5-(4'-Substituted-2'-nitroanilino)-1,2,3-triazoles as new potential potassium channel activators. I

Article abstract of DOI:10.1016/S0223-5234(00)00180-X

By the hypothesised correlation with the large conductance Ca⁺⁺- activated potassium channel (BK(Ca)) openers NS 004 and NS 1619, bearing a benzimidazolone ring, a series of new 5-(4'-substituted-2'-nitroanilino)- 1,2,3-triazoles were synthesised and tested on in vitro isolated vascular preparation. The compounds were prepared starting from the appropriately substituted 2-nitro-phenylazides by 1,3-dipolar cycloaddition reaction to cyanoacetamide and following Dimroth isomerisation of the corresponding 1- arylsubstituted-5-amino-1,2,3-triazoles. The analogous 5-(4'-substituted-2'- amino-anilino)-1,2,3-triazoles were also prepared to assess the role of the nitro group in the pharmacophoric model. Almost all the nitro compounds showed a vasorelaxant activity on endothelium-denuded rat aortic rings with a potency comparable to that recorded for the reference compound NS 1619. Such a vasorelaxing activity was significantly reduced by the increase of the level of membrane depolarisation and by the potassium channel blocker 4- aminopyridine with a pharmacodynamic behaviour consistent with a potassium channel activation. (C) 2000 Editions scientifiques et medicales Elsevier SAS.

Full text of DOI:10.1016/S0223-5234(00)00180-X

Eur. J. Med. Chem. 35 (2000) 715−720
© 2000 Éditions scientifiques et médicales Elsevier SAS. All rights reserved
S022352340000180X/FLA
Preliminary communication
5-(4′-Substituted-2′-nitroanilino)-1,2,3-triazoles as new potential potassium
channel activators. I
Giuliana Biagia, Vincenzo Calderonea, Irene Giorgia, Oreste Livia*, Valerio Scartonia, Barbara
Baragattib, Enrica Martinottib
aDipartimento di Scienze Farmaceutiche, Università degli Studi di Pisa, via Bonanno 6, 56126 Pisa, Italy
bDipartimento di Psichiatria, Neurobiologia, Farmacologia e Biotecnologie, Università degli Studi di Pisa, via Bonanno 6, 56126 Pisa, Italy
Received 12 October 1999; revised 1 February 2000; accepted 6 March 2000
Abstract – By the hypothesised correlation with the large conductance Ca++-activated potassium channel (BKCa) openers NS 004 and NS
1619, bearing a benzimidazolone ring, a series of new 5-(4′-substituted-2′-nitroanilino)-1,2,3-triazoles were synthesised and tested on in vitro
isolated vascular preparation. The compounds were prepared starting from the appropriately substituted 2-nitro-phenylazides by 1,3-dipolar
cycloaddition reaction to cyanoacetamide and following Dimroth isomerisation of the corresponding 1-arylsubstituted-5-amino-1,2,3-
triazoles. The analogous 5-(4′-substituted-2′-amino-anilino)-1,2,3-triazoles were also prepared to assess the role of the nitro group in the
pharmacophoric model. Almost all the nitro compounds showed a vasorelaxant activity on endothelium-denuded rat aortic rings with a potency
comparable to that recorded for the reference compound NS 1619. Such a vasorelaxing activity was significantly reduced by the increase of
the level of membrane depolarisation and by the potassium channel blocker 4-aminopyridine with a pharmacodynamic behaviour consistent
with a potassium channel activation. © 2000 Éditions scientifiques et médicales Elsevier SAS
1,2,3-triazoles / potassium channel activators
1. Introduction
preliminarily tested on in vitro isolated endothelium
denuded rat aortic preparations, showing strong vasore-
laxing properties [5].
In previous papers we reported synthesis and structure
evidence of new 1-(1,2,3-triazol-4-yl)-benzotriazoles [1]
and of analogous 1-(1,2,3-triazol-4-yl)-benzimidazolones
[2]. These structures attracted our interest for medicinal
chemistry because they may be correlated to NS 004 and
NS 1619 [3], which are two of the few actually available
openers of the large conductance Ca++-activated potas-
sium channels (BKCa). Such a pharmacological activity
represents an interesting approach in the treatment of
several cardiovascular disorders without the typical side-
effects of classical vasodilators [4].
The preparation of both series of compounds, triazolyl-
benzotriazoles or triazolyl-benzimidazolones, followed a
synthetic route based upon common 5-anilino-triazole
intermediates, consisting of a structure which could also
be correlated to the above mentioned openers of the BKCa
channels (figure 1). Therefore some intermediates were
2. Chemistry
This paper describes the synthesis and pharmacologi-
cal evaluation of a series of 5-(4′-substituted-2′-
nitroanilino)-1,2,3-triazoles (2a–g) and of the correspond-
ing aminoderivatives 3a–g (figure 2).
All the azides, excluding the 4-sec-butyl-2-nitro-
phenylazide 1g, have been described in the literature and
were prepared from the appropriate primary aromatic
amines via diazonium salt and treatment with sodium
azide: 2-nitrophenylazide 1a [6], 4-methyl-2-nitrophenyl-
azide 1b [7], 4-methoxy-2-nitrophenylazide 1c [8],
4-chloro-2-nitrophenylazide
1d
[9],
4-fluoro-2-
nitrophenylazide 1e [10] and 4-trifluoromethyl-2-
nitrophenylazide 1f [10].
The 1,3-dipolar cycloaddition reaction of the appropri-
ate azide 1a–g to the activated methylene compound
* Correspondence and reprints: livi@farm.unipi.it

716
G. Biagi et al. / Eur. J. Med. Chem. 35 (2000) 715–720
The structures of all the new compounds were assigned
on the basis of the well-known reaction mechanisms, as
shown in previous works [1] and [2], and were confirmed
by analytical and spectroscopic data.
3. Pharmacology
There is actually strong evidence of a hyperpolarisa-
tion of the vascular smooth muscle cells, mediated by
drugs activating the BKCa potassium channels, whose
presence has been widely demonstrated (by functional
and/or electrophysiological investigations) in many cir-
culatory areas, such as the rabbit pulmonary artery [11],
the rat portal vein [12] and the rat aorta [13]. Thus, the
functional evaluation of a vasorelaxing activity of the
tested compounds was chosen as a preliminary screening
method to unmask a possible potassium channel opening
effect.
4. Results and discussion
Almost all the compounds bearing the nitro group on
the benzene ring, and the carboxamido substituent on the
1,2,3-triazole heterocyclic moiety, showed vasorelaxing
properties (table II). These effects consisted of a full
abolition of the contractile tone induced by the depola-
rising stimulus (administration of KCl 20 mM), with
potency orders of magnitude weakly lower or also com-
parable to that recorded for the reference compound
NS1619. Only the fluorine-substituted compound 2e,
belonging to this series, showed an unusual profile of
activity, with a very low efficacy (< 50%) that made the
calculation of the potency value impossible. This last
experimental observation could be probably explained by
the strong electronic influence of the fluorine atom,
determining a dramatic negative impact on the effective-
ness of the molecule; however, more detailed hypotheses
are, actually, premature, because of the few available
data. Other preliminary observations of the structure–ac-
tivity relationships allowed us to observe that the absence
of any substituent in position 4 of the benzene group (2a)
led to a decrease of potency (but not of efficacy), whilst
the presence of the sec-buthylic substituent could be
correlated with the highest value of potency, recorded for
the compound 2g. This preliminary experimental data
could also suggest a direct correlation between the
potency and the steric hindrance or the lipophilicity in
this position; whilst the electronegativity of the substitu-
ent would appear to exert a negative influence on the
potency.
Figure 1. Chemical structures and two orthogonal hypothetical
tridimensional overlaps of NS1619 (grey lines) and 5-(4-
trifluoromethyl-2-nitroanilino)4-carboxamido-1,2,3-triazole
(black lines).
cyanoacetamide was carried out in the usual manner in
anhydrous ethanol in the presence of sodium ethoxide at
room temperature for 1 night (figure 2).
Generally
the
insoluble
1-(4′-substituted-2′-
nitrophenyl)-4-carboxamido-5-amino-1H-1,2,3-triazole
intermediate was not isolated, but it was directly con-
verted to the desired Dimroth isomer by addition of a
small amount of 10% sodium hydroxide solution and
heating of the reaction mixture.
The Dimroth isomers 2a–g, which turned acidic be-
cause of the presence of the triazole proton, went in the
aqueous alkaline solution and were isolated by acidifica-
tion after previous washing with an organic solvent (table
I). The amino derivatives 3a–g (figure 2) were obtained
by catalytic hydrogenation of the corresponding nitro
derivatives 2a–g at room temperature and pressure. As
previously described [1], the carboxylic acid 4 and the
corresponding decarboxylated compound 5 (figure 3)
were obtained by alkaline hydrolysis of 1a and heating of
the acid 4 in DMF, respectively.
The replacement of the carboxamido substituent, bound
to the triazole ring in the active compound 2a, by the

G. Biagi et al. / Eur. J. Med. Chem. 35 (2000) 715–720
717
Figure 2. Preparation of
compounds 2 and 3.
carboxylic group (4) or by the hydrogen atom (5),
determined an almost complete fall of the vasorelaxant
activity.
experimental condition, significant decreases of potency
(IC50 = 4.22 ± 0.018) and of efficacy (Emax = 84 ± 4%)
could be observed, fitting the profile of response expected
for a potassium channel opener. Surprisingly, the refer-
ence benzimidazolone compound NS1619 did not show
any significant decrease of activity when administered to
preparations pre-contracted by KCl 80 mM. This anoma-
lous behaviour could be due to different ancillary mecha-
nisms of actions, whose existence has been already
suggested by the literature [13]. Furthermore, the vasore-
laxing activity of 2f was also significantly inhibited
(IC50 = 4.25 ± 0.025, Emax = 71 ± 3) by 4-aminopyridine
(3 mM), a widely used non-selective blocker of several
potassium channel subtypes.
It is concluded that the 5-(4′-substituted-2′-nitro-
anilino)-4-carboxamido-1,2,3-triazole derivatives can rep-
resent a new chemical class of vasodilators. The first
preliminary experimental observation can suggest a po-
tassium channel opening mechanistic hypothesis. An
accurate pharmacological characterisation of the target of
these drugs is actually in progress by means of functional
and electrophysiological tools.
To assess the importance of the presence of the oxygen
atom(s) of the nitro group, clearly mimicking the position
(and the function of an H-bond acceptor?) of the carbonyl
group of the benzimidazolone reference compounds, the
nitro group was replaced by the corresponding amino
group. All the tested compounds possessing this struc-
tural characteristic (3b, c, f and g) were devoid of any
significant vasorelaxing activity.
Finally, a further investigation of the possible potas-
sium channel opening mechanism of action was also
performed in the isolated vessels. It is widely reported
that drugs, acting through the potassium channel activa-
tion, possess the highest degree of ‘depolarisation-
sensitive’ inhibition of the functional responses among
the several classes of vasodilators [14]. Therefore, the
selected compound 2f (IC50 = 4.79 ± 0.089, Emax = 100)
was also tested on isolated aortae, whose contractile tone
was induced by a higher level of membrane depolarisa-
tion (due to the administration of KCl 80 mM). In such an

718
G. Biagi et al. / Eur. J. Med. Chem. 35 (2000) 715–720
Table I. Chemical and physical properties of derivatives 2 and 31.
Compound
Yield (%)
Crystall.
solvent
M.p. (°C)
Mass. M+
m/z. base
peak
Elemental
analysis
Calcd./found
H
C
N
2b
2c
2d
2e
2f
56
31
25
30
28
46
83
86
91
84
96
91
EtOH
AcOH
MeOH
MeOH
MeOH
MeOH
EtOH
278–282
dec
271–274
dec
303–307
dec
285–298
dec
253–257
174–176
189–191
191–194
178–182
202–205
168–171
165–168
262
278
282
266
316
304
232
248
252
236
286
274
216
232
236
44
C10H10N6O3
C10H10N6O4
C9H7N6O3Cl
C9H7N6O3F
C10H7N6O3F3
C13H16N6O3
C10H12N6O
C10H12N6O2
C9H9N6OCl
C9H9N6OF
45.80
45.68
43.17
43.27
38.25
38.43
40.61
40.57
37.99
38.04
51.31
51.12
51.72
51.85
48.38
48.36
42.78
42.89
45.76
45.73
41.96
42.29
56.92
56.64
3.84
3.51
3.62
3.90
2.50
2.46
2.65
2.71
2.23
2.29
5.30
5.06
5.21
5.18
4.87
4.56
3.59
3.31
3.84
3.51
3.17
2.97
6.61
6.84
32.05
32.15
30.21
30.51
29.73
29.79
31.57
31.51
26.58
26.71
27.62
27.29
36.19
36.10
33.85
34.17
33.26
32.99
35.58
35.69
29.36
29.03
30.64
30.31
270
44
2g
3b
3c
3d
3e
3f
44
EtOH
133
44
H2O
H2O
44
H2O
286
44
C10H9N6OF3
C13H18N6O
3g
MeOH/ H2O
1 Compounds 2a and 3a are reported in the literature [1].
5. Experimental protocols
(C, H, N) were within ± 0.4% of the theoretical values
and were performed on a Carlo Erba Elemental Analyzer
Mod. 1106 apparatus. TLC data were obtained with
Riedel de Haen, 37360 DC-Karten F254, 0.2 mm, eluting
with a 1:3 AcOEt/petroleum ether mixture. Petroleum
ether corresponds to the fraction boiling at 40–60 °C.
5.1. Chemistry
Melting points were determined on a Kofler hot-stage
and are uncorrected. IR spectra in nujol mulls were
recorded on a Mattson Genesis series FTIR spectrometer.
1H-NMR spectra were recorded with a Varian CFT-20
spectrometer in DMSO-d6 in d units, using TMS as the
internal standard. Mass spectra were performed with a
Hewlett Packard MS/System 5988. Elemental analyses
Table II. Potency (pIC50) and efficacy (Emax %) values of the
vasorelaxing compounds 2a–g1.
Compound
pIC50
Emax %
2a
2b
2c
2d
2e
2f
2g
NS 1619
3.98 ± 0.066
4.73 ± 0.11
4.55 ± 0.11
4.61 ± 0.048
not calculated
4.79 ± 0.089
5.14 ± 0.049
5.30 ± 0.29
100
100
100
100
41 ± 1
100
100
100
1 The tested compounds 3b, c, f, g, 4 and 5 did not show significant
vasorelaxant effects.
Figure 3. Previously prepared compounds 4 and 5.

G. Biagi et al. / Eur. J. Med. Chem. 35 (2000) 715–720
719
Short distillations were performed with a Buchi GKR 50
tubular oven.
5.1.4. 4-Carboxamido-5-(4-
substituted-2-nitroanilino)-1,2,3-triazoles 2b–g
To a stirred solution of sodium ethoxide (0.41 g,
18 mmol of Na) in 25 mL of absolute ethanol, 1.50 g
(18 mmol) of cyanoacetamide were added. After
15–20 min the suspension was cooled in an ice-bath
(–10 °C) and 15 mmoles of the appropriate 4-substituted-
2-nitrophenylazide (1b, 1c, 1d, 1e, 1f or 1g) in 25 mL of
absolute ethanol were added drop by drop keeping the
temperature < 0–5 °C. After 1 h the ice-bath was removed
and stirring was continued at room temperature for
1 night. The solvent was evaporated under reduced pres-
sure, H2O and 3–5 mL of 10% NaOH were added and the
mixture was heated under reflux for 15–20 min. After
cooling, the reaction mixture was paper filtered from the
formed nitroaniline and by acidification (pH 1–2) of the
filtrate the title compounds precipitated and were col-
lected by filtration (table I).
Compounds 2a and 3a are reported in the literature [1].
5.1.1. 4-sec-Butyl-2-nitro-acetanilide
To an ice-cooled (0–5 °C) and stirred solution of
4-sec-butyl-acetanilide [15] (4.50 g, 23.5 mmol) in
100 mL of AcOH, 30 mL of fuming HNO3 (d 1.482) were
added dropwise. The ice-bath was removed, the mixture
was heated at 50 °C for 6 h and, after cooling, it was
poured into crushed ice. The suspension was alkalinised
(pH 10) with solid Na2CO3 and extracted with CHCl3.
After evaporation of the solvent, the residue was stirred
with AcOEt/petroleum ether 1:9 mixture and the in-
soluble material, consisting of the starting solid, was
filtered off. This separation could be achieved by flash-
chromatography through a silica gel column, eluting with
AcOEt/petroleum ether 1:3 mixture. Evaporation of the
solvent gave the title compound as an oil (4.03 g, yield
72%) which could be purified by short distillation
at 160–170 °C/2 mm Hg; m.p. 14–16 °C (Lit. [16] 15–
17 °C). Anal.: C12H16N2O3. Found: C 61.05, H 6.62, N
11.52; calculated: C 61.02, H 6.78, N 11.86. IR (cm–1):
1 514 and 1 360 (NO2). MS (m/z,): 236 [M+], 43 [100].
5.1.5. 4-Carboxamido-5-(4-
substituted-2-aminoanilino)-1,2,3-triazoles 3b–g
To a solution or suspension of the appropriate nitro-
derivative 2b–g (5.0 mmol) in ≈ 200 mL of MeOH, 10%
Pd/C (0.100 g) was added and the mixture was hydroge-
nated at room temperature and pressure. The catalyst was
filtered off, washed with hot MeOH and the filtrate was
evaporated in vacuo to give the title compounds (table I).
5.1.2. 4-sec-Butyl-2-nitroaniline
5.2. Pharmacology
A solution of 4-sec-butyl-2-nitro-acetanilide (4.0 g,
17 mmol) in 40 mL of 50% H2SO4 was heated under
reflux for 1 night. After cooling the mixture was alkalin-
ised (pH 10) with 10% NaOH and extracted with
CHCl3. The solution, dried (MgSO4) and evaporated in
vacuo providing the title compound as an oil (3.0 g, yield
92%) which was purified by short distillation at
165–170 °C/2.5 mm Hg). Anal.: C10 H14N2O2. Found: C
62.08, H 7.03, N 14.08; calculated: C 61.86, H 7.22, N
14.43. IR (cm–1): 3 490 and 3 371 (NH2). MS (m/z): 194
[M+], 165 [100].
All the procedures, performed on experimental ani-
mals, were carried out following the guidelines of the
European Community Council Directive 86-609.
To determine a possible vasodilator mechanism of
action, the compounds were tested on isolated thoracic
aortae of male normotensive Wistar rats (250–350 g). The
rats were killed by cervical dislocation under light ether
anaesthesia and bled. The aortae were immediately ex-
cised, freed of extraneous tissues and the endothelium
was removed by gently rubbing the intimal surface of the
vessels. Aortic rings were suspended, under a preload of
2 g, in 10 mL organ baths containing Tyrode solution
(composition of saline in mM: NaCl 136.8; KCl 2.95;
CaCl2 1.80; MgSO4·7H2O 1.05; NaH2PO4 0.41; NaHCO3
11.9; glucose 5.5), thermostated at 37 °C and continu-
ously bubbled with a mixture of O2 (95%) and CO2 (5%).
Changes in tension were recorded by means of an
isometric transducer (Basile mod. 7005), connected with
a unirecord microdynamometer (Basile mod. 7050).
After an equilibration period of 60 min the endothelial
integrity was confirmed by acetylcholine (ACh) (55 µM)-
induced relaxation of norepinephrine (NE, 1 µM)-
precontracted tissues. A relaxation < 20% of the NE-
5.1.3. 4-sec-Butyl-2-nitrophenylazide 1g
To a stirred and cooled (0–5 °C) solution of 4-sec-
butyl-2-nitroaniline (3.0 g, 15.5 mmol) in 100 mL of 18%
HCl, 10 mL of aqueous solution of NaNO2 (1.30 g,
18.8 mmol) were added drop by drop. After 15–20 min, a
solution of NaN3 (1.20 g, 18.8 mmol) in 10 mL of H2O
was added dropwise under vigorous stirring. After ap-
proximately 1 h the solution was extracted with Et2O and
the organic layer was dried (MgSO4) and evaporated to
give the title compound as an oil which was used without
purification: 2.60 g, yield 76%; IR (cm–1): 2 130 (N3).

720
G. Biagi et al. / Eur. J. Med. Chem. 35 (2000) 715–720
induced contraction was considered representative of an
acceptable lack of the endothelial layer, while the organs
showing a relaxation ≥ 20% (i.e. significant presence of
the endothelium), were not used in the experimental
procedures. 30–40 min after the confirmation of the
endothelium removal, the aortic preparations were con-
tracted by treatment with a single concentration of KCl
(20 mM) and when the contraction reached a stable
plateau, 3-fold increasing concentrations of the tested
compounds and of the reference compound NS 1619
(10 nM to 1 mM) were added cumulatively. In parallel
sets of experiments, to investigate the influence of a
higher level of depolarisation on the responses evoked by
the tested compound, the aortic preparations were con-
tracted by KCl 80 mM. Then, 3-fold increasing concen-
trations of the compounds (10 nM to 1 mM) were added
cumulatively. Furthermore, the vasorelaxing activity was
also evaluated on KCl 20 mM-contracted vessels, in the
presence of the potassium channel blocker
4-aminopyridine (3 mM).
Preliminary experiments showed that both the KCl (20
and 80 mM)-induced contractions remained constant in a
stable tonic state for at least 40 min. Norepinephrine
hydrochloride (Sigma), acetylcholine chloride (Sigma),
4-aminopyridine and KCl were dissolved in bi-distilled
water. All the synthesised 1,2,3-triazole derivatives and
the reference compound NS 1619 (RBI) were dissolved
(10 mM) in NaOH 0.1 N. All the further dilutions were
performed in bi-distilled water. All the solutions were
prepared immediately before the pharmacological experi-
mental procedures. Previous experiments showed a com-
plete ineffectiveness of the administration of the vehicle.
substantially induce a maximal effect in endothelium
denuded aortic rings.
The parameter of potency of the vasorelaxing effects
was expressed as pIC50, representing the negative loga-
rithm of the vasodilator molar concentration determining
a half reduction of the contractile tone induced by the
contractile agent. The above parameter was calculated by
means of a non-linear regression analysis of the sigmoi-
dal concentration–response curves (computer program:
GraphPad Prism).
All the results are expressed as mean SEM of 4–6
experiments. The statistical comparison of experimental
data was performed by two-tailed Student t-test and
Anova. A value of P < 0.05 was considered as represen-
tative of significant differences.
References
[1] Biagi G., Giorgi I., Livi O., Scartoni V., Velo S., Barili P.L.,
J. Heterocycl. Chem. 33 (1996) 1847–1853.
[2] Bertelli L., Biagi G., Calderone V., Giorgi I., Livi O., Scartoni V.,
Barili P.L., J. Heterocycl. Chem. (in press).
[3] Olesen S.P., Jensen L.H., Moldt P., European Patent Application
(1974), EPA 617023.
[4] Cook N.S., Potassium Channels: Structure, Classification, Function
and Therapeutic Potential, Ellis Horwood Ltd., Southampton, 1990.
[5] Calderone V., Nieri P., Baragatti B., Martinotti E., Livi O., 6th Joint
Meeting of the Italian, Hungarian and Polish Pharmacological
Societies, May 14–16, 1998, Pisa, p. 46.
[6] Smith P.A.S., Boyer J.H., Benzofurazan oxide, in: Schreiber R.S.
(Ed.), Organic Syntheses, Wiley, NY, 1951, pp. 14–16.
[7] Tanaka Y., Velen S.R., Miller S.I., Tetrahedron 29 (1973) 3271–3283.
[8] Dyall L.K., Kemp J.E., Aust. J. Chem. 20 (1967) 1395–1402.
[9] Dyall L.K., Aust. J. Chem. 39 (1986) 89–101.
[10] Ghosh P.B., Ternai B., Whitehouse M.W., J. Med. Chem. 15 (1972)
5.3. Data analysis
255–260.
[11] Vandier C., Bonnet P., Eur. J. Pharmacol. 295 (1996) 53–60.
The efficacy of the vasorelaxing responses was ex-
pressed as maximal relaxant effect (Emax), calculated as a
% of the contractile tone developed by the smooth muscle
preparation, after the depolarising stimulus induced by
KCl 20 or 80 mM. Previous experiments demonstrated an
almost complete quantitative equivalence between the
contractile responses evoked by the two different concen-
trations of KCl, since the concentration 20 mM could
[12] Edwards G., Niederste-Hollenberg A., Schneider J., Noack T.,
Weston A.H., Br. J. Pharmacol. 113 (1994) 1538–1547.
[13] Sargent C.A., Grover G.J., Antonaccio M.J., McCullough J.R.,
J. Pharmacol. Exp. Ther. 266 (1993) 1422–1429.
[14] Magnon M., Calderone V., Floch A., Cavero I., Naunyn Schmiede-
berg’s Arch. Pharmacol. 358 (1998) 452–463.
[15] Reilly J., Hickinbottom W.J., J. Chem. Soc. 117 (1920) 103–137.
[16] Itaya M., Takai Y., Suzuki K., Yakugaku Zasshi 81 (1961) 921–923;
Chem. Abstr. 56 (1962) 3405.