Angewandte
Chemie
in the reactions with aromatic or aliphatic aldehydes under
similar conditions,[10] we assumed that the origin of this
decomposition would be the relatively low pKa of the in situ
generated secondary propargylic alkoxide,[13] which might
prevent the facile proton abstraction from the aminophos-
phonium cation. Rather, the intermediate would attack the
cationic phosphorus center, thus triggering the collapse of 1a.
To gain information about this presumably unique behavior
of the b-nitro secondary propargylic alkoxide, we prepared
the Henry adduct of 2a and 2-nitropropane, 4-methyl-4-nitro-
1-phenylpent-1-yn-3-ol, and treated it with the iminophos-
phorane preformed from 1c·Cl and KOtBu in tetrahydro-
furan. 31P NMR spectroscopic analysis of the resulting
mixture at À988C revealed that the original sharp signal of
iminophosphorane at d = 47.2 ppm had disappeared and a
new broad signal was observed upfield (d = 37.7 ppm),
corresponding to the tetraaminophosphonium cation,[10,12]
which confirmed the formation of the corresponding amino-
phosphonium alkoxide by an acid–base reaction.[14] This
observation is quite different to the result obtained with the
benzaldehyde-derived Henry adduct, 2-methyl-2-nitro-1-phe-
nylpropan-1-ol,[15] and supports our assumption regarding the
influence of pKa of the product 4. Interestingly, our attempts
to tackle this problem found that the catalyst decomposition
pathway could be suppressed by adding N,N-dimethylform-
amide as a co-solvent and that the conversion of the starting
ynal 2a improved gradually as the concentration of N,N-
dimethylformamide was increased, at the slight expense of
stereoselectivity (Table 1, entries 2 and 3). This trend is
probably because the overall polarity of the solvent system
has a beneficial effect on the stabilization of the amino-
phosphonium alkoxide intermediate.[16] These optimized
condition allowed us to investigate the effect of the aromatic
substituents of 1 (Ar) on the diastereo- and enantioselectiv-
ities (Table 1, entries 4–7), and steric bulk, rather than the
electronic effects of a para substituent on the aromatic
groups, seemed to be important. Eventually, the use of
1d·Cl, which contains para-tolyl groups, afforded quantitative
formation of 4a with an anti/syn ratio of greater than 20:1; the
enantiomeric excess of the anti isomer was determined to be
98% (Table 1, entry 6).
Further experiments were then conducted to probe the
substrate scope. The representative results listed in Table 2
show not only the generality of this reaction but also the
intriguing relationship between the structures of the reactants
and the concentration of N,N-dimethylformamide required
for efficient catalysis. In the reactions of aromatic ynals, the
electronic properties of the terminal aromatic substituents
affected the estimation of the suitable amount of N,N-
dimethylformamide, which again agrees with our conjecture
that the pKa of the in situ generated secondary propargylic
alkoxide is related to the decomposition of 1. For instance, the
amount of N,N-dimethylformamide can be reduced in the
coupling of para-anisyl-substituted ynal 2b with 3a, and
should be increased when electron-withdrawing para-chloro-
phenyl-substituted ynal 2c is employed (Table 2, entry 1
versus 2). In both cases, high levels of diastereo- and
enantioselectivities were maintained. On the other hand,
the addition of 5% (v/v) of N,N-dimethylformamide was
generally sufficient for facilitating a smooth reaction with
aliphatic ynals, and excellent stereoselectivities were uni-
formly obtained irrespective of their steric demands (Table 2,
entries 3–6). In addition to nitroethane (3a), other nitro-
alkanes were used as the nucleophilic component by simply
tuning the N,N-dimethylformamide concentration according
to their polarity (Table 2, entries 7 and 8).
The synthetic utility of this catalytic, highly anti-selective
and enantioselective direct Henry reaction of ynals was
demonstrated by its application to the concise asymmetric
syntheses of several naturally occurring, biologically active
compounds (Scheme 2). The treatment of 2-dodecynal (2h)
and nitroethane (3a) with 5 mol% of ent-1d·Cl/KOtBu in
tetrahydrofuran/N,N-dimethylformamide (95:5) at À788C for
30 minutes resulted in the predominant formation of anti-b-
nitro propargylic alcohol 4h (anti/syn = > 20:1) in 94% yield
and 98% ee. Subsequent simultaneous reduction of the triple
Table 2: Substrate profile.[a]
Entry
R (2)
R’ (3)
DMF
[v/v %]
Yield
[%][b]
anti/syn[c]
ee
4
[%][d]
1
2
3
4
5
6
7
8
p-MeO-C6H4 (2b)
p-Cl-C6H4 (2c)
Me(CH2)5 (2d)
cyclohexyl (2e)
PhCH2O(CH2)2 (2 f)
tBuMe2SiO(CH2)2 (2g)
Ph (2a)
Me (3a)
Me (3a)
Me (3a)
Me (3a)
Me (3a)
Me (3a)
H (3b)
5
15
5
5
5
5
5
15
>99
94
>99
89
>99
98
92
>20:1
16:1
>20:1
>20:1
14:1
>20:1
–
10:1
99
95
98
98
98
>99
88
93
4b
4c
4d
4e
4 f
4g
4i
Ph (2a)
Et (3c)
92
4j
[a] The reaction was performed with 0.20 mmol of 2 and 2.0 mmol of 3 in the presence of 1d·Cl/KOtBu (5 mol%) in THF/DMF (2.0 mL) at À788C
under an argon atmosphere. [b] Yield of isolated product. [c] Diastereomeric ratio (d.r.) was determined by 1H NMR (400 MHz) spectroscopic analysis
of an aliquot of the crude reaction mixture. [d] Enantiomeric excess was determined by HPLC on a chiral stationary phase. Absolute and relative
configurations of 4b were determined from the establishment of the total synthesis (Scheme 2) and others were assigned by comparison to 4b and
4h.
Angew. Chem. Int. Ed. 2010, 49, 7562 –7565
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7563