A non-singlet oxygen mediated reaction photoinduced by phenalenone, a universal reference for singlet oxygen sensitization

Abstract

Photosensitized oxidations involve the generation of radicals (type I mechanism) and/or the production of singlet molecular oxygen (¹O₂) (type II mechanism). Phenalenone (PN) is one of the well-known type II photosensitizers and is widely used as a ¹O₂ reference sensitizer. In this work, we have studied the interaction of tryptophan (Trp) with the excited states of PN, the production of ¹O₂ in the presence of Trp, the formation of radicals and the kinetics of the photodegradation of Trp under different experimental conditions. In contrast to what is usually assumed, we have proven that PN is also able to act as a type I photosensitizer. Moreover, we have demonstrated that the predominant mechanism of the photosensitization of Trp by PN involves an electron transfer initiated-process.

---

Introduction

A photosensitized reaction is defined as a photochemical alteration occurring in one molecular entity as a result of initial absorption of radiation by another molecular entity called a photosensitizer.¹ The biological and medical importance of photosensitized reactions is mostly related to their participation in processes involved in the generation of skin cancer^2,3 and to several applications in disinfection^4,5 and photodynamic therapy (PDT) for the treatment of cancer and other pathologies.^6,7

The chemical changes in biological components resulting from photosensitized reactions can take place through different mechanisms. Energy transfer from the triplet state of the photosensitizer can generate excited states in the target molecule.⁸ Photosensitized oxidations also contribute to biological damage induced by UV radiation. These processes involve the generation of radicals (type I), e.g., via electron transfer or hydrogen abstraction, and/or the production of singlet molecular oxygen (O₂(¹Δ_g), denoted throughout as ¹O₂) (type II).⁹

Singlet oxygen is an important oxidizing intermediate in chemical processes and one of the main reactive oxygen species responsible for the damaging effects of light on biological systems (photodynamic effects).^10,11 Photosensitization is primarily responsible for the production of ¹O₂ in vivo.¹² In this process, ¹O₂ is most often produced by energy transfer from the excited triplet state of a sensitizer to dissolved molecular oxygen (O₂). Currently, it is accepted that the photosensitization of proteins occurs mainly through the reactions of ¹O₂ with tryptophan (Trp), tyrosine, histidine, methionine and cysteine side chains.¹³

Several compounds are used as model photosensitizers for the study of ¹O₂-mediated reactions. In this context, 1H-phenalen-1-one (1H-benzonaphthen-1-one or 7-perinaphthenone, denoted throughout as phenalenone (PN)), is a unique photosensitizer due to its quantum yield of ¹O₂ production (Φ_Δ, defined as the number of ¹O₂ molecules produced per absorbed quantum of light) close to unity.^14–16 In addition, PN is photostable in most solvents and has a very low ability to deactivate ¹O₂. These properties have made PN one of the well-known ¹O₂ photosensitizers.¹⁷ In fact, it is widely used as a reference for the determination of Φ_Δ values and for investigating oxidations of biological compounds by ¹O₂.

PN is an aromatic ketone whose chemical structure and absorption spectrum are shown in Fig. 1. This compound is present in the environment¹⁸ and some plants produce secondary metabolites containing its skeleton¹⁹ as a response to pathogen attacks. These phenalenone derivatives would act as ¹O₂-photosensitizers able to use solar energy for defense.²⁰ The PN photodynamic activity and that of its derivatives have been demonstrated early,^21,22 and it has recently been used as standard for singlet oxygen-mediated damage to DNA²³ and oxidation of amino acids²⁴ and proteins.²⁵

Fig. 1 Molecular structure of PN and Trp, and the corresponding absorption spectra in air-equilibrated aqueous solutions; solid line: PN; dashed-dotted lines: Trp.

A study using electron paramagnetic resonance (EPR) analysis revealed in the 1960's that radicals are formed when solutions of PN in 2-propanol were exposed to UV radiation (λ > 300 nm).²⁶ Later, it was demonstrated that, despite low quantum yields, significant photochemical decomposition of PN occurred under steady-state irradiation in air equilibrated 1,4-dioxane and N,N′-dimethylacetamide, mainly due to hydrogen abstraction from the solvent by its triplet excited state.²⁷ In 2004, Flors and Nonell demonstrated that the triplet excited state of PN is quenched by the typical electron donor 1,4-diazabicyclo-[2.2.2]octane (DABCO) with a rate constant approaching the diffusion control limit.²⁸ Moreover, in the same work evidence of the formation of the PN radical anion (PN˙⁻) was provided.

Despite the information mentioned in the previous paragraph, to the best of our knowledge, no study on the capability of PN to act as a photosensitizer through electron transfer-initiated reactions has been published up-to now and PN is systematically used as a purely type II photosensitizer. Therefore, we investigated the mechanism of the oxidation of Trp photosensitized by PN. We chose Trp as an oxidizable substrate because it is a compound able to be photooxidized through both mechanisms, it is highly soluble in water and may be easily quantified by chromatographic methods. In this work, we have studied the interaction of Trp with the excited states of PN, the production of ¹O₂ in the presence of Trp, the formation of radicals and the kinetics of the photodegradation of Trp under different experimental conditions.

Experimental section

General

Tryptophan (Trp) (Sigma Chemical Co) was of the highest purity available (>98%) and was used without further purification. 1H-Phenalen-1-one (PN) (perinaphthenone, Merck) was purified as indicated in ref. 14. Other chemicals were also provided by Sigma.

Measurements of pH were performed with a pH-meter PHM220 (Radiometer Copenhagen) combined with a pH electrode pHC2011-8 (Radiometer Analytical). The pH of the aqueous solutions was adjusted by adding very small aliquots (a few μL) of concentrated (0.1–2 M) HCl or NaOH solutions using a micropipette. In experiments using D₂O as a solvent, D₂O (>99.9%; Sigma-Aldrich), DCl (99.5%; Aldrich) in D₂O, and NaOD (Euriso-Top CEA, France) in D₂O were employed.

Singlet oxygen (¹O₂) studies

The experiments were carried out at room temperature using D₂O as a solvent since the lifetime of ¹O₂ (τ_Δ) is much longer in D₂O than in H₂O.^29,30

Steady-state experiments, equipment 1. The experiments were performed using a NIR PMT Module H10330-45 (Hamamatsu) coupled to a single-photon-counting equipment FL3 TCSPC-SP (Horiba Jobin Yvon).³¹ The sample solution (0.8 mL) in a quartz cell (1 cm × 0.4 cm) was irradiated with a CW 450W Xenon source through an excitation monochromator (FL-1004 with a 1200 groove per mm, 330 nm blaze grating). The luminescence in the near-infrared (NIR) region, after passing through an emission monochromator (FL-1004 with a 600 groove per mm, 1000 nm blaze grating), was detected at 90° with respect to the incident beam using the NIR PMT and counted by a FluoroHub-B module. Corrected emission spectra obtained by excitation at 365 nm were recorded between 950 and 1400 nm, and the total integrated ¹O₂ phosphorescence intensities (I_P) were calculated by integration of the emission band centered at ca. 1270 nm.

Steady-state experiments, equipment 2. Singlet oxygen measurements upon continuous monochromatic excitation of the sensitizer were also performed on a custom-built equipment described elsewhere.^32,33 Briefly, the solutions in a quartz cell (1 cm × 1 cm) were irradiated with a Xe–Hg arc lamp (1 kW) through a water filter, focusing optics and a monochromator. The ¹O₂ emission was collected at right angle with respect to the incident beam using a mirror, chopped, and after passing through a focusing lens and appropriate filters (cut-off at 1000 nm, interference at 1270 nm) was detected using a cooled NIR photomultiplier (Hamamatsu R5509 PMT), a lock-in amplifier and an oscilloscope.

Time-resolved experiments. A phosphorimeter (FL-1040), including a pulsed xenon lamp (0.05–25 Hz, 3 μs at full-width half-maximum) and gating electronics to control the size and temporal displacement of the detection windows, was used. The other optical devices were the same as those described in “Steady-state experiments, equipment 1”. In this case the emission at 1270 nm was registered as a function of time.

Steady-state irradiation

Irradiation set-up. Aqueous solutions containing Trp and PN were irradiated in 1 cm path length quartz cells at room temperature with a Rayonet RPR lamp emitting at 350 nm (bandwidth ∼20 nm) (Southern N. E. Ultraviolet Co.). The experiments were performed in the presence and absence of dissolved O₂. Experiments with air-equilibrated solutions were performed in open quartz cells without bubbling, whereas argon and oxygen-saturated solutions were obtained by bubbling for 20 min with the corresponding gas (Linde, purity > 99.998%), previously water saturated.

Actinometry. Aberchrome 540 (Aberchromics Ltd) was used as an actinometer for the measurements of the incident photon flux (q_p,0) at the excitation wavelength. A value of 1.4 (±0.1) × 10⁻⁵ einstein L⁻¹ s⁻¹ was obtained for q_p,0 at 350 nm in the irradiation setup described in the previous paragraph. The method for the determination of q_p,0 has been described in detail elsewhere.^34,35 Values of the photon flux absorbed (q_p,a) were calculated from q_p,0 according to the Lambert–Beer law (q_p,a = q_p,0 (1–10^−A), where A is the absorbance of the reactant at the excitation wavelength).

UV/Visible spectrophotometry. Electronic absorption spectra were recorded on a Shimadzu UV-1800 spectrophotometer, using quartz cells of 1 cm optical path length. The absorption spectra of the solutions were recorded at regular intervals of irradiation time.

High-performance liquid chromatography. A high-performance liquid chromatograph Prominence from Shimadzu (solvent delivery module LC-20AT, on-line degasser DGU-20A5, communications bus module CBM-20, auto sampler SIL-20A HT, column oven CTO-10AS VP and photodiode array detector SPD-M20A) was employed for monitoring the reaction. A Synergi Polar-RP column (ether-linked phenyl phase with polar end capping, 150 × 4.6 mm, 4 μm, Phenomenex) was used for product separation. Aqueous solutions of AcNH₄ (10 mM, pH = 6.5) were used as mobile phase.

Determination of O₂ concentration. The O₂ consumption during irradiation was measured with an O₂-selective electrode (Consort c932). The solutions and the electrode were placed in a closed glass cell of 130 mL.

Detection and quantification of H₂O₂. H₂O₂ was determined by its reaction with 4-aminophenazone and phenol catalyzed by the enzyme peroxidase to yield 4-(p-benzoquinone monoimino)phenazone, which is detected by its absorbance in the visible region.^36,37 This assay has high sensitivity and specificity due to the intense absorbance of the product at 505 nm and the enzymatic catalysis, respectively. The reactants were purchased from Wiener Laboratorios SAIC (cholesterol kit). Briefly, 500 μL of irradiated solution was added to 600 μL of reagent. The absorbance of the resulting mixture at 505 nm was measured after 30 min at room temperature, under conditions of reduced environmental light, using the reagent as a blank. Aqueous H₂O₂ solutions prepared from commercial standards were employed for obtaining the corresponding calibration curves.

Laser flash photolysis

The lifetimes of the PN triplet excited state in Ar- and O₂-saturated and in air-equilibrated D₂O solutions were determined by laser flash photolysis. A Nd:YAG laser (GCR 130-1, pulse width 9 ns, 355 nm) was used for sample irradiation. The absorbance of PN at 355 nm was 0.5. The monitoring system consisted of a 150 W pulsed Xe arc lamp, a R928 photomultiplier and a 05-109 Spectra Kinetics Applied Photophysics monochromator. Signals were digitized by a HP54522A oscilloscope. Samples were irradiated in quartz cells; solutions were over-oxygenated or deoxygenated by continuous bubbling with O₂ or Ar.

Electron paramagnetic resonance-spin trapping experiments

EPR spectra were collected on a Bruker ESP 500E spectrometer. Samples were irradiated with a Rayonet RPR3500 lamp. The following instrumental settings were employed for the measurements: microwave power: 20 mW; field modulation amplitude: 0.1 mT; field modulation frequency: 100 kHz; microwave frequency: 9.77 GHz.

EPR-spin trapping detection. Nitrones are common reagents for the detection and identification of transient radicals due to their ability to form persistent radical adducts that are detectable and finger printable by EPR spectroscopy.^38,39 In our experiments, 5,5-dimethyl-1-pyrroline-N-oxide (DMPO) from Sigma was used as spin trap.⁴⁰ Samples (1 mL) contained PN (Abs_{(365 nm)} = 0.8), 3 × 10⁻⁴ M Trp, buffer Tris/HCl, pH = 7.0 and DMPO (5 × 10⁻⁴ M). The irradiation was performed in air-equilibrated and O₂-saturated solutions. Samples in quartz flat cells were exposed to UV-A irradiation (Rayonet RPR lamp, vide supra) at room temperature. EPR spectra were recorded at various irradiation times.

Results and discussion

Quenching of ¹O₂ and of the triplet excited state of PN by Trp

The photosensitized formation of singlet oxygen (¹O₂) by 1H-phenalen-1-one (PN) involves energy transfer from the triplet excited state of PN (³PN*) to dissolved molecular oxygen (³O₂) (reactions (1) and (2)).

¹O₂ relaxes to its ground state (³O₂) through solvent induced radiationless and radiative pathways (reactions (3) and (4)). It may also be deactivated by a physical quencher (reaction (5)) and/or oxidize an acceptor molecule (reaction (6)).

Steady-state and time resolved analyses of ¹O₂ were performed irradiating PN at 365 nm in D₂O in the presence of various tryptophan (Trp) concentrations and registering the phosphorescence emission of ¹O₂ in the near-infrared (NIR) region. As can be inferred from the corresponding absorption spectra (Fig. 1), under these experimental conditions, PN was excited whereas Trp did not absorb radiation.

Time-resolved singlet oxygen measurements

In time-resolved experiments, first-order kinetics were observed for the decays of the ¹O₂ emission, following PN excitation in the presence of different concentrations of Trp. The corresponding ¹O₂ lifetimes (τ_Δ) were calculated by fitting the emission decay with eqn (7):where S(t) is the signal registered by the NIR detector (proportional to the ¹O₂ concentration at a given time t) and S_i is the pre-exponential factor (proportional to the initial ¹O₂ concentration (t = 0, end of flash)). It should be noted that eqn (7) is only valid if the ¹O₂ lifetime (τ_Δ) is much longer than that of the sensitizer triplet excited state (τ_T) under the experimental conditions used.⁴¹ This is the case for PN in air-equilibrated D₂O solutions: τ_T was measured by laser flash photolysis and found to be approx. 1.7 μs, compared to a τ_Δ value of 59 (±2) μs measured in this work, in good agreement with literature values.^30,31

In the presence of a quencher (Q), τ_Δ is given by:

τ_Δ = 1/(k^Δ_d + k^Δ_t[Q]) (8)

where k^Δ_d (= 1/τ⁰_Δ) is the rate constant of ¹O₂ deactivation by the solvent (s⁻¹, with k^Δ_e ≪ k^Δ_d, in most solvents (reactions (3) and (4))^42,43 and τ⁰_Δ is the ¹O₂ lifetime in the absence of Q), k^Δ_t (L mol⁻¹ s⁻¹) is the rate constant of ¹O₂ total (physical and chemical) quenching by Q (k^Δ_t = k^Δ_r + k^Δ_q, reactions (5) and (6)). It should be noted that there is no significant ¹O₂ quenching by the sensitizer itself in the case of PN in the range of concentrations generally used (k^Δ_t [PN] ≪ k^Δ_d).¹⁵

The analysis showed that τ_Δ decreased as a function of the Trp concentration (Fig. 2). This result was expected since Trp is able to deactivate ¹O₂ by physical quenching (reaction (5)) and to trap ¹O₂ by chemical reaction (reaction (6)).⁴⁴

Fig. 2 Time-resolved ¹O₂ experiments: quenching of ¹O₂ emission by Trp. Experiments performed in D₂O solutions containing PN and Trp at pD 6.2; excitation wavelength 365 nm, A (365 nm) = 1.07. (a) Selected NIR emission decays in the absence and in the presence of various concentrations of Trp that appear above each decay (μM); (b) Stern–Volmer plot of the ¹O₂ lifetimes (τ_Δ), calculated analyzing the NIR ¹O₂ luminescence decays (eqn (7)); (c) Stern–Volmer plot of the initial ¹O₂ concentration, estimated through the pre-exponential factor (S_i, eqn (7)).

According to eqn (8), ¹O₂ quenching by Trp (Q) should follow a linear Stern–Volmer behavior (eqn (9)).

τ⁰_Δ/τ_Δ = 1 + τ⁰_Δk^Δ_t[Trp] (9)

Fig. 2 shows that the corresponding plot (eqn (9)) was linear with a slope K_SV = 2690 L mol⁻¹, and a value of 4.4 (±0.4) × 10⁷ L mol⁻¹ s⁻¹ was obtained for k^Δ_t by Trp, which is of the same order of magnitude as a previously published value (3.2 × 10⁷ L mol⁻¹ s⁻¹).⁴⁴

The emission signal at t = 0 (S_i, eqn (7)) and, consequently, the amount of ¹O₂ initially produced by energy transfer from ³PN* to ³O₂ (reaction (2)), decreased with the increase of the Trp concentration (Fig. 2a). This fact reveals that, besides quenching by dissolved ³O₂ (reaction (2)) and deactivation by intersystem crossing to the ground state (reaction (10)), ³PN* may be quenched by Trp (reactions (11) and (12)). Note that the quenching of the singlet excited state of PN (¹PN*) by Trp would also result in a decrease of the ³PN* concentration and consequently in a decrease in the ¹O₂ production. However, the quenching of ¹PN* by Trp can be discarded at the concentrations of Trp used because the lifetime of ¹PN^* is very short, due to the efficient intersystem crossing to yield ³PN*.⁴⁵

Under these conditions, ³PN* quenching by Trp competes with energy transfer from ³PN* to ³O₂ (reaction (2)), and therefore the efficiency of energy transfer (ϕ_et) and thus the quantum yield of ¹O₂ production (Φ_Δ) decreases as the Trp concentration increases (eqn (13)).

where Φ_T is the quantum yield of ³PN* formation, k^T_d (s⁻¹) is the rate constant of ³PN^* deactivation in the absence of quencher (reaction (10)), k_et (L mol⁻¹ s⁻¹) is the rate constant of the energy transfer from ³PN* to ³O₂ (reaction (2)) and k^T_t (L mol⁻¹ s⁻¹) is the rate constant of ³PN* total quenching by Trp (k^T_t = k^T_q + k^T_r, reactions (11) and (12)).

The ratio of the Φ_Δ values in the absence and in the presence of Trp is thus given by:

where τ⁰_T is the lifetime of ³PN* in the absence of Trp in the air-equilibrated solution (). Since the initial ¹O₂ emission signal (S_i, eqn (7)) is proportional to the initial ¹O₂ concentration, and therefore to Φ_Δ, the quenching efficiency of ³PN* by Trp may be evaluated from the Stern–Volmer analysis of the variation of S_i as a function of the Trp concentration (eqn (14), with S⁰_i/S_i = Φ⁰_Δ/Φ_Δ).

In agreement with eqn (14), the Stern–Volmer plot of S⁰_i/S_i vs. [Trp] was linear (Fig. 2c). Taking into account the τ_T value measured in air-equilibrated D₂O solution (1.7 μs), calculation of k_t^T from the slope of the Stern–Volmer plot (eqn (14)), K_SV = τ⁰_Tk^T_t = 9900 L mol⁻¹ yielded a value of 6 (±1) × 10⁹ L mol⁻¹ s⁻¹. This value is higher than the rate constant of ³PN* quenching by ³O₂ measured in this work by laser flash photolysis (k_et = 2.0 (±0.2) × 10⁹ L mol⁻¹ s⁻¹).

Steady-state singlet oxygen measurements

In another set of experiments, the emission of ¹O₂ was recorded at a constant PN concentration and different concentrations of Trp (0 to 400 μM) at pD 6.2. The measurements were carried out in independent experiments using equipment 1 and 2, described in the Experimental section. Emission spectra registered with equipment 1 showed a strong decrease in the ¹O₂ phosphorescence with the increase of the Trp concentration, the wavelength of the emission maximum remaining unchanged (Inset Fig. 3).

Fig. 3 Steady-state ¹O₂ experiments. Quenching of the emission of ¹O₂ by Trp. Experiments performed in D₂O solutions containing PN and Trp at pD 6.2. Stern–Volmer plot of the phosphorescence intensities (I_P); excitation wavelength 365 nm, experiments performed using equipments 1 (▼) and 2 (▲), described in the Experimental section, solid line: fit using eqn (16) and the parameters obtained in time-resolved experiments (Fig. 2). Inset: selected NIR emission spectra registered in the absence and in the presence of Trp at various concentrations (corresponding values indicated above each spectrum (μM)).

The corresponding Stern–Volmer plot of the emission intensity (I_p) (I⁰_p/I_p vs. [Trp]) was not linear. The same result was observed for the signals registered using equipment 2 (Fig. 3). This behavior is again a consequence of quenching of both ¹O₂ and ³PN* by Trp. Indeed, under continuous irradiation, I_p is proportional to the steady-state concentration of ¹O₂ ([¹O₂]_ss). Taking into account the reactions involved in the production and quenching of ¹O₂ (eqn (1)–(6)) and applying the quasi-stationary hypothesis to the concentrations of excited states, [¹O₂]_ss may be expressed as:

[¹O₂]_ss = q_p,aΦ_Δτ_Δ (15)

where q_p,a (einstein L⁻¹ s⁻¹) is the photon flux absorbed by the sensitizer (see Experimental section).

In a given solvent and at a constant concentration of the sensitizer PN (photon flux absorbed by PN constant), a Stern–Volmer analysis of the quenching of I_p by Trp leads to the following relation by combining eqn (9), (14) and (15):

I⁰_p/I_p = [¹O₂]_ss⁰/[¹O₂]_ss = (Φ⁰_Δτ⁰_Δ)/(Φ_Δτ_Δ) = (1 + τ⁰_Δk^Δ_t[Trp]) (1 + τ⁰_Tk^T_t[Trp]) (16)

where the superscript “0” refers to the values in the absence of Trp in air-equilibrated solutions.

The plot of I⁰_P/I_P vs. [Trp] can be fitted with a second order polynomial, in agreement with eqn (16). Using the values of the various constants determined by time-resolved measurements, a reasonable fit of the non linear Stern–Volmer plot (I⁰_p/I_p) could be obtained (Fig. 3).

Taking into account the results presented in this section, we can conclude that Trp is able to interact with both ¹O₂ and ³PN*. Moreover the quenching of ³PN* is more efficient and, subsequently, occurs at lower Trp concentrations.

Evidence for the involvement of an electron transfer process

Singlet oxygen studies demonstrated that Trp quenches ³PN*, but no information about the nature of that interaction was provided. As mentioned above the quenching may consist in a physical process (reaction (11)) and/or a chemical reaction (reaction (12)). If the latter takes place, the process might involve an electron transfer to yield the corresponding radical ion pair (reaction (17)). The Trp radical cation may deprotonate to the tryptophanyl radical (Trp(–H)˙) (pKa = 4.1) (reaction (18)).⁴⁶ It is well-established that O₂ readily quenches an organic radical anion to produce the superoxide anion (O₂˙⁻) (reaction (19)).^47,48

The free energy change (ΔG) of reaction (17) can be estimated with eqn (20):¹

ΔG = F [E_{(Trp˙⁺/Trp)} − E_(PN/PN˙⁻) − (e_o²/εR_D⁺A⁻)] − ΔE_0,0 (20)

where E_{(Trp˙⁺/Trp)} and E_(PN/PN˙⁻) are the standard electron potentials of electron donor and acceptor, respectively. These values have already been reported for Trp (E_{(Trp˙⁺/Trp)} = 1.015 V vs. NHE⁴⁹) and PN (E_(PN/PN˙⁻) = −0.86 V vs. NHE²⁸). ΔE_0,0 is the energy of the triplet excited state of PN and has been estimated from its phosphorescence spectra (ΔE_T = 185 kJ mol⁻¹).^15,17 The term e_o²/εR_D⁺A⁻ is the solvation energy of an ion pair D⁺A⁻ and can be ignored in the case of strong polar solvents. The calculated ΔG value was −4.1 kJ mol⁻¹, thus indicating that electron transfer from Trp to the triplet excited state of ³PN* can spontaneously occur.

To investigate further if an electron transfer process takes place when solutions containing PN and Trp are exposed to UV-A irradiation, electron paramagnetic resonance (EPR) experiments were performed in the presence of a spin trap (5,5-dimethyl-1-pyrroline N-oxide, DMPO).

No EPR signal was registered after irradiation of an air-equilibrated solutions (pH 7.0) containing PN (A (365 nm) = 0.8), Trp (3 × 10⁻⁴ M) and DMPO (5 × 10⁻² M) (Fig. 4a). However, the irradiation of the same solution, previously bubbled with N₂, led to the immediate formation of an EPR signal, which increased with the irradiation time (Fig. 4). The solution containing DMPO as a spin trap presents a six line spectrum having hyperfine coupling constants of a_N = 15.8 G, a_H = 23.8 G (g = 2.0060). This EPR spectrum corresponds to a carbon-centered radical.⁵⁰ Moreover, our values match well those reported for the radical resulting from the one-electron oxidation of Trp by Br₂˙⁻ under anaerobic conditions,⁵¹ which allows us to assign the EPR spectrum registered in our reaction system to a Trp radical.

Fig. 4 EPR experiments performed in aqueous solutions containing PN (A (365 nm) = 0.8), Trp (3 × 10⁻⁴ M) and DMPO (5 × 10⁻² M) as spin trap; excitation wavelength 350 nm, buffer Tris/HCl, pH = 7.0. (a) Signal registered after 15 minutes of irradiation in air-equilibrated solutions; (b) signal registered in a N₂-saturated solution before irradiation; (c) signal registered in a N₂-saturated solution after 15 minutes of irradiation.

To confirm that the registered EPR spectra correspond to a radical generated in the photosensitized process, several control experiments were carried out. First, no EPR signal was detected in solutions containing PN, Trp and DMPO before irradiation (solutions were kept in the dark for 30 minutes). In addition, no EPR signal could be recorded after irradiation of solutions containing: (i) Trp and DMPO; (ii) PN and DMPO; (iii) only DMPO.

Therefore, the results presented in this section demonstrate that upon UV-A irradiation aqueous solutions containing PN and Trp produce radicals, in particular Trp radicals, which in turn demonstrate that electron transfer takes place from Trp to excited PN. As mentioned in the Introduction, the formation of PN radicals upon excitation in the presence of the electron donor DABCO was reported.²⁸ However, it is worth mentioning that DABCO is more easily oxidized than Trp (E_{(DABCO˙⁺/DABCO)} < E_{(Trp˙⁺/Trp)}). Therefore, it can be concluded that ³PN* is able to oxidize not only typical organic electron donors, but also oxidizable biomolecules.

Oxidation of Trp photosensitized by PN

The efficiency of Trp oxidation photosensitized by PN was investigated under different experimental conditions. Air-equilibrated aqueous solutions containing the photosensitizer and the substrate were exposed to UV-A (350 nm) radiation for various periods of time. From the corresponding absorption spectra (Fig. 1), it can be inferred that only PN was excited. The samples were analyzed by UV-visible spectrophotometry, HPLC and the concentration of H₂O₂ produced was measured (see Experimental section). Significant changes in the absorption spectra of the solutions were registered after irradiation (Fig. 1S, ESI†) and the concentration profile of Trp, determined by HPLC, showed a decrease of the Trp concentration as a function of irradiation time (Fig. 5). In addition, H₂O₂ was generated and its concentration increased during irradiation (Fig. 5). The detected H₂O₂ would then be the product of the spontaneous disproportionation of O₂˙⁻ with its conjugated acid, the hydroperoxyl radical (HO₂˙) (reactions (21) and (22)).⁵²

Fig. 5 Evolution of the Trp and H₂O₂ concentrations in air-equilibrated, O₂-saturated and Ar-saturated aqueous solutions (pH 5.5) of PN (A (350 nm) = 0.5) and Trp (100 μM) exposed to UV-A radiation (350 nm) as a function of elapsed time.

To better elucidate the role of O₂ in the photosensitization of Trp by PN, photolysis experiments were carried out in deoxygenated solutions and the results were compared with those performed in air-equilibrated solutions (Fig. 5). HPLC measurements showed that, in the experiments lacking oxygen, the Trp concentration did not decrease significantly, within experimental error. Control experiments were also carried out on Trp (100 μM) solutions (pH 5.5) in the absence of PN. As expected, no consumption of the amino acid was detected, thus excluding the possibility of product formation by spurious direct excitation of Trp.

In experiments performed with a different optical geometry, an O₂-selective electrode (Experimental section) was used to monitor the decrease of the O₂ concentration upon irradiation of air-equilibrated solutions containing Trp (400 μM) and PN (A (350 nm) = 0.5) at pH 5.5 (Fig. 6). Control experiments in the absence of Trp, showed that the consumption of O₂ resulting from the photolysis of PN itself was negligible in comparison with that observed in the presence of Trp (Fig. 6).

Fig. 6 Evolution of the O₂ concentration in irradiated solutions (pH = 5.5) containing Trp (400 μM) and PN (A (350 nm) = 0.50) as a function of time (●). Control experiment: irradiation of a solution of PN (A (350 nm) = 0.5) at pH 5.5 in the absence of Trp (Δ).

Mechanistic analysis

The results presented so far clearly demonstrate that PN photosensitizes the oxidation of Trp under UV-A irradiation in air-equilibrated aqueous solutions. Taking into account that: (i) the process needs O₂ to take place, (ii) PN is accepted as a pure type II photosensitizer due to its very high efficiency of ¹O₂ production upon UV-A irradiation (see Introduction) and (iii) it has been previously demonstrated that Trp is easily oxidized by ¹O₂,⁴⁴ it may be assumed that the photooxidation of Trp is a ¹O₂-mediated reaction. However, in the previous sections it was also demonstrated that ³PN* is quenched by Trp and that Trp radicals are formed in a process photosensitized by PN. Does the electron transfer process lead to a significant consumption of Trp or is it negligible compared to ¹O₂-mediated oxidation? This question cannot be answered considering the results shown so far. Therefore additional experiments were carried out to explore this point.

In another set of photolysis experiments, O₂-saturated solutions were irradiated and the results were compared with those performed in air-equilibrated solutions. The concentration profiles obtained by HPLC clearly showed that the rate of Trp disappearance and the rate of H₂O₂ production were greater in air-equilibrated than in O₂-saturated solutions (Fig. 5). These results are not consistent with a process wherein Trp would be consumed only by reaction with ¹O₂. However, we have shown (section “Quenching of ¹O₂ and of the triplet excited state of PN by Trp”) that ³PN* is also quenched by Trp, this process competing with quenching of ³PN* by O₂ to yield ¹O₂ (reaction (2)). In oxygen-saturated solutions, the latter reaction competes more efficiently with ³PN* quenching by Trp than in air-equilibrated solution and the proportion of ³PN* reacting with Trp is smaller. Consequently, the rate of Trp consumption is also slower (Fig. 5), since the rate constant of reaction (2) (k_et) is lower than that of ³PN* quenching by Trp (k^T_t, reactions (11) and (12)).

Calculation of Φ_Δ in air-equilibrated solutions using eqn (14) yields values of 0.49 and 0.20 at Trp concentrations of 100 μM and 400 μM, respectively, a large decrease compared to the value reported in the absence of quencher (Φ⁰_Δ = 0.97).^14,15 Subsequently, if PN would oxidize Trp only via a type II mechanism and Trp acted as a pure physical quencher of ³PN*, a reduced consumption of the amino acid should be observed when the Trp concentration increases. Therefore, the experiment corresponding to Fig. 5 was repeated, but at a higher Trp concentration (photon flux absorbed by PN remained the same). As shown in Fig. 7, Trp was still consumed in this experiment and the rate of its consumption ((d[Trp]/dt)_i,exp, Table 1) was faster than that registered at lower Trp concentration (Fig. 5), confirming that PN is able to photosensitize the oxidation of Trp via a type I mechanism (reaction (17)).

Fig. 7 Evolution of the Trp concentration in air-equilibrated aqueous solution (pH 5.5) of PN (A (350 nm) = 0.5) and Trp (390 μM) exposed to UV-A radiation (350 nm) as a function of elapsed time.

Table 1 Initial rates of Trp consumption: experimental values (d[Trp]/dt)_i,exp (air equilibrated aqueous solutions, pH 5.5, steady-state irradiation at 350 nm) and calculated values (d[Trp]/dt)_i,Δ due to the chemical reaction between ¹O₂ and Trp^a

[Trp]i (μM)	(d[dTrp]/dt)i,Δ (10^−7 M s^−1)	(d[Trp]/dt)i,exp (10^−7 M s^−1)
a HPLC was used to determine (d[Trp]/dt)i,exp, and values of (d[Trp]/dt)i,Δ were calculated from eqn (24).
100	0.22	4.3
390	0.33	7.5

---

The contribution of ¹O₂ can be estimated considering the reported value of the rate constant of the chemical reaction between ¹O₂ and Trp (k^Trp_r = 1.3 × 10⁷ M⁻¹ s⁻¹).⁴⁴ The initial rate of the oxidation of Trp by ¹O₂ ((d[Trp]/dt)_i,Δ) was calculated using eqn (23):

Combining eqn (15), (16) and (23), (d[Trp]/dt)_i,Δ can be estimated as a function of the Trp concentration (eqn (24))

In air-equilibrated solutions, the values of (d[Trp]/dt)_i,Δ calculated from eqn (24) were much lower than the initial rate of Trp consumption experimentally determined by HPLC analysis ((d[Trp]/dt)_i,exp, Table 1). This analysis clearly demonstrates that the oxidation of Trp by ¹O₂ does not contribute significantly to the consumption of Trp, with calculated initial rates not exceeding 5% of the experimentally determined initial rates (Table 1). Therefore, the electron transfer process (reaction (17)) is largely dominant, at least in air-equilibrated solutions.

Therefore, taking into account all the results presented, the general mechanism summarized in Fig. 8 can be proposed. After UV-A excitation of PN and formation of its triplet excited state (³PN*), three reaction pathways compete for the deactivation of the latter: intersystem crossing to singlet ground state, energy transfer to O₂ leading to the regeneration of Ptr and the production of ¹O₂, and electron transfer between Trp and ³PN* yielding the corresponding pair of radical ions (PN˙⁻ and Trp˙⁺). In the following step, the electron transfer from PN˙⁻ to O₂ regenerates PN and forms O₂˙⁻, which undergoes disproportionation into H₂O₂ and O₂. Finally, the reaction of Trp˙⁺/Trp(–H)˙ with O₂ leads to the formation of products. In the absence of O₂ recombination of PN˙⁻ and Trp˙⁺ recovers the reactants, Fig. 8) and, therefore no consumption of the amino acid is observed under anaerobic conditions. Besides, in oxygen-saturated solutions the quenching of ³PN* by O₂ competes with the electron transfer process and, consequently, the rate of Trp consumption is slower than in air-equilibrated solutions.

Fig. 8 Mechanism of the oxidation of tryptophan (Trp) photosensitized by phenalenone (PN).

Conclusions

Although photosensitized oxidations may involve the generation of radicals (type I mechanism) and/or the production of singlet molecular oxygen (¹O₂) (type II mechanism), it is generally accepted that the photosensitization of proteins occurs mainly through the reactions of ¹O₂ with tryptophan (Trp) and other amino acids. In this work, we have investigated the mechanism of the oxidation of Trp photosensitized by phenalenone (PN), one of the well-known type II photosensitizers, widely used as a ¹O₂ reference sensitizer. We have studied the interaction of Trp with the excited states of PN, the production of ¹O₂ in the presence of Trp, the formation of radicals and the kinetics of the photodegradation of Trp under different experimental conditions. From the results obtained, it could be established that ¹O₂ played a minor role and that the predominant mechanism of the photosensitization of Trp by PN involved an electron transfer process (type I mechanism).

A photosensitized reaction is currently accepted to be a ¹O₂-mediated oxidation (type II mechanism) if (i) the reaction needs O₂ to take place, (ii) the photosensitizer generates ¹O₂ upon irradiation, (iii) the target molecule is sensitive to¹O₂. Although this assumption seems to be quite obvious and trivial, in this work we have presented evidence against it. The predominant type I mechanism of the oxidation of Trp photosensitized by PN, which is systematically used as a purely type II photosensitizer, calls into question many reported photooxidation mechanisms and even the role of ¹O₂ in processes involved in the photodynamic effects of a variety of photosensitizers in living systems.

Acknowledgements

The present work was partially supported by Consejo Nacional de Investigaciones Cientificas y Técnicas (CONICET-Grant PIP 112-200901-00425) and the Universidad Nacional de La Plata (UNLP-Grant X586). C.C. thanks the CONICET for graduate research fellowships. A. H. T. thanks the Royal Society of Chemistry for a Journals Grant for International Authors. The authors thank the Centre National de la Recherche Scientifique (CNRS, France) for supporting their collaboration through a Programme de Coopération Scientifique (PICS/CNRS no. PICS05920). The authors are grateful to Dr Claire Richard (UMR CNRS 6296, Université Blaise Pascal, Aubière, France) for the laser flash photolysis measurements. A. H. T. and C. L. are research members of CONICET.

References

1. S. E. Braslavsky, Pure Appl. Chem., 2007, 79, 293–465 and references therein.
2. F. Ghissassi El, R. Baan and K. Straif, Lancet Oncol., 2009, 10, 751–752.
3. J. Cadet, T. Douki, J. L. Ravanat and P. Di Mascio, Photochem. Photobiol. Sci., 2009, 8, 903–911.
4. F. Comanjón, L. Villén, D. García-Fresnadillo and G. Orellana, Environ. Sci. Technol., 2008, 42, 301–307.
5. F. Manjón, D. García-Fresnadillo and G. Orellana, Photochem. Photobiol. Sci., 2009, 8, 926–932.
6. A. P. Castano, P. Mroz and M. R. Hamblin, Nat. Rev. Cancer, 2006, 6, 535–545.
7. P. Babilas, S. Schreml, M. Landthaler and R.-M. Szeimies, Photodermatol., Photoimmunol. Photomed., 2010, 26, 118–132.
8. J. Cadet, E. Sage and T. Douki, Mutat. Res., 2005, 571, 3–17.
9. C. S. Foote, Photochem. Photobiol., 1991, 54, 659.
10. M. C. De Rosa and R. J. Crutchley, Coord. Chem. Rev., 2002, 351, 233–234.
11. J. Cadet, J. L. Ravanat, G. R. Martinez, M. H. G. Medeiros and P. Di Mascio, Photochem. Photobiol., 2006, 82, 1219–1225.
12. M. Tarr and D. P. Valenzeno, Photochem. Photobiol. Sci., 2003, 2, 355–361.
13. D. I. Pattison, A. S. Rahmantoa and M. J. Davies, Photochem. Photobiol. Sci., 2012, 11, 38–53.
14. E. Oliveros, P. Suardi-Muraseco, T. Aminian-Saghafi, A. M. Braun and H. J. Hansen, Helv. Chim. Acta, 1991, 74, 79–90.
15. R. Schmidt, C. Tanielian, R. Dunsbach and C. Wolff, J. Photochem. Photobiol., A, 1994, 79, 11–17.
16. C. Martí, O. Jürgens, O. Cuenca, M. Casals and S. Nonell, J. Photochem. Photobiol., A, 1996, 97, 11–18.
17. C. Flors and S. Nonell, Helv. Chim. Acta, 2001, 84, 2533–2539.
18. K. Winters, J. C. Batterton and C. Van Baalen, Environ. Sci. Technol., 1977, 11, 270–272.
19. J. G. Luis, W. Quiñones, F. Echeverri, T. A. Grillo, M. P. Kishi, F. Garcia-Garcia, F. Torres and G. Cardona, Phytochemistry, 1996, 41, 753–757.
20. C. Flors and S. Nonell, Acc. Chem. Res., 2006, 39, 293–300.
21. J. M. Kornfeld and J. M. Edwards, Biochim. Biophys. Acta, 1972, 286, 88–90.
22. T. Takamura-Enya, R. Ishii and Y. Oda, Mutagenesis, 2011, 26, 499–505.
23. A. Belvedere, F. Boscá, A. Catalfo, M. C. Cuquerella, G. De Guidi and M. A. Miranda, Chem. Res. Toxicol., 2002, 15, 1142–1149.
24. A. Posadaz, N. M. Correa, M. A. Biasutti and N. A. García, Photochem. Photobiol., 2010, 86, 96–103.
25. M. A. Biasutti, A. Posadaz and N. A. García, J. Pept. Res., 2003, 62, 11–18.
26. G. P. Rabold, K. H. Bar-Eli, E. Reid and K. Weiss, J. Chem. Phys., 1965, 42, 2438–2447.
27. E. Oliveros, S. H. Bossmann, S. Nonell, C. Martí, G. Heit, G. Tröscher, A. Neuner, C. Martínez and A. M. Braun, New J. Chem., 1999, 23, 85–93.
28. C. Flors and S. Nonell, J. Photochem. Photobiol., A, 2004, 163, 9–12.
29. P. R. Ogilby and C. S. Foote, J. Am. Chem. Soc., 1983, 105, 3423–3430.
30. F. Wilkinson, H. P. Helman and A. B. Ross, J. Phys. Chem., 1995, 24, 663–677.
31. M. P. Serrano, M. Vignoni, M. L. Dántola, E. Oliveros, C. Lorente and A. H. Thomas, Phys. Chem. Chem. Phys., 2011, 13, 7419–7425.
32. E. Oliveros, P. Murasecco-Suardi, A. M. Braun and H.-J. Hansen, Methods in Enzymology, Carotenoids, ed. L. Packer, Academic Press, San Diego, CA, 1992, vol. 213, pp. 420–429.
33. F. M. Cabrerizo, M. Laura Dántola, G. Petroselli, A. L. Capparelli, A. H. Thomas, A. M. Braun, C. Lorente and E. Oliveros, Photochem. Photobiol., 2007, 83, 526–534.
34. A. M. Braun, M. T. Maurette and E. Oliveros, in Photochemical Technology, Wiley, Chichester, 1991, ch. 2, pp. 85–88.
35. H. J. Kuhn, S. E. Braslavsky and R. Schmidt, Pure Appl. Chem., 2004, 76, 2105–2146.
36. C. C. Allain, L. S. Poon, C. S. G. Chan, W. Richmond and P. C. Fu, Clin. Chem., 1974, 20, 470–475.
37. H. M. Flegg, Ann. Clin. Biochem., 1973, 10, 79–84.
38. E. Janzen, Acc. Chem. Res., 1971, 4, 31–40.
39. F. A. Villamena and J. L. Zweier, Antioxid. Redox Signaling, 2004, 6, 619–629.
40. F. A. Villamena, J. Phys. Chem. A, 2009, 113, 6398–6403.
41. S. Nonell and S. Braslavsky, in Methods in Enzymology, Academic Press, San Diego, CA, 2000, vol. 319, pp. 37–49.
42. R. D. Scurlock, S. Nonell, S. E. Braslavsky and P. R. Ogilby, J. Phys. Chem., 1995, 99, 3521–3526.
43. L. A. Martinez, C. G. Martinez, B. B. Klopotek, J. Lang, A. Neuner, A. M. Braun and E. Oliveros, J. Photochem. Photobiol., B, 2000, 58, 94–107.
44. A. Michaeli and J. Feitelson, Photochem. Photobiol., 1994, 59, 284–289.
45. M. C. Daza, M. Doerr, S. Salzmann, C. M Marian and W. Thiel, Phys. Chem. Chem. Phys., 2009, 11, 1688–1696.
46. S. Solar, N. Gtoff, P. S. Surdhar, D. A. Armstrong and A. Singh, J. Phys. Chem., 1991, 95, 3639–3643.
47. E. K. Hodgson and I. Fridovich, Photochem. Photobiol., 1973, 18, 451–455.
48. J. Eriksen, C. S. Foote and T. L. Parker, J. Am. Chem. Soc., 1977, 99, 6455–6456.
49. A. Harriman, J. Phys. Chem., 1987, 91, 6102–6104.
50. G. R. Buetner, Free Radical Biol. Med., 1997, 3, 259–303.
51. L. P. Candeias, P. Wardman and M. P. Mason, Biophys. Chem., 1997, 67, 229–237.
52. B. H. J. Bielski, D. E. Cabelli, R. L. Arudi and A. B. Ross, J. Phys. Chem., 1985, 14, 1041–1100.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c3ra46867d

---