Asymmetric Baylis-Hillman reactions: Catalysis using a chiral pyrrolizidine base

Article abstract of DOI:10.1039/a806115g

A novel chiral pyrrolizidine base 5 derived from L-proline promotes the Baylis-Hillman reaction of ethyl and methyl vinyl ketones with electron deficient aromatic aldehydes with moderate levels of enantiomeric excess.

Full text of DOI:10.1039/a806115g

Asymmetric Baylis–Hillman reactions: catalysis using a chiral pyrrolizidine
base
Anthony G. M. Barrett,* Andrew S. Cook and Akio Kamimura
Department of Chemistry, Imperial College of Science, Technology and Medicine, South Kensington, London, UK
SW7 2AY. E-mail: m.stow@ic.ac.uk
Received (in Liverpool, UK) 3rd August 1998, Accepted 16th October 1998
A novel chiral pyrrolizidine base 5 derived from
L
-proline
2-O2NC6H4, R2 = Et) (Table 1) which was formed in variable
yield and enantioselectivity.§ The yield of the reaction was
significantly improved by cooling and no significant decrease in
rate was observed until 250 °C. Leahy has reported unusual
temperature dependence on conversions in the DABCO medi-
ated Baylis–Hillman reaction of acrylate esters with alde-
hydes.10 Although, the yield of the reaction was optimum at
240 °C, enantioselectivities were superior at higher tem-
peratures with a maximum value at 220 °C (47% ee). Since
Aggarwal has reported rate enhancements on the use of
lanthanide triflates in the Baylis–Hillman reaction,11 a series of
Lewis acid co-catalysts were examined in the synthesis of 6.
Amongst diverse metal salts examined, those of the alkali
metals, in particular sodium, were the most effective additives.
A series of aldehydes were allowed to react with ethyl or methyl
promotes the Baylis–Hillman reaction of ethyl and methyl
vinyl ketones with electron deficient aromatic aldehydes
with moderate levels of enantiomeric excess.
The Baylis–Hillman reaction is a convenient process for the
preparation of a b-hydroxy-a-methylene ketone, nitrile, ester,
etc. in one step from an a,b-unsaturated ketone, acrylonitrile or
an acrylic ester and an aldehyde.1 The reaction is mediated by a
tertiary amine, and DABCO (diazabicyclo[2.2.2]octane) is the
most common catalyst employed. Whilst the Baylis–Hillman
reaction of chiral aldehydes or chiral Michael acceptors has
been shown to proceed, in some cases, with high diastereo-
selectivities, the development of chiral catalysts for the Baylis–
Hillman reaction is less well developed. Hirama and Markó
have respectively reported the use of chiral derivatives of
diazabicyclo[2.2.2]octane2 and of quinidine or cinchonine3 as
enantioselective catalysts. However, these authors observed
only modest levels of enantioselectivities (11–47 and 6–45% ee,
respectively) and the requirement to use elevated pressures
(3–10 Kbar) to ensure acceptable conversions. We have
previously published a two step procedure to effect the
enantioselective (50–96% ee) Baylis–Hillman reaction of an
aldehyde with an a-methylene ketone via a tandem Michael
addition–aldol reaction of (phenylthio)- or (phenylselenyl)-
trimethylsilane catalysed by a chiral borane Lewis acid
followed by oxidative elimination.4 Recently, Soai and co-
workers reported the use of (S)-BINAP as a catalyst for the
enantioselective (9–44% ee) Baylis–Hillman reaction of pyr-
imidine-5-carbaldehydes with acrylate esters.5 This work has
prompted us to report the use of pyrrolizidine (1-azabicyclo-
[3.3.0]heptane) derivatives as alternative chiral catalysts. We
considered that such amines may function as efficient catalysts
for the Baylis–Hillman reaction on account of their enhanced
basicity relative to common tertiary amines6 and the accessi-
bility of the nitrogen lone pair. We were concerned, in our
design, to seek to alleviate the known slow kinetics of the
DABCO catalysed Baylis–Hillman reaction.
vinyl ketone in the presence of amine 5 (10 mol%) and 1
NaBF4 or NaBPh4¶ in MeCN at 220 °C (Table 2). The
M
i
OH
CO2Me
N
N
Boc
Boc
2 (81%)
1
H
ii,iii
iv
N
O
3 (73%)
O2N
O2N
H
N
H
N
H
H
v
HO
HO
O
4 (18%, 76% other isomers)
5 (53%)
Scheme 1 Reagents and conditions: i, Swern oxidation, CH2Cl2, then
Ph3PNCHCO2Me; ii, Raney Ni, H2 (40 psi), MeOH; iii, HCl, EtOAc, 0 °C,
then NaOMe, MeOH; iv, Lithium 2,2,6,6-tetramethylpiperidide, THF, 278
to 230 °C, then 4-O2NC6H4CHO, BF3·OEt2, 278 to 220 °C; v, BH3·SMe2,
THF, reflux, then TsOH, MeOH, reflux, then K2CO3, MeOH, reflux.
Swern oxidation of Boc-
-prolinol7 1 and direct Wittig
L
homologation gave ester 28 (81%) (Scheme 1). Subsequent
hydrogenation over Raney nickel, deprotection of the Boc
group, under acidic conditions, and lactamisation gave the
pyrrolizidinone9 3 (73%). Aldol reaction of lactam 3 with
4-nitrobenzaldehyde in the presence of BF3·OEt2 gave a
mixture of four b-hydroxy lactams (94%). The less polar
component consisted of a single crystalline diastereoisomer 4†‡
which was readily isolated in 16% yield. The remaining three
diastereoisomers co-chromatographed and were not separable
at the lactam oxidation stage. Finally, BH3·SMe2 mediated
reduction gave the desired pyrrolizidine 5 which was initially
isolated as the robust borane adduct but which could be
converted into the free base following sequential reflux with
methanolic TsOH and K2CO3.
Table 1 Temperature variation in the Baylis–Hillman reaction of 2-nitro-
benzaldehyde with ethyl vinyl ketone
O
OH
O
5
Et
2-O2NC6H4
Et
2-O2NC6H4CHO +
(10 mol%)
T/°C
Yield (%)
Ee (%)
t/d
Solvent
25
4
27
21
57
50
53
93
9a
37
42
30
47
31
26
21
3
3
3
2
2
2
3
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
EtCN
210
220
230
240
275
The Baylis–Hillman reaction of ethyl vinyl ketone with
2-nitrobenzaldehyde was examined in MeCN or EtCN solution
in the presence of pyrrolizidine 5 (10 mol%) at variable
temperatures (275 to 25 °C). All reactions gave rise to the
a This slow reaction was stopped before reaching completion.
corresponding b-hydroxy-a-methylene ketone
6
(R1
=
Chem. Commun., 1998, 2533–2534
2533

View Online
Table 2 Baylis–Hillman reactions of aldehydes with ethyl or methyl vinyl
ketones
Wellcome Research Ltd for the most generous endowment (to
A. G. M. B.); the Wolfson Foundation for establishing the
Wolfson Centre for Organic Chemistry in Medical Science at
Imperial College; and George O’Doherty, Oswy Pereira and
D. Christopher Braddock for helpful discussions.
O
OH
O
5 (10 mol%)
R2
R1CHO
+
R1
R2
NaBF4, MeCN
–40 °C
6
Notes and references
R1
R2
Yield (%) Ee (%)a t/h
† All new compounds were fully authenticated by spectroscopic data and
microanalysis and/or HRMS.
‡ The structure of 4 and related b-hydroxy lactams and pyrrolizidines,
which were confirmed by X-ray crystallography, will be reported
elsewhere.
2-O2NC6H4
2-O2NC6H4
2-FC6H4
Et
Me
Et
Et
Et
Et
Et
Et
Et
Et
71
71
31
58
63
51
83
93
63
17
67
53
63
72
71
37
21
49
70
39
18
18
48
14
72
18
14
12
18
48
2-ClC6H4
2-BrC6H4
3-O2NC6H4
2-Pyridyl
3-Pyridyl
4-Quinolinylb
4-O2NC6H4
a
§ Enantioselectivities of all reactions were determined by HPLC analyses
on Chiralcel OD-H and AD columns and, in some cases, by 1H NMR
spectroscopy in the presence of the chiral the shift reagent Eu(tfc)3.
¶ For the use of lithium salts to enhance the rate of reactions proceeding via
ionic intermediates, see ref. 12.
∑ General experimental procedure (Table 2, entry 4): 6 (R1 = 2-ClC6H4, R2
= Et)·NaBF4 (27 mg, 0.25 mmol) followed by 2-chlorobenzaldehyde (26
ml, 0.22 mmol) were added to a stirred suspension of amine 5 (5 mg, 0.019
mmol) in MeCN (0.25 ml) under nitrogen at –40 °C. The reaction mixture
was stirred for a further 10 min when ethyl vinyl ketone (19 ml, 0.19 mol)
was added and stirring continued for 24 h. The mixture was concentrated in
vacuo and the residue chromatographed (1:6 EtOAc–hexanes, Rf 0.3) to
yield the title compound 6 (R1 = 2-ClC6H4, R2 = Et) (24.8 mg, 58%) as a
colourless oil.
b
Determined by HPLC analyses (Chiralcel OD-H and AD). Reaction
carried out using NaBPh4 not NaBF4.
corresponding b-hydroxy-a-methylene ketones 6 were isolated
in modest to excellent yields (17–93%) and with acceptable
levels of enantioselectivity (21–72% ee).∑ The absolute config-
uration of ketone 6 (R1 = 2-O2NC6H4, R2 = Me) was
determined to be R by comparison of the sign of its specific
rotation with that of the antipodal ketone.** The absolute
stereochemistry of the other b-hydroxy-a-methylene ketones 6
were assigned by analogy since all were laevorotatory. It is
reasonable to speculate that the reaction may proceed via the
intermediate 7 rather than the more sterically congested system
8. The importance of hydroxy substitution on enhancing the rate
** The absolute stereochemistry of the major adduct 6 (R1 = 2-O2NC6H4,
R2 = Me) was determined by comparison of the sign of the specific rotation
with literature data on (4R)-4-hydroxy-3-methylene-4-(2-nitrophenyl)bu-
tan-2-one, which was prepared by an enantioselective Baylis–Hillman
reaction (11–42% ee) and by the partial kinetic resolution of the racemic
compound by Sharpless epoxidation using
2).
L-(+)-diethyl tartrate (see ref.
1 S. E. Drewes and G. H. P. Roos, Tetrahedron, 1988, 44, 4653; D.
Basavaiah, P. D. Rao and R. S. Hyma, Tetrahedron, 1996, 52, 8001; E.
Ciganek, Org. React., 1997, 51, 201; L. J. Brzezinski, S. Rafel and J. W.
Leahy, J. Am. Chem. Soc., 1997, 119, 4317.
2 T. Oishi, H. Oguri and M. Hirama, Tetrahedron: Asymmetry, 1995, 6,
1241.
BF4–
–OH +N
O
M+
NO2
H
H
Et
3 I. E. Markó, P.R. Giles and N. J. Hindley, Tetrahedron, 1997, 53,
1015.
O
R
7
4 A. G. M. Barrett and A. Kamimura, J. Chem. Soc., Chem. Commun.,
1995, 1755.
BF4–
5 T. Hayase, T. Shibata, K. Soai and Y. Wakatsuki, Chem. Commun.,
1998, 1271.
6 M. Ikeda, T. Sato and H. Ishibashi, Heterocycles, 1988, 27, 1465; J.
Zabicky, in The Chemistry of Amino Group, ed. S. Patai, Wiley,
London, 1968, p 132.
–O +N
O
O
M+
NO2
H
Et
H
R
8
7 A. F. Spatola, M. K. Anwer, A. L. Rockwell and L. M. Gierasch, J. Am.
Chem. Soc., 1986, 108, 825.
of Baylis–Hillman reactions is well known1 and is exemplified
by the fact that 3-hydroxyquinuclidine is a superior catalyst to
quinuclidine. Such an effect may also be of significance in both
the catalysis by pyrrolizidine 5 and the enhancement of
enantioselectivity in the presence of sodium ions. Further
studies of this variant of the Baylis–Hillman reaction are now
being investigated in our laboratory.
8 S. Le Coz, A. Mann, F. Thareau and M. Taddei, Heterocycles, 1993, 36,
2073.
9 R. Grote, A. Zeeck, J. Stümpfel and H. Zähner, Liebigs Ann. Chem.,
1990, 525.
10 S. Rafel and J. W. Leahy, J. Org. Chem., 1997, 62, 1521.
11 V. K. Aggarwal, G. J. Tarver and R. McCague, Chem. Commun., 1996,
2713.
12 P. A. Grieco, Aldrichim. Acta, 1991, 24, 59.
We thank Zeneca, Chiroscience, the EPSRC and the DTI for
generous support under the Link Asymmetric Scheme; Glaxo-
Communication 8/06115G
2534
Chem Commun., 1998, 2533–2534