Dedicated to the memory of Prof. Carlos F. Barbas III
The construction of chiral quaternary stereocenters bearing an amine moiety represents
an important reaction in synthetic organic chemistry due to the range of compounds
possessing such a structure in nature, most of them having biological and pharmaceutical
activity.[1 ] In this sense, a wide variety of methods has been developed to gain access to these
motifs. Among them, the asymmetric electrophilic amination of prochiral carbonyl compounds
employing diazocarboxylates as nitrogen source is a simple and straightforward method
since the latter reagents are bench-stable and readily available.[2 ]
Particularly interesting is the catalytic asymmetric α-amination of prochiral 1,3-dicarbonyl
compounds, since the highly functionalized resulting structures can be further transformed
and elaborated.[1 ]
[2 ]
[3 ] In this regard, several strategies have recently been developed to accomplish this
transformation.[3 ] Thus, since the pioneer work of Jørgensen and co-workers using a copper(II)-box
catalytic system[4 ] different methods, not only metal-catalyzed,[5 ] but also employing organocatalysts[6 ]
[7 ]
[8 ] have been reported.
Table 1 Catalyst Screeninga
Entry
Catalyst
Conversion (%)b
ee (%)c
1
I
>95
68
2
II
>95
90
3
III
>95
75
4
IV
>95
85
5
V
90
10
6
VI
80
rac
7
VII
90
50
8
VIII
15
rac
9
IX
70
65
10
X
<15
n.d.
a Reaction conditions: 1a (0.10 mmol), 2a (0.15 mmol), catalyst (10 mol%) in toluene (1 mL).
b Determined by 1 H NMR analysis from the reaction crude.
c Determined by chiral HPLC (Daicel Chiralpak IA, see Supporting Information for details).
Recently, we have been interested in the use of trans -cyclohexanediamine benzimidazole derivatives as hydrogen-bonding organocatalysts
in various organic transformations.[9 ] Therefore, we decided to explore the performance of these catalysts in the electrophilic
amination of 1,3-dicarbonyl compounds. The results of this study are disclosed herein.
First, the search for the appropriate catalyst to carry out this reaction was tackled
using ethyl 2-oxocyclopentanecarboxylate (1a ) and di-tert -butylazodicarboxylate (2a ) as model substrates (Table [1 ]) and different trans -cyclohexanediamine benzimidazole derivatives I –VIII (Table [1 ], entries 1–8). The more basic catalysts I –IV afforded the corresponding amination product 3aa in high conversions and enantioselectivities (Table [1 ], entries 1–4), reaching up to 90% ee in the case of dimethylamino derivative II (Table [1 ], entry 2). The presence of less basic nitrogen in the catalysts, as is the case
of V and VI , resulted in a dramatic drop of enantioselection (Table [1 ], entries 5 and 6). Next, bis(2-aminobenzoimidazole) derivatives VII and VIII were also evaluated, but poorer results were observed in both conversion and enantioselectivity
(Table [1 ], entries 7 and 8). Finally, for the sake of comparison, Takemoto’s thiourea catalyst
IX and the bisthiourea X were also evaluated but moderate conversion and enantioselectivity and low conversion
were observed, respectively (Table [1 ], entries 9 and 10)
Once the organocatalyst screening revealed that benzimidazole II provided the best results, further optimization of reaction conditions was performed
(Table [2 ]). Firstly, different solvents were tested (Table [2 ], entries 1–7) obtaining the best results in terms of both conversion and enantioselectivity
when toluene, diethyl ether and hexane were employed (Table [2 ], entries 1, 3 and 6). With these solvents the temperature influence was evaluated.
Thus, at 0 °C, the same results were observed (Table [2 ], entries 8–10) and lowering the temperature to –20 °C resulted in lower conversions
with enantioselectivity remaining the same. At this point, and since the influence
of the temperature was negligible, we decided to continue the optimization at room
temperature, using diethyl ether as solvent for solubility reasons. For substrate
efficiency, we carried out the reaction using 1.05 equivalents of 2a and the same results were observed (Table [2 ], entry 11). Next, the effect of concentration of 1a was studied, and exactly the same results were obtained using 0.2 M and 0.05 M reaction
solutions (Table [2 ], entries 12 and 13); therefore we chose the latter as optimal concentration. Finally,
we tried to reduce the amount of catalyst and we observed that not only 5 mol% (Table
[2 ], entry 14), but also as low as 1 mol% of catalyst loading was enough to promote
the reaction with full conversion and excellent enantioselectivity (Table [2 ], entry 15).
Table 2 Optimization of Reaction Parametersa
Entry
Solvent
Temp (°C)
Conv. (%)b
ee (%)c
1
toluene
25
>95
90
2
CH2 Cl2
25
45
90
3
Et2 O
25
>95
92
4
THF
25
90
86
5
TBME
25
95
87
6
hexane
25
>95
92
7
MeOH
25
90
52
8
toluene
0
>95
90
9
Et2 O
0
>95
92
10
hexane
0
95
92
11d
Et2 O
25
>95
92
12d,e
Et2 O
25
>95
92
13d,f
Et2 O
25
>95
92
14d,e,g
Et2 O
25
>95
92
15d,e,h
Et2 O
25
>95
92
a Reaction conditions: 1a (0.10 mmol), 2a (0.15 mmol), II (10 mol%) in solvent (1.0 mL).
b Determined by 1 H NMR analysis from the reaction crude.
c Determined by chiral HPLC (Daicel Chiralpak IA, see Supporting Information for details).
d Amount of 2a used was 0.105 mmol (1.05 equiv).
e Volume of Et2 O ([1a ] = 0.2 M) used was 0.5 mL.
f Volume of Et2 O ([1a ] = 0.05 M) used was 2 mL.
g Conditions: 5 mol% of II , 0.5 mL of Et2 O.
h Conditions:1 mol% of II , 0.5 mL of Et2 O.
Then, with the optimal reaction conditions established (Table [2 ], entry 15), we decided to study the influence of the diazocarboxylate structure
(Scheme [1 ]). Thus, β-keto ester 1a was allowed to react with different alkyl diazocarboxylates 2b –d but in all cases the results turned out to be worse than that in the case of 2a .
Scheme 1 Study of different diazocarboxylates
With the reaction parameters optimized we next explored substrate scope (Table [3 ]).[10 ] First, cyclic β-keto esters were examined. As previously noted 1a yielded the desired product in high yields and with 92% ee (Table [3 ], entry 1). Surprisingly, when the six-membered analogue was submitted to the optimal
reaction conditions it failed completely, even when higher catalyst loadings were
investigated (Table [3 ], entry 2). The use of benzocondensed substrate 1c rendered the amination product 3ca in high yield and moderate enantioselectivity (Table [3 ], entry 3). In contrast, high optical purity along with high yield were obtained
with keto ester 1d (Table [3 ], entry 4). Cyclic β-amido ester was also examined but a disappointingly low enantioselectivity
was obtained despite several reaction conditions tested (Table [3 ], entry 5).
Next, the more reactive cyclic 1,3-diketones were considered. The five-membered ring
diketone 1f was firstly tested obtaining good yield and moderate enantioselectivity (Table [3 ], entry 6). In this case, the yield was slightly increased by using 5 mol% catalyst
loading. As already observed in the case of keto esters, the six-membered 1,3-diketone
1g afforded low conversions, regardless of the reaction conditions tested (Table [3 ], entry 7). The corresponding benzocondensed analogues 1h and 1i were also evaluated and, in both cases, gave high yields although moderate enantioselectivities
for the corresponding amination products were achieved (Table [3 ], entry 7). In both cases a slight increase of the optical purity was observed by
lowering the temperature. Finally, compounds 1j and 1k gave rise to the corresponding amination products, 3ja and 3ka respectively, in good yields but with poor ee values.
Table 3 Substrate Scopea
Entry
1
3
Yield (%)b
ee (%)c
1
1a
3aa
99
92
2
1b
3ba
<10
n.d.
3
1c
3ca
97
45
4
1d
3da
98
88
5
1e
3ea
87
20
6
1f
3fa
75 (82)d
50
7
1g
3ga
<10
n.d.
8
1h
3ha
88 (70)e
27 (35)e
9
1i
3ia
70 (68)f
48 (54)f
10
1j
3ja
66
26
11
1k
3ka
89
25
a Unless otherwise stated, the reaction conditions were: 1a (0.20 mmol), 2a (0.21 mmol), II (1 mol%) in Et2 O (1 mL), 25 °C.
b Isolated yield after column chromatography.
c Determined by chiral HPLC (see Supporting Information for details).
d The reaction was carried out using 5 mol% of II .
e The reaction was carried out at –20 °C.
f The reaction was carried out at –50 °C.
Different linear β-keto ester and 1,3-diketones were also evaluated but, despite our
efforts, racemic mixtures were obtained in all the cases.
Regarding the reaction mechanism, and based on previous computational and experimental
studies carried out by our research group employing identical catalysts for the asymmetric
conjugate addition of 1,3-dicarbonyl compounds onto nitroalkenes,[9a ] we propose the catalytic cycle depicted in Scheme [2 ] in which benzimidazole II can act as a bifunctional organocatalyst. Thus, II can act initially as a base, forming the corresponding 1,3-dicarbonyl compound enolate,
that can coordinate through hydrogen-bonding to the catalyst, as depicted in intermediate
A . Then, the protonated dimethylamino moiety can activate the diazocarboxylate and
hence facilitate the enantioselective attack of the enolate (intermediate B), releasing
the corresponding amination product and regenerating the organocatalyst II .
Scheme 2 Proposed catalytic cycle
It is worthy of note that the S -configured amination product seems to be obtained when (R ,R )-II is employed. This assumption was taken from a specific rotation comparison between
product 3aa and the values reported in literature.[11 ]
In conclusion, we have demonstrated that chiral trans -cyclohexanediamine benzimidazole derivative II is a suitable and effective organocatalyst for the asymmetric electrophilic amination
of cyclic 1,3-dicarbonyl compounds. The corresponding amination products are obtained
in the majority of the cases with high yields and moderate to high enantioselectivities
using just 1 mol% of catalyst loading. In addition, a bifunctional role of the catalyst
is assumed due to a presumably dual hydrogen-bond activation of both the 1,3-dicarbonyl
compound and the diazocarboxylate.
