Organic Letters
Letter
Attention then turned to potentially troublesome enolizable
aliphatic aldehydes. The reaction using octanal proceeded
smoothly, giving ethyl decanoate (5l) in good yield. Similarly,
ethyl dodecanoate (5m) was prepared from decanal, and
branched aldehyde 2-phenylpropanal gave 5n, both in very
good yield. The piperidinyl scaffold, which represents the third
Scheme 3. Chemoselectivity in the Tandem Wittig
Hydrogenation with Ylides 2 and 7−10
1
0
most abundant ring in small molecule drugs,
was
incorporated through the reaction of N-Boc-piperidine-4-
carboxaldehyde, to give 5o. To our delight, glycine derived
N-Boc-glycinal was also well-tolerated providing protected
GABA analogue 5p. Enals were also shown as viable substrates,
with concomitant hydrogenation of the internal olefin; trans-
cinnamaldehyde underwent tandem Wittig hydrogenation to
give 5q in 89% yield. We were also interested to test the
enantiointegrity of α-chiral aldehydes in the tandem reaction.
An ideal substrate identified for this test was N-Boc-isoleucinal,
as α-amino aldehydes are known to undergo facile
1
4
racemization. On subjecting freshly prepared N-Boc-
isoleucinal to the tandem Wittig hydrogenation with ylide 2,
the desired N-Boc-γ-amino ester 5r was isolated in 80% yield
as a single diastereomer, with no epimerization. Garners
1
5
aldehyde, a highly versatile chiral synthon was also
demonstrated to undergo the Wittig hydrogenation reaction,
16
providing 5s in 81% yield.
Having demonstrated the scope of the tandem Wittig
hydrogenation on a broad set of aldehydes with ylide 2, we
next tested other stabilized ylides (Scheme 3A). Under the
optimized conditions for ylide 2, β-keto ester 6a was accessed
from the corresponding ylide 7 in 92% yield. For the less
reactive methyl ketone ylide 8, gentle heating and excess ylide
was necessary. Under these slightly modified conditions with 8,
benzaldehyde and p-anisaldehyde gave 6b and 6c, respectively,
in exceptional yields. Nitrile ylide 9 was also successfully used
in the tandem reaction, providing 6d and 6e from p-
anisaldehyde and trans-cinnamaldehyde, respectively. Interest-
ingly, substrates 6a-e contain reducible functional groups (i.e.,
ketone, nitrile); however, no side-products arising from
competitive reduction pathways were observed.
a
b
c
Reaction temp = 45 °C. From trans-cinnamaldehyde. From p-
nitrobenzaldehyde.
Intrigued by the chemoselectivity displayed, we decided to
investigate substrates with potentially reducible/reductively
labile moieties (Scheme 3B). Returning to ethyl ester ylide 2,
the reaction of p-cyanobenzaldehyde gave 6f in 70% yield.
Again, no nitrile reduction was observed. Indeed, no
reduction of the acetyl group was observed on using p-
acetylbenzaldehyde, and 6g was isolated in excellent yield.
Under the reaction conditions, nitro-reduction is not sup-
pressed, however, and p-nitrobenzaldehyde gave aniline 6h in
1
9
hydrogenolysis. Thus, we chose the formation of 6j as the
model reaction (Table 2). First, we compared the distribution
of products obtained from a one-pot, tandem Wittig
hydrogenation, versus a two-step procedure, which involved
isolation and purification of the intermediate enoate 11 prior
to hydrogenation. After 17 h, the one-pot, tandem reaction of
p-benzyloxybenzaldehyde with ylide 2 gave only 6j (entry 1).
1
7
8
6
9% yield. o-Fluorobenzaldehyde was well-tolerated, providing
i in good yield with no apparent hydrodefluorination. Some
18
other halogenated aldehydes were not suitable substrates for
this procedure. The lability of the benzyl protecting group was
next tested using p-benzyloxybenzaldehyde (Scheme 3C).
Surprisingly, no hydrogenolysis of the benzyl group was
observed and 6j was isolated in 86% yield. We were able to
further demonstrate the stability of the benzyl group in the
tandem Wittig hydrogenation using benzyl ester ylide 10,
providing 6k and 6l in good yields.
We then decided that the somewhat unpredicted chemo-
selectivity associated with a number of substrates in the
tandem Wittig hydrogenation warranted some investigation. In
particular, we were surprised that substrates 6j−l did not
undergo debenzylation, as the reaction conditions (Pd/C,
protic solvent) would in fact be considered ideal for benzyl
a
b
entry
procedure
additive
11:6j:5d
c
1
2
3
4
one-pot
0:100:0
0:6:94
28:72:0
0:77:23
two-pot
two-pot
two-pot
2 (1.2 equiv)
Ph PO (1.2 equiv)
3
a
b
1
c
0
.25 mmol scale. Ratio determined by H NMR. Tandem Wittig
hydrogenation of p-benzyloxybenzaldehyde with 1.2 equiv of ylide 2
and 20 mol % DMA.
C
Org. Lett. XXXX, XXX, XXX−XXX