Asymmetric epoxidation catalyzed by esters of α-hydroxy-8-oxabicyclo[3.2.1]octan-3-one

Article abstract of DOI:10.1016/S0957-4166(01)00495-5

Several esters of α-hydroxy-8-oxabicyclo[3.2.1]octan-3-one were prepared and tested as catalysts for alkene epoxidation by Oxone.

Full text of DOI:10.1016/S0957-4166(01)00495-5

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 12 (2001) 2779–2781
Asymmetric epoxidation catalyzed by esters of
a-hydroxy-8-oxabicyclo[3.2.1]octan-3-one
Alan Armstrong,a,* William O. Mossb and Jonathan R. Reevesa
aDepartment of Chemistry, Imperial College, London SW7 2AY, UK
bAstraZeneca R&D Charnwood, Loughborough, Leicestershire, UK
Received 30 October 2001; accepted 5 November 2001
Abstract—Several esters of a-hydroxy-8-oxabicyclo[3.2.1]octan-3-one were prepared and tested as catalysts for alkene epoxidation
by Oxone®. © 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction
2. Results and discussion
Asymmetric alkene epoxidation with Oxone® catalyzed
by chiral ketones, via dioxirane intermediates, is
providing extremely promising results.1 A major chal-
lenge is the development of active catalysts that do not
suffer decomposition by Baeyer–Villiger reaction.2 We
have identified bicyclo[3.2.1]octan-3-ones as a class of
conformationally well defined catalysts with low intrin-
sic tendency to undergo decomposition.3 Our initial
study indicated that a-fluoro-N-carbethoxytropinone 1
afforded high activity, allowing use at low loadings at
pH 7.5.3a Promising enantioselectivities were also
recorded.3a This work was followed by a study of the
corresponding oxabicyclic compound 2 which provided
higher enantioselectivity in the epoxidation of (E)-stil-
bene.3b Additionally, we found that replacement of the
a-fluoro substituent with an acetoxy group provided
further improvement.3b Herein, we report further stud-
ies on asymmetric epoxidation promoted by these
promising oxabicyclic ketones, focusing on the effect of
modifying the a-ester substituent.
Since a-acetoxyketone 4a gave promising enantioselec-
tivity in the epoxidation of (E)-stilbene,3b we decided to
use the parent alcohol 33b as a starting point for the
preparation of several ester derivatives 4.5 We were
particularly interested in aromatic esters in view of the
possibility of p–p interactions with aromatic alkene
substrates; accordingly, we prepared both electron-rich
and electron-poor esters with the substituents at either
the meta- or para-positions in order to probe their
effect on epoxidation enantioselectivities. The results
for the epoxidation of (E)-stilbene using these racemic
ketones (20 mol%) under the Yang4 Oxone®/CH3CN/
H2O system (pH ca. 7.5) are shown in Table 1. The
trifluoroacetate 4b showed low conversion (entry 2) due
to hydrolysis under the reaction conditions. All of the
aromatic esters examined afforded lower conversion
than the acetate 4a (entries 1, 3–7), although they
appeared to be stable under the reaction conditions.
Perhaps surprisingly, there appeared to be little differ-
ence between electron-poor and electron-rich aromatics.
O
a R=CH3
b R=CF3
c R=Ph
d R=p-NO2Ph
e R=m-NO2Ph
f R=p-MeOPh
g R=m-MeOPh
O
X
Y
O
R
1 X=NCO2Et, Y=F
2 X=O, Y=F
3 X=O, Y=OH
H
H
O
O
4
* Corresponding author. E-mail: a.armstrong@ic.ac.uk
0957-4166/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(01)00495-5

2780
A. Armstrong et al. / Tetrahedron: Asymmetry 12 (2001) 2779–2781
Table 1. Epoxidation of (E)-stilbene catalyzed by racemic
ketones 4a–ga
substitution patterns: (E)-stilbene, styrene and a-methyl-
styrene (Table 2). The last two compounds represent
terminal alkenes, which are still problematic for asym-
metric epoxidation.7 The esters were assumed to have the
same enantiomeric excess as the parent alcohol, and none
of the solid esters (4c, 4d and 4f) were recrystallized.
Table 2 includes an ‘e.e.max’ value, the expected product
e.e. with enantiomerically pure catalyst based on the
assumption that there is a linear relationship between
catalyst e.e. and product e.e. We have established that
this is indeed the case in the epoxidation of (E)-stilbene
catalyzed by fluoroketone 28a and by 4a.8b For the
epoxidation of (E)-stilbene, all of the aromatic esters
4c–g afforded lower e.e. than the acetate 4a (entries 1,
2–6). Amongst the aromatic esters, the nitroaromatics
provided higher e.e. than the methoxy analogues (com-
pare entries 3 and 4 to 5 and 6). A similar trend was noted
in the epoxidation of styrene. For a-methylstyrene,
selectivities were low in all cases but the methoxyaromat-
ics gave marginally better results than the other catalysts.
Entry
Ketone
R
Time (h)
Conversionb
1
2
3
4
5
6
7
4a
4b
4c
4d
4e
4f
CH3
CF3
Ph
p-NO2Ph
m-NO2Ph
p-MeOPh
m-MeOPh
24
24
24
24
24
24
24
85
3
36
39
46
53
33
4g
a Alkene (0.1 mmol), Oxone® (1.0 mmol KHSO5), ketone (0.02
mmol), NaHCO3 (1.55 mmol), CH3CN (1.5 ml), aq. Na2EDTA (1
ml of 0.4 mmol dm−3 solution), rt.
b Estimated by integration of the crude 1H NMR spectrum.
Starting from alcohol 3 of 72% e.e. obtained by chiral
base chemistry as described earlier,3b,6 the esters were
prepared in enantiomerically enriched form and evalu-
ated for asymmetric epoxidation of alkenes with different
Table 2. Epoxidation of aromatic alkenes catalyzed by non-racemic ketones 4a–ga
Entry
Ketone
R
(E)-Stilbeneb
Conversione
Styrenec
a-Methylstyrened
Conversione
E.e.max
Conversione
E.e.max
f,g
f,h
f,i
E.e.max
1
2
3
4
5
6
4a
4c
4d
4e
4f
CH3
Ph
p-NO2Ph
m-NO2Ph
p-MeOPh
m-MeOPh
85
39
41
46
53
24
93
75
78
81
67
46
100
100
100
100
100
100
48
36
35
40
31
25
100
100
100
100
75
10
11
15
10
19
26
4g
65
a Alkene (0.1 mmol), Oxone® (1.0 mmol KHSO5), ketone (0.02 mmol), NaHCO3 (1.55 mmol), CH3CN (1.5 ml), aq. Na2EDTA (1 ml of 0.4 mmol
dm−3 solution), rt, 24 h.
b The major isomer had (R,R)-configuration in each case.9
c The major isomer had (R)-configuration in each case.7
d The major isomer had (S)-configuration in each case.7
e Estimated by integration of the crude 1H NMR spectrum.
f E.e.max=100×product e.e./ketone e.e. All catalysts were derived from alcohol 3 of 72% e.e.
g Measured by chiral HPLC (Chiralcel OD column using 10% isopropylalcohol/hexane as eluent).
h Measured by chiral GC (Chiraldex GTA column using a thermal ramp 50–180°C).
i Measured by chiral HPLC (Chiralcel OD column using 0.8% iso-propylalcohol/hexane as eluent).
Table 3. Asymmetric epoxidation of alkenes catalyzed by non-racemic ketone 4aa
c
Entry
Alkene
Conversionb
E.e.max
Product configuration
1
2
3
4
5
6
7
8
Styrene
a-Methylstyrene
(E)-Stilbene
b-Methylstyrene
Phenylcyclohexene
Phenylstilbene
Methyl (E)-cinnamate
2-Cyclohexenone
100
100
85
100
89
71
4
51
48
10
93
70
82
98
84
3
(R)d
(S)e
(R,R)e
(R,R)f
(R,R)f
(R,R)e
n.d.g
(R,R)h
a Alkene (0.1 mmol), Oxone® (1.0 mmol KHSO5), ketone (0.02 mmol), NaHCO3 (1.55 mmol), CH3CN (1.5 mL), aq. Na2EDTA (1 ml of 0.4 mmol
dm−3 solution), rt, 24 h.
b Estimated by integration of the crude 1H NMR spectrum.
c E.e.max=100×product e.e./ketone e.e. Catalyst 4a used was of 82% e.e. in all cases, except for entries 2, 5 and 6 where 4a was of 72% e.e.
d Product e.e. and configuration were determined by chiral GC.7
e Product e.e. and configuration determined by chiral HPLC analysis.7
f Product e.e. and configuration determined by chiral shift reagent 1H NMR analysis.9
g Configuration not determined.
h Product e.e. determined by GC; configuration determined by comparing the measured optical rotation with the reported one.10

A. Armstrong et al. / Tetrahedron: Asymmetry 12 (2001) 2779–2781
2781
Since the a-acetoxyketone 4a appeared to be the best of
the esters examined, we investigated its use in the
epoxidation of a wider range of alkenes (Table 3). The
results are highly encouraging given that they were
obtained at room temperature and without optimiza-
tion of solvent or reaction pH. Importantly, catalyst 4a
does not appear to undergo Baeyer–Villiger decomposi-
tion under the reaction conditions. Very high e.e.max
was observed for epoxidation of phenylstilbene (98%,
entry 6). Trans- or trisubstituted aromatic olefins gener-
ally afforded products with good e.e.max (entries 3–6).
Since there are few examples of the asymmetric epoxi-
dation of electron-poor alkenes with chiral dioxiranes,
we tested methyl (E)-cinnamate. Although good e.e.max
was observed, the conversion was very low (entry 7).
2-Cyclohexenone proved more reactive but the enan-
tioselectivity was very poor (entry 8).
We gratefully acknowledge generous unrestricted sup-
port from Pfizer, Merck and Bristol-Myers Squibb.
References
1. Review: Frohn, M.; Shi, Y. Synthesis 2000, 1979–2000.
2. (a) Armstrong, A.; Hayter, B. R. Tetrahedron 1999, 55,
11119–11126; (b) Tian, H.; She, X.; Shi, Y. Org. Lett.
2001, 3, 715–718.
3. (a) Armstrong, A.; Hayter, B. R. Chem. Commun. 1998,
621–622; (b) Armstrong, A.; Hayter, B. R; Moss, W. O.;
Reeves, J. R.; Wailes, J. S. Tetrahedron: Asymmetry
2000, 11, 2057–2061.
4. Yang, D.; Wong, M.-K.; Yip, Y.-C. J. Org. Chem. 1995,
60, 3887–3889.
5. Esters were prepared under standard conditions
(RCOCl, pyridine). All gave satisfactory spectroscopic
data.
6. Bunn, B. J.; Cox, P. J.; Simpkins, N. S. Tetrahedron
1993, 49, 207–218.
7. Tian, H.; She, X.; Xu, J.; Shi, Y. Org. Lett. 2001, 3,
1929–1931.
8. (a) Hayter, B. R. Ph.D. Thesis, University of Notting-
ham, 1998; (b) a graph of catalyst e.e. against product
e.e. (9 data points) was linear with R2=0.9991. Reeves,
J. R. Unpublished results.
In summary, we have further investigated the use of
esters of a-hydroxy-bicyclo[3.2.1]octan-3-one as cata-
lysts in the Oxone® epoxidation of alkenes. The highest
e.e.max was obtained using the a-acetoxy-oxabicycle 4a
and the scope of this catalyst was explored. We are
continuing to examine further derivatives of this ring
system as well as addressing the efficient preparation of
4a in enantiomerically pure form.
9. Tu, Y.; Wang, Z.-X.; Shi, Y. J. Am. Chem. Soc. 1996,
118, 9806–9807.
Acknowledgements
10. Sato, T.; Watanabe, M.; Honda, N.; Fujisawa, T. Chem.
Lett. 1984, 1175–1176.
We thank the EPSRC and AstraZeneca (CASE award
to J.R.R.) and the EPRSC (GR/M84534) for funding.