Organic Letters
Letter
Table 1. Optimization of the Diastereoselective
Peroxidation Reaction
Scheme 2. Representative Synthesis and Peroxidation of
Enantiopure Lactones
a
PhSiH3
(mol %)
yield syn-7
yield 7a
a
a
entry solvent
dr 7
(%)
(%)
1
2
3
4
PhCF3
PhCF3
PhCF3
MeCN
0
5
10
0
94 : 6
97 : 3
98 : 2
98 : 2
63
66
57
73
68
10
11
12
8
7
peroxide syn-7 while maintaining high diastereoselectivity, even
without PhSiH3. Because the addition of 5 mol % of PhSiH3
resulted in lower yields (entry 5), this additive was excluded
from the optimal conditions (entry 4).
The purification method also required optimization. The use
of Co(thd)2 in cobalt-mediated peroxidation reactions results
in the formation of significant quantities of cobalt-containing
impurities that are difficult to separate from the desired
product.24 The use of high surface-area silica (Davisil-grade)
was critical for the separation of the major silyl peroxide from
the cobalt-containing impurities. After one purification, the
product was isolated without residual cobalt-containing
a
Peroxidation reaction conditions: alkene (15 or 16, 1 equiv), Et3SiH
(2 equiv), and Co(thd)2 (10 mol %) in MeCN (0.3 M) under a
balloon of O2, 3 h.
1
impurities, as evidenced by H NMR spectroscopy and visual
inspection of the resulting white solid. Purification using
standard-grade silica gel caused significant decomposition of
the product,15 likely due to an acid-catalyzed25 degradation
pathway that formed methyl ketone 9 (Figure 2).26
The overall sequence employed for the preparation of
enantiomerically pure products from racemic Baylis−Hillman
adducts is demonstrated in Scheme 2 for a substrate that bears
a side chain similar to those found in peroxide-containing
natural products.27 Ester 12, which was prepared from 3,5,5-
trimethylhexanal (10) and methyl acrylate (11), was hydro-
lyzed to give the corresponding hydroxyacid 13 (dr 50:50).
After recrystallization, a single diastereomer of hydroxyacid 13
was obtained in 33% yield. Silylation of hydroxyacid 13 with
(Me3Si)2NH provided compound 14, which was immediately
coupled to (−)-menthone (5) with catalytic quantities of
trimethylsilyl trifluoromethanesulfonate (Me3SiOTf).10−13,28
The resulting diastereomeric lactones 15 and 16 were
separated by flash chromatography.
Although the sequence was generally continued with the
major diastereomer of each lactone, the minor diastereomeric
lactone can also be used to prepare enantiomerically pure
products. When lactones 15 and 16 were subjected separately
to the optimized peroxidation conditions, the corresponding
silyl peroxide products syn-17 and syn-18 were formed in
similar diastereomeric ratios (Scheme 2). X-ray crystallo-
Figure 2. Methyl ketone decomposition product.
graphic analysis was used to establish the configuration of the
major silyl peroxide product syn-18.29 Comparison of the H
1
NMR spectra of the silyl peroxide products syn-17 and syn-18
showed that they shared the same relative configuration.29
These experiments demonstrate that the configuration at the
allylic stereocenter, not the menthone auxiliary, controlled the
stereochemical outcome of the reaction. Torsional effects30
during the addition of molecular O2 to the planar, stabilized
radical31,32 likely dictate the stereochemical outcome.15
A range of Baylis−Hillman adduct-derived alkenes under-
went this peroxidation reaction with high diastereoselectivity
(Scheme 3). The diastereoselectivity of the peroxidation
depended upon the size of the alkyl side chain. The
peroxidations of β-branched substrates, such as 6 and 16,
occurred with the highest diastereomeric ratios (silyl peroxides
7 and 18, dr ≥ 98:2). The reactions of lactones with less
B
Org. Lett. XXXX, XXX, XXX−XXX