Examination of homo-[3 + 2]-dipolar cycloaddition: Mechanistic insight into regio-and diastereoselectivity

Article abstract of DOI:10.1021/jo702073w

(Chemical Equation Presented) The reaction of 2,3-disubstituted-1,1- cyclopropanediesters with nitrones under Lewis acid conditions produces tetrahydro-1,2-oxazines in which the cis/trans relationship of the cyclopropanes is not conserved. Reacting nitrones with 2,3-cis-disubstituted cyclopropanes lead to 5,6-trans-oxazines, and 2,3-trans-disubstituted cyclopropanes lead to 5,6-cis-oxazines. This observed stereochemical inversion provides evidence for a stepwise annulation mechanism in the preparation of tetrahydro-1,2-oxazines.

Full text of DOI:10.1021/jo702073w

in energy.5 Herein, we present a study of the reaction mechanism
that relies on the use of a modified cyclopropane diester as a
mechanistic probe.
Examination of Homo-[3 + 2]-Dipolar
Cycloaddition: Mechanistic Insight into Regio-
and Diastereoselectivity
Figure 1 shows a simple reaction/energy diagram of a
stepwise and a concerted reaction between a nitrone and a 1,1-
cyclopropanediester. In the concerted scenario, the cyclopropyl
substituent R3 would necessarily be exo to the nitrone to satisfy
the observed 3,6-cis relationship observed in the adducts (since
R1 and R2 are trans in the nitrone). If this mechanism was
operational, R3 and R2 would end up cis to Ha in the adduct. In
a stepwise mechanism, we envision a nucleophilic ring opening
of the cyclopropane (with inversion) by the nitrone oxygen
resulting in a zwitterionic intermediate. Ring closure of the
malonic anion onto the iminium moiety would result in an
adduct in which R3 is cis to Hb. Of course, in our previous
studies, the adducts from each mechanistic pathway were
indistinguishable. We reasoned, then, that the positioning of an
additional substituent on the cyclopropane vicinal to both R3
and the diester moiety would allow, upon examination of the
adducts, insight into the concerted or stepwise nature of the
reaction.
Avedis Karadeolian and Michael A. Kerr*
Department of Chemistry, The UniVersity of Western Ontario,
London, Ontario, Canada, N6A 5B7
makerr@uwo.ca
ReceiVed September 20, 2007
The reaction of 2,3-disubstituted-1,1-cyclopropanediesters
with nitrones under Lewis acid conditions produces tetrahy-
dro-1,2-oxazines in which the cis/trans relationship of the
cyclopropanes is not conserved. Reacting nitrones with 2,3-
cis-disubstituted cyclopropanes lead to 5,6-trans-oxazines,
and 2,3-trans-disubstituted cyclopropanes lead to 5,6-cis-
oxazines. This observed stereochemical inversion provides
evidence for a stepwise annulation mechanism in the
preparation of tetrahydro-1,2-oxazines.
Recently, we reported the reaction of nitrones 1 with 1,1-
cyclopropanediesters 2 to form tetrahydro-1,2-oxazines 3 in what
we termed a homo-[3 + 2]-dipolar cycloaddition.1 This moniker
(as opposed to a [3 + 3] cycloaddition) is preferred since one
of the carbons in the cyclopropane is an electronic bystander
in the reaction. While this reaction has been developed by us2
and others3 over the last several years, and has been shown to
be a powerful tool in total synthesis,4 the exact mechanism of
the reaction remains unclear.
FIGURE 1. Stepwise and concerted mechanisms for cycloaddition.
Scheme 1 shows the preparation of two cyclopropane sub-
strates for our study. Dihydroxylation of trans-â-methylstyrene
4 gave diol 5, which upon treatment with thionyl chloride gave
cyclic sulfite 6. Oxidation to sulfate 7 gave a suitable substrate
for conversion to cyclopropane 8 via double displacement with
the anion derived from dimethyl malonate. The diastereomeric
cyclopropane 11 was prepared directly by carbenoid insertion
of diazomalonate 10 to cis-â-methylstyrene 9.
We recently subjected the reaction to DFT analysis to probe
the likely course of the reaction only to find global minima for
both a concerted and a stepwise fashion, which were very close
* Corresponding author. Fax: (519) 661-3022.
(1) Young, I. S.; Kerr, M. A. Angew. Chem., Int. Ed. 2003, 42, 3023.
(2) (a) Ganton, M. D.; Kerr, M. A. J. Org. Chem. 2004, 69, 8554. (b)
Young, I. S.; Kerr, M. A. Org. Lett. 2004, 6, 139.
(3) (a) Kang, Y.-B.; Sun, X.-L.; Tang, Y. Angew. Chem., Int. Ed. 2007,
46, 3918. (b) Sibi, M. P.; Zhihua, M.; Jasperse, C. P. J. Am. Chem. Soc.
2005, 127, 5764.
With cyclopropanes 8 and 11 in hand, we proceeded to subject
them to cycloaddition reactions with two representative nitrones
(4) (a) Young, I. S.; Kerr, M. A. J. Am. Chem. Soc. 2007, 129, 1465. (b)
Carson, C. A.; Kerr, M. A. Angew. Chem., Int. Ed. 2006, 45, 6560.
(5) Wanapun, D.; Van Gorp, K. A.; Mosey, N. J.; Kerr, M. A.; Woo, T.
K. Can. J. Chem. 2005, 83, 1752-1767.
10.1021/jo702073w CCC: $37.00 © 2007 American Chemical Society
Published on Web 11/14/2007
J. Org. Chem. 2007, 72, 10251-10253
10251

SCHEME 1. Synthesis of trans- and cis-Disubstituted
Cyclopropanes
SCHEME 3. Reaction of Cyclopropane 8 with Nitrones 12
and 13
SCHEME 2. Reaction of Cyclopropane 11 with Nitrones 12
and 13
amount of regioisomer 16 was observed in reactions of nitrone
12. This is surprising since we have come to expect that the
carbocation-stabilizing ability of the cyclopropane substituent
vicinal to the diester moiety is key to promoting ring opening
at that position.
Cyclopropane 8 behaved quite differently in the cycloaddition
reactions, most notably in its sluggishness. The results are shown
in Scheme 3.6 In contrast to reactions with 11, conditions
involving refluxing toluene were necessary to effect appreciable
reaction and the yields were somewhat diminished. In these
reactions, all isolated adducts bore a cis relationship between
substituents at the 5- and 6-positions, indicating again that the
reaction had proceeded with inversion of configuration during
the cyclopropane ring opening event. The most striking result
to us, however, was the fact that for the first time ever, the
3,6-trans adduct was the major diastereomer isolated. In essence,
then, the extra substituent, which is on a carbon not electroni-
cally involved in the reaction, influenced the stereochemical
relationship.
Figure 2 presents a reasonable rationale for the stereochemical
distribution of products from both cyclopropane 8 and 11. In
the case of cyclopropane 11, nitrone attack on the cyclopropane
results in intermediate I, in which the phenyl and methyl groups
originating from the cyclopropane can both be equatorial in the
proposed chair-like conformation. Intermediate I then proceeds
uneventfully to the product with the expected and observed
stereochemical outcome.
Cyclopropane 8, on the other hand, would result in intermedi-
ate II upon reaction with the nitrone, necessarily putting the
methyl group in an axial orientation. This should result in an
unfavorable 1,3-diaxial interaction with the phenyl substituent
from the nitrone. The formation of this higher energy intermedi-
ate would imply a slower rate of reaction. Relief of the
unfavorable 1,3-diaxial interaction could be achieved upon
changing to a boat conformation III. Ring closure from III
would produce the unusual 3,6-trans-tetrahydro-1,2-oxazines
observed.7
12 and 13. Aside from their ease of preparation, these nitrones
were chosen because they bear an aryl and an alkyl nitrogen
substituent, respectively. Scheme 2 shows the reaction of
cyclopropane 11 with nitrones 12 and 13.
Several observations from Scheme 2 are worth noting. In all
cases, the groups at the 5- and 6-positions of the tetrahydroox-
azine ring had a trans relationship, indicating that ring opening
of the cyclopropane had occurred with inversion of configura-
tion.6 Although slightly lower than usual for this transformation,
the overall yields of adducts were reasonably good. As expected
the 3,6-cis diastereomer was the major isomer although a
significant amount of the 3,6-trans isomer 18 was isolated in
the reaction with nitrone 13. Last it is interesting that a small
(6) Stereochemistry of all compounds was unambiguously confirmed by
either X-ray crystallographic analysis or NOE analysis. See Supporting
Information for details.
10252 J. Org. Chem., Vol. 72, No. 26, 2007

CHART 1. Cycloadducts from Deuterium Labeled
Cyclopropanes
FIGURE 2. Mechanistic rationale for observed product mixture.
SCHEME 4. Preparation of Cyclopropanes 26 and 29
forcing the adoption of a boat-like transition state. When the
cycloadditions were repeated with deuterium labeled cyclopro-
panes, thus removing the steric effects, the expected adducts
were formed in excellent yields where clean inversion of
stereochemistry had occurred at the cyclopropyl carbon under-
going attack. These findings further strengthen the postulation
of a stepwise mechanism as being the mode of reaction for the
cycloaddition.
Experimental Section
To remove the steric effects imposed by the methyl group
on the cyclopropanes, we prepared two cyclopropanes 26 and
29, which bore a deuterium label cis and trans to the phenyl
moiety, respectively (Scheme 4). Dihydroxylation of cis-â-
deuterostyrene 248 followed by bis-mesylation provided 25,
which was a suitable substrate for cyclopropane formation with
dimethylmalonate and sodium hydride. Cyclopropane 29 was
prepared in the same manner, however, from trans-â-deu-
terostyrene 28, which in turn was available from hydride
reduction of phenylacetylene 27 and a deuterium quench.
As expected, cycloadditions employing cyclopropanes 26 and
29 proceeded in excellent yields as seen from our previous work.
Chart 1 summarizes the results. In all cases, the cyclopropane
ring opening proceeded with inversion of configuration. That
is, the trans-cyclopropane 29 resulted in cis adducts 32 and 33,
while cis-cyclopropane 26 yielded trans adducts 30 and 31.
In conclusion, by utilizing cis/trans-disubstituted cyclopro-
panes in cycloadditions with nitrones, we have shown that the
reaction proceeds through a stepwise mechanism with inversion
of stereochemistry at the cyclopropane carbon undergoing attack
by the nitrone oxygen. The adducts obtained from cis-cyclo-
propane 11 yield a cis 3,6-substituted adduct, while those
obtained from trans-cyclopropane 8 yield adducts with a trans
3,6 relationship as the major products. This observation is
attributed to the unfavorable nature of a chair-like transition
cis-Cyclopropane 11 (0.1 g, 0.4 mmol) and nitrone 12 (0.16 g,
0.8 mmol) were dissolved in 5 mL of toluene. Yb(OTf)3 (0.025 g,
0.04 mmol) was added to the solution, and the reaction mixture
was heated to 60 °C for 16 h. The reaction was poured into water
and extracted with EtOAc, and the combined layers were washed
with brine and then dried with anhydrous MgSO4. The crude
mixture was purified by flash column chromatography, and a
mixture of diastereomers (1:0.06:0.06, totaling 0.102 g, 57%) was
isolated as a yellow solid. The major isomer was recrystallized from
a solution of EtOAc/hexanes as large white crystals. 1H NMR (600
MHz, CDCl3): δ ) 7.63 (d, 2H, J ) 7.2 Hz), 7.54 (d, 2H, J ) 6.6
Hz), 7.46 (t, 2H, J ) 7.2 Hz), 7.41 (t, 1H, J ) 6.6 Hz), 7.25 (t,
2H, J ) 7.2 Hz), 7.21 (t, 1H, J ) 7.2 Hz), 7.10 (t, 2H, J ) 7.2
Hz), 6.99 (d, 2H, J ) 7.8 Hz), 6.77 (t, 1H, J ) 7.2 Hz), 5.67 (s,
1H), 4.82 (d, 1H, J ) 10.8 Hz), 3.90 (s, 3H), 3.42 (s, 3H), 3.25
(dq, 1H, J ) 10.8, 6.6 Hz), 0.92 (d, 3H, J ) 6.6 Hz); 13C NMR
(100 MHz, CDCl3): δ ) 170.2, 168.7, 148.3, 137.9, 135.6, 130.2,
128.7, 128.6, 128.4, 128.1, 128.02, 127.98, 121.2, 115.5, 85.0, 67.4,
62.6, 52.5, 52.2, 35.6, 14.2; IR (thin film): νmax ) 3063, 3032,
3002, 2979, 2952, 2917, 2887, 2849, 1735, 1598, 1492, 1456, 1436,
1253, 1230, 1043, 910, 754, 733, 700, 650, 605; HRMS calcd for
C27H27NO5 445.1889, found 445.1882 amu, mp ) 161-163 °C.
Acknowledgment. We thank the Natural Sciences and
Engineering Research Council of Canada for generous funding
of this research. We are grateful to Doug Hairsine for performing
MS analyses and Dr. Michael Jennings for X-ray data.
Supporting Information Available: Complete experimental
procedures as well as 1H NMR and 13C NMR, IR, MS, and crystal
structure data for 14, 17, 20, and 22. This material is available free
of charge via the Internet at http://pubs.acs.org.
(7) It should be noted that we are assuming that the stereochemical
integrity of the nitrone iminium bond (Z-geometry) is maintained during
the course of the reaction. While the Z-geometry is preferred for nitrones,
it cannot be ruled out that the iminium geometry is fluxional under the
reaction conditions.
(8) Dolbier, W. S.; Wicks, G. E. J. Am. Chem. Soc. 1985, 107, 3626.
JO702073W
J. Org. Chem, Vol. 72, No. 26, 2007 10253