Differentiation between [1,2,4]triazolo[1,5-a] pyrimidine and [1,2,4]triazolo[4,3-a]- Pyrimidine regioisomers by 1H-15N HMBC experiments

Article abstract of DOI:10.1002/mrc.2634

The condensation of malonoaldehyde derivatives with either a 3-amino-[1,2,4]-triazole or a 3,5-diamino-[1,2,4]-triazole precursor was studied. In agreement with previous reports, two different bicycles, namely, bearing the regioisomeric [1,2,4]triazolo[1,

Full text of DOI:10.1002/mrc.2634

Research Article
Received: 12 April 2010
Revised: 17 May 2010
Accepted: 20 May 2010
Published online in Wiley Interscience: 29 June 2010
(www.interscience.com) DOI 10.1002/mrc.2634
Differentiation between [1,2,4]triazolo[1,5-a]
pyrimidine and [1,2,4]triazolo[4,3-a]-
pyrimidine regioisomers by 1H–15N HMBC
experiments
Antonio Salgado,∗ Carmen Varela, Ana María García Collazo and
Paolo Pevarello
The condensation of malonoaldehyde derivatives with either a 3-amino-[1,2,4]-triazole or a 3,5-diamino-[1,2,4]-triazole
precursor was studied. In agreement with previous reports, two different bicycles, namely, bearing the regioisomeric
[1,2,4]triazolo[1,5-a]pyrimidine (1) or[1,2,4] triazolo [4,3-a]pyrimidine (2) structural surrogates, could be obtained. We found
that, depending on the triazole precursor, only one regioisomer resulted, either of the 1 or 2 series. We also observed that these
two structural surrogates could be unambiguously differentiated by indirectly measuring their 15N chemical shifts by 1H–15
N
HMBC experiments. The occasional conversion of [1,2,4]triazolo[4,3-a]pyrimidines to the [1,2,4]triazolo[1,5-a]pyrimidine
counterparts could be unequivocally determined by 15N NMR data. Copyright ꢀ 2010 John Wiley & Sons, Ltd.
c
Keywords: NMR; 1H; 13C; 15N; HMBC; [1,2,4]triazolo[1,5-a]pyrimidines; [1,2,4]triazolo[4,3-a]pyrimidines; regioisomerism; isomerisation
Introduction
CHO
Y
N
N
N
N
N
N
CHO
X
H2N
HN
[1,2,4]-Triazolo[1,5-a]pyrimidines (1, Fig. 1) can be regarded as
the aza-analogues of purines.[1] They have received a growing
interest in the past years due to their important pharmaceutical
properties.[1–5] For instance, the anti-ischaemic drug Trapidil[2]
and the recently isolated antibiotic Essramycin[3] have been found
to have the [1,2,4]-triazolo[1,5-a]pyrimidine parent structure (1).
Most of the reported syntheses of these compounds are based on
thereactionofa1,3-divalentsynthonwitheithera3-amino-[1,2,4]-
triazole[6] or a 2-hydrazinopyrimidine precursor.[7] In both cases,
it is commonly accepted that the [1,2,4]triazolo[4,3-a]pyrimidine
system (2) is formed first, but isomerises to the thermodynamically
more stable [1,5-a] isomer (1), either thermally or under acidic
or alkaline conditions. This reaction, known as the Dimroth
rearrangement, has been reported to occur via ring opening of
the six-membered cycle in a formal [3,5] sigmatropic shift.[8–10]
It should be noted that less work has been devoted to [1,2,4]-
triazolo[4,3-a]pyrimidines (2), perhaps due to their intrinsic lower
stability.[11–17]
and/or
X
N
N
N
N
Y
Y
X
X = H, NHR
1 (X = H)
(Y = H, NHR, Br)
2 (X H, NHR)
(Y = H, NHR, Br)
Scheme 1. Condensation between malonoaldehydes and 3-amino-[1,2,4]-
triazole or 3,5-diamino-[1,2,4]-triazole precursors.[1]
always a single compound was obtained. Yet, some doubt arose
when it came to distinguishing between the two possible parent
structures 1 and 2. We found that one- and two-dimensional 1H
and 13C chemical shift data would be difficult to assess, as 1H–13
C
HMBC experiments for comparatively small cyclic systems can
be rather deceitful. Furthermore, as 1H and 13C chemical shifts
strongly depend on the substituents attached to the core scaffold,
these data can be sometimes complicated to interpret. On top of
that, in fused, rigid aromatic systems like the title compounds,
nuclear Overhauser effect (NOE) experiments may not afford
conclusive results for those bearing few hydrogen atoms. More
useful NMR experiments would be those that yield information
about carbon–carbon or even carbon–nitrogen connectivities.
These experiments are of very low sensitivity (INADEQUATE) or
present processing problems, such as covariance methods.[19,20]
Results and Discussion
As part of our medicinal chemistry programme, we decided
to prepare compounds of the parent structure [1,2,4]triazolo[4,
3-a]pyrimidine (2). It had been described previously that direct
condensation of malonoaldehyde derivatives with either a 3-
amino-[1,2,4]-triazole or a 3,5-diamino-[1,2,4]-triazole precursor
invariably gives a mixture of the two possible regioisomers 1 and
2 (Scheme 1).[8,13,18]
∗
Correspondence to: Antonio Salgado, Department of Medicinal Chemistry,
Centro Nacional de Investigaciones Oncologicas (CNIO), Melchor Fernandez
Almagro 3, 28029 Madrid, Spain. E-mail: asalgado@cnio.es
´
´
When we attempted this condensation with either mono- or
diaminotriazoles and malonate-type compounds, we found that
Department of Medicinal Chemistry, Centro Nacional de Investigaciones
´
´
Oncologicas (CNIO), Melchor Fernandez Almagro 3, 28029 Madrid, Spain
c
Magn. Reson. Chem. 2010, 48, 614–622
Copyright ꢀ 2010 John Wiley & Sons, Ltd.

Regioisomer differentiation by 1H–15N HMBC experiments
8
R7 8 N
4 N 3
N
1
N
1N
N2
large range (up to 1500 ppm). With the advent of high-field mag-
nets, cryoprobes and the use of pulsed gradients for coherence
selection, indirect 15N detection methods with natural isotopic
abundance have become more common.[22] It is interesting to
note that some regioisomerism problems have been resolved via
1H–15N HMBC.[24–30]
8a
8a
R7
R6
R2
4 N
N
R6
R3
R5
R5
2
1
In order to determine whether our compounds have
the [1,2,4]triazolo[1,5-a]pyrimidine (1) or the [1,2,4]triazolo[4,
3-a]pyrimidine (2) structure, we decided to obtain their 15N
NMR data from 1H–15N HMBC experiments. 15N NMR data
of [1,2,4]triazolo[1,5-a]pyrimidine (1) derivatives have been
reported,[1,8,31,32] but to the best of our knowledge, no 15N NMR
data of [1,2,4]triazolo[4,3-a]pyrimidine derivatives other than the
bare compound 2a have been published.[31]
The structures of the prepared compounds are shown in Fig. 2.
Compound 1a was synthesised as described by Breitmaier et al.,[8]
and compound 2a was prepared by following the method of
Paudler[33] with minor modifications. Compound 1b had already
been described by Rusinov et al.[34] All other compounds are new
and gave satisfactory mass spectrometry (MS) and NMR data.
The obtained1H and 13C NMR data are shown in Tables 1 and 2.
HSQC, HMBC and NOE experiments were run when appropriate to
confirm the structures.
Figure 1. Parent structures of 1,2,4-triazolo[1,5-a]pyrimidines (1) and 1,2,
4-triazolo[4,3-a]pyrimidines (2).
In these cases, it may become mandatory to resort to X-ray crystal
analyses, which may not always be feasible.
Some differentiation studies between [1,2,4]triazolo[4,
3-a]pyrimidines and [1,2,4]triazolo[1,5-a]pyrimidines had been
carried out in as early as 1975.[12] The authors based
their structural analysis on 13C NMR data stating that C-
2
in the [1,5-a] series (1) should be more deshielded
than C-3 in the [4,3-a] analogues (2). Other authors dis-
tinctly identified [1,2,4]-triazolo[1,5-a]pyrimidines and [1,2,4]-
triazolo[4,3-a]pyrimidines from their 1H data.[13] In an-
other study, the structures of 2-methyl-[1,2,4]triazolo[1,
5-a]pyrimidine and 3-methyl-[1,2,4]triazolo[4,3-a]pyrimidine were
distinguished on the basis of their 13C NMR data.[21] Again, the
C-3 carbon of 3-methyl-[1,2,4]triazolo[4,3-a]pyrimidine showed
As it can be seen from Tables 1 and 2, 1H and 13C NMR data do
not really allow us to clearly distinguish between both isomers, as
chemical shifts can substantially vary (up to 30 ppm in 13C NMR)
depending on the substituents. Nevertheless, some regularity can
a
lower frequency than C-2 of 2-methyl-[1,2,4]triazolo[1,
5-a]pyrimidine. Nevertheless, these authors pointed out that iso-
mer differentiation on the basis of spectroscopic data was by no
means straightforward and that in many cases disagreement had
been encountered among the published data.
15N NMR spectroscopy, in particular long-range 1H–15N het-
eronuclear multiple-bond correlation (1H–15N HMBC) experi-
ments, can constitute a useful alternative.[22,23] Unlike the 13C
counterparts, these experiments have been rarely employed due
to the low gyromagnetic ratio and natural abundance of the 15N
nuclide. Yet, despite these shortcomings, 15N NMR spectroscopy
has the advantage that 15N chemical shifts span over quite a
1
be seen with the H NMR chemical shifts in both regioisomeric
series. For compounds 1c and 1d, the presence of an amine
at C-6 shields H-5 as compared to the parent compound 1a
(1.03 ppm difference for 1d). For the 6-bromo derivative 1b, H-5
also resonates at higher field than 1a (0.45 ppm difference). The
chemicalshiftofH-2andH-6hydrogenatomsofcompounds1b–d
varies upon substitution at C-6 but to a lesser extent than H-5. The
presence of a second substituent at C-5 (1e) shields H-2 and H-7.
In the [1,2,4]triazolo[4,3-a]pyrimidine series, the amino group at
N
(a)
N
(b)
N
N
N
N
R2
N
N
R6
R6
R3
R5
compound
compound
2a
R2
H
R5
R3
R6
H
R6
H
1a
H
H
1b
2b
2c
2d
3a
H
H
H
H
NH2
NH
Br
Br
Br
N
1c
H
H
O
NH
Cl
NH
1d
COMe
Cl
Cl
NH
1e
H
H
CO2H
Cl
F
3b
H
H
OMe
Figure 2. (a) Structure of compounds 1a to 1e and 3b. (b) Structure of compounds 2a to 2d and 3a.
c
Magn. Reson. Chem. 2010, 48, 614–622
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/mrc

A. Salgado et al.
Table 1. 1H chemical shifts of 1a–e, 2a–d and 3a–b
1H chemical shifts (ppm), nJ(1H-1H) (Hz)
H-2
H-3
–
H-5
H-6
H-7
Other H
–
1a
8.69 (s)
9.43 (dd)
3J 6.7, 4J 4.2
8.98 (s)
7.38 (dd)
3J 4.3, 1.8
8.91 (dd)
3J 4.2, 4J 1.8
8.81 (s)
1ba
1c
8.46 (s)
8.50 (s)
8.43 (s)
–
–
–
–
–
–
–
8.83 (s)
8.98 (s)
3.19 (4H, m), 3.78 (4H, m)
1d
8.40 (d)
8.63 (d)
4.10–4.60 (1H, br s), 4.35 (2H, s), 7.42 (1H, d, 2J 8.1),
7.61 (1H, d, 2J 8.1), 7.71 (1H, s)
4J 2.6
4J 2.6
1eb
8.26 (s)
–
–
–
8.56 (s)
4.33 (2H, d, 2J 6.5), 5.88 (1H, t, 2J 6.6), 7.25 (1H, d, 2J
8.2), 7.49 (1H, d, 2J 8.3), 7.52 (1H, s), 7.62 (2H, d, 2J
8.2), 8.08 (2H, d, 2J 8.2)
2a
2b
2c
–
–
–
9.25 (s)
9.02 (dd)
3J 6.9, 4J 1.9
9.39 (d)
7.13 (dd)
3J 6.9, 4J 3.8
–
8.77 (dd)
3J 3.8, 4J 1.9
8.55 (d)
–
–
–
6.55 (2H, broad s)
4J 1.9
4J 1.9
9.44 (d)
–
–
8.58 (d)
4.47 (2H, d, 2J 6.3), 7.22 (1H, t, 2J 7.2), 7.30 (2H, dd, 2J
7.5, 7.2), 7.35 (2H, d, 2J 7.5), 7.71 (1H, t, 2J 6.3)
4J 2.2
4J 2.2
2d
–
–
9.51 (d)
8.95 (d)
2.67 (3H, s), 4.51 (2H, d, 2J 6.3), 7.23 (1H, t, 2J 7.2), 7.31
(1H, dd, 2J 7.6, 7.2), 7.38 (2H, d, 2J 7.6), 7.63 (1H, t, 2J
7.6), 7.68 (1H, t, 2J 6.3), 7.95 (1H, d, 2J 7.6), 8.05 (1H,
d, 2J 7.6), 8.33 (1H, s)
4J 2.4
4J 2.4
3a
–
–
–
–
9.01 (dd)
7.17 (dd)
8.78 (dd)
7.49 (2H, dd, 2J 8.3, 3JHF 8.3), 8.01 (2H, dd, 2J 8.3, 4JHF
5.5)
3J 6.9, 4J 1.6
8.91 (s)
3J 6.9, 4.0
7.16 (dd)
3J 5.5, 3.2
3J 6.9, 4J 4.0
8.76 (s)
3ba
3.86 (3H, s), 7.01 (2H, d, 2J 8.2), 8.19 (2H, d, 2J 8.2)
3J 5.5
3J 3.2
a In CDCl3/MeOD (4 : 1).
b Structure confirmed by X-ray crystal analysis.
Table 2. 13C chemical shifts of 1a–e, 2a–d and 3a–b
13C chemical shifts (ppm)
C-2
C-3
C-5
C-6
C-7
C-8a
Other C
1a
1ba
1c
156.3
156.3
155.5
154.2
–
–
–
–
138.1
135.3
121.3
114.1
111.3
105.7
138.8
135.6
156.0
155.4
150.6
149.2
155.1
153.3
151.0
148.8
–
–
49.5, 66.1
1d
45.1, 48.5, 127.8, 129.5, 130.5, 131.0,
139.6
1eb
154.5
–
129.7
131.5
145.6
150.7
45.7, 127.4, 128.7, 129.3, 129.8,
130.1, 130.5, 131.0, 137.9, 140.8,
145.5, 169.3
2a
2b
2c
2d
–
–
–
–
135.8
168.0
168.2
170.5
133.9
134.8
135.2
135.1
109.9
106.2
102.5
123.2
155.4
152.0
152.7
153.8
152.8
153.7
154.2
157.1
–
–
45.2, 126.7, 127.1, 128.3, 139.9
29.63, 48.15, 129.3, 129.7, 129.9,
130.9, 132.1, 133.8, 134.7, 136.7,
140.3, 142.8, 200.5
3a
–
144.8
133.6
135.7
110.8
110.2
155.5
154.8
154.2
155.7
116.3 (d, 2JC−F = 22 Hz), 122.7 (d,
4JC−F = 3 Hz), 130.6 (d,
3JC−F = 9 Hz), 163.1 (d,
1JC−F = 248 Hz)
3ba
165.6
–
54.2, 114.7, 122.1, 128.9, 161.8
a In CDCl3 –MeOD 4 : 1.
b Structure confirmed by X-ray crystal analysis.
c
www.interscience.wiley.com/journal/mrc
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2010, 48, 614–622

Regioisomer differentiation by 1H–15N HMBC experiments
C-3 (2b–d) shifts the NMR signal of H-5 to higher frequencies as
compared to 2a (0.49 ppm difference for 2d). On the other hand, in
compounds 2a–d, the H-7 signal is less sensitive to the presence
of the amino group at C-3.
Table 3. 15N chemical shifts of 1a–e, 2a–d and 3a–b
15N chemical shifts (ppm)
N-1
N-2
N-3
N-4
N-8
Other N
The 13C NMR signals that show greater regularity in their
chemical shifts are C-2 for 1a–e and C-3 for 2b–d, respectively,
resonating at δ = 154–156 and 167–170 ppm. Interestingly, the
measured 13C chemical shift of C-3 of the fully unsubstituted
compound 2a is smaller (δ = 135.8 ppm). Other carbon atoms
are found to be more sensitive to the presence and nature of
the substituents. In the [1,2,4]triazolo[1,5-a]pyrimidine series, C-5
and C-6 resonate between δ = 135–138 and 105–111 ppm for
derivatives 1a and 1b, respectively, but for those derivatives
having an amino group at C-6 (1c–e), C-5 and C-6 are shielded
anddeshielded,respectively.Inthe[1,2,4]triazolo[4,3-a]pyrimidine
series, the carbon atom showing greater sensitivity towards
substitution at C-6 is C-5, which resonates between δ = 133
and 135 ppm. As to C-6, it can be seen (Table 2) that in compounds
2a–c, all those having an amino group at C-3 or no substitution
at all (2a), C-6 resonates between δ = 101 and 110 ppm. In the
case of 2d, with an aryl substituent at C-6 and an amino function
at C-3, the measured chemical shift for C-6 is noticeably larger
(δ = 123 ppm). The carbon atoms C-7 and C-8a followed a more
definite pattern, appearing at δ = ca 155 ppm in both series.
Consequently, the comparatively large dispersion of 13C chemical
shift values made regioisomer differentiation difficult.
With these results in hand, we found that structural assignment
based only on 1H and 13C NMR data could be rather deceitful.
Thus, we turned our attention to 1H–15N HMBC experiments.
We expected that valuable structural information about the
regioisomericsystems1and2wouldbeobtainedfromlong-range
1H–15N correlations. Traditionally, 15N chemical shifts in nitrogen-
containing aromatic systems have been discussed in terms of
two different kinds of nitrogen atoms, the so-called pyrrole-like
and pyridine-like.[29,35] It is well known that, due to their higher
π-electron density, pyrrole-like nitrogen atoms are substantially
more shielded compared to their pyridine-like counterparts (ca
50–100 ppm).[36]
1H–15N HMBC experiments were run to measure the 15N
chemical shifts of compounds 1a–e and 2a–d indirectly (Table 3).
We observed that the 15N chemical shifts of all compounds
followed a far more distinctive and regular pattern than those
due to the 13C signals. As expected, four 15N signals for the
aromatic nitrogen atoms could be detected for each compound.
Surprisingly, we observed that, depending on the triazole
precursor employed in the syntheses (Scheme 1), the chemical
shift values of these four aromatic 15N NMR signals were distinctly
and uniformly distributed (Fig. 3a and 3b).
When 3-amino-[1,2,4]-triazole was used, the four aromatic 15N
resonances of 1b–e came in two narrow frequency regions, at
δ = 225–230 ppm (N-1 and N-4) and at δ = 275–278 ppm
(N-3 and N-8). The values of the encountered 15N chemical
shifts were very similar to those reported for 1a and for other
[1,2,4]triazolo[1,5-a]pyrimidine derivatives.[8,37] This 15N chem-
ical shift distribution pattern permitted us to conclude that
1b–e had the [1,2,4]triazolo[1,5-a]pyrimidine parent structure.
The X-ray crystal analysis of 1e (Fig. 4) fully corroborated the
[1,2,4]triazolo[1,5-a]pyrimidine structure. (CCDC 766900 contains
the supplementary crystallographic data for this paper. These
data can be obtained free of charge from The Cambridge Crystal-
lographic Data Centre via www.ccdc.cam.ac.uk/data request/cif.)
On the other hand, for compounds 2b–d, which were synthesised
1a
231
227
230
228
229
296
228
227
224
n.d.
n.d.
–
–
275
271
276
274
273
–
227
223
225
225
225
183
198
195
196
179
226
278
278
278
278
270
277
266
266
264
277
274
–
1ba
–
1c
–
58 (C6–N)
57 (C3–NH)
53 (C6–NH)
–
57 (C3–NH)c
66 (C3–NH)
66 (C3–NH)
–
1d
1eb
2a
–
–
331
231
229
227
n.d.
–
2b
2c
–
–
2d
3a
3ba
–
–
n.d.
–
a In CDCl3 –MeOD 4 : 1.
b Structure confirmed by X-ray crystal analysis.
c Determined with an HMQC experiment.
from 3,5-diamino-[1,2,4]-triazole precursors, the four aromatic 15
N
signals were distributed differently to 1b–e, with one signal at
ca δ = 195 ppm (N-4), two other signals at ca δ = 225–230 ppm
(N-1 and N-2), and the fourth signal (N-8) at ca δ = 270 ppm. It was
assumed that this other 15N chemical shift distribution pattern
corresponded to the [1,2,4]triazolo[4,3-a]pyrimidine series. That
was supported by a 1D NOESY experiment of 2d. Unequivocal
NOE interaction between H-5 (δ = 9.51 ppm) and the methylene
of the benzylamino group linked at C-3 (δ = 4.51 ppm) con-
firmed the [1,2,4]triazolo[4,3-a]pyrimidine structure of 2d (Figs 3
and 4).
In the [1,2,4]triazolo[1,5-a]pyrimidine series (1a–e), we ob-
servedthattheN-8chemicalshiftrangedfromδ = 270to278 ppm.
This value fits to a pyridine-like nitrogen atom in a six-membered
ring.[36] The N-1 atoms also gave a signal with a very uniform
chemical shift (δ = 227–231 ppm), consistent with a pyridine-like
nitrogen atom but in a five-membered ring. The bridge-head N-4
atoms, which contribute with a full electron pair to the π system,
are clearly pyrrole-like, although they are substantially deshielded
(δ = 223–227 ppm) when compared to what would be expected
for a common pyrrole nitrogen. This was assumed to be due to the
effect of the neighbouring N-3 atoms, which resonate between
δ = 271 and 276 ppm. As to compounds 2b–d, the 15N chemi-
cal shift values also show remarkable uniformity, although some
differences with the [1,2,4]triazolo[1,5-a]pyrimidines 1a–e are no-
ticed. The N-4 atoms of 2b–d are comparatively more shielded
(δ = 195–198 ppm vs δ = 223–227 ppm for 1a–e), in agreement
with a pyrrole-like nitrogen without adjacent nitrogen atoms. The
N-8 atoms give signals ranging from δ = 264 to 266 ppm, slightly
more shielded than in 1a-e. The chemical shifts of N-1 and N-2
are very similar in magnitude (δ = 223–231 ppm), both fitting
to imino nitrogens. It should be noticed that the presence of an
adjacent nitrogen atom (N-2) in 2b–d does not result in a sub-
stantial modification of the chemical shifts of the N-1 atoms, when
compared to 1a–e.
Unlike the 13C counterparts, it is noteworthy that these two
15N signal distribution patterns are essentially not dependent on
the substituents attached to the core scaffold. This emphasises the
relevance of 15N NMR as a tool for regioisomer differentiation. The
only exception to this trend was the unsubstituted compound
c
Magn. Reson. Chem. 2010, 48, 614–622
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/mrc

A. Salgado et al.
2a whose N-1 and N-2 15N chemical shifts (δ = 296 and
331 ppm, respectively) differed greatly from the rest of the series
(δ = 223 and 224 ppm for 2b). This can be attributed to the
increased π-electron density of the triazole ring due to the
exocyclic amino group. The N-8 nucleus (δ = 277 ppm) was also
deshielded when compared to the rest of the series (δ = 266 ppm
for 2b), whereas N-4 was more shielded (δ = 183 ppm vs
δ = 198 ppm for 2b). This could be due to the plausible
electron-donating effect of the substituents at C-3 of the
[1,2,4]triazolo[4,3-a]pyrimidine core scaffold. To explore further
the 15N chemical shifts of other [1,2,4]triazolo[4,3-a]pyrimidines
devoid of an exocyclic amino group, compounds 3a and 3b
were prepared from 2-hydrazinopyrimidine precursors (Scheme
2).[17,38] These reactions are expected to give the [1,2,4]triazolo[4,
3-a]pyrimidine system exclusively and, consequently, their 15N
chemical shifts should be similar to those of 2a.
The13Cand15NNMRdataofthesetwocompoundsareshownin
Tables 2 and 3. The 1H–15N HMBC experiment of 3a permitted to
spot two 15N resonances at δ = 179 and 277 ppm. The magnitude
of the chemical shift of these two signals agreed with pyrrole-type
and a pyridine-type nitrogen atoms, respectively. These values
are also very similar in magnitude to those of N-4 and N-8 of
2a and support the [1,2,4]triazolo[4,3-a]pyrimidine structure of
3a. Additional evidence comes from the 13C NMR data, as the
resonances due to the C-5, C-6, C-7 and C-8a atoms of compounds
2a and 3a are essentially identical. Some differences are seen
only with the chemical shifts of C-3 (δ = 135.8 ppm for 2a and
δ = 144.8 ppm for 3a) due to the substitution in 3a. On the
other hand, intense NOE interaction between the 1H resonances
at δ = 8.80 and 7.75 ppm, respectively, attributed to H-5 and
H-2,6 of the p-fluorophenyl substituent at C-3, corroborate the
[1,2,4]triazolo[4,3-a]pyrimidine structure of 3a.
Yet, the encountered 15N chemical shift values of 3b (δ =
226 and 274 ppm) are not consistent with a [1,2,4]triazolo[4,
3-a]pyrimidine derivative such as 2a or 3a. Instead, these
chemical shift values are much like those of the [1,2,4]triazolo[1,
5-a]pyrimidine derivatives 1a–e. The 13C chemical shifts of 3b
are also quite similar to those of 1a, with the exception of
the signal due to the C-2 atom (δ = 156.3 ppm for 1a and
δ = 165.6 ppm for 3b), probably caused by the substitution
at C-2 of 3b. Moreover, 1D NOESY experiments irradiating at
δ = 8.91 ppm (H-5) did not result in any NOE interaction with
the p-methoxyphenyl ring of 3b. All these results support the fact
that 3b belongs to the [1,2,4]triazolo[1,5-a]pyrimidine series. The
presence of the [1,2,4]triazolo[1,5-a]pyrimidine structure instead
of the expected [4,3-a] counterpart can be explained by a Dimroth
rearrangementthatgivesrisetothemorestable[1,2,4]triazolo[1,5-
a]pyrimidinescaffold.[11] Thesefindingsillustratehow15NNMRcan
Figure 3. (a) 1H–15N HMBC spectrum of compound 2d (DMSO-d6, 700 MHz) recorded with an evolution delay τ of 62.5 ms for the optimal detection of
long-range JH–N of 8 Hz. (b) 1H–15N HMBC spectrum of compound 1d (DMSO-d6, 700 MHz) recorded with an evolution delay τ of 166 ms for the optimal
detection of long-range JH–N of 3 Hz. A 1H–15N four-bond correlation between H-7 and N-1 can be seen.
c
www.interscience.wiley.com/journal/mrc
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2010, 48, 614–622

Regioisomer differentiation by 1H–15N HMBC experiments
Figure 3. (Continued).
CHO
Y
N
N
H
N
N
N
N
N
NHNH2
N
N
N
Y
Y
3a (Y = F)
3b (Y = OMe)
Scheme 2. Planned synthesis for compounds 3a and 3b.[17,38]
be used to verify the structure of some reported [1,2,4]triazolo[4,
3-a]pyrimidine compounds,[12,14–17] in particular those cases
prone to undergo Dimroth-type isomerisation. As we have seen
in the case of 3b, in which a rearrangement to a [1,2,4]triazolo[1,
5-a]pyrimidine had occurred, 15N NMR data can easily settle this
point (Fig. 5).
Finally, from the regioselectivity found in the condensation re-
actions of malonoaldehydes with either 3-amino-[1,2,4]-triazole or
3,5-diamino-[1,2,4]-triazole precursors (Scheme 1), we rationalised
that structures 1 result after an acid-mediated Dimroth-type re-
arrangement of a previously formed derivative of the parent
structure 2 (Scheme 3). For 2b–d, however, protonation occurs at
the exocyclic amino group, and thus no isomerisation to 1 takes
place. It should be mentioned at this point that other authors
Figure 4. X-ray crystal analysis of compound 1e. (CCDC 766900 contains
the supplementary crystallographic data for this paper. These data can be
obtained free of charge from The Cambridge Crystallographic Data Centre
via www.ccdc.cam.ac.uk/data request/cif).
c
Magn. Reson. Chem. 2010, 48, 614–622
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/mrc

A. Salgado et al.
f1 and 16 transients per increment (total experiment time 45 min).
Delays were adjusted for one-bond (1JHC) couplings of 145 Hz
N
N
N
N
N
N
N
1
and multiple-bond (nJHC) couplings of 10 Hz. The H–15N HMBC
OMe
N
experiments were carried out in duplicate, with delays adjusted
for one-bond (1JHN) couplings of 90 Hz and multiple-bond (nJHN
)
couplings of either 8 or 3 Hz. In both cases, the spectral sizes
were 8389.26 Hz (11.98 ppm) × 39 062.50 Hz (550.4 ppm) with 512
increments in f1 and 32 transients per increment (total experiment
time 6 h 28 min). All HMBC spectra were presented in magnitude
mode. Proton-decoupled 13C NMR spectra were recorded on a
Bruker AVANCE II 300 spectrometer fitted with a QNP 300 MHz
S1 probe and operating at 300.13 and 75.47 MHz (1H and 13C
frequencies, respectively). All NMR data were processed with the
MestReNova software (version 6.0.4–5850, Mestrelab Research
S. L., Santiago de Compostela, Spain). ESI+ high-resolution mass
spectra (HRMS) were obtained with a Bruker Daltonics maXis
UHR-TOF system.
F
3a
3b
Figure 5. Structure of compounds 3a and 3b.
N
N
N
+
N
N
N
N
N
H+
N
N+
H
N
HN
2a
N
N
N
N
- H+
N
+
N
N
N
H
1a
Preparation of [1,2,4]triazolo[1,5-a]pyrimidine (1a)
N
N
N
H+
N
3-Amino-[1,2,4]-triazole (1.0 g, 11.89 mmol), 1,1,3,3-tetra-
methoxypropane (2.16 ml, 13.08 mmol) and AcOH (10 ml) were
refluxed for 1 h.[8] Evaporation of excess acetic acid and column
chromatography (silica, 2–20% MeOH in CH2Cl2) gave 1a as a
yellow solid (84%). 1H NMR (700 MHz, DMSO-d6) (Table 1).
no ring opening
N
N
N
N
R
R
+NH2R
NHR
2b-d
Scheme 3. Rearrangement of [1,2,4]-triazolo[4,3-a]pyrimidines.
Preparation of 6-bromo-[1,2,4]triazolo[1,5-a]pyrimidine (1b)
isolated 3-amino-[1,2,4]triazolo[4,3-a]pyrimidines from the con-
densation of a 3,5-diamino-[1,2,4]-triazole and pentane-2,4-dione,
and that conversion to the corresponding [1,2,4]triazolo[1,5-
a]pyrimidine was achieved in another step.[39]
3-Amino-[1,2,4]-triazole(1.0 g,11.89 mmol)andbromomalonalde-
hyde (2.33 g, 15.46 mmol) were refluxed in EtOH (10 ml) for 4 h.[35]
Solvent evaporation and column chromatography (silica, 0–20%
MeOH in CH2Cl2) gave 1b as a pale yellow solid (84%). 1H
NMR (700 MHz, CDCl3-MeOD 4 : 1) (Table 1). 13C NMR (75 MHz,
CDCl3 –MeOD 4 : 1) (Table 2).
Conclusions
The structural determination of some [1,2,4]triazolo[4,3-
a]pyrimidine and [1,2,4]triazolo[1,5-a]pyrimidine derivatives was
unambiguously achieved from 15N NMR. The observed isomeri-
sation of [1,2,4]triazolo[4,3-a]pyrimidine structure to the regioi-
someric [1,2,4]triazolo[1,5-a]pyrimidine scaffold was rationalised
on the basis of a Dimroth rearrangement. The result of this iso-
merisation was clarified by measuring the 15N chemical shifts. The
obtained results validate 1H–15N HMBC as a useful tool for regioi-
somer differentiation. Further studies, including comparison of
experimental 15N chemical shifts to theoretical data, are currently
in progress.
Preparation of 6-(morpholin-4-yl)-[1,2,4]triazolo[1,
5-a]pyrimidine (1c)
1b (0.1 g, 0.5 mmol) was refluxed in morpholine (2.5 ml) for 3 h.
Removalofexcessmorpholineandcolumnchromatography(silica,
0–20% MeOH in CH2Cl2) gave 1c as a white solid (73%). 1H
NMR (700 MHz, DMSO-d6) (Table 1). 13C NMR (75 MHz, DMSO-
d6) (Table 2). HRMS (ESI+): C9H11N5O+ requires m/z 205.0964,
C9H12N5O+ (M + H+) requires m/z 206.1036; found 206.1017.
Preparation of 6-(3,4-dichlorophenylmethyl)amino-[1,2,4]
triazolo[1,5-a]pyrimidine (1d)
1b (0.05 g, 0.25 mmol) and 3,4-dichlorobenzylamine (0.15 ml,
1.13 mmol) were stirred at 100 ◦C for 10 min. After cooling,
conventional work-up (CH2Cl2) and column chromatography
(silica, 0–20% MeOH in CH2Cl2) gave 1d as a white solid (23%).
1H NMR (700 MHz, DMSO-d6) (Table 1). 1+3C NMR (75 MHz, DMSO-
d6) (Table 2). HRMS (ESI+): C12H9Cl2N5 requires m/z 293.0235,
C12H10Cl2N5+ (M + H+) requires m/z 294.0308; found 294.0330.
Experimental
All samples were dissolved in DMSO-d6 unless otherwise stated.
The concentration ranged from 60 to 200 mM. Spectra were
recorded at 25 ◦C on a Bruker AVANCE 700 spectrometer fitted
with a QXI 700 MHz S4 probe and operating at 700.13, 176.05 or
70.94 MHz (1H, 13C and 15N frequencies, respectively). Chemical
shifts are referred to external TMS (1H and 13C, 0.00 ppm) or
liquid NH3 (15N, 0.00 ppm). The 1H 90◦ hard pulse was calculated
for every sample. No loss of signal resolution due to tentative
aggregation effects was observed in the concentration range
used here. The standard Bruker pulse sequence hmbcgplpndqf
was used for the 1H–13C and 1H–15N HMBC experiments. The
spectral sizes for the 1H–13C HMBC experiments were 8389.26 Hz
(11.98 ppm) × 39 062.50 Hz (221.86 ppm) with 128 increments in
Preparation of 5-(4-carboxyphenyl)-6-(3,
4-dichlorophenylmethyl)amino-[1,2,4]triazolo[1,5-
a]pyrimidine (1e)
To 1d (100 mg, 0.34 mmol) in CHCl3 (2 ml) was added N-bromo
Succinimide (61 mg, 0.34 mmol). After 10 min stirring at room
temperature (RT) for 10 min, the mixture was extracted with
CHCl3:iPrOH (3 × 10 ml). Purification by column chromatography
c
www.interscience.wiley.com/journal/mrc
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2010, 48, 614–622

Regioisomer differentiation by 1H–15N HMBC experiments
(silica, 0–20% MeOH in CH2Cl2) gave a pale yellow solid (50 mg).
This solid was immediately suspended in water (1 ml), and 3-
acetylphenylboronic acid (0.029 g, 0.17 mmol), Pd(PPh3)4 (0.015 g,
0.013 mmol), sodium carbonate (0.043 g, 0.40 mmol) and TBAB
(0.043 g, 0.13 mmol) were added. Microwave irradiation (120 ◦C,
200 W) for 1 h and column chromatography (silica, 0–20% MeOH
filtration and dissolved in dichloromethane (8 ml). Iodobenzene
diacetate(380 mg, 1.17 mmol)wasaddedportion-wiseover5 min.
Stirring for 1 h at RT, solvent evaporation and trituration of the
solid residue with petroleum ether afforded 3a as a pale yel-
low solid (51%). 1H NMR (700 MHz, DMSO-d6) (Table 1). 13C NMR
+
(75 MHz, DMSO-d6) (Table 2). HRMS (ESI+): C11H7FN4 requires
1
in CH2Cl2) afforded 2e as a yellowish solid (4% overall yield). H
m/z 214.0655, C11H8FN4+ (M + H+) requires m/z 215.0727; found
NMR (700 MHz, DMSO-d6) (Table 1). 13C+NMR (75 MHz, DMSO-d6)
(Table 2). HRMS (ESI+): C19H13Cl2N5O2 requires m/z 413.0446,
C19H14Cl2N5O2+ (M + H+) requires m/z 414.0519; found 414.0531.
215.0745.
Preparation of 3-(4-methoxyphenyl)-[1,2,4]triazolo[1,
5-a]pyrimidine (3b)
Preparation of [1,2,4]triazolo[4,3-a]pyrimidine (2a)
2-Chloropyrimidine (0.5 g, 4.37 mmol, 1.0 equiv.) and 4-
methoxybenzhydrazide (0.725 g, 4.37 mmol) were suspended in
absolute EtOH (7 ml) and refluxed for 24 h. After cooling, a pale yel-
low solid (530 mg) was collected by filtration. This compound was
suspended in xylene (5 ml), phosphorus oxychloride (1 ml) was
added and the mixture was refluxed for 3 h. Careful basification
with 32% ammonia, extraction with Et2O and column chromatog-
raphy (silica, 2–15% MeOH in CH2Cl2) afforded 3b as a white solid
(19%, combined yield). 1H NMR (700 MHz, CDCl3 –MeOD 4 : 1)
(Table 1). 13C NMR (75 MHz, CDCl3 –MeOD 4 : 1) (Table 2). HRMS
(ESI+): C12H10N4O+ requires m/z 226.0855, C12H11N4O+ (M + H+)
requires m/z 227.0927; found 227.0935.
2-Hydrazinopyrimidine (0.5 g, 4.54 mmol, 1 equiv.) was suspended
intriethylorthoformate(4 ml)andthereactionmixturewasheated
at 110 ◦C for 45 min (open to the air to allow evaporation of
EtOH).[33] On cooling, the resulting yellow suspension was filtered
and rinsed with EtOH to give 2a (66%) as a yellow solid. 1H NMR
(700 MHz, DMSO-d6) (Table 1).
Preparation of 3-amino-6-bromo-[1,2,4]triazolo[4,
3-a]pyrimidine (2b)
3,5-Diamino-[1,2,4]-triazole (0.05 g, 0.51 mmol) and bromomalon-
aldehyde (0.1 g, 0.66 mmol) in EtOH (1 ml) were irradiated with
microwaves for 30 min (200 W, 130 ◦C). Column chromatography
(silica, 0–30% MeOH in CH2Cl2) gave 2b as a yellowish solid (33%).
1H NMR (700 MHz, DMSO-d6) (Table 1).+13C NMR (75 MHz, DMSO-
d6) (Table 2). HRMS (ESI+): C5H4BrN5 requires m/z 212.9650,
C5H5BrN5+ (M + H+) requires m/z 213.9723; found 213.9710.
Acknowledgements
We thank C. Dalvit, T. Parella and M. Gairí for their helpful
discussions, and K. Ashman and F. García for running the
HRMS experiments. AS thanks the Laboratori de Ressona`ncia de
Barcelona and the University of Barcelona for a training fellowship.
Part of this work was presented at the SMASH 2008 conference at
Santa Fe (NM, USA, September 2008).
Preparation of 3-benzylamino-6-bromo-[1,2,4]triazolo[4,
3-a]pyrimidine (2c)
5-Amino-3-benzylamino-[1,2,4]triazole (0.5 g, 2.64 mmol), bromo-
malonaldehyde (0.5 g, 3.17 mmol) and EtOH (5 ml) in acetonitrile
(20 ml) were refluxed for 15 h. Solvent evaporation and column
chromatography (silica, 0–30% MeOH in CH2Cl2) gave 2c as a
beige solid (51%). 1H NMR (700 MHz, DMSO-d6) (Table 1+). 13C NMR
(75 MHz, DMSO-d6) (Table 2). HRMS (ESI+): C12H10BrN5 requires
m/z 303.0120, C12H11BrN5+ (M+H+)requiresm/z 304.0192;found
304.0180.
References
[1] G. Fischer, Adv. Heterocycl. Chem. 2008, 95, 143.
[2] A. Maresta, M. Balducelli, L. Cantini, A. Casari, R. Chioin, M. Fabbri,
A. Fontanelli, P. A. Monici Preti, S. Repetto, S De Servi, E. Varani,
Circulation 1994, 20, 2710.
[3] M. M. A. El-Gendy, M. Shaaban, K. A. Shaaban, A. M. El-Bondkly,
H. Laatsch, J. Antibiot. 2008, 61, 149.
[4] M. R. Shaaban, T. S. Saleh, A. M. Farag, Heterocycles 2007, 71, 1765.
[5] A. M. Farag, Heterocycles 2008, 75, 3005.
[6] S. A. Yamashkin, N. Y. Kucherenko, M. A. Yurovskaya, Chem.
Preparation of 6-(3-acetylphenyl)-3-benzylamino-
[1,2,4]triazolo[4,3-a]pyrimidine (2d)
Heterocycl. Compd. 1997, 33, 499.
[7] E. M. Gizatullina, V. G. Kartsev, Chem. Heterocycl. Compd. 1993, 29,
1369.
[8] M. Ku¨nstlinger, E. Breitmaier. Synthesis 1983, 44.
[9] D. J. Brown, T. Nagamatsu, Aust. J. Chem. 1977, 30, 2515.
[10] E. S. H. El Ashry, Y. El Kilany, N. Rashed, H. Assafir, Adv. Heterocycl.
Chem. 1999, 75, 79.
[11] M. A. E. Shaban, A. E. A. Morgaan, Adv. Heterocycl. Chem. 1999, 73,
131.
2c (0.2 g, 0.66 mmol), 3-acetylphenylboronic acid (0.14 g,
0.86 mmol), Pd(PPh3)4 (0.076 g, 0.066 mmol), sodium carbonate
(0.21 g, 1.97 mmol) and tetrabutylammonium bromide (0.21 g,
0.66 mmol) in H2O (4 ml) were irradiated with microwaves (120 ◦C,
200 W) for 50 min. Conventional work-up (CH2Cl2) and column
chromatography (silica, 0–20% MeOH in CH2Cl2) gave 2d as a
white solid (11%). 1H NMR (700 MHz, DMSO-d6) (Table 1). 13C NMR
(75 MHz, DMSO-d6) (Table 2) HRMS (ESI+): C20H17N5O+ requires
m/z 343.1433, C20H18N5O+ (M + H+) requires m/z 344.1506; found
344.1495.
[12] T. Novinson, T. Okabe, R. K. Robins, P. Dea,J.Heterocycl.Chem. 1975,
12, 1187.
[13] T. La Noce, A. M. Giuliani, Tetrahedron 1978, 34, 2927.
[14] A. S. Shawali, A. H. Elghandour, A. R. Sayed, Synth. Commun. 2001,
31, 731.
ˇ
ˇ
[15] M. I. MusˇiS, B. VerSek, Synth. Commun. 2001, 31, 1511.
[16] M. R. Shaaban, T. M. A. Eldebss, A. F. Darweesh, A. M. Farag,
J. Heterocycl. Chem. 2008, 45, 1739.
Preparation of 3-(4-fluorophenyl)-[1,2,4]triazolo[4,
3-a]pyrimidine (3a)
[17] O. Prakash, V. Bhardwaj, R. Kumar, P. Tyagi, A. R. Aneja, Eur. J. Med.
Chem. 2004, 39, 1073.
2-Hydrazinopyrimidine (0.5 g, 4.54 mmol) and 4-fluoro-
benzaldehyde (0.51 ml, 4.77 mmol) were stirred in absolute EtOH
(15 ml) for 3 h at RT. A white solid (255 mg) was collected by
[18] C. Bulow, K. Hass, Berichte. 1909, 42, 4638.
[19] G. E. Martin, P. A. Irish, B. D. Hilton, K. A. Blinov, A. J. Williams, Magn.
Reson. Chem. 2007, 45, 624.
c
Magn. Reson. Chem. 2010, 48, 614–622
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
www.interscience.wiley.com/journal/mrc

A. Salgado et al.
[20] G. E. Martin,B. D. Hilton,K. A. Blinov,A. J. Williams,J.Nat.Prod.2007,
70, 1966.
[31] L. Stefaniak, M. Witanowski, U. Mahmoud, J. D. Roberts, G. A. Webb,
J. Crystallogr. Spectrosc. Res. 1989, 19, 159.
[21] H. S. El Khadem, J. Kawai, D. L. Swartz, Heterocycles 1989, 28, 239.
[22] G. E. Martin, C. E. Hadden, J. Nat. Prod. 2000, 63, 543.
[23] R. Marek, A. Lycka, E. Kolehmainen, E. Sieva¨ne, J. Tousek, Curr. Org.
Chem. 2007, 11, 1154.
[24] G. E. Martin, A. J. Williams, Annu. Rep. NMR Spectrosc. 2005, 55, 1.
[25] I. Delso, T. Tejero, Tetrahedron Lett. 2007, 48, 4101.
[26] M. Ko¨ck, J. Junker, T. Lindel, Org. Lett. 1999, 1, 2041.
[27] S. Fukuzawa, S. Matsunaga, N. Fusetami, Tetrahedron Lett. 1996, 37,
1447.
[32] E. Szlyk, A. Grodzicki, L. Pazderski, E. Bednarek, B. Kamienski,
Polyhedron 2000, 19, 965.
[33] W. W. Paudler, L. S. Helmick, J. Heterocycl. Chem. 1966, 3, 269.
[34] G. L. Rusinov, P. V. Plekhanov, A. U. Ponomareva, O. N. Chupakhin,
Mendeleev Commun. 1999, 233.
[35] M. Witanowski, L. Stefaniak, G. A. Webb, Ann. Rep. NMR Spectrosc.
1993, 25, 1.
[36] G. C. Levi, R. L. Lichter, in Nitrogen-15 Nuclear Magnetic Resonance
Spectroscopy, Wiley: New York, 1979, pp 114.
[28] P. A. McDonnell, A. D. Gauthier, M. P. Ferro, Magn. Reson. Chem.
1998, 36, 35.
[37] E. Szlyk, L. Pazderski, I. Lakomska, L. Kozerski, J. Sitkowski, Magn.
Reson. Chem. 2002, 40, 529.
[29] N. Kawamura,
H. Naganawa, T. Takeuchi, J. Antibiot. 1996, 49, 651.
[30] P. M. Chaudhary, S. R. Chavan, M. Kavitha,
R. Sawa,
Y. Takahashi,
K. Isshiki,
T. Sawa,
[38] M. Ciesielski, D. Pufky, M. Do¨ring, Tetrahedron 2005, 61, 5942.
[39] B. C. Bishop, H. Marley, P. N. Preston, S. H. B. Wright. J. Chem. Soc.,
Perkin Trans. 1, 1999, 1527.
S. P. Maybhate,
S. R. Deshpande, A. P. Likhite, P. R. Rajamohanan, Magn. Reson.
Chem. 2008, 46, 1168.
c
www.interscience.wiley.com/journal/mrc
Copyright ꢀ 2010 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2010, 48, 614–622