First 1,3-dipolar cycloaddition of Z-α-phenyl-N-methylnitrone with allylic fluorides: A stereoselective route to enantiopure fluorine-containing isoxazolidines and amino polyols

Article abstract of DOI:10.1016/j.tetasy.2003.11.008

Enantiopure fluorinated isoxazolidines and amino polyols were obtained from the 1,3-dipolar cycloaddition of allylic fluorides and Z-α-phenyl-N- methylnitrone under environment-friendly conditions.

Full text of DOI:10.1016/j.tetasy.2003.11.008

Tetrahedron:
Asymmetry
Tetrahedron: Asymmetry 15 (2004) 245–250
First 1,3-dipolar cycloaddition of Z-a-phenyl-N-methylnitrone with
allylic fluorides: a stereoselective route to enantiopure
fluorine-containing isoxazolidines and amino polyols
Luca Bernardi, Bianca F. Bonini, Mauro Comes-Franchini,* Mariafrancesca Fochi,
Mahena Folegatti, Stefano Grilli, Andrea Mazzanti and Alfredo Ricci
Dipartimento di Chimica Organica ‘A. Mangini’, Universitꢀa di Bologna, Viale Risorgimento 4, 40136 Bologna, Italy
Received 24 September 2003; accepted 4 November 2003
Abstract—Enantiopure fluorinated isoxazolidines and amino polyols were obtained from the 1,3-dipolar cycloaddition of allylic
fluorides and Z-a-phenyl-N-methylnitrone under environment-friendly conditions.
ꢀ 2003 Elsevier Ltd. All right reserved.
1. Introduction
2. Results and discussion
The introduction of a fluorine atom into biologically
active compounds often induces strong modifications of
their chemical, physical and biological properties. Due
to this, the interest in fluorine-containing compounds
has greatly increased over the last few decades,1 since
enantiopure fluorine-containing heterocycles and fluo-
rinated amino polyols2 have shown great potential as
drug candidates.
The reaction of 1 with racemic methyl 2-[fluoro-
(phenyl)methyl]acrylate 25a has been investigated as a
model reaction, with the final aim of applying this
protocol to allylic fluorides derived from the natural
pool. In toluene as solvent under In(OTf)3 catalysis
(5 mol %), complete conversion of the starting materials
was achieved after 24 h at reflux. Analysis of the crude
mixtures using 19F NMR showed three new sets of
doublets for the CH–F signals, thus indicating the for-
mation of three cycloadducts 3a–c, in a 10:60:30 ratio.
As far as the regiochemistry was concerned, after sepa-
ration of the three diastereoisomers,6 obtained in 46%
overall yield, NMR analysis (COSY and gHSQC) con-
firmed that the three products were 5,50-disubstituted
isoxazolidines in agreement with the majority of the
literature findings.7
As part of our ongoing interest in the stereoselective
synthesis of enantiopure isoxazolidines containing a
carbon–fluorine centre, we herein report on the 1,3-
dipolar cycloaddition3 of Z-a-phenyl-N-methylnitrone 1
with allylic fluorides, focusing on the use of environ-
ment-friendly conditions,4 such as microwave irradia-
tion and metal triflate catalysis. Moreover, since the
reductive opening of isoxazolidine is a well known
method for obtaining amino alcohols, this protocol
might provide an easy entry to enantiopure fluorinated
amino polyols. Our approach is based on the utilization
of allylic fluorides as dipolarophiles. Very little is known
at present about the effect of these moieties on stereo-
selective organic transformations.5
Stereochemical assignment of the three adducts was
based upon 1D-NOE experiments. Selective excitation8
of the CH(F) signal in the first isolated adduct 3a
revealed a positive NOE effect only on one of the two
diastereotopic CH2 signals (H4a in Scheme 1). This result
is compatible with a structure in which the phenyl group
and the CH(F)Ph group are on the same side of the ring
3a (exo-structure). On the other hand, excitation of the
CH(F) signals in the remaining adducts, 3b and 3c,
showed NOE enhancements of the benzylic proton H3
and of the other diastereotopic CH (H4b) in line with
arrangements in which the phenyl group and the
CH(F)Ph group are on the opposite side of the ring. The
Keywords: Allylic fluorides; Nitrone; 1,3-Dipolar cycloaddition;
Isoxazolidines; Amino polyols.
* Corresponding author. Tel.: +39-05-12093626; fax: +39-05-120936-
54; e-mail: comes@ms.fci.unibo.it
0957-4166/$ - see front matter ꢀ 2003 Elsevier Ltd. All right reserved.
doi:10.1016/j.tetasy.2003.11.008

246
L. Bernardi et al. / Tetrahedron: Asymmetry 15 (2004) 245–250
ronmentally benign solvents for transition metal cata-
lyzed organic transformations. Catalytic amounts of
triflates (5 mol %) were used throughout. As shown in
Table 1 neither the yields nor the exo/endo ratio could be
substantially improved by running the reaction in dif-
ferent media.13 A remarkably shorter reaction time was
obtained under microwave irradiation without affecting
the yields (entries 2, 4, 6). A substantial improvement
was finally achieved when the reaction of 1 with 2 was
performed under microwave activation without any
solvent in the presence of a catalytic amount of metal
triflates. In two cases (entries 7 and 8), the cycloaddition
proceeded smoothly to afford 3a–c in 72–75% yields
after 15 min.
Scheme 1.
The final goal of this research was the development of
asymmetric routes to unknown enantiomerically pure
fluorinated isoxazolidines and to fluorinated amino
polyols. For this, we applied this synthetic protocol to
an enantiopure allylic fluoride. (S)-2-Benzyloxyprop-
anal, a very useful building block for stereoselective
synthesis, was the starting material of choice. Its Baylis–
Hilmann14 reaction with methyl acrylate in DMF gave,
after 14 days, an 87% yield of the allylic alcohols 4 in an
anti/syn ratio of 75:25.15 Fluorination of the isolated
anti-alcohol with DAST in CH2Cl2 yielded the enan-
tiopure allylic fluoride 5 in 82% yield with retention of
configuration5a (Scheme 2).
unambiguous assignment of the two diastereotopic CH
(H4a and H4b) was possible because of the long range
(4J) coupling constant of H4a with the fluorine atom.
This coupling constant can be observed only when the
ÔWÕ path is available.10 Since 3b and 3c differ only in
the stereochemistry at the carbon–fluorine bond and the
starting allylic fluoride 2 being racemic, an overall endo/
exo ratio of 90:10 can be calculated for this reaction.
A screening aimed at the optimization of the reaction
conditions was then performed with the results shown in
Table 1.
Cycloaddition of 5 with 1 in the presence of 5% In(OTf)3
took place under the optimized solvent-free conditions
and microwave irradiation at 60 W for 25 min to give a
mixture of the 5,50-isoxazolidine isomers 6a–b in 61%
overall yield in a 90:10 ratio (Scheme 3).
The use of microwave irradiation as an alternative
method of heating reaction mixtures in cycloaddition
reactions has been reported to dramatically reduce
reaction times and affect product ratios and yields.10 We
applied this technique to the reaction shown in Scheme 1
performed in various reaction media such as water, ionic
liquids (RTILs) and under solvent-free conditions with
the aim of finding the optimal reaction conditions
leading to 3a–c. The use of these reaction media was
inspired by the well known acceleration of cyclo-
additions in water4;11 and the recently reported great
potential of ionic liquids (RTILs)12 as novel and envi-
Semi-preparative HPLC allowed the separation of the
two compounds with the relative assignment of 6a and
6b obtained from NOE data. Selective saturation9 of the
HF hydrogen shows a positive NOE on one of the two
diastereotopic HÕs at position 4 (Ha in Fig. 1). Selective
irradiation of the latter reveals an NOE on the ortho
Table 1. 1,3-Dipolar cycloaddition reactions of 1 with 2
Entry
Solvent
MW (W)
T (ꢁC)/Time
Yield (%)
Lewis acid (5%)
Ratio 3a/3b/3c
1
2
3
4
5
6
7
8
Toluene
Toluene
H2O
––
120
––
90
––
90
60
60
110/24 h
110/10 min
rt/72 h
46
48
50
49
19
39
73
75
In(OTf)3
In(OTf)3
Yb(OTf)3
Yb(OTf)3
In(OTf)3
In(OTf)3
Sc(OTf)3
In(OTf)3
10:60:30
10:60:30
10:60:30
10:60:30
10:60:30
10:60:30
10:60:30
10:60:30
H2O
100/7 min
80/16 h
BimimBF4
Toluene/BimimBF4
None
165/10 min
100/15 min
100/15 min
None
Scheme 2.

L. Bernardi et al. / Tetrahedron: Asymmetry 15 (2004) 245–250
247
Scheme 3.
hydrogens of the phenyl group in one of the two dia-
stereoisomers 6a (in Fig. 1), and on the H at position 3
(HN) for the other diastereoisomer 6b (Fig. 1).
allylic oxygen and the nitrone oxygen.16 Accordingly,
even though our understanding is at a preliminary level,
we propose that also the strong electron-withdrawing
fluorine atom might also increase the exo-preference in
the transition state. Further studies are currently in
progress to rationalize the chemical behaviour of allylic
fluorides in 1,3-dipolar cycloadditions.
Reductive opening of the cycloadducts with LiAlH4 in
THF took place cleanly in 2 h affording 7a and 7b in 80–
81% yields in which the ester functionality was also
reduced (Scheme 4). These fluorinated enantiopure
amino polyols are important structures containing up to
four stereocentres, with one of them being quaternary.
Figure 1.
These results imply an exo-structure for the major iso-
mer 6a and an endo-structure for the minor isomer 6b.
The absolute structure can be confirmed by the follow-
ing NOE data. In the case of 6a, selective saturation of
Ha shows NOE effects also on HF and HO signals; sat-
uration of HF shows NOE effects on the signal of HO
and on the methyl group. NOE ratios show that the
distances Ha–HF and Ha–HO are very similar, as well as
the distances HF–HO and HF–Me. Of the four obtain-
able diastereoisomers, only the 5(R)-CHF(R) structure
matched the experimental NOE data (the exo relation-
ship of the cycle is defined, and one stereogenic carbon,
indicated as (S) in Fig. 1, is fixed). By applying the exo
relationships, the configuration (3R,5R,R,S) can be
assigned to 6a. The same approach was used for 6b and
structure (3S,5R,R,S) assigned. This further confirms
the retention of configuration in the DAST fluorination.
Scheme 4.
3. Conclusions
It has been claimed that the stereoselectivity in the 1,3-
dipolar cycloaddition is very much affected by the
structure of the substrates.3 The explanation for the
reversed stereoselectivity observed for 6a–b in compar-
ison with 3a–c is currently under investigation, while as
far as the influence of the fluorine on the reactivity and
stereoselectivity is concerned, we can only rely on
comparison with other structurally similar, non-fluori-
nated dipolarophiles. It has been reported that chiral
allylic alcohols and ethers react with nitrones to give the
5,50-disubstituted cycloadducts.16 The stereochemical
outcome of these cycloadditions is rationalized by an
Ôinside alkoxyÕ transition state model that always leads
to exo/endo mixtures with a preference for the exo-cyc-
loadduct. This outcome was ascribed to the electroneg-
ativity of the OR group in the allylic position, which
prefers the inside to the outside position in the transition
state to avoid the repulsive interaction between the
In summary, the cycloadditions of the allylic fluorides
with nitrone 1 reported herein allow the synthesis of
enantiopure fluorine-containing isoxazolidines and
amino polyols. To the best of our knowledge, this is the
first 1,3-dipolar cycloaddition using allylic fluorides. The
optimization of the reaction has been achieved under
environment-friendly conditions and its extension to the
synthesis of other fluorinated compounds of pharma-
ceutical interest is currently being investigated.
4. Experimental
Melting points (uncorrected) were determined with a
Buchi melting point apparatus. H NMR and 13C NMR
1
€
spectra were recorded using CDCl3 solutions at 300, 400

248
L. Bernardi et al. / Tetrahedron: Asymmetry 15 (2004) 245–250
and 600 MHz for 1H and 75.46, 100.6 and 150.92 MHz
for 13C. Chemical shifts (d) are reported in ppm relative
4.2. General procedure for the 1,3-dipolar cycloaddition
of 2
1
to CHCl3 (d ¼ 7:26 for H and d ¼ 77:0 for 13C or 7.24
and 128.0 ppm for C6D6) and relative to CF3C6H5
2-[Fluoro(phenyl)methyl]acrylate
2
(150 mg, 0.77
1
(d ¼ )163.0 ppm for 19F). J values are given in Hz. H
mmol), In(OTf)3 (22 mg, 0.04 mmol) and Z-a-phenyl-
N-methylnitrone 1 (104 mg, 0.77 mmol) were mixed in a
cylindrical Pyrex sealed vessel. The mixture was intro-
duced into the reactor and irradiated as described in
Table 1. In the cases where toluene or water was used as
solvent, 2 ml of them was added. After cooling, the
mixture was extracted with ethyl acetate (3 · 20 ml),
washed with water (2 · 20 ml) and dried over anhydrous
sodium sulfate. After filtration, removal of the solvent
and purification by column chromatography on silica
gel (CH2Cl2/EtOAc 80:1), the three cycloadducts 3a–c
were obtained as pale yellow oils and fully characterized.
NMR and 13C NMR spectral assignments were made by
DEPT, gCOSY and gHSQC experiments. IR spectra
were recorded on a Perkin–Elmer model 257 grating
spectrometer. Mass spectra were obtained using a VG
7070-E spectrometer at an ionizing voltage of 70 eV.
HPLC analysis has been performed with a Kromasil
column KR100-5SIL (25 cm, id ¼ 4.6 mm) and semi-
preparative separation with a Kromasil 100-10 silica
20
(25 cm, id ¼ 20 mm). ½aꢀ values were determined with
D
Perkin–Elmer polarimeter 341. Reactions under micro-
wave irradiation have been run in a monomode reactor,
with a Synthwave 402 Prolabo focused MW at
2.45 GHz. Reactions were conducted in oven-dried
(120 ꢁC) glassware under a positive Ar atmosphere.
Transfers of anhydrous solvents or mixtures were
accomplished with oven-dried syringes/septum tech-
niques. THF was distilled from sodium/benzophenone
just prior to use and stored under Ar. Toluene was
distilled from sodium. Et2O was distilled from phos-
phorus pentoxide. CH2Cl2 was passed through basic
alumina and distilled from CaH2 prior to use. Other
solvents were purified by standard procedures. Light
petroleum ether refers to the fraction with bp 40–60 ꢁC.
The reactions were monitored by TLC performed on
silica gel plates (Baker-flex IB2-F). Column chroma-
tography was performed with Merck silica gel 60 (70–
230 mesh). Preparative thick layer chromatography was
carried out on glass plates using a 1 mm layer of Merck
silica gel 60 Pf 254. All chemicals were used as obtained
or purified by distillation as needed. Lithium aluminium
hydride (1 M THF) was purchased from Aldrich.
4.2.1. exo-Methyl 5-[fluoro(phenyl)methyl]-2-methyl-3-
phenyltetrahydro-5- isoxazolecarboxylate 3a. m=z (ESI):
352 (Mþ+Na). (Found: C, 69.37; H, 6.19; N, 4.33;
1
C19H20FNO3 requires C, 69.29; H, 6.12; N, 4.25.) H
NMR (300 MHz, CDCl3): d 2.43 (s, 3H, CH3–N), 2.55
(dd, 1H, J ¼ 13.3 Hz, J ¼ 10.6 Hz, CH2), 2.75 (dd, 1H,
J ¼ 13.3 Hz, J ¼ 6.50 Hz, CH2), 2.90 (dd, 1H,
J ¼ 10.6 Hz, J ¼ 6.5 Hz, CH), 3.36 (s, 3H, CH3O), 5.67
(d, 1H, J ¼ 45.1 Hz, CH–F), 6.92–7.35 (m, 10H,
ArCH) ppm. 13C NMR (75.46 MHz, CDCl3, +25 ꢁC): d
43.3 (CH3–N), 45.1 (CH2), 52.6 (CH3O), 72.5 (CH–N),
85.9 (d, J ¼ 22.6 Hz, Cq), 94.4 (d, J ¼ 183.9 Hz, CH–F),
127.1–128.9 (ArCH), 134.7 (d, J ¼ 21.9 Hz, Ar–Cq),
137.0 (Ar–Cq), 170.8 (C@O) ppm. 19F NMR (283 MHz,
CDCl3): d )174.8 (d, J ¼ 45.1 Hz) ppm.
4.2.2. endo-Methyl 5-[fluoro(phenyl)methyl]-2-methyl-3-
phenyltetrahydro-5-isoxazolecarboxylate 3b. m=z (ESI):
352 (Mþ+Na). 1H NMR (300 MHz, CDCl3): d 2.39
(s, 3H, CH3–N), 2.68 (dd, 1H, J ¼ 13.2 Hz, J ¼ 8.4 Hz,
CH2), 2.75 (dd, 1H, J ¼ 13.2 Hz, J ¼ 1.6 Hz, CH2), 2.96
(t, 1H, J ¼ 8.4 Hz, CH), 3.40 (s, 3H, CH3O), 5.64 (d, 1H,
J ¼ 46.0 Hz, CH–F), 6.95–7.15 (m, 10H, ArCH) ppm.
13C NMR (50.30 MHz, CDCl3): d 42.5 (CH3–N), 43.5
(CH2), 52.4 (CH3O), 72.9 (CH–N), 86.1 (d, J ¼ 22.6 Hz,
Cq), 93.9 (d, J ¼ 183.5 Hz, CH–F), 127.2–129.2 (ArCH),
134.5 (d, J ¼ 20.5 Hz, ArCq), 137.5 (ArCq), 172.1
(C@O) ppm. 19F NMR (283 MHz, CDCl3): d )181.05
(d, J ¼ 46.0 Hz) ppm.
4.1. Methyl 2-[(1R,2S)-2-(benzyloxy)-1-fluoropropyl]-
acrylate 2
To a stirred solution of DAST (1.0 ml, 7.6 mmol) in
5.0 ml of CH2Cl2 at )78 ꢁC under argon, methyl
2-[hydroxy(phenyl)methyl]acrylate (1.46 g, 7.6 mmol) in
0.5 ml of CH2Cl2 was added. Stirring was continued at
)78 ꢁC for 5 min and then for 30 min at room temper-
ature. The reaction was diluted with CH2Cl2 (20 ml) and
washed with water, NaHCO3 and brine. The resulting
organic layer was dried over Na2SO4, filtered and con-
centrated in vacuo to give a dark orange oil, which was
purified by silica gel column chromatography (petro-
leum ether/Et2O 10:1) to give 736 mg (57%) of 2 as a pale
yellow oil. IR (KBr): m 3036, 1728, 1100–1000 cmꢁ1. m=z
4.2.3. endo-Methyl 5-[fluoro(phenyl)methyl]-2-methyl-3-
phenyltetrahydro-5-isoxazolecarboxylate 3c. m=z (ESI):
1
352 (Mþ+Na). H NMR (300 MHz, CDCl3, +25 ꢁC): d
2.43 (s, 3H, CH3–N), 2.74 (dd, 1H, J ¼ 12.6 Hz,
J ¼ 7.8 Hz, CH2), 2.92 (dd, 1H, J ¼ 12.6 Hz, J ¼ 9.4 Hz,
CH2), 3.03 (t, 1H, J ¼ 7.6 Hz, CH), 3.31 (s, 3H, CH3O),
5.82 (d, 1H, J ¼ 45.7 Hz, CH–F), 6.90–7.20 (m, 10H,
ArCH) ppm. 13C NMR (50.3 MHz, CDCl3): d 42.6
(CH3–N), 43.7 (CH2), 52.9 (CH3O), 72.7 (CH–N), 85.9
(d, J ¼ 29.6 Hz, Cq), 94.4 (d, J ¼ 182.6 Hz, CH–F),
127.9–129.9 (ArCH), 133.6 (Ar–Cq), 136.5 (Ar–Cq),
172.2 (C@O) ppm. 19F NMR (283 MHz, CDCl3): d
)172.59 (J ¼ 45.7 Hz) ppm.
1
(ESI): 217 (Mþ+Na). H NMR (300 MHz, CDCl3): d
3.37 (s, 3H, COOCH3), 6.04 (s, 1H, CH2@), 6.42 (d, 1H,
J ¼ 46.0 Hz, CH–Ph), 6.47 (d, 1H, J ¼ 2.0 Hz, CH2@),
7.35 (s, 5H, ArCH) ppm. 13C NMR (125.7 MHz,
CDCl3): d 51.9 (CH3O), 90.7 (d, J ¼ 173.4 Hz, CH–F),
127.1–128.9 (ArCH), 137.4 (d, J ¼ 20.5 Hz, Cq),
139.4 (d, J ¼ 22.5 Hz, Cq), 165.2 (d, J ¼ 6.2 Hz,
COOCH3) ppm. 19F NMR (470.38 MHz, CDCl3):
d )171.48 ppm.

L. Bernardi et al. / Tetrahedron: Asymmetry 15 (2004) 245–250
249
4.3. Methyl 2-[2(S)-(benzyloxy)-1-fluoropropyl]acrylate 5
CH2–O), 5.17 (dd, J ¼ 47.8 Hz, J ¼ 7.6 Hz, 1H, CH–F),
7.28–7.33 (m, 2H, Ph), 7.35–7.39 (m, 4H, Ph), 7.47–7.48
(m, 2H, Ph), 7.60–7.62 (m, 2H, Ph). 13C NMR (C6D6,
100.6 MHz): d 15.7 (d, J ¼ 5.0 Hz, CH3–CH), 42.3 (N–
CH3), 42.6 (CH2), 52.1 (CH3–O), 70.8 (CH2–O), 73.6
(CH–Ph), 73.8 (d, J ¼ 25.9 Hz, CH–Me), 85.4 (d,
J ¼ 20.1 Hz, O–C–C@O), 93.3 (d, J ¼ 183.9 Hz, CH–F),
127.7, 127.9, 128.3, 128.4, 128.5, 128.8, 138.2, 139.2,
171.9 (d, J ¼ 7.6 Hz, CO2Me) ppm. 19F NMR (C6D6,
282 MHz): d )198.6 (m, J ¼ 50.2 Hz).
To a solution of (S)-2-benzyloxypropanal (3.22 g,
19.7 mmol) in DMF (20 ml), methyl acrylate (2.66 ml,
29.5 mmol) and DABCO (2.42 g, 21.7 mmol) were added
and the mixture allowed to stir at room temperature for
14 days. The reaction mixture was diluted with EtOAc
(50 ml) and washed with water (2 · 20 ml). The com-
bined organic layers were dried over Na2SO4 and
evaporated under reduced pressure. The residue was
purified by column chromatography on silica gel
(petroleum ether/CH2Cl2/EtOAc 10:10:1) to afford
4.28 g (87%) of 4 in a 75:25 anti/syn ratio. To a stirred
solution of the isolated anti-4 (1.28 g, 5.15 mmol) in
dichloromethane (7.0 ml) at 0 ꢁC, DAST (0.60 ml,
5.15 mmol) was added dropwise as a neat liquid. After
1 min, the cold bath was removed and the reaction
mixture allowed to reach )30 ꢁC. Upon completion (6 h)
the reaction was worked up by the slow addition of
sodium bicarbonate solution (2 ml) followed by washing
with water (3 ml). Pure product 5 (1.06 mg, 82%) was
isolated by column chromatography on silica gel
4.4.2. endo-Methyl (3S,5R)-5-[(1R,2S)-2-(benzyloxy)-1-
fluoropropyl]-2-methyl-3-phenyltetrahydro-5-isoxazolecar-
boxylate 6b. HPLC (hexane/i-PrOH 99.5:0.5): elution
time 13.99 min. ½aꢀ ¼ +7.95 (c ¼ 1:0, MeOH). m=z
D
(ESI): 410 (Mþ+Na). 1H NMR (C6D6, 600 MHz): d 1.52
(dd, J ¼ 6.2 Hz, J ¼ 3.0 Hz, 3H, CH3–CH), 2.80 (s, 3H,
N–CH3), 3.05 (dd, J ¼ 12.4 Hz, J ¼ 11.0 Hz, 1H, CH2),
3.24 (dd, J ¼ 12.4 Hz, J ¼ 6.0 Hz, 1H, CH2), 3.54 (s, 3H,
CH3–O), 3.88–3.94 (bm, 1H, CH–Ph), 4.02–4.07 (m, 1H,
CH3–CH), 4.46 (d, J ¼ 11.4 Hz, 1H, CH2–O), 4.53 (d,
J ¼ 11.4 Hz, 1H, CH2–O), 5.32 (dd, J ¼ 47 Hz,
J ¼ 6.5 Hz, 1H, CH–F), 7.29–7.34 (m, 4H, Ph), 7.36–
7.39 (m, 2H, Ph), 7.46–7.49 (m, 4H, Ph). 13C NMR
(C6D6, 100.6 MHz): d 15.8 (d, J ¼ 6.1 Hz, CH3–CH),
43.6 (N–CH3), 44.2 (CH2), 51.9 (CH3–O), 70.8 (CH2–
O), 73.1 (CH–Ph), 73.7 (d, J ¼ 25.1 Hz, CH–Me), 84.8
(d, J ¼ 20.1 Hz, O–C–C@O), 95.4 (d, J ¼ 184.7 Hz, CH–
F), 127.7, 127.8, 128.1, 128.2, 128.4, 128.8, 138.1, 138.5,
170.7 (d, J ¼ 7.6 Hz, CO2Me) ppm. 19F NMR (C6D6,
282 MHz): d )197.4 (m, J ¼ 45.9 Hz).
(petroleum ether/Et2O 8:1). ½aꢀ ¼ +3.15 (c ¼ 4:33,
D
CHCl3). m=z (ESI): 275 (Mþ+Na). (Found: C, 66.71; H,
6.87; C14H17FO3 requires C, 66.65; H, 6.79.) H NMR
1
(600 MHz, CDCl3): d 1.18 (dd, J ¼ 6.5 Hz, J ¼ 1.5 Hz,
3H, CH3–CH), 3.73 (s, 3H, CH3–O), 3.76–3.84 (m, 1H,
CH–O), 4.63 (s, 2H, CH2–Ph), 5.48 (m, J ¼ 47.1 Hz, 1H,
CH–F), 5.99 (dd, J ¼ 1.5 Hz, J ¼ 1.0 Hz, 1H, CH2@),
6.42 (ddd, J ¼ 3.4 Hz, J ¼ 1.0 Hz, J ¼ 1.0 Hz, 1H,
CH2@), 7.26–7.36 (m, 5H, C6H5). 13C NMR
(100.6 MHz, CDCl3): d 13.3 (d, J ¼ 6.9 Hz, CH3–CH),
51.9 (CH3O), 70.9 (CH2–Ph), 75.1 (d, J ¼ 22.9 Hz, CH–
O), 91.1 (d, J ¼ 180.9 Hz, CH–F), 127.1 (d, J ¼ 9.9 Hz,
CH2@), 127.6 (C6H5), 127.7 (C6H5), 128.3 (C6H5), 136.8
(C@CH2), 138.2 (ArCH), 165.4 (CO2Me). 19F NMR
4.5. Lithium aluminium hydride opening reactions:
(2R,3R,4S)-4-(benzyloxy)-3-fluoro-2-[(2R)-2-(methyl-
amino)-2-phenylethyl]-1,2-pentanediol 7a
(282 MHz, CDCl3):
J ¼ 46.6 Hz).
d
)200.25 (dd, J ¼ 21.7 Hz,
Cycloadduct 6a (190 mg, 0.57 mmol) was dissolved in
THF (2 ml) and cooled to 0 ꢁC. LiAlH4 (2.3 ml,
2.3 mmol) was slowly added and the mixture stirred for
10 min after which the flask was warmed to room tem-
perature and allowed to stir for a further 1 h. The
reaction was quenched at 0 ꢁC with the slow addition of
20% NaOH (2 ml) and stirring continued for 2 h. The
aqueous solution was extracted three times with ethyl
acetate. The organic layers were dried over Na2SO4,
filtered and evaporated to give a light yellow oil which
was purified by silica gel column chromatography
(EtOAc/petroleum ether 2:1) to give 7a (59 mg, 81%) as
4.4. 1,3-Dipolar cycloaddition with chiral allylic fluoride 5
The cycloaddition was performed following the general
procedure starting from 5 (200 mg, 0.79 mmol), In(OTf)3
(44 mg, 0.079 mmol) and
1 (107 mg, 0.79 mmol).
Microwave irradiation (60 W) for 25 min gave 185 mg
(61%) of 6a/6b in a 95:5 ratio. After column chromato-
graphy (petroleum ether/EtOAc 4:1), the two cyclo-
adducts were isolated together and subjected to
semi-preparative HPLC separation.
a colourless oil. ½aꢀ ¼ +6.3 (c ¼ 0:5, MeOH). m=z (ESI):
D
384 (Mþ+Na). HRMS for C21H28FNO3 [M+NHþ4 ]
requires 361.2053, found 361.2051. 1H NMR (C6D6,
600 MHz): d 1.25 (dd, J ¼ 6.0 Hz, J ¼ 2.8 Hz, 3H, CH3–
CH), 2.28–2.35 (m, 1H, CH2CHN), 2.45 (s, 3H, NMe),
2.55–2.75 (m, 3H, CHN+CH2CHN), 3.60–3.65 (m, 1H,
CHMe), 3.80 (t, 2H, J ¼ 8.9 Hz, CH2O), 3.90 (bs, 3H,
2OH+NH), 4.22 (dd, J ¼ 12.0 Hz, 1H, CH2Ph), 4.30
(dd, 1H, J ¼ 12.0 Hz, CH2Ph), 4.95 (d, 1H, J ¼ 44.1 Hz,
CHF), 7.05–7.40 (m, 10H, ArCH) ppm. 13C NMR
(C6D6, 100.6 MHz): d 12.6 (d, J ¼ 6.0 Hz, CH3–CH),
41.6 (CH2), 42.6 (N–CH3), 66.6 (CH2OH), 73.8 (CHN),
83.1 (Cq, d, J ¼ 25.0 Hz), 73.1 (CH–Ph), 73.1 (d,
4.4.1. exo-Methyl (3R,5R)-5-[(1R,2S)-2-(benzyloxy)-1-
fluoropropyl]-2-methyl-3-phenyltetrahydro-5-isoxazolecar-
boxylate 6a. HPLC (hexane/i-PrOH 99.5:0.5): elution
time 9.78 min. ½aꢀ ¼ +7.3 (c ¼ 1:0, MeOH). m=z (ESI):
D
410 (Mþ+Na). (Found: C, 68.31; H, 6.84; N, 3.70;
1
C22H26FNO4 requires C, 68.20; H, 6.76; N, 3.62.) H
NMR (C6D6, 600 MHz):
d
1.40 (dd, J ¼ 6.1 Hz,
J ¼ 3.1 Hz, 3H, CH3–CH), 2.74 (s, 3H, N–CH3), 3.01–
3.04 (m, 1H, CH2), 3.47–3.51 (m, 1H, CH2), 3.52 (s, 3H,
CH3–O), 3.68–3.76 (m, 2H, CH–Ph+CH–CH3), 4.33 (d,
J ¼ 11.1 Hz, 1H, CH2–O), 4.45 (d, J ¼ 11.1 Hz, 1H,

250
L. Bernardi et al. / Tetrahedron: Asymmetry 15 (2004) 245–250
5. (a) Brown, J. M.; Comes-Franchini, M.; Farrington, E.
Chem. Commun. 1998, 277–278; (b) Gree, D.; Vallerie, L.;
Gree, R. J. Org. Chem. 2001, 66, 2374–2387; (c) Greedy,
B.; Paris, J.-M.; Vidal, T.; Gouverneur, V. Angew. Chem.,
Int. Ed. Engl. 2003, 42, 3291–3294.
J ¼ 23.3 Hz, CH–Me), 91.6 (d, J ¼ 188.7 Hz, CH–F),
127.0–128.6 (ArCH), 135.1 (ArCq) ppm. 19F NMR
(C6D6, 282 MHz): d )181.0–181.6 (m).
6. Despite a careful investigation, we did not find any trace
of the fourth possible exo-cycloadduct.
7. (a) Huisgen, R. In 1,3-Dipolar Cycloaddition Chemistry;
Padwa, A., Ed.; Wiley Interscience: New York, 1984; Vol.
Acknowledgements
ꢁ
1; (b) Herrera, R.; Nagarajan, N.; Morales, M.; Mendez,
ꢁ
We acknowledge the financial support by the National
Project ÔStereoselezione in Sintesi Organica. Metodo-
logie ed ApplicazioniÕ 2002–2003 and from the RTN
project ÔDesign, Analysis and Computation for Catalytic
Organic ReactionsÕ (contract HPRN-CT-2001-00172).
Dr. John M. Brown of the Dyson Perrins Laboratory
(Oxford, UK) is also acknowledged for useful discus-
sions.
ꢁ
F.; Jimenez-Vasquez, H. A.; Zepeda, L.-G.; Tamariz, J.
J. Org. Chem. 2001, 66, 1252–1263.
8. Obtained at 600 MHz using a DPFGSE-NOE sequence
with either a 25 or 75 Hz Ôr-snobÕ pulse and a mixing time
of 2 s. See: Claridge, T. D. W. In High-Resolution NMR
Techniques in Organic Chemistry; Pergamon Press: Am-
sterdam, 1999 Shaka, A. J.; Van, Q. N.; Smith, E. M. J.
Magn. Res. p 191–198.
9. Wasylingen, R. E.; Barfield, M. J. Am. Chem. Soc. 1975,
97, 4545–4552.
10. de La Hoz, A.; Diaz-Ortis, A.; Moreno, A.; Langa, F. Eur.
J. Org. Chem. 2000, 3573–3659.
References and notes
11. Ribe, S.; Wipf, P. Chem. Commun. 2001, 299–307.
12. Wilkes, J. S. Green Chemistry 2002, 4, 73–80.
13. In particular, serious problems were encountered with
RTILs under microwave irradiation because of difficult
temperature control (entry 5). The use of ionic liquids as
aids for microwave heating of non-polar solvents turned
out to slightly improve the isolated yield of 3a–c (entry 6).
See: Leadbeater, N. E.; Terenius, H. M. J. Org. Chem.
2002, 67, 3145–3148.
1. (a) Enantiocontrolled Synthesis of Fluoro-Organic Com-
pounds. Stereochemical Challenges and Biomedicinal Tar-
gets; Soloshonok, V. A., Ed.; Wiley: USA, 1999; (b)
Myers, A. G.; Barbay, J. K.; Zhang, B. J. J. Am. Chem.
Soc. 2001, 123, 7207–7219.
2. Asensio, A.; Bravo, P.; Crucianelli, M.; Farina, A.;
Fustero, S.; Soler, J. G.; Meille, S. V.; Panzeri, W.; Viani,
F.; Volonterio, A.; Zanda, M. Eur. J. Org. Chem. 2001,
1449–1458.
14. Aggarwal, V. D.; Dean, D. K.; Meren, A.; Williams, R.
J. Org. Chem. 2002, 67, 510–514.
3. Gothelf, K. V. In Cycloaddition Reaction in Organic
Synthesis; Kobayashi, S., Jorgensen, K. V., Eds.; Wiley-
VCH: New York, 2001, Chapter 6.
4. Green Chemistry: Frontiers in Benign Chemical Synthesis
and Processes; Anastas, P. T., Williamson, T. C., Eds.;
Oxford University Press: Oxford, 1998.
15. Bernardi, A.; Cardani, S.; Colombo, S.; Poli, G.; Schim-
perna, G.; Scolastico, C. J. Org. Chem. 1988, 52, 888–891.
16. Ito, M.; Maeda, M.; Kibayashi, C. Tetrahedron Lett. 1992,
33, 3765–3768; Annunziata, R.; Cinquini, M.; Cozzi, F.;
Giaroni, P.; Raimondi, L. Tetrahedron Lett. 1991, 32,
1659–1662.