In addition to carboxylic acids and carboxamides as chelating
groups, we have also prepared a novel class of imidazoleephenolate
precatalysts, with a representative example given in Scheme 5.
Benzylation of o-hydroxyacetophenone 21 with sodium hydride
and benzyl bromide proceeded smoothly to afford benzyl ether 22,
followed by a-bromination with NBS under solvent-free conditions.32
The resulting bromoketone 23 was treated with sodium
azide in DMF to afford azide 24. A protocol was developed that
enabled the azide reduction and isolation of the sensitive amino
ketone 25 cleanly as the HCl salt. After several low yielding and
messy Staudinger reduction reactions, the azide was cleanly hydrogenated
using Pd/C poisoned with diphenylsulfide, according to
the report by Sajiki.33 The reaction was performed in acidic methanol
to ensure that the resulting amine would immediately be
protonated to circumvent intermolecular reactions with the acetophenone.
The aminoketone 25 was then coupled with N-Boc-Lphenylalanine
to obtain amide 26. Heating with ammonium acetate
facilitated a cyclodehydration reaction to yield the desired
imidazole 27, which was globally deprotected with hydrogen and
Pd/C, followed by HCl, to provide the primary amine precatalyst
28a. The N-methyl phenylalanine-derived precatalyst 29 and the
proline-derived precatalyst 30 were synthesized in a similar manner
(see Supplementary data for details).the aldol reaction between propionaldehyde and 4-
nitrobenzaldehyde. The carboxamide-based precatalysts gave the
best combination of reactivity and enantioselectivity when used
with certain metal salts, including zinc triflate and lanthanide (III)
salts. The optimal catalysts were also selective for the anti stereoisomer,
with enantiomeric excess of up to 91% observed. One of the
optimal catalysts, 18ceZn(OTf)2, was studied with a series of controls
to support the hypothesis that it acts as a true hybrid catalyst to
promote the desired aldol reaction.
Despite these proof of concept results, our first-generation hybrid
catalysts at this stage offer no obvious benefits over simpler
organocatalysts for direct aldol reactions. Though the catalysts we
have identified are tolerant of air and moisture and are highly robust
(preliminary results show prolonged catalyst activity well beyond
24 h), the substrate scope is very limited, and they suffer the same
disadvantage of organocatalysts, namely sluggish reactions and/or
high catalyst loading. The relatively low catalyst activity is also
consistent with some previously reported hybrid catalysts for the
direct aldol reaction.24 We believe that alternative geometries for
hybrid systems may lead to more reactive catalysts, and computational
and synthetic efforts are underway to explore nextgeneration
hybrid systems with well-defined ligand/metal geometries
for the aldol and other carbonecarbon bond forming reactions