COMMUNICATION
unsaturated a-ketoester 2 generates acyclic intermediate A;
2) Subsequent intramolecular nucleophilic addition of the
enolate oxygen of A to the C5 atom of the azlactone subunit
results in the formation of intermediate B as a formal [4+2]
cycloaddition product; and 3) The consecutive process is ter-
minated by ring-opening of the azlactone moiety, to give an
a-amino-d-lactone derivative 3 with a sugar framework.
Herein, we report on the diastereoselective and enantiose-
lective cycloaddition reaction of 1 with 2, catalyzed by the
axially chiral guanidine 4, a chiral Brønsted base catalyst de-
veloped within our group.[7,8]
An initial investigation was conducted to explore the opti-
mal aromatic substituent (Ar) at the C2 atom of azlactone
1.[9] Substituted azlactones 1a–1h were treated with a b,g-
unsaturated a-ketoester 2a with guanidine (R)-4 (2 mol%)
as the base catalyst at room temperature in THF. As shown
in Table 1, the desired cycloaddition products 3 were ob-
tained in high yields, irrespective of the steric and electronic
properties of the Ar substituent. However, the Ar substitu-
ent displayed a marked influence on both the enantio- and
diastereoselectivity (Table 1, entries 1–8). Although the re-
action provided the cis isomers predominantly in all cases,
the introduction of electron-donating groups at the 3- and 5-
positions led to a slight decrease in selectivity for the cis
isomer (Table 1, entries 5 and 6). In contrast, dramatic posi-
tional effects on the aromatic ring were observed in terms
of enantioselectivity. Introduction of the electron-donating
methoxy group at the 4-position resulted in the formation of
racemic products (Table 1, entries 2 and 8) whereas the
slightly electron-donating methyl and electron-withdrawing
trifluoromethyl substituents gave moderate selectivity
(Table 1, entries 3 and 4). The highest enantioselectivity was
observed for the 3,5-dimethoxyphenyl substituent, albeit at
the expense of diastereoselectivity (Table 1, entry 5).
Among the aryl substituents tested, the azlactone with an
unsubstituted phenyl group was optimal in terms of the
enantio- and diastereoselectivity (Table 1, entry 1).
Although cis-3aa was obtained with high diastereoselec-
tivity (Table 1, entry 1), the enantioselectivity remained
moderate despite thorough screening of the Ar substituent.
Hence, we next optimized the reaction conditions by chang-
ing the solvent and reaction temperature (Table 2). Delight-
fully, the enantioselectivity could be slightly improved with
acyclic ethers as solvents, which gave an increase in cis se-
lectivity (Table 2, entries 3–5). Among the acyclic ethers
tested, diethyl ether displayed the highest enantio- and dia-
stereoselectivity (Table 2, entry 3). Further optimization of
the reaction conditions by lowering the reaction tempera-
ture resulted in an increase in enantioselectivity (Table 2,
entries 6 and 7). However, in the reaction conducted at
À608C for 2 h, a considerable amount of the acyclic product
5 (the Michael addition product resulting from protonation
1
of intermediate A, see Scheme 1) was detected by H NMR
spectroscopic analysis of the crude product (Table 2,
entry 7).[10] Compound 5 was transformed to the desired
cyclic product 3aa by warming the reaction to room temper-
ature, after the azlactone 1a had been consumed completely
Table 1. Enantioselective and diastereoselective cycloaddition reaction of
1 with 2a catalyzed by an axially chiral guanidine (R)-4.[a]
Table 2. Optimization of the cycloaddition reaction of 1a with 2a, cata-
lyzed by (R)-4.[a]
Entry Solvent
T
Yield [%][b] cis/trans[c] ee [%][d]
1
2
3
4
5
6
7
8
THF
RT
RT
RT
RT
94
99
97
90
97
92
86
93:7
91:9
97:3
95:5
96:4
96:4
95:5
99:1
72
69
80
79
79
88
90
91
Entry
1
3
t [h]
Yield [%][b]
cis/trans[c]
ee [%][d]
DME[e]
Et2O
1
2
3
4
5
6
7
8
1a
1b
1c
1d
1e
1 f
1g
1h
3aa
3ba
3ca
3da
3ea
3 fa
3ga
3ha
2
6
1.5
6
4.5
2
94
95
98
92
93
70
95
97
93:7
93:7
91:9
96:4
80:20
79:21
95:5
94:6
72
6
tBuOMe
CPME[f]
Et2O
Et2O
Et2O
69
48
73
59
62
<1
RT
À208C
À608C
À608C to RT[g] 99
4.5
1.5
[a] Unless otherwise noted, all reactions were carried out by using (R)-4
(0.002 mmol, 2 mol%), 1a (0.10 mmol), and 2a (0.11 mmol, 1.1 equiv) in
the indicated solvent (0.5 mL) for 2 h. [b] Yield of the isolated product.
[c] Determined by 1H NMR spectroscopy. [d] Determined by chiral sta-
tionary-phase HPLC analysis of the major cis isomer. [e] 1,2-Dimethoxy-
ethane. [f] Cyclopentyl methyl ether. [g] Carried out at À608C for 2 h,
then warmed to room temperature and stirred for 1 h.
[a] Unless otherwise noted, all reactions were carried out with (R)-1
(0.002 mmol, 2 mol%), 1 (0.10 mmol), and 2a (0.11 mmol, 1.1 equiv) in
THF (0.5 mL) at room temperature. [b] Yield of isolated product. [c] De-
termined by 1H NMR spectroscopic analysis. [d] Determined by chiral
stationary phase HPLC analysis of the major cis isomer.
Chem. Eur. J. 2011, 17, 1760 – 1763
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1761