Enantiopure Benzofused Cyclic Ethers
COMMUNICATION
signments of these new compounds were based on a series
of NMR studies. Additionally, the structures of compounds
3n, 3o, and 3q were confirmed by single-crystal X-ray struc-
ture analysis.[10] Surprisingly, we observed the opposite chiral
induction when starting from alkynols 1l,m, derived from d-
lactate, or from alkynols 1n–r, derived from l-amino acids.
A tentative mechanism for the formation of compounds 3,
based on a tandem sequence involving a 6-exo-cycloisomeri-
zation reaction followed by an intramolecular hydroaryla-
tion process, is presented in Scheme 2. Thus, the reaction is
room temperature. After 1 h, complete conversion to 3d
was observed.
An intriguing question related to the mechanism of this
process is the identity of the catalytic species responsible for
the two catalytic cycles proposed in Scheme 2. In principle,
both Lewis and Brønsted acids could catalyze the hydroal-
koxylation and hydroarylation reactions. Traces of Brønsted
acids could be present in the reaction media coming from
the platinum or gold complexes (or the solvent), and these
protic acids could be the responsible for both or at least one
of the catalytic cycles proposed in Scheme 2.[11] To shed light
on this issue, the alkynol 1d was subjected to a range of re-
action conditions. No reaction was observed in the absence
of platinum or gold complexes or by treatment of 1d with
Brønsted acids, which demonstrates the necessity of the
platinum or gold complexes at least for the hydroalkoxyla-
tion reaction. To eliminate the possibility of competing
Brønsted acid catalysis in the second catalytic cycle (the hy-
droarylation reaction), we performed some control experi-
ments by using a non-poisoning base. Thus, when compound
1d was allowed to react with 5 mol% PtCl4 in a dichlorome-
thane solution containing 2.5 mol% of the strong phosphor-
ine base BEMP (2-tert-butylimino-2-diethylamino-1,3-di-
methyl-perhydro-1,3,2-diazaphosphorine), we observed the
formation of the expected product 3d after 1 h at room tem-
perature. Additionally, we performed an experiment in
which we treated 1d with 5 mol% PtCl4 in dichloromethane
at À208C and once we observed the complete conversion
(by GC-MS
) to the intermediate iso-2d, we added
2.5 mol% BEMP. The solution was warmed to room tem-
perature, and after 1 h complete conversion to the final
product 3d was observed. All these results suggest that re-
sidual Brønsted acids are not responsible, because these
would be quenched by BEMP (at 2.5 mol%). Accordingly,
we believe that the gold or platinum catalytic species are in-
volved in both catalytic cycles proposed in Scheme 2.[12]
To verify the stereochemistry observed in the products de-
rived from d-lactate and l-amino acids (see Table 2), we fo-
cused on the structure of the corresponding intermediate 6.
Thus, for d-lactate derived products 3l,m we suppose that
the reactions proceed via the intermediates 6l,m
(Scheme 2). As shown, in this structure the methyl group is
placed in a pseudo-equatorial position. However, to explain
the formation of compounds 3n–r, derived from l-amino
acids we suppose that the reactions proceed via intermedi-
ates 6n–r, in which the R group is placed in a pseudo-axial
position.[13]
In summary, we have developed a highly efficient and
general method for the diastereoselective synthesis of
benzo-fused eight-membered carbo- and heterocycles. The
method is based on a new tandem gold- or platinum-cata-
lyzed hydroalkoxylation–hydroarylation reaction. The
tandem sequence described allows the straightforward and
efficient synthesis of complex final products. The number of
points of diversity present in the reaction products, together
with the simplicity of the starting materials, makes this
tandem reaction a powerful method for both library genera-
Scheme 2. Proposed mechanism for the formation of benzofused bicyclo-
ACHTREUNG[3.3.1]nonanes 3 from alkynols 1.
initiated by coordination of the metallic complex to the
triple bond of the starting alkynol 1 to form intermediate 4.
Intramolecular addition of the hydroxy group to the internal
carbon of the triple bond generates 5. Protodemetalation of
the latter affords the enol ether 2 and releases the catalytic
species. Once enol ether 2 is formed, it enters the second
catalytic cycle. Thus, after an initial coordination of the cata-
lyst to the double bond of the enol ether 2, the oxonium in-
termediate 6 is formed. Further nucleophilic attackof the
phenyl group affords the intermediate 7. Interestingly, the
stereochemistry of all new stereocenters is fixed by the axial
position of the benzyl group required for the cyclization
step. Finally, from 7, a rearomatization and a protodemeta-
lation step lead to the final product 3, regenerating the cata-
lytic species.
The role of enol ethers 2 as intermediates of the reaction
was supported by the isolation of enol ether iso-2d (see
Scheme 2) when an experiment was performed with alkynol
1d in dichloromethane at À208C and using 5 mol% PtCl4 as
catalyst. To corroborate the formation of 3d from iso-2d we
treated the latter with 5 mol% PtCl4 in dichloromethane at
Chem. Eur. J. 2008, 14, 4153 – 4156
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4155