H. Sato, K. Hiroi / Tetrahedron Letters 47 (2006) 5793–5796
Table 2. Thermal rearrangements of iminocyclopropenes 3 into pyrroles (4 and 5)
5795
Entry
R1
R2
Products
Ratio of 4:5a
Yields (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Ph
Ph
Ph
Ph
Ph
Ph
p-FC6H4
p-FC6H4
p-FC6H4
p-FC6H4
p-FC6H4
p-FC6H4
p-FC6H4
n-C6H13
n-C6H13
o-MeC6H4
p-MeC6H4
p-EtOC6H4
p-FC6H4
PhCH2
PhCH2CH2
n-C4H9
i-Pr
4a/5a
4b/5b
4c/5c
4d/5d
4e/5e
4f/5f
90:10
92:8
94:6
95:5
91:9
25:75
93:7
91:9
93:7
97:3
75
53
72
46
67
36
56
33
44
43
43
53
24
80
64
21
18
Ph
4g/5g
4h/5h
4i/5i
p-EtOC6H4
p-FC6H4
PhCH2
PhCH2CH2
n-C4H9
i-Pr
PhCH2
i-Pr
PhCH2
PhCH2
4j/5j
4k/5k
4l/5l
94:6
91:9
4m/5m
4n/5n
4o/5o
4p/5p
4q/5q
10:90
58:42
12:88
>99:1
95:5
a The ratio was determined by 1H NMR analysis.
In the case of 3n (n-hexyl substituent), the reaction via 6
would slightly decrease due to steric reason by the n-hex-
yl group to give pyrroles 4n and 5n with a ratio of 58:42,
respectively.
The structure of the products 4 and 5 was determined
by the NMR spectral analysis. The a-methylene of
benzyl group, methyl ester at C3 and methyl at C2
of 5c appear at 5.07, 3.67 and 2.45, respectively.8
The hydrogens at C5 of 5g and C4 of 4g appear at
6.68 and 6.74, respectively.9 The regiochemistry of 4g
was determined by NOE observed between a hydrogen
at C4 and a methoxy group, and also HMBC between
the hydrogen at C4 and the ester carbonyl carbon. The
regiochemistry of the product 4m and 5m was deter-
mined by NMR spectral analysis; the hydrogens at
C5 of 5m and C4 of 4m appear at 6.60 and 6.43,
respectively.9
Dramatic solvent effects were observed in the above
reaction (Table 1). We examined solvent effects, by
employing, instead of toluene, acetonitrile, chloroform,
dichloromethane, dimethyl sulfoxide (DMSO), N,N-
dimethylformamide (DMF), benzene, xylene, N,N-
dimethylacetamide (DMA), or HMPA as a solvent.
The results are summarized in Table 1. As listed in Table
1, it should be noted that inversion of the regioselectivity
in the product was observed, depending upon the sol-
vent used. With the use of a slightly or strongly polar
solvent such as acetonitrile, DMSO, or DMF, an
amount of 5c was increased in acetonitrile with a 74:26
ratio of 4c and 5c, and 5c was obtained as a major prod-
uct in DMSO, DMF and DMA with 23:77 to 31:69 ratio
of 4c and 5c, respectively, even though the yield was not
so high (54–23%). In another case using 3e in DMSO,
the inversion of regiochemistry in products was
observed with the same ratio (4e:5e = 8:92, 30% yield).
The experimental results obtained, including the regio-
selectivity and the solvent effects mentioned above, are
rationalized as follows. If the reaction proceeds via a
concerted mechanism, it will be particularly hard to
rationalize the dramatic solvent effects mentioned
above. Therefore, on the basis of this fact, we can
undoubtedly anticipate the existence of ionic intermedi-
ates such as 6 and 7, which are formed by fission of the
cyclopropene rings, followed by cyclization affording 4
and 5, respectively. In the cyclization, if the steric factors
are seriously important, the thermal reaction will
provide 5 as a major product. Otherwise, if the chemical
stability of the intermediary sp2 carbocation is a crucial
factor for cyclization, 4 will be provided as a major
product. Based on the results described earlier, it should
be concluded that the most important factor in this reac-
tion will be the stability of the sp2 carbocation produced.
So, with the use of DMSO and DMF as a solvent, steric
interference will be increased by solvation of the carbo-
cation coordinated with the anionic solvents to disturb
the reaction at the stage 6, and, therefore, the formation
of 5 will be preferred with the solvent. However, the
use of HMPA as a solvent was not effective for presen-
tation of high regioselectivity, due to decrease of the
coordination ability to alkenyl carbocations presumably
by the steric reason. The steric effects by N-substituent
(i-propyl) in the imino groups will be rationalized along
this line (Scheme 2).
Use of other primary amines (2a–g) (5.0 equiv), instead
of benzylamine, in the above reaction was studied under
the same reaction conditions (in toluene, 80 °C, and
12 h), and the results are listed in Table 2. It shows that
moderate yields of 4a–q and 5a–q were obtained with
high regioselectivity (75:25 to >99:1). The steric bulki-
ness of N-substituents at the imino groups seriously af-
fected the cyclization stage. We have studied the effects
of N-substituents in 3, changing the substituents such
as benzyl, phenethyl, p-methoxyphenyl, p-fluorophenyl,
n-butyl, i-propyl and phenyl groups. As we expected,
no big difference was observed in the series of aromatic
substituents and alkyl substituents. However, it should
be noted that the regiochemistry of the products was in-
versed in the case of i-propyl group as shown in Table 2.
An i-propyl group significantly affected the regiochem-
istry on cyclization, with the preferential formations of
5f over 4f (75:25) and 5m over 4m (90:10) (Table 2).