An investigation of the Lewis acid mediated 1,3-dipolar cycloaddition between N-benzyl-C-(2-pyridyl)nitrone and allylic alcohol. Direct entry to isoxazolidinyl C-nucleosides

Article abstract of DOI:10.1039/b304112c

The cycloaddition reaction of N-benzyl C-(2-pyridyl) nitrone with allylic alcohol has been carried out to obtain the corresponding 2-benzyl-3-(2-pyridyl)-5-hydroxymethylisoxazolidine. The influence of Lewis acids in the reaction has been studied and a complete 3,5-regioselectivity and cis diastereoselectivity was observed when the reaction was carried out with 1.0 equiv of AgOTf, [Ag(OClO₃)(PPh₂Me)] or Zn(OTf)₂. Insight into the mechanism of the reaction has been obtained by isolating and characterizing (X-ray) the intermediate complexes. Also, a model based on both experimental and theoretical results is proposed.

Full text of DOI:10.1039/b304112c

View Article Online / Journal Homepage / Table of Contents for this issue
An investigation of the Lewis acid mediated 1,3-dipolar
cycloaddition between N-benzyl-C-(2-pyridyl)nitrone and allylic
alcohol. Direct entry to isoxazolidinyl C-nucleosides†
Pedro Merino,*a Tomas Tejero,a Mariano Laguna,b Elena Cerrada,b Ana Morenob and
Jose A. Lopezb
a Departamento de Quimica Organica Facultad de Ciencias, ICMA, Universidad de Zaragoza,
CSIC, E-50009 Zaragoza, Aragon, Spain. E-mail: pmerino@posta.unizar.es;
Fax: ϩ34 976 762075; Tel: ϩ34 976 762075
b Departamento de Quimica Inorganica Facultad de Ciencias, ICMA, Universidad de Zaragoza,
CSIC, E-50009 Zaragoza, Aragon, Spain
Received 16th April 2003, Accepted 20th May 2003
First published as an Advance Article on the web 4th June 2003
The cycloaddition reaction of N-benzyl C-(2-pyridyl) nitrone with allylic alcohol has been carried out to obtain the
corresponding 2-benzyl-3-(2-pyridyl)-5-hydroxymethylisoxazolidine. The influence of Lewis acids in the reaction has
been studied and a complete 3,5-regioselectivity and cis diastereoselectivity was observed when the reaction was
carried out with 1.0 equiv of AgOTf, [Ag(OClO3)(PPh2Me)] or Zn(OTf )2. Insight into the mechanism of the reaction
has been obtained by isolating and characterizing (X-ray) the intermediate complexes. Also, a model based on both
experimental and theoretical results is proposed.
interesting isoxazolidinyl C-nucleoside analogues 1.11 In this
Introduction
context, several C-nucleosides with different hetero- and carb-
The 1,3-dipolar cycloaddition of nitrones with alkenes is a
ocyclic rings have been found to possess promising biological
powerful synthetic method that has found numerous appli-
activities.12
cations in the synthesis of a variety of carbon frameworks.1
Compound 1 (R = Bn) can be easily obtained in one step by
a cycloaddition reaction between the corresponding 2-pyridyl
Several laboratories have focused on the development of Lewis
acid reagents to modulate the selectivity of the reaction.2 Such a
nitrone and allylic alcohol, and using the appropriate Lewis
modulation can be achieved either by complexation of the
acid. In this paper we also present a detailed study of that
dipolarophile3 (typically, in cycloadditions with electron-poor
reaction, including isolation and full characterization of
alkenes) or by nitrone activation4 through previous complex-
the intermediate complexes.13 A theoretical study, at a semi-
ation.5 For the latter it has been observed that the presence
empirical (PM3) level, of the reaction has also been carried out.
of a coordinating group at the alpha position is needed to
achieve the desired complexation (Scheme 1). Usually, such
a coordinating group consists of a carbonyl moiety (ester or
ketone), and typically, several cycloaddition reactions of carb-
onyl conjugated nitrones with allylic alcohols have been
reported to show high selectivities when activated with several
Lewis acids.6 However, the coordinating group X, can also be
part of a heterocyclic ring, thus increasing the range of nitrones
which can be used as suitable substrates.7
Scheme 1
Results and discussion
As part of our continuing efforts to develop novel strategies
for preparing heterocyclic nucleosides8 (i.e. nucleoside ana-
logues in which the furanose ring has been replaced by a differ-
ent heterocyclic ring),9 we have investigated the applicability of
1,3-dipolar cycloaddition reactions of nitrones to the synthesis
of isoxazolidinyl nucleosides.10 We now wish to report our
findings on Lewis acid modulated cycloadditions leading to
The N-benzyl-C-(2-pyridyl)nitrone 2 employed for the pres-
ent investigation was obtained by condensation of 2-formyl-
pyridine and N-benzylhydroxylamine.14 The results of the
reaction illustrated in Scheme 2 are collected in Table 1.
Cycloaddition of 2 with allylic alcohol took seven days in
acetone at reflux to give a 70 : 30 mixture of cis and trans
adducts, respectively (entry 1). On the other hand, the addition
of silver salts (entries 2 and 3) only afforded the cis isomer with
notable increasing of the rate of the reaction. With Zn salts,
only in the case of Zn(OTf )2 an even higher effect was observed
† Electronic supplementary information (ESI) available: optimized
geometries (PDB format). See http://www.rsc.org/suppdata/ob/b3/
b304112c/
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 3 3 6 – 2 3 4 2
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
2336

View Article Online
Table 1 Cycloadditions of nitrone 2 with allylic alcohola
Experimental section). In addition nitrone 2 and complexes 5
and 6 were submitted to a single-crystal X-ray structure analy-
sis15 which allowed the complete structural identification.16
Thus, silver complexes 4 and 5 were obtained as white solids
when 2-PyBN 2 was added to a diethyl ether solution of the
corresponding silver() salts AgTfO and [Ag(OClO3)(PPh2Me)],
respectively. Their IR spectra showed absorptions that can be
assigned to covalent triflate groups17 in the case of 4 and ionic18
perchlorate for 5. The presence of covalent triflate groups was
confirmed by 19F{1H} NMR, spectra in which a singlet at
Ϫ77.5 ppm, characteristic of OSO2CF3 groups, was observed.19
Time
(days)
Yieldc
(%)
Entry
Lewis acid
cis : transb
1
2
3
4
5
None
AgOTf
[Ag(OClO3)PPh2Me]
Zn(OTf )2
ZnBr2
7
3.5
5
3
5
70 : 30
>95 : 5d
>95 : 5d
>95 : 5d
>95 : 5d
90
100
92
100
100
a Reactions carried out on a 1 mmol scale; nitrone–Lewis acid–
dipolarophile 1 : 1 : 10. b The diastereomeric ratio was determined
by integration of the appropriate signals. c Yield of the mixture of
diastereomers after purification by column chromatography. d Only one
isomer was detected by 1H and 13C NMR.
1
The H NMR spectra exhibit a downfield displacement of
the nitrone resonances in both cases, owing to the coordination
to the metallic centre. The 31P{1H} NMR of 5 displayed a doub-
let of doublets as a consequence of the 107Ag (51.82% isotopic
abundance) and 109Ag (48.18%) nucleus coupling. The LSIMS
mass spectra (nitrobenzyl alcohol as matrix) of 4 showed the
parent peak with loss of one triflate molecule at m/z (%) =
787 (10) which confirmed the proposed dinuclear formulation.
In the case of 5 only fragmentation peaks are observed.
The molecular structure of complex 5 is shown in Fig. 1. The
molecule consists of two N-benzyl-C-(2-pyridyl) nitrone phos-
phine silver() molecules connected through the oxygen atom
from the nitrone. In each unit the nitrone is acting as a chelate
ligand through both N,O heteroatoms. The geometry around
the silver is tetrahedral distorted with bond angles of 146.9(2)Њ,
N(1)–Ag–P and 88.66(19), O(1)–Ag–O(1Ј). Bond lengths and
angles in the nitrone molecule are similar to those found in the
free nitrone.16 Thus, the distances in the N–O and N᎐C units in
᎐
Scheme 2
7 are 1.309(8) and 1.295(11) Å respectively and 1.300(3) and
1.299(3) Å in 2. Only a small variation is observed in the C–N
distance from 1.488(3) in the free nitrone to 1.535(11) Å in the
silver complex. There are two different Ag–O bond lengths in
the molecule: Ag–O(1) = 2.381(5) and Ag–O(1Ј) = 2.510(6) Å.
This fact has been observed in other dinuclear silver() deriv-
atives with bridging oxygen ligands.20 The Ag–N and Ag–P
distances of 2.306(7) and 2.338(2) Å respectively are similar to
those found in the literature.21 On the basis of this structure and
the mass spectrum of complex 4, we can propose a similar
dinuclear disposition with covalent triflates for this complex.
(entry 4), the addition of ZnBr2 (entry 5) leading to a slower
reaction; in both cases only the cis isomer was obtained. This
behaviour is in good agreement with the results of previously
reported Lewis acid promoted cycloaddition reactions with
allylic alcohols.6 Thus, carrying out the reaction in the presence
of 1.0 equiv of Zn(OTf )2 the cis adduct 5a was obtained in
quantitative yield as a single isomer.
Isolation and characterization of intermediate complexes
In order to obtain information concerning the structures of the
intermediate nitrone–metal complexes, nitrone 2 was treated in
the absence of the dipolarophile with [Ag(OClO3)(PPh2Me)],
AgTfO, ZnBr2 and Zn(OTf )2 (Scheme 3). Complexes 4–7
were isolated and characterized by spectroscopic methods (see
Fig. 1 ORTEP diagram of complex 5. Displacement ellipsoids drawn
at the 50% probability level. Hydrogen atoms have been omitted for
clarity.
Complexes 6 and 7 can be isolated as air-stable solids. The
presence of covalent triflate in 7 can be confirmed by IR and
1
19F{1H} NMR spectra. The H NMR spectra in both com-
plexes show a downfield displacement when compared with the
Scheme 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 3 3 6 – 2 3 4 2
2337

View Article Online
free nitrone as occurred in silver compounds. As stated above
the structure of complex 6 was confirmed by X-ray diffraction.
The molecule (Fig. 2) shows a tetrahedral distorted disposition
for the metallic centre [N(1)–Zn–Br(1) = 106.62(11)Њ, O–Zn–
Br(2) = 113.43(10)Њ]. The distances in the nitrone skeleton are
similar to those found in the free nitrone, however, the bond
angle N(2)–C(6)–C(5) of 128.2(4)Њ is slightly larger than that
found in 2 [N–C–C = 126.3(2)Њ], which can be assigned to the
formation of the chelate. In addition, it should be mentioned
that the distance between the zinc atom and the bromine trans
to the pyridine [Zn–Br(1) = 2.3641(8) Å] is 0.034 Å larger than
between the zinc atom and the bromine trans to the oxygen
from the N–O group [Zn–Br(2) = 2.3300(8) Å]. This situation
has been also observed in N,O-chelating aldehyde zinc
complexes four-, five- or even six-coordinated.22
Fig. 2 ORTEP diagram of complex 6. Displacement ellipsoids drawn
at the 50% probability level. Hydrogen atoms have been omitted for
clarity.
Scheme 4
been calculated in detail. Thus, calculation of the whole reac-
tion path profiles starting from the optimized geometry of the
nitrone was carried out. Critical points have been located with-
out any geometry restriction and have been characterized
through the calculation of the force constants matrix by ensur-
ing that they correspond to minima or saddle points on the
PES; i.e., they have zero or one and only one imaginary fre-
quency, respectively. For transition structures the C–C and C–O
distances represent the reaction coordinates. All the reaction
paths have been checked by tracing the intrinsic reaction co-
ordinate (IRC)28 from the transition structure to the lower
energy structures it connects. In all cases, it could be verified
that the transition states proceed from the reactants and give
rise to the products. Optimized geometries (in PDB format) of
all stationery points are available from the supplementary
material.†
Firstly, we have investigated the reaction promoted by zinc()
triflate outlined in Scheme 4. The formation of complex 7 from
nitrone 2 is exothermic (Ϫ20.81 kcal molϪ1) and it takes place
without energy barrier. After nucleophilic substitution of a
triflate group by allylic alcohol, complex 8 is formed. The
formation of 8 is endothermic (by 7.06 kcal molϪ1) which
is in agreement with the fact that energy must be given to the
reaction (it only works at reflux, lower temperatures affording
rather low conversion rates or no reaction at all).
The only possible transition state TS1 for the intramolecular
reaction was located. The IRC analysis showed as starting
and final points of the reaction the corresponding complex 8
and product P1c, respectively. The energies (kcal molϪ1) of the
corresponding stationary points are given in Table 2 and
the optimized structures are shown in Fig. 3.
Alternatively, it is also possible to consider an external
delivery of the allylic alcohol to the complex 7 as illustrated in
Scheme 5.
In this case, it might even become apparent that, in the exo
approach, an opportunity for coordination between zinc and
allylic alcohol may emerge, such a possibility has been rejected
by calculations. Hence, a tetracoordinated zinc atom is always
preferred. The two possible transition structures TS2x and
TS2n corresponding to exo and endo attacks, respectively were
located. The energy values for reactants (complex 7 and allylic
To confirm that complexes 4–7 were real intermediates of the
reactions they were dissolved in acetone and treated with the
corresponding dipolarophiles. In all reactions the obtained
results were identical to those observed when the Lewis acid
was added directly to the reaction as indicated above. Moreover,
once compounds 4–7 were identified, it was possible to isolate
them from the reaction mixtures after short periods of time
(less than 6 h), thus confirming the in situ formation of such
complexes.
At this point it is reasonable to speculate on the coordination
of allylic alcohol in the complex as a function of the number of
ligands. Thus, for the reaction in the presence of Zn(OTf )2, it is
possible to invoke substitution of a triflate group by the allylic
alcohol giving rise to the reactive intermediate 8 (Scheme 4).
This intermediate immediately evolves to give complex P1c
which, after work-up, affords the final cis adduct 3a.
According to this mechanism, only an exo approach is
possible and thus the complete cis selectivity is rationalized
with transition state TS1. A similar process may be invoked for
activation with silver salts. In that case, however, the allylic
alcohol might dissociate complexes 4 and 5 to form the
corresponding intermediates with similar structure to 8.
Mechanistic considerations
The validity of the hypothesis outlined in Scheme 4 was tested
by carrying out semiempirical calculations (PM3) of the whole
process.23 Semiempirical methods are intended for studying
large and complicated molecular systems of interest for which a
complete application of the ab initio methods is generally pro-
hibitive in terms of computational effort.24 Furthermore, when
several mechanisms can be invoked for related processes the
problem of determining the operating mechanism in each case
can be most conveniently and economically tackled following a
semiempirical approach at the first stages of the computational
study, considering the limitations of the underlying approx-
imations.25 All calculations were performed with the MOPAC
package26 using the PM3 method.27 The only simplification was
the replacement of the N-benzyl group by a methyl group. The
potential energy surfaces (PESs) for the studied processes have
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 3 3 6 – 2 3 4 2
2338

View Article Online
Table 2 Heats of formation (PM3) of the reagents, starting com-
plexes, transition structure and product for the intramolecular path
shown in Scheme 4
Energy (kcal molϪ1
)
Relative energy to reactants
2
43.47
Ϫ438.63
Ϫ31.30
Zn(OTf )2
Allylic alcohol
7
Ϫ415.97
Ϫ183.14
Ϫ137.46
Ϫ213.69
Ϫ20.81a
Ϫ16.85b
42.42c
8
TS1
P1c
Ϫ30.83c
a The energy of the reactants has been calculated as the sum of that of
the nitrone and zinc() triflate, i.e.: Ϫ395.16 kcal molϪ1 b The energy of
.
the reactants has been calculated as the sum of that of the nitrone,
zinc() triflate and allylic alcohol, from which the energy corresponding
to a molecule of triflic acid (Ϫ260.33 kcal molϪ1) has been subtracted,
i.e.: Ϫ166.13 kcal molϪ1 c Complex 8 has been considered as the start-
.
ing point.
Table 3 Heats of formation (PM3) of the reagents, starting com-
plexes, transition structure and product for the intramolecular path
shown in Scheme 4
Energy (kcal molϪ1
)
Relative energy to reactantsa
Scheme 5
7
Ϫ415.97
Ϫ31.30
Ϫ404.49
Ϫ405.85
Ϫ485.83
Ϫ486.42
Allylic alcohol
TS2x
42.78
41.42
Ϫ38.56
Ϫ39.15
TS2n
P2c
P2t
a The energy of the reactants has been calculated as the sum of complex
7 and allylic alcohol, i.e.: Ϫ447.27 kcal molϪ1
.
alcohol), transition structures and diastereomeric products are
given in Table 3. The optimized structures for transition struc-
tures and final adducts are shown in Fig. 4. In this case, the
preferential formation of the cis adduct (from an exo attack) is
not well-predicted although close values are obtained for the
addition by the two channels.
For the purpose of comparison, the complete profiles of the
two approaches are shown in Fig. 5. The compared energy
values are, in all cases, relative values. Admittedly, the energy
Fig.
4
Optimized structures of complexes, transition state and
product for the intramolecular path depicted in Scheme 5 (hydrogen
atoms have been omitted for clarity).
values for the two approaches are too close and fall into
the range of the experimental error (less than 5 kcal molϪ1
)
considering that the reaction is carried out in acetone at reflux
(ca. 330 K). However, the fact that only one isomer is obtained
supports the hypothesis of the intramolecular mechanism,
since the values found for the TS’s corresponding to the inter-
molecular mechanism predict mixtures of isomers. These
results clearly indicate that semiempirical calculations are not
precise enough to describe the system and higher level calcu-
lations should be made in order to evaluate quantitatively the
energetic differences.
Conclusions
In summary, we have demonstrated that the intermediate com-
plexes formed in the cycloaddition of N-benzyl-C-(2-pyridyl)
nitrone with allylic alcohol when metals complexes are added
to promote the reaction, can be isolated and characterized by
X-ray crystallography. It has also been shown that the com-
Fig.
3
Optimized structures of complexes, transition state and
product for the intramolecular path depicted in Scheme 4 (hydrogen
atoms have been omitted for clarity).
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 3 3 6 – 2 3 4 2
2339

View Article Online
residue was purified by preparative, centrifugally accelerated,
radial, thin-layer chromatography (Chromatotron®).
With Lewis acid. To a 0.1 M solution of the nitrone 2
(0.212 g, 1 mmol) in acetone (10 mL) was added the corre-
sponding metal complex (1 mmol). The resulting mixture
was stirred at ambient temperature for 15 min and then allylic
alcohol (0.58 g, 10 mmol) was added and the mixture
was heated at reflux under Ar atmosphere for the time indicated
in Table 1. The solvent was evaporated and the products
ratio was established by NMR analysis. The residue was puri-
fied by preparative, centrifugally accelerated, radial, thin-layer
chromatography (Chromatotron®).
3a: (0.265 g, 98%; Table 1, entry 4); oil (Found: C, 71.3; H,
6.5; N, 10.6. Calc. for C16H18N2O6: C, 71.1; H, 6.7; N, 10.4%).
δH(CDCl3) 2.53 (dt, 1H, J 12.9 and 5.6 Hz, H4a), 2.88 (dt, 1H,
J 12.9 and 8.6 Hz, H4b), 3.52 (br s, 1H, ex. D2O, CH2OH ), 3.67
(dd, 1H, J 4.8 and 12.0 Hz, CH2OH), 3.81 (dd, 1H, J 2.6 and
12.0 Hz, CH2OH), 3.93 (d, 1H, J 13.5 Hz, NCH2Ph), 3.99 (d,
1H, J 13.5 Hz, NCH2Ph), 4.19 (dd, 1H, J 5.6 and 8.6 Hz, H3),
4.49 (m, 1H, H5), 7.16 (m, 1H, PyH5), 7.39 (m, 1H, PyH3), 7.40
(m, 5H, Ph), 7.63 (m, 1H, PyH4), 8.50 (m, 1H, PyH6).
δC(CDCl3) 36.6 (C4), 60.2 (NCH2Ph), 63.7 (CH2OH), 64.0 (C3),
69.8 (C5), 121.7 (PyC5), 122.4 (PyC3), 127.1 (PhCH), 128.1
(PhCH × 2), 128.6 (PhCH × 2), 136.9 (PyC4), 137.0 (PhC),
148.7 (PyC6), 159.9 (PyC2).
Fig.
5 Reaction profiles for both intra- (continuous line) and
intermolecular (dotted line) paths illustrated in Schemes 4 and 5,
respectively. See notes of Tables 2 and 3 for definitions of relative
energies in each case.
3b: (0.073 g, 27%; Table 1 entry 1); oil (Found: C, 70.9; H,
6.4; N, 10.2. Calc. for C16H18N2O6: C, 71.1; H, 6.7; N, 10.4%).
δH(CDCl3) 2.61 (m, 2H, H4a and H4b), 3.58 (dd, 1H, J 4.8 and
11.7 Hz, CH2OH), 3.60 (br s, 1H, ex. D2O, CH2OH ), 3.72 (dd,
1H, J 2.9 and 11.7 Hz, CH2OH), 3.93 (d, 1H, J 13.5 Hz,
NCH2Ph), 3.99 (d, 1H, J 13.5 Hz, NCH2Ph), 4.11 (t, 1H, J 7.4
Hz, H3), 4.28 (m, 1H, H5), 7.12 (m, 1H, PyH5), 7.29 (m, 5H,
Ph), 7.54 (m, 1H, PyH3), 7.62 (m, 1H, PyH4), 8.48 (m, 1H,
PyH6). δC(CDCl3) 38.3 (C4), 61.1 (NCH2Ph), 63.7 (CH2OH),
64.0 (C3), 70.9 (C5), 122.5 (PyC5), 124.4 (PyC3), 127.3 (PhCH),
128.3 (PhCH × 2), 129.0 (PhCH × 2), 137.0 (PyC4), 137.1
(PhC), 149.4 (PyC6), 160.2 (PyC2).
plexes are real intermediates of the reaction. The semiempirical
study of the reaction does not agree with the experimental
observations. However, intramolecular cycloadditions remain
as a possibility since it has been demonstrated that it is
energetically possible. Although an intermolecular approach is
predicted, the energy differences should not be considered
quantitatively due to both limitations of the method and the
proximity of energy values for a reaction carried out at 330 K.
The extension of the method leading to asymmetric induction
is currently the object of intensive investigation.
Experimental
[Ag(OTf)(2-PyBN)]2 4
General methods
To a diethyl ether solution (20 mL) of nitrone 2 (2-PyBN)
(0.042 g, 0.2 mmol) under Argon atmosphere was added
[AgOTf] (0.042 g, 0.2 mmol). A solid precipitated immediately
which was filtered off and washed with diethyl ether (2 × 5 mL)
to give 0.071 g, (75%) as a white solid. Mp 116 ЊC (from diethyl
ether) (Found: C, 35.49; H, 2.75; S, 6.61. Calc. for C28H24N4-
The reaction flasks and other glass equipment were heated in
an oven at 130 ЊC overnight and assembled in a stream of Ar.
All reactions were monitored by TLC on silica gel 60 F254; the
position of the spots was detected with 254 nm UV light or
by spraying with one of the following staining systems: 50%
methanolic sulfuric acid, 5% ethanolic phosphomolybdic acid
and iodine. Preparative centrifugally accelerated radial thin-
layer chromatography (PCAR-TLC) was performed with a
Chromatotron® Model 7924 T (Harrison Research, Palo Alto,
CA, USA) and with solvents that were distilled prior to use; the
rotors (1 or 2 mm layer thickness) were coated with silica gel
Merck grade type 7749, TLC grade, with binder and fluor-
escence indicator (Aldrich 34,644–6) and the eluting solvents
O Ag F S : C, 35.84; H, 2.58; S, 6.83%). IR(Nujol): ν (C᎐N)
᎐
8
2
6 2
1593, ν (TfO) 1290, 1237, 1167 cmϪ1. δH(acetone-d6): 5.23 (s,
4H), 7.30–7.35 (m, 6H), 7.50–7.65 (m, 6H), 8.15 (dt, 2H, J 1.0
and 7.8 Hz), 8.49 (s, 2H), 8.75–8.85 (m, 4H). δF(acetone-d6)
Ϫ77.5 (s). δC(acetone-d6) 76.6, 130.4, 131.0, 133.8, 134.0, 134.4,
135.0, 139.0, 139.6, 144.5, 154.6, 157.5, 210.8. m/z (LSISMϩ)
787 (10%) [M Ϫ TfO]ϩ; 575 (8) [M Ϫ TfO Ϫ (2-PyBN)]ϩ; 531
(33) [Ag(2-PyBN)2]ϩ, 319 (100) [Ag(2-PyBN)]ϩ.
were delivered by the pump at a flow-rate of 0.5–1.5 mL minϪ1
.
1
Melting points were uncorrected. H, 13C, 19F and 31P NMR
spectra were recorded on a Varian Unity or on a Bruker 300
instrument in CDCl3. Elemental analysis were performed on
a Perkin Elmer 240B microanalyzer. Pyridyl nitrone 2 was
prepared as described.16
[Ag(2-PyBN)(PPh2Me)]2(ClO4)2 5
To a diethyl ether solution (20 mL) of nitrone 2 (2-PyBN)
(0.042 g, 0.2 mmol) under an argon atmosphere was added
[Ag(OClO3)(PPh2Me)] (0.081 g, 0.2 mmol). After 12 h of stir-
ring the obtained solid was filtered off and washed with diethyl
ether (2 × 5 mL) to give 5 (0.109 g, 88%) as a white solid. Mp:
172 ЊC (from diethyl ether) (Found: C, 51.40; H, 3.99; N,
2.10. Calc. for C54H52N2O10Ag2Cl2P2: C, 51.41; H, 4.23; N,
[(3S*,5R*)-2-Benzyl-3-pyridin-2-ylisoxazolidin-5-yl]methanol
3a and [(3R*,5R*)-2-benzyl-3-pyridin-2-ylisoxazolidin-5-yl]-
methanol 3b
2.26%). IR(Nujol): ν (C᎐N) 1599, ν (ClO ) 1097, 624 cmϪ1
.
᎐
4
Without Lewis acid. To a solution of the nitrone 2 (0.212 g,
1 mmol) in acetone (10 mL) was added allylic alcohol (0.58 g,
10 mmol) and the resulting mixture was heated at reflux under
Ar atmosphere for seven days. The solvent was evaporated and
the products ratio was established by NMR analysis. The
δH(CDCl3) 1.93 (d, 6H, J 7.6 Hz), 5.24 (s, 4H), 5.29–5.53
(m, 36H), 7.72 (s, 2H), 8.17–8.20 (m, 4H). δP(CDCl3, Ϫ60 ЊC)
Ϫ2.8 (dd, J(107Ag–P) 734 Hz, J(109Ag–P) 846 Hz). δC(CDCl3)
13.1, 13.4, 71.8, 125.5, 128.8, 129.2, 129.3, 129.8, 131.0, 131.1,
132.3, 132.5, 132.7, 134.6, 139.3. m/z (LSISMϩ) 518 (83%)
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 3 3 6 – 2 3 4 2
2340

View Article Online
[Ag(2-PyBN)(PPh2Me)]ϩ; 507 (60) [Ag(PPh2Me)2]ϩ; 307 (100)
[AgPPh2Me]ϩ.
2353–2360; (e) W. W. Ellis, L-S. Gavrilova, A. L. Rheingold and
B. Bosnish, Organometallics, 1999, 18, 332–338; ( f ) S. M. Mullins,
R. G. Bergman and J. Arnold, Organometallics, 1999, 18, 4465–
4467.
[ZnBr2(2-PyBN)] 6 and [Zn(OTf)2(2-PyBN)] 7
6 (a) O. Tamura, N. Mita, K. Gotanda, K. Yamada, T. Nakano,
R. Katagiri and M. Sakamoto, Heterocycles, 1997, 46, 95–99;
(b) Y. Ukaji, K. Taniguchi, K. Sada and K. Inomata, Chem. Lett.,
1997, 547–548; (c) S. Kanemasa and T. Tsuruoka, Chem. Lett., 1995,
49–50; (d ) S. Kanemasa, T. Tsuruoka and H. Yamamoto,
Tetrahedron Lett., 1995, 36, 5019–5022.
7 The use of chelating hetaryl nitrones as ligands has been
scarcely investigated. See: (a) M. L. Kahn, J.-P. Sutter, S. Golhen,
P. Guionneau, L. Ouahab, O. Kahn and D. Chasseau, J. Am. Chem.
Soc., 2000, 122, 3413–3421; (b) E. G. Petkova, R. D. Lampeka,
M. V. Gorichko and K. V. Domasevitch, Polyhedron, 2001, 20, 747–
753; (c) P. Das, M. Boruah, N. Kumari, M. Sharma, D. Konwar
and D. K. Dutta, J. Mol. Catal. A: Chem., 2002, 178, 283–287;
(d ) F. A. Villamena, M.-H. Dickman and D. R. Crist, Inorg. Chem.,
1998, 37, 1446–1453.
8 (a) P. Merino, S. Franco, N. Garces, F. L. Merchan and T. Tejero,
Chem. Commun., 1998, 493–494; (b) P. Merino, S. Franco,
F. L. Merchan and T. Tejero, Tetrahedron Lett., 1998, 39, 6411–6414;
(c) P. Merino, E. M. Del Alamo, M. Bona, S. Franco, F. L. Merchan,
T. Tejero and O. Vieceli, Tetrahedron Lett., 2000, 41, 9239–9243.
9 For reviews illustrating the biological interest of heterocyclic
nucleosides see: (a) T. Mansour and R. Storer, Curr. Pharm. Des.,
1997, 3, 227–264; (b) P. Merino, Curr. Med. Chem. Anti-Infect.
Agents, 2002, 1, 389–411.
To an acetone solution (20 mL) of nitrone 2 (2-PyBN) (0.042 g,
0.2 mmol) under an argon atmosphere was added ZnBr2 (0.045
g, 0.2 mmol) or Zn(TfO)2 (0.073 g, 0.2 mmol). After 4 h of
stirring the solution was concentrated under pressure and the
residue was triturated with diethyl ether to give a white solid
which was filtered off and washed with diethyl ether (2 × 5 mL).
6: (0.075 g, 85%). white solid; mp:249–250 ЊC (from diethyl
ether) (Found: C, 35.50; H, 2.60; N, 6.64. Calc. for C13H12N2-
OBr Zn: C, 35.73; H, 2.77; N, 6.41%). IR(Nujol): ν (C᎐N) 1617
᎐
2
cmϪ1. δH(acetone-d6) 5.40 (s, 2H), 7.41–7.45 (m, 3H), 7.60–7.64
(m, 2H), 8.00 (t, 1H, J 6.2 Hz), 8.07 (d, 1H, J 7.8 Hz), 8.41
(dt, 1H, J 1.7 and 7.8 Hz), 8.76 (d, 1H, J 4.5 Hz), 8.80 (s, 1H).
δC(acetone-d6) 149.6, 141.2, 137.8, 132.8, 130.0, 129.5, 129.1,
128.4, 71.2. m/z (LSISMϩ) 567 (37%) [M Ϫ Br ϩ (2-PyBN)]ϩ;
355 (66) [M Ϫ Br]ϩ.
7: (0.104 g, 90%). white solid; mp 253–255 ЊC (from diethyl
ether) (Found: C, 31.69; H, 2.28; N, 4.50; S, 11.12. Calcd. for
C15H12N2O7F6S2Zn: C, 31.31; H, 2.10; N, 4.87; S, 11.14%).
IR(Nujol): ν (C᎐N) 1617, ν (TfO) 1288, 1246, 1188 cmϪ1
.
᎐
δH(acetone-d6) 5.37 (s, 2H), 7.42–7.52 (m, 5H), 7.83 (t, 1H, J 6.6
Hz), 8.04 (d, 1H, J 7.5 Hz), 8.40 (t, 1H, J 7.8 Hz), 8.78 (br s,
2H). δC(acetone-d6) 150.6, 150.5, 147.6, 141.5, 138.0, 133.2,
130.9, 130.9, 130.1, 129.5, 128.4, 71.1. δF(acetone-d6) Ϫ74.6 (s).
m/z (LSISMϩ) 637 (100%) [M Ϫ TfO ϩ (2-PyBN)]ϩ; 425 (42)
[M Ϫ TfO]ϩ.
10 (a) P. Merino, S. Franco, F. L. Merchan and T. Tejero, J. Org. Chem.,
2000, 65, 5575–5589; (b) P. Merino, E. M. Del Alamo, S. Franco,
F. L. Merchan, A. Simon and T. Tejero, Tetrahedron: Asymmetry,
2000, 11, 1543–1546.
11 To the best of our knowledge only one report has been
recently published concerning isoxazolidinyl C-nucleosides. See:
U. Chiacchio, A. Corsaro, J. A. Mates, P. Merino, A. Piperno,
A. Rescifina, G. Romeo, R. Romeo and T. Tejero, Tetrahedron, 2003,
59, 4733–4738.
12 Pyrrolidinyl C-nucleosides: (a) R. W. Miles, P. C. Tyler, R. H.
Furneaux, C. K. Bagdassarian and V. L. Schramm, Biochemistry,
1998, 37, 8615–8621; (b) W. Shi, C. M. Li, P. C. Tyler, R. H.
Furneaux, S. M. Cahill, M. E. Girvin, C. Grubmerer, V. L. Schramm
and S. C. Almo, Biochemistry, 1999, 38, 9872–9880; (c) R. W. Miles,
P. C. Tyler, G. B. Evans, R. H. Furneaux, D. W. Parkin and
V. L. Schramm, Biochemistry, 1999, 38, 13147–13154; (d ) G. B.
Evans, R. H. Furneaux, G. J. Gainsford, V. L. Schramm and
P. C. Tyler, Tetrahedron, 2000, 56, 3053–3062; Carbocyclic
C-nucleosides: (e) B. K. Chun, G. Y. Song and C. K. Chu, J. Org.
Chem., 2001, 66, 4852–4858; ( f ) N. Katagiri, T. Haneda,
E. Hayasaka, N. Watanabe and C. Kaneko, J. Org. Chem., 1988, 53,
226–227; (g) S. Hildbrand, C. Leuman and R. Scheffold, Helv. Chim.
Acta, 1996, 79, 702–709.
13 To the best of our knowledge, there is only one report in the
literature in which an intermediate nitrone–metal complex has been
isolated and characterized by X-ray analysis. See ref. 5e. For other
references in which nitrones are used as ligands in metal complexes
see: (a) W. Kliegel, J. Metge, S. J. Rettig and J. Trotter, Can. J. Chem.,
1998, 76, 1082–1092; (b) F. A. Villamena and D. R. Crist, J. Chem.
Soc., Dalton Trans., 1998, 4055–4064 See also ref. 7.
Acknowledgements
We thank the Spanish Ministry of Science and Technology
and FEDER Program (Projects CASANDRA: BQU2001-2428
and PB98-0542) and the Government of Aragon (Project
P116-2001) for their support of our programs.
References
1 (a) J. J. Tuffariello, in 1,3-Dipolar Cycloaddition Chemistry, ed. A.
Padwa, John Wiley & Sons, New York, 1984, vol. 2, p. 83; (b) P. N.
Confalone and E. M. Huie, Org. React., 1988, 36, 1–42; (c) K. G. B.
Torssell, Nitrile Oxides, Nitrones and Nitronates in Organic
Synthesis, VCH, Weinheim, 1988.
2 (a) S. Kanemasa, N. Ueno and H. Shirahase, Tetrahedron Lett.,
2002, 43, 657–660; and references cited therein. For reviews see:
(b) K. V. Gothelf and K. A. Jorgensen, Chem. Commun., 2000,
1449–1458; (c) K. V. Gothelf and K. A. Jorgensen, Chem. Rev., 1998,
98, 863–909.
3 See inter alia: (a) G. Faita, A. Paio, P. Quadrelli, F. Rancati and
P. Seneci, Tetrahedron, 2001, 57, 8313–8332; (b) G. Desimoni,
G. Faita, M. Mella, P. Righetti and M. Sema, Tetrahedron, 1999,
55, 8509–8524; (c) K. Hori, H. Kodama, T. Ohta and I. Furukawa,
J. Org. Chem., 1999, 64, 5017–5023; (d ) S. Kanemasa,
Y. Oderaotoshi, J. Tanaka and E. Wada, J. Am. Chem. Soc., 1998,
120, 12355–12356; (e) K. V. Gothelf, R. G. Hazell and K. A.
Jorgensen, J. Org. Chem., 1998, 63, 5483–5488; ( f ) K. B. Jensen,
K. V. Gothelf, R. G. Hazell and K. A. Jorgensen, J. Org. Chem.,
1997, 62, 2471–2477; (g) K. V. Gothelf, R. G. Hazell and K. A.
Jorgensen, J. Org. Chem., 1996, 61, 346–355 For a recent example
with coordinated organonitriles see: G. Wagner, A. J. L. Rombeiro
and V. Y. Kukushkin, J. Am. Chem. Soc., 2000, 122, 3106–3111.
4 According to the classical FMO theory metal complexation of
electron-poor dipolarophiles enhances their reactivity towards
nitrones. On the other hand, in cycloadditions of nitrones with
electron-rich alkenes (inverse demand reactions) nitrone complexes
should be more reactive.
14 R. F. Borch, M. D. Berstein and M. D. Durst, J. Am. Chem. Soc.,
1971, 93, 2897–2904.
15 The graphic views showed in Figs. 1–4 were made with ORTEP3
software. Copyright by L. J. Farrugia, University of Glasgow. 1997–
2000. CCDC reference numbers 208653 and 208654. See http://
www.rsc.org/suppdata/ob/b3/b304112c/ for crystallographic data in
.cif or other electronic format.
16 For a preliminary communication on the crystal structure of nitrone
2 and complex 4 see: P. Merino, S. Anoro, E. Cerrada, M. Laguna,
A. Moreno and T. Tejero, Molecules, 2001, 6, 208–220.
17 (a) G. A. Lawrence, Chem. Rev., 1986, 86, 17–89; (b) D. H. Johnston
and D. F. Shriver, Inorg. Chem., 1993, 32, 1045–1047; (c) R. Terroba,
M. B. Hursthouse, M. Laguna and A. Mendía, Polyhedron, 1999, 18,
807–810.
18 G. Scherhay and M. D. Spicer, J. Chem. Soc., Dalton Trans., 2000,
1237–1238.
19 P. J. Stang, Y.-H. Hwang and A. M. Arif, Organometallics, 1992, 11,
231–237.
5 (a) P. Bayon, P. De March, M. Figueredo, J. Font and J. Medrano,
Tetrahedron: Asymmetry, 2000, 11, 4269–4278; (b) P. Bayon, P. De
March, M. Espinosa, M. Figueredo and J. Font, Tetrahedron:
Asymmetry, 2000, 11, 1757–1765; (c) K. B. Jensen, M. Roberson and
K. A. Jorgensen, J. Org. Chem., 2000, 65, 9080–9084; (d ) K. B.
Jensen, R. G. Hazell and K. A. Jorgensen, J. Org. Chem., 1999, 64,
20 (a) A. Cingolani, F. Marcehti, C. Pettinari, B. W. Skelton and
A. H. White, J. Chem. Soc., Dalton Trans., 1999, 4047–4055;
(b) G. Smith, B. A. Cloutt, D. E. Lynch, K. A. Byriel and C. H. L.
Kennard, Inorg. Chem., 1998, 37, 3236–3242; (c) N. J. Calos, C. H. L.
Kennard, T. C. W. Mak and G. Smith, Aust. J. Chem., 1989, 42,
2047–2052.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 3 3 6 – 2 3 4 2
2341

View Article Online
21 (a) P. C. Healy, N. K. Mills and A. H. White, J. Chem. Soc.,
Dalton Trans., 1985, 111; (b) M. Bardají, O. Crespo, A. Laguna and
A. K. Fischer, Inorg. Chim. Acta., 2000, 304, 7–16; (c) E. Effendy,
G. G. Lobbia, M. Pellei, C. Pettinari, C. Santini, B. W. Skelton and
A. H. White, Inorg. Chim. Acta, 2001, 315, 153–162.
22 B. Müller and H. Vahrenkamp, Eur. J. Inorg. Chem., 1999, 137–
144.
23 For a recent theoretical study on nitrone complexes with Lewis acids
see: J. Tanaka and S. Kanemasa, Tetrahedron, 2001, 57, 899–905.
24 Several comparative studies have suggested semiempirical methods
(a) A. Sbai, V. Branchadell, R. M. Ortuño and A. Oliva, J. Org.
Chem., 1997, 62, 3049–3054; (b) R. Annunziata, M. Benaglia,
M. Cinquini, F. Cozzi and L. Raimondi, J. Org. Chem., 1995, 60,
4697–4706.
25 J. J. P. Stewart, in Reviews in computational chemistry, eds. K. B.
Lipkowitz and D. B. Boyd, VCH, New York, 1990, pp. 45–81.
26 MOPAC 2000 package was used as implemented in the
ChemOffice 2002 suite of programs. CambridgeSoft corporation
(Cambridge, USA).
27 J. J. P. Stewart, J. Comput. Chem., 1989, 10, 20.
28 K. Fukui, J. Phys. Chem., 1970, 74, 4161.
as
a good choice for studying cycloaddition processes. See:
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 3 3 6 – 2 3 4 2
2342