Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Our attempts to alter the selectivity of this reaction began
with protection of the secondary amine of 1 followed by
addition to cyanoalkyne 2. This reaction cleanly gave the
Michael adduct; however, attempts to remove a variety of
protective groups [allyl, benzyl, silyl] in the presence of the
nitrile failed. During these investigations we noticed that the
addition of base affected the distribution of reaction prod-
ucts.7 After optimization, we have found that using 2 equiv
of NaHMDS in 5% MeCN/THF completely reverses the
selectivity of the reaction, giving 4 as the major product in
a 97:3 ratio (Entry 1, Table 1).
involving NaHMDS; when 100% THF is employed as
solvent, the reaction gives many side products. Furthermore,
if less than 2 equiv of NaHMDS is used, the reaction
proceeds well but seems to be nonchemoselective. This
information has led us to hypothesize that the dianion of the
amidine is formed and that the acetonitrile aids in solubility
of the dianion.
Cyclic amidines seem to react more quickly and give
higher yields. Preliminary investigation indicates that the ring
closure step is slower in the case of acyclic amidines. The
reaction is also tolerant of substituents on the phenyl ring of
the cyanoalkyne,3 as shown in Table 2.
Table 1. Reaction of 3-Phenyl-2-propynenitrile with Amidines
Table 2. Reaction of 2-Iminopiperidine with Cyanoalkynes
We have extended this methodology to include 2-imino-
4-phenylpyrrolidine8 (20, entry 2), as well as unsymmetrical
monosubstituted acyclic amidines (entries 3 and 4). All of
the cases were run under the same conditions9 and show a
similar reversal of selectivity in the presence of base.10,11
Addition of MeCN is crucial to the success of the reaction
The reactions recorded in entries 4 and 8, however,
behaved somewhat differently than the other examples. In
the absence of base, the selectivity and yield both degraded
substantially. We believe this is due to the steric congestion
associated with the ring closing step to form compounds 9
and 17. After the Michael addition occurs, the aryl group at
position 6 (pyrimidine numbering) and the substituent on
N-1 must come into close proximity in order to close the
(6) Ratios were measured by HPLC with UV detection at 214 nM. Data
for compound 3‚TFA: 1H NMR (500 MHz, CDCl3) δ 9.42 (br s, 1 H),
7.58-7.50 (m, 3 H), 7.38-7.35 (m, 2 H), 6.81 (s, 1 H), 6.73 (br s, 1 H),
3.84 (t, J ) 5.8 Hz, 2 H), 3.10 (t, J ) 6.6 Hz, 2 H), 2.03-1.94 (m, 4 H)
ppm; 13C NMR (125 MHz, CDCl3) δ 163.5, 163.0, 157.4, 131.5, 130.6,
129.5, 128.3, 106.6, 50.3, 31.5, 21.8, 18.1 ppm; HRMS m/z 226.1353 [(M
+ H)+, calcd for C14H16N3 226.1339]. Data for compound 4‚TFA: 1H NMR
(500 MHz, CDCl3) δ 9.74 (br s, 1 H), 9.18 (br s, 1 H), 8.04 (dd, J ) 7.8,
1.5 Hz, 2 H), 7.56-7.43 (m, 4 H), 4.10 (t, J ) 6.2 Hz, 2 H), 3.13 (t, J )
6.5 Hz, 2 H), 2.18-2.11 (m, 2 H), 2.02-1.98 (m, 2 H) ppm; 13C NMR
(125 MHz, CDCl3) δ 161.2, 158.3, 157.7, 132.2, 129.1, 129.0, 127.5, 101.4,
47.0, 32.3, 21.4, 18.1 ppm; HRMS m/z 226.1351 [(M + H)+, calcd for
C14H16N3 226.1339].
(8) The starting amidine for entry 2 (20) was prepared from 4-phenyl-
2-pyrrolidinone (19)12 as shown:
(7) Similar observations in the case of amidine additions to â-ketoesters
have been made (see ref 1a); however, the conditions reported for selectivity
reversal [EtONa, EtOH, reflux], when applied to our system, gave primarily
Michael addition of ethoxide to the cyanoalkyne.
3390
Org. Lett., Vol. 2, No. 21, 2000