Spin trapping endogenous radicals in MC-1010 cells: Evidence for hydroxyl radical and carbon-centered ascorbyl radical adducts

Incubation of MC-1010 cells with the spin-trapping agent 5,5-dimethyl-1-pyrroline 1-oxide (DMPO) followed by brief treatment with the solid oxidant lead dioxide (PbO₂) yielded, after filtration, a cell-free solution that contained two nitroxyl adducts. The first was the hydroxyl radical adduct, 5,5-dimethyl-2-hydroxypyrrolidine-1-oxyl (DMPO-OH), which formed immediately upon PbO₂ oxidation. The second had a 6-line EPR spectrum typical of a carbon-centered radical (A_N=15.9 G; A_H=22.4 G) and formed more slowly. No radical signals were detected in the absence of either cells or PbO₂ treatment. The 6-line spectrum could be duplicated in model systems that contained ascorbate, DMPO and DMPO-OH, where the latter was formed from hydroxyl radicals generated by sonolysis or the cleavage of hydrogen peroxide with Fe²⁺ (Fenton reaction). In addition, enrichment of MC-1010 cells with ascorbate prior to spin trapping yielded the 6-line EPR spectrum as the principal adduct following PbO₂ oxidation and filtration. These results suggest that ascorbate reacted with DMPO-OH to form a carbon-centered ascorbyl radical that was subsequently trapped by DMPO. The requirement for mild oxidation to detect the hydroxyl radical adduct suggests that DMPO-OH formed in the cells was reduced to an EPR-silent form (i.e., the hydroxylamine derivative). Alternatively, the hydroxylamine derivative was the species initially formed. The evidence for endogenous hydroxyl radical formation in unstimulated leukocytes may be relevant to the leukemic nature of the MC-1010 cell line. The spin trapping of the ascorbyl radical is the first report of formation of the carbon-centered ascorbyl radical by means other than pulse radiolysis. Unless it is spin trapped, the carbon-centered ascorbyl radical immediately rearranges to the more stable oxygen-centered species that is passive to spin trapping and characterized by the well-known EPR doublet of A_H4=1.8 G.Abbreviation EPR Electron Paramagnetic Resonance     

Studies on the photolytic breakdown of hydroperoxides and peroxidized fatty acids by using electron spin resonance spectroscopy. Spin trapping of alkoxyl and peroxyl radicals in organic solvents.

Spin trapping using 5,5-dimethyl-1-pyrroline N-oxide (DMPO) has been used to detect and distinguish between the carbon-centred, alkoxyl, and peroxyl radicals produced during the photolytic decomposition of hydroperoxides. Photolysis of tert-butyl and cumene hydroperoxides, and peroxidized fatty acids, in toluene, with low levels of u.v. light, is shown to lead to the initial production of alkoxyl radicals by homolysis of the oxygen-oxygen bond. Subsequent reaction of these radicals with excess hydroperoxide leads, by hydrogen abstraction, to the production of peroxyl radicals that can be detected as their corresponding adducts with the spin trap. Subsequent breakdown of these adducts produces alkoxyl radicals and a further species that is believed to be the oxidized spin-trap radical 5,5-dimethyl-1-pyrrolidone-2-oxyl. No evidence was obtained at low hydroperoxide concentrations, with either the cumene or lipid alkoxyl radicals, for the occurrence of beta-scission reactions; the production of low levels of carbon-centred radicals is believed to be due to the alternative reactions of hydrogen abstraction, ring closure, and/or 1,2 hydrogen shifts. Analogous experiments with 3,3,5,5-tetramethyl-1-pyrroline N-oxide (TMPO) led only to the trapping of alkoxyl radicals with no evidence for peroxyl radical adducts, this is presumably due to a decreased rate of radical addition because of increased steric hindrance.     

Spin adducts of superoxide,alkoxyl, and lipid-derived radicals with EMPO and its derivatives

The compound 5-(ethoxycarbonyl)-5-methyl-1-pyrroline N-oxide (EMPO) is a hydrophilic cyclic nitrone spin trap, which, in contrast to DMPO, forms a relatively stable superoxide adduct (t(1/2)=8.6 min) with an EPR spectrum similar to the respective DMPO adduct. In order to find the optimal degree of lipophilicity of this novel type of spin trap with respect to the detection of radicals formed during lipid peroxidation, the ethoxy group of EMPO was replaced by alkoxy substituents of increasing chain length, leading to the methoxy- (MeMPO), 1-propoxy- (PrMPO), 1-butoxy- (BuMPO), and 1-octyloxy- (OcMPO) derivatives of EMPO. The stability of their superoxide adducts was found to be strongly dependent on the size of the alkoxycarbonyl group. Increasing chain length of the alkoxyl substituent decreased the stability of alkoxyl radical adducts of MeMPO, EMPO, and PrMPO, but increased the stability of OcMPO adducts. The stability of alkoxyl radical adducts of BuMPO, on the other hand, were practically independent of the size of the alkoxyl group. Detection of lipid alkoxyl radicals formed by peroxidizing linoleic acid in a stationary system was therefore only possible with the most lipophilic spin trap, OcMPO. However, with the more hydrophilic spin traps MeMPO, EMPO, PrMPO, and BuMPO optimal EPR signal intensity could be obtained when a slow-flow system was used. Thus, within this series EMPO is the best spin trap for the detection of superoxide; OcMPO, on the other hand, is most suitable for the detection of lipid alkoxyl radicals.     

Reassignment of organic peroxyl radical adducts.

The study of the important role of peroxyl radicals in biological systems is limited by their difficult detection with direct electron spin resonance (ESR). Many ESR spectra were assigned to 5,5-dimethyl-1-pyrroline N-oxide (DMPO)/peroxyl radical adducts based only on the close similarity of their ESR spectra to that of DMPO/superoxide radical adduct in conjunction with their insensitivity to superoxide dismutase, which distinguishes the radical adduct from DMPO/superoxide radical adduct. Later, the spin-trapping literature reported that DMPO/peroxyl radical adducts have virtually the same hyperfine coupling constants as synthesized alkoxyl radical adducts, raising the issue of the correct assignment of peroxyl radical adducts. However, using 17O-isotope labelling, the methylperoxyl and methoxyl radical adducts should be distinguishable. We have reinvestigated the spin trapping of the methylperoxyl radical. The methylperoxyl radical was generated in aerobic solution with 17O-molecular oxygen either in a Fenton system with dimethylsulfoxide or in a chloroperoxidase system with tert-butyl hydroperoxide. Two different spin traps, DMPO and 2,2,4-trimethyl-2H-imidazole-1-oxide (TMIO), were used to trap methylperoxyl radical. 17O-labelled methanol was used to synthesize methoxyl radical adducts by nucleophylic addition. It was shown that the 17O hyperfine coupling constants of radical adducts formed in methylperoxyl radical-generating systems are identical to that of the methoxyl radical adduct. Therefore, methylperoxyl radical-producing systems form detectable methoxyl radical adduct, but not detectable methylperoxyl radical adducts at room temperature. One of the possible mechanisms is the decomposition of peroxyl radical adduct with the formation of secondary alkoxyl radical adduct. These results allow us to reinterpret previously published data reporting detection of peroxyl radical adducts. We suggest that detection of 17O-alkoxyl radical adduct from 17O-labelled molecular oxygen can be used as indirect evidence for peroxyl radical generation.     

Spin trapping: ESR parameters of spin adducts

Spin trapping has become a valuable tool for the study of free radicals in biology and medicine. The electron spin resonance hyperfine splitting constants of spin adducts of interest in this area are tabulated. The entries also contain a brief comment on the source of the radical trapped.     

Spin trapping of lipid radicals with DEPMPO-derived spin traps: detection of superoxide, alkyl and alkoxyl radicals in aqueous and lipid phase

The spin trap 5-(diethoxyphosphoryl)-5-methyl-1-pyrroline N-oxide (DEPMPO) forms a superoxide adduct with a half-life of almost 15 min. DEPMPO is very hydrophilic and its use for the detection of radicals in the lipid phase (lipid-derived radicals and superoxide generated in the lipid phase) is therefore limited due to its very low concentration in the lipid phase. For the detection of lipid-derived radicals, three derivatives of DEPMPO with increasing degree of lipid solubility have been investigated: 5-(di-n-propoxyphosphoryl)-5-methyl-1-pyrroline N-oxide (DPPMPO), 5-(di-n-butoxyphosphoryl)-5-methyl-1-pyrroline N-oxide (DBPMPO), and 5-(bis-(2-ethylhexyloxy)phosphoryl)-5-methyl-1-pyrroline N-oxide (DEHPMPO). As compared with the spin trap DMPO, the half-lives of the respective superoxide adducts were clearly higher in aqueous solutions of the spin traps, which facilitates qualitative ESR measurements. The stability of the superoxide spin adducts formed with the various lipophilic spin traps in aqueous buffer were similar to those observed with DEPMPO (half-life: 7-11 min.). In model experiments using Fe(3+)-catalyzed nucleophilic addition of methanol or tert-butanol to the respective spin trap the respective alkoxyl radical adducts were formed in aqueous solution as transient species in the presence of high concentrations of the alcohol. Upon dilution with water the alkoxyl group was substituted by water, giving the respective hydroxyl adduct of the spin trap. Care must therefore be taken when Fenton-type reactions are used for the generation of radicals such as the use of Fe(2+) complexes with phosphate or DTPA or inactivation of iron by addition of "Desferal" (Novarti's Pharma GmbH, Vienna, Austria) after a short incubation time. Addition of Fe(2+) under anaerobic conditions to an aqueous suspension of linoleic acid hydroperoxide and the spin trap resulted in the detection of three different species: a carbon-centered radical adduct, an acyl radical adduct, and the hydroxyl adduct. In the presence of oxygen a different species was observed with DEPMPO, DPPMPO, and DBPMPO, which was only slightly suppressed upon the addition of SOD, possibly the respective spin adduct of either the alkylperoxyl radical or, in analogy to DMPO, a secondary alkoxyl radical.     

Alkyl free radicals from the beta-scission of fatty acid alkoxyl radicals as detected by spin trapping in a lipoxygenase system

2-Methyl-2-nitrosopropane (tNB)-radical adducts from incubation mixtures of fatty acids and soybean lipoxygenase in borate buffer (pH 9.0) were measured by electron paramagnetic resonance (EPR). In addition to the previously reported six-line signal of secondary carbon-centered radicals (RCHR'), a weak signal submerged in the baseline was detected after the peroxidation phase was finished. We propose that this radical is a decomposition product formed via beta-scission of fatty acid alkoxyl radicals. EPR spectra of tNB-radical adducts formed in mixtures of either linoleic acid, arachidonic acid, or 15-hydroperoxyeicosatetraenoic acid with lipoxygenase exhibited hyperfine structure characteristic of tNB/.CH2CH2-R with hyperfine coupling constants: aN = 17.1 G; aH beta = 11.2 G (2H); and aH gamma = 0.6 G (2H). In the case of linolenic acid, this radical tNB/.CH=CH-R' with hyperfine coupling constants: aN = 17.1 G; aH beta = 10.9 G (2H); aH gamma = 1.1 G; and aH delta = 0.5 G. In accord with the decomposition scheme of hydroperoxides derived from unsaturated fatty acids, the radical adducts tNB/.CH2CH2-R and tNB/.CH2-CH=CH-R' were assigned as the pentyl and 2-pentenyl radicals, respectively.     

Spin trapping of lipid-derived radicals in liposomes

Electron-spin resonance-spin trapping has been used to detect lipid-derived radicals in liposomes. Using the lipid-soluble spin trap 2-methyl-nitrosopropane (MNP), we have detected both the lipid and hydrogen-atom spin adducts in liposomes composed of a fully saturated phospholipid (dimyristoylphosphatidylcholine, DMPC) with various mol fractions of unsaturated phospholipid (1-palmitoyl-2-arachidonoylphosphatidylcholine, PAPC) or fatty acid (arachidonic acid, AA). The lipid-derived spin adduct formed during autoxidation of liposomes was separated by thin-layer chromatography and found to co-migrate with the product(s) formed by direct addition of MNP to the corresponding unsaturated lipid or fatty acid. Both the MNP-PAPC and MNP-AA spin adducts showed some restriction of rotational motion when in the liposome bilayer (rotational correlation times 0.72 and 0.69.10(-9) s, respectively), and nitrogen hyperfine coupling constants (14.94-14.96 G) consistent with a hydrophobic localization. Radical versus non-radical mechanisms of spin adduct formation during liposome autoxidation were separated using alpha-tocopherol as a radical scavenger. The utility of nitroso spin traps in trapping of radicals in liposomes is discussed.     

Superoxide and peroxyl radical generation from the reduction of polyunsaturated fatty acid hydroperoxides by soybean lipoxygenase.

Soybean lipoxygenase is shown to catalyze the breakdown of polyunsaturated fatty acid hydroperoxides to produce superoxide radical anion as detected by spin trapping with 5,5-dimethyl-1-pyrroline-N-oxide (DMPO). In addition to the DMPO/superoxide radical adduct, the adducts of peroxyl, acyl, carbon-centered, and hydroxyl radicals were identified in incubations containing linoleic acid and lipoxygenase. These DMPO radical adducts were observed just prior to the system becoming anaerobic. Only a carbon-centered radical adduct was observed under anaerobic conditions. The superoxide radical production required the presence of fatty acid substrates, fatty acid hydroperoxides, active lipoxygenase, and molecular oxygen. Superoxide radical production was inhibited when nordihydroguaiaretic acid, butylated hydroxytoluene, or butylated hydroxyanisole was added to the incubation mixtures. We propose that polyunsaturated fatty acid hydroperoxides are reduced to form alkoxyl radicals and that after an intramolecular rearrangement, the resulting hydroxyalkyl radical reacts with oxygen, forming a peroxyl radical which subsequently eliminates superoxide radical anion.     

Detection of peroxyl and alkoxyl radicals produced by reaction of hydroperoxides with rat liver microsomal fractions.

E.s.r. spin trapping using the spin trap 5,5-dimethyl-1-pyrroline N-oxide (DMPO) was used to detect peroxyl, alkoxyl and carbon-centred radicals produced by reaction of t-butyl hydroperoxide (tBuOOH) with rat liver microsomal fraction. The similarity of the hyperfine coupling constants of the peroxyl and alkoxyl radical adducts to those obtained previously with isolated enzymes suggests that these species are the tBuOO. and tBuO. adducts. The effects of metal-ion chelators, heat denaturation, enzyme inhibitors and reducing equivalents demonstrate that these species arise from reaction of tBuOOH with a haem enzyme such as cytochrome P-450 or cytochrome b5. In the absence of NADPH or NADH the previously undetected peroxyl radical adduct is the major species observed. In the presence of these reducing equivalents the alkoxyl and carbon-centred radical adducts predominate, which is in accord with product studies on similar systems. These results demonstrate that both reductive and oxidative decomposition of tBuOOH can occur in rat liver microsomal fraction with the reductive pathway favoured in the presence of NADH or NADPH.     

Lipid peroxyl radical intermediates in the peroxidation of polyunsaturated fatty acids by lipoxygenase. Direct electron spin resonance investigations

Lipid peroxyl radicals resulting from the peroxidation of polyunsaturated fatty acids by soybean lipoxygenase were directly detected by the method of rapid mixing, continuous-flow electron spin resonance spectroscopy. When air-saturated borate buffer (pH 9.0) containing linoleic acid or arachidonate acid was mixed with lipoxygenase, fatty acid-derived peroxyl free radicals were readily detected; these radicals have a characteristic g-value of 2.014. An organic free radical (g = 2.004) was also detected; this may be the carbon-centered fatty acid free radical that is the precursor of the peroxyl free radical. The ESR spectrum of this species was not resolved, so the identification of this free radical was not possible. Fatty acids without at least two double bonds (e.g. stearic acid and oleic acid) did not give the corresponding peroxyl free radicals, suggesting that the formation of bisallylic carbon-centered radicals precedes peroxyl radical formation. The 3.8-G doublet feature of the fatty acid peroxyl spectrum was proven (by selective deuteration) to be a hyperfine coupling due to a gamma-hydrogen that originated as a vinylic hydrogen of arachidonate. Arachidonate peroxyl radical formation was shown to be dependent on the substrate, active lipoxygenase, and molecular oxygen. Antioxidants are known to protect polyunsaturated fatty acids from peroxidation by scavenging peroxyl radicals and thus breaking the free radical chain reaction. Therefore, the peroxyl signal intensity from micellar arachidonate solutions was monitored as a function of the antioxidant concentration. The reaction of the peroxyl free radical with Trolox C was shown to be 10 times slower than that with vitamin E. The vitamin E and Trolox C phenoxyl radicals that resulted from scavenging the peroxyl radical were also detected.     

Strand cleavage of supercoiled DNA by water-soluble peroxyl radicals. The overlooked importance of peroxyl radical charge

It is well established that the peroxyl radicals formed during the thermal decomposition of 2,2'-azobis(amidinopropane), ABAP, in oxygenated water can cleave double-stranded DNA, from which fact it has been concluded that peroxyl radicals, as a general class, can induce DNA strand scission. However, the ABAP-derived radicals are positively charged, and DNA is a negatively charged polyanion. Moreover, the relatively small and, therefore, free to diffuse peroxyl radicals likely to be formed in vivo will generally be negatively charged or neutral. Plasmid supercoiled DNA [pBR 322, 4361 base pairs (bp)] was reacted with known, equal fluxes of two positively charged peroxyl radicals, a negatively charged peroxyl radical, and a neutral peroxyl radical. The two positively charged peroxyl radicals degraded >/=80% of the supercoiled pBR 322 at a flux of 4 radicals/bp, but the negatively charged and neutral peroxyl radicals had no significant effect even at a flux as high as 24 radicals/bp. The same lack of effect on the DNA was also observed with high fluxes of superoxide/hydroperoxyl radicals. Similar results were obtained with another supercoiled DNA, pUC 19, except that pUC 19 is somewhat more sensitive to strand scission by positively charged peroxyl radicals than pBR 322. We conclude that most of the peroxyl radicals likely to be formed in vivo have little or no ability to induce DNA strand scission and that the potential role of electrostatics in radical/DNA reactions should always be considered.     

Detection of peroxyl and alkoxyl radicals produced by reaction of hydroperoxides with heme-proteins by electron spin resonance spectroscopy

ESR spin trapping using the spin trap 5,5-dimethyl-1-pyrroline N-oxide (DMPO) has been used to directly detect alkoxyl radicals (with hyperfine coupling constants aN 1.488, aH 1.600 mT and aN 1.488, aH 1.504 mT for the tBuO. and PhC(CH3)2O. adducts, respectively) and peroxyl radicals (aN 1.448, aH 1.088, aH 0.130 mT and aN 1.456, aH 1.064, aH 0.128 mT for the tBuOO. and PhC(CH3)2OO. adducts, respectively) produced from t-butyl or cumene hydroperoxides by a variety of heme-containing substances (purified cytochrome P-450, metmyoglobin, oxyhemoglobin, methemoglobin, cytochrome c, catalase, horseradish peroxidase) and the model compound hematin. The observed species exhibit a complicated dependence on reagent concentrations and time, with maximum concentrations of the peroxyl radical adducts being observed immediately after mixing of the hydroperoxide with low concentrations of the heme-compound. Experiments with inhibitors (CN-, N3-, CO, metyrapone and imidazole) suggest that the major mechanism of peroxyl radical production involves high-valence-state iron complexes in a reaction analogous to the classical peroxidase pathway. The production of alkoxyl radicals is shown to arise mainly from the breakdown of peroxyl radical spin adducts, with direct production from the hydroperoxide being a relatively minor process.     

Spin traps: in vitro toxicity and stability of radical adducts

We have evaluated the effects of DMPO, CMPO, EMPO, BMPO, and DEPMPO on functioning CHO cells and the stability of the radical adducts in the presence of cells. The potential toxic effects of the spin traps were measured by two estimates of cell viability (trypan blue exclusion and colony formation) and one of cell function (rate of oxygen consumption). We also studied the effects of the spin traps on colony formation in a second cell line, 9L tumor cells. Toxicity varied with the type of cell line and the parameter that was measured. In aqueous solutions the order of stability for all spin adducts was SO(3) > OH > CH(3), while in cell suspensions it was SO(3) > OH approximately CH(3). The radical adducts of the new spin traps have significantly increased stability as compared to DMPO. These results indicate that the new spin traps potentially offer increased stability of spin adducts in functioning cells. It also is clear that it is necessary to carry out appropriate studies of the stability and toxicity in the system that is to be studied for any particular use of these spin traps. It then should be feasible to select the spin trap(s) best suited for the proposed study.     

Oxygen radical chemistry of polyunsaturated fatty acids

Polyunsaturated fatty acids (PUFA) are readily susceptible to autoxidation. A chain oxidation of PUFA is initiated by hydrogen abstraction from allylic or bis-allylic positions leading to oxygenation and subsequent formation of peroxyl radicals. In media of low hydrogen-donating capacity the peroxyl radical is free to react further by competitive pathways resulting in cyclic peroxides, double bond isomerization and formation of dimers and oligomers. In the presence of good hydrogen donators, such as alpha-tocopherol or PUFA themselves, the peroxyl radical abstracts hydrogen to furnish PUFA hydroperoxides. Given the proper conditions or catalysts, the hydroperoxides are prone to further transformations by free radical routes. Homolytic cleavage of the hydroperoxy group can afford either a peroxyl radical or an alkoxyl radical. The products of peroxyl radicals are identical to those obtained during autoxidation of PUFA; that is, it makes no difference whether the peroxyl radical is generated in the process of autoxidation or from a performed hydroperoxide. Of particular interest is the intramolecular rearrangement of peroxyl radicals to furnish cyclic peroxides and prostaglandin-like bicyclo endoperoxides. Other principal peroxyl radical reactions are the beta-scission of O2, intermolecular addition and self-combination. Alkoxyl radicals of PUFA, contrary to popular belief, do not significantly abstract hydrogens, but rather are channeled into epoxide formation through intramolecular rearrangement. Other significant reactions of PUFA alkoxyl radicals are beta-scission of the fatty chain and possibly the formation of ether-linked dimers and oligomers. Although homolytic reactions of PUFA hydroperoxides have received the most attention, hydroperoxides are also susceptible to heterolytic transformations, such as nucleophilic displacement and acid-catalyzed rearrangement.     

Detection of lipid radicals by electron paramagnetic resonance spin trapping using intact cells enriched with polyunsaturated fatty acid.

Electron paramagnetic resonance (EPR) spin trapping was used to detect lipid-derived free radicals generated by iron-induced oxidative stress in intact cells. Using the spin trap alpha-(4-pyridyl 1-oxide)-N-tert-butylnitrone (POBN), carbon-centered radical adducts were detected. These lipid-derived free radicals were formed during incubation of ferrous iron with U937 cells that were enriched with docosahexaenoic acid (22:6n-3). The EPR spectra exhibited apparent hyperfine splittings characteristic of a POBN/alkyl radical, aN = 15.63 +/- 0.06 G and aH = 2.66 +/- 0.03 G, generated as a result of beta-scission of alkoxyl radicals. Spin adduct formation depended on the FeSO4 content of the incubation medium and the number of 22:6-enriched cells present; when the cells were enriched with oleic acid (18:1n-9), spin adducts were not detected. This is the first direct demonstration, using EPR, of a lipid-derived radical formed in intact cells in response to oxidant stress.     

A microplate-based colorimetric assay of the total peroxyl radical trapping capability of human plasma

We developed a colorimetric assay estimating the radical-scavenging activity of human plasma. The test is based on a measure, in 96-well microplates at 450 nm, of the bleaching of carotenoid crocin by peroxyl radicals generated during thermal decomposition of 2, 2'-azobis-(2-amidinopopane) dihydrochloride (ABAP). The inhibition of this bleaching is a function of the antioxidant power of substances added to incubation mixture. We determined the optimal conditions for a sensitive, rapid, and reproducible assay of 50% inhibitory capacity (IC50) of a range of antioxidant substances and of plasma. Only a total of 200 microl of plasma is required in a complete dose-inhibition curve. The IC50 of normal human plasma resulted of 2.70 microl of plasma/250 microl assay volume. The total antioxidant capability (TAC) of plasma was defined as the reciprocal of IC50 and its value in a group of 19 healthy adults resulted in 0. 369 +/- 0.06. Intraassay and interassay coefficients of variation of plasma TAC were 6.13 and 4.80%, respectively. Measurement of samples with different uric acid concentration showed that antioxidant activity of uric acid accounts for approximately two-thirds of TAC.     

Spin trapping and protein cross-linking of the lactoperoxidase protein radical

Lactoperoxidase (LPO) reacts with H(2)O(2) to sequentially give two Compound I intermediates: the first with a ferryl (Fe(IV)=O) species and a porphyrin radical cation, and the second with the same ferryl species and a presumed protein radical. However, little actual evidence is available for the protein radical. We report here that LPO reacts with the spin trap 3,5-dibromo-4-nitroso-benzenesulfonic acid to give a 1:1 protein-bound radical adduct. Furthermore, LPO undergoes the H(2)O(2)-dependent formation of dimeric and trimeric products. Proteolytic digestion and mass spectrometric analysis indicates that the dimer is held together by a dityrosine link between Tyr-289 in each of two LPO molecules. The dimer retains full catalytic activity and reacts to the same extent with the spin trap, indicating that the spin trap reacts with a radical center other than Tyr-289. The monomeric protein recovered from incubations of LPO with H(2)O(2) is fully active but no longer forms dimers when incubated with H(2)O(2), clear evidence that it has also been structurally modified. Myeloperoxidase, a naturally dimeric protein, and eosinophil peroxidase do not undergo H(2)O(2)-dependent oligomerization. Analysis of the interface in the LPO dimers indicates that the same protein surface is involved in LPO dimerization as in the normal formation of myeloperoxidase dimers. Oligomerization of LPO alters its physical properties and may alter its ability to interact with macromolecular substrates.