Synthesis and NMR characterization of cis and trans decahydro-2a,4a,6a,8a-tetraazacyclopent[fg]acenaphthylene. Solid state structure of the trans stereoisomer. Modelling studies

Article abstract of DOI:10.1016/j.tet.2007.04.036

The reduction with Vitride of the cis and trans isomers of octahydro-2a,4a,6a,8a-tetraazacyclopenten[fg]acenaphthylene 1,2-diones (5 and 6) and octahydro-2a,4a,6a,8a-tetraazacyclopenten[fg]acenaphthylene 3,4-diones (7 and 8) led to the cis 9 and the so far unknown trans 10 isomers of decahydro-2a,4a,6a,8a-tetraazacyclopent[fg]acenaphthylene. The stereochemistry of 9 and 10 was determined by comparison of the NMR coupling constants of the proton signals of the ethinic bridge. The results were confirmed by the solid state structure of the trans isomer (10) determined by single crystal X-ray diffraction. Given that the trans bis-aminal species 2 and 4 can undergo rearrangement, i.e. stereoisomerization into the corresponding cis ones (1 and 3, respectively), the conformational behaviour of species 1-4 was investigated by means of both molecular mechanics (MM) and quantum chemical (QC) approaches. The theoretical study was completed by including diamides 5-8 and the final products 9 and 10.

Full text of DOI:10.1016/j.tet.2007.04.036

Tetrahedron 63 (2007) 6915–6923
Synthesis and NMR characterization of cis and trans decahydro-
2a,4a,6a,8a-tetraazacyclopent[fg]acenaphthylene. Solid state
structure of the trans stereoisomer. Modelling studies
a
a
a
Maria Argese,a, Marino Brocchetta, Mario De Miranda, Andrea Ferraris,
*
Paolo Dapporto,b Paola Paolib and Patrizia Rossib
aMilano Research Centre, Bracco Imaging spa, via E. Folli 50, 20134 Milano, Italy
bDipartimento di Energetica ‘Sergio Stecco’, Universitaꢀ di Firenze, via S. Marta 3, 50139 Firenze, Italy
Received 15 December 2006; revised 26 March 2007; accepted 12 April 2007
Available online 19 April 2007
Abstract—The reduction with VitrideÒ of the cis and trans isomers of octahydro-2a,4a,6a,8a-tetraazacyclopenten[fg]acenaphthylene 1,2-
diones (5 and 6) and octahydro-2a,4a,6a,8a-tetraazacyclopenten[fg]acenaphthylene 3,4-diones (7 and 8) led to the cis 9 and the so far
unknown trans 10 isomers of decahydro-2a,4a,6a,8a-tetraazacyclopent[fg]acenaphthylene. The stereochemistry of 9 and 10 was determined
by comparison of the NMR coupling constants of the proton signals of the ethinic bridge. The results were confirmed by the solid state struc-
ture of the trans isomer (10) determined by single crystal X-ray diffraction. Given that the trans bis-aminal species 2 and 4 can undergo re-
arrangement, i.e. stereoisomerization into the corresponding cis ones (1 and 3, respectively), the conformational behaviour of species 1–4 was
investigated by means of both molecular mechanics (MM) and quantum chemical (QC) approaches. The theoretical study was completed by
including diamides 5–8 and the final products 9 and 10.
Ó 2007 Elsevier Ltd. All rights reserved.
1. Introduction
HN
NH
CHO
CHO
+
In a recent article,1 as well as in a previous patent,2 we
reported that, by amidation of a mixture containing the bis-
aminals 1–4 with diethyl oxalate (DEO) in EtOH at reflux,
the related mixture containing cis octahydro-2a,4a,6a,8a-
tetraazacyclopent[fg]acenaphthylene 1,2-dione (5) and cis
octahydro-2a,4a,6a,8a-tetraazacyclopent[fg]acenaphthylene
3,4-dione (7) is obtained (Scheme 1). The result is a conse-
quence of a rearrangement, which transforms 2 and 4 into 1
and 3, respectively, before the amidation.1,3 Accordingly, the
trans isomers 6 (trans octahydro-2a,4a,6a,8a-tetraazacyclo-
pent[fg]acenaphthylene 1,2-dione) and 8 (trans octahydro-
2a,4a,6a,8a-tetraazacyclopent[fg]acenaphthylene 3,4-dione)
are not formed. In the presence of 2-pyridinol also, the trans
compounds are absent, but when the reaction is catalyzed
with MeONa, the amidation becomes faster than the stereo-
isomerization, and along with 5 and 7, compounds 6 and 8
are also obtained.1 The latter, after reduction of the amide
carbonyls, could obviously disclose route to 10, a compound
not yet reported in the literature.4 Such a synthetic strategy
has never been used in the past, as compounds 6 and 8
were unknown.
NH2 H2N
H
H
H
H
N
N
N
N
N
N
NH
NH
N
N
NH
NH
COOEt
COOEt
+
+
+
+
N
H
N
H
N
H
N
H
H
H
H
H
1
2
3
4
H
H
H
H
H
N
N
N
N
N
N
O
N
N
N
N
N
N
O
O
Red.
+
+
N
N
O
N
N
H
H
H
O
O
O
O
8
7
5
6
(R,R + S,S)
H
H
N
N
N
N
+
N
N
N
N
H
H
9
10
Scheme 1.
2. Results and discussion
The reduction with VitrideÒ,7 of the amide carbonyls of
* Corresponding author. E-mail: maria.argese@poste.it
a mixture containing the cis isomers 5 and 71 led to a product,
0040–4020/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2007.04.036

6916
M. Argese et al. / Tetrahedron 63 (2007) 6915–6923
which was recovered in good yield by a quite simple proce-
dure. We were confident that this compound corresponded
to cis decahydro-2a,4a,6a,8a-tetraazacyclopent[fg]acenaph-
thylene (9). Analogously, we could use either one of the
previously prepared trans isomers 6 or 81 for the synthesis
of the trans tetracycle, but due to the low quantity of these
product at disposal, we decided to reduce a mixture contain-
ing 5–8, which could be easily prepared.1 The subsequent
chromatographic separation of the reduced mixture af-
forded, apart from 9, a new compound having the same
mass of 9 but different mp, TLC Rf and GC retention time.
To this product we tentatively assigned the structure of
trans decahydro-2a,4a,6a,8a-tetraazacyclopent[fg]acenaph-
thylene (10).
3.0
4.5
4.0
3.5
2.5
ppm
Taking advantage of the NMR experience acquired with
compounds 5–8, whose stereochemistry was known by sin-
gle crystal X-ray diffraction,1,3 we could attribute the correct
stereochemistry to 9 and 10, even before the solid state struc-
ture determination of 10 by X-ray diffraction. Indeed, the
discrimination between cis and trans isomers was achieved
by the analysis of the coupling constants (J) of the proton
signals of the ethinic bridge. The latter showed a doublet
multiplicity for compounds 5 and 6 and the J values could
be easily deduced, while for the symmetric compounds 7
and 8 the CH signals were singlets. Accordingly, the direct
calculation of J was prevented, but as it is known that the
analysis can also be performed on the signals of rare isotope
satellites,8 we turned our attention to the 13C satellite analy-
J=9.6Hz
4.7
4.6
4.5
4.4
ppm
Figure 2. J value for the ethinic protons of 8.
1
sis of the H NMR signals. The presence of a 13C atom in
although not the only one, is the most important factor af-
fecting the J values. Reliable results in structural analysis
are to be expected from the comparison of closely related
species, as in our case in which a large coupling constant
is associated with the trans structure with a dihedral angle
close to 180ꢀ, while a small one with a cis structure with a di-
hedral angle in the 40ꢀ range.
only one of the halves produces a loss of molecular symme-
try, which enables the determination of the coupling constant
between otherwise equivalent protons (Figs. 1 and 2). The J
values of the cis compounds 5 and 7 resulted lower than
those of the related trans isomers 6 and 8 (Table 1). This is
in agreement with the theory,9 which correlates the vicinal
proton coupling constant to the dihedral angle. The latter,
The same study allowed us to obtain the coupling constant
of 9 (Fig. 3), while for compound 10 the 13C satellites of
the CH signals unfortunately overlapped the CH2 signals.
1
Accordingly, a H–13C reverse HSQC 2D experiment was
performed using the INVIEAGSSI pulse program from
Bruker library and removing the GARP decoupling during
acquisition so that the cross-peaks between carbon and
proton resulted in correspondence with the 13C satellites in
the proton spectrum (f2 dimension). The experiment gave
a phase sensitive spectrum, which was properly phased using
the 2D Bruker phase routine and a resolution sufficient to
show that the multiplicity pattern was obtained using the fol-
lowing parameters: fid resolution¼78.6 Hz (f1)/0.1 Hz (f2);
matrix size¼128 (f1)ꢁ8192 (f2). Extracting the row of the
desired cross-peak from the 2D spectrum, it was possible
to obtain the J value for 10, which resulted in the range of
the trans isomers (Fig. 4) (Table 1). In order to verify the ap-
propriateness of the method, the same experiment with the
same parameters was run for compound 9. The J value
4.5
4.0
3.5
3.0
2.5
ppm
J=4.5Hz
Table 1. J values (Hz) found for the proton signals of the ethinic bridge of
compounds 5–10
5
6
7
8
9
10
4.4
4.3
4.2
4.1
ppm
4.6
6.2
4.5
9.6
3.0
9.2
Figure 1. J value for the ethinic protons of 7.

M. Argese et al. / Tetrahedron 63 (2007) 6915–6923
6917
3.4
3.2
3.0
2.8
2.6
2.4
ppm
J=3.0Hz
Figure 5. J value for the ethinic protons of 9 (from the 2D experiment).
Table 2). Table 3 lists experimental (X-ray data) and calcu-
lated (quantum chemical) bond distances.
3.25
3.20
3.15
3.10
3.05
3.00
ppm
Figure 3. J value for the ethinic protons of 9.
Both the six- and five-membered rings show their usual con-
formation, that is, chair and envelope, respectively. The po-
sition, i.e. axial (A) or equatorial (E) of the bound –CH2
groups belonging to the five-membered ring results in the
EAAE isomer (see Section 4.4.1).
obtained from the 2D method (3.1 Hz) (Fig. 5) resulted very
close to the one (3.0 Hz) obtained with the direct measure-
ment in the 1D spectrum and was related to a cis compound.
The further and definitive proof of the stereochemistry of 10
was achieved when we finally obtained crystals suitable for
the solid state structure determination by X-ray diffraction.
The analysis of the experimental bond distances (Table 3)
˚
˚
shorter than the CH2–CH2 ones (d2, mean value 1.54 A).
reveals that the CH–CH bond (d1¼1.496(2) A) is by far
Figure 6 shows an ORTEP3 view of compound 10, which is
in the asymmetric unit and interacts via a network of weak
hydrogen bonds10 with its symmetry related images (see
C(7)
C(8)
ppm
30
N(1)
N(4)
C(6)
C(9)
C(1)
40
50
C(5)
J=9.2Hz
60
C(10)
C(2)
70
N(3)
N(2)
80
C(3)
C(4)
90
100
ppm
4.0
3.5
3.0
2.5
2.0
Figure 6. ORTEP3 view of compound 10. (Thermal ellipsoids are set to
30% probability).
Figure 4. J value for the ethinic protons of 10.

6918
M. Argese et al. / Tetrahedron 63 (2007) 6915–6923
Table 2. Intermolecular contacts found in the crystal lattice of 10
Donor–H/acceptor Donor/acceptor H/acceptor Donor–H/
orientation to conjugate with the s* C(1)–C(2), C(5)–C(6)
and C(9)–C(10) bonds. The estimation of the relative ener-
getic importance of these hyper-conjugative effects suggests
that while the axial lone pairs on N(1)/N(4) interact mainly
with the s*C(9)–H [vs the C(1)–H, C(6)–H], the equatorial
lone pairs on N(2)/N(3) conjugate preferentially with the s*
C(1)–C(2) and C(5)–C(6) [vs C(9)–C(10)]. It is noteworthy
that in the optimized trans tetraazadecalin isomers (EEEE,
AAAA and EAAE) the different n(N)/s*(C–Y) inter-
actions are energetically equivalent and, accordingly, no
significant differences affect the C–N and C–C bond
˚
distance (A)
˚
distance (A) acceptor
angle (ꢀ)
C(1)–H(1a)/N(3)a
3.817(2)
2.83(2)
2.98(2)
2.89(2)
2.69(2)
2.97(2)
2.89(2)
2.82(2)
2.53(2)
168(1)
162(1)
126(1)
135(1)
131(1)
133(1)
163(1)
162(1)
C(2)–H(2b)/N(4)b 3.957(2)
C(7)–H(7a)/N(2)c
C(8)–H(8a)/N(2)c
C(4)–H(4a)/N(3)d
C(3)–H(3a)/N(3)d
C(4)–H(4a)/N(2)d
C(5)–H(5a)/N(1)e
3.583(2)
3.490(3)
3.711(3)
3.651(2)
3.792(3)
3.455(2)
˚
distances (mean values 1.45 and 1.53 A, respectively). The
a
¼x+1/2, ꢂy+1/2, z+1/2.
¼x+1, y, z.
comparison of the results of NBO analyses on optimized
2EAAE and 2AEEA isomers (see Scheme 2) suggests that
it is the CH2–CH2 bridge in equatorial position, which
causes the asymmetry in the conjugation effects and, as
a consequence, in the geometry. In fact in the 2AEEA spe-
cies the hyper-conjugative interactions, as well as the C–N
and C–C bond distances are all comparable. On the contrary
in the 2EAAE isomer there are energetic and geometrical
differences, a trend is consistent with that observed in 10.
A further support to the importance of the equatorial position
of the CH2–CH2 bridge in determining an asymmetry in
the tricyclic molecule was obtained from an ab initio optimi-
zation of two different isomers of dodecahydroacenaph-
thylene: the AA and EE ones (the acronyms refer to the
position of the –CH2 groupings of the five-membered ring
with respect to the tertiary carbon atoms). In the AA isomer
the CH–CH and CH2–CH2 are almost identical, while in the
molecule bearing an ‘equatorial’ five-membered ring we ob-
served the same trend found in 10.
b
c
d
e
¼xꢂ1/2, ꢂy+1/2, z+1/2.
¼ꢂx+1, ꢂy+1, ꢂz.
¼ꢂx+1/2, y+1/2, ꢂz+1/2.
The opposite, i.e. d1>d2, holds in fragments featuring
a 1,4,5,8-tetraazadecalin (TAD) core11 retrieved in the Cam-
bridge Structural Database.12 The CH–N bonds are shorter
than the CH2–N ones, a trend particularly significant for
the bonds about N(1) and N(4), which, as a whole, appear
shorter than those about the corresponding N(2) and N(3)
atoms. The equivalent couples N(1)/N(4) and N(2)/N(3)
differ from the positions of their lone pair and, correspond-
ingly, of the –CH2 grouping of the five-membered ring:
axial–equatorial (A–E), respectively, on the first couple;
equatorial–axial (E–A), respectively, on the latter.
Maybe stereoelectronic effects, such as the anomeric one,13
could explain such structural features, which are maintained
also in the HF and DFT optimized molecule (see Table 3).
Given that the trans bis-aminals species 2 and 4 (Scheme 1)
can undergo rearrangement, i.e. stereoisomerization into the
corresponding cis ones (1 and 3, respectively) and that, to the
best of our knowledge, only one tricyclic species has been
characterized in the solid state14 (it shows a cis-junction be-
tween the six-membered rings), we decided to investigate
the energetic and the conformational behaviour of species
1–4 by means of both molecular mechanics (MM) and quan-
tum mechanics (QM) approaches. The theoretical study was
completed also by including diamides 5–8 and the final
products 9 and 10.
In order to find a rationale for these structural differences
featuring the nitrogen atoms N(1)/N(4) and N(2)/N(3), a nat-
ural bond analysis (NBO) was performed on the HF opti-
mized molecule. Actually, the lone pairs on the nitrogen
atoms are interested in hyper-conjugative n(N)/s*(C–Y)
interactions mainly within the six-membered rings. In fact,
the lone pairs on N(1) and N(4) are aligned in an antiperipla-
nar geometry and interact with the axial C–H bonds of C(1),
C(6) and C(9), while those on N(2) and N(3) have the correct
˚
Table 3. Selected bond distances (A) for 10 as obtained from X-ray diffrac-
tion and modelling studies
A bibliographic search revealed that the conformational be-
haviour of TAD, together with some derivatives, has been in-
vestigated by means of NMR and MM.15 Its intramolecular
dynamics can be rationalized according to the following
types of motion: trans–cis isomerization (by means of ami-
nal ring-chain tautomerism), ring inversion (in cis isomers
only) and nitrogen inversion. The X-ray structure of the
trans-fused species has also been reported.16 Obviously,
molecules 1–10 can undergo less significant conformational
changes with respect to TAD, given the constraints imposed
by the five-membered ring/ring closure. However, steric in-
teractions of peri-substituents, lone pair–lone pair, np–s*
interactions (as already pointed out) and NH/N hydrogen
bonds are still active, although not in all the molecules,
and should influence their dynamics as well as their solid
state 3D arrangement.
X-ray
diffraction
HF/6-31G**
(full opt)
B3LYP/6-31G**
(full opt)
N(1)–C(1)
N(1)–C(8)
N(1)–C(9)
N(2)–C(2)
N(2)–C(3)
N(2)–C(10)
N(3)–C(4)
N(3)–C(5)
N(3)–C(10)
N(4)–C(6)
N(4)–C(7)
N(4)–C(9)
C(1)–C(2)
C(3)–C(4)
C(5)–C(6)
C(7)–C(8)
C(9)–C(10)
1.463(2)
1.476(3)
1.438(2)
1.475(2)
1.480(2)
1.459(2)
1.485(2)
1.473(3)
1.458(2)
1.465(2)
1.478(2)
1.438(2)
1.544(3)
1.541(3)
1.541(3)
1.552(3)
1.496(2)
1.450
1.460
1.425
1.460
1.461
1.441
1.461
1.460
1.441
1.450
1.460
1.425
1.549
1.550
1.549
1.560
1.501
1.461
1.474
1.439
1.471
1.476
1.456
1.476
1.471
1.456
1.461
1.474
1.439
1.559
1.557
1.559
1.568
1.513
Molecular dynamics in vacuum showed that the conforma-
tional space accessible for 1–10 is quite limited as provided

M. Argese et al. / Tetrahedron 63 (2007) 6915–6923
6919
(E)
(E)
(E)
(E)
(E)
(E)
(E)
H
H
H
N1
N4
N1
N4
N1
N4
1
N2 N3
H
N2 N3
H
N2 N3
H
(E)
(E)
(E)
(E)
(E)
(E)
(A)
(A)
(A)
H
H
H
H
H
H
H
N1 N4
1 EEEE
1 EAAE
1 EAEE
N2 N3
H
(E)
(E)
(E)
(E)
(A)
(E)
(E)
N4H
H
H
H
N1
N2
N4H
N1
N2
N4H
N1
N2
9
3
N3H
N3H
N3H
H
H
H
(E)
(E)
(A)
(E)
(A)
3 EEEE
3 EEAA
3 EEAE
(E)
(E) (E)
(E) (E)
(E) (A)
(A) (A)
(A) (A)
(A)
H
H
H
H
H
H
N1 N4
N1 N4
N1 N4
N1 N4
N1 N4
N1 N4
2
N2 N3
H
N2 N3
N2 N3
H
N2 N3
N2 N3
N2 N3
H
H
H
H
(E)
(E) (A)
(A) (E)
(A) (A)
(A) (A)
(E) (E)
(E)
H
H
H
H
H
H
H
H
H
H
H
H
(E)
N1 N4
2 EEEE
(E)
(A)
2 EAAE
2 AAAA
2 AAEA
2 AEEA
2 EEAE
H
N2 N3
H
(A)
(E)
(E)
N4H
(E)
(A)
H
(E)
(E)
N4H
(E)
(A)
N4H
10
H
H
H
H
H
N1
N1
N2
N4H
N1
N2
N1
N2
4
N2
(A)
N3H
(E)
N3H
N3H
(A)
N3H
(E)
H
H
(A)
(A)
(A)
(A)
4 EAEA
4 EAEE
4 EAAE
4 EAAA
Scheme 2. Different isomers of the tricyclic intermediates 1–4 were generated according to the procedure in Section 4.4.1 (i). Letters (A/E) refer to the
conformations on N atoms.
by the rather small number of conformers spotted during the
simulations. As expected on the basis of their stereochemis-
try (sp2 hybridized nitrogen atoms) diamides 5–8 appear as
the most constrained: molecule 6 results the most rigid
(only two different conformers were classified, rms devia-
Monte Carlo conformational searches, including the effect
of the solvent (water) carried out for all compounds, showed
a similar behaviour, with the cis-fused isomers more flexible
than the trans-fused ones. Again, considering the filtering
conditions reported in the Section 4, molecule 6 appeared
the most rigid (1 conformer selected) and molecule 1 the
most flexible (30 conformers). From an energetic point of
view, all the cis-fused molecules were more stable than the
corresponding trans ones. In all cases, HF geometry optimi-
zations did not produce any relevant change in the 3D ar-
rangement of the molecules with respect to their starting
geometry. The most stable isomer (Table 4) of each tricyclic
intermediate (1–4) has the two hydrogen atoms bound to the
secondary nitrogens in an A/E disposition, thus optimizing
˚
tion 0.45 A). On the contrary molecule 1 shows the largest
variability: 11 different conformers were found (max rms
deviation 0.67 A) differing up to 12 kcal molꢂ1. As a general
˚
trend, cis-fused isomers appear more flexible than the corre-
sponding trans one. In most cases, the different conformers
recognized can be ascribed to nitrogen inversion and, more
rarely, to ring inversion processes. With some exception (3
and 4), cis-fused molecules appear more stable than the
corresponding trans ones.
Table 4. Energy differences (cis/trans couples relative energies, kcal molꢂ1) for the intermediates and the final compounds as obtained from quantum chemical
calculations
a b
,
Compound
Relative energy
HF/6-31G** (full opt)
Relative energy
MP2/6-31G**//HF/6-31G**
Relative energy
B3LYP/6-31G** (full opt)
Relative energy
MP2/6-31G**//B3LYP/6-31G**
1
2
3
4
5
6
7
8
9
10
0.00
1.62
0.00
1.14
0.00
0.41
0.00
1.17
0.00
6.53
0.00
2.98
0.00
3.24
0.34
0.00
0.00
2.65
0.00
7.00
0.00
1.40
0.00
0.61
1.15
0.00
0.00
1.22
0.00
5.71
0.00
1.52
1.43
0.00
0.00
0.46
0.00
1.12
0.00
6.51
a
For the compounds 1–4 and 9 the most stable conformers obtained from modelling (details in Section 4.4.1) were: EAEE (1); EEAE (2); EEAE (3); EAEA
(HF optimization) and EEEE (DFT optimization) (4) and EEEE (9).
For compounds 5–81,3 and 10 (Fig. 5) the data are referred to the conformations deduced by X-ray diffraction analysis.
b

6920
M. Argese et al. / Tetrahedron 63 (2007) 6915–6923
their relative distance as well as that separating the nitrogen
lone pairs, hence the non-bonding interactions.
(1 mL) of a 20 mg/mL solution in toluene containing ace-
naphthene (10 mg/mL) as an internal standard. IR spectra
were recorded on a Perkin–Elmer 882 spectrophotometer,
using potassium bromide disks. MS spectra were acquired
At DFT level all the minima were equivalent to the HF ones,
with one exception: the most stable conformer for the com-
pound 4 differed for N2 and N4 inversions, corresponding to
the EEEE isomer. However, the E/E disposition of the hydro-
gen atoms bound to secondary nitrogens guaranteed their
relative maximum distance, as well as one of the respective
lone pairs.
on
a TSQ700 ThermoFinnigan Spectrometer using
CH3OH as the solvent. H and 13C NMR spectra were re-
corded at 298 K in D2O at 400.13 and 100.61 MHz, respec-
tively, with a Bruker DRX 400 spectrometer. In order to have
a complete assignment of the structures, 2D spectra were re-
corded using 1H–1H COSY45, HMQC and HMBC standard
pulse sequences. The chemical shifts are given in d units
(ppm) relative to TMS (d¼0). Elemental analyses were car-
ried out at the Redox Laboratories (Monza, Milano, Italy).
1
In general, QC results (Table 4) confirm the MM trend, i.e.
cis-fused products appear lower in energy than the corre-
sponding trans ones as already found for TAD.15 Since this
tendency is less significant for the 3 and 4 couples (no
TAD core) and it is definitely dubious for the 5 and 6 ones
(where the TAD core is somewhat perturbed by the presence
of sp2 atoms in the five-membered cycle),1 perhaps the cis-
versus trans-junction preference is an intrinsic feature of
the TAD core.
4.2. Synthetic methods
4.2.1. cis Decahydro-2a,4a,6a,8a-tetraazacyclopent[fg]-
acenaphthylene (9). A mixture containing 5 and 7 in about
2/1 ratio1 (10.0 g; 0.045 mol) was added to a 70% toluene
solution of VitrideÒ,7 (50.4 g; 0.174 mol) diluted with tolu-
ene (100 mL) and stirred at 40 ꢀC under a nitrogen atmo-
sphere. In about 10 min the suspension was heated to
112 ꢀC and turned into a clear solution. After 1 h in the
same conditions, when the conversion (GC) reached 97%,
the mixture was cooled to rt and 5% aq NaOH (29 mL)
was cautiously added dropwise. The two phases were sepa-
rated and the viscous aqueous phase was extracted with tol-
uene (2ꢁ25+50 mL) then the combined organic phases were
evaporated to a residue. The latter was dissolved in water
(40 mL) and the solution was loaded onto a strong cation ex-
change resin (Amberlite AmberjetÒ, 80 mL) column, which
was eluted with water to neutrality. The subsequent percola-
tion of 2.5% NH4OH (450 mL) afforded a basic eluate,
which was evaporated to dryness. The residue was extracted
with hexane (90 mL) and the organic phase was evaporated
to afford 9 (7.1 g; 84%). Mp 90–92 ꢀC.17 TLC Rf 0.8; GC
3. Conclusion
A new synthetic strategy, starting from diones 5–8, allowed
us to synthesize compound 9 and the so far unknown com-
pound 10, whose stereochemistry was assigned by solid state
X-ray diffraction.
The analysis of the NMR coupling constants for the hydro-
gens of the ethinic bridge afforded an independent tool for
the attribution of the stereochemistry to compounds 5–10.
In the solid state structure, 10 shows an EAAE conformation
on the nitrogen atoms. The shortening of the CH–CH bond
distance, with respect to the CH2–CH2 was ascribed to the
equatorial position of the CH2–CH2 bridge of the five-mem-
bered ring (NBO analysis on the ab initio optimized mole-
cule). The conformational space accessible to molecules
1–10 is quite limited both in vacuum and in solvent (water),
diamides 5–8 being the most constrained. As a general trend,
cis-junctions are associated with a larger conformational
freedom and a greater energy stability with respect to the
trans ones. The TAD core on its own could account for the
latter energetic trend.
1
100% (rt 8.5 min); IR (KBr) n 2938, 1465 cmꢂ1; H NMR
(400 MHz, D2O) d 2.46 (m, 4H: CH2), 2.61 (br m, 4H:
CH2), 2.80 (m, 4H: CH2), 2.89 (m, 4H: CH2), 2.99 (s, 2H:
CH); 13C NMR (100 MHz, D2O) d 49.7 (CH2), 50.6
(CH2), 76.6 (CH); MS m/z (ESI) 195 (M+H)+, 217
(M+Na)+; Anal. Calcd for C10H18N4: C, 61.82; H, 9.34; N,
28.84. Found: C, 62.02; H, 9.38; N, 29.07.
4.2.2. trans Decahydro-2a,4a,6a,8a-tetraazacyclopent-
[fg]acenaphthylene (10). A mixture containing 5 (61.5%),
6 (16.4%), 7 (3.4%) and 8 (18.7%)1 (8.0 g; 0.036 mol) was
reduced under the conditions used to obtain 9. The final
residue (5.4 g containing 9 and 10 in 2.5/1 ratio,19 as deter-
mined by GC) was subjected to silica gel chromatography
(CHCl3/CH3OH 9/1) to afford 10 (0.7 g; 28%20). Mp 136–
139 ꢀC;17 TLC Rf 0.54; GC 100% (rt 10.4 min); IR (KBr)
4. Experimental
4.1. General
All reagents and solvents, obtained from commercial sour-
ces, were used without further purification. Melting points
1
n 2940, 1484 cmꢂ1; H NMR (400 MHz, D2O) d 2.73 (m,
ꢀ
€
( C, uncorrected) were measured with a Buchi 510 instru-
4H: CH2), 2.81 (m, 8H: CH2), 2.87 (s, 2H: CH), 3.10 (m,
4H: CH2); 13C NMR (100 MHz, D2O) d 47.6 (CH2), 48.2
(CH2), 75.6 (CH); MS m/z (ESI) 195 (M+H)+, 217
(M+Na)+; Anal. Calcd for C10H18N4: C, 61.82; H, 9.34; N,
28.84. Found: C, 61.65; H, 9.46; N, 28.65.
ment. TLC was carried out on 60 F254 silica gel plates using
CHCl3/CH3OH/25% NH4OH, 6/3/1 as the eluent and iodine
vapours for the visualization. GC analyses were performed
with a Hewlett–Packard HP 5890 instrument using: CP Sil
19 CB column (25 mꢁ0.32 mm; film thickness 0.2 mm);
He pressure 20 psi; injector and detector (FID) temperatures
250 and 275 ꢀC, respectively; oven temperature timetable as
follows: first isotherm at 120 ꢀC for 5 min; 15 ꢀC/min ramp
to 260 ꢀC; second isotherm at 260 ꢀC for 12 min; injection
4.3. X-ray crystallographic study
Crystals of 10 used for X-ray diffraction analysis were
grown at rt from a saturated solution in hexane. Crystals of

M. Argese et al. / Tetrahedron 63 (2007) 6915–6923
Table 5. Crystal data and structure refinement for 10
6921
secondary nitrogen atoms refers to the axial (A) or
equatorial (E) position of the bound hydrogen atom.
For the tertiary nitrogen atoms the A or E label arises
from the position of the bound –CH2 group belonging
to the five-membered ring as found in the solid state
structure. In the isomers nomenclature the first letter
refers to the conformation on the N(1) atom and then
it proceeds counterclockwise along the fused rings.
(ii) Starting models (1–4 and 9) were obtained from Monte
Carlo searches [details in Section 4.4.2 (ii)].
Empirical formula
Formula weight
T (K)
C10H18N4
194.28
120
0.71073
monoclinic, P21/n
a¼7.6341(9); b¼11.457(1),
b¼96.873(9); c¼11.172(1)
970.1(2)
4, 1.330
0.084
˚
l (A)
Crystal system, space groꢀup
˚
Unit cell dimensions (A,
)
3
˚
Volume (A )
Z, dcalcd (g cmꢂ3
)
m(mmꢂ1
F(000)
)
424
0.4ꢁ0.5ꢁ0.55
Crystal size (mm)
2q Range for data collection (ꢀ)
Reflections collected/unique
Refinement method
8.92–51.94
4.4.2. Molecular mechanics calculations. Two parallel
strategies were used to explore the conformational space
accessible to molecules 1–10 in vacuum and in solvent.
7510/1639 [R(int)¼0.0446]
Full-matrix least-squares on F2
1639/0/199
Data/restraints/parameters
Final R indices [I>2s(I)]
R indices (all data)
R1¼0.0395, wR2¼0.0714
(i) Molecular dynamics (MD) simulations were performed
in vacuum at 1000 K (CVFF force field) using the MSI
software programs Insight II and Discover28 (version
98.0). The Verlet leapfrog algorithm, with a time step
of 1 fs, was used for integration of equations of motion
in all simulations. Before starting each MD simulation,
the starting models, derived as outlined in Sec-
tion 4.4.1 (5–8, 10) and 4.4.1 (i) (1–4, 9), were
R1¼0.0750, wR2¼0.0788
9, either as free base from hexane, or as the hydrochloride
salt from methanol or acetonitrile, were obtained but unfor-
tunately, in any cases, they resulted as twinned crystals,
whose structure was not solvable by single crystal X-ray
diffraction. Intensity data collection for compound 10 was
performed by using an Oxford Diffraction XCalibur diffrac-
tometer equipped with a CCD area detector using the radia-
optimized until the convergence criterion was met
˚
(<0.01 kcal molꢂ1 ꢂ1). For every MD run the system
A
˚
tion Mo Ka (l¼0.7107 A). Diffraction data were collected
was allowed to equilibrate for 10 ps and then a 1000 ps
dynamic was performed. Snapshot conformations were
collected for every pico second, minimized and ranked
accordingly to their energy contents. Then, for each
molecule, conformers, differing in energy for
1 kcal molꢂ1, were selected and clustered in four inter-
vals on the basis of their structural similarity.
with the CrysAlis CCD program21 and reduced with the
CrysAlis RED program.22 Absorption correction was per-
formed with the program ABSPACK in CrysAlis RED.
Structure was then solved using the SIR97 program23 and
refined by full-matrix least squares against F2 using all
data (SHELX 97).24 All the non-hydrogen atoms were re-
fined anisotropically. All the hydrogen atoms were found
in the Fourier difference map. Geometrical calculations
were performed by PARST9725 and molecular plots were
produced by the program ORTEP3.26
(ii) The Monte Carlo method,29 as implemented in
Macromodel software program (version 9.0) in Maestro
(version 7.0)30 was used to perform conformational
searches for all the molecules. The Amber * force
field,31 and, as an implicit solvation treatment, the GB/
SA solution model32 were employed. For each structure,
a global search was performed, with a maximum num-
ber of 5000 Monte Carlo steps: the conformers were
minimized by the Truncated Newton Conjugate Gradi-
Crystallographic data and refinement parameters are re-
ported in Table 5.
Crystallographic data for compound 10 have been deposited
with the Cambridge Structural Data Centre as supplemen-
tary publication number CCDC 628339.
ent method,33 with a derivative convergence criterion
˚
at a value of 0.01 kcal molꢂ1 ꢂ1; they were then fil-
A
tered using an energy cut-off of 12 kcal molꢂ1 and
˚
a threshold of 0.25 A as the maximum distance between
corresponding heavy atoms after superposition.
4.4. Molecular modelling studies
4.4.1. General strategy. The starting geometries of mole-
cules 5–81,3 and 10 were those of their solid state structures.
For compounds 1–4 and 9, whose X-ray structural parame-
ters were not available, starting geometries were generated
using different approaches, (i) and (ii), for comparative
purposes.
4.4.3. Quantum chemical calculations. Parallel QC calcu-
lations, (i) and (ii), were carried out on molecules 1–10.
(i) The geometry of the starting models obtained as
described in Section 4.4.1 (5–8 and 10) and 4.4.1 (i)
(1–4 and 9) was optimized using the GAUSSIAN 03
(Revision B.05)34 package implemented on a personal
computer.
(i) For 1, 3 and 9, the atomic coordinates of the cis decahy-
dro-2a,4a,6a,8a-tetraazacyclopent[fg]acenaphthylene-
2a,6a-bis(borane)27 (YAHTOQ refcode) backbone were
used, while the input geometries of 2 and 4 were derived
from that of the parent molecule 10 (Scheme 2). In prin-
ciple, for the tricyclic molecules 1–4 several conforma-
tional isomers should be taken into account, depending
on the axial or equatorial position (A or E, respectively)
of the hydrogen atoms bound to the secondary nitrogen
atoms (see Scheme 2). The A or E letter labelling the
In all cases the basis set was 6-31G(d,p).35 The
reliability of the stationary points, found by means of
HF-SCF calculations and the Berny algorithm,36 was
assessed by the evaluation of the vibrational frequen-
cies. Then single point MP237 calculations [6-
31G(d,p), basis set] were performed on the optimized
structures. NBO analyses were performed using the
NBO 5.0 program.38

6922
M. Argese et al. / Tetrahedron 63 (2007) 6915–6923
15. M€uller, R.; von Philipsborn, W.; Schleifer, L.; Aped, P.; Fuchs,
B. Tetrahedron 1991, 47, 1013–1036.
16. Sanger, I.; Lerner, H.-W.; Bolte, M. Acta Crystallogr., Sect. E
2004, 60, 1847–1848.
17. In a previous article,18 the mp 91–92 ꢀC was reported for a
decahydro-2a,4a,6a,8a-tetraazacyclopent[fg]acenaphthylene
to which the trans stereochemistry was tentatively assigned as
the compound was synthesized from trans 1,4,5,8-tetraazade-
calin under conditions, which should not have induced epime-
rization. Our data disagree, as they attribute such mp to the cis
isomer and mp 136–139 ꢀC to the trans isomer.
(ii) To take into account the effect of electron correlation
on the molecular geometry, further calculations at DFT
level were carried out on the starting models obtained
from Monte Carlo method [Section 4.4.2 (ii)]. The
Spartan’0239 software on a SGI Indigo 2 IMPACT
10000 workstation was used. Gas-phase geometry
optimizations were performed by the Becke’s Three
Parameter Hybrid functional40 using the LYP Correla-
tional functional41 [B3LYP method, 6-31G(d,p) basis
set]; again, a single point energy calculation at MP2
level [6-31G(d,p), basis set] was performed on each
resulting structure.
18. Ferrari, M.; Giovenzana, G. B.; Palmisano, G.; Sisti, M. Synth.
Commun. 2000, 30, 15–21.
19. The ratio between the cis and trans isomers, which was 1.8/1 in
the starting material, became 2.5/1 after the reduction because
in the mixtures containing 5–8 isomerization phenomena in-
duced by the alkaline pH occur, which favour the increase of
the cis compounds.1
References and notes
1. Argese, M.; Brocchetta, M.; De Miranda, M.; Paoli, P.; Perego,
F.; Rossi, P. Tetrahedron 2006, 62, 6855–6861.
2. Argese, M.; Brocchetta, M.; Manfredi, G.; Rebasti, F.; Ripa, G.
WO 01/79207, 2001.
20. Calculated on the mol (0.0126) of 6 and 8, which can theoret-
ically afford 10.
21. CrysAlis CCD, Oxford Diffraction: Version 1.171.pre23-10
beta (release 21.06.2004 CrysAlis171.NET) (compiled June
21 2004,12:00:08), Abingdon, Oxfordshire, England.
22. CrysAlis RED, Oxford Diffraction: Version 1.171.pre23-10
beta (release 21.06.2004 CrysAlis171.NET) (compiled June
21 2004,12:00:08), Abingdon, Oxfordshire, England.
23. Altomare, A.; Cascarano, G. L.; Giacovazzo, C.; Guagliardi,
A.; Burla, M. C.; Polidori, G.; Camalli, M. J. Appl.
Crystallogr. 1999, 32, 115–119.
ꢀ
3. Herve, G.; Bernard, H.; Toupet, L.; Handel, H. Eur. J. Org.
Chem. 2000, 33–35.
4. Although theoretical calculations appearing in the lit.5
and confirmed by those reported in the present article,
anticipated that 10 was ‘perfectly feasible’, it still remained
a ‘challenging synthetic target’ until the trans isomers 6 and
8 were obtained and proved valuable starting materials for its
preparation. After 9 and 10 were isolated and characterized,
we revised some alkylations performed in the past with ethanes
bearing two leaving groups in the 1,2 positions on mixtures
containing 1–4.6 The analysis of the GC chromatograms,
acquired with the same method (see Section 4) able to separate
9 and 10, showed that, apart from 9, small percentages of a peak
with the same retention time of 10 was evidenced. Obviously,
a MS–GC study should be undertaken to confirm the presence
of 10.
€
24. Sheldrick, G. M. SHELX 97; University of Gottingen:
€
Gottingen, Germany, 1997.
25. Nardelli, M. J. Appl. Crystallogr. 1995, 28, 659.
26. Farrugia, L. J. J. Appl. Crystallogr. 1997, 30, 565.
27. Rojas-Lima, S.; Farfan, N.; Santillan, R.; Castillo, D.;
Sosa-Torres, M. E.; Hopfl, M. Tetrahedron 2000, 56, 6427–
6433.
28. Biosym/MSI, San Diego, CA.
5. Galasso, V.; Chuburu, F.; Handel, H.; le Baccon, M.; Jones, D.
THEOCHEM 2002, 582, 187–193.
29. Chang, G.; Guida, W. C.; Still, W. C. J. Am. Chem. Soc. 1989,
111, 4379–4386.
30. Maestro; Schrodinger: Portland, OR, 2005.
31. (a) Kollman, P. A.; Weiner, S. J.; Case, D. A.; Singh, C.; Ghio,
C.; Alagona, G.; Profeta, S.; Weiner, P. J. Am. Chem. Soc. 1984,
106, 765–784; (b) Weiner, S.; Kollman, P. A. J. Comput. Chem.
1986, 7, 230–252.
6. Argese, M.; Ripa, G.; Scala, A.; Valle, V. WO 97/49691, 1997.
7. Trademark for sodium dihydrobis(2-methoxyethoxy)alumi-
nate. The compound is largely used for industrial purposes
because of the strong reducing power that makes it similar to
LiAlH4 but without the drawbacks of the latter, such as pyro-
phoric nature, short shelf life and limited solubility.
32. Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T.
J. Am. Chem. Soc. 1990, 112, 6127–6129.
33. Ponder, J. W.; Richards, F. M. J. Comput. Chem. 1987, 8, 1016–
1024.
8. Buevich, A.; Chan, T.-M.; Wang, C. H.; McPhail, A. T.;
Ganguly, H. K. Magn. Reson. Chem. 2005, 43, 187–199.
9. (a) Karplus, M. J. Am. Chem. Soc. 1963, 85, 2870–2871; (b)
Jardetzky, C. D. J. Am. Chem. Soc. 1961, 83, 2919–2920.
10. Desiraju, G. R.; Steiner, T. The Weak Hydrogen Bond, IUCr
34. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.;
Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.;
Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi,
M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.;
Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.;
Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.;
Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.;
Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski,
J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.;
Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels,
A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck,
A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui,
Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov,
Monographs
Publications: New York, NY, 1999.
on
Crystallography;
Oxford
Science
11. Crystal data of compounds containing the TAD core were
retrieved from the Cambridge Structural Database (CSD, v.
5.27).12 Only the non-ionic molecules having three-coordi-
nated nitrogens were considered. Among the nine retrieved
structures, two are trans-fused and seven have cis-fused rings.
In eight of these structures the d1 distance is longer than the
d2 one.
12. Allen, F. H. Acta Crystallogr. 2002, B58, 380–388.
13. Cuevas, G.; Juaristi, E. J. Am. Chem. Soc. 2002, 124, 13088–
13096.
ꢀ
14. Herve, G.; Bernard, H.; Le Bris, N.; Yaouanc, J. J.; Handel, H.;
Toupet, L. Tetrahedron Lett. 1998, 39, 6861–6864.

M. Argese et al. / Tetrahedron 63 (2007) 6915–6923
6923
B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople,
J. A. Gaussian 03, Revision B05; Gaussian: Pittsburgh, PA,
2003.
37. Frisch, M. J.; Head-Gordon, M.; Pople, J. A. Chem. Phys. Lett.
1990, 166, 281–289.
38. Glendening, E. D.; Badenhoop, J. K.; Reed, A. E.; Carpenter,
J. E.; Bohmann, J. A.; Morales, C. M.; Weinhold, F. NBO
5.0; Theoretical Chemical Institute, University of Wisconsin:
Madison, WI, 2001.
35. Gordon, M. S. Chem. Phys. Lett. 1980, 76, 163–168.
36. Peng, C.; Ayala, P. Y.; Schlegel, H. B.; Frisch, M. J. J. Comput.
Chem. 1996, 17, 49–56.
39. Spartan’02; Wavefunction: Irvine, CA, 2002.
40. Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
41. Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1998, 37, 785–789.