Benextramine and derivatives as novel human monoamine oxidases inhibitors: an integrated approach

Article abstract of DOI:10.1111/febs.14994

The two human monoamine oxidase isoforms (namely MAO A and MAO B) are enzymes involved in the catabolism of monoamines, including neurotransmitters, and for this reason are well-known and attractive pharmacological targets in neuropsychiatric and neurodeg

Full text of DOI:10.1111/febs.14994

Received Date : 28-Feb-2019
Revised Date : 02-Jun-2019
Accepted Date : 08-Jul-2019
Benextramine and derivatives as novel human monoamine oxidases
inhibitors: an integrated approach
Maria Luisa Di Paolo1,2, Giorgio Cozza1, Andrea Milelli3, Francesca Zonta4, Stefania Sarno4,
Elirosa Minniti5, Fulvio Ursini1, Michela Rosini5, and Anna Minarini5
1Department of Molecular Medicine, University of Padova, Padova, Italy,
marialuisa.dipaolo@unipd.it
2 Consorzio Interuniversitario “Istituto Nazionale Biostrutture e Biosistemi” Roma, Italy
3Department for Life Quality Studies, Alma Mater Studiorum-University of Bologna, Rimini, Italy
4Department of Biomedical Sciences, University of Padova, Padova, Italy
5Department of Pharmacy and Biotechnology, Alma Mater Studiorum-University of Bologna,
Bologna, Italy
Corresponding authors
Maria Luisa Di Paolo, Department of Molecular Medicine, Biological Chemistry
section, Via G. Colombo 3, University of Padova, 35131 Padova, Italy; tel: ++39-049-
8276119; e-mail: marialuisa.dipaolo@unipd.it;
https://www.medicinamolecolare.unipd.it/
Anna Minarini, Department of Pharmacy and Biotechnology, Via Belmeloro 6, Alma Mater
Studiorum-University of Bologna, 40126 Bologna, Italy; tel: ++39-051-2099709;
anna.minarini@unibo.it; http://www.fabit.unibo.it/
Giorgio Cozza, Department of Molecular Medicine, Biological Chemistry section, Via G.
Colombo 3, University of Padova, 35131 Padova, Italy; tel: ++39-049-8276154; e-mail:
giorgio.cozza@unipd.it; https://www.medicinamolecolare.unipd.it/
Running title: Benextramine derivatives are MAOs inhibitors
Article type : Original Articles
Keywords: benextramine; monoamine oxidases; inhibitors; polyamines analogues, docking
studies;
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1111/febs.14994
This article is protected by copyright. All rights reserved.

Abbreviations: AD, Alzheimer’s disease; BZA, benzylamine; DC50: death concentration
50%; DMSO, dimethyl sulphoxide; DTT, 1,4-dithiothreitol; hr-MAO, human recombinant
monoamine oxidase; MTDL, multi-target-directed ligand; NEM, N-ethylmaleimide; Par,
pargyline; PD, Parkinson’s disease; p-Tyr, p-tyramine; TMS, tetramethylsilane.
Abstract
The two human MAO isoforms (namely MAO A and MAO B) are enzymes involved in the
catabolism of monoamines, including neurotransmitters, and for this reason are well-known
and attractive pharmacological targets in neuropsychiatric and neurodegenerative diseases,
for which novel pharmacological approaches are necessary. Benextramine is a tetraamine
disulfide mainly known as irreversibleα-adrenergic antagonist, but able to hit additional
targets involved in neurodegeneration. As the molecular structures of monoamine oxidases
contain nine cysteine residues, the aim of this study was to evaluate benextramine and eleven
structurally related polyamine disulfides as potential MAO inhibitors. Most of the compounds
were found to induce irreversible inactivation of MAOs with inactivation potency depending
on both the polyamine structure and the enzyme isoform. The more effective compounds
generally showed preference for MAO B. Structure-activity relationships studies revealed the
key role played by the disulfide core of these molecules in the inactivation mechanism.
Docking experiments pointed to Cys323, in MAO A, and Cys172, in MAO B, as target of
this type of inhibitors thus suggesting that their covalent binding inside the MAO active site
sterically impedes the entrance of substrate towards the FAD cofactor. The effectiveness of
benextramine in inactivating MAOs was demonstrated in SH-SY5Y neuroblastoma cell line.
These results demonstrated for the first time that benextramine and its derivatives can
inactivate human MAOs exploiting a mechanism different from that of the classical MAO
inhibitors and could be a starting point for the development of pharmacological tools in
neurodegenerative diseases.
INTRODUCTION
Monoamine oxidases (MAOs) are mitochondrial FAD-containing enzymes that catalyse the
oxidative deamination of biogenic amine neurotransmitters (dopamine, noradrenaline,
adrenaline and serotonin), and a wide variety of xenobiotic amines, including therapeutic
drugs.
The two MAO isoenzymes, MAO A and MAO B (70% of sequence identity), differ for their
substrate specificities, inhibitor sensitivities and tissue localization [1, 2].
Being involved in the catabolism of neurotransmitters, MAOs are well-known
pharmacological targets in various neurological, psychiatric and neurodegenerative diseases,
such as Parkinson’s disease (PD) and Alzheimer’s disease (AD) [3, 4]. Indeed, MAO
inhibitors decrease the catabolism of neurotransmitters and the subsequent generation of
corresponding aldehydes, ammonia and hydrogen peroxide (reaction products), which have
been linked to conditions of oxidative stress that can contribute to the onset of various
diseases [5].
Different MAO inhibitors are or have been used in clinical practice as antidepressant and in
the treatment of PD [6, 7]. Plenty of evidences suggest that the therapeutic potential of MAO
inhibitors (such as deprenyl and rasagiline) are also related to their ability to hit other
biochemical targets [8]. Indeed, their therapeutic actions in the treatment of
neurodegenerative disorders have been linked to the reduction of the oxidative stress,
This article is protected by copyright. All rights reserved.

regulation of gene expression and induction of antiapoptotic and prosurvival factors [3, 4, 8-
11]. The development of multi-target-directed ligands (MTDLs), such as ladostigil, able to hit
simultaneously MAO B and acetylcholinesterase (AChE), strengthens their therapeutic
potential in AD and other neurodegenerative disease [12, 13]. Additionally, in recent years
MAOs (mainly MAO A) emerged as potential therapeutic targets in other pathologies, such
as prostate cancer and cardiovascular diseases [14, 15], opening up new horizons for MAO
inhibitors.
In search of novel types of MAO inhibitors, our previous studies highlighted that polyamine
based analogues are capable to inhibit human MAOs [16, 17]. In particular, ITC-BzDD (N-
benzyl-12-isothiocyanato-N-methyldodecan-1-amine) was found to act as MAO B
competitive inhibitor, and as MAO A irreversible inhibitor (Fig.1) [17].
By exploring the molecular structure of MAOs, it emerged that both the isoforms contain
nine Cys residues, one bound to the FAD-cofactor and eight in the reduced form: kinetic,
structural and site-directed mutagenesis studies supported that some of these Cys may be
important in maintaining the fully functional enzyme [18-20]. To probe the role of Cys
residues on MAOs activity, the most used compounds were the “classical” N-ethylmaleimide
(NEM) and its derivatives [19], and 2,2’dipiridyl disulfide [21]. Importantly, MAO A was
generally found more easily inactivated by thiol oxidation compared to MAO B [19].
Given these premises, benextramine (1, Fig.1), characterized by a cystamine core in its
structure, was herein selected to be studied on MAOs. Compound 1 is a tetraamine disulfide
mostly known as an irreversible α-adrenergic antagonist [22-26], also able to hit additional
targets involved in neurodegenerative processes, such as muscarinic receptors and AChE
enzyme [27, 28]. It was demonstrated that compound 1 exerts its action towards the α-
adrenergic receptor thanks to a disulfide-thiol interchange reaction between the disulfide
moiety of the cystamine core and a Cys residue of the receptor, leading to a disulfide covalent
bond formation. In this context, emerging studies emphasise the role of Cys residues as redox
molecular switches in various signalling pathways and pathologies [29, 30], such as
neurodegeneration [31]. Although compound 1 has been used as a template to design
polyamine analogues as MTDLs against neurodegenerative diseases [27, 28], no data is
available on its effects on MAOs.
Thus, the aim of this study was to evaluate compound 1 and a series of structurally related
polyamine disulfides (2-12) (Table 1) as human MAO inhibitors. The results obtained by
exploiting a multidisciplinary approach including kinetic studies integrated with docking
experiments and in cell validation, showed that most of the tested compounds induced
irreversible inactivation of both MAO isoforms. Compound 1 showed the highest kinact value
for MAO A, and the “short” analogue 12, was found to be the most potent MAO B
inactivator. By the comparison of the inactivation potency (kinact/KIE), compounds 1 and 9
were found to be MAO B selective. The importance of the disulfide bridge in the mechanism
of inactivation of these molecules was demonstrated by testing carbon analogues of
compound 1. Docking experiments highlighted the role of Cys323 (in MAO A) and Cys172
(in MAO B) as targets of compound 1 and its derivatives. This investigation led to discovery
of a new class of MAO inhibitors with an unprecedented mechanism of action. This
discovery may have important implication in the design of new MAO inhibitors as
therapeutic agents or bio-pharmacological investigation tools.
This article is protected by copyright. All rights reserved.

RESULTS
Rational of compound selection
In order to investigate the structural features required for an optimal interaction of a
polyamine disulfide scaffold with the MAO A and MAO B active sites, compound 1, together
with polyamine disulfide analogues 2-12, (Table 1), were tested as potential MAO A and B
substrates and inhibitors. The structural modifications applied to compound 1 concerned the
variation of the length of the methylene linkers between the nitrogen atoms (compounds 2-5),
the replacement of the o-methoxybenzyl groups on the terminal amine functions with
catechol (compound 6 and 9), pyrrole (compound 8) or pyridine (compound 7) rings, and a
decrease of the number of amine functions in the polyamine backbone, leading to the
asymmetrically substituted triamine (compound 10) and to diamines (compounds 11 and 12).
To boost the crucial role of the disulfide moiety in MAO interaction, Met-666 and
methoctramine, two carbon analogue of compound 1 (Fig. 1) and the new diamine 13 (Table
1), featuring the hydroxylated analogue of compound 1 truncated in two halves, were also
included in the study.
Screening of the effects of compound 1 and its analogues on monoamine oxidase activity
To evaluate the potential effects of the polyamine analogues on human recombinant -MAOs
(hr-MAOs), compounds 1-13 (Table 1) were firstly evaluated as potential substrates. No
significant MAO activity was detected testing all the analogues in the 0.1-1 mM range of
concentration, with the exception of the compound 10, characterized by a free primary amino
group. However, compound 10 behaved as a very poor substrate for MAO A (enzyme
activity less than 1% compared to 1 mM, (p-Tyr).
Consequently, compounds 1-13, Met-666 (carbon analogue of 1), cystamine (core moiety of
1), and NEM (reference-standard reactive compound for thiols) were tested as potential
inhibitors.
The kinetic parameters (Km and Vmax) of MAO A and MAO B were determined in the
presence of 100 M compounds added in the assay solution immediately before the substrate.
The results are shown in Table 2 as the ratio between the Vmax or Km values with respect to
control samples (in absence of any compounds). Most of these molecules affect both Vmax
and Km, depending on the MAO isoform. The double-reciprocal plots reported as examples in
Fig. 2 indicate that the most effective compounds decreased mainly MAO B substrate affinity
(Fig. 2B) (Km/Km0  1.6-2 with compounds 1 and 9 and Km/Km0  6 with compound 4, Table
2) whereas, in the case of MAO A, both the Vmax and/or Km values were affected (Fig. 2A).
To understand the mechanism of action of these polyamine disulfides, their time-dependent
inhibitory activity was investigated. MAOs were pre-incubated with 100 M of each
compound, in absence of substrate, and the residual enzyme activity was determined at
various times, after dilution and addition of the substrate. Residual enzyme activity was
mostly found to decrease with increasing incubation time, suggesting that most of the
polyamine disulfides behaved as time-dependent MAO inactivators (see example of
compound 1 in Fig 3 and the calculated apparent inactivation constant values, kobs’ in Table
2). Moreover, in our experimental conditions, these compounds were more potent than NEM
and cystamine, being compounds 1, 4 and 12 the most effective molecules. In agreement with
previous studies [19], NEM was found to inactivate MAO A faster than MAO B, whereas
cystamine behaved as a very weak MAO B inactivator. Moreover, the substitution of the
cystamine core of compound 10 with the ethanolamine moiety, leading to compound 13,
determined the loss of the inactivation capability toward both the MAO isoforms.
This article is protected by copyright. All rights reserved.

Concerning the length of the spacer between the inner and outer amine functions of the
tetraamine backbone, compound 2 and 3, respectively, the shorter and longer homologous of
compound 1, showed almost the same MAO inactivation profile. Increasing the length of the
spacer up to 12 methylene units, as in compound 4, resulted in a significant rise of the
inactivation potency, suggesting that more hydrophobic chains boost the reactivity against
MAOs. Elongation of the cystamine core of compound 2 with two methylene units led to
compound 5, resulting in a significant decrease of MAO inhibitory ability, particularly for
MAO A, with an apparent inversion of selectivity (5 is more effective on MAO B and
compound 2 is more effective on MAO A). The replacement of the terminal o-
methoxybenzyl groups of compound 1 with other aromatic rings resulted in a reduction of the
activity: compound 7, bearing the pyridine ring at the two terminal ends, is the less potent
compound of the series. The substitution with a pyrrole ring has similar effect to that with the
catechol ring (compound 6) on MAO B; however, the introduction of the catechol moiety
increased the selectivity for MAO B compared to the lead compound 1. This apparent MAO
B selectivity and decreased MAO A inhibitory activity was further enhanced in compound 9,
characterized both by a “lengthened” cystamine core and the catechol groups at the terminal
ends.
The reduction in the number of amine functions in the polyamine backbone, as in the
asymmetric triamine 10 and in the diamine 11, decreased their inactivation capability mainly
towards MAO A in comparison with compound 1. Compound 12, where the o-
methoxybenzyl groups of compound 11 are replaced with o-methoxyphenoxyethyl moieties,
turned out to be the second more potent compound in inactivating MAO B.
On this basis, compounds 1, 4 and 12 (the most potent ones), compound 9 (the most selective
for MAO B) and the triamine 10 were selected for a more detailed kinetic characterization as
time-dependent inhibitors.
Kinetic characterization of the most potent and selective compounds
The inactivation rate constant values were determined at various concentrations of
compounds 1, 4, 9, 10 and 12 with both MAOs isoforms. From the plots kobs’ vs inhibitor
concentrations, saturation kinetics were found with compounds 1, 9, 10 and 12. For these
molecules, the two steps mechanism of inactivation, which consider a first step of formation
of the reversible complex (EI) followed by the second step of the chemical reaction with the
production of the irreversibly inactivated enzyme (E-X*) (Eqn.1-2, in Material and Methods),
was found to fit the experimental data. In Fig. 4 the analysis of the kinetic data obtained with
1 vs MAOs (panels A-D) and with 12 vs MAO B (panels E,F) are shown as examples. The
calculated kinact (inactivation constant rate) and KIE (inhibition constant that represents the
enzyme-inhibitor dissociation constant) are reported in Table 3.
In the case of compound 4, saturation was not reached (example in Fig. 4, panels G, H), thus,
from the slope of the plots kobs’ vs inhibitor concentration, only the kinact/KIE parameter was
calculated. Notably, also inactivation of MAO A by NEM showed a similar non-saturable
behaviour (Fig. 4, panels I, J), in agreement with other studies [21]. Under our experimental
conditions, inactivation of MAO A by NEM (kinact/KIE = 400 min-1 M-1) is about one order of
magnitude slower respect to compounds 1, 4 and 12.
From Table 3 it appears that all the tested compounds show higher affinity (lower KIE values)
for MAO B than for MAO A, resulting compound 1 the most potent MAO B inhibitor (KIE =
6 M for MAO B vs KIE = 175 M for MAO A). On the other side, the highest inactivation
rate constants (at saturating concentration of compounds) were found with compound 12
(kinact = 1.42 min-1 vs MAO B) and 1 (kinact = 0.98 min-1 vs MAO A). Consequently, the best
inhibitory potency (kinact/KIE), which represents the rate of inactivation under not-saturating
This article is protected by copyright. All rights reserved.

concentration of inhibitor, was found with compound 12 (kinact/KIE =109 x103 M-1min-1 vs
MAO B), followed by compound 4 (kinact/KIE =51x103 M-1min-1 vs MAO A). By considering
kinact/KIE, the most selective MAO B inhibitors were the catechol derivative 9 (MAO A:MAO
B 1:27) and compound 1 (MAO A:MAO B 1:7), but only 1 maintained a high inactivation
potency for both MAO isoforms.
The unsymmetrical substituted cystamine derivative 10, lacking one of the N-(2-
methoxybenzyl) hexanamine moieties of 1 was found to inactivate both MAOs with lower
selectivity and potency than its precursor 1. In particular, the kinact value for MAO A is halved
and its kinact/KIE for MAO B is decreased by a factor of 5, in comparison to 1.
Inhibition and reversibility mechanisms of compound 1
To understand the role of the disulfide bond in inactivating MAOs, we focused our
investigation on compound 1, showing the best balance between selectivity and inhibitory
potency vs MAO isoforms. After incubation and reduction with 1,4-dithiothreitol (DTT), the
disulfide bond of compound 1 was broken in half producing two molecules of thiols. The
inactivation capability was tested by comparing the effect of 100 M compound 1 and of 100
M of DTT pre-reduced compound 1 on MAOs activity. Intriguingly, the products derived
by the reduction of compound 1 inactivated MAO A and MAO B about 5 and 8 times faster,
respectively than the entire molecule (Fig. 5).
To note that cystamine showed no increase in the inactivation rate constant vs MAOs after
DTT treatment that led to cysteamine formation. These results suggested that the formation of
a disulfide bond between the sulfhydryl group of the reduced form of compound 1 and a Cys
residue of MAOs could be a possible mechanism of MAOs inactivation and that this
interaction takes place faster with respect to the one with the entire disulfide molecule.
To evaluate the reversibility of the chemical bond between compound 1 and MAOs, the
inactivated enzymes were extensively dialyzed. After dialysis, no recovery of MAO A
activity was detected, whereas, in the case of MAO B, about 41% of activity was found
(Table 4), suggesting that the irreversible inhibition induced by compound 1 versus MAO B
is less efficient, as demonstrated also by the lower kinact (see Table 3). The irreversible and
partial reversible inactivation of MAO A and MAO B, respectively, were confirmed also by
using compound 4, bearing the longest chain within this series (data not shown).
To note that, after inactivation by compound 1, MAOs enzymatic activity was not recovered
by DTT treatment (Table 4). This behaviour suggested that the disulfide bridge formed
between compound 1 and a Cys of MAOs, could be buried, and, consequently, DTT-
inaccessible.
To further confirm the role of the disulfide moiety of 1 in inactivating MAOs, Met-666, the
carbon analogue of compound 1, was assayed. Unlike compound 1, if Met-666 is incubated
with MAOs, a decrease in the enzyme activity over time was not observed (Table 2).
Moreover, if tested without pre-incubation, Met-666 and compound 1 (100 M) showed
similar effects on Km (Table 2, first columns). These results demonstrated that the substitution
of the disulfide moiety of the irreversible inhibitor 1 with two methylene groups, as in Met-
666, led to a weak reversible inhibitor of MAOs. In particular, Met-666 decreased Vmax (KI =
790 ± 190 M) and increased Km values (KEI>1mM) of MAO A (Fig. 6A) and increased Km
values of MAO B (KEI = 540 ± 130 M (Fig. 6B). Similar results were obtained also with
methoctramine (Met-686, molecular structure in Fig. 1), another carbon analogue of 1,
characterized by an eight methylene inner chain.
All these data confirmed the key role played by the disulfide moiety in the irreversible
mechanism of human MAO inactivation by compound 1.
This article is protected by copyright. All rights reserved.

Protection experiments to prove that polyamine disulfides are “active-site-directed
inhibitors”
The reported experimental evidence clearly demonstrated that compound 1 and its analogues
inactivate MAOs through the disulfide moiety; however, it remains unclear if the mechanism
of inactivation involves Cys located in the MAO active site and if the polyamine disulfides
are to be considered active-site-directed inhibitors. To answer to this question, the ability of a
competitive inhibitor or of a substrate to protect against compound 1 inactivation was
evaluated [32].
First of all, inactivation by 1 was performed in the presence of the competitive inhibitor isatin
[2, 33]. The apparent inactivation rate constant of MAOs (k’obs) obtained with the lead
compound 1 was slowed down in the presence of isatin (Table 5), by a factor of 6 and 4 in the
case of MAO A and B, respectively. Similar results were obtained also with compound 4, the
longest analogue of the series.
Furthermore, the effect of the substrate on the inactivation rate was evaluated by using the
compounds with the highest inactivation potency (kinact/KIE), compound 4 for MAO A, and
compound 12 for MAO B. In the presence of saturating concentrations of substrate (1 mM p-
Tyr for MAO A and 10 mM benzylamine for MAO B) both compounds 4 and 12 lost most of
their capability to inactivate MAOs (Table 5).
These results strongly support the fact that the active site of MAOs is the main target of
compound 1 and its analogues, thus involving Cys residues located in the catalytic cavity.
Docking studies
The above reported biochemical investigations suggested that the binding of compound 1, 4
and 12 might occur in the cavity of MAOs active site, through the formation of a covalent
disulfide bond with a Cys of MAOs. With the aim to evaluate this hypothesis and to
determine the binding mode of these compounds, docking studies were carried out using
MAO A and MAO B crystal structures.
Before the docking procedure, a Site Finder strategy (see Material and Methods) was applied
to select all the possible binding sites of MAO A and MAO B. Interestingly, in both cases the
most suitable interaction zone was superimposed with the catalytic binding site of the two
enzymes, accordingly to the mechanism of action of the molecules above discussed. The
catalytic sites of both MAO A and MAO B are characterized by a hydrophobic tunnel and a
small hydrophilic region located close to the FAD cofactor. Several cysteines are located
inside the catalytic tunnel but only a couple of them are exposed towards the catalytic site, in
particular, MAO A Cys323 and MAO B Cys172, located in the opposite side of the two
cavities, are not conserved. Intriguingly, MAO A Cys323 (Thr314 in MAO B) is located in a
motif almost identical except for the Cys to Thr substitution (MAO A: DYCGCMIIED;
MAO B: DYCGTMIIDG); similarly, MAO B Cys172 (Asn171 in MAO A) belongs to a
very conserved sequence (MAO B: FVNLCVTAE; MAO A: FVNINVTSE). To note that
even if one single cysteine is conserved in both MAO A and MAO B (Cys321 and Cys312,
respectively), these amino acids are not accessible, accordingly to the Site finder approach.
Indeed, MAO A Cys321 is shielded by Cys323 and Leu97, while MAO B Cys312 by Thr314
and Leu88.
To obtain structural information about the residues involved in the interaction inside the
active site of MAOs and, since, under reductive conditions, compound 1 resulted in a more
efficient MAO inactivator, compounds 1, 9 and 12 were docked against MAO A and MAO B
crystal structures as half-molecules, representing the thiol derivatives.
This article is protected by copyright. All rights reserved.

Docking results suggested that thiol molecules derived from the disulfide compounds 1 and
12 bind MAO A and MAO B active sites by exploiting two different hydrophobic cavities,
located inside the tunnel, close each other but not superimposable. In particular, in MAO A,
the two compounds insert their o-methoxybenzyl head into a hydrophobic region assembled
by Pro113, Leu97, Ile 325 and the alkyl chain of Gln99 (Fig. 7A and C). On the other side,
the aromatic heads of compounds 1 and 12 fit into a hydrophobic cavity built by Trp119,
Leu164, Ile199, Leu167 in MAO B (Fig. 7B and D). While the heads of the two compounds
are structurally conserved, sharing the same binding pocket in MAO A or in MAO B, the
thiol tails present huge differences in length and chemical composition, thus reflecting an
unrelated binding motif towards MAO A Cys323 and MAO B Cys172. Indeed, the shorter
thiol chain of compound 12 is perfectly directed versus MAO A Cys323 by the hydrophobic
interaction with Phe208 and Ile335, further stabilized by the hydrogen bond between the
amino group of compound 12 and the carbonyl group of Phe208 (2.7 Å, Fig. 7A). In MAO B
Ile316, Tyr326 and Leu171 address the sulfhydryl group of the thiol compound deriving from
12 towards Cys172 (Fig. 7B). Interestingly, the docking pose versus MAO B is partially
superimposable with the crystallographic pose of pioglitazone (PDB code: 4A79, Fig S1 A).
To note that the in silico pK values of the two complexes MAO A/compound 12 and MAO
B/compound 12 are very similar (7.5 and 7.8, respectively).
On the contrary, the bulkier chain of compound 1 is not able to efficiently interact with MAO
A Cys323 due to the short distance between Cys323 and the hydrophobic cavity bearing the
methoxyphenyl ring of 1 (Fig. 7C). However, in the case of MAO B, compound 1 can easily
accommodate his chain directing it versus Cys172, by exploiting the same hydrophobic
interactions discussed for compound 12 (Fig. 7D). The difference binding poses of compound
1 against MAO A and B are reflected also on the in silico pK values. Indeed, while MAO A/1
complex present a pK of 6.2, MAO B/1 seems more efficient with a value (pK = 8.0) close to
the one observed for MAO B/12 complex.
In the case of compound 9, the docking pose presents substantial differences. The substitution
of the o-methoxyl group of the benzyl moiety with two hydroxyl groups drives the compound
to interact in a more hydrophilic and deep region of MAO catalytic tunnel by performing
direct contact with the FAD molecule (Fig. 7E and F). This interaction is further stabilized by
hydrophobic contacts with MAO A Phe352 and Tyr407 as well as with MAO B Phe343 and
Tyr398 (Fig. 7E and F). The thiol chain of compound 9 is directed versus MAO A Cys323
and MAO B Cys172, being the interaction with the first one (Fig. 7E), apparently less
efficacious compared to the second one (Fig. 7F). This event is also reflected in the in silico
pK values (MAO A/9 pK = 6.5; MAO B/9 pK = 7.1), even if they are close to the limit of
significance cut off.
Interestingly, the in silico pK values above reported showed the same trends and selectivity
as the KIE determined by the kinetic approach: compound 12 binds with similar “affinity” to
both MAOs, while compounds 1 and 9 show higher affinity for MAO B. To note that also the
whole polyamine disulfides considered in the in silico analysis are able to bind to MAO A
and MAO B. However, the pK values calculated suggested that the binding of the whole
molecules to MAOs are less efficient compared to those of the corresponding thiols,
supporting the results obtained in the biochemical experiments.
In summary, the docking results suggests that compounds 1, 9 and 12 can efficiently interact
with the catalytic sites of MAO A and MAO B, directing their thiol chains versus specific
cysteines. To confirm the compatibility of the docking poses with the possible formation of a
disulfide bridge, covalent docking has been performed for compound 1 versus MAO A and B
(Fig. S1 B and C). The results obtained are superimposable with the previously reported for
non-covalent docking; slight differences are explainable by the application of an optimization
process to the final covalent complexes.
This article is protected by copyright. All rights reserved.

Toxicity evaluation and in-cell MAO inhibition of compound 1
After the kinetic and computational studies on the interaction of compound 1 and derivatives
with human-recombinant MAOs, the effect of this novel type of MAO inhibitors on a suitable
cellular model has been prioritized. Since the previously identified targets of compound 1 are
involved in neurodegenerative diseases, the SH-SY5Y neuroblastoma cells, one of the most
used model of neuronal function [34], were chosen as cellular model. To note that this cell
line, as most of the neuronal cell models, expresses mainly MAO A [35] (see also
https://www.proteinatlas.org). MAOs activity was preliminary tested in cellular lysates, using
pre-incubation with clorgyline (irreversible and specific inhibitor of MAO-A) and pargyline
(irreversible inhibitor of both MAO A and MAO B). Clorgyline was able to completely
abolish MAOs activity, confirming that MAO A is the main type of MAO isoform expressed
in SH-SY5Y cells, as previously reported [35].
Before evaluating the effect on MAOs activity, the cytotoxicity of the compounds was tested.
The most efficient compounds 1 and 12 were used for these studies. Their effects were
compared to that of methoctramine, a polyamine analogue lacking the disulfide moiety,
already studied on the SH-SY5Y cells [36].
The cells were incubated for 12 and 24 h in the presence of increasing concentration of the
above reported compounds and then the cell viability was determined. From the DC50 (Death
Concentration 50) values calculated from plots shown in Fig. 8, it clearly appears that
compound 1 has a much lower cytotoxic effect than compound 12 (DC50 = 252  28 M vs
DC50 = 25  7 M and DC50 = 208  10 M vs DC50 = 22  4 M with compound 12, at 12
and 24 h, respectively); the DC50 = 72  10 M obtained with methoctramine (at 24 h) was in
good agreement with the previously reported value [36]. On this basis, only the less toxic
compound 1, was tested as potential MAO inhibitor on the cellular system; pargyline was
also assayed as standard MAO inhibitor. After incubation of cells with compound 1 for 3 and
24 h, a significant decrease in MAOs activity was found (Fig. 9A). After 24 h, an IC50 of
about 13 M was found, a concentration too low to induce significant citotoxicity. No
decrease in MAO A protein content was found by western blotting and immunostaining
analysis (Fig. 9B). From Fig. 9B, an increase of about 30% of MAO A protein was observed
at the highest compound 1 concentration (250 M). An increase in MAO A expression was
previously reported in the same cell line with other MAO inhibitors (selegiline and rasagiline,
MAO B inhibitors) [37].
The above reported results confirm the effectiveness of compound 1 in the inactivation of
MAO A in a cellular system, at not cytotoxic concentrations.
DISCUSSION
This study demonstrated for the first time that benextramine (1) and its derivatives can
inactivate human MAO A and MAO B isoforms, exploiting a mechanism of inactivation
different from that of the classical MAO inhibitors. While the traditional MAO inhibitors act
by targeting the FAD cofactor, the mechanism herein proposed for benextramine involves a
time-dependent inactivation by the formation of disulfide bridge with the thiol group of a Cys
of the enzyme.
The inactivation rate constant values with compounds 1-13 were found dependent on: a) the
polyamine disulfide skeleton, b) the aromatic substituent on the two terminal ends of the
molecule and c) the MAO isoform. In particular, the o-methoxybenzyl moieties at the
terminal ends of the tetraamine disulfide skeleton are responsible of a stronger potency than
other substitutions; on the other side, the elongation of two methylene groups in the inner
This article is protected by copyright. All rights reserved.

cystamine core, decreases the inactivation potency versus MAO A to favour the selectivity
towards MAO B.
By considering kinact/KIE (index of inactivation efficiency), the diamine 12, the shortest
analogue of the series, was found to be the most potent MAO B inactivator (kinact/KIE = 1x105
M-1min-1), and the tetraamine 4, the longest one, appears to be the more potent MAO A
inactivator (kinact/KIE = 57x103 M-1min-1); however, both the compounds are poorly selective.
A good selectivity for MAO B was found for compounds 1 and 9 (MAO A: MAO B, 1:7 and
1:27, respectively).
By comparing these kinetic results with that previously found with ITC-BzDD (Fig. 1) [17],
compounds 1, 4 and 12 resulted more potent than ITC-BzDD as MAO A inactivator in terms
of kinact/KIE or kinact. Furthermore, although ITC-BzDD (Fig. 1) behaves with a different
inhibition mechanism, that is as competitive inhibitor of MAO B, its KIE value towards MAO
B is comparable with those found with the inactivator compounds 1 and 12.
Moreover, by comparing the kinetic behaviour of the standard and selective inhibitors
targeting the FAD cofactor, clorgyline for MAO A and deprenyl for MAO B, 1 analogues
show a lower inactivation efficiency (kinact/KIE, in the range 106-108 M-1min-1 for clorgyline
and deprenyl vs kinact/KIxE, in the range 104-105 M-1min-1), but good inactivation rate constant
values (kinact about 0.9 min-1 with clorgyline and 0.08 min-1 with deprenyl vs kinact = 1-1.4 min
-1 with 1 and 12). Additionally, the inactivation efficiency of 1 analogues is higher than that
of other thiol reactive reagents (2,2'-dipyridyl disulfide, kinact/Ki 1.2-0.3 x 103 M-1 min-1 for
MAO A [21] and NEM, kinact/Ki 0.4 x 103 M-1 min-1). These data suggest that the mechanism
of chemical inactivation by compound 1 derivatives could be considered as a potential new
mode of inactivating MAOs.
The inactivation by compound 1 was found irreversible for MAO A and partially reversible
for MAO B. This partial reversibility is probably explained by the stronger but reversible
physical interaction of compound 1 with MAO B and by a concomitant slower chemical
reactivity or reversibility of the chemical reaction (kinact = 0.25 min-1 for MAO B vs 0.98 min-
1 in MAO A).
By the comparison of the inhibitory behaviour of Met-666 vs compound 1 and compound 13
vs compound 10, respectively, it was demonstrated the crucial role played by the disulfide
bridge in the inactivation process. Indeed, in absence of this feature, inactivation did not
occur and the compounds act as poor reversible inhibitor (Met-666), suggesting the formation
of a disulfide bond between compound 1 and MAOs.
Previous structural studies demonstrated the presence of eight reduced Cys residues in both
MAO proteins [19] and other authors reported that the mutation/chemical modification of
some of them may affect enzyme activity [20, 21]; so the Cys residues of MAOs are the best
candidates to react with the disulfide moiety of compound 1 and its derivatives.
The protection by inactivation in the presence of a competitive inhibitor or of a substrate
strongly supported that these compounds target a Cys residue of MAO active site cavity. The
docking experiments pointed to the not conserved Cys323 (in MAO A) and Cys172 (in MAO
B) as targets (Fig. 7): the covalent disulfide bond formed between the molecules and these
Cys residues of MAO active sites should impede/occlude the entrance of substrate towards
the FAD-cofactor, resulting in enzyme inactivation.
However, alternative binding or additional binding to Cys at remote sites, such as on the
enzyme surface, may not be excluded, even if the not recovery of enzyme activity, after DDT
treatment suggests that the Cys involved in the binding of compound 1 and its analogues are
solvent inaccessible.
This article is protected by copyright. All rights reserved.

Therefore, concerning the mechanism of MAO inactivation by compound 1 derivatives, two
ways are here proposed: a) the pre-reduced disulfide (half-molecule of the compound) could
react with the Cys of MAO active site via a sulfenic acid intermediate (–SOH), which might
form inside the active site in air-exposed aqueous solutions [38]; b) the disulfide core of the
compound could react with the Cys-SH of MAO active site through a classical reaction of
thiol-disulfide interchange [38].
The first way is supported by the extremely efficacious enzyme inactivation by the DTT-
reduced compounds 1 and 10, which demonstrated that the free -SH groups formed are much
more efficient than the whole disulfide-containing molecule in inactivating both MAOs.
Under physiological conditions, in a cellular system, the presence of reducing agents, such as
glutathione, could speed up the reduction of compound 1 analogues before reacting with
MAO (bound to the mitochondrial membrane).
On the other side, the known flexibility of the molecular structure of the enzyme [40] and of
the polyamine backbone might allow the disulfide compound/Cys interaction in the active
site, through a thiol-disulfide interchange reaction, even if docking experiments suggested
better binding affinities for the thiol compound (half molecule), compared to the disulfide.
Indeed, these results could be impaired by the semiflexible docking protocol, in which the
compounds enjoy full degree of freedom, while the protein is completely rigid and not
subjected to induced fitting.
The efficacy of compound 1 in inactivating MAO A was confirmed also in a model of
neuronal cells, where inactivation occurs at concentration one order of magnitude lower than
its cytotoxic concentration (IC50 = 13 M vs DC50 = 208 M).
In conclusion, the main finding of this study is that human MAOs are novel targets for the
tetraamine disulfide benextramine and its derivatives, exploiting a novel inactivation
mechanism of MAOs.
Starting from the above reported results, further studies are in progress to design novel
analogues with a better selectivity and reactivity than compound 1, for a future potential
development as pharmacological tools in neurodegenerative diseases.
MATERIALS AND METHODS
Chemicals and reagents
All chemicals were of analytical grade and purchased from Fluka-Sigma-Aldrich S.r.l. (Milan,
Italy) with the exception of Amplex Red reagent (10-acetyl-3,7-dihydroxyphenoxazine),
purchased from Molecular Probes/Invitrogen (Invitrogen s.r.l, San Giuliano Milanese (MI,
Italy).
Human recombinant MAO A and MAO B expressed in baculovirus infected BT1 cells (5
mg/mL) and horseradish peroxidases were purchased from Fluka-Sigma-Aldrich s.r.l. (Italy).
Rabbit polyclonal to MAO A was from Abcam and anti-actin was from Sigma-Aldrich S.r.L.
(Milan, Italy)
This article is protected by copyright. All rights reserved.

Synthesis of compounds 1-13
Compounds 1-12 were synthesized following improved procedures respect to those previously
reported [22-24, 26, 39]. We describe here the modified synthesis of compound 10 and the
synthesis of the new compound 13 (Scheme 1). The detailed procedure is reported in the SI
(NMR spectra in Fig.S2).
Briefly, 2-methoxybenzaldeyde was subjected to a classical reductive amination with 6-
aminohexanol to furnish compound 14. The secondary nitrogen was protected with Boc (15)
and the hydroxy functionality was activated with tosyl chloride (16). Reaction between 16
and cystamine led to intermediate 17, which was subjected to acidic deprotection to remove
Boc group giving compound 10 as hydrochloride salt. Reaction between 16 and ethanolamine
led to compound 18, which was subjected to acidic deprotection to give compound 13 as
hydrochloride salt. Benextramine can be obtained adopting the same procedure described for
17, by using 16 and cystamine in 2:1 equivalent ratio, followed by deprotection of Boc
groups.
Monoamine oxidase activity experiments
Amine oxidase assay methods
Amine oxidase activity was determined by measuring H2O2 generation rate by a peroxidase-
coupled continuous assay. The Amplex Red reagent was used as fluorogenic substrate for
horseradish peroxidase [40]. No significant interference of the various polyamine analogues
on fluorescence intensity was observed.
All experiments were carried out in 50 mM Hepes-NaOH, 50 mM KCl, 120 mM NaCl, pH
7.4 and at 37°C.
Assays were carried out in a final volume of 800 l, in the presence of Amplex Red (100
M) and horseradish peroxidase type II (5 UmL-1). Benzylamine (BZA) and p-tyramine (p-
Tyr) were used as substrates for MAO B and MAO A, respectively. Initial velocities were
determined by measuring the increase in fluorescence intensity (exc = 563 nm and em = 586
nm); H2O2 generation rate was calculated from the change in fluorescence intensity, by
means of calibration curves obtained by serial dilution of stock solution of H2O2. In particular,
since the fluorescence intensity generated by the system peroxidase-Amplex Red-hydrogen
peroxide, decreases in the presence of high tyramine concentrations, the reaction rates of
MAO A, at various tyramine concentrations, were calculated using specific calibration curves
built for each concentration of the substrate used for the enzyme activity assay. If not
differently reported, hr-MAO A and hr-MAO B concentrations in the assay solutions were 3
and 6 g/ml, respectively.
In experiments carried out in the presence of DTT, which was found to interfere with the
Amplex Red- peroxidase assay method, MAO activity was determined by the method using
kynuramine as substrate [41]. Briefly: enzyme stock solutions, after incubation in the presence
of DTT (with or without inhibitor), were diluted with assay buffer to get a final concentration
0.003 mg/mL for MAO-A and 0.006 mg/mL for MAO-B; 300 M kynuramine (final
concentration) was added to the 200 l of assay solution (0.1 M potassium phosphate buffer,
pH 7.4, 0.1 mM EDTA). After 30 min incubation at 37°C, the reaction was stopped by
addition of 2 M NaOH (80 l) and 480 L of distilled water. Kynuramine deaminated by
MAO spontaneously cyclizes to give 4-hydroxyquinoline, the amount of which was
determined by the fluorescence intensity of the peak of its emission spectra (exc = 330 nm and
This article is protected by copyright. All rights reserved.

em 330-530 nm), using a specific calibration curve built with the standard 4-
hydroxyquinoline.
This kynuramine assay was also used for the assay of MAO activity in SH-SY5Y cell lysates
(25 L of cell lysates in 200 L of the assay solution). MAO activity was expressed as
pmol4HQ min-1 mg protein-1. To verify the presence of MAO A activity in this cells, SH-SY5Y
lysates, were pre-incubated for 15 min at 37 °C with clorgyline (1 M, to inhibit MAO A) and
pargyline (250 M, to inhibit both MAO A and MAO B); then the substrate (kynuramine) was
added and the residual MAO activity was determined.
Inactivation of MAOs by compound 1 pre-treated with DTT
Compound 1 (1mM) was pre-incubated with DTT (20 mM) in Hepes buffer (pH 7.4) for 15
min at 37°C. In these conditions, the disulfide bond is reduced breaking compound 1 in two
half-molecules characterized by the presence of thiol groups. After reduction, experiments of
inactivation of MAOs were carried out using 100 M of compound 1, or 100 M of DTT-
reduced compound 1. MAOs without any compounds and MAOs with only DTT (at 2 mM
final concentration, as in the sample with the reduced compound 1), were run under the same
experimental conditions and were taken as control samples. MAOs activity studies were
carried out using the assay method with kynuramine as substrate.
The effect of DTT-oxidation product on MAOs, was evaluated using DTT-pretreated
cystamine under the same experimental conditions.
Reversibility experiments
MAO A and MAO B (0.025 mg/mL) were incubated in the presence of compound 1 or 4 200
M. When the residual activity was decreased and constant at its minimum value (about zero
for MAO A and about 30%, in the case of MAO B), the samples were extensively dialyzed
against several changes of buffer (100 mM potassium phosphate, pH 7.20, at 4 °C). Control
samples of MAOs were run simultaneously, under the same experimental conditions. After
dialysis, the residual MAO activity and protein concentration were determined, to calculate
the residual specific activity (nmolH2O2 min-1 mg-1 protein).
The reversibility was also tested by adding DTT to the MAOs previously inactivated by 1.
DTT was added at a final concentration of 10 mM, for about 20 min, at 36°C. The residual
enzyme activity was tested after dilution (1:40) of the enzyme- 1 solution, before adding
substrate. Control samples of enzyme, in the presence and absence of DTT were also run and
treated under the same experimental conditions to verify the effect of DTT on the free
enzyme.
Protein assay
The protein content of samples was measured by the Bradford method, with bovine serum
albumin as standard [42].
This article is protected by copyright. All rights reserved.

In silico analysis
The crystal structure of human MAO A and MAO B were retrieved from the Protein Data
Bank (PDB codes: 2Z5X and 4A79, respectively) and processed in order to remove ligands
and water molecules. Hydrogen atoms were added to the protein structures using standard
geometries with the MOE program [43]. To minimize contacts between hydrogens, the
structures were subjected to Amber99 force-field minimization until the rms (root mean
square) of conjugate gradient was <0.1 kcal•mol−1•Å−1 (1 Å = 0.1 nm) keeping the heavy
atoms fixed at their crystallographic positions. Compound 1, and the selected derivatives 12
and 9 were built using MOE, considering the molecules in both oxidized and reduced states;
Gasteiger partial charges (MOE Suite) were added.
Before the molecular docking procedure a Site Finder Approach [43] was performed; this
tool, by identifying regions of tight atomic packing (using Alpha Shapes, a generalization of
convex hulls developed in [44], allows the identification of pockets and surface sites and was
successfully applied in different target analysis [45, 46]. In the case of MAO A and MAO B,
the procedure suggests a cavity, superimposable with the catalytic tunnel of both the
enzymes, as the best interaction zone for the molecules considered in this work.
Using the Site Finder indications, a preparation step for the docking procedure was performed
with Glide package (Schrodinger suite) [47]. A set of co-crystallized inhibitors of MAO A
and MAO B were used to optimize the docking protocol, determining a final Glide grid (outer
grid dimensions: 14Å x 28Å x 16Å for MAO A; 18Å x 25Å x 15Å for MAO B). Compounds
1, 12 and 9 were docked in the defined sites and grids by using the XP (extra precision)
procedure. The in silico association constant (pK), estimated using MOE suite, was optimized
exploiting the crystal structures available during the docking preparation phase. It is
expressed as the summation of a directional hydrogen-bonding term, a directional
hydrophobic interaction term, and an entropic term; higher values indicate more affinity with
a significance cut off of 0.5.
Covalent docking of compound 1 versus MAO A and MAO B was performed by using
CovDock tool [47]. The optimization of the final covalent complexes was performed using
Prime [47].
Cell cultures
Human SH-SY5Y neuroblastoma cells (A.T.C.C., Manassas, VA) were cultured in DMEM
medium (Sigma-Aldrich) supplemented with 10% fetal bovine serum, 2 mM L-glutamine,
100 U/mL penicillin and 100 mg/mL streptomycin (Sigma-Aldrich) and maintained in a 5%
CO2 humidified atmosphere at 37 °C and grown until they reached 70% confluence for a
maximum of 20 passages.
After cell treatment with the various compounds, cells were gathered, washed with phosphate
saline buffer (PBS) and after centrifugation pellets were lysated in 20 mM Hepes-NaOH, pH
7.5, 1mM ethylenediaminetetraacetic acid (EDTA) and protease inhibitor cocktail (about 10
million cells/1mL of lysis buffer) and frozen in liquid nitrogen. Samples were stored at -
80 °C until use.
This article is protected by copyright. All rights reserved.

Cell viability assay
Log-phase cells were seeded into the 96-wells plate (15.000 cells/well) and were allowed for
the attachment overnight followed by the treatment with increasing concentrations of
compound 1 and methoctramine for 12 h and 24 h. Cellular proliferation was quantified with
colorimetric methods based on the metabolic reduction of the soluble yellow 3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) dye to its insoluble formazan
[48]. The absorbance was read on Titertek Multiskan Plus plate reader (Flow Laboratories,
Sutton, U.K.) using 540 nm wavelength.
Results were normalized to time point zero (t = 0 h), and the concentration inducing 50% of
cell death (DC50) was calculated with respect to DMSO treated controls, with Prism 7.0a
software.
Western blot analysis
Cells treated at the reported concentrations of 1 and pargyline for 3h and 24h were
trypsinized and lysed by addition of 20 mM Hepes (pH 7.5), 1 mM EDTA and protease and
phosphatase inhibitor cocktails and frozen in liquid nitrogen. Protein concentration was
determined by the Bradford method. 30 g of protein lysates were subjected to 9% SDS-
PAGE, blotted on Immobilon-P membranes (Merck-Millipore), processed in western blot
with the indicated primary antibodies and detected by enhanced chemiluminescent detection
system (ECL). Re-probing of a blot with new primary antibodies was done after 3 washing
with phosphate buffered saline with 0.1% Tween-20 (Sigma-Aldrich).
Immunostained bands were quantified by means of a Kodak-Image-Station 4000MM-PRO
and analysis with Carestream Molecular Imaging software (New-Haven, CT).
Differences among treatments were evaluated by analysis of variance (ANOVA) followed by
Bonferroni post-test, using Prism 7.0a; p<0.05 was considered significant.
Kinetic and statistical analysis
Steady-state kinetic parameters (Vmax and Km) were calculated by fitting the Michaelis-
Menten equation to the experimental data (initial rate of reactions vs substrate concentrations)
by non-linear regression analysis, with Sigma Plot software, version 9.0 (Jandel Scientific,
San Rafael, CA, USA) and the value of the kinetic parameter obtained from the best fit and
its S.E. are reported.
Time-dependent inhibition studies. Recombinant MAO A and MAO B (0.02-0.05 mg/mL) in
the Hepes buffer, were pre-incubated for periods from 0 to about 50 minutes, at 37 °C in the
presence and in the absence of inhibitor (control sample). After various time intervals of pre-
incubation (at least 5 times after mixing of enzyme and inhibitor), the residual MAOs enzyme
activity of samples was determined: an aliquot of sample was withdrawn and diluted in the
assay buffer (50-fold dilution) and, after 1 minute, substrate was added under “saturating”
condition, (p-Tyr 1 mM and BZA 10 mM for MAO A and B, respectively).
All time-dependent inhibition experiments were monitored relative to a control sample,
which was run exactly as the experiment except without inhibitor added (only an equivalent
volume of solvent used for the tested inhibitor, was added).
This article is protected by copyright. All rights reserved.

The enzyme activity of this control sample (v0) was set to 100% and for each time point the
ratio between residual enzyme activities (v+I) relative to the control sample (v0), was
calculated. The “apparent inactivation rate constant” values (kobs’), for each inhibitor
concentration, were calculated by linear regression analysis of the data of logarithm of
residual activity (ln (v+I/v0)) versus time. The kobs’ values were plotted versus [I] and, in the
case of a saturation behaviour, the Kitz and Wilson model, shown in the following equation,
was found to be appropriate to fit the time-dependent inhibition data (kobs’ vs inhibitor
concentration):
KEI
kinact
EI
(1)
EI*
E + I
In this scheme, a rapid reversible interaction between free enzyme (E) and inhibitor (I) is
followed by a slower, irreversible reaction, which transforms the reversible enzyme-inhibitor
complex (EI) into an irreversible enzyme inhibitor complex (EI*). The inactivation rate
constant (kinact) and the enzyme-inhibitor dissociation constant (KIE), were calculated by
fitting eqn. 2 to the data of “apparent inactivation rate constants” (kobs’) versus [I]0 (non-
linear regression analysis, with Sigma Plot software, version 9.0)
kinact[I]0
k '
=
obs
(KEI [I]0 )
(2)
If saturation kinetics is not observed, only the apparent bimolecular rate constant kinact/KIE
was calculated from the slope of the plot of k obs’ against [I]0.
Unless stated otherwise, the correlation coefficient for linear regression was 0.98 or greater.
All experiments were repeated independently at least three times, and the experimental data
were expressed as the mean ± standard deviation (SD).
Statistical analyses of data from cellular experiments were conducted using Student's t-test.
Values of *p < 0.05 or **p < 0.01 were defined as statistically significant. All analyses were
carried out with the GraphPad Prism program (GraphPad Software, Inc., San Diego, CA).
Error bars on all graphs consider 95% confidence intervals.
For calculating the IC50 value of compound 1 in the cellular system, curve fitting was
performed by non-linear regression analysis using the standard “dose-response” curve
equation and the Sigma Plot software.
Acknowledgments The Authors thank the “International Polyamine Foundation - ONLUS”
for the availability to look up in the Polyamines documentation and the Modelling Section
(MMS, University of Padova) for the kind support.
This research was supported by institutional grants from the University of Padova, Italy
(60A06-1707/15 to M.L.D.P. and COZZ_SID18_01 to G.C.) and by the institutional grant
from the University of Bologna, Italy "University of Bologna - RFO" to A.M.
Conflict of interest: the authors declare they have no conflicts of interests with the contents
of this article.
This article is protected by copyright. All rights reserved.

Author contributions
MLDP, AM (Anna Minarini) and GC, conceived the project; MLDP conducted all kinetic
investigations, analysed all the results and wrote the manuscript; GC conducted the
computational studies, analysed and discussed all the results and participated in writing the
manuscript; AM (Andrea Milelli) contributed to the synthesis of compounds and to writing
the paper; FZ carried out cell cultures and analysis of cellular data; SS contributed to design
and analyse all cellular experiments; EM carried out the synthesis of compounds; FU
contributed to the analysis and discussion of the results; MR contributed to the analysis of
results; AM (Anna Minarini) designed the compounds, carried out their synthesis, analysed
and discussed all the results and participated in writing the manuscript. All authors have
given approval to the final version of the manuscript.
References
1.
Edmondson DE & Binda C (2018) Monoamine Oxidases. Subcell Biochem 87, 117-
139.
2.
Binda C, Li M, Hubalek F, Restelli N, Edmondson DE & Mattevi A (2003) Insights
into the mode of inhibition of human mitochondrial monoamine oxidase B from high-
resolution crystal structures. Proc Natl Acad Sci U S A 100, 9750-9755.
3.
Alzheimer's disease (Review). Mol Med Rep 9, 1533-1541.
4. Youdim MB, Edmondson D & Tipton KF (2006) The therapeutic potential of
monoamine oxidase inhibitors. Nat Rev Neurosci 7, 295-309.
5. Bortolato M, Chen K & Shih JC (2008) Monoamine oxidase inactivation: from
pathophysiology to therapeutics. Adv Drug Deliv Rev 60, 1527-1533.
6. Finberg JP (2014) Update on the pharmacology of selective inhibitors of MAO-A and
Cai Z (2014) Monoamine oxidase inhibitors: promising therapeutic agents for
MAO-B: focus on modulation of CNS monoamine neurotransmitter release. Pharmacol Ther
143, 133-152.
7.
Neuropsychopharmacol Biol Psychiatry 69, 112-124.
8. Szoko E, Tabi T, Riederer P, Vecsei L & Magyar K (2018) Pharmacological aspects
Fisar
Z
(2016) Drugs related to monoamine oxidase activity. Prog
of the neuroprotective effects of irreversible MAO-B inhibitors, selegiline and rasagiline, in
Parkinson's disease. J Neural Transm (Vienna) 125, 1735-1749.
9.
Youdim MB, Bar Am O, Yogev-Falach M, Weinreb O, Maruyama W, Naoi M &
Amit T (2005) Rasagiline: neurodegeneration, neuroprotection, and mitochondrial
permeability transition. J Neurosci Res 79, 172-179.
10.
Weinreb O, Amit T, Bar-Am O & Youdim MB (2007) Induction of neurotrophic
factors GDNF and BDNF associated with the mechanism of neurorescue action of rasagiline
and ladostigil: new insights and implications for therapy. Ann N Y Acad Sci 1122, 155-168.
11.
Weinstock M, Kirschbaum-Slager N, Lazarovici P, Bejar C, Youdim MB & Shoham
S (2001) Neuroprotective effects of novel cholinesterase inhibitors derived from rasagiline as
potential anti-Alzheimer drugs. Ann N Y Acad Sci 939, 148-161.
12.
Ramsay RR & Tipton KF (2017) Assessment of Enzyme Inhibition: A Review with
Examples from the Development of Monoamine Oxidase and Cholinesterase Inhibitory
Drugs. Molecules 22.
13.
Weinreb O, Amit T, Bar-Am O & Youdim MB (2016) Neuroprotective effects of
multifaceted hybrid agents targeting MAO, cholinesterase, iron and beta-amyloid in ageing
and Alzheimer's disease. Br J Pharmacol 173, 2080-2094.
This article is protected by copyright. All rights reserved.

14.
Deshwal S, Di Sante M, Di Lisa F & Kaludercic N (2017) Emerging role of
monoamine oxidase as a therapeutic target for cardiovascular disease. Curr Opin Pharmacol
33, 64-69.
15.
Wu JB, Shao C, Li X, Li Q, Hu P, Shi C, Li Y, Chen YT, Yin F, Liao CP, et al.
(2014) Monoamine oxidase A mediates prostate tumorigenesis and cancer metastasis. J Clin
Invest 124, 2891-2908.
16.
Bonaiuto E, Minarini A, Tumiatti V, Milelli A, Lunelli M, Pegoraro M, Rizzoli V &
Di Paolo ML (2012) Synthetic polyamines as potential amine oxidase inhibitors: a
preliminary study. Amino Acids 42, 913-928.
17.
Bonaiuto E, Milelli A, Cozza G, Tumiatti V, Marchetti C, Agostinelli E, Fimognari C,
Hrelia P, Minarini A & Di Paolo ML (2013) Novel polyamine analogues: from substrates
towards potential inhibitors of monoamine oxidases. Eur J Med Chem 70, 88-101.
18.
A and B: role of cysteines. Mol Pharmacol 43, 888-893.
19. Hubalek F, Pohl J & Edmondson DE (2003) Structural comparison of human
Wu HF, Chen K & Shih JC (1993) Site-directed mutagenesis of monoamine oxidase
monoamine oxidases A and B: mass spectrometry monitoring of cysteine reactivities. J Biol
Chem 278, 28612-28618.
20.
Vintem AP, Price NT, Silverman RB & Ramsay RR (2005) Mutation of surface
cysteine 374 to alanine in monoamine oxidase A alters substrate turnover and inactivation by
cyclopropylamines. Bioorg Med Chem 13, 3487-3495.
21.
monoamine oxidase type A from human placenta. J Biol Chem 260, 13199-13207.
22. Melchiorre C, Yong MS, Benfey BG & Belleau B (1978) Molecular properties of the
Weyler W & Salach JI (1985) Purification and properties of mitochondrial
adrenergic alpha receptor. 2. Optimum covalent inhibition by two different prototypes of
polyamine disulfides. J Med Chem 21, 1126-1132.
23.
Quaglia W, Brasili L, Cristalli G, Giardina D, Picchio MT & Melchiorre C (1988)
Structure-activity relationships among benextramine-related tetraamine disulfides. Chain
length effect on alpha-adrenoreceptor blocking activity. J Med Chem 31, 1861-1866.
24.
Ueda Y, Melchiorre C, Lippert B, Belleau B, Chona S & Triggle DJ (1978) Molecular
properties of the adrenergic alpha-receptor. I--Structural requirements for specific covalent
occupancy by N,N'-bis--(5-aminopentyl)cystamine derivatives. Farmaco Sci 33, 479-495.
25.
Angeli P, Brasili L, Brancia E, Giardina D, Quaglia W & Melchiorre C (1985)
Structure-activity relationships among benextramine-related tetraamine disulfides at
peripheral alpha-adrenoreceptors. J Med Chem 28, 1643-1647.
26.
Giardina D, Brasili L, Melchiorre C, Belleau B & Benfey BG (1981) Molecular
properties of the adrenergic a-receptor. 6. Optimum carbon chain-length between the inner
nitrogen and the sulfur of tetramine disulfides. Eur J Med Chem - Chimica Therapeutica 16
569-571.
27.
Melchiorre C, Bolognesi ML, Minarini A, Rosini M & Tumiatti V (2010) Polyamines
in drug discovery: from the universal template approach to the multitarget-directed ligand
design strategy. J Med Chem 53, 5906-5914.
28.
Minarini A, Milelli A, Tumiatti V, Rosini M, Bolognesi ML & Melchiorre C (2010)
Synthetic polyamines: an overview of their multiple biological activities. Amino Acids 38,
383-392.
29.
Targets of Disease. Front Mol Neurosci 10, 167.
30. Rossin F, Villella VR, D'Eletto M, Farrace MG, Esposito S, Ferrari E, Monzani R,
Fra A, Yoboue ED & Sitia R (2017) Cysteines as Redox Molecular Switches and
Occhigrossi L, Pagliarini V, Sette C, et al. (2018) TG2 regulates the heat-shock response by
the post-translational modification of HSF1. EMBO Rep 19.
This article is protected by copyright. All rights reserved.

31.
Minarini A, Milelli A, Tumiatti V, Rosini M, Simoni E, Bolognesi ML, Andrisano V,
Bartolini M, Motori E, Angeloni C, et al. (2012) Cystamine-tacrine dimer: a new multi-
target-directed ligand as potential therapeutic agent for Alzheimer's disease treatment.
Neuropharmacology 62, 997-1003.
32.
of enzymes. Methods Enzymol 87, 469-499.
33. Hubalek F, Binda C, Khalil A, Li M, Mattevi A, Castagnoli N & Edmondson DE
Plapp BV (1982) Application of affinity labeling for studying structure and function
(2005) Demonstration of isoleucine 199 as a structural determinant for the selective inhibition
of human monoamine oxidase B by specific reversible inhibitors. The Journal of biological
chemistry 280, 15761-15766.
34.
disease research: a systematic review. Mol Neurodegener 12, 10.
35. Fitzgerald JC, Ufer C, De Girolamo LA, Kuhn H & Billett EE (2007) Monoamine
Xicoy H, Wieringa B & Martens GJM (2017) The SH-SY5Y cell line in Parkinson's
oxidase-A modulates apoptotic cell death induced by staurosporine in human neuroblastoma
cells. Journal of neurochemistry 103, 2189-2199.
36.
Zini M, Passariello CL, Gottardi D, Cetrullo S, Flamigni F, Pignatti C, Minarini A,
Tumiatti V, Milelli A, Melchiorre C, et al. (2009) Cytotoxicity of methoctramine and
methoctramine-related polyamines. Chem Biol Interact 181, 409-416.
37.
Inaba-Hasegawa K, Akao Y, Maruyama W & Naoi M (2013) Rasagiline and
selegiline, inhibitors of type B monoamine oxidase, induce type A monoamine oxidase in
human SH-SY5Y cells. J Neural Transm (Vienna) 120, 435-444.
38.
substitution and thiol oxidation-mediated pathways. Antioxid Redox Signal 18, 1623-1641.
39. Chiarini A, Minarini A, Budriesi R & Melchiorre C (1990) Molecular properties of
Nagy P (2013) Kinetics and mechanisms of thiol-disulfide exchange covering direct
the histamine H2-receptor. Covalent inhibition by tetraamine disulfides. Farmaco 45, 1001-
1011.
40.
continuous measurement of monoamine oxidase activity. Anal Biochem 253, 169-174.
41. Santillo MF, Liu Y, Ferguson M, Vohra SN & Wiesenfeld PL (2014) Inhibition of
Zhou M & Panchuk-Voloshina N (1997) A one-step fluorometric method for the
monoamine oxidase (MAO) by beta-carbolines and their interactions in live neuronal (PC12)
and liver (HuH-7 and MH1C1) cells. Toxicol In Vitro 28, 403-410.
42.
quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 72, 248-254.
43. Molecular Operating Environment (MOE 2010.10) CCG, Inc, 1255 University St.,
Suite 1600, Montreal, Quebec, Canada, H3B 3X3.
44. Edelsbrunner H (1955) Smooth surfaces for multi-scale shape representation, .
Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram
Foundations of Software Technology and Theoretical Computer Science Springer Berlin,
Heidelberg, 391–412
45.
Cozza G, Zanin S, Sarno S, Costa E, Girardi C, Ribaudo G, Salvi M, Zagotto G,
Ruzzene M & Pinna LA (2015) Design, validation and efficacy of bisubstrate inhibitors
specifically affecting ecto-CK2 kinase activity. Biochem J 471, 415-430.
46.
Cozza G, Rossetto M, Bosello-Travain V, Maiorino M, Roveri A, Toppo S, Zaccarin
M, Zennaro L & Ursini F (2017) Glutathione peroxidase 4-catalyzed reduction of lipid
hydroperoxides in membranes: The polar head of membrane phospholipids binds the enzyme
and addresses the fatty acid hydroperoxide group toward the redox center. Free Radic Biol
Med 112, 1-11.
47.
Friesner RA, Banks JL, Murphy RB, Halgren TA, Klicic JJ, Mainz DT, Repasky MP,
Knoll EH, Shelley M, Perry JK, et al. (2004) Glide: a new approach for rapid, accurate
docking and scoring. 1. Method and assessment of docking accuracy. J Med Chem 47, 1739-
1749.
This article is protected by copyright. All rights reserved.

48.
Twentyman PR, Fox NE & Rees JK (1989) Chemosensitivity testing of fresh
leukaemia cells using the MTT colorimetric assay. Br J Haematol 71, 19-24.
Supporting information
Additional supporting information may be found online in the Supporting Information section
at the end of the article.
Synthesis of compounds 10 and 13.
Fig. S1. Superimposition between compound 12 and pioglitazone in MAO B; covalent
docking of benextramine (1) against MAOs.
Fig. S2. NMR spectra of compounds 10 and 13.
1
A) H and 13C NMR spectra of compound 10
1
B) H and 13C NMR spectra of compound 13
This article is protected by copyright. All rights reserved.

Table 1. Molecular structures of benextramine (1) and its analogues.
Compound
Structure
Benextramine (1)
Modification of the length of methylene linkers between the amino groups
2
3
4
Elongation of the cystamine core
5
Replacemtent of the o-methoxybenzyl groups
6
7
8
9
Decrease in the number of amino groups
10
11
This article is protected by copyright. All rights reserved.

12
Replacement of the disulfide core
13
MET666
Cystamine
NEM
For each group of molecules, a brief description of the rational modifications is reported. The
reference compounds cystamine (core moiety) and NEM (-SH reactive compound) are also
shown.
This article is protected by copyright. All rights reserved.

Table 2. Preliminary screening of compounds 1-13 as potential inhibitors vs hr-MAOs.
Compound hr-MAO A
s
hr-MAO B
Vmax/Vmax0
0.90
hr-MAO A hr-MAO B
Vmax Vmax0
/
Km/Km0
Km/Km0
kobs’
(min-1)
kobs’
(min-1)
(100 M)
0.77
1.5
1.89
1
0.3500.039
0.2170.031
Effect of the length of methylene linkers between the nitrogen atoms
0.85
0.9
0.95
0.72
0.47
2.21
2.02
6.3
2
3
4
0.3870.012
0.3780.043
5.81.0
0.1970.070
0.1050.027
2.430.50
0.78
1.23
n.d.(a)
n.d. (a)
Effect of the elongation of the cystamine core
1.28 0.88 1.55
0.93
5
0.0540.005
0.1180.003
Effect of the replacement of the o- methoxybenzyl groups
1.0
1.04
1.0
1.20
1.20
1.10
1.7
1
1
1.2
2.02
1.20
1.6
0.060±0.06
0.0030.01
0.0780.013
0.2840.100
6
7
8
0.0390.003
0.0170.002
0.0770.010
0.0240.005
1.0
1
1.0
9
Effect of the decrease in the number of amine functions
0.76
0.87
0.64
1.08
0.8
0.92
1
1.25
1.14
1.87
10
11
12
0.2070.027
0.1660.017
0.4500.105
0.1760.020
0.33 0.10
0.94
0.77
0.9400.280
Effect of the replacement of the disulfide core
0.0030.002(b 0.0010.001
0.90
1.10
1.5
1
1
1.10
1.74
1
13
)
0.0030.002(b 0.0010.001
1
Met-666
CYSTAMINE
NEM
1
1
1
)
1
0.0020.001(b 0.0100.002
)
1
0.8
1.2
0.0270.003
0.0050.002
Steady-state kinetic parameters in the absence of compounds were: Vmax0,p-Tyr = 110 ± 9
nM/min and Km0,p-Tyr = 165± 30 for MAO A; Vmax0,BZA = 109 ± 11 nM/ and Km0,BZA = 280
± 45 for MAO B. The “apparent inactivation rate constant” values (kobs’) of the various
compounds for MAOs were determined by the linear regression analysis of data of residual
activity (ln(v+I/vo) vs time)(r≥0.98). Parameter ± S.D are reported (n=3)
(a)n.d.: not determinable, MAO A was immediately inactivated.
(b)
k ’ equivalent to that of control samples in absence of compound
obs
This article is protected by copyright. All rights reserved.

Table 3. Inactivation rate constant and inhibition constant values of the most effective
polyamine disulfides.
Selectivity
MAO A/MAO B
Compoun
d
MAO A
MAO B
kinact
KIE
(M)
kinact/KIE
(M-1min-
kinact
KIE
(M)
kinact/KIE
kinact kinact/KIE
(min-1)
(min-1)
(M-1min-1)
1
)
0.98±0.0
1
5.6±0.1
x103
0.25±0.0
1
40.0± 0.4
x103
175±14
n.d.
6.3±0.6
n.d.
3.9:1
n.d.
1:7.1
2.8:1
1:27
1
4
9
51 ±7
18.3±6
x103(a)
n.d.
n.d.
x103(a)
0.04±0.0
1
0.5±0.1
x103
0.19±0.0
1
13.8±1.0
x103
77±25
14±2
1:4.8
0.40±0.0
1
4.0±0.5
x103
0.22±0.0
1
7.8± 1.7
x103
97±9
20±4
30±5
13±2
1:1.9
1:3.9
10
12
1.8:1
1:2.5
0.57±0.0
4
28.0±7.0 1.42±0.0
x103
109.0±17.0
x103
6
The kinetic parameters of inactivation were calculated fitting equation 2 to experimental data
(k’obs of inactivation vs inhibitor concentrations), by non-linear regression analysis (parameter
± standard error values are shown).
n.d.: not determinable, saturation was not reached in the explored compound 4 concentration
range (2-100 M)
(a)This parameter is the slope value obtained by the linear regression of the data of the plot
k’obs vs compound 4 concentrations.
This article is protected by copyright. All rights reserved.

Table 4. Irreversibility of the inactivation of MAOs by compound 1
RECOVERED MAO ACTIVITY
( %)
A) Effect of dialysis on MAOs inactivated by compound 1
Sample
hr-MAO A
hr-MAO B
100a
100b
Control
0
41±5
Compound 1-inactivated MAO
B) Effect of DTT on MAOs inactivated by compound 1
100
100
11011
364
Control
Control + DTT
876
0
0
Compound 1-inactivated MAO
Compound 1- inactivated MAO after DTT
244
A) Effect of dialysis: MAOs were inactivated by compounds 1 (200 M) and then extensively
dialyzed; after dialysis, the recovered MAO activity was measured. A control MAO samples
was run under the same experimental conditions.
B) Effect of DTT: MAO A and MAO B were inactivated by pre-incubation with compound 1
(200 M for 15 min at 36°C); then DTT 10 mM was added. After additional 20 min of
incubation, samples were diluted and the residual MAO activity was tested, using the
kynuramine assay method (final concentration in the assay: compound 1 5 M and DTT 0.25
mM)
Means ±SD are shown (n=3).
-1
a MAO A specific activity: 30 nmol min-1 mgprotein
b MAO B specific activity: 16 nmol min-1 mgprotein
-1
This article is protected by copyright. All rights reserved.

Table 5. Protective effect of a competitive inhibitor and of the substrate on inactivation
of MAOs by compounds 1 and its derivatives.
A) Effect of the competitive inhibitor Isatin
kobs’
kobs’+ISATIN
[Isatin]a
M)
Enzyme
[Compound]
kobs’ (min-1)
- isatin + isatin
[1]
100
12
5.8
3.8
MAO A
MAO B
100 M
100 M
[4]
0.3500.039 0.0600.005
0.2170.031 0.0570.007
100
12
5.8±1.4
3.0±0.3
2
MAO A
MAO B
100 M
100 M
2.35±0.52
0.52±0.08
4.5
B) Effect of the substrate
Substrate
kobs’ (min-1)
Enzyme
[Compound]
[4]
+ substrate
- substrate
p-Tyr
MAO A
MAO B
5 M
5 M
[12]
0.920.14
0.530.07
No inactivation
BZA
0.310.03
No inactivation
No inactivation
p-Tyr
MAO A
MAO B
5 M
5 M
0.3490.08
0.1050.01
BZA
A) Effect of the competitive inhibitor isatin on inactivation by compound 1 and 4.
Inactivation experiments were performed under standard conditions, in the presence and
absence of isatin.
B) Effect of substrate on inactivation of MAOs by compounds 4 and 12. Inactivation
experiments were performed under the standard conditions, in the presence and absence of
substrate ([S]>>Km: 1 mM p-Tyr for MAO A and 10 mM BZA)
The “apparent inactivation rate constant” values (kobs’) for MAOs were determined by linear
regression analysis of data of residual activity (ln(v+I/vo)) vs time. Error bars represent the SD
of the parameters (n = 3 replicates).
a)Isatin concentrations were 6-7 times higher than its corresponding KI for the free enzyme
([E-Isatin]/[E]>6). In our experimental conditions: Ki = 16±3 M and Ki = 2.3±0.5 M for
MAO A and MAO B, respectively, in agreement with those previously reported [33].
This article is protected by copyright. All rights reserved.

Scheme 1. i) a. o-OMe-benzaldehyde, EtOH, reflux, 5 h, b. NaBH4, EtOH, rt, , 16h,79%
yield; ii) Boc2O, CH2Cl2, rt, 16 h, quantitative yield; iii) tosyl chloride, Et3N, DMAP,
CH2Cl2, 0 °C, 1.5 h, 81% yield; iv) cystamine (for 17) or ethanolamine (for 18), DMF, Et3N,
rt, 3 days, 36-60% yield; v) 4M HCl/dioxane, 0 °C, 40 min, 81% yield. Boc = -COOC(CH3)3;
Ts = -SO2C6H4CH
This article is protected by copyright. All rights reserved.

Fig. 1. Structures of the prototype of the tetraamine disulfides, benextramine (1), its carbon
analogues (Met-666 and methoctramine) and of the previously studied MAO inhibitor ITC-
BzDD [17].
Fig. 2. Effect of compounds 1, 4, 9 and 12 on kinetic parameters (Km and Vmax) of hr-MAOs.
Lineweaver-Burk plots of oxidative deamination of substrate catalysed by hr-MAO A (A)
and hr-MAO B (B), in the absence (●) and presence of 100 M compounds 1 (●),4 (■), 9
(▲) or 12 (▲). Steady-state kinetic experiments were carried out without pre-incubating the
enzyme with the compound. Solid lines were calculated by linear-regression analysis of data
(r > 0.99).
This article is protected by copyright. All rights reserved.

Fig. 3. Time dependent inactivation of MAOs by compound 1
Decay of MAO activity in the presence of 100 M compound 1 vs pre-incubation time; solid
lines are the results of the best-fitting of the equation for an exponential decay at two
parameters (V(t=0) and kobs’) to the experimental data (non-linear regression analysis, with
Sigma Plot software, version 9.0; r ≥ 0.99). No decay of MAO activity was found in control
samples (dashed lines).
p-Tyr 1 mM and BZA 10 mM were used as substrate to assay the residual MAO A and MAO
B activity, respectively. Error bars represent the S.D. of the data (n = 3 replicates).
This article is protected by copyright. All rights reserved.

This article is protected by copyright. All rights reserved.