Reaction of the lithio-derivative of methoxyallene with hydrazones. Part 2: Formation of 3-pyrrolines and azetidines; synthetic and mechanistic aspects

Article abstract of DOI:10.1016/S0040-4020(01)00031-X

The reaction of α-lithiomethoxyallene with aromatic hydrazones leads to 3-pyrrolines when run in THF. In the case of SAMP-hydrazones, this reaction occurs with almost complete stereodifferentiation. With other hydrazones (dimethyl, piperidinyl, morpholinyl), the 3-pyrrolines are accompanied by azetidines. However, formation of the latter compounds is reversible, since they are transformed to 3-pyrrolines when the time and temperature of the reaction are increased, but the rate of this transformation depends on the substituents of the terminal nitrogen. A reaction mechanism is proposed which involves intermolecular or intramolecular electron transfers from the lithium amide leading to hydrazinyl radicals. The relative stabilities of these intermediates may then explain the role of the substituents of the terminal nitrogen in the formation of azetidines by a 4-exo-dig radical process.

Full text of DOI:10.1016/S0040-4020(01)00031-X

TETRAHEDRON
Pergamon
Tetrahedron 57 (2001) 1951±1960
Reaction of the lithio-derivative of methoxyallene with
hydrazones. Part 2: Formation of 3-pyrrolines and azetidines;
synthetic and mechanistic aspects
Â
Âp
Valerie Breuil-Desvergnes and Jacques Gore
 Ã
Laboratoire de Chimie Organique I, associe au CNRS, University Claude Bernard, UMR 5622, Bat 308, CPE-Lyon,
43 Bd du 11 novembre 1918, 69622 Villeurbanne, France
Received 2 August 2000; revised 7 December 2000; accepted 4 January 2001
AbstractÐThe reaction of a-lithiomethoxyallene with aromatic hydrazones leads to 3-pyrrolines when run in THF. In the case of SAMP-
hydrazones, this reaction occurs with almost complete stereodifferentiation. With other hydrazones (dimethyl, piperidinyl, morpholinyl), the
3-pyrrolines are accompanied by azetidines. However, formation of the latter compounds is reversible, since they are transformed to
3-pyrrolines when the time and temperature of the reaction are increased, but the rate of this transformation depends on the substituents
of the terminal nitrogen. A reaction mechanism is proposed which involves intermolecular or intramolecular electron transfers from the
lithium amide leading to hydrazinyl radicals. The relative stabilities of these intermediates may then explain the role of the substituents of the
terminal nitrogen in the formation of azetidines by a 4-exo-dig radical process. q 2001 Elsevier Science Ltd. All rights reserved.
In the preceding paper,1 we have shown that the reaction of
the lithio-derivative of methoxyallene with aldehyde hydra-
zones leads, when the solvent is ether, to a-allenic hydra-
zines 1. These compounds, when treated in THF with
2 equiv. of butyllithium, are cleanly transformed in the
case of aromatic hydrazines (Rˆphenyl, naphthyl, etc.) to
N-dialkylaminopyrrolines 2. The cyclization process
appears to be particularly easy in the case of SAMP-hydra-
zines which are transformed to 2a, regardless of the ali-
phatic or aromatic nature of R. Still more interesting is the
almost complete diastereoselection observed in the case of
these SAMP-hydrazones which permits, after hydrogen-
olysis of the N±N bond, the 3-pyrrolines 3 to be obtained
with high enantiomeric purity (Scheme 1).
In this paper, we will provide evidence to show that the same
reaction of hydrazones with the lithio-derivative of methoxy-
allene leads, in many cases, directly to the N-dialkyl-
aminopyrrolines 2 when run in THF. We will discuss the
in¯uence of structural and experimental factors, which
favour cyclization and ®nally propose a reaction mechanism
based on the intermediate formation of a hydrazinyl radical.
Scheme 1.
Keywords: allenes; pyrrolines; azetidines; radical cyclizations; hydrazinyl radicals.
Corresponding author. Tel.: 133-4-72-44-81-35; fax: 133-4-72-43-12-14; e-mail: gore@univ-lyon1.fr
p
0040±4020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(01)00031-X

Â
V. Breuil-Desvergnes, J. Gore / Tetrahedron 57 (2001) 1951±1960
1952
Table 1.
1. Results
1.1. Reaction in THF at 2258C
The ®rst reactions were made by adding, at 2788C, the
preformed lithio-derivative2 in THF to a solution of the
hydrazone in the same solvent. After 2 h at 2788C,
the temperature was raised to 2258C and maintained at
that temperature for 16 h. After work up, the pyrrolines 2
were isolated by ¯ash chromatography; the allenic hydra-
zines 1 are fairly unstable under these chromatographic
conditions, but the crude product is generally of good purity
Hydrazone
Product
Yield
(%)a
D.e.
(%)
ArˆPh
88
.99
1
as shown by H and 13C NMR spectroscopy.
Under these conditions, only the reactions involving SAMP-
hydrazones lead to the N-dialkylamino pyrrolines 2a and
cyclization appears to be limited to aryl-hydrazones. The
results of these reactions, listed in Table 1, show that the
yields, based on the starting aromatic aldehydes, are gener-
ally good and that the diastereoselection is complete in all
Arˆb-naphthyl
76
.99
1
cases, as analysed by GC±MS, H and 13C spectroscopy.
The only exception was 2a3, for which GC±MS analysis
revealed the presence of 0.4% of the other diastereomer.
Based on the model proposed by Enders et al. as well as
on our previous results1 (X-ray analysis of 2a1), the S con-
®guration was assigned to the stereogenic carbon.
Arˆp-CH3Ph
79
.99b
All the reactions carried out with other hydrazones, under
the same conditions, lead to the corresponding a-allenyl
hydrazines 1 or to a mixture of them with cyclized products.
Aliphatic SAMP-hydrazones are uniformly converted to 1
with modest yield [RˆEt: 38%; RˆnC5H11: 30%; RˆisoPr:
60%; Rˆt-But: 50%] but with d.e. .95%.
Arˆp-CH3OPh
75
.99
The behaviour of morpholinyl hydrazones is more complex
since they lead, for the aromatic ones, to a mixture of the
allenylhydrazines 1b, of the N-morpholinyl pyrrolines 2b
and of N-morpholinyl azetidines 4b (Scheme 2). As found
for SAMP-hydrazones, the aliphatic morpholinyl hydra-
zones are converted to the a-allenyl hydrazines [RˆEt:
85%; RˆisoPr: 85%; RˆtBu: 75%]. It is noteworthy that,
in this series, when the reaction was run with cyclohexanone
hydrazone, the corresponding hydrazine was obtained with
a 60% yield (the corresponding SAMP-hydrazone is almost
inert under similar conditions).
Arˆp-FPh
69
.99
a
Yields were determined by comparing the amount of 2a isolated by ¯ash
chromatography to that of the starting aldehydes ArCHO.
Scheme 2.

Â
V. Breuil-Desvergnes, J. Gore / Tetrahedron 57 (2001) 1951±1960
1953
In the morpholinyl series, the reaction run at room tempera-
ture still leads to a mixture of the allenyl hydrazine 1b, of
the N-morpholinyl pyrroline 2b and of the N-morpholinyl
azetidine 4b (see Scheme 2) with small variations in com-
position, compared to the products obtained in the reaction
run at 2258C. For example, in the case of the phenylhydra-
zone, the reaction at room temperature for 16 h produces
1b1 (5%), 2b1(66%) and 4b1 (10%). If the reaction time is
increased to 96 h the allenyl hydrazine 1b1 is no longer
observed and the product is composed of 2b1 (59%) and
4b1 (15%). However, the increase of the temperature to
508C has a more spectacular effect since 2b1 is obtained
exclusively. The same conditions were applied to other
aromatic hydrazones, leading to morpholinyl pyrrolines 2b
(Scheme 5).
Scheme 3.
The dimethylhydrazones as well as their piperidinyl
homologues are converted only into the allenyl hydrazines
1, regardless of the nature of R (Scheme 3).
The results of these reactions, run at 2258C, show that two
factors facilitate the cyclization: the nature of the hydrazino
group (SAMP-hydrazones are by far the best substrates for
obtaining the pyrrolines 2) and the aromatic nature of R. The
formation of the azetidine 4b in the case of aromatic
morpholinyl hydrazones has also to be underlined.
On the contrary, the increase in temperature has no effect
when aliphatic hydrazones are used: the a-allenylhydrazine
1.2. Reaction in THF at higher temperatures
Scheme 4.
To try to obtain the N-dialkylamino-3-pyrrolines 2 prefer-
entially, experiments were repeated at higher temperatures.
Aliphatic SAMP-hydrazones were converted to the corre-
sponding 2a when treated in THF with 6 equiv. of a-lithio-
methoxyallene for 16 h at 258C (Scheme 4). As was found in
the case of their aromatic counterparts, the yield are good
and the diastereoselection is high.
These results show that, in this series, the pyrrolines 2a are
generally obtained. Consequently, the parent 3-pyrrolines 3
can probably be formed with a high enantiomeric excess
(>95%), after hydrogenolysis of the N±N bond.1
Scheme 5.
Scheme 6.

Â
V. Breuil-Desvergnes, J. Gore / Tetrahedron 57 (2001) 1951±1960
1954
1b is the only reaction product formed at room temperature
as well as at 508C.
place more readily in THF than in ether and is probably
much easier in the case of SAMP-hydrazinide, where the
lithium can be coordinated intramolecularly by the methoxy
group of the SMP moiety.
The results of the reactions of dimethylhydrazones as well as
piperidinyl hydrazones are slightly different, except for
alkyl hydrazones which are transformed only to a-allenyl
hydrazines 1 regardless of the temperature (225 to 508C).
In both the cases, with aromatic hydrazones, a mixture of
pyrrolines 2 and azetidines 4 is obtained at room tempera-
ture as well as at 508C. The proportions of these compounds
are only slightly in¯uenced by the increase of temperature.
Unfortunately, it was impossible to obtain pure 2 by
increasing the reaction time at 508C. If the ratio 2:4 is
increased, when longer times are used (see below), there
is a parallel formation of tars which complicates the puri®-
cation of 2 and decreases the yield of the process.
Figure 1.
The problem remains, however, of determining the mode of
cyclization of 6 to 2 and eventually to 4. If we take
Baldwin's rules3 into consideration, a concerted 4-exo-dig
process can be postulated for the formation of 4 from 6.
On the contrary a concerted 5-endo-trig process for the
transformation of 6 to 2 has to be disregarded. Generally,
this cyclization of allenyl amines or homologues to nitrogen
heterocycles (aziridines to piperidines) is promoted by
electrophilic species (H1, Ag1, Cu1) or by palladium
complexes. In the ®rst case an intermediate, allylic cation,
is generally proposed which is transformed into 5- or
6-membered rings by apparent exo-trig4 or endo-trig5
processes.
High selectivity was observed only with b-naphthyl hydra-
zones which are converted at 508C (almost) exclusively to
N-dimethylamino- or N-piperidinyl-3-pyrrolines 2 (Scheme
6).
2. Discussion
In this paper, and as previously reported,1 the reaction of
a-lithiomethoxyallene with hydrazones can produce,
depending on the conditions, either the expected a-allenyl
hydrazines 1 or cyclized products N-dialkylamino-3-pyrro-
lines 2 or even N-dialkylamino azetidines 4. We have also
shown that 1 is converted to the cyclized products 2 in
certain cases and, eventually, 4 when treated in THF by
2 equiv. of n-butyllithium. This last result proves that the
lithium hydrazinide 5 is able to cyclize only in THF; it is
stable in Et2O at 2258C, where its hydrolysis produces the
hydrazines 1.
Similarly, the cyclization catalysed by palladium (II)
complexes has, depending on the substrate and on the
complex involved, either a p-allylpalladium5 or a s-vinyl-
palladium6,7 intermediate which can be converted,
apparently by endo-dig or exo-trig processes, into 3- to
6-membered rings.
The only cyclization of an amide formed by the reaction of a
base with an a-allenyl amine has only very recently been
described: it leads cleanly to pyrrolines or pyrroles depend-
ing on the conditions.8 This cyclization, which appears to be
closely related to the transformation 1!2, has not been
explained from a mechanistic point of view.
Even in THF, the cyclization is not general and its behaviour
is in¯uenced by temperature and mainly by two structural
factors: the substituents of the terminal nitrogen of the
hydrazine and the nature, aromatic or aliphatic, of the
starting aldehyde. From the presented results, it appears
that cyclization, essentially to 2, is easier when SAMP-
hydrazones of aromatic aldehydes are involved.
On the contrary, the cyclization of a-allenic alcohols to
dihydrofuranes and/or epoxides has been described by
Magnus et al.9 It occurs only in DMSO and was explained
by electron and hydrogen transfers between the substrate
and the solvent.
Hence, it would appear that cyclization is favoured by a
partial or total dissociation of the N±Li bond of 5, giving
the anion 6 (Fig. 1). This dissociation, as expected, takes
In spite of some differences (we never observed aziridines,
Scheme 7.

Â
V. Breuil-Desvergnes, J. Gore / Tetrahedron 57 (2001) 1951±1960
1955
Table 2.
Time (h)
T (8C)
1d1
2d1
4d1
12
24
60
6
20
45
225
25
25
50
50
50
100%
25%
e
±
±
±
±
±
45%
50%
50%
68%
85%
30%
50%
50%
32%
15%
and Magnus et al. did not obtain 4-membered rings), it may
be possible to propose, for the cyclization 5!214, a similar
mechanism with a single electron transfer from the nitrogen
to the allenyl moiety (Scheme 7).
the reaction by GC. For this experiment, the reaction of the
piperidinyl hydrazone of benzaldehyde with 6 equiv. of
a-lithiomethoxyallene in THF was chosen since it gives
rise, depending on the conditions, to three possible products
1d1, 2d1 and 4d1. The results, given in Table 2, show the
percentages of 1d, 2d and 4d, formed at different reaction
times and temperatures in the same reaction mixture.
Samples were analysed by GC using dodecane as internal
reference and the proportions of each product were deter-
mined after standardization with puri®ed compounds. The
evolution of the composition shows clearly that (i) lithium
allenyl hydrazide 6 is completely transformed to the
cyclized products 2d1 and 4d1 with the increase in time
and temperature, and (ii) transformation of 4d1 to 2d1 is
observed when these two factors are increased, thus showing
the reversibility of the formation of the 4-membered ring.
This proposal seems to agree with the formation of the
5-membered ring, but is less appropriate with that of the
4-membered ring. On one hand, the canonical form 7b,
which is evidently more stable than 7a, is probably more
representative of the structure of 7 and, on the other hand, 2
is favoured to 4 by 21±34 kcal/mol depending on the sub-
stituents R of the terminal nitrogen (hyperchem software):
such a mechanism would lead essentially to the pyrrolines 2.
Consequently, two questions remain concerning (i) the
mechanism for the formation of 4 and (ii) the nature of
the electron transfer agent which was supposed to be
DMSO in the Magnus mechanism.
As it appears that the cyclization is favoured in the direct
reaction of lithioallene with hydrazones in THF, compared
to that of n-butyllithium with hydrazines 1 in the same
solvent, the question of the role of the excess lithioallene
in the cyclization process had to be addressed. To determine
this effect, the reaction of n-BuLi with the piperidinyl
In an attempt to answer these two questions, two sets of
additional reactions were undertaken.
The in¯uence of experimental conditions (time and
temperature) was more carefully determined by following
Table 3.
Co-reagent
±
Time (h)
60
T (8C)
1d2
2d2
4d2
25
100%
±
±
2 equiv.
16
16
6
225
25
50
100%
±
±
±
66%
97%
±
34%
3%
id 6 equiv.
CuCl2 10%
16
16
25
25
±
±
100%
89%
±
11%

Â
V. Breuil-Desvergnes, J. Gore / Tetrahedron 57 (2001) 1951±1960
1956
Scheme 8.
naphthylhydrazine 1d2 was run under different conditions as
depicted in Table 3.
b-allenyl hydrazinyl radical 8 can be envisaged. These
hydrazinyl radicals are not well-known species, and it is
only recently that they were mentioned in the literature.11
Comparatively, the cyclization of carbon radicals by a
reversible 4 exo-trig process is known.12 Indeed, it has
been shown that the radical cyclization of N-vinyl a-halo-
geno amides leads to b-lactams at room temperature while it
gives g-lactams in re¯uxing toluene.13 This result has been
explained by a kinetic 4-exo process which reverses at
higher temperature leading to the more stable 5-membered
ring isomer. Lastly, the cyclization of the b-allenyl carbon
radical was described and it gives vinylcyclopropane14 in
low yield. Comparatively, the cyclization of 3,4-penta-
dien-1-yllithium gives exclusively a-methyl-lithiocyclo-
butene while the corresponding Grignard reagent is
transformed to its vinylcyclopropyl isomer.15 It is note-
worthy that the formation of cyclopentene was not described
in the reports of Crandall et al.
The results show that the a-lithiomethoxyallene plays a role
in the cyclization process, which evolves with the increase
of temperature and eventually with that of time, leading to
the formation of the pyrroline 2d2 only. As a catalytic
quantity of cupric chloride exerts the same in¯uence, it
appears that the allenic organometallic is probably an
electron transfer agent in the cyclization process, trans-
forming ®rst the amide 6 to the hydrazinyl radical 8,
which is then reduced to 7 which then cyclizes preferentially
to 2 (Scheme 8).
Even if the formation of the radical anion of a-lithiomethoxy-
allene is surprising, it is supported by the formation of a
C3H2Li2 species by treatment of allene with 2 equiv. of
n-BuLi, one possible structure for the species being bis
lithioallene.10 If the methoxyallene moiety is able to
accommodate two negative charges, it is reasonable to
suppose that a three-electron species has also a certain
stability.
In spite of these two lines of evidence (the chemistry of
hydrazinyl radicals is not well known and the cyclization
of b-allenyl radicals or organometallics gives troublesome
results), we make the hypothesis that the azetidines 4 come
from a reversible 4-exo-dig cyclization of the intermediate
8. Then, depending on the nature of the R and R0 sub-
stituents, the reversibility of the process can operate at
As mentioned before, the cyclization of 7 would lead prefer-
entially to the pyrroline 2, but its evolution to the azetidine 4
is far less probable. In comparison, a radical cyclization of
Scheme 9.

Â
V. Breuil-Desvergnes, J. Gore / Tetrahedron 57 (2001) 1951±1960
1957
These different behaviours can probably be related to the
basicity of the NR2 group. From literature data,18 it appears
that, if dimethylamine and piperidine have about the same
basicity constants, morpholine is, depending on the solvent
and temperature, 102±103 times less basic. This loss of basi-
city is probably due to the electron withdrawing character of
the b-carbon±oxygen bond since (2-methoxyethyl)methyl-
amine has a basicity of the same magnitude as morpholine.
It is consequently reasonable to think that SMP, which also
has a b-carbon±oxygen bond, has a basicity in the same
range as that of morpholine. On this basis, it appears that
cyclization, particularly cyclization to 2, is favoured by a
less basic terminal nitrogen. The difference between
SAMP-hydrazones and morphinyl hydrazones could be
the consequence of the coordination of the lithium cation
by the methoxymethyl group which favours the dissociation
of the N±Li bond of 6 giving 10, where the basicity of
the pyrroline nitrogen is decreased by the electron-with-
drawing effect of the coordinated ether functionality
(Scheme 10).
different rates depending on the stability of 8. It can give the
pyrrolines 2, either by a disfavoured 5-endo process, or
more probably via 7 due to a double electron transfer
(Scheme 9).
In this hypothesis, the relative proportions of 2 and 4 at a
given temperature would re¯ect the relative stability of the
radicals 8 and 9. Since the stability of 9 seems to be inde-
pendent of the substituents R and R0, the main factor
governing these proportions and the rate of ring opening
of 9 would be the stability of the hydrazinyl radical 8,
which can be in¯uenced by the following two factors.
The aliphatic or aromatic nature of R0. In the case of an
aromatic substituent, the nitrogen radical can be stabilized
by homoconjugation, due to its homobenzylic character.16
Hence, cyclizations are easier in the case of the reaction of
aromatic hydrazones since the transformation of 2 is
favoured for a b-naphthyl substituent compared to a phenyl
substituent (see Scheme 6). The increase of the stability of 7
and 8 can then explain either the easier formation of 7 from
6 or the faster reversibility of the 4-exo-dig process 8!9.
The nature of the R substituent can also in¯uence the
stability of the nitrogen radical of 7 and 8 by varying the
electron donating ability of the terminal nitrogen.
The stability of a-amino carbon radicals is governed by
the substituents of amino groups: it is known that the
dimethylamino group stabilizes a vicinal radical by about
10 kcal/mol compared to an NH2 group.17 It is reasonable to
think that the same effect can be obtained with a hydrazinyl
radical which would be better stabilized by a more basic
terminal nitrogen. In our case, large differences in the
behaviour of the cyclization process were observed depend-
ing on the nature of the starting hydrazones:
Scheme 10.
With these considerations, one explanation for all the
experimental results would be that the amide 6 has, in
THF, two possibilities: it can give rise, directly by intra-
molecular electron-transfer, to the bis-radical anion 7
which rapidly cyclizes to the 5-membered ring or it can
transfer the electron to a-lithiomethoxyallene (see Scheme
9) giving the allenyl hydrazinyl radical which cyclizes
reversibly to the 4-membered ring (Scheme 11).
² in order to observe cyclization with dimethyl and piperi-
dinyl hydrazones, the reaction must be run at room
temperature where it gives rise to a mixture of 2 and 4,
with a slow evolution at higher temperature and longer
time towards the preferential formation of 2 (Scheme 6
and Tables 2 and 3).
² with aromatic morpholinyl hydrazones, the cyclization is
partially observed at 2258C giving both 2 and 4, but the
reaction at 508C leads only to 2.
² with SAMP-hydrazones, the cyclization only to pyrro-
lines 2 is easier since it occurs at 2258 with aromatic
hydrazones (Table 1) and at room temperature with their
aliphatic counterparts. In this series, the azetidine 4 was
never observed.
As mentioned before, both processes are favoured when R0
is aromatic but it seems that 8 is more stable when it is
substituted by a basic nitrogen and it is only in that case
that the 4-exo-dig process can be observed. In other words, for
the intermolecular electron transfer between 6 and a-lithio-
methoxyallene to take place, the centred nitrogen radical has
to be well stabilized by the terminal amino group. In other
cases, the intramolecular electron transfer giving 7 is
favoured to a large extent.
Scheme 11.

Â
V. Breuil-Desvergnes, J. Gore / Tetrahedron 57 (2001) 1951±1960
1958
3. Conclusion
4.1.3. N-(2)-(S)-2-Methoxymethylpyrrolidinyl-(S)-2-(p-
¯uorophenyl)-3-methoxy-2,5-dihydropyrrole (2a5).
20
The results presented here and in our previous report1 show
that the reaction of a-lithiomethoxyallene with hydrazones
constitutes an easy entry in the preparation of 5-membered
nitrogen heterocycles since the nucleophilic addition can be
followed by the cyclization of the lithium amide on the
allenic linkage. The main interest from a synthetic point
of view relies on its almost complete diastereoselectivity
when SAMP-hydrazones are used, allowing then the for-
mation of enantiopure 3-pyrrolines.
Yield: 69%, Rfˆ0.27 (95/5 EP/AcOEt), [a]D ˆ1112
(cˆ1.55, CHCl3). IR (®lm): nˆ3020, 2990±2910, 1660,
1
1610, 1450. H NMR (CDCl3, 300 MHz): dˆ1.25, 1.51±
1.61 (m, 2H), 1.64±1.69 (m, 2H), 2.64 (m, 1H), 2.78±2.93
(m, 2H), 3.21 (s, 3H), 3.23±3.37 (m, 2H), 3.51 (s, 3H),
3.68±3.95 (m, 2H), 4.59 (m, 1H), 4.95 (m, 1H), 6.95±7.01
(m, 2H), 7.34±7.39 (m, 2H). 13C NMR (50 MHz): dˆ21.07
(CH2), 26.2 (CH2), 44.25 (CH2), 54.7 (CH2), 56.6 (CH3),
59.5 (CH), 63.8 (CH), 74.4 (CH2), 90.3 (CH), 114.8 (CH),
115.1 (CH), 129.5 (CH), 129.6 (CH), 138.6 (C), 138.7 (C),
157.0 (C). EI MS m/z (%): 306 (8), 190 (14), 177 (32), 147
(20), 109 (12), 95 (6), 71 (100), 45 (44), 28 (12).
C17H23N2O2F (306.4): calculated C 66.6, H 7.6, N 9.1;
found C 66.82, H 7.66, N 9.22.
The study of the cyclization process with variations of the
substituents on the hydrazones has shown that a second
mode of cyclization is possible, leading to azetidines by a
reversible radical 4-exo-dig process. The evolution of the
reaction has been explained by intramolecular or inter-
molecular electron transfers, developing hydrazinyl radicals
which are relatively new reaction intermediates.
4.1.4. N-Morpholino-2-(p-methoxyphenyl)-3-methoxy-
2,5-dihydropyrrole (2b2). Yield: 74%, Rfˆ0.33 (85/15
EP/AcOEt). IR (®lm): nˆ3030, 2980±2820, 1660, 1450.
1H NMR (CDCl3, 300 MHz): dˆ2.69 (m, 4H), 3.51 (s,
3H), 3.67 (m, 4H), 3.79 (s, 3H), 4.56 (m, 1H), 4.87 (m,
1H), 6.84±6.88 (m, 2H), 7.30±7.34 (m, 2H). 13C NMR
(50 MHz): dˆ50.0 (CH2), 50.2 (CH2), 55.2 (CH3), 56.6
(CH3), 64.5 (CH), 67.3 (CH2), 89.7 (CH), 113.5 (CH),
129.1 (CH), 134.5 (C), 157.4 (C), 159.9 (C). EI MS; m/z
(%): 290 (32), 190 (100), 159 (57), 147 (10), 115 (10), 91
(13), 56 (17), 43 (34), 28 (21). HRMS EI: theoretical:
290.16304; measured: 290.16301.
4. Experimental
4.1. General procedure
See preceding paper.
4.1.1. Preparation of N-dialkylamino-3-pyrrolines 2.
Only compounds not already described in the preceding
paper1 are described herein.
4.1.5. N-Dimethylamino-2-p-methylphenyl-3-methoxy-
2,5-dihydropyrrole (2c3). Yield: 43%, Rfˆ0.29 (93/7 EP/
1
AcOEt). IR (®lm): nˆ3040, 2960±2780, 1660, 1450. H
a-Lithiomethoxyallene was synthetized in situ by addition
of an equivalent of n-butyllithium (2.5 M hexane solution)
to a solution of 5 mmol methoxyallene in THF (5 ml) at
2408C. The solution was chilled to 2788C. After 20 min,
a solution of hydrazone (0.833 mmol in 2 ml of THF) was
added dropwise to the a-lithiomethoxyallene solution. The
mixture was warmed to 2208C and, after 16 h at this
temperature, it was quenched with 5 ml of water and
extracted with ether (2£5 ml). The combined organic phases
were dried (Na2SO4) and concentrated. The crude product
was puri®ed by FC on neutral alumina gel (Merck 90 type
II±III: 0.063±0.20 mm) with a mixture of petroleum ether
EP/AcOEt as a solvent to give pure 2 and/or 4 as an oil. Its
purity was controlled by TLC on alumina using phospho-
molybdic acid as chemical tracer.
NMR (CDCl3, 300 MHz): dˆ2.34 (s, 3H), 2.36 (s, 6H),
3.53 (s, 3H), 3.76 (m, 2H), 4.55 (m, 1H), 4.78 (m, 1H),
7.15 (m, 2H), 7.30 (m, 2H). 13C NMR (50 MHz): dˆ22.6
(CH3), 40.8 (CH3), 48.2 (CH2), 60.4 (CH3), 65.5 (CH), 89.4
(CH), 127.9 (CH), 128.9 (CH), 136.7 (C), 139.4 (C), 157.5
(C). EI MS; m/z (%): 174 (39), 159 (81), 144 (28), 115 (27),
105 (12), 91 (37), 77 (19), 43 (100), 27 (20). HRMS EI:
theoretical: 232.15756; measured: 232.15760.
4.1.6. N-Dimethylamino-2-(p-methoxyphenyl)-3-methoxy-
2,5-dihydropyrrole (2c4). Yield: 35%, Rfˆ0.36 (85/15 EP/
1
AcOEt). IR (®lm): nˆ3030, 2970±2760, 1660, 1450. H
NMR (CDCl3, 300 MHz): dˆ2.38 (s, 6H), 3.52 (s, 3H),
3.78 (s, 3H), 4.54 (m, 1H), 4.77 (m, 1H), 6.86 (m, 2H),
7.35 (m, 2H). 13C NMR (50 MHz): dˆ40.8 (CH3), 47.9
(CH2), 55.2 (CH3), 56.6 (CH3), 65.3 (CH), 89.3 (CH),
113.66 (CH), 129.0 (CH), 134.5 (C), 157.5 (C), 158.9 (C).
EI MS m/z (%): 248 (100), 190 (68), 175 (37), 59 (40), 147
(24), 43 (20). HRMS EI: theoretical: 248.15247; measured:
248.15243.
4.1.2. N-(2)-(S)-2-Methoxymethylpyrrolidinyl-(S)-2-(p-
(2a4).
methoxyphenyl)-3-methoxy-2,5-dihydropyrrole
20
Yield: 75%, Rfˆ0.11 (97/3 EP/AcOEt), [a]D ˆ1132
(cˆ1.55, CHCl3). IR (®lm): nˆ3020, 2980±2810, 1660,
1
1610, 1460. H NMR (CDCl3, 300 MHz): dˆ1.21, 1.56
(2m, 2H), 1.66 (m, 2H), 2.65 (m, 1H), 2.88 (m, 2H), 3.21
(s, 3H), 3.26±3.34 (m, 2H), 3.52 (s, 3H), 3.71±3.89 (m, 2H),
3.78 (s, 3H), 4.59 (m, 1H), 4.93 (m, 1H), 6.82±6.86 (m, 2H),
7.30±7.34 (m, 2H). 13C NMR (50 MHz): dˆ21.08 (CH2),
26.04 (CH2), 43.81 (CH2), 55.0 (CH2), 56.5 (CH3), 59.0
(CH3), 59.6 (CH), 63.6 (CH), 74.4 (CH2), 90.1 (CH),
113.6 (2CH), 129.1 (2CH), 134.8 (C), 157.5 (C), 158.9
(C). EI MS m/z (%): 318 (100), 273 (84), 190 (22), 175
(15), 159 (49), 91 (6), 71 (43). C18H26N2O3 (318.4): calcu-
lated C 67.9, H 8.2, N 8.8; found C 68.37, H 8.23, N 8.14.
4.1.7. N-Piperidino-2-b-naphthyl-3-methoxy-2,5-dihydro-
pyrrole (2d2). Yield: 78%, Rfˆ0.48 (90/10 EP/AcOEt). IR
(®lm): nˆ3030, 2930±2820, 1660, 1450. 1H NMR (CDCl3,
300 MHz): dˆ1.30 (m, 2H), 1.55 (m, 4H), 2.68 (m, 4H),
3.54 (s, 3H), 4.63 (m, 1H), 5.18 (m, 1H), 7.45 (m, 2H), 7.66
(s, 1H), 7.84 (m, 4H). 13C NMR (50 MHz): dˆ24.4 (CH2),
26.4 (CH2), 51.0 (CH2), 56.6 (CH3), 64.6 (CH), 90.1 (CH),
125.4 (CH), 125.7 (CH), 126.4 (CH), 126.9 (CH), 127.6
(CH), 127.7 (CH), 128.0 (CH), 133.1 (C), 133.4 (C),

Â
V. Breuil-Desvergnes, J. Gore / Tetrahedron 57 (2001) 1951±1960
1959
140.9 (C), 157.3 (C). EI MS m/z (%): 308 (41), 209 (29), 179
(41), 165 (31), 127 (7), 98 (8), 70 (17), 42 (100). HRMS CI:
theoretical: 309.19663; measured: 309.19675.
4-methyl-azacyclobut-3-ene (4c4). Yield: 38%, Rfˆ0.42
(85/15 EP/AcOEt). IR (®lm): nˆ3290, 3020±2790, 1605,
1
1505, 1450. H NMR (CDCl3, 300 MHz): dˆ2.15 (s, 3H),
2.60 (s, 6H), 3.65 (s, 3H), 3.83 (s, 3H), 6.17 (s, 1H), 6.87 (m,
2H), 7.66 (m, 2H). 13C NMR (50 MHz): dˆ14.74 (CH3),
47.33 (CH3), 55.24 (CH3), 58.32 (CH3), 113.83 (CH),
116.67 (CH), 127.89 (C), 130.75 (CH), 153.20 (C), 158.88
(C), 159.25 (C). EI MS m/z (%): 248 (41), 233 (33), 178
(11), 164 (100), 148 (14), 121 (10), 91 (7), 77 (7), 44 (16).
HRMS EI: theoretical: 248.15247; measured: 248.15166.
4.1.8. N-Piperidino-2-(p-methylphenyl)-3-methoxy-2,5-
dihydropyrrole (2d3). Yield: 43%, Rfˆ0.58 (96/4 EP/
1
AcOEt). IR (®lm): nˆ3030, 2950±2760, 1660, 1460. H
NMR (CDCl3, 300 MHz): dˆ1.52 (m, 2H), 2.34 (s, 3H),
2.60 (m, 4H), 2.81 (m, 4H), 3.51 (s, 3H), 3.75±3.92 (m,
2H), 4.55 (m, 1H), 4.95 (m, 1H), 7.13 (m, 2H), 7.32 (m,
2H). 13C NMR (50 MHz): dˆ21.6 (CH3), 24.7 (CH2), 26.7
(CH2), 51.7 (CH2), 56.6 (CH2), 57.0 (CH3), 64.4 (CH), 90.2
(CH), 128.4 (CH), 129.0 (CH), 132.7 (C), 137.6 (C), 157.9
(C). EI MS m/z (%): 272 (25), 174 (9), 159 (24), 143 (11),
115 (10), 91 (13), 70 (23), 55 (48), 42 (100). HRMS EI:
theoretical: 272.18886; measured: 272.18875.
4.1.13. N-Piperidino-(2-p-methylphenyl)-3-methoxy-4-
methyl-azacyclobut-3-ene (4d3). Yield: 38%, Rfˆ0.64
1
(96/4 EP/AcOEt). H NMR (CDCl3, 300 MHz): dˆ1.30
(m, 2H), 1.58±1.72 (m, 2H), 2.16 (s, 3H), 2.33 (s, 3H),
3.69 (s, 3H), 6.19 (s, 1H), 7.13 (m, 2H), 7.60 (m, 2H). 13C
NMR (50 MHz): dˆ15.30 (CH3), 21.62 (CH3), 23.34
(CH2), 25.65 (CH2), 51.18 (CH2), 58.74 (CH3), 117.35
(CH), 128.98 (C), 129.14 (CH), 136.89 (C), 140.59 (C),
154.62 (C), 159.79 (C). EI MS m/z (%): 272 (8), 257 (5),
188 (100), 167 (5), 132 (13), 105 (8), 78 (10), 55 (13), 42
(26). HRMS EI: theoretical: 272.18886; measured:
272.18899.
4.1.9. N-Piperidino-2-(p-methoxyphenyl)-3-methoxy-2,5-
dihydropyrrole (2d4). Yield: 38%, Rfˆ0.39 (90/10 EP/
1
AcOEt). IR (®lm): nˆ3030, 2920±2730, 1660, 1460. H
NMR (CDCl3, 300 MHz): dˆ1.55 (m, 2H), 1.72 (m, 2H),
2.75 (t, 2H, Jˆ5.18 Hz), 3.55 (s, 3H), 3.78 (s, 3H), 3.82 (m,
2H), 4.56 (m, 1H), 4.92 (m, 1H), 6.87 (m, 2H), 7.32 (m, 2H).
13C NMR (50 MHz): dˆ23.9 (CH2), 25.4 (CH2), 51.2
(CH2), 55.2 (CH3), 56.6 (CH3), 56.8 (CH2), 63.8 (CH),
89.8 (CH), 113.8 (CH), 129.1 (CH), 135.3 (C), 157.4 (C),
158.8 (C). EI MS m/z (%): 288 (53), 190 (51), 159 (40), 91
(11), 70 (15), 55 (44), 42 (100). HRMS EI: theoretical:
288.18378; measured: 288.18384.
4.1.14. N-Piperidino-(2-p-methoxyphenyl)-3-methoxy-4-
methyl-azacyclobut-3-ene (4d4). Yield: 38%, Rfˆ0.44
1
(90/10 EP/AcOEt). H NMR (CDCl3, 300 MHz): dˆ1.32
(m, 2H), 1.55 (m, 2H), 2.17 (s, 3H), 2.61 (m, 2H), 3.68 (s,
3H), 3.88 (s, 3H), 6.18 (s, 1H), 6.88 (m, 2H), 7.69 (m, 2H).
13C NMR (50 MHz): dˆ14.84 (CH3), 24.4 (CH2), 26.38
(CH2), 50.84 (CH2), 55.23 (CH3), 58.31 (CH3), 113.82
(CH), 116.72 (CH), 127.96 (C), 130.75 (CH), 153.35 (C),
158.72 (C), 159.45 (C). EI MS m/z (%): 288 (3), 204 (100),
167 (8), 148 (8), 121 (11), 91 (11), 55 (9), 42 (18). HRMS
EI: theoretical: 288.18378; measured: 288.18384.
Compounds 4 obtained were mixed with compounds 2.
They were isolated when possible; in the other cases (4d3
and 4d4), they were fully characterized by use of 2D NMR:
COSY, HSQC and HMBC techniques from the mixture
4d312d3 and 4d412d4.
4.1.10. N-Morpholino-2-phenyl-3-methoxy-4-methyl-aza-
cyclobut-3-ene (4b1). Yield: 22%, Rfˆ0.24 (97/3 EP/
Acknowledgements
1
AcOEt). IR (®lm): nˆ3020, 2960±2810, 1610, 1440. H
The ef®cient help of Dr Philippe Compain at the beginning
Â
of the study was greatly appreciated. Valerie Breuil-
Desvergnes thanks MENSR for a scholarship.
NMR (CDCl3, 300 MHz): dˆ2.18 (s, 3H), 2.88 (t, 4H,
Jˆ4.41 Hz), 3.69 (s, 3H), 3.85 (t, 4H), 6.24 (s, 1H), 7.30
(m, 3H), 7.68 (m, 2H). 13C NMR (50 MHz): dˆ14.99
(CH3), 55.18 (CH2), 58.55 (CH3), 66.36 (CH2), 117.44
(CH), 127.52 (CH), 128.41 (CH), 134.92 (C), 157.4 (C),
154.48 (C), 160.59 (C). EI MS m/z (%): 260 (12), 245 (6),
201 (4), 188 (5), 176 (100), 118 (25), 90 (23), 77 (3), 56
(17), 28 (13). HRMS CI: theoretical: 261.160303;
measured: 261.161252.
References
Â
1. Breuil-Desvergnes, V.; Gore, J. Tetrahedron, 2001, 57, 1939±
1950.
2. Hoff, Th.; Brandsma, L.; Arens, J. F. Recl. Trav. Chim. Pays-
Bas 1968, 87, 917±924.
3. Baldwin, J. J. Chem. Soc., Chem. Commun. 1976, 736±740.
4.1.11. N-Dimethylamino-(2-p-methylphenyl)-3-methoxy-
4-methyl-azacyclobut-3-ene (4c3). Yield: 35%, Rfˆ0.39
(93/7 EP/AcOEt). IR (®lm): nˆ3030, 2970±2790, 1510,
Â
4. Arseniyadis, S.; Gore, J. Tetrahedron Lett. 1983, 24, 3997±
1
1470. H NMR (CDCl3, 300 MHz): dˆ2.15 (s, 3H), 2.34
4000. Arredondo, V.; McDonald, F.; Marks, T. J. Am. Chem.
Soc. 1998, 120, 4871±4872. Goldstein, S.; Overman, L.;
Rabinowitz, M. J. Org. Chem. 1992, 57, 1179±1190.
(s, 3H), 2.60 (s, 6H), 3.53 (s, 3H), 6.18 (s, 1H), 7.13 (m, 2H),
7.60 (m, 2H). 13C NMR (50 MHz): dˆ14.83 (CH3), 21.33
(CH3), 47.31 (CH3), 58.37 (CH3), 116.85 (CH), 129.11
(CH), 129.31 (CH), 129.32 (CH), 132.28 (C), 137.21 (C),
154.11 (C), 159.17 (C). EI MS m/z (%): 232 (3), 217 (4), 187
(2), 148 (100), 132 (6), 115 (4), 104 (7), 78 (9), 42 (12), 28
(6). HRMS EI: theoretical: 232.15756; measured:
232.15762.
5. Claesson, A.; Sahlberg, C.; Luthman, K. Acta Chem. Scand.
1979, 309±310. Grimaldi, J.; Cormons, A. Tetrahedron Lett.
1986, 27, 5089±5090. Jayaprakash, K.; Venkatachalam, C.;
Balasubramanian, K. Tetrahedron Lett. 1999, 40, 6493±6496.
6. Fox, D.; Gallagher, T. Tetrahedron 1990, 48, 4697±4710.
Meguro, M.; Yamamoto, Y. Tetrahedron Lett. 1998, 39,
5421±5424. Ohno, H.; Toda, A.; Miwa, Y.; Taga, T.;
Osawa, E.; Yamaoka, Y.; Futjii, N.; Ibuka, T. J. Org. Chem.
4.1.12. N-Dimethylamino-(2-p-methoxyphenyl)-3-methoxy-

Â
V. Breuil-Desvergnes, J. Gore / Tetrahedron 57 (2001) 1951±1960
1960
È
1999, 3, 324±326. Kang, S. K.; Baõk, T. G.; Kulak, A. Synlett
1999, 3, 324±326.
13. Ishibashi, H.; Higuchi, M.; Ohba, M.; Ikeda, M. Tetrahedron
Lett. 1998, 39, 75±78.
14. Crandall, J. K.; Tindell, G. L.; Mannade, A. Tetrahedron Lett.
1982, 23, 3769±3772.
7. Karstens, W.; Rutjes, F.; Hiemstra, H. Tetrahedron Lett. 1997,
38, 6275±6278.
15. Crandall, J. K.; Ayers, T. A. J. Org. Chem. 1992, 57, 2993±
2995.
8. Okala Amombo, M.; Hausherr, A.; Reissig, H.-U. Synlett
1999, 12, 1871±1874.
16. Motherwell, W. B.; Crich, D. Free Radical Chain Reactions in
Organic Synthesis; Academic: London, 1992 (p. 163).
 Á
17. P. Vogel, Chimie organique, Methodes et modeles, de Boeck
9. Magnus, P.; Albaugh-Robertson, P. J. Chem. Soc., Chem.
Commun. 1984, 804±806.
10. Hooz, J.; Calzada, J. G.; MacMaster, D. Tetrahedron Lett.
1985, 26, 271±274. Cabezas, J. A.; Alvarez, L. X.
Tetrahedron Lett. 1998, 39, 3935±3938.
Â
Universite, 1997, p. 317.
18. Kosuke, I. Acid±Base Dissociation Constants in Dipolar
Aprotic Solvents; Blackwell Scienti®c: Oxford, 1990. Bernas-
coni, C. F.; Panda, M. J. Org. Chem. 1987, 52, 3042±3050.
Cook, A. G. Enamines: Synthesis Structure and Reactions;
Dekker: New York, 1969 (p. 63).
È
11. Miyabe, H.; Shibata, R.; Sangawa, M.; Ushiro, C.; Naõto, T.
Tetrahedron 1998, 54, 11431±11444.
12. Crandall, J. K.; Tindell, G. L.; Mammade, A. Tetrahedron
Lett. 1982, 23, 3769±3772.