aromatic aldehydes selectivity is moderate to low, as can be
seen for reactions of nitriles 1-4 with benzaldehyde 5a
(Table 1, aldols 6a-9a, column 4). In an attempt to improve
The relative stereochemistry of the aldols was determined
according to our previously described vicinal (3J2-3) coupling
constant method,7 or in the case of anti-7a by X-ray
crystallography. In the cases of syn-9b,c, additional confir-
mation of the relative stereochemistry was obtained by
conversion of the aldols to the cyclic carbamate derivatives
cis-10b,c (Scheme 1).7,9 1H NMR spectroscopy confirmed
Table 1. Effect of HMPA on Aldol Reaction
Diastereoselectivity of Arylacetonitrilesa
Scheme 1a
anti:syn (% yield)
0 equiv
6 equiv
aldol
Ar
Ph
R
of HMPA
of HMPA
6a
7a
8a
9a
9b
9c
9d
9e
9f
Ph
Ph
Ph
Ph
85:15 (70)b,c
88:12 (56)
70:30 (89)b,c
72:28 (75)c
58:42 (39)c
69:31 (29)c
63:37 (55)c
57:43 (44)c
50:50 (46)c
38:62 (39)c
>98:2 (87)c
88:12 (81)c
42:58 (69)
90:10 (68)
36:64 (82)
8:92 (76)
8:92 (80)d
5:95 (73)
7:93 (81)
4:96 (72)
8:92 (73)
4:96 (68)
>98:2 (64)
71:29 (83)
mesityl
2-Naph
1-Naph
1-Naph
1-Naph
1-Naph
1-Naph
1-Naph
1-Naph
1-Naph
1-Naph
1-Naph
2-Naph
a Reagents: (a)1:1 LiAlH4/AlCl3, Et2O, rt, 6 h; (b) (Cl3CO)2CO,
Et3N, CH2Cl2, -78 to 0 °C.
2-Me-C6H4
4-Me-C6H4
4-Cl-C6H4
4-MeO-C6H4
t-Bu
9g
9h
9i
the cis relationship of the ring substituents in each case. The
preferred conformer of the cis-carbamates 10 depends on
the relative steric demand of the C-5 and C-6 substituents.9
If the C-5 and C-6 substituents have equal steric demand
(as for cis-10b), then the C-5 axial conformer dominates, to
minimize 1,3-diaxial interactions.
c-C6H11
a Reactions were performed in THF at -78 °C at 0.025 M; 30 min after
addition of the aldehyde, reactions were quenched at -78 °C by the addition
of NH4Cl (aq). For reactions with HMPA, aldehydes were added 30 min
after addition of HMPA (6 equiv) to the lithiated nitrile. The yields reported
are isolated yields after chromatography, unless otherwise noted. b Data from
ref 7. c NMR yield. d The yield at 30 min was quite low; this reaction was
quenched 4 h after addition of aldehyde.
The syn-selectivity observed for arylacetonitriles 1, 3, and
4 with aromatic aldehydes in the presence of HMPA is quite
unusual. Previously, syn-selectivity with benzaldehyde has
been observed only for an aliphatic nitrile, valeronitrile.9 As
a first step toward determining the role of HMPA in reversing
the selectivity of these aldol reactions, the extent of aldolate
equilibration under the two reaction conditions was exam-
ined. We have previously used this technique to establish
that the aldol reaction of 1 and cyclohexanecarboxaldehyde
5i is essentially irreversible over 30 min at -78 °C.10 Thus,
diastereomerically enriched samples of aldols 7a, 8a, and
9c were resubjected to 1 equiv of LDA at -78 °C for 30
min, in the presence or absence of HMPA, followed by the
normal quench (Table 2). In the absence of HMPA, diaste-
reomerically enriched samples of these aldols show little
(e5%) change, suggesting that aldol reaction under these
conditions is predominantly kinetically controlled.
the selectivity, a number of additives were examined, and
HMPA was found to cause a reversal of diastereoselectivity
in three of the four cases (Table 1, aldols 6a-9a, cf. columns
4 and 5). Most striking were the reactions with 1-naphthyl-
acetonitrile 4. In pure THF, diastereoselectivities with
aromatic aldehydes 5a-g were very low. However, in the
presence of 6 equiv of HMPA, anti:syn selectivities ranging
from 8:92 to 4:96 were obtained (Table 1, aldols 9a-g,
column 5).8 The addition of HMPA also led to significant
increases in the yield of the aldol products. Experiments to
determine how much HMPA is needed for high syn-
selectivity were then carried out. For aldols 9e,g use of 2 or
6 equiv of HMPA gave almost identical levels of syn-
selectivity; for 9a-d,f use of 2 equiv of HMPA gave slightly
lower levels of syn-selectivity. To explore the generality of
this HMPA-induced change in diastereoselectivity, reactions
with aliphatic aldehydes 5h and 5i were performed. Unfor-
tunately, in the presence of HMPA these reactions remained
anti-selective (Table 1, aldols 9h,i).
However, aldolate equilibration experiments in the pres-
ence of HMPA show significant change in the diastereomer
ratio (15% (7a), 68% (8a), 88% (9c)). In the case of 8a and
9a, the final anti:syn ratios attained closely approximate those
observed in the corresponding HMPA-mediated aldol reac-
tions (cf. Table 1). This important observation indicates that
(8) Thus far HMPA is uniquely effective. In reactions of 4 and 5a, LiCl,
DMSO, TMEDA, and the HMPA substitutes DMPU and dimethylacetamide
failed to induce syn-selectivity. Use of sodium hexamethylsilazide as base,
with or without added HMPA, resulted in poor syn-selectivity.
(9) Carlier, P. R.; Lo, K.-M.; Lo, M. M.-C.; Lo, P. C.-K.; Lo, C. W.-S.
J. Org. Chem. 1997, 62, 6316-6321.
(10) Carlier, P. R.; Lo, K.-M. J. Org. Chem. 1994, 59, 4053-4055.
2444
Org. Lett., Vol. 2, No. 16, 2000