Antitubercular polyhalogenated phenothiazines and phenoselenazine with reduced binding to CNS receptors

Article abstract of DOI:10.1016/j.ejmech.2020.112420

Targeting energy metabolism in Mycobacterium tuberculosis (Mtb) is a new paradigm in the search for innovative anti-TB drugs. NADH:menaquinone oxidoreductase is a non-proton translocating type II NADH dehydrogenase (NDH-2) that is an essential enzyme in the respiratory chain of Mtb and is not found in mammalian mitochondria. Phenothiazines (PTZs) represent one of the most known class of NDH-2 inhibitors, but their use as anti-TB drugs is currently limited by the wide range of potentially serious off-target effects. In this work, we designed and synthesized a series of new PTZs by decorating the scaffold in an unconventional way, introducing various halogen atoms. By replacing the sulfur atom with selenium, a dibromophenoselenazine 20 was also synthesized. Among the synthesized poly-halogenated PTZs (HPTZs), dibromo and tetrachloro derivatives 9 and 11, along with the phenoselenazine 20, emerged with a better anti-TB profile than the therapeutic thioridazine (TZ). They targeted non-replicating Mtb, were bactericidal, and synergized with rifampin and bedaquiline. Moreover, their anti-TB activity was found to be related to the NDH-2 inhibition. Most important, they showed a markedly reduced affinity to dopaminergic and serotonergic receptors respect to the TZ. From this work emerged, for the first time, as the poly-halogenation of the PTZ core, while permitting to maintain good anti-TB profile could conceivably lead to fewer CNS side-effects risk, making more tangible the use of PTZs for this alternative therapeutic application.

Full text of DOI:10.1016/j.ejmech.2020.112420

Journal Pre-proof
Antitubercular polyhalogenated phenothiazines and phenoselenazine with reduced
binding to CNS receptors
Maria Giulia Nizi, Jenny Desantis, Yoshio Nakatani, Serena Massari, Maria Angela
Mazzarella, Gauri Shetye, Stefano Sabatini, Maria Letizia Barreca, Giuseppe
Manfroni, Tommaso Felicetti, Rowena Rushton-Green, Kiel Hards, Gniewomir Latacz,
Grzegorz Satała, Andrzej J. Bojarski, Violetta Cecchetti, Michal H. Kolář, Jadwiga
Handzlik, Gregory M. Cook, Scott G. Franzblau, Oriana Tabarrini
PII:
S0223-5234(20)30391-3
DOI:
https://doi.org/10.1016/j.ejmech.2020.112420
EJMECH 112420
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 24 January 2020
Revised Date: 21 March 2020
Accepted Date: 2 May 2020
Please cite this article as: M.G. Nizi, J. Desantis, Y. Nakatani, S. Massari, M.A. Mazzarella, G.
Shetye, S. Sabatini, M.L. Barreca, G. Manfroni, T. Felicetti, R. Rushton-Green, K. Hards, G. Latacz,
G. Satała, A.J. Bojarski, V. Cecchetti, M.H. Kolář, J. Handzlik, G.M. Cook, S.G. Franzblau, O.
Tabarrini, Antitubercular polyhalogenated phenothiazines and phenoselenazine with reduced binding
to CNS receptors, European Journal of Medicinal Chemistry (2020), doi: https://doi.org/10.1016/
j.ejmech.2020.112420.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
©
2020 Published by Elsevier Masson SAS.

Antitubercular Polyhalogenated Phenothiazines and Phenoselenazine with
Reduced Binding to CNS Receptors
¶
¶,#
‡
¶
Maria Giulia Nizi, Jenny Desantis, Yoshio Nakatani, Serena Massari, Maria Angela
¶
§
¶
¶
¶
Mazzarella, Gauri Shetye, Stefano Sabatini, Maria Letizia Barreca, Giuseppe Manfroni,
¶
‡
‡
ƺ
⊥
Tommaso Felicetti, Rowena Rushton-Green, Kiel Hards, Gniewomir Latacz, Grzegorz Satała,
ƺ
¶
†
‡
⊥
Andrzej J. Bojarski, Violetta Cecchetti, Michal H. Kolář, Jadwiga Handzlik, Gregory M Cook,
§
¶,
Scott G. Franzblau, and Oriana Tabarrini. *
¶
Department of Pharmaceutical Sciences, University of Perugia, Via del Liceo 1, Perugia 06100,
Italy.
§
Institute for Tuberculosis Research, College of Pharmacy, University of Illinois at Chicago,
Chicago, Illinois, USA.
‡
Department of Microbiology and Immunology, University of Otago, Dunedin, New Zealand
†
Department of Physical Chemistry, University of Chemistry and Technology, Technicka 5, 16628
Prague, Czech Republic, ORCID: 0000-0003-0841-944X
⊥Department of Technology and Biotechnology of Drugs, Faculty of Pharmacy, Jagiellonian
University Medical College, Medyczna 9, 30-688 Kraków, Poland
ƺ
Department of Medicinal Chemistry May Institute of Pharmacology, Polish Academy of Sciences,
Smętna 12, 31-343 Kraków, Poland
#
Present address: Department of Chemistry, Biology and Biotechnology, University of Perugia,
06123 Perugia, Italy

ABSTRACT
Targeting energy metabolism in Mycobacterium tuberculosis (Mtb) is a new paradigm in the search
for innovative anti-TB drugs. NADH:menaquinone oxidoreductase is a non-proton translocating
type II NADH dehydrogenase (NDH-2) that is an essential enzyme in the respiratory chain of Mtb
and is not found in mammalian mitochondria. Phenothiazines (PTZs) represent one of the most
known class of NDH-2 inhibitors, but their use as anti-TB drugs is currently limited by the wide
range of potentially serious off-target effects.
In this work, we designed and synthesized a series of new PTZs by decorating the scaffold in an
unconventional way, introducing various halogen atoms. By replacing the sulfur atom with
selenium, a dibromophenoselenazine 20 was also synthesized. Among the synthesized poly-
halogenated PTZs (HPTZs), dibromo and tetrachloro derivatives 9 and 11, along with the
phenoselenazine 20, emerged with a better anti-TB profile than the therapeutic thioridazine (TZ).
They targeted non-replicating Mtb, were bactericidal, and synergized with rifampicin and
bedaquiline. Moreover, their anti-TB activity was found to be related to the NDH-2 inhibition. Most
important, they showed a markedly reduced affinity to dopaminergic and serotonergic receptors
respect to the TZ.
From this work emerged, for the first time, as the poly-halogenation of the PTZ core, while
permitting to maintain good anti-TB profile could conceivably lead to fewer CNS side-effects risk,
making more tangible the use of PTZs for this alternative therapeutic application.
Keywords: Phenothiazines, Tuberculosis, NDH-2, Phenoselenazine, Respiratory chain,
Halogenated-phenothiazines.

INTRODUCTION
Phenothiazines (PTZs) are well known tricyclic compounds that show antifungal, anthelmintic,
antihistaminic, antiviral, and antibacterial activities.[1] Their sole medical application is for the
treatment of psychotic disorders, with chlorpromazine (CPZ) and fluphenazine included by the
World Health Organization (WHO) among the Essential Medicines.[2] The very versatile structure
th
of PTZs makes them one of the most promiscuous lead-structures of the 20 century. However,
with appropriate structural changes, potent and selective molecules against a defined biological
target could be generated and with this aim many analogues continue to appear in the literature.[3–
5]
In recent years, a structurally specific and directional non-covalent interaction between a halogen
atom and a Lewis base or an electron-rich moiety, has gained attention of the medicinal chemistry
community. This so called halogen bond (X-bond)[6,7] has been found in biomolecular complexes
including those with pharmaceutical relevance.[8] Intriguingly, all of the X-bonded complexes
feature protein-ligand interactions.[9] A survey over the Protein Data Bank revealed that X-bonds
do not exist in nucleic acid complexes, despite the favorable electrostatic properties of DNA and
RNA.[10] It was argued that this is observed due to a low number of known nucleic-acid binders
bearing a halogen atom, and that the role of X-bond in nucleic acid complexes remains to be
clarified.
Searching for viral nucleic acid Trans-activation Response (TAR) RNA binders as innovative anti-
HIV-1 agents,[11–14] we have recently explored the possibility to increase the ability of a small
molecule to interact with (TAR) RNA through X-bond. Among the known TAR RNA binders, we
selected PTZs both for their well-known pharmacological profiles and the synthetic feasibility.
From a molecular dynamics simulations,[15] a set of mono and poly HPTZs functionalized with
chlorine, bromine and/or fluorine atoms in various positions of the scaffold emerged, that was
consequently synthesized and tested for their anti-HIV-1 activity. Since the compounds resulted
inactive, the TAR RNA research line was not pursued.

Despite a large body of literature about PTZs, only rare examples of poly HPTZs have been
reported,[16,17] prompting us to re-purpose them for an alternative therapeutic application, i.e.
tuberculosis (TB). TB is a major public health threat worldwide. In 2018 alone 1.5 million people
died from TB.[18] The discovery of new anti-TB compounds to treat and minimize the
development of multi, extensively and totally drug-resistant strains (MDR, XDR, TDR) is definitely
a priority.[19–21] Importantly, the PTZs inhibit Mycobacterium tuberculosis (Mtb) growth by
targeting the type II NADH dehydrogenase (NDH-2).[22] NDH-2 is a membrane-bound bi-
+
substrate enzyme that catalyzes the cytoplasmic oxidation of NADH to NAD , followed by the
reduction of quinone to quinol in the membrane in the presence of flavin adenine dinucleotide
(FAD) co‐factor. It is the primary entry point of electrons derived from NADH into the
mycobacterial respiratory chain thus playing a central role in Mtb for generating ATP through
oxidative phosphorylation.[23,24] This key function along with its absence in mammalian
mitochondria make it an attractive target for drug development. The entire mycobacterial
respiratory chain is an important source of antimycobacterial drug targets,[25,26] as demonstrated
by the FDA‐approval of bedaquiline (BDQ) that targets F F -ATP synthase.[27]
1
0
Various chemical compounds have been identified as NDH-2 inhibitors although none have
progressed to drug development.[28] Among the therapeutic PTZs, the most potent antituberculars
CPZ and thioridazine (TZ) inhibited purified NDH-2 with an IC50 of 70 µM and 11 µM,
respectively.[29]
Although the antipsychotic PTZs are reported to inhibit the growth of Mtb, they are definitely less
potent than clinical TB drugs, but of even greater consequence they target neurological pathways
with associated side effects. Nevertheless, their applicability to the treatment of TB is under
consideration.[30,31] TZ was used on a compassionate basis to treat XDR-TB patients curing the
majority of them.[32] In addition, CPZ synergized with rifampicin and streptomycin.[33] Thus, they
have a great potential to be used in the fight of TB, but additional efforts are required to increase
potency and mainly reduce the CNS effects.

With this aim, in the present study we profiled our first set of HPTZs (compounds 1-5, Table 1) for
the anti-TB activity. Based on the preliminary biological results, additional compounds were
designed maintaining the best halogen substitution pattern around the scaffold while studying the N-
10 alkyl side chain (compounds 6-19, Table 1). In a more drastic structural modification the sulfur
atom was replaced by a selenium giving the 3,7-dibromophenoselenazine 20 (Table 1).
For the best compounds, an in-depth characterization of the anti-TB profile and mechanism of
action was performed along with the evaluation of their affinity towards a panel of dopaminergic
and serotonergic receptors as well as metabolic stability and hepatotoxicity risk in vitro.
RESULTS AND DISCUSSION
Chemistry
The synthesis of most of the HPTZs was accomplished starting from the commercial 10H-
phenothiazine 21, which was halogenated in the planned positions under different conditions
(
Scheme 1). On the other hand, 1-fluoro-10H-phenothiazine 22[16] was prepared by thionation of 2-
fluoro-N-phenylaniline with sulfur and iodine at high temperature as previously described (Scheme
). Starting from synthone 21, the 3,7-dichloro derivative 24[17] was prepared by reaction with N-
1
chlorosuccinimide (NCS) in DMF at room temperature (rt). By using the same condition but
increasing the equivalents of NCS, the synthesis of 1,3,7-trichloro derivative 25 and 1,3,7,9-
tetrachloro derivative 26 was attempted, but unsuccessfully. Thus, derivatives 25 and 26 were
prepared through an alternative procedure. In particular, in an attempt to prepare derivative 24 by
following a synthetic procedure previously reported[34] and entailing the chlorination of 21 with 3
equivalents PCl in CHCl at reflux, derivative 25 was obtained instead of 24. By exploiting the
5
3
same chlorinating agent and increasing its amount to 5 equivalents, derivative 26 was also prepared.
The synthesis of 3,7-dibromo derivative 27[16] and 1-fluoro-3,7-dibromo derivative 28[35] was
accomplished, as previously described, by bromination of compounds 21 and 22, respectively, by
using 2.5 equivalents of Br in glacial acetic acid at rt.
2

The successive N-alkylation reaction of the commercial 3-chloro-10H-phenothiazine 23 and
synthesized halogenated synthones 24-28 with the appropriate (chloroalkyl)amine chains was
performed in the presence of NaH in dry DMF at 80 °C affording target compounds as a
sticky semisolid. In order to obtain them solids, the corresponding hydrochloride salts 1,[36] 2-15,
17 and 18 were prepared by bubbling HCl in ethereal solution of the compounds. Only for
compounds 16, obtained already as a solid, the corresponding hydrochloride salt was not prepared.
a
Scheme 1
a
Reagents and conditions: (i) S , I , from 150 °C to 210 °C; (ii) NCS, DMF, from 0 °C to rt; (iii)
8
2
PCl , CHCl , reflux; (iv) Br , glacial AcOH, rt; (v) (chloroalkyl)amino chain hydrochloride, NaH,
5
3
2
dry DMF, 80 °C; (vi) HCl, Et O, rt.
2
The synthesis of 3,7-dibromo derivative 19 was performed by a convergent approach entailing the
preparation of the main fragments 29 and 30[37] (Scheme 2). In particular, the 3,7-dibromo-10-(2-
chloroacetyl)-10H-phenothiazine 29 was synthesized by reacting synthone 27 with chloroacetyl

chloride in toluene at 80 °C. The thiadiazol-2-amino derivative 30[37] was obtained as reported in
literature by treating the 4-fluorobenzoic acid with thiosemicarbazide in the presence of POCl at 75
3
°
C. Finally, the nucleophilic reaction between 29 and 30 was performed in ethanol at 80 °C
affording target compound 19.
a
Scheme 2
a
Reagents and conditions: (i) 2-chloroacetyl chloride, toluene, 80 °C; (ii) POCl , 75 °C; (iii)
3
ethanol, 80 °C.
The synthesis of phenoselenazine derivative 20 (Scheme 3) was accomplished starting from
synthone 31,[38] which was prepared as reported in literature reacting diphenylamine at 150 °C in a
pressure vial, in the presence of selenium, selenium dioxide and iodine, using sulfolane as solvent.
Even if the yield was very low, 16%, any modification of this procedure gave the 31 even with
lower yield.
The successive bromination, N-alkylation and salification reactions, performed according to the
phenothiazine analogue 9 (Scheme 1), gave target compound 20, through intermediate 32.
a
Scheme 3

a
Reagents and conditions: (i) Se, SeO , I , sulfolane, 150 °C, pressure vial; (ii) Br , glacial AcOH,
2
2
2
rt; (iii) 2-(2-chloroethyl)-1-methylpiperidine hydrochloride, NaH, dry DMF, 80 °C; (iv) HCl, Et O,
2
rt.
Structure-(anti-Mtb)Activity Relationship of HPTZs.
All the synthesized compounds were initially tested against Mtb strain H37Rv through the
Microplate Alamar blue assay (MABA), and in parallel for the cytotoxicity toward mammalian
Vero cells, using the PTZs TZ, CPZ, along with Rifampin (RMP), Isoniazid (INH) and Linezolid
(
LIZ) as reference compounds. As shown in Table 1, the first set of HPTZs (compounds 1-5)
displayed MIC values ranging from 2.38 to 6.59 μg/mL, with the 3,7-dichloro and 3,7-dibromo
derivatives 2 and 4, respectively, which emerged as the most active compounds with MIC values in
the same order as TZ. Analogously, the cytotoxicity was in the same range as TZ and CPZ with the
dibromo derivative 4 that showed the best profile. The therapeutic PTZs as well as almost all of the
PTZs endowed with anti-TB activity are mono-substituted on the PTZ ring and the substituent is
exclusively placed at the C-2 position. On the other hand, from this small set of compounds
emerged that the presence of two halogens imparted the strongest anti-TB activity. Thus, we sought
to further investigate these compounds designing and synthesizing additional analogues. Initially,
the N,N-dimethylpropan-1-amine side chain at the N-10 position was replaced by the 2-(1-
methylpiperidin-2-yl)ethyl chain characterizing the TZ, synthesizing the analogues 6-10. The
tetrachloro phenothiazine 11 was also prepared. Comparing the MIC values of the two series of
analogues, it emerged that the presence of the 2-(1-methylpiperidin-2-yl)ethyl chain imparted a
better anti-TB activity. The dibromo and tetrachloro derivatives 9 and 11 emerged with a very
interesting MIC value of 1.50 μg/mL, with 11 that was the most selective.

Selecting the dibromo and tetrachloro PTZ scaffolds, we further enlarged the chemical space at the
N-10 position. In particular, N-propylpyrrolidine and N-propylmorpholine were inserted
synthesizing compounds 12-15. Shorter N-10 side chains were also introduced to further
functionalize the dibromo scaffold (compounds 16 -18).
From the anti-Mtb evaluation, clear SAR insights emerged. In particular, the pyrrolidine ring on the
N-10 side chain granted good activity (MICs of about 1.50 μg/mL) when coupled to both the PTZ
scaffolds (compounds 12 and 14). Also when linked to a shorter side chain (compound 16), it
conferred good activity although significantly increasing the cytotoxicity. On the other hand, the
activity of the morpholine derivatives strongly depends on the length of the side chain as well as by
the halogen substitution pattern on the PTZ scaffold. Indeed, while the dibromo derivative 15
showed a good activity coupled with the lack of cytotoxicity up to 50 μg/mL, derivatives 13 and 17
were completely inactive. The presence of a methyl-2-pyridine gave the inactive compound 18,
indicating that an aromatic group is not tolerated at this position at least when so near to the PTZ
core. Exploiting a side chain previously reported in an anti-Mtb PTZ,[39] derivative 19 was also
prepared but was inactive.
Finally, a more incisive structural modification entailed the replacement of the sulfur atom by a
selenium synthesizing the phenoselenazine 20. Although many aspects of selenium physiological
and pathological roles need to be clarified, over the last years there has been a growing interest in
the synthesis of Se-containing compounds yielding promising biological activities such as
antioxidant, anticancer, anti-inflammatory, and anti-infective.[40,41] A very good result was
obtained by this bioisosteric approach, indeed compound 20 showed an anti-TB profile that was
superimposable to that of the sulfur analogue 9. Therefore, phenoselenazine scaffold emerged as a
new chemotype that under proper functionalization could be exploitable not only in the anti-TB but
also in other therapeutic fields.

An in-depth investigation was undertaken for some selected derivatives to define the target, further
investigate the anti-TB profile, evaluate the central effects, and preliminary study the
pharmacokinetic profile.
Table 1. Structure, Antitubercular activity and Cytotoxicity of HPTZs.
MABA
MIC
IC Vero
50
a
b
Compd
R
1
R
3
R
7
R
9
cells
(
μg/mL)
(μg/mL)
1
H
H
Cl
H
F
Cl
Cl
Cl
Br
Br
Cl
Cl
Cl
Br
Br
Cl
H
Cl
Cl
Br
Br
H
-
-
-
-
-
6.59
35.76
2
3
4
5
6
7
8
9
2.51
4.46
2.38
5.70
5.48
2.40
2.67
1.50
2.65
1.55
16.73
16.75
26.03
34.67
30.78
17.90
19.20
18.26
17.70
26.94
H
H
Cl
H
F
H
H
H
H
H
Cl
Cl
Cl
Br
Br
Cl
10
11
Cl
R10
Cl
S
Cl
12
1.50
17.97
N
c
13
>
50
NT
Cl
R10 Cl
14
15
16
17
1.45
5.52
2.57
>100
17.07
>50
9.42
>100

1
1
8
9
>100
>50
>100
c
NT
20
2.35
17.57
d
TZ
2.40
4.41
17.79
26.52
e
CPZ
f
RMP
0.01
0.25
0.87
>100
g
c
INH
NT
h
c
LIZ
NT
a
Antitubercular activity of the HPTZs through Microplate Alamar Blue Assay (MABA) assay against Mtb
H37Rv strain. The Minimum Inhibitory Concentration (MIC) is the minimum concentration of the compound
b
required to complete inhibition of bacterial growth. The IC value represents the compound concentration
50
c
d
e
f
that causes a decrease of Vero cells viability of 50%. NT = not tested Thioridazine; Chlorpromazine;
Rifampin; Isoniazid; Linezolid. All compounds are hydrochloride with the exception of 16 and 19.
g
h
NDH-2 inhibitory activity. Both NADH oxidation and oxygen consumption assays using inverted
2
membrane vesicles (IMVs) from wild-type M. smegmatis mc 4517[42] were conducted to assess
inhibition activities against NDH-2 of some HPTZs variously functionalized (compounds 4, 6, 7, 9,
and 11). NADH oxidation assay showed that compounds 7, 9, and 11 were potential NDH-2
inhibitors with IC values of 6.5, 4.7, and 3.6, (μg/mL), respectively (Table 2, Figure S1), on the
5
0
same order of that reported for TZ (IC of 4.5 μg/mL).[29] Importantly, these compounds inhibited
5
0
NADH-coupled respiratory activity to ~90% at 50 µM. To validate NDH-2 as the target of our
HPTZs, they were tested additionally against IMVs containing overexpressed Mtb H37Rv NDH-2
(
TBNDH).[42] Similar inhibitory activity was observed for 9 and 11, of which 11 appeared the
most potent with an IC50 of 4.5 (μg/mL). In these assays, 2-heptyl-4-hydroxyquinoline-N-oxide
HQNO), a compound that targets the quinone-binding site of NDH-2 exhibited weak inhibitory
activity.[43] Being the NDH-2 inhibitory activity proportional to the anti-TB activity, it seems
(

emerging as the primary target responsible for the anti-TB activity of the HPTZs. However,
additional targets cannot be excluded considering that an interference with a large number of targets
essential for Mtb life is reported to be responsible for the anti-TB activity of PTZs.[3,21]

Table 2. Inhibition of NDH-2 by selected HPTZs
Wild-type
95% Confidence
TBNDH overexpressed
95% Confidence
Reduction
in oxygen
Reduction
in oxygen
consumption
(% inhibition)
IC50
2c
IC50
(μg/mL)
2c
Compd
a
b
R
a
b
R
(
μg/mL)
Interval
consumption
Interval
d
d
(
% inhibition)
e
4
6
7
9
NC
NC
NC
NC
NC
69.4 ± 9.4
47.5 ± 2.5
85.4 ± 2.9
93.7 ± 1.7
92.4 ± 1.9
18.1 ± 12.0
NC
NC
NC
11.2
4.5
NC
NC
NC
NC
NC
40.7 ± 7.2
3.6 ± 12.5
59.3 ± 2.1
84.8 ± 4.0
74.2 ± 1.8
NC
6.5
4.7
3.6
NC
7.6-5.6
5.7-3.7
4.1-3.2
NC
0.97
0.99
0.99
NC
NC
45.2-2.8
5.1–4.0
NC
0.95
0.99
NC
11
f
HQNO
NC
29.2 ± 6.7
a
b
The IC value represents the compound concentration that causes an inhibition of the 50% NDH-2 activity. Thioridazine showed IC 4.5 (μg/mL) [29]; The
50
50 =
c
d
Confidence Interval is a range referred to the IC values expressed in (μg/mL); Goodness of fits; The inhibition % was performed at concentration of 50 μM of
each compound; Not calculated; 2-Heptyl-4-hydroxyquinoline-N-oxide; All experiments were performed in triplicate.
50
e
f

Bactericidal activity. Many NDH-2 inhibitors are reported to produce bactericidal effect on
Mtb.[44,45] Thus, we assayed compounds 9 and 11 for their bactericidal activity at 4 and 7 days
through MABA assay, including rifampicin (RMP) as a reference drug. They showed minimum
bactericidal concentration (MBC) of 3.1 μg/mL at both times (Table 3), that is only two-fold higher
than the MICs and thus they can be considered to be bactericidal,[46] with the potential to prevent
the infection reactivation.
Anti-Mtb activity under low oxygen condition. Mtb is capable of establishing persistent infection
in the host, in which it adapts to depletion of available oxygen and nutrients and enters a stage of
non-replicating persistence. It is widely accepted that the physiological state of non-replicating
persistence subpopulation of mycobacteria is responsible for the antimycobacterial tolerance, as
well as the key to shortening the long treatment regimen. Unfortunately, most of the first- and
second-line drugs are inactive against the Mtb dormant species.[47] Thus, in order to fully define
the real potential of HPTZs we have also determined the anti-TB activity of compounds 9 and 11
under hypoxic conditions using low oxygen recovery assay (LORA).[48] TZ, as well as the other
antitubercular agents RMP, INH, LIZ, were tested in parallel. As shown in Table 3, the tested
compounds maintained a good activity, which is slightly lower than those measured in replicating
Mtb. Of note, both compounds are better than TZ , definitely better than INH, while comparable to
LIZ.

Table 3. Minimum bactericidal concentration (MBC)
and low oxygen recovery assay (LORA) for selected
HTPZs.
a
MABA MBC (μg/mL)
LORA MIC
(μg/mL)
Compd
b
Incubation Incubation
4
days
7 days
9
3.1
3.1
2.96
2.89
11
3.1
3.1
NT
0.06
NT
NT
d
c
TZ
NT
5.29 – 5.70
0.08
e
RMP
0.06
NT
f
INH
>128
g
LIZ
NT
2.46
a
Microplate Alamar Blue Assay (MABA) MBC was performed
b
against Mtb H37Rv after 4 and 7 days incubation; The LORA
MIC was defined as the lowest drug concentration that inhibited ≥
9
0% of the luminescence compared to the drug-free control,
determined after 10 days of anaerobic incubation and 28h of
c
d
e
aerobic incubation (recovery).
Rifampin; Isoniazid; Linezolid.
Not tested;
Thioridazine;
f
g
Synergistic activity. Finally, we verified if our HPTZs showed synergistic effect with some first
and second line anti-TB drugs, in agreement with what has been reported for the therapeutic PTZs.
[
49,50] To this end, a checkerboard growth-inhibition assay at various concentrations of compounds
and 11 was performed. As shown in Table 4, when combined with RMP and BDQ, the tested
9
compounds showed a fractional inhibitory concentration (FIC) index ranging from 0.45 to 0.78,
which being < 1 indicates a synergistic effect. The synergism with BDQ is particularly interesting
showing that the inhibition of two different enzymes both involved in the ATP synthesis, could
result in an efficient in vivo anti-TB activity. An additive effect was instead observed when the
compounds were combined with moxifloxacin, in contrast to TZ that instead synergized with this
fluoroquinolone. An antagonistic effect was finally measured for all the tested compounds when
combined with INH, LIZ and Pretomanid (PRT).

Table 4. Fractional inhibitory concentration (FIC) index measurement of selected
HTPZs in combination with anti-TB drugs
.
Expected MIC
Observed MIC
(µM)
a
FIC
(
µM)
Anti-TB MIC
drugs (µM)
9
10
TZ
9
10
TZ
9
10
TZ
b
RMP
0.01 3.03 2.82 8.26
0.72 3.35 3.15 8.58
1.80 3.90 3.68 9.12
0.19 3.12 2.90 8.35
0.22 3.12 2.92 8.36
0.05 3.04 2.84 8.28
1.45
6.27
6.73
3.04
11.38
1.37
2.21
6.03
6.22
3.13
5.93
1.41
3.32
15.27
16.55
4.01
0.48 0.78
1.87 1.92
1.72 1.69
0.97 1.08
3.65 2.03
0.45 0.50
0.40
1.78
1.82
0.48
1.89
0.19
c
INH
d
LIZ
e
MOX
f
PRT
15.79
1.58
g
BDQ
a
The FIC index was calculated as: FIC = (MIC1 combination/MIC1 alone) + (MIC2 combination / MIC 2 alone) where 1
and 2 were the two combined compounds; FIC <1 synergistic effect; FIC = 1 additive effect; FIC >1
b
c
d
e
f
g
antagonistic effect.[51,52]
Bedaquiline.
Rifampin;
Isoniazid;
Linezolid;
Moxifloxacin;
Pretomanid;
CNS side-effects risk. The potential use of therapeutic PTZs as alternative anti-TB drugs is
definitely limited by their cognitive side effects derived from the ability to recognize various
postsynaptic receptors, of which dopaminergic and serotonergic are those mainly responsible for
their pharmacological profile.[53] Based on the well-delineated SAR, the PTZ core has to be
functionalized with an electron withdrawing substituent at the C-2 position to achieve a central
activity. Having a set of anti-TB PTZs that does not fit this SAR, it was mandatory to assay their
potential CNS effects. Thus, the affinity of the most active derivatives 9 and 11 towards selected
dopaminergic (D ) and serotonergic (5-HT , 5-HT and 5-HT ) receptors was examined in the
2
2A
6
7
radioligand binding assays, using cloned receptors and four suitable radioligands (Table 5). The
new chemotype phenoselenazine 20 was also studied while TZ was included as reference D -agent
2
with significant affinities also for the serotonin receptors (K < 200 nM). In these assays, potent
i
ligands display K values lower than 10 nM, the value 200 nM > K > 1 μM indicates a mean to
i
i
weak affinity, while K > 1μM can be considered as very low activity.
i
Table 5. Affinity of selected HPTZs towards dopaminergic and
serotonergic receptors.

a
K (nM)
i
Compd
9
D2
688
639
367
8
5-HT2A
1749
1244
1762
84
5-HT6
758
387
222
71
5-HT7
2149
1003
2340
181
1
2
1
0
b
TZ
a
Affinities tested experimentally in the radioligand binding assays. The
binding affinity constant, K , expressed as the average of at least two
i
3
3
independent experiments. Radioligands used: [ H]-Raclopride (D ); [ H]-
Ketanserin (5-HT2A ), [ H]-LSD (5-HT ), [ H]-5-CT (5-HT ). Thioridazine.
2
3
3
b
6
7
The tested compounds showed rather low or very low affinities for the panel of central GPCRs,
definitely lower than that shown by TZ. In the case of D receptor, our HPTZs showed a 45-80 fold
2
reduced affinity with respect to TZ. The reduction of the 5-HT R affinity in 15-21 fold respect to
2
A
TZ, with values that give a very low risk of interaction with this GPCR, seems to be promising as
the agonistic action towards 5-HT R is responsible for dangerous CNS-side effects, i.e.
2
A
psychedelic and hallucinogenic ones. The relatively lowest decrease of the HPTZs’ affinity
compared to TZ (3-11 fold) was observed in the case of 5-HT R, however, taking into account the
6
role of 5-HT R in CNS, this result does not worsen their pharmacological profile. Although this
6
serotonin receptor subtype is almost exclusively distributed in CNS, lines of evidence hardly
describe any undesirable or toxic effects mediated via 5-HT R. Paradoxically, similar and beneficial
6
neuroprotective effects, i.e. antidepressive, anxiolytic and procognitive ones, have been confirmed
for both agonists and antagonists of 5-HT R. Thus, the moderate but statistically significant affinity
6
observed for 5-HT R, in particular for the selenocompound 20, does not disqualify these HPTZs in
6
search for new anti-TB drugs.
While from the whole anti-TB biological characterization the poly HPTZs emerged as only slightly
better than TZ, their very low central receptor affinity definitely differentiate them from the
commercial PTZs highlighting their superior anti-TB therapeutic potential.

Metabolic stability and hepatotoxicity. To further characterize the HPTZs, metabolic stability and
a toxicity risk were evaluated for compounds 9, 11 and 20, in comparison to TZ. Hepatotoxicity as
the probable toxicity risk for the HPTZs was selected on the basis of previous lines of evidence
indicating hepatotoxic effects of TZ.[54,55] To estimate the number of metabolites and probable
biotransformation ways, compounds were incubated with human liver microsomes (HLMs) for 120
min and the reaction mixture was analyzed by LC/MS, according to the previously described
protocols.[56–59] Verapamil was used as metabolically instable reference compound (Figure S2).
Analyzing the spectra, it clearly emerged as the HPTZs (Figure S4-S6) are more stable than TZ
(
Figure S3), which gave three metabolic products related to hydroxylated or S-oxidized metabolites
M1-M3, m/z = 387.25).[60] In particular, only the dibromo derivative 9 underwent to a slight
(
metabolism with the formation of one metabolite, probably obtained after debromination (M1, m/z
405.20) (Figure S4). The same reaction was not observed for the other dibromo derivative
=
phenoselenazine 20 (Figure S6) in which the presence of a selenium atom seems to improve the
stability.
For the same compounds, the hepatotoxicity was studied evaluating the viability of hepatoma
HepG2 cell line after 48 h of incubation. Doxorubicin and the toxin carbonyl cyanide 3-
chlorophenyl-hydrazone (CCCP, 10 µM) were used as comparators. As shown in Figure 1, either
TZ or HPTZs displayed lower hepatotoxic effects than those of DOX, being non-toxic at the
concentration of 1 μM. At 10 µM concentration, compound 9 showed a certain toxicity analogously
to TZ, while phenoselenazine 20 was less toxic. Tetrachloroderivative 11 was the least hepatotoxic
causing toxic effect only at the concentration of 50 μM.

Figure 1. The effect of doxorubicin (DX, 1 µM), mitochondrial toxin carbonyl cyanide 3-
chlorophenyl-hydrazone (CCCP, 10 µM) and tested compounds (9, 11, and 20 and TZ) on
hepatoma HepG2 cell line viability after 48 h of incubation.
CONCLUSIONS
PTZs show broad spectrum of activities including the anti-TB property, but the impossibility to
separate the central effects from the other pharmacological activities relegated them to the treatment
of psychosis.
In this work, by initially re-purposing a small set of poly HPTZs, we designed and synthesized
additional analogues that led to identify compounds showing anti-TB properties superimposable or
slightly better than those shown by TZ. Most importantly, they had a markedly lower affinity to
CNS receptors and thus potentially decreased central effects. The PTZ poly-halogenation is a great
structural novelty within this class of compounds that, independently from the biological activity,
usually bear only one substituent at the C-2 position. The identification of the anti-TB
phenoselenazine 20, which is also devoid of central receptors affinity, adds another important
insight to the PTZ SAR.
The best compounds, such as dibromo and tetrachloro PTZs 9 and 11, showed MIC values against
Mtb strain H37Rv in the low micromolar range coupled with positive SI values. They were

bactericidal and showed good activity also against non-replicating Mtb, with the potential to prevent
the infection reactivation and shortening the long treatment regimen. In addition, they synergized
with RMP and BDQ, were metabolically stable, and showed hepatotoxicity only at highest
concentrations. Of note, HPTZs inhibited both oxygen consumption and NADH oxidation, two
essential energetic processes in the growth and survival of mycobacteria, indicating NDH-2 as their
putative target. The inhibition of elements of the oxidative phosphorylation pathway and the entire
energy metabolism in mycobacteria has recently emerged as an innovative approach to treat
dormant or latent mycobacteria, as shown by the approval of BDQ. In this context, the inhibition of
NDH-2 could represent an alternative valid way to interrupt the mycobacterial respiratory chain.
Unfortunately, the unavailability of TB structural data of NDH-2 did not permit us to elucidate the
exact role of X-bonds in HPTZs mechanism of action, that could contribute to further optimize
these compounds.
In conclusion, we have shown as poly-halogenation of the PTZs is a valid approach to switch off
the central receptor affinity. However, to actually define their anti-TB applicability, HPTZs with
higher selectivity index should be achieved while ruling out the side-effects characterizing all the
clinical PTZs.
EXPERIMETAL SECTION
General Chemistry. All reactions were routinely checked by TLC on silica gel 60F254 (Merck) and
visualized by using UV or iodine. Flash column chromatography (FCC) separations were carried
out on Merck silica gel 60 (mesh 230-400). The purity of all final compounds was over 95% as
determined by HPLC analysis monitored at 254 nm. HPLC analysis were performed on HPLC
Waters LC Module I Plus, equipped with XTerra MS C18 column reversed-phase (3.5 μ spherical
hybrid, 4.6 mm x 150 mm, 3.5 μm particle size) at 25 °C and were carried out at 25 °C according to
1
the method described below. Compounds structures were confirmed by H NMR (200 or 400MHz)
1
3
and C NMR (101 MHz) spectra recorded on Bruker Avance DPX-200 or DPX-400 spectrometers

by using chloroform (δ = 7.26) or dimethylsulfoxide (δ = 2.48) as an internal standard. Chemical
shifts are given in ppm (δ) and the spectral data are consistent with the assigned structures. Data are
reported as: chemical shift (multiplicity, coupling constants where applicable, number of hydrogen
atoms). Abbreviations are: s (singlet), d (doublet), t (triplet), q (quartet), qui (quintuplet) dd (doublet
of doublet), dt (double of triplet), tt (triplet of triplet), m (multiplet), br (broad signal). Coupling
constant (J) quoted in Hertz (Hz) to the nearest 0.1 Hz. After extraction, organic solutions were
dried over anhydrous Na SO , filtered, and concentrated with Büchi rotary evaporator at reduced
2
4
pressure. Yields are of purified product and were not optimized.
Reagents and solvents were purchased from common commercial suppliers and were used without
further purification. All starting materials were commercially available unless otherwise indicated.
3,7-Dichloro-10H-phenothiazine (24).[17] NCS (0.737 g, 5.52 mmol) was added portionwise to a
solution of 21 (0.500 g, 2.50 mmol) in DMF (5.00 mL) maintained at 0 °C. The reaction mixture
was stirred for 2 h at this temperature and then let to stir at rt overnight. Then, the reaction mixture
was poured into ice/water and extracted with Et O. The organic layers were evaporated to dryness
2
affording a residue, which was purified by FCC (Cyclohexane/EtOAc 80:20) to give 24 as green
1
solid (0.2 g, 31% yield). H NMR (200 MHz, DMSO-d ): δ 6.64 (d, J = 9.1 Hz, 2H, aromatic CH),
6
7.05–7.01 (m, 4H, aromatic CH), 8.85 (s, 1H, NH).
1
,3,7-Trichloro-10H-phenothiazine (25). PCl (3.13 g, 15.06 mmol) was added to a solution of 21
5
(
1.00 g, 5.02 mmol) in CHCl (40 mL) and the reaction mixture was refluxed overnight. The CHCl3
3
was eliminated by distillation under vacuum and the crude product was poured into ice/water and
stirred for 20 min. The precipitate obtained was filtered, washed with EtOH and purified by FCC
1
(
cyclohexane/ EtOAc 98:2) to give 25 as a teal green solid (1.51 g, 99% yield). H NMR (200 MHz,
CDCl ): δ 6.29 (bs, 1H, NH), 6.50 (d, J = 8.2 Hz, 1H, aromatic CH), 6.81 (dd, J = 2.2 and 0.4 Hz,
3
1
2
H, aromatic CH), 6.91 (m, 1H, aromatic CH), 6.96 (d, J = 2.3 Hz, 1H, aromatic CH), 7.00 (d, J =
.2 Hz, 1H, aromatic CH).

1,3,7,9-Tetrachloro-10H-phenothiazine (26). The title compound was synthesized according to
the procedure reported for the synthesis of 25 but by using five equivalents of PCl . After filtration,
5
compound 26 was obtained in 36% yield as a grey solid and it was used without further purification.
1
H NMR (200 MHz, CDCl ): δ 6.89 (d, J = 2.2 Hz, 2H, aromatic CH), 7.14 (d, J = 2.2 Hz, 2H,
3
aromatic CH), 11.20 (bs, 1H, NH).
General procedure for the alkylation reaction and salification (Method A). A solution of the
appropriate phenothiazine or phenoselenazine (1.0 equiv) in dry DMF was slowly added at 0 °C to a
suspension of NaH (5.0 equiv) in dry DMF. The reaction mixture was stirred at rt for 30 min. Then,
a solution of the appropriate (chloroalkyl)amine chain (2.0 equiv) in dry DMF was added dropwise
and the reaction mixture was maintained at 80 °C until no starting material was detected by TLC.
After cooling, the reaction mixture was partitioned between EtOAc and water and extracted with
EtOAc. The organic layers were evaporated to dryness obtaining an oil that was purified by FCC as
described below. With the exception of compound 15, each compound was dissolved in Et O and
2
HCl gas was bubbled into the solution. The mixture was then filtered or evaporated to dryness
depending on whether the hydrochloride salt was obtained as a solid or as an oil, respectively,
furnishing target compounds.
3-(3-Chloro-10H-phenothiazin-10-yl)-N,N-dimethylpropan-1-amine hydrochloride (1).[36]
The title compound was prepared starting from 23 following Method A (2h) and by using 3-chloro-
N,N-dimethylpropan-1-amine hydrochloride. The compound was purified by FCC (CHCl /MeOH
3
from 98:2 to 95:5) and obtained as a brown oil in 59% yield; after salification, 1 was obtained by
1
filtration as a white solid in 66% yield. H NMR (200 MHz, CDCl ): δ 2.25-2.38 (m, 2H,
3
CH CH CH ), 2.61 (s, 6H, CH ), 2.97-3.08 (m, 2H, CH CH CH ), 4.01 (t, J = 5.7 Hz, 2H,
2
2
2
3
2
2
2
CH CH CH ), 6.82-6.95 (m, 4H, aromatic CH), 6.98-7.06 (m, 1H, aromatic CH), 7.12-7.21 (m, 2H,
2
2
2
1
3
aromatic CH), 12.49 (bs, 1H, NH). C NMR (101 MHz, CDCl ): 21.66, 42.99, 44.33, 55.77,
3
1
9
16.37, 116.46, 123.21, 123.79, 124.64, 125.91, 127.92, 128.36, 133.71, 143.66, 146.02. Purity:
9.54% (HPLC, R :2.479 min; Method: solvent A, CH CN; solvent B, FA solution (0.1% v/v),
t
3

+
gradient from 60% to 20% of FA solution; 15 min). HRMS-ESI: m/z (M + H) calcd for
C H ClN S 319.10357, found 319.1043.
17
19
2
3-(3,7-Dichloro-10H-phenothiazin-10-yl)-N,N-dimethylpropan-1-amine hydrochloride (2). The
title compound was prepared starting from 24[17] following Method A (1h) and by using 3-chloro-
N,N-dimethylpropan-1-amine hydrochloride. The compound was purified by FCC (CHCl /MeOH
3
95:5) and obtained as a blue oil in 47% yield; after salification, the solution was evaporated to
1
dryness, triturated with EtOAc, and filtered to give 2 as a light-yellow solid in 18% yield. H NMR
(
400 MHz, CDCl ): δ 2.30-2.47 (m, 2H, CH CH CH ), 2.66 (s, 6H, CH ), 2.85-3.04 (m, 2H,
3 2 2 2 3
CH CH CH ), 4.04-4.12 (m, 2H, CH CH CH ), 6.80 (d, J = 7.6 Hz, 2H, aromatic CH), 7.15-7.18
2
2
2
2
2
2
1
3
(
m, 4H, aromatic CH), 12.59 (bs, 1H, NH). C NMR (101 MHz, CDCl ): 21.84, 43.15, 44.56,
3
5
5.86, 116.89, 127.19, 127.48, 127.80, 128.64, 142.97. Purity: 98.95% (HPLC, R :2.848 min;
t
Method: solvent A, CH CN; solvent B, FA solution (0.1% v/v), gradient from 80% to 20% of FA
3
+
solution; 12 min). HRMS-ESI: m/z (M + H) calcd for C H Cl N S 353.06460, found 353.0652.
1
7
18
2
2
N,N-Dimethyl-3-(1,3,7-trichloro-10H-phenthiazin-10-yl)-propan-1-amine hydrochloride (3).
The title compound was prepared starting from 25 following Method A (2h) and by using 3-chloro-
N,N-dimethylpropan-1-amine hydrochloride. The compound was purified by FCC (CHCl /MeOH
3
97:3) and obtained as a brown oil in 52% yield; after salification, the solution was evaporated to
1
dryness, triturated with EtOAc, and filtered to give 3 as a white solid in 38% yield. H NMR (400
MHz, CDCl ): δ 2.01-2.26 (m, 2H, CH CH CH ), 2.63 (s, 6H, CH ), 2.85-3.05 (m, 2H,
3
2
2
2
3
CH CH CH ), 4.06-4.09 (m, 2H, CH CH CH ), 7.00 (d, J = 8.5 Hz, 1H, aromatic CH), 7.08 (d, J =
2
2
2
2
2
2
2.2 Hz, 1H, aromatic CH), 7.14 (d, J = 2.0 Hz, 1H, aromatic CH), 7.20 (dd, J = 2.0 and 8.3 Hz, 1H,
1
3
aromatic CH), 7.25 (d, J = 2.1 Hz, 1H, aromatic CH), 12.56 (bs, 1H, NH). C NMR (101 MHz,
CDCl ): δ 23.42, 50.65, 55.86, 122.17, 126.06, 127.03, 127.59, 128.34, 129.75, 129.87, 130.16,
3
1
30.36, 134.30, 139.35, 143.28. Purity: 98.52% (HPLC, R :4.193 min; Method: solvent A, CH CN;
t
3

solvent B, FA solution (0.1% v/v), isocratic 80% of FA solution; 12 min). HRMS-ESI: m/z (M +
+
H) calcd for C H Cl N S 387.02563, found 387.0262.
1
7
17
3
2
3-(3,7-Dibromo-10H-phenothiazin-10-yl)-N,N-dimethylpropan-1-amine hydrochloride (4).
The title compound was prepared starting from 27[16] following Method A (3h) and by using 3-
chloro-N,N-dimethylpropan-1-amine hydrochloride. The compound was purified by FCC
(
CHCl /MeOH from 99:1 to 95:5) and obtained as green oil in 44% yield; after salification, 4 was
3
1
obtained by filtration as a white solid in 49% yield. H NMR (200 MHz, CDCl ): δ 2.30-2.37 (m,
3
2
H, CH CH CH ), 2.64 (s, 6H, CH ), 2.72-3.01 (m, 2H, CH CH CH ), 4.00-4.10 (m, 2H,
2 2 2 3 2 2 2
CH CH CH ), 6.73 (d, J = 8.0 Hz, 2H, aromatic CH), 7.21-7.25 (m, 4H, aromatic CH), 10.30 (bs,
2
2
2
1
3
1
1
H, NH). C NMR (101 MHz, CDCl ): δ 21.85, 43.21, 44.53, 55.89, 115.91, 117.34, 127.55,
3
30.27, 130.71, 143.44. Purity: 97.72% (HPLC, R :3.595 min; Method: solvent A, CH CN; solvent
t
3
+
B, FA solution (0.1% v/v), isocratic 90% of FA solution; 15 min). HRMS-ESI: m/z (M + H) calcd
for C H Br N S 440.96357, found 440.9637.
1
7
18
2
2
3-(3,7-Dibromo-1-fluoro-10H-phenothiazin-10-yl)-N,N-dimethylpropan-1-amine
hydrochloride (5). The title compound was prepared starting from 28[16] following Method A (1h)
and by using 3-chloro-N,N-dimethylpropan-1-amine hydrochloride. The compound was purified by
FCC (CHCl /MeOH 95:5) and obtained as a brown oil in 54% yield; after salification, 5 was
3
1
obtained by filtration as a white solid in 57% yield. H NMR (200 MHz, DMSO-d ) δ 1.81-1.88 (m,
6
2
H, CH CH CH ), 2.57 (s, 6H, CH ), 2.94 (t, J = 7.9 Hz, 2H, CH CH CH ), 3.88-3.91 (m, 2H,
2 2 2 3 2 2 2
CH CH CH ), 7.09 (d, J = 8.3 Hz, 1H, aromatic H ), 7.29-7.30 (m, 1H, aromatic CH), 7.34-7.48
2
2
2
9
1
3
(
m, 3H, aromatic CH), 10.04 (bs, 1H, NH). C NMR (101 MHz, CDCl ) δ 22.86, 43.11, 48.23,
3
5
1
5.83, 116.18, 116.51, 119.23, 119.54, 126.10, 128.65, 130.07, 130.44, 131.09, 132.51, 144.01,
53.42. Purity: 98.87% (HPLC, R :2.675 min; Method: solvent A, CH CN; solvent B, FA solution
t
3
+
(
0.1% v/v), isocratic 60% of FA solution; 15 min). HRMS-ESI: m/z (M + H) calcd for
C H Br FN S 458.95415, found 458.9541.
17
17
2
2

3
-Chloro-10-[2-(1-methylpiperidin-2-yl)ethyl]-10H-phenothiazine hydrochloride (6). The title
compound was prepared starting from 23 following Method A (3h) and by using 2-(2-chloroethyl)-
-methylpiperidine hydrochloride. The compound was purified by FCC (CHCl /MeOH 98:2) and
1
3
obtained as an orange oil in 80% yield; after salification, 6 was obtained by filtration as a white
1
solid in 47% yield. H NMR (400 MHz, DMSO-d ) δ 1.25-1.53 (m, 1H, piperidine CH ), 1.54-1.86
6
2
(
m, 4H, piperidine CH ), 1.86-2.13 (m, 2H, CH CH ), 2.19-2.29 (m, 1H, piperidine CH ), 2.60 (s,
2 2 2 2
3
H, CH ), 2.86-3.02 (m, 1H, piperidine CH), 3.02-3.22 (m, 2H, piperidine CH ), 3.82-4.15 (m, 2H,
3
2
CH CH ), 6.96-7.00 (m, 2H, aromatic CH), 7.05-7.07 (m, 1H, aromatic CH), 7.11 (s, 1H, aromatic
2
2
13
CH), 7.15-7.29 (m, 3H, aromatic CH), 10.75 (bs, 1H, NH). C NMR (101 MHz, DMSO-d ): δ
6
2
1
2.01, 22.77, 27.99, 28.30, 40.52, 43.42, 55.44, 63.14, 116.43, 116.93, 122.88, 123.60, 123.73,
24.41, 127.86, 128.28, 128.70, 132.99, 144.09, 146.75. Purity: 99.87% (HPLC, R : 3.138 min;
t
Method: solvent A, CH CN; solvent B, FA solution (0.1% v/v), gradient from 80% to 20% of FA
3
solution; 12 min).
3,7-Dichloro-10-[2-(1-methylpiperidin-2-yl)ethyl]-10H-phenothiazine hydrochloride (7). The
title compound was prepared starting from 24[17] following Method A (5h) and by using 2-(2-
chloroethyl)-1-methylpiperidine hydrochloride. The compound was purified by FCC
(
CHCl /MeOH 98:2) and obtained as a blue oil in 68% yield; after salification, the solution was
3
evaporated to dryness, triturated with a mixture of CHCl MeOH, and filtered to give 7 as a light-
3/
1
yellow solid in 57% yield. H NMR (400 MHz, DMSO-d ): δ 1.32-1.42 (m, 1H, piperidine CH ),
6
2
1
1
3
1
5
.50-1.68 (m, 4H, piperidine CH ), 1.84-2.06 (m, 3H, piperidine CH and CH CH ), 2.20-2.29 (m,
2 2 2 2
H, piperidine CH), 2.58 (s, 3H, CH ), 2.79-2.85 and 2.94-3.16 (m, each 1H, piperidine CH ), 3.78-
3
2
.99 (m, 2H, NCH CH ), 7.05 (d, J = 8.6 Hz, 2H, aromatic CH), 7.24-7.30 (m, 4H, aromatic CH),
2
2
1
3
0.50 (bs, 1H, NH). C NMR (101 MHz, DMSO-d ): δ 21.97, 22.73, 27.72, 28.13, 43.53, 55.38,
6
9.13, 62.97, 117.82, 126.07, 127.00, 127.16, 127.99, 143.63. Purity: 95.00% (HPLC, R :2.159
t
min; Method: solvent A, CH CN; solvent B, FA solution (0.1% v/v), gradient from 80% to 20% of
3
FA solution; 12 min).

1,3,7-Trichloro-10-[2-(1-methylpiperidin-2-yl)ethyl]-10H-phenothiazine hydrochloride (8).
The title compound was prepared starting from 25 following Method A (3h) and by using 2-(2-
chloroethyl)-1-methylpiperidine hydrochloride. The compound was purified by FCC
(
CHCl /MeOH 98:2) and obtained as a brown oil in 91% yield; after salification, the solution was
3
evaporated to dryness, triturated with CHCl and filtered to give 8 as a light-yellow solid in 59%
3
1
yield. H NMR (400 MHz, DMSO-d ): δ 1.18-1.70 (m, 7H, piperidine CH and CH CH ), 1.83-2.06
6
2
2
2
(
m, 1H, piperidine CH), 2.62-2.80 and 2.80-2.97 (m, each 1H, piperidine CH ), 3.30 (s, 3H, CH ),
2 3
3
.75-4.09 (m, 2H, NCH CH ), 7.23 (d, J = 8.5 Hz, 1H aromatic CH), 7.30 (d, J = 8.5 Hz, 1H,
2 2
aromatic CH), 7.30-7.40 (m, 2H, aromatic CH), 7.50 (d, J = 1.5 Hz, 1H, aromatic CH), 8.30 (s, 1H,
1
3
NH). C NMR (101 MHz, DMSO-d ): δ 21.97, 22.72, 28.18, 29.43, 40.49, 50.48, 55.49, 62.54,
6
123.17, 126.37, 126.93, 127.18, 128.37, 128.66, 129.19, 129.57, 130.58, 134.91. Purity: 99.77%
(
HPLC, R :2.717 min; Method: solvent A, CH CN; solvent B, FA solution (0.1% v/v), isocratic
t
3
6
3
0% of FA solution; 12 min).
,7-Dibromo-10-[2-(1-methylpiperidin-2-yl)ethyl]-10H-phenothiazine hydrochloride (9). The
title compound was prepared starting from 27[16] following Method A (3h) and by using 2-(2-
chloroethyl)-1-methylpiperidine hydrochloride. The compound was purified by FCC
(
CHCl /MeOH from 98:2 to 97:3) and obtained as a green oil in 40% yield; after salification, 9 was
3
1
obtained by filtration as a white solid in 44% yield. H NMR (400 MHz, DMSO-d
₆
): δ 1.19-1.94
(
m, 7H, piperidine CH and CH CH ), 2.19-2.37 (m, 1H, piperidine CH ), 2.59 (s, 3H, CH ), 2.87-
2 2 2 2 3
2
.94 (m, 1H, piperidine CH), 3.04-3.25 (m, 2H, piperidine CH ), 3.83-4.13 (m, 2H, CH CH ), 7.01
2 2 2
1
3
(
d, J = 8.6 Hz, 2H, aromatic CH), 7.36-7.40 (m, 4H, aromatic CH), 10.64 (bs, 1H, NH). C NMR
(
101 MHz, DMSO-d ): δ 21.94, 22.79, 27.74, 28.16, 40.50, 43.46, 55.38, 63.03, 114.88, 118.27,
6
1
26.50, 129.71, 130.86, 144.03. Purity: 99.35% (HPLC, R :2.966 min; Method: solvent A, CH CN;
t
3
solvent B, FA solution (0.1% v/v), isocratic 90% of FA solution; 15 min).
,7-Dibromo-1-fluoro-10-[2-(1-methylpiperidin-2-yl)ethyl]-10H-phenothiazine hydrochloride
10). The title compound was prepared starting from 28[35] following Method A (3h) and by using
3
(

2-(2-chloroethyl)-1-methylpiperidine hydrochloride. The compound was purified by FCC
(
CHCl /MeOH 95:5) and obtained as a yellow semisolid in 55% yield; after salification, 10 was
3
1
obtained by filtration as a white solid in 70% yield. H NMR (400 MHz, DMSO-d
₆
): δ 1.15-1.97
(
m, 6H, piperidine CH and CH CH ), 2.12-2.35 and 2.72-2.85 (m, each 1H, piperidine CH ), 2.99-
2
2
2
2
3
.15 (m, 2H, piperidine CH ), 3.27-3.53 (m, 1H, piperidine CH), 3.90-4.25 (m, 2H, CH CH ), 7.13
2
2
2
(d, J = 8.6 Hz, 1H, aromatic CH), 7.33 (m, 1H, aromatic CH), 7.40-7.48 (m, 3H, aromatic CH),
1
3
1
5
0.51 (bs, 1H, NH). C NMR (101 MHz, DMSO-d ) δ 22.01, 22.71, 28.28, 28.95, 40.29, 47.52,
6
5.44, 62.75, 115.43 (d, J = 10.2 Hz,), 115.68, 119.54 (d, J = 24.9 Hz), 120.56, 126.21, 128.55 (d,
J = 12.7 Hz), 129.77, 131.11, 132.62, 144.61, 153.24 (d, J = 250. 5 Hz). Purity: 99.71% (HPLC,
R :3.289 min; Method: solvent A, CH CN; solvent B, FA solution (0.1% v/v), isocratic 70% of FA
t
3
solution; 15 min).
1
,3,7,9-Tetrachloro-10-[2-(1-methylpiperidin-2-yl)ethyl]-10H-phenothiazine
11). The title compound was prepared starting from 26 following Method A (overnight) and by
using 2-(2-chloroethyl)-1-methylpiperidine hydrochloride. The compound was purified by FCC
CHCl /MeOH 98:2) and obtained as an orange oil in 29% yield; after salification, the solution was
hydrochloride
(
(
3
evaporated to dryness, triturated with EtOAc and filtered to give 11 as a light-pink solid in 50%
1
yield. H NMR (400 MHz, DMSO-d
₆
): δ 1.50-1.76 (m, 6H, piperidine CH and CH CH ), 1.87-
2
2
2
2.14 (m, 1H, piperidine CH ), 2.51 (s, 3H, CH ), 2.69-2.80 (m, 1H, piperidine CH ), 2.94-2.99 (m,
2 3 2
2
H, piperidine CH ), 3.16-3.27 (m, 1H, piperidine CH), 3.64-3.91 (m, 2H, CH CH ), 7.52 (s, 2H,
2
2
2
1
3
aromatic CH), 7.61 (s, 2H, aromatic CH), 10.30 (bs, 1H, NH). C NMR (101 MHz, DMSO-d ): δ
6
2
1
1.98, 22.66, 28.11, 29.73, 40.13, 50.88, 55.49, 62.41, 126.44, 129.83, 130.45, 136.36, 136.45,
39.33, 139.89. Purity: 99.40% (HPLC, R :9.553 min; Method: solvent A, CH CN; solvent B, FA
t
3
solution (0.1% v/v), isocratic 80% of FA solution; 20 min).
,3,7,9-Tetrachloro-10-(3-pyrrolidin-1-ylpropyl)-10H-phenothiazine hydrochloride (12). The
1
title compound was prepared starting from 26 following Method A (4h) and by using 1-(3-
chloropropyl)pyrrolidine hydrochloride. The compound was purified by FCC (CHCl /MeOH 97:3)
3

and obtained as a brown oil in 73% yield; after salification, 12 was obtained by filtration as a brown
1
solid in 53% yield. H NMR (400 MHz, DMSO-d ): δ 1.73-1.81 (m, 4H, pyrrolidine CH ), 1.83-
6
2
1.91 (m, 2H, CH CH CH ), 2.75-2.84 and 2.97-3.03 (m, each 2H, pyrrolidine CH ), 3.28-3.38 (m,
2 2 2 2
2
H, CH CH CH ), 3.72 (t, J = 7.6 Hz, 2H, CH CH CH ), 7.51 (d, J = 2.3, 2H, aromatic CH), 7.61
2
2
2
2
2
2
1
3
(
d, J = 2.2, 2H, aromatic CH), 10.21 (bs, 1H, NH). C NMR (101 MHz, DMSO-d ,): δ 22.92,
6
25.27, 51.68, 51.86, 53.29, 126.34, 129.25, 129.90, 130.46, 136.36, 139.61. Purity: 98.17% (HPLC,
R :19.248 min; Method: solvent A, CH CN; solvent B, FA solution (0.1% v/v), gradient from 90%
t
3
to 40% of FA solution; 25 min).
1,3,7,9-Tetrachloro-10-(3-morpholin-4-ylpropyl)-10H-phenothiazine hydrochloride (13). The
title compound was prepared starting from 26 following Method A (5h) and by using 4-(3-
chloropropyl)morpholine
hydrochloride.
The
compound
was
purified
by
FCC
(
CH Cl /Cyclohexane/EtOAc 80:10:10) and obtained as a brown oil in 81% yield; after salification,
2 2
1
1
1
3 was obtained by filtration as a white solid in 40% yield. H NMR (400 MHz, DMSO-d ): δ 1.77-
6
.85 (m, 2H, CH CH CH ), 2.83-2.91 (m, 2H, morpholine CH ), 2.94-3.00 (m, 2H, morpholine
2
2
2
2
CH ), 3.19-3.22 (m, 2H, CH CH CH ), 3.63 (t, J = 11.9 Hz, 2H, morpholine CH ), 3.68 (t, J = 7.6
2
2
2
2
2
Hz, 2H, morpholine CH ), 3.82-3.85 (m, 2H, CH CH CH ), 7.50-7.51 (m, 2H, aromatic CH), 7.60-
2
2
2
2
1
3
7
.61 (m, 2H, aromatic CH), 10.59 (bs, 1H, NH). C NMR (101 MHz, DMSO-d ): δ 23.09, 51.42,
6
51.98, 53.79, 63.57, 126.43, 129.31, 129.97, 130.56, 136.51, 139.64. Purity: 99.58% (HPLC,
R :18.769 min; Method: solvent A, CH CN; solvent B, DEA solution (0.1% v/v), gradient from
t
3
30% to 5% of DEA solution; 35 min).
3,7-Dibromo-10-(3-pyrrolidin-1ylpropyl)-10H-phenothiazine hydrochloride (14). The title
compound was prepared starting from 27[16] following Method A (4h) and by using 1-(3-
chloropropyl)pyrrolidine. The compound was purified by FCC (CHCl /MeOH 90:10) and obtained
3
as a brown oil in 64% yield; after salification, 14 was obtained by filtration as a grey solid in 18%
1
yield. H NMR (400 MHz, DMSO-d ): δ 1.77-1.80 (m, 2H, pyrrolidine CH ), 1.83-1.91 (m, 2H,
6
2
pyrrolidine CH ), 1.96-2.03 (m, 2H, CH CH CH ), 2.85-2.92 (m, 2H, pyrrolidine CH ), 3.10-3.15
2
2
2
2
2