Organic Letters
Letter
a
Table 1. Optimization of the Reaction Conditions
1
a/2a
3a
b
c
b
c
entry
ligand
metal salts
Ni(OTf)2
Ni(OTf)2
yield (%)
1a:2a
ee (%)
yield (%)
ee (%)
1
2
3
4
L-PiCHPh2
L-PimtBu
L-PiPr3
L-PiPr3
L-PiPr3
52
69
89
75
50
84
26/74
5/95
99/1
98/2
93/7
99/1
29/56
−/11
8/−
29
32
14
24
46
−
92
38
98
99
97
−
Ni(OTf)2
Ni(ClO ) ·6H O
26/−
93/−
0/−
4
2
2
d
5
Ni(ClO ) ·6H O
4
2
2
6
tBu-box
Ni(OTf)2
a
Unless otherwise noted, the reactions were performed with ligand/metal salts (1:1, 10 mol %), 1a (0.1 mmol), and Et N (1.2 equiv) in EtOAc
3
b
c
(
1.0 mL) at 50 °C for 48 h. Isolated yield, and the ratio of 1a to 2a in the bracket represents was determined via HPLC analysis. Determined by
d
HPLC analysis on a chiral stationary phase. The reaction was performed in 0.3 mL of EtOAc.
and stereoselectivity was the foreseeable challenge in
developing such a synergistic kinetic resolution/[2,3]-Wittig
rearrangement reaction having several competitive side
reactions. On the other hand, the propargylic ethers show
low reactivity in that the alkyne sp centers distort the
transition-state geometries. Herein, we wish to disclose the
slightly better isolated yield for allene (entry 4). To our
delight, when the reaction was performed with a higher
concentration, the α-propargyloxy ether 1a was recovered in
50% yield (1a/2a, 93/7) with 93% ee, and allene 3a was
separated in 46% yield with 97% ee (entry 5; see SI for further
II
details). In comparison, the Ni /bisoxazoline catalyst was not
II
results in this endeavor. A chiral N,N′-dioxide-Ni com-
able to promote the titled rearrangement reaction (entry 6).
Having identified the optimized reaction conditions in hand,
we investigated the substrate scope of the reaction. Due to the
separation problem, only the allene products 3 were isolated
and fully characterized. As depicted in Scheme 2, varying the
diester groups affected the reactivity. Substrate 1a with methyl
ester and substrate 1b bearing ethyl ester had higher activity
than substrate 1c with sterically congested tert-butyl ester, but
their products 3a−3c were isolated in high enantioselectivities
(97−99% ee). Aryl substituents at the terminal position α-
propargyloxy ethers were investigated, and the position of the
substituents had a limited impact on the yield and selectivity
(3e−3g, 43−45% yields, and 97−99% ee). Substrates either
with an electron-donating substituent (methyl or methoxyl) or
with an electron-withdrawing substituent (chloro, ester, and
trifluoromethyl) underwent kinetic resolution/asymmetric
[2,3]-Wittig rearrangement smoothly and afforded the
corresponding α-allenyl alcohols (3d, 3h−j) in good yields
(42−47%) with high enantioselectivities (91−98% ee).
Moreover, 1-naphthyl-substituted product 3k and 2-thiophen-
yl-containing product 3l were formed with good ee values
1
5,16
plex
was found to be efficient in promoting asymmetric
propargyl [2,3]-Wittig rearrangement via kinetic resolution of
α-propargyloxy dicarbonyl compounds with high to excellent
enantioselectivities. This methodology was further applied in
the asymmetric synthesis of a chiral 2,5-dihydrofuran
derivative.
Initially, we commenced our investigation with the kinetic
resolution of racemic α-propargyloxy dicarbonyl compound 1a
by [2,3]-Wittig rearrangement (Table 1). L-PiCHPh /Ni-
2
6c
(
OTf) , which is efficient in our previous work, was tested in
2
EtOAc (1.0 mL) at 50 °C in the presence of Et N. [2,3]-Wittig
3
rearrangement of the racemic compound 1a took place
smoothly, giving the corresponding chiral α-hydroxyallene 3a
in 29% yield with 92% ee (entry 1). Meanwhile, the unreacted
α-propargyloxy ether 1a along with dihydrofuran byproduct 2a
was isolated together (the ratio of 1a to 2a, ca. 26:74, 29% ee
17
for 1a, 56% ee for 2a). Due to a similar R value, column
f
separation of 1a and 2a was unsuccessful. Next, other metal
2
no better results were given (see SI, p S4 for further details).
Then, chiral N,N′-dioxide ligands were examined (entries 2
and 3), and it was found that both the amino acid skeleton and
steric hindrance of amide moiety of the ligands had a
significant influence on the regio- and enantioselectivity.
When L-PimtBu was used, the separated major compound
was 2a (1a:2a = 5:95, 69% yield in total, 11% ee for 2a), and
allene 3a was isolated in 32% yield with 38% ee (entry 2). In
2
(90% ee and 95% ee). Regardless of the steric hindrance of R
groups at the stereogenic center (isopropyl, n-butyl, and
benzyl), all these propargylic ethers could give excellent
enantioselectivities (3m−3o, 96%−98% ee). The absolute
crystallographic analysis (CCDC 1951577).
To further demonstrate the potential of this protocol, we
first carried out the gram-scale successive kinetic resolution/
[2,3]-Wittig rearrangement of racemic 4-phenylbut-3-yn-2-ol
derived 1a for the synthesis of both enantiomers of chiral α-
allenyl alcohol 3a (Scheme 3a). By treatment of 8.0 mmol of
sharp contrast, the use of L-PiPr prevented the formation of
3
dihydrofuran 2a, albeit the allene 3a was afforded in 14% yield
with 98% ee (entry 3). Switching Ni(OTf) to Ni(ClO ) ·
2
4 2
6
H O resulted in the improved ee value for recovered 1a and a
2
B
Org. Lett. XXXX, XXX, XXX−XXX
Xu, Xi
