Angewandte
Chemie
Table 2: Reaction scope.[a]
bisoxazoline ligands unveiled that iPr-BOX gave the best
enantioselectivity (entry 2),[8] and encouraged us to further
elaborate this framework. To our delight, the sidearm
strategy[10,11] of introducing a pendant group at the bridging
carbon atom of the bisoxazoline ligand proved to be quite
effective in this reaction, and the enantioselectivity of the
reaction was gradually improved to 97% ee by increasing the
steric hindrance at the para position of the aryl sidearm
groups (entries 3–7). It is worth noting that the racemic D–A
cyclopropanes 2a and 2b were used in this reaction, thus
suggesting that the current reaction proceeds with a dynamic
kinetic resolution[9b–d] of the cyclopropanes. Reaction accel-
eration was consistently observed with the sidearm-modified
ligands (entries 3–7 versus entries 1 and 2). In contrast,
alteration of the ester group of 2 indicated that a large ester
group is also important to the enantioselectivity with 2-
adamantyl (Ad) being the best (entry 7 versus 8).[12] It is
noteworthy that the TBDPS protecting group proved to be
crucial to suppress the formation of the ring-opened product 4
and ensure a high yield of the desired [3+2] product (entries 9
and 10). For example, in the reactions of the TBDPS enol
ether 1a, only a trace or a small (< 10%) amount of the ring-
opened product can be detected by proton NMR spectrosco-
py, which is in contrast to reaction of the TIPS-protected enol
ether 1b which exclusively produced the ring-opened product
(entry 9). The formation of 4 may result from the hydrolysis of
the carboxonium intermediate before the ring closure or the
abstraction of its a proton followed by desilylation.[5,6] In
addition, other metal salts such as Cu(SbF6)2, Cu(OTf)2,
CuBr2, Ni(ClO4)2, and Co(ClO4)2 were also examined. Only
Cu(SbF6)2 and Cu(OTf)2 showed a good activity, but resulted
in slightly lower enantioselectivities.[8]
Entry
1
R2
P
Yield [%]
(d.r.)[b]
ee
[%][c]
1
2
4-MeOC6H4 (2a)
2-thiophenyl (2c)
(3,4,5-MeO)3C6H2 (2d) 3ad 78 (91:9)
3aa 85 (87:13)
3ac 62 (87:13)
97
93
99
92
3[d]
4
=
CH CHPh (2e)
3ae 74 (80:20)
5
4-MeOC6H4 (2a)
4-MeOC6H4 (2a)
4-MeOC6H4 (2a)
3da 94 (81:19)
91
92
6[e]
3ea 70 (85:15)
7
3 fa 61 (>99:1) 95
8
9
4-MeOC6H4(2a)
2-thiophenyl (2c)
3ga 87 (>99:1) 93
3gc 83 (>99:1) 93
10
(3,4,5-MeO)3C6H2 (2d) 3gd 81 (>99:1) 97
11[f,14]
12[g]
CH CHPh (2e)
2-furyl (2 f)
3ge 45 (>99:1) 95
3gf 68 (>99:1) 92
=
13[e,14]
4-MeOC6H4 (2a)
3ha 45 (>95:5) 98[h]
Having the optimized reaction conditions, we moved to
the investigation of the reaction scope. As shown in Table 2,
apart from a PMP-substituted D–A cyclopropane, 2-thio-
phenyl- (2c), 3,4,5-trimethoxyphenyl- (TMP, 2d), and
alkenyl-substituted (2e) cyclopropanes also reacted well
with 1a, thus giving the desired fused cyclic products in
good to excellent enantiomeric excesses (entries 1–4). The
reaction of the TMP-substituted cyclopropane was slower,
and could probably be ascribed to the steric hindrance of the
TMP group or the competing bidentate coordination of the
methoxy groups to the copper (entry 3). Notably, both
conjugate and isolated dienol-ether-type substrates (1d and
1e) also react smoothly under the standard reaction con-
ditions, thus affording carbon–carbon double-bond-contain-
ing [5.3.0]bicyclic products, with the potential for additional
transformation of this functionality (entries 5 and 6). To our
great pleasure, the current reaction works quite well with six-
and five-membered substrates (entries 7–13). Thus, a range of
[n.3.0] (n = 3–5) 3a-hydroxy bicycles in high enantiomeric
purity are accessible by this reaction. Remarkably, in the
reactions with five- and six-membered enol silyl ethers,
perfect diastereoselectivities (> 99:1) were consistently
observed (entries 7–12). In fact, in these cases only one
diastereoisomer was detected by 1H NMR analysis of the
crude reaction mixture. In addition, the ß-disubstituted enol
silyl ether 1h is also a suitable substrate with high enantio-
selectivity, and notably the corresponding product (3ha)
[a] Reaction conditions: [Cu(ClO4)2·6H2O] (0.02 mmol), L7
(0.022 mmol), 1 (0.30 mmol), and 2 (0.20 mmol) in 2.0 mL of CH2Cl2 at
308C with 4 ꢀ M.S. and N2. [b] Yield of the isolated product. A trace
amount of ring-opened product 4 was observed. The d.r. values (cis/
1
trans) were determined by H NMR spectroscopy. [c] Values for the cis
products as determined by HPLC using a chiral stationary phase.
[d] Using 3 equiv of 1a. [e] Cu(SbF6)2. [f] Cu(OTf)2, Ref. [14]. [g] Cu-
(OTf)2, 08C, in 1,1,2,2-tertrachloroethane. [h] When Cu(SbF6)2 was used,
the yield was 70% with 91% ee. Ad=2-adamantyl, M.S.=molecular
sieves, Tf=trifluoromethanesulfonyl.
contains three contiguous quaternary centers, two of which
are at the ring juncture (entry 13). The configurations of the
products 3aa, 3da, and 3gc were determined by X-ray crystal
structure analysis,[13] and the configurations of other products
were assigned by analogy.
The success with monocyclic enol silyl ethers encouraged
us to extend the current method to benzocyclic enol silyl
ethers, which can deliver the benzene-fused cyclic products.
As shown in Scheme 1, the reaction works extraordinarily
well with these benzene-fused substrates, such as 3,4-dihy-
dronaphthalen-1-one- and 1-indanone-derived silyl enol
ethers (5a–f), and different substitution patterns such as 6-
bromo, 6-fluoro, 7-methyl, and 5-methyl substituents are
tolerated under the standard reaction conditions. In all cases
high yield (80–98%) and complete diastereoselectivity
Angew. Chem. Int. Ed. 2013, 52, 4004 –4007
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim