1-(Arylalkyl)quinolizidine Derivatives and Thio-Isosteric Analogues as Ligands for Sigma Receptors

Article abstract of DOI:10.1002/hlca.200490055

A set of 1-(arylalkyl)quinolizidines, isosteric thioanalogues, and variously functionalized congeners were synthesized (see 1-25) and tested for affinity to sigma 1 and sigma 2 receptor subtypes, by displacing [ ³H]- (+)-pentazocine and [³

Full text of DOI:10.1002/hlca.200490055

580
Helvetica Chimica Acta Vol. 87 (2004)
1-(Arylalkyl)quinolizidine Derivatives and Thio-Isosteric Analogues as
Ligands for Sigma Receptors
by Anna Sparatore*a), Federica Novellib), and Fabio Sparatoreb)
a
¡
) Istituto di Chimica Farmaceutica, Universita di Milano, Viale Abruzzi 42, I-20131 Milano
(tel.: ‡390250317554; fax: ‡390250317565; e-mail: anna.sparatore@unimi.it)
b
¡
) Dipartimento di Scienze Farmaceutiche, Universita di Genova, Viale Benedetto XV 3, I-16132 Genova
A set of 1-(arylalkyl)quinolizidines, isosteric thioanalogues, and variously functionalized congeners were
synthesized (see 1 25) and tested for affinity to sigma 1 and sigma 2 receptor subtypes, by displacing [3H]- (‡)-
pentazocine and [3H]DTG from guinea pig brain and rat brain preparations, respectively. All compounds
exhibited a good affinity for the s1 subtype, with subnanomolar Ki values for the best of them, while only modest
or poor affinity for the s2 subtype was observed (Tables 1 and 2). Some structureÀactivity relationships were
put forward.
Introduction. Sigma receptors occur in at least two classes of binding sites, namely
s1 and s2, which are widely distributed in the central nerve system (CNS) and in several
peripheral tissues [1][2] and also expressed in some human and rodent tumor cell lines
[3][4]. The functional roles of the two receptor subtypes are being progressively
defined, particularly in the pathogenesis of psychiatric and motor disorders, but also
outside of the nervous system [5 7], which accounts for the large number of attempts
to obtain selective sigma binding-site ligands. Many structurally unrelated compounds
have been described as being able to bind to sigma receptors, but only few display high
affinity and selectivity for the sigma receptor subtypes [8 13].
Recently we described two novel types of ligands for sigma receptors that
correspond, respectively, to the general structures A-a or A-b ((1R, 9aR) or (1S, 9aR)-
1-(2-arylethyl)octahydro-2H-quinolizines) and B (1'-substituted-(spiro[1,2,4-benzo-
triazine-3(4H),4'-piperidines])) [14][15]. Many compounds of both types exhibited
high affinity for the s1 receptor subtype, with Ki in the low nm range, while the affinity
for s2 subtype of compounds so far tested was ten to more than 100 times lower. In type
B compounds, the lengthening of the aliphatic chain from one to four CH2 units did not
change significantly the affinity, which was, however, improved when five CH2 units are
present. On the other hand, in a set of N-(arylalkyl)piperidines [8][9][16], of which
compounds of type A-a or A-b could be considered −closed× analogues with an imposed
conformation, the affinity and the selectivity for the s1 subtype increased with the
increasing distance between the aromatic ring and the basic N-atom.
Therefore, we deemed it worthwhile to investigate whether the latter structural
feature would improve affinity for the s1 subtype also in type A-a and A-b compounds
in spite of the rigidity and bulkiness of the terminal quinolizine ring. Thus additional
compounds whose aromatic and octahydro quinolizine (quinolizidine) nuclei are
separated by one to four C- and/or S-atoms were prepared, taking into account the

Helvetica Chimica Acta Vol. 87 (2004)
581
well-known bioisosterism of a CH2 unit to a S-atom. These compounds correspond to
the general formulae C-a and C-b (absolute configurations; 11 is racemic), which
include also the previously described compounds of types A-a and A-b, i.e., compounds
1 18.
To further investigate the observed [14] negative influence of the presence of
oxygenated functions on the affinity for the s1 receptor and, more generally, to
investigate the influence of the nature of the chain that links the quinolizidine and
benzene moieties, a group of miscellaneous compounds (i.e., 19 23) was also tested.
Compounds 20 [17] and 22 and 23 [18] were already described by us. Finally, the
claimed [8][16] existence in the s1 receptor of a bulk-tolerating region at a given
distance from the proton-donor site was explored by testing the two previously described
quinolizidine derivatives 24 [19] and 25 [20], bearing two aromatic moieties linked to
the same C-atom, which is separated from the basic N-atom by three or four atoms.

582
Helvetica Chimica Acta Vol. 87 (2004)
In this paper, the synthesis and the biological evaluation of compounds 1 25 are
reported.
Syntheses. For the preparation of compound C-a with m ˆ n ˆ 0, i.e., of 1, the
coupling of w-chlorolupinane (ˆ(1R,9aR)-1-(chloromethyl)octahydro-2H-quinoli-
zine) with PhMgBr was firstly assayed (Scheme 1), but the reaction gave only a
modest yield of the expected (1S,9aR)-octahydro-1-(phenylmethyl)-2H-quinolizine.
Probably due to steric hindrance, the rate of the coupling reaction was very low, and the
competitive intra- and/or intermolecular quaternarization of chlorolupinane prevailed.
Scheme 1
Therefore, another possibility was explored, although, unavoidably, a mixture of
stereoisomers resulted: 1-quinolizidinone (octahydroquinolizin-1-one) was reacted
with (4-fluorobenzyl)magnesium chloride, followed by dehydration of the formed
tertiary alcohol and reduction of the unsaturated compound (Scheme 2: only one of the

Helvetica Chimica Acta Vol. 87 (2004)
583
possible stereoisomers is represented). Quinolizidin-1-one exists predominantly (90%)
in the trans fused chair-chair conformation [21] and reacts with Grignard reagents to
give mixtures of epimeric alcohols. In the reaction with arylmagnesium bromides, the
axial alcohols largely prevail [22][23], while, with MeMgI, the ratio of epimeric
alcohols is inverted [24]. In our reaction of quinolizidin-1-one with (4-fluorobenzyl)-
magnesium chloride, only a single racemate was isolated, whose sturdy resistance to
dehydration suggests it to be the equatorial alcohol 26. However, by reacting this
alcohol with P2O5 in phosphoric acid at 1658, an unsolitary (by TLC) unsaturated
compound 27 was finally obtained. (1H-NMR: d 3.3 (s, 1.8 H, CH2 between Ar and
CˆC); 5.38 5.47 (m, 0.2 H, olefinic H), suggesting that, in the dehydration, the
angular H-atom was preferentially eliminated to form the enamine 27.
Scheme 2
The presence of a tetrasubstituted CˆC bond in 27 accounts for the very slow
absorption of 1 mol of H2 in the presence of 10% Pd/C, giving, apparently, a single
saturated compound (by TLC and NMR). The protons of the quinolizidinylmethylene
moiety of this saturated compound give rise, in the 1H-NMR spectrum, to a sequence of
multiplets between d 0.80 and 3.10, whose overall outline is quite similar to that which
characterizes several epi-lupinane derivatives, thus suggesting that the 4-fluorobenzyl
residue is linked equatorially to the quinolizidine nucleus. Therefore, the compound
obtained is tentatively assigned as the racemate of (1RS,9aSR)-1-[(4-fluorophenyl)-
methyl]-octahydro-2H-quinolizine (11).
To obtain the (1R,9aR)-octahydro-1-(3-phenylpropyl)-2H-quinolizine (4), the
cross-coupling between chlorolupinane and PhEtMgBr was attempted, but it failed
completely even in the presence of the CuI catalyst suggested by Tamura et al. [25][26].
On the contrary, good results were obtained when the BnMgCl was reacted with w-
cyanolupinane (ˆ(1S,9aR)-octahydro-2H-quinolizine-1-acetonitrile), and the result-
ing ketone 19 was reduced with hydrazine under HuangÀMinlon conditions [27]
(Scheme 3). To avoid the drastic conditions required by this reaction, the reduction of
the corresponding ketone tosylhydrazone with NaBH4 suggested by Caglioti [28] was
initially attempted, but only the corresponding alcohol (as mixture of diastereoisom-
ers) was obtained.

584
Helvetica Chimica Acta Vol. 87 (2004)
Scheme 3
S-Aryl- or S-(arylalkyl)thiolupinines (5 9) and -epithiolupinines (14 18)
[(1R,9aR)- and (1S,9aR)-1-[(arylthio)methyl]- or (1-{[(w-arylalkyl)thio]methyl}octa-
hydro-2H-quinolizines], respectively, were obtained, without any particular difficulty,
by reacting w-chlorolupinane or w-chloroepilupinane with benzene- and 4-fluoroben-
zenethiol, benzene- and 4-chlorobenzenemethanethiol, and benzeneethanethiol
(Scheme 4). Due to higher steric hindrance, the reaction with the axial chlorolupinane
was sluggish and gave the expected sulfides 5 9 in lower yields than the reaction with
the equatorial w-chloroepilupinane ( ! 14 18).
Scheme 4
S-(4-Fluorophenethyl)thiolupinine (ˆ(1R,9aR)-1-({[2-(4-fluorophenyl)ethyl]thio}-
methyl)octahydro-2H-quinolizine; 10) was of particular interest for comparing its
affinity for s receptors with that of the previously described S-(4-fluorophen-
acyl)thiolupinine (ˆ1-(4-fluorophenyl)-2-{[(1R,9aR)-(octahydro-2H-quinolizin-1-yl)-
methyl]thio}ethanone; 20) [14][17]. Since the 4-fluorobenzeneethanethiol is as yet, far
unknown and unavailable, 10 was obtained by reacting 4-fluorophenethyl bromide with
thiolupinine (ˆ(1R,9aR)-octahydro-2H-quinolizine-1-methanethiol) [17] (Scheme 5).
Finally, 1-(4-fluorophenyl)-2-({(1R,9aR)-octahydro-2H-quinolizin-1-yl]methyl}thio)-
ethanol was prepared by reduction of S-(4-fluorophenacyl)thiolupinine (20) with

Helvetica Chimica Acta Vol. 87 (2004)
585
Scheme 5
NaBH4; the alcohol 21 was obtained as a mixture of diastereoisomers. No attempts to
resolve the mixture were made, and it was tested as such (Scheme 5).
Biological Evaluation and Discussion.
Novel and previously described com-
pounds 1 25 were tested in vitro to evaluate their affinities for the s1 receptor subtypes
through the displacement of [3H]- (‡)-pentazocine from guinea pig brain preparations.
Selected compounds were also assayed for affinity to s2, 5 HT2A, and D2 receptor
subtypes by displacing [3H]DTG, [3H]ketanserin and [3H]nemonapride, respectively,
from rat brain (s2, 5 HT2A), and rat striatum (D2) preparations.
Results of binding assays for compounds 1 18 to s1 and s2 receptor subtypes are
collected in Table 1, while results concerning compounds 19 25 are collected in
Table 2. The reported data show that all compounds exhibit good affinity for the s1
receptor subtype, with subnanomolar Ki values for the best of them. On the contrary, a
clear trend for modest or poor affinity to the s2 subtype was observed, although only
representative compounds were assayed in this case.
Affinity for the s1 subtype increased with the increasing distance between the
benzene ring and the quinolizidine moieties, however, also the nature of the connecting
chain influenced the increase.
Indeed, while the expected bioisosterism between a S-atom and a CH2 group was, in
principle, confirmed, the actual bioequivalence of these groups is lower when the S-
atom is directly linked to the benzene ring as in compounds 5, 6, 14, and 15. Moreover,
the introduction of an oxygenated function (hydroxy or oxo) at the aliphatic chain
produced a decrease in affinity (compare compounds 4 and 10 with 19, 20, and 21), but
to a lesser extent than previously observed for compounds of structure A [14]. Thus, the
negative effect of the presence of an oxygenated function seems to be largely
counterbalanced by the positive effect of the elongation of the aliphatic chain. Similar

586
Helvetica Chimica Acta Vol. 87 (2004)
Table 1. Binding Affinities of 1 18 for s1 and s2 Receptor Subtypes
Kia) [nm]
Ligand
Ratio Ki s2/Ki s1
s1
s2
(1R)-Substituted quinolizidines
1
35.2c)
38.0d)
6.6e)
2b)
30
)
0
7.9
33.3
g
3b)
220g)
4
5
6
7
8
9
10
4.2c)
89.0f)
13.2c)
3.2f)
463h)
35.1
1.0 f)
186h)
186.0
247.1
233.3
0.63f)
0.63f)
155.7h)
147h)
(1S)-Substituted quinolizidines
(Æ)-11
12b)
13b)
14
15
16
11.0c)
6.5d)
h
40
)
5
114.9
350.5
3.7e)
1297i)
10.0f)
4.4c)
578h)
131.4
0.52f)
0.23f)
0.38f)
17
18
233.6h)
138.4h)
1015.6
364.2
a) Means of duplicate experiments: each value differed from the mean by less than 10%. b) Previously
published data [14], included for comparison. Ki (nm) of reference ligand haloperidol used in s1 experiments:
c) 4.4. d) 3.64. e) 2.01. f) 1.45. Ki (nm) of reference ligand haloperidol used in s2 experiments: g) 91.0 . h) 78.7.
i) 88.2.
considerations apply also to the 1-lupinine esters 22 and 23 (4-chlorophenoxy- and 4-
chlorophenylthioacetate, resp.), which still exhibited Ki of 37 and 24 nm, respectively.
Thus, the intercalation of an ester function might represent an easy way to increase the
distance between the quinolizidine and benzene moieties, overcoming the difficulties of
synthesizing long-chain arylalkyl quinolizidines.
It is worth noting that compounds somewhat similar to 22, and 23, such as 2-(4-
chlorophenoxy- or chlorophenylthio)propanoic and -butanoic acid esters of tropanol,
(ˆ(3-endo)-8-methyl-8-azabicyclo[3.2.1]octan-3-ol) have been recently found
[29][30] to exhibit rather poor affinity for the s1 receptor subtype. Such differences
in affinity might be due to the different bicyclic aminoalcohol, or possibly to the
branching of the aliphatic chain in the relevant esters.
All compounds bearing the arylalkyl chain in the b-position at C(1) of the
quinolizidine ring displayed higher affinities than those bearing the same chain in the a-
position, as already observed for the 1b-(arylethyl)quinolizidine derivatives A-b [14].
However, the difference in affinity in each couple of epimers decreased with the
increasing length of the chain. The introduction of F or Cl at the para-position of the
phenyl ring strongly increased the affinity for the s1 receptor subtype, but also this
effect was quenched by the increasing length of the aliphatic chain; thus, compounds 9
and 10 exhibited the same K1 value (0.63 nm).

Helvetica Chimica Acta Vol. 87 (2004)
587
Table 2. Binding Affinities of 19 25 for the s1
Receptor Subtype
Ligand
Kia) (nm)
19
20
21
22
23
24
25
10.2b)
9.0c)
9.2b)
36.9b)
24.2b)
49c)
6.4b)
a) Mean of duplicate experiments: each value
differed from the mean by less than 10%. b) Ki
(nm) of reference ligand haloperidol 1.45. c) Ki
(nm) of reference ligand ifenprodil 1.4 (see
[14]).
The affinity-enhancing effect of the mentioned structural features (length of chain,
equatorial position of the substituent at the quinolizidine moiety, presence of a halogen
atom on the aromatic ring) showed additive character; when (1R,9aR)-octahydro-1-
[(phenylthio)methyl]-2H-quinolizine (5; Ki ˆ 89 nm) and (1S,9aR)-octahydro-1-({[4-
chlorophenyl)methyl]thio}methyl)-2H-quinolizine (17; Ki ˆ 0.23 nm) are compared, a
387-fold increase in affinity is observed.
Finally, for compounds 24 and 25, each bearing two fluorophenyl residues, a high
affinity for the s1 subtype was still observed, albeit lower than that of the related
compounds 3 and 7, that each bear only a single halobenzene moiety. The increasing
distance between the basic N-atom and the aromatic part of the molecule produces a
significant increase in affinity also in these compounds, which exhibit a Ki of 49 and
6.4 nm, respectively.
Taken together, these results support further the s1-receptor pharmacophore model
proposed by Glennon and co-workers [8][9][31], which consists of a primary
hydrophobic binding site situated 6 10ä from the basic N-atom binding site, with a
secondary binding site, associated with a region of bulk tolerance, situated 2.5 3.9 ä
from the amine site. Thus, the aromatic part of our ligands 1 23 could interact with
either the primary or secondary hydrophobic site of the receptor, depending on the
effective distance from the basic N-atom, taking into account the possible folding of the
aliphatic chain. On the other hand, the two 4-fluorophenyl rings of compounds 24 and
25 should better accommodate in the secondary hydrophobic bulk-tolerating region.
Concerning the affinity for the s2 receptor subtype, compounds tested so far exhibit
Ki values that were one to more than two orders of magnitude lower than those found
for the s1 subtype. Moreover, the above structural features, which improved the affinity
for the s1 subtype, particularly the equatorial position of the substituent at the
quinolizidine moiety and the presence of a halogen atom, did not play an identical role
in the case of the s2 subtype.
Compound 17, which displays the highest affinity for the s1 subtype and the highest
selectivity vs. the s2 subtype, was also tested for affinity to serotonin 5-HT2A and
dopamine D2 receptor subtypes that are implicated, though in different measures, in the

588
Helvetica Chimica Acta Vol. 87 (2004)
activity of conventional and atypical neuroleptic drugs. As was previously observed for
compounds 12 and 13 [14], in this case only modest affinity for 5-HT2A (Ki ˆ 183 nm)
was found, while affinity for the D2 subtype was once more very poor, with only 40%
inhibition of [3H]nemonapride binding at a 10 mm concentration.
Conclusions. Several functions of s1 receptors have been progressively uncovered,
but the interest in selective ligands for this receptor subtype is now addressed mainly
because of their neuroprotective and anti-amnesic potentials in aging-related
pathologies [5]. Recently, 4-phenyl-1-(4-phenylbutyl)piperidine was found to afford
neuroprotection in experimental stroke, through attenuation of neuronal nitric oxide
synthase activity and ischemia-evoked NO production [32].
The (arylalkyl)quinolizidines, their isosteric thioanalogs, and the variously func-
tionalized congeners presently considered (1 25) behave as good ligands for the s1
receptor subtype with subnanomolar binding constants for the best compounds, while
they display a manifold lower, or definitely poor, affinity for s2, 5 HT2A, and D2
receptor subtypes. Therefore, this class of quinolizidine derivatives deserves further
investigation toward development of still more potent and selective compounds.
The proper length of the benzene-quinolizidine-connecting aliphatic chain repre-
sents an important structural feature that defines a good ligand. On the other hand, the
demonstrated good affinity of thio ethers and esters permit facile construction of new
sets of quinolizidine derivatives with aliphatic chains of variable length and nature,
overcoming the difficulties of synthesizing long chain (arylalkyl)quinolizidines. In this
manner, a larger molecular diversity can also be introduced in the new ligands, which
will be useful for studying the receptor requirements around the central proton donor
site.
¡
The financial support from the Italian −Ministero dell×Istruzione, dell×Universita e Ricerca× (MIUR) is
gratefully acknowledged.
Experimental Part
General. All commercially available solvents and reagents were used without further purification, unless
otherwise stated. CC ˆ column chromatography. M.p.s: B¸chi apparatus; uncorrected. IR Spectra: Perkin-
Elmer Paragon-1000-PC spectrophotometer; KBr pellets for solid, and neat for liquid; nÄ in cmÀ1 1H-NMR
.
Spectra: Varian Gemini-200 spectrometer; CDCl3 with Me4Si as internal standard; d in ppm, J in Hz; Q ˆ
octahydroquinolizine ring. Elemental analyses were performed on a Carlo-Erba EA-1110 CHNSÀO instrument
in the Microanalysis Laboratory of the Department of Pharmaceutical Sciences of Genoa University.
(1S,9aR)-Octahydro-1-(phenylmethyl)-2H-quinolizine (1). To a soln. of w-chlorolupinane [33] (1.69 g,
9 mmol) in dry Et2O (5 ml), 3m phenylmagnesium bromide soln. in Et2O (Aldrich; 5 ml, 15 mmol) was added,
and the soln. was heated under reflux for 17 h. After cooling, 1n HCl (25 ml) was added dropwise, and the acid
soln. was extracted with Et2O. The aq. phase was alkalinized with 2n NaOH and extracted with Et2O. After
evaporation, the residue was distilled at 115 1208 (air-bath temp.)/0.05 Torr: 1 (0.28 g, 13.6%). Oil which
solidified on standing. M.p. 30 33 8. 1H-NMR (CDCl3): 1.10 2.20( m, 14 H, Q); 2.60 2.80( m, 1 H, ArCH2Q);
2.80 3.00( m, 1 H of ArCH2Q, 2 Ha near N of Q); 7.10 7.40( m, 5 arom. H). Anal. calc. for C16H23N: C 83.78,
H 10.11, N 6.11; found: C 83.55, H 10.26, N 6.10.
(Æ)-(1RS,9aRS)-1-[(4-Fluorophenyl)methyl]octahydro-2H-quinolizin-1-ol (26). A soln. of freshly distilled
4-fluorobenzyl chloride (1.5 g, 10.4 mmol) in dry Et2O (20ml) was added dropwise to 0.25 g (10mmol) of Mg-
turnings covered with dry Et2O, and the mixture was refluxed under N2 until all Mg was dissolved. A soln. of
octahydro-2H-quinolizin-1-one [34] (1.6 g, 10mmol) in dry Et 2O (10ml) was added, and the soln. was refluxed
for 18 h. The soln. was extracted with 0.1n HCl, and the acid soln. was alkalinized and extracted with Et2O. After

Helvetica Chimica Acta Vol. 87 (2004)
589
evaporation, the oily residue (2.43 g) was purified by CC (neutral alumina (1:15), CH2Cl2, then CH2Cl2
containing 0.5% (v/v) of MeOH): 26 (2.08 g, 78.6%). Oil that soon solidified. Repeated crystallization from
1
hexane gave 1.12 g (42.4%) of 26. M.p. 80 81 8. H-NMR (CDCl3): 1.00 2.20 (m, 13 H, Q); 2.45 (dd, J ˆ 14,
1 H, ArCH2); 2.60 2.80( dm, 1 H, ArCH2); 2.80 3.05 ( m, 2 Heq near N of Q, OH collapsing with D2O); 6.90
7.05 ( m, 2 arom. H); 7.10 7.25 ( m, 2 arom. H). Anal. calc. for C16H22FNO: C 72.97, H 8.42, N 5.32; found:
C 73.27, H 8.32, N 5.41.
1-[(4-Fluorophenyl)methyl]-3,4,6,7,8,9-hexahydro-2H-quinolizine (27). Phosphorous pentoxide (1.5 g) was
added to a soln. of 26 (1.1 g, 4.2 mmol) in polyphosphoric acid (PPA; 10g). The mixture was heated to 165 8 and
stirred at 1658 for 1 h. After cooling, H2O (30ml) was added, and the mixture neutralized with Na 2CO3 and
alkalinized to pH 10by 6 n NaOH. By extraction with Et2O, impure 27 (0.93 g, 90%) was obtained. Light
brownish oil. TLC (alumina, CH2Cl2): very minor spot just behind the main spot. IR (neat): no OH band.
1H-NMR (CDCl3): 1.00 3.10 (m, 14 H, Q); 3.30( s, 1.8 H, ArCH2CˆC); 5.38 5.47 (m, 0.2 H, > CˆCH);
6.80 7.20( m, 4 arom. H).
(Æ)-(1RS,9aSR)-1-[(4-Fluorophenyl)methyl]octahydro-2H-quinolizine (11). Compound 27 (0.90 g) was
dissolved in EtOH (20 ml) and hydrogenated at r.t. and atmospheric pressure over 10% Pd/C (0.5 g) until the
absorption of H2 ceased. Catalyst and solvent were removed, and the residue was purified twice by CC (neutral
alumina (1:20), CH2Cl2) pure (TLC) 11 (0.57 g, 62.8%). Colorless oil. 1H-NMR (CDCl3): 0.80 2.20 (m, 14 H,
Q); 2.60 3.10( m, therein dd, J ˆ 14, 11, 4 H, 2 Heq. near N of Q, ArCH2Q); 6.85 7.20( m, 4 arom. H). Anal.
calc. for C16H22FN: C 77.69, H 8.96, N 5.66; found: C 77.41, H 9.20, N 5.54.
1-[(1S,9aR)-Octahydro-2H-quinolizin-1-yl]-3-phenylpropan-2-one (19). This ketone was first described in
[35]. A soln. of freshly distilled w-cyanolupinane [36] (1.78 g, 10mmol) in a few ml of dry Et 2O was added to 1m
benzylmagnesium chloride soln. in Et2O (Aldrich; 18 ml, 18 mmol), previously diluted with dry Et2O (10ml).
The mixture was heated under reflux for 11 h and then, under ice cooling, 1n HCl (38 ml) was added, and the
acid soln. was further washed with Et2O. The aq. phase was alkalinized with 30% KOH soln. and extracted with
Et2O, the extract evaporated, and the residue crystallized from petroleum ether: 19 (2.13 g, 78.6%). M.p. 73
748. IR (KBr): 1690(C ˆO). Anal. calc. for C18H25NO: C 79.66, H 9.29, N 5.16; found: C 79.44, H 9.35, N 5.37.
(1R,9aR)-Octahydro-1-(3-phenylpropyl)-2H-quinolizine (4). Na (0.19 g, 8.26 mmol) was dissolved in
diethylene glycol (6 ml), hydrazine monohydrate (0.39 ml, 7.7 mmol) and 19 (0.75 g, 2.76 mmol) were added in
this order, and the mixture was heated at 205 2108 for 4 h. After cooling, H2O was added, the soln. extracted
with toluene, the org. phase washed with H2O, dried (Na2SO4), and evaporated, and the residue bulb-to-bulb
distilled at 130 145 8 (air bath temp/0.06 Torr): 4 (0.57 g, 80%). Colorless oil. 1H-NMR (CDCl3): 1.15 2.05
(m, 16 H of Q, CH2Q); 2.50 2.65 ( m, CH2CH2CH2); 2.72 2.88 (m, PhCH2); 7.10 7.35 ( m, 5 arom. H). Anal.
calc. for C18H27N: C 83.99, H 10.57, N 5.44; found: C 84.05, H 10.60, N 5.70.
Hydrochloride 4 ¥ HCl: M.p. 149 1518. Anal. calc. for C18H27N ¥ HCl ¥ 0.25 H2O: C 72.45, H 9.63, N 4.69;
found: C 72.66, H 9.82, N 4.87.
(1S,9aR)-Octahydro-a-(phenylmethyl)-2H-quinolizine-1-ethanol (28). Tosylhydrazide (ˆ4-methylben-
zenesulfonic acid hydrazide; 0.616 g, 3.24 mmol) was added to a soln. of 19 (0.447 g, 1.65 mmol) in MeOH
(34 ml), which was heated under reflux for 3.5 h. NaBH4 (0.625 g; 16.2 mmol) was added to the ice-cooled soln.,
which was subsequently refluxed for 3 h. The solvent was evaporated, the residue taken up in H2O and extracted
with Et2O, the org. phase evaporated and the residue distilled at 1778 (air-bath temp./0.06 Torr) diastereoisomer
mixture 28 (0.36 g, 80%). Oil. IR (neat): 3390 (OH), no CˆO band. 1H-NMR (CDCl3): 1.15 2.10( m, 16 H of
Q ring QCH2); 2.58 2.74 (m, 1 H, ArCH2); 2.75 2.92 (m, 1 H of ArCH2, 2 Ha near N of Q); 3.80 3.95, 4.00
4.15 (2m, CHOH); 6.40(br. s, OH, collapsing with D2O); 7.10 7.38 ( m, 5 arom. H). Anal. calc. for C18H27NO:
C 79.07, H 9.95, N 5.12; found: C 79.17, H 10.05, N 5.30.
(1R,9aR)- and (1S,9aR)-1-[(Arylthio)methyl]- or 1-{[(w-Arylalkyl)thio]methyl}-Octahydro-2H-quinoli-
zines 5 9 and 14 18, resp.: General Method. To a soln. of 4-substituted benzenethiol or 4-substituted benzene
alkanethiol (6 mmol) in abs. EtOH (3 6 ml), ground NaOH (6 mmol) was added, and the mixture was heated
to reflux under N2. After a clear soln. was obtained, w-chlorolupinane [33] or w-chloroepilupinane [37] (1.13 g,
6 mmol) was added, and the soln. was further refluxed for 6 h under N2. To obtain 18, w-bromoepilupinane [38]
was used. The mixture was diluted with H2O, acidified with 1m HCl to pH 2, and washed with Et2O to remove the
unreacted thiol. The aq. soln. was then strongly alkalinized with 4m NaOH and extracted with Et2O, the extract
evaporated, and the residue distilled at 0.04 0.06 Torr. At 80 908 (air-bath temp.), the unreacted halo
compound was removed, while the target compound distilled at 140 170 8. In the case of 15 and 17, after
removing the chloroepilupinane, the residues were crystallized from petroleum ether and dry Et2O, respectively.
(1R,9aR)-Octahydro-1-[(phenylthio)methyl]-2H-quinolizine (5): Yield 68%. M.p. 23 24.58. B.p. 139
1478/0.04 Torr. 1H-NMR (CDCl3): 1.10 2.15 ( m, 14 H, Q); 2.75 2.95 (m, 2 Ha near N of Q); 3.07 (dd,

590
Helvetica Chimica Acta Vol. 87 (2004)
J ˆ 12.64, 9.88, 1 H, QCH2SPh); 3.25 (dd, J ˆ 12.64, 4.32, 1 H QCH2SPh); 7.10 7.45 ( m, 5 arom. H). Anal. calc.
for C16H23NS: C 73.53, H 8.87, N 5.36, S 12.24; found: C 73.65, H 8.80, N 5.50, S 11.90.
(1R,9aR)-1-{[(4-Fluorophenyl)thio]methyl}octahydro-2H-quinolizine (6): Yield 40%. Oil that solidified
on standing. B.p. 148 1528/0.05 Torr. 1H-NMR (CDCl3): 1.10 2.15 ( m, 14 H, Q); 2.75 2.95 (m, 2 Ha near N of
Q); 3.03 (dd, J ˆ 12.64, 9.88, 1 H, QCH2SAr); 3.18 (dd, J ˆ 12.64, 4.32, 1 H, QCH2SAr); 6.90 7.10, 7.25 7.45
(2m, each 2 arom. H (p-subst.)). Anal. calc. for C16H22FNS: C 68.77, H 7.94, N 5.01, S 11.48; found: C 68.51,
H 7.90, N 5.03, S 11.11.
(1R,9aR)-Octahydro-1-{[(phenylmethyl)thio]methyl}-2H-quinolizine (7): Yield 70%. Oil. B.p. 145 1528/
0.05 Torr. 1H-NMR (CDCl3): 1.10 2.10 ( m, 14 H, Q); 2.55 2.85 (m, 2 Ha near N of Q, QCH2S); 3.7
(s, PhCH2S); 7.15 7.40( m, 5 arom. H). Anal. calc. for C17H25NS: C 74.14, H 9.15, N 5.09, S 11.62; found:
C 74.36, H 9.05, N 5.20, S 11.31.
(1R,9aR)-1-{{[(4-Chlorophenyl)methyl]thio}methyl}octahydro-2H-quinolizine (8): Yield 72%. M.p. 29
308. B.p. 168 1718/0.09 Torr. 1H-NMR (CDCl3): 1.05 2.00 (m, 14 H, Q); 2.50 2.83 ( m, 2 Ha near N of Q,
CH2S); 3.60( s, PhCH2S); 7.15 7.35 (m, 4 arom. H (p-subst.)). Anal. calc. for C17H24ClNS: C 65.88, H 7.81,
N 4.52, S 10.35; found: C 65.92, H 7.73, N 4.46, S 10.48.
(1R,9aR)-Octahydro-1-{[(2-phenylethyl)thio]methyl}-2H-quinolizine (9): Yield 33%. Oil. B.p. 160 167 8/
0.05 Torr. 1H-NMR (CDCl3): 1.18 2.05 (m, 14 H, Q); 2.65 2.94 (m, CH2CH2SCH2Q, 2 Ha near N of Q); 7.18
7.35 (m, 5 arom. H).
Hydrochloride 9 ¥ HCl: M.p. 113 1148. Anal. calc. for C18H27NS ¥ HCl: C 66.33, H 8.66, N 4.30, S 9.84;
found: C 66.15, H 8.74, N 4.29, S 9.51.
(1S,9aR)-Octahydro-1-[(phenylthio)methyl]-2H-quinolizine (14): Yield 94%. M.p. 59 60.58. B.p. 130
1398/0.05 Torr. 1H-NMR (CDCl3): 1.00 2.15 (m, 14 H, Q); 2.70 2.90( m, therein dd at 2.78, J ˆ 12.48, 7.63, 2 Ha
near N of Q, 1 H of QCH2S); 3.16 (dd, J ˆ 12.48, 2.95, 1 H, QCH2S); 7.10 7.40( m, 5 arom. H). Anal. calc. for
C16H23NS: C 73.53, H 8.87, N 5.36, S 12.24; found: C 73.60, H 8.90, N 5.55, S 11.80.
(1S,9aR)-1-{[(4-Fluorophenyl)thio]methyl}octahydro-2H-quinolizine (15): Yield 93%. M.p. 63 63.58.
1H-NMR (CDCl3): 0.95 2.20 (m, 14 H, Q); 2.60 2.90 ( m, therein dd at 2.71, J ˆ 12.48, 7.63, 2 Ha near N of
Q, 1 H of QCH2SAr); 3.08 (dd, J ˆ 12.48, 2.95, 1 H, QCH2SAr); 6.90 7.10, 7.25 7.45 (2 m, each 2 arom. H (p-
subst.)). Anal. calc. for C16H22FNS: C 68.77, H 7.94, N 5.01, S 11.48; found: C 68.84, H 7.96, N 5.04, S 10.98.
(1S,9aR)-Octahydro-1-{[(phenylmethyl)thio]methyl}-2H-quinolizine (16): Yield 92%. M.p. 44.5 458. B.p.
152 1618/0.04 Torr. 1H-NMR (CDCl3): 0.90 2.15 (m, 14 H, Q); 2.27 (dd, J ˆ 12.6, 7.6, 1 H, QCH2S); 2.59
(dd, J ˆ 12.6, 3.2, 1 H, QCH2S); 2.70 2.90( m, 2 Ha near N of Q); 3.68 (s, PhCH2S); 7.20 7.45 ( m, 5 arom. H).
Anal. calc. for C17H25NS: C 74.14, H 9.15, N 5.09, S 11.62; found: C 73.98, H 9.30, N 5.07, S 11.80.
(1S,9aR)-1-{{[(4-Chlorophenyl)methyl]thio}methyl}octahydro-2H-quinolizine (17): Yield 96%. M.p. 92
1
92.58. H-NMR (CDCl3): 0.90 2.15 (m, 14 H, Q); 2.27 (dd, J ˆ 12.6, 3.0, 1 H, QCH2S); 2.59 (dd, J ˆ 12.6, 7.6,
1 H, QCH2S); 2.70 2.90 ( m, 2 Ha near N of Q); 3.68 (s, ArCH2S); 7.20 7.40 ( m, 4 arom. H). Anal. calc. for
C17H24ClNS: C 65.88, H 7.81, N 4.52, S 10.35; found: C 66.21, H 7.61, N 4.55, S 10.00.
(1S,9aR)-Octahydro-1-{[(2-phenylethyl)thio]methyl}-2H-quinolizine (18): Yield 90%. M.p. 45 468. B.p.
160 165 8/0.05 Torr. 1H-NMR (CDCl3): 1.05 2.10 (m, 14 H, Q); 2.40( dd, J ˆ 12.6, 7.6, 1 H, QCH2S); 2.66 2.96
(m, 7 H, ArCH2CH2S, 1 H of QCH2S, 2 Ha near N of Q); 7.16 7.34 (m, 5 arom. H). Anal. calc. for C18H27NS:
C 74.68, H 9.40, N 4.84, S 11.08; found: C 74.50, H 9.45, N 4.95, S 10.98.
(1R,9aR)-1-{{[2-(4-Fluorophenyl)ethyl]thio}methyl}octahydro-2H-quinolizine (10). A mixture of 4-fluo-
robenzeneethanol (2 g) and 48% hydrobromic acid (18 ml) was heated under reflux for 3 h. After cooling, H2O
was added, and the soln. was extracted with Et2O. The org. phase was shaken with 10% NaHCO3 soln., then with
H2O, dried, and evaporated. The 4-fluorophenethyl bromide was distilled (1208/10Torr).
A soln. of 4-fluorophenethyl bromide (1.41 g, 7 mmol) in DMF (3 ml) was introduced into an Aldrich
pressure tube flushed with N2; freshly distilled thiolupinine [17] (1.29 g, 7 mmol) was added and the sealed tube
heated at 1408 for 18 h. DMF was evaporated, and the residue was dissolved in CH2Cl2 and treated with 0.5n
HCl to remove the unreacted thiolupinine, while the title compound×s hydrohalides remained in the org. phase.
The solvent was evaporated, the residue dissolved in 0.5n HCl, and the acid soln. washed with Et2O. The aq.
phase was alkalinized and extracted with Et2O, the extract evaporated, and the residue distilled at 1608 (air-bath
temp.)/0.01 Torr: 10 (1.13 g, 52.6%). Oil. 1H-NMR (CDCl3): 1.05 2.15 (m, 14 H, Q); 2.60 3.00 ( m, 8 H,
CH2CH2SCH2Q, 2 Ha near N of Q); 6.90 7.10, 7.10 7.25 (2 m, each 2 arom. H (p-subst)). Anal. calc. for
C18H26FNS: C 70.32, H 8.52, N 4.56, S 10.43; found: C 70.52, H 8.90, N 4.53, S 10.08.
4-Fluoro-a-{{{[(1R,9aR)-octahydro-2H-quinolizin-1-yl]methyl}thio}methyl}benzenemethanol (21). To a
soln. of S-(4-fluorophenacyl)thiolupinine [17] (0.3 g, 0.93 mmol) in EtOH/H2O 8 :2 (10ml), NaBH 4 (0.045 g)
was added and the soln. heated under reflux for 3 h. The solvent was evaporated, and the residue was taken up in

Helvetica Chimica Acta Vol. 87 (2004)
591
H2O and filtered. The collected product was dried (0.25 g) and crystallized from pentane: diastereoisomer
mixture 21 (0.16 g, 53.3%). M.p. 64 778. Anal. calc. for C18H26FNOS: C 66.84, H 8.10, N 4.33, S 9.91; found:
C 66.74, H 8.25, N 4.38, S 9.70.
REFERENCES
[1] M. Abou-Gharbia, S. Y. Ablordeppey, R. A. Glennon, Annu. Rep. Med. Chem. 1993, 28, 1.
[2] S. D. Wyrick, R. G. Booth, Drugs Future 1995, 20, 1033.
[3] B. J. Vilner, C. S. John, W. D. Bowen, Cancer Res. 1995, 55, 408.
[4] R. H. Mach, C. R. Smith, J. Al-Nabulsi, B. R. Whirrett, S. R. Childers, K. T. Wheeler, Cancer Res. 1997, 57,
156.
[5] T. Maurice, B. P. Lockhart, Prog. Neuro-Psychopharmacol. Biol. Psychiat. 1997, 21, 69.
[6] W. D. Bowen, Pharm. Acta Helv. 2000, 74, 211.
[7] K. W. Crawford, W. D. Bowen, Cancer Res. 2002, 62, 313.
[8] R. A. Glennon, S. Y. Ablordeppey, A. M. Ismaill, M. B. El-Ashmawy, J. B. Fisher, K. B. Howie, J. Med.
Chem. 1994, 37, 1214.
[9] S. Y. Ablordeppey, J. B. Fisher, R. A. Glennon, Bioorg. Med. Chem. 2000, 8, 2105.
[10] W. Quaglia, M. Giannella, A. Piergentili, M. Pigini, L. Brasili, R. Di Toro, L. Rossetti, S. Spampinato, C.
Melchiorre, J. Med. Chem. 1998, 41, 1557.
[11] F. Berardi, S. Santoro, P. Perrone, V. Tortorella, S. Govoni, L. Lucchi, J. Med. Chem. 1998, 41, 3940.
[12] A. Marrazzo, O. Prezzavento, L. Pasquinucci, F. Vittorio, G. Ronsisvalle, Farmaco 2001, 56, 181.
[13] R. H. Mach, B. Yang, L. Wu, R. J. Kuhner, B. R. Whirrett, T. West, Med. Chem. Res. 2001, 10, 339.
[14] A. Sparatore, F. Novelli, F. Sparatore, Med. Chem. Res. 2002, 11, 1.
[15] F. Novelli, F. Sparatore, Farmaco 2002, 57, 871.
[16] S. Y. Ablordeppey, J. B. Fisher, K. J. Burke Howie, R. A. Glennon, Med. Chem. Res. 1992, 2, 368.
[17] F. Novelli, F. Sparatore, Farmaco 1993, 48, 1021.
[18] A. Sparatore, F. Sparatore, Farmaco 2001, 56, 169.
[19] I. Vazzana, F. Novelli, F. Sparatore, A. Sparatore, G. Fadda, C. Manca, Farmaco 1994, 49, 105.
[20] F. Novelli, B. Tasso, F. Sparatore, Farmaco 1999, 54, 354.
[21] S. F. Mason, K. Schofield, R. J. Wells, Proc. Chem. Soc. 1963, 337.
[22] J. D. England, J. Sam, J. Heterocycl. Chem. 1966, 3, 482.
[23] D. L. Temple, J. Sam, J. Heterocycl. Chem. 1970, 7, 847.
[24] D. L. Temple, J. Sam, J. Heterocycl. Chem. 1968, 5, 441.
[25] M. Tamura, J. Kochi, Synthesis 1971, 303.
[26] M. Tamura, J. Kochi, J. Organomet. Chem. 1971, 31, 289.
[27] H. Minlon, J. Am. Chem. Soc. 1946, 68, 7487.
[28] L. Caglioti, Tetrahedron 1966, 22, 487.
[29] R. H. Mach, L. Wu, T. West, B. R. Whirrett, S. R. Childers, Life Sci. 1999, 64, PL 131.
[30] O. Prezzavento, F. Gualtieri, A. Marrazzo, M. N. Romanelli, G. Ronsisvalle, E. Teodori, Arch. Pharm.
Pharm. Med. Chem. 2002, 335, 39.
[31] S. Y. Ablordeppey, M. El-Ashmawy, J. B. Fisher, R. A. Glennon, Eur. J. Med. Chem. 1998, 33, 625.
[32] T. Goyagi, S. Goto, A. Bhardway, V. L. Dawson, P. D. Hurn, J. R. Kirsch, Stroke 2001, 32, 1613.
[33] F. Sparatore, V. Boido, P. Preziosi, E. Miele, G. De Natale, Farmaco, Ed. Sci. 1969, 24, 587.
[34] W. A. Reckhow, D. S. Tarbell, J. Am. Chem. Soc. 1952, 74, 4960.
[35] M. Ercoli, Ph. D. Thesis, University of Genoa, 1995.
[36] G. R. Clemo, J. Rudinger, J. Chem. Soc. 1951, 2714.
[37] V. Boido, A. Sparatore, Farmaco, Ed. Sci. 1982, 37, 63.
[38] V. Boido, A. Boido, C. Boido-Canu, F. Sparatore, Farmaco, Ed. Sci. 1979, 34, 673.
Received August 21, 2003