2
216
A. M. Ghanim et al. / Tetrahedron Letters 57 (2016) 2215–2218
NH
2
O
2
CO H
N
PhCHO, Ac O
Ph
N
2
O
O
N
N
N
K CO , 100 oC, 1 h
N
N
N
NH
HN
2
3
N
HO
Ph
85%
Ph
N
O
N
H
O
O
7
8
1
2
3
HO
OH
2
0% aq. NaOH,
2
H N
NH
reflux, 2.5 h.
Figure 1. The 1,2,4-triazine nucleus and modifications.
HN
NH
S
6
N
Ph
NH
2
O
Ph
including 12, such resonances occurred in the ranges d
and d 60–69 ppm, respectively.
H
5.0–6.2
O
N
NH
S
CO
2
AcOH, reflux
5%
2
C
6
Fortunately, a brief solvent study established that substituting
:1 aqueous ethanol for acetonitrile resulted in complete selectiv-
10
9
1
Scheme 1. Synthesis of the starting material 10.
ity in favour of the S-alkylated product 11. This had good levels of
both chemical yields and regioselectivity for a range of primary
allylic halides, as set out in Table 1.18 No traces of N-alkylated
products arising from attack at the pendant N-amino group were
ever observed.
Having achieved this first objective, we then wondered if the
corresponding N-alkylated derivatives [cf. 12] could be obtained
by carrying out a thio-Claisen rearrangement19 of the initial ally-
lated products 13, without any interference from the pendant
Ph
O
Ph
O
Ph
O
N
N
N
N
N
N
NH
MeCN
N
N
+
S
S
S
NH2
NH2
NH2
~
3:1
Cl
1
0
11
12
amino group, which potentially could attack the new allyl group
Scheme 2. Initial alkylation attempt in MeCN.
0
by an S
N
2 mechanism. Initially, we examined purely thermal acti-
vation, the simplest way in which such rearrangements can be trig-
gered. It was soon established that relatively high temperatures
were required to trigger such thio-Claisen rearrangements with
prolonged heating in xylene at 120–140 °C delivering reasonable
isolated yields of the N-allylated products 14a–g (Table 2).20 How-
ever, we were concerned that prolonged exposure to such high
temperatures could preclude the use of more sensitive substrates
in future studies.
(tosyl chloride, CH Cl , Et N or pyridine) and to similar reactions
2
2
3
with chloroformates. However, we were successful in adding a
benzyl group by exposing precursors 13f,g to benzyl bromide in
aqueous ethanol containing 1.5 equiv of potassium hydroxide
(Scheme 4).
Despite the presence of what should be a readily protonated
amino group in both the initial precursors 13f,g and the N-benzy-
lated derivatives 16a,b, we examined the possibility of carrying out
acid-catalysed cyclisations of such substrates on the grounds that
these all also contain easily protonated side chains, which could
result in the formation of a tertiary carbenium ion. We also
included the homologous prenylated precursor 13h in this group
for the same reasons. We observed complete disappearance of all
these precursors when exposed to an equivalent of trifluo-
romethanesulfonic (triflic) acid in toluene at reflux for 1 h.
Milder conditions are available by using a palladium-based cat-
alyst to induce such rearrangements.21 We were pleased to find
that these were successful when the precursors 13 were heated
at reflux in dry tetrahydrofuran for somewhat briefer periods
2
2
(
Table 2). Isolated yields were generally higher or at least compa-
rable with the purely thermal procedure, except in the two exam-
ples where the rearrangement involved migration of a 1,1-
disubstituted alkene (entries f, g). Various attempts to improve
such returns by using higher boiling ether solvents failed. At pre-
sent, it is not clear why these two cases were inefficient. The one
exception to the foregoing positive results was the complete failure
of the prenyl-substituted derivative 13h to undergo rearrangement
using either the purely thermal or the palladium-catalysed method
to give the isomeric species 15 (Scheme 3). Excessive steric
requirements may be responsible for this.
2
4
In the case of precursor 16a, the product was separated by
1
column chromatography in 80% yield. H NMR data revealed the
disappearance of the terminal alkene protons and clearly showed
the continued presence of an NH group as a triplet but which
was shifted from d 5.37 to d 5.83, together with a six-proton res-
H
H
onance at dH 1.46, due to freely rotating geminal methyl groups
3
attached to an sp carbon. Crucially, two new resonances at dH
Having succeeded in alkylating both the sulfur and the ring
nitrogen centres, we then tried to derivatize the pendant amino
group in the initial alkylation products 13 in anticipation of being
able to effect acid-catalysed cyclisation of such modified amino
groups onto the newly introduced alkene side chains.23
We were surprised to find that attempts to add typical amine
protecting groups to this likely reactive amine group, which can
be considered as a monoacylated hydrazine functional group, were
unsuccessful. Substrates including the S-allyl and S-cinnamyl
derivatives 13a,b proved inert to ‘standard’ tosylation conditions
3.02 (2H, d, J = 9.1 Hz) and dH 0.98 (1H, t, J = 9.1 Hz) pointed to
the presence of a new CH SH thiol group and hence to the forma-
2
tion of the rearranged structure 20c, the origin of which is sug-
gested in Scheme 5.
As anticipated, an initial protonation would give rise to a ter-
tiary carbenium ion 17, trapping of which by a ring imine nitrogen
would then lead to the iminium species 18, stabilised by a number
of resonance forms, which could then be neutralised by addition of
adventitious water to give a hemiacetal species 19. This would
then collapse to give the observed products 20. Throughout this
pathway, it is likely that the benzylamine group is also protonated;
this feature has been omitted for clarity.
R
O
N
N
R
O
O
H
2
N
This rearrangement is general for this structural type, at least in
the case of the five substrates tested [13f,g,h; 16a,b], as set out in
Table 3. The yields quoted are for reactions which were followed
by TLC until complete disappearance of the starting material.
Lower temperatures, such as that of refluxing dichloromethane,
were insufficient to trigger the rearrangement. In the context of
NH
NH
+
S
OH
HN
NH
S
NH
2
2
4
5
6
Figure 2. 1,2,4-Triazin-3-thiono-5-one synthesis.