[3 + 2] Cycloadditions between α,β-Unsaturated Esters or Nitroalkenes and Camphor-Derived Oxazoline N-Oxides: Experimental and Theoretical Study

Article abstract of DOI:10.1021/jo980326e

Oxazoline N-oxide 2 in the presence of α,β-unsaturated esters or nitroalkenes afforded the corresponding adducts with excellent regio- and stereoselectivities depending of the substitution pattern on the dipolarophile. Especially, inversion of regioselectivity was observed by switching from a 1,2-disubstituted to a 1,1-disubstituted α,β-unsaturated ester. This was not observed when nitroalkenes are used as dipolarophiles. An attempt to rationalize the observed selectivities by semiempirical calculations at RHF AM1 level is described. The cycloadduct 5e has been used as the starting material in a total synthesis of the pheromone (-)-frontalin 12.

Full text of DOI:10.1021/jo980326e

J . Org. Chem. 1998, 63, 5123-5128
5123
[3 + 2] Cycloa d d ition s betw een r,â-Un sa tu r a ted Ester s or
Nitr oa lk en es a n d Ca m p h or -Der ived Oxa zolin e N-Oxid es:
Exp er im en ta l a n d Th eor etica l Stu d y
Cyrille Kouklovsky,† Olivier Dirat,† Thierry Berranger,† Yves Langlois,*,†
Marie Elise Tran-Huu-Dau,‡ and Claude Riche‡
Laboratoire de Synthe`se des Substances Naturelles, Associe´ au CNRS, ICMO, Baˆtiment 410,
Universite´ de Paris-Sud, 91440, Orsay, France and Institut de Chimie des Substances Naturelles, CNRS,
91190, Gif sur Yvette, France
Received February 20, 1998
Oxazoline N-oxide 2 in the presence of R,â-unsaturated esters or nitroalkenes afforded the
corresponding adducts with excellent regio- and stereoselectivities depending of the substitution
pattern on the dipolarophile. Especially, inversion of regioselectivity was observed by switching
from a 1,2-disubstituted to a 1,1-disubstituted R,â-unsaturated ester. This was not observed when
nitroalkenes are used as dipolarophiles. An attempt to rationalize the observed selectivities by
semiempirical calculations at RHF AM1 level is described. The cycloadduct 5e has been used as
the starting material in a total synthesis of the pheromone (-)-frontalin 12.
In tr od u ction
Asymmetric [3 + 2] cycloadditions of nitrile oxides and
analogues thereof with alkenes have been developed in
the recent years as a convenient alternative to the aldol
addition reaction.1 Recently, we have introduced the use
of enantiomerically pure, camphor-derived oxazoline
N-oxides as dipoles for the asymmetric [3 + 2] cycload-
dition with a great variety of reactive alkenes2 (Figure
1). This methodology, combined with a new method for
the hydrolysis of the chiral auxiliary, has led to the
preparation of aldol-type adducts,2 and has been applied
to the synthesis of â-lactones3 and (+)-carbovir.4 In an
extensive study concerning the scope and applications of
this cycloaddition reaction, we focused on the use of R,â-
unsaturated esters and nitroalkenes since these very
reactive dipolarophiles give rise to useful synthetic
intermediates after functional group manipulation and
hydrolysis. In the present communication, we wish to
report our observations of the regio- and stereoselectivity
obtained with diversely substituted R,â-unsaturated
esters and nitroalkenes as well as theoretical calcula-
tions.
F igu r e 1.
carbonate, at 45 °C in either dichloromethane or toluene,
to form the oxazoline N-oxide 2. After 4 h, the dipolaro-
phile 3a -g was added, and the mixture was stirred at
the appropriate temperature and time (see Table 1).
After usual workup and purification, the cycloadducts
4a -g and/or 5a -g were isolated. The experimental
results are summarized in Table 1.
The reaction proceeds in moderate to good yield (except
for entry g) and proved to be highly regio- and stereose-
lective, giving in most cases a single product, except for
entries a, d, and g. NMR analysis of samples taken
during the course of the reaction showed no modification
of product distribution in the crude mixture. No byprod-
uct arising from side-reaction (apart from decomposition
of the dipole 2), e.g. anionic polymerization of the
dipolarophile, radical coupling, or rearrangement, could
be detected. The excess dipolarophile (especially nitro
compounds) may be recovered if necessary in nearly
quantitative yield. Therefore, the reaction products can
be considered as kinetic products issued from a cycload-
dition transition state, in which the dipolarophile ap-
proaches from the less hindered bottom face of the dipole.
To rationalize the experimental observations, the
calculation of the frontier molecular orbital energies and
Resu lts a n d Discu ssion
In a general procedure, 3-(hydroxyamino)isoborneol
hydrochloride (1)5 was treated with trimethyl orthofor-
mate6 (4 equiv) in the presence of 1 equiv of calcium
* Corresponding author. Fax: (33) 01 69 15 46 79; email: langlois@
icmo.u-psud.fr.
† Universite´ de Paris-Sud.
‡ Institut de Chimie des Substances Naturelles.
(1) (a) Kozikowski, A. P. Acc. Chem. Res. 1984, 17, 410. (b) Curran,
D. P. Advances in Cycloaddition; J AI: Greenwich, 1988; Vol. 1. (c)
Torssell, K. B. G. Nitrile Oxides, Nitrones and Nitronates in Organic
Synthesis; VCH Publishers: Weinheim, 1988. (d) Carruthers, W.
Cycloaddition Reactions in Organic Synthesis; Pergamon Press: Ox-
ford, 1990.
(2) Berranger T.; Langlois, Y. J . Org. Chem. 1995, 60, 1720.
(3) Dirat, O.; Berranger, T.; Langlois, Y. Synlett 1995, 935.
(4) Berranger, T.; Langlois, Y. Tetrahedron Lett. 1995, 36, 5523.
(5) Baldwin, S. W.; McFayden, R. B.; Aube´, J .; Wilson, J . D.
Tetrahedron Lett. 1991, 32, 4431.
(6) (a) Ashburn, S. P.; Coates, R. M. J . Org. Chem. 1984, 49, 3127.
(b) Ashburn, S. P.; Coates, R. M. J . Org. Chem. 1985, 50, 3076.
S0022-3263(98)00326-0 CCC: $15.00 © 1998 American Chemical Society
Published on Web 07/01/1998

5124 J . Org. Chem., Vol. 63, No. 15, 1998
Kouklovsky et al.
Sch em e 1
Sch em e 2
Ta ble 1
Ta ble 2. F MO En er gy Levels a n d Atom Coefficien ts in
Dip ole 2
yield,
%
endo/
entry EWG
R1 R2
conditions
4/5a
exoa,b
Ea
O1
N2
C3
Ob
a
b
c
d
e
f
CO2Me
NO2
H
H
H
Me
H
CH2Cl2, reflux, 3 h
53
70
80
58
65
75
16
85/15 75/25
>95/5 >95/5
>95/5 >95/5
>95/5 90/10
>5/95 >95/5
>95/5 >95/5
50/50 >95/5
HOMO
LUMO
-8.58
+0.56
+0.64
+0.37
-0.31
-0.59
-0.57
+0.65
+0.36
-0.18
Bun CH2Cl2, reflux, 2 h
CO2But
NO2
Pr
H
H
toluene, 80 °C, 16 h
CH2Cl2, reflux, 4 h
toluene, 60 °C, 2 h
a
The E(HOMO) and E(LUMO) are in eV. b The O atom adjacent
CO2Me Me
NO2
CO2But Me Bun toluene, 80 °C, 16 h
to C3.
Me Bun toluene, 80 °C, 16 h
g
of the 2,4,4-trimethyloxazoline N-oxide, respectively: this
successfully explains the good reactivity of 2. On the
basis of these results, one might anticipate that oxazoline
N-oxides such as 2 would show better reactivities and
selectivities than the corresponding nitrones in the [3 +
2] cycloaddition reactions.
a
b
Determined by NMR analysis of the crude product. For the
major regiosomer in entries a-f; for both regioisomers in entry g.
Frontier orbital energies and corresponding coefficients
of N-oxide 2 and dipolarophiles 3a -g are reported,
respectively, in Tables 2 and 3 (Figure 3, Tables 2 and
3). The relative disposition of the frontier orbitals for
dipole 2 (HOMO ) -8.58 eV; LUMO ) +0.56 eV) and
dipolarophiles 3a -g (HOMO about -10, -11 eV; LUMO
about 0, -1 eV) suggests that, on the basis of the
narrower HOMO-LUMO gap, the reactions are predomi-
nantly HOMO dipole-controlled reactions. However, it
can be noted that the difference between O and C
coefficients is more important in the LUMO (+0.37,
+0.65, respectively) than in the HOMO (+0.64, -0.57,
respectively) of the dipole 2; consequently, in some cases,
the LUMO dipole-HOMO dipolarophile interaction should
not be neglected.
Experimentally, nitroalkenes show greater reactivity
than R,â-unsaturated esters, as judged by reaction tem-
peratures and times, and reactivity decreases when the
substitution on double bond increases. The high electron-
withdrawing character of the nitro group allows cycload-
dition of 2-nitro-2-heptene 3f to proceed with a good yield
under relatively mild conditions (Table 1, entry f). In
contrast, cycloaddition with the 1,1,2-trisubstituted un-
saturated ester 3g gives a disappointing 16% yield under
the same conditions. FMO analysis (Table 3) shows that
the substitution of the CO2Me group in the dipolarophiles
3c, 3e and 3g by a NO2 group as in 3b, 3d , and 3f,
respectively, lowers to a much greater extent the LUMO
energy level than the HOMO one (HOMO/LUMO )
-10.51/-0.01, -10.47/+0.04 and -9.93/+0.17 eV instead
of -11.27/-0.87, -11.21/-0.79 and -10.68/-0.73 eV).
This fact explains the great reactivity of nitroalkenes
F igu r e 2.
coefficients of the reactants were performed at the RHF/
AM1 level7 using MOPAC program (Version 5.0).8
Coates and co-workers6 have rationalized the enhanced
reactivity of the 2,4,4-trimethyloxazoline N-oxide com-
pared to the related 2,5,5-trimethylpyrroline N-oxide in
terms of frontier orbital theory9 (Figure 2). To compare
with our results, we have performed FMO calculations
of these two compounds at the AM1 level. The presence
of the endocyclic oxygen atom in the oxazoline N-oxide
reduces the frontier orbital separation, compared to the
pyrrolidine N-oxide: it shifts the HOMO to higher energy
(-8.51 to -8.41 eV) and the LUMO to lower energy
(+0.69 to +0.59 eV). This allows for stronger interactions
between the dipole HOMO and the dipolarophile LUMO
as well as the dipole LUMO and the dipolarophile
HOMO. Calculations applied on dipole 2 show that the
HOMO (-8.58 eV) and LUMO (+0.56 eV) of 2 are slightly
lower than the HOMO (-8.41 eV) and LUMO (+0.59 eV)
(7) Dewar, M. J . S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J . J . P.
J . Am. Chem. Soc. 1985, 107, 3902.
(8) Stewart, J . J . P. MOPAC5.0, QCPE program number 455 and
J . Comput. Chem. 1989, 101, 221.
(9) Coates and co-workers have used molecular orbitals calculated
at CNDO and EH levels by Houk and co-workers: (a) Houk, K. N.
Acc. Chem. Res. 1975, 8, 361. (b) Houk, K. N.; Sims, J .; Duke, R. E.;
Strozier, R. W.; George, J . K. J . Am. Chem. Soc. 1973, 95, 7287. (c)
Houk, K. N.; Sims, J .; Watts, C. R.; Luskus, L. J . Am. Chem. Soc. 1973,
95, 7301.

{3 + 2} Cycloadditions
J . Org. Chem., Vol. 63, No. 15, 1998 5125
Ta ble 3. F MO En er gy Levels a n d Atom Coefficien ts in Dip ola r op h iles 3a -g
entry
a
EWG
R1
H
R2
H
Ea
C1
C2
C or N
Ob
Oc
CO2Me
HOMO
LUMO
HOMO
LUMO
HOMO
LUMO
HOMO
LUMO
HOMO
LUMO
HOMO
LUMO
HOMO
LUMO
-11.08
+0.001
-11.27
-0.87
+0.68
-0.48
+0.69
-0.43
+0.67
-0.48
+0.62
-0.43
+0.61
-0.48
+0.63
-0.43
+0.62
-0.49
+0.65
+0.66
+0.55
+0.63
+0.57
+0.66
+0.63
+0.62
+0.64
+0.66
+0.55
+0.62
+0.57
+0.66
-0.05
-0.43
-0.03
-0.40
-0.02
-0.40
-0.02
-0.43
-0.02
-0.42
-0.01
-0.41
0.
-0.30
+0.34
-0.15
+0.33
-0.27
+0.33
-0.15
+0.34
-0.24
+0.34
-0.15
+0.33
-0.24
+0.31
-0.10
+0.17
-0.19
+0.33
-0.10
+0.16
-0.13
+0.34
-0.07
+0.16
-0.14
+0.33
-0.07
+0.15
b
c
d
e
f
NO2
H
Me
Me
H
CO2Me
NO2
H
-10.51
-0.01
Me
Me
Me
Me
-11.21
-0.79
CO2Me
NO2
H
-10.47
+0.04
Et
Et
-10.68
-0.73
g
CO2But
-9.93
+0.17
-0.39
a
b
The E(HOMO) and E(LUMO) are in eV. The O atom in cis position compared to the C-C double bond. c The O atom in trans position
compared to the C-C double bond.
electron-withdrawing nitro substituent in the dipolaro-
philes considerably lowers the energy level of the LUMO
orbital (about -0.7, -0.9 eV), favoring HOMO dipole-
controlled reactions. Therefore, as mentioned above, in
contrast with 3b and 3f, the larger coefficient is at
unsubstituted carbon C2 in the HOMO of 3d , but the
regioselectivity is not affected, in contrast with the esters,
F igu r e 3.
because only HOMO dipole-controlled reactions should
occur and lead to formation of regioisomers 4. The
which should react, predominantly with their LUMO.
Moreover, the addition of an electron-releasing alkyl
results of FMO calculations support the experimental
substituent in the dipolarophile shifts its HOMO and
data.
LUMO to higher energy. Independently of steric effects,
In both reactions, endo and facial selectivities were
complete, except in the reaction with 2-nitropropene
(Table 1, entry d) where a small amount of the exo isomer
is also obtained. This minor compound is easily removed
by chromatography. FMO analysis (Table 3) reveals that
the coefficients at the central nitrogen and at the en-
docyclic oxygen in the dipole HOMO (-0.3 and +0.4,
respectively) and at the carbon (or nitrogen) and at the
oxygen of the COOR (or NO2) group in the dipolarophile
LUMO (about -0.4 and +0.3, respectively) have the same
sign. This suggests that this attractive interaction can
lead to stabilizing secondary orbital interaction and
explains the preferred endo selectivity in those cycload-
ditions.
These calculations point out that the origin of the
regioselectivity is certainly the stabilization in the transi-
tion state due to the favorable overlap of the π orbitals,
that the preferred endo stereoselectivity is due to the
stabilizing secondary interactions between the pz of the
central nitrogen and endocyclic oxygen in the dipole and
the pz of the CO2 (or NO2) group of the dipolarophile,
and that the reversal of regioselectivity with methyl
methacrylate might be the consequence of the dipole
LUMO-dipolarophile HOMO interaction which becomes
more important with the addition of an alkyl group.
Moreover, the regioselectivity with methyl methacrylate
could also be explained by steric effects with transition
structure model considerations: the interaction of the exo
methyl substituent on the alkene with the hydrogen
borne by the C3 of the dipole probably disfavors the endo
4-substituted cycloadduct by comparison with the endo
5-substituted one; however, this interaction is not well
evaluated in AM1 calculations. It is possible that the
induced torsion in the developing five-membered ring in
the transition state destabilizes the formation of the
cycloadduct 4e, leading to the formation of 5e, arising
from a less hindered transition state. The reversal of
regioselectivity is not observed in the cycloaddition with
2-nitropropene 3d (in which the larger coefficient in the
this is consistent with the observed experimental de-
crease of reactivity when substitution on the double bond
increases.
Some striking differences may also be observed in the
selectivity of the cycloadditions with esters. Although
reaction with methyl acrylate (Table 1, entry a) gives
moderate regio- and stereoselectivities, reaction with
disubstituted alkenes such as tert-butyl 2-pentenoate 3c
and methyl methacrylate 3e occurred with complete
regioselectivity, but to give the opposite regioisomers
(Table 1, entries c and e, Scheme 2). FMO analysis shows
that both HOMO or LUMO dipole-controlled reactions
with the dipolarophiles 3a , 3b, 3c, 3f, 3g (larger coef-
ficient at carbon C2 in the LUMO and larger coefficient
at carbon C1 in the HOMO) should lead to 4-substituted
cycloadducts (Table 2). For 3d and 3e (larger coefficient
at unsubstituted carbon C2 in the LUMO and in the
HOMO), the control of regioselectivity by the dipole
HOMO (larger coefficient on the O heteroatom) and the
dipolarophile LUMO (larger coefficient is at unsubsti-
tuted carbon C2) will lead to 4-substituted cycloadduct,
but the control of regioselectivity by the dipole LUMO
(larger coefficient on the C atom) and the dipolarophile
HOMO (larger coefficient is at unsubstituted C2 carbon)
will lead to 5-substituted cycloadduct. As mentioned
above, on one hand, the addition of an electron-releasing
alkyl substituent in the dipolarophile shifts its HOMO
and LUMO to higher energy and on the other hand, in
the HOMO of 3e in contrast with 3c, the larger coefficient
is at the C2 unsubstituted carbon. Reversal of regiose-
lectivity with methyl methacrylate (Table 1, entry e)
could be the consequence of the dipole LUMO-dipolaro-
phile HOMO interaction which becomes more important
with the addition of an alkyl group (Table 3).
In contrast with esters, nitroalkenes react with the
dipole 2 (entries b, d, and f) with consistent regioselectivity
to give the cycloadducts of type 4, whatever the substitu-
tion pattern on the double bond. This remarkable selec-
tivity may be explained by the fact that the strongly

5126 J . Org. Chem., Vol. 63, No. 15, 1998
Kouklovsky et al.
Sch em e 3
HOMO is at the C2 unsubstituted carbon as in 3e)
because the dipole HOMO-dipolarophile LUMO interac-
tion remains predominant, this being due to the high
electron-withdrawing character of the nitro group. More-
over, the presence of a small amount of the exo isomer
supports the above mentioned steric effect of the exo
substituent on the nitroalkene with the hydrogen borne
by the C3 carbon of the dipole (Figure 4).
yield (overall yield for two steps), together with the
recovered ketol 10 (80% recovery). The lower yield of 9
is due to partial elimination of the tertiary alcohol; this
side-reaction becomes more important when hydrolysis
reaction time is extended. This aldehyde has already
been prepared (as its antipode) by Monneret and co-
workers and used as an intermediate in the total syn-
thesis of frontalin.11 Therefore the remaining steps have
been carried out as described by these authors: treat-
ment of 9 with (triphenylphosphoranylidene)acetone in
refuxing acetonitrile gave the unsaturated ketone 11 in
71% yield (with 15% recovered starting material). Hy-
drogenation reaction (1 atm, Pd/C) gave in quantitative
yield (-)-frontalin 12, whose spectral data were in
complete agreement to those reported in the literature.
This short and stereoselective synthesis of frontalin 12
has shown the potential of cycloaddition reactions with
oxazoline N-oxide in the preparation of small, densely
functionalized chiral molecules such as the aldehyde 9.
F igu r e 4.
Con clu sion
Hyd r olysis of 5e. Tota l Syn th esis of F r on ta lin .
Despite being obtained with an “reversed” regioselectiv-
ity, the cycloadduct 5e has an interesting potential in
synthesis since the introduction of a chiral tertiary
alcohol with two different neighboring functional groups
could be useful for the assembly of polyoxygenated
molecules. We have therefore undertaken the hydrolysis
of the chiral auxiliary in 5e and used the corresponding
product in total synthesis of the natural pheromone (-)-
frontalin 1210 (Scheme 3).
Previous studies showed it is possible to reduce the
ester function in cycloadducts leaving the N-O bond
unaffected; therefore, reduction of 5e (LiAlH4, Et2O)
followed by protection of the primary alcohol 6 as its
benzyl ether (BnBr, nBu4I, THF/DMF) gave compound
7 in an excellent yield and in a complete chemoselective
fashion. This product was subjected to the oxidative
hydrolysis procedure developed in our laboratory:2 treat-
ment with excess m-chloroperoxybenzoic acid to give the
intermediate nitrone 8, followed by acidic hydrolysis with
2 N HCl in THF afforded the aldehyde 9 in 71% isolated
The experimental results obtained in this study have
revealed the great versatility of the asymmetric [3 + 2]
cycloadditions of oxazoline N-oxides to give good yields
and excellent regio- and stereoselectivities with disub-
stituted R,â-unsaturated esters or with diversely substi-
tuted nitroalkenes. In comparison, nitrile oxides are
poorly regioselective with these dipolarophiles.1 The use
of cycloadduct 5e in the synthesis of the natural product
fontalin 12 has illustrated the application of these
cycloadditions in synthesis. Furthermore, it has given
additional confirmation of the structure and purity of the
starting cycloadduct. Other synthetic applications of
asymmetric [3 + 2] cycloadditions with oxazoline N-
oxides are currently under investigation in our labora-
tory.
Exp er im en ta l Section
Gen er a lities. 1H and 13C NMR spectra were recorded at
200 or 250 MHz and 50 or 62.5 MHz, respectively. Optical
rotations were recorded at 25 °C. Chromatographic purifica-
tions were performed on 230-400 mesh silica gel (Merck 9385)
using the indicated solvent system. Dichloromethane, aceto-
nitrile, DMF, and trimethyl orthoacetate were distilled from
(10) Isolation: Kinzer, G.W., Fentimar, A. F., J r.; Page, T. F., J r.;
Foltz, R. L.; Vite´, J . P.; Pitman, G. B. Nature 1969, 221, 477.
Syntheses: Nishimura, Y.; Mori, K. Eur. J . Org. Chem. 1998, 233.
Chen, S. T.; Fang, J . M. J . Org. Chem. 1997, 62, 4349; Tanigushi, T.;
Nakamura, K.; Ogasawara, K. Synthesis 1997, 509 and references
therein.
(11) Trinh, M.-C.; Florent, J . C.; Monneret, C. Tetrahedron 1988,
44, 6633.

{3 + 2} Cycloadditions
J . Org. Chem., Vol. 63, No. 15, 1998 5127
calcium hydride. Toluene, diethyl ether, and THF were
distilled from sodium metal/benzophenone ketyl. Chloroform
used for optical measurements was filtered through basic
alumina before use. tert-Butyl 2-hexenoate and tert-butyl
2-methyl-2-pentenoate were prepared by esterification of the
commercially available acids. 2-Nitropropene was prepared
by dehydration (phthalic anhydride, 150 °C) of commercial
2-nitro-1-propanol. 1-Nitro-1-hexene and 2-nitro-2-heptene
were prepared by Henry condensation (Al2O3) of nitromethane
and nitroethane with valeraldehyde, followed by dehydration
(PPh3, CCl4). Methyl acrylate and methyl methacrylate were
obtained from commercial sources. All nonaqueous reactions
were performed under an argon atmosphere using oven-dried
glassware.
Gen er a l Exp er im en ta l P r oced u r e for th e Cycloa d d i-
tion Rea ction s. A suspension of 3-(hydroxylamino)isoborneol
hydrochloride 1 (1 equiv) and powdered calcium carbonate (l
equiv) in either dichloromethane or toluene (5 mL/mmol) was
heated at 45 °C with stirring, and trimethyl orthoacetate (4
equiv) was added with syringe. The mixture was stirred at
45 °C for 4 h, and then the dipolarophile was added (4 equiv).
The mixture was vigorously stirred at the appropriate tem-
perature and time (see Table 1). The completion of the
reaction was monitored by TLC. After cooling to room tem-
perature, the mixture was filtered through a pad of Celite and
the filtrate concentrated in vacuo. The residue was purified
by flash chromatography on silica gel (10% ethyl acetate/
heptane) to give the pure cycloadduct.
NMR (50 MHz, CDCl3): δ (ppm) 167.7, 98 1, 89.1, 80.9, 77.5,
75.2, 58.4, 48.6, 47.7, 45.3, 33.7, 30.8, 27.5, 25.9, 21.6, 18.6,
18.4, 13.6, 10.1; mass (IC, NH3): 366 (MH+); [R]20 ) -129 (c
D
) 2.5, CHCl3). Anal. Calcd for C21H35NO4, MW: 365.25659;
Calcd: C: 68.99%; H: 9.66%; N: 3.83%. Found: C: 68.66%;
H: 9.66%; N: 3.91%.
(3S,3a S,4a S,5R,8S,8a R)-3-Nitr o-3,5,10,10-tetr a m eth yl-
5,8-m et h a n ooct a h yd r o-2H -isoxa zolo[3,2-b]b en zoxa zole
(4d ). Reaction was performed in refluxing dichloromethane
for 4 h after the addition of 2-nitropropene. The crude product
was purified by chromatography. First to elute was 4d (Rf:
0.35), followed by a small amount of exo-4d (Rf: 0.29). Yield:
58%. 1H NMR (250 MHz, CDCl3): δ (ppm) 4.94 (lH, s), 4.73
(lH, d), 3.94 (lH, d, J ) 7.5 Hz), 3.79 (lH, d, J ) l0 Hz), 3.42
(lH, d, J ) 7.5 Hz), 2.09 (lH, d, J ) 4.5 Hz), 1.78 (3H, s), 1.80
and 1.65 (4H, 2m), 0.87, 0.83, 0.77; 13C NMR (62.5 MHz,
CDCl3): δ (ppm) 102.8, 94.9, 91.4, 75.6, 72.1, 49.2, 48.9, 45.8,
31.4, 25.5, 22.8, 22.6, 18.9, 10.8; mass (IC NH3): m/z: 301
(MNH4+), 283 (MH+); [R]20D ) -202.1 (c ) 1.89, CHCl3). Anal.
Calcd for C14H22N2O4, MW: 282.34244; Calcd: C: 59.56%; H:
7.85%; N: 9.92%. Found: C: 59.71%; H: 8.11%; N: 9.42%.
(2S,3aS,4aS,5R,8S,8aR)-2,5,10,10-Tetramethyl-5,8-methano-
octa h yd r o-2H-isoxa zolo[3,2-b]ben zoxa zole-2-ca r boxylic
Acid Meth yl Ester (5e). Reaction was performed in toluene
at 80 °C for 2 h after adding methyl methacrylate. Yield: 65%.
The product 5e was obtained as a single stereomer. 1H NMR
(250 MHz, CDCl3): δ (ppm) 5.48 (lH, dd, J ) 4.2 and 6.6 Hz,),
3.92 (lH, d, J ) 7.5 Hz), 3.80 (3H, s), 3.65 (lH, d, J ) 7.5 Hz),
2.85 (lH, dd, J gem ) 13.5 Hz and J trans ) 4.2 Hz), 2.25 (lH, dd,
J gem ) 13.5 Hz and J cis ) 6-6 Hz), 2.05 (lH, d, J ) 4.5 Hz),
1.70 and 1.40 (4H, 2m), 1.48 (3H, s), 0.95, 0.94, and 0.79 (9H,
3s); 13C NMR (50 MHz, CDCl3): δ (ppm) 174.0, 98.1, 88.4, 83.9,
75.9, 52.5, 49.0, 48.1, 45.6, 44.3, 31.5, 25.5, 24.5, 22.2,19.0, 10.7;
(3S,3a S,4a S,5R,8S,8a R)-5,10,10-Tr im et h yl-5,8-m et h a -
n oocta h yd r o-2H-isoxa zolo[3,2-b]ben zoxa zole-3-ca r boxyl-
ic Acid Meth yl Ester (4a ). Reaction was performed in
refluxing dichloromethane for 4 h after adding methyl acrylate.
1H NMR analysis of the crude product showed the presence of
three cycloadducts 4a , 5a , and exo-4a in the ratio (60:20:15)
according to the integration of the C3a proton (4a : 5.51 ppm;
5a : 5.37 ppm; exo-4a : 5.29 ppm). The combined yield is 53%.
Careful chromatographic purification allows isolation of some
pure 4a (Rf: 0.20) together with mixed fractions with 5a and
exo-4a (Rf: 0.18). 1H NMR (250 MHz, CDCl3): δ (ppm) 5.51
(lH, d, J ) 7.5 Hz), 4.16 (lH, dd, J ) 8 and 10 Hz), 3.95 (lH,
dd, J ) 1 and 10 Hz), 3.88 (lH, d, J ) 7 Hz), 3.72 (3H, s), 3.42
(lH, ddd, J ) 1, 7.5 and 10 Hz), 3.22 (lH, d, J ) 7 Hz), 2.09
(lH, d, J ) 4,5 Hz), 1.68 and 1.35 (4H, 2m), 0.91 and 0.73 (9H,
2s); 13C NMR (62.5 MHz, CDCl3): δ (ppm) 169.3, 98.1, 90.2,
75.5, 66.9, 53.2, 52.4, 49.2, 48.7, 46.0, 31.5, 25.6, 22.3 19.0,
mass (IC NH3): m/z 296 (MH+); [R]20 ) -32.2 (c ) 1.42,
D
CHCl3). Anal. Calcd for C16H25NO4, MW: 295.3482; Calcd:
C: 65.06%; H: 8.53%; N: 4.74%. Found: C: 64.77%; H:
8.71%; N: 4.77%.
(2S,3S,3a S,4a S,5R,8S,8a R)-2-Bu tyl-3-n itr o-3,5,10,10-tr i-
m et h yl-5,8-m et h a n ooct a h yd r o-2H -isoxa zolo[3,2-b]b en -
zoxa zole (4f). Reaction was performed at 80 °C in toluene
for 16 h after adding 2-nitro-2-heptene. Yield: 75%. The
product 4f was obtained as a single stereomer. 1H NMR (250
MHz, CDCl3): δ (ppm) 4.96 (lH, s, C3a-H), 4.77 (lH, t, J ) 6
Hz, C2-H), 4.00 (lH, d, J ) 7.5 Hz), 3.45 (lH, d, J ) 7.5 Hz),
2.10 (lH, d, J ) 4.5 Hz), 1.63 (3H, s), 1.80-1.30 (10H, m), 0.90
(3H, t, J ) 7 Hz), 0.90, 0.85, and 0.80 (9H, 3s); 13C NMR (50
MHz, CDCl3): δ (ppm) 103.9, 95.9, 91.2, 79.6, 76.6, 49.3, 48.7,
45.6, 31.3, 28.2, 26.5, 25.4, 22.6, 22.1, 18.7, 17.8, 13.8, 10.7;
10.9; mass (IC NH3): 282 (MH+); [R]20 ) -121.6 (c ) 1.53,
D
CHCl3). Anal. Calcd for C15H23NO4, MW: 281.35486; Cald:
C: 64.04%; H: 8.45%; N: 4.98%. Found: C: 64.21%; H:
8.45%; N: 4.82%.
mass (IC NH3): m/z: 339 (MH+); [R]20 ) -172 (c ) 1.1,
D
(2S,3S,3a S,4a S,5R,8S,8a R)-2-Bu t yl-3-n it r o-5,10,10-t r i-
m et h yl-5,8-m et h a n ooct a h yd r o-2H -isoxa zolo[3,2-b]b en -
zoxa zole (4b). Reaction was performed in refluxing dichlo-
romethane for 2 h after adding l-nitro-1-hexene; yield: 70%.
The product 4b was obtained as a single stereomer. This
product was not very stable on storage, and therefore combus-
tion analysis could not be obtained. 1H NMR (250 MHz,
CDCl3): δ (ppm) 5.55 (1H, d, J ) 7 Hz), 4.80 (lH, dd, J ) 7
and 8 Hz), 4.65 (lH, dt, J ) 6 and 8 Hz), 4.05 (lH, d, J ) 7.5
Hz, 3.45 (lH, d, J ) 7.5 Hz), 2.10 (lH, d, J ) 4.5 Hz), 1.80-
1.30 (10H, m), 0.90 (3H, t, J ) 7 Hz), 0.95, 0.88 and 0.80 (9H,
3s); 13C NMR (50 MHz, CDCl3): δ (ppm) 96.4, 92.5, 91.4, 78.0,
75.7, 49.1, 48.7, 45.8, 31.3, 30.6, 27.3, 25.2, 22.4, 22.0, 18.7,
CHCl3). Anal. Calcd for C18H30N2O4, MW: 338.22054; Cal-
cd: C: 63.88%; H: 8.93%; N: 8.28%. Found: C: 64.23%; H:
9.24%; N: 8.41%.
(2S,3S,3aS,4aS,5R,8S,8aR)-2-Eth yl-3,5,10,10-tetr am eth yl-
5,8-m eth a n oocta h yd r o-2H-isoxa zolo[3,2-b]ben zoxa zole-
3-car boxylic Acid 1,1-Dim eth yleth yl Ester (4g) an d (2R,3S,
3aS,4aS,5R,8S,8aR)-3-Ethyl-2,5,10,10-tetramethyl-5,8-methano-
oct a h yd r o-2H -isoxa zolo[3,2b]b en zoxa zole-2-ca r b oxylic
Acid l,l-Dim eth yleth yl Ester (5g). Reaction was performed
in toluene at 80 °C. The reaction was stopped 16 h after
adding (E)-tert-butyl 2-methyl-2-pentenoate (no further evolu-
tion and degradation products were observed). Purification
of the crude product gave first 4g (yield: 8%; Rf: 0.37), followed
by 5g (yield: 8%; Rf: 0.28).
13.7, 10.6; mass (IC NH3): m/z: 325 (MH+); [R]20 ) 161 (c )
D
1.47, CHCl3); high-resolution mass spectrum, calcd for
4g: 1H NMR (250 MHz, CDCl3): δ (ppm) 4.84 (lH, s), 4.33
(lH, dd, J ) 7 Hz), 3.87 (lH, d, J ) 7.5 Hz), 3.35 (lH, d, J ) 7.5
Hz), 2.04 (lH, d, J ) 4.5 Hz), 1.67 and 1.49-1.35 (6H, m), 1.42
(9H, s, But), 1.15 (3H, s), 0.92 (3H, t, J ) 7 Hz), 0.84 and 0.76
(9H, 2s); 13C NMR (62.5 MHz, CDCl3): δ (ppm) 170.8, 107.3,
89.9, 81.7, 81.4, 76.9, 59.4, 49.5, 48.6, 45.7, 31.6, 28.1, 25.8,
C
17H28N2O4; calcd: 324.20489; found: 324.20491.
(2S,3S,3a S,4a S,5R,8S,8a R)-5,10,10-Tr im eth yl-2-p r op yl-
5,8-m eth a n oocta h yd r o-2H-isoxa zolo[3,2-b]ben zoxa zole-
3-ca r boxylic Acid 1,1-Dim eth yleth yl Ester (4c). Reaction
was performed in toluene at 80 °C for 18 h after adding tert-
butyl 2-hexenoate. Yield: 80%. The product 4c was obtained
as a single stereomer. 1H NMR (250 MHz, CDCl3): δ (ppm)
5.37 (lH, d, J ) 7.5 Hz), 4.35 (lH, dt, J ) 10.5 and 6 Hz), 3.72
(lH, d, J ) 7.5 Hz), 3.25 (lH, d, J ) 7.5 Hz), 2.95 (lH, dd, J )
10.5 and 7.5 Hz), 2.02 (lH, d, J ) 4.5 Hz), 1.60-1.10 (8H, m),
1.42 (9H, s), 0.95 (3H, t), 0.90 and 0.71 (6H and 3H, 2s); l3C
22.3, 20.7, 19.0, 11.1, 10.9; mass (IC NH3): 366 (MH+); [R]20
D
) -125.3 (c ) 1, CHCl3); high-resolution mass spectrum, calcd
for C21H35NO4: calcd: 365.2566; found: 365.2563.
5g: 1H NMR (250 MHz, CDCl3): δ (ppm) 4.91 (lH, d, J )
7.5 Hz), 4.01 (lH, d, J ) 7.5 Hz), 3.66 (lH, d, J ) 7.5 Hz), 2.48
(lH, dt), 2.01 (lH, d, J ) 4.5 Hz), 1.63-1.41 (6H, m), 1.43 (9H,

5128 J . Org. Chem., Vol. 63, No. 15, 1998
Kouklovsky et al.
s), 1.28 (3H, s), 0.95 (3H, masked t), 0.93 and 0.87 (9H, 2s);
13C NMR (62.5 MHz, CDCl3): δ (ppm) 172.9, 103.1, 87.9, 87.4,
81.9, 75.4, 52.6, 48.8, 48.1, 46.4, 31.9, 28.0, 25.4, 22.3, 20.7,
solid (crude yield >95%), which was used immediately in the
next reaction.
The crude nitrone 8 was dissolved in THF (5 mL) with
strirring, and a 2 N hydrochloric acid solution (5 mL) was
added (exothermic reaction). The turbid solution cleared
within 10 min whereupon it was neutralized with saturated
sodium carbonate solution and extracted with ethyl acetate.
The organic phase was washed with water and brine, dried
(Na2SO4), filtered, and concentrated in vacuo. The crude
product was purified by silica gel (20% ethyl acetate/heptane);
first to elute was the recovered ketol 10 (80% recovery),
followed by the aldehyde 9 (332 mg; 71% yield) which was
obtained as a yellow oil.
19.7, 19.6, 12.3, 10.9; mass (IC NH3): 366 (MH+); [R]20
)
D
-49.6 (c ) 1, CHCl3); high-resolution mass spectrum, calcd
for C21H35NO4: calcd: 365.2566; found: 365.2573.
(2S,3a S,4a S,5R,8S,8a R)-2-(H yd r oxym et h yl)-2,5,10,10-
Tetr a m eth yl-5,8-m eth a n o-octa h yd r o-2H-isoxa zolo[3,2-b]-
ben zoxa zole (6). A solution of the cycloadduct 5e (1.06 g,
3,6 mmol) in diethyl ether (30 mL) was cooled to 0 °C, and
lithium aluminum hydride (402 mg, 10.1 mmol, 3 equiv) was
added portionwise. The gray suspension was stirred at 0 °C
for 30 min and then hydrolyzed by careful addition of a
saturated aqueous sodium sulfate solution until a white solid
separates from the colorless solution. The mixture was filtered
through a pad of Celite, rinsing with ether; the filtrate was
washed with brine, dried (Na2SO4), filtered, and concentrated
in vacuo to give the alcohol 6 as a white solid, sufficientely
pure for the next reaction (893 mg; yield: 93%). An analytical
sample was obtained by chromatography.
1H NMR (200 MHz, C6D6): δ (ppm) 9.62 (1H, t, J ) 2 Hz),
7.31 (5H, broad s), 4.18 (2H, s), 3.04 (2H, d, J ) 2 Hz), 2.70
(1H, broad s, exchangeable with D2O), 2.22 (1H, ddd, J gem
)
16 Hz and J 1-2 ) 2 Hz), 1.16 (3H, s); 13C NMR (50 MHz,
C6D6): δ (ppm) 202.1, 138.5, 128.7, 128.0, 77.5, 73.5, 71.6, 52.3,
25.1; mass (IC NH3): m/z 208 (MH+); [R]20 ) -8.5 (c ) 0.85,
D
CHCl3) (lit:11 -8 (c ) 1, CHCl3)). Anal. Calcd for C12H16O3,
MW: 267.1834; Calcd: C: 69.21%; H: 7.74%. Found: C:
68.86%; H: 7.47%.
1H NMR (250 MHz, CDCl3): δ (ppm) 5.48 (1H, dd, J ) 6.6
and 7 Hz), 4.02 (lH, d, J ) 7.5 Hz), 3.47 (2H, broad s), 3.33
(lH, d, J ) 8 Hz), 2.59 (1H, exchangeable with D2O), 2.20 (lH,
dd, J gem ) 13.5 Hz and J trans ) 7 Hz), 1.96 (lH, d, J ) 4.5 Hz),
1.89 (lH, dd, J gem ) 13.5 Hz and J cis ) 6.6 Hz), 1.60, 1.31, and
0.85 (4H, 3m), 1.14 (3H, s2), 0.90, 0.87 and 0.72 (9H, 3s); 13C
NMR (50 MHz, CDCl3): δ (ppm) 98.5, 87.8, 85.7, 78.8, 66.5,
48.7, 48.0, 46.3, 39.6, 31.7, 25.5, 23.6, 22.2, 19.4, 10.8; mass
(S)-6-Hyd r oxy-6-m eth yl-7-[(p h en ylm eth yl)oxy]-3-h ep -
ten -2-on e (11). A solution of the aldehyde 9 (270 mg, 1.3
mmol) and (triphenylphosphoranylidene)acetone (656 mg; 2.08
mmol, 1.6 equiv) in acetonitrile (5 mL) was stirred at reflux
for 3 h and then cooled and concentrated in vacuo. The residue
was purified by chromatography (35% ethyl acetate/heptane);
first to elute was the recovered starting material (40 mg, 15%
recovery), followed by the ketone 11 which was isolated as a
pale yellow oil (227 mg; yield 71%).
(IC NH3): m/z 268 (MH+); [R]20 ) -84.3 (c ) 1.57, CHCl3).
D
Anal. Calcd for C15H25NO3, MW: 267.1834; Calcd: C: 67.37%;
H: 9.43%; N: 5.24%. Found: C: 67.28%; H: 8.97%; N: 5.53%.
1H NMR (250 MHz, CDCl3): δ (ppm) 7.31 (5H, broad s), 6.80
(2S,3aS,4aS,5R,8S,8aR)-2-[[(P h en ylm eth yl)oxy]m eth yl]-
2,5,10,10-tetr am eth yl-5,8-m eth an ooctah ydr o-2H-isoxazolo-
[3,2-b]ben zoxa zole (7). A suspension of oil-free sodium
hydride (134 mg, 4.18 mmol, 1.25 equiv) in THF/DMF (15 mL/5
mL) was cooled to 0 °C with stirring, and a solution of the
alcohol 6 (893 mg, 3.34 mmol) in THF (10 mL plus 5 mL rinse)
was added dropwise. The mixture was stirred 30 min at 0
°C, and then a catalytic amount of tetra-n-butylammonium
iodide was added, followed by benzyl bromide (0.5 mL, 4.18
mmol, 1.25 equiv). The cooling bath was removed, and the
white suspension was stirred at room temperature for 4 h. The
reaction mixture was quenched by addition of saturated
ammonium chloride solution, poured onto ice, and extracted
with diethyl ether (100 mL). The organic phase was washed
successively with 5% NaHCO3 solution, water, and brine and
then dried (Na2SO4), filtered, and concentrated in vacuo. The
crude product was purified by silica gel chromatography (20%
ethyl acetate/heptane) to give the benzyl ether 7 as a nearly
colorless oil (994 mg; yield: 85%). 1H NMR (250 MHz,
CDCl3): δ (ppm) 7.30 (5H, broad s), 5.48 (1H, dd, J ) 6.6 and
7 Hz), 4.61 (2H, AB system with J ) 4 Hz), 3.92 (lH, d, J )
7.5 Hz), 3.49 (lH, d, J ) 8 Hz), 3.44 (2H, s), 2.33 (lH, dd, J gem
) 13.5 Hz and J trans ) 7 Hz), 1.98 (lH, partially masked dd,
J gem ) 13.5 Hz and J cis ) 6.6 Hz), 1.95 (lH, d, J ) 4.5 Hz),
1.57 and 1.39 (4H, 2m), 1.26 (3H, s), 0.94, 0.89 and 0.77 (9H,
3s); 13C NMR (62.5 MHz, CDCl3): δ (ppm) 183.1, 128.3, 128.0,
127.6, 127.5, 98.0, 87.7, 84.9, 76.0, 74.4, 73.3, 48.5, 47.7, 46.1,
40.5, 31.6, 25.1, 24.4, 22.1, 19.3, 10.7; mass (EI 70 eV): m/z
(1H, dd, J trans ) 16 Hz and J 4-5 ) 7 Hz), 6.05 (1H, d, J trans
)
16 Hz), 4.53 (2H, s), 3.30 (2H, s), 2.53 (1H, broad s, exchange-
able with D2O), 2.44 (2H, m), 2.20 (3H, s), 1.16 (3H, s); 13C
NMR (62.5 MHz, CDCl3): δ (ppm) 198.7, 143.9, 137.8, 134.1,
128.5, 127.9, 127.8, 76.9, 73.5, 72.0, 42.4, 26.8, 24.1; mass (EI
70 eV): m/z 246 (M+•), 231 (M+• - Me), 91 (100%); [R]20
)
D
-14.4 (c ) 1.51, CHCl3). Anal. Calcd for C15H20O3, MW:
248.32485; Calcd: C: 72.55%; H: 8.12%. Found: C: 72.38%;
H: 8.14%.
(1S,5R)-1,5-Dim eth yl-6,8-dioxabicyclo[3.2.1]octan e ((-)-
fr on ta lin ) (12). A solution of the ketone 11 (183 mg, 0.74
mmol) in methanol (2 mL) containing palladium on carbon (30
mg) was stirred overnight under 1 atm of hydrogen. The black
suspension was filtered through Celite, and water was added
to the filtrate which was extracted with dichloromethane (3
× 10 mL). The combined organic extracts were dried (Na2-
SO4) and filtered. Careful evaporation of the solvent gave the
title compound as a colorless liquid (103 mg, 99% yield). 1H
NMR (250 MHz, CDCl3): δ (ppm): 1.33 (3H, s, Me-C5), 1.45
(3H, s, Me-C1), 1.52-1.67 and 1.75-1.90 (5H and 1H, m, C2-H
× 2, C3-H × 2 and C4-H × 2), 3.46 (1H, dd, one of C7-H), 3.92
(1H, dd, one of C7-H); in C6D6, the signal of Me-C5 shifts to 1.1
ppm and the signal of Me-C1 to 1.53 ppm; in the presence of 1
equiv of Eu(hfc)3, the latter signal shifts to a single peak at
2.40 ppm. No peak corresponding to the other enantiomer
could be detected. Therefore, synthetic 12 was considered to
be enantiomerically pure within the detection range of 1H
NMR spectroscopy. 13C NMR (62.5 MHz, CDCl3): δ (ppm)
108.2 (C5), 80.1 (C1), 74.3 (C7), 34.6 and 34.0 (C2 and C4), 24.8
357 (MH+•), 328, 236, 196, 195, 152, 107, 91 (100%); [R]20
)
D
-61.5 (c ) 0.95, CHCl3). Anal. Calcd for C22H31NO3, MW:
357.4973; Calcd: C: 73.92%; H: 8.75%; N: 3.92%; Found: C:
73.84%; H: 8.51%; N: 3.83%.
(Me-C5), 23.2 (Me-C1), 18.1 (C3). [R]20 ) -51.2 (c ) 2.23,
D
Et2O); lit.10 -53 (c ) 1.33, Et2O); mass (EI 70 eV): m/z: 142
(M+•), 127; HRMS, calcd for C8H14O2, MW: 142.0994; found:
142.0991.
(S)-3-H yd r oxy-3-m et h yl-4-[(p h en ylm et h yl)oxy]b u t a -
n a l (9). m-Chloroperoxybenzoic acid (1.38 g, 8 mmol, 3.5
equiv) was added portionwise to a stirred solution of the benzyl
ether 7 (812 mg, 2.27 mmol) in diethyl ether (25 mL). The
clear and colorless solution was stirred at room temperature
for 1 h and then quenched by addition of an aqueous 10%
sodium bicarbonate/10% sodium thiosulfate solution and
vigorous stirring for 20 min. The mixture was decanted, and
the organic phase was washed with saturated sodium carbon-
ate solution and then with brine, dried (Na2SO4), filtered, and
concentrated in vacuo to give the crude nitrone 8 as a white
Su p p or tin g In for m a tion Ava ila ble: NMR spectra of all
cycloadducts and of compounds 6, 7, 9, 11, and 12 (28 pages).
This material is contained in libraries on microfiche, im-
mediately follows this article in the microfilm version of the
journal, and can be ordered from the ACS; see any current
masthead page for ordering information.
J O980326E