Full Paper
in reaction time. Total consumption of the starting material
was observed after 3hours, and oxazine 2a was isolated in
57% yield, again with the acetyl derivative 5a (Table 1,
entry 3). The use of a different silver salt (i.e., silver triflimide)
led to comparable results (Table 1, entry 4). At this point, it is
important to mention that TLC analysis revealed that product
5a was derived from oxazine 2a and can be assumed to form
from traces of water present in the reaction medium. To avoid
this hydrolysis, the reaction was performed in the presence of
molecular sieves (3 ꢅ), which led to a significant improvement
in the yield (Table 1, entry 5). Although the yield determined
by GC-MS was 83%, the yield of the isolated oxazine 2a was
72%, thus indicating that this compound is labile and decom-
poses during purification. Substrate 1a was inert either in the
presence of the silver salt or triflic acid, thus indicating the role
of gold in catalyzing the reaction (Table 1, entries 6 and 7). The
next step of our study was the evaluation of the solvent in the
process. The use of toluene or acetonitrile as solvents led to
a decrease in efficiency (Table 1, entries 8 and 9). However, the
use of methanol gave a surprising result, in which the exclu-
sive formation of dihydropyridone 4a was observed in excel-
lent yield after 3 hours (Table 1, entry 10). The use of silver tri-
flimide gave comparable results, although with longer reaction
times (Table 1, entry 11). On the other hand, it took 24 hours to
complete the process and a mixture of 4a and 5a was ob-
tained when AuCl was used as catalyst (Table 1, entry 12).
With these results, we proceeded to study the behavior of
other homopropargylic amides under the optimized condi-
tions. Thus, when homopropargyl amides 1 were treated with
[AuClPPh3] in combination with AgOTf in dichloromethane in
the presence of molecular sieves, a clean formation of oxazines
2 was observed (Table 2, entries 1–7). As mentioned before,
these compounds partially decompose during purification, but
obtaining good yields of the isolated products was possible.
Hydrolysis of oxazines 2 took place in almost quantitative
yields in all cases to render the alkyne hydrolysis products 5
(Table 2, entries 1–7). It seems that the hydration of the triple
bond under gold catalysis is assisted by the carbonyl group.
The solvent played a crucial role in the process, and the exclu-
sive formation of dihydropyridones 4 in methanol was ob-
served in good-to-excellent yields (Table 2, entries 1–7).
Table 2. Preparation of oxazines 2, dihydropyridones 4, and acetyl deriva-
tives 5.
Entry
R1
R2
R3
Yield 2
Yield 4
Yield 5
[%]
[%]
[%]
1
2
3
4
5
6
7
8
CF3
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
CO2Et
Ph
Me
Me
Me
Me
Et
Me
Me
Me
2a 72 (83)[a]
2b 70 (86)[a]
2c 80 (94)[a]
2d 88 (98)[a]
2e 61 (95)[a]
2 f 76
4a 93
4b 93
4c 80
4d 94
4e 70[b]
4 f 75
4g 76
–
5a 99
5b 99
5c 99
5d 99
5e 99
5 f 84
5g 99[c]
–
CF2Ph
CF2Bn
CF2allyl
CF3
Ph
CF3
2g 74
[d]
Ph
H
–
[a] The yield determined by GC-MS is given in brackets. [b] The formation
of 28% of 5e was also observed. [c] To carry out the hydrolysis in this
case, it was necessary to add 5 mol% of the gold(I) salt to activate the ox-
azine for the nucleophilic attack. [d] The substrate remained unaltered in
methanol.[8]
aza-Wittig reaction with chiral phosphazene 7 and the corre-
sponding iminoester 8 was treated in situ with propargyl bro-
mide and activated zinc in DMF under Barbier-type conditions.
The addition of propargylzinc was completely selective, thus
affording homopropargylic sulfinylamine 9 as a single diaste-
reoisomer. Protecting-group removal and acetylation led to the
starting amide 1a in enantiomerically pure form.[11] Reaction of
1a with [AuOTfPPh3] gave oxazine 2a in dichloromethane,
acetyl derivative 5a upon hydrolysis, and dihydropyridone 4a
in methanol (Scheme 2). Compound 4a contains a cyclic qua-
ternary a-amino acid unit, which infers more relevance to the
asymmetric process.
In all cases, a quaternary center is present in the starting
amides 1. The reactivity of substrates that arise from non-fluo-
rinated aldimines (Table 1, entry 8) has been described
before;[8g] although the formation of the corresponding oxa-
zine has been reported, these compounds were unstable and
were subjected to further transformations without isolation.
Additionally, this substrate was inert in the presence of the
gold salt in methanol, even in the presence of acidic additives,
thus preventing the complexation of gold and the basic nitro-
gen atom (Table 2, entry 8). It is clear that the presence of the
quaternary center plays a significant role in the process, but
we do not have an explanation so far for this different
behavior.
As mentioned before, hydration products 5 come from oxa-
zines 2. In the same way, we observed by TLC analysis that the
first step in the reaction performed in methanol was once
again the formation of oxazines 2, thus indicating that these
compounds are also intermediates in the formation of dihydro-
pyridones 4. To prove this observation and the role of the gold
salt in the process, several experiments were performed. First,
when oxazine 2a was dissolved in dry methanol, it remained
unaltered after 24 hours at room temperature (Table 3,
entry 1). When a catalytic amount of triflic acid was added to
the reaction mixture, total consumption of the starting materi-
al was observed after 6 hours, thus giving rise to 4a in 61%
yield (Table 3, entry 2). On the other hand, the reaction in the
presence of AgOTf did not proceed, thus recovering the start-
ing material untouched after 24 hours at room temperature
The next step of our study was an asymmetric version of the
sequence. To this end, sulfinylimines were used as chiral auxil-
iaries. Thus, ethyl trifluoropyruvate (6) was subjected to an
Chem. Eur. J. 2014, 20, 1 – 7
3
ꢃ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ