Photodynamic treatment and H 2 O 2 -induced oxidative stress result in different patterns of cellular protein oxidation Dmitri V. Sakharov 1 , Anton Bunschoten 1 , Huib van Weelden 2 and Karel W. A. Wirtz 1 1 Department of Biochemistry of Lipids, Centre for Biomembranes and Lipid Enzymology, Institute of Biomembranes, Utrecht University, the Netherlands; 2 Department of Photodermatology, University Medical Center Utrecht, the Netherlands Photodynamic treatment (PDT) is an emerging therapeutic procedure for the management of cancer, based on the use of photosensitizers, compounds that generate highly reactive oxygen species (ROS) on irradiation with visible light. The ROS generated may oxidize a variety of bio- molecules within the cell, loaded with a photosensitizer. The high reactivity of these ROS restricts their radius of action to 5–20 nm from the site of their generation. We studied oxidation of intracellular proteins during PDT using the ROS-sensitive probe acetyl-tyramine-fluorescein (acetylTyr-Fluo). This probe labels cellular proteins, which become oxidized at tyrosine residues under the conditions of oxidative stress in a reaction similar to dityrosine for- mation. The fluorescein-labeled proteins can be visualized after gel electrophoresis and subsequent Western blotting using the antibody against fluorescein. We found that PDT of rat or human fibroblasts, loaded with the photosensitizer Hypocrellin A, resulted in labeling of a set of intracellular proteins that was different from that observed on treatment of the cells with H 2 O 2 . This difference in labeling patterns was confirmed by 2D electrophoresis, showing that a lim- ited, yet distinctly different, set of proteins is oxidized under either condition of oxidative stress. By matching the Western blot with the silver-stained protein map, we infer that a-tubulin and b-tubulin are targets of PDT-induced protein oxidation. H 2 O 2 treatment resulted in labeling of endoplasmic reticulum proteins. Under conditions in which the extent of protein oxidation was comparable, PDT caused massive apoptosis, whereas H 2 O 2 treatment had no effect on cell survival. This suggests that the oxidative stress generated by PDT with Hypocrellin A activates apoptotic pathways, which are insensitive to H 2 O 2 treatment. We hypothesize that the pattern of protein oxidation observed with Hypocrellin A reflects the intracellular localization of the photosensitizer. The application of acetylTyr-Fluo may be useful for characterizing protein targets of oxidation by PDT with various photosensitizers. Keywords: apoptosis; Hypocrellin A; photodynamic treat- ment; protein oxidation; tubulin. Photodynamic therapy is an emerging modality for the treatment of cancer [1]. It is based on the killing of tumor cells by light-activatable photosensitive compounds, or photosensitizers. In the presence of oxygen, the combination of visible light and a photosensitizer causes generation of singlet oxygen and other cytotoxic reactive oxygen species (ROS), such as superoxide anions and the extremely reactive hydroxy radical [2]. The higher uptake of photosensitizers by cancerous tissues compared with normal tissues, and the possibility of local illumination of tumors are essential for selective eradication of tumor cells with photodynamic treatment (PDT). The mode of cell death by PDT may be either apoptosis or necrosis, depending on the nature and concentration of the photosensitizer and the amount of irradiation [3,4]. Although the signaling pathways activated in response to PDT are partly delineated and the sequence of apoptotic events induced by PDT is well described [3–8], the specific cellular targets of PDT and critical early events involved in triggering PDT-induced apoptosis are not clear [4,9]. Different photosensitizers have different intracellular locali- zations. Singlet oxygen and hydroxy radical, the most reactive photodynamically generated species, have extre- mely short lifetimes (less than 1 ls) in the intracellular environment, and therefore their sphere of influence is very small, not more than 20 nm from the site of their generation [2]. In this way, the intracellular localization of the photosensitizer determines the areas of its photodynamic action. Molecular targets of oxidation may therefore also vary depending on the localization of the photosensitizer [2,3]. PDT may damage proteins, lipids, DNA, and a variety of small molecules in the cell [2]. Recent data [10] suggest that cellular proteins are a likely key target for toxicity mediated by singlet oxygen. Oxidative assault may cause modifica- tions of the side chains of amino acids within a protein. In particular, the side chains of cysteine, histidine, methionine, tryptophan and tyrosine are susceptible to ROS-induced modifications [2,11,12]. Modifications of tyrosine are of particular interest, because it is critically involved in intra- cellular signal transduction via tyrosine phosphorylation. Correspondence to D. V. Sakharov, CBLE, Utrecht University, PO Box 80.054, 3508 TB Utrecht, the Netherlands. Fax: + 31 30 2533151, Tel.: + 31 30 2532852, E-mail: d.sakharov@chem.uu.nl Abbreviations: PDT, photodynamic treatment; ROS, reactive oxygen species; acetylTyr-Fluo, acetyl-tyramine-fluorescein; ECL, enhanced chemiluminescence. (Received 29 July 2003, revised 8 October 2003, accepted 21 October 2003) Eur. J. Biochem. 270, 4859–4865 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03885.x Interactions of tyrosine with ROS may result in generation of tyrosyl radicals, which can dimerize to yield dityrosine [13,14]. It has been shown that tyrosyl radicals, and eventually dityrosine, are formed as a result of PDT of tyrosine, probably in a reaction mediated by singlet oxygen [15]. We have recently developed a technique that utilizes a probe, acetyl-tyramine-fluorescein (acetylTyr-Fluo), which allows detection and identification of intracellular proteins that become oxidized at tyrosine residues under the conditions of oxidative stress [16,17]. Using this technique, we have shown that the proteins of the endoplasmic reticulum are the major targets of oxidation induced by treatment of cells with H 2 O 2 [18]. In this study, we used this technique in combination with 2D electrophoresis to assess the oxidation of proteins in cells subjected to PDT with the photosensitizer Hypocrellin A. Hypocrellins are stucturally related to polycyclic quinones. They show extremely high phototoxicity towards tumors and viruses and are being explored for a variety of therapeutic applications [19–21]. We have found that PDT with Hypocrellin A oxidizes a distinct set of cellular proteins, including tubulins, which are not oxidized by treatment of the cells with H 2 O 2 . Experimental procedures Materials Hypocrellin A and Hoechst 33342 were purchased from Molecular Probes (Leiden, the Netherlands). Rose Bengal, carbonic anhydrase, H 2 O 2 , propidium iodide and tyramine were from Sigma. Polyclonal antibody against fluorescein, conjugated to horseradish peroxidase, was purchased from Biogenesis (Poole, Dorset, UK). Tyramine-fluorescein (Tyr- Fluo) and acetylTyr-Fluo were synthesized as described elsewhere [16]. Photodynamic treatment Specimens containing either cells or solutions of purified components were illuminated with visible light from a slide projector equipped with a 250 W tungsten lamp. The purple and blue part of the light spectrum with k< 470 nm was cut off by a short-cut filter. The fluence rate in the irradiation area was 10 mWÆcm )2 . To reach the fluence of 1JÆcm )2 and 2 JÆcm )2 , the specimens were irradiated for 100 s and 200 s, respectively. The fluence rate was measured with a specially modified and calibrated photometer (Waldmann AG, Schwenningen, Germany). Assessment of dityramine formation caused by PDT Photosensitizers at a final concentration of 10 l M were added to a well of a plastic culture plate containing 1 m M tyramine in NaCl/P i , pH 7.4. The wells were irradiated with visible light as described above. Dityramine formation was assessed by measuring a characteristic fluorescence signal of dityramine (excitation maximum at 315 nm, emission maximum at 405 nm). Some of the samples were also analyzed by electrospray MS using a Quattro Ultima mass spectrometer. Photodynamic labeling of a model protein with the Tyr-Fluo probe A solution containing 0.4 mgÆmL )1 carbonic anhydrase and 10 l M Tyr-Fluo in NaCl/P i was irradiated with visible light in either the presence or absence of a photosensitizer (Rose Bengal or Hypocrellin A, 10 l M ). The samples were subjected to SDS/PAGE and Western blotting with anti- body against fluorescein. Cell culture and PDT Rat-1 fibroblasts or adult normal human dermal fibro- blasts were cultured in Dulbecco’s modified Eagle’s medium with 7.5% fetal bovine serum at 5% CO 2 (v/v) in the presence of penicillin and streptomycin. The experiments were performed with 70–80% confluent cells growing in 10 cm Petri dishes. For the experiments involving microscopy, the cells were grown in glass- bottomed 3.5-cm dishes (Willco Wells, Amsterdam, the Netherlands). Most of the experiments were performed with Rat-1 fibroblasts, which are easier to culture. Key experiments, in particular those involving 2D-PAGE, were also repeated with human fibroblasts, because their detailed protein map has been published. Hypocrellin A was loaded into the cells in the culture medium for 3 h. Then the medium with photosensitizer was removed, and the cells were incubated for 15 min with acetylTyr-Fluo (5 l M )inNaCl/P i supplemented with 0.9 m M CaCl 2 , 0.5 m M MgCl 2 ,and5m M glucose (NaCl/P i +). After removal of NaCl/P i + containing acetylTyr-Fluo, fresh NaCl/P i + was added and the cells were irradiated with visible light as described above. Immediately after irradi- ation, the cells were rinsed with a salt-free isotonic buffer (0.25 M sucrose, 1 m M EDTA, and 20 m M Tris/HCl, pH 7.4) and lysed in buffer containing 20 m M Tris/HCl (pH 7.4), 1 m M EDTA, 1% Triton X-100 and a cocktail of protease inhibitors (Sigma P-8340) diluted 1 : 40. In some experiments, cells loaded with acetylTyr-Fluo as described above were treated with H 2 O 2 in NaCl/P i +for 15 min and lysed. Detection of cellular proteins susceptible to oxidation Cell lysates were subjected to either SDS/PAGE under redu- cing conditions in 10% polyacrylamide gels or 2D-PAGE. Isoelectrofocusing, the first step of the 2D-PAGE, was per- formed on 11 cm-long Bio-Rad IPG strips (ReadyStrip TM ), pH 3–10, according to the manufacturer’s instructions, using a Protean IEF Cell. SDS/PAGE in the second direction was run under reducing conditions in 15% polyacrylamide gel with 0.08% bisacrylamide. 1D PAGE gels were blotted on to a nitrocellulose membrane and subjected to Western blotting with peroxidase-conjugated antibody against fluorescein to detect the Tyr-Fluo-labeled proteins. An enhanced chemiluminescence (ECL) kit from Bio-Rad was used to visualize the labeled spots. 2D gels were either stained with silver or subjected to Western blotting, as described above for 1D gels. After blotting of the two-dimensional gels (before blocking of the membrane and application of the antibody), the membranes were stained with Ponceau Red and scanned. 4860 D. V. Sakharov et al.(Eur. J. Biochem. 270) Ó FEBS 2003 To colocalize the labeled spots on the ECL film with the spots on the silver-stained gels, a composite image file was created, containing the spots labeled with fluorescein (oxidized proteins, detected by ECL after Western blotting) and 7–10 major spots visible on Ponceau-stained mem- branes. PDQUEST software was used to edit the images of silver-stained gels and spots from the ECL films. Adobe Photoshop software was used to rescale the images and fit the major spots of the silver-stained gel to the corresponding Ponceau-stained spots on the membrane. In this way, it was possible to match the ECL-detected spots to the corres- ponding spots on the silver-stained gels. MS (peptide mass fingerprints of trypsin digests of the spots of interest obtained with MALDI-TOF, followed by a database search with the Mascot software for peptide mapping result) and matching of our protein maps to the published protein maps of human fibroblasts, available at the Human 2D-PAGE Databases of the Danish Centre for Human Genome Research (http://cancer.proteomics.dk), were used to identify the protein spots of interest in the silver-stained gels. Fluorescence/confocal microscopy Nikon Eclipse TE2000-U microscope, equipped with both conventional fluorescence appliances and confocal laser scanning C1 unit, was used in this study. Hypocrellin A distribution before and after PDT was assessed using the confocal mode with excitation at 543 nm from a HeNe laser. For the immunofluorescence detection of tubulin, the cells subjected to PDT were briefly incubated with propi- dium iodide for 3 min, fixed with methanol at )20 °Cfor 5 min, permeabilized with 0.1% (v/v) Triton X-100 in NaCl/P i for 15 min, and stained with Cy3-labeled tubulin antibody (Sigma). For assessment of the viability, the cells were stained with a mixture of Hoechst 33342 and propidium iodide (both at 2 lgÆmL )1 in the culture medium), and fluorescence images were taken using the conventional fluorescence mode. Cell morphology was documented by differential interference contrast. Results In this study we focused on the detection of tyrosine oxidation of the intracellular proteins on oxidative stress induced by PDT of the cells. A tyrosine analogue, tyramine, coupled covalently to fluorescein (Tyr-Fluo), was used as a probe to label the cellular proteins susceptible to this type of oxidative modification. On oxidation of the tyramine moiety by ROS, tyramine is converted into a tyrosyl radical that can form crosslinks Fig. 1. Photosensitized formation of dityramine. Asolutionof1m M tyramine was irradiated with visible light (2 JÆcm )2 ) in either the presence or absence of a photosensitizer (Rose Bengal or Hypocrel- lin A, 10 l M ). Formation of dityramine was assessed by measuring fluorescence with the characteristic spectra of dityramine (excitation maximum at 315 nm, emission maximum at 405 nm). 1, Rose Bengal with light; 2, Hypocrellin A with light; 3, Rose Bengal without light; 4, Hypocrellin A without light; 5, no photosensitizer with light. Fig. 2. Photosensitized labeling of carbonic anhydrase with tyramine- fluorescein. A solution containing 0.4 mgÆmL )1 carbonic anhydrase and 10 l M Tyr-Fluo was irradiated with visible light (2 JÆcm )2 )in either the presence or absence of a photosensitizer (Rose Bengal or Hypocrellin A, 10 l M ). The samples were subjected to SDS/PAGE and Western blotting with an antibody against fluorescein. Lane 1, no photosensitizer; 2, no photosensitizer with light; 3, Rose Bengal; 4, Rose Bengal with light; 5, Hypocrellin A; 6, Hypocrellin A with light. Fig. 3. Labeling of cellular proteins on PDT and treatment with H 2 O 2 . Lane 1, control cells loaded with acetylTyr-Fluo, no treatment; 2, cells were loaded with acetylTyr-Fluo and irradiated with visible light at 1JÆcm )2 (no photosensitizer control). Lanes 3 and 4, cells were loaded with Hypocrellin A (1 l M and 2 l M , respectively), then with acetyl- Tyr-Fluo, and were finally irradiated with visible light (1 JÆcm )2 ). Lane 5, cells were loaded with 2 l M Hypocrellin A, then with acetyl- Tyr-Fluo, and were not irradiated (no light control). Lane 6, cells were loaded with acetylTyr-Fluo and then treated with 200 l M H 2 O 2 .Cell lysates were subjected to SDS/PAGE and Western blotting with antibody against fluorescein. Ó FEBS 2003 Protein oxidation on photodynamic treatment (Eur. J. Biochem. 270) 4861 with oxidized tyrosine residues in a target protein. In the first experiments, we assessed whether PDT can cause dityrosine (dityramine) formation, and the covalent coup- ling of the Tyr-Fluo to a model protein. Figure 1 shows that dityramine is formed on PDT of a solution of tyramine with either Rose Bengal or Hypocrel- lin A as photosensitizer. Dityramine formation was docu- mented by generation of a fluorescent signal with a characteristic spectrum (maximum of the excitation spec- trum at 315 nm and a maximum of the emission spectrum at 405 nm). MS (not shown) also confirmed generation of dityramine on PDT with Hypocrellin A. No dityramine was formed in the absence of either light or photosensitizer. Figure 2 shows that PDT in the presence of either Rose Bengal or Hypocrellin A causes labeling of carbonic anhydrase with Tyr-Fluo. Labeling was dependent on the concentration of the photosensitizer used (not shown). Rose Bengal caused stronger labeling than Hypocrellin A. In further experiments, Hypocrellin A was used because Rose Bengal does not accumulate in the cell. Irradiation of rat fibroblasts, loaded with both Hypo- crellin A and acetylTyr-Fluo, resulted in the labeling of cellular proteins, as shown in Fig. 3. PDT-induced protein labeling was dependent on the concentration of the photosensitizer (Fig. 3, lanes 3 and 4) and the dose of irradiation (not shown). The pattern of labeling in the cells treated with PDT was different from that obtained with the cells treated with H 2 O 2 . 2D-PAGE in combination with Western blotting was applied to resolve the difference in the protein labeling. These experiments were performed with both rat (not shown) and human fibroblasts with similar results. 2D-PAGE images obtained with human fibroblasts are presented in Fig. 4. Only a limited number of proteins were labeled on PDT and H 2 O 2 treatment, but the patterns of protein labeling were distinctly different (Fig. 4A,B). Matching the blot with the protein map shows that PDT caused labeling of a-tubulin and b-tubulin (spots 1 and 2). The minor spot 3 probably reflects labeling of a small fraction of actin. The rest of the spots remain to be identified. We could not detect any labeled spots in the control samples obtained from cells either loaded with the photosensitizer but not irradiated or irradiated in the absence of the photosensitizer. As for treatment with H 2 O 2 , the labeling pattern agreed with the results of our previous study [18], which showed labeling of endoplasmic reticulum proteins (Bip, spot 4; PDI, spot 5; GPP58, spot 6). Careful assessment of the general changes of the protein map on PDT was beyond the scope of this study. Under the conditions of the experiment presented in Fig. 4, the protein map did not change dramatically, although some of the spots in the silver-stained gels were upregulated or down- regulated in PDT-treated samples. PDT at higher concen- trations of the photosensitizer had a dramatic effect on the protein map (not shown): many spots either disappeared or were spread along the horizontal axis of the gel. This was probably a result of photodynamic crosslinking of proteins [22,23]. Under these conditions, PDT resulted in rapid cell death (not shown). Figure 5A shows the subcellular localization of Hypo- crellin A in rat fibroblasts before irradiation. In agreement with other studies (reviewed in [19]), Hypocrellin A locali- zed mainly in lysosomes. We observed that it was also present throughout the cytoplasm, although to a lesser extent. Some of the photosensitizers have been shown to rapidly redistribute within the cell under irradiation, for instance to leak from lysosomes to cytosol [24,25]. It was not the case under the conditions used in this study. Under the conditions used in the experiment presented in Fig. 4, the distribution of Hypocrellin A did not change during and immediately after irradiation (not shown), implying that oxidation of cytoskeletal proteins is not a result of acute Fig. 4. 2D-PAGE detection of oxidized pro- teins in cells treated with PDT or H 2 O 2 . (A,C) Human fibroblasts were loaded with 1 l M Hypocrellin A, then with acetylTyr-Fluo, and were finally irradiated with visible light (1 JÆcm )2 ); (B,D) Cells were loaded with ace- tylTyr-Fluo and exposed to 200 l M H 2 O 2 for 10 min. Cell lysates were subjected to 2D- PAGE and either analyzed for the presence of oxidized proteins by Western blotting with antibody against fluorescein, or stained with silver. Oxidized proteins detected by Western blotting are shown in (A) and (B). Silver stainingisshownin(C)and(D)inblue superimposed with the spots of oxidized pro- teins shown in red. Protein labels: 1, a-tubulin; 2, b-tubulin;3,actin;4,PDI;5,BiP;6, GRP58. 4862 D. V. Sakharov et al.(Eur. J. Biochem. 270) Ó FEBS 2003 leakage of the photosensitizer from the sites of its primary localization into the cytosol. The results presented in Fig. 4 indicate that tubulin is a direct target of oxidation on PDT. To follow the fate of the microtubule network, we used immunofluorescence. Micro- tubule organization was already disturbed 5 min after PDT. At 1 l M Hypocrellin A, the tubulin network became less regular and less sharp (Fig. 5C) than in control cells (Fig. 5B). At a higher concentration of the photosensitizer, the microtubules were completely destroyed (Fig. 5D). Under the latter conditions (2 l M Hypocrellin A), the cells were not yet dead 5 min after PDT, as judged by the absence of staining with propidium iodide, but after 1 h most of the cells were dead through necrosis. OnPDTat1l M Hypocrellin A (conditions used in the experiment presented in Figs 4 and 5C), most cells became apoptotic 4 h after irradiation (Fig. 6A,C,E). Quantitative analysis of three independent experiments showed that only 6 ± 4% (mean ± SD) of the cells remained alive (normal cellular and nuclear morphology, no propidium iodide staining), 68 ± 28% were apoptotic (blebbing, condensed or fragmented nucleus, no propidium iodide staining), and 26 ± 16% were necrotic (characteristic necrotic morphology, propidium iodide staining of the nucleus). In the light-only and photosensitizer-only controls, there were practically no dead cells (less than 2% necrotic, no apoptotic cells). In contrast with PDT, treatment with H 2 O 2 did not result in noticeable cell death after 4 h (Fig. 6B,D,F) or 24 h (not shown). Discussion Oxidative stress induced by PDT can affect several types of biomacromolecules including proteins, lipids, and DNA [2]. A substantial body of evidence indicates that the cellular proteins are the key target of ROS-mediated toxicity [11,12,26] including singlet oxygen-mediated toxicity [10,26]. Oxidation of cellular proteins in response to PDT may be crucially involved in the mechanisms of PDT- induced cell death. Although a number of particular intracellular proteins have been shown to be modified as a result of PDT [27–29], little work has been done at the level of the whole cellular proteome in response to PDT. In the only available paper, Grebenova et al. [30] showed that a number of protein spots in the proteomic map of the HL60 cell lysates are significantly reduced after subjection of the cells to PDT Fig. 5. Distribution of Hypocrellin A in Rat-1 fibroblasts and effect of PDT on the microtubule network. (A) Rat-1 fibroblasts were loaded with Hypocrellin A under the conditions described in the legend to Fig. 4. The confocal image shows Hypocrellin A distribution before irradiation. No considerable change in the localization of Hypocrel- lin A was observed after irradiation (not shown). (B–D) Cells were loaded with 0 l M (B), 1 l M (C), or 2 l M (D) Hypocrellin A, irradiated with visible light (1 JÆcm )2 ), fixed with cold methanol 5 min after irradiation and stained with Cy3-labeled antibody against tubulin. Bar: 20 lm. Fig. 6. PDT, but not H 2 O 2 treatment, induces apoptosis. Rat-1 fibro- blasts were treated with either PDT (A,C,E) or H 2 O 2 (B,D,F) under the conditions described in the legend to Fig. 4, incubated in a CO 2 incubator for 4 h and stained with a mixture of Hoechst 33342 and propidium iodide. Differential interference contrast images (A,B) show apoptotic morphology (blebbing) in the most of the cells treated with PDT (A), but not in the cells treated with H 2 O 2 (B). Hoechst 33342 staining (C,D) allows the distinction between normal cells (large evenly stained nucleus, indicated with No) and apoptotic cells (condensed or fragmented nucleus, indicated with Ap). Staining with propidium iodide (E,F) indicates dead cells with permeabilized plasma membrane. Bar: 20 lm. Ó FEBS 2003 Protein oxidation on photodynamic treatment (Eur. J. Biochem. 270) 4863 with 5-aminolevulinic acid. In our study, we combined the proteomics approach with detection of proteins oxidized in response to PDT. We used a technique that utilizes an intracellular oxidation-sensitive probe, acetylTyr-Fluo, which labels proteins susceptible to oxidation at tyrosine residues. In a purified system we have shown that dityramine formation, the reaction essential for Tyr-Fluo labeling of proteins, can be induced by PDT of tyramine solution with the photosensitizers Hypocrellin A and Rose Bengal. Fur- thermore, a model protein was labeled with Tyr-Fluo by PDT with the same photosensitizers. Furthermore, in the cells, protein oxidation was observed, which was dependent on the concentration of the photosensitizer and on the illumination. 2D electrophoresis was further applied to determine which proteins are oxidized on PDT. We have previously shown that treatment of cells with H 2 O 2 causes oxidation of proteins localized in the endo- plasmic reticulum. This has been suggested to be a consequence of the specific redox status of the endoplasmic reticulum, facilitating local generation of radicals capable of inducing tyrosyl radical formation [31]. In this study, we observed a different pattern of protein labeling on PDT of cells loaded with Hypocrellin A. We hypothesize that this pattern reflects the cellular localization of Hypocrellin A. Hypocrellin A is a moderately hydrophobic substance, which localizes mainly to the membranes of various organelles. Labeling of cytoskeletal proteins (a-tubulin and b-tubulin, and slight labeling of actin) suggests that the cytoplasmic compartment is exposed to the oxidative stress generated by PDT with Hypocrellin A. This is in agreement with the partial presence of the photosensitizer throughout the cytoplasm (Fig. 5A). In a number of papers [32–36], deleterious effects of PDT on the microtubules have been documented. Under our experimental conditions (1 l M Hypocrellin A, irradiation at 1JÆcm )2 ), the microtubules were partly depolymerized immediately after PDT (Fig. 5C). Inactivation of the microtubules leads to the inability of the photosensitized cells to form functional mitotic spindles and finally results in the arrest at the G2/M phase of the cell cycle and subsequent apoptosis [32]. It has been hypothesized that the micro- tubules may be damaged within the radius of action of singlet oxygen in close proximity to the organelles in which photosensitizers accumulate (lysosomes, mitochondria, endoplasmic reticulum) [32,34]. Alternatively, an indirect mechanism has been suggested involving release of calcium caused by photodynamic insult and subsequent calcium- induced microtubule depolymerization [36]. In this paper, we show that PDT with Hypocrellin A results in direct oxidative modification of tubulin, and we hypothesize that this modification may be responsible for the PDT-induced impairment of microtubules. Further studies, including those in a purified system (reconstituted microtubules), will be needed to determine the sites of the oxidative modifica- tions within the tubulin molecule, and to elucidate the role of these modifications in the functional damage to tubulin. Interestingly, for the two modes of oxidative stress (PDT and H 2 O 2 treatment), the relationships between overall protein oxidation and cell death were dramatically different. For instance, treatment with 200 l M H 2 O 2 resulted in profound protein oxidation, but caused no cell death. PDT with 1 l M Hypocrellin A and illumination at 1 JÆcm )2 resulted in comparable protein oxidation (Figs 3 and 4), but the cells became massively apoptotic. This implies that the total degree of protein oxidation is not a critical determinant for the onset of apoptosis. Oxidation of endoplasmic reticulum proteins, occurring on treatment with H 2 O 2 , appears not to be critical for cell survival. Rather, oxidation of particular proteins in particular subcellular sites deter- mines the onset of apoptosis. Oxidation of other biomol- ecules, for instance lipid peroxidation, may also trigger cell death, mostly through rather unspecific mechanisms invol- ving damage to the cellular membranes. In contrast, oxidation of particular proteins may activate specific signaling pathways that regulate cell death or survival [27,29], which may be important at sublethal doses of PDT. In conclusion, we have shown for the first time that the pattern of intracellular protein oxidation depends on the kind of oxidative stress exerted. The methodology described here offers the possibility to identify the proteins oxidized under various forms of oxidative stress, including PDT with various photosensitizers localized to different cellular com- partments. It is hoped that this will allow the identification of photosensitizer-specific protein targets and will help to further elucidate the mechanisms of PDT-induced cell death. Acknowledgements The study was supported by NWO/ZON MW grant No 901-03-097. We are grateful to E. Romijn and C. Versluis for performing MS measurements, and to C. L. H. Guikers for assistance with PDT experiments. References 1. Dolmans, D.E., Fukumura, E. & Jain, R.K. (2003) Photodynamic therapy for cancer. Nat. Rev. Cancer 3, 380–387. 2. Sobolev, A.S., Jans, D.A. & Rozenkranz, A.A. (2000) Targeted intracellular delivery of photosensitizers. Prog. Biophys. Mol. Biol. 73, 51–90. 3. Moor, A.C.E. (2000) Signalling pathways in cell death and sur- vival after photodynamic therapy. J. Photochem. Photobiol. B 57, 1–13. 4. Vantieghem, A., Assefa, Z., Vandenabeele, P., Declercq, W., Courtois, S., Vandenheede, J.R., Merlevede, W., de Witte, P. & Agostinis, P. (1988) Hypericin-induced photosensitization of HeLa cells leads to apoptosis or necrosis. Involvement of cyto- chrome c and procaspase-3 activation in the mechanism of apoptosis. FEBS Lett. 40, 19–24. 5. Chan, W.H., Yu, J.S. & Yang, S.D. (2000) Apoptotic signalling cascade in photosensitized human epidermal carcinoma A431 cells: involvement of singlet oxygen, c-Jun N-terminal kinase, caspase-3 and p21-activated kinase 2. Biochem. J. 351, 221–232. 6. Granville, D.J., Carthy, C.M., Jiang, J.G., McManus, B.M., Matroulle,J.Y.,Piette,J.&Hunt,D.W.C.(2000)Nuclearfactor- jB activation by the photochemotherapeutic agent verteporfin. Blood 95, 256–262. 7. Assefa, Z., Vantieghem, A., Declercq, W., Vandenabeele, P., Vandenheede, J.R., Merlevede, W., de Witte, P. & Agostinis, P. (1999) The activation of the c-Jun N-terminal kinase and p38 mitogen-activated protein kinase signalling pathways protects HeLa cells from apoptosis following photodynamic therapy with hypericin. J. Biol. Chem. 274, 8788–8796. 8. Klotz, L.O., Fritsch, C., Briviba, K., Tsacmacidis, N., Schliess, F. & Sies, H. (1998) Activation of JNK and p38 but not ERK MAP 4864 D. V. Sakharov et al.(Eur. J. Biochem. 270) Ó FEBS 2003 kinases in human skin cells by 5-aminolevulinate-photodynamic therapy. Cancer Res. 58, 4297–4300. 9. Agostinis,P.,Vantieghem,A.,Merlevede,W.&deWitte,P.A.M. (2002) Hypericin in cancer treatment: more light on the way. Int. J. Biochem. Cell Biol. 34, 221–241. 10. Schafer, F.Q. & Buettner, G.R. (1999) Singlet oxygen toxicity is cell line-dependent: a study of lipid peroxidation in nine leukemia cell lines. Photochem. Photobiol. 70, 858–867. 11. Dean, R.T., Fu, S., Stocker, R. & Davies, M.J. (1997) Biochem- istry and pathology of radical-mediated protein oxidation. Bio- chem. J. 324, 1–18. 12. Berlett, B.S. & Stadtman, E.R. (1997) Protein oxidation in aging, disease, and oxidative stress. J. Biol. Chem. 272, 20313–20316. 13. Heinecke, J.W., Li, W., Daehnke, H.L. 3rd & Goldstein, J.A. (1993) Dityrosine, a specific marker of oxidation, is synthesized by the myeloperoxidase-hydrogen peroxide system of human neu- trophils and macrophages. J. Biol. Chem. 268, 4069–4077. 14. Pfeiffer, S., Schmidt, K. & Mayer, B. (2000) Dityrosine formation outcompetes tyrosine nitration at low steady-state concentrations of peroxynitrite. Implications for tyrosine modification by nitric oxide/superoxide in vivo. J. Biol. Chem. 275, 6346–6352. 15. Pecci, L., Montefoschi, G., Antonucci, A., Costa, M., Fontana, M. & Cavallini, D. (2001) Formation of nitrotyrosine by Methy- lene Blue photosensitized oxidation of tyrosine in the presence of nitrite. Biochem. Biophys. Res. Commun. 289, 305–309. 16. Van der Vlies, D., Wirtz, K.W.A. & Pap, E.H.W. (2001) Detection of protein oxidation in Rat-1 fibroblasts by fluorescently labeled tyramine. Biochemistry 40, 7783–7788. 17. Czapski, G.A., Avram, D., Sakharov, D.V., Wirtz, K.W., Stros- znajder, J.B. & Pap, E.H. (2002) Activated neutrophils oxidize extracellular proteins of endothelial cells in culture: effect of nitric oxide donors. Biochem. J. 365, 897–902. 18. Van der Vlies, D., Pap, E.H.W., Post, J.A., Celis, J.E. & Wirtz, K.W.A. (2002) Endoplasmic reticulum resident proteins of normal human dermal fibroblasts are the major targets for oxidative stress induced by hydrogen peroxide. Biochem. J. 366, 825–830. 19. Diwu, Z. (1995) Novel therapeutic and diagnostic applica- tions of hypocrellins and hypericins. Photochem. Photobiol. 61, 529–539. 20. Deininger, M.H., Weinschenk, T., Morgalla, M.H., Meyermann, R. & Schluesener, H.J. (2002) Release of regulators of angiogen- esis following Hypocrellin-A and -B photodynamic therapy of human brain tumor cells. Biochem. Biophys. Res. Commun. 298, 520–530. 21. Khoobehi, B., Grinstead, R. & Passos, E. (2002) Experimental photodynamic effects of hypocrellin A on the choriocapillaris. Ophthalmic Surg. Lasers 33, 207–213. 22. Verweij, H., Dubbelman, T.M. & Van Steveninck, J. (1981) Photodynamic protein cross-linking. Biochim. Biophys. Acta 647, 87–94. 23. Spikes, J.D., Shen, H.R., Kopeckova, P. & Kopecek, J. (1999) Photodynamic crosslinking of proteins. Kinetics of the FMN- and rose bengal-sensitized photooxidation and intermolecular crosslinking of model tyrosine-containing N-(2-hydroxypropyl) methacrylamide copolymers. Photochem. Photobiol. 70, 130–137. 24. Moan, J., Berg, K., Anholt, H. & Madslien, K. (1994) Sulfonated aluminium phthalocyanines as sensitizers for photochemotherapy. Effects of small light doses on localization of the dye fluorescence and photosensitivity in V79 cells. Int. J. Cancer 58, 865–870. 25. Georgakoudi, I. & Foster, T.H. (1998) Effects of the subcellular redistribution of two nile blue derivatives on photodynamic oxy- gen consumption. Photochem. Photobiol. 68, 115–122. 26. Davies, M.J. (2003) Singlet oxygen-mediated damage to proteins and its consequences. Biochem. Biophys. Res. Commun. 305, 761–770. 27. Agostinis, P., Vandenbogaerde, A., Donella-Deana, A., Pinna, L.A., Lee, K.T., Goris, J., Merlevede, W., Vandenheede, J.R. & De Witte, P. (1995) Photosensitized inhibition of growth factor- regulated protein kinases by hypericin. Biochem. Pharmacol. 49, 1615–1622. 28. Gantchev, T.G. & van Lier, J.E. (1995) Catalase inactivation following photosensitization with tetrasulfonated metallophthalo- cyanines. Photochem. Photobiol. 62, 123–134. 29. Usuda, J., Chiu, S M., Murphy, E.S., Lam, M., Nieminen, A L. & Oleinick, N.L. (2003) Domain-dependent photodamage to Bcl-2. A membrane anchorage region is needed to form the target of phthalocyanine photosensitization. J. Biol. Chem. 278, 2021– 2029. 30. Grebenova, D., Halada, P., Stulik, J., Havlicek, V. & Hrkal, Z. (2000) Protein changes in HL60 leukemia cells associated with 5-aminolevulinic acid-based photodynamic therapy. Early effects on endoplasmic reticulum chaperones. Photochem. Photobiol. 72, 16–22. 31. Hwang, C., Sinskey, A.J. & Lodish, H.F. (1992) Oxidized redox state of glutathione in the endoplasmic reticulum. Science 257, 1496–1502. 32. Vantieghem, A., Hu, Y., Assefa, Z., Piette, J., Vandenheede, J.R., Merlevede, W., de Witte, P. & Agostinis, P. (2002) Phoshorylation of Bcl-2 in G 2 /M phase-arrested cells following photodynamic therapy with hypericin involves a CDK1-mediated signal and delays the onset of apoptosis. J. Biol. Chem. 277, 37718–37731. 33. Berg, K. & Moan, J. (1997) Lysosomes and microtubules as tar- gets for photochemotherapy of cancer. Photochem. Photobiol. 65, 403–409. 34. Stockert, J.C., Juarranz, A., Villanueva, A. & Canete, M. (1996) Photodynamic damage to HeLa cell microtubules induced by thiazine dyes. Cancer Chemother. Pharmacol. 39, 167–169. 35. Lee, C., Wu, S.S. & Chen, L.B. (1995) Photosensitization by 3,3¢- dihexiloxacarcocyanide iodide: specific disruption of microtubules and inactivation of organelle motility. Cancer Res. 5, 2063–2069. 36. Sporn, L.A. & Foster, T.H. (1992) Photofrin and light induces microtubule depolymerization in cultured human endothelial cells. Cancer Res. 52, 3443–3448. Ó FEBS 2003 Protein oxidation on photodynamic treatment (Eur. J. Biochem. 270) 4865 . (1999) Photodynamic crosslinking of proteins. Kinetics of the FMN- and rose bengal-sensitized photooxidation and intermolecular crosslinking of model tyrosine-containing. infer that a-tubulin and b-tubulin are targets of PDT-induced protein oxidation. H 2 O 2 treatment resulted in labeling of endoplasmic reticulum proteins. Under