Organic Letters
Letter
carbonate or the alkoxide anion generated during the
formation of the π-allyl rhodium intermediate could be used
as the mild base to activate the neutral pronucleophiles. This
small amount of gradually released base is expected to
L2 with methyl and benzyl respectively gave the desired
product 3aa in low yield, branch/linear ratio, and enantiose-
lectivity (entries 1 and 2). A large amount of side product 3aa′
10
was obtained. However, L3 and L4 with bulkier isopropyl
and phenyl groups lead to moderate yields, above the 20:1
branch/linear ratio and 98% ee for 3aa (entries 3 and 4). To
our delight, full conversion of the allylic carbonate 1a was
observed when L5 with a tert-butyl group was applied (entry
B
selectively remove the H in the pyridinium salts to achieve
the interrupted Tschitschibabin reaction. Herein, we report a
Rh/NPN*-catalyzed highly regio- and enantioselective allylic
alkylation of pyridinium salts and base-promoted cyclization
sequence to prepare 3-substituted indolizines (Scheme 1C).
Chiral 3-allylindolizines bearing aliphatic groups could be
synthesized from 2-alkylpyridines and α-bromoketones. Above
the 20:1 branch/linear ratio, up to a 96% yield and 99% ee
could be obtained with the help of the chiral bisoxazoline-
phosphine ligand with tert-butyl groups on the oxazoline rings.
We started the study with racemic allylic carbonate 1a and 2-
methyl-1-(2-oxo-2-phenylethyl)pyridin-1-ium bromide 2a as
the model substrates (Table 1). The reactions were conducted
5
), which might be caused by the noncovalent dispersion
11
interaction between the ligand and the substrate. Compound
aa was isolated in 95% yield, >20:1 b/l ratio, and 99% ee. The
3
1
effect of the R group on the phosphine atom was further
investigated. For highly effective ligands with a tert-butyl group,
the electron-donating or electron-withdrawing groups at the R
1
position have little effect on the outcome of the reactions
(
entries 6 and 7). Nevertheless, an inferior effect of the 4-OMe
group in L8 was observed by comparison to L4 with the
2
isopropyl group at R (entry 8). The reaction at room
a
Table 1. Optimization of the Reaction Conditions
temperature was less efficient (entry 9). Higher temperature
(
60 °C) caused the faster consumption of 2a to form more
cyclization byproduct 3aa′ (entry 10). CH CN was proven to
3
be superior compared to other solvents examined.
With the optimized reaction conditions in hand (entry 5,
Table 1), we then investigated the scope of allylic carbonates
(
Scheme 2). The reactions of allylic carbonates with simple
methyl, phenylethyl, and benzylether protected alkyl groups
afford the desired chiral 3-allylindolizines in high yields and
99% ee (3ba, 3ca, and 3da). The formation of 3ba could be
conducted in 5 mmol scale without erosion of reactivity and
a
Scheme 2. Reaction Scope of Allylic Carbonates
b
c
d
e
entry
L*
3aa/3aa′
yield (%)
ee (%)
B/L
1
2
3
4
5
6
7
8
L1
L2
L3
L4
L5
L6
L7
L8
L5
L5
0.25:1
0.24:1
1.38:1
1.05:1
3.85:1
2.99:1
2.64:1
0.88:1
2.15:1
1.75:1
23
21
65
61
95
88
85
51
82
76
33
23
98
98
99
>99
99
98
99
99
3:1
2:1
>20:1
>20:1
>20:1
13:1
>20:1
12:1
>20:1
>20:1
f
9
1
g
0
a
All reactions were run with 2.5 mol % catalyst precursor and 6 mol %
ligand on a 0.3 mmol scale at 40 °C for 24 h unless otherwise noted,
and the reactions with 4 equiv Cs CO were conducted in the
2
3
b
1
c
presence of air. The ratios were determined by H NMR. Yield of
isolated product. The enantiomeric excess values were determined
by HPLC analysis with a chiral column. The ratios of branch
products to linear products were determined by H NMR. The
d
e
1
f
g
reaction was carried out at room temperature. The reaction was
a
All reactions were run with 2.5 mol % catalyst precursor and 6 mol %
carried out at 60 °C.
ligand on a 0.3 mmol scale at 40 °C for 24 h unless otherwise noted,
and the reactions with 4 equiv of Cs CO were conducted in the
presence of air. Reaction at 5 mmol scale. The ee value was
determined after the hydroboration/oxidation sequence. 1.5 equiv of
allyl carbonate, 1.0 equiv of 2a, and 3 equiv of BSA were used and the
reaction was run for 72 h. 1.5 equiv of allyl carbonate, 1.0 equiv of 2a,
and 3 equiv of BSA were used and the reaction was run for 48 h. 2.0
equiv of allyl carbonate, 1.0 equiv of 2a, and 3 equiv of BSA were used
2
3
b
c
with 2.5 mol % [Rh(cod)Cl] and 6 mol % chiral NPN ligand
d
2
in acetonitrile (0.15 M) at 40 °C for 24 h. A 4 equiv amount of
Cs CO was then added into the reaction tube to promote the
e
2
3
2
f
cyclization. The substituent R on the oxazoline rings in the
NPN ligands has a significant influence on the reactivity and
selectivity of the reaction. Reactions in the presence of L1 and
and the reaction was run for 72 h.
B
Org. Lett. XXXX, XXX, XXX−XXX
Li, Ke
