Synthesis of (Z)-N-(2-amino-1,2-dicyanovinyl)formamide O-alkyloximes and a study of their cyclization in the presence of base

Article abstract of DOI:10.1039/a901332f

The title compounds (3) have been prepared in high yield by reaction of (Z)-N-(2-amino-1,2-dicyanovinyl)formimidate with an alkoxyamine NH₂OR (R = CH₂Ph, Me). These compounds cyclize to the corresponding 5-amino-4-cyanoimidazoles 4 by reaction with ethanolic NaOH solution. In ethyl acetate and using DBU as a base, amidoximes 3 follow an unexpected cyclization pathway leading to a pyrimidine structure 5. Single-crystal X-ray structures have been obtained for both the amidoxime 3a and the pyrimidine 5a (R = CH₂Ph). A strong intramolecular H-bond in the amidoxime structure, which was identified both in the solid state and in solution, may be responsible for the unusual cyclization pattern observed in these compounds.

Full text of DOI:10.1039/a901332f

View Online / Journal Homepage / Table of Contents for this issue
Synthesis of (Z)-N-(2-amino-1,2-dicyanovinyl)formamide
O-alkyloximes and a study of their cyclization in the
presence of base
Brian L. Booth,a Flora A. T. Costa,b Zahid Mahmood,a Robin G. Pritcharda and
M. Fernanda Proença*b
a Department of Chemistry, UMIST, PO Box 88, Manchester, UK M60 1QD
b Departamento de Química, Universidade do Minho, Campus de Gualtar, 4710 Braga, Portugal
Received (in Cambridge) 17th February 1999, Accepted 17th May 1999
The title compounds (3) have been prepared in high yield by reaction of (Z)-N-(2-amino-1,2-dicyanovinyl)-
formimidate with an alkoxyamine NH2OR (R = CH2Ph, Me). These compounds cyclize to the corresponding
5-amino-4-cyanoimidazoles 4 by reaction with ethanolic NaOH solution. In ethyl acetate and using DBU as a
base, amidoximes 3 follow an unexpected cyclization pathway leading to a pyrimidine structure 5. Single-crystal
X-ray structures have been obtained for both the amidoxime 3a and the pyrimidine 5a (R = CH2Ph). A strong
intramolecular H-bond in the amidoxime structure, which was identified both in the solid state and in solution,
may be responsible for the unusual cyclization pattern observed in these compounds.
N
H
H
EtO
H
H
2
Introduction
RO
i)
A variety of (Z)-N-(2-amino-1,2-dicyanovinyl)formamidines1–6
and formamidrazones7–10 have been previously prepared in our
group, and these compounds have proved to be versatile
reagents in the synthesis of nitrogen heterocycles, including
6-carbamoylpurines and 1,2-dihydropurines, imidazoles,
pyrroles and pyrrolo-triazepines. These results prompted us to
investigate the synthesis of the corresponding O-alkyl-
amidoximes, as they could be used as the starting material for
the synthesis of N-alkoxypurines which are analogues of anti-
cancer and anti-viral agents.11 Considering that the imidazole
ring has been traditionally obtained when the amidines or
amidrazones were treated with base, similar reaction conditions
were used with the amidoximes that were isolated.
N
4
CN
CN
N
CN
CN
+
NH2OR.HCl
2
H2N
H2N
5
1
3a R = CH2Ph
b R = CH3
iii)
ii)
OR
N
N
CN
4
NH2
CN
5
4
2
2
N
N
RO
5
NH2
6
NH
Results and discussion
4a R = CH2Ph
b R = CH3
5a R = CH2Ph
b R = CH3
In a small scale reaction, the commercially available hydro-
chloride salt of O-benzylhydroxylamine was caused to react
with ethyl (Z)-N-(2-amino-1,2-dicyanovinyl)formimidate lead-
ing to diaminomaleonitrile as the only product (as shown by
TLC) after 5 min at room temperature in ethanolic solution.
Neutralisation of the ammonium salt with either one equiv-
alent of DBU or with 1 M aqueous sodium hydroxide solution,
followed by reaction of the O-benzylhydroxylamine with imid-
ate 1 gave the amidoxime 3a in good yield (Scheme 1). When
this reaction is carried out in a small volume of ethanol some
diaminomaleonitrile formation is still observed, but the desired
product 3a can be obtained in 70% isolated yield after chrom-
atography. Reaction in THF is very slow at room temperature
(4 days), but a clean reaction ensues and a 97% isolated yield of
pure 3a can be obtained after simple filtration. The most effect-
ive conditions for the preparation of 3a are to reflux an equi-
molar mixture of 1 and O-benzylhydroxylamine in diethyl
ether. On cooling, the product precipitates from the ether to
give a 97% yield of pure 3a after 3 h. The first procedure was
also used for the synthesis of amidoxime 3b, which was isolated
in 65% yield (Table 1).
iv)
H
OR
OR
OH
N
N
N
N
N
N
v)
vi)
OEt
N
N
N
N
N
CN
NH2
9a R = CH2Ph
NH2
8a R = CH2Ph
10
Scheme 1 Reagents and conditions: i, EtOH, RT, 2 h–2 days; ii, NaOH,
EtOH, RT; iii, DBU, ¼ equiv.; ethyl acetate, RT; iv, Ac2O, CH(OEt)3,
reflux, 16 h; v, NH3(aq), MeOH, RT, 2 h; vi, HBr 32% in AcOH, steam
bath, 3.5 h.
twisted, preventing extended conjugation [bond length N(2)–
C(1) = 1.45
Å and conformation angle C(2)–C(1)–N(2)–
C(10) = Ϫ140Њ]. The phenyl group is oriented perpendicularly
to the plane of the amidine ring [conformation angle O(1)–
C(9)–C(8)–C(3) = Ϫ95Њ] and the N(1)–C(10) and C(10)–N(2)
bonds are both short (1.29 Å and 1.33 Å respectively) which is
indicative of delocalized double bond character. In the solid
state there is strong intramolecular bonding between O(1) and
The structure of amidoxime 3a was confirmed by X-ray
crystallography (Fig. 1). Bond angle and torsion angle analysis
indicate that the amidine and diaminomaleonitrile units are
J. Chem. Soc., Perkin Trans. 1, 1999, 1853–1858
1853

View Online
Table 1 Microanalytical, mp and mass spectroscopic data for compounds 3, 4 and 5
Found (%) (Required)
Compound
(Formula)
Mp/ЊC
(decomp.)
Yield
(%)
C
H
N
m/z (EI)
3a
143–145
70a
97b
97c
65
59.8
(59.8)
4.7
(4.6)
29.2
(29.0)
241 (39.9%, Mϩ)
224 (100)
(C12H11N5O)
3b
141–142
112–115
154–155
200–209
192–194
—
165.0652
(165.0651)f
4.8
(4.7)
4.4
(4.3)
4.7
(4.6)
4.2
(4.2)
5.3
(4.9)
4.5
(4.6)
3.2
(3.3)
165 (16.1%, Mϩ)
59 (100)
(C6H7N5O)
4a
63d
82e
16
61.9
(61.7)
43.5
(43.5)
59.8
(59.8)
43.7
(43.6)
57.7
(58.3)
59.8
(59.8)
39.5
25.9
(26.2)
40.4
(40.6)
28.6
(29.0)
42.0
(42.4)
18.3
(18.4)
28.7
(29.0)
46.1
214 (25.9%, Mϩ)
91 (100)
(C11H10N4O)
4b
138 (40.1%, Mϩ)
107 (100)
(C5H6N4O)
5a
81
90
82
80
75
242 (100%, M ϩ 1ϩ)g
(C12H11N5O)
5b
165 (7.0%, Mϩ)
92 (100)
(C6H7N5O)
8
271 (100%, M ϩ 1ϩ)g
(C14H14N4O2ؒH2O)
9
165–166
330–331
242 (100%, M ϩ 1ϩ)g
152 (M ϩ 1)ϩa
(C12H11N5O)
10
(C5H5N5O)
(39.7)
(46.8)
a Using Procedure 1. b Using Procedure 2. c Using Procedure 3. d Using 2 M solution of KOH in ethanol. e Using a large excess of 1 M aqueous KOH
in solution. f High resolution mass spectroscopy. g Fast atom bombardment.
Fig. 1 X-Ray crystal structure of (Z)-N-(2-amino-1,2-dicyanovinyl)
formamide O-benzyloxime 3a.
the hydrogen on N(2), responsible for a stable Z configuration
of the amidine moiety. The bond angle N(1)–C(10)–N(2) of
125.7Њ is considered a typical feature of the Z configuration in
amidines, in contrast with the E configuration where this bond
angle is less than 124Њ.12 The adjacent angles O(1)–N(1)–C(10)
and C(10)–N(2)–H(2) measured 108.4Њ and 125Њ respectively.
The IR spectra of the amidoximes 3 show intense bands in
Fig. 2 X-Ray crystal structure of N-benzyloxy-5-amino-4-cyano-6-
imino-1,6-dihydropyrimidine 5a.
of 82%. The product precipitates from solution and can be
isolated by simple filtration and washing.
Imidazole 4b is the only product in solution after almost two
days at room temperature, but the solid could only be obtained
after dry flash chromatography. The poor isolated yield (17%)
reflects the difficulty in removing this compound from the
column, even when using a large volume of diethyl ether.
When DBU was added to a solution of amidoximes 3 in ethyl
acetate, a different product immediately precipitated from the
reaction mixture and has been identified as the pyrimidine 5 by
single crystal X-ray structure analysis of 5a. From Fig. 2 it can
be seen that the pyrimidine ring is planar and parallel to the
aromatic ring. The C(6)–N(18) and C(2)–N(3) bonds are short
(1.273 Å and 1.277 Å respectively) indicative of double bond
character. This must be affecting the bond angles in their vicin-
ity as C(5)–C(6)–N(1) measures 112.2Њ, C(6)–N(1)–C(2) meas-
ures 125.5Њ and N(1)–C(6)–N(3) measures 123.3Њ [expected for
an unsubstituted pyrimidine ring: 121.3Њ, 115Њ and 129.7Њ
respectively].13 The IR spectra of all compounds show a
the 3200–3400 cmϪ1 region (ν NH) and both C᎐N groups are
᎐
᎐
visible at 2248 and 2202 cmϪ1 for 3a and 2232 and 2213 cmϪ1 for
1
3b. In the H NMR spectrum, the chemical shift value for the
C(2) proton is δ 6.8 ppm for 3a and 3b, coupling with the
adjacent N–H with coupling constants of 9 Hz and 10.2 Hz
respectively. This result seems to indicate that the Z-anti con-
figuration is maintained in DMSO solution, in contrast with
the analogous N-aryl amidines where the E-syn configuration
is preferred [C(2)–H appears in the range δ 7.9–8.3 ppm and
J CH–NH≅5 Hz].
The amidoximes 3a and 3b cyclized in the presence of
ethanolic potassium or sodium hydroxide leading to the corre-
sponding 5-amino-4-cyanoimidazoles 4. Imidazole 4a is formed
after only 15 min at room temperature, precipitating out of the
reaction mixture in 63% yield. The use of a large excess of 1 M
aqueous KOH in this reaction affords 4a in an improved yield
1854
J. Chem. Soc., Perkin Trans. 1, 1999, 1853–1858

View Online
Table 2 1H NMR spectroscopic data for compounds 3, 4 and 5
Compound
δ([2H6]Me2SO; 300 MHz)
3a
3b
4a
4b
5a
8.1 (d, J = 9 Hz, 1H, NH), 7.05 (s, 2H, NH2), 6.8 (d, J = 9 Hz, 1H, CH), 5.0 (s, 2H, CH2), 7.2–7.4 (m, 5H, Ph)
8.1 (d, J = 10 Hz, 1H, NH), 7.0 (s, 2H, NH2), 6.8 (d, J = 10 Hz, CH), 3.7 (s, 3H, CH3)
7.28 (s, 1H, CH), 6.56 (s, 2H, NH2), 5.14 (s, 2H, CH2), 7.4–7.51 (m, 5H, Ph)
7.58 (s, 1H, CH), 6.60 (s, 2H, NH2), 3.93 (s, 3H, CH3)
A: 7.7 (s, 1H, CH), 7.5 (s, 1H, NH), 6.6 (s, 2H, NH2), 5.1 (s, 2H, CH2), 7.3–7.5 (m, 5H, Ar)
B: 8.1 (s, 1H, NH), 7.6 (s, 1H, CH), 6.9 (s, 2H, NH2), 5.2 (s, 2H, CH2), 7.3–7.5 (m, 5H, Ar)
A:B, 3:1 ratio
5b
A: 7.9 (s, 1H, CH), 7.6 (s, 1H, NH), 6.6, (s, 2H, NH2), 3.9 (s, 3H, CH3)
B: 8.0 (s, 1H, NH), 7.8 (s, 1H, CH), 6.9 (s, 2H, NH2), 3.9 (s, 3H, CH3)
A:B, 5:1 ratio
8
9
8.3 (s, 1H, CH), 7.95 (s, 1H, CH), 7.36–7.45 (m, 5H, Ph), 5.22 (s, 2H, CH2), 4.28 (q, 2H, J = 7 Hz, CH2), 1.28 (t, 3H, CH3)
8.25 (s, 1H, CH), 8.15 (s, 1H, CH), 7.25–7.38 (m, 5H, Ph), 7.2 (br s, 2H, NH2), 5.4 (s, 2H, CH2)
Table 3 13C NMR spectroscopic data for compounds 3, 4 and 5
δC([2H6]Me2SO; 75 MHz)
Compound
C(2)
C(4)
C(5)
C(6)
CN
R
3a
3b
140.2
140.1
94.7
95.7
120.1
120.7
—
—
116.5
114.8
115.8
117.4
117.2
117.0
117.1
118.2
119.6
—
74.4 (CH2), 138.3, 128.5, 127.7, 127.5
62.0 (CH3)
4a
4b
5aa
5bb
8
128.5
127.2
136.5
136.8
146.3
158.3
86.4
86.1
97.5
98.6
135.9
150.4
144.3
143.7
142.7
144.1
136.4
120.1
—
—
144.7
144.4
—
80.8 (CH2), 133.6, 130.6, 129.9, 129.0
67.1 (CH3)
78.4 (CH2), 130.2, 129.4, 128.7, 133.4
66.1 (CH3)
18.2 (CH3), 67.9 (CH2), 85.7, (OCH2), 133.9, 134.0, 134.2, 165.5 (C᎐N)
85.8 (CH2), 133.7, 134.5, 134.8, 141.9, 150.0
᎐
9
161.3
a Only one set of bands is visible in the spectrum. b The data recorded belongs to the most abundant species, A. For B, the only signal which is visible
in the spectrum corresponds to C(2) at δ 138.4 ppm.
medium–strong band at 2210 cmϪ1 for the C᎐N stretching
identified as the imidazole14 led to a complex mixture. 1H NMR
spectroscopic examination of the oil showed the presence of the
bands assigned by Watson to the imidazole 4a, clearly indicat-
ing that this is a different compound.
᎐
᎐
vibration. The corresponding carbon atom can be identified in
the 13C NMR at 117.1 (5a) and 118.2 (5b) ppm. Only one set of
bands is visible in the 13C NMR spectrum of each compound,
1
1
which is in contrast with the data registered in the H NMR,
The H NMR data reported by Watson indicate a C(2)–H
where two sets of bands are always present in a 3:1 ratio (5a)
and 5:1 ratio (5b).
The structure of the 5-amino-4-cyano imidazoles 4 was
proton at δ 6.84 and NH2 protons at δ 5.00 ppm (cf. δ 7.28
and 6.56 ppm respectively for compound 4a). In earlier work1
we have described a large number of substituted 5-amino-4-
cyanoimidazole derivatives and in all cases the C(2)–H proton
of the imidazole ring appears within a relatively narrow range
of δ 7.15 to 7.65 ppm. A chemical shift of δ 6.84 ppm is
unprecedented. In contrast, the CHN2 proton of the amidines
3a and 3b, which appear in the NMR spectrum at δ 6.80 ppm
(see Table 2) agree well with the figure quoted by Watson.14 For
this reason, we feel that the most plausible structure for the
compound isolated by Watson is the amidine 4c, i.e. an isomer
assigned on the basis of elemental analysis and spectroscopic
᎐
data. The C᎐N stretching vibration is a typical feature in the IR
᎐
spectrum of these compounds and corresponds to an intense
band at 2215 cmϪ1 (4a) and 2211 cmϪ1 (4b). A strong band at
around 1650 cmϪ1 is visible in both spectra and may be assigned
to the C᎐N stretching vibration. In the 1H NMR spectrum, the
᎐
NH2 group always shows up as a singlet in the δ 6.5–6.6 ppm
region and this seems to be a typical feature of the 5-amino-4-
cyanoimidazoles that have been prepared (see Table 2).1,2,4,10
The C–H proton gives a sharp singlet at δ 7.28 (4a) and δ 7.58
(4b) ppm. All the bands in the 13C NMR are sharp, with the
signals for C(4) and C(5) around δ 86 and 144 ppm respectively
(see Table 3).
The synthesis of imidazole 4a has been previously reported
by Watson14 from the reaction of aminomalononitrile toluene-
p-sulfonate and triethyl orthoformate to form the imino ether
tosylate, which on reflux with an equivalent of benzyloxyamine
in ether gave the product as white crystals in only 19% yield.
Considering that both the melting point and spectroscopic data
for this compound were different from the values now reported,
the experimental procedure described in the literature was care-
fully reproduced. In the complex reaction mixture it was pos-
sible to identify the presence of imidazole 4a (by TLC) and this
compound was selectively isolated in less than 1% yield. The 1H
NMR and IR data on this compound, compare exactly with the
values that were obtained in this present work for the solid
isolated in the cyclization of amidine 3a by alcoholic sodium
hydroxide solution. Attempts to isolate the product previously
OCH2Ph
N
OCH2Ph
N
NH2
N
H
H
H
CN
CN
N
H
N
CN
H
4c
11
and precursor to the imidazole 4a. The signal attributed by him
to an NH2 group could arise by the accidental equivalence of
the CH(CN)2 proton and the NH proton. Such an amidine
might be expected to tautomerise in solution and coupling
between the CH and NH protons might not be observed.
The different behaviour of the amidines 3a and b in the pres-
ence of KOH and DBU requires explanation. In the presence of
a relatively strong base KOH in ethanol or water there is good
evidence to indicate that the amidines 3 are deprotonated to
form an anion. So, for example, when NaOH (2 × 10Ϫ2 mol
dmϪ3) in ethanol was added to a solution of amidine 3a
J. Chem. Soc., Perkin Trans. 1, 1999, 1853–1858
1855

View Online
Fig. 3 Changes in the UV spectrum of amidine 3a (1.53 × 10Ϫ5 mol
dmϪ3 in ethanol, 2 cm3) by addition of 0.25 cm3 NaOH (2 × 10Ϫ2 mol
dmϪ3) in ethanol: (a) 3a in the absence of base, (b) immediately after the
addition of NaOH in ethanol, (b–j) spectra recorded at an interval of
10 min, (k) spectrum recorded after 24 h.
Fig. 4 Changes in the UV spectrum of amidine 3a (4.15 × 10Ϫ5 mol
dmϪ3 in ethyl acetate, 2 cm3) by addition of DBU (10 µl): (a) 3a in the
absence of DBU, (b) immediately after the addition of 10 µl of DBU,
(b–g) spectra recorded at an interval of 60 s, (h) spectrum recorded 300
s after (g), (h–m) spectra recorded at an interval of 200 s.
(1.53 × 10Ϫ5 mol dmϪ3) in ethanol (2 ml) the UV absorption
band with a maximum at λmax (EtOH)/nm 320 (ε/dm3 molϪ1
cmϪ1 17 400) for the amidine rapidly shifts to λmax (EtOH)/nm
360 for the anion. This absorption band gradually loses inten-
sity and after 24 h at room temperature the only UV signal
visible is that of the imidazole 4a λmax (EtOH)/nm 240 (ε/dm3
molϪ1 cmϪ1 13 800) (see Fig. 3). In a separate experiment it has
been shown that the band at 360 nm for the anion reverts to the
initial amidine signal at 320 nm upon addition of hydrochloric
acid. It is clear from this experiment that the formation of the
anion is both rapid and quantitative in the presence of HOϪ
ion, and that cyclisation to the unstable imidazole 7 and, hence,
the isolated product 4a is a much slower reaction. It is likely
that the anion adopts the conformation shown in Scheme 2, in
(ε/dm3 molϪ1 cmϪ1 8 000), 280 (5 200), 345 (13 800)] were the
only ones observed. There was no evidence for a band at 360
nm for the anion. The fact that cyclisation to an imidazole does
not occur in this case with DBU is a little surprising as DBU
does still catalyse cyclisation of amidines having N-alkyl,
-benzyl and -aryl substituents to form imidazoles of type 7 in
high yields.1,2,4,10 Perhaps, the difference of the O-alkyl substitu-
ent increases the basicity of amidine or the difference may be
associated with the strong intramolecular H-bonding between
the N–H and the O atom observed in the X-ray structure of the
amidine. The formation of pyrimidine 5a requires rotation
about the C᎐C bond of the amidine as shown in Scheme 3 and
᎐
it is not clear what role, if any, the DBU plays in accelerating
this rotation. The DBU may also catalyse tautomerism in the
amidine and this may play a part in determining this unusual
outcome (see Scheme 3). Further work is being carried out to
establish the mechanism of pyrimidine formation.
N
H
PhCH2O
N
CN
CN
H
H2N
N
H
H
H
H
OCH2Ph
H
OCH2Ph
N H
PhCH2O
3
N
N
CN
CN
DBU
NaOH
HCl
N
CN
H2N
NC
N
EtOAc
H2N
OCH2Ph
CN
H2N
CN
3
N
–
H
N
H
PhCH2O
rotation
N
–
N
CN
CN
C
N
H
OCH2Ph
H
N
N
CN
H2N
NC
NH2
N
C
slow
N
N
6
PhCH2O
NH2
NH
H2N
CN
5a
OCH2Ph
OCH2Ph
NH2
NH2
Scheme 3
N
N
4
N
N
7
NH
Watson14 converted his supposed 4a into 9-hydroxyadenine
10 via a two step procedure involving heating with triethyl
orthoformate, followed by treatment with ammonia. He identi-
fied the product of this reaction as 9-benzyloxyadenine 9a,
which was then debenzylated with 32% aqueous hydrogen
bromide in glacial acetic acid to give 9-hydroxyadenine. We
have repeated this sequence with our compound 4a, and after
treatment with triethyl orthoformate in glacial acetic acid the
imidate intermediate 8a was isolated and fully characterised,
before treatment with ammonia to produce 9a in 80% yield. The
1H NMR spectrum of 9a showed some significant differences
from those reported by Watson.14 In particular this can be
observed in the chemical shift values of the 8-H proton and the
NH2 protons: [δH reported14 (60 MHz, d6-DMSO) 8.40 (s, 1H,
2-H), 7.40 (s, 1H, 8-H), 7.33 (s, 5H, Ph), 6.63 (s, 2 H, NH2),
5.37 (s, 2 H, OCH2). δH our work (300 MHz, d6-DMSO) 8.25 (s,
CN
CN
Scheme 2
which the lone pair of electrons on the nitrogen atoms are in an
anti-position to minimise repulsion.
In the presence of the weaker base DBU in ethyl acetate or
diethyl ether, the concentration of the anion will be very low
and cyclisation to an imidazole will be very slow. This has been
demonstrated by addition of DBU (10 µl) to a solution of
amidine 3a (4.15 × 10Ϫ5 mol dmϪ3) in ethyl acetate (2 ml). The
UV absorption band with a maximum at λmax (EtOAc)/nm 310
(ε/dm3 molϪ1 cmϪ1 17 400) for the amidine gradually reduces in
intensity (see Fig. 4) and, after 30 min at room temperature
absorption bands for the pyrimidine [λmax (EtOAc)/nm 328
1856
J. Chem. Soc., Perkin Trans. 1, 1999, 1853–1858

View Online
Table 4 Crystal data and details of refinement for compounds 3a and
ture factor calculation in idealized positions, and were
5a
assigned isotropic thermal parameters which were 20%
greater than the equivalent B value of the atom to which they
were bonded.
Compound
3a
5a
CCDC reference number 207/334.
See http://www.rsc.org/suppdata/p1/1999/1853 for crystallo-
graphic files in .cif format.
Formula
M
C12H11N5O
241.26
tetragonal
0.40 × 0.15 × 0.15
293 K
C12H11N5O
241.26
monoclinic
0.35 × 0.25 × 0.20
293 K
Crystal system
Crystal size/mm
Temperature of
data collection
Space group
a/Å
Synthesis of (Z)-N-(2-amino-1,2-dicyanovinyl)formamide
O-benzyloxime (3a)
P421c
19.028(4)
P21
Procedure 1. A suspension of O-benzylhydroxylamine (1.02
g, 6.39 mmol) and DBU (156 µl, 6.39 mmol) in ethanol was
stirred at room temperature until a homogeneous solution was
obtained. The solvent was concentrated to low volume in the
rotary evaporator and ethyl N-[(Z)-2-amino-1,2-dicyanovinyl]-
formimidate (1.05 g, 6.39 mmol) was added. The mixture was
stirred at room temperature for 2 h until TLC indicated that
all the imidate had been consumed. The title compound was
isolated by dry flash chromatography using diethyl ether as
eluant (1.08 g, 4.48 mmol, 70%).
8.3838(10)
6.6380(10)
10.580(2)
93.35(2)
587.8(2)
2
1.363
252
0.94
0.71069
2–25Њ
b/Å
c/Å
7.014(3)
β/Њ
U/Å3
2540(1)
8
1.262
1008
0.81
0.71069
2–25Њ
Z
Dc/g cmϪ3
F(000)
µ/cmϪ1
λ/Å
θ range for data
collection/Њ
Index ranges
0 ≤ h ≤ 22
0 ≤ k ≤ 22
0 ≤ l ≤ 8
1291
Ϫ7 ≤ h ≤ 9
0 ≤ k ≤ 7
Ϫ12 ≤ l ≤ 12
2162
Procedure 2. O-Benzylhydroxylamine hydrochloride (2.6 g,
16.2 mmol) was neutralised with 1 M aqueous sodium hydrox-
ide and the solution was then extracted with diethyl ether. The
extract was dried (MgSO4) and filtered and the solution was
then added dropwise to a refluxing solution of ethyl (Z)-N-(2-
amino-1,2-dicyanovinyl)formimidate (2.7 g, 16.2 mmol) in
diethyl ether. After complete addition, reflux was continued for
3 h then on cooling the title compound precipitated as an analyt-
ically pure solid (3.8 g, 15.7 mmol, 97%).
Measured reflections
Independent reflections
1291
1132
(Rint = 0.019)
R1 = 0.034
wR2 = 0.106
R1 = 0.043
wR2 = 0.115
Final R [I > 2σ(I)]
R (all data)
R1 = 0.047
wR2 = 0.128
R1 = 0.200
wR2 = 0.187
Procedure 3. O-Benzylhydroxylamine (2.0 g, 16.2 mmol) and
the formimidate (2.7 g, 16.2 mmol) were stirred at room tem-
perature in dry THF (20 ml) until TLC (EtOAc–hexane, 1:1)
showed complete reaction (4 days). Removal of the solvent and
addition of diethyl ether gave the product (3.79 g, 15.7 mmol,
97%).
1H, 2-H), 8.15 (s, 1H, 8-H), 7.25–7.38 (m, 5H, Ph), 7.20 (br s,
2H, NH2), 5.37 (s, 2H, OCH2) ppm]. In our experience the H
1
NMR spectra of adenine derivatives invariably show the 8-H
proton in the range of δ 8.00 to 8.40 ppm, with the 2-H proton
at lower field. Watson’s value of δ 7.40 is more typical of an
imidazole derivative (vide supra). These differences can be
understood if Watson’s 9-benzoyladenine is, in fact, the inter-
mediate 11 and the band he assigns to 2-H, δ 8.40 ppm, is then
attributable to the formamidine proton. This structure would
also explain the observed differences in the NH2 protons. It is
noticeable, that our compound 9 and Watson’s compound have
the same mp, and perhaps cyclisation occurs to give 9 on
heating.
Synthesis of (Z)-N-(2-amino-1,2-dicyanovinyl)formamide
O-methyloxime (3b)
Using Procedure 1, O-methylhydroxylamine hydrochloride
(0.38 g, 4.60 mmol) was neutralised with DBU in ethanol
and, after reduction in the volume, caused to react with the
formimidate (0.75 g, 4.60 mmol) to give, after 2 days, the title
compound (0.49 g, 3.0 mmol, 65%) isolated after flash chrom-
atography.
Debenzylation of either 9a or, what we now believe to be 11,
result in the same compound, namely 9-hydroxyadenine imply-
ing that heating under strong acid conditions results in initial
cyclisation of 11 to 9a, prior to debenzylation.
Synthesis of 5-amino-4-cyano-1-benzyloxyimidazole (4a)
Procedure 1. The amidoxime 3a (2.43g, 10 mmol) was solubil-
ized in a 2 M solution of KOH in ethanol (20 cm3) and the
mixture was stirred at room temperature for 15 min. The
brownish solid that precipitated from solution was filtered and
washed with water and diethyl ether. The cream solid was iden-
tified as the title compound (1.36 g, 6.3 mmol, 63%) by elemental
analysis and spectroscopic data.
Experimental
Ethyl (Z)-N-(2-amino-1,2-dicyanovinyl)formimidate 1 used in
this work was prepared by a procedure described previously.15
IR spectra were recorded on a Perkin-Elmer 1600 series
FTIR spectrometer, 1H and 13C NMR spectra on a Bruker XL
300 spectrometer and mass spectra on a GC-MS Automass 120
or on a Kratos Concept instrument.
Procedure 2. The amidoxime 3a (3.0 g, 12.5 mmol) was added
to an excess of 1 M aqueous KOH solution (120 ml) and after 5
min, the title compound precipitated as a white solid (2.18 g,
10.18 mmol, 74%) which was filtered, washed with diethyl ether
and dried under vacuum.
Crystallography
Crystal data and refinement details for compounds 3a and 5a
are presented in Table 4. Both crystals were mounted on a
glass fibre. All measurements were made on a Rigaku AFC6S
diffractometer employing graphite monochromated Mo-Kα
radiation. The data were collected at a temperature of
23 1 ЊC using the ω–2θ scanning technique to a maximum
2θ value of 50.0Њ. The structures were solved by direct
methods using MITHRIL16 and refined using DIRDIF.17 The
non-hydrogen atoms were refined anisotropically. Hydrogen
atoms were refined isotropically or were included in the struc-
Synthesis of 5-amino-4-cyano-1-methoxyimidazole (4b)
The amidoxime 3b (0.34 g, 2.06 mmol) was solubilized in a 0.18
M solution of NaOH in ethanol (20 cm3) and the solution was
stirred at 4 ЊC for 17 hours and then at room temperature for 24
hours. The solid product was isolated by dry flash chrom-
atography using diethyl ether as eluant. The cream solid was
J. Chem. Soc., Perkin Trans. 1, 1999, 1853–1858
1857

View Online
identified as the title compound (0.05 g, 0.36 mmol, 17%) by
pound was filtered off and was washed with several portions of
elemental analysis and spectroscopic data.
diethyl ether. The solid was dissolved in hot dilute aqueous
ammonia solution (24 ml), treated with charcoal, filtered, and
then addition of glacial acetic acid precipitated the product as
white crystals (0.09 g, 75%), which were collected, washed with
water, ethanol and diethyl ether, before drying under vacuum.
Synthesis of 5-amino-1-benzyloxy-4-cyano-6-imino-1,6-dihydro-
pyrimidine (5a)
Procedure 1. DBU (53 µl, 0.35 mmol) was added to a solution
of the amidoxime 3a (0.25 g, 1.04 mmol) in ethyl acetate (2 ml)
and the solution was stirred at room temperature. After ca. 3
min, a cream solid started to precipitate out of solution and the
suspension was stirred for a further 15 min, after which the
solid was filtered and washed with diethyl ether to give 5a (0.21
g, 0.84 mmol, 80%).
Acknowledgements
Thanks are due to PRAXIS XXI for a studentship award to
Flora Costa (PRAXIS XXI/BD/5135/95) and to the University
of Minho and the Foundation for Science and Technology,
Portugal, for financial support. Thanks are also due to the ORS
and the Charles Wallace Foundation for an award to Zahid
Mahmood.
Procedure 2. A suspension of 3a (3.0 g, 12.5 mmol) in diethyl
ether (50 ml) was heated to reflux temperature and when the
solution became homogeneous DBU (10 µl) was added drop-
wise. After 10 min the title compound precipitated as a white
solid (1.48 g, 6.16 mmol, 74%) and was washed with cold
diethyl ether and dried.
References
1 M. J. Alves, B. L. Booth, O. K. Al-Duaij, P. Eastwood, L. Nezhat,
M. F. Proença and A. S. Ramos, J. Chem. Res. (S), 1993, 402;
J. Chem. Res. (M), 1993, 2701.
2 B. L. Booth, A. M. Dias and M. F. Proença, J. Chem. Soc., Perkin
Trans. 1, 1992, 2119.
3 M. J. Alves, B. L. Booth, P. Eastwood, R. G. Pritchard and M. F.
Proença, J. Chem. Soc., Chem. Commun., 1993, 834.
4 M. J. Alves, B. L. Booth and M. F. Proença, J. Heterocyc. Chem.,
1994, 31, 345.
5 M. J. Alves, O. K. Al-Duaij, B. L. Booth, M. A. Carvalho, P. R.
Eastwood and M. F. Proença, J. Chem. Soc., Perkin Trans. 1, 1994,
3571.
6 M. J. Alves, B. L. Booth, M. A. Carvalho, R. G. Pritchard and
M. F. Proença, J. Heterocycl. Chem., 1997, 34, 739.
7 M. J. Alves, B. L. Booth, A. P. Freitas and M. F. Proença, J. Chem.
Soc., Perkin Trans. 1, 1992, 913.
Synthesis of 5-amino-1-methoxy-4-cyano-6-imino-1,6-dihydro-
pyrimidine (5b)
DBU (50 µl, 0.30 mmol) was added to a solution of the
amidoxime 3b (0.15 g, 0.90 mmol) in ethyl acetate (1 ml) and the
solution was stirred at room temperature. After ca. 3 min, a
cream solid started to precipitate out of solution and the sus-
pension was stirred for a further 15 min, after which the solid
was filtered and washed with diethyl ether to give 5b (0.13 g,
0.80 mmol, 90%).
Synthesis of 1-benzyloxy-4-cyano-5-ethoxyformimidoyl-
imidazole (8a)
8 B. L. Booth, A. P. Freitas and M. F. Proença, J. Chem. Res. (S),
1993, 260; J. Chem. Res. (M), 1993, 1640.
9 B. L. Booth, A. P. Freitas and M. F. Proença, J. Heterocycl. Chem.,
1995, 32, 457.
10 M. J. Alves, B. L. Booth, A. P. Freitas and M. F. Proença, J. Chem.
Soc., Perkin Trans. 1, 1992, 913.
11 (a) M. R. Harden, C. A. Parkin and P. G. Wyatt, Tetrahedron Lett.,
1988, 29, 701; (b) M. R. Harden and R. L. Jarvest, J. Chem. Soc.,
Perkin Trans. 1, 1990, 1705.
12 T. M. Krygowski and K. Wozniak, in The Chemistry of Amidines
and Imidates, ed. S. Patai and Z. Rappoport, John Wiley & Sons,
New York, 1991, vol. 2, p. 143.
Acetic anhydride (1.3 ml, 2.34 mmol) and triethyl orthoformate
(4.66 ml, 2.3 mmol) were added to 4a (0.5 g, 2.34 mmol) and
the mixture was heated under reflux for 16 h. Removal of the
volatiles on a rotary evaporator followed by addition of hexane
to the residue gave the title compound as a white solid (0.52 g,
1.93 mmol, 82%).
Synthesis of 9-benzyloxyadenine (9a)
13 D. J. Brown, in Comprehensive Heterocyclic Chemistry, ed. A. J.
Boulton and A. McKillop, Pergamon Press, New York, 1984, vol. 3,
p. 59.
14 A. A. Watson, J. Org. Chem., 1977, 42, 1610.
15 D. W. Woodward, USP 2 534 331/1950.
16 C. J. Gilmore, MITHRIL—an integrated direct methods computer
program, J. Appl. Crystallogr., 1984, 17, 42.
Aqueous ammonia solution (35 µl, 1.85 mmol) was added
dropwise to a solution of 8a (0.5 g, 1.85 mmol) in methanol (10
ml) and the mixture was stirred for 2 h, when the title compound
(0.36 g, 1.49 mmol, 80%) precipitated out and was filtered,
washed with diethyl ether, and dried.
17 P. T. Beurskens, DIRDIF: direct methods for difference structures—
an automatic procedure for phase extension and refinement of
difference structure factors. Technical report 1984/1 Crystallography
Laboratory, Toernooiveld, 6525 Ed Nijmegan, Netherlands.
Synthesis of 9-hydroxyadenine (10)
Compound 9 (0.2 g, 0.83 mmol) was added to 32% hydrogen
bromide in glacial acetic acid (4 ml) and the solution was
warmed on a steam bath for 3.5 h. The reaction mixture was
then cooled and the hydrogen bromide salt of the title com-
Paper 9/01332F
1858
J. Chem. Soc., Perkin Trans. 1, 1999, 1853–1858