Regioselective nucleophilic opening of azetidinium ions

Article abstract of DOI:10.1055/s-2005-871537

Azetidinium ions bearing different substitution patterns were reacted with nitrogen nucleophiles (sodium azide and benzylamine) and oxygenated nucleophiles (sodium acetate and alkoxides). High regioselectivity of nucleophilic opening was observed in both

Full text of DOI:10.1055/s-2005-871537

1666
LETTER
Regioselective Nucleophilic Opening of Azetidinium Ions
H
ighly Regios
r
elective
a
Opening of
n
A
zetidinium
ç
T
riflates ois Couty,* François Durrat, Gwilherm Evano
SIRCOB, UMR 8086, Université de Versailles, 45, Avenue des Etats-Unis, 78035, Versailles Cedex, France
E-mail: couty@chimie.uvsq.fr
Received 14 April 2005
We recently reported a very efficient access to enantio-
merically pure functionalized azetidines, using b-amino
alcohols as starting materials.7 This allowed us to prepare
a range of azetidines bearing different substitution pat-
Abstract: Azetidinium ions bearing different substitution patterns
were reacted with nitrogen nucleophiles (sodium azide and benzyl-
amine) and oxygenated nucleophiles (sodium acetate and alkox-
ides). High regioselectivity of nucleophilic opening was observed in
both cases: the nucleophile reacting on the unsubstituted carbon in terns and permitted us to undertake different studies of
their reactivity.8 We wish to present in this Letter our pre-
liminary results concerning the nucleophilic opening of a
series of azetidinium ions which demonstrates that high
regioselectivity can be attained in such processes, there-
the case of a-substituted azetidinium salts and on the carbon bearing
an ester or cyano moiety in the case of a,a¢-substituted azetidinium
salts.
Key words: azetidiniums, nitrogen, heterocycles, ring opening,
asymmetric synthesis
fore considerably enhancing their synthetic usefulness.
The azetidinium triflate salts 1–10 necessary to conduct
this study were prepared by alkylation of the correspond-
The reactivity of small cycles towards nucleophilic open-
ing azetidine with methyltrifluoromethanesulfonate
ing is a cornerstone of organic synthesis. Although the
(Figure 2). The yield of this reaction was uniformly high
ring-opening of aziridines1 and aziridinium ions2 are well
and this alkylation proceeded with high stereoselectivity
established synthetic tools, this reaction with their higher
in the case of N-benzyl azetidines.8c The azetidine precur-
homologues azetidines and azetidinium ions has been un-
sors are derived from either (1R,2S)-ephedrine or (R)-phe-
der-utilized. In this area, after the pioneering work of
nyl glycinol and have been described earlier,7,8 expect the
Gaertner3 and Leonard4 in the early sixties showing that
precursor of salt 10. These azetidinium ions can be divid-
these strained ammonium ions can react with a range of
ed into two classes: a-substituted compounds, derived
nucleophiles leading to ring opening, Illuminati et al.5
from (R)-phenylglycinol (1–5) and bearing different moi-
clearly brought into light through an elegant study the par-
eties at C-2, and a,a¢-disubstituted azetidiniums (6–10),
ticular reactivity of these cyclic ammoniums compared to
bearing different moieties at C-2, and a methyl substituent
their higher homologues: in these ions, the ring strain
at C-4. The latter compounds are derived from (1R,2S)-
strongly favors the substitution reaction versus the classi-
ephedrine, or (1R,2S)-norephedrine (for 10).
cal Hofmann elimination. Notwithstanding, the use in
With a set of enantiomerically pure azetidinium salts
showing various a,a¢-substitution patterns in hand, we
first examined the reactivity of these compounds towards
nitrogen nucleophiles such as sodium azide and benzyl-
amine. Thus, compounds 1, 2, 4–8 were reacted in DMF
at room temperature with an excess of sodium azide (5
equiv) and compounds 1 and 6 were reacted with benzyl-
amine. The structure of the obtained adducts is detailed
hereafter (Table 1).
synthesis of these original building blocks has not been
developed and only few and isolated examples report the
use of these ammonium ions in synthesis.6 This is most
probably due to the lack of easy access to these heterocy-
cles, particularly in enantiomerically pure form. In addi-
tion, issues with regioselectivity of the nucleophilic
opening of a-substituted or a,a¢-disubstituted azetidinium
ions have seldom been addressed (Figure 1): scarce exam-
ples report the regioselective opening with a carboxy-
late,6m an amine6f or a phenol6d at the unsubstituted carbon
(R1 = H, path a), or the regioselective opening at a benzyl-
ic position (R1 = H, R3 = Ph, path b),6d but no systematic
study has been conducted so far.
Important trends concerning the opening of these aziridin-
ium ions can be outlined through these experiments. First,
these reactions occur in a stereospecific way through an
exclusive SN2 pathway, as shown by the diastereoisomer-
ic homogeneity of the ring-opened products. The yields
are excellent in most cases and no competitive elimination
was detected as a side reaction. The regioselectivity of the
opening is high and strongly depends on the substitution
pattern of the ring: while attack of the nucleophile occurs
on the unsubstituted side in the case of phenyl glycinol de-
rived azetidinium ions (entries 1, 2, 7), this regioselectiv-
ity is reversed in the case of disubstituted substrates
derived from ephedrine. In those cases, the nucleophile at-
tacks the carbon bearing the CN or CO2Et moiety (entries
4–6, 8). The substrate 5, bearing an alkene substituent, is
R2
R1
R3
+
N
R5
b
R4
a
Nu:
Figure 1
SYNLETT 2005, No. 11, pp 1666–1670
0
7
.0
7
.2
0
0
5
Advanced online publication: 14.06.2005
DOI: 10.1055/s-2005-871537; Art ID: D08505ST
© Georg Thieme Verlag Stuttgart · New York

LETTER
Regioselective Nucleophilic Opening of Azetidinium Ions
1667
Ph
N
Ph
N
Ph
Ph
Ph
OH
Ph
CO2Et
CN
N
Ph
N
CO2Et
N
Me
Bn
Me
Bn
Me
Me
Bn
Bn
Me
Bn
1 (90%)
2 (99%)
3 (94%)
4 (96%)
5 (100%)
Ph
Ph
Ph
Ph
Ph
Me
Me
CO2Et
Me
CN Me
CN
Me
CN
OH
N
N
N
N
N
OH
Me
Me
Me
Me
Me
Me
Me
Me
Me
10 (86%)
9 (89%)
6 (90%)
7 (90%)
8 (89%)
Figure 2
a special case. With this compound, exclusive SN2¢ reac- We studied next the reaction of oxygen nucleophiles, and
tion was observed, and yielded the trisubstituted alkene 15 sodium acetate as well as benzyl alkoxide were selected
as a mixture of isomers. This reaction gave a more com- for this study as external nucleophiles. We also experi-
plex mixture in case of salt 4.
mented with an alkoxide as an internal nucleophile, start-
ing from azetidinium ions 3, 9 and 10. The results are
summarized in Table 2.
Regioselectivity of these openings was confirmed through
chemical transformation (reduction of the azido group) on
11, 17 and 19. Upon reduction of 11, lactam 26 was ob- The same trend of regioselectivity is observed: the acetate
tained after spontaneous lactamization and N-debenzyl- anion exclusively reacts at the unsubstituted carbon (entry
ation. In contrast, the same reaction carried out on 17 and 1), with a regioselectivity even higher than in the previous
19 yield a-aminoester or a-aminonitrile, respectively, the case involving an azide anion (compare entry 2 in Table 1
latter being protected as a carbamate (Scheme 1). Careful and entry 1 in Table 2). In the case of a disubstituted sub-
analysis of 1H NMR spectra (relevant chemical shifts are strate, the acetate anion regioselectively attacks the car-
depicted in Scheme 1) unambiguously allowed for identi- bon bearing the ester, but, with this less reactive
fication of the regioisomer formed during these reactions. nucleophile, the yield is lower than with an azide anion
and we had to switch to the more nucleophilic cesium salt
(entries 3, 4) to induce the opening. However, even under
these conditions, the yield of different trials was not con-
Me
sistent. Benzyl alkoxide was inactive in an intermolecular
Ph
H
Ph
N
H2, Pd/C cat.
79%
reaction (entry 2), but alkoxides generated with the aim of
inducing intramolecular ring-opening proceeded well and
gave the corresponding unstable epoxides 33 and 34
(entries 5, 6) if the alkoxide was generated from an hy-
droxymethyl substituent at C-2. The degradation observed
in entry 7, involving an hydroxylated side chain branched
on the nitrogen of the azetidine ring can be explained by
the requirement of a difficult 5-endo-tet process for the
intramolecular ring opening.9
N3
CO2Et
Bn
O
N
H
N
Me
26
11
Ph
Ph
H2, Pd/C cat.
84%
Me2N
CO2Et
NH2
Me2N
CO2Et
Me
Me
N3
27
Ph
17
Although the general trends are well delineated by this set
of experiments, i.e., regioselective attack on the unsubsti-
tuted site and regioselective attack on the site substituted
by an electron-withdrawing group in case of a,a¢-disubsti-
tuted salts, more subtle influences on the efficiency of this
reaction need further comments. This is the case with en-
tries 5 and 6 in Table 1, highlighting a strong influence of
the relative stereochemistry of the substituents on the sub-
strate. The very efficient reaction of the 2,3-cis azetidini-
um salt 8, compared to the 2,3-trans isomer 7, giving the
corresponding opened product 19 in only 27% yield in a
reproducible way, can be explained by considering the
Ph
H2, Pd/C cat.
Me2N
Ha
CN
NHR
Me2N
CN
Hb
Me
Me
19
N3
28a: R = H (δ Ha: 3.29, δ Hb: 4.18)
28b: R = Boc (δ Ha: 3.33, δ Hb: 4.75)
Boc2O, 61% overall
Scheme 1
Synlett 2005, No. 11, 1666–1670 © Thieme Stuttgart · New York

1668
F. Couty et al.
LETTER
Table 1 Reaction of Azetidinium Ions with Nitrogen Nucleophiles
Entry
1
Substrate
Conditions
Products
Yield (%)a
Quant.
C-4 Attack
C-2 Attack
1
NaN3 (5 equiv), DMF
Ph
Ph
Me
N3
CO2Et
Bn
N
CO2Et
+
93:7
Bn
N
N3
Me
Ph
12
11
2
2
NaN3 (5 equiv), DMF
Quant.
Ph
Me
N3
CN
Bn
N
CH2NH2
+
Bn
82:12
N
N3
Me
14
13
3
4
5
6
NaN3 (5 equiv), DMF
NaN3 (5 equiv), DMF
88
Me
Ph
Ph
N
N3
Bn
15 6:4 isomeric mixture
Quant.
Ph
Ph
Me
N
CO2Et
N3
CO2Et
NMe2
+
Me
14:86
Me
N3
Me
Me
Me
17
16
5
6
7
7
8
2
NaN3 (5 equiv), DMF
NaN3 (5 equiv), DMF
27
Ph
Ph
Ph
N3
CN
NMe2
Me2N
CN
CN
CN
+
>2:98
Me
N3
18
19
Quant.
Ph
Me
N3
CN
NMe2
+
N
Me
>2:98
Me
N3
20
21
Benzylamine (5 equiv),
CH2Cl2
89
Ph
Ph
Me
BnHN
CN
Bn
N
+
Bn
>98:2
N
NHBn
Me
Ph
23
22
8
6
Benzylamine (5 equiv),
CH2Cl2
68
Ph
Me
BnHN
CO2Et
N
CONHBn
NHBn
+
Me
>2:98
Me
NMe2
Me
24
25
a Yield of pure isolated product.
conformation of the involved substrate. Figure 3 illus- the cyano moiety. In this case, the kinetics of the reaction
trates the conformational equilibrium in diasteroisomeric should be less hampered by an unfavorable pre-conforma-
salts 7 and 8. In compound 7, the equilibrium should tional equilibrium.
strongly favor conformer 7a, in which all substituents are
in pseudo-equatorial positions. However, this conformer
appears to be less reactive towards nucleophilic opening,
In conclusion, we have shown that high regioselectivities
of nucleophilic ring-opening can be obtained with these
original and readily accessible electrophilic synthons.10
due to steric shielding of the aromatic ring in the trajectory
Generalization with other nucleophiles and synthetic ap-
of the nucleophile. In this case, it appears that the strongly
plications of this methodology are under investigation.
disfavored conformer 7b, presenting severe 1,3-diaxial
interactions between the substituents, is the reactive one.
Acknowledgment
If the diastereomeric salt 8 is now considered, conformer
8b, which is also the reactive conformer for the same rea-
sons, is now more stable due to the equatorial position of
The CNRS is acknowledged for generous financial support (ATIP
jeune équipe).
Synlett 2005, No. 11, 1666–1670 © Thieme Stuttgart · New York

LETTER
Regioselective Nucleophilic Opening of Azetidinium Ions
1669
Table 2 Reaction of Azetidinium Ions with Oxygenated Nucleophiles
Entry
1
Substrate
Conditions
Products
Yield (%)a
Quant.
C-4 Attack
C-2 Attack
Ph
2
AcONa (5 equiv), DMF
Me
N
Ph
AcO
CN
Bn
CN
OAc
+
Bn
>98:2
N
Me
30
29
2
3
4
2
6
6
PhCH2ONa (5 equiv), DMF
AcONa (5 equiv), DMF
AcOCs (5 equiv), DMF
–
–
Decomposition
Decomposition
21–67
Ph
Ph
AcO
CO2Et
Me2N
CO2Et
OAc
+
>2:98
Me
NMe2
Me
32
31
b
5
6
3
9
KHMDS (1.2 equiv),
THF, –78 °C to r.t.
–
Me
N
Ph
Ph
Bn
O
33
b
KHMDS (1.2 equiv),
THF, –78 °C to r.t.
–
Ph
Me2N
O
Me
34
7
10
KHMDS (1.2 equiv),
THF, –78 °C to r.t.
Decomposition
–
a Yield of pure isolated product.
b These epoxides were not purified since they were unstable on silica gel. Exclusive formation of the epoxide was observed in the crude reaction
mixture.
–
3555; and references cited therein. (c) Rowlands, G. J.;
Barnes, W. K. Tetrahedron Lett. 2004, 45, 5347.
(d) Periasamy, M.; Seenivasaperumal, M.; Dharma Rao, V.
Tetrahedron: Asymmetry 2004, 15, 3847.
unshielded attack of N3
CN
Me
Me
CN
+
N
Me
Me
Me
N+
(3) (a) Gaertner, V. R. Tetrahedron Lett. 1967, 4, 343.
(b) Gaertner, V. R. J. Org. Chem. 1968, 33, 523.
(4) Leonard, N. J.; Durand, D. A. J. Org. Chem. 1968, 33, 1322.
(5) Cospito, G.; Illuminati, G.; Lilloci, C.; Petride, H. J. Org.
Chem. 1981, 46, 2944.
–
shielded attack of N3
Me
7b
7a
(6) (a) O’Brien, P.; Phillips, D. W.; Towers, T. D. Tetrahedron
Lett. 2002, 43, 7333. (b) Concellón, J. M.; Riego, E.;
Bernard, P. L. Org. Lett. 2002, 4, 1303. (c) Concellón, J.
M.; Bernard, P. L.; Pérez-Andrés, J. A. Tetrahedron Lett.
2000, 41, 1231. (d) Higgins, R. H.; Faircloth, W. J.;
Baughman, R. G.; Eaton, Q. L. J. Org. Chem. 1994, 59,
2172. (e) Guidicelli, M.-B.; Picq, D.; Anker, D. Tetrahedron
Lett. 1992, 48, 6033. (f) Kane, M. P.; Szmuszkovicz, J. J.
Org. Chem. 1981, 46, 3728. (g) Heller, M.; Bernstein, S. J.
Org. Chem. 1971, 36, 1386. (h) Jeziorna, A.; Heliꢀski, J.;
Krawiecka, B. Synthesis 2003, 288. (i) Jeziorna-Bakalarz,
A.; Heliꢀski, J.; Krawiecka, B. J. Chem. Soc., Perkin Trans.
1 2001, 1086. (j) Heliꢀski, J.; Skrzypczynski, Z.; Michalski,
J. Tetrahedron Lett. 1995, 36, 9201. (k) Jeziorna, A.;
Heliꢀski, J.; Krawiecka, B. Tetrahedron Lett. 2003, 44,
3239. (l) Bakalarz, A.; Heliꢀski, J.; Krawiecka, B.;
Michalski, J.; Potrzebowski, M. J. Tetrahedron 1999, 55,
12211. (m) Sudo, A.; Iitaka, Y.; Endo, T. J. Polym. Sci., Part
A: Polym. Chem. 2002, 40, 1912.
–
unshielded attack of N3
Me
N+
Me
CN
Me
Me
Me
N+
Me
CN
–
shielded attack of N3
8b
8a
Figure 3
References
(1) Hu, E. X. Tetrahedron 2004, 60, 2701.
(2) For recent examples, see: (a) Mulvihill, M. J.; Cesario, C.;
Smith, V.; Beck, P.; Nigro, A. J. Org. Chem. 2004, 69,
5124. (b) Chuang, T.-H.; Sharpless, K. B. Org. Lett. 2000, 2,
Synlett 2005, No. 11, 1666–1670 © Thieme Stuttgart · New York

1670
F. Couty et al.
LETTER
(7) (a) Agami, C.; Couty, F.; Evano, G. Tetrahedron:
Asymmetry 2002, 13, 297. (b) Carlin-Sinclair, A.; Couty, F.;
Rabasso, N. Synlett 2003, 726. (c) Agami, C.; Couty, F.;
Rabasso, N. Tetrahedron Lett. 2002, 43, 4633.
(8) (a) Couty, F.; Durrat, F.; Prim, D. Tetrahedron Lett. 2003,
44, 5209. (b) Couty, F.; Evano, G.; Prim, D.; Marrot, J. Eur.
J. Org. Chem. 2004, 3893. (c) Couty, F.; Durrat, F.; Evano,
G.; Prim, D. Tetrahedron Lett. 2004, 45, 7525.
(Cipso Ph), 133.1, 131.9, 130.5, 129.9, 129.8, 128.4 (CH Ph),
128.3 (Cipso Ph), 112.3 (CN), 69.3, 69.2 (C-4, C-6), 65.3 (C-
2), 46.9 (C-5), 39.6 (C-3) ppm. Anal. Calcd for
C19H19F3N2O3S: C, 55.33; H, 4.64; N, 6.79. Found: C, 55.23;
H, 4.66; N, 6.74.
Typical procedure for the reaction of an azetidinium salt
with sodium azide is given below:
To a solution of azetidinium triflate (2.0 mmol) in DMF (10
mL) was added sodium azide (650 mg, 10.0 mmol) and the
suspension was stirred overnight at r.t. Partition between
Et2O and H2O was followed by usual work-up. The crude
residue was examined by NMR, and the major compound
was purified by flash chromatography.
(9) Baldwin, J. E.; Thomas, R. C.; Kruse, L. I.; Silberman, L. J.
Org. Chem. 1977, 42, 3846.
(10) All new compounds were characterized by 1H NMR, 13
C
NMR spectroscopy, mass spectra analysis, and for most
relevant compounds, by elementary analysis. Typical
procedure for the preparation of an azetidinium salt is given
below:
To a solution of the azetidine (2.0 mmol) in CH2Cl2 (10 mL),
cooled at 0 °C, was added dropwise methyltrifluoro-
methanesulfonate (0.45 mL, 4 mmol). The reaction mixture
was stirred for 1 h at r.t., and the solvent was evaporated
under reduced pressure. The crude salt was washed
thoroughly with small quantities of dry Et2O, and dried
under vacuum.
Compound 11: purified by flash chromatography (Et2O–
cyclohexane, 10:90, 20:80, 40:60); yield 93%; clear oil;
Rf = 0.61 (Et2O–petroleum ether, 3:7); [a]D25 –90 (c 0.75,
CH2Cl2). MS (IC NH3 Pos): m/z = 353 [M + H]+. 1H NMR
[300 MHz, (CD3)2CO]: d = 7.38–7.19 (m, 10 H, Ph), 4.67 (d,
J = 4.2 Hz, 1 H, H-2), 4.17 (q, J = 7.1 Hz, 2 H, H-7), 3.66 (d,
J = 13.1 Hz, 1 H, H-6), 3.54–3.47 (m, 2 H, H-3, H-6¢), 3.01
(t, J = 11.9 Hz, 1 H, H-4), 2.47 (dd, J = 4.2, 12.3 Hz, 1 H, H-
4¢), 2.28 (s, 3 H, H-5), 1.22 (t, J = 7.1 Hz, 3 H, H-8) ppm. 13
C
Compound 2: yield 99%; [a]D25 –34 (c 0.5, acetone); mp
116 °C. MS (ESI Pos): m/z = 263.2 [M – OTf]+. 1H NMR
[300 MHz, (CD3)2CO]: d = 7.89–7.82 (m, 2 H, Ph), 7.67–
7.42 (m, 8 H, Ph), 6.34 (d, J = 9.3 Hz, 1 H, H-2), 5.31–5.14
(m, 4 H, H-3, H-4, H-6, H-6¢), 4.91–4.81 (m, 1 H, H-4¢), 3.70
(s, 3 H, H-5) ppm. 13C NMR [75 MHz, (CD3)2CO]: d = 133.9
NMR [75 MHz, (CD3)2CO]: d = 170.2 (C-1), 138.9, 138.1
(Cipso Ph), 129.2, 128.9, 128.5, 128.4, 127.6, 127.3 (CH Ph),
64.3 (C-2), 62.9 (C-6), 61.7 (C-7), 59.3 (C-4), 45.2 (C-3),
42.5 (C-5), 14.2 (C-8) ppm. Anal. Calcd for C20H24N4O2: C,
68.16; H, 6.86; N, 15.90. Found: C, 68.03; H, 6.91; N, 15.80.
Synlett 2005, No. 11, 1666–1670 © Thieme Stuttgart · New York