Ring opening of 2-phenylazetidines with allylsilanes

Article abstract of DOI:10.1016/j.tetlet.2006.01.108

N-Activated 2-phenylazetidines were opened regioselectively at the benzylic carbon with various allylsilanes or propargylsilane in the presence of BF ₃?Et₂O, providing amino olefins, precursors of biomolecules such as phenyl-homo-kai

Full text of DOI:10.1016/j.tetlet.2006.01.108

Tetrahedron Letters 47 (2006) 2205–2208
Ring opening of 2-phenylazetidines with allylsilanes
`
Mathias Domostoj, Ioana Ungureanu, Angele Schoenfelder,
z
*
´
Philippe Klotz and Andre Mann
Laboratoire de Pharmacochimie de la Communication Cellulaire UMR 7175/LC, Faculte´ de Pharmacie, 74 route du Rhin,
BP 60024, 67401 Illkirch, France
Received 24 November 2005; revised 13 January 2006; accepted 20 January 2006
Available online 10 February 2006
Abstract—N-Activated 2-phenylazetidines were opened regioselectively at the benzylic carbon with various allylsilanes or proparg-
ylsilane in the presence of BF3ÆEt2O, providing amino olefins, precursors of biomolecules such as phenyl-homo-kainoids.
Ó 2006 Elsevier Ltd. All rights reserved.
Over the years the reactivity and/or use of aziridines and
azetidines have been addressed by several groups.1–13
Shortly we and others have reported on new reactivities
of N-activated 2-phenylaziridines and 2-phenylazeti-
dines towards various nucleophiles. These three and
four membered heterocycles have been identified,
respectively, as masked 1,3 and 1,4 dipoles in the pres-
ence of selected dipolarophiles.14–18 Recently we focused
on the reactivity of 2-phenylazetidines (1a–b) as electro-
philes towards various allylsilanes, and report here our
preliminary results on the successful ring opening of
the azetidines 1a–b together with relevant applications
towards the preparation of biomolecules.
internal collapse of the silicon stabilized b-carbenium
on the transient amide anion. A similar result was
observed with the corresponding aziridine.7 It is note-
worthy that the silylated piperidines 3a–a0 could be
transformed quantitatively into 4a via a brief heating
in a TBAF solution. Interestingly if the initial reaction
mixture was aged overnight only the allylated adduct
4a was isolated, in situ desilylation being promoted by
the fluorinated Lewis acid. This observation will sim-
plify the purification step, because at that time the ally-
lated adducts were our targets for further applications
(vide infra). Very similar results in reaction rates and
yields were obtained with 1b, the nosyl activated azeti-
dine, in this case we isolated the adducts 3b–b0 and 4b,
respectively, the piperidine and the allylated adducts.
The mixture of piperidines 3b–b0 is cleanly converted
to 4b by a TBAF treatment or more conveniently with
the overnight ageing procedure. As anticipated the tosyl
or nosyl activations in 1a or 1b are similarly effective for
the azetidine ring opening with allylsilane 2. Conversely
when different Lewis acids (TiCl4, MgBr2, ZnBr2) were
experimented, we observed severe contaminations with
byproducts, mainly halogenated adducts at the benzyl
carbon. Therefore BF3ÆEt2O was favoured as a good
compromise between electronic assistance and intrinsic
reactivity. The temperature seemed not to be a critical
factor but we found that the ideal conditions are to
mix azetidine (1a) and allylsilane (2) at ꢀ78 °C, followed
by dropwise addition of the Lewis acid. The mixture was
then left to warm up to room temperature overnight.
The remarkable regioselectivity in the ring opening at
the benzylic carbon is best explained by the release of
the cyclic constraint and the polarization of the benzylic
C–N bond with the Lewis acid.20
The two starting azetidines 1a (NTs) and 1b (NNos),
selected for our study can be readily obtained from styrene
following reported procedures.19 The reaction of 1a or
1b with trimethylallylsilane 2 was performed at room
temperature in CH2Cl2 in the presence of BF3ÆOEt2
(Scheme 1). After 30 min reaction time, no starting
azetidine was apparent on TLC. From the reaction mix-
ture two adducts could be isolated by column chroma-
tography, and identified as 3a–a0,
a mixture of
inseparable silylated piperidines, and 4a the open chain
allylated adduct in a ratio 3a–a0/4a: 1/4. This first experi-
ment established that azetidine 1a is ring opened with
allylsilane at the benzylic carbon via a SE20 pathway
where azetidine 1a is the electrophilic substrate (Eq.
1). The piperidine adducts 3a–a0 is resulting from an
*
Corresponding author. Tel.: +33 03 90 24 42 27; fax: +33 03 90 24 43
10; e-mail: andre.mann@pharma.u-strasbg.fr
z Philippe Klotz died suddenly on June 9, 2005.
0040-4039/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.01.108

2206
M. Domostoj et al. / Tetrahedron Letters 47 (2006) 2205–2208
ii
even if propargylsilane is known to be less nucleophilic
Ph
N
Ph
Ph
than allylsilane, the allenyl adducts 10a and 10b were
isolated in acceptable yields. These data demonstrated
that azetidines of type 1a or 1b are of great value as
masked electrophiles.
i
+
+
ArO SN
(1)
2
SiMe
3
NHSO Ar
SiMe
2
3
SO Ar
2
2
3a-a'
1a Ar = 4-MePh
4a (70%)
4b (57%)
(14%)
(20%)
1b
3b-b'
Ar = 4-NO2Ph
To explain the regioselective ring opening of azetidines
1a–b, with allylsilanes or propargylsilane, we propose
the mechanism depicted in Scheme 2. The Lewis acid
(BF3ÆEt2O) may have a double function (i) first in
coordinating the sulfonyl appendage that stretch out
the benzylic C–N bond leading to the electrophilic
substitution; (ii) second in stabilizing the transient amide
in intermediate A, which can then evolve to a mixture of
3a and 4a, or solely to 4a, respectively, if the reaction is
quenched after 1 h, or left overnight (Scheme 2). Then
we decided to prepare more complex allylsilanes 11,
12, 14 and 1621 in order to extend the scope of the
reaction and to apply it to the synthesis of scaffolds
occurring in biomolecules (Eqs. 5–8). Each reaction
with these allylsilanes deserves special comments. With
bis-allylsilane 11 only adduct 4a was obtained, the
expected cooperativity and/or double addition of the
two allylsilane residues embedded in 11 was not
detected. Considering the peculiar reactivity of azetidine
1a with methylenecyclohexane,15 allysilane 11 is combin-
ing two potential nucleophilic centres one on the alkene
and one on the allylsilane. Therefore a multi-path
reaction was awaited for 11 towards 1a, via two putative
carbenium ions on the b- or c-carbon next to the silicon
atom.
Ph
i
(2)
1a and 1b
+
SiMe
3
ArO SHN
5
2
6a-a'
(51%)
(64%)
6b-b'
Ph
Cl
i
Cl
(3)
1a
SiMe
+
3
ArO SHN
2
7
8a (68%)
Ph
i
SiMe
3
•
(4)
+
1a or 1b
ArO SHN
9
2
10a
10b
(54%)
(66%)
Ph
i
+
1a
SiMe
2
(5)
(6)
NHSO Ar
2
(72%)
4a
11
Ph NHSO Ar
2
SiMe
3
i
1a or 1b
+
12
(61%)
Ph
13a
Indeed the reaction proceeded, but only adduct 13a
could be characterize, accounting that a carbenium
stabilized by the silicon atom ranked over a tertiary
one. Similarly as in our work with aziridines azetidines
1a and 1b were treated with cyclopentenyl- (14) and
cyclohexenyltrimethysilane (16), the allylation worked
efficiently, delivering inseparable mixtures (always in
1/1 ratio) of 15a–a0, 15b–b0 and 17a–a0, respectively, in
65% and 70% yield. Indeed we found a practical and
efficient reaction for the regioselective ring opening of
azetidines with allylsilanes. As we are currently inter-
ested in the preparation of kainoids,22 we considered
that 15a–a0 may constitute ideal starting materials for
the obtention of homo-kainoids (Scheme 3). Kainoids
have been used as successful probes to understand the
excitatory functions of glutamic acid in the mammalian
central nervous system.23–25 But there is still a demand
for more selective ligands towards the kainate receptors,
especially for antagonists, and homo-kainoids may be
potential candidates. Indeed following our previous
reported sequence for the synthesis of phenyl-kainic
acid, we used 15a–a0 towards the formal synthesis of
i
+
(7)
1a or 1b
SiMe
3
ArO SHN
2
15a-a' (69%)
14
(55%)
15b-b'
Ph
i
+
1a
(8)
SiMe
3
ArO SHN
2
16
17a-a'
(49%)
Scheme 1. Reagents and conditions: (i) BF3Et2O, CH2Cl2 from
ꢀ78 °C to rt; (ii) TBAF in THF reflux 1 h.
Next the reactivity of easily accessible allylsilanes
namely the E crotylsilane (5), or 2-chloromethyl-3-tri-
methylsilylpropene (7) towards 1a–b was explored (Eqs. 2
and 3). In each case the expected transformations were
operative and delivered the allylated adducts 6a–a0,
6b–b0, mixtures of diastereomers (1/1 ratio), and 8a. In
a complementary study, we decided to oppose proparg-
ylsilane (9) to azetidine 1a and 1b (Eq. 4). Surprisingly
H O
2
3a-a'
4a
+
4a
Ph
NSO Ar
2
O
N
Ar
Ph
S
O
Me Si
BF
3
3
F
Me Si
3
2
1a
A
Scheme 2. Proposed mechanism for the regioselective opening of azetidine 1a.

M. Domostoj et al. / Tetrahedron Letters 47 (2006) 2205–2208
2207
Ph
N
Ph
9. Rutjes, F. J. T.; Tjen, K. C. M.; Wolf, L. B.; Karstens, W.
F. J.; Schoemaker, H. E.; Hiemstra, H. Org. Lett. 1999, 1,
717–720.
10. Alcaide, B.; Almendros, P.; Aragoncillo, C.; Salgado, N.
R. J. Org. Chem. 1999, 64, 9596–9604.
11. Couty, F.; Durrat, F.; Prim, D. Tetrahedron Lett. 2003,
44, 5209–5212.
12. For recent review on aziridines see: Hu, X. E. Tetrahedron
2004, 60, 2701–2748.
13. For a recent review on azetidines see: Couty, F.; Evano,
G.; Prim, D. Mini-Rev. Org. Chem. 2004, 1, 133–148.
14. Ungureanu, I.; Klotz, P.; Mann, A. Angew. Chem., Int.
Ed. 2000, 39, 4615–4616.
ii, iii
CO Me
2
CO Me
N
2
68%
SO Ar
2
SO Ar
i
2
18a
16a (exo)
35%
30%
15a-a'
Ph
Ph
N
ii, iii
i
CO Me
2
57%
N
CO Me
2
SO Ar
2
SO Ar
2
17a (endo)
19a
R
Ph
CO H
2
CO H
2
15. Ungureanu, I.; Klotz, P.; Schoenfelder, A.; Mann, A.
Chem. Commun. 2001, 958–959.
16. Prasad, B. A. B.; Pandey, G.; Singh, V. K. Tetrahedron
Lett. 2004, 45, 1137–1141.
CO H
2
N
CO H
2
N
H
H
Homo-kainoids
Phenyl-kainic acid
17. Prasad, B. A. B.; Bisal, A.; Singh, V. K. Org. Lett. 2004, 6,
4829–4831.
18. Yadav, V. K.; Siramurthy, V. J. Am. Chem. Soc. 2005,
127, 16366–16367.
Scheme 3. Reagents and conditions: (i) Pd(OAc)2, O2, DMSO, 40 °C,
4 h, separation via chromatography; (ii) RuCl3, NaIO4; (iii) CH2N2 in
Et2O.
19. Begley, M.; Combie, L.; Haigh, D.; Jones, R.; Osborne, S.;
Webster, R. J. Chem. Soc., Perkin Trans. 1 1993, 2027–
2046.
20. In our hands azetidine 1a did not react with allylmagne-
sium bromide, no reaction was observed.
21. The allylsilanes have been prepared following reported
procedures. Compound 5: D’Aniello, F.; Falorni, M.;
Mann, A.; Taddei, M. Tetrahedron: Asymmetry 1996, 7,
1217–1226; Compound 11: Nasiak, L. D.; Post, H. W. J.
Org. Chem. 1959, 24, 489–492; Compound 12: Fleming, I.;
Paterson, I. Synthesis 1979, 446–448; Compounds 14 and
16: Ashe, A. J. Am. Chem. Soc. 1973, 95, 818–821.
22. Schneider, M. R.; Klotz, P.; Ungureanu, I.; Mann, A.;
Wermuth, C. G. Tetrahedron Lett. 1999, 40, 3873–
3876.
23. For the importance of kainoids see: Monaghan, D. T.;
Bridges, R. J.; Cotman, C. W. Ann. Dev. Pharmacol.
Toxicol. 1989, 29, 365–402.
24. For a review on kainoids see: Parsons, A. F. Terahedron
1996, 52, 4149–4174.
25. For a review on excitatory amino acids see: Braeuner-
Osborne, H.; Egebjerg, J.; Nielsen, E. O.; Madsen, U.;
Kroggsgaard-Larsen, P. J. Med. Chem. 2000, 43, 2609–
2645.
26. Larock, R. C.; Hightower, T. R.; Hasvold, L.; Peterson,
K. P. J. Org. Chem. 1996, 61, 3584–3585.
phenyl-homo-kainic acid. Using the hydroamination
reaction assisted by Pd(II),26 the mixture of diastereo-
mers (15a–a0) was converted into bicyclic adducts 16a
and 17a, which could be separated by column chroma-
tography and identified by extensive NMR (COSY
and NOESY experiments) to be, respectively, the exo
(16a) and endo (17a) adducts with a cis ring junction.
Finally, a single crystal X-ray analysis of 17a confirmed
our NMR analysis.27 Then each diastereomer could be
transformed into the fully protected phenyl-homokai-
noids 18a–19a in a two step sequence: Sharpless oxida-
tion and diazomethane esterification, awaiting final
deprotection.28,29
In conclusion in this work on N-activated azetidines, we
have demonstrated that the regioselective ring opening
with allysilanes is realized at the benzylic carbon under
smooth reaction conditions in a good yielding sequence.
Our work is probably giving a new vitality to the azeti-
dine chemistry, indeed some of the described adducts
have a real chemical potential. We are currently using
the above methodology towards the synthesis of biolo-
gical active compounds.
27. The X-ray structure and coordinates for compound 17a
have been deposited at the Cambridge Crystallographic
Database under the following code CCDC 290233.
28. Typical experimental conditions for the preparation of 4a. A
mixture of azetidine 1a (100 mg, 0.35 mmol) and trimeth-
ylallylsilane 2 (85 ll, 0.53 mmol, 1 equiv) in CH2Cl2 (5 ml)
is cooled at ꢀ78 °C and a solution of BF3ÆEt2O (100 ll,
0.8 mmol, 2.3 equiv) in CH2Cl2 (1 ml) is added dropwise.
After 1 h, H2O (2 ml) was added and the mixture was left
overnight at room temperature. More water was added
and the organic layer was separated, washed with brine,
dried with Na2SO4 and concentrated in vacuo. The residue
was purified by column chromatography on silica gel
eluting with heptane/Et2O 8:2 to yield 4a (70%).
29. Selected physical data: Compound 4a mp 76–78 °C; IR
(KBr) m (cmꢀ1) 3270, 3090, 1320, 1170; 1H NMR
(300 MHz, CDCl3, 25 °C) d 7.68 (d, J 8.5 Hz, 2H) 7.38–
7.12 (m, 5H), 7.11–7.02 (m, 2H), 5.68–5.52 (m, 1H), 4.95
(d, J 6 Hz, 1H), 4.73 (br s, 1H), 2.74–2.85 (m, 2H), 4.86–
4.83 (m, 1H), 2.62–2.58 (m, 1H), 2.42 (s, 3H), 2.29 (dt, J
7.0 and 1.5 Hz, 2H), 1.97–1.60 (m, 2H); 13C NMR (75 Hz,
CDCl3, 25 °C) d 143.7, 143.4, 137.1, 136.4, 129.8, 128.7,
References and notes
1. Gensler, W. J.; Koehler, R. W. J. Org. Chem. 1962, 27,
2754–2762.
2. Testa, E.; Fontanella, L.; Aresi, V. Liebigs Ann. Chem.
1964, 676, 160–167.
3. Hassner, A.; Wiegand, N. J. Org. Chem. 1986, 51, 3652–
3656.
4. Roberto, D.; Alper, H. J. Am. Chem. Soc. 1989, 111,
7539–7543.
5. Ojima, I.; Zhao, M.; Yamato, T.; Nakahashi, K.;
Yamashita, M.; Abe, R. J. Org. Chem. 1991, 56, 5263–
5277.
6. Satake, A.; Ishii, H.; Shimizu, I.; Inoue, Y.; Hasegawa, H.;
Yamamoto, A. Tetrahedron 1995, 51, 5331–5340.
7. Schneider, M. R.; Mann, A.; Taddei, M. Tetrahedron Lett.
1996, 37, 8493–8496.
8. O’Neil, I. A.; Potter, A. J. Chem. Commun. 1998, 1487–
1488.

2208
M. Domostoj et al. / Tetrahedron Letters 47 (2006) 2205–2208
127.7, 127.2, 126.6, 116.6, 43.1, 41.6, 41.3, 35.6, 21.7; m/z
1H), 5.07 (s, 1H), 3.97 (d, J 13 Hz, 1H), 2.91 (td, J 12 and
3 Hz, 1H), 2.47 (s, 3H), 2.38 (m, 2H), 2.17 (dd, J 165 and
2.5 Hz, 1H), 1.96 (dd, J 16 and 2.5 Hz, 1H), 1.59 (m, 2H);
13C NMR (75 MHz, CDCl3, 25 °C) d 21.6, 23, 32, 38.6,
41, 62.9, 1216.4, 127.3, 127.5, 128.5, 129.7, 135.3, 137,
143.2, 143.3; m/z (FAB-MS) 354 (M+H).
(FAB-MS) [M+H+] 330.
Compound 13a, waxy solid; IR (KBr) m (cmꢀ1) 3360, 1640,
1600, 1170; 1H NMR (300 MHz, CDCl3, 25 °C) d 7.64 (d,
J 8.3 Hz, 2H), 7.34–7.11 (m, 5H), 7.08–6.97 (m, 2H), 5.49
(dd, J 11 and 17.7 Hz, 1H), 5.25 (dd, J 1.5 and 11.3 Hz,
1H), 4.87 (dd, J 1.5 and 18 Hz, 1H), 4.28 (m, 1H), 2.81–
2.68 (m, 1H), 2.64–2.53 (m, 1H), 2.43 (s, 3H), 2.44–2.36
(m, 1H), 2.05–1.87 (m, 1H), 1.85–1.04 (m, 11H); 13C NMR
(75 Hz, CDCl3, 25 °C) d 143.6, 143, 140.8, 137.4, 130,
128.2, 127.5, 126.9, 116.4, 54.2, 42.4, 38.9, 29.6, 26.8, 22.6,
21.9; m/z (FAB-MS) [M+H+] 398.
Compound 18a waxy solid; IR (KBr) m (cmꢀ1) 3015, 2950,
1710, 1580, 1170; 1H NMR (300 MHz, CDCl3, 25 °C) d
7.82–8.73 (m, 2H), 7.51–7.43 (m, 2H), 7.37–7.18 (m, 3H),
7.17–6.99 (m, 2H), 4.61 (d, J 5.5 Hz, 1H), 4.18–4.01 (m,
1H), 3.62 (s, 3H), 3.42 (s, 3H), 3.58–3.39 (m, 1H), 3.18–
2.89 (m, 2H), 2.79–2.58 (m, 1H), 2.45 (s, 3H), 2.39–2.24
(m, 1H), 2.19–2.09 (m, 2H); 13C NMR (75 MHz, CDCl3,
25 °C) d 33.1, 33.8, 40.8, 49.4, 50.5, 51.9, 52.2, 59.0, 64.7,
124.3, 124.5, 127.2, 127.4, 128.3, 128.5, 128.8, 140, 140.7,
145.4, 152.8, 170.2, 172.5; m/z (FAB-MS) 446 (M+H+).
Compound 19a waxy solid; IR (KBr) m (cmꢀ1) 3010, 2980,
1715, 1510, 1175; 1H NMR (300 MHz, CDCl3, 25 °C) d
7.88 (d, J 8.6, 2H), 7.83 (d, J 9.0, 2H), 7.32–7.23 (m, 3H),
7.07 (d, J 6.7, 2H), 5.13 (d, J 4.9 Hz, 1H), 4.02 (d, J 2.6, 1H),
3.71–3.38 (m, 7H), 2.72–2.52 (m, 2H), 2.47 (s, 3H) 2.17–1.99
(m, 2H), 1.97–1.77 (m, 2H); 13C NMR (75 MHz, CDCl3,
25 °C) d 28.5, 33.8, 35.2, 39.9, 49.0, 41.9, 43.5, 52.4, 54.1,
126.5, 126.8, 127.1, 127.7, 128.0, 128.5, 128.9, 143.1, 144,
144.6, 155.1, 171.9, 172.5; m/z (FAB-MS) 446 (M+H+).
Compound 16a (exo), mp 123–125 °C; IR (KBr) m (cmꢀ1
)
3080, 1610, 1500, 1175; 1H NMR (300 MHz, CDCl3,
25 °C) d 7.76 (d, J 8 Hz, 2H), 7.20–7.36 (m, 5H), 7.09 (d, J
7 Hz, 2H), 5.91 (m, 1H), 5.57 (m, 1H), 4.79 (d, J 5.5 Hz,
1H), 3.48 (m, 1H), 3.30 (m, 1H), 2.90 (m, 1H), 2.76 (m,
1H), 2.46 (s, 3H), 2.25 (m, 1H), 1.93 (m, 2H). 13C NMR
(75 MHz, CDCl3, 25 °C) d 21.6, 23.0, 32, 38.6, 41.0, 62.9,
126.4, 127. 3, 127.5, 128.5, 129.7, 135.3, 137, 143.2, 143.3;
m/z (FAB-MS (M+H+).
Compound 17a (endo) mp 85–88 °C; IR (KBr) m cmꢀ1
)
3080, 1615, 1520, 1180; 1H NMR (300 MHz, CDCl3,
25 °C) d 7.68 (d, J 8.5 Hz, 2H), 7.35 (d, J 8 Hz, 2H), 7.18–
7.31 (m, 3H), 7.08 (d, J 7 Hz, 2H), 5.80 (m, 1H), 5.38 (m,