Address reprint requests to Russel J. Reiter, Department of Cellular and Structural Biology, The University of Texas Health Science Center at San Antonio, Mail Code 7762, 7703 Floyd Curl Drive, San Antonio, TX 78229-3900, USA. E-mail: firstname.lastname@example.org
Abstract: Melatonin and classic antioxidants possess the capacity to scavenge ABTSḃ+ with IC50s of 4, 11, 15.5, 15.5, 17 and 21 μm for melatonin, glutathione, vitamin C, trolox, NADH and NADPH, respectively. In terms of scavenging ABTSḃ+, melatonin exhibits a different profile than that of the classic antioxidants. Classic antioxidants scavenge one or less ABTSḃ+, while each melatonin molecule can scavenge more than one ABTSḃ+, probably with a maximum of four. Classic antioxidants do not synergize when combined in terms of scavenging ABTSḃ+. However, a synergistic action is observed when melatonin is combined with any of the classic antioxidants. Cyclic voltammetry indicates that melatonin donates an electron at the potential of 715 mV. The scavenging mechanism of melatonin on ABTSḃ+ may involve multiple-electron donations via intermediates through a stepwise process. Intermediates including the melatoninyl cation radical, the melatoninyl neutral radical and cyclic 3-hydroxymelatonin (cyclic 3-OHM) and N1-acetyl-N2-formyl-5-methoxykynuramine (AFMK) seem to participate in these reactions. More interestingly, the pH of the solution dramatically modifies the ABTSḃ+ scavenging capacity of melatonin while pH changes have no measurable influence on the scavenging activity of classic antioxidants. An acidic pH markedly reduces the ABTSḃ+ scavenging capacity of melatonin while an increased pH promotes the interaction of melatonin and ABTSḃ+. The major melatonin metabolites that develop when melatonin interacts with ABTSḃ+ are cyclic 3-OHM and AFMK. Cyclic 3-OHM is the intermediate between melatonin and AFMK, and cyclic 3-OHM also has the ability to scavenge ABTSḃ+. Melatonin and the metabolites which are generated via the interaction of melatonin with ABTSḃ+, i.e. the melatoninyl cation radical, melatoninyl neutral radical and cyclic 3-OHM, all scavenge ABTSḃ+. This unique cascade action of melatonin, in terms of scavenging, increases its efficiency to neutralized ABTSḃ+; this contrasts with the effects of the classic antioxidants.
Melatonin is a highly evolutionarily conserved molecule that has possibly existed for billions of years . Poeggeler et al.  were the first to identify melatonin in a unicellular organism, a marine alga (Gonyaulax polyedra). The concentrations of melatonin in G. polyedra are very high (relative to values measured in mammalian blood) and they exhibit a cycle with highest levels during the scotophase and lowest levels during the photophase, just as in the mammalian pineal gland. In 1995, Manchester et al.  observed that melatonin also exists in an ancient-type prokaryote, a photosynthetic bacterium, Rhodospirillum rubrum. These observations suggest that melatonin might have been present during the evolution of the most primitive life forms. Melatonin has also been found in plants [4, 5], yeast and other low-rank organisms . Melatonin is endogenously produced in all vertebrates from fish to humans . Thus, it seems obvious that melatonin exists in all organisms and is involved at all evolutionary levels.
In addition to its other well-known physiological functions, melatonin is a potent free radical scavenger and a broad spectrum antioxidant [8, 9]. It has been speculated that a primary function of melatonin may include protecting organisms from intrinsically or environmentally induced oxidative stress. A large number of publications have shown that melatonin detoxifies a variety of oxidants including ḃOH, O2, NOḃ, ONOO−, HOCl, the hemoglobin oxoferryl radical and possibly LOOḃ [10–13], thereby, reducing oxidative damage both under in vitro and in vivo conditions [14–17]. There are, however, rather few studies which relate to the mechanistic examination of the free radical scavenging activity of melatonin. Thus, it is virtually unknown how melatonin scavenges such a variety of free radicals and the nature of the resulting metabolites has been only sparingly investigated. Also, whether the scavenging mechanisms of melatonin differ from those of the classic antioxidants remain to be elucidated.
In this study, the ABTSḃ+ scavenging capacities of melatonin and several classic antioxidants including vitamin C, trolox (a water-soluble vitamin E analog), glutathione, NADH and NADPH were compared. In addition, the metabolites which were generated via the interaction of melatonin and ABTSḃ+ were also analyzed as to their scavenging capacity.
Materials and methods
Melatonin (100% chromatographically pure) was a gift from the Helsinn Chemical Co. (Biasca, Switzerland). All other reagents and thin layer chromatographic (TLC) plates (silica gel on polyester, fluorescent indicator, layer of 250 μm and 20 × 20 cm) were purchased from Sigma (St Louis, MO, USA). Cyclic 3-hydroxymelatonin (C-3-OHM) and N1-acetyl-N2-formyl-5-methoxykynuramine (AFMK) were synthesized in our laboratory.
Cyclic voltammetry (CV) measurements of melatonin
CV was used to test whether melatonin possesses the ability to donate an electron . Highly pure melatonin was dissolved in methanol and diluted with phosphate buffer (100 mm pH 7.4) at a final concentration of 50 mm. One milliliter of the melatonin solution was placed in a test cell into which the three electrodes were introduced: the working electrode (3.3 mm in diameter, glassy carbon), the reference electrode (Ag/AgCl), and the auxiliary electrode (platinum wire). A potential was applied linearly to the working electrode at a constant rate (100 mV/s) either toward the positive potential (evaluation of reducing equivalents) or toward the negative potential (evaluation of oxidizing species). An electrochemical working station CV-50W (Bioanalytical Systems, West Lafayette, IN, USA) was used. During operation of the CV, a potential current curve was recorded (cyclic voltammogram). The applied potential ranged from −0.4 to 1.2 V at a rate of 100 mV/s versus the Ag/AGCl reference electrode. The anodic waves were recorded and analyzed.
ABTSḃ+ generation and determination
The ABTSḃ+ was generated by incubating 2,2′-azino-bis-2-ethybenz-thiazoline-6-sulfonic acid (ABTS) (7 mm) with potassium persulfate (2.45 mm) in purified water . The incubation was carried out at room temperature and the samples were in the dark for at least 18 hr. The stock solution of ABTSḃ+ was serially diluted with potassium phosphate buffer (50 mm, pH 7.4) to the range of 1.25–20 μm and measured spectrophotometrically (DU530 Life Science UV/Vis Spectrophotometer, (Beckman, Fullerton, CA, USA) at a wavelength of 415 nm to establish a standard curve. Temperature was controlled at 30°C using a built-in Peltier Temperature Control Module. The concentrations of ABTSḃ+ were calculated from the molar extinction coefficient of ABTSḃ+ at 415 nm (ε = 3.6 × 104 mol−1.cm−1) .
Assay protocol to determine the ABTSḃ+ scavenging activities of melatonin and other antioxidants
Melatonin, vitamin C, trolox, glutathione, NADH and NADPH, at different concentrations (total volume of 5 μL) were added to 1 mL of 20 μm ABTSḃ+ solution, respectively. After the addition of either melatonin or another antioxidant to the ABTSḃ+ solution, complete mixing of reactants was achieved by bubbling three to four times using plastic pipettes (2–3 s was required to complete this procedure). The optical absorbance of ABTSḃ+ at 415 nm was measured before the addition of these agents and at 1, 4, 8, and 12 min after addition. Each individual treatment was repeated five to six times and the results of the experiments were compared.
To further explore the ABTSḃ+ scavenging mechanisms of melatonin, the synergistic effects of melatonin with other classic antioxidants or between the classic antioxidants themselves were examined. During these experiments, the concentration of ABTSḃ+ was 20 μm, that of melatonin was 2.5 μm and that of other antioxidants was 5 μm. A preliminary study showed that low doses of the tested substances (2.5 μm melatonin and 5.0 μm other antioxidants) were more sensitive when examining the synergistic actions of these agents. In these studies, two series of experiments were carried out. Series 1 tested the overall synergistic effects of melatonin with other antioxidants and between the classic antioxidants themselves. Briefly, for each pair of tested substances one antioxidant was incubated with ABTSḃ+ to examine its ABTSḃ+ scavenging capacity individually. Then the tested pair of substances was co-incubated with ABTSḃ+ to examine whether their ABTSḃ+ scavenging capacity was greater or less than the sum of their individual values. For example, for the melatonin and vitamin C pair, 2.5 μm melatonin or 5 μm vitamin C were incubated with 20 μm ABTSḃ+; this was repeated for melatonin plus each of the other antioxidants. Potential synergistic effects were calculated using the method of Webb .
In series 2, the kinetics of the synergistic effects of melatonin with other antioxidants were examined. In earlier studies, it was found that the ABTSḃ+ scavenging activities of all classic antioxidants occurred almost exclusively during the first minute (the earliest time point tested for the reaction was at 1 min; thus, it is possible that the reaction was completed earlier than 1 min). The interaction of ABTSḃ+ with melatonin, however, continued for hours. Thus, in this series, 2.5 μm melatonin with 5 μm vitamin C, 5 μm trolox, 5 μm glutathione, 5 μm NADH were incubated with 20 μm ABTSḃ+. After 1 min of incubation (as indicated above the classic antioxidants had completed their reaction with ABTSḃ+ within the first minute, and therefore it was assumed that the reaction after one 1-min period was attributable to melatonin), the kinetics of the absorbance changes at 415 nm were measured. The results are expressed as a rate constant of absorbance decay in ABTSḃ+/min. In control experiments, 2.5 μm melatonin was incubated with 20 μm ABTSḃ+ for 1 min and the kinetic measurement was recorded. The results obtained when melatonin was used alone and when melatonin was co-incubated with a second antioxidant were compared to determine whether there were synergistic effects.
Determination of pH effects on melatonin and classic antioxidants in scavenging ABTSḃ+
It was first determined that a pH ranging from 5.0 to 9.3 had no significant effect on the absorbance of ABTSḃ+ at 415 nm. Subsequently, different concentrations of melatonin (from 0.25 to 640 μm), vitamin C, or vitamin E (from 1 to 25 μm) were incubated with 20 μm ABTSḃ+ at a pH 5.0, 7.4 or 9.3, respectively. The absorbance at 415 nm was measured before the addition of antioxidants and at 1, 4, 8 and 12 min after the addition of each agents. Each procedure was repeated four times.
Measurements of the metabolites found during interaction of melatonin with ABTSḃ+
Thin layer chromotography and high-performance liquid chromatography-electrochemical detector (HPLC-ECD) were used to identify the consumption of melatonin as well as the formation of metabolites resulting from the interaction of melatonin with ABTSḃ+.
For the TLC analysis, 5 mL (2 mm) ABTSḃ+ was incubated with 0.1 ml (100 mm) melatonin (in equimolar concentrations) at 30°C, pH 7.4 for 4 hr. The mixtures were extracted with 20 mL chloroform and washed trice with NaCl-saturated saline. Chloroform was evaporated under a vacuum. The dried residues were redissolved in methanol and separated by TLC. Melatonin, purified AFMK or C-3-OHM served as the standards, respectively. The separated metabolites were detected by UV light at a wavelength of 254 nm and photographed using a digital camera. The density of each substance was quantified using a computer program (‘Scion Image for Windows’ NIH Image for PC).
For HPLC analysis, 20 μm melatonin was incubated with 20 μm ABTSḃ+. At 0.5, 1.0, 6.0 and 12.0 min, absorbance was measured at 415 nm, respectively, for identifying the concentrations of ABTSḃ+, and at each time, a 10-μL solution was sampled and immediately added to 490 μL HPLC mobile phase solution (this immediately stopped the reaction of melatonin with ABTSḃ+ because of the highly acidic pH of the mobile phase). Thirty microliters of this mixture was injected into an HPLC-ECD system to simultaneously measure melatonin and AFMK levels.
The ESA HPLC system consisted of an eight-channel CoulArray 5600 detector connected to a column (4.6 × 250 mm, Partisil 5 μm OD53; Waters, Milford, MA, USA). The mobile phase was constituted with pH 4.5 potassium phosphate buffer (100 mm) and acetonitrile (20%). Before use the mobile phase was filtered with 0.22 μm membrane and degassed by sonication and vacuum for 20 min. The prepared mobile phase was then delivered using ESA 580 solvent delivery system with a flow rate of 1 ml/min. Before sample assay, the mobile phase was preoxidized under a potential of 900 mV to decrease the baseline noise using the ESA Gardstat apparatus. Commercially available melatonin and synthesized AFMK served as standards. The applied potentials for melatonin and AFMK were 700 and 900 mV, respectively.
Assay to determine whether cyclic 3-OHM scavenges ABTSḃ+
In this study, we tested whether cyclic 3-OHM also possessed ABTSḃ+ scavenging activity. Cyclic 3-OHM (20 μm or 40 μm) was incubated with 20 μm ABTSḃ+ at 30°C pH 7.4 for 12 min. The absorbance was measured at 415 nm before the addition of C-3-OHM and at 1, 4, 8 and 12 min after its addition. Each treatment was repeated four times.
Data are expressed as mean ± S.E.M. When three or more groups were involved in a study, a one-way-analysis of variance (anova) was used followed by the Student–Newman–Keuls t-test. When only two groups were compared, a Student's t-test was used. A P value <0.05 was considered statistically significant.
The voltammogram generation during CV analysis indicated that in an electric field, melatonin, like the classic antioxidants, generated an anodic wave; for melatonin this wave appeared at Ep(a) 715 mV during the forward scan (Fig. 1). No wave was detected during the reverse scan. The presence of an anodic wave on CV indicates the ability of melatonin to donate an electron and to function as a reductive agent.
The absorbance values of ABTSḃ+ at the wavelength of 415 nm exhibited a strong positive relationship with the concentrations of ABTSḃ+, when the concentrations of ABTSḃ+ were 20 μm or less. The standard curve of the concentrations of ABTSḃ+ and the corresponding absorbance values at a wavelength of 415 nm was linear and the correlation coefficient between them was essentially 1 (data not shown).
The results indicate that melatonin, vitamin C, trolox, glutathione, NADH and NADPH individually scavenge ABTSḃ+ and that the scavenging capacities are concentration-dependent, i.e. increasing antioxidant levels correlated with decreasing ABTSḃ+ concentrations (Fig. 2A). The profiles of the interaction of classic antioxidants with ABTSḃ+ indicate that the majority of reactions of these agents with ABTSḃ+ occurred within <1 min (the earliest test point). After 1 min, the reactions are completed as reflected by the unchanged ABTSḃ+ absorbance at 415 nm from 2 to 12 min (Fig. 2B). The stoichiometric analysis implied that one classic antioxidant scavenges one ABTSḃ+ or less. This was shown by the IC50 values (the concentration required to inhibit 50% of the reaction) for glutathione (11 μm), vitamin C (15.5 μm), trolox (15.5 μm), NADH (17 μm) and NADPH, (21 μm), respectively, in scavenging 20 μm ABTSḃ+. Melatonin, however, exhibited a different profile from the classic antioxidants. The interaction of melatonin with ABTSḃ+ showed two phases, i.e. an initial rapid phase and a phase of progressive reduction (Fig. 2B). The stoichiometric analysis indicated that one melatonin molecule scavenged more than one ABTSḃ+. The IC50 of melatonin to scavenge 20 μm ABTSḃ+ was 4 μm. Furthermore, when the reaction time is sufficiently long (2 hr), 5 μm melatonin almost completely scavenged 20 μm ABTSḃ+ (Fig. 2C). In this case, one melatonin molecule presumably scavenged up to four ABTSḃ+.
Melatonin exhibits a synergistic effect with both vitamin C or trolox in terms of scavenging ABTSḃ+ (Table 1). Conversely, among vitamin C, vitamin E and NADH no synergistic effects were observed in terms of reducing ABTSḃ+ (Table 2).
Table 1. Tests for synergistic effects between melatonin and vitamin C or between melatonin and trolox
In these studies, the concentration of ABTS.+ was 20 μm and the observation period was 12 min. The percentage values represent relative reduction of ABTS·+ for a single antioxidant or a combination of two. Data are mean ± S.E.M. for five determinations; *P < 0.01 values of i1,2 versus (i1 + i2 − i1·i2) as mentioned.
10.2 ± 0.5%
5.8 ± 1.2%
33.0 ± 2.4%
51.7 ± 1.6%*
31.5 ± 3.7%
52.8 ± 2.3%*
Table 2. Tests for synergistic effects between vitamin C and trolox or between vitamin C and NADH
Substances Vitamin C
In these studies, the concentration of ABTS·+ was 20 μm and the observation period was 12 min. The percentage values represent relative reduction of ABTS·+ for a single antioxidant or combinations of them. Data are mean ± S.E.M. for five determinations; the results show that combinations of classic antioxidants do not exhibit synergistic effects.
4.6 + 0.9%
4.5 ± 0.9%
11.2 ± 0.4%
13.4 ± 0.4%
11.6 ± 0.5%
15.9 ± 1.2%
Synergistic effect was determined using the following formula (Webb 1963).
where i1 is the percentage relative reduction of ABTSḃ+ by melatonin, i2 is the percentage relative reduction of ABTSḃ+ by another antioxidant, i1,2 is the percentage relative reduction of ABTSḃ+ with a combination of melatonin and another antioxidant, and i1 ḃ i2 is the product of i1 and i2. Among the classic antioxidants, i1 and i2 represent different antioxidants, respectively. If the i1,2 is > i1 + i2 − i1ḃi2 a synergistic effect is accepted . If the i1,2≤i1 + i2 − i1ḃi2 a synergistic effect does not exist.
Another approach also indicated that melatonin possesses synergistic effects with classic antioxidants in scavenging ABTS.+. This is reflected by the increased ABTS.+ scavenging rate of melatonin during the progressive reduction phase when it was combined with other antioxidants. When ABTS.+ was 20 μm, 5 μm melatonin scavenged these ABTS.+ with a rate constant of 0.38 ± 0.02 μm/min during the progressive reduction phase. The rate constant was increased 40–60% when an additional 5 μm of a second antioxidant was combined with melatonin (Fig. 3). This increase was statistically significant. As mentioned above, the progressive reduction phase of ABTS.+ is a unique property of melatonin but not of the classic antioxidants.
The pH changes did not modify the ABTSḃ+ scavenging abilities of the classic antioxidants, vitamin C and trolox (Fig. 4A and B). The pH changes, however, dramatically modified the ABTSḃ+ scavenging ability of melatonin. When the pH was reduced from a physiological level (7.4) to 5.0, the ABTSḃ+ scavenging capacity of melatonin was reduced roughly 40-fold (Fig. 4C). Conversely, when the pH was increased from 7.4 to 9.3, this accelerated the ABTSḃ+ scavenging rate of melatonin significantly but the scavenging efficacy of melatonin during a 12-min period was not modified significantly (Fig. 5).
HPLC-ECD analysis showed that when 15 μm melatonin was incubated with 20 μm ABTSḃ+ at 30°C and pH 7.4, an equal number of moles of melatonin and ABTSḃ+ were consumed during the initial 30 s. Thereafter until 12 min, melatonin consumption was limited, however, ABTSḃ+ consumption continued until completion (10 μm was further consumed) (Fig. 6).
A TLC analysis revealed that the major metabolites during the interaction of melatonin with ABTSḃ+ were C-3-OHM and AFMK (Fig. 7). When the melatonin concentration is equivalent to that of ABTSḃ+, the production of C-3-OHM is predominant. A density analysis revealed that the yield of C-3-OHM was threefold more than that of AFMK under these conditions (Fig. 7). When the melatonin concentration is much lower than that of ABTSḃ+, the dominant metabolite is AFMK. The results of the HPLC-ECD analysis indicate that when 20 μm melatonin was incubated with 20 μm ABTSḃ+ for 1 min, roughly 5 μm AFMK was produced and thereafter AFMK levels were not changed significantly during the 12-min period (Fig. 8A and B).
A major melatonin metabolite, C-3-OHM, also exhibited the ability to scavenge ABTSḃ+. The scavenging profile of C-3-OHM was similar to that of melatonin and with both a rapid and progressive reduction during a 12-min incubation period. The ABTSḃ+ scavenging activity of C-3-OHM was also concentration-dependent (Fig. 9). The efficacy of C-3-OHM to scavenge ABTSḃ+ appeared to be less than that of melatonin. This is reflected in the higher concentrations of C-3-OHM required to achieve the same effects as with melatonin.
As melatonin scavenges a variety of radicals, herein we conducted a study using ABTSḃ+ to further explore the potential mechanisms of the interaction of melatonin with this reactant. ABTSḃ+ has been successfully used to evaluate the total antioxidant capacity of fluids . It was believed that the interaction of antioxidants with ABTSḃ+ results in blanching of this blue-colored cation radical. The degree of blanching of ABTSḃ+ is determined by the scavenging activities of the antioxidants present. The advantage of using ABTSḃ+ to study mechanisms of antioxidants lies in its long half-life. Once it is formed, ABTSḃ+ can be sustained for days . This feature provides an opportunity to directly examine changes in its concentration rather than testing for the by-products of radical reactions such as with ḃOH which requires a spin trapping product. The original ABTSḃ+ assay was based on the activation of metamyoglobin with H2O2 in the presence of ABTS to produce the radical cation and in the presence or absence of antioxidants. This has been criticized on the basis that the faster reacting antioxidants might also contribute to the reduction of the ferryl myoglobin radical. A more suitable format for this assay is a blanching method in which the radical is generated directly in a stable form prior to reaction with putative antioxidants. An improved technique for generation of ABTSḃ+ was described by Re et al. . The blue/green ABTSḃ+ formation was via the reaction between ABTS and potassium persulfate. In this system, ABTSḃ+ is the only reactive agent and this rules out the influence of other radicals when the ABTSḃ+ scavenging activity of an antioxidant is tested.
Poeggeler et al.  compared the ABTSḃ+ scavenging capacities of melatonin to vitamin C, trolox and glutathione. Their ABTSḃ+ generating system, however, contained both ABTSḃ+ and ḃOH. This made it difficult to deduce the exact mechanism of interaction of melatonin with ABTSḃ+. To the best of our knowledge, this is the first study to compare the ABTSḃ+ scavenging activity of melatonin with that of the classic antioxidants including vitamin C, trolox, glutathione, NADH and NADPH using the system employed here.
Vitamin C and vitamin E are the well-known chain-breaking antioxidants which are exclusively obtained from the diet. They represent a large portion of the total naturally occurring antioxidants present in the diet. Glutathione is one of the most important intracellular antioxidants and is a representative of all the sulfhydryl (-SH) containing antioxidants. NADH and NADPH are also classified as antioxidants due to the fact that they directly scavenge a variety of free radicals  and they can also provide a hydrogen to recycle other antioxidants such as glutathione.
Our results show that both melatonin and the classic antioxidants effectively scavenge ABTSḃ+. Stoichiometric analysis revealed that for all classic antioxidants, one molecule scavenges one or less ABTSḃ+ as reflected by the IC50 of vitamin C (15.5 μm), trolox (15.5 μm), glutathione (11 μm), NADH (17 μm) and NADPH (21 μm), respectively, in a solution containing 20 μm ABTSḃ+. However, one molecule of melatonin scavenges more than one ABTSḃ+. The IC50 for melatonin when scavenging ABTSḃ+ is 4 μm. Under conditions where 5 μm melatonin was incubated with 20 μm ABTSḃ+, given sufficient time, the 5 μm melatonin completely scavenged 20 μm ABTSḃ+ (Fig. 2C). This suggests that one melatonin molecule might scavenge up to four ABTSḃ+.
Another obvious difference between melatonin and the classic antioxidants is their ABTSḃ+ scavenging profiles. For all classic antioxidants, the scavenging activity on ABTSḃ+ occurred within the first minute. Thereafter, there is no further scavenging action as indicated by the stable ABTSḃ+ levels during the 2–12-min observation period. This finding is consistent with the results of Re et al. . For melatonin, the ABTSḃ+ scavenging activity occurred progressively during the 12-min observation period. The interaction of melatonin with ABTSḃ+ exhibited two apparent reaction phases (an initial rapid and a progressive reduction phase). During the initial phase (within the first minute) the reaction properties of melatonin are similar to those of the classic antioxidants. In contrast to the classic antioxidants, melatonin progressively scavenges ABTSḃ+ during the subsequent 11 min with a relatively slower rate than that during the first minute (the phase of progressive reduction, Fig. 2B).
Due to the different efficacies and scavenging profiles of melatonin compared with the classic antioxidants, it is speculated that the classic antioxidants tested share a similar mechanism for scavenging ABTSḃ+; this is probably via hydrogen donation. Melatonin, however, may have multiple electron donation reactions when scavenging ABTSḃ+, probably mediated by different intermediates including the melatoninyl cation radical in the initial step. This hypothesis is supported by the recent publication of Tesoriere et al. . They observed a stepwise two-electron donation process for melatonin in scavenging the hemoglobin oxoferryl radical.
To further examine whether melatonin and classic antioxidants exhibit different mechanisms in scavenging ABTSḃ+, a synergistic study including melatonin in combination with another antioxidant was performed. This additive process occurs when two classic antioxidants are incubated with ABTSḃ+ (Table 1). However, when melatonin was combined with either vitamin C or trolox and incubated with ABTSḃ+, the synergistic effect was obvious (Table 2) when evaluated using the method of Webb .
Interestingly, the synergistic effect primarily occurred during the phase of progressive reduction. This phase is believed to be exclusively a result of melatonin because scavenging by classic antioxidants lacks this phase. The ABTSḃ+ scavenging rate of melatonin in the phase of progressive reduction was increased roughly by 50% following the addition of classic antioxidants including vitamin C, trolox, glutathione or NADH (Fig. 3).
An explanation of the synergistic effect of melatonin with classic antioxidants is that the classic antioxidants may recycle melatonin intermediates such as the melatonin cation or neutral radicals back to melatonin. As the melatonin molecule seems to scavenging more ABTSḃ+ than does a classic antioxidant, leaving out a classic antioxidant to recycle a melatonin molecule would be a highly efficient means for scavenging ABTSḃ+ and would result in the synergistic effect that was noted.
Melatonin recycling has been proposed by Mahal et al.  and Olsen et al. . Based on their experimental observations, both groups speculated that vitamin C or NADH may recycle melatonin from the melatonin cation radical and the mechanism might involve electron transport.
Collectively, the data indicate that melatonin possesses a mechanism to scavenge ABTSḃ+ which appears different from that of the classic antioxidants. The initial step seems to involve one electron donation to form the melatoninyl cation radical. To test this hypothesis, solutions with a range of pHs were examined in terms of the ABTSḃ+ scavenging ability of melatonin as well as the classic antioxidants. We hypothesized that pH changes would not modify the ABTSḃ+ scavenging ability of classic antioxidants as their scavenging activities do not involve one electron donation and cation radical formation. The results indeed showed that the pH change did not modify the ABTSḃ+ scavenging capacity of the classic antioxidants (Fig. 4A and B).
As for melatonin, it was predicted that a changing pH would influence its ABTSḃ+ scavenging capacity. It is documented that the melatoninyl cation radical and melatoninyl neutral radical are in equilibrium in the aqueous phase . When the pH decreases, the equilibrium is disturbed in favor of the melatoninyl cation radical. When the pH increases more melatoninyl neutral radicals are generated. The pKa of indole and substituted indolyl radicals is typically between 4.5 and 6.5 . In all experiments, to mimic a physiological situation, the selected pH was 7.4. At this pH, the melatoninyl cation radical generated by scavenging ABTSḃ+ in the initial step would be transformed into the melatoninyl neutral radical. As a result, the melatoninyl neutral radical would react with another ABTSḃ+ and neutralize it more easily than would the melatonin cation radical. This could explain, in part, how melatonin scavenges more than one ABTSḃ+ at pH 7.4. Conversely, under acidic conditions, the melatoninyl cation radical would dominate. Because both the ABTSḃ+ and melatoninyl cation radical bear positive charges, the electrical repelling force would prohibit or slow their interaction. Thus, decreasing the pH would dramatically reduce the ABTSḃ+ scavenging ability of melatonin.
The experimental results were as anticipated. When the pH was reduced from 7.4 to 5.0, the ABTSḃ+ scavenging capacity of melatonin decreased by roughly 40-fold (Fig. 4C). When the pH was increased from 7.4 to 9.0, the efficiency of melatonin to scavenge ABTSḃ+ was not significantly altered. The reaction rate of melatonin with ABTSḃ+, however, was increased significantly at the higher pH (Fig. 5). This was expected considering that a rise in pH would increase the production of melatoninyl neutral radical and promote the interaction of melatonin with ABTSḃ+. Regardless of how fast melatonin interacts with ABTSḃ+ under high pH conditions, however, one melatonin molecule has a fixed capacity to scavenge a fixed number of ABTSḃ+. These pH studies indirectly support the speculation that the initial step of melatonin to scavenge ABTSḃ+ involves electron donation and the formation of the melatoninyl cation radical.
Obviously, identifying metabolites that result as a consequence of the interaction of melatonin and ABTSḃ+ would help to further clarify the mechanisms involved. The metabolite analysis using a TLC method showed that the interaction of melatonin and ABTSḃ+ produced two major metabolites, i.e. cyclic 3-OHM which is also the metabolite resulting from the interaction of melatonin with hydroxyl radical  and AFMK (Fig. 7). When the molecular ratio of melatonin to ABTSḃ+ was 1:1, the amount of cyclic 3-OHM was threefold in excess of that of AFMK. At the ratio of 0.5:1 (melatonin:ABTSḃ+), the majority metabolite is AFMK. These results are similar to those of Tesoriere et al.  using a hemoglobin oxoferryl radical and melatonin reaction system. They also observed that excess oxidants convert cyclic 3-OHM to AFMK.
Considering this, an HPLC-ECD method was developed to detect AFMK and melatonin simultaneously. This method, combined with a spectrophotometric measurement of ABTSḃ+, allowed us to examine the relationships of the consumption of melatonin and ABTSḃ+ as well as AFMK generation. The results showed that during the first minute, melatonin and ABTSḃ+ were consumed roughly in equal molar concentrations. Thereafter, however, melatonin consumption was limited but the ABTSḃ+ continued to be consumed (Fig. 6). This observation supports the idea that melatonin and its metabolites donate more than one electron (the multi-electron donation hypothesis). It is speculated that during the initial phase, melatonin per se takes part in the reaction by donating an electron to form the melatoninyl cation radical and to neutralize an ABTSḃ+. This is reflected by the equimolar consumption of melatonin and ABTSḃ+. Thereafter, ABTSḃ+ is scavenged by metabolic intermediates of melatonin. This is consistent with the limited melatonin consumption and the continuing ABTSḃ+ reduction during the progressive phase of the reaction. This feature was even obvious when 5 μm melatonin reacted with 20 μm ABTSḃ+. After the first minute of incubation, there was no detectable melatonin as measured by HPLC-ECD; yet the levels of ABTSḃ+ continued to decrease until it had almost completely disappeared. Interestingly, AFMK generation was not changed significantly between the initial phase and the phase of progressive reduction when 20 μm melatonin was co-incubated with 20 μm ABTSḃ+. This was not surprising because the TLC analysis indicated that, at this 1:1 concentration ratio, the major metabolite is cyclic 3-OHM. As predicted, when the pH was decreased to 5.0, melatonin consumption was limited and AFMK formation was diminished. AFMK formation in the ABTSḃ+ system seems to be strongly modified by changes in pH or the existence of transient melatoninyl neutral radicals.
As melatonin metabolite analysis revealed that cyclic 3-OHM was one of the intermediates between melatonin and its final metabolite, AFMK, we investigated whether cyclic 3-OHM participates in ABTSḃ+ scavenging during the progressive reduction phase. When cyclic 3-OHM was incubated with ABTSḃ+ the scavenging activity of cyclic 3-OHM was reflected in the persistent decrease in ABTSḃ+ levels; this scavenging activity was concentration-dependent (Fig. 9). The ABTSḃ+ scavenging profile of cyclic 3-OHM is similar to the profile of the progressive reduction phase of melatonin. The findings show that cyclic 3-OHM indeed participates in the ABTSḃ+ scavenging action during the phase of progressive reduction. This explains how a single melatonin molecule scavenges more than one ABTSḃ+.
Based on the experimental observations described above, we propose the scheme summarized in Fig. 10 to explain the interactions of melatonin and ABTSḃ+. According to this scheme, melatonin donates an electron to scavenge an ABTSḃ+ and forms the melatoninyl cation radical. At physiological pH, the melatoninyl cation radical transforms into a melatoninyl neutral radical and then interacts with a second ABTSḃ+ to form cyclic 3-OHM. If all ABTSḃ+ have been scavenged, the C-3-OHM would be the final melatonin metabolite. In the presence of additional ABTSḃ+s however, cyclic 3-OHM would scavenge two ABTSḃ+ and form AFMK; this would require more than one reaction step. Thus, one melatonin molecule would scavenge up to four ABTSḃ+s via several stepwise reactions. Clearly, this pathway is an assumption based on the experimental observations and currently available data and requires further experimental confirmation. This proposed pathway of the interaction of melatonin with ABTSḃ+ stoichometrically fits well with the experimental results presented in Fig. 10.
The fact that melatonin as well as its metabolites function as free radical scavengers  is defined as a radical scavenging cascade . This cascade reaction not only increases the efficiency of melatonin to scavenge free radicals but also limits prooxidative actions which are a common feature of most classic antioxidants.
DX Tan was supported by NIH training grant T32AG00165-13; JC Mayo has a post-doctoral training research fellowship from FICYT; RM Sainz has a Fullbright Grant financial sponsorship of the Spanish Ministry of Education, Culture & Sports; S Lopez-Burillo was supported by a grant from the Spanish Ministry of Education, Culture & Sports.