Biotransformations in low-boiling hydrofluorocarbon solvents

Article abstract of DOI:10.1002/anie.200460082

Solvent solutions: Low-boiling hydrofluorocarbons (see examples) are excellent solvents for lipase-catalyzed reactions and ideal replacements for conventional organic solvents and supercritical fluids as media for nonaqueous biotransformations. Notably increased rates, yields, and enantioselectivities were observed with the model kinetic resolution of (±)-1-phenylethanol and the desymmetrization of weso-2-cyclopentene-1 ,4-diol.

Full text of DOI:10.1002/anie.200460082

Angewandte
Chemie
Fluorinated Solvents
Biotransformations in Low-Boiling
Hydrofluorocarbon Solvents**
Simon Saul, Stuart Corr, and Jason Micklefield*
The scope of biotransformations, particularly in the prepara-
tion of homochiral precursors, has been considerably
extended through the use of enzymes in organic solvents.[1]
Some of the advantages of organic solvents include ease of
recovery of products and enzymes, increased solubility and
rate of transformation of more lipophilic substrates, increased
lifetime of enzymes, increased chemo-, regio-, and enantio-
selectivity through control of solvent physical properties
(solvent engineering), and control of thermodynamic equi-
libria in favor of synthesis rather than hydrolysis.[1] More
recently the use of enzymes in nonaqueous media has been
extended to include supercritical fluids[2] and ionic liquids[3]
with tuneable solvent properties. The use of these solvents
reduces the quantities of waste volatile organic compounds
(VOCs), which is an important step in the direction of “green
chemistry”. One group of potential solvents for biotransfor-
[*] S. Saul, Dr. J. Micklefield
School of Chemistry
The University of Manchester
Institute of Science and Technology (UMIST)
PO Box 88, Manchester M601QD (UK)
Fax: (+44)161-200-4484
E-mail: j.micklefield@manchester.ac.uk
Dr. S. Corr
INEOS Fluor Ltd.
PO Box 13
The Heath, Runcorn, Cheshire WA7 4QF (UK)
[**] This work is supported by INEOS Fluor Ltd.
Supportinginformation for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2004, 43, 5519 –5523
DOI: 10.1002/anie.200460082
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5519

Communications
mations that has received little attention to date is that of
pressurized liquids with normal boiling points below, and
critical temperatures above, room temperature. These fluids
are easily handled at moderate pressures and their relative
volatility should allow ready removal of solvent residues from
the products, in contrast to the situation with many ionic
liquids. Such fluids can be readily compressed and reliquified
in a closed system, thereby allowing the solvents to be
recycled and reused with minimal losses into the environ-
ment, which is an important factor in the quest for greener
chemical processes.
Here we report the first investigation of biotransforma-
tions in liquid-phase, low-boiling hydrofluorocarbon solvents
(HFCs).[4] HFCs are generally of low toxicity, do not have
ozone depletion potential, and are not classed as VOCs.
Many, such as 1,1,1,2-tetrafluoroethane (R-134a)and
1,1,1,2,3,3,3-heptafluoropropane (R-227ea), are also non-
flammable. Several HFCs are used as replacements for the
chlorofluorocarbons (CFCs)and hydrochlorofluorocarbons
(HCFCs)in the refrigeration industry and are manufactured
globally on a large scale to high purity. Both R-134a and R-
227ea are also manufactured to current good manufacturing
practice (cGMP)standards for use in metered-dose inhaler
applications in the pharmaceutical industry. The HFCs thus
have the potential to be environmentally benign and eco-
nomically feasible alternatives to conventional organic sol-
vents and supercritical fluids. Whilst
Scheme 1. Lipase-catalyzed kinetic resolution of racemic 1-phenyletha-
nol (1) and desymmetrization of meso-2-cyclopentene-1,4-diol (4).
with vinyl acetate 2, catalyzed by immobilized lipase B from
Candida antarctica (Novozym 435), was carried out in the five
different media and the progress of the reactions was
monitored by GC. We found that plastic-coated glass aerosol
bottles (10 mL), with teflon-coated aerosol valves that could
be crimp sealed, were ideal for small-scale HFC-based
reactions and were able to withstand the moderate pressures
(up to 19 bar for R-32 at 308C)[5] without any loss of solvent.
The results (Table 1)clearly show that the reactions in the
a number of the properties of HFCs
Table 1: Kinetic resolution of 1-phenylethanol (1) catalyzed by Novozym 435.
Solvent t[a] [h] Conv. [%] ee (S)-1 [%] ee (R)-3 [%] Initial rates[b] [nmolminꢁ1 mgꢁ1
aw <0.01 aw ꢂ0.43 aw ꢂ0.58 aw ꢂ0.75
are also common to supercritical
fluids,[4] the moderate absolute pres-
sure of liquid-phase HFCs[5] pre-
cludes the need for expensive, spe-
cialized high-pressure reaction
equipment. A further advantage of
using liquid HFCs is their polarities,
which are comparable to those of
moderately polar organic solvents
such as tetrahydrofuran (THF)and
dichloromethane.[6a] This increase in
polarity over solvents such as
hexane and supercritical CO2 may
improve the solubility of a range of
]
R-32
5
3.5
4
8
35
50
49
49
46
49
>99
96
96
85
96
>99
>99
>99
>99
>99
n.d.[c]
n.d.
325
227
51
n.d.
n.d.
387
241
64
n.d.
n.d.
407
260
68
n.d.
n.d.
358
228
60
R-227ea
R-134a
hexane
MTBE
[a] The time point when no further reaction was evident (that is, the rate of the reaction was approaching
zero). [b] Initial rates are given in units of nmol of product 3 per minute per mgof enzyme and were
determined from the slope of the time-course measurements between 0 and 5% conversion for an
approximate water activity (aw). [c] n.d.=not determined.
desirable substrates and avoid the need for the use of polar
cosolvents, which are difficult to remove. Despite this
increased polarity, HFCs are relatively hydrophobic,[6b]
which should ensure that enzymes retain their essential
active-site water molecules and activity.[7]
HFCs are superior, both in terms of rate and degree of
conversion observed. In the case of R-32, a near-perfect
resolution was achieved with approximately 50% conversion
after 5 h resulting in a virtually equimolar mixture of
homochiral product ester (R)-3 and unreacted alcohol (S)-1.
Moreover this reaction was easily scaled up by using a one-
liter aluminum reaction vessel (see the Supporting Informa-
tion). By starting with (ꢀ )-1-phenylethanol (1.24 g) in R-32
(100 mL), ester (R)-3 (772 mg)was isolated in 46% yield and
99% ee, as determined by chiral GC, along with alcohol (S)-1
(595 mg)in 48% yield and 99% ee, after 5 h.
It is well known that the activity of enzymes in organic
solvents depends on the thermodynamic water activity (aw)of
the system.[10] The reactions described here were all carried
out in anhydrous solvents[8] where the aw value is close to zero,
thereby enabling a fair comparison between different media.
However, in order to examine how the thermodynamic water
Initially we chose to investigate the model lipase-cata-
lyzed kinetic resolution of (ꢀ )-1-phenylethanol (rac-1,
[8]
Scheme 1)in anhydrous
R-134a, R-227ea, and difluoro-
methane (R-32). It is widely accepted that transesterification
reactions catalyzed by lipases are most efficient in apolar
hydrophobic solvents because more polar solvents can strip
the enzymes of their essential water.[7] We therefore chose to
compare the resolution of 1 in HFCs with the reaction carried
out under identical conditions in anhydrous[8] hexane and
methyl tert-butyl ether (MTBE), both of which have been
shown to be good solvents for this biotransformation.[9]
Accordingly, the acylation of (ꢀ )-1-phenylethanol (rac-1)
5520
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 5519 –5523

Angewandte
Chemie
activity affects the lipase-catalyzed acylation of (ꢀ )-1-phe-
nylethanol (rac-1)in an HFC solvent (R-134a)in comparison
with the reaction in conventional organic solvents, the initial
rates of the reaction were measured at different water
activities.[10b] From these measurements it can be seen
(Table 1)that the initial rates follow the typical bell-shaped
profile,[10a,c] with a maximum rate attained at aw ꢂ 0.58 in all
solvents. Notably, the initial rates in R-134a are always higher
than those in hexane or MTBE, irrespective of the aw value.
The desymmetrization of prochiral or meso-diols by
monoacylation catalyzed by various hydrolases in nonaqu-
eous solvents can give an enantiomerically pure product in
near quantitative yield. As a result, these biotransformations
are now widely employed. In order to further investigate the
potential of the HFCs as media for biotransformations, the
model lipase-catalyzed desymmetrization of meso-2-cyclo-
pentene-1,4-diol (4)with vinyl acetate 2 was studied. The
acyclated product (1R,4S)-1-hydroxy-2-cyclopentene-4-ace-
tate (5)and its enantiomer are starting materials for the
synthesis of various cyclopentenoid natural products, such as
prostaglandins, prostacyclins, and thromboxanes.[11] Previous
investigations have found that THF with Et3N as an additive
is the solvent system of choice for this biotransformation.[11a,b]
Accordingly, the reaction of meso-diol 4 with vinyl acetate,
catalyzed by the lipases from Pseudomonas cepacia and
C. antarctica (Novozym 435), was carried out in anhydrous R-
134a, R-227ea, and R-32 and compared with identical trans-
formations in anhydrous THF with and without added Et3N.
The time courses for all reactions were followed and the
results shown in Table 2 indicate the time point at which the
A typical time-course plot for the desymmetrization of
meso-diol 4 with Novozym 435 in R-134a is shown in Figure 1.
The desymmetrization process comprises two reactions. The
Figure 1. Time-course plot of the desymmetrization of meso-2-cyclo-
pentene-1,4-diol (4) catalyzed by Novozym 435 in R-134a. The graph
~
shows the percentage of meso-diol 4 remaining( ), the percentage
&
^
*
yields of 5 ( ) and 6 ( ), and the ee value of product 5 ( ) versus
time.
first reaction produces the monoacylated product 5 (or its
enantiomer). In the second step, which constitutes per se a
kinetic resolution, the monoacetate product is subject to a
second acylation that produces the diacetate 6. As a result of
this, and as is evident from the time-course plot, when all of
the diol is consumed the yield of the monoacetate product 5
begins to drop while its ee value continues to increase due to
the kinetic resolution of the second acylation step. Thus, the
final yield of nearly optically pure monoacetate product 5 is
dependent on the enantioselectivity of the enzyme in the first
acylation step. Clearly, greater enantioselectivities in the first
acylation step are achieved when the reaction is undertaken in
the HFCs than when it is performed in THF/Et3N. Further-
more, inspection of the time-course plots for all five solvent
systems clearly shows that the rate of reaction with Novo-
zym 435 is much lower in THF/Et3N or THFalone, with a half
life (t1/2)of approximately 8–12 h, than in the HFCs ( t1/2 = 1.5–
2 h), with the highest rate being observed in R-227ea. With
the P. cepacia lipase the t1/2 value for the reactions in the HFCs
was 0.5–1.5 h, which can be compared with the values of 2.5 h
for THF/Et3N and 3.5 h for THFalone. The desymmetrization
of meso-diol 4 (500 mg)by using Novozym 435 was also
successfully carried out on a preparative scale in a one-liter
aluminum vessel containing R-227ea (500 mL); this resulted
in the isolation of product 5 (425 mg, 59%)in 99% ee, as
determined by chiral GC, along with diacetate 6 (372 mg,
40%), after 3 h.
Table 2: Desymmetrization of meso-2-cyclopentene-1,4-diol (4) catalyzed
by lipase from P. cepacia and Novozym 435.
Solvent
P. cepacia lipase
Novozym 435
t [h][a] Yield 5 [%] ee 5 [%] t [h][a] Yield 5 [%] ee 5 [%]
R-134a
R-227ea
R-32
THF/Et3N 17
THF 17
3.5 53
>99
>99
>99
>99
38
4
3
55
61
>99
>99
>99
91
3
5
42
58
43
53
5.5 55
48
48
42
56
40
[a] The time point at which the maximum ee value of monoacetate
product 5 was achieved.
maximum ee value was achieved for the monoacetate product
5. With both lipases it is clear that enzyme activity is far
greater in the HFCs than in THF/Et3N or in THFalone. In the
case of P. cepacia lipase comparable or higher yields, up to
58%, of enantiopure monoacetate 5 are achieved in around a
quarter of the time. For Novozym 435 the use of HFCs
similarly increases the yield of 5 with dramatically improved
rates and increases the enantioselectivity from 40% ee in
THF or 91% ee in THF/Et3N to > 99% ee in all of the HFCs.
It is also interesting to note that R-227ea gives the lowest
yield of 5 (42%)with P. cepacia lipase but gives the highest
yield of 5 (61%)with Novozym 435. Clearly HFCs, like
conventional solvents, differ in their physical properties (for
example, dielectric constants and dipole moments).[6a] Thus,
whilst one HFC may be the best solvent for one particular
enzyme, it does not necessarily follow that it will be the best
for another.
Finally, in order to explore the utility of the HFCs as
media for other classes of enzyme-catalyzed reactions, the
model transesterifications of N-acetyl and N-trifluoroacetyl
phenylalanine propyl esters rac-7 and rac-8 (Scheme 2)with
methanol, catalyzed by subtilisin Carlsberg protease, were
investigated in anhydrous solvents.[12] In this case, R-134a
gave the highest rates and yields of enantiopure methyl esters
(S)-9 and (S)-10 and was marginally better than hexane, the
Angew. Chem. Int. Ed. 2004, 43, 5519 –5523
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5521

Communications
and immediately charged with the HFC (5.0 mL), then the reaction
mixture was stirred magnetically at room temperature. Samples
(approximately 20 mL)were discharged periodically through the
aerosol valve, dissolved in CH2Cl2 (100 mL), and analyzed by GC, or
chiral GC where appropriate. Reactions in hexane and MTBE were
carried out in an identical fashion except a Supelco graduated screw-
top vial (7 mL)was used as the reaction vessel and samples (1 mL)
were withdrawn periodically for analysis by using a Hamilton syringe.
For initial rate measurements the reactions were carried out in
identical fashion, except 1 mg (10 U)of Novozym 435 was used with
the same substrate concentrations and total volume (5 mL). The aw
values of hexane or MTBE reaction mixtures with Novozym 435 were
all adjusted by preequilibration with saturated salt solutions as
described previously.[10b] The aw value of R-134a reaction mixtures
with Novozym 435 was adjusted by mixing appropriate ratios, by
weight, of anhydrous and water-saturated R-134a. The relationship
between molar ratio of water and aw value is not always linear so the
water activities in R-134a are only approximate. The large-scale
kinetic resolution of rac-1 (1.24 g, 10.2 mmol)was carried out in a 1-L
aluminum reaction vessel with Novozym 435 (190 mg, 1900 U), vinyl
acetate (17.3 g, 0.201 mol), and R-32 (100 mL). After 5 h, the reaction
was vented and the resulting mixture was separated by silica gel
column chromatography, eluting with 10!1% diethyl ether in
petroleum ether, to give ester (R)-3 (772 mg, 46%; [a]D = +111 (c =
2.0, CH3OH); literature value:[13a] [a]D = +114 (c = 2.0, CH3OH))
and alcohol (S)-1 (595 mg, 48%; [a]D = +41 (c = 2.0, CH3OH);
literature value:[13a] [a]D = +45 (c = 2.0, CH3OH)). Both products
were identical by 1H and 13C NMR spectroscopy to commercial
samples as well as to the characterization data in a previous report.[13b]
The desymmetrization of meso-diol 4 in HFCs (5 mL)was carried
out and analyzed as described above with 4 (5.0 mg, 0.05 mmol), vinyl
acetate (86.1 mg, 1.00 mmol), and P. cepacia lipase (5.0 mg, 0.463 U)
or Novozym 435 (1.0 mg, 10 U)in aerosol bottles. The reaction was
also carried out in an identical fashion with and without Et3N
(10.1 mg, 0.1 mmol)in anhydrous THF (5 mL.) The large-scale
desymmetrization of 4 (500 mg, 4.99 mmol)was carried out in a 1-L
aluminum vessel containing R-227ea (500 mL), vinyl acetate (8.61 g,
0.10 mol), and Novozym 435 (100 mg, 1000 U). After 5 h, the reaction
mixture was purified by silica gel column chromatography, eluting
with hexane/ethyl acetate (2:1), to give monoacetate 5 (425 mg, 59%;
[a]D = +65 (c = 1.0, CHCl3); literature value:[11a] [a]D = +66 (c =
1.0, CHCl3)) and diactetate 6 (372 mg, 40%). Both products were
identical by 1H and 13C NMR spectroscopy to a commercial samples
as well as to the characterization data in a previous report.[11] The
kinetic resolutions of N-acetyl and N-trifluoroacetyl phenylalanine
propyl esters rac-7 and rac-8 were carried out as before in aerosol
bottles containing anhydrous HFCs (4 mL)or in vials containing
anhydrous conventional solvents (4 mL). Reaction mixtures con-
tained rac-7 (9.9 mg, 0.04 mmol)or rac-8 (12.1 mg, 0.04 mmol),
MeOH (25.6 mg, 0.80 mmol), and subtilisin Carlsberg (4.0 mg,
0.042 U).
Scheme 2. Transesterification of racemic N-protected phenylalanine
propyl esters catalyzed by subtilisin Carlsberg.
best conventional solvent examined, but significantly better
than the more polar solvents, THF and acetonitrile (Table 3).
Transesterifications in R-32, on the other hand, exhibited
Table 3: Summary of results from the transesterification of N-protected
phenylalanine propyl esters.
Solvent
rac-7 (R=CH3)
rac-8 (R=CF3)
t [h] Yield 9 [%] ee 9 [%] t [h]
Yield 10 [%] ee 10 [%]
R-134a 19 23
R-32 19 13
hexane 19 20
>99
>99
>99
>99
>99
72
33
>99
>99
>99
72 (19)[a] 10 (10)[a]
72
72
72
23
1.3
0
THF
19
8
4
n.d.[b]
CH3CN 19
–
[a] The reaction effectively ceased after 19 h. [b] n.d.=not determined.
similar initial rates to the reactions in hexane but stopped at
lower yields of (S)-9 (13%)and ( S)-10 (10%). This is,
however, still better than the results for THF and acetonitrile.
In summary, we have demonstrated the benefits and
potential of HFCs as solvents for biotransformations. In the
kinetic resolution of model secondary alcohol rac-1 signifi-
cant increases in rate and product yield were demonstrated in
the HFC reactions compared to reactions in the lipase
solvents of choice, hexane and MTBE. The desymmetrization
of a model meso-diol 4 was also achieved with substantially
increased rates, yields, and enantioselectivities in the HFCs in
comparison with the results in the typical organic solvent
system.[11] It is possible that the improved rates of reaction
observed are due, in part, to the low viscosity and the
consequently increased solute diffusivity in the HFCs, which
are mid way between those observed for a typical organic
solvent on one hand and a supercritical fluid on the other.[4]
Finally, the benefits of HFCs are not limited to lipase-
catalyzed reactions, as demonstrated by the improved activity
of the subtilisin Carlsberg protease in R-134a. Indeed, recent
findings, which will be reported in due course, demonstrate
that hydroxynitrile lyases also display activity in the HFC
solvents.
Received: March 22, 2004
Revised: June 11, 2004
Keywords: biotransformations · desymmetrization ·
.
hydrofluorocarbons · kinetic resolution · lipases
Experimental Section
Immobilized lipase B from Candida antarctica (Novozym 435)with a
specific activity of 10000 Ugꢁ1 was purchased from Fluka. Lipase
from Pseudomonas cepacia (92.6 Ugꢁ1)and subtilisin Carlsberg
protease (10.5 Ugꢁ1)were purchased from the Sigma Chemical Co.
In all experiments enzymes were used straight from the bottle, unless
otherwise stated.
[1] a)A. M. Klibanov, Nature 2001, 409, 241 – 246; b)A. Schmidt,
J. S. Dordick, B. Hauer, A. Kiener, M. Wubbolts, B. Witholt,
Nature 2001, 409, 258 – 268; c)G. Correa, S. Riva, Angew. Chem.
2000, 112, 2312 – 2310; Angew. Chem. Int. Ed. 2000, 39, 2226 –
2254; d)S. M. Roberts, J. Chem. Soc. Perkin Trans. 1 2000, 611 –
633; e)S. M. Roberts, J. Chem. Soc. Perkin Trans. 1 1999, 121;
f)C. H. Wong, Science 1989, 244, 1145 – 1152; g)C.-S. Chen, C. J.
For the kinetic resolution of racemic 1-phenylethanol (1)in the
HFCs, Novozym 435 (9.5 mg, 95 U)was added to
0.50 mmol)and vinyl acetate 2 (861.0 mg, 10.0 mmol)in a plastic-
coated aerosol bottle (10 mL). The aerosol was capped, crimp sealed,
1 (61.0 mg,
5522
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 5519 –5523

Angewandte
Chemie
Sih, Angew. Chem. 1989, 101, 711 – 723; Angew. Chem. Int. Ed.
Engl. 1989, 28, 695 – 707.
[2] a)A. J. Mesiano, E. J. Beckman, A. J. Russell, Chem. Rev. 1999,
99, 623 – 633; b)T. Matsuda, T. Harada, K. Nakamura, Chem.
Commun. 2000, 1367 – 1368; c)T. Mori, M. Li, A. Kobayashi, Y.
Okahata, J. Am. Chem. Soc. 2002, 124, 1188 – 1189.
[3] a)M. Erbeldinger, A. J. Mesiano, A. J. Russell, Biotechnol. Prog.
2000, 10, 1129 – 1131; b)F. Van Rantwijk, R. M. Lau, R. A.
Sheldon, Trends Biotechnol. 2003, 21, 131 – 138; c)S. Park, R. J.
Kazlauskas, Curr. Opin. Biotechnol. 2003, 14, 432 – 437.
[4] S. Corr, J. Fluorine Chem. 2002, 118, 55 – 67.
[5] The boiling points and absolute pressures at room temperature
(208C)of liquid phase HFCs used in this study: 1,1,1,2-
tetrafluoroethane (R-134a): 26.18C, 5.7 bar; difluoromethane
(R-32): 51.78C, 14.8 bar; 1,1,1,2,3,3,3-heptafluoropropane (R-
227ea): 15.68C, 3.9 bar.
[6] a)Selected dielectric constants ( e)and dipole moments ( m [D]):
R-134a: e 9.5, m 2.05; R-32: e 8.2, m 1.98; R-227ea: e 4.1, m 0.93;
hexane: e 1.9, m 0.08; THF: e 7.61, m 1.63; CH2Cl2: e 9.08, m 1.55;[4]
b)R-134a, R-32, and R-227ea are all totally immiscible with
water. The solubility of water in R-134a is around 1100 ppm by
weight at 218C.
[7] a)G. Kirchner, M. P. Scollar, A. M. Klibanov, J. Am. Chem. Soc.
1985, 107, 7072 – 7076; b)A. Zaks, A. M. Klibanov, Proc. Natl.
Acad. Sci. USA 1985, 82, 3192 – 3196.
[8] Anhydrous solvents were used throughout. The only water in the
reaction mixture comes from the enzyme. For example, Novo-
zym 435 contains 14 mg of water per mg of enzyme, as
determined by Karl Fischer titration (GRS2000). The majority
of the water will remain tightly associated with the enzyme, so
reaction mixtures have low thermodynamic water activities
(aw < 0.1).
[9] a)S. H. Schꢀfer, N. Kaftzik, P. Wasserscheid, U. Kragl, Chem.
Commun. 2001, 425 – 426; b)M. Persson, U. T. Bornscheuer, J.
Mol. Catal. B 2003, 22, 21 – 27.
[10] a)P. J. Halling, Enzyme Microb. Technol. 1994, 16, 178 – 206;
b)G. A. Hutcheon, P. J. Halling, B. D. Moore, Methods Enzy-
mol. 1997, 286, 465 – 472; c)G. Bell, A. E. M. Janssen, P. J.
Halling, Enzyme Microb. Technol. 1997, 20, 471 – 477.
[11] a)F. Theil, H. Schick, G. Winter, G. Reck, Tetrahedron 1991, 47,
7569 – 7582; b)S. R. Ghorpade, R. K. Kharul, R. R. Joshi, U. R.
Kalkote, T. Ravindranathan, Tetrahedron: Asymmetry 1999, 10,
891 – 899; c)C. R. Johnson, S. J. Bis, Tetrahedron Lett. 1992, 33,
7287 – 7290.
[12] a)J. Partridge, G. A. Hutcheon, B. D. Moore, P. J. Halling, J. Am.
Chem. Soc. 1996, 118, 12873 – 12877; b)G. A. Hutcheon, M. C.
Parker, A. James, B. D. Moore, Chem. Commun. 1997, 931 – 932;
c)K. Kawashiro, H. Sugahara, S. Sugiyama, H. Hayashi,
Biotechnol. Bioeng. 1997, 53, 26 – 31.
[13] a)M. Bakker, A. S. Spruijt, F. van Rantwijk, R. A. Sheldon,
Tetrahedron: Asymmetry 2000, 11, 1801 – 1808; b)Y. Kita, Y.
Takebe, K. Murata, T. Naka, S. Akai, J. Org. Chem. 2000, 65, 83 –
88.
Angew. Chem. Int. Ed. 2004, 43, 5519 –5523
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5523