Studies on the stability of the cyclobutane β-aminoacid skeleton: A cautionary tale

Article abstract of DOI:10.1016/j.tetlet.2004.01.084

The 2-amino-1-cyclobutanecarboxylic acid skeleton undergoes facile retro-Mannich type ring opening in solution, which may lead to unexpected by-products during its synthesis or manipulation.

Full text of DOI:10.1016/j.tetlet.2004.01.084

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 2359–2361
Studies on the stability of the cyclobutane b-aminoacid skeleton:
a cautionary tale
David J. Aitken,* Christine Gauzy and Elisabeth Pereira
Laboratoire SEESIB-CNRS, Dꢀepartement de Chimie, Universitꢀe Blaise Pascal—Clermont-Ferrand II, 24 Avenue des Landais,
63177 Aubiꢁere Cedex, France
Received 12 December 2003; accepted 19 January 2004
Abstract—The 2-amino-1-cyclobutanecarboxylic acid skeleton undergoes facile retro-Mannich type ring opening in solution, which
may lead to unexpected by-products during its synthesis or manipulation.
Ó 2004 Elsevier Ltd. All rights reserved.
~
As a contribution to the rapidly expanding family of
constrained b-aminoacids, we recently described the
short and efficient synthesis of ( )-cis-2-amino-1-cyclo-
butanecarboxylic acid 1 via a [2+2] photocycloaddition
strategy (Scheme 1).1 Our NMR spectroscopic data for
product ( )-1 were consistent with those reported in
The final step of the Ortuno groupÕs synthesis was the
catalytic hydrogenolysis of the benzylcarbamate deriv-
ative ())-2. We prepared racemic 2 from our own sam-
ples of ( )-1, and submitted it to hydrogenolysis
(Scheme 2; Table 1).
1982 by Kennewell et al.,2 but differed from those
Under the proposed literature conditions (2 atm H2,
10% Pd–C, MeOH, rt, 16 h), a mixture of three amino-
acid products was obtained, of which the least abundant
one presented NMR data identical to those of the
expected product 1. One of the two remaining com-
pounds was found to be spectroscopically identical with
3
~
published more recently by the Ortuno group for the
(1R,2S) enantiomer, described as having a specific
optical rotation of )9. This was rather confusing, since
both recent syntheses seemed to be validated by X-ray
crystallographic studies: for our part, with a benzamide
derivative of the final product,1 and with a dipeptide
derivative of the immediate synthetic precursor of ())-1
in the other case.3 Given the recent interest in such
structures,1–4 we decided to carry out a detailed study to
clear up this ambiguity, and have thus revealed the
remarkable ease with which the parent skeleton may
undergo ring opening.
H
CO2H
CO2H
NH2
H
H
CO2H
H
( )
hydrogenolysis
HN
-
1
NHR
-
( )
( )
1
2
(R = H)
CO2H
CO2H
4
H2N
-
(R = Cbz)
3
Scheme 2.
O
O
H
H
H
H
CO2H
NH2
hν
NH
O
NH
Table 1
N
H
O
N
H
R
Compound
Conditions
Results
% ( )-1–% 3–% 4
( ) -
1
Cbz
H
( )-2
( )-1
( )-1
A*
A
16–35–49
4–36–60
0–25–75
Scheme 1.
H
B
Conditions A: 2 atm H2, 10% Pd–C, MeOH–H2O, rt, 16 h (*: MeOH
only). Conditions B: 3 atm H2, 20% Pd(OH)2–C, MeOH–H2O, rt,
3 days.
Keywords: b-Aminoacid; Cyclobutane; Retro-Mannich reaction.
* Corresponding author. Tel.: +33-4-73-40-74-84; fax: 33-4-73-40-77-
17; e-mail: aitken@chimie.univ-bpclermont.fr
0040-4039/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2004.01.084

2360
D. J. Aitken et al. / Tetrahedron Letters 45 (2004) 2359–2361
OH
OH
CO2H
CO2H
Ph
NH2
(c)
(d)
(b)
(a)
HN
Boc
N
Boc
N
N
Ph
2
I
OH
CO2H
CO2H
OH
OH
4
Scheme 3. Reagents and conditions: (a) K2CO3, EtOH, reflux, 5 h, 69%; (b) H2, Pd(OH)2–C, Boc2O, EtOAc–EtOH, rt, 24 h, 69%; (c) PDC, DMF, rt,
6 h, 43%; (d) TFA, CH2Cl2 rt, 30min, 80%.
a commercial sample of 5-aminopentanoic acid, 3, while
the most abundant component was identified as 4, by
comparison with an authentic sample prepared accord-
ing to Scheme 3.
We tested the stability of the b-aminoacid ( )-1 under
~
hydrogenation conditions (Table 1). Using OrtunoÕs
conditions, most of the sample was transformed into a
mixture of 3 and 4 after 16 h; under more forcing con-
ditions (3 atm H2, 20% Pd(OH)2–C, 3 days) the trans-
formation was complete.
Of the three compounds, the NMR chemical shifts
obtained (D2O solution) for 45 are the closest to those
previously published for ())-1.3 These results clearly
suggest that the product previously thought to be ())-1
was probably a mixture of ())-1, 3 and 4, with the latter
predominating; the observation of a nonzero optical
rotation indicates that at least some of the expected
compound was present. Hydrogenation alone does not
explain the formation of 3 and 4 from 2. Indeed,
hydrogenolysis of a cyclobutane ring to give an acyclic
product usually requires forcing conditions. In a recent
example of the four-membered ringÕs stability, straight-
forward hydrogenolytic N- and/or O-debenzylations of
cis-2-aminocyclobutanols were carried out without ring
opening.6 Further investigation was therefore required.
Figure 1. J-Modulated 13C NMR spectra of fresh 5 (above) and aged
5 (below).
In order to pursue our studies, we used the a-methyl-
benzyl derivative 5,7 which incorporates a useful
Ômolecular locationÕ moiety as its N-substituent. An
aqueous solution of 5 was allowed to stand at room
temperature for 24 h. After lyophilisation, the solid
O
COO
HO
O
push-pull
H2O
H2O
HO
HN
NH
Ph
NH
Ph
1
residue was analysed in D2O solution by H and 13C
Ph
5
NMR spectroscopy. The sample had been completely
transformed, giving a complex series of signals, which
indicated, notably, the disappearance of the C1 and C2
cyclobutane ring methine centres, and the appearance of
two new types of signal corresponding to an aldehyde
(dH 9.11 ppm; dC 198.9 ppm) and an iminium (dH
6.64 ppm; dC 159.8 ppm). The J-modulated 13C spectra
of 5 and of the sample obtained from 5 by ageing are
presented in Figure 1 (above and below, respectively).
8
COOH
O
NH2
Ph
+
7
6
Scheme 4.
pullÕ induced ring opening to generate an enolate.8
Reprotonation of the latter by the solvent would lead to
8 and hence to the final mixture. Support for this
argument was obtained by ageing 5 for 24 h in D2O
solution. The same NMR spectral data were observed in
this case, except that the carbon atom giving rise to the
We interpreted these data in terms of an equilibrium
mixture of three components: a-methylbenzylamine 6, 5-
oxopentanoic acid 7, and the iminium 8 obtained by
condensation of 6 and 7 (Scheme 4). Formation of the
iminium could be explained by a spontaneous Ôpush-

D. J. Aitken et al. / Tetrahedron Letters 45 (2004) 2359–2361
2361
signal at dC 36.1 ppm (putatively a- to the carboxylate)
was monodeuterated.
formally a retro-Mannich type process and is reminis-
cent of certain azo-de Mayo fragmentations of bicyclic
b-aminoketones.10 This appears to be the first demon-
stration of analogous behaviour by the cyclobutane
b-aminoacid skeleton.11 In the light of these results,
appropriate care is cautioned in the preparation and use
of such compounds and in the interpretation of physico-
chemical data obtained or reported for them.
Confirmation of the above hypothesis was obtained
independently as follows. 5-Oxopentanoic acid 7 was
prepared in two steps from commercial 2-chlorocyclo-
pentanone in an adaptation of previous work9 (Scheme
5). Equimolar amounts of 7 and commercial ( )-
a-methylbenzylamine were dissolved together in D2O,
and the NMR spectra recorded after 30min. Data were
identical to those obtained for the sample of 5, which
had been aged in aqueous solution (for example, the
J-modulated 13C NMR spectrum was identical to the
lower one in Fig. 1).
Acknowledgements
We thank the MENRT for a grant (to C.G.) and the
CNRS for ÔJeune EquipeÕ (AIP) funding. We are also
grateful to D. Lefeuvre for some initial experiments.
The facile formation of 3 and 4 from 1 or 2 under
hydrogenation conditions is thus rationalised: the
iminium intermediate formed in solution by ring open-
ing can be simply reduced to the corresponding amine 3.
Alternatively, it can undergo hydrolysis to give the
corresponding aldehyde 7, which condenses with
already-formed amine 3 giving a new iminium, which is
in turn reduced to 4 (Scheme 6).
References and notes
1. Aitken, D. J.; Gauzy, C.; Pereira, E. Tetrahedron Lett.
2002, 43, 6177–6179.
2. Kennewell, P. D.; Matharu, S. S.; Taylor, J. B.; West-
wood, R.; Sammes, P. G. J. Chem. Soc., Perkin Trans. 1
1982, 2563–2570.
This work illustrates and underlines the propensity for
the cyclobutane b-aminoacid skeleton to undergo ring
opening in mild conditions. The proposed mechanism is
ꢀ
3. Martın-Vila, M.; Muray, E.; Aguado, G. P.; Alvarez-
Larena, A.; Branchadell, V.; Minguillon, C.; Giralt, E.;
ꢁ
ꢀ
~
Ortuno, R. M. Tetrahedron: Asymmetry 2000, 11, 3569–
3584.
O
O
ꢀ
4. (a) Izquierdo, S.; Martın-Vila, M.; Moglioni, A. G.;
Branchadell, V.; Ortuno, R. M. Tetrahedron: Asymmetry
ꢁ
COOH
CHO
(a)
(b)
~
Cl
OH
2002, 13, 2403–2405; (b) Panda, J.; Ghosh, S.; Ghosh, S. J.
Chem. Soc., Perkin Trans. 1 2001, 3013–3016; (c) Yuan, P.;
Driscoll, M. R.; Raymond, S. J.; Hansen, D. E.; Blatchley,
R. A. Tetrahedron Lett. 1994, 35, 6195–6198; (d) Mitsudo,
T.; Zhang, S.-W.; Satake, N.; Kondo, T.; Watanabe, Y.
Tetrahedron Lett. 1992, 33, 5533–5536.
7
Scheme 5. Reagents and conditions: (a) H2O, 100 °C, 1 h, 83%; (b)
NaIO4, THF–H2O, rt, 16 h, 100%.
5. Spectroscopic data for compound 4: dH (D2O) 1.66 (8H,
m), 2.31 (4H, t, J ¼ 6:8 Hz), 3.04 (4H, t, J ¼ 6:8 Hz); dC
(D2O) 21.9 (CH2), 25.1 (CH2), 35.0(CH 2), 47.1 (CH2),
180.7 (Cq); ESMS m=z 218 [MH]þ.
COOH
COO
O
HO
H2N
H2, cat.
6. Bisel, P.; Breitling, E.; Frahm, A. W. Eur. J. Org. Chem.
1998, 729–733.
NH2
3
NH2
7. Compound 5 was prepared from racemic 1-(a-methylbenz-
yl)uracil, via a [2+2] photocyclisation reaction with
ethylene then hydrolysis of the heterocyclic moiety,
according to Ref. 1, and was used as a single (racemic)
diastereomer. Full details of the interest of this and similar
structures in the asymmetric synthesis of 1 will appear in a
future full paper.
8. Ring opening processes of donor–acceptor substituted
cyclopropanes are well documented: (a) Reissig, H.-U.;
Zimmer, R. Chem. Rev. 2003, 103, 1151–1196; (b) Gnad,
F.; Reiser, O. Chem. Rev. 2003, 103, 1603–1623.
1
H2O
COOH
COOH
NH
3
O
7
COOH
9. (a) Godchot, M.; Taboury, F. C.R. Hebd. Sꢀeances Acad.
Sci. 1913, 156, 332–334; (b) Floresca, R.; Kurihara, M.;
Watt, D. S. J. Org. Chem. 1993, 58, 2196–2200.
10. For leading references on aza-de Mayo reactions, see the
appropriate sections in: (a) Winkler, J. D.; Bowen, C. M.;
Liotta, F. Chem. Rev. 1995, 95, 2003–2020; (b) Crimmins,
M. T. Chem. Rev. 1988, 88, 1435–1473.
CO2H
H2, cat.
HN
11. Very recently the ring expansion of a strained tricyclic 4-
aminocyclobutenecarboxylate system was described: Mis-
lin, G. L.; Miesch, M. J. Org. Chem. 2003, 68, 433–441;
See also: Adembri, G.; Donati, D.; Fusi, S.; Ponticelli, F.
J. Chem. Soc., Perkin Trans. 1 1992, 2033–2038.
4
CO2H
Scheme 6.