Letter
Table 1. Optimization of the Reaction Conditions
Figure 1. Selected medicinally important spirocyclohexaneoxindole
and spiro-δ-lactone skeletons.
carried out the systemic examination of substrates using well-
9
defined chiral N,N′-dioxide/metal salt complex as the catalyst.
Herein, we report our results in this area that chiral
spirocyclohexaneoxindoles were readily available from (E)-
alkenyloxindoles via a highly diastereo- and enantioselective [4
a
a
a
a
entry
1
2
1
ligand
yield (%)
dr (%)
ee (%)
1a
1a
1a
1a
1a
1a
1a
1a
1b
1c
1c
L-PrPr2
L-PrPr2
55 (3aa)
32 (3aa)
50 (3aa)
32 (3aa)
56 (3aa)
59 (3aa)
76 (3aa)
64 (3aa)
75 (3ba)
72 (3ca)
99 (3ca)
85:15
84:16
>19:1
92:8
>19:1
94:6
92:8
94:6
>19:1
>19:1
>19:1
−39
0
−34
−48
61
b
+
2] cycloaddition. A hetero-Diels−Alder reaction occurred at
10
3
4
5
6
7
L-PiPr
2
the CO bond of (E)-dioxopyrrolidines, affording spiro-δ-
lactones with good results (up to 85% yield, 97% ee). In
addition, a [2 + 2] cycloaddition performed from isatin-derived
L-RaPr2
L-RaEt2
L-RaMe
L-RaPh
3a
2
72
0
73
87
92
93
imine rather than an aza-[4 + 2] process generated a β-
lactam adduct in high yield with moderate enantiomeric excess
c
8
L-RaMe2
L-RaMe2
L-RaMe2
L-RaMe2
(
ca. 60% ee).
,
cd
9
1
1
Initially, the ring-opening/cycloaddition of cyclobutenone
c
c
,
,
d
d
0
1
2
a with (E)-alkenyloxindole 1a was selected as the model
,
e
reaction to optimize the reaction conditions (Table 1). A
variety of metal salts were examined by coordinating with L-
a
The reactions were performed with L/Dy(OTf) (10 mol %), 1a
3
(
0.10 mmol), 2a (1.0 equiv), and 4 Å MS (50 mg) in DCE (0.1 M) at
proline derived N,N′-dioxide L-PrPr in DCE at 60 °C (see the
2
6
0 °C for 16 h. Isolated yield of the product. The dr values were
SI for more details). It was found that the complex of
1
determined by H NMR analysis, and ee values were determined by
Dy(OTf) with L-PrPr could promote the reaction smoothly
b
3
2
HPLC analysis on a chiral stationary phase. Sc(OTf) instead of
Dy(OTf) . 4 Å MS (80 mg). At 50 °C for 48 h. DCE (0.2 M), 2a
(
3
to afford the corresponding [4 + 2] cycloaddition product
3aa) in 55% yield, 85:15 dr and 39% ee (Table 1, entry 1). In
c
d
e
3
(
1.5 equiv).
sharp contrast, Sc(OTf) as the metal precursor only gave the
3
racemic product (Table 1, entry 2). As such, Dy(OTf) was
3
chosen as the central metal to screen the chiral N,N′-dioxide
increase of the yield (99%) with 93% ee (Table 1, entry 11).
Therefore, the optimal reaction conditions were established as
1c, 2a (1.5 equiv), Dy(OTf) /L-RaMe (1:1, 10 mol %), 4 Å
ligands. It was indicated that L-RaPr derived from L-ramipril
2
was superior to L-PrPr and L-pipecolic acid derived L-PiPr in
2
2
3
2
terms of enantioselectivities (Table 1, entry 4 vs entries 1 and
). Interestingly, decreasing the steric hindrance of the amide
MS, in DCE (0.2 M) at 50 °C for 48 h.
3
Under the optimized reaction conditions, the scope of (E)
-alkenyloxindoles was then investigated with cyclobutenone 2a.
As depicted in Table 2, the position of substitution on the
phenyl ring had a significant influence on the enantioselectivity
of the reaction. First, variant of the groups at C5- and C6-
positions in (E)-alkenyloxindoles exhibited that regardless of
electronic nature of the substituents all of the reactions
proceeded well, delivering the desired products 3da−3ka in
good yields (86−99%) with moderate to high enantioselectiv-
ities (79−92% ee) (Table 2, entries 2−9). However, when C4-
and C7-substituted (E)-alkenyloxindoles 1l−1n were subjected
into the current system, only moderate enantioselectivities
(48−79% ee) were observed, probably due to the effect of
i
substituents from 2,6- Pr C H to 2,6-Et C H (L-RaEt ) or
2
6
3
2
6
3
2
2
,6-Me C H (L-RaMe ) led to higher enantioselectivity, but
2 6 3 2
with reversal configuration, and L-RaMe exhibited a better
2
performance (Table 1, entry 6, 59% yield, 94:6 dr, and 72%
1
1
ee). However, the use of L-RaPh with a simple phenyl group
delivered a racemic product (Table 1, entry 7), implying the
important role of the 2,6-substituents at the phenyl group.
Increasing the amount of 4 Å molecular sieves enhanced both
the isolated yield and enantioselectivity (Table 1, entry 8 vs
entry 6). Other parameters such as solvents and additives were
investigated as well; however, no better results were achieved
(
see the SI for details). When the substrate 1b bearing a large
tert-butyl ester group was used instead of 1a at 50 °C for 48 h,
the corresponding spirocyclohexaneoxindole 3ba was obtained
in 75% yield and improved enantiomeric excess (87% ee;
Table 1, entry 9). To our delight, a better result was afforded
when 1c with a 3-fluoro-substituted benzoyl group at nitrogen
was employed (92% ee; Table 1, entry 10). Finally, enhancing
the reaction concentration (0.2 M) and using slightly excessive
amount of cyclobutenone 2a (1.5 equiv) resulted in an obvious
steric hindrance (Table 2, entries 10−12). Additionally, when
1
R was CO Et or COPh, the product 3oa or 3pa was obtained
2
in good results (3oa, 88% yield, 86% ee; 3pa, 99% yield, 91%
ee; Table 2, entries 13 and 14). The gram-scale reaction of 1c
(3.0 mmol) and 2a (4.5 mmol) proceeded well under the
optimized reaction conditions, delivering the corresponding
product 3ca in 99% yield (>19:1 dr, 1.53 g) with 93% ee
(Table 2, entry 1). Furthermore, the absolute configuration of
B
Org. Lett. XXXX, XXX, XXX−XXX
Luo, Yao
