Coumarin-Tagged Dinuclear Trithiolato-Bridged Ruthenium(II)?Arene Complexes: Photophysical Properties and Antiparasitic Activity

Article abstract of DOI:10.1002/cbic.202000174

The synthesis, characterization, photophysical and biological properties of 13 new conjugate coumarin-diruthenium(II)?arene complexes against Toxoplasma gondii are presented. For all conjugate organometallic unit/coumarins, an almost complete loss of fluorescence efficacy was observed. However, the nature of the fluorophore, the type of bonding, the presence and length of a linker between the coumarin dye and the ruthenium(II) moiety, and the number of dye units influenced their biological properties. The in vitro activity against a transgenic T. gondii strain grown in human foreskin fibroblasts (HFF) leads to IC₅₀ values for T. gondii β-gal from 105 to 735 nM. Of note is that nine compounds displayed lower IC₅₀ than the standard drug pyrimethamine. One compound applied at its IC₅₀ did not affect B-cell proliferation but had an impact on T-cell proliferation in murine splenocyte cultures. Transmission electron microscopy of T. gondii β-gal-infected HFF showed that treatment predominantly affected the parasites’ mitochondrion.

Full text of DOI:10.1002/cbic.202000174

Combining Chemistry and Biology
Accepted Article
Title: Coumarin-Tagged Dinuclear Trithiolato-Bridged Ruthenium(II)-
Arene Complexes – Photophysical Properties and Antiparasitic
Activity
Authors: Oksana Desiatkina, Emilia Păunescu, Martin Mösching,
Nicoleta Anghel, Ghalia Boubaker, Yosra Amdouni, Andrew
Hemphill, and Julien Furrer
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemBioChem 10.1002/cbic.202000174
Link to VoR: https://doi.org/10.1002/cbic.202000174
01/2020

10.1002/cbic.202000174
ChemBioChem
Coumarin-Tagged Dinuclear Trithiolato-Bridged Ruthenium(II)-Arene Complexes –
Photophysical Properties and Antiparasitic Activity
Oksana Desiatkina[a], Emilia Păunescu[a]*, Martin Mösching,[a] Nicoleta Anghel[b], Ghalia
Boubaker[b], Yosra Amdouni[b], Andrew Hemphill[b]*, Julien Furrer[a]*
[a]Department of Chemistry and Biochemistry, University of Bern, Freiestrasse 3, 3012 Bern,
Switzerland.
[b]Institute of Parasitology, Vetsuisse Faculty, University of Bern, Länggass-Strasse 122, 3012
Bern, Switzerland.
paunescu_emilia@yahoo.com
julien.furrer@dcb.unibe.ch
andrew.hemphill@vetsuisse.unibe.ch
*Corresponding authors:
Chemistry:
Emilia Păunescu, paunescu_emilia@yahoo.com
Julien Furrer, Tel. +41-31-6314383, julien.furrer@dcb.unibe.ch
https://furrer.dcb.unibe.ch/
ORCID: 0000-0003-2096-0618
Parasitology:
Andrew Hemphill, Tel. +41-31-6312384, Fax +41-31-6312477,
andrew.hemphill@vetsuisse.unibe.ch
http://www.ipa.vetsuisse.unibe.ch/ueber_uns/personen/prof_dr_hemphill_andrew/index_ger.html
ORCID: 0000-0002-0622-2128
1
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Abstract
The synthesis, characterization, photophysical and biological properties against T. gondii of 13 new
conjugates coumarin-di-ruthenium(II)-arene complexes are presented. For all conjugates
organometallic unit: coumarin an almost complete loss of fluorescence efficacy was observed.
However, the nature of the fluorophore, the type of bonding, the presence and length of a linker
between the coumarin dye and the ruthenium(II) moiety, and the number of dye units did influence
their biological properties. The in vitro activity against a transgenic T. gondii strain grown in human
foreskin fibroblasts (HFF) leads to IC50 values for T. gondii β-gal ranged from 105 nM to 735 nM.
Of note, 9 compounds displayed lower IC50 than the standard drug pyrimethamine. One compound
applied at its IC50 did not affect B cell proliferation but had an impact on T cell proliferation in murine
splenocyte cultures. TEM of T. gondii β-gal infected HFF showed that treatment predominantly
affected the parasites’ mitochondrion.
Keywords: cationic trithiolato dinuclear ruthenium(II)arene complexes, coumarin, Toxoplasma
gondii, antiparasitic efficacy, cytotoxicity, TEM.
2
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Introduction
The use of ruthenium complexes as potential chemotherapeutics is an active area of research for
almost two decades.1-3 Developed initially as a potential alternative to platinum based anticancer
drugs,4, 5 ruthenium(II)-arene complexes were also considered for other pharmacological properties,
particularly as antiparasitic,6-13 and antibacterial compounds.14-17 A special class of ruthenium(II)-
arene complexes is constituted by symmetric18, 19 and ‘mixed’20 cationic trithiolato-bridged dinuclear
ruthenium(II)-arene complexes (general formula [(η6-arene)2Ru2(μ2-SR)3]+ and [(η6-arene)2Ru2(μ2-
SR1)2(μ2-SR2)]+, respectively). The high cytotoxicity against human cancer cells shown by this type
of compounds and, more interestingly, their apparent ability to circumvent platinum-drug
resistance,21 encouraged the development of several libraries of complexes containing this scaffold
as potential biological active compounds. For example, IC50 values as low as 30 nM against A2780
(human ovarian cancer) cells and also against their cisplatin-resistant variant A2780cisR were
measured for compounds B and C (Figure 1), while the less lipophilic complex A exhibited lower
cytotoxicity (IC50 of 130 nM and 80 nM on A2780 and A2780cisR, respectively).
Figure 1. Structure of cationic trithiolato di-ruthenium(II)-arene complexes presenting high
cytotoxicity against cancer cells and efficacy against various parasites.
The mechanisms by which these trithiolato dinuclear ruthenium complexes exert their cytotoxicity
are still not clearly determined. Due to their lipophilic nature, these trithiolato dinuclear ruthenium
complexes are only sparingly soluble in water. In contrast, the di-ruthenium trithiolato moiety is
stable, does neither hydrolyze nor interact with biomolecules such as amino acids (other than
cysteine), nucleotides, or glucose.21 This inertness towards hydrolysis and ligand substitution can be
attributed to the lack of the typical very labile Ru(II)-Cl bonds or carboxylate ligands that are present
in many other Ru(II)-arene complexes. Nevertheless, weak binding to various proteins (human serum
albumin, (HsA), transferrin (Tf), cytochrome c (Cyt c), ubiquitin (Ub), and myoglobin (Mb)) was
demonstrated, albeit through hydrogen bonds and π-π interactions and not through covalent bonds.21
Interestingly, catalytic oxidation of bio-relevant reducing agents including cysteine (Cys) and
glutathione (GSH) to form cystine and GSSG, respectively, was demonstrated.22 Although no
3
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
correlation between the in vitro cytotoxicity and the catalytic activity on the oxidation reaction of
glutathione could be observed,19, 22 it is nevertheless considered that this process can be at least
partially responsible for the decreased cancer cell survival.
Recently, inductively coupled plasma mass spectrometry (ICP-MS) showed that complexes A and C
specifically target the mitochondrion in A2780 (human ovarian cancer) cells, with up to 97% of the
Ru content being found in the mitochondrial fraction.23 Complex C was taken up by A2780 cells
more efficiently than A, which parallels with the respective lipophilicity of the two compounds.
Lately, trithiolato di-ruthenium compounds were reported to exhibit interesting antiparasitic
properties against Toxoplasma gondii,24 Neospora caninum23 and Trypanosoma brucei.25 Thus,
complex A presents an IC50 value of 34 nM against the apicomplexan parasite Toxoplasma gondii
cultured in HFF (human foreskin fibroblast) monolayers,24 while compound B exhibits an IC50 of
4 nM against Trypanosoma brucei bloodstream forms.25 Viability of non-infected HFF cells treated
with 2.5 µM of complex A was decreased to 63%, however complex B did not affect the vitality when
applied at the same concentration.24 In Toxoplasma, Neospora and in Trypanosoma, TEM
(transmission electron microscopy) demonstrated that trithiolato di-ruthenium compounds affected
the structural integrity of the mitochondrion after few hours of treatment, and led to a more
pronounced destruction of tachyzoites at later timepoints.24 In T. brucei, these compounds altered the
mitochondrial membrane potential, while other organelles and structural elements of the parasites
remained largely unaffected.25
The intracellular fate of these trithiolato dinuclear ruthenium complexes, and how they exert
anticancer and antiparasitic effects, are unknown, and thus investigating their cellular traffic and
possible mechanisms of action has become a priority. In vitro/in vivo tracking of organometallic drugs
using traceable compounds represents a promising approach.26, 27 Various conjugates obtained by
anchoring an imaging probe to the biologically active metal-organic moiety were shown to facilitate the
intracellular tracing of the metal-based drugs. 28-30 However, the modification of a drug with a fluorophore
tag can strongly influence its physico-chemical properties, and thus activity and behavior. It was already
shown that anchoring different fluorophores (coumarin, BODIPY, porphyrin) on the same therapeutic
organometallic moiety influences cellular internalization and accumulation.28-30 Thus, the optimization of
the conjugate structure (the organometallic drug onto which the fluorophore probe is already
anchored/tethered) should be considered from the beginning.
Microscopy studies of BODIPY,29, 31 coumarin32 and naphthalimide33-tagged ruthenium conjugates
demonstrated that these conjugates could be applied for investigating the subcellular localization of
half-sandwich ruthenium(II)-arene complexes. Due to their good photophysical properties, chemical
stability, and ease of synthetic modifications, coumarin derivatives represent an important class of
4
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
fluorescent probes for biological imaging.34, 35 Additionally, compounds presenting a coumarin
scaffold display broad pharmacological activities and their potential as anticancer agents has received
a lot of interest.36-40
In the particular case of traceable organometallic drugs, two strategies can be considered: (i) direct
coordination of the (coumarin) fluorophore to the metal center41, 42 or (ii) anchoring of the dye to
more elaborated ligands.32, 43, 44 The nature of the metal center as well as the type of the ligand can
strongly affect the photophysical properties of the hybrid molecule. Some data on the development
of trackable anticancer agents based on metal complexes were recently reviewed.26,45 Conjugates
combining metalorganic units with covalently linked coumarins have been shown to be versatile tools
for imaging in the case of various fluorophore-tagged platinum,43, 44, 46 ruthenium,32 gold28, 47, 46 or
iridium48 complexes. Some examples of coumarin modified organometallic compounds are presented
in Figure 2.
Figure 2. Structure of various organometallic moiety-coumarin conjugates.
Coumarin-tagged ruthenium(II)-arene complex D was shown to be useful for bioimaging.32 Laser
scanning confocal microscope (LSCM) showed that compound D was internalized by HCT-116 cells
(human colorectal cancer, IC50 = 66 μM), causing intracellular fluorescence after 3 h of incubation
when applied at 30 μM. The process occurred in a concentration-dependent manner. Similar results
were obtained with Pt(IV) conjugate E,44 which was considered as potential tracking agent due to the
presence of a coumarin moiety. Nevertheless, neither for conjugate D or E the physical properties (in
terms of fluorescence quantum yield ΦF) were reported.
Compound F is a three-in-one hybrid containing both an androgen receptor (AR) binding ligand and
a coumarin unit connected to a Pt(IV) moiety.46 The fluorescence quantum yield of F was weaker
5
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
than that of the respective coumarin ligand (ΦF = 0.072 vs 0.32). F was effectively internalized and
visualized by LSCM in LNCaP (AR+) cells (androgen-sensitive human prostate adenocarcinoma
cells overexpressing androgen receptor) after 4 h of incubation when applied at 20 μM. Derivative G,
a Pt(II) complex of a coumarin modified diethylenetriamine, was developed as a selective probe,
fluorescent agent and inorganic medicinal agent.43 However, the fluorescence intensity of the
coumarin moiety was reduced by 90% in this conjugate, thus no microscopy studies on cellular uptake
were reported. Several gold(I) complexes bearing coumarin units were described.28,47, 49 For example
gold(I) carbene complex H displayed poor photophysical properties (fluorescence quantum yield,
ΦF = 2%), the fluorescence quenching being attributed to a photo induced electron transfer between
the coumarin and the carbene.47 Nevertheless fluorescence confocal microscopy demonstrated the
uptake of H in the nuclei of A2780 (human ovarian cancer) cells (IC50 = 12 µM) at a concentration
of 50 µM. In contrast, gold(I) phosphine complex I showed interesting photophysical properties
(ΦF = 83%), and two-photon fluorescence microscopy of MDA-MB-231 (human breast
adenocarcinoma) cells showed this compounds to accumulate at the plasma membrane as small
18, 50
aggregates.28 Interestingly, reports on coumarin trithiolato decorated compounds such as J
focused only on the anticancer activity of the compounds and not on their photophysical properties.
The easy synthesis of symmetric and mixed trithiolato dinuclear ruthenium complexes presenting
functional groups that further allow ‘chemistry on the complex’ opens the possibility to prepare
various conjugates with the molecules of interest e.g. fluorophores, drugs or biomolecules (amino
acids, fatty acids, polyamines, carbohydrates). Hybrid molecules with anticancer-drug chlorambucil51
and short peptides52 were already reported as potential strategy to enhance the anticancer activity.
In this study we report the design and synthesis of a library of new trithiolato-bridged dinuclear
ruthenium(II)-arene organometallic conjugates in which coumarin fluorophore moieties were
anchored as pendant arms on the bridge thiol(s). A systematic assessment of the influence of various
structural features of the conjugates (nature of the fluorophore, type of bonding (ester vs amide),
presence and length of a linker between the coumarin dye and the binuclear ruthenium(II) moiety,
number of dye units) upon the photophysical and biological properties was performed. All
compounds of interest (coumarin-labelled conjugates, non-modified thiolato-bridged dinuclear
ruthenium(II)-arene complexes and free dyes) were evaluated with respect to their activity against
the transgenic strain T. gondii β-gal grown in HFF host cells. In addition, the cytotoxicity in non-
infected HFF was assessed. For selected compounds, the potential to impair immunity was assessed
using B and T cell proliferation assays, and TEM was carried out to evaluate structural alterations in
treated parasites.
6
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Results and Discussion
Synthesis and characterization of the compounds
In this study, the influence of various structural features of the coumarin conjugates (nature of the
fluorophore, type of bonding (ester vs amide), presence and length of a linker/spacer between the dye
and the di-ruthenium moiety, number of the fluorophore units) upon the photophysical and biological
properties was evaluated.
Coumarin derivatives present a wide structural diversity due to the diff erent substitution possibilities
within their basic motif that contains a benzene ring fused to an α-pyrone. Analogues substituted with
electron-donating groups in position 7 such as 7-methoxy and 7-diethylamino coumarins are
frequently used fluorophores. The two coumarin dyes considered for this study, namely 7-
(diethylamino)-2-oxo-2H-chromene-3-carboxylic acid (Dye1-CO2H) and 11-oxo-2,3,6,7-tetrahydro-
1H,5H,11H-pyrano[2,3-f]pyrido[3,2,1-ij]quinoline-10-carboxylic acid (also known as butterfly
coumarin 343, Dye2-CO2H) present easy derivatizable carboxylic group. The use of these particular
fluorophores is justified by their good quantum yields, appropriate absorption/emission wavelengths
(different emission colors), lipophilicity, lack of net ionic charge, photostability, and relatively small
size.34, 35 It is well known that the photophysical properties of the coumarin derivatives can be tuned
with small changes in the substituents and their position.53 If these two coumarins present very similar
structure, previous reports associated the blockage of the free rotation of the C-N bond in Dye2-
CO2H, which constrains the nitrogen lone pair to a maximal interaction with the aromatic rings, with
an improvement of the photophysical properties compared to Dye1-CO2H.54, 55 Nevertheless, if this
interaction results in higher quantum yield in aqueous solution for Dye2-CO2H than the
diethylamino-coumarin Dye1-CO2H, the coumarin substituent is more bulky and hydrophobic.
A SAR study56 performed on a library of trithiolato-bridged dinuclear ruthenium(II)-arene
compounds showed that the nature of the substituents in the para position of the other two thiol
ligands influence the biological properties of the complexes. Some of the compounds were developed
also in an analogue series in which the bulky hydrophobic tBu substituents were replaced by
hydrophobic polar CF3 groups, in order to determine the role of this type of structural modifications
upon the biological activity of the conjugates.
The first series of compounds were functionalized with one coumarin unit anchored to one of the
bridged thiols, providing a ratio metal-organic moiety: dye of 1:1. To access this type of analogues,
the previously reported56 dinuclear trithiolato ruthenium(II)-arene intermediates, bearing one hydroxy
2a/b, amino 3a/b and carboxylic acid group 4a/b, were prepared starting from the ruthenium dimer
([Ru(η6-p-MeC6H4Pri)Cl]2Cl2) in two synthetic steps (Scheme 1).20 The obtainment of this type of
7
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
mixed trithiolato complexes is facilitated by the easy synthesis of the dinuclear dithiolato
intermediates 1a/b which were isolated in high yields. In the case of the carboxylate functionalized
trithiolato complexes 4a/b a mixture of CH2Cl2/acetone (10:1 (v/v) was used as solvent to avoid
esterification side reaction catalyzed by the chlorhydric acid resulted during the reaction).
Scheme 1. Synthesis of the dinuclear dithiolato 1a/b and OH, NH2, CH2CO2H functionalized
trithiolato ruthenium(II)-arene intermediates 2a/b, 3a/b and 4a/b.
If intermediates 2a/b and 3a/b could be directly coupled with the two dyes considered Dye1-CO2H
and Dye2-CO2H via ester and, respectively, amide bonds (Scheme 3), the carboxylate functionalized
trithiolato complexes 4a/b offer the possibility to introduce linkers of variable length between the
organometallic moiety and the dye. To this end, the linker modified coumarins 8, 9 and 10 were
synthesized in two steps staring from the corresponding coumarins (Dye1-CO2H or Dye1-CO2H)
and the appropriate monoprotected diamine (Scheme 2).
Scheme 2. Synthesis of the coumarin precursors containing an amino spacer 8, 9 and 10.
The coupling reactions of 2a/b and 3a/b with fluorophores Dye1-CO2H and Dye2-CO2H afforded
the ester 11, 12a/b and amide 13, 14a/b conjugates in medium to good yields (65-99%) (Scheme 3).
The coumarin hybrids 15, 16a/b and 17a/b containing an amino spacer were obtained in reasonable
yields (12-69%) by reacting the carboxylate functionalized trithiolato complexes 4a/b with the
judiciously functionalized fluorophores 8, 9 and 10 (Scheme 4). The esterification reactions were
realized using 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI) as coupling agent and 4-
8
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
dimethylaminopyridine (DMAP) as catalyst. For the amide bond formation, the coupling reactions
were conducted in the presence of HOBt (1-hydroxybenzotriazole) and EDCI as coupling agents and
DIPEA (N,N-diisopropylethylamine) as basic catalyst. As previously reported in other conjugation
reactions run on the trithiolato di-ruthenium complexes, the organometallic scaffold is stable to ligand
exchange in the reaction conditions used.51,52
Scheme 3. Synthesis of the coumarin-based ester 11, 12a/b (top) and amide 13, 14a/b (bottom)
conjugates.
9
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Scheme 4. Synthesis of the coumarin conjugates 15, 16a/b and 17a/b containing an amino spacer.
A second family of conjugates with organometallic unit: fluorophore (dye) ratios of 1:2 and 1:3 was
also synthesized (Scheme 5) and their photophysical and biological properties were evaluated.
10
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Scheme 5. Synthesis of conjugates 20 and 22 bearing two and, respectively, three fluorophore units.
Conjugate 20 functionalized with two coumarin units was synthesized in three steps starting from
([Ru(η6-p-MeC6H4Pri)Cl]2Cl2). If the synthesis of the dithiolato intermediates 1a/b, obtained by the
complexation of less active benzylic thiols, could be easily controlled by the judicious use of the
reagents ratio, the use of more reactive aromatic thiol of 4-mercaptophenol led to complex mixture
of the desired dithiolato product 18, accompanied by trithiolato and monothiolato complexes, as well
as unreacted ruthenium dimer. Nevertheless, the next two steps: (i) complexation of a third thiol with
the obtainment of mixed trithiolato compound 19 bearing two hydroxy groups, and (ii) subsequent
esterification reaction with Dye1-CO2H leading to di-coumarin conjugate 20 proceed smoothly, with
isolated yields of 96% and 75%, respectively (Scheme 5).
The tri-substituted conjugate 22 was obtained in two steps, the synthesis of the symmetric trihydroxy
intermediate 21 (previously reported57) followed by the anchoring of three coumarin dye units, and
was isolated in good yield (61%).
11
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
All compounds were characterized by 1H, 19F (when appropriate), and 13C NMR spectroscopy, mass
spectrometry, and elemental analysis (see Supporting information for full details).
For example, the 1H NMR spectra of hydroxy 2a and amine 3a intermediates differ from those of the
respective ester and amide conjugates 11/12a and 13/14a. After anchoring the coumarin unit via an
ester bond, for 11 and 12a the two protons in the α position to the ligand C-O are observed at slightly
lower frequencies, by ca. ΔδH ≈ 0.05 ppm, while the protons in the β position with respect to the C-
O bond are shifted towards higher frequencies ΔδH ≈ 0.32 ppm. In the 13C NMR spectra of esters 11
and 12a the corresponding C atoms α to the ligand C-O are observed at higher frequencies than in
hydroxy intermediate 2a, ΔδC ≈ 5.6 ppm. Following esterification, a strong effect is observed on the
ligand C-O, which is shifted to lower frequencies with ΔδC ≈ 8.6 ppm compared to C-OH 2a. In the
case of amides 13 and 14a, the signal corresponding to the two protons in the α position to the ligand
C-NH are strongly shifted to higher frequencies, by ca. ΔδH ≈ 0.96 ppm, while a similar but less
important effect of only ΔδH ≈ 0.25 ppm is observed in the case of the corresponding protons in the
β position with respect to the C-NH. The 13C NMR spectra of amides 13 and 14a shows the C atoms
α to the ligand C-NH at higher frequencies than in amine 3a with ΔδC ≈ 4.8 ppm. After amide bond
formation, the signal corresponding to the ligand C-NH is strongly shifted to lower frequencies with
ΔδC ≈ 9.6 ppm compared to the respective signal/peak C-NH2 in compound 3a. Less important shifts
are observed in the 1H NMR spectra of amides 15, 16a and 17a compared to acid intermediate 3a for
the corresponding protons α to the ligand C-CH2CONH due to the presence of the methylene spacer.
Interesting, in the 1H NMR spectra of the di-hydroxy intermediate 19 and of the di-ester 20, two sets
of signals/peaks corresponding to the two 4-mercaptophenolic ligands are observed, indicating that
they are not equivalent in the NMR time scale measurement. Electrospray ionization mass
spectrometry (ESI-MS) corroborated the spectroscopic data. The spectra of the dithiolato
intermediates contain a parent ion peak attributable to the [M-Cl]+ ion, formed after the loss of one
of the labile chlorine ligands. The ESI-MS of the cationic trithiolato complexes exhibit a strong peak
corresponding to [M-Cl]+ ion.
The compounds containing the trithiolato scaffold (intermediates as well as conjugates) are stable
towards ligand exchange in highly complexing solvent as DMSO-d6 even after elongated storage at
0°C (see Figures S2−S4 for the spectra corresponding to the compounds 2a-4a and 11-17a,
Supporting Information). This is essential to validate the biological tests for which compounds are
used as stock solutions in DMSO.
The solid state structure of symmetric intermediate 21 containing three free hydroxy groups was
established by single crystal X-ray diffraction analysis (Figure 3), confirming the expected molecular
structure.
12
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
X-ray crystallography
The crystal structure of the symmetric complex 21 was established in the solid state by single-crystal
X-ray diffraction (ORTEP representations shown in Figure 3), confirming the expected structure.
Selected bond lengths and angles are presented in Table S2. Data collection and refinement
parameters are given in Table S1.
Figure 3. ORTEP representation of complex 21 (thermal ellipsoids are 50% equiprobability
envelopes, and H atoms are spheres of arbitrary diameter; the asymmetric unit contains also one
CH2Cl2 molecule).
The Ru2S3 unit forms a trigonal-bipyramidal framework; no metal–metal bond are present, the
corresponding Ru(1)-Ru(2) distance being 3.341 Å. In the cation, the values of Ru–S bonds as well
as the Ru–S–Ru angles (Table S2) are similar to those found in other symmetric p-cymene derivatives
reported previously.18
In the network of 21 a complex interplay of H-bonding interactions involving the chlorine anions and
the three OH groups of the complex in observed (Table S3). Each Cl- ion interacts with the OH groups
from three different cationic complexes. A dimeric organization is saw, which is mediated by four H-
bonding interactions: two chlorine anions bridge two di-ruthenium units at the level of their respective
two hydroxy groups (Figure S1). The remaining hydroxy group of each of the symmetric di-
ruthenium complexes is involved in H-bonding interactions with another Cl- ion. Thus, all three
hydroxy groups of the symmetric trithiolato di-nuclear ruthenium(II)-arene complex 21 are involved
in intermolecular H-bonding interactions, two of them interact with two chlorine anions and lead to
the formation of a dimer with another di-ruthenium unit, while the third OH group interacts via H-
13
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
bonds with other Cl- anion present. In network these intermolecular H-bonding interactions lead to
further arrangements e.g. with the formation of cycles involving up to four cationic complexes.
Photophysical Characterization
The photophysical properties of coumarin containing compounds investigated in this study, namely
the starting Dye1-CO2H and Dye2-CO2H, coumarin-based amino intermediates 5-7, and ester and
amide conjugates with the trithiolato ruthenium(II)-p-cymene scaffold 11-17a/b, 20 and 22, were
studied in CHCl3 and EtOH at r.t. and are summarized in Tables 1 and S4. No solvatochromism was
observed. The absorption and emission spectra of representative compounds for 10 μM solutions in
CHCl3 and/or EtOH are comparatively presented in Figures 4-5 and S5-S7.
14
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Table 1. Photophysical data of compounds 5-17a/b, 20 and 22 in CHCl3 at r.t.
풆풎
풂풃풔
ε
흀
풎풂풙
Δλ
(nm)
24
ΦF
(%)
75[a]
184
158
186
175
170
4
흀
풎풂풙
Compound
(M-1cm-1)
(nm)
557
458
474
449
448
463
455
472
474
458
478
480
454
449
449
462
465
457
459
(nm)
533
rhodamine 6G[a]
62696.6
431
43074.2
27
Dye1-CO2H
449
47372.9
25
Dye2-CO2H
418.5
42971.9
30.5
30.5
29.5
24
5
417.5
47042.6
6
433.5
41504.7
7
431, 245
448, 245.5
447, 247.5
439.5, 246
456, 245.5
454.5, 248
418.5, 245.5
417, 245.5
417, 248
433.5, 245.5
433.5, 248
431.5, 248.5
432, 246
57147.6, 64027.8
59950.3, 64535.9
57647.7, 63481.4
73039.4,68176.6
68769.4, 62124.7
57340.8, 72892.3
37245.8, 63190.6
39772.3, 65900.6
50370.3, 34021.4
36532.7, 64378.6
46995.0, 32315.8
105690.0, 57402.0
136190.0, 55390,8
11
24
1
12a
12b
13
27
0.4
0.3
2
18.5
22
14a
14b
15
25.5
35.5
32
2
3
3
16a
16b
17a
17b
20
32
2
28.5
31.5
25.5
27
3
3
0.3
0.4
22
[a]Values taken from ref. 58
15
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Figure 4. UV-Vis absorption (left) and emission spectra (right) of rhodamine 6G, Dye2-CO2H,
intermediate 7 and the corresponding ester 12a and amide 14a, 17a conjugates, at 10 µM in CHCl3.
Figure 5. UV-Vis absorption (left) and emission spectra (right) of rhodamine 6G, Dye2-CO2H,
intermediate 7 and the corresponding ester 12b and amide 14b, 17b conjugates, at 10 µM in CHCl3.
The absorption spectra of all di-ruthenium unit: coumarin conjugates 11-17a/b, 20 and 22 present a
similar profile (Figures 4-5 and S5-S7). In both solvents, strong peaks corresponding to the coumarin
fragment are observed in the 410-460 nm region. The signals in the 200-300 nm range, associated to
the trithiolato di-nuclear ruthenium(II)-arene moiety, are better resolved in the spectra measured in
CHCl3 compared to those in EtOH. For 11 and 12a/b, the direct attachment of the trithiolato di-
ruthenium moiety to coumarins via ester bonds induced no shifts of the absorption peaks, while in
16
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
case of amides 13 and 14a/b a slight bathochromic shift (Δλ ca. 10 nm) was observed. In contrast,
the presence of the di-amino linker in conjugates 15-17a/b led to a slender hypsochromic shift
(Δλ ~ 12-16 nm).
When excited at 405 nm, all coumarin-containing compounds 5-17a/b, 20 and 22 emit in the blue
range (450–490 nm). The emission spectra of conjugates 11-17a/b, presenting a 1:1 ratio
organometallic unit: coumarin, show similar profiles (Figures 4-5 and S5-S7) with almost complete
fluorescence quenching, only slightly less pronounced for solutions in CHCl3 compared to those in
EtOH. This loss of fluorescence efficacy was independent of the nature of coumarin (Dye1-CO2H or
Dye2-CO2H), the type of bond (ester or amide) or presence of a di-amino linker between the two
moieties. From this compound library, the highest calculated fluorescence quantum yields remain
very modest (ΦF = ca. 3%). Of note, similar but less pronounced quenching effects were observed in
other organometallic coumarin-based conjugates.26, 32, 43
For all coumarin conjugates (11-17a/b, 20 and 22) a small bathochromic shift of the fluorescence
maximum is observed in EtOH compared to CHCl3 (Δλ ca. 10 nm). The dramatic fluorescence
intensity change depending on the presence, or not, of the trithiolato di-ruthenium unit could be
valorized to monitor the stability of the conjugates towards hydrolysis and their behavior in vitro.
Further structural optimization (as for example the introduction of longer or more rigid spacers
between the di-ruthenium and fluorophore moieties) is required in order for this type of conjugates to
be used also as trackable theranostics for the trithiolato di-nuclear ruthenium(II)-arene fragment.
In the case of conjugates 20 and 22 with organometallic moiety: dye ratios of 1:2 and, respectively,
of 1:3, a proportional increase was observed for the coumarin absorbance signal (maximum at
λ = 431-432 nm), which parallels the number of the attached dye units. No significant effect was
observed for the emission signal at 457-459 nm, corresponding to the trithiolato di-ruthenium(II)-
arene unit/moiety.
Coumarin intermediates 5-10 and conjugates bearing di-amino linkers 15-17a/b present slightly
larger Stokes shifts compared to free Dye1-CO2H and Dye2-CO2H and other hybrid molecules.
Calculated Stokes shifts values range between 22-36 nm for experiments performed using solutions
in CHCl3, and in 31-49 nm interval when EtOH was used as solvent.
In both solvents, CHCl3 and EtOH, a linear dependence of the absorbance and emission intensity with
concentration was determined (data not shown). This is observed even in the case of the conjugate 22
bearing three Dye1-CO2H units (spectra of 22 at various concentrations in CHCl3 are summarized in
Figure 6 and Figure S7).
17
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Figure 6. UV-Vis absorption (left) and emission spectra (right) of tri-coumarin ester conjugate 22 at
various concentrations in CHCl3.
Photostability
As the fluorescence intensity dramatically drops upon the addition of the trithiolato di-ruthenium unit
(coumarin-ruthenium conjugates 11-17a/b, 20 and 22 compared to free Dye1-CO2H and Dye2-
CO2H), we used this difference to monitor the stability of the hybrid molecules. To verify the
compounds’ bleaching sensitivity but also stability at r.t. under air, in the presence of a polar solvent
that can induce solvolysis, 10 µM solutions of 12b, 14b, 16b, 17b, 20 and 22 in EtOH were
maintained under indoor light for an extended time period. Absorption and emission spectra were
monitored after 24 h, 48 h and one-week of light exposure, the measurements being made in the same
conditions as previously described. Absorption spectra of all compounds remained unchanged after
48 h. After one week of light exposure a minor decrease of intensity can be observed for the
absorbance signal in the 350-480 nm range, corresponding to the coumarin fragment.
In the emission spectra, notable changes were observed only for ester conjugates 12b, 20, and 22
(data not shown). A similar effect was noticed for the amide conjugates 14b and 17b (Figure 7), but
only after one week of light exposure, while no spectral changes were observed for amide 16b. This
increase of the emission signal can be attributed to a partial solvolysis of the ester or amide bonds
present on the conjugates with the release of the respective coumarin dyes. Nevertheless, considering
the time scale, and seen the very high fluorescence efficacy of Dye1-CO2H, Dye2-CO2H and
corresponding coumarin intermediates 5-10 compared to the emission of the conjugates (fluorescence
18
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
almost entirely quenched in the hybrid molecule), we can conclude that the coumarin-organometallic
conjugates present high stability in the conditions used for this experiment.
Figure 7. UV-Vis absorption (left) and emission spectra (right) of trithiolato ruthenium(II)-p-cymene
complex 14b after 0 h, 24 h, 48 h, and one week (168 h) of exposure to indoor light at 10 µM in
EtOH.
Biological activity of the coumarin conjugates against Toxoplasma gondii
Toxoplasmosis, one of the most common global zoonotic diseases, is caused by the protozoan parasite
Toxoplasma gondii.59 This parasite can infect virtually all warm-blooded animals on the planet and
has very high zoonotic potential. Up to one third of the human population is infected with T. gondii,
but only a small fraction of infected individuals exhibits clinical signs. The economic impact of
T. gondii is enormous, in that it causes severe losses in a wide range of wild and domestic animals,
including most animals used in food production. In general, T. gondii infestation remains without
clinical symptoms in immune competent individuals, and no treatment is required. However, in
humans, infection has been linked to neuropsychiatric disease, and upon immunosuppression, or
primary infection during pregnancy, T. gondii can cause toxoplasmosis, a life-threatening disease
affecting both humans but also animals, which can lead to severe pathology including fetal
malformation and abortion.60 Current standard treatment options for toxoplasmosis include macrolide
antibiotics and sulfonamides,61 which inhibit protein biosynthesis and intermediary metabolism in the
apicoplast, a prokaryote-like organelle that is unique to apicomplexans.62 However, these treatments
are often characterized by adverse side effects and do not act in a parasiticidal manner. The
19
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
development of novel treatment options that specifically target the parasite is therefore of prime
importance.
The compounds presented in this study were screened for biological activity in vitro against T. gondii
β-gal, a transgenic strain that constitutively expresses β-galactosidase, which is grown in HFF (human
foreskin fibroblast) monolayers. In addition, the effects on non-infected HFF host cells were assessed.
For the primary screening, cell cultures were exposed during 3 days to 1 µM and 0.1 µM of each
compound (including non-modified thiolato-bridged dinuclear ruthenium(II)-arene complexes 2a/b,
3a/b, 4a/b, coumarin-labeled conjugates 11-17a/b, 20 and 22, free dyes Dye1-CO2H and Dye2-
CO2H, and corresponding coumarin-based intermediates). The viability of HFF cultures following
drug treatments was measured by alamarBlue assay, and the proliferation of T. gondii was quantified
by measuring β-galactosidase activity. The results of this primary screening are presented as
percentage in relation to untreated control cultures in Table S5 (Supporting information). The results
obtained at concentration of 0.1 µM and 1 µM of tested compound for T. gondii and HFF are
presented in Figure 8, in relation to controls (CTR), namely HFF treated with 0.1% DMSO exhibiting
100% viability, and T. gondii β-gal tachyzoites treated with 0.1% DMSO showing 100%
proliferation.
Figure 8. In vitro activities of all compounds at 0.1 µM and 1 µM on HFF viability (A) and T. gondii
β-gal tachyzoites proliferation (B), in relation to treatments with 0.1% DMSO. For each assay,
standard deviations were calculated from triplicates.
20
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
No clear structure-activity relationship could be identified. The observed activity/cytotoxicity
appeared to be the result of an interplay of various structural parameters that influence the cellular
internalization and further interactions with biomolecules. The trithiolato dinuclear derivatives are
mono-cationic complexes with a substantial molecular weight (> 950 g/mol). Anchoring of
hydrophobic coumarin analogues as pendant arms on one of the thiol ligands, and the nature of the
substituents present on the other two thiols, influence the physico-chemical properties of the
molecule.
Compounds 2a/2b and 3a/3b bearing one hydroxy or amino group presented considerable activity
against T. gondii β-gal at 1 µM, but also impaired HFF cell viability. The same was observed in the
case of dihydroxy compound 19. Acid functionalized trithiolato complexes 4a and 4b did not affect
the viability of HFF monolayers, but did also not affect tachyzoite proliferation at 1 µM. None of the
coumarin-based intermediates 5-10 (Table S5 and Figure S8 in Supporting information) impacted on
HFF viability or T. gondii β-gal proliferation.
The nature of the substituents on the other two thiol ligands, hydrophobic-bulky tert-butyl versus
hydrophobic-polar trifluoromethyl, had a considerable influence in the case of the hydroxyl
derivatives 2a/2b compared to the amino analogues 3a/3b. This structural feature (tBu vs CF3)
appeared to affect the activities of the ester compounds 12a/12b and the amide compounds 14a/14b
in which the coumarin is directly connected to the thiol ligand. In some cases, (11 vs 13, 12a vs 14a,),
the amide conjugates exhibited lower antiparasitic activity compared to the esters for the same
coumarin substituent.
Dye1-CO2H -functionalized compounds 13, 15, and 16a presented a similar cytotoxicity profile for
HFF at 0.1 µM and 1 µM. The introduction of the linker augmented the measured activity against the
parasite. For compounds bearing Dye2-CO2H moieties, the introduction of a linker between the di-
ruthenium scaffold and the coumarin led to a substantially increased antiparasitic activity in the case
of tBu analogue 17a compared to 14a; however, a similar effect was not observed for CF3 analogues
17b and 14b. Compounds 13, 14a and 22 are only poorly active against T. gondii β-gal at 1 µM, and
the last two compounds did not notably affect HFF cell viability. In the series of Dye1-CO2H
functionalized ester compounds 11, 20 and 22 no noteworthy correlation between the number of
coumarin units and the measured biological activity was observed. At 1 µM, 11 displayed substantial
HFF toxicity, while 20 and 22 did not. As shown in Figure 8, Dye1-CO2H and Dye2-CO2H did not
affect viability of HFF when applied at 0.1 µM or 1 µM. However, proliferation of T. gondii relative
to the rate of β-galactosidase activity was decreased following treatment with 0.1 µM of Dye1-CO2H
(53%) and Dye2-CO2H (69%) (Table S5). These results may suggest that the in vitro anti-
Toxoplasma activity of these coumarin dyes may be lost due to solubility issues.
21
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Based on this preliminary screening, 13 compounds were selected for determination of the IC50 values
against T. gondii β-gal. These were compounds that, when applied at 1 µM, allowed viability values
for HFF at 45% or more, while they also inhibited T. gondii proliferation by 94% or more. The results
are shown in Table 2 and Figure S8, and the respective dose-response curves are shown in Figure S9.
Table 2. IC50 values (µM) against T. gondii β-gal tachyzoite proliferation and effects on HFF
viability at 2.5 µM, for selected compounds and pyrimethamine (as positive control).
HFF viability at
Compound
IC50 (µM)
[LS; LI][a]
SE[b]
SD[d]
2.5 µM (%)[c]
0.117
0.336
0.153
0.135
0.105
0.298
0.391
0.127
0.735
0.243
0.203
0.115
0.377
0.326
[0.098; 0.139]
[0.323; 0.35]
[0.127; 0.185]
[0.105; 0.174]
[0.099; 0.111]
[0.292; 0.305]
[0.351; 0.437]
[0.170; 0.095]
[0.467; 1.156]
[0.190; 0.311]
[0.113; 0.366]
[0.098; 0.135]
[0.36; 0.39]
0.0510
0.0088
0.0488
0.0562
0.0230
0.0051
0.0236
0.129
56
72
51
53
58
28
71
4
6
6
5
4
4
1
5
1
7
5
7
4
4
6
2a*
2b
3a*
3b
12a*
12b
14b
16a
0.1155
0.0905
0.1459
0.0447
0.0137
0.0518
73
25
54
2
16b
17a
17b
19
20
71
99
[0.288; 0.396]
Pyrimethamine
[a]Values at 95% confidence interval (CI); LS (limit superior) and LI (limit inferior) are the
upper limit and the lower limit of the CI, respectively. [b]The standard error (SE) of the estimate
represents the average distance that the observed values fall from the regression line (T. gondii
β-gal). [c] Control HFF treated only with 0.25% DMSO exhibited 100% viability. [d] The
standard deviation of the mean of six replicate experiments (HFF). *indicates the three
compounds 2a, 3a and 12a selected for assessing their potential to interfere in splenocyte
proliferation in vitro
The most active compound against T. gondii β-gal was 12a, with an IC50 value of 0.105 µM. This
compound was not toxic at 1 µM, but when applied at 2.5 µM it reduced HFF viability to 58%. Higher
levels of HFF viability impairment were noted for compounds 12b, 17a and especially 19 (98%
22
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
reduction in HFF viability). The difference in antiparasitic activity between compounds 12a and 12b
(IC50 values of 0.105 µM and 0.298 µM, respectively) highlights the importance of the physico-
chemical properties of the substituents present on the other thiol ligands, tBu vs CF3. As shown in
Table 2, unmodified hydroxy compound 2a and amino compounds 3a and 3b, also presented
considerable antiparasitic activities (IC50 values of 0.117 µM, 0.153 µM and 0.135 µM,
respectively), furthermore, their impact on HFF viability at 2.5 µM was less pronounced compared
to 17a and 19.
Three selected compounds (2a, 3a and 12a), applied at their respective IC50 against T. gondii, were
further assessed with respect to their potential to interfere in splenocyte proliferation in vitro. Isolated
murine splenocytes from healthy mice were stimulated with Concanavalin A (ConA) to induce T cell
proliferation or with bacterial lipopolysaccharide (LPS) to induce B cell proliferation, either in the
presence or absence of tested compounds 2a, 3a and 12a. Measurements of proliferation were done
using an assay that quantifies the incorporation of 5-bromo-2’-deoxy-uridine (BrdU) into the DNA
of replicating cells. As seen in Figure 9, 3a and 12a significantly interfered in T cell proliferative
responses, which resulted in a reduction of BrdU incorporation by 25% and 31%, respectively, while
2a did not affect T cell proliferation. Compounds 2a and 3a significantly impacted the proliferation
of B cell (only 48% and 68%, respectively, of BrdU incorporation compared to the control), while
12a did not exhibit significant proliferation inhibition of antibody producing cells. Thus, 3a affected
both B and T cell proliferation, while 2a and 12a impaired the proliferative capacity of only one type
of immune cells; B cells for 2a (humoral immunity) and T cells for 12a (cellular immunity).
Figure 9. Inhibitory effect of selected compounds 2a, 3a and 12a ConA- (A) and LPS-induced (B)
proliferative activity of mouse splenocytes. Bars represent standard deviation from the mean of four
replicates. A 100% of proliferation is attributed to controls (ConA and LPS); values are percent of
proliferation compared to control (*) p < 0.01 in relation to controls (ConA or LPS).
23
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
A reduction in the capacity of T and B cells to respond to external signals by proliferative responses
could also indicate a potential risk of impaired immunity. However, a reduction in cellular
proliferation does not automatically imply a reduced metabolic activity or reduction in cellular
viability. Thus, in addition to proliferative responses of B and T cells upon LPS and ConA
stimulation, we also assessed the effects of 2a, 3a and 12a on the viability of splenocytes employing
the alamarBlue assay, which allows, similar to what we have tested in HFF, to quantify metabolic
activity (Figure 10). Measurements taken each hour during the first 5 h after adding the substrate
resazurin showed that none of the compounds impaired the metabolic activity of ConA stimulated
T cells, but compounds 2a and 3a significantly impaired the viability of B cells upon stimulation with
LPS. 12a did not show any interference in metabolic activity of neither B nor T cells. The lack of
viability impairment of splenocytes due to exposure with 12a observed here indicates that this
compound is a promising drug candidate for future in vivo studies in the mouse model, to combat
T. gondii infection without impairing the cellular and humoral immune response.
24
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Figure 10. Determination of cytotoxic effects of 2a, 3a and 12a in ConA- (A) and LPS (B) induced
splenocytes by alamarBlue assay. Bars represent the mean emission of 4 replicates ± standard
deviation. In both experiments and for each time point a 100% of metabolic activity was attributed to
the control sample (splenocytes induced with ConA or LPS), then the percentage of metabolic activity
in relation to the controls was calculated, and is indicated in the graph for each sample. Significance
with p < 0.001 was observed between LPS and LPS + 2a, and LPS + 3a for all five time points.
The ultrastructural changes induced by compounds 12a (IC50 of 0.105 M against T. gondii, HFF
viability of 58% at 2.5 µM) and 17a (IC50 of 0.243 M against T. gondii, HFF viability of 24% at
2.5 µM) were further studied by TEM. HFF monolayers were infected with T. gondii β-gal
tachyzoites and after 24 h drug treatments (500 nM of each compound, a concentration that did not
25
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
notably affect the host cell) were initiated. Samples were fixed and processed after 6 h, 24 h and 48 h.
Non-treated control cultures are shown in Figure 11 A-C. A shows a sample fixed after 6 h post-
invasion, B and C have been fixed after 36 h and 60 h post-invasion, respectively, and the increase in
number of parasites illustrates the proliferation that takes place within the host cell. Once invaded,
T. gondii tachyzoites are localized within a parasitophorous vacuole (PV), surrounded by a
parasitophorous vacuole membrane (PVM), which is essentially host cell surface membrane-derived
and modified by the parasite following invasion. The mitochondrion exhibits a characteristic electron
dense matrix containing numerous cristae. Tachyzoites treated with 500 nM 12a are shown in
Figure 12. At 6 h after initiation of treatment, parasites do not exhibit massive alterations, however,
PVs usually contained only 1-3 tachyzoites. The PVM was still clearly discernible, and the secretory
organelles such as rhoptries, micronemes, and dense granules, as well as the mitochondria remained
largely unaltered (Figure 12 A-B). At 24 h, first ultrastructural changes were noted within the
mitochondrial matrix, which started to lose its characteristic electron-dense matrix. These
mitochondrial alterations became progressively more pronounced at 24 h (Figure 12 D and E) and
48 h (Figure 12 F, G), resulting in tachyzoites that were completely devoid of a mitochondrion, but
exhibited large, seemingly empty, vacuoles instead. While it is not clear whether these effects were
reversible, it is conceivable that this extensive vacuolization would eventually lead to parasite death.
Similar results were seen for 17a, (Figure S10), although the effects after 24 h and 48 h were slightly
less pronounced. Overall, these findings mirror previously reported structural alterations induced by
ruthenium complexes reported in T. gondii, N. caninum and in T. brucei, and in the latter it was
recently shown that active complexes strongly impaired the mitochondrial membrane potential,23-25
which indicates that ruthenium complexes interfere in the energy metabolism of these parasites.
However, while oxidative phosphorylation resulting in the generation of ATP is the major function
of mitochondria, these organelles are also involved in other crucial processes, including cell cycle
regulation, t-RNA- and protein-import, mitochondrial protein translation, alternative oxidase, acetate
production for cytosolic and mitochondrial fatty acid biosynthesis, amino acid metabolism and
calcium homeostasis, and they are involved in the steps leading to programmed cell death. How, and
to what extent, these compounds, actually target mitochondrial functions, or whether other targets are
also involved, remains to be elucidated
26
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Figure 11. TEM of untreated T. gondii-β-gal tachyzoites fixed at 6 h (A), 24 h (B) and 48 h (C) post
infection. A shows a single tachyzoite, located within a parasitoporous vacuole that is delineated by
a parasitophorous vacuole membrane (arrows). B Proliferation of tachyzoites takes place within the
vacuole, which occupies a substantial part of the host cell cytoplasm. C shows two neighboring
parasitophorous vacuoles, delineated by arrows, located within a HFF host cell, both containing
numerous newly formed tachyzoites. Note the mitochondrion (mito) with an electron dense matrix in
A; nuc = tachyzoite nucleus, dg = dense granule, hcc = host cell cytoplasm; hcn = host cell nucleus.
27
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
Figure 12. TEM of T. gondii-β-gal tachyzoites treated with 500 nM of 12a during (6 h (A, B), 24 h
(C-E) and 48 h (F-H). No alterations (A) or only very slight changes in the electron dense
mitochondrial matrix (mito in B) are detected after 6 h. C-E show that alterations were much more
pronounced after 24 h of treatment, C is a low magnification overview, D and E represent distinct
parts of C shown at higher magnification. F is a low magnification view of a parasitophorous vacuole
of tachyzoites treated for 48 h, G and H represent respective high magnifications. Note the increased
vacuolization (marked with *) and the absence of any mitochondrial matrix after 24-48 h of treatment.
PV = parasitophorous vacuole; mito = mitochondrion; nuc = nucleus, con = conoid; rop = rhoptries;
dg = dense granule; mic = microneme; arrows point towards the parasitophorous vacuole membrane.
28
This article is protected by copyright. All rights reserved.

10.1002/cbic.202000174
ChemBioChem
In order to assess the potential use of the new conjugates coumarin-diruthenium unit as cellular
trackable probes fluorescence microscopy assays were performed of HFF cells treated with either
20 μM Dye2-CO2H or 20 μM of 12a, both counterstained with an antibody directed against tubulin
and NucRed, a fluorescent nuclear label (see results and experimental in Supporting information).
However, while anti-tubulin readily stained the cytoskeleton of HFF and NucRed indicated the
nucleus, no labeling could be observed with neither Dye2-CO2H nor 12a. In order to obtain trithiolato
di-nuclear ruthenium conjugates suitable to be used as theranostics, the attachment of other
fluorophores should be considered.
CONCLUDING REMARKS
A library of 13 new trithiolato-bridged dinuclear ruthenium(II)-arene organometallic conjugates in
which coumarin moieties were anchored to the bridged thiol(s), have been designed and synthesized.
Irrespective to the structural variations considered, for all conjugates organometallic unit: coumarin
an almost complete loss of fluorescence efficacy was observed. However, the nature of the
fluorophore, the type of bonding, the presence and length of a linker between the coumarin dye and
the dinuclear ruthenium(II) moiety, and the number of dye units did influence the biological
properties of these compounds. These modifications also affect the toxicity of these compounds
against human fibroblasts, and impact the activity against the apicomplexan parasite T. gondii, grown
in these cells in vitro. For selected compounds (2a, 3a and 12a), applied at their T. gondii-IC50, the
effect on the proliferative responses of splenocytes upon ConA (for T cell) and LPS (for B cell)
stimulation was investigated, and the influence on the viability/metabolic activity of B and T cells
in vitro was assessed. With IC50 values ranging from 105 nM to 735 nM, and seen that 9 compounds
displayed lower IC50 than the standard drug pyrimethamine, suggest that these class of compounds is
promising. In particular, compound 12a did not affect the metabolic activity of B or T cells in vitro,
is therefore not expected to impair immunity, and thus represents a promising compound for future
in vivo assessment in toxoplasmosis mouse models.
29
This article is protected by copyright. All rights reserved.