Sulfide-BF3·OEt2 mediated Baylis-Hillman reactions

Article abstract of DOI:10.1016/S0040-4039(02)02053-1

A sulfide-BF₃·OEt₂ mediated Baylis-Hillman reaction has been developed in which the sulfide acts via attack onto the activated alkene. The use of a chiral sulfide gives rise to Baylis-Hillman adducts with up to 53% ee.

Full text of DOI:10.1016/S0040-4039(02)02053-1

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 8219–8222
Sulfide–BF ·OEt mediated Baylis–Hillman reactions
3
2
Louise M. Walsh, Caroline L. Winn and Jonathan M. Goodman*
Department of Chemistry, University of Cambridge, Lensfield Road, Cambridge CB2 1EW, UK
Received 2 August 2002; revised 10 September 2002; accepted 19 September 2002
Abstract—A sulfide–BF ·OEt mediated Baylis–Hillman reaction has been developed in which the sulfide acts via attack onto the
3
2
activated alkene. The use of a chiral sulfide gives rise to Baylis–Hillman adducts with up to 53% ee. © 2002 Elsevier Science Ltd.
All rights reserved.
7
2b,d,e
The coupling of an activated alkene and an aldehyde,
the Baylis–Hillman reaction, has been the focus of
much research interest. Kataoka et al. recently devel-
(cyclohexenone) but is needed at −78°C (MVK).
We decided to clarify at which point the sulfide is
required. We found, in contrast to results reported by
Shi et al., that the reaction of methyl vinyl ketone and
p-nitrobenzaldehyde in the presence of TiCl was facile
in the absence of a Lewis base, even at −90°C. Since
our aim was to use chiral sulfides to lead to an asym-
1–3
oped the sulfide/Lewis acid mediated Baylis–Hillman
4
,5
reaction. The Lewis acid must be present for these
reactions to proceed, in contrast to amine promoted
reactions where a Lewis acid is not always required.
The most commonly used Lewis acid for the sulfide
mediated reactions is TiCl4 but more recently
BBr ·DMS and BCl ·DMS have also been successfully
4
metric reaction, we decided that TiCl was an unsuit-
4
able Lewis acid for our needs. We set out to develop a
sulfide/Lewis acid mediated Baylis–Hillman reaction in
which the sulfide was unambiguously required and ide-
ally acting via attack on to the activated alkene.
3
3
6
employed. Although it seems clear that the Lewis acid
is required in these reactions, it is less clear whether the
sulfide itself is necessary. The reaction of p-nitroben-
zaldehyde and cyclohexenone is effected by the addition
We chose to use Lewis acids which did not bear chlo-
ride, bromide or iodide ligands as we anticipated that,
in these cases, the halide ion could easily be liberated,
perhaps even in the absence of a sulfide, and subse-
quently bring about the Baylis–Hillman reaction.
7
of TiCl4 at room temperature. However, Shi has
reported that the reaction of methyl vinyl ketone and
p-nitrobenzaldehyde in the presence of TiCl at −78°C
4
does not proceed unless an additional Lewis base is
2
b,d,e
present.
This poses the question: what is the role of
the sulfide in the sulfide/Lewis acid mediated Baylis–
Hillman reaction?
Table 1. Tetrahydrothiophene–BF ·OEt mediated Baylis–
3
2
Hillman reaction of methyl vinyl ketone with various
10
It was originally thought that the sulfide/TiCl medi-
aldehydes
4
ated Baylis–Hillman reactions proceeded via Michael
attack of the sulfide onto the activated alkene but it is
now believed that these reactions proceed via attack of
a halide ion, released from the Lewis acid by the
8
sulfide. Supporting evidence for this was that the use
of BF ·OEt , where the release of a halide ion from the
3
2
Lewis acid is far less likely, had thus far been
unsuccessful.
Entry
R
Yield (%)
4
1
2
3
4
5
6
p-NO C H
4
50
48
52
45
45
43
2
6
The literature suggests that a sulfide is not needed for
TiCl4 Baylis–Hillman reactions at room temperature
p-ClC H
6
4
Ph
p-MeOC H
6
4
Et
*
Corresponding
author.
Fax:
+44-1223-335352;
e-mail:
PhCH CH
2
2
j.m.goodman@ch.cam.ac.uk
0
040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02053-1

8
220
L. M. Walsh et al. / Tetrahedron Letters 43 (2002) 8219–8222
Indeed, the reaction of cyclohexenone and various
with aromatic aldehydes. Baylis–Hillman reactions with
aliphatic aldehydes in the presence of Et AlI and no
simple aliphatic aldehydes such as propionaldehyde are
2
11
additional Lewis base at 0°C is complete in only 20
often troublesome and low yielding so this is a partic-
ularly pleasing result.
9
minutes.
We studied the reaction of p-nitrobenzaldehyde and
methyl vinyl ketone both with and without tetra-
hydrothiophene with various Lewis acids. Kataoka et
al. have previously reported that BF ·OEt is an ineffec-
The analogous reaction of p-nitrobenzaldehyde and the
less reactive enone, cyclohexenone, also yields the
desired Baylis–Hillman adduct, albeit in a reduced yield
of 11% (Scheme 1).
3
2
tive Lewis acid for both the dimethylsulfide mediated
Baylis–Hillman reaction of p-nitrobenzaldehyde and
Scheme 2 illustrates the most likely mechanistic path-
way for this reaction; the sulfide directly attacking the
methyl vinyl ketone and the resultant species then
attacking the aldehyde. Treatment with base leads to
elimination of the sulfide to yield the Baylis–Hillman
adduct.
4
b
cyclohexenone and also for the 2,6-diphenyl-4H-
thiopyran-4-one mediated reaction of p-nitrobenzalde-
4
c
hyde and methyl vinyl ketone. We found BF .OEt2
3
and tetrahydrothiophene to be effective for this trans-
formation and that this reaction is rapid. The reaction
of 1 equiv. of p-nitrobenzaldehyde and 3 equiv. of
methyl vinyl ketone in the presence of 1.2 equiv. of
Evidence to support this pathway was obtained by
quenching the reaction with dilute HCl instead of base.
Under these conditions elimination of the sulfide did
not occur and 2 was recovered from the aqueous
extracts, formed from the acid mediated dehydration of
tetrahydrothiophene and 1.5 equiv. BF ·OEt at 0°C for
3
2
just 30 minutes followed by the addition of Et N gave
3
a 50% yield of the Baylis–Hillman adduct (Table 1,
entry 1). Moreover, there is no reaction in the absence
of sulfide.
1
. It seems likely, therefore, that the sulfide is acting in
a similar fashion to the amine in an amine promoted
It is necessary to use 1.5 equiv. of BF ·OEt to obtain
3
2
Baylis–Hillman reaction.
optimum yields in this reaction. The use of just 1 equiv.
of BF ·OEt gives a 20% yield of the Baylis–Hillman
3
2
Since we believed the sulfide to be attacking the methyl
vinyl ketone we envisaged that the use of a chiral
sulfide would result in an asymmetric version of this
reaction.
adduct. Increasing the amount of Lewis acid to 2
equiv., however, does not improve on the 50% yield.
This sulfide–BF ·OEt mediated Baylis–Hillman reac-
3
2
tion is very versatile with respect to the aldehyde used:
the use of both aromatic and aliphatic aldehydes lead-
ing to successful reactions (Table 1). Particularly note-
worthy are the results with the aliphatic enolisable
aldehydes (entries 5 and 6). No side products were
obtained and the yields are similar to those observed
The C symmetric tricyclic sulfide 3 has recently been
2
12
synthesised in our group and, used in the one-pot
sulfur-ylide epoxidation reaction, gives epoxides in
13
>
94% ee. We therefore decided to use this sulfide in
the Baylis–Hillman reaction. We were pleased to find
that the chiral sulfide successfully mediated the reaction
giving enantioenhanced Baylis–Hillman adducts (Table
2). The sulfide could be recovered after each reaction.
The Baylis–Hillman adduct obtained from the reaction
of p-nitrobenzaldehyde and methyl vinyl ketone using
the chiral sulfide 3 at 0°C was found to have an
enantiomeric excess of 21% (Table 2, entry 1). By
Scheme 1.
Scheme 2.

L. M. Walsh et al. / Tetrahedron Letters 43 (2002) 8219–8222
8221
Table 2. Asymmetric Baylis–Hillman reaction of methyl vinyl ketone with various aldehydes
Entry
R
Temp. (°C)
Time (min)
Yield (%)
Ee (%)a,b
1
2
3
4
5
6
7
p-NO C H
0
−78
−78
−78
−78
−78
−78
30
1
5
30
30
120
120
37
8
21
53
49
46
28
23
14
2
6
4
4
4
4
p-NO C H
2
6
p-NO C H
38
48
18
60
41
2
6
p-NO C H
2
6
Et
Et
PhCH CH
2
2
a
Determined by HPLC analysis using a Chiralcel™ OD column eluting with hexane/propan-2-ol (R=p-NO C H , 9:1; R=Et, 49:1; R=
2
6
4
PhCH CH , 19:1).
2
2
b
All [h]D values are negative. When R=p-NO C H the absolute configuration is known by comparison with the literature value ([h] =−12.1,
2
6
4
D
c=0.53).14
lowering the temperature of the reaction to −78°C, the
ee could be increased. A reaction time of 1 minute led
to an ee of 53% but in only 8% yield (Table 2, entry 2).
Increasing the reaction time results in an increase in
yield with a slight decrease in ee; a reaction time of 30
minutes at −78°C gave a 48% yield of the Baylis–Hill-
man adduct with 46% ee.
aldehyde can therefore only occur from one side of this
intermediate. Reaction with the aldehyde via an
extended transition state 6 gives rise to the R enan-
tiomer as shown in Fig. 1. This serves as a qualitative
model to explain the selectivity and could help design
other chiral sulfides for further studies.
17
In summary, we have developed a Baylis–Hillman reac-
tion in which the sulfide directly participates via
Michael addition to the a,b-unsaturated ketone. The
reaction is rapid and a number of aldehydes can be
used with methyl vinyl ketone to give the Baylis–Hill-
man adducts in moderate yields. The use of a chiral
sulfide gives Baylis–Hillman adducts with ee’s of up to
53%.
By comparison of the sign of the specific rotation of
Baylis–Hillman adduct 4 (R=p-NO C H ) with the
2
6
5
1
4
literature value the selectivity of the reaction was
found to be in favour of the R enantiomer. In order to
rationalise this selectivity, MacroModel MM2*
1
5,16
calculations
were undertaken on the sulfonium spe-
cies 5. RHF/6-31G** calculations verified that the Z
enol is preferred over the E enol. The calculations
revealed a lowest energy conformation in which the
reactive enolate chain attached to the sulfur points
away from the bulky sulfide substituent. This in turn
leads to one face of the enolate being shielded by the
other side of the tricyclic sulfide. Attack onto the
Acknowledgements
We would like to thank the EPSRC for financial
support.
References
1
. For reviews see (a) Ciganek, E. Org. React. 1997, 51, 201;
b) Basavaiah, D.; Rao, P.; Hyma, R. S. Tetrahedron
996, 52, 8001; (c) Drewes, S. E.; Roos, G. H. P. Tetra-
(
1
hedron 1988, 44, 4653.
2
. (a) Leadbeater, N. E.; Van der Pol, C. J. Chem. Soc.,
Perkin Trans. 1 2001, 2831; (b) Shi, M.; Jiang, J.-K.; Cui,
S.-C.; Feng, Y.-S. J. Chem. Soc., Perkin Trans. 1 2001,
390; (c) Han, Z.; Uehira, S.; Shinokubo, H.; Oshima, K.
J. Org. Chem. 2001, 66, 7854; (d) Shi, M.; Feng, Y.-S. J.
Org. Chem. 2001, 66, 406; (e) Shi, M.; Jiang, J.-K.; Cui,
S.-C. Tetrahedron 2001, 57, 7343.
3
. Examples of enantioselective reactions include (a)
Iwabuchi, Y.; Nakatani, M.; Yokoyama, N.;
Hatakeyama, S. J. Am. Chem. Soc. 1999, 121, 10219; (b)
Barrett, A. G. M.; Kamimura, A. J. Chem. Soc., Chem.
Commun. 1995, 1755.
Figure 1. Lowest energy conformation of 5 and its approach
to benzaldehyde via an extended transition state 6.

8
222
L. M. Walsh et al. / Tetrahedron Letters 43 (2002) 8219–8222
4. (a) Kataoka, T.; Iwama, T.; Tsujiyama, S. Chem. Com- stirred for a further 10 min whilst warming to room
mun. 1998, 197; (b) Kataoka, T.; Iwama, T.; Tsujiyama,
S.; Iwamura, T.; Watanabe, S. Tetrahedron 1998, 54,
temperature. The solution was washed with dilute HCl,
saturated NaHCO3 and brine, dried over MgSO4 and
concentrated in vacuo to give the crude product which
was purified by column chromatography.
1
1813; (c) Iwama, T.; Kinoshita, H.; Kataoka, T. Tetra-
hedron Lett. 1999, 40, 3741.
5
. For examples of enantioselective sulfide mediated Baylis–
Hillman reactions, see: (a) Kataoka, T.; Iwama, T.; Tsu-
jiyama, S.; Kanematsu, K.; Iwamura, T.; Watanabe, S.
Chem. Lett. 1999, 257; (b) Iwama, T.; Tsujiyama, S.;
Kinoshita, H.; Kanematsu, K.; Tsurukami, Y.; Iwamura,
T.; Watanabe, S.; Kataoka, T. Chem. Pharm. Bull. 1999,
1
1
1. For example, Auvray, P.; Knochel, P.; Normant, J. F.
Tetrahedron 1988, 44, 6095.
2. (a) Rao, A. V. R.; Reddy, K. A.; Gurjar, M. K.; Kunwar,
A. C. J. Chem. Soc., Chem. Commun. 1988, 1273; (b)
Winn, C. L.; Goodman, J. M. Tetrahedron Lett. 2001, 42,
7091.
4
7, 956.
1
1
1
3. Winn, C. L.; Bellenie, B. R.; Goodman, J. M. Tetra-
6
. Iwamura, T.; Fujita, M.; Kawakita, T.; Kinoshita, S.;
Wantanabe, S.; Kataoka, T. Tetrahedron 2001, 57, 5455.
. Li, G.; Wei, H.-X.; Gao, J.; Caputo, T. Tetrahedron Lett.
hedron Lett. 2002, 43, 5427.
4. Oishi, T.; Oguri, H.; Hirama, M. Tetrahedron: Asymme-
try 1995, 6, 1241.
7
2
000, 41, 1.
5. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.;
Liskamp, R.; Lipton, M.; Caufield, C.; Chang, C.; Hen-
drickson, T.; Still, W. C. J. Comp. Chem. 1990, 11, 440.
6. Allinger, N. L. J. Am. Chem. Soc. 1977, 99, 440.
7. For early discussions on open transition states, see: (a)
Murata, S.; Suzuki, M.; Noyori, R. J. Am. Chem. Soc.
8
9
. Shi, M.; Jiang, J.-K. Tetrahedron 2000, 56, 4793.
. Itoh, A.; Ozawa, S.; Oshima, K.; Nozaki, H. Bull. Chem.
Soc. Jpn. 1981, 54, 274.
1
1
1
0. Typical procedure: To a solution of aldehyde (1 mmol),
methyl vinyl ketone (3 mmol) and tetrahydrothiophene
(
(
1.2 mmol) in DCM (5 ml) at 0°C was added BF ·OEt2
1.5 mmol). After stirring the reaction for 30 min at this
3
1980, 102, 3248; (b) Yamamoto, Y.; Yatagai, H.; Naruta,
temperature Et N (1 mmol) was added and the mixture
Y.; Maruyama, K. J. Am. Chem. Soc. 1980, 102, 7109.
3