Rearrangement of dialkyl-2-(azetidin-3-yl)propane-1,3-dioates-a structural revision

Article abstract of DOI:10.3987/COM-08-11463

The rearrangement reaction of dialkyl 2-[1-diphenylmethyl)azetidin-3-yl]propane-1,3-dioates during saponification was studied. Contrary to a previous report postulating the formation of a 4-(hydroxy-methyl)-2-oxo-pyrrolidine-3-carboxylic acid derivative, our analytical data showed that the rearrangement reaction led to a lactone ring. This structural revision is based on elemental analysis, IR and two-dimensional NMR studies.

Full text of DOI:10.3987/COM-08-11463

HETEROCYCLES, Vol. 75, No. 12, 2008
2981
HETEROCYCLES, Vol. 75, No. 12, 2008, pp. 2981 - 2988. © The Japan Institute of Heterocyclic Chemistry
Received, 13th June, 2008, Accepted, 14th July, 2008, Published online, 17th July, 2008. COM-08-11463
REARRANGEMENT OF DIALKYL-2-(AZETIDIN-3-YL)PROPANE-
1,3-DIOATES – A STRUCTURAL REVISION
Mark R. Faust, Cornelia E. Hoesl, and Klaus T. Wanner*
Department of Pharmacy – Center for Drug Research, LMU Munich, Butenandtstr.
5-13, D-81377 Munich, Germany. Klaus.Wanner@cup.uni-muenchen.de
Abstract – The rearrangement reaction of dialkyl 2-[1-diphenylmethyl)azetidin-
3-yl]propane-1,3-dioates during saponification was studied. Contrary to a
previous report postulating the formation of
a
4-(hydroxy-methyl)-
2-oxo-pyrrolidine-3-carboxylic acid derivative, our analytical data showed that
the rearrangement reaction led to a lactone ring. This structural revision is based
on elemental analysis, IR and two-dimensional NMR studies.
INTRODUCTION
Cyclic amino acids have attracted considerable interest in the field of synthetic and medicinal chemistry
due to their structural rigidity. As they induce conformational restrictions they were incorporated into
peptides and peptidomimetics with the intention to create novel bioactive compounds and to perform
structural and biomechanistic investigations.1 Cyclic analogues of γ-amino butyric acid (GABA), such as
nipecotic acid and guvacine, have been found to exhibit significant in vitro inhibitory affinity to the
GABA transport proteins.2 Since a low GABA concentration in the synaptic cleft is associated with the
development of epilepsy it can be counteracted by inhibition of the GABA transport proteins. As part of
an ongoing project directed towards the synthesis and biological evaluation of azetidine derivatives as
conformationally constrained GABA uptake inhibitors we planned to employ azetidin-3-ylacetic acid (1)
as a synthetic precursor.
In search of a preparative method, we found a procedure recently published by Carruthers et al.3 The
synthetic design employed therein started with N-protected diethyl 2-(azetidin-3-yl)propane-1,3-dioate
derivatives 2. The article describes how the saponification of 2a - which contains a tert-butyloxycarbonyl
group - followed by decarboxylation and subsequent cleavage of the ester and carbamate group by acidic

2982
HETEROCYCLES, Vol. 75, No. 12, 2008
hydrolysis affords azetidin-3-ylacetic acid (1), as expected, in an overall yield of 18%.3 Carruthers et al.3
also used compound 2b, incorporating a benzhydryl protection group, as starting material for the
preparation of 1. In contrast to the reaction employing 2a, saponification of 2b did not result in the
formation of the corresponding free dicarboxylic acid of 1 as expected. Instead, a compound was
produced by a rearrangement reaction which the authors, based on the structure of the product of a
subsequent transformation reaction, identified as 3. The outcome of this reaction was explained by the
presence of a basic nitrogen atom in compound 2b, which is absent in compound 2a. This was thought to
be the cause of an intramolecular acylation of the ring nitrogen by one of the ester functions present in the
molecule which initiates a rearrangement reaction. However, analytical data we obtained when
conducting this reaction demonstrate that the structural assignment of the rearrangement product 3 has to
be revised. Based on our analytical investigations, we report in this article that though a rearrangement
reaction does occur the product formed is not the lactam 3 but the lactone 4.4
CO2H
O
O
CO2H
CO2Et
HO2C
EtO2C
O
Ph
N
OH
N
H
N
R
Ph
H
N
Ph
+
-
H
Cl
Ph
2a: R = Boc
2b: R = CHPh2
4
3
1
Scheme 1
RESULTS AND DISCUSSION
For this study we prepared the required compound 2b starting with epichlorhydrin 5 (Scheme 2).
Following the procedure published by Carruthers et al.3 reacting 5 with diphenylmethanamine yielded the
benzhydryl protected azetidine 6 in one step. Compound 6 was then subjected to a mesylation reaction to
afford compound 7 still following the synthetic route described in literature.3 However, instead of the
nucleophilic substitution of the mesyl group using LiBr and displacement of the bromine with the diethyl
malonate anion described there (Scheme 2) we attempted to simplify and optimise this procedure by
directly treating mesyl compound 7 with the diethyl malonate anion (generated by deprotonation of
diethyl
malonate
with
NaH).
This
way,
we
did
indeed
obtain
diethyl
2-[1-(diphenylmethyl)azetidin-3-yl]propane-1,3-dioate (2b) in a 94% yield (Scheme 2).

HETEROCYCLES, Vol. 75, No. 12, 2008
2983
Me
O S O
O
CO2Et
EtO2C
OH
N
Br
N
O
d
a
b
c
N
N
81%3
51%3
Ph
Ph
Cl
Ph
Ph
Ph
Ph
Ph
Ph
2b
5
6
7
8
e
94%
Scheme 2. aPh2CHNH2, MeOH bMsCl, pyridine cLiBr, Li2CO3, THF d,eNaH, CH2(CO2Et)2
The next step, the saponification of compound 2b, was performed exactly as described by the
experimental procedure of Carruthers et al.3 The product obtained this way was characterized by IR,
elemental analysis and one- and two-dimensional NMR experiments. However, to our surprise the data
obtained by us, though identical with those reported by Carruthers et al., were not in accord with the
structure of the cyclic amide 3.
CO2H
O
O
c
structure as proposed
by Carrutherset al.
Ph
N
Ph
N
OH
OH
Ph
Ph
a, b
2b
3
9
O
O
O
O
CO2H
d
e
revised structures
O
O
H
N
H
NH
Ph
N
Ph
Ph
Ph
+
+
H
Ph
H
Cl-
Cl-
Ph
11
10
4
Scheme 3. aKOH, EtOH, H2O bacidification to pH 4.5 (1 M HCl) c130 °C, 2 x 10-3 bar, 120 min d(1)
130 °C, 2 x 10-3 bar, 2 h; (2) CC (n-heptane/EtOAc = 80/20 + 0.5% ethyldimethylamine) eHCl
The IR spectrum of the product shows an absorbance at 1768 cm-1 and 1715 cm-1, whereas the absorption
of a lactam group in a five-membered ring and a carboxylic group should only show absorbance at

2984
HETEROCYCLES, Vol. 75, No. 12, 2008
approximately 1720 cm-1 and 1700 cm-1,5 respectively. The absorption of five-membered cyclic lactones,
however, is generally observed at a wave number of approximately 1770 cm-1. To further verify the
structure of the product obtained we performed an elemental analysis. Since the lactone derivative 4 bears
a secondary amine functionality, we expected to obtain the corresponding ammonium salt following
acidification to pH 4.5 with 1 M HCl. In contrast, structure 3 is naturally not able to form a hydrochloride.
We found that the elemental analysis for the compound isolated is in accordance with the empirical
formula of hydrochloride 4. Furthermore, decarboxylation of the rearrangement product following the
procedure of Carruthers et al.3 led to a product (10) which after purification by column chromatography
(n-heptane/EtOAc = 80/20 + 0.5% ethyldimethylamine) and subsequent treatment with hydrogen chloride
was also found, according to elemental analysis, to form a hydrochloride 11. Compound 10 also spotted
on a TLC plate gave a positive response with Dragendorff’s reagent,6 which further proves the presence
of a basic nitrogen atom. Most importantly, the IR spectrum of 10 shows an absorbance at 1752 cm-1.
This clearly points to the presence of a lactone moiety in compound 10 just as it is present in its precursor
4. Next, we performed a multidimensional NMR analysis of compounds 4 and 10. The data for 4 were
also in accord with the lactone structure proposed by us. However, the NMR spectra contained two sets of
signals, one with high the other with very low intensity. This is likely to be a result of the presence of the
two diastereomeric forms compound 4 can take. For simplicity's sake we therefore decided to analyze the
NMR spektra of 10 which does not have diastereomeric forms. In the 1H NMR spectrum of compound 10
the ring methylene hydrogen atoms were observed at 4.13 ppm and 4.44 ppm. These chemical shift values
correspond to a structure with a methylene group located next to a lactone moiety and not to a lactam
functionality. The hydrogen atoms of a methylene group bound to a lactam moiety are expected to be at a
higher field. Furthermore, the coupling constants, 3J = 9.1/4.8 and 9.1/6.7, respectively, are evidence of a
methylene group as part of a ring system and not of an aliphatic side chain. Analysis of 1H,13C-HMBC
corroborates our structural assignment (Scheme 4). In theory, for compound 9 long range correlation (3J)
would be observed between the aliphatic hydrogen atom of Ph2CH and both, the lactam carbon atom as well
as the carbonyl atom of NCH2 in addition to the signals to be attributed to coupling with the aromatic ring
carbons. However, only one signal resulting from the coupling of the hydrogen atom of Ph2CH (δ = 4.78
ppm) with the carbon atom of NHCH2 (δ = 50.5 ppm) was detected besides the aromatic ring correlations.
Likewise, a cross-peak between the aliphatic carbon atom of Ph2CH (δ = 67.6 ppm) and the hydrogens
attached to the carbon of the NHCH2 moiety (δ = 2.57–2.75 ppm) was seen. In summary, based on our
findings the compound formed from by a rearrangement and subsequent decarboxylation reaction must be
10 whereas 9 proposed by Carruthers et al. has to be ruled out.

HETEROCYCLES, Vol. 75, No. 12, 2008
2985
O
O
OH
NH
O
N
Ph
Ph
hypothetical
observed
1H,13C-HMBC coupling of 9
1H,13C-HMBC coupling of 10
Scheme 4. Expected and observed 1H,13C-HMBC couplings for compounds 9 and 10, respectively.
Carruthers suggested that the ring expansion to lactam 3 may occur due to the presence of a basic nitrogen
which is attacked by a carbonyl group of the malonate moiety. The charged bicyclic intermediate
subesquently fragments to lactam 3. However, since we found that the rearrangement product is the lactone
4, it seems safe to assume that under the saponification conditions the ester functions are cleaved. Then, in
the next step, as soon as one of the carboxylate groups has formed, it attacks the C-2 position of the
azetidine ring. As a result of this intermolecular substitution reaction the azetidine ring is cleaved and a
γ-lactone moiety formed. From there it takes only a reprotonation step and a further ester hydrolysis, if one
of the ester functions should still be intact, to get to the isolated compound 4.
-
XOC
N
CO2Et
EtO2C
CO2
XOC
Ph
O
OH-/H2O
O
4
-
N
N
Ph
Ph
Ph
Ph
Ph
2b
X = O- or OEt
Scheme 5. Postulated mechanism for the formation of lactone derivative 4
CONCLUSION
In this article we present a structural revision of the product obtained after saponification of diethyl
2-[1-diphenylmethyl)azetidin-3-yl]propane-1,3-dioate (2b) which appears to be the lactone 4 and not the

2986
HETEROCYCLES, Vol. 75, No. 12, 2008
lactam 3. In addition, we describe
a
straightforward, optimized synthesis of diethyl
2-[1-diphenylmethyl)azetidin-3-yl]propane-1,3-dioate (2b) in three steps starting from epichlorhydrin
(overall yield: 46%).
EXPERIMENTAL
Methods and Materials. Anhydrous reactions were carried out in vacuum dried glassware under nitrogen
atmosphere. THF and Et2O were distilled from sodium metal/benzophenone ketyl prior to use. Other
common solvents for column chromatography were freshly distilled before use. Purchased chemical
reagents were used without further purification. TLC plates were made from silica gel 60 F254 on aluminum
sheets (Merck). Compounds were stained with 5% (NH4)6Mo7O24⋅4H2O, 0.2% Ce(SO4)2⋅4H2O and 5%
conc. H2SO4. Unless otherwise stated, Merck silica gel (mesh 230-400) was used as stationary phase for
flash chromatography (CC). Melting points: mp (uncorrected) were determined with a Büchi 510 Melting
Point apparatus. Elementary analysis: Elementaranalysator Rapid (Heraeus). IR spectroscopy: FT-IR
Spectrometer 1600 and Paragon 1000 (Perkin Elmer). Mass spectrometry: Mass Spectrometer 5989 A with
59980 B particle beam LC/MS interface (Hewlett Packard). NMR spectroscopy: NMR spectra were
recorded on JNMR-GX (Jeol, 400 MHz and 500 MHz) with TMS as internal standard and integrated with
the program of NMR-software Nuts (2D Version 5.097, Acorn NMR, 1995).
Diethyl 2-[1-(diphenylmethyl)azetidin-3-yl]propane-1,3-dioate (2b). Diethylmalonat (21 mL, 139
mmol, 2.2 equiv.) was added to a slurry of NaH (3.0 g, 126 mmol, 2.0 equiv.) in THF (20 mL) at 0 °C.
The reaction mixture was stirred for 15 min. Following addition of compound 73 (20.0 g, 63 mmol ) in
THF (30 mL), the reaction mixture was refluxed for 18 h. Subsequently, the reaction was quenched by
pouring the reaction mixture into ice. The resulting mixture was extracted with CH2Cl2 (3 x 50 mL), the
combined organic extracts were dried (MgSO4) and concentrated in vacuo.
2b: 22.6 g (94%). Analytical data were in accord with those reported in literature.3
4-[(Diphenylmethylamino)methyl]-2-oxotetrahydrofuran-3-carboxylic acid hydrochloride (4). KOH
(8.64 mmol, 4.4 equiv.) was added to a solution of compound 2b (748 mg, 1.96 mmol) in EtOH (7 mL)
and H2O (1.5 mL). The reaction mixture was stirred for 16 h. Following acidification to pH 4.5 using 2 N
HCl, the solvent was evaporated. The resulting residue was washed with water (3 x 5 mL) and dried in
vacuo. (yield: 511 mg; 72%).

HETEROCYCLES, Vol. 75, No. 12, 2008
2987
4: 511 mg (72%); colorless crystals, mp 106°C. 1H NMR (DMSO-d6): δ = 2.58 ppm (d, J = 6.8 Hz, 2 H,
NH2CH2), 2.93–3.05 (m, 1 H, CHCH2), 3.47 (d, J = 8.1 Hz, 1 H, CHCO), 4.03 (t, J = 8.1 Hz, 1 H, OCH2),
4.43 (t, J = 8.1 Hz, 1 H, OCH2), 4.83 (s, 1 H, CHPh2), 7.16-7.23 (m, 2 H, CHar), 7.26-7.34 (m, 4 H, CHar),
13
7.38-7.46 (m, 4 H, CHar). C NMR (DMSO-d6): δ = 40.1 ppm (CHCH2), 49.8 (NCH2), 51.0 (CHCO),
66.4 (CHPh2), 70.1 (OCH2), 126.6 (CHar), 126.9 (CHar), 128.2 (CHar), 143.9 (Cq), 169.3 (CO), 173.1
= 3423 cm-1, 3031, 2965, 2869, 2586, 1768, 1715, 1610, 1455, 1374, 1159, 1024,
~
ν
(CO). IR (KBr):
+
749, 700. MS (CI, CH5 ); m/z (%): 282 [M++1-HCl-CO2] (45), 167 (100); MS (EI, 70eV); m/z (%): 281
[M+-HCl-CO2] (5), 204(35), 167 (100). Anal. Calcd for C19H20ClNO4 (361.8): C, 63.07; H, 5.57; N, 3.87.
Found. C, 63.01; H, 5.42; N, 3.86.
4-[(Diphenylmethylamino)methyl]-4,5-dihydrofuran-2(3H)-one (10). Compound 4 (391 mg, 1.08
mmol) was heated to 130 °C at 2 x 10-3 bar for 120 min. Purification by CC (n-heptane/EtOAc = 80:20 +
0.5% ethyldimethylamine) .
10: 275 mg (91%); colorless crystals, mp 109 °C. Rf = 0.24 (n-heptane/EtOAc = 80:20 + 0.5%
1
ethyldimethylamine). H NMR (CDCl3): δ = 1.72 ppm (s br, 1 H, NH), 2.32 (dd, J = 16.4, 4.8 Hz, 1 H,
CH2CO), 2.57-2.75 (m, 4 H, NCH2, CHCH2, CH2CO), 4.13 (dd, J = 9.1, 4.8 Hz, 1 H, OCH2), 4.44 (dd, J
= 9.1, 6.7 Hz, 1 H, OCH2), 4.78 (s, 1 H, CHPh2), 7.20-7.40 (m, 10 H, CHar). 13C NMR (CDCl3): δ = 32.6
ppm (CH2CO), 35.8 (CHCH2), 50.5 (CH2N), 67.6 (CHPh2), 71.9 (OCH2), 127.1 (CHar), 127.2 (CHar),
= 3312 cm-1, 3024, 2918,
~
ν
127.3 (CHar), 128.6 (CHar), 143.4 (Cq), 143.6 (Cq),177.0 (CO). IR (KBr):
+
2849, 1752, 1492, 1451, 1193, 1000, 747, 705. MS (CI, CH5 ); m/z (%): 282 [M++1] (100), 167 (60); MS
(EI, 70eV): m/z (%): 281 [M+] (3), 204 (39), 167 (100). Anal. Calcd for C18H19NO2 (281.4): C, 76.84; H,
6.81; N, 4.98. Found: C, 76.61; H, 6.72; N, 4.87.
4-[(Diphenylmethylamino)methyl]-4,5-dihydrofuran-2(3H)-one hydrochloride (11). Compound 10
(102 mg, 0.36 mmol) in Et2O/CH2Cl2 (50:50, 5 mL) was treated with anhydrous, gaseous HCl. The
resulting compound 11 was dried in vacuo.
11: 115 mg (~100%); colorless crystals, mp 223°C. 1H NMR (CDCl3): δ = 2.00 ppm (dd, J = 21.1, 1 H,
10.7 Hz, CH2CO), 2.58 (dd, J = 21.1, 9.0 Hz, 1 H, CH2CO), 2.51-2.63 (m, 1 H, CHCH2), 2.81 (m, 2 H,
CH2N), 3.77 (dd, J = 9.7, 6.0 Hz, 1 H, OCH2), 4.21 (dd, J = 9.7, 7.0 Hz, 1 H, OCH2), 5.08 (s, 1 H,
+
= 2930 cm-1, 2743, 1788,
~
ν
CHCPh2), 7.34-7.66 (m, 10 H, CHar), 10.53 (s br, 2 H, NH2 ). IR (KBr):
1576, 1497, 1454, 1172, 1031, 748, 703, 564. MS (EI, 70eV); m/z (%): 281 [M+-HCl] (3), 204 (30), 167
+
(100); MS (CI, CH5 ); m/z (%): 282 [M++1-HCl] (100), 167 (47). Anal. Calcd for C18H20NO2Cl (317.8):
C, 68.03; H, 6.34; N, 4.41; Cl, 11.16. Found: C, 67.82; H, 6.44; N, 4.28; Cl, 10.93.

2988
HETEROCYCLES, Vol. 75, No. 12, 2008
REFERENCES
1. K.-H. Park and M. J. Kurth, Tetrahedron, 2002, 58, 8629.
2. S. Høg, J. R. Greenwood, K. B. Madsen, O. M. Larsson, B. Frølund, A. Schouesboe, P.
Krogsgaard-Larsen, and R. P. Clausen, Curr. Top. Med. Chem., 2006, 6, 1861.
3. N. Carruthers, S.-C. Wong, and T.-M. Chan, J. Chem. Res. Synop., 1996, 9, 430.
4. For related ring cleavage reaction see: M. Vuilhorgne, A. Commerçon, and S. Mignani, Chem. Lett.,
1999, 28, 605.
5. M. Hesse, H. Meier, and B. Zeeh, Spektroskopische Methoden in der organischen Chemie, Thieme
Verlag, Stuttgart, 1991, 106.
6. H. Wagner, L. Hörhammer, and P. Wolff, Biochem. Z., 1961, 334, 175.