K. Geyl et al. / Tetrahedron Letters 60 (2019) 151108
3
compounds decomposed in H2O2/AcOH and were also not oxidized
by m-CPBA in CH2Cl2. Therefore, the reaction sequence was altered
and the N-oxidation was performed at the first stage. The subse-
quent oxadiazole core assembly was carried out under mild condi-
tions according to our previous developed procedures (Scheme 2)
[51–53].
3a was obtained in an improved isolated yield (76%) (Table 1, entry
13). The structure of product 3a was established by NMR spec-
troscopy and HRMS. In addition to 3a the N-oxide reduction pro-
duct – 5-methyl-3-(pyridin-4-yl)-1,2,4-oxadiazole – was isolated
(ꢀ10%) from the reaction mixture. We then studied the possibility
of increasing the reaction scale under these conditions. An increase
in the reagent amounts by 4.5 times (Table 1, entry 14) did not lead
to a noticeable decrease in the product yield, which indicates the
scalability of the proposed procedure.
Disappointedly, the acid-catalyzed CAH functionalization with
dialkylcyanamides reported by Rassadin and co-workers [28] was
not suitable in the case of oxadiazoles due to decomposition.
Therefore, we had to develop a new methodology. Initial optimiza-
tion showed that the reaction needed to be carried out in a solvent
(Table 1, entries 1–8). Although complete conversion of the N-
oxide and isolation of the corresponding urea 3a was achieved in
moderate yield (68%) using excess dimethylcyanamide (DMCA,
Table 1, entry 2) as the solvent, this method is too wasteful and
expensive for the practical use. MeCN was recognized as the most
appropriate solvent for this reaction. However, due to moderate N-
oxide conversion (66%), the optimization was continued by study-
ing the effect of the temperature and the reaction time (Table 1,
entries 9 and 10). Unfortunately, these efforts did not provide the
desired result despite a significant increase in the conversion of
the starting N-oxide. Increasing the temperature (up to 80 °C), as
well as prolonging the reaction time, led to an increase in the
impurities formed. According to our observations, the latter are
formed due to 1,2,4-oxadiazole ring decomposition. Then we tried
to vary the DMCA and MsOH amounts (Table 1, entries 11–13). An
increase in the DMCA amount used to 2 equiv. led to an increased
conversion (Table 1, entry 11), while the use of a larger excess had
no effect (Table 1, entry 12). The reaction with 1.5 equiv. of MsOH
resulted in full (98%) conversion of N-oxide 1a and the desired urea
Finally, we examined the possibility of using microwave heat-
ing and found that this does not noticeably affect the speed of
the process (Table 1, entry 15).
Used these optimized conditions, we verified the scope of the
reaction using a number of 1,2,4- and 1,3,4-oxadiazolyl-substi-
tuted N-oxides 1 as well as selected commercially available
dialkylcyanamides 2 (Scheme 3). Firstly, we replaced DMCA 2a
with other dialkylcyanamides (diethylcyanamide 2b, 1-
piperidinecarbonitrile 2c, 4-morpholinecarbonitrile 2d) in the
reaction with 1a. In the case of diethylcyanamide the yield of pro-
duct 3b decreased, whereas ureas (3c,d) were obtained in good
yields. Further, a number of structural variations related to the
1,2,4-oxadiazole ring, including replacement of the substituents
at positions 3 and 5 of the heterocycle, were explored. There were
no observed effects from the replacement of Me with Ph (3e), as
well as the relocation of substituents (3f,g). In contrast, the change
from 4-pyridine to 3-pyridine led to a lower yield (3h). However,
the most intriguing results were obtained in cases of 2-pyridine
substrates. 2-(5-Phenyl-1,2,4-oxadiazol-3-yl)pyridine 1-oxide 1e
reacted with cyanamide 2a providing the desired urea 3i in moder-
ate yield (63%), whereas N-oxides 1f and 1g, bearing a 1,2,4-oxadi-
Scheme 3. Reaction scope with various N-oxides 1 and dialkylcyanamides 2.