 ϭ 0.126, max. residual electron
density 0.45 (Ϫ0.42) e·AϪ3; hydrogen atoms calculated and refined
Ϫ11.7 (c ϭ 1.0, ethyl acetate); 30% ee (chiral GC, after saponifi-
˚
1
cation). H NMR: δ ϭ 1.36Ϫ1.49 (m, 2 H), 1.72Ϫ1.80 (m, 2 H),
as riding atoms.
1.90Ϫ2.38 (m, 4 H), 2.06 (s, 3 H), 3.54Ϫ3.59 (m, 1 H), 3.83Ϫ3.91
(m, 2 H), 4.68Ϫ4.76 (m, 1 H) ppm. 13C NMR: δ ϭ 21.1 (t), 21.3
(q), 22.5 (t), 22.6 (t), 25.1 (t), 68.1 (d), 70.0 (d), 70.9 (d), 170.6 (s)
ppm. MS (GC/MS): m/z (%) ϭ 200 (8), 182 (4), 139 (15), 121 (6),
112 (6), 104 (6), 95 (28), 79 (35), 67 (35). C10H16O4 (200.2): calcd.
C 59.98, H 8.05; found C 60.13, H 8.11.
2-Oxa-6-thiatricyclo[3.3.1.13,7]decane-4,8-diol [(؎)-11]: C8H12O3S,
M ϭ 188.24, colorless crystal 0.60 ϫ 0.40 ϫ 0.20 mm, a ϭ 6.465(1),
˚
b ϭ 6.957(1), c ϭ 10.139(1) A, α ϭ 88.64(1)°, β ϭ 74.16(1)°, γ ϭ
3
71.23(1)°, V ϭ 414.3(1) A , ρcalcd. ϭ 1.509 g cmϪ3, µ ϭ 31.90 cmϪ1
,
˚
empirical absorption correction (0.233 Յ T Յ 0.528), Z ϭ 2, tri-
˚
¯
clinic, space group P1 (No. 2), λ ϭ 1.54178 A, T ϭ 223 K, ω/2θ
scans, 1838 reflections collected (Ϫh, Ϯk, Ϯl), [(sinθ)/λ] ϭ 0.62
A
(1R,2S,5R,6S)-(؊)-2,6-Diacetoxy-9-oxabicyclo[3.3.1]nonane [(؊)-
16]: Yield 0.22 g (17%). M.p. 111 °C (ethyl acetate, ref.[13] M.p.
113Ϫ114 °C for the racemic compound). [α]2D0 ϭ Ϫ29.3 (c ϭ 1.0,
ethyl acetate), 92% ee (chiral GC after saponification). 13C NMR:
δ ϭ 21.2 (q), 21.7 (t), 22.6 (t), 69.6 (d), 70.6 (d), 170.6 (s) ppm. MS
(GC/MS): m/z (%) ϭ 242 (2) [Mϩ], 183 (10), 139 (10), 122 (18),
Ϫ1, 1680 independent (Rint ϭ 0.044) and 1657 observed reflec-
˚
tions [I Ն 2σ(I)], 112 refined parameters, R ϭ 0.038, wR2 ϭ 0.102,
max. residual electron density 0.46 (Ϫ0.26) e·AϪ3; hydrogen atoms
˚
calculated and refined as riding atoms.
104 (6), 94 (32), 81 (13), 67 (15), 55 (9), 43 (100). C12H18O5 (242.3): (1S,2S,4S,5S,7S,8S)-(؉)-2-Oxa-6-thiatricyclo[3.3.1.13,7]decane-4,8-
calcd. C 59.49, H 7.49; found C 59.58, H 7.73. A 100 MHz 1H
NMR spectrum of the racemate is known in the literature.[13]
diol [(؉)-11]: C8H12O3S, M ϭ 188.24, colorless crystal 0.60 ϫ 0.50
˚
ϫ 0.20 mm, a ϭ 9.122(1), b ϭ 10.207(1), c ϭ 17.925(1) A, V ϭ
2190
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 2181Ϫ2192

Synthesis of Enantiopure 9-Oxabicyclononanediol Derivatives
FULL PAPER
F. Theil, Chem. Rev. 1995, 95, 2203Ϫ2227.
[17]
[18]
[19]
[20]
3
1669.0(3) A , ρcalcd. ϭ 1.498 g cmϪ3, µ ϭ 31.68 cmϪ1, empirical
˚
K. Faber, S. Riva, Synthesis 1992, 895Ϫ910.
absorption correction (0.236 Յ T Յ 0.530), Z ϭ 8, orthorhombic,
˚
W. Bohland, C. Frößl, M. Lorenz, Synthesis 1991, 1049Ϫ1072.
K. Hegemann, H. Schimanski, U. Höweler, G. Haufe, Tetra-
hedron Lett. 2003, 44, 2225Ϫ2229.
space group P212121 (no. 19), λ ϭ 1.54178 A, T ϭ 223 K, ω/2θ
scans, 1956 reflections collected (Ϫh, Ϫk, Ϫl), [(sinθ)/λ] ϭ 0.66
A
Ϫ1, 1956 independent and 1943 observed reflections [I Ն 2σ(I)],
˚
[21]
[22]
E. Heymann, W. Junge, Eur. J. Biochem. 1979, 95, 509Ϫ518.
L. K. P. Lam, C. M. Brown, B. De Jeso, L. Lym, E. J. Toone,
J. B. Jones, J. Am. Chem. Soc. 1988, 110, 4409Ϫ4411.
A. M. Klibanov, CHEMTECH 1986, 354Ϫ359.
A. M. Klibanov, Acc. Chem. Res. 1990, 23, 114Ϫ120.
Enzymatic Reactions in Organic Media (Eds.: A. M. P. Kos-
kinen, A. M. Klibanov), Blackie Academic Professional, New
York, 1996.
222 refined parameters, R ϭ 0.049, wR2 ϭ 0.137, Flack parameter
0.06(3), max. residual electron density 0.74 (Ϫ0.55) e·AϪ3; hydro-
˚
gen atoms calculated and refined as riding atoms.
[23]
[24]
[25]
(1S,2R,5S,6R)-(؉)-9-Oxabicyclo[3.3.1]nonane-2,6-diol
[(؉)-12]:
C8H14O2.923S0.077, M ϭ 159.43, colorless crystal 0.50 ϫ 0.30 ϫ
3
˚
˚
0.20 mm, a ϭ 11.223(5), c ϭ 6.434(3) A, V ϭ 810.4(6) A , ρcalcd. ϭ
1.307 g cmϪ3, µ ϭ 9.87 cmϪ1, empirical absorption correction
(0.638 Յ T Յ 0.827), Z ϭ 8, tetragonal, space group P41212 (No.
[26]
[27]
[28]
[29]
[30]
K. Faber, Biotransformations in Organic Chemistry, 4th ed.,
Springer, Berlin, 2000, pp. 334Ϫ418.
˚
92), λ ϭ 1.54178 A, T ϭ 223 K, ω/2θ scans, 970 reflections col-
G. Carrea, S. Riva, Angew. Chem. 2000, 112, 2312Ϫ2341; An-
gew. Chem. Int. Ed. 2000, 39, 2226Ϫ2254.
A. Sattler, G. Haufe, Tetrahedron: Asymmetry 1995, 6,
lected (Ϫh, ϩk, ϩl), [(sinθ)/λ] ϭ 0.62 AϪ1, 823 independent (Rint ϭ
˚
0.025) and 693 observed reflections [I Ն 2σ(I)], 59 refined param-
eters, R ϭ 0.036, wR2 ϭ 0.096, Flack parameter 0.2(2), max. re-
2841Ϫ2848.
sidual electron density 0.18 (Ϫ0.11) e·AϪ3. The crystal contained
O. Goj, A. Burchardt, G. Haufe, Tetrahedron: Asymmetry 1997,
8, 399Ϫ408.
R. Skupin, T. G. Cooper, R. Fröhlich, J. Prigge, G. Haufe,
˚
7.7(3)% of the sulfur-bridged starting material, disorder refined
with PART command, hydrogen atoms calculated and refined as
riding atoms.
Tetrahedron: Asymmetry 1997, 8, 2453Ϫ2464.
[31]
[32]
D. Wölker, G. Haufe, J. Org. Chem. 2002, 67, 3015Ϫ3021.
This discrepancy is due to the fact that combustion of the diols
in the flame ionization detector (FID) of the gas chromato-
graph produces less ions than the mono- or diacetates and
hence their presence is underestimated. Investigating a 1:1
weight-by-weight mixture of compounds 2 and 3 on the one
hand and 6 and 7 or 4 and 5 on the other hand, we determined
an error of about Ϯ2%.
CCDC-215356Ϫ215362 contain the supplementary crystallo-
graphic data for this paper. These data can be obtained free of
charge at www.ccdc.cam.ac.uk/conts/retrieving.html [or from the
Cambridge Crystallographic Data Centre, 12 Union Road,
Cambridge CB2 1EZ, UK; Fax: ϩ44-1223-336-033: E-mail:
deposit@ccdc.cam.ac.uk].
[33]
The E value (enantiomeric ratio) describes the selectivity of a
resolution, see: C.-S. Chen, Y. Fujimoto, G. Gierdaukas, C. J.
Sih, J. Am. Chem. Soc. 1982, 104, 7294Ϫ2799.
Acknowledgments
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
B. Berger, C. G. Rabiller, K. Königsberger, K. Faber, H.
Griengl, Tetrahedron: Asymmetry 1990, 1, 541Ϫ546.
T. Oberhauser, K. Faber, H. Griengl, Tetrahedron 1985, 45,
1679Ϫ1682.
R. J. Kazlauskas, A. N. E. Weissfloch, A. T. Rappaport, L. A.
Cuccia, J. Org. Chem. 1991, 56, 2656Ϫ2665.
S. N. Ahmed, R. J. Kazlauskas, A. H. Morinville, P. Grochul-
ski, J. D. Schrag, M. Cygler, Biocatalysis 1994, 9, 209Ϫ225.
R. S. Cahn, C. K. Ingold, V. Prelog, Angew. Chem. 1966, 78,
413Ϫ447; Angew. Chem. Int. Ed. Engl. 1966, 5, 385Ϫ419.
V. Prelog, G. Helmchen, Angew. Chem. 1982, 94, 614Ϫ631; An-
gew. Chem. Int. Ed. Engl. 1982, 21, 567Ϫ584.
K. Naemura, T. Matsumura, M. Komatsu, Y. Hirose, H. Chi-
kamatsu, J. Chem. Soc., Chem. Commun. 1988, 239Ϫ241.
S. Behr, K. Hegemann, H. Schimanski, G. Haufe, manuscript
in preparation.
Generous financial support by the Deutsche Forschungsgemein-
schaft and the Fonds der Chemischen Industrie is gratefully
acknowledged.
[1]
Enzymes as Catalysts in Organic Synthesis (Ed.: M. Schneider),
NATO ASI Series, vol. 178, Reidel, Dortrecht, 1986.
H. G. Davies, R. H. Green, D. R. Kelly, S. M. Roberts, Bio-
[2]
transformations in Preparative Organic Chemistry, Academic
Press, London, 1989.
Microbial Reagents in Organic Synthesis (Ed.: S. Servi), NATO
[3]
ASI Series, vol. 381, Reidel, Dortrecht, 1992.
C. H. Wong, G. M. Whitesides, Enzymes in Synthetic Organic
[4]
Chemistry, Pergamon Press, Oxford, 1994.
Enzyme Catalysis in Organic Synthesis 2nd (Eds.: K. Drauz, H.
[5]
Waldmann), Wiley-VCH, Weinheim, 2002.
[6]
H. B. Kagan, T. P. Dang, J. Am. Chem. Soc. 1972, 94,
R. Bishop, D. C. Craig, I. G. Dance, S. Kim, M. D. A. I. Mal-
lick, K. C. Pich, M. L. Scudder, Supramol. Chem. 1993, 1,
171Ϫ178.
K. C. Pich, R. Bishop, D. C. Craig, I. G. Dance, A. D. Rae,
M. L. Scudder, Struct. Chem. 1993, 4, 41Ϫ51.
6429Ϫ6433.
[7]
J. K. Whitesell, Chem. Rev. 1989, 89, 1581Ϫ1590.
[8]
A. Alexakis, P. Mangeney, Tetrahedron: Asymmetry 1990, 1,
[43]
[44]
477Ϫ511.
[9]
R. Noyori, Asymmetric Catalysis in Organic Synthesis, Wiley &
R. Bishop, D. C. Craig, I. G. Dance, M. L. Scudder, A. T. Ung,
Supramol. Chem. 1993, 2, 123Ϫ131.
G. Haufe, Tetrahedron Lett. 1984, 25, 4365Ϫ4368.
H. Stetter, H. J. Meissner, Tetrahedron Lett. 1966, 4599Ϫ4602.
C. Ganter, K. Wicker, W. Zwahlen, K. Schaffner-Sabba, Helv.
Chim. Acta 1970, 53, 1618Ϫ1629.
H. J. Franz, W. Höbold, R. Höhn, G. Müller-Hagen, R. Müller,
W. Pritzkow, J. Prakt. Chem. 1970, 312, 622Ϫ634.
J. N. Labows Jr., D. Swern, J. Org. Chem. 1972, 37, 3004Ϫ3006.
N. J. Averina, N. S. Zefirov, P. P. Kadzyauskas, S. V. Rogozina,
N. K. Sadovaya, N. M. Soldatov, Zh. Org. Khim. (Engl.
Transl.) 1974, 10, 1453Ϫ1455; Chem. Abstr. 1974, 81, 105341n.
G. Haufe, E. Kleinpeter, M. Mühlstädt, J. Graefe, Monatsh.
Chem. 1978, 109, 575Ϫ585.
Sons, New York, Chichester, 1994.
[10]
J. A. Nieman, B. A. Keay, Tetrahedron: Asymmetry 1996, 7,
[45]
[46]
[47]
3521Ϫ3526.
[11]
P. Lafont, G. Vivant, Franz. Pat. 1336187, 1963 [Chem. Abstr.
1963, 60, 2803].
[12]
A. H. Frazer, US Pat. 3347826, 1967 [Chem. Abstr. 1968, 68,
[48]
P2819].
[13]
R. O. Duthaler, K. Wicker, P. Ackermann, C. Ganter, Helv.
[49]
[50]
Chim. Acta 1972, 55, 1809Ϫ1829.
[14]
P. E. Eaton, R. Millikan, Synthesis 1990, 483Ϫ484.
[15]
A. C. Cope, B. S. Fisher, W. Funke, J. M. McIntosh, M. A.
McKervey, J. Org. Chem. 1969, 34, 2231Ϫ2234.
E. Schoffers, A. Golebiowski, C. R. Johnson, Tetrahedron
[16]
[51]
1996, 52, 3769Ϫ3826.
Eur. J. Org. Chem. 2004, 2181Ϫ2192
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2191

K. Hegemann, R. Fröhlich, G. Haufe
FULL PAPER
[52]
[59]
R. M. Scarbourgh Jr., A. B. Smith III, W. E. Barnette, K. C.
H. Stetter, H. J. Meissner, W. D. Last, Chem. Ber. 1968, 101,
2889Ϫ2893.
H. Stetter, K. Heckel, Chem. Ber. 1973, 106, 339Ϫ348.
A. Takahashi, M. Aso, M. Tanaka, H. Suemune, Tetrahedron
2000, 56, 1999Ϫ2006.
C. Ganter, K. Wicker, Helv. Chim. Acta 1968, 51, 1599Ϫ1603.
T. Oberhauser, K. Faber, H. Griengl, Tetrahedron 1989, 45,
1679Ϫ1682.
G. M. Sheldrick, Acta Crystallogr., Sect. A 1990, 46, 467Ϫ473.
G. M. Sheldrick, University of Göttingen, 1997.
E. Keller, Universität Freiburg, 1997.
Nicolaou, J. Org. Chem. 1979, 44, 1742Ϫ1744.
K. C. Nicolaou, D. A. Claremon, W. E. Barnette, S. P. Seitz,
[53]
[60]
[61]
J. Am. Chem. Soc. 1979, 101, 3704Ϫ3706.
A. Toshimizu, T. Aoai, S. Uemura, M. Okano, J. Org. Chem.
[54]
[62]
[63]
1981, 46, 3021Ϫ3026.
A. Toshimizu, T. Aoai, H. Owada, S. Uemura, M. Okano,
[55]
Tetrahedron 1985, 41, 5301Ϫ5306.
G. Haufe, G. Alvernhe, D. Anker, A. Laurent, J. Fluorine
[56]
[64]
[65]
[66]
[67]
Chem. 1990, 46, 83Ϫ95.
G. Haufe, G. Alvernhe, D. Anker, A. Laurent, C. Saluzzo, J.
[57]
Org. Chem. 1992, 57, 714Ϫ719.
R. Wartchow, U. Albrecht, H. M. R. Hoffmann, Z. Kristallogr.
1996, 211, 328Ϫ328.
[58]
K. Hegemann, B. Wibbeling, G. Haufe, J. Prakt. Chem. 1998,
340, 455Ϫ458.
Received December 16, 2003
2192
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 2181Ϫ2192