Hydrogen Peroxide Formation by Reaction of Peroxynitrite with HEPES and Related Tertiary Amines

Organic amine-based buffer compounds such as HEPES (Good’s buffers) are commonly applied in experimental systems, including those where the biological effects of peroxynitrite are studied. In such studies 3-morpholinosydnonimineN-ethylcarbamide (SIN-1), a compound that simultaneously releases nitric oxide (⋅NO) and superoxide (O⨪2), is often used as a source for peroxynitrite. Whereas in mere phosphate buffer H2O2 formation from 1.5 mmSIN-1 was low (∼15 μm), incubation of SIN-1 with Good’s buffer compounds resulted in continuous H2O2 formation. After 2 h of incubation of 1.5 mm SIN-1 with 20 mm HEPES about 190 μm H2O2 were formed. The same amount of H2O2 could be achieved from 1.5 mm SIN-1 by action of superoxide dismutase in the absence of HEPES. The increased H2O2 level, however, could not be related to a superoxide dismutase or to a NO scavenger activity of HEPES. On the other hand, SIN-1-mediated oxidation of both dihydrorhodamine 123 and deoxyribose as well as peroxynitrite-dependent nitration ofp-hydroxyphenylacetic acid were strongly inhibited by 20 mm HEPES. Furthermore, the peroxynitrite scavenger tryptophan significantly reduced H2O2 formation from SIN-1-HEPES interactions. These observations suggest that peroxynitrite is the initiator for the enhanced formation of H2O2. Likewise, authentic peroxynitrite (1 mm) also induced the formation of both O⨪2 and H2O2 upon addition to HEPES (400 mm)-containing solutions in a pH (4.5–7.5)-dependent manner. In accordance with previous reports it was found that at pH ≥5 oxygen is released in the decay of peroxynitrite. As a consequence, peroxynitrite(1 mm)-induced H2O2 formation (∼80 μm at pH 7.5) also occurred under hypoxic conditions. In the presence of bicarbonate/carbon dioxide (20 mm/5%) the production of H2O2 from the reaction of HEPES with peroxynitrite was even further stimulated. Addition of SIN-1 or authentic peroxynitrite to solutions of Good’s buffers resulted in the formation of piperazine-derived radical cations as detected by ESR spectroscopy. These findings suggest a mechanism for H2O2 formation in which peroxynitrite (or any strong oxidant derived from it) initially oxidizes the tertiary amine buffer compounds in a one-electron step. Subsequent deprotonation and reaction of the intermediate α-amino alkyl radicals with molecular oxygen leads to the formation of O⨪2, from which H2O2 is produced by dismutation. Hence, HEPES and similar organic buffers should be avoided in studies of oxidative compounds. Furthermore, this mechanism of H2O2formation must be regarded to be a rather general one for biological systems where sufficiently strong oxidants may interact with various biologically relevant amino-type molecules, such as ATP, creatine, or nucleic acids.

The term "peroxynitrite" is commonly used to describe the equilibrium mixture of oxoperoxonitrate(1Ϫ) (ONOO Ϫ ) and its conjugated acid, hydrogen oxoperoxonitrate(1Ϫ) (peroxynitrous acid, ONOOH). Peroxynitrite is a strong oxidant formed in the diffusion-controlled reaction of superoxide (O 2 . ) and nitric oxide (nitrogen monoxide, ⅐ NO) (k ϭ 3.9 -6.7 ϫ 10 9 M Ϫ1 s Ϫ1 ) (1,2). Peroxynitrite has been suggested to play a major role in many pathological processes like atherosclerosis (3) and stroke (4). The pathological activity of ONOO Ϫ /ONOOH is assumed to result from its ability to attack various biological targets, including protein-and non-protein sulfhydryls (5), DNA (6), low density lipoproteins (3), or membrane phospholipids (7). A favored method to generate peroxynitrite for experimental purposes is to use SIN-1 1 (8). This compound decays in solution in the presence of oxygen with simultaneous release of ⅐ NO and O 2 . in a 1:1 stoichiometry (9). Consequently, SIN-1 has been shown to attack many biological targets in nearly the same manner as authentic peroxynitrite (3). The formation of ONOO Ϫ from SIN-1 can be suppressed by the enzyme superoxide dismutase (SOD), resulting in the formation of ⅐ NO and hydrogen peroxide (H 2 O 2 ) as major products (10).
In obvious contradiction to the assumption of a central role of ONOO Ϫ /ONOOH in cell injuring processes, almost complete protection from SIN-1 cytotoxicity in experiments with rat liver endothelial cells and Fu5 rat hepatoma cells was provided by catalase but not by SOD (11,12). Since catalase does not effectively react with peroxynitrite (13), these results strongly suggest a participation of H 2 O 2 in SIN-1-mediated cytotoxicity rather than a participation of ONOO Ϫ /ONOOH. Indeed, formation of H 2 O 2 from SIN-1 was observed under certain experimental conditions (12). We (14) have recently demonstrated that the formation of H 2 O 2 , and consequently the protection exerted by catalase, decisively depends on the presence of the organic buffer compound HEPES in the incubation medium. In the absence of this "Good's buffer" neither H 2 O 2 was formed nor was catalase protective. The question, however, how HEPES and similar Good's buffers (15,16) mediate the formation of H 2 O 2 from SIN-1 remained open. The present study aims at the elucidation of the underlying chemical mechanism.
Solutions-Care was taken to exclude possible contamination by bicarbonate/carbon dioxide. Doubly distilled water was bubbled (2 liters/min) with synthetic air at room temperature for 20 min. This water was used for synthesis of oxoperoxonitrate(1Ϫ), NaOH (0.01-0.5 N), and for all other solutions. Potassium phosphate buffer (50 mM) containing DTPA (0.1 mM) was prepared freshly each day. The pH was adjusted to 7.5 at 37°C, and the solution was again bubbled (2 liters/min) with synthetic air (normoxia) or with nitrogen (hypoxia) or with the carbon dioxide mixture for 20 min. In the case of bubbling with the CO 2 mixture, the pH must be readjusted to 7.5. SIN-1 and spermine NONOate solutions were prepared as 100ϫ stock solutions at 4°C in 50 mM KH 2 PO 4 and 10 mM NaOH, respectively, and used within 15 min. DHR123 200ϫ stock solution was prepared in water-free, nitrogenpurged dimethylformamide and stored in the dark at Ϫ20°C.
Experimental Conditions-SIN-1 and spermine NONOate (final pH 7.5) were added to 1 ml of phosphate buffer and incubated in 12-well cell culture plates (volume of each well 7 ml, Falcon, Heidelberg, Germany). For the detection of H 2 O 2 with the catalase assay, SIN-1 was added to 10 ml of buffer and incubated in tissue culture dishes (75 ml, Falcon, Heidelberg, Germany). Under normoxic conditions these plates/dishes were placed in an air-tight anaerobic vessel (10 liters). During the first 15 min of each experiment these vessels were flushed (5 liters/min) with synthetic air in a warming incubator (Heraeus, Hanau, Germany). In the presence of HCO 3 Ϫ /CO 2 the plates/dishes were placed in an incubator for cell culture (37°C, humidified atmosphere of 95% authentic air and 5% CO 2 , Labotect, Göttingen, Germany). The experiments with authentic peroxynitrite were performed in reaction tubes (1.4 ml, Eppendorf, Hamburg, Germany) by using the drop-tube Vortex mixer  . This electrode was also used to detect the production of O 2 from decomposed peroxynitrite. Superoxide radicals were determined by using the modified ferricytochrome c 3ϩ reduction technique of McCord and Fridovich (19). Peroxynitrite was vortexed to the reaction solution in the absence and presence of various concentrations of HEPES. Cytochrome c (20 M) or cytochrome c plus SOD (100 units/ml) were added 2 min after peroxynitrite addition. This time lag guaranteed that neither cytochrome c nor SOD reacted directly with peroxynitrite. The resulting mixture was stored for 40 min at 37°C. The amount of cytochrome c reduction was determined by reading the absorbance at 550 nm (⌬⑀ 550 ϭ 21,000 M Ϫ1 s Ϫ1 ) (20). The amount of reduced cytochrome c in the presence of SOD was subtracted from the amount of reduced cytochrome c in the absence of SOD for each HEPES concentration. This difference was used to calculate the amount of trapped O 2 . .

Capillary Zone Electrophoresis
Measurements-SIN-1 and SIN-1C were quantified on a Beckman P/ACE 5000 apparatus. Separation conditions for SIN-1 and SIN-1C were as follows: fused silica capillary (50-cm effective length, 75 M internal diameter), hydrodynamic injection for 5 s, temperature 23°C, voltage 20 kV, normal polarity, UV detection at 280 nm, and 100 mM potassium phosphate, pH 6.3, as electrolyte system.
ESR Measurements-ESR spectra were recorded at ambient temperature on a Bruker ESP300E X-band spectrometer (Bruker, Rheinstetten, Germany) equipped with a TM 110 wide bore cavity. Solutions were prepared from 1 ml of the buffer solution (pH 7.5) containing HEPES or PIPES (both 400 mM) and SIN-1 (10 mM). Alternatively, 200 l of authentic peroxynitrite (500 mM) and 1.8 ml of the buffer solution (pH 7.5) containing HEPES or PIPES (both 1 M) were vortexed. The reaction solutions were quickly transferred to an aqueous solution quartz cell (Willmad, Buena, NY). The first spectra were run as fast as possible, i.e. within 1 min, and then in 5-min intervals. Recording conditions were as follows: microwave frequency, 9.8 GHz; modulation, 0.04 mT; signal gain, 5 ϫ 10 5 ; sweep range, 20 mT; sweep time, 4 min. Spectrum simulation was performed using the WinSim program (21).
Determination of NO Scavenger Activity-The spermine NONOate (25 M)-driven reduction of oxyhemoglobin (70 M) in the absence and presence of HEPES (1-20 mM) was used. Formation of methemoglobin was quantified spectrophotometrically at 578 nm (⑀ M ϭ 12100 M Ϫ1 cm Ϫ1 ) (12). Ϫ /CO 2 (20 mM/5%). H 2 O 2 production was quantified using the catalase assay. Data are means Ϯ S.D. of three experiments performed in duplicate. In all experiments described here 0.1 mM DTPA was present to inhibit the influence of heavy metal contaminations (25). DTPA (0.1 mM) itself did not affect H 2 O 2 formation, neither in the presence nor in the absence of HEPES (data not shown).

Peroxynitrite from SIN-1 Decomposition H 2 O 2 Production from SIN-1 in the Presence of Good's Buffers-In
The rate of H 2 O 2 generation in the presence of 20 mM HEPES was comparable to the rate of H 2 O 2 production in the presence of 100 units/ml SOD (Fig. 1). 100 units/ml was already the optimal SOD activity to induce maximal H 2 O 2 production from SIN-1 (Fig. 2B). Both decreasing the SOD activity to 1 unit/ml or increasing it to 10,000 units/ml strongly decreased H 2 O 2 formation, in full agreement with results reported by Gergel et al. (10). Thus, 20 mM HEPES already stimulated H 2 O 2 formation from SIN-1 to the maximum yield that could be obtained by SOD.
Other Good's buffers at concentrations of 20 mM were also able to induce H 2 O 2 formation from SIN-1 (Table I). The highest yield of H 2 O 2 after 2 h was observed with EPPS (ϳ207 M). HEPES, POPSO, and PIPES induced moderately less H 2 O 2 formation from SIN-1. The non-piperazine-type organic buffer triethanolamine (20 mM) also stimulated some production of H 2 O 2 from SIN-1, about 30% of the maximum amount obtained from HEPES.
SOD-like Activity-Due to the fact that in the presence of HEPES H 2 O 2 was formed at virtually the same rate as could maximally be achieved by SOD ( Fig. 1 and Fig. 2, A and B (27), we used the spermine NONOate-driven reduction of oxyhemoglobin to verify ⅐ NO scavenger properties of HEPES. However, HEPES (1-20 mM) did not inhibit methemoglobin generation (data not shown). Involvement of Peroxynitrite-Since the foregoing experiments showed that the HEPES-dependent H 2 O 2 formation from SIN-1 is neither mediated by a SOD-like activity of HEPES nor by ⅐ NO-HEPES interactions, we concluded that reaction between peroxynitrite and HEPES should be responsible for the stimulation of H 2 O 2 generation.
We (14) and others (28) have found that HEPES and SOD influence the decomposition of SIN-1. To exclude possible artifacts by an altered decomposition of this compound in the presence of HEPES or SOD, we performed the subsequent experiments as end point determinations. DHR123 is oxidized by peroxynitrite to RH123 in the presence of DTPA but neither by O 2 . nor ⅐ NO alone (29), and TBARS are formed from deoxyribose only by very strong oxidants like hydroxyl radicals (30) and peroxynitrite (31). The influence of HEPES (20 mM) and SOD (100 units/ml) on the SIN-1-driven oxidation of both DHR123 and deoxyribose is shown in Fig. 3  Influence of Tryptophan and Bicarbonate/Carbon Dioxide on H 2 O 2 Formation-To support the foregoing conclusion further, we attempted to scavenge peroxynitrite with tryptophan, which has been reported to scavenge peroxynitrite with a rate constant of k ϭ 184 M Ϫ1 s Ϫ1 (35) but should not react with ⅐ NO or O 2 . at significant rates. Accordingly, tryptophan (10 mM) inhibited the HEPES (5 and 10 mM)-stimulated H 2 O 2 production from SIN-1 by about 50% (Fig. 4A). The inhibitory effect of tryptophan decreased with increasing concentrations of HEPES and was abolished at 50 mM HEPES.
A maximum value of about 220 M H 2 O 2 was reached at 20 mM HEPES. Contrary to the bicarbonate-free system, however, higher HEPES concentrations again diminished the H 2 O 2 level. In the presence of HCO 3 Ϫ /CO 2 , tryptophan (10 mM) inhibited the HEPES (5, 10, and 20 mM)-mediated H 2 O 2 production from SIN-1 somewhat more pronounced than in its absence (Fig. 4B). These results favor the view that peroxynitrite-derived oxidants, e.g. bicarbonate radical, nitrogen dioxide, nitryl cation, that have been postulated to be produced in the presence of carbon dioxide (37) also might be able to induce H 2 O 2 production from HEPES.  SIN-1 (Fig. 4B).

Authentic Peroxynitrite
With regard to the short lifetime of peroxynitrite (t1 ⁄2 Ϸ 1 s at 37°C) H 2 O 2 production by action of authentic peroxynitrite     Formation-According to our proposed mechanism (see "Discussion") atmospheric oxygen is transformed to H 2 O 2 ; hence, H 2 O 2 production should be critically dependent on the oxygen concentration. However, the yield of H 2 O 2 was found to be largely independent on the oxygen level (Fig. 6). Somewhat surprisingly, even under hypoxic conditions about 88% of the maximum yield of H 2 O 2 was produced within 10 s. Only the minor (12%) long term production of H 2 O 2 appeared to be dependent on the oxygen level.
Oxygen and H 2 O 2 Formation under Hypoxic Conditions-The apparent independence of H 2 O 2 production on the level of dissolved oxygen implied that either peroxynitrite is directly converted to H 2 O 2 or that a substantial amount of oxygen is released in the decay of peroxynitrite. In fact, oxygen release from peroxynitrite has been reported previously (7,38,39). To check on this, we measured the release of O 2 from peroxynitrite in the presence and absence of HEPES (400 mM) (Table II). In the absence of HEPES we observed at pH 7.5 the formation of oxygen, in agreement with the above report (39). In mere phosphate buffer an oxygen production of 175 Ϯ 7.9 M was measured within 10 s after addition of peroxynitrite (1 mM). However, in the presence of DPTA (0.1 mM) the maximum yield of oxygen dropped to 132.8 Ϯ 7.8 M, in agreement with data published by Radi et al. (7). In any case, oxygen is produced at a level sufficient to account for the observed yield of H 2 O 2 (78.9 Ϯ 2.9 M). On the other hand, release of oxygen from peroxynitrite was abolished in the presence of 400 mM HEPES. In mere phosphate buffer at pH 5 the release of oxygen was reduced by about 90% (17.0 Ϯ 3.9 M). Consequently, H 2 O 2 production was strongly suppressed to 11.8 Ϯ 0.9 M in the presence of HEPES under these hypoxic conditions (Table II) (Table III). These yields (8 and 19%) are virtually identical to those found for nitration of tyrosine (ϳ6 -7 and 19%) at the same pH (40). In the presence of SOD (100 units/ ml) the yields of 3-NO 2 -4-HPA increased slightly by about 10%, in line with reports by Beckman et al. (41). The addition of HEPES strongly decreased the peroxynitrite-mediated nitration of p-HPA. At 20 mM HEPES formation of 3-NO 2 -4-HPA was reduced by 73% in the absence of HCO 3 Ϫ /CO 2 and by 64% in its presence. Increasing the HEPES concentration to 400 mM further inhibited 3-NO 2 -4-HPA formation under both conditions by approximately 90%. The effect of HEPES on the 3-NO 2 -4-HPA formation from peroxynitrite in HCO 3 Ϫ /CO 2 -free solution was half-maximal at 6 Ϯ 1 mM, in excellent agreement with the half-maximally stimulation of H 2 O 2 production (see above). In the presence of HCO 3 Ϫ /CO 2 the effect of HEPES on the half-maximal yield of H 2 O 2 formation was almost unchanged (5 Ϯ 1 mM), whereas nitration of p-HPA was halfmaximal at 9 Ϯ 1 mM. These results suggest that HEPES has been attacked by the same reactive intermediates that are responsible for the nitration of p-HPA.
Radicals from Good's Buffers-In previous studies (42, 43) fairly persistent ESR spectra have been observed from piperazine buffer compounds by action of various oxidants like Fe(II)/ H 2 O 2 or polymeric iron/O 2 . The spectra were reasonably attributed to the radical cations expected to be generated by oneelectron oxidation of the piperazine compounds (42), although the assignment appeared tentative as no detailed analysis of the spectra was performed. We therefore checked, by ESR spectroscopy, the ability of peroxynitrite to mediate radical formation from HEPES and PIPES, both by SIN-1 (10 mM) and by authentic peroxynitrite (50 mM) solutions. After mixing of the reactants by vortexing, we observed "instantaneously" (i.e. within 1 min) weak, multi-line ESR spectra (shown for authentic peroxynitrite in Fig. 7). In some experiments a 2-4-fold increase of the signal intensity was observed during the run. Both the HEPES-and PIPES-derived spectra were virtually identical to those reported in the literature (42,43). In agreement with the previous observations the signals decayed with half-lifes of about 10 -15 min at 20°C. The spectrum obtained from HEPES was too weak to be evaluated in detail, but the PIPES-derived spectrum was sufficiently intense to be completely analyzed. Elaborate analysis by spectral simulation revealed that the ESR spectrum must be interpreted as a superposition of the two radical cations R1 and R2 (Fig. 8).
The presence of at least two radicals was further indicated by slight temporal changes of the overall spectral shape, indicating somewhat different lifetimes of both species. A similar superposition has been observed previously for HEPES-derived ESR spectra (43). The presence of a small amount of a possible further oxidized species, R3, cannot be completely excluded. The identification of R1, R2 as radical cations unambiguously follows from their ESR spectral properties, that is the even number of interacting nuclei, the g factor, and the magnitude of the corresponding hyperfine splittings, all of which are strikingly similar to the data known for other piperazine and dihydropiperazine radical cations in organic solvents (44 -46 Formation-Only the amino unprotonated forms of HEPES, PIPES, etc. are amenable to oneelectron oxidation to yield radicals R1-R3 (Fig. 8). Thus, H 2 O 2 formation was expected to be pH-dependent. In the absence of HEPES a constant, low amount of H 2 O 2 (ϳ7 M) was found within the pH range 4 -8.5 after vortexing of peroxynitrite (1 mM) solutions (Fig. 10). At pH 4 no change of H 2 O 2 level was detected in the presence of HEPES (400 mM). However, with increasing pH H 2 O 2 formation strongly increased to reach a maximum value of about 90 M at pH 7.5. Further increase of the pH again diminished the production of H 2 O 2 . The same pH dependence on H 2 O 2 formation, with maxima at pH 7.5, was also observed for EPPS and MOPS (data not shown). The pH-dependent half-maximal production of H 2 O 2 correlated well with the pK a of the respective buffer compound (Table IV). In agreement with the proposal that the unprotonated forms of the tertiary amines are oxidized by peroxynitrite, the quaternary amino compound tetramethylammonium chloride (400 mM) did not stimulate H 2 O 2 formation (Table IV). DISCUSSION Reactions between authentic or in situ generated peroxynitrite and HEPES or PIPES result in the formation of HEPESand PIPES-derived radical cation species (R1 and R2) (Fig. 8) as shown by ESR spectroscopy (Fig. 7). A reasonable mechanism for the formation of R1 and R2 is displayed in Fig. 11, where [Ox] stands for the action of any sufficiently strong FIG. 11. Radical cation formation from Good's buffers by oneelectron oxidants. In the first step one-electron oxidation of the piperazine compounds produces the amine radical cation R1 which subsequently undergoes ␣-deprotonation, a common decay path for such types of radical cations (47, 58). The putative ␣-aminoalkyl radical R4 certainly is too short-lived to be observed under the conditions of the ESR experiment (see below). Further oxidation of R4 to yield R2 also is feasible, in particular in the presence of O 2 (47, 58). A further reaction of R2 to give R3 could not be verified by the ESR spectra but cannot be ruled out. oxidant in the system, e.g. peroxynitrite and peroxynitritederived oxidants formed in the absence and presence of carbon dioxide, without specifying the actual reactive species. Fig. 11 also provides the starting point for the interpretation of O 2 . formation in our system. It is common knowledge that carbon-centered radicals react rapidly (k Ϸ2 ϫ 10 9 M Ϫ1 s Ϫ1 ) with molecular oxygen to give peroxyl radicals (47, 48  49,50). In fact, this property of electron-rich alkyl radicals has recently been used as the basis for the development of a thermal superoxide source (51). By analogy, we propose that a similar mechanism is operative in the production of O 2 . from Good's buffers by attack of authentic or SIN-1-generated peroxynitrite (Fig. 12). Thus, it would appear that the rate of O 2 . production is governed by the first two steps (electron transfer and deprotonation) of the reaction sequence. On the other hand, R4 may also be generated directly from the parent compound by hydrogen abstraction by suitably reactive radicals X ⅐ , e.g. hydroxyl or peroxyl radicals. In line with this assumption, formation of R1 and R4 has been postulated in the autoxidation of DNA/Cu 2ϩ /H 2 O 2 systems in the presence of HEPES and PIPES (52). Further consequences arise from the foregoing. First, O 2 . /H 2 O 2 production is expected to be pH-dependent, decreasing with decreasing pH and reflecting the respective pK a of the organic buffer compound. This hypothesis was confirmed by the data of Fig. 10 and  (Tables I and IV) than the bis-sulfonylated POPSO and PIPES, but, vice versa, the HEPES-derived radical showed a much lower ESR signal intensity (steadystate concentration) than the PIPES-derived one.
There is no contradiction in the fact that the ESR spectra of radical cations R1 and R2 have been detected although their reaction with oxygen is extremely fast. What we detected by ESR is just the small "excess" amount of R1 and R2 after all of the dissolved oxygen has been consumed. This explains the sometimes observed short term growth of the ESR signals immediately after mixing of the reactants.
The results presented here clearly demonstrate that the interaction of peroxynitrite and piperazine-type buffers, regardless of the presence of HCO 3 Ϫ /CO 2 , may lead to serious consequences concerning the investigation of peroxynitritedriven actions under physiological conditions, e.g. necrosis, apoptosis, inhibition of enzymes, formation of metabolites, etc. There are several reports in the literature in which similar effects of the interaction of HEPES with strong oxidants other than peroxynitrite are mentioned, although no satisfying explanation has been given. For example, vanadyl induces hemolysis of vitamin E-deficient erythrocytes in HEPES buffer but not in phosphate buffer (55); HEPES stimulates hydroxyl radical generation significantly in the presence of both copper ions and H 2 O 2 (56), and HEPES promotes also hypochlorous acidinduced oxidation of ferrocyanide very efficiently (57). In conclusion, it must be emphasized that O 2 . /H 2 O 2 formation according to the above mechanism (Fig. 12) is not restricted to peroxynitrite as an oxidant and not to piperazine buffer compounds as targets but should be regarded as a general pathway for compounds from which electron-rich alkyl radicals (preferably ␣-amino or ␣,␣-dialkoxyl radicals) can be generated. Accordingly, H 2 O 2 formation in the range of 75 M has been found in the reaction of peroxynitrite (1 mM) with other tertiary amines (400 mM), viz. triethylamine and triethanolamine. 2 Generally speaking, any oxidant strong enough to oxidize certain tertiary amines in a one-electron step would be able to initiate the above reaction sequence. Furthermore, any other 2 M. Kirsch, H. G. Korth, R. Sustmann, and H. de Groot, unpublished observations. FIG. 12. Proposed mechanism of hydrogen peroxide formation from piperazine-based buffer compounds and peroxynitrite or other oxidants. As already shown in Fig. 11 the reaction sequence is initiated by one-electron oxidation of the piperazine compound by peroxynitrite. The ␣-aminoalkyl radical R4 formed by proton loss from the initial radical cation R1 would react at a close to diffusion-controlled rate with oxygen. Although not detected by ESR, radical R1 may likewise undergo deprotonation at the side chains to give exocyclic ␣-amino radicals R4, which would react with O 2 in the same manner. Rapid fragmentation of the so-formed peroxyl radicals R5, R5 produces O 2 . and the cationic species 6.
The latter would be rapidly trapped by reaction with water or other nucleophiles (47) and/or further oxidized to the radical cation R2. reaction, preferably hydrogen abstraction, that generates ␣-amino alkyl or ␣,␣-dialkoxyl radicals would induce the same process. Because tertiary amine groups as targets for strong oxidants are present in a variety of biological molecules, e.g. in ATP, creatine, or nucleic acids, we propose that Fig. 12 for O 2 . /H 2 O 2 production must be considered to be a general mechanism under in vivo conditions. This certainly sheds a new light on a number of studies aimed at the pathophysiological effects of various oxidative species, as actually hydrogen peroxide might have been the true operating agent.