Organic Letters
Letter
a
ketols, and various 2-aryl-2-hydroxy cyclohexanones were
readily afforded in moderate to good yields with high
enantioselectivities. Moreover, under slightly modified con-
ditions, the asymmetric isomerizations of acyclic α-hydroxy
aldehydes and α-iminols were achieved as well.
Scheme 2. Substrate Scope of Cyclic α-Ketols 1
Initially, 1-benzoylcyclopentanol 1a was selected as the model
substrate to optimize the reaction conditions (Table 1).6,14 First,
Table 1. Optimization of Reaction Conditions for the
Asymmetric Rearrangement of Cyclic α-Ketol 1a
a
b
c
entry
metal salt
ligand
yield (%)
ee (%)
1
2
3
4
5
6
Fe(OTf)3
In(OTf)3
Al(OTf)3
Al(OTf)3
Al(OTf)3
Al(OTf)3
Al(OTf)3
Al(OTf)3
Al(OTf)3
L3-PiMe2
L3-PiMe2
L3-PiMe2
L3-PiPr2
L3-PrPr2
L3-RaPr2
L3-PiPr2
L3-PiPr2
L3-PiPr2
48
49
47
50
39
28
69
99
98
53
27
55
87
85
69
87
91
91
a
The reactions were carried out with Al(OTf)3/L3-PiPr2 (1:1, 5 mol
%) and 1 (0.1 mmol) with H2O (3 μL) in BrCH2CH2Br (1.0 mL) at
40 °C for the indicated time. Isolated yields. Enantiomeric excess (ee)
was determined by HPLC analysis on a chiral stationary phase.
d
7
8
9
de
,
def
, ,
a
Unless otherwise noted, the reactions were performed with the metal
salt/ligand (1:1, 10 mol %) and 1a (0.1 mmol) in THF (1.0 mL) at
electron-withdrawing group (for instance, CF3) at the para-
position (2f) or methyl substituent at the meta-position (2h) in
the substrate led to diminished yields (59% yield and 67% yield,
respectively). To get more insight into the reaction, the product
2f was subjected to the standard reaction system, and the very
trace of starting compound 1f was observed; the ee value did not
change.4 It is worth mentioning that 2-naphthyl-containing
substrate 1j proceeded the migration process smoothly under
the standard conditions, affording the desired product 2j in
decent yield (63%) with good enantiomeric excess (87% ee).
Meanwhile, piperonyl-substituted α-hydroxy ketone was suit-
able as well, and the corresponding product 2k was isolated in
89% yield with 88% ee. For the reaction of 1l with the
disubstituted aryl ring, elevated reaction temperature (60 °C)
was necessary, and product 2l was obtained with good outcomes
(97% yield, 91% ee). The absolute configurations of the
products 2i and 2l were both determined to be (S) configuration
by X-ray crystallography analysis, which were consistent with the
configurations of chiral 2a, 2e, 2f, and 2j, by comparing the
optical rotation data in the previous report.16 The configurations
of other products were assigned by comparing with the CD
spectrum of compound 2i.
Encouraged by the satisfactory results of cyclic α-ketol, we
attempted to apply the current system to the rearrangement of
α-hydroxy aldehydes 3 (Scheme 3). Unfortunately, only a trace
amount of the desired rearranged product was afforded with
11% ee under the above optimized reaction conditions. After
modification of the reaction parameters, including the use of L3-
PiMe2 as the ligand and Br2CHCHBr2 as solvent, lower reaction
temperature (30 °C), and addition of 4 Å MS (10 mg), the yield
and enantioselectivity of this transformation reached a
satisfactory level (54% yield, 82% ee) after 48 h (see details in
b
c
40 °C for 24 h. Yield of isolated product. Determined by HPLC
analysis on a chiral stationary phase. BrCH2CH2Br (1.0 mL) was
used as the solvent. H2O (3 μL) was added. 5 mol % of catalyst
d
e
f
loading.
different metal salts were investigated in the presence of N,N′-
dioxide L3-PiMe2. It was pleasant to find that in situ formed
chiral Lewis acid catalysts could promote the rearrangement of
compound 1a under mild conditions,6 and Al(OTf)3 gave better
results than Fe(OTf)3 and In(OTf)3 in terms of enantiose-
lectivity (entries 1−3, 55% ee vs 53% ee, 27% ee). To our
delight, increasing the steric hindrance of the amide substituents
from 2,6-Me2C6H3 to 2,6-iPr2C6H3 resulted in higher
enantioselectivity (entry 4, 87% ee vs 55% ee). Then,
representative N,N′-dioxides with different skeletons were
examined. L3-PiPr2 derived from L-pipecolinic acid was proved
to be superior to L-proline-derived L3-PrPr2 or L-ramipril-based
L3-RaPr2 (entries 5 and 6, 87% ee vs 85% ee, 69% ee). The
subsequent survey of solvents indicated that BrCH2CH2Br
could provide an improved yield compared to THF (entry 7,
69% yield vs 50% yield; for more details, see SI, Page 7).
Interestingly, the addition of water (3 μL, ca. 1.7 equiv) into the
reaction system led to further improvement in the yield and
enantiomeric excess (entry 8, 99% yield, 91% ee.).15 When the
reaction was carried out with 5 mol % catalyst loading, both the
yield and enantioselectivity were maintained (entry 9).
With the optimal reaction conditions in hand, the substrate
scope of 1-aroylcyclopentanols was next examined (Scheme 2).
It was found that the substituent pattern and the electronic
property of the aryl moiety displayed a limited influence on the
enantioselectivity (2a−2l, 84−92% ee). However, the strong
B
Org. Lett. XXXX, XXX, XXX−XXX