N. G. Moon, A. M. Harned / Tetrahedron Letters 54 (2013) 2960–2963
2961
Table 1
Table 2
Optimization of reaction conditions
Cyclization of different N-allyl amides
% Yielda
Entry
1
R1
Compound
Time
2 h
% Yielda
Entry
Additive (equiv)
Solvent
Temp (°C)
1
2
3
4
–
5:1 AcOH/Ac2O
5:1 AcOH/Ac2O
5:1 AcOH/Ac2O
5:1 AcOH/Ac2O
CH2Cl2/10 eq. AcOH
CH2Cl2 (no AcOH)
CH2Cl2
50
50
50
25
25
25
25
23b
45c
86
64
84
BF3ÁOEt2 (0.1)
BF3ÁOEt2 (1.2)
BF3ÁOEt2 (1.2)
BF3ÁOEt2 (1.2)
BF3ÁOEt2 (1.2)
BF3ÁOEt2 (1.2)
b
77
5d
6e
7f
2
3
4
c
2 h
68
85
80
55
N.R.g
a
b
c
d
e
f
After chromatographic purification.
d
e
2 h
31% conversion after 24 h by 1H NMR analysis of the crude product.
68% conversion after 24 h by 1H NMR analysis of the crude product.
Reaction time = 2 h.
2.5 h
Reaction time = 5 h.
Without PhI(OAc)2.
g
No reaction by TLC and 1H NMR.
5
6
f
5.5 h
2 h
44
75
observations, speakstothepossiblemechanisticpathwaysresponsi-
ble for this transformation, vide infra. Lastly, it should be noted that
the optimal reaction conditions (entry 5) also involve workup with
aqueous ammonia. This is necessary to remove the BF3ÁOEt2, which
appears to form a strong adduct with the oxazoline product (see
Supporting information for details).
g
7
8
9
10
11
(E)-PhCH@CH–
Me–
t-Bu–
CF3–
EtO2CCH2–
h
i
j
k
l
2.5 h
2 h
2 h
2 h
2 h
74
24
57
decomp.
trace
b
c
d,e
Having identified an optimal set of reaction conditions, we then
investigated the influence of the R1 group on reaction efficiency
(Table 2). Electron-rich and electron-poor aromatic substituents
were tolerated and had little influence on the reaction time and
overall yield (entries 1 and 2). Similar observations were made
with halogenated substrates (entries 3 and 4). Heteroaromatic sub-
stituents are tolerated, but the efficiency is dependent on their
identity. Cyclization of the 2-pyridyl substrate 1f (entry 5) was
sluggish and resulted in lower yield. This is likely due to an addi-
tional Lewis basic moiety being present. Oxidation of the pyridine
to an N-oxide is another possible side reaction that would lead to
diminished yield. In contrast, cyclization of the 3-furyl substrate
1g (entry 6) was quite efficient and no Friedel–Crafts-type prod-
ucts were observed. The 2-styryl substrate 1h (entry 7) cyclized
smoothly, but the corresponding vinyl substrate (not shown) re-
sulted in general decomposition.
a
b
c
After chromatographic purification, unless otherwise noted.
48% yield by 1H NMR using mesitylene as an internal standard.
78% yield by 1H NMR using mesitylene as an internal standard.
3 equiv BF3ÁOEt2 used.
d
e
<10% yield by 1H NMR.
Some notable limitations were also encountered during this
preliminary study. The cyclization of substrates 1i and 1j (entries
8 and 9) proceeded with high conversion and no side reactions.
However, product isolation was quite difficult. Both products were
much more polar than those discussed above, and both lack an
effective chromophore. This made identification by TLC and chro-
matographic purification very challenging. Finally, the attempted
cyclization of substrates 1k and 1l resulted in more serious prob-
lems. Only general decomposition was observed with 1k. This is
likely due to the diminished nucleophilicity of the trifluroaceta-
mide group. With substrate 1l, only trace amounts of the desired
product were observed along with substantial decomposition.
Competitive coordination of BF3ÁOEt2 to the b-dicarbonyl moiety
present in this starting material is one possible explanation for this
failure.
Scheme 1. Investigating the diastereoselectivity.
transferring its stereochemical information and oxazoline 2n was
formed as the major diastereomer. The relative configuration was
determined to be trans by NOE and coupling constant analysis.17
The expected mechanism for this transformation is shown in
Scheme 2. The dramatic influence of BF3ÁOEt2 on cyclization effi-
ciency (Table 1, entries 1–3) suggests that PIDA is first converted
to electrophilic aryliodinium ion 3. Interaction of iodonium ion 3
with the alkene in 1 will generate either an activated olefin com-
plex (4a) or a cyclic iodonium ion (4b). When R2 = H, attack by
the amide in a 5-exo fashion will give primary alkyl iodane 5.
The enhanced leaving group ability of the iodine(III) nucleus18 will
make nucleophilic attack by a second nucleophile (in this case
acetate) quite favorable. This SN2-like, bimolecular reductive
elimination forms the oxazoline and liberates acetate and iodoben-
zene. We made several attempts at cyclizing substrates with
internal alkenes. Unfortunately, these efforts resulted in complex
The substrates shown in Scheme 1 were prepared in order to
investigate the diastereoselectivity of the cyclization event. Cycli-
zation of valinamide 1m proceeded in high yield. Not surprisingly,
the remote stereocenter in this substrate had little influence and
oxazoline 2m was formed as
a 1:1 mixture of inseparable
diastereomers. The branched substrate 1n was more successful in