90% yield, entry 7) but also exhibited good to excellent
stereocontrol. For entries 4 and 9, the enantioselectivity was
reversed because the absolute configurations of the chiral
ligands were opposite. The addition of activated 4 Å
molecular sieves did not have an obvious effect on the
reaction (entry 8). The loading of Lewis acid could be
reduced to as low as 20 mol % with no loss in ee (entries 7
vs 9), and up to 91% ee was obtained for the ene cyclization
of 1c. When [Cu((S,S)-Ph-box)](OTf)2 and Sc[(R,R)-Ph-
pybox)](OTf)3 were used as the catalyst, the diastereomeric
ratio was greater than 50:1 (entries 4 and 7-9).
tions between substituents on the olefinic CdC double bond
and the phenyl groups of the chiral ligand (S,S)-Ph-box, the
ene cyclization from the re-face (transition states A and B)
should be more favorable than that from the si-face (not
shown). In addition, because of the lack of steric interactions
between methyl substituent on the olefinic CdC double bond
and the phenyl group, transition state B would be favored
over A, resulting in the cyclization product of (1R,2R)-
configuration and a cis relationship between the 1-hydroxyl
group and the 2-alkyl group.
Several other substrates 1d-f were tested under the
aforementioned carbonyl ene cyclization conditions (Table
3). In the presence of 0.2 equiv of Lewis acid Cu(OTf)2
These ene cyclizations were found to be solvent dependent.
For catalyst [Cu((S,S)-Ph-box)](OTf)2, CH2Cl2 was a better
solvent than Et2O (entries 9 vs 10), whereas compound 2c
was not obtained in THF (entry 11). The counterions8 also
affected the catalyst efficiency of the Cu(II) Lewis acids.
When the counterion was changed from OTf- to noncoor-
dinating SbF6-, the yield of 2c decreased dramatically (entries
8 vs 12). Compared to Cu[(S,S)-t-Bu-box](OTf)2, the use of
Table 3. Asymmetric Carbonyl Ene Reaction of 1d-fa
9
catalyst Cu[(S,S)-t-Bu-box)(H2O)2] (SbF6)2 did not give
much improvement to the yield of 2c (entries 6 vs 13).
Therefore, in CH2Cl2 at room temperature, the catalysts [Cu-
10
11
((S,S)-Ph-box)](OTf)2 and Sc[(R,R)-Ph-pybox)](OTf)3
were found to be efficient for the intramolecular carbonyl
ene reactions of 1c.
The observed stereoselectivity may be explained by
invoking the transition state models proposed by Jørgensen
et al. for the intermolecular carbonyl ene reactions (Figure
1).12 The [Cu((S,S)-Ph-box)](OTf)2 complex is assumed to
time
ligand
(equiv)
yield of
dr
ee of 2
entry substrate (h)
2 (%)b
(2:3)b
(%)c
1
2
3
4
5
6
7
8
9
1d
1df
1d
1eg
1e
1e
1fg
1f
24
9
55 (89)d,e 51:11
(S,S)-L1 (0.55) 87 (95)d >50:1
75
71
93
97
24 (S,S)-L1 (0.22) 78 (91)d
46:1
7.3:1
24:1
1.3:1
8:1
4
4
76
(S,S)-L1 (0.22) 91
12 (R,R)-L1 (0.22) 54
5
5
87
86
98.3
99.3
98.3
(S,S)-L1 (0.22) 94
34:1
1.3:1
1f
12 (R,R)-L1 (0.22) 56
a Unless otherwise indicated, all reactions were carried out at room
temperature in CH2Cl2 with 0.1-0.2 mmol of substrate and 0.2 equiv of
b
Cu(OTf)2. 1H NMR yield with R-methyl stilbene as the internal standard.
1
Ratio of 2 and 3 was determined by H NMR analysis of crude products.
Compounds 2 and 3 were separable by flash column chromatography.
c Enantiomeric excess of 2 was determined by HPLC analysis using a Chiral
OD or AD column. Relative configuration of 2d was determined by the
analysis of NOESY spectra of its diol derivative (see Supporting Informa-
tion), but its absolute configuration was not determined. Absolute configura-
tions of 2e/2f as (1R,2R,5R)- and 3e/3f as (1S,2S,5R)- were determined by
NOESY analysis. d Percentage conversion in parentheses. e Byproduct 4d
was isolated in 27% yield. f Performed with 0.5 equiv of Lewis acid. g Ee
values of 1e and 1f were not determined.
Figure 1. Proposed transition-state model for the {Cu[(S,S)-Ph-
box]} complex-promoted carbonyl ene cyclization reaction of 1c.
chelate with the dicarbonyl moiety of the substrate in a
tetrahedral-like geometry.12a,c Considering the steric interac-
(8) For studies on the counterion effects of copper(II) complexes, see:
Evans, D. A.; Murry, J. A.; Van Matt, P.; Norcross, R. D.; Miller, S. J.
Angew. Chem., Int. Ed. Engl. 1995, 34, 798-800.
(9) Evans, D. A.; Peterson, G. S.; Johnson, J. S.; Barnes, D. M.; Campos,
K. R.; Woerpel, K. A. J. Org. Chem. 1998, 63, 4541-4544.
(10) Same catalyst gave excellent enantioselectivity in the intermolecular
carbonyl ene reactions; see refs 1f,j and 13a.
without ligand, the ene cyclization of 1d gave cyclopentane
products 2d and 3d in poor yield (55%), together with a
double bond-rearranged product 4d (entry 1). However, 0.5
equiv of chiral Lewis acid [Cu((S,S)-Ph-box)](OTf)2 cata-
lyzed the cyclization of 1d in good yield and ee (87 and
75%, respectively; entry 2). Reducing the loading of the
chiral Lewis acid led to a slightly decreased yield (78%)
and ee (71%) (entries 2 vs 3). Furthermore, the addition of
chiral ligands gave improved diastereoselectivity (entries
1-3).
(11) Evans, D. A.; Sweeney, Z. K.; Rovis, T.; Tedrow, J. S. J. Am. Chem.
Soc. 2001, 123, 12095-12096.
(12) For discussion on the metal center geometry of chiral bis(oxazoline)-
copper(II) complexes, see: (a) Johannsen, M.; Jørgensen, K. A. J. Org.
Chem. 1995, 60, 5757-5762. (b) Evans, D. A.; Johnson, J. S.; Burgey, C.
S.; Campos, K. R. Tetrahedron Lett. 1999, 40, 2879-2882. (c) Thorhauge,
J.; Roberson, M.; Hazell, R. G.; Jørgensen, K. A. Chem. Eur. J. 2002, 8,
1888-1898.
Org. Lett., Vol. 5, No. 20, 2003
3751