Pyrroles from 1,2-cyclopropanediamines and aldehydes

Article abstract of DOI:10.1002/(SICI)1099-0690(199808)1998:8<1645::AID-EJOC1645>3.0.CO;2-H

Mechanistic investigations by means of proton spectroscopy detected intermediates and uncovered the course of reactions in acetate-buffered [D₄]methanol of primary cyclopropanediamines cis- and trans-2a with benzaldehyde (3a) or 2,2-dimethylpropanal (3b), of secondary cyclopropanediamines cis- and trans-2b with 3b, and of the ringmethylated cyclopropanediamine trans-14a and the aromatic aldehydes 3a and c. This study provided the basis of an expedient synthesis of pyrroles which takes place under exceptionally mild conditions. Irrespective of the configuration, primary (2a·2HBr) and secondary cyclopropanediammonium dibromides 2b and c·2HBr that are devoid of ring substituents react with aromatic aldehydes 3a, e-h, cinnamic aldehyde (3i), and 3b to afford 2-substituted (8a, b) and 1,2-disubstituted pyrroles (8c-i), respectively. The 3-substituted secondary transcyclopropanediammonium dibromides 24·2HBr furnish 1,2,4-trisubstituted pyrroles 25. While the primary 1-methylcyclopropanediammonium dibromide trans-14a·2HBr reacts regioselectively with 3a and c to produce only 2,3-substituted pyrroles 19a, c, the corresponding secondary dibromide trans-14c·2HBr gives rise to the formation of mixtures of 1,2,3- (22) and 1,2,5-trisubstituted pyrroles 23. The key step of pyrrole formation from 1,2-cyclopropanediamines and aldehydes is the ring expansion of intermediate monoiminium ions of type 5 via azomethine ylides (E,Z)-6 to yield dihydropyrrolium ions 7.

Full text of DOI:10.1002/(SICI)1099-0690(199808)1998:8<1645::AID-EJOC1645>3.0.CO;2-H

FULL PAPER
Cyclopropanediamines, 7[᭛]
Pyrroles from 1,2-Cyclopropanediamines and Aldehydes
Wolfgang von der Saal*[°], Robert Reinhardt, Josef Stawitz, and Helmut Quast*
Institut für Organische Chemie der Universität Würzburg,
Am Hubland, D-97074 Würzburg, Germany
Received March 25, 1998
Keywords: Azomethine ylides / Nitrogen heterocycles / Ring expansion / Small ring systems / Synthetic methods
Mechanistic investigations by means of proton spectroscopy
detected intermediates and uncovered the course of
reactions in acetate-buffered [D4]methanol of primary
substituted (8a, b) and 1,2-disubstituted pyrroles (8c–i),
respectively. The 3-substituted secondary trans-
cyclopropanediammonium dibromides 24·2HBr furnish 1,2,4-
cyclopropanediamines cis- and trans-2a with benzaldehyde trisubstituted pyrroles 25. While the primary 1-
(3a) or 2,2-dimethylpropanal (3b), of secondary cyclopro- methylcyclopropanediammonium dibromide trans-14a·2HBr
panediamines cis- and trans-2b with 3b, and of the ring- reacts regioselectively with 3a and c to produce only 2,3-
methylated cyclopropanediamine trans-14a and the aromatic substituted pyrroles 19a, c, the corresponding secondary
aldehydes 3a and c. This study provided the basis of an dibromide trans-14c·2HBr gives rise to the formation of
expedient synthesis of pyrroles which takes place under mixtures of 1,2,3- (22) and 1,2,5-trisubstituted pyrroles 23.
exceptionally mild conditions. Irrespective of the The key step of pyrrole formation from 1,2-cyclo-
configuration,
primary
(2a·2HBr)
and
secondary
propanediamines and aldehydes is the ring expansion of
cyclopropanediammonium dibromides 2b and c·2HBr that
intermediate monoiminium ions of type 5 via azomethine
are devoid of ring substituents react with aromatic aldehydes ylides (E,Z)-6 to yield dihydropyrrolium ions 7.
3a, e–h, cinnamic aldehyde (3i), and 3b to afford 2-
Introduction
cis-4 to dihydrodiazepine 10. An intriguing observation was
that benzaldehyde reacted with the primary cyclopropanedi-
ammonium dibromides cis- or trans-2a·2HBr in acetate-
buffered methanol at room temperature to give diphenyl-
dihydrodiazepine 11a in 47% yield, while, under identical
conditions, the (secondary) N,NЈ-dibenzyl derivative trans-
Various reactions of 1,2-cyclopropanediammonium dibro-
mides in aqueous or methanol buffers proceed via common
intermediates, i.e. azomethine ylides. These reactions com-
prise the cis-trans isomerization of cis-cyclopropanediam-
monium dibromide (cis-2a·2HBr)[1], the autocatalytic and
ketone-catalyzed conversion of 1-methyl-trans-cyclopro-
panediammonium dibromide (trans-14a·2HBr) into 4-
aminobutanone (20)[2], the formation of cis-2,3-substituted
2,3-dihydro-1H-1,4-diazepines (11) from cyclopropanedia-
mines and aromatic aldehydes[3], and the formation of 2-
aryl-1-benzylpyrroles from N,NЈ-dibenzylcyclopropanedia-
mines and aromatic aldehydes[1][4]. We observed, using pro-
ton spectroscopy, that solutions in [D4]methanol of cis- or
trans-2a·2HBr, benzaldehyde (3a, 2 mol), and sodium acet-
ate (2 mol) equilibrated within seconds at room temperature
to afford identical mixtures consisting mainly of the di-
hydrodiazepinium ion 11a·Hϩ and the trans-bisimine trans-
4a[5]. The results were interpreted in terms of the reactions
and equilibria depicted in Scheme 1.
2c·2HBr gave 1-benzyl-2-phenylpyrrole (8d) in 78% yield[4]
.
No dihydrodiazepines were found under the conditions of
pyrrole formation, most probably because the N,NЈ-di-
benzylbisiminium dication, required as precursor for a
Cope rearrangement, did not arise in the nearly neutral me-
dium. Likewise, no trace of a pyrrole was detected under
the conditions of dihydrodiazepine formation. This result is
probably due to a dramatic difference in rates of the two
irreversible steps, of which the Cope rearrangement of (s-
cis)-cis-4a is much faster than the cyclization of azomethine
ylide (E,Z)-6a to dihydropyrrole 7a.
Scheme 1 offered opportunities to test this assumption:
The route, which eventually would lead to dihydrodiazep-
ines, might be redirected to pyrrole formation if it were pos-
sible to reduce the rate of the Cope rearrangement without
affecting the rate of pyrrole formation. This can be con-
ceived in two ways, either by reducing the concentration of
the cis-bisimine cis-4, or by lowering the rate constant of
the Cope rearrangement. Here we report that both stra-
tegies can be used to favor the formation of pyrroles. These
observations not only confirmed Scheme 1, but also un-
covered an expedient method for the preparation of such
pyrroles in good yields under very mild conditions.
There are only two irreversible steps involved, viz. the
intramolecular nucleophilic attack in (E,Z)-6 of the azome-
thine ylide moiety at the imine carbon atom to yield the
dihydropyrrole 7, and the Cope rearrangement of (s-cis)-
[᭛] Part 6: Ref.[2]. The results are part of the dissertations by J.
Stawitz, 1978, W. von der Saal, 1983, and R. Reinhardt, 1985,
University of Würzburg.
[°] Present address: Boehringer Mannheim GmbH, Sandhofer Str.
116, D-68305 Mannheim, Germany.
Eur. J. Org. Chem. 1998, 1645Ϫ1652
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1434Ϫ193X/98/0808Ϫ1645 $ 17.50ϩ.50/0
1645

W. von der Saal, R. Reinhardt, J. Stawitz, H. Quast
FULL PAPER
Scheme 1. Equilibria and irreversible reactions to pyrroles 8 and dihydrodiazepines 11 of 1,2-cyclopropanediammonium dibromides
2·2HBr and aldehydes 3
Mechanistic Results
Three phases could be distinguished. The first ended five
minutes after mixing trans-2a·2HBr, benzaldehyde, and so-
dium acetate in [D4]methanol in the ratio 1:2:2. Dihydrodi-
azepine 11a was the predominating species and the cyclo-
propanediamines had equilibrated to yield a ratio of trans-
2a/cis-2a of 4:1 which did not change anymore. Signals of
pyrrole 8a were not yet observable but those of two tran-
sient species which are most probably the imines trans- and
cis-1a. They are characterized by singlets at δ ϭ 8.6 and 8.7
(CH of benzylidene groups) and unresolved multiplets at
δ ϭ 1.5 and 2.9 from intact cyclopropane rings. After the
end of the second phase, viz. fifteen minutes, these signals
had disappeared and the mixture consisted mainly of un-
changed trans-2a (40%) and dihydrodiazepine 11a (40%).
Minor components were the diastereomerized diamine cis-
2a (10%) and indeed 2-phenylpyrrole (8a, 10%). Bisimines
4a were never detected.
Reactions at Low Concentrations of Benzaldehyde: 2-Phenylpyrrole
From Primary Cyclopropanediamines
Scheme 1 suggested that pyrrole 8a might be formed if
the concentration of cis-bisimine cis-4a is lowered relative
to those of the monoimines 5a. We envisaged that this could
be achieved by reducing the concentration of benzaldehyde.
In previous experiments, we employed two moles of an aro-
matic aldehyde and one mol of a cyclopropanediammonium
dibromide because two moles of the aldehyde are included
in the dihydrodiazepines 11, and two moles are consumed
in the pyrrole forming reactions as well, viz. one for the
pyrroles 8 themselves, and the second one for the Schiff
bases 9, which accompany the pyrroles. Since Schiff bases
equilibrate with aldehydes and amines, pyrrole formation is
conceivable to occur with just one mol of aldehyde, how-
ever, which eventually is completely incorporated in the pyr-
role.
When trans-2a·2HBr was exchanged for its cis isomer
For scrutiny of the dihydrodiazepine/pyrrole competition, cis-2a·2HBr, the resulting mixtures had exactly the same
we monitored the reaction of cis- and trans-2a with benzal- composition. A third experiment started from bisimine
dehyde by proton spectroscopy. This was possible because trans-4a dissolved in [D4]methanol to which hydroxylam-
the proton spectra of cis- and trans-2a[6], trans-4a[3], and ine hydrochloride, sodium acetate, and acetic acid were
11a[3] were known from previous work, while that of 8a was added. The rationale behind this experiment was trapping
recorded from an authentic sample (Table 2). In the spectra of benzaldehyde as oxime and thus producing monoimine
taken from the reaction mixtures, the signals of the cyclo- trans-5a from bisimine trans-4a. Indeed the resulting mix-
propane protons were clearly separated and this allowed the tures closely resembled those of the first two experi-
assignment of structures to some intermediates.
ments.
1646
Eur. J. Org. Chem. 1998, 1645Ϫ1652

Cyclopropanediamines, 7
FULL PAPER
During the slowest phase (twelve hours), the relative with a broadened X2 part. Neither this unknown product
amounts of the cyclopropanediamines decreased further nor trans-1b or cis-2a were present when the amount of 2,2-
and the proton signals of 8a changed slightly: those of the dimethylpropanal had been doubled. In that case mainly
phenyl protons increased, but the signals of both 3-H and bisimine trans-4b, the hemiacetal of the aldehyde and the
5-H were overlaid with doublets (J ϭ 2.7 Hz), while the solvent[10], and a small amount of pyrrole 8b were observed
signal of 4-H remained unchanged. Obviously, a new, isoto- after one day. As expected, the signals of dihydrodiazepine
pomeric pyrrole emerged with 4-H exchanged for deu- 11b never emerged in these experiments.
terium, viz. 4-[D]8a. The final mixture consisted of 11a,
trans-2a, 4-[D]8a, 8a, and cis-2a in the ratio 4:2:2:1:1.
The broadened X2 part of the ABX2 multiplet of the un-
known compound is indicative of an exchange process. Be-
These experiments confirm Scheme 1. The reactions be- cause the 1H-NMR parameters resemble those of cis-4b, the
tween cyclopropanediamines 2a and benzaldehyde afford unknown compound is most probably the cis imine cis-1b
mainly dihydrodiazepine 11a. At the onset, it arises fast, undergoing a degenerate rearrangement via 2,4-diazabicy-
but finally, when there is only a small fraction of the alde- clo[3.1.0]hexane 12b. An additional broadening of the sig-
hyde left in form of the monoimines 1a/5a, then pyrrole nal of the aldimine proton (δ ϭ 7.57) suggests the presence
formation can compete with the route via the Cope re- of some 12b equilibrating rapidly with cis-1b, most probably
arrangement. This observation led to the development of a the exo diastereomer exo-12b. The overall dynamic process
convenient synthesis of 2-arylpyrroles (vide infra).
is a two-site exchange for the methylene protons and the
The slow formation of 4-[D]8a is indicative of the inter- methine protons CHtBu, but a three-site exchange for the
mediate dihydropyrrolium ion 7a, whose low concentration cyclopropane methine protons. This is yet another example
precluded observation by proton spectroscopy. Control of the ring-chain equilibria involving imidazolidines, which
experiments showed that the proton 4-H of 8a is not ex- have been thoroughly investigated by proton spec-
changed under the reaction conditions during a period of troscopy[11] and kinetic measurements[12]
time as long as two weeks. Neither is deuterium incorpor-
ated at the carbon atoms of equilibrating cyclopropanedia-
mines trans- and cis-2a or dihydrodiazepine 11a. Evidently,
H/D exchange occurs exclusively at the α position of the
iminium group of 7a, which eventually becomes C-4 of pyr-
role 8a. Apparently, elimination of ammonia from 7a to
afford 8a is slow enough to allow the H/D exchange under
the conditions employed.
.
Lowering the Rate of the Cope Rearrangement ؊
2-tert-Butylpyrrole From Primary Cyclopropanediamines and
2,2-Dimethylpropanal
A second way to favor the formation of pyrroles is of-
fered by retardation of the Cope rearrangement with the
help of steric strain. The Cope rearrangement of cis bisaryl-
2-tert-Butyl-1-methylpyrrole From Secondary
Cyclopropanediamines and 2,2-Dimethylpropanal
ideneamines cis-4 is so fast that their isolation from the re-
Treatment of the N,NЈ-dimethylcyclopropanediammon-
action of cis-cyclopropanediamines and aromatic aldehydes ium dibromide trans-2b·2HBr with 2,2-dimethylpropanal
is impossible. In contrast, the encumbered aldehyde 2,2-di- yielded in one hour pyrrole 8c, small amounts of trans- and
methylpropanal (3b) affords cis-bisimine cis-4b and its Cope cis-2b, and two new products in the ratio 7:3. The proton
rearrangement to 10b occurs only at temperatures above spectrum of each compound exhibits an ABX2 multiplet
100°C[7]. Under these conditions, trans-4b does not re- characteristic of the cyclopropane protons of a free cis-1,2-
arrange. The dramatic decrease in the rate constant may be cyclopropanediamine[6] and a singlet from two equivalent
anticipated because formation of the new bond C2ϪC3 of N-methyl groups. Furthermore, the proton spectrum of the
10b requires an eclipsed conformation of the two tert-butyl major product shows the singlet of a methine proton (δ ϭ
groups, which is strongly destabilized[8]. As before, we 2.70), which is apparently obscured by other signals in the
1
monitored the conversions by proton spectroscopy spectrum of the minor product. On the basis of the H-
(Table 2).
NMR evidence we assigned the bicyclic structures exo- and
Solutions of trans-2a·2HBr and 2,2-dimethylpropanal in endo-12c to the new compounds. After one day, the main
acetate-buffered [D4]methanol yielded immediately the rap- products were pyrrole 8c and imine 9c. Two moles of the
idly equilibrating pair of tautomers trans-1b and trans-5b aldehyde gave rise to essentially the same results. Treatment
(ABXY spectrum of cyclopropane ring protons), of which of cis-2b·2HBr with 2,2-dimethylpropanal produced im-
the former is greatly favored[9], and bisimine trans-4b. After mediately large amounts of only exo-12c, which exhibited
one day, the fraction of trans-4b had increased and some very broad signals. After one hour, signals of endo-12c em-
cis-2a had formed as well as two new products. One was erged at higher field but remained much less intense than
pyrrole 8b, the other an unknown compound characterized those of exo-12c (10:1). In addition, trans diamine trans-2b,
by an ABX2-type spectrum of cyclopropane ring protons pyrrole 8c, and imine 9c were detected.
Eur. J. Org. Chem. 1998, 1645Ϫ1652
1647

W. von der Saal, R. Reinhardt, J. Stawitz, H. Quast
FULL PAPER
In contrast to the reactions of the primary cyclopropane- closure is much faster than the hypothetical recyclization
diamines cis- and trans-2a with 2,2-dimethylpropanal, the 16a Ǟ cis-13a, which would result in a transǞcis isomeri-
analogous conversions of the secondary cyclopropanediam- zation.
ines cis- and trans-2b did not allow detection of monoim-
The lack of pyrrole 21a in this multistep scenario be-
ines, e.g. cis- and trans-5c, as intermediates en route to the comes plausible if one considers the relative stabilities of
observed bicyclic products 12c. According to the face of the the azomethine ylides 16a and 18a. Strain introduced by
aldimine carbon atom of cis-5c, which is attacked by the the methyl group into the planar azomethine ylide moiety
methylamino group, either exo- or endo-12c is produced. and the higher energy of the aldiminium group render 18a
The major diastereomer is assigned the exo configuration less stable than 16a and hence disfavor this reaction chan-
because the endo configuration should be disfavored by nel. Similar results were observed in the autocatalytic and
steric interaction between the tert-butyl group and the endo ketone-catalyzed reactions of trans-14a·2HBr in aqueous
proton 6-H.
buffers which eventually led to 4-aminobutanone (20). Also
in this study, the final result was dominated by the relative
stabilities of the two ketimines, derived from the ketone and
trans-14a, and their acyclic isomers, namely azomethine yl-
ides such as 17, which was hydrolyzed to afford 20[2]. In the
present experiments, performed in [D4]methanol as solvent,
this hypothetical hydrolysis product of 16a was not ob-
served, however, although the presence of small amounts of
deuterium oxide could not be excluded.
The implied dramatic difference in the reactivities be-
tween the deceptively similar azomethine ylides 16a and 17
became obvious, when trans-14a·2HBr was treated with the
water-soluble aromatic aldehyde 3c in a buffered aqueous
medium which should facilitate hydrolysis of 16c. When the
concentrations of the reactants and of sodium acetate were
kept similar to those used in [D4]methanol solution, only
pyrrole 19c formed immediately but no trace of 4-aminobu-
tanone (20) or pyrrole 21c. In a competition experiment,
the concentration of aldehyde 3c was lowered to a fifth of
the concentration of trans-14a·2HBr. The resulting solution
consisted of pyrrole 19c and the hydrolysis product 20 (1:4).
Obviously, the aldehyde 3c and the corresponding fraction
of the diamine trans-14a·2HBr reacted quickly to give pyr-
role 19c, and the remaining diamine was converted more
2-Aryl-3-methylpyrroles From the Primary
1-Methylcyclopropanediamine trans-14a and Aromatic Aldehydes
Formation of 2,4-diazabicyclo[3.1.0]hexanes 12 in the slowly into 20.
experiments described above was not unexpected since bi-
cyclic compounds often result as byproducts in reactions of
Preparative Results
cis-1,2-substituted cyclopropanes. They may even predomi-
Treatment of N,NЈ-dibenzylcyclopropanediammonium
nate if the cyclopropanes bear additional substituents at C- dibromides cis- and trans-2c·2HBr with aromatic aldehydes
1 and/or C-2[13]. For example, we observed previously that 3a, eϪh or cinnamic aldehyde 3i in acetate-buffered meth-
the carboxy groups of 1-methyl- and 1,2-dimethyl-cis-cyclo- anol or ethanol at ambient temperature afforded the 2-sub-
propanedicarboxylic acid (or derivatives thereof) interact stituted N-benzylpyrroles 8dϪ8i[4]. In the reactions of the
more readily than those of cis-cyclopropanedicarboxylic primary cyclopropanediamines cis- and trans-2a with 2,2-
acid itself[6]. It appeared, therefore, interesting to study the dimethylpropanal, pyrrole 8b was only a minor product, but
reaction of 1-methylcyclopropanediamine trans-14a with N,NЈ-dimethylcyclopropanediamine trans-2b gave mainly N-
benzaldehyde.
methylpyrrole 8c which was isolated in 32% yield. Unlike
We expected to observe formation of a 2,4-diazabicy- the primary 1-methylcyclopropanediammonium dibromide
clo[3.1.0]hexane of type 12, which is a derivative of the hith- trans-14a·2HBr, which gave a single pyrrole with benzal-
erto unknown cyclopropanediamine cis-14a. Surprisingly, dehyde (3a Ǟ 19a) and with the ortho-sulfonic acid derived
3-methyl-2-phenylpyrrole (19a) had formed quantitatively thereof (3c Ǟ 19c, vide supra), the corresponding N,NЈ-
already after five minutes. No trace of a dihydrodiazepine dibenzyl derivative trans-14c·2HBr reacted with aromatic al-
nor of the isomeric pyrrole, viz. 5-methyl-2-phenylpyrrole dehydes (3e, h) to furnish a mixture of isomeric pyrroles,
(21a), was detected. This result may be rationalized in terms viz. 22e, h (major isomers) and 23e, h (minor isomers)[4]
.
of an initial rapid equilibration of trans-14a and benzal- This demonstrates that combinations of subtle effects tune
dehyde with the imines trans-13a and trans-15a[14]. Only the the delicate balance between the various reaction paths. The
former eventually yields a pyrrole (19a) via cyclization of 3-substituted cyclopropanediamines 24 gave exclusively the
an intermediate azomethine ylide (16a). Obviously, this ring corresponding 1,2,4-substituted pyrroles 25.
1648
Eur. J. Org. Chem. 1998, 1645Ϫ1652

Cyclopropanediamines, 7
FULL PAPER
out the reaction period: After a dilute solution of benzal-
dehyde in methanol had been added very slowly to a solu-
tion of trans-2a·2HBr in acetate-buffered methanol, pyrrole
8a could be isolated in 64% yield.
Discussion
The present study revealed several parameters which gov-
ern the ultimate fate of 1,2-cyclopropanediammonium di-
bromides in the reactions with aldehydes in methanol buf-
fers. (i) The ratio of monoimines 5 and cis bisimines cis-4
determines the observed ratio of pyrrole 8 vs. dihydrodia-
zepine 11. The observation that both, pyrrole 8a and di-
hydrodiazepine 11a, arise simultaneously from trans-cyclo-
propanediamine (trans-2a) under certain conditions indi-
cated that both are formed via a common intermediate, i.e.
an azomethine ylide (6). It also suggested the opportunity
to direct the multistep sequence either toward formation of
2-arylpyrroles 8 or toward formation of dihydrodiazepines
11, simply by keeping the concentration of the aromatic al-
dehyde low or high, respectively. (ii) The second factor is
the rate of the Cope rearrangement of the cis bisimines cis-
4 Ǟ 10. The encumbered bisimine cis-4b rearranges to 10b
only at high temperature. Accordingly, in acetate-buffered
methanol at room temperature, trans-2a reacted with 2,2-
dimethylpropanal to give, among other products, pyrrole 8b
The experimental technique employed in the preparation but no trace of dihydrodiazepine 11b. (iii) The third param-
of pyrroles from secondary cyclopropanediamines, viz. sim- eter is rooted in the substitution pattern of the 1,2-cyclopro-
ple addition of an excess of the aldehyde to solutions of panediamines. The N,NЈ-dimethyl derivatives 2b reacted
cyclopropanediammonium dibromides in acetate-buffered more rapidly than their primary analogs 2a with 2,2-di-
methanol, fails in the case of primary cyclopropanediam- methylpropanal to form a pyrrole (8c). The ring-methylated
ines. Instead of pyrroles, dihydrodiazepines 11 arise exclus- derivative trans-14a reacted even faster. Only one of two
ively except for 2,2-dimethylpropanal (3b) which affords bis- possible pyrroles, i.e. 19a, arose via azomethine ylide 16a.
imines 4b. The mechanistic study detailed above showed, No dihydrodiazepine was observed. This observation al-
however, that, at low concentrations of benzaldehyde, cis- lowed the connection of the present study with a previous
and trans-2a yielded a small amount of pyrrole 8a besides one, carried out in aqueous buffers, where the hydrolysis
dihydrodiazepine 11a. These observations were now ex- product 2-aminobutanone (20) was formed from the similar
ploited for a convenient synthesis of 2-phenylpyrrole (8a) azomethine ylide 17. Azomethine ylides 16 that are derived
by keeping the concentration of benzaldehyde low through- from an aromatic aldehyde undergo ring closure to a five-
Table 1. Starting material and experimental conditions in the synthesis of pyrroles from 1,2-cyclopropanediammonium dibromides and
aldehydes in acetate-buffered methanol; yields, melting points, solvents used for recrystallization, and boiling points
Compound
Starting
material
Time
Yield
[%]
M.p. [°C]
(Solvent)
B.p. [°C][a]
/Torr
Ref.
[b]
[b]
[b]
[4]
8a
8c
8d
8e
trans-2a·2HBr
trans-2b·2HBr
trans-2c·2HBr
trans-2c·2HBr
3a
3b
3a
3e
16 h
4 d
4 h
2 d
64
135
90Ϫ110/10Ϫ2
30/10Ϫ3
32
78
103Ϫ105/10Ϫ3
51[c]
104Ϫ105
(1. MeOH, 2. hexane)
125 (MeOH)
[4]
[4]
[4]
[4]
[b]
[4]
[4]
[4]
[4]
[1]
8f
trans-2c·2HBr
cis-2c·2HBr
trans-2c·2HBr
trans-2c·2HBr
trans-2c·2HBr
trans-14c·2HBr
trans-14c·2HBr
3f
3f
3g
3h
3i
3h
3e
4 h
75
65
55
95
54
57
38
29
62
71
3 h
8g
22 h
3 d
90/10Ϫ2
8h
67Ϫ68
8i
2.5 h
2 d
oil[d]
22h, 23h (7:3)
90/10Ϫ2
22e
23e
25a
25b
2 d
79Ϫ80 (petroleum ether)
119Ϫ120 (petroleum ether)
24a·2HBr
24b·2HBr
3e
3i
4 h
1 h
110/10Ϫ2
141Ϫ142 (MeOH/C6H6)
[a] Bath temperature during distillation in a sublimation apparatus with a cold finger (Ϫ78°C). Ϫ [b] This work. Ϫ [c] Yield after fractiona-
ting crystallization; the yield of the mixture of 8e and 9e (R1 ϭ CH2Ph, R2 ϭ 4-NO2ϪC6H4) was 94%. Ϫ [d] Purified by preparative TLC.
Eur. J. Org. Chem. 1998, 1645Ϫ1652
1649

W. von der Saal, R. Reinhardt, J. Stawitz, H. Quast
FULL PAPER
aq. NaOH (2 , 100 ml). The organic layer was dried with Na2SO4,
and the solvent was distilled in vacuo to yield colorless crystals
(10.5 g, 47%), m.p. 180Ϫ182°C (from methanol) (ref.[3]
182Ϫ183°C). Ϫ No trace of 8a could be detected with EhrlichЈs
pyrrole test[21] and proton spectroscopy in the residue, which was
obtained by addition of water to the mother liquor, acidification
to pH ϭ 5 and extraction with CCl4 followed by distillation of the
solvent and some 3a at 40°C/10Ϫ2 Torr.
membered ring so fast that hydrolysis cannot compete even
in aqueous buffers.
The synthesis of pyrroles presented here is related to the
pyrrole formation from 1-(1-piperidino)cyclopropyl keti-
mines which is catalyzed by tetrafluoroboric acid[15]. De-
spite a wealth of methods for the preparation of pyrroles[16]
there is a continued interest in novel approaches in general
and especially for 2-arylpyrroles[17]. The present synthesis
compares favorably to existing methods as far as it converts
aldehydes into pyrroles under very mild, neutral conditions,
viz. in buffered methanol solutions at room temperature.
These conditions render it ideal for use in combinatorial
Thermolysis of cis-4b ؊ cis-2,3-Di-tert-butyl-2,3-dihydro-1H-dia-
zepine (11b): Under Ar, a solution of cis-4b (252 mg, 1.20 mmol)
in dry toluene (15 ml) was heated under reflux for 4 d. The solvent
was distilled in vacuo until the final volume was 3 ml. Cooling of
the solution at 0°C yielded a crystalline precipitate which was col-
lected by filtration and recrystallized from toluene to give colorless
needles (124 mg, 49%), m.p. 203Ϫ205°C (subl. above 180°C). The
chemistry[18]
.
Financial support of this work by the Deutsche Forschungsge-
meinschaft and the Fonds der Chemischen Industrie, Frankfurt am
Main, is gratefully acknowlegded.
mother liquors contained mainly unchanged cis-4b. Ϫ IR (Nujol):
˜
ν ϭ 3190 cmϪ1 (NH), 1610 (CϭN). Ϫ UV (tetrahydrofuran): λmax
1
(ε) ϭ 301 nm (6600). Ϫ H NMR (CF3CO2D, which provides the
cation 1,4,6-[D3] 11bϩ): δ ϭ 1.10 (tBu), 1.25 (tBu), 3.13 (3-H), 4.43
(2-H), 7.61 (5-H), 7.73 (7-H). Ϫ MS; m/z (%): 208 (4) [Mϩ], 193
(11) [M Ϫ Me] 151 (11) [M Ϫ C4H9], 81 (100).
Experimental Section
General: Starting material, yields, and physical data: Table 1. Ϫ
1H NMR: Table 2. Ϫ 13C NMR: Table 3. Ϫ Molecular formulae
1
and masses and elemental analyses: Table 4. Ϫ H NMR: Bruker
Mechanistic Experiments
WM 400 and Varian EM 390. Spectra that were recorded for solu-
tions in acetate-buffered [D4]methanol were standardized with so-
dium 3-(trimethylsilyl)propanesulfonate (0.3 mg)[19]. The param-
eters of spectra of higher order were optimized with the program
LAOCOON III[20]. Ϫ 13C NMR: Bruker WM 400. Ϫ MS (70 eV):
Varian MAT CH 7.
1,2-Cyclopropanediammonium Dibromides and Aldehydes in Ace-
tate-Buffered [D4]Methanol: Cyclopropanediammonium dibro-
mide (0.10 mmol) and sodium acetate (a mmol) were dissolved in
[D4]methanol (0.7 ml) in 5 mm NMR sample tubes, followed by
addition of the aldehydes (b mmol). Ϫ (a) cis- or trans-2a.2HBr;
a ϭ 0.21; 3a, b ϭ 0.091. Ϫ (b) cis- or trans-2a·2HBr or cis- or trans-
2c·2HBr; a ϭ 0.22; 3b, b ϭ 0.10 or 0.20. Ϫ (c) trans-14a·2HBr; a ϭ
0.21; 3a or 3c, b ϭ 0.22.
Starting Material and Reference Compounds
3a was distilled prior to use. 3b was purified via the bisulfite
adduct and distilled[10]. 1,2-Cyclopropanediammonium dibromides
cis- and trans-2a؊c·2HBr, trans-14a, c·2HBr, 24a, b·2HBr were pre-
trans-4a (22.1 mg, 0.089 mmol), sodium acetate (16.3 mg, 0.20
mmol), and hydroxylamine hydrochloride (6.0 mg, 0.09 mmol) were
dissolved in [D4]methanol (0.7 ml). The reaction was initiated by
the addition of acetic acid (5 µl, 0.09 mmol).
pared as described[6]
.
cis-N,NЈ-Bis(2,2-dimethylpropylidene)-1,2-cyclopropanediamine
(cis-4b): Aqueous KOH (42%, 30 ml) was added to a suspension
of cis-2a·2HBr (7.02 g, 30.0 mmol) in diethyl ether (50 ml). The
ether layer was separated, and the aquous layer was extracted with
ether (10 ϫ 50 ml). The combined organic layers were dried with
powdered KOH for 15 min, and 3b (6.03 g, 70 mmol) was added.
The solution was kept for 19 h at 20°C, the solvent was distilled in
vacuo, and the residue (yellow oil) at 20Ϫ30°C/10Ϫ5 Torr to yield
Preparative Experiments
2-Phenylpyrrole (8a): A solution of 3a (1.09 g, 10.0 mmol) in
methanol (100 ml) was added dropwise during 9 h to a stirred solu-
tion of trans-2a·2HBr (2.34 g, 10.0 mmol) and sodium acetate (1.64
g, 20.0 mmol) in methanol (100 ml). After 16 h, the solvent was
distilled in vacuo, and the residue was extracted with diethyl ether
(4 ϫ 50 ml). The solvent was distilled in vacuo, and the solid resi-
due was sublimated at 90Ϫ110°C bath temp./10Ϫ2 Torr at a cold
finger (10°C), colorless crystals, m.p. 135°C (ref.[22] m.p. 137°C).
a colorless oil (4.31 g, 69%), which was stored at Ϫ25°C. Ϫ IR
˜
(film) ν ϭ 1660 cmϪ1 (CϭN). Ϫ 13C NMR (CDCl3): δ ϭ 14.4
(CH2), 27.1 (CMe3), 36.1 (Me), 46.1 (CH), 170.5 (NϭCH). Ϫ MS;
m/z (%): 208 (0.2) [Mϩ], 151 (1) [Mϩ Ϫ C4H9], 83 (100).
2-(1,1-Dimethylethyl)-1-methylpyrrole (8c): A solution of trans-
2a·2HBr (4.00 g, 15.3 mmol), sodium acetate (2.71 g, 33.0 mmol),
and 3b (2.63 g, 30.5 mmol) in methanol (120 ml) was stirred for 4
d in the dark and poured into sat. aq. KH2PO4 (250 ml). The mix-
ture was extracted with CH2Cl2 (5 ϫ 100 ml). The combined or-
ganic layers were dried with K2CO3. The solvent was distilled in
vacuo and the residue sublimed at 30°C bath temp./10Ϫ3 Torr at a
trans-N,NЈ-Bis(2,2-dimethylpropylidene)-1,2-cyclopro-
panediamine (trans-4b) was prepared, as described for cis-4b, from
˜
trans-2a·2HBr in 79% yield as colorless oil. Ϫ IR (film): ν ϭ 1655
cmϪ1 (CϭN). Ϫ MS; m/z (%): 208 (0.5) [Mϩ], 151 (78) [M Ϫ C4H9],
83 (100).
Attempted Thermolysis of trans-4b: trans-4b (200 mg, 1 mmol) cold finger (Ϫ50°C) to yield a yellow oil, which was purified by
was distilled (40°C, 10Ϫ5 Torr) into a vial, which was sealed at 10Ϫ5
bulb-to-bulb condensation at 30°C bath temp./10Ϫ3 Torr to afford
Torr and heated to 100°C for 4 d. Trans-4b was recovered un- a colorless oil. Ϫ 13C NMR (CDCl3): δ ϭ 30.3 (Me), 31.8 (CMe3),
1
changed (GC, IR, H NMR).
36.5 (NMe), 105.0 (C-3), 105.7 (C-4), 123.5 (C-5), 141.1 (C-2). Ϫ
MS; m/z (%): 137 (30) [Mϩ], 122 (100) [M Ϫ Me], 107 (16) [M Ϫ
C2H6], 94 (14) [M Ϫ C3H7].
cis-2,3-Dihydro-2,3-diphenyl-1H-diazepine (11a): A mixture of
trans-2a·2HBr (9.59 g, 41.0 mmol), sodium acetate (7.38 g, 90.0
mmol), and 3a (9.66 g, 91.0 mmol) in methanol (320 ml) was stirred
Synthesis of Pyrroles Ϫ General Procedure: Solutions of cyclo-
for 3 d at 20°C. The solvent was distilled in vacuo. The residue was propanediammonium dibromide (6 mmol), sodium acetate (16
dissolved in CH2Cl2 (150 ml), and the solution was extracted with mmol) and aldehyde (3, 13 mmol) in 30 ml methanol were stirred
1650
Eur. J. Org. Chem. 1998, 1645Ϫ1652

Cyclopropanediamines, 7
FULL PAPER
Table 2. Chemical shifts (δ values) and coupling constants (J, [Hz]) in 400-MHz proton spectra
Compound Alkyl groups
CHϭN
Ring protons
3Ј-H[a]
2J
3Jcis
3Jtrans
3J1,2
[b]
1-H, 2-H
3-H
cis-2a
2.54
2.63
1.02
1.01
1.02
1.07
1.12
1.27
1.25
1.29
1.29
0.57
0.68
1.43
Ϫ7.1
7.8
4.4
M
M
M
M
T
[c]
trans-2a
cis-2b
Me
2.60
2.58
1.01
2.55
Ϫ7.1
Ϫ7.2
7.6
8.3
7.3
4.5
4.9
5.1
trans-2b
cis-4b
Me
2.61
2.2
2.1
tBu
7.58
3.08
Ϫ6.1
[c]
trans-4b
tBu[d]
7.72
3.11
M
T
[c]
1.03
7.64
7.57 (br.)
7.81
3.07
cis-1b
trans-1b
tBu[d]
tBu[d]
3.02 (br.)
2.82, 3.18
1.10[e]
1.27
Ϫ7.2
Ϫ11.5
7.6
4.4
M
M
[f]
[f]
1-H,
5-H
3-H
6-H
6Ј-H
[g]
[g]
exo-12c
tBu
Me
tBu
Me
0.90
2.49
0.82
2.31
3.02
2.70
0.12
Ϫ6.2
Ϫ6.7
6.0
5.8
3.0
2.9
M
M
[g]
endo-12c
2.90
Ϫ0.09
Aryl-H
3-H
4-H
5-H
3J34
4J35
3J45
[g]
8a
6.44
6.51
6.13
6.28
6.79
6.71
3.4
3.5
1.5
1.5
2.7
2.7
M
T
7.20 (1H),
7.35 (2H),
7.45 (2H)
8b
tBu[g]
5.73
5.76
5.80
5.87
5.91
5.82
6.48
6.45
6.46
3.4
3.5
3.6
1.8
1.5
2.0
2.8
2.75
2.8
M
C
M
ref.[23]
8c
1.27
1.32
3.71
5.16
tBu
Me
8i
CH2Ph
6.8Ϫ7.3[h]
6.56
6.23
6.71
3.7
1.7
2.7
T
(E)-PhCHϭCH
19a
19c
Me
Me
2.22
2.15
7.2Ϫ7.4
7.2Ϫ7.6
5.99
6.02
6.70
6.74
2.4
2.4
M
M
[a] 3Ј-H and 6Ј-H are trans to 1-H. Ϫ [b] Solvent: M ϭ acetate-buffered [D4]methanol, T ϭ [D]trichloromethane, C ϭ tetrachloromethane.
[d]
Ϫ [c] Coupling constants could not be determined (unresolved multiplet). Ϫ The signal of the tert-butyl group was not unambiguously
[e]
identified. Ϫ The ABX2 spectrum of the cyclopropane ring protons is the result of rapid exchange via diazabicyclo[3.1.0]hexane 12b.
[f] 3
3
3
3
[g]
[h]
Ϫ
J
:3J1,3 ϭ 8.8, J2,3Ј ϭ 8.8, Jtrans:3J1,3Ј ϭ 4.2, J2,3 ϭ 4.4 Hz. Ϫ The signal was obscured. Ϫ Including the vinyl protons.
cis
Table 3. Chemical shifts (δ values) in carbon-13 spectra
[a]
Compound
Cyclopropane ring
Alkyl groups
CH2
CH
CϭN
trans-4b
cis-4b
17.5
14.4
50.1
46.1
169.6
170.5
tBu
tBu
35.8, 27.0
36.1, 27.1
T
T
pyrrole ring
phenyl
C-2
C-3
106.6
C-4
C-5
ϭCHϪ[b]
i-C
8a
8c
8d
133.4
110.4
120.3
123.5
124.7
125.0
126.9
130.1
135.5
D
T
T
141.1
141.1
105.0
109.6
105.7
110.3
tBu
NMe
CH2
31.8, 30.3
36.5
51.9
127.8
128.3
128.6
112.1
122.2
128.3
123.3
125.9
126.4
126.9
117.0
120.8
124.2
125.4
125.8
129.6
129.8
130.2
129.0
130.5
135.4
8f
8i
139.4
131.9
107.4
107.1
107.8
108.9
126.8
117.1
120.6
134.3
149.2
137.8
138.1
NMe
CH2
39.7
49.6
D
T
127.6
128.6
128.8
CH2
50.6
25b
138.7
103.9
123.7
125.2
126.5
126.7
127.2
128.5
132.1
135.0
137.2
CH2
49.5
D
[a]
[b]
Solvent: T ϭ [D]trichloromethane; D ϭ [D6]dimethyl sulfoxide. Ϫ The vinyl carbon atoms of 8i and 25b are included.
during the periods of time listed in Table 1. The conversions were
monitored by TLC [Al2O3, petroleum ether (50Ϫ70°C)]. The sol-
vent was distilled in vacuo. The residue was extracted with diethyl
ether. The organic layer was dried with MgSO4 and the solvent
Eur. J. Org. Chem. 1998, 1645Ϫ1652
1651

W. von der Saal, R. Reinhardt, J. Stawitz, H. Quast
FULL PAPER
[7]
[8]
H. Quast, J. Stawitz, Tetrahedron Lett. 1977, 2709Ϫ2712.
E. L. Eliel, S. H. Wilen, Stereochemistry of Organic Compounds,
1st ed., Wiley-Interscience, New York, 1994, p. 705.
Iminium ions are more acidic by three orders of magnitude than
the corresponding ammonium ions: W. P. Jencks, Acc. Chem.
Res. 1976, 9, 425Ϫ432.
H. Quast, H. Leybach, Chem. Ber. 1991, 124, 2105Ϫ2112.
J. B. Lambert, G. Wang, D. E. Huseland, L. C. Takiff, J. Org.
Chem. 1987, 52, 68Ϫ71; J. B. Lambert, M. W. Majchrzak, J.
Am. Chem. Soc. 1980, 102, 3588Ϫ3591; 1979, 101, 1048Ϫ1049.
T. H. Fife, A. M. Pellino, J. Am. Chem. Soc. 1981, 103,
1201Ϫ1207; 1980, 102, 3062Ϫ3071; T. H. Fife, J. E. C. Hutch-
ins, A. M. Pellino, J. Am. Chem. Soc. 1978, 100, 6455Ϫ6462; T.
H. Fife, J. E. C. Hutchins, J. Am. Chem. Soc. 1976, 98,
2536Ϫ2543.
L. L. McCoy, G. W. Nachtigall, J. Am. Chem. Soc. 1963, 85,
1321Ϫ1324
A. Bruylants, E. Feytmants-de Medicis in The chemistry of the
carbon-nitrogen double bond (Ed.: S. Patai), Interscience, Lon-
don, 1970, 465Ϫ504.
H. H. Wasserman, J. M. Fukuyama, Tetrahedron Lett. 1991, 48,
7127Ϫ7130; H. H. Wasserman, R. P. Dion, Heterocycles 1989,
28, 629Ϫ631; H. H. Wasserman, R. P. Dion, Tetrahedron Lett.
1983, 24, 3409Ϫ3412.
R. J. Sundberg, P. V. Nguyen, Prog. Hetercycl. Chem. 1994, 6,
110Ϫ128; R. J. Sundberg, Prog. Heterocycl. Chem. 1992,
81Ϫ94; T. L. Gilchrist, Contemp. Org. Synth. 1994, 1, 205Ϫ217.
Ϫ For a review on the formation of pyrroles from cyclopro-
panes, see: O. G. Kulinkovich, Russ. Chem. Rev. 1993, 62,
839Ϫ850.
Table 4. Molecular formulae and masses, and elemental analyses
[9]
Compound Formula
Mol.
mass
Elemental analyses
C
H
N
[10]
[11]
cis-4b
C13H24N2
208.4 calcd. 74.95 11.71 13.45
found 74.89 11.60 13.48
trans-4b
11b
8a
found 75.00 11.60 13.68
found 74.82 11.70 13.56
[12]
C10H9N
143.2 calcd. 83.88 6.34 9.78
found 83.54 6.45 9.89
8c
8d
8e
8f
C9H15N
137.2 calcd. 78.78 11.02 10.21
found 78.91 11.28 10.12
C17H15N
C17H14N2O2
C19H20N2
C16H14N2
C20H21N
C19H17N
C21H23N
C18H16N2O2
233.3 calcd. 87.52 6.46 6.00
found 87.84 6.68 6.04
[13]
[14]
278.3 calcd. 73.36 5.07 10.07
found 72.87 5.25 9.93
276.4 calcd. 82.57 7.29 10.14
found 81.70 7.48 9.87
[15]
[16]
8g
8h
8i
234.3 calcd. 82.02 6.02 11.96
found 82.11 5.98 11.29
275.4 calcd. 87.23 7.69 5.08
found 87.22 7.86 5.05
259.4 calcd. 87.99 6.61 5.40
found 87.66 6.73 5.33
22h, 23h
(7:3)
22e
289.4 calcd. 87.15 8.01 4.84
found 86.83 8.11 4.95
292.3 calcd. 73.95 5.52 9.58
found 74.25 5.59 9.56
[17]
P. Nagafuji, M. Cushman, J. Org. Chem. 1996, 61, 4999Ϫ5003;
H. Ogawa, T. Aoyama, T. Shioiri, Heterocycles 1996, 42, 75Ϫ82;
A. R. Katrizky, J. Li, J. Org. Chem. 1996, 61, 1624Ϫ1628; J.
23e
25a
25b
found 74.61 5.97 9.40
found 73.30 5.11 9.31
C25H21N
335.4 calcd. 89.51 6.31 4.18
found 89.63 6.38 4.04
´
´
Barluenga, M. Tomas, V. Kouznetsov, A. Suarez-Sobrino, E.
Rubio, J. Org. Chem. 1996, 61, 2185Ϫ2190; I. V. Magedov, A.
V. Magedov, A. V. Kornienko, T. O. Zotova, V. N. Drozd, Tetra-
hedron Lett. 1995, 36, 4619Ϫ4622; T. Masquelin, D. Obrecht,
Synthesis 1995, 276Ϫ284; A. Fürstner, H. Weintritt, A. Hup-
perts, J. Org. Chem. 1995, 60, 6637Ϫ6641; K. D. Barnes, R.
Ward, J. Heterocycl. Chem. 1995, 32, 871Ϫ874; A. R. Katrizky,
H.-X. Chang, S. V. Verin, Tetrahedron Lett. 1995, 36, 343Ϫ346;
S. Mataka, H. Kitagawa, O. Misumi, M. Tashiro, K. Kamata,
J. Chem. Soc. Chem. Commun. 1993, 670Ϫ671.
A. Nefzi, J. M. Ostresh, R. A. Houghton, Chem. Rev. 1997,
97, 449Ϫ472.
distilled in vacuo. Purification of the residue by preparative TLC
or flash chromatography was followed by distillation or subli-
mation of the product at a cold finger (Ϫ78°C) and/or recrystalli-
zation from the solvents listed in Table 1.
[18]
[1]
W. von der Saal, H. Quast, J. Org. Chem. 1995, 60, 4024Ϫ4029.
[2]
[19]
[20]
W. von der Saal, H. Quast, J. Org. Chem. 1996, 61, 4920Ϫ4928.
G. V. D. Tiers, R. I. Coon, J. Org. Chem. 1961, 26, 2097Ϫ2098.
A. A. Bothner-By, S. Castellano in Computer Programs for
Chemistry (Ed. D. F. DeTar), vol. 1, 1st ed., Benjamin, New
York, 1968.
[3]
H. Quast, H.-M. Seidenspinner, J. Stawitz, Liebigs Ann. Chem.
1983, 1207Ϫ1229.
[4]
Preliminary communication: H. Quast, W. von der Saal, J. Sta-
[21]
[22]
[23]
witz, Angew. Chem. 1981, 93, 619Ϫ620; Angew. Chem. Int. Ed.
Engl. 1981, 20, 588Ϫ589.
H. Fischer, H. Orth, Die Chemie des Pyrrols, 1st ed., Akademi-
sche Verlagsanstalt, Leipzig, 1934, p 66.
[5]
H. Quast, W. von der Saal, W. Chem. Ber. 1984, 117,
H. Berner, G. Schulz, H. Reinshagen, Monatsh. Chem. 1977,
108, 285Ϫ297.
1591Ϫ1596.
[6]
W. von der Saal, R. Reinhardt, H.-M. Seidenspinner, J. Stawitz,
G. A. Cordell, J. Org. Chem. 1975, 40, 3161Ϫ3169.
[98121]
H. Quast, Liebigs Ann. Chem. 1994, 569Ϫ579.
1652
Eur. J. Org. Chem. 1998, 1645Ϫ1652