Antioxidative Reactivity of L-Ascorbic Acid and D-Isoascorbic Acid Species towards Reduction of Hexachloroiridate (IV)

,e pair [IrCl6]/[IrCl6] has been demonstrated to be a good redox probe in biological systems while L-ascorbic acid (AA) is one of the most important antioxidants. D-isoascorbic acid (IAA) is an epimer of AA and is widely used as an antioxidant in various foods, beverages, meat, and fisher products. Reductions of [IrCl6] by AA and IAA have been analyzed kinetically and mechanistically in this work. ,e reductions strictly follow overall second-order kinetics and the observed second-order rate constants were collected in the pH region of 0≤ pH≤ 2.33 at 25.0°C. Spectrophotometric titration experiments revealed a welldefined 1 : 2 stoichiometry, namely Δ[AA] :Δ[Ir(IV)] or Δ[IAA] :Δ[Ir(IV)]� 1 : 2, indicating that L-dehydroascorbic acid (DHA) and D-dehydroisoascorbic acid (DHIA) were the oxidation products of AA and IAA, respectively. A reaction mechanism is suggested involving parallel reactions of [IrCl6] with three protolysis species of AA/IAA (fully protonated, monoanionic, and dianionic forms) as the rate-determining steps and formation of ascorbic/isoascorbic and ascorbate/isoascorbate radicals; in each of the steps, [IrCl6] acquires an electron via an outer-sphere electron transfer mode. Rate constants of the rate-determining steps have been derived or estimated.,e fully protonated forms of AA and IAA display virtually identical reactivity whereas ascorbate and isoascorbate monoanions have a significant reactivity difference. ,e ascorbate and isoascorbate dianions are extremely reactive and their reactions with [IrCl6] proceed with the diffusion-controlled rate. ,e species versus pH and the species reactivity versus pH distribution diagrams were constructed endowing that the ascorbate/isoascorbate monoanionic form dominated the total reactivity at physiological pH. In addition, the value of pKa1 � 3.74± 0.05 for IAA at 25.0°C and 1.0M ionic strength was determined in this work.


Introduction
Vitamin C is a general name for all the compounds which possess the fully or partially biological activity of L-ascorbic acid (AA) [1][2][3][4][5][6]. It includes AA itself, L-dehydroascorbic acid (DHA), the esters of AA such as ascorbyl palmitate, and D-isoascorbic acid (IAA); the latter two are the synthetic compounds [1][2][3]. Vitamin C is widely found in plants (vegetables and fruits in particular) and in animals, and nowadays it is also available as caplets, tablets, capsules, and drink mixes and in multivitamin and antioxidant formulations [1][2][3]. Biologically and/or medically, extensive investigations of vitamin C have been carried out [7,8]. When vitamin C is utilized in food and nutrients, its antioxidative property is often considered to be an important factor [9,10]. IAA is an epimer of AA and widely used as an antioxidant in various foods, beverages, meat, and fisher products [2,[11][12][13][14][15]. In this regard, several analytical methods have been developed to determine content of IAA in foodstuffs, together with those of AA and their oxidized forms [2,[11][12][13][14][15]. Scheme 1 provides the structures of AA, IAA and their oxidized forms.
As one of the most important antioxidants, the kinetics and mechanisms of vitamin C/ascorbic acid reactions with various oxidants have been explored extensively [9,10,[16][17][18][19][20]; the oxidants include reactive oxygen species such as HO 2 /O 2 radicals and peroxynitrite, and some metal ions/complexes [16][17][18][19]. Among the metal ions/complexes, single electron oxidants apparently dominate the exploration [9,10]. We have had an interest in the study of the redox pair [IrCl 6 ] 2-/[IrCl 6 ] 3due to its salient characters: (i) [IrCl 6 ] 2has two strong absorption bands (around 420 and 488 nm, vide infra) and [IrCl 6 ] 3does not possess these two bands, conferring a convenient monitoring of the pair. (ii) [IrCl 6 ] 2is very stable and meta-stable from pH 0 to 10, enabling us to measure the rate constants in such a wide pH range [20][21][22]. (iii) It has been used as a redox probe to discriminate DNA monolayers and to obtain chemical information of oxidative stress in biological matrixes [23,24]. e antioxidative reactivity of AA and IAA towards reduction of [IrCl 6 ] 2has been investigated in this work. Previously, the kinetics of oxidation of AA by [IrCl 6 ] 2was studied [25,26], but only was carried out in a very narrow acidity region (namely, 0.20 M ≤ [HClO 4 ] ≤ 1.00 M). On the other hand, the redox reaction between [IrCl 6 ] 2and IAA has never been investigated. In fact, the oxidation kinetic and mechanistic aspects of the IAA oxidations were poorly understood due to a very limited study [19]. e purposes of the present work were to study the redox reactions in a pH range as wide as possible, to derive the rate constants of the rate-determining steps in a good accuracy, to propose a convincing reaction mechanism, and to find out the antioxidative difference between AA and IAA.  [27]. For buffer solutions in the pH range from 2.28 to 6.26, the combinations of the buffering pairs H 3 PO 4 /NaH 2 PO 4 , AcOH/ NaOAc, and NaH 2 PO 4 /Na 2 HPO 4 (0.15-0.2 M) were employed. All the buffers were tuned to an ionic strength of 1.0 M by use of NaClO 4 ·H 2 O; their pH values were measured by an Accumet Basic AB150 Plus pH meter, equipped with an Accumet accuTupH ® combination pH electrode ( ermoFisher). e electrode was calibrated using standard buffers of pH 4.00, 7.00, and 10.00 immediately before pH measurements. ese buffer solutions were only used weekly.

Kinetic Experiments.
Stock solutions of about 10 mM [IrCl 6 ] 2were prepared daily by dissolving the desired amount of Na 2 IrCl 6 ·6H 2 O in a solution containing 0.90 M NaClO 4 , 0.09 M NaCl, and 0.01 M HCl; these solutions were only used daily. Stock solutions of AA/IAA were prepared by adding the desired amount of AA/IAA to a buffer solution of specific pH. Each of the solutions was bubbled with pure nitrogen for 5 min and was only employed for ca.

Spectral Insight into the Reaction Courses.
Although an outer-sphere electron transfer was suggested for the oxidation of AA by [IrCl 6 ] 2-, the reaction course was not exploited previously [25,26]. e rapid scan spectra were thus recorded for the oxidations of AA/IAA by [IrCl 6 ] 2in this work; Figure 1 displays such spectra in the case of AA. Clearly the absorption bands around 488, 420, and 306 nm, which are the typical ones of [IrCl 6 ] 2-, remained upshifted; in addition, no new absorption bands emerged during the reaction course. Moreover, the kinetic traces recorded at these bands could be well described by single exponentials, giving rise to the values of pseudo-first-order constants k obsd . e above reaction characters indicate that a simple electron transfer took place without complications such as an adduct formation or faster substitution reactions on [IrCl 6 ] 2before the rate-determining step(s). It is thus concluded that the redox reaction is indeed first-order in [Ir(IV)]. e rapid scan spectra in the case of IAA are very similar to those shown in Figure 1.

Second-Order Kinetics.
Still under pseudo-first-order conditions, the reaction rates were determined when [AA] tot /[IAA] tot was varied in the region 0.10 ≤ [AA] tot / [IAA] tot ≤ 2.0 mM; for this concentration region, the reaction media had enough buffering capacities to ascertain that no pH changes could be caused by the variations of the concentrations of the reductants. e values of k obsd as functions of [AA] tot or [IAA] tot and of pH were collected; for each concentration, the k obsd value is reported as an average of the at least 5 duplicate runs and the standard deviations were usually less than 5%. Plots of k obsd versus [AA] tot are displayed in Figure 2 and versus [IAA] tot are shown in Figure 3; undoubtedly the linearity of the plots is very good, and no significant intercepts are observable. Hence, the redox reactions are first-order in [AA] tot or [IAA] tot , rendering overall second-order kinetics as described by equations (1a) and (1b), where k′ stands for the observed secondorder rate constants.
Values of k′ were calculated from the plots in Figures 2 and 3 and are summarized in Table 1. In addition, plots of logk′ versus pH are shown in Figure 4 (data points). e stopped-flow spectrometer was employed essentially to its up limit for following the redox reactions.

e Reaction Stoichiometry.
e stoichiometric ratios were determined in a medium of 0.10 M HClO 4 and 0.90 M NaClO 4 by use of spectrophotometric titration method which was demonstrated to be a very useful approach [28][29][30][31] is is perhaps not surprising because commonly AA is oxidized to DHA and IAA is oxidized to D-dehydroisoascorbic acid (DHIA), cf. structures in Scheme 1. e stoichiometric reactions in the present reaction systems can be expressed by 3.4. Acid Dissociation Constants of AA and IAA. Acid dissociation constants of AA were reported to be pK a1 � 3.96 [32] and pK a2 � 11.24 [27]   ionic strength. UV-Vis spectra were recorded for 0.10 mM IAA in a series of buffer solutions covering a pH range from 2.28 to 6.26; before recording each spectrum, a fresh IAA solution and the corresponding buffer solution as reference were thermoequilibrated at 25.0°C for about 10 min. e spectra recorded in pH 2.28, 3.89, and 6.26 buffers are given in the upper panel of Figure 6; an rough isosbestic point in the figure indicates two absorbing species are present. e wavelength of 266 nm was then chosen for the measurements of absorption values. e measured values as a function of pH are given in the lower panel of Figure 6 (data points). Equation (3) was utilized to simulate the data by use of a nonlinear least-squares method [29,30], where ε 1 and ε 2 pertain to the molar absorptivities of fully protonated and monoanionic forms of IAA, respectively.
e simulation confers a good fit (cf. the lower panel of Figure 6) and affords pK a1 � 3.74 ± 0.05 at 25.0°C and 1.0 M ionic strength; this value is virtually identical to that reported earlier albeit with a difference in ionic strength [19]. Since the pK a2 value was expected to be very close to that of AA and far from the pH region used in this work, we did not determine it.

Reaction Mechanism.
e observed second-order rate constants in Table 1 increased more than 100 times when pH was changed from 0.16 to 2.33 for both AA and IAA, indicating that their monoanionic forms have a much higher reactivity than the fully protonated forms. e increasing trends are anticipated to continue, but the reactions became too fast to  follow by the stopped-flow spectrometer when pH > 2. 33. In analogy, the fully deprotonated species of AA and IAA are anticipated to have the highest reactivity among their 3 protolysis species [17]. Since the attributes of the rapid scan spectra suggest that the electron transfer reactions between [IrCl 6 ] 2and AA/IAA undergo directly, it is logical to assume that all the protolysis species of AA/IAA in Scheme 2 will reduce [IrCl 6 ] 2-. A reaction mechanism is proposed as delineated in Scheme 2 which involves the three protolysis species of AA/IAA reacting with [IrCl 6 ] 2in parallel; the parallel reactions as indicated by k 1 -k 3 are the rate-determining steps. In each of the rate-determining steps, a single electron transfer takes place, generating a free radical species. us, three different free radicals as highly reactive transients are likely involved in the mechanism, cf. the possible structures of the free radicals in Scheme 2 [9,10,33]. Each of the free radicals is reacting with another [IrCl 6 ] 2in subsequently fast reaction leading to formation of DHA or DHIA.

Evaluation of the Rate Constants.
e rate law of equation (4) When equation (4) is compared with equation (1a), it confers the expression of k′: Equation (5) was used to simulate the k′-pH dependence data by use of a weighted nonlinear least-squares method [34]. In the simulation of the AA reaction, pK a1 � 3.96 and pK a2 � 11.24 were used as direct inputs and k 1 , k 2, and k 3 were treated as tunable parameters. e simulation resulted in a set of values (k 1 � 611 ± 67 M −1 s −1 , k 2 � 2.2 × 10 7 M −1 s −1 , and k 3 � 3.3 × 10 15 M −1 s −1 ); however, the value of k 3 is over the diffusion-controlled rate constant (ca. 10 10 M −1 s −1 in aqueous solution) [35]. Since the k 2 >> k 1 , the k 3 value is expected to be also several orders of magnitude higher than k 2 [17]. It is thus reasonable to assume that k 3 takes a diffusion-controlled rate constant [35]. Given an assumption that k 3 � 10 10 M −1 s −1 , the simulation was executed again and the simulated resulted is shown in the upper panel of Figure 4, concurrently providing k 1 � 451 ± 59 M −1 s −1 and When equation (6) was utilized to simulate the data, an identical fit and identical values of k 1 and k 2 were obtained (Table 2), indicating that even k 3 takes a diffusion-controlled rate constant; the contribution from ascorbate dianion in the pH region used in this work is negligible. is is understandable because the pK a2 value of AA is far from the pH region studied. Nevertheless, a well-defined k 2 value is obtained.
For the reaction of IAA, the above simulations were also performed and the simulated results were very similar. Consequently, the simulation of the k′-pH dependence data by equation (6) confers an excellent fit (the lower panel of Figure 4) by use of pK a1 � 3.74 determined in this work; the values of k 1 and k 2 acquired from the fit are listed in Table 2. For both AA and IAA, the relatively large errors associated with k 1 values are ascribed to the much higher values of k 2 , which dominates the total reactivity in the pH region studied.
e oxidation kinetics of ascorbic acid by [IrCl 6 ] 2was studied only in a narrow acidity region (0.20 M ≤ [HClO 4 ] ≤ 1.00 M); the derived rate constants were k 1 < 200 M −1 s −1 and k 2 � 1.4 × 10 7 M −1 s −1 at 20.0°C and 1.0 M ionic strength from this region (a correction of a stoichiometric factor was made) [25]. ese values were not well determined due to the very narrow acidity region used and are significantly smaller than those obtained in this work; our values derived from a much wider pH region are more accurate and of course more  reliable. Moreover, the k 3 value is estimated to be a rate constant of the diffusion-controlled one in aqueous solution.

Antioxidative Reactivity of the AA and IAA Species.
e acquired k 1 and k 2 values for IAA provide a direct comparison with AA. While the k 1 values are essentially identical within the experimental errors, the k 2 values have a clear difference, namely k 2 (AA) ≈ 2k 2 (IAA); this reactivity difference is ascribed probably to the slightly lower pK a1 value of IAA. e reactivity difference suggests that the antioxidative power of IAA is weaker than that of AA in the acidic to neutral media where the ascorbate monoanions dominate their respective populations (vide infra).
For both AA and IAA, the kinetic characters observed in the present work bolster the conclusion that the redox reactions take place via an outer sphere electron-transfer mode [25,26]. It is clearly shown from the data in Table 2 that k 1 << k 2 << k 3 . In a more visualized way, the species population versus pH distribution diagram and the species reactivity versus pH distribution diagram were contrarily constructed for the AA reaction [36], which are displayed in Figure 7 (the diagrams for the IAA reaction are very similar and are not shown). A few attributes are discerned from the diagrams: (a) the fully protonated AA exists between pH 0 and 6, but it only contributes slightly to the total reactivity in   an acidic region (pH from 0 to 2). (b) At the biological matrix of pH 7.0-7.4, the ascorbate monoanion dominates both the species population and the total reactivity. (c) e ascorbate dianion exists only in basic media (pH > 10), but it comes into play as early as pH 6.5 and takes a leading role when pH > 9. e reactivity of AA/IAA species toward reductions of single electron oxidants involved in our body may show similar characters observed above. e significance of the above analysis is that when AA and IAA are used as additives in foods and drinks, these foods and drinks will pass through our digestive systems, where stomach and intestines are in frontlines possessing an acidic environment (pH range from 2 to 5), and the monoanionic forms of AA and IAA may take a leading role in reductions of free radicals or single electron oxidants. Consequently, the antioxidative reactivity of IAA is only about half that of AA.

Conclusions
e reduction reactions of [IrCl 6 ] 2by AA and IAA have been analyzed kinetically and mechanistically by use of rapid scan and stopped-flow spectral techniques. e reactions strictly follow overall second-order kinetics. e proposed reaction mechanism involves three parallel reactions being the rate-determining steps; in each of the reactions [IrCl 6 ] 2acquires an electron via an outer-sphere electron transfer mode. Rate constants of the rate-determining steps have been derived or estimated which are confidently reliable. A direct comparison between reactivity of AA and IAA demonstrates that IAA has about a half antioxidative reactivity of that of AA in slightly acidic to neutral media. e antioxidative data acquired for AA and IAA can be used as a reference for comparisons with other biologically and biomedically important antioxidants at which we are studying. In addition, pK a1 � 3.74 ± 0.05 at 25.0°C and 1.0 M ionic strength for IAA has been determined in this work.

Data Availability
All the data supporting the results are included in the manuscript.

Conflicts of Interest
e authors declare that they have no conflicts of interest.