Angewandte
Communications
Chemie
ter, we questioned whether substrate activation through
C2 and C3 provided 3Aa with good diastereomeric ratios and
multiple hydrogen-bonding interactions might facilitate
deprotonation by a mild Brønsted base to give an a-keto
enolate with a specific configuration. Recently, we introduced
ureidopeptide-based Brønsted bases as a new subfamily of
ee values, but the transformations required prolonged reac-
tion times (entries 2–5). We were gratified to observe that
ureidopeptide-based Brønsted bases, which might display up
[18]
to three hydrogen-bonding interactions,
cross-aldol addition most effectively.
promoted this
[14]
organic catalysts bearing several hydrogen-bond donors.
Herein we report the utility of these newly developed
Brønsted bases by documenting the first direct catalytic
enantioselective cross-aldol reaction of a-keto amides with
As the results in Table 1 show, the catalysts C4 and C5
promoted complete conversion into 3Aa within 48 hours
(entries 6 and 7), whereas the catalyst C6 achieved a some-
what higher level of stereocontrol in a relatively shorter
reaction time (entry 8). Importantly, under these reaction
conditions self-condensation of either 1A or 2a were not
detected. Also, neither dehydrated nor lactonized aldol
products were observed. Further experiments revealed an
improvement in diastereoselectivity by increasing the aro-
matic character of the substituent in the a-keto amide. The
adduct 3Ba was produced, with increased diastereoselectivity
(90:10 d.r.), whilst maintaining the enantioselectivity (92%
ee), from the a-keto amide 1B (entry 9). In addition, the N-
methylated catalyst C7 behaved similarly to both C2 and C3,
and provided the aldol 3Aa in 30% conversion (d.r.75:25,
86% ee) after 48 hours.
[15]
either enolizable or non-enolizable aldehydes.
In contrast to the progress achieved in enamine based
aldol reactions, cross-aldol couplings mediated by Brønsted
[16]
bases appear to be more challenging to establish. Probably,
the most effective systems to date involve highly electrophilic
[17]
carbonyl acceptors such as 1,2-dicarbonyl compounds. We
initiated this work by evaluating several known Brønsted
bases for the reaction of the a-keto amide 1A and hydro-
cinnamaldehyde (2a; Table 1). The experiments soon
revealed that, indeed, the Brønsted base was the key for
success. For example, when using quinine, quinidine, and
(
DHQ) PYR, no aldol product (3Aa) was observed after
2
2
4 hours, either at À408C or 08C. While the squaramide C1
was also ineffective in terms of reactivity and selectivity
A representative selection of aldehydes was subjected to
the optimized reaction conditions to produce the aldol
adducts 3B for which diastereomeric ratios and enantiomeric
excesses were determined (Table 2). To avoid a-epimeriza-
(
entry 1), the bifunctional thiourea-tertiary amine catalysts
[
19]
[a]
tion during purification, each crude reaction mixture of the
products 3B was submitted to reduction with l-selectride.
The reduction proceeded cleanly at À788C and with essen-
tially complete stereoselectivity to give the corresponding
syn,syn 1,2,3-triols 4B in 60–72% yields upon isolation after
two steps. As the data in Table 2 show, results were
consistently good. Short alkyl chain aldehydes (e.g., propa-
nal), longer chain aldehydes (e.g., hexanal and heptanal), b-
branched isovareldehyde, and even aldehydes bearing side
chains with functional groups (e.g., alkene, ester, carbamate,
and ether) participate satisfactorily, thus giving enantiomeric
excesses of up to 96%. In contrast, diastereomeric ratios
seemed to decrease as the length of the alkyl chain in the
aldehyde increased (compare 4Bb, 4Bc, and 4Bd), and with
the presence of a-substitution (4Bj and 4Bk) while main-
taining high enantiomeric excesses. In addition, 3 mmol scale
reactions proceeded without any detrimental effect in the
reaction outcome (4Ba and 4Bf).
Table 1: Catalyst screening for the aldol reaction between 1A and 2a.
[
b]
[c]
[d]
Entry
1
Cat.
t [h]
Conv. [%]
d.r.
ee [%]
The relative and absolute configurations of the major
syn,syn enantiomer were determined by X-ray crystallo-
graphic analysis of 4Bd (Scheme 2a) and a uniform reaction
[
e]
1
2
3
4
5
6
7
8
9
1A
1A
1A
1A
1A
1A
1A
1A
1B
1A
C1
C2
C2
C3
C3
C4
C5
C6
C6
C7
48
48
96
48
96
48
48
36
36
48
30
40
60
60
80
>95
>95
>95
>95
>30
44:56
83:17
85:15
96:4
64
90
92
94
94
92
90
92
92
86
[20]
mechanism for the aldol reaction was assumed. Taking into
account the diastereo- and enantioselectivity observed, the
capacity of the ureidopeptide-based catalysts to mainly
produce syn-configured adducts might be consistent with
the generation, as a result of electrostatic and hydrogen-
bonding interactions, of a more stabilized Z enolate which
preferentially approaches the Si face of the aldehyde (Sche-
me 2b). Although we still have no evidence of the actual
96:4
75:25
79:21
86:14
90:10
75:25
1
0
[
a] Reactions conducted on a 0.2 mmol scale in 0.5 mL of CH Cl (mol
2 2
[21]
1
mode of substrate–catalyst interaction, the fact that reac-
tions with common Brønsted base catalysts, as well as with C7,
were significantly less efficient supports the beneficial effect
of multiple hydrogen-bonding interactions to boost reactivity.
ratio of 1A/2a, 1: 1.2). [b] Determined by H NMR spectroscopy.
[
[
1
c] Determined by H NMR spectroscopy and corroborated by HPLC.
d] The ee value of the major diastereomer determined by chiral-phase
HPLC. [e] 50% conversion after 96 h.
Angew. Chem. Int. Ed. 2016, 55, 3364 –3368
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3365