Key words
asymmetric synthesis - Brønsted acid catalysis - electrophilic amination - α-branched
ketones - enol catalysis - azodicarboxylates
Enantiomerically pure α-amino ketones and α-amino alcohols are important substructures
of natural products and pharmaceuticals.[1] For example, ketamine is used as an anesthetic and currently under investigation
for the treatment of depression.[2] Recent studies revealed that its (S)-isomer is four times more active than its enantiomer, making new asymmetric methodologies
for the synthesis of ketamine even more relevant.[3] Recently, catalytic asymmetric electrophilic amination reactions of carbonyl compounds
have been developed. The reported protocols employ azodicarboxylates, N-hydroxycarbamates, and nitrosobenzene as amination reagents.[4] Several activation modes have successfully been applied and typically take advantage
of highly reactive starting materials, such as β-keto esters, α-cyano carbonyls, α-fluorinated
ketones, nitroacetates, or oxindoles.[4]
[5] The direct amination of unactivated carbonyl compounds has been successfully achieved
with both aldehydes and ketones by enamine catalysis.[6] However, when α-branched ketones are employed, chiral amine catalysts give poor
results because of the sterically constrained enamine intermediate. Furthermore, aminocatalysis
preferentially forms the kinetic enamine, thus restricting access to valuable enantioenriched
ketones bearing a quaternary stereocenter. One of the rare examples which afford these
compounds from unactivated α-alkyl ketones was reported by Terada and co-workers by
employing tetralone derivatives as substrates and a chiral organosuperbase as catalyst.[7]
Recently, our group proposed a solution to this limitation of aminocatalysis by shifting
to enol catalysis with the development of a chiral phosphoric acid catalyzed asymmetric
Michael reaction of α-branched ketones with enones.[8] This approach directly suggests the use of various other electrophiles (X=Y; Scheme
[1]). Brønsted acid catalyzed α-aminations of aromatic nucleophiles have been reported
but, the direct amination of α-branched ketones has been completely unknown.[9]
[10] Here, we report a chiral phosphoric acid catalyzed α-amination of α-branched cyclic
ketones using azodicarboxylates as the electrophilic nitrogen source.[11]
Scheme 1 Proposed catalytic cycle for the α-functionalization of ketones by enol catalysis
Table 1 Optimization of the Brønsted Acid Catalyzed α-Amination of α-Branched Ketonesa
|
|
Entry
|
Catalyst
|
Solvent
|
Conversionb
|
erc
|
|
1
|
4a
|
CH2Cl2 (0.5 M)
|
21%
|
94:6
|
|
2
|
4b
|
CH2Cl2 (0.5 M)
|
39%
|
93:7
|
|
3
|
4c
|
CH2Cl2 (0.5 M)
|
96%
|
98:2
|
|
4
|
4c
|
MeCN (1.0 M)
|
93%
|
98.5:1.5
|
|
5d,e
|
4c
|
MeCN (2.0 M)
|
(85%)
|
99:1
|
a Reaction conditions: 1a (0.05 mmol), 2 (0.1 mmol), 4 (5 mol%), solvent (0.1 mL), r.t., 24 h.
b Determined by 1H NMR using triphenylmethane as an internal standard, with isolated yield given in
parentheses.
c Determined by HPLC on a chiral stationary phase.
e Amount of 2 used was 0.06 mmol.
f Reaction was carried out on a 0.2-mmol scale.
Scheme 2 Substrate scope of the Brønsted acid catalyzed asymmetric α-amination of α-branched
ketones. Reaction conditions: 1 (0.2 mmol), 2 (0.24 mmol), 4c (5 mol%), MeCN (0.1 mL), r.t. or 40 °C, 24–60 h. a Reaction was carried out on a 1.0 mmol scale.
We began our studies by employing 2-methyl cyclohexanone (1a) as the model substrate and dibenzyl azodicarboxylate (DBAD; 2a) as the electrophile (Table [1]). When (S)-TRIP (4a) was used as catalyst, 3a was obtained in excellent enantioselectivity; however. only poor conversion was observed
(Table [1], entry 1). After screening several phosphoric acids bearing different substituents
in the 3,3′-positions, catalyst 4c proved to be superior, in terms of both reactivity and selectivity (Table [1], entry 3). Switching the solvent from dichloromethane to acetonitrile and raising
the concentration of the substrate to 1 M led to a slight increase of the enantiomeric
ratio and a beneficial effect on the conversion (Table 1, entry 4). The optimized
reaction conditions, found by further increasing the concentration to 2 M and decreasing
the amount of 2a to 1.2 equivalents, afforded the desired product 3a in 85% isolated yield and an enantiomeric ratio of 99:1 (Table [1], entry 5). Interestingly, decreasing the catalyst loading to 1 mol% had no influence
on enantioselectivity, but lower conversions were obtained. When neat conditions were
employed, a strong increase of reactivity was observed, albeit the enantioselectivity
was diminished (for the full optimization, see the Supporting Information).
With the optimal reaction conditions in hand, we focused our attention on the scope
of the Brønsted acid catalyzed α-amination of α-branched ketones (Scheme [2]). The reaction proved to be rather general; in addition to DBAD (2a), DEAD (2b) and DIAD (2c) could also be used as electrophilic aminating reagents, giving the corresponding
α-hydrazino ketones in moderate to good yields and excellent enantioselectivities
(3a–c in Scheme [2]). To our delight, a large variety of substituents in the 2-position were tolerated
under the reaction conditions, affording in all cases the desired products in good
yields and excellent enantioselectivities (3d–i). Interestingly, when cyclohexanone itself was used as a starting material, the monoaminated
product 3d was obtained, and no racemization occurred under the reaction conditions, further
supporting the generality of enol catalysis. 2-Alkyl-substituted cyclopentanones also
reacted smoothly (3j,k). When the ring size of the cyclic ketones was increased, a lower reactivities and
enantioselectivities were observed, presumably due to a slow and hindered enolization
(3l,m). The scalability of the reaction was proven by reacting 2-phenylcyclohexanone (1c) and DEAD (2b) on a 1.0 mmol scale to give product 3n in 77% yield and an enantiomeric ratio of 97.5:2.5. 1-Indanone- and 1-tetralone-derived
substrates represent a current limitation of the methodology as only low reactivity
and enantioselectivity were observed (3o and 3p). The absolute configuration of the products was assigned based on known compound
3d.[12]
To further prove the utility of our transformation, preliminary attempts to cleave
the N–N bond by employing a slightly modified literature procedure[13] afforded the desired product 6 in 35% yield and without any loss of enantiopurity (Scheme [3]).
Scheme 3 Application of the Brønsted acid catalyzed α-amination of α-branched ketones to the
synthesis of carbamate-protected α-amino ketones
In summary, we have developed an asymmetric α-amination of α-branched cyclic ketones
by enol catalysis.[14] Employing a chiral phosphoric acid as the catalyst and various azodicarboxylates,
the desired carbamate-protected cyclic α-hydrazino ketones were obtained in good to
excellent yields (40–99%) and enantioselectivities (enantiomeric ratios from 60:40
to 99:1). The feasibility of the N–N bond cleavage under mild redox neutral conditions
was also proven. The current protocol expands the scope of enol catalysis, thus further
opening new reaction pathways in asymmetric catalysis.
