Tandem copper-catalyzed enantioselective allylation-metathesis

Article abstract of DOI:10.1021/ol0269244

(matrix presented) Grignard reagents undergo enantioselective (up to 86% ee) copper-catalyzed S_N2′ substitution on achiral allylic chlorides. The reaction is wide in scope for both the Grignard reagent and the allylic substrate. The resulting terminal alkene could be submitted to intra-or intermolecular metathesis to afford new chiral synthons. The experimental conditions are compatible with a one-pot overall substitution - metathesis procedure without loss of enantioselectivity.

Full text of DOI:10.1021/ol0269244

ORGANIC
LETTERS
2002
Vol. 4, No. 23
4147-4149
Tandem Copper-Catalyzed
Enantioselective Allylation−Metathesis
Alexandre Alexakis* and Karine Croset
Department of Organic Chemistry, UniVersity of GeneVa, 30, Quai Ernest Ansermet,
1211 GeneVa 4, Switzerland
alexandre.alexakis@chiorg.unige.ch
Received September 18, 2002
ABSTRACT
Grignard reagents undergo enantioselective (up to 86% ee) copper-catalyzed SN2′ substitution on achiral allylic chlorides. The reaction is wide
in scope for both the Grignard reagent and the allylic substrate. The resulting terminal alkene could be submitted to intra- or intermolecular
metathesis to afford new chiral synthons. The experimental conditions are compatible with a one-pot overall substitution−metathesis procedure
without loss of enantioselectivity.
The allylic substitution reaction is a useful organic trans-
formation, provided the regio-, stereo-, and chemoselectivities
could be controlled. This control is usually provided by the
type of metal catalyst, by the nucleophile, and by the leaving
group.1 In the field of asymmetric synthesis, although
spectacular results have been achieved with Pd,2 only recent
attention has been paid to Cu, despite the fact that it allows
the best γ-regioselectivity.3 In addition, Cu allows the use
of Grignard or organozinc reagents as nucleophiles.
chiral source, whereas all other authors have used an external
chiral ligand (Scheme 1).
Scheme 1
Enantioselective copper-catalyzed γ-allylations have been
disclosed, with R2Zn, by Knochel,4 Hoveyda,5 and Feringa,6
and with RMgX, by us7 and by Van Koten and Ba¨ckvall.8
The latter authors have worked with a chiral Cu thiolate, as
In our first report,7 we showed that Grignard reagents can
afford high enantioselectivity only with aryl-substituted
(cinnamyl-type) allylic chlorides. CuCN was essential for
the control of the γ-regioselectivity, and the best ligand was
4 (Figure 1). Since that report, we have disclosed new chiral
ligands, 5-7, with induced atropoisomerism,9 and we have
(1) Magid, R. M. Tetrahedron 1980, 36, 1901.
(2) Trost, B. M.; van Vranken D. L. Chem. ReV. 1996, 96, 395.
(3) Karlstro¨m, A. S. E.; Ba¨ckvall, J.-E. In Modern Organocopper
Chemistry; Krause, N., Ed.; Wiley-VCH: Weinheim, 2001; p 259.
(4) (a) Du¨bner, F.; Knochel, P. Angew. Chem., Int. Ed. 1999, 38, 379.
(b) Du¨bner, F.; Knochel, P. Tetrahedron Lett. 2000, 41, 9233.
(5) Luchaco-Cullis, C. A.; Mizutani, H.; Murphy, K. E.; Hoveyda, A.
H. Angew. Chem., Int. Ed. 2001, 40, 1456.
(8) (a) van Klaveren, M.; Personn, E. S. M.; del Villar, A.; Grove, D.
M.; Ba¨ckvall, J.-E.; van Koten, G. Tetrahedron Lett. 1995, 36, 3059. (b)
Karltro¨m, A. S. E.; Huerta, F. F.; Meuzelaar, G. J.; Ba¨ckvall, J.-E Synlett
2001, 923.
(6) Malda, H.; van Zijl, A. W.; Arnold, L. A.; Feringa, B. L. Org. Lett.
2001, 3, 1169.
(7) Alexakis, A.; Malan, C.; Lea, L.; Benhaim, C.; Fournioux, X. Synlett
2001, 927.
(9) Alexakis, A.; Rosset, S.; Allamand, J.; March, S.; Guillen, F.;
Benhaim, C. Synlett 2001, 1375.
10.1021/ol0269244 CCC: $22.00 © 2002 American Chemical Society
Published on Web 10/16/2002

lower enantioselectivities. Finally, the known ligands 8 and
911 were also screened and gave moderate or very low ee’s.
It should be pointed out that, in the absence of ligand,
CuTC alone gave 90% of R-substituted product, whereas the
presence of a phosphorus ligand reverses the trend toward
γ-regioselectivity (Scheme 2). This effect of a donor ligand
on the regioselectivity is unprecedented in the Cu literature.12
Scheme 2
Figure 1. Chiral ligands used in this work.
found that other experimental conditions were more ap-
propriate (Cu carboxylates, Et2O) for the conjugate addition
with phosphorus ligands.10 We expected that these new
conditions, together with new ligands, could afford much
better enantioselectivities, and wider substrate and RMgX
tolerance.
Keeping our previous best ligand 4 but changing the
copper salt to CuTC (copper thiophene 2-carboxylate)
improved the enantioselectivity on cinnamyl chloride to 82%,
instead of 73% (Table 1, entries 1 and 2). Shifting to the
Table 2 shows the results obtained with various Grignard
reagents. It should be recalled that in all these examples
Table 2. Allylic Substitution of Cinnamyl Chloride with
Various Grignard Reagentsa
entry ligand
Grignard
EtMgBr
% convnb
γ/R
% ee
1
2
4
5
4
5
4
5
4
5
4
5
4
5
100 (97%) 96/4
100 (96%) 92/8
95
83
100 (96%) 80/20 75, R
100
100
82, R
79, S
3
4
5
6
MeMgBr
31/69 61, R
12/88 62, S
Table 1. Allylic Substitution of Cinnamyl Chloride with
EtMgBr and Various Chiral Ligands, According to Scheme 1
n-PrMgBr
59/41 69, S
91/9 61, R
7
8
9
3-butenyl-MgBr
entry ligand
Cu salt
solvent T (°C) % convn γ/R % ee
100 (94%) 71/29 48, S
4-pentenyl-MgBr 100 90/10 72, R
100 (92%) 73/27 64, S
100 88/12 46, R
100 (98%) 90/10 83, S
1
2
3
4
5a
6
7
8
9
4
4
5
5
5
5
5
5
5
5
5
6
7
8
9
CuCN
CH2Cl2 -78
CH2Cl2 -78
CH2Cl2 -78
CH2Cl2 -78
CH2Cl2 -78
CH2Cl2 -55b
100
100
100
100
100
100
47
96/4 73, R
96/4 82, R
100/0 6, S
91/9 79, S
84/16 75, S
90/10 68, S
2/92
34/66 58, S
92/8 70, S
94/6 70, S
68/32 60, S
38/62 54, S
89/11 68, S
90/10 49, R
85/15 7, R
CuTC
CuCN
CuTC
CuTC
CuTC
CuTC
CuTC
CuTC
CuOAc2
10
11
12
i-PrMgBr
a All reactions were done with 1% CuTC + 1% ligand, in CH2Cl2, at
-78 °C. b In parentheses: isolated yield of the mixture of regioisomers.
Et2O
-55b
toluene -55b
CHCl3 -55b
CH2Cl2 -55b
50
100
100
100
100
100
100
100
CuCN does not allow high enantioselectivities, whereas
CuTC does. Cinnamyl chloride reacts with n-propyl, 3-
butenyl, and 4-pentenyl Grignard reagents in a similar way
than EtMgBr. The regio- and the enantioselectivity are
usually better with the TADDOL-derived ligand 4 than with
5. Only with a secondary Grignard, i-PrMgBr, ligand 5 is
clearly superior, giving the best ee’s (83%, entry 12).
As shown in Scheme 3, the results obtained with i-PrMgBr
are steadily high (up to 86% ee) with a range of cinnamyl-
type chlorides. This is a clear improvement over our previous
results.7
The use of CuTC and the new ligands allow extension to
alkyl substituents on the allylic substrate. Thus, the saturated
analogue of cinnamyl chloride 10 could be enantioselectively
substituted (Table 3, entry 8) with up to 74% ee. This is the
first time that a catalytic system, with Grignard reagents, is
10
11
12
13
14
15
CuNaphth CH2Cl2 -55b
CuTC
CuTC
CuTC
CuTC
CH2Cl2 -55b
CH2Cl2 -55b
CH2Cl2 -55b
CH2Cl2 -55b
b
a With cinnamyl bromide instead of chloride. 5% of Cu salt and ligand.
new ligand 5 showed a dramatic difference between the use
of CuCN (6% ee) and CuTC (79% ee). In this case, cinnamyl
bromide (entry 5) affords results similar to those with the
chloride. Et2O and toluene (entries 7 and 8) are not good
solvents, whereas chloroform (entry 9) is similar to dichlo-
romethane. CuOAc2 (entry 10) is as efficient as CuTC,
whereas copper naphthenate gives a diminished regioselec-
tivity. The other biphenol-type ligands 6 and 7 gave slightly
(11) Feringa, B. L. Acc. Chem. Res. 2000, 33, 346.
(12) Breit, B.; Demel, P. In Modern Organocopper Chemistry; Krause,
N., Ed.; Wiley-VCH: Weinheim, 2001; p 188.
(10) Alexakis, A.; Benhaim, C.; Rosset, S.; Humam, M. J. Am. Chem.
Soc. 2002, 124, 5262.
4148
Org. Lett., Vol. 4, No. 23, 2002

Scheme 3
Scheme 4
ethyl acrylate and 5% Grubbs catalyst were added to the
reaction mixture, after completion of the allylic substitution
(Scheme 4). After 24 h at 40 °C, the reaction was completed,
and 12 was obtained in 53% isolated yield and an ee of 82%,
identical to intermediate 2.
The intramolecular version was equally successful with
the 4-pentenyl- and 3-butenyl-substituted substrates affording,
quantitatively, the cyclized product in just 1 h at room
temperature. The RCM compounds could be easily separated
from the SN2 regioisomer and obtained in good isolated yield
with the same enantioselectivity as the intermediate diene
(Scheme 5).
equally effective with both aromatic and aliphatic substituents
on the allylic substrate. Table 3 summarizes our results with
several ligands. As seen previously, TADDOL-derived ligand
4 is not efficient in this series. Biphenol ligands 5 and 7
work equally well with high ee’s (68%). Similarly, ligand 8
and its diastereomer 8′ afford high ee’s but show a match/
mismatch effect,11 as for the conjugate addition.
Table 3. Allylic Substitution of
(3-Chloropropenyl)cyclohexane by i-PrMgBr and 1% CuTC as
Catalysta
Scheme 5
entry
ligand
Cu salt
% convnb
γ/R
% ee
1
2
3
4
5
6
7
8
9
4
4
5
5
6
7
8
8
CuCN
CuTC
CuCN
CuTC
CuTC
CuTC
CuCN
CuTC
CuTC
CuTC
100
100
100
100 (95%)
100
100
100
100
100
96/4
83/17
99/1
99/1
18/82
98/2
99/1
99/1
97/3
96/4
2, R
13, R
51, S
68, S
18, S
68, S
49, R
74, R
63, S
12, R
In conclusion, thanks to new Cu salts and new ligands,
we have extended the enantioselective Cu-catalyzed γ-allylic
substitution with Grignard reagents to several aryl- and alkyl-
substituted substrates. The ee values are the best known for
Grignard reagents, which are cheaper and more available
commercially than diorganozinc reagents. We have also
pointed out the unknown effect of the ligand on the
regioselectivity. Finally, we have demonstrated the synthetic
versatility of the method by carrying out a tandem substitu-
tion-metathesis reaction in a one pot procedure.
8′a
9
10
100
a 8′ is the (Sa,SS) diastereomer of ligand 8. b In parentheses: isolated
yield of the mixture of regioisomers.
All the above allylic substitution reactions end up with a
terminal vinyl group. This group is usually further trans-
formed in a second step, by cleavage of the double bond or
hydroborated and oxidized. Another possibility, not yet
explored, could be the metathesis reaction. Accordingly, the
intermolecular reaction was attempted, according to Grubbs,13
without loss of optical purity, giving 11, a useful new chiral
synthon (Scheme 4).
An interesting aspect would be to be able to perform the
whole procedure in one pot. However, nothing was known
about the compatibility of Grubbs catalyst with excess
Grignard reagent and copper salts. Accordingly, 3 equiv of
Acknowledgment. We thank Stephane Rosset and Gae¨lle
Lisci for help and the Swiss National Research Foundation
No. 20-61891.00 and COST action D12/0022/99 for financial
support.
Supporting Information Available: Experimental details
and spectral data for all compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
(13) Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18.
Org. Lett., Vol. 4, No. 23, 2002
OL0269244
4149