Organic Letters
Letter
a
Table 1. Optimization of Reaction Conditions
b
entry
catalyst
ZnBr2
ZnBr2
ZnCl2
FeCl3
Sc(OTf)3
Yb(OTf)3
In(OTf)3
Cu(ClO4)2·6H2O
Zn(ClO4)2·6H2O
ZnBr2
time (h)
yield (conversion) (%)
c
1
24
24
24
24
24
24
24
24
24
48
48
17
48
41 (100)
74 (92)
51 (86)
39 (87)
65 (91)
69 (90)
63 (88)
54 (99)
64 (83)
2
3
4
5
6
7
8
9
d
10
69 (100)
e
d
11
12
13
ZnBr2
ZnBr2
ZnBr2
81 (100)
e f
d
,
77 (100)
g
d
83 (100)
a
Reaction conditions: 1a (0.10 M, 1.0 equiv), 2a (1.1 equiv), catalyst
b
(10 mol %), and 4 Å MS (100.0 mg) in DCM (1.0 mL), 25 °C. The
conversion and the yield were determined by the H NMR spectrum
1
of the crude mixture with 1,1,2,2-tetrachloroethane (TTCE) as an
c
d
e
internal standard. Run without 4 Å MS. Isolated yield. Run with
f
g
0.24 M of 1a in DCM (1.0 mL). Run at 35 °C. Run with 0.24 M of
2a (1.0 equiv) and 1a (1.1 equiv) in DCM (1.0 mL).
single-crystal X-ray analysis.14 Note that the relative stereo-
chemical outcome between the ester group and the electron-
donating substituent in the alkene components was completely
reversed, when compared to those reported in the work of the
France group,8a possibly influenced by the cyclic structure.
Moreover, the addition of 4 Å MS improved the yield to 74%,
although the conversion of 1a was only 92% (Table 1, entry 2).
Other Lewis acids were then screened in the presence of 4 Å
MS. While Yb(OTf)3 showed an activity comparable to that of
zinc bromide, other frequently used Lewis acidic metal
chlorides, triflates, and perchlorate hydrates resulted in inferior
yields of 3aa with similar conversions (Table 1, entries 3−9).
Within an extended time, 1a was fully consumed; however, the
yield of 3aa changed slightly (Table 1, entry 10). A screening
of the concentration of 1a showed that the reaction performed
at a higher concentration gave 3aa in 81% isolated yield (Table
1, entry 11).14 Warming the reaction at 35 °C remarkably
accelerated the reaction, but a slight decrease in the yield was
observed (Table 1, entry 12). To our delight, 3aa was also
isolated in good yield when N-acryloyl indole 2a, which is the
more-complex C5 component, was used as the limiting reagent
(Table 1, entry 13). No other diastereomers were observed in
the reactions examined.
With the optimal reaction conditions in hand, the scope of
the [5 + 2] annulation with various N-tosyl enamides was
studied. As shown in Figure 2, N-tosylated 2,3-dihydropyrroles
with 4-alkyl substituents, such as methyl (1a), ethyl (1b), n-
propyl (1c), allyl (1d), and benzyl (1e), reacted smoothly with
2a at 25 °C, giving the corresponding azepino[1,2-a]indoles
(3aa−3ea) as the single diastereomer in good to excellent
yields. The allyl group in 3da could allow for further
elaboration. Interestingly, N-tosyl-2,3-dihydropyrrole (1f) was
also suitable for the [5 + 2] annulation, but 3fa was obtained in
Figure 2. Scope for the [5 + 2] annulation: Unless otherwise
specified, 1 (0.26 mmol), 2 (0.24 mmol), ZnBr2 (10 mol %), and 4 Å
MS (100.0 mg) in DCM (1.0 mL), 25 °C, 48 h. Isolated yield.
[Footnote a indicates conditions including In(OTf)3 (15 mol %) and
4 Å MS (30.0 mg).]
57% yield with a moderate diastereoselectivity, indicating the
important role of the 4-substituents in the cyclic N-sulfonyl
enamides for the excellent diastereomeric ratio (dr) selectivity.
Unfortunately, 4-arylated N-sulfonyl 2,3-dihydropyrroles were
unreactive, possibly because of the reduced reactivity of the
double bond caused by the conjugate effect of aryls. In
addition, this annulation was sensitive to the ring size of the
cyclic enamide. For example, the reaction with the six-
membered ring enamide 1g was much slower even warmed
at 35 °C. Moreover, like the acyclic sulfonyl enamides, the
reaction with the seven-membered ring enamide N-tosyl-6-
methyl-2,3,4,5-tetrahydro-1H-azepine produced a complicated
mixture.
A wide range of functional groups on the indole ring of N-
acryloyl indoles 2 were well-tolerated in the ZnBr2-catalyzed [5
+ 2] annulation (Figure 2). Specifically, 3-methylated
substrates 2 with halo (2b, 2e−2g, and 2j), methyl (2c and
2h), and methoxy (2d and 2i) substituents on the 4-, 5-, or 6-
positions of the indole ring reacted readily with 1a under the
optimal reaction conditions, diastereoselectively affording the
corresponding products (3ab−3aj) in yields up to 90%.
Moreover, the 3-methyl group in 2a could be replaced by ethyl
(2k) or the β-functionalized ethyl groups (2l−2n) without the
significant loss of the product yields. Interestingly, the 3-
unsubstituted indole 2o was less reactive and only 51% yield of
3ao was obtained. Given the promising further cyclization of
the tethered alkyl halides in 3am and 3an with the pyrrolidine
B
Org. Lett. XXXX, XXX, XXX−XXX