Chem. Res. Toxicol. 2004, 17, 1227-1235
1227
UV Photolysis of 3-Nitrotyrosine Generates Highly Oxidizing Species: A Potential Source of Photooxidative Stress Thomas Nauser,† Willem H. Koppenol,† Jill Pelling,‡ and Christian Scho¨neich*,†,§ Laboratory for Inorganic Chemistry, ETH Zu¨ rich, CH-8093 Zu¨ rich, Switzerland, Department of Pathology & Laboratory Medicine, University of Kansas Medical Center, Kansas City, Kansas 66160-7410, and Department of Pharmaceutical Chemistry, University of Kansas, Lawrence, Kansas 66047 Received May 18, 2004
Laser flash photolysis at 266 nm of 3-nitrotyrosine and N-acetyl-3-nitrotyrosine ethyl ester generates an oxidizing species, which shows all of the characteristics of a hydroxyl radical. This species reacts with Br- to yield Br2•-, via an intermediate, that is kinetically identified as HOBr•-. Moreover, the formation of Br2•- can be suppressed by methanol; competition kinetics yield relative rate constants for the reaction of the reactive species with Br- and methanol that are similar to those for the hydroxyl radical. Parallel time-resolved UV/vis spectroscopy suggests the formation of phenoxyl radicals, consistent with the formation of hydroxyl radicals. Laser flash photolysis at 355 nm also generates reactive intermediates that oxidize Br- to Br2•- but appear not to be hydroxyl radicals.
Introduction
Scheme 1
3-NY1
The modification of tyrosine (Tyr) to (Scheme 1, reaction 1) represents an important NO•-dependent process during biological conditions associated with oxidative stress (1-3). Several important proteins have been shown to accumulate 3-NY under pathological conditions, such as Mn superoxide dismutase during allograft rejection (4), Tyr hydroxylase in an animal model of Parkinson’s disease (5), prostaglandin synthase in cardiovascular dysfunction (6), and the tumor suppressor protein p53 during cancer (7). A role for NO• in tumor progression has been established (8), and the mere incubation of human MCF-7 cells with an NO• donor, S-nitrosoglutathione, resulted in a significant accumulation of 3-NY on p53, monitored by Western blotting (9). Evidence was provided that p53 restricts the expression of nitric oxide synthase-2 (NOS-2), an important source of elevated levels of NO• during inflammatory conditions (8). The NO•-dependent modification of p53 (likely through peroxynitrite) causes structural alterations and the incorporation of 3-NY, which correlates with a loss of the affinity of p53 toward DNA (7). Hence, the NO•-dependent inactivation of p53 could potentially lead to a further increase of NO• levels if the modified p53 is unable to control the expression of NOS-2. An important correlation was provided by Kato et al. (10), who demonstrated that patients with 3-NY-positive esophageal cancers suffer from a lower survival rate as compared to patients with * To whom correspondence should be addressed. E-mail: schoneic@ ku.edu. † ETH Zu ¨ rich. ‡ University of Kansas Medical Center. § University of Kansas. 1 Abbreviations: HO•, hydroxyl radical; LFP, laser flash photolysis; MALDI, matrix-assisted laser desorption ionization; MeOH, methanol; MS, mass spectrometry; NANTE, N-acetyl-3-nitrotyrosine ethyl ester; NO•, nitrogen monoxide; 3-NY, 3-nitrotyrosine; UVB, ultraviolet radiation B.
3-NY-negative cancers. In the skin, exposure to either UVB radiation or sunlight results in the formation of 3-NY, suggesting that NO•-dependent pathways play a role in the photooxidative damage of the skin (11). No systematic studies on the biophysical and chemical consequences of 3-NY formation in proteins are available. Through the introduction of the 3-nitro group, the phenolic pKa of Tyr shifts to ca. 7.1, measured for several model peptides (12). Therefore, around physiological pH, a significant fraction of protein-bound 3-NY may exist in its deprotonated form, introducing an additional negative charge into the protein sequence. Such an additional negative charge may lead to conformational and functional changes of the affected protein(s). In this paper, we will establish an additional, potentially harmful role of 3-NY as a photosensitizer in light-exposed tissues: The photolysis of 3-NY results in the energydependent formation of at least two highly oxidizing species, of which one shows the characteristics of OH•. Nitroaromatic compounds have long been utilized as photosensitizers, and a manifold of organic photoreactions has been established, such as substitutions, isomerizations, and rearrangements (13). A recent MS study employing MALDI has demonstrated the photochemical
10.1021/tx049862u CCC: $27.50 © 2004 American Chemical Society Published on Web 08/14/2004
1228
Chem. Res. Toxicol., Vol. 17, No. 9, 2004
Nauser et al.
Figure 1. UV spectrum of 1.9 × 10-4 M NANTE in 2 × 10-2 M phosphate buffer (pH 6.9). Indicated are the regions of UVA, UVB, and terrestrial UVB.
net release of oxygen from 3-NY during the MS experiment, rationalized through the photoreduction of 3-NY followed by the elimination of water (14). In contrast, on the basis of product studies, Alif et al. have proposed the photolytic elimination of O•- from 2-nitrophenolate (15). The present paper provides direct kinetic evidence for the photolytic elimination of O•- and/or HO• from 3-NY and a peptide analogue, NANTE, after LFP at 266 nm in aqueous solution, pH 6.9. The 266 nm wavelength is absorbed by the high energy side of the ca. 281 nm absorption maximum of NANTE at pH 6.9 (Figure 1). The fact that this peak reaches far into the region of the terrestrial UVB (295-320 nm) (16, 17) underlines the physiological importance of this experiment. In contrast, LFP at 355 nm does not generate O•-/ HO• but a different oxidizing intermediate.
Experimental Section Materials. NANTE and 3-NY were purchased from Sigma (Buchs, Switzerland). All chemicals were of the highest commercially available purity. LFP. LFP experiments were carried out with an Applied Photophysics LKS 50 instrument, as described (18). The solutions were made up with Millipore Q water in Schlenk tubes. Deoxygenation was achieved by the use of a vacuum (membrane pump) followed by saturation with either Ar or N2O. This process was repeated at least three times. For some LFP experiments and the determination of quantum yields, the solutions were transferred into gas tight Hamilton syringes and pumped with a syringe pump through glass capillaries into a quartz sample cell with a 1 cm optical path length and a 0.1 cm depth. The irradiated cell volume was 0.1 mL, where the laser beam had a diameter of 1 cm. Reproducible photolysis yields required a thorough rinse of the flow cell after each laser pulse as some photoproducts appeared to adsorb on the cell walls, without which a drop in signal intensity was observed in subsequent experiments. Similar observations have been reported for steady state photolysis experiments of nitroaromatic compounds (19). Some LFP experiments were performed in a standard fluorescence cell with a 1 cm optical path length and a 1 cm depth, which was closed by a Youngs type valve. Because of different dose depths in both cells, the absolute photolytic yields of both cells cannot be directly compared with each other. However, the relative photolytic yields were similar within experimental error. Pulse radiolysis experiments were performed with a 2 MeV Febetron 705 accelerator. The set up and data acquisition have been described (20).
Results Photolysis of NANTE at 266 nm at Neutral to Alkaline pH. The optical spectrum of 1.9 × 10-4 M NANTE in 10-2 M potassium phosphate, pH 6.9, is shown in Figure 1. The LFP of an Ar-saturated aqueous solution of 9.4 × 10-5 M NANTE in 2 × 10-2 M potassium phosphate, pH 6.9, with 266 nm light results in photobleaching at 420 nm (-∆420) and a new absorbance in the 290-330 nm region (representatively quantified at 320 nm; ∆320) immediately after the laser pulse (data not shown). Only a negligible net change in absorbance was observed around 360 nm (∆360). A similar result was obtained with an Ar-saturated aqueous solution of 9.4 × 10-5 NANTE in 2 × 10-2 M potassium borate, pH 8.58.9, and with 9.4 × 10-5 3-NY in 2 × 10-2 M potassium borate, pH 8.5-8.9. The LFP of an O2-saturated solution of 9.4 × 10-5 NANTE in 2 × 10-2 M potassium borate, pH 8.5, showed a similar ∆320 but a ca. 20% lower -∆420. No differences in the absence or presence of 1% (v/v) MeOH were observed for both, -∆420 and ∆320, during LFP of an Ar-saturated solution of 9.4 × 10-5 M NANTE and 2 × 10-2 M potassium borate, pH 8.5. These initial measurements of absorbance changes only serve to indicate a photochemical reaction of NANTE at neutral to alkaline pH. Identification of reactive intermediates under these conditions is not possible because of simultaneous photobleaching and absorbance increase. A better spectral analysis of reactive intermediates is possible at acidic pH (vide infra). The addition of 2 × 10-3 M KBr to an Ar-saturated solution of 9.4 × 10-5 M NANTE in 2 × 10-2 M potassium phosphate, pH 6.9, results in a strong absorbance with λmax ) 360 nm, characteristic for the formation of Br2•(21) (Figure 2, trace a). The build-up of this absorbance is completely suppressed by the addition of 0.25 M MeOH (Figure 2, trace c). On the other hand, the 360 nm absorbance is recovered upon decreasing the MeOH concentration to 0.025 M and increasing the concentration of KBr to 4 × 10-3 M (Figure 2, trace b). These characteristics indicate the formation of a reactive species capable of oxidizing Br- to Br• (reactions 2 and 3) and reactive toward MeOH (reaction 4).
Ox + Br- f Ox•- + Br•
(2)
UV Photolysis of 3-Nitrotyrosine
Chem. Res. Toxicol., Vol. 17, No. 9, 2004 1229
Figure 2. LFP (266 nm) of an Ar-saturated solution containing 9.4 × 10-5 M NANTE in 2 × 10-2 M potassium phosphate (pH 6.9) and (a) 2 × 10-3 M KBr, (b) 4 × 10-3 M KBr and 0.025 M MeOH, and (c) 2 × 10-3 M KBr and 0.25 M MeOH. The absorbance was recorded at 360 nm. Negative absorbance values after the disappearance of Br2•- indicate bleaching due to the photochemical reaction of NANTE.
Br• + Br- ) Br2•-
(3)
Ox + CH3OH f products
(4)
Br• + CH3OH f HBr + •CH2OH
(5)
Br•
A direct reaction of with MeOH (reaction 5) can be excluded based on k3 ) 1.0 × 1010 M-1 s-1 (21) and k5 ) 3 × 105 M-1 s-1 (22). The decay of Br2•- is of second-order kinetics independent of the concentration of added MeOH. This observation indicates a rather low reactivity of Br2•toward NANTE, consistent with pulse radiolysis experiments (vide infra). Figure 3 shows an absorbance vs time profile for the formation of Br2•- after LFP of an Arsaturated solution of 9.4 × 10-5 M NANTE in 10-2 M potassium phosphate, pH 6.7, containing 4 × 10-3 M KBr (note that an accurate recording of the trace in Figure 3 was only possible starting at 0.1 µs after LFP due to electronic noise within the first 0.1 µs). The trace is fit according to first-order kinetics with t1/2 ) 150 ns (k ) 4.6 × 106 s-1). This value corresponds well to the published rate constant for the decomposition of HOBr•into HO- and Br• (rection 7; k7 ) 4.2 × 106 s-1) (21) and suggests that Br2•- is formed through hydroxyl radicals via reactions 6 and 7, followed by reaction 3. In this sequence, at neutral pH, reaction 7 represents the ratedetermining step.
HO• + Br- f HOBr•-
(6)
HOBr•- f HO- + Br•
(7)
Table 1 displays the photolytic yields of Br2•- as a function of the KBr concentration. The yields steadily increase until ca. 4 × 10-3 M KBr, likely reflecting the competition of KBr and unreacted NANTE for HO• radicals. A further slight increase in the Br2•- yields is noted for high KBr concentrations between 0.1 and 0.5 M. Correction for Two-Photon Processes at 266 nm. It is known that high intensities of laser light (>1011 W
m-2) can ionize water through two-photon absorption according to the general reaction 8 (23, 24).
H2O + 2hν f H2O* f eaq-, H3O+, HO•, H•
(8)
Experimental evidence will show that these processes are likely of lower importance under our experimental conditions. We obtained significant yields of Br2•- at laser intensities way below 1011 W m-2. Figure 4 displays the yields of Br2•- obtained after 266 nm LFP of 9.4 × 10-5 M NANTE with laser energies of 110 (trace a) and 9 mJ (trace b) per pulse. Trace b was obtained with a laser intensity of 1.8 × 1010 W m-2, i.e., a laser intensity an order of magnitude lower than usually required for the two-photon ionization of water. More importantly, Go¨rner and Nikogosyan showed that HO• radical-dependent products under conditions of two-photon absorption of water increased ca. 1.8-fold when solutions were saturated with N2O instead of Ar (due to the conversion of hydrated electrons into HO• radicals; reaction 9) (24).
eaq- + N2O + H2O f N2 + HO• + HO-
(9)
Figure 5 displays the relative photolytic yields of Br2•for 4 × 10-3 M KBr in Ar-saturated 10-2 M phosphate buffer, pH 6.9, alone (trace a), in an Ar-saturated containing 9.4 × 10-5 M NANTE and 10-2 M phosphate (trace b), and in an N2O-saturated solution containing 9.4 × 10-5 M NANTE and 10-2 M phosphate (trace c). Clearly, we do not see a 1.8-fold increase of the photolytical yields of Br2•- under saturation with N2O (trace c) as compared to Ar (trace b) even with high laser intensities of 2.2 × 1011 W m-2. Importantly, trace a only reflects the yield of Br2•- via two-photon ionization of water in the absence of NANTE. In the presence of NANTE, most of the incident light is absorbed by NANTE so that the probability for a two-photon ionization of water is significantly lower. Therefore, in Figure 5, trace a cannot be subtracted from trace b. In a first approximation, we assume that trace b is predominantly the result of NANTE photolysis and that N2O saturation enhances the
1230
Chem. Res. Toxicol., Vol. 17, No. 9, 2004
Nauser et al.
Figure 3. Absorbance vs time profile for the formation of Br2•- after LFP of an Ar-saturated solution of 9.4 × 10-5 M NANTE in 10-2 M potassium phosphate (pH 6.7) containing 4 × 10-3 M KBr. The profile is fit according to first-order kinetics with t1/2 ) 150 ns (k ) 4.6 × 106 s-1). The residuals of the first-order fit are shown around the x-axis. Table 1. Experimental Yields of Br2•- after 266 nm LFP (110 mJ Laser Energy) of an Ar-Saturated Aqueous Solution Containing 9.4 × 10-5 M NANTE, 2 × 10-2 Phosphate (pH 6.9), and Various Concentrations of KBr [KBr] (10-3 M)
A360
[Br2•-] (10-6 M)
0.2 0.5 2.0 4.0 100 200 500
0.0124 0.017 0.026 0.031 0.042 0.042 0.049
1.25 1.71 2.64 3.17 4.26 4.23 4.96
yield of HO• radicals ca. 1.3-fold. This fact could indicate some photoionization of NANTE generating hydrated electrons, which are converted into HO• radicals by reaction with N2O (reaction 9). More evidence for a singlephoton process is derived from Br2•- quantum yields at low laser intensities, as described below. Quantum Yield of Br2•- at Low Laser Intensities. Quantum yields of Br2•- were representatively determined for 266 nm LFP with laser energies of 9 (1.8 × 1010 W m-2) and 30 mJ/pulse (6 × 1010 W m2), according to eq I.
Φ)
cproducts cabsorbedphotons
(I)
In eq I, cabsorbed photons ) ctotalphotons (1 - transmission of sample at 266 nm); that is
cabsorbedphotons ) Epulse × λ (1 - 10-[NANTE] × 266 × llaser) h × c × NA × Virradiated (II) In eq II, the symbols/abbreviations are defined as follows: Epulse, laser energy; h, Planck constant; c, concentration of absorber; NA, Avogadro’s number; Virradiated, irradiated sample volume; and llaser, optical path length within the cell. The quantum yields for Br2•- formation during 266 nm LFP of an Ar-saturated aqueous solution, containing 9.4 × 10-5 M NANTE and 2 × 10-2 M phosphate buffer, pH 6.9, are calculated to Φ ) 0.04
(30 mJ laser energy) and 0.045 (9 mJ laser energy). The close agreement of quantum yields for both laser energies is consistent with a single-photon process. Photolysis at 266 nm of NANTE at Acidic pH. In 2 × 10-3 M H2SO4, pH 2.7, the UV spectrum of NANTE shows two maxima at 277 and 358 nm, respectively. The LFP of an air-saturated solution containing 2 × 10-3 M H2SO4, 9.4 × 10-5 NANTE, and 4 × 10-3 M KBr with 266 nm light results in the formation of Br2•-. Figure 6 displays a competition plot, obtained with 110 mJ laser energies, where A0 and A are the absorbances at 360 nm for [Br2•-]0 and [Br2•-], which represent the yields of Br2•in the absence and presence of various concentrations of MeOH. The experimental slope of the least-squares fit for the plot of A0/A vs [MeOH]/[KBr] gives mexp ) 0.15 ( 0.03. A comparable plot was obtained for laser energies of 30 mJ, where mexp ) 0.11 ( 0.02 (data not shown). In Scheme 2, all reactions leading to the formation of Br2•- or the oxidation of MeOH by free hydroxyl radicals are shown, together with the known rate constants (21), which lead to the general eq III for the formation of Br2•in the presence of MeOH.
A0 )1+ A
k11[MeOH]
(III)
k7 + k10[H+]
k6[Br-]
k-6 + k7 + k10[H+]
From eq III, a calculated slope mcalcd ) k11/k6(k7 + k10[H+])/(k-6 + k7 + k10[H+] ) 0.12 is calculated, which corresponds well to the average value of the experimental slopes, mexp,av ) 0.13, obtained for both laser energies. The close agreement of mexp and mcalcd is consistent with the oxidizing species being a free HO•. At pH 2.7, NANTE shows significantly less absorbance at λ > 400 nm as compared with NANTE at pH g 6.9. Hence, the LFP experiments at pH 2.7 give the possibility to characterize intermediates in this spectral region. Figure 7 shows the UV/vis spectrum recorded 120-320 ns after 266 nm LFP of 9.4 × 10-5 M NANTE in an airsaturated solution, pH 2.7. The spectrum shows two absorbance maxima with λmax ca. 300 nm and λmax ) 430 nm. Approximately 75% of the initial 310 nm absorbance
UV Photolysis of 3-Nitrotyrosine
Chem. Res. Toxicol., Vol. 17, No. 9, 2004 1231
Figure 4. Yields of Br2•- obtained after 266 nm LFP with laser energies of 110 (trace a) and 9 mJ (trace b) per pulse in Arsaturated aqueous solutions, containing 9.4 × 10-5 NANTE, 2 × 10-2 M phosphate buffer (pH 6.7), and 4 × 10-3 M KBr.
Figure 5. Relative yields of Br2•- obtained after 266 nm LFP of (a) Ar-saturated aqueous solution containing 2 × 10-2 M phosphate buffer (pH 6.7) and 4 × 10-3 M KBr; (b) Ar-saturated aqueous solution containing 9.4 × 10-5 NANTE, 2 × 10-2 M phosphate buffer (pH 6.7), and 4 × 10-3 M KBr; and (c) N2O-saturated aqueous solution containing 9.4 × 10-5 NANTE, 2 × 10-2 M phosphate buffer (pH 6.7), and 4 × 10-3 M KBr.
decays within 3 µs (t1/2 ≈ 520 ns; k ≈ 1.3 × 106 s-1), whereas the remainder of the absorbance is stable over the following 15 µs. In contrast, the 430 nm absorbance decays significantly slower. The 430 nm absorbance would be consistent with the formation of phenoxyl radicals (25). In the 400-540 nm region, para-substituted phenoxyl radicals show absorbances with absorbance coefficients between 1000 and 5000 M-1 cm-1 (25). If for the NANTE-derived phenoxyl radical the absorbance coefficient were around the mean of this range, i.e., 430 ≈ 3000 M-1 cm-1, the experimental absorbance of 0.01 (obtained at pH 2.7 with a laser energy of 0.11 J) would yield a phenoxyl radical concentration of 3.3 × 10-6 M. This would nicely correspond to 3.1 × 10-6 M Br2•measured in the presence of 4 × 10-3 KBr at pH 6.9 with
a laser energy of 0.11 J (Figure 4, trace a). One might, however, expect that part of the hydroxyl radicals would generate additional phenoxyl radicals in a subsequent process (reaction 12). Hence, we would expect that 430 < 3000 M-1 cm-1 (vide infra).
R-Ar-OH + HO• f R-Ar-O• + H2O
(12)
To obtain reference spectra, we reacted NANTE in a pulse radiolysis experiment directly with hydroxyl radicals or SO4•-, a typical one-electron oxidant. Figure 8, trace a, displays the UV spectrum obtained for the reaction of NANTE with hydroxyl radicals, recorded 20 µs after pulse irradiation of an N2O-saturated aqueous solution containing 3 × 10-4 NANTE and 10-4 M HClO4.
1232
Chem. Res. Toxicol., Vol. 17, No. 9, 2004
Nauser et al.
Figure 6. Plot of A0/A vs [CH3OH]/[KBr] for Ar-saturated aqueous solutions containing 9.4 × 10-5 NANTE, 2 × 10-2 M phosphate buffer (pH 6.7), and various concentrations of MeOH and KBr.
Scheme 2
This spectrum is characterized by two absorption maxima around 340 and 430 nm. Taking G(HO•) ) 5.5 (26), we calculate that 430 ) 2480 M-1 cm-1 (based on an applied dose of 22 Gy, a measured absorbance of A430 ) 0.031, and the assumption that 100% of the hydroxyl radicals contribute to the formation of the 430 nm absorbance). On the other hand, the reaction of NANTE with SO4•leads to a different spectrum with an absorption maximum at 409 nm, displayed in Figure 8, trace b. At 20 µs after the pulse, the reaction of SO4•- with NANTE is completed (t1/2 ≈ 4 µs under our reaction conditions: 3 × 10-4 M NANTE, 0.1 M K2S2O8, 0.1 M tert-butyl alcohol, 10-4 M HClO4, Ar), based on parallel kinetic experiments following the disappearance of SO4•-, which has a wellcharacterized absorbance at 450 nm (27). A comparison with Figure 7 reveals that the photochemically generated transient absorbing at 430 nm strongly resembles the species formed after the reaction of NANTE with hydroxyl radicals. On the other hand, the spectra displayed in Figures 7 and 8, trace a, differ in the region
