Synthesis of 3-hydroxy- and 3-carboxy-Δ2-isoxazoline amino acids and evaluation of their interaction with GABA receptors and transporters

Article abstract of DOI:10.1002/ejoc.200600628

A series of 3-hydroxy- and 3-carboxy-Δ²-isoxazoline amino acids, structurally related to the GABA_A agonist THIP and to the GABA uptake inhibitors THPO and exo-THPO, was prepared by means of synthetic strategies involving the 1,3-dipolar cycloaddition of nitrile oxides. All derivatives were submitted to a pharmacological investigation at both GABA receptors and transporters. Unfortunately, all amino acids were devoid of activity except for a very low affinity at GABA_A receptors, displayed by compounds 8 and 17. Wiley-VCH Verlag GmbH & Co. KGaA, 2006.

Full text of DOI:10.1002/ejoc.200600628

FULL PAPER
DOI: 10.1002/ejoc.200600628
Synthesis of 3-Hydroxy- and 3-Carboxy-∆2-isoxazoline Amino Acids and
Evaluation of Their Interaction with GABA Receptors and Transporters
Paola Conti,*[a] Marco De Amici,[a] Andrea Pinto,[a] Lucia Tamborini,[a]
Giovanni Grazioso,[a] Bente Frølund,[b] Birgitte Nielsen,[b] Christian Thomsen,[c]
Bjarke Ebert,[c] and Carlo De Micheli[a]
Keywords: Neurotransmitters / Amino acids / ∆2-Isoxazoline amino acids / GABA receptors and transporters / Binding
affinity
A series of 3-hydroxy- and 3-carboxy-∆2-isoxazoline amino
acids, structurally related to the GABAA agonist THIP and
to the GABA uptake inhibitors THPO and exo-THPO, was
prepared by means of synthetic strategies involving the 1,3-
dipolar cycloaddition of nitrile oxides. All derivatives were
submitted to a pharmacological investigation at both GABA
receptors and transporters. Unfortunately, all amino acids
were devoid of activity except for a very low affinity at GA-
BAA receptors, displayed by compounds 8 and 17.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
Introduction
γ-Aminobutyric acid (GABA, 1, Figure 1) is the main
inhibitory neurotransmitter in the central nervous system.
A decrease in GABAergic neurotransmission seems to be
involved in neurological pathologies such as epilepsy,[1,2]
anxiety[3] and pain.[4] GABA interacts with three classes of
receptors: the ionotropic GABAA and GABAC receptors,
which are ligand-gated chloride-ion channels and cause in-
hibition of neuronal firing, and the metabotropic GABAB
receptors, which are G-protein-coupled receptors.[5–7]
GABA is removed from the synaptic cleft by high affinity
sodium-dependent GABA transporters located both in neu-
rons and in glial cells.
The approaches used to modulate GABAergic trans-
mission include direct agonism at the GABA receptors or
inhibition of GABA reuptake into neuronal and glial cell
bodies. Among GABA uptake inhibitors, particularly rel-
evant from a therapeutic perspective are those showing
selectivity for the glial transporters vs. the neuronal trans-
porters since they increase the amount of GABA taken up
by neuronal carriers, consequently, they enhance the release
of GABA into the synaptic cleft. Selective GABAA receptor
ligands can be obtained by conformational rigidification of
Figure 1. Model compounds.
the endogenous neurotransmitter. A typical example is rep-
resented by muscimol[8] (2, Figure 1), a natural compound
found in different species of mushrooms, in which the
propanoic acid moiety is bioisosterically replaced by the 3-
hydroxyisoxazole group. Besides its good activity as a
GABA agonist, muscimol also interacts with the GABA
uptake system. A further rigidification of muscimol led to
the bicyclic derivative THIP[9] (3, Figure 1), which also pro-
vides a noticeable GABAA agonist activity but, at variance
with muscimol, is devoid of any activity at GABA trans-
porters. THIP (Gaboxadol) is currently in phase III clinical
trials as a hypnotic drug.[10] While THIP behaves as a highly
selective GABAA receptor agonist with functional selectiv-
[a] Istituto di Chimica Farmaceutica e Tossicologica “Pietro Pra-
tesi”, Università degli Studi di Milano,
Viale Abruzzi 42, 20131 Milano, Italy
Fax: +39-02-50317565
E-mail: paola.conti@unimi.it
[b] Department of Medicinal Chemistry, The Danish University of
Pharmaceutical Sciences,
2 Universitetsparken, 2100 Copenhagen, Denmark
[c] H. Lundbeck A/S, Research Denmark,
Ottiliavej 9, 2500 Valby, Denmark
Eur. J. Org. Chem. 2006, 5533–5542
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5533

P. Conti, et al.
FULL PAPER
Figure 2. Target compounds.
ity for δ-containing extrasynaptic receptors,[11] its re- strained GABA-mimetic. Similarly, whereas amino acid 12
gioisomer
4,5,6,7-tetrahydroisoxazolo[4,5-c]pyridine-3-ol has been designed as the lower homologue of both 8 and 9,
(THPO,[12] 4, Figure 1) turned out to be a GABA uptake derivative 13 can be seen either as a higher homologue of
inhibitor, essentially devoid of any activity at GABA recep- 12 or as another constrained conformer of GABA. Finally,
tors. Due to its unexpected selectivity profile, THPO has amino acids 14 and 15 were devised as GABA-mimetics, in
been used as a model compound for the design of new which the embedded conformations of the amino acid chain
GABA uptake inhibitors selective for glial transporters (e.g. were different from that of derivatives 11 and 13. On the
exo-THPO 5 and its N-methyl analogue 6, Figure 1).[13]
other hand, amino acids 16 and 17 represent new GABA
In order to study the structure-activity relationships of homologues (Figure 2). The new amino acids 8–17 were
THIP, some authors prepared and tested the corresponding submitted to binding assays at both GABAA and GABAB
isoxazoline derivative 8 (cis-DH-THIP, Figure 1).[14] The bio- receptors and GABA transporters.
isosteric replacement of the 3-hydroxyisoxazole moiety-
with the 3-hydroxyisoxazoline ring was successfully applied
to muscimol, in fact, its 4,5-dihydro derivative [(R,S)-4,5-
Results and Discussion
dihydromuscimol (7), Figure 1] turned out to be almost
equipotent with the lead compound.[15] Interestingly, while
The key step of the synthetic pathway used to prepare
the (S)-(+) isomer mainly activated the GABAA receptors, amino acids 8 and 9 was the 1,3-dipolar cycloaddition of
the inhibition of GABA transporters was exclusively due to bromonitrile oxide,[16] generated in situ by the treatment of
the (R)-(–) isomer,[15] indicating the influence of the stereo- dibromoformaldoxime with solid NaHCO3, to 1-(tert-bu-
chemistry on the pharmacological profile.
toxycarbonyl)-1,2,3,6-tetrahydropyridine 18, in turn ob-
The synthetic strategy designed to prepare amino acid 8 tained from commercially available 1,2,3,6-tetrahydropyri-
yielded also its regioisomer 9, which is structurally related dine following a literature procedure.[14] The cycloaddition
to THPO (Figure 2).[14] However, due to the low efficiency step (Scheme 1) gave rise to an inseparable mixture of the
of the applied synthetic approach, the authors were able to two regioisomers 19a and 19b in an almost equimolar ratio.
isolate only small amounts of derivative 8, whereas 9 was The subsequent substitution of the 3-bromo group with the
not obtained in a pure form and, consequently, its pharma- 3-hydroxy moiety, achieved with 1  aqueous NaOH in di-
cological activity and selectivity were not evaluated. Since oxane at 80 °C, afforded derivatives 20a and 20b. These in-
the pharmacological profile of isoxazoline amino acid 9 was termediates could be separated by silica gel column
not defined and, furthermore, the activity of its regioisomer chromatography and were then converted into final amino
8 at GABA transporters was not determined, we devised a acids 8 and 9 through a standard removal of the N-tert-
new efficient preparation of both amino acids 8 and 9. As butoxy group. Starting from dipolarophile 18, the synthesis
a further development of this research, we also designed, described above allowed us to obtain both isomeric amino
prepared and tested a group of analogues and homologues acids 8 and 9 in reasonable overall yields of 17% and 22%,
of amino acids 8 and 9. In particular, amino acids 10 and respectively. The present synthesis constitutes a significant
11 (Figure 2) are higher homologues of 8 and 9, and deriva- improvement over the previously reported protocol,[14]
tive 11 could also be considered a conformationally con- which produced the two regioisomers in less than 1% over-
5534
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5533–5542

Synthesis of ∆2-Isoxazoline Amino Acids
FULL PAPER
all yield. The structural assignments of intermediate and
target derivatives are based on 1H NMR and COSY experi-
ments.
Scheme 2. a: NaHCO3, AcOEt, ∆, 7 d, b: 30% TFA/CH2Cl2, c: 1 
NaOH, EtOH, d: 2  HCl , e: Amberlite IR-120 (H+ form), 10%
py/H2O.
Scheme 1. a: NaHCO3, AcOEt, 3 d, b: 1  NaOH, dioxane, 80 °C,
c: 30% TFA/CH2Cl2, d: Amberlite IR-120 (H+ form), 10% py/H2O.
Amino acids 10 and 11 (Scheme 2) were in turn prepared et al.,[19] the cycloaddition of nitrile oxides to allylic second-
with the same dipolarophile 18 and ethyl 2-chloro-2-(hy- ary amides proceeds with a high degree of regio- and ste-
droxyimino)acetate[17] as the precursor of the desired 1,3- reocontrol, leading to the syn isomer as the major product
dipole. Since, in this case, the cycloaddition proceeded very (70–90%) with the amide group facing the oxygen of the
slowly at room temperature, the reaction was carried out in heterocyclic ring. In most of the reported examples, this ste-
refluxing AcOEt. As in the previous case, cycloadducts 21a reoisomer was contaminated by marginal amounts of the
and 21b were collected as an inseparable mixture in an al- two anti isomers. Similarly, we observed a high degree of
most equimolar ratio. The chromatographic separation of stereocontrol when an allylic carbamate (protected as
the two regioisomers was made possible after removal of NHBoc) was employed as the dipolarophile in 1,3-dipolar
the N-Boc group at the stage of amino esters 22a and 22b. cycloaddition with nitrile oxides.[20] Since we were inter-
These intermediates were then converted into amino acids ested in a strategy able to generate all four stereoisomers in
10 and 11 by alkaline hydrolysis of the ester group, followed comparable amounts, we employed cyclopenten-1-ol as the
by acidification and purification by cation-exchange col- dipolarophile. As described by Caramella et al.,[21] benzoni-
umn chromatography. Also, in this case, structural assign- trile oxide (BNO) reacts with cyclopenten-1-ol 28 to yield
ments of the regioisomers are based on 1H NMR and all four possible stereoisomers 29a, 29b, 29c and 29d in a
COSY experiments.
12:53:5:30 ratio (Scheme 4). The ratio among the isomers is
Similarly, amino acids 12 and 13 were prepared with the the result of a subtle balance of polar, steric and hydrogen-
strategies described above with 1-(tert-butoxycarbonyl)-3- bonding effects. In our hands, the outcome of the cycload-
pyrroline[18] as the dipolarophile (Scheme 3). The final com- dition of ethoxycarbonylformonitrile oxide to cyclopenten-
pound 12 was purified as its zwitterion by ion-exchange col- 1-ol turned out to be even more advantageous in terms of
umn chromatography with Amberlite IR-120 (H+ form), stereoisomer distribution, since we isolated the four cy-
and 10% aqueous pyridine as the eluent. Conversely, its cloadducts 30a–d in
a
relative ratio of 23:44:17:16
homologue 13 was isolated as the corresponding crystalline (Scheme 4). The structural assignments of the four stereo-
trifluoroacetate.
isomers are based on 1H NMR and COSY experiments.
To prepare amino acids 14–17, we took into account the For the 4-substituted derivatives, when H3a and H4 are in
cycloaddition of ethoxycarbonylformonitrile oxide to N- a trans relative stereochemistry (e.g. compound 30b), H3a
Boc-2-cyclopentenylamine. However, as reported by Curran resonates as a doublet (J = 9.2 Hz), whereas in the corre-
Eur. J. Org. Chem. 2006, 5533–5542
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5535

P. Conti, et al.
FULL PAPER
Scheme 3. a: Br2CN=OH, NaHCO3, AcOEt, 3 d, b: EtOOCC(Cl)N=OH, NaHCO3, AcOEt, 5 d, c: 1  NaOH, dioxane, 80 °C, d: 1 
NaOH, EtOH, e: 30% TFA/CH2Cl2, f: Amberlite IR-120 (H+ form), 10% py/H2O.
Scheme 4. a: MeSO2Cl, TEA, CH2Cl2, b: NaN3, DMSO, 80 °C, c: H2-Pd/C, Boc2O, EtOH, d: 20% K2CO3 (aq.), EtOH, e: 30% TFA/
CH2Cl2.
5536
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5533–5542

Synthesis of ∆2-Isoxazoline Amino Acids
FULL PAPER
sponding cis stereoisomer (e.g. 30c), H3a appears as a doub- Table 1. Molecular parameters describing the pharmacophoric dis-
tance N–Cω of reference (2–5 and 7) and investigated (8, 9 and 11–
let of doublets (J = 8.4 and 8.4 Hz). Analogously, in the
15) compounds.
case of the 6-substituted derivatives, H6a is a doublet in 30a
(J = 8.8 Hz) and a doublet of doublets in 30d (J = 5.1 and
Compounds[a]
∆E (kcal/mol)[b]
N–Cω distance [Å][c]
9.2 Hz). The four isomers were separated by flash
chromatography and then transformed into the desired 2B
amino acids following the reaction sequence reported in
Scheme 4. After activation of the secondary alcohol as the
2A
0.00
0.74
0.00
1.14
0.00
0.96
0.00
0.30
0.00
1.36
0.00
0.18
0.00
0.30
0.00
0.98
0.00
1.90
0.00
2.06
0.00
1.73
0.00
0.98
4.52
4.31
4.20
4.12
3.86
3.80
3.34
3.15
3.75
4.29
3.77
3.98
3.73
3.31
4.80
4.75
3.04
3.54
4.08
4.65
3.07
3.16
4.49
3.76
3A
3B
4A
mesylate, each derivative was submitted to nucleophilic sub-
4B
stitution with sodium azide, which proceeded with complete
inversion of configuration. The intermediate azides were
not characterized but directly submitted to catalytic hydro-
genation in the presence of di-tert-butyl dicarbonate to give
the N-Boc-protected compounds 31a–d. As reported for the
5A
5B
7A
7B
8A
8B
corresponding hydroxy derivatives, the relative stereochem- 9A
istry was assigned by means of 1H NMR and COSY experi-
9B
11A
ments. Alkaline hydrolysis of their ester group followed by
deprotection of the amino group with trifluoroacetic acid
afforded the final amino acids 14–17 as trifluoroacetates.
All new compounds 8–17 were pharmacologically char-
acterized in receptor binding studies on rat brain membrane
preparations. The affinities for GABAA and GABAB recep-
tor sites, using [3H]muscimol and [3H]GABA, respectively,
11B
12A
12B
13A
13B
14A
14B
15A
were determined with previously described procedures.[22] 15B
All the compounds under study displayed Ki and IC50 val-
ues higher than 100 µ except for amino acids 8 (Ki = 14 µ
at GABAA receptors and IC50 = 86 µ at GABAB recep-
tors) and 17 (Ki = 38 µ at GABAA receptors). Further-
more, derivatives 8–17 were unable to inhibit [3H]GABA
uptake when assayed at 1 m concentration in rat brain
synaptosomes, with GABA (1 m) and Tiagabine (10 µ)
[a] A corresponds to the most favoured conformation at room tem-
perature whereas B is the second most favoured populated con-
former. [b] ∆E is the difference in energy between the two optimized
conformations, calculated in a polarizable conductor-like solvation
model (C-PMC/H2O). [c] See text for the definition of the N–Cω
distance.
as controls. Finally, the amino acids were tested as modula- conformations of (S,S)-8. As reported in Figure 3 (part A),
tors of spontaneous activity in the rat cortical wedge prepa- the isoxazoline rings and the amino groups of the two com-
ration. All derivatives were assayed at 30 µ alone or in pounds are properly fitted. Therefore, the drop in affinity
combination with a concentration of GABA capable of elic- of 8 compared to 7 might be attributed to a steric repulsion
iting a 20% response. All the compounds were inactive between C4 and the complementary receptor site. Quite
either as direct activators or modulators of GABA re- similarly, the low affinity observed for 4-methylmuscimol
sponses.
(IC50 Ͼ 100 µ)[24] in comparison with muscimol had been
To account for the pharmacological data reported above, mainly attributed to a steric bump of the methyl group lo-
we carried out a conformational analysis on selected deriva- cated in position 4. Indeed, the C4 of 8 is located in the
tives 8, 9 and 11–15, which share the same N–Cω through- space occupied by the 4-methyl group of 4-methylmuscimol,
bond path (where N is the nitrogen of the amino group, as clearly shown by superimposing their populated confor-
and Cω corresponds either to the C3 of the 3-hydroxyisox- mations (Figure 3, B). As far as the activity at GABA trans-
azoline (or 3-hydroxyisoxazole) moiety or the carbon of the porters is considered, even though the N–Cω distance mea-
carboxylate group) with reference compounds 2–5 and 7. sured for derivatives 9 and 12 falls in the same range of that
Theoretical calculations were performed at the B3LYP/6- of reference compounds 4 and 5, respectively (Table 1), the
31G(d) level, as implemented in GAUSSIAN03.[23] All the two compounds were inactive as GABA uptake inhibitors.
compounds showed two populated conformations, tagged
In summary, we developed a new efficient strategy to
as A and B (Table 1), which were analyzed by examining synthesize amino acids 8 and 9, which are structurally re-
the inter-atomic distance among the two ionisable centres, lated to THIP 3 and THPO 4, respectively. Furthermore, a
N and Cω.
group of conformationally rigidified amino acids was pre-
Inspection of the parameters gathered in Table 1 reveals pared, in which the length of the amino acidic chain was
that amino acid 8, the only derivative with a detectable modulated to check how critical such a parameter is for
binding to GABA receptors, can be compared, in terms of biological activity. Our results give evidence that the dis-
N–Cω distance, to the model compound 7. Consequently, tance between the two pharmacophoric groups alone is in-
since (S)-7 is the enantiomer of dihydromuscimol acting as adequate to rationalize the interaction with both the GABA
a competitive GABAA agonist (Ki = 0.004 µ),[15] we super- uptake system and the GABA receptor sites. Crystal struc-
imposed its most populated conformer onto the A and B tures of suitable ligands inside the transporter or receptor
Eur. J. Org. Chem. 2006, 5533–5542
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5537

P. Conti, et al.
FULL PAPER
The two regioisomers were separated by column chromatography
(silica gel, petroleum ether/AcOEt, 3:7) to give 0.688 g (43% yield)
of (Ϯ)-20a and 0.776 g (49% yield) of (Ϯ)-20b.
(؎)-20a: Colourless oil. Rf (petroleum ether/AcOEt, 3:7) = 0.22.
1H NMR (300 MHz, [D6]DMSO, 100 °C): δ = 1.42 (s, 9 H), 1.66–
1.90 (m, 2 H), 2.86 (ddd, J = 7.5, 7.5, 7.5 Hz, 1 H), 3.16–3.34 (m,
2 H), 3.44 (dd, J = 4.4, 14.5 Hz, 1 H), 3.61 (dd, J = 4.4, 14.5 Hz,
1 H), 4.56 (ddd, J = 4.4, 4.4, 7.5 Hz, 1 H), 10.90 (br. s, 1 H).
C11H18N2O4 (242.27): calcd. C 54.53, H 7.49, N 11.56; found C
54.52, H 7.53, N 11.22.
Figure 3. (A) Superimposition of (S,S)-8A, (S,S)-8B and (S)-7A.
(B) Superimposition of (S,S)-8A, (S,S)-8B and 4-methylmuscimol.
Carbon atoms of conformers 8A and 8B are depicted in green
whereas those of (S)-7 are in white-grey and those of 4-methylmus-
cimol are in yellow. Pictures were acquired by SYBYL7.2 (Tripos
Inc., 1699 South Hanley Rd., St. Louis, Missouri, 63144, USA).
(؎)-20b: Colourless oil. Rf (petroleum ether/AcOEt, 3:7) = 0.35.
1H NMR (300 MHz, [D6]DMSO, 100 °C): δ = 1.42 (s, 9 H), 1.68–
1.92 (m, 2 H), 2.85 (ddd, J = 5.8, 5.8, 7.7 Hz, 1 H), 3.18 (ddd, J =
4.8, 6.6, 13.2 Hz, 1 H), 3.36–3.48 (m, 2 H), 3.68 (dd, J = 5.8,
13.5 Hz, 1 H), 4.72 (ddd, J = 5.5, 5.5, 7.7 Hz, 1 H), 11.21 (br. s, 1
H). C11H18N2O4 (242.27): calcd. C 54.53, H 7.49, N 11.56; found
C 54.88, H 7.76, N 11.25.
cavities would shed light on the relevant interactions at the
molecular level and promote the rational design of novel
targeted compounds and are, therefore, highly desirable.
(؎)-3a,4,5,6,7,7a-Hexahydroisoxazolo[5,4-c]pyridin-3-ol
[(؎)-8]:
Compound (Ϯ)-20a (0.688 g, 2.84 mmol) was treated with
CF3COOH (30% CH2Cl2 solution, 7.2 mL, 28.4 mmol) at 0 °C.
The solution was stirred at room temp. for 3 h. The volatiles were
removed under vacuum, and the residue was dissolved in H2O and
purified by cation-exchange column chromatography with Am-
berlite IR-120 (H+ form). The aqueous solution was slowly eluted
onto the resin, and the column was washed with H2O until the pH
was neutral. The compound was then eluted off the resin with 10%
aqueous pyridine, and the product-containing fractions (detected
with ninhydrin on a TLC plate) were combined and concentrated
under vacuum. The residue was taken up with ethanol, filtered,
washed sequentially with ethanol and Et2O and dried in vacuo at
50 °C to give amino acid (Ϯ)-8 (0.198 g, 49% yield). White prisms,
m.p. dec. Ͼ197 °C, Rf (nBuOH/H2O/CH3COOH, 4:2:1) = 0.37. 1H
NMR (300 MHz, [D6]DMSO): δ = 1.52–1.70 (m, 2 H), 2.38–2.58
(m, 2 H), 2.65 (dd, J = 6.2, 13.5 Hz, 1 H), 2.72–2.81 (m, 1 H), 2.88
(dd, J = 5.1, 13.5 Hz, 1 H), 3.30 (br. s, 2 H), 4.31–4.40 (m, 1 H).
C6H10N2O2 (142.16): calcd. C 50.69, H 7.09, N 19.71; found C
50.76, H 7.32, N 19.49.
Experimental Section
Material and Methods: 1H NMR spectra were recorded with a Var-
ian 300 (300 MHz) spectrometer at 20 °C. Chemical shifts (δ) are
expressed in ppm and coupling constants (J) in Hertz. HPLC
analyses were performed with a Jasco PU-980 pump, equipped with
a Jasco UV-975 UV/Vis detector and a Lichrosphere Si 60 (Merck)
column. TLC analyses were performed with commercial silica gel
60 F254 aluminium sheets, and spots were further evidenced by
spraying with a dilute alkaline potassium permanganate solution
or with ninhydrin. Melting points were determined with a Büchi
apparatus and are uncorrected. All reagents were purchased from
Aldrich. 1-(tert-Butoxycarbonyl)-1,2,3,6-tetrahydropyridine 18,[14]
1-(tert-butoxycarbonyl)-3-pyrroline 23,[18] (Ϯ)-cyclopent-2-en-1-ol
(Ϯ)-28,[25] dibromoformaldoxime[16] and ethyl 2-chloro-2-(hy-
droxyimino)acetate[17] were prepared according to literature pro-
cedures.
(؎)-3a,4,5,6,7,7a-Hexahydroisoxazolo[4,5-c]pyridin-3-ol
[(؎)-9]:
Compound (Ϯ)-20b (0.776 g, 3.21 mmol) was treated as described
for (Ϯ)-20a to give amino acid (Ϯ)-9 (0.255 g, 56% yield). White
prisms, m.p. dec. 177–179 °C, Rf (nBuOH/H2O/CH3COOH, 4:2:1)
tert-Butyl (؎)-3-Hydroxy-4,5,7,7a-tetrahydroisoxazolo[5,4-c]pyrid-
ine-6(3aH)-carboxylate [(؎)-20a] and tert-Butyl (؎)-3-Hydroxy-
3a,6,7,7a-tetrahydroisoxazolo[4,5-c]pyridine-5(4H)-carboxylate [(؎)-
1
= 0.42. H NMR (300 MHz, [D6]DMSO): δ = 1.52–1.63 (m, 1 H),
20b]: To
a solution of 1-(tert-butoxycarbonyl)-1,2,3,6-tetra-
1.66–1.77 (m, 1 H), 2.45 (ddd, J = 4.0, 7.7, 12.8 Hz, 1 H), 2.59
(ddd, J = 6.2, 6.2, 6.2 Hz, 1 H), 2.65 (ddd, J = 4.0, 7.3, 12.8 Hz, 1
H), 2.80 (m, 2 H), 3.32 (br. s, 2 H), 4.58 (ddd, J = 6.2, 6.2, 6.2 Hz,
1 H). C6H10N2O2 (142.16): calcd. C 50.69, H 7.09, N 19.71; found
C 50.39, H 7.15, N 19.51.
hydropyridine (18) (1.5 g, 8.2 mmol) in AcOEt (30 mL) was added
dibromoformaldoxime (3.3 g, 16.4 mmol) and solid NaHCO3 (7 g).
The mixture was vigorously stirred for 3 d. The progress of the
reaction was monitored by TLC (petroleum ether/AcOEt, 4:1).
H2O (15 mL) was added to the reaction mixture, and the organic
layer was separated. The aqueous phase was further extracted with
AcOEt (10 mL), and the combined organic layers were dried with
anhydrous Na2SO4. The crude material, obtained after evaporation
of the solvent, was purified by column chromatography (silica gel,
petroleum ether/AcOEt, 4:1) to give 2.0 g (80% yield) of an insepa-
rable mixture of cycloadducts (Ϯ)-19a and b, which were directly
submitted to the next step.
Ethyl (؎)-3a,4,5,6,7,7a-Hexahydroisoxazolo[5,4-c]pyridin-3-carbox-
ylate [(؎)-22a] and Ethyl (؎)-3a,4,5,6,7,7a-Hexahydroisoxazolo[4,5-
c]pyridin-3-carboxylate [(؎)-22b]: To a solution of 1-(tert-butoxy-
carbonyl)-1,2,3,6-tetrahydropyridine 18 (1.5 g, 8.2 mmol) in AcOEt
(30 mL) was added ethyl 2-chloro-2-(hydroxyimino)acetate (2.48 g,
16.4 mmol) and solid NaHCO3 (7 g). The mixture was vigorously
stirred and refluxed for 7 d. The progress of the reaction was moni-
tored by TLC (petroleum ether/AcOEt, 4:1). H2O (10 mL) was
added to the reaction mixture, and the organic layer was separated.
The aqueous phase was further extracted with AcOEt, and the
combined organic layers were dried with anhydrous Na2SO4. The
crude material, obtained after evaporation of the solvent, was puri-
fied by column chromatography (silica gel, petroleum ether/AcOEt,
4:1) to give 0.733 g (30% yield) of an inseparable mixture of cy-
The mixture of (Ϯ)-19a and b (2.0 g, 6.55 mmol) was dissolved in
dioxane (16 mL), and NaOH (1 , 33 mL) was added. The reaction
was stirred and heated at 80 °C for 24 h. The progress of the reac-
tion was monitored by TLC (petroleum ether/AcOEt, 7:3). The
aqueous phase was made acidic with 2  HCl and extracted with
CH2Cl2 (3ϫ10 mL). The combined organic extracts were dried
with anhydrous Na2SO4 and concentrated under reduced pressure.
5538
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5533–5542

Synthesis of ∆2-Isoxazoline Amino Acids
FULL PAPER
cloadducts (Ϯ)-21a–b, which were directly submitted to the next
step.
layer was separated. The aqueous phase was further extracted with
AcOEt (5 mL), and the combined organic layers were dried with
anhydrous Na2SO4. The crude material, obtained after evaporation
of the solvent, was purified by column chromatography (silica gel,
petroleum ether/AcOEt, 7:3) and recrystallization from 2-propanol
to give compound (Ϯ)-24 (1.4 g, 82% yield). White prisms, m.p.
50–52 °C, Rf (cyclohexane/AcOEt, 70:30) = 0.35. 1H NMR
(300 MHz, CDCl3): δ = 1.43 (s, 9 H), 3.37 (dd, J = 6.1, 13.5 Hz, 1
H), 3.51 (dd, J = 4.4, 13.5 Hz, 1 H), 3.80–4.10 (m, 3 H), 5.12 (br.
dd, J = 5.4, 7.9 Hz, 1 H). C10H15BrN2O3 (291.14): calcd. C 41.25,
H 5.19, N 9.62; found C 41.60, H 5.34, N 9.44.
The mixture of (Ϯ)-21a–b (0.733 g, 2.46 mmol) was treated with
CF3COOH (30% CH2Cl2 solution , 6.3 mL, 24.6 mmol) at 0 °C.
The solution was stirred at room temp. for 3 h. The disappearance
of the starting material was monitored by TLC (petroleum ether/
AcOEt, 7:3). The volatiles were removed under vacuum, the residue
was dissolved in 5% aqueous NaHCO3 (60 mL) and the aqueous
phase was extracted with AcOEt (3ϫ20 mL). The combined or-
ganic extracts were dried with anhydrous Na2SO4 and concentrated
under reduced pressure. The two regioisomers were separated by
column chromatography (silica gel, CH2Cl2/MeOH, 95:5) to give
0.182 g (37% yield) of (Ϯ)-22a and 0.250 g (51% yield) of (Ϯ)-22b.
tert-Butyl (؎)-3-Hydroxy-3a,4,6,6a-tetrahydro-5H-pyrrolo[3,4-d]-
isoxazole-5-carboxylate [(؎)-25]: To a solution of (Ϯ)-24 (1.4 g,
4.8 mmol) in dioxane (12 mL) was added NaOH (1 , 24 mL), and
the mixture was stirred and heated at 80 °C for 24 h. The aqueous
phase was washed twice with Et2O and then made acidic with 2 
HCl and extracted with AcOEt (4ϫ3 mL). The organic phase was
dried (Na2SO4), the solvent was evaporated under reduced pres-
sure, and the crude material was recrystallized (AcOEt) to give (Ϯ)-
25 (1.01 g, 92% yield). White prisms, m.p. 128–130 °C, Rf (CHCl3/
MeOH, 9:1 + 1 % CF3COOH) = 0.84. 1H NMR (300 MHz,
CDCl3): δ = 1.43 (s, 9 H), 3.38 (ddd, J = 3.3, 8.4, 8.4 Hz, 1 H),
3.52–79 (m, 3 H), 3.90 (br. d, J = 11.0 Hz, 1 H), 5.19 (ddd, J =
3.3, 6.2, 7.7 Hz, 1 H). C10H16N2O4 (228.25): calcd. C 52.62, H 7.07,
N 12.27; found C 52.69, H 7.14, N 11.82.
(؎)-22a: White prisms from EtOH/Et2O, m.p. 124–126 °C, Rf
(CH2Cl2/MeOH, 95:5) = 0.19. 1H NMR (300 MHz, CDCl3): δ =
1.31 (t, J = 7.1 Hz, 3 H), 1.94–2.12 (m, 2 H), 2.60 (ddd, J = 2.4,
13.2, 13.2 Hz, 1 H), 2.70 (br. s, 1 H), 2.98 (ddd, J = 4.1, 4.1,
13.2 Hz, 1 H), 3.07 (dd, J = 3.2, 15.2 Hz, 1 H), 3.32 (ddd, J = 8.2,
8.2, 8.2 Hz, 1 H), 3.39–3.48 (m, 1 H), 4.24–4.34 (m, 2 H), 4.35–
4.42 (m, 1 H). C9H14N2O3 (198.22): calcd. C 54.53, H 7.12, N
14.13, found C 54.72, H 7.15, N 13.98.
(؎)-22b: Yellow oil. Rf (CH2Cl2/MeOH, 95:5) = 0.28. 1H NMR
(300 MHz, CDCl3): δ = 1.36 (t, J = 7.0 Hz, 3 H), 1.92–2.05 (m, 1
H), 2.06–2.16 (m, 1 H), 2.30 (br. s, 1 H), 2.65 (dd, J = 8.4, 11.7 Hz,
1 H), 2.79 (ddd, J = 4.0, 10.6, 12.5 Hz, 1 H), 2.94 (ddd, J = 4.0,
4.0, 12.5 Hz, 1 H), 3.22–3.38 (m, 2 H), 4.26–4.38 (m, 2 H), 4.64–
4.72 (m, 1 H). C9H14N2O3 (198.22): calcd. C 54.53, H 7.12, N
14.13; found C 54.89, H 7.27, N 13.74.
(؎)-4,5,6,6a-Tetrahydro-3aH-pyrrolo[3,4-d]isoxazol-3-ol [(؎)-12]:
Compound (Ϯ)-25 (1.01 g, 4.42 mmol) was treated with
CF3COOH (30% CH2Cl2 solution, 11 mL, 44.2 mmol) at 0 °C. The
solution was stirred at room temp. for 3 h. The volatiles were re-
moved under vacuum, and the residue was dissolved in H2O and
purified by cation-exchange column chromatography with Am-
berlite IR-120 (H+ form), as reported above. The residue was taken
up with ethanol, filtered, washed sequentially with ethanol and
Et2O and dried in vacuo at 50 °C to give amino acid (Ϯ)-12
(0.323 g, 57% yield). White prisms, m.p. dec. 180 °C, Rf (nBuOH/
H2O/CH3COOH, 4:2:1) = 0.37. 1H NMR (300 MHz, [D6]DMSO
+ 1 drop CF3COOD): δ = 3.32–3.44 (m, 2 H), 3.48–3.58 (m, 2 H),
3.65 (ddd, J = 1.9, 8.3, 8.3 Hz, 1 H), 5.26 (ddd, J = 1.5, 5.4, 8.3 Hz,
1 H). C5H8N2O2 (128.13): calcd. C 46.87, H 6.29, N 21.86; found
C 46.68, H 6.42 N 21.52.
(؎)-3a,4,5,6,7,7a-Hexahydroisoxazolo[5,4-c]pyridine-3-carboxylic
Acid [(؎)-10]: To a solution of (Ϯ)-22a (0.182 mg, 0.92 mmol) in
EtOH (3 mL) was added NaOH (1  1.4 mL), and the mixture was
stirred at room temp. for 12 h. After removing the solvent under
reduced pressure, the solution was made acidic with 2  HCl and
submitted to cation-exchange column chromatography with Am-
berlite IR-120 (H+ form) following the procedure described above.
The residue was taken up with methanol, filtered, washed sequen-
tially with methanol and Et2O and dried in vacuo at 50 °C to give
amino acid (Ϯ)-10 (0.069 g, 44% yield). White prisms, m.p. dec.
1
105–110 °C, Rf (nBuOH/H2O/CH3COOH, 4:2:1) = 0.26. H NMR
(300 MHz, D2O): δ = 1.55 (dddd, J = 4.4, 8.4, 9.5, 15.4 Hz, 1 H),
2.05 (dddd, J = 3.7, 6.9, 6.9, 15.4 Hz, 1 H), 2.93 (ddd, J = 3.7, 9.5,
13.2 Hz, 1 H), 3.08 (ddd, J = 4.4, 6.9, 13.2 Hz, 1 H), 3.29 (dd, J =
2.9, 14.7 Hz, 1 H), 3.36–3.45 (m, 1 H), 3.54 (dd, J = 2.2, 14.7 Hz,
1 H), 4.62–4.68 (m, 1 H). C7H10N2O3 (170.17): calcd. C 49.41, H
5.92, N 16.46; found C 49.12, H 6.06, N 16.14.
5-tert-Butyl 3-Ethyl (؎)-3a,4,6,6a-Tetrahydro-5H-pyrrolo[3,4-d]-
isoxazole-3,5-dicarboxylate [(؎)-26]: To a solution of 1-(tert-
butoxycarbonyl)-3-pyrroline 23 (1 g, 5.9 mmol) in AcOEt (20 mL)
was added ethyl 2-chloro-2-(hydroxyimino)acetate (1.79 g,
11.8 mmol) and solid NaHCO3 (5 g). The mixture was vigorously
stirred for 5 d. The progress of the reaction was monitored by TLC
(petroleum ether/AcOEt, 4:1). H2O (10 mL) was added to the reac-
tion mixture, and the organic layer was separated. The aqueous
phase was further extracted with AcOEt (5 mL), and the combined
organic layers were dried with anhydrous Na2SO4. The crude mate-
rial, obtained after evaporation of the solvent, was purified by col-
umn chromatography (silica gel, petroleum ether/AcOEt, 4:1) to
give compound (Ϯ)-26 (0.705 g, 42% yield). Yellow oil, Rf (petro-
(؎)-3a,4,5,6,7,7a-Hexahydroisoxazolo[4,5-c]pyridine-3-carboxylic
Acid [(؎)-11]: Regioisomer (Ϯ)-22b (0.250 g, 1.26 mmol) was
treated as described for (Ϯ)-22a to give amino acid (Ϯ)-11 (0.102 g,
48% yield). White prisms, m.p. dec. 119–121 °C, Rf (nBuOH/H2O/
1
CH3COOH, 4:2:1) = 0.30. H NMR (300 MHz, D2O): δ = 2.06–
2.14 (m, 2 H), 3.0–3.22 (m, 3 H), 3.37 (dd, J = 6.6, 13.5 Hz, 1 H),
3.60 (ddd, J = 6.6, 8.0, 8.0 Hz, 1 H), 4.75 (ddd, J = 4.0, 4.0, 8.0 Hz,
1 H). C7H10N2O3 (170.17): calcd. C 49.41, H 5.92, N 16.46; found
C 49.04, H 5.99, N 16.34.
1
leum ether/AcOEt, 4:1) = 0.18. H NMR (300 MHz, CDCl3): δ =
1.35 (t, J = 7.2 Hz, 3 H), 1.42 (s, 9 H), 3.41–3.54 (m, 2 H), 3.79–
4.08 (m, 1 H), 3.80–4.10 (m, 2 H), 4.35 (q, J = 7.2 Hz, 2 H), 5.32
(dd, J = 5.4, 9.6 Hz, 1 H). C13H20N2O5 (284.31): calcd. C 54.92, H
7.09, N 9.85; found C 54.90, H 7.19, N 9.49.
tert-Butyl (؎)-3-Bromo-3a,4,6,6a-tetrahydro-5H-pyrrolo[3,4-d]isox-
azole-5-carboxylate [(؎)-24]: To a solution of 1-(tert-butoxycar-
bonyl)-3-pyrroline 23 (1 g, 5.9 mmol) in AcOEt (20 mL) was added
dibromoformaldoxime (2.4 g, 11.8 mmol) and solid NaHCO3 (5 g).
The mixture was vigorously stirred for 3 d. The progress of the
reaction was monitored by TLC (petroleum ether/AcOEt, 7:3).
H2O (10 mL) was added to the reaction mixture, and the organic
(؎)-5-(tert-Butoxycarbonyl)-4,5,6,6a-tetrahydro-3aH-pyrrolo[3,4-d]-
isoxazole-3-carboxylic Acid [(؎)-27]: To a solution of (Ϯ)-26
(0.705 g, 2.48 mmol) in EtOH (8 mL) was added NaOH (1 ,
3.7 mL), and the mixture was stirred at room temp. for 3 h. The
Eur. J. Org. Chem. 2006, 5533–5542
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5539

P. Conti, et al.
FULL PAPER
volatiles were removed under vacuum, and the aqueous phase was
extracted with Et2O (2ϫ3 mL), made acidic with 2  HCl and ex-
tracted with AcOEt (4 ϫ 3 mL). The organic phase was dried
(Na2SO4), the solvent was evaporated under reduced pressure, and
the crude material was recrystallized (AcOEt) to give (Ϯ)-27
(0.55 g, 87 % yield). White prisms, m.p. dec. 134–137 °C, Rf
(CHCl3 /MeOH, 9:1 + 1 % CF3 COOH) = 0.34. 1 H NMR
(300 MHz, CDCl3): δ = 1.46 (s, 9 H), 3.40–3.59 (m, 2 H), 3.85–
4.20 (m, 3 H), 4.40–5.10 (m, 2 H), 4.75 (br. s, 1 H), 5.39 (dd, J =
5.2, 9.4 Hz, 1 H). C11H16N2O5 (256.26): calcd. C 51.56, H 6.29, N
10.93; found C 51.72, H 6.37, N 10.66.
H), 3.73 (d, J = 9.2 Hz, 1 H), 4.28–4.40 (m, 2 H), 4.45 (d, J =
4.0 Hz, 1 H), 5.36 (dd, J = 5.5, 9.2 Hz, 1 H). C9H13NO4 (199.20):
calcd. C 54.26, H 6.58, N 7.03; found C 54.38, H 6.73, N 6.85.
(؎)-30c: Recrystallizes from 2-propanol as white prisms, m.p.
55 °C, Rf (cyclohexane/AcOEt, 1:1) = 0.36, HPLC: petroleum
ether/AcOEt, 85:15, 1 mL/min., λ = 254 nm, tR = 14.72 min, 1H
NMR (300 MHz, CDCl3): δ = 1.37 (t, J = 7.3 Hz, 3 H), 1.59 (dddd,
J = 7.0, 9.2, 12.4, 12.4 Hz, 1 H), 1.78–1.92 (m, 1 H), 2.00 (dddd, J
= 2.6, 6.2, 6.2, 12.4 Hz, 1 H), 2.10–2-20 (m, 1 H), 2.54 (br. s, 1 H),
3.92 (dd, J = 7.7, 9.2 Hz, 1 H), 4.34 (q, J = 7.3 Hz, 2 H), 4.49 (ddd,
J = 6.2, 7.7, 9.2 Hz, 1 H), 5.23 (ddd, J = 1.1, 5.1, 9.2 Hz, 1 H).
C9H13NO4 (199.20): calcd. C 54.26, H 6.58, N 7.03; found C 54.45,
H 6.78, N 6.81.
(؎)-3-Carboxy-4,5,6,6a-tetrahydro-3aH-pyrrolo[3,4-d]isoxazol-5-
ium Trifluoroacetate [(؎)-13]: Compound (Ϯ)-27 (0.55 g,
1.93 mmol) was treated with CF3COOH (30% CH2Cl2 solution,
4.9 mL, 19.3 mmol) at 0 °C. The solution was stirred at room temp.
for 3 h. The volatiles were removed under vacuum, and the residue
was taken up with methanol, filtered, washed sequentially with
methanol and Et2O and dried in vacuo at 50 °C to give amino acid
(Ϯ)-13 as the corresponding trifluoroacetate (0.365 g, 70% yield).
White prisms, m.p. dec. 140–144 °C, Rf (nBuOH/H2O/CH3COOH,
4:2:1) = 0.27. 1H NMR (300 MHz, [D6]DMSO + 1 drop
CF3COOD): δ = 3.32–3.46 (m, 2 H), 3.56–3.70 (m, 2 H), 4.30 (m,
1 H), 5.47 (dd, J = 4.8, 9.5 Hz, 1 H). C8H9F3N2O5 (270.16): calcd.
C 35.57, H 3.36, N 10.37; found C 35.21, H 3.39, N 10.09.
(؎)-30d: Yellow oil, Rf (cyclohexane/AcOEt, 1:1) = 0.33, HPLC:
petroleum ether/AcOEt, 85:15, 1 mL/min, λ = 254 nm, tR
=
18.50 min, 1H NMR (300 MHz, CDCl3): δ = 1.36 (t, J = 7.0 Hz, 3
H), 1.35–1.46 (m, 1 H), 1.76–1.91 (m, 1 H), 1.92–2.02 (m, 2 H),
2.30 (br. s, 1 H), 3.89 (ddd, J = 2.2, 9.2, 9.2 Hz, 1 H), 4.17 (ddd, J
= 5.1, 5.1, 10.2 Hz, 1 H), 4.29–4.38 (m, 2 H), 5.02 (dd, J = 5.1,
9.2 Hz, 1 H). C9H13NO4 (199.20): calcd. C 54.26, H 6.58, N 7.03;
found C 54.28, H 6.77, N 6.71.
Ethyl (؎)-(3aS*,6S*,6aR*)-6-[(tert-Butoxycarbonyl)amino]-
4,5,6,6a-tetrahydro-3aH-cyclopenta[d]isoxazole-3-carboxylate [(؎)-
31a]: To a solution of (Ϯ)-30a (500 mg, 2.51 mmol) in CH2Cl2
(10 mL), cooled to 0 °C, were added triethylamine (0.524 mL,
3.765 mmol) and methanesulfonyl chloride (0.292 mL,
3.765 mmol). After stirring at room temp. for 30 min, the solution
was sequentially washed with 2  HCl, 5% aqueous NaHCO3 and
brine. The organic phase was dried with anhydrous Na2SO4, and
the solvent was evaporated under reduced pressure. The crude ma-
terial was dissolved in DMSO (7 mL), and NaN3 (1.63 g,
25.1 mmol) was added. The mixture was stirred and heated at 80 °C
for 5 h. The progress of the reaction was monitored by TLC (petro-
leum ether/AcOEt, 4:1). The reaction mixture was partitioned be-
tween H2O and Et2O, and the aqueous phase was further extracted
with Et2O. The combined organic layers were dried with anhydrous
Na2SO4, and the solvent was evaporated under reduced pressure.
Ethyl (؎)-(3aS*,6R*,6aR*)-6-Hydroxy-4,5,6,6a-tetrahydro-3aH-cy-
clopenta[d]isoxazole-3-carboxylate [(؎)-30a], Ethyl (؎)-
(3aR*,4S*,6aS*)-4-Hydroxy-4,5,6,6a-tetrahydro-3aH-cyclopenta-
[d]isoxazole-3-carboxylate [(؎)-30b], Ethyl (؎)-(3aR*,4R*,6aS*)-4-
Hydroxy-4,5,6,6a-tetrahydro-3aH-cyclopenta[d]isoxazole-3-carbox-
ylate [(؎)-30c], and Ethyl (؎)-(3aS*,6S*,6aR*)-6-hydroxy-4,5,6,6a-
tetrahydro-3aH-cyclopenta[d]isoxazole-3-carboxylate [(؎)-30d]: To
a solution of (Ϯ)-cyclopent-2-en-1-ol (Ϯ)-28 (3 g, 35.7 mmol) in
AcOEt (60 mL) was added ethyl 2-chloro-2-(hydroxyimino)acetate
(5.4 g, 35.7 mmol) and solid NaHCO3 (15 g). The mixture was vig-
orously stirred for 4 d. Another portion of ethyl 2-chloro-2-(hy-
droxyimino)acetate (5.4 g, 35.7 mmol) was added, and the mixture
was stirred for an additional 3 d. The progress of the reaction was
monitored by TLC (petroleum ether/AcOEt, 7:3). H2O (30 mL)
was added to the reaction mixture, and the organic layer was sepa-
rated. The aqueous phase was further extracted with AcOEt
(20 mL), and the combined organic layers were dried with anhy-
drous Na2SO4. The crude material, obtained after evaporation of
the solvent, was purified by flash chromatography (silica gel, petro-
leum ether/AcOEt, 7:3) to give 0.8 g of (Ϯ)-30a, 2 g of (Ϯ)-30b,
0.6 g of (Ϯ)-30c, 0.87 g of (Ϯ)-30d, 1.1 g of (Ϯ)-30a + (Ϯ)-30b,
0.55 g of (Ϯ)-30b + (Ϯ)-30c, and 0.2 g of (Ϯ)-30c + (Ϯ)-30d. Over-
all yield: 87%.
The desired azide was purified by column chromatography (silica
gel, petroleum ether/AcOEt, 9:1) to give 430 mg (1.92 mmol) of a
yellow oil, which was dissolved in EtOH (20 mL) and submitted to
catalytic hydrogenation in the presence of 5% Pd/C (80 mg) and
di-tert-butyl dicarbonate (628 mg, 2.88 mmol). The progress of the
reaction was monitored by TLC (petroleum ether/AcOEt, 4:1). Af-
ter 24 h, the catalyst was filtered off, and the solvent was evapo-
rated under reduced pressure. The crude material was purified by
column chromatography (silica gel, petroleum ether/AcOEt, 4:1)
and then recrystallized from 2-propanol to give compound (Ϯ)-31a
(444 mg, 1.49 mmol, 59% overall yield). White prisms, m.p. 93–
95 °C, Rf (cyclohexane/AcOEt, 7:3) = 0.40, 1H NMR (300 MHz,
CDCl3): δ = 1.32 (t, J = 7.0 Hz, 3 H), 1.40 (s, 9 H), 1.38–1.50 (m,
1 H), 1.77 (ddd, J = 5.9, 8.0, 13.9 Hz, 1 H), 1.87–2.04 (m, 2 H),
3.85 (dd, J = 8.5, 8.5 Hz, 1 H), 4.00–4.14 (m, 1 H), 4.22–4.34 (m,
2 H), 4.94–5.04 (m, 2 H). C14H22N2O5 (298.33): calcd. C 56.36, H
7.43, N 9.39; found C 56.49, H 7.64, N 9.06.
(؎)-30a: Yellow oil, Rf (cyclohexane/AcOEt, 1:1) = 0.47, HPLC:
petroleum ether/AcOEt, 85:15, 1 mL/min, λ = 254 nm, tR
=
1
8.07 min, H NMR (300 MHz, CDCl3): δ = 1.34 (t, J = 7.3 Hz, 3
H), 1.54 (dddd, J = 4.4, 7.3, 13.2, 13.2 Hz, 1 H), 1.79 (ddd, J =
1.1, 6.2, 13.2 Hz, 1 H), 1.96 (dd, J = 7.3, 13.2 Hz, 1 H), 2.22 (dddd,
J = 6.2, 8.8, 13.2, 13.2 Hz, 1 H), 2.31 (br. s, 1 H), 3.96 (dd, J =
8.8, 8.8 Hz, 1 H), 4.26–4.38 (m, 2 H), 4.36 (dd, J = 1.1, 4.4 Hz, 1
H), 4.97 (d, J = 8.8 Hz, 1 H). C9H13NO4 (199.20): calcd. C 54.26,
H 6.58, N 7.03; found C 54.28, H 6.69, N 6.73.
Ethyl (؎)-(3aS*,4R*,6aS*)-4-[(tert-Butoxycarbonyl)amino]-
4,5,6,6a-tetrahydro-3aH-cyclopenta[d]isoxazole-3-carboxylate [(؎)-
31b]: (Ϯ)-31b was prepared from (Ϯ)-30b in a comparable yield
to that described for (Ϯ)-31a. White prisms, m.p. 164–165 °C, Rf
(؎)-30b: Recrystallizes from 2-propanol as white prisms, m.p. 111–
112 °C, Rf (cyclohexane/AcOEt, 1:1) = 0.40, HPLC: petroleum
ether/AcOEt, 85:15, 1 mL/min, λ = 254 nm, tR = 9.46 min, 1H
NMR (300 MHz, CDCl3): δ = 1.36 (t, J = 7.0 Hz, 3 H), 1.64 (dddd,
1
(cyclohexane/AcOEt, 7:3) = 0.31, H NMR (300 MHz, CDCl3): δ
= 1.37 (t, J = 7.3 Hz, 3 H), 1.44 (s, 9 H), 1.55–1.62 (m, 1 H), 1.76–
J = 4.0, 7.0, 13.5, 13.5 Hz, 1 H), 1.81 (dd, J = 6.6, 13.5 Hz, 1 H), 1.90 (m, 1 H), 2.07 (ddd, J = 6.2, 6.2, 12.4 Hz, 1 H), 2.18 (dd, J =
2.13 (dd, J = 7.0, 13.5 Hz, 1 H), 2.20–2.33 (m, 1 H), 2.45 (br. s, 1 6.2, 14.3 Hz, 1 H), 4.08(dd, J = 8.4, 8.4 Hz, 1 H), 4.16–4.44 (m, 3
5540
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5533–5542

Synthesis of ∆2-Isoxazoline Amino Acids
H), 4.70–4.80 (m, 1 H), 5.24 (dd, J = 5.1, 8.4 Hz, 1 H). C14H22N2O5 (Ϯ)-16. White prisms, m.p. dec. 82–89 °C, Rf (nBuOH/H2O/
FULL PAPER
(298.33): calcd. C 56.36, H 7.43, N 9.39; found C 56.46, H 7.55, N
9.17.
CH3COOH, 4:2:1) = 0.35. 1H NMR (300 MHz, D2O): δ = 1.78–
1.90 (m, 2 H), 2.00–2.20 (m, 2 H), 3.85 (m, 1 H), 3.93 (d, J =
9.5 Hz, 1 H), 5.32 (ddd, J = 2.2, 5.1, 7.3 Hz, 1 H). C9H11F3N2O5
(284.19): calcd. C 38.04, H 3.90, N 9.86; found C 37.99, H 4.13, N
9.69.
Ethyl (؎)-(3aS*,4S*,6aS*)-4-[(tert-Butoxycarbonyl)amino]-4,5,6,6a-
tetrahydro-3aH-cyclopenta[d]isoxazole-3-carboxylate [(؎)-31c]: (Ϯ)-
31c was prepared from (Ϯ)-30c in a comparable yield to that de-
scribed for (Ϯ)-31a. White prisms, m.p. 86–88 °C, Rf (cyclohexane/
(؎)-(3aS*,6R*,6aR*)-3-Carboxy-4,5,6,6a-tetrahydro-3aH-cyclopen-
ta[d]isoxazole-6-ammonium Trifluoroacetate [(؎)-17]: (Ϯ)-17 was
prepared from (Ϯ)-31d in a comparable yield to that described for
(Ϯ)-16. White prisms, m.p. dec. 100–107 °C, Rf (nBuOH/H2O/
CH3COOH, 4:2:1) = 0.25. 1H NMR (300 MHz, D2O): δ = 1.70–
1.98 (m, 3 H), 2.02–2.40 (m, 1 H), 3.67 (m, 1 H), 4.06 (ddd, J =
3.7, 9.2, 9.2 Hz, 1 H), 5.07 (dd, J = 2.2, 9.2 Hz, 1 H). C9H11F3N2O5
(284.19): calcd. C 38.04, H 3.90, N 9.86; found C 37.85, H 4.02, N
9.49.
1
AcOEt, 7:3) = 0.35, H NMR (300 MHz, CDCl3): δ = 1.36 (t, J =
7.3 Hz, 3 H), 1.44 (s, 9 H), 1.70–1.82 (m, 2 H), 2.06–2.16 (m, 2 H),
3.78 (d, J = 9.2 Hz, 1 H), 4.14–4.21 (m, 1 H), 4.28–4.40 (m, 2 H),
4.60–4.80 (m, 1 H), 5.30 (ddd, J = 2.6, 4.4, 9.2 Hz, 1 H).
C14H22N2O5 (298.33): calcd. C 56.36, H 7.43, N 9.39; found C
56.38, H 7.60, N 9.05.
Ethyl (؎)-(3aS*,6R*,6aR*)-6-[(tert-Butoxycarbonyl)amino]-
4,5,6,6a-tetrahydro-3aH-cyclopenta[d]isoxazole-3-carboxylate [(؎)-
31d]: (Ϯ)-31d was prepared from (Ϯ)-30d in a similar yield as that
described for (Ϯ)-31a. White prisms, m.p. 108–109 °C, Rf (cyclohex-
Binding Assays: The receptor binding technique for determining
the affinities for GABAA and GABAB receptors was performed on
rat brain membrane preparations with either [3H]muscimol or [3H]-
GABA as the radioligands, as described previously.[22]
1
ane/AcOEt, 7:3) = 0.37, H NMR (300 MHz, CDCl3): δ = 1.34 (t,
J = 7.3 Hz, 3 H), 1.40 (s, 9 H), 1.60–1.78 (m, 2 H), 1.90–2.15 (m,
2 H), 3.89 (ddd, J = 1.8, 8.8, 8.8 Hz, 1 H), 3.99–4.12 (m, 1 H),
4.20–4.38 (m, 2 H), 4.60–4.80 (m, 1 H), 5.05 (d, J = 8.8 Hz, 1 H).
[3H]GABA Uptake Into Cortical Synaptosomes: Fresh cortex from
male Wistar rats (150–200 g) was homogenized with 0.4  sucrose
with a glass/teflon homogenizer. The homogenate was centrifuged
at 1000ϫg for 10 min, the pellet was discarded and the supernatant
was centrifuged at 40000ϫg for 20 min, and this washing step was
repeated once. The final pellet was homogenized in assay buffer
(120 m NaCl, 5 m KCl, 1.8 m CaCl2, 1.2 m MgSO4, 10 m
-glucose, 1.2 m NaH2PO4, 25 m NaHCO3, 10 m HEPES, pH
7.4 and 40 times the pellet volume). Aliquots (50 µL) of uptake
buffer/test compound solution and tissue (100 µL, ≈ 2.5 mg origi-
nal tissue/well) were added to 96-well plates and preincubated
2 min at 25 °C before adding 3H-GABA (50 µL, 1 nM tracer +
25 nM unlabelled GABA) and further incubating for 5 min at
25 °C. Samples were filtered directly onto Whatman GF/B glass
fibre filters under vacuum and immediately washed with 0.9%
NaCl (3ϫ0.5 sec.) with a Packard UniTomTec cell harvester.
Bound radioactivity was measured with a liquid scintillation
counter (Trilux short, Wallac) with a liquid scintillation cocktail
(Optiphase “Supermix”, Perkin–Elmer). Non-specific uptake was
defined as the uptake in the presence of 1 m GABA and ac-
counted for 4–5% of total uptake.
C
14H22N2O5 (298.33): calcd. C 56.36, H 7.43, N 9.39; found C
56.44, H 7.58, N 9.19.
(؎)-(3aS*,6S*,6aR*)-3-Carboxy-4,5,6,6a-tetrahydro-3aH-cyclo-
penta[d]isoxazole-6-ammonium Trifluoroacetate [(؎)-16]. Step A:
Compound (Ϯ)-31a (400 mg, 1.34 mmol) was dissolved in EtOH
(10 mL) and treated with aqueous K2CO3 (20%, 10 mL). The solu-
tion was stirred at room temp. overnight, the volatiles were re-
moved under vacuum, and the aqueous phase was extracted once
with CH2Cl2 (5 mL). The aqueous phase was then made acidic with
2  HCl and extracted with CH2Cl2 (4ϫ5 mL). The combined or-
ganic layers were dried with anhydrous Na2SO4, and the solvent
was evaporated under vacuum to give the crude carboxylic acid,
which was directly submitted to the next step.
Step B: The crude carboxylic acid obtained from the previous step
(293 mg, 1.085 mmol) was treated with CF3COOH (30% CH2Cl2
solution, 2.8 mL, 10.85 mmol) at 0 °C. The solution was stirred at
room temp. for 3 h. The volatiles were removed under vacuum, and
the residue was taken up with methanol, filtered, washed sequen-
tially with methanol and Et2O and dried in vacuo at 50 °C to give
amino acid (Ϯ)-16 as the corresponding trifluoroacetate (136 mg,
0.48 mmol, 36 % overall yield). White prisms, m.p. 132–137 °C
(dec.), Rf (nBuOH/H2O/CH3COOH, 4:2:1) = 0.21. 1H NMR
(300 MHz, D2O): δ = 1.37 (dddd, J = 7.1, 12.4, 12.4, 12.4 Hz, 1
H), 1.82 (dddd, J = 5.9, 8.4, 12.4, 12.4 Hz, 1 H), 1.95 (dd, J = 7.1,
12.4 Hz, 1 H), 2.04 (ddd, J = 6.2, 6.2, 12.4 Hz, 1 H), 3.58 (ddd, J
= 6.2, 6.2, 12.4 Hz, 1 H), 3.99 (dd, J = 8.4, 8.4 Hz, 1 H), 5.06 (dd,
J = 6.2, 8.4 Hz, 1 H). C9H11F3N2O5 (284.19): calcd. C 38.04, H
3.90, N 9.86; found C 37.69, H 3.80, N 9.54.
GABA Functional Assay in Rat Cortical Wedge: The rat cortical
wedge preparation was carried out according to a previously pub-
lished method.[6] The wedges were left for development of sponta-
neous activity for 2–3 h. Characterization of effects on the sponta-
neous activity was initiated when the spontaneous activity was
more than 30 spikes per 12 min and stable over a 30 min period.
The number of population responses (spikes) per 12 min was
counted. Compounds were applied in superfusion buffer, and the
wedges were superfused for 20 min. The number of spikes during
the last 12 min of drug application were counted, and the frequency
(spikes per min) were calculated. The relative frequency, calculated
as the ratio between frequencies in the presence and absence of
compound, were calculated. Compounds were tested alone or in
combination with 20 µ GABA (corresponding to EC20 for
GABA).
(؎)-(3aS*,4R*,6aS*)-3-Carboxy-4,5,6,6a-tetrahydro-3aH-cyclo-
penta[d]isoxazole-4-ammonium Trifluoroacetate [(؎)-14]: (Ϯ)-14
was prepared from (Ϯ)-31b in a similar yield as that described for
(Ϯ)-16. White prisms, m.p. dec. 132–138 °C, Rf (nBuOH/H2O/
CH3COOH, 4:2:1) = 0.35. 1H NMR (300 MHz, D2O): δ = 1.45–
1.52 (m, 1 H), 1.76–1.93 (m, 1 H), 2.00–2.14 (m, 2 H), 3.72 (m, 1
H), 4.05 (dd, J = 8.8, 8.8 Hz, 1 H), 5.27 (dd, J = 4.8, 8.8 Hz, 1 H).
C9H11F3N2O5 (284.19): calcd. C 38.04, H 3.90, N 9.86, found C
37.74, H 3.99, N 9.51.
Theoretical Calculations: The conformational space of all the ana-
lyzed compounds was explored with the DFT/B3LYP approach at
the 6-31G(d) level as implemented in GAUSSIAN03.[23] We fully
optimized all the starting geometries deriving from the pseudorota-
tional path of the five- and six-membered carbocyclic rings and, in
(؎)-(3aS*,4S*,6aS*)-3-Carboxy-4,5,6,6a-tetrahydro-3aH-cyclo-
penta[d]isoxazole-4-ammonium Trifluoroacetate [(؎)-15]: (Ϯ)-15 was some instances, from rotation around the exocyclic single bonds.
prepared from (Ϯ)-31c in a comparable yield to that described for The energies of the conformations were recalculated in a polariz-
Eur. J. Org. Chem. 2006, 5533–5542
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5541

P. Conti, et al.
FULL PAPER
able conductor-like solvation model (C-PCM)[26] to obtain values
compatible with aqueous solutions.
[17]
[18]
[19]
[20]
A. P. Kozikowski, M. Adamcz, J. Org. Chem. 1983, 48, 366–
372.
Y. Tsuzuki, K. Chiba, K. Mizuno, K. Tomita, K. Suzuki, Tet-
rahedron: Asymmetry 2001, 12, 2989–2997.
D. P. Curran, A. G. Scott, C. Sung-Mo, Heterocycles 1993, 35,
1371–1395.
P. Conti, M. De Amici, S. Joppolo di Ventimiglia, T. B.
Stensbøl, U. Madsen, H. Bräuner-Osborne, E. Russo, G.
De Sarro, G. Bruno, C. De Micheli, J. Med. Chem. 2003, 46,
3102–3108.
Acknowledgments
This research has been financially supported by a grant from the
University of Milan (FIRST, 2004). Marianne Faber and Charlotte
Fredsøe Hundahl are acknowledged for their skilled technical as-
sistance.
[21]
P. Caramella, G. Cellerino, Tetrahedron Lett. 1974, 2, 229–232.
[22] B. Frølund, L. S. Jensen, L. Guandalini, C. Canillo, H. T. Ves-
tergaard, U. Kristiansen, B. Nielsen, T. B. Stensbøl, U.
Madsen, P. Krogsgaard-Larsen, T. Liljefors, J. Med. Chem.
2005, 48, 427–439.
[1] G. Sperk, S. Furtinger, C. Schwarzer, S. Pirker Advances in Ex-
perimental Medicine & Biology. 2004, 548, 92–103.
[2] G. B. Richerson, Y. Wu, Advances in Experimental Medicine &
Biology 2004, 548, 76–91.
[23] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr., T.
Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar,
J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N.
Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y.
Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P.
Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R.
Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R.
Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morok-
uma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzew-
ski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas, D. K.
Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V.
Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B.
Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L.
Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A.
Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, GAUSSIAN 03,
revision C.02, Gaussian, Inc., Wallingford, CT, 2004.
[24] B. Frølund, L. Tagmose, T. Liljefors, T. Bryan Stensbøl, C.
Engblom, U. Kristiansen, P. Krogsgaard-Larsen, J. Med.
Chem. 2000, 43, 4930 –4933.
[3] R. B. Lydiard, J. Clin. Psychiatry 2003, 64, 21–27 (Suppl. 3).
[4] I. W. Tremont-Lukats, C. Megeff, M. M. Backonja, Drugs
2000, 60, 1029–1052.
[5] S. J. Enna, N. G. Bowery (Eds.), The GABA Receptors, The Hu-
mana Press: Totowa, NJ, 1997.
[6] P. Krogsgaard-Larsen, B. Frolund, U. Kristiansen, K. Fryden-
vang, B. Ebert, Eur. J. Pharm. Sci. 1997, 5, 355–384.
[7] M. Chebib, G. A. R. Johnston, J. Med. Chem. 2000, 43, 1427–
1447.
[8] P. Krogsgaard-Larsen, G. A. R. Johnston, D. R. Curtis, C. J. A.
Game, R. M. McCulloch, J. Neurochem. 1975, 25, 803–809.
[9] P. Krogsgaard-Larsen, G. A. R. Johnston, D. Lodge, D. R.
Curtis, Nature 1977, 268, 53–55.
[10] J. Faulhaber, A. Steiger, M. Lancel, Psychopharmacology
(Berlin) 1997, 130, 285–291.
[11] S. I. Stο´rustovu, B. Ebert, J. Pharmacol. Exp. Ther. 2006, 316,
1351–1359.
[12] P. Krogsgaard-Larsen, G. A. R. Johnston, J. Neurochem. 1975,
25, 797–802.
[13] E. Falk, J. Perregaard, B. Frølund, B. Søkilde, A. Buur, L. M.
Hansen, K. Frydenvang, L. Brehm, T. Bolvig, O. M. Larsson,
C. Sanchez, H. S. White, A. Schousboe, P. Krogsgaard-Larsen,
J. Med. Chem. 1999, 42, 5402–5414.
[14] R. Nordmann, P. Graff, R. Maurer, B. H. Gähwiler, J. Med.
Chem. 1985, 28, 1109–1111.
[15] P. Krogsgaard-Larsen, L. Nielsen, E. Falch, D. R. Curtis, J.
Med. Chem. 1985, 28, 1612–1617.
[25] Synthesis and ring-opening reactions of the diastereoisomeric
cis- and trans-epoxides derived from 3-(benzyloxy)cyclopentene
and 2-(benzyloxy)-2,5-dihydrofuran P. Crotti, V. Di Bussolo, L.
Favero, F. Macchia, M. Pineschi, Eur. J. Org. Chem. 1998,
1675–1686.
[26] V. Barone, M. Cossi, J. Phys. Chem. A 1998, 102, 1995–2001.
[16] D. M. Vyas, Y. Chiang, T. W. Doyle, Tetrahedron Lett. 1984,
25, 487–490.
Received: July 20, 2006
Published Online: October 18, 2006
5542
www.eurjoc.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2006, 5533–5542