Oxidation of Catalase by Singlet Oxygen*

Different bands of catalase activity in zymograms (Cat-1a-Cat-1e) appear during Neurospora crassa development and under stress conditions. Here we demonstrate that singlet oxygen modifies Cat-1a, giving rise to a sequential shift in electrophoretic mobility, similar to the one observed in vivo. Purified Cat-1a was modified with singlet oxygen generated from a photosensitization reaction; even when the reaction was separated from the enzyme by an air barrier, a condition in which only singlet oxygen can reach the enzyme by diffusion. Modification of Cat-1a was hindered when reducing agents or singlet oxygen scavengers were present in the photosensitization reaction. The sequential modification of the four monomers gave rise to five active catalase conformers with more acidic isoelectric points. The pI of purified Cat-1a-Cat-1e decreased progressively, and a similar shift in pI was observed as Cat-1a was modified by singlet oxygen. No further change was detected once Cat-1e was reached. Catalase modification was traced to a three-step reaction of the heme. The heme of Cat-1a gave rise to three additional heme peaks in a high performance liquid chromatography when modified to Cat-1c. Full oxidation to Cat-1e shifted all peaks into a single one. Absorbance spectra were consistent with an increase in asymmetry as heme was modified. Bacterial, fungal, plant, and animal catalases were all susceptible to modification by singlet oxygen, indicating that this is a general feature of the enzyme that could explain in part the variety of catalases seen in several organisms and the modifications observed in some catalases. Modification of catalases during development and under stress could indicate in vivo generation of singlet oxygen.

Different bands of catalase activity in zymograms (Cat-1a-Cat-1e) appear during Neurospora crassa development and under stress conditions. Here we demonstrate that singlet oxygen modifies Cat-1a, giving rise to a sequential shift in electrophoretic mobility, similar to the one observed in vivo. Purified Cat-1a was modified with singlet oxygen generated from a photosensitization reaction; even when the reaction was separated from the enzyme by an air barrier, a condition in which only singlet oxygen can reach the enzyme by diffusion. Modification of Cat-1a was hindered when reducing agents or singlet oxygen scavengers were present in the photosensitization reaction. The sequential modification of the four monomers gave rise to five active catalase conformers with more acidic isoelectric points. The pI of purified Cat-1a-Cat-1e decreased progressively, and a similar shift in pI was observed as Cat-1a was modified by singlet oxygen. No further change was detected once Cat-1e was reached. Catalase modification was traced to a three-step reaction of the heme. The heme of Cat-1a gave rise to three additional heme peaks in a high performance liquid chromatography when modified to Cat-1c. Full oxidation to Cat-1e shifted all peaks into a single one. Absorbance spectra were consistent with an increase in asymmetry as heme was modified. Bacterial, fungal, plant, and animal catalases were all susceptible to modification by singlet oxygen, indicating that this is a general feature of the enzyme that could explain in part the variety of catalases seen in several organisms and the modifications observed in some catalases. Modification of catalases during development and under stress could indicate in vivo generation of singlet oxygen.
Photosynthetic evolution of dioxygen into the atmosphere and its subsequent accumulation led to formation of an ozone layer in the stratosphere, which permitted the dispersal of microorganisms around earth by absorbing damaging ultraviolet radiation from the sun (1,2). Atmospheric dioxygen also led to the evolution of adaptation mechanisms to live with a poisonous gas (3,4).
The electron affinity of O 2 makes it a reactive compound. Furthermore, dioxygen generates more reactive intermediates in its sequential univalent reduction to water. Likewise, singlet oxygen, excited states of O 2 , are highly reactive species. They arise upon absorption of radiation by O 2 , either directly or through prevalent cellular compounds such as tetrapyrrols, flavins, pterins, chlorophylls, and retinoids. Reactive oxygen species (ROS) 1 are inevitably produced in cells under aerobic or microaerobic conditions (4).
Primeval cells, which originated in an anoxic environment (2), had either to hide from O 2 , or to evolve mechanisms for efficient reduction of entering O 2 , disposal of ROS, and sequestering of transition metals, which participate in reducing O 2 into ROS. Respiration gave early aerobic heterotrophs a new source of energy, but more than anaerobes made them depend on the availability of reduced carbon, indispensable for O 2 reduction. Thus, since the generation of an oxidant atmosphere, the threat of damaging effects caused by ROS has been prevalent for all organisms. This threat was further utilized by some species that developed devices for the controlled production of ROS, as seen in many host-parasite relationships (5)(6)(7).
Because dioxygen and its unavoidable ROS have been of such importance for survival, cells developed mechanisms to exquisitely detect the presence of O 2 and ROS for regulation of metabolism and antioxidant responses (8 -13). Most interestingly, ROS became signals used by cells to regulate growth or proliferation (14), cell differentiation (15)(16)(17)(18) and death (14,19).
ROS are related to the arrest of growth and the start of cell differentiation. We have detected a hyperoxidant state at the start of all three morphogenetic transitions of Neurospora crassa asexual development (conidiation) (15-17, 20 -23). Increased generation of ROS leads to specific oxidation of some enzymes and massive protein oxidation and degradation (20,24,25). Specific modifications induced by ROS have been detected in protein that bind Fe(II) directly (24,26), bind a Fe(II) chelate complex (25), or have a noncatalytic iron-sulfur cluster (27). These modifications inactivate the enzymes and make them more susceptible to endogenous proteolytic activity (27)(28)(29)(30).
Safe disposal of H 2 O 2 in cells is carried out by catalases and peroxidases. Hydrogen peroxide is formed mainly from dismutation of superoxide, which is generated by O 2 capturing an electron, usually from electron transport chains. Hydrogen peroxide is also a product of some oxidases. Being uncharged and not very reactive, it can diffuse between cell compartments. Its toxicity is traced to the formation of the hydroxyl radical upon capture of an electron, for instance, from Fe(II) or Cu(I). The hydroxyl radical, one of the most reactive species known, reacts immediately with almost any cellular compound giving rise to alterations, such as modified or broken proteins and nucleic acids (4). Thus, it is understandable why catalases are one of the most efficient enzymes known (31). It is so efficient that it cannot be saturated by H 2 O 2 at any concentration.
Catalases are prevalent in most organisms (4). Many microorganisms have more than one catalase, and in some a catalase is related to cell differentiation (32)(33)(34)(35)(36)(37)(38)(39)(40)(41)(42)(43)(44)(45). Two catalase genes have been cloned in Aspergillus nidulans, one expressed only for asexual spores (conidia) (39,40). A third catalase activity has been recently found. 2 Neurospora crassa has two 3 or three (46) catalase genes. The cat-1 gene mapped to complementation group IIIR has been cloned and partially sequenced (72). It codes for a homotetramer of 320-kDa molecular mass. Cat-1 is present in the whole vegetative life cycle of N. crassa. 3 The cat-2 gene, assigned to complementation group VIIR, has not been cloned. It is expressed under stress conditions and transiently during the formation of conidia. 3 Catalase-specific activity increases stepwise with each morphogenetic transition of the N. crassa conidiation process. 3 Cat-1 was modified in these transitions and under stress conditions. 3 Since we have demonstrated enzyme inactivation due to ROS-specific oxidation under cell differentiation and stress conditions (24,25), it was important to determine if antioxidant enzymes such as catalases were also vulnerable to in vivo alteration by ROS. As predicted (15,16), these enzymes are susceptible to oxidation by ROS but not affected in their activity.
We show that Cat-1 was oxidized through a sequential reaction of the four monomers with singlet oxygen, giving rise to active catalase conformers with more acidic isoelectric points. Modification could be traced to a three-step reaction of the heme with singlet oxygen. Catalases from different organisms were similarly modified by singlet oxygen, indicating a general feature of the enzyme that could explain in part the variety of catalases seen in several organisms (47)(48)(49)(50)(51) and the observed modifications in some catalases (52)(53)(54)(55).

EXPERIMENTAL PROCEDURES
Strains and Cultures-N. crassa, wild type strain 74-OR23-1A from the Fungal Genetic Stock Center, was grown at 30°C in slants of agar minimal medium of Vogel, supplemented with 1.5% sucrose. A. nidulans, wild type strain FGSC26 (biA1; veA1; Fungal Genetic Stock Center), was grown for 18 h at 200 rpm, from a inoculum of conidia (10 5 /ml), in Kä fer's minimal nitrate medium, supplemented with 2.5 g/liter biotin and 1% glucose. Saccharomyces cerevisiae, wild type strain S288C, was grown in yeast extract/peptone/dextrose liquid medium at 30°C for 40 h at 250 rpm. Streptomyces coelicolor, wild type strain A3(2) J801, was grown 10 days at 200 rpm, from an inoculated Hopwood's minimal medium with an initial culture, grown from spores, in the same medium.
Catalase Activity-Catalase activity was measured by determining the initial rate of dioxygen production with a Clark microelectrode (56). Reaction was started by injecting catalase, usually 5 l or less, into a sealed chamber filled with 2 ml of 10 mM of H 2 O 2 in 10 mM phosphate buffer (PB), pH 7.8 (adjusted by mixing Na 2 HPO 4 and KH 2 PO 4 solutions). Units are defined as micromoles of O 2 produced per min per mg of protein under these conditions. Activity was measured in samples just before loading them on a gel for electrophoresis.
Catalase activity in polyacrylamide gels was determined by incubating the gel after electrophoresis, 5 min in 5% methanol and then, after rinsing three times with tap water, 10 min in 10 mM H 2 O 2 . The gel, rinsed with tap water, was incubated in a 1/1 mixture of freshly prepared 2% potassium ferric cyanide and 2% ferric chloride. Blue color developed in the gel except at zones where H 2 O 2 was decomposed by catalase (57). Staining was stopped by soaking the gel in a 10% acetic acid and 5% methanol solution.
Purification of Catalase-N. crassa, A. nidulans, S. cerevisiae, and S. coelicolor cells were homogenized 5 times for 30 s in a Bead-Beater with glass beads (710 -1180 m for the fungi and 150 -212 m for the bacterium) at a ratio of 1 g of dried weight per 7.5 ml of 20 mM Hepes, pH 7.2, containing 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, and 0.1 mM deferriferrioxamine B mesylate (Desferal). Sunflower or maize seeds were ground in the same buffer in a mortar.
Human catalase was extracted from blood clots, homogenized in a Potter-Elvejem homogenizer in the same buffer. Cell extracts were centrifuged 20 min at 6,000 ϫ g and 4°C. Supernatants were heated 5 min at 60°C and frozen at Ϫ20°C. After thawing, extracts were centrifuged as before, and the resulting supernatants were used for catalase determinations. These supernatants contained usually 85% of the initial catalase activity. Different catalase activities were separated by electrophoresis and electroelution (see below). Electroeluted catalases from N. crassa, analyzed by polyacrylamide gel electrophoresis (PAGE) under denaturing conditions, showed a high degree of purification judged by the heavy 80,000-kDa catalase band stained with Coomassie Brilliant Blue, and a low number of faint bands in the gel.
For heme extraction, Cat-1a from N. crassa conidiating aerial mycelium was purified to near homogeneity with a method to be described in detail elsewhere. Briefly, the cell extract was frozen and thawed twice, and after centrifugation, the supernatant was precipitated with acetone. The precipitate was resuspended in the same buffer and centrifuged, and the supernatant was fractionated with ammonium sulfate. The 35% ammonium sulfate precipitated fraction, containing most of the activity, was resuspended in 3 ml of 0.5 M ammonium sulfate and 50 mM PB (solution PS), and then 1 volume of phenyl-Sepharose CL-4B, equilibrated in the same solution, was added. The resulting sludge was stirred for 30 min at room temperature, washed with 50 ml of solution PS, and then loaded onto a small column. The enzyme was eluted with 20 ml of 50 mM PB and 20 ml of 10 mM PB. Fractions having most of the activity were pooled and concentrated with an Amicon ultrafiltration cell, holding a YM30 filter, and applying N 2 at a pressure of 25 p.s.i. The filter was washed with 1 volume of 10 mM PB which was added to the concentrated enzyme.
Electrophoresis-Two types of PAGE were used, minigels and preparative gels. Minigels were used for detection of catalase activity bands. They were 8 ϫ 9 cm and 0.75 mm thick, 8% polyacrylamide 0.2% bis-acrylamide, made according to the Laemmli procedure (58), but without adding ␤-mercaptoethanol or boiling the samples. Gels were run at 150 V for 2.5 h and immediately after electrophoresis stained for catalase activity.
For the preparative procedure a supernatant of a cell extract was loaded on a 12 ϫ 16 cm, 2 mm thick, 6.5% polyacrylamide gel, 0.15% bis-acrylamide, having cast a single pit. Electrophoresis was run at 200 V for 4 h, at 4°C. After electrophoresis, a slice from one side of the gel was stained for catalase activity and used to cut out slices of the gel containing the different catalases. These were then electroeluted using a "Little Blue Tank" from ISCO. Gel slices in pieces were loaded into microtraps containing electrophoresis buffer diluted 10 times. Electroelution was done at 3 W for 3 h at 4°C. Before recovering the enzymes, electric current was inverted for 10 s. Electroeluted catalases were stored in aliquots at Ϫ20°C.
Electrofocusing of catalases was done in a 14 ϫ 16 cm, 5.5% polyacrylamide gel, 0.75 mm thick, containing the following concentration and pH ranges of ampholytes, 0.75% of pH 3.5-10.0, 5% of pH 4.0 -6.6, 1.7% of pH 5.0 -7.0, and 0.8% of pH 2.5-5.0. Acetic acid, 20 mM, was used as anolyte and 20 mM sodium hydroxide as catolyte. A mixture of 1% of pH 3.5-10.0 and 5% of pH 4.0 -6.5 ampholytes in 15% glycerol was added to samples in a ratio 1:3, and these were loaded onto the gel after a prerun of 15 min at 20 W. Electrofocusing was at 25 W, for 3 h at 25°C. The gel was then stained for catalase activity.
To determine isoelectric point of the different catalase activities, a lateral strip from top to bottom of the gel was cut into 2-mm slices, each slice was incubated overnight at 22°C in 1 ml of distilled water, and the pH of each sample was determined.
Photosensitization Reactions-Two systems were used. 1) Catalase (770 g/ml) in 10 mM PB containing 10 mM uric acid was subjected to oxidation by singlet oxygen generated with riboflavin (1.37 mM) and a fluorescent light source (0.4 W/cm 2 ). A 96-well tissue-culture plate with 30 l/well was illuminated for 4 h at room temperature. Riboflavin was eliminated by filtration and washing with 15 ml of 50 mM PB through Millipore Ultrafree-CL cellulose (5,000 nominal molecular weight limit) and the enzyme was used for heme extraction and analysis. 2) A drop of 15 l of purified catalase (40 units) in electroeluting buffer (2.5 mM Tris, pH 8.9, 19 mM glycine) was suspended from a micropipette tip in a 2-ml vial in close vicinity to a filter paper impregnated with a photoactive substance, usually rose Bengal; riboflavin, acridine orange, methylene blue, or toluidine blue were also used. Illumination was done for 3 h with an incandescent light bulb (100 W) at 40 cm distance (see Fig. 4A).
Heme Extraction and HPLC Analysis-The heme of purified Cat-1 (770 g) was extracted under dim light with acid acetone. Catalase preparations used for heme extraction were immediately after purification, after 1 or 3 weeks of storage, or after 1 week of storage modified by photosensitization with riboflavin and light. The acetone extract was dried with a stream of N 2 . Heme was dissolved in 50 l of acetonitrile containing 0.5% trifluoracetic acid, and a 0.5% trifluoracetic acid solution was slowly added to obtain a acetonitrile/water proportion of 60/40.

RESULTS
Catalase Modification Is Oxygen-dependent-The electrophoretic mobility of purified Cat-1a changed during storage due to a gain in negative charges, giving rise to the modified Cat-1b to Cat-1e. 3 In vivo, Cat-1c and Cat-1e appeared or became prominent during the recurrent hyperoxidant state of the conidiation process and under stress conditions. 3 We considered that ROS could react with Cat-1a in vivo and in vitro to give Cat-1c and Cat-1e. Indeed, there was a modification of Cat-1a when stored for 1 week under air or pure dioxygen (Fig.  1, lane 3 and 4, respectively), and no modification was observed when stored under argon (Fig. 1, lane 2).
Catalase Is Modified by Singlet Dioxygen-Cat-1a, incubated in the presence of riboflavin and light, exhibited a time-dependent shift in electrophoretic mobility to Cat-1e ( Fig. 2A). Cat-1a mobility did not change with light alone (lane 2), nor did its specific activity after 24 h of intense illumination (not shown). Photosensitization reactions give rise to singlet oxygen (mainly 1 ⌬gO 2 ), but in presence of a reducing agent the main product is superoxide ion. Reducing agents, such as thiols, clearly inhibited the shift from Cat-1a to Cat-1e, but other reducing agents, dimethyl sulfoxide (Me 2 SO) and NADPH, were less effective (Fig. 2B). Added superoxide dismutase did not affect the change in mobility brought about by photosensitization with riboflavin alone (Fig. 2, compare lane 5 of A with lane 1 of B). These results suggests that singlet oxygen caused the Cat-1a modification.
A hydroxyl radical may be formed under our experimental conditions. However, this species did not change the mobility of Cat-1a, but caused instead enzyme inactivation in about 3 h. To avoid Cat-1a inactivation by hydroxyl radical, uric acid was used in the experiments shown in Fig. 2. Uric acid is a good hydroxyl radical scavenger and does not react significantly with singlet oxygen (4). In the absence of uric acid the enzyme was inactivated (Fig. 2B, lane 2).
Modification in electrophoretic mobility of Cat-1a was due to an increase in net negative charges. 3 Thus, we compared the isoelectric point of purified Cat-1a to Cat-1e with the isoelectric point of the Cat-1a modified by photosensitization with riboflavin. Cat-1a to Cat-1e were purified from a nondenaturing gel; Cat-1b was the enzyme electroeluted from the zone of the gel between Cat-1a and Cat-1c. The pI of purified catalases decreased progressively from Cat-1a (pI ϭ 5.45) to Cat-1e (pI ϭ 5.25) (Fig. 3A, lanes 1-4), corresponding to a 0.05 change between successive conformers with the exception of Cat-1c. Illuminated Cat-1a with riboflavin showed a similar modification in pI (Fig. 3A, lanes 5-8). No further modification was observed on continuing the photosensitization reaction once the pI of Cat-1e was reached (lane 8 and 9). Purified Cat-1c exhibited two very different isoelectric points, 5.80 and 4.72 (Fig. 3A, lane 3). A possible explanation for these results is offered in Fig. 3B (see below).
Results shown up to now indicate the participation of singlet oxygen in the modification of Cat-1a. However, the excited state of the photosensitizer or a degradation product from it could react with the enzyme. Hartman et al. (59) showed that singlet oxygen is the only ROS that can diffuse through air. Thus, we incubated Cat-1a with a source of singlet oxygen at a distance of less than 2 mm (Fig. 4A). Under these conditions, photoactive compounds changed the electrophoretic mobility of Cat-1a. Rose Bengal was more effective than acridine orange, methylene blue, riboflavin, and toluidine blue in Fig. 4B. Differences were due to the incandescent light source used, because in similar experiments, using a halogen light source at the same light intensity, the sample containing toluidine blue was the most effective (not shown).
Catalases from Different Organisms Are also Modified by Singlet Dioxygen-A shift in electrophoretic mobility was observed when Cat-2 of N. crassa was subjected for 3 h to a source of singlet dioxygen (Fig. 5A, lanes 3-6). This modified Cat-2 coincided with the mobility of Cat-2Ј found in conidia and the one formed during storage of purified Cat-2 (not shown). Purified enzymes from other fungi, catalase A from A. nidulans conidia (Fig. 5B) and catalase T from S. cerevisiae (Fig. 5C) were also susceptible to modification by singlet oxygen (lanes 3-6). Purified HPII from E. coli showed a similar modification with riboflavin and light (Fig. 5D, lanes 3-6). Modification of . Superoxide dismutase (0.5 g/ml) and uric acid (10 mM) were used in A-C to scavenge superoxide and the hydroxyl radical, respectively. these catalases was impaired by histidine or 5-ASA (Fig. 5,  A-D, lanes 7 and 8). A partially purified cell extract from S. coelicolor showed up to five bands of catalase activity. The band with the lowest mobility shifted to a higher mobility when illuminated in the presence of riboflavin (Fig. 5E, lanes 3-6); this shift was impaired by His or 5-ASA (lanes 7 and 8). It is noted that a band with intermediate mobility appeared with incubation; the appearance of this band was still present in the sample with His (lane 7). None of these catalases were photo-inhibited (Fig. 5, A-E, lanes 1 and 2). In contrast, the commercial bovine liver catalase and purified human erythrocyte catalase were photoinactivated (lanes 1 and 2 in Fig. 5, F and G, respectively) (60, 61) and modified by singlet oxygen (lanes  3-6). However, the modification by singlet oxygen did not increase this inactivation (compare lanes 2 and 6). Catalases in a partially purified cell extract from Helianthus annuus and Zea mays seeds were modified by singlet oxygen (Fig. 5, H and I,  respectively). Because these catalases were photoinactivated, the extracts were illuminated only for 5-30 min (lanes 2-8) (60). Illumination alone of the sunflower cell extract was sufficient to cause a shift in catalase electrophoretic mobility (Fig.  5H, lanes 1 and 2), but this change in mobility was more prominent in the presence of singlet oxygen (lanes 3-6). 5-ASA inhibited the sunflower catalase activity but not the Z. mays enzyme (Fig. 5, H and I, lane 8). His was effective in preventing the modification in both plant catalases (lane 7). In summary, the modification by singlet oxygen giving rise to more acidic  4. A, scheme of the system used to expose catalase to a pure source of singlet oxygen. Singlet oxygen is the only reactive oxygen species in this system that can diffuse through air. B, zymogram of purified Cat-1a after exposure to a pure source of singlet oxygen. c, control of Cat-1a illuminated 3 h in the absence of photosensitizing agent; in the presence of acridine orange, ao; methylene blue, mb; riboflavin, rf; and rose Bengal, rb. active conformers seems to be a universal characteristic of catalases.
The Heme of Cat-1a Is Modified by Singlet Oxygen-Heme from Cat-1a was isolated with acid acetone and analyzed by HPLC. The main heme peak from Cat-1a (peak a in Fig. 6A) migrated together with a protoporphyrin IX standard. Three additional smaller peaks were observed: two peaks eluted few minutes earlier (b and c), indicating less hydrophobic porphyrins, and one (peak d) migrated close to unmodified heme. In Cat-1a, modified during 1 week of storage to give mainly Cat-1c (Fig. 7A, lane 2), the main heme peak decreased and the other three increased (Fig. 6B). This was more evident after storage for 3 weeks (Fig. 6C). When the 1-week stored catalase was modified by photosensitization with riboflavin, peaks a, b, and c disappeared, and peak d increased (Fig. 6D). There was no detectable change in the polypeptides from the acid acetone extracts when analyzed by sodium dodecyl sulfate-PAGE (Fig.  7B). Thus, heme from the Cat-1a, modified during storage or by singlet oxygen, changed by a three-step reaction: the unmodified heme peak a gave rise to b, c, and d; the fully modified heme presented only peak d.
Cat-1 Modification Results in an Increase in Heme Asymmetry-Heme absorbance spectra from Cat-1a and Cat-1e were run in the absence or presence of imidazole. Heme extracted from Cat-1a showed a Soret peak at 369 nm and the usual small peak at 592 nm (Fig. 8A). Imidazole shifted the Soret peak to 408 nm, but its absorbance was low as compared with the hemin standard when shifted from 395 to 412 nm with imidazole (Fig. 8C). Hexa-coordination of the Fe(III)protoheme IX (heme b) shifts the Soret peak from a high spin (395 nm) to a low spin (412 nm) with an increase in absorbance. The different Soret peaks and the low absorbance at 408 nm indicates that heme in Cat-1a is not heme b. Cat-1e, modified by photosensitization, showed a Soret peak at 363 nm that shifted to 408 nm with imidazole, but its absorbance was even smaller than the one of the unmodified heme (Fig. 8B). Absorbance ratio of the Soret peaks (ϩimidazole/Ϫimidazole) was 1.45 for the unmodified heme and 2.65 for the modified one. These results are consistent with an asymmetric heme in Cat-1a, and an increase in asymmetry with heme modification.

Singlet Oxygen Modified Cat-1a-Modification of purified
Cat-1a was dependent on O 2 ; no modification occurred under argon. Singlet oxygen generated by photosensitization reactions brought about a rapid sequential shift in electrophoretic mobility of purified Cat-1a, similar to the ones observed in vivo. 3 Modification of catalase by photosensitization reactions was hindered by reducing agents, such as thiols, conditions in which the main product is superoxide instead of singlet oxygen. Singlet oxygen scavengers also impaired the modification of catalase. The best singlet oxygen scavengers were 5-ASA, histidine and tryptophan; the last two have the highest secondorder rate constant for reaction with singlet oxygen (4). Besides reacting with singlet oxygen, 5-ASA could bind to the enzyme and hinder the access of dioxygen to the heme; in fact, we have observed inhibition of Cat-1a by 5-ASA. 4 The electrophoretic mobility of Cat-1a did not change with light alone, even when its heme was excited for 2 h with 400 nm monochromatic light of a fluorometer (not shown). Modification of Cat-1a was still observed when the photosensitization reaction was separated from the enzyme by an air barrier, demonstrating that singlet 4 A. Díaz-Vilchis, and W. Hansberg, unpublished results. oxygen was in fact the oxidizing agent. Dioxygen in its basal triplet state, albeit in an infrequent reaction, would be the reacting species during storage of catalase, producing the same modification of catalase. The reaction would be orders of magnitude more frequent with dioxygen in an activated singlet state, producing a rapid modification of catalase.
Activity of the Modified Catalase-Cat-1 did not change its specific activity when modified by singlet oxygen. A similar K m of Cat-1a and Cat-1c and inhibition by 5-ASA were detected. 4 Two reports suggested that catalase activity from beef liver was inhibited by singlet oxygen, formed from photosensitization with tetrasulfonated metallophthalocyanines (62) and by rose Bengal (63). Both studies report increased inactivation of catalase in the presence of deuterated water, which is known to increase the half live of singlet oxygen. In our experiments, catalases from beef liver, human erythrocyte and plants were modified by singlet oxygen, but modification did not increase their photoinactivation (Fig. 5). It could be that the inactivation reported by these authors is due to hydroxyl radical rather than to singlet oxygen. This is in accordance with the protecting effect of mannitol, the EPR signals, and the polymerization of the enzyme reported in one of these studies (62). Results with deuterated water could be explained if singlet oxygen modified catalase is more susceptible to inactivation by hydroxyl radicals than the unmodified enzyme. A difference in conformation of the modified enzyme is indicated by the two very different pI of Cat-1c as compared with Cat-1a. Also, increased inactivation of Cat-1c by acid as compared with Cat-1a suggested a difference in conformation. 3 A slow inactivation of bovine liver catalase by superoxide has also been reported (64). Since most activity was present after 3 h of photosensitization reaction in the absence of superoxide dismutase (Fig. 2B), Cat-1a was either insensitive to superoxide, or modification was too slow to be detected in our system.
The Modifying Reaction-The electrophoretic mobility of the Cat-1 conformers changes due to a gain in net negative charges. 3 Modification of Cat-1a shifted the pI from 5.45 to 5.25 of the fully modified Cat-1e, a change of 0.05 pH unit per modified monomer, if one assumes a sequential modification of all monomers in the tetramer (Fig. 3). Cat-1c has two modified monomers and thus should have a pI of 5.35 (5.45 Ϫ 2 ϫ 0.05). Unexpectedly, Cat-1c showed two divergent isoelectric points (5.80 and 4.72). The purified Cat-1c obtained under different conditions presented these two forms. We interpret these data as the result of conformational changes arising from the structure of catalases, which are formed by a pair of dimers. The active sites of the enzyme are in the interface between two tightly bound monomers in each dimer (65,66). Modification of two monomers in the tetramer could generate two forms: one would have a fully modified and a nonmodified dimer; the other would have two partially modified dimers. Each structure could have a different conformation (different pI) but, upon modification of a third monomer, both will give the same structure (same pI). This model is under current investigation.
The probable modification site in Cat-1 was the heme group. Cat-1a heme eluted together with a protoporphyrin IX standard in a HPLC, although its visible absorbance spectra with imidazole suggested a different heme. Modification of Cat-1a to Cat-1c diminished this heme peak giving rise to two more hydrophilic peaks and a slightly more hydrophobic one. Cat-1e, the completely modified enzyme, showed only the most hydrophobic heme peak. These results indicate that the heme was modified by dioxygen in what seems to be a three-step reaction. Modification probably increased the asymmetry of the heme as suggested by the absorbance spectra. Breakage of the ␤-bridge between pyrroles II and III, as seen in 35% and 50% of the heme molecules of P. mirabilis catalase (55) and the bovine catalase (31), is not compatible with the shift of the Soret peak observed in the presence of imidazole.
Singlet oxygen is a reactive species that could hydroxylate the porphyrin ring. Hydroxylation of heme by singlet oxygen would explain the two less hydrophobic peaks in the HPLC. The heme b incorporated into the HPII of Escherichia coli is hydroxylated spontaneously into cis-heme d without a change in activity (52). A catalase from Penicillium vitale has a similar heme d (53). The heme of a N. crassa catalase, probable Cat-1, which is the main activity, was described as a chlorin (67). Thus, hydroxylation of heme by singlet oxygen could be part of the modification reaction observed in the heme of Cat-1c and Cat-1e. Besides hydroxylation, another heme modification introducing negative charges is required to explain changes in pI. The exact reaction of singlet oxygen with heme is under current investigation.
Other Possible Sources for Cat-1a Modification-Three different catalase genes were reported for N. crassa based on catalase electrophoretic mobility and its segregation after crossing different wild type strains (46). The authors assumed no post-translational modification of catalases. In view of our results showing a change in electrophoretic mobility by reaction with singlet oxygen, arguments for three catalase genes are questionable. Because two of the presumed catalases were shown to map very close to each other in the right arm of chromosome III (46), they could be two forms of the same catalase, probably Cat-1a and Cat-1c. Deletion of the cat-1 gene should give a definitive answer.
A post-translational modification of a methionine to methionine sulfone has been detected in close proximity of the active site of the P. mirabilis PR catalase (54). This modification could be related to an oxidative modification of methionine by singlet oxygen, since methionine is one of the most reactive amino acids with singlet oxygen (4). Because most other catalases have a valine in place of the methionine sulfone of the P. mirabilis catalase, this modification is probably exceptional. N. crassa Cat-1 also has a valine residue in this position (72).
Deamination of glutamine and asparagine is often a cause of increased electrophoretic mobility in proteins (68). We have incubated Cat-1a in different alkaline buffers where deamination have been observed. However, modification of Cat-1a during storage was actually lower in alkaline buffers as compared with acid ones. We have discarded electrophoretic mobility changes due to modification of Cat-1 cysteines, by reaction of thiols and disulfides with Cat-1a, and also formation of methionine sulfoxide, by incubating Cat-1a in acid Me 2 SO (69). A change due to partial proteolysis was also dismissed because no protease activity could be detected in the purified Cat-1a after 18 h in the presence of azocasein (70). Modification of the polypeptide is improbable since the modified and unmodified polypeptides migrated similarly in a PAGE under denatured conditions.
Singlet Oxygen Formed under Stress Conditions as Possible Source of Catalase Isoforms in Different Organisms-Different catalase activity bands have been detected in Bacillus subtilis that could arise from other catalases when the bacterium was induced to sporulate (47). In S. coelicolor only one gene was detected (48), but up to six catalase bands in a zymogram were observed in cell extracts; activity and number of catalases increased with the stationary and cell differentiation phases (69). In Z. mays seeds five catalase bands appeared rapidly upon germination that were interpreted to be due to interchange of monomers between two isoenzymes having a different pI (49). Eight charge isoforms appeared in the peroxisomes of sunflower cotyledons during growth after germination and some isoforms became more prominent with light (50). Interestingly, it has been recently reported that erythrocyte catalase from HIV(ϩ) patients had an acidic isoform that was not present in uninfected individuals (71).
Bacterial, fungal, plant, and animal catalases all share the property of been susceptible to singlet oxygen modification giving rise to active enzyme conformers with a higher migration rate toward the cathode than the unmodified enzyme. The physiological meaning of catalase modification by singlet oxygen is at the moment not clear, but its appearance could be indicative of stress conditions. In fact, catalase shift in electrophoretic mobility could be used as a marker for the presence of singlet oxygen in the cell. N. crassa under stress conditions, where one would expect an increased generation of singlet oxygen, had an increased Cat-1a modification rate. 3 Modification of Cat-1 also occurs during N. crassa conidiation and spore germination, 3 indicating increased generation of singlet oxygen during these developmental transitions.
Because formation of singlet oxygen is probably inevitable under some stress conditions, the reaction of singlet oxygen with the O 2 site of the enzyme could also be unavoidable. Thus, what could have been selected in catalase is, not avoidance of modification by singlet oxygen but, a protein structure that keeps the enzyme active despite its inevitable modification by singlet oxygen. It is also possible that nature has found another use for this specific heme modification.