New analogues of epiboxidine incorporating the 4,5-dihydroisoxazole nucleus: Synthesis, binding affinity at neuronal nicotinic acetylcholine receptors, and molecular modeling investigations

Article abstract of DOI:10.1002/cbdv.200800077

A group of novel 4,5-dihydro-3-methylisoxazolyl derivatives, structurally related to epiboxidine (= 1R,4S,6S)-6-(3-methylisoxazol-5-yl)-7-azabicyclo[2.2.1]heptane), was prepared via 1,3-dipolar cyclo-addition of acetonitrile oxide to different olefins. Target compounds 1a and 1b, 2a and 2b, 3, 4, and 5 were tested for affinity at neuronal nicotinic heteromeric (α4β2) and homomeric (α7) acetylcholine receptors. Notably, diastereoisomers 1a and 1b were characterized by a massive drop of the affinity at the α4β2 subtypes (K_i values spanning the range 4.3-126 μM), when compared with that of epiboxidine (K_i= 0.6 nM). Therefore, the replacement of the 3-methylisoxazole ring of epiboxidine with the 4,5-dihydro-3-methylisoxazole nucleus is detrimental for the affinity at α4β2 receptors. A comparable lack of affinity/selectivity for the two nAChR subtypes under study was evidenced for the remaining epiboxidine-related dihydroisoxazole derivatives 2a and 2b, and 3-5. Diastereoisomers 1a and 1b, and spirocyclic derivative 3 were docked into molecular models of the receptor subtypes under study, and their binding mode was compared with that of reference ligands endowed with high binding affinity.

Full text of DOI:10.1002/cbdv.200800077

244
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
New Analogues of Epiboxidine Incorporating the 4,5-Dihydroisoxazole
Nucleus: Synthesis, Binding Affinity at Neuronal Nicotinic Acetylcholine
Receptors, and Molecular Modeling Investigations
by Clelia Dallanoce*a), Pietro Magronea), Paola Bazzaa), Giovanni Graziosoa), Luca Rizzia),
Loredana Rigantib), Cecilia Gottib), Francesco Clementib), Karla Frydenvangc), and
Marco De Amicia)
a
`
) Istituto di Chimica Farmaceutica e Tossicologica ꢀPietro Pratesiꢁ, Universita degli Studi di Milano,
Via Mangiagalli 25, I-20133 Milano (phone: þ39-02-50319327; fax: þ39-02-50319326;
e-mail: clelia.dallanoce@unimi.it)
b) CNR, Istituto di Neuroscienze, Farmacologia Cellulare e Molecolare e Dipartimento Farmacologia,
`
Chemioterapia e Tossicologia Medica, Universita degli Studi di Milano, Via Vanvitelli 32, I-20129 Milano
c) Department of Medicinal Chemistry, Faculty of Pharmaceutical Sciences, University of Copenhagen,
Universitetsparken 2, DK-2200 Copenhagen
A group of novel 4,5-dihydro-3-methylisoxazolyl derivatives, structurally related to epiboxidine (¼
(1R,4S,6S)-6-(3-methylisoxazol-5-yl)-7-azabicyclo[2.2.1]heptane), was prepared via 1,3-dipolar cyclo-
addition of acetonitrile oxide to different olefins. Target compounds 1a and 1b, 2a and 2b, 3, 4, and 5 were
tested for affinity at neuronal nicotinic heteromeric (a4b2) and homomeric (a7) acetylcholine receptors.
Notably, diastereoisomers 1a and 1b were characterized by a massive drop of the affinity at the a4b2
subtypes (Ki values spanning the range 4.3–126 mm), when compared with that of epiboxidine (Ki ¼
0.6 nm). Therefore, the replacement of the 3-methylisoxazole ring of epiboxidine with the 4,5-dihydro-3-
methylisoxazole nucleus is detrimental for the affinity at a4b2 receptors. A comparable lack of affinity/
selectivity for the two nAChR subtypes under study was evidenced for the remaining epiboxidine-related
dihydroisoxazole derivatives 2a and 2b, and 3–5. Diastereoisomers 1a and 1b, and spirocyclic derivative 3
were docked into molecular models of the receptor subtypes under study, and their binding mode was
compared with that of reference ligands endowed with high binding affinity.
Introduction. – Neuronal nicotinic acetylcholine receptors (nAChRs) belong to a
heterogeneous family of ligand-operated ion channels differently expressed in the
central and peripheral nervous systems [1–3]. The receptors are composed of five
separate transmembrane proteins (subunits), each containing a long N-terminal
extracellular domain, four membrane spanning a helices (M1–M4), a large cytoplas-
matic loop linking M3 and M4 sequences, and a short C-terminal extracellular domain
[4]. The resulting transmembrane ion channel is assembled in a pentameric
architecture to form a pore which is permeable to cations such as Naþ, Kþ, or Ca2þ
.
Of the 17 different nAChR subunits which have been cloned so far, twelve (a2 to
a10, and b2 to b4) were found in the neuronal receptors of vertebrates [5]. Different
subunit combinations characterize the nicotinic receptors expressed within the diverse
areas of the central nervous system (CNS) [6][7], and affect their biophysical and
functional properties, such as ion selectivity, conductance, mean open-channel time,
rate of desensitization, as well as the sensitivity to neurotoxins [4]. On the other hand,
ꢂ 2009 Verlag Helvetica Chimica Acta AG, Zꢃrich

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
245
ongoing investigations on the pharmacology and neurobiology of CNS nAChRs
allowed a better understanding of their involvement in various neuropsychiatric
pathologies, such as Alzheimerꢁs and Parkinsonꢁs diseases, mild cognitive impairment
(MCI), Touretteꢁs syndrome, schizophrenia, depression, anxiety, attention-deficit
hyperactivity disorder (ADHD), and nicotine addiction [1].
More than 90% of the heteromeric channels localized in the CNS contain the a4
and b2 subunits, whereas the most abundant homomeric channel is the a7 pentamer
[3][8]. As a consequence, research efforts have mainly focused on the discovery of
selective a4b2 or a7 ligands, to ascertain the physiological and pathophysiological
relevance of these receptor subtypes. Even though a large number of nicotinic agonists
and noncompetitive antagonists has been reported, very few of them are subtype-
selective [9].
However, over recent years, the advancement of potentially useful nAChR-based
therapeutics took advantage of the availability of molecular models for the different
nicotinic channels and the elaboration of computational protocols capable to
quantitatively estimate the binding free energy of selective ligands. Interesting
examples of this strategy are represented by a comparative theoretical study performed
on a set of selective agonists of both a4b2 and a7 nAChR subtypes [10], and by the
results on a new molecular model of the homomeric a7 channel, developed by our
research group, coupled with a computational method that allowed us to reliably
evaluate the ligand-receptor recognition process of a7 nAChR agonists [11].
As far as the nAChR ligands selectively activating the a4b2 subtype are considered,
a significant impulse to their development was given by the discovery of epibatidine
(Fig. 1), a toxin isolated from the skin of the poisonous Ecuadorian frog Epipedobates
tricolor [12], which possesses an analgesic potency ca. 100 times higher than that of
morphine. The potent analgesic activity of epibatidine, which is 30 times higher than
that of nicotine, is due to its high affinity for the a4b2 nAChR subtype [13].
Unfortunately, the therapeutic use of epibatidine as an analgesic drug is hampered by
its very narrow therapeutic index, due to the inability to discriminate among the
different nAChR subtypes [14]. Nonetheless, epibatidine has been used as a lead for
the synthesis of analogues provided with an improved pharmacological profile [15]. A
number of modifications to the structural elements of epibatidine, i.e., functionalization
and enlargement of the bicyclic skeleton, variation of the substituent at the N-atom
and/or its position within the bicyclic structure [2][4], replacement of the pyridyl ring
with different heteroaryl moieties, has been set forth to yield a variety of novel
epibatidine-related compounds. Among them, (ꢀ)-epiboxidine (Fig. 1), characterized
by the presence of a 3-methylisoxazole ring, emerged as a potent nicotinic receptor
agonist, tenfold less potent than epibatidine as antinociceptive agent but about 20-fold
less toxic [16]. More recently, homo-epiboxidine and a series of epiboxidine
homologues were synthesized and tested in binding as well as cultured cells assays [17].
To explore the relevance of the 3-methylisoxazole moiety of epiboxidine in the
interaction with nAChRs, we prepared and tested the two stereoisomeric 4,5-dihydro-
3-methylisoxazole analogues 1a and 1b, and their structurally related tertiary alcohols
2a and 2b (Fig. 1). Further compounds bearing the 4,5-dihydro-3-methylisoxazole
nucleus were investigated, i.e., derivatives 3 and 4 (Fig. 1), in which the 7-
azabicyclo[2.2.1]heptane system and the 4,5-dihydroisoxazole ring of isomers 1 are

246
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
Fig. 1. Structures of model and target derivatives
either spiro-linked or fused with the azanorbornane core. Finally, to evaluate the
impact of the 7-azabicyclo[2.2.1]heptane skeleton on the affinity profile, we removed
the ethylene bridge of 4 to give its simplified analogue 5 (Fig. 1).
This study aimed at further deepening the role exerted by the 4,5-dihydroisoxazole
ring in the process of molecular recognition/activation of nAChR subtypes [18][19].
Here, we describe the synthesis of the racemic forms of compounds 1–5, the estimation
of their binding affinity at a4b2 and a7 nAChR subtypes, and the qualitative evaluation
of the biological results by means of a molecular-modeling analysis.
Results and Discussion. – The synthetic route to the closest analogues of
epiboxidine 1a and 1b, and 2a and 2b is depicted in Schemes 1 and 2. Known tert-
butyl 2-oxo-7-azabicyclo[2.2.1]heptane-7-carboxylate (6) [20] was reacted with vinyl-
magnesium bromide to yield alcohol 7 exclusively. The relative configuration of 7 was
assigned on the assumption that the nucleophile approaches from the least hindered
exo-face. A comparable result was obtained with ketones structurally related to 6 [21].
Alcohol 7 underwent a 1,3-dipolar cycloaddition reaction with in situ generated
acetonitrile oxide (Scheme 1) to give a 2 :3 mixture (HPLC analysis, 65% overall yield)
of stereoisomeric dihydroisoxazoles 9a and 9b. 1-Chloroacetaldehyde oxime (8) [22],
the stable precursor of the 1,3-dipole acetonitrile oxide, was prepared by treating
acetaldehyde oxime with benzyl(trimethyl)ammonium tetrachloroiodate (BTMA·
ICl4), according to a recently published procedure [23]. The two cycloadducts were
easily separated by column chromatography, and their relative configurations were

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
247
unequivocally assigned by an X-ray analysis on isomer 9b, as detailed in Fig. 2. The
desired final compounds 2a and 2b were obtained as crystalline salts after standard
removal of the N-Boc protection, followed by treatment with fumaric acid (Scheme 1).
Fig. 2. Molecular structure (ORTEP-3) [24] of compound 9b illustrating the relative configuration.
Displacement ellipsoids of the non-H-atoms are shown at 50% probability level.
Intermediate alcohols 9a and 9b were then transformed into the corresponding
deoxygenated analogues 11a and 11b via xanthates 10a and 10b [25], respectively, by
applying the classic Barton–McCombie protocol [26] (Scheme 2). The spatial arrange-
ment of the dihydroisoxazole moiety was unaffected by the chemical treatments as
1
testified by the maintenance of the H-NMR patterns on passing from 9a to 11a and
from 9b to 11b, respectively. The subsequent preparation of 1a and 1b was achieved
according to the reactions previously applied to their analogues 2a and 2b.
Finally, the 4,5-dihydro-3-methylisoxazolyl derivatives 3, 4, and 5 were synthesized
according to the three reaction sequences illustrated in Scheme 3. Alkene 12 [18], tert-
butyl 7-azabicyclo[2.2.1]hept-2-ene-7-carboxylate (13) [27], and tert-butyl 2,5-dihydro-
pyrrole-1-carboxylate 14 [28] were all synthesized according to literature methods.
Olefins 12, 13, and 14 underwent cycloaddition with acetonitrile oxide to yield the
expected cycloadducts 15, 16, and 17, respectively. The reactive olefin 13, stirred with 8
and Et3N in CH2Cl2 at room temperature, provided 16 in 87% yield. Conversely, the
sluggish dipolarophiles 12 and 14 needed to be refluxed in ClCH2CH2Cl for a few days
with time-to-time addition of further amounts of chloroacetaldehyde oxime and Et3N.
Cycloadducts 15 and 17 were hence isolated in 11 and 27% yields, respectively.

248
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
Scheme 1
a) Vinylmagnesium bromide (1m in THF), 08. b) Et3N, CH2Cl2. c) 30% CF3COOH, CH2Cl2.
As expected, acetonitrile oxide added to the less hindered face of bicyclic olefin 13
yielding the exo-cycloadduct 16 exclusively. The observed exo-facial selectivity is
controlled by the endo-pyramidalization of the olefinic H-atoms in the dipolarophile
[18][29]. The structural assignment of compound 16 was based on the comparison of
1
relevant H-NMR data with those of closely related structural analogues [18]. By
taking into account the stereochemical outcome of reactions carried out on olefins
structurally related to 12 [18][30], we assigned the exo-configuration even to the
spirocyclic 4,5-dihydroisoxazole 15, isolated as single isomer from the cycloaddition
step. Sequential treatment of intermediates 15, 16, and 17 with CF3COOH and fumaric
acid provided the expected epiboxidine analogues 3, 4, and 5, respectively, as crystalline
derivatives.
The target compounds 1a and 1b, 2a and 2b, 3, 4, and 5 were assessed for binding
affinity at rat a4b2 and a7 nAChR subtypes, using [3H]epibatidine and [125I]-a-
bungarotoxin as radioligands. The Ki values were calculated from the competition
curves by means of the LIGAND program [31]. The data reported in the Table
evidence that the replacement of the 3-methylisoxazole ring of epiboxidine with the
4,5-dihydro-3-methylisoxazole nucleus causes a sharp drop in the affinity at the a4b2
nAChRs. Indeed, the two stereoisomers 1a and 1b show an affinity, at this receptor
subtype, four orders of magnitude lower than that reported for epiboxidine (Ki values 9
and 4.3 mm, resp., vs. 0.6 nm).
A parallel trend, with a further loss in the binding affinity, was observed on passing
to the OH-containing analogues 2a and 2b (Ki values 46 and 126 mm, resp.).

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
249
Scheme 2
a) NaH, CS2/MeI, THF, 08. b) 2,2’-Azobis[isobutyronitrile] (AIBN), Bu3SnH, toluene, reflux. c) 30%
CF3COOH, CH2Cl2.
Scheme 3
a) Et3N, CH2Cl2. b) Et3N, ClCH2CH2Cl. c) 30% CF3COOH, CH2Cl2.

250
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
Table. Binding Affinities (Ki [mm]) of 1a, 1b, 2a, 2b, 3, 4, and 5 to a4b2 and a7 nAChR Subtypes
a4b2
[3H]Epibatidine
Ki [mm] (% CV)
a7
[
125I]-a-BgTx
Ki [mm] (% CV)
Epibatidine
a-BgTx
0.000032
n.d.a)
0.0008
0.0009
(ꢀ)-Epiboxidine
0.0006b)
9 (16)
n.d.c)
1a
1b
2a
2b
3
32 (42)
40 (36)
77(33)
32 (23)
15 (19)
>500
4.3 (17)
46 (22)
126 (25)
4 (19)
197 (32)
83 (36)
4
5
13 (33)
a) Not determined. b) See [14]. c) Unknown value.
Unfortunately, we are not able to comment on the results obtained with the homomeric
a7 nAChRs, since the data on epiboxidine, to the best of our knowledge, have never
been reported. We can only notice that the two couples of structurally related
diastereoisomers are provided with a comparable affinity in the micromolar range (Ki
values 32–77 mm). Indeed, the overall affinity profile of compounds 1a and 1b, and 2a
and 2b appears to be marginally influenced by the relative configurations at the two
stereogenic centers.
On the other hand, either the spiro-linkage or the fusion of the 4,5-dihydro-3-
methylisoxazole and the azanorbornane component parts turned out to be unfruitful.
Compound 3 is a nonselective ligand with a low binding affinity at both a4b2 and a7
subtypes (Ki values 4 and 15 mm, resp.), whereas derivative 4 is characterized by an
affinity at both receptors significantly lower than that displayed by 3 (Ki values 197 and
>500 mm, resp.). Finally, removal of the ethylene bridge on passing from derivative 4 to
its analogue 5 causes an increase in the distance between the cationic site, i.e., the NH
group, and the H-bond acceptor site, i.e., the 4,5-dihydroisoxazole moiety, which
increased appreciably the affinity at the a7 subtype and reverted the a4b2/a7 selectivity
of the related compounds.
Three selected ligands, i.e., the two isomers 1a and 1b, and the spirocyclic derivative
3, were inspected by means of a molecular-modeling analysis. To this end, we used our
model of the a7 nAChRs [11] and a recently developed model of the a4b2 nAChRs
[32]. We chose (ꢁ)-epibatidine (Fig. 3,c and d) and the yet unknown enantiomer of
epiboxidine (Fig. 3,a and b) having the same absolute configuration of (ꢁ)-epibatidine
as the reference ligands for the a7 and a4b2 receptors, respectively. Similarly, we
docked into the models the enantiomers of compounds 1a and 1b (Fig. 3,a and c), and 3
(Fig. 3,b and d) whose absolute configuration matches that of (ꢁ)-epibatidine.
The high-affinity ligands for the a4b2 subtype, such as (ꢁ)-epibatidine and
epiboxidine, give rise to a network of most favorable molecular contacts with the
receptor complementary sites, such as i) a H-bond involving the NH group of the ligand
and the C¼O group of the Trp147 residue, ii) a network of cation–p interactions with

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
251
Fig. 3. Binding of compounds (ꢁ)-Epibatidine, Epiboxidine, 1a, 1b, and 3 in the a4b2-nAChR (a and b)
and a7-nAChR (c and d) active sites. Receptor model residues are depicted as stick model, colored
according to the atom code (C-atoms in white). Ligands are depicted as stick model, colored according to
the atom code: C-atoms in cyan for epiboxidine (a and b) and (ꢁ)-epibatidine (c and d), in magenta for
1a, green for 1b, and pink for 3.
the so-called ꢀaromatic boxꢁ (Trp55, Trp147, Tyr188, Tyr195), and iii) a tight
hydrophobic interaction of the ligand heteroaromatic moiety with the Leu119,
Phe117, and Val109 residues.
Conversely, a moderate or negligible binding affinity of a ligand is accounted for by
a loss in the number and nature of complementary molecular contacts with the receptor
and/or by the presence of unfavorable steric interactions. As far as the docking of

252
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
isomers 1a and 1b into the a4b2 receptor model is compared with that of epiboxidine,
replacement of the heteroaromatic isoxazole group with the flexible 4,5-dihydroisox-
azole ring brings about a distortion of the azanorbornane moiety, which is poorly
compatible for both isomers 1a and 1b with the sterically demanding hydrophobic triad
Leu119, Phe117, and Val109 (Fig. 3,a). In this respect, the docking of both isomers 1a
and 1b into the a7 receptor binding site is further negatively affected, since an even
more hindered Leu108 replaces the Val109 residue located on the a4b2 receptor
(Fig. 3,c).
As far as the binding of compound 3 is considered, the spirocyclic junction
engenders a steric clash between the 4,5-dihydroisoxazole moiety and the Leucine
residues found in both receptors (Leu119 for the a4b2 subtype, and Leu118 for the a7
subtype), as depicted in Fig. 3,b and d. Therefore, at variance with the reference
compounds, the azanorbornane nucleus of 3 is forced to adopt a conformation which
considerably hampers the interaction with the above mentioned ꢀaromatic boxꢁ
residues. Such a different orientation of ligand 3 in the receptor binding clefts emerges
as the leading cause of the drop in affinity at both receptor subtypes.
Conclusions. – In summary, the replacement of the 3-methylisoxazole ring of (ꢀ)-
epiboxidine with the 4,5-dihydro-3-methylisoxazole nucleus to generate the two
diastereoisomers 1a and 1b led to a dramatic loss of affinity at a4b2 nAChRs, thus
confirming the essential role played by the heteroaromatic fragment for an effective
interaction with the complementary receptor subsites. A molecular-modeling analysis
allowed us to clarify that the 4,5-dihydroisoxazole nucleus is precluding a favorable
insertion of both diastereoisomers in a hydrophobic cavity of the receptor. Comparable
results were also obtained when 1a and 1b were docked into a model for the a7
nAChRs. More drastic modifications of the epiboxidine molecular skeleton involving
insertion of the 4,5-dihydro-3-methylisoxazole moiety, i.e., compounds 3–5, gave also
rise to unsatisfactory results, since the three derivatives behaved as low-affinity and
nonselective nAChR ligands.
Experimental Part
General. Reagents and solvents for syntheses and crystallizations were reagent-grade and were used
without further purification. Solvents for extraction and flash chromatography (FC) were of technical
grade and were used without further purification. Solvents for HPLC were purchased from Sigma-
Aldrich. Compounds 6 [20], 12 [20], 13 [27], and 14 [28] were prepared according to reported procedures.
TLC: Al-backed sheets coated with silica gel 60 F254; spots were further evidenced by spraying with a dil.
alkaline KMnO4 soln. M.p.: Bꢀchi B 540 melting-point apparatus; uncorrected. HPLC: Jasco PU-980
equipped with a UV/VIS detector Jasco UV-975, Lichrosphere Si60 (Merck). 1H-NMR Spectra: Varian
Mercury 300 spectrometer, in CDCl3 solns. (unless otherwise indicated); d in ppm rel. to Me4Si, J in Hz.
Microanalyses (C,H,N) of new compounds agreed with the theoretical values within ꢀ0.4%.
Syntheses. tert-Butyl 2-Ethenyl-2-hydroxy-7-azabicyclo[2.2.1]heptane-7-carboxylate (7). tert-Butyl
2-oxo-7-azabicyclo[2.2.1]heptane-7-carboxylate (6 [20]; 450 mg, 2.13 mmol) was dissolved in anh. THF
(20 ml) and cooled to 08. Vinylmagnesium bromide (1.0m in THF, 4.3 ml, 4.3 mmol) was added dropwise
under a flow of anh. N2, and the mixture was stirred at 08 for 30 min. After addition of a sat. soln. NH4Cl
(20 ml), the mixture was extracted with AcOEt (3ꢂ10 ml). The collected org. layers were dried
(Na2SO4), filtered, and evaporated at reduced pressure. The residue was purified by FC (SiO2; petroleum
ether/AcOEt 4 :1) and then crystallized from hexane to afford pure 7 (380 mg, 75%). Colorless prisms.

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
253
M.p. 116–1178. Rf (petroleum ether/AcOEt 7:3) 0.64. 1H-NMR: 1.44 (s, 9 H); 1.50–1.70 (m, 3 H); 1.71–
1.89 (m, 2 H); 2.04–2.14 (m, 1 H); 2.22–2.37 (m, 1 H); 3.94 (br. s, 1 H); 4.21 (br. s, 1 H); 5.04 (dd, J ¼ 0.7,
10.6, 1 H); 5.22 (dd, J ¼ 0.7, 16.9, 1 H); 6.12 (dd, J ¼ 10.6, 16.9, 1 H). Anal. calc. for C13H21NO3 (239.31):
C 65.25, H 8.84, N 5.85; found: C 65.49, H 8.61, N 6.02.
1-Chloroacetadehyde Oxime (8). Acetaldehyde oxime (670 mg, 11.4 mmol) was dissolved in CH2Cl2
(60 ml), and benzyl(trimethyl)ammonium tetrachloroiodate (BTMA·ICl4) (4.77 g, 11.4 mmol) [23] was
added. The greenish suspension dissolved within 10 min after vigorous stirring at r.t. to result in a yellow
soln. After stirring for additional 30 min, the mixture was diluted with Et2O (280 ml), and the resulting
precipitate was filtered off. Evaporation of the filtrate in vacuo gave 8 as a yellow oil, which was used in
the following step without further purification.
tert-Butyl 2-(4,5-Dihydro-3-methylisoxazol-5-yl)-2-hydroxy-7-azabicyclo[2.2.1]heptane-7-carboxy-
lates (9a and 9b). To a soln. of 7 (1.0 g, 4.18 mmol) in CH2Cl2 (33 ml) 8 (600 mg, 6.42 mmol) and Et3N
(900 ml, 6.42 mmol) were added. The mixture was stirred for 3 d at r.t., while further aliquots of 8 (total
amount 3.0 g, 32.1 mmol) and Et3N (total amount 4.5 ml, 32.1 mmol) were added portionwise. The
mixture was diluted with AcOEt (30 ml) and sequentially washed with a sat. NaHCO3 soln. (3ꢂ20 ml),
H2O (1ꢂ20 ml), and brine (1ꢂ20 ml). The residual org. layer was dried (Na2SO4), filtered, and
evaporated at reduced pressure to afford a crude mixture of 9a and 9b, which were separated by FC
(SiO2; petroleum ether/AcOEt 4 :1). Diastereoisomers 9a (240 mg) and 9b (350 mg) were obtained in
48% overall yield. The HPLC analysis of the crude mixture confirmed a 9a/9b ratio of 2 :3. The reported
relative configurations resulted from the X-ray analysis performed on isomer 9b, as depicted in Fig. 2 and
described below.
(1R*,2R*,4S*)-tert-Butyl
2-[(5S*)-4,5-Dihydro-3-methylisoxazol-5-yl]-2-hydroxy-7-azabicy-
clo[2.2.1]heptane-7-carboxylate (9a). Colorless prisms from (i-Pr)2O. M.p. 112–1148. Rf (petroleum
ether/AcOEt 1:1) 0.43. HPLC (mobile phase: petroleum ether/AcOEt 1:1, flow rate: 0.5 ml/min): tR
8.84 min. 1H-NMR: 1.43 (s, 9 H); 1.52–1.70 (m, 4 H); 1.79 (br. s, 1 H); 1.99 (s, 3 H); 2.14–2.32 (m, 2 H);
2.90–3.15 (m, 2 H); 3.89 (br. s, 1 H); 4.21 (br. s, 1 H); 4.59 (t, J ¼ 9.6, 1 H). Anal. calc. for C15H24N2O4
(296.36): C 60.79, H 8.16, N 9.45; found: C 60.91, H 7.95, N 9.32.
(1R*,2R*,4S*)-tert-Butyl
2-[(5R*)-4,5-Dihydro-3-methylisoxazol-5-yl]-2-hydroxy-7-azabicy-
clo[2.2.1]heptane-7-carboxylate (9b). Colorless prisms from (i-Pr)2O. M.p. 150–1518 Rf (petroleum
ether/AcOEt 1:1) 0.26. HPLC (mobile phase: petroleum ether/AcOEt 1:1, flow rate: 0.5 ml/min): tR
12.59 min. 1H-NMR: 1.44 (s, 9 H); 1.50–1.87 (m, 6 H); 1.98 (s, 3 H); 2.15–2.27 (m, 1 H); 2.91 (d, J ¼ 10.1,
2 H); 4.12–4.24 (m, 2 H); 4.68 (t, J ¼ 10.1, 1 H). Anal. calc. for C15H24N2O4 (296.36): C 60.79, H 8.16, N
9.45; found: C 60.72, H 8.07, N 9.37.
(1R*,2R*,4S*)-2-[(5S*)-3-Methylisoxazol-5-yl]-7-azabicyclo-[2.2.1]heptan-2-ol (2a). Isomer 9a
(142 mg, 0.48 mmol) was dissolved in CH2Cl2 (0.8 ml), and CF3COOH (0.37 ml, 4.80 mmol) was added
dropwise. After stirring for 2 h at r.t., the mixture was evaporated in vacuo, H2O (2 ml) was added, and
the aq. layer was washed with Et2O (3ꢂ3 ml). The residual aq. phase was basified to pH 10 with solid
K2CO3 and extracted with AcOEt (6ꢂ3 ml). The collected org. layers were dried (Na2SO4), filtered, and
evaporated. The free amine was purified by crystallization from Et2O/CH2Cl2 1:1 to afford 2a (59 mg,
63%). White solid. M.p. 109–1118 (dec.). Rf (CH2Cl2/MeOH/NH3 ·H2O 8 :2 :0.1) 0.55. 1H-NMR: 1.21 (d,
J¼12.8, 1 H); 1.34–1.47 (m, 1 H); 1.54–1.64 (m, 2 H); 1.81 (br. s, 2 H); 1.97 (s, 3 H); 2.03 (dd, J¼5.5,
12.5, 1 H); 2.21–2.32 (m, 1 H); 2.94 (dd, J ¼ 11.0, 17.2, 1 H); 3.07 (dd, J ¼ 8.8, 17.2, 1 H); 3.32 (dd, J¼8.8,
11.0, 1 H); 3.66 (br. s, 1 H); 4.72 (dd, J¼8.8, 11.0, 1 H).
To a soln. of 2a (59 mg, 0.30 mmol) in MeOH (1.5 ml) was added a soln. of fumaric acid (35 mg,
0.30 mmol) in MeOH (1.5 ml). After stirring overnight at r.t., the mixture was concentrated under
reduced pressure to afford quantitatively the corresponding fumarate which was crystallized.
1
2a·C4H4O4. Colorless prisms from i-PrOH. M.p. 135–1378. H-NMR (CD3OD): 1.56 (d, J¼13.9,
1 H); 1.73–1.87 (m, 2 H); 1.98 (s, 3 H); 2.00–2.12 (m, 2 H); 2.44–2.56 (m, 1 H); 2.94 (dd, J ¼ 8.4, 17.9,
1 H); 3.14 (dd, J ¼ 11.0, 17.9, 1 H); 4.04 (d, J¼3.7, 1 H); 4.13 (t, J¼5.1, 1 H); 4.68 (dd, J¼8.4, 11.0, 1 H);
6.68 (s, 2 H). Anal. calc. for C14H20N2O6 (312.32): C 53.84, H 6.45, N 8.97; found: C 53.68, H 6.59, N 9.11.
(1R*,2R*,4S*)-2-[(5R*)-4,5-Dihydro-3-methylisoxazol-5-yl]-7-azabicyclo[2.2.1]heptan-2-ol (2b).
Isomer 9b (180 mg, 0.61 mmol) was reacted with CF3COOH according to the procedure described for
9a. The free amine 2b, obtained in comparable yield (80 mg, 67%), was crystallized from AcOEt to

254
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
afford a colorless crystalline compound. M.p. 125–1278 (dec.). Rf (CH2Cl2/MeOH/NH3 ·H2O 8 :2 :0.1)
0.45. 1H-NMR: 1.23 (d, J¼12.1, 1 H); 1.37–1.49 (m, 1 H); 1.52–1.65 (m, 2 H); 1.96 (s, 3 H); 2.02 (br. s,
3 H); 2.18–2.29 (m, 1 H); 2.93 (d, J ¼ 9.3, 2 H); 3.57–3.67 (m, 2 H); 4.8 (t, J¼9.3, 1 H).
The corresponding fumarate was obtained as previously described for 2a.
2b·1/2 C4H4O4. Colorless prisms from i-PrOH. M.p. 134–1378. 1H-NMR (CD3OD): 1.55 (d, J¼13.5,
1 H); 1.67–1.82 (m, 2 H); 1.84–1.95 (m, 2 H); 1.97 (s, 3 H); 2.34–2.46 (m, 1 H); 3.00 (dd, J¼8.7, 17.4,
1 H); 3.07 (dd, J¼10.6, 17.4, 1 H); 3.96–4.05 (m, 2 H); 4.64 (dd, J¼8.7, 10.6, 1 H); 6.66 (s, 1 H). Anal.
calc. for C12H18N2O4 (254.28): C 56.68, H 7.13, N 11.02; found: C 56.43, H 7.30, N 10.87.
(1R*,2R*,4S*)-tert-Butyl 2-[(5S*)-4,5-Dihydro-3-methylisoxazol-5-yl]-2-{[(methylsulfanyl)thiocar-
bonyl]oxy}-7-azabicyclo[2.2.1]heptane-7-carboxylate (10a). A suspension of NaH (110 mg, 4.58 mmol)
in dry THF (10 ml) was cooled at 08 under a flow of anh. N2. To the mixture was added dropwise a soln. of
9a (466 mg, 1.57 mmol) in dry THF (5 ml). The mixture was stirred at 08 for 10 min, then was allowed to
warm to r.t. and stirred for additional 15 min. After cooling again to 08, CS2 (775 ml, 12.87 mmol) was
added dropwise, and the red-orange soln. was stirred for 30 min at r.t., then cooled to 08, and MeI
(1.56 ml, 25.12 mmol) was added. The mixture was allowed to warm to r.t. and then stirred overnight. A
sat. aq. NH4Cl soln. (20 ml) was added, and the separated aq. layer was extracted with AcOEt (3ꢂ
10 ml). The pooled org. phases were dried (Na2SO4), filtered, and evaporated. The residue was purified
by FC (SiO2; petroleum ether/AcOEt 4 :1) to afford 10a as a yellow viscous oil (243 mg, 40%). Rf
(petroleum ether/AcOEt 7:3) 0.35. 1H-NMR: 1.45 (s, 9 H); 1.52–1.75 (m, 4 H); 1.76–1.90 (m, 1 H); 1.97
(s, 3 H); 2.21–2.34 (m, 1 H); 2.52 (s, 3 H); 2.93 (dd, J¼11.5, 17.3, 1 H); 3.05–3.28 (m, 1 H); 4.27 (br. s,
1 H); 4.81 (br. s, 1 H); 5.99 (t, J¼10.6, 1 H). Anal. calc. for C17H26N2O4S2 (386.53): C 52.82, H 6.78, N
7.25; found: C 53.07, H 6.55, N 7.50.
(1R*,2S*,4S*)-tert-Butyl 2-[(5S*)-4,5-Dihydro-3-methylisoxazol-5-yl]-7-azabicyclo[2.2.1]heptane-7-
carboxylate (11a). To a soln. of 2,2’-azobis[isobutyronitrile] (AIBN; 180 mg, 1.10 mmol) in toluene
(65 ml) was cautiously added a soln. of BuSnH (2.5 ml, 9.28 mmol) in toluene (40 ml). To the mixture
heated at reflux under anh. N2, a soln. of 10a (450 mg, 1.16 mmol) in toluene (20 ml) was added dropwise
over 20 min. After stirring at reflux for 6 h, the crude mixture was concentrated at reduced pressure, and
the residue was submitted to FC (SiO2; petroleum ether/AcOEt 3 :2) to give 11a (200 mg, 61%). Pale
yellow oil. Rf (petroleum ether/AcOEt 1:1) 0.21. 1H-NMR: 1.22–1.36 (m, 2 H); 1.44 (s, 9 H); 1.64–1.88
(m, 4 H); 1.98 (s, 3 H); 2.21–2.35 (m, 1 H); 2.58 (dd, J¼8.4, 16.8, 1 H); 3.01 (dd, J¼9.9, 16.8, 1 H); 4.18
(br. s, 1 H); 4.29 (br. s, 1 H); 4.32–4.44 (m, 1 H). Anal. calc. for C15H24N2O3 (280.36): C 64.26, H 8.63, N
9.99; found: C 64.43, H 8.55, N 9.72.
(1R*,2S*,4S*)-2-[(5S*)-4,5-Dihydro-3-methylisoxazol-5-yl]-7-azabicyclo-[2.2.1]heptane (1a). Inter-
mediate 11a (190 mg, 0.68 mmol) was reacted with CF3COOH, according to the protocol applied to 9a, to
afford 1a (88 mg, 72%). Colorless oil. Rf (CH2Cl2/MeOH 4 :1) 0.29. 1H-NMR: 1.21–1.69 (m, 5 H); 1.75–
1.92 (m, 1 H); 1.98 (s, 3 H); 2.09–2.21 (m, 1 H); 2.59 (dd, J¼8.8, 16.8, 1 H); 2.85–3.08 (m, 1 H); 3.45–
3.76 (m, 2 H); 4.29–4.42 (m, 1 H).
The corresponding fumarate was obtained as previously described for 2a.
1a·3/4 C4H4O4. Colorless prisms from i-PrOH/MeOH 2 :1. M.p. 177–1798. 1H-NMR (CD3OD): 1.56
(dd, J¼5.1, 12.8, 1 H); 1.71–1.93 (m, 3 H); 1.97 (s, 3 H); 2.05–2.32 (m, 2 H); 2.40–2.53 (m, 1 H); 2.72
(dd, J¼8.4, 17.2, 1 H); 3.18 (dd, J¼10.3, 17.2, 1 H); 4.12–4.25 (m, 2 H); 4.55–4.78 (m, 1 H); 6.67 (s,
1.5 H). Anal. calc. for C13H19N2O4 (267.30): C 56.75, H 6.80, N 9.45; found: C 57.02, H 6.67, N 9.53.
(1R*,2R*,4S*)-tert-Butyl 2-[(5R*)-4,5-Dihydro-3-methyl-4,5-dihydroisoxazol-5-yl]-2-{[(methylsul-
fanyl)thiocarbonyl]oxy}-7-azabicyclo[2.2.1]heptane-7-carboxylate (10b). Isomer 9b (410 mg, 1.30 mmol)
was reacted according to the protocol reported for 9a. The crude mixture was purified by FC (SiO2;
petroleum ether/AcOEt 4 :1) to afford 10b (250 mg, 47%). Grey amorphous solid. Rf (petroleum ether/
1
AcOEt 7:3) 0.25. H-NMR: 1.45 (s, 9 H); 1.50–1.72 (m, 4 H); 1.75–1.91 (m, 1 H); 1.96 (s, 3 H); 2.18–
2.38 (m, 1 H); 2.51 (s, 3 H); 2.95 (dd, J¼11.4, 17.3, 1 H); 3.22 (dd, J¼9.5, 17.3, 1 H); 4.22 (br. s, 1 H); 5.02
(br. s, 1 H); 5.99 (t, J¼10.4, 1 H). Anal. calc. for C17H26N2O4S2 (386.53): C 52.82, H 6.78, N 7.25; found: C
52.94, H 6.66, N 7.43.
(1R*,2S*,4S*)-tert-Butyl 2-[(5R*)-4,5-Dihydro-3-methylisoxazol-5-yl]-7-azabicyclo[2.2.1]heptane-
7-carboxylate (11b). Intermediate 10b (464 mg, 1.20 mmol) was reacted according to the procedure
applied to 11a. The crude mixture was purified by FC (SiO2; petroleum ether/AcOEt 3 :2) to afford 11b

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
255
as a yellow oil (195 mg, 58%). Rf (petroleum ether/AcOEt 1:1) 0.18. 1H-NMR: 1.17–1.34 (m, 2 H); 1.43
(s, 9 H); 1.62–1.87 (m, 4 H); 1.97 (s, 3 H); 2.18–2.33 (m, 1 H); 2.58 (dd, J¼8.8, 16.8, 1 H); 2.98 (dd, J¼
8.4, 16.8, 1 H); 4.05 (br. s, 1 H); 4.16 (br. s, 1 H); 4.10–4.46 (m, 1 H). Anal. calc. for C15H24N2O3 (280.36):
C 64.26, H 8.63, N 9.99; found: C 64.37, H 8.57, N 10.21.
(1R*,2S*,4S*)-2-[(5R*)-4,5-Dihydro-3-methylisoxazol-5-yl]-7-azabicyclo[2.2.1]heptane (1b). Inter-
mediate 11b (196 mg, 0.70 mmol) was reacted with CF3COOH, according to the protocol reported for 9a
to afford 1b (82 mg, 65%). Pale yellow oil. Rf (CH2Cl2/MeOH 4 :1) 0.20. 1H-NMR: 1.27–1.70 (m, 5 H);
1.75–1.92 (m, 1 H); 1.97 (s, 3 H); 2.07–2.21 (m, 1 H); 2.58 (dd, J¼9.1, 16.8, 1 H); 2.89–3.05 (m, 1 H);
3.50–3.81 (m, 2 H); 4.26–4.43 (m, 1 H).
The corresponding fumarate was obtained as previously described for 2a.
1b·3/4 C4H4O4. Colorless prisms from i-PrOH/MeOH 2 :1. M.p. 162–1658 (dec.). 1H-NMR
(CD3OD): 1.56 (dd, J¼5.1, 13.2, 1 H); 1.69–1.94 (m, 3 H); 1.97 (s, 3 H); 2.05–2.24 (m, 2 H); 2.39–2.52
(m, 1 H); 2.72 (dd, J¼8.3, 17.3, 1 H); 3.19 (dd, J¼10.6, 17.3, 1 H); 4.12–4.25 (m, 2 H); 4.55–4.72 (m,
1 H); 6.68 (s, 1.5 H). Anal. calc. for C13H19N2O4 (267.30): C 56.75, H 6.80, N 9.45; found: C 56.57, H 7.05, N
9.61.
tert-Butyl 3’-Methyl-spiro[7-azabicyclo[2.2.1]heptane-2,5’-[4’H]isoxazole]-7-carboxylate (15). To a
soln. of 8 (600 mg, 6.42 mmol) in ClCH2CH2Cl (30 ml) were sequentially added 12 [20] (837 mg,
4.0 mmol) and Et3N (900 ml, 6.42 mmol). The mixture was heated at reflux for 4 d, while further amounts
of 8 (935 mg, 10 mmol) and Et3N (1.4 ml, 10 mmol) were added portionwise. After addition of H2O
(20 ml), the separated aq. phase was extracted with CH2Cl2 (6ꢂ20 ml). The collected org. layers were
dried (Na2SO4), filtered, and evaporated. The crude mixture was purified by FC (SiO2; petroleum ether/
AcOEt 1:1) to provide 15 (117 mg, 11%). Yellow oil. Rf (petroleum ether/AcOEt 3 :7) 0.40. 1H-NMR:
1.42 (s, 9 H); 1.56 (d, J¼11.3, 1 H); 1.46–1.82 (m, 4 H); 1.97 (s, 3 H); 2.32 (dd, J¼3.5, 12.2, 1 H); 2.78 (d,
J¼18.0, 1 H); 2.99 (d, J¼18.0, 1 H); 4.12 (br. s, 1 H); 4.28 (br. s, 1 H). Anal. calc. for C14H22N2O3
(266.34): C 63.13, H 8.33, N 10.52; found: C 62.91, H 8.54, N 10.43.
3’-Methyl-spiro[7-azabicyclo[2.2.1]heptane-2,5’-[4’H]-isoxazole] (3). Reaction of 15 (110 mg,
0.41 mmol) with CF3COOH according to the procedure reported for 9a afforded 3 (55 mg, 80%).
1
Colorless oil. Rf (CH2Cl2/MeOH 9 :1) 0.32. H-NMR: 1.22–1.35 (m, 2 H); 1.58–1.70 (m, 3 H); 1.97 (s,
3 H); 2.07 (dd, J¼5.1, 13.9, 1 H); 2.87 (d, J¼17.2, 1 H); 2.97 (d, J¼17.2, 1 H); 3.44 (d, J¼4.0, 1 H); 3.66
(t, J¼4.0, 1 H).
The corresponding fumarate was obtained as previously described for 2a.
3·C4H4O4. Colorless prisms from MeOH. M.p. 149–1518. 1H-NMR (CD3OD): 1.65–1.80 (m, 1 H);
1.82–1.94 (m, 3 H); 1.98 (s, 3 H); 2.08 (d, J¼14.3, 1 H); 2.35 (dd, J¼4.4, 14.3, 1 H); 3.12 (d, J¼17.6,
1 H); 3.27 (d, J¼17.6, 1 H); 4.04 (br. s, 1 H); 4.21 (br. s, 1 H); 6.68 (s, 2 H). Anal. calc. for C13H18N2O5
(282.29): C 55.31, H 6.43, N 9.92; found: C 55.53, H 6.18, N 9.70.
tert-Butyl 5-Methyl-3-oxa-4,10-diazatricyclo[5.2.1.02,6]dec-4-ene-10-carboxylate (16). Alkene 13 [27]
(295 mg, 1.48 mmol) and 8 (187 mg, 2 mmol) were dissolved in CH2Cl2 (10 ml). To the stirred soln., Et3N
(280 ml, 2.0 mmol) was added dropwise, and the mixture was further stirred for 4 h at r.t. After pouring
into H2O (10 ml) and the usual workup, the residue was purified by FC (SiO2; petroleum ether/AcOEt
1:1) to afford 16 (325 mg, 87%) as a crystalline derivative (colorless leaflets from hexane). M.p. 79–808.
1
Rf (petroleum ether/AcOEt 3 :7) 0.54. H-NMR: 1.20–1.40 (m, 2 H); 1.43 (s, 9 H); 1.61 (s, 1 H); 1.70–
1.88 (m, 2 H); 2.00 (s, 3 H); 3.23 (d, J¼8.1, 1 H); 4.35–4.50 (m, 1 H); 4.51–4.71 (m, 1 H). Anal. calc. for
C13H20N2O3 (252.31): C 61.88, H 7.99, N 11.10; found: C 61.67, H 7.80, N 11.38.
5-Methyl-3-oxa-4,10-diazatricyclo[5.2.1.02,6]dec-4-ene (4). Cycloadduct 16 (280 mg, 1.11 mmol) was
reacted with CF3COOH as previously described to give 4 (137 mg, 81%). Pale yellow oil. Rf (CH2Cl2/
MeOH 9 :1) 0.35. 1H-NMR: 1.10–1.38 (m, 2 H); 1.57–1.75 (m, 2 H); 1.99 (s, 3 H); 3.17 (d, J¼7.7, 1 H ) ;
3.61 (d, J¼4.0, 1 H); 3.75 (d, J¼4.8, 1 H); 4.61 (d, J¼7.7, 1 H ) .
The corresponding fumarate was obtained as previously described for 2a.
1
4·C4H4O4. Colorless prisms from AcOEt/MeOH 1:1. M.p. 154–1578. H-NMR (CD3OD): 1.55–
1.77 (m, 2 H); 1.78–1.93 (m, 2 H); 1.99 (s, 3 H); 3.62 (d, J¼8.4, 1 H); 4.26 (dd, J¼4.2, 15.5, 2 H); 4.80 (d,
J¼8.4, 1 H); 6.69 (s, 2 H). Anal. calc. for C12H16N2O5 (268.27): C 53.73, H 6.01, N 10.44; found: C 55.95,
H 6.12, N 10.27.

256
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
tert-Butyl 4,5,6,6a-Tetrahydro-3-methyl-3aH-pyrrolo[3,4-d]isoxazole-5-carboxylate (17). Alkene 14
[28] (677 mg, 4 mmol) was reacted with 8 according to the protocol for the synthesis of 15. After the usual
workup, the crude mixture was purified by FC (SiO2; petroleum ether/AcOEt 2 :3) to afford 17 (245 mg,
27%). Yellow oil. Rf (petroleum ether/AcOEt 3 :7) 0.60. 1H-NMR: 1.43 (s, 9 H); 2.00 (s, 3 H); 3.35–3.58
(m, 2 H); 3.70 (t, J ¼ 8.7, 2 H); 3.82 (d, J¼12.4, 1 H); 5.09–5.19 (m, 1 H). Anal. calc. for C11H18N2O3
(226.27): C 58.39, H 8.02, N 12.38; found: C 58.60, H 7.98, N 12.13.
4,5,6,6a-Tetrahydro-3-methyl-3aH-pyrrolo[3,4-d]isoxazole (5). Reaction of 17 (190 mg, 0.84 mmol)
1
with CF3COOH produced 5 (95 mg, 90%). Yellow oil. Rf (CH2Cl2/MeOH 4 :1) 0.30. H-NMR: 1.97 (s,
3 H); 2.80 (dd, J¼3.7, 13.5, 1 H); 2.88 (dd, J¼7.7, 12.8, 1 H); 3.21 (d, J¼12.8, 1 H); 3.40 (d, J¼13.5, 1 H);
3.63 (t, J¼8.8, 1 H); 5.13 (dd, J¼3.7, 8.8, 1 H).
The corresponding fumarate was obtained as previously described for 2a.
5·3/2 C4H4O4. Colorless prisms from hexane/MeOH 1:1. M.p. 156–1578. 1H-NMR (CD3OD): 1.99
(s, 3 H); 3.42 (dd, J¼5.1, 12.8, 2 H); 3.66 (d, J¼12.8, 2 H); 4.08 (t, J¼8.8, 1 H); 5.29 (dd, J¼5.1, 8.8,
1 H); 6.75 (s, 3 H). Anal. calc. for C12H16N2O7 (300.26): C 48.00, H 5.37, N 9.33; found: C 47.75, H 5.51, N
9.12.
X-Ray Crystallographic Analysis. Structure Determination of Compound (ꢀ)-9b. Colorless single
crystals were obtained from a soln. in (i-Pr)2O. Crystal data: C15H24N2O4, Mr 296.36, monoclinic, space
group P21/c (No. 14), a¼6.0800(7) ꢄ, b¼8.2100(8) ꢄ, c¼31.959(3) ꢄ, b¼91.085(10)8, V¼
1595.0(3) ꢄ3, Z¼4, Dc ¼1.234 Mg mꢁ3, F(000)¼640, m(MoKa)¼0.089 mmꢁ1, T¼122.0(5) K, crystal
dimensions 0.1ꢂ0.15ꢂ0.4 mm.
Data Collection and Processing. Diffraction data were collected on an Enraf-Nonius KappaCCD
diffractometer using graphite monochromated MoKa radiation (l 0.71073 ꢄ) [33][34]. The reflections
were measured in the range ꢁ7ꢃhꢃ7, ꢁ10ꢃkꢃ10, ꢁ41ꢃlꢃ41, (2.558<q<27.508). Data were
reduced using the program EvalCCD [35]. A total of 34141 reflections were averaged according to the
point-group symmetry 2/m resulting in 3652 unique reflections (Rint ¼0.0472 on F2o ).
Structure Solution and Refinement. The structure was solved by the direct method using the
programme SHELXS97 [36][37] and refined using the programme SHELXL97 [38]. Full-matrix least-
squares refinement on F2 was performed, minimizing Sw(Fo2 ꢁFc2 )2, with anisotropic displacement
parameters for the non-H-atoms. The positions of the H-atoms were located on intermediate difference
electron-density maps. Nearly all H-atoms are included in calculated positions, riding on the parent
atoms with fixed isotropic displacement parameters. Only the positions of H-atoms connected to
stereogenic C-atoms are refined with fixed isotropic displacement parameters. The refinement (206
parameters, 3652 reflections) converged at RF ¼0.0360, wRF2 ¼0.0920 for 2833 reflections with Fo >
4s(Fo); w¼1/[s2(Fo2 )þ(0.0495P)2 þ0.3240P], where P¼(Fo2 þ2F2c )/3; S¼1.052. In the final difference
Fourier map, maximum and minimum electron densities were 0.291 and ꢁ0.225 e ꢄꢁ3, resp. Complex
atomic scattering factors for neutral atoms were as incorporated in SHELXL97 [38][39].
Further details of crystal structure for compound (ꢀ)-9b have been deposited with the Cambridge
Crystallographic Data Centre1).
Receptor Binding Assay. Membranes binding of [3H]Epibatidine and [125I]-a-Bungarotoxin. The
cortex tissues were dissected, immediately frozen on dry ice and stored at –808 for later use. In each
experiment, the cortex tissues from two rats were homogenized in 10 ml of a buffer soln. (50 mm Na3PO4,
1m NaCl, 2 mm EDTA, 2 mm EGTA, and 2 mm PMSF, pH 7.4) using a potter homogenizer; the
homogenates were then diluted and centrifuged at 60,000g for 1.5 h. The total membrane homoge-
nization, dilution, and centrifugation procedures were performed twice, then the pellets were collected,
rapidly rinsed with a buffer soln. (50 mm Tris·HCl, 120 mm NaCl, 5 mm KCl, 1 mm MgCl2, 2.5 mm CaCl2,
and 2 mm PMSF, pH 7), and resuspended in the same buffer containing a mixture of 20 mg/ml of each of
the following protease inhibitors: leupeptin, bestatin, pepstatin A, and aprotinin.
[3H]Epibatidine Binding. (ꢀ)-[3H]Epibatidine with a specific activity of 56–60 Ci/mmol was
purchased from Perkin-Elmer (Boston, MA); the nonradioactive a-bungarotoxin, nicotine, and
epibatidine were purchased from Sigma. It has previously been reported that [3H]epibatidine also binds
1
)
CCDC-668587 contains the supplementary crystallographic data for this work. These data can be
obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif.

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
257
to a-bungarotoxin-binding receptors with nm affinity [14b]. To prevent the binding of [3H]epibatidine to
the a-bungarotoxin-binding receptors, the membrane homogenates were preincubated with 2 mm a-
bungarotoxin and then with [3H]epibatidine. The saturation experiments were performed by incubating
aliquots of cortex membrane homogenates with 0.01–2.5 nm (ꢀ)-[3H]epibatidine overnight at 48.
Nonspecific binding was determined in parallel by means of incubation in the presence of 100 nm
unlabelled epibatidine. At the end of the incubation, the samples were filtered on a GFC filter soaked
in 0.5% polyethylenimine and washed with 15 ml of a buffer soln. (10 mm Na3PO4, 50 mm NaCl, pH 7.4),
and the filters were counted in a b counter.
[
125I]-a-Bungarotoxin Binding. The saturation binding experiments were performed using aliquots of
cortex membrane homogenates incubated overnight with concentrations ranging from 0.1–10 nm of
125I]-a-bungarotoxin (specific activity 200–213 Ci/mmol, Amersham) at r.t. Nonspecific binding was
[
determined in parallel by means of incubation in the presence of 1 mm unlabelled a-bungarotoxin. After
incubation, the samples were filtered as described above, and the bound radioactivity was directly
counted in a g counter.
nACh Receptor Affinity of Derivatives 1a and 1b, 2a and 2b, and 3–5. The inhibition of radioligand
binding by epibatidine, nicotine, and the test compounds was measured by preincubating cortex
homogenates with increasing doses (10 pm –10 mm) of the reference nicotinic agonists, epibatidine or
nicotine, and the drug to be tested for 30 min at r.t., followed by overnight incubation with a final
concentration of 0.075 nm [3H]epibatidine or 1 nm [125I]- a-bungarotoxin at the same temp. as those used
for the saturation experiments. These ligand concentrations were used for the competition binding
experiments, because they are within the range of the KD values of the ligands for the two different classes
of nAChRs. For each compound, the experimental data obtained from the three saturation and three
competition binding experiments were analyzed by means of a nonlinear least-square procedure using
the LIGAND program as described by Munson and Rodbard [31]. The binding parameters were
calculated by simultaneously fitting three independent saturation experiments, and the Ki values were
determined by fitting the data of three independent competition experiments. The errors in the KD and Ki
values of the simultaneous fits were calculated using the LIGAND software, and expressed as percentage
coefficients of variation (% CV). When final compound concentrations up to 500 mm did not inhibit
radioligand binding, the Ki value was defined as being >500 mm based on the Cheng and Prusoffꢁs
equation [40].
Molecular Modeling. Ligands docked into the receptor binding clefts were built by Sybyl 8.0 (Tripos
Inc., 1699 South Hanley Rd., St. Louis, MO 63144) and preliminarily minimized at the HF/6-31g* level as
implemented in Gaussian 03 [41]. The amino groups were considered in the ionized form to better
simulate the physiological conditions.
We made use of our recently published a7 nAChR [11] molecular model and of a model of a4b2
nAChRs elaborated at the University of Florence [32]. Docking experiments of the tested ligands were
performed by means of the program GOLD 3.2 [42]; the goldscore fitness function, and the distribution
of torsion angles were chosen as indicators of the quality of the docking results. Van der Waals and H-
bonding radii were set at 4.0 and 3.0 ꢄ, resp., while genetic algorithm parameters were kept at the default
value. The resulting complexes were then optimized by means of a molecular-mechanics method
implemented in the Sybyl 8.0 software. Figures were acquired by the PyMOL software [43].
This research was financially supported by the Italian FIRB grant # RBNE03FH5y002, the PRIN
grant # 2005054943 (F. C. , C. G., and M. D. A.), and the Fondazione Cariplo grant # 2006/0882 (F. C. and
M. D. A.). We are indebted to Prof. Paola Gratteri and co-workers, University of Florence, for the
availability of their recently developed model of nicotinic a4b2 receptors. The authors wish to thank Mr.
F. Hansen, Center for Crystallographic Studies, University of Copenhagen, for the technical assistance
with the X-ray data collection, and Dr. Maria De Mieri for her contribution to the synthetic organic
chemistry section.

258
CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
REFERENCES
[1] A. Singh, A. Potter, P. Newhouse, IDrugs 2004, 7, 1096.
[2] A. A. Jensen, B. Frølund, T. Liljefors, P. Krogsgaard-Larsen, J. Med. Chem. 2005, 48, 4705.
[3] M. N. Romanelli, P. Gratteri, L. Guandalini, E. Martini, C. Bonaccini, F. Gualtieri, ChemMedChem
2007, 2, 746.
[4] M. N. Romanelli, F. Gualtieri, Med. Chem. Rev. 2003, 23, 393, and refs. cit. therein.
[5] J. Lindstrom, ꢀThe Structure of Neuronal Nicotinic Receptorsꢁ, in ꢀ Handbook of Experimental
Pharmacology: Neuronal Nicotinic Receptorsꢁ, Eds. F. Clementi, C. Gotti, D. Fornasari, Springer-
Verlag, Berlin, 2000, p. 101.
[6] S. N. Millar, Biochem. Soc. Trans. 2003, 31, 869.
[7] C. Gotti, M. Zoli, F. Clementi, Trends Pharmacol. Sci. 2006, 27, 482.
[8] J. D. Schmitt, Curr. Med. Chem. 2000, 7, 749.
[9] G. K. Lloyd, M. Williams, J. Pharmacol. Exp. Ther. 2000, 292, 461.
[10] X. Huang, F. Zheng, X. Chen, P. A. Crooks, L. P. Dwoskin, C.-G. Zhan, J. Med. Chem. 2006, 49, 7661.
[11] G. Grazioso, A. Cavalli, M. De Amici, M. Recanatini, C. De Micheli, J. Comput. Chem. 2008, 29,
2593.
[12] T. F. Spande, H. M. Garraffo, M. W. Edwards, H. J. C. Yeh, L. Pannell, J. W. Daly, J. Am. Chem. Soc.
1992, 114, 3475.
[13] T. C. Li, C. G. Qian, J. Eckman, D. F. Huang, T. Y. Shen, Bioorg. Med. Chem. Lett. 1993, 3,
2759.
[14] a) J. P. Sullivan, A. W. Bannon, CNS Drug Rev. 1996, 2, 21; b) V. Gerzanich, X. Peng, F. Wang, G.
Wells, R. Anand, S. Fletcher, J. Lindstrom, Mol. Pharmacol. 1995, 48, 774; c) J. P. Sullivan, M. W.
Decker, J. D. Brioni, D. L. Donnelly-Roberts, D. A. Anderson, A. W. Bannon, C. H. Kang, P.
Adams, M. Piattoni-Kaplan, M. J. Buckley, M. Gopalakrishnan, M. Williams, S. P. Arneric, J.
Pharmacol. Exp. Ther. 1994, 271, 624.
[15] J. E. Tønder, P. H. Olesen, Curr. Med. Chem. 2001, 8, 651, and refs. cit. therein; Z. L. Wei, Y. Xiao,
H. Yuan, M. Baydyuk, P. A. Petukhov, J. L. Musachio, K. J. Kellar, A. P. Kozikowski, J. Med. Chem.
2005, 48, 1721; J. W. Daly, J. Med. Chem. 2003, 46, 445, and refs. cit. therein.
[16] B. Badio, H. M. Garraffo, C. V. Plummer, W. L. Padgett, J. W. Daly, Eur. J. Pharmacol. 1997, 321,
189.
[17] J. Cheng, S. Izenwasser, C. Zhang, S. Zhang, D. Wade, M. L. Trudell, Bioorg. Med. Chem. Lett. 2004,
14, 1775; R. W. Fitch, X.-F. Pei, Y. Kaneko, T. Gupta, D. Shi, I. Federova, J. W. Daly, Bioorg. Med.
Chem. 2004, 12, 179.
[18] C. Dallanoce, P. Bazza, G. Grazioso, M. De Amici, C. Gotti, L. Riganti, F. Clementi, C. De Micheli,
Eur. J. Org. Chem. 2006, 3746.
[19] C. De Micheli, M. De Amici, C. Dallanoce, F. Clementi, C. Gotti, PCT Int. Appl. WO 2008000469,
2008.
[20] C. Zhang, M. L. Trudell, J. Org. Chem. 1996, 61, 7189.
[21] P. Sundararaman, W. Herz, J. Org. Chem. 1977, 42, 813; G. Rꢃedi, D. N. Laikov, H. J. Hansen, Helv.
Chim. Acta 2004, 87, 1990.
[22] J. C. Zhang, D. P. Curran, J. Chem. Soc., Perkin Trans. 1 1991, 2627.
[23] S. Kanemasa, H. Matsuda, A. Kamimura, T. Kakinami, Tetrahedron 2000, 56, 1057.
[24] L. J. Farrugia, ORTEP-3 for Windows, J. Appl. Crystallogr. 1997, 30, 565.
[25] R. M. Williams, J. Cao, H. Tsujishima, R. J. Cox, J. Am. Chem. Soc. 2003, 125, 12172; M. G. Banwell,
A. J. Edwards, G. J. Harfoot, K. A. Jolliffe, Tetrahedron 2004, 60, 535.
[26] W. Hartwing, Tetrahedron 1983, 39, 2609.
[27] F. I. Carroll, F. Liang, H. A. Navarro, L. E. Brieaddy, P. Abraham, M. I. Damaj, B. R. Martin, J.
Med. Chem. 2001, 44, 2229.
[28] S. Brandꢅnge, B. Rodriguez, Synthesis 1988, 347; P. Conti, C. Dallanoce, M. De Amici, C. De
Micheli, R. Fruttero, Tetrahedron 1999, 55, 5623.
[29] C. De Micheli, R. Gandolfi, R. Oberti, J. Org. Chem. 1980, 45, 1209; R. Gandolfi, G. Tonoletti, A.
Rastelli, M. Bagatti, J. Org. Chem. 1993, 58, 6038.

CHEMISTRY & BIODIVERSITY – Vol. 6 (2009)
259
[30] D. P. Curran, J. Am. Chem. Soc. 1983, 105, 5826; H. C. Brown, W. J. Hammar, J. H. Kawakami, I.
Rothberg, D. L. Vander Jagt, J. Am. Chem. Soc. 1967, 89, 6381.
[31] P. J. Munson, D. Rodbard, Anal. Biochem. 1980, 107, 220.
[32] P. Gratteri (University of Florence), personal communication, 2008.
[33] COLLECT, B. V. Nonius, Delft, 1999.
[34] A. J. M. Duisenberg, J. Appl. Crystallogr. 1992, 25, 92.
[35] A. J. M. Duisenberg, EvalCCD, Ph.D. Thesis, University of Utrecht, Utrecht, 1998.
[36] G. M. Sheldrick, SHELXS97, Program for the Solution of Crystal Structure, University of
Gçttingen, Gçttingen, 1997.
[37] G. M. Sheldrick, Acta Crystallogr., Sect. A 1990, 46, 467.
[38] G. M. Sheldrick, SHELXL97, Program for the Refinement of Crystal Structures, University of
Gçttingen, Gçttingen, 1997.
[39] A. J. C. Wilson, in ꢀInternational Tables for Crystallographyꢁ, Kluwer Academic Publishers,
Dordrecht, 1995, Vol. C, Tables 4.2.6.8 and 6.1.1.4.
[40] Y.-C. Cheng, W. H. Prusoff, Biochem. Pharmacol. 1973, 22, 3099.
[41] 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. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, 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, S. Baboul, J. Clifford, B. B. Cioslowski, G. Stefanov, A. Liu, P.
Liashenko, I. Piskorz, A. G. 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 B.04, Gaussian, Inc., Wallingford CT, 2004.
[42] Gold v. 3.2, Cambridge Crystallographic Data Centre, Cambridge, UK.
[43] W. L. DeLano, The PyMOL Molecular Graphics System, DeLano Scientific, Palo Alto, 2002; http://
www.pymol.org.
Received February 20, 2008