toluene at -40 °C over 20 h occurred cleanly to form the
R)-amino nitrile 3a in 96% yield and with 93:7 R/S
selectivity (86% ee). This reaction is quite general for
N-benzhydrylimines of aromatic aldehydes as shown by the
results which are collected in Table 1. The yields given in
poor (0-25%) enantiomeric purity. The N-(2,3:6,7-diben-
zocycloheptadienyl) analog of 2a afforded amino nitrile of
77% ee which is definitely below that with 2a itself.
Similarly, the placement of para substituents on the benz-
hydryl group of 2a resulted in products of somewhat reduced
(
3
enantiomeric excess (p-Me, OMe, Cl, and CF : 78, 70, 68,
and 59% ee, respectively, for the analogs of 3a).
The catalytic action of the bicyclic guanidine 1 can be
understood simply in terms of the mechanism for the
conversion of 2a to 3a which is outlined in Scheme 2. In
Table 1. Conversion of ArCHdNCH(C6H5)2 (2) to
ArCH(CN)NHCH(C6H5)2 (3) by Reaction with HCN in the Presence of
1
0 mol % of 1
2
a -j
Ar )
T( °C) t (h) % yield
% ee
3a -j
a
a
b
c
d
e
f
g
h
i
Ph
Ph
p-tolyl
3,5-xylyl
o-tolyl
4-t-Bu-Ph
4-TBSO-Ph
4-MeO-Ph
4-F-Ph
4-Cl-Ph
1-naphthyl
-40
-20
-40
-40
-20
-40
-40
-40
-40
-20
-20
20
8
96
99
96
96
88
80
98
99
97
88
90
86
82
80
79
50
85
88
84
86
81
76
(R)-3a
(R)-3a
(R)-3b
(R)-3c
3d
Scheme 2
20
16
12
72
38
28
23
20
12
3e
(R)-3f
(R)-3g
(R)-3h
(R)-3i
(R)-3j
j
5
Table 1 refer to isolated pure R-amino nitrile. The guanidine
catalyst 1 was easily separated from the crude reaction
mixture by extraction with oxalic acid and recovered for
5
reuse. The amino nitriles 3 upon heating at reflux with 6 N
HCl underwent benzhydryl cleavage and CN f COOH
conversion to form cleanly the corresponding (R)-aryl-
the first step of the cycle, HCN hydrogen bonds to the
catalyst 1, generating a guanidinium cyanide complex which
can serve as a hydrogen bond donor to the aldimine 2a
forming the pre-transition-state termolecular assembly shown
(4). Finally, attack by cyanide ion within the ion pair on the
hydrogen-bond-activated aldimine affords the Strecker prod-
uct (R)-3a. It is likely that the last step is rate-limiting since
hydrogen bond making and breaking are normally relatively
fast. It is also relevant that no kinetic H/D isotope effect on
the reaction rate was detected in experiments which com-
pared the velocity of the catalytic Strecker reaction with DCN
and HCN.
2d
glycines, the absolute configurations of which were deter-
mined by measurement of optical rotation. Enantiomeric
ratios were evaluated by HPLC analysis using chiral
5
columns. It is important to note that in the absence of
catalyst 1 there was no reaction between HCN and the
aldimines 2 in toluene at 10 °C or below and that N-methyl-1
is entirely inactive as a catalyst.
The N-benzhydryl subunit of the aldimine substrates 2 is
critical to the realization of good enantioselectivity. N-
Benzyl- or N-(9′-fluorenyl) aldimine analogs of 2 underwent
reaction with HCN in the presence of 10 mol % of 1 in
toluene at -10 to -20 °C to afford Strecker products of
The catalytic cycle which is summarized in Scheme 2
provides insights with regard to the origin of enantioselec-
tivity for the conversion of 2 to 3 in the presence of catalyst
(4) (3R,7R)-3,7-Diphenyl-1,4,6-triazabicyclo[3.3.0]oct-4-ene (1): color-
less solid, mp 159-160 °C; Rf 0.28 (1:10:90 NH4OH/MeOH/CH2Cl2);
1
and HCN. Modeling of the imine 2a hydrogen bonded to
23
[
R] D +23.8 (c 0.58, CHCl3); IR (thin film) 1452, 1492, 1674, 2829, 2926,
027, 3058, 3106, 3156, 3205 cm-1; H NMR (400 MHz, CDCl3) δ 7.37
1
the guanidinium in pre-transition-state assembly 4 indicates
the possibility of a three-dimensional arrangement depicted
in Figure 1 where the cyanide is positioned to attack the re
face of the imine carbon, a trajectory that leads to the
predominating enantiomer (R)-3a. A proximal phenyl group
of the catalyst can undergo π-stacking with one of the
benzhydryl phenyls, while the si face of the imine carbon is
blocked by the other phenyl of the benzhydryl group. At
the same time the aryl π-conjugated to the imine of 2a is
accommodated in a vacant quadrant of the guanidine face
and can experience van der Waals attractions with the
guanidine core and distal phenyl edge of catalyst 1. Rotation
of imine 2a by 180° about the H-N bond to expose the si
face to attack removes the van der Waals attractions and
3
(
d, 4 H, J ) 7.0 Hz), 7.32 (t, 4 H, J ) 7.1 Hz), 7.25 (t, 2 H, J ) 2.5 Hz),
6
.37 (bs, 1 H), 5.17 (t, 2 H, J ) 6.5 Hz), 3.50 (t, 2 H, J ) 7.8 Hz), 3.03
1
3
(
dd, 2 H, J ) 6.3, 7.8 Hz) ppm; C NMR (100 MHz, CDCl3) δ 169.3,
+
1
2
43.0, 128.5, 127.4, 126.4, 68.0, 57.2 ppm; CIMS 278(30) [M + NH4 ],
+
64(100) [M + H ].
(5) To a clear solution (0.2 M) of aldimine (0.18 mmol) and guanidine
1
(0.018 mmol) in toluene under N2 was added liquid HCN (0.36 mmol)
via a precooled 50-µL gastight syringe. Upon consumption of starting
material as indicated by TLC analysis, the reaction mixture was concentrated
in vacuo, acidified by addition of oxalic acid (0.018 mmol) in water (5
mL), and extracted with ether (3 × 15 mL). The combined extracts were
washed with brine, dried over MgSO4, and concentrated to give amino nitrile
3
that could be further purified by silica gel chromatography (1% ethyl
acetate-hexanes). Basification with 1 N NaOH of the oxalic acid wash
and extraction with ethyl acetate and concentration allows recovery of 1
(
80-90% yield). Enantioselectivity was analyzed by chiral HPLC with
Chiralpak AD, Chiralpak AS or Chiralcel OJ columns with 2-propanol-
hexanes as eluent.
158
Org. Lett., Vol. 1, No. 1, 1999