but also the aminal substrate.6 We next examined the influence
of several typical bases on the reaction of 1a because the
recovery of benzaldehyde was considered to be attributed to
the interaction of 1a with zinc chloride or zinc bromide
generated in situ. As expected, the yield of 3aa increased
dramatically when diisopropylamine base was used (entry 3).
For this reaction, the amounts of reagents are important: the
use of 1.5 or 2.0 mol equiv amounts of reagents (2a, Zn, TMSCl,
and i-Pr2NH) against 1a resulted in the moderate yield of 3aa
(entries 4 and 5).7 Other typical bases were also effective, except
pyridine, giving 3aa in moderate to good yields (entries 6-10).
We also found that zinc was more effective than other metals.
Surprisingly, no allylation took place with manganese, tin, and
samarium, except indium (entries 11-14). Furthermore, this
allylation was affected by the substituent bearing silyl chloride.
As expected, both sterically hindered TESCl and TBSCl
impeded the reaction to give 3aa in low to moderate yields
(entries 15 and 16). The use of tetrachlorosilane, which may
be expected to work as a strong Lewis acid, led to 3aa in good
yield (entry 17). Moreover, the formation of 1a is also achieved
in THF and Et2O, in which a Barbier-type reaction took place,
giving 3aa in 80% and 72% yields, respectively (entries 18-22).
The reaction system with TMSCl and diisopropylamine was
useful for other aminal derivatives (Table 2). Similar treatment
of aminals 1 gave the corresponding homoallylamines 3 in
moderate to good yields (entries 1-5). Although the formation
of homoallylamine took place slightly in the case of 1g derived
from morpholine, this problem was circumvented by the
treatment of 1g at 50 °C for 24 h (entry 6). In dimethyl
derivative 1h, the product was purified by Kugelrohr distillation
because dimethyl derivatives 3ha decomposed during purifica-
tion on silica gel columns (entry 7). The allylation of aliphatic
aminals (1i and 1j)8 proceeded efficiently in THF rather than
benzene because of their poor solubility in benzene (in benzene:
3ia 56%, 3ja 30%, respectively, entries 8 and 9).
TABLE 2. Reaction of 1 with Allylzinc Bromide in the Presence
of TMSCl and i-Pr2NHa
This reaction procedure is applicable to several highly reactive
alkyl bromides. The results are presented in Table 3. Treatment
of 1a with 2b or 2c at room temperature for 6 h gave the
corresponding homoallylamines (3ab and 3ac) in moderate to
good yields (entries 1 and 2).9 The crotyl derivatives (2d and
2e) were also used in this reaction, which displayed no marked
stereoselectivity (entries 3 and 4).9 The reaction of 1a took place
in the case of benzyl bromide (2f) to give 3af in 46% yield
(entry 5). The use of R-bromoacetate (2g) and R-bromonitrile
(2h), a Reformatsky-type reaction, does not present a problem,
giving the corresponding ꢀ-amino ester (3ag) and ꢀ-amino nitrile
(3ah) in 68% and 52% yields, respectively (entries 6 and 7).
a Reaction conditions: 1 (2.0 mmol), 2a (6.0 mmol), Zn (6.0 mmol),
TMSCl (6.0 mmol), i-Pr2NH (12.0 mmol), benzene (10 mL), rt, 6 h,
under N2. b Isolated yield. c The reaction was carried out at 50 °C for
24 h. d Product (3ha) was purified by Kugelrohr distillation. e THF was
used instead of benzene.
Recently, ꢀ-amino acids have been shown to have biologically
important properties and have been used as intermediates in
natural product syntheses.10 We finally achieved direct access
to ꢀ-amino carbonyl derivatives from aminals.
Our direct access to ꢀ-amino ester from aminal was applied
to the synthesis of butaverine (3ai), which is an antispasmodic
agent bearing a ꢀ-amino ester.11 For example, the preparation
of 1a in the presence of alumina and subsequent reaction of 1a
with n-propyl R-bromoacetate under these reaction conditions
afforded 3ai in 66% yield in two steps from benzaldehyde
(Scheme 1). Reportedly, butaverine (3ai) had been synthesized
(4) References for aminals and related compounds using Barbier- and
Reformatsky-type reaction, see: (a) Katritzky, A. R.; Manju, K.; Singh, S. K.;
Meher, N. K. Tetrahedron 2005, 61, 2555–2581. (b) Tan, C. Y. K.; Wainman,
D.; Weaver, D. F. Bioorg. Med. Chem. 2003, 11, 113–122. (c) Wang, X.; Li, J.;
Zhang, Y. Synth. Commun. 2003, 33, 3575–3582. (d) Katritzky, A. R.; Shobana,
N.; Harris, P. A. Tetrahedron Lett. 1991, 32, 4247–4248.
(5) For recently selected examples using Barbier- and Reformatsky-type
reaction, see: (a) Petrini, M.; Profeta, R.; Righi, P. J. Org. Chem. 2002, 67,
4530–4535. (b) Choucair, B.; Leon, H.; Mire, M. A.; Lebreton, C.; Mosset, P.
Org. Lett. 2000, 2, 1851–1853. (c) Saidi, R. M.; Khalaji, R. H.; Ipaktschi, J.
J. Chem. Soc., Perkin Trans. 1 1997, 1, 1983–1986.
(6) Picotin, G.; Miginiac, P. J. Org. Chem. 1987, 52, 4796–4798.
(7) The silylated amines (N-trimethylsilylpiperidine and N-trimethylsilyldi-
isopropylamine) and allylated piperidine (3-piperidino-1-propene) were observed
by GC-MS, which were removed by hydrolysis and evaporation. In contrast,
allylated diisopropylamine was not observed.
(10) Juaristi, E. EnantioselectiVe Synthesis of ꢀ-amino Acids, Wiley-VCH:
New York, 1997.
(11) (a) Cerbai, G.; Di Paco, G. F. Boll. Chim. Farm. 1966, 105, 45–53. (b)
Pacheco, H.; Dreux, M.; Beauvillain, A. Bull. Soc. Chim. Fr. 1962, 1379–1387.
(c) Pollard, C. B.; Mattson, G. C. J. Am. Chem. Soc. 1956, 78, 4089–4090.
(8) Aminals (1i and 1j) were prepared according to a modified procedure of
a known method, see: Katritzky, A. R.; Drewniak, M. J. Chem. Soc., Perkin
Trans. 1 1988, 2339–2344.
(9) No regioisomer was detected from 1H NMR of crude product.
J. Org. Chem. Vol. 73, No. 22, 2008 9189