E ﬀ ect of cysteine thiols on the catalytic and anticancer activity of Ru( II ) sulfonyl-ethylenediamine complexes

* a We have synthesized a series of novel substituted sulfonyl ethylenediamine (en) Ru II arene complexes 1 – 8 of [( η 6 -arene)Ru(R 1 -SO 2 -EnBz)X], where the arene is benzene, HO(CH 2 ) 2 O-phenyl or biphenyl (biph), X = Cl or I, and R 1 is phenyl, 4-Me-phenyl, 4-NO 2 -phenyl or dansyl. The ‘ piano-stool ’ structure of complex 3 , [( η 6 -biph)Ru(4-Me-phenyl-SO 2 -EnBz)I], was con ﬁ rmed by X-ray crystallography. The p K *a values of their aqua adducts were determined to be high (9.1 to 9.7). Complexes 1 – 8 have antiproliferative activity against human A2780 ovarian, and A549 lung cancer cells with IC 50 values ranging from 4.1 to >50 µM, although, remarkably, complex 7 [( η 6 -biph)Ru(phenyl-SO 2 -EnBz)Cl] was inactive towards A2780 cells, but as potent as the clinical drug cisplatin towards A549 cells. All these complexes also showed catalytic activity in transfer hydrogenation (TH) of NAD + to NADH with sodium formate as hydride donor, with TOFs in the range of 2.5 – 9.7 h − 1 . The complexes reacted rapidly with the thiols glutathione (GSH) and N -acetyl- L -cysteine (NAC), forming dinuclear bridged complexes [( η 6 -biph) 2 Ru 2 (GS) 3 ] 2 − or [( η 6 -biph) 2 Ru 2 (NAC-H) 3 ] 2 − , with the liberation of the diamine ligand which was detected by LC-MS. In addition, the switching on of ﬂ uorescence for complex 8 in aqueous solution con ﬁ rmed release of the chelated DsEnBz ligand in reactions with these thiols. Reactions with GSH hampered the catalytic TH of NAD + to NADH due to the decomposition of the complexes. Co-administration to cells of complex 2 [( η 6 -biph)Ru(4-Me-phenyl-SO 2 -EnBz)Cl] with L -buthionine sulfoximine ( L -BSO), an inhibitor of GSH synthesis, partially restored the anticancer activity towards A2780 ovarian cancer cells. Complex 2 caused a concentration-dependent G1 phase cell cycle arrest, and induced a signi ﬁ cant level of reactive oxygen species (ROS) in A2780 human ovarian cancer cells. The amount of induced ROS decreased with increase in GSH concentration, perhaps due to the formation of the dinuclear Ru-SG complex.


Introduction
Studies of the antiproliferative activity of organometallic complexes, especially the Ir, 1-3 Os, 4 Rh 5,6 and Ru 7, 8 complexes, have been stimulated by the clinical success of platinum anticancer drugs such as cisplatin, carboplatin and oxaliplatin. [9][10][11] The activity of organometallic half-sandwich complexes can be tuned by the choice of the arene, chelating and monodentate ligands. 12,13 In recent years, the possibility of using synthetic organometallic complexes as catalysts for transfer hydrogenation (TH) reactions in cells has emerged as a potential way of modulating intracellular redox states and inducing cancer cellkilling by a mechanism which can circumvent resistance. 14,15 Noyori-type Ru II complex JS2 [(η 6 -p-cym)Ru(TsEn)Cl] exhibits potent catalytic TH activity in the reduction of co-enzyme NAD + to NADH. Up to a 90% decrease in cancer cell viability was observed in the presence of non-toxic doses of sodium formate (Fig. 1), with the ratio of NAD + /NADH greatly decreasing in cancer cells. 16 The analogue complex [(η 6 -p-cym)Ru (TsEnBz)Cl] shows enhanced activity in the TH of NAD + with formate as hydride donor, 17 however, the reaction appeared to be hampered by interaction of the Ru II complex with thiol-containing molecules, including the tripeptide glutathione, γ-L-Glu-L-Cys-Gly (GSH, Fig. 1). 17 We have therefore explored the role of GSH in mediating reactions of these half-sandwich organometallic catalysts in this work.
GSH is an important tripeptide that exists ubiquitously in all eukaryotic cells (in mM concentrations); it can be oxidized, e.g. to GSSG, to protect cells from being damaged by reactive oxygen species (ROS, metabolic side products). [18][19][20] GSH can also interact with metal complexes in cells, and many organometallic complexes are thiophilic. 21 By taking advantage of such interactions, thiols have been used as switch-on probes to trigger luminescence or fluorescence in cells, by either reacting with a probe or by displacement of fluorescent ligands, which can be used to map their distributions in cells. 22, 23 We have found that the tethered Ru II complex [Ru(η 6 -Ph(CH 2 ) 3ethylenediamine-N-Ts)Cl] can rapidly react with GSH to form a Ru-SG adduct which can decompose slowly. 24 In the present work, we have studied the effect of the amino acid cysteine and the tripeptide glutathione on the catalytic and anticancer activity of Ru II sulfonyl ethylenediamine complexes 1-8 [(η 6 -arene)Ru(R 1 -SO 2 -EnBz)X], where the arene is benzene, HO(CH 2 ) 2 O-phenyl, or biphenyl, and R 1 is various sulfonyl substituents ( Table 1). The catalytic TH reduction of NAD + to NADH using sodium formate as hydride source was studied, as well as reactions of 1-8 with the tripeptide GSH and N-acetyl-L-cysteine (NAC) (Fig. 1), and the effect of GSH on the TH of NAD + . The anticancer activity of complexes 1-8 towards A2780 ovarian and A549 lung human cancer cells was determined together with the effect of coadministration with GSH, NAC, and the redox modulator L-buthionine sulfoximine (L-BSO), an inhibitor of GSH biosynthesis. Cell cycle arrest and the influence of GSH on induction of ROS were also investigated. The study revealed an interesting and unusual role for GSH in the biological activity of this class of organometallic transfer hydrogenation catalysts.

X-ray crystal structure
A crystal of complex 3 [(η 6 -biph)Ru(4-Me-phenyl-SO 2 -EnBz)I] suitable for X-ray crystallographic analysis was obtained by the slow diffusion of diethyl ether into a saturated methanol solution of complex 3 at ambient temperature. Selected bond lengths and angles for complex 3 are given in Table 2, crystallography data are listed in Tables S1 and S2, † and the molecular structure is shown in Fig. 2. Complex 3 adopts a pseudotetrahedral geometry with a η 6 -phenyl ring on one face of the metal centre occupying 3 coordination sites. The sulphonamide nitrogen of the ethylenediamine ligand is deprotonated and bound as a monoanionic bidentate ligand to Ru, together with an iodido ligand completing the coordination sphere of the complex. pK * a determination The pK * a values ( pK a determined in deuterated solvent) of the aqua adducts of complexes 1, 2 and 4-7 were determined by    (8) titration over the pH* range from 2 to 12. The 1 H NMR chemical shifts of protons of the sulfonyl phenyls as a function pH were fitted to the Henderson-Hasselbalch equation (Fig. S26 in the ESI †). All the pK * a values of these aqua complexes are in the range of 9.10-9.75 (Table 3).

Nucleobase binding
The interaction of complex 2 with model nucleobase 9-ethylguanine (9-EtG) was studied by 1 H NMR spectroscopy (Fig. S27 in the ESI †). The reactions were performed by titrating a solution of complex 2 (2 mM, 10% MeOD-d 4 in D 2 O) with 9-EtG in D 2 O at 310 K, in 0.5 mol equiv. steps. Complex 2 reacted rapidly with 9-EtG, and binding appeared to be complete within 5 min. Adduct formation was confirmed by following a new set of peaks, with slow exchange on the NMR time scale between the aqua adduct of 2 and its 9-EtG adduct. Peaks for complex 2 disappeared when 1.0 mol equiv. of 9-EtG was added, indicating strong binding and completion of the reaction ( Fig. S27 in the ESI †).

Reactions with glutathione
Reactions of complex 2 with GSH were investigated via a series of concentration-dependent experiments and monitored by 1 H NMR, at pH* 7.2, ( pH* = pH meter reading in deuterated solvent), 310 K. Complex 2 (2 mM) and GSH were mixed in the mol ratio of 1 : X, where X = 1, 2, 5, 10, respectively, in the mixed solvent of MeOD-d 4 and D 2 O, 1 : 9 (v/v). Reaction of complex 2 with 1.0 mol equiv. GSH led to the disappearance of 1 H NMR peaks for 2 in the biphenyl ligand region of the spectrum within 10 min, and generated a new set of peaks shifted to lower field ( Fig. S28 in the ESI †). No further change in this region was observed with further addition of GSH (2-10 mol equiv.). However, the products were difficult to identify based on the 1 H NMR spectra alone (Fig. S28 in the ESI †).

Identification of GSH/NAC adducts by LC-MS
HPLC and LC-MS were used to elucidate the nature of the products from reactions of complex 2 with GSH. Similar reactions with N-acetyl-L-cysteine (NAC) were studied for comparison.
Complex 2 (2 mM in MeOH/H 2 O, 1 : 9 (v/v)) and GSH or NAC (10 mol equiv., in H 2 O) were mixed in a vial and pre-incubated at 310 K for 24 h ( pH 7.10 ± 0.1). As can be seen from Fig. S29, † the reactions proceeded with >95% and 100% conversions to form the Ru II -SG and Ru II -NAC adducts, respectively, as determined by HPLC. HPLC peak p4, assignable to complex 2, disappeared after 24 h co-incubation at 310 K, with two new peaks p1 and p2 emerging (Fig. S29 in the ESI †). Subsequently, reactions were studied by LC-MS using the same conditions. MS peaks for dinuclear complexes [(η 6 -biph) 2 Ru 2 (GS) 3 ] 2− 2a and [(η 6 -biph) 2 Ru 2 (NAC-H) 3 ] 2− 2b are assignable to HPLC peaks p1 and p2, respectively. Displaced free chelating TsEnBz ligand was detected as peak p3. The peak assignments are listed in Table S3 in the ESI. † Next, the isolation of the dinuclear complexes 2a and 2b was attempted by HPLC using a ZORBAX Eclipse XDB-C18 Semi-preparative column (9.4 × 250 mm). Complex 2b was collected and characterized by 1 H NMR (Fig. S30 in the ESI †). A high resolution MS peak at 715.6011 m/z was observed, which corresponds to [(η 6 -biph) 2 Ru 2 (GS) 3 +4H] 2+ (2a in Fig. S31 in the ESI †), and the high resolution peak at 998.0339 m/z can be assigned to [(η 6 -biph) 2 Ru 2 (NAC) 3 ] + (2b in Fig. S32 in the ESI †). Ellipsoids are shown at the 50% probability level. All hydrogen atoms have been omitted for clarity. The catalytic efficiency of complex 2 was little affected, with the TOF decreasing slightly from 7.9 ± 0.4 h −1 to 6.29 ± 0.53 h −1 when 0.2 mol equiv. GSH was present. However, the TOF dropped dramatically to 0.91 ± 0.43 h −1 when 0.5 mol equiv. GSH was co-administered. The reaction totally stopped when 1 mol equiv. or more GSH was added, probably due to the completing reaction of 2 with GSH.
Fluorescence-detected thiol-triggered chelated ligand release from complex 8 The fluorescence of the chelated phenylsulfonyl ethylenediamine dansyl-ligand DsEnBz was fully quenched when it was present as a chelated ligand in complex 8 (Table 1) The reaction was initially detected by UV-vis spectroscopy at 310 K, pH 7 (Fig. 3a). Next, a ca. 200-fold increase in emission intensity was observed on adding GSH (10 mol equiv.) to an aqueous solution of complex 8, pH 7, 310 K, on excitation at 350 nm (Fig. 3b). NAC induced a stronger increase in fluorescence under the same conditions (ca. 1.7-fold stronger than GSH). When excited at 405 nm, the emission intensity was relatively lower, only about 40-fold intensity for GSH and 60-fold for NAC compared to complex 8 alone, Fig. 3c. In order to confirm the importance of the thiol groups in these reactions, complex 8 was reacted with the amino acids L-leucine and L-tryptophan, and the thiol-containing molecule 1-butanethiol. No increase in fluorescence was observed when 8 was mixed with thiol-free amino acids; however, a relatively strong fluorescence emission was found when 1-butanethiol was added, indicating the key role of the thiol group.
Effect of L-buthionine sulfoximine on antiproliferative activity L-Buthionine sulfoximine (L-BSO) is a specific inhibitor of γ-glutamylcysteine synthetase, 26 an enzyme involved in the biosynthesis of GSH. Treatment with L-BSO can scavenge the intracellular GSH levels up to 40%, which effectively hampers

Paper
Dalton Transactions cellular GSH synthesis. 27 The anticancer activity of complex 2 against A549 human lung cancer cells after co-incubation with L-BSO was determined, to provide insights into the role of GSH in antiproliferative activity. Complex 2 was coadministered with three different concentrations of L-BSO: 1, 5 and 50 µM. After 24 h co-incubation with L-BSO (at concentrations of 1 and 5 µM), the antiproliferative activity of 2 (IC 50 , ca. 13 µM) remained unchanged; but increased by 61% to 8.3 ± 0.5 µM when co-treated with 50 µM L-BSO.

Effect of GSH and NAC on anticancer activity
GSH often acts as a detoxification agent for metal-based drugs in cells, and some drug-resistant cancer cells are capable of generating higher levels of GSH to circumvent damage. 21 N-Acetyl-L-cysteine (NAC) is scavenger of reactive oxygen species (ROS), which can block cisplatin related caspase-3 activation and cell apoptosis. 28 Since complex 2 can react rapidly with GSH to form the dimer 2a [(η 6 -biph) 2 Ru 2 (GS) 3 ] 2− , co-administration of complex 2 with GSH (5, 10 and 50 µM) was studied, to investigate the effect of GSH on the antiproliferative activity in A2780 human ovarian carcinoma cells. Cells were incubated with three concentrations of GSH (5, 10 and 50 µM) as controls. The results indicated that GSH exposure only is not toxic towards A2780 cancer cells, Table S4 (in the ESI †). After 72 h of recovery time in drug-free medium, cell survival was evaluated using the sulforhodamine B colorimetric assay. As shown in Fig. 4, the antiproliferative activity decreased gradually with increase of GSH concentration, giving IC 50 values of 22.41 ± 1.25, 29.9 ± 2.1 and >50 µM towards A2780 cells and, 27.33 ± 0.54, 43.93 ± 3.54 and >50 µM towards A549 cancer cells, for GSH concentrations of 5, 10, and 50 µM, respectively.
Co-treatment with complex 2 and NAC gave rise to a similar trend, in which the anticancer activity decreased with an increase in NAC concentration (under similar conditions as those for GSH above), IC 50 values are 25.8 ± 0.9, 39.9 ± 0.9 and >50 µM, for NAC concentrations of 5, 10 and 50 µM, respectively.

Cell cycle arrest
A cell cycle analysis for A2780 human ovarian cancer cells treated with the representative complex 2 was performed using propidium iodide staining and flow cytometry. A2780 cancer cells were incubated with IC 50 or 2 × IC 50 concentrations of complex 2 for 24 h. In comparison to negative control population, complex 2 increased cell cycle arrest at the G1 phase when the drug concentration increased from IC 50 to 2 × IC 50 , Fig. 5.

Reactive oxygen species (ROS)
The level of reactive oxygen species (ROS) induced by complex 2 and GSH in combination with 2 in A2780 human ovarian cancer cells was determined at the IC 50 concentration of the complex by flow cytometry fluorescence analysis (Fig. 6). These experiments were carried out following previously described protocols. 29 A2780 cancer cells were treated at a fixed IC 50 concentration of 2 and GSH (0.5 or 5 µM) without any recovery time. The total level of oxidative stress (including H 2 O 2 , peroxy and hydroxyl radicals, peroxynitrite, and NO) in the FITC-A

Discussion
The X-ray crystal structure of complex 3 [(η 6 -biph)Ru(4-Mephenyl-SO 2 -EnBz)I] shows that it has the typical half-sandwich 'piano-stool' geometry ( Fig. 2), similar to related Ru II complexes. 25 In aqueous solution the ability of these halido sulfonylethylenediamine complexes to undergo hydrolysis and gene-rate aqua adducts, 17,25,26 decreases with the halido ligand in the order of Cl ≈ Br > I. Halido complexes with strong δ-donor ligands like en (ethylenediamine) and acac (acetylacetone), hydrolyse quickly and produce basic hydroxido adducts ( pK a > 7), while the π-acceptor ligands like azopyridine, hydrolyse slowly and give more acidic aqua complexes. 26,31 The aqua adducts of the catalysts studied here with high pK * a values of 9.1-9.7 (Table 3), would be predominantly protonated in physiological media ( pH 7.4). Aqua adducts are usually much more reactive than the corresponding hydroxido adducts and more readily undergo substitution reactions, which is favourable for a catalytic centre.

Paper Dalton Transactions
As an inhibitor of the enzyme γ-glutamyl cysteine synthetase, L-BSO can limit the cellular synthesis of GSH, and enhance ROS levels to induce cell apoptosis. Co-treatment of organometallic Ru II or Os II complexes with L-BSO has been developed as a strategy to overcome GSH mediated detoxification of drugs. 46,47 For example, L-BSO can restore CDDP activity against several CDDP-resistant cancer cell lines. L-BSO has been shown to cause a significant reduction in A2780 cellular GSH levels (ca. 50% with 5 μM L-BSO) and a significant enhancement of anticancer activity towards ovarian cancer cells upon co-administration of organo-Os complex [Os(η 6 -pcym)( p-NMe 2 -Azpy)I]PF 6 with L-BSO (5 µM dose), with 87% improvement in anticancer activity at an equipotent 2 × IC 50 concentration of the complex. 27,28 However, such restoration of antiproliferative activity by L-BSO only occurs when a complex is already biologically active. 27 In the present work, enhancement of the anticancer activity against A549 cancer cells was observed at an L-BSO concentration of 50 µM with IC 50 decreasing from ca. 13 to 8.3 µM; with levels of GSH reduced to ca. 61%. 28 High L-BSO concentrations probably interfere with cellular GSH synthesis, while excessive GSH might react with the complex to destroy catalytic activity but promote the generation of biologically active dinuclear Ru(II) species. 42 The cell cycle arrest study of complex 2 in A2780 cancer cells revealed a dose-dependent cell population increase in G1 phase (66.7 ± 1.5% to 75.2 ± 0.2% and 80 ± 2% at IC 50 and 2 × IC 50 concentrations), but a cell population depletion in G2/M and S phase, which implies that complex 2 is less likely to have DNA as a target site, in agreement with the previous study. 16 DNA-targeted compounds normally cause cell accumulation in the S phase or G2/M phase, e.g. cisplatin. 48 Reactive Oxygen Species (ROS) display important roles in cell metabolism. As respiratory side-products of mitochondria, over-production from ROS damages proteins or causes oxidation of DNA nucleobases to induce cell apoptosis, and ROSmediated apoptotic signalling is usually associated with reduction of cytosol or mitochondrial GSH levels. 49,50 Organo Ir, Os and Ru complexes have been widely reported as anticancer agents which can induce cell apoptosis via ROS-involving pathways. 28,51-53 Complex 2 can induce significant amounts of superoxide in A2780 ovarian cancer cells (up to 16% of the cell population, Table S5 in the ESI †). Co-administration of complex 2 with GSH reduces both superoxide levels and antiproliferative activity against A2780 human ovarian cancer cells. The levels of superoxide showed an inverse relationship with the concentrations of GSH added (Table S5 in the ESI †), and a similar trend in antiproliferative activity with GSH concentration was also observed. This might imply that ROS play an important role in killing cancer cells for these complexes.

Conclusions
A series of Ru II sulfonyl-substituted ethylenediamine complexes of the type [(η 6 -arene)Ru(R 1 -SO 2 -EnBz)X] (where the arene is benzene, biphenyl or HOCH 2 CH 2 O-phenyl, R 1 is 4-Me-phenyl, phenyl, 4-F-phenyl, 4-NO 2 -phenyl or Dansyl, and X is halide) were synthesized and fully characterized. The halfsandwich structure of complex 3 was confirmed by X-ray crystallography. These complexes retained high potency in transfer hydrogenation of coenzyme NAD + with formate as hydride donor. The introduction of the substituted arene HOCH 2 CH 2 O-phenyl (complex 4) significantly improved the water solubility of the complex, albeit with reduced catalytic activity.
Complex 2 [(η 6 -biph)Ru(4-Me-phenyl-SO 2 -EnBz)Cl] bound strongly to 9-ethylguanine to form the 2-9-EtG-N7 adduct. However, complex 2 caused only G1 cell cycle arrest in a concentration-dependent manner, and is unlikely to target DNA in cells. Complex 2 exhibited a high affinity for GSH and NAC, to form the Ru-thiolate bridged dimers [(η 6 -biph) 2 Ru 2 (GS) 3 ] 2− and [(η 6 -biph) 2 Ru 2 (NAC-H) 3 ] 2− . Co-incubation of complex 2 with increasing GSH concentrations effectively reduced induction of reactive oxygen species in cells, and decreased the antiproliferative activity significantly. Such reactions with GSH lead to the release of the chelated diamine ligand, and hence trigger the fluorescence of the free dansyl ligand upon release from complex 8 [(η 6 -biph)Ru(DsEnBz)Cl], which might provide a basis for a study of such release in cells. Reactions of Ru II complexes with GSH to form the dinuclear species may offer a new class of TH catalysts that can also form cytotoxic thiolatebridged complexes in cells. Future work on strategies to control the rate and extent of reactions of these catalysts with GSH, might lead to new concepts in the design of this class of multi-targeting candidate metallodrugs.

Materials
Dansyl chloride was purchased from Sigma-Aldrich. The Ru IIη 6 -arene precursor dimers were prepared following literature methods, 25,54 as were the ligands (synthesis is presented in ESI †). 17 The solvents used for NMR spectroscopy were purchased from Sigma-Aldrich and Cambridge Isotope Laboratories Inc. Non-dried solvents used in syntheses were obtained from Fisher Scientific. Glutathione and N-acetyl-Lcysteine were purchased from Fisher Scientific.

X-ray crystallography
Diffraction data for complex 3 were collected on an Oxford Diffraction Gemini four-circle system with an AtlasS2 CCD area detector. The structure of complex 3 was refined by full-matrix least-squares against F 2 using Olex2 55 and was solved by with the ShelXT 56 structure solution program using Intrinsic Phasing and refined with the ShelXL 57 refinement package using Least Squares minimisation. The atoms from the sulphonamide nitrogen to the end of the chain (C10 C11 N12 C13) were modelled as disordered over two positions related by a small ruffle in the chain. The occupancy of the two positions was linked to a free variable which refined to 86 : 14. The minor component was refined isotopically. The NH of the major component was located in a difference map though both it and the NH of the minor position were placed at calculated positions for the rest of the refinement. The data were processed by the modelling program Mercury 3.8. X-ray crystallographic data for complex 3 have been deposited in the Cambridge Crystallographic Data Centre under the accession number CCDC 2117792. †

In vitro growth inhibition assays
The antiproliferative activity of complexes 1-8 was determined against A2780 human ovarian and A549 human lung cancer cells. Briefly, 5000 cells per well were seeded in 96-well plates. The plates were left to pre-incubate with drug-free medium at 310 K for 48 h before adding different concentrations of the tested compounds. Exact drug (Ru) concentrations were determined by ICP-OES. A drug exposure period of 24 h was allowed. After this, supernatants were removed by suction and each well was washed with PBS. A further 72 h were allowed for the cells to recover in drug-free medium at 310 K. The Sulforhodamine B (SRB) assay was used to determine cell viability. IC 50 values, as the concentration that causes 50% cell death, were determined as duplicates of triplicates in two independent sets of experiments and their standard deviations were calculated. For cell growth inhibition by GSH/NAC with complex 2, GSH/NAC at concentrations of 5, 10 and 50 µM were added to the cells first, followed by complex 2 (within 10 min), and the cytotoxicity was monitored using the SRB assay described above. After 24 h co-incubation with complex 2 and GSH, cancer cell viability was assessed after washing the cells with PBS.

Cell cycle arrest
Approximately 1.5 × 10 6 per well of A2780 human ovarian cancer cells were cultured in a six-well plate and pre-incubated in drug-free media at 310 K for 24 h, after which complex 2 at equipotent IC 50 concentration were added. After drug exposure for 24 h, supernatants were removed by suction and cells were washed with PBS. Then A2780 cells were harvested using trypsin-EDTA and fixed with cold 70% ethanol for 2 h. DNA staining was obtained by re-suspension of cell pellets in PBS (containing propidium iodide (PI) and RNase). Cell pellets were washed and re-suspended in PBS before being analysed in a Becton Dickinson FACScan flow cytometer using excitation of DNA-bound PI at 536 nm, with emission at 617 nm. Data were processed with Flowjo software.

ROS determination
ROS/superoxide induction in A2780 cells induced by complex 2 was determined using the Total ROS/Superoxide detection kit (Enzo-Life Sciences) according to the instructions. The analysis was performed via Flow cytometry. Generally, 1.0 × 10 6 of A2780 cells per well were seeded in the six-well plate and then pre-incubated in drug-free media at 310 K for 24 h (under 5% CO 2 humidified conditions), and then drugs were added to triplicates wells at equipotent IC 50 concentration. After 24 h of drug exposure, supernatants were removed by suction and cells were washed with PBS and harvested. Cell pellets were then re-suspended in PBS buffer containing the orange/green fluorescent reagents to achieve cell staining. Cells were analysed in a BD LSR II flow cytometer (488 nm laser) using FITC-A channel: 575/26 nm for the oxidative stress and PE-A channel: 530/30 nm for superoxide detection. Data were gated using positive-stained ( pyocyanin positive control), untreatedstained and untreated-unstained control samples, acquired as instrumental triplicates by using Flowjo V10 for Windows software. All samples were kept under dark conditions to avoid light-induced ROS production.

Author contributions
F. C., S. B., A. H., and P. J. S. designed the project. F. C. carried out the synthesis and characterisation of ligands and complexes, determined the pK * a values, UV-vis spectra and turnover frequencies. I. R.-C. and J.-I. S. carried out the antiproliferative cell studies and related biochemical assays. G. J. C. carried out the X-ray crystallography. L. S. the high resolution mass spectra work, and I. P. the NMR work. All authors contributed to the writing of the paper.

Conflicts of interest
There are no conflicts to declare.