Synthesis and characterization of η6-p-cymene ruthenium(II) complexes containing alkyl- and methoxy-substituted triarylphosphines

Article abstract of DOI:10.1016/j.jorganchem.2020.121599

Neutral Ru(II)-arene complexes [Ru(η⁶‐p-cymene)(PAr₃)Cl₂], PAr₃ where Ar = 3,5-((CH₃)₃C)₂C₆H₃–, 3,5-(CH₃)₂C₆H₃–, 4-CH₃O-3,5-(CH₃)₂C₆H₂– and 4-CH₃O-C₆H₄– were synthetized and fully characterized including single-crystal X-ray diffraction. The ligands have empirical logP_ow = 10.29, 7.38, 7.08, 5.43 (vs. PPh₃ = 5.14). They have sufficiently low solubility to form Langmuir monolayers on water in accordance with overall molecular sizes. Density functional theory (DFT) calculations were performed to determine the molecular and electronic structures and to interpret the results of voltammetry and UV–Vis spectroscopy. Cytotoxicity assays were evaluated against MRC-5 (non-tumoral lung), A549 (lung cancer) and MDA-MB-231 (breast cancer) cell lines. The complexes with Ar = 3,5-((CH₃)₃C)₂C₆H₃–, 4-CH₃O-3,5-(CH₃)₂C₆H₂– and 4-CH₃O-C₆H₄– show high cytotoxicity, whilst the first complex is inactive towards the (healthy) MRC-5 cell line, indicating a degree of selectivity.

Full text of DOI:10.1016/j.jorganchem.2020.121599

Journal Pre-proof
Synthesis and characterization of η6-p-cymene ruthenium(II)
complexes containing alkyl- and methoxy-substituted
triarylphosphines
˜
´
Ivelise Dimbarre Lao Guimaraes , Flavia Marszaukowski ,
´
Renan Ribeiro , Sergio Ricardo de Lazaro , Katia Mara de Oliveira ,
Alzir Azevedo Batista , Patr´ıcia Castellen , Ellen Wrobel ,
´
´
Jarem Raul Garcia , Rene T. Boere , Karen Wohnrath
PII:
S0022-328X(20)30502-7
DOI:
Reference:
https://doi.org/10.1016/j.jorganchem.2020.121599
JOM 121599
To appear in:
Journal of Organometallic Chemistry
Received date:
Revised date:
Accepted date:
19 August 2020
27 October 2020
29 October 2020
˜
´
Please cite this article as:
Ivelise Dimbarre Lao Guimaraes ,
Flavia Marszaukowski ,
´
Renan Ribeiro ,
Sergio Ricardo de Lazaro ,
Katia Mara de Oliveira ,
Alzir Azevedo Batista ,
´
´
Patr´ıcia Castellen ,
Ellen Wrobel , Jarem Raul Garcia ,
Rene T. Boere ,
Karen Wohnrath ,
Synthesis and characterization of η6-p-cymene ruthenium(II) complexes containing alkyl-
and methoxy-substituted triarylphosphines, Journal of Organometallic Chemistry (2020), doi:
https://doi.org/10.1016/j.jorganchem.2020.121599
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier B.V.

Highlights
 A rationally selected series of PAr3 more basic than PPh3
 Phosphines enhance the lipophilicity of neutral 6-arene Ru complexes
 Langmuir films demonstrate how geometries are dominated by bulky phosphines
 Redox-stability window exceeding the reduction and oxidation of aqueous media
 Cytotoxicity against healthy and cancerous cell lines shows partial selectivity
1

Synthesis and characterization of η6-p-cymene
ruthenium(II) complexes containing alkyl- and
methoxy-substituted triarylphosphines
Ivelise Dimbarre Lao Guimarães,a Flávia Marszaukowski,a Renan Ribeiro,b Sergio Ricardo
de Lazaro,a Kátia Mara de Oliveira,c Alzir Azevedo Batista,c Patrícia Castellen,d Ellen
Wrobel,a Jarem Raul Garcia,a René T. Boeré*e and Karen Wohnrath*a
a Departamento de Química, Universidade Estadual de Ponta Grossa, Av. Gal. Carlos
Cavalcanti, 4748, Uvaranas, CEP 84030-900, Ponta Grossa, PR, Brasil.
b Departamento de Química, Universidade do Estado de Minas Gerais, Av. Paraná, 3001,
Jardim Belvedere, CEP 35501-170, Divinópolis, MG, Brasil.
c Departamento de Química, Universidade Federal de São Carlos, Rodovia Washington Luis
s/n, km 235, CEP 13565-905, São Carlos, SP, Brasil.
d Departamento de Engenharia Agronômica, Instituto Federal Catarinense, Rua das Rosas,
s/n, Vila Nova, CEP: 88965-000, Santa Rosa do Sul, SC, Brasil.
e Department of Chemistry and Biochemistry, University of Lethbridge, 4401 University
Drive, Lethbridge, Alberta, Canada T1K 3M4.
Ruthenium-arene complexes, triarylphosphines, cyclic voltammetry, cytotoxicity assays,
lipophilicity, Langmuir monolayers.
2

ABSTRACT
Neutral Ru(II)-arene complexes [Ru(η6‐p-cymene)(PAr3)Cl2], PAr3 where Ar = 3,5-
((CH3)3C)2C6H3–, 3,5-(CH3)2C6H3–, 4-CH3O-3,5-(CH3)2C6H2– and 4-CH3O-C6H4– were
synthetized and fully characterized including single-crystal X-ray diffraction. The ligands
have empirical logPow = 10.29, 7.38, 7.08, 5.43 (vs. PPh3 = 5.14). They have sufficiently low
solubility to form Langmuir monolayers on water in accordance with overall molecular sizes.
Density functional theory (DFT) calculations were performed to determine the molecular and
electronic structures and to interpret the results of voltammetry and UV-Vis spectroscopy.
Cytotoxicity assays were evaluated against MRC-5 (non-tumoral lung), A549 (lung cancer)
and MDA-MB-231 (breast cancer) cell lines. The complexes with Ar = 3,5-((CH3)3C)2C6H3–,
4-CH3O-3,5-(CH3)2C6H2– and 4-CH3O-C6H4– show high cytotoxicity, whilst the first
complex is inactive towards the (healthy) MRC-5 cell line, indicating a degree of selectivity.
3

1. Introduction
The use of ruthenium complexes for the design of new metallodrugs for cancer
chemotherapy is high and there is a strong current interest in compounds where the metal is
present in its 2+ oxidation state.1-3 As antineoplastic drugs, ruthenium complexes are very
promising, mainly for use against tumors that develop resistance to cisplatin, [PtCl2(NH3)2].
The two main classes of Ru(II) complexes under current development are ruthenium
polypyridyls, [Ru(NN)3]2+, which are generally inert towards ligand substitution, and more
labile 6-arene complexes [Ru(6-arene)L3]n+.2 The 6-arene ligand is often considered to
stabilize the ruthenium 2+ oxidation state under physiological conditions. Many different
kinds of arene-Ru(II) complexes have been investigated but the design strategies aim that at
least one L is labile (often Cl−) while the remainder may be monodentate or chelating ligands
with N, O or S donor atoms.1 Neutral molecules are considered prodrugs relying on
hydrolysis of e.g. halide to develop charge in situ, but complexes that are already cationic are
also common. With one key exception, the so-called RAPTA (Ruthenium-Arene-PTA)
family of complexes which have promising antineoplastic activity and are based on the
monodentate 1,3,5-triaza-7-phosphatricyclo-[3.3.1.1]decane (PTA) ligand, organophosphines
(PR3) have been rarely used as donor ligands.4 The RAPTA-C group of complexes has been
most investigated, because though shown to have low cytoxicity to tumor cells in vitro, they
demonstrate strong antimetastatic activity in vivo.1
PTA is both a strong base and quite water-soluble and is exceptional in these features
compared to the majority of PR3 employed in coordination chemistry. However, there is now
a surge of interest in the use of triarylphosphines in 6-arene Ru(II) complexes for cytotoxic
properties.5-19 The most common themes of this recent work include simply using PPh3 as a
stabilizing ligand with activity occurring through the remaining ligands at Ru.11,13,14,20
Another theme has been using the strong  donor properties of PAr3 to link a bioactive
4

fragment that is attached at the 4-position of one phosphine aryl group.7-9 PPh3 has been
described as a ‘highly lipophilic’ phosphine that greatly increases cytotoxicity (IC50 values in
the ηM range) in some cationic complexes albeit at the cost of an inherent lower solubility in
water.20
Chart 1. Structures of the ligands and complexes used in this work
Here we report on the first [Ru(η6-p-cymene)(L)Cl2] complexes based on a rationally
selected series of comparable PAr3 ligands, L1 – L4 (Chart 1), which are all more basic than
PPh3,21,22 so as to give stable bonding to Ru(II), but which differ systematically in their
expected lipophilicity (L1 > L2 > L3 > L4) based on empirical substituent factors (Table 1).
The ligands and corresponding ligand oxides (LO1 – LO4) were also synthesized and have
been structurally characterized.19,21 Complexes 1 – 4 have been prepared and fully
characterized structurally and spectroscopically. The cytotoxic assays of these neutral
phosphine complexes were evaluated in vitro against the MRC-5 (non-tumoral lung), A549
(lung cancer), and MDA-MB-231 (breast cancer) cell lines. IC50 in the μM range were
observed, even for the most lipophilic complexes and a significant metal-effect was
demonstrated by greater activity of 1 – 4 compared to L1 – L4 and LO1 – LO4. Here we
report a first analysis of structure-property relationships for neutral 1 – 4, including
5

comparisons to high-level DFT computations on [Ru(η6-p-cymene)(PPh3)Cl2] 5, formation
of Langmuir films and antiproliferative activity. Lipophilicity can be strongly correlated with
cell-uptake for chemotherapeutic agents;12,20,24,25 the hope here is that this effect may
compensate for lower solubility and contribute to the ongoing quest to establish the mode(s)
of action of organometallic Ru complexes in cell toxicity.
Table 1. Empirical logP octanol-water partition coefficient (logPow) dataa
PPh3
L1
L2 L3 L4
ALOGPS 2.1 b 10.51 7.45 6.90 5.86 5.46
Molinspector c 10.06 7.30 7.25 4.99 4.82
Average
10.29 7.38 7.08 5.43 5.15
a
Empirical computed logPow determined from the 3D geometries (SMILES).b
http://www.vcclab.org/lab/alogps/. c https://www.molinspiration.com/cgi-bin/properties.
2 Results and discussion
2.1 Synthesis, characterization and reactivity of complexes
Our prior experience with highly substituted triarylphosphines bearing 2,6-
diisopropylphenyl (Dipp) groups demonstrated less pyramidal geometries and more easily
oxidized electronic structures for the series P(Dipp)3, PDipp2Ph, a, PDippPh2, b (Scheme
1).26,27 Like the well-known bulky phosphine PMes3 (Mes = 2,4,6-trimethylphenyl), c, these
phosphines have limited abilities to coordinate to metals and reactivity with [{Ru(η6‐p-
cymene)(-Cl)Cl}2] has not yet been reported. It was found that reactions of a, b, and c with
the dimeric precursor are unsuccessful (Scheme 1), even when heated under reflux (as
monitored by 31P NMR spectroscopy). The phosphine b with only a single Dipp group is
known to form complexes with Cu(I) but is still too sterically encumbered to form the desired
[Ru(η6‐p‑cymene)(PAr3)Cl2] complex.27
6

Scheme 1. Synthesis schemes for successful (1 - 4) and failed (a – c) reactions
By contrast, the reactions of the 3,5-disubstituted series (L1 – L4) in 2:1 ratio with the
dinuclear precursor [{Ru(η6‐p-cymene)(-Cl)Cl}2], were rapid and complete at room
temperature (Scheme 1), yielding organometallic complexes of general formula [Ru(η6‐
p‑cymene)(PAr3)Cl2]: 1 (Ar = 3,5-di-tertbutylphenyl), 2 (Ar = 3,5-dimethylphenyl), 3 (Ar =
4-methoxy-3,5-dimethylphenyl) and 4 (Ar = 4-methoxyphenyl). All reactions were conducted
in degassed dichloromethane at RT under N2 and resulted in good yields (70-95%). The
complexes were recrystallized by preparing concentrated CH2Cl2 solutions and diluting by
diffusion using n-pentane (1) or diethyl ether (2 – 4) at RT in a sealed vessel protected from
13
31
light. Complexes 1 – 4 have been fully characterized by 1D (1H, C{1H} and P{1H}) and
2D (1H–1H gCOSY), (1H–13C gHSQC) and (1H–13C gHMBC) NMR experiments (see
Experimental Methods and Figs. S21-S24 in Supporting Information), as well as by EA, UV-
vis absorption spectroscopy and vibrational (IR and Raman) spectroscopy.
Complexes 1 - 3, which all have 3,5-substituents, are unpublished whereas the synthesis of
4 is already reported but the complex was not structurally characterized.20 The synthesized
complexes are red/orange solids, stable in light and in air, with high purity, as evidenced by
NMR spectra and elemental analysis, which are consistent with the proposed formulae for
their structures. In addition, 1 – 4 are soluble in acetonitrile, dichloromethane and
7

chloroform, and almost insoluble in water. For qualitative solubilities in a variety of organic
solvents, see Table S1.
2.2 Stability in solution and reactivity studies
To be sure that the complexes are stable to the protocol used for the in vitro cell tests for
cytotoxicity, NMR spectra were undertaken in d6-DMSO. Recent reports indicate that neutral
[Ru(η6‐p‑cymene)LCl2] complexes where L are various monodentate nitrogen ligands are
rapidly converted to the known species [Ru(η6‐p‑cymene){S(O)CH3)2}Cl2] with
displacement of L.28-32 Apart from typical solvent shifts for the switch from CDCl3 to d6-
DMSO, no changes are observed in the NMR spectra that are indicative of reaction with the
solvent under conditions of neat DMSO at RT (Fig. S1-S2).
Scheme 2. Reactivity studies conducted on complex 4.
In order to determine stability towards hydrolysis, 4 was also reacted at RT with aqueous
methanol (10% H2O v:v) in presence of AgPF6 (Scheme 2, ) and no precipitation of the
AgCl salt by displacement of chloride was observed. Therefore, no evidence of reaction is
detected from the 1H NMR spectrum (Fig. S3). Further tests of the stability of the complexes
towards DMSO was undertaken on 4 by refluxing with methanol solutions containing 1 – 3
equivalents of DMSO. Surprisingly, this shows by NMR (Fig. S4) to result in the
displacement of phosphine L4 by the solvent affording a neutral complex, which has not been
further characterized (Scheme 2, ). In this case, no precipitation of AgCl was observed,
consistent with spectroscopic evidence of partial displacement of the phosphine ligand rather
than chlorides from the structure. In summary, it appears that complexes 1–4 are stable
towards hydrolysis and DMSO, except when heated to high temperatures.
8

2.3 X-ray crystallography
Crystals of complexes 1 – 4 were obtained in several ways, first by layering CDCl3
solutions with pentane at -30 °C (NMR tube method; 2∙2CHCl3, 3∙1.5CHCl3 and 4∙0.5CHCl3)
and later, after bulk recrystallization for purification of the complexes, from evaporation of
pentane (1) or methanol (2) solutions or by vapor diffusion of diethyl ether into concentrated
solutions of the complexes in CH2Cl2 (3 and 4). In all, seven crystal structures were
determined. For simplicity, displacement ellipsoids plots of the molecular structures are
shown in Fig. 1 only for the solvent-free species 1 – 4 and key comparative structural data are
compiled in Table 2. Full crystallographic details and selected bond lengths (Å) and angles
(°) are summarized in Tables S2-S4 in the Supporting Information. The crystal structures
display considerable complexity, which is detailed exhaustively in the Supporting
Information (Fig S5-S13), and fully reliable models could be refined in all cases.
All show the expected pseudo-octahedral coordination at Ru (Fig. 1) with the η6‐cymene,
phosphine and two chloride ligands.6,11,33 The structure images in Fig. 1 have been accurately
scaled relative to one another, showing the considerable variation in overall size as well as in
the atomic displacements in the lattice at similar temperatures (noticeably higher for 1).
Importantly, while the local environment at the metal is typical of such “three-legged piano-
stool” structures, the molecular shapes are dominated by the PAr3 resulting in globular shapes
for 1 and 2 (Fig. 1 left) compared to overall triangular in 3 and 4 (see Fig. S14 in the
Supporting Information). The distal steric effects from the exo-oriented perimeter substituents
described previously for L1 – L4 are clearly visible in the structure diagrams.21 These
structural features strongly influence Langmuir films formed from layers of the pure
complexes on a water surface (see below).
9

Figure 1. (Left) displacement ellipsoids plots omitting H atoms of complexes 1 (30%
probability) and 2-4 (50% probability) in their crystal structures showing key atom number
labels and (right) ‘space filling’ plots of each complex, scaled to relative sizes. Details: 1
t
shows only residue-1 (of in all 6, see SI) omitting disordered Bu; 2 and 4 omit the minor
components of disorder.
Average geometric parameters comparing the most reliable structures of 1 – 4 are provided
in Table 2 and are compared to UB3LYP/(6-31+G(d), Ru:SPK-DZCD) DFT calculations on
[Ru(η6‐p‑cymene)(PPh3)Cl2], 5, which serves as a simplified electronic model for 1 – 4. The
Ru–P bond distances range from 2.366(8) to 2.426(2) Å and average to 2.386(24) Å. Values
10

are within standard uncertainty (s.u.) at the 99% confidence level of the averages of
2.345(52) for all such bonds reported in the Cambridge Structure Database (CSD, Version
2.40, November 2018) and 2.345(31) Å in Ru(6-p- cymene) phosphine complexes.34
Table 2. Geometry parameters of complexes 1-4 in comparison with computed 5.
Complexes
Ru-P
2∙2CHCl3
1
3
4
5c
2.366(8)
2.416(4)
2.408(2)
2.426(2)
2.3824(6)
2.3693(12)
2.4101(13)
2.4250(11)
2.216(2)
262.72
2.4648
2.4783
2.4767
2.29(3)
264.44
Ru-Cl1
2.409(10) 2.4153(7)
Ru-Cl2
2.426(1)
2.23(5)
266.61
2.4189(6)
2.21(3)
Ru-C (average) a 2.215(2)
Ʃ of angles b
263.72
265.63
a
b
Average of bond lengths Ru-C (Å) on (η6‐p‑cymene) ligand. Ʃ of angles (°) = P—Ru—Cl1, P—Ru—Cl2
and Cl1—Ru—Cl2. [Ru(η6‐p‑cymene)(PPh3)Cl2] complex 5 from DFT calculations, see text; a crystal
c
structure of 5 has been reported.35
Regarding 1, the ligand L1, oxide LO1 as well as sulfide and selenide have been reported.19
Beyond this only one coordination complex is reported in the literature, [L12PdI2], a trans
square-planar complex (CSD refcode: ABIDEX) in which the phosphine has a very similar
conformation to that found in 1 with one of the aryl rings ‘flat’.36 For 2, the ligand L2 and
oxide LO2 have been crystallographically characterized.23 There are five reported structures
of metal complexes and the two with Ru may be compared. The complex with CSD refcode
COQDET has the composition [RuCl3L22(OH2)1/2]2 and is dimerized with 2-fold rotational
symmetry.37 The complex XARCOJ is [RuCl2L22(NN)] with cis-orientation of the
phosphines and a back-bone substituted 1,2-diaminoethane type ligand.38 The average Ru-P
distances in these two structures of 2.280(1) and 2.318(18) are both significantly shorter than
2.426(2) Å in 2 at the 99% confidence level. Similarly, apart from structures L3 and LO3,23
only a single complex of L3 is available for comparison with 3, in which the phosphine is
11

coordinated to Ir in a bimetallic cluster (CSD refcode: TUTJAV).39 The structures of 4 and
4∙0.5CHCl3 are most closely related to [Ru(η6‐C6H6)Cl2L4] (CSD refcode: DAFNIJ)40 and to
[Ru(η6‐CH3C6H5)L4Cl2] (CSD refcode: SITHEJ)41 which are the only other reported
structures for (η6-arene)RuCl2 complexes of this phosphine. The Ru–P, Ru–Cl and P–Cipso
bonds in these related complexes are averageable to 2.362(6), 2.430(8) and 1.830(4) Å at the
99% confidence level.
2.4 Electronic structure from DFT calculations on a model system
DFT calculations at the UB3LYP/(6-31+G(d),Ru:SPK-DZCD) level of theory 42 on model
complex 5, were used to verify the electronic structure, electronic absorption spectra and
voltammetric behavior of complexes 1-4. Molecular geometries for ionized states with −1, +1
and +2 charges were also fully optimized in the gas phase and with a polarizable continuum
model (PCM) for CH2Cl2. The DFT-computed geometry for the neutral complex is in
excellent agreement for the parameters common to all five complexes (see Tables S6 and S16
in Supporting Information). The energies and topologies of the frontier molecular orbitals
(FMOs) are depicted in Fig. 3 (and S17 in Supporting Information). The HOMO is dominated
by Ru d (44.5 %) and Cl p-lone-pair (39.2 %) character, is formally antibonding to Cl, but
has bonding character to the p-cymene ligand. Much the same is true for HOMO-1 (41.0 and
45.6 %) and HOMO-2 (37.3 and 53.0 %) for Ru d and Cl p-lone-pair character, respectively.
Together, these three MOs are close in energy and correspond approximately to the t2g set of
Ru d-orbitals for an ideal octahedral structure. As for any transition metal with π-base ligands
(halogens, alkoxides, RS- and the like), these MOs have filled chlorine non-bonding electron
contributions. Moreover, HOMO-3 has primarily Cl π-non-bonding electron contributions of
a symmetry combination that is not matched by any d orbital on Ru (82.9 %). The LUMO
and LUMO+1 have Ru-p-cymene * and Ru-Cl σ* character. For the LUMO, the
12

composition was calculated to be Ru d (49.5 %), Cl p (10.0 %) and p-cymene (8.1 %),
whereas the LUMO+1 has a composition of Ru d (55.6 %), Cl p (14.7 %) and p-cymene
(10.8 %). These two MOs correspond approximately to the eg set of Ru d-orbitals.
Figure 2. FMO energy levels and Kohn-Sham orbital surface representations for the FMOs
from dispersion-corrected DFT/TZVP/RPE0 calculations on the model complex 5.
2.4 Electronic absorption spectroscopy
The electronic absorption spectra of the complexes 1-4 were recorded in acetonitrile,
dichloromethane and chloroform (Fig. 3, see Fig. S19 and Table S7), at 10-4 and 10-5 M. The
spectra of all four complexes are very similar with “shoulders” of high intensity in the UV
region, centered at about 248-291 nm (ɛ = 85700-4500 L∙mol-1∙cm-1), which can be assigned
to intraligand π–π* transitions of triarylphosphine and p-cymene ligands. Lower-energy and
lower intensity bands in the range at 370-376 nm (ɛ = 4300-940 L∙mol-1∙cm-1) are attributed
to MLCT transitions from Ru to π* ligand orbitals. A third region centered closer to 500 nm
occurs in all the spectra as shoulders. These involve transitions between t2g and eg levels of
13

Ru. The assignments are based on TD-DFT calculations performed in a PCM solvent model
(CH3CN) on model complex 5 (Table S8 and Fig. S20).43,44 The strong similarity of the
spectra of 1 - 4 confirms the validity of using 5 as an electronic model system in computation.
Figure 3. UV-Vis spectra of complexes 1 – 4 in CH3CN. The range in the main window is
300-700 nm with an inset expanded to 240 - 350 nm.
2.5 Electrochemical characterization
To study the redox stability of 1-4 under expected physiological conditions, cyclic
voltammetry (CV) experiments were undertaken using, however, CH2Cl2 as the medium to
extend to a wider voltage range than is possible in water, as has previously been reported for
related kinds of complexes.9 Representative voltammetric data are presented in Fig. 4 and
Table 3. CVs for 1-4 have been recorded upon scanning from +1.5 to −1.8 V over 100 mV∙s-1
(Fig. 4), 0 to 1.5 V and 0 to −1.8 V (v = 100-2000 mV∙s-1) (see as S24-30 in the Supporting
Information). The voltammograms for all complexes show very similar behavior and redox
I
potentials. Firstly, the anodic region exhibits a quasi-reversible redox process (퐸푝푎 /퐸푝푐I)
commonly assigned to a one-electron oxidation (consistent with the DFT calculations), i.e.
removal of a HOMO electron, formally a Ru(II/III) couple, with potential values ranging
from 퐸푝푎I = +1.13 to +1.15 V, Em = 1.08 to 1.09 V. The same assignment is in accordance
14

with similar complexes where the range of the redox potential Ru(II/III) couple is Em = 1.11
to 1.15 V.9
Figure 4. Cyclic voltammograms of 1 - 4 at 2.0∙10-3 mol∙L-1 in TBAP/CH2Cl2 0.1 mol∙L-1 vs.
Ag/AgCl; obtained at 100 mV∙s-1.
Table 3 Voltammetric data for complexes 1 – 4.
퐼푝 퐼푝 ∆퐸푝 /Vb 퐸푝푐 /V 퐸푝푐III /V
⁄
Ewind/Vc
I
퐸푚 /V a
퐸푝푎 /V 퐸푝푎푐I /V
I
I
I
II
푐
푎
Complex
1.13
1.14
1.14
1.15
1.03
1.03
1.04
1.04
1.08
1.08
1.09
1.09
0.55
0.62
0.62
0.10
0.11
0.10
0.11
-1.63
-1.54
-1.53
-1.52
-1.83
-1.73
-1.72
-1.69
2.96
2.87
2.86
2.84
1
2
3
4
0.59
c
a 퐸푚 = (퐸푝푎 + 퐸푝푐)/2 ; b ∆퐸푝 = |퐸푝푎 − 퐸푝푐|; 퐸푤푖푛푑 = 퐸푝푎I − 퐸푝푐III
.
In contrast to the quasi-reversible oxidation processes, scans into the cathodic region
II
show two electrochemically irreversible reductions independent of oxidation II, 퐸푝푐 at −1.52
to −1.63 V and III, 퐸푝푐III at −1.69 to −1.83 V. We attribute II and III, based on prior experience
45
in our laboratory,28 by literature precedents
and by the DFT calculations, where these
processes which involve reduction to Ru(II/I), followed by rapid displacement of a chloride
15

anion (EC processes). All redox processes are in good qualitative agreement with the
electronic structure determined for model complex 5. These hypotheses were further
confirmed from a detailed examination of the optimized DFT studies of the charged states
using a PCM solvation model (see Table S6 in Supporting Information).
Most importantly, the voltammetric studies provide evidence of an extensive redox stability
I
window (Ewind = 2.84 to 2.96 V), (Table 3), defined by the difference between 퐸푝푎 in the
III
anodic region and 퐸푝푐 in the cathodic region.46 This range of values demonstrates the
stability of ruthenium in the oxidation state (2+) which greatly exceeds the physiological
range of E, between ~ +1V to −1V vs. SHE. Thus, the series of complexes 1-4 are expected
to be stable as Ru(II) in vivo, provided that the p-cymene ligand remains coordinated.45
2.6 Langmuir monolayers of complexes 1 – 4
The low solubility of complexes 1-4 in water and PBS buffer (pH 7.4, simulating
physiological conditions) provides an opportunity to investigate Langmuir monolayers of
these organometallics. Complexes 1-4 were spread on PBS buffer subphase and compressed
in order to obtain Langmuir monolayers. Surface pressure-area (-A) and surface potential-
area (V-A) isotherms are illustrated in Fig. 5. The aim of the Langmuir isotherms studies
was to evaluate the shape, stability and thermodynamic behavior of the floating monolayer.47
We are investigating the reorientation, existence of the phase transition and conformational
transformation of the complexes 1-4 on air-water interface to bring us closer to understanding
how these complexes behave at an aqueous interface.
16

Figure 5. (a) Surface pressure vs. area or (insert) compressional modulus-surface pressure
and (b) surface potential vs. area isotherms for 1-4 on PBS buffer (pH = 7.4) subphase.
The isotherms for 1-4 all increase surface pressures to values above zero, implying that
these lipophilic complexes are able to form Langmuir monolayers on a PBS buffer subphase;
significantly, although not amphipathic molecules, they have high and uniform calculated
dipole moments (7.8 to 8.6 Debye; Table 4 and Fig. S14 in Supporting Information). -A
isotherms for 1-4 display that these complexes occupy different molecular areas, obtained by
extrapolating the curve to zero pressure (A0) (Table 4). Interestingly, these values are close to
the areas estimated from v.d.Waals’ radii taken from their crystal structures (Acs) for the very
globular complexes 1 and 2 (Fig. 1) if treated as close packed spheres (the CP area factor is
17

0.91, resulting in effective areas of 161 and 124 Å2 for 1 and 2 when applied to Acs). On the
other hand, the measured areas of 3 and 4 are about half that estimated from the overall
molecular radii of 3 and 4. Consideration of the space-filling models (Fig. 1) indicates
dominant triangular shapes for these complexes, and close-packed trigonal pyramids can be
expected to pack much more densely than spheres.
Table 4. Characteristic parameters of the surface pressure and potential-mean molecular area
isotherms for pure monolayers of 1 – 4 a
μ (Debye) a A0 (Å2)
Acsb (Å2)
177†
퐶푠−1 (mN m-1)
 (mN m-1) V (V)
7.97
7.50
7.93
8.63
143
134
84
25
20
29
30
0.65
0.62
0.63
0.47
89
53
47
27
1
2
3
4
136†
145‡
59
129‡
a From PBE0/Def2TZVP calculations at the crystallographic geometries (see Table S4 in
Supporting Information). b Area of the complexes obtained from crystal structure geometries;
† CP spheres factor = 0.91; ‡ trigonal factor = 0.5 (see discussion).
-A isotherms of 3 and 4 display an additional plateau at 17 and 10 mN m-1, respectively,
which may be related to a phase transition or even to the beginning of the monolayer collapse
process. This further suggests that the layers formed from these –OMe may be more confused
from competing v.d.Waals interactions and dipolar attraction to water molecules, causing
both the higher measured surface density and instability of the monolayers to the applied
pressure.
Taking into account the phase transitions and elasticity of the monolayers as a function of
lateral packing pressure, the compressional moduli were calculated by 퐶푠−1 = −퐴(휕휋/휕퐴)T,
where A is the molecular area at pressure . As observed from 퐶푠−1 data (Table 4 and insert to
Fig. 5a), the interfacial elasticity decreases from 1 to 4. As the surface compressional
modulus of an insoluble monolayer is a measure of the film stiffness,48 the higher value of
퐶푠−1 for 1 means that its films have a lower elasticity; the less flexible monolayers are likely
18

due to the bulky globular structure of 1 (see Fig. 1) leading to close-packed spheres. In
contrast, more flexible monolayers of 4 probably reflect less-efficient packing from the more
angular geometry or from bi-stability induced by competition between dipolar organization
and H-bonding of water with the CH3O oxygen.
V-A isotherms for 1 - 4 display a decrease of the maximum potential, respectively (Table
4). The initial surface potentials are not zero at large molecular areas for complexes 1 – 3,
which may indicate the occurrence of molecular aggregation at the interface from the start.49
The higher maximum surface potential for 1 also reveals that the much larger tBu substituents
lead to an increased capacitance of monolayers formed by 1.
2.7 In vitro cytotoxicity assays
The in vitro cytotoxicity of 1 – 4 was tested against cell lines of human tumor (MDA-MB-
231 and A549) and one non-tumor cell line (MRC-5) using the colorimetric MTT method.
The cytotoxicity of cisplatin was determined under the same conditions as control. As
additional controls, the corresponding free phosphines L1 - L4, as well as ligand oxides LO1
- LO4,21,23 have been tested against the same cell lines since it can be expected that the
phosphines will spontaneously undergo oxidation in the cell test conditions should they be
released from the complex for any reason. IC50 values could be determined in most cases
(Table 5, Fig. 6) but only for 1, 2 and 4 (hence only corresponding ligand values are
reported.)
19

Table 5. IC50 data from in vitro MTT assays for complexes and phosphine ligands.
Tested compounds
IC50 (µM) for indicated cell lines a
MDA-MB-231b
A549 c
26.3 ± 1.6
1.5 ± 0.1
14.7 ± 0.3
> 150
MRC-5 d
> 100
1 e
2
28.0 ± 2.3
3.1 ± 0.1
1.8 ± 0.3
9.4 ± 0.8
118.1 ± 5.6
78.0 ± 12.2
> 150
15.5 ± 0.9
85.5 ± 12.5
118.4 ± 10.6
177.9 ± 24.5
74.2 ± 1.4
51.1 ± 1.8
150.5 ± 4.5
>50 g
4
L1
> 150
L2
> 150
L4
86.9 ± 8.4
168.9 ± 4.0
108.0 ± 8.0
>50 g
55.7 ± 7.3
69.9 ± 8.8
> 150
LO1
LO2
LO4
f
Ru-PPh3
-
12.7 ± 0.2 h
50.6 ± 0.2 h
Cisplatin i
2.4 ± 0.2
11.5 ± 1.2
29.1 ± 0.8
a
Complexes 1, 2, 4 could be tested; all data from triplicate measurements. The
corresponding free phosphines, L1, L2, L4, as well as phosphine oxides LO1, LO2, LO4,
were also tested against the indicated cell lines. Stock solutions made up in DMSO unless
indicated otherwise. b MDA-MB-231 – Human breast adenocarcinoma. c A549 – Human lung
d
e
f
cancer, MRC-5 – Human non-tumoral lung. Stock made in EtOH. Ru-PPh3 = [Ru(η6‐p-
cymene)(PPh3)Cl2]. g Ref. 13. h Ref. 16. i Stock solution in DMF.
20

Figure 6. Cytotoxicity assessment against cell lines: (MDA-MB-23) Mammary
adenocarcinoma, (A549) lung cancer and (MRC-5) lung cells, 48h incubation, for clinical
trials of complexes 1, 2 and 4.
These complexes show promising behavior for cytotoxic activity in human cancer cell lines
and with significantly higher activities than the respective free ligands L1 – L4 (IC50 > 100-
150 μM) as well as the ligand oxides LO1 - LO4, indicating a significant metal effect.
Complexes 2 and 4 are active against tumor cells (MDA-MB-231 and A549), but also
cytotoxic against non-tumor cells. Moreover, 2 has average IC50 values significantly lower
than for the reference drug cisplatin against the A549 cell line. Also, although 1 appears less
active than either 2 or 4, it does not show activity against the non-tumor cells line MRC-5
which is an indication of possible selective attack on cancerous cells. Intriguingly, all
attempts to test activity for complex 3 failed due to visible crystallization of the complex after
addition of DMSO stock solutions to the cell media. This accords well with the experience of
the chemists that 3 demonstrates high crystallinity during crystal growth studies, especially
compared to 1 and 2. The success in the MTT tests for the remaining complexes despite
21

effective insolubility in pure water indicates that processes in the cell growth media
contribute to maintaining dissolution, for example from interaction with the high
concentration of BSA. The film-forming abilities of 1 and 2 may also contribute. The results
obtained could emphasize the importance of lipophilicity of the ligand in the design of new
metallodrugs, with regards to changes in structure. Small variations in the ligands can affect
the activity of these complexes. Here, we have different forms of substitution in the ligands
considered to be of high lipophilicity and the empirical results are in accordance with the
lipophilicity expected for the ligands (L1 > L2 > L3 > L4) (Table 1).
The activity of these complexes 1, 2 and 4 against different cancer cell lines can be
compared to results reported for other complexes with phosphine ligands. Complexes of
t
phosphine ligands such as [RuCl2(η6-p-cymene)(κP-Ph2PR)] where [R= 4-C6H4OSiMe2 Bu,
t
4-C6H4Br, OC (=O)CHCl2, OPh, O(2-C6H4OSiMe2 Bu) had their antitumor activity and
cytotoxicity determined against a breast cancer cell line (MDA-MB-231), ovarian cancer
(A2780) and human skin fibroblasts (HSF). These complexes showed considerable cytotoxic
activity, with lower IC50 values (9.2 ± 1.1 to 29.1 ± 1.4 µM) compared to cisplatin (59.4
µM).6 Some studies report a relationship between the increased lipophilicity of these
phosphine ligands and cytotoxic action. Parveen et al. compared in vitro cytotoxicity assays
of metal complexes of the more lipophilic PPh3 against analogous complexes based on the
PTA ligand (1,3,5-triaza-7-phosphadamantane).20 With the triphenylphosphine (PPh3) ligand
coordinated to the Ru (n6-p-cymene) [3-hydroxy-4(1H)-pyr(id)one]] complex the IC50 values
are reported as 1.58 ± 0.2 µM, while the PTA complex was inactive (IC50> 150 µM). Other
neutral ruthenium complexes with various substituted phosphines have been reported to be
active against tumor cells (Table S12).
3 Experimental methods
22

3.1 General methods
All reactions were carried out under a nitrogen atmosphere using standard Schlenk
techniques. RuCl3.3H2O was purchased from Strem, α-phellandrene, tris-(3,5-
dimethylphenyl)phosphine (L2), tris-(p-methoxy-3,5-dimethylphenyl)phosphine (L3) tris-(p-
methoxyphenyl)phosphine (L4), as well tetrabutylammonium perchlorate (TBAP), were
purchased from Sigma-Aldrich. The tetrahydrofuran (THF) solvent used in the synthesis of
the tris-(3,5-di-tert-butylphenyl)phosphine (L1) was dried under reflux over sodium metal
and benzophenone under a nitrogen atmosphere. HPLC grade dichloromethane (CH2Cl2)
used in the electrochemical measurements (Panreac AppliChem), was treated by distillation
51
from calcium hydride.50 [Ru(p-cymene)(-Cl)Cl]2
and
tris-(3,5-di-tert-
butylphenyl)phosphine (L1)21 were prepared by the previous published methods.
Elemental analyses for C, H and N were performed with a Vario Micro-cube Elemental
Analyzer. FT-IR spectra were recorded on a Bruker Tensor 37 Spectrometer in the range of
4000–400 cm-1. Raman spectra (4000–400 cm-1) were obtained with a Bruker Senterra
dispersive Raman microscope (Fig 31 and Table S13 in SI). UV-Vis spectra of the complexes
(0.1 mmol solutions) were recorded on a Varian Cary spectrophotometer using a quartz cell
in the range 200–900 nm in CH2Cl2. 1D (31P{1H}, 1H, 13C{1H},) and 2D (1H–1H gCOSY,1H–
13C gHSQC, and 1H–13C gHMBC) NMR experiments were recorded in CDCl3-d6 on a Bruker
1
Avance II 300 MHz spectrometer. The H and 31P{1H} NMR chemical shifts are given in
ppm, relative to tretramethylsilane (TMS) and H3PO4 85% at 0 ppm (internal reference),
respectively. The 31P{1H}, 1H and 13C NMR spectra were acquired with 16, 200 and 64 scans,
respectively; spectral widths (sw) of 6393.862 and 25510.203 Hz; relaxation delays (d1) of 1
1
1
1
and 0.2 s; and 90° and 80° pulse lengths. The H–1H gCOSY, H–13C gHSQC and H–13C
gHMBC, spectra were acquired with 8, 16, and 8 scans and spectral widths of 4595.588 for
1
13
F2 and 18852.455 for F1, respectively. All H and C signal assignments in 1-4 have been
23

confirmed by these 2D NMR techniques. Conductivity values were obtained using an
InfolabWTW TetraCon® 325 conductivity bridge in a thermostated bath held at 25.0 °C. The
measurements were performed in acetonitrile solution (1.00∙10-3 mol∙L-1) after mixing and
after 24 h to determine if solvolysis was a factor for the voltammetric and electronic
spectroscopic experiments. A solution of [nBu4N][ClO4] (1∙10-3 mol∙L-1) in CH3CN was used
as the 1:1 electrolyte standard, where the conductivity value for this solution was 197.1
mS∙cm-1.52
3.2 Syntheses
General procedure for synthesis of complexes 1-4
Synthesis consisted of stirring the solution of the precursor complex [Ru(p-cymene)(-
Cl)Cl]2 with the ligand PAr3 in 1:2 ratio in degassed CH2Cl2 (50 mL) for 1 h at room
temperature under N2 atmosphere. After removal of the solvent, to the residue formed, was
added 50 mL of diethyl ether or pentane under stirring for 30 min. The precipitate formed
was filtered, washed with diethyl ether, and dried under reduced pressure. Suitable crystals of
1 were grown in a saturated solution of methanol and pure crystalline samples of 2, 3 and 4
were grown from a saturated solution of CH2Cl2 with slow infusion of diethyl ether.
[Ru(p-cymene)P(m-C(CH3)3-C6H3)3Cl2] 1: Orange solid. Yield: 0.69 g, (72%). MM: 905.13
g mol-1. Elemental analysis calc. for C52H77Cl2PRu (%): C, 68.97; H, 8.57. Found: C, 69.26;
H, 8.28. M.p.: 168-175 °C (dec.). Λm(CH3CN) = 32.9 µS cm-1 (24 h, 25ºC). 1H NMR (300.13
3
MHz, CDCl3): δ H5 0.97 (d, 6H, JHH = 6.9 Hz); H7 1.25 (s, 54 H); H1 1.80 (s, 3H); H4 2.48
(sept, 1H, 3JHH = 7.5 Hz); H2 5.11 (d, 2H, 3JHH = 5.1 Hz); H3 5.17 (d, 2H, 3JHH = 6.0 Hz); H8
2
13
7.36 (s, 3H), H6 7.74 (d, 6H, JPH = 11.1 Hz). C NMR (75.48 MHz, CDCl3: δ Ca 17.82 (s),
24

Cg 22.19 (s); Cf 29.80 (s); Cl 31.52 (s); Ck 35.07 (s); Cd 86.50 (d, 1JPC = 5.40 Hz); Cc 88.65 (d,
1JPC = 3.91 Hz); Cb 96.52 (s); Ce 109.17 (d, 1JPC = 1.51 Hz); Cm 123.68 (s, 3H, 4JPC = 2.3 Hz);
3
Ci 128.79 (d, 2JPC = 9.81 Hz); Ch 133.81 (d, 1JPC = 44.45 Hz); Cj 149.66 (d, JPC = 9.58 Hz).
31P NMR (121.49 MHz, CDCl3): δ 27.20. UV-Vis (CH2Cl2, max/nm) (εmax/L mol−1cm−1):
284sh (8320), 368 (2617), 511 (792). IR (Diamond ATR): 3047 w, 2951 vs, 2903 m, 2861 m,
1575 m, 1472 s, 1419 s, 1359 s, 1286 w, 1244 s, 1196 w, 1135 s, 1051 w, 1022 w, 925 w, 872
m, 803 w, 775 m, 710 vs, 595 vs, 571 m, 529 m, 469 vs, 427 vs.
[Ru(p-cymene)P(m-CH3-C6H3)3Cl2] 2: Orange solid. Yield: 0.88 g, (95%). MM: 652.65 g
mol-1. Elemental analysis calc. for C34H41Cl2PRu (%): C, 62.57; H, 6.33. Found: C, 63.12; H,
1
6.14. M.p.: 186-196 °C (dec.). Λm(CH3CN) = 55.8 µS cm-1 (24 h, 25ºC). H NMR (300.13
3
MHz, CDCl3): δ H5 1.04 (d, 6H, JHH = 6.9 Hz); H1 1.96 (s, 3H); H7 2.28 (s, 18H); H4 2.75
(sept, 1H, 3JHH = 6.6 Hz); H3 4.99 (d, 2H, 3JHH = 6.0 Hz); H2 5.09 (d, 2H, 3JHH = 5.7 Hz); H8
6.99 (s, 3H); H6 7.46 (d, 6H, 2JPH = 10.8 Hz). 13C NMR (75.48 MHz, CDCl3): Ca 17.58 (s), Ck
1
21.49 (s); Cg 21.84 (s); Cf 30.18 (s), Cd 85.77 (d, CH, 1JPC = 6.0 Hz); Cc 90.67 (d, JPC = 4.0
Hz); Cb 95.00 (s); Ce 110.00 (s); Ci 131.90 (d, 2JPC = 11 Hz); Cl 131.86 (d, 4JPC = 2.8 Hz), Ch
1
3
31
133.91 (d, JPC = 45.3 Hz); Cj 137.14 (d, JPC = 9.2 Hz). P NMR (121.49 MHz, CDCl3): δ
24.20. UV-Vis (CH2Cl2, max/nm) (εmax/L mol−1cm−1): 289sh (10820), 371 (3198), 510
(3500). IR (Diamond ATR): 3039 w, 2961 w, 2909 w, 2850 w, 1579 w, 1463 m, 1410 m,
1369 m, 1264 w, 1172 w, 1125 m, 1032 w, 897 w, 851 s, 792 w, 693 vs, 554 vs, 525 m, 484
s, 466 m, 437 vs, 408 s, 398 s.
25

[Ru(p-cymene)P(p-MeO-m-CH3-C6H2)3Cl2] 3: Red solid. Yield: 0.92 g, (93%). MM:
742.73 g mol-1. Elemental analysis calc. for C37H47Cl2O3PRu (%): C, 59.83; H, 6.38. Found:
1
C, 60.21; H, 6.40. M.p.: 199-209 °C (dec.). Λm(CH3CN) = 58.0 µS cm-1 (24 h, 25ºC). H
3
NMR (300.13 MHz, CDCl3): δ H5 0.99 (d, 6H, JHH = 7.2 Hz); H1 1.95 (s, 3H); H7 2.22 (s,
18H); H4 2.68 (sept, 3JHH = 6.9 Hz); H8 3.71 (s, 9H); H3 4.97 (d, 2H, 3JHH = 6.0 Hz); H2 5.08
(d, 2H, 3JHH = 6.3 Hz); H6 7.44 (d, 6H, 2JPH = 10.5 Hz). 13C NMR (75.48 MHz, CDCl3): δ Ck
1
16.31 (s); Ca 17.60 (s), Cg 21.77 (s); Cf 30.08 (s), Cm 59.59 (s); Cd 85.98 (d, JPC = 6.0 Hz); Cc
1
1
90.49 (d, JPC = 4.0 Hz); Cb 94.77 (s); Ce 109.24 (s); Ch 129.02 (d, JPC = 47 Hz); Cj 130.26
3
2
31
(d, JPC = 11 Hz); Ci 134.85 (d, JPC = 10 Hz); Cl 158.49 (d, JPC = 3 Hz). P NMR (121.49
MHz, CDCl3): δ 22.00. UV-Vis (CH2Cl2, max/nm) (εmax/L mol−1cm−1): 249sh (40140), 277sh
(21040), 366 (13180), 510 (3500). IR (Diamond ATR): 3041 w, 2957 w, 2923 w, 2862 w,
1585 w, 1474 m, 1387 m, 1276 s, 1208 s, 1115 vs, 1053 vs, 1003 vs, 861 s, 799 w, 759 w,
693 w, 669 w, 613 vs, 576 m, 539 m, 517 m, 452 s, 384 s.
[Ru(p-cymene)P(p-MeO-C6H4)3Cl2] 4: Red solid. Yield: 0.99 g, (90%). MM: 658.53 g mol-
1. Elemental analysis calc. for C31H35Cl2O3PRu (%): C, 56.54; H, 5.35. Found: C, 57.23; H,
1
5.20. M.p.: 176-187 °C (dec.). Λm(CH3CN) = 53.8 µS cm-1 (24 h, 25ºC). H NMR (300.13
3
3
MHz, CDCl3): δ H5 1.13 (d, 6H, JHH = 6.9 Hz); H1 1.85 (s, 3H) ; H4 2.89 (sept, 1H, JHH
=
3
6.8 Hz) ; H8 3.80 (s, 9H); H2 4.94 (d, 2H, 3JHH = 5.1 Hz); H3 5.22 (d, 2H JHH = 6.3 Hz); H7
3
3
3
13
6.86 (d, 6H, JHH = 9.0 Hz); H6 7.71 (d of d, 6H, JHH= 7.2 Hz; JPH= 9.0 Hz) . C NMR
1
(75.48 MHz, CDCl3): δ Ca 17.86 (s); Cg 21.93(s); Cf 30.30 (s); Cl 55.27 (s,); Cd 87.36 (d, JPC
1
1
= 5.3 Hz); Cc 88.52 (d, JPC = 3.0 Hz), Cb 95.76 (s); Ce 111.03 (d, JPC = 3.8 Hz); Ci 113.54 (s,
1
3
2JPC = 10.94 Hz); Ch 125.34 (d, JPC = 51 Hz); Cj 135.83 (d, JPC = 10.72 Hz); Ck 160.90 (d,
4JPC = 2.3 Hz). 31P NMR (121.49) MHz, CDCl3): δ 23.70. UV-Vis (CH2Cl2, max/nm) (εmax/L
26

mol−1cm−1): 250sh (42180) 274sh (21780), 370 (14736), 509 (5010). IR (Diamond ATR):
3048 w, 2997 w, 2990 w, 2970 w, 2937 w, 2833 w, 1585 s, 1566 m, 1497 s, 1462 m, 1439 m,
1407 w, 1381 w, 1282 s, 1251 vs, 1186 s, 1174 s, 1089 vs, 1018 vs, 718 w, 645 m, 625 m,
610 m, 537 vs, 526 vs, 494 s, 458 s, 423 vs.
3.3 Langmuir monolayers
Ru complexes 1-4 solutions (5∙10-4 mol L-1) were prepared in chloroform. Surface pressure
and surface potential experiments were performed with a KSV Langmuir trough (KN 1003
model, KSV Instruments, Ltd., Helsinki, Finland) with a maximum surface area of 273 cm2.
The trough is equipped with a Wilhelmy plate (chromatography paper Whatman Chr1) as
surface pressure sensor, and with a KSV Kelvin probe as surface potential sensor, using the
vibrating plate method. The experiments were performed at ~20 ºC. Langmuir monolayers
were formed by spreading Ru complexes on PBS buffer (pH 7.4, 150 mM NaCl), and after an
equilibration time of 15 min, the monolayers were compressed at a speed of 10 mm min-1.
Each isotherm experiment was repeated at least three times.
3.4 Single Crystal X-ray Crystallography
Suitable crystals for X-ray diffraction were grown initially by layering pentane onto CDCl3
solutions employed for NMR and cooling overnight at -30 °C. Suitable crystals 2∙2CHCl3,
3∙1.5CHCl3 and 4∙0.5CHCl3 were grown in this way but all were shown to contain CDCl3. As
part of the purification and bulk recrystallizations required to prepare samples suitable for
biological tests, solvent-free crystals of all the complexes were also obtained; only these
solvent-free materials were used for the cytotoxicity and Langmuir film investigations.
Crystals of 1 from slow evaporation of a pentane solution, 2 by slow evaporation of a
methanol solution; both 3 and 4 from diethyl ether diffusion into a CH2Cl2 solution of the
complexes in a large, sealed Schlenk tube.
27

Crystals were selected under a microscope and mounted on a 200 μm MiTeGen loop using
Paratone™ oil, and cooled using the diffractometer Cryostream 800 device. Data was
collected on a Rigaku-Oxford SuperNova dual (Mo, Cu) wavelength instrument equipped
with a Pilatus 200K HPC detector. All data collection, reduction and integration were
undertaken using CrysAlisPro 1.171.38.41 (Rigaku Oxford Diffraction, 2015). Empirical
absorption corrections were undertaken with the Scale 3 Abspack routine within CrysAlisPro.
53
Structures were solved and refined in the Olex2 2.1 environment using SHELXT for data
54
solution
and SHELXL for model refinement.54 Final data analysis and structure
visualization was undertaken with Mercury CSD 4.0.0.56
3.5 Electronic structure of the model complex 5 from DFT and TD-DFT computations
Ab-initio calculations using Density Functional Theory (DFT) in gas phase, zero Kelvin,
vacuum, dichloromethane and acetonitrile solvent conditions were carried out to verify the
accuracy of the molecular geometry of complex 5. In this study, all calculations were
performed on GAUSSIAN09 program and calculation parameters were B3LYP functional,
Gaussian-type 6-31+G(d) basis set for C, H, N, P, and F atoms and Sapporo double-zeta
(SPK-DZCD) basis set for Ru atom.42 To optimize the molecular geometries, solvent effect
was inserted from Polarized Continuum Method (PCM); 10-6 Hartree and 10-8 Hartree for
SCF based on Hartree-Fock formalism and total energy criteria, respectively. UV-Vis
experiments were clarified from Frontier Molecular Orbitals (FMOs) contribution for neutral
molecule because they can localize electronic densities where there is a possibility of
electronic excitation. The most important FMOs for excitation energies are the highest-
occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO);
however, other orbitals close in energy are sometimes important, especially when (near-
)degenerate levels are involved mainly. Molecular geometries for oxidized states such as -1,
28

+1 and +2 charges were optimized to elucidate voltammetry results, including ligand
displacement processes. Dipole moments of 1 – 4 were calculated at the crystal structure
geometries using dispersion-corrected RPE0/TZVP DFT calculations.
3.6 Cyclic voltammetry
Cyclic voltammetry measurements were performed at RT using an Autolab
potentiostat/galvanostat (Type III, Metrohm-Eco Chemie). These experiments were
performed using dichloromethane (CH2Cl2) of HPLC grade (Panreac AppliChem), containing
0.10 M Bu4NClO4 (tetrabutylammonium perchlorate-PTBA) as a supporting electrolyte and a
three electrode electrochemical cell: 3 mm2 glassy carbon working electrode, platinum
auxiliary electrode and Ag/AgCl reference electrode (in a Luggin capillary) in
dichloromethane. Under these conditions, ferrocene (Fc) is oxidized at +0.375 V and the data
is corrected for the Fc+/0 scale. Analyses were performed in solution with analyte
concentrations of 2.10-3 mol L-1, under an N2 atmosphere, over variable scan rates from 50-
2000 mV s-1, in both cathodic and anodic initial scan directions.
3.7 Cell culture and In vitro cytotoxicity assays
The cytotoxicity of the ruthenium complexes were evaluated against three cell lines, MDA-
MB-231 (ATCC: HTB-26, Breast adenocarcinoma), A549 (ATCC: CCL-185, Lung cancer)
and MRC-5 (ATCC: CCL-171, Non-tumoral lung), that were grown in Dulbecco's Modified
Eagle Medium (DMEM) culture containing 10 % fetal bovine serum (FBS), penicillin (100
IU/ml), streptomycin (100 mg∙mL−1) and L-glutamine (2 mM). All cell lines were kept in a
humid oven with 5% CO2 and a constant temperature of 37 ºC in cell culture bottles. Cells
were trypsinized, counted and plated on 96-well culture plates (Corning Costar) (1.5 x 104
cells/well for MDA-MB-231 and A549, and 1∙104 cells/well for MRC-5). The plates were
29