Photoinduced Reactions of cis,trans,cis-[Pt^IV(N₃)₂(OH)₂(NH₃)₂] with 1-Methylimidazole

Abstract

The photodecomposition of cis,trans,cis-[Pt^IV(N₃)₂(OH)₂(NH₃)₂] in phosphate buffered saline (PBS), as well as in the presence of 1-methylimidazole (1-MeIm), induced by UVA light (centered at λ = 365 nm) has been studied by multinuclear NMR spectroscopy. We show that photoreduction, photoisomerisation and trans-labilization pathways are involved. The photodecomposition pathway in PBS which involves azide release, as detected by ¹⁴N NMR spectroscopy, appears to differ from that in acidic aqueous conditions, where N₂ is a product. A number of trans-{N-Pt^II-NH₃} species were also observed as photoproducts, as well as the release of free ammonia with a corresponding increase in pH. Oxygen was also detected as a product in solution. In the presence of 1-methylimidazole, surprisingly the major photoproduct was the tetra-substituted Pt^II complex [Pt^II(1-MeIm-N³)₄]²⁺ (structure confirmed by crystallography), even at a Pt:1-MeIm molar ratio of 1:1, together with cis- and trans-[Pt^II(NH₃)₂(1-MeIm-N³)₂]²⁺ as minor products. In these photoinduced 1-MeIm reactions, free ammonia, azide and oxygen were also detected. The results from this study illustrate that photoinduced reactions of platinum complexes can lead to novel reaction pathways, and therefore to new cytotoxic mechanisms in cancer cells.

Introduction

Photoactivation provides an interesting mechanism for reduction of Pt^IV to Pt^II. Pt^IV complexes can be made photolabile by choosing appropriate ligands, for example Pt^IV-azido complexes are well-known to be photoactive.[1,2] Such platinum complexes can then undergo a variety of photochemical reactions, including isomerization, substitution and reduction. A possible application of such processes is for therapeutic purposes and the development of light-sensitive Pt^IV pro-drugs that can be photoactivated (e.g. with laser light) to active antitumor agents directly at the site of the tumor.[3,4] Such site-selective activation would be expected to lead to a decrease in side-effects and toxicity, and an increase in the therapeutic index. Hence, the clinical usefulness of platinum antitumor agents might be expanded. Therefore, we have developed a series of complexes of the form [Pt^IV(N₃)₂(OH)₂(amine)₂], such as cis,trans,cis-[Pt^IV(N₃)₂(OH)₂(NH₃)₂] (1) which exhibit little or no dark toxicity, but bind to DNA upon photoactivation.[3,4]

Transcription mapping studies with treated plasmid DNA have shown that the platination sites resemble those from reaction with cisplatin (2) in dark conditions.[5] However, these complexes appear to kill cancer cells via a mechanism different from those of the classical platinum-based drugs.[6,7] In the photoreduction of Pt^IV-diazido precursors to their Pt^II counter-parts, a possible mechanism involves the release of azido ligands as N₃^• radicals. Azidyl radicals are very unstable and rapidly decompose in water into molecular nitrogen (N₂), thus preventing reoxidation of the platinum center (Scheme 1).[8]

Scheme 1.

Possible mechanism for the photoreduction of Pt^IV-diazido complexes.

We report here studies of the photodecomposition pathways in phosphate buffered saline (PBS) solution of complex 1 both on its own and in the presence of the nitrogen heterocycle 1-methylimidazole (1-MeIm). Imidazole fragments are an important part of purines, such as guanosine, xanthine, theophylline, 8-methylcaffeine, and histidyl residues in proteins. For example, in the enzyme carbonic anhydrase, the active site involves coordination of Zn^II to the imidazole ring nitrogen of histidine,[9] and in the serine proteases, the pH-dependent protonation-deprotonation of the imidazole ring of histidine plays an important role in the mechanism of peptide bond cleavage.[10] In addition to inter- and intrastrand crosslinks with DNA (involving especially binding to N⁷ of guanine bases), Pt^II complexes also form DNA-protein crosslinks involving histidine-, cysteine- and methionine-rich proteins.[11-13] Thus, 1-methylimidazole is a useful model for studies of photoinduced reactions of these Pt^IV-diazido complexes.

In particular, we have used here multinuclear NMR methods, including ¹⁴N NMR spectroscopy, to track photoreaction pathways involving the azido ligands. The elucidation of such photochemical pathways aids the understanding of the biological mechanism of action for this group of complexes.

Results

Photochemical behavior of complex 1 in PBS solution

First, we recorded a 1D ¹⁴N{¹H} NMR spectrum of the unlabelled complex 1 in PBS solution in dark conditions (Figure 1a). Peaks a (353.3 ppm), b (227.4 ppm) and c (160.5 ppm) are assigned to the terminal, central and Pt^IV-bound nitrogen atoms, respectively, of the coordinated azido ligands. The upfield peak d at −40.8 ppm (¹J(¹⁹⁵Pt-¹⁴N) = 180 Hz) is consistent with an ammine ligand bound to a Pt^IV metal center trans to a nitrogen donor ligand.[14] Upon irradiation of the sample with UVA light for 10 min (λ_max = 365 nm, P = 7.3 mW cm⁻² (P = power level), dose = 4.4 J cm⁻²) at 310 K, new ¹⁴N signals appeared (Figure 1b). Peaks a (353.4 ppm), b (227.5 ppm) and c (161.1 ppm) for the azido ligands of 1 were still present but with reduced intensities, and the new peaks a’ (76.9 ppm) and b’ (225.7 ppm) appeared which are assignable to the terminal and central nitrogen atoms of free azide (N₃⁻) in solution, in accordance with literature data.[15] New peaks d’ (−40.9 ppm) and d” (−50.8 ppm) are assignable to ammine groups coordinated to a Pt^IV center trans to nitrogen and oxygen donor ligands, respectively.[14] Intriguingly, no peaks for either N₂ (sharp peak expected at ca. 287 ppm[15]) or for Pt^II-NH₃ species (expected in the range −60 to −90 ppm)[14] were observed. Remarkably, similar results were obtained upon irradiation at 310 K with either UVA light for 2 h at a much lower power (λ_max = 365 nm, P = 0.6 mW cm⁻², dose = 4.5 J cm⁻²), or with UV-Visible light (λ = 250-700 nm) for 6 h (P = 30.0 mW cm⁻², dose = 648.0 J cm⁻²) (data not shown).

Figure 1.

1D ¹⁴N{¹H} NMR spectra of complex 1 (12.6 mm in 90% PBS/10% D₂O, Ar-saturated, initial pH 7.4) a) in the dark, b) after 10 min of irradiation (λ_max = 365 nm, P = 7.3 mW cm⁻², dose = 4.4 J cm⁻²) at 310 K, c) after 2 h of irradiation (λ_max = 365 nm, P = 7.3 mW cm⁻², dose = 56.6 J cm⁻², final pH 10.1) at 310 K. d) 1D ¹⁴N{¹H} NMR spectrum of complex ¹⁵N₂-1 (14.8 mm in 90% H₂O/10% D₂O, Ar-saturated, initial pH 7.4) after 2 h of irradiation (λ_max = 365 nm, P = 7.2 mW cm⁻², dose = 52.0 J cm⁻², final pH 9.9) at 310 K.

Assignments: a, b, c, coordinated azido ligand (δ = 353.3 (Pt^IV-NNN), 227.4 (Pt^IV-NNN), 160.5 ppm (Pt^IV-NNN)) as labeled in structure; a’, b’, free azide (N₃⁻; δ = 76.9 (NNN), 225.7 ppm (NNN)); d, chemical shift consistent with t-{N-Pt^IV-NH₃} species (δ = −40.8 ppm); d’, d”, chemical shifts consistent with t-{N/O-Pt^IV-NH₃} species (δ = −40.9 ppm, −50.8 ppm); e’, free ammonia (NH₃; δ = −20.9 ppm); f’, chemical shift consistent with t-{N/O-Pt^II-NH₃} species (δ = −69.6 ppm).

When UVA irradiation was prolonged for 2 h, the final pH of the solution increased to 10.1 (despite the initial attempt to buffer the solution with 10 mm phosphate), and no ¹⁴N peaks assignable to either coordinated azides or N₂ were detected. The two new sharp signals at 77.0 ppm (a’) and 225.6 ppm (b’) (Figure 1c) are due to the terminal and central nitrogen atoms, respectively, of free N₃⁻ in solution, whereas the chemical shift of peak f’ (−69.6 ppm) is consistent with an NH₃-Pt^II fragment trans to N- and/or O-donor ligands (broad, perhaps due to overlapping peaks).[14] Finally, new peak e’ (−20.9 ppm) is due to the presence of free ammonia in solution, consistent with the increase of pH upon irradiation. Some precipitate formed during the photoreaction and is likely to arise from polymeric hydroxo-/oxo-bridged Pt species, which arise partly as a result of ammonia release, hydrolysis and deprotonation of aqua/hydroxo ligands.[16]

A similar experiment was carried out using the ¹⁵NH₃-labelled analogue of complex 1 (¹⁵N₂-1). The 1D ¹⁴N{¹H} NMR spectrum of ¹⁵N₂-1 (14.8 mm in 90% PBS/10% D₂O, Ar-saturated, initial pH 7.4) obtained after 2 h of irradiation (λ_max = 365 nm, P = 7.2 mW cm⁻², dose = 52.0 J cm⁻²) at 310 K is shown in Figure 1d. Again, the pH increased (to 9.9), insoluble precipitate formed during the photoreaction, and the two new peaks (a’ and b’) due to free azide in solution were detected. As expected, no ¹⁴N NMR signals assignable to Pt-NH₃ moieties were detected, as the coordinated ammine groups were ¹⁵N-labelled. After performing ¹⁴N NMR experiments, the pH of the solution was re-adjusted to 3.1 with dilute HClO₄ to ensure slow exchange of NH protons, and 1D ¹H and 2D [¹H,¹⁵N] HSQC NMR experiments were carried out. Only one major Pt^II-ammine species was detected at δ(¹H/¹⁵N) = 4.05/−68.7 ppm (¹J(¹⁵N-H) = 72 Hz, ¹J(¹⁹⁵Pt-¹⁵N) = 282 Hz, ²J(¹⁹⁵Pt-H) = 56 Hz), which is consistent with a trans-{Pt^II(NH₃)₂} fragment.[17,18] This suggests that photoisomerization can accompany the main photoreduction process, and no other Pt^II or Pt^IV species were observed. In addition, a new peak at δ(¹H/¹⁵N) = 7.12/−1.90 ppm was detected, assignable to free ¹⁵NH₄⁺ in solution, which is consistent with the release of some ¹⁵N-labelled ammine ligands upon irradiation. The possible formation of Pt-phosphate adducts was ruled out by performing 1D ³¹P{¹H} NMR experiments, in which only one ³¹P signal was observed at 1.2 ppm due to the free buffer itself. Remarkably, very similar results were obtained when the photoreactions were carried out in phosphate buffer in the absence of Cl⁻ ions, thus ruling out the formation of Pt-chlorido derivatives and the possible role of chlorides in the photoredox process (data not shown).

Evolution of oxygen upon UVA irradiation of complex 1

In order to explore other potential photoreduction pathways, we investigated the possible evolution of gaseous dioxygen after irradiation by means of a Po₂ electrode. A 10.56 mm solution of complex 1 in PBS solution (initial pH 7.4) was Ar-saturated for 30 min and irradiated for 2 h (λ_max = 365 nm, P = 6.3 mW cm⁻², dose = 45.6 J cm⁻²) at 310 K. After irradiation, the pH increased to 10.0 and a Po₂ value of 64 mmHg was recorded.

Photochemical behavior of complex 1 in PBS solution in presence of 1-methylimidazole

First we studied the pH dependence of the ¹⁴N NMR shifts of 1-methylimidazole (1-MeIm). The imidazole ring is a particularly important constituent of biological systems (e.g. a histidyl side chain in proteins) primarily because of its acid-base equilibrium with a unique pK_a value (7.1) near to neutrality in aqueous solution.[19] This gives the imidazole ring the capability, in many biological fluids, of acting as a proton donor or acceptor and, perhaps most importantly, as a ligand which can coordinate to metal ions.

¹⁴N NMR shifts for N¹ and N³ of 1-methylimidazole in aqueous solution as a function of pH are plotted in Figure 2a. As expected, protonation of 1-methylimidazole at N³ produces a large shift of the N³ resonance (ca. 71 ppm upfield) and a comparatively small shift of N¹ resonance (ca. 7 ppm downfield) relative to the deprotonated species. The pK_a of 7.15 ± 0.06 is in good agreement with the reported value of 7.2.[19] At low pH, ¹⁴N¹ and ¹⁴N³ resonances essentially superimpose, as well as the ¹H NMR peaks of H⁴ and H⁵ protons (Figure 2b), in agreement with reported data.[19]

Figure 2.

a) pH dependence of ¹⁴N NMR shifts of N¹ and N³ of 1-methylimidazole. The pK_a value was determined as described in the Experimental Section. b) Representative 1D ¹H and ¹⁴N{¹H} NMR spectra of 1-methylimidazole (62.7 mm in 90% H₂O/10% D₂O) at pH 2.5, 7.6 and 9.8.

The interaction between the ¹⁵N-labelled analogue of complex 1 (¹⁵N₂-1) and 1-methylimidazole in PBS solution was investigated. A 8.6 mm solution of ¹⁵N₂-1 in 90% PBS/10% D₂O was prepared and 1-methylimidazole added (molar ratio = 1:1.02) at a starting pH of 8.7 (no adjustment). After degassing the solution with Ar for 30 min, the sample was irradiated for 2 h (λ_max = 365 nm, P = 6.4 mW cm⁻², dose = 46.1 J cm⁻²) at 310 K, and the photoproducts were characterized by NMR spectroscopy. The solution, initially pale yellow, turned colorless after irradiation, and no precipitate formed. As already observed after irradiating complex ¹⁵N₂-1 on its own, the pH increased to 10.2. In the 1D ¹⁴N{¹H} NMR spectrum, several signals were detected. The two new sharp signals at 77.0 ppm (a’) and 225.7 ppm (b’) (Figure 3) are due to the terminal and central nitrogen atoms of free N₃⁻ in solution, respectively, whereas new peaks g (221.1 ppm) and h (141.9) are assigned to the N³ and N¹ atoms of free 1-methylimidazole. The new broad signal i in the range 150-123 ppm is assignable to the N³ and N¹ atoms of Pt-coordinated 1-methylimidazole. As previously observed, no peaks assignable to either coordinated azides or N₂ were detected. On the contrary, evolution of gaseous dioxygen was again detected potentiometrically soon after irradiation, giving a Po₂ value of 50 mmHg.

Figure 3.

1D ¹⁴N{¹H} NMR spectrum of complex ¹⁵N₂-1 (8.6 mm) in presence of 1-methylimidazole (molar ratio 1:1.02) in 90% PBS/10% D₂O (Ar-saturated, initial pH 8.7) after 2 h of irradiation (λ_max = 365 nm, P = 6.4 mW cm⁻², dose = 46.1 J cm⁻², final pH 10.2) at 310 K.

Assignments: a’, b’, free azide (N₃⁻) in solution (δ = 77.0 (NNN,), 225.7 ppm (NNN,)); g (N³), h (N¹), free 1-methylimidazole in solution (δ = 221.1, 141.9 ppm); i, broad signal indicating N³ and N¹ atoms of Pt-coordinated 1-methylimidazole (δ = 150-123 ppm).

Six different 1-methylimidazole-containing species were detected with the aid of 1D ¹H (Figure 4), 2D [¹H,¹H] NOESY (Figure 5), [¹H,¹³C] HSQC, and [¹H,¹⁵N] HMBC (Figure 6) NMR spectroscopy. Their chemical shifts are summarized in Table 1. The major species (γ) can be identified as [Pt^II(1-MeIm-N³)₄]²⁺, supported by the X-ray structure of crystals obtained under similar experimental conditions (see Supporting Information). The crystal structure of this complex has been reported previously by Clement and co-workers (CCDC ref. code: HIGVUO01).[20] The presence of ClO₄⁻ as the counter-ion arises from the use of perchloric acid for pH adjustment. This square-planar complex contains Pt^II coordinated to four 1-methylimidazole ligands via N³. The crystallographic data (Space group P −1, a = 8.07470(10) Å, b = 9.12370(10) Å, c = 9.3090(2) Å, α = 71.9540(10)°, β = 65.9410(10)°, γ = 83.3230(10)°) are in close agreement with the published data. Species α and β can be assigned to trans- and cis-[Pt^II(NH₃)₂(1-MeIm-N³)₂]²⁺, respectively,[21-23] whereas the minor species δ and ε have not been identified. For comparison purposes, each of the three identified Pt^II species (α, β and γ) were synthesized individually, according to literature procedures,[24] and their NMR spectra (Figures S1-S4) compared to the peaks in the spectrum of the irradiated solution. The presence of all three species was confirmed via this method (see Table 1 for chemical shift assignments). Surprisingly, unreacted 1-methylimidazole (ζ) was detected as a major component of the investigated solution, even though the reactants were mixed in a 1:1 molar ratio. The large upfield shift of the ¹⁵N³ resonance (ca. 90 ppm, compared to δ(¹⁵N³) of free 1-methylimidazole at 221.08 ppm) observed for all the new 1-methylimidazole species is consistent with coordination of 1-methylimidazole to platinum taking place via the N³ atom.[25] The possible coordination of phosphates from the buffer was also ruled out, as in the 1D ³¹P{¹H} NMR spectrum only one peak was detected at 3.84 ppm, arising from the PBS solution itself.

Figure 4.

1D ¹H NMR spectrum of complex ¹⁵N₂-1 (8.6 mm) in presence of 1-methylimidazole (molar ratio 1:1.02) in 90% PBS/10% D₂O (Ar-saturated, initial pH 8.7) after 2 h of irradiation (λ_max = 365 nm, P = 6.4 mW cm⁻², dose = 46.1 J cm⁻², final pH 10.2) at 310 K. See Table 1 for species assignment.

Figure 5.

Part of the 2D [¹H,¹H] NOESY NMR spectrum of complex ¹⁵N₂-1 (8.6 mm) in presence of 1-methylimidazole (molar ratio 1:1.02) in 90% PBS/10% D₂O (Ar-saturated, initial pH 8.7) after 2 h of irradiation (λ_max = 365 nm, P = 6.4 mW cm⁻², dose = 46.1 J cm⁻², final pH 10.2) at 310 K. See Table 1 for species assignment.

Figure 6.

2D [¹H,¹⁵N] HMBC NMR spectrum of complex ¹⁵N₂-1 (8.6 mm) in presence of 1-methylimidazole (molar ratio 1:1.02) in 90% PBS/10% D₂O (Ar-saturated, initial pH 8.7) after 2 h of irradiation (λ_max = 365 nm, P = 6.4 mW cm⁻², dose = 46.1 J cm⁻², final pH 10.2) at 310 K. See Table 1 for species assignment.

Table 1.

¹H, ¹³C and ¹⁵N chemical shifts of the peaks produced after 2 h of irradiation (λ_max = 365 nm, P = 6.4 mW cm⁻², dose = 46.1 J cm⁻², 310 K) of complex ¹⁵N₂-1 (8.6 mm) in presence of 1-methylimidazole (molar ratio 1:1.02) in 90% PBS/10% D₂O (Ar-saturated, pH 8.7 (initial)/10.2 (final)), and of the individually synthesized Pt^II-(1-MeIm) complexes (see Figures 4, 5, 6 and Supporting Information)

Detected Species	Chemical shift [ppm]
	δ(CH3/CH3)	δ(C2H/C2H)	δ(C4H/C4H)	δ(C5H/C5H)	δ(N1)	δ(N3)
α(t-[PtII(NH3)2(1-MeIm-N3)2]2+)	3.78/35.02	7.99/140.01	7.14/128.94	7.23/123.47	146.42	127.48
β(c-[PtII(NH3)2(1-MeIm-N3)2]2+)	3.76/34.96	7.83/139.53	7.02/127.57	7.16/122.04	144.86	n.d.
γ([Pt(1-MeIm-N3)4]2+)	3.73/35.04	7.76/140.08	6.85/128.87	7.15/123.13	146.03	131.09
δ	3.68/34.89	7.71/140.02	6.87/129.56	7.08/122.92	n.d.	n.d.
ε	3.67/34.99	7.64/139.23	6.76/128.32	7.07/122.98	145.35	133.82
ζ(1-MeIm)	3.72/33.73	7.63/139.40	7.01/128.26	7.12/122.10	141.74	221.08
Synthesized Complexes*
t-[PtII(NH3)2(1-MeIm-N3)2]2+	3.82/35.03	8.05/139.94	7.19/129.00	7.26/123.51	-	-
c-[PtII(NH3)2(1-MeIm-N3)2]2+	3.73/34.96	7.78/139.55	7.04/128.55	7.16/123.19	-	-
[Pt(1-MeIm-N3)4]2+	3.73/35.04	7.77/139.93	6.86/128.81	7.16/123.05	-	-

^*In D₂O.

Once the pH of the investigated sample was re-adjusted to 2.6, 1D ¹H and 2D [¹H,¹⁵N] HSQC NMR experiments were carried out. Based on chemical shifts (Table 2),[14] five peaks assignable to Pt^II species were detected, all consistent with coordinated NH₃ groups trans to N-donor ligands. In agreement with results obtained by irradiating complex ¹⁵N₂-1 on its own under the same experimental conditions, there was no evidence of Pt^IV-ammine species, and free ¹⁵NH₄⁺ in solution was detected (δ(¹H/¹⁵N) = 7.12/−1.90 ppm), due to the release of ammine ligands upon irradiation.

Table 2.

¹H and ¹⁵N chemical shifts and assignments of the Pt-NH₃ peaks produced after 2 h of irradiation (λ_max = 365 nm, P = 6.4 mW cm⁻², dose = 46.1 J cm⁻², 310 K) of complex ¹⁵N₂-1 (8.6 mm) in presence of 1-methylimidazole (molar ratio 1:1.02) in 90% PBS/10% D₂O (Ar-saturated, pH 8.7 (initial)/10.2 (final; re-adjusted to 2.6)). Data obtained from a 2D [¹H,¹⁵N] HSQC NMR experiment.

Chemical shifts [ppm]		Assignment
δ(1H)	δ(15N)
4.23	−67.00	t-{N-PtII-15NH3}
4.05	−65.05	t-{N-PtII-15NH3}
4.04	−65.12	t-{N-PtII-15NH3}
4.25	−63.45	t-{N-PtII-15NH3}
4.04	−61.46	t-{N-PtII-15NH3}
7.12	−1.90	15NH4+

Discussion

The NMR spectroscopic data obtained during photoactivation of complex 1 allow reaction pathways to be proposed (Scheme 2). According to the literature, in addition to photoredox processes, azido complexes can undergo photosubstitution reactions in which the azido ligand is substituted, e.g. by a solvent molecule.[8] This is in agreement with the ¹⁴N spectrum of 1 in PBS recorded after only 10 min of irradiation with UVA light (Figure 1b), in which the detected ¹⁴N NMR signals are consistent with the presence of Pt^IV species of the type cis,trans,cis-[Pt^IV(N₃)(H₂O/OH)(OH)₂(NH₃)₂] and free N₃⁻. Thus, the first step of such photoreactions under these conditions appears to be the release of at least one azido ligand, followed by its replacement with a water molecule which deprotonates on coordination due to the increase in pH of the solution. However, we cannot rule out the simultaneous and/or subsequent release of, at least, one ammine ligand before reduction of Pt^IV to Pt^II takes place. A reported study of [M(N₃)(NH₃)₅]²⁺, where M = Cr or Co,[26-28] found that the first photochemical step is the formation of trans-[M(N₃)(H₂O)(NH₃)₄]²⁺ caused by the photolabilization of the ammine ligand trans to N₃⁻ due to a strong trans-effect of the azido ligand. Thus, depending on the experimental conditions, the generated aqua complexes may undergo subsequent redox (M = Co) or substitution (M = Cr) reactions. Also, it has been reported recently that the photoreduction of some Pt^IV-dichlorotetrammine complexes is accompanied by the release of ammonia.[29] In our case, the loss of ammine ligands has been widely confirmed by NMR spectroscopy, thus explaining the formation of insoluble hydroxo-/oxo-bridged polymerized species and the increase of basicity. It is surprising that although photoreduction of the starting Pt^IV-diazido complexes in PBS solution occurs, the azido ligands are not oxidized to N₂. This behavior differs from that observed previously in which irradiation of complex 1 in acidic aqueous solution led to several Pt^II species together with the evolution of gaseous dinitrogen via the formation of nitrene intermediates.[30] This emphasizes that the observed photoreaction pathways are highly dependent on the solution conditions.

Scheme 2.

Possible pathways for photoreactions of complex 1 in PBS. Dashed arrows refer to reactions in presence of 1-methylimidazole, and curly brackets are used to denote transient intermediates. The photodecomposition pathways of 1 in acidic aqueous solution have been reported previously;[30] they involve not only photoreduction to Pt^II and O₂ evolution, but N₂ release via the formation of nitrene intermediates, in contrast to N₃⁻ release as detected in the present work.

After replacement of both azido ligands, photoreduction may occur by one-electron transfers from each of two hydroxo groups coordinated to the Pt^IV center leading to the corresponding Pt^II species together with the formation of hydroxyl radicals. When generated at high local concentrations, hydroxyl radicals readily dimerize to H₂O₂.[31] On irradiation with UV light, hydrogen peroxide is known to decompose by disproportionation to water and gaseous dioxygen (O₂), with a maximum rate at pH = 11.6.[32] After each irradiation experiment, the solution was checked for hydrogen peroxide content by means of peroxide test sticks, but no H₂O₂ was detected. However, since in our system pH increases to ca. 10.1 on irradiation, the decomposition of hydrogen peroxide is expected to be favored. Although the evolution of O₂ was only measured qualitatively, the experimental evidence does support the hypothesis of the involvement of reactive oxygen-containing species (as shown in Scheme 2), which are subsequently oxidized to O₂ while Pt^IV complexes are reduced to the corresponding Pt^II photoproducts. The detection of O₂ as a photoproduct suggests that a variety of photodecomposition pathways in additions to that shown in Scheme 1 can occur, depending on the actual experimental conditions.

Finally, it is worth noting that on irradiation of complex 1 and its ¹⁵N-labelled analogue ¹⁵N₂-1, the only detected Pt^II photoproduct has NMR parameters consistent with a trans-diammine species. This could result from photoisomerization processes which lead to the formation of the more thermodynamically stable trans isomer. On absorption of a photon, some complexes may undergo rearrangement of their structure. Photoisomerization of Pt^IV-am(m)ine compounds has been previously observed;[29,33,34] in addition, a number of photoisomerizations of Pt^II complexes, including cis-[Pt^II(N₃)₂(PPh₃)₂] and cis-[PtCl₂(N-amp)₂] (amp = diethyl aminomethylphosphonate), are known to occur from an excited state by a twisting mechanism, i.e. without bond-breaking and bond-forming.[35,36] Our experimental evidence does not show whether cis to trans conversion occurs before or after reduction takes place.

It is intriguing that when irradiation studies of complex 1 were carried out in the presence of 1 molar equivalent of 1-methylimidazole, unreacted 1-methylimidazole was still present in the final solution. If the first step of the investigated photoreaction involves the replacement of the azido ligands with water molecules, then the subsequent photoreduction process would lead to the formation of Pt^II-aqua species. Given that in our system the pH increases to 10.2 on irradiation, coordinated H₂O molecules are expected to be readily deprotonated, thus leading to Pt^II species in which replacement of hydroxo ligands by free imidazole is unlikely to occur.[18,37] In fact, some of the peaks detected in the 2D [¹H,¹⁵N] HSQC NMR spectrum might be assignable to Pt^II-hydroxo species containing trans-{Pt^II(NH₃)₂} fragments, thus accounting for the lack of coordination by the free 1-methylimidazole. The fact that a major photoproduct is [Pt^II(1-MeIm-N³)₄]²⁺, confirms the high affinity of complex 1 toward imidazole rings upon irradiation, and suggests that coordination by 1-methylimidazole and consequent replacement of the NH₃ ligands is favored as long as the pH of the solution is low enough to give rise to coordinated aqua ligands which are easily substituted. However, our experimental evidence does not indicate whether coordination of 1-methylimidazole to platinum occurs before or after reduction takes place.

Conclusions

Pt^IV-diazido complexes show promise as anticancer agents. They are relatively non-toxic in the dark and can be photoactivated by light to produce new species which are cytotoxic to cancer cells by a novel mechanism of action.[6,7] This may involve the formation of unusual platinated lesions on DNA,[5,38] and initial work has shown that the rapid reactions with guanine derivatives can occur often with seemingly simple pathways in which Pt^II-guanine adducts are accompanied by the release of N₂. However, it is also apparent that the reaction pathways which accompany these photoactivated processes are dependent on the solution conditions (e.g. pH, buffer); for example, nitrene intermediates can be trapped in the presence of dimethylsulfide.[30] In the present work, we have shown that photoinduced release of azide from cis,trans,cis-[Pt^IV(N₃)₂(OH)₂(NH₃)₂] is facile in phosphate buffered saline solution, and that photoinduced isomerization can occur.

Our investigation of these reaction pathways has benefited from the use of ¹⁴N NMR spectroscopy (99.63% natural abundance, receptivity 1.0×10⁻³ relative to ¹H). There are relatively few reports of the use of this technique in the study of platinum drugs in the literature.[39] The major drawback is ¹⁴N quadrupolar relaxation (¹⁴N nuclear spin (I) = 1), which dominates when the environment of ¹⁴N has a low symmetry and can lead to very broad lines. More commonly, ¹⁵N NMR (I = ½) has been employed in the study of platinum anticancer drugs, despite the low receptivity (3.85×10⁻⁶ relative to ¹H), as for most compounds of interest ¹⁵N-labelled derivatives can be prepared from readily available starting materials.[14,37] A combination of these two NMR techniques has provided us with new insights into the mechanism of photoactivation and photodecomposition pathways.

Our work on photoinduced reactions of cis,trans,cis-[Pt^IV(N₃)₂(OH)₂(NH₃)₂] in the presence of 1-methylimidazole suggests that histidyl side chains in proteins could be a target for attack in cells. Intriguingly, under the conditions of our experiments, ammonia release was readily observed together with O₂ production, as well as azide displacement. The formation of the tetrakis adduct even at low Pt:1-methylimidazole ratios (1:1) was surprising and illustrates further that photoactivation can lead to unexpected reaction pathways. This is encouraging for the discovery of anticancer drugs with novel mechanisms of action, drugs which may not be cross-resistant to existing chemical agents.

Experimental Section

Materials

K₂[Pt^IICl₄] was purchased from Precious Metals Online; KI, NaN₃, KOH, and NaCl from Fisher Scientific; AgNO₃, NH₄Cl, ¹⁵NH₄Cl, 1-methylimidazole, phosphate buffered saline tablets (1 tablet in 200 mL of water: 0.01 m phosphate pH 7.4, 0.138 m NaCl, 0.0027 m KCl), 1,4-dioxane, and D₂O from Sigma-Aldrich; H₂O₂ (30%) from Polabo; and HClO₄ from Fisons.

Warning! Heavy metal azido complexes are known to be shock-sensitive detonators. We encountered no problems in this study, but these materials should be handled with extreme caution, especially not to put pressure on them in the crystalline form.

Syntheses

Complex cis,trans,cis-[Pt^IV(N₃)₂(OH)₂(NH₃)₂] (1) and the corresponding ¹⁵N-ammine labeled analogue cis,trans,cis-[Pt^IV(N₃)₂(OH)₂(¹⁵NH₃)₂] (¹⁵N₂-1) were synthesized as previously described.[16,38] For ¹⁵N₂-1 ¹⁵NH₄Cl was used as the source of ¹⁵N. Yield: 78-80%. ¹H NMR (500.1 MHz, 90% H₂O/10% D₂O, pH 5.4, TSP): δ = 5.12 ppm (¹J(¹⁵N-H) = 73 Hz, ²J(¹⁹⁵Pt-H) = 46 Hz; Pt^IV-¹⁵NH₃); 2D [¹H,¹⁵N] HSQC NMR (500.1/50.7 MHz, 90% H₂O/10% D₂O, pH 5.4, ¹⁵NH₄Cl): δ = −40.8 ppm (¹J(¹⁹⁵Pt-¹⁵N) = 261 Hz; Pt^IV-¹⁵NH₃); ¹⁴N{¹H} NMR (36.1 MHz, 90% PBS/10% D₂O, pH 7.4, NH₄Cl): δ = 353.3 (Pt^IV-NNN), 227.4 (Pt^IV-NNN), 160.5 (Pt^IV-NNN), −40.8 ppm (¹J(¹⁹⁵Pt-¹⁴N) = 180 Hz; Pt^IV-NH₃).

Complexes cis-[Pt^II(NH₃)₂(1-MeIm-N³)₂]Cl₂, trans-[Pt^II(NH₃)₂(1-MeIm-N³)₂]Cl₂ and [Pt^II(1-MeIm-N³)₄]Cl₂ were synthesized according to literature procedures.[24] See Table 1 and Supporting Information for NMR characterization.

NMR spectroscopy

All NMR spectra were acquired in 90% PBS/10% D₂O (unless otherwise stated) at 298 K on a Bruker DPX300 spectrometer with a QNP [¹H,¹³C,¹⁹F,³¹P] probe head, a Bruker DPX400 spectrometer using a DUL [¹H,¹³C] probe-head, a Bruker DMX500 spectrometer using a TBI [¹H,¹³C,X] probe-head and a Bruker Avance 600 spectrometer using a TXI [¹H,¹³C,¹⁵N] cryoprobe-head, all equipped with z-field gradients. Data processing was carried out using XWIN-NMR version 3.5 (DMX500 and Avance 600) Bruker U.K. Ltd.) and MestReC version 4.9.9.9 (Mestrelab Research S.L.).

Typical acquisition parameters for 1D ¹H NMR spectra (¹H: 400.12, 500.13 or 599.92 MHz): 16 transients, spectral width 7.5 kHz, using 32k data points and a delay time of 2.0 s. Water suppression was achieved using a 55.0 dB power level presaturation. Spectra were processed using exponential weighting with a resolution of 0.5 Hz and a line broadening threshold of 0.1 Hz.

Typical acquisition parameters for 1D ¹⁴N{¹H} NMR spectra (¹⁴N: 36.13 MHz): 64k transients, spectral width 20 kHz, using 32k data points and a delay time of 0.25 s. Sequences were optimized for ¹J(¹⁴N-H) = 60 Hz, and ¹H decoupling was achieved using WALTZ16 pulse sequence. Spectra were processed using exponential weighting with a resolution of 0.6 Hz and a line broadening threshold of 10 Hz.

Typical acquisition parameters for 1D ¹³C{¹H} PENDANT NMR spectra (¹³C: 100.60 MHz): 512 transients, spectral width 25 kHz, using 64k data points and a delay time of 2 s. Sequences were optimized for 1J(¹³C-H) = 145 Hz, and ¹H decoupling was achieved using WALTZ16 pulse sequence. Spectra were processed using exponential weighting with a resolution of 2 Hz, and a line broadening threshold of 2.8 Hz.

Typical acquisition parameters for 1D ³¹P{¹H} NMR spectra (³¹P: 202.46 MHz): 512 transients, spectral width 20 kHz, using 32k data points and a delay time of 1.5 s. ¹H decoupling was achieved using WALTZ16 pulse sequence. Spectra were processed using exponential weighting with a resolution of 0.6 Hz and a line broadening threshold of 10 Hz.

Typical acquisition parameters for 2D [¹H,¹H] NOESY NMR spectra (¹H: 500.13 MHz): 1k transients of 32 scans/block, spectral width 5.0/5.0 kHz, using 1k/1k data points and a delay time of 1.5 s. Spectra were processed using sine weighting with a resolution of 4.8/9.8 Hz and a line broadening threshold of 0.8/1.5 Hz.

Typical acquisition parameters for 2D [¹H,¹³C] HSQC NMR spectra (¹H: 500.13/¹³C: 125.77 MHz): 2k transients of 64 scans/block, spectral width 5.4/17.6 kHz, using 2k/2k data points and a delay time of 1.7 s. Sequences were optimized for ¹J(¹³C-H) = 145 Hz, and ¹H decoupling was achieved using WALTZ16 pulse sequence. Spectra were processed using cosine-squared weighting with a resolution of 2.7/34.4 Hz and a line broadening threshold of 0.7/2.5 Hz. ¹H chemical shifts were referenced to TSP via internal 1,4-dioxane at 3.764 ppm.

Typical acquisition parameters for 2D [¹H,¹⁵N] HSQC NMR spectra (¹H: 500.13/¹⁵N: 50.70 MHz): 512 transients of 16 scans/block, spectral width 7.5/4.6 kHz, using 2k/2k data points and a delay time of 1.5 s. Sequences were optimized for ¹J(¹⁵N-H) = 76 Hz, and ¹H decoupling was achieved using WALTZ16 pulse sequence. Spectra were processed using cosine-square weighting with a resolution of 2.2/3.7 Hz and a line broadening threshold of 0.7/1.0 Hz.

Typical acquisition parameters for 2D [¹H,¹⁵N] HMBC NMR spectra (¹H = 599.92/¹⁵N: 60.79 MHz): 512 transients of 32 scans/block, spectral width 6.0/6.1 kHz, using 2k/2k data points and a delay time of 1.5 s. Sequences were optimized for ¹J(¹⁵N-H) = 76 Hz/ⁿJ(¹⁵N-H) = 6 Hz, and ¹H decoupling was achieved using WALTZ16 pulse sequence. Water suppression was achieved using a 47.0 dB power level presaturation. Spectra were processed using cosine-square weighting with a resolution of 2.9/6.2 Hz and a line broadening threshold of 0.7/1.0 Hz.

¹H chemical shifts were referenced to TSP (sodium 3-(trimethylsilyl)-[2,2,3,3-D₄]-proprionate) via internal 1,4-dioxane at 3.764 ppm, ¹³C to TMS (tetramethylsilane) via internal 1,4-dioxane at 67.19 ppm, ¹⁴N and ¹⁵N to external ¹⁵NH₄Cl 1.5 m in 1 m HCl at 0.00 ppm and ³¹P to external H₃PO₄ 85% in D₂O at 0.00 ppm.

pH measurement

pH values were measured at ambient temperature directly in the NMR tube with a Corning 145 pH-meter equipped with an Aldrich microcombination electrode, calibrated with standard Aldrich buffer solutions at pH 4, 7 and 10.

pK_a determination

The experimental pH titration data were fitted to the following formula derived from the Henderson-Hasselbalch equation:

$$\delta =\frac{{\delta }_A\cdot {10}^{-\mathit{pH}}+{\delta }_B\cdot {10}^{-{\mathit{pK}}_a}}{{10}^{-\mathit{pH}}+{10}^{-{\mathit{pK}}_a}}$$

where K_a is the dissociation constant for species 1-MeImH⁺, δ_A, and δ_B are the limiting chemical shifts of species 1-MeImH⁺ and 1-MeIm, respectively. The pH titration curves (¹⁴N chemical shifts of both N¹ and N³ vs. pH) were fitted using the program ORIGIN version 6.0 (Microcal Software Ltd.).

Photoreactions

Samples for the photochemical studies were 8-15 mm solutions of the investigated Pt^IV complexes (on their own or containing 1-methylimidazole in 1:1 molar ratio) in 0.6 mL 90% PBS/10% D₂O. Before irradiation, the samples were transferred into 5 mm screw-cap NMR tubes (Wilmad) provided with a teflon/silicone septum, and Ar-saturated for 30 min. NMR samples were then irradiated for 2 h (unless otherwise stated) at 310 K using a LZC-ICH2 photoreactor (Luzchem Research Inc.) equipped with a temperature controller and 16 LZC-UVA lamps (Hitachi, λ_max = 365 nm) or 16 LZC-VIS UV-Visible light lamps (Sylvania cool white, λ = 250-700 nm) with no other sources of light filtration. The spectral outputs for both sources are shown in the Supporting Information (Figure S5). The power levels (P = 6.4-7.2 mW cm⁻² (UVA), P = 30.0 mW cm⁻² (UV-Visible)) were monitored and assessed using the appropriate probe window, calibrated against an OAI-306 UV power meter from Optical Associates, Inc.; the delivered radiation dose (J cm⁻²) is defined as irradiance (P, W cm⁻²) × time (s). NMR spectra were then recorded soon after irradiation.

Po₂ measurement

Po₂ values were measured with a Corning 240 pH-meter (switched to read millivolts) equipped with a Lazar DO-166-NP dissolved oxygen needle probe. Calibration was carried out at 295 K and a barometric pressure of 761 mmHg using a 0.1 g L⁻¹ NaCl solution in double-distilled water (made freshly) giving at equilibrium with atmospheric dioxygen a Po₂ value of 161 mmHg (theoretical: 163 mmHg). The electrode was then set up to read Po₂ directly in mm of Hg partial pressure.

Supplementary Material

01

Figure S1. a) 1D ¹H and b) 1D ¹³C{¹H} PENDANT NMR spectra of trans-[Pt(NH₃)₂(1-MeIm)₂]Cl₂ in D₂O. ¹H chemical shifts were referenced to TSP (sodium 3-(trimethylsilyl)-[2,2,3,3-D₄]-proprionate) via internal 1,4-dioxane (*) at 3.764 ppm, and ¹³C chemical shifts to TMS (tetramethylsilane) via internal 1,4-dioxane (*) at 67.19 ppm.

Assignments: δ(¹H/¹³C) =3.82/35.03 (CH₃), 7.19/129.00 (CH⁴), 7.26/123.51 (CH⁵), 8.05/139.94 ppm (CH²).

Figure S2. a) 1D ¹H and b) 1D ¹³C{¹H} PENDANT NMR spectra of cis-[Pt(NH₃)₂(1-MeIm)₂]Cl₂ in D₂O. ¹H chemical shifts were referenced to TSP (sodium 3-(trimethylsilyl)-[2,2,3,3-D₄]-proprionate) via internal 1,4-dioxane (*) at 3.764 ppm, and ¹³C chemical shifts to TMS (tetramethylsilane) via internal 1,4-dioxane (*) at 67.19 ppm.

Assignments: δ(¹H/¹³C) =3.73/34.96 (CH₃), 7.04/128.55 (CH⁴), 7.16/123.19 (CH⁵), 7.78/139.55 ppm (CH²).

Figure S3. a) 1D ¹H and b) 1D ¹³C{¹H} PENDANT NMR spectra of [Pt(1-MeIm)₄]Cl₂ in D₂O. ¹H chemical shifts were referenced to TSP (sodium 3-(trimethylsilyl)-[2,2,3,3-D₄]-proprionate) via internal 1,4-dioxane (*) at 3.764 ppm, and ¹³C chemical shifts to TMS (tetramethylsilane) via internal 1,4-dioxane (*) at 67.19 ppm.

Assignments: δ(¹H/¹³C) =3.73/35.04 (CH₃), 6.86/128.81 (CH⁴), 7.16/123.05 (CH⁵), 7.77/139.93 ppm (CH²).

Figure S4. Ortep diagram and numbering scheme for [Pt(1-MeIm)₄](ClO₄)₂ (50% probability ellipsoids). The hydrogens have been omitted for clarity. This structure determined by S. Parsons and S. Moggach is the same as that reported previously (O. Clement et al., J. Am. Chem. Soc. 1996, 118, 612–620) and the reported atom labeling scheme is used above.

Figure S5. Spectral outputs of LZC-UVA (Hitachi) and LZC-VIS (Sylvania cool white) lamps. (Data taken from manufacturer’s manual (Luzchem Research Inc.))

Intro Structures.