Angewandte
Chemie
[
a]
deficient azides (e.g., MsN , CbzN ) were also effective at
Table 3: Oxime and azide chemoselectivity.
3
3
providing the corresponding amidines, yields were generally
lower (ca. 20%) than that observed with TsN . Not surpris-
3
ingly, non-stabilized azides (e.g., BnN ) led to exclusive
3
isolation of rearranged phosphinite as a result of the higher
temperatures required for phosphazene generation. The
formation of the amidines 3 depicted in Table 2 is consistent
with a Beckmann-like ligation in which the anti-periplanar
oxime substituent migrates preferentially.
To gain insight into the stereochemical outcome of the
migration event, we chose to evaluate our phosphite-medi-
ated ligation approach using the a-chiral menthone-derived
[
c]
Entry Additive
5
3b
yield [%]
3b/5
[
b]
[
c]
1
H
O
2
49
4:1
oxime 1j [Eq. (1)]. Exposure of 1j and TsN to ClP(dmp-ol)
3
[d]
2
3
PhSO NH
85
40
>99:1
2
2
[
c]
5a
3:1
[a] Reaction conditions: 0.3 mmol of 1b, 0.3 mmol of ClP(dmp-ol),
0
.3 mmol of Et N, 0.3 mmol of TsN , and 0.3 mmol of additive in PhMe
3
3
1
(0.2m). [b] Yield of isolated product. [c] Ratio determined by H NMR
(600 MHz) spectroscopy. [d] Ratio determined by HPLC.
and Et N in PhMe at 08C yielded the N-sulfonyl phosphor-
3
[
18]
amidine 3j in 58% yield as a single regio- and stereoisomer.
nucleophile (Table 3). Treatment of 1b and TsN with ClP-
3
It is important to note that under the reaction conditions,
a epimerization of the chiral oxime is not observed, and the
retention of absolute configuration at the site of migration is
consistent with a concerted [1,2] shift.
(dmp-ol), Et N, and H O (1 equiv) led to the formation of 3b
3
2
and 5 in a ratio of 4:1 (Table 3, entry 1). It is unlikely that 5 is
generated by in situ amidine hydrolysis as independent
treatment of 3b under the reaction conditions led solely to
While our results point to a distinct steric influence on the
migratory preference of oxime substituents, the relative
reactivity of electron-rich and electron-deficient aryl oximes
suggest that electronic factors may also impact the rate of
rearrangement. Speculating that we could exploit this dispa-
rate reactivity to kinetically resolve a mixture of oxime
isomers bearing electronically dissimilar, yet sterically con-
gruent substituents, we examined the oxime 1k as a 1:1 E/Z
mixture of isomers in our azide ligation approach toward
amidines. Exposure to our optimized reaction conditions led
exclusively to migration of the more electron-rich anti-
periplanar aryl ring to give the amidine 3k in 30% yield
recovered amidine. The presence of H O also resulted in
a less efficient ligation, thus giving 3b in only 49% yield.
2
However, the addition of PhSO NH yielded exclusively the
2
2
azide ligation product 3b in 85% yield (Table 3, entry 2).
Performing the reaction in the presence of p-methoxybenzoic
acid did not lead to sulfonyl imide formation consistent with
our recently reported chlorophosphite-mediated Staudinger-
[
11b]
like ligation (Table 3, entry 3).
However, 3b and 5a, likely
arising from carboxylic acid addition to a nitrilium cation
intermediate and subsequent hydrolysis, were obtained in
a 3:1 ratio. Our observation that the addition of a radical
inhibitor did not hinder the formation of 3b would indicate
[17]
(
Scheme 3). The corresponding Z-oxime isomer proceeded
a non-radical mechanism. While these results demonstrate
a distinct preference for azide ligation indicative of an
intermediate iminophosphorane (2), CÀN bond formation
no further than azide reduction to give the iminophosphorane
[
19]
(Z)-2k in 36% yield.
[
20]
In an effort to evaluate azide chemoselectivity we
likely occurs through a polar intermolecular process.
examined the ligation event in the presence of an added
To verify the intermediacy of 2 we again exploited the
tempered reactivity of the electron-deficient aryl oxime 1l in
the isolation of phosphazene 2l (Scheme 4). Treatment of 1l
with ClP(dmp-ol), Et N, and TsN at room temperature for
3
3
3
0 minutes gave 2l in 85% yield as a free-flowing solid whose
structure was unequivocally ascertained by single-crystal X-
[
17]
ray diffraction.
Monitoring of the reaction over time
showed the appearance and subsequent disappearance of 2l
en route to 3l in 40% overall yield. This observation provides
compelling evidence that 2l is a chemically and kinetically
[
17]
viable intermediate under the reaction conditions.
Based on these composite results, our working mecha-
nistic hypothesis involves initial phosphinylation of the oxime
1
and subsequent Staudinger-like reduction of TsN to give
3
Scheme 3. Kinetic resolution of oxime isomers.
iminophosphorane 2 (Scheme 5). A stereo- and regioselective
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!