Regio- and stereochemistry of [2 + 2] photocycloadditions of imines to alkenes: A computational and experimental study

Article abstract of DOI:10.1021/jo8015242

(Chemical Equation Presented) The [2 + 2] photocycloaddition of isoxazolines to alkenes has been studied by means of CASPT2/6-31G*// CASSCF/6-31G*. The reaction outcome is influenced by the relative ratio of imine deactivation and photocycloaddition. Anal

Full text of DOI:10.1021/jo8015242

Regio- and Stereochemistry of [2 + 2] Photocycloadditions of Imines
to Alkenes: A Computational and Experimental Study
Diego Sampedro,* Alberto Soldevilla, Pedro J. Campos, Roberto Ruiz, and
Miguel A. Rodr´ıguez
Departamento de Qu´ımica, UniVersidad de La Rioja, Grupo de S´ıntesis Qu´ımica de La Rioja, Unidad
Asociada al C.S.I.C., Madre de Dios, 51, 26006 Logron˜o, Spain
diego.sampedro@unirioja.es
ReceiVed July 11, 2008
The [2 + 2] photocycloaddition of isoxazolines to alkenes has been studied by means of CASPT2/6-
31G*//CASSCF/6-31G*. The reaction outcome is influenced by the relative ratio of imine deactivation
and photocycloaddition. Analysis of the conical intersection points involved in the photoreaction shows
that fast deactivation is prevented when an electron-withdrawing group is placed in any position that can
affect the imine moiety. Computational data predict that the photoreaction will be regiospecific but without
stereoselectivity. Furthermore, the favored regioisomer will be different for alkenes with electron-
withdrawing or electron-releasing substituents. The results of a complementary experimental study correlate
well with the computational data. Several conclusions included in the present work could prove useful
for the generalization of the photocycloaddition of imines.
Introduction
imine moiety constrained in a five- or six-membered ring, but
few other structural requirements are clear. For instance, reports
of photocycloadditions of different olefins to oxadiazoles,7
azauracils,8 isoindolones, oxazolinones,9 and isoxazolines10 have
appeared. In order to shed some light on the reaction mechanism,
we carried out a computational study of the photocycloaddition
of imines to alkenes.11 This step is necessary before this
reactivity can be converted into a more general synthetic tool.
Our initial results11 showed that there are some structural
requirements beyond the confinement of the imine moiety in a
cycle for a successful photocycloaddition. Whereas pyrroline
(1) and isoxazoline (2) share the cyclic structure, our calculations
showed that photocycloaddition of 1 to ethylene is hampered
by fast deactivation to the ground state, whereas in 2 deactivation
is not that fast due to the presence of an energy barrier in the
excited state. Similar results were obtained for compounds for
[2 + 2] Photocycloadditions of olefins to CdC, CdO, and
CdS bonds to yield four-membered cycles are now well-
documented reactions and are useful tools in the synthesis of
cyclobutanes,1 oxetanes,2,3 and thietanes,4 respectively. How-
ever, similar photocycloaddition reactions of olefins to imines
are less common.5 Several isolated examples of this type of
reactivity have been reported and reviewed,6 although a
systematic study to explain such different behavior has yet to
be carried out. One of the required features of the CdN bond
to yield successful photocycloadditions to olefins is to have the
(1) Fleming, S. A. In Synthetic Organic Photochemistry; Griesbeck, A. G.,
Mattay, J., Eds.; Marcel Dekker: New York, 2005; p 141.
(2) Abe, M. In CRC Handbook of Photochemistry and Photobiology, 2nd
ed.; Horspool, W. M., Lenci, F., Eds.; CRC Press: Boca Raton, FL, 2004; Chapter
62.
(3) Griesbeck, A. G.; Bondock, S. In CRC Handbook of Photochemistry and
Photobiology, 2nd ed.; Horspool, W. M., Lenci, F., Eds.; CRC Press: Boca Raton,
FL, 2004; Chapter 59.
(4) Padwa, A.; Jacquez, M. N.; Schmidt, A. Org. Lett. 2001, 3, 1781–1783.
(5) Padwa, A. Chem. ReV. 1977, 77, 37.
(6) Nishio, T. In CRC Handbook of Organic Photochemistry and Photobi-
ology; Horspool, W., Ed.; CRC Press: Boca Raton, 1995; p 841.
(7) Oe, K.; Tashiro, M.; Tsuge, O. J. Org. Chem. 1977, 42, 1496.
(8) Swenton, J. S.; Hyatt, J. A. J. Am. Chem. Soc. 1974, 96, 4879.
(9) Anderson, D. R.; Keute, J. S.; Koch, T. H.; Moseley, R. H. J. Am. Chem.
Soc. 1977, 99, 6332.
(10) Kawamura, Y.; Kumagai, T.; Mukai, T. Chem. Lett. 1985, 1937.
(11) Sampedro, D. ChemPhysChem 2006, 7, 2456–2459.
10.1021/jo8015242 CCC: $40.75
Published on Web 10/01/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 8331–8336 8331

Sampedro et al.
Results and Discussion
On the basis of our previous results, it is clear that the key
feature of these reactions is the competition between fast
deactivation and photocycloaddition, with the relative ratio
controlled by the energy of the CI points involved. Compounds
3 and 4, in which fast deactivation (3) and photocycloaddition
(4) were detected experimentally, show potential energy surfaces
(PES) consistent with these facts (Figure 1). When the energy
of the CI point is low enough, fast deactivation can occur and
photocycloaddition, although energetically available, does not
take place. In contrast, when the CI point has a higher energy,
the system has to spend some time in the excited state, thus
allowing the reaction to occur. Accordingly, if we manage to
understand the geometrical aspects that affect the CI point and
its relative energy, we could gain a deeper understanding of
the reaction mechanism, and this information could be used to
predict the feasibility of any reaction and even design a system
that is able to successfully photocycloadd to olefins. With this
aim in mind, we analyzed the CI points for 3 and 4 (see Fig-
ure 2).
It can be seen that most of the geometric parameters for the
two CI points are very similar, with differences no greater than
0.003 Å. Clearly these small differences in the geometry can
hardly account for such different reactivity, in that 3 is known
not to photocycloadd to olefins and 4 yields azetidines after
irradiation. However, one of the distances, marked in red in
Figure 2, is sufficiently different in the two structures to allow
an explanation to be put forward. The key feature of these
geometries is the triangular shape of the CdN-O moiety, which
is planar in the ground state. This deformation causes the
ground-state energy to increase while the excited state is
stabilized, a situation that leads to the CI. The only distinction
between 3 and 4 is the cyano group. The electron-withdrawing
effect of the CN transmitted by the phenyl ring causes a partial
positive charge to appear on the carbon atom of the CdN-O
moiety in 4. This partial positive charge interacts with the O
atom, causing the C-O distance to shorten from 1.830 Å in 3
to 1.818 Å in 4. Perhaps this deformation contributes to an
increase in the energy of the CI point, thus allowing the system
to spend some time in the excited state and yield photocycload-
dition. To check that the enlargement of the active space is not
essential, we carried out further calculations on 4 including the
π orbitals of the cyano group (CAS(14,11)). Both geometries
and relative energies remained almost unaltered.
FIGURE 1. Critical points along the potential energy surfaces of
isoxazolines 3 and 4 (top).
which the experimental outcome is available. In the case of
3-phenyl-2-isoxazoline (3), fast deactivation takes place, whereas
3-(p-cyanophenyl)-2-isoxazoline (4) successfully undergoes
cycloaddition to alkenes. The reason behind this behavior seems
to lie not only in the feasibility of the photocycloaddition but
also in the competition between this reaction and fast deactiva-
tion to the ground state. We wish to report here the results of
a theoretical study into the effect of alkene substitution and
regiochemistry issues in the [2 + 2] photocycloaddition of
alkenes to isoxazolines and an experimental study carried out
to confirm the theoretical data.
Computational Details. All critical points were computed
using fully unconstrained ab initio quantum chemical computa-
tions in the framework of a CASPT2//CASSCF strategy.12 This
process requires the reaction coordinate to be computed at the
complete active space self-consistent field (CASSCF) level of
theory and the corresponding energy profile to be re-evaluated
at the multiconfigurational second order Møller-Plesset per-
turbation theory level (here we used the CASPT2 method
implemented in MOLCAS-6.4)13 to take into account the effect
of electron dynamic correlation. All computations were carried
out at the CASSCF level with the 6-31G* basis set. The active
space was chosen depending on the compounds considered but
in each case included the π and π* orbitals of the imine and
alkene moieties and the N lone pair. Depending on the alkene
considered, the active space also included ketone π and π*
orbitals, N or O lone pairs. The zeroth order wave function used
in the single-point CASPT2 calculations needed for the re-
evaluation of the energy profile is a three root (S0, S1, S2) state
average CASSCF wave function with the 6-31G* basis set. The
same type of wave function was used where necessary in order
to avoid convergence problems. The structure of the conical
intersection (CI) funnels associated with each path were
optimized by applying the methodology included in GAUSSIAN
03.14
Having assessed the effect of substitution on the CI geometry
and thus on the course of the reaction, we aimed to study the
regioselectivity of the reaction. For this purpose, we calculated
the critical points for the photocycloaddition of 2 to methyl vinyl
ether. We chose 2 for this part of the study because of its similar
behavior and deformation in the excited state as 411 and its
smaller size. It has been reported15 that a singlet excited state
is involved in the photochemistry of this kind of compound.
(14) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery Jr., J. A.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin, K. N.;
Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi, R.;
Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.; Petersson,
G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.;
Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts,
R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B.;
Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.;
Replogle, E. S.; Pople, J. A. Gaussian 03, reVision C.02; Gaussian Inc.:
Pittsburgh, PA, 2003.
(12) Olivucci, M.; Sinicropi, A. In Computational Photochemistry; Olivucci,
M., Ed.; Elsevier: Amsterdam, 2005.
(13) Andersson, K.; Barisz, M.; Bernhardsson, A.; Blomberg, M. R. A.;
Cooper, D. L.; Fleig, T.; Fu¨lscher, M. P.; de Graaf, C.; Hess, B. A.; Karlstro¨m,
G.; Lindh, R.; Malmqvist, P.-Å.; Neogra´dy, P.; Olsen, J.; Roos, B. O.;
Schmmelpfennid, B.; Schu¨ltz, M.; Sadlej, A. J.; Schu¨tz, M.; Seijo, L.; Serrano-
Andre´s, L.; Siegbahn, P. E. M.; Stårling, J.; Thorsteinsson, T.; Veryazov, V.;
Widmark, P.-O. MOLCAS, Version 6.4; University of Lund: Lund, Sweden, 2003.
8332 J. Org. Chem. Vol. 73, No. 21, 2008

[2 + 2] Photocycloadditions of Imines to Alkenes
FIGURE 2. Optimized CI geometries and charges for 3 and 4.
TABLE 1. CASPT2/6-31G* Relative Energies for Critical Points
along the Potential Energy Surface for the Reaction of 2 with
Methyl Vinyl Ether
SCHEME 1. Model Compounds
structure
FC
state
energya
S0
S1
S2
S1
S0
S1
S0
S1
S0
S1
S0
S1
S0
S0
S0
S0
0.0
158.4
203.2
124.5
68.9
76.0
85.3
88.7
69.6
76.8
82.5
85.5
-9.8
-4.2
-4.6
-8.0
Min S1
CI-a
CI-b
CI-c
CI-d
SCHEME 2. Synthesized 3-Arylisoxazolines
Prod-a
Prod-b
Prod-c
Prod-d
a CASPT2 energy (kcal/mol) relative to the ground-state minimum.
Accordingly, we found no experimental evidence of the
participation of triplet states. The results are shown in Table 1
and Figure 3.
As can be seen, four approaches between the two molecules
are possible, which implies the presence of up to four different
CI points and four different photoproducts. These four pos-
sibilities correspond to two regioisomers, each with endo and
exo descriptions available. Geometries for the CI points reflect
the concerted reaction between the imine and alkene moieties.
Both CdC (ca. 1.43 Å) and CdN (ca. 1.36 Å) distances increase
compared to the ground state, and pyramidalization is clear in
the atoms involved. The forming C-C bond in the azetidine
has distances of 2.09 Å in CI-a and CI-c (lower in energy) and
2.17 Å in CI-b and CI-d. The forming N-C bond varies from
ca. 2.18 Å in CI-a and CI-c to ca. 2.05 Å in CI-b and CI-d.
Comparison of the energetics for the formation of these four
possible compounds reveals that, although all four are achiev-
able, there is a clear preference of at least 10 kcal/mol for one
regioisomer (CI-a, CI-c) with respect to the other (CI-b, CI-
FIGURE 3. Optimized CI geometries and CASPT2 relative energies (S0/S1, kcal/mol) for the reaction of 2 with methyl vinyl ether. Atomic charges
are also shown.
J. Org. Chem. Vol. 73, No. 21, 2008 8333

Sampedro et al.
TABLE 2. CASPT2/6-31G* Relative Energies for Critical Points
along the Potential Energy Surface for the Reaction of 2 with
Dimethyl Vinyl Amine
for each regioisomer. Thus, the computational data predict that
a preference for the endo or the exo products is not expected
and Prod-a and Prod-c are equally probable. However, the
CASPT2 correction to the CASSCF energy causes an energy
gap between states to appear that could conceal the relative
stability of the endo and exo approaches. It should be noted
that conclusions drawn from the study of 2 could be extrapolated
to other isoxazolines with the same photochemical behavior
(such as 4).
We found it of interest to confirm the prediction from the
calculations by means of experimental photoreactions. We
synthesized several 3-arylisoxazolines16 (see Scheme 2) and
carried out irradiation experiments in the presence of an excess
of furan using acetonitrile as the solvent.
structure
FC
state
energya
S0
S1
S2
S1
S0
S1
S0
S1
S0
S0
0.0
140.8
200.2
112.8
86.7
92.5
60.7
69.0
1.1
Min S1
CI-e
CI-f
Prod-e
Prod-f
0.3
a CASPT2 energy (kcal/mol) relative to the ground-state minimum.
In agreement with the theoretical predictions, photoadducts
could not be detected upon irradiation of the phenyl or
p-methoxyphenyl derivatives (5a and 5b, respectively) and
furan, but the presence of an electron-withdrawing group on
the phenyl ring (5c-5f) enables the reaction to proceed (Scheme
3). The photoreaction led to the formation of only one
regioisomer but as two diastereoisomer pairs, which were
obtained in pure form by column chromatography on silica gel.
The configuration was assigned on the basis of X-ray structure
(6f, see Supporting Information) and difference NOESY (6′f)
experiments. Interestingly, the reaction led to the formation of
exo cycloadduct products. Although the formation of the endo
products cannot be excluded, we were unable to isolate them
or even detect them in the photomixture. This fact is very
important because we were unable to predict the stereochemical
outcome of the reaction from our computational results.
According to our calculations, the electron-withdrawing effect
of the cyano group, transmitted along the phenyl ring, leads to
a certain partial positive charge on the CI point, and this favors
the photocycloaddition. This effect does not depend on the
position of the cyano group as far as the geometries and energies
for the structures of the meta and para isomers are very similar,
although experimentally we found a higher yield when the cyano
group was in the meta position.
FIGURE 4. Optimized CI geometries and CASPT2 relative energies
(kcal/mol) for the reaction of 2 with dimethyl vinyl amine.
TABLE 3. CASPT2/6-31G* Relative Energies for Critical Points
along the Potential Energy Surface for the Reaction of 2 with
Acrylaldehyde
structure
FC
state
energya
S0
S1
S2
S1
S0
S1
S0
S1
S0
S0
0.0
122.6
191.0
74.4
73.2
74.8
79.8
84.0
3.4
Min S1
CI-g
CI-h
Prod-g
Prod-h
-6.4
a CASPT2 energy (kcal/mol) relative to the ground-state minimum.
The next step was to extend the study of the photocycload-
dition to different olefins. This served not only to check the
experimental feasibility of azetidines with diverse substitution
patterns but also to obtain information on the effect that the
alkene substitution has on the course of the reaction. We
therefore calculated the critical points on the potential energy
surface for the photocycloaddition of 2 to dimethyl vinyl amine,
an N-substituted olefin. The results are shown in Table 2 and
Figure 4.
SCHEME 3. Irradiation of 3-Arylisoxazolines in the
Presence of Furan
Similar results were found for the dimethyl vinyl amine, and
this is not surprising given that both electronic and steric factors
are analogous to those in methyl vinyl ether. For the sake of
simplicity, comparison of endo versus exo photocycloaddition
was not considered. Two regioisomers are still possible in this
reaction, but as in the case of O-substituted olefins, one of them
should be formed preferentially as a result of interactions
between partial charges. The geometries for the CI points have
parameters similar to those obtained in the case of methyl vinyl
ether. The distances for the CdC (ca. 1.44 Å) and CdN (ca.
1.36 Å) bonds together with the pyramidalization observed are
consistent with a concerted reaction. The forming C-C bond
d). Thus, for the reaction of 2 with methyl vinyl ether (and by
extension any O-substituted alkene with similar electronic and
steric factors), the experimental outcome should show mainly
one regioisomer, represented by Prod-a and Prod-c. The reason
behind this seems to be related to the partial charges in the atoms
involved. In the isoxazoline CdN moiety, the nitrogen atom is
electron-rich and thus will preferentially interact with the
electron-poor carbon in the olefin moiety to yield Prod-a and
Prod-c (Figure 3). For each regioisomer, two different pos-
sibilities are still available, i.e., the endo and exo approaches.
As can be seen from the results shown in Figure 3 and Table 1,
there is little difference, if any, between these two approaches
(15) Kumagai, T.; Shimizu, K.; Kawamura, Y.; Mukai, T. Tetrahedron 1981,
37, 3365–3376.
(16) Oppolzer, W.; Kingma, A. J.; Pillai, S. K. Tetrahedron Lett. 1991, 4893.
8334 J. Org. Chem. Vol. 73, No. 21, 2008

[2 + 2] Photocycloadditions of Imines to Alkenes
FIGURE 5. Optimized CI geometries and CASPT2 relative energies (S0/S1, kcal/mol) for the reaction of 2 with acrylaldehyde. Charge distributions
are also shown.
SCHEME 4. Irradiation of Isoxazoline 5d in the Presence of Ethyl Acrylate
in the azetidines have distances of 2.18 Å in CI-e (higher in
energy) and 1.94 Å in CI-f, while the forming N-C bond varies
from 2.02 Å in CI-e to 2.33 Å in CI-f.
A very different outcome could result when the electronic
character of the olefin is sufficiently different.17 To test this
hypothesis, we calculated the critical points for the photocy-
cloaddition of 2 to acrylaldehyde. The results of these calcula-
tions are shown in Table 3 and Figure 5.
The results described here allow an explanation for some
previous data10 and could be used to shed light on the
mechanism of the photocycloaddition of imines. Some structural
requirements for isoxazolines are made clear, but further
computational and experimental effort will be needed to
generalize this reaction to other imines.
Conclusions
Once again, two different regioisomers are available. In this
case the regioisomer predicted to be formed predominantly is
different from the one formed for methyl vinyl ether and
dimethyl vinyl amine. The reason for the preference, however,
remains the same: the electrostatic interactions account for the
regiochemistry. Once again, very similar geometries were found
for the CI points. Values of ca. 1.44 Å were obtained for the
CdC distances and ca. 1.36Å for the CdN. The forming C-C
bond in the azetidine has a distance of 2.27 Å in CI-g (lower
in energy) and 2.16 Å in CI-h, while the forming N-C bond
has a distance of 1.97 Å in CI-g and 2.00 Å in CI-h.
We therefore investigated the irradiation of 5c and 5d in the
presence of an alkene bearing an electron-withdrawing group,
in this case ethyl acrylate (Scheme 4). In agreement with the
theoretical calculations, the reaction led to the formation of only
one regioisomer, which correlates with CI-g, but three diaste-
reoisomer pairs, one corresponding to the endo addition and
two to the exo addition. This is in contrast with the reaction
with furan, where only the exo products were isolated (Scheme
3). From the reaction mixture only the endo diastereomer pair
(7′′) was obtained in pure form after column chromatography
on silica gel. The configuration was assigned on the basis of
difference NOESY experiments (7′′c) and the 1H NMR chemical
shifts of the signal corresponding to HCOOEt, since these ethyl
protons are shielded by the diamagnetic anisotropy of the phenyl
ring (7 and 7′).18
The [2 + 2] photocycloaddition of alkenes to imines has been
explored by computational methods. Interpretations based on
the energy and geometry of critical points along the potential
energy surface enabled the identification of some structural
requirements of the imine moiety needed to yield azetidines.
We also studied the regio- and stereochemistry of the reaction
using a model compound for the imine and different olefins. It
was found that the regiochemistry is completely dependent on
the alkene used, as illustrated by the fact that the regioisomer
obtained is different when olefins substituted by electron-
withdrawing or electron-releasing groups are involved. A
preference for a particular stereoisomer was not found, and as
a consequence, mixtures of stereoisomers were predicted. To
test the computational data, we carried out a complementary
experimental study. A high correlation was found between data
obtained in the calculations and in the experiments, a finding
that supports the mechanistic proposals. Our results also provide
an explanation for previous experimental data and could be used
as a guide for the generalization of the photocycloaddition of
imines to alkenes.
Experimental Section
Irradiation of 3-Arylisoxazolines 5. General Procedure.
3-Arylisoxazolines 5 (0.6 mmol) and 10 equiv of furan or ethyl
acrylate (6 mmol) were dissolved in 50 mL of deoxygenated
acetonitrile and irradiated at room temperature under Ar atmosphere
through Pyrex glass with a 400 W medium-pressure mercury lamp,
until 5 was consumed (TLC, hexane/AcOEt, 9:1). Solvents were
then removed with a rotary evaporator, and the products were
separated by column chromatography (silica gel, hexane/AcOEt,
9:1).
(17) The experimental outcome in the case of N-substituted olefins should
be the same that the one found for O-substituted olefins, as predicted by the
theoretical data. Thus, experimental exploration of the regio- and stereochemistry
is well represented by methyl vinyl ether. Nevertheless, we tried to carry out
the reaction with N-vinyl pyrrolidinone to further confirm the mechanism.
Unfortunately, although some reaction took place, we found the product to be
photoactive and no conclusive data could be obtained from the reaction mixture.
Further efforts are currently under way.
(2S*,6S*,7S*,9R*)-9-Butyl-7-(4-cyanophenyl)-3,10-dioxa-1-
azatricyclo[5.3.0.02,6]deca-4-ene (6d). Rf ) 0.30 (silica gel, hexane/
AcOEt, 9:1), yield 16 mg, 9%; 1H NMR (300 MHz, CDCl3) δ 7.65
(18) Silverstein, R. M.; Webster, F. X. Spectrometric Identification of Organic
Compounds, 6th ed.; Wiley: New York, 1998.
J. Org. Chem. Vol. 73, No. 21, 2008 8335

Sampedro et al.
(d, J ) 8.5 Hz, 2H), 7.54 (d, J ) 8.5 Hz, 2H), 6.35-6.33 (m, 1H),
5.61 (d, J ) 7 Hz, 1H), 4.74 (m, 1H), 4.06-4.02 (m, 1H),
4.05-3.96 (m, 1H), 2.63 (dd, J ) 13, 5.5 Hz, 1H), 2.40 (dd, J )
13, 10 Hz, 1H), 1.85-1.77 (m, 1H), 1.65-1.58 (m, 1H), 1.44-1.26
(m, 4H), 0.90 (t, J ) 6 Hz, 3H); 13C NMR (300 MHz, CDCl3) δ
148.9, 147.3, 132.1, 126.6, 118.7, 110.6, 102.1, 100.0, 83.2, 79.0,
49.7, 49.0, 33.6, 28.5, 22.6, 13.9; ES(+) m/z 297.2 (M + 1).
HRMS(+) m/z 229.1339 ([(M - furan)H+], calcd 229.1341. Anal.
Calcd for C18H20N2O2: C, 72.95; H, 6.80; N, 9.45. Found: C, 73.05;
H, 6.85; N, 9.36.
MHz, CDCl3) δ 148.4, 148.2, 132.2, 126.3, 118.8, 110.6, 102.1,
96.4, 82.2, 81.5, 49.5, 47.3, 32.5, 28.5, 22.6, 13.9; ES(+) m/z 297.2
(M + 1). HRMS(+) m/z 229.1339 ([(M - furan)H+], calcd
229.1341. Anal. Calcd for C18H20N2O2: C, 72.95; H, 6.80; N, 9.45.
Found: C, 73.10; H, 6.96; N, 9.38.
Acknowledgment. We thank the Spanish MEC (CTQ2007-
64197) for financial support. D.S. is financed by the Ramo´n y
Cajal program from the MEC. R.R. thanks the Comunidad
Autonoma de La Rioja for his fellowship.
(2S*,6S*,7S*,9S*)-9-Butyl-7-(4-cyanophenyl)-3,10-dioxa-1-
azatricyclo[5.3.0.02,6]deca-4-ene (6′d). Rf ) 0.15 (silica gel,
hexane/AcOEt, 9:1), yield 46 mg, 26%; 1H NMR (300 MHz,
CDCl3) δ 7.64 (d, J ) 8.5 Hz, 2H), 7.46 (d, J ) 8.5 Hz, 2H),
6.35-6.33 (m, 1H), 5.73 (d, J ) 6.5 Hz, 1H), 4.73-4.71 (m, 1H),
4.70-4.58 (m, 1H), 3.92-3.86 (m, 1H), 2.83 (dd, J ) 12, 5 Hz,
1H), 1.89 (dd, J ) 12, 11 Hz, 1H), 1.75-1.67 (m, 1H), 1.57-1.49
(m, 1H), 1.42-1.25 (m, 4H), 0.92-0.87 (m, 3H); 13C NMR (300
Supporting Information Available: Experimental proce-
dures, characterization data for new compounds, X-ray structure
details for compound 6g in CIF format, and Cartesian coordi-
nates for the calculated geometries. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO8015242
8336 J. Org. Chem. Vol. 73, No. 21, 2008