Synthesis and oxidant properties of phase 1 benzepine N-oxides of common antipsychotic drugs

Article abstract of DOI:10.1055/s-0033-1338519

There is increasing evidence that cell constituents are oxidized by widely used antipsychotic drugs but until now the underlying chemistry has remained unclear. It is well known that such drugs readily undergo N-oxidation as a first key metabolic step. To gain insight into the problem, the tertiary phase 1 N-oxides of clozapine, olanzapine, quetiapine, and zotepine were synthesized, together with the N,S-dioxides of quetiapine and zotepine. These N-oxides were then subjected to well-established chemical transformations to test their oxidant properties in group VIII transition-metal-catalyzed reactions. In the osmium tetroxide catalyzed dihydroxylation of styrene or cinnamyl alcohol and in the tetrapropylammonium perruthenate catalyzed oxidation of cinnamyl alcohol, the benzepine N-oxides could be used as replacements for the standard oxidant, N-methylmorpholine N-oxide (NMO) with varying degrees of efficiency. From a chemical point of view, clozapine N-oxide displayed a comparable oxidation power to NMO, characterizing the benzepines as oxygen carriers. Moreover, quetiapine was found to be an excellent double oxygen acceptor, undergoing initial N-oxidation and subsequent S-oxidation. It is therefore worthwhile considering whether oxidative damage to the human body might be related to the potential redox properties of common antipsychotic drugs. Georg Thieme Verlag Stuttgart ? New York.

Full text of DOI:10.1055/s-0033-1338519

PAPER
▌2875
paper
Synthesis and Oxidant Properties of Phase 1 Benzepine N-Oxides of Common
Antipsychotic Drugs
N-Oxides of Common Antipsychotic Benzepine Drugs
Jochen Körber,a Stefan Löffler,b Dieter Schollmeyer,a Udo Nubbemeyer*a
a
Institut für Organische Chemie, Johannes Gutenberg-Universität, 55128 Mainz, Germany
Fax +49(6131)3924533; E-mail: nubbemey@uni-mainz.de
b
Department of Psychiatry and Psychotherapy, Sana Klinikum Offenbach, Teaching Hospital of Goethe University, 63069 Offenbach,
Germany
Received: 28.05.2013; Accepted after revision: 20.07.2013
CN and other antipsychotics can induce oxidative stress
Abstract: There is increasing evidence that cell constituents are ox-
and injury to mammalian cells by increasing levels of ox-
idized by widely used antipsychotic drugs but until now the under-
idation of lipids and proteins.3,4
lying chemistry has remained unclear. It is well known that such
drugs readily undergo N-oxidation as a first key metabolic step. To
gain insight into the problem, the tertiary phase 1 N-oxides of clo-
zapine, olanzapine, quetiapine, and zotepine were synthesized, to-
ine N4-oxide (CNO),5 in the designer receptors exclusive-
gether with the N,S-dioxides of quetiapine and zotepine. These N-
oxides were then subjected to well-established chemical transfor-
Recently, a further central role has been found for CN, in
the form of its second major phase 1 intermediate clozap-
ly activated by designer drugs (DREADD) technique. In
this technique, muscarinic G protein-coupled receptors
(GPCRs) are mutated so that their ability to bind to their
natural ligands is lost. This induces nanomolar potency
for the xenobiotic CNO, which otherwise displays weak
pharmacological activity. This procedure allows the cre-
ation of examinable specific synthetic pathways in vivo
with minimal invasiveness. Although GPCR drugs, gene
knockouts, and other techniques provided diverse in-
mations to test their oxidant properties in group VIII transition-
metal-catalyzed reactions. In the osmium tetroxide catalyzed dihy-
droxylation of styrene or cinnamyl alcohol and in the tetrapropyl-
ammonium perruthenate catalyzed oxidation of cinnamyl alcohol,
the benzepine N-oxides could be used as replacements for the stan-
dard oxidant, N-methylmorpholine N-oxide (NMO) with varying
degrees of efficiency. From a chemical point of view, clozapine N-
oxide displayed a comparable oxidation power to NMO, character-
izing the benzepines as oxygen carriers. Moreover, quetiapine was
found to be an excellent double oxygen acceptor, undergoing initial sights, many basic lessons can only be learned by building
N-oxidation and subsequent S-oxidation. It is therefore worthwhile
considering whether oxidative damage to the human body might be
related to the potential redox properties of common antipsychotic
ligand, because: (i) the parent compound CN has a high
drugs.
new GPCR signaling pathways in intact cells or in freely
behaving animals.6 CNO is often selected as the synthetic
affinity with muscarinic receptors and, hence, relatively
few mutations are required to render CNO a potent ago-
Key words: oxidations, catalysis, transition metals, benzepines, di-
hydroxylations, medicinal chemistry, drugs
nist; (ii) CNO appears to be highly bioavailable in rodents
and humans;5c,7 and (iii), importantly, CNO lacks appre-
ciable affinities (<1 μM) for neuroreceptors.8 Despite the
limited number of studies, contrary findings,5b,c and one
critical comment by a DREADD pioneer,6 DREADD re-
searchers continue to treat CNO as a compound that is al-
most biologically and pharmacologically inert.9
Antipsychotic drugs are used to alleviate psychotic symp-
toms. Despite their metabolic side effects, the benzepines
clozapine (CN), olanzapine (ON), quetiapine (QN), and
zotepine (ZN) represent one of the most beneficial groups
of such drugs. For example, CN is uniquely effective in
treatment-refractory schizophrenia and suicidal ideation.
The other drugs, ON, QN, and ZN, were developed as an-
alogues of CN with the aim of retaining its pharmacolog-
ical design and superior efficacy while eliminating its
potential to cause adverse reactions (for example, poten-
tially life-threatening agranulocytosis occurs in about 1%
of patients). Although these aims were only partly
achieved, the CN analogues are nevertheless useful as real
alternatives or co-medicaments in current psychopharma-
cotherapeutic practice.1
To examine this presumed inertness in the context of clo-
zapine-associated oxidative modifications of cellular key
constituents, we first synthesized the tertiary phase 1 N-
oxides of CN and its congeners ON, QN, and ZN. Next,
we compared the N-oxides with the established oxidant N-
methylmorpholine-N-oxide (NMO)10 in terms of their co-
oxidative effects in two well-known specific chemical ox-
idation reactions catalyzed by two transition-metal cata-
lysts: osmium tetroxide11 and tetrapropylammonium
perruthenate (TPAP).12 The resulting insight into the reac-
tivity/inertness of benzepines and their N-oxides might be
helpful as a basis for rationalizing the biotransformations
of benzepine drugs in the human body.
Oxidative damage to DNA, proteins, or lipids are forms of
cellular injury that can cause cellular dysfunction without
necessarily leading to cell death or tissue degeneration.2
SYNTHESIS 2013, 45, 2875–2887
Advanced online publication: 15.08.2013
0
0
3
9
-
7
8
8
1
1
4
3
7
-
2
1
0
X
DOI: 10.1055/s-0033-1338519; Art ID: SS-2013-T0380-OP
© Georg Thieme Verlag Stuttgart · New York

2876
J. Körber et al.
PAPER
Oxidation of Benzodiazepines and Dibenzothiepines
Me
N
Me
O
N
O
Our first investigations focused on the oxidation of the
benzepines under defined conditions (Scheme 1 and
Scheme 2).
N
N
S
HO
Cl
S
By using reaction conditions similar to those described by
VanRheenen in 1978,10 the benzepines were treated with
an excess of 35% aqueous hydrogen peroxide. Despite the
poor solubility of ON in water, its oxidation was highly
exothermic and the mixture boiled after some minutes.
QN is moderately soluble in water, and its oxidation was
accompanied by modest warming of the solution. In con-
trast, CN dissolved immediately, but its oxidation was not
accompanied by a significant change in temperature.
Workup of the respective reaction mixtures permitted the
isolation of CNO and quetiapine N19-oxide (QNO), albeit
in low yields (12% for CNO and trace amounts for QNO).
Olanzapine N-oxide (ONO) could not be isolated. In con-
trast, direct oxidation of QN gave quetiapine N19,S-diox-
ide (QNSO2) in 87% yield, showing that a stable and
isolable product is obtained under the oxidation condi-
tions used.
QN
ZN
i or ii or iii
ii or iii
0–97%
–O
N+
0–81%
–O
Me
Me
N+
O
N
O
N
S
HO
Cl
S
QNO
ZNO
i or ii or iii
ii or iii
35–87%
–O
11–92%
–O
Me
Me
N+
N+
O
N
O
N
HO
Cl
S+
S+
Me
Me
–O
–O
N
N
QNSO2
ZNSO2
N
N
N
N
Cl
Scheme 2 Oxidation of thiepine neuroleptics: quetiapine (QN), que-
tiapine N19-oxide (QNO), quetiapine N19,S-dioxide (QNSO2), zot-
epine (ZN), zotepine N20-oxide (ZNO), zotepine N,S-dioxide
(ZNSO2). Reaction conditions: i) 35% aq H2O2 (~40 equiv), H2O, 0–
23 °C, 18 h; yields: QNO, 0%; QNSO2, 87%; ii) 35% aq H2O2 (~2–3
equiv), MeOH, 23 °C, 20 h; yields: QNO, 61%; ZNO, 56%; QNSO2,
35%; ZNSO2, 11%; iii) MCPBA (1 equiv), MeOH (QN) or CHCl3
(ZN), 23 °C, 10 min.; yields: QNO, 97%; ZNO, 81%; QNSO2, 83%;
ZNSO2, 92%.
N
H
N
H
S
CN
ON
i or ii or iii
i or ii or iii
12–94%
Me
0–79%
Me
N+
O–
O–
N+
N
N
S
N
N
Cl
also occurred in the case of thiepines, and quetiapine
N19,S-dioxide (QNSO2) and zotepine N20,S-dioxide
(ZNSO2) were isolated in yields of 35% and 11%, respec-
tively. Finally, CN, the least-sensitive substrate, gave the
best results, and CNO was isolated in 79% yield
(Schemes 1 and 2).16
N
H
N
H
CNO
ONO
Scheme 1 Oxidation of diazepine neuroleptics: clozapine (CN), clo-
zapine N4-oxide (CNO), olanzapine (ON), olanzapine N4-oxide
(ONO). Reaction conditions: i) 35% aq H2O2 (~40 equiv), H2O, 0–
23 °C, 18 h; yields: CNO, 12%; ONO, 0%; ii) 35% aq H2O2 (~2–3
equiv), MeOH, 23 °C, 20 h; yields: CNO, 79%; ONO, 17%; iii)
MCPBA (1 equiv), MeOH (CN) or CHCl3 (ON), 23 °C, 10 min;
yields: CNO, 94%; ONO, 79%.
Unregulated oxidation of benzepines could be avoided by
the use of one equivalent of m-chloroperoxybenzoic acid
(MCPBA) in an appropriate solvent.17 The reaction of
ON, the most sensitive substrate, gave the desired oxide
ONO in 79% yield. The less reactive dibenzodiazepine
CN gave the corresponding oxide CNO in 94% yield.18
The oxidations of the thiepane-derived compounds gave
the corresponding N-oxides as the primary products;
QNO was isolated in 97% yield and ZNO in 81% yield.
Furthermore, treatment of these N-oxides with a second
equivalent of MCPBA gave the corresponding N,S-diox-
ides QNSO2 and ZNSO2 in 86% and 92% yield, respec-
tively. CNO and ONO decomposed on treatment with
two equivalents of MCPBA. Overall, the oxidation of
benzepines with MCPBA gave the best results in terms of
selectivity and yield. Detailed information is provided in
Table 1.
With the intention of suppressing the formation of multi-
ple products, we performed a second series of experi-
ments by using a slight excess (2–3 equivalents) of
aqueous 35% hydrogen peroxide in methanol. Again, ON,
the most reactive substrate, underwent rapid oxidative
degradation and the desired N-oxide ONO was isolated in
only 17% yield.13 Thiepane-substituted QN and ZN gave
the corresponding N-oxides QNO and ZNO in 61% and
56% yield, respectively.14,15 Furthermore, S-oxidation
Synthesis 2013, 45, 2875–2887
© Georg Thieme Verlag Stuttgart · New York

PAPER
N-Oxides of Common Antipsychotic Benzepine Drugs
2877
Table 1 Synthesis of N-Oxides and N,S-dioxides
Entry
Reactant
Product
aq H2O2 (~40 equiv) aq H2O2 (~2–3 equiv), MeOH MCPBA (1 equiv)
Yield (%)
Yield (%)
Yield (%)
1
2
3
CN
ON
QN
CNO
ONO
12
–
79
17
61
35
–
94
79a
97
–
QNO
–
QNSO2
87
–
QNSO2
4
5
QNO
ZN
83
81a
–
b
ZNO
–
56
11
–
b
ZNSO2
–
ZNSO2
6
ZNO
–
92
a Solvent CHCl3.
b Zotepine was not available.
The N-oxides and the N,S-dioxides were characterized by group attached to the oxygen in ZNO remained in almost
1
1
means of H, 13C, COSY, HSQC, and HMBC NMR ex- the original position found in ZN. In addition, H NMR
periments and by high-resolution mass spectrometry.
spectra of all N-oxides showed a downfield shift of the α-
CH group by about 1 ppm and of the β-CH group by about
0.3 ppm. Finally, X-ray analyses of QNO and ZNO un-
equivocally confirmed their structures and the chemose-
lectivity of the N-oxidation reactions (Figure 1 and Figure
2).19
Oxidation at the 4-position of the piperazine in CNO and
ONO led to significant changes in the 13C NMR shifts at
the adjacent carbon centers. Both CNO and ONO showed
a downfield shift of the methyl group by about 14 ppm,
and the α-ring CH2 groups underwent a downfield shift of
nearly 10 ppm. In contrast, the β-ring CH2 groups were As expected, the thiepine N,S-dioxides showed additional
shifted upfield by about 4 ppm in CNO and 5 ppm in characteristic 13C shifts in the tricyclic core fragment. Fur-
ONO. The signals of the other carbon centers of the bicy- thermore, the NMR spectra of QNSO2 showed a double
clic cores showed little, if any, change. The N-oxides of set of peaks as a result of the presence of two slowly inter-
the thiepines QN and ZN showed a related shift pattern. converting conformers.20 In QNSO2 (both conformers)
The α-side chain CH2 group of QNO showed a downfield and ZNSO2, the aromatic α-carbon atoms directly bound
shift of nearly 13 ppm, and the signal of the α-ring CH2 to the S-oxide moiety showed downfield shifts of about 7
group moved about 12 ppm downfield. Again, the β-ring ppm. All the β-carbon atoms showed upfield shifts. The
CH2 groups were shifted upfield by about 4 ppm in the
side chain and by about 3 to 6.5 ppm in the ring. In con-
trast, ZNO was characterized by α-carbon downfield
shifts only (CH3: 16 ppm, CH2: 13.5 ppm); the β-CH2
Figure 1 X-ray crystal structure of QNO (2c)
Figure 2 X-ray crystal structure of ZNO (2d)
Synthesis 2013, 45, 2875–2887
© Georg Thieme Verlag Stuttgart · New York

2878
J. Körber et al.
PAPER
R
O–
N+
Me O–
N+
Me
N+
Me
N+
Me
O–
Me
O–
N
Het
Me
O
Het
N-methyl-
piperazine
N-oxide (2)
N-methyl-
morpholine
N-oxide
dimethylamine
N-oxide (2d,f)
trimethylamine
N-oxide
Het = benzodiazepinyl or
dibenzothiazepinyl
Het = dibenzothiepinyl
R = Me, HO(CH2)2O(CH2)2
Figure 4 Well-known chemical reagents N-methylmorpholine N-
oxide (NMO) and trimethylamine N-oxide and structurally and func-
tionally related benzodiazepine N-oxides
epine N-oxides QNO and ZNO underwent a more
complex oxygen transfer. In addition to the expected
amines QN and ZN, the corresponding S-oxides QSO21
and ZSO, as well as smaller amounts of the dioxides
QNSO2 and ZNSO2, were produced in the osmium- and
ruthenium-catalyzed oxygen-transfer reactions. An over-
view of these reactions is given in Scheme 3.
Our first series of experiments focused on the use of benz-
epine N-oxides as a replacement for NMO in the osmium
tetroxide catalyzed dihydroxylation of styrene (1)
(Scheme 3 and Scheme 4).22
Figure 3 X-ray crystal structure of ZNSO2 (2f)
aromatic CH-groups in ZNSO2 and QNSO2 (conformer
1) were shifted upfield by about 11–12 ppm. The corre-
sponding C atoms in QNSO2 (conformer 2) were shifted
only 5 and 8 ppm upfield, respectively. The aromatic C at-
oms next to the thiepine side chain showed a downfield
shift of about 11 to 14 ppm in all the dioxides. Significant
differences were observed in the upfield shifts for the ar-
omatic C–N centers in the two conformers of QNSO2
(conformer 1: 7 ppm, conformer 2: 4 ppm). The aromatic
carbon atom adjacent to the enol ether carbon in ZNSO2
showed an upfield shift of about 9 ppm, whereas the enol
ether carbon (γ- position) showed a shift of about 4 ppm.
In addition, the 1H NMR spectra showed a downfield shift
of the aromatic β-(SO)-CH group by about 0.3 ppm. Fi-
nally, the structure of ZNSO2 was unequivocally con-
firmed by X-ray crystal structure analysis (Figure 3).19
All reactions were carried out on a 0.2-mmol scale of sty-
rene (1) with a 2.5-fold excess of the N-oxide 2 to give 1-
phenylethane-1,2-diol (3) and the corresponding products
4. In all cases the conversion was terminated after about
20 hours, and the resulting product mixture was analyzed
by NMR spectroscopy and by high-performance liquid
chromatography. Diol 3 was obtained in all the experi-
ments (Scheme 4; for a HPLC chromatogram, see the
Supporting Information), indicating that all the N-oxides
are active oxidants.23 The fate and behavior of the N-ox-
ides was investigated by analyzing the HPLC chromato-
grams in a qualitative fashion. The semiquantitative
details are presented in Schemes 3 and 4 and in Table 2.
In the dihydroxylation of styrene, CNO was the most ef-
ficient oxidant in terms of conversion of the olefin and re-
generation of the benzepine. In contrast, ONO showed
slow reaction rates and poor regeneration of ON, and it is
possible that some unreacted ONO remained. The thi-
epine derivatives QNO and ZNO behaved as oxidants of
medium activity. Despite the consumption of the starting
oxidant, QNO regenerated no QN. Instead, QSO (major)
and QNSO2 (minor) were the predominant products, indi-
cating that an efficient N–S oxygen shift had occurred.
The reaction of ZNO showed a less distinctive oxygen
shift. Here, regenerated ZN was the major product, and
ZNSO2 and ZSO were detected as minor products. Addi-
tionally, ZNSO2 (as a mixture containing ~13% of ZNO)
could also be used as the oxidant. The regeneration of
ZSO showed that the N-oxide moiety is the most reactive
oxygen donor. However, the minor reactant ZNO re-
mained untouched, demonstrating the superior reactivity
of the dioxide ZNSO2 in terms of its oxidation power.
Oxidant Reactivity of Benzepine N-Oxides
We surmised that the benzepine N-oxides might be useful
as oxidants, as they share features with structurally related
well-known chemical reagents such as N-methylmorpho-
line N-oxide (NMO) and trimethylamine N-oxide (Figure
4). These reagents participate in such oxidation reactions
as the osmium tetroxide (OsO4)-catalyzed dihydroxyl-
ation of olefins or allylic alcohols and the TPAP-catalyzed
oxidation of primary or secondary alcohols to give alde-
hydes and ketones, respectively. Consequently, we tested
the benzepine N-oxides as replacements for the original
N-oxide oxidants (Figure 4).11,12
In most cases, the benzodiazepine N-oxide derivatives
CNO and ONO regenerated the corresponding amines
CN and ON, respectively, and therefore displayed a sim-
ilar reactivity to that of NMO. In contrast, the dibenzothi-
Synthesis 2013, 45, 2875–2887
© Georg Thieme Verlag Stuttgart · New York

PAPER
N-Oxides of Common Antipsychotic Benzepine Drugs
2879
Me
N+
Me
O–
N+
Me
Me
N
O–
N
N
N
N
N
N
N
N
N
Cl
Cl
i or ii or iii
i or ii or iii
N
H
N
H
N
H
N
H
S
S
CN (4a)
ONO (2b)
ON (4b)
CNO (2a)
O–
N+
O–
N+
N
N
O
O
O
O
N
N
N
N
N
N
N
i or ii or iii
N
HO
HO
HO
HO
+
+
S+
–O
S+
–O
S
S
QNSO2 (2e)
QNO (2c)
QN (4c)
QSO (4e)
O–
O–
Me
N+
Me
Me
Me
N+
Me
O
Me
O
Me
N
N
Me
O
O
i or ii or iii
+
+
Cl
Cl
Cl
Cl
S+
O–
S+
O–
S
S
ZNSO2 (2f)
ZNO (2d)
ZN (4d)
ZSO (4f)
Scheme 3 Product patterns obtained by using benzepine N-oxides as oxidants in osmium- and ruthenium-catalyzed oxygen-transfer reactions.
Reaction conditions: i) styrene (1; 1 equiv), OsO4 (0.2 mol%), N-oxide (2.5 equiv), t-BuOH–H2O; ii) cinnamyl alcohol (5; 1 equiv), OsO4 (0.2
mol%), N-oxide (2.5 equiv), t-BuOH–H2O; iii) cinnamyl alcohol (5; 1 equiv), (Pr4N)RuO4 (5 mol%), N-oxide (1.5 equiv), MS, CH2Cl2.
Having derived a qualitative reactivity chart, we repeated
HO
OH
Me
N+
R2
Me
N
the sequence in a more quantitative manner using cin-
namyl alcohol [5; (2E)-3-phenylprop-2-en-1-ol] as the
starting material.24 All reactions were carried out on a 0.2-
mmol scale of alcohol 5 with a 2.5-fold excess of the ap-
propriate N-oxide 2 to give 1-phenylpropane-1,2,3-triol
(6) and the corresponding products 4. In all cases, the con-
version was terminated after about 20 hours and the out-
come of the reaction was analyzed by preparative HPLC.
In each experiment, triol 6, benzepine 4, and unreacted
benzepine N-oxide 2 were separated from the side prod-
ucts; details are presented in Scheme 3, Scheme 5 and Table 3.
R1
O–
R1
OsO4 (0.2 mol%)
+
+
t-BuOH–H2O (1:1)
R2
2.5 equiv
1
2a–d
2e,f
3
4a–d
4e,f
Scheme 4 OsO4/benzepine N-oxide oxidation of styrene (1); for
products 4, see Scheme 3. The thiepine N-oxides underwent partial S-
oxidation to give the N,S-dioxides 2e and 2f; these gave the corre-
sponding S-oxides 4e and 4f on oxygen transfer.
Table 2 Oxidation of Styrene (1) by Osmium Tetroxide and an N-Oxide
Entrya
N-Oxide
Yield (%) of 3
Yield (%) of
Yield (%) of
Other products
2
benzepine 4
1
2
CNO
ONO
8
4
36
53
13%
54%
21%
3
4
QNO
ZNO
6
4
21
50
–
9% QNSO2
62% QSO
28
7% ZNSO2
5% ZSO
a
ZNSO2
5
6
25
–
38
–
13% ZNO
6b
NMO
65
–
a An 87:13 mixture of ZNSO2 and ZNO was used as the starting material. Obviously, the reactant ZNO remained unchanged in this experiment.
b The NMO-mediated oxidation proved the usefulness of the reaction conditions.
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2013, 45, 2875–2887

2880
J. Körber et al.
PAPER
Table 3 Oxidation of Cinnamyl Alcohol (5) by Osmium Tetroxide and an N-Oxide
Entry
N-Oxide
Yield (%) of triol 6
Yielda (%) of N-oxide 2
Yielda (%) of benzepine 4 Other products
1
CNO
ONO
QNO
ZNO
NMO
73
0b
38
53
0
84
30
71
71
–
–
2
–
3
29
42
78
QSO (32%)
4
32
–
–
–
5c
a Yields based on recovered starting material.
b 60% of cinnamyl alcohol (5) was recovered.
c The NMO-mediated oxidation proved the usefulness of the reaction conditions.
tions were carried out on a 0.25-mmol scale of the alcohol
5 with a 1.5-fold excess of the appropriate N-oxide 2 to
give cinnamaldehyde (7) and the corresponding products
4. In each case, the conversion was terminated after 19
hours and the outcome of the reaction was analyzed by 1H
and 13C NMR spectroscopy. All runs were characterized
by smooth conversions, and no side-products were found.
The ratio of residual cinnamyl alcohol (5), cinnamalde-
hyde (7), and the product benzepine 4 were determined by
integrating the characteristic peaks (δ = 4.27 for 5, δ =
OH
+
OH
OH
HO
Me
N+
R2
R1
O–
OsO4 (0.2 mol%)
Me
N
R1
+
t-BuOH–H2O (1:1)
R2
2.5 equiv
5
2a–d
2e
6
4a–d
4e
Scheme 5 OsO4/N-oxide oxidation of cinnamyl alcohol (5); for
products 4, see Scheme 3. QNO underwent partial S-oxidation to give
2e, which was converted into S-oxide 4e upon oxygen transfer.
1
9.67 for 7, and the characteristic signals for 4) in the H
NMR spectra. Details are given in Scheme 3, Scheme 6
and Table 4.
Overall, the results for the dihydroxylation of cinnamyl
alcohol (5) confirmed the findings from the oxidation of
styrene (1). Again, CNO showed nearly the same oxidant
quality as NMO. ONO gave only moderate amounts of
ON, indicating that its oxidation power is weak. QNO un-
derwent complete N-deoxygenation and QN (major) and
QSO (minor) were isolated as product thiazepines. In
contrast to the first series, ZNO was reduced to ZN, and
none of the S-oxide ZSO was detected.
OH
+
H
O
Me
N+
R2
R1
O–
Pr4NRuO4 (5 mol%)
CH2Cl2, MS
Me
N
R1
+
R2
2.5 equiv
5
2a–d
2e
7
4a–d
4e
In a third series of experiments, we tested the TPAP/N-ox-
ide oxidation of cinnamyl alcohol (5) to cinnamaldehyde
(7) as second type of reaction to ensure that oxidations by
benzepine N-oxides have a broad scope.25 All the reac-
Scheme 6 Pr4NRuO4/N-oxide oxidation of cinnamyl alcohol (5); for
products 4, see Scheme 3. QNO underwent partial S-oxidation to
form 2e; this gave the S-oxide 4e upon oxygen transfer.
Table 4 Oxidation of Cinnamyl Alcohol (5) by Tetrapropylammonium Perruthenate and an N-Oxide
Entry
N-Oxide
Proportion
Proportion
Proportion of
Proportion of
of alcohol 5a
of aldehyde 7a
benzepine 4a
other productsa
1
CNO
ONO
QNO
ZNO
NMO
1
50
1
90
2
–
2
1.5
1
–
3
1.7
7
1.9
14
–
1.4 QSO
4
1
–
–
5b
1
35
a Ratios based on recovered starting material.
b The NMO-mediated oxidation confirmed the usefulness of the reaction conditions.
Synthesis 2013, 45, 2875–2887
© Georg Thieme Verlag Stuttgart · New York

PAPER
N-Oxides of Common Antipsychotic Benzepine Drugs
2881
The results for the oxidation of cinnamyl alcohol (5) with (Scheme 7). Presumably, the NH center of the antiaromat-
TPAP to give cinnamaldehyde (7) confirmed the trend in ic benzodiazepine core was converted into the hydroxyl-
reactivity of the benzepine N-oxides. CNO was the most amine (CNO2), which underwent subsequent elimination
reactive reagent, giving a nearly quantitative conversion of water. Any resulting amidinium salt might have under-
of the alcohol after 19 hours. ONO was the least potent gone final hydrolysis with nucleophile addition/rearoma-
oxidant and only 40% of alcohol 5 was converted into the tization,27 ring opening, and irreversible degradation.28
corresponding aldehyde 7 after the standard reaction time. Comparison of the reaction rates of CN and ON showed
Analogously to the second series, the ZNO-supported re- that the latter undergoes more rapid degradation. It seems
action showed a nearly 90% consumption of the reactant reasonable, that the 4-chloroaniline moiety in CN de-
alcohol 5 after 19 hours, and ZN was the sole detectable creases the nucleophilicity of the NH, leading to a some-
benzepine; no S-oxide ZSO was formed. QNO displayed what slower second oxidation. In contrast, the electron-
moderate activity; presumably, the competing oxygen rich ON undergoes exothermic multiple oxidation with
transfer of oxygen from nitrogen to sulfur to form QSO immediate ring cleavage (Scheme 7).
reduced the rate of formation of the aldehyde 7. However,
the absence of any QNSO2 indicated that QNO has an ex-
cellent oxidation power with respect to the N-oxide moi-
The dibenzothiepine neuroleptics QNO and ZNO under-
go a second oxidation to give stable dioxides. Obviously,
the thiepine sulfur atom is converted into the correspond-
ety. Obviously, the formation of QNSO2 as an
ing sulfoxide, delivering the N,S-dioxides QNSO2 and
intermediate provides much more efficient N-deoxygen-
ZNSO2 (Scheme 8). Analogously to the benzodiazepine
ation in comparison with the reactant QNO.
series, the electron-rich QNO undergoes rapid S-oxida-
These highly efficient N-deoxygenation reactions permit- tion whereas the less electron-rich 4-chlorophenyl-substi-
ted the isolation and characterization of the thiepine S-ox- tuted ZNO shows slower formation of the S-oxide.
ides. In the 13C NMR spectra of QSO and ZSO, the Because of the bulk of the sulfoxide group, no elimination
carbon shifts for the piperazine moiety were nearly the of water occurs to induce further cleavage. Presumably, a
same as those found in the original QN, indicating the third oxidation is necessary to achieve final cleavage,
presence of a tertiary amine function. For ZSO, the chem-
ical shifts of the carbon centers around the dimethylamine
moiety were nearly consistent with those of ZN, indicat-
ing the presence of a nonoxygenated nitrogen center. In
contrast, the peak pattern of the thiepine core containing
an S-oxide moiety showed carbon shifts that were nearly
identical with those of QNSO2 and ZNSO2, demonstrat-
ing conservation of the sulfoxide functions.
Polonovski reaction
to desmethyl-CN
Me
N+
Me
O–
N
1st oxidn
N
N
N
N
Cl
Cl
redn
Four standard benzepine antipsychotic drugs CN, ON,
QN, and ZN were subjected to selected oxidizing condi-
tions. The use of excess of hydrogen peroxide caused al-
most complete degradation of benzepines other than QN,
which was converted into the almost-stable dioxide
QNSO2. Milder conditions gave the N-oxides CNO,
QNO, and ZNO as the major products; the thiepines QN
and ZN additionally gave a proportion of the correspond-
ing dioxides QNSO2 and ZNSO2. Again, most of the sen-
sitive ON was destroyed, and only 17% of the desired
ONO was obtained. Stepwise selective oxidations with an
equimolar ratio of the stronger oxidant, MCPBA, gave the
oxides CNO, ONO, QNO, and ZNO in 81–97% yield.
Furthermore, the thiepines QNO and ZNO could be sub-
jected to a second oxidation step to give the corresponding
dioxides QNSO2 and ZNSO2 chemoselectively. In all
cases, the exocyclic nitrogen [the alkylated nitrogen in the
4-position of the piperazines in CN, ON, and QN and the
(dimethylamino)methyl group in ZN] was the best nu-
cleophile; the initial (kinetic) attack of the oxidant oc-
curred on the lone pair of this basic amine center in
accordance with the scale of nucleophilicity proposed by
Mayr.26
N
H
N
H
CN
CNO
2nd oxidn
?
Me
N+
Me
N+
O–
O–
H+
N+
N
N
N
N
Cl
Cl
– H2O
N
HO
CNO2
cleavage/nucleophile addition
Me
N+
Me
N
O–
1st oxidn
N
N
2
N
N
redn
N
H
N
H
S
S
ON
ONO
nd oxidn
cleavage/nucleophile addition
In the case of the benzodiazepines, the second oxidation
showed a critical difference. CN and ON underwent rapid
cleavage/degradation in the presence of excess oxidant
Scheme 7 Stepwise oxidation of benzodiazepines CN and ON and
reduction of the N-oxide moieties in CNO and ONO
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2013, 45, 2875–2887

2882
J. Körber et al.
PAPER
which might take place either through formation of the and rapid oxidation, and the resulting ONO showed slow
sulfone or through attack by oxygen at the amidine deoxygenation. In contrast, the less easily oxidized clo-
(QNSO2) or the enol ether (ZNSO2) group (Scheme 8).29 zapine CN proved to be a much better oxidant, and CNO
showed an oxidant behavior similar to that of NMO in an
Analysis of the experiments in which the benzepine N-ox-
artificial chemical environment. However, simple varia-
ides were used as oxidants in osmium- and ruthenium-cat-
tion of the conditions permitted the use of CN as a reduc-
alyzed reactions gave a reasonable reactivity pattern. In
the case of the benzodiazepine N-oxides CNO and ONO,
the latter was found to be a very weak oxidant and only
tant (the presence of an oxygen donor) and of CNO as an
oxidant (in the presence of an oxygen acceptor).30
low conversion rates were observed in the osmium tetrox- ZNO can be regarded as an analogue of CNO. As a result
ide-catalyzed dihydroxylations and the ruthenium-cata- of the presence of the chloro substituent on the dibenzo-
lyzed formation of aldehyde 7. In several runs, a dark thiepine core, the sulfur atom is less electron-rich and
color formed suggesting some reduction of the metal ox- therefore less easily oxidized. Consequently, ZNO acts as
ides to the elemental metals. Overall, ON underwent easy an oxygen donor in the dihydroxylation of cinnamyl alco-
hol (5) to give triol 6 and in the oxidation of 5 to give enal
7. The oxidation of the less-reactive styrene (1) results in
the formation of a more complicated product pattern. Ob-
Polonovski reaction
to desalkyl-CN
viously, the sulfur atom of ZNO is a better oxygen accep-
tor than is the double bond of styrene. Therefore, some
–O
N+
N
O
O
dioxide ZNSO2 is formed as a result of disproportionation
of ZNO to form ZN and ZNSO2. Because ZNSO2 also
displays N-oxygen donor properties, the dioxide under-
goes N-deoxygenation to give the corresponding S-oxide
ZSO.31
1st oxidn
redn
N
N
N
S
N
S
HO
HO
QNO
QN
In the case of QNO, the more electron-rich sulfur tends to
act as a much more efficient oxygen acceptor. Conse-
quently, in all the oxidations of styrene (1) and cinnamyl
alcohol (5) using QNO, the formation of QNSO2 is the
major competing process, occurring through dispropor-
tionation of QNO into QN and QNSO2).32 Analogously to
the zotepine ZN series, the N-oxide moiety of QNSO2
also acts as an oxidant, giving the corresponding QSO ef-
ficiently through N-oxygen transfer.31,33 Because the S-
oxide QSO was obtained as a major product in all the test
oxidation reactions, QNSO2 must have oxidant qualities
that are comparable to those of NMO, CNO, and ZNO.
From a chemical point of view, QN can behave as an ex-
cellent double oxygen acceptor, undergoing a cascade of
sequential N-oxidation and S-oxidation reactions. QNO is
a medium-strength oxygen donor, as the TPAP-catalyzed
oxidation of cinnamyl alcohol (5) to cinnamaldehyde (7)
resulted in more efficient N-deoxygenation (the ratio of
N-deoxygenation to S-oxidation was 1.9:1.4). In contrast,
the corresponding dioxide QNSO2, with a less electron-
rich dibenzothiazepine S-oxide core, is a much more effi-
cient oxidant, undergoing N-deoxygenation in both the
osmium- and ruthenium-catalyzed oxidation reactions.
2
nd oxidn
–O
N
N+
O
O
N
N
oxidn
redn
N
N
HO
HO
S+
–O
S+
–O
QNSO2
QSO
3rd oxidn
Polonovski reaction
to desmethyl-ZN
cleavage
–O
N+
Me
Me
Me
N
Me
1st oxid.
reduct.
O
O
Cl
Cl
Me
Cl
S
S
ZNO
ZN
2
nd oxidn
–O
Me
N
N+
In comparing the chemical reactivity of these benzepine
drugs with reported results for their xenobiotic transfor-
mations in mammals, some analogies and differences be-
come apparent. Whereas the N-oxides CNO, ONO, and
ZNO are known to be more-or-less stable metabolites of
the parent drugs, only traces of metabolic QNO have been
reported.14 In contrast, only QSO has been described as a
predominant oxidation product of QN.34 Correlation of
these findings with the present results suggests that the S-
oxide might be formed as a result of a rapid N,S-dioxy-
genation of QN and efficient N-deoxygenation of the re-
active QNSO2.
Me
Me
O
O
oxidn
redn
Cl
S+
–O
S+
–O
ZNSO2
ZSO
3
rd oxidn
cleavage
Scheme 8 Stepwise oxidation/cleavage of benzothiepines QN and
ZN and reduction of the N-oxide moieties in QNO, ZNO, QNSO2,
and ZNSO2
Synthesis 2013, 45, 2875–2887
© Georg Thieme Verlag Stuttgart · New York

PAPER
N-Oxides of Common Antipsychotic Benzepine Drugs
2883
In accord with the reports of Henning and co-workers,35
both oxidation of CN and deoxygenation of CNO can oc-
cur: clozapine can be easily oxidized to give one major
metabolite, CNO, but retroreduction can also be a favored
process in the presence of a catalyst and an appropriate
oxygen acceptor. In principle, analogous behavior was
found for all of the antipsychotic drugs studied, despite
some differences in their reactivities. In our hands, none
of the N-oxides was a stable metabolite under chemical
conditions and we can hazard a guess that some related in-
stability of the N-oxides might also occur in the human
body. Given that (i) clozapine CN and (at least) its major
metabolites (N4-desmethylclozapine and CNO) can cross
biological membranes;35 (ii) some of the benzepine N-ox-
ides represent major catabolites15 or minor metabolic
routes in mammals;15,17,36 and (iii) transition-metal ions
are present in most compartments of the mammalian
body, the potential activities of the N-oxides CNO, ONO,
QNO, and ZNO (as well as those of QNSO2 and ZNSO2)
as oxygen donors (oxidants) might shed light on the
degradation of these compounds and extend the range of
sources of oxygen for oxygen-dependent processes in
mammalian tissues.5c,37,38
System P3 consisted of two Knauer Smartline 1000 pumps connect-
ed to a Knauer Smartline Autosampler 3950, a Phenomenex Gemini
C18 (110–5 21, 2 × 250 mm) column, a Knauer Smartline UV De-
tector 2550 (λ = 254 nm), and a Knauer Smartline RI-Detector
2300. Other conditions for preparative chromatography are noted
below.
Oxidation of the Benzepines
General Procedure A (Neat 35% Aqueous H2O2): 35% aq H2O2 (5
mL) was added to the appropriate benzepine (500 mg, 1.3–1.6
mmol), and the mixture was stirred at r.t. In the case of highly exo-
thermic reactions, the temperature rise was modulated by cooling
with ice–water after about 50 min. The mixture was stirred for 18 h,
cooled to 0 °C, and treated with Na2SO3 (5 g) to destroy excess
H2O2. After 4 h, H2O (10 mL) and MeOH (10 mL) were added, and
any remaining solid Na2SO3 was removed by filtration. The sol-
vents were removed under reduced pressure and the crude product
was purified by HPLC. The yields are reported in Table 1. HPLC
data are given below.
General Procedure B [35% Aqueous H2O2 (2–3 Equivalents) in
MeOH]: A soln of the benzepine (0.50 mmol) in MeOH (5 mL) was
cooled to 0 °C and 35% aq H2O2 (0.1 mL, 1.2 mmol) was added
slowly. The mixture was stirred at r.t. for 20 h and then the solvents
were removed under reduced pressure. The crude product was puri-
fied by HPLC. The yields are reported in Table 1. HPLC data are
given below.
General Procedure C (MCPBA in CHCl3 or MeOH): The benzepine
or corresponding N-oxide (≤8.9 mmol) was dissolved CHCl3 or
MeOH (75 mL). A soln of MCPBA (1 equiv) in CHCl3 or MeOH
(10 mL) was added dropwise and the mixture was stirred for 10 min
at r.t. Et3N (2 equiv) was added, and the mixture was stirred for a
further 10 min. The solvents were then removed under reduced
pressure and the crude product was purified by HPLC. The yields
are reported in Table 1. HPLC data are given below.
Furthermore, a careful reassessment is required of the as-
sumption that N-oxides such as CNO represent biologi-
cally inert synthetic ligands that display reliable and
selective receptor-activation properties. Multiple func-
tions, for example, those occurring after CN regeneration
and oxygen transfer, have to be excluded to maintain the
prerequisites of the DREADD technique on a firm basis.
Clozapine N4-Oxide (CNO)
Pale yellow slowly crystallizing oil; mp 199–206 °C with decompo-
sition (colour changed to red).
When necessary, reaction solvents were dried by standard proce-
dures before use. All reactions that involved moisture- or air-sensi-
tive reagents were carried out under argon. 1H NMR, 13C NMR, and
2D spectra (COSY, HSQC, HMBC, and NOESY) were recorded at
r.t. on a Bruker ARX400 or AV400 spectrometer in CDCl3 with re-
sidual CHCl3 as the internal standard or in CD3OD with residual
CH3OD as the internal standard. IR spectra were recorded with a
Jasco FT/IR-400 plus spectrometer with a single-reflection horizon-
tal ATR (ZnSe) window. FD mass spectra were recorded by using
a Finnigan MAT 95 spectrometer. High-resolution mass spectra
were recorded with a Waters Q Tof Ultima 3 Micromass spectrom-
eter. The progress of reactions was monitored by TLC on Al sheets
coated with 60 F254 silica gel (Merck or Macherey & Nagel).
Prepared from CN (508 mg, 1.55 mmol) by following General Pro-
cedure A; yield: 62 mg (0.18 mmol, 12%).
Prepared from CN (164 mg, 0.5 mmol) by following General Pro-
cedure B; yield: 126 mg (0.4 mmol, 79%).
Prepared from CN (2.00 g, 6.1 mmol) by following General Proce-
dure C; yield: 1.97 g (5.8 mmol, 94%).
Analytical HPLC: System A1; flow rate: 1 mL/min; mobile phase:
MeOH–H2O– 25% aq NH4OH (60:40:0.35); tR = 10.08 min; Sys-
tem A1; flow rate: 2 mL/min; mobile phase MeOH–H2O–25% aq
NH4OH (75:25:0.35); tR = 2.47 min.
Preparative HPLC: System P1; flow rate: 15 mL/min; mobile phase:
The analytical HPLC systems used to analyze the products were a
Knauer HPLC pump 64 connected to a Phenomenex Gemini-NX
C18 (110- 4, 6 × 250 mm) column, a Knauer UV/VIS-FILTER pho-
tometer (λ = 254 nm), and a Knauer differential refractometer (Sys-
tem A1) or a Knauer Smartline Pump 1000 connected to a Knauer
Smartline Autosampler 3950, a Phenomenex Gemini-NX C18
(110-5 4, 6 × 250 mm) column, a Knauer Smartline UV Detector
2550 (λ = 254 nm), and a Knauer Smartline RI-Detector 2300 (Sys-
tem A2). Other conditions for analytical chromatography are noted
below.
MeOH–H2O–25% aq NH4OH (75:25:0.35); tR = 5.6 min.
IR (ATR): 3262 (br w), 2944 (w), 2819 (w), 1604 (m), 1566 (m),
1461 (m), 1430 (w), 1401 (w), 11378 (w), 1303 (w), 1282 (w), 1260
(w), 1232 (m), 1215 (w), 1159 (w), 1120 (w), 1099 (w), 1083 (w),
1031 (m), 970 (m), 930 (w), 879 (w), 845 (w), 816 (m), 749 (s), 695
(w), 669 (w) cm–1.
1H NMR (400 MHz, CD3OD): δ = 3.08–3.15 (m, 2 H, H-2), 3.18 (s,
3 H, H-1), 3.48–3.59 (m, 2 H, H-2), 3.65–3.77 (m, 2 H, H-3), 3.71–
3.93 (m, 2 H, H-3), 6.81 (d, J = 8.4 Hz, 1 H, H-12), 6.86 (dd, J = 8.4,
2.3 Hz, 1 H, H-13), 6.98 (d, J = 2.3 Hz, 1 H, H-15), 7.00 (dd,
J = 7.9, 1.0 Hz, 1 H, H-9), 7.02 (dt, J = 7.5, 7.5, 1.0 Hz, 1 H, H-7),
7.27 (dd, J = 7.7, 1.3 Hz, 1 H, H-6), 7.34 (dt, J = 7.8, 7.7, 1.5 Hz, 1
H, H-8).
Various preparative HPLC Systems were used to purify the sub-
stances. System P1 consisted of a Knauer MPLC pump connected
to a Phenomenex Gemini C18 (110–5 21, 2 × 250 mm) column, a
Knauer Variable Wavelength Monitor (λ = 254 nm), and a Bischoff
RI-detector 8110. System P2 consisted of a Knauer WellChrom
Preparative Pump K-1800 connected to a Phenomenex Gemini-NX
C18 (110–5, 30 × 250mm) column, a Knauer Variable Wavelength
Monitor (λ = 254 nm), and a Knauer Differential Refractometer.
13C NMR (100 MHz, CD3OD): δ = 43.2 (br, C-3), 60.1 (C-1), 65.8
(C-2), 121.4 (C-9), 121.6 (C-12), 123.8 (C-5), 124.2 (C-7), 124.9
(C-13), 127.4 (C-15), 129.5 (C-14), 131.1 (C-6), 133.7 (C-8), 142.8
(C-16), 143.2 (C-11), 155.4 (C-10), 164.0 (C-4).
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2013, 45, 2875–2887

2884
J. Körber et al.
PAPER
HRMS (ESI): m/z [M + H]+ calcd for C18H20ClN4O: 343.1326;
found: 343.1333.
1 H, H-17), 7.23–7.34 (m, 3 H, H-9, H-10,H-11), 7.37 (dd, J = 7.7,
1.1 Hz, 1 H, H-15), 7.47 (d, J = 7.1 Hz, 1 H, H-12).
13C NMR (100 MHz, CDCl3): δ = 39.6 (br, C-6′′), 43.2 (br, C-6′),
60.4 (C-1), 63.6 (C-3), 64.1 (br, C-5′′), 65.0 (C-5′), 70.7 (C-4), 72.4
(C-2), 123.2 (C-16), 125.0 (C-18), 127.6 (C-14), 128.5 (C-9 or C-
10), 128.7 (C-9 or C-10), 129.1 (C-17), 131.1 (C-11), 132.0 (C-12
or C-15), 132.1 (C-12 or C-15), 133.3 (C-8), 139.6 (C-13), 148.2
(C-19), 159.8 (C-7).
Olanzapine N4-Oxide (ONO)
Orange solid; mp 183–187 °C with decomposition (colour changed
to violet).
Prepared from ON (501 mg, 1.6 mmol) by following General Pro-
cedure A: no ONO found.
Prepared from ON (156 mg (0.5 mmol) by following General Pro-
cedure B; yield: 28 mg (0.09 mmol, 17%).
Quetiapine N19,S-Dioxide (QNSO2)
Colourless sticky foam (no defined melting point).
Prepared from ON (1.0 g, 3.2 mmol) by following General Proce-
dure C; yield: 759 mg (2.3 mmol, 72%). In the absence of Et3N, the
corresponding 3-chlorobenzoic acid salt (ONO·HOBz′) was isolat-
ed; yield: 124 mg (0.26 mmol, 78%). For details of the purification
and analytical data for ONO-HOBz′, see the Supporting Informa-
tion.
Prepared from QN (506 mg, 1.32 mmol) by following General Pro-
cedure A; yield: 475 mg (1.14 mmol, 87%).
Prepared from QN (235 mg, 0.61 mmol) by following General Pro-
cedure B; yield: 90 mg (0.22 mmol, 35%).
Prepared from QN (461 mg, 1.15 mmol) by following General Pro-
Analytical HPLC: System A1; flow rate: 2 mL/min; mobile phase:
cedure C; yield: 399 mg (0.96 mmol, 83%).
MeOH–H2O–25% aq NH4OH (75:25:0.35); tR = 2.10 min.
Analytical HPLC: System A1; flow rate: 2 mL/min; mobile phase:
Preparative HPLC: System P1; flow rate: 12 mL/min; mobile phase:
MeOH–H2O– 25% aq NH4OH (60:40:0.35); tR = 10.0 min; System
P1; flow rate: 21 mL/min; mobile phase: MeOH–H2O (70:30);
tR = 4.0 min.
MeOH–H2O–25% aq NH4OH (75:25:0.35); tR = 1.63 min.
Preparative HPLC: System P3; flow rate: 16 mL/min; mobile phase:
MeOH–H2O– 25% aq NH4OH (60:40:0.35); tR = 5.55 min (major
conformer), 6.05 min (minor conformer); System P1; flow rate: 16
mL/min, mobile phase: MeOH–H2O–25% aq NH4OH (50:50:0.35);
tR = 11.3 min (major conformer), 12.8 min (minor conformer).
IR (ATR): 3223 (br m), 2920 (w), 2818 (w), 1590 (s), 1509 (w),
1467 (m), 1396 (m), 1285 (w), 1263 (m), 1219 (m), 1175 (w), 1137
(m), 1119 (w), 1032 (s), 981 (w), 950 (m), 927 (m), 837 (w), 817
(w), 757 (s), 674 (w) cm–1.
1H NMR (400 MHz, CD3OD): δ = 2.30 (d, J = 1.1 Hz, 3 H, H-15),
3.14 (br d, J = 11.4 Hz, 2 H, H-2eq), 3.19 (s, 3 H, H-1), 3.51 (dt,
J = 11.7, 11.7, 3.1 Hz, 2 H, H-2ax), 3.65–3.74 (m, 2 H, H-3), 3.91
(br d, J = 13.9 Hz, 2 H, H-3), 6.38 (d, J = 1.2 Hz, 1 H, H-6), 6.67–
6.72 (m, 1 H), 6.87–6.94 (m, 3 H).
13C NMR (100 MHz, CD3OD): δ = 15.3 (C-15), 42.9 (C-3), 60.1
(C-1), 66.0 (C-2), 119.0 (C-5), 120.1 (C-10), 123.6 (C-6), 125.3 (C-
12), 125.8 (C-11), 128.9 (C-13), 130.7 (C-7), 141.6 (C-14), 145.6
(C-9), 156.4 (C-8), 159.3 (C-4).
IR (ATR): 3356 (br w), 2935 (w), 2820 (w), 1592 (m), 1577 (s),
1556 (s), 1451 (m), 1420 (m), 1303 (m), 1284 (w), 1265 (m), 1244
(w), 1166 (w), 1113 (m), 1079 (s), 1028 (s), 949 (w), 922 (w), 829
(w), 767 (s), 741 (m), 713 (w), 685 (m), 668 (w) cm–1.
1H NMR (400 MHz, CDCl3): δ (major conformer) = 2.95–3.41 (m,
2 H, H-5), 3.26–3.50 (m, 2 H, H-5′), 3.30–3.42 (m, 2 H, H-4), 3.50–
3.57 (m, 2 H, H-2), 3.63–3.70 (m, 2 H, H-1), 3.99–4.13 (m, 2 H, H-
3), 6.96 (dd, J = 1.0, 7.8 Hz, 1 H, H-18), 7.14 (dt, J = 1.0, 7.6, 7.6
Hz, 1 H, H-16), 7.22 (dt, J = 1.4, 7.7, 7.7 Hz, 1 H, H-17), 7.33 (dd,
J = 0.8, 7.6 Hz, 1 H, H-10), 7.41 (dt, J = 0.8, 7.4, 7.4 Hz, 1 H, H-9),
7.56 (dd, J = 1.2, 7.6 Hz, 1 H, H-15), 7.60 (t, J = 7.6, 7.6 Hz, 1 H,
H-11), 7.80 (d, J = 7.7 Hz, 1 H, H-12); δ (minor conformer) = 2.93–
3.01 (m, H-5), 3.12–3.21 (m, H-5′), 3.26–3.30 (m, H-4), 3.27–3.38
(m, H-6), 3.42–3.52 (m, H-5), 3.45–3.50 (m, H-2), 3.54–3.64 (m, H-
5′), 3.57–3.63 (m, H-1), 3.92–3.97 (m, H-3), 3.95–4.07 (m, H-6′),
4.14–4.25 (m, H-6), 4.69–4.79 (m, H-6′), 6.92–7.00 (m, 1 H, H-16),
7.09 (dd, J = 0.8, 8.0 Hz, 1 H, H-18), 7.30 (dt, J = 1.5, 7.8, 7.8 Hz,
1 H, H-17), 7.41–7.51 (m, 2 H, H-9, H-15), 7.51–7.54 (m, 2 H, H-
10, H-11), 7.62 (d, J = 8.5 Hz, 1 H, H-12); H6 (major conformer)
occurred as a very broad 4 H integral between 3.4 and 5.0 ppm. Be-
cause of extensive overlaps with the major conformer, the integra-
tions of the minor conformer peaks are omitted.
HRMS (ESI): m/z [M + H]+ calcd for C17H21N4OS: 329.1436;
found: 329.1432.
Quetiapine N19-Oxide (QNO)
Colourless foam; mp 106–110 °C.
Prepared from QN (506 mg, 1.32 mmol) by following General Pro-
cedure A: no QNO found.
Prepared from QN (235 mg, 0.61 mmol) by following General Pro-
cedure B; yield: 149 mg (0.37 mmol, 61%).
Prepared from QN (3.40 g, 8.9 mmol) by following General Proce-
13C NMR (100 MHz, CDCl3): δ (major conformer) = 39.2 (br, C-
6′), 43.3 (br, C-6), 60.4 (C-1), 63.6 (C-3), 64.2 (C-5′), 64.9 (C-5),
70.5 (C-4), 72.5 (C-2), 119.8 (C-12), 120.1 (C-15), 121.9 (C-8),
124.3 (C-16), 124.6 (C-18), 128.0 (C-10), 130.3 (C-9), 130.3 (C-
17), 132.1 (C-11), 135.1 (C-14), 141.7 (C-19), 147.5 (C-13), 156.6
(C-7); δ (minor conformer) = 39.3 (C-6′), 44.5 (C-6), 60.6 (C-1),
63.5 (C-5′), 63.8 (C-3), 65.9 (C-5), 71.0 (C-4), 72.4 (C-2), 122.4 (C-
16), 124.3 (C-8), 124.4 (C-12), 126.9 (C-18), 127.2 (C-15), 129.5
(C-10), 130.9 (C-9), 132.1 (C-11), 132.6 (C-17), 133.0 (C-14),
144.3 (C-19), 146.9 (C-13), 157.7 (C-7).
dure C; yield: 3.45 g (8.6 mmol, 97%).
Analytical HPLC: System A1; flow rate: 2 mL/min; mobile phase:
MeOH–H2O–25% aq NH4OH (75:25:0.35); tR = 2.56 min.
Preparative HPLC: System P1; flow rate: 15 mL/min; mobile phase:
MeOH–H2O–25% aq NH4OH (75:25:0.35); tR = 6.0 min.
Recrystallization from toluene–H2O–MeOH (~40:1:1) gave
QNO·2MeOH as a colorless block; mp 114 °C. The structure of
QNO·2MeOH was confirmed by X-ray crystallography.19
IR (ATR): 3056 (w), 3000 (w), 2952 (w), 2869 (w), 1597 (m), 1577
(m), 1557 (m), 1473 (w), 1454 (w), 1397 (m, br), 1303 (m), 1261
(m), 1245 (m), 1216 (w), 1166 (w), 1123 (m, br), 1079 (m, br), 1038
(m), 947 (w), 922 (w), 832 (w), 741 (s), 687 (w), 664 (m) cm–1.
1H NMR (400 MHz, CDCl3): δ = 3.10–3.30 (m, 3 H, H-5), 3.28–
3.35 (m, 2 H, H-4), 3.45–3.58 (m, 3 H, H-2, H-5), 3.61–3.67 (m, 2
H, H-1), 3.84–3.96 (m, 1 H, H-6), 3.96–4.14 (m, 3 H, H-3, H-6),
4.39–4.85 (m, 2 H, H6), 6.88 (dt, J = 7.6, 7.6, 1.2 Hz, 1 H, H-16),
7.03 (dd, J = 7.9, 1.1 Hz, 1 H, H-18), 7.15 (dt, J = 7.6, 7.6, 1.2 Hz,
HRMS (ESI): m/z [M + Na]+ calcd for C21H25N3NaO4S: 438.1436;
found: 438.1464.
Zotepine N20-Oxide (ZNO)
Colourless solid; mp 110–115 °C.
Prepared from ZN·xHCl (1.7 g, 4.6 mmol) by following General
Procedure B; yield: 898 mg (2.6 mmol, 56%).
Prepared from ZN (1.97 g, 6.0 mmol) by following General Proce-
dure C; yield: 1.69 g (4.9 mmol, 81%).
Synthesis 2013, 45, 2875–2887
© Georg Thieme Verlag Stuttgart · New York

PAPER
N-Oxides of Common Antipsychotic Benzepine Drugs
2885
Analytical HPLC: System A1; flow rate: 2 mL/min; mobile phase:
MeOH–H2O–25% aq NH4OH (75:25:0.35); tR = 4.29 min.
for 19 h at r.t. 10% aq Na2SO3 (1 mL) was added and stirring was
continued for 1–2 h. The aqueous layer was extracted with CH2Cl2
(3 × 3 mL) and the organic layers were combined, dried (MgSO4),
and concentrated under reduced pressure.
Preparative HPLC: System P1; flow rate: 20 mL/min; mobile phase
MeOH–H2O (70:30); tR = 12.8 min.
The yields of 1-phenylethane-1,2-diol (3) obtained by using the var-
ious N-oxides 2 were compared by means of analytical HPLC. Ali-
quots were injected and the resulting peak areas were integrated to
provide qualitative information.
Recrystallization from toluene–H2O (~40:1) gave ZNO·2H2O as
colorless plates; mp 117.5 °C. The structure of ZNO·2H2O was
confirmed by X-ray crystallography.19
IR (ATR): (ZNO·2H2O): 3311 (br m), 2940 (w), 2826 (w), 1622
(m), 1577 (w), 1550 (w), 1463 (m), 1432 (w), 1396 (w), 1357 (w),
1277 (w), 1256 (w), 1226 (m), 1199 (s), 1140 (m), 1116 (s), 1099
(s), 1059 (w), 1028 (s), 956 (w), 917 (w), 889 (w), 816 (s), 754 (s),
735 (w), 715 (w), 689 (w), 658 (m) cm–1.
1H NMR (400 MHz, CDCl3): δ = 3.37 (s, 6 H, H-1), 3.78–3.81 (m,
2 H, H-2), 4.68–4.72 (m, 2 H, H-3), 6.48 (s, 1 H, H-17), 7.20 (ddd,
J = 2.6, 6.3, 7.5 Hz, 1 H, H-13), 7.22–7.27 (m, 2 H, H-14, H-15),
7.29 (dd, J = 2.4, 8.3 Hz, 1 H, H-8), 7.44 (s, 1 H, H-6), 7.46 (d,
J = 6.6 Hz, 1 H, H-9), 7.47 (dd, J = 1.2, 7.6 Hz, 1 H, H-12).
Analytical HPLC (qualitative): System A2; flow rate; 1 mL/min;
mobile phase: MeOH–H2O (50:50); System A2; flow rate: 1
mL/min; mobile phase: MeOH–H2O (70:30).
The products of the styrene reactions were determined by means of
analytical HPLC with co-injection of reference substances, and by
1H and 13C NMR of the crude product; yields were determined inte-
grating peak areas obtained from the RI detector.
Analytic HPLC (quantitative): System A1; flow rate: 2 mL/min;
mobile phase: MeOH–H2O–25% aq NH4OH (70:30:0.35).
13C NMR (100 MHz, CDCl3): δ = 60.0 (C-1), 62.4 (C-3), 69.7 (C-
2), 108.7 (C-17), 127.4 (C-6), 127.9 (C-13), 128.5 (C-14), 129.4 (C-
15), 130.1 (C-8), 132.2 (C-12), 133.3 (C-11), 133.7 (C-9), 134.3 (C-
7 or C-10), 134.5 (C-7 or C-10), 138.1 (C-16), 139.3 (C-5), 155.2
(C-4).
HRMS (ESI): m/z [M + H]+ calcd for C18H19ClNO2S: 348.0825;
found: 348.0818.
The yields from (E)-cinnamyl alcohol (5) were measured after pre-
parative HPLC.
Preparative HPLC: System P2; flow rate: 20 mL/min; mobile phase:
MeOH–H2O–25% aq NH4OH (70:30:0.35).
For details of the use of the benzepine N-oxides as oxidants, see the
Supporting Information.
Tetrapropylammonium Perruthenate-Catalyzed Oxidation of
(E)-Cinnamyl Alcohol (5) by Using N-Oxides 2 as Co-oxidants:
General Procedure E
Zotepine N20,S-Dioxide (ZNSO2)
Pale brownish solid; mp 105–110 °C.
(E)-Cinnamyl alcohol (5; 34 mg, 0.25mmol) was added to a soln of
N-oxide 2 (375 μmol) in CH2Cl2 (2 mL) containing 4-Å MS (100
mg) under argon, and the mixture was stirred for 30 min. 1 mL of a
soln of Pr4N(RuO4) (105 mg) in anhyd CH2Cl2 (10 mL) containing
4-Å MS (750 mg) was added and the mixture was stirred for 19 h at
r.t. 10% aq Na2SO3 (2 mL) was added, and the mixture was stirred
for a further 1–2 h. The aqueous layer was then extracted with
CH2Cl2 (3 × 3 mL), and the organic layers were combined, dried
(MgSO4), and concentrated under reduced pressure. The yield was
determined by analyzing the 1H NMR spectra by integration of the
characteristic peaks for cinnamyl alcohol (5) (δ = 4.3 ppm, 2 H,
CH2OH) and cinnamaldehyde (7) (δ = 9.7 ppm, 1 H, CH=O) and
one signal for the corresponding amine 4. If necessary, signals for
the residual N-oxide 2 and any S-oxide were also integrated.
Prepared from ZN·xHCl (1.7 g, 4.6 mmol) by following General
Procedure B; yield: 188 mg (0.5 mmol, 11%).
Prepared from ZNO (515 mg, 1.4 mmol) by following General Pro-
cedure C; yield: 471 mg (1.3 mmol, 93%).
Analytical HPLC: System A1; flow rate: 2 mL/min; mobile phase:
MeOH–H2O–25% aq NH4OH (75:25:0.35); tR = 1.97 min.
Preparative HPLC: System P1; flow rate: 12 mL/min; mobile phase:
MeOH–H2O (60:40); tR = 12.1 min; System P1; flow rate: 20
mL/min; mobile phase: MeOH–H2O–25% aq NH4OH (75:25:0.35);
tR = 5.2 min.
Recrystallization from toluene–H2O (~40:1) gave ZNSO2·2H2O as
colorless plates; mp 111 °C. The structure of ZNSO2·2H2O was
confirmed by X-ray crystallography.19
Quetiapine S-Oxide (QSO)
IR (ATR): 3223 (br m), 3052 (w), 2953 (w), 1615 (m), 1583 (w),
1552 (m), 1471 (m), 1458 (m), 1434 (w), 1395 (w), 1355 (m), 1260
(m), 1226 (m), 1202 (s), 1119 (s), 1095 (s), 1077 (s), 1038 (s), 955
(m), 889 (m), 825 (s), 757 (s), 734 (s), 710 (s), 657 (w) cm–1.
1H NMR (400 MHz, CDCl3): δ = 3.28 (s, 3 H, H-1), 3.36 (s, 3 H, H-
1), 3.73 (t, J = 4.5, 4.5 Hz, 2 H, H-2), 4.60–4.66 (m, 1 H, H-3),
4.72–4.78 (m, 1 H, H-3), 6.51 (s, 1 H, H-17), 7.26 (dd, J = 1.4, 7.7
Hz, 1 H, H-15), 7.30 (dt, J = 1.3, 7.4, 7.4 Hz, 1 H, H-14), 7.44 (dt,
J = 1.4, 7.5, 7.5 Hz, 1 H, H-13), 7.49 (d, J = 2.0 Hz, 1 H, H-6), 7.53
(dd, J = 2.1, 8.5 Hz, 1 H, H-8), 7.72 (dd, J = 1.1, 7.9 Hz, 1 H, H-12),
7.76 (d, J = 8.4 Hz, 1 H, H-9).
13C NMR (100 MHz, CDCl3): δ = 59.9 (C-1), 62.5 (C-3), 69.2 (C-
2), 105.2 (C-17), 120.8 (C-12), 122.2 (C-9), 126.6 (C-6), 128.5 (C-
13), 128.6 (C-16), 128.8 (C-15), 129.2 (C-5), 129.5 (C-14), 131.0
(C-8), 135.7 (C-7), 140.1 (C-11), 141.4 (C-10), 152.4 (C-4).
HRMS (ESI): m/z [M + H]+ calcd for C18H19ClNO3S: 364.0774;
found: 364.0768.
Analytical HPLC: System A1; flow rate: 2 mL/min; mobile phase:
MeOH–H2O–25% aq NH4OH (75:25:0.35); tR = 2.03 min; System
A1; flow rate: 2 mL/min, mobile phase: MeOH–H2O–25% aq
NH4OH (60:40:0.35); tR = 4.11 min; System A1; flow rate: 2
mL/min; mobile phase MeOH–H2O–25% aq NH4OH (50:50:0.35);
tR = 9.90 min.
Preparative HPLC: System P1; flow rate: 18 mL/min; mobile phase
MeOH–H2O (50:50); tR = 28.0 min.
Pale green viscous oil.
IR (ATR): 3437 (br w), 3054 (w), 2928 (w), 2858 (w), 2821 (w),
1591 (m), 1577 (s), 1554 (s), 1452 (m), 1422 (m), 1367 (w), 1352
(w), 1306 (m), 1286 (w), 1259 (m), 1243 (m), 1122 (m), 1085 (m),
1038 (m), 1014 (s), 909 (w), 831 (w), 767 (m), 739 (m), 687 (w) cm–1.
1H NMR (400 MHz, CDCl3): δ = 2.55 (br s, 1 H, H-5), 2.64 (br s, 1
H, H-5), 2.64 (t, J = 5.3, 5.3 Hz, 2 H, H-4), 3.61 (m, 2 H, H-2), 3.66
(t, J = 5.3, 5.3 Hz, 2 H, H-3), 3.70 (m, 2 H, H-1), 7.00 (dd, J = 0.8,
7.8 Hz, 1 H, H-18), 7.16 (dt, J = 1.0, 7.5, 7.5 Hz, 1 H, H-16), 7.26
(dt, J = 1.5, 7.7, 7.7 Hz, 1 H, H-17), 7.34 (dd, J = 0.8, 7.6 Hz, 1 H,
H-9), 7.44 (dt, J = 0.9, 7.5, 7.5 Hz, 1 H, H-10), 7.58 (dd, J = 1.4, 7.8
Hz, 1 H, H-15), 7.62 (dt, J = 0.9, 7.6, 7.6 Hz, 1 H, H-11), 7.84 (dd,
J = 0.7, 7.9 Hz, 1 H, H-12); remaining H5 and H6 occurred as a very
broad 6 H integral between 2.8 and 4.0 ppm.
Osmium Tetroxide-Catalyzed Dihydroxylations With N-Oxides
as Co-oxidants: General Procedure D
Styrene (1) or (E)-cinnamyl alcohol (5) (0.2 mmol) and N-oxide 2
(0.5 mmol) were dissolved or suspended in t-BuOH–H2O (1:1; 5
mL), and the mixture was stirred for 15 min at r.t. A 4% aq soln of
OsO4 (5 μL) was injected into the mixture, which was then stirred
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2013, 45, 2875–2887

2886
J. Körber et al.
PAPER
13C NMR (100 MHz, CDCl3): δ = 45.8 (br, C-6), 52.9 (br, C-5),
57.8 (C-4), 61.8 (C-1), 67.4 (C-3), 72.4 (C-2), 119.7 (C-12), 120.2
(C-15), 122.6 (C-8), 123.8 (C-16), 124.8 (C-18), 128.0 (C-9), 130.0
(C-10), 130.3 (C-17), 131.8 (C-11), 135.5 (C-14), 142.4 (C-19),
147.7 (C-13), 157.3 (C-7).
(5) (a) Gauch, R.; Michaelis, W. Farmaco, Ed. Prat. 1971, 26,
667. (b) Pirmohamed, M.; Williams, D.; Madden, S.;
Templeton, E.; Park, B. K. J. Pharmacol. Exp. Ther. 1995,
272, 984. (c) Chang, W. H.; Lin, S. K.; Lane, H. Y.; Wei, F.
C.; Hu, W. H.; Lam, Y. W.; Jann, M. W. Prog.
HRMS (ESI): m/z [M + H]+ calcd for C21H26N3O3S: 400.1695;
found: 400.1690.
Neuropsychopharmacol. Biol. Psychiatry 1998, 22, 723.
(d) Bun, H.; Disdier, B.; Aubert, C.; Catalin, J. Fundam.
Clin. Pharmacol. 1999, 13, 577. (e) Aitchison, K. J.; Jann,
M. W.; Zhao, J. H.; Sakai, T.; Zaher, H.; Wolff, K.; Collier,
D. A.; Kerwin, R. W.; Gonzalez, F. J. J. Psychopharmacol.
(London U. K.) 2000, 14, 353. (f) Heiser, P.; Schüler-
Springorum, M.; Schulte, E.; Hausmann, C.; Remschmidt,
H.; Krieg, J.-C.; Vedder, H. Eur. J. Pharmacol. 2003, 476,
167.
Zotepine S-Oxide (ZSO)
Analytical HPLC: System A1; flow rate: 2 mL/min; mobile phase:
MeOH–H2O–25% aq NH4OH (75:25:0.35); tR = 4.68 min.
Preparative HPLC: System P1; flow rate: 21 mL/min; mobile phase:
MeOH–H2O–25% aq NH4OH (75:25:0.35); tR = 10.22 min.
Colourless solid; mp 128–130 °C (very slow decomposition, colour
changed to brown).
(6) Conklin, B. R. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 4777.
(7) (a) Bender, D.; Holschbach, M.; Stöcklin, G. Nucl. Med.
Biol. 1994, 21, 921. (b) Ray, R. S.; Corcoran, A. E.; Brust,
R. D.; Kim, J. C.; Richerson, G. B.; Nattie, E.; Dymecki, S.
M. Science (Washington D.C.) 2011, 333, 637. For a
comment on (b), see: (c) Löffler, S.; Körber, J.;
IR (ATR): 3054 (w), 2976 (w), 2946 (w), 2821 (m), 2771 (m), 1613
(m), 1583 (w), 1552 (m), 1470 (m), 1400 (w), 1353 (m), 1261 (m),
1227 (m), 1202 (m), 1141 (m), 1119 (s), 1096 (m), 1078 (s), 1039
(s), 992 (w), 963 (w), 914 (w), 875 (w), 827 (s), 756 (m), 735 (m),
710 (m), 688 (w) cm–1.
1H NMR (400 MHz, CDCl3): δ = 7.81 (d, J = 8.5 Hz, 1 H, H-9),
7.78 (dd, J = 1.0, 8.0 Hz, 1 H, H-12), 7.66 (d, J = 2.1 Hz, 1 H, H-6),
7.58 (dd, J = 2.1, 8.5 Hz, 1 H, H-8), 7.49 (dt, J = 1.1, 7.6, 7.6 Hz, 1
H, H-13), 7.35 (dt, J = 1.3, 7.4, 7.5 Hz, 1 H, H-14), 7.27 (d, J = 7.4
Hz, 1 H, H-15), 6.41 (s, 1 H, H-17), 4.18 (t, J = 5.7, J = 5.7 Hz, 2 H,
H-3), 2.93–2.78 (m, 2 H, H-2), 2.40 (s, 6 H, H-1).
13C NMR (100 MHz, CDCl3): δ = 153.7 (C-4), 141.3 (C-10), 140.4
(C-11), 135.8 (C-7), 131.0 (C-8), 129.8 (C-5), 129.4 (C-14), 129.2
(C-16), 128.6 (C-15), 128.3 (C-13), 127.3 (C-6), 122.1 (C-9), 121.0
(C-12), 104.4 (C-17), 66.7 (C-3), 57.9 (C-2), 45.9 (C-1).
HRMS (ESI): m/z [M + H]+ calcd for C18H19ClNO2S: 348.0825;
found: 348.0836.
Nubbemeyer, U.; Fehsel, K. Science (Washington D.C.)
2012, 337, 646. For the response to the comment, see:
(d) Dymecki, S. M.; Ray, R. R.; Brust, R. D.; Corcoran, A.
E.; Kim, J. C.; Richerson, G. B.; Nattie, E. Science
(Washington D.C.) 2012, 337, 646.
(8) Armbruster, B. N.; Li, X.; Pausch, M. H.; Herlitze, S.; Roth,
B. L. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 5163.
(9) Some authors have even mentioned the presumed inertness
of CNO in the title of a publication: see ref. 8.
(10) (a) VanRheenen, V.; Cha, D. Y.; Hartley, W. M. Org. Synth.
Coll. Vol. VI; Wiley: London, 1988. (b) Chappuis, G.;
Tamm, C. Helv. Chim. Acta 1982, 65, 521.
(11) (a) Jacobsen, E. N.; Markó, I.; Mungall, W. S.; Schröder, G.;
Sharpless, K. B. J. Am. Chem. Soc. 1988, 110, 1968.
(b) Wai, J. S. M.; Marko, L.; Svendsen, J. S.; Finn, M. G.;
Jacobsen, E. N.; Sharpless, K. B. J. Am. Chem. Soc. 1989,
111, 1123; see also ref. 10.
Acknowledgment
(12) (a) Griffith, W. P.; Ley, S. V.; Whitcombe, G. P.; White, A.
D. J. Chem. Soc., Chem. Commun. 1987, 1625. (b) Ley, S.
V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis
1994, 639.
(13) For a synthesis of ONO, see: Thatipalli, P.; Kumar, R.;
Bulusu, C.; Chakka, R.; Padi, P. R.; Yerra, A.; Bollikonda,
S. ARKIVOC 2008, (xi), 195.
We are grateful to Christa Sick, Erik Sick, and Peter Sick for their
financial support of this work.
Supporting Information for this article is available online at
http://www.thieme-connect.com/ejournals/toc/synthesis.SnuIomprifgtooatrinSugpIoi
n
nfomartit
(14) For a synthesis of QNO under neutral conditions, see:
Kumar, S. K.; Reddy, K. T.; Kanth, A. V.; Omprakash, G.;
Dubey, P. K. Org. Chem.: Indian J. 2012, 8, 164.
(15) For a synthesis of ZNO, see: Shiraga, T.; Kaneko, H.;
Iwasaki, K.; Tozuka, Z.; Suzuki, A.; Hata, T. Xenobiotica
1999, 29, 217.
(16) For a synthesis of CNO, see: Garipelli, N.; Reddy, B. M.;
Jithan, A. V. Int. J. Pharm. Sci. Nanotechnol. 2010, 2, 762.
(17) Calligaro, D. O.; Fairhurst, J.; Hotten, T. M.; Moore, N. A.;
Tupper, D. E. Bioorg. Med. Chem. Lett. 1997, 7, 25.
(18) The oxidation of CNO and ONO with MCPBA at 50%
conversion gave unknown polar oxidation products, mostly
through decomposition.
References
(1) (a) Pirmohamed, P.; Park, K. CNS Drugs 1997, 2, 139.
(b) Müller, W. E. In Therapie Psychischer Erkrankungen;
Möller, H.-J., Ed.; 3rd ed.; Thieme: Stuttgart, 2006, 182.
(2) Nishioka, N.; Arnold, S. E. Am. J. Geriatr. Psychiatry 2004,
12, 167.
(3) (a) Reinke, A.; Martins, M. R.; Lima, M. S.; Moreira, J. C.;
Dal-Pizzol, F.; Quevedo, J. Neurosci. Lett. 2004, 372, 157.
(b) Walss-Bass, C.; Weintraub, S. T.; Hatch, J.; Mintz, J.;
Chaudhuri, A. R. Int. J. Neuropsychopharmacol. 2008, 11,
1097. (c) Baig, M. R.; Navaira, E.; Escamilla, M.; Raventos,
H.; Walss-Bass, C. J. Psychiatr. Pract. 2010, 16, 325.
(4) Moreover, CN induced up-regulation of the proapoptotic
gene bax α and the production of an excess of superoxide
anion in circulating granulocytes from all treated patients,
whereas neither septicemia-induced respiratory bursts in
neutrophils nor ON induced such effects; see: (a) Loeffler,
S.; Fehsel, K.; Henning, U.; Fischer, J.; Agelink, M.; Kolb-
Bachofen, V.; Klimke, A. Pharmacopsychiatry 2003, 36, 37.
(b) Fehsel, K.; Loeffler, S.; Krieger, K.; Henning, U.;
Agelink, M.; Kolb-Bachofen, V.; Klimke, A. J. Clin.
Psychopharmacol. 2005, 25, 419.
(19) Crystallographic data for compounds QNO, ZNO, and
ZNSO2 have been deposited with the accession numbers
CCDC 907672, 907670, and 907671, respectively, and can
be obtained free of charge from the Cambridge
Crystallographic Data Centre, 12 Union Road, Cambridge
CB2 1EZ, UK; Fax: +44(1223)336033; E-mail:
deposit@ccdc.cam.ac.uk; Web site:
www.ccdc.cam.ac.uk/conts/retrieving.html.
(20) If we consider the exocyclic C–N bond between the C atom
of the thiazepine and N atom of the piperazine as a partial
Synthesis 2013, 45, 2875–2887
© Georg Thieme Verlag Stuttgart · New York

PAPER
double bond (amidine mesomerism), the N-oxide-derived
N-Oxides of Common Antipsychotic Benzepine Drugs
2887
(29) The third nucleophilic position in QN might be the
dibenzothiepine ring nitrogen center, by analogy to N-
piperazine–thiazepine moiety can form a chiral axis
(racemate). In combination with the chiral S-oxide
(racemate), this led to the presence of two slowly
interconverting diastereomeric conformers that appeared as
a double set of peaks in the spectra. The two conformers
were successfully separated by preparative HPLC. On
standing, the separated conformers underwent slow
equilibration to regenerate the original conformer mixture.
(21) For syntheses of QSO, see: (a) Warawa, E. J.; Migler, B. M.;
Ohnmacht, C. J.; Needles, A. L.; Gatos, G. C.; McLaren, F.
M.; Nelson, C. L.; Kirkland, K. M. J. Med. Chem. 2001, 44,
372. (b) Sibinski, R. Cent. Eur. J. Chem. 2012, 10, 232.
(22) (a) Moreno-Dorado, F. J.; Guerra, F. M.; Ortega, M. J.;
Zubía, E.; Massanet, G. M. Tetrahedron: Asymmetry 2003,
14, 503. (b) Branco, L. C.; Afonso, C. A. M. J. Org. Chem.
2004, 69, 4381. (c) Kobayashi, S.; Endo, M.; Nagayama, S.
J. Am. Chem. Soc. 1999, 121, 11229. For data on diol 3, see,
for example: (d) Huang, K.; Wang, H.; Stepanenko, V.; De
Jesús, M.; Torruellas, C.; Correa, W.; Ortiz-Marciales, M.
J. Org. Chem. 2011, 76, 1883.
(23) HPLC analyses showed that diol 3 was obtained in low
yields. This might have been a consequence of the presence
of residual styrene (1) (not analyzed), the long reaction time,
or the presence of excess oxidant permitting subsequent
transformations of 3, such as glycol cleavage. Our
investigation focused on the fate and behavior of the N-
oxides.
(24) (a) By analogy to: Kang, B.; Chang, S.; Decker, S.; Britton,
R. Org. Lett. 2010, 12, 1716. For data on triol 6, see, for
example: (b) Wang, Z.; Cui, Y.-T.; Xu, Z.-B.; Qu, J. J. Org.
Chem. 2008, 73, 2270.
(25) (a) By analogy to: Ley, S. V.; Ramarao, C.; Smith, M. D.
J. Chem. Soc. Chem. Commun. 2001, 2278. For data on
cinnamaldehyde (7), see, for example: (b) Koenning, D.;
Hiller, W.; Christmann, M. Org. Lett. 2012, 14, 5258.
(26) For the nucleophilicity of piperazine, morpholine,
piperidine, and anilines, see: (a) Botzel, F.; Chu, Y. C.;
Mayr, H. J. Org. Chem. 2007, 72, 3679. For the nucleo-
philicity of N-methylmorpholine and N-methylpiperidine,
see: (b) Phan, T. B.; Brugst, M.; Mayr, H. Angew. Chem. Int.
Ed. 2006, 45, 3869; Angew. Chem. 2006, 118, 3954 .
(c) Kanzian, T.; Nigst, T. A.; Maier, A.; Pichl, S.; Mayr, H.
Eur. J. Org. Chem. 2009, 6379. (d) Ammer, J.; Baidya, M.;
Kobayashi, S.; Mayr, H. J. Phys. Org. Chem. 2010, 23, 1029.
In the presence of additional protons, the most nucleophilic
and most basic position can be blocked (protonated).
Oxidation might then be directed towards the N5 (diazepine)
and S5 (thiepine) sites, inducing alternative reaction
cascades.
alkylation; see: Noskov, V. G.; Kalinina, L. N.; Noskova, M.
N.; Kruglyak, Y. L.; Strukov, O. G.; Kurochkin, V. K.
Pharm. Chem. J. (Engl. Trans.) 1998, 32, 210.
(30) Generally, deoxygenation might proceed by demethylation,
following a competing pathway. In the presence of a metal
oxide (especially Fe2+), a Polonovski-type elimination
delivers an iminium salt, the final hydrolysis leads to the
well-known desmethylclozapine through elimination of
formaldehyde. For related demethylations, see:
(a) Schildan, A.; Schirrmacher, R.; Samochocki, M.;
Christner, C.; Maelicke, A.; Rösch, F. J. Labelled Compd.
Radiopharm. 2001, 44, 247. (b) Rosenau, T.; Potthast, A.;
Kosma, P. Tetrahedron 2002, 58, 9809. (c) Herlem, D.;
Martin, M.-T.; Thal, C.; Guillou, C. Bioorg. Med. Chem.
Lett. 2003, 13, 2389. (d) Mary, A.; Renko, D. Z.; Guillou,
C.; Thal, C. Tetrahedron Lett. 1997, 38, 5151.
(31) Because of their less electron-rich dibenzothiepine S-oxide
tricyclic cores, ZNSO2 and QNSO2 could be rated as more
powerful oxidants in relation to N-oxide deoxygenation than
the less-oxidized ZNO and QNO.
(32) Furthermore, QNSO2, as the better oxygen donor, tends to
convert QN into QNO as follows: QNSO2 + QN →
QSO + QNO. The analogous process is possible within the
ZN series: ZNSO2 + ZN → ZSO + ZNO.
(33) Two S-oxides have been described; see: DeVane, C. L.;
Nemeroff, C. B. Clin. Pharmacokinet. 2001, 40, 509.
(34) We succeeded in oxidizing QN directly to QSO in the
presence of a hemifumarate buffer. The N4-nitrogen
underwent protonation–deactivation, and the oxidation was
directed towards the still-nucleophilic sulfur atom; see:
Mittapelli, V.; Vadali, L.; Sivalakshmi, D. A.;
Suryanarayana, M. V. Rasayan J. Chem. 2010, 3, 677.
(35) (a) Henning, U.; Krieger, K.; Loeffler, S.; Klimke, A.
Pharmacopsychiatry 2003, 36, 233; abstract 116.
(b) Henning, U.; Krieger, K.; Klimke, A.
Pharmacopsychiatry 2003, 36, 233; abstract 115.
(36) Dain, J. G.; Nicoletti, J.; Ballard, F. Drug Metab. Dispos.
1997, 25, 603.
(37) (a) Meier, J. Br. J. Pharmacol. 1975, 53, 440. (b) Jann, M.
W.; Lam, Y. W. F.; Chang, W.-H. Arch. Int. Pharmacodyn.
Ther. 1994, 328, 243. (c) Pirmohamed, M.; Williams, D.;
Madden, S.; Templeton, E.; Park, B. K. J. Pharmacol. Exp.
Ther. 1995, 272, 984.
(38) CNO and the other N-oxides can act as powerful oxidants
and, hence, the parent drugs can be regarded as vehicles of
varying efficiency for atomic oxygen. In this way, the
current study might contribute to a basic understanding of
oxidative effects of the antioxidant clozapine CN, see also:
(a) Dalla Libera, A.; Scutari, G.; Boscolo, R.; Rigobello, M.
P.; Bindoli, A. Free Radical Res. 1998, 29, 151. For some
examples of reductions, see: (b) Kitamura, S.; Sugihara, K.;
Tatsumi, K. Drug Metab. Dispos. 1998, 27, 92. (c) Parte, P.;
Kupfer, D. Drug Metab. Dispos. 2005, 33, 1446.
(27) Dragovic, S.; Gunness, P.; Ingelman-Sundberg, M.;
Vermeulen, N. P. E.; Commandeur, J. N. M. Drug Metab.
Dispos. 2013, 41, 651.
(28) For oxidative metabolism of drugs, see: Poraj-Kobielska,
M.; Kinne, M.; Ullrich, R.; Scheibner, K.; Kayser, G.;
Hammel, K. E.; Hofrichter, M. Biochem. Pharmacol. 2011,
82, 789.
© Georg Thieme Verlag Stuttgart · New York
Synthesis 2013, 45, 2875–2887