Angewandte
Chemie
synthesis,[6h,i] but ketones with a-proton pKa values of ca. 18
require amine activation by enamine formation. The relative
reactivities of these compounds allowed us to define this
functional pKa barrier.[1] As illustrated in Figure 1, the change
from an oxyphenol ester to a thiophenol ester results in a
reduction in the pKa value of the a-proton of an ester by
approximately 2 units, just at the borderline for nucleophile
activation using the currently available amine organocata-
lysts. Thus, we sought appropriate thioesters that might be
used to reduce the pKa value of the a-proton to a suitable
range for organocatalysis. In elegant experiments, Um and
Drueckhammer studied the kinetic acidities of the a-proton
of a series of thioesters.[7] These studies revealed that
trifluoroethyl thioesters have a-proton exchange rates in
toluene/triethylamine solutions that are approximately ten
times faster than phenyl thioesters, suggesting that the
pKa value of the a-proton of trifluoroethyl thioesters might
be close to those of malonate diesters, making them candidate
ester nucleophiles in catalytic direct asymmetric synthesis.
Since our 2000 report,[3c] a variety of pyrrolidine-based
iminium catalysts of Michael additions have broadened the
scope and efficiency of the iminium-based Michael reac-
tion.[3,6] As illustrated in Scheme 2, we screened a number of
these organocatalysts together with benzoic acid co-catalyst,
for catalysis of the Michael addition of thioester 1a to
cinnamaldehyde 2a. A number of the pyrrolidine-based
catalysts were effective. In particular, the trialkylsilyl-pro-
tected diarylprolinols (4–6) pioneered by Hayashi and co-
workers and Jørgensen and co-workers provided the product
in excellent chemical yield and with high enantioselectivity.[6]
With respect to our hypothesis concerning the pKa window
accessible for catalysis with these catalysts, the analogous
phenylthioester of 1a was unreactive, whereas the trifluoro-
ethyl thioester was reactive.
Catalyst 4 was the most promising of those evaluated with
respect to chemical yield and ee values; we therefore set out
to optimize this reaction. A variety of solvents and co-
catalysts were evaluated (see Table 1S in the Supporting
Information). Nonpolar solvents were ineffective in the
absence of a co-catalyst (Table 1S, entries 2–4); however,
the polar protic solvents, such as methanol and ethanol, were
highly effective and provided the desired product in 95% and
83% yield, respectively, on the basis of conversion after
48 hours at room temperature (Table 1S, entries 9 and 10).
Surprisingly, isopropanol provided less than 5% conversion.
We next studied a variety of acid co-catalysts in methanol. We
found that the co-catalysts benzoic acid, 5-methyl-1H-tetra-
zole, or water were equally effective, whereas trifluoroacetic
acid poisoned the reaction (Table 1S, entries 12–15).
To determine the scope of this reaction we synthesized a
variety trifluoroethyl thioesters derived from arylacetic acids,
and analyzed the products and the yields of the addition
reactions to a,b-unsaturated aldehydes (Table 1). We noted
significant electronic effects: faster reactions were observed
with thioesters derived from arylacetic acids having electron-
withdrawing groups, whereas reaction of thioesters derived
from arylacetic acids with electron-donating groups were
significantly slower (compare Table 1, entries 4 and 8). With
cinnamaldehyde as the Michael acceptor substrate, the yields
of the isolated products ranged from 46 to 88% and the
enantioselectivities were up to 98%. Although most of these
reactions were performed at room temperature, we were able
to demonstrate in the p-nitrophenylacetic acid ester case that
the ee value of the product could be improved from 66 to 91%
by simply performing the reaction at 08C (compare Table 1,
entries 4 and 6). With the highly reactive p-nitrophenylacetic
acid ester substrate, we observed that the reaction was
reversible during extended reaction times, ultimately eroding
the ee value of the product (compare Table 1, entries 4 and 5).
Other less active thioesters such as p-methoxyphenylacetic
acid and napthylacetic acid esters (Table 1, entries 8 and 9)
were studied by monitoring the reaction under extended
reaction times or with excess catalyst, and no erosion of the
ee values was noted. The trifluoroethyl thioester of propionic
acid was not reactive under these conditions, suggesting that
an aromatic functionality was required to bring the pKa value
of the a-proton down to a functional range.
Modification of the Michael acceptor substrate was also
possible and the o-methoxycinnamaldehyde acceptor pro-
vided product with 96% ee (Table 1, entry 12). Crotonalde-
hyde was significantly less effective, and Michael product 3k
was obtained with only 54% ee. It is known that catalyst 4
may be a poor catalyst for crotonaldehyde-based Michael
reactions.[6h] The overall diastereoselectivity of these reac-
tions was modest and the relative configurations of the
products were assigned after conversion of the phenylacetic
acid product (Table 1, entry 7) into a known product (see the
Scheme 2. Evaluation oforganocatalysts together with a benzoic acid
co-catalyst, of r catalysis ofthe Michael addition ofthioester 1a to
cinnamaldehyde 2a. The yields of 3a are shown below the respective
catalysts. [a] Reaction was performed in a 1.0m solution in the
presence of15 mol% of 7 and benzoic acid.
Angew. Chem. Int. Ed. 2008, 47, 4588 –4591ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim