ꢀ
ꢀ
J. Staron, W. Pietrus, R. Bugno et al.
European Journal of Medicinal Chemistry 220 (2021) 113533
HBD [28]. The QTAIM analysis showed that X-HBD interaction be-
tween Cl and hydroxyl group of T497 has the highest values of
energy stabilization (ꢁ1.7, ꢁ5.0 kcal/mol, for 7 and 42, respec-
tively). This feature of the chlorine atom translate to the formation
of a very strong X-HBD interactions, which were found to be
responsible for the high binding affinity of 3,4-diCl analogs. In spite
of the complex nature of X-HBD, pure HB and XB can also be
observed in the analyzed L-P complexes. In that case, importance of
XBs increase along with the halogen atom size, with simultaneous
decreased contribution of HBs (only 6% for iodine analog 9)
(Table S5, Supplementary Information). For all of the discussed
interactions the dispersion energy is a vital component, however its
share is strongly related to the type of the interaction. It is worth to
note at this point, that the obtained QM refined poses constitute of
the most optimum ligand-receptor mutual orientation and are
probably poorly populated in the real time ligand-receptor in-
teractions. Despite that, this type of analysis still allows to deeply
investigate the binding of a ligand to the receptor.
T497. This type of interaction was found to be a hybrid of XB and HB
interactions, with the latter accounting for a higher share of the
interaction energy. Subsequently, the use of an XSAR analysis
allowed to translate this to a general rule governing the binding of
SERT ligands. Especially interesting was the observation that SERT
ligands with heavier halogens bind via a markedly different,
“switched” binding mode than the parent trifluoromethyl analogs.
To the best of our knowledge, no paper to date has exclusively re-
ported two distinct binding modes for SERT ligands. Finally, the
given hypothesis of halogen bond type interactions with the re-
ceptor was confirmed by the experimental measurement of the
affinity constant for 3,4-dichloro analogs of fluoxetine and fluvox-
amine, which were the most active among all of the synthesized
compounds. The presented manuscript states as an example of the
successful use of molecular modeling for the analysis and optimi-
zation of compound activity. The utilized tool, XSAR analysis, pre-
viously used for the 5-HT7R [22] and D4R ligands [22,23], proved to
be universal and applicable for a broader type of biological target.
We anticipate that such a methodology could be successfully uti-
lized for the structure optimization of any halogen-containing
biologically active compound.
All of the above observations lead to the hypothesis that starting
from bromine, the halogen-receptor interaction becomes a domi-
nant factor influencing the activity of these two SSRIs. Chlorine
analogs 6 and 35 also possess similar binding affinity (SERT
Ki ¼ 55 nM and Ki ¼ 61 nM, respectively), but the effect is less
profound. The trifluoromethyl group interactions with the protein
are weak and indirect; thus, structural differences of fluoxetine and
Declaration of competing interest
The authors declare that they have no known competing
financial interests or personal relationships that could have
appeared to influence the work reported in this paper.
fluvoxamine produce
a very distinct binding affinity (SERT
Ki ¼ 31 nM and Ki ¼ 458 nM, respectively). The ability of heavier
halogen atoms to form a direct interaction causes different chem-
ical scaffolds of these compounds to lose their decisive impact on
the activity. The molecular modeling experiments showed two
distinct binding modes (Fig. 3). Here, the increase in activity
brought by the introduction of a halogen atom can be explained on
the ground of the formed halogen atom interactions, which are
possible only in the alternative binding mode (Fig. 3). Most
frequently, these interactions were formed with either T497 or
E493. To date, only one halogen bond in the SERT complex has been
reported. In 2019, Abramyan et al. showed a complex of bromo
paroxetine with SERT, where the bromine atom pointed in the di-
rection of the E493 and T497 backbone oxygens. However, no
further analysis was performed [31]. Therefore, we performed an
XSAR analysis, which showed explicitly that E493 was the most
frequently targeted amino acid in terms of halogen bonding with
SERT. Moreover, the proximity of T497 and E493 allowed for the
formation of two strong interactions with two vicinal halogen
atoms. Such an optimal arrangement in some cases resulted in a
spectacular thousand-fold increase in biological activity (Table 4).
This finding supported our hypothesis for the alternative binding
mode of SERT ligands bearing heavier halogen atoms. Finally, two
synthesized 3,4-dichloro analogs of fluoxetine and fluvoxamine
confirmed the theoretical assumption and showed a marked in-
crease in activity. Fluoxetine analog 42 was 3-fold more active and
fluvoxamine analog 46 was 1.7-fold more active than the best-
binding 4-iodo analogs.
Acknowledgments
The study was partially supported by grant SONATA UMO-2016/
21/D/NZ7/00620 from the Polish National Science Centre and
POWR.03.02.00e00-I013/16 from the European Union. The study
was performed with the use of PLGrid infrastructure. WP ac-
knowledges the support of InterDokMed project no.
POWR.03.02.00e00-I013.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
Methodology
Structureeactivity relationship datasets for halogenated analogs
An algorithm to find all pairs containing halogenated and cor-
responding unsubstituted structures (called the XSAR library) was
developed and used for the 5-HT7 and D4 targets in our previous
studies [22]. To describe the influence of halogenation on the bio-
logical activity of the unsubstituted (parent) molecule, the Xeffect
parameter was calculated as an activity (extracted from the
ChEMBL database during the generation of the XSAR library) ratio
of the parent compound to its halogenated derivative (an Xeffect
between 0 and 1 denotes a decrease in the activity upon haloge-
nation, and an Xeffect greater than 1 means the fold of activity
increased after halogen substitution).
5. Conclusions
In the present study, a series of fluoxetine and fluvoxamine
analogs with varying patterns of halogen substitution were syn-
thesized, and their biological activities were measured. Among the
synthesized initial series of compounds, a classical order of
increasing activity from chlorine to iodine was observed. Intro-
duced additional fluorine atoms did not have a substantial influ-
ence on the affinity for the receptor. Subsequent thorough in silico
analysis of the binding modes indicated the possibility of the for-
mation favorable X-HBD interactions with the residues E493 and
Identification of halogen bonding hot spots for SERT
Privileged amino acids (i.e., hot spots) for halogen bonding were
identified using a procedure including the following steps: clus-
tering the halogenated analogs representing each XSAR set, using
the centroids of the clusters to tune the SERT binding site by an
induced-fit docking procedure, QPLD docking of the XSAR library to
9