Communications
Table 1: Asymmetric [3+2] annulations of MBH adducts with isatin-
MBH adducts are tremendously useful intermediates in
derived activated alkenes.[a]
organic synthesis. In the past few years, Lu and co-workers
pioneered the use of the MBH adducts in [3+2] cyclizations
and various cycloadditions.[12] However, an asymmetric cyclo-
addition employing the MBH adducts is yet to be devel-
oped.[13] Compared with allenes and alkynes, the MBH
adducts are much more challenging substrates for such
annulations; their lower reactivity makes the development
of an asymmetric catalytic annulation process particularly
difficult. Given the high nucleophilicity observed for our
amino-acid-derived chiral phosphines,[8n,10] we reasoned that
our catalysts may be suitable for such an activation. Herein,
we document the first highly enantioselective [3+2] cyclo-
addition between the MBH carbonates and isatin-derived
tetrasubstituted activated alkenes, thereby creating 3-spiro-
cyclopentene-2-oxindoles containing two contiguous quater-
nary centers.
The [3+2] cycloaddition between isatin-derived a,a-
dicyanoalkene[14] 7 and the MBH carbonate 8 was selected
as a model reaction, and the catalytic effects of various amino-
acid-derived phosphines were examined (Table 1). To induce
effective stereochemical control, potential interactions of the
catalyst with the hydrogen-bond acceptor groups in the isatin
derivatives seem to be important, thus, a number of amino-
acid-based phosphines having various Brønsted acid moieties
were prepared. l-Valine-derived phosphine sulfonamide 1
showed poor catalytic activity (entry 1). In the presence of
valine-derived amide 2 or carbamate 3, the reaction pro-
ceeded smoothly, thus affording the desired cyclization
product in excellent yield, albeit with low stereoselectivity
(entries 2 and 3). Valine-derived phosphine thiourea 4
displayed much-improved catalytic effects, and the products
were obtained in excellent yield and with moderate enantio-
selectivity. However, the ratio of the a to g regioisomer
remained very poor (entry 4). With our previous success in
threonine-based catalytic systems,[9] we then focused on
threonine-derived phosphine thiourea catalysts. To our
delight, the enantioselectivity was significantly improved
when the threonine core was introduced, and only one
diastereomer was observed. Common sterically hindered
silyloxy groups[15] all proved to be effective, and the catalyst
5d having a TIPS group gave slightly better results (entries 5–
8). Replacement of the thiourea moiety in the catalyst with a
urea resulted in a faster reaction, but with slightly decreased
enantioselectivity (entry 9). Notably, the thiourea catalyst
containing a 3,5-di-trifluoromethylphenyl group was not more
effective than 4-fluorophenyl-derived thiourea, and the latter
was thus chosen as it is more economical (entry 5 versus 10).
To improve the regioselectivity of the reaction, we next
employed oxindoles with different N-alkyl groups. Gratify-
ingly, the introduction of a benzyl group on the oxindole
nitrogen atom led to substantially enhanced regioselectivity
(entry 11). When the N-para-methoxybenzyl (PMB) was
employed, the cyclization products were formed with an g
to a ratio of 14:1 (entry 12). The utilization of N-trityl
substrate 7d, however, resulted in very poor conversion,
probably as a result of the steric hindrance introduced by the
trityl group (entry 13). Variation of the ester groups in the
MBH adducts led to further improvement; ethyl ester proved
Entry Cat.
7
8
Solvent
t
Yield
g/a[c]
ee
[h] [%][b]
[%][d]
1
1
7a
7a
7a
7a
7a
7a
7a
7a
7a
7a
8a THF
8a THF
8a THF
8a THF
8a THF
8a THF
8a THF
8a THF
8a THF
8a THF
12
12
12
12
12
12
12
12
6
12
18
18
48
18
48
48
18
18
18
18
48
18
<30
90
92
94
93
89
83
93
92
89
86
84
<30
83
90
n.r.
92
89
70
–
–
2
2
1.5:1 10
3
3
1.5:1 16
1.5:1 35
2:1 71
2:1 70
2:1 69
2:1 74
2:1 68
2:1 70
13:1 75
14:1 77
4
4
5
6
7
8
5a
5b
5c
5d
5e
6
5d
5d
5d
5d
5d
5d
5d
5d
5d
5d
5d
5d
5d
5d
9
10
11
12
13
14
15
16
17[e]
18[e]
19[e]
20[e]
21[e]
22[e]
23[e]
24[e]
7b 8a THF
7c 8a THF
7d 8a THF
–
–
7c
7c
7c
7c
7c
7c
7c
7c
7c
7c
7c
8b THF
8c THF
8d THF
8c THF
8c
8c
8c
8c
8c
8c
8c
14:1 80
13:1 85
–
14:1 86
7:1 88
2:1 63
–
CH2Cl2
CH3CN
toluene
DMF
Et2O
CH3OH 48
CHCl3 18
90
trace
91
trace
93
>20:1 90
–
–
>20:1 85
–
–
13:1 96
[a] Reactions conditions: 7 (0.05 mmol), 8 (0.075 mmol), catalyst
(0.005 mmol), and solvent (0.5 mL) under Ar. [b] Yield of the isolated
product. [c] The g/a ratios were determined by H NMR analysis of the
crude products. [d] The ee value of the major regioisomer was deter-
mined by HPLC analysis using a chiral stationary phase. [e] 3 ꢀ molecular
sieves (30 mg) were added. Boc=tert-butoxycarbonyl, DMF=N,N’-
dimethylformamide, n.r.=no reaction, TBDPS=tert-butyldiphenylsilyl,
TBS=tert-butyldimethylsilyl, TDS=thexyldimethylsilyl, THF=tetrahy-
drofuran, TIPS=triisopropylsilyl, Trt=trityl.
1
to be superior to the methyl ester (entry 14), and the
utilization of the MBH adduct with the tert-butyl ester further
increased the ee value of the cyclization product to 85%
(entry 15). The carbonate group in the MBH adducts is
important for the observed reactivity, since the corresponding
MBH acetate was found to be inactive in the cycloaddition
(entry 16). The addition of 3 ꢀ molecular sieves to the
7838
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 7837 –7841