Synthesis and reactivity of enantiomerically pure N-alkyl-2-alkenyl azetidinium salts

Article abstract of DOI:10.1016/j.tetlet.2004.07.113

A synthesis of stereodefined enantiomerically pure 2-alkenyl azetidines is described using Wittig olefination as key step. The quaternary triflate ammonium salts of these heterocycles were prepared in a stereoselective way and treatment of these azetidini

Full text of DOI:10.1016/j.tetlet.2004.07.113

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 7525–7528
Synthesis and reactivity of enantiomerically pure N-alkyl-2-alkenyl
azetidinium salts
Franc¸ois Couty,* Franc¸ois Durrat, Gwilherm Evano and Damien Prim
´ ´
SIRCOB, UMR CNRS 8086, Universite de Versailles, 45, avenue des Etats-Unis, 78035, Versailles Cedex, France
Received 13 July 2004; accepted 24 July 2004
Abstract—A synthesis of stereodefined enantiomerically pure 2-alkenyl azetidines is described using Wittig olefination as key step.
The quaternary triflate ammonium salts of these heterocycles were prepared in a stereoselective way and treatment of these azetid-
inium salts with a base (KHMDS or PhLi) induced a regioselective Stevens rearrangement leading to a 3-alkenyl pyrrolidine. An
unprecedented SN20 reaction involving phenyllithium as nucleophile and an ammonium as leaving group was observed in one case.
Ó 2004 Elsevier Ltd. All rights reserved.
In the course of studies on the synthesis and reactivity of
non-racemic azetidines,1 we wish to report herein a con-
venient synthesis of stereodefined enantiomerically pure
N-alkyl-2-alkenyl azetidines of general structure 1,
which, to our knowledge, have not been reported to
date.2 Preliminary investigations on the reactivity of
these strained cyclic allylamines were aimed in order to
induce ring expansions1e,3 of these compounds that
could eventually be used as key steps for the stereoselec-
tive synthesis of larger nitrogen heterocycles.
were all prepared in good yields starting from the corre-
sponding 2-cyano azetidines, for which we recently
described an efficient synthesis starting from enantio-
merically pure b-amino alcohols.1a The key step for
the alkene formation involves a Wittig olefination per-
formed either on ketones 8–10 or on epimeric aldehydes
11–12 to afford 2-alkenyl azetidines 13–18.
The ketones were prepared by addition of an organo-
lithium reagent onto the nitrile moiety, as previously
described,1c whereas aldehydes 11 and 12 were produced
by Swern oxidations of the corresponding primary alco-
hols 6 and 7. Gratifyingly, these aldehydes were
produced without any detectable epimerization, though
a-amino aldehydes are known to have low configura-
tional stabilities.4 This can be attributed to the strain
of the four-membered ring, that disfavors the formation
of an enol.5 Unsaturated (E) ester 13 was stereoselec-
tively prepared in a one-pot (Swern+Wittig) sequence
from 7, and needed to be purified immediately after
This work describes a regioselective Stevens rearrange-
ment performed on the azetidinium triflate derivatives
2 of these heterocyles, leading to 3-alkenyl pyrrolidines
3, and delineates the suitable experimental conditions
and stereochemical requirements in 2 needed to induce
this reaction.
2-Alkenyl azetidines 13–19, showing different substitu-
tion pattern on the heterocycle and on the alkene moiety
Stevens rearrangement
+
N
N
N
3
2
1
TfO-
Keywords: Azetidines; Ring expansions.
*
Corresponding author. Tel.: +33 1 39254455; fax: +33 1 39254452; e-mail: couty@chimie.uvsq.fr
0040-4039/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.07.113

7526
F. Couty et al. / Tetrahedron Letters 45 (2004) 7525–7528
the reaction, since a partial isomerization of the alkene
into the (Z) isomer was observed with stored crude sam-
ples of 13. Finally, reduction of 13 with DIBAH gave
allylic alcohol 19. It should be mentioned however that
a hindered ketone (8: R = t-Bu) failed to give the corre-
sponding alkene upon a trial of Wittig olefination
(Scheme 1).
The results presented above deserve some comments.
First, the high stereoselectivity observed during the for-
mation of the trifluoromethanesulfonates starting from
trans-N-benzyl-2,3-substituted azetidines 14, 15, 18,
and 19 can be rationalized by considering the four differ-
ent conformers and invertomers A–D of the azetidine
ring depicted in Scheme 2. Among these, conformer C
appears to be the more stable since all the substituents
on the ring are in a pseudo-equatorial position. Further-
more, the lone pair of the nitrogen atom is not sterically
crowded for the subsequent alkylation. The selectivity
can therefore be explained by the favored alkylation of
this conformer. On the opposite, when 2,3-cis com-
pound 17 is considered, conformational analysis does
not clearly show that one among the four conformers
is more stable or reactive compared to the other ones.
In this case, the alkylation probably proceeds on several
conformers, which can explain the low diastereoselectiv-
ity observed for the alkylation of 17.
Having in hand a different set of alkenyl azetidines, the
ammonium triflate salts of these heterocycles were next
prepared upon reaction with methyltrifluoromethane-
sulfonate.6 In all cases, except for the 2,3-cis compound
17 leading to ammonium 23, this alkylation proceeded
in a highly diastereoselective way, and gave ammoniums
20–25 depicted in Scheme 1. In order to induce a Stevens
rearrangement in these compounds, diastereoisomeri-
cally pure N-benzyl triflates 20, 21, 24, and 25 were then
treated at ꢀ78°C with KHMDS or PhLi in THF and
gave pyrrolidines 26–29, in fair (unidentified minor by
products were produced starting from 20) to good
yields, resulting from a regioselective rearrangement.
These compounds were obtained as mixtures of isomers
at C-2, which could be separated by flash chromatogra-
phy. When the mixture of ammonium isomers 23 result-
ing from the reaction of 17 with MeOTf was treated with
KHMDS, a more complex mixture of at least four differ-
ent products was formed (Table 1).
The Stevens rearrangement of ammonium ylides has
been studied in much detail and was described once
involving azetidinium substrates.7 It is admitted that this
rearrangement occurs through diradical species in a sol-
vent cage, after formation of an intermediate ammo-
nium ylide. It is also known that configuration of the
carbon migrating center is retained.8 The outcome of
Ph
Ph
Ph
Ph
O
R
a
e
CN
+
N
Bn
TfO-
Bn
N
N
Bn
N
Bn
R
R
4
Me
8: R = Ph
9: R = Me
14: R = Ph
15: R = Me
20: R = Ph
21: R = Me
refs. 2a,c
Ph
Ph
O
Ph
Ph
Me
CN
a
Me
e
e
Me
Me
+
N
Me
N
Me
N
Me
TfO-
N
Ph
Ph
Ph
5
Me
Me
10
16
22
Ph
Ph
Ph
Ph
b
a
TfO-
+
N
O
N
Bn
OH
N
N
Bn
6
Me Bn
Bn
Ph
11
17
23
(de = 12%)
Ph
Ph
Ph
b
a
e
e
TfO-
Me
+
N
O
N
N
Bn
OH
N
Bn
Bn
Bn
7
Bn
12
18
24
Ph
Ph
Ph
c
d
TfO-
Me
+
CO2Et
N
N
N
OH
OH
13
19
Bn
Bn
25
Scheme 1. Reagents and conditions: (a) trimethylphosphonium iodide, BuLi, ꢀ78°C, THF, 79% (14), 62% (15), 86% (16), 62% from (6), (17), 67%
(18); (b) Swern oxidation, 93% (12); (c) Swern oxidation, then (C6H5)3PCH@CHCO2Et, 84%; (d) DIBAH, THF, ꢀ78°C, quant.; (e) MeOTf, DCM,
0°C, 93% (20), 83% (21), 94% (22), quant. (23), 98% (24), 95% (25).

F. Couty et al. / Tetrahedron Letters 45 (2004) 7525–7528
Table 1. Reaction of 2-alkenyl azetidinium triflates in basic medium
7527
R'
R
Ph
KHMDS or PhLi
Ph
20, 21, 24-25
N
26-30
Bn
Substrate
Conditions
Product
Yield (%)
de (%)
20
21
24
25
KHMDS
KHMDS
KHMDS
PhLi
26 (R = Ph, R0 = H)
27 (R = Me, R0 = H)
28 (R = H, R0 = H)
39
94
83
69
8
72
6
29 (R = Ph, R0 = CH2OH)
8
Ph
Ph
Ph
Ph
Ph
Ph
Bn
PhLi
80%
H
N
Ph
Ph
N
TfO-
+
N
Bn
N
Ph
R1
R1
Me
R2
20
30
B
D
A
R2
Scheme 4. Unexpected ring opening occurs with compound 20.
Ph
Ph
Bn
R2
N
R2
N
R1
R1
C
Bn
This reaction was unexpected since nitrogen-derived
moieties have scarcely been used as leaving groups in
SN20 reactions involving organometallic reagent and
are restricted to particular cases in which the leaving
group ability of the nitrogen atom is increased by strain
release such as in 2-vinyl-9a or 2-ethynyl-9b N-tosyl azir-
idines. This is the case here, with the strain of the azet-
idine ring, but to our knowledge, ammonium acting as
leaving group in SN20 processes involving organometal-
lic nucleophiles has no precedent.10 Interestingly, alkene
30 is produced as a single isomer, whose configuration
was not determined. It should be noted that this reaction
occurred only with ammonium 20, in which the double
bound is substituted by a phenyl group. This can give
some insights into the mechanism of this reaction that
might create a transient partial negative charge on the
alkenyl carbon, which is stabilized in case of substrate
20 by the aromatic ring.
Scheme 2. Conformational analysis for 2,3-trans azetidines.
the reaction described herein fits well with these points:
the high chemoselectivity of deprotonation at the benz-
ylic center and the total regioselectivity of the migration
involving the carbon bearing an alkenyl moiety accounts
for the production of two isomeric pyrrolidines at C-2.
This low diastereoselectivity might be the result of a
non-selective deprotonation (formation of Z and E
intermediate ylides) at the methylene carbon of the
benzyl group (Scheme 3).
Altogether, the reaction occurs very rapidly at low tem-
perature, is unsensitive to the substitution pattern of the
alkene present in the substrate and gives no b-elimina-
tion product. The nature of the employed base is, how-
ever, important: while KHMDS worked well in all cases,
phenyllithium, frequently used as a base in similar reac-
tions,7 gave erratic results: it afforded the rearrangement
product in case of 25, but extensive degradation was
observed with substrates 21 and 24 and a surprising
outcome was obtained with 20. In this last case,
SN20 addition of the organolithium reagent occurred
on the alkene, resulting in an opening of the azetidine
to give 30 (Scheme 4).
The relative stereochemistries of the ammonium salts
20–21 and 24–25, that is a trans relationship between
the N-benzyl substituent and the adjacent alkenyl moi-
ety, were determined by NMR using a NOESY experi-
ment performed with 21. Indeed, 20, 24, and 25
display the same 2,3-trans substitution pattern on the
azetidine ring as in 21, and the alkylation step producing
these compounds is expected to proceed with the same
stereoselectivity. It should be noted that because of the
R2
R1
Ph
N
Me 1
Ph
N
Me 1
Ph
Ph
N
Me 1
Ph
Ph
Ph
Ph
Ph
Base
Ph
N
Me 1
N
R
R2
R
R2
R
R2
R
R2
Me
26-29
Solvent Cage
Scheme 3. Mechanism of the Stevens rearrangement of azetidines 20, 21, 24, 25.

7528
F. Couty et al. / Tetrahedron Letters 45 (2004) 7525–7528
Wolf, H.; Karstens, W. F. J.; Schoemaker, H. E.;
+
X
Hiemstra, H. Org. Lett. 1999, 1, 717–720; (b) Anzai, M.;
Toda, A.; Ohno, H.; Takemoto, Y.; Fujii, N.; Ibuka, T.
Tetrahedron Lett. 1999, 40, 7393–7397; (c) Ohno, H.;
Anzai, M.; Toda, A.; Ohishi, S.; Fujii, N.; Tanaka, T.;
Takemoto, Y.; Ibuka, T. J. Org. Chem. 2001, 66,
4904–4914.
N
N
-
Ph
Ph
Figure 1.
3. Hesse, M. Ring Enlargements in Organic Chemistry; VCH:
Weinheim, New York, Basel, Cambridge, 1991.
4. Reetz, M. T.; Drewes, M. W.; Schmitz, A. Angew. Chem.,
Int. Ed. Engl. 1987, 26, 1141–1143.
5. A similar observation involving aziridine 2-carboxaldehy-
des was recently reported, see: Lee, W. K.; Ha, H.-Y.
Aldrichim. Acta 2003, 36, 57–63.
low conformational mobility brought by the four-mem-
bered ring, this trans disposition between the ylide and
the alkene precludes the possibility of a2,3 sigmatropic
shift (Fig. 1) that would produce a tetrahydroazepine.11
The more complex mixture obtained with ammoniums
23, in which the ylide is in part cis to the alkene, suggest
that this process might be operative and highlight the
importance of the relative configuration of the substitu-
ents in the starting azetidine for the success of the
Stevens rearrangement.
6. Iodomethane was ineffective in this reaction probably due
to ring opening of the ammonium by the halide, see:
´
´ `
Concellon, J. M.; Bernard, P. L.; Perez-Andres, J. A.
Tetrahedron Lett. 2000, 41, 1231–1234.
7. Anderson, A. G.; Wills, M. T. J. Org. Chem. 1968, 33,
3046–3050; (b) Wills, M. T.; Wills, I. E.; Von Dollen, L.;
Butler, B. L.; Porter, J. J. Org. Chem. 1980, 45, 2489–2498;
For recent exemples on the use of the Stevens rearrange-
ment in synthesis, see inter alia: (c) Saba, A. Tetrahedron
Lett. 2003, 44, 2895–2898; (d) Vanecko, J. A.; West, F. G.
Org. Lett. 2002, 4, 2813–2816; (e) Padwa, A.; Beall, L. S.;
Eidell, C. K.; Worsencroft, K. J. J. Org. Chem. 2001, 66,
2414–2421.
In summary, we have decribed herein a synthesis of
enantiomerically pure 2-alkenyl azetidines and a regiose-
lective Stevens rearrangement of their ammonium deriv-
atives. Further work is in progress in our group in order
to delineate the scope of these reactions.
8. (a) Brewster, J. H.; Kline, M. W. J. Am. Chem. Soc. 1952,
74, 5179–5182; (b) Scho¨llkopf, U.; Ludwig, U.; Oster-
mann, G.; Patsch, M. Tetrahedron Lett. 1969, 39,
3415–3418.
9. (a) Aoyama, H.; Mimura, N.; Ono, H.; Ishii, K.; Toda, A.;
Tanamura, H.; Otaka, A.; Fujii, N.; Ibuka, T. Tetrahedron
Lett. 1997, 38, 7383–7386; (b) Ohno, H.; Toda, A.; Fujii,
N.; Takemoto, Y.; Tanaka, T.; Ibuka, T. Tetrahedron
1997, 56, 2811–2820.
Acknowledgements
CNRS is acknowledged for generous support (soutien
´
jeune equipe).
References and notes
10. For an isolated example of reductive opening of boron–
amine complexes through a SN20 pathway see: Chabouni,
R.; Laurent, A.; Marquet, B. Tetrahedron Lett. 1976, 10,
757–758; For a discussion on the mechanism of the
intramolecular reaction involving an alkoxide as nucleo-
phile and an ammonium as leaving group, see: Ollis, W.
D.; Sutherland, I. O.; Thebtaranonth, Y. J. Chem. Soc.,
Perkin Trans 1, 1981, 1981–1993.
11. For examples of such rearrangement involving five- and
six-membered nitrogen heterocycles see (a) Vedejs, E.;
Hagen, J. P.; Roach, B. L.; Spear, K. L. J. Org. Chem.
1978, 43, 1185–1190; (b) Vedejs, E.; Arco, M.; Powell, D.
W.; Renga, J. M.; Singer, S. P. J. Org. Chem. 1978, 43,
4831–4837.
1. (a) Agami, C.; Couty, F.; Evano, G. Tetrahedron: Asym-
metry 2002, 13, 297–302; (b) Agami, C.; Couty, F.;
Rabasso, N. Tetrahedron Lett. 2002, 43, 4633–4636; (c)
Couty, F.; Prim, D. Tetrahedron: Asymmetry 2002, 13,
2619–2624; (d) Carlin-Sinclair, A.; Couty, F.; Rabasso, N.
Synlett 2003, 726–729; (e) Couty, F.; Durrat, F.; Prim, D.
Tetrahedron Lett. 2003, 44, 5209–5213; (f) Couty, F.;
Evano, G.; Rabasso, N. Tetrahedron: Asymmetry 2003,
14, 2407–2412; (g) Couty, F.; Evano, G.; Prim, D. Mini-
Rev. Org. Chem. 2004, 1, 133–148.
2. A synthesis of N-tosyl-2-alkenyl azetidines has been
recently reported almost simultaneously by Hiemstra
and Ibuka: (a) Rutjes, F. P. J. T.; Tjen, K. C. M. F.;