Regioselective 6-endo cyclization of 5-carbomethoxy-5-hexenyl radicals: A convenient synthesis of derivatives of the 1-azabicyclo[2.2.1]heptyl system

Article abstract of DOI:10.1021/ol990557n

(equation presented) Ring closure of the 5-carbomethoxy-5-hexenyl radical is governed largely by the polar effect, and as predicted by frontier molecular orbital considerations, endo cyclization predominates, leading to cyclohexyl rather than cyclopentyl-based products. In cyclization of the corresponding β-ammonio species 18, stereoelectronic effects do not distinquish between attack of the radical center at C3 or C4, each of which represents a 5-exo ring closure. The radical 18 is found to cyclize with great rapidity and with high stereoselectivity to give bicyclo[2.2.1]heptane products in accordance with expectation based on polar effects; this transformation represents an excellent entry to the physiologically important bicyclic ester 17.

Full text of DOI:10.1021/ol990557n

ORGANIC
LETTERS
1999
Vol. 1, No. 3
363-365
Regioselective 6-Endo Cyclization of
5-Carbomethoxy-5-hexenyl Radicals:
A Convenient Synthesis of Derivatives
of the 1-Azabicyclo[2.2.1]heptyl System
Ernest W. Della,* Chris Kostakis, and Paul A. Smith
Department of Chemistry, Flinders UniVersity, Bedford Park, South Australia 5042
ern.della@flinders.edu.au
Received April 5, 1999
ABSTRACT
Ring closure of the 5-carbomethoxy-5-hexenyl radical is governed largely by the polar effect, and as predicted by frontier molecular orbital
considerations, endo cyclization predominates, leading to cyclohexyl rather than cyclopentyl-based products. In cyclization of the corresponding
â-ammonio species 18, stereoelectronic effects do not distinquish between attack of the radical center at C3 or C4, each of which represents
a 5-exo ring closure. The radical 18 is found to cyclize with great rapidity and with high stereoselectivity to give bicyclo[2.2.1]heptane products
in accordance with expectation based on polar effects; this transformation represents an excellent entry to the physiologically important
bicyclic ester 17.
Intensive investigations over many years have placed the
understanding of the factors which influence the regiochem-
istry of ring closure of the 5-hexenyl radical 1 on a firm
foundation.1 It is generally accepted that this reaction has
an early transition state and is under kinetic control.1d The
overwhelming preference for 5-exo cyclization to give the
cyclopentylmethyl radical 2 over 6-endo ring closure, which
gives 3, is a consequence of the interplay of three main
factors, the stereoelectronic, polar, and steric effects. All three
effects favor production of 2. The regiochemistry of cycliza-
tion of substituted 5-hexenyl radicals provides a striking
example in which additional steric effects are thought to
assume a much more prominent role in controlling the mode
of ring closure.1e,2 Rearrangement of the 5-methyl-5-hexenyl
radical 4, for example, is observed to afford a 40:60 ratio of
the radicals 5:6. This reversal of regioselectivity has been
ascribed to a decrease in the rate of 5-exo cyclization; the
rate of formation of 6 has been shown to be essentially
unaffected by the attached methyl group.
(1) For leading references, see: (a) Beckwith, A. L. J. Chem. Soc. ReV.
1993, 143. (b) RajanBabu, T. V. Acc. Chem. Res. 1991, 24, 139. (c)
Beckwith, A. L. J. Tetrahedron 1981, 37, 3073. (d) Giese, B. Angew. Chem.,
Int. Ed. Engl. 1983, 22, 753. (e) Julia, M. Acc. Chem. Res. 1971, 4, 386.
(e) Giese, B. Radicals. In Organic Synthesis: Formation of Carbon Carbon
Bonds; Baldwin, J. E., Ed.; Pergamon: New York, 1986.
(2) Beckwith, A. L. J.; Moad, G. J. Chem. Soc., Chem. Commun. 1974,
472.
10.1021/ol990557n CCC: $18.00 © 1999 American Chemical Society
Published on Web 07/15/1999

We wish to report our observations on the behavior of
the 5-carbomethoxy-5-hexenyl radical 7. A priori, stereo-
electronic factors predict C5 to be the preferred site of attack
by the radical center in ring closure of 7. In terms of steric
effects, it seems reasonable to assume, on the basis of the
conformational energies of the ester and methyl groups in
cyclohexane, that the less bulky ester group would have a
smaller influence on the regiochemistry. The expectation is,
therefore, that the combined effect of these two factors on
ring closure of 7 should lead to an approximately equal
mixture of 8 and 9. The polar effect, however, is seen to
predict a high preference for the 6-endo product 9. Consid-
eration of the appropriate frontier orbitals of the reactants,1d
viz., the (nucleophilic) radical SOMO and the alkene LUMO,
reveals that the most favorable interaction occurs between
C1 and C6 because of the greater magnitude of the orbital
coefficient and hence greater concentration of spin on C6
compared to that on C5.
exo ring closure in the radical 4, there is no apparent
attenuation in rate as a result of a steric effect of the COOMe
group in the rate of exo cyclization of the radical 7; indeed,
this mode of ring closure occurs more rapidly in 7 than in
the parent radical 1.
We have recently been investigating ring closure of R-
and â-ammonio-substituted 5-hexenyl radicals as a possible
entry into heterocyclic compounds.5 We have observed that
the radical 15, for example, undergoes cyclization with high
stereoselectivity, giving 16 exclusively.5c One of the major
objectives in our ongoing investigation is the synthesis of
3-carb(om)ethoxy-1-azoniabicyclo[2.2.1]heptane bromide (17).
The methyl ester (17, R ) CH3) is of considerable interest
because, together with its N-demethylated derivative, it has
been shown6 to possess important physiological properties.
We have employed the Barton ester 13 of the half ester
123 as precursor to the radical 7. A solution of tributyltin
hydride (1.05 equiv) in benzene containing a catalytic amount
of AIBN was added over 15 min to a solution of the ester
13 in benzene maintained at 80 °C and illuminated by a
tungsten lamp (300 W). After being allowed to stand for a
further hour under these conditions, the mixture was quenched
and worked up. Analysis of the product by GC revealed the
presence of a 15:85 mixture of the cyclic esters 10 and 11.
The product of reduction, 14, was not detected. This
observation is strongly indicative of a more dominant role
exerted by the polar effect in directing the mode of ring
closure of the radical 7.
We were particularly interested in exploiting the observed
preference for 6-endo cyclization in the radical 7 combined
with the facility for ring closure of 15 as the basis of a
potential synthetic route to the ester 17. The target radical
selected to affect this synthesis was the cyclopentenyl species
18.
The radical 18 has rather interesting structural ramifica-
tions compared to its acyclic analogue 7. In fact, 18 is
uniquely placed because there is no distinction between C3
and C4 as a potential site for 5-exo attack of the radical
center. According to stereoelectronic considerations, there-
fore, there is no expected differentiation for ring closure at
C3 or C4 in 18. Furthermore, both steric and polar factors
We also undertook a kinetic study of the rate of cyclization
of 7, the results of which are collated in Table 1; included
Table 1. Experimental Kinetic Data for Cyclization of the
5-Hexenyl Radicals 1, 4, and 7
radical
exo:endoa
k(exo)b (s-1
)
k(endo)b (s-1
)
ref
1
4
7
97:3
40:60
15:85
2.3 × 105
5.3 × 103
2.9 × 105
4.1 × 103
9.0 × 103
2.2 × 106
4
2
c
a 80 °C. b 25 °C. c This work.
for comparison are the corresponding data for the 5-hexenyl
and 5-methyl-5-hexenyl radicals 1 and 4. Inspection of the
data displayed in Table 1 reveals a number of intriguing
phenomena. First, the rate of endo cyclization of 7 is ca.
seven times faster than that for the exo mode. This provides
strong support for the predictions presented above for the
regiochemistry of ring closure of 7 and demonstrates the
dominance of the polar effect in its cyclization. Second, the
presence of the ester substituent is seen to have a spectacular
effect on the kinetics of cyclization. Thus, comparison of
the rates of endo ring closure for the three species 1, 4, and
7 shows that reaction of 7 occurs several orders of magnitude
faster than that of 1 or 4. Third, it is noteworthy that, unlike
the retarding influence of the methyl group on the rate of
strongly favor formation of 19 over 20 and, as observed
above in connection with the acyclic analogue 7, the polar
364
Org. Lett., Vol. 1, No. 3, 1999

factor is particularly effective. On these grounds it was
predicted that the bicyclic ester 17 rather than its isomer 23
would be the major product to be derived from treatment of
the selected precursor 21 with tributyltin hydride.
To test these hypotheses, synthesis of the precursor 21
was accomplished by treatment of the known7 amine 22 with
1,2-dibromomethane. Exposure of the derived salt 218 to
tributyltin hydride in 2-methyl-2-butanol solution under the
conditions referred to above afforded an 85% yield of
3-carbethoxy-1-azoniabicyclo[2.2.1]heptane bromide (17, R
) CH2CH3). 1H and 13C NMR spectral analysis of the crude
product revealed that it was essentially pure. None of the
isomeric species 23, nor the product of reduction 24, was
detected. The ester 17 (R ) CH2CH3) was obtained as a
10:90 mixture of the exo and endo isomers. The predomi-
nance of the endo ester is consistent with the expected exo
delivery of hydrogen by Bu3SnH at a C2 trigonal carbon in
the norbornyl system 19.
In conclusion, it is seen that ring closure of the 5-car-
bomethoxy-5-hexenyl radical 7 is influenced largely by polar
effects and endo closure predominates. When extended to
the 5-carbethoxy radical 18, it is found that cyclization occurs
with high stereoselectivity, in agreement with theoretical
predictions. In view of these observations, we suggest that
this transformation represents an excellent entry to the
important bicyclic ester 17.
(3) Microanalytical and spectral characterization of the half ester 12 and
all intermediate compounds involved in its synthesis from 2-carbomethoxy-
cyclohexanone will be presented in the full paper.
(4) For leading reference, see: Beckwith, A. L. J.; Schiesser, C. H.
Tetrahedron 1985, 41, 3925.
(5) (a) Della, E. W.; Smith, P. A. J. Org. Chem. 1999, 64, 2110. (b)
Della, E. W.; Knill, A. M.; Smith, P. A, J. Chem. Soc., Chem Commun.
1996, 1637. (c) Della, E. W.; Knill, A. M. Tetrahedron Lett. 1996, 37,
5805. (d) Della, E. W.; Knill, A. M. J. Org. Chem. 1996, 61, 7529.
(6) For leading reference, see: Jenkins, S. M.; Wadsworth, H. J.;
Bromidge, S.; Orlek, B. S.; Wyman, P. A.; Riley, G. J.; Hawkins, J. J.
Med. Chem. 1992, 35, 2392.
Acknowledgment. We thank the Australian Research
Council for financial support.
(7) Mattocks, A. R. J. Chem. Soc., Perkin Trans. 2 1978, 896.
(8) Full characterization of 21 and 17 will be presented in the full paper.
OL990557N
Org. Lett., Vol. 1, No. 3, 1999
365