Hydroacridines XVIII [1]. Synthesis and NMR spectroscopic investigation of (4aα,8aβ,9aα,10aβ)-tetradecahydroacridine and some of its derivatives

Article abstract of DOI:10.1007/PL00000108

The reductive amination of (R*,R*)-2,2′-methylene-bis-cyclohexanone (1) with methylamine and potassium borohydride affords a mixture of (4aα,8aβ,9aα,10aβ)- and (4aα,8aα,9aβ,10aα)-tetradecahydro-10-methylacridine (2, 3) in a ratio of approximately 1.3:1 in 57% overall yield. By N-demethylation of 2, via the N-nitrosamine 4 the first synthesis of (4aα,8aβ,9aα,10aβ)-tetradecahydroacridine (5) could be performed. The relative configurations and conformations of compounds 2-5 as well as the barrier of conformational inversion of 5 (ΔG^#₃₀₀ = 55.5 ± 0.4 kJ · mol^-1) were determined by NMR spectroscopy.

Full text of DOI:10.1007/PL00000108

Monatshefte fuÈr Chemie 129, 515±522 (1998)
Hydroacridines XVIII [1]. Synthesis
and NMR Spectroscopic Investigation of
(4aa,8ab,9aa,10ab)-Tetradecahydroacridine
and Some of its Derivatives
Francisc Potmischil1;Ã, Friedrich W. Vierhapper2,
and Hermann Kalchhauser2
1
Department of Organic Chemistry, University of Bucharest, RO-70031 Bucharest, Romania
2
Institute of Organic Chemistry, University of Vienna, A-1090 Vienna, Austria
Summary. The reductive amination of (R*,R*)-2,20-methylene-bis-cyclohexanone (1) with
methylamine and potassium borohydride affords a mixture of (4aꢀ,8aꢁ,9aꢀ,10aꢁ)- and
(4aꢀ,8aꢀ,9aꢁ,10aꢀ)-tetradecahydro-10-methylacridine (2, 3) in a ratio of approximately 1.3:1 in
57% overall yield. By N-demethylation of 2, via the N-nitrosamine 4 the ®rst synthesis of
(4aꢀ,8aꢁ,9aꢀ,10aꢁ)-tetradecahydroacridine (5) could be performed. The relative con®gurations
and conformations of compounds 2±5 as well as the barrier of conformational inversion of 5
ꢀÁG#300 ˆ 55:5 Æ 0:4 kJ Á molꢁ1† were determined by NMR spectroscopy.
Keywords. Acridines, tetradecahydro; 13C NMR; Conformational analysis; N-Nitrosamines.
Hydroacridine, 18. Mitt. [1]. Synthese und NMR-spektroskopische Untersuchung von
(4aa,8ab,9aa,10ab)-Tetrahydroacridin und einigen seiner Derivate
Zusammenfassung. Die reduktive Aminierung von (R*,R*)-2,20-Methylen-bis-cyclohexanon (1)
mit Methylamin und Kaliumborhydrid ergibt in einer Gesamtausbeute von 57% ein Gemisch aus
È
(4aꢀ,8aꢁ,9aꢀ,10aꢁ)- und (4aꢀ,8aꢀ,9aꢁ,10aꢀ)-Tetradecahydro-10-methylacridin (2, 3) im Verhalt-
nis von ca. 1.3:1. Durch N-Demethylierung von 2 gelang uÈber das Nitrosamin 4 die erste Synthese
von (4aꢀ,8aꢁ,9aꢀ,10aꢁ)-Tetradecahydroacridin (5). Die relativen Kon®gurationen und die Kon-
formationen der Verbindungen 2±5 sowie die Aktivierungsenergie der Konformationsumwandlung
von 5 ꢀÁG#300 ˆ 55:5 Æ 0:4 kJÁmolꢁ1† wurden NMR-spektroskopisch bestimmt.
Introduction
The reductive amination of (R*,R*)-2,20-methylene-bis-cyclohexanone (1) with
MeNH2 and KBH4 has been reported to yield an N-methylperhydroacridine by
Vysotskii [2]; its steric structure could not be speci®ed by the author. As by-products,
several stereoisomeric 2,20-methylene-bis-cyclohexanols have been reported.
* Corresponding author

516
F. Potmischil et al.
Scheme 1
By repetition of the procedure of Vysotskii we were able to demonstrate that the
described reaction yields a mixture of two stereoisomeric N-methylperhydroacri-
dines (Scheme 1) which could be identi®ed as (4aꢀ,8aꢁ,9aꢀ,10aꢁ)-tetrahydro-10-
methylacridine (2; the compound referred to in Ref. [2]) and (4aꢀ,8aꢀ,9aꢁ,10aꢀ)-
tetrahydro-10-methylacridine (3), respectively. As deduced from 13C NMR
spectroscopy, 2 was also present in the product mixture resulting from catalytic
hydroamination of 1 with MeNH2 and H2/Ni-Ru [3]; it could not be isolated in this
case, though.
We have now been able to isolate and characterize pure 2. Our data support the
assumption that, prior to Vysotskii's work [2], compound 2 has occured at least
twice in the literature as a component of reaction mixtures and without distinct
proof of its structure. In the ®rst case, mild hydrogenation of cis-1,2,3,4,4a,9,9a,10-
octahydro-10-methylacridine at platinum afforded presumably 2 (which was
referred to as N-methyl-ꢂ-perhydroacridine by the authors) in a mixture with its
cis-cisoid-trans isomer [4] from which only a small amount of its hydroiodide
could be isolated whose melting point (224±226ꢂC [4]) agrees fairly well with that
of the authentic hydroiodide prepared in the present work. The second case deals
with reductive amination of 1 with MeNH2 and HCOOH [5, 6] which has been
described to yield a mixture of four stereoisomeric N-methylperhydroacridines,
three of them being already known to the literature [5±7]. The fourth one, not
isolated, was characterized by the proton chemical shift of its N-CH3 group in the
1H NMR spectrum of the product mixture only [5, 6] which corresponds to that of
the authentic compound 2.

Hydroacridines
517
Scheme 2
In addition to the isolation and characterization of 2, we were able to obtain
(4aꢀ,8aꢁ,9aꢀ,10aꢁ)-tetradecahydroacridine (5) by N-demethylation of 2 via
(4aꢀ,8aꢁ,9aꢀ,10aꢁ)-tetradecahydro-10-nitrosoacridine (4, Scheme 2). To our
knowledge, this is the ®rst synthesis of compound 5 reported so far.
Results and Discussion
The reductive amination of 1 with MeNH2 and KBH4 was performed according to
Ref. [2], however, using a modi®ed workup procedure (see Experimental). After
removal of the 2,20-methylene-bis-cyclohexanols (referred to as ``neutral products''
1
in Scheme 1), the H NMR spectrum of the crude aminic fraction exhibited two
sharp N-CH3 resonances (CCl4; ꢂ ˆ 2.16 and 2.07 ppm, respectively) demonstrat-
ing the presence of two isomeric N-methylperhydroacridines. After separation, the
isomer with the N-CH3 peak at 2.16 ppm turned out to be the compound described
by Vysotskii; based on its 13C NMR spectrum it was assigned structure 2. Due to its
conformational mobility originating from the two cis junctions, only two sharp
signals are present in the room temperature 13C NMR spectrum of 2 (N-CH3:
39.42 ppm, C-9: 30.86 ppm; Ref. [3]: 39.78 and 31.08 ppm); the remaining signals
1
suffer from considerable line broadening. The second isomer whose H NMR
spectrum has been reported repeatedly in the literature (ꢂ(N-CH3) in CCl4: 2.07 [5,
6], 2.10 [7] ppm) was identi®ed as compound 3 by comparison of its 13C NMR
spectrum with that of an authentic sample [7].
During the reductive amination applied in this investigation, the con®gurations
at the tertiary carbon atoms in ꢀ positions to the carbonyl groups remain
unaffected. Therefore, the con®gurations of the resulting amines are related to
those of the starting diketone in a de®ned way. Obviously, conformer 1a leads to
amine 2, and conformer 1b gives rise to amine 3 (Scheme 1). Thus, only transoid
perhydroacridines are to be expected from diketone 1. Besides 2 and 3, the

518
F. Potmischil et al.
formation of a trans-transoid-trans-N-methylperhydroacridine (so far unknown)
cannot be excluded. The formation of this compound is, however, defavoured by
the fact that its piperidine ring should have a twisted boat form (like the central
ring in trans-transoid-trans-perhydroanthracene [8]). Based on the present
experience with 1, it might be predicted that the same reaction applied to
(R*,S*)-2,20-methylene-bis-cyclohexanone would result in a mixture of trans-
cisoid-trans, cis-cisoid-trans, and, possibly, cis-cisoid-cis (sterically less favoured)
N-methylperhydroacridines.
The possibility to isolate and purify compound 2 prompted us to strive for its
parent heterocycle 5 for which no synthesis has been reported so far. It could ®nally
be obtained in very poor overall yield by demethylation of 2 via N-nitrosamine 4
(Scheme 2), a strategy already successfully applied to other stereoisomers of N-
methylperhydroacridine [5±7]. To reduce losses, no puri®cation of intermediate 4
was carried out prior to denitrosation. For characterization purposes, a pure sample
of 4 was synthesized by nitrosation of 5. The structures of both 4 and 5 were
1
con®rmed by their H and 13C NMR spectra.
13C NMR spectroscopy of compounds 4 and 5
The 13C NMR spectrum of compound 5 has been recorded both at room and low
temperatures. At room temperature, only C-9 ± being the sole carbon atom that,
owing to symmetry, preserves its chemical shift during the fast conformational
inversion process of 5 (Scheme 2) ± gives rise to a sharp resonance (ꢂ ˆ 30.48 ppm).
Lowering the temperature causes further broadening of all signals except that of C-
9, and at 300 K coalescence is observed. Further temperature decrease results in 13
sharp peaks. The complete assignment, achieved from a 2D INADEQUATE
spectrum at 250 K, has been published previously [9]; to facilitate comparison, the
shift values are included in Table 1. From the low temperature 13C NMR data, a
free energy of activation of ÁG3#00 ˆ 55:5 Æ 0:4 kJ Á molꢁ1 and a rate constant of
k300 ꢃ 1370 sꢁ1 can be obtained for the chair-chair conformational interchange of
5. For the carbocyclic analogue of 5 (cis-transoid-cis-perhydroanthrancene) in
CHCl3, a value of ÁG# ˆ 60:5 Æ 0:4 kJ Á molꢁ1 has been reported [10].
318
The 13C NMR spectrum of N-nitrosamine 4 exhibits 13 sharp signals,
indicating conformational homogeneity even at room temperature. This fact can be
rationalized taking into account two effects: (i) a drastic steric interaction between
the N-nitroso oxygen atom and the C-4/C-5 methylene groups in the case of an
equatorially oriented linkage with respect to the piperidine ring (termed ``Aꢀ1;3†
strain'' by Johnson and Malhotra who have for the ®rst time analyzed its
conformational implications with respect to cyclohexane derivatives containing
exocyclic double bonds [11]), and (ii) the high rotational energy barrier of the N-N
bond in aliphatic N-nitrosamines (ÁG# ꢃ 96 kJ Á molꢁ1, [12] (temperature not
given)) which is almost twice as high than that determined above for the ring
inversion of 5. Thus, for the lowest energy conformation of 4 an all-chair
conformation with the N-O bond oriented anti relative to C-10a (and, at the same
time, to the equatorial C-5 methylene group) and syn to C-4a is to be expected (cf.
Scheme 2). In this arrangement, the axial C-4 methylene group does not interfere
with the nitroso oxygen atom. The conformation deduced above is strongly

Hydroacridines
519
Table 1. 13C chemical shifts (CDCl3; ꢂ in ppm from
internal TMS) and selected shift differences (cf. Text;
Áꢂ ˆ ꢂ(4) ± ꢂ(5)) of compounds 4 and 5; data for
compound 5 are taken from Ref. [9]; values with the
same superscript may be interchanged
4
5
Áꢂ
C-1
29.471
19.672
24.913
23.00
49.72
27.80
20.422
25.883
25.613
36.91
29.981
27.83
55.63
30.87
20.07
25.89
25.54
55.08
32.06
20.02
26.29
25.12
35.73
30.06
29.08
46.21
C-2
C-3
C-4
ꢁ2.54
ꢁ5.36
ꢁ4.26
C-4a
C-5
C-6
C-7
C-8
C-8a
C-9
1.18
C-9a
C-10a
ꢁ1.25
9.42
supported by a comparison of the 13C NMR chemical shifts of 4 (Table 1) with
those given in Ref. [13] for the syn and anti conformers of N-nitroso-cis-
decahydroquinoline (conformers B±s and B±a, compound 2) and N-nitroso-2ꢁ-
methyl-cis-decahydroquinoline (conformers B±s and A, compound 6).
Most relevant for the conformation of 4 are the 13C NMR signals of the four
tertiary carbons C-4a, C-8a, C-9a, and C-10a which were identi®ed by means of an
APT spectrum. A discrimination of ꢂ(C-4a) from ꢂ(C-10a) as well as of ꢂ(C-8a)
from ꢂ(C-9a) may well be achieved by a comparison of the chemical shift
differences of these carbon atoms when going from amine 5 to N-nitrosamine 4 (cf.
Table 1) with the appropriate differences in the system cis-decahydroquinoline
[14]/syn- and anti-N-nitroso-cis-decahydroquinoline [13]. In each of these N-
nitrosamine conformers, the ꢀ-carbons syn to N-O experience an up®eld shift of
more than 4 ppm, whereas the corresponding anti ꢀ-carbons are deshielded by
about 7 ppm with respect to the parent amine. This striking contrast in the
behaviour of the ꢀ-syn and ꢀ-anti carbon atoms leaves no doubt that the signal at
49.72 ppm has to be assigned to C-4a, whereas the resonance at 55.63 ppm must be
ascribed to C-10a.
The same effect, although to a much lesser extent, is exerted on the syn and anti
carbon atoms in the ꢁ positions of the piperidine ring of decahydroquinolines by
N-nitrosation. Therefore, the signals at 36.91 and 27.83 ppm could be unambigu-
ously assigned to C-8a and C-9a, respectively. In addition, the resonances of C-4
and C-5 were assigned by direct comparison with the 13C NMR shifts of
appropriate conformers of N-nitroso-cis-decahydroquinoline [13].

520
F. Potmischil et al.
Experimental
General
Melting points were determined using a Boetius hot-plate microscope and are uncorrected. 1H NMR
spectra in CCl4 were aquired with a Varian EM 360/390 NMR spectrometer (60 MHz, CW mode);
all other NMR spectra were measured on a Bruker AM 400 WB NMR spectrometer (1H:
400.13 MHz, 13C: 100.62 MHz) equipped with an Aspect 3000 computer as 0.4±1 M CDCl3 solutions
in 5 mm tubes. All NMR spectra are referenced to internal TMS. 13C NMR spectra were recorded in
the APT mode (attached proton test, [15, 16]) employing a relaxation delay of 3 s. Data processing
was performed on a work station (Bruker X 32) running the UXNMR software.
Syntheses
(R*,R*)-2,20-Methylene-bis-cyclohexanone (1, m.p.: 58ꢂC) was synthesized, separated, and puri®ed
as described earlier [17].
(4aꢀ,8aꢁ,9aꢀ,10aꢁ)-Tetradecahydro-10-methylacridine (2) and
(4aꢀ,8aꢀ,9aꢁ,10aꢀ)-Tetradecahydro-10-methylacridine (3)
10.4 g (50 mmol) of 1 were subjected to reductive amination with 36 ml of 30% aqueous MeNH2 and
5.4 g (100 mmol) of KBH4 according to Ref. [2].
The reaction mixture was then acidi®ed with concentrated HCl (pH ˆ 2). The precipitated
inorganic chlorides were removed by ®ltration and washed with MeOH; the volatile components of
the combined ®ltrates were removed on a steam bath applying vacuum towards the end of the
evaporation procedure. To the residue, 200 ml H2O were added, and the mixture was shaken on a
steam bath for 15 min. Then, the aqueous extract was separated from a viscous, water insoluble phase
(termed ``neutral products''; cf. Scheme 1). The viscous product was washed with additional 40 ml of
H2O, and the combined aqueous solutions were alkalized with 40% aqueous NaOH to pH 12. The
resulting oily organic layer which crystallized partly upon cooling was extracted twice with diethyl
ether (50‡30 ml); the etheral solution was dried over K2CO3, and an aliquot was taken for 1H NMR
spectroscopy. Two sharp singlets (N-CH3) at 2.16 and 2.07 ppm, respectively (CCl4, internal TMS)
were observed in the spectrum. After removal of the diethyl ether, the remainder of the etheral
solution afforded 5.9 g of an oily residue (57% overall yield of N-methylperhydroacridines) which
crystallized in part. After ®ltration, 2.7 g (26%) of crude solid and 3.2 g of oil were obtained.
Trituration with acetone resulted in a melting point of 58±64ꢂC for the crude solid, and its 13C NMR
spectrum was identical with that of 2 as reported previously [3]. In the 13C NMR spectrum of the oily
fraction, the two characteristic signals of 2 (C-9 and N-CH3) were present in addition to the fourteen
signals of an authentic sample of 3 [7]. From the relative intensities of the N-CH3 signals (39.41 ppm
for 2, 36.45 ppm for 3; CDCl3), an approximate product distribution of 2:3 ˆ 20:80 could be
estimated. Thus, the total yield of 2 amounts to about 32% and that of 3 to about 25%.
Compound 2 was puri®ed via its hydroiodide. Crude 2 obtained as described above (m.p.: 58±
64ꢂC) was dissolved in warm acetone. The solution was acidi®ed with 67% HI until a pH value of 3
was reached, and formation of hydroiodide crystals could soon be observed. After 24 h the
hydroiodide was ®ltered, washed with acetone, and recrystallized from EtOH. The puri®ed
hydroiodide was mixed with an excess of 40% aqueous NaOH, and the mixture was subjected to
steam distillation. Towards the end of the distillation, cooling had to be interrupted to allow melting
of 2 which had solidi®ed in the condenser.
M.p.: 65±66ꢂC (Ref. [2]: m.p.: 63±64ꢂC); 1H NMR (298 K, CDCl3): ꢂ ˆ 2.25 ppm (s, N-CH3), all
other signals broadened; 13C NMR (298 K, CDCl3): ꢂ ˆ 39.42 (N-CH3), 30.86 (C-9) ppm, all other
signals broadened; picrate: m.p.: 194±196ꢂC (MeOH; Ref. [2]: m.p.: 191±192ꢂC); hydroiodide: m.p.:

Hydroacridines
521
227±229ꢂC (EtOH); methiodide (prepared at room temperature in acetone): m.p.: 278±279ꢂC (not
recrystallized).
Compound 3 was not isolated in pure form from the oily fraction. 13C NMR (298 K, CDCl3):
ꢂ ˆ 70.33 (C-4a), 63.61 (C-10a), 39.45, 37.72, 37.06, 36.45 (N-CH3), 33.79, 30.92, 30.67, 27.46,
27.01, 26.15, 25.83, 19.79 ppm; a complete signal assignment will be presented in one of the
forthcoming papers of this series.
The neutral products were not worked up further.
(4aꢀ,8aꢁ,9aꢀ,10aꢁ)-Tetrahydroacridine (5)
To a solution of 5.6 g (27 mmol) of 2 in 70 ml AcOH and 120 ml H2O, a solution of 40 g (580 mmol)
of NaNO2 in 50 ml H2O was added dropwise during 0.5 h on the steam bath. After the addition was
complete, heating was continued for one more hour. Subsequently, another batch of 70 ml AcOH and
120 ml H2O was added, and a solution of 80 g (1.6 mol) NaNO2 in 100 ml H2O was dropped to the
mixture over 0.5 h, followed by heating for 1 h. After cooling, the oily product formed was extracted
three times with 250 ml diethyl ether each. The etheral extracts were washed with 5Â150 ml 10%
aqueous NaOH and dried over Na2SO4. The ether was removed, and the residue (crude 4) was
re¯uxed for 2 h with 50 ml of 36% HCl. The excess of HCl was then removed under reduced
pressure, and the residue was rendered strongly alkaline by addition of 40% aqueous NaOH. Steam
distillation afforded pure 5 which crystallized completely in the condenser. It was recovered by
dissolution in a small amount of diethyl ether and subsequent evaporation of the solvent.
Yield: 0.35 g (6.7%); long colourless prisms; m.p.: 116±117ꢂC; 1H NMR (250 K, CDCl3):
ꢂ ˆ 3.03 (m, ꢃ1=2 ˆ 5.2 Hz, 1H, H-10a), 2.88 (dt, J ˆ 12.4, 4.4 Hz, 1H, H-4a) ppm; C13H23N (193.3);
calcd.: N 7.24; found: N 7.11; picrate (prepared in MeOH, crystallized after complete evaporation of
the solvent): m.p.: 163±165ꢂC.
(4aꢀ,8aꢁ,9aꢀ,10aꢁ)-Tetradecahydro-10-nitrosoacridine (4)
To a solution of 145 mg (0.75 mmol) of 5 in 1 ml AcOH and 2 ml H2O, a solution of 150 mg
(2 mmol) of NaNO2 in 2 ml H2O was added dropwise on the steam bath, and heating was continued
for 10 min. A yellow oil was formed which quickly crystallized upon cooling, affording an almost
white solid. The latter was separated by ®ltration, washed with H2O, dissolved in a minimal amount
of acetone at room temperature, and reprecipitated by addition of an excess of a diluted aqueous
KHCO3 solution ( pH ˆ 8.5). After ®ltration and washing to neutral pH with H2O, pure 4 was
obtained.
Yield: 158 mg (94%); m.p.: 100ꢂC; 1H NMR (298 K, CDCl3): ꢂ ˆ 5.10 (dt, J ˆ 12.5, 4.4 Hz, 1H,
H-4a), 4.07 (m, ꢃ1=2 ˆ 8.4 Hz, 1H, H-10a) ppm; C13H22N2O (222.3); calcd.: N 12.60; found: N 12.53.
Acknowledgements
È
Financial support of the NMR measurements by the Fonds zur Forderung der wissenschaftlichen
È
Forschung in Osterreich (Project P 6537 C) is gratefully acknowledged.
References
[1] For part 17, see Potmischil F, Herzog H, Buddrus J (1998) Magn Reson Chem 36 (in press)
[2] Vysotskii VI (1970) Khim Geterosikl Soedin 1970: 1236
[3] Kriven'ko AP, Nikolaeva TG, Yudovich LM, Komyagin NT, Yanovskii AI, Struchkov YT,
Kharchenko VG (1987) Khim Geterosikl Soedin 1987: 1645
[4] Masamune T, Ohno M, Takemura K, Ohuchi S (1968) Bull Chem Soc Japan 41: 2558

522
F. Potmischil et al.: Hydroacridines
[5] Barbulescu N, Potmischil F (1969) Tetrahedron Lett 60: 5275
[6] Barbulescu N, Potmischil F (1970) Rev Roum Chim 15: 1601
[7] Barbulescu N, Potmischil F (1970) Liebigs Ann Chem 735: 132
[8] Johnson WS (1953) J Am Chem Soc 75: 1498
[9] Potmischil F, Kalchhauser H (1994) Magn Reson Chem 32: 563
[10] Vanhee P, De Pessemier F, Anteunis M, Tavernier D (1979) Rec Trav Chim Pays-Bas 98: 294 and
references cited therein
[11] Johnson F, Malhotra SK (1965) J Am Chem Soc 87: 5492; Malhotra SK, Johnson F (1965) J Am
Chem Soc 87: 5493
[12] Cerioni G, Giumanini AG, Verardo G, Dahn H (1994) Magn Reson Chem 32: 46 and references
cited therein
[13] Vierhapper FW (1980) Monatsh Chem 111: 551
[14] Vierhapper FW, Eliel EL (1977) J Org Chem 42: 51
[15] Brown DW, Nakashima TT, Rabenstein DL (1981) J Magn Reson 45: 302
[16] Pratt SL, Shoolery JN (1982) J Magn Reson 46: 535
[17] Barbulescu N, Potmischil F (1969) Rev Roum Chim 14: 1427
Received December 12, 1997. Accepted December 22, 1997