The bicarbonate/carbon dioxide pair increases hydrogen peroxide-mediated hyperoxidation of human peroxiredoxin 1

2-Cys peroxiredoxins (Prxs) rapidly reduce H2O2, thereby acting as antioxidants and also as sensors and transmitters of H2O2 signals in cells. Interestingly, eukaryotic 2-Cys Prxs lose their peroxidase activity at high H2O2 levels. Under these conditions, H2O2 oxidizes the sulfenic acid derivative of the Prx peroxidatic Cys (CPSOH) to the sulfinate (CPSO2−) and sulfonated (CPSO3−) forms, redirecting the CPSOH intermediate from the catalytic cycle to the hyperoxidation/inactivation pathway. The susceptibility of 2-Cys Prxs to hyperoxidation varies greatly and depends on structural features that affect the lifetime of the CPSOH intermediate. Among the human Prxs, Prx1 has an intermediate susceptibility to H2O2 and was selected here to investigate the effect of a physiological concentration of HCO3−/CO2 (25 mm) on its hyperoxidation. Immunoblotting and kinetic and MS/MS experiments revealed that HCO3−/CO2 increases Prx1 hyperoxidation and inactivation both in the presence of excess H2O2 and during enzymatic (NADPH/thioredoxin reductase/thioredoxin) and chemical (DTT) turnover. We hypothesized that the stimulating effect of HCO3−/CO2 was due to HCO4−, a peroxide present in equilibrated solutions of H2O2 and HCO3−/CO2. Indeed, additional experiments and calculations uncovered that HCO4− oxidizes CPSOH to CPSO2− with a second-order rate constant 2 orders of magnitude higher than that of H2O2 ((1.5 ± 0.1) × 105 and (2.9 ± 0.2) × 103 m−1·s−1, respectively) and that HCO4− is 250 times more efficient than H2O2 at inactivating 1% Prx1 per turnover. The fact that the biologically ubiquitous HCO3−/CO2 pair stimulates Prx1 hyperoxidation and inactivation bears relevance to Prx1 functions beyond its antioxidant activity.

Peroxiredoxins (Prx) 4 are a family of abundant Cys-dependent peroxidases that rapidly react with peroxides, constituting an important antioxidant defense and acting as sensors and transmitters of H 2 O 2 signals in cells (1)(2)(3)(4). There are six human Prxs (Prx1 to Prx6) that vary in their intracellular location and catalytic mechanisms (5). Prx1 to Prx4 are typical 2-Cys Prx enzymes that in the reduced state assemble into a decameric/ duodecameric toroid, which is the most active form of the enzyme (6,7). The peroxidase activity of 2-Cys Prxs depends on a fully-conserved Cys residue, which has a low pK a value, and it is known as the peroxidatic cysteine (C P SH). This residue rapidly reacts with peroxides reducing them while being oxidized to the sulfenic acid derivative (C P SOH). The minimal functional unit for 2-Cys Prxs is a homodimer, and during the catalytic cycle, a second Cys residue, the so-called resolving cysteine (C R ), reacts with C P SOH of the adjacent monomer forming a head-to-tail disulfide (C P -C R ). This Prx intermolecular disulfide is reduced by the thioredoxin-thioredoxin reductase-NADPH system (NADPH/TrxR/Trx), completing the peroxidase cycle (Fig. 1A). Because a partial unfolding of the ␣-helix containing C P SH is required for the formation of the disulfide, C P SOH can be further oxidized to the sulfinate (C P SO 2 Ϫ ) and, subsequently, to the sulfonated (C P SO 3 Ϫ ) forms if the oxidant concentration is sufficiently high (see Fig. 1A). These processes are called hyperoxidation and lead to Prx inactivation because C P SO 2 Ϫ and C P SO 3 Ϫ cannot be reduced by the NADPH/TrxR/Trx system or by other thiol reductants (1,2,8).
Prxs hyperoxidation and inactivation may constitute a pathway to additional functions and activities of these enzymes. For instance, hyperoxidized Prx1 and Prx2 switch the peroxidase activity to a high-molecular-weight holdase (9 -11) that can also recruit chaperones to protein aggregates formed on H 2 O 2 exposure (12). Also, a cellular deficit in peroxidase activity due to Prx hyperoxidation may potentiate signaling pathways that depend on H 2 O 2 -mediated oxidation of proteins containing Cys residues that are less reactive toward the oxidant than the Prxs C P residues (13)(14)(15). Additionally, hyperoxidized Prx may act as a cellular stress signal interrupted by the ATP-dependent reduction of CysSO 2 Ϫ back to CysSOH by sulfiredoxin (16 -18). Furthermore, hyperoxidized Prxs do not consume reduced Trx, preserving this oxidoreductase to other essential cellular activities (19 -21). All of these possible consequences of Prx hyperoxidation are currently under active investigation.
The susceptibility of 2-Cys Prxs to hyperoxidation and inactivation varies greatly and depends on structural features that dictate the C P SOH intermediate lifetime. In other words, a long-lived intermediate will have more time to react with a second hydroperoxide molecule before disulfide formation (Fig. 1A) (8,13,22). Typically, bacterial Prxs are more resistant to H 2 O 2 -mediated hyperoxidation than their mammalian counterparts (8,13). This enhanced susceptibility of the latter is attributed to the presence of Tyr-Phe (YF) and Gly-Gly-Leu-Gly (GGLG) motifs near the enzyme-active site. However, whereas human Prx1, Prx2, and Prx3 exhibit high sequence homology and contain the YF and GGLG motifs, each enzyme displays a different degree of susceptibility to H 2 O 2 (Prx2 Ͼ Prx1 Ͼ Ͼ Prx3) (23), and additional motifs underlying susceptibility to hyperoxidation have been reported (24).
Because of the reported intermediate susceptibility of Prx1 to H 2 O 2 -mediated hyperoxidation, the enzyme was selected here for investigating the possible effect of the HCO 3 Ϫ /CO 2 pair on its hyperoxidation. The HCO 3 Ϫ /CO 2 pair constitutes the main physiological buffer, and it is long known to accelerate the peroxidation of several biological targets, including biothiols (25,26). The biothiols investigated so far included GSH and albumin (27), papain and members of the protein-tyrosine phosphatase family (PTP1B and SHP-2) (28), and alkyl hydroperoxide reductase E (AhpE) (29). Each of these biothiols reacts with H 2 O 2 with second-order rate constant values ranging from 1.2 ϫ 10 0 to 2.1 ϫ 10 4 M Ϫ1 ⅐s Ϫ1 , which are considerably lower than those reported for the C P of Prx1 (0.38 -1.1 ϫ 10 8 M Ϫ1 ⅐s Ϫ1 ) (22,30). Although this extremely high second-order rate constant argues against HCO 3 Ϫ /CO 2 having an influence on the first step of the Prx1 catalytic cycle, the results from this study demonstrate that a physiological concentration of HCO 3 Ϫ /CO 2 (25 mM) stimulates H 2 O 2 -mediated Prx1 hyperoxidation and inactivation.

/CO 2 effect on Prx1 hyperoxidation mediated by H 2 O 2
To test whether HCO 3 Ϫ /CO 2 influenced Prx1 hyperoxidation by H 2 O 2 , we first performed direct experiments. Reduced Prx1 (2.5 M) was treated with different concentrations of H 2 O 2 (2.5-100 M) in the absence and presence of HCO 3 Ϫ /CO 2 (25 mM). The incubations were performed in phosphate buffer (50 mM) containing DTPA (0.1 mM), pH 7.4, for 5 min at room temperature. Addition of catalase interrupted the reactions and NEM addition alkylated the remaining thiols (see "Experimental procedures"). Next, the samples were submitted to nonreducing SDS-PAGE and either stained with Coomassie Blue or transferred to membranes for Western blot analysis, using an anti-PrxSO 2 H/PrxSO 3 H antibody (Fig. 1B, left and right 1B, right panel). Prx1-hyperoxidized dimers indicate that one C P is involved in intermolecular disulfide, whereas the other is hyperoxidized. In the presence of HCO 3 Ϫ /CO 2 (25 mM), Prx1 hyperoxidation was detectable at lower concentrations of H 2 O 2 and was always more pronounced than in the absence of HCO 3 Ϫ /CO 2 (Fig. 1B). In addition to Prx1 monomer and dimer bands, weak higher-molecular-weight Prx1 bands are evident in the protein gel (Fig. 1B, left panel). These bands correspond to oligomeric disulfides, which can be formed upon Prx1 oxidation because of its Cys 83 residue (23,31). The intensity of these bands increased with the employed H 2 O 2 concentrations, particularly in the presence of HCO 3 Ϫ /CO 2 , but these bands were hardly observed in the Western blotting at these H 2 O 2 concentrations. These experiments clearly show that HCO 3 Ϫ / CO 2 (25 mM) increases Prx1 hyperoxidation by small excesses of H 2 O 2 (4 -40 times) after a short incubation time (5 min) (Fig. 1B).

/CO 2 effect on Prx1 hyperoxidation and inactivation during enzymatic turnover
More biologically relevant was to examine whether the HCO 3 Ϫ /CO 2 pair stimulates Prx1 hyperoxidation and inactivation during its peroxidase activity sustained by NADPH/TrxR/ Trx (enzymatic turnover) (Fig. 1A). To this end, we performed ؊ /CO 2 on Prx1 hyperoxidation (B). A, 2-Cys Prx is represented by its minimal functional unit, which is a homodimer, with one of the peroxidatic (C P S Ϫ ) and one of the resolving Cys (C R SH) shown. The catalytic cycle starts with the rapid reaction of the peroxidatic cysteine with hydroperoxides to reduce them, while being oxidized to the sulfenic acid form (C P SOH). Next, the resolving Cys reacts with C P SOH of the adjacent monomer forming a head-to-tail disulfide (C P S-SC R ). This oxidized form of Prx is reduced by the NADPH/Trx/TrxR system, completing the peroxidase cycle. Because a partial unfolding of the ␣-helix containing C P is required for the formation of the disulfide, C P SOH can be further oxidized to the sulfinate (C P SO 2 Ϫ ) and, subsequently, to the sulfonated (C P SO 3 Ϫ ) forms if the oxidant concentration is sufficiently high. These processes are called hyperoxidation and lead to inactivation of the peroxidase activity of the enzyme. B, representative effects of HCO 3 Ϫ /CO 2 on H 2 O 2 -mediated Prx1 hyperoxidation revealed by nonreducing SDS-PAGE (lower panel) and Western blot analysis using an anti-PrxSO 2 H/PrxSO 3 H antibody (upper panel). The incubations contained Prx1 (2.5 M), H 2 O 2 (2.5-100 M) in the absence and presence of HCO 3 Ϫ /CO 2 (25 mM) in phosphate buffer (50 mM) containing DTPA (0.1 mM), pH 7.4. After a 5-min incubation at room temperature, the samples were treated and analyzed as described under "Experimental procedures." The shown experiment is representative of three independent experiments. Ϫ /CO 2 (25 mM), pH 7.0, at 30°C (Fig. 2, A-C) (32). The NADPH consumption rates varied with time, presenting a curvature that increased with the concentration of H 2 O 2 ( Fig. 2A), a behavior consistent with Prx1 inactivation during turnover (8,32). At all H 2 O 2 concentrations tested, the curvature increased in the presence of HCO 3 Ϫ /CO 2 (25 mM), but the initial rate remained unchanged (Fig. 2, B and C). These experiments indicated that HCO 3 Ϫ /CO 2 lowers the concentration of H 2 O 2 required to inactivate Prx1 during turnover (Fig. 2D) Control experiments showed that NADPH consumption was marginal in the absence of any of the enzymatic components (Prx1, Trx, and TrxR), confirming the usual low reactivity of the NADPH/TrxR/Trx system toward H 2 O 2 (33). Still, to exclude Trx or TrxR inactivation under the experimental conditions of Fig. 2, we ran parallel experiments monitoring the insulin reductase activity of the NADPH/TrxR/Trx system (34). Addition of NADPH to insulin/TrxR/Trx resulted in a time-dependent decrease in the absorbance at 340 nm due to NADPH consumption up to the time when insulin precipitation leads to a temporal absorbance increase (Fig. 3A, black curve). Such behavior was practically unaffected by the presence of either H 2 O 2 (0.5 mM) (Fig. 3A, blue curve) or HCO 3 Ϫ /CO 2 (25 mM) (Fig. 3A, gray curve). H 2 O 2 (0.5 mM) plus HCO 3 Ϫ /CO 2 (25 mM) ( Fig. 3A, red curve) presented a slight effect, particularly at long incubation times. Taken together, these control experiments (Fig. 3A) exclude a considerable influence of Trx or TrxR inactivation on the kinetic data obtained in the presence of Prx1 (Fig. 2). To prove that Prx1 inactivation during enzymatic turnover occurred in parallel with its hyperoxidation, the incubations of Fig. 2 with 0.5 or 1.0 mM H 2 O 2 in the presence and absence of HCO 3 Ϫ /CO 2 (25 mM) were analyzed by nonreducing SDS-PAGE and Western blotting using the anti-PrxSO 2 H/PrxSO 3 H antibody after 15 or 30 min of incubation. We performed several of these experiments, first only Western blottings because of the low Prx1 concentration in the mixtures (Fig. S1) but also Western blottings along with protein blots revealed by silver staining (Fig. 3B). All the experiments performed showed that Prx1 was hyperoxidized in a H 2 O 2 -dependent manner and that such hyperoxidation was always increased in the presence of HCO 3 Ϫ /CO 2 (25 mM), whereas the controls showed marginal or no Prx1 hyperoxidation. Fig. 3B displays a representative experiment with both blots, in which the low molecular weight Trx  Ϫ /CO 2 (25 mM), pH 7.0; the controls with HCO 3 Ϫ /CO 2 (25 mM) are also shown. After 15 min at 30°C, the samples were treated and analyzed as described under "Experimental procedures." The shown experiment is representative of several similar experiments, including some performed with an antibody from a different commercial source (Fig. S1).

HCO 3
؊ /CO 2 increases Prx1 hyperoxidation (11.2 kDa) migrated out of the gel (12%). Prx1 incubated with the buffer in the absence or the presence of HCO 3 Ϫ /CO 2 (25 mM) presented bands corresponding to Prx1 dimers, indicating oxidation during the incubation but no hyperoxidation (Fig. 3B, right and left panels, respectively). Prx1 incubated with the reducing system (NADPH/TrxR/Trx) in the absence or presence of HCO 3 Ϫ /CO 2 presented strong monomer and faint dimer bands that showed practically no hyperoxidation (Fig.  3B, left and right panels, respectively). In contrast, Prx1 incubated with H 2 O 2 (0.5 mM) and the reducing system presented a strong monomer band along with faint dimer and oligomeric disulfide bands, all of which were partially hyperoxidized (Fig.  3B, left and right panels, respectively). In the presence of HCO 3 Ϫ /CO 2 (25 mM), the intensity of all these hyperoxidized bands increased considerably (Fig. 3B). Taken together, kinetic ( Fig. 2) and Western blotting experiments (Figs. 3B and Fig. S1) show that HCO 3 Ϫ /CO 2 increases Prx1 hyperoxidation and consequent inactivation during turnover. Such a HCO 3 Ϫ /CO 2 effect was further confirmed by MS/MS experiments.

Nano-ESI-Q-TOF-MS/MS analysis of tryptic digests of Prx1 after peroxidase activity sustained by DTT
To simplify the MS/MS analyses, Prx1 turnover was sustained by DTT (8), rather than by the enzymatic reducing system (NADPH/TrxR/Trx). Prx1 (5 M) was incubated with DTT (10 mM) in the absence (control) or presence of H 2 O 2 (1 mM) or H 2 O 2 (1 mM) plus HCO 3 Ϫ /CO 2 (25 mM) in phosphate buffer (50 mM) containing DTPA (0.1 mM), pH 7.4, for 16 h at 30°C. Thereafter, unreacted thiols were blocked with NEM (25 mM) in a denaturing medium. Then, DTT (50 mM) was added to eliminate excess NEM (35). Following the removal of excess reagents, the samples were digested with trypsin and subjected to nano-ESI-Q-TOF-MS/MS analysis as described under "Experimental procedures" (36,37). The nano-ESI-Q-TOF-MS/MS analysis of the tryptic digests of all the Prx1 samples yielded sequence coverage of Ն 88% (data not shown). Full MS/MS analyses focused on data obtained for the Prx1 peptides containing Cys residues (Cys 52 (C P ), Cys 173 (C R ), Cys 71 , and Cys 83 ), which are the major H 2 O 2 targets (38 -40).
The first point to emphasize is that we obtained reasonable nano-ESI-Q-TOF-MS/MS spectra of most ions corresponding to peptides containing Cys residues and their modified forms, i.e. alkylated peptides (corresponding to the reduced -SH form), and hyperoxidized peptides (corresponding to both -SO 2 H and -SO 3 H forms). For instance, Fig. 4 shows representative nano-ESI-Q-TOF-MS/MS spectra of the ions with three charges corresponding to the peptide 38 (Table 1). Depending on the sample, ions from different peptides for each Cys residue and for each modification were characterized. As an example, the ions corresponding to the peptides containing the modified peroxidatic Prx1 residue (Cys 52 ) appeared not only as the three charged ions corresponding to 38 YVVFFFY-PLDFTFVCPTEIIAFSDR 62 (Fig. 4) but also, depending on the modification, as ions corresponding to miscleaved peptides spanning from residues 36 to 62 (Fig. S2), 38 to 68, and 36 to 67, with some ions also appearing with four charges ( Table 1).
The relative yields of all the identified Prx1 peptides in the performed incubations (control, H 2 O 2 -treated, and H 2 O 2 / HCO 3 Ϫ /CO 2 -treated) were calculated. To this end, the experimental procedures and the MS analyses were independently repeated three times. The intensity of the MS of each ion cor-

/CO 2 increases Prx1 hyperoxidation
responding to the modified peptide identified in each experiment was calculated based on the ratio between the extracted ion chromatogram of the modified peptide (XIC) and the total ion chromatogram (TIC) for each peptide. The extracted ion chromatogram of the modified peptide and the TIC/TIC value of each specific ion was averaged for the three independent experiments. The sum of the average XIC/TIC values for all the ions corresponding to each specific Cys residue (Cys 52 , Cys 173 , Cys 71 , and Cys 83 ) and to each specific modification (-NEM, -SO 2 H, or -SO 3 H) is displayed in Fig. 5, permitting comparison of the samples. This comparison showed that during DTTsustained Prx1 turnover, the alkylated C P residue was the only one that decreased in a statistically significant manner while being significantly hyperoxidized to the -SO 2 H and -SO 3 H forms in control (black bars), H 2 O 2 -treated (blue bars), and H 2 O 2 /HCO 3 Ϫ /CO 2 -treated samples (red bars) (Fig. 5). Relevantly, the decrease in alkylated C P and increase in hyperoxidized forms were more pronounced in the presence (red bar) than in absence (blue bar) of HCO 3 Ϫ /CO 2 (Fig. 5A). Although hyperoxidized forms of the C R residue showed statistically significant increases in hyperoxidation, significant changes were not detected for the alkylated form (Fig. 5B). The other Prx1 Cys residues (Cys 71 and Cys 83 ) did not present significant differences in the alkylated form, and hyperoxidized forms were not detectable (Fig. 5C). The above comparison was limited to ions of similar peptides because each peptide behaves differently in LC-MS/MS analyses (41,42). Although we added TIC/ XIC values of ions corresponding to the same modified peptide miscleaved or not, the basic peptide is the same and should behave similarly. Indeed, the intensity of all the ions corresponding to Prx1 C P peptides were ϳ100 times lower than those of the C R and Cys 71 and Cys 81 domains (Fig. 5). This is likely due to the higher hydrophobicity of the tryptic peptides of Prx1 C P domain (retention time around 66 min under our experimental conditions) as compared with those of C R domain (retention time around 25 min under our experimental conditions). Indeed, hydrophobicity is a major factor influencing detection of a peptide by LC-MS/MS (37,41,42). Overall, the LC-MS/MS results (Figs. 4 and 5; Table 1) were consistent with the kinetic (Fig. 2) and Western blotting results (Fig. 3B) in demonstrating that physiological concentrations of HCO 3 Ϫ / CO 2 increase Prx1 hyperoxidation and inactivation during turnover. 4 ؊ in the stimulating effect of HCO 3 ؊

/CO 2 on Prx1 hyperoxidation
The better investigated accelerating effects of HCO 3 Ϫ /CO 2 on H 2 O 2 -mediated oxidations have been attributed to peroxymonocarbonate (HCO 4 Ϫ ), a two-electron oxidant stronger than H 2 O 2 that is present in equilibrated solutions of H 2 O 2 and HCO 3 Ϫ /CO 2 in levels corresponding roughly to 1% of total H 2 O 2 concentration (Equation 1) ( Fig. S4; Table S1) (26,43,44) (see also "Discussion"). Likewise, the kinetic data and the calculations presented below indicate that HCO 4 Ϫ is primarily responsible for the stimulating effect of HCO 3 Ϫ /CO 2 on Prx1 hyperoxidation mediated by H 2 O 2 , Indeed, the kinetic data obtained from experiments monitoring Prx1 enzymatic turnover, in the presence and absence of HCO 3 Ϫ / CO 2 (Fig. 2, A-C), were used to calculate C 1% hyperoxidation , which is the concentration of H 2 O 2 that inactivates 1% of Prx1 per turnover (see "Experimental procedures") (8). The C 1% hyperoxidation values were found to be 335 Ϯ 18 and 101 Ϯ 4 M H 2 O 2 in the absence and presence of HCO 3 Ϫ /CO 2 (25 mM), respectively (Fig. 2D). Therefore, three times less H 2 O 2 is necessary to inactivate 1% Prx1 per turnover in the presence of physiological concentrations of the  (32). More recently, a similar value (50 M) was reported as a confirmation of the previous value, but the displayed data indicate a quite different value (Fig. S5) (24). Because the C 1% hyperoxidation value is not expected to vary with the reducing system (8), the use of Trx and TrxR from Saccharomyces cerevisiae in our experiments cannot account for the divergent values obtained here and previously, and neither altered the Trx nor the TrxR activities because they are maintained (Fig. 3A), particularly in the time frame of the kinetic experiments employed to calculate C 1% hyperoxidation (up to 60 s) (see "Experimental procedures"). Although we cannot explain these divergent values, the obtained value of 335 Ϯ 18 M H 2 O 2 is quite consistent with the data (Fig. 2, A-C) and constraints to determine C 1% hyperoxidation (8) but was considerably higher than that obtained under the same conditions in the presence of HCO 3 Ϫ /CO 2 (25 mM) (101 Ϯ 4) M) (Fig. 2D). The data in Fig. 2D were further explored to estimate the C 1% hyperoxidation for HCO 4 Ϫ . To this end, we calculated the concentration of HCO 4 Ϫ present in the incubations containing H 2 O 2 plus HCO 3 Ϫ /CO 2 using their corresponding concentrations and the equilibrium constant of Equation 1 (43). To each calculated HCO 4 Ϫ concentration, we obtained the f inactivation value by subtracting the corresponding f inactivation value obtained in the presence of HCO 3 Ϫ /CO 2 from the one obtained in its absence. The plot of the calculated values of f inactivation against the calculated HCO 4 Ϫ concentrations resulted in a straight line, from which the C 1% hyperoxidation for HCO 4 Ϫ was estimated to be 1.3 Ϯ 0.1 M (Fig. 6A). This value indicates that ϳ250 times less HCO 4 Ϫ is required to inactivate 1% Prx1 per turnover as compared with H 2 O 2 . Therefore, the quite small concentrations of HCO 4 Ϫ present in mixtures of H 2 O 2 and HCO 3 Ϫ /CO 2 (Fig. S4) are sufficient to increase Prx1 hyperoxidation considerably.
The above conclusion was confirmed by direct kinetic studies of Prx1 hyperoxidation mediated by H 2 O 2 in the presence and absence of HCO 3 Ϫ /CO 2 monitored by changes in the intrinsic fluorescence of the enzyme. Upon reaction of reduced Prx1 with H 2 O 2 , changes in the intrinsic fluorescence of the enzyme occur (22,30). An extremely rapid decrease in Prx1 fluorescence is followed by a considerably slower fluorescence increase (Fig. 6B, inset). The rapid phase is linearly dependent on H 2 O 2 concentration, corresponding to the oxidation of the Prx1 C P S Ϫ residue to C P SOH (Fig. 1A). The slow phase does not depend on oxidant concentration at low H 2 O 2 concentrations (Fig. 6B, inset) and corresponds to disulfide formation (Fig. 1A) (22,30). At higher H 2 O 2 concentrations (10 -80 times excess over Prx1), the slow phase became dependent on oxidant concentrations and was recently attributed to the oxidation of C P SOH to C P SO 2 Ϫ , i.e. to the first step of Prx1 hyperoxidation (Fig. 1A) (45). We concur with the latter conclusion because we performed similar experiments while examining the influence of HCO 3 Ϫ /CO 2 on Prx1 hyperoxidation and obtained similar data (Fig. 6, B and C). The reaction of reduced Prx1 (2.5 M) with H 2 O 2 (2.5 M) displays the usual profile (Fig. 6B, inset). Using a 10 times or higher H 2 O 2 excess over enzyme, the oxi-

/CO 2 increases Prx1 hyperoxidation
dation of C P S Ϫ to C P SOH is too rapid to be monitored, but a slow increase in Prx1 fluorescence is observed and is linearly dependent on oxidant concentration (Fig. 6B). The k obs value for each H 2 O 2 concentration was obtained and plotted against H 2 O 2 concentration (Fig. 6C). From the slope of the straight line, the second-order rate constant of the first step of Prx1 hyperoxidation was calculated to be 2.9 Ϯ 0.2 ϫ 10 3 M Ϫ1 ⅐s Ϫ1 , a value that is in good agreement with the previously reported value (1.8 ϫ 10 3 M Ϫ1 ⅐s Ϫ1 ) (45). Additionally, the y-intercept of the plot shown in Fig. 5C provided the first-order rate constant of disulfide formation (13.4 s Ϫ1 ), also in line with previously reported values (22,30). Moreover, experiments performed with a high fixed concentration of H 2 O 2 (0.5 mM) and variable concentrations of HCO 3 Ϫ /CO 2 (25-100 mM) showed a stimulating effect of the pair (data not shown). These experiments permitted the calculation of the second-order rate constant for the oxidation of C P SOH to C P SO 2 Ϫ by HCO 4 Ϫ (k ϭ (1.5 Ϯ 0.1) ϫ 10 5 M Ϫ1 ⅐s Ϫ1 ) (Fig. 6D). The value of the y-intercept in Fig. 6D (15.1 Ϯ 0.2 s Ϫ1 ) corresponds to the k obs value of the oxidation of Prx1 C P SOH by 0.5 mM H 2 O 2 (15.0 Ϯ 0.2 s Ϫ1 ) (Fig. 6C), attest-ing to the consistency of the overall data. These results show that HCO 4 Ϫ oxidizes Prx1 C P SOH to C P SO 2 Ϫ ϳ2 orders of magnitude faster than H 2 O 2 . Taken together, the results presented herein (Figs. 2 and 6) provide evidence for the involvement of HCO 4 Ϫ in the stimulating effect of HCO 3 Ϫ /CO 2 on Prx1 hyperoxidation mediated by H 2 O 2 .
In contrast, the oxidation of Prx1 C P by H 2 O 2 is extremely rapid (Fig. 6B) (k ϭ 0.38 -1.1 ϫ 10 8 M Ϫ1 ⅐s Ϫ1 ) (22,30), making it unlikely that an eventual effect of HCO 3 Ϫ /CO 2 could be quantified. To support this suggestion, we determined the secondorder rate of the reaction of Prx1 C P with H 2 O 2 in the absence and presence of HCO 3 Ϫ /CO 2 (25 mM) in phosphate buffer (50 mM) containing DTPA (0.1 mM), pH 7.0, by competition experiments with horseradish peroxidase (HRP) (46, 47). These experiments were performed in a plate reader and are considered preliminary. Nevertheless, the values obtained in the absence or presence of HCO 3 Ϫ /CO 2 were similar (Fig. S6) and in line with previously reported values at pH 7.4 (22,30).

Discussion
The in vivo concentrations of CO 2 in mammalian tissues are generally high due to its continuous metabolic formation. Approximately 1.3 mM CO 2 in equilibrium with HCO 3 Ϫ (25 and 14.4 mM in extra and intracellular media, respectively) constitutes the main physiological buffer (48). From the physiologically ubiquitous HCO 3 Ϫ /CO 2 pair, a central role for CO 2 in peroxynitrite-mediated oxidation of biological targets was firmly established (49). The roles of the HCO 3 Ϫ / CO 2 pair on other biologically relevant oxidations remain less investigated (25,26).
In this work, we studied the effect of HCO 3 Ϫ /CO 2 on H 2 O 2mediated Prx1 hyperoxidation and inactivation. Western blotting experiments showed that HCO 3 Ϫ /CO 2 (25 mM) increases Prx1 hyperoxidation promoted by the addition of a small excess of H 2 O 2 (up to 40 times) (Fig. 1B) Table 1). These experiments permitted direct detection of Prx1 hyperoxidized forms (-SO 2 H and/or -SO 3 H), complementing the kinetic and Western blotting analyses.
The MS/MS data showed minor modifications of Prx1 turned over by DTT in the absence of oxidants (Fig. 5, black  bars). The samples treated with oxidants showed a major hyperoxidation of the Prx1 C P residue, rendering both C P SO 2 H and C P SO 3 H (Fig. 5A, blue and red bars). Additionally, C P hyperoxidation was higher in the presence of H 2 O 2 plus HCO 3 Ϫ /CO 2 (Fig. 5A, red bars) than in the presence of only H 2 O 2 (Fig. 5A, blue bars) as attested by the significant decrease of alkylated C P and the significant increase of hyperoxidized C P forms. In all samples, Prx1 C R and Cys 71 and Cys 83 residues were either marginally or undetectably hyperoxidized and did not show significant differences in the alkylated Prx1 present in control or treated samples (Fig. 5). In the incubations containing H 2 O 2 plus HCO 3 Ϫ /CO 2 , the hyperoxidized C P SO 3 H form

/CO 2 increases Prx1 hyperoxidation
increased more than the C P SO 2 H form as compared with the control sample (3.1 and 7.7 times, respectively). The same trend was not observed in the incubations containing H 2 O 2 alone (blue bars), even considering the large deviation of the C P SO 3 H yield (Fig. 5A). Such a trend suggests that HCO 3 Ϫ /CO 2 may also increase the oxidation of C P SO 2 H to C P SO 3 H, but additional experiments are required to substantiate this suggestion because such oxidation may have occurred during the long sample workup. Nevertheless, the obtained MS/MS data confirmed that HCO 3 Ϫ /CO 2 increases Prx1 C P hyperoxidation. Based on previous studies (26), we hypothesized that the stimulating effect of HCO 3 Ϫ /CO 2 on Prx1 hyperoxidation and inactivation mediated by H 2 O 2 was most likely due to the presence of HCO 4 Ϫ in equilibrated solutions of H 2 O 2 and HCO 3 Ϫ / CO 2 (Fig. 7A, Fig. S3, and Table S1). At neutral pH, HCO 4 Ϫ formation mostly depends on the slow perhydration of CO 2 , whereas the addition of deprotonated H 2 O 2 (HOO Ϫ ) to CO 2 becomes increasingly important at higher pH values (Fig. 7A) (43). Relevantly, there is an important exception to the limited studies about the effects of HCO 3 Ϫ /CO 2 on biological oxidations. This is the case of H 2 O 2 -mediated oxidation of biothiols not particularly reactive toward H 2 O 2 (k ϭ 10 0 -10 4 M Ϫ1 ⅐s Ϫ1 ), such as BSA (27), PTP1B (28), and AhpE (29), among others. In each of these examples, a stimulatory effect of HCO 3 Ϫ /CO 2 on thiol oxidation was observed and attributed to HCO 4 Ϫ . The calculation of second-order rate constants of these thiol peroxidations in the presence and absence of HCO 3 Ϫ /CO 2 showed that HCO 4 Ϫ reacted with second-order rate constants 2-3 orders of magnitude higher than those of H 2 O 2 (Ref. 26 and references therein). These results are in agreement with the trend, based on S n 2 reactivity between thiolates and peroxides, that these reactions are expected to proceed faster when they produce a less basic leaving group, and CO 3 2Ϫ (pK a ϭ 10.3) is significantly less basic than HO Ϫ (pK a ϭ 15.7). Relevantly, the C P residues of typical 2Cys-Prxs, such as Prx1 and Prx2, do not follow this trend and react with H 2 O 2 with higher rate constants than with peroxynitrite, which has the most acidic leaving group (NO 2 Ϫ ) (pK a ϭ 3.2) among biological peroxides (29,40). Such a comparison supports the view that Prx1 and Prx2 are specialized in reducing and sensing H 2 O 2 rather than other peroxides (22,45,50).
To support the involvement of HCO 4 Ϫ in the stimulatory effect of HCO 3 Ϫ /CO 2 on H 2 O 2 -mediated Prx1 hyperoxidation, we compared HCO 4 Ϫ and H 2 O 2 with regard to C 1% hyperoxidation and second-order rate constant of the oxidation of C P SOH to C P SO 2 Ϫ (Figs. 2D and 6).  (Fig. 2D) (see "Results") and HCO 4 Ϫ (Fig. 6A) were determined as 335 Ϯ 18 and 1.30 Ϯ 0.1 M under our experimental conditions. This indicates that HCO 4 Ϫ is ϳ250 times more effective than H 2 O 2 in inactivating 1% Prx1 per turnover. This 2-3 orders of magnitude difference is similar to the difference observed in the second-order rate constants of biothiols of low reactivity reacting with H 2 O 2 or HCO 4 Ϫ (26), indicating that HCO 4 Ϫ is responsible for the stimulating effect of HCO 3 Ϫ /CO 2 on Prx1 hyperoxidation. This indication was further supported by determination of the second-order rate constant of the first step of Prx1 hyperoxidation (C P SOH to C P SO 2 Ϫ ) by monitoring changes in the intrinsic fluorescence of the enzyme (see "Results") (Fig. 6, B and D). The values found were (2.9 Ϯ 0.2) ϫ 10 3 and (1.5 Ϯ 0.1) ϫ 10 5 M Ϫ1 ⅐s Ϫ1 for H 2 O 2 (Fig. 6C) and HCO 4 Ϫ (Fig. 6D), respectively. Therefore, Prx1 C P SOH reacts with HCO 4 Ϫ with a second-order rate constant 2 orders of magnitude higher than with H 2 O 2 . Previous work determined the second-order rate constants of the reaction of the C P SOH of AhpE, a bacterial 1-Cys Prx, with H 2 O 2 and HCO 4 Ϫ as 4.0 ϫ 10 1 and 2.1 ϫ 10 3 M Ϫ1 ⅐s Ϫ1 , respectively, which represents a similar 2 orders of magnitude difference (29). Additionally, the second-order rate constant we determined for H 2 O 2 herein is in good agreement with a value recently reported (45). Taken together, these results (Fig. 6) strongly support the involvement of HCO 4 Ϫ in the stimulating effects of HCO 3 Ϫ /CO 2 on Prx1 hyperoxidation mediated by H 2 O 2 . In summary, our data show that a physiological concentration of HCO 3 Ϫ /CO 2 (25 mM) increases H 2 O 2 -mediated Prx1 hyperoxidation and inactivation of its peroxidase activity Ϫ formation from H 2 O 2 and HCO 3 Ϫ /CO 2 (A) and of the mechanism proposed for the stimulatory effect of HCO 3 Ϫ /CO 2 on H 2 O 2 -mediated Prx1 hyperoxidation and inactivation is shown. A, mechanism of HCO 4 Ϫ formation from H 2 O 2 /HOO Ϫ was from Richardson and co-workers (43); the rate constants of all proposed reactions (Table S1) were used to perform the computer simulation HCO 4 Ϫ formation from H 2 O 2 (0.5 mM) and HCO 3 Ϫ /CO 2 (25 mM) displayed in Fig. S4. B, stimulatory effect of HCO 3 Ϫ /CO 2 on H 2 O 2 -mediated Prx1 hyperoxidation is attributed to HCO 4 Ϫ . 2-Cys Prx is represented by its minimal functional unit, which is a homodimer, with one of the peroxidatic (C P S Ϫ ) and one of the resolving Cys (C R SH) shown. HCO 4 Ϫ coexists in H 2 O 2 solutions containing HCO 3 Ϫ /CO 2 (Fig. 6A) and oxidizes C P SOH to C P SO 2 Ϫ with a second-order rate constant 2 orders of magnitude higher than that of H 2 O 2 . HCO 4 Ϫ should not influence the first step of the Prx1 catalytic cycle, because the reduction of H 2 O 2 and concomitant oxidation of C P S Ϫ to C P SOH is extremely rapid. Prx1 hyperoxidation is an inactivation pathway that depends on specific Prx1 structural features that determine the lifetime of the C P SOH intermediate.

Because HCO 4
Ϫ is considerably more reactive toward the C P SOH intermediate than H 2 O 2 , it will redirect higher levels of C P SOH from the catalytic cycle to the hyperoxidation and inactivation pathway. Our data suggested but did not prove that HCO 4 Ϫ oxidizes C P SO 2 H to C P SO 3 H more rapidly than H 2 O 2 justifying the use of broken lines in the second step of the hyperoxidation pathway (see text).

/CO 2 increases Prx1 hyperoxidation
through the intermediacy of HCO 4 Ϫ (Fig. 7B). This peroxide is present at low levels (roughly 1% of total H 2 O 2 concentration) in equilibrated solutions of H 2 O 2 and HCO 3 Ϫ /CO 2 (Equation 1) (Fig. S4 and Fig. 7A) (26,43) but oxidizes C P SOH to C P SO 2 H with a second-order rate constant 2 orders of magnitude greater than that of H 2 O 2 (Fig. 6, C and D). It is unlikely that HCO 4 Ϫ influences and/or alters the first step of the Prx1 catalytic cycle, because H 2 O 2 reduction and concomitant C P S Ϫ to C P SOH oxidation is extremely rapid. Indeed, preliminary experiments did not show significant differences in the secondorder rate constant of this reaction determined in the absence or presence of HCO 3 Ϫ /CO 2 (Fig. S6). The Prx1 hyperoxidation and inactivation pathway depends on structural features of the enzyme that determine the lifetime of the C P SOH intermediate.

However, because HCO 4
Ϫ is considerably more reactive than H 2 O 2 toward the C P SOH intermediate, HCO 4 Ϫ will redirect higher levels of the intermediate from the catalytic cycle to the hyperoxidation and inactivation pathway. Whether HCO 4 Ϫ oxidizes C P SO 2 Ϫ to C P SO 3 Ϫ more rapidly than H 2 O 2 was suggested by our data but not proved (see above). However, this enhanced reactivity is likely due to the higher reactivity of HCO 4 Ϫ in S n 2 oxidations as compared with H 2 O 2 (26,40). The mechanism displayed in Fig. 7B illustrates a possible explanation for why the H 2 O 2 -mediated hyperoxidation of 2-Cys Prxs is apparently more efficient in cells than predicted from in vitro experiments (23). Indeed, cells are usually maintained in a 5% CO 2 atmosphere, although in vitro studies are rarely performed in HCO 3 Ϫ /CO 2 buffers. The participation of HCO 4 Ϫ in biological oxidations has been overlooked in the literature because of its slow formation at neutral pH and the low concentration it attains at equilibrium (Fig. 7A and Fig. S4). However, H 2 O 2 and CO 2 are continuously produced in cells and are the precursors of HCO 4 Ϫ . Therefore, at each cellular circumstance, the concentrations of H 2 O 2 , CO 2 , and HCO 4 Ϫ are likely to be at steady-state levels. In cases where HCO 4 Ϫ reacts with an important biological target with a second-order rate constant 2 to 3 orders of magnitude higher than that of H 2 O 2 , a cellular response is likely to ensue. A potential example is the increased Prx1 hyperoxidation and peroxidase activity inactivation intermediated by HCO 4 Ϫ as reported here. 5 Upon Prx1 hyperoxidation, the antioxidant and redox relay functions of the enzyme decline, but other actions may rise, such as the chaperone-like activity (9 -11), recruitment of chaperones toward protein aggregates (12), potentiation of redox signaling pathways mediated by Cys-based proteins that are poorly reactive toward H 2 O 2 (13,13,14), and maintenance of Trx-dependent activities (19 -21).

Materials
Unless stated otherwise, all chemicals were purchased from Sigma and Merck-Millipore (Burlington, MA) and were analytical grade or better. Trypsin Gold MS grade was purchased from Promega (Madison, WI). The H 2 O 2 solutions were prepared from stock immediately before use, and their concentrations were determined spectrophotometrically by reaction with HRP to produce compound I (⌬⑀ 403 ϭ 5.4 ϫ 10 4 M Ϫ1 cm Ϫ1 ) (47). All solutions and buffers were prepared with Milli-Q water (Millipore, Billerica, MA) and treated with Chelex-100 resin.

Expression and purification of recombinant proteins (Prx1, Trx, and TrxR)
Human Prx1 was cloned into pET-17b, expressed in Escherichia coli strain BL21(DE3), and purified as described previously (30). The fractions with high Prx1 purity were pooled and concentrated with an Amicon Ultra filter (10 kDa) (Millipore) with exchange to phosphate buffer, pH 7.4. Prx1 was aliquoted and stocked under argon. S. cerevisiae Trx1 was cloned into pET-17b, expressed in E. coli strain BL21, and purified as described previously (51) with the following modifications. To the supernatant of nucleic acid precipitation, DTT (100 mM final concentration) was added, and the sample was heated at 90°C for 15 min. Then, the sample was centrifuged at 15,000 ϫ g at 4°C for 20 min, and the supernatant was loaded into a Mono Q column (GE Healthcare) equilibrated with Tris-HCl (20 mM), pH 7.4. The proteins were eluted with a rising gradient of 0 -1 M NaCl. Trx1 eluted at ϳ20% NaCl. The collected fractions were analyzed by SDS-PAGE, and those with high Trx1 purity were pooled and concentrated on an Amicon filter (3 kDa) (Millipore) with buffer exchanged to phosphate buffer (20 mM), pH 7.4, and maintained in the refrigerator. His-tagged S. cerevisiae TrxR1 was cloned into pPROEX-1, expressed in E. coli DH5 strain, and purified as described previously (52). After chromatography, the fractions containing high-purity Trx1 were selected by SDS-PAGE, pooled together, and dialyzed against phosphate buffer (20 mM), pH 7.4. The dialyzed sample was concentrated on an Amicon filter (3 kDa) (Millipore) and maintained in the refrigerator. The concentration of the proteins was determined at 280 nm using the ⑀ values provided by the ProtParam tool (https://www.expasy.org); reduced Prx1 (⑀ 280 nm ϭ 18,450 M Ϫ1 cm Ϫ1 ), reduced Trx1 (⑀ 280 nm ϭ 9,970 M Ϫ1 cm Ϫ1 ), and oxidized TrxR1 (⑀ 280 nm ϭ 24,660 M Ϫ1 cm Ϫ1 ).

Prx1 thiol reduction and quantification
Immediately before most experiments, Prx1 was reduced with a 16 times excess of DTT under an argon atmosphere for 90 min at 37°C. Excess DTT and other low molecular weight products were removed by filtration on an Amicon filter (10 kDa) (Millipore) and washed with 20 mM phosphate buffer containing 0.1 mM DTPA, pH 7.4. Thiol quantification was performed spectrophotometrically after treatment of reduced Prx1 with excess of 4,4Ј-dithiodipyridine (0.5 mM) for 15 min (⑀ 324 ϭ 21,400 M Ϫ1 cm Ϫ1 ) (53). The thiol content from reduced Prx1 was typically (3.5 Ϯ 0.2) thiols/protein.

Incubation conditions
To study the effect of the HCO 3 Ϫ /CO 2 pair on Prx1 hyperoxidation mediated by H 2 O 2 by different methodologies, the incubation conditions were standardized to ensure reproducibility due to the sluggishness of HCO 3 Ϫ /CO 2 and HCO 3 Ϫ /H 2 O 2 equil-5 While this manuscript was under review, the manuscript "Bicarbonate is essential for protein-tyrosine phosphatase 1B (PTP1B) oxidation and cellular signaling through EGF-triggered phosphorylation cascades" was published on line (54).

HCO 3
؊ /CO 2 increases Prx1 hyperoxidation ibrations ( Fig. S3 and Table S1) (43). The employed buffers were prepared with and without HCO 3 Ϫ (50 mM), and the desired final pH was adjusted and the flasks maintained closed to minimize CO 2 loss. In the absence of HCO 3 Ϫ /CO 2 , the reactions were initiated by adding H 2 O 2 to the other reagents in the working buffer. Unless otherwise stated, the reactions performed in the presence of HCO 3 Ϫ /CO 2 were started by mixing equal volumes of twice the concentration of H 2 O 2 preincubated with 50 mM HCO 3 Ϫ /CO 2 for 15 min to twice the concentration of all the other reagents in the working buffer to give the desired final concentrations. The final concentration of HCO 3 Ϫ /CO 2 was 25 mM except in the stopped-flow experiments (see below).
In another series of experiments, Western blotting analyses were used to monitor Prx1 hyperoxidation after enzymatic turnover. Prx1 (0.6 M) decomposed H 2 O 2 (0.5 or 1.0 mM) in the presence of NADPH (0.2 mM), Trx1 (2.5 M), and TrxR1 (80 nM) (8,32). The incubations were performed Hepes-NaOH (50 mM), pH 7.0, containing 1 mM EDTA in the absence or presence of HCO 3 Ϫ /CO 2 (25 mM) for 30 min at 30°C. The reactions were interrupted by catalase addition (0.05 mg/ml). Remaining thiols were alkylated with NEM (30 mM), and the samples (260 ng of protein per lane) were submitted to nonreducing SDS-PAGE and Western blotting experiments as described above. Some of these experiments, including the one shown in Fig. 3B, were performed similarly but with the following modifications. Protein gel was stained with silver due to the low Prx1 content. After blocking the polyvinylidene difluoride membrane with 5% milk, they were incubated with an antibody to Prx-SO 2 H/ SO 3 (1:2500) (YIF-LF-PA0004 Abfrontier); next, the membrane was incubated with an anti-rabbit KPL IgG peroxidase antibody (1:10,000) (04-15-06) (Sera Care; Milford, MA). The immunoreactivity was detected by chemiluminescence, and the bands were visualized in a photodocumentation system (Uvitec Alli-ance 4.7) using the software Nine Alliance 9.7 (Uvitec Ltd., Cambridge, UK).

Insulin reductase activity of NADPH/TrxR/Trx
Insulin (0.16 mM) was incubated with NADPH (0.2 mM), Trx (2.5 M), and TrxR (80 nM) in Hepes-NaOH (50 mM) containing 1 mM EDTA, pH 7.0, at 30°C (34) in an Infinite M200 plate reader (Tecan, Männedorf, Switzerland). NADPH oxidation and insulin precipitation were followed at 340 nm with 10-s readings. Similar experiments were performed in the presence of H 2 O 2 (0.5 mM), HCO 3 Ϫ /CO 2 (25 mM), or H 2 O 2 (0.5 mM) plus HCO 3 Ϫ /CO 2 (25 mM) to examine the influence of each on the insulin reductase activity of the NADPH/TrxR/Trx system. Thereafter, the samples were treated with guanidinium chloride (6 M) and NEM (25 mM) in Tris-HCl (50 mM), pH 8.0, for 15 min at 37°C. Then, DTT (50 mM) was added in each sample to eliminate excess NEM (35). To remove all the low-molecularweight compounds, the samples were washed and concentrated with an Amicon Ultra filter (10 kDa) (Millipore) with exchange to NH 4 HCO 3 (50 mM). Protein contents of the samples were determined by the Bradford assay, and the samples were digested with trypsin Gold (protein/trypsin ratio of 50 g/1 g) for 12 h at 37°C. Then, a second aliquot of trypsin (protein/ trypsin ratio of 50 g/1 g) was added, and the samples were further incubated for 12 h at 37°C (36,37). The hydrolysates were dried in a speed vacuum and re-dissolved in Milli-Q water containing 0.1% formic acid, desalted, and concentrated with a ZipTipC18 (Millipore). The peptides were eluted with 85% acetonitrile, dried, and maintained at Ϫ80°C until analysis. On the day of analysis, the samples were resuspended in 50 l of water containing 0.1% formic acid and 2% acetonitrile and left under rotation (600 rpm) for 4 h at 4°C. Each sample was diluted 20 times, and 1 l was loaded onto an ACQUITY UPLC-C18 (20 mm ϫ 180 m; 5 m) nano-HPLC column (Waters) coupled to a Hybrid quadrupole-TOF-LC-MS/MS mass spectrometer (TOF-6600) with nanospray source (AB Sciex, Concord, ON, CA) (36,37). A chromatographic run of 85 min was employed with a 400-nl/min flow rate. The eluents were solvent A (water HCO 3 ؊ /CO 2 increases Prx1 hyperoxidation with 0.1% formic acid) and solvent B (acetonitrile with 0.1% formic acid). For chromatographic separation, the employed linear gradient was 99% A and 1% B up to 35% B for 60 min, 35-85% solvent B for 61 min, 85% B for 81 min, and returning to 99% A and 1% B for 85 min. The conditions for the nanospray source were as follows: capillary, 2.4 kV; dry heater and 100°C. The Analyst TF software (version 1.7.1) and Peak view software (version 2.2) (AB Sciex) were employed for data acquisition and processing. The MASCOT (version 2.5) (Matrix Science Ltd., London, UK) and Biotools (version 4.3) (Bruker Daltonics, Bremen, Germany) were employed for the analysis of the modifications and retention times. The mass tolerance in all experiments was Յ10 ppm for MS analysis and Յ0.05 Da for MSMS analysis; the false discovery rate was Յ1.0%. Peak sequencing was confirmed using the Peak view software (version 2.2).

/CO 2
The C 1% hyperoxidation values were determined as described by Nelson et al. (8). Briefly, the initial rate of NADPH oxidation at each peroxide concentration was calculated by fitting the first 10 s of NADPH consumption versus time curve to a straight line. Prx1 turnover was calculated by dividing the initial rate by the employed Prx1 concentration (0.6 M). The inactivation constant (k inactivation ), was calculated by fitting 60 s of NADPH consumption versus time curve to Equation 2, where y is the absorbance value; t is time, k is k inactivation (exponential decay); and b is the initial rate (linear decay); the a, b, and c parameters fluctuate during the fitting, but the final b value should be quite close to the previously calculated initial rate, y ϭ ae kt ϩ bt ϩ c (Eq. 2) The fitting provides k inactivation , from which the fraction of inactivation (f inactivation ) at each peroxide concentration is calculated from Equation 3, f inactivation ϭ k inactivation /turnover All the values obtained in this work fulfilled these requirements.

Prx1 hyperoxidation followed by the intrinsic fluorescence of the enzyme
The reaction of reduced Prx with H 2 O 2 has been followed by the changes in the intrinsic fluorescence of the enzyme and permitted the determination of the second-order rate constant of the oxidation of the peroxidatic Cys (C P ) of the enzyme by H 2 O 2 , the first-order rate constant of disulfide formation (22,30), and, more recently, the second-order rate constant of Prx1 C P SOH oxidation to Prx1 C P SO 2 Ϫ by H 2 O 2 (45). Here, these studies were confirmed and extended to determine the secondorder rate constant of Prx1 C P SOH oxidation to Prx1 C P SO 2 Ϫ by HCO 4 Ϫ . The experiments were performed on a stopped-flow spectrometer (mixing time Ͻ3 ms) (Applied Photophysics SX-18MV (Leatherhead, Surrey, UK) by following the intrinsic Prx1 (2.5 M) fluorescence increase ( excitation ϭ 295 nm; emission Ͼ320 nm) with time upon addition of high concentrations of H 2 O 2 (0.025-2 mM) or at a fixed H 2 O 2 concentration (0.5 mM) in the presence of variable concentrations of HCO 3 Ϫ / CO 2 (25 mM-200 mM). The reactions were performed in phosphate buffer (50 mM) containing 0.1 mM DTPA, pH 7.4, at 25°C. The k obs values were determined by fitting the traces to a single exponential equation using the OriginPro 8.0 software. At least three independent experiments for each substrate concentration were performed to calculate k obs . The apparent secondorder rate constant of Prx1 C P SOH oxidation to Prx1 C P SO 2 Ϫ by H 2 O 2 was determined from the slope of k obs values plotted against H 2 O 2 concentrations, using linear least-squares regression analysis. The apparent second-order rate constant of Prx1 C P SOH oxidation to Prx1 C P SO 2 Ϫ by HCO 4 Ϫ was determined from the slope of k obs obtained from the experiments performed with a fixed concentration of H 2 O 2 (0.5 mM) and variable concentrations of HCO 3 Ϫ /CO 2 (25-200 mM) plotted against the HCO 4 Ϫ concentrations calculated from Equation 1.

Determination of the second-order rate constant of the reaction between Prx1 C P SH and H 2 O 2
The second-order rate constant in the absence and presence of HCO 3 Ϫ /CO 2 (25 mM) was determined by competition experiments with HRP type VI (Sigma), as described previously (46,47).

Statistical analysis
When pertinent, the data are expressed as the mean Ϯ S.D. of values determined in at least three independent experiments. Statistical significance was calculated employing the one-way ANOVA and Bonferroni post-test using GraphPad 5.0 software (GraphPad Software, Inc.).