S. S. Gratia et al. / Tetrahedron Letters 55 (2014) 448–452
449
R'
in good yields. However, we were only able to produce the known
2, 5-diphenylthiazoline 1e16 in very low yields (results not shown).
When 1-nonene was reacted with benzamide a 1:1 mixture of 5-
heptyl-2-phenyloxazoline 4a and 4-heptyl-2-phenyloxazoline 4a0
were obtained in good yields (Table 1, entry 4). 1H NMR of 5-hep-
NH2COR'
N
1A,1B,1D,2A,2D
3A,3B,3D,4A-A',
4B-B', 4D-D'.
O
NIS/Propionitrile
4Å MS/ 45oC
R
tyl-2-phenyloxazoline gave
a distinct multiplet for H-5 at
Ph
4.72 ppm and a doublet of doublet for H-4 at 4.12 (dd, J = 7.0 Hz,
J = 6.50 Hz, 1H, CHOCH2N) ppm and 3.67 (dd, J = 7.5 Hz,
J = 7.00 Hz, 1H, CHOCH2N) ppm. The regioisomer 4-heptyl-2-phe-
nyloxazoline gave a distinct doublet of doublet at 4.49 (dd,
J = 8.50 Hz, J = 7.50 Hz, 1H, CHNCH2O) ppm and 4.05 (dd,
J = 8.25 Hz, J = 8.00 Hz,1H, CHNCH2O) ppm for H-5 and a multiplet
at 4.28 ppm for H-4 (Scheme 2). The same chemical shift pattern
was observed for the rest of the corresponding 5-heptyl and 4-hep-
tyl substituted oxazolines. Furthermore, 1:1 ratios were also ob-
tained for compounds 4b, 4b0, 4d, 4d0 (Table 1, entry 4). The
regioisomers were easily separated by column chromatography
with the exception of compounds 4d and 4d0 that resulted in an
inseparable mixture. When hydroxybenzamide was used in the
reaction the dihydro-1, 4, 2-dioxazine 4c was obtained as a single
product albeit in low yields. Unfortunately very poor yields were
also obtained with thiobenzamide and the preparation of 4e was
not further pursued.
The use of vinylanisole 2 and vinylpyridine 3 allowed for the for-
mation of the known compounds 2a, 2e, 3a, 3b, 3d, 3e (Table 1, en-
tries 2 and 3).16,17 The yields ranged from good to low, especially for
compounds 2e, 3b, and 3d. During the synthesis of these com-
pounds multiple spots were observed during TLC analysis. The
iodination of mono substituted benzene using NIS and acetoni-
trile18 as well as NIS in the presence of trifluoroacetic acid19 has
been previously reported. It is possible that some of the byproducts
obtained in these reactions may have resulted from the halogena-
tion of the benzene ring via electrophilic aromatic substitution.
We decided to test this hypothesis by mixing vinylanisole with
2.0 equiv of NIS in propionitrile at 45 °C for two hours (Scheme 3).
After column chromatography 1H NMR analysis confirmed the for-
mation of compound 9 only. This result does not surprise us since
the addition of the succinimide ion in the absence of a strong nucle-
ophile has been previously reported using NIS and glucal.20 No sub-
stitution occurred in the benzene ring and compound 9 was never
isolated in any of our reaction trials with alkenes 2 and 3.
N
NH2CSPh
1E,3E.
S
R
R
NIS/Propionitrile
4Å MS/ 45oC
R= 1)C6H5
N
Ph
2)p-CH3OC6H4
O
3)C5H4N
1C,2C,3C,4C.
HONHCOPh
4)CH3(CH2)6
O
R
R'=Ph,NHPh,C4H3S
NIS/Propionitrile
4Å MS/ 45oC
Scheme 1. Targeted N-heterocycles.
thiazolines, and the less abundant dihydro dioxazine via the addi-
tion of amides and N-hydroxyamides to alkenes and unsaturated
sugars in the presence of NIS and propionitrile at 45 °C. In this
study we show that access to oxazolines and other heterocycles
is possible using alkenes and an excess of both NIS and the amide
at 45 °C. The overall yields and stereoselectivity obtained for the
formation of oxazolines when using electron rich alkenes are com-
parable with currently published one-pot procedures. Previously
we attempted to use the electron deficient glycals, tri-O-acetyl-
glucal and tri-O-acetyl- -galactal in a two-step iodoamidation/
cyclization to form the oxazolines. Yields were poor and mixtures
of products were obtained.10 When tri-O-acetyl-
-glucal, tri-O-
acetyl- -galactal, and tri-O-methyl- -glucal were employed in the
new one-pot reaction the N-glycooxazolines were obtained exclu-
sively. Compounds 1a, 1c, 2a, 2d, 2e, 3c were selected to be
screened against a panel of enzymes as part of Eli-Lilly’s open inno-
vation drug discovery program. All of the compounds were inactive
with the exception of 1c which exhibited moderate inhibition
against hexokinase II (hHK2) with a 70% inhibition and IC50
D-
D
D
D
D
>20 lM.
Results and discussion
Based on these results one can envision the formation of the
oxazolines and thiazolines via the addition of the nitrogen of the
amide to the benzylic carbon of aryl alkenes. Migration of the
nitrogen via an acyl aziridine followed by O-alkylation will lead
to ring closure and formation of the 2,5-disubstituted oxazoline/
thiazoline. While a free radical mechanism is possible it is not
likely to occur in this case since reaction yields were not affected
when conducted in the presence of light. However, the migration
of amides through an aziridine intermediate is well documented.21
Therefore we hypothesized that the formation of the dihydro
dioxazine will follow a similar pathway through the initial forma-
tion of intermediate 11 (Scheme 4). Displacement of the iodine fol-
lowed by O-alkylation will lead to ring closure and formation of our
target compound 13 exclusively. The electron deficient 5,6-dihydro-
2H-pyran-2-one and trimethyl silyl acrylate were also used in this
study but gave very sluggish results and their use was discontinued.
As part of our program aimed at the synthesis of novel glycosi-
dase inhibitors we decided to revisit the direct preparation of the
Compounds 1a–e (Table 1, entry 1) were prepared by reacting
benzamide,
thiophene-2-carboxamide,
thiobenzamide,
N-
hydroxybenzamide, and phenylurea with styrene in the presence
of N-iodosuccinimide (NIS).12 Hence, styrene (1 equiv) was mixed
in freshly distilled propionitrile (4.0 mL) followed by the addition
of the amide (3 equiv). The mixture was heated to 45 °C until the
reaction became homogenous. Pre-activated 4 Å molecular sieves
powder was added to the reaction mixture to avoid inadvertent
addition of water across the double bond and formation of halo-
hydrins. Finally, NIS (2 equiv) was added to the reaction mixture.
The reaction was kept at 45 °C and after 2 h TLC analysis showed
complete disappearance of the starting material. The mixture
was quenched with deionized water and diluted in dichlorometh-
ane (50 mL). The crude was filtered to remove the molecular sieves
and subsequently washed with saturated Na2S2O3 (100 mL) and
distilled water (3 ꢀ 100 mL). Reactions were conducted in the
presence and absence of light and no changes were observed in
product yields. The best overall yields were obtained when 1-non-
ene 4 and styrene 1 were used in the reaction (Table 1, entries 1
and 4). Our NMR data for compounds 1a13 and 1d14 matched the
data previously reported confirming the formation of the 2, 5-
diphenyloxazoline 1a and N, 5-diphenyl-4, 5-dihydrooxazol-2-
amine 1d. The known compounds 1b15 and 1c were also obtained
N-glycooxazolines using the electron deficient tri-O-acetyl-
D-glu-
cal 5 and tri-O-acetyl- -galactal 6 (Table 1, entry 6). Acetate pro-
D
tecting groups are easily removed under basic conditions making
them a very desirable protecting group in carbohydrate synthesis.
When benzamide was used in the reaction the N-glycooxazolines
5a, 6a were obtained exclusively in fair yields. Similar results were
obtained when thiophene-2-carboxamide was employed in the