A new entry to the synthesis of substituted azetidines: [2+2] cycloaddition reaction of four-membered endocyclic enamides to ketenes

Article abstract of DOI:10.1016/j.tetlet.2006.06.174

The first example of a [2+2] cycloaddition reaction of a four-membered endocyclic enamide (2-azetine) to dichloroketene is described and constitutes a new entry to the synthesis of substituted azetidines. Preliminary studies concerning the Baeyer-Villiger

Full text of DOI:10.1016/j.tetlet.2006.06.174

Tetrahedron Letters 47 (2006) 6377–6380
A new entry to the synthesis of substituted azetidines:
[2+2] cycloaddition reaction of four-membered endocyclic
enamides to ketenes
Antonio Carlos B. Burtoloso and Carlos Roque D. Correia*
Instituto de Quı´mica, Universidade Estadual de Campinas, UNICAMP, C.P. 6154, CEP. 13083-970, Campinas, Sa˜o Paulo, Brazil
Received 8 February 2006; revised 30 June 2006; accepted 30 June 2006
Available online 24 July 2006
Abstract—The first example of a [2+2] cycloaddition reaction of a four-membered endocyclic enamide (2-azetine) to dichloroketene
is described and constitutes a new entry to the synthesis of substituted azetidines. Preliminary studies concerning the Baeyer–Villiger
oxidation of the [2+2] cycloadduct revealed an unusual regioselectivity. The synthesis of a new azetidine-3-carboxylic acid derivative
from the [2+2] cycloadduct is also presented.
Ó 2006 Elsevier Ltd. All rights reserved.
Azetidines are an interesting class of four-membered
nitrogen-containing heterocycles which have been
receiving much attention from the chemical community
recently. Some of the reasons for this increasing interest
are the applications of azetidines in biological chemistry
and in the field of chiral ligands (Fig. 1).1,2
methodologies. This scenario can be attributed mainly
to the scarcity of natural azetidines when compared to
other cyclic amines such as pyrrolidines and piperidines.
Although the literature reports a reasonable number of
studies on the synthesis of azetidines,1,2 most of these
studies are based on an intramolecular nucleophilic sub-
stitution for the construction of the four-membered ring.
Therefore, new protocols are necessary and desirable.
Despite this interest, azetidines still do not have the de-
served attention regarding the development of synthetic
During the last few years, we have been involved with
the development of different methodologies for the syn-
thesis of 4, 5 and 6-membered nitrogen and oxygen het-
erocycles.3 Of these, the [2+2] cycloaddition reaction of
five-membered endocyclic enecarbamates and enamides
with ketenes has been particularly effective (Scheme 1).
Employing this [2+2] cycloaddition methodology, we
accomplished the synthesis of many nitrogen hetero-
cycles such as the Geissman–Waiss lactone, indolizidine
and pyrrolizidine alkaloids and conformationally
restricted analogues of aspartic and glutamic acids.3d–g
OH
R2
OH
N
H
HO
R1
Penaresidin A, R1 = Me, R2 = H
Penaresidin B, R1 = H, R2 = Me
COOH
OH
COOH
N
COOH
R
R
N
H
N
N
HO
R
R
OH
chiral ligands
Mugineic acid
Cl
N
O
C
indolizidine and
pyrrolizidine alkaloids
O
HO2C
N H
O
ABT-594
R2
R3
R2
R3
Geissman-Waiss
lactone
azetidine-3-carboxylic acid
N
N
N
H
O
R1
glutamate
and aspartate
analogues
Figure 1. Some examples of natural and synthetic azetidines.2
O
R1
R2, R3=H, Cl, alkyl
*
Scheme 1. [2+2] cycloaddition of five-membered enecarbamates with
Corresponding author. Tel.: +55 19 37883114; e-mail: roque@iqm.
ketenes.
unicamp.br
0040-4039/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.06.174

6378
A. C. B. Burtoloso, C. R. D. Correia / Tetrahedron Letters 47 (2006) 6377–6380
The [2+2] cycloaddition of five-membered enamides and
enecarbamates with ketenes has shown great applicabil-
ity. However, the same reaction involving strained four-
membered nitrogen analogues constitutes a considerable
challenge.4 A successful cycloaddition would potentially
open a new entry for the synthesis of substituted azeti-
dines (Scheme 2).
chloride in 86–91% overall yield.7d The synthesis of the
2-azetine 1 was then concluded by the addition of potas-
sium tert-butoxide in a hot solution of the N-pivaloyl
mesylate 6 in tert-butanol.7d
With the 2-azetine 1 in hand, we started our study of its
[2+2] cycloaddition reaction with dichloroketene. After
some experimentation, we found that best results for
the cycloaddition reaction can be obtained with the slow
addition of dichloroketene to a solution of 2-azetine 1 in
a mixture of cyclohexane and benzene (10:1 ratio)
(Scheme 5). In view of its great instability, the cyclo-
addition adduct 2 could not be purified and handled
properly. After many fruitless attempts to convert the
crude cycloaddition product into a stable derivative,8
we discovered that the immediate reduction of cyclobu-
tanone 2 with NaBH4 in THF was the best choice.9
Using these conditions, cyclobutanol 7 could be prepared
in 54% yield after two steps from 2-azetine 1.10
Herein, we describe our preliminary study of the [2+2]
cycloaddition reaction of four-membered endocyclic
enamides and enecarbamates with ketenes and the suc-
cessful synthesis of azabicyclic cyclobutanone 2 from
the four-membered endocyclic enamide 1 (Scheme 3).
We also investigated the Baeyer–Villiger oxidation of
the highly strained azabicyclic system 2 by converting
it to the lactone 3, an advanced precursor of substituted
azetidine-3-carboxylic acids. The b-amino acid azeti-
dine-3-carboxylic acid and its analogues possess interest-
ing pharmacological properties in peptidomimetics.5
Azetidine-3-carboxylic acid has been used by the Merck
Research Laboratories for the preparation of many
pharmacologically active compounds, including tryptase
inhibitors, procollagen C-proteinase inhibitors and
growth hormone secretagogues, among others.6
Next, we investigated the removal of the chlorine atoms
by employing a mild dechlorination procedure (H2, Pd/
C, 45 psi) described in the literature.11 Interestingly, to-
gether with the desired dechlorinated alcohol 8, we also
observed the formation of piperidine 9 as the major
reduction product. Piperidine 9 was probably formed
from a very facile hydrogenolysis of the strained azabi-
cyclic cyclobutanol 7. The use of more mild conditions
such as lower temperatures and lower pressures did
not suppress the formation of piperidine 9. However,
the use of the classic radical dechlorination procedure
using n-Bu3SnH/AIBN12 circumvented this hydrogeno-
lysis problem and provided the desired azabicyclo
[2.2.0] hexanol 8 in 65–85% yields, which was smoothly
oxidized to cyclobutanone 10 with PCC.13 We then car-
ried out the Baeyer–Villiger (BV) oxidation of azabicy-
clic cyclobutanone 10 (Scheme 6). Unfortunately,
Baeyer–Villiger oxidation of ketone 10 with m-CPBA
in CH2Cl2 consistently furnished low yields of a mixture
of lactones 3 and 11. After some attempts to improve
yields, we found that sequential addition of PCC and
m-CPBA to a CH2Cl2 solution of cyclobutanol 8, at
0 °C in basic medium, was the best experimental condi-
tion to generate the BV adducts. In this case, lactones 3
We started the investigation of the [2+2] cycloaddition
reaction by preparing the non-substituted 2-azetine 1
from azetidin-3-ol
4 (Scheme 4) using protocols
described in the literature.7 Azetidin-3-ol 4 was prepared
by the reaction of benzhydrylamine and epichloridrine7a
and was subsequently converted to the mesylate 5 in
quantitative yield.7b Next, the benzhydryl group in
mesylate 5 was removed by hydrogenolysis7b,c and the
nitrogen atom protected in the presence of pivaloyl
R1
R2
R3
R1
NH
[2+2] cycloaddition
with ketenes
N
R
O
substituted azetidines
Scheme 2. Synthesis of substituted azetidines.
O
H
O
O
C
H
H
H
O
HO
Cl
Cl
Cl
N
N
O
Cl
Cl
b
a
N
N
H
N
Cl
Cl
N
O
H
O
3
2
H
O
Cl
2-azetine 1
H
O
O
2
7
O
1
unstable
(mixture of isomers)
Scheme 3. Application of the [2+2] cycloaddition reaction.
α β
: =95:5)
(
c
OH
N
d
H
H
HO
HO
MsO
MsO
b
H
H
HO
c
a
+
H
N
N
N
N
N
Ph
N
Cl
Ph
9
O
Ph
O
O
8 O
5
4
6
Ph
2-azetine 1
8
O
major product
Scheme 4. Synthesis of 2-azetine 1 from 3-azetidinol 4. Reagents and
conditions: (a) Et3N, MsCl, C6H6, 0 °C-rt, 3 h. Then, HCl/Et2O 0 °C,
100%. (b) Pd(OH)2, H2, 55 psi, 3,5 h. Then, Et3N, CH2Cl2, PivCl,
À78 °C-rt, 4 h, 86–91%. (c) t-BuOÀK+/t-BuOH, 60 °C, 6 h,
quantitative.
Scheme 5. Synthesis of cyclobutanol 8 from enamide 1. Reagents and
conditions: (a) Cl2CCO, C6H12/C6H6, 2 h, rt. (b) NaBH4, MeOH,
À21 °C, 5 min, 54% over two steps. (c) H2, Pd/C, MeOH, K2CO3, 2 h,
70% (8 + 9). (d) AIBN, n-Bu3SnH, C6H6, reflux, 24 h, 65–85%.

A. C. B. Burtoloso, C. R. D. Correia / Tetrahedron Letters 47 (2006) 6377–6380
6379
O
HO
OTBS
OTBS
H
a
N
N
O
N
17
Cbz (-)
H
H
Cbz
(-) 15
D -17 (
H
H
O
HO
H
H
α
[ ]D -8 (c 0.14, C6H6)
a
b
+
O
3
N
α
[
]
c
0.33, CHCl3)
N
O
O
8
10 O
O
HO
N
OTBS
H
O
N
11
Cbz
(+) 16
D +41 ( 1.90, CHCl3)
Scheme 6. Baeyer–Villiger oxidation of cyclobutanone 10. Reagents
and conditions: (a) PCC, CH2Cl2, NaOAc, molecular sieves, 0 °C, 1 h.
(b) m-CPBA, NaHCO3, 0 °C, 15 min, 45–55% over two steps.
α
[
]
c
Scheme 8. Synthesis of the chiral endocyclic enecarbamate 17.
Reagents and conditions: (a) Et3N, CH2Cl2, MsCl, reflux, 24 h or
POCl3, Et3N, C6H6, 24 h, (ꢀ40%).
and 11 were isolated in 45–55% yields over the two steps
as a 3:1 regioisomeric mixture. Recrystallization of this
mixture increased the ratio to 5:1.
OH
R2
OH
Interestingly, the major isomer 3 was not the expected
lactone considering the Baeyer–Villiger oxidation rules.
The use of other procedures such as urea hydrogen
peroxide complex (UHP)14 with acetic anhydride and
m-CPBA in neutral or acid medium did not change this
trend of the BV oxidation to furnish lactone 3 as the
major product. Another interesting point is the very fast
Baeyer–Villiger oxidation of cyclobutanone 10 in the
presence of PCC (10 min with PCC and 6 h in the ab-
sence of it). We hypothesize that a chromium species is
functioning as a Lewis acid or as a peroxide metal com-
plex during the Baeyer–Villiger oxidation step, making it
faster. Some efforts are under way in our laboratories
to understand not only this apparent catalysis, but also
the unusual regioselectivity of this Baeyer–Villiger
oxidation.
N
H
HO
R1
Penaresidin A, R1 = Me, R2 = H
Penaresidin B, R1 = H, R2 = Me
O
N
TBSO
N
TBSO
Cbz
17
Cbz
(-)
Scheme 9. Strategy for the synthesis of azetine alkaloids from
endocyclic enecarbamate 17.
thesize 17 from the cis-3-azetidinol 16 (Scheme 8). The
preparation of chiral enecarbamate 17, opens a new en-
try to the synthesis of chiral fully-substituted azetidines
as outlined in Scheme 9 (ongoing studies aiming at the
synthesis of azetidine alkaloids).
These results are even more surprising when we compare
these BV results with those obtained previously3e,h from
aza-cyclobutanone 12 (Scheme 7), in which the expected
lactone 13 was obtained as the major isomer. Neverthe-
less, lactone 3 constitutes an interesting intermediate to
the synthesis of substituted azetidines. Since the [2+2]
cycloaddition reaction furnished satisfactory results
with enamide 1, its employment in the reaction with
more complex enamides might constitute a potential
route to fully-substituted azetidines.
In summary, we have demonstrated for the first time the
[2+2] cycloaddition reaction between a ketene and a
four-membered endocyclic enamide. Baeyer–Villiger
reaction on the [2+2] cycloaddition adduct 10 furnished
the unusual regioisomeric lactone 3 as the major prod-
uct. Bicyclic lactone 3 poses as an advanced intermediate
for the synthesis of substituted 3-azetidine-carboxylic
acids and others substituted azetidines. Furthermore,
the synthesis of enecarbamate (À)-17 opens a new entry
to the synthesis of chiral fully substituted azetines.
In spite of the fact that there is no report in the literature
dealing with the synthesis of more complex four-mem-
bered enamides or enecarbamates, we were able to pre-
pare the four-membered chiral enecarbamate 17 from
trans-3-azetidinol 15. Curiously, the trans relationship
between the hydroxyl group and the alkyl substituent
is crucial for effective elimination since we could not syn-
Acknowledgements
We thank FAPESP and CNPq for financial support and
fellowships. We also thank Professors M. E. Jung (Uni-
versity of California—Los Angeles) and P. J. Stevenson
(Queens University of Belfast) for experimental details
concerning the synthesis of 2-azetine 1.
O
H
H
H
H
H
O
O
a or b
O
+
13
O
N
P
N 14
P
N
12
H
P
major isomer
Supplementary data
Scheme 7. Baeyer–Villiger oxidation of azabicyclobutanone 12.
Reagents and conditions: (a) (P = Cbz): m-CPBA, NaHCO3, CH2Cl2,
rt, 0,5 h, 90% (13:14 = 1:0). (b) (P = Boc): H2O2, AcOH, 0 °C, 3 h,
80% (13:14 = 7:1).
1H NMR, 13C NMR, IV and MS for all new com-
pounds. The supplementary data are available online

6380
A. C. B. Burtoloso, C. R. D. Correia / Tetrahedron Letters 47 (2006) 6377–6380
n-Bu3SnH in C6H6 under reflux were fruitless. Also, a
direct Baeyer–Villiger oxidation of 2 did not provide the
expected dichlorolactone.
with the paper in ScienceDirect, at doi:10.1016/j.tetlet.
2006.06.174.
9. The use of MeOH as solvent causes ring open of bicyclic
cyclobutanone 2 to furnish a methyl ester, together with
cyclobutanol 7 (ratio 1:1).
References and notes
10. Typical experimental procedure for the [2+2] cycloaddi-
tion reaction, and reduction of the unstable cyclobutanone
2 to cyclobutanol 7: 108.5 mg (0.78 mmol) of 2-azetine 1
was dissolved in a dry mixture of 10:1 cyclohexane/
benzene (5.5 mL), under argon. Next, 0.38 mL
(2.74 mmol) of dry Et3N was added under vigorous
stirring, followed by the slow addition (3 mL/h) of the
ketene precursor (1.95 mmol of dichloroacetyl chloride
dissolved in 3.3 mL of a 10:1 mixture of cyclohexane/
benzene). After the addition of ꢀ2.0 mL, most of 2-azetine
1 had already been consumed, as indicated by TLC
analysis (KMnO4 stain). The reaction mixture was filtered
to remove Et3NHCl, which was washed a few times with a
1:1 mixture of cyclohexane/benzene. The combined
organic layers were washed with cold water (eliminate
some remaining Et3NHCl), dried over Na2SO4 and
evaporated in vacuo to furnish the crude cyclobutanone
2 as an yellowish oil. (1H NMR (300 MHz, benzene-d6):
d = 4.73 (br s, 1H), 3.65 (dd, J = 8.8, 2.9 Hz, 1H), 3.55 (t,
J = 8.8 Hz, 1H), 3.08 (ddd, J = 8.8, 4.4, 2.9 Hz, 1H), 0.98
(s, 9H). IR (cmÀ1): 2974, 1810, 1642. MS-ESI: 254 (M+1,
1. Katritzky, A. R.; Pozharskii, A. F. Handbook of Hetero-
cyclic Chemistry, 2nd ed.; Elsevier Science: Amsterdam,
2000.
2. Couty, F.; Evano, G.; Prim, D. Mini-Reviews in Organic
Chemistry 2004, 1, 133, and references cited therein.
3. For some examples, see: (a) Burtoloso, A. C. B.; Correia,
C. R. D. Tetrahedron Lett. 2004, 45, 3355; (b) Burtoloso,
A. C. B.; Correia, C. R. D. J. Organomet. Chem. 2005, 690,
5636; (c) Burtoloso, A. C. B.; Correia, C. R. D. Synlett
2005, 1559; (d) Correia, C. R. D.; Faria, A. R.; Carvalho,
E. S. Tetrahedron Lett. 1995, 36, 5109; (e) Correia, C. R.
D.; Faria, A. R.; Matos, C. R. R. Tetrahedron Lett. 1993,
34, 27; (f) Correia, C. R. D.; Carpes, M. J. S.; Miranda, P.
C. M. L. Tetrahedron Lett. 1997, 38, 1869; (g) Correia, C.
R. D.; Miranda, P. C. M. L. Tetrahedron Lett. 1999, 40,
´
7735; (h) Ambrosio, J. C. L.; Santos, R. H. A.; Correia, C.
R. D. J. Braz. Chem. Soc. 2003, 14, 27; (i) Severino, E. A.;
Correia, C. R. D. Org. Lett. 2000, 2, 3039.
4. To the best of our knowledge, the only example of a [2+2]
cycloaddition reaction of a four-membered olefin to a
ketene was described by Conia and co-workers reporting
the reaction of cyclobutene to dichloroketene (Fadel, A.;
2
37Cl), 252 (M+1, 37Cl + 35Cl), 250 (M+1, 2 35Cl),
57, 55). Cyclobutanone 2 was immediately dissolved in
20 mL of THF and cooled to À21 °C. Next, 150 mg of
NaBH4 (5 equiv) was added to this solution and after
15 min of stirring, saturated ammonium chloride was
added. The organic solvent was removed in vacuo, and the
aqueous phase extracted with CH2Cl2. The combined
organic layers were dried over Na2SO4, filtered, and
evaporated in vacuo. The resulting oil was purified by flash
column chromatography to give 106 mg of cyclobutanol 7
(54% yield) as a white solid. 1H NMR (major isomer)
(300 MHz, tetrachloroethane-d2, 100 °C): d = 4.96 (m,
1H), 4.70 (d, J = 6.6 Hz, 1H), 4.56 (dd, J = 8.5, 1.9 Hz,
1H), 4.30 (t, J = 8.5 Hz, 1H), 3.36 (m, 1H), 3.02 (br s, 1H,
OH), 1.25 (s, 9H). IR (cmÀ1): 3299, 2969, 2925, 1608,
1480, 1425, 1365, 1149, 782. MS-ESI: 256 (M+1, 2 37Cl),
254 (M+1, 37Cl + 35Cl), 252 (M+1, 2 35Cl), 200, 198, 196,
57. TLC: Rf = 0.51, AcOEt. HRMS m/z calcd for
C10H15Cl2NO2 251.04798, found 251.04898.
Salaun, J.; Conia, J. M. Tetrahedron, 1983, 39, 1567).
¨
However, examples employing four-membered nitrogen
analogues (2-azetines) are not described in the literature.
5. (a) Barrett, D.; Tanaka, A.; Harada, K.; Ohki, H.; Watabe,
E.; Katsuyuki, M.; Ikeda, F. Bioorg. Med. Chem. Lett.
2001, 11, 479; (b) Macdonald, S. J. F.; Dowle, M. D.;
Harrison, L. A.; Clarke, G. D. E.; Inglis, G. G. A.;
Johnson, M. R.; Shah, P.; Smith, R. A.; Amour, A.;
Fleetwood, G.; Humphreys, D. C.; Molloy, C. R.; Dixon,
M.; Godward, R. E.; Wonacott, A. J.; Singh, O. M. P.;
Hodgson, S. T.; Hardy, G. W. J. Med. Chem. 2002, 45,
3878; (c) Fotsch, C.; Han, N.; Arasasingham, P.; Bo, Y.;
Carmouche, M.; Chen, N.; Davis, J.; Goldberg, M. H.;
Hale, C.; Hsieh, F.; Kelly, M. G.; Liu, Q.; Norman, M. H.;
Smith, D. M.; Stec, M.; Tamayo, N.; Xi, N.; Xu, S.;
Bannon, A. W.; Baumgartner, J. W. Bioorg. Med. Chem.
Lett. 2005, 15, 1623.
6. Miller, R. A.; Lang, F.; Marcune, B.; Zewge, D.; Song, Z.
J.; Karady, S. Synth. Commun. 2003, 33, 3347.
11. (a) Ketelaar, P. E. F.; Staring, E. G. J.; Wynberg, H.
Tetrahedron Lett. 1985, 26, 4665; (b) Tennyson, R.; Romo,
D. J. Org. Chem. 2000, 65, 7248.
12. Kunz, H.; Ganz, I. Synthesis 1994, 1353.
13. Collington, E. W.; Finch, H.; Montana, J. G.; Taylor, R.
J. K. J. Chem. Soc., Perkin Trans. 1 1990, 1839.
7. (a) Gaertner, V. R. Tetrahedron Lett. 1966, 4691; (b)
Anderson, A. G.; Lok, R. J. Org. Chem. 1972, 37, 3953; (c)
Masuda, K.; Okutani, T.; Kaneko, T. Chem. Pharm. Bull.
1974, 22, 1490; (d) Jung, M. E.; Choi, Y. M. J. Org. Chem.
1991, 56, 6729.
14. Cooper, M. S.; Heaney, H.; Newbold, A. J.; Sanderson,
W. R. Synlett 1990, 533.
8. Attempts to remove the chlorine atoms of cycloadduct
2 employing Zn/Cu, NH4Cl in MeOH or AIBN, and