ARCHIVES
OF BIOCHEMISTRY
Vol. 298, No. 2, November
AND
BIOPHYSICS
1, pp. 431-437,1992
Peroxynitrite-Mediated Tyrosine Nitration Catalyzed by Superoxide Dismutase Harry Craig
Ischiropoulos,* Ling Zhu, * Jun Chen,* D. Smith,$ and Joseph S. Beckman*pl
Departments The University
of *Anesthesiology and TPhysics and $The of Alabama at Birmingham, Birmingham,
Received
2, 1992, and in revised
April
form
June
Michael Center for Alabama,
Nitric oxide (NO) is produced by endothelium, neurons, hepatocytes, neutrophils, and macrophages from the ‘To whom correspondence should be addressed at Department Anesthesiology, University of Alabama at Birmingham, 619 South Street, Birmingham, AL 35233. Fax: (205) 934-7437. $5.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
James
Macromolecular 35294
C. Martin,? Crystallography,
15, 1992
Peroxynitrite (ONOO-), the reaction product of superoxide (0;) and nitric oxide (NO), may be a major cytotoxic agent produced during inflammation, sepsis, and ischemia/reperfusion. Bovine Cu,Zn superoxide dismutase reacted with peroxynitrite to form a stable yellow protein-bound adduct identified as nitrotyrosine. The uvvisible spectrum of the peroxynitrite-modified superoxide dismutase was highly pH dependent, exhibiting a peak at 438 nm at alkaline pH that shifts to 356 nm at acidic pH. An equivalent uv-visible spectrum was obtained by Cu,Zn superoxide dismutase treated with tetranitromethane. The Raman spectrum of authentic nitrotyrosine was contained in the spectrum of peroxynitrite-modified Cu,Zn superoxide dismutase. The reaction was specific for peroxynitrite because no significant amounts of nitrotyrosine were formed with nitric oxide (NO), nitrogen dioxide (NOg), nitrite (NO;), or nitrate (NO;). Removal of the copper from the Cu,Zn superoxide dismutase prevented formation of nitrotyrosine by peroxynitrite. The mechanism appears to involve peroxynitrite initially reacting with the active site copper to form an intermediate with the reactivity of nitronium ion (NO,+), which then nitrates tyrosine on a second molecule of superoxide dismutase. In the absence of exogenous phenolics, the rate of nitration of tyrosine followed second-order kinetics with respect to Cu,Zn superoxide dismutase concentration, proceeding at a rate of 1.0 + 0.1 M-l. s-l. Peroxynitrite-mediated nitration of tyrosine was also observed with the Mn and Fe superoxide dismutases as well as other copper-containing proteins. 0 1992 Academic Press, Inc.
0003.9861/92
Tsai,?
of 19th
NADPH-dependent oxidative deamination of L-arginine (l-6). Nitric oxide rapidly reacts with superoxide (7) at rate of at least 3.7 X lo7 M-l * s-l to form the powerful oxidant peroxynitrite [ONOO-; Ref. (S)]. Sources of superoxide rapidly inactivate nitric oxide produced by cells and tissues (9, lo), and superoxide dismutase is well known to stabilize the biological activity of nitric oxide (10-12). Thus, the stabilization of nitric oxide by superoxide dismutase suggeststhat peroxynitrite can be formed in Go. Nitric oxide does not form a detectable complex with the copper of Cu,Zn superoxide dismutase (13), though Murphy and Sies (14) have recently reported that reduced Cu,Zn superoxide dismutase converts nitric oxide to nitroxyl anion (NO-). Although nitroxyl anion is proposed to have the biological activity of endothelium-derived relaxing factor (14), nitroxyl anion also reacts with molecular oxygen to give peroxynitrite (15). Peroxynitrite has a pK, of 6.8 and a half-life of under 1 s under physiological conditions. When protonated, the resulting peroxynitrous acid (ONOOH) decomposes to form potent and toxic oxidants with the reactivity of hydroxyl radical and nitrogen dioxide (16). Peroxynitrite also directly oxidizes sulfhydryl groups at a IOOO-fold greater rate than hydrogen peroxide (17) and initiates lipid peroxidation without a requirement for a transition metal (18). Thus, peroxynitrite is highly reactive and potentially a major cytotoxic agent. Pathological conditions such as activation of inflammatory cells, sepsis, and reperfusion of ischemic tissue may induce tissues to simultaneously produce superoxide and nitric oxide, leading to the formation of peroxynitrite. In the accompanying paper, we found that peroxynitrite was formed by activated alveolar macrophages (19). Recently, the luminol chemiluminescence from activated Kupffer cells was proposed to be dependent upon peroxynitrite formation (20). In experimental studies of ischemia in heart, brain, and other organs, intravenous administration of superoxide dismutase has been shown 431
432
ISCHIROPOULOS
to reduce injury when the tissue is reperfused (21-24). One mechanism whereby superoxide dismutase might be protective is by scavenging superoxide before it reacts with nitric oxide to form peroxynitrite (25). While studying the reactions of nitric oxide with superoxide in the presence of bovine Cu,Zn superoxide dismutase, we found that a stable yellow adduct of superoxide dismutase was formed at neutral to alkaline pH. The adduct was crystallized and electron density maps calculated from the X-ray diffraction data showed a clear increase in electron density at the ortho position of the sole tyrosine residue 108 of the enzyme located 18 A from the Cu in the active site (26). We report here that the yellow adduct is due to the formation of 3-nitrotyrosine of the tyrosine residue 108, resulting from the reaction of peroxynitrite with the active site of superoxide dismutase to produce a nitrating species resembling nitronium ion (NO,+). MATERIALS
AND
METHODS
Peroxynitrite was synthesized in a quenched flow reactor as previously described (16), but was not treated with manganese dioxide to avoid metal contamination. Bovine erythrocyte Cu,Zn superoxide dismutase was of pharmaceutical grade and generously provided by Griinenthal, Inc. (Aachen, FDR). Superoxide scavenging activity was measured by the inhibition of cytochrome c reduction by xanthine oxidase (27). Peroxynitrite-modified superoxide dismutase was prepared by reacting l20 mg/ml superoxide dismutase with three to five additions of 1 mM peroxynitrite in 50 mM phosphate buffer, pH 7.4. Residual nitrite and nitrate were removed with a 20 X l-cm Sephadex G25 column equilibrated with 50 mM potassium phosphate, pH 7.4. The nitrotyrosine adduct of native superoxide dismutase was prepared by treating 5 mg/ ml superoxide dismutase in 50 mM sodium bicarbonate, pH 9, with 10 mM tetranitromethane (Aldrich, Milwaukee, WI) for 30 min. The yellow product was then dialyzed exhaustively against 50 mM potassium phosphate, pH 7.4, and diluted in the appropriate buffer to 2 mg/ml prior to recording the uv-visible spectrum. Superoxide dismutase was reduced with borohydride as described by Viglino et al. (28). The copper-free enzyme was prepared by dialyzing reduced superoxide dismutase against 50 mM KCN in 100 mM potassium phosphate, pH 8.0, for 10 h (29). The cyanide was subsequently removed by dialysis against three changes of 50 mM potassium phosphate, pH 7.4. The HsO*-phenylglyoxal-inactivated Cu,Zn SOD was prepared by methods reported previously (30,31). In brief, 2 mg/ml Cu,Zn superoxide dismutase at pH 9.0 was treated with 1 mM HrO, followed by the addition of 10 mM phenylglyoxal for 1 h. The enzyme was then eluted from an Excellulose GF-5 column (Pierce, Rockford, IL) equilibrated with 50 mM potassium phosphate, pH 7.4. Protein concentrations were determined by the bicinchoninic acid method (Pierce, Rockford, IL) using Cu,Zn superoxide dismutase as a standard. The superoxide dismutase was standardized by using the extinction coefficient of bovine Cu,Zn superoxide dismutase at es68nm = 10.3 mM-’ * cm-’ (30). ~scherichi~ cdi Mn superoxide dismutase and Fe superoxide dismutase, bovine plasma ceruloplasmin, Limuluspolyphemus homolymph type VIII hemocyanin, and Pseudomonas aeruginosa azurin were obtained from Sigma (St. Louis, MO). Laser Raman spectroscopy. Raman spectra were obtained with a custom system consisting of a Spex 1401 double monochromater, a cooled RCA C31034 photomultiplier tube, and a Stanford Research System Model SR400 gated photon counter. A Macintosh II computer equipped with a National Instruments General Purpose Interface Bus (GPIB) interface card controlled the monochromator scanning and data acquisition. Raman spectra were taken at room temperature using the excitation line at 488 nm generated by a l’llargon ion laser (Spectra Physics).
ET
AL.
Spectra were determined of peroxynitrite-treated vored resonant Raman
at pH 10 where the 438 nm absorption band superoxide dismutase was most intense and faenhancement of the nitrotyrosine derivative.
Polyacrylamide gel electrophoresis. Polyacrylamide gal electrophoresis (PAGE)* was performed on homogeneous 20% native gels using a PhastSystem (Pharmacia, Piscataway, NJ) with 200 ng protein per lane. The buffer of the gels was 0.112 M acetate, 0.112 M Tris, pH 6.5, and of the buffer strips 0.88 M L-alanine, 0.25 Tris, pH 8.8. Upon completion of the electrophoresis (70 V * h), the gels were stained with Coomassie blue. SDS-PAGE was performed on 12.5% gels using samples briefly boiled in the presence of 2.5% SDS and 5% /3-mercaptoethanol. Nitrogen dioxide and nitric oxide measurement. Nitric oxide and nitrogen dioxide were detected using a chemiluminescent nitric oxide detector (Antek Instruments, Houston, TX). A continuous stream of helium gas forced the nitrogen dioxide and nitric oxide generated in the reaction chamber to the inlet of the detector which was maintained under 10 Torr of vacuum. Nitrogen dioxide was converted to nitric oxide by a furnace heated at 55O’C and placed before the gas entered the detector. The amount of nitric oxide was quantified using authentic nitric oxide solutions (1.7 mM) at room temperature. To examine the reaction of nitric oxide with Cu,Zn superoxide dismutase, 1 mg Cu,Zn superoxide dismutase in 1 ml of 50 mM phosphate buffer, pH 7.5, was rigorously bubbled with argon in a gas-tight vial. Gaseous nitric oxide (4.5 pmol) was injected with a gas-tight syringe into the 1 ml Cu,Zn superoxide dismutase solution and allowed to react for at least 10 min. Nitrogen dioxide was prepared by allowing nitric oxide to react first with oxygen and then injected into sealed vials of superoxide dismutase as described above except the vials were not bubbled with argon. The uv-visible spectrum was then recorded at acidic and alkaline pH. Stopped-flow studies. Stopped-flow measurements were performed using a Hi-Tech Scientific spectrophotometer with a mixing time less than 2 ms. The decomposition of peroxynitrite was followed at 302 nm at different pH by rapidly mixing alkaline peroxynitrite solutions with Cu,Zn superoxide dismutase in 100 mM phosphate buffer. The nitration of tyrosine 108 was followed at 440 nm. Phosphate buffers and water were passed through charcoal to remove organics and Chelex 100 to remove metals and were also vigorously degassed to eliminate complications of bicarbonate interference. Statistics. Data are reported as means f STD with significant differences determined by one-way analysis of variance (ANOVA) using the least significant difference post hoc test (JMP program from SAS, Cary, NC) on a Macintosh.
RESULTS Ultraviolet-visible spectra. The spectra of peroxynitrite-modified superoxide dismutase were highly pH-dependent, consistent with the presence of 3-nitrotyrosine (Figs. 1 and 2). Spectral shifts exhibited a pK, of 7.8 for the disappearance of a peak in the alkaline range at 438 = 4300 + 160 M-’ . cm-‘), which shifted to 356 nm (~438~~ nm in acidic pH (e356nm = 3600 + 400 M-l . cm-‘). Similar spectra were obtained by treating native superoxide dismutase with tetranitromethane, which is well-known to produce 3-nitrotyrosine (32). The pK, of 3-nitrotyrosine itself was at 7.35 with c350nm= 3400 M-’ . cm-’ in acid and * cm-’ in alkali. The pH-dependent at ~438 nm = 4200 M-l peaks at 356 and 438 nm of peroxynitrite-modified superoxide dismutase disappeared with reduction by dithi’ Abbreviations sodium dodecyl mutase.
used: PAGE, polyacrylamide gel electrophoresis; sulfate; PEG, polyethylene glycol; SOD, superoxide
SDS, dis-
SUPEROXIDE
DISMUTASE-CATALYZED
433
NITRATION 8CUM
0.1
0
500
1000
300
350
400
450
500
550
600
Wavelength (nm) FIG. 1. Ultraviolet-visible spectra of peroxynitriteand tetranitromethane-modified Cu,Zn superoxide dismutase. Dashed lines are spectra of 2 mg/ml of peroxynitrite-modified superoxide dismutase and solid lines are the spectra of 2 mg/ml of tetranitromethane-modified superoxide dismutase in 50 mM phosphate, pH 6.0 (absorbance maxima, 356 nm), and 50 mM bicarbonate, pH 10 (absorbance maxima, 438 nm).
onite, but not ascorbate, glutathione, or borohydride. thionite reduces the nitro group on nitrotyrosine colorless amine (32).
Dito a
Laser Raman spectra of Cu,Zn superoxide dismutase. The Raman spectra of peroxynitrite-modified superoxide dismutase contained the Raman spectrum of authentic 3-nitrotyrosine (Fig. 3). Tentative band assignments based upon spectra in the literature are given in Table I (33). The only significant differences between nitrotyrosine and peroxynitrite-modified superoxide dismutase were broadening of the symmetric stretching band
FIG. 3. Laser Raman spectra of 0.5 mM nitrotyrosine (spectrum a) and 2 mg/ml bovine Cu,Zn superoxide dismutase modified with peroxynitrite (spectrum b), both in 50 mM bicarbonate, pH 10. Spectra were recorded with the following parameters: laser excitation line, 466 nm; laser output power, 400-500 mW; slit height for Spex 1401 monochromator, 10 mm; slitwidths of the Spex 1401 monochromator, 225250-225; resolution, 5 cm-‘; scan rate, 1 wavenumber * s-i. channel-‘.
of the nitro-group at 1336 cm-’ and splitting of the symmetric ring-breathing vibration peak at 817 cm-‘. The symmetric NO2 stretching band also showed a shoulder at 1319 cm-l. Similar observations were reported for the NO2 stretching band of nitrotyrosine in lysozyme nitrated by tetranitromethane (33). These subtle differences may be attributable to constraint of nitrotyrosine within the superoxide dismutase protein. Electrophoretic studies. The reaction of peroxynitrite with Cu,Zn superoxide dismutase produced a series of more negatively charged superoxide dismutase variants
TABLE 0.20
Comparison Assignments Dismutase
8 B f 2 2
I
of the Observed Raman of Peroxynitrite-Modified versus 3-Nitro-L-Tyrosine
Frequencies and Band Cu,Zn Superoxide at pH
10.3
i Peroxynitrite-modified superoxide dismutase
0.10
-
0.05
-
0.00
2ooo
Wave Number (cm-‘)
0.0
0.15
1500
1
’ 2
4
6
8
10
12
PH FIG. 2. The pK, of peroxynitrite-modified estimated from changes in absorbance dismutase modified with peroxynitrite as a function of pH.
superoxide dismutase was of 1 mg/ml Cu,Zn superoxide at 356 nm (A) and 438 nm (m)
1622 1538 1447 1413 1336 1319 1247 1176 1072 926 817 790 396 Note. sym, symmetric; plane; -, not observed.
Nitrotyrosine 1622 1538 1447 1413 1336 1247 1072 926 817 400 as, anti-symmetric;
Band CNO2 C -C C-C NOz NO2 C-OH CC-H C-C-C sym Fermi C-C-C
assignment
C stretching as stretching stretching stretching sym stretching sym stretching stretching NOa stretching ip bending ip bending ring breathing resonance op bending
ip, in plane;
op, out of
434
ISCHIROPOULOS
1234561234567 FIG. 4. Electrophoretic banding patterns of 200 ng bovine Cu,Zn superoxide dismutase on 20% native PAGE stained with Coomassie blue. Gel A, lane 1, bovine Cu,Zn superoxide dismutase; 2, bovine Cu,Zn superoxide dismutase modified by 10 mM peroxynitrite; 3-6, bovine Cu,Zn superoxide dismutase modified by 10 mM peroxynitrite in the presence of 0.5, 1, 5, and 10 mM phenol. Gel B, lane 1, bovine Cu,Zn superoxide dismutase; 2, bovine Cu,Zn superoxide dismutase modified by 5 mM peroxynitrite; 3-7, bovine Cu,Zn superoxide dismutase modified by 5 mM peroxynitrite in the presence of 1,5,10,25, and 50 mg/ml lysozyme.
ET
AL.
oxide dismutase followed pseudo-first-order kinetics. However, a small deviation occurred at later times when overlapping absorbance from the nitrotyrosine formed became significant. The formation of nitrotyrosine followed at 440 nm was linear when plotted as l/(Ainf, - A,) versus time, indicating that the reaction obeyed a secondorder rate law with respect to Cu,Zn superoxide dismutase concentration (Fig. 5). The slope of the line gave an estimated rate constant of kz = 1.0 f 0.1 M-l * s-l (n = 6) for the self nitration of Cu,Zn superoxide dismutase. The reaction between two superoxide dismutase dimers was confirmed by showing that a catalytic amount of na1.0
0.8
5 2
on native polyacrylamide gel (Fig. 4). Similar changes in the electrophoretic mobility were observed with the tetranitromethane-modified enzyme (data not shown). Peroxynitrite treatment of Cu,Zn superoxide dismutase did not appear to degrade the protein since SDS-PAGE under reduced conditions showed a single band with molecular weight of 16,000 (data not shown). Hydroxyl radical scavengers, such as 50 mM dimethylsulfoxide, did not affect the extent of superoxide dismutase modification. However, some of the apparent electrophoretic modifications to superoxide dismutase could be reversed by adding either phenol or lysozyme, implying that the peroxynitrite-mediated modification of superoxide dismutase was competitively inhibited by phenolic compounds and tyrosinecontaining molecules (Fig. 4). If a solution of phenol plus superoxide dismutase was treated with peroxynitrite and then separated by size-exclusion chromatography, only the phenol peak was nitrated. Reactions with modified CyZn superoxide dismutases. Copper-deficient superoxide dismutase did not form a nitrated adduct and was devoid of superoxidescavenging activity. Restoration of copper to the active site regenerated normal enzymatic activity for both superoxide and peroxynitrite. Treatment of Cu,Zn superoxide dismutase with Hz02 and phenylglyoxal inhibited the superoxide scavenging activity by >99%, but the modified enzyme retained the peroxynitrite nitrating activity. Cu,Zn superoxide dismutase with 70% of the lysine residues modified by conjugation with monomethoxypolyethylene glycol (average molecular mass5000) formed a nitrated adduct when reacted with peroxynitrite. All available tyrosine residues were nitrated (63 PM nitrotyrosine) by the reaction of 1 mg/ml PEG-Cu,Zn superoxide dismutase with 2 mM peroxynitrite. Second-order reaction rate. The disappearance of peroxynitrite at 302 nm in the presence of Cu,Zn super-
a
0.4
0.2 j
0.6 10
20
Time
0.0
30
40
1 0
50
(se.@
! 5
10
15
20
25
Time (set) 30.0 25.0 VT
20.0
3
i
15.0
g < 5
10.0 5.0 0.0
Time (set) FIG. 5. sition at from the dismutase bance of tyrosine 302 nm.
(a) Stopped-flow detection of 1 mM peroxynitrite decompopH 8.0, 37°C followed at 302 nm and nitrotyrosine formation reaction of 1 mM peroxynitrite and 6 mg/ml Cu,Zn superoxide followed at 440 nm. This pH was used to increase the absornitrotyrosine at 440 nm. (b) Second-order rate plot of nitroformation. (c) Semilog plot for peroxynitrite decomposition at
SUPEROXIDE
DISMUTASE-CATALYZED
tive Cu,Zn superoxide dismutase nitrated Cu-depleted, Zn superoxide dismutase. Fifty five percent of the tyrosine residues (there are two tyrosine residues per superoxide dismutase dimer) of 78 f.&M Cu-depleted, Zn superoxide dismutase was nitrated by 2 mM peroxynitrite in the presence of 8 PM Cu,Zn superoxide dismutase. Only 9% of Cu-depleted, Zn superoxide dismutase was nitrated in the absence of Cu,Zn superoxide dismutase. Reaction of CyZn superoxide dismutase with nitric oxide, nitrogen dioxide, nitrite, and nitrate. The reaction of either oxidized or reduced Cu,Zn superoxide dismutase with up to 4 mM nitric oxide or nitrogen dioxide in phosphate buffer at pH 7.4 did not result in a significant formation of nitrotyrosine. Nitrotyrosine was not formed after the reaction of either oxidized or reduced Cu,Zn superoxide dismutase with up to 63 mM nitrite or nitrate at pH 7.4. Furthermore, the production of nitrogen dioxide from 1 mM peroxynitrite in the presence of Cu,Zn superoxide dismutase in phosphate buffer, pH 7.4, was only 2.0 -C 0.5 FM, which was about the same as in buffer alone. In contrast, the same amount of peroxynitrite added to 100 mM DMSO to trap hydroxyl radical increased nitrogen dioxide to 39 + 2 PM in the presence or absence of superoxide dismutase. The yield of nitrotyrosine formed by 1 mM peroxynitrite reacted with 2 mg/ml Cu,Zn superoxide dismutase was the same (94 PM) in the presence of 100 mM DMSO as in its absence even though the concentration of nitrogen dioxide was 20-fold greater. Manganese and iron superoxide dismutase. Self nitration of tyrosine residues was also catalyzed by Mn superoxide dismutase and Fe superoxide dismutase. Six of the seven available tyrosines of the E. coli Mn superoxide dismutase were nitrated by 12 mM peroxynitrite. However, unlike Cu,Zn superoxide dismutase, Mn superoxide dismutase and Fe superoxide dismutase were inactivated by peroxynitrite. The enzymatic activity for superoxide of 1 mg/ml Cu,Zn superoxide dismutase was unaffected after treatment with up to 10 mM peroxynitrite as measured by standard superoxide dismutase assays based upon in-
TABLE
II
Superoxide Scavenging Activity after Reaction with Peroxynitrite -Peroxynitrite Cu,Zn superoxide dismutase Mn superoxide dismutase Fe superoxide dismutase Note.
1 mg/ml
mM phosphate,
was determined 4. * Statistically
4600 k 260 3200 + 170 4300 f 440
+Peroxynitrite 4700 f 390 970 + 230* 1100 + 240*
protein was reacted with 1 mM peroxynitrite in 100 pH 7.4. Superoxide dismutase activity (U/mg protein) by the xanthine oxidase cytochrome c reduction, n = significant
at
P < 0.05.
435
NITRATION
TABLE
III
Reaction of Peroxynitrite with Copper Containing Proteins Tyrosine
content bM)’
Azurin (40) Cu,Zn SOD Hemocyanin Ceruloplasmin
(84) (357)* (471)
Nitrotyrosine (PM) 17 83 85 18
f + + +
%nitration 1 2 1 1
Note. Nitrotyrosine formation after reacting the proteins peroxynitrite in phosphate, pH 7.5 (n = 3). “The tyrosine content was estimated from the known sequences. * The tyrosine content of hemocyanin is an approximation protein forms aggregates of unknown size.
43 100 24 4 with amino
2 mM acid
since the
hibition of cytochrome c reduction. At the same protein concentration, Mn superoxide dismutase was inactivated by 70% and Fe superoxide dismutase by 75% after reacting with 1 mM peroxynitrite (Table II). Higher peroxynitrite concentrations (>5 mM) inhibited more than 95% the activity of Mn superoxide dismutase and Fe superoxide dismutase. In dilute Cu,Zn superoxide dismutase solutions (10 pg/ml), up to 30% of activity was lost when greater than 10 mM peroxynitrite was added. Reaction of peroxynitrite with other copper-containing proteins. The reaction of peroxynitrite with coppercontaining proteins in phosphate buffer, pH 7.4, resulted in the formation of nitrotyrosine evidenced by the appearance of a peak between 420 and 440 nm at alkaline pH. Table III shows that 2 mM peroxynitrite nitrated all the tyrosine residues in Cu,Zn superoxide dismutase, 43% of the tyrosine residues in azurin, approximately 24% in hemocyanin, and only 4% in ceruloplasmin. DISCUSSION Peroxynitrite reacts with Cu,Zn, Mn, and Fe superoxide dismutases to generate a potent nitrating agent that modifies tyrosine residues. Nitration of tyrosine residue 108 in bovine Cu,Zn superoxide dismutase was the only apparent modification in the X-ray structure (26). However, other less visible alterations most likely occurred randomly to other amino acids resulting in the multiple charged bands observed on the native gel electrophoresis (Fig. 4). The nitrating agent was not nitric oxide nor nitrogen dioxide because they were not significant products of the superoxide dismutase reaction and adding up to 4 mM concentrations of either compound did not result in significant nitration of tyrosine 108. Furthermore, removal of the copper from Cu,Zn superoxide dismutase prevented nitration, while nitrating activity was restored by replacing copper in the active site. To account for the essential role of copper in the
436
ISCHIROPOULOS
active site of superoxide dismutase, we propose that peroxynitrite anion is attracted to the active site of superoxide dismutase by the same electrostatic field that attracts the negatively charged superoxide anion (Fig. 6). The active site of Cu,Zn superoxide dismutase contains a deep hydrophobic pocket shaped to accommodate superoxide with the copper atom sitting at the base (34,35). Thus, peroxynitrite in the trans-configuration could fit into the active site with the O-O group sitting in the pocket and the -N=O remaining exposed to solvent. Once in the active site, one electron from peroxynitrite may be transferred to the copper to form a cuprous intermediate. ONOO-
+ Cuf2,Zn
SOD *ON00
-
Cu+l,Zn
SOD
We have recently estimated that the redox potential at pH 7.0 between the nitrosodioxyl radical (ON00 * ) and peroxynitrite anion to be 0.43 V (36), which is close to the reported midpoint potential of 0.42 V for the copper in superoxide dismutase (34, 37). Thus, peroxynitrite could reduce the active site copper. During the second step in the catalytic cycle of superoxide dismutation, a hydrogen ion is donated from one of the histidines involved in binding the copper. A similar event could occur with peroxynitrite, initiating the decomposition of peroxynitrite to form nitronium ion while leaving a hydroxide ion bound to the copper. ON00
-
Cu+‘,Zn
SOD + H+ * NO;
+ HO -
Cu+‘,Zn
SOD
Addition of a second hydrogen ion from the solvent will release water from the copper, regenerating native cupric superoxide dismutase. Nitronium ion is a strong oxidant [E,’ = 1.6 V, Ref. (36)] and the active agent in fuming nitric acid (a caustic mixture of sulfuric and nitric acid) commonly used to prepare nitrophenols in organic chemistry. We have recently calculated that the energy for heterolytic cleavage of peroxynitrite to nitronium ion and water requires approximately 13 kcal. mol-l at pH 7.0 (36). In an accompanying paper (38) the activation energy for phenolic nitration by peroxynitrite catalyzed by Fe+3EDTA was found to be 12 kcal. malll. How could generation of nitronium ion at the active site result in nitration of tyrosine 108 located 18 A distant from the copper active site? Possibly, the positively charged nitronium ion could be repelled by the same electrostatic field that attracts superoxide and peroxynitrite into the active site and then be guided to tyrosine 108 by the adjacent negatively charged glutamate residue 107. However, nitronium ion reacts rapidly with water to yield nitrate and thus might not survive long enough to reach tyrosine. Furthermore, such a mechanism also does not
ET
AL.
-0,0/N
3
s
9 OXN
FIG. 6. Proposed model for the generation of a strong nitrating in the active site of Cu,Zn superoxide dismutase. The superoxide mutase schematic is derived from Getzoff et al. (34).
agent dis-
account for the second-order kinetics of nitration observed with respect to superoxide dismutase concentration. The second-order kinetics suggest that a transient intermediate is formed in or near the active site, which nitrates tyrosine upon collision with a second superoxide dismutase molecule. The intermediate may simply be peroxynitrite bound to copper in the active site. A second possibility is that nitronium ion reacts with the negatively charged glutamate residue 131 located only 8 A from the active site copper to form a highly reactive nitrocarboxylate adduct (R-C02-N02). Glutamate 131 is a highly conserved amino acid in the Cu,Zn superoxide dismutase family, which helps generate the electrostatic field that guides superoxide into the active site (34, 39). The negative charged carboxylic acid is oriented toward the active site in an ideal position to react with nitronium released from the heterolytic cleavage of peroxynitrite bound to the copper (Fig. 6). Nitrocarboxylates are well-known nitrating agents for the preparation of nitrophenols. In the absence of exogenous substrates, the nitroglutamate could slowly nitrate tyrosine 108 on other superoxide dismutase molecules by a second-order reaction rate proceeding at 1.0 f 0.1 M-’ * s-l. In support of the above mechanism, we found that a small amount of Cu,Zn superoxide dismutase will catalyze the nitration of tyrosine residues on Cu-depleted,Zn superoxide dismutase. Furthermore, the electrophoretic modifications to peroxynitrite caused by peroxynitrite were competitively inhibited by exogenous phenol and by lysozyme (Fig. 4). Peroxynitrite did not inactivate bovine Cu,Zn superoxide dismutase whereas E. coli Mn and Fe superoxide dismutases rapidly and irreversibly lost activity (Table II). The inactivation may be due to nitration of a tyrosine in the active site (residues 34 or 36 in the bacterial Mn superoxide dismutase sequences), which is involved in catalysis of superoxide dismutation (40). The resulting steric constraint and additional partial negative charge in the active site would then inhibit further reaction with superoxide. The active site tyrosine in Mn superoxide dis-
SUPEROXIDE
DISMUTASE-CATALYZED
mutase is resistant to nitration by tetranitromethane (411, suggesting that peroxynitrite could also be a useful alternative for protein modification. Superoxide dismutase catalyzes the nitration of many phenolics or tyrosine by peroxynitrite, which is further characterized in the second paper in this series. This nitrating reaction may be pathologically important by nitrating critical tyrosines on proteins important for cell regulation. Furthermore, superoxide dismutase-catalyzed phenolic nitration offers a useful assay for peroxynitrite. The competing reaction with superoxide dismutation can be minimized by treatment of Cu,Zn superoxide dismutase with HzOz and phenylglyoxal, while the Cu-depleted, Zn superoxide dismutase can serve as an inactive control. This assay method has allowed us to measure peroxynitrite formed by activated rat alveolar macrophages (19). ACKNOWLEDGMENTS This work was supported by a grant from the National Institute of Health HL 46407 and a Grant-in-Aid from the American Heart Association. J. S. Beckman is an Established Investigator of the American Heart Association. We thank Drs. H. C. Cheung and S-H. Lin for their assistance in the stopped flow experiments, C. Gunn for assisting with the detection of nitrogen dioxide, and Dr. R. Radi (University of the Republic, Montevideo, Uruguay) for helpful discussions.
REFERENCES 1. Palmer,
R. M. J., Ferrige,
A. G., and Moncada,
S. (1987)
Nature
327,524-526. 2. Garthwaite, J., Charles, S. L., and Chess-Williams, R. (1988) Nature 336,385-388. 3. Curran, R. D., Billiar, T. R., Stuehr, D. J., and Hofmann, K. (1989) J. Exp. Med. 170,1769-1774. 4. Schmidt, H. H. H. W., Seifert, R., and Bohme, E. (1989) FEBS Lett 244,357-360. 5. Hibbs, J. B., Jr., Taintor, R., and Vavrin, Z. (1987) Science 236, 473-476. 6. Marletta, M. A., Yoon, P. S., Iyengar, R., Leaf, C. D., and Wishnok, J. S. (1988) 7. Blough,
Biochemistry
0. C. (1985)
Znorg.
Chem.
24,
3504-
3505. 8. Saran, mun.
M., Michel,
C., and Bors,
W. (1990)
Free Radical
Res. Com-
10,221-226.
9. Gryglewski,
R. J., Palmer,
R. M. J., and Moncada,
S. (1986)
Nature
320,454-456. 10. Rubanyi, G. M., and Vanhoutte, H822-H827. 11. Moncada, S., Palmer, Natl. Acad. Sci. USA
P. M. (1986)
Am. J. Physiol.
250,
R. J. (1986)
Proc.
R. M. J., and Gryglewski, 83,9164-9168.
12. Miigge, A., Elwell, J. H., Peterson, Am. J. Physiol. 260, C219-C225.
T. E., and Harrison,
13. Gorren, A. C. F., de Boer, E., and Wever, Acta 916,38-47.
R. (1987)
B&him.
D. G. (1991) Biophys.
437
14. Murphy, M., and Sies, H. (1991) Proc. N&l. Acad. Sci. USA 88, 10860-10864. 15. Donald, C. E., Hughes, M. N., Thompson, J. M., and Bonner, F. T. (1986) Znorg. Chem. 25,2676-2677. 16. Beckman, J. S., Beckman, T. W., Chen, J., Marshall, P. M., and Freeman, B. A. (1990) Proc. Natl. Acad. Sci. USA 87, 1621-1624. 17. Radi, R., Beckman, J. S., Bush, K. M., and Freeman, B. A. (1991) J. Biol. Chem. 266, 4244-4250. 18. Radi, R., Beckman, J. S., Bush, K. M., and Freeman, B. A. (1991) Arch. Biochem. Biophys. 288,481-487. 19. Ischiropoulos, H., Zhu, L., and Beckman, J. S. (1992) Arch. Biochem. Biophys. 298,446-451. 20. Wang, J-F., Komarov, P., Sies, H., and de Groot, H. (1991) B&hem. J. 279,311-314. 21. Tamura, Y., Liguo, C., Driscoll, E. M., Jr., Hoff, P. T., Freeman, B. A., Gallagher, K. P., and Lucchesi, B. R. (1988) Circ. Res. 63, 944-959. 22. Liu, T. H., Beckman, J. S., Freeman, B. A., Hogan, E. L., and Hsu, C. Y. (1989) Am. J. Physiol. 256, H589-H593. 23. Imaizumi, S., Wollworth, V., Fishman, R. A., and Chan, P. H. (1990) Stroke 21, 1312-1317. 24. Kawamoto, S., Inoue, M., Tashiro, S., Morino, Y., and Miyauchi, Y. (1990) Arch. Biochem. Biophys. 277, 160-165. 25. Beckman, J. S. (1990) Nature 346, 27-28. 26. Smith, C. D., Carson, M., van der Woerd, M., Chen, J., Ischiropoulos, H., and Beckman, J. S. Submitted for publication. 27. McCord, J. M., and Fridovich, I. (1969) J. Biol. Chem. 244, 6049-
6055. 28. Viglino,
P., Scarpa, M., Cocco, D., and Rigo, A. (1985) Biochem. J. 229,87-90. 29. Rotilio, G., Calabrese, L., Bossa, F., Barra, D., Finazzi-Agro, A., and Mondovi, B. (1972) Biochemistry 11,2182-2186. 30. Beyer, W. F., Jr., Fridovich, I., Mullenbach, G. T., and Hallewell, R. (1987) J. Biol. Chem. 262,11182-11187. 31. Hodgson, E. K., and Fridovich, I. (1975) Biochemistry 14, 5294-
5299. 32. Glazer, 33. 34. 35.
27,8706-8711.
N. V., and Zafiriou,
NITRATION
36. 37.
A. N., Delange, R. J., and Sigman, D. S. (1975) Chemical Modification of Proteins, pp. 95-99, North-Holland, Amsterdam. Izzo, G. E., Jordan, F., and Mendelsohn, R. (1982) J. Am. Chem. Sot. 104,3178-3182. Tainer, J. A., Getzoff, E. D., Richardson, J. S., and Richardson, D. C. (1983) Nature 306, 284-287. Getzoff, E. D., Tainer, J. A., Weiner, P. K., Kollman, P. A., Richardson, J. S., and Richardson, D. C. (1983) Nature 306, 287-290. Koppenol, W. H., Moreno, J. J., Pryor, W. A., Ischiropoulos, H., and Beckman, J. S. Chem. Res. Toxicol., in press. Lawrence, G. D., and Sawyer, D. T. (1979) Biochemistry 18,3045-
3050. 38. Beckman,
J. S., Ischiropoulos, H., Zhu, L., Chen, J., van der Woerd, M., Smith, C. D., Martin, J. C., and Tsai, M. (1992) Arch. Biochem. Biophys. 298,438-445. 39. Koppenol, W. H. (1981) in Oxygen and Oxyradicals in Chemistry and Biology (Rodgers, M. A. J., and Powers, E. L., Eds.), pp. 671674, Academic Press, New York. 40. Stallings, W. C., Pattridge, K. A., Strong, R. K., and Ludwig, M. L. (1985) J. Biol. Chem. 260, 16424-16432. 41. Borders, C. L., Jr., Horton, P. J., and Beyer, W. F., Jr. (1989) Arch. Biochem. Biophys. 268, 74-80.
OF BIOCHEMISTRY
Vol. 298, No. 2, November
AND
BIOPHYSICS
1, pp. 431-437,1992
Peroxynitrite-Mediated Tyrosine Nitration Catalyzed by Superoxide Dismutase Harry Craig
Ischiropoulos,* Ling Zhu, * Jun Chen,* D. Smith,$ and Joseph S. Beckman*pl
Departments The University
of *Anesthesiology and TPhysics and $The of Alabama at Birmingham, Birmingham,
Received
2, 1992, and in revised
April
form
June
Michael Center for Alabama,
Nitric oxide (NO) is produced by endothelium, neurons, hepatocytes, neutrophils, and macrophages from the ‘To whom correspondence should be addressed at Department Anesthesiology, University of Alabama at Birmingham, 619 South Street, Birmingham, AL 35233. Fax: (205) 934-7437. $5.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
James
Macromolecular 35294
C. Martin,? Crystallography,
15, 1992
Peroxynitrite (ONOO-), the reaction product of superoxide (0;) and nitric oxide (NO), may be a major cytotoxic agent produced during inflammation, sepsis, and ischemia/reperfusion. Bovine Cu,Zn superoxide dismutase reacted with peroxynitrite to form a stable yellow protein-bound adduct identified as nitrotyrosine. The uvvisible spectrum of the peroxynitrite-modified superoxide dismutase was highly pH dependent, exhibiting a peak at 438 nm at alkaline pH that shifts to 356 nm at acidic pH. An equivalent uv-visible spectrum was obtained by Cu,Zn superoxide dismutase treated with tetranitromethane. The Raman spectrum of authentic nitrotyrosine was contained in the spectrum of peroxynitrite-modified Cu,Zn superoxide dismutase. The reaction was specific for peroxynitrite because no significant amounts of nitrotyrosine were formed with nitric oxide (NO), nitrogen dioxide (NOg), nitrite (NO;), or nitrate (NO;). Removal of the copper from the Cu,Zn superoxide dismutase prevented formation of nitrotyrosine by peroxynitrite. The mechanism appears to involve peroxynitrite initially reacting with the active site copper to form an intermediate with the reactivity of nitronium ion (NO,+), which then nitrates tyrosine on a second molecule of superoxide dismutase. In the absence of exogenous phenolics, the rate of nitration of tyrosine followed second-order kinetics with respect to Cu,Zn superoxide dismutase concentration, proceeding at a rate of 1.0 + 0.1 M-l. s-l. Peroxynitrite-mediated nitration of tyrosine was also observed with the Mn and Fe superoxide dismutases as well as other copper-containing proteins. 0 1992 Academic Press, Inc.
0003.9861/92
Tsai,?
of 19th
NADPH-dependent oxidative deamination of L-arginine (l-6). Nitric oxide rapidly reacts with superoxide (7) at rate of at least 3.7 X lo7 M-l * s-l to form the powerful oxidant peroxynitrite [ONOO-; Ref. (S)]. Sources of superoxide rapidly inactivate nitric oxide produced by cells and tissues (9, lo), and superoxide dismutase is well known to stabilize the biological activity of nitric oxide (10-12). Thus, the stabilization of nitric oxide by superoxide dismutase suggeststhat peroxynitrite can be formed in Go. Nitric oxide does not form a detectable complex with the copper of Cu,Zn superoxide dismutase (13), though Murphy and Sies (14) have recently reported that reduced Cu,Zn superoxide dismutase converts nitric oxide to nitroxyl anion (NO-). Although nitroxyl anion is proposed to have the biological activity of endothelium-derived relaxing factor (14), nitroxyl anion also reacts with molecular oxygen to give peroxynitrite (15). Peroxynitrite has a pK, of 6.8 and a half-life of under 1 s under physiological conditions. When protonated, the resulting peroxynitrous acid (ONOOH) decomposes to form potent and toxic oxidants with the reactivity of hydroxyl radical and nitrogen dioxide (16). Peroxynitrite also directly oxidizes sulfhydryl groups at a IOOO-fold greater rate than hydrogen peroxide (17) and initiates lipid peroxidation without a requirement for a transition metal (18). Thus, peroxynitrite is highly reactive and potentially a major cytotoxic agent. Pathological conditions such as activation of inflammatory cells, sepsis, and reperfusion of ischemic tissue may induce tissues to simultaneously produce superoxide and nitric oxide, leading to the formation of peroxynitrite. In the accompanying paper, we found that peroxynitrite was formed by activated alveolar macrophages (19). Recently, the luminol chemiluminescence from activated Kupffer cells was proposed to be dependent upon peroxynitrite formation (20). In experimental studies of ischemia in heart, brain, and other organs, intravenous administration of superoxide dismutase has been shown 431
432
ISCHIROPOULOS
to reduce injury when the tissue is reperfused (21-24). One mechanism whereby superoxide dismutase might be protective is by scavenging superoxide before it reacts with nitric oxide to form peroxynitrite (25). While studying the reactions of nitric oxide with superoxide in the presence of bovine Cu,Zn superoxide dismutase, we found that a stable yellow adduct of superoxide dismutase was formed at neutral to alkaline pH. The adduct was crystallized and electron density maps calculated from the X-ray diffraction data showed a clear increase in electron density at the ortho position of the sole tyrosine residue 108 of the enzyme located 18 A from the Cu in the active site (26). We report here that the yellow adduct is due to the formation of 3-nitrotyrosine of the tyrosine residue 108, resulting from the reaction of peroxynitrite with the active site of superoxide dismutase to produce a nitrating species resembling nitronium ion (NO,+). MATERIALS
AND
METHODS
Peroxynitrite was synthesized in a quenched flow reactor as previously described (16), but was not treated with manganese dioxide to avoid metal contamination. Bovine erythrocyte Cu,Zn superoxide dismutase was of pharmaceutical grade and generously provided by Griinenthal, Inc. (Aachen, FDR). Superoxide scavenging activity was measured by the inhibition of cytochrome c reduction by xanthine oxidase (27). Peroxynitrite-modified superoxide dismutase was prepared by reacting l20 mg/ml superoxide dismutase with three to five additions of 1 mM peroxynitrite in 50 mM phosphate buffer, pH 7.4. Residual nitrite and nitrate were removed with a 20 X l-cm Sephadex G25 column equilibrated with 50 mM potassium phosphate, pH 7.4. The nitrotyrosine adduct of native superoxide dismutase was prepared by treating 5 mg/ ml superoxide dismutase in 50 mM sodium bicarbonate, pH 9, with 10 mM tetranitromethane (Aldrich, Milwaukee, WI) for 30 min. The yellow product was then dialyzed exhaustively against 50 mM potassium phosphate, pH 7.4, and diluted in the appropriate buffer to 2 mg/ml prior to recording the uv-visible spectrum. Superoxide dismutase was reduced with borohydride as described by Viglino et al. (28). The copper-free enzyme was prepared by dialyzing reduced superoxide dismutase against 50 mM KCN in 100 mM potassium phosphate, pH 8.0, for 10 h (29). The cyanide was subsequently removed by dialysis against three changes of 50 mM potassium phosphate, pH 7.4. The HsO*-phenylglyoxal-inactivated Cu,Zn SOD was prepared by methods reported previously (30,31). In brief, 2 mg/ml Cu,Zn superoxide dismutase at pH 9.0 was treated with 1 mM HrO, followed by the addition of 10 mM phenylglyoxal for 1 h. The enzyme was then eluted from an Excellulose GF-5 column (Pierce, Rockford, IL) equilibrated with 50 mM potassium phosphate, pH 7.4. Protein concentrations were determined by the bicinchoninic acid method (Pierce, Rockford, IL) using Cu,Zn superoxide dismutase as a standard. The superoxide dismutase was standardized by using the extinction coefficient of bovine Cu,Zn superoxide dismutase at es68nm = 10.3 mM-’ * cm-’ (30). ~scherichi~ cdi Mn superoxide dismutase and Fe superoxide dismutase, bovine plasma ceruloplasmin, Limuluspolyphemus homolymph type VIII hemocyanin, and Pseudomonas aeruginosa azurin were obtained from Sigma (St. Louis, MO). Laser Raman spectroscopy. Raman spectra were obtained with a custom system consisting of a Spex 1401 double monochromater, a cooled RCA C31034 photomultiplier tube, and a Stanford Research System Model SR400 gated photon counter. A Macintosh II computer equipped with a National Instruments General Purpose Interface Bus (GPIB) interface card controlled the monochromator scanning and data acquisition. Raman spectra were taken at room temperature using the excitation line at 488 nm generated by a l’llargon ion laser (Spectra Physics).
ET
AL.
Spectra were determined of peroxynitrite-treated vored resonant Raman
at pH 10 where the 438 nm absorption band superoxide dismutase was most intense and faenhancement of the nitrotyrosine derivative.
Polyacrylamide gel electrophoresis. Polyacrylamide gal electrophoresis (PAGE)* was performed on homogeneous 20% native gels using a PhastSystem (Pharmacia, Piscataway, NJ) with 200 ng protein per lane. The buffer of the gels was 0.112 M acetate, 0.112 M Tris, pH 6.5, and of the buffer strips 0.88 M L-alanine, 0.25 Tris, pH 8.8. Upon completion of the electrophoresis (70 V * h), the gels were stained with Coomassie blue. SDS-PAGE was performed on 12.5% gels using samples briefly boiled in the presence of 2.5% SDS and 5% /3-mercaptoethanol. Nitrogen dioxide and nitric oxide measurement. Nitric oxide and nitrogen dioxide were detected using a chemiluminescent nitric oxide detector (Antek Instruments, Houston, TX). A continuous stream of helium gas forced the nitrogen dioxide and nitric oxide generated in the reaction chamber to the inlet of the detector which was maintained under 10 Torr of vacuum. Nitrogen dioxide was converted to nitric oxide by a furnace heated at 55O’C and placed before the gas entered the detector. The amount of nitric oxide was quantified using authentic nitric oxide solutions (1.7 mM) at room temperature. To examine the reaction of nitric oxide with Cu,Zn superoxide dismutase, 1 mg Cu,Zn superoxide dismutase in 1 ml of 50 mM phosphate buffer, pH 7.5, was rigorously bubbled with argon in a gas-tight vial. Gaseous nitric oxide (4.5 pmol) was injected with a gas-tight syringe into the 1 ml Cu,Zn superoxide dismutase solution and allowed to react for at least 10 min. Nitrogen dioxide was prepared by allowing nitric oxide to react first with oxygen and then injected into sealed vials of superoxide dismutase as described above except the vials were not bubbled with argon. The uv-visible spectrum was then recorded at acidic and alkaline pH. Stopped-flow studies. Stopped-flow measurements were performed using a Hi-Tech Scientific spectrophotometer with a mixing time less than 2 ms. The decomposition of peroxynitrite was followed at 302 nm at different pH by rapidly mixing alkaline peroxynitrite solutions with Cu,Zn superoxide dismutase in 100 mM phosphate buffer. The nitration of tyrosine 108 was followed at 440 nm. Phosphate buffers and water were passed through charcoal to remove organics and Chelex 100 to remove metals and were also vigorously degassed to eliminate complications of bicarbonate interference. Statistics. Data are reported as means f STD with significant differences determined by one-way analysis of variance (ANOVA) using the least significant difference post hoc test (JMP program from SAS, Cary, NC) on a Macintosh.
RESULTS Ultraviolet-visible spectra. The spectra of peroxynitrite-modified superoxide dismutase were highly pH-dependent, consistent with the presence of 3-nitrotyrosine (Figs. 1 and 2). Spectral shifts exhibited a pK, of 7.8 for the disappearance of a peak in the alkaline range at 438 = 4300 + 160 M-’ . cm-‘), which shifted to 356 nm (~438~~ nm in acidic pH (e356nm = 3600 + 400 M-l . cm-‘). Similar spectra were obtained by treating native superoxide dismutase with tetranitromethane, which is well-known to produce 3-nitrotyrosine (32). The pK, of 3-nitrotyrosine itself was at 7.35 with c350nm= 3400 M-’ . cm-’ in acid and * cm-’ in alkali. The pH-dependent at ~438 nm = 4200 M-l peaks at 356 and 438 nm of peroxynitrite-modified superoxide dismutase disappeared with reduction by dithi’ Abbreviations sodium dodecyl mutase.
used: PAGE, polyacrylamide gel electrophoresis; sulfate; PEG, polyethylene glycol; SOD, superoxide
SDS, dis-
SUPEROXIDE
DISMUTASE-CATALYZED
433
NITRATION 8CUM
0.1
0
500
1000
300
350
400
450
500
550
600
Wavelength (nm) FIG. 1. Ultraviolet-visible spectra of peroxynitriteand tetranitromethane-modified Cu,Zn superoxide dismutase. Dashed lines are spectra of 2 mg/ml of peroxynitrite-modified superoxide dismutase and solid lines are the spectra of 2 mg/ml of tetranitromethane-modified superoxide dismutase in 50 mM phosphate, pH 6.0 (absorbance maxima, 356 nm), and 50 mM bicarbonate, pH 10 (absorbance maxima, 438 nm).
onite, but not ascorbate, glutathione, or borohydride. thionite reduces the nitro group on nitrotyrosine colorless amine (32).
Dito a
Laser Raman spectra of Cu,Zn superoxide dismutase. The Raman spectra of peroxynitrite-modified superoxide dismutase contained the Raman spectrum of authentic 3-nitrotyrosine (Fig. 3). Tentative band assignments based upon spectra in the literature are given in Table I (33). The only significant differences between nitrotyrosine and peroxynitrite-modified superoxide dismutase were broadening of the symmetric stretching band
FIG. 3. Laser Raman spectra of 0.5 mM nitrotyrosine (spectrum a) and 2 mg/ml bovine Cu,Zn superoxide dismutase modified with peroxynitrite (spectrum b), both in 50 mM bicarbonate, pH 10. Spectra were recorded with the following parameters: laser excitation line, 466 nm; laser output power, 400-500 mW; slit height for Spex 1401 monochromator, 10 mm; slitwidths of the Spex 1401 monochromator, 225250-225; resolution, 5 cm-‘; scan rate, 1 wavenumber * s-i. channel-‘.
of the nitro-group at 1336 cm-’ and splitting of the symmetric ring-breathing vibration peak at 817 cm-‘. The symmetric NO2 stretching band also showed a shoulder at 1319 cm-l. Similar observations were reported for the NO2 stretching band of nitrotyrosine in lysozyme nitrated by tetranitromethane (33). These subtle differences may be attributable to constraint of nitrotyrosine within the superoxide dismutase protein. Electrophoretic studies. The reaction of peroxynitrite with Cu,Zn superoxide dismutase produced a series of more negatively charged superoxide dismutase variants
TABLE 0.20
Comparison Assignments Dismutase
8 B f 2 2
I
of the Observed Raman of Peroxynitrite-Modified versus 3-Nitro-L-Tyrosine
Frequencies and Band Cu,Zn Superoxide at pH
10.3
i Peroxynitrite-modified superoxide dismutase
0.10
-
0.05
-
0.00
2ooo
Wave Number (cm-‘)
0.0
0.15
1500
1
’ 2
4
6
8
10
12
PH FIG. 2. The pK, of peroxynitrite-modified estimated from changes in absorbance dismutase modified with peroxynitrite as a function of pH.
superoxide dismutase was of 1 mg/ml Cu,Zn superoxide at 356 nm (A) and 438 nm (m)
1622 1538 1447 1413 1336 1319 1247 1176 1072 926 817 790 396 Note. sym, symmetric; plane; -, not observed.
Nitrotyrosine 1622 1538 1447 1413 1336 1247 1072 926 817 400 as, anti-symmetric;
Band CNO2 C -C C-C NOz NO2 C-OH CC-H C-C-C sym Fermi C-C-C
assignment
C stretching as stretching stretching stretching sym stretching sym stretching stretching NOa stretching ip bending ip bending ring breathing resonance op bending
ip, in plane;
op, out of
434
ISCHIROPOULOS
1234561234567 FIG. 4. Electrophoretic banding patterns of 200 ng bovine Cu,Zn superoxide dismutase on 20% native PAGE stained with Coomassie blue. Gel A, lane 1, bovine Cu,Zn superoxide dismutase; 2, bovine Cu,Zn superoxide dismutase modified by 10 mM peroxynitrite; 3-6, bovine Cu,Zn superoxide dismutase modified by 10 mM peroxynitrite in the presence of 0.5, 1, 5, and 10 mM phenol. Gel B, lane 1, bovine Cu,Zn superoxide dismutase; 2, bovine Cu,Zn superoxide dismutase modified by 5 mM peroxynitrite; 3-7, bovine Cu,Zn superoxide dismutase modified by 5 mM peroxynitrite in the presence of 1,5,10,25, and 50 mg/ml lysozyme.
ET
AL.
oxide dismutase followed pseudo-first-order kinetics. However, a small deviation occurred at later times when overlapping absorbance from the nitrotyrosine formed became significant. The formation of nitrotyrosine followed at 440 nm was linear when plotted as l/(Ainf, - A,) versus time, indicating that the reaction obeyed a secondorder rate law with respect to Cu,Zn superoxide dismutase concentration (Fig. 5). The slope of the line gave an estimated rate constant of kz = 1.0 f 0.1 M-l * s-l (n = 6) for the self nitration of Cu,Zn superoxide dismutase. The reaction between two superoxide dismutase dimers was confirmed by showing that a catalytic amount of na1.0
0.8
5 2
on native polyacrylamide gel (Fig. 4). Similar changes in the electrophoretic mobility were observed with the tetranitromethane-modified enzyme (data not shown). Peroxynitrite treatment of Cu,Zn superoxide dismutase did not appear to degrade the protein since SDS-PAGE under reduced conditions showed a single band with molecular weight of 16,000 (data not shown). Hydroxyl radical scavengers, such as 50 mM dimethylsulfoxide, did not affect the extent of superoxide dismutase modification. However, some of the apparent electrophoretic modifications to superoxide dismutase could be reversed by adding either phenol or lysozyme, implying that the peroxynitrite-mediated modification of superoxide dismutase was competitively inhibited by phenolic compounds and tyrosinecontaining molecules (Fig. 4). If a solution of phenol plus superoxide dismutase was treated with peroxynitrite and then separated by size-exclusion chromatography, only the phenol peak was nitrated. Reactions with modified CyZn superoxide dismutases. Copper-deficient superoxide dismutase did not form a nitrated adduct and was devoid of superoxidescavenging activity. Restoration of copper to the active site regenerated normal enzymatic activity for both superoxide and peroxynitrite. Treatment of Cu,Zn superoxide dismutase with Hz02 and phenylglyoxal inhibited the superoxide scavenging activity by >99%, but the modified enzyme retained the peroxynitrite nitrating activity. Cu,Zn superoxide dismutase with 70% of the lysine residues modified by conjugation with monomethoxypolyethylene glycol (average molecular mass5000) formed a nitrated adduct when reacted with peroxynitrite. All available tyrosine residues were nitrated (63 PM nitrotyrosine) by the reaction of 1 mg/ml PEG-Cu,Zn superoxide dismutase with 2 mM peroxynitrite. Second-order reaction rate. The disappearance of peroxynitrite at 302 nm in the presence of Cu,Zn super-
a
0.4
0.2 j
0.6 10
20
Time
0.0
30
40
1 0
50
(se.@
! 5
10
15
20
25
Time (set) 30.0 25.0 VT
20.0
3
i
15.0
g < 5
10.0 5.0 0.0
Time (set) FIG. 5. sition at from the dismutase bance of tyrosine 302 nm.
(a) Stopped-flow detection of 1 mM peroxynitrite decompopH 8.0, 37°C followed at 302 nm and nitrotyrosine formation reaction of 1 mM peroxynitrite and 6 mg/ml Cu,Zn superoxide followed at 440 nm. This pH was used to increase the absornitrotyrosine at 440 nm. (b) Second-order rate plot of nitroformation. (c) Semilog plot for peroxynitrite decomposition at
SUPEROXIDE
DISMUTASE-CATALYZED
tive Cu,Zn superoxide dismutase nitrated Cu-depleted, Zn superoxide dismutase. Fifty five percent of the tyrosine residues (there are two tyrosine residues per superoxide dismutase dimer) of 78 f.&M Cu-depleted, Zn superoxide dismutase was nitrated by 2 mM peroxynitrite in the presence of 8 PM Cu,Zn superoxide dismutase. Only 9% of Cu-depleted, Zn superoxide dismutase was nitrated in the absence of Cu,Zn superoxide dismutase. Reaction of CyZn superoxide dismutase with nitric oxide, nitrogen dioxide, nitrite, and nitrate. The reaction of either oxidized or reduced Cu,Zn superoxide dismutase with up to 4 mM nitric oxide or nitrogen dioxide in phosphate buffer at pH 7.4 did not result in a significant formation of nitrotyrosine. Nitrotyrosine was not formed after the reaction of either oxidized or reduced Cu,Zn superoxide dismutase with up to 63 mM nitrite or nitrate at pH 7.4. Furthermore, the production of nitrogen dioxide from 1 mM peroxynitrite in the presence of Cu,Zn superoxide dismutase in phosphate buffer, pH 7.4, was only 2.0 -C 0.5 FM, which was about the same as in buffer alone. In contrast, the same amount of peroxynitrite added to 100 mM DMSO to trap hydroxyl radical increased nitrogen dioxide to 39 + 2 PM in the presence or absence of superoxide dismutase. The yield of nitrotyrosine formed by 1 mM peroxynitrite reacted with 2 mg/ml Cu,Zn superoxide dismutase was the same (94 PM) in the presence of 100 mM DMSO as in its absence even though the concentration of nitrogen dioxide was 20-fold greater. Manganese and iron superoxide dismutase. Self nitration of tyrosine residues was also catalyzed by Mn superoxide dismutase and Fe superoxide dismutase. Six of the seven available tyrosines of the E. coli Mn superoxide dismutase were nitrated by 12 mM peroxynitrite. However, unlike Cu,Zn superoxide dismutase, Mn superoxide dismutase and Fe superoxide dismutase were inactivated by peroxynitrite. The enzymatic activity for superoxide of 1 mg/ml Cu,Zn superoxide dismutase was unaffected after treatment with up to 10 mM peroxynitrite as measured by standard superoxide dismutase assays based upon in-
TABLE
II
Superoxide Scavenging Activity after Reaction with Peroxynitrite -Peroxynitrite Cu,Zn superoxide dismutase Mn superoxide dismutase Fe superoxide dismutase Note.
1 mg/ml
mM phosphate,
was determined 4. * Statistically
4600 k 260 3200 + 170 4300 f 440
+Peroxynitrite 4700 f 390 970 + 230* 1100 + 240*
protein was reacted with 1 mM peroxynitrite in 100 pH 7.4. Superoxide dismutase activity (U/mg protein) by the xanthine oxidase cytochrome c reduction, n = significant
at
P < 0.05.
435
NITRATION
TABLE
III
Reaction of Peroxynitrite with Copper Containing Proteins Tyrosine
content bM)’
Azurin (40) Cu,Zn SOD Hemocyanin Ceruloplasmin
(84) (357)* (471)
Nitrotyrosine (PM) 17 83 85 18
f + + +
%nitration 1 2 1 1
Note. Nitrotyrosine formation after reacting the proteins peroxynitrite in phosphate, pH 7.5 (n = 3). “The tyrosine content was estimated from the known sequences. * The tyrosine content of hemocyanin is an approximation protein forms aggregates of unknown size.
43 100 24 4 with amino
2 mM acid
since the
hibition of cytochrome c reduction. At the same protein concentration, Mn superoxide dismutase was inactivated by 70% and Fe superoxide dismutase by 75% after reacting with 1 mM peroxynitrite (Table II). Higher peroxynitrite concentrations (>5 mM) inhibited more than 95% the activity of Mn superoxide dismutase and Fe superoxide dismutase. In dilute Cu,Zn superoxide dismutase solutions (10 pg/ml), up to 30% of activity was lost when greater than 10 mM peroxynitrite was added. Reaction of peroxynitrite with other copper-containing proteins. The reaction of peroxynitrite with coppercontaining proteins in phosphate buffer, pH 7.4, resulted in the formation of nitrotyrosine evidenced by the appearance of a peak between 420 and 440 nm at alkaline pH. Table III shows that 2 mM peroxynitrite nitrated all the tyrosine residues in Cu,Zn superoxide dismutase, 43% of the tyrosine residues in azurin, approximately 24% in hemocyanin, and only 4% in ceruloplasmin. DISCUSSION Peroxynitrite reacts with Cu,Zn, Mn, and Fe superoxide dismutases to generate a potent nitrating agent that modifies tyrosine residues. Nitration of tyrosine residue 108 in bovine Cu,Zn superoxide dismutase was the only apparent modification in the X-ray structure (26). However, other less visible alterations most likely occurred randomly to other amino acids resulting in the multiple charged bands observed on the native gel electrophoresis (Fig. 4). The nitrating agent was not nitric oxide nor nitrogen dioxide because they were not significant products of the superoxide dismutase reaction and adding up to 4 mM concentrations of either compound did not result in significant nitration of tyrosine 108. Furthermore, removal of the copper from Cu,Zn superoxide dismutase prevented nitration, while nitrating activity was restored by replacing copper in the active site. To account for the essential role of copper in the
436
ISCHIROPOULOS
active site of superoxide dismutase, we propose that peroxynitrite anion is attracted to the active site of superoxide dismutase by the same electrostatic field that attracts the negatively charged superoxide anion (Fig. 6). The active site of Cu,Zn superoxide dismutase contains a deep hydrophobic pocket shaped to accommodate superoxide with the copper atom sitting at the base (34,35). Thus, peroxynitrite in the trans-configuration could fit into the active site with the O-O group sitting in the pocket and the -N=O remaining exposed to solvent. Once in the active site, one electron from peroxynitrite may be transferred to the copper to form a cuprous intermediate. ONOO-
+ Cuf2,Zn
SOD *ON00
-
Cu+l,Zn
SOD
We have recently estimated that the redox potential at pH 7.0 between the nitrosodioxyl radical (ON00 * ) and peroxynitrite anion to be 0.43 V (36), which is close to the reported midpoint potential of 0.42 V for the copper in superoxide dismutase (34, 37). Thus, peroxynitrite could reduce the active site copper. During the second step in the catalytic cycle of superoxide dismutation, a hydrogen ion is donated from one of the histidines involved in binding the copper. A similar event could occur with peroxynitrite, initiating the decomposition of peroxynitrite to form nitronium ion while leaving a hydroxide ion bound to the copper. ON00
-
Cu+‘,Zn
SOD + H+ * NO;
+ HO -
Cu+‘,Zn
SOD
Addition of a second hydrogen ion from the solvent will release water from the copper, regenerating native cupric superoxide dismutase. Nitronium ion is a strong oxidant [E,’ = 1.6 V, Ref. (36)] and the active agent in fuming nitric acid (a caustic mixture of sulfuric and nitric acid) commonly used to prepare nitrophenols in organic chemistry. We have recently calculated that the energy for heterolytic cleavage of peroxynitrite to nitronium ion and water requires approximately 13 kcal. mol-l at pH 7.0 (36). In an accompanying paper (38) the activation energy for phenolic nitration by peroxynitrite catalyzed by Fe+3EDTA was found to be 12 kcal. malll. How could generation of nitronium ion at the active site result in nitration of tyrosine 108 located 18 A distant from the copper active site? Possibly, the positively charged nitronium ion could be repelled by the same electrostatic field that attracts superoxide and peroxynitrite into the active site and then be guided to tyrosine 108 by the adjacent negatively charged glutamate residue 107. However, nitronium ion reacts rapidly with water to yield nitrate and thus might not survive long enough to reach tyrosine. Furthermore, such a mechanism also does not
ET
AL.
-0,0/N
3
s
9 OXN
FIG. 6. Proposed model for the generation of a strong nitrating in the active site of Cu,Zn superoxide dismutase. The superoxide mutase schematic is derived from Getzoff et al. (34).
agent dis-
account for the second-order kinetics of nitration observed with respect to superoxide dismutase concentration. The second-order kinetics suggest that a transient intermediate is formed in or near the active site, which nitrates tyrosine upon collision with a second superoxide dismutase molecule. The intermediate may simply be peroxynitrite bound to copper in the active site. A second possibility is that nitronium ion reacts with the negatively charged glutamate residue 131 located only 8 A from the active site copper to form a highly reactive nitrocarboxylate adduct (R-C02-N02). Glutamate 131 is a highly conserved amino acid in the Cu,Zn superoxide dismutase family, which helps generate the electrostatic field that guides superoxide into the active site (34, 39). The negative charged carboxylic acid is oriented toward the active site in an ideal position to react with nitronium released from the heterolytic cleavage of peroxynitrite bound to the copper (Fig. 6). Nitrocarboxylates are well-known nitrating agents for the preparation of nitrophenols. In the absence of exogenous substrates, the nitroglutamate could slowly nitrate tyrosine 108 on other superoxide dismutase molecules by a second-order reaction rate proceeding at 1.0 f 0.1 M-’ * s-l. In support of the above mechanism, we found that a small amount of Cu,Zn superoxide dismutase will catalyze the nitration of tyrosine residues on Cu-depleted,Zn superoxide dismutase. Furthermore, the electrophoretic modifications to peroxynitrite caused by peroxynitrite were competitively inhibited by exogenous phenol and by lysozyme (Fig. 4). Peroxynitrite did not inactivate bovine Cu,Zn superoxide dismutase whereas E. coli Mn and Fe superoxide dismutases rapidly and irreversibly lost activity (Table II). The inactivation may be due to nitration of a tyrosine in the active site (residues 34 or 36 in the bacterial Mn superoxide dismutase sequences), which is involved in catalysis of superoxide dismutation (40). The resulting steric constraint and additional partial negative charge in the active site would then inhibit further reaction with superoxide. The active site tyrosine in Mn superoxide dis-
SUPEROXIDE
DISMUTASE-CATALYZED
mutase is resistant to nitration by tetranitromethane (411, suggesting that peroxynitrite could also be a useful alternative for protein modification. Superoxide dismutase catalyzes the nitration of many phenolics or tyrosine by peroxynitrite, which is further characterized in the second paper in this series. This nitrating reaction may be pathologically important by nitrating critical tyrosines on proteins important for cell regulation. Furthermore, superoxide dismutase-catalyzed phenolic nitration offers a useful assay for peroxynitrite. The competing reaction with superoxide dismutation can be minimized by treatment of Cu,Zn superoxide dismutase with HzOz and phenylglyoxal, while the Cu-depleted, Zn superoxide dismutase can serve as an inactive control. This assay method has allowed us to measure peroxynitrite formed by activated rat alveolar macrophages (19). ACKNOWLEDGMENTS This work was supported by a grant from the National Institute of Health HL 46407 and a Grant-in-Aid from the American Heart Association. J. S. Beckman is an Established Investigator of the American Heart Association. We thank Drs. H. C. Cheung and S-H. Lin for their assistance in the stopped flow experiments, C. Gunn for assisting with the detection of nitrogen dioxide, and Dr. R. Radi (University of the Republic, Montevideo, Uruguay) for helpful discussions.
REFERENCES 1. Palmer,
R. M. J., Ferrige,
A. G., and Moncada,
S. (1987)
Nature
327,524-526. 2. Garthwaite, J., Charles, S. L., and Chess-Williams, R. (1988) Nature 336,385-388. 3. Curran, R. D., Billiar, T. R., Stuehr, D. J., and Hofmann, K. (1989) J. Exp. Med. 170,1769-1774. 4. Schmidt, H. H. H. W., Seifert, R., and Bohme, E. (1989) FEBS Lett 244,357-360. 5. Hibbs, J. B., Jr., Taintor, R., and Vavrin, Z. (1987) Science 236, 473-476. 6. Marletta, M. A., Yoon, P. S., Iyengar, R., Leaf, C. D., and Wishnok, J. S. (1988) 7. Blough,
Biochemistry
0. C. (1985)
Znorg.
Chem.
24,
3504-
3505. 8. Saran, mun.
M., Michel,
C., and Bors,
W. (1990)
Free Radical
Res. Com-
10,221-226.
9. Gryglewski,
R. J., Palmer,
R. M. J., and Moncada,
S. (1986)
Nature
320,454-456. 10. Rubanyi, G. M., and Vanhoutte, H822-H827. 11. Moncada, S., Palmer, Natl. Acad. Sci. USA
P. M. (1986)
Am. J. Physiol.
250,
R. J. (1986)
Proc.
R. M. J., and Gryglewski, 83,9164-9168.
12. Miigge, A., Elwell, J. H., Peterson, Am. J. Physiol. 260, C219-C225.
T. E., and Harrison,
13. Gorren, A. C. F., de Boer, E., and Wever, Acta 916,38-47.
R. (1987)
B&him.
D. G. (1991) Biophys.
437
14. Murphy, M., and Sies, H. (1991) Proc. N&l. Acad. Sci. USA 88, 10860-10864. 15. Donald, C. E., Hughes, M. N., Thompson, J. M., and Bonner, F. T. (1986) Znorg. Chem. 25,2676-2677. 16. Beckman, J. S., Beckman, T. W., Chen, J., Marshall, P. M., and Freeman, B. A. (1990) Proc. Natl. Acad. Sci. USA 87, 1621-1624. 17. Radi, R., Beckman, J. S., Bush, K. M., and Freeman, B. A. (1991) J. Biol. Chem. 266, 4244-4250. 18. Radi, R., Beckman, J. S., Bush, K. M., and Freeman, B. A. (1991) Arch. Biochem. Biophys. 288,481-487. 19. Ischiropoulos, H., Zhu, L., and Beckman, J. S. (1992) Arch. Biochem. Biophys. 298,446-451. 20. Wang, J-F., Komarov, P., Sies, H., and de Groot, H. (1991) B&hem. J. 279,311-314. 21. Tamura, Y., Liguo, C., Driscoll, E. M., Jr., Hoff, P. T., Freeman, B. A., Gallagher, K. P., and Lucchesi, B. R. (1988) Circ. Res. 63, 944-959. 22. Liu, T. H., Beckman, J. S., Freeman, B. A., Hogan, E. L., and Hsu, C. Y. (1989) Am. J. Physiol. 256, H589-H593. 23. Imaizumi, S., Wollworth, V., Fishman, R. A., and Chan, P. H. (1990) Stroke 21, 1312-1317. 24. Kawamoto, S., Inoue, M., Tashiro, S., Morino, Y., and Miyauchi, Y. (1990) Arch. Biochem. Biophys. 277, 160-165. 25. Beckman, J. S. (1990) Nature 346, 27-28. 26. Smith, C. D., Carson, M., van der Woerd, M., Chen, J., Ischiropoulos, H., and Beckman, J. S. Submitted for publication. 27. McCord, J. M., and Fridovich, I. (1969) J. Biol. Chem. 244, 6049-
6055. 28. Viglino,
P., Scarpa, M., Cocco, D., and Rigo, A. (1985) Biochem. J. 229,87-90. 29. Rotilio, G., Calabrese, L., Bossa, F., Barra, D., Finazzi-Agro, A., and Mondovi, B. (1972) Biochemistry 11,2182-2186. 30. Beyer, W. F., Jr., Fridovich, I., Mullenbach, G. T., and Hallewell, R. (1987) J. Biol. Chem. 262,11182-11187. 31. Hodgson, E. K., and Fridovich, I. (1975) Biochemistry 14, 5294-
5299. 32. Glazer, 33. 34. 35.
27,8706-8711.
N. V., and Zafiriou,
NITRATION
36. 37.
A. N., Delange, R. J., and Sigman, D. S. (1975) Chemical Modification of Proteins, pp. 95-99, North-Holland, Amsterdam. Izzo, G. E., Jordan, F., and Mendelsohn, R. (1982) J. Am. Chem. Sot. 104,3178-3182. Tainer, J. A., Getzoff, E. D., Richardson, J. S., and Richardson, D. C. (1983) Nature 306, 284-287. Getzoff, E. D., Tainer, J. A., Weiner, P. K., Kollman, P. A., Richardson, J. S., and Richardson, D. C. (1983) Nature 306, 287-290. Koppenol, W. H., Moreno, J. J., Pryor, W. A., Ischiropoulos, H., and Beckman, J. S. Chem. Res. Toxicol., in press. Lawrence, G. D., and Sawyer, D. T. (1979) Biochemistry 18,3045-
3050. 38. Beckman,
J. S., Ischiropoulos, H., Zhu, L., Chen, J., van der Woerd, M., Smith, C. D., Martin, J. C., and Tsai, M. (1992) Arch. Biochem. Biophys. 298,438-445. 39. Koppenol, W. H. (1981) in Oxygen and Oxyradicals in Chemistry and Biology (Rodgers, M. A. J., and Powers, E. L., Eds.), pp. 671674, Academic Press, New York. 40. Stallings, W. C., Pattridge, K. A., Strong, R. K., and Ludwig, M. L. (1985) J. Biol. Chem. 260, 16424-16432. 41. Borders, C. L., Jr., Horton, P. J., and Beyer, W. F., Jr. (1989) Arch. Biochem. Biophys. 268, 74-80.
