Organic Letters
Letter
Herein we describe the regiospecific trifluoromethylative
spirocyclization (CF3-spirocyclization) of cyclic alkenes bearing
an amide pendant mediated by photoredox catalysis, i.e.,
photoredox-catalyzed oxytrifluoromethylation,4c,10e,l,11a,g to be
a versatile protocol for access to CF3-containing spirooxazolines
and spirooxazines. In addition, this is the first report on
predominant formation of anti-diastereomers of CF3-containing
spirocyclic compounds.13
Scheme 2. Scope of the Present Photocatalytic CF3-
Spirocyclization of Allylic Amides (2)
a b c
, ,
We initially examined the photocatalytic reaction of N-((3,4-
dihydronaphthalen-1-yl)methyl)benzamide (2a) with Umemo-
to’s reagent 1 in the presence of 5 mol % [Ru(bpy)3](PF6)2 in
acetone-d6 at room temperature under visible light irradiation
with blue LED lamps (λmax = 425 15 nm). After 1.5 h, 2a was
completely converted but the corresponding spirooxazoline 3a
was not formed at all (Table 1, entry 1). Instead, a N-protonated
Table 1. Optimization of the Photocatalytic CF3-
a
Spirocyclization of Allylic Amide (2a)
b
b
entry
base
none
solvent
temp
rt
% yield of 3a
dr
c
1
2
3
4
5
acetone-d6
acetone-d6
acetone-d6
acetone-d6
CD2Cl2
−
−
−
−
c
K2CO3
rt
2,6-lutidine
2,6-lutidine
2,6-lutidine
2,6-lutidine
none
rt
91
91
90
0
77:23
84:16
74:26
−
−78 °C
−78 °C
−78 °C
−78 °C
−78 °C
d
6
acetone-d6
acetone-d6
acetone-d6
e
7
e
0
−
f
f
8
2,6-lutidine
28, 37
74:26
a
Reaction conditions: A mixture of [Ru(bpy)3](PF6)2 (1.25 μmol, 5
mol %), 1 (28 μmol, 1.1 equiv), 2a (25 μmol, 1.0 equiv), and solvent
a
b
b
Reaction conditions: see the Supporting Information. Yields were
(0.4 mL) was irradiated by 3 W blue LEDs (λ = 425 15 nm). Yields
c
and diastereomer ratios (dr) were determined by 1H NMR
obtained after purification. Diastereomer ratios (dr) were determined
d
by 1H and 19F NMR spectra of crude reaction mixtures. CH2Cl2 was
c
spectroscopy using SiEt4 as an internal standard. Protonated 3a was
e
d
e
f
used as a solvent due to solubility of substrate. Reaction was carried
out in the presence of 1.7 equiv of 1 at 0 °C for 4 h.
formed. In the dark. No photocatalyst. Reaction time = 8 h.
compound type of 3a was likely to be formed (see the Supporting
Information). Therefore, addition of a base, K2CO3 and 2,6-
lutidine, was tested. K2CO3 showed a result similar to entry 1
possibly due to its low solubility (entry 2). In contrast, to our
delight, 2,6-lutidine yielded the product 3a in a good yield (91%)
with good diastereoselectivity (77:23 dr) (entry 3). Next, to
improve the diastereoselectivity, the reaction was conducted at
−78 °C (in a dry ice−methanol bath), resulting in better
diastereoselectivity (84:16 dr) (entry 4). Use of CD2Cl2 as a
solvent also afforded the product smoothly (90% yield) but with
a slightly lower diastereoselectivity (74:26 dr) (entry 5). Finally,
the reaction did not proceed at all either in the dark or in the
absence of a photocatalyst (entries 6 and 7). To our surprise, 2,6-
lutidine also promoted the reaction to some extent even in the
absence of the photocatalyst (entries 7 and 8).14 But, even after a
longer reaction time, the yield was much lower than that obtained
by photoredox catalysis, suggesting that the reaction promoted
by 2,6-lutidine is not the main reaction pathway of the present
photocatalysis.
First, substituents on the amide pendants were explored.
Substrates with benzene rings bearing halogens, Br (2b) and I
(2c), and an electron-donating group, MeO (2d), and aliphatic
t
groups, Me (2e) and Bu (2f), afforded the corresponding
spirooxazoline products (3b−f) in good to high yields (59−
86%) with moderate to good diastereoselectivities (72:28−86:14
dr). Amide functionalities with an alkoxy group, OEt (2g), a
trichloromethyl group (2h), a naphthyl group (2i), and a pyridyl
group (2j) tolerated the present photocatalytic system (3g−3j:
52−85%, 75:25−86:14 dr). Remarkably, the alkene with a bulky
mesitylamide pendant (2k) gave 3k in a 73% yield with excellent
diastereoselectivity (97:3 dr). Next, structures of cyclic alkenes
were investigated. Reactions of N-((2H-benzopyran-4-yl)-
methyl)benzamide (2l) and N-((2H-thiocromen-4-yl)methyl-
benzamide (2m) proceeded without any retardation regardless
of the ether and sulfide functionalities to give the corresponding
CF3-spirocyclized products in good yields (3l: 71%, 88:12 dr and
3m: 64%, 89:11 dr) in a diastereoselective manner, respectively.
The reaction of N-((1H-inden-3-yl)methyl)benzamide (2n) also
afforded a good yield of the CF3-containing spirooxazoline (3n:
77%) but with slightly lower diastereoselectivity (77:23 dr)
compared to the dihydronaphthalenyl skeleton (3a). A
tetrasubstituted alkene, ((1H-2-methyl-inden-3-yl)methyl)-
With the optimal reaction conditions in hand, the preparative
scale experiments were performed. Then, we found that the
catalyst loading can be reduced to 0.5 mol %. The obtained CF3-
containing spirooxazolines 3 were summarized in Scheme 2.
B
Org. Lett. XXXX, XXX, XXX−XXX