Furoxan rearrangement of some pyridofuroxan derivatives studied by 1H, 13C, 14N, 15N and 17O NMR spectroscopy

Article abstract of DOI:10.1002/1099-1395(200008)13:8<480::AID-POC268>3.0.CO;2-Z

Pyridofuroxan ([1,2,5]oxodiazolo[3,4-b]pyridine 1-oxide) undergoes isomerization between the N1-oxide and N3-oxide forms which can be observed by the ¹H, ¹³C and ¹⁵N NMR spectroscopy but not by ¹⁴N and ¹⁷O NMR at ambient and low temperatures. The rearrangement becomes slower at low temperatures and at 233 K ¹H NMR signals for the two structures become observable. ¹H, ¹³C and ¹⁵N chemical shifts and ¹H-¹H, ¹³C-¹H and ¹³C-¹³C coupling constants are used to characterize both forms in the equilibrium mixture. From the ¹H NMR integrals at 233 K equilibrium constants are calculated. Protonation studies using trifluoroacetic acid as a solvent showed the favoured site of protonation to be the pyridine N4 nitrogen atom. DFT shielding calculations are reported for the ¹³C, ¹⁵N and ¹⁷O nuclei which support the assignments given. From the point of view of structural changes, ¹J_CC data for 8-nitrotetrazolo[1,5-a]pyridine and o-nitroaminopyridine as precursors of the pyridofuroxans are given for comparison purposes. X-ray diffraction data on 5-methoxypyridofuroxan support the structural results obtained from the NMR investigations. Copyright

Full text of DOI:10.1002/1099-1395(200008)13:8<480::AID-POC268>3.0.CO;2-Z

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2000; 13: 480–488
Furoxan rearrangement of some pyridofuroxan derivatives
13 14
15
studied by 1H, C, N, N and 17O NMR spectroscopy
1
1
1
P. Cmoch, * B. Kamienski, K. Kamienska-Trela, L. Stefaniak1† and G. A. Webb2
Â
Â
1Institute of Organic Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, P.O.B. 58, 01-224 Warsaw 42, Poland
2Department of Chemistry, University of Surrey, Guildford, Surrey GH2 5XH, UK
Received 1 February 2000; revised 17 March 2000; accepted 22 March 2000
ABSTRACT: Pyridofuroxan ([1,2,5]oxodiazolo[3,4-b]pyridine 1-oxide) undergoes isomerization between the N1-
oxide and N3-oxide forms which can be observed by the 1H, 13C and 15N NMR spectroscopy but not by 14N and 17
O
1
NMR at ambient and low temperatures. The rearrangement becomes slower at low temperatures and at 233 K H
1
1
NMR signals for the two structures become observable. H, 13C and 15N chemical shifts and H–1H, 13C–1H and
13C–13C coupling constants are used to characterize both forms in the equilibrium mixture. From the H NMR
1
integrals at 233 K equilibrium constants are calculated. Protonation studies using trifluoroacetic acid as a solvent
showed the favoured site of protonation to be the pyridine N4 nitrogen atom. DFT shielding calculations are reported
for the 13C, 15N and 17O nuclei which support the assignments given. From the point of view of structural changes,
1JCC data for 8-nitrotetrazolo[1,5-a]pyridine and o-nitroaminopyridine as precursors of the pyridofuroxans are given
for comparison purposes. X-ray diffraction data on 5-methoxypyridofuroxan support the structural results obtained
from the NMR investigations. Copyright 2000 John Wiley & Sons, Ltd.
1
KEYWORDS: furoxan rearrangement; H, 13C, 14N, 15N and 17O NMR; protonation; DFT calculations
INTRODUCTION
measurement nor the unequivocal assignment of the
signals present in these spectra is an easy task.
In order to obtain a more detailed characterization of
pyridofuroxan ([1,2,5]oxodiazolo[3,4-b]pyridine 1-
oxide) and some of its derivatives by NMR studies, we
have become interested in the elucidation of isomeriza-
tion of these compounds.1,2 The rearrangement involves
oxygen migration between the N1 and N3 atoms and
valence isomerization which occurs most probably via o-
dinitrosopyridine as the intermediate step (Fig. 1). Such a
mechanism has been proposed by Dunkin et al.3 and
Murata and Tomioka4 for an analogous rearrangement of
benzofuroxan. However, no systematic studies devoted to
substituent effects on the pyridofuroxan equilibrium have
been performed so far and the number of NMR data
published for this group of compounds is limited. The 1H
NMR characteristics has been published by Boulton et
al.5 for pyridofuroxan itself and for its four methyl
EXPERIMENTAL
Syntheses. Pyridofuroxan and its derivatives were
obtained as the result of oxidation of the corresponding
o-nitroaminopyridines (compounds 1–4, 6 and 8) or in
the thermolysis of corresponding nitrotetrazolo[1,5-
a]pyridines (compounds 1 and 5) (Fig. 2).2 Compounds
5, 9 and 10 were prepared according to procedure
described previously (Fig. 2).6 Selectively labelled (in
positions 3 and 4) compound 1 was prepared by the
thermolysis reaction of selectively labelled compound 11
(see Fig. 4). Data for unknown substituted pyridofurox-
ans are as follows.
1
derivatives. H and 13C spectra have been measured for
6-Chloropyridofuroxan (6). MS: m/z 173, 171, 157, 155,
143, 141, 129, 127, 113, 111, 88, 86, 76. Analysis:
calculated for C5H2N3O2Cl, H 1.17, C 34.99, N 24.90;
found, H 1.08, C 34.80, N 24.78%. IR (KBr): 1618, 1572,
six variously substituted pyridofuroxans by Lowe-Ma et
al.6 However, no 15N NMR spectra have been published
so far, which is not surprising, since neither the
1518, 1503, 1492, 1418, 1338, 1238, 1169, 1050, 1030,
*Correspondence to: P. Cmoch, Institute of Organic Chemistry, Polish
Academy of Sciences, Kasprzaka 44/52, P.O.B. 58, 01-224 Warsaw
42, Poland.
1
928, 873, 755, 698, 586 cm
.
E-mail: piocmo@ichf.edu.pl
5-Methoxypyridofuroxan (7). 2-Chloro-3-nitro-6-meth-
oxypyridine (1.0 g) and sodium azide (0.38 g) were
dissolved in 10% aqueous ethanol (50 ml) at ambient
† Deceased October 1998.
Contract/grant sponsor: Polish State Committee for Scientific
Research; Contract/grant number: 3 TO9A 010 12.
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 480–488

FUROXAN REARRANGEMENT OF PYRIDOFUROXAN DERIVATIVES
481
Figure 1. Valence tautomerism scheme for condensed furoxans
Figure 2. Numbering schemes and structures of the compounds studied. In the case of compounds 12 and 13 the numbering do
not correspond to the IUPAC system
temperature, and 10% hydrochloric acid (5 ml) was
added. The solution was then heated under reflux for
48 h. Evaporation to dryness, addition of water (20 ml)
and filtration gave a yellow solid residue. Recristalliza-
tion from ethanol–chloroform (50:50) gave 7 as pale
yellow needles (0.62 g), 70%. MS: m/z 167, 151, 137,
121, 96, 94, 80, 64. M.p.: 139–140°C. Analysis:
calculated for C6H5N3O3, H 2.99, C 43.11, N 25.15;
found, H 2.98, C 43.14, N 24.96%. IR (KBr): 1617, 1599,
1536, 1516, 1491, 1466, 1408, 1338, 1278, 1231, 1188,
acetone solutions of compounds studied were between
0.5 and 1 mol dm 3. The temperature at which the
spectra were measured was 233 K with exception of
compounds 7, 9 and 10, for which the temperature was
303 K.
Standard experimental conditions and standard Bruker
1
programs for XHCORRD (optimized for JCH = 160–
170 Hz, INVGATE and INADEQUATE (optimized for
1
1
65 Hz) for JCC were used. Moreover, GHMBC H–15N
correlation experiments were used to confirm the proper
1
1119, 1029, 992, 958, 832, 806, 759, 723, 690, 673,
assignment of the 15N signals. The H and 13C spectra
1
567 cm
.
data are given relative to the TMS signal at 0.0 ppm at
233 K. Nitromethane, whose signal is at 0.0 ppm, was
used as an external standard for the 15N NMR spectra. No
correction of the calibrations for the 15N NMR spectra at
low temperatures was made. For 17O NMR measure-
ments at 303 K external H2O was used as the standard. IR
spectra in KBr pellets were measured using a Perkin
Elmer 2000 FT-IR spectrometer and mass spectra were
measured using an ADM-604 instrument.
7-Chloropyridofuroxan (8). MS: m/z 173, 171, 157, 155,
143, 141, 127, 125, 113, 111, 90, 88, 86, 76. Analysis:
calculated for C5H2N3O2Cl, H 1.17, C 34.99, N 24.90;
found, H 1.10, C 34.82, N 24.84%. IR (KBr): 1608, 1577,
1529, 1511, 1496, 1416, 1331, 1196, 1044, 1012, 949,
1
853, 615, 552 cm
.
Spectra. Bruker AM 500 and DRX 500 spectrometers
operating at 500.13, 125.76, 36.45, 50.68 and 67.80 MHz
for 1H, 13C, 14N, 15N and 17O, respectively were used for
the measurement of all the spectra. The concentrations of
Calculations. DFT nuclear shielding calculations were
made using the deMon7 program with experimental and
optimized molecular geometries.
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 480–488

482
P. C. CMOCH ET AL.
Table 1. 1H NMR chemical shifts (ppm) and 1H±1H coupling constants (Hz) in acetone-d6 (T = 233 K) of the compounds studied
1 I
1 II
2 I
2 II
3 I
3 II
4 I
4 II
5 I
H5
8.99
7.53
8.16
—
3.80
9.05
1.60
8.80
7.69
8.39
—
3.75
9.25
1.35
—
7.40
7.98
2.69
—
9.16
—
—
8.82
—
7.80
2.54
—
8.62
—
8.03
2.56
—
8.70
7.23
—
2.62
4.03
—
8.59
7.43
—
2.65
3.96
—
9.59
—
8.89
—
—
—
2.38
H6
7.56
8.22
H7
CH3
H5–H6
H6–H7
H5–H7
a
—
—
9.30
—
—
—
—
—
b
b
—
—
5 II
6 I
6 II
7 Ic
8 I
8 II
9 Ic
10 I Tc,d
10 I Ac,d
H5
H6
H7
CH3
H5–H6
H6–H7
H5–H7
9.36
—
9.09
—
—
—
2.05
8.91
—
8.38
—
—
—
2.28
8.73
—
8.59
—
—
—
2.06
—
6.94
7.89
4.08
—
9.39
—
8.84
7.60
—
8.71
7.86
—
—
7.23
7.00
—
—
9.35
—
—
7.73
7.63
—
—
9.80
—
—
6.70
7.71
—
—
9.30
—
—
—
4.34
—
—
4.29
—
—
a
Overlapped with signal of methyl group of 2 I.
b Not determined because of coupling constants with methyl group,
c
Measured at 303 K.
d T = tetrazole form and A = azide form of 10 I.
X-ray structure of 5-methoxypyridofuroxan (7). The
crystal system of 5-methoxypyridofuroxan (7) was
orthorhombic and the space group was Pna21, Z = 2;
In order to determine whether I or II is the major form
present in acetone solution, we followed the approach of
Boulton et al.5 and analysed the chemical shifts for the
C3a and C7a carbon atoms in compound 1. Fortunately,
we observed in the low-temperature 13C spectrum of this
compound two pairs of signals of different intensity. By
considering the chemical shifts of these nuclei in
benzofuroxan, which are 153.0 and 114.7 ppm for C3a
and C7a, respectively,8 and using additivity rules for the
introduction of the nitrogen atom in the benzene ring one
can predict the chemical shift values in I and II. The
additivity increments are deduced from a comparison of
the relevant carbon chemical shifts of naphthalene and
quinoline. This shows that the introduction of the
nitrogen atom in quinoline results in a chemical shift
increase of 16.1 ppm for the a-carbon (C3a) and a
decrease of 4.4 ppm for the b-carbon (C7a). The
predicted values for I are 170 and 110 ppm for C3a and
C7a, respectively. These values are in close agreement
with the observed chemical shifts at 160 and 110 ppm.
Following the same reasoning one obtains 136 and
144 ppm for C3a and C7a, respectively in form II. The
observed signals at 126 and 148 ppm are in reasonable
agreement with this prediction. Since the pair of signals
at 160 and 110 ppm are significantly stronger than those
at 126 and 148 ppm we conclude that form I is present to
a larger extent than is form II in acetone solution.
unit cell parameters a = 7.4470(1), b = 14.935(3) and
3
˚
˚
˚
c = 6.3630(1) A, V = 107.7 2 A , ꢀ(Cu Ka) = 1.54178 A,
Dc = 1.569 mg m 3. Data were collected at 293 K on an
Enraf-Nonius MACH3 diffractometer using the ! 2ꢁ
scan technique for the ꢁ range 5.93–73.24°. Among 555
measured reflections, 555 were independent (Rint
=
0.000). The structure was solved by direct methods (G.
M. Sheldrick, SHELXS-93, Program for Crystal Struc-
ture Determination, University of Go¨ttingen, 1993) and
refined against F2(hkl) with full-matrix least-squares (G.
M. Sheldrick, SHELXL-93, Program for Crystal Struc-
ture Refinement, University of Go¨ttingen, 1993). Hydro-
gen atoms were included in the refinement at their
calculated positions. The final residuals were
R1 = 0.0755, wR2 = 0.1496 (R indicates for all data:
R1 = 0.0755, wR2 = 0.1496). Bond lengths, angles and
torsions are given in Table 5. Other parameters are
available on request.
RESULTS AND DISCUSSION
The compounds studied are presented in Fig. 2. The
spectra were measured under different conditions, i.e. at
ambient and low temperatures and in protic and aprotic
solvents. The 1H, 13C, 15N and 17O NMR data are given in
Tables 1, 2, 3 and 4, respectively. An assignment of the
signals in the 13C NMR spectra was made unequivocally
on the basis of corresponding INADEQATE spectra
which were measured for compounds 1, 4, 6 and 7 (Table
2).
As shown in Fig. 2, the present investigations involved
10 compounds, the majority of which are found to exhibit
equilibrium. Those which do not are compounds 7, 9 and
10. For those compounds which do exhibit equilibrium in
solution the predominant form is I with the exception of
compound 8, where form II predominates. X-ray
diffraction data (Table 5) on crystals of compound 7
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 480–488

FUROXAN REARRANGEMENT OF PYRIDOFUROXAN DERIVATIVES
483
Table 2. 13C NMR chemical shifts (ppm) and 13C±1H and 13C±13C coupling constants (Hz) of the compounds studied in acetone-
d6 solution at T = 233 K
1 I
1 II
2 I
2 II
3 I
3 II
4 I
4 II
5 I
a
C7a
C3a
C5
C6
C7
CH3
C5–H5
C6–H6
C7–H7
C7a–C3a
C5–C6
C6–C7
C7–C7a
109.6
160.4
161.7
125.8
123.3
—
185.7
172.1
179.2
65.5
148.4
126.2
157.3
129.2
127.7
—
108.3
160.2
171.5
127.6
122.4
26.1
—
—
—
—
108.9
160.0
164.5
136.6
119.0
19.0
148.4
124.6
160.0
140.0
123.8
19.2
109.7
160.2
161.0
124.4
137.1
15.5
149.3
125.3
156.9
127.3
139.8
15.4
107.0
157.7
153.7
141.4
122.0
—
a
a
a
a
—
25.7
—
a
187.9
170.0
176.7
—
183.7
—
185.0
—
183.8
168.8
—
186.3
167.3
—
—
a
a
170.5
174.3
—
—
a
a
—
176.4
173.5
—
a
b
b
b
b
a
a
b
—
—
—
—
—
—
—
—
—
—
—
—
—
—
b
b
b
b
b
48.8
61.4
64.1
49.6
61.1
62.8
—
—
49.6
62.9
64.0
50.6
b
b
b
b
a
b
—
—
—
b
b
b
b
a
b
—
—
—
—
—
—
5 II
6 I
6 II
7 Ic
8 I
8 II
9 Ia
10 I Ac,d
10 I Tc,d
a
C7a
C3a
C5
C6
—
—
—
—
—
—
—
—
109.2
158.4
160.9
131.6
121.0
—
193.6
—
183.0
148.0
124.4
156.3
135.3
125.3
—
196.1
—
180.8
107.9
159.6
168.8
121.1
125.2
55.5
—
175.3
174.9
108.3
160.8
160.8
125.0
131.2
—
188.7
175.4
—
147.5
127.3
156.2
128.3
134.8
—
191.3
173.6
—
106.9
160.4
167.7
129.6
119.7
—
107.5
157.8
162.9
120.6
124.7
—
—
174.8
178.5
107.0
144.1
149.9
116.1
119.3
—
—
181.5
181.3
a
a
a
a
C7
a
CH3
C5–H5
C6–H6
C7–H7
a
a
—
a
a
—
a
a
—
—
a
Not determined because of low concentration.
b Not recorded.
c
Measured at 303 K.
d T = tetrazole form of and A = azide form of 10 I.
(Fig. 3) show that in the solid state this compound exists
entirely in form I, thus supporting the finding of the 13C
NMR measurements, especially for the assignment of the
C3a and C7a signals which are typical of the N1-oxide
form I.
The pyridofuroxan molecule is almost planar (Table
5). The bond lengths and bond angles of the furoxan
portion are similar to the values mentioned pre-
viously,6,9,10 although some differences caused by the
introduction of a nitrogen atom and methoxy group in the
benzofuroxan molecule are also discernible. The pyridine
ring has typical bonds reflecting the presence of double
bonds and thus the presence of furoxan functionality. The
˚
C5–N4 bond is shorter than the reported mean of 1.337 A
for pyridine C—N bonds,6 whereas the C5—C6 and
C7—C7a bonds are both long and C6—C7 is short
compared with the aromatic C—C bond length of
˚
1.380 A found in other pyridines.
Table 3. 15N NMR chemical shifts (ppm) of the compounds studied in acetone-d6 solution at T = 233 K
1 I
1 II
4.4
20.0
98.0
12.5
2 I
2 II
3 I
3 II
4 I
4 II
6.0
19.5
106.1
13.5
5 I
16.5
—
87.9
11.4
a
a
a
N1
N3
N4
N4–H5
19.4
6.4
90.8
12.2
—
—
—
—
—
—
—
20.7
—
—
18.6
7.5
98.6
13.0
a
a
a
a
a
a
—
a
91.3
12.0
99.5
12.6
—
5 II
6 I
6 II
7 Ib
8 I
8 II
9 Ib
10 IAb,c
10 ITb,c
a
a
a
N1
N3
N4
—
—
20.2
5.0
86.1
11.8
—
24.9
17.8
7.3
94.5
13.1
6.5
19.0
101.3
13.3
31.2
27.6
160.5
—
—
22.3
29.8
158.8
—
a
a
—
17.3
163.4
—
13.9
140.4
—
92.1
11.8
93.8
12.0
N4–H5
a
1
Not recorded or not observed in H–15N GHMBC experiment.
b Measured at 303 K. 9 I: —N
=
P = 255.6 ppm.
c
T = tetrazole form and A = azide form of 10 I, 10 I A: N1' = 261.7 ppm, N2' = 146.7 ppm, N3' = 136.6 ppm. 10 I T: N1' = 56.9 ppm,
N2' = ‡23.2 ppm, N3' = 28.6 ppm.
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 480–488

484
P. C. CMOCH ET AL.
Table 4. 17O NMR chemical shifts (ppm) in acetone-d6 solution at T = 303 K for some of the compounds studied
1
4
5
6
7
10 I Ta
Oexo
406 (230)
509 (270)
—
409 (260)
506 (290)
—
433 (310)
526 (390)
411 (290)
518 (340)
—
388 (290)
492 (350)
100 (250)
406 (300)
502 (500)
—
O2
b
NO2/OCH3
—
a
T = tetrazole form of 10 I.
b Invisible because overlapping with acetone signal.
The 15N NMR spectra of the compounds which
undergo the equilibrium consist of two sets of signals
of different intensity, three resonances each. Their
assignment was made in the following way using
compound 1 as an example. The 15N NMR spectrum of
this compound has three signals at 6.4, 19.4 and
90.8 ppm, which, owing to their intensity, obviously
belong to the form I. Their assignment was made as
follows. The signal at 90.8 ppm is fairly typical of the
pyridine-type nitrogen atom11 and is assigned to N4 of
structure I. In order to distinguish between the other two
signals we also recorded the 14N NMR spectrum. The
signal appearing at 19.4 ppm is much sharper in the 14N
NMR spectrum than that at 6.4 ppm. The N-oxide
nitrogen atoms invariably give a fairly sharp 14N NMR
signal due to the presence of the residual charge on the
nitrogen atom producing a fairly small electric field
gradient at the site of quadrupolar nuclei.12 Conse-
quently, the signal at 19.4 ppm is attributed to N1 in
structure I and the signal at 6.4 ppm to N3. In the case
of structure II the signal at 98.0 ppm must be assigned
to N4, whereas the signals at 4.4 ppm and 20.0 ppm
are assigned to N1 and N3, respectively.
Figure 3. Crystallographic structure (ORTP projection) of 5-
methoxypyridofuroxan (7)
To assist in the assignment of the 15N signals, a sample
Table 5. Bond lengths and angles of compound 7
˚
Bond
Bond length (A)
Bond
Bond angle (°)
Bond
Torsion (°)
N1–Oexo
N1–O2
O2–N3
N3–C3a
C3a–N4
N4–C5
C5–C6
1.215(4)
1.446(5)
1.389(6)
1.318(5)
1.364(5)
1.294(6)
1.461(6)
1.328(6)
1.422(7)
1.418(6)
1.330(6)
1.00
C5–N4–C3a
C5–O8–C9
113.5(4)
117.6(4)
109.7(1)
105.5(1)
112.4(6)
103.4(1)
144.8(1)
105.4(4)
135.1(5)
118.2(4)
106.4(4)
120.8(4)
126.3(4)
112.7(4)
124.0(4)
112.3(4)
123.7(4)
108.8(3)
114.5(4)
122.8(3)
122.7(3)
106.9(5)
131.3(5)
120.4(4)
120.3(4)
119.8(2)
119.8(2)
C3a–N4–C5–O8
C3a–N4–C5–C6
C9–O8–C5–N4
C9–O8–C5–C6
O2–N3–C3a–N4
O2–N3–C3a–C7a
C5–N4–C3a–N3
C5–N4–C3a–C7a
C3a–N3–O2–N1
Oexo–N1–O2–N3
C7a–N1–O2–N3
176.7(1)
2(2)
O8–C9–H9a
O8–C9–H9b
H9a–C9–C9b
O8–C9–H9c
H9a–C9–H9b
C3a–N3–O2
0(2)
176.0(1)
179.5(1)
1(2)
177.3(1)
3(2)
3(2)
180.0(1)
5(2)
179(2)
5(2)
15(3)
170.8(1)
4(2)
177(2)
171.4(1)
9(2)
C6–C7
C7–C7a
C7a–C3a
N1–C7a
C6–H6
C7–H7
C5–O8
O8–C9
C9–H9a
C9–H9b
C9–H9c
O
exo–N1–C7a
O
exo–N1–O2
C7a–N1–O2
N4–C5–O8
N4–C5–C6
O8–C5–C6
N3–C3a–N4
N3–C3a–C7a
N4–C3a–C7a
N3–O2–N1
C6–C7–C7a
C6–C7–H7
C7a–C7–H7
N1–C7a–C3a
N1–C7a–C7
C3a–C7a–C7
C7–C6–C5
C7–C6–H6
C5–C6–H6
O
exo–N1–C7a–C3a
O2–N1–C7a–C3a
O1–N1–C7a–C7
O2–N1–C7a–C7
N3–C3a–C7a–N1
N4–C3a–C7a–N1
N3–C3a–C7a–C7
N4–C3a–C7a–C7
C6–C7–C7a–N1
C6–C7–C7a–C3a
C7a–C7–C6–C5
N4–C5–C6–C7
1.00
1.325(5)
1.453(6)
1.07(3)
1.13(3)
1.05(2)
177(2)
13(2)
11(3)
7(3)
178(2)
O8–C5–C6–C7
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 480–488

FUROXAN REARRANGEMENT OF PYRIDOFUROXAN DERIVATIVES
485
Figure 4. Two possible mechanisms for the reaction between 2-chloro-3-nitropyridine and selectively labelled potassium azide,
K15NN15N
individually enriched at N3 and N4 was prepared as
shown in Fig. 4 and its spectrum analysed. The synthesis
of selectively 3- and 4-15N-labelled pyridofuroxan, 1,
proceeds according to the following steps.
However, as was shown in our recent study on this class
of compounds, this assignment should be reversed.15 It is
therefore reasonable to assume that also in the case of
pyridofuroxans the exocyclic oxygen is shielded with
respect to the ring oxygen atom (see also the section
where the relevant 17O NMR DFT calculations are
discussed).
In the first step the reaction between 2-chloro-3-
nitropyridine and the doubly labelled potassium azide
K15NN15N takes place, leading to compound 11, whose
labelling depends on the mechanism involved: the 1,3-
di-15N-8-nitrotetrazolo[1,5-a]pyridine isotopomer, 11a
(45%), is formed as the result of an SN nucleophilic
substitution mechanism and 2,4-di-15N-8-nitrotetrazo-
lo[1,5-a]pyridine isotopomer, 11b (55%), is obtained
when the ANRORC (addition of nucleophile, ring
opening and ring closure) mechanism takes place (Fig.
4). The ratio of these two isotopomers (45:55) can be
easily estimated from the integration of the signals in the
15N NMR spectrum (in DMSO) of the compound 11.
Two signals at 67.3 and 30.0 ppm are characteristic of
the tetrazole nitrogens N1 and N3, respectively, which
correspond to isotopomer 11a, and two signals at ‡20.7
and 122.9 ppm are due to the N2 and N4 atoms of the
isotopomer 11b.13 After thermolysis of labelled com-
pound 11, two isotopomers of compound 1 are formed:
3-15N (45%) and 4-15N (55%).
In order to obtain some independent proof of the
assignments of the signals in the spectra of the
compounds studied, we performed DFT calculations of
the nuclear shieldings for the parent compound 1 (in both
forms I and II) and compound 7. The results are
presented in Figs 5, 6 and 7 where, respectively, plots
of observed 13C, 15N and 17O NMR chemical shifts
against nuclear shieldings calculated by the density
functional method (DFT)7 are presented. In the DFT
calculations we employed the PW91 functional and the
IGLO III basis set. The correlation coefficients and
standard deviations for the least-squares data fitting are
given in Figs 5–7 and are very satisfactory in each case.
These results support our assignments of the 13C, 15N and
17O NMR spectra. It is also worth mentioning that not
only are the trends reproduced correctly but also
agreement between the calculated and experimental data
is satisfactory.
Inspection of the data in Table 6 reveals that
compounds 1–6 and 8 exist in acetone solution in the
form of two tautomers, and for compounds 1–6 the
equilibrium between these tautomers is strongly shifted
towards the N1-oxide form, I. The compounds in
question are the parent pyridofuroxan 1, its 5-, 6- and
7-methyl derivatives, 2–4, the 6-nitro derivative, 5, and
the 6-chloro derivative, 6. In the case of the 7-chloro
derivative (8), the equilibrium is displaced in the opposite
direction with form II being favoured. The KI/II values
vary from 0.05 in compound 5, where the strong electron-
withdrawing NO2 group is attached to carbon 6, to 0.15
On lowering the temperature of the mixture of
isotopomers of 1 in acetone solution, two sets of signals
are observed. The stronger signals at 6.4 (N3) and
90.8 ppm (N4) are characteristic of the N1-oxide form
I, whereas two weaker signals at 20.0 and 98.0 ppm
correspond to atoms N3 and N4, respectively, of form II.
These observations support the nitrogen signal assign-
ments based on the 14N NMR relative linewidths and
1H–15N correlation measurements.
An assignment of the signals in the 17O NMR spectrum
presents some problems. From an 17O NMR study on
benzofuroxan it is reported that the exocyclic oxygen is
deshielded with respect to the ring oxygen atom.14
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 480–488

486
P. C. CMOCH ET AL.
Figure 7. Plot of calculated 17O absolute shieldings against
Figure 5. Plot of calculated 13C absolute shieldings against
experimental 17O NMR chemical shifts
experimental 13C NMR chemical shifts
for compound 6, where the electron-donating chlorine
atom is present in position 6. These results show that the
substituents at position 6 have a relatively small influence
regardless of their nature on the position of the
equilibrium. This is in agreement with the previously
published data.16
In the case of compounds with a substituent in position
5, both forms I and II were observed only in the case of
the methyl derivative 2. The presence of substituents
containing electronegative atoms leads most obviously to
destabilization of the N3-oxide form and, as a result, only
form I is found in the case of compounds 7, 9 and 10. It
should be mentioned that in the case of compound 10,
owing to the presence of the azide group, a more complex
equilibrium may exist.13 As shown in Table 6, the KI/II
data for compounds 4 and 8 are significantly different
(0.18 and 1.22, respectively). This difference arises from
a change in the substituent at position 7 from a methyl
group to a chlorine atom. The change in KI/II is
undoubtedly not due to steric effects but rather to
electronic differences between these two substituents.
As suggested by Boulton et al.,5 weak hydrogen bonding
and thus accentuation of hyperconjugation of the 7-
methyl group with the aza-nitrogen are of small
importance because there is no meaningful change in
the 17O NMR chemical shifts for compound 4 (Table 4).
Finally, we have also become interested in the effects
1
of protonation on the compounds studied. The H, 13C
and 15N NMR spectra of compounds 1, 3, 4, 6, 7 and 10
were measured in the presence of excess trifluoroacetic
acid (TFA) and the results are shown in Table 7. The
main conclusion which can be drawn from an analysis of
the data is that under acidic conditions the equilibrium is
totally shifted towards isomer I. First, only one set of the
signals is observed in the spectra of all compounds
studied when measured at 269 K. Second, the signal
corresponding to carbon C7a appears in the 13C NMR
spectra of all protonated samples at ca. 110 ppm, which is
very close to the position of this signal in the spectra of
non-protonated compounds, ca. 108 ppm. These results
provide unquestionable evidence that, as mentioned
above, the protonated compounds exist exclusively in
the form I. Additionally, according to the mesomeric
structures given in Fig. 8, protonation of the molecules
under consideration should strongly stabilize the N1-
oxide (form I). It is worth noting again at this point that
Table 6. Equilibrium constants for the N1-N3-oxide rearran-
gement
Compound
Constant
Compound
Constant
1
2
3
4
5
0.10 Æ 0.005
0.08 Æ 0.005
0.12 Æ 0.005
0.18 Æ 0.005
0.05 Æ 0.005
6
7
0.15 Æ 0.005
0.00
8
1.22 Æ 0.005
0.00
9
10
Figure 6. Plot of calculated 15N absolute shieldings against
0.00
experimental 15N NMR chemical shifts
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 480–488

FUROXAN REARRANGEMENT OF PYRIDOFUROXAN DERIVATIVES
487
Table 7. 1H, 13C and 15N NMR chemical shifts for some protonated (in TFA) and alkylated (in acetone, T = 269 K) compounds
1
1a
8.56
7.04
7.81
1 etb
9.78
8.11
9.11
3
4
4 etb
9.50
6
7
10 Tc
H5
H6
H7
CH3
C7a
C3a
C5
C6
C7
N1
N3
N4
8.65
7.09
7.69
8.46
—
8.44
6.92
—
2.29
—
—
—
—
—
17.6
21.7
8.02
—
7.15
—
109.6
155.2
160.7
132.6
122.1
23.1
—
6.42
7.30
3.66
109.1
157.3
169.7
121.8
124.9
30.7
30.0
191.9
—
7.34
7.43
7.87
7.59
1.98
110.9
150.7
161.1
136.4
129.0
21.2
—
1.64
114.4
—
—
—
—
d
111.2
151.0
158.7
123.3
133.0
18.2
111.0
152.7
158.2
123.1
129.7
114.2
150.9
161.4
125.1
136.5
17.2
107.5
144.1
150.3
114.3
121.4
21.8
d
d
d
d
150.7
159.2
123.7
154.5
e
f
—
—
e
e
f
e
17.0
175.2
13.6
150.5
—
—
173.6
—
—
116.1
30.4
158.2
f
188.9
212.0
—
a
At 303 K
b ethylated form of compounds 1 and 4.
c 15N NMR chemical shifts of remaining nitrogen nuclei of tetrazole form (T) of compound 10: N1' = 78.6 ppm, N2' = ‡7.4 ppm, N3' = 31.2 ppm.
d Not recorded.
e
1
Not observed in H–15N GHMBC experiment.
f Not recorded because of instability of the ethylated derivative.
the C7a signals in the spectra of isomer II appear in the
region of 148 ppm. A further comparison of the data
obtained for the protonated compounds with those
derived for the non-protonated compounds shows that
the protonation takes place on the nitrogen atom of the
pyridine ring, i.e. on N4. This results in very strong
shielding of this atom in comparison with the non-
protonated molecule. Its magnitude depends dramatically
on the substituents attached to the pyridine ring and
varies from 30 ppm in compound 6 up to 110 ppm in
compound 4. In the first case the chlorine atom in position
6 probably strongly decreases the electron density at N4,
weakening the protonation effect considerably, and the
methyl group at position 7, in contrast, increases the
electron density at N4, strengthening the shielding effect
caused by protonation of this atom. The changes
occurring at N4 are accompanied by shielding of carbon
C3a and deshielding of C7, 9 and 10 ppm, respectively.
All these values are very close to the protonation effects
in pyridine itself and its corresponding derivatives. Thus,
for example the corresponding D(ꢂ) values in unsub-
in TFA in comparison with those determined for the non-
protonated compound may be considered as a result
typical of protonation of the pyridine ring.17
For labelled compound 1 we were also able to measure
the 15N NMR spectrum in TFA at 303 K. At this
temperature the N3 and N4 signals are shifted to higher
frequencies, by 3.4 and 22.0 ppm, respectively, which
indicates that the proton–nitrogen interaction is, as
expected, weaker at higher than at lower temperatures.
It is also of interest that similar effects to those
described above are observed upon N4 alkylation of the
compounds studied. Thus, a nitrogen shielding increase
of about 100 ppm has been observed in the N4-ethyl
derivatives of compounds 1 and 4 for nitrogen N4 and
deshielding of carbon C7 by 10 ppm (see Table 7).
In order to obtain a deeper insight into the electron
1
distribution in pyridofuroxans, we measured also JCC
couplings for compound 1 and its protonated form, and
for the compounds which are pyridofuroxan precursors,
11, 12 and 13. These results are given in Table 8.
Thermolysis of compound 11 and oxidation of com-
pounds 12 and 13 (Fig. 2) provide a change in electron
distribution and result in bond localization. The most
sensitive NMR parameters to these changes are 1JCC and
nitrogen chemical shifts (Table 8). Especially visible are
the changes of coupling constants (C5–C6) between
stituted pyridine are
115.6 ppm,
7.7 ppm and
‡12.4 ppm for N4, C2 and C4, respectively (J Wiench,P
Cmoch, L Stefaniak and GA Webb. J. Mol. Struct. in
1
1
preparation). Also, an increase of JC5C6 and JC3aC7a
couplings of 5 Hz observed in the spectrum of 1 recorded
Figure 8. Mesomeric structures of pyridofuroxan
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 480–488

488
P. C. CMOCH ET AL.
1
Table 8. JCC coupling constants for compounds 1, 11, 12,
13 and the protonated form of 1
Acknowledgements
Support for this work was partially provided by a grant
from the Polish State Committee for Scientific Research,
No. 3 T09A 010 12.
1
1a
11b
12
13
c
C7a–C3a
C5–C6
C6–C7
C7–C7a
65.5
48.8
61.4
64.1
71.2
53.7
58.8
65.8
—
71.2
51.4
56.0
67.8
76.2
50.5
59.5
55.8
65.5
52.7
71.9
a
For protonated form.
Numbering of atoms for compounds 11, 12 and 13 in Fig. 2; for
b
compound 11, corresponding couplings: C8–C8a = C7a–C3a, C5–
REFERENCES
C6 = C5–C6, C6–C7 = C6–C7 and C7–C8 = C7–C7a.
c
Not observed.
1. Gosco A, Boulton AJ. Adv. Heterocycl. Chem. 1981; 29: 251.
2. Paton RM. In Comprehensive Heterocyclic Chemistry Part II, vol.
4, Katrizky AR (ed). Pergamon Press: New York, 1996; 263.
3. Dunkin IR, Lynch MA, Boulton AJ, Henderson N. J. Chem. Soc.,
Chem. Commun. 1992; 1178.
4. Murata S, Tomioka H. Chem. Lett. 1992; 55.
5. Boulton AJ, Halls PJ, Katrizky AR. J. Chem. Soc. B 1970; 636.
6. Lowe-Ma ChK, Nissan RA, Wilson WS. J. Org. Chem. 1990; 55:
3755.
compounds 11 and 1. The value of 1JC5C6 is 65.5 Hz for
11 and 48.8 Hz for 1, showing the changes in bond
localization for these two different compounds.
7. Malkin VG, Malkina OL, Casida ME, Salahub DR. J. Am. Chem.
Soc. 1994; 116: 5898.
CONCLUSIONS
8. Anet FAL, Yavari I. Org. Magn. Reson. 1976; 8: 158.
9. Allen FH, Kennard O, Watson DG, Bremmer L, Orpen AG, Taylor
R. J. Chem. Soc., Perkin Trans. 2 1987; 1.
10. Britton D, Olson JM. Acta Crystallogr., Sect. B 1979; 35: 3076.
11. Witanowski M, Stefaniak L, Webb GA. Annu. Rep. NMR
Spectrosc., 1993; 25: 282.
12. Witanowski M, Stefaniak L, Webb GA. Nitrogen NMR. Plenum
Press: London, 1973.
13. Cmoch P, Wiench J, Stefaniak L, Webb GA. J. Mol. Struct. 1999;
510: 165.
14. Christ HA, Diehl P, Schneider HR, Dohn H. Helv. Chim. Acta
1961; 44: 865.
15. Cmoch P, Wiench J, Stefaniak L, Webb GA. Spectrochim. Acta,
Part B 1999; 55: 2207.
16. Katritzky AR, Gordeev MF. Heterocycles, 1993; 35: 483.
17. Denisov AYu, Mamatiuk WI, Szkurko OP. Khim. Geterotsikl.
Soedin. 1988; 1243.
1
We conclude that H, 13C and 15N NMR parameters are
very well placed to provide information on the furoxan
type of equilibrium and also on possible mechanism of
furoxan formation. By means of these data it is possible
to distinguish between the two structures (N1-oxide and
N3-oxide) in equilibrium. A significant shielding increase
by more 100 ppm for N4 reveals that this atom is the site
of protonation and alkylation. Changes in 1JCC values and
nitrogen chemical shifts can be used to show that bond
localization between pyridofuroxans and their precursors
is completely different. X-ray results and DFT molecular
orbital calculations can be useful in supporting NMR
signal assignments.
Copyright 2000 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2000; 13: 480–488