← Back

Cytotoxicity, cellular localization and photophysical properties of Re(I) tricarbonyl complexes bound to cysteine and its derivatives.

PMID: 32583226
JBIC Journal of Biological Inorganic Chemistry https://doi.org/10.1007/s00775-020-01798-9 ORIGINAL PAPER Cytotoxicity, cellular localization and photophysical properties of Re(I) tricarbonyl complexes bound to cysteine and its derivatives Miles S. Capper1 · Alejandra Enriquez Garcia1 · Nicolas Macia1 · Barry Lai2 · Jian‑Bin Lin1 · Masaharu Nomura3 · Amir Alihosseinzadeh4 · Sathish Ponnurangam4 · Belinda Heyne1 · Carrie S. Shemanko5 · Farideh Jalilehvand1 Received: 29 April 2020 / Accepted: 8 June 2020 © Society for Biological Inorganic Chemistry (SBIC) 2020 Abstract The potential chemotherapeutic properties coupled to photochemical transitions make the family of fac-[Re(CO) (N,N) 3 X]0/+ (N,N = a bidentate diimine such as 2,2′-bipyridine (bpy); X = halide, H O, pyridine derivatives, PR , etc.) complexes 2 3 of special interest. We have investigated reactions of the aqua complex fac-[Re(CO) (bpy)(H O)](CF SO ) (1) with poten- 3 2 3 3 tial anticancer activity with the amino acid l-cysteine ( H 2 Cys), and its derivative N-acetyl-l-cysteine ( H 2 NAC), as well as the tripeptide glutathione (H A), under physiological conditions (pH 7.4, 37 °C), to model the interaction of 1 with thiol- 3 containing proteins and enzymes, and the impact of such coordination on its photophysical properties and cytotoxicity. We report the syntheses and characterization of fac-[Re(CO) (bpy)(HCys)]·0.5H O (2), Na(fac-[Re(CO) (bpy)(NAC)]) (3), and 3 2 3 Na(fac-[Re(CO) (bpy)(HA)])·H O (4) using extended X-ray absorption spectroscopy, IR and NMR spectroscopy, electrospray 3 2 ionization spectrometry, as well as the crystal structure of {fac-[Re(CO) (bpy)(HCys)]} ·9H O (2 + 1.75 H O). The emission 3 4 2 2 spectrum of 1 displays a variance in Stokes shift upon coordination of l-cysteine and N-acetyl-l-cysteine. Laser excitation at λ = 355 nm of methanol solutions of 1–3 was followed by measuring their ability to produce singlet oxygen (1O ) using direct 2 detection methods. The cytotoxicity of 1 and its cysteine-bound complex 2 was assessed using the MDA-MB-231 breast cancer cell line, showing that the replacement of the aqua ligand on 1 with l-cysteine significantly reduced the cytotoxicity of the Re(I) tricarbonyl complex. Probing the cellular localization of 1 and 2 using X-ray fluorescence microscopy revealed an accumulation of 1 in the nuclear and/or perinuclear region, whereas the accumulation of 2 was considerably reduced, potentially explaining its reduced cytotoxicity. Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s0077 5-020-01798 -9) contains supplementary material, which is available to authorized users. Extended author information available on the last page of the article Vol.:(0112 33456789) JBIC Journal of Biological Inorganic Chemistry Graphic abstract Replacing the aqua ligand with cysteine in the antitumor active fac-[Re(CO) (bpy)(H O)](CF SO ) complex significantly 3 2 3 3 reduced its cellular accumulation and cytotoxicity against the MDA-MB-213 breast cancer cell line, shifted its maximum emission to considerably higher energies, and decreased its fluorescence quantum yield. Keywords Re(I) tricarbonyl complexes · Cysteine and thiol-containing biomolecules · Photochemistry · Cytotoxicity · X-ray fluorescence miscorscopy Abbreviations the unwanted toxic side effects and prevalence of platinum- ESI–MS Electrospray ionization mass spectrometry resistance mechanisms [1]. Treatment of cancer using pho- EXAFS Extended X-ray absorption fine structure todynamic therapy (PDT) and photo-activated chemother- spectroscopy apy (PACT) has been a rapidly developing area of research XFM X-ray fluorescence microscopy due to its multimodal approach [2–4]. In PDT, a nontoxic PDT P hotodynamic therapy photosensitizer (PS) is used with light (usually visible/near PACT P hoto-activated chemotherapy infrared) and oxygen to generate cytotoxic products inducing MT M etallothionein localized cell death. PDT initiates apoptosis through differ- GC-TCD Gas chromatography with thermal conductiv- ent mechanisms which involve the generation of oxidative ity detection stress, damaging mitochondria and vascular system, leading HOMO Highest occupied molecular orbital to tumour inflammation, and eliciting an immune response LUMO Lowest unoccupied molecular orbital against the tumour cells [3, 5, 6]. TD-DFT Time-dependent density functional theory PACT offers an oxygen-independent alternative to PDT MLLCT Metal–ligand to ligand charge transfer that is particularly useful when considering the characteristic ILCT Intra-ligand charge transfer hypoxic environment in cancer cells [7]. Photoactivation can bpy 2 ,2′-Bipyridine induce ligand substitution reactions in PACT agents, leading phen 1 ,10-Phenanthroline to the release of a small cytotoxic molecule such as CO or dmphen 2 ,9-Dimethyl-1,10-phenantroline NO [8–11]. The antiproliferative effect of CO is thought to DMEM Dulbecco’s Modified Eagles Medium be caused by its influence on mitochondria activity in cancer PBS P hosphate-buffered saline cells, through binding to cytochrome c oxidase and inhibit- IC Half-maximal inhibitory concentration ing its activity, generating reactive oxygen species and radi- 50 cals, and accelerating cellular bioenergetics by high oxygen consumption, thus leading to metabolical exhaustion of the Introduction cancer cell [12, 13]. A combined therapy approach can help in circumventing unwanted side effects, such as toxicity and Despite the great success of platinum-based chemothera- resistance effects associated with some of the currently used peutics, drugs with a different mode of action other than the chemotherapeutics such as cisplatin [14]. platinum antineoplastic family are highly desirable due to 1 3 JBIC Journal of Biological Inorganic Chemistry Complexes of transition metals with d3 and d6 (low spin) investigations on the reaction products of fac-[Re(CO) (bpy) 3 electronic configurations are great candidates as potential (H 2 O)](CF 3 SO 3 ) (1) with l-cysteine ( H 2 Cys), N-acetyl-l- PDT and PACT agents due to their relatively high kinetic cysteine ( H NAC) and glutathione (GSH, denoted H A in 2 3 inertness and photophysical properties, particularly for d6 its triprotonated form), and discuss the impact of such coor- metal complexes that can have a diverse range of possible dination on the cellular localization, cytotoxicity and photo- excited states and light-induced transitions that enhance physical properties of the Re(I) complex formed. their photochemical properties [8, 15–18]. For exam- ple, Re(I) organometallic compounds and in particular, diiminerhenium(I) tricarbonyl complexes fac-[Re(CO) (N,N) Experimental section 3 X]0/+ have been studied as potential anticancer therapeutics; Materials for many their cellular distribution could be probed due to their intrinsic luminescence properties, and for some, their antiproliferative activity exceeded that of cisplatin [19]. Ford All chemicals and solvents were of reagent grade or higher. et al. demonstrated that irradiation of fac-[Re(CO) (bpy) Rhenium pentacarbonyl chloride (98%), 2, 2′-bipyridine, 3 (P(CH OH) )](CF SO ) (bpy = 2,2′-bipyridine) in the near- tris(2,2′-bipyridine)ruthenium(II) chloride hexahydrate, 2 3 3 3 UV region (λ = 405 nm) initiated the photo-induced release l-cysteine, N-acetyl-l-cysteine and glutathione were pur- of axially coordinated CO, monitoring the intracellular chased from Sigma-Aldrich and used without further puri- location of the complex using confocal microscopy [20]. fication. Degassed water was prepared by bubbling argon Kastl et al. introduced a novel class of photosensitizers in through boiled distilled water until cooled down to room the rhenium(I) tricarbonyl complexes that was based on the temperature. The oxygen-free aqueous NaOH solution metallo-pyridocarbazole scaffold, which displayed visible (1.0 M) was prepared by dissolving solid NaOH in degassed light-induced (λ ≥ 505 nm) antiproliferative activity [21]. water. Dry toluene was prepared by refluxing toluene with Leonidova et al. identified two N,N-bis(quinolinoyl) Re(I) sodium metal and benzophenone under argon [29]. The pH tricarbonyl complex derivatives as excellent 1O generators of solutions was measured with a calibrated Thermo Scien- 2 in a lipophilic environment with quantum yields as high tific Orion Star pH meter using standard buffers. as ~ 75%, showing significant enhancement in their cyto- toxicity upon UV-A irradiation (λ = 350 nm) [22]. Cell culture The rhenium(I) tricarbonyl entity can bind to guanine, forming fac-[Re(CO) (guanine) L]0/+ complexes (L = H O, MDA-MB-231 breast cancer cells were acquired from the 3 2 2 Br; guanine = 7-methyl- and 9-methylguanine) with two American Type Culture Collection (ATCC) and the pas- different conformations: head-to-head (HH) and head- sage used within 6 months. Dulbecco’s Modified Eagles to-tail (HT) around the Re centre that are in equilibrium Medium (DMEM) (Invitrogen/Gibco) was supplemented in solution. This study suggets that similar to cisplatin, with heat-inactivated foetal bovine serum (10% v/v; Sigma [Re(CO) 3 (H 2 O) 3 ]+ may target DNA [23]. Cellular tar- Life Sciences), l-glutamine (2 mM, Invitrogen), penicillin gets other than DNA can include metallothioneins (MT), (100 units/mL) and streptomycin (100 μg/mL). Cells were which are cysteine-rich proteins with high affinity for zinc incubated at 37 °C in a 5% CO -humidified incubator and 2 and heavy metals such as Re(I) and Tc(I) [24]. Interaction were sub-cultured every 3–4 days. between Re(CO) + complexes and MT could be important 3 Syntheses when treating human breast tumors, as they accumulate unu- sually high level of zinc that can lead to apoptosis, and over- expression of MT may occur as a method of protection of fac-[Re(CO) 3 (bpy)(H 2 O)](CF 3 SO 3 ) abbreviated as Re-Aq malignant breast cancer cells from zinc hyper-accumulation (1) The complex was prepared following a previously [25, 26]. described method [30], through the synthesis of fac- Both in vivo and in vitro anticancer activities of a [Re(CO) (bpy)Cl]; see Electronic Supplementary Materials. 3 series of diimine rhenium(I) tricarbonyl aqua complexes, A solution mixture of fac-[Re(CO) (bpy)Cl] (2.69 mmol) 3 fac-[Re(CO) (N,N)(H O)]+, have recently been reported and AgCF SO (2.69 mmol) in 100 mL acetone was refluxed 3 2 3 3 [27, 28]. The fac-[Re(CO) (dmphen)(H O)]+ complex at 70 °C for 2 h under argon. The white AgCl precipitate was 3 2 (dmphen = 2,9-dimethyl-1,10-phenantroline) showed high removed by filtering, and the filtrate was concentrated to affinity for binding to 9-ethylguanine, N-acetyl-cysteine dryness using a rotary evaporator. The resulting crude yel- and N-acetyl-histidine, but not for other amino acids such low solid was recrystallized using an acetone/H O (1:3 v/v) 2 as methionine, glycine and serine. Cysteine residues of pro- mixture and left in a desiccator under vacuum for 4 days. The teins and enzymes could be a potential target and binding yellow crystals formed were filtered and washed with diethyl site for these complexes. Here, we report the results of our ether. The unit cell parameters of these crystals were the 1 3 JBIC Journal of Biological Inorganic Chemistry same as previously reported [31]. Elemental anal. calcd for in water and displayed poor stability in a range of solvents, (fac-[Re(CO) (bpy)(H O)](CF SO ) ( ReC H N O SF ): undergoing rapid decomposition. Elemental anal. calcd for 3 2 3 3 14 10 2 7 3 %C 28.33, %H 1.70, %N 4.72. Found: %C 28.43, %H Na(fac-[Re(CO) (bpy)(NAC)]) ( ReC H N O SNa): %C 3 18 15 3 6 1.69, %N 4.68; yield 94%. 1H NMR (600 MHz, D O) δ 35.41, %H 2.48, %N 6.88. Found: %C 35.79, %H 2.54, %N 2 H 9.14 (d, J = 4.8 Hz, 2H), 8.52 (d, J = 8.2 Hz, 2H), 8.31 (td, 6.77; yield 54%. IR: ṽ = 2003, 1907, 1869 cm−1. For 1H co J = 7.9, 1.5 Hz, 2H), 7.73 (t, J = 7.0 Hz, 2H). 13C {1H} NMR NMR measurement, a small portion of the filtrate (when (600 MHz, D O) δ 196.9 (CO ), 192.8 (CO ), 156.4 (C1, separating the crude solid and solution) was passed through 2 C eq ax C1′), 154.1 (C3, C3′), 141.4 (C5, C5′), 128.2 (C4, C4′), a Sephadex G-15 size exclusion column chromatography to 124.3 (C6, C6′). IR ṽ = 2033, 1914 cm−1. separate unreacted 1 and excess N-acetylcysteine. The eluted co fac-[Re(CO) 3 (bpy)(HCys)]·0.5H 2 O abbreviated as Re- orange solution was evaporated to dryness and then dis- HCys (2) Solid l-cysteine (2.53 mmol) was added to a solved in D 2 O and further analyzed by 1H NMR. (400 MHz, completely dissolved, yellow solution of 1 (0.253 mmol) D O) δ 9.11 (d, J = 5.4 Hz, 2H), 8.52 (d, J = 8.3 Hz, 2H), 2 H in 20 mL degassed water under argon atmosphere to pre- 8.25 (t, J = 8.1 Hz, 2H), 7.70 (t, J = 8.0 Hz, 2H), 3.98 (dd, vent oxidation of cysteine thiol group. The pH of the solu- J = 9.3, 3.8 Hz, 1H), 2.80 (dd, J = 12.8, 4.0 Hz, 1H), 2.62 tion (4.55) was adjusted to 7.40 by dropwise addition of (dd, J = 13.2, 10.0 Hz, 1 H), 1.97 (s, 3H). 1.0 M degassed NaOH, then allowing the orange solution to Na(fac-[Re(CO) 3 (bpy)(HA)])·H 2 O abbreviated as Re- stir for 48 h at 37 °C. The resulting precipitate was filtered GSH (4) Solid glutathione (H A; 1.68 mmol) was added to 3 and washed with water. The crude solid, which contained a completely dissolved solution of 1 (0.17 mmol) in 18 mL residual sodium triflate by-product was further purified by degassed water under argon (pH = 2.59). The pH of the dissolving it in methanol and filtering the solution. The fil- solution was raised to 7.40, using 1.0 M degassed NaOH. trate was then concentrated to dryness using a rotary evapo- The orange solution was left stirring for 48 h at 37 °C, rator; the orange solid was dried under vacuum at 65 °C and then evaporated to dryness using a rotary evaporator. with calcium chloride as the drying agent. The electrospray Methanol was added to the remaining solid, which dis- ionization (ESI) mass spectrum of the final product in (−) solved the Re-GSH complex but only partially the excess ion mode did not show a peak at − m/z = 148.95 amu for the glutathione. The remaining solid glutathione was removed C F SO − ion in the final product. Elemental anal. calcd for by filtering, followed by evaporation of the filtrate with 3 3 fac-[Re(CO) (bpy)(HCys)]·0.5H O (ReC H N O S): %C rotary evaporator. This cycle of adding MeOH, filtering 3 2 16 15 3 5.5 34.59, %H 2.72, %N 7.56 (1.62% H O). Found: %C 34.25, and evaporating was repeated six times. The remaining 2 %H 2.42, %N 7.42 (TG: 1.57% H O); yield 58%. Complex solid was then dissolved in 1–2 mL water. Size exclusion 2 2 is insoluble in water but soluble in methanol; it was ini- column chromatography (Sephadex G-15) was used to fur- tially stable in CD OD although some decomposition was ther separate excess glutathione from the Re(I)-glutathione 3 observed over a period of 7 days. 1H NMR (600 MHz, product. The orange band on the column was separated, and C D OD) δ 9.07 (dddd, J = 11.2, 5.5, 1.6, 0.8 Hz, 2H), the resulting solution was again concentrated and passed 3 H 8.58 (dt, J = 8.2, 1.1 Hz, 2H), 8.22 (td, J = 7.9, 1.6 Hz, 2H), through the column. This process was repeated three times 7.66 (dddd, J = 7.8, 5.5, 2.5, 1.2 Hz, 2H), 3.27 (dd, J = 10.7, to ensure separation of unreacted glutathione. The orange 3.6 Hz, 1H), 2.99 (dd, J = 13.3, 3.7 Hz, 1H), 2.61 (dd, solution was rotary evaporated to dryness giving a dark J = 13.3, 10.6 Hz, 1H). 13C {1H} NMR (600 MHz, CD OD) red/orange solid. Despite all the efforts for purifying the 3 δ 198.8 ( CO ), 190.1 ( CO ), 172.3 (C12), 156.0 (C1, C1′), sample, the ESI-mass spectrum of the final product still C eq ax 153.3 (C3, C3′), 139.6 (C5, C5′), 127.6 (C4, C4′), 124.2 showed a small peak for CF SO − (− m/z = 148.95 amu; 3 3 (C6, C6′), 59.0 (C11), 31.4 (C10). IR: ṽ = 2005, 1907, Figure S2), leading to deviations observed in the elemen- co 1877 cm−1. tal analysis: calcd for Na(fac-[Re(CO) (bpy)(HA)])·H O 3 2 Na(fac-[Re(CO) 3 (bpy)(NAC)]) abbreviated as Re-NAC (ReC 23 H 25 N 5 O 10 SNa): %C 35.75, %H 3.26, %N 9.06 (2.33% (3) Following the above procedure, N-acetyl-l-cysteine H 2 O). Found: %C 32.54, %H 3.60, %N 8.12 (TG: 3.0% (1.68 mmol) was mixed with 1 (0.17 mmol) in 20 mL H O). 1H NMR (600 MHz, D O) δ 9.07 (dddd, J = 9.9, 2 2 H degassed water under argon, adjusting the solution pH (2.11) 5.5, 1.6, 0.8 Hz, 2H), 8.48 (dq, J = 8.5, 1.1 Hz, 2H), 8.21 (tt, to 7.40. After stirring the mixture for 48 h at 37 °C, the J = 8.1, 1.3 Hz, 2H), 7.65 (dddd, J = 7.7, 5.4, 4.1, 1.2 Hz, precipitate was removed by filtering, and a small portion of 2H), 4.09 (dd, J = 8.8, 5.0 Hz, 1H; C H), 3.72–3.68 (m, 11 the filtrate was set aside for the 1H NMR measurement (see 2H; C H ,H ), 3.63 (d, J = 17.3 Hz, 1H; C H), 2.77 (dd, 13 A B 18 below). The crude solid was dissolved in methanol and then J = 13.2, 5.0 Hz, 1H; C H ), 2.60 (dd, J = 13.8, 8.8 Hz, 1H; 10 B filtered; the filtrate was evaporated to dryness. To remove C H ), 2.41 (t, J = 7.8 Hz, 2H; C H ,H ), 2.10–2.03 (m, 10 A 16 A B traces of Na(CF SO ), the solid was suspended in water and 2H; C H ,H ).13C NMR (151 MHz, D O) δ 198.7 ( CO ), 3 3 17 A B 2 C eq sonicated for a few minutes. The dark-orange solid product 190.5 ( CO ), 176.1 (C ), 174.6 (C ), 173.9 (C ), 172.2 ax 14 15 19 (3) was filtered and dried under vacuum. It was insoluble ( C ), 155.3 (C1, C1′), 153.0 (C3, C3′), 139.7 (C5, C5′), 12 1 3 JBIC Journal of Biological Inorganic Chemistry 127.4 (C4, C4′), 123.9 (C6, C6′), 57.4 ( C ), 54.2 ( C ), 43.3 [Re(CO) (bpy)(H O)](CF SO ) (1) and {fac-[Re(CO) (bpy) 11 18 3 2 3 3 3 ( C ), 31.5 (C ), 30.4 (C ), 26.3 (C ). (HCys)]} ·9H O (2 + 1.75 H O), using the FEFF 7.0 pro- 13 16 10 17 4 2 2 gram to obtain ab initio calculated amplitude f (k), phase eff i Physical measurements and methods shift ϕij(k), and mean free path λ(k) functions in Eq. (1) [37, 38]. Single-crystal X-ray diffraction Single crystals of {fac- N ⋅S2(k) [ u R si e n (C g O th ) 3 e ( b s p a y m )( e H r C e y a s c ) t ] i } o 4 n ·9 c H o 2 n O d i ( t 2 i o + n 1 s . 7 a 5 s H d 2 e O cr ) i w be e d re a g b r o ow ve n . 𝜒(k)=∑ i i k⋅R 0 2 i | | f eff (k) | | i exp ( −2k2𝜎 i 2 ) Instead of being continuously stirred at 37 °C, the flask was exp −2R∕Λ(k) sin 2kR +𝜙 (k) [ ] [ i ij ] (1) sealed under Ar and left in the fridge (4 °C). After 5 days, orange crystals had formed. Single-crystal X-ray diffraction Least-squares curve fittings of theoretically simulated data were collected at 173 K on a Bruker APEX-II CCD dif- EXAFS oscillations χ(k) to the k3-weighted, unfiltered fractometer using graphite-monochromated Cu K α radiation experimental EXAFS spectra were performed over k-range (λ = 1.5406 Å). The single crystal was mounted on a nylon 3.5–15.3 Å−1 (12.8 Å−1 for 1), by keeping the coordination cryo loops with paratone oil. The data integration and reduc- number (N) constant, and allowing the bond distance (R) i tion were processed with the INTEGRATE program of the and the Debye–Waller parameter (σ2) to be refined for each APEX3 software [32]. Multi-Scan absorption correction was backscattering path. The amplitude reduction factor (S 2) 0 applied to the collected reflections. Using Olex2 [33], the and ΔE (a common value for all paths) were allowed to 0 structure was solved with the ShelXT [34] structure solution float. The accuracy of the bond distances is within ± 0.02 Å program using intrinsic phasing and refined with the ShelXL for Re-S, and ± 0.04 Å for Re-(C/O/N). Estimated error lim- [35] refinement package using least squares minimization. its of the Debye–Waller parameters for the Re-C and Re-S All non-hydrogen atoms were refined anisotropically. The paths are ± 0.001 Å2; for the Re-(N/O) single scattering and water hydrogen atoms were located in difference Fourier Re-C-O (n = 3) multiple scattering paths ± 0.002 Å2. leg maps and refined by using the HTAB command. The organic Electronic spectroscopy UV–Vis spectra were recorded hydrogen atoms were generated geometrically. The refine- using a Cary 50 spectrophotometer. Samples were measured ment parameters are summarized in Table S1. in a quartz cuvette (or in a Schlenk cuvette [39]) with 1 cm X-ray absorption spectroscopy (XAS) The Re L 3 -edge path length, using water or methanol as a reference. The X-ray absorption spectra were recorded in transmission molar extinction coefficients (ε) of compounds 1–3 were mode at bending magnetic beamline 9C of the Photon Fac- evaluated from absorbance data using Beer–Lambert law. tory of the High Energy Accelerator Research Organization Fluorescence emission spectroscopy Fluorescence emis- (Tsukuba, Japan) under dedicated conditions (2.5 GeV with sion spectra were carried out using a Photon Technology critical energy of 4 keV, 430–450 mA). Monochromatiza- International QuantaMaster 400 steady-state spectrofluor- tion was achieved by means of a Si(111) double-crystal imeter (Horiba Scientific). Fluorescence quantum yields of monochromator, detuning the beam intensity to 50% at I 0 . 1–3 were determined via a relative method, as described The first ion chamber (I 0 ) was filled with nitrogen and the elsewhere [40], using [Ru(bpy) 3 ]Cl 2 in air-equilibriated second one (I 1 ) with (50% N 2 + 50% Ar). The energy scale MeOH (ϕ = 0.31) [41], or quinine sulfate in 0.1 N H 2 SO 4 was calibrated by placing a platinum foil in front of the (ϕ = 0.55) [42] as standards. The absorbance at the excitation third ion chamber and assigning its first inflection point to wavelength 290 nm for quinine sulfate and 2, and at 360 nm 11,564.0 eV. The solid samples 1–4 were finely ground and for [Ru(bpy) ]Cl , 1 and 3 was kept below 0.05. 3 2 mixed with boron nitride, then mounted in a 1 mm Al frame, 1O 2 direct detection Phosphorescence kinetic traces of using Mylar tape as window material. Three scans were col- 1O at 1270 nm were acquired with a time-resolved near- 2 lected for each sample at room temperature. infrared (TR-NIR) detection system in air-equilibrated meth- The extended X-ray absorption fine structure (EXAFS) anol. A diode-pumped Nd:YAG laser (FTSS355-Q3, Cry- oscillations were extracted using the WinXAS 3.1 pro- Las) tuned to 355 nm at 1 kHz repetition rate was used as the gram [36], by subtracting the pre-edge background with excitation source for all the complexes. A 1150 nm cut-on a first-order polynomial, followed by edge step normali- long pass filter (FEL1150, ThorLabs) and a 1064 nm notch zation. The energy unit was converted to k (Å−1), where filter (NF1064-44, ThorLabs) were mounted side by side at k = [(8π2m e /h2)(E − E 0 )]1/2, using the threshold energy of the entry port of a monochromator (Digikröm CM110 1/8 m, E 0 = 10,535.7–10,536.0 eV for 1–4. A seven-segment cubic Spectral Products). A TE-cooled PMT (model H10330A-45, spline was then subtracted in the post-edge region to obtain Hamamatsu) working at − 908 V was used as the detector the EXAFS oscillation. at the exit port of the monochromator. The PMT output was The EXAFS model functions, χ(k), were constructed amplified to a voltage pulse using a 1.1 GHz preamplifier with structural information from the crystal structures of 1 3 JBIC Journal of Biological Inorganic Chemistry module (PAM-102-T, PicoQuant GmbH) connected to a (PBS) before being treated with 10% v/v alamarBlue rea- multichannel scaler (TimeHarp 260-Nano, PicoQuant). The gent and left at 37 °C in a 5% CO -humidified incubator 2 signal monitored at 1270 nm was collected for 200 s with for 2.5 h. After incubation, fluorescence was measured on 256 ns resolution for each time-resolved 1O emission curves a SpectraMax M2e Multi-detection reader using excitation 2 recorded. The time-resolved curves were Chi-squared fitted and emission wavelengths of 565–575 nm and 585–595 nm, to Eq. (2) using Prism 7.0 (GraphPad Software Inc.) with the respectively. Using Prism GraphPad Prism Version 5.03 lifetime of 1O (τ ), the lifetime of the PS (τ ) and the signal software, cell viability was determined as a percentage of the 2 Δ T strength (S ) as free parameters. The goodness of the fit- fluorescence of cells when treated with 1 or 2 with respect 0 tings was assessed by the residual plots. 1O quantum yields to the control (DMEM) or vehicle control (1.5% EtOH in 2 were calculated using tris(2,2-bipyridine)ruthenium(II), DMEM in the case of 2). By curve-fitting plots of cell viabil- [Ru(bpy) ]2+ (ϕ = 0.86 in air-equilibriated MeOH) as the ity (%) vs. log of drug concentration, half-maximal inhibi- 3 Δ reference [43]. tory concentration (IC ) values were calculated. All data 50 were gathered from three independent experiments using 𝜏 −t −t S(t)=S Δ (exp( )−exp( )) nine biological replicates and their standard deviations were 0 𝜏 −𝜏 𝜏 𝜏 (2) ( Δ T) Δ T calculated. X-ray fluorescence microscopy (XFM): sample prepa- CO release detection Possible release of CO gas upon ration Stock solutions (800 μM) of 1 and 2 were pre- irradiating solutions of 1–3 with light (λ = 365 nm) was pared in DMEM (solution 2 contained 1.5% EtOH for monitored using UV–Vis spectroscopy, and gas chro- solubility purposes), and were diluted to 20 μM with matography with thermal conductivity detection (GC- DMEM. As described elsewhere [44], cells were grown TCD) using a solution of fac-[Re(CO) (bpy)(P(CH OH) ] 3 2 3 on 1.5 mm × 1.5 mm × 500 nm Si N windows, by seeding 3 4 (CF SO ) in degassed PBS (pH 7.4) as a standard [30]. 3 3 225,000 cells per well in six-well plates which were then Solutions of 1, 2 and 3 (50 μM, 5 mL) in degassed MeOH allowed to attach at 37 °C in a 5% CO -humidified incuba- 2 were injected in a Schlenk cuvette [39] (internal volume of tor for 24 h. Using 20 μM solutions of 1 or 2 in DMEM, or 23.6 mL) under argon. Each sample was then exposed to only DMEM as control, cells were treated for a period of 6 h. λ = 365 nm light (7.105 mW/cm−2) for 2 hours, measuring Upon removal of the medium, cells were washed with PBS its UV–Vis spectrum at 20 min intervals. This process was and then fixed with a 4% paraformaldehyde (prepared fresh repeated with the GC-TCD method, but after each 20 min in Dulbecco’s PBS (D-PBS)) solution for 1 h in an incubator interval, a gastight syringe was used to remove a 100 μL at 37 °C and 5% C O , and then washed with a 100 mM solu- 2 sample from the headspace of the cell (in Schlenk cuvette) tion of ammonium acetate (in ultrapure water) two times, through a special screw-top fitted with a three-layer lami- and air-dried overnight while covered [45, 46]. nated silicone GC septum. This sample was analyzed by a XFM data collection and data analysis Details of data GC-TCD using an Agilent 6890 N gas chromatograph fitted collection and data analysis for XFM have been described with a Carboxen 1010 PLOT fused silica capillary column elsewhere [44]. XFM images of the cells were obtained (L × I.D. = 30 m × 0.53 mm, average thickness 30 μm). The at undulator beamline 2-ID-D of the Advanced Photon GC inlet temperature was 230 °C. The oven was held at Source (APS), Illinois, USA, using an incident energy of 35 °C for 14 min, then ramped at 20 °C/min to 245 °C, and 12.8 keV using a double-crystal monochromator. Using finally held at 245 °C for 10 min. The retention time for CO two Fresnel zone plates, the beam size was focused to is about 11.2 min, and that for CO is 21 min [30]. Oxygen 2 0.35 μm; the images were obtained with a spatial resolution and nitrogen from ambient air have retention times of about of 0.5 × 0.5 μm in step-scan mode and 1 s dwell time. Using 8.2 and 8.4 min, respectively [30]. The ChemStation auto- a single element silicon drift energy dispersive detector integration function was used to calculate the peak area for (Vortex EX, SII Nano-technology, Northridge, CA), X-ray CO. fluorescent spectra were collected from samples under a Cell viability The cytotoxicity of complexes 1 and 2, as He atmosphere set at 75º to the incident beam. Four to five well as cisplatin, were measured using the alamarBlue Cell individual cells per sample were selected using an optical Viability Reagent Assay (Invitrogen). The MDA-MB-231 microscope. breast cancer cell line was seeded at 7500 cells per well By fitting the raw fluorescence emission spectra to Gauss- in black-walled 96-well plates for 48 h at 37 °C in a 5% ians at each spatial point, elemental maps (in units of μg C O 2 -humidified incubator. Fresh stock solutions of 1 and c m−2), as well as regions of interest (ROIs), were gener- cisplatin were prepared in DMEM, whereas for 2, 1.5% ated [47, 48]. To obtain Re images, its L β emission line EtOH in DMEM was used as the vehicle. Drug exposure (10,010.0 eV) was used for fitting, as its L α line (8586.2 eV) times were 24 and 48 h. Medium was removed using suction overlaps with the Zn K α (8638.9 eV) fluorescence line (Fig- and each well was washed with phosphate buffered saline ure S12). Comparison of X-ray fluorescence intensity to 1 3 JBIC Journal of Biological Inorganic Chemistry those of the thin-film standards, NBS-1832 and NBS-1833, from the National Bureau of Standards (Gaithersburg, MD) allowed for quantification. All data analysis was performed using MAPS software [49, 50]. Results and discussions Characterization of Re(I)‑thiolate complexes Using yellow crystals of fac-[Re(CO) (bpy)(H O)](CF SO ) 3 2 3 3 (1), we successfully prepared three orange/dark-orange Re(I) tricarbonyl thiolate complexes with bound l-cysteinate, N-acetyl-l-cysteinate or glutathione ligands under physio- logical conditions (pH = 7.4 and 37 °C): fac-[Re(CO) (bpy) Fig. 1 Crystal structure of {fac-[Re(CO)(bpy)(HCys)]}·9HO 3 3 4 2 (HCys)]·0.5H O (2) and Na(fac-[Re(CO) (bpy)(NAC)]) (2 + 1.75 H 2 O). Water molecules have been omitted for clarity 2 3 (3), both of which have poor water solubility, and Na(fac- [Re(CO) (bpy)(HA)])·H O (4). These complexes were char- state of the carboxylate (–COO−) and the amine (as –NH +) 3 2 3 acterized using 1H-, 13C-NMR (for 2 and 4), IR and X-ray groups in the crystal structure of {fac-[Re(CO) (bpy) 3 absorption spectroscopy, as well as electrospray ionization (HCys)]} ·9H O was determined based on the neutrality of 4 2 mass spectrometry (ESI–MS). Scheme 1 provides an over- the complex, its elemental analysis, and the fact that there view for the syntheses of these Re(I) thiolate compounds. was no counter ion (e.g., N a+ or C F SO −) present in the unit 3 3 Single crystals of {fac-[Re(CO) (bpy)(HCys)]} ·9H O cell. Since the complex was crystallized at pH 7.4, formation 3 4 2 (2 + 1.75 H O) were obtained by mixing the reactants in of ( H O)[Re(CO) (bpy)(Cys)] or [Re(CO) (bpy)(H Cys)] 2 3 3 3 2 water, adjusting the pH to 7.4 and then leaving the reac- (OH) is not feasible. tion mixture at 4 °C (see the “Experimental Section”). After A survey of the Cambridge Structural Database (CSD) 5 days, small orange crystals had formed which were suit- [51] shows that the Re–S bond distances of 2.503(6) and able for single-crystal X-ray diffraction (see Fig. 1). Crystal- 2.496 (10) Å in the two fac-[Re(CO) (bpy)(HCys)] entities 3 lographic data, and selected interatomic distances and bond are similar to the average Re-S distance in Re(CO) (N,N)+ 3 angles are reported in Tables S1 and S2. The protonation complexes with S-donor ligands: 2.509 ± 0.023 Å Scheme 1 Synthesis of fac-[Re(CO) 3 (bpy)(H 2 O)](CF 3 SO 3 ) (1) and in red denote the carbon atoms associated with the 13C NMR signals its subsequent reaction products (2–4) with thiol-containing biomol- described in the “Experimental Section” ecules l-cysteine, N-acetyl-l-cysteine, and glutathione. The numbers 1 3 JBIC Journal of Biological Inorganic Chemistry (Table S3). However, the mean Re-C bond length analysis (Re-C 1.93 Å, Re–N 2.18 Å, Re–S 2.48 Å; Table 1), (axial) (1.925 ± 0.005 Å; Table S3) in these complexes is slightly with those from its crystal structure (1.90 Å, 2.14 Å, 2.50 Å, longer than the corresponding distances [1.879(24) Å and respectively) shows that the accuracy of these distances is 1.909 (20) Å] in the Re(I) cysteinate complex 2, and also within ± 0.02 Å for Re-S and ± 0.04 Å for Re–(C/N) bonds. longer than the reported Re-C distance in the Re-Aq The Debye–Waller parameter (σ2) values for the Re–C (axial) complex (1), 1.882(10) Å [31]. and Re–(N/O) scattering paths in the EXAFS fitting model The k3-weighted Re L -edge X-ray absorption fine struc- of 1 seem reasonable (0.0021 ± 0.001 Å2 and 0.0055 ± 0.002 3 ture (EXAFS) spectra for complexes 1–4, and their corre- Å2, respectively; Table 1), while that of the longer Re–C–O sponding Fourier transforms (FTs) are shown in Fig. 2. The (n = 3) path is relatively small (0.0022 ± 0.002 Å2). In the leg FTs of the Re(I)–thiolate complexes (2–4) display a small EXAFS model fitting of Re(CO) bound to a TiO surface, 3 2 peak at ~ 2.2 Å (not corrected for phase shift) associated with a similar σ2 value was reported for the corresponding mul- the Re–S scattering that is absent in the FT of the Re-aqua tiple scattering path [53]. By replacing H O with a cystein- 2 complex 1, fac-[Re(CO) (bpy)(H O)](CF SO ). This dif- ate ligand in 2, and introducing a Re–S path in its EXAFS 3 2 3 3 ference shows that contradictory to an earlier report [52], model fitting, an unreasonably small σ2 value was obtained Re L -edge EXAFS spectroscopy would be sensitive to the for the Re–N path (Table 1), which is probably due to the 3 nature of the axial ligand X in fac-[Re(CO) (N,N)X]0/+ com- interference between their EXAFS oscillations (see Figure 3 plexes, if high-quality data (k > 11 Å−1) are obtained. S1). Comparing the average bond distances in the first coor- The Re–GSH complex (4) shows a very similar first dination sphere of Re(I) in 2 obtained from EXAFS data coordination sphere as the Re–HCys complex (2) with Fig. 2 k3-weighted model fitted Re L-edge EXAFS spectra of 3 Re(I) complexes 1–4, and their corresponding Fourier trans- forms (see Table 1) Table 1 Least squares curve- Re-C Re-(N/O) Re-C-O (n = 3) Re-S S2 ΔE R fitting results for Re L -edge leg 0 0 3 EXAFS spectra of solids 1–4 N R σ2 N R σ2 N R σ2 N R σ2 (see Fig. 2) 1 3 f 1.88 0.0021 3 f 2.12 0.0055 6 f 3.05 0.0022 1.08 1.7 32.9 2a 3 f 1.93 0.0037 2 f 2.18 0.0005 6 f 3.08 0.0023 1 f 2.48 0.0029 0.94 8.6 21.8 3 3 f 1.91 0.0029 2 f 2.16 0.0027 6 f 3.08 0.0025 1 f 2.47 0.0078 0.95 7.1 28.7 4 3 f 1.92 0.0027 2 f 2.17 0.0010 6 f 3.08 0.0023 1 f 2.48 0.0037 0.95 7.4 24.7 The accuracy of the average distances is within ± 0.02 Å for Re–S and ± 0.04 Å for Re–(C/N/O) f fixed, R fitting residual a See the contribution of each scattering path in the model fit in Figure S1 1 3 JBIC Journal of Biological Inorganic Chemistry comparable average bond distances obtained from the similar frequencies as those of the Re–HCys complex (2): EXAFS data analysis; see Table 1. The ESI-mass spec- 2003 cm−1 (A′(1)), 1907 (A″), and 1869 cm−1 ((A′(2)) (see tra of this complex, in both (−) and ( +) ion modes, dis- Figure S5). EXAFS data analysis of this complex, however, played intense peaks at − m/z = 732.08 and + m/z = 756.07 showed that the mean Re–S distance has relatively high amu {[Re(CO) (bpy)(HA)]− + Na+ + H+}+, indicating that variation (σ2 = 0.0078 ± 0.0010 Å2; Table 1) compared with 3 [Re(CO) (bpy)(HA)]− is the dominating species in this sam- complexes 2 and 4 (σ2 = 0.003–0.004 Å2). A larger σ2 value 3 ple (see Figure S2 and Table S4). for the Re–S bond in 3 reflects a larger structural disorder, The 13C NMR spectra of both Re–HCys (2) and Re–GSH or vibrational disorder along the Re–S bond, relative to 2 (4) complexes displayed a signal for the Cys C β atom ( C 10 ) at and 4, even though all three compounds have similar mean δ = 31.4 (in CD OD) and 30.4 ppm (in D O), respectively, Re–S distances (2.47–2.48 Å; Table 1). For similar type c 3 2 being somewhat deshielded relative to those of the pure of bonds (e.g., Re–S), there is a correlation between the ligands in D O (pD = 7.8 [54], reading pH = 7.4): δ = 26.6 bond length and vibrational frequency, force constant and, 2 c and 26.0, respectively (see Figures S3 (a) and S4 (a)). This therefore, bond strength: The longer the bond, the lower the downfield shift of Δ(δ ) ~ 4–5 ppm is associated with the vibrational frequency; the force constant would be smaller C coordination of their cysteine thiolate groups to the Re(I) and the bond would be weaker. Thus, Re–S bonds with a ions. Similar shifts have been previously observed for Zn(II), higher vibrational/ structural disorder may indicate a weaker Hg(II), Pb(II), Cd(II) ions bound to cysteine and glutathione Re–S bond in the Re–NAC complex (3) than in 2 and 4. [55]. The 1H NMR signals, however, of the two hydrogen It would also explain the rapid decomposition (i.e. ligand atoms ( H , H ) on the Cys C atoms (C ) in pure cysteine (NAC2−)–solvent exchange) of 3 in CD OD (Figure S6), A B β 10 3 and glutathione (δ = 2.9–3.0 ppm; pD = 7.8 in D O) showed and other organic solvents (CD CN, C DCl , C D COCD ), H 2 3 3 3 3 an upfield shift of Δ(δ ) ~ 0.3–0.4 ppm when bound to Re(I) while the Re–HCys complex (2) was stable in d4-methanol. H ions in complexes 2 and 4 [see Figures S3 (a) and S4 (a)]. Therefore, the small structural difference between cysteine This shift could be explained by the shielding effect imposed and N-acetylcysteine appears to reflect a significant effect by the bipyridine ring on the C hydrogen atoms (H and on the overall stability of the complex. It is noteworthy 10 A H ). As shown in the crystal structure of 2 (Fig. 1), the side that based on previous NMR data of 1 in CD OD, the fac- B 3 chain of the cysteinate ligands adopt a “folded” confor- [Re(CO) (bpy)(H O)]+ complex is able to establish an equi- 3 2 mation, where these C hydrogen atoms (H and H ) get librium in the ligand exchange between H O and C D OD, as 10 A B 2 3 close to, or are stacked above the bipyridine aromatic rings. evidenced by observing two sets of bipyridine peaks in the Similar intramolecular interaction and folding of the axial aromatic region. The addition of D O shifted the equilibrium 2 ligand’s extended chain over the bipyridine ring has previ- to the intended product of 1, observed through coalescence ously been observed for the [Re(CO) (bpy)NC(CH ) CH ]+ of the signals [27]. 3 2 n 3 complex ions [56]. The characteristic C≡O vibrational stretching bands for Photophysical properties the Re–HCys complex (2) appeared at lower frequencies in the FT-IR spectra relative to those of the Re–Aq com- The three pure Re(I) complexes [Re(CO) (bpy) 3 plex (1). For example, the totally symmetric in-phase ν(CO) X]+/0/− (X = H O (1), H Cys− (2), N AC2− (3)) were further 2 vibrational resonance, denoted as A′(1), appeared at 2033 characterized by recording their UV–Vis absorption, excita- and 2005 cm−1 for 1 and 2, recpectively. Similar shifts to tion and emission spectra, and evaluating their fluorescence lower wavenumbers were observed in the asymmetric vibra- quantum yield (Φ ) and 1O quantum yield (Φ ) in air-equil- F 2 Δ tional frequency of the equatorial CO ligands (A″ mode, ibrated MeOH at 298 K (Table 2). Absorbance data of 1 1907 cm−1) and the totally symmetric out-of-phase ν(CO) were previously reported in DMF showing the absorption vibration (A′(2) mode, 1877 cm−1); see Figure S5 [57]. The and emission maxima at λ = 350 nm and λ = 540 nm, max em shift of the C≡O stretching vibrations to lower frequencies recpectively [58]. for 2 could be attributed to the increased electron density on The electronic absorption spectra of complexes 1, 2 and Re(I) ion from the coordinated thiolate group, which is then 3 display an intense band at 243 nm (Fig. 3), which is inde- back donated to the π* orbital of carbonyl ligands, making pendent of the type of axial ligand X. It has been attrib- the C≡O bond weaker [57]. uted to the bipyridine localized intra-ligand (1IL) π → π* As for the Re–NAC complex (3), small ESI-mass peaks transition [59, 60]. Another intense absorption band was could be observed at − m/z = 588.02 amu and + m/z = 590.04 observed in the 290–300 nm region, similar to that of fac- amu, which were attributed to[Re(CO) (bpy)(NAC)]− and [Re(CO) (bpy)(pyridine 4-thiolate)] in CH Cl at 295 nm, 3 3 2 2 its protonated form [Re(CO) (bpy)(NAC) + 2H+]+, respec- which was assigned by the authors to 1IL n → π* transition 3 tively (see Figure S2 and Table S4). In the FT-IR spec- in the aromatic ligands [61]. However, time-dependent trum of 3, the carbonyl stretching bands ν(CO) appeared at density functional theory (TD-DFT) calculations of singlet 1 3 JBIC Journal of Biological Inorganic Chemistry Table 2 Photophysical properties of complexes 1—3 Complex λ (ε)/nm (104 M−1 cm−1)a λ /nmb Φ b Φ (%)c max em F Δ 1 243 (1.8), 297 (1.32), 303 (1.3), 317 (0.935), 356 (0.325) 587 0.037 ± 0.004 15.0 ± 1.8 2 243 (1.8), 291 (1.25), 370 (0.226), 475 (0.05) 458 0.0010 ± 0.0002 1.2 ± 0.1 3 244 (2.0), 291 (1.41), 360 (0.33), 484 (0.055) 608 0.007 ± 0.001 8.5 ± 0.9 a Electronic absorption spectra were measured in MeOH at room temperature; λ = maximum absorption; molar absorption coefficient (ε) val- max ues were determined by measuring the spectra of each complex at different concentrations (see Figure S7) b In air-equilibrated MeOH at 298 K; λ = maximum emission; Φ = fluorescence quantum yield em F c Φ = 1O quantum yield in air-equilibrated MeOH at 298 K Δ 2 electronic transitions for fac-[Re(CO) (bpy)Cl] showed that between those of 1 and 2, indicating that solvent molecules 3 its high-energy, strong absorption bands < 350 nm mainly can replace the N AC2− ligands, forming fac-[Re(CO) (bpy) 3 originate from 1IL π → π* [62, 63]. Stonayov et al. assigned (CH OH)]+ (Eq. 3): 3 its peak at 292 nm (solvent = 4:1 v/v EtOH: MeOH) to fac− Re(CO) (bpy)(NAC) − Na++CH OH ligand-centered π → π*,[62] while El-Nahas et al. attributed [ 3 ] 3 its absorption peaks at 293 and 317 nm (solvent = DMF) →fac− [ Re(CO) 3 (bpy) ( CH 3 OH )] + Na+ ( NAC2− ) (3) to Re(CO) Cl → bpy transitions [63], originating from 3 HOMO-1 and HOMO-2 with mixed Re (dπ), CO (π*) and Cl This is another indication that complex 3 is unstable in (pπ) contributions, and LUMOs mainly having bpy π* char- MeOH, which is consistent with our earlier observation from acter [57, 62]. In the present study, the change in the axial its 1H NMR spectrum in CD 3 OD (Figure S6). ligand X has caused a blue shift in the peak position from Emission and excitation spectra of the Re(I) complexes 297 nm for 1 (X = H O) to 291 nm for 2 and 3 (X = HCys−, 1, 2 and 3 were also measured in air-equilibrated MeOH 2 NAC2−), where there is higher electron density on the Re(I) at 298 K (Fig. 4). The emission spectrum of the fac- ion from the electron-donating thiolate (−SR) ligands. As [Re(CO) 3 (bpy)(H 2 O)](CF 3 SO 3 ) complex (1) displayed an such, Re(CO) 3 (SR) → bpy transitions may contribute in this intense, structureless broad peak (λ em = 587 nm) that is con- band. sistent with the typical emission of diimine Re(I) tricarbonyl There are also less intense, low-energy absorption bands in the 325–550 nm region, originating from several 1CT (sin- glet charge transfer) transitions. Complex 1 shows a broad absorption feature with λ = 356 nm (tailing to ~ 480 nm). max A similar absorption was observed for fac-[Re(CO) (bpy) 3 Cl] at λ = 370 nm (tailing to ~ 430 nm), which based on max TD-DFT calculations was assigned to Re(CO) Cl → bpy 3 1CT transition (HOMO-1 → LUMO (bpy π*)) [63, 64], also described as metal–ligand to ligand charge transfer (1MLLCT) [62]. The stabilizing effect of the coordinated water molecule to Re(I) ion on the HOMO energy can explain the hypsochrmoic shift of the absorption maximum observed for complex 1 compared to fac-[Re(CO) (bpy)Cl]. 3 Substitution of H O with cysteinate ligand leads to sig- 2 nificant changes in the 325–550 nm region. Instead of a clear band, complex 2 displays a shoulder at ~ 370 nm, which is followed by two broad weak absorptions in the visible region at ~ 475 and ~ 520 nm. These changes are reminiscent of the ones observed for rhenium(I) tricarbonyl complexes with arenethiolate axial ligands [65]. As reported earlier, these low-energy bands are responsible for the orange-red color of the Re(I) thiolate complex compared to the yellow com- plex 1, and have been assigned to another charge transfer Fig. 3 UV–Vis absorption spectra of the Re(I) complexes 1–3 in state [65]. Interestingly, Fig. 3 shows that the Re(I)–NAC MeOH (C = 5.0 × 10–5 M); inset presents details in the 325–600 nm complex (3) has absorption features (λ max = 360, 484 nm) region Re 1 3 JBIC Journal of Biological Inorganic Chemistry Our data suggest that the weak emission of 2 mainly originates from the MLCT as indicated by the excitation spectra which are centered around the same wavelengths for all the complexes (Fig. 4). It has been shown for a series of [Re(CO) (N,N)(L)]+/0 complexes that the higher the electron 3 density on the Re(I) ion, the lower the energy of the 3MLCT emission band [66, 67]. Since the CO stretching vibrations of 2 appeared at lower frequencies than those of 1 (Figure S5), the electron density on the Re(I) ion is higher in 2 (from the π-donor S-cysteinate ligand). Therefore the 3MLCT tran- sitions for 2 are expected to occur at lower energies/ higher wavelengths than those of 1 (λ = 587 nm). However, the em experimental emission spectrum of 2 shows the opposite (λ = 458 nm), having a very short fluorescence lifetime em (Table S5). Taken together, these data suggest that the emis- sion of Re–HCys complex (2) could potentially originate from the 1MLCT, rather than the typical 3MLCT. The exact assignment of the emission band for 2 requires further theo- retical calculations that is beyond the scope of this study. The excitation spectrum of 2 exhibits vibronic bands that are associated with the bipyridine ligand. Indeed, the dif- ference in wavenumbers of the peak maxima is in the range of 1077–1246 cm−1, which corresponds to in-plane ring stretching and C–H in-plane bending of the bipyridine [68]. Interestingly, the emission spectrum of the Re–NAC Fig. 4 Excitation (solid line) and emission (dashed line) spectra complex (3) displayed two luminescence bands (Fig. 4): a of 1–3 in air-equilibrated MeOH at 298 K. (Exciation wavelength: low-intensity broad peak centered at 442 nm, which is close λ ex = 360 nm for 1 and 3, 290 nm for 2) to the maximum fluorescence observed for the Re–cystein- ate complex (2), and an intense broad peak at 608 nm that complexes and has been ascribed to originate from a tri- is red-shifted relative to the 3MLCT phosphorescence of plet metal-to-ligand charge-transfer (3MLCT) states [27]. the Re–Aq complex (1), and can be assigned to the fac- Upon substitution of H O with the cysteinate ligand, the [Re(CO) (bpy)(CH OH)]+ species. Observing these two 2 3 3 maximum emission for the Re–HCys complex (2) appears emission bands for 3 is consistent with its electronic absorp- at λ = 458 nm. This complex is almost non-emissive as tion features being in-between those of 1 and 2 (Fig. 3), em indicated by the extremely low fluorescence quantum and the complexity of its 1H NMR spectrum in CD OD 3 yield (Table 2), which is associated to a very short fluo- due to ligand (NAC2−)–solvent exchange (Eq. 3; Figure rescence lifetime (Table S5). While surprising, this result S6). The red shift of the 3MLCT band to 608 nm could be is not unprecedented as a similar effect has previously explained by the higher electron density on Re(I) ion in 3 been observed for the emission properties of an arenethi- [66], since CH OH is a slightly better σ donor than H O due 3 2 olate complex, fac-[Re(CO) (Pyta)(SPhOCH )], where to the inductive effect of the methyl group. Furthermore, 3 3 Pyta = 4-(2′-pyridyl)-1,2,3-triazole, and S PhOCH = 4-meth- the slight increase in the fluorescence quantum yield of 3 3 oxybenzenethiolate (λ = 450 nm) [65]. The emission bands (Φ = 0.007 ± 0.001) could be related to generation of fac- em F for similar arenethiolate complexes appeared at lower ener- [Re(CO) (bpy)(CH OH)]+ in MeOH solution, as discussed 3 3 gies, implying that they were highly dependent on the nature earlier (Eq. 3). of the thiolate and diimine (N,N) ligands. Based on DFT Singlet oxygen quantum yield We further examined the calculations, He et al. suggested that these emissions may ability of the complexes 1, 2 and 3 to produce 1O using 2 have a mixed 3MLCT/3LLCT (ligand-to-ligand charge trans- direct detection method, i.e. by exciting the samples with fer) character, originating from a π (S/benzenethiolate)/ Re a 355 nm laser source and measuring the near-IR lumines- (d ) → π*(N,N) transition [65]. Similary, Fernández-Moreira cence of 1O at 1270 nm (see Figure S8). Singlet oxygen π 2 et al. proposed that the low-intensity emission of fac- quantum yields for 1–3 could only be obtained in air-equil- [Re(CO) (bpy)(pyridine 4-thiolate)] (λ = 572 nm) could ibrated MeOH (Table 2). No signal could be detected for 3 em be assigned to the 3MLCT transition, involving a possible 1–3 in 1:1 H O:MeOH, neither for 1 dissolved in water. The 2 non-radiative secondary process (LLCT) [61]. quality of regressions was visually assessed by analysing the 1 3 JBIC Journal of Biological Inorganic Chemistry residuals plotted in Figure S8, which are randomly scattered were not efficient trans-directors, and, therefore, did not around zero, indicating that the bi-exponential model (Eq. 2) induce CO release. fits well to the data. The fitted singlet oxygen lifetimes (τ ) Δ Cytotoxicity and cellular localization are ca. 10 ± 1 μs for the three complexes 1–3, which corre- sponds to the expected value in air-equilibrated MeOH [69]. Complex 1 was the best singlet oxygen producer with a Cell viability Complexes 1 and 2 were chosen for further cell 1O quantum yield of Φ = 15.0 ± 1.8%. Upon cysteine coor- viability assessment, tested against the MDA-MB-231 breast 2 Δ dination in complex 2, a weak 1O signal could be detected cancer cell line using the alamarBlue assay. Cisplatin was 2 (Φ = 1.2 ± 0.1%), indicating that coordination of a thiolate also measured for comparison. The instability of complex Δ ligand influenced the ability of the Re(CO) (bpy)+ species to 3 in various solvents, and impurity of complex 4 led to no 3 generate 1O . This is a significant observation, because in a further investigation on these complexes. Experiments were 2 complex biological environment, the aqua ligand of 1 would performed at the time points of 24 and 48 h (Fig. 5, Table 3). be susceptible to substitution with a variety of thiol-contain- Complex 1 displayed cytotoxicity and an IC of 38 ± 6 μM 50 ing biomolecules. Complex 3 appeared to be a slightly bet- over the course of 24 h, which appears to outperform cis- ter 1O generator (Φ = 8.5 ± 0.9%) than 2, which could be platin that does not show significant cytotoxicity at 24 h. 2 Δ attributed to the formation of fac-[Re(CO) (bpy)(CH OH)]+ At 48 h, the IC value of 1 dropped to 26 ± 3 μM, whereas 3 3 50 Na+ (NAC2−) in MeOH solution (see Eq. 3). that of cisplatin was 11 ± 2 μM (Table 3). An IC value of 50 15 ± 5 μM using the HeLa cell line for 1 has been previ- Detection of CO release ously reported (over 72 h) [27]. Since complex 1 displayed cytotoxic activity in the dark, its photocytotoxicity was not To further investigate the photochemistry of Re–HCys (2) measured, despite its ability to produce 1O upon irradia- 2 and Re–NAC (3) complexes, solutions of 2 (12.5 μM) and 3 tion. An ideal photosensitizer would not be toxic in the dark, (20 μM) in deoxygenated methanol were placed in a Schlenk rendering applicability of complex 1 as a photosensitizer cuvette under argon atmosphere, and then irradiated using a doubtful. Despite this, its activity as a chemotherapeutic 365 nm UV-A hand lamp, measuring their UV–Vis spectra drug is somewhat promising, beside other complexes with at 20-min time intervals [see Figure S9(a)]. The changes in the general formula, fac-[Re(CO) (N,N)(H O)]+ [27]. 3 2 IL and MLCT bands in the absorption spectra, as well as When considering 2, the replacement of the aqua ligand observation of isosbestic points, would indicate conversion in 1 with l-cysteine appears to have a significant impact of one species to another through the release of a CO ligand on cytotoxicity. At both time points, IC values of over 50 [20]. Although the spectral changes observed for complexes 100 μM were recorded for 2. An earlier study showed that 2 and 3 were minimal, a closer look at these changes after the co-administration of l-cysteine with cisplatin reduced 2 h irradiation shows that their spectral features resemble the number of apoptotic cells in rats, indicating that it can that of the Re–Aq complex (1) shown in Fig. 3; see the act as a protecting agent [70]. N-acetyl-l-cysteine, known for comparison in Figure S9(b). This could be attributed to the its antioxidant properties, has been shown to have protect- exchange between the Re-bound thiolate ligands and solvent ing effects against cisplatin-induced toxicities and damages (MeOH) molecules over 2 h photolysis. The release of CO [71–73]. Our results in this study suggest that the formation was also measured using gas chromatography with a thermal of a Re–sulfur bond by l-cysteine can reduce the cytotoxic- conductivity detector (GC-TCD). Solutions of 1, 2 and 3 ity, and, therefore, the efficacy of complex 1. (50 μM) in deoxygenated MeOH were irradiated using the Cellular localization same apparatus over a period of 2 h, with minimal to no CO being detected (see Table S6). Earlier studies on [Re(CO) (N,N)L]+/0 complexes have The cellular accumulation of complexes 1 and 2 was 3 shown that the nature of the ligand trans to CO , such assessed and their elemental distribution maps were (axial) as phosphine-based ligands, can influence the release of the obtained using X-ray fluorescence microscopy (XFM) [74]. axially bound CO, due to their trans-directing abilities [20, For imaging techniques such as confocal or optical fluores- 30]. We used fac-[Re(bpy)(CO) (P(CH OH) )](CF SO ) as cence microscopy, it is essential for the sample to emit vis- 3 2 3 3 3 reference, and repeated the same experiment using degassed ible fluorescence light upon excitation [75]. However, the Φ F PBS as solvent (Figure S10), which led to near identical values for complexes 1 and 2 (specially for 2) were minimal results as reported in the literature (λ = 365 nm; sol- (Table 2). Moreover, there have been reports of quenching of irradiation vent = PBS, pH = 7.4) [30], indicating that the experimental the luminescence properties of Re organometallic in the cel- setup and apparatus used were sufficient for monitoring CO lular environment [76, 77]. Different with confocal or optical release. In the present study, the H Cys− and NAC2− ligands fluorescence microscopy, XFM does not rely on lumines- cence properties of the complexes. Figures 6 and S11(a) 1 3 JBIC Journal of Biological Inorganic Chemistry Fig. 5 Dose–response curve of 1, 2 and cisplatin at 24 h and 48 h. replicates. Data were analyzed with a one-way ANOVA, followed by Data are represented with error bars showing their corresponding Dunnet post-tests from comparisons between treated and control cells standard deviations from three independent experiments with nine show elemental distribution maps of MDA-MB-231 breast and in the DNA backbone, respectively [46, 78]. When com- cancer cells (1-A, 1-B and 1-C) treated for 6 h with a 20 μM paring the Zn and Re maps in cells treated with 1, there solution of 1 in DMEM, with X-ray fluorescence ( K and L ) appears to be some overlap between the Zn and Re signals α β peaks of different elements shown for a sample cell in Figure in the nuclear regions of 1-A and 1-B, indicating that these S12 (top). Rhenium accumulation in the nuclear and/or peri- two elements to some extent co-localize in these cells. Rhe- nuclear region of these cells is evident. The nuclear region nium accumulation mainly occurs in the perinuclear area of can be identified via morphology in the optical image, as 1-C cell (rather than the nuclear region); see Figure S11(a). well as by the presence of zinc and phosphorous, which are A recent report on cellular accumulation of a Re(I) predominantly found in the nucleus in zinc finger proteins tricarbonyl tetrazolato complex, fac-[Re(CO) (phen)L] 3 1 3 JBIC Journal of Biological Inorganic Chemistry Table 3 IC 50 values of 1, 2 Complex IC (μM) a (phen = 1,10′-phenanthroline; L = 5-(4-iodophenyl)tetraa- and cisplatin in MDA-MB-231 50 zolate), also revealed using the XFM technique that Re was breast cancer cells (see Fig. 5) 24 h 48 h localized within the diffuse reticular network in the nuclear 1 38 ± 6 26 ± 3 and/or perinuclear region in 22Rv1 cells, and that the com- 2 > 100 > 100 plex remained intact after uptake. Earlier studies had shown Cisplatin > 100 11 ± 2 that cellular localization of such complexes varied depend- ing on the substitution group on the aryltetrazolate ligand a Data are represented as I C 50 [78]. values ± standard deviations Another imaging study by Wilson’s group, based on from three independent experi- ments with nine replicates confocal fluorescence microscopy, suggested that a poten- tial antitumor agent Re(I) tricarbonyl complex similar to 1, i.e., fac-[Re(CO) (dmphen)(H O)](CF SO ), was distributed 3 2 3 3 Fig. 6 Optical micrographs (top left), and XRF elemental distribution area densities (quantified from standards and expressed in μg cm−2) map of MDA-MB-231 cells treated for 6 h with 1 (cell 1-A), 2 (cell are given in the bottom of each map 2-A) and DMEM as a control (Control-A). The maximum elemental 1 3 JBIC Journal of Biological Inorganic Chemistry throughout the cytosol of HeLa cells, being primarily local- The inconsistency between the results of these two studies ized in lysosomes and highly accumulated in the outer mem- was explained based on the sensitivity and dependency of branes of cytoplasmic vacuoles [27]. However, the same Re(I) luminescence properties to its speciation, which can group recently reported, based on an inductively coupled be changed upon its interaction with endogenous ligands, plasma mass spectrometry (ICP-MS) study on ovarian can- thus limiting the effectiveness and accuracy of data from cer cells, that this complex accumulates mainly in mitochon- confocal fluorescence microscopy. Our photophysical data in dria, with a small detectable amount found in nuclei [28]. “Photophysical properties” provide strong evidence to attest Fig. 7 Intracellular content of Re, P, Zn, Ca, Cu, S, K, Cl and Fe n = 5). Error bars represent standard deviations. Data were analyzed obtained by quantification using XFM as compared with the nuclear with a one-way ANOVA followed by Tukey post-tests: *p < 0.05; content of MDA-MB-231 cells treated for 6 h with control (green, **p < 0.01; ***p < 0.001 from comparisons between treated and con- number of cells tested with XFM: n = 5) as well as 20 µM solutions trol cell/nucleus regions of 1 in DMEM (blue, n = 4) and 2 in 1.5% EtOH in DMEM (orange, 1 3 JBIC Journal of Biological Inorganic Chemistry to the above explanation: the fluorescence quantum yield Chemotherapeutic drugs are typically administered intra- of 1 changed from 0.037 ± 0.004 to 0.0010 ± 0.0002 upon venously, resulting in exposure to blood plasma and a variety replacing H O in 1 with a cysteinate ligand in 2 (Table 2). of amino acids and biomolecules, which may interact with 2 The above studies on cellular localization of fac- the drug before it reaches the cellular environment [81–83]. [Re(CO) (N,N)L]0/+ complexes show that the change in This highlights the importance of studies (such as this one) 3 diimine (N,N) or axial L ligands can alter the extent and/ or focusing on interactions between potential anticancer agents the site of cellular accumulation, which is important when and biomolecules, and how such interactions would affect targeting cellular organelles other than DNA, offering alter- their activity and cellular accumulation. native modes of cell death [14, 79], helping overcome issues such as platinum-resistance mechanisms [80]. It also high- Conclusion lights the importance of using alternate techniques, such as synchrotron-based XFM, to ascertain the cellular accumu- lation of metal-based drugs. In the present study, it is par- This study reports the syntheses of three complexes formed ticularly difficult to pinpoint the exact position of complex from the reaction of a known anticancer active complex, 1 in the cells in Figs. 6 and S9(a), other than being present fac-[Re(CO) (bpy)(H O)](CF SO ) (1), with thiol-con- 3 2 3 3 in the nuclear and/or perinuclear region, as the distribution taining biomolecules, i.e. the amino acid l-cysteine, the images themselves are two-dimensional projection of three tripeptide glutathione, and N-acetyl-l-cysteine, which is a dimensional objects in dried cells [78]. This also applies to potent antioxidant with clinical applications. The crystal quantification of elemental concentrations shown in Fig. 7, structure of {fac-[Re(CO) (bpy)(HCys)]} ·9H O (2 + 1.75 3 4 2 which includes the area both over and under the nucleus such H O) is also reported. Rhenium L -edge EXAFS spectros- 2 3 as endoplasmic reticulum [78]. copy confirmed a similar local structure for Re(I) ions in its When looking at intracellular content of different ele- glutathione and N-acetylcysteine complexes in solid state. ments in all cells treated with 1, an increase in K and Cl ions Complexes with cysteine and N-acetylcysteine were insolu- can be seen when compared to cells treated with DMEM, ble in water, while the Re–glutathione complex easily dis- or 2 (Fig. 7). A possible explanation for such increase is the solved in water at physiological pH, making its purification use of PBS during sample preparation, thus exposing cells to process quite challenging. Cell viability experiments against higher concentrations of K and Cl ions. We also observed a the MDA-MB-231 breast cancer cell line using 1, 2 and cis- significant decrease in copper accumulation in cells treated platin showed 1 had a significant increase in cytotoxicity with 1 and 2, with respect to control. This observation is over the course of 24 h. When measured again at a 48 h time independent of the cytotoxicity results. point, the activity of cisplatin had surpassed that of 1 pos- When considering the cytotoxicity and IC value of sibly indicating a different mechanism of action between the 50 2, which was significantly reduced in comparison to 1 two complexes. At both time points, the cytotoxic activity (Table 3), the elemental distribution maps offer an explana- of 2 (dissolved in 1.5% EtOH) was significantly reduced, tion. In Figs. 6 and S11 (b), the Re signal in cells treated highlighting the effect of replacing the aqua ligand in 1 with with 2 is significantly low, with no strong overlap with Zn l-cysteine. The cellular localization was monitored using in the nuclear region. For example, maximum elemental synchrotron-based X-ray fluorescence microscopy (XFM), densities of Re in cells 1-A and 2-A in Fig. 6 are 0.127 and which showed 1 accumulated in the nuclear and/or peri- 0.0125 μg cm−2, respectively. The quantification histograms nuclear regions, partially co-localizing with Zn, while the (Fig. 7) show increased amounts of Re in the cells treated accumulation of 2 into the cell was considerably lower and with complex 1 (number of cells tested with XFM: n = 4) as may be hindered, explaining its reduced cytotoxic activity. compared to complex 2, which has little to no Re in a range The photophysical properties of complexes 1, 2 and of cells treated (n = 5). This difference may indicate that the Na[Re(CO) (bpy)(NAC)] (3) were also assessed in metha- 3 replacement of the aqua ligand in 1 with cysteinate (in 2) nol. The excitation and emission spectra displayed a large hinders its cellular accumulation and the passage through variance in Stokes shift between all three complexes. the cell membrane, a phenomenone that could be related The major change in the emission spectra of 1 and the to their charge difference (1 has positive change, and 2 is Re–cysteine complex 2 highlights the impact of complex neutral). An earlier study on a series of Re(I) tricarbonyl formation with biomolecules on emission properties and vis- indolato complexes with strong light-induced antiprolifera- ibility by confocal/fluorescence microscopy. The generation tive activity emphasized on the importance of their neutral of 1O in MeOH was observed for complexes 1 and 3, but not 2 charge and membrane localization on their photocytotoxic for 2. This again indicates that the coordination of cysteine effect [21]. In the current study, although the Re–HCys com- to [Re(CO) (bpy)]+ entity deactivates 1O generation. The 3 2 plex (2) is neutral, its cellular accumulation is very limited release of CO upon irradiation with UV-A light (λ = 365 nm) ( IC > 100 μM). was not observed for any of the complexes, showing that 50 1 3 JBIC Journal of Biological Inorganic Chemistry these thiol-containing biomolecules or the aqua ligand are 6. Hasan T, Ortel B, Solban N, Pogue B (2006) In: Kufe DW, Bast poor trans-directors for activating their trans CO group, and, RCJ, Hait WN, Hong WK, Pollock RE, Weichselbaum RR, Hol- land JF, Frei EI (eds) Cancer medicine. BC Decker Inc., Hamilton, therefore, their Re(I) diimine tricarbonyl complexes would pp 537–548 not be efficient as a PACT agent. 7. Muz B, de la Puente P, Azab F, Azab AK (2015) Hypoxia 3:83–92 8. Farrer NJ, Salassa L, Sadler PJ (2009) Dalton Trans Accession codes 48:10690–10701 9. Bonnet S (2018) Dalton Trans 47:10330–10343 10. Schatzschneider U (2017) Chapter 6—Metal complexes as deliv- CCDC 1981028 contains the supplementary crystallo- ery systems for CO, NO, and H S to explore the signaling network 2 graphic data for this paper. These data can be obtained free of small-molecule messengers. In: Lo KK-W (ed) Inorganic and of charge via www.ccdc.cam.ac.uk/data_reques t/cif, or by organometallic transition metal complexes with biological mol- ecules and living cells. Academic Press, pp 181–204 emailing data_request@ccdc.cam.ac.uk, or by contacting 11. Faizan M, Muhammad N, Niazi KUK, Hu Y, Wang Y, Wu Y, Sun The Cambridge Crystallographic Data Centre, 12 Union H, Liu R, Dong W, Zhang W, Gao Z (2019) Materials 12:1643 Road, Cambridge CB2 1EZ, UK; fax: + 44 1223 336033. 12. Wegiel B, Gallo D, Csizmadia E, Harris C, Belcher J, Vercellotti GM, Penacho N, Seth P, Sukhatme V, Ahmed A, Pandolfi PP, Acknowledgements We express our sincere appreciation to Mr. Wade Helczynski L, Bjartell A, Persson JL, Otterbein LE (2013) Cancer Res 74:7009–7021 White at the instrumentation facility at the Department of Chemistry 13. Zuckerbraun BS, Chin BY, Bilban M, d’Avila JDC, Rao J, Billiar for his assistance with the ESI–MS measurements, and to Ms. Valerie Brunskill for measuring the 13C and 1H NMR spectra of cysteine and TR, Otterbein LE (2007) FASEB J. 21:1099–1106 14. Gao P, Pan W, Li N, Tang B (2019) ACS Appl Mater Inter glutathione solutions. A.E.G acknowledges University of Calgary Eyes 11:26529–26558 High, and Faculty of Science Dean’s Open Competitions Doctoral 15. Bruijnincx PC, Sadler PJ (2008) Curr Opin Chem Biol 12:197–206 Scholarships. N.M. acknowledges NSERC for an Alexander Graham 16. Imberti C, Zhang P, Huang H, Sadler PJ (2020) Angew Chem Int Bell Canada Graduate Scholarship-Doctoral and Alberta Innovates for Edn 59:61–73 a Nanotechnology Doctoral Scholarship. This work was financially 17. McKenzie LK, Bryant HE, Weinstein JA (2019) Coor Chem Rev supported by the Natural Science and Engineering Research Council 379:2–29 of Canada (NSERC), the Canadian Cancer Society, the Canadian Foun- 18. Friederike R, Wiktor S (2017) Curr Med Chem 24:4905–4950 dation for Innovation (CFI), Department of Innovation and Science of 19. Leonidova A, Gasser G (2014) ACS Chem Biol 9:2180–2193 Province of Alberta. X-ray absorption spectra were measured at the 20. Pierri AE, Pallaoro A, Wu G, Ford PC (2012) J Am Chem Soc Photon Factory (PF; proposal no. 2018G563) and X-ray fluorescence 134:18197–18200 microscopy data were collected at the Advanced Photon Source (APS; 21. Kastl A, Dieckmann S, Wähler K, Völker T, Kastl L, Merkel AL, proposal no. 53103). Use of the Advanced Photon Source was sup- Vultur A, Shannan B, Harms K, Ocker M, Parak WJ, Herlyn M, ported by the U.S. Department of Energy, Office of Science, Office Meggers E (2013) Chem Med Chem 8:924–927 of Basic Energy Sciences, under contract No. DE-AC02-06CH11357. 22. Leonidova A, Pierroz V, Rubbiani R, Heier J, Ferrari S, Gasser G (2014) Dalton Trans 43:4287–4294 Funding The following funding is acknowledge: Natural Science and 23. Zobi F, Blacque O, Schmalle HW, Spingler B, Alberto R (2004) Engineering Research Council of Canada (Grant no. RGPIN 2016- Inorg Chem 43:2087–2096 04546 to FJ and RGPIN 2018-04773 to CS), Canadian Cancer Society 24. Lecina J, Palacios O, Atrian S, Capdevila M, Suades J (2015) J (Grant no. 300072 to CS); Canadian Foundation for Innovation (Grant Biol Inorg Chem 20:465–474 no. 9479 to FJ); Department of Innovation and Science of Province of 25. Lopez V, Foolad F, Kelleher SL (2011) Cancer Lett 304:41–51 Alberta (Grant to FJ). 26. Alam S, Kelleher SL (2012) Nutrients 4:875–903 27. Knopf KM, Murphy BL, MacMillan SN, Baskin JM, Barr MP, Compliance with ethical standards Boros E, Wilson JJ (2017) J Am Chem Soc 139:14302–14314 28. Konkankit CC, King AP, Knopf KM, Southard TL, Wilson JJ Conflict of interest The authors declare that they have no conflict of (2019) ACS Med Chem Lett 10:822–827 29. Williams DBG, Lawton M (2010) J Org Chem 75:8351–8354 interest. 30. Marker SC, MacMillan SN, Zipfel WR, Li Z, Ford PC, Wilson JJ (2018) Inorg Chem 57:1311–1331 31. Salignac B, Grundler PV, Cayemittes S, Frey U, Scopelliti R, References Merbach AE, Hedinger R, Hegetschweiler K, Alberto R, Prinz U (2003) Inorg Chem 42:3516–3526 32. APEX3 crystallography software suite. Bruker AXS Inc., Madison 1. Galluzzi L, Senovilla L, Vitale I, Michels J, Martins I, Kepp O, (2016) Castedo M, Kroemer G (2011) Oncogene 31:1869–1883 33. Dolomanov OV, Bourhis LJ, Gildea RJ, Howard JA, Puschmann 2. Szaciłowski K, Macyk W, Drzewiecka-Matuszek A, Brindell M, H (2009) J Appl Crystallogr 42:339–341 Stochel G (2005) Chem Rev 105:2647–2694 34. Sheldrick GM (2015) Acta Crystallogr A 71:3–8 3. Monro S, Colón KL, Yin H, Roque J, Konda P, Gujar S, Thum- 35. Sheldrick GM (2015) Acta Crystallogr C Struc Chem 71:3–8 mel RP, Lilge L, Cameron CG, McFarland SA (2019) Chem Rev 36. Ressler T (1998) J Synchrotron Rad 5:118–122 119:797–828 37. Zabinsky SI, Rehr JJ, Ankudinov A, Albers RC, Eller MJ (1995) 4. van Straten D, Mashayekhi V, de Bruijn HS, Oliveira S, Robinson Phys Rev B 52:2995–3009 DJ (2017) Cancers (Basel) 9:19 38. Ankudinov AL, Rehr JJ (1997) Phys Rev B 56:R1712–R1716 5. Dougherty TJ, Gomer CJ, Henderson BW, Jori G, Kessel D, Kor- 39. Rimmer RD, Richter H, Ford PC (2010) Inorg Chem belik M, Moan J, Peng Q (1998) J Natl Cancer Inst 90:889–905 49:1180–1185 40. Fery-Forgues S, Lavabre D (1999) J Chem Educ 76:1260 1 3 JBIC Journal of Biological Inorganic Chemistry 41. Degen J, Reinecke K, Schmidtke H-H (1992) Chem Phys 65. He M, Ching HV, Policar C, Bertrand HC (2018) New J Chem 162:419–426 42:11312–11323 42. Sabol JE, Rockley MG (1987) J Photochem Photobiol A Chem 66. Fredericks SM, Luong JC, Wrighton MS (1979) J Am Chem Soc 40:245–257 101:7415–7417 43. Demas JN, McBride RP, Harris EW (1976) J Phys Chem 67. Ramos LD, da Cruz HM, Frin KPM (2017) Photochem Photobiol 80:2248–2253 Sci 16:459–466 44. Enriquez Garcia A, Lai B, Gopinathan SG, Harris HH, Shemanko 68. Strukl JS, Walter JL (1971) Spectrochim. Acta A. Mol. Spectrosc CS, Jalilehvand F (2019) Chem Comm 55:8223–8226 27:209–221 45. McRae R, Lai B, Fahrni CJ (2013) Metallomics 5:52–61 69. Wilkinson F, Helman WP, Ross AB (1995) J Phys Chem Ref Data 46. McRae R, Lai B, Vogt S, Fahrni CJ (2006) J Struct Biol 155:22–29 24:663–677 47. Van Epsen P (2002) In: Van Grieken RE, Markowicz AA (eds) 70. Ahmed EA, Omar HM, Elghaffar SKA, Ragb SMM, Nasser AY Handbook of X-ray spectrometry: methods and techniques. Marcel (2011) Food Chem Toxicol 49:1115–1121 Dekker Inc., New York, pp 239–339 71. Wang F, Liu S, Shen Y, Zhuang R, Xi J, Fang H, Pan X, Sun J, 48. Vogt S, Maser J, Jacobsen C (2003) J Phys IV 104:617–622 Cai Z (2014) Exp Ther Med 8:1939–1945 49. Vogt S (2003) J Phys IV 104:635–638 72. Rosic G, Selakovic D, Joksimovic J, Srejovic I, Zivkovic V, 50. Carter EA, Rayner BS, McLeod AI, Wu LE, Marshall CP, Levina Tatalović N, Orescanin-Dusic Z, Mitrovic S, Ilic M, Jakovljevic A, Aitken JB, Witting PK, Lai B, Cai Z, Vogt S, Lee YC, Chen CI, V (2016) Toxicol Lett 242:34–46 Tobin MJ, Harris HH, Lay PA (2010) Mol Biosyst 6:1316–1322 73. Huang S, You J, Wang K, Li Y, Zhang Y, Wei H, Liang X, Liu Y 51. Allen FH (2002) Acta Crystallogr B B58:380–388 (2019) Biomed Res Int 2019:4805853 52. El Nahhas A, van der Veen RM, Penfold TJ, Pham VT, Lima FA, 74. Jin Q, Paunesku T, Lai B, Gleber S-C, Chen S, Finney L, Vine D, Abela R, Blanco-Rodriguez AM, Záliš S, Vlček A, Tavernelli I, Vogt S, Woloschak G, Jacobsen C (2017) J Microsc 265:81–93 Rothlisberger U, Milne CJ, Chergui M (2013) J Phys Chem A 75. Sanderson MJ, Smith I, Parker I, Bootman MD (2014) Cold 117:361–369 Spring Harb Protoc. https: //doi.org/10.1101/pdb.top071 795 53. Lobo-Lapidus RJ, Gates BC (2010) Chem Eur J 16:11386–11398 76. Kitanovic I, Can S, Alborzinia H, Kitanovic A, Pierroz V, Leoni- 54. Koeppe RE, Stroud RM (1976) Biochemistry 15:3450–3458 dova A, Pinto A, Spingler B, Ferrari S, Molteni R, Steffen A, Met- 55. Fuhr BJ, Rabenstein DL (1973) J Am Chem Soc 95:6944–6950 zler-Nolte N, Wölfl S, Gasser G (2014) Chem Eur J 20:2496–2507 56. Reitz GA, Dressick WJ, Demas JN, DeGraff BA (1986) J Am 77. Gasser G, Neumann S, Ott I, Seitz M, Heumann R, Metzler-Nolte Chem Soc 108:5344–5345 N (2011) Eur J Inorg Chem 5471–5478 57. Vlček A (2010) In: Lees AJ (ed) Topics in organometallic chem- 78. Wedding JL, Harris HH, Bader CA, Plush SE, Mak R, Massi istry: photophysics of organometallics. Springer, Heidelberg, pp M, Brooks DA, Lai B, Vogt S, Werrett MV (2017) Metallomics 73–114 9:382–390 58. Kurz P, Probst B, Spingler B, Alberto R (2006) Eur J Inorg Chem 79. Obinna CU, Heather MW (2014) Curr Pharm Des 20:201–222 2006:2966–2974 80. Florea A-M, Büsselberg D (2011) Cancers 3:1351–1371 59. Giordano PJ, Wrighton MS (1979) J Am Chem Soc 81. Barnham KJ, Djuran MI, Murdoch PdS, Ranford JD, Sadler PJ 101:2888–2897 (1996) Inorg. Chem 35:1065–1072 60. Worl LA, Duesing R, Chen P, Ciana LD, Meyer TJ (1991) J Chem 82. Ramachandran E, Senthil Raja D, Bhuvanesh NSP, Natarajan K Soc Dalton Trans 849–858 (2012) Dalton Trans 41:13308–13323 61. Fernández-Moreira V, Sastre-Martín H (2017) Inorg Chim Acta 83. Lukey MJ, Katt WP, Cerione RA (2017) Drug Discov Today 460:127–133 22:796–804 62. Stoyanov SR, Villegas JM, Cruz AJ, Lockyear LL, Reibenspies JH, Rillema DP (2005) J Chem Theory Comput 1:95–106 Publisher’s Note Springer Nature remains neutral with regard to 63. El Nahhas A, Consani C, Blanco-Rodríguez AM, Lancaster KM, jurisdictional claims in published maps and institutional affiliations. Braem O, Cannizzo A, Towrie M, Clark IP, Záliš S, Chergui M (2011) Inorg Chem 50:2932–2943 64. Cannizzo A, Blanco-Rodríguez AM, El-Nahhas A, Šebera J, Záliš S, An V, Chergui M (2008) J Am Chem Soc 130:8967–8974 Affiliations Miles S. Capper1 · Alejandra Enriquez Garcia1 · Nicolas Macia1 · Barry Lai2 · Jian‑Bin Lin1 · Masaharu Nomura3 · Amir Alihosseinzadeh4 · Sathish Ponnurangam4 · Belinda Heyne1 · Carrie S. Shemanko5 · Farideh Jalilehvand1 * Farideh Jalilehvand 3 Institute for Materials Structure Science, High faridehj@ucalgary.ca Energy Accelerator Research Organization, Oho 1-1, Tsukuba 305-0801, Japan 1 Department of Chemistry, University of Calgary, Calgary, 4 Department of Chemical and Petroleum Engineering, AB T2N 1N4, Canada University of Calgary, Calgary, AB T2N 1N4, Canada 2 Advanced Photon Source, X-Ray Science Division, Argonne 5 Department of Biological Sciences, University of Calgary, National Laboratory, Argonne, IL 60439, USA Calgary, AB T2N 1N4, Canada 1 3