Catalytic electronic activation: Indirect Sharpless asymmetric epoxidation of enals

Article abstract of DOI:10.1016/S0957-4166(02)00100-3

The asymmetric epoxidation of allylic cyanohydrins using the Sharpless kinetic resolution (SKR) reaction was explored. The SKR methodology was extended to enals via in-situ conversion into the corresponding cyanohydrin. The prospect of achieving a dynamic

Full text of DOI:10.1016/S0957-4166(02)00100-3

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 13 (2002) 317–323
Catalytic electronic activation: indirect Sharpless asymmetric
epoxidation of enals
Phillip J. Black, Kerry Jenkins† and Jonathan M. J. Williams*
Department of Chemistry, University of Bath, Bath BA2 7AY, UK
Received 8 February 2002; accepted 21 February 2002
Abstract—The asymmetric epoxidation of allylic cyanohydrins using the Sharpless kinetic resolution (SKR) reaction was explored.
The SKR methodology was extended to enals via in-situ conversion into the corresponding cyanohydrin. The prospect of
achieving a dynamic Sharpless kinetic resolution was investigated and the enantioselectivity of the SKR of (E)-2-hydroxy-3-
methyl-4-phenyl-3-butenenitrile 1 studied. Reversible hydrocyanation of a-methyl cinnamaldehyde was developed and a one-pot
hydrocyanation/epoxidation process achieved. © 2002 Elsevier Science Ltd. All rights reserved.
1. Introduction
2. Results and discussion
With increasing worldwide interest in domino and cas-
cade reactions, we have recently proposed an indirect
addition of nucleophiles to allylic alcohols using the
concept of catalytic electronic activation.1 In this cur-
rent study we envisaged the prospect of applying transi-
tion metal-mediated epoxidation chemistry to an
electron-deficient a,b-unsaturated aldehyde by tempo-
rary conversion into the electron-rich and functionally
compatible allylic cyanohydrin. This had particular
appeal to us because of the potential application to the
Sharpless asymmetric epoxidation reaction (Scheme 1).2
2.1. Ti(OiPr)4-mediated epoxidation of the allylic
cyanohydrin 1
As anticipated, cyanohydrin 1 underwent smooth epox-
idation with Ti(OiPr)4 and tert-butyl hydroperoxide to
give the epoxycyanohydrin
diastereomers (Scheme 2). The major diastereomer has
2 as a mixture of
Electron-deficient alkenes are less susceptible to elec-
trophilic epoxidation than their electron rich analogues.
Furthermore, transition-metal mediated epoxidations
can benefit from neighbouring group participation by
suitably placed coordinating groups. With these facts in
mind, the appeal of the proposed scheme is apparent.
Hydrocyanation of the electron-poor enal substrate
should furnish the ‘activated’ racemic allylic cyanohy-
drin.3 If the hydrocyanation can be made reversible and
rapid with respect to the rate of epoxidation of the
mismatched cyanohydrin enantiomer, one would expect
to effect a dynamic kinetic resolution with respect to
the cyanohydrin stereogenic centre.4–6 Subsequent dehy-
drocyanation should furnish the epoxy aldehyde and
potentially enable catalytic turnover of cyanide.
Scheme 1. Catalytic electronic activation: indirect Sharpless
asymmetric epoxidation of enals.
* Corresponding author. Tel.: +44 (0) 1225 826138; fax: +44 (0) 1225
826 231; e-mail: j.m.j.williams@bath.ac.uk
† Present address: Millennium Pharmaceuticals Limited, Granta Park,
Great Abington, Cambridge CB1 6ET, UK.
Scheme 2. Reagents and conditions: 1 equiv. Ti(OiPr)4, 2
equiv. tBuOOH, CH2Cl2 (0.10 M), −16°C, 83%, 65% d.e.
0957-4166/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(02)00100-3

318
P. J. Black et al. / Tetrahedron: Asymmetry 13 (2002) 317–323
been tentatively assigned as syn in analogy with related
epoxidation reactions of secondary allylic alcohols.
has been unambiguously assigned by derivatisation to
the known epoxide of a-methylcinnamyl alcohol. At
48% conversion the epoxide had 87% e.e., whilst at 61%
conversion the residual cyanohydrin 1 was essentially
enantiomerically pure (99% e.e.). Monitoring the
stereoselectivity of the reaction against time showed
that initially (i.e. prior to 50% conversion), predomi-
nantly syn-2 with (R)-epoxide configuration was
formed. Allowing the reaction to proceed beyond 50%
conversion resulted in reduction of enantiomeric excess
with respect to the epoxide and dramatic slowing of the
reaction rate (Scheme 3, Graphs 1 and 2).
2.2. Sharpless kinetic resolution of racemic allylic
cyanohydrin ( )-1
A novel Sharpless kinetic resolution7 of cyanohydrin 1
using (+)-diethyl tartrate (DET) led to the preferential
formation of the (2R)-epoxide (Scheme 3). The absolute
stereochemistry of the cyanohydrin has been inferred
using the predictive model for enantioselectivity pro-
posed by Sharpless.1 However, the epoxide stereocentre
Scheme 3. Reagents and conditions: 1 equiv. Ti(OiPr)4, 1 equiv. (+)-DET, 2 equiv. tBuOOH, CH2Cl2 (0.1 M), −16°C.
Graph 1. Enantiomeric excess of epoxide (syn-2+anti-2) versus conversion.
Graph 2. Conversion of cyanohydrin 1 versus time.

P. J. Black et al. / Tetrahedron: Asymmetry 13 (2002) 317–323
319
Table 1. Hydrocyanation of a-methyl cinnamaldehyde in the presence of titanium tetraisopropoxide
Entry
ACH (mol%)
Base (mol%)
Ti(OiPr)4 (mol%)
Yield, 1 h (%)
Yield, 24 h (%)
Yield, 72 h (%)
1
2
3
4
5
6
7
100
100
100
100
100
100
500
–
100
–
–
100
100
100
100
B1
18
3
9
B1
9
9
19
15
41
38
32
62
36
17
14
41
39
31
63
iPr2NEt (20)
NaCN (20)
iPr2NEt (20)
NaCN (20)
NaCN (10)
NaCN (20)
32
2.3. Hydrocyanation of the enal
2.5. One-pot hydrocyanation/epoxidation
Acetone cyanohydrin (ACH) was found to be a conve-
nient source of ‘HCN’ for the hydrocyanation of a-methyl
cinnamaldehyde in the presence of Ti(OiPr)4 (Table 1,
entry 1). Furthermore, the rate of hydrocyanation was
enhanced when a suitable base (e.g. R3N or NaCN) was
present. The presence of both base and Ti(OiPr)4 (entries
4 and 5) appeared to have a beneficial synergistic effect
over using base alone (entries 2 and 3) (Scheme 4).
Attempts to effect the domino asymmetric epoxidation
of a-methyl cinnamaldehyde were moderately successful
and it was possible to convert a-methyl cinnamalde-
hyde indirectly into epoxide 2 through an SKR reaction
on the hydrocyanated starting material generated in
situ (Scheme 6, Table 2).
Scheme 6. One-pot hydrocyanation/epoxidation of a-methyl
cinnamaldehyde.
Scheme 4. Reagents and conditions: 1 equiv. Ti(OiPr)4, base,
ACH, CH2Cl2 (0.25 M).
The initial results obtained using Methods A and B
were very encouraging;10 however, the complete cata-
lytic cycle (Scheme 1) was not realised (i.e. no
epoxyaldehyde 3 observed), although the principle of
substrate activation has clearly been demonstrated.
More disappointing were the relatively modest enan-
tiomeric excesses (36–38%) obtained. This is thought to
reflect the lower enantioselectivity of the SKR at mod-
erately high temperatures (7 and 22°C c.f. −16°C)b and
may also reflect a slow rate of cyanohydrin racemisa-
tion relative to the rate of epoxidation. This could then
result in a competing (less enantioselective) epoxidation
of the mismatched enantiomer.
2.4. Racemisation and reversible hydrocyanation
In order to realise the full potential of the proposed system
we needed to ensure that hydrocyanation was reversible.
We chose to investigate this through using enantiomer-
ically enriched cyanohydrin 1, which was prepared by
asymmetric trimethylsilylcyanation of a-methyl cin-
namaldehyde.8 The absolute configuration of the major
enantiomer has not been determined, but has been
tentatively assigned R on the basis of analogy with similar
examples.8 The enantiomerically enriched cyanohydrin 1
(26% e.e.) was subjected to the optimised hydrocyanation
conditions (Ti(OiPr)4, ACH and base) (Scheme 5) and
subsequent racemisation occurred. This confirmed that
the addition of HCN was indeed a reversible exchange.9
Racemic cyanohydrin ( )-1 was obtained in less than 30
min using iPr2NEt (0.2 equiv.), whilst racemisation using
NaCN (0.2 equiv.) took considerably longer (22% e.e.
after 4 h; racemic within 24 h).
An attempted one-pot hydrocyanation/epoxidation
reaction using NaCN (Method C) gave epoxycyano-
hydrin 2 in 66% isolated yield11 (3:4 mixture of
diastereomers). However, further analysis of this reac-
tion revealed that after only 5 min 2 was formed as a
racemic mixture (30% conversion). This suggested that
the reaction is accelerated via an achiral reaction mani-
fold and furthermore, the analysis also revealed an
initial build up of cyanohydrin 1, which subsequently
reduced with concomitant formation of epoxycyano-
hydrin 2. This would appear to rule out a pathway
involving direct (nucleophilic) epoxidation of the enal.
Additionally, no epoxidation is observed when a-
methyl cinnamaldehyde is treated with NaCN (1.0
equiv.)/tBuOOH (2.0 equiv.) or NaCN (1.0 equiv.)/
tBuOOH (2.0 equiv.)/Ti(OiPr)4 (1.0 equiv.). However,
Scheme 5. Reagents and conditions: 1 equiv. Ti(OiPr)4, 0.2
equiv. NaCN or iPr2NEt, 1 equiv. ACH, CH2Cl2 (0.25 M).

320
P. J. Black et al. / Tetrahedron: Asymmetry 13 (2002) 317–323
Table 2. Reagents and conditions for the one-pot hydrocyanation/epoxidation of a-methyl cinnamaldehyde
Methoda
ACH
Base (mol%)
Ti(OiPr)4
(+)-DET
Temp.
t (h)
Yield cyanohydrin 2/e.e. (R)-epoxide
(mol%)
(mol%)
(mol%)
(°C)
A
B
C
D
200
200
500
–
EtNiPr2 (200)
EtNiPr2 (200)
NaCN (100)
NaCN (100)
100
100
100
200
110
110
110
120
7b
22
22
192
47
23
46% (38%, 37% d.e.)
39% (36%, 40% d.e.)
66% (0%, 7% d.e.)
22
25
50%c (0%, 0% d.e.); 25% epoxyaldehyde 3;
25% starting material
a All the experiments were carried out in CH2Cl2 (0.25 M) using 2 equiv. tBuOOH.
b No product was formed using these conditions at −16°C; even after 312 h.
c Conversion determined by 1H NMR on the isolated mixture.
hydrocyanation/epoxidation of a-methyl cinnamalde-
hyde was achieved under the latter conditions when
(+)-diethyl tartrate was present (Method D). H NMR
analysis of the crude product mixture revealed epoxy
cyanohydrin 2 (50%) epoxy aldehyde 3 (25%) along
with unreacted starting material (25%). Unfortunately
the epoxide of 2 was again racemic. It is unclear why
diethyl tartrate is crucial for this reaction, but it is
proposed that it could function as a proton source in
the absence of acetone cyanohydrin.
red spectra were recorded on a Perkin–Elmer 1600
series FTIR spectrometer as thinly dispersed films
(from CH2Cl2) between sodium chloride plates. Low
resolution (EI+/FAB+) and accurate mass spectra were
obtained using a Finnigan MAT 8340 mass spectrome-
ter. Thin-layer chromatography was carried out using
Macherey–Nagal, Polygram® SIL G/UV254 pre-coated
plastic backed plates. The plates were visualised using
ultraviolet light (u=254 nm) and/or KMnO4 solution.
High-pressure column chromatography was carried out
using a Thermo Separation Products Spectraseries P200
HPLC with a UV100 detector and Chromjet integrator.
Chiralcel OD (0.46 cm Diameter×25 cm) chiral station-
ary phase was used.
1
3. Conclusion
The first Sharpless kinetic resolution of an allylic
cyanohydrin and conditions for reversible hydrocyana-
tion of an enal that are compatible with the SKR
reaction have been demonstrated. Catalytic electronic
activation of the enal substrate has enabled a successful
asymmetric epoxidation of an enal in a one-pot
sequence. However, the poor rates of hydrocyanation
at the lower temperatures, which are optimal for asym-
metric epoxidation, do need to be addressed. The cata-
lytic turnover of the nucleophile (in this case cyanide)
also remains a desirable goal.
4.2. Ti(OiPr)4-mediated epoxidation of (E)-2-hydroxy-
3-methyl-4-phenyl-3-butenenitrile 1
A solution of cyanohydrin 1 (702 mg, 4.053 mmol) in
,
CH2Cl2 (40 mL) with 4 A molecular sieves (404 mg),
was stirred at −5°C under a nitrogen atmosphere. Tita-
nium tetraisopropoxide (1.20 mL, 4.065 mmol) was
then added. After 5 min stirring, a solution of tert-butyl
hydroperoxide (5 M in decane, 1.62 mL, 8.1 mmol) was
added and the reaction mixture was transferred to a
freezer at −16°C. The reaction was quenched after 24 h
using aq. FeSO4/tartaric acid solution (250 g dm−3/100
g dm−3, 1.0 mL) and was filtered through a pad of
Celite®/Na2SO4. The product was purified on pre-aci-
dified SiO2 column using gradient elution (petroleum
ether/diethyl ether, 20:1–7:3) to give 2 as a blue-grey oil
(635 mg, 83%, 65% d.e. syn-diastereomer assumed
major).
4. Experimental
4.1. General
All glassware was oven dried and cooled in a desiccator
(P2O5 desiccant) prior to use. Dichloromethane
(CH2Cl2) was distilled from calcium hydride and stored
,
over 4 A molecular sieves. Commercially supplied
4.3. syn-2-Hydroxy-3-methyl-3-oxiranyl-4-phenyl
butanenitrile 2
reagents were used as supplied. Flash column chro-
matography was carried out using Merck 60 silica gel
(70–240 mm). In order to prevent dehydrocyanation, all
cyanohydrin products were purified on pre-acidified
(petroleum ether/diethyl ether/acetic acid, 20:1:1) silica
gel columns. This is particularly important for
cyanohydrin 1. 1H, 13C and 19F NMR spectra were
recorded on either a Jeol GX270 or EX400 spectrome-
ter. Chemical shifts are reported downfield in parts per
million (ppm) from a tetramethylsilane reference. Infra-
wmax (liquid film): 3424, 2256, 1063 cm−1; lH (270 MHz,
CDCl3): 1.28 (s, 3H, CH3), 3.67 (s br, 1H, OH), 4.34 (s,
1H, PhCH(O)C), 4.60 (s, 1H, CH(OH)CN), 7.26–7.39
(m, 5H, Ph); lC (67.8 MHz, CDCl3): 12.5 (CH3), 61.0
(CH(OH)CN), 62.8 (PhCH(-O-)C(CH3)CH), 64.7
(PhCH(-O-)C(CH3)CH), 117.2 (ArC), 126.3 (2 ArCH),
128.2 (p-ArCH), 128.3 (2 ArCH), 133.4 (CH(OH)CN);
+
+
’
’
MS (70 eV): m/z (%) 189 Da (M , 1%), 162 (M −

P. J. Black et al. / Tetrahedron: Asymmetry 13 (2002) 317–323
321
HCN, 4%), 43 (100%): acc. mass EI+ found, 189.0790
Da, C11H11NO2 pred., 189.07897 Da; Rf=0.40
(petroleum ether/diethyl ether, 1:1).
in 1 (lH=6.14, 6.10) relative to the sum of the
cyanohydrin a-methine protons in 2 (lH=5.42, 5.57,
5.57, 5.60).
4.4. anti-2-Hydroxy-3-methyl-3-oxiranyl-4-phenyl
4.5.1. 1-(R)-Derived Moshers ester. lH (270 MHz,
CDCl3): 1.87 (s, 3H, CꢀCCH3), 3.62 (d, 3H, J=1.10
butanenitrile 2a
wmax (liquid film): 3424, 2256, 1063 cm−1; lH NMR (270
MHz, CDCl3): 1.27 (s, 3H, CH3), 3.53 (t, 1H, J=9.0
Hz, CHOH), 4.29 (s, 1H, PhCH(O)C), 4.63 (d, 1H,
J=9.0 Hz, CH(OH)CN), 7.26–7.39 (m, 5H, Ph); lC
(67.8 MHz, CDCl3): 12.4 (CH3), 60.0 (CH(OH)CN),
62.7 (PhCH(-O-)C(CH3)CH), 64.5 (PhCH(-O-)C-
(CH3)CH), 117.2 (ArC), 126.4 (2 ArCH), 128.3 (p-
ArCH), 128.4 (2 ArCH), 133.5 (CH(OH)CN); MS (70
Hz,
OCH3),
6.14
(d,
1H,
J=0.73
Hz,
CꢀCCH(CN)O2C), 6.83 (s, 1H, PhCHꢀC), 7.23–7.53
(m, 5H, Ph); lF (376 MHz, CDCl3): −72.09 (s, CF3).
4.5.2. 1-(S)-derived Moshers ester. lH (270 MHz,
CDCl3): 1.86 (s, 3H, CꢀCCH3), 3.56 (d, 3H, J=1.10
Hz,
OCH3),
6.10
(d,
1H,
J=0.74
Hz,
CꢀCCH(CN)O2C), 6.88 (s, 1H, PhCHꢀC), 7.23–7.53
(m, 5H, Ph); lF (376 MHz, CDCl3): −72.22 (s, CF3).
+
+
’
’
eV): m/z (%) 189 Da (M , 1%), 162 (M −HCN, 4%),
43 (100%): acc. mass EI+ found, 189.0790 Da,
C11H11NO2 pred., 189.07897 Da; Rf=0.48 (petroleum
ether/diethyl ether, 1:1).
4.5.3. 2-syn-(R)-Derived Moshers ester. lH (270 MHz,
CDCl3): 1.13 (s, 3H, CH3), 3.60 (d, 3H, J=1.22 Hz,
OCH3), 4.09 (s, 1H, PhCH(-O-)C), 5.57 (s, 1H,
CCH(CN)O2C), 7.18–7.52 (m, 5H, Ph); lF (376 MHz,
CDCl3): −72.05 (s, CF3).
4.5. Sharpless kinetic resolution of ( )-(E)-2-hydroxy-3-
methyl-4-phenyl-3-butenenitrile 1
4.5.4. 2-syn-(S)-Derived Moshers ester. lH (270 MHz,
CDCl3): 1.27 (s, 3H, CH3), 3.60 (d, 3H, J=1.22 Hz,
OCH3), 4.20 (s, 1H, PhCH(-O-)C), 5.60 (s, 1H,
CCH(CN)O2C), 7.18–7.52 (m, 5H, Ph); lF (376 MHz,
CDCl3): −72.14 (s, CF3).
Titanium tetraisopropoxide (1.19 mL, 4.031 mmol) was
added to a cooled (−16°C) nitrogen purged solution of
(+)-diethyl tartrate (0.69 mL, 4.032 mmol) in CH2Cl2
,
(32 mL) containing 4 A molecular sieves (368 mg).
After stirring the solution for 10 min a solution of
cyanohydrin 1 (698 mg, 4.030 mmol) in CH2Cl2 (8 mL)
was added dropwise. After a stirring the mixture for a
further 10 min, a solution of tert-butyl hydroperoxide
(5 M in decane, 1.61 mL, 8.05 mmol) was added.
(i) Aliquot quench and work-up: For the purposes of
monitoring the kinetic resolution we removed
aliquots (1.5 mL) of reaction mixture which were
immediately quenched in vigorously stirred CH2Cl2
(5 mL) containing aq. FeSO4/tartaric acid solution
(250 g dm−3/100 g dm−3, five drops). After 30 min the
mixture was filtered through Celite®/MgSO4 with
CH2Cl2 (2×5 mL) washings. The combined CH2Cl2
phase was extracted once with aq. FeSO4/tartaric
acid solution (250 g dm−3/100 g dm−3, 5 mL) and
dried over MgSO4. Removal of the solvent in vacuo
yielded the crude product mixture, which was taken
through to the next step without further purification.
(ii) Preparation of Moshers esters: A freshly prepared
solution of dicyclohexylcarbodiimide (0.165 mmol),
(R)-(−)-a-methoxy-a-(trifluoromethyl)phenylacetic
acid (0.165 mmol) and 4-(N,N-dimethylamino)-
pyridine (0.033 mmol) in CH2Cl2 (2.2 mL) was added
to the crude reaction mixture aliquots. After 24 h the
reactions were diluted with petrol (5 mL) and filtered
through a pipette containing Celite®. The crude mix-
ture was subjected to column chromatography on
pre-acidified SiO2 (petroleum ether/diethyl ether,
20:1–15:1).12 The enantiomeric ratio for 1 was deter-
mined by integration of the diastereomeric cyanohy-
drin a-methine protons (lH=6.14, 6.10) whilst the
enantiomeric and diastereomeric ratios for 2 were
determined by integration of the 4-oxyranyl methine
protons (lH=4.09, 4.17, 4.20, 4.25). The relative
conversion of 1 to 2 was determined by the sum of
the integration of the cyanohydrin a-methine protons
4.5.5. 2-anti-(R)-Derived Moshers ester. lH (270 MHz,
CDCl3): 1.24 (s, 3H, CH3), 3.56 (d, 3H, J=1.22 Hz,
OCH3), 4.25 (s, 1H, PhCH(-O-)C), 5.42 (s, 1H,
CCH(CN)O2C), 7.18–7.52 (m, 5H, Ph); lF (376 MHz,
CDCl3): −72.26 (s, CF3).
4.5.6. 2-anti-(S)-Derived Moshers ester. lH (270 MHz,
CDCl3): 1.26 (s, 3H, CH3), 3.56 (d, 3H, J=1.22 Hz,
OCH3), 4.17 (s, 1H, PhCH(-O-)C), 5.57 (s, 1H,
CCH(CN)O2C), 7.18–7.52 (m, 5H, Ph); lF (376 MHz,
CDCl3): −72.14 (s, CF3).
4.6. Hydrocyanation study–monitoring the hydrocyana-
tion of a-methyl cinnamaldehyde
4.6.1. General protocol. a-Methyl cinnamaldehyde
(0.070 mL, 0.501 mmol) was added to a stirred solu-
tion/suspension of base in CH2Cl2 (2.0 mL). The mix-
ture was purged with nitrogen and titanium
tetraisopropoxide (0.150 mL, 0.508 mmol) was added
where necessary. This was immediately followed by
addition of acetone cyanohydrin. To monitor the
hydrocyanation, aliquots (0.050 mL) of the reaction
mixture were taken and immediately quenched in
biphasic diethyl ether (1.0 mL)/aq. tartaric acid (10%
w/v, 0.20 mL) which were shaken vigorously for a few
seconds before being allowed to stand. A sample of
ethereal phase (0.010–0.020 mL) was removed and dis-
solved in hexane/iPrOH (9:1, 1.5 mL) containing
MgSO4. These solutions were subjected to HPLC anal-
ysis (Chiralcel OD column with hexane/iPrOH (9:1)
eluent, 1.00 cm3 min−1 flow rate, u=254 nm). a-Methyl
cinnamaldehyde tR 6.9 min (signal integration multi-
plied by a factor of 1.145 to give relative molarity with

322
P. J. Black et al. / Tetrahedron: Asymmetry 13 (2002) 317–323
respect to 1), 1-(S) tR 10.7 min, 1-(R) tR 11.6 min. See
allowed to stir at −16°C for 2 h prior to adding
tert-butyl hydroperoxide (5 M in decane, 0.200 mL,
1.00 mmol). The reaction mixture was then trans-
ferred to a refrigerator at 7°C. The reaction was
quenched after 8 days with aq. FeSO4/tartaric acid
solution (250 g dm−3/100 g dm−3, 0.3 mL) and was
filtered through a pad of Celite®/Na2SO4. The
organic phase was extracted with aq. FeSO4/tartaric
acid solution (250 g dm−3/100 g dm−3, 3×30 mL)
and dried over Na2SO4. The product was purified
on pre-acidified SiO2 using gradient elution
(petroleum/diethyl ether, 6:1–4:1). a-Methyl cin-
namaldehyde was recovered (16 mg, 22%). Epoxy-
cyanohydrin 2 was isolated as a 1:1 molar mixture
with ethyl diisopropyl tartrate (94 mg mixture; cal-
culated 43 mg 2, 46%, 37% d.e. assumed syn, 38%
e.e. (R)-epoxide determined by conversion to Mosh-
ers ester).
(ii) To a nitrogen purged flask containing NaCN
(123 mg, 2.510 mmol) was added a pre-formed solu-
tion of (+)-diethyl tartrate (0.47 mL, 2.747 mmol)
and titanium tetraisopropoxide (0.74 mL, 2.507
mmol) in CH2Cl2 (10 mL). To this stirred solution
was added a-methyl cinnamaldehyde (0.35 mL,
2.507 mmol) followed by acetone cyanohydrin (1.15
mL, 12.59 mmol) and then tert-butyl hydroperoxide
(5 M in decane, 1.00 mL, 5.00 mmol). The reaction
was quenched after 24 h with aq. FeSO4/tartaric
acid solution (250 g dm−3/100 g dm−3, 1.0 mL) and
filtered through a pad of Celite®/Na2SO4. The
organic phase was extracted with aq. FeSO4/tartaric
acid solution (250 g dm−3/100 g dm−3, 3×30 mL)
and dried over Na2SO4. The product was purified
on pre-acidified SiO2 using gradient elution
(petroleum ether/diethyl ether, 20:1–7:3). Epoxy-
cyanohydrin 2 was isolated as viscous oil (312 mg,
66%, 7% d.e. assumed anti, 0% e.e. epoxide deter-
mined by conversion to Moshers ester).
Table 1 for tabulated results.
4.6.2. (E)-2-Hydroxy-3-methyl-4-phenyl-3-butenenitrile
1. wmax (liquid film): 3419, 2248, 1600, 1491, 1447,
1014, 700 cm−1; lH (270 MHz, CDCl3): 2.04 (d, 3H,
J=1.3 Hz, CH3), 3.77 (s br, 1H, OH), 5.03 (d, 1H,
J=4.6 Hz, CH(OH)CN), 6.80 (s, 1H, PhCHꢀC), 7.26–
7.41 (m, 5H, Ph); lC (67.8 MHz, CDCl3): 14.1 (CH3),
66.8 (CH(OH)CN), 118.4 (ArC), 127.5 (ArCH), 128.3
(2 ArCH), 128.9 (2 ArCH), 130.0 (PhCHꢀC), 131.6
(CꢀC(CH3)CH), 135.7 (CH(OH)CN); MS (70 eV): m/
+
+
’
’
z (%) 173 Da (M , 11%), 145 (M −HCN, 100%):
acc. mass EI+ found, 173.08309 Da, C11H11NO pred.,
173.08406 Da; Rf=0.50 (petroleum ether/diethyl ether,
1:1).
4.7. Racemisation of (E)-2(R)-hydroxy-3-methyl-4-
phenyl-3-butenenitrile 1
4.7.1. General protocol. To a nitrogen-purged vessel
containing base (0.2 equiv.) and acetone cyanohydrin
(1.0 equiv.) was added a solution of enantiomerically
enriched 1 (0.25 M in CH2Cl2, 1.0 equiv.) followed by
titanium tetraisopropoxide (1.0 equiv.). To monitor
the racemisation, aliquots (0.050 mL) of the reaction
mixture were taken and immediately quenched in
biphasic diethyl ether (1.0 mL)/aq. tartaric acid (10%
w/v, 0.20 mL) which were shaken vigorously for a few
seconds before being allowing to stand. A sample of
ethereal phase (0.010–0.020 mL) was removed and dis-
solved in hexane/iPrOH (9:1, 1.5 mL) containing
MgSO4. These solutions were subjected to HPLC anal-
ysis (vide supra).
4.7.2. Racemisation using iPr2NEt. As above, using
iPr2NEt (0.0140 mL, 0.080 mmol) and acetone
cyanohydrin (0.0370 mL, 0.405 mmol), 1 in CH2Cl2
(0.25 M, 1.60 mL, 0.400 mmol) and titanium tetra-
isopropoxide (0.120 mL, 0.406 mmol). A racemic mix-
ture was observed within 30 min (33% a-methyl
cinnamaldehyde formed).
(iii) To a nitrogen purged suspension of NaCN (25
mg, 0.510 mmol) in CH2Cl2 (2.0 mL) was added
titanium tetraisopropoxide (0.150 mL, 0.508 mmol)
and (+)-diethyl tartrate (0.10 mL, 0.584 mmol).
After stirring the solution for 10 min, a-methyl cin-
namaldehyde (0.070 mL, 0.501 mmol) was added
4.7.3. Racemisation using NaCN. As above, using
NaCN (5 mg, 0.102 mmol) and acetone cyanohydrin
(0.0460 mL, 0.504 mmol), 1 in CH2Cl2 (0.25 M, 2.00
mL, 0.500 mmol) and titanium tetraisopropoxide
(0.150 mL, 0.508 mmol). A racemic mixture was
observed within 24 h (22% a-methyl cinnamaldehyde
formed).
followed by tert-butyl hydroperoxide (5
M in
decane, 0.200 mL, 1.00 mmol). The reaction was
quenched after 25 h with aq. FeSO4/tartaric acid
solution (250 g dm−3/100 g dm−3, 0.2 mL) and
filtered through a pad of Celite®/Na2SO4. The
organic phase was extracted with aq. FeSO4/tartaric
acid solution (250 g dm−3/100 g dm−3, 3×30 mL)
and dried over Na2SO4. The product was eluted
4.8. One-pot hydrocyanation/asymmetric epoxidation of
a-methyl cinnamaldehyde
through
a pre-acidified (petroleum ether/diethyl
ether, 7:3) SiO2 column to yield clear oil (40 mg).
1H NMR revealed a mixture of epoxycyanohydrin
2/ epoxyaldehyde 3/a-methyl cinnamaldehyde
(48:26:26). This corresponds to epoxycyanohydrin 2
(21 mg, 22%, 1:1 d.r., <5% e.e. epoxide determined
by conversion to Moshers ester); epoxyaldehyde 3
(10 mg, 12%, e.e. not determined); a-methyl cin-
namaldehyde (9 mg, 12%). Further purification of 2
on a pre-acidified (petroleum ether/diethyl ether,
20:1–7:3) SiO2 column gave 2 (17 mg, 18%) as a
clear oil.
(i) To a cooled (−16°C), nitrogen purged solution of
(+)-diethyl tartrate (0.096 mL, 0.561 mmol) in
CH2Cl2 (2.0 mL) was added titanium tetra-
isopropoxide (0.150 mL, 0.508 mmol). After stirring
for 15 min, a-methyl cinnamaldehyde (0.070 mL,
0.501 mmol) was added, followed by iPr2NEt
(0.0170 mL, 0.098 mmol) and then acetone cyanohy-
drin (0.092 mL, 1.01 mmol). The mixture was

P. J. Black et al. / Tetrahedron: Asymmetry 13 (2002) 317–323
323
4.9. (E)-2-Methyl-3-phenyl oxiranecarbaldehyde 313
6. Ward, R. S. Tetrahedron 1995, 6, 1475–1490.
7. Gao, Y.; Hanson, R. M.; Klunder, J. M.; Ko, S. Y.;
Masamune, H.; Sharpless, K. B. J. Am. Chem. Soc. 1987,
109, 5765–5780.
8. Masahiko, H.; Matsuda, T.; Oguni, N. J. Chem. Soc.,
Perkin Trans. 1 1992, 3135–3140.
9. Racemisation under base catalysed asymmetric hydrocya-
nations has been previously reported: (a) Oku, J; Ito, N.;
Inoue, S. Makromol. Chem. 1982, 183, 579–586; (b) Jack-
son, W. R.; Jayatilake, G. S.; Matthews, B. R.; Wilshire,
C. Aust. J. Chem. 1988, 41, 203–213.
10. In this reaction, epoxycyanohydrin 2 was isolated as a 1:1
mixture with diisopropyl tartrate, a by-product from
transesterification under the reaction conditions.
11. No significant transesterification was observed in this
reaction; even at room temperature.
wmax (liquid film): 1730 (CꢀO), 890, 850 cm−1; lH (270
MHz, CDCl3): 1.22 (s, 3H, CH3), 4.31 (s, 1H,
PhCH(O)C), 7.26–7.43 (m, 5H, Ph), 9.10 (s, 1H, CHO);
lC (67.8 MHz, CDCl3): 9.2 (CH3), 60.4 (PhCH(-O-)C),
65.0 (CH(-O-)C(CH3)CHO), 126.5 (2 ArCH), 128.4 (2
ArCH), 128.5 (p-ArCH), 132.9 (ArC), 199.1 (CHO);
+
’
+
MS (70 eV): m/z (%) 162 Da (M , 55%), 91 (C7H7 ,
100%): acc. mass EI+ found, 162.068390 Da, C10H10O2
pred., 162.068080 Da; Rf=0.65 (petroleum ether/
diethyl ether, 1:1).
References
1. Black, P. J.; Harris, W.; Williams, J. M. J. Angew. Chem.,
Int. Ed. 2001, 113, 4607–4608.
2. Katsuki, T.; Sharpless, K. B. J. Am. Chem. Soc. 1980,
102, 5976–5978.
3. For a review on cyanohydrins: North, M. Synlett 1993,
807–820.
4. Caddick, S.; Jenkins, K. Chem. Soc. Rev. 1996, 447–456.
5. Noyori, R.; Tokunaga, M.; Kitamura, M. Bull. Chem.
Soc. Jpn. 1995, 68, 36–55.
12. TLC of the crude reaction mixtures revealed a compo-
nent consistent with remaining starting material but fur-
ther analysis showed this to be co-running mono Moshers
ester of diethyl tartrate, Rf=0.40 (petroleum ether/diethyl
ether, 1:1), which was easily removed by flash column
chromatography. No diester was observed.
13. Oshima, M.; Yamazaki, H.; Shimizu, I.; Nisar, M.; Tsuji,
J. J. Am. Chem. Soc. 1989, 111, 6280–6287.