S. Lopez et al. / Tetrahedron Letters 50 (2009) 6022–6024
6023
ing acetylamino compound 8 and the 2-acetyl derivative of 6, from
which 8 could be isolated in 38% overall yield on crystallisation.
The Vilsmeier–Haack reagent was used to convert 8 into the re-
quired dichloronitropyrimidine 1 in very good yield, and even
though compound 5 was still present, they could be separated eas-
ily by flash chromatography (Scheme 2).
O
O
Ac2O,
KNO3/H2SO4
rt, 2 h
cat. H2SO4
NO2
Cl
HN
H2N
HN
H2N
90 °C, 1 h
N
Cl
N
6
7
Considering these results, a question arises: why is the nitro
group displaced in the Vilsmeier–Haack reaction in compound 3
but not in 8. In heteroaromatic systems, the substitution of a nitro
group by chlorine is an unusual reaction and, therefore, very little
information regarding this particular type of substitution, and its
mechanism could be obtained from the literature; however, two
alternative explanations can be given. The displacement of the ni-
tro group in compound 3 will be promoted by an electron-with-
drawing group at the 2-position. This can arise from either
protonation of the 2-amino group, or, if the 2-aminomethylene-
dimethylamino group had been introduced first, by reaction of this
with a proton or other electrophilic species derived from the re-
agent. A 2-acetylamino group, as in 8, would not be as basic, and
it would not protonate or promote the replacement reaction at
the 4-position.
Having obtained trichloro compound 5, and considering the
versatility of this derivative we studied its reactivity and report
some of the results obtained. Reaction of 5 with sodium methoxide
in methanol at room temperature gave 2-amino-4,5-dichloro-6-
methoxypyrimidine (9) (Scheme 3), the structure of which was
confirmed by X-ray analysis (Figure 1). This structure corroborated
the unusual displacement of the nitro group by chlorine in com-
pound 5. After introduction of the methoxy group, we were unable
to replace the second chlorine in 9. Neither 2-amino-5-chloro-4,6-
dimethoxypyrimidine, nor 4-azido-5-chloro-6-methoxy analogues
were obtained under quite forcing conditions.
O
N
Cl
POCl3, DMF,
NO2
Cl
NO2
Cl
100 °C, 1 h
N
N
AcHN
H2N
N
8 (38%)
1 (82%)
Scheme 2. Synthesis of 2-amino-4,6-dichloro-5-nitropyrimidine.
the reaction mixture to stand after pouring onto aqueous ammo-
nia, gave a product which was not the expected dichloronitro com-
pound 1, but rather the trichloro compound 4 in which the 2-
amino group was protected with a methylenedimethylamino
group.12 When the reaction mixture was poured into water and al-
lowed to stand, this protecting group was removed and 2-amino-
4,5,6-trichloropyrimidine (5) was obtained. In addition, when
intermediate 4 was stirred in acid, compound 5 was formed in a
moderate yield.
Subsequent experimentation suggested that, for the optimal con-
ditions for nitro group displacement by chloride, the best ratio of 3 to
the reagents, POCl3/DMF, was 1:8:4.4, rather than that suggested
earlier for a similar reaction with other pyrimidines.14 These exper-
iments confirmed that the replacement of the nitro group occurred
during the reaction with the Vilsmeier–Haack reagent rather than
in the subsequent hydrolysis step. Even though this type of aromatic
nucleophilic substitution had been reported previously in the diaz-
otisation of several benzene derivatives,15 to our knowledge, it has
never been described for the halogenation of pyrimidines. Based
on the review of Bunnett and Zahler,15 it seems feasible that, initially
the 4,6-(keto)hydroxy groups in compound 3 are substituted by
chlorine, as expected, and then these two chlorine atoms could facil-
itate subsequent substitution of the 5-nitro group.
Pfleiderer9 managed to prepare 4-alkylamino-2-amino-6-alk-
oxypyrimidines in two steps from 2-amino-4,6-dichloropyrimidine
though the introduction of the alkoxy group in the second stage re-
quired more vigorous conditions than those we tried. Interestingly,
we were able to react 5 with one equivalent of azide (Scheme 3) to
give a product, the IR spectrum of which (Nujol) demonstrated a
peak at 2142 cmÀ1 which suggests that it exists as the azido com-
pound 10a in the solid state.
Having failed to prepare 1 from compound 3, we turned our
attention back to the work of Temple et al.6 These authors reported
compound 1 as a minor impurity without giving experimental de-
tails. Therefore, we were able to optimise the conditions for the
nitration of 6 to give 7, though the product always contained ca.
10% of the starting material. It was impossible to separate the
two compounds as they were both insoluble in suitable organic
solvents. However, acetylation of the mixture gave the correspond-
In solution (DMSO-d6), the 1H NMR and 13C NMR spectra sug-
gest that bicyclic tetrazolo[1,5-c]pyrimidine 10b is the main com-
ponent in an 8:1 mixture with 10a. The latter spectrum, in
particular, is quite different from the other substituted pyrimidines
synthesised, and can be assigned by analogy.16,17
The mono azido compound 10a/b was also prepared by the ac-
tion of nitrous acid on the dichloro hydrazide 11 prepared by us.
OMe
Cl
Cl
Cl
1 eq NaN3,
NaOMe, MeOH,
rt, overnight
Cl
Cl
Cl
Cl
Cl
Cl
N
DMSO, rt, 48 h
N
N
N
N
(58%)
H2N
N
H2N
N
H2N
N
N
N
N
H2N
N
N
9 (95%)
5
N
10a
10b
6 eq NaN3,
DMSO, rt, 48 h
NH2NH2,
rt, 2 h
i) NaNO2, H2O
HCl, 0 °C
ii) 20 °C, 2 h
(56%)
Cl
N
N3
N
N3
N
Cl
Cl
Cl
N
H2N
NH
NH2
H2N
N
N3
H2N
N
N
N
N
11 (84%)
12a/b (67%)
Scheme 3. Reactivity of 2-amino-4,5,6-trichloropyrimidine 5.