Angewandte
Communications
Chemie
Table 2: Substrate scope.[16]
provided a 10% yield of 3, but only after a prolonged reaction
time (entry 3). Several other salts were screened but AgOTf
proved optimal.[15,16] These data (no reaction without a Ag-
salt and entries 1–4) suggest that dicationic-PdII is required
and AgOTf likely improves the p-acidity.[15] Further experi-
ments on the solvent showed that molecular sieves diminished
the reactivity (entry 7) and, after an extensive solvent
screen,[16] it was found that use of distilled acetone with
0.5 equiv H2O struck an excellent balance between reactivity
and selectivity to provide 2a in 82% yield and 98% ee
(entry 9). These conditions were deemed optimal. At this
stage, two additional experiments should be mentioned.
Firstly, it should be noted that the corresponding trans-
substrate does not function in the reaction. Perhaps more
importantly, when non-racemic product 2a was subjected to
a Pd0-catalyst, rapidly racemization was observed, consistent
with our expectations.[16,17]
The substrate scope was explored to probe generality
(Table 2). A variety of substituted benzoic acids with varying
electronic and steric properties were competent in the
lactonization process, providing access to phthalides in good
to excellent yields and eeꢀs with minimal exception. Substrates
with both EWGs and EDGs functioned quite well to provide
the products in excellent ee (2b–g,i,l,m). In addition to small
scale experiments, the reaction was also carried out on the
gram scale with 2a and 1 mmol scale with 2l, and both were
formed in comparable yield and ee. This range of substrates is
particularly attractive since phthalides with EDGs are present
in most of the natural products of the family.[4e] Additionally,
it has been reported that EWGs on the phthalide enhance the
biological activity.[4f] It should also be noted that the method
doesnꢀt suffer from the same pKa requirements as the
Overman trichloroacetimidate method[6e] and more acidic
substrates like the salicylic acid derived-substrates leading to
2i and 2e functioned well in the reaction. Despite this, some
limitations are as follows: substrates with 5-substitution (2j),
a fused thiophene (2o), as well as a substituted olefin (2p)
were not tolerated. A substituted allyl alcohol was reactive,
and provided 2k in 88% yield, but a racemic substrate was
employed, and this led to racemic product signifying that a p-
allyl is not formed. It was also possible to form a 6-membered
lactone 2n, but further optimalization will be needed to
increase the selectivity.
[a] 1 g scale; [b] Solvent=distilled MeOAc (0.1 M) at rt; [c] Racemic
allylic alcohol substrate employed. [d] Reaction carried out at 1 mmol
scale. [e] Solvent=distilled acetone:TFE (4:1, 0.1 M); [f] Solvent=dis-
tilled MeOAc (0.1 M) at 08C.
In addition to phthalides, the scope was expanded to g-
lactones widely found in terpenes and alkaloids such as the
guaianolides and others.[4a,18] The reaction functions partic-
ularly well for fused a,b-unsaturated g-lactones (2q–s,u), all
exhibiting > 90% ee. Introduction of a CH3 at the a and b
positions of the aliphatic acid resulted in the nearly optically
pure (98% ee) g-butenolide 2w which is an important plant
germination signalling molecule.[19] In addition, 2x containing
a highly electrophilic exomethylene moiety was isolated in
78% and 90% ee, but this required additional optimization.
This is particularly significant since it is prone to racemiza-
tion.[20]
substrates described above have utilized a hydroxyl group as
a leaving group and we wondered how the reaction might
function with alternative, and perhaps better leaving
groups.[3a,21] To this end, substrates 1aa’–1ad’ were prepared
and subjected to the conditions. As can be seen in Table 3, and
as would be expected, these substrates completely suppressed
formation of 3 and 2a was produced exclusively, albeit with an
erosion of the enantioselectivity. The reactivity of alkoxyl
substrates 1aa’ and 1ab’ (entries 2,3) was decreased and
required 48 hours to proceed to completion but 1ac’ and 1ad’
showed only partial conversion after this prolonged reaction
time (entries 4,5). It should be noted that analyses of the
reaction mixtures show only product and unreacted starting
material, with no other by-products.
These data demonstrate that it is possible to identify
a catalyst system to replace a poor leaving group in an allylic
substrate and enantioselectively form a product containing an
ostensibly better leaving group without racemization. The
Angew. Chem. Int. Ed. 2021, 60, 1 – 7
ꢀ 2021 Wiley-VCH GmbH
3
These are not the final page numbers!