Full Paper
transformations. To implement our assumption, the disulfoni-
mide C2 has been employed in the model reaction (Table 1,
entry 3). To our delight, the desired product 3a was obtained
in almost quantitative yield with a significant increase in enan-
tioselectivity (95% yield, 56% ee). Encouraged by this proof-of-
principle result, a systematic exploration of different reaction
conditions was conducted. The subsequent solvent screening
revealed toluene to be the optimal choice (see the Supporting
Information). The rate of the uncatalyzed racemic reaction was
also found to be moderate (ꢀ25% yield in 48 h), indicating
the influence of the background reaction in diminishing the
enantioselectivity through the catalytic pathway. To prevent
this situation, we turned our attention to decreasing the reac-
tion temperature to 08C; although the chemical yield dropped
to 63% with poor enantioselectivity (17% ee, Table 1, entry 4).
Screening of other disulfonimides (Table 1, entries 5–7) showed
that the use of C3 was encouraging as the selectivity was ele-
vated to 58%. At this juncture, two principle parameters could
be taken into account to improve the enantioselectivity:
a) steric and electronic remodulation of the catalyst and b) re-
structuring the ketimine substrate by installing other protect-
ing groups.
Under the optimized conditions, the scope of the enantiose-
lective Mukaiyama–Mannich reaction of the N-trityl-protected
isatin-derived N-Boc ketimines catalyzed by disulfonimide C2
was next investigated (Scheme 2). Ketimines possessing both
electron-rich and electron-withdrawing substituents at the 5, 6,
and 7 positions of the indolinone ring were examined in this
newly developed protocol. The effect of various alkoxy sub-
stituents on the silyl ketene acetals was also considered. Re-
markably, all the reactions proceeded smoothly and the de-
sired 2-oxoindolinyl-b3,3-amino esters 3c–3x were obtained in
excellent yields with outstanding enantiopurities. In general, all
the reactions were very clean, energy-efficient, and could
easily be followed by visual color check of the reaction mix-
ture. Notably, with a tert-butyl substituent on the silyl ketene
acetal, the reaction was not successful at all because of the
steric hindrance. It is also worth noting that the tetrasubstitut-
ed silyl ketene acetal derived from methyl iso-butyrate does
not respond to the reaction with the ketamine 1c under the
optimized conditions or even at higher temperature (up to
808C).
To further demonstrate the practicability and robustness of
the developed protocol, a gram-scale synthesis was conducted
with 2.0 g (4.124 mmol) of isatin-derived ketimine 1c and com-
mercially available silyl ketene acetal 2b in the presence of
1000 ppm (0.10 mol%) of disulfonimide catalyst C2 (Table 2,
entry 1). We are pleased to report that the reaction proceeded
smoothly and led to the final product 3l in excellent yield with
outstanding enantioselectivity (96% yield, >99% ee). It is note-
worthy that further reduction of the catalyst loading to
250 ppm (0.025 mol%) could still successfully catalyze the Mu-
kaiyama–Mannich reaction of ketimine 1c (500 mg batch,
1.031 mmol) in full conversion with good enantioselection
albeit with longer reaction time (5 days; Table 2, entries 2–3).
To the best of our knowledge, this is the lowest catalyst load-
ing achieved to date for the Mukaiyama–Mannich reaction
with excellent enantioselectivity. This chiral disulfonimide cata-
lyzed Mukaiyama acetate Mannich reaction followed the analo-
gous ACDC reaction mechanism reported in the literature.[13b]
To demonstrate the utility of the newly developed enantio-
selective Mukaiyama–Mannich protocol, we pursued the
formal synthesis of (+)-AG-041R, a potential cholecystokinin-B/
gastrin receptor antagonist (Scheme 3). Starting from the enan-
tiopure building block 3l, the single-step deprotection of both
the Boc and trityl groups by the employment of TFA in di-
We envisioned that reconstructing the ketimine system
could be the superior choice as the catalyst (C2) is readily
available in addition to its potency in ACDC catalysis. Pleasing-
ly, our speculation was in harmony with the following experi-
mental results. When ketimine 1c, containing a triphenylmeth-
yl (trityl=Tr) group, was employed as the substrate, the de-
sired product 3c was obtained in excellent yield with out-
standing enantioselectivity (95% yield, >99.5% ee) within
5 min at room temperature (Table 1, entry 11). Catalyst C2 was
deemed to be the optimal choice owing to its enhanced reac-
tivity compared with catalysts C1 and C3 (Table 1, entries 10
and 12). The disulfonimide catalyst C2 was also found to be su-
perior to catalyst C1 when applying the ketimine 1a (Table 1,
entries 2 and 3), but the catalyst C1 was suited better for the
ketamine 1b in terms of enantioselectivity although the reac-
tion was faster with catalyst C2 (Table 1, entries 8 and 9). The
concomitant steric effect of the N-trityl functionality was a
clear advantage for achieving high enantioselectivity. The re-
sults also led us to consider the possibility of additional non-
covalent interactions between the catalyst (C2) and the trityl
group of the imine substrate during the reaction.
Although organocatalysis has already been recognized as
the third pillar in modern asymmetric catalysis toward the total
synthesis of natural and pharmaceutical products, one criticism
is that organocatalytic processes commonly require a large cat-
alyst loading in comparison with the metal-catalyzed pathways
to furnish the reactions useful. It is noteworthy that very few
enantioselective organocatalytic CÀC bond-forming reactions
have been reported as using ultra-low catalyst loadings
(<0.5 mol%) or even with ppm and sub-ppm levels of load-
ing.[14] Fortunately, we were able to determine that only
0.25 mol% catalyst loading was sufficient to afford the desired
2-oxoindolinyl-b3,3-amino ester 3c in excellent yield while
maintaining the high enantioselectivity (Table 1, entries 13–14,
for details please see the Supporting Information).
chloromethane afforded the N-unprotected 2-oxoindolinyl-b3,3
-
amino ester 7. The compound 7 could be converted into
(+)-AG-041R by following the reported literature procedure,
thereby accomplishing the formal synthesis of (+)-AG-041R.[10]
The obtained 2-oxoindolinyl-b3,3-amino ester 7 could also serve
as a precursor for the synthesis of spiro-b-lactam 11, a skeleton
of the natural product chartellines. The absolute configuration
of the compound 3l was determined as the (R)-configuration
by correlating the sign of the specific rotation of 7 with the lit-
erature data (for details see the Supporting Information).[10]
The absolute configuration of the other compounds was tenta-
tively assigned by analogy.
Chem. Eur. J. 2019, 25, 1 – 7
3
ꢀ 2019 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ