A 3.6 Hz four-bond W-coupling constant is observed
between H4R and H6R. NOEs are observed between both
equatorial methyl groups and both H6R and H6â. In the major
azide 14a, a NOE is observed between H6R and the C5-Me
group and between H6â and the C1-Me group, indicating
that the methyl groups are trans. Oxidative cyclization of
159 proceeds more rapidly but gives only 31% (43% based
on recovered 15) of cyclic azide 16, which is isolated as the
enol tautomer.
Oxidative cyclization of 1710 provides an 87% yield of
an inseparable 9:1 mixture of 5-exo cyclization product 18
and 6-endo cyclization product 19, as expected.10 Hydroge-
nation7 (50 psi) of the crude mixture over 10% Pd/C in EtOH
provides bicyclic lactam 20 in 42% overall yield.
and halides that cyclization of tertiary radical 22 gives 23
•
as a 2:1 mixture of exo and endo CH2 radicals11 that will
lead to the observed 2:1 mixture of exo product 25a and
endo product 25b. Analysis of the 13C NMR spectra confirms
the stereochemical assignment.11 C8 and the C5-Me group
are shifted upfield from δ 48 and 25 in 25b to δ 43 and 21
in 25a due to shielding by the exo gauche CH2N3 group.
Similarly, C4 is shifted upfield from δ 41 in 25a to δ 33 in
25b due to shielding by the endo gauche CH2N3 group.
Initially, the reaction was run for 20 h at 50-55 °C as
with other cyclizations. Very little 24a is observed under
these reaction conditions. Cyclohexanone 24a, with the azide
and 2-propenyl group on the same face of the cyclohexane
ring, undergoes a thermal 1,3-dipolar cycloaddition to give
a triazoline that reacts further to give a complex mixture of
products.12 Pure 24a decomposes on heating for 12 h in
MeOH at 50-55 °C. Monocyclic azide 24b and bicyclic
azides 25a and 25b are stable to these conditions.
Mn(III)-based oxidative cyclization of 21 with Cu(OAc)2
yields only the alkene formed by oxidative elimination from
bicyclic radical 23. No monocyclic products are obtained
even though Cu(II) reacts with radicals with rate constants
of >106 sec-1.4 Since monocyclic azides 24 are the major
product, reaction of an azide anion with tertiary radical 22
must occur with a rate constant of close to 107 sec-1.
We then explored the effect of additives and other sources
of azide on the ratio of monocyclic products 24 to bicyclic
products 25, which is 65:35 in MeOH. To our surprise, a
91:9 mixture of 24 and 25 is formed with 5 equiv of TMSN3,
instead of NaN3, at 50 °C for 2 h. We hypothesized that
TMSN3 reacts with MeOH to give MeOTMS and HN3 and
that the increased percentage of 24 formed with TMSN3
might be due to the decrease in pH. Carrying out the
oxidative cyclization of 21 with 5 equiv of NaN3 and 5 equiv
of various acids and bases established that this was the case.
At low pH (ClCH2CO2H), the same 91:9 mixture is obtained
as with TMSN3. Formation of 24 is less favored with HOAc
(79:21), while bases NaHCO3 (55:45), NaOAc (48:52), and
Na2CO3 (48:52) favor the formation of bicyclic adducts 25.
Alkyl radicals are nucleophilic and may therefore react more
rapidly with HN3 than with azide anion.
Several intriguing features of this reaction emerge from
an examination of the tandem oxidative cyclization-azide
trapping of â-keto ester 21.11 Oxidative cyclization in MeOH
at 25 °C for 4 days yields monocyclic azides 24a (28%)
and 24b (11%) and bicyclic azides 25a (16%) and 25b (8%)
(Scheme 5). The stereochemistry of the monocyclic azides
Scheme 5
We were disappointed to find that oxidative cyclization
of 266 in MeOH containing 5 equiv of ClCH2CO2H affords
only 11% of azide 29 (Scheme 6). Even poorer results were
obtained without ClCH2CO2H. This result was quite surpris-
ing because oxidative cyclization of 26 in HOAc with
oxidative termination by Cu(OAc)2 affords 78% of 27.6
However, oxidative cyclization of 26 in MeOH with Cu-
(OAc)2 proceeds in poor yield, indicating that, for reasons
that are obscure, the use of HOAc as the solvent is important
for this cyclization. Oxidative cyclization in HOAc with 5
equiv of NaN3 affords a mixture of 24% of azide 28, 18%
of azide 29, and 17% of reduction product 30. The structure
of 30 was established by preparation of an authentic sample
by hydrogenation of 27. The stereochemistry of 28 and 29
24a and 24b was assigned by analogy to azides 14a and
14b. In 14b and 24b, the downfield H3 absorbs as a ddd (J
) 14, 14, 6) at δ 2.99-3.02, while the downfield H3 of 14a
and 24a absorbs as a multiplet at δ 2.6-2.7. We have
previously established by trapping radical 23 with hydrogen
(9) Snider, B. B.; Patricia, J. J. J. Org. Chem. 1989, 54, 38-46.
(10) Snider, B. B.; McCarthy, B. A. J. Org. Chem. 1993, 58, 6217-
6223.
(11) Dombroski, M. A.; Kates, S. A.; Snider, B. B. J. Am. Chem. Soc.
1990, 112, 2759-2767.
(12) (a) Laatsch, H.; Ernst, B. P.; Noltemeyer, M. Liebigs Ann. Chem.
1996, 731-737. (b) Eguchi, S.; Suzuki, T.; Toi, N.; Sasaki, T. Heterocycles
1990, 30, 237-240.
Org. Lett., Vol. 6, No. 8, 2004
1267