J. Guernon et al. / Tetrahedron Letters 55 (2014) 2134–2137
2135
Table 1
Optimization of the addition of 2 to metal enolate of 8
O
S
HN
S
N
F
O
O
O
a
b,c
Br
Br
O
O
S
HN
S
N
O
N
F
O
8
O
F
F
F
Br
Br
F
2
N
3
base, THF
O
F
F
2
7
O
O
S
HN
S
HN
O
OH
d
Entry
Conditions
Isolated yield (%)
Br
Br
H
N
1
2
3
4
4
5
6
LDA, ClTi(Oi-Pr)3, À78 °C
0
0
O
F
F
F
F
LDA, ClTi(Oi-Pr)3, À78 °C to À40 °C
LDA, À78 °C
5
4
15
25
27
31
59
LDA, À78 °C to À40 °C
KHMDS, À78 °C
e
1
KHMDS, À78 °C to À40 °C
n-BuLi, À78 °C to À40 °C
Scheme 1. Reagents and conditions: (a) LDA (2 equiv), THF, À78 °C, ClTi(Oi-Pr)3
(2 equiv), À78 °C, 72%; (b) LiAlH4, ether, 0 °C, 82%; (c) Swern oxidation, 74%; (d) 4-
bromo-3,5-dimethylisoxazole, n-BuLi, À78 °C or magnesium, iodine, rt, THF; (e) 4 N
HCl, dioxane, <5% yield.
prepare the amine-free lithium enolate of ketone 8. Treatment of 8
with 1.05 equiv of n-BuLi in THF at À78 °C brought about enoliza-
tion with no appreciable formation of the adduct due to nucleo-
philic addition of n-BuLi. Reaction of this lithium enolate with
ketimine 2 at À78 °C for 30 min and then at À40 °C for 30 min gave
adduct 7 in 59% yield, a significant improvement over the reaction
with LDA or KHMDS.10 Again, the addition proceeded with high
diastereoselectivity (>20:1). The addition of the lithium enolate
to ketimine 2 was sluggish at À78 °C, so it was necessary to warm
the reaction mixture to À40 °C and maintain at that temperature
for 30 min. Also worth noting is the impact of the concentration.
In the studies reported by Davis et al., dilute concentration
(0.02 M) was typically used in the ketone enolate addition to aldi-
mines (to minimize self aldol condensation of ketones). This would
limit its applications on large scale reactions. In our case, similar
yields were obtained when the reaction was run between 0.02
and 0.5 M. Presumably, self condensation was minimized due to
complete enolate formation. The high concentration conditions al-
lowed us to make adduct 7 on a large scale (20 g).
O
O
S
HN
S
HN
O
O
b
a
Br
F
Br
F
3
N
OMe
N
O
F
F
7
6
Scheme 2. Reagents and conditions: (a) N,O-dimethyl-hydroxylamine hydrochlo-
ride (5 equiv), n-BuLi (10 equiv), THF, À60 °C, 60%; (b) 4-bromo-3,5-dimethylisox-
azole, n-BuLi or magnesium, iodine, <3%.
O
S
HN
S
O
N
F
O
+
O
N
base
Br
F
Br
N
O
O
F
F
8
7
2
In order to demonstrate the scope of the method, the n-BuLi-
generated lithium enolate of ketone 8 was added to a variety of
ketimines derived from aryl methyl ketones, and the results are
summarized in Table 2. In most cases, these reactions were run
only once as a sufficient quantity was obtained for our medicinal
chemistry analog purposes, and the yields were not optimized.
The additions worked well for ketimines bearing both electron-
withdrawing and electron-donating groups. The reaction condi-
tions are mild enough to tolerate a variety of functional groups
including nitro (entry 11) and nitrile (entries 6–7 and 15).
We also applied the addition of the n-BuLi-generated lithium
enolate of ketone 5 to ketimines derived from heteroaryl methyl
ketones (Table 3). The ketimines bearing all isomeric pyridyl
groups gave 39–51% yields (entries 1, 2, 4–9) with the exception
of 2-fluoro-pyridin-3-yl (19%, entry 3). The ketimine derived from
1-(pyrazin-2-yl)ethanone furnished the desired adduct in moder-
ate yield (31%, entry 10). Ketimines bearing halogen-substituted
thiophenes gave various yields: the highest (43%, entry 13) ob-
tained from 3-chloro-thiophen-2-yl, while the lowest from 2,5-
di-chloro-thiophen-3-yl (9%, entry 12). The N-protected pyrrole
heteroaryl also gave a moderate yield (32%, entry 15). Both oxa-
zole-4-yl (entry 16) and thiazol-4-yl (entry 17) delivered the best
yields in the heteroaryl series (56% and 63%, respectively). How-
ever, the unsubstituted isoxazole (entry 19) failed to provide any
desired adduct due to decomposition under basic conditions.
1,2,3-Thiadiazol-4-yl worked (8%, entry 20), but the yield was
low again due to a stability issue.
Scheme 3.
the more covalent titanium enolate. When ketimine 2 was added
to this enolate solution at À78 °C or À40 °C, no formation of the de-
sired adduct 7 was observed, and only the starting materials were
recovered. At higher temperatures, significant decomposition of
ketone 8 occurred.
We next turned to the lithium and potassium enolates which
were successfully added to aldimines in the studies reported by
Davis et al.4–6 Reaction of ketone 8 with KHMDS at À78 °C followed
by addition of ketimine 2 gave the desired adduct in 27–31% iso-
lated yields along with some recovered ketimine (Table 1). A some-
what lower yield (15–25%) was obtained when LDA was used to
generate the lithium enolate. Nevertheless, the addition was highly
diastereoslective as the other diastereomer was virtually undetect-
able by both LC/MS and 1H NMR analysis of the crude reaction mix-
ture. We speculated that the poor yield may partially result from
the proton transfer from the diisopropylamine or bis(trimethyl-
silyl)amine (HMDS) (released from the enolate formation) to the
metal enolate. This problem can be obviated by treatment of an
LDA-generated enolate with another equiv of n-BuLi to remove
the problematic NH proton.7 This methodology has been used to
improve the efficiency of enolate reactions such as alkylation8
and phenylselenation.9 Another approach is to generate amine-free
metal enolates. In many reactions with electrophiles, use of amine-
free enolates gives better results.7 With this in mind, we sought to