All substrates underwent the Nazarov cyclization with
complete regioselectivity, usually placing the double bond
on the side of the alkoxy substituents. In the case of 9r,
however, the cyclization afforded 10r wherein the double
bond resides in the more substituted position.
Note that trienone 9f and terminally disubstituted substrates
9g,l reacted cleanly to afford vinyl cyclopentenone 10f and
products featuring quaternary stereocenters 10g,l, respec-
tively. Dihydrodioxine 9t gave compound 10t in good yield.
While many substrates underwent the reaction with
reasonable rates in dichloromethane, we found that aceto-
nitrile was a better choice for less reactive, sterically more
hindered substrates. This polar, coordinating solvent, which
is rarely used in combination with aluminum chloride, may
facilitate proton-transfer steps or catalyst turnover.
Figure 3. X-ray structure of compound 12.
Remarkably, 2-ethoxypentadienones 9p-s underwent the
reaction at a considerably slower rate than their dihydro-
pyranyl counterparts under otherwise identical conditions.10
The nature of the solvent was found to be less important in
these cases.
chloride, 9m cyclized to afford a 1:4 mixture of pyrano-
indenones 13a,b in good overall yield.
Due to their ability to engage in bidendate binding and
their high reactivity, substrates 9 should lend themselves well
toward catalytic asymmetric synthesis. So far, attempts to
develop a catalytic asymmetric version of the Nazarov
cyclization have only been moderately successful. For
instance, alkoxydienone 9j underwent Nazarov cyclization
in the presence of 20 mol % chiral scandium triflate pybox
complex 14 to afford enantiomerically enriched tricycle 10j
in 53% yield and 61% ee (Scheme 5).11 The enantiomeric
The diastereoselectivity of the cyclization was further
investigated using substrates 9h, 9i, and 9m (Scheme 4). The
Scheme 4
Scheme 5
excess was found to be very dependent on the solvent, with
THF providing the highest value. No attempts were made
to elucidate the absolute configuration of the major enanti-
omer. Note that 10j cannot easily undergo racemization,
which was found to be a persistent problem with other
substrates.
Although the enantiomeric excess is modest, this reaction
represents, to the best of our knowledge, the first catalytic
asymmetric Nazarov cyclization reported in the literature.9
In summary, the development of an efficient Nazarov
cyclization involving a highly reactive class of substrates
has been described. Its products include highly functionalized
heterobicyclic and -tricyclic compounds that could serve as
valuable synthetic building blocks.
4,5-disubstituted dienone 9h underwent fast cyclization with
the anticipated low diastereoselectivity to afford a 1.5:1
mixture of the cis and trans products 11a,b. By contrast,
substrate 9i gave exclusively cis diastereomer 12. The X-ray
crystal structure of 12 is shown in Figure 3.
Induced diastereoselection was investigated using com-
pound 9m, which could be obtained in two straightforward
steps from perillaldehyde and dihydropyran (see Supporting
Information). In the presence of 10 mol % aluminum
(9) Concomitant to our work, the Tius group has studied Nazarov
cyclizations of 2-alkoxy-1,4-penten-3-ones and catalytic asymmetric vari-
ants thereof. Tius, M. A.; Bee, C.; Leclerc, E. Org. Lett. 2003, 5, 4927-
4930. For a related system, see: He, W.; Sun, X.; Frontier, A. J. J. Am.
Chem. Soc. 2003, 125, 14278-14279.
(10) Compound 10p has been previously described: Katritzky, A. R.;
Zhang, G. F.; Jiang, J. L. J. Org. Chem. 1995, 60, 7605-7611.
(11) For the use of scandium pybox complexes in asymmetric synthesis,
see: (a) Evans, D. A.; Wu, J. J. Am. Chem. Soc. 2003, 125, 10162-10163.
(b) Evans, D. A.; Masse, C. E.; Wu, J. Org. Lett. 2002, 4, 3375-3378.
Org. Lett., Vol. 5, No. 26, 2003
4933