Neutral iridium(III) complexes bearing BODIPY−substituted N−heterocyclic carbene (NHC) ligands: synthesis, photophysics, in vitro theranostic photodynamic therapy, and antimicrobial activity

Abstract

The synthesis, photophysics, and photobiological activities of a series of novel neutral heteroleptic cyclometalated iridium(III) complexes incorporating boron dipyrromethene (BODIPY) substituted N-heterocyclic carbene (NHC) ligands (Ir1 – Ir5) are reported. The effect of the substitution position of BODIPY on the NHC ligands, either on C4 of the phenyl ring (Ir1 – Ir3) or C5 of the benzimidazole unit (Ir4 and Ir5), and its linker type (single or triple bond) on the photophysical properties was studied. Ir1 − Ir5 exhibited BODIPY−localized intense ¹IL (intraligand transition) / ¹MLCT (metal-to-ligand charge transfer) absorption at 530–543 nm and ¹IL/³CT (charge transfer) emission at 582–610 nm. The nanosecond transient absorption results revealed that the lowest triplet excited states of these complexes were the BODIPY−localized ³π,π* states. Complexes Ir4 and Ir5 exhibited blue-shifted ¹IL absorption and ¹IL/³CT emission bands compared to the corresponding absorption and emission bands in complexes Ir1 and Ir3. However, replacing the methyl substituents on N3 of benzimidazole in complexes Ir1 and Ir4 with oligoether substituents in Ir3 and Ir5, respectively, did not impact the energies of the low-energy absorption and emission bands in the corresponding complexes. Water-soluble complexes Ir3 and Ir5 have been explored as photosensitizers for in vitro photodynamic therapy (PDT) effects toward human SKMEL28 melanoma cells. Ir3 showed no dark cytotoxicity (EC₅₀ >300 μM) but good photocytotoxic activity (9.66 ± 0.28 μM), whereas Ir5 exhibited a higher dark cytotoxicity (20.2 ± 1.26 μM) and excellent photocytotoxicity (0.15 ± 0.01 μM). The phototherapeutic indices with visible light (400–700 nm) activation were >31 for Ir3 and 135 for Ir5. Ir3 and Ir5 displayed ¹O₂ quantum yields of 38% and 22% in CH₃CN, respectively, upon 450 nm excitation. Ir5 was more effective at generating reactive oxygen species (ROS) in vitro. Ir5 was also active against Staphylococcus aureus upon visible light activation, with a phototherapeutic index of >15 and EC₅₀ value of 6.67 μM. These photobiological activities demonstrated that these neutral Ir(III) complexes are promising in vitro PDT reagents, and substitution at C5 on the benzimidazole group of the NHC ligand was superior to C4 substitution on the phenyl ring.

Graphical abstract

Neutral Ir(III) complexes bearing BODIPY−substituted N−heterocyclic carbene ligands exhibited moderate to strong photodynamic therapeutic effects toward SKMEL28 cells and antimicrobial activity upon broadband visible light activation.

1. Introduction

Photodynamic therapy (PDT) is an emerging cancer treatment modality that combines a photosensitizer (PS), oxygen, and light to destroy tumors and tumor vasculature and induce an immune response.^1–5 PDT is inherently selective because toxicity is confined to regions where the PS, oxygen, and light overlap spatiotemporally.^3,6,7 During PDT, reactive oxygen species (ROS) are formed through energy (type II) or electron transfer (type I) reactions that take place between ground state molecular oxygen and the excited triplet state of the PS. The most important mediator of the PDT effect is thought to be singlet oxygen (¹O₂) produced through the type II PDT mechanism, but superoxide (O₂^-•) or hydroxyl radical (^•OH), generated by the type I PDT mechanism may also play a role.

Certain transition metal complexes are among the numerous PSs that have been explored due to their (1) interesting photophysical properties, (2) reduced photobleaching, and (3) improved kinetic stabilities (compared to organic PSs).⁸ Among the metal complexes, Ru(II) systems are probably the most widely studied for PDT.^9,10 Our Ru(II) complex TLD1433,^10,11 which incorporates a π-expansive ligand derived from imidazo[4,5-f][1,10]phenanthroline appended to an α-terthienyl unit, just successfully completed a Phase 1b clinical trial for treating bladder cancer with PDT (ClinicalTrials.gov Identifier: NCT03053635). This example demonstrates the utility of lowest-energy π,π* triplet states, in this case centered on the oligothienyl unit, with extremely long lifetimes for highly efficient photosensitization.

Some transition-metal complexes possess the distinct advantage of having singlet and triplet excited states that can be tuned independently for optimization of their respective characteristics when appropriate π-conjugated ligands are chosen. We have reported that it is possible to shift the singlet charge-transfer absorption to longer wavelengths (for activation by red wavelengths of light) by extending the π-conjugation of the organic ligands, while allowing the lowest triplet excited state (T₁) to be localized on the π-expansive organic ligands.¹² The π,π* nature of the lowest triplet excited state results in very long triplet excited state lifetimes, which have the potential to facilitate ROS production even at low oxygen tension. Unlike organic PSs, metal complexes possess high quantum yields for triplet state formation due to the heavy atom effect (metal-induced rapid ISC), and their triplet excited states are typically long-lived. All of these properties are desirable features for next-generation PSs.

Ir(III) complexes are also of interest because some have exhibited high efficiency for ROS production via electron or energy transfer.¹⁴ Many Ir(III) complexes possess high quantum yields for triplet excited state formation and long-lived triplet excited states for efficient ROS generation even under hypoxia.¹³ It has been reported that cationic Ir(III) complexes can target mitochondria,^14–18 lysosomes,^19,20 the endoplasmic reticulum,^21,22 or nuclei^23,24 in a variety of cancer cell lines. A mitochondria-targeted Ir(III) complex PS was reported to show improved PDT effects under hypoxia.¹⁴ These complexes can also display bright intracellular luminesce, imparting theranostic capacity.^{13–20,23–25} Therefore, Ir(III) complexes have emerged as a new platform for theranostic PDT applications. However, most of the currently studied Ir(III) complexes absorb light maximally in the UV to blue spectral regions, wavelengths of lesser interest for PDT because of their shallow tissue penetration due to tissue scattering and absorption.²⁶

Although it is possible to red-shift the charge transfer ground-state absorption bands of the cyclometalating monocationic Ir(III) complexes [Ir(ĈN)₂(N^N)]⁺ (where N^N refers to the diimine ligand and ĈN refers to the cyclometalating ligand) to longer wavelength by introducing electron-withdrawing substituents on the diimine ligands²⁷ or through benzannulation at the diimine ligands, the molar extinction coefficients of the charge transfer absorption bands are quite low and the lowest triplet excited state (T₁) lifetimes become much shorter (tens of ns).^28–32 In contrast, it has been shown that bichromophoric transition-metal complexes with π-conjugated organic chromophores result in long-lived ³π,π* T₁ states localized at the organic chromophore with simultaneous red-shifting of the ground-state absorption to the longer wavelengths. Metal coordination improves π-conjugation across the organic ligands through planarization, facilitating the interactions between ligands. It also increases the quantum yield of triplet excited-state formation via heavy atom induced ISC. Such a strategy has been applied in TLD1433¹⁰ and other Ru(II) complexes for PDT applications,^11,33–39 but is relatively rare for Ir(III) systems.^40–44

Borondipyrromethene (BODIPY) and its derivatives, a class of strong light−harvesting fluorophores with facile wavelength tunability, have been extensively explored as potential PSs for PDT applications in recently years.^45–48 For example, Dong and co-workers have demonstrated that nanoparticles encapsulating BODIPY or aza-BODIPY derivatives bearing diethylaminophenyl or dimethylaminophenyl substituents exhibited very strong pH-triggered photodynamic and photothermal therapy for Hela cells and Hela tumor bearing nude mice.^47,48 Moreover, BODIPY appears as a promising moiety to be attached to a monodentate or polydentate ligand to tune the absorption wavelength of the transition-metal PSs to the desired spectral region.^39–41,49 However, BODIPY-Ir(III) dyads for PDT applications are not well-studied.^40,41,49 One of the few examples used BODIPY tethered cationic cyclometalated Ir(III) dyads as PSs for PDT and bioimaging applications, and their in vitro PDT effects were marginal due to high dark toxicity (IC₅₀ = 8.16 – 16.70 μM toward 1121 or LLC cell lines). Phototherapeutic index (PI) values (1.3–3.8) were very small.^40,41

Recently, N−heterocyclic carbene (NHC) based iridium(III) complexes have attracted great interest.^50–54 Unlike the numerous well−studied monocationic cyclometalated Ir(III) complexes [Ir(ĈN)₂(N^N)]⁺ (where N^N refers to the diimine ligand and ĈN refers to the cyclometalating ligand), monoanionic NHC ligands, such as 1-phenyl-3-alkylbenzimidazole, as the N^N ligand gives rise to neutral complexes.^54,55 To date, NHC-Ir(III) complexes have largely been studied for their catalytic⁵⁰ and optoelectronic applications.^51–54 Biological activities were only reported on monocationic Ir(III) complexes with bis(alkylated-NHC) ligands,^15,50 which possessed high dark toxicity and could only be activated by UV light (365 nm). To date, the biological activities of the neutral Ir(ĈN)₂(NHC) complexes remain unexplored.

Herein, a series of neutral iridium(III) complexes bearing BODIPY−NHC ligands (Chart 1) were designed and synthesized to explore their photophysical properties and in vitro PDT effects. These complexes all contain benzo[h]quinoline (bhq) as the cyclometalating ligands, but the NHC ligand varies from 1−(4−BODIPY−ethynylphenyl)-3−methyl−1H−benzo[d]imidazole−3−iumiodide (L1), 1−(4−BODIPY−phenyl)−3−methyl−1H−benzo[d]imidazole−3−iumiodide (L2), 1−(4−BODIPY−ethynylphenyl)−3−(2−(2−(2−methoxyethoxy)ethoxy)ethyl−1H−benzo[d]imidazol−3−iumiodide (L3), 5−(4−BODIPY−ethynylphenyl)−3−methyl−1H−benzo[d]imidazol− 3−iumiodide (L4), to 5−(4−BODIPY−ethynylphenyl)−3−(2−(2−(2− methoxyethoxy)ethoxy)ethyl−1H−benzo[d]imidazol−3−iumiodide (L5). BODIPY was incorporated into the NHC ligand scaffold at different sites through a single bond (Ir2) or a triple bond (Ir1 and Ir3−Ir5) in order to red-shift the absorption of the complexes to longer wavelengths. An oligoether chain, 2−(2−(2−methoxyethoxy)ethoxy)ethyl, was appended to Ir3 and Ir5 to increase the hydrophilicity of the neutral complexes. The photophysical properties of these five complexes were systematically investigated by various spectroscopic methods and simulated by density functional theory (DFT) calculations. The in vitro theranostic PDT effects of the two water-soluble complexes Ir3 and Ir5 were investigated. The impact of the BODIPY attachment site on the photophysics and PDT effects of these dyads was explored.

Chart 1.

Structures of BODIPY-NHC-Ir(III) Complexes Ir1–Ir5.

2. Experiments section

2.1. Materials and synthesis

All reagents and solvents were purchased from Alfa Aesar and VWR International and used as received. Al₂O₃ gels (activated, neutral) and silica gels (230−400 mesh) for column chromatography were purchased from Sorbent Technology. Benzo[h]quinoline (ĈN ligand) was obtained from Alfa−Aesar and its Ir(III) μ−chloro−bridged dimers [Ir(ĈN)₂Cl]₂ was synthesized according to the literature procedure.⁵⁶ Compounds 1-(4-bromophenyl)-1H-benzimidazole (compound 1 in Scheme 1),⁵⁷ 4-bromo-2-nitro-N-phenylaminobenzene (compound 6 in Scheme 1),⁵⁸ 4-bromo-N¹-phenyl-1,2-benzenediamine (compound 7 in Scheme 1),⁵⁹ and 2−iodo−BODIPY⁶⁰ were synthesized following the reported procedures. The synthetic route for the synthesis of ligands L1-L5 and complexes Ir1-Ir5 are provided in Scheme 1. ¹H NMR, high−resolution mass spectrometry (HRMS), and elemental analyses were used to characterize these complexes. ¹H NMR spectra were obtained on a Varian Oxford−400 or Bruker−400 spectrometer in CDCl₃ using tetramethylsilane (TMS) as the internal standard. Bruker BioTOF III mass spectrometer was used for ESI−HRMS analyses. NuMega Resonance Laboratories, Inc. in San Diego, California conducted elemental analyses.

Scheme 1.

Synthetic Route for Ligands L1−L5 and Complexes Ir1−Ir5. Reagents and conditions: (i) 1−bromo−4−iodobenzene, CuI, Cs₂CO₃, 1,10-phenanthroline, DMF, 110°C, 40 hrs.; (ii) CuI, PdCl₂(PPh₃)₂, ethynyltrimethylsilane, TEA/dioxane, 100 °C, 1 hr.; (iii) TBAF, THF, 0°C, 1 hr.; (iv) 2−Iodo−BODIPY, CuI, Pd(PPh₃)₄, TEA, THF, 50 °C, 16 hrs.; (v) triisopropyl borate, ⁿBuLi, anhydrous THF, −78 °C, 1 hr., then r.t. overnight, followed by HCl, 100 °C, 1 hr.; (vi) 2−iodo−BODIPY, K₂CO₃, Pd(PPh₃)₄, toluene, THF, water, 70 °C, overnight; (vii) NaOAc, aniline, 160 °C, 7 hrs.; (viii) Na₂S, S, water, reflux, 6 hrs.; (ix) formic acid, 100 °C, overnight; (x) MeI, THF, 100 °C, 24 hrs.; (xi) 1−iodo−3,6,9−trioxadecane, toluene, 100 °C, 96 hrs.; (xii) [Ir(benzo[h]quinoline)₂(μ−Cl)]₂, Ag₂O, 1,2−dichloroethane, reflux, 24 hrs.

2.2. Synthesis of precursor compounds 2−5 and 8−10

Compound 2.

A solution of 1 (544 mg, 2 mmol), CuI (19 mg, 0.1 mmol), PdCl₂(PPh₃)₂ (70 mg, 0.1 mmol) in trimethylamine (TEA)/dioxane (2:1, v/v) was degassed with N₂ at room temperature for 20 minutes. Then, ethynyltrimethylsilane (0.86 mL, 6 mmol) was added and the resulting reaction mixture was heated at 100 °C for 1 hour under N₂. After cooling to room temperature, the mixture was concentrated and purified by column chromatography over silica gel using hexane:ethyl acetate = 1:2 (v/v) as the eluent to give product as pale yellow solid. This pale yellow solid (500 mg, 1.8 mmol) was dissolved in THF (10 ml), and THF solution of tetrabutylammonium fluoride (TBAF) (10 mL, 1 M in THF) was added at 0 °C. The reaction mixture was stirred at room temperature for 1 hour. Then the solution was diluted with ethyl acetate (EA) and washed with water and brine consequently. The organic layer was dried over Na₂SO₄ and concentrated. The residue was purified by column chromatography over silica gel using hexane:ethyl acetate= 1:2 (v/v) as the eluent to give compound 2 as pale yellow solid (360 mg, 81%). ¹H NMR (400 MHz, CDCl₃) δ 8.14 (s, 1H), 7.91 (dd, J = 6.4, 2.7 Hz, 1H), 7.73 (d, J = 8.4 Hz, 2H), 7.57 (dd, J = 6.6, 2.7 Hz, 1H), 7.53 (d, J = 8.4 Hz, 2H), 7.43−7.34 (m, 2H), 3.22 (s, 1H).

Compound 3.

The solution of 2-iodo−BODIPY (200 mg, 0.44 mmol), compound 2 (97 mg, 0.44 mmol), CuI (17 mg, 0.09 mmol), and Pd(PPh₃)₄ (57 mg, 0.049 mmol) in 10 mL degassed mixed solvent (THF/TEA = 4/1) was heated to 50 ^oC and stirred for 16 hours in dark. After cooling to room temperature, the solvent was removed under reduced pressure and the residue was extracted with CH₂Cl₂ and washed with brine. After removal of the solvent from the organic layer, the crude product was purified by flash column chromatography with CH₂Cl₂/acetone (v/v = 50/1) being used as the eluent to obtain a dark red solid (200 mg, 83%). ¹H NMR (400 MHz, with CDCl₃) δ 8.14 (s, 1H), 7.94−7.86 (m, 1H), 7.67 (d, J = 8.4 Hz, 2H), 7.62−7.44 (m, 6H), 7.43−7.30 (m, 4H), 6.08 (s, 1H), 2.75 (s, 3H), 2.62 (s, 3H), 1.55 (s, 3H), 1.44 (s, 3H).

Compound 4.

Under −78 °C, ⁿBuLi (0.5 mL, 2.5 M in hexane) was added dropwise to the solution of compound 1 (151 mg, 0.56 mmol) and B(OⁱPr)₃ (157 mg, 0.83 mmol) in THF (10 mL). The mixture was stirred at −78°C for 1 h, and then allowed to warm up to room temperature, and continued stirring for overnight. The reaction was quenched by addition of concentrated aqueous NH₄Cl (10 mL), and the volatiles were evaporated in vacuum. Then, HCl (10 mL, 37% aq.) was added to the solution and the resultant mixture was heated to reflux for 1 h. After cooling to room temperature, the mixture was carefully adjusted to pH=7 with NaOH. The white suspension was extracted with CH₂Cl₂. After removal of the solvent, the residue was dried under vacuum. The crude product was purified by column chromatography on silica gel and eluted with acetone to obtain white solid as the product (47 mg, 36%). ¹H NMR (400 MHz, DMSO) δ 8.60 (s, 1H), 8.24 (s, 1H), 8.04 (d, J = 8.0 Hz, 2H), 7.79 (d, J = 7.4 Hz, 1H), 7.67 (d, J = 7.9 Hz, 2H), 7.34 (dt, J = 13.6, 6.8 Hz, 2H).

Compound 5.

A solution of compound 4 (44 mg, 0.18 mmol), 2-iodo-BODIPY (56 mg, 0.12 mmol), and K₂CO₃ (166 mg, 1.2 mmol) in toluene (8 mL), THF (8 mL), and water (2 mL) was degassed with N₂ at room temperature for 20 minutes. After that, Pd(PPh₃)₄ (3 mg, 0.0025 mmol) was added, and the mixture was heated to 70 °C for overnight. After evaporation of the volatiles under reduced pressure, water (10 mL) was added, and the mixture was extracted with CH₂Cl₂ (3×10 mL). The combined organic layers were washed with brine (100 mL), and dried over anhydrous MgSO₄. The solvent was then removed, and the crude product was purified by column chromatography on silica gel using hexane/ethyl acetate (v/v = 1:1) as the eluent to obtain orange solid as the target compound (79 mg, 78%). ¹H NMR (400 MHz, CDCl₃) δ 8.13 (d, J = 7.5 Hz, 1H), 7.92−7.84 (m, 1H), 7.62−7.45 (m, 7H), 7.40−7.29 (m, 5H), 6.04 (s, 1H), 2.61 (s, 3H), 2.59 (s, 3H), 1.41 (s, 3H), 1.36 (s, 3H).

Compound 8.

The mixture of compound 7 (880 mg, 3.36 mmol) and formic acid (20 mL) was stirred at 100 °C for overnight. The reaction mixture was cooled to room temperature and concentrated under reduced pressure to afford a crude solid. The crude solid was partitioned between ethyl acetate (250 mL) and NH₃/H₂O (25 mL). The ethyl acetate layer was combined, dried with Na₂SO₄, and then concentrated under reduced pressure. The residue was purified by column chromatography (silica gel, eluted with CH₂Cl₂) to get the target compound as off−white solid (860 mg, 98%). ¹H NMR (400 MHz, CDCl₃) δ 8.10 (s, 1H), 8.02 (d, J = 1.5 Hz, 1H), 7.59 (dd, J = 8.2, 7.4 Hz, 2H), 7.52−7.47 (m, 3H), 7.46−7.36 (m, 2H).

Compound 9.

Following the similar synthetic procedure for compound 2, compound 9 was obtained with 87% yield using compound 8 as the starting material. ¹H NMR (400 MHz, CDCl₃) δ 8.13 (s, 1H), 8.06−7.96 (m, 1H), 7.59 (dd, J = 10.0, 5.5 Hz, 2H), 7.49 (dd, J = 11.7, 5.7 Hz, 4H), 7.42 (dt, J = 13.5, 5.1 Hz, 1H), 3.07 (s, 1H).

Compound 10.

Following the similar synthetic procedure for compound 3, compound 10 was obtained with 41% yield using compound 9 as the starting material. ¹H NMR (500 MHz, CDCl₃) δ 8.14 (s, 1H), 7.97 (s, 1H), 7.59 (t, J = 7.7 Hz, 2H), 7.54−7.45 (m, 7H), 7.42 (d, J = 8.4 Hz, 1H), 7.33−7.28 (m, 2H), 6.03 (s, 1H), 2.74 (s, 3H), 2.59 (s, 3H), 1.54 (s, 3H), 1.41 (s, 3H).

2.3. General synthetic procedure for ligands L1−L5.

In a round−bottom flask equipped with a gastight Teflon cap, the precursor NHCs compound 3, 5, or 10 (1 mmol) was dissolved in 10 mL of THF (for synthesizing L1, L2, or L4) or in 1 mL of toluene (for synthesizing L3 and L5). Then 1.5 mmol CH₃I (for synthesizing L1, L2, or L4) or 1−iodo−3,6,9−trioxadecane (for synthesizing L3 and L5) was added. The mixture was heated to 100 °C for 24 h. After the reaction, the solution was allowed to cool to room temperature. The formed precipitate was collected by filtration and washed with THF, dichloroethane, and Et₂O (10 mL × 2) to obtain the corresponding target compound.

L1. A dark red powder was obtained as the product (yield: 73%). ¹H NMR (400 MHz, DMSO) δ 10.16 (s, 1H), 8.18 (d, J = 8.4 Hz, 1H), 8.05−7.71 (m, 8H), 7.65 (d, J = 3.7 Hz, 2H), 7.46−7.40 (m, 2H), 6.39 (s, 1H), 4.20 (s, 3H), 2.66 (s, 3H), 2.56 (s, 3H), 1.53 (s, 3H), 1.43 (s, 3H).

L2. A red powder was obtained as the product (yield: 67%). ¹H NMR (400 MHz, DMSO) δ 10.15 (s, 1H), 8.16 (d, J = 8.2 Hz, 1H), 7.94−7.69 (m, 6H), 7.63 (dd, J = 12.4, 7.5 Hz, 4H), 7.46 (d, J = 7.6 Hz, 2H), 6.29 (s, 1H), 4.17 (s, 3H), 2.52 (s, 3H), 2.50 (s, 3H), 1.38 (s, 3H), 1.36 (s, 3H).

L3. A dark red powder was obtained as the product (yield: 53%). ¹H NMR (500 MHz, CDCl₃) δ 10.72 (s, 1H), 7.97 (d, J = 7.6 Hz, 1H), 7.88 (d, J = 8.7 Hz, 2H), 7.74−7.66 (m, 4H), 7.53 (dd, J = 5.0, 1.9 Hz, 2H), 7.33−7.27 (m, 4H), 6.07 (s, 1H), 5.09−5.03 (m, 2H), 4.22−4.14 (m, 2H), 3.74 (dd, J = 5.5, 3.4 Hz, 2H), 3.60 (dd, J = 5.4, 3.5 Hz, 2H), 3.54 (dd, J = 5.6, 3.7 Hz, 2H), 3.41 (dd, J = 5.6, 3.7 Hz, 2H), 3.28 (s, 3H), 2.73 (s, 3H), 2.60 (s, 3H), 1.53 (s, 3H), 1.42 (s, 3H).

L4. A dark red powder was obtained as the product (yield: 58%). ¹H NMR (400 MHz, DMSO) δ 10.17 (s, 1H), 8.36 (s, 1H), 7.82−7.73 (m, 7H), 7.62 (d, J = 3.6 Hz, 3H), 7.53−7.34 (m, 2H), 6.37 (s, 1H), 4.17 (s, 3H), 2.66 (s, 3H), 2.54 (s, 3H), 1.52 (s, 3H), 1.41 (s, 3H).

L5. A dark red powder was obtained as the product (yield: 65%). ¹H NMR (400 MHz, CDCl₃) δ 10.61 (s, 1H), 8.19 (s, 1H), 7.97−7.88 (m, 3H), 7.79−7.68 (m, 4H), 7.64−7.54 (m, 4H), 7.35−7.30 (m, 1H), 6.10 (s, 1H), 5.08−4.98 (m, 2H), 4.24−4.14 (m, 2H), 3.81−3.72 (m, 2H), 3.67−3.61 (m, 2H), 3.54 (dd, J = 5.4, 3.6 Hz, 2H), 3.45 (dd, J = 5.7, 3.7 Hz, 2H), 3.30 (s, 3H), 2.75 (s, 3H), 2.63 (s, 3H), 1.56 (s, 3H), 1.45 (s, 3H).

2.4. General synthetic procedure for iridium complexes Ir1-Ir5.

The mixture of the NHC ligand (L1-L5, 0.12 mmol), [Ir(benzo[h]quinoline)₂(μ-Cl)]₂ (76 mg, 0.06 mmol), Ag₂O (56 mg, 0.24 mmol), and 1,2-dichloroethane (15 mL) was heated to reflux in the dark for 24 h. After the mixture was cooled to room temperature, the solvent was removed under reduced pressure to afford the crude product. Purification of the crude product was carried out on silica gel column chromatography eluted with CH₂Cl₂. The obtained product was further recrystallized in CH₂Cl₂ / hexane to give the desired pure complex.

Ir1. A dark purple powder was obtained as the product (yield: 32%). ¹H NMR (400 MHz, CDCl₃) δ 8.39 (d, J = 5.2 Hz, 1H), 8.27 (d, J = 5.2 Hz, 1H), 8.17 (d, J = 8.2 Hz, 1H), 8.03 (dd, J = 12.1, 7.6 Hz, 2H), 7.85−7.66 (m, 3H), 7.57−7.46 (m, 4H), 7.39−7.21 (m, 9H), 7.18−6.99 (m, 4H), 6.87 (s, 1H), 6.74 (d, J = 6.9 Hz, 1H), 6.33 (d, J = 7.0 Hz, 1H), 5.98 (s, 1H), 3.26 (s, 3H), 2.55 (s, 3H), 2.53 (s, 3H), 1.36 (s, 3H), 1.33 (s, 3H). ESI-HRMS (m/z): calcd. for [C₆₁H₄₄BF₂IrN₆+H]⁺, 1103.3396; found, 1103.3384. Anal. Calcd for C₆₁H₄₄BF₂IrN₆·5CH₂Cl₂·2.4C₆H₁₄ (C₆H₁₄: hexane): C, 55.71; H, 5.09; N, 4.85. Found: C, 55.37; H, 5.42; N, 5.15.

Ir2. A red powder was obtained as the product (yield: 30%). ¹H NMR (400 MHz, CDCl₃) δ 8.40 (d, J = 5.2 Hz, 1H), 8.30 (d, J = 5.5 Hz, 1H), 8.18 (d, J = 8.1 Hz, 1H), 8.04 (d, J = 8.0 Hz, 1H), 7.97 (d, J = 7.8 Hz, 1H), 7.83 (d, J = 7.9 Hz, 1H), 7.78 (d, J = 8.7 Hz, 1H), 7.69 (d, J = 8.7 Hz, 1H), 7.60−7.40 (m, 4H), 7.35 (t, J = 8.1 Hz, 2H), 7.22 (d, J = 6.6 Hz, 6H), 7.10 (ddd, J = 22.5, 14.7, 6.6 Hz, 4H), 6.89 (d, J = 6.9 Hz, 1H), 6.79 (d, J = 8.3 Hz, 1H), 6.43 (s, 1H), 6.26 (d, J = 7.3 Hz, 1H), 5.91 (s, 1H), 3.28 (s, 3H), 2.52 (s, 3H), 2.15 (s, 3H), 1.33 (s, 3H), 1.26 (s, 3H). ESI-HRMS (m/z): calcd. for [C₅₉H₄₄BF₂IrN₆+H]⁺, 1079.3396; found, 1079.3389. Anal. Calcd for C₅₉H₄₄BF₂IrN₆·0.4CH₂Cl₂·0.6C₆H₁₄: C, 65.02; H, 4.61; N, 7.22. Found: C, 64.62; H, 4.92; N, 7.19.

Ir3. A dark purple powder was obtained as the product (yield: 56%). ¹H NMR (400 MHz, CDCl₃) δ 9.36 (d, J = 5.4 Hz, 1H), 8.39 (d, J = 5.9 Hz, 1H), 8.27 (dd, J = 12.3, 7.0 Hz, 2H), 8.16 (d, J = 8.5 Hz, 1H), 8.03 (dd, J = 16.6, 8.0 Hz, 2H), 7.83 (d, J = 8.3 Hz, 1H), 7.74 (t, J = 8.2 Hz, 2H), 7.68 (d, J = 8.8 Hz, 1H), 7.56−7.47 (m, 3H), 7.39−7.26 (m, 4H), 7.25−7.00 (m, 6H), 6.87 (d, J = 1.8 Hz, 1H), 6.81 (t, J = 7.6 Hz, 1H), 6.68 (d, J = 7.0 Hz, 1H), 6.28 (d, J = 7.0 Hz, 1H), 5.98 (d, J = 7.1 Hz, 1H), 4.37−4.17 (m, 2H), 3.75 (dd, J = 10.5, 5.7 Hz, 2H), 3.39−3.32 (m, 2H), 3.29 (s, 3H), 3.02−2.83 (m, 2H), 2.75−2.59 (m, 2H), 2.59 (s, 3H), 2.52 (s, 3H), 2.40 (dd, J = 4.1, 2.8 Hz, 2H), 1.36 (s, 3H), 1.32 (s, 3H). ESI-HRMS (m/z): calcd. for [C₆₇H₅₆BF₂IrN₆O₃+H]⁺, 1235.4183; found, 1235.4164. Anal. Calcd for C₆₇H₅₆BF₂IrN₆O₃·1.6CH₂Cl₂·0.3C₆H₁₄: C, 60.57; H, 4.58; N, 6.02. Found: C, 60.74; H, 4.55; N, 5.75.

Ir4. A dark purple powder was obtained as the product (yield: 57%). ¹H NMR (400 MHz, CDCl₃) δ 8.27 (d, J = 4.6 Hz, 1H), 8.20 (d, J = 5.5 Hz, 1H), 8.06 (d, J = 8.7 Hz, 1H), 8.02−7.93 (m, 2H), 7.79−7.68 (m, 3H), 7.54−7.42 (m, 5H), 7.38 (t, J = 8.4 Hz, 1H), 7.30 (dd, J = 11.9, 7.4 Hz, 2H), 7.22 (dd, J = 6.5, 3.1 Hz, 3H), 7.11−6.95 (m, 5H), 6.75 (d, J = 5.7 Hz, 1H), 6.68 (dd, J = 11.5, 7.1 Hz, 2H), 6.28 (d, J = 7.1 Hz, 1H), 5.98 (s, 1H), 3.19 (s, 3H), 2.64 (s, 3H), 2.52 (s, 3H), 1.43 (s, 3H), 1.34 (s, 3H). ESI-HRMS (m/z): calcd. for [C₆₁H₄₄BF₂IrN₆+H]⁺, 1103.3396; found, 1103.3385. Anal. Calcd for C₆₁H₄₄BF₂IrN₆·4.8CH₂Cl₂·2C₆H₁₄: C, 55.55; H, 4.89; N, 5.00. Found: C, 55.46; H, 5.14; N, 5.26.

Ir5. A dark purple powder was obtained as the product (yield: 65%). ¹H NMR (400 MHz, CDCl₃) δ 8.34 (d, J = 5.5 Hz, 1H), 8.25 (d, J = 5.4 Hz, 1H), 8.11 (d, J = 8.5 Hz, 1H), 8.07−8.03 (m, 1H), 8.02−7.98 (m, 1H), 7.83 (dd, J = 14.7, 8.1 Hz, 1H), 7.75 (dd, J = 8.9, 2.0 Hz, 2H), 7.67 (d, J = 1.9 Hz, 1H), 7.59−7.46 (m, 5H), 7.46−7.27 (m, 5H), 7.17−7.00 (m, 5H), 6.80 (dd, J = 5.1, 3.6 Hz, 1H), 6.74 (t, J = 7.6 Hz, 1H), 6.67 (dd, J = 6.2, 2.5 Hz, 1H), 6.29 (dd, J = 6.3, 3.3 Hz, 1H), 6.04 (s, 1H), 4.31−4.24 (m, 2H), 3.77−3.71 (m, 2H), 3.42−3.37 (m, 2H), 3.33−3.29 (m, 2H), 3.27 (s, 3H), 3.25−3.22 (m, 2H), 2.96−2.85 (m, 2H), 2.70 (s, 3H), 2.58 (s, 3H), 1.50 (s, 3H), 1.40 (s, 3H). ESI-HRMS (m/z): calcd. for [C₆₇H₅₆BF₂N₆O₃Ir+H]⁺, 1235.4183; found, 1235.4169. Anal. Calcd for C₆₇H₅₆BF₂IrN₆O₃.0.3H₂O: C, 64.92; H, 4.60; N, 6.78. Found: C, 64.97; H, 4.77; N, 6.49.

2.5. Photophysical studies.

The spectroscopic grade solvents used for photophysical studies were purchased from VWR International and used as received without further purification. The ultraviolet−visible (UV−vis) absorption spectra of complexes Ir1-Ir5 were measured on a Varian Cary 50 spectrophotometer. Steady−state emission spectra of complexes Ir1-Ir5 were recorded using a HORIBA FluoroMax 4 fluorometer/phosphorometer. The relative actinometry method was used for measuring the emission quantum yields of Ir1-Ir5. [Ru(bpy)₃]Cl₂ in degassed acetonitrile (λ_max = 436 nm, Φ_em = 0.097)⁶¹ was used as the reference. The nanosecond transient absorption (TA) measurements, i.e. the TA spectra, triplet lifetimes, and quantum yields, were carried out on a laser flash photolysis spectrometer (Edinburgh LP920). The excitation source was the third−harmonic output (355 nm) of a Quantel Brilliant Nd:YAG laser (pulse duration, 4.1 ns; repetition rate was 1 Hz). Before measurement, each sample solution was purged with nitrogen for 40 min. The singlet depletion method was followed to calculate the triplet excited−state molar extinction coefficients (ε_T),⁶² and the relative actinometry method⁶³ was used to estimate the triplet excited−state quantum yields using SiNc in benzene as the reference (ε₅₉₀ = 70,000 M⁻¹cm⁻¹, Φ_T = 0.20).⁶⁴

2.6. Computational methodology.

The ground state geometries were optimized using the Density Functional Theory (DFT)⁶⁵ employing PBE1 functional⁶⁶ and the mixed basis set, where LANL2DZ⁶⁷ was used for Ir(III) to incorporate the effective core potential for a heavy element and 6–31G* ⁶⁸ basis set was used for all remaining atoms. The solvent effects were modeled by the Conductor-like Polarizable Continuum Model (CPCM) reaction field method⁶⁹ for toluene.

Linear response time dependent DFT (TDDFT)⁷⁰ was employed to calculate excited state properties of the complexes by using the same functional and basis set that were used for the ground state calculations. It was previously reported for other Ir(III) complexes that both the ground and excited state calculations using this methodology show good agreement with the experimental results,⁷¹ which defined our choice of the methodology. To meet the same energy range of the experimental absorption spectra, the lowest 80 states were calculated by TDDFT. The thermal linewidths of spectra were obtained using Gaussian function with the broadening parameter of 0.08 eV, which well fits to the line shape of experimental absorption spectra.

To calculate the emission energies, we have optimized the triplet excited state using the analytical gradient method based on TDDFT, within PBE1 functional and the mixed LANL2DZ/6–31G* basis set. To get a better understanding of the nature of the transitions, natural transition orbitals (NTOs)⁷² were calculated, which is the compact representation of the excited electron-hole pair obtained from the transition density matrices calculated by TDDFT. NTOs were visualized by VMD⁷³ software using 0.02 iso resolution. All quantum chemical calculations were performed using the Gaussian09^® software package.⁷⁴

2.7. Photobiological activity studies.

The details of the cellular assays, cell culture, cytotoxicity and photocytotoxicity cell assays, confocal microscopy, bacterial survival assays, and in vitro ROS generation measurements are provided in the Supporting Information.

3. Results and discussion

3.1. Electronic absorption.

The UV−vis absorption spectra of Ir1-Ir5 were measured in toluene, CH₃CN, THF, and CH₂Cl₂ at room temperature. The spectra in toluene (the preferred solvent due to the high solubility of the neutral complexes in it) are displayed in Fig. 1, and the normalized spectra in the other solvents are provided in ESI Fig. S1. The absorption band maxima and molar extinction coefficients are summarized in Table 1. Ir1-Ir5 exhibited strong absorption bands with vibronic structure in the region of 280−460 nm, and broad, intense absorption bands centered between 500–600 nm. The energies and spectral features of these low-energy absorption bands resemble those of the 1,3,5,7-tetramethyl-8-phenyl-BODIPY.⁷⁵ However, the band maxima of Ir1-Ir5 are approximately 30–40 nm red-shifted compared to the ¹π,π* absorption band of the 1,3,5,7-tetramethyl-8-phenyl-BODIPY, and these bands are broader than that of 1,3,5,7-tetramethyl-8-phenyl-BODIPY. These characteristics imply that the NHC ligand interacts with the BODIPY motif and there could be some charge transfer character in this band (see discussion vide infra).

Fig. 1.

UV−vis absorption spectra of Ir1−Ir5 in toluene.

Table 1.

Photophysical parameters for complexes Ir1−Ir5.

	λabs/nm (log ε /L·mol−1·cm−1) a	λem/nm (τem/μs); Φem b	λT1−Tn/nm (τTA/μs; log εT1−Tn/ L·mol−1·cm−1); ΦT c	ΦΔ d (λex/nm)
Ir1	294 (4.86), 335 (4.68), 352 (4.56), 398 (4.40), 430 (4.33), 543 (4.90)	610 (4.96); 0.041	457 (42.1; -e), 679 (48.0; 4.13); 0.51	n.d.f
Ir2	286 (4.78), 347 (4.46), 398 (4.29), 530 (4.87)	583 (3.12); 0.063	458 (24.0; -e), 681 (28.7; 3.83); 0.86	n.d.f
Ir3	290 (4.86), 335 (4.69), 352 (4.57), 397 (4.42), 430 (4.33), 543 (4.88)	610 (5.26); 0.047	455 (26.1; -e), 690 (32.1; 4.11); 0.51	0.38 (450); 0.37 (534)
Ir4	322 (4.74), 350 (4.64), 396 (4.40), 535 (4.88)	587 (4.77); 0.034	449 (28.0; -e), 699 (30.9; 4.09); 0.39	n.d.f
Ir5	309 (4.87), 350 (4.73), 395 (4.50), 535 (4.86)	587 (4.89); 0.010	452 (52.9; -e), 690 (55.2; 4.06); 0.40	0.11 (352); 0.22 (450); 0.08 (534)

^aAbsorption band maxima (λ_abs) and molar extinction coefficients (log ε) in toluene at room temperature.

^bEmission band maxima (λ_em), lifetimes (τ_em), and quantum yields (Φ_em) in toluene at room temperature, c = 1×10⁻⁵ mol/L. The reference used was a degassed acetonitrile solution of [Ru(bpy)₃]Cl₂ (Φ_em = 0.097, λ_ex = 436 nm)⁶¹ The emission lifetimes reported are the longer-lived ³ILCT/³MLCT phosphorescence lifetimes. The short-lived ¹IL fluorescence lifetimes were unable to be reliably measured due to the resolution of our instrument.

^cNanosecond TA band maxima (λ_T1−Tn), triplet excited−state lifetimes (τ_TA), triplet extinction coefficients (log ε_T1−Tn), and quantum yields (Φ_T) measured in toluene at room temperature. SiNc in benzene (ε₅₉₀ = 7×10⁴ L mol⁻¹cm⁻¹, Φ_T = 0.20)⁶⁴ was used as the reference for calculating Φ_T.

^dSinglet oxygen quantum yields (Φ_Δ) were measured in acetonitrile and are corrected to within ±5%.

^eNot determined due to strong ground-state absorption, which does not satisfy the condition to apply singlet depletion method for estimation of ε_T1−Tn.

^fNot determined.

Inserting a C≡C bond between the BODIPY motif and the phenyl ring in complex Ir1 led to a red-shift of approximately 13-nm compared to that of complex Ir2. This difference was attributed to the extended π-conjugation within the NHC-BODIPY ligand of Ir1. Varying the point of attachment of BODIPY on the NHC ligand also showed a noticeable effect on the energy of the low-energy absorption band. Complex Ir4, with BODIPY attached at C5 of the benzimidazole ring blue shifted the low-energy absorption band by approximately 8 nm compared to the corresponding band in Ir1, where BODIPY is appended at C4 of the phenyl ring. Replacement the methyl substituents on C3 of benzimidazole in complexes Ir1 and Ir4 by oligoether substituents as in Ir3 and Ir5 did not impact the energy of the low-energy absorption band but slightly attenuated the molar extinction coefficients for this transition.

To better understand the nature of the different absorption bands, time-dependent density functional theory (TDDFT) calculations were carried out for complexes Ir1−Ir5 in toluene. As shown in ESI Fig. S1, the calculated spectra matched well with the experimental spectra. The natural transition orbitals (NTOs) corresponding to the major transitions contributing to the major absorption bands of Ir1−Ir5 are presented in Table 2, and in ESI Tables S1 and S2.

Table 2.

NTOs for lowest-energy transitions of complexes Ir1-Ir5 in toluene.

	Sn	Hole	Electron
Ir1	S1 533 nm f = 0.633
Ir2	S1 480 nm f = 0.507
Ir3	S1 531 nm f = 0.647
Ir4	S1 515 nm f = 0.655
Ir5	S1 513 nm f = 0.641

As the NTOs in Table 2 indicated, the electrons of the S₁ states were almost exclusively localized on the BODIPY moiety, while the holes were delocalized on the entire BODIPY−substituted NHC ligands and on the metal d orbitals. Therefore, the S₁ states in Ir1−Ir5 have the mixed ¹π,π* / ¹ILCT (intraligand charge transfer, π(NHC)→π*(BODIPY)) / ¹MLCT (metal-to-ligand charge transfer, d(Ir)→π*(BODIPY)) configurations, which contributed to the intense low-energy absorption bands. For the absorption band(s) in the 370–450 nm region, the NTOs in ESI Table S1 suggest major ligand-to-ligand charge transfer (¹LLCT, π(bhq)→π*(BODIPY)) / ¹MLCT transitions, admixing with some BODIPY-NHC localized ¹IL (intraligand transition, i.e. ¹π,π* / ¹ILCT) characters. In contrast, according to the NTOs shown in Table S2, the high energy, strong absorption bands at 280–370 can be predominantly assigned to the spin-allowed bhq or BODIPY-NHC ligand-centered ¹π,π* transitions, mixed with ¹LLCT, ¹MLCT, and minor ¹ILCT configurations.

3.2. Photoluminescence.

The steady−state emission spectra for complexes Ir1-Ir5 were measured in degassed toluene, CH₃CN, THF, and dichloromethane at room temperature. The normalized emission spectra in toluene are presented in Fig. 2, and the emission maxima and lifetimes are compiled in Table 1. The normalized emission spectra and the emission data in other solvents are provided in ESI Fig. S3 and Table S3. As shown in Fig. 2, all of the complexes exhibited broad and structureless luminescence in toluene ranging from 582 to 610 nm, which were mirror images to their corresponding ¹IL absorption bands. The emission signals showed bi-exponential decays, with a short lifetimes of several nanoseconds and a longer lifetimes of several microseconds. The emission was only partially quenched by air. These characteristics imply that the emission of these complexes could be mixed with ¹IL fluorescence and charge transfer phosphorescence (³CT). This fluorescence/phosphorescence mixing has been reported for a cationic dinuclear Ir(III) complex linked by BODIPY.⁴¹

Fig. 2.

Normalized experimental emission spectra of Ir1−Ir5 in deaerated toluene at room temperature. The excitation wavelength was 543 nm for Ir1 and Ir3, 530 nm for Ir2, and 535 nm for Ir4 and Ir5. The open-headed lines represent the calculated S₁ fluorescence and the solid-headed lines represent the calculated T₂ phosphorescence. The inset shows the comparison of emission intensity in air-saturated and deaerated toluene solutions for Ir1.

Similar to the trend observed from the UV-vis absorption experiments, the emission maxima of Ir1 and Ir3 were identical to those of Ir4 and Ir5, respectively. This indicates that the oligoether chain did not impact the energies of the emitting states in these complexes. In contrast, the emission of Ir2 with BODIPY singly-bonded to NHC showed a pronounced blue-shift (27 nm) compared to that of Ir1, with the C≡C linker, due to the shorter π-conjugation and the reduced co-planarity between BODIPY and NHC ligand in Ir2. The attachment position of BODIPY on the NHC ligand also affected the emission energy as it did for the UV-vis absorption, namely, the emission maxima of Ir4 and Ir5 displayed a blue-shift of 23 nm with respect to those of Ir1 and Ir3.

To assign the nature of the emitting states for these complexes, TDDFT calculations were performed based on the optimized singlet triplet excited state geometries. The singlet and triplet excited-state NTOs obtained for Ir1-Ir5 are compiled in Table 3 for S₁ and T₂ states, and in ESI Table S4 for T₁ states. Because neither the calculated T₁ energies nor the energy trend matched the experimental results, and the emission lifetimes were much shorter than the lifetimes obtained from the ns transient absorption measurement (which will be discussed in the next section), we believe that the phosphorescence components of the observed emission from these complexes are from the T₂ states. Based on the NTOs provided in Table 3, the fluorescent S₁ states have mixed ¹ILCT/¹π,π*/¹MLCT configurations, while the T₂ states are predominantly ³ILCT/³MLCT configurations mixed with minor ³π,π* character. The S₁ states and the T₂ states are in energetic proximity, therefore, they both contribute to the observed emission. However, these two states are not in thermal equilibrium because the emission decay was bi-exponential, with a short component of less than 10 ns and a longer lifetime of several microseconds.

Table 3.

NTOs corresponding to the optimized S₁ and T₂ states of complexes Ir1–Ir5 in toluene.

S1				T2
	Energy	Hole	Electron	Energy	Hole	Electron
Ir1	647 nm			623 nm
Ir2	598 nm			538 nm
Ir3	644 nm			621 nm
Ir4	609 nm			586 nm
Ir5	605 nm			584 nm

3.3. Transient absorption (TA).

To further understand the triplet excited state characteristics of Ir1-Ir5, especially the non-emitting T₁ states in these complexes, nanosecond TA were investigated in deaerated toluene at room temperature. The TA spectra of Ir1-Ir5 at zero-time delay upon 355-nm excitation are presented in Fig. 3. The TA parameters, such as the TA band maxima, triplet excited−state lifetimes and quantum yields, and triplet extinction coefficients, are compiled in Table 1. Fitting of the TA signals revealed a long-lived species (ca. 28–55 μs). In addition, the TA spectral features of Ir1-Ir5 were similar, with ground-state bleaching arising from the BODIPY-NHC localized ¹IL absorption. These spectra are also similar to that of the iodo-BODIPY reported in the literature.⁷⁵ Therefore, the triplet excited states that produced the observed TA should have the same nature for all of the complexes, likely from the BODIPY localized ³π,π* state. The lifetimes of Ir1-Ir5 obtained from the decay of the TA signals are distinctively different from their emission lifetimes, indicating that the excited states observed by TA in these complexes are different from their emitting excited states. The long TA lifetimes were assigned to excited states with ³π,π* characters. The NTOs for T₁ states in ESI Table S4 confirmed that the T₁ states for these complexes are of ³π,π* configuration and localized exclusively on the BODIPY motif.

Fig. 3.

Nanosecond transient difference absorption spectra of complexes Ir1−Ir5 in toluene at room temperature immediately after 355 nm laser pulse excitation. A₃₅₅ = 0.4 in a 1-cm cuvette.

3.4. Singlet oxygen generation.

Ir3 and Ir5 were designed with oligoether substituents as R3 (Chart 1) to increase the hydrophilicities of these neutral Ir(III) NHC complexes for improved aqueous solubility (which is a desirable property for in vitro biological testing). Our detailed photophysical and computational analyses of Ir3 and Ir5 (compared to their corresponding methyl-substituted Ir2 and Ir4 relatives) indicated that the addition of the oligoether groups did not alter the general excited state dynamics observed for these systems. Therefore, the more water-soluble Ir3 and Ir5 were further investigated for their photobiological activities.

The abilities of Ir3 and Ir5 to generate singlet oxygen (¹O₂) were assessed as quantum yields for ¹O₂ production (Φ_Δ) in CH₃CN relative to [Ru(bpy)₃](PF₆)₂ as the standard (Φ_Δ = 56%). The direct method was employed, whereby sensitized ¹O₂ was detected by its phosphorescence centered at 1268 nm. CH₃CN was used as the solvent since ¹O₂ phosphorescence is quenched in aqueous solution.⁷⁶ The value of Φ_Δ for Ir3 was near 37% and largely independent of excitation wavelength, whereas that for Ir5 ranged from 8–22% as a function of excitation wavelength. Excitation at 352, 450, and 534 nm yielded values for Φ_Δ of 11, 22, and 8%, respectively.

Ir3 was the more efficient ¹O₂ generator at all wavelengths investigated, which could be attributed, in part, to the fact that its quantum yield for triplet state formation is larger (51 versus 40%). Clearly, the substitution position of the BODIPY chromophore, either on the meso-phenyl versus the benzimidazole portion of the NHC ligand, impacts the excited state dynamics and photophysical properties of the complexes. These structural differences may, therefore, be expected to also alter their photobiological properties.

3.5. Cytotoxicity and photocytotoxicity toward cancer cells.

Based on their abilities to generate ¹O₂ under cell-free conditions, albeit with modest efficiency, it was anticipated that both Ir3 and Ir5 would give rise to in vitro photobiological effects. The human melanoma cell line SKMEL28 was used to probe the cytotoxicities and photocytotoxicities of Ir3 and Ir5. Briefly, cells were dosed 1 nM-300 μM of Ir3 or Ir5 (dissolved in water supplemented with 10% DMSO (v/v) and serially diluted with phosphate buffered saline (PBS)) and incubated for 16 h before receiving a dark treatment or a light treatment. The illumination condition was 100 J cm⁻² of either broadband visible or monochromatic red (625 nm) light. After a dark or light treatment, the cells were incubated for 48 h, treated with the resazurin cell viability indicator,⁷⁷ and incubated an additional 2–4 h. The relative cell viabilities for the dark and light treatments of cells dosed with complex were quantified by the metabolic reduction of resazurin to its fluorescent product resorufin (Fig. 4, Table 4).

Fig. 4.

In vitro dose-response curves for complexes Ir3 (left) and Ir5 (right) in SKMEL28 cells treated in the dark (black) and with visible (blue) or red (red) light activation.

Table 4.

Comparison of EC₅₀ values (μM) for SKMEL28 cancer cells dosed with complexes Ir3 and Ir5.

	Dark	Visa	PIb	Redc	PIb
Ir3	> 300	9.66 ± 0.28	>31	53.7 ± 2.1	>6
Ir5	20.2 ± 1.3	0.15 ± 0.01	135	13.0 ± 0.1	2

^aVis−PDT: 16 h drug-to-light interval followed by 100 J cm⁻² broadband visible light (400–700 nm) irradiation,

^bPI = phototherapeutic index (ratio of dark EC₅₀ to visible-light EC₅₀),

^cRed-PDT: 16 h drug-to-light interval followed by 100 J cm⁻² light irradiation with 625 nm LEDs.

Ir3 was nontoxic toward SKMEL28 cells (EC₅₀ >300 μM), while Ir5 was substantially more cytotoxic (EC₅₀ = 20 μM). With broadband visible light activation, both complexes were photocytotoxic. The photocytotoxicy of Ir5 was submicromolar, with an EC₅₀ value of 150 nM and PI of 135, while the visible-light EC₅₀ value for Ir3 was approximately 10 μM and its PI was 31. The photocytotoxicities of both complexes with red light were substantially attenuated (5-fold for Ir3 and over 85-fold for Ir5), giving rise to much smaller PI values (6 for Ir3 and 2 for Ir5).

Despite having much higher cytotoxicity in the dark, Ir5 was the better in vitro photosensitizing agent according to both the magnitude of its photocytotoxicity and its phototherapeutic margin. This was initially surprising since the value of Φ_Δ for Ir3 was significantly larger than that for Ir5. This discrepancy could be due to differences in ¹O₂ production under cell-free conditions versus the in vitro experiment, or that reactive oxygen species (ROS) other than ¹O₂ may be involved in the phototoxic mechanism, or due to different cell uptake.

To probe for the involvement of ROS, more broadly defined, as a possible source of the photocytotoxic activity for Ir3 and Ir5, SKMEL28 cells were preincubated with the cell-permeable and highly sensitive 2’,7’-dichlorodihydrofluorescein diacetate (DCFDA) fluorogenic dye before performing (photo)cytotoxicity assays (Fig. 5). The compounds were tested in the range of 4 pM to 126 μM, and a sub-lethal light dose of 50 J cm⁻² broadband visible light was used for the light condition with a reduced incubation time of 30 min between compound addition and light treatment. DCFDA is deacetylated by cellular esterases to a nonfluorescent compound, which is subsequently oxidized by ROS to the highly fluorescent DCF product that can be detected by its characteristic emission.⁷⁸ ROS that can be detected by this method include superoxide anion (O₂^•–), hydrogen peroxide (H₂O₂), hydroxyl radical (HO^•), and singlet oxygen (¹O₂), which react directly or indirectly with the deacetylated probe.⁷⁹

Fig. 5.

Reactive oxygen species assay results for SKMEL28 cells treated with Ir3 (a), Ir5 (b), or the positive control tert-butyl hydrogen peroxide TBHP (c) using DCFDA as a ROS probe. Cells were treated in the dark (black bars) or with 50 J cm⁻¹ visible light (blue bars). ROS production was measured at 120 min post-treatment.

Dark treatments were included to quantify baseline ROS levels under the assay conditions for comparison, and tert-butyl hydrogen peroxide (TBHP) was used as a positive control (Fig. 5c). The signal produced by TBHP was relatively weak, but sufficient, owing to the short incubation that was necessary to ensure a sub-lethal treatment with photoactivated Ir3 and Ir5. ROS generation with Ir3 in the dark was minimal at all concentrations tested, while cells treated with Ir5 in the dark showed elevated ROS levels at concentrations greater than 1 μM. This observation may account for the cytotoxicity induced by Ir5 in the absence of a light trigger, whereby Ir5 was at least 15× more cytotoxic (in the dark) than Ir3.

Light-treated complexes Ir3 and Ir5 increased the detected fluorescence from the DCF product in a dose-dependent manner over the dark controls and also in comparison to the positive control THBP (Fig. 5). The highest concentration of light-treated Ir5 (126 μM) showed a slight decrease in the DCF fluorescence relative to the second highest concentration (39 μM). This attenuation is typical when the cells begin to lose viability, which is why it is important to choose a sub-lethal dose for ROS quantification. Ir5 produced significantly more DCF fluorescence than Ir3 at the highest concentrations tested.

It is possible that the much larger PI measured for Ir5 (compared to Ir3) may stem from its more efficient light-mediated ROS production. ROS other than ¹O₂ could be involved based on the lower value of Φ_Δ measured for Ir5 under the cell-free condition. However, the level of DCF fluorescence is only raised at relatively high PS concentrations, which seems to indicate the absence of a good correlation between photocytotoxicity and widespread redox imbalance. It seems more likely that the large difference in phototoxic effects for Ir3 and Ir5 stems from factors that were not directly probed as part of this study, including differences in the biological target(s) for each PS and possibly contact-mediated interactions with such targets.⁸⁰

Confocal microscopy was used to observe the qualitative effects of Ir3 and Ir5 on SKMEL28 cells with a dark or visible light treatment of 50 J cm⁻² (Fig. 6). A concentration of 50 μM was chosen because a marked difference in the dark cytotoxicity of Ir3 and Ir5 at this concentration was anticipated (since 50 μM is greater than the dark EC₅₀ value of Ir3, but not of Ir5). The light treatment was chosen to be 50 J cm⁻² (half the dose of the cell cytotoxicity assays) and imaging was done at 15 min post-treatment to ensure that cells were imaged under sub-lethal conditions, where a comparison of cellular morphologies and uptake might be made.

Fig. 6.

Laser scanning confocal microscopy images of SKMEL28 cells dosed with 50 μM Ir3 (top row) or Ir5 (bottom row) in the dark (a) or with 50 J cm⁻² visible light (b).

The differences in the ³MLCT emission quantum yields for Ir3 and Ir5 were apparent in the dark confocal images, where cellular uptake by Ir3 was detectable as a very weak signal but no signal was detected for Ir5. Qualitatively, there appeared to be more cellular debris from dead and dying cells treated with Ir5 in the dark (Fig. 6a, bottom row) but emission was not detected. Both compounds caused a general change in the morphology of a large fraction of the cells treated in the dark, from the typical elongated and spindle shape to spherical.

With light activation, intracellular emission from Ir3 was much more pronounced (Fig. 6b, top row). If the signal intensity is proportional to concentration, then photoactivated uptake of Ir3 results in accumulation in the cytoplasm and mitochondria but not in the nuclei. Interestingly, Ir5 produced luminescence in the cellular debris from dead and dying cells that were treated with visible light and very faint luminescence from the intact cells. Images of cells treated with Ir3 and light did not exhibit the dark clumps of cellular debris that were observed in the images of Ir5, possibly reflecting the differences in photocytotoxicities and/or aggregation properties (in the presence of cellular remains) of these two complexes.

Photoactivated uptake also led to a larger concentration of Ir5 in cells (assuming that signal intensity is proportional to concentration), making its detection possible. Given that luminescence from Ir3 is almost five-fold more efficient than that from Ir5 (and this difference could be magnified in cells), the confocal imaging experiment cannot definitively establish that cellular uptake of Ir3 higher than Ir5. In addition, the Stokes shift associated with Ir3 is larger than that of Ir5, where inner-filter effects could contribute to the lower signal intensity associated with Ir5. It would be expected that the higher dark toxicity of Ir5 would be correlated with better cellular uptake, but additional intracellular compound quantification methods were not carried out as part of this study.

3.6. Cytotoxicity and photocytotoxicity toward bacteria.

The antimicrobial photobiological activities of Ir3 and Ir5 were also assessed using Staphylococcus aureus (S. aureus) growing as planktonic cultures (Table 5, Fig. 7). The complexes were tested at concentrations between 0.4 and 100 μM, and the light treatments were approximately 35 J cm⁻² of visible light. Despite showing photobiological activity against SKMEL28 melanoma cells, Ir3 was inactive against S. aureus both in the dark and with a light treatment. On the other hand, Ir5 was nontoxic to S. aureus in the dark but phototoxic with a visible light EC₅₀ of approximately 7 μM (PI > 15). Relatively few examples of metal complexes as PSs for antimicrobial PDT (aPDT) have been reported,^38,81–84 making Ir5 notable in that respect.

Table 5.

Comparison of EC₅₀ values (μM) for Staphylococcus aureus dosed with Ir3 or Ir5.

	Dark	Visa	PIb
Ir3	>100	>100	-
Ir5	>100	6.67 ± 0.07	>15

^a35 J cm⁻² broadband visible light irradiation;

^bPI = phototherapeutic index (ratio of dark EC₅₀ to visible-light EC₅₀).

Fig. 7.

Bacterial cell survival dose-response bar graphs for complexes Ir3 (a) and Ir5 (b) in S. aureus, treated in the dark (black) or with 35 J cm⁻¹ visible (blue) light. In panel b, the bacterial survival rates were zero when Ir5 concentrations reached ≥ 12.5 μM, thus no blue bars were shown for those concentrations.

The observation that Ir3 was less photobiologically active (inactive) than Ir5 against S. aureus agreed with the trend in cancer cells, further highlighting the discrepancy between the ¹O₂ quantum yields and phototoxicities for these two complexes. Why Ir5 is less effective than Ir3 is not known at this time, but it appears that factors other than ¹O₂ are also important for the phototoxic mechanism against S. aureus. The substitution position of the BODIPY unit on the NHC ligand clearly plays a major role in determining the biological and photobiological activities against the cell lines investigated in this study. It is plausible that their structural differences affect membrane permeabilization and contact-dependent chemical reactions such as lipid photooxidation as demonstrated by Baptista and coworkers.⁸⁰ Confirming this would be a critical next step for rational design of subsequent-generation PSs of this class.

4. Conclusions

New neutral Ir(III) complexes bearing BODIPY-substituted NHC ligands are reported. Their photophysical characteristics were systematically investigated via spectroscopic methods and DFT calculations. All complexes exhibited BODIPY-localized intense ¹IL/¹MLCT absorption at 530–543 nm and ¹IL/³CT emission at 582–610 nm. However, the lowest triplet excited state of these complexes are the BODIPY−localized ³π,π* states. It was demonstrated that the position of the BODIPY pendant on the NHC ligand impacted both the ¹IL/¹MLCT absorption and ¹IL/³CT emission bands. Complexes Ir1 and Ir3, with the BODIPY-ethynyl motif attached at C4 of the phenyl ring of the NHC ligand, caused a red-shift of the ¹IL/¹MLCT absorption and ¹IL/³CT emission bands compared to those in Ir4 and Ir5 that have BODIPY-ethynyl at C5 of the benzimidazole unit of the NHC ligand. Meanwhile, both the lowest singlet excited state and the emitting states of Ir1 were lowered compared to those in Ir2, due to the extended π−conjugation induced by the ethynyl linker in Ir1. In contrast, replacing the methyl substituents on C3 of benzimidazole in complexes Ir1 and Ir4 by oligoether substituents in Ir3 and Ir5, respectively, did not impact the energies of the lowest singlet and emitting excited states in the corresponding complexes.

Ir3 and Ir5 were photobiologically active toward SKMEL28 melanoma cells with visible light activation, with Ir5 possessing a much larger PI and higher photocytotoxicity. The photobiological trends in cancer cells did not correlate with cell-free ¹O₂ quantum yields. The DCFDA assay for intracellular ROS detection argued against a correlation between phototoxicity and widespread redox imbalance, but did show that Ir5 was more effective at ROS production at the high PS concentrations. Ir5 was also photobiologically active toward S. aureus, while Ir3 was not. While ROS other than ¹O₂ may play a role in the phototoxic mechanism toward both cancer cells and bacterial cells, we hypothesize that factors such as cell membrane permeabilization and contact-mediated photooxidation reactions could be more important.

These studies indicate that the substitution position of BODIPY on the NHC ligand plays a profound role in the cytotoxicity and photocytotoxicity of this new class of complexes. C4-phenyl substitution of BODIPY compared to substitution at C5 of benzimidazole leads to lower ¹O₂ quantum yields but more effective phototoxic effects. Obviously, the impact of BODIPY substitution position at the NHC ligand is more pronounced on the photobiological activities than on the photophysical properties. Future studies are aimed at quantifying the interaction of the Ir3 and Ir5 with cells to better understand their stark differences in photocytotoxicity, which is a critical next-step toward establishing structure-activity relationships and expanding this new class of neutral Ir(III) NHC complexes.