Influence of ester chain length, enzyme, and physical parameters on lipase-catalysed hydrolyses of meso-oxiranedimethanol esters. Part 2

Article abstract of DOI:10.1016/S0957-4166(00)00276-7

A range of meso-oxiranedimethanol diesters were prepared and subjected to hydrolysis by a selection of microbial and porcine-derived lipases. The best chain length was then subjected to a wider range of lipases. The physical reaction parameters were furth

Full text of DOI:10.1016/S0957-4166(00)00276-7

Tetrahedron: Asymmetry 11 (2000) 3197±3209
In¯uence of ester chain length, enzyme, and physical
parameters on lipase-catalysed hydrolyses of
meso-oxiranedimethanol esters. Part 2
,y
Jonathan D. Moseley* and James Staunton
University Chemical Laboratories, Lens®eld Road, Cambridge, CB2 1EW, UK
Received 28 June 2000; accepted 11 July 2000
Abstract
A range of meso-oxiranedimethanol diesters were prepared and subjected to hydrolysis by a selection of
microbial and porcine-derived lipases. The best chain length was then subjected to a wider range of lipases.
The physical reaction parameters were further investigated for the best enzyme/substrate combination. The
results revealed the e€ects of ester chain length, choice of enzyme, and physical reaction parameters on the
enzymic hydrolysis. Using a combination of the these three variables, it is demonstrated that this is a
versatile approach for the optimisation of an enzymic hydrolysis for a given substrate. Additionally, we
also report on an ecient, multigram synthesis of oxiranedimethanol diesters from cheap achiral
precursors, their successful enzymic hydrolyses to alcohols in good chemical yield and e.e., and their
conversion to synthetically useful aldehydes in high yield. Both enantiomeric series can readily be accessed
from a single enzymic hydrolysis result. # 2000 Elsevier Science Ltd. All rights reserved.
1
. Introduction
In our initial communication, we reported on the desymmetrisation of meso-diesters 3 by
1
enzymic hydrolysis, and demonstrated that the ester side chain can be tailored to improve the
enantiomeric excess (e.e.) of the resulting alcohols 4. Since both ester side chains are discarded in
subsequent manipulations, there is considerable scope for further optimisation of this hydrolysis
by choice of the ester or other parameters. We now report on the extension of our investigations
on the preferred butyrate diester 3d which was subjected to hydrolysis by an extensive range of
enzymes. The physical parameters were then varied for the preferred substrate/enzyme combination.
*
Corresponding author. Tel: +44 117 938 5601; e-mail: jonathan.moseley@astrazeneca.com
Present address: AstraZeneca, Process Research and Development, Avlon Works, Severn Road, Hallen, Bristol BS10
ZE, UK.
y
7
0957-4166/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(00)00276-7

3
198
J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
Lastly, we detail the simple transformations required to generate further useful synthetic inter-
mediates, and access to both enantiomeric series.
2
. Results and discussion
2
.1. Preparation of substrates for enzymic hydrolysis
A range of meso-diesters 3a±i was readily prepared by esteri®cation of the known epoxide 2
derived by m-CPBA epoxidation of cis-butene-1,4-diol 1 adapted from the method of Schloss2
(Scheme 1). An alternative epoxidation method was also tried using m-CPBA in acetonitrile
instead of chloroform. Although this method gave a nominally higher yield of diol 2 (89% versus
3
4
6
0%), it performed poorly in the esteri®cation step giving a lower overall yield for the two steps.
Schloss' epoxidation method was adequate on moderate scale-up and so this was used in pref-
erence. Surprisingly, the majority of the simple alkyl esters were unknown at the time of this
work, although some more exotic derivatives had been prepared by Schloss.
Scheme 1. R=a, Me; b, Et; c, iPr; d, nPr; e, nPen; f, tBu; g, tAmyl; h, nHep; i Ph
2
.2. Variation of enzymes
In our earlier communication, we reported the e€ects on the hydrolysis of varying the side
chain substitution for the propionate 3b, iso-butyrate 3c, n-butyrate 3d and hexanoate 3e diesters.
1
5
A limited range of enzymes was used, including the more common LPF, PLE and PPL. In
summary, it was concluded that the butyrate ester has an ideal chain length for many lipases,
although the results for the propionate ester were also worth considering. The iso-butyrate and
1
hexanoate esters gave only moderate results. Other workers have achieved high e.e.s with the
,7
acetate diester 3a.6
We therefore chose to assay the butyrate diester 3d against a wider range of commercially
5
available microbial and porcine derived lipases and esterases. The e.e.s were determined from the
8
speci®c rotations of the puri®ed alcohol 4d and correlated against Mosher's ester formation of a
number of representative samples, the integration of the 19F NMR signals agreeing within þ1%
of the speci®c rotation ®gures. The results are collected in Table 1, and where the same enzyme
9
has been used, are generally similar to those of Marples. The results varied from e€ectively
racemic hydrolyses (entries 1±3 and 5) through to those displaying good enantiotopic
discrimination in up to 85% e.e. (entries 8 and 11). The majority of the lipases, and all of the best

J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
3199
Table 1
Enzymic hydrolysis of butyrate diester 3d with di€erent enzymes
results, yielded the (+)-(2S,3R) enantiomer 4d in good chemical yield (62±76%). Only one lipase
gave a fair result for the opposite enantiomer (LPR, 56%), and this was by far the slowest
hydrolysis (entry 6). PLE was also slow and as expected did not cope well with a four carbon
10
chain. The best results were achieved for LRD and PPL (Fluka). Interestingly LRD, which is an
extracellular fungal lipase that selectively hydrolyses triacylglycerols, has itself been the subject of
mutagenesis studies to alter its side chain length selectivity over its regioselectivity.11 Further
work with this lipase was not continued due to its higher cost (£500/g against £0.30/g for PPL)
and given the consistently good results achieved with PPL, we continued the investigation of
physical parameters with this lipase alone.
2
.3. Variation of physical parameters
We then investigated the e€ects of several physical parameters on the hydrolysis of butyrate
diester 3d speci®cally with PPL (Fluka) including pH, temperature, concentration, enzyme
6
equivalents, and the use of phosphate bu€ers and co-solvent (Table 2). Grandjean had
conducted a limited study of the ®rst of these parameters and found that generally for the
butyrate (and acetate) diesters, a moderate lowering of both the pH (from 7.0 to 6.0) and
ꢀ
temperature (from 20 to 4 C) resulted in a 5±10% improvement in e.e. This was not the case in
our hands (entries 2, 3 and 8) with no signi®cant improvement seen, although the results were
e€ectively no worse, being within 5%.12 The use of dilute phosphate bu€er did however result in
a detectably worse result at two pHs (entries 4 and 5). This may have been due to the higher ionic
strength of the aqueous solvent forcing a less than ideal performance from the lipophilic lipase.

3
200
J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
Table 2
Enzymic hydrolysis of butyrate diester 3d with PPL under di€erent reaction conditions
Doubling the concentration (entry 6) resulted in a predictable halving of the reaction time, but
ꢀ
with no bene®t on the e.e. Increasing the temperature to 40 C (entry 7) also reduced the reaction
time, but the resulting e.e. was lower.
A co-solvent of methanol (20% v/v with water) was tried with PPL and the hexanoate diester
ꢀ
3
e, a low melting solid (m.p. 28±29 C). However, an identical result in e.e. of 65% was achieved
1
in similar reaction time (4 hours) to the control preparation without methanol, but with the yield
being lower (25% versus 42%), probably due to mechanical loss on work-up in the solubilising
liquors. The use of a co-solvent therefore made no di€erence to the overall reaction pro®le.
Portionwise addition was also used due to concern over enzyme inactivation through either
poisoning or mechanical degradation (due to grinding by the magnetic stirrer),13 but no
improvement over the standard result was achieved (entry 9), indicating that this was not a
factor. Finally, as a control, re-screening with the sample of Fluka PPL after 12 months storage
ꢀ
at 4 C but without special precautions against moisture etc. gave a similar result (entry 10) also
indicating no degradation over this period.
2
.4. Derivatisation of hydrolysis products
We converted the individual alcohols (+)-4c, (+)-4d and (+)-4e to the common alcohol 6 and
aldehyde 7 via the protected esters 5c±e, respectively. This was achieved by TBDMS-protection,
mild alkaline hydrolysis and Swern oxidation14 (Scheme 2). This provided a further check on e.e.
values by comparison of the [ꢀ] 's across the series for the common products 6 and 7, and also
D
with other literature sources with which they were found to be in good agreement.15 This facile
reaction sequence could be achieved in 89% yield over the three steps without the need for
chromatography in a semi-telescoped fashion on the crude oils (conducted on the butyrate half
ester (+)-4d). The largest scale hydrolysis undertaken on this work was 37 mmol, but there is no

J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
3201
Scheme 2.
obvious reason why this could not have been scaled up much further, as there appeared to be no
agitation or physical form issues. Larger scale Swern oxidations would also not have been
expected to present a problem. Finally, although the highest e.e.s achieved in this work were 80±84%,
Grandjean has shown that e.e.s of these levels can be improved to e€ectively enantiomeric purity
in one crystallisation for 4d, irrespective of the potential improvement that might be achieved
with other later intermediates i.e. 5, 6 and 7.
Alcohol 6 is e€ectively the enantiopode of the (+)-4 alcohols but with a di€erent protecting
group, thus demonstrating that both enantiomeric series can be readily accessed from a single
hydrolysis result, irrespective of the result of the enzymic hydrolysis. Direct oxidation of (+)-4d to
the useful aldehyde 8 proved unsatisfactory under a wide range of mild oxidation conditions
(
i.e. TPAP, MnO , Mo€at, Corey±Kim, PCC/PDC all gave <25% yields) (Scheme 3). The best
2
16
yields for the isolated material were achieved using the Dess±Martin reagent (50% impure) or
Swern oxidation (45%, also impure). The product appears to be unstable to chromatographic
puri®cation since Pale17 has generated the aldehyde 8 by modi®ed Swern oxidation and reacted it
in situ in higher overall yields.
Scheme 3.
3
. Conclusions
In summary, we conclude our study of the in¯uence of side chain substitution on the result of
enzymic hydrolyses, showing that the side chain can be tailored to give potentially the best result
for a given molecule. Investigation of both a range of enzymes and physical parameters provides
a further option for optimisation. If the reverse trans-esteri®cation process is considered as
well,1
8,19
then there are multiple options for optimisation, making this approach ideal for process
development work, perhaps best conducted under a factorial experimental design (FED)
approach.
We also report on an ecient, multigram synthesis of oxiranedimethanol diesters 3 from cheap
achiral precursors; their successful enzymic hydrolyses in good chemical yield and e.e. to the

3
202
J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
alcohols 4; and demonstrate facile access to both enantiomeric series through conversion to
synthetically useful derivatives 6 and 7. These manipulations were carried out on a multigram scale,
without puri®cation of the intermediates, to give the valuable chiral aldehyde 7 in 90% over three steps.
4
. Experimental
4
.1. General methods
Optical rotations were measured on a Perkin±Elmer 241 polarimeter using the sodium D line at
5
point apparatus and are uncorrected. Infra red spectra were run on Perkin±Elmer 297 or 1310
È
89 nm. Melting points were determined on a Reichert hot-stage apparatus or a Buchi 510 melting
^1
spectrophotometers between 4000 and 600 cm and only readily assigned absorbances above
1500 cm^ are listed. The following abbreviations are used: s=strong, m=medium, w=weak,
1
ꢁ
1
b=broad. H NMR spectra were recorded on Bruker WM-250 (250 MHz), AC-250 (250 MHz),
or AM-400 (400 MHz) spectrometers. Chemical shifts are given in ppm relative to tetra-
methylsilane at ꢁ=0, with tetramethysilane or the residual non-deuterated solvent signal as
internal standard. In addition to the conventional abbreviations for multiplicity, the following
have been used for ®rst order spectra: qu=quintet, sx=sextet, sp=septet, m=multiplet,
b=broad. 13C NMR spectra were recorded on Bruker AM-400 (100.6 MHz) or AC-250 (62.9
MHz) spectrometers and were proton decoupled. Chemical shifts are given in ppm relative to
tetramethylsilane at ꢁ=0. An `attached proton test' was routinely carried out on all samples to
aid in the assignment of spectra. 19F NMR spectra were recorded on a Bruker WM-250 spectro-
meter at 235 MHz with proton decoupling, and are given in ppm relative to CFCl as external
3
+
standard. Low resolution electron impact (EI ) mass spectra were recorded on Kratos MS902 (ex
+
A.E.I.) or MS 890 spectrometers. Fast atom bombardment (FAB ) mass spectra (carrier gas Xe)
+
were recorded on Kratos MS 50 or MS 890 spectrometers. High resolution EI mass spectra were
recorded on a Kratos MS 30 spectrometer. Elemental analyses (C and H) were performed by the
`
in-house' analytical service.
Analytical TLC was carried out on commercially prepared plates coated with 0.25 mm of self-
indicating Merck Kieselgel 60 F254 and were developed using a 5% solution of phosphomolybdic
acid in ethanol, in addition to visualisation by UV and I . Preparative ¯ash silica chromatog-
2
raphy was performed using Merck Kieselgel 60 (230±400 mesh). Enzymes were purchased from
5
Fluka or Sigma as noted. All other reagents and solvents were puri®ed and dried as necessary
according to standard procedures.
4
.2. cis-2,3-Epoxybutane-1,4-diol 22
cis-2-Butene-1,4-diol 1 (5.7 ml, 69 mmol) and m-CPBA (21.7 g of 50±55% technical grade sta-
ꢀ
bilised with water, 69 mmol) were stirred in a solution of ether (80 ml) at 5 C in the dark for 3
hours. The resulting white precipitate was collected by ®ltration, washed with ice-cold ether (20
ꢀ
ml maximum) and dried in vacuo to give a hygroscopic white powder (4.45 g, 62%). Mp 50±52 C
2
ꢀ
lit., 60%, mp 50±52 C). H NMR (250 MHz, D O) 3.86 (2H, dd, JGem=12.6, JVic=2.9 Hz,
2
1
(
+
2
1
4
ÂOCHH), 3.60 (2H, dd, JGem=12.6, JVic=2.9 Hz, 2ÂOCHH), 3.32 (2H, m, epoxide); MS (EI )
+
04 (M , 5), 87 (18), 73 (66), 61 (100). Anal. calcd for C H O : C, 45.97; H, 7.63. Found: C,
4
8
3
6.14; H, 7.76.

J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
3203
4
.3. General preparation of diesters 3a±i
The following procedure, exempli®ed by the preparation of 3d below, was applied to epoxy diol
with the appropriate acid chloride to generate in an analogous fashion the following diesters,
2
for which physical and spectroscopic data is supplied.
4
.3.1. cis-2,3-Epoxybutane-1,4-dibutyrate 3d
Epoxy diol 2 (2.0 g, 19.2 mmol), pyridine (7.5 ml, 90 mmol) and DMAP (trace) were dissolved
ꢀ
in dichloromethane (40 ml) and cooled to ^5 C. Butyryl chloride (6.0 ml, 57.7 mmol) was added
dropwise over 15 min keeping the temperature below 5 C. The reaction was allowed to warm to
ꢀ
rt overnight and quenched with ice (5 g). After stirring for 2±3 hours at rt, the solution was
diluted with dichlromethane (40 ml), washed sequentially with 1.0 M H SO (3Â40 ml), saturated
2
4
NaHCO (2Â40 ml), dried (Na SO ) and concentrated to an orange oil (7 g). The crude oil was
3
2
4
puri®ed by ¯ash silica gel chromatography in 4:1 hexane:ether (R 0.28) to yield the title
f
ꢀ
compound as a yellow oil or a low melting solid (4.1 g, 87%). Mp 13±14 C (neat); IR (CHCl )
3
2
2
^
1 1
900m, 2840m, 1737s cm ; H NMR (250 MHz, CDCl ) 4.33 (2H, dd, JGem=12.3, JVic=4.0 Hz,
3
ÂOCHH), 4.10 (2H, dd, JGem=12.3, JVic=6.8 Hz, 2ÂOCHH), 3.27 (2H, m, epoxide), 2.33 (4H,
t, J=7.4 Hz, 2ÂCH CO), 1.65 (4H, sx, J=7.4 Hz, 2ÂCH CH ), 0.94 (6H, t, J=7.4 Hz, 2ÂCH );
2
2
3
3
1
3
C NMR (100.6 MHz, CDCl ) 173.23 (CˆO), 61.97 (CH O), 53.35 (epoxide), 35.84 (CH CO),
3
2
2
+
1
for C H O : C, 58.99; H, 8.27. Found: C, 59.10; H, 8.38.
8.30 (CH CH ), 13.59 (CH ); MS (EI ) 245 (M+1, 3%), 229 (15), 201 (45), 157 (100). Anal. calcd
2
3
3
12
20
5
4
.3.2. cis-2,3-Epoxybutane-1,4-diacetate 3a
A pale yellow solid, 87%. R 0.48 (3:2 hexane:EtOAc); mp 40±42 C (neat); IR (CHCl ) 1743s
ꢀ
f
3
^
1 1
cm ; H NMR (400 MHz, CDCl ) 4.31 (2H, dd, JGem=12.3, JVic=3.8 Hz, 2ÂOCHH), 4.08 (2H,
3
13
dd, JGem=12.3, JVic=6.7 Hz, 2ÂOCHH), 3.27 (2H, m, epoxide), 2.09 (6H, s, 2ÂMe); C NMR
(
100.6 MHz, CDCl ) 170.59 (CˆO), 62.16 (CH O), 53.26 (epoxide), 35.84 (CH CO), 20.67 (CH );
3
2
2
3
+
MS (FAB ) 189 (M+1, 100%), 154 (50), 137 (65), 129 (25), 107 (18), 87 (20), 69 (18). Anal. calcd
for C H O : C, 51.05; H, 6.30. Found: C, 51.05; H, 6.44.
8
12
5
4
.3.3. cis-2,3-Epoxybutane-1,4-dipropionate 3b
A pale yellow solid, 75%. R 0.24 (2:1 hexane:ether); mp 24±26 C (neat); IR (CHCl ) 1739s
ꢀ
f
3
^
1 1
cm ; H NMR (250 MHz, CDCl ) 4.35 (2H, dd, JGem=12.3, JVic=4.0 Hz, 2ÂOCHH), 4.11 (2H,
3
dd, JGem=12.3, JVic=6.8 Hz, 2ÂOCHH), 3.27±3.33 (2H, m, epoxide), 2.40 (4H, q, J=7.5 Hz,
13
2ÂCH CH ), 1.17 (6H, t, J=7.5 Hz, 2ÂCH CH ); C NMR (100.6 MHz, CDCl ) 173.96 (CˆO),
6
1
2
3
2
3
3
+
2.01 (CH O), 53.28 (epoxide), 27.21 (CH CO), 8.89 (CH ); MS (FAB ) 217 (M+1, 100%), 173 (8),
2
2
3
43 (34), 69 (18). Anal. calcd for C H O : C, 55.32; H, 7.62. Found: C, 55.53; H, 7.47.
10
16
5
4
.3.4. cis-2,3-Epoxybutane-1,4-diisobutyrate 3c
A pale yellow oil, 78%. R 0.28 (4:1 hexane:ether); IR (CHCl ) 2900w, 1734s cm ; H NMR
^1 1
f
3
(
400 MHz, CDCl ) 4.31 (2H, dd, JGem=12.3, JVic=4.1 Hz, 2ÂOCHH), 4.11 (2H, dd,
3
JGem=12.3, JVic=6.5 Hz, 2ÂOCHH), 3.28 (2H, m, epoxide), 2.59 (2H, sp, J=6.9 Hz,
1
3
2
(
ÂMe CH), 1.16 (12H, d, J=6.9 Hz, 2Â(CH ) CH); C NMR (100.6 MHz, CDCl ) 176.74
2
3 2
3
+ +
CˆO), 62.08 (CH O), 53.32 (epoxide), 33.81 (Me CH), 18.90 (CH ); MS (EI ) 244 (M , 2%),
2
2
3
2
Found: C, 59.22; H, 8.48.
01 (20), 157 (100), 143 (100), 114 (25), 71 (100). Anal. calcd for C H O : C, 58.99; H, 8.27.
12 20 5

3
204
J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
4
.3.5. cis-2,3-Epoxybutane-1,4-dihexanoate 3e
A yellow solid, 72%. R 0.21 (5:1 hexane:ether); mp 28.5±29.5 C; IR (CHCl ) 2930m, 2865m,
ꢀ
f
3
^
1 1
1
4
736s cm ; H NMR (250 MHz, CDCl ) 4.32 (2H, dd, JGem=12.4, JVic=3.9 Hz, 2ÂOCHH),
3
.09 (2H, dd, JGem=12.4, JVic=6.8 Hz, 2ÂOCHH), 3.27 (2H, m, epoxide), 2.34 (4H, t, J=7.5
Hz, 2ÂCH CO), 1.63 (4H, qu, J=7.4 Hz, 2ÂCH CH CO), 1.29 (8H, m, 2ÂCH CH ), 0.88 (6H,
2
2
2
2
2
13
t, J=6.7 Hz, 2ÂCH ); C NMR (62.9 MHz, CDCl ) 173.42 (CˆO), 61.948 (CH O), 53.36 (epoxide),
3
3
2
+
3
3.97, 31.22, 24.51 and 22.25 (4ÂCH ), 13.84 (CH ); MS (FAB ) 301 (M+1, 100%), 185 (17), 154
2
3
(82), 137 (78), 136 (76). Anal. calcd for C H O : C, 64.07; H, 9.48. Found: C, 63.96; H, 9.41.
16 28 5
4
.3.6. cis-2,3-Epoxybutane-1,4-di-tert-butyrate 3f
A pale yellow oil, 55%. R 0.34 (4:1 hexane:ether); IR (CHCl ) 2988m, 2973m, 2874m, 1732s
f
3
^
1 1
cm ; H NMR (250 MHz, CDCl ) 4.31 (2H, dd, JGem=12.3, JVic=4.1 Hz, 2ÂOCHH), 4.10 (2H,
3
13
dd, JGem=12.3, JVic=6.7 Hz, 2ÂOCHH), 3.27 (2H, m, epoxide), 1.21 (18H, s, 2ÂtBu); C NMR
(
62.9 MHz, CDCl ) 178.10 (CˆO), 62.14 (CH O), 53.17 (epoxide), 38.66 (CMe ), 27.02 (CH );
3
2
3
3
+
MS (EI ) 273 (M+1, 57%), 171 (77), 157 (53), 116 (25), 101 (20), 85 (51), 69 (37), 57 (100). Anal.
calcd for C H O : C, 61.83; H, 9.06. Found: C, 61.73; H, 8.90.
14
24
5
4
.3.7. cis-2,3-Epoxybutane-1,4-di-tert-butylacetate 3g
A yellow oil, 46%. R 0.32 (4:1 hexane:ether). Anal. calcd for C H O : C, 63.71; H, 9.46.
Found: C, 63.96; H, 9.41.
f
16 28
5
4
.3.8. cis-2,3-Epoxybutane-1,4-dioctanoate 3h
A white solid, 56%. R 0.23 (4:1 hexane:ether); mp 47±48 C; IR (CHCl ) 2920s, 2860m, 1735s
ꢀ
f
3
^
1 1
cm ; H NMR (250 MHz, CDCl ) 4.33 (2H, dd, JGem=12.3, JVic=4.0 Hz, 2ÂOCHH), 4.09 (2H,
3
dd, JGem=12.3, JVic=6.8 Hz, 2ÂOCHH), 3.27 (2H, m, epoxide), 2.34 (4H, t, J=7.5 Hz,
2
ÂCH CO), 1.62 (4H, qu, J=7.4 Hz, 2ÂCH CH CO), 1.27 (16H, m, 2Â(CH ) ), 0.86 (6H, t,
2
2
2
2 4
13
J=6.7 Hz, 2ÂCH ); C NMR (100.6 MHz, CDCl ) 173.44 (CˆO), 61.99 (CH O), 53.35
3
3
2
+
(
epoxide), 34.00 (CH CO), 31.60, 29.03, 28.87, 24.82 and 22.56 (5ÂCH ), 14.03 (CH ); MS (EI )
2
2
3
3
calcd for C H O : C, 67.57; H, 10.22. Found: C, 67.36; H, 10.20.
57 (M+1, 2%), 285 (14), 213 (19), 199 (100), 188 (13), 141 (16), 127 (100), 99 (11), 83 (99). Anal.
20
36
5
4
.3.9. cis-2,3-Epoxybutane-1,4-dibenzoate 3i
A white powder, 63%. R 0.18 (4:1 hexane:ether); mp 118±119 C (neat). Anal. calcd for
C H O : C, 69.18; H, 4.98. Found: C, 69.21; H, 5.17.
ꢀ
f
1
8
16
5
4
.4. General method for Mosher's ester formation8
Samples (20 mg) of the enzyme hydrolysis products (alcohols 4b±e) were converted to their
Mosher's ester derivatives as follows: The alcohol (0.15 mmol) was stirred for between 3 and 24
hours with DMAP (18 mg, 0.15 mmol), Et N (100 ml) and (+)-(S)-MTPA-Cl (30 ml) in dichloro-
3
Ê
methane (0.5 ml) dried with a few 3 A molecular sieves. The reaction was quenched with 3-di-
methylaminopropylamine (50 ml) and applied directly to a small ¯ash silica gel chromatography
column, eluting with EtOAc±hexane mixtures (typical values, R 0.28 in 4:1 hexane:EtOAc), taking
care to collect all high-running materials. The resulting combined fractions were assayed by
f
19
F
NMR spectroscopy to determine the e.e. from the relative integration of the CF signal. F NMR
19
3
(
235 MHz, CDCl ) ^72.17 and ^72.22 (varying ratios).
3

J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
3205
4
.5. General procedure for enzymic hydrolyses
The reactions were carried out in an appropriately sized wide-necked beaker on the open
bench. Rapid stirring was achieved by a ¯at magnetic stirrer bar. A pH meter electrode and a
thermometer were immersed in the solution at all times. Careful positioning of these monitors
was necessary to avoid the liquid's vortex, and also to avoid splashing, which over the long
reaction times could result in substantial mechanical loss of solution volume and hence reagents
and product. Glass-distilled water, adjusted to appropriate pH with either a few drops of 1.0 M
NaOH or 0.1 M KH PO solutions after the addition of the appropriate diester 3b±e, was used.
2
4
In a few cases a 0.1 M K HPO bu€ered solution adjusted to pH 7.0 was used. On the 5.0 mmol
2
4
scale, the following quantities of enzymes were used: microbial lipases (Fluka), 40 mg except LPF
and LRD, 10 mg; PPL (Fluka), 50 mg; CCL and PPL (Sigma), 100 mg; PLE (Sigma), 3 ml.
The reactions started with the addition of the enzymes to the rapidly stirred suspension. A 1.0
M NaOH solution was added by burette to maintain the pH at close to or below the appropriate
pH (usually pH 7.0). On addition of 1 equivalent of alkali, the reaction mixture was quickly
extracted into an organic solvent to halt the reaction. The crude concentrates were then puri®ed
by ¯ash silica gel chromatography. One standard reaction is given below for the dibutyrate ester
3
KH PO adjusted to the required pH with a few drops of 1.0 M NaOH. The portionwise addition
d as an example. The hydrolyses run in phosphate bu€er (Table 2, entries 4 and 5) used 0.1 M
2
4
reaction (Table 2, entry 9) was run at 37 mmol, and eight 40 mg portions of PPL were added at
half hour intervals.
4
.5.1. (+)-(2S,3R)-4-Butanoyloxy-2,3-epoxybutan-1-ol 4d
A suspension of diester 3d (1.22 g, 5.0 mmol) and PPL (Fluka, 50 mg) were vigourously stirred
at rt in glass-distilled water (40 ml) adjusted to pH 7.0. A 1.0 M NaOH solution (5.0 ml, 5.0
mmol) was titrated into the reaction mixture to maintain the pH at or just below 7.0 until nearly
the full equivalent of NaOH had been consumed. The aqueous solution was promptly extracted
with EtOAc (6Â30 ml). The combined organic extracts were washed once with a saturated brine
solution (30 ml), dried (Na SO ) and concentrated to dryness (763 mg). The crude oil was puri®ed
2
4
by ¯ash silica gel chromatography in 2:1 ether:hexane (R 0.22) to yield the title compound as a
f
2
D
3
pale yellow oil or a low melting solid (627 mg, 73%). ‰ꢀŠ +15.1 (c 1.03, CHCl , 84% e.e.); mp
3
ꢀ ꢀ
6
20±21 C (neat) (lit. 20þ3 C, pentane±CH Cl ); IR (CHCl ) 3200±3600bs, 2940m, 2870m, 1725s
cm ; H NMR (400 MHz, CDCl ) 4.32 (1H, dd, JGem=12.4, JVic=5.3 Hz) and 4.14 (1H, dd,
2
2
3
^
1 1
3
JGem=12.4, JVic=5.6 Hz, CO CH ), 3.82 (1H, dd, JGem=12.4, JVic=5.3 Hz) and 3.78 (1H, dd,
2
2
JGem=12.7, JVic=5.4 Hz, CH OH), 3.18±3.27 (2H, m, epoxide), 2.34 (2H, t, J=7.4 Hz, CH CO),
2
2
1
3
2
(
(
.25 (ꢁ1H, bs, OH), 1.66 (2H, sx, J=7.4 Hz, CH CH ), 0.95 (3H, t, J=7.4 Hz, CH ); C NMR
3 2
2
3
3
100.6 MHz, CDCl ) 173.80 (CˆO), 61.60 and 60.15 (2ÂCH O), 56.14 (C-2), 53.77 (C-3), 35.86
+
CH CO), 18.27 (CH CH ), 13.55 (CH CH ); MS (FAB ) 175 (M+1, 100). Anal. calcd for
2
2
3
2
3
C H O : C, 54.88; H, 8.01. Found: C, 55.15; H, 8.12.
8
14
4
4
.5.2. (+)-(2S,3R)-4-iso-Butanoyloxy-2,3-epoxybutan-1-ol 4c
23
A pale yellow oil. ‰ꢀŠ +14.4 (c 1.04, CHCl , 25% e.e.); R 0.14 (2:1 ether:hexane); oil; IR
D
CHCl ) 3500bm, 2930w, 2870w, 1720s cm ; H NMR (250 MHz, CDCl ) 4.32 (1H, dd,
3
f
^1
1
3 3
(
JGem=12.3, JVic=5.3 Hz) and 4.14 (1H, dd, JGem=12.2, JVic=5.5 Hz, CO CH ), 3.88 (2H, d,
2
2
J=4.5 Hz, CH OH), 3.20 (2H, m, epoxide), 2.55 (1H, sp, J=7.0 Hz, (CH ) CH), 2.30 (1H, bs,
2
3 2
13
OH), 1.70 (6H, d, J=7.0 Hz, (CH ) CH); C NMR (100.6 MHz, CDCl ) 177.33 (CˆO), 61.54
3
2
3

3
206
J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
+
and 60.19 (2ÂCH O), 56.09 (C-2), 53.32 (C-3), 33.67 (Me CH), 18.90 (CH ); MS (EI ) 175 (M+1,
2
2
3
4
Found: C, 54.61; H, 7.85.
), 161 (62), 143 (85), 131 (100), 89 (100), 71 (100). Anal. calcd for C H O : C, 54.88; H, 8.01.
8 14 4
4
.5.3. (+)-(2S,3R)-4-Hexanoyloxy-2,3-epoxybutan-1-ol 4e
23
A low melting yellow solid. ‰ꢀŠ +7.5 (c 0.59, CH Cl , 65% e.e.); R 0.31 (2:1 ether:hexane); mp
D
2
2
f
ꢀ
8±29 C (neat); IR (CHCl ) 3400±3650bm, 2920m, 2860m, 1730s cm ; H NMR (250 MHz,
3
^1 1
2
CDCl ) 4.34 (1H, dd, JGem=12.3, JVic=5.4 Hz) and 4.15 (1H, dd, JGem=12.2, JVic=5.6 Hz,
3
CO CH ), 3.85 (1H, dd, JGem=12.5, JVic=5.5 Hz) and 3.79 (1H, dd, JGem=12.5, JVic=5.2 Hz,
2
2
CH OH), 3.19±3.29 (2H, m, epoxide), 2.45±2.65 (1H, bs, OH), 2.37 (2H, t, J=7.5 Hz, CH CO),
2
2
13
1
.65 (2H, qu, J=7.4 Hz, CH CH CO), 1.31 (4H, m, CH CH ), 0.91 (3H, t, J=6.4 Hz, CH );
3
C
2
2
2
2
NMR (100.6 MHz, CDCl ) 174.02 (CˆO), 61.56 and 60.15 (2ÂCH O), 56.14 (C-2), 53.77 (C-3),
3
2
+
3
4.02 (CH CO), 31.12, 24.51, 22.26 (CH ), 13.87 (CH ); MS (FAB ) 203 (M+1, 39), 176 (10), 150
2
2
3
(
19). Anal. calcd for C H O : C, 59.37; H, 8.99. Found: C, 59.64; H, 8.91.
10 18 4
4
.6. General preparation of silyl esters 5c±e
The following procedure for the preparation of silyl esters 5c±e is exempli®ed below for 5d;
only physical and spectroscopic data are provided for the other examples.
4
.6.1. (+)-(2R,3S)-4-(tert-Butyldimethylsilyloxy)-1-butanoloxy-2,3-butan-1-ol 5d
Alcohol 4d (4.84g, 27.8 mmol) was stirred with TBDMS-Cl (5.02g, 33.3 mmol) and imidazole
(3.78g, 55.6 mmol) in DMF (70 ml) for 30 min. The mixture was diluted with hexane (150 ml),
washed with a 5% NaHCO solution (75 ml) and separated. The aqueous layer was further
3
extracted with hexane (4Â80 ml) and the combined organic portions dried and concentrated to a
colourless oil (7.77 g, 97%). The product was often used without further puri®cation. However,
for analytical purposes it could be puri®ed by ¯ash silica chromatography in 9:1 hexane:ether (R
f
1
D
8
^1 1
0.25). ‰ꢀŠ +8.5 (c 1.07, CH Cl , 83% e.e.); oil; IR (CHCl ) 2957s, 2932s, 2858m, 1733s cm ; H
NMR (250 MHz, CDCl ) 4.34 (1H, dd, JGem=12.3, JVic=3.6 Hz) and 4.05 (1H, dd, JGem=12.3,
2
2
3
3
JVic=7.0 Hz, CH CO), 3.81 (1H, dd, JGem=11.9, JVic=4.8 Hz) and 3.75 (1H, dd, JGem=11.8,
2
JVic=5.3 Hz, CH OSi), 3.14±3.26 (2H, m, epoxide), 2.33 (2H, t, J=7.4 Hz, CH CO), 1.66 (2H,
2
2
sx, J=7.4 Hz, CH CH ), 0.95 (3H, t, J=7.4 Hz, CH CH ), 0.89 (9H, s, tBu), 0.08 (3H, s) and
2
3
2
3
1
3
0
5
(
2
.07 (3H, s, Me Si); C NMR (62.9 MHz, CDCl ) 173.39 (CˆO), 62.50 and 61.38 (2ÂCH O),
2
3
2
6.46 (C-2), 53.60 (C-3), 35.95 (CH CO), 25.82 (Me C), 18.36 (CH CH ), 18.27 (Me C), 13.63
2
CH CH ), ^5.30 and ^5.42 (Me Si); MS (EI ) 289 (M+1, 100%), 273 (64), 257 (16), 245 (22),
31 (100), 143 (100). HRMS (EI ) C H O Si requires M+H, 289.1835. Found: M+H,
3
2
3
3
+
2
3
2
+
14
29
4
2
89.1826.
4
.6.2. (+)-(2R,3S)-4-(tert-Butyldimethylsilyloxy)-1-iso-butanoyl-2,3-epoxybutan-1-ol 5c
2
5
Data for the crude material: R 0.36 (4:1 hexane:ether); oil; ‰ꢀŠ +4.6 (c 1.40, CH Cl , 25%
f
D
e.e.); IR (CHCl ) 2982w, 2955w, 2931m, 2857m, 1732s cm ; H NMR (250 MHz, CDCl ) 4.34
2
2
^
1 1
3
3
(1H, dd, JGem=12.3, JVic=3.6 Hz) and 4.05 (1H, dd, JGem=12.3, JVic=6.9 Hz, CH OSi), 3.82
2
(1H, dd, JGem=11.8, JVic=4.7 Hz) and 3.75 (1H, dd, JGem=11.8, JVic=5.3 Hz, CH OCO), 3.15±
2
3
0
.27 (2H, m, epoxide), 2.60 (1H, sp, J=7.0 Hz, (CH ) CH), 1.18 (6H, J=7.0 Hz, (CH ) CH),
3 2 3 2
13
.89 (9H, s,tBu), 0.08 (3H, s) and 0.07 (3H, s, Me Si); C NMR (62.9 MHz, CDCl ) 176.77
2
3
(
CˆO), 62.47 and 61.27 (2ÂCH O), 56.28 (C-2), 53.49 (C-3), 33.73 (Me CH), 25.69 (Me C),
2
2
3

J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
3207
+
1
8.82 and 18.79 (Me CH), 18.14 (Me C), ^5.43 and ^5.55 (Me Si); MS (EI ) 289 (M+1, 72%),
2
3
2
2
31 (48), 185 (36), 143 (100).
4
.6.3. (+)-(2R,3S)-4-(tert-Butyldimethylsilyloxy)1-hexanoyl-2,3-epoxybutan-1-ol 5e
Data for the crude material: R 0.39 (4:1 hexane:ether); oil; [ꢀ] n.d.; IR (CHCl ) 2930m,
f
D
3
^
1 1
2
860m, 1735 cm ; H NMR (250 MHz, CDCl ) 4.34 (1H, dd, JGem=12.3, JVic=3.6 Hz) and 4.05
3
(1H, dd, JGem=12.3, JVic=6.9 Hz, CH OSi), 3.82 (1H, dd, JGem=11.8, JVic=4.7 Hz) and 3.75
2
(1H, dd, JGem=11.8, JVic=5.3 Hz, CH OCO), 3.17±3.28 (2H, m, epoxide), 2.35 (2H, t, J=7.4
Hz, CH CO), 1.64 (2H, qu, J=7.4 Hz, CH CH CO), 1.32 (4H, m, CH CH ), 0.89 (3H, t, J=6.5
2
2
2
2
2
2
Hz, CH ), (9H, s, tBu), 0.08 (3H, s) and 0.07 (3H, s, Me Si).
3
2
4
.7. (+)-(2R,3S)-4-(tert-Butyldimethylsilyloxy)-2,3-epoxybutan-1-ol 6
The crude silyl ester 5d (1.0 g, 3.47 mmol) was stirred in a sealed ¯ask with concentrated `0.88'
aqueous ammonia solution (520 ml, 10.3 mmol) and methanol (20 ml) at rt for 3 days or at
re¯ux for 8 hours. The crude concentrate was puri®ed by ¯ash silica chromatography in 1:1
1
D
9
hexane:ether to yield the title compound as a light yellow oil (701 mg, 92%). R 0.28. ‰ꢀŠ +9.9 (c
f
^1 1
1.40, CH Cl , 84% e.e.); oil; IR (CHCl ) 3300±3600bm, 2940s, 2865s cm ; H NMR (250 MHz,
2 2 3
CDCl ) 3.90 (1H, dd, JGem=11.8, JVic=5.5 Hz, CHHOSi), 3.73±3.80 (2H, m, CH OH), 3.71 (1H,
3
2
dd, JGem=11.8, JVic=5.4 Hz, CHHOSi), 3.15±3.25 (2H, m, epoxide), 2.29 (1H, bt, J=6.5 Hz,
OH), 0.89 (9H, s, tBu), 0.08 (3H, s) and 0.07 (3H, s, Me Si); C NMR (100.6 MHz, CDCl ) 61.58
13
2
3
and 60.83 (2ÂCH O), 56.29 and 55.96 (epoxide), 25.79 (Me C), 18.23 (Me C), ^5.23 and ^5.47
2
3
3
+
(
Me Si); MS (FAB ) 219 (M+1, 46%). Anal. calcd for C H O Si: C, 54.97; H, 10.26. Found: C,
2
10 22
3
5
4.99; H, 10.17.
4
.7.1. Compound 6 from hydrolsis of silyl ester 5c
Silyl ester 5c was hydrolysed according to the preceding method described for 5d above to yield
23
the silyl-protected alcohol 6 in 90%. ‰ꢀŠ +6.7 (c 2.18, CH Cl , 25% e.e.). Other data for 6 as
D
2
2
described in Section 4.7.
4
.7.2. Compound 6 from hydrolsis of silyl ester 5e
Silyl ester 5e was hydrolysed according to the preceding method described for 5d above to yield
23
the silyl-protected alcohol 6 in 53%. ‰ꢀŠ +7.9 (c 1.04, CH Cl , 65% e.e.). Other data for 6 as
D
2
2
described in Section 4.7.
15
4
.8. (^)-(2R,3S)-4-(tert-Butyldimethylsilyloxy)-2,3-epoxybutan-1-al 7
Ê
All solutions were dried over 4 A molecular sieves for 3±4 hours before addition to the main
reaction vessel. Oxalyl chloride (2.2 ml, 25.4 mmol) in dichloromethane (150 ml) was cooled to
ꢀ
^70 C and DMSO (3.9 ml, 55.4 mmol) in dichloromethane (15 ml) was added dropwise over 10
min. The solution was stirred for a further 10 min before the alcohol 6 (5.05 g, 23.1 mmol) in
dichloromethane (25 ml) was likewise added over 10 min. The reaction was quenched after 1 hour
by the careful addition of triethylamine (16.1 ml, 115.5 mmol), maintaining the temperature
ꢀ
below ^70 C and further stirring for 25 min. The mixture was warmed to rt and water (100 ml)
added with stirring for 30 min. HCl (1.0 M, 100 ml) was added and the mixture separated. The
organic layer was washed consecutively with 100 ml portions of water, 5% NaHCO , and water,
3

3
208
J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
2
1
dried and concentrated to a pale yellow oil (4.51 g, 92%), ‰ꢀŠ (crude) ^59.8 (c 0.70, CH Cl ).
D
The material was often used without further puri®cation. Analytically pure material could be
2
2
1
8
obtained by ¯ash silica chromatography in 3:1 hexane:ether (R 0.29). ‰ꢀŠ (pure) ^73.0 (c 0.93,
f
D
1 1
^
CH Cl , 84% e.e.); oil; IR (CHCl ) 2910m, 2845s, 2735w, 1715s cm ; H NMR (250 MHz,
2
2
3
CDCl ) 9.49 (1H d, J=4.4 Hz, CHO), 4.01 (1H, dd, JGem=12.3, JVic=2.6 Hz) and 3.93 (1H, dd,
3
JGem=12.3, J =3.5 Hz, CH OSi), 3.36±3.45 (2H, m, epoxide), 0.87 (9H, s, tBu), 0.06 (6H, s
2
Vic
1
3
Me Si); C NMR (62.9 MHz, CDCl ) 197.99 (CˆO), 60.23 (C-2), 59.87 (C-4), 57.69 (C-3), 25.78
2
3
+
(
Me C), 18.31 (Me C), ^5.48 (Me Si); MS (FAB ) 217 (M+1, 10), 215 (24), 201 (26), 187 (11), 159
3
3
2
(
37), 75 (100). Anal. calcd for C H O Si: C, 55.50; H, 9.42. Found: C, 55.50; H, 9.33.
10
20
3
4
.9. (2R,3R)-4-Butyryloxy-2,3-epoxybutan-1-al 8
Alcohol 4d (523 mg, 3.0 mmol) was oxidised by Swern14 oxidation according to the method of
Section 4.8 to generate the title compound as a crude oil, contaminated with ꢁ25% unreacted
1
alcohol as determined by H NMR integration and TLC. Estimated yield of title compound by
1
H NMR, 45%. Data for crude product: R 0.28 (1:1 hexane:ether); [ꢀ] n.d.; oil; IR (CHCl )
f
D
3
^
1 1
1
JGem=12.6, JVic=4.3 Hz) and 4.27 (1H, dd, JGem=12.6, JVic=5.6 Hz, CH O), 3.51 (1H, ddd,
738bs, 1715bs cm ; H NMR (400 MHz, CDCl ) 9.46 (1H, d, J=4.8 Hz, CHO), 4.34 (1H, dd,
3
2
J2 =4.8, J =4.3 and 5.6 Hz, H-3), 3.45 (1H, t, J=4.8 Hz, H-2), 2.29 (2H, t, J=7.4 Hz,
�3
2
3�4
CH CO), 1.62 (2H, sx, J=7.4 Hz, CH CH ), 0.91 (3H, t, J=7.4 Hz, CH ); C NMR (100.6
13
2
3
3
MHz, CDCl ) 196.78 (CHO), 172.99 (CˆO), 60.64 (CH O), 57.19 (C-2), 55.89 (C-3), 35.70
3
2
(
CH CO), 18.20 (CH CH ), 13.55 (CH ).
2
2
3
3
Acknowledgements
We are grateful to the SERC for generous ®nance (to J.D.M.). J.D.M. thanks David R. Lewis
for helpful discussions at the time of this work, former colleagues at Merck Sharp and Dohme
(Terlings Park) for facilitating this research, and colleagues at AstraZeneca (Avlon) for helpful
comments on the preparation of this manuscript.
References
1
2
3
4
. Moseley, J. D.; Staunton, J. Tetrahedron: Asymmetry 1998, 9, 3619±3623.
. Schloss, J. V.; Hartman, F. C. Bioorg. Chem. 1980, 9, 217±226.
. Nelson, W. L.; Freeman, D. S.; Vincenzi, F. F. J. Med. Chem. 1976, 19, 153±158.
. Aside from poor performance in the esteri®cation step, the inferior quality was typi®ed by its less crystalline
ꢀ
ꢀ
appearance and lower melting point (47±49 C in our hands by Nelson's method against 50±52 C using Schloss'
ꢀ
method; Nelson reported 44±47 C). This may be accounted for by the fact that the material was isolated from
water by freeze drying, probably resulting in an amorphous or less crystalline form, and possibly retained water.
. All enzymes are commercially available from Fluka (F) except CCL and PLE available from Sigma (S), and PPL
from both. Abbreviations used: CCL, lipase from Candida cylindracea; LAN, lipase from Aspergillus niger; LCL,
lipase from Candida lipolytica; LMJ, lipase from Mucor javanicus; LPF, lipase from Pseudomonas ¯uorescens;
LPR, lipase from Penicillium roqueforti; LRA, lipase from Rhizopus arrhizus; LRD, lipase from Rhizopus delemar;
LRN, lipase from Rhizopus niveus; PLE, pig liver esterase; PPL porcine pancreatic lipase.
5
6. Grandjean, D.; Pale, P.; Chuche, J. Tetrahedron Lett. 1991, 32, 3043±3046.
7. Brevet, J.-L.; Mori, K. Synthesis 1992, 1007±1012.

J. D. Moseley, J. Staunton / Tetrahedron: Asymmetry 11 (2000) 3197±3209
3209
8
9
. Dale, J. A.; Dull, D. L.; Mosher, H. S. J. Org. Chem. 1969, 34, 2543±2549.
. Corser, D. A.; Dart, R. K.; Marples, B. A. J. Chem. Res. (S) 1992, 66±67.
1
1
1
0. The Sigma derived PPL was moderately inferior, perhaps due to the source or method of isolation.
1. Joerger, R. D.; Haas, M. J. Lipids 1994, 29, 377.
2. In general, our results were 5±10% lower in e.e. compared to Grandjean although they mirrored their results in
all other respects. We o€er three possibilities for the observed di€erences; (i) di€erences in enzyme source (Fluka
versus Sigma); (ii) di€erences in the w/w quantity of enzyme used; and (iii) our preparations were run to 100%
conversion, theirs to a maximum of 90%.
6
1
1
1
1
1
1
1
3. Kvittingen, L.; Sjursnes, B.; Halling, P.; Anthonsen, T. Tetrahedron 1992, 48, 5259±5264.
4. Omura, K.; Swern, D. Tetrahedron 1978, 34, 1651±1660.
5. Roush, W. R.; Straub, J. A.; van Nieuwenhze, M. S. J. Org. Chem. 1991, 56, 1636±1648.
6. Dess, D. B.; Martin, J. C. J. Org.Chem. 1983, 48, 4155±4156.
7. Grandjean, D.; Pale, P.; Chuche, J. Tetrahedron Lett. 1992, 33, 5355±5358.
È
8. Vanttinen, E.; Kanerva, L. T. Tetrahedron: Asymmetry 1992, 3, 1529±1532.
9. Trincone, A.; Pagnotta, E. Tetrahedron: Asymmetry 1996, 7, 2773±2774.