Peroxidatic oxidation of catecholamines. A kinetic electron spin resonance investigation using the spin stabilization approach.

Using spin stabilization, ESR measurements have been made of o-semiquinone production from the horseradish peroxidase-H2O2 oxidation of catecholamine substrates. The termination rate constant for semiquinones stabilized with Zn2+ at pH 5 is about 10(4) times smaller than for uncomplexed semiquinones at neutral pH. Stabilization allows steady state concentrations of semiquinones to be obtained. The duration of the steady state is dependent upon the concentrations of enzyme, hydrogen peroxide, and catecholamine substrate. The relative reactivity of the substrates 3,4-dihydroxyphenylalanine, norepinephrine, and dopamine at pH 5 is 1:8:40. The effects of phenol and ascorbate were studied and shown to be consistent with scavenging of phenoxyl radicals by catecholamine and semiquinone radicals by ascorbate, respectively.

Using spin stabilization, ESR measurements have been made of o-semiquinone production from the horseradish peroxidase-H202 oxidation of catecholamine substrates. The termination rate constant for semiquinones stabilized with Zn2+ at pH 5 is about lo4 times smaller than for uncomplexed semiquinones at neutral pH. Stabilization allows steady state concentrations of semiquinones to be obtained. The duration of the steady state is dependent upon the concentrations of enzyme, hydrogen peroxide, and catecholamine substrate. The relative reactivity of the substrates 3,4-dihydroxyphenylalanine, norepinephrine, and dopamine at pH 5 is 1:8:40. The effects of phenol and ascorbate were studied and shown to be consistent with scavenging of phenoxy1 radicals by catecholamine and semiquinone radicals by ascorbate, respectively.
o-Semiquinone radicals are intermediates during the 1electron oxidation of cathechols and catecholamines. For example, autoxidation and enzymatic oxidation of these materials to give free radicals are well known (1,2). The toxicity of catechols and catecholamines generally is felt to be related to production during oxidation of semiquinones and toxic molecular products such as o-quinones (3-6). Except at high pH, semiquinone radicals are transient (7), decaying rapidly via disproportionation to give the catechol and o-quinone. Thus in a system where rates of radical production are low, steady state radical concentrations can be below the level for detection by ESR (-0.1 PM). For this reason, most ESR studies of semiquinone radical production in enzymatic systems (8) have utilized high enzyme concentrations and a rapid flow system in order to have high rates of radical production and to minimize the associated rapid depletion of starting materials.
We previously showed (9) that steady state radical concentrations can be greatly enhanced in the presence of complexing metal ions, a procedure that has been termed spin stabilization (10, 11). In a preliminary communication ( l l ) , we demonstrated detection of semiquinones during the peroxidatic oxidation of dopa' and related catecholamines in static systems using Zn2+ as a stabilizing metal ion. The levels of Zn2+ ions used had no major effect on enzyme activity. Stabilized radicals also have been identified from the oxidation of epinephrine and its analogs (12). We now report kinetic data for the peroxidase-catalyzed oxidation of three catecholamine substrates: dopa, norepinephrine, and dopamine. Effects of phenol and ascorbate on the oxidation of norepinephrine also are reported.

MATERIALS AND METHODS
Catecholamines, phenol, and ascorbate were obtained either from Sigma or Aldrich and were used as received. Solutions containing 2-27 mM catecholamine were made up in acetate buffer, prepared by the addition of zinc acetate (Fisher) to 0.2 M acetic acid, followed by dropwise addition of either 1 M NaOH or glacial acetic acid to bring the buffer to the desired pH.
Enzymatic oxidation of catecholamines was with horseradish peroxidase and H202 as previously described (11). Semiquinone radicals were detected in steady state concentrations in a static system. The concentration of horseradish peroxidase in stock solutions was calculated using a molar absorptivity coefficient of 2.02 X IO6 M" cm" at 403 nm (13). Incubations for ESR studies were made anaerobic by purging with nitrogen gas. Photooxidation of catecholamines was by UV irradiation with an Eimac VIX3OOUV 300-watt xenon arc as previously described (9). For kinetic studies on uncomplexed radicals the solutions were slowly flowed through the cavity to prevent depletion of starting material. For complexed radicals flow was not used in kinetic studies since the radical lifetime is relatively long (longer than the residence time in the cavity under our usual flow conditions). In this case a very brief exposure to light was given to minimize depletion of the catecholamine.
Electron spin resonance measurements were made at ambient temperature on solutions contained in a quartz aqueous flat cell using a Varian E-109 spectrometer operating at X band (9.5 GHz) and employing 100-kHz modulation. For kinetic measurements the light was turned on and off using a magnetic shutter assembly. Signal averaging was accomplished using a Tracor Northern NS-570A signal analyzer. Radical concentrations were estimated by double integration using the nitroxide 3-carbamoyl-2,2,5,5-tetramethyl-3-pyrroline-1-yloxy (Aldrich) as a standard. Second order rate constants were reproducible to +20%.

Kinetic Stabilization of o-Semiquinones-o-Semiquinones
are transient in neutral and acid solution. At low pH, where the semiquinone is present as the neutral species, the rate constant for second order recombination ( 2 k ) is typically -10' nates (pK, values reported for o-semiquinones range from 3.6 to 5.2 (7,14,15)), a rate constant of -IO8 M-' s" has been reported for the o-semiquinone from epinephrine (14). (Rate constants reported for p-semiquinones (15) are quite similar.) This rate constant may reflect cross-reaction between neutral and anionic semiquinones, similar to that found for 4-t-butylo-semiquinone (7), rather than a rate constant for the reaction between two anions.
Semiquinones from the catecholamines we have studied also are transient in neutral aqueous solution. UV photolysis of neutral solutions of dopa gave the steady state ESR spec-trum shown in Fig. la which is typical of dopa semiquinone (9). Time-resolved experiments (Fig. l b ) show that the radical is transient, decaying with second order kinetics (Fig. 2) under these conditions. Spin polarization phenomena (16), which are common in photolysis experiments, also are apparent immediately after turning the light on and off. The radical decay corresponds to a second order rate constant of 2. 10' M-' s-l, close to that reported for the semiquinone from epinephrine (14).
The results obtained are markedly different when complexing metal ions are included in the reaction mixture. Using Zn2+ ions at pH 5.0, the only radical detected (Fig. IC) is the semiquinone complexed with Zn2+ (9). (The ESR parameters of the complex are modified from those of the uncomplexed species because of the redistribution of spin density that occurs in the presence of the metal ion (17).) The complexed radicals are much less transient than the uncomplexed ones. Decay (Fig. Id) remains second order (Fig. 2), but the radical lifetime is now several seconds for a steady state concentration of -lo-' M. The calculated second order rate constant is 1.1 X lo4 M-' s-I for these conditions. A similar value (1.5 X lo4 M" s-') was obtained for o-benzosemiquinone (from the UV photolysis of catechol) complexed with Zn".
Complexation is thus very effective at kinetically stabilizing o-semiquinones. To achieve the steady state radical concentration obtained at pH 7 in the absence of metal ions a radical formation rate that is lo4 times lower is all that is required if the radicals are complexed. Thus with spin stabilization one can generate high radical concentrations but have a slow rate of substrate depletion. This allows the use of static, rather than flow systems.
Enzymatic Oxidation of Catecholamines-Enzymatic oxidation of catecholamines using horseradish peroxidase/H202 gives, in the presence of Zn2+ ions, high concentrations of spin-stabilized semiquinones. These radicals can be obtained in steady state concentration for prolonged periods of time. Radical concentrations of -1-20 PM were typically obtained. For the reaction rates employed, these concentrations were consistent with rate constants for radical termination of the order of those determined in the photolysis experiments. We examined the effects on the steady state of varying concentrations of enzyme, Hz02, and substrate, [Enzyme] Dependence-For peroxidatic oxidation occurring according to the general mechanism (18),  Fig. 1, b and d, respectively. The linearity of the plots is indicative of decay by second order kinetics. The gradient of the plots is the second order rate constant (212) for radical recombination.
the rate of production of RT can be expressed in the following Under steady state conditions so that [RTIgS cc [El" is predicted.
We tested this for several substrates, including norepinephrine (Fig. 3). It can be seen that increasing the enzyme concentration by a factor of 4 increases the steady state radical concentration by a factor of 2, implying that the spinstabilized radicals decay with second order kinetics, as was found in the photolytic system. A square root dependence on enzyme concentration also was obtained for dopa and dopamine (data not shown).
[H20J Dependence-As indicated above, steady state con- centrations of free radicals could be obtained for prolonged periods of time. We found that the duration of the steady state was strongly dependent on the hydrogen peroxide concentration, although the steady state concentration was independent of the concentration of this reagent over the range studied. These findings imply that kl[H202] >> k3[RH2] (see Equation 6) and that [RH2] (the reagent in excess) is not decreased appreciably during the depletion of [H20a] (the limiting reagent).
Zero order dependence on [H202] has been demonstrated in several systems. For example, steady state concentrations of spin-stabilized semiquinones from dopa were found to be independent of hydrogen peroxide concentration over a wide range (1 1). A significant decrease was observed only when the hydrogen peroxide concentration was reduced below -10 p~. Thus, with HzOz limiting and constant substrate concentration the duration of the steady state was linearly dependent on the initial concentration of hydrogen peroxide, [Hz02]o (Fig. 4); doubling the concentration of hydrogen peroxide exactly doubled the duration of the steady state, t., (Fig. 4a), for each of the substrates studied. This allowed an estimate of the rate of removal of hydrogen peroxide in each system, since -d[H202]/dt = u = [HzO2],Jtss. For the same substrate concentration, the rate of peroxide removal was greatest for dopamine and was lowest for dopa (Fig. 4b).
[Substrate] Dependence-A first order dependence of the rate of H20, disappearance on substrate concentration was found for the three catecholamines investigated (Fig. 5). Clearly the enzyme is not saturated with substrate under the conditions employed. With the reaction rate limited by the substrate concentration and k, > > ks the overall reaction rate is given by u = k3[E][RH2]. Values of k3 calculated on this basis are given in Table I. Effects of Phenol and Ascorbate-Phenol and ascorbate are examples of the oxidogenic and redogenic substances described by Ohnishi et al. (20). Their effects are to, respectively, increase and decrease the rate of oxidation of other peroxidase substrates. It has been proposed (20) that these effects are a result of free radical reactions. Thus, phenol, a good peroxidase substrate (21), is suggested to be oxidized to the phenoxyl radical which subsequently oxidizes the other substrate (e.g. as in Reactions 7 and 8). Ascorbate, itself a poor peroxidase substrate (21), is suggested to reduce the substrate radical back to the parent compound (e,g. Reactions 9 and 10). However, few examples of such reactions are known. The phenoxyl radical is known to react rapidly with ascorbate (22), but reaction with catecholamines has not been reported.   Reaction between o-semiquinone and ascorbate was reported to be too slow to measure by the pulse radiolysis technique (22), although there is ESR evidence for the reduction by ascorbate of a semiquinone radical from 6-hydroxydopa (23). (7) PhO' + RH, * PhOH + R' + H+. (8) PhOH HRP/HzOz PhO.,
Using spin stabilization the effects of phenol and ascorbate on the concentrations of catecholamine semiquinones are readily studied. Low concentrations of phenol markedly increase the steady state semiquinone radical concentration and decrease the duration of the steady state (Table 11), i.e. the rates of radical production and peroxide removal are both increased. For 1.75 mM phenol the increase is by a factor of 13 in each case. Assuming that the above mechanism holds, i.e. that the increased rate of removal of HzO, results from oxidation of phenol, then a Lineweaver-Burk plot for the oxidation of phenol can be constructed (Fig. 6).   . a, b, and c, effect of ascorbate on the level of semiquinone radical detected during horseradish peroxidase-HZOz oxidation of norepinephrine. Ordinate is ESR signal amplitude (proportional to free radical concentration). Abscissa is reaction time. The time at which enzyme was added is indicated by the arrows. Individual measurements of At were reproducible to +.lo%. Conditions: 6 mM norepinephrine, 28 nM horseradish peroxidase, 0.28 mM HzOZ, 227 mM Zn2+ in acetic acid-acetate buffer, pH 5.0. The lag in semiquinone detection that is observed is proportional to the initial concentration of ascorbate (d).
Low concentrations of ascorbate resulted in a lag time for semiquinone detection (Fig. 7, a-c) that was proportional to the concentration of ascorbate added (Fig. 7d). Assuming that the lag time corresponds to removal of ascorbate (hydrogen peroxide and catecholamine substrate are in excess), the rate of ascorbate removal can be calculated. The rate determined consistent with quantitative oxidation of ascorbate by HzOz.
We verified in photolysis experiments that reactions of phenoxyl radicals with catechols, and of semiquinones with ascorbate, indeed occur. With low concentrations of catechol (a model for the catecholamines) a small steady state concentration of the semiquinone (a(2H) = 3.92, 0.48 G; g = 2.0039 (17, 25)) is detected during UV photolysis (Fig. 8a). The radical is formed from catechol by a mixture of photoionization and photohomolysis, as was also found for dopa (26). In the presence of excess phenol, where some of the UV light is absorbed by the phenol so that phenoxyl radicals are formed (27), the steady state concentration of semiquinones is greatly increased (Fig. 86), implying that the phenoxyl radical oxidizes catechol to the semiquinone. Experiments with ascorbate show an effect in the opposite direction. Small amounts of ascorbate (insufficient to effectively compete with catechol for the incident light) greatly reduce steady state levels (Fig.   8, c and d) of the semiquinone, and a weak ESR spectrum of the ascorbyl radical (a(2H) = 1.76 G, g = 2.0052 (28)) ( Fig.  8d, arrows) can be detected. DISCUSSION We have shown that the major effect of metal ions on semiquinones is to kinetically stabilize the free radical. Spin stabilization provides a very effective means for detecting very low rates of semiquinone radical formation. Assuming a termination rate constant of -IO4 M" s" and a radical detection limit of M, radicals can be detected by ESR at formation rates as low as 10"' M s-'.
In the present study we have used Zn2+ as the stabilizing metal ion in weakly acid solution (pH 5). However, reactions at other pH values are easily studied with an appropriate choice of metal ion and buffer system. For example, Mg+ in tris buffer is effective at pH 7.5. ' In the peroxidase system, where Zn2+ does not markedly B. Kalyanaraman, R. C. Sealy, and K. Sivarajah, manuscript in preparation.
affect the enzyme activity (ll), reaction kinetics can easily be determined by ESR in a static system. The use of rapid flow systems with large volumes of reagents and high enzyme concentrations is thus not necessary if spin stabilization is employed.
The kinetic data that we have obtained at pH 5 may be compared with those for the same substrates at pH 7-8 based on observations of product formation (29) and for smaller molecules such as catechol (21) (Table I). A pH dependence of the rate constants for reaction of Compound I1 with substrates was expected based on earlier kinetic data; enzyme groups with pK, -5.7 and 8. 6 have been implicated (30). Table I shows that in general the rate constant decreases (i) with increasing substrate bulk and (ii) with decreasing pH. Thus, if we compare dopa with catechol, the rate constant for dopa is about 400 times smaller than catechol at around neutral pH. At pH 5 the rate constant is about 10 times smaller than at pH 8. A similar trend is apparent from data for tyrosine (31, 32) and simple phenols (21), where tyrosine is -1000 times less reactive than simple phenols at neutral pH and is about 5 times less reactive at pH 5 than at pH 8. The rate constant for reaction with dopa at pH 5 (4.5 X 10' M" s-') is very close to that for reaction with tyrosine at this pH (4 x 10' M" s-'), as might be expected given their structural similarities.
Dopamine and norepinephrine are considerably better substrates for horseradish peroxidase than is dopa, both at pH 8 and at pH 5. Whereas norepinephrine, like dopa and tyrosine, shows a large effect of pH, this is not so for dopamine, where the reactivity at pH 5 is close to that measured at pH 8.
For more reactive substrates such as phenol and o-dianisidine (33), where it is evident that saturation occurs, a modification of the general mechanism expressed in Equations 1-4 is necessary, since the existence of an enzyme-substrate complex is implied. The existence of such a complex has been demonstrated spectrophotometrically in the case of p-cresol (30).