Having identified conditions suitable for the one-step
conversion of dienals to pyridines, efforts to determine the
scope of this synthetic methodology were undertaken. Thus,
a variety of diynol substrates of type 2 were constructed,
and the cycloisomerization-6π-cyclization sequence applied
(eq 3).
Table 2. Pyridines via the
Cycloisomerization-6π-Electrocyclization Sequence
The use of alkynes linked by heteroatoms was of particular
interest, as they can be potentially cleaved at a later stage,
allowing for further functionalization of the pyridine skeleton
at the 3 and 4 positions. We found that heteroatoms are
indeed tolerated (Table 2, entries 3, 4, and 5). The reactivity
of substrates without R-branching at the propargylic position
(R2 ) Me, Et) was also demonstrated (Table 2, entries 7-9).
As part of our effort to explore the scope of the cycloi-
somerization-6π-cyclization sequence, an examination of
the effect of ring size was undertaken. Tethers which
cycloisomerize to form 6-membered rings have been shown
to give slightly lower yields in the ruthenium-catalyzed
cycloisomerization than those which form 5-membered
rings.7 We were pleased to find that 2f (Table 1, entry 6)
underwent the cycloisomerization-6π-cyclization, in high
yield for both steps.
Previously, the only isolable products from the cyclization
of secondary propargylic alcohols had the E-olefin config-
uration at the γ,δ-double bond. It was postulated that
elimination of the hyroxyl group may occur stereoselectively,
where the observed E/Z ratio is dictated by steric interactions
between the R2 group and the cyclopentane ring.7 In a
number of the substrates examined, we observed for the first
time products containing the external olefin in a mixture of
E- and Z-olefin geometries. In some of the examples explored
herein the differential A-1,3 strain between the olefin
substituent and the annealed ring vs the ligands on Ru appear
to be more competitive because of olefin’s smaller size. The
6π-cyclizations of these substrates proceeded with complete
consumption of starting material (Table 2, entries 5-9), the
double bond mixtures having no noticeable effect on the yield
of the pyridine obtained in the subsequent 6π-cyclization.
Ketones are known to generate oximes as a mixture of
isomers, and we wondered whether this mixture would
impact the 6π-cycloisomerization reaction.11 Internal diynols
2d, 2e, 2i, and 2j were prepared and submitted to the
cycloisomerization reaction to give the desired ketones in
high yields (Table 2, entries 4, 5, 9, and 10). The formation
of pyridines from dienones 3d, 3e, and 3i took place readily,
forming the corresponding pyridines 6d, 6e, and 6i in good
yields. With use of the conditions of Table 1, entry 2, the
reaction of dienone 3j required higher temperatures and
longer times to provide the corresponding pyridine 6j, in a
reduced yield of 67%. The best yields of 6j were obtained
when the azatriene was first isolated, and then submitted to
the 6π-cyclization in trifluorotoluene, in which case an 85%
yield was obtained (Table 2, entry 10).
a See ref 7. b Methods: (i) NH2OH‚HCl, NaOAc, EtOH, 90 °C, 30 min,
(ii) PhCF3, microwave, 220-240 °C, 4 h. c Method: NH2OH‚HCl, NaOAc,
EtOH, microwave, 150 °C, 1-1.5 h. d Method: NH2OH‚HCl, NaOAc,
EtOH, 90 °C, 5-24 h. e E ) CO2Me.
(11) If one considers the disrotatory model for thermal 6π-electrocy-
clizations, only the E-isomer would position the hydrogen and the
N-hydroxyl group which are to be eliminated trans (anti-periplanar) to each
other in the resulting dihydropyridine. The Z-isomer might also introduce
steric strain that would prevent the substituents from attaining the required
planar confirmation.
We also examined substrates containing a terminal olefin
(Table 2, entry 4). As reported previously, primary alcohols
cycloisomerize to yield both the expected cycloisomerized
product 3d and the formally hydrated product 7. When dienal
Org. Lett., Vol. 9, No. 8, 2007
1475