Unusual intramolecular cyclization of tris(β-oximinoalkyl)amines. the first synthesis of 1,4,6,10-tetraazaadamantanes

Article abstract of DOI:10.1021/ol9015157

An unusual intramolecular cyclizatlon of tris(β-oximinoalkyl)amines 1 Into 4,6,10-trihydroxy-1,4,6,10-tetraazaadamantanes 2 was discovered. Compounds 2 are related to a previously unknown type of heteroadamantanes that contain the cage isomeric to urotrop

Full text of DOI:10.1021/ol9015157

ORGANIC
LETTERS
2009
Vol. 11, No. 18
4072-4075
Unusual Intramolecular Cyclization
of Tris(ꢀ-oximinoalkyl)amines.
The First Synthesis of
1,4,6,10-Tetraazaadamantanes
Artem N. Semakin,† Alexey Yu. Sukhorukov,‡ Alexey V. Lesiv,‡ Sema L. Ioffe,*,‡
Konstantin A. Lyssenko,§ Yulia V. Nelyubina,§ and Vladimir A. Tartakovsky‡
Higher Chemical College of Russian Academy of Sciences, 125047, Miusskaya str.,
9, Moscow, Russian Federation, N.D. Zelinsky Institute of Organic Chemistry, 119991,
Leninsky prosp. 47, Moscow, Russian Federation, and A.N. NesmeyanoV Institute of
Organoelement Compounds, VaViloV Str., 28, 119991, Moscow, Russian Federation
iof@ioc.ac.ru
Received July 2, 2009
ABSTRACT
An unusual intramolecular cyclization of tris(ꢀ-oximinoalkyl)amines 1 into 4,6,10-trihydroxy-1,4,6,10-tetraazaadamantanes 2 was discovered.
Compounds 2 are related to a previously unknown type of heteroadamantanes that contain the cage isomeric to urotropin. A simple three-step
synthesis of tetraazaadamantanes 2 and their N-substituted derivatives 3 and 4 from ammonia and aliphatic nitro compounds via the intermediacy
of available tris-oximes 1 was developed.
Recently, we reported a simple and efficient synthesis of
tris(ꢀ-oximinoalkyl)amines 1 from secondary nitro com-
Scheme 1. Intramolecular Cyclization of Tris-oxime 1a
pounds RCH(NO2)CH3 and ammonia.1 In the course of a
systematic study on the reactivity of these interesting but
rather poorly explored compounds, we discovered a quite
unusual intramolecular cyclization of tris-oxime 1a (R )
CH3) to 4,6,10-trihydroxy-1,4,6,10-tetraazaadamantane 2a
(Scheme 1). Such a reaction of poly-oximes has never been
observed before.2 However, a few examples of successful
cyclization of tris-imines and some other nitrogen-containing
† Higher Chemical Colledge.
‡ N.D. Zelinsky Institute of Organic Chemistry.
§ A.N. Nesmeyanov Institute of Organoelement Compounds.
(1) Semakin, A. N.; Sukhorukov, A. Yu.; Lesiv, A. V.; Khomutova,
Yu. A.; Ioffe, S. L.; Lyssenko, K. A. Synthesis 2007, 2862.
(2) Among closely related reactions, only two intermolecular trimer-
izations of oximes were reported: (a) Gardent, J. Ann. Chim. (Paris) 1955,
10, 413. (b) Sacca, A.; Freni, M. Gazz. Chim. Ital. 1956, 86, 199.
derivatives of tris-carbonyl compounds of general formula
X[CH2C(O)R]3 (X ) AlkC or heteroatom) as well as
cyclohexanes bearing three -CH2C(O)R groups in the C1,
C3, and C5 positions are known. These reactions were
10.1021/ol9015157 CCC: $40.75
Published on Web 08/14/2009
 2009 American Chemical Society

applied to the synthesis of some heteroadamantanes3 and
other heterocage structures (wurzitanes and isowurzitanes4).
Tetraazaadamantane 2a is assigned to an unknown class
of 1,4,6,10-tetraazaadamantanes. The unsubstituted 1,4,6,10-
tetraazaadamantane is a structural isomer of well-known
1,3,5,7-tetraazaadamantane (urotropin) that has found various
applications.5 Although the procedure for the synthesis of
urotropin itself from ammonia and formaldehyde is very
efficient,5,6a even the simplest C-substituted derivatives of
urotropin are almost unknown, apparently, because of their
relative thermodynamic instability.6b,c
From this standpoint, the synthesis of tetraazaadamantanes
of type 2 could be considered as an urgent but not quite
simple task. The comparison of physical and chemical
properties of heteroadamantanes of types 2 and urotropin
represents a fundamental interest. In addition, tetraazaada-
mantanes 2 or their derivatives could find some important
applications like urotropin. Therefore, it seems reasonable
to carry out a detailed study of the intramolecular cyclization
of tris-oximes 1 in order to optimize the procedure for
preparation of adamantanes 2.7 This study was performed
using tris-oxime 1a as a model compound.
Presumably, the generation of heteroadamantanes 2, which
are related to aminals, is a reversible reaction like the
synthesis of urotropin.6a In any case, reflux of adamantane
2a solution in water resulted in its full conversion to tris-
oxime 1a, while on the contrary, heating of tris-oxime 1a
up to a melting point (for 20 min) or in boiling water solution
(for 4-5 h), did not furnish the adamantane 2a. This indicates
that the open-chain form 1a is more thermodynamically
favored at high temperature than the cage structure 2a,
possibly due to a large contribution of an entropy factor in
the position of the equilibrium 1/2. In that case, in order
to shift the equilibrium to the adamantane structure 2, one
needs to minimize the entropy contribution by reducing the
reaction temperature. This effect can be attained by ac-
celerating the cyclization of 1 using different types of
additional promoting reagents (for details see the discussion
of Table 1).
An alternative way for shifting the equilibrium 1/2 to
the right may be the increasing of thermodynamic preference
of adamantane structure 2a, for example, by a quaternization
of the nitrogen atom in the bridgehead position (see the
discussions of Tables 1 and 2).
Expected effects of various factors on the rate of the
process 1 f 2 and on the position of the equilibrium 1/2
are briefly discussed below.
Action of Nucleophiles (Route (a) in Scheme 1). Addi-
tion of a nucleophile to the double C,N bond of 1a increases
the nucleophilicity of the nitrogen atom of this oximino group
in the reaction with a carbon atom of the neighboring
oximino fragment. The formation of the adamantane cage
is completed after the elimination of the nucleophile.
Action of Electrophiles (Routes b and b′ in Scheme
1). Brønsted acids (HX) and other electrophiles can promote
the formation of adamantanes 2 by two different pathways. First,
protonation of the nitrogen atom of the oximino group (route
(b)) provides an imminium cation, which can react with the
nitrogen atom of another oximino group. Two subsequent
cyclizations and elimination of proton from cage ammonium
cation furnish target adamantane 2a. However, protonation of
a more basic sp3 nitrogen atom of the tris-oxime 1a (route (b′))
leading to a corresponding ammonium cation seems to be more
likely. The generation of ammonium cation increases the
electrophilicity of carbon atoms of double C,N bonds due to a
-I-effect and favors the desired cyclization process. At the same
time, quaternization of the nitrogen may affect the relative
disposition of reacting oximino fragments.8 However, one
should take into account a possibility of deoximation as a side
reaction promoted by acids.9
Action of Transition-Metal Salts (Route (c) in
Scheme 1). The coordination of metal ion Mn+ with oxime
groups of tris-oxime 1a may bring oxime groups closer to each
other, thereby promoting their cyclotrimerization to give target
adamantane 2a.
The isomerization of model tris-oxime 1a into adamantane
2a with different types of co-reagents was studied. The results
are summarized in Table 1.
Positive results could be achieved with all types of reagents
employed. In reactions with nucleophiles (ammonia and sulfite
anion in H2O-MeOH) the target adamantane 2a was obtained
in moderate yield (entries 1 and 4, Table 1). When more basic
reagents (NaOH and NaCN) were employed (entries 2 and 3,
Table 1) only traces of adamantane 2a were detected after full
conversion of initial oxime 1a. Possibly, this could be caused
by a deoximation of 1a in alkaline media.9
Brønsted acids (HX) promote the formation of adamantane
2a (or its salts 2a·HX) as well (entries 5-7, Table 1).
However, when strong acids (HCl or TFA) are employed,
besides the respective salts of target adamantane 2a·HX (yield
30-40%), complex mixtures of unidentified products and
corresponding salts of hydroxylamine are generated. This
result indicates that the hydrolysis of initial oxime 1a or
adamantane 2a takes place substantially under these condi-
tions.
(3) (a) Quast, H.; Berneth, C.-P. Chem. Ber. 1983, 116, 1345. (b) Lukes,
R.; Syhora, K. Chem. Listy 1952, 46, 731. (c) Diaz-Alvarez, A. E.; Crochet,
P.; Zablocka, M.; Duhayon, C.; Cadierno, V.; Gimeno, J.; Majoral, J. P.
AdV. Synth. Catal. 2006, 348, 1671. (d) Kraemer, R.; Navech, J.; Mathis,
F.; Majoral, J. P. Tetrahedron 1976, 32, 2633.
(4) (a) Nielsen, A. T.; Christian, S. L.; Moore, D. W. J. Org. Chem.
1987, 52, 1656. (b) Izumi, H.; Setokuchi, O.; Shimizu, Y. J. Org. Chem.
1997, 62, 1173. (c) Izumi, H.; Setokuchi, O.; Shimizu, Y. J. Chem. Soc.,
Perkin Trans. 1 1998, 1925.
(5) For application of urotropin, see: Ullmann’s encyclopedia of
industrial chemistry; Wiley-VCH: Verlag, 2002.
(6) (a) Nielsen, A. T.; Moore, D. W.; Ogan, M. C.; Atkins, R. L. J.
Org. Chem. 1979, 44, 1678. (b) Kamal, A.; Ahmad, A.; Qureshi, A. Au.
Tetrahedron 1963, 19, 869. (c) Nielsen, A. T.; Chafin, A. P.; Christian,
S. L.; Moore, D. W.; Naddler, M. P.; Nissan, R. A.; Vanderah, D. J.
Tetrahedron 1998, 54, 11793.
(7) An alternative scheme for the synthesis of 2, an intermolecular
condensation NH3 + 3NH2OH + 3 RCOCH2LG f 2 (LG ) leaving group)
similar to a known synthesis of urotropin6 seems to be quite problematic
because it is a multicomponent reaction proceeding via many reversible
transformations of very unstable intermediates.
(8) According to the X-ray data, oximino fragments of tris-oxime 1a
are remote from each other in the crystal: Goldcamp, M. J.; Krause Bauer,
J. A.; Baldwin, M. J. Acta Crystallogr. Sect. E: Struct. Rep. Online 2002,
E58, 1354.
(9) Corsaro, A.; Chiaccio, U.; Pictara, V. Synthesis 2001, 1903.
Org. Lett., Vol. 11, No. 18, 2009
4073

the equilibrium is shifted to the corresponding adamantanes 2.
On the contrary, no evidence of adamantane structure is
observed in H NMR spectra of reaction mixtures for entries
5-7 in Table 2. Adamantanes 2a,b can be isolated from the
respective reaction mixtures as individual compounds.
Table 1. Intramolecular Cyclization of Tris-oxime 1a under
Action of Various Activating Reagents
1
amount of
time yield of
no.
reagent; conditions
reagent (equiv) (h) 2a (%)
1
2
3
4
5
6
7
8
9
NH3; MeOH/H2O, 60 °C
NaOH; MeOH/H2O, 60 °C
NaCN; MeOH/H2O, 60 °C
Na2SO3; MeOH/H2O, 20 °C
HCl; MeOH/H2O, 20 °C
TFA; MeOH/H2O, 20 °C
AcOH; MeOH/H2O, 20 °C
NiCl2; H2O, 20 °C
∼150
1.0
0.3
1.0
1.0
1.0
3.0
1.0
1.0
1.0
1.0
1.0
0.05
1.0
4.0 65a
2.5 tracesb,c
3.0 0c
Table 2. Cyclization of Tris-oximes 1a-g to the Corresponding
Azaadamantane Derivatives 2-4
48 34a,d
0.3 30b,e
0.3 40b,f
24 95a
48 0g
time (h)
yield (%)
3a 4a
no.
1
R
1f2 1f3 1f4
2
CuCl2; H2O, 20 °C
1.0 0g
1.0 0g
1
2
3
4
5
6
7
a
b
c
d
e
f
Me
H
Et
24
4.0
72
72
72
72
72
24
4.0
24
24
24
48
48
24
24
24
24
48
48
48
95a 100 94
10 ZnCl2; H2O, 20 °C
1.0 0g
89a 80
35b 86
59c 78
72
94
11 FeCl3; H2O, 20 °C
12 CoCl2; H2O, 20 °C
13 CoCl2; 1 equiv HCl, H2O, 20 °C
14 NiCl2; HCl/H2O, 20 °C
a On isolated product. From H NMR spectra with internal standard.
c Complex mixtures of products were obtained. d Conversion of 1a ∼60%.
e Product 2a·HCl (3a). f Product 2a·TFA. g Only stable metallo-complexes
of tris-oxime 1a were identified. h Possibly, HCl is generated due to a partial
hydrolysis of CoCl2.
72 26a,e,h
0.5 80a,e
1.0 0g
(CH2)2CO2Me
Bn
Ph
CO2Et
95
89
n.r.
n.r.
d
b
1
n.r.
n.r.
n.r.
d
d
g
a On isolated product. b Recovery of 1c: 55%; isolated product 2c
contains ∼20% of 1c (NMR). c Recovery of 1d: 36%; isolated product 2d
contains ∼10% of 1d (NMR). d Only products of decomposition of 1. n.r.
) no reaction.
Reactions of transition-metal salts with tris-oxime 1a
usually led to an instantaneous and quantitative formation
of stable metallo-complexes.10 However, in the case of CoCl2
(entry 12, Table 1) a small amount of adamantane salt 2a·HCl
(3a) was isolated by crystallization from the reaction mixture.
Optimization allowed us to realize the synthesis of 3a on a
preparative scale by using a catalytic amount of CoCl2 and
1 equiv of HCl (entry 13, Table 1). It is noteworthy that the
above-mentioned process is the first example of azomethine
cyclotrimerization successfully catalyzed by a transition-
metal compound.
Thus, the simplest and most convenient procedure for
the cyclization 1a f 2a is the treatment of tris-oxime 1a
with 3 equiv of acetic acid in aqueous methanol media at
ambient temperature (entry 7, Table 1). This procedure
was used for the cyclization of a series of “symmetrically-
substituted” tris-oximes 1b-g, bearing substituents with
different steric and electronic demands at the oximino
group (Scheme 2, Table 2).
Evidently, quaternization of the nitrogen atom in the bridge-
head position of adamantanes 2 by means of protonation (salts
3) or benzylation (salts 4) significantly stabilizes the adamantane
cage and, accordingly, allows the equilibrium to shift to
adamantanes. The procedure for the synthesis of salts 3 includes
treatment of the solution of initial tris-oxime 1 in aqueous
methanol with AcOH for the time indicated in Table 2 followed
by addition of 1 equiv of HCl (transformation 1 f 3). The
target salts 3a-d were isolated in high yields. However, under
these conditions oximes 1e-g gave only complex mixtures,
possibly, consisting of tris-oxime salts and the products of their
deoximation.
Under the conditions of the synthesis of N-benzyl salts 4
deoximation cannot occur, and therefore, BnBr is added just
after the acetic acid. Interestingly, adamantane 4e could be
obtained using this procedure too (entry 5, Table 2), but no
cyclization of tris-oximes 1f and 1g was observed.11
Using the method illustrated in Schemes 1 and 2 only
“symmetrically substituted” adamantanes 2 bearing three identi-
cal substituents R can be synthesized. The example of tris-oxime
1h demonstrates that the cyclization of tris-oximes 1 f 2 has
more general scope (Scheme 3). However, the lack of conve-
nient approaches to the synthesis of “unsymmetrically substi-
Scheme 2
.
Intramolecular Cyclization and Quaternization of
Tris-oximes 1a-g
Scheme 3
. Synthesis of “Unsymmetrically Substituted”
Adamantane 2h from Tris-oxime 1h
We believe that under conditions pointed in Scheme 2, the
equilibrium 1/2 is achieved. The target adamantanes 2 and
initial tris-oximes 1 are separated by liquid chromatography
followed by NMR control. The position of equilibrium strongly
depends on the substituent R. For tris-oximes 1a, 1b, and 1d
4074
Org. Lett., Vol. 11, No. 18, 2009

tuted” tris-oximes (similar to 1h) makes a systematic study of
their intramolecular cyclization difficult at the moment.12
The presence of heteroadamantane cage in compounds
2-4 was established by X-ray diffraction analysis of
adamantane 2a derivatives: tartrate 5a tetrahydrate and benzyl
salt 4a dihydrate (Figure 1).
Supporting Information). Thus, the geometry of the tet-
raazaadamantane cage is mainly governed by the stereoelec-
tronic interactions rather than steric reasons.
The brutto-formulas of stable individual products 2a,h,
3a-d, and 4a-e were confirmed by the elemental analysis
data.14 The NMR spectra of adamantanes 2a-d,h and their
derivatives 3a-d and 4a-e have some characteristic fea-
tures. The signals of atoms of adamantane cage and atoms
connected to them with one or two bonds are shifted to higher
field and appreciably widened in comparison to the corre-
sponding signals of tris-oximes 1. Such widening should be
due to exchange processes with a barrier of 10-15 kcal/
mol that may be the slow inversion of nitrogen atoms in
bridge positions and/or the restricted rotation around N-O
bonds.15 According to DFT calculations of four possible
conformers of adamantane 2a, the most stable are those found
in crystals of derivatives 4a·2H2O and 5a·4H2O. Conformers
with equatorial-axial-axial and equatorial-equatorial-axial OH
groups are energetically almost equal, while the difference
in energy between these conformers and those with all axial
(equatorial) OH groups is 1.1 kcal/mol (3.4 kcal/mol).
However, low solubility of adamantanes 2-4 in most of
organic solvents makes a detailed study of the stereodynamic
processes in these objects problematic.16
Figure 1. General view of azaadamatane cation in the crystals of
tartrate 5a tetrahydrate (a) and bromide 4a dihydrate (b) in the
representation of atoms via thermal ellipsoids (p ) 50%).
According to the X-ray data the adamantane moieties in
5a·4H2O and 4a·2H2O correspond to two conformers with
equatorial-axial-axial and equatorial-equatorial-axial disposi-
tions of hydroxyl groups. Examination of the geometry of
the tetraazaadamantane moiety in both crystalline salts
revealed the marked distinction between C-N bonds in the
triazine cycle (1.469(4)-1.499(4) Å in 5a·4H2O and
1.462(5)-1.501(5) Å in 4a·2H2O). It is the systematic
shortening of N2-C5 and N3-C4 bonds (1.469(4) and
1.479(4) Å) in 5a·4H2O and N3-C4 and N3-C6 ones
(1.462(5) and 1.466(4) Å) in 4a·2H2O. Moreover, some of
the C-C bonds also appear to be elongated. This is,
apparently, indicative of the competitive anomeric interac-
tions lpeq(N)fσ*(C-N) and lpax(N)fσ*(C-C),13 i.e., of the
charge transfer from the lone pairs of nitrogen atoms to the
σ* orbitals of above-mentioned C-N and/or C-C bonds.
The latter agrees with the results of the quantum-chemical
calculations for all of the four isomers of 2a (see the
Thus, a novel reaction of intramolecular cyclization of
tris(ꢀ-oximinoalkyl)amines, which leads to previously un-
known 4,6,10-trihydroxy-1,4,6,10-tetraazaadamantanes 2 pos-
sessing a heteroadamantane cage isomeric to urotropin, was
discovered and studied. This reaction was used as a key stage
in a three-step synthesis of adamantanes 2-4 from ammonia
and available aliphatic nitro compounds.
Acknowledgment. We are grateful for the support of the
Russian Foundation for Basic Research (Grant No. 09-03-
00676-a).
Supporting Information Available: Experimental pro-
cedures, characterization data of products, DFT calculations,
and X-ray crystallographic data (CIF). This material is
available free of charge via the Internet at http://pubs.acs.org.
OL9015157
(10) For generation of such complexes, see ref 1 and works cited therein.
(11) Formation of salts 3 and 4, of course, may proceed via quaterniza-
tion of sp3 nitrogen atom in initial tris-oximes 1.
(14) The brutto-formula of 2b was indirectly confirmed by elemental
analysis of its derivatives 3b and 4b.
(15) Raban, M.; Kost D. In Acyclic Organonitrogen Stereodynamics;
Lambert, J. B., Takeuchi Y., Eds.; VCH Publishers: New York, 1991; pp
57-88.
(12) Our strategy for the synthesis of “unsymmetrical” tris-oximes 1
from aliphatic nitro compounds will be published in the near future.
(13) Bushmarinov, I. S.; Antipin, M. Y.; Akhmetova, V. R.; Nadyrgu-
lova, G. R.; Lyssenko, K. A. J. Phys. Chem. A 2008, 112, 5017.
(16) The problems of stereodynamics in adamantanes 2 as well as the
equilibrium 1/2 will be discussed in a separate publication.
Org. Lett., Vol. 11, No. 18, 2009
4075